Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins
Geological Society Special Publications Se...
73 downloads
1556 Views
32MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins
Geological Society Special Publications Series Editor A . J .
FLEET
GEOLOGICAL SOCIETY SPECIAL PUBLICATION NO. 78
Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins EDITED
BY
JOHN PARNELL School of Geosciences The Queen's University of Belfast, UK
1994 Published by The Geological Society London
THE GEOLOGICAL SOCIETY
The Society was founded in 1807 as the Geological Society of London and is the oldest geological society in the world. It received its Royal Charter in 1825 for the purpose of 'investigating the mineral structure of the Earth'. The Society is Britain's national society for geology with a Membership of 7500 (1993). It has countrywide coverage and approximately 1000 members reside overseas. The Society is responsible for all aspects of the geological sciences including professional matters. The Society has its own publishing house which produces the Society's international journals, books and maps, and which acts as the European distributor for publications of the American Association of Petroleum Geologists and the Geological Society of America. Fellowship is open to those holding a recognized honours degree in geology or cognate subject and who have at least two years relevant postgraduate experience, or who have not less than six years relevant experience in geology or a cognate subject. A Fellow who has not less than five years relevant postgraduate experience in the practice of geology may apply for validation and subject to approval, may be able to use the designatory letters C. Geol (Chartered Geologist). Further information about the Society is available from the Membership Manager, The Geological Society, Burlington House, Piccadilly, London WlV 0JU, UK. Published by The Geological Society from: The Geological Society Publishing House Unit 7 Brassmill Enterprise Centre Brassmill Lane Bath BAI 3JN UK (Orders: Tel. 0225 445046 Fax 0225 442836) First published 1994 © The Geological Society 1994. All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No paragraph of this publication may be reproduced, copied or transmitted save with the provisions of the Copyright Licensing Agency, 90 Tottenham Court Road, London WlP 9HE, UK. Users registered with Copyright Clearance Center, 27 Congress Street, Salem, MA 01970, USA: the item-fee code for this publication is 0305-8719/94 $7.00. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN 1-897799-05-5 Typeset by Type Study, Scarborough Printed by Alden Press, Oxford, UK
Distributors USA AAPG Bookstore PO Box 979 Tulsa Oklahoma 74101-0979 USA (Orders: Tel. (918)584-2555 Fax (918)584-0469) Australia Australian Mineral Foundation 63 Conyngham Street Glenside South Australia 5065 Australia (Orders: Tel. (08)379-0444 Fax (08)379-4634) India Affiliated East-West Press PVT Ltd G-l/16 Ansari Road New Delhi 110 002 India (Orders: Tel. (1 !)327-9113 Fax (11)326-0538) Japan Kanda Book Trading Company Tanikawa Bldg 3-2 Kanda Surugadai Chiyoda-Ku Tokyo 101 Japan (Orders: Tel. (03) 3255-3497 Fax (03) 3255-3495)
Contents
Preface
vii
FYFE, W.S. The water inventory of earth: fluids and tectonics Large-scale fluid flow
VAN BALEN, R. & CLOETINGH, S. Tectonic control of the sedimentary record and stress-induced fluid flow: constraints from basin modelling
9
DEMING, D. Fluid flow and heat transport in the upper continental crust
27
JEssoP, A.M. & MAJOROWlCZ,J.A. Fluid flow and heat transfer in sedimentary basins
43
PmLLIpS, G.N., WILLIAMS, P.J. & DE JoNfi, G. The nature of metamorphic fluids and significance for metal exploration
55
Deformation and fluid flow
SmSON, R.H. Crustal stress, faulting and fluid flow
69
MUIR WOOD, R. Earthquakes, strain-cycling and the mobilization of fluids
85
KNIPE, R.J. & MCCAIG, A.M. Microstructural and microchemical consequences of fluid flow in deforming rocks
99
STEPHENSON,E.L., MALTMAN,A.J. & KNIPE,R.J. Fluid flow in actively deforming sediments:
ll3
'dynamic permeability' in accretionary prisms Fluid flow and reservoir evolution
BJORLYKKE,K. Fluid-flow processes and diagenesis in sedimentary basins
127
RINGROSE, P.S. t~ CORBET/',P.W.M. Controls on two-phase fluid flow in heterogeneous
141
sandstones Fluid chemistry; metal-organic interactions
HANOR, J.S. Origin of saline fluids in sedimentary basins
151
GIORDANO,T.H. & KHARAKA,Y.K. Organic ligand distribution and speciation in sedimentary basin brines, diagenetic fluids and related ore solutions
175
FILBY, R.H. Origin and nature of trace element species in crude oils, bitumens and kerogens: implications for correlation and other geochemical studies
203
NICHOLSON, K. Fluid chemistry and hydrological regimes in geothermal systems: a possible link between gold-depositing and hydrocarbon-bearing aqueous systems
221
Fluid evolution: migration and precipitation of hydrocarbons and metals
MANN, U. An integrated approach to the study of primary petroleum migration
233
SIMONEIT,B.R.T. Organic matter alteration and fluid migration in hydrothermal systems
261
PARNELL, J. Hydrocarbons and other fluids: paragenesis, interactions and exploration potential inferred from petrographic studies
275
vi
CONTENTS
FOWLER, A.D. The role of geopressure zones in the formation of hydrothermal Pb-Zn Mississippi Valley type mineralization in sedimentary basins
293
METCALFE, R., ROCHELLE, C.A., SAVAGE,D. • HIGGO, J.W. Fluid-rock interactions during continental red bed diagenesis: implications for theoretical models of mineralization in sedimentary basins
301
Tracers of fluid evolution
DUDDY, I.R., GREEN, P.F., BRAY, R.J. & HEGARTY,K.A. Recognition of the thermal effects of fluid flow in sedimentary basins
325
BALLENTINE,C.J. & O'NIONS,R.K. The use of natural He, Ne and Ar isotopes to study
347
hydrocarbon-related fluid provenance, migration and mass balance in sedimentary basins
Geofluids: introduction JOHN PARNELL
School of Geosciences, Queen's University of Belfast, Belfast B T71NN, UK
Geological fluids are a central theme linking the petrography and chemistry of all rock types, deformation processes on the microscopic to the continental scale, and the concentration of economic resources. The fundamental importance of fluid migration and evolution to rock composition and structure is reflected in a growing interest in fluid processes, including a series of successful conferences on water-rock interaction (Kharaka & Maest 1992). The papers in this volume are intended to give a state-of-theart review of the whole spectrum of geofluids research. In an introductory account, Fyfe summarizes the water inventory of the planet Earth, and emphasizes the importance of quantifying water fluxes at all levels within the crust, and in particular the fluxes resulting from subduction and continental collision. The relationship between fluid migration and heat flow helps to explain why fluid fluxes are of fundamental importance to the formation of mineral resources.
Large-scale fluid flow In the past decade several models have been proposed for large-scale fluid flow at continental margins and across continental interiors. Van Balen & Cloetingh describe the constraints imposed by basin modelling upon theories for tectonic control of the sedimentary record and stress-induced fluid flow. They use a dynamic numerical model to investigate the effect of short-term variations in the level of intraplate stresses on fluid flow and sedimentation patterns. Increases in stress strongly influence the hydrodynamic regime during the post-rift phase of basins by causing an increase in meteoric water influx and compactional flow. Deming shows that the magnitude of convective heat transport through the upper continental crust is exponentially dependent upon fluid velocity and depth of fluid circulation. The major mechanisms for fluid flow in the upper crust are topography, fluid released by sediment compaction, phase changes or metamorphism, and free convection. The most effective mechanism for transporting heat is topographically-driven flow,
which may be responsible for some ore-forming processes in the North American mid-continent. Jessop & Majorowicz emphasize that heat transport through fluid flow is as effective as conduction, leading to wide contrasts in both lateral and vertical heat flow in some basins. This applies not just to basins above sea level with an obvious topographic driving force, but also to some sub-sea basins. Phillips et al. review the nature of metamorphic fluids, and their role in ore formation. Metamorphic fluids in many gold provinces are dominated by water, carbon dioxide and hydrogen sulphide, with low salinity, reflecting an abundance of mica, carbonate and sulphide in the rocks. A less common metamorphic fluid involving evaporite-bearing sequences is saline, with the potential to transport both gold and base metals.
Deformation and fluid flow Several contributions examine the interrelationships between deformation and fluid flow, in which fluids help to enable deformation, and faulting helps to facilitate fluid migration. Sibson discusses crustal stress, faulting and fluid flow. Deviatoric stress exerts both static and dynamic effects on rock permeability and fluid flow, modulating flow systems in the Earth's crust. Textural evidence from hydrothermal veins suggests that fluid flow in fault-related fracture systems generally occurs episodically, and that stress cycling effects may be widespread. Muir Wood shows how empirical observations of hydrological changes following major earthquakes allow the prediction of subsurface fluid flow during active tectonism. The type of change is dependent on the style of fault displacement: normal faults displace large volumes of fluid from the crust, while reverse faults draw fluids into the crust. Knipe & McCaig review the interactions between deformation and fluid flow. They show that the various deformation mechanisms possible in rocks have different effects on fluid flow which depend upon the associated volume changes. Microstructural analysis of deformed rocks provides information on fluid flow pathways, fluid chemistry and the amount of fluid involved. Stephenson et al.
viii
PREFACE
discuss the importance of inter-related fluid flow and deformation during the evolution of accretionary prisms. With the aid of experimental data, they show that the permeability measured in an actively deforming material contains a dynamic component in addition to the classical notion of capacity to transmit fluid.
Fluid flow and reservoir evolution Inevitably some of the most detailed studies of fluid flow in the past decade have been those related to the evolution of hydrocarbon reservoirs. Bj0rlykke relates the flow of fluids through basins to transport of heat and dissolved ions, and consequent diagenetic reactions. The greatest potential for transporting mass and creating secondary porosity is through meteoric water flow, as the flow rate may be very substantially greater than typical compaction-driven flow. The flow behaviour of immiscible fluids in permeable sandstones is assessed by Ringrose & Corbett, who show that capillary forces result in significant amounts of both trapping and bypassing of the non-wetting phase. In a typical water-wet oil/water system the amount of trapped oil varies between 38% and 65% depending upon the patterns of rock heterogeneity.
Fluid chemistry; metal-organic interactions Research on the chemistry of groundwaters and hydrothermal fluids in sedimentary basins has highlighted the role of organic species in complexing with organic species. Hanor reviews the origins of saline brines in sedimentary basins. Thermodynamic buffering by silicatecarbonate-(halide) mineral assemblages is a first-order control on subsurface fluid compositions. Where fluid composition is rockbuffered its ultimate origin may be obscured by its most recent history; however some nonbuffered components, such as chlorine and bromine, can be useful in providing information on the original end-member fluid compositions. Giordano & Kharaka discuss the diagenetic processes involving dissolved acids. The dissolved acids are important as a control on pH and buffer capacity, as organic ligands to form aqueous complexes with metals and other inorganic species, as reducing agents controlling the Eh of fluids, and through breakdown as a source of carbon dioxide and hydrocarbon species. Filby describes the origin and nature of trace element species in crude oils, bitumens and
kerogens. Nickel and vanadium metalloporphyrins are formed during sedimentation/early diagenesis of oil source rocks, and the relative abundances of the metals are related to depositional environment. Complexes of other trace elements in crude oils may be primary, including products of mineral-kerogen reactions, or secondary, from interactions between oils with mineral matter or formation waters during migration, maturation or biodegradation. Nicholson describes compositions of geothermal fluids which range from gold-depositing dilute waters to saline, oilfield brines, and proposes that techniques used to study active geothermal systems may be applicable to both gold exploration and hydrocarbon reservoir modelling.
Fluid evolution: migration and precipitation of hydrocarbons and metals Fluids have a fundamental role as the agents of migration and concentration of hydrocarbons and metals. Mann presents an approach to the study of primary petroleum migration, integrating sedimentological, petrophysical, organic geochemical and numerical modelling methods. Primary migration probably proceeds through diffusion into pore/fracture systems where a petroleum bulk phase develops, possibly with aqueous solutions. Simoneit describes the alteration of sedimentary organic matter to petroleum hydrocarbons by reductive reactions in modern hydrothermal systems. This alteration occurs under high pressure, over a wide temperature range, and in a very brief geological time. Petroleum generation, expulsion and migration occurs as a single continuous process during hydrothermal activity. Parnell shows how paragenetic relationships between hydrocarbons and inorganic minerals provide information on the relative timing of hydrocarbon migration and the migration of other fluids. Co-migration of hydrocarbons and aqueous fluids is evinced by the occurrence of large quantities of authigenic silicate minerals within some hydrocarbon residues. Fowler describes the fluid pathways and driving mechanisms for lead-zinc mineralizing brines, and considers the possible role of overpressuring in their formation. Fluid flow must be fast to advect the heat needed in mineralizing fluids from deep within a basin. In shale-dominated basins, overpressured zones immediately below platform carbonate rocks can be a proximal source of heat and metals. The shales act as thermal barriers, and high fluid pressure ruptures the overlying rocks to provide vertical pathways for hot mineralizing brines
PREFACE into carbonate host rocks. Metcalfe et al. describe the chemistry of fluid-rock interactions during continental red bed diagenesis. Models for fluid evolution in this environment can be useful in understanding how the relatively high content of heavy metals in ferromagnesian/ aluminosilicate detritus can be concentrated into red bed-hosted ore deposits.
Tracers of fluid evolution Several highly specialized analytical techniques have evolved which help to trace the pathways and consequences of fluid flow. Duddy et al. explain how apatite fission track analysis can be used to determine palaeotemperature histories. T e m p e r a t u r e ~ l e p t h profiles can be used to distinguish the effects of flow of hot fluid through a basin from heating due to the simple conduction of basal heat flow. Ballentine & O'Nions show how the relative abundance of the rare gases helium, neon and argon in crustal, mantle and atmosphere-derived components of fluids can be distinguished according to their isotopic distribution. This data provides information on the physical processes experienced in the fluid and, combined with mass balance calculations, can be used to constrain fluid provenance and transport.
ix
The Geofluids '93 conference, from which this volume developed, was an initiative in collaboration with Steve Lawrence (Quad Consulting) and Chris Cornford (IGI Ltd), and included a special session on Deformation and Fluid Flow organized by Rob Knipe (University of Leeds). The organizers were ably supported by a committee including Sally Cornford, Bert Kennedy, Richard Bray, Graham Harman, Janet James, Les Oldham, Des Horscroft and Dave Naylor, and also the editorial committee (editor, Aiastair Ruffeii and Norman Moles) and the staff of Quad Consulting and IGI Ltd. The conference was supported by The Geological Society of London, The Institution of Mining and Metallurgy and The Institute of Petroleum. Support from these bodies reflects the wide significance of fluids research and the potential for interchange of approaches between the hydrocarbon and minerals industries. This volume should help to foster a greater appreciation of how geofluids research can contribute to the understanding of the evolution of sedimentary basins and their resource potential.
Reference KHARAKA, Y.F. & MAEST, A.S. 1992. Water-Rock Interaction (2 vols). Balkema, Rotterdam.
The water inventory of the Earth: fluids and tectonics W.S. FYFE
Department of Earth Sciences, University of Western Ontario, London, Ontario, Canada N6A 5B7 Abstract: As the human population continues to expand, to approach ten billion, there will be an increasing demand upon all Earth's resources. Of particular importance will be resources related to energy, soil and water, all of which involve geofluids. Problems associated with waste disposal also involve systems for protecting near-surface water quality. There is an urgent need to quantify the fluid inventory and fluid dynamics of the near surface, while understanding volatiles and their deep recycling is fundamental to our understanding of the major dynamic processes of the Earth.
Over the past century, the growth of human population (now almost 100 million per year), supported by diverse and complex technologies, and increasing expectations of quality of life, have placed vast new demands on Earth resources, which come mainly from the atmosphere, hydrosphere, soil and the top few kilometres of the solid crust. At this time, we are concerned with the need for careful management of our fossil carbon fuel resources from which we derive most of our energy and our usable water resources, which, in many nations, are approaching limits on the local scale. There is little doubt that oil-gas (and coal) will eventually be replaced by other energy sources, such as solar, geothermal and biomass sources. However a limiting resource for human occupation of Earth may well be water. There is an urgent need for the best possible inventories of all such resources, and a growing need to integrate the knowledge from all involved with the study of geofluids. It was, perhaps, with the development of observations from space that we acquired a new sense of the limits and fragility of our wet planet. Studies of crust and hydrosphere history have also revealed that, unlike our nearest planetary neighbours, for about 4 billion years of recorded history, the Earth has always had a massive hydrosphere which has never boiled or totally frozen (Broecker 1985). Earth has some remarkable environmental buffer systems, which we still do not adequately understand. We also know that micro-organisms have been present for essentially all of our recorded history, and new studies (Schopf 1992) suggest that photosynthetic organisms may have been present much earlier than has been suggested by other workers. The outer surface and the top, porouspermeable layers of Earth are wet, and the
recent search for micro-organisms at depth (stimulated by consideration of deep biocorrosion of nuclear waste systems), indicates that rocks may well host life at all places where temperatures do not exceed 100°C (Pederson 1993). Studies of surface heat-flow patterns (for example, the spectacular new results from the German deep drilling, KTB) clearly show that, in the outer layers of Earth, fluids transport a substantial amount of heat out of the crust. The work of Straus & Schubert (1977) showed that the adiabatic gradient for fluid convection in a porous medium is almost always exceed in crustal situations. Studies of fluid convection through the oceanic crust, oceanic heat flow studies and the famous hot and cold discharge systems have shown that the sea floor crust is water-cooled, with something like half the thermal energy removed by fluid convection (Fyfe & Lonsdale 1981; Boul6gne & Pflumio 1992). Another growing world problem, which has focused our attention on deep and near surface fluids and their motions, is that of disposal of wastes of all types (urban, chemical, agricultural, n u c l e a r . . . ) , and the scale of the problems which will be associated with a human population of 10 billion is vast. In most cases, the available technologies are not adequate.
The inventory In general, we have only a moderate knowledge of the global inventory of water-rich fluids. Most accessible, the oceans and ground water are the most massive (Berner & Berner 1987). Of much smaller mass is water in the ice caps (which we must leave alone!) and surface waters in rivers and lakes. There is a massive quantity of water
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 1-7.
2
W.S. FYFE
(similar to the ocean mass), contained in the hydrous mineral phases of the crust. While the inventory of volatiles fixed in the crust (H20, CO2, N compounds, halogen compounds) are often not well-quantified, we know that change in the P - T conditions of their environment may lead to their evolution or fixation, processes which often lead to chemical transport related to formation of mineral resources. The rock mechanics of the crust would be very different on a planet without water (Fyfe et al. 1978). However, below the accessible crust, the inventory of volatiles in the deep Earth, mantle and core, is very uncertain. Thompson (1992) has recently reviewed the situation of water in the mantle. There are many ways in which water and other volatiles may be held in the mantle. Thus Fyfe (1970) discusses the substitution of SiO44- by ( 0 H ) 4 4- at high pressures and possible tetrahedral CO44-, Pawley et al. (1993) discuss hydrogen in stishovite and Schrauder & Navon (1993) report solid carbon dioxide in diamond. But, while there is evidence that the core must contain some elements of low atomic number, the candidates (O, H, C, etc.), are less certain. Anderson (1993) has recently reviewed the problem of helium isotopes in deep crustal systems. He proposed that recycling from the surface from solar inputs may be more important than deep primordial sources (see also Fyfe 1987).
The global water cycle Most discussions of the global water cycle are restricted to consideration of atmospherebiosphere-hydrosphere interactions within the top few kilometres of the planet. Certainly, in terms of usable water for humans, these systems dominate our fluid resources. The global state of water resources has recently been reviewed by Postal (1992). But, if we broaden our interests to all geochemical fluxes related to fluid motions and hydrosphere evolution, and particularly the flux of bio-essential elements (Mn, Fe, Co, Cu, Ni, Zn, etc.) into the oceans, it is necessary to consider much deeper interactions (see Gillet 1993). If we wish to quantify the total evolution of the hydrosphere and all volatile systems, over geologic time, the deep systems must be considered. All phenomena which involve mantle convection, and the magmatic-tectonicmetamorphic expressions of such convection (plate tectonics) result in volatile transport.
Rising magma convection cells We are increasingly aware that the nature of the surface of our planet is largely controlled by
heat-mass transfer in the mantle, even down to the core-mantle boundary, and by erosional processes involving surface atmospherehydrosphere-biosphere interactions. The major sites of rising mantle convection cells are the great ocean ridge systems, where new basaltic crust is formed. It is also recognized that the same processes, but certainly with different spacing and higher intensity, operated in the ancient crust (see Kroner & Lager 1992; Fyfe 1974). The water cooling processes which transfer almost half the energy from the ridge systems were first modeled by Lister (1977) and Davis & Lister (1977), whose general ideas have been well confirmed by a host of later direct observations. When magmas cool, they contract, and zones of high porosity and permeability must be common. Boul6gne & Pflumio (1992) report a bulk porosity of 15-20% in ocean crust, and a mean hydrothermal flux from the ridges of about 40 km 3 a -~ and, for the overall ocean floor flux, 148kin 3 a -~. Given that the ocean mass is 1.4 x 1021kg, this implies recycling of the entire ocean mass through the sea floor systems in about 10 million years; a massive exchange process. The discharged fluids (hot or warm) are enriched in many species (CO2, CH4, HzS, H2, SiO2, Li, Mn, Fe, Cu, Zn, etc.), and this process makes a major contribution to the geochemistry of ocean water, sediments, and to the nutrient fluxes for the marine biomass. In addition, many of the great sulphide ore deposits with Cu, Zn (Ag, Au) are related to such processes over geological time. A process of great significance to the global water inventory is the hydration of the seafloor crust that must accompany the cooling process. Altered seafloor basalts are highly hydrated, the limiting case being the transformation of peridotites to serpentinites (see MacDonald & Fyfe 1985). While the ridge processes are intense, slow convective cooling occurs over the entire ocean floor regions, and the early ideas of Hess (Fowler 1990) on massive serpentinization appear confirmed. The total mass of water contained in the ocean crust is similar to that in the ocean and, in addition, large quantities of CO2, ammonium and halogen compounds must be present, but this inventory is not well known (Fyfe & Lonsdale 1981). The hydrothermal processes associated with hot spot phenomena, and with the great continental flood basalts are less well studied. Of the present Earth's surface, almost 7% has been influenced by recent hot spot events (Fyfe 1992a). When the mass of some flood basalt events is considered, it even seems possible that
WATER INVENTORY OF THE EARTH cooling processes could perturb atmospheric oxygen (Fyfe 1990).
Subduction and fluids: subduction-induced mantle convection There is very little ocean floor ophiolitic crust older than 200 million years. Ocean crust is subducted at a rate almost equal to its rate of formation, but the materials subducted are very different from those which form the ridges. The original basaltic crust and gabbro-peridotite basement has been changed to a spilitic crust (with H20, CO2, S, U, etc.) and the gabbroperidotite basement has been partially converted to amphibolite-serpentinite before subduction (Fyfe 1992b). In the last few years, we have also come to recognize that pelagic sediments may be subducted on a scale of cubic kilometres per year. This concept, originally proposed by Gilluly (1971), once strongly opposed by those involved with isotope systematics, has now become respectable because of the direct observation of trenches and the structure of lithosphere near trenches. Thus, Hilde & Uyeda (1983) and Uyeda (1983) and the more recent Kaiko Project (Lallemand et al. 1986; Le Pichon 1986) have clearly shown that, when the lithosphere bends, it cracks in the upper part and forms horst and graben structures which fill with sediments. If there is not enough sediment to fill the structure, tectonic erosion of the overplate occurs: Japan is being tectonically eroded and underthrusted. Such studies clearly show that initially light materials which are tectonically trapped may move towards the mantle (Lallemand & Malavieille 1992). It is of note that there is little evidence for sediment subduction in the case of the Northern Cascadia subduction zone (Davis & Hyndman 1989). Consideration of volatiles shows that, for species like H20 and CO2, the recycling of the major reservoirs occurs with time constants of the order of a billion years, at the present rate of subduction. The Earth has a hydrosphere, so that return flow must be moderately efficient. But the present processes, and their scales, should warn us that any purely steady-state model of ocean volumes and other volatile reservoirs may be inadequate. Thermodynamics tells us that heating bodies degass, while cooling bodies adsorb gases. Thus if a cooling Earth continues to convect, the ocean volume might be expected to diminish. Is this what has happened on Mars (Carr 1987)? Processes involved in the return flow of volatiles are complex. At the initial stages of
3
thrusting, pore fluids are literally squeezed out and pass up the thrust structures, reducing friction on the thrusts (Anderson 1981). Exotic fauna, originally characteristic of ridge vents, have now been found in deep trenches, but the fluids are cold, and would not transport large quantities of silica or metals. Recently, the Canadian Lithoprobe Project (Yorath et al. 1985) has been studying the structure of the subduction of the Juan de Fuca plate beneath Vancouver Island. There are complex fault structures above the subducting slab. Seismic studies have revealed the complexity of the structures beneath the continental edge. The very young and active faults provide evidence for fluid flow transporting hydrocarbons, Mn-Fe-Ag, and related species. Recently, ODP Leg 110 Scientific Party (1987) reported gases including methane being produced, presumably by the reprocessing of subducted organic debris. Lewis et al. (1988) have described the thermal influences of slab dewatering. Once the preliminary compression stage has passed, metamorphic processes will dominate, eventually leading to the formation of eclogites from basalts; kyanite and garnet-bearing rocks from pelagic sediments, and even kyanitecoesite-pyrope rocks (Chopin 1984). At this stage, fluids may hydrofracture their way to the surface along faults. Some fluid may be carried to very great depths in minerals like phlogopite, a natural product of the metamorphism of K-bearing spilite or pelagic sediments in an ultramafic mantle environment. When deep degassing occurs, with hotter mantle above, a new chain of events must occur. Water injected into this hotter mantle (water which will essentially be a soup of SiOz-alkalis and trace metals) will soften the mantle, and lead to convection and plume formation. As plumes of contaminated mantle rise, they will melt to produce the contaminated basalts we call andesites. Thus, mantle convective motions are induced by the rising fluids. The small amount of fluids are amplified into a much larger heat flow process, which eventually forms the mountain chains of the volcanic areas of Andean type. In a general way, every gram of fluid introduced will probably lead to something like 100-1000 times the mass of volcanic rock. The injected fluids have led to a process which drains energy and mass out of the overlying mantle wedge. Studies of heat flow and electrical conductivity across subduction zones show the scale of this energy transfer process, catalysed by subduction. Almost one third of the continental crust of
4
W.S. FYFE
the Americas has been influenced by recent subduction events. But we should not forget that it is the volatile loading in the sea floor environment that has led to this process. It would not occur on a dry planet. Once mantle plume processes start above a subduction zone, an array of fluid-mass-energy transfer processes occur (see Rice 1985; Fyfe 1987): (i) basaltic andesite rises, extrudes, intrudes and underplates continental crust; (ii) the crust melts, producing granitic plutons and acid volcanics; (iii) the basal crust undergoes progressive metamorphism; (iv) magmas mix, and complex hybrids are produced near the Moho region; (v) ultra-high-temperature gases are injected into the base of the crust from the andesite magmas and assimilated dense crustal components, which founder in the underplate magmas; (vi) high-level plutons and volcanics are watercooled by deep groundwaters in the high heat flow near-surface environments; (vii) high topography is created, and deep ground water circulation through fractured intrusives and porous volcanics must follow. The total fluid fluxes which must result from the entire array are impressive. We are not at a stage to quantify such fluxes, but we can make some order of magnitude calculations. For example, if we consider the western Americas, where the plutonic-volcanic terrains extend for about 20000km with a width of 500km, and assume that a 5 km thickness of crust has been 'granitized' or melted, the total acid igneous mass is about 5 x 107km 3. Given a 108 year cycle, pluton production is 0.5 km3a -~. Andesite production is about 2km3a -~ (Thorpe 1982). Thus, about 2.5 km 3 of magma can be watercooled per year. The fluid fluxes will thus be about 20% of that of ocean ridges. In this case, there is no question that the fluxes of fluids, just as the volcanic eruptions, will not be steady state, as plutons rise and Tamboras erupt. There will be large spatial variation of fluid flux, fluctuations which may influence the global temperature for years, or even cause mass extinctions (Officer et al. 1987; Grove 1988). It is the knowledge of these events, their intensity and frequency distribution which are needed for the IGBP. Recently, we have become interested in the problem of dewatering of slabs, and the possibility of fluidized bed injection associated with such processes. Le Pichon et al. (oral comm.
1990) showed recent observations of the large (30km 3) warm mud volcanoes being extruded from the Barbados accretionary complex. Barriga et al. (1992) have described a number of tectonic regions where such processes may occur, and A. Ribeiro has named the process 'eduction', where fluidized beds are injected from mantle depths. One of the common consequences of thrusting is the rapid burial of fluid-rich rocks, including water-saturated sediments and metamorphic rocks with variable water contents. Physico-chemical processes that range from compaction to prograde metamorphism will tend to produce less hydrated rocks plus water. Hydraulic fracturing is a common mechanism when impermeable rocks cap fluidrich zones, and fast and concentrated fluid flow may generate vein mineral deposits (Fyfe & Kerrich 1985). Tectonically-induced fluid generation, at all lithospheric levels, can lead to rock fluidization and injection. These processes are well known by sedimentologists and neotectonics specialists, as they are responsible for intrusive sediments and seismically-induced mud and sand volcanoes. This process is suggested to explain some of the blueschisteclogite-serpentine associations of California and even, perhaps, the coesite rocks of the Italian Alps (see Barriga et al. 1992; Maekawa et al. 1993).
Continental collisions and associated strikeslip faults At the present time, most subduction processes involve the underthrusting of oceanic crust in situations near continental margins. But, periodically, a continent is moved into the zone of subduction. Molnar & Gray (1979) considered the problem of the possible subduction of such a continental edge, and concluded that it was indeed possible in terms of density relations. Over the past 50 Ma major collision events have occurred, and are still proceeding at a rate of 5 cm a ~ in the Himalayas. While slightly different models of detail occur, there is no question that India is currently being thrust under Asia (Allegre et al. 1984; Barazangi & Ni 1982). In this region, a section of crust, 1500 x 3000 kin, has been doubled in thickness. Present seismic results provide evidence for a jagged Moho at 60-80km depth with 10km steps. The region appears to be highly electrically conductive (Pham et al. 1986), with active zones of conductive fluids or melts. The area of thickened crust is similar in area to about 60% of continental Australia, and a section of crust of
WATER INVENTORY OF THE EARTH Australian size has been reworked in the process. In Fyfe (1986) some aspects of the problem of fluid fluxes when crust on this scale is heated and compressed due to the thrusting and shortening processes is considered. Essentially, the underthrust rocks will be dehydrated, lose CO2 and other volatiles, and eventually melt. The young plutons of Himalayas show such geochemical features (e.g. extreme initial 87Sr/86Sr etc.) as would be expected. The quantities of metamorphic water which may be expelled up fracture zones and thrust planes may be similar to the mass of the present day ice caps. If thick carbonate sequences are involved, CO2 release could perturb the atmosphere and its greenhouse effect. A particularly interesting situation can occur if large salt basins are involved in which even ocean salinity could be influenced (we tend to forget the frequency and scale of large continental salt deposits). In a collision event of this scale, a vast perturbation in the volatile element flux, from H20 to salt, hydrocarbons and even elements such as Hg and As, etc., is likely. It is also likely that the present rise of the 87Sr/S6Sr ratio of the oceans is related to enhanced weathering of evolved continental crust at high elevations (Burke et al. 1982). It is clear that a collisional event of this magnitude will change the global environment, by altering global wind and ocean current patterns. But there may be even more subtle and fluctuating influences on global ocean chemistry. Such processes are not quantitatively understood, but the record of change must exist in ocean sediments. On a smaller scale, similar phenomena may occur on regional strike-slip faults, when these form plate boundaries. Thus, the present models of structure on the great Alpine Fault of New Zealand shows major regions of over-ride with thick crust (Allis 1981). The strike-slip process requires fluids for lubrication, and the override processes will produce the necessary fluids. As Oliver (1986) has suggested, when major thrusts occur, there must be a huge 'squeegie' effect ahead of all overthickening tectonics. As the load moves forward (e.g. Tibet over India), a vast front of fluid expulsion must advance before the thrusts. Fluids would be typical pore fluids, zeolite facies fluids and salt and hydrocarbons, if appropriate rocks are present. Thus, there could be fluid pulses of highly different geochemistry as a thrust develops. For a thrust front of Himalayan scale (3000 km), moving at 10cma x, the fluid expulsion rate could attain 0.5 km 3 a ~. If this fluid is rich in hydrocarbons or salt, local influences could be dramatic. It is
5
also unlikely that the thrust motions would be steady state. It is interesting to note that the crustal granites formed during these thickening processes may also lead to the formation of large deep aquifers. Thus, the rising plutons and dykes described by Le Fort et al. (1987) and Thakur (1987) may act as excellent aquifers once they cool, crystallize, and contract and thus direct fluid flow. We have described these phenomena in Saudi Arabia, where acid dykes coming off plutons are almost totally converted to epidote by flooding with descending meteoric fluids (Marzouki et al. 1979). Plutons will cause local fracturing as is well shown by Drummond et al. (1988) in their description of granodiorites being converted to tonalites by salt-water flooding.
The impact of high elevations on fluid flows Subduction and collision processes produce extensive belts of high elevation on our planet. These are normally associated with vast fault and thrust structures, and discontinuities associated with magma activity, plutonism and dyke emplacement. Deep gravity driven flow in such systems is now becoming well documented (e.g. Nesbitt & Muehlenbachs 1989; Rye & Bradbury 1988). The scale of this giant hydrogeological process, which may float frontal thrusts, can obviously be significant (Fyfe 1986). High elevations also drive rapid erosion and ocean sedimentation. As Milliman & Meade (1983) estimate, present continental erosion moves about 1.7 x 1016g a-x. Given the mass of continental crust of 1.6 x 1025g, erosion could reprocess all crust in a billion years. A metre of rain per year could dissolve the Himalayas to sea level in 100 million years! Very thick (5-10km) accumulation of sediment are common near many of the great delta systems of the world (Burk & Drake 1974). For example, such sediment piles are common around the entire margin of Brazil, the Amazon delta regions, and the great Bengal Fan system. For example, what happens when ocean floor crust, with a thick deep serpentine layer, is loaded with over 10 km of sediment? At thermal equilibration, the base of the now 20 km section could reach 600°C. The peridotites will deserpentinize, and hot, highly reduced fluids will rise, transporting a range of metals. Faulting in such regions is ubiquitous. The chemistry of systems where reduced hot fluids interact with organic-rich cover sediments along fault systems must be highly anomalous. Hot seeps along such sediment sections are common (Grassle 1985).
6
W.S. FYFE
There is a great need for submersible observations in these systems. There is an obvious final question: does such loading and metamorphism of the oceanic basement provide the guide for future sites of subduction?
Concluding statement As understanding environmental change, and providing human resources, become great challenges for modern Earth science, the necessity to quantify the total global water inventory, and the inventory of other volatiles, is essential. W e must better understand the surface biD-nutrient cycles, and the impact of water use on climate changes. For example, we need models of climate impact when all rivers are d a m m e d and evaporation and evapotranspiration replaces runoff. Water, water quality, and life are inextricably connected. A n intriguing question is whether the mass of the oceans has been constant over geological time? If water is being subducted in large quantities, and if the planet is cooling, will the mantle slowly absorb this volatile? Or, as Frank (1990) has suggested (a not popular view), do we receive constant additions of water via small comets. It is time we visited Mars, and studied a planet where Gaia has failed. I would like to dedicate this brief review to Geoffrey Brown of the Open University, who lost his life while in Columbia on a volcano watch. Geoff, a wonderful student and colleague, worked with me in Manchester, to do some of the classic work on the melting of dry metamorphic rocks.
References ALLEGRE, J.C., COURTILLOT, V. 8£ 33 others. 1984. Structure and evolution of the Himalayan-Tibet orogenic belt. Nature, 307, 17-22. ALLIS, R.G. 1981. Continental Underthrusting beneath the Southern Alps of New Zealand. Geology, 9,303-307. ANDERSON, D.L. 1993. Helium -3 from the mantle: primordial signal or cosmic dust? Science, 261, 170-176. ANDERSON, R.N. 1981. Surprises from Glomar Challenger. Nature, 293,261-262. BARAZANGI,M. & NI, J. 1982. Velocities and propagation of Pn and Sn beneath the Himalayan arc and Tibetan plateau: Possible evidence of underthrusting of India continental lithosphere beneath Tibet. Geology, 10, 179-185. BARRIGA, F.J., FYFE, W.S., LANDEFELD, L.A., MUNHA, J. & RraEmo, A. 1992. Mantle education: tectonic fluidization at depth. Earth Science Reviews, 32, 123-130. BERNER, E.K. & BERNER, R.A. 1987. The Global Water Cycle. Prentice Hall inc. New Jersey.
BOULOGNE,J. & PFLUMIO,C. 1992. Global chemical fluxes and fluid flow through the seafloor. In K.J. Hsu and J. THIEDE, (eds) Use and misuse of the seafloor. John Wiley and Sons, New York, 305-318. BROECKER, W.S. 1985. How to Build a Habitable Planet, Eldigio Press, New York. BURr, C.A. & DRAKE, C.L. 1974. The geology of continental margins. Springer-Verlag. BURKE,W.H., DENISON,R.E., HETHERINGTON,F.A., KOEPNICK,R.B., NELSON,H.F. & OTa, J.B. 1982. Variation of sea water 87Sr/86Srthroughout Phaerozoic time. Geology, 10, 516-519. CARR,M.H. 1987. Water on Mars. Nature, 326, 30-35. CHOPIN, C. 1984. Coesite and pure pyrope in highgrade blueschists of the Western Alps: a first record and some consequences. Contributions to Mineralogy and Petrology, 86, 107-118. DAVIS, E.E. & HYNDMAN,R.D. 1989. Accretion and recent deformation of sediments along the Northern Cascadia subduction zone. Geological Society of America Bulletin, 101, 1465-1480. DAVIS, E.E. & LISTER, C.R.B. 1977. Heat flow measurements over the Juan de Fuca Ridge: evidence for widespread hydrothermal circulation in a highly heat transportive crust. Journal of Geophysical Research, 82, 4845--4860. DRUMMOND,M.S., RAGLAND,P.C. & WESLOWSKI,D. 1986. An example of trondhjemite genesis by means of alkali metasomatism: Rockford granite, Alabama Appalachians. Contributions to Mineralogy and Petrology, 93, 98-113. FOWLER,C.M.R. 1990. The Solid Earth. Cambridge University Press, Cambridge. FRANK,L.A. 1990. The big splash. Avon Books, New York. FYFE, W.S. 1970. Lattice energies, phase transformations and volatiles in the Earth's mantle. Physics Earth and Planetary Interiors, 3, 196-200. 1974. Archaean tectonics. Nature, 249, 338-399. 1986. Fluids in deep continental crust. In BARAZANGI, M. & BROWN,L. (eds) Reflection Seismology: The Continental Crust. American Geophysics Union, Geodynamics series, 14, 33-39. 1987. The fluid inventory of the crust and its influence on crustal dynamics. In FRITz, P. & FRAPE, S.K. (eds) Saline waters and gases in crystalline rocks. Geological Association of Canada, Special Paper, 33, 1-3. 1990. Geosphere forcing: plate tectonics and the biosphere. Palaeogeography, PalaeoclimatoIogy, Palaeoecology, 89, 185-181. 1992a. Magma underplating of continental crust. Journal of Volcanology and Geothermal Research, 50, 33-40. 1992b. Geosphere Interactions On A Convecting Planet: Mixing and Separation. In HUTZIN~ER,O. (ed.) The Handbook of Environmental Chemistry, 1, part F. Springer Verlag Berlin Heidelberg, 1-26. & KERRICR, R. 1985. Fluids and Thrusting. Chemical Geology, 49,353-362. & LONSDALE,P. 1981. Ocean floor hydrothermal
WATER INVENTORY OF THE EARTH activity, in EMILIAN, C. (ed.), The Oceanic Lithosphere, John Wiley and Sons, 589-638. , PRICE, N.J. & THOMPSON, A.B. 1978. Fluids in the Earth's Crust. Elsevier, Amsterdam. GILLET, P. 1993. L'eau du manteau terrestre. La Recherche, 255,676-685. GILLULY, J. 1971. Plate tectonics and magmatic evolution. Geological Society of America Bulletin, 82, 2387-2396. GRASSLE, J.F. 1985. Hydrothermal vent animals: Distribution and biology. Science, 229, 85-400. GROVE, J.M. 1988. The Little Ice Age. Methuen, London. HILDE, T.W.C. • UYEDA,S. 1983. Convergence and subduction. Tectonophysics, 99, 85-400. KRONER,A. & LAYER,P.W. 1992. Crust formation and plate motion in the early Archean. Nature, 256, 1405-1411. LALLEMAND, S. t~ MALAVIEILLE,J. 1992. L'erosion profondes des continents. La Recherche, 23, 1388--1397. LALLEMANT, S., LALLEMAND,S., JOVIET, L. & HUCHON,P., 1986. Kaiko: l'exploration des losses du Japan. La Recherche, 17, 1344-1357. LE FORT, P., CUNEY, M., DENIEL, C., FRANCELANORD, C., SHEPPARD,S.M.F., UPRETI, B.N. & VIDAL, P., 1987. Crustal generation of the Himalayan leucogranites. Tectonophysics, 134, 39-57. LE PICHON,X. 1986. Kaiko, voyage aus extremites de la met. Editions Odile Jacob, Paris. LEWIS, B.T.R., BENTOWSKI, W.H., DAVIS, E.E., HYNDMAN,R.D., SOUTHER,J.G. & WRmHT, J.A. 1988. Subduction of the Juan de Fuca plate: thermal consequences. Journal of Geophysical Research, 93, 15,207-225. LISTER, C.R.B. 1977. Qualitative models of spreading center processes, including hydrothermal penetration. Tectonophysics, 37,203-218. MACDONALD, A.H. & FYFE, W.S. 1985. Rates of serpentization in seafloor environments. Tectonophysics, 116,123-132. MAEKAWA, H., SHOZUI, M., ISHII, T., FRYER, P. & PEARCH, J.A. 1993. Blueschist metamorphism in an active subduction zone. Nature, 364,520-523. MARZOUKI, F., FYFE, W.S. & KERRICH, R. 1979. Epidotization of diorites at A1 Hadah, Saudi Arabia: fluid influx into a cooling pluton. Contri-
butions to Mineralogy and Petrology, 68,281-284. MILLIMAN, J.D. & MEADE, R.H. 1983. World-wide delivery of river sediments to the oceans. Journal of Geology, 91, 1-21. MOLNAR, P. & GRAY, D. 1988. Subduction of continental lithosphere: some constraints and uncertainties. Geology, 7, 58-63. NESBITr, B.E. & MUEHLENBACHS, K. 1989. Origins and movement of fluids during deformation and
7
metamorphism in the Canadian Cordillera. Science, 245,733-736. ODP LEG I IO SOEN~FIC PARTY, 1987. Expulsion of fluids from depth along a subduction-zone decollement horizon. Nature, 326,785-788. OFFICER, C.B., HALLAM,A., DRAKE,C.L. & DEVINE, J.D. 1987. Late cretaceous and parysmal Cretaceous/Tertiary extinctions. Nature, 326, 143149. OLIVER, J. 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-104. PAWLEY, A.R., MCMILLAN, P.F. & HOLLOWAY,J.R. 1993. Hydrogen in stishovite, with implications for mantle water content. Science, 261, 10241026. PEDERSEN,K. 1993. The deep subterranean biosphere. Earth Science Reviews, 34,243-260. PHAM, V.N., BOYER,D., THEOME,P., YAN, X.C., LI, L. & JIN, G.V. 1986. Partial melting zones in southern Tibet from magnetotelluric results. Nature, 319,310-312. POSTAL, S. 1992. Last oasis-facing water scarcity. W.W. Norton and Co. New York. RICE, A. 1985. The mechanism of the Mt. St. Helens eruption and speculations regarding Soret effects in planetary dynamics. Geophysical Surveys, 7, 303-384. RYE, D.M & BRADBURY,H.J. 1988. Fluid flow in the crust: an example from a pyrenean thrust ramp. American Journal of Science, 288,197-235. ScnovF, J.W. 1992. Microfossils of the Early Archean Apex chart: new evidence of the antiquity of life. Science, 260,640-646. SCHRAUDER, M. & NAVON, O. 1993. Solid carbon dioxide in a natural diamond. Nature, 365, 42--44. STRAUS, J.M. & SCHUBERT,G. 1977. Thermal convection of water in a porous medium: effect of temperature- and pressure-dependent thermodynamic and transport properties. Journal of Geophysical Research, 82, 325-333. THAKUR, V.C. 1987. Plate tectonic interpretation of the Western Himalayas. Tectonophysics, 134, 91-102. THOMPSON, A.B. 1992. Water in the Earth's upper mantle. Nature, 358,295-300. THORPE, R.S. 1982. Andesites. John Wiley and Sons. UYEDA, S. 1983. Comparative subductology. Episodes, 1983, 19-24. YORATH, C.J., GREEN, A.G., CLOSES,R.M., SUTHERLAND GROWN, A., BRANDON, M.T., KANASEWlCH, E.R., HYNDMAN, R.D. & SPENCER, C. 1985. Lithoprobe southern Vancouver Island: seismic reflection sees through Wrangellia to the Juan de Fuca plate. Geology, 13,759-762.
Tectonic control of the sedimentary record and stress-induced fluid flow: constraints from basin modelling R. V A N B A L E N & S. C L O E T I N G H Tectonics~Structural Geology Group, Vrije Universiteit, De Boelelaan 1085, 1081 H V A m s t e r d a m , The Netherlands
Abstract: Many basins show deviations from the simple subsidence pattern predicted by the
stretching model for basin evolution. Incorporating finite strength of the lithosphere during rifting and stresses acting on the lithosphere during the post-rift phase of basin evolution can successfully explain these deviations. Finite strength of the lithosphere during rifting causes flexurally supported rift-shoulders. The erosion of these rift-shoulders induces both additional uplift at the basin margin and sediment loading-induced subsidence in the basin centre, causing tilting of syn-rift sediments. A 'break-up' unconformity results from the progressive diminishing of this tilting towards the end of the rifting period, separating the syn- and post-rift sediments. As the amount of erosion is primarily controlled by climatic conditions, basins having the same tectonic history but, for example, located at different lattitudes will not have developed the same pattern of stratigraphic fill. The same effect is expected if another mechanism causes the uplift of the basin margin, as, for example, a change in the level of intraplate stress. Plate reorganizations are the main cause for changes in the level of intraplate stress. We demonstrate that the in-plane stress variations affect the fluid flow regime in rifted basins, with possible implications for the diagenesis of sediments, primary migration of hydrocarbons, faulting and localization of economic resources. Examples include the North Sea and Pannonian Basin. An increase in the level of compressive stress causes flank uplift and basin centre subsidence, inducing a contemporaneous increase of meteroic water influx and compaction-driven fluid overpressures. An increase in the level of tensile inplane stress induces the opposite effects.
Over the last few years considerable progress has been made in quantifying the effects of forces originating from plate boundaries on differential uplift and subsidence and the stratigraphic record in rifted sedimentary basins (Cloetingh et al. 1985, 1989; CIoetingh & Kooi 1992). In fact, in many cases it appears to be very difficult to discriminate between eustatic and stress induced on- and offlap patterns in the stratigraphy (Kooi & Cloetingh 1992; Reemst et al. 1994). The influence of intraplate stresses is not restricted to rifted basins; stresses also affect the stratigraphy of foreland basins in a way comparable to the effect of eustatic sealevel changes (e.g. Peper et al. 1992, 1994; Posamantier & Allen 1993). The effect of inplane stress variations on the compaction-driven fluid flow system in rifted basins can be considerable, as shown by numerical modelling by Van Balen & Cloetingh (1993). Fluid overpressures and flow velocities are altered by as much as 30%. This can account for important phenomena as hydrofracturing and episodic dewatering with associated thermal, diagenetic and migration processes. Observations of present-day stress fields by
earthquake focal mechanism studies, in situ stress measurements and break-out analyses (Zoback 1992; Miiller et al. 1992) shows that a large number of basins formed in an extensional regime, such as the North Sea (Kooi & Cloetingh 1989) and the Pannonian Basin (Horvfith 1993), are in a situation of present-day compression. The most popular model for extensional basin formation, the McKenzie stretching model, has been extensively applied to the North Sea and Pannonian Basin (Sclater & Christie 1980; Sclater et al. 1980; Royden & D6venyi 1988). The main reason for this has been the observed thinning of crustal and sub-crustal lithosphere under these basins, compatible with the predictions of the stretching model. The stretching model also predicts a decay of subsidence during the post-rift phase of extensional basin evolution. This is, however, not consistent with recent observations of anomalous subsidence during the Plio-Quaternary times in the North Sea Central Graben and Pannonian Basin, showing the need for a new generation of basin formation models incorporating constraints on lithosphere stress and rheology (Cloetingh et al. 1993).
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 9-26.
10
R. VAN BALEN & S. CLOETINGH -20"
30'
S.-.l'q~',m.~Ic fP,oc'b.~ /
Su-br~kouts
m,m~,OIF O w O ~ Ovom* I n g n
E~lr~rd S~lm ~ ~mm~na s~m ~ ~Ike ~ ~ Rqmw unwo~ sblm ~
Da~a
0 t28 252 378 mmlm~:=lil=.m~
N
i.N~m" c:~r.N, iif~Emlm¢
~i~
Fig. 1. The stress map for Europe compiled by the World Stress Map team (modified after Miiller et al. 1992). The map shows the directions of the maximum horizontal stress. Clearly shown are the consistent NW-SE orientation in the North Sea Basin area, and the interference of the NW-SE and NE-SW (the Dinaride trend) directions in the Pannonian Basin area. The stress field for Europe (see Fig. 1) shows two main directions for the maximum horizontal stress: a N W - S E orientation in northwestern Europe and a N E - S W orientation in southeastern Europe. The first stems from the ridge push in the Atlantic ocean, the second from the
subduction processes in the Aegean area (MOiler et al. 1992). Recent investigations have revealed the basinwide existence of a giant normal fault network in the Palaeogene successions of the North Sea Basin (Cartwright 1993, pers. comm.), which is related to over-
BASIN MODELLING CONSTRAINTS pressuring. The timing of the normal faulting coincides with a period of acceleration of Pliocene-Recent subsidence in the North Sea Basin (Cloetingh et al. 1990). This acceleration of subsidence can be explained by an increasing compressive intra-plate stress (Kooi & Cloetingh 1989). The model of Van Balen & Cloetingh (1993) can, therefore, explain the overpressuring which caused the activation of the normal faults in the Palaeogene successions. The Pannonian Basin is a large intra-montane basin developed inside the Alpine chain. The basin originated from a combination of tectonic escape (Ratschbacher et al. 1991) and subduction related processes (Horvfith 1993). The main rifting period for this basin was in the MidMiocene (Badenian). Horvfith & Cloetingh (pers. comm.) have proposed a model in which an increase in the compressive stress field explains the observed anomalous Late Pliocene - Quaternary subsidence in the Pannonian Basin. In the present paper we discuss the effect of this change in magnitude of the compressive stress on the compaction driven fluid flow regime in the two deepest sub-basins in the Pannonian Basin. A full presentation of the results of a forward modelling study of the stratigraphy and fluid flow in the Pannonian Basin will be presented elsewhere (Van Balen et al. 1994).
Intraplate stresses and anomalous
11
2" c o m p r e s s i o n ~
tension
^W:
rift-shoulde~
I
_1
-- --0.5
I
rifted _ .
~ '
--0 -0.5 --3
E v
hinterland
E qJ
-
0
-
3
"o
lithospheric necking during rifting:
,4:,',i:i:i:!:!:i,'i:i'i:i
-0
necking depth"',',',,,','"',"'"""";'"","
E v e-
"0
- 30 :;sub-crust?.:'" isostas~-'-:-~-'.-T-:-~.-:-: -'.
surface expression after isostatic rebound: rifted
rift-sh0ulder'~- - 3 ,-,
subsidence of rifted basins
The stretching model, mathematically formulated by McKenzie (1978), can account for the long term post-rift subsidence (thermal sagging), but does not incorporate rift-basin topography (and, therefore, the distribution of syn-rift deposits) and observed short term deviations of the subsidence pattern in the post-rift phase. Anomalous Pliocene-Quaternary subsidence, observed in a large number of basins in the Northern Atlantic/Mediterranean region (Cloetingh et al. 1990; Cloetingh & Kooi 1992; Zijerveld et al. 1992), is consistent with the model proposed by Cloetingh et al. (1985, 1989). In this model in-plane forces, originating from processes operating at plate boundaries (slabpull, ridge push), modulate the bending of the lithosphere beneath extensional basins (see also Cloetingh & Kooi 1992). The surface expression of this process is differential uplift and subsidence operating on a long wavelength (up to several hundreds of kilometres). The vertical loads acting on the lithosphere determine which parts of the basin are going to be uplifted and which will subside. For models invoking a
3"o
(b) Fig. 2. (a) The effect of changes in the lithospheric stress regime on the subsidence and uplift pattern in a rifted basin. An increase in the level of compressive stress induces flank uplift and basin center subsidence, whereas an increase in the level of tensile stress induces the opposite differential movements. (b) Due to finite strength during rifting the lithosphere necks around its strongest part. Depending on the depth of necking, the isostatic rebound forces can be directed upwards, causing flexurally supported rift-shoulders. shallow depth of lithospheric necking, in the case of compressive forces, the largest downward-pointing lithospheric loads (located under the deepest part of the basin) coincide with areas of enhanced subsidence, while uplifted areas are located where the loads are the least (at the basin flanks). In the case of tensile forces, crustal movements are oppositely directed (see Fig. 2a)
12
R. VAN BALEN & S. CLOETINGH
Such crustal movements are associated with onlap and offlap patterns, bounding system tracts (e.g. Van Wagoner et al. 1988), and can therefore equally well explain relative sea-level changes inferred from these patterns as, for example, glacio-eustatic variations (Cloetingh 1991; Posamentier & Allen 1993) and sediment supply (Schlager 1993). Regional intraplate stress level changes can occur on time scales varying between a few tens to a few million years (Philip 1987). N u m e r i c a l model Basin formation model
Our numerical model for basin formation and evolution is based on the stretching principle, originally proposed by McKenzie (1978). In this model, the lithosphere is kinematically thinned during the rifting time, causing a closer spacing of the geotherms (the geothermal gradient is increased). Relaxation of the disturbed temperature field causes cooling of the rocks, which therefore contract, giving rise to subsidence. The model employs different extension factors for the crustal and mantle part of the lithosphere (Royden & Keen 1980). An important modification of the pure-shear stretching model was incorporated by Kooi et al. (1992) in forward stratigraphic modelling, based on inferences by Braun & Beaumont (1989) and Weissel & Karner (1989). In this model the lithosphere has finite strength during rifting (see Fig. 2b). The actual depth of necking is controlled by bulk rheological properties of the lithosphere. The average response to extension in a multi-layer temperature-dependent rheology with strong layers in the upper crust and upper mantle (Carter & Tsenn 1987; Cloetingh & Banda 1992), therefore, can be cast in terms of a level of necking at lower crustal levels, which itself acts as a detachment level in which major faults sole out (Kooi et al. 1992; Van der Beek et al. 1994). Dynamical analyses of rifted basin formation (Braun & Beaumont 1989; Bassi et al. 1993; Dunbar & Sawyer 1989) show that the lithosphere necks around the strongest part of the crust (i.e. a mid-crustal level), even when vertically orientated crustal or mantle weakness zones are taken into account. Kinematic modelling of rift-shoulder evolution constrained by fission-track data suggests necking depths between 10 and 15 km for the Saudi Arabian Red Sea margin and the Transantarctic Mountains (Van der Beek et al. 1994). Modelling of Neogene western Mediterranean extensional basins formed in collisional regimes including
the Gulf de Lions (Kooi et al. 1992) and the Valencia Trough (Janssen et al. 1993) also yield a best fit for necking depths at mid-crustal levels. In contrast, modelling of the Tyrrhenian Sea Basin points to a deep level of necking at a lower crustal level, around 25km depth (Spadini & Cloetingh pers. comm.). McKenzie's original model assumes that local isostasy during rifting creates a topographic 'hole'. Here we adopt a model in which this feature results from lithospheric necking (see Fig. 2b). Models not invoking finite-strength of the lithosphere require larger amounts of lithospheric thinning, leading to an overestimation for paleo-heatflow (Kooi et al. 1992). This has major implications for the determination of the oil-window as the amount of heat available for hydrocarbon maturation is overestimated (Kooi & Cloetingh 1989). Another effect of introducing necking and flexural isostasy is that, depending on the depth of necking, the vertical loads acting on the lithosphere after stretching can be directed upwards. If the depth of necking is 'sufficiently deep', the basin is overdeep and the lithosphere is in an upward state of flexure, instead of downward. Riflshoulders are supported flexurally by this mechanism (see Fig. 2b). Rift-shoulder erosion
During their uplift, the rift-shoulders will be eroded, causing additional isostatic uplift of the basin margin. At the same time, the erosion products are deposited in the basin, inducing isostatic subsidence (see Fig. 3). Because riftshoulder topography is created by an isostatic uplift caused by lithospheric thinning, this process is particularly important during the syn-rift phase of basin evolution. decreasing
decreasing
K
trans
erosion
.... ============================================================
Fig. 3. Syn-rift erosion of rift-shoulders causes an
additional isostatic uplift at the basin margin due to unloading. At the same time, the increased sediment loading enhances the subsidence in the nearby basin.
BASIN MODELLING CONSTRAINTS In order to quantify the effect of rift-shoulder erosion on basin evolution, sedimentation and erosion are treated as dynamical processes which are modelled numerically as a diffusion problem. The governing equation is (Kenyon & Turcotte 1985; Syvitski et al. 1988; Flemings & Jordan 1989; Peper etal. 1994): Oh 32h Ot - Kt .... Ox---7
(1)
This equation expresses the law of conservation of mass, whose justification of which depends on the transport medium (see Angevine et al. 1990; Kenyon & Turcotte 1985). The transportation coefficient, Kt..... depends on the transport medium and the sediment type. Values of Kt.... can be derived theoretically, although in some cases this is very difficult. In our modelling we have applied empirical values from the literature (e.g. Flemings & Jordan 1989). In general Kt .... increases basinward until the shelf, from where it decreases basinward. The diffusion equation is solved numerically using a one dimensional implicit finite-difference scheme. Two simulation runs are presented in Fig. 4a and b, which illustrate the difference in basin development predicted by models incorporating different amounts of riftshoulder erosion. The same parameters were used for the simulations, the only difference being the presence or absence of erosion of the riftshoulders. During the syn-rift phase, sediments are eroded from the riftshoulders, whereas for the post-rift phase of basin evolution an external source for the sediment is assumed. The stratigraphy predicted by a model ignoring the effects of erosion is shown in Fig. 4a. The stratigraphy is characterized by a continuous onlap of sediments onto the rift-shoulder, with a very minor horizontal distance effected by the shoreline migration. The inclusion of erosion (see Fig. 4b) has two main effects compared with no erosion. The syn-rift sediments, which are derived from the eroded rift-shoulder, experience major syn-rift tilting. This tilting is caused by both erosion-produced additional uplift at the basin margin and enhanced sediment loading in the basin (see Fig. 3). As the amount of tilting diminishes progressively towards the end of the rifting period, a 'break-up' unconformity separates the syn- and post-rift sediments. As shown by Fig. 4b, because of erosion of the riftshoulder, a platform area has developed at the basin margin, causing the post-rift sediments to onlap over a larger horizontal distance. The modelling shows the importance of including
13
riftshoulder erosion in forward basin analyses. As shown by a comparison of Fig. 4a and Fig. 4b, models using the same set of stretching factors can lead to quite different predictions for the basin fill patterns. The 'amount' of rift-shoulder erosion (the value of Kt.... at the riftshoulder area) is dependent amongst others on climatic conditions. If the climate is humid, erosion will be very efficient, whereas in arid climatic conditions erosion is absent. Therefore, the differences in stratigraphy shown in Fig. 4a and 4b could be indicators for respectively dry and wet climatic conditions during rifting. Eustatic sea-level changes (Haq et al. 1987), in-plane stress variations (Cloetingh et al. 1985) and sediment supply (Schlager 1993) are known to influence the development of stratigraphic sequences. However, our results show that climate-controlled erosion of rift-shoulders affects the syn-rift basin fill pattern through its coupling with a tectonic process, i.e. tilting due to isostatic uplift in the eroding area and subsidence in the nearby depocentre. R e s p o n s e o f the basins to vertical a n d h o r i z o n t a l loads
The response of the lithosphere to vertical and horizontal loads, adopting a thin elastic plate analogy for the mechanically strong part of the lithosphere is given by: 02 OX2
O2W O2...~W ( D(x)--~x 2 ) + F 0X 2 + (Pa-Pfin)gw (X) = q(x)
D-
(2)
Ec T~3 12 (1 - vz)
(for notation see Table 1). The effective elastic thickness (Te) of the lithosphere is taken to be equal to the depth of the 450°C isotherm (Kooi et al. 1992), and, therefore, varies in space and time during basin evolution. Temperature calculations are performed on a finite-difference mesh, which is deformed due to the thinning and extension process. The mesh also deforms due to the thermal contraction, which removes the need to make the Boussinesq approximation. In the case of the Pannonian Basin we found that models invoking this approximation underestimate thermally induced subsidence (up to 800m). The finite difference calculations are fully implicit in the vertical direction and explicit in the horizontal direction. The boundary conditions for the temperature calculations are zero heatflow
14
R. VAN BALEN & S. CLOETINGH I
T
i
r
T
--7
on,ap
°
! •
i
E'-" I1) "EI C~l
0
1O0
200 distance (km)
300
0
1O0
200 distance (kin)
300
A
E'-"
v" v
Q.
Fig. 4. (a) Dynamic sedimentation modelling result showing the stratigraphy in the absence of rift-shoulder erosion, representing dry climatic conditions. Each stratigraphic interval represents 2 Ma. The stratigraphy shows a minor onlap to the rift-shoulder. (b) Result of the dynamical sedimentation modelling incorporating rift-shoulder erosion, representing wet climatic conditions. The onlap of sediments to the basin margin is considerable. The syn-rift deposits show tilting due to the couple of additional uplift at the margin and subsidence in the basin. A 'break-up' unconformity results from the ceasure of the tilting. across the side boundaries of the domain, 0 ° C at the top and a user specified (usually 1350 ° C) temperature at the depth of the base of the pre-stretch lithosphere. Sedimentation rates are determined using water depth profiles; i.e. the bathymetry as a function of time is given as an input parameter. The modelling approach taken allows for differ-
ent sediments to be deposited, which obey different hydraulic characteristics and compaction trends.
Fluidflow modelling The module which calculates fluid overpressures and sediment porosities has been extensively
BASIN MODELLING CONSTRAINTS Table 1.
Notation
D Ec
flexural rigidity (N m) Youngs modulus of lithospheric material (Pa) horizontal force (N) total fluid pressure (Pa) lithostatic or overburden pressure (Pa) effective pressure (Pa) effective elastic thickness (m) displacement due to compaction, relative to a fixed point (m) parameters for empirical laws relating porosity to effective pressure and permeability to porosity acceleration of gravity (m/s 2) permeability (m 2) horizontal and vertical permeability (m 2) vertical load due to sediments and seawater (Pa) time (s) Poisson's ratio vertical deflection (m) fluid overpressure, pressure above hydrostatic (Pa) fluid compressibility (Pa-~) porosity, porosity at sedimentation density of asthenosphere (kg/m 3) density of crust (kg/m 3) density of basin fill (kg/m 3) density of grains (kg/m 3) fluid density (kg/m 3) fluid viscosity (Pa s) divergence, gradient
F Pn
PHxH Peff Te 1I...1 a. b, c g k khor, kvert
q t v w qb 13 qb, ~b0 Pa Pc pfill Ps Pn tx VT, V
described in Van Balen & Cloetingh (1993). In this approach porosity is represented by a sediment dependent exponential function of the effective pressure (Shi & Wang 1986):
+z = +oe-be°"z z
(3)
z
eo.= f, (1-+Op~dz+f (+z)o.gdz- P~ The permeability is a function of porosity. Anisotropy is accounted for, as shown in equations 4 (see Bethke 1985). khor = 10 -"+b+,
kver = c. khor
(4)
In contrast to Van Balen & Cloetingh (1993), each finite-element can represent a mixture of sediment types. The overall porosity for the finite-element is given by the volumetric mean porosity. The overall permeabilities in the direction parallel and perpendicular to the strata are determined by considering the analogous problem of electrical resistors wired parallel or in
15
series (permeabilities are 'inverse' resistors, see equations 5). Khortot = aKhorsl + (1 --a) Khors2 1/gvertot =
(5)
a/gversl + (1-a)/Kvcrs2
a = volume percentage of sediment sl. Assuming incompressible grains, fluid overpressure is determined (see Bethke 1985; Van Balen & Cloetingh 1993) by:
+13 0(,~ + 0°Vzm)= VT{_~[V(O)]) at 1 0+ ( 1 - + ) Ot
(6)
The differential equations are solved using Galerkin finite elements. The resulting equations are mutually dependent. This non-linearity is solved using a Picard iteration. The three modules (basin subsidence, sedimentation and fluid flow) are each executed every time step. Because the fluid flow calculations require the smallest step size, this module contains an automatic time step-size adjuster. A typical time step in the simulation is 1000 years.
lntraplate stress variations and compactiondriven fluid flow Van Balen & Cloetingh (1993) have proposed a model for the effect of changing inplane stresses on the fluid flow regime in rifted basins, assuming a very shallow level of necking (see Fig. 5). In this model, an increase in the level of compressive intraplate stress induces basin flank uplift and basin center subsidence. This leads to enhanced sediment exposure at the basin margin and an increase in the gravity potential. Therefore, under humid conditions an increase of meteoric water influx will occur. At the same time, sedimentation rates in a basinward position have amplified (due to an increase of sediment accumulation space and hinterland weathering), inducing an increase in compaction-driven fluid velocities and overpressures. The opposite processes are expected during an increase in the level of tensile intraplate stress, or a drop in the magnitude of compressive stress. Numerical simulations support this hypothesis. The stress-induced perturbations in fluid velocities and overpressures were found to be up to the order of 30%. Because intraplate stress events have a finite duration, stresses are ultimately relaxed, leading to differential motions opposite to those during the increase in the level of stress. Therefore, the intraplate stress-induced perturbation of the
16
R. VAN BALEN & S. CLOETINGH
~
~ w S ° ° ~ ~ + , ,, (m) ~ -soo ....
t
--.,.
I . . . . . . . . . . . . .
L I I I
COMPRESSION
t I
1000 m). The deltaic sediments are better constrained, due to the geometry of delta systems. We need a palaeo-water depth maximum of 1300 m in order to make the model fit. Although huge, it seems plausible. The lake-level variations we have applied are the same as the eustatic sealevel changes given by Haq et al. (1987) with the exception of the Messinian where we included a 150 m sea level drop and subsequent rise in the Late Pliocene to Quaternary. During this time interval there is, according to Haq et al. (1987) an eustatic sea-level rise. We found that during this time interval there cannot have been a lake-level rise in the Pannonian Basin, because this would result in onlapping of sediments on the basement highs, which is not observed. We have therefore kept the lake level fixed since the Late Pliocene. The increase in the level of compressive intraplate stress started at 3.9 Ma, reaching a value of 300MPa at 2.4Ma and subsequently dropping to a present-day value of 200 MPa. As shown by an inspection of Fig. 9, the overall stratigraphy and basin shape predicted by the model compares well with the observed
BASIN MODELLING CONSTRAINTS
21
E¢~ ¢-
"0
tD
0
1O0
200 distance (km)
Fig. 9. The stratigraphy as derived by our forward modelling. The insets show the adopted lake level curve and the palaeo-water depth curve for the deepest part of the Mako trough (the position of the Hod-I well).
large scale pattern in the basin geometry and basin fill. Observed differences in the fine structure of the predicted and observed stratigraphy are probably caused by simplifications in the model assumptions and partly due to uncertainties in the available data. The latter include unknown lake-level changes, previous intraplate stress events, and the assumption that lithostratigraphic units are equivalent to chronostratigraphic units. Although this is approximately true for the older units, this might be wrong for the AP unit, explaining why in contrast to observation, in the model the AP unit onlaps on the basin margin.
Table 3. Numerical values adopted in fluid flow
Compaction-driven fluid flow modelling in the Pannonian Basin. Using our forward strati-
36MPa, due to the increased sedimentation rates above the two sub-basins (2.4Ma). The prediction for the current amount of overpressure is depicted in Fig. 10c. The maximum overpressure reaches 41 MPa, which is in excellent agreement with observations in the Hod-I well (Szalay 1982, 1988). The overpressures are maximal in the DB unit, as it is the least permeable. The overpressures decrease in the deeper B unit. This has implications for the hydrocarbon migration, as discussed below. The B6k6s Basin has developed a slightly less overpressure than the Mako Trough.
graphic model we have made an analysis of the compaction-driven fluid flow system in the Pannonian Basin. The composition of the stratigraphic units is shown in Table 2. The empirical parameters for porosity and permeability laws are depicted in Table 3. The results for three different time slices are shown in Fig. 10. The fluid overpressures prior to the compressive stress event are shown in Fig. 10a. The maximum overpressure is approximately 21MPa, reflecting the dramatic increase in sedimentation rates during the deposition of the DS unit due to the arrival of the Pannonian delta in the deep sub-basin (5.5 Ma). The effect of the increase of compressive stress on the fluid overpressures is shown in Fig. 10b. The maximum overpressure has dramatically increased to
calculations Sediment type
Parameter
Value
Sandstone
4) logkhor
0.475e -45 I°-~Peff -- 18 + 8+ 1.62 2.7gcm -3 0.65e -s's2 t°-SPeff -21 + 8+
khor/kv~rt Marl
p~ + Iogkhor khor/kvert P~
11.7 2.7g cm-3
Discussion and conclusions Models incorporating the finite strength of the lithosphere during rifting and changes in the stress regime during the post-rift phase of
22
R. VAN BALEN & S. CLOETINGH
(a)
A
EOJ
t~)
0
100
200 distance (km)
(b) o
,,e t-i1) "0
1O0
200
distance (kin) extensional basins can successfully explain a number of important features recognized in the basin fill, not compatible with models invoking local isostasy for the basin formation and equating post-rift subsidence to thermal subsidence. Better understanding of the rheology of the lithosphere and its interplay with erosion processes and stress changes enhances the quality of predictions generated by forward models for basin stratigraphy and fluid flow. In particular, the modelling has demonstrated that an increase in the level of compressive stress or a decrease in the amount of tensile stress can lead to substantial flank uplift and basin centre
subsidence. Therefore, more meteoric water will infiltrate from the margins into the basin, due to increased sediment exposure and gravity potential. At the same time, sedimentation rates in a basinward position increase, leading to an increase in compaction-driven fluid overpressuring and flow velocities. An increase of tensile stress or a decrease of compressive stress have the opposite effects. As intraplate stresses can change sign and magnitude on short time scales, the fluid regime also can change very rapidly due to this mechanism. This process triggers faulting and fracturing, alters heatflow and diagenesis (Horbury & Robinson 1993), enhances primary
BASIN MODELLING CONSTRAINTS
23
(c)
ECq ,¢ ¢Q~ "o
0
100
200
distance (krn) Fig. 10. The results from the compaction-driven fluid flow modelling. Thin lines represent the stratigraphic layering, thick lines are isobars (1, 10, 20, 30 and 40 MPa). (a) Fluid overpressures at 5.5 Ma, showing the effect of the increased sedimentation rates due to the arrival of the Pannonian delta. The maximum overpressure is 21 MPa. (b) The overpressuring regime at 2.4 Ma, demonstrating the impact of the stress-induced subsidence on the fluid overpressures, which have increased to a maximum value of 36 MPa. (c) The present overpressures as predicted by our modelling. The maximum value of 41 MPa is in excellent agreement with observations. migration of hydrocarbons and transportation of metal-rich brines in a pulse-like way. The forward modelling of fluid flow in the Pannonian Basin has shown that an increase in the level of compressive stress during the late stage evolution of this extensional basin has altered the compaction-driven fluid overpressuring substantially. Fluid overpressures have almost doubled due to the effect of the increase of compressive stress. This has a number of implications for the diagenesis of sediments and primary migration of hydrocarbons. The hydrocarbon generation started between 9 and 6 Ma (Horv~th etal. 1987; Szalay 1988), implying that during the stress-induced change of the fluid regime in the Pannonian Basin the hydrocarbons were mature and actively expelled. The dramatic increase of overpressure has probably enhanced the primary migration considerably. Our modelling also shows that the overpressures decrease with depth in the deepest part of the basin, enhancing primary migration into this unit from the overlying sediments, enhancing the prospectivity of the deepest unit. This research was funded by the IBS (Integrated Basin Studies) project, part of the JOULE I1 research programme funded by the Commission of European Communities (contract nr. JOU2-CT 92-0110). F. Horvgtth and L. Lenkey made significant contributions
to the modelling of the Pannonian Basin. M. G61ke kindly provided us with the European stress map (Fig. 1). IBS contribution No. 7 Publication No. 8 of the Netherlands Research School of Sedimentary Geology.
References ANGEVINE, C.L., HELLER, P. & PAOLA, C. 1990.
Quantitative sedimentary basin evolution. American Association of Petroleum Geologists, Continuing Education Course Note Series, 32. BASSI, G., KEEN, C.E. & POTTER,P. 1993. Contrasting styles of rifting: models and examples from the eastern Canadian margin. Tectonics, 12,639-655. BECKER, A. 1993. Contemporary state of stress and neotectonic deformation in the CarpathianPannonian region. Terra Nova, 5,375-388. B~RCZl, I. & PHILLIPS, R.L. 1985. Processes and depositional environments within Neogene deltaic-lacustrine sediments, Pannonian Basin, southeast Hungary. Geophysical Transactions, 31, 55-74. BETHKE, C.M. 1985. A numerical model of compaction-driven groundwater flow and heat transfer and its application to the paleohydrology of intracratonic sedimentary basins. Journal of Geophysical Research, 90, 6817-6828. --,
HARRISON, W.J., UPSON, C. • ALTANER, S.P.
1988. Supercomputer analyses of sedimentary basins. Science, 239,261-267.
24
R. VAN BALEN & S. CLOETINGH
DEMETRESCU,C. 8z POLONIC,G. 1989. The evolution of the Pannonian Depression (Romanian sector) as derived from subsidence and heat flow data. Tectonophysics, 164,287-299. DOGLIONI, C. 1991. A proposal for the kinematic modelling of W-dipping subductions - possible applications to the Tyrrhenian-Appenine system. Terra Nova, 3,423--434. DUNBAR, J.A. & SAWYER,D.S. 1989. HOW preexisting weaknesses control the style of rifting. Journal of Geophysical Research, 94, 7278-7292. FLEMINGS, P.B. & JORDAN, T. 1989. A synthetic stratigraphic model of foreland basin development. Journal of Geophysical Research, 94, 3851-3866. GLUYAS, J.G., GRANT, S.M. & ROBINSON,A.G. 1993. Geochemical evidence for a temporal control on sandstone cementation. In: HORBURY, A.D. & ROBINSON, A.G. (eds), Diagenesis and Basin Development, AAPG Studies in Geology Series, 36, 23-33. HARRISON, W.J. & TEMPLE, R.N. 1993. Diagenetic pathways in sedimentary basins. In: HORBURY, A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development, AAPG Studies in Geology Series, 36, 69-86. HAQ, B.U., HARDENBOL,J. & VAIL, P. 1987. Chronology of fluctuating sea levels since the Triassic. Science, 235, 1156-1167. HORBURY, A.D. & ROBINSON, A.G. (eds) 1993. Diagenesis and Basin Development, AAPG Studies in Geology Series, 36. HORV~,TH, F. 1985. Thickness of Neogene-Quaternary basin fill (map). Hungarian Geological Institute, Budapest, Hungary. 1993. Towards a kinematic model for the Pannonian Basin. Tectonophysics, 226,333-357. & BERCKHEMER,H. 1982. Mediterranean backarc basins. In: BERCKHEMER, H & HSU, K. (eds) Alpine-Mediterranean geodynamics. Geodynamic Series 7, American Geophysical Union and Geological Society of America, Boulder Colorado, 141-173. - - , DOVENYI,P., PAPP, S., SZALAY,A. & HARGITAY, M. 1987. Tectonic control on migration and trapping of hydrocarbons in the Pannonian Basin. In: DOLIGEZ, B. (ed.) Migration of hydrocarbons in sedimentary basins, Editions Technip, Paris, 667-681. --, SZALY, A. & ROYDEN, L. 1988. Subsidence, thermal, and maturation history of the Great Hungarian Plain. In: ROYDEN, L.H. & HORV~,TH,F. (eds) The Pannonian Basin: A Study in Basin Evolution, American Association of Petroleum Geologists, Memoirs, 45,355-372. JANSSEN, M.E., TORNI~,M., CLOETINGH,S. & BANDA, E. 1993. Pliocene uplift of the eastern Iberian margin: inferences from quantitative modelling of 208,221-241. the Valencia Trough. Earth and Planetary Science DANK, V. 1988. Petroleum geology of the Pannonian Basin, Hungary: an overview. In: ROYDEN, Letters, 119,585-598. JENSEN, R.P. & DOR~, A.G. 1993. A recent NorL.H. & HORV~TH, F. (eds) The Pannonian wegian Shelf heating event - fact or fantasy? In: Basin: A Study in Basin Evolution. American DORI~, A.G. et al. (eds) Basin Modelling: AdAssociation of Petroleum Geologists Memoirs, vances and Applications. Norwegian Petroleum 45,319-331.
