Acid Rain Research: Do We Have Enough Answers? (Studies in Environmental Science)

ACID RAIN RESEARCH: DO W E HAVE ENOUGH ANSWERS? This Page Intentionally Left Blank Studies in Environmental Science...

Author: G.J. Heij | J.W. Erisman

8 downloads 692 Views 26MB 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

DOWNLOAD PDF

ACID RAIN RESEARCH: DO W E HAVE ENOUGH ANSWERS?

This Page Intentionally Left Blank

Studies in Environmental Science 64

ACID RAIN RESEARCH: DO WE HAVE ENOUGH ANSWERS? Proceedings of a Speciality Conference, ‘s-Hertogenbosch, The Netherlands, 10-12 October 1994

Edited by:

G.J. Heij and J.W. Erisman National Institute of Public Health and the Environment, F!O. Box 1,3720 BA Bilthoven, The Netherlands

ELSEVIER Amsterdam Lausanne N e w York Oxford. Shannon Tokyo 1995

-

-

-

ELSEVIER SCIENCE B.V. Sarah Burgerhartstraat 25 P.O. Box 21 1, 1 0 0 0 AE Amsterdam, The Netherlands

ISBN 0-444-82038-8

0 1995 Elsevier Science B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the Publisher, Elsevier Science B.V., Copyright & Permissions Department, P.O. Box 521, 1000 A M Amsterdam, The Netherlands. Special regulations for readers in the U.S.A. This publication has been registered with the Copyright Clearance Center Inc. (CCC), Salem, Massachusetts. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the U.S.A. All copyright questions, including photocopying outside the U.S.A., should be referred t o the copyright owner, Elsevier Science B.V., unless otherwise specified.

No responsibility is assumed by the Publisher for any injury and/or damage t o persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Printed in The Netherlands on acid-free paper

V

FOREWORD "his book represents the Proceedings of the International Specialty Conference, "AcidRain Research; do we have enough answers",held for about 120 scientists from 15 countries 10 - 12 October 1994 in 's-Hertogenbosch in the Netherlands. The conference proved a valuable conclusion to the co-ordinated research on acidification in the Netherlands, lasting from the beginning of 1985 to the end of 1994. Directly following the conference, an international team of experts in the field reviewed the research of the third and last phase of the Dutch Priority Programme on Acidification. The main results of the first two phases including a scientific review were published in the Elsevier series on Studies in Environmental Sciences, no. 46 (Heij and Schneider, 1991), while the results of the third phase of the programme, including the review team's report, will also be published in the same series. "he Specialty Conference focused on: Atmospheric deposition Effects of acid deposition on forest ecosystems in the Netherlands Future of acidification research. Atmospheric deposition has been a major research issue in several national and international research programmes. The aim of the Dutch Priority Programme on Acidification in this field was to assess acid, nitrogen and base-cation deposition loads to forest and heathland, and to compare these loads with critical deposition values to determine exceedances. As the critical loads concept is applied to ecosystems, deposition fluxes must also be assessed a t the ecosystem level. During the conference, special attention was given to the following subjects: trace gases, chaired by David Fowler (Institute of Terrestrial Ecology, UK); ammonia, chaired by Willem Asman (National Environmental Research Institute, Denmark) and particle deposition, chaired by J a n Willem Erisman (National Institute of Public Health and Environmental Protection, the Netherlands). Other topics, such as wet deposition, fog and cloud-water deposition, important for obtaining a n overall assessment of deposition loads to ecosystems and soils, were discussed in a session on generalisation chaired by Bruce Hicks (National Oceanic and Atmospheric Administration, USA). At the end of a long-term research programme the question usually arising is: Do we have enough answers, or are we generating new problems to keep our research going .......? Final results and conclusions of the Dutch research on forest stands and forest soils were presented and discussed in that light in a session chaired by BertJan Heij (National Institute of Public Health and Environmental Proteciton, the Netherlands). The session on "Futureof acidification research" on the last day of the conference brought up the question of whether present day knowledge and research trends have attracted sufficient support for decision-making purposes. This session was chaired by Ellis Cowling (College of Forest Resources, North Carolina State University, USA). Future acidification research has to be combined with research on other environmental topics, such as climate change, landuse changes or ecosystem dynamics, incorporating all relevant stress factors. A special session on these topics was chaired by Tomas Paces (Czech Geological Survey, Czech Republic).

vi

Each chairman summarized the main conclusions of his session. These conclusions, and an answer to the question: Do we have enough answers? are listed in the chapter titled Conclusions. The proceedings of the Conference start with the opening statements by Andre van Alphen and are followed by the papers presented during the different sessions along with about 30 posters for explanation to the visiting scientists. Posters were divided into the following topics: critical loads I exceedances, wet deposition I throughfall, dry deposition I concentrations and a miscellaneous session.

ACKNOWLEDGEMENTS The editors would like to express their gratitude for the outstanding effort of all the chairmen and, in particular, the organizing chairman, Toni Schneider. Ottelien van Steenis not only carried out the organisation of the Conference but also operated the Registration and Information desk, with the excellent help of Marianne Vonk. Last but not least, Ottelien also took care of all the preparatory work for the Proceedings. Ottelien and Marianne are therefore greately acknowledged for their contributions. We hope that although the Conference focused on the three topics: atmospheric input, summary of Dutch Acidification research results and the future of acidification research, the excellent calibre of work and new initiatives on acidification research as a whole, reflected in the proceedings, will be of value to both research scientists and policy makers.

vii

CONCLUSIONS Deposition 1. Trace gases - Fluxes of the acidifying compounds NO2, NH3, SO2 and aerosols to forests and short vegetation have not only been measured directly, overcoming important uncertainties in methods and interpretation, but also have heen monitored over long periods. This work provides the basis for greatly improved accuracy of input estimates of pollutants to forests and the landscape in The Netherlands and across Europe. 2. Ammonia / ammonium

- The highest uncertainty in estimates of NH3 deposition is caused by uncertainties in temporal and spatial variations in NH3 emissions.

- The conversion rate of NH3 to N H 4 + aerosol is not known accurately. It is likely that it shows temporal and spatial variations, that e.g. depend on the concentration of acidic compounds in the atmosphere. This information should be known as it determines where NH, will be deposited, because the dry deposition velocity of NH3 is much larger than the dry deposition velocity of aerosol. For that reason reduction of emissions of acidic compounds in the air only could lead to a change in the dry deposition pattern of NH3. The concentration of NH3 at the surface of vegetation and seawater determines partly the flux of NH3 to or from the surface. It should be taken into consideration in transport modelling. NH3 emissions from agricultural crops could be potentially important in the growing season.

m+

-

3. Particle deposition - Dry deposition of particles to forests has often been underestimated until now. Furthermore, the role of particles in regulating water layer (chemistry) on vegetation and thus influencing gaseous dry deposition is important.

4. Generalization - Deposition should be determined at a scale that enables the estimation of risk for ecosystem damage. Furthermore, most important factors determining deposition (edge effects, slopes, topography, roughness transition zones, etc.) should be taken into account in estimating input to sensitive ecosystems. For model development it is necessary to obtain key parameters by field experiments and validate the models by further field measurements. Effects of acid deposition 1. No direct relationship exists between tree health and acid deposition. 2. Atmospheric deposition of N and S compounds on forests leads to:

viii

- changes in vegetation composition into the direction of nitrogen-loving species and monocultures;

- high concentration of Al and NO3 in soil solution and groundwater and to loss of biodiversity in non forest ecosystems. 3. Ozone has a significant adverse impact on plants. Not only crops, but also forest trees can be affected. The impact on natural vegetation is largely unknown as yet. In The Netherlands the contribution of NO, to the total nitrogen deposition is currently less than 20%.But its adverse impact through formation of ozone must not be neglected. 4. The impact of atmospheric deposition on forest trees should be evaluated in terms of risk rather than in terms of visible effects. The future 1. Global climate change and land use change will influence acidification processes; 2. A shift is necessary from effect oriented to system oriented research; 3. Ecologists, studying acidification effects, have to include climate factors; 4. Scientific uncertainties have to be reported explicitly. 5. Long term monitoring programmes are necessary to evaluate effects of acidification and of policy actions. 6. "Local" processes are largely unknown (especially for N). Knowledge on "local" processes will improve knowledge on causal relations.

ix

CONTENTS

Foreword

V

Acknowledgements

vi

Conclusions

vii

OPENING SESSION

1

Opening Remarks A. van Alphen, Chairman Steering Committee Acidification Research

3

ATMOSPHERIC DEPOSITION Session I Trace aases Long term measurements of SO, dry deposition over vegetation and soil and comparisons with models D. Fowler, C. Flechard, R.L. Storeton-West, M.A. Sutton, K.J. Hargreaves and R.I. Smith

7 9

The use of the gradient method to monitor trace gas fluxes over forest: flux-profile functions for ozone and heat J. Duyzer and H. Weststrate

21

Deposition of nitrogen oxides and ozone to Danish forest sites K. Pilegaard, N.O. Jensen and P. Hummelshj

31

Monitoring dry deposition fluxes of SO, and NO,: analysis of errors M.G. Mermen, J.E.M. Hogenkamp, H.J.M.A. Zwart and J.W. Erisman

41

Session 11 Ammonia /Ammonium Ammonia and ammonium in the atmosphere: present knowledge and recommendations for further research W.A.H. Asman

53

Measurement and modelling of ammonia exchange over arable surfaces M.A. Sutton, J.K. Burkhardt, D. Guerin and D. Fowler

71

Preliminary Validation of ammonia emission data using a combination of monitoring and modelling J.M.M. Aben, P.S.C. Heuberger, R.C. Acharya, and A.L.M. Dekkers

81

55

X

Deposition network of the Federal Environmental Agency (UBA) Results and trends D. Kallweit

91

The influence of ammonium nitrate equilibrium on the measurement of exchange fluxes of ammonia and nitric acid Y. Zhang, H. ten Brink, S. Slanina and P. Wyers

103

Session 111 Particle deposition Particle deposition to forests J.W. Erisman, G. Draaijers, J. Duyzer, P. Hofschreuder, N. van Leeuwen, F. Romer, W. Ruijgrok and P. Wyers

113 115

Deposition of aerosol to coniferous forest G.P. Wyers, A.C. Veltkamp, M, Geusebroek, A. Wayers and J.J. Mols

127

Microscopic processes governing the deposition of trace gases and particles to vegetation surfaces J. Burkhardt

139

The atmospheric input of inorganic nitrogen and sulphur by dry deposition of aerosol particles to a spruce stand K. Peters and G. Bruckner-Schatt

149

Session IV

Generalization: total atmospheric deuosition and soil loads: measurements and models On the determination of total deposition to remote areas B.B. Hicks

161

Quantifying the scale dependence in estimates of wet and dry deposition and the implications for critical load exceedances R.I. Smith, D. Fowler, and K.R. Bull

175

Uncertainties associated with the inferential modelling of trace gas dry deposition: A comparison of four models with observations from four surface types J.R. Brook and J. Padro

187

EDACS: European deposition maps of acidifying components on a small scale W.A.J. van Pul, C.J.M. Potma, E.P. van Leeuwen, G.P.J. Draaijers and J.W. Erisman

197

163

xi EFFECTS OF ACID DEPOSITION ON FOREST ECOSYSTEMS IN THE NETHERLANDS Effects of acid deuosition on forest ecosystems in The Netherlands Assessment and evaluation of critical levels for 0, and NH, E. Steingrover, T. Dueck and L.G. van der Eerden

21 1

Experimental manipulations: forest ecosystem responses to changes in water, nutrients and atmospheric loads A.W. Boxman, P.H.B. de Visser and J.G.M. Roelofs

223

Effects of acid deposition on forest ecosystems in The Netherlands: analysis of the Speuld Douglas fir site H. van Grinsven, B.J. Groenenberg, K. van Heerden, H. Kros, F. Mohren, C. van der Salm, E. Steingrover, A. Tiktak and J.R. van de Veen

245

Large scale impacts of acid deposition on forests and forest soils in The Netherlands W. de Vries, E.E.J.M. Leeters, C.M.A. Hendriks, H. van Dobben, J. van den Burg and L.J.M. Bouwmans

26 1

Ecological effects of atmospheric deposition on non-forest ecosystems in Western Europe R. Bobbink and J.G.M. Roelofs

279

Evaluation; integration H.F. Van Dobben

293

Session V

213

FUTURE OF ACIDIFICATION RESEARCH Todav’s knowledge; is it sufficientfor tomorrow’s decision makina uuruoses Lessons learned in acidification research: Implications for future environmental research and assessments E. Cowling

305

Exceedence, damage and area minimisation approaches to integrated acidic deposition modelling C. Gough, J.J. Kuylensstierna, P. Bailey and M.J. Chadwick

321

Reliability of environmental information obtained by modelling and monitoring J.A. Hoekstra, J.C.H. van Eijkeren, A.L.M. Dekkers, B.J. de Haan, P.S.C. Heuberger, P.H.M. Janssen, A.U.C.J. van Beurden, A.A.M. Kusse and M.J.H. Pastoors

333

Session VI

307

xii Session VII Future research: combination with other environmental touics Future of acidification research T. Paces

34 1 343

Potential ecological risk due to acidification of heavy industrialized areas the Upper Silesia case A. Worsztynowicz and W. Mill

353

Acidification and metal mobilization: effects of land use changes on Cd mobility P.F. Rornkens and W. de Vries

367

Acidification interacting with global changes: research to manage drifting systems R.S.A.R. van Rompaey

381

POSTERS Session VIII Critical loads / Exceedances Critical loads of heavy metals for European forest soils G.J. Reinds, J. Bril, W. de Vries, J.E. Groenenberg and A. Breeuwsma

385 3 87

Setting critical loads of acidity for dystrophic peat - a new approach E.J. Wilson, R.A. Skeffington, C.J. Downer, E. Maltby, P. Immirzi and C. Swanson

3 89

A comparison of models for the assessment of critical loads on different scales of observation R.J.M. Lenz, S. Mendler and R. Stary

393

Session LY Wet deaosition / Throuahfall Deposition and leaching at forest stands in Sumava Mts. J. Kubiznhkovii, J. Kubiznhk and 0. Rauch

395 397

Precipitation input of inorganic chemicals in the S. Vitale pine stand of Ravenna (Italy) T. Georgiadis, F. Fortezza, L. Alberti, P. Rossini and V. Strocchi

399

Trends of some components of wet deposition in East Germany after the unification E. Briiggeman and W. Rolle

403

Eight years studying bulk and wet deposition in Spanish Basque country D. Encinas and H. Casado

407

Scavenging of gases during growth of ice crystals G. Santachiara, F. Prodi and F. Vivarelli

41 1

...

Xlll

Contribution of root-derived sulphur to sulphate in throughfall in a Douglas fir forest A.C. Veltkamp, M. Geusebroek and G.P. Wyers

413

Fate of nitrogen in Spruce and Pine ecosystems T. Staszweski, S. Godzik and J. Szdzuj

417

Acid deposition: data from the Swiss Alps S. Braun and W. Fluckiger

419

Dry deposition to bulk samplers underneath a roof in a spruce (picea abies Karst.) forest M. Bredemeier and M. Meyer

423

Deposition on air pollutants to forest ecosystems along pollution and climatic gradients in Poland S. Godzik, T. Staszweksi and J. Szdzuj

425

Session X Dn,deDosition / Concentrations Immission and dry deposition of SO, and ozone lee side of the conurbation of Leipzig in Eastern Europe G. Spindler and W. Rolle

427 429

Dry deposition of acidifying and alkaline particles to forests: model and experimental results compared W. Ruijgrok

43 1

The measurement of ammonia in the National Air Quality Monitoring Network (LML) (1) Instrumentation and network set-up B.G. van Elzakker, E. Buijsman, G.P. Wyers and R.P. Otjes

439

The measurement of ammonia in the National Air Quality Monitoring Network CML) (2) Results and performance B.G. van Elzakker, J.T. Schippers, J. Stuiver and G.J.B.M. van Uden

443

Fine resolution modelling of ammonia dry deposition in Great Britain R.J. Singles, M.A. Sutton and K.J. Weston

449

Fog deposition measurements on Douglas Fir forest A.T. Vermeulen, G.P. Wyers, F.G. Romer, G.P.J. Draaijers, N.F.M. van Leeuwen and J.W. Erisman

453

The contribution of canopy exchange to differences observed between atmospheric deposition and throughfall fluxes G.P.J. Draaijers, J.W. Erisman, N.F.M. van Leeuwen, F.G. Romer, B.H. te Winker and G.P. Wyers

455

xiv Dry deposition monitoring SO,, NH, and NO, over a coniferous forest J. Hogenkamp, J.W. Erisman, M. Mermen, E. Kemkers, A. van Pul, G. Draaijers, J. Duyzer and P. Wyers

457

Gas deposition of sulphur dioxide on the territory of the Czech Republic in 1991 M. Zapletal

459

Session X I Miscellaneous Forest condition in Europe and North America: What have we leant over the past ten years? H. Visser

463 465

Stomata1 regulation in field-grown Douglas-fir W.W.P. Jans and E.G. Steingrover

473

Carbon partitioning in Douglas-fir E.G. Steingrover, M. Posma and W.W.P. Jans

475

Decreasing concentration of air pollutants and the rate of dry and wet acidic deposition at three forestry monitoring stations in Hungary L. Horvhth, Gy. Baranka and E.Gy. Fiihrer

477

The characteristics of acid precipitation in Southern China Y. Bai and X. Tang

483

The response of peat wetland methane emissions to temperature, water table and sulphate deposition D. Fowler, J. MacDonald, I.D. Leith, K.J. Hargreaves and R. Martynoga

485

Annex I

List of participants

489

OPENING SESSION


3

ACID RAIN RESEARCH CONFERENCE, October 10-12, 1994 Opening remarks by Andr6 van Alphen Deputy director Air and Energy, Ministry of Housing, Spatial Planning and Environment, P.O. Box 30945, 2500 GX, The Hague, The Netherlands Ladies and Gentlemen, Addressing this conference, while deputizing for Joris Al, the Chairman of the Steering Committee on Acidification Research, is a rather confusing experience. Of course I feel privileged to do so, but I realise that Joris Al, who chaired the entire third phase of the Dutch acidification research programme, is far more experienced in this matter than I. However, I have sufficient knowledge of the subject to know that there is another, more factual, reason for these mixed feelings. This Symposium marks the completion of a ten-year acidification research programme in the Netherlands, at a moment when the call for hard and conclusive scientific evidence coincides with the most drastic cut in research budgets ever. That is why I really feel confused. In the next ten minutes I intend to focus on three elements of the problem: -the desire for hard evidence; -the end of a research programme of long standing; -future acidification research.

To start with the first, I must say that the sub-title of the conference "Do we have enough answers?" is a perfect description of the policymakers' dilemma: we know a lot about acidification, but is it all the knowledge we need for policy purposes and would more knowledge lead to policy measures that are not only easier but also better? When looking into that question there are two points which have to be stressed. Firstly, results from scientific research can never be a substitute for policy decisions. Both scientists and policymakers may regard this as a truism, but it is still worthwhile repeating it now and then. Secondly, one should not forget that environmental problems like acidification, with a great lapse of time between the onset of the effects and evidence of damage, can only be dealt with on the basis of a risk approach. Action should be taken on the basis of the risk that acidification results in harmful effects which, if we postpone action until damage is apparent, will probably be irreversible. Since the start of acidification abatement the approach has been to start by gradually reducing emissions while intensifying research into acidification to consolidate the scientific basis for action. Of course the intention was to ensure that the timing of more drastic measures coincided with the development of further scientific substantiation of the acidification issue. In real life however, scientific knowledge develops more capriciously: not only does our

4

understanding of a great number of issues increase in the course of time, but other issues, which seemed clear in the beginning, pose new problems. This is what is happening at present with respect to ammonia. These questions are not necessarily fundamentale, but they require clarification to prevent the main issue from being confused. It is entirely understandable that when the social and economic consequences of abatement measures become extremely serious, those who have to pay for these consequences will have difficully accepting any validation that is unclear. Nevertheless, I do not expect new research programmes to result in a break-through in our knowledge about the risks of acidification. As I mentioned before, science cannot choose between protection of the environment and nature on the one hand and social and economic consequences of emission reduction measures on the other. It is the task of the government to weigh these interests. To enable government to come to a balanced decision, it is of great importance to be as clear as possible about what is known with respect to the risks of acidification and the reliability of that knowledge, using the tremendous amount of scientific information now available. The project team has the important and difficult task to create such clearness after this conference and the subsequent international scientific review process. And then the second element of the problem that causes the confusion: the end of the coordinated acidification research programme. A tremendous amount of knowledge has been generated in a unique cooperative venture involving numerous scientific institutes and scientists. Scientific cooperation on a research programme financed all these years by an equally unique form of cooperation between government and industry. An almost countless number of publications in scientific periodicals and theses has resulted from this acidification research programme over the years. I am sure that the international review will confirm that the research in the third phase of the research programme was of high quality, as it was in the first and second phases. Dutch scientists play an importante role in improving scientific understanding of acidifica-tion at a European level, and Dutch research has made an important contribution to the development of national and international acidification control strategies. Its role is exemplified by the development of the scientific basis for the recently signed UN-ECE Second Sulphur Protocol. Although this acidification research programme is now coming to an end, there are still questions unanswered and undoubtedly new questions on specific aspects of the acidification process will arise when our national acidification policy is implemented further. In addition, acidification abatement at a European level, preparing "second generation protocols", will require further scientific support. That brings me to the third element of the problem: the future of acidification research. To generate answers to remaining questions at a national and international level, further investigations are certainly required. However, there is no specific need to incorporate relevant studies in a new coordinated research programme. It is considered sufficient that institutes which investigate acidification further can apply for government funding, competing with other air pollution research projects. In principle there is nothing wrong with such a development. However, it is alarming that at

5

the time the third phase of the acidification research programme comes to an end, the budget for acidification research of my ministry is drastically cut. Not much financial support for further acidification studies can be given when budgets considered necessary for the years to come are cut by 50% or more; as this is the probable outlook for 1995 and after. An important area of environmental research has been developed. After a decade of financial support by government and industry, it seems that acidification research is to be left to sink or swim on its own. Will it be possible to carry on or do we have to fear for the decline of research facilities in the Netherlands? I don't know. But I do not think it necessary to end on a gloomy note. Because of the position the acidification issue has acquired in scientific research, because of the enthusiasm and dedication of the scientists and because I believe that in the end its importance will once more be recognized at a political level, but above all, because scientists are extremely creative in raising funds, I have confidence in a positive future for acidification research and in the continuation of cooperation between scientists and policymakers.

I wish you all a successful conference here in Den Bosch, where on earlier occasions other memorable meetings on acidification have been hosted. But, I hope especially that your stay in this charming city and in the Netherlands proves to be a very pleasant one.


ATMOSPHERIC DEPOSITION SESSION I TRACE GASES


G.J.Heij and J.W Erisman (Editors). Acid Rain Research: Do we have enough answers? 0 1995 Elsevier Science BCI All rights reserved.

9

LONG TERM MEASUREMENTSOF SO, DRY DEPOSITION OVER VEGETATION AND SOIL AND COMPARISONS WITH MODELS

D. Fowler, C. Flechard, R.L. Storeton-West, M.A. Sutton, K.J.Hargreaves and R.I. Smith Institute of Terrestrial Ecology, Bush Estate, Penicuik, Midlothian, EH26 OQB, Scotland.

Abstract A semi-continuous series of measurements of SO, fluxes above soil, wheat and sugar beet has been used to quantify the major components of canopy and surface resistance in a wide range of conditions. The data show that over dry cereal canopy, the marked diurnal cycle in canopy resistance is regulated primarily by changes in stomata1 resistance. Good agreement between a process based model of SO, deposition and the field data in dry conditions has been obtained. For a dew and rain wetted crop, the canopy resistance is decreased from 80 s m-l to 55 s rn-'. Longterm (4 month) median deposition velocity (V,) was 7.2 mm s" for wheat. For sugar beet, similar results are obtained with median vd of 5.7 mm s-'. There is also a clear dependence of r, on SO, concentrationwith r, decreasing from 100 s m-' at 4 pg m-3 SO, to 40 s m-l at 20 pg m-3for both crops. For bare soil, canopy r, is small with a median value of 15 s rn-' for dry soil and 5 s m-' for wet soil. The widespread assumption of very small r, for wet surfaces is clearly an oversimplificationin current models. The influence of these new findings for annual dry deposition estimates will be discussed. For both crop surfaces vd increased in the presence of surface water for dew precipitation by typically 30%. The increased affinity of wet leaf surfaces for SO, uptake w a s equivalent to a 40% decrease in the canopy resistance r,.

-

INTRODUCTION The turbulent deposition of SO, to terrestrial surfaces represents the major removal process for boundary layer SO, over large areas of Europe. Even in the high rainfall regions dry deposition may contribute 20-30% of deposited sulphur. While wet deposition is routinely monitored throughout Europe and North America to provide estimates of wet deposition, the majority of countries rely on simple parameterization of dry deposition using monitored SO, concentrations and deposition velocities from the literature. In North America, the dry deposition monitoring network is in fact a series of monitoring stations for meteorological variables and SO, concentration for which deposition rates are inferred, the measurements do not provide flux to the surface directly.

10

Our understanding of dry deposition process in the field was provided by micrometeorological studies of fluxes of SO, to natural surfaces. The early work in the 1970's by Garland (1977) and Fowler and Unsworth (1979) amongst others, quantified the relative importance of surface and atmospheric processes in regulating deposition rates for a range of natural surfaces and atmospheric conditions. These analyses have been extended to provide annual dry deposition estimates on the basis of multiple resistance models Hicks et al. (1987), and similar approaches have been widely applied to estimate annual SO, inputs to catchments, regions or countries (Fowler, 1980; RGAR, 1990). Although it was widely considered that rates of dry deposition could not be continuously monitored the recent work by Erisman et al. (1993) shows that instrumentation has developed t o a point where simple flux-gradient approaches may be used to estimate dry deposition continuously at sites suitable for application of micrometeorologicalmethods. The measurements by Erisman et al. demonstrated the importance of surface water on foliage on rates of SO, deposition and provide an excellent basis for the temporal extrapolation of fluxes from measured SO, concentration fields with appropriate meteorological and land use information. This paper describes an SO, flux measurement station for continuous operation and reports results of a year of measurements of fluxes at an agricultural site in the English Midlands. The results are used to demonstrate the relative importance of Werent sinks at the terrestrial surfaces in a range of conditions, and the results of measurements are compared with model estimates. THEORY AND METHODS The measurements were made at a field site in arable agricultural land close to Sutton Bonington, a University of Nottingham field station. The equipment was placed close to a field boundary and provided a fetch of between 200 and 300 m in most wind directions, excepting the 45" sector centred on 180" within which a n instrumentation cabin was placed to house the monitoring and logging equipment. The site provided winter wheat, bare soil and sugar beet according to sector and time of the year during the measurement period of April 1993 until June 1994. Concentrations of SO, were monitored using a high sensitivity pulsed fluorescence monitor (TECO 43s) which sampled filtered air along heated sample lines at two heights in the surface layer and up to a height of 2.3 m. Wind speeds were measured at three heights and air temperature at two heights using sensitive cup-anemometers (Vector Instruments) and miniature thermocouples (Campbell Scientific) respectively. Precipitation, surface wetness, wind direction and total radiation were also monitored continuously at the site. All of the instrumentation was controlled and logged using a Campbell 21X micrologger. In addition to the long-term continuous measurements, 2 campaigns each of 2 weeks duration of measurement at the site were also made using a 5-point gradient system for SO,, temperature and windspeed and a Campbell instruments Bowen ratio system. The field data were analysed to provide fluxes of SO, (F,) calculated from

11

where k i s Von Karmans constant, u is windspeed, z height, x is SO, concentration, d the zero place displacement and ($,&J1 is a stability factor to correct fluxes for the effects of thermal stratification of the surface layers of the atmosphere. A detailed account o f the standard micrometeorological theory is outside the scope of this short paper a n d may be found in m o m (1976) or Monteith & Unsworth (1990). The sequential sampling system provided a measure of vertical gradients during periods w i t h constant SO, concentration. However, non-stationarity in SO, concentration introduced an important source of uncertainty in the sequential sampling system, and contributed an average of 18% of the uncertainties in V,. To overcome these effects the results were detrended and running median values were obtained for SO, concentration ( x ) , deposition velocity (Vd),the flux F, and r, with a time constant of 3 hours. For the continuous monitoring of fluxes there are a range of conditions which result in erroneous fluxes. It was necessary therefore to filter out data obtained when the wind direction was within the sector occupied by the instrument cabin, when atmospheric stability effects were too large to be corrected by standard approaches, and in this case data were rejected when Monin-Obukhov lengths were < 5m. For periods during which SO, concentration changed rapidly, non-stationarity effects were important, but these also provided important data for the investigation of processes and whenever possible, corrections to the data set to overcome sequential sampling errors were made and the data were accepted.

SO, concentrations at the site The half-hour mean values of SO, concentrations for one year of monitoring are, as is usual, log normally distributed. The mean concentration is 12.8 yg m-3, geometric mean 6.1 yg m-3 a n d geometric standard deviation is 3.5 as shown in Fig. 1. There is a clear diurnal cycle in SO, concentration at the site with a rninimum at F ~ ( K 5) 0300 GMT and at 1300 maximum GMT. Similarly, windspeed shows a -2 -1 0 1 2 3 marked diurnal cycle =f-=@a lb with a minimum at 0500 G M T and a maximum at 1400 Figure 1. The log-normal distributionof SO, concentrations(April 1993 - April 1994).

