Perspectives for Agronomy (Developments in Crop Science, Volume 25)

PERSPECTIVES FOR AGRONOMY Adopting Ecological Principles and Managing Resource Use Developments in Crop Science Volum...

Author: M.K. van Ittersum | S.C. van de Geijn

34 downloads 1065 Views 24MB 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

PERSPECTIVES FOR AGRONOMY Adopting Ecological Principles and Managing Resource Use

Developments in Crop Science Volume 1 Oil Palm Research, edited by R.H.V. Corley, J.J. Hardon and B.J. Wood Volume 2 Application of Mutation Breeding Methods in the Improvement of Vegetatively Propagated Crops, by C. Broertjes and A.M. van Harten Volume 3 Wheat Studies, by H. Kihara Volume 4 The Biology and Control of Weeds in Sugarcane, by S.Y. Peng Volume 5 Plant Tissue Culture" Theory and Practice, by S.S. Bhojwani and M.K. Razdan Volume 6 Trace Elements in Plants, by M.Ya. Shkolnik Volume 7 Biology of Rice, edited by S. Tsunoda and N. Takahashi Volume 8 Processes and Control of Plant Senescence, Y.Y. Leshem, A.H. Halevy and C. Frenkel Volume 9 Taigu Genic Male-Sterile Wheat, edited by Deng Jingyang Volume 10 Cultivating Edible Fungi, edited by P.J. Wuest, D.J. Royse and R.B. Beelman Volume 11 Sugar Improvement through Breeding, edited by D.J. Heinz Volume 12 Applied Mutation Breeding for Vegetatively Propagated Crops, by C. Broertjes and A.M. van Harten Volume 13 Yield Formation in the Main Field Crops, by J. Petr, V. Cern~, and L. Hrugka Volume 14 Origin of Cultivated Rice, by H. Oka Volume 15 Nutritional Disorders of Cultivated Plants, edited by W. Bergmann Volume 16 Hop Production, edited by V. Ryb~t~ek Volume 17 Principles and Methods of Plant Breeding, by S. Borojevi~ Volume 18 Experimental Morphogenesis and Integration of Plants, by J. Seb/mek, Z. Sladl~ and S. Proch~tzka Volume 19 Plant Tissue Culture: Applications and Limitations, by S.S. Bhojwani Volume 20 Weather and Yield, edited by J. Petr Volume 21 Plant Physiology, edited by J. Sebfinek Volume 22 Reproductive Adaption of Rice to Environmental Stress, by Y. Takeoka, A.A. Mamum, T. Wada and P.B. Kaufman Volume 23 Natural Rubber: Biology, Cultivation and Technology, edited by M.R. Sethuraj and N.M. Mathew Volume 24 Irrigated Forage Production, by A. Dovrat Volume 25 Perspectives for Agronomy, edited by M.K. van Ittersum and S.C. van de Geijn v

Developments in Crop Science 25

PERSPECTIVES FOR AGRONOMY Adopting Ecological Principles and Managing Resource Use Proceedings of the 4th Congress of the European Society for Agronomy, Veldhoven and Wageningen, The Netherlands, 7-1 1 July 1996 Edited by

M.K. VAN ITTERSUM Wageningen Agricultural University, Department of Theoretical Production Ecology, P.O.Box 430, 6700 AK Wageningen, The Netherlands

S.C. VAN DE GEIJN Research Institute for Agrobiology and Soil Fertility (AB-DLO), P.O. Box 14, 6700 AA Wageningen, The Netherlands

ELSEVIER

Amsterdam - Lausanne - New York - Oxford - Shannon - Singapore - Tokyo 1997

ELSEVIER SCIENCE B.V. Sara Burgerhartstraat 25 P.O. Box 21 l, 1000 AE Amsterdam, The Netherlands

Partly reprintedfrom the European Journal of Agronomy, Vol. 7/1-3

ISBN 0 444 82852 4 © 1997, 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 or 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 AM Amsterdam, The Netherlands. Special regulations for readers in the U.S.A.-This publication has been registered with the Copyright Clearance Center Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923. 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 other copyright questions, including photocopying outside of the U.S.A., should be referred to the copyright owner, Elsevier Science B.V., unless otherwise specified. No responsibility is assumed by the publisher for any injury and/or damage to 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. This book is printed on acid-free paper Transferred to Digital Printing 2006

Preface Since the Second World War, agricultural development has been characterised by a strong increase in land and labour productivity in large parts of Europe (Porceddu and Rabbinge, 1997), strongly stimulated by policy (e.g., Common Agricultural Policy of the EU). The level of self-sufficiency has been surpassed for nearly all agricultural products. The strong increase in productivity leading to the present situation characterised by over-production, has been attended by changes in farm structure, a decrease in agricultural employment, social, budgetary and economic problems at the regional and European Union level and by environmental deterioration. This situation contrasts with a situation of relatively low productivity, socio-economic and environmental problems in mainly Southern and Eastern parts of Europe, and with a major problem and at the same time a challenge for the next century at a global level: food production and security for a doubled population with a more affluent diet. Agricultural development has evolved from an activity with mainly one-dimensional, productivity aims, into a multi-dimensional issue with environmental, agricultural, economic and social objectives. Agriculture should adopt ecological principles and it should optimize the use of resources, i.e., agriculture should 'ecologize', and at the same time it should meet social and economic objectives. Agricultural development is an important issue within the framework of what is called sustainable development. The notion of sustainable development calls for explicit consideration of each of the mentioned objectives and for consideration of the problem at several aggregation levels. The situation in (a part of) Europe cannot be isolated from the situation at country or global scale; problems and solutions at field level should also consider, for instance, plant and crop rotation level.

Agriculture with such broadened objectives requires new systems at a range of aggregation levels. It requires different analyses, synthesis of knowledge and a different type of intervention at the policy level. Agriculture with broadened objectives requires a different agronomy. It calls for detailed knowledge concerning the functioning and production of agricultural plants and crops and their ecological relationships. In addition, it calls for synthesis and design of new ideotypes and genotypes, new production technologies, cropping systems, farming systems and agroecological land use systems. Basic knowledge at field, plant and lower levels of integration should be used and synthesized in the design of new systems at higher levels that meet a set of explicit objectives. These new systems should then be evaluated for their effectiveness at the various levels. This type of agronomic research will often be of an interdisciplinary nature with agronomists working together with breeders, physiologists, soil scientists, economists and sociologists. To fulfil this new role, agronomy has a range of sophisticated tools at its disposal. To fully exploit the potential of these tools, they should not be used separately, but in combination. A new agronomy should tailor the tools to the type of questions and benefit from the synergism of: empirical and experimental research, be it in laboratories, climate chambers, greenhouses or in the field to diagnose and analyse problems and to test new designs; mathematical modelling techniques to summarize knowledge, to test hypotheses and to identify knowledge gaps; - prototyping on experimental and commercial farms to design and implement new crop rotations and farming systems; and - model-based explorations to improve systems -

-

vi

understanding and identify a wide range of options. During the Fourth Congress of the European Society for Agronomy, held in Veldhoven-Wageningen, The Netherlands, 7-11 July 1996, the new perspective for agronomy emerged. Various keynote addresses, session themes, and oral and poster contributions demonstrated the need for a new role of agronomy and its tools (Van Ittersum et al., 1996). The special issue of the European Journal of Agronomy and the Proceedings Book of the Fourth ESACongress (Van Ittersum and van de Geijn, 1997) present a set of case studies illustrating the various agronomic tools that can be used for specific questions. The case studies are grouped in sections illustrating relevant subquestions in developing an agriculture with broadened objectives. The papers were selected such that the various subquestions were represented in the Proceedings. This implies a non-random sample from the contributions during the Congress, since the number of contributions addressing the level of cropping system, farm and agricultural land use was limited. Nevertheless, we think that agronomy should consider these levels of scale in its analysis and design because questions of stakeholders often concern these levels. After an introductory paper on the role of agronomy in research and education in Europe, the second section presents case studies addressing issues concerning agricultural land use, food security and environment. The next set of papers addresses crop physiological aspects in relation to growth factors such as radiation, CO 2, temperature and water. Experimental research and simulation modelling are used in mutual interaction. One important outlet for the generated and integrated knowledge is the ideotyping of crops. Improving resource-use efficiency in agriculture positively affects economic, environmental and agricultural objectives. Many papers presented during the Congress have directly or indirectly addressed this issue. A set of papers particularly focusing on nutrients and organic matter is presented in this volume. Again, a combination of experimental and modelling research is used to enhance understanding of the system and identify options for improvement. The final section addresses the design of integrated and ecological arable farming systems. Prototyping is

put forward as a promising tool to design and implement new farming systems. Indicators are presented that support evaluation of farming systems. Finally the contribution of model-based explorations in developing new farming systems is considered. A discussion follows on the notion that development of sustainable farming systems is not a matter of one-way research delivery, but rather a process in which researchers and target groups should cooperate, learn and develop in true interaction. This applies especially as the operationalisation of sustainability requires value-driven choices calling for a continuous interaction between society, its organisations and farmers on the one hand, and the scientists and designers on the other. We hope that the activities of the European Society for Agronomy and the Proceedings of its Fourth Congress will stimulate to serve the new perspectives of agronomy, i.e., to adopt ecological principles, to optimally manage the use of resources and to meet social and economic objectives.

Martin K. van Ittersum Wageningen Agricultural University Department of Theoretical Production Ecology P.O. Box 430 6700 AK Wageningen The Netherlands

Siebe C. van de Geijn Research Institute for Agrobiology and Soil Fertility (AB-DLO) P.O. Box 14 6700 AA Wageningen The Netherlands

References Porceddu, E. and Rabbinge, R., 1997. Role of research and education in the developmentof agriculture in Europe. Eur. J. Agron., 7: 1-13. Van Ittersum, M.K., Venner, G.E.G.T., van de Geijn, S.C. and Jetten, T.H. (Editors), 1996. Book of Abstracts, Volume I and II. Fourth Congress of European Society of Agronomy, 7-11 July, 1996, Veldhoven, The Netherlands, 736 pp. Van Ittersum, M.K. and van de Geijn, S.C. (Editors), 1997. Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use. Elsevier Science, Amsterdam, in press.

vii

Acknowledgements The editors gratefully acknowledge logistic and technical assistance of Loes Helbers, Irene Gosselink and Guido Venner in processing the flow of manuscripts. We also sincerely thank the reviewers of the manuscripts for contributing to the maintenance of the scientific standard of the papers: D. Auclair L. Bastiaans F. Bonciarelli C.J.H. Booij M.G.R. Cannell J.G. Conijn J.B. Dent P. Dijkstra M. Donatelli P.A.I. Ehlert B. Gerowitt J. Goudriaan D.J. Greenwood P. Gregory J.J.R. Groot J. Hassink A.J. Haverkort G. Hoogenboom C. Jambert B.H. Janssen S.C. Jarvis J. Kubat P.J. Kuikman E.A. Lantinga D.W. Lawlor J.F. Ledent D.K.L. MacKerron L. 't Mannetje H. Meinke K. Mengel J.M. Meynard S. Mikkelsen M.I. Minguez

G.M.J. Mohren J.D. Mumford H. Naber J.J. Neeteson B. Nicolardot S.E. Ogilvy J.E. Olesen F.W.T. Penning de Vries J.R. Porter G. Russell M. Schenk H. Schnyder J.J. Schr6der C.A. Shand A.L. Smit J.-F. Soussana J.H.J. Spiertz E.A. Stockdale P.C. Struik H.F.M. Ten Berge N. Van Breemen P. Van Halteren H. Van Keulen E.N. Van Loo M.A. Van Oijen W.H. Van Riemsdijk J.A. Van Veen J. Vos D.C. Whitehead A.P. Whitmore F.G. Wijnands D. Younie J.C. Zadoks

This Page Intentionally Left Blank

ix

Table of contents Preface ............................................................................................................................................................. v Acknowledgements ................................................................................................... ................................ vii Section I INTRODUCTION

Role of research and education in the development of agriculture in Europe E. Porceddu and R. Rabbinge .......................................................................................................................

3

Section 2 AGRICULTURAL LAND USE, FOOD SECURITY AND ENVIRONMENT

Land use transformation in Africa: three determinants for balancing food security with natural resource utilization P.A. Sanchez and R.R.B. Leakey .................................................................................................................... 19

Agro-ecological characterisation, food production and security P. Bullock. ..................................................................................................................................................... 29

The potential benefits of agroforestry in the Sahel and other semi-arid regions H. Breman and J.J. Kessler. .......................................................................................................................... 39

Chemical crop protection research and development in Europe R. Neumann .................................................................................................................................................... 49

Emissions of CO2, CH4 and N20 from pasture on drained peat soils in the Netherlands C.A. Langeveld, R. Segers, B.O.M. Dirks, A. van den Pol-van Dasselaar, G.L. Velthof and A. Hensen .......... 57

Section 3 CROP PHYSIOLOGY AND IDEOTYPING

Effects of CO2 and temperature on growth and yield of crops of winter wheat over four seasons G.R. Batts, J.I.L. Morison, R.H. Ellis, P. Hadley and TR. Wheeler. ........................................................... 67

Use of in-field measurements of green leaf area and incident radiation to estimate the effects of yellow rust epidemics on the yield of winter wheat R.J. Bryson, N.D. Paveley, W.S. Clark, R. Sylvester-Bradley and R.K. Scott. ................................................... 77

Simulating light regime and intercrop yields in coconut based farming systems 3'. Dauzat and M.N. Eroy ................................................................................................................................. 87

Improving wheat simulation capabilities in Australia from a cropping systems perspective: water and nitrogen effects on spring wheat in a semi-arid environment H. Meinke, G.L. Hammer, H. van Keulen, R. Rabbinge and B.A. Keating. .................................................... 99

Comparison of CropSyst performance for water management in southwestern France using submodels of different levels of complexity C. O. Stockle, M. Cabelguenne and P. Debaecke ......................................................................................... 113

Root growth of three onion cultivars A.D. Bosch Serra, M. Bonet Torrens, F. Domingo Oliv~ and M.A. Melines Pagbs ...................................... 123

Interspecific variability of plant water status and leaf morphogenesis in temperate forage grasses under summer water deficit J.-L. Durand, F. Gastal, S. Etchebest, A.-C. Bonnet and M. Ghesqui~re .................................................... 135 Evaluation of sunflower (Helianthus annuus, L.) genotypes differing in early vigour using a simulation

model F. Agiiera, F.J. Villalobos and F. Orgaz ...................................................................................................... 145

Options of breeding for greater maize yields in the tropics A. Elings, J.W. White and G.O. Edmeades .................................................................................................. 155

Section 4 MANAGING RESOURCE USE

Nitrogen budgets of three experimental and two commercial dairy farms in the Netherlands J.J. Neeteson and J. Hassink ........................................................................................................................ 171

Resource use at the cropping system level P. C Struik and F. Bonciarelli ...................................................................................................................... 179

The efficient use of solar radiation, water and nitrogen in arable farming: matching supply and demand of genotypes A.J. Haverkort, H. van Keulen and M.I. Minguez .......................................................................................... 191

Soil-plant nitrogen dynamics: what concepts are required? E.A. Stockdale, J.L. Gaunt and J. Vos ......................................................................................................... 201

Modeling crop nitrogen requirements" a critical analysis C 0. Stockle and P. Debaeke ....................................................................................................................... 217

Maize production in a grass mulch system - seasonal patterns of indicators of the nitrogen status of maize B. Fell, S.V. Garibay, H.U. Ammon and P. Stamp ........................................................................................ 227 Nitrogen transformations after the spreading of pig slurry on bare soil and ryegrass ~5N-labelled ammonium T. Morvan, Ph. Leterme, G.G. Arsene and B. Mary ...................................................................................... 237 Size and density fractionation of soil organic matter and the physical capacity of soils to protect organic matter J. Hassink, A.P. Whitmore and J. Kub6t ....................................................................................................... 245 Characterization of dissolved organic carbon in cleared forest soils converted to maize cultivation L. Delprat, P. Chassin, M. Lin~res and C. Jambert. ...................................................................................... 257 Analysis of impact of farming practices on dynamics of soil organic matter in northern China H.S. Yang and B.H. Janssen .......................................................................................................................... 267 Agronomic measures for better utilization of soil and fertilizer phosphates K. Mengel. ................................................................................................................................................... 277 Section 5 DESIGNING FARMING SYSTEMS

A methodical way of prototying integrated and ecological arable farming systems (I/EAFS) in interaction with pilot farms P. Vereijken ................................................................................................................................................. 293

The Logarden project: development of an ecological and an integrated arable farming system CA. Helander. ............................................................................................................................................. 309

xi

Integrated crop protection and environment exposure to pesticides: methodes to reduce use and impact of pesticides in arable farming F. G. Wijnands ..............................................................................................................................................

319

Use of agro-ecological indicators for the evaluation of farming systems C. Bockstaller, P. Girardin and H.M.G. van der Werf ................................................................................ 329

Model-based explorations to support development of substainable farming systems: case studies from France and the Netherlands W.A.H. Rossing, J.M. Meynard and M.K. van lttersum ............................................................................... 339

Learning for substainable agriculture B.M. Somers .................................................................................................................................................

353

Author Index ............................................................................................................................................... 361 Subject index ............................................................................................................................................... 363


Section 1 INTRODUCTION Role of research and education in the development of agriculture in Europe E. Porceddu and R. Rabbinge ........................................................................................................................ 3 Reprinted from the European Journal of Agronomy 7 (1997) 1-13


© 1997 ElsevierScience B.V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

Role of research and education in the development of agriculture in Europe E. Porceddu a'*, R. Rabbinge b aDepartment of Agrobiology and Agrochemistry, University of Tuscia, Via S. Camillo De Lellis, O1100 Viterbo, Italy bDepartment of Theoretical Production Ecology, WageningenAgricultural University, P.O. Box 430, 6700 AK Wageningen, The Netherlands

Accepted 16 June 1997

Abstract Agricultural research and education in Europe has played a major role in the advancement of agriculture and land use during the last century. The scientific basis of agriculture has been strengthened and the use of insight, knowledge and expertise in farmers' fields is widely adopted. As a result of this development, productivity per hectare, efficiency and efficacy of use of external inputs has increased considerably. During the last decades, objectives of agriculture and land use have broadened and this illustrates the need for further ecotechnological knowledge and insight to reach, in a balanced way, multiple goals of agriculture, productivity, protection of the environment, nature conservation and development. Research and education have to be developed in that direction to make agricultural science and technology more responsive to changing societal demandsl The disciplinary scientific quality and depth should develop in tandem with integrating problem-oriented multidisciplinary research activities. Systems approaches may serve as an instrument to that goal. © 1997 Elsevier Science B.V. Keywords: Systems approaches; Reorientation agricultural research; Broadened objectives

I. Introduction Since mankind started to exploit plants and other organisms to fulfill the changing demands for food and other needs, agriculture has been an important activity of mankind. The continuous evolution and manipulation of organisms and their function has been based on empiricism, knowledge, insight and expertise. The development has accelerated during the last century, when the scientific basis for agriculture strengthened and its simultaneous implementation became possible. The last decades have seen a broadening of aims and objectives of agriculture and

an increased importance of various environmental and socio-economic constraints. The ways that development took place, the change from old concepts to new perspectives and the challenges of agricultural research in general and more specifically of agronomy, are described in this introductory article to the Proceedings of the 4th ESA conference. The possibilities to reach those aims, and the new institutions and concepts which recently were introduced and adopted, are described.

2. Historical setting

* Corresponding author. Reprinted from the European Journal of Agronomy 7 (1997) 1-13

Scientific developments in agriculture during the

19th century were dominated by two scientists: Justus von Liebig and Gregorius Mendel. Their discoveries in the field of plant nutrition and the laws of heredity opened entirely new roads. Even today they still inspire scientific progress in fields very distant from the ones in which they originally made their contributions. A careful examination of their scientific experiences leads us to understand that, contrary to what is often reported, they worked in a lively and stimulating environment. It laid the basis for agricultural sciences and the institutionalization of research in agriculture; an event that Whitehead (1925) considered as the greatest innovation of the 19th century. The experiments carded out by private farmers, including the introduction of new crops, the establishment of chairs of agriculture at several university level teaching institutions, and the initiation in 1786 of the first public experimental farm near Braunschweig and a farm school in Hamburg, gave birth to agricultural research and experimental institutes in most European countries. The forestry school and the pomology institute of Weihenstephan were set up in 1803, followed by the research and teaching institutes at Hohenheim in 1818 and at Moglin in 1819. By the end of the century, Germany could boast as many as 87 research institutes, many of them in Prussia (Nomisma, 1996). The movement soon spread to France, where, at the initiative of private societies, farm schools were set up in Roville, near Nancy in 1822, Le Saussaie in 1828, and Bechelbon in Alsace in 1834. The first public institution became eventually operational in Versailles in 1848. In total there were 61 experimental stations in France at the end of the century (Nomisma, 1996). In the United Kingdom, the first experimental institution was set up in Rothamsted in 1844 as a private foundation, a form it was to retain until the end of the century; it was later joined by similar organizations in Woburn and Pumpherstone. By the end of the century agricultural colleges had been established at Bangor in 1889, Leeds in 1890, Aberystwyth in 1891, Nottingham in 1893, Reading in 1893, Cambridge in 1894 and Wye in 1894 (Speedy, 1994). In Italy, a private school opened in Tuscany in 1834 with a program of theoretical and practical lectures. Six years later it was transformed into a university level institution attached to the University of Pisa

(Coppini and Volpi, 1991). Within a few years of the unification of the country, in 1866, the University of Naples started planning a Faculty of Agriculture. It eventually became operational in 1872, though by that time it had been preceded by a Higher School of Agriculture, set up in 1870 by the University of Milan. Three agricultural experimental stations were set up in the same year. The last institution set up during last century in Italy was the Higher School of Agriculture at the University of Perugia in 1896. In Spain teaching started in 1855 at the Agricultural Central School in Aranjuez, close to Madrid, which had been promoted by the Ministry of Agriculture (J.A. Cubero, in litteris). In many countries of Eastern Europe similar developments took place, although in most cases much later. During 18501900 many National Agricultural Research Systems (NARS), including academic education and extension, were developed. However, during the last decade of the 19th century, many of these systems suffered due to considerable contraction of the agricultural sector. The story of the agricultural research institutes was very different in Denmark and the Netherlands, where the stimulus for extensive teaching and experimentation activities was provided by the great agricultural crisis of the last quarter of the 19th century. While other countries adopted protectionistic measures, these two countries reacted by rising the level of cultivation and agronomic techniques of their farmers. As a result, their competitive capacity increased. Agricultural research in Denmark was entrusted to four experimental stations, while teaching was concentrated at the Higher School of Agriculture in Copenhagen. In the Netherlands, the Faculty of Agriculture, then still a Higher School, was founded at Wageningen in 1876. Very quickly it gave rise to four distinct bodies: the Higher School of Agriculture, the School of Horticulture, the Agricultural Secondary School, and the Agricultural Colonial School. Agricultural extension was started and an intensive system of experimental stations and farmer field experiments initiated (Eveleens and Rabbinge, 1994). This is not the place for attempting a full-scale historical review of Europe's agricultural research, experimentation, and teaching institutions, but the few words devoted to their origin demonstrate not only the fervor that existed in Europe for agricultural

research and teaching, but also that this fervor made it possible to develop and exploit certain processes. In no more than a century these lead to greater transformations in agriculture than the ones that had occurred in the two preceding millennia. Both in arable farming and dairy farming, increases in land and labor productivity in the last 100 years were dramatic, compared with all ages before. In Italy, wheat production, which at the beginning of the century was of the order of 1 t/ ha, is today four times as great; the cultivation of 1 ha of wheat today absorbs 4 man days or 30 h of labor, while as many as 70 man days were needed only 70 years ago. Similar developments occurred in other industrialized countries. Average yield level or wheat in the Netherlands increased from 1.5 t/ha to 8.5 t/ha and labor requirement decreased from 300 to 15 man hours/ha (de Wit et al., 1987). At present yield levels of 10 t/ha are not exceptional. Similar developments can be observed in other crops. The increased productivity coupled with an expansion of the agricultural land allowed most of the European countries to produce enough food commodities and meet the increasing demand through population growth and diet change. Severe famine crisis could thus be prevented. Also, emigration, which drained

more than 40 million Europeans during the 19th century, could be prevented in this century.

3. Recent developments in agricultural productivity and employment Events during the last quarter of the present century, a period even the youngest scientists have experienced at first hand, are well documented. At the end of the 1960s the number of people served by a single European Union farmer was only half the present number (Table 1), and the amount of produce provided by such a farmer was half of what it is today (Table 2). This has left an ever larger number of people free to be engaged in other activities, to leave the rural areas and settle in towns (Table 1). In the same 25 years, agricultural land has decreased and within this acreage the arable crop land has increased, the yield per unit of land has increased by approximately 50%, and yield per unit of labor has tripled (Table 2). These facts made it possible for agricultural commodity prices to lag behind the increase in the general cost of living, enabling an ever larger number of people to enjoy

Table 1 Total and active population in EU countries. Situation in 1973 and trends 1970-1993. (Source: Eurostat, 1994; FAO, 1995) Country

Denmark Finland Sweden Austria Belgium France Germany Ireland Netherlands United Kingdom Greece Italy Portugal Spain

1993

1993 vs. 1970

Total popul, (1000)

Rural Rural/ T o t a l Agric. Agric./ popul, total labor labor total (1000) popul. (1000) (1000) popul. labor labor (%) (%)

Inhab/ arg. labor (#)

Total popul, (%)

Rural Total popul, labor (%) (%)

Agri./ tot. % labor (%)

Inhab./ arg. labor (%)

5155 5017 8633 7846 10251 56848 63694 3555 15 158 57908

644 2011 1356 3199 243 14400 7704 1512 1716 6341

12 40 16 41 2 25 12 43 II II

2567 2030 3912 3570 3744 21908 36 Ill 1149 6640 25 348

131 174 139 245 99 1195 1272 157 256 522

5l 8.6 3.6 6.9 2.6 5.5 3.5 13.7 3.9 2.1

39 29 62 32 104 48 50 23 59 Ill

105 109 107 105 103 100 105 121 ll6 104

64 88 89 89 37 98 68 106 95 97

111 93 100 Ill 98 105 134 103 145 100

48 34 41 41 55 41 62 52 78 73

202 320 260 270 190 270 170 230 150 140

10055 57 812 10300 39390

3671 17604 6732 8134

37 30 65 21

3715 20267 4464 11 868

791 1488 516 1212

21.3 7.3 11.6 10.2

13 39 20 33

ll4 107 ll4 117

88 92 100 71

98 105 139 96

45 40 53 32

260 260 220 360

Table 2 Total land and utilized agricultural area in EU countries. Situation in 1973 and trends 1970-1993. (Source: Eurostat, 1994; FAO, 1995) Country

1993 Total area (1000 ha)

Denmark Finland Sweden Austria Belgium France Germany Ireland Netherlands United Kingdom Greece Italy Portugal Spain

1993 vs. 1970 (2)/(1) (%)

Crop area/ (2) (%)

1

Utilized agric, area (1000 ha) 2

(2) (%)

(4) (%)

(3) 1970 (%)

(5) (%)

Output/ area (%)

Output/ labor (%)

4

(2)/ labor (ha) 5

3

4306 33 699 44 759 8385 3050 54 883 24871 7031 3694 24419

2751 2610 3359 3482 1412 30 217 17 162 4450 1997 17 178

64 8 8 42 46 55 69 63 54 70

92 96 83 41 58 59 68 17 46 35

19,2 11,6 19,2 15,6 17,8 21,3 10,1 29,1 8,3 30,3

92 87 97 89 88 91 126 92 91 89

94 94 91 85 96 94 145 65 105 84

90 91 89 43 53 58 60 24 40 37

183 190 173 208 207 207 177 162 128 121

153 137 135 144 154 136 143 158 187 152

282 263 232 297 320 280 255 255 239 165

13 196 30 098 9202 50 508

5785 16 800 3829 26 398

43 56 42 52

51 54 58 58

9,3 9,3 5,6 18,2

65 87 78 73

81 64 51 75

41 73 89 56

155 202 140 194

151 145 102 180

233 292 143 348

higher food standards and a better lifestyle. Today there are problems of overproduction. The increasing yield per unit of land, as result of research innovations and use of external inputs, which have been progressively introduced and adopted by farmers, is continuing. Greater income has also led to a greater demand for goods and services of high elasticity with respect to income, among them a cleaner environment, stricter requirements for food quality and the way food is produced. Environmental quality is nowadays considered as a 'superior good'. Its demand rises in proportion to income growth, while food production is an 'inferior good', the demand for which falls in proportion to increases in income (Engel' s Law). Demand on food quality increasingly dominates food quantity while concern for environmental health and safety grows. Natural resources are no longer perceived by policy communities as merely the medium to produce more and cheaper food but rather in terms of local and global ecosystem functioning. Ecosystem maintenance, bio-diversity, recharge of ground water, clean air and bequeath value are important topics, as agronomists well know. Agriculture is also a service to society, for example, as a means to preserve the landscape. However, there

is no doubt that such a service to society requires greater attention in planning and financing structures, and can only play an additive role for the farmer community. It cannot replace farming as an economic activity. Interventions using farmers income compensation, as it was recently experimented in the Netherlands, where a number of farmers in certain areas are subsidized to maintain landraces and primitive varieties, utilizing traditional technologies, may be seen as examples. National attitudes are nowadays dominated by the effects of market globalisation, which modifies trade flows, reallocates production and consumption at the world level, thus influencing agricultural activities and natural environment and, quite generally, setting the scene for any development in rural areas.