BIELIK, M. 1988. Analyses of the stripped gravity map of the Pannonian Basin. Geologica Carpathica, 39, 99-108. BRAUN, J. & BEAUMONT, C. 1989. A physical explanation of the relationship between flank uplifts and the breakup unconformity at rifted continental margins. Geology, 17,760-764. CARTER,N.L. & TSENN, M.C. 1987. Flow properties of continental lithosphere. Tectonophysics, 136, 27-63. CARTWRIGHT, J.A. 1993. Giant mud cracks in the Paleogene of the North Sea: evidence for basin scale hydrofracturing of overpressured shale sequences, extended abstract. In: PARNELL, J., RUFEELL, A.H. & MOLES, N.R. (eds) Geofluids '93, Torquay. 4-7 May 1993, 172-174. CLAYTON, J.L., SPENCER,C.W., KONCS, I. & SZALAY, A. 1990. Origin and migration of hydrocarbon gases and carbon dioxide, B6k6s Basin, southeastern Hungary. Organic Geochemistry, 15, 233-247. CLOETINGH,S. 1991. Tectonics and sea level changes a controversy? In: MULLER, D., WEISSEL, H. & MCKENZIE, D. (eds) Controversies in modern geology: a survey of recent developments in sedimentation and tectonics. Academic Press, London, 249--277. -& BANDA, E. 1992. Mechanical structure of Europe's lithosphere. In: BLUNDELL, D., MULLER, S & FREEMAN, R. (eds) A continent revealed: The European Geotraverse. Cambridge University Press, 80-91. & KooI, H. 1992. Intraplate stresses and dynamical aspects of rift basins. Tectonophysics, 215,167-185. --, GRADSTEIN, F., KOOI, H., GRANT, A. & KAMINSKI,M. 1990. Plate reorganization: a cause for rapid Neogene subsidence and sedimentation around the North Atlantic? Journal of the Geological Society, London, 147,495-506. , KooI, H. & GROENEWOUD,W. 1989. Intraplate stress and sedimentary basin evolution. In: PRICE, R.A. (ed.) Origin and Evolution of Sedimentary basins and these energy and mineral resources. American Geophysical Union, Geophysical Monographs, 48, 1-16. , McOuEEN, H. & LAMBECK, K. 1985. On a tectonic mechanism for regional sea level variations. Earth and Planetary Science Letters, 75, 31-61. - - , SASSI, W. & HOVATH,F. (eds) 1993. The origin of sedimentary basins: inferences from quantitative modelling and basin analyses. Tectonophysics, 226, 1-504. CSONTOS, L., NAGYMAROSY, A., HORVATH, F. & KovAc, M. 1992. Tertiary evolution of the Intra-Carpathian area: a model. Tectonophysics,
BASIN MODELLING CONSTRAINTS Society (NPF), Special Publications, 3, Elsevier, Amsterdam, 85-106. JUH~,SZ, G. 1991. Lithostratigraphical and sedimentological framework of the Pannonian (s.1.) sedimentary sequence in the Hungarian Plain (Alf61d), Eastern Hungary. Acta Geologica Hungarica, 34, 53-72. KAZM~R, M. 1990. Birth, life and death of the Pannonian Lake. Palaeogeography, Palaeoclimatology, Palaeoecology, 79,171-188. KENYON, P.M. & TURCOTrE, D.L. 1985. Morphology of a delta prograding by bulk sediment transport. Geological Society of America Bulletin, 96, 14571465. KNIPE, R.J. 1993. The influence of fault zone processes and diagenesis on fluid flow. In: HORBURY,A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development, AAPG Studies in Geology Series, 36,135-151. Kool, H. & CLOETINGH, S. 1989. Intraplate stresses and the tectono-stratigraphic evolution of the central North Sea. In: TANKARD, A.J. & BALKWILL, H. (eds) Extensional tectonics and stratigraphy of the North Atlantic margins. American Association of Petroleum Geologists Memoir, 46,541-558. &~ 1992. Lithospheric necking and regional isostasy at extensional basins 2. Stress induced vertical motions and relative sea level changes. Journal of Geophysical Research, 97, 1757317591. , -& BURRUS, J. 1992. Lithospheric necking and regional isostasy at extensional basins 1. Subsidence and gravity modelling with an application to the Gulf of Lions margin (SE France). Journal of Geophysical Research, 97, 1755317571. L~,Z.~RESCU, V., CORNEA, I., R.g.DULESCU, F. & POPESCU, M. 1983. Moho surface and recent crustal movements in Romania: geodynamic connections. Proceedings of the Institute of Geology and Geophysics in Bucarest, 63, 83-91. MCKENZIE, D.P. 1970. Some remarks on the development of sedimentary basins. Earth and Planetary Science Letters, 40, 25-31. MATHCK, R.E., PHILIPS, R.L. & RUMPLER, J. 1988. Seismic stratigraphy and depositional framework of sedimentary rocks in the Pannonian Basin, southeastern Hungary. In: ROYDEN, L.H. & HORV~.TH, F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45, 117-146. MOLLER, B., ZOBACK, M.L., FvcrIS, K., MASTIN, L., GREGERSEN, S., PAVONI, N., STEPHANSON,O. & LJUNGGREN, C. 1992. Regional patterns of tectonic stress in Europe. Journal of Geophysical Research, 97, 11783-11803. PEPER, T., BEEKMAN, F. & CLOETINGH, S. 1992. Consequences of thrusting and intraplate stress for foreland basin development and peripheral areas. Geophysical Journal International, l l l , 104-126. , VAN BALEN, R. & CLOETINCH, S. 1994. Implications of orogeny, intraplate stress variations -
-
25
and eustatic sealevel change for foreland basin stratigraphy - inferences from numerical modelling. In: DOROBEK,S. & Ross, G. (eds) Stratigraphy and foreland basins. SEPM Special volume, in press. PHILIP, H. 1987. Plio-Quaternary evolution of the stress field in Mediterranean zones of subduction and collision. Annales Geophysicae, 5b, 301-320. POSAMENTIER,H.W. & ALLEN, G.P. 1993. Variability of the sequence stratigraphic model: effects of local basin factors. Sedimentary Geology, 86, 91-109. RATSCHBACHER,L., FRISCH,W., LINZER,H. & MERLE, O. 1991. Lateral extrusion in the Eastern Alps, part 2: structural analyses. Tectonics, 10, 257271. REEMST, P., CLOETINGH, S. & FANAVOLL, S. 1994. Tectono-stratigraphic modelling of Cenozoic uplift and erosion in the SW Barents Sea. Marine and Petroleum Geology, in press. R6NAI, A. 1974. Size of Quaternary movements in Hungary's area. Acta Geologica Academiae Scientiarum Hungaricae, 18, 39--44. 1985. Limnic and terrestial sedimentation and the N/Q boundary in the Pannonian Basin. In: KRETZOI, M. & P~csl, M. (eds) Problems of the Neogene and Quaternary. Akad6miai Kiad6, Budapest, 000-4)00. ROVDEN, L.H. & D6V~NVI, P. 1988. Variations in extensional styles at depth across the Pannonian Basin. In: ROYDEN, L.H. & HORVATH, F. (eds)
The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,235-255. & KEEN, C.E. 1980. Rifting processes and thermal evolution of the continental margin of eastern Canada determined from subsidence curves. Earth and Planetary Science Letters, 51, 343-361. --, HORVATH, F. & NAGYMAROSY,A. 1983. Evolution of the Pannonian Basin system 2: subsidence and thermal history. Tectonics, 2, 91-137. RUMPLER, J. & HORV.~TH, F. 1988. Some representative seismic reflection lines from the Pannonian Basin and their structural interpretation. In: ROYDEN, L.H. & HORV~,TU,F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,153-171. SARKOVI(~, M., STANKOVI(~,S. & MILOSLAVLJEVR~,S. 1991. Petroleum geology of the southeastern part of the Pannonian Basin. Terra Nova, 3, 550-554. SCHLAGER, W. 1993. Accommodation and supply - a dual control on stratigraphic sequences. Sedimentary Geology, 86,111-136. SCLATER, J.G. & CHRISTIE, P.A.F. 1980. Continental stretching: an explanation of the .post-midCretaceous subsidence of the central North-Sea Basin. Journal of Geophysical Research, 85, 3711-3739. , ROYDEN, L., HORV~TH, F., BURCHFIEL, B.C., SEMKEN,S. & STEGENA,L. 1980. The formation of the Intra-Carpathian basins as determined from
26
R. VAN BALEN & S. CLOETINGH
subsidence data. Earth and Planetary Science Letters, 51,139-162. Sin, Y. & WANG, C. 1986. Pore pressure generation in sedimentary basins: overloading versus aquathermal. Journal of Geophysical Research, 91, 21532162. STEGENA, L., G~CZY, B. & HORVXTH, F. 1975. Late Cenozoic evolution of the Pannonian Basin. Tectonophysics, 26, 71-90. SWITSKI, J.P.M., SMrrH, J.N., CALABRESE,E.A. & BOUDREAU, B.P. 1988. Basin sedimentation and the growth of prograding deltas. Jounral of Geophysical Research, 93, 6895-6908. SZALAY,A. 1982. A rekonstrukci6s szemldletii fOldtani
kutatds lehetOs~gei a szdnhidrogdn-perspektivdk elOrejelz~sdben [Possibilities of the reconstruction of basin evolution in the prediction of hydrocarbon prospects]. PhD thesis, Hungarian Academy of Sciences, Budapest. 1988. Maturation and migration of hydrocarbons, SE Pannonian Basin. In: ROYDEN, L.H. & HORV~,TH, F. (eds) The Pannonian Basin: A Study in Basin Evolution. American Association of Petroleum Geologists Memoirs, 45,347-354. TARI, G., HORV.~TH,F. d(z RUMPLER,J. 1992. Styles of extension in the Pannonian Basin. Tectonophysics, 208,203-219. VAN BALEN,R. ¢~zCLOETINGH,S. 1993. Stress-induced fluid flow in rifted basins. In: HORBURY, A.D. & ROBINSON, A.G. (eds) Diagenesis and Basin Development. AAPG Studies in Geology Series, 36, 87-98.
~,
LENKEY,L., CLOETINGH,S. • HORVXTH,F. 1994. Two-dimensional modelling of stratigraphy and fluid flow in the Pannonian Basin. Marine and Petroleum Geology, in press. VAN DER BEEK, P., CLOETINGH,S. & ANDRIESSEN,P. 1994. Mechanisms of extensional basin formation and vertical motions at rift flanks: Constraints from tectonic modelling and fission-track thermochronology. Earth and Planetary Science Letters, 121, 417-433. VAN WAGONER, J., POSAMENTIER, H.W., MITCHUM, R.M., VAIL, P.R., SARG, J.F., LOUTIT, T.S. & HARDENBOL, J. 1988. An overview of fundamentals of sequence stratigraphy and key definitions. Sea-level changes - an integrated approach, The Society of Economic Paleontologists and Mineralogists Special Publications 42, 40-45. WEISSEL, J.K. & KARNER, G.D. 1989. Flexural uplift of rift flanks due to mechanical unloading of the lithosphere during extension. Journal of Geophysical Research, 94, 13919-13950. ZOBACK,M.L. 1992. First and second order patterns of stress in the lithosphere: The World Stress Map Project. Journal of Geophysical Research, 97, I 1703-11728. ZIJERVELD, L., STEPHENSON, R.A., CLOETINGH, S., DUIN, E. & VAN DEN BERG, M. 1992. Subsidence analysis and modelling of the Roer Valley graben (Southeastern Netherlands). Tectonophysics,
208,159-171.
Fluid flow and heat transport in the upper continental crust DAVID
DEMING
School o f Geology & Geophysics, University o f Oklahoma, Norman, O K 73019-0628, USA
Abstract: The upper 10-15 km of the continental crust is saturated with aqueous brines and is sufficiently permeable to allow the circulation of these fluids. The most important driving forces for fluid flow in the continental crust are topography gradients, sediment compaction and diagenesis, and buoyancy forces. Topographically-driven flow is generally the most efficient mechanism for mass and heat transport, and therefore has recently been favoured in theories of the origin of Mississippi Valley-type lead-zinc deposits. The magnitude of heat transport by fluid movement in the crust has an exponential dependence upon the depth of circulation. In areas of the continental crust with appreciable topography gradients (c.0.01) and permeabilities greater than c. 10-17m2, the background thermal state will likely be appreciably perturbed by groundwater movement. Fluid circulation in the upper crust is both continuous and pervasive due to topography and fluid-density gradients. The old conception of the continental crust as an unchanging body of solid rock should be replaced by a paradigm that recognizes the continental crust as a two-component system; a solid framework which continuously evolves through thermal, chemical, and mechanical interaction with crustal fluids.
Importance of fluid and heat flow With the advent of plate tectonics, the concept of a static Earth was replaced by a new picture of the Earth's crust as a domain caught up in a process of continual evolutionary change. An increasing knowledge of the importance of fluids in the Earth's crust is now invalidating the old concept of the crust as being composed of unchanging bodies of solid rock. As geological fluids move through the crust they enter into chemical reactions with solid components of the crust, mobilize and transport both heat and mass, alter the geologic and geophysical properties of crustal rocks, and create economic concentrations of important ore minerals. Fluids play an important role in nearly all geological processes and are ubiquitous in the continental crust to depths of at least 10 to 15km (see Table 1 and Nesbitt & Muehlenbachs 1989, 1991; Zoback et al. 1988; N R C 1990; Oliver 1986, 1992). In the new paradigm, the continental crust is viewed as a two-component system: a solid framework that evolves through interactions with geological fluids. A convenient division may be made between hydrological regimes in the upper and lower continental crust. In the upper continental crust, the permeability of sedimentary and crystalline rocks may be sufficiently high to allow continuous flow systems that may conceivably persist for tens of millions of years. Below depths of
Table 1. Possibleindicators offluids in the continental crust (after Nur & Walder, 1990, p. 114) Indicator
Depth range
Water table Deep wells Reservoir induced seismicity Crustal low velocity zones Crustal electrical conductivity zones Oxygen isotopes Metamorphism Crack healing and sealing Hydrothermal ore deposits Crustal seismic attenuation zones Low stress on faults Silicic volcanism Fluid inclusions
0-2 km 0-12 km 0-12 km 7-12 km 10-12 km 0-12 km >20 km ? >5 km 7-12 km 0-10 km near surface ?
approximately 10-20km, however, the continental crust is thought to be essentially impermeable. Open fractures probably cannot exist below the brittle-ductile transition, and unfractured crystalline rocks are otherwise as impermeable as window glass. Nevertheless, there is considerable geochemical evidence from examination of exhumed terrains that large amounts of fluid have circulated through the lower crust (Walther 1990). One way in which the inferences of large fluid fluxes can be reconciled with the rheological requirements of low permeability is episodic hydrofracturing (Nur & Walder 1990).
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 27-42.
27
28
D. DEMING S A L I N I T Y (%o) 0 0
200 I
I
400 I
if
• - ~-..
I
500
i
• m
•
m
E a:: I000
v
.m
•m
a
•
•w
•
•
w'lwm
1500-
3
w
i,
2000
.
'
' '
.
.
.
.
.
.
'
II0
'
....
I
. . . . . . . .
I00
I000
Total Dissolved Solids (mgll)
4 Fig. 1. Maximum observed salinities of pore fluids from selected sedimentary basins (after Hanor 1979; Ranganathan & Hanor 1987). Theoretical calculations based upon the physical properties of aqueous solutions (Norton & Knapp 1977; Norton 1984, 1990; Nur & Walder 1990) have shown the feasibility of this mechanism, and there is field evidence that cyclic flow mechanisms probably exist near magmatic intrusions in the upper crust (Titley 1990). Nature of fluids in the crust Fluids found in the continental crust are primarily aqueous brines; most data concerning these brines come from sedimentary basins. The salinity of aqueous brines from sedimentary basins is commonly found to increase with increasing depth (Fig. 1) to a maximum of more than 40% solids by weight (Dickey 1969; Hanor 1979). Stable isotope studies indicate that most of the formation fluids found in sedimentary basins are not connate water trapped at the time of sediment deposition, but of meteoric origin (Clayton et al. 1966; Kharaka 1986). The presence of juvenile fluids in the crust is relatively rare (White 1974; Truesdell & Hulston 1980). In comparison to sedimentary environments there is less known about the nature of fluids in crystalline rocks, no doubt due to the scarcity of boreholes and the limited availability of fluid samples. One of the best data sets that is available comes from a study of crustal fluids in the Canadian Shield. In the late 1970s a number of relatively shallow (c. 1600m, maximum depth) boreholes were drilled in crystalline rocks of the Canadian Shield for the purpose of studying possible sites for the disposal of nuclear-fuel waste products. A study of fluid samples from
Fig. 2. Salinities of pore fluids from crystalline rocks in the Canadian Shield (after Frape & Fritz 1987). these boreholes showed that the dominant fluid
was a highly saline Ca-Na-C1 brine. Total dissolved solids content was found to increase exponentially with depth, from fresh water near-surface (c.100m) to c.150g1-1 at c.1500m depth (Fig. 2). Had the boreholes penetrated depths exceeding c. 1600 m, salinity presumably would have continued to increase. However, the rate of increase could not have been maintained as extrapolation of the observed trend would have exceeded saturation values near 2000m depth. Analyses showed the isotopic signature of brines recovered from the Canadian boreholes to be very much different from fluids from sedimentary or geothermal environments, evidently reflecting a long history of intense fluid-rock interactions (Frape & Fritz 1987).
Mathematical description of fluid flow Hydraulic head
Fluid movement in porous media occurs in response to head gradients and buoyancy forces. Hydraulic head h is mechanical energy per unit mass of fluid, equivalent to the height above an arbitrary datum to which fluid will rise in a tube, otherwise known as the potentiometric surface (Hubbert 1940). It is convenient to divide head into an elevation and pressure term, h = z + P/pg
(1)
where h is hydraulic head, z is elevation above an arbitrary datum, P is fluid pressure, P is fluid density, and g is the acceleration due to gravity (Freeze & Cherry 1979). A persistent fallacy, first corrected by Hubbert (1940), is that fluids flow from regions of high pressure to low pressure. As can be seen by inspection of
FLUID FLOW AND HEAT TRANSPORT equation (1), head gradient is proportional to pressure gradient only for the special case where elevation is constant. Thus fluids do not necessarily flow in the direction of decreasing pressure. For example, water in a drinking glass does not spontaneously flow from the bottom to top of the glass. Equation of flow
Neglecting inertial forces, a continuum (volumeaveraged) description of the laminar flow of an inhomogeneous fluid through an anisotropic porous medium is (Bear 1972) q = - ( E / I x ) [VP+pgVz]
(2)
where q is the Darcy velocity or specific discharge (volumetric flow rate per unit time divided by cross-sectional_ area perpendicular to flow direction), E is the permeability tensor, Ix is fluid dynamic-viscosity, VP is the gradient of fluid pressure, z is elevation, and g, the acceleration due to gravity, is assumed to act in the z-direction. Under the assumptions stated above, equation (2) always holds true, regardless if flow is steady-state or transient. The Darcy velocity q is distinguished from linear velocity, v, which is the average speed with which individual water molecules move through a porous medium. In a porous medium of fractional porosity qb, linear velocity v is related to Darcy velocity q by v=q/+.
(3)
By defining an equivalent fresh-water hydraulic head h, h = z + P/pog
(4)
equation (2) can be rewritten as q=-(~pog/ix)[vh+P-P°Vz] Po
(5)
where Po is a reference fluid density. Thus fluid flow results not only from a head gradient Vh, but also fluid-density gradients which give rise to buoyancy forces. It follows that in the general case of a fluid of variable density, head is not a potential field as q is not exactly proportional to Vh (Bear 1972). However, if an assumption of constant density fluid is made, equation (5) reduces to the familiar form of Darcy's Law (Darcy 1856; Hubbert 1969) q = - (~ 9og/tx)Vh
(6)
q = - KVh
(7)
or
where K, the hydraulic conductivity tensor, is a
29
function of both the porous medium (rock) and the fluid moving through it. Hydraulic conductivity
Some of the factors that determine hydraulic conductivity may vary over orders of magnitude in different geological settings; others tend to be relatively constant throughout the crust. Fluid density, p, is determined by temperature, salinity, and pressure. Of these three factors, temperature and salinity are usually more important. For example, the density of pure water decreases by c.8% as temperature increases from 25 to 150°C at constant pressure. Assuming that this increase of temperature corresponds to a depth increase of c.5km (25°Ckm-1), the corresponding increase of pressure under hydrostatic conditions is c.50 MPa. An increase in pressure of 50 MPa at 25°C results in c.2% increase in water density. Thus the decrease in density due to thermal effects is about four times larger than the density increase due to pressure effects. In sedimentary basins, the decrease in fluid density due to increased temperature is usually more than offset by increases in salinity that result in fluid density increasing with increasing depth. However, in the absence of appreciable salinity gradients, fluid columns in the crust are potentially unstable and may be susceptible to convective overturn. The acceleration due to gravity, g, is essentially constant over the Earth's surface, varying from c.9.83 ms -2 at the poles, to c.9.78 ms -z at the equator for elevations at sea level and below. Dynamic fluid viscosity, Ix, is strongly temperature-dependent. An empirical formula for the dynamic viscosity of pure water is Ix = 2.414 x 10-5 x 10 (248"37/(T+133"15)) (8) where tx has units of k g m - l s -1 and T is temperature in degrees Celsius (Touloukian et al. 1975). Dynamic viscosity decreases by approximately an order of magnitude as temperature increases from 0 to 150° C. All other factors being equal, the reduction of viscosity with increasing temperature tends to promote deep flow in the crust (Smith & Chapman 1983). The effect of salinity is to increase the dynamic viscosity, but the magnitude of the effect is generally smaller than the temperature dependence (see Phillips et al. 1981). Generally the most significant factor that determines hydraulic conductivity is permeability, k. The permeability of Earth materials varies over more than 11 orders of magnitude, from values as high as 10-9 m 2 for
30
D. DEMING
karst limestone, to values lower than 10-2°m 2 for some shales and unfractured crystalline rocks. Presumably the average permeability of lithologies such as salt may be even lower than 10-2°m 2. One of the primary tenets of hydrogeology is that there are no completely impermeable rocks or geological materials (Bredehoeft et al. 1982). A possible exception is ice-rich permafrost, in that any water attempting to move through such a layer simply freezes. Rock permeability is highly variable. For example, permeability measurements on cores from sedimentary basins that sample geological horizons with uniform lithologies commonly vary over five orders of magnitude. Rock permeability also changes with scale (Brace 1980, 1984; Garven 1986; Ciauser 1992). Estimates of permeability made from pumping tests in wells commonly yield values one to two orders of magnitude greater than those derived from measurements on core samples. Core measurements may fail to sample fracture networks, the presence of which can dramatically increase the effective permeability. At the present time, it is not well understood if permeability also increases from the well scale (101-103m) to the regional scale (105-106m). Much of what we know about the change of permeability with scale comes from the studies of sedimentary basins in the central and western US by Bredehoeft and his colleagues at the US Geological Survey. Bredehoeft et al. (1983) found that for their numerical models of regional groundwater flow to yield acceptable matches to hydraulic head data, it was necessary to specify the permeability of the Cretaceous shale confining layer in South Dakota as one to three orders of magnitude higher than values derived from in situ well tests and laboratory measurements on cores. Similarly, Belitz & Bredehoeft (1990) found the regional scale permeability of the shale confining layer in the Bighorn Basin in the western US to be 1000 to 10000 times higher than would be inferred from laboratory measurements. The dramatic increase from the laboratory to regional scale was attributed to extensive high-angle thrust faulting. In contrast, Belitz & Bredehoeft (1988, 1990) estimated that the permeability of the shale confining layer in the Denver Basin of the western US was in the range 10-19-10-2°m 2, and exhibited no significant increase from the laboratory to regional scales. From these three studies, it would appear that scale dependence of confining layer permeabilities is entirely dependent upon fracturing and faulting, and thus the regional geology and tectonic setting. Clauser (1992) reviewed data from crystalline
rocks and noted that permeability increases approximately three orders of magnitude from the core (10-2-10 -1 m) to the well scale (101102m), but found no corresponding increase from the well to the regional scale (103-105 m). Deming (1993) studied regional-scale (c.300 km) groundwater flow and concomitant heat transport in the North Slope Basin, Alaska. Simulations of coupled heat and fluid flow in an idealized model showed that the general trend of near-surface conductive heat flow estimated from borehole data could be reproduced by numerical models incorporating permeability data from core measurements. Thus, it would appear in the case of the North Slope Basin that there is no apparent increase of permeability from the core to the regional scale. However, the direct comparison of core data with largescale indirect estimates is complicated by considerations such as possible bias in selection of intervals for core measurements and the choice of an appropriate algorithm for averaging core data (e.g., arithmetic, geometric, or harmonic mean?). These considerations made it difficult to draw conclusive inferences in this instance. Diffusion equation
Through the implicit assumption of constantdensity fluid, hydraulic head can be approximated as a hydraulic potential (or pseudopotential) for the purpose of obtaining insight into geological problems for which an exact answer is unobtainable. In this case, there are no buoyancy forces and the transient movement of fluids in the crust is described by the diffusion equation. In complex and heterogeneous geological environments, the assumption of constant-density fluid is nearly always violated, but the diffusion equation nevertheless is a useful approximation that can be used to obtain insight into the physics of fluid flow in the crust. For a porous medium that is both homogeneous and isotropic, the diffusion equation is Oh _ g V2h Ot Ss
(9)
where h is hydraulic head, t is time, K is hydraulic conductivity, V2h is the curvature of the head field, and Ss is a material property called specific storage. Specific storage refers to the physical ability of a porous medium to store fluid, and is largely determined by the compressibility of the porous medium itself and the compressibility of the fluid which saturates it S~= pg(c~ + CB)
(10)
FLUID FLOW AND HEAT TRANSPORT where P is fluid density, g is the acceleration due to gravity, + is porosity, c~ is porous medium compressibility, and B is fluid compressibility. Freeze & Cherry (1979) give the compressibility of water as 4.4 x 10-mPa -~, and cite a range of compressibilities for consolidated rocks that range from 10 -9 to 10-~IPa -1. Ge & Garven (1992) reviewed porous medium compressibilities for consolidated rocks and found most shale compressibilities to be of the order of 10 -9 Pa -~ , with other lithologies (sandstone, limestone, granite, quartzite, etc.) commonly of the order of 10-m-10 -~1Pa -1. Neuzil (1986) pointed out that most compressibility data reflect measurements in the laboratory at human time scales, and therefore may underestimate in situ compressibilities. Estimates of in situ compressibilities for sedimentary rocks derived from porosity-depth relationships commonly are one to three orders of magnitude higher than laboratory data. However, porosity reduction in the subsurface may not solely reflect mechanical compaction, as porosity may also be reduced by chemical processes such as mineralization. At the present time, it appears that our best estimate is that most rocks fall in the range 10 - u -< et 1) are required at all depths for hydraulic extension fractures to develop in compressional stress regimes where o.v = o.3. Localization of hydrofracture arrays to the immediate vicinity of fault zones suggests that, in many cases, the fault zones are themselves the principal conduits for the migration of overpressured fluids (Cox et al. 1991). The development of hydrofracture dilatancy in compressional regimes that are strongly fluid overpressured is an accompaniment to extreme fault-valve action (see below). Grain-scale microcrack dilatancy and associated particulate flow (approximately equivalent to sand-pile dilatancy) may also develop in fluid overpressured rock masses where Xv~ 1, and the effective mean stress, 6.'= ( 6 . - P f ) ~ 0 (Borraidale 1981; Cox & Etheridge 1989; Fischer & Paterson 1989; Knipe 1989). Mixed dilatancy effects. In most natural settings it is likely that more than one of these various dilatancy mechanisms may be operating at a given time. For example, consider the coupled variation of both shear and mean stress around thrust and normal faults as illustrated in Fig. 4. In the case of the thrust fault, any dilatancy related to increasing shear stress during fault loading is opposed by the coupled rise in mean stress, whilst any post-failure tendency for crack closure from reduced shear stress is counteracted by the lowered mean stress. In the case of normal faults, however, development of dilatancy during loading is favoured both by the increasing shear stress and by the coupled reduction in mean stress. Post-failure, reduced shear stress and increased mean stress both
CRUSTAL STRESS, FAULTING & FLUID FLOW
77
ru/pture nucleation ~'/
dilational jog
compressional bend
~
"
~
" ~'~ compressional jog
Fig. 5. Seismotectonic carbon of an irregular rupture trace (not to scale), showing areas of enhanced (+ and cross-hatched) and reduced ( - and diagonal hachures) mean stress arising from a rupture propagating from left to right. Note that the response of isolated fault bends depends on the direction of rupture propagation. Diagram represents a map view of a strike-slip fault, or a cross-section through a dip-slip fault.
contribute to crack closure. On these arguments, dilatancy effects throughout the fault loading cycle should be more pronounced in extensional rather than compressional stress regimes. However, another factor to be considered is that at the same depth and fluid pressure level (kv value), the shear stress required for frictional reactivation of a thrust fault is about four times that needed to reactivate a normal fault (Sibson 1974). Effects from shear stress related dilatancy are therefore likely to be more pronounced in the vicinity of thrust faults. Until the details of the stress levels driving faulting and the constitutive laws for the different dilatancy mechanisms are fully resolved, it is impossible to evaluate the contributions of the various mechanisms to fluid redistribution in the crust. Of critical concern, because it affects the volume of fluid redistribution by dilatancy pumping, is the question of whether cyclic dilatational strains are generally localized to the material within fault zones, or whether they extend over broad regions of the surrounding crust. While there is considerable geological evidence for localized dilatancy within shear zones at all crustal levels (Hobbs et al. 1990), documentation of microstructures demonstrating cyclic dilatancy on a regional scale is generally lacking. Thus, the extent to which dilatancy pumping of one form or another may redistribute fluids, and in particular may draw fluids down from the near-surface to levels of the crust undergoing prograde metamorphism, as proposed by McCaig (1988), remains unresolved.
Post-seismic fluid redistribution Rupturing of an irregular fault surface leads to abrupt post-failure changes in mean stress
localised around the structural irregularities (Segall & Pollard 1980; Sibson 1989) (Fig. 5). As a consequence, there is a tendency for fluids to be redistributed from areas of raised mean stress to areas of lowered mean stress (Nur & Booker 1972). Fluids are driven out of compressional jogs and bends where mean stress increases post-failure, while the most intense fluid influx, coupled with aftershock activity, is concentrated in regions of sharply reduced mean stress such as dilational jogs and bends. At these dilational sites, rapid slip transfer during rupture propagation causes abrupt local reductions in fluid pressure below ambient (hydrostatic?) values. In some instances, the induced suctional forces may lead to rupture arrest (Sibson 1985, 1989). Dilational fault jogs and bends thus act essentially as suction pumps and are often characterized by multiply recemented wallrock breccias resulting from repeated hydraulic implosion. Larger dilational structures typically comprise a mesh of extension veins, implosion breccias, and subsidiary shears in the form of a Hill fault/ fracture mesh. Active geothermal systems are commonly localized within dilational sites in extensional and transtensional fault systems. Where transected by faults in the top kilometre of the seismogenic zone (the epithermal environment), pressure reductions from slip transfer may induce episodes of boiling within the hydrothermal system, triggering mineral deposition throughout the phase of aftershock activity (Sibson 1987). The association of epithermal mineralization with the high levels of extensional/transtensional fault systems thus arises from the coincidence of this 'boiling horizon' with the near-surface region where extensive hydrofracturing may occur under hydrostatic fluid pressure conditions (Sibson 1992a).
78
R.H. SIBSON
(TECTONIC'~ -i-~ ~LOADING'J ~
I FAULT I I INSTABILITY I I ~ = C +/~s(an-Pr)
I
• R U P T U, R IxN G
//
(PERMEABILITY) k DESTRUCTION) Self"Se aling "
/
~CREATION OF "~ rHYDROTHERMAL'~ FRACTURE Fluid Discharge PRECIPITATION Pf
Decrease
Fig. 6. Schematic representation of fault-valve activity, illustrating the dependence of frictional fault failure on the cycling of both tectonic shear stress and fluid pressure (after Sibson 1992b). Fault-valve action in o v e r p r e s s u r e d crust
From the time of Hubbert & Rubey's (1959) seminal paper on the mechanics of low-angle thrusting, it has become increasingly clear that overpressured fluids play a critical role in faulting, and may be especially important in the case of faults that are unfavourably oriented for frictional reactivation (Cello & Nut 1988; Sibson 1990a). Recent oil-field studies have demonstrated the widespread occurrence of fluid pressure compartments in sedimentary basins, and analysis of oil migration paths suggests that overpressured compartments become breached from time to time with episodic discharge of overpressured fluids (Hunt 1990; Powley 1990). There is also now good evidence that, in at least some areas, seismic rupturing is occurring within overpressured portions of the crust. Fault-valve action thus occurs wherever ruptures transect suprahydrostatic gradients in fluid pressure and breach impermeable barriers, leading to upwards fluid discharge along the transient permeability of the fault zone and local reversion towards a hydrostatic fluid pressure gradient before hydrothermal self-sealing of the rupture zone occurs, and fluid overpressures rebuild at depth. Under such circumstances, the timing of successive failure episodes is controlled by the cycling of both tectonic shear stress and fluid pressure through the interseismic period (Fig. 6). A key issue, affecting the volume of fluid available for post-failure discharge and the time for fluid replenishment of the fault zone in the interseismic period, is the extent to which fault zones are locally overpressured in relation to the surrounding crust (Fig. 7). On mechanical grounds, a case can be made that valve action is most likely to be prevalent in compressional or transpressional fault systems. Steep reverse faults are particularly likely to be
effective as fault-valves because they are severely misoriented for reactivation, and frictional shear failure can only initiate when slightly supralithostatic fluid pressures are attained (Sibson etal. 1988; Sibson 1990b). Under these conditions, arrays of subhorizontal extension fractures develop by hydraulic fracturing adjacent to the fault (hydrofracture dilatancy) prior to failure, to form a lithostatically pressured reservoir that, together with associated grain-scale dilatancy, may contain a substantial fluid volume (Fig. 8). Fault failure and rupturing through the upper crust then allows fluid from the overpressured reservoir to drain upwards, with fluid pressures dropping rapidly towards hydrostatic values. The discharge may be accompanied by phase separation (if CO2 is present), rapid mineral precipitation and hydrothermal self-sealing of the fault. Fluid pressure then rebuilds towards the critical value needed to trigger the next episode of fault slip, and the cycle repeats. Such faults may give rise to extreme fault-valve action involving the discharge of large fluid volumes, with fluid pressure potentially cycling between pre-failure lithostatic and post-failure hydrostatic levels. Extreme valve action appears responsible for many mesothermal Au-quartz lodes hosted in steep reverse or reverse-oblique faults, such as the Mother Lode vein system described previously (Sibson et al. 1988; Cox et al. 1991; Boullier & Robert 1992). A typical vein assemblage consists of flat-lying extensional veins, inferred to have developed by hydraulic fracturing prior to episodes of fault slip, in mutual cross-cutting relationships with steep fault-veins formed during episodic upwards discharge along the steeply dipping fault zone (Fig. 8). The hosting shear zones are typically of mixed discontinuous-continuous ('brittle-ductile') character and developed under greenschist
CRUSTAL STRESS, FAULTING & FLUID FLOW
FLUID PRESSURES INSIDE FAULT ZONE
I
!1
I
I
FLUID PRESSURES OUTSIDE FAULT ZONE ~'V
1.0
I
I
i
I
1.0 i i
I I
I I
79
I
m
¢¢
:i ".n~ 's
....
,,,///////~;,~,,~
I W
¢)" a
Fig. 7. Hypothetical plots of pore-fluid factor, ~,,, versus depth, illustrating pre-failure and post-failure fluid pressure gradients inside a fault zone undergoing fault-valve action in relation to possible gradients (curves A, B, C, D) in the surrounding rock mass. Curve A represents an end-member case where the pressure distributions inside and outside the fault zone are identical. Curves B, C, and D represent different degrees of localization of fluid overpressure within the fault zone, with D representing a crust that is hydrostatically pressured throughout the seismogenic zone. Conceivably, curves A-D could reflect fluid pressure gradients with increasing distance from a major fault zone.
facies metamorphic conditions at crustal levels corresponding to the base of the seismogenic zone (depths of c . 1 0 ± 5 k m ) with the vein material precipitated from low salinity, mixed H20-CO2 fluids (Robert & Kelly 1987). Localization of the prefailure arrays of hydraulic extension fractures to the vicinity of the fault zones (generally to within tens to hundreds of metres) suggests that the fault zones are locally overpressured with respect to the surrounding crust, but the degree of comparative overpressuring is unclear (see Fig. 7). Valving action leading to the formation of mesothermal vein systems may also occur at depth within strikeslip fault systems in the vicinity of compressional jogs and bends (Fig. 5). This is in direct contrast to epithermal mineralization which typically
develops in extensional/transtensional fault systems (Sibson 1992a) While extreme valve action (involving the episodic discharge of large fluid volumes and the formation of major hydrothermal vein systems) may be comparatively rare and restricted to specific tectonic settings, the frequent presence of minor syn-tectonic veins in fault zones (e.g. Chester et al. 1993), and the growing evidence from fluid inclusion studies for fluid pressure cycling (e.g. Parry & Bruhn 1990; Mullis 1990), suggest that basic fault-valve activity is widespread in the earth's crust. Even minor valving action involving small fluid volumes is likely to have a considerable effect on the nucleation and recurrence of earthquakes (Sibson 1992b). The depositional environment of mesothermal vein
80
R.H. SIBSON
~'~. ,~ ZO N E I +~'~(~1 ~'~,rupture nucleatiosinte,~!,,l,IIl,III
~3
Pfl I
"h'yd~s6tic TIME
Fig. 8. Schematic diagram of extreme fault-valve action and associated fluid-pressure cycling on a
high-angle reverse fault. Mesothermal gold-quartz lodes form in the region of large rupture nucleation and intense fluid pressure cycling near the base of the seismogenic zone.
systems also suggests that the region towards the base of the seismogenic zone may in general represent a time-dependent interface between hydrostatically and lithostatically pressured portions of the crust, as represented schematically in Fig. 9. General fault zone model The absence of a localized heat flow anomaly centred on the trace of the San Andreas strike-slip fault in California limits the timeaveraged frictional shear resistance in the seismogenic zone to < 2 0 M P a (Lachenbruch & McGarr 1990). Additional evidence for a very weak fault zone comes from the accumulation of data suggesting that the San Andreas fault is extremely unfavourably oriented for frictional reactivation within the regional stress field, lying at 65-85 ° to the greatest compressive stress (Mount & Suppe 1987). Present knowledge of the frictional characteristics of rock materials suggests that, despite the problem of containment, fluid overpressures are the most probable weakening mechanism that could account for this low frictional strength. As a consequence, a
range of models accounting for the generation and maintenance of fluid overpressures in transcrustal fault zones have recently been proposed. A key issue in the different models is whether the high fluid pressures are derived from the fault zone acting as a migratory conduit for overpressured fluids (Sibson 1990a; Stark & Stark 1991; Rice 1992), or whether the fluid overpressures are continually regenerated from essentially the same fluid volume during cyclical loading (Byerlee 1990, 1993; Sleep & Blanpied 1992). The development of extensive hydrothermal veining in fault zones, especially the gold--quartz mineralization precipitated at structural levels corresponding to the lower half of the seismogenic zone, provides geological evidence that fault zones, in at least some circumstances, are acting as migratory conduits for the passage of rather substantial fluid volumes (Kerrich 1986). If fluid overpressures and associated faultvalve action leading to fluid pressure cycling are as widespread as evidence is beginning to suggest, then fluid pressure gradients within transcrustal fault zones must be regarded as time-dependent, affecting the shear resistance profiles derived from rheological modelling. Figure 9 is an attempt to illustrate the effects on rheology and fault-strength for a transcrustal fault zone that are likely to arise from fluid pressure cycling associated with fault-valve action. Both the depth and amplitude of the peak shear resistance become time-dependent. In addition, the integrated strength of the fault zone is at a minimum pre-failure and reaches its maximum value at the end of the post-failure discharge phase, before self-sealing occurs and fluid pressure starts to reaccumulate. Some of the implications of such a model for rupture nucleation and recurrence have been explored by Sibson (1992b) Discussion While static stress fields may induce directional permeability in the rock mass, textural evidence from hydrothermal veins hosted in fault/fracture systems suggests that fluid flux through such features is commonly episodic. Moreover, it is difficult to account for the great fluid mobility in the vicinity of fault zones in terms of static stress fields and constant permeability. It appears that in many circumstances fault and fracture permeability must continually be renewed for them to remain as effective channelways. Stress cycling thus appears as the great of fluid flow in deforming upper crust, affecting flow systems driven by long-term hydraulic
modulator
CRUSTAL STRESS, FAULTING & FLUID FLOW T, °C m
FAULT ROCK METAMORPHISM
0 m
FAULT STRENGTH
0
100
200
I
I
300 MPa
FLUID PRESSURE
0
100 ~,~. I
I
200
300
400
I
I
,
~ \.~~..~~
Hydrothermal Alteration -- 100
!:i:!:i:i:i:!q:i:i:i:!:!:! Zeolite
81
.
.
.
z, km
500MPa I
0
-
hydrostatic regime
.
s -
Facies
--200 !:i:i:i:i:i:!:i:i:i:i:!:i:!
\X ~
PrehnitePumpellyite Facies
_.o --
400
iiiiiiii!iiiiiiiill iiiiiiii Greenschist Facies
_,oo ii ililiiiiiiiiiiiiilil
x
~
~
stfailure
\
\
.
\ \ prefailure
suprahydrostatic regime
lO
-
,.. .~'v < o 9• ~ ~u.,,< !
~f
-!
pos
~fithostatic regime
20-
( ~'v > 0.9 ) Amphibolite -- 600
Facies
\
25-
\ Fig. 9. General fault zone model incorporating fluid pressure cycling, assuming a linear geotherm of 25° C km-1. The extent of the different fluid pressure regimes and their relationship to metamorphic environment is conjectural. potentials such as topography, intrusive heat sources, metamorphic dewatering, etc. A considerable range of mechanisms tied to the earthquake stress cycle may contribute to fluid redistribution around fault zones. 'Suction pump' effects at dilational jogs and bends have characteristic structural associations, as have the vein assemblages that result from 'extreme fault-valve action'. However, with the exception of hydrofracture dilatancy linked to extreme valve action, the relative importance for largescale fluid redistribution of the various pumping mechanisms involving pre-failure dilatancy and post-failure fluid expulsion remains uncertain. From consideration of hydrothermal veining at the scale of individual outcrops, it may be difficult or impossible to discriminate between the different mechanisms for fluid redistribution. In evaluating the possibilities, critical account must be taken of the structural site, the mode of faulting at the time of hydrothermal deposition, the level of exposure, and the evidence for a predominantly hydrostatic or suprahydrostatic fluid pressure regime. Similar difficulties arise when attempting to attribute post-seismic discharge at the ground surface to a particular fluid redistribution mechanism (e.g. Sibson 1981; Rojstaczer & Wolf 1992; Muir Wood & King 1993). Systematic monitoring of discharge chemistry over extended time periods
may help to distinguish varying depths of origin for the discharging fluids, and allow discrimination between the different mechanisms. It is, however', becoming increasingly evident that fluid redistribution tied to the earthquake stress cycle has application to the development of fault-hosted mineralization and perhaps also to the migration of hydrocarbons in certain tectonic settings (e.g. Burley et al. 1989). Gold-quartz veins, especially, record the episodic passage of substantial volumes of aqueous fluid through fault zones and allied fracture networks (Cox et al. 1991; Boullier & Robert 1992). Regions near the top and the bottom of the seismogenic zone are favoured sites for the development of fault-hosted mineralization, with fluctuations in stress and permeability acting as triggers for episodes of hydrothermal precipitation at different stages of the fault loading cycle (Sibson 1992a). Rapid slip transfer across upwelling hydrothermal systems hosted within dilational jogs and bends in extensional and transtensional fault systems leads to abrupt pressure reductions, triggering episodes of boiling within the epithermal environment and mineral deposition throughout the phase of aftershock activity. Disseminated epithermal mineralization may also develop from the to-and-fro passage of fluids driven by cyclic loading within intensely fractured portions of
82
R.H. SIBSON
the uppermost crust. At deeper levels in compressional/transpressional fault systems, intense valving action towards the base of the seismogenic zone gives rise to high-amplitude fluid pressure cycling and the formation of mesotherreal gold-quartz lodes. There is, therefore, mounting evidence for mechanical involvement of fluids with all stages of the earthquake cycle. In the lower half of the seismogenic zone, the competition between creation and destruction of fracture permeability plays a critical role in the accumulation of fluid overpressures, affecting earthquake nucleation and recurrence, while transient fluid pressure reductions at dilational jogs and bends contribute to rupture arrest and aftershock activity (Sibson 1992b). The cyclical character of fluid pressure gradients arising from fault-valve action on seismically active faults has the implication that rheological models of fault zones and crustal shear strength profiles must also be considered time-dependent, with integrated shear strength at a minimum pre-failure, but attaining a maximum value postfailure at the end of the discharge phase. Many of the ideas expressed here have arisen through interaction and conversations over many years with N. Brown, J. Moore, H. Poulsen, N. Price, and F. Robert, while S. Cox, I. Main, J.-R. Grasso and an anonymous reviewer provided constructive criticism of the manuscript. I thank the organizers of Geofluids '93 for the financial support which made my participation in the conference possible.
References ANDERSON,E.M. 1951. The Dynamics of Faulting and Dyke Formation with Application to Britain. 2nd edn, Oliver & Boyd, Edinburgh. AVDIN, A. & JOHNSON, A.M. 1978. Development of faults as zones of deformation bands and as slip surfaces in sandstones. Pure and Applied Geophysics, 116,932-942. BATZLE, M.L. & SIMMONS, G. 1977. Geothermal systems: rocks, fluids, fractures. American Geophysical Union Monograph, 20,233-242. BETHKE, C.M. 1986. Hydrologic constraints on the genesis of the Upper Mississippi Valley mineral district from Illinois basin brines. Economic Geology, 81,233-249. BORRAIDALE, G.J. 1981. Particulate flow and the generation of cleavage. Tectonophysics, 72,306321. BOULLIER, A-M. & ROBERT, F. 1992. Paleoseismic events recorded in Archean gold-quartz vein networks, Val d'Or, Abitibi, Quebec. Journal of Structural Geology, 14, 161-179. BRACE, W.F. 1980. Permeability of crystalline and argillaceous rocks. International Journal of Rock Mechanics and Mining Science, 17,241-251.
1984. Permeability of crystalline rocks: new in situ measurements. Journal of Geophysical Research, 89, 4327-4330. BRANTLEV, S. 1992. The effect of fluid chemistry on quartz microcrack lifetimes. Earth and Planetary Science Letters, 113,145-156. BURLEY,S.D., MULLIS,J. & MATTER,A. 1989. Timing diagenesis in the Tartan Reservoir (UK North Sea): constraints from combined cathodoluminescence microscopy and fluid inclusion studies. Marine and Petroleum Geology, 6, 98-120. BYERLEE, J.D. 1978. Friction of rocks. Pure and Applied Geophysics, 116,615--626. 1990. Friction, overpressure and fault-normal compression. Geophysical Research Letters, 17, 2109-2112. -1993. Model for episodic flow of high-pressure water in fault zones before earthquakes. Geology, 21,303-306. CARRIO-SCHAFFHAUSER, E., RAYNAUD, S., LATIERE, H.J. & MAZEROLLE,F. 1990. Propagation and localisation of stylolites in limestones. In: KNIPE, R.J. & RUTTER,E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 193-199. CATHLES,L.M. 1990. Scale and effects of fluid flow in the upper crust. Science, 248,323-329. CELLO,G. & NUR, A. 1988. Emplacement of foreland thrust systems. Tectonics, 7,261-271. CHESTER, F.M., EVANS, J.P. & BIEGEL, R.L. 1993. Internal structure and weakening mechanisms of the San Andreas Fault. Journal of Geophysical Research, 98,771-786. Cox, S.F. & Etheridge, M.A. 1989. Coupled grainscale dilatancy and mass transfer during deformation at high fluid pressures: examples from Mount Lyell, Tasmania. Journal of Structural Geology, 11,147-162. ~, WALL,V.J., ETHERIDGE,M.A. & POTTER,T.F. 1991. Deformational and metamorphic processes in the formation of mesothermal vein-hosted gold deposits - examples from the Lachlan Fold Belt in central Victoria, Australia. Ore Geology Reviews, 6,391-423. CRAMPIN, S., EVANS, R. & ATKINSON, B.K. 1984. Earthquake prediction: a new physical basis. Geophysical Journal of the Royal Astronomical Society, 76, 147-156. DAS, S. & SCHOLZ, C.H. 1981. Theory of timedependent rupture in the earth. Journal of Geophysical Research, 86, 6038-6051. Du ROUCHET, J. 1981. Stress fields, a key to oil migration. American Association of Petroleum Geologists Bulletin, 65, 74-85. ETHERIDGE,M.A., WALL,V.J., Cox, S.F. & VERNON, R.H. 1984. High fluid pressures during regional metamorphism and deformation:implicationsfor mass transport and deformation mechanisms. Journal of Geophysical Research, 89, 4344-4358. FISCHER, G.J. & PATERSON, M.S. 1989. Dilatancy during rock deformation at high temperatures and pressures. Journal of Geophysical Research, 94, 17,607-17,617. FLETCHER, R.C. & POLLARD, D.D. 1981. Anticrack
CRUSTAL STRESS, FAULTING & FLUID FLOW model for pressure solution surfaces. Geology, 9, 419-424. FYFE,W.S. 1990. Plate tectonics and the hydrosphere. In: NESBITT, B.E. (ed.) Fluids in Tectonically Active Regimes of the Continental Crust. Mineralogical Association of Canada Short Course Handbooks, 18,299-312. , PRICE, N.J. & THOMPSON, A.B. 1978. Fluids in the Earth's Crust, Elsevier, Amsterdam. GOLD, T. & SOTER,S. 1984. Fluid ascent through the solid lithosphere and its relation to earthquakes. Pure and Applied Geophysics, 122,492-529. GRINDLEY,G.W. & BROWNE,P.R.L. 1976. Structural and hydrological factors controlling the permeability of some hot-water geothermal fields. In: Proceedings of the 2nd United Nations Symposium on Development and Use of Geothermal Resources. New York, 377-386. GROSHONG,R.H. 1988. Low-temperature deformation mechanisms and their interpretation. Geological Society of America Bulletin, 100, 1329-1360. HICKMAN,S.H. 1991. Stress in the lithosphere and the strength of active faults. Review of Geophysics Supplement, 29,759-775. HILL, D.P. 1977. A model for earthquake swarms. Journal of Geophysical Research, 82, 1347-1352. - - , REASENBERG,P.A. & 29 OTHERS1993. Seismicity remotely triggered by the magnitude 7.3 Landers earthquake. Science, 260, 1617-1623. HoBas, B.E., MULHAUS, H-B. & ORD, A. 1990. Instability, softening, and localization of deformation. In: KNIPE, R.J. & RUTrER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics, Geological Society, London, Special Publications, 54, 143-166. HOOPER, E.C.D. 1991. Fluid migration along growth faults in compacting sediments. Journal of Petroleum Geology, 14, 161-180. HUBBERT, M.K. & RUBEY, W.W. 1959. Role of fluid pressure in the mechanics of overthrust faulting. Geological Society of America Bulletin, 70, 115205. HULIN, C.D. 1925. Structural control of ore deposits. Economic Geology, 24, 15-49. HUNT, J.M. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. American Association of Petroleum Geologists Bulletin, 74, 1-12. IRWIN, W.P. & BARNES,I. 1980. Tectonic relations of carbon dioxide discharges and earthquakes. Journal of Geophysical Research, 8 5 , 3115-3121. KERRICH, R. 1986. Fluid infiltration into fault zones: chemical, isotopic, and mechanical effects. Pure and Applied Geophysics, 124,226-268. KNIPE, R.J. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. KNOPF, A. 1929. The Mother Lode System of California. US Geological Survey Professional Paper, 157. LACHENBRUCH,A.H. & MCGARR, A. 1990. In: WALLACE, R.E. (ed.) The San Andreas Fault System, California. Stress and Heat Flow. US Geological Survey Professional Paper, 1515, 261-277.
83
MCCAIG, A.M. 1988. Deep fluid circulation in fault zones. Geology, 16, 867-870. McCLAY, K. 1977. Pressure solution and Coble creep in rocks. Journal of the Geological Society, London, 134, 57-70. MARONE, C., RALEIGH, C.B. & SCHOLZ, C.H. 1990. Frictional behavior and constitutive modeling of simulated fault gouge. Journal of Geophysical Research, 95, 7007-7025. MORROW,C., LOCKNER,D., MOORE,D. & BYERLEE,J. 1981. Permeability of granite in a temperature gradient. Journal of Geophysical Research, 86, 3002-3008. MUIR WOOD, R. & KING, G.C.P. 1993. Hydrological signatures of earthquake strain. Journal of Geophysical Research, 98, 22035-22068. MOUNT, V.S. & SUPPE,J. 1987. State of stress near the San Andreas fault: implications for wrench tectonics. Geology, 15, 1143-1146. MULLIS, J. 1988. Rapid subsidence and upthrusting in the Northern Appenines, deduced by fluid inclusion studies in quartz crystals from Porretta Terme. Schweizerische Mineralogische und Petrographische Mitteilungen, 68,157-170. NEWHOUSE, W.H. 1942. Ore Deposits as Related to Structural Features. Princeton University Press, Princeton, New Jersey. NICHOLSON, C. & WESSON, R.L. 1990. Earthquake hazard associated with deep well injection - a report to the U.S. Environmental Protection Agency. US Geological Survey Bulletin, 1951. NORTON, D.L. 1982. Fluid and heat transport phenomena typical of copper-bearing pluton environments, south-eastern Arizona. In: TrrLEY, S.R. (ed.) Advances in Geology of the Porphyry Copper Deposits, Southwestern North America. University of Arizona Press, Tucson Arizona, 59-72. NUR, A. 1972. Dilatancy, pore fluids, and premonitory variations of t/tp travel times. Bulletin of the Seismological Society of America, 62, 1217-1222. - 1975. A note on the constitutive law for dilatancy. Pure and Applied Geophysics, 133,197-206. - & BOOKER, J.R. 1972. Aftershocks caused by pore-fluid flow? Science, 175,885-887. & WALDER,J. 1990. Time-dependent hydraulics of the Earth's crust. In: The Role of Fluids in Crustal Processes. National Academy of Sciences, Washington, DC, 113-127. OLIVER, J. 1986. Fluids expelled tectonically from orogenic belts: their role in hydrocarbon migration and other geologic phenomena. Geology, 14, 99-102. PARRY, W.T. & BRUHN, R.L. 1990. Fluid pressure transients on seismogenic normal faults. Tectonophysics, 179,335-344. POWER, W.L., TULLIS, T.E., BROWN, S., BOITNOTT, G.N. & SCHOLZ,C.H. 1987. Roughness of natural fault surfaces. Geophysical Research Letters, 14, 29-32. POWLEY, D.E. 1990. Pressures and hydrogeology in petroleum basins. Earth Science Reviews, 29, 215-226. RAMSAY,J.G. 1980. The crack-seal mechanism of rock deformation. Nature, 284,135-139.
84
R.H. SIBSON REDWINE, L. 1981. Hypothesis combining dilation, natural hydraulic fracturing, and dolomitization to explain petroleum reservoirs in Monterey Shale, Santa Maria area, California. In: GARRISON, R.E., DOUGLAS, R.G., PISOA~O, K.E., ISAACS, C.M. & INGLE, J.C. (eds) The Monterey
Formation and Related Siliceous Rocks of California. Special Publications of the Society of Economic Paleontologists and Mineralogists, 221-248. RICE, J.R. 1992. Fault stress states, pore pressure distributions, and the weakness of the San Andreas fault. In: EVANS, B. & WONG, T-F. (eds)
Fault Mechanics and Transport Properties of Rocks. Academic Press, New York, 475-503. ~& CLEARY,M.P. 1976. Some basic stress diffusion solutions for fluid-saturated elastic porous media with compressible constituents. Reviews of Geophysics and Space Physics, 14,227-241. ROBERT, F. & KELLY,W.C. 1987. Ore-forming fluids in Archean gold-bearing quartz veins at the Sigma Mine, Abitibi greenstone belt, Quebec, Canada. Economic Geology, 82, 56--74. ROBERTS, G. 1991. Structural controls on fluid migration through the Rencurel thrust zone, Vercors, French Sub-Alpine Chains. In: ENGLAND, W.T. & FLEET, A.J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59,245-262. ROJSTACZER, S. & WOLF, S. 1992. Permeability changes associated with large earthquakes: an example from Loma Prieta, California. Geology, 2 0 , 211-214. RUDNICKI, J . W . & H s u , T-C. 1988. Pore pressure changes induced by slip on permeable and impermeable faults. Journal of Geophysical Research, 93, 3275-3285. SCHOLZ, C.H., SYKES,L.R. & AGGARWAL,Y.P. 1973. Earthquake prediction: a physical basis. Science, 181,803-810. SECOR, D.T. 1965. Role of fluid pressure in jointing. American Journal of Science, 263,633-646. -& POLLARD, D.D. 1975. On the stability of open hydraulic fractures in the earth's crust. Geophysical Research Letters, 2,510-513. SEGALL, P. & POLLARD, D.D. 1980. Mechanics of discontinuous faults. Journal of Geophysical Research, 85, 4337-4350. SIBSON, R.H. 1974. Frictional constraints on thrust, wrench and normal faults. Nature, 249, 542544. -1981. Fluid flow accompanying faulting: field evidence and models. In: SIMPSON, D.W. & RICHARDS, P.G. (eds) Earthquake Prediction: an International Review. American Geophysical Union, Maurice Ewing Series, 4, 593-603. 1983. Continental fault structure and the shallow earthquake source. Journal of the Geological Society, London, 140,741-767. - 1985. Stopping of earthquake ruptures at dilational fault jogs. Nature, 316,248-251. 1987. Earthquake rupturing as a hydrothermal mineralising agent. Geology, 15,701-704.
1989. Earthquake faulting as a structural process.
Journal of Structural Geology, 11, 1-14. 1990a. Rupture nucleation on unfavorably oriented faults. Bulletin of the Seismological Society of America, 80, 1580-1604. 1990b. Conditions for fault-valve behaviour. In: KNIPE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 15-28. 1991. Loading of faults to failure. Bulletin of the Seismological Society of America, 81, 2493-2497. 1992a. Earthquake faulting, induced fluid flow, and fault-hosted gold-quartz mineralization. In: BARTHOLOMEW, M.J., HYNDMAN, D.W., MOGK, D.W. & MASON, R. (eds) Basement Tectonics 8:
Characterization and Comparison of Ancient and Mesozoic Continental Margins. Proceedings of the 8th International Conference on Basement Tectonics, Butte, Montana, Kluwer, Dordrecht, 603-614. 1992b. Implications of fault-valve behaviour for rupture nucleation and recurrence. Tectonophysics, 21 l, 283-293. 1993. Load-strengthening versus load-weakening faulting. Journal of Structural Geology, 15, 123-128. ~, MOORE, J.McM. & RANKIN, A.H. 1975. Seismic pumping - a hydrothermal fluid transport mechanism. Journal of the Geological Society, London, 131,653-659. ~, ROBERT, F. & POULSEN, K.H. 1988. High-angle reverse faults, fluid pressure cycling and mesothermal gold-quartz deposits. Geology, 16, 551555. SLEEP, N.H. & BLANPtED, M.L. 1992. Creep, compaction, and the weak rheology of major faults. Nature, 359,687-692. SMITH, D.A. 1980. Sealing and non-sealing faults in Louisiana Gulf Coast Salt Basin. American Association of Petroleum Geologists Bulletin, 64, 145-172. SMITH, D.L.& EVANS, B. 1984. Diffusional crack healing in quartz. Journal of Geophysical Research, 89, 4125-4135. SPRUNT, E.S. & NUR, A. 1977. Destruction of porosity through pressure solution. Geophysics, 42, 726741. STARK, C.P. & STARK,J.A. 1991. Seismic fluids and percolation theory. Journal of Geophysical Research, 96, 8417-8426. TORGERSON, T. 1990. Crustal-scale fluid transport: magnitude and mechanisms. LOS, Transactions of the American Geophysical Union, 71, l, 4, 13. 1991. Crustal fluid flow: continuous or episodic.
LOS, Transactions of the American Geophysical Union, 72, 18--19. WALDER,J. & NUR, A. 1984. Porosity reduction and crustal pore pressure development. Journal of Geophysical Research, 89, 11,539-11,548. WOOD, S.H., WURTS, C., BALLENGER, N., SHALEEN, M. & TOTORICA,D. 1985. The Borah Peak, Idaho, earthquake of October 28, 1983: hydrologic effects. Earthquake Spectra, 2,125-150.
Earthquakes, strain-cycling and the mobilization of fluids R. M U I R W O O D
E Q E , Newbridge House, Clapton, Gloucestershire GL54 2 L G , UK Abstract: Employing empirical observations of the hydrological changes that follow major
earthquakes, it becomes possible to predict subsurface fluid flows that accompany tectonic activity. Coseismic hydrological changes are found to be critically dependent on the style of fault displacement. Normal faults involve post-seismic compressional elastic rebound and displace large volumes of fluid from the crust: rainfall equivalent discharges have been found to exceed 100 mm close to the fault and remain above 10 mm at distances greater than 50 km; the total volume of water released in two M7 normal fault earthquakes in western USA was 0.3-0.5 km 3. In contrast, reverse fault earthquakes involve extensional elastic rebound that increases crustal porosity, drawing fluids into the crust. The magnitude and distribution of the water-discharge for normal fault earthquakes has been compared with deformation models calibrated from seismic and geodetic information, and found to correlate with the crustal volume strain down to a depth of up to 5 km. These results suggest that fluid-filled cracks are ubiquitous throughout the brittle continental crust, and that these cracks open and close through the earthquake cycle. By repeatedly drawing fluids into microcracks distributed throughout a very large volume of rock, strain-cycling can achieve very significant chemical fractionation. Different classes of tectonics impose particular configurations of fractures and according to the origin of fluid recharge, lead to a range of different types and styles of mineralization. Seismic strain-cycling can also produce primary and secondary oil migration within a single causative mechanism.
A study has been undertaken to collect, and where possible quantify, hydrological changes accompanying numerous (more than 200) earthquakes in many different parts of the world (Muir Wood & King 1993). In order to resolve such changes it is necessary to focus on areas where there is hydrogeological communication from the surface to depth. Such communication is most obvious where there are high permeability deep-sourced conduits, releasing thermal springs, but has also been demonstrated more widely from certain types of reservoirinduced seismicity in which earthquakes are triggered by a raised fluid pressure at the surface (Simpson et al. 1988). (Focusing on areas in which there is no barrier between surface water and fracture flow deep in the upper crust inevitably disqualifies the collection of observations from most sedimentary basins, and other platform areas overlain by low-permeability sediments.) Within this study it was found that once coseismic hydrological effects could be isolated from an inevitable groundwater 'noise', a simple relationship could be demonstrated between the style of fault displacement and its hydrological 'signature' (Muir Wood & King 1993). In most igneous and metamorphic rock, as well as well-lithified sediments, mobile water is held
and transported in fractures. The response of these fracture systems to the coseismic strain field of different classes of earthquakes indicates that porosity is responding to the prevailing stress-state, as is implied by numerous experimental studies (Walsh 1965; Batzle et al. 1980). In a region undergoing extensional faulting, continuing strain distributed through the crust causes appropriately oriented high-angle fractures to dilate, thereby increasing crustal porosity. Pore-pressures are sustained by slow infilling of the dilated fractures with fluid (Fig. 1). At the time of a major normal fault rupture strain, formerly distributed through the crust, becomes concentrated on the fault. As a result the region of the crust that 'surrounds' the fault (ie. the footwall and hanging wall) undergoes elastic rebound in compression. In contrast, in a region undergoing compressional tectonic deformation, in the interseismic period negative strain (volume decrease) in the crust closes high-angle fractures and reduces crustal porosity. At the time of the fault rupture, as strain is transferred into fault displacement the surrounding crust undergoes elastic rebound in extension. Hence in the region around a normal fault rupture the coseismic decrease in crustal porosity should lead to the postseismic expulsion of
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin Sedimentary Basins, Geological Society Special Publication No. 78, 85-98.
85
86
R. MUIR WOOD
NORMAL FAULTING !
ti,, i.......
I
b)
t
Interseismic extension
Coseismic compressional elastic rebound
REVERSE FAULTING
d)
c)
Interseismic compression
Coseismic extensional elastic rebound
Fig. 1. For extensional faulting the interseismic period (a) is associated with crack opening and increase of effective porosity. At the time of the earthquake (b) cracks close and water is expelled. For compressional faulting the interseismic period (c) is associated with crack closure and the expulsion of water. At the time of the earthquake (d) cracks open and water is drawn in.
water. In the region surrounding a compressional fault rupture expanded porosity reduces hydraulic pressures leading to water being drawn into the crust. In both normal and reverse fault earthquakes the impact on the height of the water-table (and consequent well-levels and superficially sourced spring flows) will be dependent on how effectively this water-table is connected to the fracture-flow system at depth. River flows provide the most important resource for quantifying these post-seismic changes; by sampling changes in discharge, unrelated to precipitation, across catchments typically >100 km 2 in area, it becomes possible to average the crust on a scale whose lateral dimension is of the same order as the hydraulically conductive thickness of the crust. River discharge data can then be normalised for the area of the drainage basin to achieve a 'rainfall equivalent' discharge,
either in the form of velocity (for a daily average) or cumulative linear 'thickness'. N o r m a l v. r e v e r s e f a u l t s
For the purposes of illustration the form of these hydrological changes can be seen from the contrast between two major, surface-rupturing, high-angle dip-slip earthquakes: the 1959 M7.3 (normal) Hebgen Lake (Montana) earthquake and the 1896 M7.2 (reverse) Rikuu (North Honshu, Japan) earthquake. Many other comparable examples of normal and reverse fault earthquake hydrological effects, as well as strike-slip, oblique slip and subsurface faultrupturing events can be found discussed and mapped in Muir Wood & King (1993). (a) Three river flow profiles (with entirely separate catchments) are shown in Fig. 2 from
MOBILIZATION OF FLUIDS IN EARTHQUAKES
18 Madison river 16
m3per
rain
14 12 10 September 19.59
August
: ~ier~
river
August
30
O c l ~ Days
r~n
September 1959
l
October Days
Gailatin river
1 o I,.............................. August
, ............................. September 1959
,.................... October
1
EVays
Fig. 2. River flow data for three rivers in the vicinity of the 17 August 1959 Hebgen Lake, Montana earthquake, in the days around the time of the earthquake. Average monthly flows plotted as thin line.
87
discharges, are shown in Fig. 3. By summing the individual catchments the cumulative volume of water discharged in the region around the earthquake is equivalent to c. 0.5 km 3. (b) The 31 August 1896 Rikuu (N. Honshu) earthquake involved reverse fault surfacerupture over a distance of 36kin, with a maximum uplift of 3.5 m. This was accompanied by an antithetic reverse fault rupture located some 15 km to the east and in the hanging wall of the main fault. Hence all the near-surface faulting appears to have been reverse, implying that all near-surface elastic rebound was extensional. No changes in river flows were reported following the earthquake, but hot springs supplying bath-houses at Oshuku, Tsunagi and Osawa dried up after the earthquake, while there was a significant reduction in flow at the thermal springs at Namari and Yuda (Yamasaki 1900 and see Fig. 4). All of these lie in the hanging wall of the main fault, to the east and within 20 km of the principal antithetic fault. All the hot springs eventually re-appeared, indicating that the water-flow was temporarily recharging the additional porosity in the crust that followed the earthquake. In contrast, close to the northern end of the main fault, between the main fault and its antithetic fault, a permanent new hot spring formed at Sengan-Toge following the earthquake.
Coseismic strain models the region of southeast Montana for the period before and after the 18 August 1959 Hebgen Lake earthquake (data from USGS Water Supply Reports and Stermitz 1964). For both sets of plots average flows (observed over a period of 15 years surrounding the year following the earthquake) are shown for comparison. From the daily plots it can be seen that in all three rivers a peak in flow arrived within four days following the earthquake. There was no precipitation around this time to explain such a surge and the increase in flow that follows the earthquake was sustained relative to the trend of the monthly averaged flow curve for more than 200 days. From expected flow curves for these rivers (average flows calibrated from rivers within the same region but outside the influence of the earthquake) it is possible to assess the overall volume of the water release in each catchment. This appears to show an almost linear decay following the initial peak (Muir Wood & King 1993). Typical decay times to half peak flow are 100-150 days. These cumulative flows, normalised for the individual drainage basins around the fault into rainfall equivalent
In order to examine in more detail the expected magnitude and extent of hydrological effects, strain models of coseismic deformation have been generated using a boundary element program in which a dislocation element is introduced into an elastic medium (for details of the model see King & Ellis 1990). Figure 5 is an illustration of the dilational strain changes that accompany a normal fault displacement, both in cross-section (Fig. 5a) and plan-view at two depths in the crust (Fig 5b and c). The predominant strain changes are compressional. Strain changes of reverse fault earthquakes are largely equivalent although of opposite sign to those of normal fault earthquakes. Models for simple dip-slip earthquakes always reveal 'strain-shadows' of opposite sign beyond the ends of the fault (see Fig. 5b and c). There is some evidence from both normal and reverse fault earthquakes for hydrological effects reflecting such opposite signed strain effects in these locations (Muir Wood & King 1993). The creation of the new hot-spring at the end of the Rikuu earthquake fault corresponds with a
88
R. MUIR WOOD
cumulative 'rainfallequivalent'discharge < l mm
~
~
1-20mm
~
fault scarp ~ ' - River
20-40mm
~
[
I >40mm
catchment
Fig. 3. Cumulative excess flow following the Hebgen Lake earthquake normalised for the area of the catchments into excess rainfall equivalent (ram).
strain-change anticipated to cause an expulsion of water.
Near-fault strain The above account concerns regional, or 'farfield', coseismic strain. However, close to the fault, irregularities in fault geometry, and variations in displacement, begin to dominate the strain-field. The most notable effects accompany jogs where displacement passes from one fault plane to a parallel one (see Sibson 1986). As jogs can have a dimension of metres to kilometres, and for major earthquakes involve the transfer of several metres of displacement, self-evidently they can be subject to very significant levels of strain; potentially higher
than 10 -2. This strain can be either compressional or dilational. Within the jog the coseismic stress changes do not simply accompany elastic rebound but instead are loaded directly by adjacent fault rupture. Along some strike-slip and oblique-slip fault ruptures, where jogs communicate with the surface, these effects may cause fountain outbursts of water (as in the 28 October 1983 earthquake at Borah Peak, Idaho; Waag & Lane 1985) or a collapse of the water-table (as along parts of the Fairview Valley, Nevada fault in the 16 December 1954 earthquake; Zones 1957).
The properties of crustal drainage In the case of the Hebgen Lake earthquake the
MOBILIZATION OF FLUIDS IN EARTHQUAKES
89
se°g o ~'~ k
T
gi"
~-JOshuku ~ ,
39°30 N ~~~Yuda ~ .
~'-X~sa w ' ~ -~_- ~ '
40km',.. 141 E O
m,
1
New hot spring Decrease in flow of hot spring
O Hot spring ceased flowing
river J fault-scarp
Fig.4.
Impact on hot-springs of the 31 August 1896 Rikuu (North Honshu) M7.2 reverse fault earthquake. Barbed lines are surface-fault reverse ruptures; barbs on upthrown side of the fault.
observed discharge for a traverse perpendicular to the strike of the principal fault rupture has been compared (see Muir Wood & King 1993) with predicted discharge for a two dimensional coseismic strain model, calibrated from known parameters of fault displacement, dip and depth for the earthquake. The closest fit is with the 5 km depth prediction both in general shape and in amplitude (Fig. 6). This comparison suggests that the water that emerges at the surface is associated with fracture systems extending to considerable depth. From the fit with the model it would appear that either all the volumetric strain down to a depth of 5 km is accommodated by crack closure, or else fracture connectivity extends to greater depths and only part of the strain results in a reduction in porosity. The simplest way to describe the form of the observed flows is to assume that strainlocalization and drainage are accomplished in
the same series of planar uniform cracks, open at the surface and closed at some depth, subject to a pressure change over a depth range (Muir Wood & King 1993). The crack width is found to be the parameter most strongly controlling the decay of the flow v. time profile while the dead depth (depth to the top of the section of the crack undergoing a sudden strain) and the crack width together determine the rise time. Most of the observations are fit by models with crack widths of about 0.03 mm and effective dead depths of about 2 km. For such characteristic crack apertures, extending to a depth of 5 kin, cumulative rainfall equivalent discharges of 20-40mm imply crack separations less than 4-8 m.
Implicationsof the empiricalobservations These observations offer (effectively for the first time) the potential to explore the macroscale
90
R. M U I R W O O D -60.00
a)
,o® ~
-60 O0 -60.00
-30.00
0.0
-30.(~)
0.0
30.(X)
60.(X)
~iiiiiiiiiiiiiiiiiiiii!i~:::i :~....
..:.~:~:;:~:;~:,~:,:,:~:,:~ii~,iiiiiiliiiiiiiiiiiil ~
30.00
-60,00
6000
[Voi-iirne
Strains at ~ = - : - _ ~ the Surface
3000
30.00
0.0
60.00
[Strains ~ ~ [at 5km ~ ~ l ] ~
- -
-30.00
....-~it!iiiiiil
-30.o0
O0
0.0
30.00
30.00
60,00
60.00
b)
.....
c)
Fig. 5. Cross-section (a) and plan-views at the surface (b) and at a depth of 5 km (c) of volumetric strains modelled around a normal fault. The scale units are kilometres; slip for all the models is I m. Strain levels are shaded from a background grey to lighter shades (negative strain) and to darker shades (positive strain): strain-steps of 2 x 10 -5.
SO0
450 400
1959 Hebgen Lake earthquake
Total 'rainfall equivalent' discharge: mm
A
350
I
~ 10 km
predicted from
3OO 250
2oo~ 150 100 50 0
,
-50
-80
-60
, distance from faui! rupture:, km -40
-20
0
20
,
,
1
40
60
80
Fig. 6. Observed (1959-1960) and predicted cumulative rainfall-equivalent discharges for a cross-section perpendicular to the strike of the Hebgen Fault. A constant 6 m displacement has been modelled on the fault dipping at 45 ° and extending to 15 km, consistent with the 17 August 1959 earthquake. Solid lines are predicted from the areal strain summed to depths of 2, 5 and 10 km.
MOBILIZATION OF FLUIDS IN EARTHQUAKES
91
Mohr-Coulomb failure criterion
a)
Fig. 7. Stress-cyclingin extensional (1) and compressional (2) tectonic regimes: (a) pre-seismic, (b) post-seismic.
properties of crustal drainage. The fluvial discharge information ultimately constrains the minimum depth of the percolation threshold (the depth below which all fluids are trapped under lithostatic pressure), the aperture of the characteristic fractures that dominate crustal drainage, and the minimum separation of such fractures. The comparability of these parameters for earthquakes in different parts of the world suggests that they are a fundamental property of the upper continental crust undergoing extension. Hence these parameters are likely to be relevant to the conditions that prevailed when episodes of rifting were occurring early in the life of many sedimentary basins. From empirical hydrological data it is not so easy to quantify the same key parameters of crustal drainage and porosity changes associated with a region involved in active compressional tectonic deformation. As deeply circulating water generally makes up a small contribution to surface discharge, a reduction in this component is unquantifiable in river flows, although notable in thermal springs. A higher ambient horizontal stress will reduce average fracture apertures, closing a greater portion of the walls of fractures, sealing fluid flow, and thereby causing a greater spacing between open flowing fractures. As the stress cycling that accompanies reverse faulting never reduces horizontal stresses below the highest levels encountered (post-seismically) in a compressional environment (see Fig. 7), there is inevitably a lower proportion of strain accommodated in the opening and closing of fractures. This means that porosities, the change in porosity in response to a strain increment and vertical permeabilities will all be lower than in an extensional tectonic environment.