12

GMT. The SO, data show a pronounced sector dependence with concentrations being largest (- 15 pg m-3)with northerly winds and smallest with southerly winds 5 pg m-3,as shown in Fig. 2. The measurements of SO, deposition may be divided into three groups, according to the surfaces present during the year within the upwind fetch, as wheat, sugar beet and bare soil.

w c w h0

-

. . .

,

243.

'.

, - , '

- . ,

'

-

?I0

=I< E

. ,

,

.

- ,

'

1IU

.

150

'

HO

m Figure 2. Wind sector dependence of sulphur dioxide air concentrations at Sunon Bonington (April 1993 - April 1994).

Dry deposition on to wheat

The measurements over winter wheat were made during the period AprilJuly 1993 during which the canopy height and leaf area index increased from 30cm to -1OOcm and 1.5 to 4.5 respectively. Considering the whole data set for the period, the median of the population of hourly average data for deposition velocity was 7.2 mm 5.' whereas that of the maximum rates of deposition (Vmm)possible was 17.1 mm s-l showing that canopy resistance exerted a major control

Figure 3. FrequencydistributionofV-and V, Im. above a wheat canopy (April - July 1993). Negative values denote SO, emission - or more likely release b the atmosphere of pre-depositedSO,.

13

over rates of deposition under moist conditions (Fig. 3). The median canopy resistance rewas 81 s m-' for the whole data set. The leaf wetness sensor data were used t o show the average influence of leaf wetness on canopy resistance and deposition velocity as shown in Fig. 3. These data show, as expected, that re is significantlysmaller in wet conditions averaging 55 s m-' than in dry conditions (83 s m-') but the wet surfaces were not generally a perfect sink for the SOz. The generalized data from Fig. 3 conceal large diurnal variations in deposition velocity which result from both changes in both r,, r b and r,. Typical diurnal variations in vd for the vegetative phase of the crop growth are shown in Fig. 4 with afternoon mixima in V, of 10 to 15 mm s-' and nocturnal minima of 2 to 5 mm s-'. It is important to recognise that a significant fraction of the diurnal variation 0 fI in v d results from the decrease in ----v, windspeed and the increase in atmospheric stability during the night 1UOUl993 W30 0230 0430 0630 a30 1030 1190 I430 1630 1 8 3 2030 P:30 which lead to a marked diurnal cycle M Figure 4. Measured depositionvelocities of SO, over a wheat canopy. i n r, and rb. The effect is clear even .averaged over the April-July data set as I d shown for the wheat crop in Fig. 5. The small effects of liquid water on leaf surfaces on r, relative to those 0 : : : : : : : . : : : : . : : : : : reported by Erisman et al. (1993) may result from solution and oxidation 'b (..nil) processes in the liquid film not being sufficiently rapid to maintain or oxidise m:w, 0230w30 0 6 : a:w ~ iow 1130 I*:N 16:wIWO a130 0:)o GMr the S'" in solution as Figure 5. Mean diumalc ~ l ofe afmospheric aerodynamic (r. (Im.)) and viscous sub-layer (r,) resistances to SO, transfer above a wheat canopy (April - July 1993). Values are the water evaporates. geometric means and 95 % confidence intervals. An example of this

- I

5

:

'

:

:

14

effect is shown in Fig. 6 in which the wetting of the wheat canopy by rain initially results in a large rate of deposition but as the canopy dries after a brief shower a peak in SO, emission is observed. Clearly, the emission is consistent with loss of dissolved S'" back to the gas phase as water evaporates in warm sunny conditions, and the sequence of events is repeated following a further shower later i n the day. These effects are not the general rule but ~3002:30M.301:301:3010301230 143016:30 1P3020302230 to illustrate the do serve Qbfr point that the assumed Figure 6. SO2exchange over a wheat canopy (23/07/1993): influence of precipitation reduction in rc to and leaf W ~ Q I ~ S Son the dry&position velocity of SO,. Precipitationevents that close to zero require the occur at 11:OO and between 13:30 and 1430 result in a wetted canopy and of the chefid increased deposition velocity at 11:00-11:30 and 13:3014:00, soon after, emission occurs (negative v,) as the result of desorptionof unoxidized SO, S'" in solution to proceed during evaporation. t o SO: and for the acidity generated in the aqueous film to be neutralized by soil derived base cations or by ammonia to prevent r, increasing as the aqueous and gas phase SO, achieve an equilibrium. The process is discussed in some detail by Brimblecome (1978) and by Chameides (1987) largely on theoretical grounds but in neither case are there field data for canopy exchange of water and SO, to confirm the details of the kinetics of the processes. Such field data are necessary for a range of conditions to provide a satisfactory basis for modelling the effects of surface water on the values of rcand V, and hence long term rates of exchange. Until then, the application of average r, values obtained by experiment are the only alternative, and clear differences are shown in the data presented here and that by Erisman et al. (1993). The mean diurnal changes in V, for dry canopies are those driven mainly by changes in stomatal resistance with daytime minima for r, entirely consistent with measured stornatal resistance (rJ. These findings are consistent with the data for grassland by Erisman et al. (19931, Garland (19771, and for wheat by Fowler and Unsworth (1979). The collection of a substantial data set allows the variation in r, with a range of variables to be examined. In particular it is important to show whether the deposition velocity is strictly independent of SO, concentration. By sorting the data ~~

k(d)

! a 200

loo 1% 50

m

-* V.(ImP)

18 16

14

--

12. 10

1 -

so. c

.

( rl=*)

Figure 7. Dry deposition of SO, onto a wheat canopy (April-July 1993). Variations of canopy resistance and deposition velocity with SO, concentmtion. Reference height is lm.above the zero-plane.

-10

0

10

sq

30

u) . .'

40

50

(om?

Egure 8. Frequencydkautionof V, and V, Im. above a sugar beet canopy (May-July 1994).

by concentration the relationship between V,, r, and atmospheric resistances with SO, concentration has been investigated. The analysis shows in Fig. 7 that the atmospheric resistances are almost constant with SO, concentration in the range 1 to 20 p g m-3. However there is a clear reduction in r, with increasing SO, concentration over the range 2 to 12 g so, from 150 s rn-' at 2 p g m-3 to 50 s m-' at 12 pg SO,m'3. Dry deposition of SO, on to sugar beet An extensive data set obtained for measurements of SO, fluxes over sugar beet were obtained during the period May to July 1994. The median deposition velocity was 5.7 mm s-' whereas 1 hour of V,, was 14.6 mm s-'. Median deposition velocities to dry and wet canopies of sugar beet were 5.6 and 6.3 mm s" respectively as shown in Fig. 8. An example of the differences in behaviour of V, and r, in these measurements with those of Erisman may be seen in the data for two days over the sugar beet crop. The first day (8/6/1994)a warm

16

-Vd

I-.

....v-

.......

,..a.

10

,A -

Figure 9. Diurnal Mliations m canopyre&ance and deposition velocity of SO, over a dry sugar beet canopy (OWM/1994). z l

I

sunny early summer day, shows the common diurnal cycle in vd with mid-day maxima of 10 mm s-land small nocturnal minima of 2 m m s" in calm, strongly stable conditions (Fig. 9). By contrast, for a day during the preceding October (13/10/93) the deposition velocity shows V, very close to V,, all day, for a canopy in cool wet conditions (Fig.10). On this day there was no evidence of r, increasing with time in the presence of a wet canopy. The reason r, does not increase, and that the surface remains a 'perfect sink' for SO, is a matter of speculation but must be closely linked to the chemical processing of the dissolved SOz.

SOILS Following the harvest of the winter wheat crop the field was cultivated and provided a long period of smooth bare soil over which SO, fluxes were measured. (-.) 10 The median vd was 13 mm s-l and was quite close to the median V,, of 15 mm s-'. The underlying canopy resistance was therefore very small and averaged just 12 s m-'. The surface was behaving almost as a perfect sink for SO, and while the presence of liquid water did reduce r, as shown in Fig. 11to 4.9 s m 1the increase in deposition velocity was small (to 15 mm s-l)and was then almost equal to vmu. The very large rates of deposition to base soil were larger than those reported by Garland )1977) and by Payrissat and Beilke (1975). A consequence of the small r,

17

is that vd for the bare soil increased almost linearly with wind speed (Fig. 12). The r, values widely applied in models of SO2 dry deposition are generally much larger than those reported here and, if such values as those reported here are applicable generally, the uptake by arable land is winter will be much larger than has been assumed and landscape values for V, in winter should be significantly larger.

---

Comparison of modelled and measured deposition velocities -100 w m w w m m m w m i l a The application of resistance 1’ 11. ~requ~cydicttibutpn &tograms for measured deposition models to estimate rural SO, dry velocities (V,) and maximum deposition velocities allowed by deposition is now a standard technique turbulence (v-) Over bare soil at Bonington (JanUatY-MaY by Hicks et (1987), Fowler (1980) 1994). Figures in brackets are 95 % confidence intervals. and by Sandness (1993) and Erisman (1994). The application and methods in all cases differ although the underlying principles are common. In the method applied for the UK (RGAR 1990) the landscape is subdivided at 5 categories (arable, forest, grassland, moorland and urban). The stomatal response to light and temperature for species representative of such land uses are used to calculate the canopy resistance for water vapour, which is corrected for difisivity differences between SOz and HzO t o provide stomatal uptake for SOz. The leaf surface resistance is assumed constant and quantified by experiment. Climatological, or measured meteorological data (radiation air temperature, wind velocity and canopy heights) are then used to calculate vd which are combined with monitored SO, concentration to yield the flux. One of the most valuable products of the SO, deposition monitoring study is the large .-, data set t o check against 31 A model production. comparison of the UK dry 1 deposition model (Smith & Fowler, 1994) with the measurements reported here is presented in Fig. 13 for a dry wheat canopy in its vegetative phase. The agreement between measurements and the model Figure 12. Mean variationswith wind speed of deposition velocity for SO, over bare soil (January-May1994). Reference height is lm. above the zero-plane

18

--

0 M

00m mm um

V d WdaaaO

06.00

01.00

iom

12m 14:oo 16m 11:oo

2a.m Dim

carrr Figue 13. Dry deposition of SO, onto a wheat canopy (16/04/1993). Measurements are compared to model predictions using halfhourlyme&omb@~ and mean deposition velocities computed for a 'wet surface'day. Reference height is 1 m. above the zeroplane.

25 20

-ModdlcdVd

MtarmedVd

wetdke

----------

15 10 5 0

WOO 02:OO 04:OO 06:OO 08:OO 1O:OO 1290 1490 1600 18:OO 2000 2200

GMT 14. Dry deposition of S0,onto sugar beet (07/06/19!34). Measurements are compared to model predictions using half-hourly meteorobgical data and mean deposition velocities computed for a 'wet surface' day. Reference height is 1 m. above the zero-plane.

is excellent on average although some of the 30 minute measurements differ significantly from the model, largely as a result of non-stationarity effects. For wet surfaces, as shown earlier, the assumption of constant and small value for r, provides a poor estimate of SO, deposition onto the sugar beet canopy to wet conditions (Fig. 14). The data show consistently smaller rates of deposition when the canopy was wet. This, as described earlier, is a consequence of oversimplistic assumptions about the chemical behaviour of the S'" in the liquid film on the sugar beet canopy. The overall results of the model and measurement comparisons for the two canopies vegetation show that 1. Diurnal and seasonal cycles in V,, are simulated very well by the model for dry canopies of either sugar beet or wheat.

19

2. Wet canopies are not perfect sinks and to model the r, and V, for those more knowledge of the processes in liquid films is necessary. A fix for the model can be provided by setting r, for wet canopies at this site to 60 s m-'.

ACKNOWLEDGEMENTS The authors gratefully acknowledge the UK Department of the Environment for funding this study and Miss Fiona Greenwood of the University of Nottingham for assistance with the field measurements.

REFERENCES Brimblecombe, P. (1978). Dew as a sink for sulphur dioxide. Tellus 3 0 151-157. Chameides, W.L. (1987). Acid dew and the role of chemistry in the dry deposition of reactive gases to wetted surfaces. J. Geophys. Res. 92: 11,895-11,908. Erisman, J.W. (1993a). Monitoring the dry deposition of SO, in the Netherlands. Atmospheric Environment 27A pp 1153-1161. Erisman, J.W. (1984). Evaluation of a surface resistance parameterization of sulphur dioxide. Atmos. Environ. 28(16): pp 2583-2594. Erisman, J.W. and Van Pal, A. (1994). Parameterization of surface resistance for the quantification of atmospheric deposition of acidifying pollutants and ozone. Atmos. Environ. 28(16): pp 2595-2607. Fowler, D. and Unsworth, M.H. (1979). Turbulent transfer of sulphur dioxide to a wheat crop. Quart. J. Roy. Meteor. SOC. 105 767-783. Fowler, D. (1980). Removal of sulphur and nitrogen compounds from the atmosphere in rain and by dry deposition. Proceedings of the conference on ecological impact of acid precipitation, Norway 1980 (eds: Drablm, D. and Tollan, A.) pp 22-32, S.N.S.F. Publishers. Garland, J.A. (1977). The dry deposition of sulphur dioxide to land and water London. A364: 245-268. surfaces. Proc. Roy. SOC. Hicks, B.B., Baldocchi, D.D., Meyers, T.P., Hosker, R.D. and Matt, D.R. (1987). A preliminary multiple resistance routine for driving dry deposition velocities from measured quantities. Water, Air and Soil Pollution 36: 311-330. Monteith, J.L. and Unsworth, M.H. (1990). principles of Environmental Physics. Arnold, London. Payrissat, M. and Beilke, S. (1975). Laboratory measurements of the uptake of sulphur dioxide by different European soils. Atmos. Environ. 9: 211-217. RGAR (1990). Acidic deposition in the United Kingdom. The third report of the Review Group on acid deposition. UK Department of the Environment. Sandnes, H. (1993). Calculated budgets for airborne acidifying components in Europe, 1985, 1987, 1989, 1990, 1991 and 1992. Det Norske Meteorologiske Institutt. EMEP/MS C-W Report 1/93. Thom, A.S. (1975). Momentum, mass and heat exchange of plant canopies. In: Vegetation and the atmosphere Vol. 1principles. Ed. Monteith, J.L. pp 57-109. Academic press, London.


G.J. Heij and J.U! Erisman (Editors). Acid Rain Research: Do we have enough answers?

0 1995 Elsevier Science BV A11 rights reserved.

21

The use of the gradient method to monitor trace gas fluxes over forest: Flux-profile functions for ozone and heat. Jan Duyzer and Hilbrand Weststrate IMW TNO, P.O. Box 6011,2600 JA Delft, The Netherlands Abstract

This study aims to assess the relation between fluxes and profiles above forest for trace gases. To this purpose the flux of ozone to a Douglas fir forest was measured continuously by eddy correlation for seven months in 1993. During the same period vertical profiles of air temperature and ozone concentration were determined over the forest. In addition several turbulent parameters were recorded. From the observed temperature profiles and sensible heatfluxes flux-profile functions could be derived. Due to the scatter in the vertical profiles, flux-profile functions for ozone could not be derived with the same confidence. However, a significant difference with the observed flux-profile functions for heat could not be detected. Fluxes of ozone calculated according to the gradient method using the derived functions for heat showed very good agreement with the fluxes observed by the eddy correlation method. This result shows that the gradient method can be used with good results over forest when local flux-profile functions are used. 1. INTRODUCTION

To asses the achievement of environmental policy goals regarding the input of acidifying species by atmospheric deposition a monitoring system needs to be available. Such a monitoring system is now being developed for sulphur dioxide and ammonia in the Speulderbos, a Douglas fir stand in the centre of the Netherlands [ 11. The deposition will be monitored by the micro-meteorological gradient method. This indirect method is used because monitors fast enough to be used in the eddy correlation method are not available. The deposition flux is determined from measurements of the concentration of gases at several heights above the forest and a turbulent diffusion coefficient k2 Over low vegetation this diffusion coefficient can be determined from empirical flux-profile functions given in the literature. Because of the large roughness and the use of towers barely extending above the canopy the situation is more complex over forest. Measurements may take place in the so-called roughness layer where large deviations from classical flux-profile functions may be expected. To support calculations of the flux from the concentration gradient observed in the Speulderbos, local fluxprofile functions are required.

22 To determine the flux-profile functions over Speulderbos an automatic system capable of measuring the flux and the concentration gradient of ozone was operated continuously in a 36 m tower for nearly eight months in 1993. From these measurements flux-profile functions for heat and ozone were derived. In this paper the experimental set up is described and the results of the measurements are reported. 2. THEORY

The eddy correlation method is considered a reference method to measure the fluxes of trace gases to the surface. The average flux is equal to the covariance of the vertical component of the wind velocity ( w ) and the air concentration (c):

In order to measure the contribution of all eddies to the flux, fast response sensors are required. These are often not available for important trace gases such as sulphur dioxide and ammonia. Therefore gradient methods are often applied. Principle to the gradient method is the flux-gradient assumption. Similar to Ficks law the flux Fi of a component i is calculated from:

The deposition velocity vd is equal to: -Fi/ci. ci is the concentration at a reference height z=h-d. Where h is the height above ground and d is the so-called zero displacement height. k is von Karman’s constant (taken equal to 0.4), u* the so called friction velocity and @C(dL) is the dimensionless flux-profile relation. L is the Monin Obukhov length scale defined in [Z] When gas fluxes are measured the flux-profile functions are often taken equal to the fluxprofile function for heat @h. Empirical values for @ j are given in the literature [Z]. In principle the functions for trace gases could be different to those of heat although field experiments over grassland [3] showed reasonable agreement between turbulent exchange coefficients for heat and ozone. Over forest such measurements have not been reported. Especially the displacement height for gas could be different from the one for heat. Therefore the objective of this study was to determine @functions for a trace gas in Speulderbos. In an earlier study carried out in 1988 and 1989 in the framework of the acidification programme local flux-profile functions ( @ j ) were used together with a displacement height of 11.5 m. These functions were derived from measurements carried out at the site [4]. A height dependent correction factor a was used to correct the flux-profile functions @ for heat given by Dyer and Hicks [ 5 ] :

23

Since then the forest has grown by a few metres. As a consequence the values for the zero plane displacement height d will have changed.

a and

3. METHODS 3.1.

Descriptionof the site

The measurements were carried out in a roughly 30 year old Douglas fir stand [2]. The stand is homogeneous of an area of 2.5 ha. It is surrounded by oak and larch. The stem density is nearly 800 stems per hectare. The height of the trees was about 18 to 20 metre. The one-sided leaf area index varies over the years between 10 and 17 [6]. A diagram of the site is given in [2]. 3.2.

Instruments

The experimental set up is schematically presented in Figure 1. Central to the instrumentation is a sonic anemometer, a Kayo Denki DAT 310 with TR61 probe mounted at the 30 metre level on a boom extending some 3 metre from the tower on the South-West side. The fast response ozone monitor (7) is mounted on a smaller boom in a way that the air inlet was located 25 cm from the sonic centre. At 24, 26.5, 30 and 35 metre above the ground high accuracy temperature sensors (8) are mounted on smaller booms. Relative humidity of air is obtained from a Vaisala instrument with a capacitive sensor. Radiation instruments for net radiation (Schenk 8110), Global radiation (Li-Cor) and Skye SKP 215 sensor for Photosynthetically Active Radiation (PAR) are mounted on the 30 metre level. Air is drawn from the inlets at the same heights as the temperature sensors to a central manifold located near the instruments at the 28 m level. Using a computer controlled valve system the height from which air is drawn through the manifold can be selected. The ozone concentration is determined using a Bendix 8002 with ozone detection based upon the chemiluminescent reaction of ozone with ethylene. Using this set up the air concentration can be determined four times at each height within each cycle of 20 minutes. In [9] and [lo] several tests with the set-up are described. An experiment with all tubes sampling from the same height showed that systematic differences in the concentration observed with the different tubes were not detectable and less than 0.2%. Typically this leads to maximum errors in the deposition velocity of 0.35 m d s . Random fluctuations in V d however can be as high as a few W s .

24 3.7,rr

P I 3u rn 76 5

4

A.1

11'

4 I

1 '

Blm

n l

lq

--

-

._.-.

-.----

Re'erence fillerhead I;rau,rllt f ~ i i e rr k d r i

PT100

EMsOnielrmanrlw ptFA5 3fvrie rr,vniIvr

h-

Nel radla! mi

PAR Glcoal 'adlallcn Valvss system

a DATA

f.IUI1IiUIC

Rmp

AGJISITIOV

3.3.

Calculations of the flux-profile functions from data

For every 20 minute interval the flux-profile functions @ for heights 1 and 2 were calculated from the eddy correlation fluxes and gradients.

with 0, the potential temperature calculated from the observed air temperature T as 0, = T + ywith ythe dry adiabatic lapse rate. In order to use only good quality data measurements in the wind sector 3 15" to 45" were excluded because the measured turbulent parametres may be affected by the tower. Measurements during sunrise and sundown were also excluded. Only cases with a monotonously increasing or decreasing temperature gradient and with the absolute value of the heatflux larger 18 W/m2 were considered. The ozone flux observed using the eddy correlation method was corrected for spectral response according to [l I]. Using an experimentally determined response time for the ozone monitor of 0.1 sec the average correction was only 4% with maxima up to 8%. The flux-profile function for ozone was only calculated when the following criteria were met: the flux of 0 3 was larger than 0.1 ppb.m.s-1. the standard deviation in the 0 3 concentration measured at one height was smaller than 5%. the concentration of 0 3 is larger than 15 ppb. the error in V d due to instationarity in the 0 3 concentration is smaller than 5 r r d s [see 121.

25 The purpose of applying these additional criteria is to reduce the noise level in the calculated flux-profile functions. There is no indication in the results suggesting that the above criteria have a systematic influence on the results. 4. RESULTS AND DISCUSSION

All instruments were operated from December 1992 to September 1993. For several different reasons a large fraction of the data obtained in the wintertime appeared to be unreliable. For the analysis presented here only the data from the period April 1993 to September 1993 were used. Roughly 2500 successful twenty minute measurements of the eddy correlation flux were available. The gradient measurements yielded nearly 3500 successful runs. 4.1.

General

Figure 2 shows a typical result of the eddy correlation measurements on July 9 and 10, 1993. A diurnal cycle of the canopy resistance R,, calculated as in [4], is clearly detectable with values going down to around 70 s/m during the day and values of 500 s/m at night. The actual results are plotted without any smoothing in order to give an impression of the good quality of the data. This diurnal cycle is probably linked with uptake of 0 3 by stomata.

800

t I

0.00

I

I

0.00

I

I

I 0.00

Time (h)

Figure 2. The canopy resistance to uptake of 0 3 calculated from eddy correlation measurements at Speulderbos (July 9 and 10, 1993). The dependency of the canopy resistance on PAR is illustrated in Figure 3a. Figure 3b shows the dependency of the canopy resistance of the vapour pressure deficit. Especially the dependency of the vapour pressure deficit is quite strong, although it is important to realize

26 that several cross-correlations between the air temperature, PAR and vapour pressure deficit exist. Therefore, a more detailed interpretation is required. In an earlier study carried out in Speulderbos the effect of several parameters on the stomata1 resistance to water vapour, calculated from eddy correlation measurements, was investigated. A strong influence of vapour pressure deficit and radiation and hardly an effect of canopy temperature was also observed [ 191. -1 Rc03(s.m )

R,

o3 (s.m")

300

300

250

250-

200 150 100- fl

.' 500

rn

. .... . . .

150-

loo-,

-

50 . . ' I . . . , . . . 1 ' ' . , ' ' . , . . .

.

-

200

0

..

.m

m .

. . . . . . . . . . . . . . . . . . . . . . .

The average ozone flux was around 0.15 ppb m/s with a standard deviation of 0.12 ppb m/s. The average 0 3 concentration was equal to 35 ppb. The average deposition velocity was

7 mm/s. The median canopy resistance calculated from the measurements was 150 s/m. A more detailed statistical treatment of the date is given in [lo]. These results fit quite well with literature data. Greenhut [ 131 reports aircraft measurements over coniferous forest in southern New Jersey with an average canopy resistance of 50 to 400 s/m. Lenshow [14] reports a canopy resistance of 50 s/m. Using the eddy correlation method Wesely [15] found values varying between 150 and 400 s/m with values up to 1500 s/m at night. 4.2.

Flux-profile functions

A specific averaging procedure was used to calculate flux-profile functions. After all selection procedures around 500 twenty minute cases were left. The arithmetic average of

27 these 0 functions for ozone in a certain stability interval iYZ. shows too much scatter. The large scatter is caused by the magnitude of the concentration gradient as well as the magnitude of the fluxes. Especially at (near) neutral conditions (windy, overcast) when the ozone concentration is low the deposition flux becomes very small. Only between 24 and 35 metre the difference can be observed well above the detection limit. In order to reduce the large effect of outliers a robust statistical treatment is required. Therefore several averaging procedures were compared [lo]. The larger temperature data set showed less scatter and was used as a database to test these procedures. Best results were obtained using a robust statistical method proposed by the Analytical Methods Committee [17]. With this method the average is calculated on the basis of the 50 and 75 percentile values thereby minimising the influence of outliers. For heat the difference in the results obtained using the various procedures appears to be small. For 0 3 only the robust methods gave useful results. Based on the similarity in transport mechanisms it was assumed that the procedures used for heat could be applied for 0 3 with confidence as well. Figure 4 shows the flux-profile functions calculated for heat for the heights 30 to 35 m. A displacement height of 15 m was assumed as 75% of the height of the trees [16]. It is important to realise that the zero displacement height was now chosen to be 15 metre rather than the 11.5 metre chosen earlier. This difference is related to the growth of the forest over this period the forest which was between 0.6 to 0.9 m per year and therefore some 3 m over these years [6]. It appears that the functions can be described quite well with small corrections to the existing flux-profile functions. The values of the @h function calculated using classical equations (3) and (4)with a equal to 0.9 are also plotted. It was noted that a slightly better comparison between the observations and the functions could be obtained when the coefficient@ in the flux-profile functions was taken to be 7 rather than 5. Similar results are also reported by Bush [2]. Table 1 shows the data for the other height intervals as well. The correction factors compare quite well with the results found in the earlier study [4]. As was observed earlier the deviation from the original functions increases when the canopy is approached. Table 1 Values of a to correct flux-profile functions as in equations (3) and (4)for heat for different . a displacement hei ht of 15 m the effective height t e f f height intervals hl and h ~Assuming for which the 0.2 ; in other words, a value of r~', which is significantly larger than 0 could not only indicate that the surface resistance is larger than 0, but also that the component could be present in the surface. NH 3 is present in many surfaces. In manure or recently fertilized soil large concentrations occur and c~* will be much larger than c,, resulting in a net emission. Also in the sea c~* is not zero and can sometimes be larger than c, (Quinn et al., 1988a,b; Quinn et al., 1990; Asman et al., 1994a). In agricultural crops there NH3 is also present in the surface (see Farquhar et al., 1983 and Schj0rring, 1991 for a review) and emission can occur. NH 3 is an important intermediate in the photorespiratory N cycle, in the conversion of nitrate to amino acids, and in the breakdown of proteins. The compensation point may vary considerably during the life cycle of plants. It seems to be especially high during senescence (Harper et al., 1987; Parton et al., 1988; O'Deen, 1989; Morgan and Parton, 1989). High concentrations can also occur during grainfilling and after anthesis. Compensation points of 1-5 pg NH 3 m 3 in air are not uncommon for agricultural crops in northwestern Europe (Schjcrring, 1991). Semi-natural ecosystems often show very low compensation points (0.01 lag NH3 m 3 in air; Sutton et al., 1992a,b) Langford and Fehsenfeld (1992) found a somewhat higher compensation point of 0.1 lag NH 3 m 3 in air in a forest in Colorado. Quinn et al. (1990) found a surface concentration ranging from 0.05 - 0.36 lag NH 3 m 3 (average 0.17) for seawater in the North Pacific. Asman et al. (1994a) found an average surface concentration for the North Sea of about 0.27 lag NH 3 m 3. The compensation point in manure or intensively grazed pastures is often so high, that the airborne concentration has no influence on the flux. Most of the NH 3 emission originates from low-level sources (ground-level, stables), which results in rather high concentrations close to the earth's surface. The concentration decreases rapidly with distance due to atmospheric mixing. The flux to the earth's surface close to the source is often high. In this way more than 20% of the NH 3 emission can be dry deposited within a few kilometres from the source (Asman and van Jaarsveld, 1992). This situation differs from that for SO 2, which originates mostly from high-level sources for which dry deposition cannot take place until atmospheric turbulence has mixed the emissions down to the surface - this takes a considerable distance to achieve. A reasonable estimate of the exchange velocity of NH 3 can be found by assuming a surface roughness for momentum of 0.3 m and a surface resistance of 30 s m 1. This would give a dry deposition velocity of 22 mm s1 for a windspeed of 5 m s ~ and a surface concentration of 0. This would lead to a removal rate due to dry deposition of about 7% h -~ if the NH 3 is homogeneously mixed over a 1000 m high mixing layer. Under the same conditions, assuming a laminar boundary layer resistance of 600 s m -1 a removal rate of 0.6% h -1 can be found for NH4 +.