4. The influence of agricultural policy and trade agreements At present, the situation in EU agriculture can be depicted as follows: agricultural employment decreased by 2% per year during the last 25 years, but in a number

of member states agricultural employment still exceeds 5% (Table 1); land and labor productivity are still limited in some member countries (Fig. 1), in spite of progress during the last 25 years (Table 2); land per unit of labor is still limited in a number of countries (Table 1), and a high percentage of agricultural laborers remains (Table

•

•

25

• Gr

20 ,P • IRL

-110

,E

,I **

2); •

pro capite GPD (Gross Domestic Production) in member states is inversely proportional to the percentage of labor in agriculture (Fig. 2).

Taking into account these facts, the restructured EU-CAP (Common Agricultural Policy) envisages an agriculture with: •

more technology, but also with greater diversity; a smaller cultivated acreage, but also with more land dedicated to special uses; a more stable production, but also with a greater variety of produce; a smaller number of farmers, who, however, perform several and more diversified activities.

• • •

*Dk

*UK,B

The reform of the EU-CAP was followed by a further extension of trade agreements in the form of General Agreement on Tariffs and Trade (GATT). The most recent economic analyses (Anania et al., 1996) indicate that these agreements would not distort the EU-CAP, although the impact may not be negligible. Analyses also underscore the need that the EU

--

0

I ....

5000

I

I

10000

15000

20000

25000

30000

pro cap/h) GDP

Fig. 2. Relationbetween pro capite GrossDomesticProductionand labor force in agriculture in EU countries during 1993. agricultural system maintains its high competitiveness in terms of low cost in combination with high product quality, this in the domestic market as well as in the international one. The need represents the nodal point of the European agriculture in the near future: lower costs, including environmental costs, better quality and toxicologically safe products. These involve the efficiency of production processes and farm technologies. That requires a policy presently adopted by the European Union. Yet, it is not sufficient. The relative weight of various objectives has to be reconsidered. The choice of such a weight factor may result in considerable differences in land use in the future, as shown in an explorative study on reformulation of the Common Agricultural Policy of the European Union (Rabbinge and van Latesteijn, 1992; WRR, 1992). Using modem simulation and systems approaches, and incorporating appropriate technical information, a series of land use scenarios were generated, based on different priorities for objectives that dominate agricultural development: o

jA

B

k,

~m

•

O I&l i

]mooo

F

•

i

o 1000

2000

~

•

4000

Output per hectare (ECU's per ha)

Fig. 1. Trends in agricultural productivityin EU countries during the period 1970-1993. Lower left points 1966-1970, upper fight points 1986-1990. (Source: Eurostat, 1994; FAO, 1995).

•

free market and free trade, based on economic efficiency, i.e. maximum yield at minimum cost; regional development, with the aim of maintaining employment in agriculture at the highest possible level but within the constraints of a productive agriculture; nature and landscape conservation, by the creation of natural reserve areas separated from the agricultural ones and thus minimiz-

ing land use for agriculture and maximizing land productivity; environmental protection, minimizing the negative effects of agriculture; in other words, minimum use of pesticides and emission of nitrate, and other environmental side effects of agriculture, per unit of product or per unit of land. The results provided a framework for assessing the strategic options on which policy makers may base their decision. Indeed, the exploration of options showed very marked differences: •

•

•

as far as land use is concerned, the maximum acreage used by agriculture in the case of regional development is three times as great as the minimum associated with the free market hypothesis, but 40 out of the present 127 million hectares will nevertheless have to be taken out of use in all scenarios (Fig. 3a); as far as employment is concerned, all scenarios give rise to a further downturn of agricultural employment, but the size of the exodus may vary between 3 and 4.5 million man power units depending on the chosen scenario (Fig. 3b); quite independent of the scenario actually adopted, policy measures can successfully promote more environmentally friendly production methods by limiting the use of fertilizer and reducing the large scale use of crop protection agents, avoiding adverse effects on the environment (Fig. 3c).

The outcomes of the scenarios demonstrated the enormous challenge and chances for reorientation of the CAP, and for agricultural research. It becomes even more important to do such explorative studies when the European Union will be further extended in the near future. The considerable increase of land area, the tremendous possibilities for agriculture and the needed readjustment of the common agricultural policy in those types of situations require extensive explorations. Therefore, explorative studies on land use and agrotechnologies should become major issues on the future research agenda.

5. Agronomy and multiple goals in agriculture and land use The changes of agriculture from a purely production-oriented activity into a science based production sector, trying to meet productivity, efficiency and efficacy aims, has been of considerable importance during the last decades. Agriculture has broadened and diversified its objectives, rendering explicit and important a number of economic, environmental, social and nature-protection roles that in the past were simply passed over in silence. The concept of a good agricultural practice - where farmers produce in an economic and socio-political environment conducive to long-term conservation and use of natural resources, and do not want their source of survival and generation of income to disappear - has always been the foundation of agronomic teaching. Indeed, the same Justus von Liebig, often blamed as the father of chemistry in agriculture, way back in 1855 wrote as follows: 'The task of the farmer is not to achieve high crop yield to the detriment of the field, which only causes it to impoverish earlier. Rather, it should be in his own interest, as well as society's, to achieve high yields that are constantly increasing on a permanent basis.' (von Liebig, 1855). The concept is echoed in the 'sustainability' idea, entered into a common use thanks to Our Common Future, the report of the World Commission on the Environment and Development (1987), and defined by the Consultative Group for International Agricultural Research (TAC, 1989) as the successful management of resources for agriculture to satisfy changing human needs while maintaining or enhancing the quality of the environment and conserving natural resources. The question of sustainability, consequently, arises when the resources used for production are placed under stress, as is now widely happening in industrialized countries due to excessive irrigation, fertilizer and pesticide application, and in developing countries where increasing population pressure continues to strain product on resources. An agricultural production system is not sustainable if it leads to declining productivity, degrades the resource base, or is not economically viable. The given description of unsustainability does not necessarily lead to a precise definition of sustainabilty. This is not easy as

.AND USE 150-

"~" 100 r,-

o..,

E t~

2 30% slope) % low cation exchange capacity % high phosphorus fixation (by Fe) % sandy % vertic (black cotton soils) % gravel layer % calcareous (Fe and Zn deficiencies) % saline % alkaline % shallow to rock (2 GLAI units between 1994 and 1995) and total incident radiation levels differed (by 136 MJ m -2 between 1994 and 1995), the HAD curves were significantly displaced. In this study, HAD did not give a consistent explanation of yield over the two seasons although it did demonstrate a decreasing return from increasing green area, suggesting that yield is more closely related to absorption of solar radiation than to leaf area alone. As an integral over time, HAD has a similar disadvantage to AUDPC in that it does not differentiate between large GLAI for a short period and small GLAI for a long period (Johnson, 1987). Nor does it account for the diminishing effect of increasing canopy size on the proportion of light intercepted (Monteith and Unsworth, 1990). This

85

has implications for disease control strategies in that a large canopy may be able to tolerate some loss of green area without an economically significant effect on yield. On the other hand, any loss of green area from a small canopy could have a serious effect on yield, making protection an economic necessity. In this study, different fungicide treatments resulted in very different values of HAD but equivalent yields. The relationship of yield to HAA gave a better correlation than that with HAD in both experimental years. With relatively large canopies, only two seasons to provide variation in incident radiation, and a large proportion of the treatments giving good disease control, most of the data points in this relationship are clumped. A more thorough test of the predictive power of HAA must depend on data from a broader range of circumstances. Nevertheless, the RUE of green leaf tissue determined here was not only consistent between the two seasons, but was similar to the RUE determined from separate disease control experiments at this site (1.2 g MJ-i; Bryson e t al., 1995) as well as to the radiation conversion coefficients reported for several different crops (Monteith, 1977) Although HAA gave the best estimate of harvested yield of the three models tested, the intercept of the relationship was found to be highly sensitive to the start date taken for the period of integration. For instance, when the relationship of yield to HAA was tested from the end of May the intercepts of the equations became negative, but the slopes were unchanged. Waggoner and Berger (1987) obtained negative intercepts when they related peanut yield to HAA. They concluded that the negative intercept indicated that no peanuts were set at very small HAA values. The values of HAA presented in this paper were calculated from 20th June in both years so that they could be compared on a common basis. On 20th June both crops were assessed as being at the start of anthesis (GS61; Tottman, 1987). However, assessment of growth stages can be prone to assessor error and the distinction between the beginning and middle of anthesis is uncertain when assessments are made on a weekly basis. It is therefore possible that the wheat crops in 1994 and 1995 were at different developmental stages on this date. It would appear that the integration of intercepted radiation over time must be

combined with precise and accurate records of growth stages if the approach, based on HAA, is to lead to an improved capacity to model absolute yield, rather than just yield 'loss'. Waggoner and Berger (1987) suggest that the amount of solar radiation intercepted by the green portion of a crop canopy is all that is needed to predict crop loss. This was obviously not the case in this study. Johnson (1987) pointed out that, when used over an entire season, HAA/yield models may not account for different source-sink relationships at different crop stages. Whilst restriction of the HAA model to the period when the harvested portion of the crop was developing provided a consistent relationship in this case, there are likely to be circumstances in which sink limitation will reduce RUE. The period before flowering is particularly important in determining the sink capacity of wheat (Evans and Wardlaw, 1996) and it may be necessary to monitor growth during this earlier period if a cropbased explanation of yield variation is to prove sufficiently robust to support commercial decision-taking. HAD and HAA, as originally defined by Waggoner and Berger (1987), do not take account of ear green area. Diseases other than yellow rust may affect yield by effects on ears (Jones and Odebunmi, 1971). In conclusion, definition of crop productivity as the product of radiation interception by green leaf tissue and RUE provides a framework for understanding disease induced reductions in yield. The approach described here is not intended as a predictive tool but it is envisaged that it will lead to the development of models which may be utilised in crop management decisions. Integrals such as HAA are an improvement over the more conventional forms of disease progress analysis, but further work is needed, particularly to account for the timing of disease effects in relation to the processes of crop development.

Acknowledgements The contribution of the Home-Grown Cereals Authority and the UK Ministry of Agriculture Fisheries and Food to the funding of this research is gratefully acknowledged. Thanks are due to Dr Alan Gay for the statistical analysis, and Miss Diane Moss, Mr

86

Jess Hunt and other ADAS colleagues for technical support.

References Anonymous, 1976. Manual of plant growth stage and disease assessment keys. Ministry of Agriculture Fisheries and Food, Harpenden. Asrar, G., Fuchs, M., Kanemasu, E.T. and Hatfield, J.l., 1984. Estimating absorbed photosynthetic radiation and leaf area index from spectral reflectance in wheat. Agron. J., 76: 300-306/ Bhan, V.M. and Pande, H.K., 1966. Measurement of leaf area of rice. Agron. J., 58: 454. Bryson, R.J., Sylvester-Bradley, R., Scott, R.K. and Paveley, N.D., 1995. Reconciling the effects of yellow rust on yield of winter wheat through measurements of green leaf area and radiation interception. Aspects Appl. Biol., 42: 9-18. Campbell, G.S., 1986. Extinction coefficients for radiation in plant canopies calculated using an ellipsoidal inclination angle distribution. Agric. For. Meteorol., 36: 317-321. Campbell, C.G. and Madden, L.V., 1990. Temporal analysis of epidemics I: Description and comparison of disease progress curves. In: C.G. Campbell and L.V. Madden (Editors), Introduction to Plant Disease Epidemiology. Wiley, New York. Evans, L.T. and Wardlaw, I.F., 1996. Wheat. Chapter 22:501-518. In: E. Zamski and A.A. Schaffer (Editors), Photoassimilate Distribution in Plants and Crops: Source:Sink Relationships. Marcel Dekker, New York, 905 pp. Fry, W.E., 1975. Integrated effects of polygenic resistance and a protective fungicide on development of potato late blight. Phytopathology, 65:908-911. Gale, M.D. and Youssefian, S., 1985. Dwarfing genes in wheat. In: G.E. Russell (Editor), Progress in Plant Breeding, Vol. 1. Butterworths, Oxford. Hansen, J.G., 1991. Use of multispectral radiometry in wheat yellow rust experiments. Bulletin OEPP/EPPO 21, 651-658. James, W.C., 1974. Assessment of plant diseases and losses. Annu. Rev. Phytopathol., 12: 27-48. James, W.C. and Teng, P.S., 1979. The quantification of production constraints associated with plant diseases. Appl. Biol., 4: 201267. Johnson, K.B., 1987. Defoliation, Disease, and Growth: A Reply. Phytopathology, 77: 1495-1497. Jones, D.G. and Odebunmi, K., 1971. The epidemiology of Septoria tritici and S. nodorum IV: The effect of inoculation at different growth stages and on different plant parts. Trans. Br. Mycol. Soc., 56: 281-288. Large, E.C., 1952. The interpretation of progress curves for potato blight and other plant diseases. Plant Pathol., 1: 109-117. Lim, L.G. and Gaunt, R.E., 1981. Leaf area as a factor in disease assessment. J. Agric. Sci., 97: 481-483. Madden, L.V., Pennypacker, S.P., Antle and C.E., Kingsolver,

C.H., 1981. A loss model for crops. Phytopathology, 17: 685689. Monteith, J.L., 1977. Climate and the efficiency of crop production in Britain. Philos. Trans. R. Soc. London Ser. B, 281: 277294. Monteith, J.L. and Unsworth, M.H., 1990. Principles of Environmental Physics. Edward Arnold, London. Murray, G.M., Ellison, P.J., Watson, A. and Cullis, B.R. 1994. The relationship between wheat yield and stripe rust as affected by length of epidemic and temperature at the grain development stage of crop growth. Plant Pathol., 43: 397-405. Owen, P.C., 1968. A measuring scale for areas of cereal leaves. Exp. Agric., 4: 275-278. Parker, S.R., Shaw, M.W. and Royle, D.J., 1995. The reliability of visual estimates of disease severity on cereal leaves. Plant Pathol., 44: 856-864. Ross, G.J.S., 1987. Maximum Likelihood Program. Release 3.08. Numerical Algorithms Group, Oxford. Shaner, G. and Finney, R.E., 1977. The effect of nitrogen fertilisation on the expression of slow-mildewing resistance in Knox wheat. Phytopathology, 67: 1051-1056. Sylvester-Bradley, R., Stokes, D.T. and Scott, R.K., 1990. A physiological analysis of the diminishing response of winter wheat to applied nitrogen - I. Theory. Aspects Appl. Biol., 25: 277287. Sylvester-Bradley, R., Goodlass, G., Paveley, N.D., Clare, R.W. and Scott, R.K., 1995. Optimising the use of fertiliser N on cereals and parallels for the development of fungicide use. In: H.G. Hewitt, D. Tyson, D.W. Hollomon, J.M. Smith, W.P. Davies and K.R. Dixon (Editors), A Vital Role for Fungicides in Cereal Production. Bios, Oxford, pp 43-56. Teng, P.S., 1983. Estimating and interpreting disease intensity and loss in commercial fields. Phytopathology, 73:1587-1590. Teng, P.S., 1985. Construction of predictive models II. Forecasting crop losses. In: C.A. Gilligan (Editor), Mathematical Modelling of Crop Diseases, Academic Press, London, pp. 179-206. Tottman, D.R., 1987. The decimal code for the growth stages of cereals with illustrations. Ann. Appl. Biol., 110: 441-454. Waggoner, P.E., 1977. Simulation of modeling of plant physiological processes to predict crop yields. In: J.J. Landsberg and C.V. Cutting (Editors), Environmental Effects on Crop Physiology. Academic Press, New York, pp. 351-363. Waggoner, P.E. and Berger, R.D., 1987. Defoliation, disease and growth. Phytopathology, 77: 393-398. Webster, J.P.G., 1987. Decision theory and the economics of crop protection measures. In: K.J. Brent and R.K. Austin (Editors), Rational Pesticide Use. Proceedings of the 9th Long Ashton Symposium. Cambridge University Press, Cambridge, 348 pp. Welles, J.M. and Norman, J.M., 1991. Instrument for indirect measurement of canopy architecture. Agron. J., 83: 818-825. Whelan, H.G. and Gaunt, R.E., 1990. Yield loss:disease relationships in barley crops with different yield potentials. Proceedings of the 43rd NZ Weed and Pest Control Conference 1990, pp. 159-162.

© 1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

87

Simulating light regime and intercrop yields in coconut based farming systems J. Dauzat a'*, M.N. Eroy b aCIRAD/GERDATPlant Modeling Unit, P.O. Box 5035, Montpellier, France bDavao Research Center~Philippines Coconut Authority, Philippines, Philippines Accepted 16 June 1997

Abstract Intercropping experiments of corn and mungbean have been conducted at the Davao Research Center of the Philippines Coconut Authority under coconut stands at different densities. Yields obtained in these experiments are more or less linear functions of the photosynthetically active radiation measured under the trees. In order to extrapolate these results for other palm ages and densities, the following steps have been achieved: (1) measurement and modeling of the architecture of 5, 20 and 40 year old palms, (2) generation of virtual coconut stands, (3) simulation of light transmission using these virtual stands, (4) prediction of intercrop yields by combining the results of intercropping experiments and the simulated light transmission. The simulated light transmission under 5, 20 and 40 year old coconut stands were close enough to field measurements to consider that both computerized coconut mock-ups and radiative models are valid. Radiative simulation experiments could thus be performed in order to assess the effect of coconut density on photosynthetically active radiation (PAR) transmission as well as the effect of frond pruning. Results exhibit a nearly linear relationship between light transmission and tree density. Pruning also appears as an effective mean of increasing the light permeability of coconut stands. These results are interpreted in terms of corn and mungbean yields by combining radiative simulations and field intercropping experiments. © 1997 Elsevier Science B.V. Keywords: Plant architecture; Radiative transfers modeling; Coconut; Intercropping; Cocos nuciferal

1. Introduction An accurate modeling of the photosynthetically active radiation (PAR) regime is essential to predict the behavior of intercrops in agroforestry systems such as coconut based farming systems (CBFS) where PAR received by intercrops is commonly 1/4 to 1/3 of the PAR in open field. Intercropping experiments under coconuts in the Philippines demonstrated * Corresponding author.

that, in the absence of strong water deficit and with a proper fertilization supply, the intercrop yields are more or less linearly related to the available PAR (B6nard et al., 1996). Thus optimizing CBFS can be achieved mainly through the choice of coconut density or by frond pruning in order to get sufficient light for intercropping. Few coconut density trials exist because they are lengthy and expensive. Moreover, the density is not the only factor to be taken into account: the development of palms, their planting pattern and the radiative

Reprinted from the European Journal of Agronomy 7 (1997) 63-74

88

conditions affect the intercrops potentialities. A radiative model accounting for all these factors is thus essential for understanding and optimizing CBFS. Classical radiative modeling represents plants as simple shapes (e.g. spheres, cones, cylinders...) without taking into account the actual plant geometry inside these shapes (Brown and Pandolfo, 1969; Chiapale, 1975; Charles-Edwards and Thorpe, 1976; Li and Strahler, 1985; Riou et al., 1989) or uses global statistical canopy properties but disregard spatial arrangement between plant items (Kimes, 1984; Sinoquet, 1989). Recent 'architectural' models such as the coconut models used in this study offer a much more realistic representation of plants because they are based on their botanical description, taking into account the precise shape of plant organs as well as their spatial or geometrical organization in threedimensional space (Reffye et al., 1988; Goel et al., 1991; Aries et al., 1993; Dauzat, 1995). The recent development of software generating realistic threedimensional models of plants opens new possibilities for the radiative transfer modeling. This elicited the interest of the Plant Modeling Unit of CIRAD to develop a specific radiative software exploiting the three-dimensional information attached to the computerized plant mock-ups. Initial studies on oil palm and coconut in Ivory Coast (Girard, 1992; Dauzat, 1994) showed that radiative climate can be assessed acutely on these computer models and plant architecture variations having significant bearing on transmission can be identified.

Moreover, climatic factors, especially the quantity and variation on sky condition, can be assessed. This enables prediction as to radiative climate in a given stand considering its density, age, planting pattern and seasonal fluctuations of radiation. The horizontal distribution of light at the soil level is also assessed as illustrated in Fig. 5.

2. Material and methods 2.1. Experimental site and plant material 2.1.1. The site Field experiments have been conducted at the Davao Research Center (DRC; 07 ° 05 N, 125 ° 57' E) of the Philippines Coconut Authority (PCA). The climate is characterized by average annual rainfall of 2400 mrn/year fairly well distributed throughout the year, a relative humidity of 73-82%, a mean temperature of 27°C and annual sunshine duration of 2350 h. The gently sloping soils are well drained and their average composition is 28% sand, 31% silt and 41% clay. The pH (H20) is 6.6. 2.1.2. The coconut stands Tree description and radiative measurements have been done within three stands of LAGUNA TALL coconuts. The first stand is composed of 5 year old trees planted in a 9 x 9m triangular pattern with rows oriented North-South. The second is composed of 20

Table 1 Corn and mungbean varieties used in intercropping campaigns at the PCA Davao Research Center Crop

lntercropping campaigns 1st

2nd

3rd

4th

Corn

USM var. 6 USM var. 2 SMC 357

Mungbean

Pag-asa 7 Pag-asa 3 BPI Mg 7 BPI Mg 9 BPI Mg 60 BPI glabrous 3

USM var. 6 USM var. 2 USM var. 10 SMC 357 IPB H921 P3246 Pag-asa 7 Pag-asa 3 BPI Mg 9 Candelaria Local, Davao Sariaya

USM var. 6 USM var. 5 USM var. 8 IPB H921 P3246 XOF 62 Pag-asa 7 Pag-asa 3 BPI Mg 9 Candelaria Local, Davao Sariaya

USM var.6 USM var. 2 USM var. 5 P3246 CPX 3007 P3022 Pag-asa 3 BPI Mg 7 BPI glabrous 3 BPI Mg 9 PAEC 5 Local, Bansalan

89

year old trees planted with the same pattern as above. The third stand was planted 40 years ago in square design, with a 8 x 8m spacing. Rational felling of certain palms within the 20 year old stand 1 m above the ground created a density and lighting gradient which determined the intercropping treatments (Fig. 7): • • • •

rows, leaving a free corridor either side of the palm rows. They received the fertilizers and phytosanitary treatments usually practiced in the region. Different combinations of varieties were tested for each campaign as indicated in Table 1. More details about these intercropping trials are given in B~nard et al. (1996).

2.2. Description of coconut stands

(L1) standard interrow (control) (L2) standard interrow with greater lateral lighting (L3) thinned interrow (L4) very thinned interrow

The stand description included the description of the individual trunks (diameter, height, projection and azimuth; see Fig. l) as well as of the number of green fronds per tree. In order to assess the inter-trees variability, 50 palms were sampled within the neighborhood of the plots used for radiative measurements for each age group. Ten palms were sampled to get the phyllotaxic angle from leaf scars along the trunk. The value was controlled later on other trees by angle measurement between fronds of rank 9 and 14. The frond length (petiole and rachis) was measured

2.1.3. The intercrops Four intercropping campaigns have been practiced for corn and for mungbean (Mungo radiata) within the thinned 20 year coconut stand between 1991 and 1994. The land was tilled prior to sowing with a disc plough and ridged. The crops were planted in NorthSouth strips down the middle of the coconut inter-

level of ,~q lower frond

I t 1 I I I t I I

diameter

I •

= i

0

I ! i I

~\e'*, / ~ - r a c h i s . ~

I 1 t t I I I

an~]le i J

"trunk projection"

Fig. 1. Definition of some coconut architectural parameters.

/e-

90 on 100 dry fronds from different trees for each age group. In addition, to study intra-tree variability, all the dry fronds produced during a year have been measured on five trees per group. The frond inclination at the junction of petiole and rachis was measured on the most number of leaves on five trees per age group using an electronical clinometer. In order to model frond curvature, the height of some points along the rachis was measured. All accessible fronds (i.e. fronds of rank above 6 or 7) of 10 young palms (under 10 years old) were sampled. The leaflet number was counted on both sides of three fronds taken from the five trees selected per age group. This counting was done on 50 cm long sections of rachis in order to assess the spacing of the leaflets. Three unbroken leaflets were taken on each side of the 50 cm sections in order to measure their length, width and surface area. These measurements were obtained by an opto-electronical planimeter (LICOR leaf area meter) after masking the gaps within the lamina if ever. The vertical and horizontal angles of leaflet insertion (see Fig. 1) were measured on photos of 10 cm long segments of rachis. All the fronds of an apparent phyllotaxic spire were taken from two trees of each age group. Segments were then cut approximately every 40 cm for photos.

2.3. Modeling of coconut architecture The morphogenetic growth pattern of coconut adheres to Corner's architectural model (Hall6 et al., 1978), which is characterized by the existence of a single leaf axis with lateral inflorescence like in oil palm and papaya. Thus the description of the plant topology is quite simple, but an accurate geometrical description is needed for our purpose. Some palm features were conveniently characterized by a mean and a SD, assuming a Gaussian distribution (Dauzat and Eroy, 1995). It is the case for the parameters of the trunk (height, diameter, projection, azimuth), the frond number and their phyllotaxy, the number of leaflets on each side of the fronds. Other features have been fitted with simple functions (Dauzat and Eroy, 1995). For instance, a power function was used to fit the frond inclination at petiole end against the frond rank, a quadratic function to fit the leaflet spacing against their rank and a sinusoidal function to fit the leaflet length against their position on the rachis. The leaflet area has been fitted with a power function of their position on the rachis. The frond curvature is modeled as a flexing beam subjected to gravity by a sub-program. The two parameters of this sub-program (the Young modulus and

Fig. 2. Simulated mock-upsof 20 (unpruned~pruned),5 (unpruned) and 40 year old (unpruned/pruned)coconut trees.