S u m m a r y of principal findings These empirical findings have a number of important implications for the understanding of tectonically induced fluid flows in the upper crust. In particular they provide a synoptic, holistic and actualistic view of the sudden changes in hydrogeology that accompany earthquakes. In the past, theories of coseismic changes in hydrogeology have almost all been based on observations collected at individual exhumed outcrops that record multiple episodes of fluid movement accompanying numerous earthquakes millions of years in the past. (i) The character of fluid mobilization is primarily determined by the tectonic style of fault rupture. Normal fault earthquakes are predominant in expelling fluids. In contrast, reverse fault earthquakes cause a sudden increase in crustal porosity, although where a fault is blind, compressional deformation in an overlying anticline may also create lesser magnitudes of fluid discharge. This contrasts with much of the literature on seismic pumping that has assumed that reverse fault earthquakes are the most significant cause of fluid expulsion (see for example Sibson 1990). Part of the confusion in the literature can be traced back to the widely quoted discharge of water that followed the oblique-slip (reverse-sinistral) Kern County, California earthquake of 1952 (Briggs & Troxell 1955). However, when mapped this post-seismic discharge is found to be concentrated at the ends of the fault and in the anticlinal hanging wall, while close to the fault-rupture wells went dry. The pattern of discharge very closely mimics the oblique-slip strain-field modelled for this style of fault displacement (Muir Wood & King 1993).
92
R. MUIR WOOD
(ii) While 'seismic pumping' is an important component of these changes, as was highlighted by the seminal paper of Sibson et al. (1975), the 'seismohydraulic' alterations in hydrogeology through the seismic strain-cycle also affect permeabilities and porosities, as well as the orientations and apertures of fractures involved in fluid flow. Most of these changes are transient but in the vicinity of the fault can also be permanent (Muir Wood 1993). (iii) The volume of displaced fluid represents a significant proportion of total coseismic strain, at least down to the depth at which open fractures are connected. This volume is large; equivalent to a giant oilfield in each Magnitude 7 normal fault earthquake. (iv) The consistency that can be found between models of coseismic strain and the pattern of water release gives confidence in being able to predict the quantity and distribution of fluid displacement for a particular size, depth and style of fault-rupture. (v) Following a major normal fault earthquake the fluid is displaced from a wide region, with a radius typically greater than the length of the fault. For a rift-bounding fault this could cover the full width of the rift. This again challenges a widely held view to be found in the literature, that fluid flow following earthquakes is largely or completely concentrated on the fault that underwent rupture (see again Sibson 1990). Faults are most likely to concentrate flow in poorly-lithified low permeability sediments (see Behrmann 1991; Knipe et al. 1991), or perhaps at great depths in the crust. However in the top few kilometres of an upper crust dominated by fracture flow, if an earthquake was described solely in terms of its hydrological effects at the surface the causative fault itself would probably not be resolved. Hydrologically the earthquake reflects the volumetric strain-field, rather than the volume-conserving process of fault rupture. (vi) These hydrological effects demonstrate the fundamental role of strain-cycling in the crust and the intimate relationship between volumetric strain and the displacement of fluids. The close match between predicted strains and the volume of displaced water is only possible if the upper crust is filled with interconnected microcracks, that represent the dynamic porosity of the rock. (vii) The close connection between strain and fracture porosity proves that a larger, although far slower, change in crustal porosity will accompany a transformation in the tectonic regime. This porosity change will be most marked in the most extreme transition in horizontal stresses between an environment of
active extension to one of active compressional deformation, or vice versa.
Tectonic controls on fracture dilation
Each tectonic environment is accompanied by a specific style of strain cycling, manifest by the opening and closing of appropriately oriented fracture systems. As a result, individual tectonic styles of deformation are likely to generate particular patterns of fracture mineralization, at different stages of the strain cycle (Fig. 8). (a) In an extensional tectonic environment high-angle fractures undergo dilation as a result of continued horizontal strain. Following fault rupture these high-angle cracks partially close, expelling fluid from the rock. The cycle then passes once again into interseismic dilation. As these fractures are located in a stress-regime of reduced ambient horizontal stresses, they are likely to be highly strain-sensitive and hence to undergo relatively large changes in aperture. (b) In a region of compressional deformation any high-angle cracks become progressively closed in the interseismic period, squeezing fluids into near-horizontal cracks (orthogonal to the minimum compressional stress direction) where, if the fluid cannot escape, pressures will be lithostatic. At the time of an earthquake extensional elastic rebound will dilate highangle fractures, releasing fluid pressures in the low angle fracture set. High-angle fractures will continue to contract through the ensuing seismic cycle. High ambient horizontal stresses mean that changes in the aperture of high angle fractures will be significantly smaller than those encountered in extensional tectonic regimes. (c) Strike-slip tectonic environments tend to be dominated by lengthy fault systems, in which individual episodes of fault rupture do not extend along the whole length of the fault. In the vicinity of such faults high-angle fractures, will be subject to alternating episodes of compressional and dilational strain according to the relative disposition of the individual episodes of fault rupture. (d) Jogs can be either compressional (antidilational) or dilational in nature, and can accompany any class of faulting (Sibson 1986). Volume strain accompanying sudden displacement across a compressional jog can raise pore pressures within fractures sufficient to cause spontaneous hydrofracture, creating new, but transient, high-aperture fracture paths through the rock. In contrast, sudden displacement across dilational jogs will create large voids and can cause such a significant reduction in pore
MOBILIZATION OF FLUIDS IN EARTHQUAKES
Inter-seismic
Post-seismic
~
t -'l~ a)
93
pattern of high-angle and low-angle fractures (Fig. 8b) encountered in some gold deposits (see Boullier & Robert 1992) has been formed in a single compressional tectonic environment as also were the similarly oriented (but far shallower) bitumen vein deposits encountered in the Argentinian Andes (Carey & Parnell 1993). Implosion fault breccias, developed in dilational jogs on reverse faults (Fig. 8d), are another important source of gold mineralization (see for example Liu 1990).
Seismohydrauliccontrols of mineralization
iI iI
b)
c)
Fig.8.
Views in section of idealized fracture dilation accompanying seismic strain-cycling in a range of tectonic environments: (a) extensional, (b) compressional, (c) dilational jog, (d) compressional jog. fluid pressures that the rock implodes, generating some form of megabreccia (Sibson 1986). It is apparent that observations of fracture systems in many exposed rocks are describing one or other style of strain-cycling associated with active tectonics at some time in the past. Crack-seal vein infillings are found in many different tectonic environments (Ramsay 1980; Parry et al. 1988) although the relationship of mineralisation to the seismic cycle will vary from one style of nearby faulting to another. One can also predict that the magnitude of individual strain-cycles recorded by crack-seal veins will be determined by the depth of burial of the rock and the prevailing tectonic regime. Some tectonic environments can generate two orientations of fractures at different parts of the strain-cycle. For example the characteristic
Hydrothermal mineralization involves processes whereby some typically low-solubility ionic species, distributed throughout the rock at low concentrations, becomes enriched and concentrated in or around conduits for fluid flow (Sharp & Kyle 1988). Strain-cycling provides an important agent for concentrating mineralization as it repeatedly allows a small volume of water to come into contact with a large volume of rock. Strain-cycling also re-opens fractures, creating new voids in the rock that can form the site of further precipitation. The fundamental differences in strain cycling between extensional and compressional tectonic environments, as well as the most extreme strain effects in jogs, are all likely to have a profound impact on the conditions and styles of mineralization. Of course active tectonics implies the creation of new topography, and the disruption of thermal equilibrium through uplift and subsidence as well as the generation of high conductivity fracture systems, all of which will have a profound influence on fluid flows. However none of these macro-scale influences on fluid flows affects the micro-scale in the manner of strain-cycling. While the principal role of strain-cycling is in repeated fluid infiltration, post-seismically it may also promote significant fluid movements, which may be of particular consequence where thermally and gravitationally driven fluxes are weak or of opposite direction. In the environment of normal faulting, the sudden collapse in porosity that follows fault rupture causes fluid to be expelled from microcracks into arterial cracks connected to the surface. The upward displacement of fluids leads to a reduction of temperature reducing the solubility of most ionic species and thus leading to post-seismic precipitation. In a compressional tectonic environment the rock becomes infilled with fluid following an earthquake and slowly decreases in porosity between earthquakes. As a generality fluid is most likely to rise in the crust,
94
R. MUIR WOOD
and hence undergo a temperature reduction leading to precipitation, towards the end of the interseismic period. A most important determinant of the chemistry and nature of mineralization in all tectonic environments will be the source of the fluid that moves into the rock during the course of the strain cycle. If this fluid comes from deeper levels it will be saturated with dissolved ionic species and will precipitate within the rock; if it arrives from shallower levels then it will have the greater capability to dissolve and remove ionic species out of the rock. Much will depend on the particular hydrogeological context, although this in turn may be dominated by the tectonic location. For example in the subsiding hanging wall of a normal fault, where thick sequences of sediments accumulate, fluids rich in dissolved ions are produced from sediment compaction. The post-seismic reduction in fracture porosity could be the trigger for the discharge of overpressured fluids, as in low permeability sedimentary sequences on continental margins (Hunt 1990). However in the uplifted footwail, recharge is more likely to be dominated by meteoric surface recharge, low in dissolved ions and hence able to leach the rock in the course of strain cycling (Bruhn et al. 1990; Reynolds & Lister 1987). Repeated flushing of water over the 1000-10000 seismic cycles that make up the typical tectonic activity of a major fault, will dramatically alter rock chemistry, as has been noted from granites adjacent to strike-slip faults (Stierman 1984) and from deep shear-zones (Wayne & McCaig 1992). In a compressional tectonic environment, compaction of footwall sediments overridden by the fault may provide a source of fluids. Fluid is squeezed out between earthquakes, producing such phenomena as mud-volcanoes. Compressional activity and resulting uplift may also drive fluids laterally into adjacent foreland basins (Bethke & Marsak 1990). The sudden opening of high-angle fractures in normal and strike-slip faults overlying accretionary wedges can also allow overpressured fluids to escape upwards through hydrofracture in the immediate post-seismic period (Behrmann 1991). However the phenomenon of seismic valving of reverse faults (Sibson et al. 1988), whereby the fault moves in response to an increase in fluid pressures, and then acts as a conduit to fluid flow post-seismically, appears to be a process that only exists deep in the crust, as there is no empirical evidence for water coming out of reverse faults at the surface. In an environment of dip-slip faulting the principal post-seismic hydraulic gradient is near
vertical, approximately parallel to the fault. However across a strike-slip fault, in which volumes of rock subject to compression and dilation are located on opposing sides of the fault, pronounced fluid pressure gradients can develop to encourage horizontal flow through the fault.
Petroleum migration As noted above, in order to be able to map the hydrological consequences of an earthquake it is necessary to study regions in which there is continuity between water stored in fractured crustal rocks, and surface-discharging aquifers. However over a significant proportion of the continents, and in particular the continental shelves, there are extensive sedimentary cover sequences of very low permeability, which prevent, at least in the short-term, any connection between basement fluid pressures and near-surface aquifers. In such environments, if pore-pressure changes cannot be relieved through flow or recharge from the surface, then fluid flow resulting from interseismic and coseismic strain changes will be lateral, either within the fractured crust, fractured low permeability sediments or within some overlying aquifer or aquifers typically near the base of the sedimentary cover. The impact of this strain cycling will be determined by the composition of the available fluid, and the degree to which the system is interconnected with the surface. In the presence of suitable organic material the fluid will inevitably include hydrocarbons. Earthquakes have been observed to mobilize hydrocarbons as in the 12 May 1802 earthquake in north Italy, after which petroleum was procured from fissures near Bardi (Mallet & Mallet 1854). Clearly, if petroleum can be released at the surface post-seismically, it is also being remobilized subsurface. Petroleum generation demands the accumulation of organic-rich sediments and burial to sufficient depths to generate hydrocarbons. These hydrocarbons then have to be liberated from the source rock and migrate to some reservoir formation in which they become trapped and concentrated. The process of primary liberation from the rock, and secondary expulsion to some external reservoir has been much debated (England et al. 1987). The principal problem in achieving primary migration arises from the difficulty of hydrocarbons escaping from fine-grained sediments against very high inter-granular capillary pressures. Secondary migration requires that a
MOBILIZATION OF FLUIDS IN EARTHQUAKES
a)
low permeability source rock
b)
I . . ~ m
Fig. 9 (a) During positive (dilational) strain, fluids are drawn into widening microfractures accomplishing primary migration; (b) during negative (compressional) strain, hydrocarbon fluids are pumped from microfractures into high-permeability channels to accomplish secondary migration.
sufficient concentration of hydrocarbons has accumulated, to allow the fluid to become mobilized as a column. It is not difficult to see how seismic strain cycling provides a potentially important mechanism for achieving both primary and secondary migration as the result of a single process (Fig. 9). Perhaps the simplest way in which to avoid the problems of capillary resistance to primary migration is to generate new microfracture void space in the rocks in which hydrocarbons can accumulate. In well-lithified sediments such microfractures will form as a response to strain-cycling in response to active tectonics. Strain-cycling allows microcracks to open and close repeatedly, drawing hydrocarbons out of the rock matrix, and then pumping them to a larger fracture system or high permeability sedimentary units. What is the evidence for such processes? Microfractures containing either bitumen, or a diagenetic mineral-fill such as calcite, are widespread in source rocks (Duppenbecker et al. 1991). Lindgreen (1987) described diagenetic
95
mineral-bearing microfractures from within the Kimmeridge Clay of the Central Graben of the North Sea with apertures of up to 0.1 mm, oriented both parallel to the bedding and at high angles, in which primary oil migration was accomplished. The orientation of such microfractures should reveal the particular tectonic regime in which they were generated. The role of tectonics in some secondary migration in the North Sea was also demonstrated by Leonard (1984) who found that a faulted connection between source and reservoir gave a fourfold increase in the chance of exploration success in chalk reservoirs in the southern Norwegian sector of the North Sea. However, once liberated from the source-rock, strain-cycling can pump hydrocarbons through pre-existing fracture systems or any suitable high permeability sedimentary formation, not just along a neighbouring active fault. In parts of the North Sea there is also evidence for Tertiary hydrothermal activity, likely to have been tectonically driven (Wayte 1993). Pulsed episodes of fluid flow, driven by strain-cycling, are likely to have the most significant impact on the migration of brines and hydrocarbons, when tectonic activity is resumed fairly late in the evolution of a sedimentary basin. In the Tertiary history of the North Sea basin there was a prominent episode of midPalaeocene extension and late Eocene-early Oligocene compressional wrench faulting (Muir Wood 1989) concurrent with an important phase of petroleum migration (Barnard & Bastow 1991). For the latter phase of compressional (or transpressional) deformation in the rift areas perhaps 100 strain-cycles are implied by the cumulative displacements of these mid-Tertiary structures, with amplitudes typically of about 200 m (see for example Pegrum & Ljones 1984).
Conclusions
Seismic strain-cycling is akin to the function of the lungs in breathing. In order to accommodate distributed strain, interconnected fracture systems must permeate throughout the rock-mass and thereby repeatedly allow a small volume of fluid to be brought into contact with a large area of rock. It is the area of the vessels involved in the infiltration and expulsion that gives the process such enormous potency. The crudest model of 'characteristic' fractures applied to the hydrological consequences of the Hebgen Lake earthquake implies a total area of the vessels from which fluid is expelled of around 10 million km 2. Repeated fluxing of fluids through the
96
R. MUIR WOOD
crack systems that permeate the brittle rockmass has the power to fractionate and concentrate many low-solubility ionic or molecular species distributed in the upper crust. Following normal fault displacement, fluids tend to migrate away from the vicinity of the fault. For the same size earthquake occurring in a sedimentary basin the volumes of displaced fluid can be predicted to be comparable to those found emerging at the surface where there is no impermeable cover (as at Hebgen Lake, involving an estimated 0.5 km 3 of fluid migrating laterally from a region 10000 km 2 in area). The flux on the margins of such an area, entirely covered by a low permeability overburden, would be more than 1000 m 3 of fluid for every metre length of perimeter; in a 10m thick aquifer with a 10% dynamic porosity, fluid would move 1 km. Following a reverse fault, displacement fluid flow will be drawn towards the fault. On the assumption that the strain 'efficiency' (porosity change divided by volumetric strain) of a reverse fault earthquake is at least 20% that of a normal fault, an M7 earthquake would draw in perhaps 100000000 m 3 of fluid. Where hydraulic conductivities in sediments lying on the crystalline basement are significantly higher than in the bedrock itself this will encourage large scale flow with the lighter fluids becoming concentrated in the sedimentary section above the reverse fault. Blind, non-outcropping reverse faults beneath thick sedimentary basins are also some of the most efficient at creating anticlinal hydrocarbon reservoir geometries. The flows created by the large-scale and sudden changes in pore-pressures accompanying seismic strain-cycling may oppose the prevailing fluid flow regimes created by gravitational or thermally-driven fluid flows. Tectonic activity could therefore drive fluids into reservoirs that might otherwise be discounted as potential drilling targets. Employing coseismic and interseismic strain models within basin analysis it should be possible to predict the magnitude and extent of these repeated episodes of fluid flow. Coseismic and interseismic strain changes could also help to explain the development of some zones of over- and under-pressurization within confined fluid reservoirs, where other explanations associated with consolidation secondary porosity reduction or temperature changes are found to be insufficient (Domenico & Palciauskas 1988). In a medium where porosity is dominated by fractures (or connected with such a medium) sealed reservoirs have the potential to act as strain-barometers recording
changes in dynamic porosity through an alteration in fluid pressure (Bodvarsson 1970). The impact on pore-fluid pressures will be determined by the dynamic porosity relative to the strain-sensitivity of the fractured medium. As with Earth-tides the most sensitive response is found in a low dynamic-porosity medium in which the fluid is distributed in strain-sensitive fracture systems. Some confined aquifers are known to produce centimetric changes in welllevels from 10 -s strain Earth-tides (Bredehoft 1967). For a confined low dynamic-porosity medium a reduction in volume of 10-4-10 -5, consequent on seismic strain-cycling, has the potential to raise pore pressures by several bars. Where the rock experiences higher levels of strain in response to the imposition of a new tectonic regime, then an even more marked change in pore-fluid pressures is possible. For example, extreme under-pressurisation can be expected in a confined, low-porosity fractured reservoir where a compressional tectonic regime passes to one of extension. Such a process could partly explain the under-pressurization of the oil-reservoirs at Kimmeridge, Dorset, England (Selley & Stoneley 1987) where mid-Tertiary compressional deformation has passed into Plio-Quaternary extension. Strain models in this paper were produced by G. King at IPG, Strasbourg, whose collaboration is gratefully acknowledged.
References BARNARD, P. C. & BASTOW, M. A. 1991. Petroleum generation, migration, entrapment and mixing in the central and Northern North Sea. In: ENGLAND,W. A. & FLEET,A. J. (eds) Petroleum Migration. Geological Society, London, Special Publications 59,167-190. BATZLE,M. L., SIMMONS,G. & SIEGFRIED,R. W. 1980. Microcrack closure in rocks under stress: direct observation. Journal of Geophysical Research, 85, 7072-7090. BEHRMANN,J. H. 1991. Conditions for hydrofracture and the fluid permeability of accretionary wedges. Earth and Planetary Science Letters, 107, 550558. BETHKE, C. M. & MARSHAK,S. 1990. Brine migration across North America - the plate tectonics of groundwater. Annual Review Earth and Planetary Science Letters, 18,287-315. BOt)VARSSON, G. 1970. Confined fluids as strain meters, Journal of Geophysical Research, 75, 2711-2718. BOUILLIER, A-M. & ROBERT,F. 1992. Palaeoseismic events recorded in Archaean gold-quartz vein networks, Val d'Or, Abitibi, Quebec, Canada. Journal of Structural Geology, 14,161-179.
MOBILIZATION OF FLUIDS IN E A R T H Q U A K E S BREDEHOFT, J.D. 1965. Response of well-aquifer systems to earth tides. Journal of Geophysical Research, 72, 3075-3087. BRIGGS, R. C. & TROXELL, H. C. 1955. Effect of the Arvin-Tehachapi earthquake on spring and stream flow. In: Earthquakes in Kern County California, during 1952. Bulletin of the California Division of Mines, 171, 81-98. BRUHNN, R. L., YONKEE,W. A. & PARRY,W. T. 1990. Structural and fluid-chemical properties of seismogenic normal faults. Tectonophysics, 175, 139-157. CAREY, P. F. & PARNELL,J. 1993. Solid bitumen veins related to overpressuring and thrusting in the Neuquen Basin, Argentina. In: PARNELL, J., RUFFELL, A. H. & MOLES, N. R. (eds) Geofluids '93, Torquay, May 4-7, 1993, 181-185. DOMENICO, P. A. & PALCIAUSKAS,V. V. 1988. The generation and dissipation of abnormal fluid pressures in active depositional environments. In: BACK, W., ROSENSHEIN, J. S. & SEABER, P. R. (eds) Hydrogeology. Geological Society of America, The Geology of North America, 0-2, 435-445. DUPPENBECKER, S. J., DOHMEN, L. & WELTE, D. H. 1991. Numerical modelling of petroleum expulsion in two areas of the Lower Saxony Basin, Northern Germany. In: ENGLAND, W. A. & FLEET, A. J. (eds) Petroleum Migration. Geological Society, London, Special Publications, 59, 47-64. ENGLAND, W. A., MACKENZIE,A. S., MANN, D. M. & QUIGLEY, T. M. 1987. The movement and entrapment of petroleum in the subsurface. Journal of the Geological Society (London), 144, 327-347. HUNT, J. M. 1990. Generation and migration of petroleum from abnormally pressured fluid compartments. Bulletin of the American Association of Petroleum Geologists, 74, 1-12. KING, G. C. P. & ELLIS, M. 1990. The origin of large local uplift in extensional regions. Nature, 348, 689-692. KNIPE, R. J., AGAR, S. M. & PRIOR, D. J. 1991. The microstructural evolution of fluid flow paths in semi-lithified sediments from subduction complexes. Philosophical Transactions of the Royal Society of London, A355, 261-273. LEONARD, R. 1984. Generation and migration of hydrocarbons on the Southern Norwegian Shelf (Abstract). Bulletin of the American Association of Petroleum Geologists, 68,796. LINDGREEN,H. 1987. Molecular sieving and primary migration in Upper Jurassic and Cambrian claystone source rocks. In BROOKS,J. & GLENNIE, K. (eds) Petroleum Geology of North West Europe. Graham & Trotman, London, 357-364. LIu, E. 1990. The metalloctectonics in Eastern Shandong gold metallogenetic province, China. Terra Nova, 2,257-263. MALLET, R. & MALLET,J. W. 1854. Third reporton the facts of earthquake phenomena. Reports of the British Association for the Advancement of Science.
97
MUIR WOOD, R. 1989. Fifty million years of 'passive margin' deformation in North West Europe. In: GREGERSEN, S. & BASHAM, P. W. (eds) Earth-
quakes at North-Atlantic Passive Margins: Neotectonics and Postglacial Rebound. Kluwer,
-
Dordrecht, 7-36. 1993. Neohydrotectonics. Zeitschrift fur Geomorphologie, 94,275-284. & KING, G. C. P. 1993. Hydrological signatures of earthquake strain, Journal of Geophysical Research, 98, 22035-22068. PARRY, W. T., WILSON, P. N. & BRUHN, R. L. 1988. Pore-fluid chemistry and chemical reactions on the Wasatch normal fault, Utah. Geochimica et Cosmochimica Acta, 52, 2053-2063. PEGRUM, R. M. & LJONES, Z. E. 1984. 15/9 Gamma Gas Field, offshore Norway, new trap type for North Sea Basin with regional structural implications. Bulletin of the American Association of Petroleum Geologists, 68, 874-902. RAMSAY,J. G. 1980. The crack-seal mechanism of rock deformation. Nature, 284, 135-139. REYNOLDS, S. J. & LISTER, G. S. 1987. Structural aspects of fluid-rock interactions in detachment zones. Geology, 15,362-366. SELLEY,R. C. & STONELEY,R. 1987. Petroleum habitat in South Dorset. In: BROOKS, J. & GLENNIE, K. (eds) Petroleum Geology of North West Europe, Graham & Trotman, London, 139-148. SHARP, J.M. JR & KYLE, J.R. 1988. The role of ground-water processes in the formation of ore deposits. In BACK, W., ROSENSHEIN, J.S. & SEABER, P.R. (eds) Hydrogeology. Geological Society of America, The Geology of North America, 0-2,461-483. SmsoN, R. H. 1986. Brecciation processes in fault zones: inferences from earthquake rupturing. Pure and Applied Geophysics, 124, 159-173. 1990. Conditions for fault-valve behaviour. In: KNIPE, R. J. & RUTTER, E. H. (eds) Deformation Mechanisms, Rheology and Tectonics, Geological Society, London, Special Publications, 54, 15-28. ~, ROBERT, F. & POULSEN, K. H. 1988. High-angle reverse faults, fluid pressure cycling and mesothermal gold-quartz deposits. Geology, 16, 551555. SIMPSON, D. W., LEITH, W. S. & SCHOLZ,C. H. 1988. Two types of reservoir induced seismicity. Bulletin of the Seismological Society of America, 78, 2025-2040. STERMITZ, F. 1964. Effects of the Hebgen Lake earthquake on surface water. United States Geological Survey Professional Paper, 435-L, 139150. STIERMAN, D. J. 1984. Geophysical and geological evidence for fracturing, water circulation and chemical alteration in granitic rocks adjacent to major strike-slip faults. Journal of Geophysical Research, 89, 5849-5857. WAAG, C. J. & LANE, T. G. 1985. The Borah Peak, Idaho earthquake of October 28, 1983 - structural control of groundwater eruptions and sediment boil formation in the Chilly Buttes area. Earthquake Spectra, 2, 151-168. -
98
R. MUIR WOOD
WALSH, J. B. 1965. The effect of cracks on the compressibility of rocks. Journal of Geophysical Research, 70,381-389. WAYNE, D. M. • MCCAIG, A. M. 1992. Fluid flow during thrusting: a Sr isotope study. In: KHARAKA, Y. & MAEST, A. S. (eds) Water-Rock Interaction. Balkema, Rotterdam, 1559-12664. WAYTE,G. J. 1993. Evidence for the cross-formational movement of hydrothermal fluid within overpressured cells: - an example from the Triassic of the U.K. Central North Sea. In: PARNELL, J.,
RUFFELL, A. H. & MOLES, N. R. (eds) Geofluids '93, Torquay, May 4-7, 1993, 329-332. YAMASA~a, N. 1990. Das grosse japanische erdbebden im nordlichen Honshu am 31 August 1896. Petermanns Mitteilungen, 46,249-255. ZONES, C. P. 1957. Changes in hydrologic conditions in the Dixie Valley and Fairview Valley areas, Nevada, after the earthquake of December 16, 1954. Bulletin of the Seismological Society of America, 47,387-396.
Microstructural and microchemical consequences of fluid flow in deforming rocks R.J. KNIPE & A.M. McCAIG
Department o f Earth Sciences, The University, Leeds LS2 9JT, U K Abstract: The different deformation mechanisms possible in rocks impact differently on
fluid flow processes because of the range of induced volume changes associated with different deformation histories. Microstructural analysis of deformed rocks can provide crucial information for the identification of fluid flow pathways, determination of the physico-chemical properties of the fluid, quantification of the amount of fluid involved and an assessment of the variation in the open/closed nature of the fluid flow system. A critical factor in the efficiency of fluid flow in deformed rocks is the competition between the processes which maintain the connectivity of the high permeability pathways and those which close such pathways. The range of deformation processes which are involved in this competition will be different depending on the tectonic settings, the deformation conditions and the rock types involved. A brief review of the processes which interact to control fluid flow during deformation in sedimentary basins, crystalline basement under low to moderate grade metamorphism and during prograde metamorphism at moderate to high grades is given.
Identification of the pathways available for fluid flow in rocks which experience deformation, together with a detailed understanding of the interactions between fluid flow and deformation mechanisms, are important to the quantification of a number of geological processes. Recent research has highlighted that both deformation and fluid flow are likely to be localized and to occur during transient events (Sibson 1980, 1990; Carter et al. 1990; Cox et al. 1991; Sleep & Blanpied 1992; Byerlee 1993; Knipe 1993). Such complex behaviour provides an additional complication to the quantification of fluid flow processes and restricts modelling of fluid flow in rocks. The main objectives of the study of palaeofluid flow events in rocks are: • identification of the critical flow pathways along which the majority of fluid flow occurred; • determination of the physico-chemical properties of the fluid (e.g. pressure, temperature, composition and the variation of these properties with time); • characterization of the physical properties of the flow path (porosity, permeability, and time-variation of these in relation to deformation); • quantification of the amount of fluid involved (time-integrated flux) and the fluid flow rate
(flux);
•
determination of the scale characteristics of the flow event (dimensions of the volume
involved) and variations in the open/closed nature of the fluid flow system; • assessment of the relationships between fluid-induced reactions, cementation, alteration patterns and the above. There are three main types of evidence which can contribute to the characterization of palaeofluid flow events through rocks. (1) Direct observation of fluid compositions and densities from fluid inclusions (Roedder 1984; Parry & Bruhn 1986; Banks etal. 1991). (2) Physical evidence from the presence of features which require fluid flow, for example veins, cementation, stylolites and fluid inclusion trails (Ramsay 1980; Lespinasse & Cathelineau 1990). (3) Chemical evidence for flow, including vein and cement compositions, metasomatic changes in rock composition, zoning patterns of minerals and cryptic veining in recrystallised rocks (e.g. Beach 1976; Etheridge et al. 1984; Kerrich 1986; McCaig & Knipe 1990; McCaig et al. 1990). Physical indicators help to identify fluid flow pathways and are essential for assessing the mechanisms of flow, while chemical evidence is required to demonstrate that significant flow has occurred. Assessment of fluid flow characteristics and the interactions between deformation and flow can be aided by the analysis of microstructural features generated during fluid
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins,
Geological Society Special Publication No. 78, 99-111.
99
Scanning proton microprobe (SPM)
Laser ICPMS
Ion microprobe (SIMS)
Laser microprobe stable isotope analysis
Electron microprobe
Transmission electron microscope (TEM, STEM) Cathodoluminescence (CL)
Secondary electron SEM
Backscatter scanning electron microscopy (BSEM)
Optical microscopy
Technique
Rapid in situ trace element determinations at ppm level. Small spot size (1 txm) and little beam spreading in sample. Spot analyses, linescans and element maps. Non-destructive with little sample damage.
Contrast depends on trace element variations, and/or defect structure. Gives good information in monomineralic rocks (e.g. Quartz veins in quartz). Best contrast from hot cathode devices (e.g. SEM) Major element compositional variations on a 5 ixm scale. Vein compositions, zoning patterns. Good backscatter SEM on modern instruments. In situ analysis of O, C, and S isotopes in minerals. Sensitivity good, especially for carbonates. Rapidly becoming a routine technique. Isotopic zoning has been related to fluid flow. In situ trace element and isotopic measurements in minerals. Good sensitivity in favourable circumstances. Spatial resolution 2-10 i~m. In situ trace element determinations on minerals. Rapid and cheap. Detection limits c. 0.5 ppm.
First stage of most studies. Excellent mineral phase contrast. Chemical variations in carbonates enhanced by staining. Porosity identified with coloured resin. Ultrathin sections required in fine grained rocks. Information on microstructures from birefringence contrast. Universal stage gives quantitative information on crystal fabrics, fracture orientations etc. Good mineral phase and compositional contrast. Compositional variations in isotropic minerals. Information on intragranular microstructure in orientation contrast mode, linked to lattice orientation through electron channelling. Good analytical, element mapping and image analysis facilities on modern instruments. Photographic collages allow comparison of fine detail over significant areas. High resolution of surfaces. Good for details of porosity structure and occlusion in rocks affected by diagenesis and hydrothermal processes. Fracture surfaces can be imaged, and etching can reveal microstructural details. Resolution < 1 Ixm. Lattice defects imaged. Compositional variations detectable on sub p~m scale.
Uses
Table 1. Techniques f o r microstructural and microchemical analys&
Many relevant references, e.g. Wintsch & Knipe (1983); McCaig (1984); McCaig & Knipe (1990); Selverstone et al. (1992). Sharp (1992); Elsenheimer & Valley (1992); Chamberlain & Conrad (199l); Dickson et al. (1991). Giletti & Shimizu (1989); Lyon & Turner (1992). Detection limit c. 0.1% for most elements.
Fraser (1990); Wogelius et al. (1992); Jamtveit et al. (1993).
Jackson et al. (1992).
McLaren (1991) for a general review; Knipe (1982); Hippler & Knipe (1990); Philippot & Van Roemund (1992). Burley etal. (1989); Yardley & Lloyd (1989); Lloyd & Knipe (1992).
Specimen preparation difficult. Only small areas can be imaged at once, so harder to relate to other techniques. No luminescence in many situations (e.g. Fe-rich minerals). Origin of contrast sometimes not clear.
Minimum spatial resolution 100 p~m with CO2 laser. Better with Nd-YAG laser, but some minerals cannot be analysed. Many complicating factors make analyses non-routine, and difficult for certain isotopes and matrices. Minimum pit size 20--40 I~m. Lack of homogeneous mineral standards can make calibration difficult. Expensive. Data reduction still time consuming. Fine grained samples require special preparation because of beam penetration (up to 40 p~m).
e.g. papers in Marshall (1987); Knipe (1992, 1993); Evans (1990).
Many relevant references, e.g. Ramsay (1980); McCalaig (1984); Knipe (1982); Brodie & Rutter (1985); papers in Marshall (1987); Cox & Etheridge (1989); Lespinasse & Cathelineau (1990); Evans (1990). Lloyd (1987) for review of techniques. Papers in Marsh all (1987); Prior (1988); McCaig & Knipe (1990); Agar (1990); Knipe etal. (1991), and this paper for examples of application to fluid flow.
References relevant to fluid flow
Compositional information mainly from crystal form. Difficult to relate directly to optical microscopy.
Specimen damage currently limits minerals suitable for electron channelling. Resolution limited to c. 1 ~m. Will not detect veining by same phase. (e.g. quartz in quartz).
Limited scale of observation. Composition information limited. Isotropic materials cannot be imaged under crossed polars.
Limitations
MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW flow and deformation (e.g. McCaig & Knipe 1990; Knipe et al. 1991; Knipe 1993). A variety of microstructural and microchemical techniques are now available which can be applied to the study of fluid flow during deformation (Table 1) and the success of this approach is enhanced by the ability of modern research microscopes to combine different imaging signals from the same features. New techniques such as the scanning proton microprobe and laser fluorination allow small scale isotopic and trace element variations to be detected, and are likely to be increasingly applied to fluid-rock interaction problems. This short paper reviews the different deformation mechanisms possible in rocks and highlights some of the progress made in understanding the different ways in which deformation influences fluid flow. The objective is to provide a brief overview of recent research and to direct the reader at some relevant literature.
Influence of deformation mechanisms on fluid flow In this paper we are not concerned with the large scale driving forces for fluid flow but with the identification and quantification of fluid flow influenced by deformation. The general direction of flow will always be down the hydraulic gradient, whether this results from the buoyancy associated with thermal convection, from overpressures due to sedimentary or tectonic loading, or from some form of stress-related dilatancy. Deformation can influence fluid flow in two ways. (1) Strain-related changes (increases and decreases) in the host rock porosity or pore geometry will create changes in pore pressures and the pattern of fluid pressure gradients. (2) Permeability can be enhanced or restricted by the deformation (e.g. enhanced by the opening of fractures or grain boundary voids, and restricted by pressure solution, compaction and cementation). These influences are not independent. For example, the opening of an array of fractures may cause a local reduction in the fluid pressure, generating a steep pressure gradient into the dilatation site and enhancing fluid flow potential. However, the pressure drop may also lead to precipitation of cements which will reduce the permeability (cf. Etheridge et al. 1984). In the following sections we outline the consequences of the operation of different deformation mechanisms on permeability and fluid flow and give
101
examples on how microstructural analyses can be used to study these processes. Three sections are presented below, each one corresponding to one major mode of deformation.
Fracture, cataclastic flow, and independent particulate flow Fracture, cataclastic flow and independent particulate flow are the dominant deformation mechanisms in rocks and sediments in the higher, cooler levels of the crust (Groshong 1988; Mitra 1988; Knipe 1989; Lloyd & Knipe 1992). There is little doubt that fractures dominate the permeability of large parts of the upper crust, particularly in crystalline rocks at low pressure and temperature. Excellent reviews of the fracturing processes in rocks have been given by Atkinson (1987) and Scholz (1990), and a recent collection of papers in Evans & Wong (1992) reviews some of the transport properties of fault zones. The fracture strength of materials increases with increasing effective confining pressure ( 0 - c f f = 0"3 - - PH20)while the flow stress needed to activate plastic deformation decreases with increasing temperature. Hence fracture processes generally give way to plastic flow with increasing depth (Sibson 1986; Knipe 1989). However, it is now apparent that there is a considerable depth range over which fracture and ductile flow processes are both involved in deformation events (Knipe 1989, 1990; Lloyd & Knipe 1992). This may be associated with the alternate operation of deformation mechanisms during the repetitive deformation events of fault growth. In addition, brittle fracture, which is associated with rapid fracture propagation after low strains, is only one of a large variety of fracture processes. Slower crack propagation, below the critical stress needed for brittle fracture, can occur either with the aid of a concentration of plastic deformation near the crack tip or by stress corrosion processes by fluids at crack tips (see review in Scholz 1990). In general, the deviatoric stress required for fracture propagation is reduced as the pore fluid pressure is increased, since this reduces the effective pressure. Where the pore fluid pressure exceeds the least principal stress by an amount equal to the tensile strength of the material, hydraulic fracture will occur. Although the controls on these mechanisms are known the recognition of the different fracture propagation processes from microstructural evidence remains an important goal for future natural fracture mechanics.
102
R.J. KNIPE & A.M. McCAIG
Fluid flow through a fractured rock requires both the fracture array to have a suitable geometry and some threshold conditions to be exceeded in order for the array to be linked into a connected fluid migration pathway (Guegen et al. 1987; Main et al. 1990; Odling 1992; Stark & Stark 1991). Fracture/fault populations normally appear to be fractal with respect to the number of features of different size or displacement magnitude (Cowie & Scholz 1992; Gillespie et al. 1992). However, analysis of fluid flow through a fracture array also requires knowledge of the individual geometrical characteristics of the linked pathways, as not all the fractures/faults present need form part of the linked network at the time of flow, and not all the deformation features in the finite array need have formed during the same deformation event. For these reasons, establishing the history of fracture opening and identifying the types of fluids associated with fracture arrays is an important component of the analysis of fluid flow. Microstructural studies can provide information for such analysis. Most fractures which may have contributed to fluid flow in rocks are now represented by either healed micro-fractures, identified by planar secondary fluid inclusion arrays, or by veins (Lespinasse & Cathelineau 1990; Sibson et al. 1988; Boullier & Robert 1992). Except in rare cases of replacement, veins represent a finite strain event which will have increased the local porosity of the rock. The effect on the permeability and fluid flow in the rock depends on the amount, extent and duration that the fractures are open and interconnected. Repeated fracturing, dilatation and sealing by cementation can generate large and extensive veins, but may only be associated with small individual dilatation events (see Knipe & White 1979; Ramsay 1980). Other veins contain large crystals which have grown into a fluid filled cavity of considerable size (1 cm or more). The growth and maintenance of such cavities is an interesting mechanical problem (i. e. what size of open fracture can be maintained by the inherent rock strength under the imposed stress conditions?) although the development of such cavities is favoured by high fluid pressures which hold open the fractures. The presence of veins does not necessarily require either high fluid pressure or large scale fluid flow. For example the quartz vein arrays described by Knipe & White (1979) are associated with pressure solution seams which could have provided the quartz precipitated in the veins. Although some fluid flow must have taken place there is no need to invoke large scale flow
as fluid-enhanced diffusion processes could have dominated over advection. The demonstration of large scale flow through a vein array requires geochemical evidence. For example, Wayne & McCaig (1992) described isotopic data from carbonate veins within Cretaceous carbonates and mylonites from beneath the Gavarnie Thrust in the central Pyrenees. Radiogenic fluid stored in Triassic redbeds forming the basement to the thrust system was introduced into mylonites along the thrust plane, causing a general increase in the 87Sr/86Sr ratios in the deformed carbonates away from seawater-derived values. Within both undeformed and mylonitized carbonates, veins are almost invariably enriched in 8VSrcompared with the adjacent wall rocks, demonstrating that the fluid which precipitated the vein material must have equilibrated with more SVSrenriched rocks further back along the flow path. The minimum distance for the fluid transport in this case is 1.5 km, which is much larger than the average diffusion distance. Flow must have been unidirectional, with fluid moving always from more altered to less altered mylonites, and outwards into less altered wall rocks (Wayne & McCaig 1992). Abundant deformed, recrystallized veins within the mylonites suggest that the majority of fluid flow occurred during fracture events, but that these alternated with periods of quasi-plastic flow. An example of how fractures can affect fluid movement in a mylonite is illustrated in Fig. 1, which shows part of a plagioclase porphyroclast from a mylonite from Andorra in the centraleastern Pyrenees. The mylonite was generated primarily by plastic deformation associated with dynamic recrystallization at a temperature of approximately 400 ° C. The plagioclase porphyroclast (Anlv) contains more than 100 veins of more calcic plagioclase (up to An35). Plagioclase of this composition is stable in adjacent more mafic layers in the mylonite zone (see McCaig & Knipe 1990). The veins demonstrate that fluid moved across the layering for distances of at least a few millimetres. Cross-cutting relationships demonstrate that several generations of fractures are present (Fig. lb), and some of the larger veins clearly formed from several crackseal fracture events. The vein network must represent the product of between 10 and 150 discrete fracturing events, depending on how many separate veins formed during each event. In the dynamically recrystallized matrix of the same rock, trials of isolated Ca-enriched plagioclase grains have been interpreted as the remnants of similar veins crossing the matrix (McCaig & Knipe 1990). This demonstrates two
30 ~m I
I
Fig. 1. (a) Backscattered SEM montage of a plagioclase porphyroclast (An16) from a mylonite in the central Pyrenees (cf. McCaig & Knipe 1990). The porphyroclast contains multiple veins of more Ca rich plagioclase (up to An35). Veins are not present in the surrounding matrix, which consists of fine recrystallized quartz, plagioclase and muscovite. (b) Detail of an area in the lower left of (a), showing three generations of cross-cutting veins of increasingly Na-rich composition. The final generation is a marginal zone or pressure shadow to the porphyroclast.
104
R.J. KNIPE & A.M. McCAIG
3 !~!} 2;:
~
1
Fig. 2. Backscattered SEM images from a fractured dolomitic layer within mylonites beneath the Gavarnie Thrust, central Pyrenees (McCaig et al. 1993). (a) Ankerite and calcite vein in dolomite (dark phase at top edge). Note euhedral ankerite crystals built up by multiple crack-seal events (arrows), and relict screens of dolomite within ankerite. (b) Dolomite veined by Fe-dolomite and calcite, with cavity fill texture at lower left. This is evidence for storage of fluid within fracture-controlled porosity within the dolomite layer. Darkest phase is quartz. important aspects of fluid flow and deformation in mylonites. Firstly, the deformation involves an alternation of fracture and plastic flow, and secondly, fracturing of the stronger phases can create local fluid reservoirs. Knipe et al. (1991) also highlighted the importance of the creation of transient micro-fluid reservoirs for fluid flow in deforming aggregates. In the case of a mylonite the dilatation and veining associated with the hard minerals will be controlled by; (a) the amount and distribution of the hard phase, and, (b) the ability of the matrix to absorb strain created by the fracturing of the hard phases. Higher concentrations of hard phases and lower temperatures will enhance the potential for
larger, more extensive and interconnected fracture arrays for fluid flow in the myionite. On a mesoscopic scale, more competent layers within a deformation zone can also provide a reservoir for storing fluid in fractures. Figure 2 illustrates intense veining and cavity-fill textures in a dolomitic layer within calcite mylonites along the Gavarnie Thrust (McCaig et al. 1993). Complex fluid movement patterns are indicated by a wide variety of vein compositions from calcite to ankerite. Cataclastic flow involving repeated fracturing, grain size reduction and frictional grain boundary sliding is characteristic of deformation in high level fault zones (Blenkinsop & Rutter 1986; Knipe 1989; Sleep & Blanpied 1992). The movement of grains during deformation is accommodated by fracturing triggered by stress concentrations at grain contacts, and by dilatation. Independent particulate flow dominates the deformation of poorly consolidated sediments and fault gouges at low effective confining pressures, where strain is achieved by the sliding of grains past one another (Borradaile 1981). In both these cases, most deformation is likely to be associated with short periods of high strain rate. These transient pulses of deformation will induce periods of dilatation during which both the permeability and the porosity will be enhanced; i.e. the system is more open and fluid flow rates can be high. These periods will be separated by periods of low strain-rate when the system may be effectively closed. Although the general effects and consequences of these alternations in fluid communication are known, the time periods when the system is open, the decay rate of the permeability after the events and the volume affected by deformation events of different magnitude are largely unknown (see discussion in Knipe 1993). It is also unclear what controls the relative importance of fluid flow associated with the linking of fracture/fault zone dilatation over extensive areas or, fluid flow related to the slow migration of a strain (dilatation) wave which effectively transports packets of fluid with it (see Knipe et al. 1991, Fig. 4 for discussion). An attempt to assess the permeability changes which occur during deformation (i.e. quantification of the dynamic permeability) of poorly consolidated sediments has been made by a series of recent experiments (Stephenson et al. this vol.). The paper by Stephenson et al. (this volume) demonstrates the complexity of permeability variations during deformation and again highlights the ability of a fault zone to alternately store and expel fluids. This dual behaviour is also likely on a larger scale where
MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW segments of fault zones such as dilatational jogs may provide transient storage volumes for fluids (see Sibson 1990, 1993).
Diffusive mass transfer Strain associated with the removal of material from grain boundaries and interfaces experiencing high stresses gives rise to a range of mechanisms grouped together as diffusive mass transfer (Spiers & Schjutjens 1990). The processes involved can be separated into; (a) dissolution processes at the source, (b) migration of material along some pathway, and (c) the precipitation of material at sinks. The migration pathway may be a grain boundary, a thin film of fluid at grain boundaries or a pore fluid. The complexity of these processes is reflected in the range of flow laws which have been constructed for different combinations of source, migration and sink processes (see Knipe 1989 for discussion). Diffusive mass transfer (DMT) is generally associated with a loss of porosity and permeability due either to the removal of material by dissolution or by cementation in pore space. These processes are important to the creation of fault seals or barriers which restrict fluid flow (Knipe 1992, 1993). Because the distribution of such barriers and the timing of their development are important to fluid flow, an understanding of the processes and mechanisms involved in seal evolution is also important. Knipe (1993) has emphasized the role of diffusive mass transfer as a late stage process, which can be enhanced in fault zones where grain size reduction by cataclasis has already occurred. That is, the last stages of movement in a fault zone act to close the high permeability window created by dilatation associated with faulting. The cements which arise from precipitation in sealed fault zones do hold evidence (mineralogy, zoning and isotopic composition) which can aid the identification of the source of fluids involved in any fluid flow (e.g. Burley et al. 1989; Knipe 1993). The time periods when fluid flow is enhanced after large (>M6) earthquakes can be assessed from stream discharge and are usually measured in months (Muir Wood & King 1993).
Dislocation creep Dislocation creep is an intragranular deformation mechanism which relies on the movement of crystal defects (dislocations) to create crystal plasticity. The strain rate is largely dependent on the climb of dislocations which is diffusion controlled. The mechanism dominates the deformation of the middle to lower crust and
105
is a crucial component in the generation of fine grained mylonitic fault zones (Sibson 1986; Knipe 1989). Dislocation creep is not normally associated with enhanced fluid flow because any 'voids' or fractures present at the start of the deformation will lead to stress concentrations which should lead to local straining and closure of such features. However, two aspects of dislocation creep behaviour can be important to the potential for fluid flow. (a) Fluid pressures in 'voids' or fractures during dislocation creep may be close to the confining pressure so that fractures under certain conditions may propagate relatively easily (see below). (b) Under conditions where dislocation creep does not/cannot produce all the strain rate required by the deformation conditions or if there is significant grain boundary sliding, then additional strain accommodation processes may create dilatation and so enhance fluid flow. This dilatation may take the form of 'void' generation on extensional grain boundaries, where strain compatibility cannot be maintained between grains by dislocation creep. Such behaviour may be associated with deformation along a decreasing temperature and/or confining pressure path or related to a transient period of high strain-rate (see Knipe 1990) and is likely to be characteristic of deformation near the transition from quasiplastic to elastico-frictional deformation. The result is an opening of grain boundaries and the ability of fluids to penetrate the deforming aggregate. There is a gradation in the amount of porosity which can be created by this process and the effect on fluid flow will depend upon the deviation from the conditions where 'steady state' deformation by dislocation creep is possible or where strain accommodation by diffusive mass transfer prevents the creation of extended and linked fluid migration pathways. At one end of the spectrum is the behaviour where deformation conditions induce only a small deviation away from the situation where 'void' formation and grain boundary dilatation is suppressed. In this situation small, isolated but migrating grain boundary 'voids' and tubes are created which link only occasionally. The movement velocity, distribution and linking probability of the fluid inclusions will control any fluid flow. It is likely that fluid movement is via the slow migration of micro-packages of fluid which are only linked with other inclusions for short time periods. Larger deviations in the deformation conditions away from those which suppress 'void' development will lead to the creation of more extensive
106
R.J. KNIPE & A.M. McCAIG
arrays of inclusions which are linked for longer periods and therefore are more efficient fluid migration pathways. Under these conditions fluid flow may be associated with pervasive flow through the deforming aggregate although not all the boundaries need to be open at any one time. That is, while fluid-rock interactions may affect all grains, the active fluid migration pathway through the aggregate is not constant but changes as previously active routes are closed or cemented up. This dynamic creation and destruction of fluid migration pathways, where only a small percentage of the grain boundaries act as linked pathways, is important as it allows the aggregate to maintain its strength. This dynamic situation will have a specific range of conditions (strain-rates, stress levels, grain boundary migration rates and temperatures) for specific materials and will give way at lower temperatures and higher strain rates to fracturing. The microstructures which are characteristic of the migration of fluids along selective grain boundaries include the following.
(a) Reaction products from fluid-rock interactions along the selective pathway. McCaig & Knipe (1990) gave an example of this from a mylonite in the Pyrenees where the generation of new plagioclase tracked the movement of a fluid away from a possible source. The particular value of such information is that the direction of flow can be identified from the pattern of selective alteration.
Fig. 3. Backscattered SEM orientation contrast image of dynamically recrystallized calcite from a deformed vein within mylonites beneath the Gavarnie Thrust, central Pyrenees (McCaig et al. 1993). The main source of contrast is differences in lattice orientation within a monomineralic calcite aggregate (Lloyd 1987), but note also the truncated zoning patterns within recrystallized calcite grains. The similar sequence of zones everywhere in the vein indicates pervasive introduction of chemical components (probably Fe) into the grain boundaries during grain growth or grain boundary migration. Note also the microporosity along grain boundaries. Statistical analysis (L-Y. Gong, unpublished data, 1993) shows that the pores are concentrated on grain triple points, and on low stress grain boundaries in the simple shear regime (shear was dextral in the view shown, with the shear plane parallel to the top edge of the photograph).
(b) Precipitation of new phases at microdilatation sites along fluid migration pathways. Extension across grain-boundaries created by the strain incompatibilities between grains may allow the precipitation of new cements in equilibrium with the invading fluid. Examples of this include the development of fine phyllosilicates along commonly orientated grain boundaries in mylonites. McCaig (1987) described the precipitation of chlorite-quartz intergrowths in dilatant shear bands within a phyllonite from the Pyrenees. Figure 3 presents another example of pervasive fluid flow through a mylonite. The deformed calcite vein in the example shown is from mylonites associated with the Gavarnie Thrust in the Central Pyrenees (McCaig et al. 1993). The dynamically recrystallized aggregate present is typical of deformation by dislocation creep accompanied by dynamic recrystallization. The crystallographic orientation contrast images (see Lloyd 1987) also reveal the presence of small scale zoning patterns, which probably arise from minor variations in Fe content (these zoning patterns were not identifiable by spot
micro-chemical analysis in the microprobe). The zoning patterns are frequently truncated by adjacent grain boundaries, probably because of subsequent grain boundary migration. In the monomineralic aggregate these microchemical variations can only be produced by the introduction of new chemical components from an external source. Because the zoning patterns are similar all over the vein the microstructures suggest either growth on selected extensional grain boundaries throughout the aggregate during deformation, or pervasive introduction of components into migrating grain boundaries during dynamic recrystallization. The latter might occur by grain boundary diffusion away from transiently open fluid channelways. The concentration of grain boundary 'voids' on extensional grain boundaries (L-Y. Gong, unpublished data) supports this suggestion. The deformation mechanism path for the mylonite probably involved a transition from dislocation creep, which generated the fine grained aggregate, to one where diffusive mass transfer and
MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW grain boundary dilatation became more important.
Discussion The above review has outlined the range of fluid flow patterns which may be generated in rocks undergoing deformation by different mechanisms. It is clear that deformation can add high permeability flow paths to the pervasive or bulk flow possible through the undeformed rock porosity. The additional flow pathways associated with deformation are however transient and their effect on fluid flow depends on the permeability evolution of the fabrics in the tectonites, the distribution of strain associated with deformation and the ability of the deformation to link domains, layers or compartments with different fluid pressures. Knipe et al. (1991) recognized three aspects of deformation in fault zones which are important to the future assessment of fluid flow. (a) The deformation mechanism paths and the transient changes in the physical properties of the fault zone associated with displacement. (b) The deformation fabric stability; i.e. the dilatation patterns induced during deformation associated with the generation of different fabrics, the distribution of deformation within fault zones and the duration of the deformation activity. (c) The competition between cyclic or transient deformation events and continuous deformation. Enhanced fluid flow may be associated with; (i) pervasive, continuous creep, (ii) transient deformation events, or, (iii) flow along faults between or after deformation events when, despite the low strain rates, enhanced flow is still possible through the deformation fabrics present. A critical factor in the efficiency of fluid flow in deformed rocks is the competition between the processes which maintain the connectivity of the high permeability pathways and those which close such pathways. The range of deformation processes which are involved in this competition will be different depending on the tectonic setting, the deformation conditions and the rock types involved. A brief review of the processes which interact to control fluid flow during deformation in sedimentary basins, crystalline basement under low to moderate grade metamorphism and during prograde metamorphism at moderate to high grades is given below. It should be emphasized that this division and separation is artificial and interactions between these regimes are possible particularly where exchange of fluids occurs via a large crustal scale fault zone as discussed by Sibson (1986, 1992).
107
Fluid flow and deformation in sedimentary basins While subsidence and sedimentation continue, material in a sedimentary basin will follow a path of increasing pressure and temperature. Fluid pressure increases to values above hydrostatic pressure during burial by a combination of compaction, pressure solution and cementation, and the release of metamorphic fluids and hydrocarbons. The process may be accelerated by deformation if this leads to over-compaction. The extent of fluid pressure increase depends on the permeability of the route to the surface, which depends on the distribution of sealing lithologies and the distribution and sealing properties of faults. This dual role of faults, to act as fluid pathways and as flow barriers in sediments, together with the complexities of the timing of fault array evolution and fluid pressure evolution in different sediments, leads to fluid flow in sedimentary basins being controlled by the creation and destruction of complex 3D migration pathways (Knipe 1993). The influence of evolving fault plane geometries, displacement patterns, tip zone processes and fault rock evolution in controlling the juxtapositions and windows for fluid communication and diagenetic changes are discussed in Knipe (1993). Fluids in the deeper parts of sedimentary basins are frequently hypersaline brines with high concentrations of Na, Ca and C1 (Hanor 1987). Such fluids have a high potential for transporting chemical components (cf. Banks et al. 1991 ; McManus & Hanor 1993), and this may in turn enhance deformation involving diffusive mass-transfer or stress corrosion.
Deformation of crystalline basement rocks at low to moderate metamorphic grades At low temperatures deformation in these rocks is likely to be overwhelmingly dominated by fracture and cataclasis. Significant compaction will not be occurring and any metamorphic reactions will be retrograde, tending to reduce fluid pressures and permeability through the absorption of water and the precipitation of hydrous phases. Many crustal earthquakes nucleate in rocks of this type, since they generally form the strongest part of the lithosphere. Various studies have shown that fluids in rocks of this type tend to be hypersaline brines, both in the near surface (Frape & Fritz 1987), and in shear zones exhibiting cyclic fracture-creep behaviour at temperatures up to 400 ° C (Yardley et al. 1993). In some cases these fluids are
108
R.J. KNIPE & A.M. McCAIG
demonstrably formation waters derived ultimately from the Earth's surface (McCaig 1988; McCaig et al, 1990; Yardley et al. 1993). Crystalline rocks are probably strong enough to support a fracture porosity for considerable periods even if fluid pressures are below lithostatic. Fluids may move down into such rocks from overpressured sediments during deformation or alternatively overthrusting onto sediments will induce upward migration into the crystalline rocks (Kerrich 1986). In some cases, structural geometries do not appear to allow this, and mechanisms involving large scale seismic pumping have been proposed (McCaig 1988). Complex flow patterns are likely to be associated with the tip zones of sub-surface, isolated faulting events, where an asymmetrical distribution of extension and compression adjacent to the fault plane may induce a range of flow patterns between the hanging wall and the footwall (Knipe 1993). Fluid flow is also likely to be different around normal and reverse faults, since mean stress variations may favour movement of fluids into normal faults and away from thrust faults during loading associated with earthquakes (Sibson 1992). It is important to bear in mind however that geochemical evidence for fluid flow often comes from relatively minor faults and shear zones where metasomatic changes can easily be documented (e.g. Beach 1976; McCaig et al. 1990). Stress cycling and fluid flow in the vicinity of these minor structures is likely to be controlled by the earthquake cycle on regionally important seismogenic faults, and they should not be treated in isolation (McCaig 1988). At high temperature (>c.350°C) enhanced plastic deformation will lead to increased fluid pressures, which enhances the possibility of fracturing. It is an interesting paradox that ductile shear zones which contain fluid-filled inclusions or micro-fractures undergoing deformation by dislocation creep are probably also in a near-critical state for fracture. Transient fracturing will lead to enhanced porosity and permeability as discussed earlier in the paper. Even when deformation in mylonites is dominated by ductile or crystal-plastic deformation, fracturing is likely in the more resistant layers or phases and these may act as important transient fluid reservoirs in mylonites, expelling fluid into the adjacent more ductile layers when these fracture intermittently. D e f o r m a t i o n associated with p r o g r a d e m e t a m o r p h i s m at m o d e r a t e to high grades
In these circumstances, fluid pressures will generally initially be high due to the release of
metamorphic fluid and deformation will occur at low differential stresses by mixtures of crystal plasticity, diffusive mass transfer and fracture (e.g. Cox & Etheridge 1989). Some reactions may enhance deformation by inducing stresses through volume changes, creating new weak phases or by enhancing diffusion in fine-grained reaction products. Such reaction-enhanced ductility (White & Knipe 1979; Brodie & Rutter 1985) may also influence permeability behaviour, e.g. in the generation of cleavages (Knipe 198!; Cox et al. 1991) or during deformation/metamorphism of impure dolomites (Yardley & Lloyd 1989).
Conclusions The aim of this paper has been to highlight some of the progress made in recent years towards understanding the interactions between deformation and fluid flow. These interactions are complex and arise from changes in the pore structure, permeability and fluid pressure gradients in rocks during straining. Deformation generates high permeability pathways (dynamic permeability of Stephenson et al. this volume) which enhance fluid flow while the pathways are open and are associated with fluid pressure gradients. Deformation may also create changes in porosity which survive after deformation has ceased and allow continued focused flow of fluids, before eventual sealing. Important objectives of future research will be to quantify these changes, assess how extensive the pathways are, and to establish the time period that such enhanced migration pathways can contribute to fluid redistribution in rocks.
The authors would like to thank E.H. Rutter, A.J. Barker and J. Parnell for comments on the first draft.
References AGAR, S.M. 1990. Fracture evolution in the upper ocean crust: Evidence from DSDP hole 504B. In: KNII,E, R.J. & RurrER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 41-50. ATKINSON, B.K. 1987. Fracture Mechanics of Rock. Academic Press, London. BANKS, D.A., DAVIES, G.R., YARDLEY, B.W.D., McCAm, A.M. & GRANT, N.T. 1991. The chemistry of brines from an Alpine thrust system: An application of fluid inclusion analysis to the study of fluid behaviour in orogenesis. Geochimica et Cosmochimica Acta, 55, 1021-1030.
MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW BEACH, A. 1976. The interrelations of fluid transport, deformation, geochemistry and heat flow in Early Proterozoic shear zones in the Lewisian Complex. Philosophical Transactions of the Royal Society of London, A280, 569--604. BLENKINSOP,T.G. & RUTrER, E.H. 1986. Cataclastic deformation of quartzite in the Moine Thrust Zone. Journal of Structural Geology, 8, 669-682. BORRADAILE, G.J. 1981. Particulate flow and the generation of cleavage. Tectonophysics, 72, 306321. BOULLIER, A-M, & ROBERT, F. 1992. Palaeoseismic events recorded in Archean gold-quartz vein networks, Val d'Or Abitibi, Quebec. Journal of Structural Geology, 14, 161-179. BURLEY, S.D., MULLIS,J. & MATTER, A. 1989. Timing diagenesis in the Tartan Reservoir (UK North Sea): constraints from combined cathodoluminescence microscopy and fluid inclusion studies. Marine and Petroleum Geology, 6, 98-120. BRODIE, K.H. & RUYrER, E.H. 1985. On the relationship between deformation and metamorphism with special reference to the behaviour of basic rocks. In: THOMPSOr%A.B. & RUBLE, D.C. (eds) Kinetics, Textures, and Deformation. Advances in Geochemistry, 4. Springer-Verlag, New York, 138-179. BVERLEE, J.M. 1993. Model for episodic flow of high pressure water in fault zones before earthquakes. Geology, 21,303--306. CARTER, N.L., KRONENBURG,A.K., Ross, J.V. & WILTSCHKO, D.V. 1990. Control of fluids on deformation in rocks. In: KNIPE, R.J. & RuI"rER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 1-13. CHAMBERLAIN,C.P. & CONRAD,M.E. 1991. Oxygen isotope zoning in garnet. Science, 254,403-406. COWIE, P.A. & SCHOLZ, C.H. 1992. Displacementlength scaling relationships for faults: Data synthesis and discussion. Journal of Structural Geology, 14, 1149-1165. Cox, S.F. & ETHERIDGE, M.A. 1989. Coupled grainscale dilatancy and mass-transfer during deformation at high fluid pressures: Examples from Mt. Lyell, Tasmania. Journal of Structural Geology, l l , 147-162. , WALL, V.J., ETHERIDGE,M.A. & POTTER, T.F. 1991. Deformational and metamorphic processes in the formation of mesothermal vein-hosted gold deposits- examples from the Lechlan Fold Belt in Central Victoria, Australia. Ore Geology Reviews, 6,391-423. DICKSON, J.A.D., SMALLEY,P.C. & KIRKLAND,B.L. 1991. Carbon and oxygen isotopes in Pennsylvanian biogenic and abiogenic aragonite (Otero County, New Mexico): a laser microprobe study. Geochimica et Cosmochimica Acta, 55, 26072613. ELSENHEIMER,D. & VALLEY,J.W. 1992. In situ oxygen isotope analysis of feldspar and quartz by Nd:YAG laser microprobe. Chemical Geology, 101, 21-42. ETHERIDGE, M.A., WALL, V.J. & COX, S.F. 1984.
109
High fluid pressures during regional metamorphism and deformation: Implications for masstransport and deformation mechanisms. Journal of Geophysical Research, 89, 4344-4358. EVANS, B. & WON6, T-F. 1992. Fault mechanics and transport properties of rocks. Academic Press pp. 524. EVANS, J.P. 1990. Textures, deformation mechanisms and the role of fluids in the cataclastic deformation of granitic rocks. In: KNIVE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society London, Special Publications, 54, 29--40. FRAPE, S.K. & Fgrrz, P. 1987. Geochemical trends for groundwaters from the Canadian Shield. In: FRITZ, P. & FRAPE, S.K. (eds) Saline water and gases in crystalline rocks. Special Papers of the Geological Association of Canada, 33, 19-38. FRASER, D.G. 1990. Applications of the high resolution scanning proton microprobe in the earth sciences: An overview. Chemical Geology, 83, 27-37. GILLESPIE, P.A., WALSH, J.J. t~ WATrERSON, J. 1992. Limitations of dimension and displacement data from single faults and the consequences for data analysis and interpretation. Journal of Structural Geology, 14, 1157-1172. GILLETI'I, B.J. & SHIMIZU, N. 1989. Use of the ion microprobe to measure natural abundances of oxygen isotopes in minerals. In: SHANKS,W.C. III & CRISS, R.E. (eds) New Frontiers In Stable
Isotope Research: Laser Probes, Ion Probes and Small-Sample Analysis. US Geological Survey Bulletin, 1890, 129-136. GROSnONG, R.H. 1988. Low temperature deformation mechanisms and their interpretation. Geological Society of America Bulletin, 100, 1329-1360. GUEGUEN, Y., DAROT, M. & REUSCHLE, T. 1987. Permeability evolution under stress. In: DOLIGEZ, B. (ed.) Migration of Hydrocarbons in Sedimentary Basins. IFP Exploration Resources Conference, 45,281-285. HANOR, J.S. 1987. Origin and migration of subsurface sedimentary brines. SEPM short course 21. HIPPLER, S.J. • KNIPE, R.J. 1990. The evolution of cataclastic fault rocks from a pre-existing mylonite. In: KNIPE, R.J. & RUIq'ER, E.H. (eds)
Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 71-79. JACKSON, S.E., LONGERICH, H.P., DUNNING, G. & FRYER, B.J. 1992. The application of LaserAblation Microprobe - Inductively Coupled Plasma - Mass Spectrometry (LAM-ICP-MS) to in situ trace-element determination in minerals. Canadian Mineralogist, 30, 1049-1064. JAMTVEIT,B., WOGELIUS,R.A. & FRASER, D.G. 1993. Zonation patterns of skarn garnets: Records of hydrothermal system evolution. Geology, 21, 113-116. KERRICH, R.J. 1986. Fluid transport in lineaments.
Philosophical Transactions of the Royal Society of London, A317, 219-251.
110
R.J. KNIPE & A.M. McCAIG
KNIPE, R.J. 1981. The interaction of deformation and metamorphism in slates. Tectonophysics, 78, 249-272. 1989. Deformation mechanisms - recognition from natural tectonites. Journal of Structural Geology, 11,127-146. 1990. Microstructural analysis and tectonic evolution in thrust systems: examples from the Assynt region of the Moine Thrust Zone, NW Scotland. In: BARBER, D.J. & MEREDITH, P.G. (eds) Deformation Processes in Minerals, Ceramics and Rocks. The Mineralogical Society Series, 1,228-259. 1992. Faulting processes and fault seal. In: LARSEN, R.M. (ed.) Structural and Tectonic
Modelling and its Application to Petroleum Geology. NPF Special Publications, 1,325-343. -
1993. The influence of fault zone processes on fluid flow and diagenesis. In: HORBURY, E.D. & ROmNSON, A.G. (eds) Diagenesis and Basin Development. American Association of Petroleum Geologists Studies in Geology, 36, 135154. -& WHrrE, S.H. 1979. Deformation in low grade shear zones in the O.R.S. from S.W. Wales. Journal of Structural Geology, 1, 53-66. - - , AGAR, S.M. & PRIOR, D.J. 1991. The microstructural evolution of fluid flow paths in semi-lithified sediments from subduction complexes. Philo-
sophical Transactions of the Royal Society of London, A335, 261-273. LESPINASSE, M. & CATHELINEAU, M. 1990. Fluid percolations in a fault zone. A study of fluid inclusion planes (FIP). Tectonophysics, 184, 173-187. LLOYD, G.E. 1987. Atomic number and crystallographic contrast images using SEM: a review of backscattered electronic techniques. Mineralogical Magazine, 51, 3-19. & KNIPE, R.J. 1992. Deformation mechanisms accommodating faulting of quartzite under upper crustal conditions. Journal of Structural Geology, 14, 127-144. LYON, I. & TURNER, G. 1992. The Isolab ® 54 ion microprobe. Chem&al Geology, 101,197-199. MAIN, I.G., MEREDITH, P.G., SAMMONDS, P.R. & JONES, C. 1990. Influence of fractal flaw distributions on rock deformation in the brittle field. In: KNIPE, R.J. & RUTTER, E.H. (eds) Defor-
-
-
mation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 71-89. MARSHALL,J.D. (ed.) 1987. Diagenesis of Sedimentary Sequences. Geological Society of London, Special Publications, 36. MCCAIG, A.M. 1984. Fluid-rock interaction in some shear zones from the Pyrenees. Journal of Metamorphic Geology, 2,129-141. 1987. Deformation and fluid-rock interaction in metasomatic dilatant shear bands. Tectonophysics, 135,121-132. 1988. Deep fluid circulation in fault zones. Geology, 16,867-870. & KNIPE, R.J. 1990. Mass transport mechanisms
-
-
in deforming rocks: Recognition using microstructural and microchemical criteria. Geology, 18,824-827. - - , WICKHAM,S.M. & TAYLOR,H.P. JR. 1990. Deep fluid circulation in alpine shear zones, Pyrenees, France: Field and oxygen isotope studies. Contributions to Mineralogy and Petrology, 106, 41-60. - - , GONG, L-Y. & WAYNE,D.M. 1993. Mechanisms of permeability enhancement in carbonate mylonites. In: PARNELL,J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids "93 Conference Volume, 159-161. MCLAREN, A.C. 1991. Transmission electron microscopy of minerals and rocks. Cambridge Topics in Mineral Physics and Chemistry 2. Cambridge University Press. MCMANUS, K.M. & HANOR, J.S. 1993. Diagenetic evidence for massive evaporite dissolution, fluid flow, and mass transfer in the Louisiana Gulf Coast. Geology, 21,727-730. MrrRA, S. 1988. Effects of deformation mechanisms on reservoir potential in Central Appalachian overthrust belt. American Association of Petroleum Geologists, 72,536--554. MUIR WOOl), R. & KING, G.C.P. 1993. Hydrological signatures of earthquake strain. Journal of Geophysical Research, 98,000-000. ODLING, N.E. 1992. Permeability of natural and simulated fracture patterns. In: LARSON, R.M. (ed.) Structural and Tectonic Modelling and its Application to Petroleum Geology. NPF Special Publications, 1,365-381. PARRY, W.T. & BRUHN, R.L. 1986. Pore fluids and seismogenic characteristics of fault rock on the Wasatch fault, Utah. Journal of Geophysical Research, 91,730-744. PHILIPPOT, P. & VAN ROERMUND, H.L.M. 1992. Deformation processes in eclogitic rocks. Journal of Structural Geology, 14, 1059-1077. PRIOR, D.J. 1988. Fractures and retrogression in garnets from the Alpine Fault mylonites, New Zealand. Journal of the Geological Society, London, 146, 335-347. RAMSAY,J.G. 1980. The crack-seal mechanism of rock deformation. Nature, 284, 135-139. ROEDDER, E. 1984. Fluid inclusions. Reviews in Mineralogy 12. SCHOLZ, C.H. 1990. The mechanics of earthquakes and faulting. Cambridge University Press. SELVERSTONE,J., FRANZ, G., THOMAS, S. t~ GETTY, S. 1992. Fluid variability in 2 GPa eclogites as an indicator of fluid behaviour during subduction. Contributions to Mineralogy and Petrology, 112, 341-357. SHARP,Z.D. 1992. In situ laser microprobe techniques for stable isotope analysis. Chemical Geology, 101, 3-20. SIBSON, R.H. 1980. Transient discontinuities in ductile shear zones. Journal of Structural Geology, 2, 165-171. 1986. Earthquakes and rock deformation in crustal fault zones. Annual Reviews in the Earth and Planetary Sciences, 14,149-175. 1990. Conditions of fault-valve behaviour. In: -
-
MICROSTRUCTURAL CONSEQUENCES OF FLUID FLOW KNIPE, R.J. & RUTTER, E.H. (eds) Deformation Mechanisms, Rheology and Tectonics. Geological Society, London, Special Publications, 54, 15-28. 1992. Implications of fault-valve behaviour in rupture nucleation and recurrence. Tectonophysics, 211,283-293. 1993. Crustal stress, faulting, and fluid flow. In: PARNELL, J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 137-140. ~, ROBERT,F. & POULSEN, K.H. 1988. High angle reverse faults, fluid pressure cycling & mesothermal gold-quartz deposits. Geology, 16,551-555. SLEEP, N.H. & BLANPIED, M.L. 1992. Creep, compaction and the weak theology of major fault zones. Nature, 359,687--692. SPIERS, C.J. & SCHUTJENS, P.M.T.M. 1990. Densification of crystalline aggregates by fluid phase diffusional creep. In: BARBER,D. & MEREDITH, P.G. (eds) Deformation of Materials. Mineralogical Society Series, 1,334-352. STARK, C.P. & STARK,J.A. 1990. Seismic fluids & percolation theory. Journal of Geophysical Research, 96, 8417-26. STEPHENSON, E.L., MALTMAN, A.J. & KNIPE, R.J. 1993. Fluid flow and microstructure in deforming sediments. In: PARNELL, J., RUFFELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 148-150.
111
WAYNE, D.M. & McCAIc, A.M. 1992. Fluid flow during thrusting: A Sr isotope study. In: KHARAKA, Y. & MAEST, A.S. (eds) Water-rock interaction. Balkema, Rotterdam, 1559-1664. WINTSCH, R.P. & KNIPE,R.J. 1983. Growth of a zoned plagioclase porphyroblast in a mylonite. Geology, 11,360-363. WHITE, S.H. & KNIPE, R.J. 1979. Transformational and reaction enhanced ductility. Journal of the Geological Society, London, 135, 513-516. WOGELIUS, R.A., FRASER, D.G., FELTHAM, D.J. & WHITEMAN, M.I. 1992. Trace element zoning in dolomite: Proton microprobe data and thermodynamic constraints on fluid composition. Geochimica et Cosmochimica Acta, 56,319-334. YARDLEY, B.W.D. & LLOYD, G.E. 1989. An application of cathodoluminescencemicroscopy to the study of textures and reactions in high-grade marbles from Connemara, Ireland. Geological Magazine, 126,333-337. ~, BANKS, D.A. & MUNZ, I.A. 1993. Fluid penetration into crystalline crust: Evidence from the halogen chemistry of inclusion fluids. In: PARNELL, J., RUEEELL, A.H. & MOLES, N.R. (eds) Geofluids '93 Conference Volume, 350-353.
Fluid flow in actively deforming sediments: 'dynamic permeability' in accretionary prisms E.L. STEPHENSON,
A.J. MALTMAN
& R.J. KNIPE 1
Institute of Earth Studies, University of Wales, Aberystwyth, Wales SY23 3DB, UK 1 Department of Earth Sciences, University of Leeds, Leeds LS2 9JT, UK Abstract: Recent investigations into active accretionary prisms have emphasized both the importance of the deformation/fluid-flow interplay in governing fundamental aspects of prisms and the need for quantitative data on the processes. Even the basic parameter of permeability is little known in this and other geological settings where the sediments are undergoing active deformation. New experimental data are presented which show that permeability during accumulating strain, here called the dynamicpermeability, is not a static value but is highly variable. Moreover, microstructural analysis and precise determination of permeant volume reveal that this dynamic permeability is not solely the varying capacity of the deforming medium to transmit fluid, a quantity here called the Darcyanpermeability, but includes a contribution to the amount of permeant from the medium itself. In the laboratory experiments this additional contribution, called the dynamic component, arises from pore-volume fluctuations associated with microstructural changes, but in nature there may be further mechanical and chemical effects. Applying conventional methods of permeability determination to an actively deforming sediment will necessarily include this dynamic component in the measurement, a consideration relevant to a variety of geological and engineering situations.