62 It would be a good idea to model the surface exchange of NH 3 from the exchange velocity and a surface concentration. In this way both emission and dry deposition could be modelled. This is, however, possible only if the surface concentration were known. The past ten years has given us much information on the exchange velocity of NH 3 on land. More information is needed on the exchange velocity at sea, the surface concentrations of NH 3 (vegetation, sea) and its temporal variation, and on the dry deposition velocity of particulate NH4§

4. WET DEPOSITION Components can be removed by different wet deposition processes. There exist removal processes within clouds (in-cloud scavenging) and removal processes below the cloud base (below-cloud scavenging), where components are removed by falling raindrops and snowflakes. Cloud- and precipitation water are usually acidic. Consequently, most of the NH 3 taken up by the drops reacts with H § to form NH4§ It is therefore possible to distinguish between the contribution of NH 3 and particulate NH4§ only if models are used. NH 3 is a highly soluble gas. Cloud droplets are so small (about 10 tam) that they take up NH 3 rapidly. Almost all NH 3 is found in the cloud droplets after the few seconds it takes to achieve equilibrium with NH 3 in the surrounding air. The NH 3 concentration in the interstitial air has then become very low. NH4§ aerosol acts as condensation nucleus, i.e. that water vapour condenses onto aerosols when the air becomes saturated with water vapour. In this way almost all NH4§ in clouds will become part of cloud droplets. NH 3 and NH4§ are transferred rapidly to the cloud droplets, but this does not necessarily lead to their removal from the atmosphere. The removal of NH x from the atmosphere by in-cloud scavenging is therefore more determined by the dynamical and physical processes that result in precipitation formation. Raindrops of typically a radius 500 lam and are much larger than cloud drops. The time they need to fall from the cloud base to the surface is relatively short (a few minutes). Raindrops are so large that the transport of airborne NH 3 to the drops is not fast enough for equilibrium with the surrounding air to be reached before they strike the surface. Consequently they will take up NH 3 after collection, unless contact with the surrounding air is avoided. For this reason the NH4§ concentration in precipitation in agricultural areas is often too high. This is caused by uptake (dry deposition) of NH 3 to the wetted funnel of bulk collectors, which are not closed during dry periods. NH4§ aerosol is not captured very well by falling raindrops. Usually not enough information is known to model dynamical and physical processes in clouds in atmospheric transport models. Moreover, if such information were available it would often take too much cpu-time to perform the necessary calculations. For these reasons scavenging is often modelled by using so called "scavenging coefficients". The change in airborne concentration due to scavenging is then described by: c, -- c,.o e -xt

(4)

where %0 is the concentration in the air at the onset of the precipitation (mol m-3), ~, is the scavenging coefficient (s 1) and t=time (s).

63 Scavenging coefficients are used to describe the effect of in- and below-cloud scavenging separately, or sometimes an overall scavenging coefficient is used to describe the effect of both processes together. Scavenging coefficients apply to gases as well as aerosols. A general function for the scavenging coefficient is: ~, = a I b

(5)

where I is the precipitation rate (mm hl). The scavenging coefficient increases with precipitation rate. For the cloud volume (in-cloud scavenging) a is about 4x10 4 and b about 0.64 for both NH 3 and NH4§ For the below-cloud volume (below-cloud scavenging) a is about 9.9x10 5 and b about 0.62 for NH 3 (Asman, 1994b). The below-cloud scavenging coefficient of particulate NH4+ is not well known, but is less than 10.5 s1 (Pruppacher and Klett, 1978). Below-cloud scavenging occurs in the lowest few hundred metres of the atmosphere (the average cloud base height during precipitation in northwestern Europe is about 300-400 m), whereas in-cloud scavenging takes place in a much larger volume. Below-cloud scavenging is less efficient than in-cloud scavenging, both for NH3 and NH4§ aerosol. Despite the much larger concentration of NH 3 near the surface in areas where net-emission is occurring (F_~sman et al., 1988), in-cloud scavenging contributes the largest fraction to the NH4§ concentration in precipitation. Computations for Denmark show that the contributions of the different processes to the NH4§ concentration in precipitation is." in-cloud scavenging of NH 3 15%, in-cloud scavenging of NH4§ 77%, below-cloud scavenging of NH 3 6% and belowcloud scavenging of NH4§ 2% (Asman and Jensen, 1993). During precipitation periods NH 3 and NH4 + are removed very efficiently (on the order of 75% h l) from the atmosphere, much more so than the removal due to dry deposition. But as precipitation occurs only 5-10% of the time in northwestern Europe, the total amount of NI-t~ wet deposited is not necessarily larger than the total amount dry deposited. Maps with the annual wet deposition of NH~ in Europe are presented in Buijsman and Erisman (1988) and Schaug et al. (1993). The wet removal of NI-Ix in statistical transport models can be modelled well if information is known on precipitation statistics (length of dry and wet periods). In other transport models it is necessary to know the same type of information, i.e. the fraction of the area that is exposed to precipitation if the whole area is supposed to be wet when precipitation is collected in one sampler. Usually only part of an air mass is exposed to precipitation. In this part NH x will be removed almost entirely after one hour of exposure. The concentration in the dry part will not change much and can be transported to other areas. More information on this type of precipitation statistics is needed.

5. REACTION In northwestem Europe most NH 3 reacts with acid aerosols that contain sulphuric acid (H2SO4). This reaction has been investigated in the laboratory (Robbins and Cadle, 1958; Baldwin and Golden, 1979; Huntzicker et al., 1980; McMurry et al., 1983). The reaction proceeds rapidly at high relative humidity and will take only a few seconds. The reaction rate is limited by the rate at which NH 3 diffuses to the acidic particle. At low relative

64 humidity the reaction is slower because only 10-40% of the collisions of NH 3 with a particle lead to a reaction (Huntzicker et al. 1980; McMurry et al., 1983). The reaction proceeds faster when the particles involved are small, because the diffusion is then faster. The pseudofirst-order reaction rate decreases with the degree of neutralization of the particle. Moreover, the reactions by which H2SO4-containing aerosols are formed could also limit the uptake of NH 3 if all of the acid is already neutralized. Not enough is known at present about these possibilities to quantify the reaction rate that actually takes place. A minor part of NH 3 reacts with gaseous nitric acid (HNO3) and gaseous hydrochloric, acid (HC1) to form particulate NH4NO 3 or NH4C1, which is part of the aerosols that contain other components (Stelson et al., 1979; Stelson and Seinfeld, 1982a,b,c; Pio and Harrison, 1987; Allen et al., 1989). Unlike the reaction with H2SOn-containing aerosol, which is a one-way reaction, these reactions can occur in both directions" N H 3 + H N O 3 ~ NH4NO 3

and

NH~ + HC1 ~ NH4C1

(6)

The fact that these reactions are in equilibrium means that if the concentration of one or two of the gaseous components becomes very low the component in aerosol form will dissociate. NH 3 can also react with OH, O and O(~D). Levine et al. (1980) found that the reaction with OH was most important. When adopting a constant and relatively high OHconcentration of 4x 106 molecules cm 3 (Logan et al., 1981) a pseudo-first-order reaction rate of 5.4x10 7 s ~ for NH 3 is found. This value is much lower than the pseudo-first-order rate for the reaction with H2SO4-containing aerosol, HNO3 and HC1 (see below). These reactions are therefore usually neglected in regional transport models. It is also possible to estimate the reaction rates from field measurements, but it is then necessary to make several assumptions, which may not always be valid. Lenhard and Gravenhorst (1980), Erisman et al. (1988) and Harrison and Kitto (1992) measured a pseudofirst order reaction rate of NH 3 between 10 4 and 10 .6 s1. The reaction rate can also be estimated by changing the rate in atmospheric transport models until the best agreement is obtained with measured concentrations of NH 3 and NH4+ in air and NH 4 in precipitation. Asman and Janssen (1987) found in this way a pseudo-first-order reaction rate of 8x10 s s ~. A value of this order gives good results even for annual average concentrations on a global scale according to Dentener and Crutzen (1993), who include all relevant processes in a more sophisticated way. This reaction rate is about 30% h ~, which is much greater than the oxidation rates of sulphur dioxide (SO2; of the order 1% h 1) or nitrogen dioxide (NO2; of the order 4% hl). It should be noted that the rate derived in this way is an annual average value. The rate may well show diurnal and seasonal variations which may depend on the local meteorological and chemical conditions. As a consequence, even the annual average reaction rate may show geographical variations. The dry deposition velocity of NH 3 is large compared to that of particulate NH4§ A high reaction rate, such as has been found, favours particulate NH4+ over NH 3. As a consequence, the dry deposition velocity of NH x as a whole is lower than if the reaction rate were lower. The lower dry deposition velocity of NH4§ promotes also long-range transport of NH x. It is rather unsatisfactory that no more information is known on the reaction rate of NH 3. More information on the reaction mechanisms and the temporal and spatial variability of the reaction rate is needed.

65 6. M O D E L L I N G A historical overview of the modelling of NI-I~ is given by Asman (1994a). Most of the NH3 is emitted from a large number of scattered low-level sources (ground-level, stables). As a consequence the NH3 concentration shows an extremely high spatial variability. Thousands of stations are needed to get a reliable average concentration for a country. This of course is not possible and that is why models are urgently needed in this case. But a good spatial resolution in a model can be achieved only if there is an emission inventory with a high spatial resolution is available. The results of a model can then be verified with measurements at a limited number of stations in areas with different emission densities (Asman and van Jaarsveld, 1992).

::..

"~..i:~i-~-.-...

'.':

. . . . '

.

:

"..':'

:.'"

9

iii~i.~iii~ii!!~ii:iii'iiiii

_

.:::::-

9

" : ;

:!'I!~!!~!!M .......... ........... : -..-::::: 9. . . . . . . . 9

9

.

.

.

.

.

.

.::.:~.:! . . . . . . . . . . . . . . . .

.

i

.

.

....

..

- .

~9. : : : .....

..

:

8|

iiiiiiiii !iiiiiiiii !iiiiiiiJi utiJiiiil l/ (

200

(

400

(

800

(

1600

(

3200

Figure 2. Modelled total deposition of NI-I~ in Europe (mol ha 1 year1); 1000 mol ha 1 a1 = 14 kg N ha ~ year a. (Reprinted from Asman and van Jaarsveld, 1992. Copyright 1992, with kind permission from Elsevier Science Ltd., The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK).

66 Models are also needed to calculate import/export balances for areas. The reaction product NH4+-containing aerosol is transported over long distances. This results in concentrations in air and precipitation (the contribution of NH4§ aerosol to precipitation is larger than that of NH 3) which do not show a high spatial variability. This means that a good impression of the concentration field can be obtained from measurements. Normally atmospheric transport models are developed either for short ranges (< 10 kin) or for long ranges (500 - 2000 kin). A transport model for NH~ should be capable of both, at least if realistic NH 3 concentrations and dry depositions are needed. The average value of the dry deposition of NH 3 in a grid element in a model (e.g. a 150x150 krn 2 area in the EMEP model) is not incorrect. It is, however, not representative of the deposition of a nature area that covers only part of the grid element. Some attempts have been made to correct the EMEP model outputs for differences in deposition to ecosystems within the 150x150 km 2 grid element used (Hettelingh et al., 1991). This is done by redistributing the deposition within the grid element in such a way that the total deposition to the whole grid element does not change. In this way mass is conserved (see also discussions in Sutton et al., 1993b), but due to the local character of NH 3, this approach is unlikely to give correct results near important emission areas, i.e. those areas where dry deposition of NH 3 is larger than wet deposition of NI-Ir Models can also be used to compute the import or export of NH~ to or from countries. It is almost impossible to infer such transboundary fluxes from measurement data. They can also be used to estimate the contribution of in- and below-cloud scavenging of NH 3 and NH4§ aerosol to the wet deposition of NH~. This cannot be derived from measurements. Moreover, models can be used to estimate historical or future depositions by using emissions representative for these periods. In this case one should check that the emissions of other components like SO 2, NO~ and hydrocarbons have not changed so much that the reaction rate of NH3 to particulate NH4§ is altered considerably (Asman et al., 1988). Most models are able to reproduce measured concentrations and depositions reasonably well, except for NH 3, which can be handled by only a few models and then not even very well. Figure 2 shows the total NI-I~ deposition in Europe (sum of wet and dry deposition of NH3 and NH4§ Model results show that in northwestern Europe 44% of the emitted NH 3 is dry deposited as NH3, 6% is wet deposited as the contribution of NH3 to the wet deposition of NI-I~, 14% is dry deposited as NH4§ aerosol and 36% is wet deposited as the contribution of NH4§ aerosol to the wet deposition of NI-I~ (Asman and van Jaarsveld, 1992). There is a need for atmospheric transport models for NI-I~ that have a spatial resolution that is sufficient to calculate realistic NH 3 concentrations. Such models should also be able to take into account that the dry deposition velocity differs from one surface to another, as otherwise e.g. no realistic dry deposition of NH 3 to forests can be calculated.

7. CONCLUSIONS The most important conclusion which can be drawn from the information presented here is that deposition of NHx takes place mainly in two forms, namely dry deposition of NH3 close to the source and wet deposition of NH~ at larger distances from the source contributed by the NH4§ aerosol. This indicates that the possibility exists of reducing the deposition of NH~ to nature areas close to those with a high NH3 emission density. This can be done to

67 some extent by selectively reducing NH 3 emissions close to these areas. This strategy does not work at long distances from important source areas, however. In these areas deposition of NH x can be reduced only by cutting down all the emissions in a much larger area. A policy to reduce the deposition of NI-I~ should also take into account the contribution of NOx and its reaction products to the total nitrogen deposition, as well as protection of other parts of the environment (soil, ground water) from emissions when atmospheric emissions are reduced. There exist good technical possibilities for reducing NH 3 emissions. A reduction can be obtained by altering the factors which lead to high emissions: a. Reducing the nitrogen content of the animal food in such a way that optimal nutrition is obtained for the stage of development of the animal under consideration. Nowadays the same food is often used for animals of almost all ages. b. Adding some amino acids to the animal food for non-ruminants (pigs etc.), so that no overdoses of other amino acids are needed. c. Prescribing housing and storage systems which give the lowest losses. There exist considerable differences in the emission rate per animal for different housing and storage systems. d. Ploughing the manure under as soon as possible after spreading or injecting the manure in the ground in the case of grassland. f. Spreading of manure and application of fertilizers under meteorological conditions (low temperatures, just prior to the onset of precipitation) which favour low emissions. Apart from these measures, emissions of NH 3 can be reduced by taking more technical measures, such as biofiltring of the air coming from stables. These measures are more expensive, however. But it is most important that the nitrogen cycle for a country, region or farm be as balanced as possible. This means a sharply reduced import of nitrogen-containing animal food, limitations to the number of animals ha 1 and in general, eliminating waste of such a precious element as nitrogen, leading to fine-tuning of the nitrogen supply to match the nitrogen demand of the crops. The conversion of nitrogen from manure and fertilizers into plant products is rather efficient (70+10%; Isermann, 1993). This efficiency can be increased by the adding of amino acids to the food, as mentioned earlier. The fact that humans in western Europe eat (animal) protein far in excess of their requirements than they need, leads also to a high production of animal proteins with all the ill consequences for the environment. Reduction of the share of animal proteins in the human diet could therefore also have beneficial environmental consequences. Although much more is known about the atmospheric behaviour of NI-I~ than 10 years ago, far from all essential information is known. Good progress has been made on the exchange velocity of NH 3 and also on the development of methods for continuously measuring NH 3. No models are, however, presently available that give good results for dry deposition of NH 3 to forests. Further information is needed on: a. Geographical differences in emission factors. b. Diurnal and seasonal variation in the NH 3 emission. c. Emission inventories for some countries with a high or medium emission density are needed. d. Emissions from plants and cities and other "minor sources". e. Exchange velocity of NH3 at sea. f. Surface concentration of NH 3 for vegetation and seawater and its temporal variation.

68 g. The dry deposition velocity of particulate NH4§ h. Precipitation statistics needed in atmospheric transport models. i. The reaction of NH 3 to NH4+: mechanisms, rate and its temporal and spatial variation. Moreover, atmospheric transport models for NI-I~ should be developed that have a spatial resolution that is sufficient to calculate realistic NH 3 concentrations and the possibility of having different dry deposition velocities for different surfaces.

8. REFERENCES

Allen, A.G., Harrison, R.M. and Erisman, J.W. (1989) Atmospheric Environment 23, 15911599. Andersen, H.V., Hovmand, M.F., HummelshOj, P. and Jensen, N.O. (1993) Atmospheric Environment 27A, 189-202 (1993). Asman, W.A.H. (1992) Report no. 228471008, National Institute of Public Health and Environmental Protection (RIVM), Bilthoven, The Netherlands. Asman, W.A.H. (1994a) Nova Acta Leopoldina NF 70, No. 288, 263-297. Asman, W.A.H. (1994b) Parameterization of below-cloud scavenging of highly soluble gases under convective conditions. Paper submitted to Atmsopheric Environment. Asman, W.A.H., Drukker, B. and Janssen, A.J. (1988) Atmospheric Environment 22, 725-735. Asman, W.A.H., Harrison, R.M. and Ottley, C.J. (1994a) Estimation of the net air-sea flux of ammonia over the southern bight of the North Sea. Paper accepted for publication in Atmospheric Environment. Asman, W.A.H. and Janssen, A.J. (1987) Atmospheric Environment 21, 2099-2119. Asman, W.A.H. and Jensen, P.K. (1993) Report 26 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen, Denmark. Asman, W.A.H., Sr L.L., Berkowicz, R., Granby, K., Nielsen, H., Jensen, B. and Runge, E. (1994b) Report 35 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen. Asman, W.A.H. and Van Jaarsveld, J.A. (1992) Atmospheric Environment 26A, 445-464. Baldwin, A.C. and Golden D.M. (1979) Science 206, 562-563. Bobbink, R., Boxman, D., Fremstad, E., Heil, G., Houdijk, A. and Roelofs, J. (1992) In: Grennfelt, P. and Th6mel6f, E.: Critical loads for nitrogen, Report Nord 1992:41, Nordic Council of Ministers, Copenhagen. Buijsman, E. and Erisman, J.W. (1988) J. Atmos Chem. 6, 265-280. Buijsman, E., Maas, J.F.M. and Asman, W.A.H. (1987) Atmospheric Environment 21, 10091022. Cass, G.R., Gharib, S., Peterson, M. and Tilden, J.W. (1982) Open File Report 82-6. Environmental Quality Laboratory, California Institute of Technology, U.S.A. Davidson, C.I. and Wu, Y.-L. (1990) In: Lindberg, S.E., Page, A.L. and Norton, S.A. (Eds.) Acidic precipitation. Vol. 3. Sources, deposition, and canopy interactions, Springer, New York, U.S.A., 103-215. Dentener, F.J. and Crutzen, P.J. (1993) A three dimensional model of the global ammonia cycle. Paper submitted to J. Atmos. Chem.

69 Duyzer, J.H., Bouman, A.M.H., Diederen, H.S.M.A. and Van Aalst, R.M. (1987) Report R 87/273, TNO Division for Society, Delft, The Netherlands. Duyzer, J.H., Verhagen, H.L.M., Westrate, J.H. and Bosveld, F.C. (1992) Environ. Pollut. 75, 3-13. Erisman, J.W., Vermetten, A.W.M., Asman, W.A.H., Waijers-Ypelaan, A. and Slanina, J. (1988) Atmospheric Environment 22, 1153-1160. Erisman, J.W. and Wyers, G.P. (1993) Atmospheric Environment 27A, 1937-1949. Farquhar, G.D., Wetselaar, R. and Weir, B. (1983) Gaseous nitrogen losses from plants. In: Freney, J.R. and Simpson, J.R." Gaseous loss of nitrogen from plant-soil systems. Nijhoff, The Hague, The Netherlands, 159-180. Fowler, D. and Duyzer, J.H. (1989) In: Andrae, M.O. and Schimel, D.S. (Eds.): Exchange of trace gases between terrestrial ecosystems and the atmosphere, John Wiley, New York, U.S.A., 763-773. Harper, L.A., Sharpe, R.R., Langdale, G.E. and Giddens, J.E. (1987) Agron. J. 79, 965-973. Harrison, R.M. and Kitto, A.-M.N. (1992) J. Atm. Chem. 15, 133-143. Hettelingh, J.-P., Downing, R.J. and De Smet, P.A.M. (1991) Technical Report No. 1, Report 259101001, National Institute of Public Health and Environmental Protection, Bilthoven, The Netherlands. Huntzicker, J.J., Cary, R.A. and Ling, C.-S. (1980) Environ. Sci. Technol. 14, 819-824. Hutchinson G.L., Millington R.J. and Peters D.B. (1972) Science 175, 771-772. Htifken, K.D., Meixner, F.X. and Ehhalt, D.H. (1983) In: Pruppacher, H.R., Semonin, R.G. and Slinn, W.G.N. (Eds.) Precipitation scavenging, dry deposition, and resuspension, Elsevier, New York, U.S.A., 825-835. Isermann, K. (1990) In: Ammoniak in der Umwelt, KTBL, Darmstadt-Kranichstein, F.R.G., 1-1 to 1-76. Isermann, K. (1993) Nahrstoffbilanzen und aktuelle N~.hrstoffversorgung der BSden. RobertBosch-Stiftung GmbH, 5. Koloquium zur Bodennutzung und Bodenfruchtbarkeit: "N~flarstoffhaushalt - Kenntnisstand und Forschungsliicken", Schw~ibisch Hall, 21-22 November 1991, Sonderband Berichte iiber Landwirtschaft. Langford, A.O. and Fehsenfeld,F.C. (1992) Science 255, 581-583. Larsen, S.E., HummelshCj,P. Jensen, N.O., Edson,J.B., De Leeuw, G. and Mestayer, P.G. (1994) Report 47 Danish Sea Research Programme 90, Danish Environmental Protection Agency, Copenhagen, Denmark. Lee, D.S., Nason, P.D. and Bennett, S.L. (1992) Report AEA-EE-0328, AEA Environment and Energy, Harwell Laboratory, U.K. Lenhard, U. and Gravenhorst, G. (1980) Tellus 32, 48-55. Levine, J.S., Augustsson, T.R. and Hoell, J.M. (1980) Geophys. Res. Lett. 7, 317-320. Lindfors, V., Joffre, S.M. and Damski, J. (1991) FMI Contribution No. 4, Finnish Meteorological Insitute, Helsinki, Finland. Logan, J.A., Prather, M.J., Wofsy, S.C. and McElroy, M.B. (1981) J. Geophys. Res. 86, 72107254. McMurry, P.H., Takano, H. and Anderson, G.R. (1983) Environ. Sci. Technol. 17, 347-352. Morgan, J. A. and Parton, W.J. (1989) Crop. Sci. 29, 726-731. Nilsson, J. and Grennfelt, P. (1988) Report 1988:15, Nordic Council of Ministers, Copenhagen, Denmark. O'Deen, W.A. (1989) Agron. J. 81,980-985. Pacyna, J.M., Larssen, S. and Semb, A. (1991) Atmospheric Environment 25A, 425-439.

70

Patton, W.J., Morgan, J.A., Altenhofen, J.M. and Harper, L.A. (1988) Agron. J. 80, 419-425. Pio, C.A. and Harrison, R.M. (1987) Atmospheric Environment 21, 1243-1246. Pruppacher, H.R. and Klett, J.D. (1978) Microphysics of clouds and precipitation. Reidel, Dordrecht, The Netherlands. Quinn, P.K., Bates, T.S., Johnson, J.E., Covert, D.S and Charlson, R.J. (1990) J. Geophys. Res. 95, 16405-16416. Quinn, P.K., Charlson, R.J. and Bates, T.S. (1988a) Nature 335, 336-338. Quinn, P.K., Charlson, R.J. and Zoller, W.H. (1988b) Tellus 39B, 413-425. Robbins, R.C. and Cadle, R.D. (1958) Phys. Chem. 62, 469-471. Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M. and Jansen, J. (1985) Plant and Soil 84, 45-56. Schaug, J., Pedersen, U., Skjelmoen, J.E. and Kvalv~tgnes, I. (1993) Data report 1991. EMEP/CCC-report 4/93, Norwegian Institute for Air Research, LillestrCm, Norway. Schjcrring, J.K. (1991) In: Sharkey T.D., Mooney, H.A. and Hollamd, E.A. (Eds.): Trace gas emissions by plants. Academic Press, New York, U.S.A., 267-292. Schj0rring, J.K. Kyllingsb~ek, A., Mortenssen, J.V. and Byskov-Nielsen, S. (1993) Plant, Cell and Environ. 16, 161-167. Schlesinger, W.H. and Hartley, A.E. (1992) Biogeochemistry 15, 191-211. Stelson, A.W., Friedlander, S.K. and Seinfeld, J.H. (1979) Atmospheric Environment 13, 369371. Stelson, A.W. and Seinfeld, J.H. (1982a) Atmospheric Environment 16, 903-922. Stelson, A.W. and Seinfeld, J.H. (1982b) Atmospheric Environment 16, 993-1000. Stelson, A.W. and Seinfeld, J.H. (1982c) Atmospheric Environment 16, 2507-2514. Sutton, M.A., Asman, W.A.H. and Schjorring, J.K. (1993a) In: L6vblad G., Erisman, J.W. and Fowler, D. (Eds.): Report 1993:573, Nordic Council of Ministers, Copenhagen, Denmark, 127143. Sutton, M.A., Fowler, D., Hargreaved, K.J. and Storeton-West, R.L. (1992b) In: Angeletti, G., Beilke, S. and Slanina, J. (Eds.) Field measurements and interpretation of species related to acid deposition. Air Pollution report 39, Commission for the European Communities, Brussels, Belgium, 211-217. Sutton, M.A., Fowler, D., Smith, R.I., Eager, M. Place, C.J. and Asman, W.A.H. (1993b) Proceedings of the joint CEC/BIATEX workshop, Aveiro, Portugal, May 1993. Commission for the European Communities, Brussels, Belgium, 117-131. Sutton, M.A., Moncrieff, J.B. and Fowler (1992a) Environ. Pollut. 75, 15-24. Sutton, M.A., Pitcaim, C.E.R. and Fowler, D. (1993c) Adv. Ecol. Res. 24, 301-393. Van Breemen, N., Burrough, P.A., Velthorst, E.J., Van Dobben, H.F., De Wit, T., Ridder, T.B. and Reijnders, H.F. (1982) Nature 299, 548-550. Van der Eerden, L.J.M. (1982) Agric. Envir. 7, 223-235. Van Hove, L.W.A., Koops, A.J., Adema, E.H., Vredenburg, W.J. and Pieters, G.A. (1987) Atmospheric Environment 21, 1759-1763. Williams, R.M. (1982) Atmospheric Environment 16, 1933-1938. Wyers, G.P., Vermeulen, A.T. and Slanina, J. (1992) Environ. Pollut. 75, 25-28.

G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.

71

Measurement and modelling of ammonia exchange over arable croplands M.A. Sutton, J.K. Burkhardt, D. Guerin and D. Fowler Institute of Terrestrial Ecology, Bush Estate, Penicuick Midlothian, EH26 0QB, Scotland, UK. Abstract

Micrometeorological measurements of the exchange of atmospheric ammonia over arable land are reported. Measurements were made over bare soil and wheat canopies at early canopy closure, approaching anthesis and during early senescence. Bare soil showed a slow rate of deposition with a canopy resistance (Rc) of around 300 s m -1, while fluxes over young wheat were still dominated by fertilizer urea soil emissions (up to 170 ng m -2 s -1) nearly a month after fertilization. Summer measurements showed that vegetation processes dominated exchange, apart from immediately after fertilization with ammonium nitrate, which contributed around 5 ng m -2 s -1 to the net flux 1-2 days after application. The Summer NH 3 fluxes have been interpreted using a new micrometeorological modelling approach. The R c model is unable to simulate the mechanisms of bi-directional exchange, and a recently developed model that quantifies the 'canopy compensation point' (gc) for NH 3 was applied. In this model net fluxes are predicted as the resolution of competing leaf surface deposition and bi-directional stomatal fluxes. The model provides a simple approach to predict net NH 3 fluxes with the atmosphere, though sometimes underestimates morning NH 3 emissions. A possible explanation of this effect is that the leaf surface behaves as a capacitor for N H 3 adsorption. A revised capacitance model is developed that extends the analysis of the Xc model. The capacitance model is able to reproduce the behaviour of morning emission, though further development of both models is required to provide a description valid over longer periods. 1. I N T R O D U C T I O N

Deposition of atmospheric ammonia (NH3) and ammonium (NH4+) (collectively NHx) is an important contributor to ecosystem acidification and eutrophication (e.g. Grennfelt and Thtirneltif 1992; Sutton et al. 1993d). It is necessary to quantify the inputs of these species to sensitive ecosystems, to be able to assess their environmental impact as well as the contribution of NH x relative to other acidifying inputs. One of the main uncertainties in quantifying total inputs of NH x is the magnitude of the gaseous N H 3 dry deposition term. The uncertainty is emphasized because NH 3 may be both emitted from and deposited to land surfaces. The atmospheric nitrogen inputs to arable land may be small compared with agricultural practice. However, because of the large area extent of arable land in Europe it becomes important to quantify the net fluxes (emission and deposition) in calculating atmospheric budgets and parametrizing atmospheric transport models (e.g. Sandnes and Styve 1992; Asman and van Jaarsveld 1992; Singles et al. 1995). In this way knowledge of crop-atmosphere exchange becomes important in understanding the N inputs to semi-natural ecosystems. Measurements of NH 3 exchange over cereals have often implied that a 'compensation point' exists within plants andis important in controlling net fluxes (e.g. Farquhar et al. 1980; Schjorring 1991; Sutton et al. 1993c). It is known that ammonia plays a major role within the biochemical pathways of plants, so that for a particular intercellular NH4+ concentration, there is an equilibrium atmospheric gaseous N H 3 concentration. The term 'compensation point' is used to reflect the interpretation that this is the concentration at which metabolic consumption processes balance production, while the exchange is viewed as operating via stomata. This is referred to here as thej 'stomatal compensation point' (Zs)- Against this physiological background, it is also known that ammonia is a very soluble gas and is frequently found in micrometeorological experiments to deposit rapidly to semi-natural vegetation and this is believed to be the result of leaf surface sorption processes (Duyzer et al. 1994; Sutton et al. 1993b; Erisman and Wyers 1993). Ammonia emission from a compensation point by therefore be short circuited by deposition to leaf cuticles.