91 •

•

•

• •

•

• I ,': i ~iei iiii~ii~iiiiiiiiiiiiii~iieiii~iiiiiiiiiiiiiii!iiiiiiiiiiiiiiii~i I

•

•

•

•

•

•

6i!i!ilili!iiiiiiiiiii!iiiii!jiiiiiiiiiiiiiii!ii!iiiiiiii!iiiiii!iijiiiij •••:ii!!••t::!!!!!i!!•••ii]i !•ii •!!•!!!ii !•ii•!•ii•i i•!i!•!i!i i i!i !i!i!!i!i!i !• O O I:~'~,!:~eiiiiiiii:iliti~ili iiiiilO~ii!ii~iii~iiiiiiiiii!iliiiii!i!i!ii !

O

!ii!iiiiiiiiiiii!iiiiiiiiiiii!iiii!i!ii!ii:iiii!iiiiiiii!iiii•iiiiiiiii!ii!i!iiiii iii!i t i ii i i i! i i iii iiiii iii i ii ii!iiiii! iii ! i ii!i i iioi ii i ii ii i!!i i i i

to::iiii iiiiiiieii!iiiiiiiiii!i!iii!ii!iiii!iiiiii iiiiiiiiii!iiiiiiiiiiiiil • • J": ,:::' ~O ~!~!~!!!!iiiiiiliQiiiiiiiiiiii~i!i~ii!iiiiiiiiiii!Qiii] /:e~ ~: i!J!ii!iii!ii!!ii!iiii!iiiiiiiiilJliiiiiiiiiiiiiiiiiii!i~il !• • I:_: ::!

O

iiii;i;iiiiii:i!ii!iiii)iiii :iii i iiiiiii !iiiiii ;?iiiiiiii!!i iiiiiiiiiiii!iiiiii o i!ii!;i!iiiii:iiiiii; !iii !i!i

O

:i!iiii!iiiiii!i,iiiiiiiiiiii,iiiiiiiiii:ii?i!ii?iiiii:i:,iiiii o i::i,ii!iiiiii

•

•

lei i!i !i i !i !!!ii iiiiiiiii!!!ii!i!i!iiiii!iiiiiljiliiiiii!i!iiiii!!iiiii!iiiiiii !i!! I•

O

O t:: i

:iii~i~ii:i~ii:~:ii~.i!ii~ii:ii~iii!iiiii~:~.i:i~iii~:!.:~:ii:~;~ii~ii:i~i~:~i!?:.i~:i!igi:i~i~.~:i~i~.i~i.:~i@.~ii:i~.i:i!i~i:,!~di!;i~ii!!iii

:el;ii!i!i~!!?iiiiiiiiiiii!ili!;i;!iiiiiiii!iiiiiiiii!iiiiii!iijiiioil •

•

O

•

•

•

•

•

•

•

•

Fig. 3. Simulated scenes with triangular design (left) and square design (right). Circles indicate coconut positions and inner rectangle represent the effective area used for radiative simulations (assuming that the stand portion within this area is surrounded by identical stand portions all around). The triangular design was used with a distance between trees of 8.5, 9, 10 and 11 m and the square design was used with a distance between trees of 8, 8.5, 9 and 10 m.

meters, like the frond length or the number of leaflets, are given with the intra-tree variability. One parameter file was created for each age group. In order to test the effect of pruning on transmitted radiation we also simulated pruned trees using the same parameter files but limiting the number of fronds to 18 (Fig. 2). To analyze the effect of the planting patterns and of the tree density, we created scenes with square and triangular designs, at different spacing (Fig. 3). Scenes in square design had 56 trees with eight rows of seven coconuts. Four densities were used, 100, 123, 138 and 156/ha, corresponding to distances between trees of 10, 9, 8.5 and 8, respectively. Four

the conicity) have been fitted using a specific interactive program. 2.4. Simulation of virtual coconut stands The coconut generator program is a software which computes the tree geometry through the functions chosen for modeling. The palm generation is stochastic, i.e. restitutes the observed tree variability (Reffye et al., 1995). Thus the input parameters are specified with their variability. Some parameters pertain to the global features of the observed population with the inter-tree variability: phyllotaxic angle, height and diameter of trunk, number of fronds. Other para•

•

•

•

0

•

.... ~

•

iii! il ii ::~!!i~~: ::~::: .... ......

O

::oii:~::!::i i::i~il i iiiiilil il i~iiiii!i~iii:,ilitiiii • '::~ :' ~iiii!i.i! iii:i!i;iii!ili!!ii!ii.....~,ii~,~e', o::: t:!iii!Oi;;ii iii.:iiii.~ii!.i.i!iii.li~ •

:~~,ii:i;~e i~ i~i~i~i :',:iQ:,::ii ~i:,:si~eiiiiiiii;~i~iii; : i;;:~c.,.:!:1

fO!!!ili~s~t::!iiiii~Oiii:.ii~!~i~i i i:iOi!i!~ii !!il2::¢!:!ii ~i i i i:.i~:i:i i~it:i ,~!~

O

O O

~,~o,~: , i ..~eiil;.i~..~i!i .li ~'i !i i......... i!','~

0

%..:

O

0

iiii :::i:~i:::i:i::~":!ii!::iii:::!i:i:i: O

~.:!:::i:~:~:eii.i :!;!]:;i!.L.ii !i:i!.O!.i!:i.i:.ii£1~i!:!i!i;i :O!ii:] ~e:~i:i!,!~i~:;i~!:,:!i ! ,~e!i!i!i i i'.i:!.i:i~i !i :i~!:!:i~!iO!ii•!i i i i:i!!:i!!!;i:!

0

:el ~:~ i i! :~e~,i!iliiiili ii!i iiiii~ili~iiiiiiiii • •~::~:i~,:::?,:~iei i lf:iii!i!~ii!i.i!iilji~iiiiiiiii!ii!!ii!i@i!iiii i !i!!i~:!i

O

• .. 1:...

:i:.:

... :..:..:..: :::::::::::::::::::::::::::::::::::::::::::::::::::::::::

:.:

i~:i::Si:.?!`!:::i:!i!]:i:i~if~!i!~]!?i!]:!i.::~!!?i]fi!ii~!i!ii~!!]i!i!~i~fi!]ii!]i):~!i!iii!]i:]i!i!i!~ii[i:~!~.~!:

Oi,: ~::::!:',::i::!ilt:::i~.iO ':: !~i~i !tl! ;!ti,:ii £iti':lii• i:,'~i:i', 'tti, 'i'::!i',! • 0

f"

""

:.o! t;t:~::~ii i!~ili !6i!!!i!! i!ii! i ~ii~,ii i i ! t~t o ! !! ii eiii~i!!i','~i i!~',i !i i i~i i!i!!~!!iiii!~i!!~!!~'~1!

!.:

:!iii!iii:i:ii-:ii:i!iiiiiiiii:i.i!~i!!!:iiiii!~ii:i!:iiii!]iii!iiiii~!il

...... :iiiii!iiiiii!:!iiii!iiiii~:i!!i::i!i

...... :!:::!iiiii::iii!'!i? i!i'i!:

.i:;i~Tq:-;i::!;:;.:~ii;~i:i:f;::; i::i~:!:ii; ::i:-ii!!!!i[!i!i t::i~:i!~i. !::;i!!;~ ~if!~ i!.:~i~.~i~!if~i:~t.ii;;!::~!;i!i!~: /: i!ii.!;.:h~!.::;.:~;!1

0 ~i!:i!~~i~:.ii!~i~!ii~!0ii~i~i~!~]:!~i~ii!!ti~!!!~i~!~i!~ii~f:iii~i~i~i~i!i~.ii~:i0!~!~!ii~i::i~i~:]i~:.~::i~::~:~ii]:~ •

0

•

0

•

Fig. 4. Simulated thinned stands. Same as legend in Fig. 4 with felled trees represented by empty circles. Thinning original stands at 143 trees/ ha leads to densities of 107 (left) and 72 trees/ha (right).

92

Fig. 5. Simulation of the light transmitted under a 20 year old coconut stand at density 143 during a day. rows of seven coconuts and four rows of six coconuts were included in scenes with triangular planting. Four densities were used: 81, 115, 143 and 160 trees/ha, corresponding to planting distances of 11, 10, 9 and 8.5 m, respectively. Three other designs were simulated by removing some palms from the triangular design at density 143. The first one represents the DRC experiment (Fig. 7). The other two result from a thinning down to 107 and 72 palms/ha (Fig. 4). The same computerized trees are used for simulating stands at a given age, assuming that the coconut density and planting design do not deeply affect the tree development. Data previously collected in a coconut density experiment in C6te d'Ivoire with densities ranging from 115 to 180 trees/ha showed that this statement is acceptable (Girard, 1992). 2.5. Radiative measurements

In characterizing the radiative conditions in CBFS,

the main concern is to assess the quantity of transmitted PAR, the part of the solar spectrum used by the chloroplasts for photosynthetic conversion. The transmitted PAR under the 5, 20 and 40 year old coconut stands was measured by a set of quantum sensors at the soil level in the absence of intercrops while the incident radiation was recorded by a reference sensor placed above the canopy or in open ground. The sensors used for measuring the transmitted radiation were manufactured by the CIRAD because of their low cost as compared to the commercial sensors. They were made with amorphous silicon cells, so-called SLAMs, which have a spectral sensitivity within the 400-700 nm waveband, though a very slight overlapping with adjacent wavebands (UV and NIR) occurs. The cells were equipped with a precision resistor. The whole was sprayed with a waterproof varnish and inserted in a black case made within a nylon bar. A white cover acting as a diffuser was placed above the cell. A commercial sensor was used as a reference for the calibration of these sensors.

93

To determine not only the mean transmission rates but also to map the distribution of light, the protocol called for the use of 32 SLAM sensors placed in two adjacent elementary triangles n. The sensors were connected to a Delta-T logger programmed to read the signals every 5 s and to integrate them every 5 min. Each of the three stands had at least 3 days of continuous logging from approximately 0600 until 1700 h. 2.6. Simulation of radiative climate under coconut stands

The software developed split the sky hemisphere in 46 sectors according to the Den Dulk's 'TURTLE' model (Den Dulk, 1989). The quantity of PAR incoming from each direction is calculated in two steps: •

•

First direct and diffuse components of global radiation are calculated from the ratio of global radiation on extra-terrestrial radiation using de Jong formulas (cited by Spitters et al., 1986). The distribution of diffuse radiation within the 46 sectors is then calculated by combining the formulas of Dogniaux (1973) describing the brightness of a clear sky and of Anderson (1966) describing the brightness of a standard overcast sky. Instantaneous direct radiation is distributed into three neighboring sectors according to the spherical distance of their center to the sun direction.

The sector brightness can be calculated for a given moment or integrated on several hours or several days with a 30 mn time step. Radiative transfers within the canopy are simulated by two programs developed at the CIRAD Plant Modeling Unit (to be published): for each of the 46 sectors defined: •

The MIR program calculates the interception of the incident radiation by the vegetation elements and by the soil. Light interception can be output for each canopy element (e.g. each leaflet), for element classes (e.g. for fronds according to their rank) or for indivi-

. ,

i An elementaryMangle is the space delimitedby three neighboring trees.

•

dual plant. A map of the radiation reaching the soil can be obtained (Fig. 5). The MUSC program calculates the multiple scattering of the intercepted light at different levels of the canopy, i.e. the mutual lighting between the soil and horizontal layers of vegetation. Resulting light interception is output for each layer defined by the user and for the soil.

These two complementary programs provide a detailed radiative balance of the canopy illuminated from each direction. The RADBAL module then combines the results obtained for the different directions according to the instantaneous or integrated sector brightness. Owing to this procedure, the directional radiative exchanges have to be calculated only once and the total radiative balance can then be obtained rapidly for any radiative condition (Rapidel, 1995).

3. Results 3.1. Radiative simulations 3.1.1. In situ measurement and validation of radiative simulations The measured rate of light transmission substantially varies with age (Fig. 6). This results from a crown development (frond length, declination and number of fronds) increasing from early stages to reach its maximum on the 15th year and decreasing gradually beyond 30 years. Simulations have been run for coconut stands with bare soil, i.e. without intercrop. In the absence of data about soil and coconut leaf optical properties in the PAR domain, plausible values have been tested:

•

10, 20 and 25% for the leaf reflection and transmission coefficients (assumed to be equal)

•

5, 10 and 15% for the soil reflection coefficient.

Simulations showed that PAR at soil level depends very slightly on the chosen optical properties (Dauzat and Eroy, 1995); actually, the light fraction impinging on the soil without interception by the vegetation is much more important than the fraction scattered by

94

field data (Fig. 6, left). The diurnal evolution of PAR transmission is also correctly simulated (Fig. 6, right). Likewise, the radiative model was able to simulate the transmitted light in the complex situation of the DRC experiment (Fig. 7). The simulated values for the different treatments are quite close to the measured ones. All these results are satisfactory enough to consider that both the coconut models and the radiative models are valid and can be used for radiative simulation experiments.

50 v

c

.o .~

E

40 ! 30

m r-

20

"" IZ

10

< n

0 40

20

5

age groups measured

3.1.2. Radiative simulation experiments PAR transmission under 20 and 40 year old coconut stands has been simulated on wide range of tree density and with different planting patterns as presented in Section 2.4. For a given age, the PAR transmission is linearly related to coconut density, irrespective of the planting design (Fig. 8). The regressions of %PAR transmission vs. density give:

l l simulated

600 X.

X.

X.

'7

ffl

o

x

400t

... for 20 year old stands /X

E

=L

%PAR = 79.97-0.386 density (R 2 = 0.996)

200

%PAR=85.79-0.312 density (R 2 = 0.984)

IZ

x x.

0

. !

5

7

.

.

.

.

.

... for 40 year old stands

x

r

,

l

!

9

11

13

15

17

time of day (h) x

measured

......... simulated

Fig. 6. Upper, Comparisonof simulatedand measuredPAR transmission under coconuts for different age groups. Lower, Daily evolution of the PAR transmitted under a 20 year old coconut stand.

the vegetation toward the soil. Thus we chose arbitrarily a reflection-transmission coefficient of 20% for the leaves and a reflection coefficient of 10% for the soil. One could expect that the presence of an intercrop with an albedo higher than the albedo of the soil would modify the radiative balance of the canopy. In practice this modification is negligible: the presence of an intercrop with an albedo of 20% would barely increase the downward PAR radiation under coconuts by about 1%. The PAR transmission has been simulated for 5, 20 and 40 year old observed stands on the whole period of in situ radiative measurements. The average simulated PAR transmission is in good agreement with

It appears that removing some fronds of the trees (to limit their number at 18) can increase the light penetration by about 25-40%. As a result, the quantities of PAR transmitted under 20 year old pruned stands with densities of 143 and 156 palms/ha are more or less comparable to those obtained under unpruned palms having densities of 95 and 100 palms/ha. Enhancement of light transmission is less marked for 40 year old than for 20 year old coconuts.

3.2. Corn and mungbean yield predictions The corn and mungbean yields largely differed from one intercropping campaign to another (Fig. 9). This can be mainly attributed to the varieties grown and, to a lesser extent, to temporary water logging (BEnard et al., 1996). Besides these differences, it can be noticed that the yield response for both crops is more or less a linear function of the PAR received. In order to simulate the expected yields of corn and mungbean grown under coconuts at different densi-

95

ties, we simulated first the PAR available under coconuts. The yields were then obtained by interpolation using the experimental yield responses. For corn under 20 year old unpruned coconuts, decreasing the tree density to 50%, may mean doubling or tripling the yield (Fig. 10). The yield response of mungbean to the density is proportionally smaller and varies among the campaigns due to the shade tolerance of the varieties (Fig. 11). The pruning of the palms has a drastic effect on the corn and mungbean yield. Globally, the density and pruning effects 0

@

,L'~

©

•

60

OC:: 50 tx . Q 2 5 0 0 qJ "' 2000-~ w

I

o.........

" Si

I

..... "" b) deep, 1/3~/

, .

,

i

i

'i'

i

1

i

i

----i i 'i I I

....

c) shallow, 111

1500--I

~/- m ....

2.5. Simulations ' 2

I

I

e

a)deep.11

,,1,

m

/

° .."°"

10001"! / s . ..... s

The simulations were performed for 20 years of daily weather data of Cordoba (Spain). We assumed a wheat-sunflower rotation, 7.1 sunflower plants/m 2 and rainfed conditions, on a medium texture soil. Combination of three season lengths and seven levels of EV were tested on two sowing dates (1 January and

]I

....... s

500....... 0

J

i 4

i 6

i 8

,

i i 101214

t

2 4 6 8 101214

EARLYVIGOUR(g.plant"1) Fig. 3. Relationships between average seed yield and average EV of sunflowerat Cordoba (Spain) in the simulations. Deep soil (a,b), shallow soil (c,d). Sowing date Ill (a,c) and 1/3 (b,d). Dotted line, short season genotypes; dashed line, medium season genotypcs; solid line, long season genotypes.

150

2.4

E

2.0

~

m

1.6

; ...........

'

1.2 . ...." a) p, 1/1~ 0.8 , m m , I l

'"~ 2.4 2.0

I

i ....;" i

c) shallow, 1/1 I

d) shallow, 113!

//~m

I

1.6

1.2 0.8

//

...

S

f_-....'..S;...... ' 1

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

, i i l i I'" i' 'l i i I 2 4 6 8 10 12 14 2 4 6 8 101214 EARLY VlGOUR (g.plant -1) i

Fig. 4. Relationships between average WUE (dry matter at maturity/ET from emergence to maturity) and average EV of sunflower crop at Cordoba (Spain) in the simulations. Deep soil (a,b), shallow soil (c,d). Sowing date 1/1 (a,c) and 1/3 (b,d). Dotted line, short season genotypes; dashed line, medium season genotypes; solid line, long season genotypes.

similar to that of the early sowing date, but the upper yield was lower at 2500 kg/ha. On the shallow soil, at the first sowing date, yield ranged from 750 to 1500 kg/ha. At the second sowing date yield was constant (ca. 500 kg/ha). The response of yield to EV was more complex in the case of long season cultivars. For undroughted conditions, as is the case of deep soil and first sowing date, an intermediate EV resulted in a yield similar to a high EV (2600 kg/ha). For the other cases the response of yield to EV was very small. On the deep soil at the second sowing date yield was ca. 1250 kg/ha. On the shallow soil, yield was ca. 500 and ca. 250 kg/ha for the first and the second sowing dates, respectively.

3.2. Water-use efficiency Fig. 4 shows the relationships between average WUE in terms of above ground dry matter at maturity, and average EV. In all cases, for a given value of EV, longest season genotypes had the highest value of WUE, and for a given genotype season length, the WUE increased with EV. Furthermore, the early sowing date showed a higher WUE than the late sowing date. On the deep soil and at first sowing date, WUE ranges were 1.00-1.40, 1.40-2.00 and 1.70-2.30 g/ dm 3 for short, medium and long season length genotypes, respectively. These values were smaller at the

second sowing date (0.90-1.20, 1.20-1.80 and 1.602.00 g/dm 3 for short, medium and long season genotypes, respectively). On the shallow soil the ranges in WUE decreased from the first (0.90-1.30, 1.20-1.80 and 1.40-2.20 g/dm 3) to the second sowing date (0.80-1.20, 1.20-1.55 and 1.20-1.60 g/dm 3) for short, medium and long season length genotypes. To investigate the behaviour of WUE, the relationships between the transpiration efficiency (TE; dry matter at maturity/transpiration from sowing to maturity) and EV, and the relationships between the transpiration/ET (T/ET) ratio and EV were obtained. Fig. 5 shows that high TE was related to a high EV, except for short season length genotypes, which showed a small response to increased EV. Taking into account the equation given by Tanner and Sinclair (1983), which relates dry matter production (B) and transpiration (T): B - K x T/VPD (where K is a coefficient depending of species and environmental conditions, and VPD is the vapour pressure deficit), is easy to understand that early planting date increased TE in all the cases studied due to the lower VPD values during its growing period. Ranges of TE were similar in both soil depths for a given combination of sowing date and genotype season length. For short season genotypes at the first sowing date, the range in TE was ca. 2.80-2.50 g/dm 3, while at the second sowing date it showed a constant value of ,~ 3.6 'E 3.2 ~

¢0

~ >oz w iT', II i,i z 0

b) deep, 1/3

t

2.8 2.4

/"-ms

2.0 i

a) deep, 111~ i w i i !

36 3.2-. ~ , - - - m

;

I

I

I

I

" i t

==. 2.8-. // n ":L.. . . . . . . . . s (n 2.4..... z ~; 2.0- c)shallow, 1/1

2

I

1'o1'214

d) shallow, 113 ,,

_tl-- 1 I

i

i

s l

i

I

2 4 6 8 101214

EARLY VIGOUR (g-plant "1) Fig. 5. Relationships between average TE (dry matter at maturity/T from emergence to maturity) and average EV of sunflower crop at Cordoba (Spain) in the simulations. Deep soil (a,b), shallow soil (c,d). Sowing date l/l (a,c) and 1/3 (b,d). Dotted line, short season genotypes; dashed line, medium season genotypes; solid line, long season genotypes.

151

o.7-1

,I

°.fi'.. ..........'

/

,

./

0.4

-"" I": a) deeP' 1/11 0.3, i i i r i I

0.7 ~

0.3

i

b) deep, ~ i

I

I

, 1'I

'"C) shallOW' 1i

o.6o.~: 0.4£

/ I

...."" ......s/

.., .Y" ...... ..r""~

I

soil (ca. 0.30 to ca. 0.35 for both sowing dates), whereas in the shallow soil this response was very small at the first sowing date (ca. 0.24 for all values of EV) and negative at the second sowing date (0.260.12). Long season genotypes showed a negative response of HI to EV on the deep soil (0.32-0.26 and 0.30-0.20 at the first and second sowing dates, respectively), and a very small response in the shallow soil (0.14-0.12 and 0.10-0.08 at the first and second sowing dates, respectively).

........s

, , l , 2 4 6 8 101214

I

,

I

,

,

,

I

2 4 6 8 101214

4. Discussion

EARLY VIGOUR (g.plant -1) Fig. 6. Relationships between average T/ET ratio and average EV of sunflower crop at Cordoba (Spain) in the simulations. Deep soil (a,b), shallow soil (c,d). Sowing date 1/1 (a,c) and I/3 (b,d). Dotted line, short season genotypes; dashed line, medium season genotypes; solid line, long season genotypes.

ca. 2.20 g/dm 3. Medium season length genotypes showed ranges of ca. 2.40-3.10 and ca. 2.00-2.60 g/dm 3 at the first and second sowing dates, respectively. Long season genotypes TE's ranges were ca. 2.60-3.40 and ca. 2.20-2.65 g/dm 3 at the first and second sowing dates, respectively. The T/ET ratio (Fig. 6) showed a positive response to increased EV and genotype season length, indicating that high EV contributes to reduced soil evaporation, although the impact decreased as season length and thus, seasonal ET increased. Medium and late genotypes showed similar values in all cases, ranging from ca. 0.50 to ca. 0.67. Short season genotypes presented a wider range in T/ET ratios with values from ca. 0.32 to ca. 0.57. For all genotypes the T/ET ratio was higher in the second than in the first sowing (Fig. 6). 3.3. Harvest index

Relationships between average HI and average EV for all the conditions studied are shown in Fig. 7. Short season genotypes had a positive response of HI to increased EV for all the cases, except for the shallow soil and late sowing date, where the response was very small. On the deep soil, HI ranges were 0.25-0.35 and 0.28-0.36 at the first and second sowing dates, respectively. Medium season length genotypes showed a positive response of HI to EV for deep

The fact that above ground dry matter was positively correlated with root dry matter means that selection for high EV will not be associated with indirect selection for a poor rooting system. Differences in these relationships between the two pot experiments may be caused by differences in the harvest dates. An increase in the above ground dry matter/root dry matter ratio related with the development of the sunflower plant has been observed by Trapani et al. (1992). Regression coefficients in the linear regressions between season length and stem volume were very low and for every value of stem volume a broad range of season length was observed. Therefore, the shortening of season length produced by the increase 0.36 0.32~

/--.-'~.--" m I

m/'--.~--~.7 .....

o.24-t "-

0.20-] 0.16-1 0.12d 0.08~

i

I

a) deep, 1111 i I I I

0.28-1 .............s 0.24-1 ...."~ . . . . 0.20-1 Z 0.16-]

I

i

b) deep, 113I I I I I

I

!

0"08/ , i i i 'i J 2 4 6 8 101214

-!

"---x

m I

I J, i :--1 ! r 1 2 4 6 8 101214

EARLY VIGOUR (g.plant "1) Fig. 7. Relationships between average HI and average EV of sunflower at Cordoba (Spain) in the simulations. Deep soil (a,b), shallow soil (c,d). Sowing date 1/l (a,c) and 1/3 (b,d). Dotted line, short season genotypes; dashed line, medium season genotypes; solid line, long season genotypes.

152

of EV was very small and a selection for EV should not be related to a reduction in season length. These results are not in agreement with those of Gimenez and Fereres (1986), who found than EV was inversely correlated with season length, although they used sunflower genotypes with a narrow genetic base. A similar result to Gimenez and Fereres (1986) was found for Ceccarelli et al. (1991) with different oat genotypes. However, Regan et al. (1992) did not find a similar relation in wheat. In all the cases in the simulations, maximum yield corresponded to highest EV, independently of season length. The range of WUE values in these experiments (0.80-2.3 g/dm 3) was wider than that found by Soriano et al. (1994) for sunflower under different irrigation treatments: 1.77-2.85 g/dm 3, or by Orgaz et al. (1990): 1.16-1.36 g/dm 3, and Gimeno et al. (1989): 1.4-2.5 g/dm 3, at winter and spring sowing dates. The increase in WUE in response to EV and season length was due to a combined improvement of TE and the T/ET ratio. The higher TE of high EV plants is related to the ability to grow faster under low temperatures when VPD is low (Tanner and Sinclair, 1983), as is the case in early sowing dates. Higher VPD following late sowing dates reduces differences among plants differing in EV. Higher TE values in short season genotypes with low EV is not evident when expressed in terms of glucose instead of dry matter (data not shown). High dry matter early in plant development improves the T/ET ratio as a larger ground area is covered by the crop, hindering the passage of radiation to the soil and decreasing E from the soil surface when there is a high probability of the soil being wet by rainfall. Gregory et al. (1984) and Cooper et al. (1987) observed, in cereals, an increase in yield associated with a faster early growth and reduced evaporation from the soil surface. For a given EV, lower T/ET values of short season genotypes may be due to the seasonal change in this ratio which is low at the beginning of the growing season and increases thereafter. Thus, short season genotypes have a shorter time of high T/ET values and the final value is lower than that of long season genotypes. Similar effects to those observed with high EV may be achieved with crop management techniques. For example with early sowing dates, crops will grow under low VPD and TE will increase, although this

early sowing date is limited by slow germination rates at low temperatures, which may cause the incidence of diseases (Gimeno et al., 1989). The T/ET ratio may be improved by increasing fertilizer application, which will have a positive effect on crop growth, or by increasing planting density (Anderson, 1992). With all these techniques, adequate season length cultivar should be selected to avoid water stress during seed filling due to a quick use of water in the vegetative period. Under good conditions of water supply (deep soil, early planting) our analysis shows large differences in yield between short (1500 kg/ha) and medium or long season cultivars (2600-2900 kg/ha). This is partly the result of assuming the same planting density for all the genotypes. Villalobos et al. (1994) have shown that short season cultivars under irrigation respond to planting density up to 12 plants/ m 2. Thus, the interaction between planting density and season length may also play an important role in the performance of sunflower under the studied conditions. Results of simulations show that it is important to take into account the combination of season length of the genotype and EV to ensure water supply to the seed during the filling period. Reductions in yield as a response to increased EV were related to reductions in HI. Moreover, in the case of short season genotypes, yield is determined by water supply (soil depth), being little affected by sowing date. Decreases in sunflower HI due to water stress during the seed filling period have been observed by Gimeno et al. (1989) and Orgaz et al. (1990), and in simulations by Sadras and Villalobos (1994). They found a decrease in HI on shallow soils when genotype season length increased but no response on deep soils, which is in agreement with our results. Our analysis indicates that the optimum strategy to obtain maximum seed yield for the conditions of this study would be to use medium season length cultivars with high EV on a deep soil. Similar cultivars would be the optimum strategy on a shallow soil at the first sowing date. A different rotation to the wheat-sunflower one assumed in the simulations would change this optimum strategy. For example, a sunflower-sunflower rotation would cause a higher water use than a sunflower-wheat rotation, and short season genotypes would be the optimum choice more frequently.