The interaction between deformation and fluid flow has recently become an important topic for geologists investigating accretionary prisms. The interest has arisen because at convergent plate margins, where materials tend to be scraped from the subducting plate to build out the prism-shaped mass, accumulating wet sediments are being transformed into virtually dry rocks in an exceptionally dynamic structural environment (e.g. Langseth & Moore 1990). Enormous quantities of fluid have to be lost from the system, and their fate influences a wide range of fundamental plate-margin processes, from magmatism to mineralization, and tectonic behaviour to metamorphism. How do the fluids escape? Are some temporarily trapped? What proportion of the fluid is subducted and what routes does the remainder follow? Such questions are fundamental to an understanding of the architecture and processes of convergent plate margins. Despite this importance, knowledge of how fluids are transported through actively deforming and deformed geological materials is extremely sparse. The fundamental parameter of permeability, for example, has been measured in countless on-land rocks that are assumed to be rigid, static, and isotropic. Any effects of deformation fabrics or ongoing strain have been little explored. Reported below is an
attempt in the laboratory to measure the permeability of sediments during their active deformation, a quantity we refer to as the dynamic permeability. The results not only reveal a substantial difference between the dynamic and the static, at-rest values, but demonstrate that the deforming sediments themselves can contribute to the amount of expelled fluid. This latter observation requires that in such dynamic situations the concept of permeability cannot be treated in the conventional way. The first half of the paper is a brief review of recent progress in investigating fluid flow in actively deforming accretionary prisms. The second half discusses the results of new experiments into permeability variations during the active deformation of sediments.
Recent studies on deformation and fluid flow in accretionary prisms In the last two decades, convergent margins have blossomed from being a poorly investigated, obscurely understood plate-margin type to one with well-established architectural features, and to being a target of major international study. By the mid-1980s, it was known that at most sites of oceanic-plate subduction,
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, Geological Society Special Publication No. 78, 113-125.
113
114
E.L. STEPHENSON E T AL.
the veneer of sediments on the converging plate and those accumulating in the trench were being intermittently offscraped, in a geometry closely analogous to that well documented from on-land fold-and-thrust belts (Moore & Lundberg 1986). The main features of such prisms of accreted materials are shown in Fig. 1. Where little sediment is available for accretion, such as at sites of intra-oceanic collision, accretion is minimal (e.g. Mariana; Fryer et al. 1992), and some margins appear to show no accretion at all, with all of the oceanic plate material being subducted (e.g. off Guatemala; yon Huene et al. 1985). At some margins, accreted material seems to be removed from beneath the prism by the subducting plate, the so-called subduction erosion (e.g. off NW Japan; Langseth et al. 1981), while at others further material is added to the sole of the prism, a process called underplating (e.g. Makkran; Platt & Leggett 1985). Also, some accretionary prisms are relatively short but immensely thick while others appear thin and areally extensive, leading to variations in prism taper angles. These major differences were reviewed by von Huene (1984). A crucial factor in determining these gross tectonic differences is the behaviour of the boundary between the colliding plates, the detachment zone or decollement, and theories attempting to explain the variations in taperangle have also highlighted the importance of decollement behaviour (e.g. yon Huene & Lee 1983). All the indications are that the fluid behaviour is fundamental (e.g. Westbrook & Smith 1983). Fluid pressures appear to govern, just as in the classic explanations of on-land thrust belts, the friction at the decollement, and hence whether it can slip freely or whether the strain has to be transferred to other features such as propagating thrusts, thus thickening the prism (Davis, et al. 1983). Drilling near the decollement beneath the Northern Barbados Ridge prism during the Deep Sea Drilling Project (DSDP; Moore & Biju-Duval 1984) inadvertently indicated the existence there of near-lithostatic fluid pressures, and a direct measurement of abnormal fluid pressures was achieved during DSDP drilling off Guatemala (yon Huene 1985). Besides this general evolution of enquiry, two particular developments in the late 1980s made fluids the crucial topic of prism research. One emerged from a series of international projects using submersibles and sea-floor imaging to observe the ocean floor (e.g. Carson et al. 1991; Le Pichon et al. 1992 and following articles). The discovery of isolated biological communities and mineral crusts at the outcrops of faults seen on
seismic traces fostered the notion of focused flow, with tectonic features acting as conduits for fluid escape. The other development arose from further drilling of the Barbados prism, now as part of the Ocean Drilling Program (ODP). This drilling revealed clear evidence of channelised flow along major tectonic features. It was indicated by, for example, geochemical anomalies (e.g. enrichment of methane, 180, and 87Sr/S6Sr, and dilution of chloride; Gieskes et al. 1990), thermal effects (Langseth et al. 1990), and syntectonic mineralization (Vrolijk & Sheppard 1991) within the faults and basal decollement. Geophysical studies, particularly involving seismic velocities together with reflection amplitudes and polarities, added further evidence (e.g. Westbrook 1991). Such discoveries revealed exciting new insights into the hydrogeology of prisms, and the neatness of the data from Barbados elevated the deformation-controlled dewatering of this prism to something of a paradigm. Moreover, drilling of prism-like sediments off Peru showed reasonably similar anomalies (e.g. Kastner et al. 1991). However, the concept of tectonic conduits had also raised a host of new questions (Langseth & Moore 1990), and exposed the need for quantitative information. It was expected that the next prism to be targeted by the ODP, the Nankai Prism off SW Japan, would display features similar to Barbados and help provide numerical data on the channelized flow and related effects. In the event, however, not only did in situ measurement continue to prove difficult, but the very evidence for the tectonic focusing of flow was unclear. In fact, in some ways, the structural (Maltman et al. 1992) and other evidence (Taira et al. 1992) was more compatible with a pervasive, intergranular drainage of this prism. Good evidence did emerge that the decollement at Nankai is currently acting as a seal, efficiently separating overpressured sediments on the downgoing plate from the apparently diffusely draining prism. Yet along strike from the drilling site at Nankai, beyond several major strike-slip faults transverse to the prism, observations from a submersible showed the presence of diapirs and bioherms indicative of localized venting (Le Pichon et al. 1992) i.e. prisms can be compartmentalised into different dewatering modes. The relatively small prism drilled during ODP investigations of processes at the Chile triple junction showed, again through thermal and chemical discontinuities (Behrmann et al. 1993), that fluids were moving vigorously in the highly sheared formations, although the processes there are taking place within an unusually high
FLUID FLOW IN DEFORMING SEDIMENTS diffuse dewatering
channelised dewatering ~
.-~~, "' " " .
i'
.
ent . ! -
• d~ut,
. "• ~' /-_ ." ' . ' - . " .• ' . '- . ~ ...
.
'-
• ". ". ". •
/
\
"
~~ I " ,
.
.
i
115
"\/
~ / ~ / / _
.
.
.
.
" . " . -
.
/ /
frontal thrust
.£
/
.
i
,'~\
. x. .
\DOWNGOINGPLATE ~1
.
.
s u b d u c t i o n e r o s on I / /
/~/\..
\"
~/
t/~.-"
~
--" ~
protothrust
....:.:.:.:.:.i:.i.i.:.:.i.:.i::
underD a t i n o ~" "" " p r i s m t a p e .
,, ~ ' \. . .
.
~
...-
.
.
I
~
/ ~ / ~ \ / ~ l t " / ,
/
\l
g t
\
"..t.."
-. , \ ~ " \ .
/
/
~
/
t
/
- / ~/ " / t t i
/
--
.
Fig. 1. Schematic drawing of the toe region of a hypothetical accretionary prism, to illustrate terminology used in the text. In this typical example, the decollement is positioned within the sedimentary pile (stippled) so that the upper sediments are being accreted to the overriding plate while oceanic crust (pecked ornament) and some overlying sediment are underthrust. Stepping up of the decollement within the sedimentary section leads to subduction erosion while stepping down produces underplating. Note that the channelised and diffuse dewatering regimes, although portrayed here for convenience on adjacent hinterland thrusts, may occur at any position within a prism.
thermal gradient. The accretionary prism most recently targeted, at the Cascadia margin, provided a further example of along-strike change in dewatering behaviour, in that sites off Vancouver disclosed diffuse dewatering through pervasively fractured silts, at flow rates no more than a few millimetres per year, whereas sites further south, off Oregon, showed substantial tectonic focusing (Westbrook et al. 1993). Some thrusts carried thermogenic hydrocarbon gases in actively flowing pore fluids, and overpressures (0.2MPa at 94-146m) were successfully documented. Summaries of some relevant aspects of recently drilled prisms are given in Table 1. Some of the instruments deployed at Cascadia have been left in place for future measurement, to evaluate the episodicity of flow. The objective of obtaining in situ measurements points the way to future research in flow-deformation studies in prisms. The planned ODP drilling of the Barbados prism in 1994 will emphasize the long-term deployment of measuring devices. The goal is to quantify the amounts and temporal behaviour of fluid flow in this otherwise well-documented, actively deforming, accretionary prism. It is clear now that different accretionary prisms dewater in different ways. The behaviour may well vary spatially within the prism, and probably fluctuates through time. Quantitative understanding of all these processes is still poor.
Permeability in deforming sediments If overpressures are the key to the structural behaviour of a prism, then permeability must be the fundamental influence on their extent and duration (Moore et al. 1991). It is the permeability that quantifies the capacity of prism
sediments to dissipate fluid pressures in order to maintain an equilibrium gradient, and hence determines the extent of overpressuring. As mentioned above, although the parameter is routinely measured in a host of on-land circumstances, knowledge of permeability in any oceanic sediments is extremely sparse, let alone in examples containing deformation fabrics and those that are undergoing active strain. Taylor & Fisher (1993) reviewed the relatively few permeability data obtained from modern accretionary prisms, through laboratory measurements. They cite values ranging from 10-14 to 10-19 m 2 for cores retrieved from Nankai and, for cores from the Barbados prism (Taylor & Leonard 1990), permeabilities between 10 -15 and 10 -18 m 2. The role that deformation fabrics play in influencing these values is unclear. Evidence of preferred flow along core-scale faults was documented by Knipe (1986a) and along shear-zones by Maltman (1988). Arch & Maltman (1990) advanced an explanation of the latter observations, for zones with marginparallel grain alignments, based on the reduced tortuosity of the flow path. Brown & Moore (1993) have commented that fabric collapse accompanying such alignments should also reduce the porosity and hence reduce the alongzone permeability. Although theory holds this to be true, it does seem that in practice, whether through transient dilation or some other mechanism, shear zones can channel flow at least for some time. Core-scale shear zones can be regarded to some extent as scale models of the large-scale tectonic discontinuities of accretionary prisms (Arch & Maltman 1990), but in some prisms they themselves may play an important role in the bulk dewatering (Byrne et al. 1993).
Anomalies Localized rhodocrosite 28-36 ° , anomalies
87Sr/86Sr Mineralized veins
Thermal gradient
Chile (ODP Leg 141)
Lacks anomalies No veins 110°, linear
50 ° , anomalies
Some localized carbonate, pyrite Highly variable between 30--600 °
Diapirs, bioherms in n/a places Local minima Lacks anomalies Some anomalies at n/a thrusts
Nankai (ODP Leg 131)
Anomalies Sparse
Anomalies Anomalies
n/a
Mud volcanoes, diapirs Anomalies Anomalies
Ocean-floor venting
Cf concentrations 180 values
Peru (ODP Leg 112)
Barbados (ODP Leg 110)
54 ° , linear
51 ° mostly linear
Carbonate veinlets
Gas discharge carbonate deposits Anomalies Pore water anomalies reflect 'spatial heterogeneity of fluid flow'
Oregon (ODP Leg 146)
Linear change 'Little evidence of significant fluid flow that is confined to conduits'
None
Vancouver
Cascadia
FLUID FLOW IN DEFORMING SEDIMENTS Maltman et al. (1993) suggested that the abundance of these kinds of core-scale structures at Nankai may be part of the reason why this prism shows less marked flow along the major tectonic structures (Taira et al. 1993). All the permeability measurements mentioned above were made on laboratory specimens 'at rest', giving a value we refer to as the static permeability. Langseth et al. (1988) pointed out that such determinations may not be an accurate measure of the value within an actively deforming prism, hence the importance of attempting the difficult matter of obtaining reliable in situ measurements. The only known efforts in the laboratory to assess the effect of active strain on flow through prism materials, referred to here as the dynamic permeability, are the preliminary results we gave in Byrne et al. (1993) and the data outlined below. The former results, on cores from the Nankai prism and analogue material, revealed that the dynamic permeability may change by almost 100% at only 25% bulk strain. New data are presented below, in an attempt to refine understanding of this kind of behaviour.
Experimental method We have employed a number of innovative, precise methods, cross-checked with other laboratories, to monitor permeability changes during sediment deformation. Details of the equipment and methods are given elsewhere (Arch, Stephenson & Maltman, in prep). Briefly, the experiments were based on coupling the well-established constant-head method of permeability measurement to a triaxial test arrangement. The two important points to note here are that: (i) the constant-head technique is based on supplying permeant under a constant pressure-head and measuring its ou(flow from a specimen; and (ii) our equipment also allows the amount of inflow to be monitored, even though this would normally be expected to be the same value. Other relevant points are that the specimens are reasonably large, 54 mm diameter cylinders × (typically) 100 mm height, tests were accomplished in a few tens of hours and at room temperature, with readings being automatically taken at ninety second intervals, and the permeant was distilled water. Our work so far has centred on clays and silty clays. The samples have either been retrieved from the active Nankai accretionary prism by the ODP, or generated in the laboratory by consolidating clay/silt slurries.
Experimental results Equilibrium flow through a specimen is known to have been established when any variation of outflow with time is negligible. From Darcy's relation, and knowing the specimen dimensions and the differential head, the permeability of the
117
specimen can be calculated. A typical result of such static permeability tests is shown in Fig. 2a. None of the materials tested so far retains its static permeability value when it undergoes strain. Our results suggest that for a given sediment type the variation depends to a large extent on its state of consolidation. Under- and normally-consolidated materials show a gradual reduction in dynamic permeability with strain (e.g. Fig. 2b) whereas over-consolidated samples of the same composition show an increase (Fig. 2c). Some lightly overconsolidated specimens show, for reasons we have not yet identified, more complex fluctuations superimposed on a progressive dynamic permeability decrease (Fig. 2d). Arch, Stephenson, & Maltman (in prep) suggest that the explanation of these varying behaviours lies in the dilation and contraction of the materials at the grain scale. In summary, more tightly-packed grains, as in an overconsolidated specimen, have to dilate to accommodate the growing strain, thus increasing the permeability either transiently or permanently, whereas poorly consolidated sediments tighten the grain packing with strain, and diminish the permeability. Temporary fluctuations presumably correspond to porosity changes associated with localized reconfigurations of grains, in particular the production of shear zones in conditions of peak stress.
The concept of dynamic permeability The inferences of dilation and tighter packing mentioned above open up a new view of the concept of permeability in active materials. That is, instead of the medium simply being a matrix of grains and pores that passively allows the permeant to filter through, the material itself can add to or subtract from the amount o f fluid. In the classical view, originating with Henry Darcy, permeability is a material constant, representing the capacity of a porous medium to transmit water in response to a hydraulic gradient. There have been discussions in geology on possible physical (e.g. Peach et al. 1987) and chemical (Sample 1990) interactions between the permeant and the permeated matrix, but normally the medium is treated as inert. This is the case in our treatment of static permeability. However, we view dynamic permeability as consisting of two components: the passive matrix element mentioned above, which we propose to call the Darcyan permeability, and a dynamic component, arising from the contribution from the grain matrix itself. Fine semantic matters arise from this; we purposefully avoid referring to the
118
E.L. STEPHENSON E T AL. 1400
5E-16
1200 r~
4E-16
5E-16
~
1000
800 "~
3E-16
4E-16 600
E
E
~- 2E-16
ne 3E-16
~-
400 200
1E-16 0
|
=
=
=
i
5
10
15
20
25
2E-16 0
2
4
2E-15 1E-15
~ >,
1200
b ~
6
8
10
12
14
16
18
20
Strain %
.time - h o u r s
~ BSS~~ ' ' ~
5E-17
] d
350 300
1000
1E-15
¢~ E
800
1E-15
600
8E-16
==
._ "~
4E-17
\
250 200
3E-17 150
~
6E-16
400
=
e~_b_i,ty_ " '-.. . . . . . . . . . .p_~r~_ .. ,_~ . . . . . . . .
4E-16
E 2E-17
200
2E-16 OE+ O0 0
|
i
,
,
5
10
15
20
0 25
30
Strain %
1E-17
]I V
0
lO0
', ;-.-., ,. . . . . ~,. . . . . ; ~.,J
~
;0
;5
2'0
ca .~ m
50 0
25
Strain %
Fig. 2. Summary of various permeability behaviours. (a) Static permeability graph for a sand-kaolinite mixture. (b) Dynamic permeability of an under-consolidated sand-kaolinite sediment. The permeability shows an overall decrease with increasing strain, initially rapid but then at a reduced rate as the specimen consolidates under the applied load. (c) Dynamic permeability of an overconsolidated sand-kaolinite mixture. After initial adjustment (1% strain), the permeability shows a small decrease followed by an increase, presumably as the specimen dilates to accommodate strain. (d) Complex dynamic permeability behaviour, in a lightly overconsolidated ball-clay. Intervals of enhanced permeability are superimposed on an overall gradual decrease.
latter component as a permeability or a flux. Depending on the circumstances, the dynamic component may be positive, negative or zero. Irrespective of this contribution from the dynamic component, in an actively straining sediment the Darcyan permeability itself may well not be of constant magnitude. In a natural active situation, the dynamic component probably arises chiefly through either or both of two effects. The first is diagenesis. For example, cementation (mineral infilling of pore space) will displace some of the pore-fluid. Transformation of organic matter during burial will produce new hydrocarbons, some of which may be in the fluid state (e.g. Berner & Faber 1993). The hydrous content of minerals in the buried sediment may change. Many minerals contain water more or less loosely linked to their molecular structure, and there are numerous geological circumstances where various minerals may take up or lose
water. However, the major process of this kind within accretionary prisms will be clay dehydration. According to Tribble (1990), for example, in the Barbados area 30% of the grain matrix consists of (hydrous) smectite. After burial to temperature conditions around 60° C, the smectite transforms to illite, releasing about a third of its volume as water (Le Pichon et al. 1991). The second effect is mechanical change within the sediment. In materials consolidating or dilating, the bulk of the volume change will be accommodated by porosity adjustment, and because pore-fluids are virtually incompressible (excepting gases), these will be accompanied by fluid intake or expulsion. It is doubtful that in our short, room-temperature experiments involving distilled water, any of the diagenetic effects are being reproduced, but we believe we have detected mechanical contributions to dynamic permeability. The evidence comes from
FLUID FLOW IN DEFORMING SEDIMENTS
;
119
i-
,
Fig. 3. SEM secondary electron images of ball clays at different amounts of bulk strain. (a) Undeformed sample, showing the domainal primary consolidation fabric, here running approximately NNE-SSW, with high-angle clays between creating pockets for fluid storage. (b) Similar material, deformed to 3% strain. The primary fabric, oriented approximately NE-SW in this image, has been intensified. (e) Similar material, deformed to 7% strain. Shear zones, with intensely aligned but undamaged clays, here run NNW-SSE. Localized dilation, to allow rotation of the clay particles, will have caused temporary storage of permeant, later to be released as the clays collapse into alignment.
microstructural analysis and from precise volume-change measurements during the experiments.
Microstructural analysis Some of the test specimens have been examined microscopically in the scanning electron microscope. Summarizing the results, materials consolidated to greater pressures show more intense fabrics, defined by better alignment of the clay flakes and decreased pore space. The effect is well known from nature (e.g. Maltman 1981; Faas & Crockett 1983). Initial shortening of the experimental cylinders continues the consolidation process; Arch (1988) showed that 30% porosity ball clays can achieve up to 12% bulk strain in this way. Continuing strain causes, in line with observations from nature (Maltman 1988; Knipe 1986b) and experiment (Maitman 1987), the sediments to fail by generating arrays
of shear zones. The evidence that these microstructural changes must have contributed to the dynamic permeability is discussed in greater detail below for a selected suite of tests. A series of ball-clay test cylinders was generated in the laboratory and each deformed to different levels of total axial strain. Samples from each were compared using secondary electron imaging of fracture surfaces, oriented roughly perpendicular to the dominant microstructure of the specimen. A control sample, trimmed from a specimen prior to testing, showed a primary fabric resulting from the uniaxial consolidation of the sediment employed in the manufacture of the sample. As illustrated in Fig. 3a, the fabric is domainal, with variations in orientation and intensity of clay particle alignment. Both presumably give rise to local variations in permeability. Also present are regions with particles oriented at a high angle to the main fabric. Being sub-parallel to the
E.L. STEPHENSON ET AL.
120
_
5E-17-
stress
180 -160 -140
4E-17-
-120 E '-JO ¢0 (1)
-100
t~ ,--
3E-17 -80
t~ "13
60 0,.
2E-17-
40
g 20 1E-17 - J 0
2
4
(~
8 10 strain %
1'2
14
1'6
0
18
Fig. 4. Dynamic permeability of a lightly overconsolidated ball clay, showing marked permeability variation as strain increases. Transient increases are superimposed on an overall permeability decrease. Confining pressure = 340 KPa; differential pressure along specimen = 200 KPa; strain rate = 1.02 × 10-4 s-1. The permeability curve is a thirty-point running mean through the outflow data.
direction of consolidation loading, these particles are prompting the formation of pockets in which permeant might be stored and, being in a less mechanically stable orientation, they may be preferred sites for the initiation of shear zones when strain accumulates. The intensification of the primary fabric at low levels of bulk strain is apparent from Fig. 3b. The clay particles show a predominantly face-toface arrangement, trending approximately N E - S W in that image. The porosity of the sample has been reduced significantly as a result of the denser packing of the parallel clay flakes. This loss of pore volume must contribute fluid to the downstream, exiting permeant. Measurement of the dynamic permeability during this period will therefore not be wholly Darcyan, but overestimated by an amount proportional to this porosity reduction. The formation of discrete shear zones at higher levels of strain is illustrated in Fig. 3c. Two such zones are present, showing a sigmoidal swinging of the consolidation fabric into the zones, which are defined by a more intense parallel orientation of the clay plates and a further reduction in porosity. Note that none of the clay particles shows signs of damage. The only conceivable way in which this new arrangement could be achieved is through temporary
dilation while the clay flakes rotate into alignment. During these intervals of dilation, permeant must have been incorporated into the expanded pores, temporarily reducing the throughflow. This stored fluid would then be released in pulses when the realigned clays progressively collapsed to form the shear zone fabric. The existence of such localised, transient reservoirs is in line with the model presented by Knipe et aI. (1991). Therefore, the dynamic permeabilities portrayed in Fig. 2 are not simply registering changes in the capacity of the sediment to transmit fluid, the Darcyan permeability, but must be incorporating varying contributions to the flux from these dynamic components arising from pore-volume changes.
Volume change analysis In the Darcyan view of permeability, any change in the fluid transport capability will be accompanied by a corresponding change in the inflow and outflow of permeant, and these will be identical quantities. Both are therefore not normally considered in the determination of permeability. Our equipment allows the precise monitoring of both the inflow and outflow volumes: should any discrepancy between the two quantities be detected it must represent the
FLUID FLOW IN DEFORMING SEDIMENTS
0.020
12l
stress
180
oo,6 / o
,! !
~ 0.o12.
/
o
-, --
o>
,oo
"~'~'~ ~
o.oos
,,.L. L I I
~,l,t~,JI,
!,~AJ,t.~, I
2
x--
4
6
60
,I
h4 O
00.006004~~~!i'~i'i, )~I~i]~'r '"" ' . )f~. I'' . 0
-80
"0
0.0101
.
.
8 10 strain %
.
20 0 12
14
16
18
Fig. 5. Inflow and outflow volumes of permeant during the dynamic permeability test illustrated in Fig. 4. Each data point through which the curve is drawn represents the fluid entering or leaving the specimen during a ninety-second interval of the test.
activation of a dynamic component in the permeability. We have no reason to suspect in our experiments any change in the density of the sediment grains, so that, assuming such a component in the present experiments is wholly mechanical and not diagenetic, the volume difference must reflect a change in pore space. In other words, differences in volume between the inflow and outflow will give a quantitative assessment of the volume changes inferred from the microstructures. Figures 4 to 7 reveal the dynamic component of permeability and its significance for a ball-clay cylinder deformed to 16% axial strain. Figure 4 shows the dynamic permeability of the deforming clay, calculated in the conventional way from the outflow of permeant from the specimen. Figure 5 shows this outflow as a volume, together with the volume of inflow, which is a lesser value for most of the test. The dynamic permeability of Fig. 4 is therefore not the classical, Darcyan permeability but is incorporating a dynamic component associated with pore-volume reduction within the specimen. The magnitude of the dynamic component is given by the difference between the incoming and outgoing fluxes. During any interval for which the outflow exceeds the inflow the specimen can be taken to have decreased in volume by an amount equal to this difference (given that deformation is accommodated entirely by change in pore fluid volume). Figure 6
shows this dynamic component, expressed as the percentage specimen volume change within each 90s interval of the. entire test. Note that the majority of the data falls in the negative field, indicating that the specimen is undergoing bulk volume loss, at various rates. At the onset of axial loading, where the rate of mechanical dewatering (negative volume change) is at a maximum, the dynamic component reaches its maximum value (at 1% strain). The rate of porosity loss decreases steadily until 6% strain, where it levels off before rapidly increasing again. This latter, sudden increase in drainage rate coincides with the strain softening exhibited in the stress curve. Then, as deformation continues, the rate of volume decrease returns to a value similar to that between 2 and 6% strain, and then gradually decreases until it becomes positive at approximately 14% strain. At this point, the specimen is undergoing bulk dilation and permeant is being stored within the specimen. Here the dynamic component is negative, causing a reduction in the dynamic permeability. The result of removing the dynamic component from the dynamic permeability is given in Fig. 7. This curve is therefore a representation of the variation in Darcyan permeability with strain. Its shape resembles that of the dynamic permeability of Fig. 5, but in inverseforrn. This demonstrates that the conventional methods of determining permeability do not, when applied
122
E.L. S T E P H E N S O N E T A L .
0.002
_ stress
180
o.oo1 o~
0.000
"
-0.001
II
I
/ . . . . . . . i~.11 ~ ~
'~'~'
-160
i! 140 ql' ~ ~ ' ".!lJh --
ii ,
-120
tO
100 -0.002
E = > -0.003 .=
,
, ~1
~i
g
-80 I
'
60
-0.004
"~ =
40
-0.005
20 I
- 0 . 0 0 6 ---~
0
2
.
4
.
6
.
.
8 1'0 strain %
.
1'2
1~4
16
0
18
Fig. 6. Volume changes of the specimen during the test illustrated in Fig. 4. The changes are plotted as an interval bulk volume change, i.e. the percentage change of the initial cylinder volume within a ninety-second interval of the test. Almost all the values are negative, indicating overall volume loss, but at varying rates. Positive slopes (e.g. at 1-6% strain) indicate successively decreasing amounts of volume loss; steeper slopes represent more rapid rates of change. Decreasing rates of volume loss in the negative field may be accounted for by increasing Iocalised dilation at the same time as continuing bulk consolidation. Bulk dilation occurs only temporarily at around 14% strain. to an actively deforming material, measure its fluid transport capacity unless corrections are made for the contributions from the dynamic component. The explanation of the contrast between the Darcyan and the dynamic permeabilities lies in the pore-volume fluctuations that accompany the microstructural changes outlined earlier, and especially the localized effects associated with shear zone formation. For example, at around 1-3.5% strain, the Darcyan permeability is decreasing very slightly, because of the increased packing of the grains in response to the bulk strain. It then begins to increase as a result of dilation associated with the initial development of shear zones, which facilitate increased fluid flow. Even so, the dynamic permeability continues to decrease (Fig. 6), reflecting continuing bulk consolidation of the specimen. The localised, dilating shear zones eventually collapse (7% strain), prompting an increased total outflow, and a peak in the dynamic permeability curve (Fig. 4). The Darcyan component, however, decreases, due to the production of a low-porosity clay fabric within the shear zones, most of which are oriented at a high angle to the axial fluid flow and hence are tending to act as barriers. In ways such as these, the permeated
medium is far from being simply a matrix of static grains and pores that passively allows the movement of fluid, as is normally assumed in permeability work.
Conclusions The observations presented above demonstrate that the dynamic permeability, the permeability measured in an actively deforming material, contains a dynamic component in addition to the classical Darcyan permeability. The amount of fluid exiting from a material is not just a function of the standard Darcyan parameters, but is contributed to by the transmitting medium itself. In our experiments, the dynamic component arises from pore-volume fluctuations associated with microstructural changes during consolidation and deformation, but in nature a host of other mechanical and chemical contributions are possible. The conclusions will apply to other laboratory methods of determining permeability, such as the falling head, flow-pump, and consolidometry techniques, and also to in situ measurements of a natural dynamic situation. The most likely approach in the near future to measuring permeabilities in an accretionary prism will be
FLUID FLOW IN DEFORMING SEDIMENTS 4E-17-
_
123
stress
180 +160 140
3E-17
120 lOO
•~
80
11[''* I~arcyan permeability
0
.
2
.
.
4
. -
6
.
.
.
.
8 10 strain %
.
.
12
.
14
"0
40 r
1E-17 d
u~
20 0
16
18
Fig. 7. The Darcyan permeability of the test illustrated in Fig. 4. Differences in flux due to pore fluid fluctuations (the dynamic component) have been removed from the dynamic permeability using the data shown in Figs 5 and 6; the curve therefore shows the permeability of the material in the conventional sense. The variation in the capacity of the deforming sediment to transmit the permeant is due to microstructural changes within the specimen as a result of deformation.
an indirect derivation from the rate of fluidpressure dissipation and a variety of methods are under development, such as piezometric probes and wireline packers (Langseth et al. 1988). Any data obtained will lead to the dynamic rather than the conventional permeability. The amount of contribution from the dynamic components must ultimately be finite, but it could be significant during certain intervals of fluid flow. Our work is continuing on defining the timings and amounts involved in different circumstances. Bearing in mind that the Darcyan permeability itself varies with strain, the role of dynamic permeability becomes important in any situation where fluid flow through actively deforming sediments is being considered. The ideas have been presented here in the context of accretionary prisms, but they are relevant to numerous other geological settings. The concepts are also relevant to engineering applications, for example the use of clays for waste containment (Arch, Stephenson & Maltman, in prep.), because the linings of many repositories are not static but undergoing small but continuous mechanical movement. In all such dynamic situations, the amount of fluid leaving the system depends not only on the conventional permeability, but a contribution from the system itself.
References
ARCH, J. 1988. An experimental study of deformation microstructures in soft sediments. PhD thesis, University of Wales, Aberystwyth. MALTMAN, A.J. 1990. Anisotropic permeability and tortuosity in deformed wet sediments. Journal of Geophysical Research, 95, 9035-9045. BEHRMAN, J.H., LEWIS, S.D., MUSGRAVE,R. & 25 OTHERS. 1992. Proceedings of the Ocean Drilling Program, Initial Results, 141. - - & - - AND THE ODP LEG 14I SHIPBOARD SCIENTIFIC PARTY. 1993. Tectonic controls on migration of fluids at the South American convergent margin near the Chile triple junction. Geofluids '93 extended abstracts, Torquay, 175177. BERNER, U. & FABER, E. 1993. Light hydrocarbons in sediments of the Nankai accretionary prism (Leg 131, Site 808). Proceedings of the Ocean Drilling Program, Scientific Results, 131, College Station, Texas (Ocean Drilling Program), 185-195. BROWN,K. &:MOORE,J.C. 1993. Comment on Arch & Maltman, Anisotropic permeability and Tortuosity in deformed wet sediments. Journal of Geophysical Research, 98, 17860-17864. BYRNE,T., MALTMAN,A.J., STEPHENSON,E., Son, W. & KNIPE, R. 1993. Deformation structures and fluid flow in the toe region of the Nankai accretionary prism. Proceedings of the Ocean Drilling Program, Scientific Results, 131, 83-101.
124
E.L. STEPHENSON ET AL.
CARSON, B., HOLMES, M.L., UMSTATTD, K., STRASSER, J.C. & JOHNSON, H.P. 1991. Fluid expulsion from the Cascadia accretionary prism: evidence from porosity distribution, direct measurements, and GLORIA imagery. Philosophical Transactions of the Royal Society, A335, 331-340. DAVIS, D.M., SUPPE, J. & DAHLEN, F.A. 1983. The mechanics of fold-and-thrust belts and accretionary prisms. Journal of Geophysical Research, 88, 1153--1172. FAAS, R.W. & CROCKETT, D.S. 1983. Clay fabric development in a deep-sea core: Site 515, Deep Sea Drilling Project Leg 72. Initial Reports DSDP, 72,519-525. FRYER, P., PEARCE, J.A., STOX-_>-~ ~_~ -~-=~==_==~_-~---=~-~------: ! -_--..Nu~g Fl~w ~ ~ ! ~(precipitationof ~ - _ ~ ~:_~.~E_-_~ ~_2-'quartzdue to ~ Hard Brittle Shale :-'~:_-."cooling) E~=E: ............ -=-: -__.--==]
After some precipitation of quartz cement by advective flOW
==-==~=-==~:~.=-= =~=~=~===~ ==-~~ ~: =_=-__,__::_~_-_~c~==_=_=:=:-:=:====~_==i
Sea Floor :=-------_-_--__----_-_--_-=_-_--=_.2_------; ............................... , ~Shai ~ - _ - - - - - - - - - - 1 L~_-_-..-_-_-_-_-_-_-_-_-~ -_-_-_-_-_--.-_-_-..-_-_-_-_-_-_-~ Precipitation t_=.-------------------.--------------------=========== Flux • Sin (X
ii::iii::i!;}'.;iii:i:iii!iiiii::iii!iii:::ii!iiii;.i .: Isotherm I ~ l ~ - - - - - - - - - - - - - - I- ~------------------':
~
lsotherm
Fig. 3. (a) Illustration of pore water flow from a fracture into adjacent sandstone beds. The rate of quartz cementation will be proportional to the flux of water and the rate of cooling. Most of the quartz cementation will occur in and close to the fracture where the flux is highest. The most permeable layers would receive the highest pore water flux and become cemented up first. (b) When fluids are flowing along a fracture or fault into a sandstone, the rate of cooling is a function of the flow rate and the angle between the direction of flow and the isotherms. Most of the precipitation in the sandstone will occur near the fracture since the angle cxbetween the isotherm and the bed becomes small in the case of relatively horizontal beds.
pressure may produce exceptions to this. At p H 5-6, the solubility gradient for calcite is steeper than for quartz, and as a result, the ratio of carbonate dissolution to quartz cementation in a rising and cooling volume of pore water is high (10-300) (Bj0rlykke & Egeberg 1993). The presence of even small amounts of carbonate cement in sandstones may therefore be used as evidence against introduction of quartz cement (Fig. 2). The source of quartz cement has been the subject of considerable discussions and several authors have suggested that quartz is introduced from an outside source through fluid flow (Gluyas & Coleman 1992; Robinson & Gluyas 1992; Grant & Oxtoby 1992). In addition
130
K. BJORLYKKE
would be reduced (Fig. 3a). This is not what is commonly observed. When water flows into sandstones from faults and fractures the rate of cooling is reduced and most of the precipitation occurs close to the fracture. Particularly in the case of relatively horizontal beds, the angle (a) between the isotherm and the direction of flow will be small, reducing the rate of precipitation for a given flux (Fig. 3b). The porosity distribution in sandstone reservoirs as recorded by core analyses or based on log measurements shows that the porosity is often rather unevenly distributed. Even in sandstones with relatively low porosities, thin beds with high porosities are often observed (Fig. 4). Adjacent beds have experienced the same temperature/pressure history during burial. When they show very different porosities and permeabilities, this is evidence of a compositional and textural control on compaction and quartz cementation. This is consistent with a model where most of the quartz cement is derived internally by pressure solution and other local diagenetic reactions. Mineralogical and textural parameters may also influence rates of precipitation of silica from pore water introFig. 4. Porosity distribution in sandstones buried to duced from an outside source. If the quartz more than 4000 m depth offshore Norway. The log cement had been derived from an outside source derived porosity (CPI log) to the left shows large by pore water flow, the pore water flux would be variations in porosities in the sandstone. The several orders of magnitude higher in the permeabilities (x) measured in the core show that permeable beds than in the low permeability highly permeable layers exist in the middle of much beds. It would then be difficult to explain how more cemented sandstones. The most permeable silica was transported into the low permeability layers have permeabilities of several hundred zones before the permeable zones had lost all millidacus while most of the more well-cemented sandstones have permeabilities around I mD or less. their porosity and permeability. This transport The rate of pore water flow would be about three would have had to be by diffusion and this orders of magnitude higher in the permeable assumes highly supersaturated pore water in the sandstone. Most of the quartz cement introduced main aquifers to provide a high concentration from an external source should therefore be expected gradient. The most permeable beds must then be to precipitate in these layers. characterized by clay coatings on quartz grains or other features inhibiting quartz cementation. Sandstones will continue to be cemented with to the problem of the large fluxes required, this increasing burial and temperature until nearly model would suggest that the rate of quartz precipitation was proportional to the vertical all porosity is lost and the permeability reaches extremely low values. This is easier to explain if component of the flux. The most permeable pathway for fluid flow would then receive most the silica is derived from an internal source. The rate of quartz cementation in a sandstone of the cement which is precipitated from solution due to cooling. This could be beds with will depend to a large extent on the rate of exceptionally high porosity and permeability, or subsidence and heating, and the time factor fractures and faults. These pathways will be probably plays a role (Bloch et al. 1990). In the Gulf Coast basin, Pliocene rocks start to become cemented up before significant cement can be introduced to the less permeable rocks, which quartz cemented at about 4 km depth while this receive a much lower pore water flux. In a occurs at about 3 km in older sedimentary rocks (Harrison & Summa 1991). In the North Sea and graded bed, the high permeability zone would be the Mid-Norway shelf (Haltenbanken) the Plioeither near the top (shallow marine sediments) or near the base (i.e. fluvial channels), and most cene/Pliestocene subsidence may exceed 1 km. A reservoir may then have subsided from 3 to of the quartz cementation would be expected to occur there such that the primary differences 4 k m and much of the most intense quartz
FLUID FLOW AND DIAGENESIS cementation would have occurred over a short period of time. Periods of rapid subsidence will also be characterized by high rates of oil generation and these processes are often linked (Gluyas et al. 1993). This may be mainly due, however, to the control temperature exerts on both processes and does not support the idea that quartz cementation is episodic in nature. There is also considerable evidence that quartz cementation may continue after oil emplacement in a reservoir (Walderhaug 1990; Saigal et al. 1992). At oil saturations above 50-60% the relative permeability of water is close to zero and import of silica through flow of pore water would then be more difficult.
Thermal convection Thermal convection has been suggested by several authors as a mechanism to move solutes through the subsurface (Wood & Hewett 1982, 1984; Davis et al. 1985; Cassan et al. 1981; Haszeldine et al. 1984). Rayleigh convection requires relatively thick sequences (100-300 m) of homogeneous and porous sandstones for the critical Rayleigh number to be exceeded (Bjerlykke et al. 1988). Mathematical modelling of Rayleigh convection has shown that relatively thin shales (0.1 m) or cemented intervals within a sandstone sequence will effectively divide potentially large convection cells into smaller ones, which may then be too small to exceed the critical Rayleigh number (Bj0rlykke et al. 1988). If the critical Rayleigh number is exceeded by as little as 10%, however, pore water flow due to Rayleigh convection will be fast enough to dissolve and precipitate 10% of the quartz within 10 Ma (Palm 1990). This demonstrates that if Rayleigh convection is taking place, the effects on diagenesis would be fast enough to be significant relative to diagenetic processes, but this is probably relatively rare. Non-Rayleigh convection will always take place when the isotherms are not perfectly horizontal. Slight slopes in the isotherms will be caused by dipping beds with different conductivity (Davis et al. 1985). The velocity of non-Rayleigh convection is proportional to the slope of the isotherm but also to the height of the convection cell. Where convection is confined to a few metres-thick sandstone beds separated by shales, the flow velocities will be too slow to be significant in terms of mass transfer during diagenesis (Bjcrlykke et al. 1988). A test would be that all carbonate minerals should be dissolved where significant quartz cementation occurs and that extensive corrosion of quartz and carbonate cementation should occur in other parts of the sequence (Fig. 2).
131
The salinity distribution observed in the North Sea Basin from formation analyses and from resistivity logs suggests that there is an increase in salinity with depth towards underlying evaporites. These salinity gradients show that there is at least at the present day no large-scale convective flow, since this would have homogenised the salinity (Gran et al. 1992).
Flow of water along fractures and permeable faults In a sedimentary basin with alternating sandstones and shales, there is a high permeability contrast between the two lithologies. In basin modelling, tight shales are assumed to have permeabilities which may be as low as 10-1°D, while the permeability of deeply buried sandstones may typically be 1-0.01D (Mudford et al. 1991). Faults and fractures will, therefore, to a very large extent control cross-formational flow. Fractures are continuous void space which may be important conduits for fluid flow. They are normally formed by tension set up during tectonic deformation or due to uplift and unloading. Relative movements along faults may also have a tensional component, so that voids similar to fractures are produced. Fractures are temporary features which after some time will be destroyed by mechanical deformation or become cemented up by precipitated minerals. For a fracture to remain open the rocks around the fracture must have sufficient shear strength to resist the stress imposed by the horizontal stress. Loose sands and soft mudstones therefore have limited potentials for preserving open fractures, since these rocks do not have sufficient strength to resist the horizontal stress. Sand normally does not become significantly quartz-cemented at temperature below about 80°C (2.0-2.5 km burial) and are then loose and poorly indurated when not cemented by carbonate or other types of early cement (Bj0rlykke et al. 1992). Carbonate rocks are more commonly subjected to early cementation and frequently develop a high shear strength capable of preserving fractures at shallow depths. Fractures will cement up by two different mechanisms. (1) Transport by diffusion of solids from surrounding rocks and precipitation of minerals in to the fracture. Minerals in the rocks on both sides of the fracture are under stress from the overburden, and, in some cases, also tectonic stress. Minerals such as quartz and calcite are
132
K. BJORLYKKE Water Boiling Temp (°C) 100 150 200 250 300 '
'
0 -50
Boiling
"x
- 1 oo \~
-15o ~\
- 2 0 0 ®o "10
~
- 2 5 0 -~ - 3 0 0 "-" ~
- 350
\\ ::
-400 -450 -500
Fig. 5. Hot water flowing on fractures will be cooled rapidly near the surface by mixing with ground water and by boiling. At 200° C boiling starts at about 130 m under hydrostatic pressure. This will favour concentrated precipitation of ore minerals like galena. The generated steam may cause severe corrosion of quartz and other minerals. then more soluble than minerals in the fracture, which are only subjected to pore pressure and no stress. There is therefore a thermodynamic drive for minerals to dissolve in the surrounding rock and to precipitate in fractures. The rate of this process, however, is difficult to calculate. (2) Precipitation of minerals by fluid flow through the fracture. The rate of precipitation can, in this case, be calculated, assuming different values for fluid flow rates, the temperature field (geothermal gradients), and mineral solubilities. It is also important to attempt to constrain the fluid flow rates and total fluxes along fractures in sedimentary basins. It is therefore necessary to examine the source of fluids and the mechanisms driving fluid flow.
Precipitation of sedimentary ores Even in very rapidly subsiding sedimentary basins with high sedimentation rates like the Gulf of Mexico, compaction-driven water does not produce very steep thermal anomalies and most of the heat flow is by conduction rather than advection (Harrison & Summa 1991). This is also the case for the North Sea (Hermanrud et al. 1991). This is because the compaction-driven pore water flow rate is on average lower than the subsidence rate and because open permeable fractures are rare in softer, relatively ductile sediments (Bjorlykke 1993). Flow rates on faults and fractures not extending up to the surface or sea floor will be reduced if the flow has to continue through low-permeability shales or
mud to reach the surface. A requirement for abnormally hot fluid to flow on fractures is that the flow rate is high so that most of the heat is not lost by conduction to the surrounding rocks. Thermal convection may produce high pore water fluxes but the rate of cooling and heating would be slow and gradual. Ore minerals and quartz would be precipitated along the limb of the convection cell where cooling occurs. Convective flow would, however, only redistribute what is present in the rocks inside the convection cells. Concentrated precipitation of dissolved minerals requires a change in the chemical environment by reactions with other fluids or minerals. We have shown that mixing of fluid is not likely to be very effective in the case of compactiondriven flow. The solubility, particularly of sulphides, is very temperature dependent and rapid cooling is the most effective way of precipitating minerals like galena and other sulphides (Deloule & Turcotte 1989). At depth rapid cooling of water flowing on fractures is difficult because the heat can not be disseminated away from the fracture fast enough because of the low thermal conductivity of the rocks. Near the surface hot fluids can be cooled more rapidly because the heat loss is higher. Fractures terminating on the sea floor will cause mixing with ocean water and precipitation of sedimentary ores. Sea water also represents a source of sulphur. Intersection with an open karstic system in limestone may also cause rapid cooling and precipitation of Mississippi Valleytype ore deposits (Bj0rlykke 1993). Also, cold
FLUID FLOW AND DIAGENESIS ground water in porous sandstone may cause rapid cooling. Boiling at 200 ° C will occur at 120-130 m depth and add to the heat loss (Fig. 5). Generation of steam may cause very strong corrosion of quartz and other silicate minerals. In the case of fractures which are terminated at some depth and overlain by a several hundred metre thick sequence of soft muddy sediments, the flow rate on the fracture is limited by the permeability of the overlying sediments because the flow has to continue to the surface. This means that the permeability on the faults and fractures in the brittle rocks may not be rate-limiting for the flow rate. The highest water flow rates and the most efficient precipitation of ores are therefore likely to occur near the top of brittle rocks before thick clay-rich sediments have been deposited. Compaction-driven pore water flow in sedimentary basins does not normally produce large thermal anomalies which would cause rapid cooling of fluids. Even in sedimentary basins with high sedimentation rates like the Gulf Basin thermal anomalies are moderate (Harrison & Summa 1991). Thermal anomalies associated with salt domes and overpressured shales are small compared to hot springs in basement rocks where brittle rocks are more likely to cause brecciation and sustain permeabilities on faults and fractures.
133 RAINFALL 100 cm / yr 108cm / m.yr
__
~
--~~--J.
. ~ ~ "
".:.::.~ I A
~UPWARDS FLOW ~ ~ / " 5 km= "~,~1¢ DUE TO COMPACTION % / . / ~
~'+. . . .
"''''+++"
I -Jdif. Inspection of equations 2, 3 and 4 demonstrates that membrane filtration should be favored in hydraulic settings where ( 7 P - pg) is large and oriented in the same direction as the salinity gradient. Among the largest natural fluid pressure gradients known are those which occur in the transition from hydropressured to overpressured fluid regimes in the south Louisiana Gulf Coast. Because this transition zone usually occurs within a shaly interval, this then should be the ideal setting for the production of brines by membrane filtration. Indeed, the Louisiana Gulf
Coast has often been cited as a type field area where membrane filtration is occurring (Graf 1982). As originally pointed out by this author (Hanor 1984), however, if membrane filtration were the dominant control on pore water salinities in the Louisiana Gulf Coast, then salinities should decrease upward through the deep overpressured sequence (Figs 2 and 6). It is obvious that the observed salinities do not correspond to trends that would be generated by membrane filtration. The saltiest waters are found a b o v e the zone of high hydraulic gradients and massive mudstones. Pore water salinities actually decrease with depth within the pressure transition and the geopressured zone. Although the mudstones exhibit membrane behaviour, as evidenced by their spontaneous potential response, upward flow of fluids through the overpressured zone may be taking place through fracture porosity and faults, rather than the matrix porosity of the sediments. The membranes are thus bypassed. This author is not aware of a single well-documented field example of the production of brines by membrane filtration.
160
J.S. HANOR
Marine aerosols: brines f r o m rain water The Murray Basin, southeast Australia, provides an example of an interesting variant on the role of subaerial evaporation of sea water in the formation of brines. The Murray Basin consists of an approximately 500m thick sequence of Palaeocene to Recent sediments deposited in fluvial to marine setting (Jones et al. 1994). The upper 100 m of the central portion of the basin now contains large volumes of subsurface brines with salinities in excess of 100 g 1-1 which have major solute compositions very much like that of marine and evaporated marine waters. The isotopic composition of many of these fluids, however, is that of evaporated meteoric water. This observation has led recent workers (e.g. Hanor & Evans 1988; Jones et al. 1994) to the conclusion that the solutes in these brines have been derived from the long-term accumulation of marine aerosols. The marine salts are introduced either in dry form or as meteoric precipitation. Because evapotranspiration greatly exceeds annual rainfall there is a net accumulation of solute as brackish surface waters. These waters are further evaporated to brines in lakes in the central part of the basin. The shallow ground water flow regime is not sufficiently dynamic to displace these brines, and they have been infiltrating downward and accumulating above a regional aquitard.
Other mechanisms The occurrence of saline waters in fractured Precambrian crystalline rocks in the Canadian Shield (Frape & Fritz 1987) and Hercynican granites of the Cornwall tin mines (Edmunds et al. 1987) has prompted the suggestion that the elevated salinities in these settings are the result of hydrolysis of Cl-bearing silicates, such as micas and amphiboles. While this mechanism is reasonable for the particular waters in question, it is less likely on mass balance grounds to be the source for the large mass of CI found in most sedimentary basins. Mills & Wells (1919) concluded on the basis of the common association of salty waters in gas-producing reservoirs that the saline waters are the residue formed as a result of the evaporation of pore fluid into gases produced in the subsurface. Russell (1933) showed by simple mass balance arguments that impossibly large volumes of methane would be required to produce significant quantities of brines through evaporation. It may, however, be necessary to invoke a mechanism such as this on a local scale to account for the extraordinary salinity of the
643000mg 1-1 CaCl2-brine from the Salina formation of the Michigan Basin described by Case (1945). The salinity of this water is far in excess of that which can be produced by normal subaerial evaporation or the dissolution of most evaporites. Other mechanisms of primarily historical interest are reviewed by Hanor (1983, 1988).
Reduction in salinity Not all formation waters have elevated salinities relative to their connate precursors. Examples of hyposaline waters in normal marine sediments occur in the US Gulf Coast, particularly in the deep overpressured section (Fig. 2). The apparent reduction of salinity may be due to incursions of meteoric water and/or the in situ production of HaO by the dehydration of mineral phases or organic matter during diagenesis.
Controls on major solute and isotopic composition Major cations and p H Most subsurface waters have major cation compositions which cannot be explained by either: (1) burial or infiltration of subaeriallyevaporated marine or continental waters, or (2) by subsurface dissolution of evaporites. There is, however, a growing body of evidence which supports the hypothesis that major cation compositions in formation waters are extensively buffered or influenced by an approach toward metastable thermodynamic equilibrium with ambient sedimentary mineral phases. Although there is up to an order of magnitude spread in the absolute concentrations of some of the major cations waters at any given chloride concentration, the first order trend for each cation over most of the salinity range is that of increasing concentration with increasing salinity (Fig. 7). An exception is Na, which decreases with increasing salinities beyond approximately 300000 mg 1-1. For moderately saline waters, the monovalent cations Na and K show an approximately 1:1 slope and the divalent cations Mg, Ca, and Sr an approximately 2:1 slope with respect to total dissolved solids on log-log scatter plots. The pH of these waters decreases with increasing salinity from typical values of 7-9 in moderately saline waters to values between 3 and 4 at high salinity, i.e., there is a progressive increase in the activity of H ÷, and the waters become more acidic. Such observations for
ORIGIN OF SALINE FLUIDS 1000000'
Sodium
+ooo ..................+................ , ................ +.....- ° ° 7
10000 1000"
2~
100
'2, .
Magr~esium
1000000"
'l
100000
_..I '~
161
,oooo.........
~
~
iooo
E d,,
[
t
1000
10000
: .........
+
.
.
11 100
1000000
.
.
.
.
.
.
.
i
:
10000'
Y 1000
I~o
I T
E Ioo-
......
"
"
i
,
~
w
..... l....
1
~......
i,1" _,i~ _~ '
10-
/
i .
t
f
O0
1000
10000
100000
,1
',
.
_J
•
-1000000
.
.
.
.
.
I I
i
.
.
.
.
.
.
.
.
-; ..............
100 .................. . . . . . . . . . . . . . . . . . . . . . . . .
O
10 2:1slop i e ~ i - - - - ~
Saturation i I
.1
1000000
100
1000
10000
100000
1000000
TDS, rng/L
1000000
Strontium
100000
+
"
=. 10000"
8
"IQ.
+
+ 100000
................. . ................ :-
~"
,
pH
9
+
i 10000
TDS, mg/L
10
+ ........... ++
Calcium
10000
~
+..-. . . . . . . . .
lOO°.i iiiiiiiiiiii:i i j
1oooooo
!
i
.......
TDS, mg/L
Potassium
100000' ................ +
; ~-Z';,iJ. .....
,~
TDS, mg/L
1000000'
" "
•
. i.~J~.=-
,f ..-y,++
[
100000
+
i .'.'~
.^
Saturation I ,
.1 IO0
~
~
..................~,................!.........~
/ 2 : 1 slope
1
i
~
1oo ..................
:+
10
i
i"
7 ................................
•
!"~
.!.~
...... Ill
..J "~
"
1000'
,.~ O0
6
................
4, ................
i
•
i
E
•.,
100'
.................
~....,.
......
° .4
.......
10'
s.
~
4'
i
3 100
1F '~. '
, 1000
10000
100000
1000000
TDS, mg/L
1 .1 100
1000
10000
100000
1000000
TDS, mg/L
Fig. 7. Scatter plots showing the covarience of log Na, K, pH, Mg, Ca and Sr with log TDS for typical saline waters. Sources of data listed in text. Note the 1:1 slopes for Na and K and the 2:1 slopes for the divalent cations.
waters in the salt d o m e province of the southern Louisiana Gulf Coast led this author (Hanor 1988) to suggest that the first order controls on pore fluid compositions in this region are the
subsurface dissolution of halite and subsequent buffering of pore water compositions by multiphase silicate-carbonate mineral assemblages. Dissolution of salt alone should produce a
162
J.S. HANOR
progressively NaCl-dominated fluid without the observed significant increases in K, Mg, Ca and
106 CalculatedF}uid Compositions i0 s .
St.
Metastable thermodynamic buffering. The idea
that the compositions of subsurface brines are at least partially thermodynamically buffered is implicit in the discussion of DeSitter (1947) and was specifically invoked as far back as Carpenter & Miller's (1969) work on saline water in the Ozark Dome. Thermodynamic buffering as a control on the composition of subsurface waters has been proposed as well for other basinal waters (e.g., Merino 1975; Nesbitt 1985; Hanor 1988; Land et al. 1988; Michard & Bastide 1988; Smith & Ehrenberg 1989; Hutcheon et al. 1993 and references therein). Testing this hypothesis by evaluating the saturation state of pore fluids with respect to common aluminosilicate and carbonate minerals is complicated by the fact that necessary data are often lacking, as is the case with many of the data sets used in this paper. Take, as an example, equilibrium involving albite: NaA1Si3Os + 4H+ = Na + + A13+ + SiO2 ° + 2H20
(5)
Calculating the saturation state of an aqueous fluid with respect to albite requires having reliable values for pH and dissolved sodium, aluminium, and silica, reliable data for the Gibbs free energy of formation of albite and all aqueous species, and an adequate thermodynamic model for the non-ideal behavior of the aqueous phase. Although significant improvements have been made in thermodynamic modeling of aqueous solutions (Helgeson et al. 1981; Pitzer 1987), methods for the treatment of the non-ideal behavior of some key species, such as carbonate and silica, in hypersaline, N a - C a dominated brines are still problematic. In addition, reliable analytical data for dissolved A1 and pH are often lacking. One technique for partially obviating these problems is to invert the process and calculate the bulk solution compositions which should exist in the presence of known or assumed mineral buffers and compare these with observed compositional trends. The results of one such calculation are presented in Fig. 8, which shows the variation in the concentrations of major dissolved species in fluids as a function of dissolved chloride in waters in equilibrium with the hypothetical mineral assemblage, quartz-muscovite-albiteK feldspar-calcite-dolomite at an arbitrary temperature of 100° C, a pressure of 1 bar, and a CO2 fugacity of 10 -3 bars. These constraints, plus the concentration of chloride, are sufficient
I
-r~ 104
.... ~ " -
~"
~
103
[
-~"" I .",
.Na
K
~
101 ¢0
100 10-1 103
J 104 102 Chloride, mg/kg H20
106
Fig. 8. Results of a computer simulation showing the calculated variation in major solute composition as a function of dissolved chloride for fluids buffered by a multiphase (see text) silicate-carbonate assemblage at fixed P, T andfco2 (Hanor 1994).
to fix uniquely the chemical potentials of dissolved species in the nine-component system: Na20-K20-MgO-CaO-A1203-SiOT-CO2HC1-H20 (or, alternatively, NaC1-KC1-MgCI~CaCI2-AICI3-SiO2-CO2-HC1-H20). Allowing CO2 to be mineral-buffered (c.f., Smith & Ehrenberg 1989) requires adding an additional solid phase, for example, chlorite, to the system. Hydrolysis constants for mineral phases and for CO2(g) were taken from Bowers et al. (1984). Conversion between aqueous activities and concentrations was made using a program which utilizes a Pitzer virial equation of state for aqueous solutions (Harvie et al. 1978; Pitzer 1987). While the particular simulation pictured here does not match all of the field data precisely, there are some important general similarities, including the progressive increases in dissolved Na, K, Mg, and Ca, and the decrease in pH and alkalinity with increasing salinity (Figs 7 and 11). Lower salinity waters are Na-dominated, but divalent cations, particularly Ca, become much more important constituents as salinity increases. By changing the buffer system, fco2, and/or T, it is in fact possible to shift the compositional trends of each of the cations and produce model waters dominated by Ca. It is highly unlikely on the basis of the complex and variable petrology of sedimentary rocks and the observed range in formation water compositions at any given chloride value and temperature that a single buffer system is responsible for the range of compositions observed in natural waters. Scatter in the field data sets presented here presumably reflect varying
ORIGIN OF SALINE FLUIDS
163
Quartz saturated, 2b'~C 12"
I
Albite +
-!-
Smectite
+~ z 6
--/--
--
/
I Chlorite
O4
"iv
10-
1
Od
r"
m o
14
I
K-feldspar CD
3 .Kaol
J
6 Kaolinite
mO
Albite
Ilite 2
I
0 0
3
6
9
log a (K+/H +)
2
I 4
I 6
8
log a (Na+/H + )
Fig. 9. Phase diagrams showing mineral stability as a function of the activity ratios Na/H, K/H and Mg/H 2. Fluid composition (black circles) is buffered by an arbitrary phase assemblage. An increase of TDS by an order of magnitude in moderately saline fluids requires approximately an order of magnitude increase in the most soluble cation, Na ÷ (Fig. 7). An increase in the activity of Na ÷ of an order of magnitude requires that the activities of H ÷ and K + also each increase by an order of magnitude and that the activity of Mg2÷ increases by two orders of magnitude. Hence the decrease in pH and the ca. I : 1 and 2:1 slopes on log plots of cations v. TDS on Fig. 7.
P-T conditions, different buffer assemblages, and department from equilibrium. Egeberg & Aagaard (1989) for example, have documented differences in composition between waters from carbonate and sandstone reservoirs in the Norwegian Shelf. Some waters may not be cation-buffered at all, including some saline waters from southwest Louisiana, whose compositions can most simply be explained by congruent dissolution of bulk salt dome evaporites (Hanor 1994). The necessity for having a large number of silicate mineral phases to buffer a multicomponent fluid system may mean that some formation waters in mineralogicallysimple sandstones and carbonates, which are the usual waters sampled, are strongly influenced by reactions occurring in ambient or intercalated mudstones and shales. Covarience in cation composition with salinity. The 1:1 covarience in the log-log plots of the concentrations of monovalent cations with TDS and the 2:1 covarience of divalent cations can be explained as follows. Na is the dominant cation in many subsurface waters simply because the silicate minerals which buffer it are in general the most soluble, not because NaCi may be the ultimate source. Because Na is the dominant cation up to salinities of approximately 330 g 1-1, it follows that an order of magnitude change in
TDS should be complemented by approximately an order of magnitude covarient change in Na. Hence the log-log plot of N a v . TDS has a slope of approximately 1 : 1. Consider next Fig. 9, which shows two generic phase diagrams which relate mineral stability and aqueous fluid composition. Equilibrium with respect to the hypothetical mineral assemblage, quartz, Na-smectite, albite, illite, and chlorite fixes the activity ratios (aNa+/aH+), (aK+/aH+), and (aMg2+/(aH+)2) at any given temperature and pressure. An increase in the activity of Na + of one order of magnitude requires an increase in the activity of H + of one order of magnitude, i.e., the solution becomes more acid. The stoichiometries of the coupled hydrolysis reactions buffering fluid compositions require an increase in the activity of monovalent ions, such as K +, of one order of magnitude, and the activity of divalent cations, such as Mg 2+, of two orders of magnitude. Even though the relations between the concentrations and the activities of the individual cations are strongly non-linear at elevated ionic strengths, broadly similar trends are reflected in the concentration scatter plots, and K + increases by approximately an order of magnitude with increasing salinity, but the divalent cations Mg 2+, Ca 2+, and Sr 2+ increase by two orders of magnitude. Sr may be buffered with respect to Sr-bearing calcites.
164
J.S. HANOR 10000
Southern Arkansas ....j~
1000 NorthSea~ ~ ,~;i S"~~: !'.~..~.]. ..........~.~j . , ~ " ~'~ ,. ,,"o ........' ° a . . ~ Z
100 ~..~i~..4~.~ " ~ 133
10 1'~ .........................
Sea Water Evaporatton Trend
SouthwestLouisiana
11 ................. lOOOO t ooooo Chloride, mg/L
1oooooo
Fig. IlL Br-C1 systematics for several sedimentary basins. The solid black line is the sea water evaporation trend (Fig. 4). Sources of data given in text. These relations fail at salinities above approximately 300g 1-1, where saturation with respect to halite is reached. Here, Na concentrations become progressively smaller with increasing salinity as a result of the increase in C1. There is an abrupt increase in K + as buffering shifts from a strictly silicate-carbonate assemblage to one involving halite as well.
Chloride and bromide Not all components in sedimentary formation waters are buffered by ambient mineral phases. The single most important example is chloride. Subsurface waters having salinities of 300 g 1- i o r less are undersaturated with respect to halite, sylvite, and other chloride-bearing mineral phases, and chloride concentrations in these waters are controlled primarily by mass transport processes of advection and dispersion, not by thermodynamic equilibrium with respect to one or more mineral phases. Even in the US Gulf Coast where a compelling case can be made for the origin of chloride by subsurface dissolution of salt domes and by burial of previously halite-saturated brines, most waters have salinities well below the >300g 1-~ levels expected at saturation. An important exception includes the Michigan Basin, where some pore waters are clearly halite-saturated (Wilson & Long 1992). Bromine is similarly non-buffered, and Br-CI systematics, although potentially complex (Hanor 1988), are thus more likely to provide clues as to the identity of saline endmember water types than other pairs of solutes. The high Br/C1 ratios of waters in the Smackover Formation, USA, for example, support the hypothesis that Br-rich, subaerially produced brines are
important saline endmembers in this natural system (Moldovanyi & Walter 1992) (Fig. 10). The low Br/C1 values of waters in South Louisiana, USA, in contrast, are consistent with the idea that high salinity is derived from the dissolution of halite-dominatred salt domes, which typically have low bulk Br/C1 ratios. The waters of the Norwegian Shelf have yet a third signature, one that supports the conclusion of Egeberg & Aagaard (1989) that there has been at least some contribution from subaerially produced brines.
Bicarbonate and sulphate The alkalinity of most of the formation waters shown in Fig. 11 is contributed predominantly by bicarbonate and organic acid anions. The latter species will be discussed in more detail later in this paper. Alkalinity in general decreases with increasing salinity. There are three probable reasons for this. First, the organic acid anions are associated primarily with low-salinity waters possibly derived by dewatering of organic-rich mudstones. Second, in a calcium carbonatebuffered system, carbonate alkalinity should decrease with an increase in dissolved calcium. Finally, the increase in H + with increasing salinity shifts dissolved carbonate and bicarbonate toward carbonic acid: HCO3- q- H + ~ H2C03 °
(6)
Sulphate shows little or no systematic variation with total dissolved solids. Recent work by McManus & Hanor (1988, 1993) provides clues as to why this is the case in the South Louisiana Gulf Coast. The isotopic composition of massive carbonate and pyrite cemented sands near the West Hackberry salt stock just south of Calcasieu Parish supports the hypothesis that the sulphur in the pyrite was derived from salt stock anhydrite and the carbon in the carbonate is from a methane source. The net reaction has involved the thermogenic reduction of sulphate and oxidation of methane and the consequent precipitation of iron sulphides and calcium carbonate. The concentrations of dissolved sulphate in this area thus controlled by relative rates of: (1) release of sulphate through salt dissolution, (2) solute transport and (3) removal by reduction. The later process, of course, is further dependent upon the availability of suitable reducing agents, such as hydrocarbons.
Dissolved aluminium and silica Dissolved Al concentrations in most subsurface waters are generally less than I mg 1-1 . There are
ORIGIN OF SALINE FLUIDS 10000
Alkalin LIty
~• •
_1 1000 O)
E
" "Z~
"-#
-
i
.................... ii i "
1000
..*.. '~'2"'Wll--i ~1~ ........d"J '~"ii,"........... *4"
__..1 e~
100
Sulfate
10000
.............. , ......
-
"
•0
i
165
100
....................... +
. . . . . . . . . . . . . . . . . . . . . . .
i
.,t ~]~'t"~
!-~,~
~
Q• '.o-11
L~
'
. . . .
.........i..................
r¢i
"
, . - -
10
i
1
........
1000
lOO
•
i
i
. . . . . . . .
10000
"•7-
i
.
--
100000
......
1000000
mg/L
TDS,
1
1 lOO
........ i ........ i . . . . . . . lOOO
lOOOO
lOOOOO 1000000
TDS, mg/L
Fig. 11. Variation in alkalinity and dissolved sulfate as a function of TDS for typical saline waters.
1000
J
Amorphous Silica Saturation
~
lOO
E 0 oo
.
buffered by silicate assemblages such as kaolinite-smectite.
..~
..ai" , ; ~
.I,,,~..~ ~•~i~7• lO
~- . .
uua~ m,. Saturation
•
/
end• ° o
o
T, deg C Fig. 12. Variation is dissolved silica versus temperature for typical saline waters. Curves show concentrations at quartz and amorphous silica saturation.
insufficient data of high quality at present to establish the systematics of dissolved A1 in deep sedimentary basins. However, simply based on the observation that Al-minerals are in general more soluble in low and in high pH fluids, and that there is a decrease in pH with salinity, one might expect that greater solubilization and transport of A1 could take place in waters of either very low or very high salinity, i.e., fresh waters associated with the dehydration of mudstones and halite-saturated brines. Many authors have noted that although quartz is a nearly ubiquitous phase in sedimentary basins, most basinal waters are not in thermodynamic equilibrium with quartz (Fig. 12). It is not known whether silica values are kinetically controlled or to what extent silica may be
Rates o f reaction. The absolute rates of diagenetic reactions, R, in basinal settings are in general not well known. Recent summaries of work on a number of important chemical systems are included in the collection of papers edited by Kharaka & Maest (1992). The evidence above, however, suggests that the rates of some hydrolysis reactions involving major species are sufficiently rapid, even at modest sedimentary temperatures, that the drive toward thermodynamic equilibrium between formation waters and ambient silicate and carbonate mineral phases is an important control on composition of formation waters. Isotopic composition
Recent investigations, unfortunately beyond the scope of this review, have established the importance of a wide variety of isotopic systems in establishing the sources of components in subsurface waters and the rates of fluid flow. Examples include isotopes of H, O, C, S, B, Sr, Nd, Sm, U, Th, C1, and I (see Banner et al. 1989, 1990 for examples). Of the many isotopic systems of interest in formation waters, 6D-51so systematics has been the most extensively used in problems of determining marine versus meteoric sources for H20 and for investigating the degree of water-rock exchange, particularly of oxygen. The H- and O-is•topic composition of formation waters has been reviewed by Sheppard (1986). The characteristic shift of values to the right of the meteoric water line, and the increase in 6180 values of formation waters
166
(.)
J.S. HANOR 20
(b)
/••,SMOW #
,.
-1000
q,
E =- -2ooo
o.'.-=.~. •
Oe 0
•
o
-40
•
-3000 /
-60
oOo
" - ..'.z
*~qP"**'' "
MWL . /
-20
0
/.
•
•
0
•
nnn
•
:'.".,o
""
° •
¢
-8O
-
o•
-4000"
• ••o
,
,
-5
0 8180
-5000
5
10
,
-10
0 ,5180
Fig. 13. D-O isotopic compositions of saline waters from master data set (a). O-is•topic composition of these waters as a function of elevation and, hence, increasing temperature (b).
with increasing temperature (Fig. 13) are thought to reflect exchange with and partial buffering of ~ 8 0 by silicate and carbonate minerals. In some sedimentary basins, there is a covarience in ~D and ~180 which may reflect mixing of light waters having significant meteoric component with heavier waters derived either from: (1) marine or evaporated marine precursors or (2) from waters having undergone isotopic exchange with ambient minerals. Other isotopic systems provide additional information on the degree of rock-water exchange. The 87Sr/86Srratios of formation water, for example, are typically higher than connate sea water values for the host sediments, reflecting in most cases the introduction of radiogenic Sr from the dissolution of silicates. There is a general decrease in 8VSr/86Srvalues of formation waters in the data sets utilized here with increasing Sr (not shown). It is possible that the Sr in the Sr-enriched waters, which are also the most saline waters, has been preferentially derived from carbonates, which should have less radiogenic values than siliciclastic minerals.
Chloride as a master variable in diagenesis The importance of the chemical potential of chloride, or, alternatively, HC1, in the thermodynamic interpretation of processes involving natural aqueous solutions was recognized over 20 years ago by Helgeson (1970). This general concept has been utilized by Giggenbach (1984) and others as a charge balance constraint in the analysis of high-temperature hydrothermal
processes. With the exception of work by Hanor (1988) in the Gulf Coast and Michard & Bastide (1988) in the Paris Basin, it has not attracted as much attention in lower-temperature basinal studies. In the specific examples discussed by Helgeson, CI activity was held constant and pH allowed to vary. In sedimentary basins, however, the case can be made on the basis of the discussion above that chloride is the more important master variable and that pH is more likely to be buffered. Figure 14 illustrates an example of the potential effect that changing chloride concentration through dilution, dissolution of a halide phase, or by mixing of waters of different salinity can have on driving diagenetic exchange of major solutes. Even though on log-log plots there is a striking linearity between the concentrations of major cations and TDS (and hence chloride), the variation in concentration of a rock-buffered cation with chloride is actually non-linear. In the example shown, waters A and B are mixed, producing a water of intermediate salinity and cation composition. On the assumption that this mixed water will attempt to achieve partial equilibrium with its surroundings, substantial amounts of Na will be dissolved and K precipitated. Dispersive dilution of a K-rich brine during regional fluid migration, for example, may be a reasonable hypothesis for the K-metasomatism, which occurred during Alleghenian regional fluid migration in eastern North America (Hearn & Sutter 1985). The amounts of the various solids which should be dissolved or precipitated as a result of changes in chloride concentration can be estimated using reactionpath calculations or mass balance techniques.
ORIGIN OF SALINE FLUIDS 120000
167
30000 @
100000
o °
•
80000
Oo
•
20000
-..:
_J tD
E 60000 d Z 40000
I
20000
¢"° •• IPCb
•
•
•
.
~')
//
A
E
• • •
,'.
/ •
"-
"" Q
0
..
,oooo
•
0
,
, 100000
, 200000
!
0J 300000
CI, m g / L
,, 0
• --~JSm~
' i
•
I
llLk, k*MSP. 100000 200000
300000
CI, m g / L
Fig. 14. Diagrams showing observed variations in dissolved Na and in K with chloride for saline waters. Changes in chloride concentrations by dilution, dissolution of salt, or by dispersive mixing (for example, mixing of waters A and B to produce C) should induce chemical disequilibrium and rock-water mass transfer. In the example shown, K-silicates should be precipitated and Na-silicates destroyed.
B r i n e s a n d organic m a t t e r
A number of dissolved constituents of subsurface aqueous fluids in sedimentary basins are profoundly influenced by the burial diagenesis of organic matter. Most important are those which are either direct or indirect participants in redox reactions. In turn, aqueous fluids can play a critical role in the redistribution of organic matter, particularly hydrocarbon fluids, in sedimentary basins (England et al. 1987). The details of the geochemical evolution of organic matter during burial diagenesis are beyond the scope of this chapter. We will briefly review, however, some of the net effects that reactions involving organic compounds can have on aqueous fluid geochemistry.
A q u e o u s organic species A wide variety of dissolved organic compounds have been found in sedimentary formation waters. These include minor amounts of neutral organic molecules of hydrocarbons, such as the alkanes, cycloalkanes, and aromatics (McAuliffe 1980) and charged anionic species, such as the mono- and dicarboxylic acid anions (Carothers & Kharaka 1978).
Dissolved hydrocarbons. As noted by McAuliffe (1980), there is a marked decrease in aqueous solubility with increasing carbon number for each class of hydrocarbons present in petroleum. Trends are nearly linear on a plot of log solubility versus carbon number. For the nalkanes, which have the general formula
C,,H2,,+2, there is an expotential decrease in solubility of at least six orders of magnitude from pure methane (n = 1) to pure dodecane (n = 12). There is a pronounced flattening out of solubilities beyond a carbon number of 12, with n-alkanes having C numbers from 12 to 36 being accommodated in water to approximately the same degree, i.e., 0.001--0.01 mg 1-1 for pure compounds. The cycloalkanes are slightly more soluble than the n-alkanes for a given carbon number (McAuliffe 1980), and the aromatics are roughly two orders of magnitude more soluble than the n-alkanes over the range n = 6 to 12. Solubilities of most dissolved hydrocarbons increase with increasing temperature, but decrease with increasing salinity (Price 1976).
Organic acid anions. Many aspects of the systematics of organic acid anion production and generation are still incompletely known. Within the data sets considered in this paper, acetate, CH3COO-, is the dominant dissolved organic species and exists in concentrations of up to approximately 2000mg 1-1. The concentrations of the other organic anions are in general much less than 100 mg 1-1 . Although concentrations of acetate exceeding 10000ppm have been reported by Surdam et al. (1984), some published high values appear to be analytically suspect (Hanor et al. 1993). There is a pronounced decrease in the range of concentration of acetate with increasing salinity. The apparent preferential association of acetate with low salinity waters may reflect their production in organicrich mudstones undergoing both mineral dehydration and maturation of organic matter. Based
168
J.S. HANOR
(a)
(b) 0
2000
+ o
Formate Acetate
0o -1000 ~ _~9, o
x Propionate A i-Butyrate
1500
[]
E il~"~ X X ~ o ff -200o~° d o o o
Butyrate
"~1~ ~= 1000
.0
o
.~
o
500
Oo
o
o
o
o~
o oO
0
200000
300000
400000
o o
-5000 / 0
TDS, mg/L
o
o q,Oo%°o
OoO
0
o , 1000
500
o cbo o
O oOO
%0® -4000 "~
==100000
o oO oo o~b
,:~oo OOo
,,,
0~ 0
Io o
-3o0o,
o
_
oo '1~O
o
Acetate,
1500
2000
mg/L
Fig. 15. Scatter plots showing variation in organic acid anions as a function of salinity (a) and dissolved acetate
as a function of depth (b).
on the data portrayed in Fig. 15, it would be a mistake to conclude that acetate is a major constituent in typical subsurface brines having salinites in excess of 100000 mg !-1 . Initial work by Carothers & Kharaka (1978) on VFAs in oil-field waters from California and Texas suggested that there is a temperature control on the distribution of VFAs. Subsequent work by these and other authors, however, has demonstrated that a several order of magnitude range in concentrations exists at all temperatures below 200 ° C. This spread is well illustrated in a plot of the variation in concentration in acetate versus elevation for the data sets considered here (Fig. 15).