72 Arable croplands to show evidence of both processes operating, resulting in the direction of the net atmospheric flux changing with environmental and plant conditions. In part, this may be related to the higher nitrogen status of these ecosystems, resulting in a larger compensation point and potential to overcome cuticular uptake. Until recently, model descriptions of these processes had taken one of two lines: either to assume a solely stomatal exchange with a compensation point (mostly the chamber experiments); or to treat the exchange with a deposition velocity (Vd) and canopy resistance (Rc) model, originally designed for parametrizing deposition processes (mostly the micrometeorological experiments). A combined model to reconcile these different interpretations was proposed by Sutton and Fowler (1993). In this model, calculation of R c is replaced by its concentration analogue (Zc), referred to as the 'canopy compensation point', the magnitude of which is defined by competition for exchange between the atmosphere, cuticle and stomata. The present paper reports measurements of ammonia surface-atmosphere exchange, made using the aerodynamic gradient technique. Fluxes were measured at several stages during the year, including over bare soil and over different wheat canopies at different stages. In addition to these main measurements, a short chamber study was made to provide an indication of the contribution of soil to the net atmospheric emission from a wheat canopy. Where the plant canopy was established as being the main site of ammonia surface-atmosphere exchange, particular attention was given to investigating the processes controlling the net ammonia flux. The results are interpreted using the canopy compensation point model of Sutton and Fowler (1993) and constraints of the model identified. In particular the possibility that the leaf surface may behave as a capacitor for NH 3 adsorption/dissolution is investigated and an initial framework to parametrize this effect suggested. 2. MICROMETEOROLOGICAL THEORY

A full description of the micrometeorology applied in this study and its restrictions has been given by Sutton et al. (1993a, b), so only an outline description is given here. The flux measurements were made using the aerodynamic gradient method. In this approach the net flux (Ft) (negative fluxes denote deposition) is determined from wind (u) and concentration (Z) profiles with height above a uniform surface with extensive fetch: Ft = _ k 2

du d[ln(z-d)-~M]

dz d[ln(z-d)-~tu]

(1)

where z is height above ground, d is the displacement of ground level due vegetation, k is the von Karman constant (0.41) and ~tM, XCHare corrections for atmospheric stability. As noted in the introduction it is usual to interpret measured fluxes by calculating the deposition velocity (Vd) and canopy resistance (Rc). In principle this resistance approach is designed to describe deposition processes, since it assumes that the concentration at the absorbing surface (~surface) is zero. In this case, by analogy with Ohm's law:

Vd{z-d} = 1/Rt{z-d} = Ft/(~surfac e - ~{z-d}) = -Fg/~{z-d}

(2)

Rt{z-d} = Ra{z-d} + R b + R c

(3)

where R t is the total resistance to deposition, R a the turbulent atmosphere resistance, R b the quasi-laminar diffusion resistance. This approach is convenient since Vd is assumed to be independent of Z, and the flux may be modelled given estimates of R c and Z, together with meteorological information. Nevertheless, it is not possible without manual switches to use the model to predict bi-directional exchange. The alternative stomatal model noted above describes bi-directional fluxes satisfactorily, but ignores any parallel deposition to leaf surfaces. Here Zs is the stomatal compensation point and R s the stomatal resistance: Ft = (ks - Z{z-d})/(Ra{z-d} + Rb + Rs)

(4)

73 180" chamber over bare soil between plants

150" 120"

Flux NH3

90

r

(ng m-2 s-1) 60 30 0 -30

,:'~

160

,o

ii

140

8

!~

12

Air conc (1 m) NH3 (ug m-3)

6

t

i

12o Air conc (1 m) lOO SO2 (ug m-3) so

',

:: i i

NH3

4

i~:

6O

i

40 20 0

' " "'"

0

00:00 02:00 04:00 06:00 08:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00

Time (GMT) 2 April 1993

Figure 1. Micrometeorologicalammonia flux measurements and air concentrations of NH3 and SO2 over a young wheat canopy (14 cm). Measurementsmade 1 month after fertilization with 35 kg N ha-1 as urea. Open squares are chamber measurementsof soil emission. Given the limitations of both these approaches, in later sections the results are used to develop revised resistance models that treat both a compensation point and cuticular uptake. 3. SITE AND M E A S U R E M E N T S The measurements were made over bare soil and canopies of wheat at the field station of the University of Nottingham at Sutton-Bonington, Loughborough, UK. The site used provided extensive fetch of >200 m in most wind sectors and up to 500 m. Measurements were made in the surface layer up to 2 m above the wheat canopy. In May 1992, measurements were made above a wheat canopy (60-70 cm) approaching senescence. On May 20, 35 kg N ha -1 as ammonium nitrate was applied. A second campaign was made in March 1993, on the boundary between a bare field (harrowed soil) and a young wheat canopy (14 cm high). This wheat canopy had been fertilized with 35 kg N ha -1 as urea about 4 weeks prior to the measurements. A third campaign was made in June - July 1993, at the same site, when the wheat was approaching senescence (80 cm high), and sugar beet growing on the previously bare soil. Ammonia concentration profiles were determined using a continuous wet annular denuder system described by Wyers et al. (1993). Measurements were made at two heights, and, to reduce possible bias between denuders, the collected NH 3 was brought (as NH4+ in the stripping solution) to a common detector for analysis. Analysis of the NH4 + in this system is by membrane diffusion of NH 3 at high pH into a counter flow of deionized water, with subsequent measurement by conductivity. Sulphur dioxide air concentrations were also determined though fluxes of SO 2 are not reported here. Windspeed was measured using 6 sensitive cup anemometers (Vector instruments), and temperature and water vapour profiles measured with a fine thermocouple and dewpoint meter system. Further details of the measurement methods and data treatment have been provided by Sutton et al. (1993a). 4. RESULTS Figure 1 shows an example of the emissions detected during the Spring 1993 measurements, with fluxes up to 160 ng NH 3 m -2 s-1 and only limited net deposition occurring at night. Such large emissions are probably related to the fertilization of the soil with urea 1 month prior to the measurements. It is well known that substantial ammonia emissions occur from urea compared with other fertilizers (e.g. Whitehead and Raistrick 1990), however, it is generally

74 25 20 15

Flux NH3 (ng m-2 s-l)

10 5 0 -5 -10 -1520O 180 5

160

Air conc (lm) 4 NH3 3 (ug m-3)

14o Air conc

12o ~oo

(1 m)

SO2 (ug m-3)

6O 4O

"'!.,...':...~ 0 12:00

2O 0

16:00

20:00

O0:00

04:00

08:00

12:00

16:00

20:00

Time (GMT) 21-22 May 1992

Figure 2. Micrometeorological ammonia flux measurements and air concentrations of NH3 and $ O 2 o v e r a wheat canopy prior to anthesis. The soil was fertilized with 35 kg N ha-1 as ammonium nitrate on 20 May. generally assumed that this soil emission is complete within 1-2 weeks of fertilization, after which vegetation processes dominate exchange. The hypothesis that the emissions derived from the soil was tested using a simple chamber placed over bare soil between the plants. The flux was derived from the difference of the inlet and outlet concentrations with a flow rate of 30 1 min -1. Given the small area of the chamber (0.07 m2), possible patchiness in emissions over the field, and potential adsorption of NH 3 to chamber walls, this test was only expected to give a broad indication of the soil flux. The results, presented in Figure 1, show a very close agreement with the micrometeorological estimates. This high level of similarity is almost certainly fortuitous, though clearly demonstrates that the soil was the main source of emission. During the same campaign, measurements were made of the deposition rate to bare agricultural soil (pH 6-7). This showed a reasonably consistent pattern of deposition at a slow rate, with good data indicating an R c of around 300 s m-1. The measurements in May 1992 were made before and after fertilization with ammonium nitrate. It is known from laboratory studies that emissions from this fertilizer are expected to be small (e.g. 1 % applied N). Measurements before fertilization (1 month after previous fertilization with ammonium nitrate) showed no detectable soil flux, with vegetation dominating the exchange process. In contrast to the April 1992 measurements, only a very small soil emission was detected following fertilization. The results are shown in Figure 2. In addition to day time emissions, they show a small but significant (5 ng m -2 s-1) net emission during night time. Since stomata are closed during the night this is attributed to soil emission (supported by within-canopy concentration profiles). In both Figures 1. and 2., SO 2 concentrations are plotted against the NH 3 concentrations. As previously shown by Sutton et al. (1993a), there is an inverse relationship between the concentrations of these 2 gases. This may be related to formation of ammonium sulphate aerosol depleting the concentration of one of the gases in the presence of an excess of the other. 5. I N T E R P R E T A T I O N O F R E S U L T S . In order to simulate explain the bi-directional fluxes observed over the wheat canopy a modelling framework is required that can treat both NH 3 sources (compensation point, soil emission) and sinks (cuticular uptake, stomatal uptake) and their interaction. In particular,

75

II

II

R~

R~

F~

F~

LT Rw

IT

Rs

Kr~FQd _-z"

>0

A.

Cd

--

Rs

s>0

B.

Figure 3. Model diagrams to describe the bi-directional exchange of ammonia between a plant canopy and the atmosphere. The net flux is described as the result of the exchange between the air concentration (Z) and the 'canopy compensation point' (Zc,), which is the result of competition between exchange to the leaf surface and transfer through stomata (Rs) with a stomatal compensation point (Zs)- Deposition to the leaf surface is treated in different ways: A, as a simple resistance (Rw); B. as an adsorption capacitance (Cd) of given charge (Qd) with adsorption resistance (Rd) and a reaction rate (Kr). R a and Rb are the resistances to atmospheric turbulence. attention is given here to addressing the interaction between a stomatal compensation point, leaf cuticle uptake processes and atmospheric NH 3 concentrations. 5.1. Canopy compensation p o i n t , cuticular resistance model A resistance diagram of the model of ammonia canopy exchange proposed by Sutton and Fowler (1993) is shown in Figure 3a. The model may be used as a tool either to interpret measured net fluxes or to infer the net flux given particular conditions. The difference between this and the of model Eq. (3) is that here the canopy compensation point (2c) is calculated rather than the canopy resistance (Rc). As with R c (Eq. 3), Zc may be directly estimated from micrometeorological measurements:

~,c = ~,{ z-d} + Ft(R a{ z-d} + Rb)

(5)

In principle the total flux F t must be conserved in its component fluxes to the cuticle/water layers (Fw) and with stomata (Fs) so that: Ft = Fw + Fs Fw =-Zc/Rw;

(6) Fs = (Zs - )~c)/Rs

(7a, 7b)

Hence, having determined F t and Xc from measurements, and from measured (or modelled) R s, the value of R w may be determined given different assumed values of )~s- Alternatively, given an estimate of R w, ~s may be found. The presence of two unknowns points to the need for independent laboratory determinations of these parameters (e.g.)~s). In contrast, the model may be used in an inferential manner to explicitly estimate Zc and F t. Substitution of Eqs. (7a,b) into Eq. (6) and rearrangement provides: F t = (Zs-Zc)/Rs - Zc/Rw This may then be combined with Eq. (5) to eliminate Ft:

(8)

76 20 10

.,~176

0 Flux NH3

-10

obs

(ng m-2 s-l)

. . . . . . . model

-20 -30

7

-40 19 May 1992 -50 00:00'

30

04:00

'

lz':oo

08:00

'

16':oo

zo:~o

oo:oo

20 10 Flux NH3 (ng m-2 's-l) 0

. . . .

-10

IIli

-20 ~30

iI ] 21-22 May 1992 I

-40 12:00

16:00

20:00

00:00

,++:I'-

~

obs model (with soil emission)

....

model (with no soil emission)

04:00

08:00

12:00

16:00

20:00

Time (GMT)

Figure 4. Comparison of measured ammonia flux over wheat and that predicted by the simple (Rw) canopy compensation point model. Rw calculated according to Eq. (12) and assuming plant intercellular pH 6.8 and 100 ~rnol NH4+ 1-1. Model soil emission in the Zc calculation (Eq. 13) for 21-22 May set at 10ng m-2 s-1. %s/%c - 1 Rs

1 Rw

Zs

1

1

ZcRs

Rs

Rw

1-Z{z-d}/Zc R a {z - d } + R b

1 Ra{z-d}+R b

(9)

Z{z-d} Zc(Ra{z-d}+Rb)

(10)

From which the canopy compensation point may be given as:

[Z{z-di/(Ra {z-di+Rb)+Zs/Rs] Zc - [ ( R a { z _ a i + R b ) _ 1 + R s - 1 +Rw_l]

(11)

and the flux found from Eq. (5). Examples of the application of this model are shown in Figure 4. On the basis of comparison with laboratory studies of the relative humidity (RH) response of adsorption, a simple parametrization of R w (s m -1) was used (Sutton and Fowler 1993)" R w = 2 exp ([ 100-RH]/12)

(12)

The stomatal compensation point was calculated based on the temperature dependent Henry equilibrium using pH of 6.8 (Farquhar et al. 1980) and assuming 100 gmol NH4 + 1-1 in the leaf intercellular fluid. The latter was chosen to fit the data in Figure 4a, and is consistent with other published estimates (Farquhar et al. 1980; Sutton et al. 1993c). Applying these values to the measurements after fertilization with ammonium nitrate (Figure 4b) significantly underestimated the flux. This may be due to daily variations in R w but is also likely to be related to the presence of an additional soil emission, as indicated by the night time

77 measurements. It is possible to provide to provide a simple treatment of soil emission in the canopy model:

Zc = [)~{z-di/(Ra { z - d } +Rb)+ )Cs/Rs + Fsoil ] [ ~ 7 { z - d } + Rb)-l+Rs-l+Rw-1]

(13)

The effect of adding a soil flux in this manner is shown in Figure 4b, indicating that the daytime emissions would be consistent with an additional emission into the canopy space of 10 ng m -2 s -1. The difference between the two model estimates is not constant because of varying recapture by leaf surfaces and stomata. It should be noted, however, that introducing a soil emission in this simple way is unlikely to give precise results, because of the different physical location of the soil to the vegetation canopy. As a single layer model, it is assumed that all the exchange occurs at a single hypothetical height, which would be expected to provide errors for soil exchange processes.

5.2. Canopy compensation point- cutieular capacitance model Although it is possible to provide reasonable agreement with the measured data using the simple canopy compensation point model, the measured results sometimes show much larger emission in the morning than predicted from a temperature dependent compensation point, while emission in the evening may be smaller than be expected. Two effects that may explain this are that the concentration of NH4+ in leaves is not constant throughout the day, and/or that the leaf cuticle acts more like a capacitor for adsorption of ammonia. There is some recent support from laboratory studies that the concentration of NH4+ may vary diurnally, with larger values in the early morning (Schjc~rring, 1994, pers. comm.), and further work is required to examine this possibility and its cause. However, it is equally possible that during increasing humidity conditions the leaf surface will be able to absorb more NH 3 than when the surface is drying. In the latter case, if deposited NH4+ is not 'fixed' by reaction to form salts with low vapour pressure (e.g. [NH412SO4), it may be released back as NH 3 and contribute to net emission. An model to explore this behaviour ammonia on leaf surfaces is shown in Figure 3b. A number of studies have investigated the link between relative humidity and thickness of notional 'water-films' on leaf surfaces (Van Hove et al. 1988; Benner et al. 1992; Burkhardt and Eiden 1994), which, for a given pH, would be expected to have a defined capacitance according to Henry's law. The relation defining the capacitance (Cd) may be expressed as:

Cd = Qdl)~d

(14)

where Qd is the adsorption charge (lag m -2) and )~d the adsorption concentration (l.tg m -3) associated with the capacitor. By analogy to this relationship, an estimate of C d may be found from the Henry equilibrium constant and an equivalent water-film thickness (MH20). Using the solubility equilibria provided by Sutton et al. (1993d) gives:

Here Cd and MH20 are given in metres and T in Kelvin. On the basis of measurements on polluted leaves it is estimated that a typical value of MH20 would be 20 nm at 60% (Burkhardt, unpublished data). Using a similar humidity response to Eq. (12) and accounting for the canopy leaf area index (LA/), an initial estimate of MH20 was found as: MH20 = LM * 20 exp ([RH-60]/10)

(16)

Unlike straightforward resistance models, treating the leaf surface exchange process as a capacitance results in the flux at a given time being time dependent on previous fluxes. Hence calculation of the modelled is linked over different model time steps. It is necessary to set an initial value of either Qd or )(;d, so that for an initial time (i): )~d{i} = Qdlii/Cd

(17)

78 The flux into or out of the adsorption capacitor (Fd) is then: F d = (Zd{i}-Zc{i})/Rd

(18)

where R d is the charging resistance of the capacitor. The value of Zc{i] may be found by applying the canopy compensation point equation (Eq. 11) in slightly modified form:

[~{z-d}/(Ra {z-d} Zc =

+ R b) + Zs/Rs + zd/Ra] [(Ra{z-dI+Rb) -1 +Rs -1 +Rd -1]

(19)

The new capacitance charge (Qd{i+t}) after t seconds is then found as: Qd{i+t} = Qd{i} - (Fd.t) (20) Given the new value of Qd a revised value of Zd is found according to Eq. (17), and the process repeated for the next time step. While this parametrization will treat adsorption and desorption to cuticular water-layers, it does not provide for any net removal of NH 3 from the air by leaf surfaces. This may be accounted for by proposing a reaction flux (F r) of the stored NH4+ (e.g. to form ammonium sulphates) with a rate constant Kr (s-l):

Fr = ad. Xr where a negative value of Kr indicates deposition. Inclusion of F r into Eq. (20) provides: Qd{i+t} = Qd{i} - (Fd.t) + Fr (21) In running this model it is found that the value of R d defines the rate of charging of Qd and provides the time constant of the adsorption/desorption process. It is anticipated that R d should be larger in dry conditions (reduced access to the water layer), and this is also required to make the model run, since the model requires smaller time steps to remain stable with decreasing values of R d and C d. The initial results presented here are calculated using R d (s m -1) = 5000/C d (m). This is equivalent to a time constant of 83 minutes. An example of the application of this model is shown in Figure 5, for the measurements on 3-4 July 1993 over mature wheat (early leaf senescence). The measured fluxes on the 4 July showed a larger emission in the morning and less in the evening than would be expected by a compensation point emission through stomata. In this case the capacitance model is run assuming a leaf surface pH of 4.5 and an adsorbed NH4+ reaction rate (Kr) of-0.01 s-1. The effect of including K r is shown on the value of Qd" Estimates of the modelled flux are provided, one just accounting for the capacitance effect, and the second linking this with the compensation point exchange. The model is able to predict successfully the peak of emission on the morning of 4 July, though daily differences related to untreated factors also occur since the flux on 3 July is not well represented. The comparison between the two model flux estimates in Figure 5, shows that stomatal uptake of N H 3 desorbed by the cuticle reduces the peak emission, while later on in the day stomatal exchange contributes to the modelled net emission. 6. DISCUSSION AND CONCLUSIONS Measurements of NH 3 surface atmosphere-exchange over arable croplands show that each of stomatal, leaf surface and soil exchange processes are important in def'ming net fluxes. Where fertilizer nitrogen is added as ammonium nitrate, soil emissions are small and contributed here about 5 ng m -2 s -1 to the net flux 1-2 days after fertilization. In contrast, fertilization with urea provides much larger ammonia emissions. Chamber measurements made here supported the interpretation that the emissions were soil rather than vegetation related. An important finding was that the enhanced ammonia flux following urea fertilization may continue long after fertilization (4 weeks). Previous studies have often considered soil emissions complete after 12 weeks and may have underestimated urea emissions. Emission of ammonia from urea is a result of its hydrolysis to by urease producing ammonia and raising solution pH. In the example reported there had been tittle rain between fertilization and the measurements reported, suggesting that hydrolysis was slow. In contrast to these results, measurements over

79

2500

700 ,.,-, ..," ",

2000 Adsorption 15oo Capacitance Cd (m)

""

,300(56) stations: [] NH3,NH4, wet deposition O NH3, wet deposition NH3 only /~ NH4 only V wet deposition only

throughout the year, the number of repeats being dependent on the emission strength in the surrounding area. 2.3. Calculation of NH 3 and NH 4 concentrations and wet deposition with OPS

OPS is a Lagrangian dispersion, conversion and transport model which calculates the concentration and deposition of primary and secondary components in a receptor point due to each of the emission sources separately. It is thus linear with respect to the emissions. The model uses national mean values for roughness length, deposition velocity and conversion rate. Normally, also national mean values for meteorological conditions are used but in this study regionalised meteorological conditions were used because of the substantial influence on predicted concentrations [6,7]. The diurnal emission pattern used here was that according to Acharya [7]. The emission data for the Netherlands used in this study are the official emission data for 1992 as described by van der Hoek [8]. The emission data for the European countries were obtained from Asman [9]. 2.4. Determination of the uncertainty of model predictions

One of the sources for uncertainty of the model results is the uncertainty in the model parameters. The uncertainty of the predicted NH 3 concentration was determined by simultaneous variation of parameters using the UNCSAM package [10]. The three parameters studied are the conversion rate of ammonia to ammonium (a), the surface resistance for ammonia (rc) and the scavenging efficiency for ammonia and ammonium (s).For this study it was assumed that the value of each of these parameters was distributed homogeneously over the uncertainty region. The default values applied in the model and the boundaries of the uncertainty regions are listed in Table 1.

84 Table 1 Default value and uncertainty boundaries of some model parameters parameter a (% per hour) rc

default

boundaries

28.8

[10,50]

30

[10,100]

1.10 6

[ 105,2.10 6]

(s m -1)

s (-)

2.5. The calibration method

Foreign emissions and industrial emissions were left out of the calibration procedure. The field of the remaining agricultural and household emissions was subsequently divided into 5 regions, 4 emission regions and the rest of the Netherlands. The determination of the emission regions was based on the amount of emission from manure per 5 by 5 km grid cell. Grid cells with emissions higher than 150 000 kg were selected. This resulted in 4 clusters of grid cells being the 4 emission regions (see Figure 2). Due to the linearity of the model with respect to the emissions the following equation applies to the model outcomes: Vcp = ~

"fl ) + ~

"f2 )+ ...... +~

"f5 ) + ~

) + ops(Ef )

(1)

where Vcv is the vector with model outcomes (NH 3 and NH 4 concentrations, NH x deposition) for each of the measurement locations and parameter set p, Eah,n is the vector of grid cell emissions for region n and fn is the emission calibration factor for region n with value 1 in the default situation. The calibration now consists of finding the values of fl to f5 which minimise:

= 51w.C - m/l

(2)

where V m is the vector with measured values and w is a vector with weighting factors which correct for the different order of magnitude between ammonia, ammonium and wet deposition (set at 1, 3 and 0.015 respectively in this study). This problem is solved analytically. If one of the values fl to f4 becomes negative, the corresponding region is added to the rest of the Netherlands and the analysis is started again. Because the model parameters are also uncertain, it is subsequently searched iteratively for the values of p which minimise C p. The calibration procedure is carried out in 2 ways. The first method uses all stations in the analysis (for wet deposition only those stations where also the ammonium and/or the ammonia concentrations are measured). The second method uses all stations but one and the calibration is repeated as many times as there are stations (m), each time leaving out another station (jack-knife method). This procedure yields m values for the calibration factors and the model parameters. The inner m - 2 values are averaged and the standard deviation of the m - 2

85 Figure 2. Partitioning of the Netherlands into 4 emission regions. The partitioning is based on the emission from manure per grid cell. r-1 Eibergen region [] Zegveld region I Vredepeel region I Lunteren region

values is calculated. The jack-knife procedure was used as a way to 'validate' the results of the 'all stations' procedure.

3. RESULTS

3.1. Comparison between measured and calculated values Figure 3 shows the comparison between measured and calculated values for the ammonium concentration, the ammonia concentration and for the wet deposition of NH x. Though the correlation between calculated and measured values of NH 3 is reasonable, the calculated values for the stations in emission areas are much lower than the annual means from the measurements. For NH 4 aerosol there is hardly any correlation between calculated and

Figure 3. Comparison between measured and calculated values before calibration. predicted (pg/m3) 25

NH3

8

20 .

10

'o0

~*, 5

9

9

42

9

;0

9

1200-

9

9

.i 9

NH4

6

.

15

predicted (pg/m3)

predicted (pg/m3)

9

;5

2'0

2'5

measured (pg/m3)

0

.

. o."

,

2

.,,

.

.. o~

wetdepositi0n

900-

~

~

~

measured (pg/m3)

600-

. "~o

300-

9

o

9

0

,

,

3~0 600 900 12'00 measured (pg/m3)

86 measured values and calculated values are much higher than the measured values. For wet deposition of NH x there is a fairly good correlation between calculated and measured values. However, the calculated values are somewhat lower than the measured values over the entire range of measured values. It should be mentioned here that the poor correlation for NH 4 aerosol and the apparent underestimation of wet deposition of NH x may be caused by erroneous values for the measured data. Unlike NH 3 (next section) there has been no investigation of the reliability of measured values of NH 4 and wet NH x deposition.

3.2. Spatial representativity of measured values of NH 3 Figure 4 shows for each station the mean value of all reference measurements (averaged over space and time) and for comparison the mean value of all simultaneous measurements at the fixed point. Also indicated are the standard errors of the mean. From these measurements there is no indication of significant local influences on the fixed point. However, especially for the stations in areas with high emission density (131, 722 and 734) conditions with enhanced concentrations seem to be under represented in the reference measurements. This becomes evident when the mean value of the reference measurements at the fixed point is compared with the annual mean value of the continuous measurements at the fixed point. using the same time window (11.00-19.00 h) as with the reference measurements to avoid diurnal variation effects.

3.3 Uncertainty in the model calculations of NH 3 The results of the UNCSAM analysis are shown in Figure 5. The dots represent the complete range of the model outcomes. The outer values of this range are less probable than the inner values. It can be concluded that although the uncertainties in the three model parameters studied give rise to an appreciable uncertainty in the predicted NH 3 concentration, the gap between predicted and measured values can not be closed by the uncertainty in the parameters only.

Figure 4. Comparison between mean value of measurements made at reference points and mean value of simultaneous measurements at fixed point. The annual mean at the fixed point is also indicated.

concentration (pg/m3) 20reference

16-

~

fixed (sameperiodsas reference) fixed (wholeyear/11.00-19.00h)

128-

0

i

!

!

540

235

928

/

538

!

|

|

|

633

722

131

734

station number

87

)redicted NH3 concentration (pg/m3) 252015-

"1 I

t .

10-

.111|

o o

1'0

l's

2'0

Figure 5. Uncertainty in the calculated NH 3 concentrations due to uncertainty in 3 model parameters (conversion rate, surface resistance and scavenging efficiency.

2's

measured NH3 concentration (IJg/m3)

3.4. Calibration results The first calibration runs in which the model parameters were also calibrated showed that the value of the conversion rate a was consistently adjusted to its lowest possible value and that the value of the scavenging efficiency s was always adjusted to its highest possible value. Therefore a and s were fixed at these boundary values in order to save computer time. Because the results of the jack-knife method compare very well with the results of the 'all stations' method only the results of the jack-knife method will be presented here. In Table 2 the results of the calibration analysis are listed. When the model parameters are kept at their default values, the emissions of all four emission regions are increased by the calibration procedure. The highest adjustments are

Table 2 Calibration results with default and adapted parameter values. with default parameter values

with adapted parameter values

a

28.8

10

rc

30

62 + 11

s

106

2" 106

...........................................................................................................................................................................................................................

fl (Eib)

1.18 + 0.03

0.80 + 0.05

f2 (Lun)

1.99 + 0.04

1.56 + 0.04

f3 (Vre)

1.20 + 0.02

1.13 + 0.04

f4 (Zeg)

1.79 + 2.69

1.35 + 0.32

f.s (rest)

0.27 + 0.56

1.06 + 0.13

123

178

Etot (ktonnes)

88 Figure 6. Comparison between measured and predicted values after calibration. predicted (pg/m3)

25

predicted (pg/m3)

predicted (pg/m3)

NH4

O"

8

NH4

.

1200-

wetdeposiUon

-O

9001

1o

,r

"

oO " .

,.,

,.

o

~

600-

..

1'o 1's 2'o 2's measured (pg/m3)

O~

o

~

300-

~.