153

5. Conclusions High EV in sunflower populations is not related to a low dry matter accumulation in the root. Moreover, the negative association between EV and season length does not imply substantial differences between plants with high and low EV. We can find plants with high EV and a determined season length. Thus, selection for high EV in a sunflower breeding program to improve seed yield will not be correlated with indirect selection for a poor rooting system or a short season length. Using a modified version of O I L C R O P - S U N to analyze the effect of EV on sunflower yield under different situations we concluded that high EV is a positive character. Higher yield was related to higher EV for every environment studied as a result of increases in WUE, the T/ET ratio and TE. The optim u m strategy would be to use the genotype with a season length adequate for a given environment and with high EV. In the cases studied, m e d i u m season length genotypes produce the m a x i m u m yield, except for a late sowing date on shallow soil, when short genotypes performed the best. This analysis may be extended to determine the o p t i m u m sunflower season length at any environment taking into account soil and weather data. The physiological bases of differences in EV remain u n k n o w n and deserve further research. Future work should also explore the effect of rotations and fertilizer m a n a g e m e n t on the performance of sunflower genotypes differing in EV and season length. The analysis may be performed by linking the simulation model with a Geographic Information System to define the ideotypes for different areas within a region.

Acknowledgements The authors are grateful to nez for his c o m m e n t s on the held a pre-doctoral fellowship de Investigaciones Cientfficas

J.M. Fernandez-Martimanuscript. F. Agtiera from Consejo Superior (Spain).

References Anderson, W.K., 1992. Increasing grain yield and water use of

wheat in a rainfed Mediterranean type environment. Aust. J. Agric. Res., 43:1 - 17. Austin, R.B., 1982. Crop characteristics and the potential yield of wheat. J. Agr. Sci., 98: 447-453. Austin, R.B., 1993. Augmenting yield-based selection. In: M.D. Hayward, N.O. Bosemark and I. Romagosa (Editors), Plant Breeding: Principles and prospects. Chapman and Hall, London, pp. 391-405. Boujghagh, M., 1994. Varibilite genetique des cultivars de tournesol en semis d'hiver dans la region du sais-fes. Helia, 20: 6780. Boukerrou, L. and Rasmusson, D.D., 1990. Breeding for high biomass yield in spring barley. Crop Sci., 30: 31-35. Ceccarelli, S., Acevedo, E. and Grando, S., 1991. Breeding for yield stability in unpredictable environments: single traits, interaction between traits, and architecture of genotypes. Euphytica, 56: 169-185. Cooper, P.J.M., Keatinge, J.D.H. and Gughes, G., 1983. Crop evapotranspiration -a technique for calculation of its components by field measurements. Field Crops Res., 7: 299-312. Cooper, P.J.M., Gregory, P.J., Keatinge, J.D.H. and Brown, S.C., 1987. Effects of fertilizer, variety and location on barley production under rainfed conditions in Northern Syria. II. Soil water dynamics and crop water use. Field Crops Res., 16: 67-84. Donald, C.M., 1968. The breeding of crop ideotypes. Euphytica, 17: 385-403. Elliot, G.A. and Regan, K.L., 1993. Use of reflectance measurements to estimate early cereals biomass production on sandplain soils. Aust. J. Exp. Agric., 33: 179-183. Fereres, E., Orgaz, F. and Villalobos, F.J., 1993. Water use efficiency in sustainable agricultural systems. In: D.R. Buxton, R. Shibles, R.A. Forsberg, B.L. Blad, K.H. Asay, G.M. Paulsen and R.F. Wilson (Editors), International Crop Science. Crop Science Society of America, Inc., Madison, WI, pp. 83-89. Fernandez-Martinez, J.M., Dominguez, J., Gimenez, C. and Fereres, E., 1990. Registration of three sunflower high-oil nonrestorer germplasm populations. Crop Sci., 30: 965. Gardner, C.O., 1961. An evaluation of effects of mass selection and seed irradiation with thermal neutrons on yield of corn. Crop Sci., l: 241-245. Gimenez, C. and Fereres, E., 1986. Genetic variability in sunflower cultivars under drought. II. Growth and water relations. Aust. J. Agric. Res., 37: 583-597. Gimeno, V., Fernandez-Martinez, J.M. and Fereres, E., 1989. Winter planting as a mean of drought escape in sunflower. Field Crops Res., 22: 307-316. Gregory, P.J., Shepherd, K.D. and Cooper, P.J.M., 1984. Effects of fertilizer on root growth and water use of barley in Northern Syria. J. Agric. Sci. Camb., 103: 429-438. Loomis, R.S., Rabbinge, R. and Ng, E., 1979. Explanatory models in crop physiology. Annu. Rev. Plant Physiol., 30: 339-367. Loss, S.P. and Siddique, K.H.M., 1994. Morphological and physiological traits associated with wheat increases in Mediterranean environments. Adv. Agron., 52: 229-276. Ludlow, M.M. and Muchow, R.C., 1990. Acritical evaluation of traits for improving crop yields in water-limited environments. Adv. Agron., 43: 107-153.

154

Orgaz, F., Gimenez, C. and Fereres, E., 1990. Efficiency of water use in winter plantings of sunflower in a Mediterranean climate. In: A. Scaife (Editor), Proceedings of the First Congress of the European Society of Agronomy, Paris, France, pp. 13-14. Passioura, J.B., 1977. Grain yield, harvest index and water-use of wheat. J. Aust. Inst. Agric. Sci., 43: 117-120. Regan, K.L., Siddique, K.H.M., Turner, N.C. and Whan, B.R., 1992. Potential for increasing early vigour and total biomass in spring wheat. II. Characteristics associated with early vigour. Aust. J. Agric. Res., 43: 541-553. Richards, R.A., Lopez-Castafieda, C., Gomez-Macpherson, H. and Condon, A.G., 1993. Improving the efficiency of water use by plant breeding and molecular biology. Irrig. Sci., 14: 93104. Sadras, V.O. and Villalobos, F.J., 1994. Physiological characteristics related to yield improvement in sunflower (Helianthus annuus, L.). In: G.A. Slafer (Editor), Genetic Improvement of Field Crops. Marcel Dekker, New York, pp. 287-320. Soriano, A., Villalobos, F.J. and Fereres, E., 1994. Response of sunflower grain yield to water stress applied under different phenological stages. In: M. Borin and M. Sattin (Editors), Proceedings of the Third Congress of the European Society of Agronomy, Abano-Padova, Italy, pp. 246-247. Tanner, C.B. and Sinclair, T.R., 1983. Efficient water use in crop

production: research or re-search?. In: H.M. Taylor, W.R. Jordan and T.R. Sinclair (Editors), Limitations to Efficient Water Use in Crop Production, Am. Soc. Agron., Madison, Wl. pp. 1-27. Trapani, N., Hall, A.J., Sadras, V.O. and Vilella, F., 1992. Ontogenic changes in radiation use efficiency of sunflower (Helianthus annuus, L.) crops. Field Crops Res., 29: 301-316. Turner, N.C. and Nicolas, M.E., 1987. Drought resistance of wheat for light-textured soils in a Mediterranean climate. In: J.P. Srivastava, E. Porceddu, E. Acevedo, and S. Varma (Editors), Drought Tolerance in Winter Cereals. Wiley, New York, pp. 203-216. Villalobos, F.J. and Ritchie, J.T., 1992. The effect of the temperature on leaf emergence rates of sunflower genotypes. Field Crops Res., 29: 37-46. Villalobos, F.J., Sadras, V.O., Soriano, A. and Fereres, E., 1994. Planting density effects on dry matter partitioning and productivity of sunflower genotypes. Field Crops Res., 36: 1-11. Villalobos, F.J., Hall, A.J., Ritchie, J.T. and Orgaz, F., 1996. OILCROP-SUN: a development, growth, and yield model of the sunflower crop. Agron. J., 88: 403-415. Whan, B.R., Carlton, G.P. and Anderson, W.K., 1991. Potential for increasing early vigour and total biomass in spring wheat. I. Identification of genetics improvements. Aust. J. Agric. Res., 42:347-361.


155

Options for breeding for greater maize yields in the tropics A. Elings*, J.W. White, G.O. Edmeades The International Centerfor Maize and Wheat Improvement (CIMMYT), Lisboa 27, Apdo. Postal 6-641, 00660 Mexico, D.F, Mexico

Accepted 13 April 1997

Abstract

Options for breeding for greater maize yields in the tropics were quantitatively examined with a crop growth simulation model that was tested against field data of five genotypes in four environments. Simulations indicate that at high production levels, grain filling of maize is sink-limited, and that increasing the number of kernels per m2 through larger primary ears, prolificacy or greater plant densities, will lead to increased grain yields. On a theoretical basis, it is concluded that larger primary ears lead to greater grain yields at all plant growth rates, and that increased prolificacy leads to greater grain yields only if plant growth rate exceeds a threshold. Under nitrogen limited growing conditions, selecting for genotypes that extract more nitrogen from soils, and for delayed leaf senescence, show promise for increasing yields. For crop growth limited by moisture availability around flowering, continued selection for improved kernel set leads to greater grain yields. © 1997 Elsevier Science B.V. Keywords: Model application" Nitrogen availability; Prolificacy; Sink-source relationships; Tropical maize; Water availability; Zea mays L

1. Introduction

Average grain yields of maize (Zea mays L.), which is grown extensively in the tropical and sub-tropical environments of the developing world, have been rising since 1960; from 1.5 to 3.1 t ha -~ in West Asia and North Africa; from 1.2 to 2.8 t ha -1 in South, East and Southeast Asia, and from 1.2 to 2.0 t ha -~ in Latin America (CIMMYT, 1990). Major reasons for increased grain yields are the greater use of fertilizer and improved germplasm. Yield increase in subSaharan Africa was relatively low, viz. from 0.9 to 1.2 t ha -~. Reasons are the limited use of fertilizer *Corresponding author. Tel.: +52 5 7269091; fax: +52 5 7267559; e-mail [email protected]

and improved germplasm, and intercropping, which may have disguised greater increases in land productivity (CIMMYT, 1990). These averages hide wide variation; for example, average grain yields in marginal areas of Central America are 0.3-0.7 t ha -~ (Brizuela et al., 1993). Average experimental yields in CIMMYT's international networks varied from 4.4 t ha -~ for early tropical germplasm to 6.5 t ha -~ for late subtropical germplasm (CIMMYT, 1994), and at CIMMYT's own experimental fields 14.0 t ha -I has been obtained. There exists, therefore, a wide gap between actual and experimental yields. As demand for food will continue to rise (Rosegrant et al., 1995), actual yield levels must be increased. Increases in grain yield have generally been realized by combined adoption of new cultivars and crop management


156

methods. Whereas crop management concentrates on influencing the environmental conditions, breeding tries to modify the crop phenotype in response to the environment. To benefit optimally from improved environmental conditions, one or more matching cultivars are required, which will result in increased resource use efficiency (de Wit, 1992), expressed as kg grain per unit input, be it solar radiation, nutrients, or water. Five broad levels of agricultural production can be distinguished (Penning de Vries and van Laar, 1982). At the potential production level, growth occurs under conditions of ample supply of water and nutrients, and is determined by radiation and temperatures in interaction with the morpho-physiological make-up of the plant. Low water, soil nitrogen and perhaps phosphorus and zinc availability can limit growth for at least part of the growing season, and define production levels two, three and four, respectively. Together, they determine the attainable production level, found on around 95% of the tropical maize growing area (Edmeades et al., 1997). Actual production may be further reduced if growth is limited by the presence of weeds, pests, diseases or pollutants. Breeding for increased grain yields in tropical maize in favorable environments has so far concentrated on reduction of plant height, resulting in increased harvest index, improvement of disease and insect resistance, and reduced lodging (Dowswell et al., 1996). For environments with water and nitrogen deficits, breeding has increased partitioning to the ear, which is associated with number of ears per plant and the anthesis-silking interval (ASI), and which has resulted in increased number of grains per m E (Edmeades et al., 1997). For environments with nitrogen deficit, breeding has also reduced senescence rate (Lafitte and Edmeades, 1994a). In this paper, we quantify increases in grain yield of tropical maizes that theoretically can be expected through alteration of certain crop characteristics. The suggested selection criteria are not new, but we attempt, nonetheless, to make explicit possible selection gains. The study does not try to be exhaustive and concentrates on breeding prospects at three production levels, viz. at a high production level (cf. Boote and Tollenaar, 1994) and under nitrogen and waterlimited growing conditions, each illustrated by the performance of five cultivars.

2. Materials and methods 2.1. Field evaluations

Five maize cultivars adapted to the lowland tropics were evaluated to quantify growth characteristics. Across 8328 BN C6 is a yellow dent, late open-pollinated variety (OPV), that was selected under conditions of low soil nitrogen availability and gives relatively high grain yields under these conditions. CML247 × CML254 is a white dent, late hybrid, that is currently one of the highest yielding lowland tropical materials available from CIMMYT. Pool 16 C20 is a white dent, early maturing open-pollinated population. La Posta Sequfa C4 is a white dent, late OPV, that was selected under conditions of limited water availability. PR 8330 is a white dent, early OPV. Evaluations were conducted at CIMMYT's experiment stations near Poza Rica (PR) and Tlaltizap~in (TL). PR is located in a lowland tropical climate at 60 m altitude near the Gulf of Mexico (20°32'N, 97°26'W), and TL is located in a mid-altitude tropical climate at 940 m altitude in the central mountains of Mexico (18°4 I'N, 99°08'W). Soils are classified as sandy loam Tropofluvent (Entisol) at PR and as clay Pellustert (Vertisol) at TL. Evaluations took place during the rainy summer season of 1995 at PR and TL and during the dry winter season of 1996 at TL. Smallest, average and largest values of minimum and maximum temperatures and incoming solar radiation are given in Table 1. Treatments at PR differed primarily in soil nitrogen availability, which resulted in low, medium and high levels of leaf nitrogen concentration. The low and medium nitrogen treatments were conducted on fields that had not received chemical nitrogen for 9 and 6 years, respectively, and the high nitrogen treatment received 75 kg N ha -~ at sowing and 125 kg N ha -1 at 1 month after sowing. The 1995 experiment at TL received 150 kg N ha -! at sowing and 50 kg N ha -l at anthesis. Experiments at TL during the dry season of 1996 received 150 kg N ha -l at sowing and were conducted at three levels of moisture availability. At the 'low moisture' level irrigations were halted about 4 weeks before anthesis and resumed about mid-way through grain filling; at the 'medium moisture' level, irrigations were halted about 1 week before anthesis and not resumed; and at the 'adequate moisture' level,

157

efficiency during vegetative growth (RUEveg, in g MJ -I) were calculated as total above ground biomass at final harvest and anthesis, respectively, divided by amounts of intercepted PAR from emergence to harvest, and from emergence to anthesis, respectively. To account for variation in RUEveg due to leaf nitrogen content, RUEveg can be related to the amount of leaf nitrogen (NL; kg N ha -~ ground surface area) (Sinclair and Horie, 1989; ten Berge et al., 1994). For PR, this relation was described by fitting

irrigation was applied as needed to ensure near-optimal growth, based on long-term field experience. Genotypes were sown in three replications in randomized complete blocks, using eight-row plots of 10.5 m length at a density of 5.33 plants m -2. Row width was 0.75 m, and plant spacing within rows was 0.25 m. Eight sub-plots of ten plants each were harvested at intervals of 2-3 weeks, and after physiological maturity, a sub-plot of 30 plants was harvested. Dry weights of stems (including leaf sheaths), green leaf tissue, dead leaf tissue, panicles, ear husk leaves, ear cob and kernels, and leaf nitrogen concentration (g N g-i leaf tissue; micro-Kjeldahl) were determined at each harvest. Ten plants in the center of the sub-plot for final harvest were marked soon after emergence, and length and width of each fully expanded leaf were measured, and senesced leaves were identified, 1 day before or after a harvest. Areas of individual green leaves were calculated as length x width x 0.75 (Montgomery, 1911), from which total green leaf area (LAI; ha ha -l) was derived. Solar radiation above and below the canopy was measured on bright days every 2-3 weeks (Sunfleck Ceptometer, Decagon Devices, Inc.). Dates at which 50% of the 30 plants in the sub-plot for final harvest produced pollen (male flowering) and at which 50% of the plants showed exerted silks (female flowering) were recorded. ASI (d) was defined as the difference between these two flowering dates. Physiological maturity was defined as the date at which a black layer had formed at the base of the kernels. Intercepted photosynthetically active radiation (PARi, MJ m -2) was calculated with the crop growth model that is described below, using observed values of LAI as input. The model calculates total PAR as 50% of daily total solar irradiation. Seasonal radiation use efficiency (RUE~eas, in g MJ -l) and radiation use

RUEveg = RUEmax × (1 - e (-~xNL))

(1)

in which RUEmax is the maximum radiation use efficiency at high amounts of leaf nitrogen, and ~ is a coefficient. Because RUEveg in TL was substantially lower than in PR, TL data were not included in the regression analysis.

2.2. Simulation model The theoretical consequences for biomass and grain yield of altered crop characteristics were assessed with a maize crop growth simulation model that was based on SUCROS87 (Spitters et al., 1989). Total daily rate of canopy CO2 assimilation is calculated from the daily incoming radiation, temperature and LAI by integrating instantaneous CO2 assimilation. The light extinction coefficient (k) is an input parameter that can be varied. The rate of leaf photosynthesis at light saturation is strongly related to leaf nitrogen concentration (van Keulen and Seligman, 1987). Maintenance and growth respiration requirements are calculated on the basis of weights and chemical composition of plant organs (Penning de Vries et al., 1989). After subtraction of respiration requirements from gross assimilation, net daily growth rate is obtained. The dry matter produced is partitioned

Table 1 Smallest, average and largest values of minimumand maximumdaily temperatures,and daily solar irradiation, for the 1995 wet and 1996dry cycles at CIMMYT's experimental stations near Poza Rica (PR) and Tlaltizaplin (TL) Station

PR TL

Cycle

1995 wet 1995 wet 1996 dry

Minimum daily temperature (°C)

Maximum daily temperature (°C)

Daily radiation (MJ m-! day-I)

Min.

Ave.

Max.

Min.

Ave.

Max.

Min.

Ave.

Max.

17.8 7.0 3.0

22.4 17.0 I 1.3

25.0 20.0 24.0

26.0 27.0 18.0

32.9 30.6 32.2

36.0 33.4 38.2

7. l 14.9 3.4

20.0 27.7 25.9

28.8 34.1 36.8

158

Table 2 Total above-ground dry matter (DM, kg ha-I), grain yield (GY, kg ha-I), maximum leaf area index (LAImax, ha ha-l), leaf area index at maturity (LAImat, ha ha-~), kernels per plant (KPP), anthesis-silking interval (ASI, d), and the maximum content of leaf N during the crop cycle (NL.... ; kg N ha -~ ground surface area) of five cultivars during the 1995 wet and 1996 dry cycles at CIMMYT's experimental stations near Poza Rica (PR) and Tlaltizap~in (TL) Cycle

Growing condition

Character

La Posta Sequfa C4

Across 8328 BN C6

PR 8330

TL95 dry

Med. N

DM GY LAImax LAImat ASI KPP NL.max

12651 ~ 5286 a 3.74 0.27 0 391 43.10

11950 a 5108 a 2.95 0.40 0 388 44.12

10167 b 4610 a 2.51 0.54 0.3 361 37.86

Pool 16 C20

CML254 x CML247

Average

9654 b 4650 ~ 2.13 0.33 1.0 346 32.84

15387 6474 4.11 0.77 1.3 480 53.31

11955 5226 3.09 0.46 0.5 393 42.25

Levels of water availability TL96 dry

Low

DM GY LAImax LAlmat KPP ASI

7781 1995 3.68 0.91 205 1.7

-

-

4875 709 2.59 0.24 77 11.0

7226 781 4.17 1.54 66 20.3

6627 1162 3.48 0.90 116 11.0

TL96 dry

Intermediate

DM GY LAlma~ LAlmat KPP ASI

11603 ab 2928 ab 3.68 2.73 228 0.3

9104 b': 1822 a 3.67 1.90 155 4.3

7477 c 1734 a 2.66 1.79 142 3.0

7898 c 2446 ac 2.54 0.16 197 1.0

12508 a 3761 ~: 4.12 3.09 347 4.0

9718 2538 3.33 1.93 214 2.5

TL96 dry

Adequate

DM GY LAlmax LAImat KPP ASI

19294 a 6743 4.70 0.95 446 -0.7

-

-

14881 6650 3.06 0.01 436 0.3

20343 a 6544 5.46 1.99 429 0.7

18173 6646 4.41 0.98 437 0.1

6643 1485 2.60 0.61 1.3 150 17.41

6744 1920 2.41 0.65 2.7 184 21.94

6041 2013 2.04 0.36 2.7 182 25.31

5226 1690 2.11 0.22 3.7 159 25.18

6966 2197 2.83 0.39 1.0 240 25.45

6324 1861 2.40 0.45 2.3 183 23.06

8407 c 3535 c

8637 c 4520 ac

Levels of soil nitrogen availability PR95 wet

Low

DM GY LAImax LAImat ASI KPP

NL.max PR95 wet

Medium

DM GY LAImax LAImat ASI KPP NL....

10751 a 4849 ab 3.21 1.90 -0.7 362 39.85

12246 ab 5110 ab 3.10 1.87 0 367 39.01

2.31 0.73 1.0 266 27.08

2.25 0.74 0.7 342 28.31

13021 b 5837 b 3.63 2.08 0 410 37.13

10612 4770 2.90 1.46 0.2 349 34.28

159

Table 2 continued Cycle

Growing condition

Character

La Posta Sequfa C4

Across 8328 BN C6

PR 8330

Pool 16 C20

CML254 x CML247

Average

PR95 wet

High

DM GY LAImax LAI,,~t ASI KPP NL,max

13206 a 6044 a 3.53 2.32 -1.0 417 61.67

13474 a 5913 a 3.81 2.23 0 418 63.36

10355 b 4081 3.02 1.06 0 310 51.02

10559 b 5226 a 2.97 0.45 -0.7 371 57.68

16221 7787 4.22 2.65 -1.0 490 72.86

12771 5810 3.51 1.74 --0.5 401 61.32

Data from the same cycle with different superscripted letters differ at P < 0.05 (given only for weights).

among the various plant organs. Phenological development is tracked as a function of ambient daily average temperature. Before canopy closure, the leaf area increment is calculated from the daily average temperature, as carbohydrate production is assumed not limiting to leaf expansion. After canopy closure, the increase in leaf area is obtained from the increase in leaf weight. Integration of daily growth rates of the organs and leaf area results in dry weight increment during the growing season. Alternatively, leaf area can be made model input. The model contains an option to compute daily growth rate on the basis of leaf area index, PAR~, leaf nitrogen content, and RUE. Determination of kernels per plant (KPP) is based on work by Edmeades and Daynard (1979) and Tollenaar et al. (1992), who related KPP to plant growth rate around flowering. They indicated that at rates higher than about 6.5 g CH20 plant -1 day-l, a second ear forms in semi-prolific cultivars, and KPP is described by two intersecting hyperbolas. Important characters are the plant growth rate at which kernels begin forming on a second ear (Xi,t) and the number of kernels on the primary ear if kernels begin forming on a second ear (Yi,t). The full equations are given in Tollenaar et al. (1992). The parameters Xint and Yi,t are model inputs that can be varied, which permits simulation of semi-prolific cultivars. Maximum grain growth rate is computed from KPP and a grain filling rate of 8.5 mg kernel -~ day -l during the linear phase (derived from data in Lafitte and Edmeades, 1995). Potential grain growth rate is equal to the daily amount of dry matter available from vegetative tissue for grain growth. Actual grain growth rate is the minimum of potential and maximum grain growth rates.

Although water deficits at all times affect canopy development, they are especially critical around male and female flowering, when the ear has a relatively weak sink capacity (Schussler and Westgate, 1995), and low carbohydrate availability causes reduced ear growth and kernel abortion (Bassetti and Westgate, 1993), resulting in grain yield reduction. Reduced growth rate at flowering also causes increase of ASI, which is therefore a character associated with KPP. Bolafios and Edmeades (1993b) described the relation between ASI and KPP on the basis of well-watered and droughted experiments as: KPP =

e7.08-0.82 x (ASI+ 1.1)0.5

(2)

For a given value of ASI, the fractional reduction in kernels per plant can be estimated. Since a mechanistic approach was not possible, ASI was used as input in simulations, as a reflection of a short-term growth limitation at flowering which had no negative consequences for later crop growth.

3. Results 3.1. Field evaluations

Total above-ground biomass and grain yield of the five evaluated cultivars are presented in Table 2. Average biomass increased from 6627 to 18173 kg ha-l with increasing moisture availability (TL 1996) and from 6324 to 12771 kg ha -l with increasing soil nitrogen availability (PR 1995). Average grain yields increased from 1162 to 6646 kg ha-l, and from 1861 to 5810 kg ha -l, respectively. The late hybrid CML254 x CML247 gave greatest biomass and grain yields,

160

Table 3 Seasonal radiation use efficiency (RUE~as) and radiation use efficiency during vegetative growth (RUEveg) for four experiments conducted at Poza Rica (PR) and Tlaltizapdn (TL) in 1995 Cultivar

TL95

PR95 Low N

La Posta Sequfa C4 Across 8328 BN C6 PR 8330 Pool 16 C20 CML254 x CML247 Average

Medium N

High N

RUE~eas

RUEveg

RUEseas

RUE,,eg

RUE ....