Redox state o f subsurface brines Although free oxygen is normally not present in measurable quantities in deep subsurface environments, it is often convenient to write redox reactions with 02 as the electron acceptor and to relate redox state to the fugacity of oxygen, fo2. This will be convention followed here for many of the reactions considered. Most solid phases and waters deposited in continental and open marine sedimentary environments contain elemental components in relatively high oxidation states. Examples include Fe 3÷ in Fe-bearing mineral phases produced by weathering and S6÷ in dissolved sulphate. A notable exception is C in biologically-derived carbon compounds other than CO2. The nominal valence state for C in most such compounds and their burial derivatives varies from - 4 to 0, and organic carbon thus exists as a potential electron source for the
biologically-mediated or inorganic reduction of other elemental components in the system. An important example is the thermogenic reduction of sulphate to sulphide with hydrocarbons as the electron source (McManus & Hanor 1988). The approximate redox state of a sediment and its included pore water can sometimes be inferred from the presence of various authigenic mineral assemblages and/or the activities of dissolved species which may be in homogeneous or heterogeneous redox equilibrium. The presence of authigenic magnetite or pyrrohotite (McManus & Hanor 1988), for example, sets upper limits on probable oxygen fugacities of 10 -4° bars or less under normal basinal temperatures (Henley et al. 1984). Shock (1988) discusses evidence which suggests that there is metastable equilibrium between dissolved carboxylate and carbonate species in the oil-field brines he considered: CH3COO-
+ 202 = 2HCO3- + H +
(7)
and between dissolved carboxylate species: 2CH3CH2COO-
+ 02
=3CH3COO- +H +
(8)
but marked disequilibrium between carboxylate species and low carbon number atkanes, such as methane: CH3COO- + H + #COz + CH4
(9)
Helgeson et al. (1993) discuss evidence for the existence of metastable equilibria between higher molecular weight hydrocarbons, here represented by the liquid alkane C,,H2,,+2, and aqueous organic species, here represented by acetate, C H 3 C O O - :
ORIGIN OF SALINE FLUIDS 2C,H2,,+z(L) + (n+l)O2(g) = nCH3COO- + H + + 2H20
(10)
The above reactions, as noted by these authors, could be coupled to mineral-fluid equilibria through such reactions as: 2C,H2,,+2(L) + [(3n+ 1)/2102(8) + nCa2+ = nCaCO3 + 2nil + + nH20
(11)
and CH3COO- + 202 + 2Ca + = 2CACO3 + 4H +
(12)
As illustrated in the reactions above, many redox processes involve the production or consumption of hydrogen ions. In acidunbuffered systems, production of hydrogen ion may induce acid hydrolysis (equation 5). In pH-buffered systems, however, mineral-fluid equilibria may instead play an important role in buffering fo2, if stable or metastable equilibria exists between reduced and oxidized carbon compounds.
169
inclusions in ore and gangue minerals. Chemical analyses indicate that many of the fluids responsible for precipitating main stage mineralization in MVT deposits were Na-Ca-CI, low-sulphate brines having salinities in the range of 100-300 g kg -1 (Hanor 1979; Sverjensky 1984, 1986). It has generally been assumed that to qualify on mass balance grounds as a potential ore-forming fluid, a basinal fluid must contain on the order of 1 mg 1- i or more of the dissolved metal in question.
Metals in subsurface brines Within the data sets evaluated here Pb, Zn or Cu values are most likely to exceed 1 mg 1-1 in waters having salinities in excess of 200000mg 1-1 (Fig. 16). Such waters also have high chloride concentrations and low pHs. Possible reasons for the preferential association of metals with highly saline brines may thus include the solubilization of metals by chloride complexing and low pH, as for example: CuO + 2C1- + 2H + ~ CHCI2° 4- H2O (13)
Exchange with organic liquid phases Maturation of organic material often produces organic liquid and gas phases which may exchange components with a coexisting aqueous phase. One example is introduction of aqueous carboxylic acids by the oxidation of liquid alkanes, as noted in equation 10 above. Another is the loss of dissolved volatiles, such as methane, carbon dioxide, and hydrogen sulphide, through pressure release and degassing (Hanor 1988).
Ore-forming components in sedimentary waters
Ore-forming fluids Sedimentary formation waters have long been invoked as ore-forming fluids in a number of distinctly different geological settings. Although ore deposit classification schemes vary from author to author, general types of ore deposits which have been genetically associated with basinal fluids include (Force et al. 1991): (1) Mississippi-Valley type Pb, Zn, Cu, Ba, and F deposits; (2) shale-hosted Pb, Zn, Ba deposits; (3) rift-basin and redbed Cu deposits; (4) sandstone-hostedUdeposits. Much of the information available on the nature of the fluids which were involved in the genesis of these deposits comes from physical and chemical measurements made on fluid
High salinity by itself, however, does not guarantee a high dissolved metal content. Many of the most saline brines in which metals were measured have metal concentrations well below 1 mg 1-I. Considering the potential importance of the subject, there has been surprisingly little work done on the systematics of the distribution of metals in subsurface waters. An exception is the well-documented occurrence of Pb and Zn enriched brines of the Mississippi Gulf Coast (Carpenter et al. 1974; Kharaka et al. 1987), which provides clues as to other factors operating in the deep sedimentary environment which may control dissolved metal concentrations. A plot of modal concentrations of dissolved Pb and Zn versus stratigraphic interval (Fig. 17) supports the hypothesis of Carpenter and colleagues that these metals have been derived by reduction of iron oxide grain coatings in the Hosston Formation, a Lower Cretaceous red bed unit, possibly during reduction of iron oxide grain coatings which release Fe and other metals. Hanor (1988) has proposed that the configuration of the concentration gradients reflects vertical dispersive transport of these metals above and below this source bed. Low values in the Jurassic section below may reflect the presence of a sulphide sink. Many of the areas of the Upper Jurassic in central Mississippi contain HzS-rich natural gases, and it is possible that Pb and Zn are in the active process of being precipitated out of solution. It was early noted by White (1965) and Sawkins (1968) that the average K/Na ratio of
170
J.S. H A N O R
(b) 1000
300
o o: ;,o
100
o-. ~. %
10
°
._J
o %o,O. ,.o
.01 o
o°
o
a
200
~
s~o%~ ~%1m% ~=oO i ~ .ooi
E
• Pb (mg/L) o Zn (mg/L) Cu (mg/L)
.-I
" 100 • o
Pb Zn
&
Cu
r
.001 0
100000
200000
300000
TDS,mg/L
400000
200
0
400
600
Acetate,mg/L
800
1000
Fig. 16. Variation in dissolved metals in formation waters as a function of salinity (a) and dissolved acetate (b). Metals are generally more abundant in Ci-rich waters than in waters having high concentrations of organic acids. These data support the hypothesis that complexing by chloride is a more important control on metal concentrations.
o) 0
B
0
D~ o
Chloride
-~Wash.-Fred.-~
i/
,.- 0
Pb
[\
~~N~ n
o 0
Sligo ~ : . . v . : ~ . : , . . - . . .
.!i!!i!!!ii,i,i~to.~:i:Fimi:i,i.
3
1 /JJ
....
"Cottonvaiiey :::::::::::::::::::::
'::: "
'
'
'""
L
.............:.:.:.:.:.::::
!iiiiiiiiiiii:i:: _li>;ouano.,i!t °
I
100
200
g/I-
i k I
300 0
•
i
100
•
!
200
•
I
300
mg/L
Fig. 17. Vertical stratigraphic section showing variation in modal values of dissolved Pb and Zn by formation in the central Mississippi salt basin, USA. Data from Carpenter et al. (1974).
MVT ore fluids appeared to be higher than that of average subsurface brines. Hanor (1979) noted, however, that the elevated K/Na ratios consistent with K/Na ratios of more saline, deeper, and hotter fluids in basins. This is reflected as well in the present data set, in which the most saline waters have elevated K/Na ratios.
Organics as metal-complexing agents Interest in the role of dissolved organics as complexing agents for metals such as Pb, Zn and Cu has been prompted by the purported low solubility of metals in what have been taken to be typical formation waters. A plot of dissolved metals versus acetate for waters in the data set
ORIGIN OF SALINE FLUIDS 1000
precipitation would have sulphate concentrations less than 200mg 1-1 but could have salinities ranging anywhere from 10000 to 300000 mg 1-1.
l:~slope ! •
:.
i
100
.,i
E
d
..~ ~.~- 8 r,...e-,,&_.
10
Summary
lid
.........................
1
..°T~
1
171
10 100 S04, mg/L
1000
10000
Fig. 18. Variation in dissolved Ba as a function of dissolved sulfate for typical saline fluids. Note inverse relation reflecting possible buffering of fluid compositions by barite.
used in this paper, however, show a pronounced inverse correlation between metal content and acetate concentrations (Fig. 16). Metal concentrations well above I mg 1-1 occur only in low acetate waters. At acetate concentrations greater than approximately 50mg 1-1, metal concentrations are negligible. Because acetate is the most abundant organic anion in most of these waters, it is unlikely that the occurrence of high metal concentrations is in any way related to the concentration of acetate or any of the other organic species analyzed (see also Kharaka et al. 1987; Moldovanyi & Walter 1992). There is a simple explanation for this inverse correlation. High concentrations of organic acids are associated with less saline fluids; high concentrations of metals are associated with the most saline fluids. It is thus more likely that metal solubilities are enhanced by high chloride concentrations and low pHs than by organic anions, at least for the present data set.
Barium in subsurface brines There is no significant trend between barium and salinity, and barium, unlike its group IIA cousins Mg, Ca and Sr, is not typically buffered by silicate-carbonate equilibrium reactions. There is, however, a very rough inverse correlation between Ba and sulphate (Fig. 18) which is consistent with the hypothesis that equilibrium with respect to barite (BaSO4) may be one factor controlling Ba concentrations (see also Macpherson 1989). Not surprisingly, the highest Ba waters are low in sulphate. Fluids that would have the potential for transporting > 10mg 1-1 barium to the site of ore-mineral
Subsurface saline waters can be divided into three groups based on their salinity and anionic composition: (1) waters with anions other than C1 dominant. With possible rare exceptions, these are waters having salinities of less than 10000 mg 1-1. These include Na-HCO3 and Na-acetate waters; (2) Cl-dominated, halite-unsaturated waters. Typical salinity range 10000-250000300000 mg 1-1. These include Na-C1 waters and, at higher salinities, Na-Ca-CI waters; (3) Cl-dominated, halite-saturated waters. Typical salinities in excess of 300000mg 1-1. Ca and K become increasingly important solutes with increasing salinity. Na decreases with increasing salinity. Subaerial evaporation of marine and continental waters and the subsurface dissolution of evaporites have the potential for producing brines having both the range of salinities and dissolved chloride concentrations of most subsurface brines, but not their major cation compositions. No convincing field example of the formation of highly saline brines by membrane filtration has yet been documented. Although the concentrations of major solutes within subsurface saline waters range by a factor of ten or more at a given salinity, the broad systematic increase in dissolved Na, K, Mg, Ca and Sr and decrease in pH and alkalinity with increasing salinity support the hypothesis that the approach toward thermodynamic buffering by silicate-carbonate + (halide) mineral assemblages is a first-order control on subsurface fluid compositions, even at temperatures well below 100° C. The increasing dominance of dissolved Ca over Na and the increasing importance of K with increasing salinity, for example, can be explained simply by the stoichiometry of mineral-fluid hydrolysis, the thermodynamic nonideality of brines, and the buffering of Na concentrations by halite saturation at extreme salinities. The chemical potential of chloride or, alternatively, the aqueous concentration of anionic charge, is a master variable which ranks in importance with such other variables as pressure and temperature in driving fluid-rock exchange and controlling bulk fluid compositions. This variable in turn is controlled
172
J.S. H A N O R
largely by physical processes of fluid advection and dispersion. W h e r e the c o m p o s i t i o n of the fluid is largely r o c k - b u f f e r e d , its ultimate origin and its pathway of chemical evolution m a y be o b s c u r e d , at least in t e r m s of its m a j o r solute c o m p o s i t i o n , by its most r e c e n t P - T - X e n v i r o n m e n t . S o m e n o n - b u f f e r e d c o m p o n e n t s , such as C1 a n d Br, are m o r e likely to retain i n f o r m a t i o n on original e n d - m e m b e r fluid compositions. Dissolved organic acid anions are associated primarily with low salinity waters, but dissolved metals are preferentially f o u n d in brines having salinities in excess of 200 g 1-1 . T h e high chloride c o n c e n t r a t i o n and low p H of these waters may e n h a n c e solubilization of metals t h r o u g h chloride complexing. This work was supported in part by NSF Grants EAR-8803889 and EAR-9019342. I thank G. Bagnetto-Waters for her help with data reduction, J.A. Nunn and A. Sarkar for many valuable discussions on fluids in sedimentary basins, and A. Carpenter and W. Sanford for their helpful reviews of this script.
References BANNER, J.L., WASSERBURG,G.J., CHEN, J.H. & MOORE, C.H. 1990. 234U-238U-23°Th-232Th systematics in saline groundwaters from central Missouri. Earth and Planetary Science Letters, 101,296-312. - - , DOBSON, P.F., CARPENTER,A.B. & Moore, C.H. 1989. Isotopic and trace element constraints on the origin and evolution of saline groundwaters from central Missouri. Geochimica et Cosmochimica Acta, 53,383-398. BENNETT, S.S. & HANOR, J.S. 1987. Dynamics of subsurface salt dissolution at the Welsh Dome, Louisiana Gulf Coast. In: LERCH, I. & O'BRIEN, J.J. (eds) Dynamical Geology of salt and related structures. Academic Press, Orlando, 653-677. BERNER, E.K. & BERNER,R.A. 1987. The global water cycle, geochemistry and environment. PrenticeHall, Inc., Englewood Cliffs. BOWERS, T.S., JACKSON, K.J. & HELGESON, H.C. 1984. Equilibrium activity diagrams. SpringerVerlag, New York. BRAY, R.B. & HANOR, J.S. 1990. Spatial variations in subsurface pore fluid properties in a portion of southeast Louisiana: Implications for regional fluid flow and solute transport. Gulf Coast
Association of Geological Societies Transactions, 40, 53-64. CAROXHERS,W.W. & KnARAKA,Y.K. 1978. Aliphatic acid anions in oil-field waters - Implications for origin of natural gas. American Association of Petroleum Geologists Bulletin, 62, 2441-2453. CARPENTER,A.B. 1978. Origin and chemical evolution of brines in sedimentary basins. Oklahoma Geological Survey Circular, 79, 78-88.
& MILLER, J.C. 1969. Geochemistry of saline subsurface water, Saline County, Missouri. Chemical Geology, 4, 135-167. , TROUT, M.L. & PICKET, E.E. 1974. Preliminary report on the origin and chemical evolution of lead- and zinc-rich oilfield brines in central Mississippi. Economic Geology, 69, 1191-1206. CASE, L.C. 1945. Exceptional Silurian brine near Bay City, Michigan. American Association of Petroleum Geologists Bulletin, 29,567-570.
CONNOLLY,C.A., WALTER,L.M., BAADSGAARD,H. LONGSTAFFE, F.J. 1990. Origin and evolution of formation waters, Alberta Basin, Western Canada Sedimentary Basin. I. Chemistry. Applied Geochemistry, 5,375-396. DEMIR, 1. 1988. Studies of smectite membrane behavior: electrokinetic, osmotic, and isotopic fractionation processes at elevated temperatures. Geochimica et Cosmochimica Acta, 52,727-737. DESITrER, L.U. 1947. Diagenesis of oil-field brines.
American Association of Petroleum Geologists Bulletin, 31, 2030-2040. EDMUNDS,W.M., KAY,R.L.F., MILES,D.L. & COOK, J.M. 1987. The origin of saline groundwaters in the Carnmenellis granite, Cornwall (U.K.): further evidence from the minor and trace elements. Geological Association of Canada Special Paper, 33,127-144. EGEBERG, P.K. & AAGAARD, P. 1989. Origin and evolution of formation waters from oil fields on the Norwegian shelf. Applied Geochemistry, 4, 131-142. ENGLAND,W.A., MACKENZIE,A.S., MANN,D.M. & OUIGLEY, T.M. 1987. The movement and entrapment of petroleum fluids in the subsurface. Journal of the Geological Society, London, 144, 327-347. EUGSTER, H.P. & HARDIE,L.A. 1978. Saline lakes. In: LERMAN, A. (ed.) Lakes- Chemistry, Geology, Physics. Springer-Verlag, New York, 237-293. & JONES, B.F. 1979. Behavior of major solutes during closed basin brine evolution. American Journal of Science, 279,609-631. FISHER, J.B. & BOLES, J.R. 1990. Water-rock interaction in Tertiary sandstones, San Joaquin Basin, California, U.S.A.: Diagenetic controls on water composition. Chemical Geology, 82, 83-101. FORCE, E.R., EIDEL, J.J. & MAYNARD, J.B. 1991.
Sedimentary and diagenetic mineral deposits: a basin analysis approach to exploration. Reviews in Economic Geology, 5. FRAPE, S.K. & FRITZ, P. 1987. Geochemical trends from groundwaters from the Canadian Shield.
Geological Association of Canada Special Paper, 33, 19-38. GIGGENBACH, W.F. 1984. Mass transfer in hydrothermal alteration systems - a conceptual approach. Geochimica et Cosmochimica Acta, 48, 2693-2711. GRAE, D.L. 1982. Chemical osmosis, reverse chemical osmosis, and the origin of subsurface brines. Geochimica et Cosmochimica Acta, 46, 14311448. , MEENTS,W.F. & SHIMP,N.F. 1966. The origin of
ORIGIN OF SALINE FLUIDS
-
saline formation waters, III: Calcium chloride waters. Illinois Geological Survey Circular, 397, 60. HANOR, J.S. 1979. The sedimentary genesis of hydrothermal fluids. In: BARNES, H.L. (ed.) Geochemistry of Hydrothermal Ore Deposits (2nd ed.). John Wiley & Sons, New York, 137-172. 1983. Fifty years of development of thought on the origin and evolution of subsurface sedimentary brines. In: BOARDMAN,S.J. (ed.) Revolution in the Earth Sciences - A d v a n c e s in the Past Half Century. Kerdall/Hunt, Dubuque, 99-111. 1984. On the non-importance of membrane filtration as a mechanism for producing subsurface brines in the Louisiana Gulf Coast. Clay Mineral Society, 1984 Annual Meeting, Program with Abstracts, 55. 1987. History of thought on the origin of subsurface sedimentary brines. In: LANDA, E.R. & INCE, S. (eds) The History of Hydrology. American Geophysical Union, Washington, 81-91. -1988. Origin and migration of subsurface sedimentary brines. Society of Economic Paleontologists and Mineralogists Short Course 21. 1994. Physical and chemical controls on the composition of waters in sedimentary basins. Marine and Petroleum Geology, 11, 31-46. --, BAILEY, J.E., ROGERS, M.C. & MILNER, L.R. 1986. Regional variations in physical and chemical properties of South Louisiana oil field brines. Gulf Coast Association of Geological Societies Transactions, 36,143-149. -& EVANS, W.R. 1988. Subsurface brines from rain water: the salinization of the Murray Basin, Australia. Geological Society of America Abstracts with Program, 20, A364. -& SASSEN, R. 1990. Evidence for large-scale vertical and lateral migration of formation waters, dissolved salt, and crude oil in the Louisiana Gulf Coast. In: SCHUMACHER, D. & PERKINS, B.F. (eds) Gulf Coast Oil and Gases: Their characteristics, origin, distribution, and exploration and production significance. Proceedings 9th Annual Research Conference, Gulf Coast Section Society of Economic Paleontologists and Mineralogists Foundation, 293-296. HARVIE, C.E., MOLLER, N. & WEARE, J.H. 1984. The prediction of mineral solubilities in natural waters: The Na-K-Mg-Ca-H-CI-SO4-OHHCO3-H20 system to high ionic strengths at 25°C. Geochimica et Cosmochimica Acta, 48, 723-751. HEARN, P.P. & SUTI'ER,J.F. 1985. Authigenic potassium feldspar in Cambrian carbonates: evidence of Alleghenian brine migration. Science, 228, 15291531. HELGESON, H.C. 1970. Description and interpretation of phase relations in geochemical processes involving aqueous solutions. American Journal of Science, 268,415-438. - - , KIRKHAM,D.H. & FLOWERS,G.C. 1981. Theoretical prediction of the thermodynamic behavior of aqueous electrolytes at high pressures and -
173
temperatures, Part IV. American Journal of Science, 281, 1249-1516. - - , KNox, A.M., OWENS, C.E. & SHOCK,E.L. 1993. Petroleum, oil field brines, and authigenic mineral assemblages: are they in metastable equilibrium in hydrocarbon reservoirs? Geochimica et Cosmochimica Acta, 57, 3295-3340. HENLEY, R.W., TRUESDELL, A.H., & BARTON, P.B. 1984. Fluid-mineral equilibria in hydrothermal systems. Reviews in Economic Geology, 1. HITCHCOCK, E. 1845. Elementary Geology. Newman, New York. HUNT, T.S. 1879. Chemical and geological essays. Osgood and Co., Boston. HUTCHEON, I., SHEVALIER,M. & ABERCROMBIE,H.J. 1993. pH buffering by metastable mineral-fluid equilibria and evolution of carbon dioxide fugacity during burial diagenesis. Geochemica et Cosmochimica Acta, 57, 1017-1028. JONES, B.F., HANOR, J.S. & EVANS, W.R. 1993. Sources of dissolved salts in the central Murray Basin, Australia. Chemical Geology, I l l , 135154. KHARAKA,Y.K. & MAEST, A.S. (eds) 1992. Proceedings of the 7th International Symposium on Water-Rock Interaction. Balkema, Rotterdam. - - , CAROTHERS,W.W., LAW, L.M., LAMOTHE, P.J. & FRIES, T.L. 1987. Geochemistry of metalrich brines from central Mississippi Salt Dome basin, U.S.A. Applied Geochemistry, 2,543-562. LAND, L.S. 1987. The major ion chemistry of brines in sedimentary basins. American Institute of Physics Conference Proceedings, 154,160-179. 1992. Saline formation waters in sedimentary basins: Connate or diagenetic? In: KHARAKA,Y. & MAEST, A. (eds) Water-rock interaction. Balkema, Rotterdam, 865-868. ~, MACPHERSON, G.L. & MACK, L.E. 1988. The geochemistry of saline formation waters, Miocene, offshore Louisiana. Gulf Coast Association of Geological Societies Transactions, 38,503-511. LANE, A.C. 1908. Mine waters and their field assay. Geological Society of America Bulletin, 19, 501512. LUNDEGARD, P.D. & TREVENA, A.S. 1990. Sandstone diagenesis in the Pattani Basin (Gulf of Thailand): history of water-rock interaction and comparison with the Gulf of Mexico. Applied Geochemistry, 5,669-686. MACPHERSON, G.L. 1989. Sources of lithium and barium in Gulf of Mexico formation waters, USA. In: MILES, D.L. (ed.) Water rock interaction (WRI-6). Balkema, Rotterdam, 453-456. 1992. Regional variations in formation water chemistry: major and minor elements, Frio Formation fluids, Texas. American Association of Petroleum Geologists Bulletin, 76, 740-757. MARINE, I.W. & FRITZ, S.J. 1981. Osmotic model to explain anomalous hydraulic heads. Water Resources Research, 10, 89-107. MCAULIFFE, C.D. 1980. Oil and gas migration: chemical and physical constraints. American Association of Petroleum Geologists Studies in Geology, 10, 89-107.
174
J.S. H A N O R
MACPHERSON, G.L. 1989. Geochemistry of formation water, Plio-Pleistocene reservoirs offshore Louisiana. Transactions of the Gulf Coast Association of Geological Societies, 39, 421-430. MCCAFFREY, M.A., LAZAR, B. & HOLLAND, H.D. 1987. The evaporation path of seawater and the coprecipitation of Br- and K + with halite. Journal of Sedimentary Petrology, 57,928-937. MCMANUS, K.M. & HANOR, J.S. 1988. Calcite and iron sulfide cementation of Miocene sediments flanking the West Hackberry salt dome, Southwest Louisiana, U.S.A. Chemical Geology, 74, 99-112. - & ~ 1993. Diagenetic evidence for massive evaporite dissolution, fluid flow, and mass transfer in the Louisiana Gulf Coast. Geology, 21, 727-730. MERINO, E. 1975. Diagenesis in Tertiary sandstones from Kettleman North Dome, California - II. Interstitial solutions: distribution of aqueous species at 100°C and chemical relation to the diagenetic mineralogy. Geochimica et Cosmochimica A cta, 39, 1629-1645. MICHARD, G . & BASTIDE, J.-P. 1988. Etude geochimique de la nappe du Dogger du Bassin Parisien. Journal of Volcanism and Geothermal Resources, 35, 151-163. MILLS, R.V. & WELLS, R.C. 1919. The evaporation and concentration of waters associated with petroleum and natural gas. United States Geological Survey Bulletin, 693. MOLDOVANYI,E.P. & WALTER,L.M. 1992. Regional trends in water chemistry, Smackover Formation, southwest Arkansas: Geochemical and physical controls. American Association of Petroleum Geologists Bulletin, 76,864-894. NESBI1"r, H.W. 1985. A chemical equilibrium model for the Illinois Basin formation waters. American Journal of Science, 285,436-458. PHILLIPS, S.L., IGBENE,A., FAIR, J.A., OZBEC~:,H. & TAVANA, M. 1981. A technical databook for geothermal energy utilization. Lawrence Berkeley Laboratory Report 12810, University of California, Berkeley. PITZER, K.S. 1987. Thermodynamic model for aqueous solutions of liquid-like density. Reviews in Mineralogy, 17, 97-142. POSEY, H. & KYLE, R. 1988. Fluid-rock interactions in the salt dome environment: an introduction and review. Chemical Geology, 74, 1-24. PRICE, L.C. 1976. Aqueous solubility of petroleum as applied to its origin and primary migration.
American Association of Petroleum Geologists Bulletin, 60, 213-244. RICHARDSON, G.B. 1917. Note on Appalachian oil field brines. Economic Geology, 12, 37-41. RI~ENIJOtJSE, G. 1967. Bromide in oilfield waters and its use in determining possibilities of origins of these waters. American Association of Petroleum Geologists Bulletin, 51, 2430-2440. RUSSELL, W.L. 1933. Subsurface concentration of chloride brines. American Association of Petroleum Geologists Bulletin, 17, 1213-1228. SANFORt), W.E. & WOOD, W.W. 1991. Brine evolution and mineral deposition in hydrologically open evaporite basins. American Journal of Science, 291,687-710. SAWKINS, F.J. 1968. The significance of Na/K and C1/SO4 ratios in fluid inclusions and subsurface waters with respect to Mississippi Valley type ore deposits. Economic Geology, 63,935-942. SHEPPARD,S.M.F. 1986. Characterization and isotopic variations in natural waters. Reviews in Mineralogy, 16, 165-180. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16,886-890. SMITh, J.T. & EHRENBERG,S.N. 1989. Correlation of carbon dioxide abundance with temperature in clastic hydrocarbon reservoirs: relationship to inorganic chemical equilibrium. Marine and Petroleum Geology, 6,129-135. STUEBER, A.M. & WALTER,L.M. 1991. Origin and chemical evolution of formation waters from Silurian-Devonian strata in the Illinois Basin, USA. Geochimica et Cosmochimica Acta, 55, 309-325. SURDAM, R.C., BOESE, S. & CROSSLY, L.J. 1984. The chemistry of secondary porosity. American Association of Petroleum Geologists Memoirs, 37, 127-149. SVERJENSKY,D.A. 1984. Oil field brines as ore-forming solutions. Economic Geology, 79, 23-37. -1986. Genesis of Mississippi Valley Type ore deposits. Annual Reviews of Earth and Planetary Sciences, 14, 177-199. WASHBURNE,C.W. 1914. Chlorides in oil-field waters. American Institute of Mining Engineers Transactions, 48,687-693. WHITE, D.H. 1965. Saline waters in sedimentary rocks. American Association of Petroleum Geologists Memoirs, 4,342-366. WILSON,T.P. & LONG, D.T. 1992. Geochemistry and isotope chemistry of Michigan Basin brines: Devonian formations. Applied Geochemistry, 7 81-100.
Organic ligand distribution and speciation in sedimentary basin brines, diagenetic fluids and related ore solutions THOMAS
H. G I O R D A N O
1 & Y O U S I F K. K H A R A K A 2
I Department o f Geological Sciences, New Mexico State University, Box 3AB, Las Cruces, New Mexico 88003, USA 2 United States Geological Survey, 345 Middlefield Road, Menlo Park, California 94025, USA Abstract: The nature, distribution, and interactions of dissolved organic ligands in the shallower zones of sedimentary basins are highly variable and poorly understood. Better understood is the chemistry of dissolved aqueous carboxylic acid species in the deeper regions of sedimentary basins. In most oil-field brines, acetate is generally the dominant organic ligand and is followed in level of concentration by longer chained aliphatic monocarboxylicacid anions. The total concentration of dicarboxylic acid anions is probably lower than previously reported and likely less than 500 mg I-1 with succinate and glutarate being the dominant species. The highest concentrations of organic acid anions are present in formation waters at 80--120° C. Based on the available field and laboratory evidence, the concentrations of organic anions are controlled mainly by the rates of their generation from kerogen and their destruction thermally or by bacteria. An extensive compilation of experimentally determined stability constants and other thermodynamic data for a large number of non-humic organic ligands is available for temperatures near 25° C. However, corresponding high-temperature data are available for only a few species of interest; most notably, acetate complexes of some rock- and ore-forming metals and protonated species of several carboxylate ligands. Simulations of speciation in oil-field brines show that significant amounts of Ca, Mg, Fe, and Al can be complexed by carboxylate ligands, in the pH range 4-6, if ligand concentrations are on the order of 10-100 mg l-I. Calculated speciation in model ore fluids for Mississippi Valley-type deposits and red-bed related base metal deposits show that optimum conditions for Pb and Zn transport by organic ligand complexation are oxidized ore fluids with total reduced inorganic sulphide concentration less than 10 -9 molal.
The accumulation and alteration of organic matter within sedimentary basins and subsequent hydrocarbon generation are well documented processes (Tissot & Welte 1984). Less well understood are the roles of organic matter in the basinal processes of diagenesis and ore formation. The diverse interactions of organic matter in sedimentary diagenesis are discussed in some detail in two recent works (Gautier et al. 1985; Gautier 1986), while specific types of diagenetic interactions involving dissolved organic acids have been succinctly outlined by Kharaka et al. (1985) and Surdam & Crossey (1987). Dissolved organic acids can play an important role in sedimentary diagenesis because of the following chemical properties (Kharaka et al. 1985): (1) they can be the dominant source or sink for protons (H+); and thus, directly or indirectly control the pH and buffer capacity of subsurface waters; (2) as organic ligands, they form stable aqueous complexes with metals and other inorganic
species, thus enhancing their transport and modifying their behavior in diagenetic reactions; (3) they behave as reducing agents controlling the oxidation potential (Eh, fo2) of subsurface waters and the concentration of multivalent elements; and (4) they can be decarboxylated, thermally or with the aid of bacteria, to form carbon dioxide and hydrocarbon gases. The current interest in organic matter as an active chemical agent in ore-forming processes is based primarily on the presence of condensed organic compounds in a variety of sedimenthosted ore deposits formed at temperatures near or below 250 ° C (Giordano 1993; Parnell et al. 1993) and the inadequacy, in many cases, of inorganic mechanisms alone to satisfactorily account for transport and deposition of these ores. Most of these deposits are hosted by cratonic sedimentary rocks and were formed by hydrochemical processes within sedimentary basins. In Table 1 are listed five major ore deposit types which typically contain visible
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,175-202.
175
176
T.H. GIORDANO & Y.K. KHARAKA
Table 1. Major types of hydrochemical ore deposits containing visible organic matter and hosted by cratonic basinal sedimentary rocks (from Giordano 1993)
Ore deposit type
Tectonic environment
Host rocks
Syngenetic chemical
Intracratonic basins
Sandstone uranium
Intracratonic basins
Red-bed copper
Intracratonic basins
Mississippi Valleytype
Intracratonic basins
Sediment-hosted base metal
Incipient rift basins
Near shore marine: siliciclastic Continental: siliciclastic Continental and marginal marine: siliciclastic Shallowmarine and shoreline: carbonate and siliciclastic Shallow marine: carbonate and siliciclastic
Approx. deposition temperature (o C)
Ref.
25
1
50
2
50-100
3
75-200
4
100-275
3
( 1) Vine & Tourtel ot 1970; Tourtelot 1979; Coveney et al. 1987; Gra uch & H uyck 1990; Ripl ey et al. 1990; Schultz 1991; (2) Nash et al. 1981; Turner-Peterson & Fishman 1986; Turner-Peterson et al. 1986; Maynard 1991a; Hansley & Spirakis 1992; (3) Gustafson & Williams 1981 ; Boyle et al. 1989; Heroux et al. 1989; Landais & Meyer 1989; Ho etal. 1990; Maynard 1991b; (4) Macqueen & Powel11983; Gize 1984; Macqueen 1986; Sverjensky 1986; Price & Kyle 1986; Gize & Barnes 1987; Leventhal 1990; Henry et al. 1992. organic matter and that are unequivocally linked to sedimentary basins and related diagenetic processes. Specific roles of organic matter in ore-forming processes are similar to those for diagenesis and have been recently reviewed and outlined by Saxby (1976), Giordano (1985, 1993), Leventhal (1986), and Manning (1986). Giordano (1993) divides the potential roles of non-living organic matter in ore-forming processes into five categories: (1) aqueous complexation, (2) non-aqueous complexation, (3) substrate for microbial processes, (4) reducing or oxidizing agent, and (5) modifying the pH and other physicochemical parameters of the environment. Dissolved organic acids in ore fluids are involved to a greater or lesser extent in all five of the above processes. In this contribution, our main focus is on the aqueous complexing and protonation behavior of organic acid-ion ligands in sedimentary diagenetic fluids and related ore solutions. Organic matter may function up to magmatic temperatures as a reductant or oxidant. However, most of the roles for organic matter outlined above, including complexation and protonation, are limited to less than about 200°C (Barton 1982). The observational evidence discussed in Tissot & Welte (1984) and Giordano (1993) strongly suggests that organic matter is not a significant active agent in diagenetic and ore-forming processes operating near or above 350 + 50° C. For this reason, as well as thermal stability evidence discussed below, it is not likely that organic ligands
significantly contribute to metal and proton speciation in basinal fluids with temperatures greater than about 250 ° C. On the other hand, there is usually sufficient and often compelling evidence suggesting that organic matter is an active agent during normal diagenetic processes and in the genesis of those deposit-types listed in Table 1, as well as other low temperature (1000 mg 1-2 (as bicarbonate). Also, the acidified samples with organic anions generally give the characteristic rancid odor of butyric acid. The most definitive test in the field, however, is obtained by
2~
I
I 20
I
I 40
181
i
l 60
l
I 80
I
I00
VOLUME OF 0.050 N H2SO4 ADDED Fig. 2. Titration curves for sodium bicarbonate,
sodium acetate and an oil field sample. Note that the inflection point for oil field water is at pH ~ 3 which is similar to that of acetate but different from that of bicarbonate (from Carothers & Kharaka 1978).
examining the inflection points of the alkalinity titrations (Fig. 2). Samples with organic acid anions have an inflection point near pH 3.0 as compared to values near pH 4.5 for samples with no organic anions. The volume of acid titrant added between pH 4.5 and the inflection point of the sample gives some indication of organic alkalinity and of the concentrations of organic acid anions. Organic alkalinity, it is worth noting here, generally is much higher than bicarbonate alkalinity in some formation waters, especially those with temperatures from 80° to 120° C (Carothers & Kharaka 1978; Lundegard & Kharaka 1990). Water samples for analysis of dissolved organic species require special handling and treatment in the field (Lico el al. 1982; Kharaka etal. 1985). In order to minimize contamination, only glass, metal (e.g. stainless steel, copper) or teflon containers and tubing should be used for collection, filtration and storage of these samples. To reduce the loss of organic species by volatilization, agitation or aeration of water prior to analysis should be minimized, especially in samples with pH values lower than about 5.0. Bacterial degradation, however, is of greater concern and is likely to affect these samples unless preventive measures are taken (Kharaka et al. 1985; MacGowan & Surdam 1988). The measures should include filtering the sample through 0.1 p~m (or at least 0.45 ~m) teflon- or silver-filter paper and storage in sterilized amber glass bottles. A bactericide such as mercuric chloride (40 mg1-1 Hg), sodium azide or zephrin chloride should be added to the bottle before storage in a refrigerator at temperatures of about
182
T.H. GIORDANO & Y.K. KHARAKA
10,000 f
,
'
',I
~
r
r
I
'
r
,
-T"'
J 5000 i-
d ;ll~ \
Zone 1
Zone 2
•
-I
.J E .~ 1000-
~
soo
!
t
o
.~_ ~ 100
°° o~
5O
~.~
!
104 "ll'l
EXPLANATION ,
© • A ÷
Texas (Oligocene) California (Miocene to Cretaceous) Alaska (Permian to Triassic) Texas (Pleistocene)
I'I'L 60
i ~ ~
160 200 Subsurface
Temperature
220
(°C)
Fig. 3. Distribution of organic acid anions in oil-field waters from USA. Note that the highest concentrations are in waters from reservoirs at temperatures of 80-100°C (from Kharaka et al. 1988).
4°C. The samples should be analysed as soon as possible after returning to the laboratory. M o n o c a r b o x l i c acid a n i o n s Data on the concentration of monocarboxlic acid anions in formation waters from many sedimentary basins worldwide were reviewed by Lundegard & Kharaka (1990) to investigate the major controls on the distribution of these species. The samples were obtained from reservoir rocks ranging in age from Pennsylvanian to Pleistocene, although most were from clastic reservoirs of Cenozoic age with temperatures from 50° to 160° C. The main conclusions from this review were: (1) concentrations of acid anions in these waters are highly variable with respect to the major known controls, including subsurface temperature, age of the reservoir rock and the type and amount of kerogen in source rock; (2) the highest concentration of acetate reported is by MacGowan & Surdam (1988) at 10000 mg 1-1 (0.17 molal), but values higher than 5000 mg 1-1 are very rare; (3) maximum concentrations vary with temperature and are generally present in waters at 80°-120°C; (4) at similar temperatures, concentrations are generally lower in waters from older reservoir rocks. While variations in source material influence organic acid
anion concentrations, the main control on these concentrations appears to be the kinetics of decarboxylation (Lundegard & Kharaka 1990). The three temperature zones established by Carothers & Kharaka (1978) appear still useful in understanding the geochemistry of organic anions (Fig. 3), although some modifications are warranted, especially if data from other brines are included (Lundegard & Kharaka 1990). The concentrations of organic acid anions in zone 1 (temperature > propionate > butyrate > valerate (Carothers & Kharaka 1978; Lundegard & Kharaka 1990). Formate appears to be either absent (Carothers & Kharaka 1978) or to be present at very low concentrations in these waters (MacGowan & Surdam 1988; Fisher & Boles 1990). The relative abundance of acetate in zone 1 is highly variable with values that range from 0 to 90% of the total organic anions, with propionate being the dominant anion in waters with low acetate. The generally low concentrations of organic acid anions and the low relative abundance of acetate in the waters of zone 1 are generally attributed to the bacterial consumption of these anions, especially the acetate (Carothers & Kharaka 1978). Bacteria are known to survive at temperatures of up to about 130°C (Ehrlich 1990), but their operation is likely limited to about 80°C in the subsurface (Davis 1967; Carothers & Kharaka 1980). In a few petroleum fields the low total acid anions in formation waters from zone 1, which have anion proportions similar to zone 2, may result from dilution of waters high in anions that originated and flowed upward from zone 2 and mixed with shallow and depleted waters of zone 1 (Kharaka
50400
1186 88 9 17 bd bd 1300
42800
Acetate Propionate i-Butyrate n-Butyrate i-Valerate n-Valerate Total organic anions TDS 10000
271 35 13 5 5 bd 330
Lower McAdams Eocene 3520 141
Kettleman North Dome, CA 323-21J
912-2"
25200
bd 16 2 6 bd bd 24
Miocene 675 56
Olcese
113-28
Wheeler Ridge, CA
74WR-2"
14900
bd 11 1 3 4 5 24
Miocene 507 50
Valve
434-8
Wheeler Ridge, CA
74WR-6"
10900
4100 336 131 157 110 30 4860
Miocene 3337 149
Reef Ridge
21-15
San Emidio Nose, CA
74SEN-3"
157000
264 31 4 3 na na 302
Cretaceous 1965 68
Stanley
McCullough et al. Gas Unit 1
Soso, NI
84-MS-3 *
Pleistocene 3719 98 1525 226 26 34 19 12 1842 91000
Pleistocene 1876 55 12.4 1.9 0.0 0.0 11 0.0 25 80000
Jurassic 4350 121 15 1 2 0.0
18 333000
* From Carothers & Kharaka (1978) * From Kharaka et al. (1987). * From Kharaka et al. (1986).
na
na
Sand Sand
A-45-1
A-11-B
Jessie Allen No. 1-N
Norphlet
> Z
High Island, TX
High Island, TX
E. Nancy, NI
>
>
Z
> Z
t"
©
83-TX-12'
83-TX-2'
83-MS-15'
bd, below detection limit; na is not analysed; TDS is calculated inorganic and organic total dissolved solids; TR is trace detected.
195 TR TR TR bd bd 195
Oligocene 2624 94
Lower Weiting Oligocene 3960 154
Evans 'A' no. 1
Age Depth (m) Temp. (o C)
Bernard No. 6
Well name
Alto Loma, TX
Frio 'A'
Chocolate Bayou, TX
Field name Location
76-GG-30"
Production zone
76-GG-1 *
Sample No.
Table 2. Concentration (mg l -t) of monocarboxylic acid anions from selected petroleum fields in USA
184
T.H. GIORDANO & Y.K. KHARAKA
et al. 1986; Fisher 1987). The low concentrations may also result from low rates of generation of organic acid anions from the source rock (Lundegard & Kharaka 1990). The decrease in the concentration of organic acid anions with increasing temperature in zone 2 is generally attributed to decarboxylation of these anions (see later section for more details) to CO2 and natural gas (Kharaka et al. 1983; Drummond & Palmer 1986). The overall reaction for acetate is:
CH3COOH ~- CH4 + C02
(1)
The relative abundance of C2-C5 organic acid anions in zone 2 is likely controlled mainly by their rates of generation from precursor molecules within the parent kerogen, because the decarboxylation rates for these anions are rather similar (Palmer & Drummond 1986; Bell 1991). The decarboxylation rate for formate is much higher than that of acetate (Crossey 1991) which can explain the low concentrations of this anion in formation waters. Shock (1988, 1989) computed the stability fields of acetic and propionic acids and concluded that the concentrations of these and other acids in formation waters at temperatures >80 ° C were controlled not by decarboxylation, but by metastable redox reactions involving 02 and CO2. The reaction proposed between acetic and propionic acids was" 2C2HsCOOH + 02 ~ 3CH3COOH
(2)
Using selected values from Carothers & Kharaka (1978), Shock (1989) observed that a plot of aCH3COOHand aC2H5COOHhad a slope of 3/2 and plotted in the equilibrium stability field required by reaction (2). Using more extensive data sets for similar plots, Bell (1991) and Lundegard & Kharaka (1993) were unable to confirm the equilibrium of reaction (2) as they observed variable trends that rarely displayed a slope of 3/2 even for samples from the same sedimentary basin. Equilibrium was also not attained between acetic and propionic acids generated from hydrous pyrolysis of crude oils (Kharaka et al. 1993a). Because oxygen fugacities carry very large uncertainties for experimental as well as field situations, Kharaka et al. (1993a) computed the departure from equilibrium for the reaction: C2H5COOH + H2 ~- CH3COOH + CH4 (3) The possible attainment of stable or metastable equilibrium between organic acid anions and other aqueous and gaseous components of formation waters clearly requires additional
testing. Based on the available evidence, the concentrations of organic acid anions, in our opinion, are controlled mainly by the rates of their generation from kerogen and destruction thermally or by bacteria. Dicarboxylic acid anions In comparison with monocarboxylic anions, data on the concentrations of dicarboxylic acid anions in formation waters from sedimentary basins are much more limited, some reported values are controversial, and the total values reported range widely from 0 to 2640 mg 1-1 (Surdam et al. 1984; Kharaka et al. 1986, 1993b; Barth 1987; MacGowan & Surdam 1988, 1990b; Fisher & Boles 1990). Because of these uncertainties in the reported concentrations and because dicarboxylic acid anions generally form stronger complexes with A1, Fe, and other cations than do monocarboxylic acid anions, this section is covered in some detail. The first reported concentrations of dicarboxylic acid anions in oil-field waters were by Surdam et al. (1984). They reported values of up to 11 mg 1-1 for malonic acid and up to 26 mg 1-1 for maleic acid, but very little geological or geographical information was given. Kharaka et al. (1985, 1986) reported the concentrations of organic and inorganic species in formation waters from sandstone reservoirs of Pleistocene age from 11 petroleum wells in the High Island field, offshore Texas. They reported succinate (up to 63 mg 1-I), glutarate (up to 36 mg 1-1) and trace concentrations (1--6 mg 1-1) of C6 to C10 dicarboxylic acid anions (Table 3). The concentrations of oxalate and malonate in these waters were below the detection limit of about 1 mg 1-1. Methyl succinate and 2-methyl glutarate were identified by the GC-MS analysis employed, but these were not quantified. Detailed inorganic and organic chemical analyses of 20 formation water samples from six petroleum fields in central Mississippi Salt Dome basin were reported by Kharaka et al. (1987). The samples were obtained from sandstones and limestones of Cretaceous and Jurassic age. The total concentrations of monocarboxylic acid anions in these waters are low with less than 302 mg 1-1. Four samples were analysed by GC-MS techniques for oxalate and other (C3-C10) dicarboxylic acid anions; the concentrations obtained were all below the detection limit of about 1 mg 1-1. The concentrations of oxalate reported by Lundegard & Trevena (1990) were also below that detection limit in water samples from oil wells in the Gulf of Thailand. No other dicarboxylic acid anions were determined in these waters which contain
185
ORGANIC LIGANDS IN SUBSURFACE WATERS Table 3. Concentration (mg l-l) of dicarboxylic acids and acetate information waters from High Island field, offshore Texas (Kharaka et al. (1986) Well
Temp. (° C)
Acetate
A-10A A-14B A-45-1
66 77 98
148 222 1525
Sample 83-TX-3 83-TX-10 83-TX-12
Butanedioic Pentanedioic Hexanedioic Heptanedioic Octanedioic Nonanedioic Decanedioic acid acid acid acid acid acid acid 1.8 0.2 63
1.7 1.2 36
0.5 0.5 bd
bd 0.6 0.2
bd 0.7 5.0
bd 1.4 6.0
bd bd 1.3
bd, Below detection limit.
Table 4. Concentration (mg 1-1) of dicarboxylic acid anions and acetate in watersfrom North Coles Levee (NCL) and Paloma (P) fields, California (Kharaka et al. (1993b) Sample 92NCL-10 92NCL- 11 92P-13 92P-14
Well
Field
Production zone
Temp. (° C)
Acetate
34--31 12-31 66X-2 24X-11
NCL NCL P P
Stevens Stevens Stevens Stevens
93 95 ! 16 114
3296 2945 1281 595
abundant monocarboxylic anions (up to 1500mg1-1 acetate) and that were obtained from reservoirs with relatively high temperatures (120-177 ° C). Barth (1987) reported 38 mg 1-1 oxalate and 10 mg 1-~ malonate in one sample of water from North Sea oil wells. Because this sample was nearly devoid of monocarboxylic acid anions and dicarboxylic anions were not detected in other samples, she concluded that the measured dicarboxylic anions were not present in natural formation water, attributing them instead to contamination by organics in drilling mud. The highest concentrations of dicarboxylic acid anions are those reported by MacGowan & Surdam (1988, 1990b) for water samples from about 40 petroleum wells located mainly in the San Joaquin, Santa Maria and northern Gulf of Mexico basins. They reported values of up to 494 mgl -~ oxalate, 2540 mg 1-1 malonate and 66 mg 1-1 maleate. Well 12-31 in the North Coles Levee field, San Joaquin basin, California has the distinction of producing water with the highest reported concentration of dicarboxylic acid anions (2640 mg 1-1) (MacGowan & Surdam 1988). Because of the high organic content, this well and several wells from the North Coles Levee and the nearby Paloma field, were re-sampled by several other investigators (MacGowan & Surdam 1990b; Fisher & Boles 1990; Kharaka et al. 1993b). For samples from well 12-31, MacGowan & Surdam (1990b) reported much lower values for the concentrations of oxalate (17 mg 1-1), malonate (42 mg 1-1) and maleate (below detection limit). Fisher & Boles (1990) detected no oxalate or malonate in the
Methyl Oxalate Malonate Succinate succinate 0.07 0.16 0.10 0.05
, Mg2+, I-, CI-, Br-, 8042AI, K, Si, Mg, Ca, Na in clay minerals; trace elements in clay minerals, e.g. Sc, REE Ba as BaSO4 in drilling muds; As in bactericides; Hg in oils flashed using liquid Hg
Table 2. Variation of nickel, vanadium and iron abundances in oils and bitumens with source-rock type and depositional environment
Oil type and location Condensate; Kapuni, Taranaki Basin, NZ (CretaceousCenozoic) (ref. 1) Light oil (non marine); Tariki, Taranaki Basin, NZ (Cretaceous-Cenozoic) (ref. 1) Crude oil (marine); Gilby, Mannville Group, Alberta Basin (L. Cretaceous) (ref. 2) Heavy biodegraded oil (marine); Boscan, Venezuela (Cretaceous) (ref. 3) Bitumen, New Albany shale, Indiana (marine; MississippianDevonian) (ref. 3) Bitumen, Green River shale Mahogany Zone, CO, (lacustrine; Eocene) (ref. 4)
Ni(ppm)
50 ppm) High-medium (1-10 ppm) Low ( 150 ° C to saline, low-temperature waters ( > 1 molal, c. 150 ° C). The discriminatory temperatures are not rigid, and some workers also use the term 'intermediate' to indicate reservoir temperatures in the 120-180 ° C range. Low-temperature systems can only be used for 'direct-use' applications (e.g. heating), while high-temperature systems can be used for electricity generation as well as direct-use applications. The heat source for the system is a function of the geological and tectonic setting. If the driving heat flux is provided by a magma, then such systems are termed volcanogenic. They are invariably high-temperature systems. Heat does not, however, have to be supplied by a magma, and a geothermal system can be generated in areas of tectonic activity. For example, heat may
be supplied by the tectonic uplift of hot basement rocks, or water can be heated by unusually deep circulation created by folding of a permeable horizon or faulting. These are termed non-votcanogenic systems and can contain high or low-temperature reservoirs.
Geothermal fluids Geothermal fluids are classified according to the dominant anions. Although not a formal genetic scheme, this descriptive classification does permit some generalizations to be made on the likely origins of the waters. The compositional range and variety of geothermal fluids is primarily a consequence of their different formational processes. Primary (deep) fluids evolve through rock-water reactions, while secondary fluids are produced from steam which is formed by boiling of the deep fluid. Characteristics of the three dominant types of geothermal fluid are discussed below; more detailed consideration of geothermal fluid chemistry may be found in books by Ellis & Mahon (1977), Henley et al. (1984)
224
K. NICHOLSON
and Nicholson (1993). The conceptual structure of geothermal systems, and the inter-relationship of the fluid types is illustrated in Fig. 2.
The primary fluid The most common type of fluid found at depth in high-temperature geothermal systems is of nearneutral pH, with chloride as the dominant anion. Other waters encountered within the profile of a geothermal field are commonly derived from this deep fluid as a consequence of chemical or physical processes.
Chloride fluids. The evolution of the primary chloride fluid in dynamic, liquid-dominated systems can be summarised as follows. Meteoric waters penetrate the crust through permeable zones and circulate to depths of up to around 5-8km (Henley 1985). As they descend, they are heated, react with the host rocks and rise by convection. At depth, the fluids typically contain 1000-10 000 mg kg -1 CI at temperatures of about 350 °C. The 'soluble-group' elements are leached from the host rocks by the waters, along with other elements whose solubilities are controlled by temperature-dependent reactions (Ellis & Mahon 1977). These reactions change the primary mineralogy of the host rocks to a distinctive alteration assemblage characteristic of the fluid and its temperature. The fluids are retained within a permeable horizon forming a reservoir in which mineral-fluid equilibria are approached, and a suite of secondary alteration minerals is established. Chloride concentrations are usually in the thousands of milligrams per kilometre range, up to about 10000mg kg -I. Less commonly, CI levels exceed 100000rag kg -~ in more saline systems (e.g. Salton Sea, California, USA). In such systems formation waters or seawater may have mixed with the original chloride fluid. Other main constituents include sodium and potassium (often in a c. 10:1 concentration ratio) as the principal cations, with significant concentrations of silica (higher concentrations with increasing temperature at depth) and boron. Sulphate and bicarbonate concentrations are variable, but are commonly several orders of magnitude less than that of chloride. Carbon dioxide, and much lower levels of hydrogen sulphide, are the main dissolved gases. Areas which contain hot, large-flow springs with the greatest C1 concentration are fed more directly from the deep reservoir, and identify permeable zones within the field. However, these areas may not necessarily overlie the major upflow zone since the local topography and
lateral permeable horizons can exert a significant control on the hydrology. Such discharge features are invariably surrounded by silica sinter, which provides a guide to extinct hot springs, pools and geysers, and to subsurface temperatures in excess of c. 200°C. High-gas fields with significant bicarbonate concentrations can deposit an intimate mixture of sinter and travertine. Sub-surface silicification is commonly produced by chloride fluids, while the alteration assemblage is of argillic-propylitic type, with the following characteristic minerals: silica (amorphous silica, cristobalite, quartz) albite, adularia, illite, chlorite, epidote, zeolites, calcite, pyrite, pyrrhotite and base-metal sulphides (Browne 1978, 1991).
Origin of water and solutes. The water constituting the geothermal fluid can be derived from a number of sources. It may represent surface (meteoric) water which has gained depths of several kilometres through fractures and permeable horizons, or it can be water which was buried along with the host sediments (formation or connate waters). Other sources of water in geothermal systems have been suggested; these include waters evolved in metamorphism (metamorphic waters) and from magmas (juvenile waters), but the importance of these sources of water is uncertain (White 1957). Magmas were initially thought to be the source of the water, solutes and heat of geothermal systems. However, this model was radically changed in the early 1960s when it was demonstrated that the fluids were of dominantly meteoric origin, and solutes could be derived from rock-water reactions. Work on the isotopic signature of the fluids by Craig (1963) showed that they had the same deuterium signature as that of local meteoric water and could not be magmatic. In a series of now classic studies in aqueous geochemistry, Ellis & Mahon (1964, 1967) and Mahon (1967) demonstrated that all the solutes in geothermal fluids could be derived from reactions between the meteoric groundwater and the host lithologies. Later experiments with seawater and basalt (e.g. Bischoff et al. 1981) produced solutions of similar chemistry to seawater-influenced geothermal systems such as those in Iceland. Rock-water reaction is therefore thought to be the major source for many of the solutes, although they may also be contributed by mixing with formation waters, seawater or magmatic fluids. While there is no doubt that the geothermal fluids are of a predominantly meteoric origin, there is sufficient latitude in the isotope data to permit 5-10% of the fluid to be from an
GEOTHERMAL FLUID CHEMISTRY alternative source, possibly a magmatic fluid. Mixing with even a small amount of magmatic fluid would significantly affect the chemistry of the final geothermal fluid, and isotope determinations cannot discount a magmatic contribution subsequently diluted by meteoric waters. However, mass balance considerations using typical values for water-rock ratios, the chloride content of host rocks, the concentration of chloride in waters from high-temperature systems, the aerial extent of geothermal systems, and the duration of geothermal activity indicate that unrealistically large volumes of rock would have to be leached over the lifetime of a geothermal system. A small but significant magmatic contribution to the geothermal fluid is therefore thought to be likely. Density differences would, however, preclude any intimate mixing between meteoric waters and a magmatic fluid. If small pulses of magmatic fluid did enter the geothermal convection cell then, while not detectable isotopically, they would make a major contribution to the solute composition. Such fluids would be at temperatures in excess of 400°C and be rich in solutes such as C1, SO2 and CO2. Although the extent to which mixing may occur is uncertain, recent analytical advances now make it possible to distinguish between the isotopes of chlorine and boron. Information from these isotopes may enable a model of magmatic fluid-meteoric water mixing to be derived.
Secondary fluids As the chloride fluids leave the reservoir and ascend to the surface they may boil to create a two-phase (steam + liquid) boiling zone. The residual chloride water can discharge at the surface in hot springs or travel laterally to finally emerge many kilometres from the upflow zone. The vapours from this boiling zone may migrate to the surface independently of the liquid phase and discharge as fumaroles. Alternatively, the vapours may dissolve in groundwaters or condense in the cooler ground to form steamheated, acid sulphate and/or bicarbonate waters.
Sulphate waters. Also known as 'acid-sulphate waters', these are invariably superficial fluids formed by the condensation of geothermal gases into near-surface, oxygenated groundwater. The gases, with steam and other volatiles, were originally dissolved in the deep fluid but separated from the chloride waters following boiling at depth. They are found on the margins of a field some distance from a major upflow area, at
225
topographic levels high above the water table, in perched water tables and over boiling zones. Although usually found near the surface ( < c. 100m), sulphate waters can penetrate to depth through faults into the geothermal system. Here they are heated, take part in rock alteration reactions and mix with the ascending chloride fluids. Other sources of sulphatechloride mixed fluids are discussed by Ellis & Mahon (1977) and Nicholson (1993). Sulphate is the principal anion, and is formed by the oxidation of condensed hydrogen sulphide H2S(g) + 202(aq) = 2H+(aq) + SO42-(aq). This reaction, and the condensation of carbon dioxide, CO2(g) + H20(1) = HzCO3(aq) = H+(aq) + HCO3-(aq) = 2H+(aq) + CO32-(aq), produces protons, creating acid waters. Chloride occurs in trace amounts. Bicarbonate is either absent or at low concentrations since in very acid waters the dissolved carbonate is usually lost from solution as carbon dioxide gas (back-reaction of above equilibrium). Other volatile constituents which separate from the deep fluid on boiling may also condense into these waters (e.g. NH3, As, B) and attain significant concentrations. Nearsurface reactions between the acid waters and the surrounding county rocks may leach silica and metal cations (Na, K, Mg, Ca, A1, Fe etc.) which can thereby attain high concentrations in the waters. Acid-sulphate waters react rapidly and leach the host rocks to produce advanced argillic alteration, with kaolinite, halloysite, crystobalite and alunite as diagnostic minerals (Browne 1978, 1991). Extensive leaching of surface lithologies by these waters, or acidic steam, can produce a silica residue. This must be differentiated from silica sinter, which is a product of depositional, not alteration, processes. Anhydrite, hematitie, dickite, jarosite, pyrite, goethite-hematite mixtures and native sulphur are also commonly found.
Bicarbonate. These waters, which include those termed CO2-rich fluids and neutral bicarbonatesulphate waters, are the product of steam and gas condensation into poorly-oxygenated subsurface groundwaters. Such fluids can occur in an umbrella-shaped perched condensate zone overlying the geothermal system, and are common on the margins of fields. Bicarbonate waters found in non-volcanogenic, hightemperature systems (e.g. Hungary, Turkey and Africa) may constitute the deep reservoir fluid. The waters are of near-neutral pH as reaction with the local rocks (either in the shallow reservoir or during lateral flow) neutralises the initial acidity of these waters (see above
226
K. NICHOLSON
carbonate equilibrium). Loss of protons in such reactions produces near neutral waters with bicarbonate and sodium as the principal constituents. Sulphate may be present in variable amounts, with chloride at low concentrations or absent (Mahon et al. 1980). These waters are highly reactive, and their corrosive action on well casings needs to be taken into account in the development of a field (Hedenquist & Stewart 1985). These waters can precipitate extensive deposits of travertine (CaCO3), and can be. indicative of subsurface temperatures of below c. 150°C. An argillic alteration assemblage is typical of these waters, with the formation of clays (kaolinite, montmorillonite) and mordinite; some calcite and minor silicification may also occur (Browne 1978, 1991).
Steam formation: boiling point-depth relations. As a geothermal fluid ascends towards the surface, the pressure imposed upon it by the overlying column of water (hydrostatic pressure) will decrease. Eventually, the pressure will drop to a level which permits the dissolved gases and steam to separate from the liquid phase. This phase separation is commonly referred to as 'boiling'. It is one of the most important processes controlling the chemistry of liquid and vapour (i.e. water and steam) discharges. The relationship between boiling point and depth has been described by Haas (1971) and is illustrated in Fig. 3. The curve indicates the maximum temperature which water can attain at any given depth (or pressure), and therefore shows the depth at which a reservoir water at a given temperature will commence boiling. From this depth the boiling, or two-phase zone, can extend upwards towards the surface. The curve assumes that only hydrostatic pressure acts upon the fluid. In practice, however, it has been found that hydrodynamic pressures exist at depth in a geothermal system at about 10% above hydrostatic pressure. This excess pressure is necessary to maintain flow through the system. It is created by the buoyancy of hot water relative to cold water recharge and by a hydrostatic head in recharge waters from areas of greater relief (Grant et al. 1982; Henley 1985). This means that higher temperatures can exist at shallower depths than indicated by the curve, and therefore that boiling will occur at shallower depths. An increase in the salinity of water lowers the vapour pressure of water, raises the curve and prevents boiling until shallower depths are attained (Sutton & McNabb 1977). However, for most geothermal systems, the fluids are dilute and small changes in salinity will not
Temperature (*C) 1 O0
200
300
400
800
1200
1600
4.4% C 0 2
I
\
(1.0m) .]~
Fig. 3. Boiling-point-depth relationship under hydrostatic conditions (Henley 1985). Note that increases in salinity and gas content have opposite effects on the boiling-point-depth profile. For a fluid at a given temperature, increasing the salinity prevents the fluid boiling until shallower depths are attained; by contrast, increasing the gas content permits the fluid to boil at greater depths. Note too that increasing the salinity has only a slight effect on the boiling profile, while relatively small increases in the gas content of the fluid significantly alters the boiling-depth relationship.
significantly alter the boiling point-depth profile of the system. More significant however, is the gas content of the fluid. The presence of several weight percent gas in the fluid will depress the isotherms in a system below the usual boiling point-depth curve. This means that boiling zones in high-gas systems will appear at far greater depths than for gas-poor systems, which follow the relationship for pure water (Fig. 3).
Hydrological profile of geothermal systems An idealized thermal, hydrological and chemical profile of a high-temperature liquid-dominated dynamic system in low-relief terrain is illustrated in Fig. 2. The hydrology and distribution of discharge features is controlled by topography and vertical or horizontal permeable structures, e.g. faults and the conduits produced by hydrothermal eruptions. Vapour-fed discharge features occupy higher ground than the chloridewater springs, and include fumaroles, hot springs discharging steam-heated waters and steaming ground. These can form above boiling zones where carbon dioxide and hydrogen
GEOTHERMAL FLUID CHEMISTRY
227
Table 1. Geochemical-mineralogicalcharacteristicsof hydrological regimes within high-enthalpy geothermal systems Regime
Fluids
Palaeosurface
CI HCO3
Steam zone
SO4, HCO3
Boiling zone (two-phase)
CI + gas
Deep fluid zone
CI
Minerals/deposits Silica sinter carbonate sinter eruption breccia Kaolin, hematite silica residue, cristobalite Quartz, adularia bladed calcite * Silicification, albite adularia t, chlorite epidote, zeblites
Geochemical enrichment As, Sb precipitates Sr As, Sb, Hg, NH3
SiO2, K
* May be pseudomorphed by quartz. 3 Adularia present in greatest amounts in permeable structures. sulphide dissolve into groundwaters or steam condensates. Bicarbonate waters (CO2-rich steam heated waters) are also common on the margins of many fields. Systems in mountainous terrain show a similar profile, but the Cl-fluid rarely attains the surface near the upflow zone and undergoes up to tens of kilometres of lateral flow; additionally, the steam zone will be deeper and more extensive. The dynamic nature of the system is shown by the cycle: meteoric water descent, geothermal fluid formation, (replenished by meteoric waters descending from the recharge zone) and surface discharge of geothermal waters and vapours through springs and fumaroles (Fig. 2). However, fluid flow from depth is unlikely to follow the idealised vertical path shown in Fig. 2, and some degree of lateral flow is probable. Local topographic and permeability influences will exert an important control on the direction of flow and area of discharge. Many systems display lateral flow structures created by strong hydraulic gradients. These in turn are formed due to high relief, often with a near-surface low-permeability horizon. Cooling by conduction and groundwater mixing are reflected in the chemistry of the discharges. Even in low-relief (< c. 250 m) settings, including those typical of silicic volcanic terrain (e.g. Taupo Volcanic Zone, New Zealand), nearsurface lateral flows can extend for several kilometres. This is greatly extended in terrain of high relief ( > c . 1000m), typical of andesitic volcanoes, where flows are 10-50 km in length.
Hydrological regimes and their use in gold exploration In epithermal deposits gold is commonly found in, or just above, boiling zones and zones of fluid
mixing. Both processes lead to cooling of the fluid, loss of HzS and dissociation of the gold-complexing sulphide ligand (Seward 1973, 1991). Recognition of such zones in both geothermal and epithermal systems, together with other parallels, has bound these systems together as modern and ancient equivalents. As the geothermal fluid rises from depth it may cool adiabatically and boil, forming a two-phase zone composed of liquid water, water vapour and gases ('steam'). Since gold commonly deposits in or near the boiling zone, a favourable exploration target requires that this zone and the overlying steam zone are preserved. However, since the ore-depositing zones are relatively shallow features (often < 1 km depth), ancient geothermal systems (especially pre-Tertiary) may have been eroded to depths below the ore zone. Gold exploration programmes therefore need to be tailored towards three objectives: recognition of remnants of geothermal deposits; identifying the level of preservation of the system; location of the ore zone. These objectives can be achieved by reconstruction of the palaeohydrology of the system through hydrological regime recognition. To do this requires an appreciation of the hydrological profile of a geothermal system derived from the physical and chemical process acting on the geothermal fluid as it ascends to the surface. These processes produce hydrological zones in which each fluid type leaves diagnostic mineralogical and geochemical signatures in the surrounding lithologies through mineral-fluid reaction products (the alteration mineral assemblage), elemental enrichments and mineral deposits (Tables 1, 2). Criteria diagnostic of each hydrological regime are summarized in Table 1. Recognition of these signatures enables
228
K. NICHOLSON
Table 2. Characterbticwater-rock interaction products for contrastinggeothermalfluids CI fluid Amorphous silica, quartz, adularia, albite, illite, chlorite, epidote, zeolites, sulphides
SO4 waters
HCO3 waters
Kaolin alunite, smectite, hematite, sulphates
Kaolin, calcite, illite, montmorillorite
Note relative positions of the fluids in the hydrological structure of the system (Fig. 3). Direction of change within upflow or outflow structures Increasing depth in a system or approach to permeable upflow structure. ancient geothermal deposits to be identified and the relative position in the hydrological profile to be established.
Palaeosurface recognition If the palaeosurface can be recognised, this indicates a high level of preservation and greatly assists interpretation. This can often only be confidently done through discovery and recognition of geothermal deposits, notably silica sinters and hydrothermal eruption breccias.
Hydrothermal eruption breccias. These often present conduits for geothermal fluids and are valuable guides to locating epithermal gold exploration targets. The characteristics of these breccias have been detailed by several authors (Nairn & Solia 1989; Collar 1985; Hedenquist & Henley 1985; Nelson & Giles 1985). They are typically matrix-dominated with free-floating angular to sub-rounded clasts of local lithologies. Silica sinters. The presence of a thick sinter deposit not only indicates a palaeosurface but demonstrates that the palaeo-reservoir temperature was in excess of c. 200°C and marks areas of chloride water discharge (the golddepositing fluid, Brown 1986). Textural and geochemical characteristics which may be used to discriminate sinters from other siliceous lithologies have been described elsewhere (Nicholson 1988, 1989, 1993; Nicholson & Aquino 1989; White et al. 1989; Nicholson & Parker 1990; Parker & Nicholson 1990; Browne 1991). Surface sinter textures include striations, mushrooms, banding, ripples, terraces, circular rims, overhangs, plant material and boxworks,
while in cross-section columnar textures are distinct. Thin sections of sinter may show rods and filament textures from biogenic activity. Geochemical enrichments in B, C1, As, Sb and Hg may also be diagnostic.
Permeable (steam) zone recognition Surficial deposits such as sinters and eruption breccias commonly do not survive; furthermore, the boiling zone represents a narrow exploration target. Given the more pervasive nature of steam-rock reactions and the greater aerial extent of the steam zone (Fig. 2) it is this hydrological horizon which presents a better prospect for exploration surveys. Steam from the boiling zone may migrate to the surface independent of the liquid phase and is often able to take a more direct ascent to the surface through horizons which are impermeable to liquid water. Steam reacts with the rocks in three ways: directly, by condensation, and by heating/ dissolving in near-surface meteoric waters. Steam condensates and steam-heated waters form the acidic sulphate and bicarbonate waters found near-surface (SO4-HCO3-type, steam zone, Fig. 2; Tables 1,2). More subtle indicators of permeable zones, the consequence of fluid-rock interaction, may also be detected. Although geochemical surveys are generally less useful than mineralogical mapping in epithermal gold exploration, they are of value in the identification of permeable, upflow structures. Vapours and gases migrating directly to the surface over the boiling upflow zone can accumulate in the steam zone on the altered near-surface lithologies to form geochemical enrichments in As, Sb, NH3, Hg and possibly B (Nicholson 1993). Permeable structures can be identified by areas of the highest concentration of these species, and flow directions within a lithology may also be traced by the use of iso-concentration maps (flow direction is towards lower concentrations as more volatiles are lost from the fluid during early boiling).
Overprinting Fluctuations in the water-table, descent of sulphate fluids and expansion of the steam zone will all lead to an overprinting of alteration mineralogies (typically of sulphate and chloride fluids), complicating interpretation and hydrological reconstruction. However, using the hydrological model and criteria outlined (Tables 1, 2; Fig. 2), the level of preservation can be assessed and permeable upflow structures, which may have contained zones of boiling and
GEOTHERMAL FLUID CHEMISTRY
a
229
b
SP
I11
C o n a ~ etmling
Chloride (mg/kg)
~I~M ~
oG Chloride
"HS
(mg/kg)
Fig. 4. Example of a chloride-enthalpy mixing model. SP, steam point; R, reservoir fluid; HS, hot spring, cooling by boiling; SH, steam-heated waters; M, meteoric waters; C, C1, cooling by condensation, Well discharges which have experienced dilution plot on R-SH or R-M mixing lines.
Given that so much information on the geothermal reservoirs is obtainable from the fluid chemistry, is it possible to apply these geochemical techniques to oil-field waters to assist in the modelling of hydrocarbon reservoirs? The areas where geothermal geochemical techniques may be transferable are (i) identification of water mixing and water source correlation through the application of mixing models (ii) temporal and/or depth thermal profiles in the reservoir through the use of geothermometers (iii) correlation of formation waters.
non-producing formations through casing damage. In geothermal systems, mixing models enable dilution of the primary reservoir fluid, and the diluting fluid, to be recognized. They are used to monitor well chemistry during production to identify intrusion of waters from other formations. The most commonly employed model is a Cl-enthalpy plot (Fig. 4), but others such as silica-enthalpy, carbonate-based and statistical models have also been developed. Figure 4 shows the dilution lines to meteoric waters and steam-heated waters, both of which are lower in CI content and enthalpy than the reservoir fluid, on which diluted well chemistries would plot. Dilution trends in wells across a field can be identified by mixing models as well as temporal dilution changes in a given well. Although oil-field waters may not have a boiling component, the figure could still be used to illustrate mixing-dilution processes between waters of differing CI concentration and temperature.
Mixing models
Geothermometry
Mixing models applied to oil-field brines commonly use plots of species against Br (e.g. C1 v. Br) or oxygen-hydrogen isotope ratios to identify the composition of pre-mixing end-member waters (Egeberg & Aagaard 1989; Connolly et al. 1990). The geothermal mixing models could be similarly applied in hydrocarbon reservoirs to identify breakthrough of waters on the field margins ('edge water') and water intrusion from
The concentration of gaseous, solute and isotopic species involved in temperaturedependent equilibria in geothermal fluids have been calibrated to act as geothermometers (see Nicholson 1993 for a review). These can also be applied to oil-field water well discharges. Possible applications are: identification of water intrusion; reservoir monitoring during production and thermal profile modelling. Monitoring
gold deposition, located. This approach has been used with some success to identify palaeogeothermal systems and gold prospects in Australia and Scotland (White et al. 1989; Nicholson 1988, 1989).
Geothermal geochemical techniques and oil-field waters
230
K. NICHOLSON
Table 3. Equilibration times of some geothermometers expressed as to.5
Exchange
1 ROCK ADDITION K
t05 (at 250°C)
AlSO SO4-H20 AlSO COz--HzO A~80 H2Oo)-H20~v) AD H2-H20(1) AD CH4-H2 AD HzO~)-CH4
AI3CCH4-CO2 A13CCO2-HCO3 A34S SO4-H2S SiO2 geothermometer NaK geothermometer NaKCa geothermometer
c. 1 year 1-100 s 1-100 s 1-20 weeks 1-20 weeks 1-1000 years | 000--10 000 years 1-100 hours > 1000 years 1-100 hours 0.3 years 0.3 years
Si02 ~• IA
•B
Na
BICARBONATE S TEAM-HEATED WATERS
CI CHLORIDE WATERS
• Mg
I
I Ca
;4
• HCO3
SO4
ACID SULPHATE WATERS
the geothermometry of the produced waters may enable breakthrough of non-producing formation waters or edge waters to be recognized by unusual variations in geothermometry temperatures. The temperature indicated by geothermometers is that of the last equilibration, which is not necessarily that of the reservoir. Table 3 lists common isotope and solute geothermometers, together with estimates of equilibration times. In old reservoirs,
Fig. 5. Factor analysis model to illustrate major sources and control on solutes in geothermal well discharges from fields in the Taupo Volcanic Zone, New Zealand. Plot shows Factors 1 and 4 with interpreted sources/controls labelled (from Salvania & Nicholson 1990).
2
• RIVER WATERS
A SP~NG WATERS •
RESERVOIR WATERS
Fig. 6. Principal component analysis of freshwater chemistry. Plot of scores of Principle Components 1 and 2 discriminates between river waters, spring waters and surface reservoir waters (from Tse Tig Cheong & Nicholson 1993).
GEOTHERMAL FLUID CHEMISTRY and reservoirs with little or no dynamic recharge, all geothermometers should have attained equilibrium and each should indicate the reservoir temperature. However, as geothermometers with large t0.5 values ' r e m e m b e r ' temperatures longer than geothermometers which equilibrate rapidly, then in young fields and 'flushed' reservoirs differences in geotherm o m e t e r equilibration times may enable the thermal history of the reservoir to be established. A more general application of geothermometers, however, may be found during production from a field. Here differences in equilibration time can be used to develop a thermal profile either with depth in the area of a well, or across a field using data from several wells.
Correlation o f f o r m a t i o n waters Statistical models to identify common sources of solutes and geothermal waters was developed by Salvania & Nichoison (1990). These were based on multivariate analytical methods applied to the discharge chemistry of wells from several fields. Cluster analysis was able to group waters of common origin, while factor analysis identified the major sources and controls on solutes (Fig. 5). Furthermore, by applying factor analysis to freshwater chemistry, Tse Tig Cheong & Nicholson (1993) were able to effectively discriminate spring waters, river waters and surface reservoir waters (Fig. 6). This statistical approach would also be valuable in monitoring changes in well discharge chemistry, identification of mixing and may be applicable for the correlation of formation waters over large areas such as the North Sea. I am grateful to S. Arnorsson for valuable comments on the original manuscript.