~

~

measured (pg/m3)

0

.,* "

o

."

*

3~o 6~o 9~o 12'oo measured (pg/m3)

needed for the Lunteren and the Zegveld regions (regions 2 and 4 respectively). However, simultaneously the emission of 'the rest of the Netherlands' is strongly decreased. Consequently, the national total NH 3 emission decreases from 166 ktonnes to 123 ktonnes. This result is regarded as very unlikely. When a and s are set at their lowest respectively highest value and r c is allowed to change within certain limits, the calibration procedure doubles the value of r c in order to increase the ammonia concentration at the stations in the high-emission areas. Consequently, the emissions need not be increased that much as with the default parameter values. For the Lunteren region and the Zegveld region calibration factors are in the range 1.4-1.6 now. For the Eibergen region (region 1) the reported emissions are decreased with about 20%. The national total NH 3 emission increases from 166 ktonnes to 178 ktonnes with this calibration 'scenario'. Figure 6 shows the comparison between measured and calculated values after calibration with parameter values allowed to change. Obviously, the calibration has resulted in a much better agreement for all components (NH 3, NH 4, wet deposition of NHx). Table 3 gives the national mean values for NH 3 concentration, NH 4 aerosol concentration, dry, wet and total deposition before and after calibration with adapted parameter values. As a result of increasing the emissions, decreasing the conversion rate and increasing the surface resistance, the ammonia concentration has increased by about 50% and the ammonium

Table 3 National mean values before and after calibration with adapted parameter values before calibration

after calibration

NH3 (lag m -3)

4.36

6.60 (+51%)

NH4 (lag m -3)

5.25

3.42 (-35%)

wet deposition (mol ha -1 a -1)

499

583 (+ 17%)

dry deposition (mol ha- 1 a- 1)

1004

1024 (+2%)

total deposition (mol ha- 1 a- l)

1503

1607 (+7%)

89 concentration has decreased by about 35%. Wet deposition has increased due to the increase in scavenging efficiency. Also the increased r c and therefore the increased ammonia concentration contributes to the increase in wet deposition. This becomes evident when the wet deposition data are not used in the calibration. In that case r c stays at its default value. Dry deposition is hardly influenced by the change in emissions and parameters. The increase in ammonium concentration is counteracted by the decrease in ammonia concentration and the increase in surface resistance. Consequently, also the total deposition has not changed much.

4. SUMMARY AND DISCUSSION The calculated values of ammonia and ammonium concentrations using the emissions as reported by van der Hoek [8] and the default model parameters show considerable disagreement with the measured values. The calculated values for the concentrations of ammonia are lower than the measured ones, whereas for ammonium the opposite is valid. For wet deposition of NH x the calculated values are lower than the measured values. In spite of the use in this study of the measured data as they are now available, further investigation of the reliability of the ammonium concentrations and also the wet deposition of NH x is needed. The deviations between measured and predicted values for ammonia cannot be explained by the uncertainty in either of these. Measurements of ammonia concentrations in the surroundings of the fixed point do not prevail local influences on the fixed point. However, conditions with high emissions are probably under represented. Therefore, local influences cannot be excluded. Considering the fact that the comparison measurements with a van are very time consuming, it is suggested to develop low-cost passive methods which integrate the ammonia concentrations over a certain time span and which should be applied a year long. The uncertainties in the model parameters result in a quite large uncertainty in the calculated ammonia concentration, but not large enough to account for the difference in measured and predicted concentrations. The parameters studied are the conversion rate of ammonia to ammonium, the surface resistance for ammonia and the scavenging ratio for ammonia and ammonium. All three parameters describe the removal process and represent the main uncertainties. Not covered in this study are some additional uncertainties caused by (meteorological) parameters describing the dispersal process. Calibration with the model parameters kept at their default value leads to a very unlikely best fit for the emission distribution. The emissions for the Zegveld and Lunteren area increase (up to a factor 2) whereas the emissions for 'the rest of the Netherlands' are substantially reduced. As a consequence, the total Dutch emission decreases from 166 ktonnes to 123 ktonnes. When the uncertainty in the model parameters is taken into account the required corrections of the emissions are smaller but still considerable (up to 1.6 for the Lunteren area). In that case the values for the conversion rate and the scavenging efficiency are invariably set at their lowest respectively highest possible value and the surface resistance is about doubled. Because the regions with the high corrections do not contribute that much to the total Dutch emission, the latter increases only slightly from 166 ktonnes to 178 ktonnes. This increase is within the 'normal' uncertainty of an emission estimate. As a consequence of the adapted emission and model parameters the modelled national mean ammonia concentration increases with about 50%. However, the dry deposition of ammonia and ammonium is hardly affected because of the simultaneous decrease in the

90 ammonium concentration (about 35%) and the increase of the surface resistance. The wet deposition increases lightly (17%) due to the increase in the scavenging efficiency but also resulting from the higher ammonia concentration. Following this, the total deposition of NH x increases with about 7%. The results from this study are (probably) not independent of the way the emission field is partitioned. In a further study not the emissions of distinguished regions but the emission factors of the activities leading to emission of NH 3 will be calibrated.

5. REFERENCES

1. Bestrijdingsplan verzuring, Report nr. VROM 90213/8-89 (in Dutch), Ministry of Housing, Spatial Planning and Environment (1989) 2. Erisman J.-W., Water, Air, and Soil Pollution 71 (1993) 51. 3. Van Jaarsveld J.A., Report nr. 222501002, RIVM, Bilthoven, 1990. 4. Asman W.A.H. & Van Jaarsveld J.A., Atmospheric Environment, 26A (1992) 445. 5. Van Elzakker B.G., Buijsman E., Wyers G.P. & Otjes B., this volume. 6. Boermans G.M.F. & Erisman J.-W., Report nr. 222105002, RIVM, Bilthoven, 1993. 7. Acharia R.C., Rep. nr. H.H. 203 (M.Sc. Thesis), IHE, Delft, 1994. 8. Van der Hoek K.W., Report nr. 773004003 (in Dutch), RIVM, Bilthoven, 1994. 9. Asman W.A.H., Report nr. 228471008, RIVM, Bilthoven, 1992. 10. Janssen P.H.M., Heuberger P.S.C. & Sanders R., Environmental Software 9 (1994) 1.

Acknowledgements: the authors wish to express their thanks to J.A.van Jaarsveld for fruitful discussion of the topic, to H.S.M.A. Diederen for critically reviewing the draft version, and to J. Burn for editorial assistance.

G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.

91

D e p o s i t i o n N e t w o r k of t h e F e d e r a l E n v i r o n m e n t a l A g e n c y (UBA) - R e s u l t s and Trends D. Kallweit Umweltbundesamt, Bismarckplatz 1, 14191 Berlin, Germany Abstract The UBA network is based on longterm standardised meteorological immision and deposition measurements at national level. Deposition data of Eastern (EG) and Western (WG) Germany were compared for 1986-1993, especially at times characterised by a dramatic decrease in air pollution. Important results so far have been the strong decrease of base cations, but the decrease in SO4 deposition has not reached such a high level in EG. The current slow increase in N-deposition (nitrate, ammonia) is accompanied by a slow increase in acidification in EG. Generally, the deposition level in EG and WG both slowly reached the same level. Introduction A Commission of Experts on Environmental Matters and the German Union of Water M a n a g e m e n t and Natural Resources (DVWK) found the deposition measurement operations too diverse to reach uniform findings on deposition for Germany. These findings are based on the fact that too many factors are involved in this process: for example, different deposition collectors, sampling frequencies and chemical analytical methods. The distribution of sampling sites is irregular and regionally confined. This leads to minimising the comparison of data. Using this as an argument, the UBA network has been extended to include the wet-only deposition measurement. The special quality of this network is the combined standardised meteorological immission and deposition long-term measurements at national level using automatically operated container stations and manned stations. B a c k g r o u n d , S a m p l i n g Sites and M e t h o d s Since the beginning of the 1980s, daily-bulk measurements have been carried out at UBA-manned stations. These measurements are indicated in Table 1. With the unification of Germany, 10 stations from a total of 30 belonging to the wet-only deposition network of the Meteorological Service of the GDR were joined up to the network of the UBA. From the end of 1991, this has led to the gradual development of a national wet-only measuring programme as stated above (Figure 1).

92

container station %~ ~ measurement site 0~ wet-only site _&~F q ,

/

~ ,~r~.~ ~ ~

~ ~ .... t

~

~" Hohenwestedt t.~

~)Twixlum

- -~

O,.um

~,

I./"-jNeug~br.s~;i ~.

~,

i'

:-~_~

~ '

.~'-:c,

, ~.~Solling

Hilchenba

,A~

...... ~ ~

Waldhof

-%:

~

~,

Wies,enburg ......

'/

Lindenberg

Leinefelde Lehn ' J J Regnitzlo~:

uselbach

8

Kehlheim Ansbach Brotjacklriegel

~chauinsland

Murnauer Moo

2 Figure 1. Immission and deposition sites within the UBA Monitoring Network

93 At the moment, measurements are being carried out at 26 stations; the entire network will be completed by the end of this year. The measuring programme is intended to estimate the quantity of rain, pH, conductivity, main ions and selected heavy metals. First, the precipitation sampling was achieved by means of an ANTAS wet-only collector produced by the Meteorological Service. After a favourable comparison of deposition collectors of the former GDR and the FRG (mean deviation < ca. 5 %) the collector was replaced by NSA 181 KD produced by the Eigenbrodt firm. This wet-only collector contains a permanent cooling and sample-changing system, allowing samples to be changed every two weeks instead of weekly. The samples are analysed in the laboratory of the Institute of Energetics in Leipzig. The bulk samples are analysed at the UBA site in Schauinsland (see Table I for methods). Table 1 UBA programme of deposition measurements and analytical methods p r o g r a m m e type wet-only programme: number of stations: 30 8 manned stations 16 a u t o m a t i c a l l y operated c o n t a i n e r stations 6 stations in cooperation with the federal states sampler: wet-only NSA 181 KD Fa. Eigenbrodt frequency of sampling: weekly (Tuesday-Tuesday), 8:00 a.m. (7:00 UTC) precipitation volume, conductivity, pH 8042", NO3-, NH4+, C1Na+, Mg2+, Ca2+, K+ heavy metals: Pb, Cd, Cu, Zn, Mn bulk programme: 8 manned stations number of stations: 8 bulk ARS 721 Fa. Eigenbrodt sampler: frequency of sampling: daily, 9:00 a.m. (8:00 UTC) precipitation volume, conductivity, pH SO42-, NO3-, NH4+, C1Na+, Mg2+, Ca2+, K+ heavy metals: Pb, Cd, Cu, Zn, Mn, Fe analytical m e t h o d s wet-only/bulk programme:" parameter method detection limit 8042ion chromatography 10 ~tg 0,21~teq/1 NO3on chromatography 10 ~tg 0,16 ~eq/l NH4 + ion chromatography 10 ~g 0,55 ~teq/l NH4 + flow injection 10 ~tg 0,55 ~eq/l C1ion chromatography 10 ~tg 0,28 ~teq/l Na+ ion chromatography 10 ~g 0,43 ~teq/1 Mg 2+ ion chromatography 10 ~tg 0,25 ~eq/1 Ca2+ ion chromatography 10 ~tg 0,50 ~eq/l K+ ion chromatography 10 ~tg 0,80 ~teq/l

94 wet-only deposition: laboratory of the Institute of Energetics, Leipzig bulk deposition: laboratory at Schauinsland/UBA site The wet deposition measurements are inspected at regular intervals to improve quality (QA/QC). For instance, spotchecks are carried out at the sites. We participate in national and international intercomparisons of collectors and analytical methods. The conditions for choosing the deposition measurement site must represent: - the background level of air pollution and deposition. They must include as many types of ecosystems as possible. The sites not owned by the UBA must be suitable for long-term measurement/observation. Another aspect was to measure on sites where scientific research installations and institutions of the various L~inder (federal states) already existed. The intent was for them to complement each other. In addition, at UBA automatic stations the following m e a s u r e m e n t s were combined: - wet-only/weekly- UBA; -wet-only/daily or four-hourly s a m p l i n g - Institute of Tropospheric Research, Leipzig; -micrometeorological/inferential m e a s u r e m e n t s - F r a u n h o f e r I n s t i t u t e of Environmental Research, Garmisch Partenkirchen; - measurements of TSP total and component amounts - UBA. This cooperation is presently being conducted through the SANA Research Project (re-development of Eastern Germany). The aim of this cooperation is to develop, if possible, a method to estimate the entire deposition input for Germany using measurements and modelling and also considering the chemical processes in the atmosphere.

Results

and

discussion

For the initial evaluation, data is available from eight stations in E a s t e r n Germany (EG) and five manned stations in the Western Sector of the Federal Republic of Germany (WG) for the period 1986-1993. The deposition data of the bulk deposition UBA network (WG) was compared with the wet-only deposition data from the Meteorological Service (since 1988 wet-only) up to 1991, when UBA data (EG) was used. In fact, the bulk samples on sites with a low pollution level are analysed only after a rainfall, so that they are more comparable to wet-only samples (see Table 2). The change in emission, and in its wake the immission matrix, was carried out at short notice in the quickest possible time, especially in EG, (i.e. in the period between 1986-1993). This was accelerated by the collapse of industry and agriculture in the former GDR, which is now experiencing an upward swing.

95 Table 2 Mean values for wet deposition for 1986-1993 year RR

WG 86 87 88 89 90 91 92 93

86 87 88 89 90 91 92 93

Na+

mm

kg~a ~a

kg/ha kg/ha ~ a

1.018 1.082 1.052 751 862 736 920 906

30.59 31.09 31.65 24.54 25.94 21.16 21.10 21.11

7.52 8.04 8.43 6.79 6.51 5.61 6.07 6.08

year RR

EG

SO42- NO3- NH4 + C1-

22.20 26.22 25.61 20.70 19.27 17.49 19.54 19.21

20.39 15.92 27.21 16.26 42.00 26.65 24.62 20.80

SO4 2" NO3- NH4 + C1-

Mg2++ Ca2+ ~a

11.32 1.59 8.37 1.28 14.56 2.11 8.16 1.41 22.212.70 13.601.94 13.581.75 10.861.41 Na+

K+

H+

kg/ha ~ a

kg/ha

3.38 3.84 3.80 3.57 3.75 3.16 3.86 3.69

1.49 1.30 1.56 1.20 1.75 3.38 1.22 1.03

0.32 0.35 0.31 0.24 0.22 0.21 0.20 0.20

K+

H+

MG2++ Ca2+

mm

kg~a ~a

kg/ha kg/ha ~ a

~a

kg/ha ~ a

~a

542 576 515 396 520

63.13 66.24 64.24 41.44 37.44

7.68 8.82 8.14 6.43 6.81

1.83 2.13 2.15 1.11 1.10

12.26 14.65 14.21 7.21 5.73

2.29 2.02 2.46 1.09 0.86

0.20 0.22 0.21 0.16 0.16

437 549

17.67 13.12 3.84 7.55 22.31 16.49 5.92 6.73

0.82 0.74

1.28 1.27

0.28 0.20

19.53 19.57 20.28 14.95 15.44

8.95 10.17 13.52 8.77 8.99

4.89 4.82 6.74 4.18 4.06

4.21 0.82 3.51 0.74

W G - mean values for the deposition sites of Westerland, Waldhof, Deuselbach, Schauinsland, Brotjacklriegel E G - m e a n values for the deposition sites of Schwerin, Teterow, Neuglobsow, Lindenberg, Leipzig, Leinefelde, Schmiicke The decrease in SO2 emission (see Figure 2) is accompanied by a s i m u l t a n e o u s decrease of particulated m a t t e r emission (flying ash with Ca and Mg) and is in accordance with the mean value of SO2 immission (see Figure 3) in rural areas (ca. 35 % lower) during 1985-1988. For the same period, in urban areas this value rose to 40 to 65 %. Nevertheless, the SO2 immission in EG is three times higher t h a n in WG and leads to exceedances of permitted a n n u a l values (140 Ilg/m3 SO2) of the Technical I n s t r u c t i o n on Air QualityfrA-Luft) in Germany. The p a r t i c u l a t e d m a t t e r immission dropped by 30 % on average.

95 The annual wet deposition (see Table 2) depict a slower decrease in comparison to the immission. This is reflected especially in the case of WG, where a marked fall in immission took place between 1987-1988. But the SO4 deposition showed a decline only from the beginning of 1989. The same can be said of EG, although the decline began here in 1989/1990. This could be related to the delayed (1990) decrease of particulated matter immission.

(~o

SO2

~oo

~oo ...... ...........

"

NO2

............................................................

2500 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

~ = .................................

.

I ~

o,

l ........

,

,

,

;'

,

,

,

partlcultecl m a t t e r

NH3

25OO

290O

. . . . . . .

~

. . . . . . . . . . . . . . . . . . . . . . .

4O0-

ISO0

lOGO

,

86

,

e7

,

e~

,

89

,

9Q

gl

O"

"

e7 ~ / 7 ~ " . G .

wG eg

j

go

.1

[~KQ i w ~ ]

Figure 2

Trend of SO2, NO2, NH3 and particulated matter emission in Germany 1986-1991 (Data from UBA)

97 At the beginning of the period in question, the SO4 deposition decreased twofold in comparison to WG (see Figure 4). In EG the base Ca cations achieved a fourfold decrease and Mg also decreased somewhat more than in WG (see Figure 5). At the end of 1993 the situation was different. The SO4 deposition in EG decreased by two-or threefold, the Ca deposition likewise. Within this same period the SO4 deposition decreased by a third and the Ca deposition in WG remained unchanged. The SO 4Ca-ratio for EG is characterised by strong deviations caused by irregular decreases in SO2 and Ca emissions.

S02 36"

30 .............

!~

26 ..............

~

15 ..............

~

..............................................................................................................

.....

I

.............................................................................

~.............. 86

.................... 87

$$

e0

I 91

90

.

.

.

.

.

.

.

.

.

.

.

.

92

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

93

J

[p'-/-~eQ ~wQ

TSP 45 40 . . . . . . . . . 353025 .............. 2015 ..............

10 .............. 5 ..............

0

$6

87

v

$$

~)

q

90

IrT/-~eo ~wo

Figure 3

|

91

92

93

]

Trend in 802 and TSP immission in Germany (yearly mean concentration based on daily means)

98 The following developments have been recorded for the nitrogen compounds. The NO3 deposition decreased slightly in WG between 1988-1991 and has since either remained constant or shown a minor increase. In EG this tendency has increased only since 1993. The NH4 deposition shown a downward trend in WG. In EG the values were the same at the beginning of this period. In 1992 these deposition values were reduced by 50% and with the revitalisation of agriculture there has been a further increase. It can be noted that the acidifying capacity is in principle characterised by a slow increase in acidifying species (SO42-, NO3-, CI-, NH4+, H +) and the more decreasing input of base cations. This development indicates a slow increase in acidification (see Figure 7) and represents a danger for the ecosystems. The ration of the main parameters of wet-deposition in EG and WG is increasing to the same level (see Figure 8). This development is a result of the active environmental policy and application of measures for air purification of national level.

Conclusions

- Acidification is increasing, especially in Eastern Germany. With the long-term and standardised wet-only network demonstrated we are able to follow the trend more accurately. - I n order to estimate the total input by deposition and its effects we need measurements on fog and cloud deposition, also on dry deposition over low vegetation and forests. -

It will be necessary to combine the measurement activities of several institutes to get information on deposition in Germany, including application using model calculations.

99

EG 1,4 , 1,2-1

0,8-1 "~ 0,6-1 0,4-1 0,2-t ! 19/~

1907

1988

1989

I mm so, ~

1990

.o3 ~

1991

| 1992

1993

c,

WG 1,4 1,31,21,110,9i

0,80,7-

0,60,50,40,30,20,10

198(!

i

1987

!

19eS

,

[ m e so, ~

1989

|

1990

.03 ~

1991

I

1992

I

1993

c,

Figure 4 Yearly development of the m e a n deposition values (anions) in EG a n d WG. There is no sea-salt correction for Westerland, so the m e a n value of C1 is influenced by this factor

100

EG 1,4 1,2-

0,80,60,40,2-

1986

i

1987

i

1988

I

1989

(3,

~

H

~Mg

i

NH4

1990

i

1991

~

N.

I

I K

i

1992

i

1993

WG 1,4 1,2-

0,8-~ 0,60,40,2-

1986

!

1987

~H

!

1988

!

1989

~Mg

i

1990

i

!

1991

!

1992

!

1993

IK

Figure 5 Annual development of mean deposition value cations in EG and WG

101 S 0 4 / C a ratio 4

3,5 - ................................................................................................................................ 3"~ 2,5 -~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

~/ ..............................................

..........................................

~ .................................

1,S-

..................................

~ .................................

1-

..................................

~ ............................

................................

~j~/...........................

2-

0,5

-

,

1988

,

19:B7

,

198:8

,

1989

,

1990

~

1991

,

,

1992

1993

Figure 6 Yearly development of SO4Ca deposition

EG/WG 4 3,6-

32,521,5I0,50

,

1986

,

1987

,

1~

,

1989 1990 Jehr

,

1991

1992

1993

Figure 8 Yearly development of ratio of main ions in deposition EG/WG

102

EG

:)(5 2 15 1

H H H~

05 i 19~(5

[~

! 1987

|

19(~

198g

aoidifying ion. ~

|

1990

i 1991

btme oldiomi

~

1992

lgg3

aoidily

WG

25

2

Z

15

i~

1 057

, 198e

L~7~ - 9

'~ / ~ ~ t ~ i ~

i 1957

| lgB$

ion- ~

| lg89

!

!

lggO

lggl

b-ee oallon.

[~]

, lgg2

aoidily

1993

I

Figure 7 Yearly development of acidity (A-acidifying ions (SO4 2-, NO3-, NH4 § Cl-, H+), B-base cations (Na+, K+, Mg2+, Ca2+) acidity = A-B)

G.J. Heij and J. W. Erisman (Editors). Acid Rain Research: Do we have enough answers ? 9 1995 Elsevier Science BV. All rights reserved.

103

The influence of ammonium nitrate equilibrium on the measurement of exchange fluxes of ammonia and nitric acid Yuanhang Zhang*, Harry ten Brink, Sjaak Slanina and Paul Wyers Netherlands Energy Research Foundation, PO box 1, 1755 ZG Petten, The Netherlands

Abstract Three sets of data with a time resolution of 40 minutes at Leende, Zegveld and Speuld in the Netherlands are used to evaluate atmospheric equilibrium of ammonia and nitric acid with ammonium nitrate. The comparison between the measured concentration products Km and theoretical values Ke reveals that gaseous ammonia and nitric acid are not in simultaneous equilibrium with ammonium nitrate aerosol, because the ratios of Km to Ke are far from unity, varying from less than 0.01 to above 100. Kinetic constraints upon the attainment of equilibrium are emphasized strongly by the results obtained at the three sites. The disequilibrium phenomena can be used to explain why HNO3 can show upward gradients and to obtain indications of the errors, using gradient methods in the measurement of deposition fluxes.

1. INTRODUCTION In Europe, exchange of NH3 and HNO3 between atmosphere and biosphere is of increasing interests. However, it is very difficult to derive their representative dry deposition velocities, especially for HNO3, because of the chemical reactions involving HNO3, NH3 and NH4NO3 aerosol. Sutton et al. [ 1] found consistent negative Rc for HNO 3 in Essex, England, while Huebert et al. [2] found a steeper aerosol nitrate gradient than that of its vapor and sometimes an apparent emission ofHNO 3 in the USA. According to Kramm and Dlugi [3], if no equilibrium exists, the fluxes can not be calculated by micrometeorological methods without appropriate corrections. Therefore, it is necessary to test whether the chemical equilibrium among HNO 3, NH3 and NH4NO3 really occurs in the atmosphere. The hypothesis of atmospheric equilibrium among gaseous HNO3, NH3 and NH4NO3 aerosol was first proposed by Stelson et al. [4] in 1979, who tried to analyze field experimental data by application of the equilibrium relationship. Since then, extensive work has been done regarding thermodynamic theory and field experiments [5-17]. Field measurements in the United States, Japan and Europe showed that NH4NO3 aerosol was generally in equilibrium with its gaseous precursors [8-17], especially at temperatures above 5~ and a relative humidity less than 80%. Departure from equilibrium was mostly found under conditions of low temperature and high humidity [9,10,15,16], when the concentration product was depressed to values less than 1 ppbv2 [ 18].

104 Meanwhile the validity of this thermodynamic approach was questioned for a long time[ 18-19]. The large uncertainty in the equilibrium constant under ambient conditions, as it is derived from laboratory experiments, complicates any assessment of its importance. Numerical simulations indicated that significant deviations from equilibrium, both positive and negative, might occur as the reactions are not sufficiently fast to maintain equilibrium under conditions of changes in meteorological, emission and chemical reaction parameters. However, after Jaffe's criticism [ 19], Mozukewich [7] argued that the available thermodynamic data on this system was much better than was implied by a factor of two range and that the dissociation constant may be determined within + 12% for dry aerosol. In the Netherlands, equilibrium between HNO3, NH3 and NH4NO3 aerosol is not extensively investigated yet. Erisman et al. [ 14] and Allen et al. [ 16] reported that the concentration products of the gaseous acids and base were generally in good agreement with theoretical prediction at temperatures higher than 0~ and relative humidity less than 80%. However, large deviations from theoretical values can be observed from the data of Erisman et al. [ 14] and Allen et al. [ 16]. Recently, various kinds of denuder techniques were developed to measure NH3 and HNO3 with high accuracy [20]. Many good data sets on simultaneously measured gaseous NH3 and HNO3 concentrations as well as meteorological data were obtained in recent years with a time resolution of 40 minutes by ECN. Three sites with different characteristics in location, emission and meteorology are chosen to test whether gaseous HNO3 and NH3 are in equilibrium with NH4NO3 aerosol in the Netherlands. The results obtained indicate complications in the measurement of dry deposition fluxes of NH3 and HNO3.

2. EXPERIMENTAL Ammonium nitrate NH4NO3 is present either in solid phase or as a solution, dependent upon temperature and relative humidity in atmosphere. NH4NO3 is assumed to be in reversible equilibrium with its gas precursors NH3 and HNO3. The equilibrium constant of the reversible reaction is defined as the product of NH3 and HNO3 vapor pressure above NH4NO3 aerosol. For solid aerosol, temperature dependence of the dissociation constant Kp of ammonium nitrate is given as [7]: Ln Kp = 118.87 - 24084/T - 6.025 Ln T

(1)

with T in K and Kp in ppbv2 with an accuracy of + 12%. When relative humidity is above the deliquescence point, NH4NO3 will exist as droplets. The equilibrium constant Kp' above the droplet is given as [7]: Kp' = [P 1-P2 (1-aw) Ln P1 =-135.94 Ln P2 = -122.65 Ln P3 = -182.61

+ P3 (1-aw)2] (1 +aw) 1.73 Kp + 8763/T + 19.12 Ln T + 9969/T + 16.22 Ln T + 13875/T + 24.46 Ln T

(2) (3) (4) (5)

Where aw is water activity and Kp is the equilibrium constant of solid NH4NO3.

105 Simultaneous measurements of NH3 and HNO3 with a time resolution of 40 minutes were made at three sites in the Netherlands: Speuld (coniferous forest) for a whole year in 1989; Leende (heathland) between April 25 and May 10, 1991; and Zegveld (grass pasture) from July 13 to July 20, 1993. A detailed description of the experiments at the three sites is given elsewhere [20-22]. A brief introduction to the experiments is given below. NH3 and HNO3 were measured at two heights at Leende and Zegveld and at one height in Speuld by wet denuders. Combining chemical and meteorological data, in total 240, 400 and 894 samples were obtained at Leende, Zegveld and Speuld respectively. NH3 concentration showed a diurnal variation with a maximum at nighttime and a minimum in daytime and HNO3 concentration showed a weak diurnal variation with slightly higher values in the late a~ernoon at the three sites, shown in figure 1. NH3 fluxes were generally directed to the surface at Leende and Speuld, while NH3 showed a bi-directional flux at Zegveld. For HNO3, its fluxes were observed to be directed away from the surface at Leende and Zegveld most of the time, as shown in figure 2. As it is believed that any surface is a perfect sink for

so-~ 2o-I

/+", "+"

+ +'++ + +~

4+, . . . .

:/ooo

/+\

-o-~ L

/ x;'•

--+--Nm-H

\

.~ooo_oo,

~

"+

/

9

.~+.+

0 6 .",

3

.,,,,

a

,,,,

• •215 x /

\•

• •

• x ~

"• ~ • • ""

""•

x • • • • • x' ' x • • • • • • • • ^ • 2 1 5 •

x

0

t

\

•215 \•215215215215

116.0

116.5

•

• 2 1 5 2 1 5 2 1 5 2 1 5 2 1 5 2• 21 15 5 2 1 5 •

117.0

117.5

118.0

Julian day

Figure 1. Concentration variations ofNH 3 and HNO 3 on April 26-27, 1991 at Leende in the Netherlands (L = lowest level, H = highest level)

0.5

/,,,,

0.0

,~

\d

-,,

\?