RUEveg

RUEseas

RUEveg

1.52 1.56 1.58 1.63 1.60 1.58

1.84 2.13 1.92 1.99 2.16 2.01

1.61 1.66 1.58 1.45 1.45 1.55

2.38 2.64 2.84 2.60 2.09 2.51

1.99 2.33 1.95 2.00 1.99 2.05

3.06 2.83 2.78 2.47 2.53 2.73

2.37 2.38 2.10 2.24 2.25 2.27

3.48 3.13 3.10 3.28 3.26 3.25

except under water deficit, where the drought-adapted OPV La Posta Sequfa C4 gave greatest grain yield. The two early OPVs PR 8330 and Pool 16 C20 gave lowest biomass and grain yields. Total biomass and grain yields were associated with maximum LAI during the crop cycle (LAImax) and LAI at maturity (LmImat) (Table 2). ASI remained below 1 day at adequate soil moisture availability, and increased to 4.3 days for Across 8328 BN C6 at intermediate water availability, and to 20.3 days for CML254 ×CML247 at low water availability. ASI of the drought tolerant cultivar La Posta Sequfa was only 1.7 days at low water availability. ASI was zero or less at high nitrogen availability, and increased to average values of 0.2 and 2.3 days at medium and low nitrogen availability. Grain yield and KPP were linearly related: yield = -860 + 16.14 x KPP (r 2 = 0.97) for all experiments, and yield = -675 + 15.95 x KPP (r 2= 0.96) for TL 1995 only. Values of RUEseas at PR and TL in 1995 varied between 1.45 and 2.38 g MJ -~ (Table 3). RUEveg varied from 1.84 to 3.48 g MJ -~ among cultivars and environments and increased with increasing soil nitrogen availability. Fitting eqn (1) for experiments in PR in 1995 resulted in estimates of RUEmax of 3.26 g MJ -~, and of e of 0.1172 (Fig. 1). With the exception of the hybrid which had a relatively low RUEveg at low NL, there were few differences among cultivars. The light extinction coefficient (k) was determined by regressing fraction intercepted solar radiation (SR0 on LAI. Combined data of the five cultivars at PR and TL in 1995 could be approximated by

SR i =0.92 x (1 - e (-kxLAl))

(3)

with k = 0.53 (Fig. 2). The light extinction coefficient showed little variation among cultivars, environments and development stages. Variation in leaf nitrogen concentration among the five cultivars was low. Leaf nitrogen concentration of the cultivars in the high N treatment gradually decreased from 0.039 g g-~ at early vegetative stage to about 0.025 g g-I at maturity (Fig. 3). Leaf nitrogen concentration in the low N treatment decreased from 0.034 g g-l at the early vegetative stage to 0.015 g g-~ at flowering and to 0.011 g g-n at maturity. Cultivars in the medium N treatment and PR and in TL were characterized by intermediate levels of leaf nitrogen con3.5

00 •

3.0

"&

n

x

n

-'-' 2.5 x

2.0

~n

It i I I ! I

• La Posta Sequia C4 • Across 8328 BN C6 • PR8330 @ Pool 16C20 RUE,,~ = = CML254xCML247 3.26 X (1-e ('°1172xNL)) !""-'regression

1.5 1.0 0.5 0.0

0

20

40

60

80

leafN content(kg ha"t) Fig. I. Radiation use efficiency during the vegetative phase (RUEveg) in relationto leafN content (Nt.,kg N ha -t ground surface area) at Poza Rica (PR) and Tlaltizap~in(TL) for the 1995 wet summer season and five cultivars. The regression uses PR data only.

161

SR, = 0.92 x (1-e (°s3= uu))

0.9 0.8 0.7 ~" 0.6 =

0.5

~

0.4

•

La Posta Sequia C4

El Across BN C6

0.3

•

0.2

@ Pool 16 C20 X

0.1

PR 8330

CML247 x CML254

"--"regression

0

1

2

3

4

5

LA! (-) Fig. 2. Fraction intercepted solar irradiation (SR0 in relation to green leaf area index (LAI, ha ha -I) of five tropical cuitivars grown at three soil nitrogen levels at Poza Rica and one nitrogen level at Tlaltizapdn during the 1995 wet summer season.

centration. The maximum NL during the crop cycle (NL,max) varied from an average of 23.06 kg ha -i at low soil nitrogen availability to 61.32 kg ha-I at high soil nitrogen availability. Average NL,max at TL was 42.25 kg ha-I (Table 4). A large green leaf area was maintained after flowering at high production levels, which resulted in high potential grain filling rates. For example, simulation of grain filling processes of La Posta Sequfa C4 at high N availability in PR in 1995 shows that the potential grain growth rate was 250-300 kg ha-~ day-Z (Fig. 4). As the maximum grain growth rate was about 170 kg ha-~ day-~, the result was a sinklimited grain fill.

opment is fragmentary. Instead of rejecting the simulation model as a tool, a better approach incorporates that knowledge in data interpretation. Therefore, although simulations of crop and grain growth need improvement, the model was considered suitable to quantify expected changes in grain yield as a consequence of changes in crop characteristics. Focus was placed on genotypic and phenotypic variation, and less on the environmental variation that certainly exists over time and place in the lowland tropics. Therefore, simulations were carried out for all five cultivars, but only for the environmental conditions under which they had been evaluated.

3.3. Options to increase grain yield at high production levels 3.3.1. Radiation use efficiency A literature review by Kiniry et al. (1989) reported RUEveg ranging from 2.1 to 4.5 g MJ -l, with an average RUEveg of 3.5 g MJ -I. RUE in our own experiments varied from 1•84 to 3.48 gv~j-l among cultivars and environments, and increased with increasing soil nitrogen availability (Table 3). Crop growth at PR in 1995 was simulated using leaf nitrogen concentrations observed in the high N treatment during that season and using a high RUEveg of 4.5 g MJ -~. To reflect aging effects, RUE was assumed to decline linearly with accumulating thermal time units to 50% of its pre-anthesis value mid-way between anthesis and physiological maturity. Simulated biomass and grain 0.040 i~

PR, low N

1

3.2. Model evaluation The model is still in its developmental phase and was only calibrated and tested for the 1995 experiments, since leaf nitrogen data were not yet available for 1996, and a water balance routine was not yet in place. Total above-ground dry matter production greater than 10 t ha-1 was about 20% over-estimated, and grain yield was with about 800 kg ha-! over-estimated at low levels of soil nitrogen availability (Fig. 5). However, rankings of observed and simulated total biomass and grain yield within an experiment were similar. Crop growth simulation models are never fully accurate because our knowledge of the physiological processes underlying crop growth and devel-

17 im

I--ll--PR, medium N

0.035

}--~PR., hish N ~TL

M @

0.030 uM

u 0.025

8 M

~ o.o2o ~ o.o15 0.010 20

30

40

50

60

70

80

90

100

time (days after sowing) Fig. 3. Average leaf nitrogen concentrations of five tropical cultivars grown at three soil nitrogen levels at Poza Rica (PR) and at Tlaltizapdn (TL) during the 1995 wet summer season.

162

Table 4 Simulated increases in total above-ground dry matter production (DM, kg ha -I) and gr ain yield (GY, kg ha -!) of five cultivars during the 1995 wet summer cycle under various growing conditions at CIMMYT's experimental stations near Poza Rica (PR) and Tlaltizap~in (TL) in 1995 Cycle

Growing condition

La Posta Sequfa C4

Across PR 8330 8328 BN C6

Pool 16 C20

CML254 x CML247

Average

DM GY

17411 7921

15479 8272

16003 8289

12910 5244

15831 8628

15527 7671

DM DM DM DM DM DM DM DM DM DM DM DM

15450 15105 13703 5776 5666 5212 11140 10905 9995 12746 12512 11497

15750 15355 13793 6384 6253 5722 11409 11163 10230 13475 13236 12239

12398 12088 10831 5629 5514 5047 7921 7744 7084 11085 10853 9930

11807 11473 10168 5191 5081 4641 9355 9121 8248 11859 11581 10532

20526 20081 18388 7049 6927 6420 11658 11474 10679 17564 17269 16125

15186 14820 13377 6006 5888 5408 10297 10081 9247 13346 13090 12065

GY

7589

9064

7215

8044

13569

9096

Yint = 300, 400, 500, 600 kernels ear -! PR95 wet High N KPPm~, 417 300 GY 400 GY 500 GY 600 GY

418 4097 5302 6385 7134

310 4258 5622 7001 8271

371 3853 5056 6285 7056

533 4220 5552 6818 7878

4781 6331 7804 9232

4242 5581 6859 7914

Xint = 6.49 and 6; 5.5, 5 g CH20 day -t PR95 wet High N 6.49, 6 GY 5.5 GY 5 GY TL95 wet Medium N 6.49, 6 GY 5.5 GY 5 GY

5521 5521 5521 5616 5616 9349

5879 8753 8753 5401 5401 10054

4016 4016 7103 3793 3793 3793

5050 5050 5050 4556 4556 4556

8203 8203 8203 7305 13349 13349

Plant density = 5.33, 6.67, 8 plants m -i PR95 wet High N 5.33 DM GY 6.67 DM GY 8 DM GY

12837 5521 13970 6901 14586 7711

13807 5879 14817 8944 15835 9771

11360 4016 12461 6684 13582 7675

12208 5050 13277 8071 14415 9011

17555 8203 19115 13225 20107 14547

13553 5734 14728 8765 15759 9743

8191 3875 8608 3880 8990

8274 3130 8997 3133 9028

7918 3099 8290 3158 8643

8394 4227 8856 4240 9289

8331 3251 8750 3559 9139

RUEveg = 4.5 g MJ -! PR95 wet High N

k = 0.53, 0.5, 0.4 TL95 dry Medium N

PR95 wet Low N

PR95 wet Medium N

PR95 wet High N

Source-limited grain fill PR95 wet High N

0.53 0.5 0.4 0.53 0.5 0.4 0.53 0.5 0.4 0.53 0.5 0.4

Leaf N concentration = 0%, 5%, 10% and 25% above observed PR95 wet Low N 0% DM 8880 GY 3274 5% DM 9327 GY 3386 10% DM 9747

163

Table 4 (continued)

La Posta Sequfa C4

Across PR 8330 8328 BN C6

3473 10869 3638

3957 10011 4058

3202 9951 3292

Post-anthesis lower limit to LAI = O, 0.5, 1, 1.5 ha ha -t PR95 wet Low N 0 DM 5775 GY 2254 0.5 DM 5775 GY 2254 1.0 DM 5850 GY 2275 1.5 DM 6141 GY 2490

6383 2926 6382 2926 2522 2962 6918 3235

5632 2229 5667 2264 2866 2408 6228 2706

Cycle

Growing condition

25%

GY DM GY

Pool 16 C20

CML254 x CML247

Average

3092 9535 3190

4281 10457 4507

3601 10165 3737

5207 2004 5325 2103 5746 2433 6358 2966

7048 3948 7048 3948 7162 3985 7455 4188

6009 2672 6040 2699 6229 2813 6620 3117

Simulation data result from: increased radiation use efficiency during the vegetative phase (RUEveg) to 4.5 g MJ-j" from a reduced light extinction coefficient (k), representing more erect leaves; a source-limited grain fill; an increased maximum number of kernels on the primary ear (Y~t); a reduced daily plant growth rate at which a second ear starts forming (X~,t); increased plant density; increased leaf nitrogen concentration; and increased green LAI after silking. The number of kernels per plant observed at PR under high N (KPPma~) is given in case of variation in Xint.

yield were on the average 2756 kg ha-~ and 1861 kg ha-] , respectively, greater than the observed averages (Tables 1, and 4).

3.3.2. Light interception Our five cultivars were characterized by a relatively 300

~250

high k of 0.53. Much tropical maize germplasm is characterized by prostrate leaves; however, some new germplasm carries more erect leaves as a result of selection. We simulated growth for crops each characterized by leaf areas as observed in the field experiments, and by k-values of 0.5, 0.45 and 0.4, respectively. For the same leaf area, a better light penetration in the canopy reduced simulated biomass and grain yield at all levels of leaf nitrogen concen20000

• •

U

~150 15000

.m

U

.g

,~ 5o

"~ lO000

0

]

~ 40

50

e # ~ '/

60

70

80

90

I00

.j

5ooo

m

time (days after sewing) Fig. 4. Simulated daily amount of carbohydrates available from photosynthesis and translocation of stem reserves for grain growth (kg ha-j day-J; 0, O), and simulated maximum daily grain growth rate determined by number of kernels per ha and maximum grain growth rate (kg ha-~ day-J; II, I-I), for La Posta Sequfa C4 at high (O, O) and low (1:], O) soil nitrogen availability at Poza Rica during the 1995 wet summer season.

I

0

,

5000

t

I

I

10000

15000

20000

observed weight (kg ha"t) Fig. 5. Observed and simulated above-ground biomass ( • ) and grain weights (0) for the five cultivars evaluated during the 1995 wet summer season at Tlaltizap:tn and at Poza Rica for three levels of leaf nitrogen.

164

100 = PR, high N 1 •-II--PR, medium N I

90 80

--&-- PI~ low N

70 60 .~

50

.~ ..~

4o

~

2o

30

0

5

10

15

20

ASI (d)

Fig. 6. Average simulated relative grain yield reduction for the five cultivars evaluated during the 1995 wet summer season at Tlaltizap~in (TL) and at Poza Rica (PR) for three levels of leaf nitrogen content in relation to the anthesis-silking interval (ASI).

tration (Table 4). Simple computations of light interception profiles and photosynthesis rates in canopy layers of I ha ha -l showed that a reduction in k results in greater canopy photosynthesis only if the LAI increases to values of 5 - 6 (depending on the amount of radiation) and greater (data not given; see also Duncan, 1971). At lower LAIs, the effects of lower amounts of intercepted radiation are not sufficiently off-set by increases in radiation use efficiency.

3.3.3. Sink capacity Analysis of ear growth showed that at high production levels, grain filling is sink-limited (Fig. 4) and therefore, a greater KPP should increase grain yields. Simulated grain yields determined only by the supply of carbohydrates from photosynthesis and translocation of stem reserves, and using observed leaf areas and leaf nitrogen concentrations, vary under high N conditions from 7215 kg ha -~ for PR 8330 to 13569 kg ha -~ for the hybrid (Table 4). One way to select for greater KPP is to increase the maximum number of kernels on the primary ear. Table 4 gives relative increases in simulated grain yield at high N availability in PR in 1995 for the five cultivars as a consequence of increasing the maximum number of kernels on the primary ear from 300 to 600 kernels per ear. Xint was maintained at 6.5 CH20 plant -l day -l. On average, grain yields increased in three steps from 4242 to 7914 kg ha -~. Alternatively, prolificacy can be

selected for through reduction of the growth rate at which a second ear is formed. Simulated reduction of X~ntto 5.5 g CH20 plant -I day -l resulted in the formation of an additional ear, more KPP, and greater sink capacity for the hybrid in TL and Across 8328 BN C6 in PR. Simulated reduction of Xi,t to 5.0 g CH20 plant -l day -~ resulted in an additional ear for PR 8330 in PR and La Posta Sequfa C4 and Across 8328 BN C6 in TL. These are also the cases in which highest simulated growth rates around flowering were realized. Simulated grain yield increased if an additional ear was simulated (Table 4). The number of kernels per m 2, and thus sink capacity, can also be increased through increased plant density. Increasing in plant density from 5.53 to 6.67 and 8 plants m -2, at high leaf nitrogen concentrations and leaf areas as observed, resulted in simulated average biomass increases of 9 and 16%, respectively, and in large simulated average grain yield increases of 53 and 70%, respectively (Table 4).

3.4. Options to increase grain yield under limited nitrogen availability 3.4.1. Nitrogen uptake The effects of increased nitrogen uptake were simulated by assuming increases of 5, 10 and 25% in leaf nitrogen concentration under low nitrogen conditions in PR in 1995. This resulted in average simulated biomass increases of 5.0, 9.7, and 22.0%, respectively, and in average simulated grain yield increases of 1.2, 2.3 and 6.1%, respectively (Table 4). Predicted grain yield increases under high nitrogen conditions were lower than 1%, as grain filling is not a sourcelimited process under these conditions. 3.4.2. Leaf senescence LAI in field experiments started to decrease under low nitrogen conditions around flowering, reaching very low values at maturity, and in some cases, all leaf material had senesced by the end of the growing season. Extended leaf longevity, resulting in greater late-season LAI will be particularly effective if the flow of carbohydrates during the linear phase of the grain filling period increases as a result of this. The effect of a greater leaf longevity was simulated by making observed LAI model input, and by restricting its decline to 0.5, 1.0 and 1.5 ha ha -l after anthesis.

165

LAI took observed values, but after anthesis was not allowed to drop below the defined minimum values. In simulation studies of the low nitrogen treatment in PR in 1995, if the observed decline of LAI after anthesis was limited by a minimum value of 0.5 ha ha -l , then grain yields increased by 4.9% from 2004 to 2103 kg ha -l for Pool 16 C20 and 1.6% from 2229 to 2264 kg ha -I for PR 8330 (Table 4). These were the two cultivars with lowest LAIs throughout the growing season, and the only two cultivars with LAIs below 0.5 ha ha -~ at maturity (Table 4). For a lower limit of LAI of 1.0 ha ha -I, average grain yield increase was 5.2%, and for a lower limit of LAI of 1.5 ha ha -~, average increase was 16.6%.

3.5. Options to increase grain yield under limited water availability 3.5.1. Sink capacity of the young ear The maximum observed ASI in TL in 1996 was 20.3 days, which resulted in a yield reduction of 88%. ASIs up to 20 days were used as input in simulations, as a reflection of a short-term growth limitation at flowering which had no negative consequences for later crop growth. Simulations indicated that the relative decreases in grain yield were similar among the nitrogen levels in the PR and TL environments (Fig. 6). This is a direct consequence of the similar fractional declines in KPP in all environments and the limitations that the low numbers of KPP set on grain filling. Simulated grain yield decrease started at ASIs of 1-3 days, and reached values of about 90% at an ASI of 20 days.

4. Discussion

Leaf nitrogen contents in TL in 1995 were slightly below those of the medium nitrogen level treatment in Poza Rica in 1995, which suggests that the demand for nitrogen was not met at TL. This does not explain, however, the low value of RUEveg at TL after correction for leaf nitrogen content (Fig. 1). Some other growth limiting factor that we have not observed may have limited crop growth rate. Further experiments may indicate whether, and why, growth at this location is in general less efficient than at PR. Simulated improvements in RUE at high levels of

leaf nitrogen concentration lead to greater biomass and grain yield (Table 4). An observed RUEveg of 3.25 g MJ -~ corresponded with a biomass production of 12771 kg ha -l, and a simulated RUE~eg of 4.5 g MJ -j corresponded with an average biomass production of 15527 kg ha -I. This implies that every improvement in RUEveg of 0.1 g MJ -I corresponds to an increase in biomass production of 220 kg ha -~, and at a harvest index of 0.5, an increase in grain yield of I l0 kg ha -l. Examples of increased RUE through breeding are known. Four cycles of selection in Across 8328 BN showed increased biomass production without differences in nitrogen uptake (Lafitte and Edmeades, 1994b), and therefore a greater RUE. Another attempt at CIMMYT to select for RUE (Chapman and Edmeades, 1995) was less successful. Although differences in RUE explained differences in biomass production, selection was hampered by the few differences among cultivars, and the large genotype .'.x environment interaction. Richards (1996) mentions the contrast between wheat and barley: most increase in wheat yield has come from an increase in harvest index, whereas for barley, harvest index and biomass have equally contributed to improved yields. Tollenaar (1989) reported on increased dry matter accumulation of newer maize hybrids, and speculated that this could be partly attributed to increased tolerance to plant density stress, which would result in greater RUE. It appears therefore that at high production levels, at which moisture and nutrients are not growth limiting, direct selection for RUE offers some promise. Grain fill at high leaf nitrogen concentrations was limited by the supply of assimilates from current photosynthesis and translocation from the stem reserves (Fig. 4). Therefore, increasing the number of kernels per m 2 should increase grain yield. This can be achieved through increase of the number of kernels per plant, or through increased plant density. KPP can be increased through selection for more kernels on the primary ear before a second ear is formed (Yi,t), or through breeding for reduction of the plant growth rate at which a second ear is formed (Xi,t). As increased Yinttheoretically leads to an increased number of kernels on the primary ear at any growth rate below Xi,t, simulated increases in Yint caused greater grain yields for all cultivars, which were characterized

166 by different growth rates during sink development. Reduced Xint, however, only leads to more KPP if plant growth rate exceeds Xint, which was not always the case in simulations. Therefore, theoretically, breeding for larger primary ears appears preferable. However, larger ears may not be the best practical option for obtaining greater grain yields. Good husk cover may be difficult to preserve, and bird and insect injury may increase. Prolificacy may in that case be a better option, especially if ears are harvested by hand. Either crop growth rate during sink formation must be increased, or Xint must be reduced. Breeding for increased crop growth rate is difficult, although breeding for increased remobilization of stem reserves may be an option. Development of semi-prolific hybrids at CIMMYT (D. Beck, pers. commun.) suggests that breeding for a lower value of gin t is possible. In any case, breeding has to guard against stem lodging, which may increase through greater ear weight and reduced stalk strength if more stem reserves are translocated. Simulated increases in plant density from 5.53 to 6.67 and 8 plants m -2 resulted in large average grain yield increases of 53 and 70%, respectively (Table 4). As green LAI was held constant, this is not explained by better light interception, but by formation of more primary ears on more plants with the same amount of intercepted energy. However, this approach requires tolerance to high plant density (lodging, barrenness) and sufficiently high soil nitrogen availability to sustain a minimum plant growth rate during sink development. Reduced plant growth rate and ear abortion may be the consequence if nitrogen becomes growth limiting, which is often the case in tropical maize farming. The simulation model that was used did not account for such plant variation, and therefore may have over-estimated yield increases. Simulation of more erect leaves only results in increased crop growth at LAIs greater than 5-6, which confirms existing views (Duncan, 1971). At lower LAIs, the effects of reduced interception of radiation are not sufficiently off-set by increases in RUE. However, such high LAIs are not common in tropical maize, and, if achieved through increased plant density, typically result in lodging and barrenness. The latter is caused by inter-plant competition, which reduces the amount of assimilates per plant available for early kernel development. An effect

that the model does not account for, however, is the possible important role of the ear leaf as a preferential carbohydrate supplier to the developing ear (Edmeades et al., 1979). A better lighting of ear leaves therefore probably helps to reduce barrenness. Water or nitrogen deficit may change leaf angle and shape (e.g., leaf rolling), which affects light interception. The grain filling process was source-limited under low nitrogen experimental conditions (Fig. 4). This implies that increasing sink capacity is unlikely to result in greater grain yields. A possible avenue is to select for cultivars with a root architecture that extracts more nitrogen from the soils, presuming that this will result in higher leaf nitrogen concentrations. Our simulation studies suggested that a 1.2% increase in grain yield would result for each 5% increase in leaf nitrogen concentration. Greater nitrogen uptake, however, has been difficult to select for (Lafitte and Edmeades, 1994b). eqn (1) implies that at low values of NL more canopy nitrogen leads to a greater RUE and growth rate. This may result in a greater amount of carbohydrates available for grain filling. Although all maize cultivars had accumulated about 15 kg leaf N ha -~ ground surface area after about 4-5 weeks after sowing in all experiments, the maximum amounts of NL in green leaf tissue varied under low nitrogen conditions from 17.4 kg ha -~ for La Posta Sequfa C4 to 25.4 kg ha -1 for CML254 x CML247. We have no experimental data to explain these cultivar differences. Possibly, different root architecture plays a role in capture of nitrogen early in the season before it is lost by leaching or denitrification. Extended leaf longevity under low nitrogen conditions is particularly effective if the LAI after silking decreases and grain filling becomes source-limited, as is illustrated by the simulation studies. Breeding for delayed senescence is possible (Lafitte and Edmeades, 1994b). However, this will be at the cost of nitrogen availability for grain filling, which may reduce grain yields. Better assessments of the effects of delayed senescence require a model that incorporates a crop nitrogen balance, including within-plant nitrogen translocations. Our field data were insufficient to analyze in detail grain filling under water limited growing conditions (TL in 1996). However, the close relation between grain yield and kernels per plant (see also Bolafios

167

and Edmeades, 1993a) indicates that grain filling after mid-season drought stress is sink-limited. Increasing plant density to increase the number of ears and kernels m -2 is a risky strategy under water limited growing conditions, as the increased evaporative demand in combination with possible prolonged drought can increase end-season water deficit. Therefore, breeding preferably concentrates on increasing the number of ears (usually below one ear per plant) and kernels per plant, which can be achieved by selecting in managed drought environments for increased grain yield, reduced ASI, and increased number of ears per plant. For a given value of ASI, the fractional reduction in kernels per plant can be estimated. Simulated grain yield decrease started at ASIs of 1-3 days, and reached values of about 90% at an ASI of 20 days. The relationship described by Bolafios and Edmeades (1993b) between ASI and grain yield implies a 96% decrease in yield for an ASI of 20 days. CML254 x CML247 had at low water availability an ASI of 20.3 days, which was associated with a yield reduction of 88%. However, genetic and environmental variation in ASI is wide, which makes forecasting its behavior in a particular environment difficult. There are, to our knowledge, no data available on the relationship between photosynthesis or crop growth rate around flowering and the length of the ASI. Quantification of these relations would enable development of explanatory modules that, on the basis of photosynthesis or crop growth rate, determine both the potential number of kernels per plant and the fractional reduction due to water deficiency.

Acknowledgements We thank Srs. Lazaro Castafieda Tolentino at PR and Juan Carlos Bahena at TL, and their staff, for collecting all field data, and Drs. J. Bolafios, M. Reynolds and S. Pandey for reviewing drafts.

References Bassetti, P. and Westgate, M.E., 1993. Water deficit affects receptivity of maize silks. Crop Sci., 33: 279-282. Ten Berge, H.F.M., M.C.S. Wopereis, Riethoven J.J.M. and Drenth, H., 1994. Description of the ORYZA-0 modules. In: H. Drenth, H.F.M. ten Berge and J.J.M. Riethoven (Editors),

ORYZA Simulation Modules for Potential and Nitrogen Limited Rice Production. SARP Research Proceedings, AB-DLO, Wageningen, TPE-WAU, Wageningen, IRRI, Los Bafios, pp. 7-42. Bolafios, J. and Edmeades, G.O., 1993a. Eight cycles of selection for drought tolerance in lowland tropical maize. I. Responses in grain yield, biomass, and radiation utilization. Field Crops Res., 31: 233-252. Bolafios, J. and Edmeades, G.O., 1993b. Eight cycles of selection for drought tolerance in lowland tropical maize. II. Responses in reproductive behavior. Field Crops Res., 3 l: 253-268. Boote, K.J. and Tollenaar, M., 1994. Modeling genetic yield potential. In: K.J. Boote, J.M. Bennett, T.R. Sinclair and G.M. Paulsen (Editors), Physiology and Determination of Crop Yield. ASAJ CSSA/SSSA, Madison, WI, pp. 533-565. Brizuela, L., Zea, J.L., Aguiluz, A., Dubon, T. and Bolafios, J., 1993. Selecci6n para tolerancia a sequfa en tuxpefio C6 x BS19 a tray,s niveles de estr~s en Centro Am6rica. In: J. Bolafios, G. Safn, R. Urbina and H. Barreto (Editors), Programa Regional de Mafz para Centro Am6rica y Caribe: Sintesis de Resultados Experimentales 1992. CIMMYT-PRM, Guatamala, pp. 63-\ 66. Chapman, S.C. and Edmeades, G.O., 1995. Radiation use efficiency of lines in a tropical maize population. Paper presented at Australian Agronomy Conf. 1995. CIMMYT, 1990. 1989/90 CIMMYT World Maize Facts and Trends: Realizing the Potential of Maize in Sub-Saharan Africa. CIMMYT, Mexico, 71 pp. CIMMYT, 1994. International Maize Testing Programs: 1992 Final Report. CIMMYT, Mexico, 362 pp. de Wit, C.T., 1992. Resource use efficiency in agriculture. In: P.S. Teng and F.W.T. Penning de Vries (Editors), Systems Approaches for Agricultural Development. Elsevier, Amsterdam, pp. 125-151. Dowswell, C.R., Paliwal, R.L. and Cantrell, R.P., 1996. Maize in the Third World. Westview Press, 268 pp. Duncan, W.G., 1971, Leaf angle, leaf area and canopy photosynthesis. Crop Sci., I l: 482-485. Edmeades, G.O. and Daynard, T.B., 1979. The relationship between final yield and photosynthesis at flowering in individual maize plants. Can. J. Plant Sci., 59: 585-601. Edmeades, G.O., Daynard, T.B. and Fairey, N.A., 1979. Influence of plant density on the distribution of 14C-labelled assimilate in maize at flowering. Can. J. Plant Sci., 59: 577-584. Edmeades, G.O., B~inziger, M., Elings, A., Chapman, S.C. and Ribaut, J.-M., 1997. Recent advances in breeding for drought tolerance in maize. In: M.J. Kropff, P.S. Teng, P.K. Aggarwal, J. Bouma, B.A.M. Bouman, J.W. Jones and H.H. van Laar (Editors), Applications of Systems Approaches at the Field Level, Vol. 2, Proc. 2nd Int. Symp. Systems Approaches for Agricultural Development,6-8 December, 1995, IRRI, Los Bafios, Philippines, pp. 63-78. Kiniry, J.R., Jones, C.A., O'Toole, J.C., Blanchet, R., Cabelguenne, M. and Spanel, D.A., 1989. Radiation-use efficiency prior to grain filling for five grain-crop species. Field Crops Res., 20: 51-64. Lafitte,H.R. and Edmeades. G.O., 1994a. Improvement for toler-

168

ance to low soil nitrogen in tropical maize I. Selection criteria. Field Crops Res., 39: 1-14. Lafitte, H.R. and Edmeades, G.O., 1994b. Improvement for tolerance to low soil nitrogen in tropical maize II. Grain yield, biomass production, and N accumulation. Field Crops Res.. 39: 1525. Lafitte, H.R. and Edmeades, G.O., 1995. Stress tolerance in tropical maize is linked to constitutive changes in ear growth characteristics. Crop Sci., 35: 820-826. Montgomery, E.G., 1911. Correlation Studies in Corn. 24th Nebraska Agric. Exp. Stn. Report, Lincoln, NE, pp. 108-159. Penning de Vries, F.W.T. and van Laar, H.H., 1982. Simulation of Plant Growth and Crop Production. Simulation Monographs, PUDOC, Wageningen, The Netherlands, 308 pp. Penning de Vries, F.W.T., Jansen, D.M., Berge, H.F.M. ten, and Bakema, A., 1989. Simulation of Ecophysiological Processes of Growth in Several Annual Crops. Simulation Monographs, PUDOC, Wageningen, The Netherlands, 271 pp. Richards, D.A., 1996. Increasing the yield potential of wheat: manipulating sources and sinks. In: M.P. Reynolds, S. Rajaram and A. McNab (Editors), Increasing Yield Potential in Wheat: Breaking the Barriers. CIMMYT, Mexico, pp. 134-149. Rosegrant, M.W., Agcaoli-Sombilla, M. and Perez, N.D., 1995. Global Food Projections to 2020: Implications for Investment.