References ARMSTEAD,H.C.H. 1983. Geothermal energy. Second edition, Spon, London. BIscnoFr, J.L., RADTKE, A.S. & ROSENBAUER,R.J. 1981. Hydrothermal alteration of greywacke by brine and seawater: roles of alteration and chlorite complexing on metal solubilities at 200° C and 350° C. Economic Geology, 79, 659--676. BROWN, K. 1986. Gold deposition from geothermal discharges in New Zealand. Economic Geology, 81,979-983. BROWNE, P.R.L. 1978. Hydrothermal alteration in active geothermal fields. Annual Reviews of Earth and Planetary Science, 6,229-250. 1991. Mineralogical guides to interpreting shallow paleohydroiogy of epithermal mineral depositing environments. In: FREESTONE, D.H.,
231
BROWNE,P.R.L. & SCOTT,G.S. (eds) Proceedings 13th New Zealand Geothermal Workshop, 263269. COLLAR, R.J. 1985. Hydrothermal eruptions in the Rotokawa geothermal system, Taupo Volcanic Zone, New Zealand. Geothermal Institute Report 14, University of Auckland, New Zealand. CONNOLLY,C.A., WALTER,L.M., BAADSGAARD,H. LONGSTAFFE,F.J. 1990. Origin and evolution of formation waters, Alberta Basin, Western Canada Sedimentary Basin. I. Chemistry. Applied Geochemistry, 5,375-395. CRAIG, H. 1953. The geochemistry of stable carbon isotopes. Geochimica et Cosmochimica Acta, 3, 53--92. CYAMEX, 1979. Massive deep-sea sulphide deposits discovered in the East Pacific Rise. Nature, 277, 523-528. EDWARDS, L.M., CHILINGAR,G.V., RIEKE, H.H. & FERTL, W.H. 1982. Handbook of geothermal energy. Gulf Publishing Co., Houston. EGEBERG, P.K. & AAGAARD, P. 1989. Origin and evolution of formation waters from oil fields on the Norwegian Shelf. Applied Geochemistry, 4, 131-142. ELLIS, A.J. & MAHON,W.A.J. 1964. Natural hydrothermal systems and experimental hot-water/rock interactions. Geochimica et Cosmochimica Acta, 28, 1323-1357. & -1967. Natural hydrothermal systems and experimental hot-water/rock interactions (Part II). Geochimica et Cosmochimica Acta, 31, 519-538. & - - 1977. Chemistry and geothermal systems. Academic Press, New York. GRANT,M.A., DONALDSON,I.A. & BIXLEY,P.F. 1982. Geothermal reservoir engineering. Academic Press, New York. HAAS, J.L. 1971. The effect of salinity on the maximum thermal gradient of a hydrothermai system at hydrostatic pressure. Economic Geology, 66,940-946. HANNINGTON,M.D., HERZIG,P. & ScoTT, S.D. 1991. Auriferous hydrothermal precipitates on the modern seafloor. In: FOSTER, R.P. (ed.) Gold metallogeny and exploration. BIackie, Glasgow, 250-282. HEDENQUIST, J.W. • HENLEY, R.W. 1985. Hydrothermal eruptions in the Waiotapu geothermal system, New Zealand: Their origin, associated breccias and relation to precious metal mineraliNation. Economic Geology, 80, 1640-1668. & 1985. Natural CO~-rich steam-heated waters in the Broadlands-Ohakki geothermal system, New Zealand: Their chemistry, distribution and corrosive nature. Geothermal Resources Council Transactions, 9,245-250. HENLEY, R.W. 1985. The geothermal framework for epithermal deposits. In: BERGER,B.R. & BETHKE, P.M. (eds) Geology and geochemistry of epithermal systems. Reviews in Economic Geology, 2, Society of Economic Geologists, 1-24. 1991. Epithermal gold deposits in volcanic
232
K. NICHOLSON
terranes. In: FOSTER,R.P. (ed.) Gold metallogeny and exploration. Blackie, Glasgow, 133-164. & ELLIS, A.J. 1983. Geothermal systems ancient and modern: A geochemical review. Earth Science Reviews, 19, 1-50. --, TRUESDELL, A.H. & BARTON, P.B. 1984. Fluid-mineral equilibria in hydrothermal systems. Reviews in Economic Geology l, Society of Economic Geologists. MAHON, W.A.J. 1967. Natural hydrothermal systems and the reaction of hot water with sedimentary rocks. New Zealand Journal of Science, 10, 206-221. --, KLYEN, L.E. & RHODE, M. 1980. Neutral sodium/bicarbonate/sulphate hot waters in geothermal systems. Chinetsu (Journal Japan Geothermal Energy Association), 17, 11-24. MATRAY, J.M. & FONTES, J.CH. 1993. Multi-sourced formation waters from the Dogger aquifer; Paris basin. In: PARNELL,J., RUFFELL, A.H. & MOLES, N.H. (eds) Geofluids '93,304-308. NAIRN, I.A. & SOLIA, W. 1980. Late Quaternary hydrothermal explosion breccias at Kawerau geothermal field New Zealand. Bulletin of Volcanology, 43, 1-13. NELSON, C.E. & GILES, D.L. 1985. Hydrothermal eruption mechanisms and hot spring gold deposits. Economic Geology, 80, 1633-1639. NICHOLSON, K. 1988. Geothermal deposits in ancient terrain as a tool in epithermal gold exploration: examples from Scotland. In: MCKIBBIN, R. (ed) Proceedings lOth New Zealand Geothermal Workshop, Auckland, 151-153. 1989. Early Devonian geothermal systems in northeast Scotland: Exploration targets for epithermal gold. Geology, 17,568-571. -1993. Geothermal fluids: Chemistry and exploration techniques. Springer-Verlag, Berlin, 268pp. & AQUINO, C. 1989. Life in geothermal systems. A key to sinter formation and recognition? In: BROWNE,P.R.L. & NICHOLSON,K. (eds) Proceedings llth NZ Geothermal Workshop, Auckland University, 143-148. & PARKER,R.J. 1990. Geothermal sinter chemis-
-
-
-
-
-
-
try: Towards a diagnostic signature and a sinter geothermometer. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. & SCOTT, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland University, 97-102. PARKER, R.J. & NtCHOLSON, K. 1990. Arsenic in geothermal sinters: Determination and implications for mineral exploration. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. & ScoxT, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland University, 35-39. SALVANIA, N.V. ,~ NICHOLSON, K. 1990. Chemometrics applied to the fluid chemistry of geothermal fields in the Taupo Volcanic Zone, New Zealand. In: HARVEY, C.C., BROWNE, P.R.L., FREESTONE, D.H. &ScoTr, G.L. (eds) Proceedings 12th NZ Geothermal Workshop, Auckland, 157-163. SAUNDERS, J.A. & SWAN, C.T. 1990. Trace metal content of Mississippi oil field brines. Journal of Geochemical Exploration, 37, 171-183. SEWARD,T.M. 1973. Thio complexes of gold and the transport of gold in hydrothermal ore solutions. Geochimica et Cosmochimica Acta, 37,379-399. 1991. The hydrothermal geochemistry of gold. In: FOSTER, R.P. (ed.) Gold metallogeny and exploration. Blackie, Glasgow, 37-62. SUTTON, F.M. & McNABB, A. 1977. Boiling curves at Broadlands geothermal field, New Zealand. New Zealand Journal of Science, 20, 333-337. SVERJENSKY,D.A. 1984. Oil field brines as ore forming solutions. Economic Geology, 79, 23-37. TSE TIG CHEONG,M. • NICHOLSON,K. 1993. Identification of element associations and controls on freshwater chemistry in Grampian Region, Scotland (Abs). llth European Symposium on Environmental Geochemistry and Health, Aberystwyth, April, 1993. WHtTE, D.E. 1957. Thermal waters of volcanic origin. Geological Society of America Bulletin, 68, 16371658. WHITE, N.C., WOOD, D.G. & LEE, M.C. 1989. Epithermal sinters of Paleozoic age in north Queensland, Australia. Geology, 17,718-722.
An integrated approach to the study of primary petroleum migration ULRICH
MANN
Forschungszentrum Jiilich GmbH, Institut fiir Erd61 und Organische Geochemie, D 52425-Jiilich, Germany Abstract: The precise mechanism of primary petroleum migration has been elusive despite intensive investigation and discussion. There is more than one mechanism, and the pathways and efficiencies of individual mechanisms vary from case history to case history due to the variable abundances of micropores, macropores and fractures in source rocks as well as different sources for the build-up of a pressure gradient. Primary migration probably proceeds as diffusion through source rock micropores via mesopores to macropores and fractures. From there, a petroleum bulk phase develops, and moves along the macropore and fracture system, possibly together with aqueous solutions. In fact, many parameters influence petroleum transport out of a source rock, and all of them have seldom been checked by exploration geologists as most mature source rocks are inaccessible. In order to collect information about primary petroleum migration that is as complete as possible, an integrated approach consisting of sedimentology, petrophysics, organic geochemistry and numerical modelling should furnish geologically acceptable results: (i) by sedimentological methods, all potential primary migration pathways are identified, (ii) with the help of petrophysical methods, non-effective migration pathways are separated from effective pathways and excluded, (iii) organic geochemical results provide control and direct evidence for petroleum expulsion stage and efficiency, and (iv) numerical modelling finally quantifies all observed effects within the framework of the geological situation.
Primary petroleum migration is one of the principal processes in the formation of petroleum accumulations. Considerable progress has been achieved in the recognition of the effects and causes of primary migration effects within the last 15-20 years. Nevertheless, the process of primary migration is still under intensive scientific debate. Several contributing mechanisms exist which vary from source rock to source rock and from sedimentary basin to sedimentary basin, and as primary migration occurred in the geological past it is difficult to prove. Besides petroleum geochemistry text books (Hunt 1979; Tissot & Welte 1984), petroleum migration in source rocks has been the subject of several extensive reviews, including Durand (1983, 1988), England et al. (1987), Ungerer (1990), Price (1989), and Welte (1987). The approach for the review presented here follows the applied analytical view of the geochemist rather than that of the exploration geologist. First, a short review of the possible modes of the primary migration process itself should familiarize the reader with the present state of knowledge. Second, several relevant research topics, such as experimental simulation methods, will be addressed as keys to the understanding of the petroleum migration process. Thereafter, most
of this review is dedicated to the methodological approach: how petroleum migration in source rocks can be analysed in order to find out if, when, where and how much hydrocarbons did or did not migrate. Certainly, this approach is not limited to petroleum migration alone and includes many similarities to the analysis of other migrating subsurface fluids. It intends to demonstrate that not only modern organic geochemistry but also other geosciences disciplines are necessary to understand primary migration sensu lato. Such an integrated approach is required especially for a more precise quantitative understanding of petroleum migration in source rocks. A reliable approach has to integrate sedimentology, petrophysics, organic geochemistry, and numerical simulation in such a way that the individual methods fit the relevant case study. In order to provide an overview of the experimental techniques, analyses and individual steps of numerical simulation (according to the four disciplines mentioned), Table 1 summarizes several of the most common methods of investigation. Definition
Movement of petroleum from the source via carrier bed to the reservoir rocks is called
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78,233-260.
233
234
U. MANN
Table 1. Experimental and analytical techniques and numerical verification steps for an integrated approach to petroleum migration in source rocks Sedimentology
Petrophysics
Organic geochemistry
Numerical simulation
Inorganic petrography
Helium pycnometry
Organic petrography
Cathodo-luminescence
Gas adsorption
Fluid inclusion analysis
Absolute permeability versus stress Mercury porosimetry
Total organic carbon, pyrolysis Thermodesorption GC
Source potential, temperature history Overpressure development Flow modelling
SEM, Cryo-SEM Carbon and oxygen isotopes
Liquid and gas chromatography Wettability, contact angle Gas chromatographymeasurement mass spectrometry
migration. It is subdivided into primary migration, which is defined as the movement of oil and gas through and out of the fine-grained source rocks, and secondary migration, the movement through wider pores to the trap. The loss of petroleum from the trap is called dismigration. This paper reviews the migration which takes place directly after petroleum generation, the process of primary migration. In this review, petroleum and hydrocarbons also include NSO-compounds if not explicitly excluded. Bitumen is applied as a generic term to natural inflammable substances of variable colour, hardness, and volatility, composed principally of a mixture of organic compounds (mainly hydrocarbons). It is applied in order to indicate a residual part of petroleum which has been lost during migration (residual oil, shows). Soluble organic matter may represent that part of the bitumen which can be dissolved by an organic solvent, but in general, it represents the organic solvent extract of any rock sample.
Modes of primary migration Mechanisms Several mechanisms seem to work alone or together in different types of source rocks and sedimentary basins, depending on the source rock properties like richness, sedimentological and petrophysical attributes, and the prevailing pressure-temperature conditions as well as the potential of the geological conditions for overpressure formation and microfracturing. According to most authors, the most important form of primary migration during the main phase of oil and gas formation seems to be discrete hydrocarbon phase movements (Dickey 1975; Hunt 1979; Momper 1978; Tissot & Welte 1984; Welte 1987; Durand 1988; Ungerer 1990).
Fracture modelling Migration pathway modelling, sensitivity analysis
However, as true hydrocarbon flow cannot take place in micro- and mesopores of a source rock, yet petroleum has to reach the macropores, diffusion may be the dominating process at the beginning of the primary migration of a hydrocarbon molecule. A sketch of the possible mechanisms from generation to expulsion is presented in Fig. 1. Under standard geological conditions, diffusion alone seems several orders of magnitude too small in order to fill a petroleum reservoir (Thomas & Clouse 1990). For the same reason, the postulation by Stainforth & Reinders (1990) that active diffusion of bitumen molecules through organic matter networks represents the rate-limiting petroleum migration process may only hold for short migration distances and/or relatively poor source rocks. According to Hunt's (1979) calculations, based on Fick's law, migration paths for diffusion should be short, between 'tens to hundred of feet', in order to account for a commercial oil accumulation by diffusion in a reasonable time scale such as 10 million years. However, fluid movement can augment a diffusion process, therefore two or three migration mechanisms acting in parallel would increase the distance of migration. Diffusion is thus the most likely initial mechanism, directly after petroleum generation up to the time a pressure-driven flow takes over. Hunt (1973) and Dickey (1975) both have suggested that water may limit the free pore space sufficiently to permit oil-phase migration. Since an oil droplet would be restricted to the bulk water phase, it is conceivable that at some depth the oil in the smaller pores would occupy enough of the bulk space to be expelled by capillary forces assisted by the fluid potential gradient into a coarser-grained (= lower capillary pressure) zone of a source rock. Water is not needed to explain oil-phase migration in
PRIMARY PETROLEUM MIGRATION
235
Pathway
Pore pressure solution feature
Kerogen
Diffusion
Deaorption
Aggregation
Bulk flow
Mechanism
Fig. 1. Primary migration mechanisms for petroleum in source rocks (after Mann et al. 1991). very organic-rich rocks, as such rocks may be partially oil-wet (Hunt 1979). Completely oilwet pores in source rocks will hardly be encountered, as the estimated minimum organic content for this situation is about 30% (Byramjee 1967). Therefore, most petroleum source rocks should provide a mixed wettability. The ratio of petroleum to water during primary migration was estimated for the Western Canada Basin by Hunt (1977). He calculated that the expelled water masses should have contained about 300-1000ppm oil which requires oil-phase migration, but a few parts per million hydrocarbon in water can be explained by a solution mechanism. Data by McAuliffe (1966, 1979) and Price (1976) for molecular solution limit efficient primary migration by solution in water to the most soluble light hydrocarbon fraction such as methane, ethane and benzene. Nevertheless, the fact that aromatics show a much better solubility than naphthenes, and naphthenes a better one than paraffins, can be used to prove whether solution in water may have contributed to specific primary migration avenues. Price (1976) could show a pronounced temperature effect for the solubility of oil in water: it increases from 25 to 100° C gradually by a factor of about two, but from 100 to 180°C by a factor of about five. Hydrocarbon transport in gaseous solution (proposed by Sokolov 1948) was re-evaluated by Price (1989) and Leythaeuser & Poelchau (1991). They advocated a mechanism that overcomes the often encountered low bitumen saturation in oxygen-rich (kerogen type III) organic matter. The hydrocarbon gathering problem is solved by a diffusive partitioning of hydrocarbons by aqueous solution through shale pore water to dispersed hydrocarbon gas bubbles.
From a theoretical point of view, the transportation of individual oil droplets seems possible, as colloidal and micellar solutions have been proposed in the past by several authors (Baker 1959; Meinschein 1959; Cordell 1972). However, 'physicochemical, geochemical and geological considerations make individual droplets and bubbles or colloidal and micellar solutions highly unlikely as an effective means of transport during primary migration' (Tissot & Welte 1984, p. 323). A consideration of the mobilization of oil in primary migration must take all the major components in the pore fluids into account. The precursors and by-products of oil generation from sedimentary organic matter may contribute significantly to the pore fluids. The experimental data by Bray & Foster (1980) support a migration process where carbon dioxide and hydrocarbon gases mobilize the liquid hydrocarbons so that they can leave the source rock with any water expelled during compaction. This process should be operable in source rocks that have pore water as the principal pore fluid. Carbon dioxide from great depth may also support primary migration of hydrocarbons (Kvenvolden & Claypool 1980). Micro fracturing
In view of the fact that source rock lithologies are generally regarded as having a low permeability, geologists have considered other potential petroleum pathways. Several authors suggested that oil generation may induce microfractures in source rocks (Snarsky 1962; Tissot & Pelet 1971; Momper 1978; Palciauskas & Domenico 1980; du Rouchet 1981; Ozkaya 1988; Lehner 1991; Diippenbecker & Welte 1992). The term 'hydraulic fracturing' has been borrowed from the petroleum industry, where it
236
U. MANN
is known as a technique of well stimulation by means of pressure applied in bore holes, and it is used to describe the lithostatic failure by high pore pressure in overpressured reservoir regimes. For differentiation, a more specific and genetic term like 'expulsion fracturing' seems preferable. According to Palciauskas & Domenico (1980), a relationship may exist between the microfracture development and hydrocarbon-maturation kinetics. They conclude that, once initiated, the average rate of microfracture propagation would keep pace with the rate of sediment burial. However, microfractures can seldom be confirmed as a migration pathway by direct observations because in most clay-shale source rocks they will probably reheal completely. This reason, the necessary excess pressure for fracturing in relation to the petrophysical properties as calculated by Sandvik & Mercer (1990), and the higher ductility of source rocks (in contrast to reservoirs), might be the main arguments why many exploration geologists are still sceptical about the concept of microfractures. On the other hand, in a thick, organic-rich, rapidly heated source rock which generates petroleum at high rates, the pore pressure may be in excess of the mechanical strength of the rock and open fractures. Long geological times may greatly compensate for low permeabilities and low flow rates. Therefore, primary and secondary porosity of a source rock may be sufficient to serve as migration pathway.
Potential driving forces for primary migration To obtain petroleum flow within macropores or fractures of a source rock, a pressure gradient is needed. Several kinds of geological processes may contribute to varying degrees. The weight of the overburden can partly be transferred from the solid grains to the pore fluids. During generation, the transformation of part of the kerogen to hydrocarbons results in the partial transmission of the geostatic load from the solid phase to the liquid phase. This situation arises when the permeability of sediments is so low that the pore fluids cannot be squeezed out fast enough. Compaction seems to represent the most common case in order to provide the driving force for expulsion (Meissner 1978; Hunt 1979; du Rouchet 1981). However, several other processes may contribute to pressure generation or influence the direction of the pressure gradient. In deeper petroleum systems, flow directions of primary
migration may be controlled by the configuration and internal pressures of individual, seal-bounded fluid compartments (Vandenbrouke et al. 1983; Demaison 1984). The deeper petroleum systems, where the oil is generated, do not exhibit one basin-wide compartment with hydrostatic pressure such as in most shallow systems, but consist of a series of individual fluid compartments that are not in hydraulic pressure communication with each other nor with the overlying hydrodynamic regime (Hunt 1990). Oil and gas migrates only downwards from a source rock to an underlying permeable rock which releases pressure and allows the petroleum potential (for migration) to continually decrease. The petroleum potential is given as the sum of the overpressure, the buoyancy potential and the capillary potential (England et al. 1987). Therefore, the overpressure must be sufficiently high in order that downward flow becomes possible. The gradient required to overcome buoyancy is such that it is unlikely that petroleum is able to migrate more than 300--500 m downwards from a source rock (Mann & Mackenzie 1990). The volume increase of organic matter due to the solid-liquid conversion during hydrocarbon generation provides an internal pressure source in the source rock (Snarsky 1962; Sokolov et al. 1964; Hedberg 1974; Momper 1978). Ungerer et al. (1983) and Goff (1983) estimated that this may be as high as between 10 and 20% of the initial volume. Although clay mineral dehydration contributes a remarkable water volume, most water leaves a siliciclastic source rock prior to petroleum expulsion (Burst 1969; Perry & Hower 1972). Nevertheless, it is tempting to propose a genetic connection between smectite dehydration (illitization) and hydrocarbon migration, not only because enormous volumes of water are involved, but especially because the temperature necessary for intense oil generation in source rocks appears, in many areas, to roughly coincide with the zone of abrupt illitization. This relationship may be fortuitious, but on the other hand, a genetic relationship may exist for selected basins such as the Gulf of Mexico (Bruce 1984). However, the existence of clay-poor carbonate source rocks indicates that water from clay mineral dehydration is certainly not a requirement for primary migration (e.g. Grabowsky 1984). If oil droplets block the pore openings in a source rock, then the water trapped behind the oil will expand as the temperature rises. The increased pressure caused by the thermal expansion of water has been termed aquathermal
PRIMARY PETROLEUM MIGRATION
pressuring by Barker (1972). According to Barker's diagram (1972, fig. 1) the pressure increase because of complete isolation is significant. Although results from subsurface and compaction data from the Gulf Coast (Magara 1975) seem to demonstrate the possible presence of aquathermal pressuring effects, Daines (1982) concludes that this mechanism could be responsible for abnormal pore pressures only in shallow, impermeable sediments, particularly in areas of high geothermal gradients. Local seals of carbonate cement may form directly above and below solution seams, and may act as a pressure reinforcement for expulsion (di Primio 1990). This would point to an accelerated expulsion process from solution seams and supports Jones's (1984) thesis for an 'explosive migration' in many carbonate source rocks. Other local sealing effects formed by cementation during diagenesis may help to bring about an internal pressure build-up within a source rock and support petroleum expulsion.
Keys to the process of primary migration The simulation of migration in the laboratory represents an alternative approach for elucidating the effects of petroleum migration from geological case studies. The advantages in the laboratory are a better control over maturation, generation, degradation, organofacies and mixing effects which generally complicate the interpretation of migration effects in natural settings. On the other hand, flow or diffusion experiments in the laboratory reflect a unique optimised system in which a small quantity of oil with a well-known composition is passed through a well-defined mineral phase path at a relatively constant rate and for a very short time. It must be expected that very subtle changes in the parameter configuration of a system may change the results tremendously. All laboratory experiments are unreal, and extrapolations to the geological situation have to be accepted for what they are. As a tool to simulate generation and expulsion of petroleum at the same time, hydrous pyrolysis has become a widely used laboratory technique (introduced by Lewan et al. 1979). The technique maintains a liquid-water phase in contact with potential petroleum source rocks while they are heated at subcritical water temperatures. Although the generated and expelled oil which accumulates on the surface of the water can be quantitatively collected, the relationship of the amount and the composition of the expelled oil phase to the petrophysical properties of the rock
237
(during the actual expulsion process) and the petroleum flow itself cannot be investigated because mineralogy and petrophysical properties may change during heating. However, in contrast to many small-scale pyrolysis methods which, in general, are limited to the evaluation of the generation characteristics, hydrous pyrolysis as applied by Lewan (1985, 1987, 1992) can also consider rock properties relevant for expulsion, as for example mineral matrix or textural effects. If generation and migration are simulated together, the conclusions of numerous studies indicate that a minimum of effluents or water pressure is necessary to allow natural conditions to be reproduced and to promote hydrogenation reactions via hydrogen transfer. Observing the behaviour of a selected hydrocarbon compound or oil during permeation of a selected mineral or rock matrix with (= bulk flow) or without a pressure gradient (= molecular flow) represents one principal approach for understanding petroleum migration in source rocks. In principle, the effects of the individual expulsion parameters like pressure-temperature conditions, maturity stage of organic matter, water content, kerogen and migrating compound type, mineral matrix, petrophysical properties and others, should be investigated separately. In spite of the idealization of the system and the chance to vary only one parameter, most experiments are complicated by the fact that several mechanisms work hand-in-hand. In fact, petroleum migration in many source rocks may represent the net effect of bulk flow, diffusion (in solution or gas phase diffusion), solution (water and hydrocarbons), sorption and swelling effects (organic/organic and organic/inorganic) and possibly other processes. Therefore, although most of the quality of a migration experiment is controlled by the ability to monitor one single migration mechanisms, this makes the system increasingly unreal compared to the geological situation. In order to elucidate primary migration mechanisms, the analysis of the petroleum composition within pores of different size may represent another promising approach. Based on a relatively general approach to petroleum migration (not only source rocks) by Beletskaya and co-workers (Beletskaya 1972; and later references cited by Sajgo et al. 1983; Brukner & Vet6 1983) and by Sajgo et al. (1983), Ropertz (1993), re-evaluated this concept in more specific terms. His SPEX-method (Selective Pore EXtraction) provides step by step the individual extracts from large, small and closed pores of the same rock sample. Compositional differences were interpreted in accordance with
238
U. MANN
standard maturity parameters for oils known from organic geochemistry such as the MPI (Radke & Welte 1983) or from the relative composition of biomarkers. It was found that the composition differed especially during expulsion at the early maturity stage of source rocks, and that the maturity (according to the composition of the soluble organic matter content) increased with decreasing pore size. Ropertz (1993) concluded that at least during the early expulsion stage, petroleum migration predominantly takes place through a macropore system. This confirms earlier considerations by Momper (1978) and Honda & Magara (1982) who claimed that petroleum migration in source rocks takes place through the largest available pore throats, through so-called 'atypical' pores. Especially during early maturation of source rocks, this may produce mismatches in oilsource rock correlations (Price & Clayton 1992). With progressive generation and increasing oil saturation of the pore space, petroleum seems to move through the whole pore system, whereas the petroleum of the closed pores is hardly influenced by progressive generation (Ropertz 1993). This would mean that if no fractures or pressure solution features are present in a source rock, primary migration takes place as a channelling process via macropores. This would provide distinctive advantages for petroleum flow. The existence of such migration channels would enhance the potential for a more oil-wet expulsion pathway and better flow conditions. In addition, if selected petroleum precursor compounds, like organic acids, possess the capability to influence oil wettability in a positive way, then such channels which have experienced more intensive flow of water with the precursor compounds in solution should be most suited for petroleum expulsion.
Well logs From the explorationist's point of view, it is important to decipher first the correct direction of petroleum migration in order to interpret the drainage conditions of a potential trap correctly. The principal direction of primary migration can be either vertical or horizontal: according to Magara (1980), in relatively young sedimentary basins strong vertical potential differences are common in the intermediate depth range (1000-2500 m). In the older sedimentary rocks, such differences may have been dissipated, but shale-porosity profiles sometimes indicate conditions of undercompaction. England et al. (1987) have argued that most fluids (oil, gas, water) in rocks with permeabilities greater than
1 mD move laterally whereas most fluid movement in rocks with permeabilities less than 1 mD is vertical. However, porosity and permeability within the same source rock may vary considerably due to lithofacies changes or due to specific pathways like fractures or solution seams. This leads to an internal hydrocarbon redistribution within individual units of a source rock and/or different expulsion times for individual source beds. Diagenetically formed pathways may determine the direction of petroleum alone or in combination with the pore network, which is known from reservoir engineering as a dual porosity model. Based on comparisons of shale porosity curves from sonic logs with hydrocarbon extract data, Magara (1980) was able to provide evidence for intervals with depleted hydrocarbons which coincided with undercompacted zones or were in close association with reservoir sections. A practical model, the 'A log R technique', was developed by Passey etal. (1990) for organic richness from porosity and resistivity logs, but this method may also have capability to reveal individual hydrocarbon-depleted (= former migration avenues) or -enriched (= active migration avenues) intervals within a source rock formation. The method of Passey et al. (1990) uses an easily implemented curve overlay that is calibrated for organic richness and maturity. Direct evidence for hydrocarbon-depleted source rock intervals may be recognized from lowered resistivity values, as shown by Mann (1991) from a secondary porosity zone from the Lower Toarcian. In principle, well logs provide a fast and therefore economic quick-look method, because logs are available soon after completion or drilling. However, in many case studies, as for example in narrow interbedded shale-sand or carbonate--evaporite sequences, the use of log data for revealing migration trends will probably be insufficient due to limited depth resolution and accuracy. Especially for testing the effects of diagenetic migration pathways, a detailed approach using sedimentological and petrophysical techniques and methods will often be unavoidable. A combination of a porositypermeability and a geochemical parameter should be applied in order to reveal the correct direction for migration.
Sedimentological approach Lithofacies, as well as texture and bedding of the sedimentary rock, are the primary controlling parameters determined by and inherited from the specific depositional environment. They
PRIMARY PETROLEUM MIGRATION control the potential fluid migration pathways through the primary interparticle pore space of the rock. Compaction, chemical diagenesis, and fracturing are secondary controlling parameters during basin evolution which control the formation of specific migration pathways. Due to much more variation of the lithofacies in carbonate, evaporite and siliceous source rocks as well as due to a more complicated compaction and diagenetic history in contrast to siliciclastic source rocks, carbonate, evaporite and siliceous source rocks possess much more potential to develop migration pathways. According to the two genetically different types of migration avenues for petroleum migration in source rocks, first individual lithofacies aspects of the different source rock lithologies will be discussed, followed by various possibilities for the formation of specific migration pathways.
Lithology and lithofacies variation in source rocks During macro- and microscopic petrography, first of all the lithology-specific sedimentological properties have to be recorded. Shales, marls and argillaceous carbonates generally convey an impression of being homogeneous, dense and impermeable, except for an occasional fracture or other discontinuities. Scanning electron micrographs, however, often reveal a remarkably open fabric beside the unresolved microand mesopores, consisting of a network of regular and irregular macropores and vugs. Both detrital and secondary minerals may be present in a corroded or partly dissolved stage. Detailed optical investigations by normal light microscopy (e.g. Littke et al. 1988), incident-light fluorescence microscopy (e.g. Soeder 1990), and by scanning electron microscopy (e.g. Lindgreen 1987b) are a very effective way to gain insight by direct observation into the most permeable part of a source rock, and to reveal which features and minerals contribute to source rock porosity and permeability. Generally, nearly all detrital non-clay minerals (quartz, calcite, pyrite, dolomite siderite, halite, feldspars, barite etc.) provide permeability because their silt- and sand-size grains form larger pores. Secondary crystals of quartz or calcite commonly provide their own permeable haloes. Clusters or laminations of microfossils and silt lenses in source rocks are generally areas of relatively large pores. In carbonate and evaporite source rocks, pathways created by diagenetic redistribution of carbonate (see below) are much more common
239
and seem to provide the dominant expulsion avenues (Grabowski 1984; Pollastro & Scholle 1986; Comer & Hinch 1987; Denham & Tieh 1990; Lee 1991 ; Hofmann 1992). Good to excellent biogenic siliceous source rocks are known from the Monterey Formation from the San Joaquin basin in California (e.g. Isaacs et al. 1983). Diagenetically altered and unaltered diatoms make up silica, porcellanites (opal-A and opal-CT) and quartz cherts (e.g. Graham & Williams 1985). Such sedimentary constituents provide generally a high porosity, but at the same time a relatively low permeability. This explains why many Monterey source rocks are distinctive in yielding high S1 values during Rockeval pyrolysis (Graham & Williams 1985), an indicator of fluid hydrocarbons and/or solid bitumen, and why petroleum migration through the macropore network has not been reported from the Monterey Formation so far. Primary migration in the Monterey shales has recently been evaluated by Lee (1991). He observed oil-stained solution seams and fractures and concluded that these are the pathways for primary migration. This is very similar to the reservoirs in the Monterey Formation, where production depends on the content of detrital material (Issacs 1980) and the resulting degree of fracturability (Hornafius 1991).
Migration path ways by diagenetic dissolution In order to check potential diagenetic pathways for primary migration, the diagenetic evolution of each individual lithological and lithofacies source rock unit has to be retraced. By luminescence investigations, (trace) element analyses and by analyses of stable carbon or oxygen isotopes, it is possible to support microscopic observations about the genetic history. The composition and mineralogy of carbonate cement is controlled by the composition of pore water and the availability of particular cations. For instance, carbonate produced from the degradation of organic matter can have a distinctive mineralogy (Curtis 1978). Carbonate produced from bacterial sulphate reduction is usually iron-poor, and carbonate produced by bacterial fermentation or thermal decarboxylation can be iron-bearing or iron-rich (Irwin et al. 1977; Coleman et al. 1979). Furthermore, the iron content of carbonate cements in shaly sequences often increases with depth (Boles & Franks 1979; Irwin 1980; Matsumoto & Iijima 1981; Irwin & Hurst 1983).
240
U. MANN
Table 2. References for pressure solution features as primary migration pathways for petroleum
Formation, rock unit
Age
Locality
Lithology
Reference
Niobrara
Late Cretaceous
Denver Basin, USA Chalk, calc. shale
Austin Chalk Sunniland Limestone Niagaran
Late Cretaceous Early Cretaceous
Chalk Limestone Limestone
Budai et al. (1984)
Woodford
Shale, chert
La Luna
Late Devonian, Early Mississippian Late Cretaceous
Texas, USA SouthFlorida Basin, USA Michigan Basin, USA Oklahoma, USA
Pollastro & Scholle (1986) Grabowski (1984) Palacas (1984)
Limestone
Smackover
Late Jurassic
Maracaibo Basin, Venezuela Gulf Coast, USA
unknown formation Monterey
Triassic Mid- to Late Miocene Early Oligocene
Comer & Hinch (1987) Talukdar et al. (1987) Sassen et al. (1987), Denham & Tieh (1990) di Primio (1990) Lee (1991)
S-Unit
Silurian
Northern-Italy California, USA Mulhouse Basin, France
The carbon isotopic composition can be used to identify the origin of carbonate cements, and the oxygen isotopic composition for deduction of the precipitation temperature of the cement and the isotopic composition of the pore water at the time of precipitation (Irwin & Hurst 1983). Composite views of mudrock evolution during burial diagenesis have been proposed by Hower et al. (1976), Foscolos & Powell (1980) and Curtis (1983). According to Curtis (1983), calcite, K-feldspar, kaolinite and possibly other fine-fraction clays are potential candidates for dissolution during the 65-95°C catagenetic phase, and may thus provide secondary porosity for primary migration. Other possible causes for dissolution of carbonate may be the interaction with organic acids (Barth et al. 1988) or cooling formation waters (Bj0rlykke 1984; Giles & de Boer 1989). Due to the higher chemical reactivity of carbonate and evaporite minerals in contrast to clay minerals, several cementation, dissolution and re-precipitation generations are common in carbonate and evaporite source rocks and thus provide a great potential for primary petroleum migration pathways created by secondary porosity. A further cause for carbonate dissolution may be the intensified flow of formation waters in front of permeability barriers (Mann et al. 1990), because flow is 'concentrated through the more permeable horizons, out of the more porous mudrocks' (Curtis 1983). The nature of dewatering during
Lime mudstone Carbonates Siliciceous and calcaceous shale Evaporites
Hofmann (1992)
compaction undoubtedly occurs upwards, although not uniformly, but relative to any stratigraphic marker (Bonham 1980). Lindgreen (1985) investigated carbonate bands as primary migration pathways from the Kimmeridgian in the Central Graben of the North Sea. He found that ' . . . once the mobile phase has reached the large pores in the claystone, flow transport out of the source rock through these pores and the large ones in the carbonate bands should be the dominant process'. By a combination of X-ray diffraction, thermal analysis, M6ssbauer spectroscopy, thinsection studies, scanning microscopy, electron microprobe, and gas adsorption studies, Lindgreen (1985) was able to show the better permeability of the carbonate bands in comparison to the adj acent clay-shale. Similar carbonate bands were described by Prozorovich et al. (1973) from Volgian source rocks in western Siberia, and by Irwin (1980) from the Kimmeridge Clay of Dorset in England. Fibrous calcite veins (so-called 'beef') lying parallel to the bedding of certain sedimentary rocks may also indicate potential as primary migration pathway (Stoneley 1983). Another type of diagenetically-formed migration pathway from the Lower Toarcian in Germany was investigated by Mann (1990, 1991). Probably due to the uplift of the source rock during basin inversion, cementation of the original pathways did not take place, and it was
PRIMARY PETROLEUM MIGRATION
241
0 0
0
C a C 0 3 (%) 0 0
0
CaCo3(%) Pre-compaction
Post-compaction
Fig. 2. Stylolite formation due to re-distribution of carbonate: pre- and post-compaction stages (after Scholle & Halley 1985). possible to study relatively open migration pathways. Based on mineralogical, electronoptical and petrophysical investigations, it was possible to differentiate four diagenetically modified lithofacies types: a calcareous clayshale, a marlstone, a cemented marlstone, and a marlstone with secondary porosity. Although all facies had the same hydrocarbon generation potential, they differed in the extent of hydrocarbon expulsion which was controlled by their petrophysical properties. Pressure solution seams deserve special attention as primary migration pathways. They are known to occur in carbonate, evaporite and siliceous source rocks from the Palaeozoic to the Tertiary (Table 2). Pressure solution seams as primary migration pathways were first described by Ramsden (1952). He suspected stylolites as the principal petroleum migration avenues in carbonate source rocks from the Middle East. Stylolites can be identified on the basis of stylolitic geometry, the truncation of fossils or a general offset of marker horizons. They often develop at contacts of contrasting grain size or along pre-existing fractures (Kulander et al. 1990). There are horizontal stylolites (mainly due to overburden stress) and vertical ones (due to compressional tectonics). Stylolites with a low relief are generally called solution seams. Oilstained solution seams provide the most convincing evidence for active migration paths. But
very hydrocarbon-poor solution seams can also be diagnosed as depleted palaeopathways if hydrocarbon inclusions can be identified (see below). Before petroleum expulsion takes place, solution seams have acted as the place for concentrating organic material (together with clay minerals) which improves the source rock potential drastically (Sassen et al. 1987). Figure 2 presents such a redistribution from the pre- to the post-compaction stage. Because of these redistribution processes, it is often difficult to answer the question about the actual petroleum expulsion stage without a quantification of the compaction stage, because the initial hydrocarbon generation potential remains speculative. For mass balance of the organic material as well as for petroleum quantification during numerical simulation, the extent of the diagenetic concentration process has to be separated from primary litho- and organofacies effects. Suitable methods for determination of the volume reduction by compaction are the measurement of the maximum stylolite amplitude (Stockdale 1926; Mossop 1972), reconstruction of fossils (Ricken 1986), or comparison of the amounts of pyrite in stylolite and rock matrix (di Primio 1990). Fractures
Fractures have been reported as migration
242
U. MANN
conduits for petroleum expulsion from source rocks in case studies from the Bakken Shale (Meissner 1978), from the Lower Toarcian (Littke & Rullk6tter 1987; Littke et al. 1988; Leythaeuser et al. 1988; Mann 1990), from the La Luna formation (e.g. Talukdar et al. 1987) and from several other locations. For a quantitative approach, the number of fractures, fracture width, distribution and orientation must be recorded during the sediment-petrographical characterisation. Jochum (1988) and Jochum et al. (1991) determined the accumulated fracture width of horizontal to sub-horizontal fractures (parallel to sub-parallel to bedding) of the Lower Toarcian shale in Germany as between 1 and 3% of the total source rock thickness. Thereafter, the origin of the fractures must be assessed in order to establish a possible link to the timing of petroleum expulsion. In principle, two genetic types of fractures are differentiated: expulsion fractures, created as a result of petroleum generation, and tectonic fractures, related to the stress of a certain tectonic event. Expulsion fractures probably completely re-heal. Therefore, they will hardly be recognized at later geological times if they are not kept open due to pressure release, e.g. during later uplift of the source rock, or preserved by secondary crystallization as in the carbonate bands reported by Lindgreen (1985). Tectonic fractures as primary petroleum migration pathways may be particularly abundant in shear zones (e.g. Azevedo 1990). They may open and close several times due to activation by processes like seismic pumping (Sibson et al. 1975; Sibson 1981). Completely mineral-filled fractures should be checked for fluid inclusions (see below) and for the timing of the open fracture stage before cementation relative to petroleum migration. Cemented intersections of fractures allow the relative dating of individual fracture generations as well as the delineation of the respective stress directions (Kulander et al. 1990). In this way, the open stage of a certain fracture generation can be related to the time of a tectonic event and tied to a possible coincidence with petroleum expulsion. Nevertheless, fracture formation may be the result of a combination of effects like halokinetic doming, overpressure formation, and with possible contributions from source rock maturation, as for example in chalks from the Norwegian North Sea (Watts 1983). Lindgreen (1987b) investigated fractures from the Upper Jurassic claystone and from the Cambrian Alum shales by scanning electron microscopy and by detailed gas adsorption experiments. He found fractures parallel to, inclined to and normal to the bedding
which he related to overpressure formation and tectonics at different compaction stages. Capuano (1993) provides direct evidence of fluid flow in microfractures in geopressured shales of the Oligocene Frio Formation from Texas. Although not in a source rock, he showed a paragenetic relationship between a calcium sulphate fracture fill and later deposited organic material, and the more extensive alteration of the fracture margins which indicates that these microfractures developed in situ and supported fluid flow.
Petrophysical approach After all potential expulsion pathways have been identified by sedimentological methods, petroleum migration along these pathways has to be quantified in order to separate effective from non-effective migration pathways. This is generally based on permeability data for a source rock. The absolute gas permeability of sediments and sedimentary rocks ranges over more than 13 orders of magnitude from 10-1°mz to about 10-23m2 (Brace 1980). Fine-grained source rocks will certainly show no better permeability than 10-1Sm~ (compare Table 3). For permeability determinations of relatively tight rocks like source rocks, pulsed permeability tests are recommended due to the much shorter measurement times (Brace 1980). The values vary according to the elasticity of the rocks in relation to the overburden pressure (0-500 bars) generally between 2-4 orders of magnitude (Table 3). This sensitivity of permeability to effective stress depends upon stress history. Rocks during compaction (inelastic) are much more sensitive than rocks which are decompressed during uplift (about 10:1). This was also reported by Sandvik & Mercer (1990) when commenting on the observations of Shi & Wang (1986) and Morrow et al. (1984). Pore size and fracture width represent the main controlling agents for petroleum flow as permeability is proportional to several powers of pore size or fracture width, and it is known that very few large pore canals need be present to have a dramatic effect on permeability (e.g. Khanin 1968). Thus, it may become essential to determine the relative amounts of pore size classes present in a specific source bed, especially to differentiate between the low and higher permeability domains of a source rock, as different types of migration mechanisms may be involved at different pore size levels. Such a quantification is possible by a combination of
243
PRIMARY PETROLEUM MIGRATION Table 3. Permeability (p.D, helium) of selected source rock formations from various basins versus confining pressure for different lithologies (after Ropertz 1993) Fm
Depth (m)
Lower Saxony Basin LT C4o, with usual carbon number maxima in the mid-C20 region and no carbon number predominance (CPI* = 1.0) (Simoneit & Lonsdale 1982; Kawka & Simoneit 1987). A significant amount of an unresolved complex mixture (UCM) of branched and cyclic naphthenic hydrocarbons is also present. The generation of the complete suite of saturated biomarkers from their biological precursors (e.g., cholestanes from cholesterol) is supportive evidence for the strongly reductive process operating during initial organic matter alteration. The major resolved peaks in the aromatic/naphthenic fraction are unsubstituted polynuclear aromatic hydrocarbons (PAH, Fig. 3b). P A H become the dominant species due to their high thermal stability as well as their enhanced solubility in near- and supercritical water (e.g., Sanders 1986). The aromatic/naphthenic fractions of the Guaymas oils also contain significant amounts of N, S, and O hetero-PAH (e.g., Gieskes et al. 1988), and Diels' hydrocarbon (Simoneit et al. 1992a). The chemical compositions of the aromatic fractions indicate an origin from oxidative alteration of precursor compounds at high temperatures in the system (>300 ° C). The volatile and most of the soluble compounds (mainly C1-C10 hydrocarbons) are not effectively retained with the heavy petroleum as it solidifies at the vents on the seafloor of Guaymas Basin. Upon exiting at the seabed the fluids are often saturated with a broad range of volatile hydrocarbons (CH4 to CmH22) as well as lower concentrations of heavy ends (>C15) (Simoneit et al. 1988). Thus, the headspace gases of a Guaymas Basin mound sample (1629-A3, Fig. 4a) can be compared with the hydrocarbon content of a 308 ° C vent water, which is highly enriched in the lower alkanes (C40) gas (CH4-C10) oil (C8-C40+) asphalt (>C40)
Relative importance state: dominant/near-critical minor/supercritical minor/supercriticai
Relative concentration: high intermediate minor trace high intermediate Effect: enhanced oil solubility complete oil solubility complete oil solubility
Hydrothermal petroleum migration (a) effective in systems with high water to rock ratios (unlike conventional sedimentary basins) (b) migration bulk phase (trapped in filled veins and voids) emulsion (trapped in fluid inclusions) solution (precipitated as wax crystals in chimneys and fluid conduits)
HYDROTHERMAL PETROLEUM
271
Summary
References
In general, hydrothermal oil generation processes differ significantly from the conventionally accepted scenario for petroleum formation in sedimentary basins, where organic matter input, subsidence, geothermal maturation, oil generation, and oil migration are discrete and successive steps that occur over a long period of geological time. Conversely, in hydrothermal petroleum formation several of the steps of oil generation occur simultaneously and have been shown to complete the organic matter maturation to oil-generating processes in short periods of geological time (Table 2). At seafloor spreading axes, hydrothermal systems active under a sedimentary cover (e.g., Guaymas Basin, Escanaba Trough, etc.) generate petroleum from the generally immature organic matter in the sediments. The alteration reactions are primarily reductive and to a lesser extent oxidative, with trace formation of synthetic products (Table 2). The process occurs in systems with high water to rock ratios, where the interplay of high temperature water, carbon dioxide and methane under pressure represents the effective and efficient thermal driving force for reactions and solvent for product extraction and migration (Table 2). Migration of hydrothermal petroleum has been observed to occur as bulk phase, emulsion and solution (Table 2). It is not known to what extent the hydrothermal petroleum generation process has contributed to the formation of crude oil accumulations, but the process appears to be highly efficient for organic matter maturation, oil generation and its migration. Hydrothermal oil generation should be evaluated thoroughly to document its occurrence and implications, and to yield a better understanding of the geological and geochemical constraints for the process.
BAZYLINKSI, D.A., FARRINGTON,J.W. & JANNASCH, H.W. 1988. Hydrocarbons in surface sediments from a Guaymas Basin hydrothermal vent site. Organic Geochemistry, 12, 547-558. ~, WIRSEN, C.O. & JANNASCH,H.W. 1989. Microbial utilization of naturally occurring hydrocarbons at the Guaymas Basin vent site. Applied Environmental Microbiology, 55, 2832-2836. BISCHOFF,J.L. & ROSENBAUER,R.J. 1984. The critical point and two-phase boundary of sea water, 200--500° C. Earth and Planetary Science Letters, 68, 172-180. & 1988. Liquid-vapor relations in the critical region of the system NaCI-HzO from 380 to 415 ° C: A refined determination of the critical point and two-phase boundary of seawater. Geochimica et Cosmochimica Acta, 52, 21212126. -& PrrZER, K.S. 1989. Liquid-vapor relations from the system NaCI-HzO: Summary of the P-T-x surface from 300° to 500° C. American Journal of Science, 289,217-248. BRAULT, M. & SIMONEIT, B.R.T. 1988. Steroid and triterpenoid distributions in Bransfield Strait sediments: Hydrothermaliy-enhanced diagenetic transformations. In: Advances in Organic Geochemistry 1987. Organic Geochemistry, 13, 697705. &~ 1989. Trace petroliferous organic matter associated with hydrothermal minerals from the Mid-Atlantic Ridge at the Trans-Atlantic Geotraverse 26°N site. Journal of Geophysical Research, 94, 9791-9798. --, MARTY, J.C. & SALIOT, A. 1985. Les hydrocarbures dans le syst6me hydrothermal de la ride Est-Pacifique, ~ 13° N. Comptes Rendus de l'Acad~mie des Sciences, Paris, 301, H, 807-812. -- & 1988. Hydrocarbons in waters and particulate material from hydrothermal environments at the East Pacific Rise, 13°N. Organic Geochemistry, 12,209-219. -& SALIOT, A. 1989. Trace petroliferous organic matter associated with massive hydrothermal sulfides from the East Pacific Rise at 13° N and 21° N. Oceanologica Acta, 12,405--415. CARRANZA-EDWARDS, A., ROSALES-HOz, L., AGUAYO-CAMARGO, J.E., LOZANO-SANTACRUZ, R. & HORNELAS-OROZCO,Y. 1990. Geochemical study of hydrothermal core sediments and rocks from the Guaymas Basin, Gulf of California. In: SIMONEIT,B.R.T. (ed.) Organic Matter Alteration in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 77-82. CHEN, C.-T.A. 1981. Geothermal systems at 21° N. Science, 211,298. CLIFTON, C.G., WALTERS,C.C. & SIMONEIT, B.R.T. 1990. Hydrothermal petroleums from Yellowstone National Park, Wyoming, U.S.A. In: SIMONEIT,B.R.T. (ed.) Organic Matter in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5,169-191.
I thank the Deep Sea Drilling Project and the National Science Foundation for access to DSDP and ODP samples and participation on the DSV Alvin cruises; J. Baross, M. Brault, C.G. Clifton, B.M. Didyk, E.M. Galimov, J.M. Gieskes, W.D. Goodfellow, P.D. Jenden, O.E. Kawka, K.A. Kvenvolden, R.N. Leif, P.F. Lonsdale, A. Lorre, J.D. Love, M.A. Mazurek, J.M. Peter, R.P. Philp, P.A. Rona, E. Ruth, A. Saliot, M. Schoell, K.A. Sundell, J. Tiercelin and C.C. Waiters for samples, data and assistance, and the reviewers for helpful comments and suggestions to improve this manuscript. Funding from the National Science Foundation, Division of Ocean Sciences (Grant OCE-9002366) and the National Aeronautics and Space Administration (Grant NAGW-2833) is gratefully acknowledged.
272
B.R.T. SIMONEIT
of Guaymas Basin, Gulf of California. Journal of CONNOLLY, J.F. 1966. Solubility of hydrocarbons Geophysical Research, 90, 6695-6707. in water near the critical solution temperatures. KULM, L.D., SUESS, E., MOORE, J.C., CARSON, B., Journal of Chemical Engineering Data, 11, 13-16. LEWIS, B.T., RITGER,S.T., KAOKO,D.C., THORNCOOPER,J.E. & BRAY,E.E. 1963. A postulated role of BURG, T.M., EMBLEY, R.W., RUGH, W.D., fatty acids in petroleum formation. Geochimica et MASSOTH, G.J., LANGSETH, M.G., COCHRANE, Comochimica Acta, 29, 1113-1127. G.R. & SCAMMAN,R.L. 1986. Oregon subduction CURRAY,J.R., MOORE,D.G. & ~7 OTHERS.1982. Initial zone: Venting, fauna and carbonates. Science, Reports of the Deep Sea Drilling Project, 64. US Government Printing Office, Washington, D.C. 231,561-566. KVENVOLDEN, K.A. & SIMONEIT, B.R.T. 1990. CZOCHANSKA, Z., SHEPPARD, C.M., WESTON, R.J., WOOLHOUSE,A.D. & COOK, R.A. 1986. Organic Hydrothermally derived petroleum: Examples geochemistry of sediments in New Zealand, Part from Guaymas Basin, Gulf of California and I. A biomarker study of the petroleum seepage at Escanaba Trough, Northeast Pacific. American Association of Petroleum Geologists Bulletin, 74, the geothermal region of Waiotapu. Geochimica et Cosmochimica Acta, 50,507-515. 223-237. RAPP, J.B. & HOSTETTLER, F.D. 1990. HydroDAVIS, E., MOTrL, M. & SHIPBOARDSCIENTIFICPARTY ~ , 1992. Proceedings of the Ocean Drilling Program, carbon geochemistry of hydrothermallyInitial Reports, 139. Ocean Drilling Program, generated petroleum from Escanaba Trough, College Station, TX. offshore California. In: SIMONEIT, B.R.T. (ed.) DIDYK,B.M. & SIMONEIT,B.R.T. 1989. Hydrothermal Organic Matter Alteration in Hydrothermal Sysoil of Guaymas Basin and implications for tems - Petroleum Generation, Migration and Biogeochem~try. Applied Geochemistry, 5, petroleum formation mechanisms. Nature, 342, 65-69. 83-91. - & - - 1990. Petroleum characteristics of the oil --, MORTON, J.L., KING, J.D. & in a Guaymas Basin hydrothermal chimney. In: CLAYPOOL,G.E. 1986. Petroleum associated with SIMONEIT,B.R.T. (ed.) Organic Matter A Iteration polymetallic sulfide in sediment from Gorda in Hydrothermal Systems - Petroleum GenerRidge. Science, 234, 1231-1234. LEIF, R.N., SIMONHT, B.R.T & KVENVOLDEN,K.A. ation, Migration and Biogeochemistry. Applied Geochemistry, 5, 29-40. 1991. Simulation of hydrothermal petroleum GALIMOV,E.M. & SIMONEIT,B.R.T 1982. Geochemisgeneration by laboratory hydrous pyrolysis. In: MANNING, D.A.C. (ed.) Organic Geochemistry, try of interstitial gases in sedimentary deposits of the Gulf of California, Deep Sea Drilling Project Advances and Applications in the Natural EnLeg 64. In: CURRAY, J.R., MOORE, D.G. et al. vironment. Manchester University Press, Man(eds) Initial Reports of the Deep Sea Drilling chester, 300-303. Project, 64, US Government Printing Office, ,-- & 1992. Hydrous pyrolysis of n-C32H66 Washington, D.C., 781-788. in the presence and absence of inorganic comGIESKES, J.M., SIMONEIT,B.R.T., BROWN,T., SHAW, ponents. American Chemical Society, Division of Fuel Chemistry, 204th National Meeting, PreT., WANG, Y.-C. • MAGENHEIM, A. 1988. Hydrothermal fluids and petroleum in surface prints, 37(4), 1748-1753. sediments of Guaymas Basin, Gulf of California: LONSDALE, P. 1985. A transform continental margin rich in hydrocarbons, Gulf of California. AmeriA case study. The Canadian Mineralogist, 26, 589-602. can Association of Petroleum Geologists Bulletin, HUNT, J.M. 1979. Petroleum Geochemistry and Ge69, 1160--1180. ology. W.H. Freeman and Company. & BECKER, K. 1985. Hydrothermal plumes, hot springs, and conductive heat flow in the Southern JONES, M.L. (ed.) 1985. Hydrothermal vents of the Trough of Guaymas Basin. Earth and Planetary Eastern Pacific: An overview. Bulletin of the Science Letters, 73, 211-225. Biological Society of Washington, 6, 1-566. JOSEPHSON, J. 1982. Supercritical fluids. EnvironLOVE, J.D. & GOOD, J.M. 1970. Hydrocarbons in mental Science and Technology, 16,548A-551A. thermal areas, northwestern Yellowstone National KAWKA, O.E. ~¢. SIMONEIT,B.R.T. 1987. Survey of Park, Wyoming. US Geological Survey Prohydrothermally-generated petroleums from the fessional Paper, 7644-B.' Guaymas Basin spreading center. Organic GeoMEREWETHER, R., OLSSON, M.S. & LONSDALE, P. chemistry, II, 311-328. 1985. Acoustically detected hydrocarbon plumes rising from 2-km depths in Guaymas Basin, Gulf - &~ 1990. Polycyclic aromatic hydrocarbons in hydrothermal petroleums from the Guaymas of California. Journal of Geophysical Research, Basin spreading center. In: SIMONE|T, B.R.T. 90, 3075-3085. (ed.) Organic Matter Alteration in Hydrothermal MICHAELIS,W., JENISCH,A. & RICHNOW,H.H. 1990. Systems - Petroleum Generation, Migration and Hydrothermai petroleum generation in Red Sea Biogeochemistry. Applied Geochemistry, 5, sediments from the Kebrit and Shaban Deeps. In: 17-27. SIMONHT, B.R.T (ed.) Organic Matter Alteration KOSKL R.A., LONSDALE, P.F., SHANKS, W.C., in Hydrothermal Systems - Petroleum GenerBERNDT, M.E. & HOWE, S.S. 1985. Mineralogy ation, Migration and Biogeochemistry. Applied and geochemistry of a sediment-hosted hydroGeochemistry, 5, 103-114. thermal sulfide deposit from the southern trough PAULAITUS, M.E., KRUKONIS,V.J., KURNIK, R.T. &
H Y D R O T H E R M A L PETROLEUM REID, R.C. 1982. Supercritical fluid extraction. Reviews in Chemical Engineering, l, 179-250. PETER, J.M. 1986. Genesis of hydrothermal vent
deposits in the southern trough of Guaymas Basin, Gulf of California: A mineralogical and geochemical study. MSc thesis, University of Toronto. ~,
SIMONEIT,B.R.T., KAWKA,O.E. & ScoTr, S.D. 1990. Liquid hydrocarbon-bearing inclusions in modern hydrothermal chimneys and mounds from the southern trough of Guaymas Basin. In: SIMONErr, B.R.T. (ed.) Organic Matter Alteration
in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 51--63. ~,
PELTONEN,P., SCOTI', S.D., SIMONEIT,B.R.T. & KAWKA, O.E. 1991. Carbon-14 ages of hydrothermal petroleum and carbonate in Guaymas Basin, Gulf of California - implications for oil generation, expulsion and migration. Geology, 19,253--256. PITZER, K.S. 1986. Large-scale fluctuations and the critical behavior of dilute NaCI in H20. Journal of Physical Chemistry, 90, 1502-1504. POLAr, J.T., BAEABAN, M., PEPLOW, A. & PHLIPS, A.J. 1989. Supercritical carbon dioxide extraction of lipids from algae. In: JOHNSTON, K.P. & PENNINGER, J.M.L. (eds), Supercritical Fluid Science and Technology. American Chemical Society, Symposium Series, Washington, DC, 4 0 6 , 4 4 9 - 4 6 6 .
water. Industrial Engineering Chemistry Fundamentals, 25, 169-171. ScoT, S.D. 1985. Seafloor polymetallic sulfide deposits: Modern and ancient. Marine Mining, 5, 191-212. SHAW, R.W., BRILL, T.B., CLIFFORD, A.A., ECKERT, C.E. & FRANCK,E.U. 1991. Supercritical water, a medium for chemistry. Chemical Engineering News, Dec. 23, 26--38. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16, 886-890. 1989. Corrections to 'Organic acid metastability in sedimentary basins'. Geology, 17,572-573. 1990. Chemical constraints on the origin of organic compounds in hydrothermal systems.
-
-
Origins of Life and Evolution of the Biosphere, 20, 331-367.
SILFER,J.A., ENGEL,M.H. & MAcro, S.A. 1990. The effect of hydrothermal processes on the distribution and stereochemistry of amino acids in Recent Antarctic sediments. In: SIMONEIT, B. R.T. (ed.) Organic Matter A Iteration in Hydro-
thermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 159-167. SIMONEIT, B.R.T. 1978. The organic chemistry of marine sediments. In: RILEY, J.P. & CHESTER, R. (eds) Chemical Oceanography, 7, Academic -
Press, 233-311. 1983. Organic matter maturation and petroleum genesis: Geothermal versus hydrothermal. In:
-
PRICE, L.C. 1976. Aqueous solubility of petroleum as applied to its origin and primary migration.
The Role of Heat in the Development of Energy and Mineral Resources in the Northern Basin and Range Province. Geothermal Research Council,
American Association of Petroleum Geologists Bulletin, 60, 213-244. , WENGER,L.M., GING, T. & BLOUNT, C.W. 1983. Solubility of crude oil in methane as a function of pressure and temperature. Organic Geochemistry, 4,201-221. RONA, P.A. 1984. Hydrothermal mineralization at seafloor spreading centers. Earth Science Reviews, 20, 1-104. 1988. Hydrothermal mineralization at oceanic ridges. The Canadian Mineralogist, 26, 431-465. ~, THOMPSON, G., Mot"rE, M.J., KARSON, J.A., JENKINS,W.J., GRAHAM,D., MALLETTE,M., VON DAMM, K. & EDMOND, J.M. 1984. Hydrothermai activity at the Trans-Atlantic Geotraverse hydrothermal field, Mid-Atlantic Ridge Crest at 26 ° N. Journal of Geophysical Research, 89, 1136511377. Ross, D.S. 1984. Coal conversion in carbon monoxide - water systems. Coal Science, 3,301-337. ~, HUM, G.P., MIIN, T.-C., GREEN, T.K. & MANSINI, R. 1986. Supercritical water/CO liquefaction and a model for coal conversion. Fuel Process Technology, 12,277-285. SAKAI, H., GAMO, T., KIM, E.-S., TSUTSUMI, M., TANAKA,T., ISHIBASHI,J., WAKITA,H., YAMANO, M. & OOMORI, T. 1990. Venting of carbon dioxide-rich fluid and hydrate formation in midOkinawa Trough backarc basin. Science, 248, 1093-1096. SANDERS, N.D. 1986. Visual observation of the solubility of heavy hydrocarbons in near-critical
273
-
-
-
-
1
9
Special Reports, 13. Davis, California, 215-241. 1985a. Hydrothermal petroleum: Genesis, migration and deposition in Guaymas Basin, Gulf of California. Canadian Journal of Earth Sciences, 22, 1919-1929. Hydrothermal petroleum: Composition and utility as a biogenic carbon source. In: JONES, M.L. (ed.) Hydrothermal Vents of the Eastern 8
5
b
.
Pacific: An Overview. Bulletin of the Biological Society of Washington, 6, 49-56. 1994. Lipid/bitumen maturation by hydrothermal activity in sediments of Middle Valley, ODP Leg 139. In: MOTTL, M., DAVIES, E. et al. (eds)
Proceedings of the Ocean Drilling Program, Scientific Results, 139, in press. GALIMOV, E.M. 1984. Geochemistry of interstitial gases in Quaternary sediments of the Gulf of California. Chemical Geology, 43, 151166. & KVENVOLDEN,K.A. 1994. Comparison of 14C ages of hydrothermal petroleums. Organic Geochemistry, in press. & LONSDALE, P.F. 1982. Hydrothermal petroleum in mineralized mounds at the seabed of Guaymas Basin. Nature, 295, 198-202. --, BRAULT,M. & SALIOT,A. 1990a. Hydrocarbons associated with hydrothermal minerals, vent waters and talus on the East Pacific Rise and Mid-Atlantic Ridge. In: SIMONEIT,B.R.T. (ed.)
-
-
-
-
Organic Matter Alteration
in Hydrothermal
274
B.R.T. SIMONEIT
Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 115124. , GOODFELLOW,W.D. & FRANKLIN,J.M. 1992b. Hydrothermal petroleum at the seafloor and organic matter alteration in sediments of Middle Valley, Northern Juan de Fuca Ridge. Applied Geochemistry, 7,257-264. - - , GRIMALT, J.O., HAYES, J.M. & HARTMAN,H. 1987. LOWtemperature hydrothermal maturation of organic matter in sediments from the Atlantis II Deep, Red Sea. Geochimica et Cosmochimica Acta, 51,879-894. ~, KAWKA, O.E. & BRAULT, M. 1988. Origin of gases and condensates in the Guaymas Basin hydrothermal system. In: Origins of Methane in the Earth (ed. M. Schoell). Chemical Geology, 71, 169-182. - & WANG, G.-M. 1992a. Biomarker maturation in contemporary hydrothermal systems, alteration of immature organic matter in zero geological time. In: MOLDOWAN,J., PHILP, R.P. & ALBRECHT, P. (eds) Biological Markers in Sediments and Petroleum. Prentice Hall, Englewood Cliffs, NJ, 124-141. --, LONSDALE, P.F., EDMOND, J.M. & SHANKS, W.C., III. 1990b. Deep-water hydrocarbon seeps in Guaymas Basin, Gulf of California. In: SIMONEIT, B.R.T (ed.) Organic Matter Alteration in Hydrothermal Systems - Petroleum Generation, Migration and Biogeochemistry. Applied Geochemistry, 5, 41-49. SISKIN, M. & KATRITZKY,A.R. 1991. Reactivity of organic compounds in hot water: Geochemical and technological implications. Science, 254, 231-237. SPIKER, E.C. • SIMONEIT, B.R.T 1982. Radiocarbon dating of Recent sediments from Leg 64, Gulf of California. In: CURRAY,J.R., MOORE,D.G. et al. (eds) Initial Reports of the Deep Sea Drilling Project. US Government Printing Office, Washington, DC, 757-758. SUESS, E., SIMONErr, B.R.T, WEFER,G., WHITICAR, M.J., FISK, M., TORRES, M., HAN, M.W.,
WI~STOCK, R. & LABAN, C. 1994. Hydrothermalism in the Bransfield Strait, Antarctica. Geologische Rundschau, in preparation. TIERCELIN, J.-J., BOUL~GUE,J. & SIMONEIT, B.R.T. 1993. Hydrocarbon, sulphide and carbonate deposits related to sublacustrine hydrothermal seeps in the North Tanganyika Trough, East African Rift. In: PARNELL, J., KUCHA, H. & LANDAIS, P. (eds) Bitumens in Ore Deposits, Springer Verlag, Berlin, 96-113. - - , THOURIN, C., KALALA,T. & MONDEGEUR, A. 1989. Discovery of sublacustrine hydrothermal activity and associated massive sulfides and hydrocarbons in the north Tanganyika trough, East African Rift. Geology, 17, 1053-1056. TISSOT, B.P. & WELTE, D.H. 1984. Petroleum Formation and Occurrence: A New Approach to Oil and Gas Exploration 2nd ed. Springer Verlag, Berlin. TODHEIDE, K. 1982. Hydrothermal solutions. Berichte der Bunsengesellschaft fiir Physikalische Chemie, 8 6 , 1005-1016. TUNNICLIFFE, V. 1991. The biology of hydrothermal vents: Ecology and evolution. In: BARNES, M. (ed.) Oceanography and Marine Biology Annual Reviews, 29. Aberdeen University Press, 319407. 1992. Hydrothermai-vent communities of the deep sea. American Scientist, 80, 336-349. WELHAN, J.A. & LUPTON, J.E. 1987. Light hydrocarbon gases in Guaymas Basin hydrothermal fluids: Thermogenic versus abiogenic origin. American Association of Petroleum Geologists Bulletin, 71,215-223. WHELAN,J.K., SIMONErr, B.R.T. & TARAEA,M. 1988. C1-C8 hydrocarbons in sediments from Guaymas Basin, Gulf of California - comparison to Peru Margin, Japan Trench and California Borderlands. Organic Geochemistry, 12, 171-194. WHITICAR, M.J., SUESS, E. & WEHNER, H. 1985. Thermogenic hydrocarbons in surface sediments of the Bransfield Strait, Antarctic Peninsula. Nature, 314, 87-90.
Hydrocarbons and other fluids: paragenesis, interactions and exploration potential inferred from petrographic studies JOHN PARNELL
Department of Geology, Queen's University of Belfast, Belfast BT7 INN, UK Abstract: Residues of hydrocarbons (bitumens sensu lato) are found in a wide range of settings in addition to reservoir porosity, including microfractures, megafractures, and several types of vein porosity. Paragenetic relationships between hydrocarbons and inorganic minerals provide information on the relative timing of hydrocarbon migration and the migration of other fluids. Hydrocarbons in reservoir rocks and mineralized fractures are paragenetically late as they are in pre-existing fluid pathways. Some other types of hydrocarbon occurrence are paragenetically earlier, especially proximal to the source rock. Different types of ore mineralization (notably base metal and uranium) in which hydrocarbons occur exhibit varying parageneses which represent distinct burial/thermal histories responsible for the ore types. Co-migration of hydrocarbons and silicon-bearing fluids is evinced by the occurrence of large quantities of authigenic silicate minerals within hydrocarbon residues. Interactions between metals and hydrocarbons may be exploited in the exploration for both metals and hydrocarbons. Metal-rich fossil fuels can preserve a signature of anomalous metal concentrations in the vicinity of ore deposits. Hydrocarbons precipitated by irradiation-solidification processes about radioelement-bearing phases help to identify hydrocarbon migration pathways and in some circumstances allow isotopic dating of migration.
There are numerous records of the paragenesis of hydrocarbons (oil or oil residues; bitumens sensu lato) and inorganic mineral phases, particularly in siliciclastic reservoir rocks and to a lesser extent in mineralized veins. The enormous database of the diagenesis of sandstones prior to hydrocarbon migration, recorded to assess poroperm constraints, is also a record of the composition of the fluid which migrated through the rock before the hydrocarbon fluid. The sequence of hydrocarbons and inorganic phases represents the changing composition of fluids passing through a sampling point, but not necessarily the sequence of fluid generation, as the origins of successive fluid pulses may be at widely different depths. It is reasonable to expect that parageneses recorded proximal to the hydrocarbon source rock would include hydrocarbons at a relatively early stage compared to parageneses in a more distal setting. As an introduction to this account I review the paragenesis of hydrocarbon and other fluids in a wide variety of settings (Fig. 1), including intraformational veinlets within the source rock, deposits in overpressured systems along bedding planes, reservoir/migration pathways at proximal and distal points and involving interaction with metalliferous groundwaters, hydrothermal systems which pass through the source rock, large bitumen veins which usually originate close
to the source rock, and faults (generally thrusts) in which hydrocarbon lubrication may have facilitated displacement. The exploration potential of hydrocarbon-inorganic interactions is subsequently reviewed, particularly in the light of the paragenetic relationships between hydrocarbons and other (mostly metalliferous) fluids.
Paragenetic context of hydrocarbons in varied settings
Fracture-hosted bitumens Solid veins of bitumen (A in Fig. 1) occur in successions which have very high hydrocarbon source rock potential (high TOC, and hydrogen indices often above 500 mg g-a) and rapid burial rates, causing the generation of substantial volumes of hydrocarbon over a geologically short time period (Monson & Parnell 1992). The veins are thought to be a consequence of this hydrocarbon generation, which is in excess of that which can migrate away through intergranular pathways. This causes a pressure build-up which is ultimately released by fracturing and migration through vein systems. As hydrocarbons are the cause of the fractures, they are in most cases the first fluids to enter the fractures: solid bitumen veins generally show
FromPARNELL,J. (ed.), 1994, Geofluids:Origin, Migrationand Evolution of Fluidsin SedimentaryBasins, 275 Geological Society Special Publication No. 78,275-291.
276
J. PARNELL
~bA
..... . E ~reservoJr ~ F .... \~,,~ JJ rock
Itumen
,11
ein
/
/
groundwater
E SOURCE
hydrothermal fluid circulation
B
intraformational bitumen
G
~.
C
overpressuring
/
Fig. 1. Schematic diagram showing settings in which paragenesis of hydrocarbons (bitumens) and other phases is discussed (letter codes in text). clean contacts with their wallrocks. However in some examples there is evidence for a very early, and volumetrically limited, stage of silica precipitation; quartz crystals lining the fractures or even suspended (coeval) in the bitumen. Examples of early authigenic quartz suspended in bitumen from Utah (Monson & Parnell 1992), Oklahoma (Monson pers. comm.) and Scotland yield fluid inclusion homogenization temperatures of about 100°C. Generally, these large bitumen veins are very pure as they have not functioned as pathways for subsequent fluids. Exceptions are some carbonate-rich successions where the bitumen has been followed and vigorously shattered/displaced by calcite precipitation; examples are seen in veins in the Neuquen basin, Argentina and at Bentheim, Germany. Vein bitumens also have very limited metal contents, excepting vanadium and nickel contents which may be high due to inheritance of metal in organic molecules from the source rock. Examples are amongst the vein bitumens in the Ouachita Mountains, Oklahoma, which were formerly mined for bitumen and as a source of vanadium (Ham 1956). Other metals are not enriched within the bitumens because of lack of opportunity for them to be introduced by other fluids. Many sequences which contain source rocks contain minor intraformational fracture-fillings of bitumen (B in Fig. 1), which are generally pure or associated with calcite. The high pressures developed during excessive hydrocarbon generation may allow the pooling of hydrocarbons along bedding planes as their fluid pressure supports overburden (overpressuring, C in Fig. 1) and holds bedding surfaces apart. In such instances, pools cause source rocks to function also as reservoir rocks, as in the West Siberian Basin (Peterson & Clarke 1991). Upon
uplift, these hydrocarbon pools result in bedding-parallel bitumen veins. As in the large bitumen veins, hydrocarbons were generally the first, causative, fluid, or sometimes coeval with/followed shortly by carbonate precipitation. The calcite in these veinlets commonly exhibits spherular interfaces with the bitumen, indicating coeval precipitation and poisoning of crystal growth by the bitumen. In other cases, the carbonate can be fibrous (Fig. 2), a petrographic feature interpreted as additional evidence for overpressuring (Stoneley 1983). The generation of hydrocarbons, with or without the development of overpressuring, may cause lubrication of the source and other rocks along migration pathways. Lubrication allows structural deformation, including thrusting (D in Fig. 1) in compressional regimes. Low-angle veins are commonplace in thrusted rocks (de Roo & Weber 1992), particularly rocks which are finely laminated as are many hydrocarbon source rocks. The veins contain traces of the hydrocarbons, and some veins along thrust planes are predominantly bitumen (Fig. 3). Hydrocarbons may also enter the thrust planes at a later stage, either during deformation (e.g. Roberts 1991) or after the cessation of deformation in which case they would not have a causative role. Small- (centimetre-) scale bitumen veinlets along decollement surfaces are also often pure, or a mixture of bitumen and calcite. Again the hydrocarbons, having a causative role, are an early component of veins, predating later carbonates and quartz.
Reservoir bitumens Hydrocarbons entering reservoir rocks (E in Fig. 1) displace aqueous pore fluids, which themselves may have been subject to long-term
PETROGRAPHIC STUDIES
~:~+ } + ~ +:i:e; ; +++' + +++ ~+:++ ; +~'+++++ C,+~++ +i ~ +++~' ....>
~'~+II~..+ +
_~'+'
.... '++++ .... +~25ppb ~ g a s 15-25ppb [----~gas HCO3-, so that evaporating seawater never becomes bicarbonate-rich. However, continental evaporites often precipitate from HCO3--dominated waters leading to the precipitation of a wide range of carbonate minerals including trona (NaHCO3.NazCO3.2H20) and burkeite (Na2CO3.2Na2SO4) (Eugster 1980). It is important for comparative purposes to quantify the redox conditions of eodiagenesis. The redox conditions of diagenetic waters are often expressed in terms of redox potentials (Eh) or oxygen fugacities (/'o2). However,
natural waters rarely approach electrochemical redox equilibrium so that Eh values are difficult to interpret in terms of overall system redox states. Additionally, foz values do not always have a clear physical meaning, and although in groundwaters Eh values can often be measured, fo2 values must generally be calculated from analytical data. Furthermore, foe and Eh values do not show a simple correspondence, and for a given aqueous system, both parameters will change with temperature making comparisons between redox conditions at different temperatures imprecise. In spite of these problems, both Eh and fo2 are of practical value if used as semi-quantitative indices of redox, and enable general comparisons to be made between redox conditions in different environments. Eodiagenetic waters will include some of the most oxidizing red bed formation waters, and organic-free waters at 25°C will have redox states up to approximately Eh ~ +500 mV, due to buffering by the atmosphere, with a redox state represented by 10~eO2,bars = - 0 . 7 . However, it is likely that eodiagenetic waters originating in density-stratified saline lakes may
304
R. METCALFE E T AL.
become locally very reducing, with Eh = - 4 0 0 to - 5 0 0 m V , due to the trapping of organic matter in bottom brines (Sonnenfeld 1984). The pH conditions are likely to be quite variable, generally lying in the range 6-10, which is typical of near-surface waters (Stumm & Morgan 1981), but possibly reaching >10 in continental evaporating lakes (e.g. Sonnenfeld 1984; Darragi & Tardy 1987). In such continental settings, HCO3- is often the most abundant anion and evaporation may lead to the precipitation of carbonates and smectites, thereby reducing the Ca and Mg content of waters to the point where they are N a - K - H C O 3 - dominated. When CO2 is independently buffered, for example by the atmosphere, then the pH is governed by: HCO3- + H + = CO2(g) +
H20
6- '
~
Q
u
a
r
t
z
.
Saturation
5-
7" 4
~K-feldspar t
\,,
E +
M 3
........"\-'~,
2
A~
'
[
", g?o
i
o
"',,
(1) ]Pyrophyllite]
fc°2[H20]
[r~CO3-1[U+l
= Constant
(2)
This means that pH may reach >10 because as [HCO3-] increases during evaporation, [H +] must decrease. Where waters were of marine origin, and in the absence of microbial activity, pH conditions will generally have decreased as ionic strengths increased during evaporation, owing to changes in the activity coefficients of the carbonate species which buffer pH (Krumgalz 1980). For example, during atmosphere-equilibrated seawater evaporation at 25 ° C, the pH falls from 8.2 to 6.7 at the point when halite saturates. However, microbial processes, such as sulphate reduction may lead to an increase in pH to 8-9 (Sonnenfeld 1984). Burial diagenesis (mesodiagenesis). As a red bed is buried, it becomes isolated from near surface conditions and undergoes burial diagenesis (mesodiagenesis) with formation waters becoming progressively more reducing and of generally lower pH. It is characterized by continuing framework grain dissolution, notably of K-feldspar, plagioclase, biotite etc, transformation of early diagenetic minerals to higher temperature minerals, and precipitation of higher temperature authigenic minerals. Early diagenetic smectite transforms to iilite, leading to the liberation of aqueous silica, Fe and Mg, according to reactions such as: 1.84Ca0.025N aoiK02Mg1-15Fe2+o,sFe3+0.2AI3+ 1.25Si3.50,o(OH)2
[Smectite] + 0.232K+ + 6.72H+ = K0.6Mg025AI23Si3sO,0(OH)2 [Inite] + 2.94SiO2(aq) + 0.92Fe2+0.368Fe3+ + 1.866Mg2++ 0.046Ca2+ + 0.184Na+ + 4.2H20. (3)
0
~
-6
r
~
1
-5
-4 -3 Log [SiO2(aq)]
-2
-1
Fig. 2. Logarithmic mineral stability diagram for the K20-SiO2-AI203-H20 system, showing how variations in K ÷ and SiO2 (aq) activities, and pH in diagenetic porewaters can be a major control on authigenic phases during red bed diagenesis. Phase boundaries were calculated for 25° C and 100° C using the SUPCRT92 code (Johnson et al. 1991). As K ÷ activities increase due to mineral dissolution during burial, red bed formation waters will move generally from the kaolinite stability field (A) to the muscovite stability field (B), and then to the K-feldspar stability field (C). Subsequent influx of dilute meteoric, or acidic basinal waters may then move the fluid composition back towards the kaolinite stability field. This leads to precipitation of authigenic quartz and feldspar overgrowths by the formation waters (Fig. 2; Bath et al. 1987; Strong & Milodowski 1987; McBride 1989). The formation of quartz cement in red beds is important because it is a major control of porosities and permeabilities, and hence influences fluid flow during diagenesis. In quartz-rich sandstones authigenic quartz precipitation may be the main factor causing porosity reduction (Leder & Park 1986). The various factors affecting quartz overgrowth formation have been reviewed by McBride (1989) who concluded that the inter-relationships between burial depths, temperatures and diagenetic mineral transformations generally cause quartz precipitation at temperatures of 60-100 ° C and depths of 1-2 km. Such depths coincide with the maximum theoretical formation water flow velocities of the order of 10 cm a -1 (Leder & Park 1986).