-0.5 -

\~ -1.0

'

0

I

4

'

--+--~o3 I

8

'

I

12

'

I

16

'

I

20

'

I

24

Time Figure 2. NH 3 and HNO 3 fluxes on April 26, 1991 at Leende in the Netherlands

106 HNO3, the upward fluxes of HNO3 are probably caused by the changes in the NH3/HNO3/ NH4NO 3 system.

3. RESULTS AND DISCUSSION Figure 3 presents the measured concentration products [NH3][HNO3] in Leende, together with theoretical curves calculated for equilibrium between aerosol and gaseous compounds, expressed as a function of temperature and relative humidity. The measured products in Leende show a tendency to exceed the equilibrium condition generally (figure 3). Below a relative humidity Rh of 60%, the measured products fit theoretical prediction for solid aerosol qualitatively with a slightly systematic positive bias (positive and negative bias are defined as measured products higher or lower than the equilibrium value), while they greatly exceed theoretical predictions for droplets. In contrast to Leende, the measured products for [NH3 ][HNO3 ] in Zegveld are generally lower than theoretical prediction when Rh is below 60%. As Rh increases, the measured concentration products exceed the theoretical prediction increasingly. In Speuld, the measured concentration products can be higher or lower than

100 - -

-

---SOLID

-

--3W---95

o

+

--X--90 --V-- g5

10 t~

>

,.Q

O

90

9 []

~o~ E! o

II

0.1

(a) Leende 0.01

[

3.4

'

I

'

3.5

I

3.6

'

I

3.7

1000/T Figure 3. Measured concentration products [NH3][HNO3] and theoretical equilibrium curves as a function of temperature and humidity at Leende.

107 theoretical predictions regardless of the fact that the aerosol was in solid or in aqueous phase. The significant characteristic in Speuld is that the measured concentration products tend to be smaller than theoretical predictions at lower relative humidity. The discrepancies from equilibrium are still large if the data with relative humidity > 80% and temperature < 5~ are rejected. Lewin et al. [ 13] obtained a good agreement between measured concentration products and theoretical equilibrium constants between HNO 3, NH3 and NH4NO3 at a temperature higher than 0~ at a rural site in the Northeastern U.S. on a 24 hours basis, although their data set was relatively small. In Speuld, a rural site in the Netherlands, disequilibrium was generally observed. Recently, Mozukewich [4] offered a model to calculate the dissociation constant of NH4NO3 aerosol as an explicit function of temperature and relative humidity, allowing a more precise evaluation of the equilibrium in different temperature and relative humidity. The ratio of measured concentration product Km=[NH3 ][HNO3] to the value Ke predicted by theory is defined as a measure of whether the system is in equilibrium. If the system is in equilibrium, the ratio should be near unity. Any ratio larger or smaller than unity means that system is not in equilibrium. Figure 4 depicts relative humidity and temperature dependence of this ratio Km/Ke at Leende. The ratios Km/Ke are mostly larger than unity, but some are slightly lower than unity at low humidity and high temperature. As Rh increases, the ratio increases and discrepancy from theory also increases (figure 4a). As temperature increases, the ratio decreases with a strong linear relationship (figure 4b). At Rh < 80 %, the ratios are ranging generally between 0.1 and 100. Figure 4 clearly shows that maximum disequilibrium occurs under conditions of high humidity and low temperature, while the concentration products are slightly under equilibrium at high temperature and low humidity. The same relations between humidity, temperature and the ratio Km/Ke are obtained in Zegveld and Speuld.

1000 -

~

lOO

~

1

~

- BB

~]

0.1

' 20

I 40

'

1000 -.

lOO

r'lffi

B o

~u

I 60

I

I 80

'

I 100

(a) Relative humidity (%)

0.1

i 270

280

I 290

I 300

(b) Temperature (K)

Figure 4. The ratio of measured concentration product to equilibrium value predicted by thermodynamics as a function of relative humidity (a) and temperature (b) at Leende.

108 1000

40 "" .

'

I

'

- I ~D~

'

I ' ---e---Km/Ke

I ~~

/%

--D--~

1

100

,,g ~

g

"~

10

o

eeee

c~ .

f 9

9e

N 0

'"~

116.0

I

I

I

116.5

117.0

117.5

4-

t~ 1

0.1 118.0

Julian day Figure 5. Diurnal variation of the ratio of measured concentration product Km to theoretical prediction Ke at Leende from April 26-27, 1991.

The ratio Km/Ke shows strong diurnal variation at Leende, as shown in figure 5. The ratios are much higher than unity in the night and the early morning and slightly lower than unity in the late afternoon. This variation is consistent with NH3 concentration changes (figure 1), showing the major role of NH3 in the disequilibrium. If the accuracy of the theoretical prediction is taken into account, the conclusion is that the system ofNH3, HNO3 and NH4NO3 is strongly above equilibrium in the night and nearly in equilibrium or under equilibrium at daytime in Leende. The results obtained here are slightly different from those ofErisman et al. [ 14] who found that a reasonable agreement existed between measured products and theoretical calculation for a temperature > 0 ~ and Rh < 80% at Cabauw in the Netherlands. The explanation for the disequilibrium between aerosol and gaseous compounds is believed to be kinetic constraints preventing rapid attainment of equilibrium. Gas-particle transformation processes are not governed by thermodynamic law only and kinetic constraints play an important role on the transformation. At the three sites, NH3 concentrations are higher at night and lower at daytime with a peak in the early morning while HNO3 concentrations shows a weak diurnal variation. Thus the concentrations products of NH3 and HNO3 do not show a strong diurnal variation with a peak in the afternoon, which is predicted by thermodynamic theory based on the variation of temperature and relative humidity, as shown in figure 5. The diurnal changes in temperature and humidity alter the dissociation constant at a rate which is sufficiently rapid to prevent equilibrium between NH4NO3 aerosol and its gas precursors. Similarly, changes in precursor gas concentrations can not be accommodated sufficiently rapid to maintain equilibrium conditions [ 18]. As shown in figure 5, it seems that kinetic constraints for gaseous reactions at low temperature and for aerosol evaporation at high temperature offer a reasonable explanation for the disequilibrium. In addition to kinetic constraints, high NH3 concentrations reduce HNO3 concentrations to such a low level that large measurement uncertainties may occur, resulting in departure from theory especially at high humidity. However, the measurement uncertainties are not the main reason for the disequilibrium because the measured products do not fit the theory well at

109 temperatures above 5~ relative humidity less than 80% and measured concentration products larger than 1 ppbv2. In Zegveld and Speuld, interference by dissociation followed by deposition of the products could cause a negative departure, because the sampling height is low and the wet surface of the canopy is a strong sink for HNO3 and NH3. The formation of hydrated gases, which are not accounted for in the theoretical model, may result in underestimation of the HNO 3 concentration especially at low temperatures and high humidity.

4. A FIRST ORDER ESTIMATION OF THE INFLUENCE OF THE NH3/HNO3/ NH4NO 3 SYSTEM ON THE FLUXES OF HNO 3 AND NH 3 In Leende, the interference on NH 3 and HNO3 fluxes by dissociation ofNH4NO 3 is obvious and large, especially for HNO 3, shown in figure 2. NH3 is deposited, while HNO3 fluxes are directed away from the surface. Since HNO3 shows a very weak peak in the late afternoon, the photochemical formation of HNO3 is not a reasonable explanation to the upward gradient of HNO 3. In Zegveld, NH4NO 3 and HNO 3 were measured by thermodenuder in July, 1992. Although the precision of the data can not allow calculation of aerosol flux, they reveal that the HNO 3 gradient and NH4NO 3 concentration show an anticorrelation, see figure 6. At high NH4NO 3 concentrations, apparent HNO 3 emissions are observed and at low NH4NO 3 concentrations, HNO 3 depositions are observed. In Manndorf, Germany where high ammonium sulfate and low NH4NO 3 concentration were observed, HNO 3 usually showed a deposition flux. In Leende, nitrate concentration was very high and about 70% of the nitrate existed as NH4NO 3. As figure 5 shows, in daytime, especially in the late afternoon, the system is nearly in equilibrium or under equilibrium. Decreasing nitrate aerosol concentrations were observed in Leende, as shown in figure 7. In daytime, NH4NO 3 aerosol dissociation is favored at low level compared to higher levels due to the influence of the dissociation followed by deposition. Equal amounts of NH 3 and HNO 3 are produced this way. The HNO 3 gradient is easily changed from downward to upward direction by this mechanism because HNO3 concentrations and gradients are small. At the night, the NH3, HNO3 and NH4NO3 are strongly above equilibrium and the system has shifted to aerosol formation. HNO3 upward gradient usually did not change very much, implying that aerosol formation rate was slow, otherwise, the strong NH3 deposition gradient would change the 20

O.8

/

/ .---4-- HNO3 | --o--NH4NO3

o ~ ~

Ti!

- 0.4 - 0.0

r,,2,

/ -

0 195

200

-w; i

I ~ 205 Julian day

&

I 210

-1.2 215

Figure 6. Anti-correlation between NH4NO 3 concentration and HNO 3 gradient in July, 1992 at Zegveld in the Netherlands

110 6 I I~! NIB/SO4 I-I NO3/SO4 I

O 2

0

i

10:00

i

i

t

11:30

12 38

i

i

13:37

17:05

Time

Figure 7. Aerosol NH4+/SO42- and NO3-/SO42" changes on April 26, 1991 in Leende

(a) April 26, 8:00 4

0

D

Measured

~

Corrected ~

(21

2 O

~

Corrected

~

0

D

;

OO

;

~

;

Measured

4

,'o

,;

(b) April 26, 13:00 4 0

0

0

~

~:

4

Measured

Cot,rected

O

Corrected .

--~

OO

/ O

Measured

o

0

HNO3 concentration (ug/m3)

NH3 concentration, ug/m3

Figure 8. Diagram of the influence ofNH4NO 3 dissociation on NH 3 and HNO 3 fluxes on April 26, 1991 at Leende in the Netherlands

HNO3 gradient greatly. At present, the kinetics of the NH 3, HNO 3 and NH4NO 3 equilibrium system are not well understood and NH4NO3 aerosol deposition is not measured precisely. It is very difficult to analyze quantitatively how the equilibrium influences the fluxes of the species involved in the system. Even though large uncertainties exist, the measurements of HNO3 and NH3 concentrations do indicate that the NH3/HNO3/NH4NO 3 system influences NH 3 and HNO 3 fluxes and the upward fluxes of HNO 3 are probably caused by NH4NO 3 dissociation. Based on the assumption that the HNO 3 upward gradient is caused by NH4NO 3 dissociation, a first order estimate can be made of the influence of the NH4NO 3 dissociation to NH 3 flux. The NH 3 and HNO 3 fluxes on April 26, 1991 in Leende are used as an example. The assumption is that the background HNO 3 concentration at 3.47 m is 1 Bg/m3, which is consistent with the HNO 3 concentration on May 3-4, 1991 when HNO 3 showed a deposition flux. The dry deposition velocity Vd ofHNO 3 at 3.47 m can be obtained by the inverse of the

111 sum of Ra and Rb, assuming Rc to be zero. From Vd and the HNO 3 concentration at 3.47 m, the corrected concentration ofHNO 3 at 0.93 m can be estimated, as shown in figure 8. The corrected NH 3 concentrations at both levels can be estimated from the measured and corrected concentration of HNO 3. Figure 8 shows two extreme cases: (a) high NH 3 deposition flux and (b) high HNO 3 upward flux. In both cases, NH4NO 3 dissociation has an enormous influence on the HNO 3 gradient and sometimes changes its direction. When the NH 3 gradient is large, NH4NO 3 dissociation has little influence on the NH 3 flux. In figure 8a, where the NH 3 flux is high, the measured NH 3 flux is estimated only about 5% lower than without the correction for NH4NO 3 dissociation. In figure 8b where the HNO 3 upward flux is high and the NH 3 gradient is small, NH4NO 3 dissociation has a greater influence on both HNO 3 and NH 3 fluxes. The measured NH 3 deposition flux is estimated 76% lower than without this correction. Figure 9 shows diurnal variations of HNO 3 and NH 3 fluxes before and after the correction for NH4NO 3 dissociation. Since HNO 3 usually shows an upward gradients at Leende, NH 3 deposition fluxes are underestimated or NH 3 emission fluxes are overestimated in the late afternoon because NH4NO 3 dissociation influences the NH 3 gradient to a large extent during that period.

D

l i t

0.5 -fj~

- - D - - NH3_M - - o - - NI-t3 C

/

9

~

[]

Dn--12_

-0.5

-

-1.0

-

I

116.0

116.5

'

I

117.0

'

I

117 9

'

I

118.0

Julian day Figure 9. The influence ofNH4NO 3 dissociation on NH 3 flux on April 26-27, 1991 in Leende (NH3_M: measured NH3 flux, NH3_C" corrected NH3 flux)

5. CONCLUSION The three data sets collected in 1989, 1991 and 1993 in Speuld, Leende and Zegveld in the Netherlands are investigated to evaluate the atmospheric equilibrium ofNH3, HNO3 and NH4NO3 aerosol. A very poor agreement between measured gaseous concentration products Km=[NH3][HNO3 ] and theoretical predicted equilibrium vales Ke is found at low relative humidity 9The comparison between the measured Km and theoretical Ke reveals that NH3 and HNO3 are generally not in equilibrium with NH4NO3 aerosol, because the ratios of Km to Ke are far from unity, varying from less than 0.01 to above 100. The ratio increase with increasing relative humidity and decreasing temperature. The maximum positive departure from equilibrium usually occurs at low temperature and high relative humidity, which is consistent with most field measurement 9The maximum negative deviation usually happens at

112 high temperature and low relative humidity. Kinetic constraints upon the attainment of equilibrium for the NH3-HNO3-NH4NO3 system are emphasized by the results obtained at the three sites. Disequilibrium between NH3, HNO3 and NH4NO3 can be used to explain why HNO3 upward gradients in Leende are observed. The dissociation ofNH4NO 3 is responsible for the upward gradient of HNO 3 and results in a slight underestimation of the NH 3 deposition flux during the night. Large errors in the estimation ofNH 3 emission or deposition fluxes are caused by the influence ofNH4NO 3 dissociation in the late afternoon when NH 3 gradients are small.

Acknowledgment. The authors gratefully acknowledge Geoffrey DoUard, AEA technology, Harwell, UK for providing aerosol data at Leende in the Netherlands.

Permanent address: Center of Environmental Sciences, Peking University, Beijing 100871, P.R. China

6. REFERENCES

[1]

[21 [3] [4]

[5] [6]

[7] [8] [9] [ 10] [11] [12] [13] [14] [15] [ 16] [ 17] [18] [ 19] [20] [21 ] [22]

Sutton M. et al. (1993) Exchange of ammonia and sulphur dioxide with vegetation. EUROTRAC/BIATEX, annual report 1993, part4, Biatex, 213-224. Huebert B. J. et al., J. Geophys. Res., 93D(6) 7127-7136 (1988). Kramm G. and Dlugi R., J. Atmospheric Chemistry. (in press) (1993) Stelson A. W. et al., Atmospheric Environment 13, 369-3 71 (1979). Stelson A. W. and Seinfeld J. H., Atmospheric Environment 16, 983-992 (1982). Stelson A. W. and Seinfeld J. H., Atmospheric Environment 16, 2507-2514 (1982). Mozurkewich M., Atmospheric Environment 27A, 261-270 (1993). Doyle G. A. et al., Environmental science & technology 13, 1416-1419 (1979). Tanner R., Atmospheric Environment 16, 2935-2942 (1982). Cadle S. H. et al., Atmospheric Environment 16, 2501-2506 (1982). Harrison R. M. and Pio C. A., Tellus 35B, 155-159 (1983). Hildeman L. M. et al., Atmospheric Environment 18, 1737-1750 (1984). Lewin E. E. et al., Atmospheric Environment 20, 59-70 (1986). Erisman J. W. et al., Atmospheric Environment 22, 1153-1160 (1988). Allen A. G. and Harrison R. M., Atmospheric Environment 23, 1591-1599 (1989). Pio, C. A. and Numes, T. V., Atmospheric Environment 26A,505-512 (1992). Chang Y. S. et al., Atmospheric Environment 20, 1969-1977 (1986). Harrison R. M. and Mackenzie A. R., Atmospheric Environment 24A, 91-102 (1990). Jeffe D. (1988) Accuracy of measured ammonium nitrate equilibrium values. Atmospheric Environment 22, 2329-2330. Slanina J. et al. (1990) Acidification research at ECN. Report ECN-C-90-064, Netherlands Energy Research Foundation, Petten, The Netherlands. Duyzer, J. H. et al. (1992) The joint experiment on surface exchange of trace gases over the Leende heathland. Report TNO PU 92/015. Plantaz M. A. et al. (1993) Continuous monitoring of deposition fluxes of nitrogen compounds on low vegetation, in: Eurotrac Annual Report 1992, part 4, B IATEX.

ATMOSPHERIC DEPOSITION S E S S I O N III PARTICLE DEPOSITION



115

Particle deposition to forests Jan WiUem Erisman', Geert Draaijers', Jan Duyzerb, Peter Hofschreuderc, Niek van Leeuwen d, Ferdinand Rtimerc, Walter Ruijgrok*, Paul Wyers f

"RIVM, P.O.Box 1, 3720 BA Bilthoven, the Netherlands. b TNO, P.O.Box 6011, 2600 JA Delft, the Netherlands c Department of Air Quality, Agricultural University, WAU, P.O. Box 8129, 6700 EV Wageningen, the Netherlands d Department of Physical Geography, University of Utrecht, P.O.Box 80.115, 3508 TB Utrecht, the Netherlands c KEMA, P.O.Box 9035, 6800 ET Arnhem the Netherlands f ECN, P.O.Box 1, 1755 ZG Petten, the Netherlands

Abstract Particle deposition to forest was studied using experimental and modelling results. Results show that the deposition of particles to forests has been underestimated until now. Particle deposition makes out reasonable contribution to the total deposition of acidifying components and base cations to forests. It was estimated that at Speulder forest the contribution of dry particle deposition to the total deposition was 18% for SO4, 38% for NO3, 23% for NH4, 56% for Na, 47% for K, 69% for Ca and 65% for Mg.. Deposition of compounds via fog at the Speulder forest was estimated to be small ( 2.5 ~m) during a period of nine months [ 18]. Concentrations of HNO2, HNO3, HC1, SO4, NO3, and NH4 were determined during several days with annular denuders [19] and once every week as hourly

118 values with wet rotating denuders [20]. Canopy exchange research using labelled S compounds (S 35) has been performed by ECN [20]. Fog composition, fog droplet diameter and liquid water content were measured by KEMA and ECN during several measuring campaigns [ 18,20,21]. In this way all ingredients determining throughfall and atmospheric deposition were measured. Aerosol deposition:

~=n~canemometer xo~,

etc.

Gas deposition:

) ~

s~

anemometer probe

'11

I ~.c~

qC:~r--.............. 35 m: tunnelsampler; PM-IO; cyclone; filterpack, DFM and a wet-only sampler, ADS 34 m: S04 denuder ................ 32 m: ADS

I1 =

,,~ JT/RH "t ~

36 m, reference height:

so2.

NHS.

eddy correlation NO2 32 m, SO2, NO2

r valve system 29 m: tunnelsampler; PM-10; cyclone; filterpack, DFM sonic anemometer probe~ liquid water content sensor

FSSP

"= < ~ ................. 28 m: filterpack; mist sampler

(L.

ASASP-X, sonic anemometq

-

-= -~- ........ 28 m0 SO2, NO2, NH3

26 m: SO4 denuder, aerosol concentrations "= ~

-

................. 24 m: filterpack, wet denuder (HNO3, HNO2, HCI, SO2)

-.,

24 m, SO2, NO2, NH3

22 rn: SO4 denuder; ADS; DFM __ ASASP-X, sonic anemom~

~I~--: ............... 21 m: tunnels,ampler; PM-10; cyclone 19 m: Pb-214:High Volume sampler, (artificial) branches

J 2.5 lain 4.5 1.6 fog droplets (December) 19.4 fog droplets (February) 7.4 The factor which is important for particle deposition, next to the size distribution of particles, is the friction velocity. If the size dependency and the u, dependence of the dry particle deposition velocity is similar for the measurements and the model results, this might serve as some sort of validation of the most determining processes. Figures 2 and 3 show the u, dependence of different experimental determined Vd and modelled Vo, respectively. In general the experimental and model results show a distinct influence of the size distribution and u, on the dry deposition velocity of particles. Both figures show that Vd (2~4pb) < Vd (SO4) < Vo (NO3) < Va (fog). Furthermore, the figures show similar relations between Va and u, for three of the categories distinguished above, although the variation in measured Va per u, class can be very high [10]. The category with base cations is not considered because no u, dependent measurements were made. The deposition velocity of fog is proportional to u, 2 indicating that impaction is the most important process determining Vd. Furthermore, sedimentation is important. The Vd values of other compounds are proportional to u, or a weak function of u, (~u,~a), indicating no distinct process is dominating rather a mixture of processes is occurring simultaneously [26]. Results are in line with other investigations, showing that particle deposition to forests can be considerable with Vd of several cm s-~ [27-29]. The relationships with u, found for different particle size classes provide confidence in the experimental results. This can only be considered as indication, because errors in deposition estimates usually show a high correlation with u,, because factors influencing the error in the deposition estimates highly correlated to u,. Evaluation of the integral model results can only be done with statistical parameters, including the uncertainty in modelled and measured values. The uncertainty in model results and the sensitivity of model results on input and model parameters has extensively been described in [10]. The overall error in modelled Vd integrated over the size distribution representative for acidifying aerosols is estimated about 65% [ 10]. For base cations this error is somewhat smaller (60%) because of the contribution of the relatively well parametrised description of sedimentation. The uncertainty in model estimates is lower than or about equal to the uncertainty in measurement results, with the exception of the fog deposition measurements which are estimated to have smaller errors (20%). The fractional bias of the

121 means (the relative difference between the mean calculated and observed value) falls within these limits. The relatively large sensitivity of the model and an inaccuracy of the same order in measuring results cause that a perfect 1"1 correspondence between both cannot be expected. Statistical testing of the difference between modelled and measured values is done using the non-paramatric sign and Wilcoxon tests for paired samples. Both test revealed no significant differences between the mean values of modelled and measured fluxes or Va'S [ 10]. This is of course mainly the result of the large standard deviations in measuring and modelling results. 0.1 0.08

/

.~ 0.06 ... 0.04 >

7

0.02 0 -0.02 0

0.2

0.4

0.6

0.8

=

fog

[3

NO3 (TNO)

*

SO4 (TNO)

0.02

ol

-0.02 [ 0

0.2

0.4

0.6

0.8

U* (m/s)

=

fog

C

NO3 (TNO)

9

SO4 (TNO)

SO4 (ECN) 214Pb

Figure 3. Summary of modelled Vd versus u,. It can be concluded that there are no strong indications for a significant underestimation or overestimation of the modelled Va compared to measured values. The fact that the model is reasonably capable of describing a similar response of Va to u. as the measurements for different particle diameter ranges provides confidence in the process descriptions. Furthermore, it is concluded that particle Va to forests and probably other rough surfaces are high. Average Vd values for half a year for fine particles at Speulder forest range from 1 to 2 cm s -~ (SO4, NO3, and NH4) with daytime values being 1.3 + 1.2 cm s~ and night-time values being 1.0 + 1.4 cm s~ (SO4). Vd values for coarse particles are about 5 cm s~ with daytime values of 5.1 :!: 3.9 cm s ~ and night-time values of 4.8 + 4.0 cm s~. In comparison, for the same period, Vd values for SO2, NH3 and NO2 were 1.5, 2.5 and 0.1 cm s l, respectively. This means that aerosol Va to forest canopies in The Netherlands is underestimated until now with a factor of 2 to 3. For forests in Europe this is even higher, taking the EMEP model results [30]. It was estimated that at Speulder forest the contribution of dry particle deposition to the total deposition was 18% for SO4, 38% for NO3, 23% for NH4, 56% for Na, 47% for K, 69% for Ca and 65% for Mg.. Deposition of compounds via fog at the Speulder forest was estimated to be small (

].0 t

.

" .

.

.

.

.

.

.

.

.

.

.

.

~. ~. . , . , . ~ , , N 9

. . . . .

,,

Hour

4.0

,,-,-,

cm]8 (e)

,~ ,,

, ~

~

ADOM I

~ ~176176176176176 ~176

|

o

a

o~

.....

, ~ ,,, , ~ . ~ , ,

--~-- Multi-Layer - ~ - Big Leaf ......

/

. . . . . . . . . . . . . . . . . . . . . . . . .

0.5t 0.0

Hour

i o Obs. 1.5 cm/s

-

o

~

-

-

:,-_'~.~

.....

0.0

Hour

.

.

.

.

3.0 2.0 1.0 Lm. 9 ~ 0.0 Hour

Mean diurnal patterns in Vd (cm s'l). (a) E M E F S I. (b) E M E F S Figure 3 II. (c) V i n e y a r d . (d) C o t t o n (e) E M E F S II SOe Model performance for SO2 was generally not as good as for 03. According to the observations there was an increase in Vd during the day (Fig. 3e), but

194 no obvious midday peak. The Big Leaf model and RADM did not match this pattern. Reasons for this behaviour were discussed above. The Multi-Layer model performed surprising well with no apparent bias during periods without sunlight and only a slight low bias during the day. ADOM predicted some structure in the hourly variations of Vd and there was a tendency for a peak in the early morning hours. This m a y have been a result of surface wetness. There are a n u m b e r of other interesting features in Figures 3a-4e, which are currently being examined in more detail. For example, the Big Leaf model predicts a large peak in 03 Vd in the late afternoon and early evening during E M E F S I. This behaviour has not been fully explained, but some of it was due to the variation in Re. Figure 5 is a plot of the mean diurnal p a t t e r n in Rc+Rb. The Big Leaf model resistances were smallest at approximately 1830 LST and remained below the ADOM, Multi-Layer and RADM values until 2000 LST or later. I

Multi-Layer

o

!

- , * - - -Big Leaf

- N-ADOM

--

RADM I

s/cm ,

"~,/" .

~1

-'~~ 9

_~. . . . . . . . . . . . . . . . . . . . . . .

................ ........ .

0

Figure 4

.

1

.

.

2

~.- : -k- - ' ~

.

.

3

.

.

4

.

5

.............

oo~:9 g

/

ol- i - ~ ~

~j~.~,

.

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 Hour

Diurnal pattern in modelled Rc+Rb (s cm -1) for E M E F S I.

Mean Vds were determined for the entire durations of the field studies. These values are listed in Table 4. For 03, the n u m b e r of 30 min values t h a t were included in the average ranged from about 500 to 1000, depending upon the study. Fewer observations were available for SO2. With the exception of the Multi-Layer model during EMEFS II and ADOM over grape, the overall tendency was for the models to underestimate Vd. For 03, the average amount of bias (difference between the observed Vd and the m e a n of the 3 or 4 models) was -17% for the leafless forest, -22% for the forest with leaves, -52% for cotton and -16% for grape. For SO2 the average bias was -18%. As seen in previous figures the models differed from one another substantially. Compared to the mean Vd across all 3 or 4 models, the range of m e a n Vds for the individual models were: +19% t o - 2 9 % for EMEFS I; +59% t o - 5 2 % for EMEFS II; +14% to -16% for cotton and; +28% t o - 1 8 % for grape. The variability among models was larger for SO2 ranging from + 140% to -84%.

195 Table 4

Comparison of Overall Mean Deposition Velocities (cm s "1)

Site ....... EMEFS I EMEFS II Cotton Grape stable unstable B90-SO2

0bs. ...........Multi=Layer .................B i g L e a f ...............~ 0 M ...........~ M 0.67 0.50 0.62 0.60 0.37 0.18 0.23 0.07 0.15 0.13 0.50 0.20 0.24 0.27 0.31 0.22 0.24 0.34 Stratified by Stability (EMEFS I) 0.44 0.28 0.42 0.44 0.18 0.91 0.73 0.84 0.79 0.59 0.52 0.52 0.07 1.02 0.10

CONCLUSIONS There are a variety of approaches to modelling and/or parameterizing the deposition velocity of acidifying pollutants to surfaces. In this paper, 4 approaches were compared for 4 different sets of conditions. The a m o u n t of discrepancy between models and observations varied among models and from one surface type to the next. In addition, the agreement between models and observations decreased with increasing temporal resolution and reasons for agreement or lack of agreement varied among models. These results clearly suggest t h a t the models are not adequately representing the underlying processes involved in pollutant dry deposition. Results of comparisons, such as those presented here, can be used to improve the models. However, before models can be fully tested, there is a need for more flux m e a s u r e m e n t s for a larger variety of surface and atmospheric conditions and chemical species. More research is needed. Given the range of predictions t h a t arise from different models it is apparent t h a t for monitoring total deposition of air pollutants it will be important for all groups/countries involved to coordinate their modelling efforts.

REFERENCES

Baldocchi D.D., Hicks B.B. and C a m a r a P., 1987: A canopy stomatal resistance model for gaseous deposition to vegetated surfaces. Atmos. Envir., 21, 91-101.