Food, Agriculture and the Environment Discussion Paper 5, IFPRI, Washington, DC, 54 pp. Schussler, R.H. and Westgate, M.E., 1995. Assimilate flux determines kernel set at low water potential in maize. Crop Sci., 35: 1074-1080. Sinclair, T.R. and Horie, T., 1989. Leaf nitrogen, photosynthesis and crop radiation use efficiency: a review. Crop Sci., 29: 9098. Spitters, C.J.T., van Keulen, H. and van Kraalingen, D.W.G., 1989. A simple and universal crop growth simulator: SUCROS87. In: R. Rabbinge, S.A. Ward and H.H. van Laar (Editors), Simulation and Systems Management in Crop Protection. Simulation Monographs, Pudoc, Wageningen, pp. 147-181. Tollenaar, M., 1989. Physiological basis of genetic improvement of maize hybrids in Ontario from 1959 to 1988. Crop Sci., 31: I 19124. Tollenaar, M., Dwyer, L.M. and Stewart, D.W., 1992. Ear and kernel formation in maize hybrids representing three decades of grain yield improvement in Ontario. Crop Sci., 32: 432-438. van Keulen, H. and Seligman, N.G., 1987. Simulation of Water Use, Nitrogen Nutrition and Growth of a Spring Wheat Crop. Simulation Monographs, PUDOC, Wageningen, The Netherlands, 310 pp.

Section 4 MANAGING RESOURCE USE Nitrogen budgets of three experimental and two commercial dairy farms in the Netherlands J.J. Neeteson and d. Hassink. ....................................................................................................................... 171 Resource use at the cropping system level P. C. Struik and F. Bonciarelli ...................................................................................................................... 179 Reprinted from the European Journal of Agronomy 7 (1997) 133-143 The efficient use of solar radiation, water and nitrogen in arable farming: matching supply and demand of genotypes ,4.J. Haverkort, H. van Keulen and M.I. Minguez .......................................................................................... 191 Soil-plant nitrogen dynamics: what concepts are required? E.,4. Stockdale, J.L. Gaunt and J. Vos......................................................................................................... 201 Reprinted from the European Journal of,4gronomy 7 (1997) 145-159 Modeling crop nitrogen requirements: a critical analysis C. O. Stockle and P. Debaeke ....................................................................................................................... 217 Reprinted from the European Journal of Agronomy 7 (1997) 161-169 Maize production in a grass mulch system - seasonal patterns of indicators of the nitrogen status of maize B. Feil, S. V. Garibay, H.U. ,4mmon and P. Stamp ........................................................................................ 227 Reprinted from the European Journal of Agronomy 7 (1997) 171-179 Nitrogen transformations after the spreading of pig slurry on bare soil and ryegrass lSN-labelled ammonium T. Morvan, Ph. Leterme, G.G. Arsene and B. Mary ...................................................................................... 237 Reprinted from the European Journal of Agronomy 7 (1997) 181-188 Size and density fractionation of soil organic matter and the physical capacity of soils to protect organic matter J. Hassink, A.P. Whitmore and J. Kub6t ....................................................................................................... 245 Reprinted from the European Journal of Agronomy 7 (1997) 189-199

Characterization of dissolved organic carbon in cleared forest soils converted to maize cultivation L. Delprat, P. Chassin, M. Linbres and C. ,lambert ....................................................................................... 257 Reprinted from the European Journal of Agronomy 7 (1997) 201-210 Analysis of impact of farming practices on dynamics of soil organic matter in northern China H.S. Yang and B.H. Janssen .......................................................................................................................... 267 Reprinted from the European Journal of Agronomy 7 (1997) 211-219 Agronomic measures for better utilization of soil and fertilizer phosphates K. Mengel .................................................................................................................................................... 277 Reprinted from the European Journal of Agronomy 7 (1997) 221-233


(~ 1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

171

Nitrogen budgets of three experimental and two commercial dairy farms in the Netherlands J.J. Neeteson *, J. Hassink Research Institute for Agrobiology and Soil Fertility (AB-DLO), P.O. Box 129, NL-9750 A C Haren, The Netherlands Abstract

Results of recent Dutch research on the quantification of nitrogen (N) budgets of grazed grassland fields and dairy farms are reviewed to obtain a better quantitative insight into the contribution of the various processes in the N cycle to N losses on dairy farms. The results are also used to investigate the feasibility of possible future regulations on maximum permissible values of the N surplus on dairy farms. The N surplus of grassland fields was assumed to equal the difference between N input through fertilizer, urine and dung, and atmospheric deposition, and the N output through harvested grass. Under experimental conditions the N surplus of grassland fields was found to be high, ranging from about 200 to almost 700 kg N ha -1 yr-1. With few exceptions it was not possible to explain the surplus entirely by measured N losses and N accumulation in soil organic matter. The N surplus on the complete-farm scale, i.e. the difference between the annual input of N to the farm and the annual output from the farm through agricultural products, did not exceed 200 kg N ha -1 yr -1 on a farm where a major effort is put into matching economic and environmental demands. At commercial farms managed according to "good agricultural practice" the value of the complete-farm N surplus is higher, about 250 kg N ha -1 yr-~. There is a strong positive linear relationship between the N fertilizer application rate and the complete-farm N surplus. It is concluded that the N surplus on dairy farms can be reduced considerably by a combination of measures such as reducing N fertilizer inputs and applying limited grazing. It is technically possible to meet the most severe standards proposed for the N surplus on dairy farms.

Keywords: Nitrogen budget; Nitrogen surplus; Nitrogen loss; Dairy farms; Nitrogen management; Regulations

1. Introduction

While producing agricultural goods farmers have to meet both economic and environmental requirements. Obviously, farming should be profitable to the farmer himself. Society, however, demands that agricultural production takes places in a sustainable manner, which implies that no severe ecological damage occurs. Nitrogen (N) may have a harmful effect on the environment through nitrate leaching or runoff, nitrous oxide emission and ammonia vol-

* Corresponding author. AB-DLO, P.O. Box 129, 9750 AC Haren, The Netherlands. Telephone: +31-505337204, fax: +31-505337291.

atilization. As far as N is concerned, farm management should be directed to minimizing N losses. In other words, the difference between the total N input into the farm by fertilizers, manures, fodder, concentrates, seeds, atmospheric deposition and symbiotic N fixation and the total N output by agricultural products should be as small as possible. Future Dutch regulations on N management will be based on this difference, the N surplus. The Dutch government proposes to set maximum permissible values for the N surplus on arable land and grassland. The values proposed for 1998 are 175 and 300 kg N ha -1 yr -1 for arable land and grassland, respectively. It is intended to decrease the values stepwise to 100 kg N ha -1 yr -1 for arable land and 180 kg N ha -1 yr -1 for grassland in the year 2008.

172

This paper reviews results of recent Dutch research on the quantification of N budgets of grazed grassland fields and dairy farms to obtain a better quantitative insight into the contribution of the various processes in the N cycle to nitrogen losses on dairy farms and to investigate the technical feasibility of the criteria proposed for the N surplus. The paper focuses on dairy farms because about twothirds of the total Dutch farmland is used as grassland by dairy farmers. Moreover, since the total amount of nitrogen cycling annually on dairy farms is among the highest in agricultural production systems, a relatively small improvement in the efficiency of the system may have a great impact on decreasing N losses on the national scale. The research presented has been conducted by the Research Institute for Agrobiology and Soil Fertility (AB-DLO) in Wageningen and Haren, the Experimental Station for Cattle, Sheep and Horse Husbandry (PR) in Lelystad, the Department of Agronomy of Wageningen Agricultural University (WAU), the Winand Staring Centre for Integrated Land, Soil and Water Research (SC-DLO) in Wageningen, the Nutrient Management Institute (NMI) in Wageningen, and the Centre for Agriculture and Environment (CLM) in Utrecht.

was executed in two blocks of four paddocks of each 1-1.5 ha. Four annual rates of N fertilizer were applied: 250, 400, 550 and 700 kg N ha -1. The paddocks were stocked with spring-calving Friesian dairy cows according to a put and take continuous grazing system. Stocking rates were adjusted regularly to keep a target sward height of 6 cm. Averaged over the three-year experimental period the number of cow grazing days was about 600. Further experimental details are described by Deenen (1994). 2.2. De Meenthoeve

The experiment was carried out on a sand soil in Achterberg (province of Gelderland). The soil had been under grass for a great number of years and in 1981 the old sward was reseeded. In 1986 and 1987 a rotational grazing experiment was conducted with young steers on paddocks of 0.2 ha. Averaged over the experimental period the number of cow grazing days was about 800. Four annual rates of N fertilizer were applied" 250, 400, 550 and 700 kg N ha -1. Further experimental details are described by Deenen (1994). 2.3. De Marke

The materials and methods used to obtain the resuits discussed in this paper are only briefly presented here. Full details can be found in the original publications to which reference is made. The data presented originate from medium-term field experiments at the experimental dairy farms De Minderhoudhoeve, De Meenthoeve, De Marke and from monitoring the commercial dairy farms Kloosterboer and Achterkamp.

Detailed measurements were performed at a wet and at a dry grazed grassland site at the experimental dairy farm De Marke on a sand soil in Hengelo (province of Gelderland). The experiment was conducted during the period 1992-1994. The average annual N fertilizer application rate was 115 and 151 kg N ha -1 at the wet and dry site, respectively. The average annual N input with slurry, exclusive of dung and urine from the grazing herd, amounted to 196 and 171 kg N ha -1 to the wet and dry site, respectively. Further experimental details are described by Hack-Ten Broeke et al. (1996).

2.1. De Minderhoudhoeve

2.4. Kloosterboer

The experiment was carried out on a well drained young sedimentary calcareous silty loam soil in Swifterbant (province of Flevoland). The soil had been under grass for more than 20 years. In August 1985 the sward was reseeded with perennial ryegrass. In 1986, 1987 and 1988 a continous grazing experiment

The commercial dairy farm of Mr. Kloosterboer is located on a sand soil in Laren (province of Gelderland). In 1993/1994 the farm consisted of 21.8 ha grassland and 11.2 ha arable land used for silage maize porduction. Annual nitrogen inputs to and outputs from the farm were monitored during six

2. Materials and methods

173

years. The N bookkeeping took place from 1 May 1988 to 30 April 1994. The stocking rate was 1.8 milking cows per ha exclusive of accompanying young cattle. Annual milk production reached about 7500 kg per cow. Further farm and farm management characteristics are given by Den Boer et al. (1996).

2.5. Achterkamp The commercial dairy farm of Mr. Achterkamp is located on a clay soil in Oosterhout (province of Noord Brabant). In 1994/1995 the farm consisted of 40.7 ha grassland and 16.2 ha arable land used for silage maize production. Annual N inputs to and outputs from the farm were monitored during five years. The nitrogen bookkeeping took place from 1 May 1990 to 30 April 1995. The stocking rate was 1.3 milking cows per ha exclusive of accompanying young cattle. In 1995 the annual milk production reached about 7500 kg per cow. Further information on farm characteristics and on farm management is given by Den Boer et al. (1996). 2.6. Methods used to quantify the contribution of the

various processes in the N cycle Nitrate leaching at De Meenthoeve and De Marke

was determined with porous ceramic cups, whereas at De Minderhoudhoeve drain water was analysed. Ammonia volatilization was measured at De Minderhoudhoeve with the micrometeorological mass balance method (Bussink, 1994). At De Meenthoeve and at De Marke it was estimated on the basis of assumptions made by Biewinga et al. (1992). Denitrification was determined at De Marke in undisturbed soil samples which were incubated with acetylene (Ryden et al., 1987). Acetylene blocks the transformation of N20 to N2 and the N20 production is then a measure of the rate of denitrication. At De Minderhoudhoeve and at De Meenthoeve denitrification was estimated from the difference between soil mineral N in autumn and spring minus the amount of nitrate leached during winter. Changes in the amount of N in soil organic matter at De Minderhoudhoeve and De Meenthoeve were quantified by determining total soil organic N and the bulk density in each spring and autumn in the top 10 cm of soil and the 10-25 cm soil layer (Hassink and Neeteson, 1991). Five replicates were taken per layer, each replicate being a combination of 10 cores. At De Marke, it was quantified by determining the amount of N in the active fractions of the soil organic matter pool and the bulk density in each spring and autumn in the top 10 cm of soil and the 10-25 cm soil layer (Hassink, 1996).

Table 1 N budget (input/output, kg N ha -1 yr -1) of grazed grassland fields at De Minderhoudhoeve (loam soil) and De Meenthoeve (sand soil). The values given are average values of three (De Minderhoudhoeve) or two years (De Meenthoeve). After Hack-Ten Broeke et al. (1996) and Hassink and Neeteson (1991). Entry

N fertilizer application rate (kg N ha -1 yr -1) De Minderhoudhoeve

Fertilizer Urine Dung Atmospheric deposition Total input Output through gross grazed and mown Surplus Ammonia volatilization Denitrification Nitrate leaching Accumulation in SOM a Total losses and accumulation Unaccounted for aSOM = Soil organic matter

De Meenthoeve

250

400

550

700

250

400

550

700

251 165 70 40 526 330 196 19 8 11 250 288 -92

398 234 80 40 752 424 328 25 22 31 250 328 0

552 281 86 40 959 491 468 30 35 44 250 359 109

694 297 87 40 1118 515 603 31 44 53 250 308 295

268 279 85 40 672 405 267 47 88 59 0 196 71

406 323 86 40 855 449 406 53 96 75 0 224 182

517 343 84 40 984 466 518 56 115 134 0 305 213

672 353 80 40 1145 472 673 56 2 !6 163 0 435 238

174 Table 2 N budget (input/output, kg N ha -1 yr -1) at sites of grazed grassland at De Marke (sand soil). After Hassink et al. (1996). Entry

Dry site Wet site

Fertilizer Slurry, dung and urine Atmospheric deposition Symbiotic N fixation Total input Output through grass grazed and mown Surplus Ammonia volatilization Denitrification Nitrate leaching Accumulation in soil organic matter Accumulation in stubble and roots Total loss and accumulation Unaccounted for

151 307 50 34 542 278 264 14 11 83 101 55 264 0

115 282 50 19 466 308 158 11 27 22 73 24 157 1

The quantification of the N content in fertilizer, urine, dung, harvested grass, milk and meat and the estimation of the amount of N deposited from atmosphere are described by Deenen (1994).

3. Results

3.1. N budgets of dairy farms." the field scale In Table 1 N budgets of grazed grassland fields with various levels of N fertilizer application are presented. The budgets pertain to experimental fields at De Minderhoudhoeve (loam soil) and De Meenthoeve (sand soil). The N surplus of the fields was assumed to equal the difference between N input through fertilizer, urine and dung, and atmospheric

Table 3 Complete-farm N budgets (input/output, kg N ha -1 yr -1) of the dairy farm De Marke (sand soil). Each recorded year runs from 1 May to 30 April. After De Vries (1995) and Van Keulen et al. (1995). Entry

Farm year

37 82 53 0 49 12 5 238 65 11 21 97 141

0 123 74 0 49 12 0 258 6 68 0 74 184

deposition, and the N output through grass harvested (grazed+mown). The N surplus appeared to be high, ranging from about 200 to almost 700 kg N ha -1 yr -1. The surplus increased with increasing levels of N fertilizer. Independent of the level of N fertilizer the surplus at De Meenthoeve was about 70 kg N ha -1 higher than it was at De Minderhoudhoeve. This was due to the higher stocking rate at De Meenthoeve causing a higher N input through urine which was only partially matched by a higher N output through grass harvested. Since ammonia volatilization, nitrate leaching, and N accumulation in soil organic matter were also quantified at the fields studied, it is possible to investigate to which extent the calculated N surplus )lus (kg N/ha.w)

4OO

800 II

/

60O 400

94/95

Roughage from elsewhere Concentrates Fertilizer Manures from elsewhere Atmospheric deposition Symbiotic fixation Miscelaneous Total input Meat Milk Miscelaneous Output through produce Surplus

N sur

N sur flus (kg N/ha.W)

93/94

~lb

Y

I

300

I

2OO 100

200

0 N fertilizer application rate (kg N/ha.w)

Fig. 1. Relationship between N fertilizer application rate to grazed grassland fields and N surplus at the field scale.

100

200

300

400

N fertilizer application rate (kg N/ha.w)

Fig. 2. Relationship between N fertilizer application rate and N surplus of dairy farms.

175 Table 4 Complete-farm N budgets (input/output, kg N ha -1 yr -1) of the dairy farm Achterkamp (clay soil). Each recorded year runs from 1 May to 30 April (Den Boer et al., 1996). Entry

Farm year

Roughage from elsewhere Concentrates Fertilizer Manures from elsewhere Atmospheric deposition Symbiotic fixation Total input Meat Milk

Miscelaneous Output through produce Surplus

90/91

91/92

92/93

93/94

94/95

0 45 237 33 46 4 365 9 45 45 99 266

6 44 178 0 46 4 278 10 55 -24 41 237

30 92 222 0 46 4 394 9 55 38 102 292

12 66 239 0 46 4 367 10 52 24 86 281

1 68 187 0 46 4

could be explained in terms of losses and accumulation. The data presented in Table 1 show that averaged over the four N fertilizer application rates 80% of the surplus at De Minderhoudhoeve could be ascribed to losses and accumulation. At De Meenthoeve, however, only 62% of the surplus could be explained. The values of N accumulation at De Minderhoudhoeve in Table 1 are similar for all fertilizer treatments, because differences among treatments were non-significant, notwithstanding a lower N accumulation at the lower N fertilizer application rates. When making N budgets it is seldomly possible to take full account of the contribution of all N flows.

306 11 54 14 79 227

Detailed measurements at two sites with contrasting soil water status at the experimental farm De Marke, however, showed that it is possible to set up closed N budgets of grazed grassland fields by taking explicitly account of N accumulation in soil organic matter (Table 2). The calculated N surplus could thus be entirely explained by measured N losses and accumulation. In Fig. 1 the N surpluses given in Tables 1 and 2 are plotted against the respective N fertilizer application rates. Fig. 1 shows that there is a strong positive linear relationship. The N surplus at the field scale thus is largely determined by the level of N fertilizer application.

Table 5 Complete-farm N budgets (input/output, kg N ha -1 yr -1) of the dairy farm Kloosterboer (sand soil). Each recorded year runs from 1 May to 30 April (Den Boer et al., 1996). Entry

Roughage from elsewhere Concentrates Fertilizer Manures from elsewhere Atmospheric deposition Symbiotic fixation Total input Meat Milk

Miscdaneous Output through produce Surplus

Farm year 88/89

89/90

90/91

91/92

92/93

93/94

0 72 305 34 52 5 468 14 64 4

0 82 201 17 52 4 356 8 72 5

1 84 196 0 52 4 337 12 69 - 1 80 257

0 90 156 0 52 4 302 12 70 - 20 62 240

29 134 146 0 52 4 365 13 70 17 100 265

0 94 182 0 52 4 332 14 67 12 93 239

82

85

386

271

176

3.2. Nitrogen budgets of dairy farms: the complete-farm scale Table 3 presents the complete-farm N budgets of the experimental farm De Marke. At this farm the objective is to fulfil both economic and environmental objectives. Farm management aims at increasing the use of internal N flows within the farm in order to reduce demand for N inputs to the farm in the form of roughage and fertilizers. By doing so, the N surplus at the whole-farm scale, i.e. the difference between the annual input of N to the farm and the annual output from the farm through agricultural products, did not exceed 200 kg N ha -1 yr -1 (Table 3). On commercial farms managed according to "good agricultural practice" the value of the complete-farm N surplus is higher, about 250 kg N ha -1 yr -1 (Tables 4 and 5). Tables 4 and 5 also show that there was little between-year and between-farm variation in the N surplus. The relationship between N fertilizer application rate and the complete-farm N surplus was established from the data presented in Table 3, Tables 4 and 5. The strong positive linear relationship (Fig. 2) suggests that the N fertilizer application rate has a major impact on the N surplus.

4. Discussion

4.1. N budgets The N budgets of grazed grassland fields (field scale) and dairy farms (complete-farm scale) presented in this paper show that annual N inputs are high, several hundreds kg N ha -1. Annual N output in agricultural products, however, is low. At the field scale it was about 50% of the total annual N input (Tables 1 and 2) and at the farm scale only about 25% (Table 3, Tables 4 and 5). The low N efficiency of dairy farming systems is well-known (Van Der Meer and Van Uum-Van Lohuyzen, 1986). With few exceptions (Table 2) the difference between total N input and N output through products, the N surplus, calculated at the field scale could not be explained entirely by the measured N losses and accumulation into soil organic matter. This can be due to insufficient account that was taken of soil and

crop heterogeneity while measuring the contribution of the various processes and interpreting the data obtained. The values assessed may differ widely from the "real" values when errors are made in sampiing procedures and statistical analyses. It is also possible that relevant processes through which N losses occur have not been quantified at all, e.g. N2-emission after nitrification of ammonium (Pel et al., 1997). The data presented in Tables 1 and 2 show that nitrate leaching and in some instances denitrification are major loss mechanisms of N at grazed grassland fields. It should be noted that the values for denitrification at De Meenthoeve given in Table 1 seem to be unrealistically high. This is probably due to the indirect manner the values have been determined, i.e., from changes in soil mineral N which were corrected for nitrate leaching (see Materials and Methods). By doing so, the values for denitrification are likely to be overestimated, since all loss mechanisms during winter with the exception of nitrate leaching are then ascribed to denitrification.

4.2. N surplus The complete-farm N surpluses presented in this paper are less than 200 kg N ha -1 yr -1 when a major effort is made to reach environmental goals (Table 3) and about 250 kg N ha -1 yr -1 when good agricultural practice is followed (Tables 4 and 5). These values are considerable lower than the values of about 600 and 400 kg N ha -~ yr -~ reported by Van Keulen et al. (1995) for current intensively and extensively managed dairy farms, respectively. In 1990/1991 the average N surplus of 2099 Dutch commercial dairy farms calculated on the basis of general assumptions and some farm data available, appeared to be 419 kg N ha -1 (Bronwasser, 1992). The values calculated ranged from 159 to 618 kg N ha -1 yr -1, depending on total N application rate per ha, milk production per ha and stocking rate (Bronwasser, 1992). Other studies suggest that production intensity and/or N application rate can not be used as simple indicators for the N surplus, since there were large differences in the calculated N surplus among farms with similar production intensity and/ or N application levels (Anonymous, 1994; Daatselaar, 1989; Daatselaar et al., 1990). However, the results presented in this paper, which

177

were derived from actual measurements rather than assumptions, indicate that the N surplus largely depends on the N fertilizer application rate, both on the field scale (Fig. 1) and on the farm scale (Fig. 2).

4.3. Strategies to reduce the N surplus It is possible to reach a low N surplus at the farm level for various farming types (Anonymous, 1994). Farms with a high production intensity should make better use of the N in their internal flows of nutrients and feed. External inputs by fertilizers and feed produced elsewhere can then be lowered. Farms with a low production intensity by using no or little N fertilizer, generally reach a low N surplus. The results presented in Fig. 2 suggest that reducing N fertilizer application rates may have a large effect on lowering the N surplus. This is confirmed by model calculations of Van Der Putten and Vellinga (1996) who found a reduction of the N surplus of more than 60 kg N ha -1 yr -~ when 100 kg fertlizer N ha -1 yr -t less was applied than when the current recommendation was followed. The nitrogen surplus could be further reduced with 50 kg N ha -1 yr -1 when limited grazing instead of rotational grazing was applied together with reduced N fertilizer application rates. Adjustment of farm management can considerably reduce the N surplus. The best strategy is to take several measures simultaneously such as lower application rates of N fertilizer, limited grazing, lower stocking rates combined with increased per-cow production, and injection of slurries into the soil. Most of these measures have little effect on farmers' profitability, except limited grazing which requires extra costs for fodder ensilage and slurry collection and application (Mandersloot and Van Scheppingen, 1994).

4.4. Feasibility of governmental norms for the N surplus The Dutch governement intends to set up regulations to limit the N surplus on dairy farms. Maximum permissible values proposed range from 300 kg N ha -1 yr -1 in 1998 to 180 kg N ha -1 yr -1 in 2008. Although it may not be scientifically justified, it was a political decision to exclude symbiotic N fixation

and atmospheric deposition from the calculation of the N surplus used in the proposed legislation. The N surplus obtained at commercial farms with "good agricultural practice" (Tables 4 and 5) then reaches values of about 200 kg N ha -1 yr -1. The N surplus obtained at the experimental farm De Marke with a major effort to match both economic and environmental objectives (Table 3) then amounts to about 100 kg N ha -1 yr -1. This suggests that it is technically possible to reach the governmental norms for the N surplus. With "good agricultural practice" the most stringent criterion can almost be reached. From the results obtained at De Marke it can be expected that even the most severe criteria will not be exceeded when farm management moves somewhat from "good agricultural practice" towards management which takes more account of environmental objectives.