CONTINENTAL RED BED DIAGENESIS A consideration of the mechanisms of formation of such cements also aids our understanding of the volumes of fluid which flux through red beds during diagenesis. A major problem in explaining the formation of quartz cements in sandstones in general (not just red bed sandstones) is that there is often >10% imported silica and it is difficult to explain the large fluid fluxes required to achieve this. This subject has received considerable discussion in the literature (e.g. Leder & Park 1986; Morad & Aldahan 1987; Houseknecht 1987; Dutton & Land 1988; McBride 1989). The origins of the silica are generally not known, but likely contenders include dissolution of quartz, for instance by pressure solution, and liberation of silica during mineral reactions such as (3) above. At present there is no generally accepted source for the silica, but there is some agreement that very large volumes of water must flux through a given volume of rock and estimates range from 104-106 pore volumes. Additionally, it seems that silica must be transported over distances on the scale of kilometres in order to link possible sites of silica release to solution, to sites of cement production (McBride 1989). Burial diagenetic formation waters become progressively more reducing, and frequently reduce ferric iron. This is particularly the case when the red bed acts as a hydrocarbon reser~,oir, when conditions may become sufficiently reducing to cause the formation of pyrite from early hematite (Burley 1984). As the crystal structures of dolomite and calcite will accommodate Fe 2+ more readily than Fe 3÷, this causes late burial diagenetic carbonates to be typically ferroan, in contrast to early diagenetic carbonates (e.g. Burley 1984; Strong & Milodowski 1987). Modern deep groundwaters from red beds can place useful constraints upon the chemistry of mesodiagenetic waters. Waters from PermoTriassic red beds in the Wessex Basin, UK, reach temperatures of more than 75°C (Smith 1986) and can be considered diagenetic waters. Most deep waters in Permo-Triassic red beds of the UK are of Na-C1 type, while salinities are highly variable, but may approach halite saturation (>300 gl -a total dissolved solids (TDS); Edmunds 1986). The dominant origin of the salinity appears to be halite dissolution within the Permo-Triassic (Edmunds 1986), and it seems likely that this process has been a major control of water salinity throughout diagenesis. Neglecting localized redox variations, within individual organic-rich beds for instance, red bed sequences as a whole are likely to encounter the most reducing conditions during late burial
305
diagenesis, due to introduction of hydrocarbonbearing waters which originate in hydrocarbon source-rocks outside the red bed sequence (Fig. 1). The redox state of such waters can be estimated by assuming equilibria involving silicates to control the pH, and that equilibria involving light organic acids buffer the redox state. For example, if the activity of K ÷ is known, then the pH can be constrained by using equilibria such as: 2KA12(A1Si3Oa0)(OH)2 + 2H + + 3H20 Muscovite = 3AI2SizOs(OH)4 + 2K + Kaolinite
(4)
Using the calculated pH, the redox state of the water can be constrained with equilibria such as:
CH3COOH(aq) = CH3COO-(aq) + H+(aq)
(5)
C2HsCOOH(aq) ~- C2HsCOO-(aq) q- H+(aq)
(6)
and 3CH3COOH~aq) = 2C2UsCOOU(aq) + O2(g) (7) Appropriate concentrations of K + can be estimated from published analyses of oilfield brines (e. g. Gulf Coast USA data from Carpenter et al. 1974). The activity coefficients of K + and light organic species can then be calculated using a suitable equation of state such as the HKF equation (Helgeson et al. 1981), and the equilibrium constants for the relevant equilibria can be obtained using the computer code SUPCRT92 (Johnson et al. 1991), and thermodynamic data for organic species from Shock & Helgeson (1990). This suggests that at a temperature of 125 ° C, reducing late diagenetic waters will be slightly acidic, with a pH of c. 5 and a redox state represented by 1OgfO2,bars = C. --50 (approximately equivalent to Eh = - 2 0 0 mV). Movement and mixing of formation waters from different parts of a red bed sequence during mesodiagenesis may exert an important effect upon diagenetic mineral assemblages. For instance influxes of dilute meteoric waters which have low K+/H + activity ratios will tend to produce diagenetic kaolinite, whereas higher temperature waters have higher K + concentrations due to mineral dissolution and will produce diagenetic illite (represented by muscovite: Fig. 2). Cooling of high temperature muscovite-equilibrated waters (point C for 100°C on Fig. 2) may give rise to authigenic K-feldspar (e.g. at point C for 25°C on Fig. 2) during ascent. Diagenesis during uplift (telodiagenesis) Telodiagenesis is characterized by progressively decreasing temperatures, and formation water
306
R. METCALFE ET AL.
salinities and by progressively more oxidising conditions. This is due to the influx of dilute neutral to low pH meteoric waters which displace basinal brines and lead to evaporite dissolution and the removal of sulphate and halite cements (e.g. Burley 1984; Parnell 1992). This also causes carbonate cement and feldspar dissolution, and leads to the precipitation of hematite due to oxidation of Fe 2÷ which is liberated from dissolving ferroan carbonates. With increasing dilution, SiO2(aq) and K+(aq) activities are depressed, leading to the precipitation of kaolinite (Burley 1984; Bath et al. 1987; Strong & Milodowski 1987; Parnell 1992; Fig. 2). Modern groundwater chemistry can be used as an indicator of late diagenetic mineral transformations. Edmunds et al. (1984) and Bath et al. (1987) investigated groundwaters in the Triassic Sherwood Sandstone red bed aquifer of the East Midlands, UK. They found that invasion of meteoric waters had flushed any pre-existing Na-C1 dominated saline waters from the aquifer, and that the groundwater chemistry was dominated by reactions with carbonate cement, detrital dolomite and sulphate minerals. The pH of pure water which is equilibrated with modern atmospheric levels of CO: is c 5.6 (Stumm & Morgan 1981) and can be used as an approximate estimate of the minimum pH of the late diagenetic waters. However away from outcrop, the pH will rise owing to the interaction between the meteoric water and diagenetic carbonates in the red bed. In the Sherwood Sandstone, pH increased from c. 7 to c. 8 at greater distances from the outcrop (Edmunds et al. 1984). However, in spite of this flushing, there is still a considerable redox gradient down-dip. Edmunds et al. (1984) found that only the waters in the unconfined part of the aquifer are sufficiently oxidizing to contain dissolved oxygen, and down dip from the outcrop, Fe 2÷ and then F e 2÷ and HS- become important in controlling redox potentials, as Eh falls from +300 mV to 0 mV.
Types o f mineralization associated with red beds
Economically, copper mineralization is perhaps the most important type of red bed ore deposit. However, other heavy metals also occur within red bed sequences, including important ore deposits of U, Pb, Zn and Ag.
The occurrence and tectonic setting of 'sediment-hosted stratiform copper deposits' has been reviewed by Kirkham (1989). He concluded that ore deposits contained wholly within red beds are of relatively minor importance, and that most major ore bodies comprise disseminated sulphides in reducing sediments, such as organic-rich shales which lie immediately above continental clastic red beds. These reducing sediments acted as traps, stripping metals from relatively oxidizing, evaporite-derived brines which had previously extracted heavy metals from the red bed sediments themselves. An important control on the formation of such ore deposits is thought to be rifting (Jowett 1989; Kirkham 1989) because this acts to: (1) create an initial source of heavy metals through riftrelated volcanism; (2) cause rapid faultcontrolled sedimentation; (3) allows the development of an internal closed drainage, causing the formation of saline brines and evaporites; and (4) leads to the development of a basin and range topography and the establishment of a hydraulic regime likely to channel oxidizing brines towards overlying reduced strata during late diagenesis. In red bed-hosted ore deposits there may be evidence for a redox-related zonation of ore minerals. In general, ore deposits have a more reduced mineral assemblage than the initial host red bed, reflecting the relative solubilities of sulphide ore minerals. For example, Cu deposits may show a zonation which reflects the sequence of increasing solubility of sulphides, similar to: hematite-bearing barren rock-native copperchalcocite (Cu2S)-bornite (CusFeS4)-chalcopyrite (CuFeS2)-galena (PbS)-sphalerite (ZnS)-pyrite (FeS2) (e.g. Gustafson & Williams 1981; Kirkham 1989; Jowett 1992), although there are variations in this pattern. In red bed mineral deposits, Cu is often accompanied by Ag, and sometimes by Co, Pb, Zn and other metals. However, Gustafson & Williams (1981) comment that Pb-Zn deposits are not nearly as well linked to red beds as are Cu-deposits, and that major Pb-Zn associated with Cu only occurs at Mt Isa, Australia. Red bed ore deposits have metal abundances: C u > Z n , Pb which contrast with those of Mississippi Valley Type (MVT) ore deposits (Pb > Zn -> Cu). Uranium deposits are also an important feature of some red bed sequences and stratiform, mostly tabular deposits of the Colorado Plateau, USA are hosted by red bed sandstones (Kimberley 1979; Durrance 1986, Chapter 6).
CONTINENTAL RED BED DIAGENESIS
Distribution of metals in red beds Distribution o f metals in whole rocks Relatively few whole-rock analyses of red bed sediments have been published. Zielinski et al. (1983,1986) analysed modern red bed sediments and showed that heavy metal variations in sediments which have undergone only early diagenesis are primarily controlled by differences in sources and/or mineral fractionation during deposition. From their analyses they concluded that in young (Holocene-Pliocene) red beds in northern Baja California, there is an early diagenetic redistribution of metals on an intergranular scale, from detrital host silicates to authigenic Fe-oxides. The abundances of these Fe-oxides were found to increase with increasing sediment age, so that although there is desorption of metals from Fe-oxides as the oxides become more crystalline with age, the rock remains closed to metal movement. The Feoxide abundances are also lithology-dependent, with finer sediments having relatively high concentrations of heavy metals (Wedepohl 1978; Zielinski et al. 1983, 1986). Additionally, the early diagenetic transformation of labile minerals other than Fe-oxides, such as the alteration of plagioclase, hornblende and biotite to smectite may also be a control on the retention of heavy metals by the bulk rock. The proportion of the metal content which is mobilized, and the distance over which it is transported, is a function of fluid flow in the red bed, which in turn is porosity- and permeabilitydependent. Thus it is to be anticipated that coarse, relatively porous and permeable sediments would liberate a higher proportion of their metals to formation waters than finer sediments. This is supported by the results of experimental heavy metal leaching from red bed sediments, which suggest that during burial diagenesis, under equivalent conditions, coarse sediments and young red beds probably undergo relatively rapid removal of metals, although older red beds may be a greater source of metals due to their higher Fe-oxide abundances (Zielinski et al. 1983, 1986). Further information regarding the integrated movement of metals through red beds during burial and uplift-related diagenesis (mesodiagenesis and telodiagenesis) can be gained from whole-rock analyses of ancient red beds. Some data for the bulk content of heavy metals in red bed sediments are presented in Wedepohl (1978), and from these it seems that red bed sandstones have mean abundances of heavy metals of: Pb = 9.9 ppm (533 samples); Cu =
307
10.0ppm (382 samples); Zn = 30.9ppm (64 samples); and U abundances generally 5 molal (at halite saturation at 25 ° C ionic strengths are c. 7 molal) and therefore late burial diagenesis cannot be modelled accurately. Although the specific ion interaction model (Pitzer 1973) provides a theoretical framework for modelling such concentrated waters, the necessary data are extremely limited. A lack of Pitzer coefficients for silica means that silicate equilibria cannot be simulated. Furthermore, the data are applicable only over limited temperature ranges, with data for carbonates, for example, being limited to 25 ° C. When pH and redox states were allowed to drift from their initial values, it was found that Iogfo2,bar~ remained almost constant during burial diagenesis, whereas pH fell by about 1.5 units. This is due to the model calculating that all Fe 2÷ in the initial sediment was oxidised during early diagenesis under atmosphere-buffered conditions, so that during later diagenesis oxygen was not consumed by Fe-oxidation. In nature it is more likely that a significant proportion of the Fe will remain in the ferrous state until after burial has isolated the sediment from the atmosphere, and more reducing conditions will evolve as Fe-oxidation depletes oxygen in the formation waters. A further major problem is the prediction of geologically reasonable concentrations of metals in solution. In the example, only very small proportions of the available copper in the red bed were predicted to be released to solution (c. 1% for 1 M C1- solution), whereas all the
available Pb and Zn were liberated from the model rock. This contrasts with the experimental simulations of diagenesis performed by Zielinski et al. (1986) which indicated that up to c. 45% of Cu, but only c. 20% of Pb and Zn should be available to the solution. These experimental data suggest that 1 kg of sandstone could liberate up to c. 70 mg Cu, c. 20 mg Pb and c. 40 mg Zn (taking values for metal concentrations in sandstones from Gustafson & Williams 1981). One possible explanation is that because the model formation waters remained highly oxidizing, as discussed above, Cu is effectively locked in tenorite (Fig. 3). For this reason, additional models were run with more realistic reducing redox states (to lo~fo2,bars = - 4 0 ) being specified for the later diagenetic waters. However, these gave similar results whatever redox states were selected, with Cu concentrations remaining at c. 0.01 mg1-1, and Pb and Zn concentrations remaining at c. 20 mg 1-1 and c. 70 mg 1-1 respectively. Alternatively, the inconsistency with experimental results may possibly be due to: a poor knowledge of the actual phases which control the solubility of Pb, Zn and Cu during diagenesis; and a lack of thermodynamic data, especially high temperature data, for potential phases which might control Pb, Zn and Cu solubility. Fluid m i x i n g In order to assess the importance of fluid mixing it is necessary to estimate the proportions of fluid of different redox states which are necessary to
CONTINENTAL RED BED DIAGENESIS produce significant changes in the aqueous chemical speciation and precipitated mineral assemblages. It is instructive to consider the amount of reducing fluid which is required to mix with an oxidised, metal-bearing red bed formation water, in order to strip it of heavy metals owing to sulphide formation. The most reducing burial diagenetic water which can be envisaged is an oilfield water with a redox state imposed by organic matter. Such a redox state is appropriate for many reduced sediments which are actually associated with red bed Cu mineralisation, for example the Kupferschiefer. These conditions were modelled as illustrated in Fig. 8, by assuming that silicate equilibria controlled pH, organic acid equilibria controlled redox states (see equations 4-7 above), and that the water was equilibrated with a typical mudrock mineral assemblage such as might occur in a hydrocarbon source rock: plagioclase; illite; chlorite; calcite; quartz; galena; sphalerite; pyrite; and chalcopyrite. Salinity was specified by making the simple assumption that the solution was charge-balanced with CI-. Other parameters were fixed by using oilfield brine analyses as a guide to reasonable values. For instance, alkalinity was fixed by assuming 4000 mg 1-1 acetate (near the maximum concentration reported for oilfield brines by Carothers & Kharaka 1978). This gave a model reducing oilfield water of the composition shown in Table 3 and with a metal content of: 1.3 x 10 -3 mgl -~ Pb; 1.8 x 10 -2 mgl -~ Zn; and 6.3 x 10 -9 mgl -~ Cu. When the formation water is at its most oxidized the maximum amount of such a reducing oilfield water will be required to remove heavy metals completely. This limiting case was investigated by using the EQ6 code to model progressive addition of the reducing oilfield water to the model atmosphere-equilibrated, late burial diagenetic porewater described above. The model was terminated when sphalerite appeared as a product since the common mineral zonation in red bed ore deposits suggests that sphalerite generally occurs late in the sulphide paragenesis (e.g. Gustafson & Williams 1981). The results of this model are illustrated in Fig. 9 and Table 3, and show a sharp redox boundary when the reducing oilfield water constituted about 5% of the mixture. At this point, all of the Cu was removed from the solution, reflecting the insolubility of Cu-sulphides compared to Pb- and Zn-sulphides (Fig. 9a). This coincided with a redox boundary at which the water became dramatically more reducing, with Iogfo2,bars decreasing from c. - 5 to c. - 4 5 (Fig. 9b). The concentrations of Pb and
317
1oo!" ~
8o"
=
60-"
"~
40 :
I
cul
I
Pbl
a
I
20, o 100
10
g 80 -6 .=_ 60
I
.~
20
-cu
30
40
50 0
log fO 2
-]0 -20 ~
¢~ 4o
-30
20
-40
0 ~ # 3e-3 0 t~
. 10
20
~h muartz uscovite ematite lbite [ ] dolomite
~J
2e-3.
-°~'X:I3.---74>-~ -50 30 40 50 C
.¢///JJS
le-3. o
~0e+0 0
10 20 30 40 50 % of reducingoilfieldwater in mixture
Fig. 9. Illustration of the output from an EQ6 model for the mixing of evaporite-equilibrated late burial diagenetic porewater, and reducing organic-rich oilfield water. (a) Cu is stripped from the solution, and then Pb, followed by Zn begin to be stripped from solution as progressively larger amounts of reducing oilfield water are added to the mixture; (b) There is a dramatic change in redox state and decrease in Cu content when only c. 5% of the mixture is reducing, organic-rich oilfield water; (e) the model predicts a gangue mineral assemblage which is dominated by dolomite and barite, with trace quantities of quartz, muscovite, hematite and albite. The moles of minerals precipitating are those produced by a notional 1 kg of solution. Zn decreased only after all the Cu had been stripped from the solution, and the onset of Pb removal from solution commenced before the onset of Zn removal. Also, initially (during addition of the first 3% of reducing oilfield water) metal concentrations fell as a consequence of simple dilution (Fig. 9a). In the model, dolomite and barite were dominant predicted gangue minerals while quartz, muscovite, albite and hematite formed a minor proportion of the mineral assemblage (Fig. 9c). Trace amounts of albite and muscovite
318
R. METCALFE E T AL.
were present throughout the paragenesis, while trace quantities of hematite were predicted by the model initially, but disappeared when about 30% of the mixture was reducing oilfield water. Thus, the mixing model can reproduce the major features of many red bed-hosted copper deposits. It is significant that for this limiting case only 5% of reducing oilfield water is needed to strip all the Cu from the oxidized late burial diagenetic porewater. This implies that only a small amount of mixing between different fluids could be a major control of heavy metal mobility and ore deposit formation. The unrealistically small Cu concentrations in the initial late burial diagenetic porewater do not invalidate this conclusion because Cu-sulphides are relatively insoluble and in sulphur-rich fluids, such as those which are typical of evaporite-bearing red bed sequences, sulphur speciation is a major control on metal mobility. It is likely that formation water redox states more reducing than the atmosphere, and outside the stability field of Cu oxides (Fig. 3), would mobilize more Cu than in the model. However, such waters will also lie closer to the field of sulphide stability, so that smaller quantities of reducing water will be required to cause sufficient reduction of the system to result in Cu-sulphide precipitation.
Sorption
Equilibrium thermodynamic models are of limited applicability because they generally assume that mineral dissolution/precipitation equilibria govern the composition of the aqueous phase. However, equilibria between dissolved and sorbed species may be a major control on the composition of diagenetic waters, particularly in the case of trace metal constituents. The dominant sorbents in nature are the common hydrous oxides of Fe, AI, Mn and Si, and since Fe-oxides are a common constituent of red beds, sorption is likely to be particularly important in controlling metal mobility in red beds. Oxide and hydroxide surfaces are capable of adsorbing or dissociating H + from a surface O H group to acquire a net surface change, according to: S - OH + H + = SOH2 +
(10)
S - OH = SO- + H +
(11)
where S - O H indicates surface sites occupied by OH. These charged sites affect the electrostatic attraction for other ions on or near the surface
of the solid and create sites for further reaction for example by: S-OH+Me
+=SO-Me ++H +
(12)
where Me + is a metal cation. Ions bound only by the electrostatic forces are said to be non-specifically sorbed whereas those bound at surface sites are specifically sorbed. The abundance of these SOH2 + and SO- charged sites is solution pHsensitive, and hence the proportion of a metal cation which is specifically sorbed is a function of pH (Fig. 10), which in turn is a function of other equilibria, such as 4. Specific adsorption of cations onto oxide surfaces increases from low values to complete adsorption over a relatively small pH range to form an 'adsorption edge'. At low pH where the surface sites are generally positively charged sorption of metal cations is low, whereas at higher pH sorption increases and can result in the 100% retention of metals. At intermediate pH values sorption may cause retardation, but not complete retention of a metal. Factors other than pH which affect sorption are: (1) concentration of the metal, which will affect sorption as the number of occupied sites approaches a significant proportion of the total sites; (2) competing cations which will reduce the number of sites available to the cation of interest; (3) redox state, which controls aqueous speciation and the solubility of sorbing phases; (4) temperature, sorption generally decreasing as temperatures rise; (5) the sorbent/sorbate ratio. The importance of sorption during red bed diagenesis has been evaluated by calculating the position of the sorption edge for Cu, Pb, Zn, Co and Ag when these are sorbed by hydrous ferric oxide (FeOOH). The H Y D R A Q L code (Papelis et al. 1988) was used in conjunction with a diffuse layer model, surface complexation data from Dzombak & Morel (1990) and the HYD R A Q L database for solution species (Papelis et al. 1988). An oxide surface area of 600 m 2 g-~ was assumed and the surface densities of strong and weak sites were taken to be 0.005 and 0.2molmol -~ Fe respectively (Dzombak & Morel 1990). The effects of varying trace-metal concentration, ionic strength and surface site concentrations on the proportion of metal which is sorbed were investigated. Redox conditions were within the stability field of Cu 2+, owing to a lack of data for the cuprous ion. Under these conditions, increasing the ionic strength from
CONTINENTAL RED BED DIAGENESIS
i ~.;~,;~;~;;;I;;~S.;;;~;;~;;~;~;;;~;;;;1;;
1.0 : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : ii!iiiii!?iiiiiiiiiiiiiii?iiiiiiiiiiii~ii :::::::::::::::::::::::::::::::::::::::::
0.5
.
.
.
.
.
.
.
.
.
.
.
.
I
I
I
I
I
4.0
5.0
6.0
7.0
8.0
1
..~
frp/ /
~40t/
,,/ ~
/
/
/
/ ,
.
7 7
4.0
/
5.0
6.0
/
~40
/
1(89g/l Goethite) 10.1 M NaCI
i
i
8.0
SurfaceArea= 53.4m2/1 (0.89g/l Goethite)
.
,' // /
/
~
0.1MNaCI
co/
,
/'
20
9.0
T = 25°C
Pb p~C~_¢~ u277' / Z n / ~J /
x~60-
--n---n /,~'
7.0
.
80"
] ISurface Area= 53400m2/1
i,'~ p
120
100I C
I
9.0
~'
.of; / j 3.0
Approx~ate F,e~d
o T = 25 C
/
z4
As CuC132-
77-71 AgCl2Field
120-r t b
1001 •-~ 801
Field r ~ Approximate Copper Transport ~ 1Silver Transport
T=25°C aAg=1 aZS = 0.03 acu = 10-5 -0.5 PCO2= 0.01 barac1= 1 3.0
319
g
0 3.0
4.0
5.0
6.0
7.0
8.0
9.0
pH Fig. 10. A comparison between the fields of aqueous Cu- and Ag-chloride mobility in the absence of sorption (a), and calculated sorption edges for Pb, Cn, Zn, Co and Ag (b and c). The 'Approximate Field Copper Transport as Cu + - C1-1' in (a) is the same as the shaded field in Fig. 3. Sorption was modelled for Eh conditions consistent with Cu 2+ stability; Cu + is considered likely to sorb in a similar fashion to Ag +. It is evident that under the range of pH required for Cu- and Ag-chloride stability (a), significant proportions of Cu and Ag may be rendered immobile (b and c). In the presence of iron oxides, the fields of Cu and Ag mobility may be significantly smaller than the approximate fields of Cu- and Ag-chloride transportation in (a). 0.1 M to 1 M and varying the metal concentration by four orders of magnitude had no significant effect on the sorption curves. However increasing the surface area, or the density of sites available for sorption lowers the pH at which 100% of the metal is sorbed. Figure 10 shows the calculated graphs for sorption of Pb, Cu, Zn, Co and Ag (5 × 10 -7 M)
in 0.1 M NaC1 onto goethite with surface areas of 53400 and 53.4m z per litre of solution. The position of the pH edge is partly dependent on the sorbate/sorbent ratio. Similar graphs are obtained in each case, but the adsorption edge is shifted up by about 1.5 pH units in the lower sorbate/sorbent ratio case (Fig. 10b compared to Fig. 10c).
320
R. METCALFE E T AL.
Variable redox conditions during red bed diagenesis will not greatly affect sorption of Pb, Zn, Co and Ag, because the oxidation states of these metals will be mostly constant. In contrast, Cu can form Cu 2÷ and Cu + complexes, and redox may control the sorption of this metal, although a lack of data for Cu + prevented modelling this. Sorption of Cu + may be weaker than for Cu 2÷, because Cu ÷ probably behaves like Ag ÷ which forms strong chloride complexes (AgCI2-, AgCI32- and mgCl43-) and which is weakly sorbed at acid and neutral pH (Fig. 10b). Recently Rose & Bianchi-Mosquera (1993) carried out a series of experimental simulations of red bed conditions, and investigated sorption of Pb, Zn, Co, Ni, Cu and Ag by goethite under a range of redox conditions. Their experimental adsorption curves were similar to the modelled curves, but compared to Pb, the copper adsorption edge was at slightly lower, rather than slightly higher pH value, possibly because of differences in Eh, and hence speciation. From Fig. 10 it is clear that under conditions of pH similar to those bounding the field of Cu transport as cuprous chloride, goethite is likely to sorb all Cu 2+, and if Cu + behaves like Ag + and sorbent/sorbate ratios are suitable, may sorb significant amounts of Cu ÷. It is likely that Cu ÷ and Ag ÷ are likely to be more mobile if the adsorbing phase is hematite rather than goethite (Rose & Bianchi-Mosquera 1993). This is consistent with the release of heavy metals to solution due to the transformation of goethite to hematite during burial diagenesis.
migrate through a red bed sequence. This is because the temperature of metal mobilization is relatively unimportant in comparison with Eh-pH conditions. Current theoretical models are capable of reproducing the major features of the mineral parageneses actually observed in ore deposits, including: Cu-bearing sulphides which pre-date Pb- and Zn- sulphides; hematite which formed before the main sulphide-forming event; 'bleached' sandstones adjacent to sulphide mineralization, owing to diminished quantities of hematite; quartz overgrowths in sandstones which predated the main phase of sulphide formation; and the formation of late barite. An important conclusion from these models is that extremely small amounts of fluid mixing can exert a critical control on both diagenesis and upon ore formation. This means that even when there is limited fluid flow within a red bed sequence, due to small hydraulic head gradients and/or low permeabilities, sufficient fluid mixing might occur to cause significant variations in mineral parageneses. Sorption may be a significant control on metal cation partitioning between the rock and the fluid phase, and variations in the pH of the diagenetic fluids may cause variations in this partitioning which in turn lead to fractionation of heavy metals in solution. Our conceptual understanding of the major processes governing red bed diagenesis is relatively good, but there is a lack of realistic computer models and there are deficiencies in the data necessary for computer modelling, especially:
Conclusions
(1) uncertainty about the identities of mineral phases which control heavy metal solubility; (2) limited thermodynamic data at >25°C for likely metal solubility-controlling minerals and aqueous species; (3) a limited Pitzer coefficient dataset, in particular for silicon-bearing aqueous species, but also for all other aqueous species for the temperatures of interest; (4) a lack of kinetic data; (5) alack ofsorption data.
Red bed sequences are chemically heterogeneous, and although they largely have low initial organic contents, they may act as conduits and reservoirs for hydrocarbon-bearing fluids. Together these factors result in a unique diagenetic evolution of authigenic mineral assemblages and pore-water chemistries. Diagenesis occurs over ranges of redox and pH conditions (broadly, logfo2.bars = --0.7 to C. --50; and Eh ~ +500 mV to - 2 0 0 m V (although Eh may reach c. -500 mV in low temperature, alkaline environments); and pH c. 10 to c. 5) which are among the largest for any type of sedimentary sequence. There is a complex interplay between the evolution of red bed porosity and permeability, red bed mineralogy, and the chemical compositions of diagenetic fluids. Formation waters may evolve from metal-transporting fluids to metal precipitating fluids and back again as they
Additional uncertainties are due to current models being limited to simple scenarios such as closed system water-rock interactions which take little account of surface sorption or reaction kinetics. Future work should be directed towards developing open-system, coupled fluid flow-chemical transport models, which can incorporate a treatment of reaction kinetics and sorption.
CONTINENTAL RED BED DIAGENESIS The authors extend their thanks to P. Turner and B.A. Hofmann for helpful reviews of the manuscript. M.B. Crawford is thanked for running the H Y D R A Q L computer sorption models. The work reported here was carried out as part of the Cheshire Basin Project of the British Geological Survey. This paper is published with the permission of the Director of the British Geological Survey.
References BATH, A.H., MILODOWSKI, A.E. & STRONG, G.E. 1987. Fluid flow and diagenesis in the East Midlands Triassic sandstone aquifer: In: GOFF, J.C. • WILLIAMS, B.P.J. (eds) Fluid flow in sedimentary basins and aquifers. Geological Society, London, Special Publications, 34, 127-140. BECHTEL, A. & HOERNES, S. 1993. Stable isotopic variations in clay minerals: a key to the understanding of Kupferschiefer-type mineralization, Germany. Geochimica et Cosmochimica Acta, 57, 1799-1816. t~ POTTMANN, W. 1991. The origin of the Kupferschiefer-type mineralization in the Richelsdorf Hills, Germany, as deduced from stable isotope and organic geochemical studies. Chemical Geology, 91, 1-18. BROOKINS, D.G. 1984. Geochemical aspects of radioactive waste disposal. Springer-Verlag, New York. BRUTON, C.J. 1989. Predicting mineral dissolution and precipitation during burial: synthetic diagenetic sequences. In: JACKSON, K.J. & BOURCIER,W.L. (eds) Proceedings of the Workshop on Geochemical Modelling, CONF-8609134. Lawrence Livermore National Laboratory, Livermore, California, 111-119. BURLEY, S.D. 1984. Patterns of diagenesis in the Sherwood Sandstone Group (Triassic), United Kingdom. Clay Minerals, 19,403-440. CAROTHERS, W.W. & KHARAKA,Y.K. 1978. Aliphatic acid anions in oil-field waters-implications for origin of natural gas. American Association of Petroleum Geologists Bulletin, 12, 2441-2453. CARPENTER,A.B., TROUT,M.L. & P1CKETr,E.E. 1974. Preliminary report on the origin and chemical evolution of lead- and zinc-rich oilfield brines in central Mississippi. Economic Geology, 69,. 1191-1206. DARRAGI, F. & TARDY, Y. 1987. Authigenic trioctahedral smectites controlling pH, alkalinity, silica and magnesium concentrations in alkaline lakes. Chemical Geology, 63, 59-72. DURRANCE, E.M. 1986. Radioactivity in Geology: principles and applications. Ellis Horwood, Chichester. DUTrON, S.P. & LAND, L.S. 1988. Diagenetic history of a well-cemented quartz arenite, Lower Cretaceous, Travis Peak Formation, East Texas. Geological Society of America Bulletin, 100, 1271-1282. DZOMBAK, D.A. & MOREL, F.M.M. 1990. Surface Complexation Modelling. Wiley Interscience, New York.
321
EDMUNDS, W.M. 1986. Geochemistry of geothermal waters in the U.K. In: DOWmNG, R.A. & GRAY, D.A. (eds) Geothermal energy: the potential in the United Kingdom. Her Majesty's Stationery Ofrice, London, 111-122. ~, MILES, D.L. & COOK, J.M. 1984. A comparative study of sequential redox processes in three British aquifers. In: ERIKSSON, E. (ed.) Hydrochemical balances of freshwater systems. IAHS Publications, 50, 55-70. EUGSTER, H.P. 1980. Geochemistry of evaporitic lacustrine deposits. Annual Reviews in Earth and Planetary Sciences, 8, 35-63. -1985. Oil shales, evaporites and ore deposits. Geochimica et Cosmochimica Acta, 49,619-635. 1989. Geochemical environments of sedimenthosted Cu-Pb-Zn deposits. In: BOYLE, R.W., BROWN, A.C., JEFFERSON,C.W., JOWETT,E.C. & KIRKHAM,R.V. (eds) Sediment-hosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 111-126. FEIN, J.B. 1991. Experimental study of aluminium-, calcium- and magnesium-acetate complexing at 80° C. Geochimica et Cosmochimica Acta, 55, 955-964. FLINT, S. 1987. Diagenesis of Tertiary playa sandstones of Northern Chile; implications for Andean uplift and metallogeny. Sedimentology, 34, 11-29. FROSTICK, L.E. & REID, I. 1987. Tectonic control of desert sediments in rift basins ancient and modern. In: FROSTICK,L. & REID, I. (eds) Desert sediments: ancient and modern. Geological Society, London, Special Publications, 35, 53-68. GERTH, J. 1990. Unit cell dimensions of trace metalassociated goethites. Geochimica et Cosmochimica Acta, 54,363-371. GIORDANO, T.H. & DRUMMOND, S.E. 1991. The potentiometric determination of stability constants for zinc acetate complexes in aqueous solutions to 295 ° C. Geochimica et Cosmochimica Acta, 55, 2401-2415. GUSTAFSON, L.B. & WILLIAMS, N. 1981. Sedimenthosted stratiform deposits of copper, lead and zinc. Economic Geology, 75th Anniversary Volume, 139-178. HARRISON, W.J. & THYNE, G.D. 1992. Predictions of diagenetic reactions in the presence of organic acids. Geochimica et Cosmochimica Acta, 56, 565-586. HARTLEY, A., W.J., FLINT, S. ~¢~ TURNER, P. 1991. Analcime: a characteristic authigenic phase of Andean alluvium, northern Chile. Geological Journal, 26, 189-202. HASLAM, H.W. 8£ SANDON, P.T.S. 1991. The geochemistry of some red bed formations in the United Kingdom. British Geological Survey Technical Report WP/90/2. HELGESON, H.C., KIRKHAM,D.H. & FLOWERS, G.C. 1981. Theoretical prediction of the thermodynamic behaviour of aqueous electrolytes at high pressures and temperatures IV: calculation of activity coefficients, osmotic coefficients, and apparent molal and standard and relative partial
322
R. METCALFE ET AL.
molal properties to 5 kb and 600° C. American genesis of sediment-hosted stratiform copper Journal of Science, 281, 1241-1516. deposits. In: BOYLE, R.W., BROWN, A.C., HEM, J.D. 1989. Study and interpretation of the JEFFERSON, C.W., JOWETT, E . C . • KIRKHAM, chemical characteristics of natural water. United R.V. (eds) Sediment-hosted Stratiform Copper States Geological Survey Water-Supply Paper Deposits. Geological Association of Canada 2254. Special Papers, 36, 3-38. HENNET, R.J.C., CRERAR,D.A. & SCHWARTZ,J. 1988. KRUMGALZ, B. 1980. Salt effect on the pH of Organic complexes in hydrothermal systems. hypersaline solutions. In: NISSENBAUM,A. (ed.) Economic Geology, 83,742-764. Hypersaline brines and evaporitic environments. HOFMANN, B.A. 1990. Reduction spheroids from Developments in Sedimentology, 28, Elsevier, northern Switzerland: mineralogy, geochemistry Amsterdam. 73-82. and genetic models. Chemical Geology, 81,55-81. LANGMUIR, D. 1979. Uranium geochemistry. In: 1991. Mineralogy and geochemistry of reduction KIMBERLEY, M.M. (ed.) Shortcourse in uranium spheroids in red beds. Mineralogy and Petrology, deposits: their mineralogy and origin. Mineralogi44,107-124. cal Association of Canada Short Course Hand1992. Isolated reduction phenomena in red beds: books, 3, 17-29. a result of porewater radiolysis? In: KHARAKA, LEDER, F. & PARK, W.C. 1986. Porosity reduction in Y.K. & MAEST, A.S. (eds) Water-Rock Intersandstone by quartz overgrowth. American Asaction, 7. Balkema, Rotterdam, 503-506. sociation of Petroleum Geologists Bulletin, 70, HOLM, T.R. & CURTISS, C.D. 1990. Copper com1713-1728. plexation by natural organic matter in ground- LUNDEGARD,P.D. & LAND,L.S. 1986. Carbon dioxide water. In: MELCHIOR,D.C. & BASSETT,R.L. (eds) and organic acids: their role in porosity enhanceChemical modelling in aqueous systems If. Ameriment and cementation, Palaeogene of the Texas can Chemical Society, Washington DC, 509-518. Gulf Coast. In: GAUTIER, D.L. (ed.) Roles of HOLMES, L., CHAMBERS,A.D., IXER, R.A., TURNER, Organic matter in sediment diagenesis. Society of P. & VAUGHAN,D.J. 1983. Diagenetic processes Economic Paleontologists and Mineralogists and the mineralization in the Triassic of central Special Publications, 38, 129-147. England. Mineralium Deposita, 18,365-377. MANNING, D.A.C. 1986. Assessment of the role of HOUSEKNECHT, D.W. 1987. Intergranular pressure organic matter in ore transport processes in solution in four quartzose sandstones. Journal of low-temperature base-metal systems. TransSedimentary Petrology, 58,228-246. actions of the Institution of Mining and MetalHUANG, W-L. & LONGO, J.M. 1992. The effect of lurgy, 95, B 195-B200. organics on feldspar dissolution and the develop- McBRIDE, E.F. 1989. Quartz cement in sandstones: a ment of secondary porosity. Chemical Geology, review. Earth Science Reviews, 26, 69-112. 98,271-292. MERINO, E., MOORE, C., ORTOLEVA,P. & RIPELY, E. IXER, R.A. & VAUGHAN,D.J. 1982. The primary ore 1986. Mineral zoning in sediment-hosted coppermineralogy of the Aiderley Edge deposit, iron sulphide deposits-a quantitative kinetic apCheshire. Mineralogical Magazine, 46,485-492. proach. In: FRIEDRICH, D.H. (ed.) Geology and JOHNSON, J.W., OELKERS, E.H. & HELGESON, H.C. metallogeny of copper deposits: proceedings of the copper symposium, 27th International Geological 1991. SUPCR792: a software package for calculating the standard radial thermodynamic Congress. Special Publications of the Society for properties of minerals, gases, aqueous species, and Geology Applied to Mineral Deposits, 4, 559reactions from 1 to 5000 bars and 0° to 1000° C. 571. University of California, Berkeley. MORAD, S. & ALDAHAN, A.A. 1986. Diagenetic JOWETT, E.C. 1986. Genesis of Kupferschiefer Cu-Ag replacement of feldspars by quartz in sandstones. deposits by convective flow of Rotliegende brines Journal of Sedimentary Petrology, 57,488-493. during Triassic rifting. Economic Geology, 81, MORTON, A.C. 1984. Stability of detrital heavy 1823-1837. minerals in Tertiary sandstones from the North - 1989. Effects of continental rifting on the location Sea Basin. Clay Minerals, 19,287-308. and genesis of stratiform copper-silver deposits. NASH, J.T., GRANGER, H.C. & ADAMS, S.S. 1981. In: BOYLE, R.W., BROWN, A.C., JEFFERSON, Geology and concepts of genesis of important C.W., JowE'rr, E.C. & KIRKnAM, R.V. (eds) types of uranium deposits. Economic Geology, Sediment-hosted Stratiform Copper Deposits. 75th Anniversary Volume, 63-116. Geological Association of Canada Special NAYLOR, H., TURNER, P., VAUGHAN, D.J., BOYCE, Papers, 36, 53-66. A.J. & FALLICK, A.E. 1989. Genetic studies of - 1992. Role of organics and methane in sulfide ore red bed mineralization in the Triassic of the formation, exemplified by Kupferschiefer Cu-Ag Cheshire Basin, northwest England. Journal of deposits, Poland. Chemical Geology, 99, 51-63. the Geological Society, London, 146, 685-699. KIMBERLEY, M.M. 1979. Uranium geochemistry. In: NICKEL,E. 1973. Experimental dissolution of light and KIMBERLEY,M.M. (ed.) Short course in uranium heavy minerals in comparison with weathering deposits: their mineralogy and origin. Mineralogiand interstratal solution. Contributions to Sedical Association of Canada Short Course Handmentology, l, 1-68. books, 3,339-375. OECD NUCLEARENERGYAGENCY& INTERNATIONAL KIRKHAM, R.V. 1989. Distribution, settings, and ATOMIC ENERGY AGENCY 1986. Uranium:
CONTINENTAL RED BED DIAGENESIS
resources, production and demand. OECD, Paris. PAPELIS, C., HAYES, K. & LECKIE, J.O. 1988. HYDRA QL: A program for the computation of chemical equilibrium composition of aqueous batch systems including surface-complexation modelling of ion adsorption at the oxide~solution interface. Department of Civil Engineering, Stanford University, Stanford, California. Technical Report, 306. PARNELL, J. 1992. Hydrocarbon potential of Northern Ireland: Part III-reservoir potential of the PermoTriassic. Journal of Petroleum Geology, 15, 51-70. PrrzER, K.S. 1973. Thermodynamics of electrolytes I: Theoretical basis and general equations. Journal of Physical Chemistry, 77,268-277. ROEDDER, E. 1984. Fluid inclusions. Mineralogical Society of America, Reviews in Mineralogy, 12. ROSE, A.W. 1976. The effect of cuprous chloride complexes in the origin of red-bed copper and related deposits. Economic Geology, 71, 10361048. - 1989. Mobility of copper and other heavy metals in sedimentary environments. In: BOYLE, R.W., BROWN, A.C., JEFFERSON,C.W., JOWETI', E.C. & K~R~HAM, R.V. (eds) Sediment-hosted stratiform copper deposits. Geological Association of Canada, Special Papers, 38, 97-110. & BIANCHI-MOSQUERA,G.C. 1985. Adsorption of Cu, Ag and other metals on Fe-oxides as a control on elemental composition of stratiform and red bed copper deposits. Geological Society of America, Abstracts with Programmes, 17,702. -&~ 1993. Adsorption of Cu, Pb, Zn, Co, Ni and Ag on goethite and hematite as a control on metal mobilisation from redbeds into stratiform copper deposits. Economic Geology, 88, 12261236. SAWLOWICZ,Z. 1991. The relationship between copper mineralisation and organic matter in the Polish Kupferschiefer. In: PAGEL, M. & LEROY, J.L. (eds) Source, Transport and Deposition of Metals. Balkema. Rotterdam, 589-592. - 1992. Primary sulphide mineralisation in Cu-Fe-S zones of Kupferschiefer, Fore-Sudetic monocline, Poland. Transactions of the Institution of Mining and Metallurgy, 101, B1-B8. SEEWALD,J.S., SEYFRIED,W.E. & THORNTON,E.C. 1990. Organic-rich sediment alteration: an experimental and theoretical study at elevated temperatures and pressures. Geochimica et Cosmochimica Acta, 55,659-669. SHOCK, E.L. 1988. Organic acid metastability in sedimentary basins. Geology, 16, 886-890. & HELGESON, H.C. 1990. Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures: Standard partial molal properties of organic species. Geochimica et Cosmochimica Acta, 54,915-945. SMITH, I.F. 1986. Mesozoic basins. In: DOWNING, R.A. & GRAY, D.A. (eds) Geothermal energy the potential in the United Kingdom. Her Majesty's Stationery Office, London, 42-83.
323
SONNENFELD, P. 1984. Brines and evaporites. Academic Press, London. STOESSELL, R.K. & P1TI?MAN,E.D. 1990. Secondary porosity revisited: the chemistry of feldspar dissolution by carboxylic acids and anions. American Association of Petroleum Geologists Bulletin, 74, 1795-1805. STRONG, G.E. & MILODOWSKI,A.E. 1987. Aspects of the diagenesis of the Sherwood sandstones of the Wessex Basin and their influence on reservoir characteristics. In: MARSHALL,J. (ed.) Diagenesis of Sedimentary Sequences. Geological Society, London, Special Publications, 36,325-337. STUMM, W. • MORGAN, 1981. Aquatic chemistry. John Wiley & Sons, New York. SURDAM, R.C. & SHEPPARD, R.A. 1978. Zeolites in alkaline-lake deposits. In: SAND, L.B. & MUMIrrON, F.A. (eds) Natural zeolites: Occurrence, properties, use. Pergammon, New York, 145-174. --, Boese, S.W. & CROSSLY, L.J. 1984. The chemistry of secondary porosity. In: MCDONALD, D.A. & SURDAM,R.C. (eds) Clastic Diagenesis. American Association of Petroleum Geologists Memoirs, 37, 127-149. --, CROSSLY, L.J., HAGEN, E.S. 8z HEASLER, H.P. 1989. Organic-inorganic interactions and sandstone diagenesis. American Association of Petroleum Geologists Bulletin, 73, 1-23. SVERJENSKY,D.A. 1984. Oil field brines as ore-forming solutions. Economic Geology, 79, 23-37. -1987. The role of migrating oil field brines in the formation of sediment-hosted Cu-rich deposits. Economic Geology, 82, 1130--1141. - 1989. Chemical evolution of basinal brines that formed sediment-hosted Cu-Pb-Zn deposits. In: BOYLE, R.W., BROWN A.C., JEFFERSON, C.W., JOWET'r, E.C. & KmKHAM,R.V. (eds) Sedimenthosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 127134. TAYLOR, S.R. & MCLENNAN,S.M. 1985. The continental crust: its composition and evolution. Blackwell Scientific Publications, Oxford. THORNTON, E.C. & SEYFRIED,W.E. 1987. Reactivity of organic-rich sediment in seawater at 350 ° C, 500 bars: experimental and theoretical constraints for the Guaymas Basin hydrothermal system. Geochimica et Cosmochimica Acta, 51, 19972010. TURNER, P. 1980. Continental red beds. Developments in Sedimentology, 29. Elsevier, Oxford. WALKER,T.R. 1967. Formation of red beds in modern and ancient deserts. Geological Society of America Bulletin, 78,353-368. - 1974. Formation of red beds in moist tropical climates. Geological Society of America Bulletin, 85,633-638. - 1976. Diagenetic origin of continental red beds. In: FAERY, H. (ed.) The continental Permian in central, west and south Europe. Reidel Publishing Co., Dordrecht, 240-282. - 1989. Application of diagenetic alterations in redbeds to the origin of copper in stratiform
324
R. METCALFE ET AL.
copper deposits. In: BOYLE,R.W., BROWN,A.C., JEFFERSON, C.W., JOWETT, E.C. & KIRKHAM, R.V. (eds) Sediment-hosted Stratiform Copper Deposits. Geological Association of Canada Special Papers, 36, 85-96. & HONEA, R.M. 1969. Iron content of modern deposits in the Sonoran desert: a contribution to the origin of red beds. Geological Society of America Bulletin, 80,535-544. (~z RUNNELLS, D.D. 1984. Potential role of smectite-illite conversion in the formation of hematite pigment in red beds. Geological Society of America Abstracts with Programs, 16,685. , WAUGH, B. (~ CRONE, A.J. 1978. Diagenesis in first-cycle alluvium of Cenozoic Age, southwestern United States and northwestern Mexico. Geological Society of America Bulletin, 89, 19-32. WAUGH, B. 1978. Authigenic K-feldspar in British Permo-Triassic sandstones. Journal of the Geological Society, London, 135, 51-56.
WEDEPOHL,K.H. (ed.) 1978. Handbook of geochemistry. Springer-Verlag, Berlin. WOLERY, T.J. 1992a. EQ3/6 package overview and installation guide. Lawrence Livermore National Laboratory Report. 1992b. EQ3NR, Theoretical manual and user's guide. Lawrence Livermore National Laboratory Report. (~¢DAVELER,S.A. 1992. EQ6 Theoretical manual and user's guide. Lawrence Livermore National Laboratory Report. ZIELINSKI, R.A., BLOCH, S. & WALKER, T.R. 1983. The mobility and distribution of heavy metals during the formation of first cycle red beds. Economic Geology, 78, 1574-1589. & -1986. Mobility and distribution of heavy metals during the formation of first cycle red beds. In: MUMVrON, F.A. (ed.) Studies in Diagenesis. United States Geological Survey, Bulletin, 1578, 52-59. -
-
Recognition of the thermal effects of fluid flow in sedimentary basins I . R . D U D D Y 1, P . F . G R E E N 1, R . J . B R A Y 2 & K . A . H E G A R T Y
~
Geotrack International Pty Ltd 1 PO Box 4120, Melbourne University, Vic, 3052 Australia 2 30 Upper High Street, Thame, Oxfordshire O X 9 3EX, UK Abstract: Moving fluids are capable of transporting a large amount of heat over long distances in sedimentary basins, but the effects are often ignored when modelling the thermal evolution of sedimentary basins. If significant, the passage of these heated fluids through the sediment pile will leave a thermal signature which can be measured at the present-day using palaeotemperature determinants such as vitrinite reflectance and apatite fission track analysis (AFTATM). The interpretation of palaeotemperature-depth profiles, particularly the slope of a palaeotemperature-depth profile, i.e. the palaeogeothermal gradient, allows fluid-induced temperature profiles to be distinguished from those due to simple conduction of basal heat flow. Steady-state systems, typified by large-scale lateral fluid flow in foreland basins or where low-temperature hydrothermal circulation systems occur associated with intrusions and thick volcanic piles, are characterized by dog-length geothermal gradients, with high-interval gradients near the surface, above the shallowest aquifer, and lower interval gradients beneath an aquifer. It is argued that the classic occurrence of near vertical vitrinite reflectance-depth profiles observed in many ancient foreland basins can result from this mechanism. Thermal effects of transient fluid flow are most easily observed at the present-day, where they are characterized by low or negative geothermal gradients beneath aquifers in active geothermal systems, and the same criterion enables their recognition in ancient situations. Similarly, shallow-level igneous intrusion into porous and permeable sediments can produce observable thermal signatures further from the intrusion than will result from simple conduction, but which can be readily explained by movement of fluids heated by the intrusion. Complex steady-state profiles that may be difficult to distinguish from transient profiles without additional geological data may arise where multiple aquifers are separated by aquitards. The thermal consequences of fluid flow in two of these situations, a foreland basin and igneous intrusion into porous sediments, are illustrated by examples of thermal history reconstruction from the Pliocene Papua New Guinea Fold Belt and the Canning Basin of Australia, respectively.
Fluids, water and hydrocarbons, are fundamental constituents of sedimentary basins. They have the ability to transfer large amounts of heat (e.g. Smith & Chapman 1983) and may therefore influence the thermal evolution of the sediments within the basin, with consequent effects on maturation of source rocks (e.g. Person & Garven 1992) and also diagenetic alteration of reservoirs (e.g. Summer & Verosub 1989). The prime causes of large-scale heated fluid flow appear to be gravity-driven fluid flow (e.g. Bethke 1986; Garven 1989) and the interaction of igneous intrusions with water-saturated sediments (e.g. Einsele et al. 1980; Summer & Verosub 1989). These mechanisms seem to offer the best ways of maintaining the temperatures and moving fluids over geological time scales in excess of a million years, which are necessary for
the formation of large ore deposits and hydrocarbon accumulations and to produce a steadystate perturbation of crustal temperature profiles. Migration of ore fluids is not specifically discussed in this paper, but the methodology of thermal history reconstruction described here has been applied to Pb-Zn Mississippi Valleytype and other ore deposits (e.g. Arne 1992; Arne et al. 1991). Person & Garven (1992), in a study of the Rhine graben, modelled the influence of heated fluids on hydrocarbon source rock maturation patterns and concluded that a regional topographically driven ground water flow system could adequately explain the present-day heatflow data and the measured thermal maturation data. In a series of papers, Summer & Verosub (e.g.
From PARNELL,J. (ed.), 1994, Geofluids: Origin, Migrationand Evolution of Fluids in Sedimentary Basins, Geological Society Special Publication No. 78, 325-345.
325
326
I.R. DUDDY ET AL.
1987, 1989, 1992) have documented widespread thermal effects attributed to fluid flow in sediments under thick Tertiary volcanic sequences in Oregon. They also recognized the importance of vertical maturation profiles, discussed further in this paper, in the interpretation of a fluid flow mechanism for heating and the mistakes in overburden estimation that would result if lateral heat transfer was not recognised. Differences in vitrinite reflectance profiles have also been recorded between ground water recharge and discharge areas in the Uinta basin by Wilier & Chapman (1987). Despite the importance of fluids in heat transfer, the most common approach to modelling the thermal development of sedimentary basins for maturation assessment assumes basal heat flow as the sole source of heat. Overlooking the thermal consequences of lateral fluid flow can lead to gross inaccuracies in the assessment of basal heat flow and therefore in the reconstruction of the thermal, burial, structural and hydrocarbon generation histories of a region. We propose that the key to recognizing the thermal effects of fluid flow starts with reconstruction of the thermal history in the preserved section at a site, using a combination of techniques, the principal among them being apatite fission track analysis (AFTA TM) and vitrinite reflectance (VR). Bray et al. (1992), in a study of the UK East Midlands Shelf, have shown that determination of the palaeogeothermal gradient at the time of maximum palaeotemperatures using AFTA and VR data can provide insight into the cause of heating and subsequent cooling. For example, if heating was due solely to a period of high heat flow, the geothermal gradient at the time of maximum temperature would be higher than the present-day gradient, whereas if heating was caused solely by deep burial followed by uplift and erosion with no change in thermal gradient, the palaeogeothermal gradient would be the same as the present gradient. In this paper, we extend the interpretation of palaeogeothermal gradients to the identification of other causes of heating, specifically heating due to steady-state and transient fluid flow, illustrated with examples from two hydrocarbon provinces: the Papuan Fold Belt and the Canning Basin, Western Australia.
Thermal history reconstruction The basis of assessing the transfer of heat by the passage of fluids in a basin is reconstruction of the thermal history framework. Integration of palaeogeothermal gradients with the timing of
cooling from maximum palaeotemperatures allows reconstruction of the major features of the thermal history. Our approach to thermal history reconstruction has been described by Duddy et al. (1991), and involves three principal steps: (1) determination of maximum palaeotemperatures; (2) measurement of the time of cooling from maximum palaeotemperatures; (3) determination of palaeogeothermal gradient at the time of maximum palaeotemperatures.
Determination o f m a x i m u m palaeotemperatures Vitrinite reflectance and AFTA are the two prime techniques which can provide quantitative estimates of maximum palaeotemperature for use in thermal history reconstruction. Vitrinite reflectance measurement. The main technique for palaeotemperature measurement is provided by determination of the reflectance of the coal maceral vitrinite. The approach of Cook (1989) is used as briefly described below. Indigenous vitrinite identification is made primarily on textural grounds rather than by arbitrary analysis of histograms disconnected from the measurement process, as this allows an independent assessment to be made of the 'in situ' population from 'caved' and 're-worked' vitrinite populations. Ro(max) data is collected for both coal and dispersed organic matter (DOM). The determination of Ro(max) is more exacting than Ro(random) measurements for DOM and requires an accurately centred rotating stage, but it has the key advantage that pairs of measurements at the maximum reflectance positions (180 ° opposed) allows the surface flatness to be assessed, a key factor which may affect the quality of VR determinations. Paired Ro(max) readings which differ by more than ___5% (relative) usually indicate excessive surface relief or tilting; such readings then being rejected and the problem corrected. Simultaneous observation of the fluorescence characteristics of liptinite macerals in a VR sample significantly aids assessment of the indigenous level of maturity and hence the maximum palaeotemperature experienced by the sample. Interpretation of VR data in terms of maximum palaeotemperature. In this study, vitrinite reflectance data are interpreted on the basis of the distributed activation energy model describing
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS the evolution of VR with temperature and time, described by Burnham & Sweeney (1989). In a considerable number of wells from around the world, in a variety of settings in which A F T A has been used independently to constrain the thermal history (e.g. Bray etal. 1992), we have found that the Burnham & Sweeney (1989) model gives good agreement between predicted and observed VR data (unpublished results). As in the case of fission track annealing, it is clear from the chemical kinetic description embodied in equation 2 of Burnham & Sweeney (1989) that temperature is more important than time in controlling the increase of vitrinite reflectance.
AFTA measurement. A F T A is a kinetic method of thermal history analysis applicable to both sediments and basement rocks (e.g. Green et al. 1989a). It provides an estimate of the maximum palaeotemperature to which a rock was subjected in the range up to c. l l 0 ° C for typical geological heating rates (e.g. Green et al. 1989b). Interpretation of A F T A data from a sediment sample requires knowledge of both the present temperature (equilibrium temperature) at which the sample resided and its stratigraphic age. Comparisons between the measured confined track length distribution and present temperature, and the fission track age and the stratigraphic age enable assessment of whether a particular sample has been hotter in the past than it is at the present day, and if so, what the maximum palaeotemperature was. As discussed further below, A F T A also provides a direct estimate of when cooling from the higher palaeotemperature took place. A F T A involves measurement of fission tracks in detrital apatite grains. Fission tracks are linear zones of radiation damage within the apatite crystal lattice, created by spontaneous fission of uranium atoms, usually present at the ppm level. Several points are important in the data collection. For example, the external detector method should be used as it enables the determination of fission track ages on individual apatite grains. The lengths of confined tracks contain much more information on thermal history than projected lengths (e.g. Laslett et al. 1994) and are preferred. Further details of the collection of A F T A data are provided in Green etal. (1989a). Interpretation of AFTA data in terms of maximum palaeotemperature. Fission tracks in apatite form with a narrow distribution of lengths (around c. 16 ~m long), but once formed fission tracks begin to shorten because of the gradual repair of the radiation damage which
327
constitutes the unetched tracks. In effect, the tracks shrink from each end in a process which is known as fission track 'annealing'. The final length of each individual track is essentially determined by the maximum temperature which that track has experienced. A time difference of an order of magnitude produces a change in fission track parameters that is equivalent to a temperature change of only c. 10° C, so temperature is by far the dominant factor in determining the final fission track parameters. As temperature increases, all existing tracks shorten to a length determined by the prevailing temperature, regardless of when they were formed. If the temperature decreases, all tracks formed prior to the thermal maximum are 'frozen' at the degree of length reduction they attained at that time. Thus the length of each track can be thought of as a maximum-reading thermometer, recording the maximum temperature to which it has been subjected. At temperatures greater than c. l l 0 ° C all fission tracks are totally annealed, so that in these cases AFTA provides only a minimum estimate of maximum palaeotemperature. The maximum palaeotemperature to which a sediment has been subjected is determined using a forward modelling approach based on the quantitative description of fission track annealing derived by Laslett et al. (1987), extended to the range of apatite compositions found in sediments (unpublished data). A 'default' thermal history based on the preserved stratigraphy and the present-day temperature of the sample is used as the basis for this forward modelling, but with the addition of episodes of elevated palaeotemperatures as required to explain the data. A F F A parameters are modelled iteratively through successive thermal history scenarios in order to identify thermal histories that can account for observed parameters. The maximum palaeotemperature is adjusted until it is sufficient to account for the degree of fission track age and track length reduction shown by those tracks formed prior to the onset of cooling. Palaeotemperatures between c. 20 and l l 0 ° C can be estimated to a precision of + 10°C at lower values and + 5 ° C at values between c. 70 and 110 ° C (Green et al. 1989b).
Measurement o f the time o f cooling f r o m m a x i m u m palaeotemperatures using apatite fission track analysis AFTA offers the only direct radiometric method for quantitative determination of the time of cooling from maximum palaeotemperature up
328
I.R. DUDDY ET AL.
to c. 110° C for use in thermal history reconstruction in sedimentary basins. Such timing information is critical in determining the relative importance of multiple unconformities in a sequence and evaluating the time of oil generation and migration in relation to trap formation.
Interpretation of time of cooling from A FTA data. Since new fission tracks are produced continually through time, different tracks experience different proportions of the thermal history. Therefore the distribution of track lengths observed at the present day reflects the variation of temperature through time, while the number of tracks measures the total time over which tracks have been accumulating. The most direct information on the time of cooling of a sediment is provided by the AFTA age parameters of those samples in which the fission tracks were totally annealed after deposition. In this case only those tracks formed after cooling will be present, and if cooling occurred rapidly to temperatures less than c. 50°C, the measured fission track age will closely approximate the time of cooling. Typically the time of cooling can be estimated with an accuracy of c. + 10% for this style of thermal history. If the cooling was more protracted and/or the present temperature is greater than c. 50°C, the measured fission track age will underestimate the time of cooling, but this can be compensated for by allowance for the measured track length reduction. In those samples in which fission tracks were not totally annealed after deposition (i.e. those subjected to temperatures less than 110° C), assessment of the relative proportions of short and long tracks in the track length distribution allows the time of cooling to be determined, although with increasingly lower precision with decreasing maximum palaeotemperature. In practice, palaeotemperatures and time of cooling are estimated at the same time in an iterative process by forward modelling the fission track age and length parameters using the approach described in a series of papers (Green etal. 1986, 1989a; Laslett etal. 1987; Duddyetal. 1988). Examples of the application of the technique are given in Bray et al. (1992), Green etal. (1989b) and Miller & Duddy (1989). Integration of AFTA and VR data. Duddy et al. (1991) showed that if the Burnham & Sweeney (1989) distributed activation energy model for vitrinite reflectance is expressed in the form of an Arrhenius plot (a plot of the logarithm of time versus inverse absolute temperature), the slopes
of lines defining contours of equal vitrinite reflectance in such a plot are very similar to those describing the Laslett et al. (1987) kinetic description of annealing of fission tracks in Durango apatite. This feature of the two quite independent approaches to palaeotemperature determination means that for a particular sample, a given degree of fission track annealing in apatite of Durango composition will be associated with the same value of vitrinite reflectance, regardless of the heating rate experienced by a sample. Thus, palaeotemperature estimates based on either A F T A or VR data sets should be equivalent, regardless of the duration of heating. One practical consequence of this relationship between AFFA and VR is that, for example, a VR value of 0.7% is associated with total annealing of all fission tracks in apatite of Durango composition, and total annealing of all fission tracks in apatites of more chlorine-rich composition is accomplished between VR values of 0.7 and c. 0.9%. Furthermore, because vitrinite reflectance continues to increase progressively with increasing temperature, VR data allow direct estimation of maximum palaeotemperatures in the range where fission tracks in apatite are totally annealed (generally above c. l l0°C) and where therefore AFTA only provides minimum estimates. Maximum palaeotemperature estimates based on vitrinite reflectance data from a well in which most AFTA samples were totally annealed will allow constraints on the palaeogeothermal gradient which would not be possible from A F r A alone. In such cases, the AFTA data should allow tight constraints to be placed on the time of cooling and the cooling history, since AFTA parameters will be dominated by the effects of tracks formed after cooling from maximum palaeotemperatures. Even in situations where AFTA samples were not totally annealed, integration of AFTA and VR can allow palaeotemperature control over a greater range of depth, e.g. by combining AFTA from sand-dominated units with VR from other parts of the section, thereby providing tighter constraint on the palaeogeothermal gradient.
Determination o f palaeogeothermal gradient at the time o f m a x i m u m palaeotemperatures Estimation of palaeogeothermal gradient has been discussed by Duddy et al. (1991) and Bray et al. (1992) and proceeds from calculation of the slope of a plot of maximum palaeotemperatures
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS Temperature (*C) 0
li~'c 20
40
60
80
100
120
140
\='!,
%,
t
x
.\
E
J Pa]eogcotherma]
/ , . gradi=,-3,.
-..
•-..
~ . LS O> E3
d' Prcsmat-day / -
gradient -33"CAm
"
.,. -
, ..,
\
i ... \
'it
'.,,
'.,,, o
O V---I •
180
TYPICAL-1
~
\
160
¶•
Corrected BEI" tmasurement P a l e o t ~ t u r e data Range estimate
TD 2600 m
Point estimate Minimum estimate
329
Estimation of the amount of eroded section and heat flow. Calculating the amount of eroded section from thermal data requires either detailed knowledge of the lithologies of the removed section, or assumptions about either the thermal conductivity and heat flow variation or the palaeogeothermal gradient through the missing section. In the absence of this information, and if the palaeogeothermal gradient in the preserved section is in the normal crustal range, then a straight line projection of the palaeogeothermal gradient to an appropriate palaeo-surface temperature value will provide a reasonable estimate of the amount of eroded section (Fig. 2). This procedure contains the implicit assumptions that the lithologies in the removed section were similar to those in the preserved section (i.e. large thicknesses of lithologies with extremes of thermal conductivity like salt or coal were not formerly present), and importantly
[ Maximum estimate
Fig. I. Palaeotemperature profile constructed using observed VR and AFTA data in a typical well section which has cooled from maximum palaeotemperatures at some time since the Carboniferous. In a real case, AFTA can be used to determine the time of cooling. determined from AFTA and/or VR data in a suite of samples in a vertical depth sequence (Fig. 1). This palaeogeothermal gradient is applicable to the time immediately before the onset of cooling from maximum paleotemperatures. In addition, A F T A also allows the cooling paths of a vertical suite of samples to be determined, and thus the possibility exists of constraining a range of palaeogeothermal gradients from the time of maximum palaeotemperatures to the present day. In practice, the constraints on palaeotemperatures during the cooling history may be quite wide, so only broad, but nevertheless useful, limits may be placed on the variation of palaeogeothermal gradient during cooling. The depth range over which samples are analysed is generally the biggest factor controlling the accuracy with which the palaeogeothermal gradient can be constrained. Palaeotemperature estimates over more than 1 km, preferably 2 or 3 km, will give the best constraints on geothermal gradient, and consequently the best constraints on uplift and erosion estimates. Thus, it is important when collecting samples for thermal history reconstruction that VR samples in particular are collected from suitable lithologies from surface to total depth, not simply in target source rock horizons.
P~loo-sur~ce
temperature = 15*C •. : . 4 - - E r o s i o n e s t i m a t e = 1.2 km
g
~
I I I I
0.5
C,m,,~l 0
¢,2o,.
4o.
- Temperature (°C) "~o_ 8o, loo, 12o 14o 16o 18o '--I
-
I
I
i,
I
I
TYPICAL-I
\
0.5
~.~
. PaJeogeothezmal
1.0 0
A o',
-
1.5
.',
-\
,\
D gradient 2.0 " -33°C/kin
-
~O
O
2.5 O I----4 •
• \O
CorrectedB H T m ~ e n t Palenlcmperamre data Range ~timate Point estimate Minimum estimate
', Maximum estimate
Fig. 2. Method of estimation of the amount of uplift and erosion using palaeotemperature estimates from mixed data. This construction is only appropriate when the relevant unconformity is at the present ground surface.
330
I.R. DUDDY E T A L .
that the only source of heat is basal heat-flow. The latter assumption is critical in regions where fluid flow might be expected and may be the prime reason for grossly inaccurate erosion estimates as discussed further below. The slope of the trend of Ro% with depth (Ro%/100 m) is not equivalent to the palaeogeothermal gradient in the section at the time of maximum palaeotemperatures. Extrapolation of a log R o % / l O O m gradient to estimate the amount of section missing through uplift and erosion was suggested by Dow (1977). However, this can only be done when the kinetic relationship between Ro% and temperature and time is taken into account (as, for example, in extrapolated palaeotemperature profiles). Thus, the simple projection of a log Ro%/100 m gradient to a depositional Ro% value (usually somewhere
Thrust-Fold Belt
Foreland Platform
RECHARGE
• Well R
................ "":'-.:'.--'---.:"""...~-"-.'-.'-~"
isotherrn'~
DISCHARGE
n~i.s
~
STEADY-STATE FLUID FLOW Fig. 3. Schematic illustration of fluid movement in an active foreland basin. Ingress of cold fluid in the elevated fold/thrust belt results in gravity-driven flow discharging heated fluids on the foreland platform (after Majorowicz et al. 1985).
Heated water discharge area
Cold water recharge area
Temperature
Temperature
Q., O.
,.C; O.
c~
E)
Heat flow
Heat flow
e-
121
o
Fig. 4. Theoretical temperature and heat flow profiles for two wells located in the recharge and discharge areas on Fig. 3 (after Majorowicz et al. 1985).
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS between 0.15 and 0.25%) will not give an accurate estimate of the amount of missing section, even in the case where the relevant unconformity is at the present-day ground surface, as the slope of log Ro% with depth is not constant with increasing Ro%, even for a constant geothermal gradient. The conversion of Ro% values to palaeotemperatures using the Burnham & Sweeney (1989) kinetic relationship overcomes many of the difficulties associated with projection of log Ro% for erosion estimates and is the approach adopted here (a full discussion of the estimation of uplift and erosion from VR data will be presented elsewhere, but see also discussion in Katz et al. 1988). Once a palaeogeothermal gradient has been determined, it is then possible to go on to constrain palaeo-heat flow at the time of maximum temperatures. However, calculation of heat flow, while very commonly attempted, requires a detailed knowledge of both the thermal conductivity distribution in the stratigraphic section and the porosity reduction history. These latter two parameters are almost never available in an active exploration setting and depend to a large degree on the magnitude of the eroded section which is to be determined! Therefore, the palaeogeothermal gradient data is preferred as a more direct measure for thermal history assessment.
Fluid flow in foreland basins The expected variation in surface heat flow and temperature profiles across an active foreland basin derived by Hitchon (1984) for the Alberta Basin, Canada, and illustrated by Majorowicz et al. (1985), is illustrated in modified form in Fig. 3. The effect of gravity-driven regional cold water recharge at the advancing thrust front is to lower the temperature in the upper part of the sediment pile, with the water transporting this deeper into the pile and then laterally and upwards as the water moves onto the foreland platform and discharges. Overall, this depresses the isotherms in the recharge area and raises them over the foreland platform (Fig. 3). Thus, at locations in the former recharge area and where sedimentation is greatest, the geothermal gradient in the upper part of the section is lowered. Meanwhile, on the foreland platform where fluids are moving laterally and upwards, the geothermal gradient is raised (Fig. 4). This perturbation of the temperature profile by the lateral transfer of heat by fluid gives low measured surface heat flow in the recharge area and a high surface heat flow in the discharge
331
Temperature
.
1 km
j
Geothermal gradient higher than
I
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Temperatureincrea
\ \ \
\ \ \ \ \ \ Geothermal gradient equal to background value but shifted to higher temperatures
Fig. 5. Expected stead-state temperature profile developed in response to hot fluid flow in a shallow confined aquifer. area, even though basal heat flow across the region may be uniform (Fig. 4). In detail, complex vertical variation in palaeogeothermal gradient may develop in the region where the heated fluids are confined in distinct aquifer horizons on the passage to discharge. The expected steady-state temperature associated with a single aquifer horizon (Ziagos & Blackwell 1986) is illustrated in Fig. 5, with an elevated geothermal gradient above the aquifer and a background geothermal gradient below the aquifer but shifted to higher temperatures. Confinement of flows of similar temperature fluids to distinct aquifer horizons separated by a considerable vertical thickness of shale may result in a zero thermal gradient through the shale section (Fig. 5b). Where local cold recharge interacts with regional hot flow, it may be possible to have hotter fluids in a shallow aquifer than a deeper one, resulting in negative geothermal gradients through the intervening aquitard section (Fig. 7).
Preservation potential of fluid flow thermal signatures A thermal signature will be observed only in a preserved stratigraphic section if the present temperature at which a section resides is lower than the maximum palaeotemperature reached during the fluid flow regime.
332
I.R. DUDDY E T A L .
Temperature
Therefore, once the active development of a foreland basin ceases, the ability to observe the thermal signature of the regional fluid flow regime at the present day varies considerably depending on the location within the former basin. Recharge area. The probability of observing the low palaeogeothermal gradients due to fluid flow in the upper part of the sediment pile in the recharge area (Fig. 3) is low for two reasons. Firstly, uplift and erosion in the part of the basin close to the advancing thrust complex may be high, removing the section which experienced the low geothermal gradient. Secondly, even without uplift and erosion, cessation of cold water ingress will result in a rapid rise in geothermal gradient to the 'normal' background value, as the cold water will no longer be suppressing the transmission of the basal heat flux to the surface. Thus, present-day observations of palaeotemperature profiles in wells at such locations will reveal a normal background geothermal gradient corresponding to the present thermal regime, and the previous influence of fluid flow on the thermal gradient will have been erased. To observe palaeo-fluid flow thermal signatures in a recharge area requires rapid cooling of the section while the fluid flow system is still active, without subsequent reheating. This will require the magnitude of cooling due to uplift and erosion to be greater than the magnitude of rapid heating which must follow cessation of fluid movement. The low fluid flow-induced palaeogeothermal gradient itself makes this situation difficult to achieve. For example, for a palaeogeothermal gradient of 10°C km -~, 3 km of uplift and erosion will only cool the section by 30°C, while return to a typical background gradient of, for example, 30°C km -I will result in preservation of the evidence of maximum palaeotemperatures associated with fluid flow in only the upper 1.5 km of section. However, in practice, the maximum temperatures in this upper 1.5 km will have been between only 30 and 45°C, a temperature range where thermal indicators are least sensitive (vitrinite reflectance would be less than 0.3% over the entire interval). Thus, the true thermal history is unlikely to be resolved in such a case and a far less complicated history would be interpreted from measured data (probably in terms of maximum temperatures at the present day throughout the section and therefore no uplift and erosion). Discharge area. Observation of the thermal
Geothermalgradient higher than background value
~'~~~~f \ lkm
J= C}
\ [i!i!i~i:upper ,!i aquifer horizon: fluid at n*c~i::ilii~i:i!~iii!!!ii::!.!i !i!!:il;i~i~i::~.~:i il \ \ Geothermal gradient is zero
\
--~-
\
\ \
Geothermalgradient equal to background value bu1 shifted to higher temperatures Fig. 6. Temperature profile around two aquifers carrying fluid at the same temperature separated by a thick aquitard.
signature of fluid flow in the discharge area is potentially more straightforward. In this region, cessation of fluid flow will result in the cooling of the section as the isotherms rapidly decline through the section. Without uplift and erosion, this will preserve a VR palaeotemperature profile in the section characteristic of the fluid flow regime; in the simple case illustrated in Fig. 5, a high geothermal gradient above the aquifer horizon, and normal geothermal gradient below. Application of A F F A in the section would additionally reveal the time of cooling from maximum palaeotemperatures, which would be the time of cessation of fluid flow. Uplift and erosion following or coincident with the cessation of fluid flow would enhance the observed magnitude of cooling, but would make the elucidation of the fluid flow thermal signature more difficult, as in the case illustrated, where the characteristic high-interval gradient shallow in the section would be completely removed as a result of only 1 km of uplift and erosion. The remaining section would then only preserve a record of the normal geothermal gradient below the aquifer. Reconstruction of the burial history from such thermal data will result in the over-estimation of the amount of removed section. The more complex palaeotemperature profiles which may develop in regions of discharge, with zero or negative geothermal gradients (Figs
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS
333
D i s c h a r g e area
Temperature
Maximum burial . j . Geothermal gradient higher than
Temperature (*C) 10
50
k\
1 km
100
I
150
l
\~
I
Interval gradient
.;:::.Upper aquifer h0riz0ni f'iuidatn*C:~ :~i: ~. i::, :. :.: : Geothermal gradien
-
NN maintained at 80"C
a
\ \
i Lower aquifer horizon: fluid at < n * C
: :
\
~
Interval
\
gradient
\ ~*C/km
: :
~2. r~
r~=,o
-
no.
Geothermal gradient equal to background value bul shifted to higher temperatures
n.i~ 300c~,.
3-
\
-
\
100"C
Fig. 7. Temperature profile around two aquifers with hotter fluid flowing in the shallower aquifer resulting a negative palaeogeothermai gradient in the intervening aquitard.
6 and 7), may provide the best clue to a fluid flow origin, as such profiles will not be produced by variation in thermal conductivities in typical classic lithologies. In all of these cases, the measurement of Ro(max) and AFTA data allows accurate determination of palaeotemperatures and time of cooling in the preserved section, which will allow the time of source rock maturation to be quantitatively assessed. However, reconstruction of the burial history in fluid-controlled regimes, particularly the amount of section eroded at unconformities, is difficult and may be impossible, as discussed further below.