196 Byun D.W., 1990: "Turbulent Transfer in the Atmospheric Surface Layer." in D.A. Haugen, (ed.), Workshop in Micrometeorology, pp. 381-392. American Meteorological Society. Hicks B.B., Wesely, M.L. and Durham J.L., 1980: Critique of methods to measure dry deposition. Workshop Summary, U.S. Environmental Protection Agency EPA-600/9-80-050, NTIS Publication No. PB81-126443, 70pp. Hicks B.B., Baldocchi, D.D., Meyers T.P., Hosker R.P. Jr. and Matt D.R., 1987: A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Wat. Air and Soil Pol., 36, 311-330. Louis J., 1979: A parametric model of vertical eddy fluxes in the atmosphere. Boundary-Layer Meteorol., 17, 187-202. Meyers T.P. and Baldocchi D.D., 1988: A comparison of models for deriving dry deposition fluxes of 0 3 and SO 2 to a forest canopy, Tellus, 40B, No. 4. Padro J., den Hartog G. and Neumann, H.H., 1991: An investigation of the ADOM dry deposition module using summertime 03 measurements obove a deciduous forest. Atmos. Envir., 25A, 1689-1704. Padro J., Massman W.J., Shaw R.H., Delany A. and Oncley S.P., 1994: A comparison of some aerodynamic resistance methods using measurements over cotton and grass from the 1991 California Ozone Deposition Experiment. Boundary-Layer Meteorol., (in press). Wesely M.L., 1989: Parameterization of surgace resistance to gaseous dry deposition in regional-scale numerical models. Atmos. Envir., 23, 12931304.


197

EDACS: European Deposition maps of Acidifying Components on a Small scale

W.A.J. van Pul, C.J.M. Potma, E.P. van Leeuwen, G.P.J. Draaijers, J.W. Erisman Laboratory for Air Research, LLO National Institute of Public Health and Environmental Protection, RIVM P.O. Box 1, 3720 BA Bilthoven, the Netherlands

Abstract In this paper a description is given of the EDACS model (European Deposition of Acidifying Components on Small scale), with which the deposition of acidifying components on a small scale over Europe is calculated for 1989. The acidifying components considered in EDACS are sulphur and reduced and oxidized nitrogen compounds. Dry deposition is estimated with the inference method i.e the deposition at the surface is inferred from the concentration and the deposition velocity at the same height. The deposition velocity is calculated using a resistance model in which the transport to and absorption or uptake of a component by the surface are described. Dry deposition velocity fields over Europe are constructed from a detailed land-use map (1/6~ ~ lat/long grid, made by RIVM) and meteorological information using a detailed parameterization of the dry deposition process. These small-scale dry deposition velocity fields are combined with concentration fields from the EMEP Lagrangian long range transport model to yield dry deposition amounts on a small scale. Wet deposition is also estimated, based on measurements, to obtain a total acidifying deposition map at a European scale. These deposition fields clearly reflect the spatial detailed land-use information and the large-scale concentration pattern over Europe. The maps of the acidifying components over Europe on a small scale are made in cooperation with the EMEPkMSC-W, Oslo, Norway.

I Introduction In Europe, sulphur and reduced and oxidized nitrogen compounds are found to acidify soils and surface waters. Furthermore, nitrogen deposition causes eutrophication. The effects of acidification and eutrophication have been described extensively in the literature (e.g. Heij and Schneider, 1991, Grennfelt and ThOrnel6f, 1992). Especially the effects on ecosystems have obtained great attention due to their vulnerability. In order to protect these ecosystems critical loads have been defined above which there is an increased risk of damage to the ecosystems. The main atmospheric pathways via which emitted acidifying components reach the earth's surface are dry and wet deposition. The acidifying components can be transported over a long range up to 1000 km or more dependent on the component properties and the dry and wet deposition processes. Several models exist for estimating long-range transport of acidifying components on a European scale (e.g. EMEP: Sandnes, 1993, TREND: van Jaarsveld and Onderdelinden, 1994). The purpose of these models is to describe the relation between source

198 and receptor. The model results are used in quantifying country to country budgets, as a basis for the sulphur and nitrogen protocols and in assessments of the effects of acidification. The horizontal spatial scale on which these models operate is typically 50x50 km. However, describing the effects of acidification on the level of ecosystems, the acid load should be available at the scale of ecosystems or at scales which allow for comparison with critical loads (i.e. typically in the order of a lxl km resolution, Hettelingh et al., 1991). So the above model resolutions are not appropriate to describe the acidifying load at the level of ecosystems. A method was presented by van Pul et al., 1992a, with which the deposition of acidifying components on a small scale over Europe can be mapped. Their method was discussed at the ECE-EMEP/BIATEX Workshops on deposition at GOteborg, Sweden (LOvblad et al., 1993) and Aveiro, Portugal (Slanina et al., 1993) and accepted as currently the best available method to describe local acid deposition fluxes. In this paper the method is described and the fast, preliminary, maps of small-scale fluxes of acidifying components over Europe are presented. The calculations are made with the EDACS model (European Deposition of Acidifying Components on Small scale) in which the method is adopted. Recommendations made during the Workshops are incorporated in the model. The emphasis in the method is on modelling local scale dry deposition fluxes. A detailed parameterization of the dry deposition process for each acidifying component is based on available experimental results (EUROTRAC~IATEX project) and literature (Erisman et al. 1994a). Wet deposition is also estimated to obtain a total acidifying deposition map over Europe (van Leeuwen et al., 1994). The dry deposition maps were presented at the EUROTRAC symposium at Garmisch Partenkirchen, April 1994 by Erisman et al.(1994b) and van Pul et a/.(1994a). The maps of the acidifying components over Europe on a small scale are made in cooperation with the EMEtXaMSC-W, Oslo, Norway.

2 General descriptionof E D A C S An overview of the input for and calculation scheme of EDACS is presented in Figure 1. In EDACS the dry deposition is estimated with the inference method (Hicks, 1986). The deposition at the surface is inferred from the concentration and the deposition velocity at the same height. The deposition velocity is calculated using a resistance model in which the transport to and absorption or uptake by the surface of a component are described (see Section 3). The parameterizations are dependent on surface characteristics and other environmental and meteorological conditions. Dry deposition velocity fields, on a 6-hourly basis, are constructed from a detailed land-use map using these parameterizations along with the meteorological information (Potma, 1993). Here a RIVM data base is used which contains land-use data on a 1/6~ ~ lat/long grid over Europe (van de Velde et al. 1994). Finally the dry deposition amounts are calculated by multiplying these dry deposition velocity fields with concentration fields. In principle the concentration data can originate from measurements, model calculations or a combination of both. However, the spatial resolution of the operational European and national networks (ECE-EMEP, EUROTRAC) is too coarse to provide the necessary data and not every component is measured. This will lead to large uncertainties in the interpolated concentration fields. On a local scale, national or local networks may provide the concentration data. When the local maps are aggregated into one European map, problems

199 may arise about the inconsistency between the networks and it is foreseen that still not a full coverage of Europe can be obtained. However, for parts of Europe with small horizontal concentration gradients this can be done e.g. for UK:UK Review group on acid rain, 1990; Sweden: l_2)vblad et al. 1993, The Netherlands; Erisman 1992). In this paper the concentration data of the EMEP Lagrangian long-range transport model (hence EMEP-LRT) on a 150x150 km scale were used (as described in Sandnes, 1993) to obtain a consistent concentration field over Europe. Using calculated concentration fields, the relation between emissions and deposition is maintained and assessments or scenario studies can be made at different scales. In the inference method it is assumed that a constant flux layer is present between the reference height and the surface i.e. the atmospheric surface layer. This assumption implies that there is no significant advection in the layer and the air flow is well-adapted to the surface properties of the depositing surface and chemical reactions are not present. In that

Figure 1 Overview of the input for and calculation scheme of EDACS. For explanation see texL

200 case the deposition flux at the reference height equals the deposition flux at the surface. The adaptation of the air flow to the surface is strongly dependent on the surface roughness and the stability of the air. The choice of the reference height is a compromise between the height where the concentration is not severely affected by local deposition and is below the surface layer height. In EDACS the concentration at 50m is taken which is the lowest LRT model level above the surface. This concentration then is assumed to be representative for a certain area, here an EMEP LRT gridcell of 150x150 km, and consequently can be used in estimating the deposition to surfaces within this area. The wet deposition of acidifying components is based on measurements of the concentrations in precipitation and rain amounts (van Leeuwen et al., 1994). A concentration map over Europe was constructed from these data by kriging. A data set with interpolated values of long-term yearly precipitation amounts was used to calculate the wet deposition. The components considered here are SO 2 and SO42--aerosol (SOx), NO, NO 2 (NOx), HNO 3 and NO3--aerosol and NH 3 and NHa+-aerosol (NHx). NH x is considered to be acidifying because of the nitrification processes in the soil in which H ÷ is produced (van Breemen et alo 1982). If all above components which are deposited produce one equivalent of acid this leads to a total potential acid load which is estimated from: potential acid = 2 SO x + NOy + NH x. The actual acid load differs from the potential load because of an incomplete nitrification of NH 3 and by neutralization of the acidity by base cations. HONO, PAN and HNO 2 are not taken into account. However, the contribution of these components to the total acidifying deposition is very small (e.g. L6vblad et al., p 19).

3 Parameterization of the dry deposition velocity The dry deposition flux of gases and particles from the atmosphere to a receptor surface is governed by the concentration in air and turbulent transport processes in the boundary layer, by the chemical and physical nature of the depositing species and by the efficiency of the surface to capture or absorb gases and particles. The flux of a trace gas is given as: F : Va(z) c(z)

(1)

where c(z) is the concentration at height z and V,t is the dry deposition velocity (Chamberlain, 1966). z is the reference height above the surface: here taken as 50m. If the surface is covered with vegetation, a zero-plane displacement, d, is included: z=z-d. The absorbing surface is often assumed to have zero surface concentration. This holds only for depositing gases and not for gases that might also be emitted, such as NH 3 and NO. For these gases a non-zero surface concentration, a compensation point cp, might exist, which can be higher than the ambient concentration, in which case the gas is emitted. Here the concentration at the various surfaces, c s, is assumed to be zero for all components because of insufficient knowledge of the compensation point. The parameterization of the dry deposition velocity is based on a description of this process via a resistance analogy or Big Leaf Model (see e.g. Thom, 1975, Hicks et al., 1987, Fowler, 1978). In this resistance model the most important deposition pathways via which the component is transported and subsequently destroyed at, or taken up by the surface, are parameterized. The resistance model used here is shown in Figure 2.

201

C(z) ra

rb

rstom

rinc rext rs~

~

C~(soil) C~(ext)

rm

Cs(m)

m

Figure 2 Resistance or Big Leaf model used in EDACS. c,(soil), c,(ext) and cs(m) denote the concentration at the soil external or mesophyl surface respectively, for explanation of the other symbols see text). V d is the inverse of three resistances:

(2)

V d - ( r a + r b + r s ) -1

These three resistances indicate the three stages of transport: the aerodynamic resistance, r a, represents the resistance against turbulent transport of the component close to the surface; the quasi-laminar sublayer resistance, r b, accounts for the transport of the component through a laminar layer adjacent to the surface by molecular diffusion and the surface resistance r s for the uptake or destruction at the surface. This surface resistance is composed of the resistances of the various destruction or uptake processes at the surface. For a surface covered with vegetation this is: -the stomatal resistance, rstom , the resistance to the transport through the stomata of leaves and needles; -the mesophyl resistance, r m, the resistance of the internal plant tissues against the uptake or destruction (in a chemical way); -the cuticle resistance, reut, or external surface resistance, rext, the resistance of the exterior plant parts against the uptake or destruction of the component; -the ri~ the in-canopy aerodynamic resistance to account for the transport of air above the vegetation towards the soil and lower plant parts; -rsoa, the soil resistance, the resistance against destruction or absorption at the soil surface; These resistances which act in parallel or series are summed up to yield a (total) surface resistance, rs: rs =

-1

(rin c + rsoil )-1 + rex t + (r m + rstom)

-1] -1

(3)

202 For a water surface: r s = rwat, where rw~t is the resistance against the solution of gases in water. For bare soil: r s = rsoa and for urban areas: r s = rurban. When the surface is covered with snow r~ = rsnow.

In turn, these resistances are affected by meteorology, leaf area, stomatal physiology, soil and external leaf surface pH, and presence and chemistry of water drops and films. Especially the state of the leaf and soil surface i.e. the presence of water films and snow, is an important variable in the deposition of soluble gases such as SO 2 and NH 3. The process of dry deposition of particles of acidifying components is not very well known compared to the gaseous counterparts (Ruijgrok et a/.,1993). As a best estimate the dry deposition of particles is described using a parameterization by Wesely et al.(1985) and Erisman (1992). Recent information on the deposition of particles to forests has come available (Ruijgrok et al. 1994) and will be used in a future version of the model. The scheme used here to derive the surface resistances for SO 2, N O 2, NO, HNO 3, NH 3 is described in Erisman et al.(1994a). This scheme is based on previous publications among others Wesely (1989), l./Svblad et a1.(1993) and recent dry deposition measurements (among others in the BIATEX project of EUROTRAC). More details on the actual parameterizations used in EDACS can be found in van Pul et a1.(1995).

4 Results

The 6-hourly deposition velocity fields were averaged to daily values and multiplied with the daily EMEP/LRT concentrations. These daily dry deposition maps were summed to annual totals. In Figure 3 the dry deposition of total potential acid over Europe estimated with EDACS is shown. The dry deposition values for most components vary greatly over Europe. This is partly explained by variations in the deposition velocity caused by variations in land use and meteorological conditions over Europe. The concentration pattern of the components over Europe, which are associated with the distribution of emissions, introduce variations on a larger scale (150xl50km i.e. the EMEP grid). Large emission areas can be detected in the maps e.g. for SO 2 and NO~ this is the so-called black triangle (Eastern Germany - PolandCzech Republic), for ammonia e.g. north western Europe (The Netherlands, Denmark). The total potential acidification map which is the sum of the dry deposition and wet deposition is presented in Figure 4. This figure reflects the above mentioned variations. The relative contribution of dry and wet deposition can be observed. For instance in the Scandinavian countries the surface inhomogeneities are not represented due to the large contribution of the wet deposition which has a smooth distribution over this area. In Figure 5 the standard deviation of the total deposition of EDACS cells in an EMEP-grid cell (about 100 EDACS cells in one EMEP-grid) is given as absolute values and values relative to the average per EMEP-grid. It can be seen that the largest absolute values of the standard deviation can be found in the above described emission areas. Whereas the relative standard deviation is largest in areas with a small deposition and so variations in land use, meteorology etcetera are reflected.

V

~

~

~

~

A

I DDlnnnn

..

...

~

_

aqi,

,

_

-=.

"-~

,,,

203

g~

r~

g~

t~

Figure 4 Potential acidification map over Europe on 1/6°xl/6 ° lat/long grid for 1989 (tool ha "1 year1).

205

Figure 5 Standard deviation of the total deposition of EDACS grid cells per EMEP grid in absolute values in mol ha -1 year -~ (above) and values relative to the total deposition per EMEP-grid cell in % (below).

206

5 Uncertainties The maps shown in Section 4 have a limited accuracy and are therefore preliminary. The aim of these maps is to show the variations of the deposition of acidifying components on a small scale. Several uncertainties and shortcomings are present which need some discussion. We will adress these items here and will suggest a quantification of the uncertainties. Also some recommendations for improvements of these maps are given. One of the main uncertainties in the maps is in the simple resistance model and especially the surface resistance parameterization for estimating the dry deposition of different gases and particles. The resistance model is a simple approach for a highly variable process. It assumes a constant flux layer, i.e. there are no surface inhomogeneities, edge effects or chemical reactions. How much these simplifications contribute to the total uncertainty in the annual average deposition fluxes has not been investigated. The uncertainty in the surface resistance parameterization is the largest uncertainty in this simplified scheme. Therefore more, and more accurate, parameterizations are needed for various vegetations and surfaces. Moreover there is a lack of measurements on which these parameterizations can be based especially for southern and eastern European climates and surfaces. This is needed to obtain parameterizations for use in LRT models which are valid for the whole of Europe. Part of the uncertainty in the surface resistance parameterization is due to the mismatch between the available parameterizations for a limited number of landcover types and the landuse classifications used in the RIVM land-use data base. Surface wetness is found to be one of the major factors influencing the deposition process. In the present version of EDACS only rain and an indication of dew is used. In the next version the dew amount will be modelled in more detail using appropriate surface properties. The evaporation of rain and dew will also be parameterized. This means that an administration of the available energy and moisture flux during the day has to be made. An indicator on the presence and condition of a snow layer will also be taken into account. The overall uncertainty in the surface resistance due to the above factors is different for each component and surface type. This uncertainty, on a annual basis, is a few tens percent points but can easily exceed 100%. In the current version of EDACS, the EMEP-LRT concentration maps on a 150x150 grid are used. The uncertainty in the concentrations are estimated at 40-70% by a statistical analysis with the EMEP measurements (Krtiger, 1993). These concentrations represent the background situation in Europe. It is assumed that the concentration distribution within a grid is homogeneous. This is not the case in a grid which contains industrialised areas or many scattered sources such as of NH 3 and NO x. For such conditions, subgrid concentration variations are present and will lead to underestimates of the deposition in that grid. To obtain an indication of the errors, a small-scale, short-range model can be useful here to resolve subgrid concentration gradients for dense source areas. The uncertainty in the deposition in an EDACS grid cell due to these gradients is estimated at 25% (Berg and Schaug, 1994). The deposition in EDACS is based on the EMEP-LRT concentrations which in turn are dependent on EMEP deposition estimates. The deposition in the EMEP-LRT model and in EDACS are calculated in different ways. By using other dry deposition velocities in EDACS a mass inconsistency, between the EMEP calculated deposition and the small scale maps by EDACS, is introduced. However, if the differences in the used deposition descriptions between the two models are not very large and non-systematic over a larger region, this will

207 not lead to large mass inconsistencies. In Figure 6 a comparison between the sulphur dry deposition per country estimated by EMEP and EDACS is shown. It can be seen that on average there is a good agreement indicating that for the model area the mass consistency is not violated to a large extent. However, for some countries the deviations can be as large as 50%. To avoid this mass inconsistency it is planned to implement the deposition module in the EMEP-LRT model. In this way the calculated concentration fields are consistent with the EDACS deposition description.

2000 0 0

E 1500

.S

0 >

o

1000

0 0

w

500

QO

I

2 0 0

500

1000

1500

2000

EMEP country averages (mol/ha/y)

Figure 6 Comparison between the country averaged dry deposition of sulphur calculated with EMEP and EDACS (deposition in tool ha s year'1). In inferential modelling the concept is used that the surfaces, at which the deposition is calculated, should have a certain horizontal length. As an approximation this is about 100 times the reference height (Pasquill, 1972). The typical horizontal length scale of the surface, using 50m as the reference height, is 5 km. This means that only surfaces with larger length scales are modelled correctly. Variations in land cover on this scale are regularly present. At each surface transition the deposition is altered. Especially forest edges give rise to large enhancements of the deposition. This enhanced deposition can be dealt with in a very simplified way using correcting factors defined by e.g. van Pul et al. 1992b, Draaijers et al., 1994. Since in this land-use data base only percentages of the land-use type per grid are given and not the geographical position, only a statistical approach of the uncertainty can be carded out. For instance, the enhanced deposition at forest edges as a whole for all forest stands in the Netherlands, is estimated at 10-30% (van Pul et al., 1992b, Draaijers et al. 1994). The accuracy of the presented results depends on the availability and quality of the input data such as the land-use map and the meteorological observations. In the gridded version of the RIVM land-use data base, forest is not subdivided into deciduous and coniferous. All forest is classified as coniferous forest. This will probably lead to overestimates of the deposition

208 velocity to deciduous forests for all components during winter. However, the stomatal resistance in winter will be large due to low temperatures. So this overestimate will be somewhat leveled out. A version in which the forest data are subdivided in the above categories will be available in 1995. In this new version the quality of the land-use data for some areas in Eastern Europe will also be improved. The dry deposition is calculated on a daily basis. However, due to the daily averaging of the concentration and the deposition velocity, a loss in temporal correlation is introduced between the concentration and the deposition velocity. This error is component specific and is estimated to be smaller than 20% (van Pul et al.,1993). In the future the deposition will be calculated on a 6-hourly basis and the above error avoided. The uncertainty in the wet deposition estimates is relatively small compared to the uncertainty in the dry deposition estimates. A comparison between the derived wet deposition maps and EMEP long-range transport model results was carded out by means of calculation of differences and ratios between the grids derived by the two methods (van Leeuwen et al., 1994). In most parts of Europe deviations were found smaller than 200 mol/ha/year in absolute terms for individual components and smaller than 50% in relative terms. The uncertainty in the wet deposition is most important in areas where wet deposition is equal to or higher than dry deposition. In such areas, however, wet deposition usually shows a smooth pattern. This is not true for mountainous regions where additional deposition pathways such as fog and cloud deposition are present. Corrections can be applied to the wet deposition if local data on fog and cloud composition, occurrence and liquid water content are available (Fowler, 1991). The overall uncertainty in the depostion maps consists of the above-mentioned uncertainties. However, the uncertainty in the surface resistance and the occurence of sub-grid concentration gradients will act as the largest uncertainty sources. Given these uncertainty estimates, the uncertainty in the deposition of a component of a 1/6~ ~ lat/long gridcell is typically 100200%. However, the uncertainty in the total potential acidification map is smaller because the total acidification consists of components such as wet deposition which have a smaller uncertainty. 6 Conclusions In this paper a description is given of the EDACS model, with which the deposition of acidifying components on a small scale over Europe is calculated. Dry deposition velocity fields are constructed from a detailed land-use map (1/6~ ~ lat/long, made by RIVM) and meteorological information using a detailed parameterization of the dry deposition process. These small-scale dry deposition velocity fields are combined with air-concentration fields (taken from the EMEP Long range transport model) to yield dry deposition amounts on a small scale. Wet deposition is also estimated to obtain a total acidifying deposition map over Europe (van Leeuwen et al., 1994). These deposition fields clearly reflect the spatially detailed land-use information and the large-scale concentration pattern over Europe. With these fields a better match is obtained between the critical and actual loads when ecosystems are concerned. The presented deposition fields are preliminary because of several shortcomings present in the method and data bases. An update of the deposition fields and calculations for more recent years will be available in 1995. A more thorough uncertainty analysis of the deposition maps,

209 a validation with (throughfall and micro-meteorological) measurements and corrections on the wet deposition caused by cloud and fog deposition will also be carded out.

Acknowledgements A.Eliassen, E.Berge and H.Styve of EMEP MSC-W are thanked for their cooperation in providing the model concentration data. The ECMWF and KNMI are acknowledged for the WMO synops data. F.de Leeuw is acknowledged for his comments on the draft and J.Burn for editing the manuscript.

References Breemen,N.van, P.A.Burrough, E.J.Velthorst,H.F.van Dobben, T.de Wit,T.B.Ridder and H.F.R.Reinders, 1982. Nature,299,548-550. Chamberlain A.C., 1966. Proc. R. Soc. Lond. A290, 236-265. Draaijers,G.P.J.,R.van Ek, and W.Bleuten, 1994. Boundary-Layer Meteorology 69, 343-366. Erisman J.W., 1992. PhD-thesis. University of Utrecht, the Netherlands. Erisman, J.W.,W.A.J. van Pul and G.P.Wyers, 1994a. Atmospheric Environment Vol. 28, No. 16: 2595-2607. Erisman, J.W.,C.J.M.Potma, G.P.J.Draaijers, E. van Leeuwen and W.A.J. van Pul., 1994b. In Proceedings of EUROTRAC symposium'94, P.Borrell (Ed.), SPB Academic Publishing, The Hague, the Netherlands. Erisman J.W., 1994c. Atmospheric Environment Vol.28, No.16: 2583-2594. Erisman J.W. and Baldocchi D.D., 1994d. Tellus 46B, 159-171. Fowler D., 1978. Atmospheric Environment 12,369-373. Fowler D., J.H.Duyzer, D.D.Baldocchi, 1991. Proc. R. Soc. Edinburgh 97B,35-59. Grennfelt,P. and E.Th6rnel6f, 1992. Report from a workshop at L6keberg, Sweden, April 610, 1992. Report No. Nord 41, Nordic Council of Ministers, Copenhagen. Heij, G.J. and T.Schneider (Ed.), 1991. Studies in Environmental Science 46. Elsevier,Amsterdam. Hettelingh,J.P.,R.J.Downing and P.A.M.de Smet, 1991. CEC technical report no.1 RIVM report 259101001. Hicks, B.B.,1986. Water,Air and Soil Pollution 30: 75-90. Hicks B.B., D.D.Baldocchi, T.P.Meyers, R.P.Hosker Jr. and D.R.Matt, 1987. Water Air Soil Pollut. 36, 311-330. Hicks B.B., R.R.Draxler, D.L.Albritton, F.C.Fehsenfeld, J.M.Hales, T.P.Meyers, R.L.Vong, M.Dodge, S.E.Schwartz, R.L.Tanner, C.I.Davidson, S.E.Lindberg and M.L.Wesely, 1989. State of Science/Technology, Report no. 2. National Acid Precipitation Assessment Program. Jaarsveld, J.A. van and D.Onderdelinden, 1995. RIVM report in preparation. Kriiger,O., 1993. In: ProceeAings CEC/BIATEX Workshop 4-7 May 1993 Aveiro, Portugal. Ed. J.Slanina. pp 31-38. L6vblad, G.,J.W.Erisman and D.Fowler, 1993. Proceedings Nordic Council/EMEP/BIATEX workshop in G6tenborg 3-6 November 1992. Leeuwen, E.P. van, J.W.Erisman, G.P.J.Draaijers, C.J.M.Potma and W.A.J.van Pul, 1994. Report no. 722108008. RIVM, Bilthoven, the Netherlands. Pasquill,F., 1972. Quarterly Journal of the Royal Meteorological Society 98:469-494. Potma, C.J., 1993. RIVM report 722401001 Bilthoven, the Nethedands.

210 Pul W.A.J. van, J.W.Erisman, J.A. van Jaarsveld and F.A.A.M. de Leeuw, 1992a. In Acidification research: evaluation and policy application. (edited by T. Schneider), Studies in Environmental Science. Elsevier, Amsterdam. Pul W.A.J. van, R.M.van Aalst and J.W.Erisman, 1992b. In: J.Slanina,ed., EUROTRACK/BIATEX annual report,Garmisch-Partenkirchen, FRG 248-254. Pul W.A.J. van, J.W.Erisman, J.A. van Jaarsveld and F.A.A.M. de Leeuw, 1993. In: Proceedings CEC/BIATEX Workshop 4-7 May 1993 Aveiro, Portugal. Ed. J.Slanina. pp 95-115. Pul W.A.J. van and A.F.G. Jacobs, 1994. Boundary-Layer Meteorology 69: 83-99. Pul W.A.J. van, C.J.M.Potma, G.P.J.Draaijers, E. van Leeuwen and J.W.Erisman, 1994. In: Proceedings of EUROTRAC symposium'94, P.Borrell (Ed.), SPB Academic Publishing, The Hague, the Netherlands. Pul W.A.J. van, C.J.M.Potma, E.P. van Leeuwen, G.P.J.Draaijers and J.W.Erisman, 1995. RIVM report 722401005 Bilthoven, the Netherlands. Ruijgrok W. and C.I.Davidson C.I., 1993. In: Proceedings Nordic Council/EMEP/BIATEX workshop, GStenborg, Sweden, 3-6 November 1992. Ed. l.~vblad, G., J.W.Erisman and D.Fowler. Ruijgrok W., H.Tieben and P.Eisinga, 1994. Report 20159-KES/MLU, KEMA, Arnhem, the Netherlands. Sandnes H., 1993. EMEP report 1/93. MSC-West, Oslo, Norway. Slanina, J., G.Angeletti and S.Beilke, 1993. Proceedings CEC~IATEX Workshop 4-7 May 1993 Aveiro, Portugal. Thorn A.S., 1975. In: Vegetation and Atmosphere, pp. 58-109 (Ed. Monteith J.L.), Academic Press, London. UK Review Group on Acid Rain, 1990. Warren Spring Laboratory, Stevenage, UK. Velde van de, R.J., W.Faber, V.Katwijk, H.J.Scholten, T.J.M.Thewessen, M.Verspuy and M.Zevenbergen, 1994. report 712401001, RIVM, Bilthoven, the Netherlands. Voldner E.C., L.A.Barrie and A.Sirois, 1986. Atmospheric Environment 20,2101- 2123. Walcek, C.J., R.A.Brost, J.S. Chang and M.Wesely, 1986. Atmospheric Environment 20, 949964. Wesely M.L., D.R.Cook and R.L.Hart, 1985. J. geophys. Res. 90,2131-2143. Wesely M.L., 1989. Atmospheric Environment 23,1293-1304.

E F F E C T S OF A C I D D E P O S I T I O N O N F O R E S T E C O S Y S T E M S I N T H E NETHERLANDS SESSION V



213

A S S E S S M E N T AND E V A L U A T I O N OF C R I T I C A L LEVELS FOR O3 AND NH3

Eveliene Steingrfver', Tom Dueck b & Ludger van der Eerden b

9 IBN-DLO Institute for Forestry and Nature Research, P.O. Box 23, 6700 AA Wageningen, The Netherlands b AB-DLO Research Institute for Agrobiology and Soil Fertility, P.O. Box 14, 6700 AA Wageningen, The Netherlands

Abstract The effect of 03 on three different tree species was similar. A threshold of 40 ppb was found for Pinus and for Pseudotsuga. Over two growing seasons growth was inhibited in Pinus and Fagus saplings and total assimilation per tree was inhibited over one growing season in Pseudotsuga trees. Pinus was most sensitive to 03, but a critical level of 10 ppm.hrs is sufficient to protect the three species from O3 damage. The effect of NH3 on growth was ambiguous. Growth was unchanged in Fagus, stimulated in Pseudotsuga and inhibited in Pinus. The effect of NH3 on tree architecture and stress sensitivity was similar. Tree architecture was changed in both Fagus and Pseudotsuga and drought and frost sensitivity were increased in Pinus and Pseudotsuga. At the moment both the critical level for 03 and for N H~ are exceeded in The Netherlands in general and in selected areas in particular.