References Anonymous, 1994. Many equations and many unkowns. An analysis of farming types and differences in input-output relationships in Dutch dairy farming (in Dutch). NRLO Report 94/1. National Council for Agricultural Research, The Hague, the Netherlands, 112 pp. Biewinga, E.E., Aarts, H.F.M. and Donker, R.A., 1992. Dairy farming at severe environmental norms. Farm and research plan of the Experimental Farm for Dairy Farming and Environment (in Dutch). Report 1. De Marke, Hengelo, the Netherlands, 238 pp. Bronwasser, K., (Ed.), 1992. DELAR: An analysis of key figures in dairy farming in 1990-1991 (in Dutch). Publication 27. Informatie en Kennis Centrum Veehouderij, afdeling Rundvee-, Schapen- en Paardenhouderij. Lelystad, the Netherlands, 98 PP. Bussink, D.W., 1994. Relationships between ammonia volatilization and nitrogen fertilizer application rate, intake and excretion of herbage nitrogen by cattle on grazed swards. Fertil. Res., 38: ll-121. Daatselaar, C.H.G., 1989. An analysis of differences in nutrient budgets among dairy farms (in Dutch). Publication 3.144. Agricultural Economics Research Institute (LEI-DLO), The Hague, the Netherlands, 34 pp. Daatselaar, C.H.G., De Hoop, D.W., Prins, H. and Zaalmink, B.W., 1990. An analysis of nutrient use efficiency at dairy farms. Research Report 61. Agricultural Economics Research Institute (LEI-DLO), The Hague, the Netherlands, 98 pp. Deenen, P.A.J.G., 1994. Nitrogen use efficiency in intensive grassland farming. Thesis. Agricultural University, Wageningen, the Netherlands, 140 pp. Den Boer, D.J., Van Middelkoop, J.C. and Bussink, D.W., 1996.

178

Minimising nutrient losses in dairy farming (in Dutch). Nutrient Management Institute (NMI), Wageningen, the Netherlands, 105 pp. De "Cries, C.K., 1995. Grassland and fodder husbandry at the farm scale (in Dutch). In: H.F.M. Aarts (Editor), Weide- en Voederbouw op De Marke: op Zoek naar de Balans tussen Produktie en Emissie. Report 12. De Marke, Hengelo, the Netherlands, pp 7%89. Hack-Ten Broeke, M.J.D., Van Der Putten, AM.L, Corr6, W.J. and Hassink, J., 1996. Nitrogen losses to the environment (in Dutch). In: J.W.G.M. Loonen and W.E.M. Bach-De Wit (Editors), Stikstof in Beeld. Onderzoek inzake de mest- en ammoniakproblematiek in de veehouderij 20. Agricultural Research Department (DLO), Wageningen, the Netherlands, pp 78-98. Hassink, J., 1996. Nitrogen in stuble, microbial biomass, roots and active organic matter fractions (in Dutch). In: M.J.D. HackTen Broeke and H.F.M. Aarts (Editors), Integrale Monitoring van Stikstofstromen in Bodem en Gewas. Report 14. Experimental Station for Cattle, Sheep and Horse Husbandry (PR), Lelystad, the Netherlands, pp 55-63. Hassink, J. and Neeteson, J.J., 1991. Effect of grassland management on the amounts of soil organic N and C. Neth. J. Agric. Sci., 39: 225-236. Hassink, J., Aarts, H.F.M., Corr6, W.J, and Hack-Ten Broeke, M.J.D., 1996. Internal nitrogen flows in the soil-crop system at six observation sites (in Dutch). In: M.J.D. Hack-Ten Broeke and H.F.M. Aarts (Editors), Integrale Monitoring van Stikstofstromen in Bodem en Gewas. Report 14. Experimental Station for Cattle, Sheep and Horse Husbandry (PR), Lelystad, the Netherlands, pp 93-105. Mandersloot, F. and Van Scheppingen, A.T.J., 1994. Manure and ammonia issues at the farm scale and sector scale (in Dutch).

In: M.H.A. De Haan and N.W.M. Ogink (Editors), Naar Veehouderij en Milieu in Balans. Onderzoek inzake de mest- en ammoniakproblematiek in de veehouderij 19. Agricultural Research Department (DLO), Wageningen, the Netherlands, pp 125-146. Pel, R., Oldenhuis, R., Brand, W., Vos, A., Gottschal, J.C. and Zwart, K.B., 1997. Combined methanotrophic nitrificationdenitrification under micro-aerobic and thermophilic conditions in a model composting-system: a 15N and laC tracer study. J. Appl. Environ. Microbiol. (in press). Ryden, J.C., Skinner, J.H. and Nixon, D.J., 1987. A soil core incubation system for the field measurement of denitrification using acetylene-inhibition. Soil Biol. Biochem., 19: 753-757. Van Der Meer, H.G. and Van Uum-Van Lohuyzen, M.G., 1986. The relationship between inputs and outputs of nitrogen in intensive grassland systems. In: H.G. Van Der Meer, J.C. Ryden and G.C. Ennik (Editors), Nitrogen Fluxes in Intensive Grassland Systems. Martinus Nijhoff Publishers, Dordrecht, the Netherlands, pp 1-18. Van Der Putten, A.H.J. and Vellinga, Th.V., 1996. The effect of grassland management on the use of applied nitrogen. In: J.W.G.M. Loonen and W.E.M. Bach-De Wit (Editors), Stikstof in Beeld. Onderzoek inzake de mest- en ammoniakproblematiek in de veehouderij 20. Agricultural Research Department (DLO), Wageningen, the Netherlands, pp 36-59. Van Keulen, H., Aarts, H.F.M., Hermans, C. and De Wit, J., 1995. Prospects of Diary Farming and Environment (in Dutch). In: A.J. Haverkort and P.A. Van Der Werff (Editors), Hoe Ecologisch Kan de Landbouw Worden? AB-DLO Thema's 3. Research Institute for Agrobiology and Soil Fertility, Wageningen, the Netherlands, pp 137-144.

t~ 1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

179

Resource use at the cropping system level P.C. Struik a'*, F. Bonciarelli b aDepartment of Agronomy, Wageningen Agricultural University, Haarweg 333, 6709 RZ Wageningen, The Netherlands bIstituto di Agronomia Generale e Coltivazione Erbacee, Universita' di Perugia, Perugia, Italy

Accepted 11 July 1997

Abstract

This paper illustrates the basic ideas of good crop rotations, adequate crop husbandry and high resource-use efficiencies and some relevant ecological approaches. The use of special crops to prevent the need of high inputs of crop protectants or to reduce losses of nutrients at the level of the cropping system deserves special attention in research. Examples are given for the ecological control of soil-borne fungi, parasitic weeds, nitrogen loss and other sustainable techniques to increase the resourceuse efficiency at the cropping system level. © 1997 Elsevier Science B.V. Keywords: Cropping system; Crop rotation; Nutrient balance; Resource-use efficiency; Sustainability; Biological control; Leguminous crop; Ley crop; Nematodes; Nutrient catch crop; Residual nutrients; Soil-borne fungi; Trap crop; Weed

1. Introduction

1.1. Success leads to a new approach

About 200 years ago science started to influence agricultural practice. This resulted in a high technological level of agriculture in western Europe, associated with large inputs and high yields per hectare. The approach was so successful that, finally, it induced a co-evolution of new views on management of the 'green space', landscape, environment, resources and nature. These new views result in an agriculture which contrasts strongly with the original struggle to produce as much as 'nature would allow farmers'. Nowadays, agricultural technology and

sciences may serve other goals than 'merely' efficient food and feed production. There is an increasing pressure from governments to strive for additional goals such as maintenance of characteristic (semi-)natural and traditional, cultural landscapes and give more space and chance to 'natural' components in the agro-ecosystems. So, the hierarchy of objectives in agro-ecosystems has changed dramatically over the past few decades. This requires new heuristics for scientists and farmers. 1.2. The new approach is a balance

From an ecological point of view, sustainable agriculture should focus on a balance between the following goals:

* Corresponding author. Fax: +31 317 484575; e-mail: [email protected] Reprinted from the European Journal of Agronomy 7 (1997) 133-143

maximisation of the use of beneficial natural processes in the cropping system (such as

180

•

•

• •

•

nitrogen fixation and antagonistic or synergistic relationships between (micro)organisms). maximisation of the recycling of certain elements (nutrients, carbon, etc.). However, it must be taken into account that (almost by definition) agro-ecosystems are not fully closed systems for certain components. optimal use of internal resources, the carrying capacity of the agro-ecosystem, and the genetic potential of plants and animals. optimal management and use of variation within the system. restricted use of external resources, especially as far as they are unfriendly to nature, the environment, the user of the resource or the consumer of the end product. If external resources are needed, they should be used at maximum use efficiency (in terms of output per input; Van Ittersum and Rabbinge, 1997) and with a minimum of emission to the environment (in terms of loss per unit area per unit of time). durable use of critical resources (either internal or external), which are easily lost or damaged, such as soil and water, or which are short in the long-term, such as phosphate, energy and land.

The basis of sustainable agriculture is a good crop rotation, adequate soil and water management, and proper husbandry of the different crops in the rotation. Agronomically, farmers should aim at the minimum input of each production resource required to allow maximum utilisation of all other resources. Consequently, above a certain minimum, higher inputs of yield-increasing factors (such as water and nutrients) result in higher yields per unit area and are associated with higher efficiencies (expressed as output per unit of input) of other resources (De Wit, 1992; Rabbinge et al., 1994), but at the same time might cause large residues or emissions per unit area (Nijland et al., 1997).

1.3. Balancing requires information on a longer time scale Many processes relevant to resource-use efficiency (RUE) are so slow or long-lasting that they also have

effects at the time scale of an entire rotation. Examples are the behaviour of organic matter in the soil, the changes in the suppressiveness of a soil for a certain disease or the changes in the weed seed bank. This paper focuses on these processes. We will concentrate on crop rotation first and especially illustrate the (management of) effects on yield-reducing factors (pests, diseases, weeds) and the (management of) effects on yield-increasing factors at the cropping system level. As an example of the latter we will describe the management of nutrient cycling at the cropping system level. Thereafter, we will briefly discuss nutrient management at the crop level (the links of the rotation). Finally, the paper illustrates potential uses of crops to improve the sustainability.

2. Crop rotation Crop rotation is a more or less fixed pattern in the succession of crops on a certain field. Component crop species, frequency of each crop and the crop sequence all affect the yielding ability of the entire rotation and of each individual crop. Moreover, a certain crop rotation also means a more or less fixed pattern of management and inputs, and thus changes in the soil and water resources with time. RUEs at the crop rotation level are therefore not only determined by short-term efficiencies of component crops but also by long-term processes influenced by tillage, the different crops in the rotation and their management. These processes are reflected in the physical soil fertility (organic-matter content, water-holding capacity, etc.), the chemical soil fertility (pH, availability of nutrients, presence of heavy metals, etc.), and the biological soil fertility (presence of useful microorganisms, soil-borne pests, diseases, weeds, etc.). The carry-over effects of the crops might not always be very homogeneous over the entire field. Growing a crop may often mean increasing variability (Almekinders et al., 1995). The physical fertility is affected by each crop, the type and timing of cropping practices in each crop, and the measures taken during fallow periods to improve the physical fertility or to control the weed population. Differences in effects of crops may arise from differences in duration of the canopy cover, rooting patterns, amount of effective organic matter left after harvest, other effects on soil structure, etc. To

181

some extent, these characteristics can be manipulated. For physical soil fertility the hydrology of the site is extremely important as well as the short-term and long-term effects of management practices aiming at optimal soil and water use and soil and water conservation. Chemical soil fertility is affected by fertiliser application; the effects of crops on nutrient fixation and mobilisation, mineralisation and losses of nutrients; the effects of crops on salinisation or pH; the effects of crops on distribution or concentrations of chemical pollutants; the amount and quality of crop residues; and the rate of degradation of crop residues. The most obvious effects of crop rotation are usually found through its effect on the biological soil fertility. We will elaborate on that below, but in general one can state that the higher the frequency of crops sensitive to the same soil-borne diseases or other biological stresses (pests, weeds), the higher the need for crop protectants to control them. In contrast, the higher the diversity of crops with different positive effects on beneficial organisms, the lower the need for crop protectants. While setting out general rules for a good crop rotation one should consider at least: •

the effects of the preceding crops on physical, chemical and biological fertility of the soil;

• •

the sensitivity of the following crops to these effects; the cumulation of these effects over cropping systems.

These effects can be modelled to some extent using models such as CropSyst (Stockle and Nelson, 1994; Van Evert and Campbell, 1994) and others. In general, alternating crops with contrasting effects on the physical, chemical and biological soil fertility is usually advisable. For example, crops with a strong negative effect on the amount of effective organic matter should be alternated with crops that enhance the content of effective organic matter. Vereijken (1995) formulated some semi-quantitative rules for a good rotation. The Agronomy Institute of the University of Perugia (Italy) is co.nducting a long-term rotation experiment on the effects of crop residues since 1972, which may serve as an example for long-term effects of cropping system management on the physical, chemical and biological soil fertility. In this experiment winter wheat is grown in continuous cropping or in several rotations characterised by different wheat/ maize ratios (0.50, 0.67, 0.75, 0.80, 0.83). The crop residues are either removed or buried (with an addition of 1 kg N/100 kg of dry matter). Twenty years after the initiation of the experiment, several soil char-

Table 1 Effect of burying or removing crop residues on some physical, chemical and biological properties of the soil in a long-term rotation trial in progress in Perugia (Italy) since 1972 Crop residues

Stable aggregates (%) Atterberg index Proctor max bulk density (t m-3) Field capacity (% by weight) Organic C (%) Humified organic C (%) Total N (%) Microbial biomass (ng g-I h-i) FDA-hydrolase activity (ng g-i h-i) Dehydrogenase activity (ng TPF g-i 24 h-I) Catalase activity (mg 02 g-i h-i) Mites (#/unit soil) Collembola (#/unit soil)

Significancea

Removed

Buried

38.0 14.6 1.72 22.7 0.812 0.265 0.107 155 29.9 163 420 14.1 0.6

42.3 ]6.7 !.69 23.2 0.944 0.320 0.118 196 37.6 192 360 25.9 3.2

* * * * ** ** ** ** ** ** ** ** **

* means significant at 0.01 < P < 0.05; ** means significant at P < 0.01 (Source: Perucci et al., 1997 and unpublished data Istituto di Agronomia, Perugia)

182

acteristics were assessed. Small but significant improvements of many soil characteristics became apparent as a consequence of burying crop residues (Table 1).

3. Yield-reducing factors relating to crop rotation A crop rotation, which has been maintained for a large number of cycles, has created a certain longterm balance between soil organisms. Even with this balance, yields may be considerably lower than potentially possible, at least too low to make optimal use of the other resources. Nevertheless, when this balance is disturbed, for example by applying a certain chemical killing part of the organisms, then the population density of the non-target organisms may also change, either resulting in a positive, a negative or no effect on yield (Table 2). Table 2 shows that a higher cropping frequency of potato strongly increased the infections by soil-borne pathogens such as Verticillium dahliae (causing the early dying syndrome) and Rhizoctonia solani (causing stem canker and black scurf) in potato. In the case of Verticillium, also the cropping sequence influenced the disease patterns. Maize was a better preceding crop than sugar beet in the 1:2 rotation of potato.

Potato after sugar beet was equally infected as potato after potato. The use of nematicides greatly influenced the percentage of infested stems in all rotations and for both diseases. For the Verticillium infection, controlling the nematodes reduced the disease. There are several explanations for that, one being the fact that fewer nematodes means fewer gates for the fungus to enter the roots. For Rhizoctonia, the effect of the nematicides was opposite: control of nematodes increased the infection with Rhizoctonia, most likely because the nematicides killed the mycophagous nematodes, collemboles and other mesofauna, thus stimulating the build up of the population of the fungus, especially after maize. In contrast, the infection by Colletotrichum (another soil-borne fungus) was hardly affected by potato frequency, crop sequence or by application of nematicides. These effects, but also the effects of the nematodes and the nematicides themselves are reflected in the tuber yields (Table 2). The maximum yield loss by continuous cropping of potato was 20%. This yield loss is considerable, considering the fact that it occurred in the absence of damage caused by cyst nematodes. These results suggest that: 1. Synergistic and antagonistic effects occur in crop-

Table 2 Percentage of potato stems infected with Verticillium dahliae (averaged over 1983-1986), percentage of potato plants infected with Rhizoctonia solani (averaged over 1981-1986), percentage of potato stems infected with Colletotrichum coccodes (averaged over 1983-1986), and tuber dry matter yield (g/m2; averaged over 1981-1986) in four rotations in control plots and plots treated with nematicides (Scholte, 1989) Verticillium *( % )

Rhizoctonia a ( % )

Colletotrichum b (%)

Tuber yield~ (g/m 2)

Nematicide

No

Yes

No

Yes

No

Yes

No

Yes

Rotation P MP SP MSBBP Average**

49 39 50 21 40

34 20 38 13 26

48 22 23 9 26

62 41 32 14 37

35 29 33 28 31

32 30 36 27 31

99 131 118 152 125

122 144 154 167 147

P, potato; M, maize; S, sugar beet; B, barley. aContinuous cropping significantly higher than both 1:2 rotations, which were significantly higher than the 1:5 rotation. bNo significant rotation effects on infection by Colletotrichum. CContinuous cropping significantly lower than other rotations in almost all cases; the 1:5 rotation significantly higher than SP without nematicide and than MP with nematicide. All assessments in all years were done ca. 74 days after planting. Compiled from different chapters of Scholte (1989) based on the same longterm field experiment on a sandy soil in the Netherlands. *Rotation effect significant at P < 0.01, P + SP being significantly different from MP and MP being significantly higher than MSBBP. **Nematicide effect significant at P < 0.001 for Verticillium, Rhizoctonia, and tuber yield, but not significant for Colletotrichum.

183

ping systems. This might lead to an ecological approach of crop protection, which would be much better and durable than trying to introduce one antagonist into a complex soil-plant-microorganism system. 2. It is possible to influence such synergistic or antagonistic effects by cultural practice. 3. The level of the other inputs must be adapted to these effects and to the potential of the rotation, even though additional resources might increase the crop vigour thereby increasing the resistance against the rotational diseases. In several crops (potato, sugar beet, cereals) relevant interactions between effects of nematodes or fungi and water or nutrient supply have been shown (Darwinkel, 1980; Haverkort et al., 1989; Smit, 1996). Usually, they are consistent with the general ideas about RUE already expressed in the introduction. The mechanisms, however, are sometimes very complex and go far beyond the simple fact that a resource might make the crop less sensitive to biotic stress. In a long-term rotation experiment in progress in Perugia (Italy) since 1972, a strong effect of narrow rotation of the wheat crop was observed. The wheat yields were highest when wheat was grown after maize, whereas yields were reduced when wheat was continuously grown on the same plot. The main factor explaining the yield reduction was the so-called 'take-all syndrome' (Fusarium, Geumannomyces). This syndrome is not present every year when wheat Table 3 Take-all damage (affected stems: 0--9) on a 6-year rotation maizewheat- wheat-wheat- wheat- whe at

1st wheat b 2nd wheat 3rd wheat 4th wheat 5th wheat Continuous wheat (since 1972) R 2 in regression 'rain yield/take-all estimates'

1993

1994

1996 a

0.3 1.8 3.0 3.9 2.0 1.9

1.5 2.3 5.0 4.1 3. I 2.5

0.3 2.8 2.5 4.4 3.9 2.5

0.75

0.73

0.67

aln 1995 no damage was observed. bWheat following maize. (Source: Data from lstituto di Agronomia, Perugia.)

Table 4 Take-all attack (%) on the third cycle of wheat after different crops or fallow Previous crops (3 years before)

Lucerne Tall fescue Mixture (lucerne + tall fescue) Fallow

N-fertilisation on wheat (kg ha -I) NO

NI00

22 15 16 40

1I I 6 11

(Source: Bianchi and Bonciarelli, 1980.)

is grown, but is frequent. Its occurrence depends on the season. Table 3 shows that in the six-cycle rotation, maize followed by 5 years of wheat, damage was found in the last 4 years, except for 1995. Damage was lowest when wheat was grown immediately after maize. When wheat was cropped continuously for 5 years or more, there was a so-called 'decline' effect: the disease became less severe, because of a new balance between pathogens and their antagonists. Twothirds to three-quarters of the variation in wheat yield was accounted for by the 'take-all damage'. A late effect of rotation was described by Bianchi and Bonciarelli (1980) when wheat was repeatedly grown for 3 years on the same field, following either different crops or fallow (Table 4). During the third cycle, wheat was affected more by the take-aU syndrome when three cycles had been grown after lucerne ley or fallow than after tall fescue or a mixture of lucerne and tall fescue. Nitrogen fertilisation reduced the disease. Several studies have been carded out in Italy to investigate the effect of crop rotation on weed flora dynamics. After six cycles of a wheat-maize rotation, the total number of weed seeds in the top soil layer and the actual weed infestation were substantially reduced compared with continuous cropping of maize (Table 5). Non-chemical ways to reduce these rotational problems are the use of resistant or tolerant cultivars, organic amendments, treating or removing crop residues, adaptation of the cropping techniques thus avoiding the problem (for example by delaying sowing time) or reducing the problem by increasing the vigour of the crops, special soil tillage techniques, biological control through introducing or stimulating

184

Table 5 Number of weed seeds recorded in the top 0.15 m-soil layer and actual weed flora (in unweeded plots) after 6 years of two different rotation systems (Covarelli and Tei, 1988) Rotation

m-m-m-m-m-m w-m-w-m-w-m

Number of seeds m-"

Weeded

Unweeded

24 500 19300

55 800 i 8 920

Actual weed flora no. of weeds m-" 422 161

m, Maize; w, wheat.

antagonists, farm hygiene and growing of special crops (see below). An example of such non-chemical strategies is illustrated in Table 6. In very specific cases the build up of the population of survival structures can be reduced by haulm treatments. Scholte et al. (1996) summarised some results obtained at the Department of Agronomy, Wageningen, on the effect of removing the plant debris on the number of microsclerotia of Verticillium in the soil (Table 6). The effects visible in March 1994 were caused by treatments in 1991 and 1992: removing plant debris reduced the Verticillium inoculum of either isolate tested considerably.

4. Nutrient management strategies at the cropping system level Differences among crops and their cultivars in recommended (economically optimal) applications and nitrogen use efficiency are large; residual N is therefore very variable (for overview, see Neeteson, 1994). Residual N from commercial fertiliser, nitrogen fixation, mineralisation, deposition or organic manure will be lost or will have after-effects later in the rotation. Examples of the influence of the residual effect of N fertilisation on yield of succeeding crops are given for a long-term experiment in Italy in which wheat was grown for 3 years after different preceding crops. Table 7 shows that input of nitrogen in a 4-year ley crop of tall fescue resulted in a significant increase in the yield of the ley crop and of the succeeding unfertilised wheat crop both in the first and in the second year after the four-year ley. In the third year the effect

became negligible. Effects of nitrogen during the 4 ley-crop years were absent when lucerne was grown, both on the yield of the leguminous ley crop itself and on its after-effects. However, wheat yields were higher in the first year after lucerne than in later years. Table 8 shows grain yields of wheat grown consecutively for 2 years on the same field, as affected by two N-fertilisation levels imposed in four different, previously grown annual crops. There was an interaction between N-level and previous crop. The effect of N-fertilisation in the previous crop in improving soil fertility and thus wheat yield became already much smaller in the second year. The effect of red clover on soil fertililty, however, was still visible, at least at NO. In other experiments, the residual effect of N could only be demonstrated if excessive amounts of N were supplied to the preceding crop (Table 9). The residual N proved very sensitive to leaching on this clay-loamy soil when the winter period was wetter than usual. At the cropping system level the efficiency of nitrogen is determined by the level of input, the form and timing of input, the efficiencies of utilisation by the different component crops and the degree to which N remaining in the soil or in crop residues can be kept within the boundaries of the cropping system and can be utilised by later crops. Soil structure affects nutrient use efficiency (Van Ittersum and Rabbinge, 1997), partly by its direct effect on attainable yield, partly indirectly by its effect on root density of the crop. Efficiency is optimal when the following aims are met: 1. Maximum use of the nutrients supplied by adjustTable 6 Effects of removal of plant debris on the number of microsclerotia of Verticillium dahliae in the soil in March 1994 Isolate

PI PI FI FI

Crop sequence 1991

i 992

1993

No. cfu g-i

P PR F FR

P PR P PR

PR PR PR PR

126 51*** i 99 28***

Isolates: PI, potato isolate; FI, field bean isolate. Crop sequences: F, field bean; P, potato; R, removal of plant debris, cfu, Colony forming units. ***Significantly different from the control at P < 0.001). After Scholte et al. (1996).

185

ing the supply to the demand, by synchronisation of supply and demand, and by synlocalisation (the nutrient is available where it can be taken up). Nutrient uptake needs to be predictable to achieve this. Greenwood et al. (1990) showed that there is a close relation between above-ground dry weight in the crop and the above-ground nitrogen concentration, which is very robust over a wide range of conditions and species. Under non-limiting conditions the nitrogen demand can therefore be derived from the expected production curve. Similar relationships for other nutrients are more difficult to assess: the variation not accounted for in the case of P and K is usually much higher than in the case of N (Greenwood et al., 1980). 2. Optimal use of crop residues for the increase of soil fertility, for example by maintaining the proper C:N ratio in the soil to allow optimal rates of breakdown of organic matter. 3. Maximum reduction of emission during the periods between the main crops, e.g. by growing nutrient catch crops (see later) or by incorporating straw. 4. Proper soil tillage; soil tillage may increase the efficiency of nutrient supply, water and other resources. Microvariability in plant and soil characteristics and their interactions are crucial for a proper management of nutrients (e.g. Van Noordwijk and Wadman, 1992). Variation must be taken into account when determining agronomically optimal rates of fertiliser with minimum ecological damage. Part of the variability in for example availability of water or of nutri-

ents may persist, increase in time and interfere with other aspects of crop management. Managing variation is therefore crucial for sustainable resource management at the cropping system level (Almekinders et al., 1995). A final example of the effects of crop rotation and nitrogen supply on the long-term balances of N, but also other nutrients, i.e. P and K is given in Table 10. Continuous cropping of potato increased the need for N, P and K compared to the 1:2 rotation. Additional N increased the nitrogen surplus similarly for the two crop rotations; N was the only nutrient present in excessive amounts.

5. Nutrient management strategies per link in the crop rotation The relation between nitrogen uptake and yield is fairy fixed (see also above). For nitrogen use efficiency (defined as output per unit input) the two following crop types can be distinguished: crops without change in N-recovery (for definition see Vos, 1996b) with an increase in N-supply until the agronomically optimal level (i.e. the maximum level at which the Nrecovery is still at its best) and crops with a decrease in N-recovery with an increase in N-supply (Vos, 1996b). Beyond the agronomically optimal supply the recovery decreases with an increase in supply for both types. The type of fertiliser is relevant to the magnitude of and variation in the losses. In all cases nitrogen residues are unavoidable. The dynamics of N availability cannot be accurately pre-

Table 7 Direct effect of N-fertilisation during the ley crop on the yield of lucerne and fescue ieys and the effect of residual N observed in the first, second and third years of the following continuous wheat. Wheat crops were fertilised with 0 N N-fertilisation (kg ha -t)

DM yield (t ha -I in 4 years)

Grain yield (t ha -j) of the following continuous wheat I st year after

0 75 150 300 600

2nd year after

3rd year after

L

TF

L

TF

L

TF

L

TF

46 46 45 48 49

14 21 31 42 47

5.5 5.6 5.6 5.5 5.8

1.9 1.9 2.1 2.7 5.0

6.0 5.6 5.8 5.8 6.1

4.2 4.0 4.4 4.6 5.4

3.2 3.1 3.2 3.2 3.4

3.0 2.8 3.1 3.2 3.3

L, lucerne; TF, tall fescue. (Elaborated from Bianchi and Bonciarelli, 1980.)