\
-
140"C
Fig. 8. Steady-state temperature profile in the discharge area developed at maximum burial and controlled by a confined aquifer at I km depth. to a palaeo-surface temperature may result in unrealistically large estimates of removed section. This is illustrated in Figs 8 to 11. Figure 8 shows a theoretical palaeotemperature profile developed in a discharge area well at the time of maximum burial with an aquifer at 1 km depth carrying fluid supplied from deeper in basin at
D i s c h a r g e area Post-uplift
Paleotemperature 10 I
50 I
100 I
(*C) 150 I
Estimates of uplift and erosion in fluid flow regimes Estimating the magnitude of uplift and erosion using thermal methods (i.e., VR, AFTA, Tmax, biomarkers, CAI etc.) in fluid-flow dominated regimes is difficult and dangerous, as the palaeotemperature estimates made in the preserved section, no matter how accurate, cannot be unequivocally transferred to the removed section assuming a constant heat flow. Where low palaeogeothermal gradients (i.e. low apparent palaeo-heat flow) are observed at the present day, simple projection of these gradients
Q.
Pm~nt-day 2 ~ ge.othcrmal ~ gradient , e 30*C/kin
# % Palengenthermal gradient 30*C/kin
Fig. 9. Palaeotemperature profile observed in a discharge area well following uplift and erosion of 1 km of section.
334
I.R. DUDDY E T A L . Time Me)
"Reconstructed" paleotemperature profile
r~
o
Eroded section estimate = 2.3 km
,
l I
~
Exu'apoladonof the
/ I \ ,./,.==,,=dp.l.og-,otho,~ | J \,I" gradient through the / I \ ".,i,.~g.=.o,,',ol~c 1.0
Present-day
groundsurface
| I /~10 ,~_/y ~
~ 2-
Temperature(°C) 5.0 1 \ : ) 0 l ~1
3 150 I
4 A
Time (Ma) J= .,... t-s t~
..-.. 1
J
\\ \ ~
'
] ldlomema= of sectionuplifted :derodcd
Paleogcothe'a~al gradicm 30°C.Jkm
Fig. 10. Reconstruction of eroded section in the discharge area assuming that the geothermal gradient measured in the preserved section extended through the missing section. The calculated value eroded section value overestimates the actual value by a factor of around 2.3, with consequent apparent increase in burial rate in the period prior to cooling.
80 ° C. The palaeotemperature profile recorded in this well at the present day and following removal of I km of section is shown in Fig. 9. If it is assumed that the observed palaeotemperature result from simple deeper burial, implying a constant basal heat flow as the only source of heat, then c. 2.3 km of eroded section would be estimated (Fig. 10). This overestimation of removed section has profound effects on the reconstruction of the burial history, not only in terms of overestimation of the magnitude removed section but also in the overestimation of the burial and heating rates (Fig. 11). Therefore, it is essential that thermal data be used in conjunction with seismic sections or other information which can give a more direct estimate of depth of burial, such as shale density or sonic response (e.g., Magara 1976), chalk porosity (e.g. Hillis 1991). Even so, these techniques may also involve considerable uncertainties, particularly in recognition of the absolute baseline datum. Stratigraphic reconstructions of removed thickness should also be
=.
~2 Actual burial history
3 Fig. 11. Comparison of burial history in the discharge area reconstructed with 2.3 km of missing section assuming no lateral heating due to fluid flow, and the actual burial history with only 1 km of missing section. treated with caution, as they will commonly underestimate the true magnitude, particularly in foreland basins where the region of maximum erosion may correspond with the region of maximum deposition (e.g. Beaumont 1981). In the Alberta Basin, for instance, a number of workers have made estimates of uplift and erosion using thermal data (e.g. Hacquebard 1977; Hitchon 1984; Nurkowski 1984; Beaumont et el. 1985; England & Bustin 1986; Kalkreuth & McMechan 1988; Majorowicz et el. 1990), but no real account has been taken of either the likely non-linearity of palaeogeothermal gradients through the removed section, or the time at which the measured palaeogeothermal gradients are applicable. The estimates of Hacquebard (1977) and Nurkowski (1984) based on equilibrium moisture content of near surface coals were claimed to be non-thermal.
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS However, the good correlation between moisture content, rank and vitrinite reflectance that they report suggests that moisture content may predominantly reflect maximum palaeotemperature rather than simple burial alone. England & Bustin (1986) used vitrinite reflectance data to estimate palaeogeothermal gradients in the Alberta Basin, obtaining values in the range 8-15°C km -1, and from these, estimated amounts of uplift and erosion between c. 5 and 9 km. Majorowicz et al. (1990) have commented on problems with these geothermal gradients on other grounds, but it is clear that extrapolation of low palaeogeothermal gradients from a preserved section, however determined, will result in erosion estimates which are much too high, when lateral heat flow resulting from fluid movement such as expected in the Alberta Basin (e.g. Hitchon 1984; Garven 1989) are not considered. Magara (1976) used shale compaction data to estimate uplift and erosion across the Alberta Basin, and his values are systematically lower than those estimates from VR data, ranging between zero and c. 1400m. He suggested essentially no uplift and erosion in the east, while subsequent studies using thermal data in the same general region suggest minimum values near 1 km (e.g. Nurkowski 1984). It is possible that Magara's (1976) sonic velocity estimates suffer from a baseline datum problem as documented elsewhere (e.g. Bray et al. 1992), although no published estimates of uplift and erosion in the Alberta Basin appear firmly based. AFTA provides a method for addressing part of the problem, by constraining the time at which the palaeogeothermal gradient was operating. This provides a fixed point in time where the thermal and possible burial histories can be evaluated against regional stratigraphic data, time of thrusting, shale density compaction data etc. Limited fission track data have been collected in the Alberta Basin, but these data have not been used either to estimate independently the amounts of eroded section, nor the time of cooling from maximum palaeotemperatures (Issler et al. 1990; Willet & Issler 1992). A practical example of the difficulties associated with reconstruction of the burial history in a foreland basin is illustrated in the next section with data from a well in the Papuan Fold Belt.
A n example o f heated palaeo-fluid flow in a foreland basin, the Papuan Fold-belt The Papuan Fold-belt, a region of Jurassic to Tertiary sedimentation, was in a foreland basin
335
setting from the Mid-Miocene (Davies 1990; Home et al. 1990; Osborne 1990). Collision of the Australian plate with the Melanesian arc at around 10Ma culminated in formation of the Papuan fold and thrust belt and foreland basin on the northern margin of the Australia plate (Smith 1990). The thrust belt includes large basement-involved ramp anticlines, as well as what appears to be a thin-skinned type of deformation in the overlying sedimentary section, with a large number of thrusts and folds in what is now the Papuan Highlands apparently formed by re-activation of earlier extensional structures (Hill 1990). Major hydrocarbon discoveries have recently been made in this region, with most discoveries concentrated along the southwestern margin of the fold-belt trend (e.g. Hedinia, Iagifu) (Fig. 12). The general fold-belt stratigraphy is of Palaeozoic to Permian basement on which was deposited a mid- to late Triassic volcanic and greywacke sequence. Sedimentation from the Jurassic to early Cretaceous was dominated by fine-grained deposits derived from contemporaneous volcanism, with the reservoir facies of the late Jurassic to early Cretaceous Toro sandstone accumulating in a period of volcanic quiescence. Seal to the Toro sandstone is provided by thick shales and silts of the Ieru Formation produced by a return to contemporaneous volcanism through to the late Cretaceous. A major unconformity from the Cenomanian to the Late Oligocene between the Ieru Formation and the overlying Darai Limestone marks regional uplift and erosion associated with the opening of the Coral Sea (Smith 1990). Remnants of a syntectonic clastic depositional phase, the Late Miocene to early Pliocene Orubadi Formation overlies the Darai Limestone in some parts of the fold belt. The outcropping Darai Limestone is deeply karstifled in the fold belt, virtually eliminating the acquisition of seismic data (Lamerson 1990), with hydrocarbon exploration therefore relying largely on outcrop mapping. Published fission track studies (Hill & Gleadow 1989, 1990) have shown that the cooling and uplift and erosion which accompanied thrusting in the fold belt region occurred dominantly in the late Miocene to early Pliocene, between c. 6 and 4 Ma. In some places, the entire Mesozoic and Tertiary section was removed, exposing Permian granite-cored anticlines as at Muller and Kubor (Hill & Gleadow 1989). In other areas of the fold belt, only a portion of the Miocene Darai Limestone and perhaps the overlying early Miocene Orubadi clastics have been removed by Pliocene erosion.
336
I.R. DUDDY ET AL. N
8 8HIELD
FORELAND II
PAPUAN
FLY PLATFORM
I
FLY-ITRICKLAND DEPOGENTflE
FOLD-THflUBT
BELT
li HEOINA IAGIFU
(proJeolsd)
0
- *-*'i~
~"'"
. . .•. ,. • ,, ,i,, ~ A +i,
• 0"*
"
:" .......
*"
--
' " " "1"! . . . . . .
":'d,i
p
.........
o
N
N
-pLEr, IITO
II'E'II'T! _~.,,. .'I.+. .l~. .,,+. . *. . * * * , + + , , ~ * l+** * * * * , i"+ * * , ++!1+ 'Ii *~ '!I+ ++
::::::~+
I
' " "" ~" ":"
:/.++++i+++
~--6000
w Q 10000-,
"
ION MAP
0 !
**:** '" "1 -10000
lOOkm !
~
Llmoelone
~
MRIn reservoir (Toro Osndotono)
Fig. 12. Schematic cross-section showing the present-day structural relationship between the Fly platform and the Papuan fold thrust belt (modified from Osborne 1990). Note the thick shale seal of the Ieru Formation above the confined Toro Formation main reservoir.
Structural and stratigraphic studies indicate around a maximum of 1 to 2 km of missing late Miocene to early Pliocene section in those areas where about 1 km of the Darai Limestone is still preserved. In contrast, more than 3kin of Jurassic to Miocene section has been removed from the Muller and Kubor anticlines (Hill & Gleadow 1990). Thermal history reconstruction in a fold belt well. Figure 13 shows a palaeotemperature-depth plot derived from A F T A in four samples and VR data in three samples from Well A and a VR sample from a nearby outcrop of the Late Miocene to early Pliocene Orubadi Formation. The well penetrates a typical thickness of Miocene Darai Limestone unconformably overlying the Cretaceous Ieru Formation, a dominantly mudstone unit with a drilled thickness of about 1.3 km at this location. The Toro Sandstone reservoir target, typically ! 00 m thick, and the Jurassic Imburu Formation source rock sequence underlie the Ieru Formation, but were not reached in this well. The present-day geothermal gradient, derived from corrected BHT data, is 16°C km -1 and the surface temperature is 15° C. The data is Fig. 13 show that the maximum palaeotemperatures in the well determined from the A F T A and VR data are consistent, and comparison with the present-day temperatures
demonstrate that considerable cooling has occurred. The A F T A results further showed that cooling from maximum palaeotemperatures commenced within the last c. 5 Ma (unpublished A F T A interpretation). Palaeotemperatures over c. 3.0 km of section vary from c. 70°C to c. 100°C, giving a palaeogeothermal gradient before cooling between c. 5 and 10° C km -] . At face value, projection of this range of gradients to a palaeo-surface temperature of 15°C implies between c. 6 and 11 km of uplift and erosion of post-Miocene section. This amount of post-Miocene section is unquestionably unrealistic in this region, with structural (Hill 1990) and stratigraphic (e.g. Lamerson 1990) constraints suggesting a maximum of 1-2 km of section removed since that time. The thermal and geological data can be reconciled if the palaeogeothermal gradient through the section at the time of maximum palaeotemperatures was non-linear, with a high geothermal gradient in the shallow eroded section. In this situation, there is no independent method of constraining both the palaeogeothermal gradient and the amount of eroded section using the palaeotemperature data alone. A viable model for the depositional and uplift history of well A is shown in Fig. 14. This model assumes c. 1200 m of uplift and erosion in the Pliocene combined with the estimated geothermal gradient of 5°C km -1 below the
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS
337
Temperature (°C) Surface 15°C 4 0 0 _. .
60 •
80 100 120 140 tl~
•
|
I; °
;.0-
~
?Orubadi
I
Well A
I t I I
l ~
0.5-
I
3t
•
J L~t
Darai Lst Miocene
r±l
I t
J
t ].5-
1
l i
4~ z
I
-
L
l
I
| ii
° . .
E v
t-
2.0
".'.'2
\
"
r.'-:
gradient -5 to 9°C/kTn
° . ° °°°
(
a 2.5
° . .
[]
. ° . ,
"
ieru Fm Cretaceous
°°.
Present-day ~'~ gradient
-16 C/kin
3.0 =
f
4 D
°°° °°. °°. °°o .°. °.°
"
~
~"
41
PNG Fold belt
r
i
TC -3000m
O
C o r r e c t e d B H T measurement
•
Maximum palcotemperaturc from V R Maximum paleotemperaturc from AFTA at - 5 Ma
Fig. 13. Actual palaeotemperature profile constructed using AFTA and VR data from a Papuan fold belt well, well A. The data allows calculation of a palaeogeothermal gradient throughout the preserved Late Miocene to Early Cretaceous section, between c. - 5 and 9°C km -I.
Orubadi Formation. These parameters allow calculation of a geologically reasonable upper interval geothermal gradient of c. 45°C km -1 when projected to a palaeo-surface temperature of 15°C. For a range of 1-2km of removed section, the corresponding upper interval geothermal gradients would be c. 55 to 28°C km -1, and are equally plausible. Thus, combination of stratigraphic and structural data with the thermal data tends to support a model with migration of fluids heated to c. 70°C through a confined aquifer assumed to be in the vicinity of the boundary between the Orubadi and Darai Formations. The low palaeogeothermal gradient measured through the Darai and Ieru Formations suggests that fluid
movement also occurred in the Toro Formation aquifer just below TD in the well. Fluid temperatures in the Toro Formation would be expected to be a maximum of c. 100-120°C, given the palaeotemperatures derived from the Ieru Formation near total depth. Transient heated fluid flow in an upper aquifer alone is an alternative explanation for the data, as this could also give a near vertical vitrinite reflectance profile as discussed in the next section. This is thought unlikely, however, as evidence from the Toro Formation throughout the fold belt shows that it is a major fluid conduit for both water and oil, and sufficient time was available prior to cooling for steady-state conditions to be established.
338
I.R. DUDDY E T AL.
IPI
o
E Icl
M II
.~_ 1.0
-J ° m
D
E ~" 2.0
;~:..ii{
2 90°C " ~
1
100oo
- - -
Reconstru c ted history WellA 140
120
100
80
60
40
20
0
Time (Ma) Fig. 14. Probable reconstructed history in Papuan fold belt well, recognizing a likely fluid-flow effect on the observed palaeotemperature profile, and assuming only 1 km of section eroded since 5 Ma rather than c. 6-11 km calculated from the palaeotemperature data assuming that the only source of heat was basal heat flow. Finally, the present-day geothermal gradient in this well is itself low when compared with the regional values, and this may reflect the influence of present-day cold water flow through the Darai Limestone karst system. Therefore any assumption that the present-day thermal conditions can be applied to the past to estimate the magnitude of uplift and erosion is also extremely dangerous.
Thermal effects of intrusions transmitted by fluid flow Transient contact thermal effects surrounding igneous intrusive bodies in sediments are well documented. As a general guide, conductive heating effects in the country rocks are felt for a distance of about 1.5-3 times the thickness of the intrusive body (Carslaw & Jaeger 1952). Evidence for the transient nature of such conductive
heating effects is provided by vitrinite reflectance data from the country rock, which shows a rapid decrease in values away from the contact with the intrusive (e.g. Dow 1977; Raymond & Murchison 1988, 1991). These vitrinite reflectance profiles can be interpreted in terms of a positive palaeogeothermal gradient in the sediments above the intrusive and high negative palaeogeothermal gradient below the intrusion giving a bell-shaped profile as illustrated in Fig. 15. The bell-shape of the thermal profile associated with fluid flow in active geothermal systems has been modelled by Ziagos & Blackwell (1986). They have shown the time taken to move from transient to steady-state thermal profiles around a shallow aquifer carrying hot fluid is geologically short (less than 100000 years), so a negative palaeogeothermal gradient below an aquifer could be diagnostic of transient thermal effects (Fig. 16). Note that these transient
RECOGNIZING HOT FLUID FLOW IN SEDIMENTS
VR-derived palaeotemperatures from the underlying finer-grained section may allow short-term transient effects to be recognized. In this context, fluid inclusion-trapping temperatures from diagenetic phases in the aquifer sequence may be useful in demonstrating higher temperature in the aquifer than in the underlying sequence. However, in general, it is probable that only comparison of these perturbed gradients with regional gradients due solely to conduction of basal heat flow will allow a distinction between steady state and transient effects to be made.
Vitrinite reflectance
" ~ ~ i
High +ve paleogeothermal
e~ 121
:,'SJLL.'-:-2..2
~
I
aa 6.0 and -7%0 < ~13Cco2< -4%0 and a crustal component characterized by R/R, < 0.1 and ~13Cco2< - 15%orepresenting a fluid enriched in radiogenic 4He and principally organic-derived carbon (labelled a) (b) Superimposed upon this data, a small number of samples from the Ardennes and the Hunsriick-Taunus fault (labelled b) is significantly enriched in ]3C and shows the presence of a third component of crustal carbon in these groundwaters, such as might be liberated by the decarbonation of limestone.
problems of hydrocarbon migration and transport. Case studies from hydrocarbon gas fields in the N e o g e n e Pannonian Basin, Hungary, the Po Basin, Italy, and the Vienna Basin, Austria, provide the first systematic application of rare gases to the study of a hydrocarbon rich environment and are reviewed here (Ballentine 1991; Ballentine et al. 1991; Ballentine & O'Nions 1992; Elliot et al. 1993). Readers are also referred to the work of Hiyagon & Kennedy
(1992) which discusses the rare gas systematics of CH4-rich gas fields in the Alberta Basin, Canada. These studies illustrate how differently sourced rare gases are resolved using their isotopic composition, how they provide information about the extent of hydrocarbon/ groundwater interaction, constrain the mechanism of hydrocarbon transport, and provide an insight into the overall behaviour of the fluid regime in sedimentary basins.
350
C.J. BALLENTINE & R.K. O'NIONS
Vienna
( N~_~._~///~~?~..///~-"~;;~-I ~_v.~ zY//'/
v__/_~Budapest ~/~ ~
~3-----J~HAJD0SZOBOSZL6
I / , .--.,/
/~ ~--~
°
, ~
"~ .~"~'_~°~)erecske ~" E R l : : ~ - ~ S u b - B a s i n
)
,'
\ J
~-
basementisopach (kin
•
,,
S~c,e ''
%
~1~
~,~////~//////////////~/
(a)
R. Danube l
Tisza River
~::"
E
~
"---------~ - : : - ~
1000
v v v v V
•
~
o
1
(b)
o
o
2
C.-_-'~-_--
~
~
,
,
3
~
v
:,~,~,v I
1
I
0
50 km
100
4
5
V
v
~ ~ ~ ,, v ~
6
7
He, Ne & Ar IN HYDROCARBON FLUIDS A
B HAJDOSZOBOSZL6
1
-
~
"~ 3-
~
5-
/ /
/
Top Pre-Pannonian basement
(c)
__
GAS POOLS - t ' - - - -
/
_ ~xx/x~Tx
GAS POOL
~-
~"
DERECSKE / SUB-BASIN--
: : 12500km 3
FILL "~
:
~
,,o~m
:
:
._
Groundwater provenance and rare gas solubility Atmosphere-derived rare gases are present in many subsurface fluids. They are probably introduced into the subsurface dissolved in groundwater which has equilibrated with the atmosphere. The solubility of rare gases in groundwater and the information this provides about the provenance of groundwater is briefly reviewed. The relative solubility of rare gases in water is strongly temperature dependent (Crovetto et al. 1982; Smith 1985). The relative proportions of rare gases dissolved in water which has equilibrated with the atmosphere, containing a fixed rare gas composition, is therefore determined by the temperature of the water at equilibration. This is shown for atmosphere-derived Ne/Ar and Xe/Ar ratios in fresh water over the temperature range 0-25°C (Fig. 2). Waters which subsequently enter the subsurface, either buried as formation water or as part of aquifer recharge, retain this characteristic rare gas elemental abundance pattern. This property of rare gases has been used successfully in determining the palaeo-recharge temperature of old groundwater systems (e.g. Mazor 1972). Current analytical techniques enable the palaeorecharge temperature to be determined with a precision of + 0.5°C (e.g. Stute & Deak 1989). The solubility of rare gases in water also depends on salinity. While an increase in salinity does not greatly affect the relative solubility of the rare gases, the absolute solubility of the rare gases decreases with an increase in salinity (Smith & Kennedy 1983). This is illustrated in Fig. 3a, where the Ne/Ar ratio of dissolved atmosphere-derived rare gases is plotted against the Ar concentration in fresh water and sea water over a temperature range of 0-25 ° C. At any given temperature there is a marked decrease in the concentration of dissolved Ar with increasing salinity, but little change in the Ne/Ar ratio. This relationship has been used, for
351
Fig. 5. (a) The location of the Pannonian Basin is shown in relation to the areas of alpine uplift and associated basins. The expanded map shows the location of the Hajduszoboszlo gas field in relation to the basement isopachs, other oil and gas fields and the exposed pre-Pannonian basement (after Ballentine et al. 1991). (b) The SW-NE cross section shows the general features of the present day flow regimes. The shallower water is topographically driven and supported by aquifer recharge. The deeper regimes, which connect locally to the shallow flow, is saline and possibly connate. 1, Principle groundwater flow direction; 2, Quaternary; 3, Upper Pliocene; 4, Pliocene with saline pore water; 5, Upper to Middle Miocene; 6, Miocene volcanic; 7, Mesozoic or older (after Martel et al. 1989). (c) Schematic section across the Hajduszoboszlo gas field along the section A-B. This diagram emphasizes the efficiency of fluid focusing and the relative scales of transport, rather than identifies any particular fluid provenance (after Ballentine et al. 1991). example, to distinguish between formation waters originating as evaporational brines and those originating as seawater in the Palo Duro Basin, Texas (Zaikowski et al. 1987). Both groundwater palaeo-temperature and mass balance investigations require that the aqueous rare gas system has remained closed since it last equilibrated with the atmosphere. In some instances it is found that the groundwater contains more atmosphere-derived rare gases than can be accounted for by equilibrium solubility, and this is ascribed to the addition of excess air dissolved in the groundwater system (Heaton & Vogel 1981). On the other hand, if the groundwater equilibrates with another phase, such as oil or gas, after entering the subsurface, rare gases dissolved in the groundwater will partition into the oil or gas phase (e.g. Bosch & Mazor 1988). The rare gases remaining in the water will retain a characteristic rare gas abundance pattern (e.g. Zaikowski & Spangler 1990), depending on the phase equilibrated with the water, the extent of interaction between the two phases, and the temperature and salinity of the water during this equilibrium (Fig. 3b). In practice, however, it is often only the hydrocarbon phase, containing the atmospherederived rare gases, which has been sampled and is used to deduce the relationship between the hydrocarbon and any groundwater phase.
Mantle and deep crustal-derived gases within basin systems In addition to the atmosphere-derived rare gases dissolved in groundwater, it is also possible in many cases to resolve rare gas contributions from both mantle sources and radiogenic rare
352
C.J. BALLENTINE & R.K. O'NIONS
,.ofen
exposedbasementIWildendurnbach t..,,,,~.,., ~__~Joil & gas fields f-&
/
] ~;~"
t ~'"
I@"X
basement isopach L ~
i
I"
,,¢ (km)F_..=. ,,(#,/~" major faults 1 ~ ' , i'(
..,
.-,.
/
"~
v /
~,
~
~ / ~ . ~ [ [ /-'"~( F 4 / i / ~
//t
" 2.0, both radiogenic He and Ar are quantitatively released into any carrier phase present (Ballentine et al. 1993). It is not surprising, therefore, to find a genetic link within continental systems between He, Ar and CH4 production (Ballentine et al. 1993), He and N2 production (Tongish 1980; Jenden et al. 1988) and He and CO2 sources (Griesshaber et
The three case studies reviewed in this paper are based upon hydrocarbon gas fields from the Po (Elliot et al. 1993), Pannonian (Ballentine et al. 1991) and Vienna basins (Ballentine 1991; Ballentine & O'Nions 1992) (Figs 5, 6 & 7). The Pannonian and Vienna basin systems appear to have developed by extension in the early to mid-Miocene, creating a series of deep basins separated by shallower basement blocks (Royden 1988; Wessely 1988). Within the Pannonian Basin this was accompanied by surface vulcanism. While there is no evidence of this in the Vienna Basin, mantle 3He, which is associated with recent melt emplacement, is ubiquitous within fluids from both basins (Martel etal. 1989; Marty et al. 1992). The basins are filled by Neogene-Quaternary sediments which reach maximum depths of 7000m and 5000m in the Pannonian and Vienna Basins respectively (Berczi 1988; Jiricek & Tomek 1981). Groundwaters in the Miocene of both basins are highly saline and communicate only locally with more shallow systems (Wessely 1983; Erdelyi 1976; Ottlik et al. 1981). In distinct contrast, the Po basin is a subalpine loading basin with the sedimentary fill varying between 4 and 5 km at its northern border to more than 12km near the Appennines. Furthermore, typical of rapidly subsiding basins, geothermal gradients are generally in the range 12-20°C (Mattavelli & Novelli 1988). Consequently the organic maturity in the basin is similarly low, and below 0.5% on the vitrinite reflectance scale (Ro) to depths > 5 km (Mattavelli & Novelli et al. 1988; Pieri & Matavelli 1986). The Hajduszoboszlo gas field in the Pannonian Basin (Fig. 5) and the Dosso degli Angeli field in the Po Basin (Fig. 6) were used to investigate the detailed systematics of rare gases in single gas fields within very different geothermal and tectonic basin environments. Both are composed of a series of discrete gas pools at different depths within Neogene interbedded clays, shales and sandstones. Samples from six gas fields in the Vienna Basin, located in different stratigraphic
354
C.J. BALLENTINE & R.K. O'NIONS
Dosso Degli Angeli
J-"r
...... ..- ..... ..- .. .........
B
Pliocene-Quaternary
IT] . . . . . . Pre-Pliocene !$@Mesozoic
Fig. 7. (a) The location of the Po Basin shown in relation to the areas of alpine uplift and associated basins. he exp%ded map shows the location of the Dosso degli Angeli gas field in-relatibn to the basement isopachs, other oil and gas fields and the exposed pre-Neogene basement. (b) Schematic cross section from SW to NE across the Po Basin. The Dosso delgi Angeli gas field lies within the Pliocene-Quaternary sediments located above the faulted Mesozoic-Miocene basement. The gas field lies between 3200 and 3500 m depth. The thermal gradient in this region is particularly low c. 18°C km-' (after Elliot et al. 1993).
He, Ne & Ar IN HYDROCARBON FLUIDS c
i
=
Ii
i
,
i
i,
m
lOOO
355
l
"-" r: 2000
_o
¢" 3000
I
uO Re
8.000
b
•
iil~
•
•
•
me
;i1~
nile
i
I
II ..mml~ NII-~
__.i .,
- 2.04 and FNe/A~---> 2.03. Thus rare gas partitioning between an oil or water with a gas phase fractionates the He/Ar and Ne/Ar ratios by the same magnitude but would not be resolvable in the He/Ne ratios. Observed rare gas fractionation patterns in Fig. 10 are therefore consistent with liquid/gas phase partitioning. They are also consistent with the observation that both the Po and Pannonian basin gases are degassed from a water phase (Fig. 9). The Vienna Basin gases, however, do not have such a straightforward relationship with the groundwater system (Fig. 9a). It is entirely possible that they have been subject to interactions with an oil phase, either within closely associated oil deposits or with the source rocks, from which they are now separated.
358 (a)
C.J. BALLENTINE & R.K. O'NIONS 6
,
1
'
I
'
I
'
1
i
A ///
/I
(b)
i
2
A///I / / //T
T A
~'~ 4
////
,it
1
Z
/
O4
a}
=' T
"1,+ 2 LL
T T -T-
=
TT
1
2
3
4
5
6
i
L
v
I
ii iI
iI
0
-r
-r"
I:) I
0
4----
---
7
F (2°Ne/36Ar)atm
0
L
l
I 2
I 3
I 4
F (2°Ne/36Ar)atm
Fig. 10. Comparison of (4He/4°mr)rad (4He/21Ne),aaand (2°Ne/36Ar)atm, normalized to their predicted end-member ratios, and expressed as fractionation values (F). (a) Comparison of F(4He/4°Ar),adand F(2°Ne/36Ar)atm.All data fall on or close to the line of equal-fractionation, labelled A, suggesting that the radiogenic and atmosphere-derived rare gases must have been intimately mixed before rare gas fractionation. (b) Comparison of F(4He/mNe)raa,and F(Z°Ne/36Ar)atm. All data are within error of the line F(4Hef21Ne)~ad = 1, labelled B, indicating that no fractionation in the radiogenic 4HeP1Ne ratio is resolvable, regardless of the extent of fractionation in the atmosphere-derived component. This indicates that diffusive/kinetic fractionation processes have been relatively unimportant. The pattern of rare gas fractionation is typical of that predicted by a gas/liquid phase partitioning model.
Scale of fluid flow Some limits may be placed on the total volume of groundwater which has interacted with the natural gas accumulation by considering the volume of groundwater required to introduce the atmosphere-derived rare gases present in the gas fields. In addition, a crustal-radiogenic rare gas mass balance may place some limits on the crustal degassing history of the region. Such mass balance calculations have been made from the Hajduszoboszlo field (Ballentine et al. 1991). The volume of gas in the Hajduszoboszlo gas field is estimated at 33 km 3 gas STP. Given the average concentration of atmosphere-derived 2°Ne and 36Ar in the gas samples, it contains 7.3 × 103 m3(STP) 2°Ne and 1.3 × 104 m3(STP) 36Ar. The corresponding volume of water that must degas completely to account for the total reservoir volume of 2°Ne and 36Ar is 50 km 3 and 17km 3 respectively. This assumes that the groundwater originated as seawater (Erdelyi 1976) and that the concentrations of 36mr and 2°Ne in seawater at 25°C are 7.65 × 10 -7 cm3(STP)/g(H20) and 1.47 × 10-7 cm3(STP)/ g(HzO) respectively (Ozima & Podosek 1983). The difference between the water volumes calculated from the Ne and Ar concentrations reflects the fact that 2°Ne/36Ar ratio in the reservoir gas is not the same as that in atmosphere-equilibrated groundwater. In reality therefore, the volume of water must be much larger, as the fractionation observed
between 2°Ne and 36Ar in the reservoir gas requires that the gas/water volume ratio was very small, which in turn results in only partial transfer of the atmosphere-derived rare gases from the water to the gas phase (Ballentine 1991). The estimated reservoir pore volume is 0.367km3; these water volume estimates are larger than this by a factor of between 50 and 140. Assuming an average pore space of 5% at a depth of 3.5 km (Dovenyi & Horvath 1988), the m i n i m u m volume of water required to have interacted with the gas phase would occupy a volume of between 350 and 1000km 3 of the sedimentary rock. This can be placed in perspective by considering the Derescke sub-basin to the southeast of the Hajduszoboszlo field, which has a sedimentary fill of 3000 km 3 (Fig. 5c). Rare gases produced radiogenically within the crust also form a significant proportion of the total rare gas inventory in the Hajduszoboszlo gas field. Given the total volume of gas in the Hajduszoboszlo field and the average concentration of the radiogenic 4He, the gas field contains 23 × 106 m3(STP) 4He,ad. This amount of 4He is large compared to the volume which could have been produced by a-decay of U and Th within the reservoir itself, and even exceeds the volume expected to have been produced within the sedimentary fill of the Derescke sub-basin. The possibility that deeper regions of the crust have contributed to the 4He mass balance must be considered. For an average U,
He, Ne & Ar IN HYDROCARBON FLUIDS Th content of the upper crust of 2.8ppm and 10.7ppm (Taylor & McLennan 1985), and an average rock density of 2.7 cm 3 g-l, the volume of rock required to produce the measured 4nerad over 1, 10 and 50Ma is estimated at 1 x 105, 1 x 104 and 2 × 103 km 3 respectively. Over the time of basin formation, the equivalent of the entire crustal rare gas production down to a depth of 10km over the area of the Derescke sub-basin (Fig. 5c) is therefore required to produce the volume of 4Herad in the Hajduszoboszlo gas field.
Regional fluid flow overview The information now available from the rare gases associated with hydrocarbon reservoirs starts to provide a picture of the fluid behaviour within these different Neogene sedimentary basin systems. The large volume of radiogenic 4He observed in just the one Pannonian Basin gas field at Hajduszoboszlo, requires a mechanism of storage of these radiogenic rare gases within the deep crust. The release of radiogenic rare gases from the extensional Pannonian and Vienna basins is presumably related to increased heat flow during basin formation (O'Nions & Ballentine 1993) and is accompanied by mantlederived fluids. In addition, the radiogenic rare gas ratios of (4He/E1Ne)rad, (4He/'*°Ar)r~d and (21Ne/4°Ar)rad are close to the theoretical production ratios except where the atmospherederived (2°Ne/36Ar)atm is also fractionated by the same extent from its predicted end member value. The radiogenic rare gases have not, therefore, undergone fractionation from the end member values until after mixing with the atmosphere-derived component. This observation requires that the radiogenic rare gases are stored within the crust at close to their production ratios. Furthermore, both the release and transport of the radiogenic rare gases also occurs without any relative fractionation. Therefore, the mechanism of transport of the radiogenic rare gases to shallow regions within these extensional basins must be principally by advection and most probably as part of a single phase system. In distinct contrast, within the Po basin there is no observable radiogenic 4°Ar or 21Ne, only radiogenic 4He, while the presence of mantle derived volatiles is also unresolvable. This supports the observation by O'Nions & Oxburgh (1988) that loading of the continental lithosphere should not be accompanied by an influx of mantle fluids from depth. The presence of only radiogenic 4He with no accompanying radiogenic 4°Ar or 21Ne suggests that within this
359
environment preferential release of 4He over 4°Ar from the site of production on a local scale occurs and is not accompanied by a flux of fluid derived from any great depth. An obvious difference between the Po and the Vienna/ Pannonian basins is that the former has a geothermal gradient less than half of the others. The volume of water required to introduce the atmosphere-derived rare gases into the Pannonian Basin gas field is significantly larger than the reservoir volume. The radiogenic rare gases, mostly derived from the deeper crust, mix with air-equilibrated groundwater on a regional scale, after which the coherent fractionation of both radiogenic and atmosphere-rare gas abundance ratios occurs through liquid-gas phase partitioning. Clear evidence for kinetic fractionation of rare gases has not been found and diffusional transport of hydrocarbons must be assumed to be relatively unimportant in both the Vienna and Pannonian Basin gas fields. In the case of both the Dosso degli Angeli field in the Po basin and in the Hajduszoboszlo gas field within the Pannonian basin, the concentration of natural gas has occurred, at least in part, by dissolution and transport in groundwater. The Authors are very grateful for reviews by B. Sherwood-Lollar and B. Marty. This work has been supported by the Royal Society and EEC Contract No. JOUG-0006-UK.
References ANDREWS, J.N. 1985. The isotopic composition of radiogenic helium and its use to study groundwater movements in confined aquifers. Chemical Geology, 49,339-351. -& WmSON, G.B. 1987. The composition of dissolved gases in deep groundwaters and groundwater degassing. In: FRITZ,P. & FRAPE,S.K. (eds) Saline water and gases in crystalline rocks. Geological Society of Canada Special Papers, 33, 245-252. BALLENTINE, C.J. 1991. He, Ne and Ar isotopes as tracers in crustalfluids. PhD Thesis, University of Cambridge. & O'N[oNS, R.K. 1992. Mantle neon in Vienna Basin hydrocarbon reservoirs. Earth and Planetary Science Letters, 113,553-567. - - & - - 1993. The Use of natural He, Ne and Ar isotopes as constraints on hydrocarbon transport. In: PARKER,J.R. (ed.) Petroleum Geology of NW Europe: Proceedings of the 4th Conference. The Geological Society, London, 1339-1345. - - , OXBURGH,E.R., HORVATH,F. & DEAK,J. 1991. Rare gas constraints on hydrocarbon accumulation, crustal degassing and groundwater flow in the Pannonian basin. Earth and Planetary Science Letters, 105,229-246.
360 ~,
C.J. BALLENTINE & R.K. O'NIONS
MAZUREK, M. & GAUTSCHI, A. 1993. Rare gas constraints on fluid provenance and migration: the Palfris formation, Wellenberg (Switzerland). In: PARNELL,J., RUFEELL,A.H. & MOLES, N.R. (eds) Geofluidss '93. Torquay 4-7 May 1993, 293-298. BERCZI, I. 1988. Preliminary sedimentological investigation of a neogene depression in the Great Hungarian Plain. In: ROYDON,L.H. & HORVATH, F.H. (eds) The Pannonian Basin: a Study in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45,107-116. BOSCH, A. & MAZOR,E. 1988. Natural gas association with water and oil as depicted by atmospheric noble gases: case studies from the south eastern Mediterranean coastal plain. Earth and Planetary Science Letters, 87,338-346. CASTRO, M.C. & MARTY, B. 1993. Argon isotope measurements and radiogenic 4He/4°Ar ratios in groundwaters of the Paris Basin, France. Hydrodynamic implications. In PARNELL,J., RUFFELL, A.H. & MOLES,N.R. (eds) Geofluid '93. Torquay 4-7 May 1993,290--292. CROVE1"ro, R., FERNANDEZ-PRINI,R. & JAPAS, M.L. 1982. Solubilities of inert gases and methane in HzO and in D20 in the temperature range of 300 to 600K. Journal of Chemical Physics, 76, 1077-1086. DOVENYI, P. & HORVATH, F. 1988. A review of temperature, thermal conductivity and heat flow data for the Pannonian Basin. In: ROYDON,L.H. & HORVATH,F.H. (eds) The Pannonian Basin: a Study in Basin Evolution, Memoirs of the American Association of Petroleum Geologists, 45, 195-233. ELLIOT, T., BALLENTINE, C.J., O'NIONS, R.K. & RIccmuTo, T. 1993. Carbon, Helium, Neon and Argon isotopes in a Po basin natural gas field. Chemical Geology, 106,429-440. ERDELYi,M. 1976. Outlines of the hydrodynamics and hydrochemistry of the Pannonian Basin. Acta Geologica Academiae Scientiarium Hungarieae, 20,287-309. FREY, M., TEICHMULLER, M., TEICHMULLER, R., MULLIS, J., KUNZl, B., BREITSCHMID, A., GRUNER, U. & SCHWIZER, B. 1980. Very low grade metamorphism in external parts of the central Alps: Illite crystallinity, coal rank and fluid inclusion data. Eclogae Geologicae Helveticiae, 73,173-203. GRIESSHABER, E., O'NIONS, R.K. & OXBURGH,E.R. 1992. Helium and carbon isotope systematics in crustal fluids from the Eifel, the Rhine Graben and Black Forest, F.R.G. Chemical Geology, 99, 213-235. HEreON, T.H.E. & VOGEL,J.C. 1981. 'Excess air' in groundwater, Journal of Hydrology, 50, 201216. Hn'AGON, H. & KENNEDY, B.M. 1992. Noble gas in CH4-rich gas fields, Alberta, Canada. Geochimica et Cosmochimica Acta, 56, 1569-1589. JENDEN, P.D., KAPLAN,I.R., POREDA,R.J. & CRAIG, H. 1988. Origin of nitrogen rich gases in the Californian Great Valley: Evidence from helium,
carbon and nitrogen isotope ratios. Geochimica et Cosmochimica Acta, 52,851-861. JIRICEK, R. & TOMEK, C. 1981. Sedimentary and structural evolution of the Vienna Basin. Earth Evolution Sciences, 3--4, 195-203. KHARAKA,Y.K. & SeECHT,D.J. 1988. The solubility of noble gases in crude oil at 25-100°C. Applied Geochemistry, 3,137-144. LADWEIN, H.W. 1988. Organic geochemistry of the Vienna Basin: model for hydrocarbon generation in overthrust belts. American Association of Petroleum Geologists Bulletin, 72,586--599. MAMYRIN, B.A. & TOLSTIKHIN, I.N. 1984. Helium isotopes in nature. Elsevier Science Publications, Developments in Geochemistry, 3. MARTEL, D.J., DEAK, J., DOVENYI,P., HORVATH,F., O'NIoNS, R.K., OXBURGH,E.R., STEGNA,L. & STUTE, M. 1989. Leakage of helium from the Pannonian Basin. Nature, 432,908--912. MARTY,B. & JAMBON,A. 1987. C/3He in volatile fluxes from the solid earth; implication for carbon geodynamics. Earth and Planetary Science Letters, 83, 16-26. ~, O'NIONS, R.K., OXBURGH,E.R., MARTEL,D. & LOMBARDI, S. 1992. Helium isotopes in Alpine regions. Tectonophysics, 206, 71-78. - - , TORGERSEN,Z., MEYNIER,V., O'NIONS, R.K. & DE MARSILY,G. 1992. Helium isotope fluxes and groundwater ages in the Dogger aquifer, Paris Basin. Water Resources Research, in press. MATrAVELLI, L. & NOVELLI, L. 1988. Geochemistry and habitat of natural gases in Italy. In: MATrAVELL1, L. & NOVELLI, L. (eds) Advances in Organic Geochemistry. Organic Geochemistry, 13, 1-13. MATrARELLI, L., RICCHINTO, T., GRIGRARI, D. & SCHOELL, M. 1983. Geochemistry and habitat of natural gases in the Po Basin, Northern Italy. American Association of Petroleum Geologists Bulletin, 67, 2239-2254. MAZOR, E. 1972. Paleotemperatures and other hydrological parameters deduced from noble gases in groundwaters; Jordon Rift Valley Israel. Geochimica et Cosmochimica Acta, 36, 1321-1336. O'NIONS, R.K. & BALLENTINE, C.J. 1993. Rare gas studies of basin scale fluid movement. Philosophical transactions of the Royal Society of London, A344, 141-156. & OXBURGH,E.R. 1988. Helium, volatile fluxes and the development of continental crust. Earth and Planetary Science Letters, 90,331-347. OrrLlK, P., GALFI, J. & HORVATH,F. 1981. The low enthalpy geothermal resource of the Pannonian Basin, Hungary. In: RYBACH, L. & MUFFLER, L.J.P. (eds) Geothermal systems: principles and case histories. Wiley, New York, 221-245. OXBURGH,E.R. & O'NIONS, R.K. 1987. Helium loss, tectonics and terrestrial heat budget. Science, 237, 1583-1588. & HILL, R.I. 1986. Helium isotopes in sedimentary basins. Nature, 324,632-635. OZIMA, M. & PODOSEK,F. 1983. Noble gas geochemistry. Cambridge University Press. PIERCE, A.P., Gorr, G.B. & MYrrON, J.W. 1964. -
-
He, Ne & Ar IN H Y D R O C A R B O N FLUIDS Uranium and Helium in the Panhandle gas field, Texas, and adjacent areas. United States Geological Survey Professional Papers, 454-G. PmRI, M. & MArrAVELLI, L. 1986. Geological framework of Italian petroleum resources. American Association o f Petroleum Geologists Bulletin, 70, 103-130. PrNxl, D.L. & MARre, B. 1993. Hydrocarbon transfer in the Paris Basin (France): a Helium isotope study. In: PARNELL,J., RUFFELE,A.H. & MOLES, N.R. (eds)Geofluids '93. Torquay 4-7 May 1993, 293-298. POREDA, R., JENDEN,P.D., KAPLAN,I.R. & CRAIG, H. 1986. Mantle helium in Sacramento Basin natural gas wells. Geochimica et Cosmochimica Acta, 50, 9-33. ROYDEN, L.H. 1988. Late cenogenic tectonics of the Pannonian basin system. In: ROYDON, L.H. & HORVA'rn, F. (eds) The Pannonian Basin: a Study in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45, 27-48. SMnn, S.P. 1985. Noble gas solubilities in water at high temperature. LOS, Transactions o f the American Geophysical Union, 66,397. -& KENNEDY,B.M. 1983. The solubility of noble gases in water and in NaCI brine. Geochimica et Cosmochimica Acta, 47,503-515. STUTE, M. & DEAK, J. 1989. Environmental isotope study (14C, ~3C, 180, D, noble gases) on deep groundwater circulation systems in Hungary with reference to paleoclimate. Radiocarbon, 31, 902-918. --, SONNTAG, C., DEAK, J. & SCHLOSSER, P. 1992. Helium in deep circulating groundwater in the Great Hungarian plain: Flow dynamics and crustal and mantle helium fluxes. Geochimica et Cosmochimica Acta, 56, 2051-2067. TAYLOR,S.R. & MCLENNAN,S.M. 1985. Thecontinental crust." in composition and evolution. Blackwell Scientific, London. TONGISn, C.A. 1980. Helium - its relation to geologic systems and its occurrence with the natural gases, nitrogen, carbondioxide and Argon. Bureau of
361
Mines Report of Investigations. US Department of the Interior, 8444. TORGERSEN, T. & CLARKE, W.B. 1985. Helium accumulation in groundwater I: and evaluation of sources and the continental flux of crustal 4He in the Great Artesian Basin, Australia. Geochimica et Cosmochimica Acta, 49, 1211-1218. - & Iwv, G.N. 1985. Helium accumulation in groundwater II: a model for the accumulation of the crustal 4He degassing flux. Geochimica et Cosmochimica Acta, 49, 2445-2452. --, KENNEDY, B.M., HrVAGON, H., Cmov, K.Y., REYNOLDS, J.H. & CLARKE, W.B. 1988. Argon accumulation and the crustal degassing flux of 4°Ar in the Great Artesian Basin, Australia. Earth and Planetary Science Letters', 92, 43-56. WALTnER, J.V. & ORVILLE, P.M. 1982. Rates of metamorphism and volatile production and transport in regional metamorphism. Contributions to Mineralogy and Petrology, 79,252-257. WESSELY, G. 1983. Zur Geologie und Hydrodynamik im siidlichen Wiener-becken und seiner Randzone. Mitteitung Osterreich Geologisehe Gesellschaft, 76, 2748. - 1988. Structure and development of the Vienna Basin in Austria. In: ROYDON, L.H. & HORVATn, F. (eds) The Pannonian Basin: a Stud)' in Basin Evolution. Memoirs of the American Association of Petroleum Geologists, 45,333-346. ZAInOWSKI,A. & SeANGLER,R.R. 1990. Noble gas and methane partitioning from groundwater. An aid to natural gas exploration and reservoir evaluation. Geology, 18, 72-74. - - , KOSAt~KE,B.J. & HUBBARD,N. 1987. Noble gas composition of deep brines from the Palo Duro Basin, Texas. Geochimica et Cosmochimica Acta, 51, 73-84. ZARTMAN, R . E . , WASSERBURG, G.J. & REYNOLDS, J.H. 1961. Helium, Argon and Carbon in some natural gases. Journal of Geophysical Research, 66, 277-306.
Index
accretionary prisms dewatering 114 geological setting 113--14 acetate ions 155,157,164 concentration 182,193 stability 187, 188, 189, 190 acid-sulphates 225 advection 31 aerosols, marine 160 Ag see silver Alaska gold province 58 see also North Slope Basin Alberta Basin crude oil metal content 204 formation water salinity 154 rare gas systematics 349 uplift and erosion history 334-5 albite 134, 162,317 ailanite 308 Alleghany gold province 59 Alpine Fault 5 aluminium ion complexes 184 ion systematics 164-5 ore 178, 194 alunite 226 amino acids 178,180 amphiboles 160,308 analcime 302 andesite 3, 4 anhydrite 157,226,302 anions in formation water 154-5 see also n a m e d species
Antarctica, Bransfield Strait study 262,266 antimony in crude oils 205 apatite in red beds 308 apatite fission track analysis 326,335 use in cooling time estimation 327-8 use in palaeogeothermal gradient estimation 329 use in palaeotemperature estimation 327 aquathermal pressuring 236-7 Aquitaine Basin 47, 49 Ar/Ar isotope analysis 278 aragonite dissolution 136 architecture, modelling behaviour of 146 argon ratios 348,351,355,356-7,359 transport effects 357-8 arsenic in crude oils 205,209 asphaltic oils 203-4,208 atacamite 309, 310 Atlantis Deep 262,266 Australia sedimentary basin studies Canning Basin 339-41 Great Artesian Basin 34, 36 Murray Basin 160 Perth Basin 283 gold province studies
Cloncurry 61-6 Kalgoorlie 59 Victoria 58 Austria see Vienna Basin azurite 309 backscattered scanning electron microscopy 100 bacteria role in ore formation 281,282,309 role in reduction 157 Baffin Bay 49 Bakken Shale 242 Barbados prism 114,115,116 barite 280,293,317 barium ion distribution 171 base metal ions in ore fluids 193 basin modelling 9-12 applications method 19 results 20-1 results discussed 21-3 fluid flow 14-15 formation and evolution 12 intraplate stress 15-17 loading 13-14 Bentheim 276 bicarbonate ions 155,157, 164,225-6 Big Horn Basin 34 biodegradation 213 biomarkers 203,214 biotite in red beds 308 bismuth 280 bisulphide ions 193 bitumen biodegradation 213-14 composition 203-5 dating 285-6 defined 203,234 formation 209-10 maturation effects 213 metal interactions 205-9,281-5 migration effects 213 in MVT ore 293 occurrence authigenic mineral data 280-1 ore deposits 278-80 orogenic setting 280 reservoir 276-8 veins 275-6 role in ore exploration 286-8 bornite 306,309 Bransfield Strait 262,266 brine behaviour in basins 28, 38 defined 152 role in metal ion transport 310-11 brittle behaviour faulting 71 plastic processes compared 101 363
364 brochantite 310 bromide ions 156-7,164,205 buoyancy, role in fluid flow of 389 burkeite 303 butyrate ions 182 C isotope studies radio 268 stable 240 calcite 276,280,293 cement 131,136-7 diagenesis 239-40 formation 157 solubility 128,129 veins 240 calcium ions 155,161,194 calcrete 302 California Alleghany gold province 59 San Andreas Fault 80 San Joaquin Basin 154, 185,188,239 Canada Alberta Basin 154,204,334-5,349 Baffin Bay 49 Western Canada Basin 47-9,215,235 Canadian Shield 28, 38, 160 Canning Basin setting 339-40 thermal history 340-1 carbohydrates 180 carbon dioxide formation 175,180, 184 helium ratio 348 role in hydrothermal systems 269 carbonate anions 157 cement 131,136--7 diagenesis 239-40 eodiagenesis 303 mesodiagenesis 305 telodiagenesis 306 solubility 128,129 carboxylic acid anions 179-80 di- 184-5 mono- 182-4 Cascadia prism 115, 116 cataclastic flow 104 cathodoluminescence 100 cations in formation water 155 see also named species
cementation processes 130, 131-2,136-7 red beds 304 sources 128-9 cerussite 313 chalcocite 306 chalcopyrite 205,306,309 chelate compounds 177 see also ligands
Chile prism 114,116 chloride ions in crude oils 205 role in diagenesis 166 role in transport 309-10
INDEX systematics 155,156, 157, 164,224 enthalpy plot 229 chlorinity, role in dissolution of 127, 128 chlorite 302 chromium in crude oils 205 cinnabar 279-80 clay minerals dehydration 236--7 effect on metalloporphyrins 211 clinoptilolite 302 Cloncurry gold province metasomatism 61-2 gold provinces compared 65-6 reactions 62-5 cobalt 179,205,306 collision tectonics 4-5 Colorado 33--4 Colorado Plateau 306 compaction and fluid flow 37-8 concentration gradients 127 conduction, thermal 128 connate water 152,224 continental crust equations 28-31 fluid flow controls 33-9 heat transport 31-3 permeability 27 convection, thermal 1,2, 131,269 cooling time analysis fission track method 327-8 vitrinite reflectance method 328 coordination compounds 177 copper in metalloporphyrins 206 ores 178,179 deposition 278,282 Lubin deposits 309 red beds 301-2,306,307 redox controls 309-10 sources 169-70 Corella Formation 62, 63 Cornish Granite 160 coseismic strain 87-8 cross bedding 146 crude oil biodegradation 213-14 composition 203-5 formation 209-10 maturation 213 metal ion species 205-9 migration 213 crystobalite 226 D isotope studies 165 Darai Limestone 335,336,337 Darcy flow 245 decarbonation 55, 57, 58 decarboxylation reactions 187-8 decollement 114 deformation environmental effects metamorphic conditions 107-8 sedimentary basins 107 factors affecting 101
INDEX deformation---cont'd
mechanisms cataclastic flow 104 diffusive mass transfer 105 dislocation creep 105-7 fracture 101-4 independent particulate flow 104 methods of analysis 100 permeability effects 113, 115-17 Darcyan v. dynamic 117-19 measurements 117 relation to microstructure 119-20 relation to volume change 120-2 degassing reactions 3 dehydration 55, 57-8,236-7 Denver Basin 33-4 desulphidation 55, 58 detachment zone 114 devolatilization 55, 57-8 dewatering 4 diagenesis carbonate rocks 136-7 effect on permeability 118 effect on petroleum 239-41 migration 277-8 pore water behaviour 127-8 effect on calcite 129 effect on quartz 129-30 flux 128 low temperature 133-5 rates of reaction 165 red beds eodiagenesis 302-4 mesodiagenesis 304-5 pH controls 313 redox controls 313 telodiagenesis 305-6 thermodynamic modelling 313-16 role of fractures 131-2 role of organic matter 191-3 thermal effects 131 diagenetic/metamorphic boundary 55-6 dickite 226 diffusion equations 30-1 role in petroleum migration 127,234,243 diffusive mass transfer (DMT) 105 dilatancy and stress cycle 74, 76--7 effect of fluid pressure 76 effect of mean stress 75-6 effect of shear stress 75 dilatancy-diffusion hypothesis 74 dislocation creep 105-7 dismigration 234 disulphides 178,179 Dogger Formation 45 Doherty Formation 63 dolomite 136-7,280,293,317 dynamic fluid viscosity 29 dynamic permeability s e e permeability as a dynamic concept earthquake stress cycle and fluid flow 74, 86--7 dilatancy 74-7
365
fault valves 78-80 post-seismic redistribution 77 East African Rift 49 hydrothermal system 262,266,267,268 East Pacific Rise 262,266 eduction 4 Eh s e e redox potential electron microprobes 100 elevation, significance of 5-6 eodiagenesis 302-4 epidote 5,308 erionite 302 erosion rates 5 estimation of 329-31,333-5 Escanaba Trough hydrothermal system 262,263,264-5 petroleum migration history 267,268 evaporation index (EI) 155 evaporation processes 155-7 evaporite dissolution 157-8 explosive migration 237 expulsion efficiency 250-1 numerical modelling 251-3 expulsion fractures 236,242 extension fractures 71 fatty acids 180 fault gouge 104 fault-valves 78--80 fault/fracture mesh 72-4 faults dilatancy effects 74 fluid pressure effects 70 modelling of 80 hydrology effects 85, 86-7, 91 permeability effects 70-1 modelling of 146-7 Fe s e e iron feldspar dissolution 133-5,302,304 in red beds 308 s e e a l s o albite a l s o plagioclase ferromagnesian minerals 302 ferrous ions 193 Fischerschiefer 243 fission track analysis s e e apatite fission track analysis flood basalt 2-3 flow studies deformation effects 101 mechanisms cataclastic flow 104 diffusive mass transfer 105 dislocation creep 105-7 fracture 101-4 independent particulate flow 104 methods of analysis 100 environmental effects metamorphic 107-8 sedimentary 43,107,131 foreland basin studies 331-5 Papuan Fold Belt 335-8 mathematical model 14-15 measurement 358-9 permeability effects 113,115-17
366
INDEX
flow s t u d i e s - - - c o n t ' d measurements 117 Darcyan v. dynamic 117-22 stress cycle effects 74 mechanisms 74-80 upper crustal studies descriptive equations 28-31 mechanisms 33-9 fluid inclusions 293,309 fluids flow s e e flow studies mixing effects 316-18 pressure effects 70, 313-16 sources 55-7, 69 fluidized bed injection 4 fluorite 293 foreland basins fluid flow 331-5 heat transfer properties 46-9 Papuan fold belt 335-8 formate ions 182,188 formation water 224,231 fractionation, role in petroleum migration of 249-51 fractures cementation 131-2 effect on fluid flow 101-4,131,236,241-2 permeability effects 71 France 45-6, 47, 49, 154 Frio Formation 242,298 Fulmar Sandstone 134 fulvic acid 177-8,180 galena 179,280,306 gallate ions 188 gangue minerals 317 garnet in red beds 308 gas formation 175,184 Gavarnie Thrust mylonite 102,104,106 Germany 243,276 geochemistry of ocean water 2 geoporphyrins 206 geopressure zones 295-6 role in ore formation 296-8 geothermal energy 302 geothermal fluids 222-3 classification 223-4 bicarbonates 225-6 chlorides 224 sulphates 225 hydrological controls 226-7 effect on gold ore 227-9 effect on oilfield water 229-31 steam generation 226 zone recognition 227 water origin 224-5 geothermal gradient analysis 328-31 glutarate ions 184 goethite 226 gold 282 ions in ore fluids 193 ore formation 178,227-9 provinces fluid behaviour 58--61
regional study 62-5 granite and aquifers 5 granodiorite 5 Great Artesian Basin 34, 36 Great Plains Province 34 Green River Shale 204,206 greenstone gold province 58 greywackes 58 groundwater flow rates 358 interaction with rare gases 355-7 Guayamas Basin hydrothermal system 262-4,270 petroleum expulsion 267,268 gypsum 157,303 halite in red beds 302,303,305 halloysite 226 heat transport 1 conduction 128 convection 131 descriptive equations 31-3 s e e also thermal characteristics Hebgen Lake 86--7 helium cycling 2 gas field study 353,355 ratios in natural gases 348,359 transport 357-8 hematite 226,280 red beds 306,317-18 Hg s e e mercury Hill fault/fracture mesh 72-4 hot spots 2 Huldra Field 278 humic acid 177--8,180 Hungarian Basin 52 s e e also Pannonian Basin hydraulic conductivity 29-30 hydraulic diffusivity 31 hydraulic fracturing 235-6 hydraulic head causes 69 measurement 28-9 hydrocarbon solubility 167 hydrocarbons s e e gas also oil a l s o petroleum hydrofracturing 27-8 hydrology, effect of faults on 85, 86-7 hydrosphere buffering 1 hydrothermal eruption breccia 228 hydrothermal systems classification 261,262 effect of organic matter continental 266-7 submarine 262-6 expulsion 267-9 fluid interactions 269 hydrothermal veins 69 hydrous pyrolysis 237 hydroxides, role in sorption of 318 hypersaline waters see salinity and saline waters hyposaline waters 160 Ieru Formation 335,336,337
INDEX Illinois Basin 154,158 ores 295 illite effect on reservoir properties 278,281 formation 134 illitization 236,278,281,302,304 Imburu Formation 336 immiscible flow s e e oil-water flow independent particulate flow 104 inert gases s e e rare gases intracratonic basins 45--6 iodide ions 205 ion microprobe 100 Ireland, ore exploration in 286,288 Irish Sea Basin 278 iron ions 184, 193,204,205,206 minerals 307-8,309 ores 178,179, 194,221 isotope studies radio dating 166,278,285,286,294 stable composition 59,165-6,216,240 Italy s e e Po Basin j arosite 226 Jean d'Arc Basin 50 Juan de Fuca plate 3 juvenile water 224 K/Ar dating 278 Kaiko Project 3 Kalgoorlie gold province 59 kaolinite 225,280,306 dissolution 240 formation 133 134 in red beds 308 Kebrit Deep 262,266 Kennedy Basin 33 kerogen catagenesis 205,209 degradation 186-7 role in petroleum migration 245,247 Keuper Sandstone 278 Kimmeridge Clay 240 Kupferscheifer 309 La Luna Shale 215,242 laser ICPMS 100 laser microprobe stable isotope analysis 100 lead ores 169-70,179,221 characteristics 293 Mississippi Valley Type history of study 293-5 source model 295-7 summary 297-8 in red beds 302,306,307 transportation 310-12 ligands action 177-80 defined 177 destruction 187-8 distribution 180--6 origin 186-7 reaction thermodynamics 188-91
role in diagenesis 191-3 role in ore formation 193-4 lithofacies analysis 239 Louisiana Basin 188 formation water study 154, 159,161,164 Magellan Basin 243 magmatic water 224-5 magnesium ions 155,156, 157,161 complexes 194 magnetite in red beds 308 role in decarboxylation 187-8 malachite 309-10 maleate ions 185 maleic acid 184 malonate ions 185,188, 189,190, 193,194 malonic acid 184 manganese 205,221 mantle convection 2, 3--4 water 2 Manville Formation 48 Maracaibo Basin 215 marcasite 205 Maronan Supergroup 63 mathematical modellings e e modelling maturation 213 Melones fault zone 69 membrane filtration 158-9 mercaptans 178,179 Mercia Mudstone Group 278,307 mercury in crude oils 205,209 ions in ore fluids 193 ore deposits 279-80,288 mesodiagenesis304-5 metal-bearing minerals 307-8 metal ion complexes 175 characteristics 205-8 in crude oil 203 evolution initial 209-13 secondary 213-14 organic complexes 170-1,177,281 metal fulvate 177-8 metal humate 177-8 modelling behaviour diagenesis 191-3 ore formation 193-4 s e e a l s o metalloporphyrins transport 308-13 metal ores 169-70 exploration 286-8 metalloporphyrins 206-8 as biomarkers 214-16 stability 210-13 metamorphic fluids 3 defined 55,224 role in gold ores 65--6 Cloncurry Province 62-5 sources 58--61 devolatilization 57-8 metamorphic/diageneticboundary 55-6
367
368
INDEX
metasomatism and gold ores 65-6 Cloncurry province 62-5 meteoric water defined 224 flux 133-5 role in carbonate diagenesis 136-7 methane formation 3 rare gas mixtures 353-5 ratios 348 sources 355-7 role in hydrothermal systems 269 Mexico, Gulf of 296 mica dehydration 57 hydrolysis 160 see also biotite also muscovite Michigan Basin formation water 10,154, 164 heat transfer 45, 46 micro-organisms 1 microcracks 71 microfracturing 235-6 microstructures 119-20 Mid Atlantic Ridge 262,266 Middle Valley hydrothermal system 262,265-6 petroleum expulsion 267,268 migration 213 see also primary migration also secondary migration mineralization 93-4 red beds 313 fluid flow modelling 316-18 sorption 318--20 thermodynamic modelling 313-16 Mississippi Valley Type (MVT) ores 193,221,279,280 characteristics 293 origin history of study 293-5 source model 295-7 summary 297-8 temperature controls 309 setting 293 Missouri groundwater study 33 modelling basin applications method 19 results 20-1 results discussed 21-3 fluid flow 14-15 formation and evolution 12 intraplate stress 15-17 loading 13-14 rift shoulder erosion 12-13 diagenesis 191-3 fault zones 80 flow 252 fracturing 252-3 MVT ores 293-7 organic ligands 191--4 petroleum pathways 253,254 expulsion 251-2 oilfield water 229 red beds
fluid mixing 316-18 mineral-fluid equilibria 313-16 strain 87-8 thermal history 326 two-phase flow 144-9 Molasse Basin 243 monazite 283,308 Monterey Shales 73,239 montmorillonite 226 role in decarboxylation 187-8 mordinite 226 Mount Isa 61,306 mud volcanoes 4 Murray Basin 160 muscovite 308,317 mylonite 102--4 Na see sodium Nankai prism 114,115,116,117 143Nd/144Ndratios 216 neon ratios 351,355,356-7,359 transport effects 357-8 Neuquen Basin 276 New Albany Shale 204,206 New Zealand 5,204,216,262,266-7 nickel 179,276 effect of biodegradation 213-14 effect of maturation 213 effect of migration 213 occurrences 204-5,206,208-9,210 use as biomarker 214-16 nitrogen : helium ratios 348 North Sea Basin 9, 49,278 North Slope Basin 34-6, 37, 50-2 petroleum geochemistry 215 Nova Scotia Shelf 51 O isotope studies 165-6,240 ocean floor flux 2 ocean water geochemistry 2 volume 3 oil-water flow 141-2 equations governing 142-3 experimental study 143-4 modelling 144-6 anisotropy effects 148 pervasive faulting effects 146--7 results discussed 148-9 sediment architecture effect 146 oilfield water composition 180-1 modelling behaviour 229-31 opal phases 133 optical microscopy 100 Oregon 326 Oregon prism 116 ores deposition 278-80 exploration 286-8 formation modelling 193--4 role of organic matter 175-6 mineralogy
INDEX ores-----cont'd
relation to formation water 169-71 sedimentary 132-3 see also n a m e d m e t a l s
organic acids 167--8, 180 destruction 187-8 method of measuring 181-2 origin 186-7 see also carboxylic acid organic matter 175-6 accumulation continental 266-7 submarine 262-6 chemical behaviour compound distribution 180-6 compound origins 186-7 compound types 177-80 ligand reaction thermodynamics 188-91 destruction 187-8 effects of hydrothermal system 261,262,270 expulsion 267-9 fluid interactions 269 modelling behaviour diagenesis 191-3 ore formation 193-4 role in diagenesis 209-10 role in formation water salinity 167-9 role in metal ion complexes 170-1,281 organometallic compounds 177 organosulphur ligands 178-9 Orubadi Formation 335,336,337 osmotic flow 158 Ouachita Mts 276,280 overburden effect on migration 236 overpressure 18-19 oxalate ions 184, 185,188,189,190,193,194 oxides, role in sorption of 318 oxygen fugacity see redox potential Ozark Mt ores 295 palaeogeothermal gradient analysis 328-31 palaeosurfaces, significance of 227,228 palaeotemperature analysis fission track method 327 vitrinite reflectance 326-7 palladium ores 178,179 Pannonian Basin 9, 11 fluid flow 18-19 hydrocarbons 18-19 modelling methods 19 results 20-1 results discussed 21-3 origin 17,353 rare gas occurrence 353,358 subsidence 18 Papuan fold belt setting 335 stratigraphy 335-6 thermal history 336-8 Paris Basin formation water salinity 154 heat transfer properties 45-6 passive margin, heat flow 49, 50
Pattani Basin 154 Pb ore see lead 2°7pb/2°6pb dating 285 pelite metamorphism 57 peptides 180 peridotite deserpentinization 5 serpentinization 2 permeability deformation-created changes 101,104-5 as a dynamic concept Darcyan v. dynamic 117-19 effect of microstructure 119-20 effect of volume 120-2 introduction 113, 115-17 measurement 117 effect of diffusive mass transfer 105 effect on faults 70-1 effect on hydraulic conductivity 29-30, 37 experimental study 143-4 mathematical modelling 144--6,148--9 anisotropy effect 148 pervasive faulting effect 146-7 sediment architecture effect 146 measurements 242-5 continental crust 27 sandstone 131 shale 131 relation to stress field 71 role in immiscible fluid flow 141-2 equations governing 142-3 Perth Basin 283 Peru prism 114, 116 petroleum hydrothermal generation 261,262,270 continental 266-7 submarine 262-6 migration see primary petroleum migration pH organic matter effects 175,180 pore water 161 role in red bed diagenesis 313,316 role in sorption 320 phenols 180 Phosphoria Shale 215 plagioclase dissolution 134 in red beds 308 plastic behaviour, brittle processes compared 101 platinum 178,179,282 Po Basin 353-5 pore size classification 243 pore water characteristics 127-8 composition 151-2 flux calculation 128 low temperature reactions 133-5 mixing effects 137 porosity deformation-created changes 101 effect on cementation 130 effect of depth 19 effect on diagenesis 56 effect of diffusive mass transfer 105
369
370
INDEX
porosity----cont' d
effect of dislocation creep 105 effect of dissolution 240 porphyrins 179 as biomarkers 214-16 metallic ion links 206-7 stability 210-13 Posidonia Shale 215 post-seismic redistribution 77 potassium cations 155,161 pressure solution, role in petroleum migration of 241 pressure/temperature controls hydrothermal systems 269 Mississippi Valley Type deposits 309 role in diagenesis 314 primary petroleum migration controls diagenesis 239-41 fractures 241-2 hydrothermal system 267-9 lithology 238-9 defined 233--4 driving forces 236-7 expulsion efficiency 250-1 fractionation effects 249-50 laboratory simulation 237-8 modelling 251-3 modes 234-6 pathway analysis geochemical 246-9 petrophysica1242-6 prism taper angle 114 propionate ions 182, 189,193,194 pyrite 205,226,280,293,306 from hematite 305 as reductant 309 role in decarboxylation 187-8 pyrolysis 186-7,245 pyroxene 308 quartz 293 cement mineral 129--30, 131,134,304-5 formation water content 165 gangue mineral 317 role in decarboxlyation 187-8 solubility 128 radio-isotopes see under isotopes Rannoch Formation 148 rare gases groundwater relationships 355-7 levels in gas field study 353-5 sources 351-3 transport methods 357-8 use in fluid flow studies 358-9 Rayleigh convection 131 Rb/Sr dating 294 red beds diagenesis 313 eodiagenesis 302-4 mesodiagenesis 304-5 telodiagenesis 305-6 thermodynamic modelling 313-16 formation 281-2
mineralization 306 fluid flow 316-18 metal distribution 307-8 metal transport 308-13 sorption 318-20 origin 301 related base metal (RBRBM) deposits 193 Red Sea 262,266 Red Sea Rift 49 redox potential effect on metal ion complexes 203,215 effect on metal precipitates 281,309 effect of organic matter 175,180 formation waters 168-9 role in diagenesis 313,316 eodiagenesis 303-4 mesodiagenesis305 role in sorption 320 reduction reactions 304, 305,309 reverse osmosis 158--9 Rheingraben 34, 49,325 rift basins and heat transfer properties 49-52 Rikuu earthquake 86-7 river flow and earthquakes 86-7 Rough Rock Group 285 rutile 205 safflorite 209 Salina Formation 154, 160 salinity and saline waters composition 154, 160-6 defined 152 effect of organic matter 167-9 history of study 152-3 origin 155--60 pH effects 161 role in dissolution 127,128 role in gas solution 351 San Andreas Fault 80 San Joaquin Basin 185,188 formation water 154 porosity study 239 sandstone diagenesis 130 permeability modelling 144-6, 148-9 anisotropy effects 148 pervasive faulting effects 146-7 sediment architecture effects 146 saturation, role in petroleum migration of 245--6 Saxony Basin, Lower 243 Sb see antimony sea water evaporation 155-7 secondary electron scanning electron microscopy 100 secondary migration 234 sediments sedimentary basin studies characterization 44-5, 52 foreland 46-9 intracratonic 45--6 rift 49--52 fluid flow 43 thermal characters equations 44 time factors 43-4
INDEX seismic pumping 74 selenium in crude oils 205,209 serpentinization 2, 5 Shaban Deep 262,266 shear stress and dilatancy 74, 75 shear zones and microstructural analysis 119,120 Sherwood Sandstone Group 278,302 diagenesis 302,306 silica sinter 226,228 silver 193,282,306 Silvermines ores 288 skutterudite 209 ~47Smgeochemistry 216 smectite dehydration 236 formation 134 replacement reactions 302,304,307 sodium in crude oils 205 ions and complexes 155,161,194 solubility, factors affecting 128 sorption 318-20 role in petroleum migration 245-6 South Africa 58 South Caspian Basin 37 SPEX method 237 sphalerite 179,280,306 spilite 3 Sr dating 59,166 steam zone 226,227 stibnite 280 strain cycle 92-3 modelling 87-8 stress cycle and dilatancy 74-7 map of Europe 10 static field effects 71 strike-slip faulting 4-5 strontium dating 59, 166 ion concentrations 161 stylolites relation to stress field 71 role in petroleum migration 241 subduction 3-4 erosion 114 succinate ions 184, 191,194 suction pumps 77 sulphate anions 155,157,164,225 sulphides ions 193 ores 221,282 role of organic matter 179 role in red bed diagenesis 317 sulphur 205,226 Swiss Molasse Basin 47, 49 sylvite 302 Tanganyika Lake hydrothermal system 262, 266,267, 268 Taranaki Basin 204,216 tectonic erosion 3 tectonic fractures 242
telodiagenesis 305-6 temperature role in diagenesis 314 role in dissolution 127,128 temperature/pressure controls hydrothermal systems 269 Mississippi Valley Type deposits 309 role in diagenesis 314 tenorite 310,313 tetrapyrrole 179 thermal characteristics descriptive equations 44 sedimentary basins 52 foreland 46-9 intracratonic 45-6 rift 49-52 time effects 43--4 thermal conduction 128 thermal convection 1, 2, 131,269 thermal diffusivity 44 thermal history modelling 326 cooling time estimation fission track analysis 327-8 vitrinite reflectance 328 fluid flow analysis foreland basins 331-5 Papuan fold belt 335-8 intrusions 338-9 Canning Basin 339-41 maximum palaeotemperature estimation fission track analysis 327 vitrinite reflectance 326-7 palaeogeothermal gradient 328-31 transient v. steady state effects 341-2 thermodynamic buffering 162 thermodynamic modelling 313-16 thiols 178,179 thiophene 178, 179 thorium ores 282-4 thrusts 3, 5 titanite in red beds 308 tonalite 5 topography and fluid flow 33-7 Toro Sandstone Formation 335,336,337 total dissolved solids 152,157 total organic carbon 245 trace elements crude oil content 203-4 in porphyrins 206-8,210-13,213-14 Trans-Atlantic Geotraverse 262,266 travertine 226 trona 303 two-phase fluid flow 141-2 equations 142-3 experimental study 143-4 mathematical modelling 144-6 anisotropy effects 148 pervasive faulting effects 146-7 results discussed 148--9 sediment architecture effects 146 Tynagh ores 286 U/Pb dating 286 Uinta Basin 34, 47, 49,326
371
372 ultrafiltration 158-9 underplating 114 underthrusting 4--5 uplift estimation 32%31,333-5 uranium ores 178 effect of bitumen 278,279,282,284 in red beds 302,306, 307,309, 310 use in dating 285,286 USA basin studies Big Horn 34 Denver 33-4 Illinois 154, 158 Kennedy 33 Louisiana 154, 159, 161,164, 18 Michigan ] 0, 45, 46, 154, 164 Missouri 33 Uinta 34, 47, 49,326 Williston 34, 36, 46,243 gold provinces Alaska 58 Alleghany 59
INDEX volatiles inventory 2 volume change effects in permeability 120-2 Waiotapu 262,266-7 waste disposal problems 1 water inventory 1-2, 69 role in global cycle 2 role in hydrothermal systems 269 rote in ocean recycling 2 sources 224--5 water table 33-7 Waterstones 278 well logs, use of 238 Wessex Basin 302,305 West Siberia Basin 276 Western Canada Basin heat transfer properties 47-9 petroleum analyses 215,235 wettability, role in petroleum migration of 246 Williston Basin 34, 36, 46,243 Witwatersrand gold province 58 xenon ratios 351
valerate ions 182 vanadium ions 179,276 effect of biodegradation 213-14 effect of maturation 213 effect of migration 213 occurrence 204-5,206,210 use as biomarker 214-16 ores 278,282 Vancouver prism 115, 116 Venezuela ores 287 Victoria gold province 58 Vienna Basin 353-5 vitrinite reflectance (VR) 293,326 measurements 335 use in cooling time analysis 328 use in palaeogeothermal gradient analysis 329 use in palaeotemperature analysis 326-7 volatile fatty acid (VFA) 180
Yellowstone National Park 262,266 zeolites 302 zinc ion behaviour 178, 179,193,194,205 ores Mississippi Valley Type ore 221,293 history of study 293-5 source model 295-7 summary 297-8 red beds mineralization 302,306,307 transportation 310-t2 sources 169-70 zincite 313 zircon 205,283 in red beds 308
Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins edited by John Parnell (Queen's University of Belfast, UK)
Geological fluids are a central theme linking the petrography and chemistry of all rock types, deformation processes on the microscopic to the continental scale, and the concentration of economic resources. The fundamental importance of fluid migration and evolution to rock composition and structure is reflected in a growing interest in fluid processes, including a series of successful conferences on water-rock interaction. The papers in this volume are intended to give a review of the whole spectrum of current geofluids research. The papers include international case studies and are written by leading experts in the field. • • • • • •
First overview of fluids research Examples of commercial applications of fluids research Reviews research for the non-specialist Up-to-date extensive bibliographies Promotes methodological transfer between oil and minerals industries Explains theory behind and consequences of fluid flow processes
This volume will be of interest to geologists in the oil and minerals industries and in academia, and to hydrogeologists and geochemists.
• 168illustrations • 374 pages • 23 chapters • index
Cover illustration: Multi-phase hydrocarbon bearing fluid inclusion within a quartz crystal, Dolyhyr, Wales.