1. INTRODUCTION

The current estimation of the mean nitrogen load in the Netherlands is 35 kg ha -I y r I of which the largest proportion (16 kg ha -~ yr -~) is gaseous ammonia. The regional mean NH3 concentrations, however, may be very different from the average national concentration, resulting in a much higher exposure level for forest trees in areas with intensive live stock farming [ 1]. Ozone, another relevant air pollutant, is more evenly distributed over the Netherlands (Tab. l). The national mean concentration exceeds the phytotoxic level for crops [2], and the economic impact on Dutch crops is large 13]. The effect of 03 on mature forest trees, on forest ecosystems and on natural ecosystems in general is largely unknown. The available information is mainly acquired from experiments with seedlings and saplings 141. Two types of air quality standards are currently in use: critical levels and critical loads. Critical levels are based on exposure values, while critical loads are based on deposition values. The use of two different air quality standards is .justified, because critical levels focus on individual air pollution components and on nondepositing oxidants such as 03, whereas critical loads lump either nitrogen or

214 acidifying components together. Moreover critical levels are usually used for shortterm exposures (hours - months) and critical loads for long-term exposures (years). The aim of the present investigation was to assess the impact of NH3 and 03 on forest ecosystems at concentrations presently occurring in the Netherlands in relation to the critical level. This was done by comparing the effects on tree saplings exposed to both pollutants in OTC's (Open Top Chambers) with the effects on mature forest trees under ambient field conditions.

2. MATERIAL AND METHODS

2.1 Fumigation experiments in OTC's Two experiments were performed in OTC's, in which 3-yr-old Beech (Fagus sylvatica L.) and Scots pine (I~'nus sylvestris L.) saplings were fumigated for 15 months, largely covering two growing seasons (from June l to September 1). The OTC's have been described earlier [5]. Ammonia and ozone were injected into the air stream prior to the blower via thermal mass-flow controllers (Brooks 5850 TR). Air pollutant concentrations were sequentially monitored with an ozone analyzer (8810, Monitor Labs) and a NH3 monitor (thermoconverter model 8750 followed by a chemiluminescent NOx analyzer model 8840, both Monitor Labs) and were computercontrolled. The data were recorded with a HP data acquisition system. A duplicated range of 03 concentrations (0, 30, 60, 90, 120 and 150 #g m -3) was used in the first experiment and these concentrations were both higher and lower than current ambient concentrations in the Netherlands (Tab. l).

Table 1 Average concentrations of N H 3 and 03 (#g m 3) in 1992.

The Netherlands Wageningen, OTC's Veluwe, field

NH3

03

3.4 16 2.5

60 72 78

Trees were exposed to 03 during a 9 h day and to a third of the daytime concentrations during the remaining 15 h. To one of the 03 ranges, a concentration of 40/~g m 3 NH3 was added (24 h dayS), which is somewhat higher than the highest mean concentrations experienced in the Netherlands. In the control treatments, the NH3 concentration was also higher (15 #g m -3) than the national mean due to the tact that Wageningen is located in a region with high NH3 concentrations (Tab. 1), and the filters used to clean the air have less that 50% capacity for NH3. In the second experiment, 03 and NH3 were applied in factorial design, resulting in a triplicated set of 03 concentrations (setpoints 0, 90, and 135/~g m3), supplemented with ambient air, or with NH3 to 40 ~g m -3 and 80 /~g m 3 NH3. The water potential on 1-year-old

215 needles of P~'nus sylvestris was measured with a pressure bomb 161. The measurements were performed in six OTC's only, in all three NH3 treatments combined with filtered air and the highest 03 concentration and were performed between 08.00 and 13.00 hours in fully watered pots and again after five days without water. The soil water potential was measured daily to ensure that the drought treatment was not prolonged to the point where excessive drought injury to the trees occurred in order to relate the soil water potential to the needle water status. This paper discusses the main results on biomass production and drought sensitivity.

2.2 Correlation studies in the field All field measurements were performed in a stand of 34-year-old Douglas fir (Pseudotsuga menziesii (Mirb.) Franco L., provenance Arlington), located at the Veluwe, in the central part of the Netherlands. The Speuld site has a stand density of approximately 800. In 1993, the average tree height was 22.2 m and the average DBH was 25.4 cm 17]. The mean 03 and NH3 concentration at Speuld are shown in Tab. 1. A computer controlled field gas exchange system was installed at the site. With this system up to 16 different branch assimilation chambers can be measured continuously. A 22 m high scaffolding was built in the middle of the stand, from which the branch assimilation chambers could be mounted in 8 different trees. The chambers were ventilated with ambient air and contained one year class of needles. Temperature, relative humidity and CO2 concentration of the air entering the chambers closely resembled that of ambient outside air. A PAR sensor was mounted at the outside of each chamber. Each chamber was sampled twice every hour during 24 hours per day from March 1992 up to December 1993. The light response curve was modelled using the equation of Goudriaan I81 with the measured photosynthetic rate and photoactive radiation levels. The unexplained variance in the data could be reduced by 40% by taking into account changes in vapour pressure deficit (VPD) and the 03 concentration. In this way, the direct effects of 03, NH3 and NOx and VPD on the photosynthetic rate were statistically estimated. This paper discusses the results on CO2 assimilation, per month and per year, and on the needle nutrient status.

2.3

Exposure-response relationship for 03

The levels of exposure to 03 used here are not expressed as concentrations, but as the accumulated exposure over a threshold concentration, abbreviated with AOT [41. The AOT adds all exposures above a certain threshold concentration over the period of interest, i.e. growing season, but also over several years in which mean concentrations can strongly fluctuate over the seasons. The threshold value is the concentration above which 03 toxicity becomes evident. The AOT approach proved to be very useful for evaluation of crop loss in which relative yield reductions could be linearly fitted to AOT values. Recently, this approach has also been applied to forest trees, although reliable field data are scarce [4].

216

3. RESULTS 3.1. Fumigation experiments in OTC's Fig. 1 shows the effects of O3 and NH3 on the total biomass of P. sylvestris. The biomass of trees exposed to the lowest 3 levels of 03 was significantly higher than that in the 3 highest levels. An effect threshold for 03 is visible between 5 and 15 ppm-hrs. The 3 lowest 03 concentrations below 40 ppb were logically similar in terms of AOT40, and this was also reflected in the biomass production in those treatments, indicating the validity of the threshold for P. sylvestris. The absence of a linear exposure-response relationship is likely due to the fact that high concentrations of 03 inhibit stomatal conductance 19] and thus the uptake of 03. The UNECE and the WHO recently recommended an AOT40 of 10 ppm-hrs as a critical level for trees, at which a 10% reduction in biomass production should not be exceeded. The results shown for P. sylvestris in Fig. 1 appear to support this critical level. Current concentrations of O3 in the Netherlands however, exceed this critical level by more than a factor 2, indicating that O3 is significantly reducing the growth of P. sylvestris. Fig. 1 also indicates that at low 03 concentration, NH3 tends to inhibit growth, but growth is not further reduced at higher 03 concentrations. 1000

A OI

9OO

-NHa

0

0

Jr ..,.. O

3=

aE"

9

+NH 3

800

700

. 0

.

. 20

.

. 40

.

. 60

.

. 80

100

AOT40 (ppm-h)

Figure 1. Effect of 03 and NH3 on the total biomass production Pinus ,~ylvestris. o O3 alone; 9 - O3 + NH3. In the same fumigation experiment, F. sylvatica was found to be less sensitive to 03; a 10% reduction in biomass production was found at an AOT40 of 30 ppm-hrs, while NH3 had no effect on the total biomass. The growth of F. ,~ylvatica was differentially affected by NH3 and 03 in tree architecture rather than in biomass production (Fig. 2). Although 03 inhibited tree height, it increased stem diameter, which resulted in relatively sturdier trees. NH3 did not influence tree height and reduced stem diameter, resulting in relatively smaller trees at higher levels of 03. Lateral branch growth was also reduced by NH3 with increasing concentrations of 03, thus reducing the potential for light interception. In the second fumigation experiment, special attention was paid to the effect of NH3 and 03 on the drought sensitivity of P. sylvestris. The data in Fig. 3 show that needles of fully watered P. ,~ylvestris saplings have a significantly higher (more negative) water potential when exposed to NH3 alone than when exposed to NH3 +03.

217

100

Length increment (cm)

lO

Diameter incr.ement (cm) o

gO

9

80 70 60

o l

0

,

i

20

,

l

,

40

l

,

60

80

,,

.

~

9

9

0

1 O0

20

AOT40 (ppm-h)

160

"

40

60

80

..i

100

AOT40 (ppm-h)

Branch dry wt (g)

150 140

130

o

o

-NH3

120 9 +NH 3

110

100 90 80

"

0

20

40

60

80

'

100

AOT40 (ppm-h)

Figure 2. Effect of O3 and NH3 on mean length increment, stem diameter increment and branch dry weight of Fagus sylvatica, o = O3 alone 9 = O3 + NH,.

No difference was observed in the needle water potentials between the three treatments of only NHs. It appears that, irrespective ~)I the NH, ct)ncentrati~)n. O, inhibits stomatal conductance and thus transpiration in the absence c)f water stress. allowing the trees to maintain turgor at a lower (200-400 k Pa) needle water potential than when exposed to NH3 alone. When the trees were droughted for 5 days, the water potential increased linearly with increasing concentrations of NHs in the absence of 03. This indicates that under conditions of drought stress, N H~ disrupts the stomatal control, resulting in increased transpiration and reduced water use efficiency I! 0-i 11. When O, was added to the NH3 treatments, the effect of increasing N H, c()ncentrati~)ns was masked, and the water potential in all three Os + NH~ treatments remained lower than in treatments with NHs alone. This confirms earlier experiments, in which the reduced transpiration of various species exposed to 03 is attributed to inhibition of stomatal al~rture [91. Thus, even at the concentrations of N H3 used in this experiment, O3 appears able to reduce water loss through transpiration and is thus able to reduce drought stress.

218 1400 w a t e r e d [ - - ' ] drought A a

1160

Jr

"~ ...

920

C 0 0

=.

680

=_

=

~:

440

i

200

FA

40N

80N

03

40N+03 80N+03

Figure 3. Effect of O3 and NH3 on drought stress of Pinus sylvestris. Dark bars indicate water potential (means + SE) of watered trees and white bars that of droughted trees (n= 10).

3.2. Correlation studies in the field The vitality of the Speuld stand was characterized in 1986 according to international EC standards of defoliation and foliar discoloration. The vitality was higher than the nationwide average for P. menziesii, which was classified as less healthy for 50% of all trees 1121. The vitality was re-assessed in 1994, and was similar to the nationwide average, indicating that the vitality of the stand had decreased to the level of most other P. menziesii stands in The Netherlands, which was classified as less healthy for 80% of all trees 171. Notwithstanding the poor vitality, biomass production was high. According to yield tables for Dutch P. menziesii stands I131, based on data from the pre-intensive livestock age, annual wood volume increment was higher than expected. Speuld can be qualified as an average soil quality site (class 111) with an expected mean annual volume increment of around 13.4 m 3 ha 1. The average annual increment over the period 1987-1993 however, was 24.7 m 3 ha -1, which is even higher than the expected volume increment of 17.3 m 3 ha 1 on the best quality site class.

Table 2. Biomass partitioning in Pseudotsuga menziesii using foliage I71 and root !141 data from Speuld and from i15-161. Speuld

low productive, slow grow rate

high productive, fast grow rate

5.5

0.95-1.1

2.7-3.6

0.67

0.21-0.34

0.19-0.31

Foliage/ fine root

Foliage/ coarse root

219 The dry weight distribution in Speuld in which the N deposition is lower than the Dutch mean, was compared to that of other P. menziesii stands. Speuld can be best compared to high productive, fast growing stands of the same age. The foliage/root ratio in Speuld is exceptionally high compared to the other stands (Tab. 2). The high foliage/root ratio results from both lower root biomass and higher needle biomass. The high amount of needle biomass is also reflected in the LAi of the stand which was 10.7 in 1992. The foliage/coarse root ratio is also high (Tab. 2). The branch dry weight however, was lower compared to the other stands. The nutrient status of the needles changed significantly during the experimental period [8]. The average N concentration increased from 1.7 % in 1987 to 2 % in 1993. The optimal concentration for biomass production in P. menziesii is 1.8%. The K concentration decreased from 0.7% to 0.5 %, which is below the deficiency level of 0.6 %. The P concentration remained constant, but was also below the deficiency level of 0.14%. The N/P ratio was constantly below/above the deficiency level, while the N/K ratio increased from 2.7 to 3.7. Extrapolating the nutrient trends found over the past 7 years, the ratio of N/K can be expected to reach the deficiency level within 1-2 years. High N/K ratio's, as found in the Speuld stand, are often considered as an indicator for increased stress sensitivity, e.g. frost sensitivity !171. There seems to be little doubt that the high nitrogen status, the high productivity, the high foliage/root ratio, the high LAI and the increased stress sensitivity are caused by the high N deposition into the stand.

Table 3. Proportional reduction of the monthly CO2 assimilation due to 03 and VPD (mean over 2 needle age classes and three crown levels). 03

VPD

03

1992 April

3.3

May

VPD 1993

2.4

28.0

14.0

12.2

22.9

13.2

8.1

June

4.4

20.1

4.4

6.1

July

5.3

6.3

7.6

2.9

August

4.1

7.0

5.4

NS

Net CO2 assimilation was measured more or less constantly throughout 1992 and 1993 and was related to changes in meterological parameters and air pollution concentration. No direct effects of NH3 and/or NOx on net CO2 assimilation of P. menziesii were found. NH3 tended to stimulate assimilation, but the differences were not significant. The only obvious relationship of CO2 assimilation with air pollution was that with 03 (Tab. 3). The reduction of net CO2 assimilation by 03 on a monthly basis can be quite considerable, and was in the same range as the reduction by VPD.

220 The annual reduction of total CO2 assimilation per tree was estimated to be in the range of 3-10% in 1992. The annual reduction over 1993 will probably be higher, as the monthly reductions are higher. In order to assess the critical level for 03 for adult P. menziesii trees under ambient conditions, the reduction in biomass production by 03 has to be known. We assumed that the annual reduction in total CO2 assimilation is in the same range as the annual reduction in biomass production. Matyssek et al. however, (pers. commun.) found that a 40% reduction in total CO~ assimilation resulted in a 60% reduction in biomass.

4. EVALUATION A N D CONCLUSIONS

Despite the fact that different tree species were used, the reaction of saplings and adult trees under field conditions to high NH3 concentrations showed both similarities and discrepancies. Most of the work done on effects of NH3 on plants to date indicate that NH3 stimulates growth, but the results presented in this paper show that NH3 may also inhibit growth in the presence of low concentrations of 03 in Pinus. In Fagus however, NH3 had no effect on growth, but changed tree architecture by decreasing tree height and branch biomass. Under field conditions the high N input resulted in both increased growth and in a changed architecture. Dry weight distribution in Pseudotsuga was affected, as foliar biomass was higher and fine root and branch biomass was lower. Another similarity was the increased sensitivity. N H 3 was found to increase drought sensitivity in Pinus saplings, and both drought and frost sensitivity in adult Pseudotsuga trees under ambient conditions. NH~ is the maior part of the total N deposition in the Netherlands. The aim to reduce emissions until critical loads are not exceeded, will surely reduce NH3 levels below the critical level for NH3. The critical level proposed for NH3 is 8 and 270/,g m 3 respectively, for an annual and a 24 hrs mean !181. The annual mean is currently being exceeded in half of the Dutch areal I1|. On sites removed from the direct influence of point sources, the 24 h critical level is probably not being exceeded, not even in the Netherlands. Exceedances are frequent only on a local scale, in the first 300 meters from a point source. If environmental policy will be based on damage estimates in relation to emission and dispersion, knowledge of effect thresholds of N H3 is insufficient. Although NOx is only a small part of the total N deposition, its abatement would be very profitable because it would reduce the 03 concentration as well. High 03 concentrations resulted in decreased growth in all tree species. However, a 10% reduction in biomass production was reached at different AOT40 values (Fig. 4). It seems that the critical level for forest trees of 10 ppm.hrs is able to protect our trees from damage by 03. From the range of A OT40 values between '87-'93 it is evident that these values are exceeded in general and in the forested area of the Veluwe in particular. Timber producers may not bother about a 10% reduction in biomass by 03 if the fertilizing effect of atmospheric N deposition compensates for the losses by 03. However, this does not hold for the Dutch situation. First of all growth stimulation by N is of a temporary nature. An initial growth stimulation by N is accompanied by increasing ratio's of N to cations and following luxurious N consumption will eventually lead to a decreased growth due to severe cation deficiencies [19|. The progressing nutrient status of Speuld suggests that it will not

221 take decades to reach this situation. Furthermore, a number of forested areas in The Netherlands appears to have reached the stage of growth reduction already 1191. A O T 4 0 for 10% r e d u c t i o n of Speuld

biomass production

75

50 9-"

Douglas

9 9-,~

Beech

25

l

NL '87-'93

Scots pine

9

10

Critical Level for forest trees

0

,

I

I

I

I

,

Figure 4. AOT40 values for forest trees. The columns on the left indicate the range of 03 concentrations measured during 1987-1993 in Speuld (25m height) and as a national average (3m). The arrows indicate the AOT40 values for Pl'nus ,sylvestris, Fagus sylvatica saplings and for mature Pseudotsuga menziesii, presented in this paper in relation to the critical level proposed for forest trees.

Secondly, forests are more than trees alone. The impact of air pollution on trees (and crops) should not be seen as indicative of the unknown effects on other parts of the forest ecosystem. Effects on herbs and grasses in the undergrowth and on biodiversity of plants and animals are likely to be more pronounced 1201. The results presented in this paper suggest that all three tree species used in the experiments are adversely affected by current 03 concentrations in The Netherlands. The possibilities to quantify this damage has increased strongly in recent years, but much has yet to be done.

5. REFERENCES G.J. Heij and T. Schneider (eds.), Acidification research in the Netherlands, Elsevier, Amsterdam, 1991. T. Schneider, S.D. Lee, G.J.R. Wolters and L.D. Grant (eds.), Atmospheric Ozone Research and its Policy Implications, Elsevier, Amsterdam, (1989).

222

9 10

11 12 13 14 15 16 17 18 19 20

L.J. Van der Eerden, A.E.G. Tonneijck and J.H.M. Wijnands, Environmental Pollution, 53 (1988) 365. J. Fuhrer and B. Achermann (eds.), Critical Levels for Ozone, FAC Report no. 16, Liebefeld-Bern, 1994. Th. A. Dueck, Functional Ecology 4 (1990) 109. P.F. Scholander, H.T. Hammel, E.D. Bradstreet and E.A. Hemmingsen, Science 48 (1965) 339. E.G. Steingr/Sver and W.W.P. Jans, Dutch Priority Programme on Acidification, Report No. 793315-03, Bilthoven, 1994. F.W.T. Penning de Vries and H.H. van Laar (eds.), Simulation of plant growth and crop production, PUDOC, Wageningen, 1992 M. Pearson and T.A. Mansfield, New Phytologist 123 (1993) 351. M. Tesche and S. Feiler (eds.), Air Pollution and Interactions between Organisms in Forest Ecosystems, Proc. IUFRO-Centennial Congress, Tharandt/Dresden, 1992. L.J. Van der Eerden and M.G.F.J. P~rez-Soba, Trees-Magazine 6 (1992) 48. T.F.C. Smits and G. van Tol (eds.), IKC-NBLF Report, 1987. J.G.A. LaBastide and P.J. Faber, Stichting Bosbouwproefstation "De Dorschkamp" 11(1), 1972. A.F.M. Olsthoorn and A. Tiktak, Neth. J. Agric. Sci. 39 (1991) 61. M.A. Espinosa Bancalari and D.A. Perry, Can. J. For. Res. 17 (1987) 722. M.R. Keyes and C.C. Grier, Can. J. For. Res. 11 (1981) 599. O. Skre, Commun. of the Norwegian Forest Research Station, 40(9), 1988. L.J. Van der Eerden, Th. A. Dueck, A.C. Posthumus and A.E.G. Tonneijck, UNECE Workshop on Critical Levels, Egham, UK, 1992. J.G.M. Roelofs et al., this proceedings, 1995. R. Bobbink et al., this proceedings, 1995.

G.J. Heij and J.W. Erisman (Editors). Acid Rain Research: Do we have enough answers? © 1995 Elsevier Science BV. All rights reserved.

223

EXPERIMENTAL MANIPULATIONS: FOREST ECOSYSTEM RESPONSES TO CHANGES IN WATER, NUTRIENTS AND ATMOSPHERIC LOADS

Andries W. Boxman 1, Pieter H.B. de Visser2 and Jan G.M. Roelofs ~ 1 Department of Ecology, Research Group of Environmental Biology, University of Nijmegen, P.O. Box 9010, 6500 GL Nijmegen, The Netherlands 2 Department of Soil Science and Geology, Agricultural University of Wageningen, P.O. Box 37, 6700 AA Wageningen, The Netherlands

ABSTRACT

In four Dutch coniferous forest ecosystems water and nutrient supply, as well as atmospheric loads, were manipulated for three or more years. Four approaches were used: (1) optimal supply of water and nutrients, (2) decrease of nitrogen and sulphur loads to pre-industrial levels, (3) increase of nitrogen and sulphur loads to excess levels (120 kg N ha -1 yrl). Nutrient supply was optimized according to tree demand in optimal proportions relative to ambient nitrogen supply. Tree growth was strongly enhanced (ca. 30%) by optimal water supply but not further enhanced by nutrient additions. Water additions tended to lower nitrogen concentrations in the needles by 5-10%, probably by growth dilution. Nutrient applications improved the nutritional balance in trees of phosphorus and potassium relative to nitrogen. Exchange of applied base cations with protons and aluminium from the soil, temporarily increased the acidity of the soil solution. Large applications oi nitrogen in a Scots pine stand increased the nitrogen concentration in the needles. Excess nitrogen stimulated tree growth during the first two years, and depressed growth in the fifth treatment year. Phosphorus deficiency was induced but no visible tree damage occurred. When atmospheric deposition of nitrogen and sulphur was reduced to preindustrial levels in a nitrogen and sulphur saturated Scots pine and Douglas fir stand, a few months after reduction of the input, output to the groundwater was also strongly reduced. This implies a tight input-output coupling. As a result, leaching of aluminium and base cations (to counteract nitrate and sulphate leaching) decreased and the mineral balance in the soil solution improved. Consequently, tree health improved as shown by increased root and shoot growth and by reduced totaI-N and arginine-N concentrations as well as an improved mineral balance of the needles

224

INTRODUCTION

High atmospheric input of ammonium results in loss of base cations from the soil. due to either direct exchange at the soil absorption complex or indirect via acidification, as a result of nitrification. This may result in nutrient deficiencies in trees (Roelofs eta/., 1985; Schulze, 1989). The nutritional balance of trees can be affected even further by preferential ammonium uptake to supraoptimal levels and can be disturbed by the ion competition of ammonium, aluminum and protons on the potassium, magnesium and calcium uptake. Furthermore acid soils can have Ca/AI ratios that are unfavourable for roots. The high inputs of nitrogen can result in lower root/shoot ratios and may increase the drought sensitivity of trees. This study deals with the growth and nutrition of two tree species, Scots pine

(Pinus sy/vestris L.) and Douglas fir (Pseudotsuga menziesii (Mirb.) Franco), under influence of acidification and eutrophication. The aim was to determine whether

the

nutrient supply is growth-limiting,

under the constraint of

aluminum stress and competition with ammonium and how nutrition and root functioning are related to the water economy of the tree. Therefore, the effects of a change in input of water, nutrients and acidifying substances to the soil, on trees were quantified. Atmospheric loads were increased or decreased to study their impact on nutrient cycling and forest decline. Secondly, it was hypothesed that the effects of increased soil acidification and eutrophication on trees could be mitigated by optimization of the water and nutrient supply. Experimental variation of the water supply may elucidate the impact of water stress on tree functioning. Reducing the nitrogen input to pre-industrial levels may assess the reversibility of nitrogen saturation on the different compartments of the ecosystem. These ecosystem manipulation experiments were conducted within the EXMAN and NITREX framework (Beier and Rasmussen, 1992; Dise and Wright, 1992). MATERIALS AND METHODS

Manipulations were carried out at four forest stands in the Netherlands. In a Scots pine stand near Harclerwijk and in a Douglas fir stand near Kootwijk either water and nutrients were applied, or rates of soil acidification

were

changed by exclusion of atmospheric loads or by increasing acid loads. Irrigation of demineralized water (I plot) amounted 3 to 4 mm day 1 maximally on days without rain. Fertigation

(IF plot) consisted of a complete set of

dissolved nutrients, given very frequently and in addition to irrigation during four growing seasons. The total annual application rate was equal to the

225

estimated gross uptake in trees. Phosphorus and potassium additions were quantitatively the most important and ranged from 13 (P) and 65 (K) for Scots pine to 36 and 60 kg ha 1 yr4 for Douglas fir respectively. A roof construction above the forest floor prevented the infiltration of throughfall water, being polluted with atmospheric substances, and clean rain was irrigated below in combination with a fertigation treatment (IF+R plot). Fertilization with dissolved (NH4)2SO4 amounted to 120 kg N ha 4 yr 4 (treatment N+S). In the Harderwijk stand the treatments C (control), I, IF and N+S, in the Kootwijk stand C, I, IF and IF+R were carried out. In all treatments soil water contents and composition, tree growth, needle chemistry and needle fall were monitored. (see De Visser (1994) for details on treatments and measurements). In 1989 two research sites were established in a Scots pine stand near Ysselsteyn and in a Douglas fir stand near Speuld in which ambient throughfall water was intercepted by means of a roof, and replaced by demineralized water to which all nutrients were added in the same amount as present in the throughfall, except for nitrogen and sulphur concentrations. Underneath the roof, two plots (10xl0m) were designed to receive either clean water (roof-clean plot) or ambient throughfall (roof-control plot). Outside the roof a second control plot was established, receiving ambient throughfall (control plot). A detailed description of the sites and of the methods has been given elsewhere (Dise and Wright, 1992; Van Dijk et aL, 1992a,b; Boxman et aL, 1994; Boxman et aL,1995). For statistical analyses the software packages Systat 5.0 and Statgraphics 6.0 were used. RESULTS AND DISCUSSION

Ecosystem responses to changes in water and nutrients Tree growth and nutrition in relation to water and nutrient supply Three out of four irrigated forest stands showed a water-limited growth in the examined period. An increase of 40% for Douglas fir to 50% for Scots pine in basal area growth was observed upon irrigation of 3 to 4 mm day 1 (Figure 1; p_ 50) in plot IF of Douglas fir. Figure 2 Fertigation did not increase total Douglas fir growth over the four-year period in addition to the growth effect of irrigation alone, and in Scots pine in one out of four treatment years only (Figure 1: 1992; p_3335 molc.ha .yr is again in n o r t h - w e s t part

]as Deposition

of SOe - 1991 Units : rnOJc.~ I.yr •

EGEND

"

> lOOO SO00

-

3000-

q[--"~

. . . . . . ~ ~ - ~ ~ "--.~EE~L-.___.

Fig.

i. Gas d e p o s i t i o n

nil ...... :

~':"...

~ ~ ~ m

of SO 2 in the Czech R e p u b l i c

in 1991 (molc.ha-l.yr-l).

m

ZOO()

.5000

2&O0-

3000

20~-

2500

ISO() -

2000

lOOO-

ISO0

~OO-

IOO0

O-

SO0

~

462

of Bohemia and in certain localities of middle and east 5% share of lowest values (

Acid Rain Research: Do We Have Enough Answers? (Studies in Environmental Science)

Acid Rain Research Do We Have Enough Answers

Climate Change Research (Studies in Environmental Science)

Are we in a war Do we have an enemy

Environmental Contamination (Studies in Environmental Science)

Environmental Radioanalysis (Studies in Environmental Science)

Acid Rain - Deposition to Recovery

Acid Rain: Report number 14

Til We Have Faces

Do Tarantulas Have Teeth: Questions and Answers About Poisonous Creatures

Do Organizations Have Feelings?

Algal Biofouling (Studies in Environmental Science)

4a - Things We Do

What We Can Do

We Could Do Worse

Why Do We Pray?

We Could Do Worse

We Have Always Lived in the Castle

We Will Do It

Should We Have Faith in Central Banks

We Have Always Lived in the Castle

We Have Fed Our Sea

We Have Fed Our Sea

We Have Never Been Modern

We have never been modern

Development and Democracy: What Have We Learned and How? (Routledge Ecpr Studies in European Political Science)

We Have Never Been Modern

Houston, We Have A Problem

We Have Fed Our Sea

We Have Never Been Modern

Houston, We Have A Problem

Acid Rain Research: Do We Have Enough Answers? (Studies in Environmental Science)