186 Table 8

Table 10

Residual effect of previous crops and their N fertilisation

Nitrogen, phosphorus and potassium balances (in kg ha-~.year) in two rotations

Previous crops

Red clover NO N300 Italian rye grass NO N300 Maize NO N300 Wheat NO N300 Fallow NO N300

1st wheat (NO)

2nd wheat (NO)

1969/70

1970/71

1971/72

4.0 4.5

4.9 5.4

2.9 3.0

3.4 3.7

2.1 2.9

2.2 3.9

2.6 2.7

2.9 3.0

2.0 4.3

3.1 5.9

2.2 2.7

2.6 2.9

2.0 2.7

1.2a 1.8a

2.1 2.2

2.4 2.7

3.5 5.1

3.9 4.8

2.2 3.3

2.4 2.7

(NO, 0 kg ha-I; N300, 300 kg ha-~) on the grain yield (t ha-I) observed in the first and second years of wheat. aHeavy attack of take-all (80%) vs. negligible attacks (2-5%) following other crops than wheat. (Elaborated from Bonciarelli, 1972 and Bonciarelli and Bianchi, 1980.) dicted, not in time and not in amount; similarly crop growth and amounts of N in crop residues are still unpredictable to a large extent. Crop residues will affect the soil fertility. Depending on their C:N

Table 9 Residual N taken up after different rates of N fertilisation in preceding crops (kg ha-I) N to preceding cropsa (kg ha-t)

50 100 200 (economic rate) 400 800

Balances (kg ha-l.year)

1972/73

Nitrogen taken up by following wheat (kg ha-I) Rainy yearb

Dry yearc

0 2 7 18 36

2 12 33 80 143

aAverages of four crops: silage maize, forage maize as catch crop, sudan grass and Italian rye grass. b1968--69:635 mm rain in period October-March (152 mm more than average). ¢1969-70:323 mm rain in period October-March (160 mm less than average). (Elaborated from Bonciarelli and Monotti, 1973.)

RINI R IN2 R2NI R2N2

N

P

K

52 90 38 78

1 0 -2 -5

1 9 -35 -27

R l, Alternating a cereal and a potato crop; R2, continuous cropping of potato and at two nitrogen levels: N1, -15% recommended quantity; N2, +15% recommended quantity. Data are averages over 4 years and over two levels of input of organic matter on a sandy soil containing peat. Recommended quantity for cereal: I l0 kg ha-~, and for potato: 200 kg ha-~. Method of calculation and data are derived from Vos (1996a). ratio, soil characteristics, tillage and cropping practices and weather, the proportions of N lost or carried over to the next growing season vary considerably. The emissions can be reduced, albeit not to zero. If nitrogen emissions are kept extremely low, usually the chemical soil fertility is reduced in the long term. This may not be true for situations in which nitrogen catch crops are grown or for other nutrients.

6. Use of special crops to improve sustainability Growing of legumes (improving nitrogen and phosphorus availability); green manure crops (physical, chemical and biological soil fertility); lure, trap and killing crops (biological control or suppression); 'wake-up crops' (inducing suicidal germination or hatching); cover crops (preventing soil erosion); and nutrient catch crops (keeping nutrients available for subsequent crops) can help to improve the sustainability of the cropping system, i.e. to maintain the natural resources and the carrying capacity of the cropping system. An example of using green manure crops to increase the population of the mycophagous soil fauna, thus suppressing a soil-borne fungal disease is illustrated in Table 11. Growing oats as a green manure crop reduced the proportion of potato stems affected by Rhizoctonia stem canker, probably by stimulating the nematodes and perhaps partly by shifts in the ratios of collemboles species.

187

Table 11 Effects of the green manure crop oats on the relative number of collemboles and nematodes in the soil and on the Rhizoctonia stem canker disease index on potato in the following growing season in two field experiments Relative number of collemboles

Control Oats

1992

1994

Relative number of nematodes 1994

100 127

100 123

100 1043"

Disease index (0-100) 1992

1994

26 10"

67 51*

Field experiments lasted two years. In year 1, the soil was infested with R. solani by growing a potato crop from seed potatoes with black scurf. In the autumn of year 1 either no crop or oats was grown. In year 2, potato was planted to test the effects of the green manure. *Means significantly different from control. After Scholte et al. (1996). Even parasitic weeds can be reduced by growing trap crops that produce germination stimulants but cannot be infected by the parasite, and thus induce suicidal germination. The example is from continuous cropping of tobacco in southern India, but similar examples can be found for faba bean in the Mediterranean area. The yields of broomrape were reduced by the trap crop because the part of the seed bank that was not dormant, was lured by growing the trap crop before the tobacco was planted and could be infected (Table 12). Part of the yield advantage in tobacco may also have been caused by the green manuring effect of the trap crop. The use of nitrogen catch crops is relatively well documented. An extreme case where nitrogen catch crops are grown to prevent high N uptake by the commercial crop is found in Italy. In Umbria flue cured tobacco is an important crop on irrigated alluvial Entisols. The rather high fertility of these soils is not suitable to produce the low nicotine content which is required in that type of tobacco. To remove the excessive nitrogen from the soil before planting the tobacco, the common practice is to grow a catch crop

Table 13 Growth and N uptake by an oat catch crop before growing tobacco

Table 12 Effects of trap crops on relative above ground yield (%) of the parasitic weed broomrape (Orobanche cernua) and relative economic yield of tobacco. Data derived from Dhanapal and Struik (1996) Trap crop Sunhemp (Crotalaria juncea L.) Redgram (Cajanus cajan L.) Millsp.) Sunflower (Helianthus annuus L.) Fallow

of forage oats sown in fall (October) and mown and removed in April-May, just before transplanting the tobacco. Table 13 gives some unpublished results of a series of samplings and analyses of both oat biomass production and nitrogen uptake. More often nutrient catch crops are grown to prevent nitrate leaching. An example, showing the potential for the Netherlands, is presented in Table 14, based on data of Vos (1996b). The performance of nutrient catch crops strongly depends on the sowing time, because temperature and light intensity decline rapidly in autumn, whereas the chances of excessive water and killing night frosts increase. For all special crops grown to improve sustainability one rule is important: they have to fit in the sequence of main crops and should not interfere with necessary soil tillage. They even may facilitate soil tillage by reducing the soil water content in early spring. Especially their response to light and temperature in dependence of sowing date and their effects on water availability need further research to optimise their use.

Broomrape Tobacco 17

33 51 100

173 157 129 100

Sampling date (1996)

DM (t ha-I)

N uptake (kg ha-j)

March

1.0 1.7 2.4 4.4 5.7 7.6 8.0

26 39 53 68 84 86 86

April

May

1 15 29 12 19 26 3

Unpublished data, Istituto di Agronomia, Perugia.

188

Table 14 Average benefit of nutrient catch crops (average of different species) in a cropping system Previous crop

Reduction in N-loss from the system (g m-2)

Oats Spring wheat Potato Sugar beet

2.6 3.9 3.1 0.4

Catch crops were sown as soon as possible after the harvest of the main crops. Losses were for the period autumn year X until spring year X + 1. Based on Vos (1996b).

7. Some final remarks Tools are strongly needed to allow analytical studies on the effects of cropping system management on • • •

the productivity of each crop in the rotation; the environmental risks; and the stability, resilience and durability of the cropping system.

Investigations into options to maintain a short rotation of a crop with low self-tolerance by making use of non-chemical strategies to avoid yield-reducing conditions are also required. Ecological approaches may aim at reducing rate of multiplication of the pathogen, stimulating its antagonists, or both. In practice, variation in RUE is strongly influenced by differences in 'farming styles' among farmers, even under similar environmental conditions and financial returns (Van der Ploeg, 1990). Apparently, not only knowledge, but also skill and motivation to apply it, are relevant.

References Almekinders, C.J.M., Fresco, L.O. and Struik, P.C., 1995. The need to study and manage variation in agro-ecosystems. Neth. J. Agric. Sci., 43: 127-142. Bianchi, A.A. and Bonciarelli, F., 1980. Effect r6siduei de fertilisation azot6e appliqu6e ~t des prairies pures et mixtes de luzerne et de f6tuque 6lev6e. S6minaire Agrimed 'M6thodologie d'l~tude des Syst~mes de Culture', Toulouse, 7-9 May 1980. Bonciarelli, F., 1972. Trials on crop sequences and N fertilisation. Rivista Agron., 1972, VI(1): 44-48. Bonciarelli, F. and Bianchi, A.A., 1980. Evaluation de la fertilitd6

r6siduelle de diff6rentes cultures soumises h diff6rents niveaux de fertilisation azotEe. S6minaire Agrimed 'M6thodologie d'l~tude des Syst~mes de Culture, Toulouse, 7-9 May 1980. Bonciarelli, F. and Monotti, M., 1973. Residual effect on wheat of different rates of nitrogen applied to annual forage grasses. Rivista di Agron., VII, (2-3): 150-158. Covarelli, G. and Tei, F., 1988. Effect de la rotation culturaie sur la flore adventice du ma'fs. 8~me Colloque International sur la BioIogie, l'Ecologie et la Systematique des Mauvaises Herbes, Dijon, pp. 477-484. Darwinkel, A., 1980. Grain production of winter wheat in relation to nitrogen and diseases. I. Relationship between nitrogen dressing and yellow rust infection. Zeitsch. Acker. Pflanzen., 149: 299-308. De Wit, C.T., 1992. Resource use efficiency in agriculture. Agric. Systems, 40:125-15 I. Dhanapai, G.N. and Struik, P.C., 1996. Broomrape control in a cropping system containing bidi tobacco. J. Agron. Crop Sci., 177: 225-236. Greenwood, D.J., Barnes, A., Liu, K., Hunt, J., Cleaver, T.J. and Loquens, S.H.M., 1980. Relationships between the critical concentrations of nitrogen, phosphorus and potassium in 17 different vegetable crops and duration of growth. J Sci. Food Agric., 31: 1343-1353. Greenwood, D.J., Lemaire, G., Gosse, G., Cruz, P., Draycott, A.A. and Neeteson, J.J., 1990. Decline in percentage of N of C3 and C4 crops with increasing plant mass. Ann. Bot., 66: 425-436. Haverkort, A.J., Vos, J., Groenwold, J. and Hoekstra, O., 1989. Crop characteristics and yield reduction of potato due to biotic stress in short rotations. In: J. Vos, C.D. van Loon and G.J. Bollen (Editors), Effects of Crop Rotation on Potato Production in the Temperate Zones. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 273-290. Neeteson, J.J., 1994. Residual soil nitrate after application of nitrogen fertilizers to crops. In: D.C. Adriano, A.K. lskandar and I.P. Murraka (Editors), Advances in Environmental Sciences. Contamination of Groundwaters. Science Reviews Ltd., Northwood, pp. 347-365. Nijland, G.O., Schouls, J. and Oomen, G.J.M., 1997. The Relation Between Nutrient Application, Nutrient Uptake, Production and Nutrient Residues. Wageningen Agricultural University Papers 97-3, Wageningen (in press). Perucci, P., Bonciarelli, U., Santilocchi, R. and Bianchi, A.A., 1997. Effect of rotation, nitrogen fertilization and management of crop residues on some chemical, microbiological and biochemical properties of soil. Biol. Fertil. Soils (in press). Rabbinge, R., Diepen, C.A. van, Dijsselbloem, J., Koning, G.J.H. de, Latesteijn, H.C. van, Woltjer, E.J. and Zijl, J. van, 1994. Ground for choices: a scenario study on perspectives for rural areas in the European Community. In: L.O. Fresco, L. Stroosnijder, J. Bouma and H. van Keulen (Editors), The Future of the Land; Mobilizing and Integrating Knowledge for Land Use Options. Wiley, Chichester, UK, pp. 95-121. Scholte, K., 1989. Effects of Crop Rotation on the Incidence of Soil-Borne Pathogens and the Consequences For Potato Production. Doctoral Thesis, Wageningen Agricultural University, Wageningen, The Netherlands, 143 pp.

189

Scholte, K., Mol, L. and Lootsma, M., 1996. Control of Verticillium dahliae and Rhizoctonia solani by cultural practices. In: Abstracts of Conference Papers, Posters and Demonstrations, 13th Triennial Conference of the European Association for Potato Research, Veldhoven, The Netherlands, 14-19 July i 996, pp. 134-135. Smit, A.B., 1996. PIEteR: a Field Specific Bio-economic Production Model for Decision Support in Sugar Beet Growing. Doctoral Thesis, Wageningen Agricultural University, Wageningen, The Netherlands, 201 p. Stockle, C. and Nelson, R., 1994. CropSyst User's Manual. BSE Dep, WSU, Pullman, USA, 186 p. Van der PIoeg, J.D., 1990. Heterogeneity and styles of farming. In: J.D. van der PIoeg (Editor), Labor, Markets and Agricultural Production. Westview Press, Boulder, pp. 1-36. Van Evert, F.K. and Campbell, G.S., 1994. CropSyst: a collection of object-oriented simulation models of agricultural systems. Agron. J., 86:325-331. Van Ittersum, M.K. and Rabbinge, R., 1997. Concepts in production ecology for analysis and quantification of agricultural inputoutput combinations. Field Crops Res., 52: 197-208.

Van Noordwijk, M. and Wadman, W.P., 1992. Effects of spatial variability of nitrogen supply on environmentally acceptable nitrogen fertilizer application rates to arable crops. Neth. J. Agric. Sci., 40: 51-72. Vereijken, P., 1995. Designing and testing prototypes. 2nd Progress Report of EC Concerted Action AIR3-CT920755, Wageningen, The Netherlands, 90 pp. Vos, J., 1996a. Input and offtake of nitrogen, phosphorus and potassium in cropping systems with potato as a main crop and sugar beet and spring wheat as subsidiary crops. Eur. J. Agron., 5:105-114. Vos, J., 1996b. Nitrogen cycle related to crop production in cool and wet climates. In: R. Samulesen, B. Solsheim, K. Pithan and E. Watten-Melvaer (Editors), Crop development for the cool and wet regions of Europe. Nitrogen supply and fixation of crops for cool and wet climates. Proceedings COST 814 Workshop, Troms~, September 7-9 1995. Office for Official Publications of the EC, Luxembourg, Luxembourg, pp. 3-14.


(~ 1997 ElsevierScience B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

191

The efficient use of solar radiation, water and nitrogen in arable farming" matching supply and demand of genotypes A.J.

Haverkort ~,*, H. van Keulen ~, M.I. Minguez b

a DLO-Research Institute for Agrobiology and Soil Fertility (AB-DLO), P.O. Box 14, 6700 AA Wageningen, The Netherlands b Depto Produccion Vegetal: Fitotecnia. E.T.S. lngenieros Agronomos, Universidad Politecnica de Madrid, Ciudad Universitaria, 28040 Madrid, Spain Abstract

The length of the growing season of a crop depends firstly on the suitable period of the year when temperatures allow growth of a particular crop. The amount of intercepted solar radiation then determines the potential dry matter production. The length of the season, however, may be reduced by lack of water. The resulting yielding ability determines the nitrogen requirement of the crop. Water and nitrogen uptake may be increased through management strategies that increase water use and improve water use efficiency in semi-arid environments. This paper shows the implications of the length of the available growing season and amount of available water for the desired genotypic characteristics of a species. Subsequently, the amount of nitrogen needed to reach the highest yields provided best use is made of solar radiation and available water is discussed in quantitative terms. Finally, recent research developments are discussed such as how the optimal temporal distribution of nitrogen application can be derived from periodical assessment of its availability in the soil and more importantly from observations and calculations of amounts of nitrogen already taken up and expected nitrogen uptake till harvest. Keywords: Resource use efficiency; Solar radiation; Water use efficiency; Nitrogen use efficiency; Ideotyping; Plant breeding

1. Introduction

Consumers, producers and scientists are aware that agricultural production should optimize the use of natural resources and minimize emissions to the environment. Extraction of ground and surface water and its pollution with nitrate pose serious risks to the natural environment. To obtain potential yields (all minerals and water needed is supplied and no (a)biotic factors present) in a given environment, resources, notably water and nutrients should be applied at rates which are both economically and environmentally unacceptably high. Attainable yield (all inputs supplied at levels that are economically justified) by farmers, may be unsustainable at the long term. To realise the food and export require* Corresponding author.

ments for Europe on the whole, groundwater is threatened because of contamination with nitrogen and because of depletion of this resource. This is a serious risk for natural ecosystems. A growing proportion of food, following consumer demands is produced without chemical fertilizers and biocides but such attainable organic yields are lower than current yields and their resource-use efficiency is amenable to debate (de Wit, 1992). Yields can be expressed in terms of resource availability (water (W), solar radiation (R) and nitrogen (N)) and resource-use efficiency (E), i.e. grams of dry matter produced per unit resource used by the crop, e.g. per g water (WUE), per joule intercepted solar radiation (RUE) or per g of nitrogen taken up by the crop (NUE). Fresh crop yield (Y) is then expressed as Y = R x R UE x H I / D M C

(1)

192

Y = W × WUE × HI/DMC

(2)

Y = N X NUE X HI/DMC

(3)

where HI is the harvest index and DMC is the dry matter content of the harvested produce. The greater the availability of a resource for single factor experiments, the lower its efficiency. When more inputs are increased simultaneously, efficiency of both may increase. De Wit (1992) stated that a production factor which is in minimum supply contributes more to production, the closer other factors are to their optimum. Strategic research should be into the identification of the minimum of each production factor needed to allow maximum utilization of all other resources. Temperature, solar radiation and crop species or cultivar are the main growth defining factors determining the lengths of the available growing season and actual growth cycle and hence the potential yields. Actual yields are mainly limited by the availability of water and nutrients, especially nitrogen and by pests, diseases and weeds. Advances have been made in maximizing the availability of resources and increasing resource use efficiencies through selection of crop species (C4/C3plants) and breeding for adaptation to adverse conditions such as drought and high temperatures. Crop management practices, however, matching crop cycles with periods of low evaporative demand and concentration of limiting resources such as strategic irrigation and application of fallow (Loomis and Connor, 1992) and organic farming techniques may have great impact as well. The most efficient use of water and nitrogen is realized by supplying them in crop management aimed at yields that are close to the potential. Then other (a)biotic constraints are reduced as much as possible and provided that the financial rate of return of each unit of water and nitrogen added is still positive. Environmental (water conservation and contamination) constraints and market demands (for products from organic farming), however, often require supplies below the economic rate. The aim of this paper is to highlight and integrate recently developed promising quantitative approaches in optimizing light, water and nitrogen

use efficiencies and to discuss how these approaches may be made operational in arable farming. Firstly we illustrate that the length of the growing season of any crop is determined by its temperature requirement that then determines during which part of the year the crop can be grown. The intercepted solar radiation during that part of the year determines the potential amount of dry matter that can be produced. Genotypes able of reaching such potential yields should have a length of the growth cycle matching that of the growing season. Secondly it is shown that attainable yields under water limited (rainfed) conditions are lower and genotypes (or species) should have a shorter length of a growth cycle than under irrigated conditions. The trade off between lower yields due to reduced solar radiation in parts of the year but increased yields due to improved water use will be illustrated. Finally, the nitrogen uptake of the crop depends on the yields as dictated by light interception and water use. It will be shown that water should not be limiting if optimal use of nitrogen is required. Split dose application based on soil and newly developed crop observation techniques may offer the best scope for improved used of this resource.

2. Radiation interception and radiation use efficiency

Breeding creates genotypes with the highest economic yield, with a specific quality for a specific environment. The specific environment is mainly determined by temperature. In environments with mean monthly temperatures below 5°C or above 25°C, for instance, potato normally is not grown commercially. In subtropical regions rice is often grown during the warmest part of the year and potato (or wheat) during the cooler part of the year. This is illustrated in Fig. 1. In the Mediterranean area the growing season is limited in spring by low temperatures at planting and by high temperatures and drought towards harvest. In northern Europe low temperatures limit the length of the growing season in spring (late frosts in March, April and May but also at the end of the season temperatures are low again (early frosts in September and October). Beside temperature, other factors may determine

193

T, R, DL

•

ooe •

qbo e e

I

•

°ee e oQ

ot

•

l

•

I,

'

;potato,'-; ! e

time

I

rice------), |

0

Fig. 1. Schematical representation of the growing seasons of potato and rice related to the temperature and radiation regimes in a subtropical area. T =temperature (drawn line), R = solar radiation and DL = daylength (broken line).

the available length of the growing season. In tropical highlands, for instance, temperatures are suited for potato production throughout the year. Rainfall at the equator, however, is in two main rainy seasons six months apart, necessitating two crops per year. Theoretically this would restrict the length of the growing season to 6 months for ware potato crops as part of this crop is used as seed for the next crop because no separate seed potato production system

exists at the farmers' level. But practically this period has to be reduced by 2 to 3 months because the tubers that are harvested, start to sprout again after two to three months only. Ideal genotype (ideotypes) should then have a length of the growing cycle of 100 days. Longer cycles would lead to an imbalance of the growing and seed rest periods. In temperate climates rainfall often determines the workability of the soil and hence planting or sowing and the harvest period. Another important factor that may limit the length of the growing season and consequently the length of the growth cycle is related to market prices. Prices are often higher at the beginning of the harvest period of crops that are harvested fresh, necessitating a reduction of the length of the season. Ideotypes have a length of the growth cycle characterized by a green leaf area that maximizes interception of solar radiation during the available growing season to accumulate as much dry matter as possible. Earlier genotypes, too early divert dry matter to the harvestable parts (grains or tubers) so that not sufficient assimilates are available for the foliage that then senesces and dies. Genotypes that are too late still have full ground cover with green leaves at the end of the available growing season which is indicative of an unfavourable distribution of dry matter to the foliage and to the harvestable parts Yield (g/m2)

1SO0 Ytolal

GROUND COVER 100%

,

, ,,

.,,

,,

tulm¢

1.s to 2.s)

0%

, •

PI

Emer. Full cover

,.,

,.,

_

i

Senem:ence

1000 MJ/m2 INTERCEPTEO RADIATION

Fig. 2. Schematical representation of tuber production in potato based on ground cover (left) going from 0 (between planting (Pl) and emergence (Emer) to 100% until the onset of senescence. The slope of total (Ytotal) and tuber (Ytuber) dry matter production in the relationship between yield in g per m 2 and intercepted solar radiation in MJ per m 2 is the light use efficiency (E).

194

of the crop. To identify ideotypes with the desired length of the growth cycle first the length of the available growing season is determined as it is restricted by adverse growing conditions or market demands. Secondly, an assesment is made of the yield determining factors (temperature, radiation, daylength and cultivar) that cannot be changed nor influenced by the farmer once the crop is planted, with emphasis on the influence of such factors on the length of the growth cycle. A simple model describing growth and development of crops is based on light interception, utilization of light to produce dry matter, allocation of dry matter to the harvestable parts and of the percentage of water in the harvestable parts. Haverkort and Kooman (1997) descibed the use of crop growth modeling in breeding for genotypes with the aid of such a LINTUL model" Light Interception and Utilization of Light, based on the principle that the amount of light that is intercepted by the crop is converted into crop dry matter through a conservative light use efficiency (Equation 1). This principle is illustrated in Fig. 2 for the potato crop. The dry matter distribution pattern as influenced by temperature and photoperiod between foliage and harvestable produce determines the length of the growing season and is used to genetically match the length of the growth cycle with that of the growing season.

3. Water uptake and water use efficiency

Even in a region where it rains relatively abundantly, such as in north western Europe, lack of water is one of the factors most limiting growth and quality of crops (Haverkort and Goudriaan, 1994). The precipitation deficit during the growing season from April through September, in the Netherlands for instance, usually exceeds 100 mm. Many soils do not have sufficient water storage capacity within the rooted zone to cover this deficit if crops are not irrigated. Irrigation is often not possible because of its high costs or because of legal restrictions associated with its extraction. In southern Europe similar observations apply during the actual growing seasons in winter and spring of annual crops such as potatoes and wheat. These crops are harvested be-

fore the highest temperatures and levels of evaporative demand are reached. The efficient use of the available amount of water is of increasing concern to maintain production and quality. A dry spell at the beginning of the growing season leads to retardation of emergence and early growth. A short transient drought period in the course of the growing season, however, may only slightly reduce growth but it may strongly affect crop development and quality of the produce. A terminal drought that intensifies in the course of the growing season (Mediterranean) and long dry spells in the second half of the growing season (temperate) have the greatest influence on yield. These kind of droughts are more frequent than early drought as crop transpiration increases and water pools from the winter period become exhausted. Crops then form fewer new leaves and accelerate leaf shedding leading to a premature senescence In a series of experiments on light sandy soils with four potato cultivars which were either irrigated or not, periodic harvests were carried out to determine the value of the yield components given in Equation 1. The water stress occurred after tuber formation between mid June and mid August i.e. a long period of transient drought. The results are shown in Table 1. The losses due to drought were highest for the relatively early cultivar Darwina (-45%) and smallest for the latest cultivar Elles (-20%). Yield losses were mainly due to a reduction of intercepted radiation because of earlier senesecence. The light use efficiency was affected less than intercepted radiation. The cultivar that suffered little losses of intercepted radiation showed a stronger reduction in light use emciency (Elles -10%) than cv. Darwina (-1%) which reacted mainly to water stress by leaf shedding. About 75% of the total amount of dry matter produced by the crop, is found in the Table 1 Relative values of yield components (Equation l) of rainfed versus irrigated (=100%) plots for four potato cultivars of increased lateness (Haverkort et al., 1992) Cultivar

Y

R×

RUE ×

HI/

DMC

Desiree Darwina Mentor Elles

77 55 73 80

88 62 87 93

99 99 97 90

94 94 97 95

105 105 111 101

195

tubers by the time of crop senescence (HI = 0.75). Drought reduced the harvest index only slightly and no differences among cultivars were found. The tolerance of the cultivar Elles for drought lies in its abundant formation of foliage associated with its lateness. Whereas drought reduced its light use efficiency, the cultivar effectively overcame relatively long periods of absence of precipitation. Making use of the principle shown here Haverkort and Goudriaan (1994) quantitatively demonstrated through modeling that early cultivars of potato obtain higher yields than late cultivars in Mediterranean conditions where it does not rain in summer whereas late cultivars obtain the highest yields in temperate climates with a modest precipitation deficit in the middle of the growing season. For production of fodder, producers, beside quality, may choose from crops depending on water availability and timing of the harvest. Table 2 shows the water-use efficiency of several fodder crops. Maize seems to be a very efficient user of water expected on the basis of its assimilatory (C4) pathway. The low efficiencies of lucerne are associated with reduced dry matter accumulation at the cost of fixing nitrogen. Grass has a low water use-efficiency and is further hampered by an imperfect stomatal regulation and a low harvest index as grass partitions a relatively great amount of dry matter to its roots. Fodder beet has two advantages over many other crops" it has a low transpiration coefficient and a high harvest index. There must be other reasons than resource use efficiency why the bulk of fodder in Europe does not consist of fodder beet but rather of grass. These are related to the advantageous regular supply, protein quality and storability of grass compared to beet. Table 2 shows the water-use efficiency of maize to be about 6 g harvestable crop dry matter per litre of water used. It is, however, important to supply the crop with water at a crucial moment during its development. For potato this is during tuber formation. Haverkort et al. (1990) showed that following a short dry period at tuber initiation, yields were not much decreased but the number of tubers was decreased considerably exerting a strong influence on the quality (size grading) of the produce. Sink capacity in potato apparently did not limit growth in the range considered. For maize the situation is dif-

Table 2 Water-use efficiencies (g dry matter harvestable produce per litre of water used by the crop) of some fodder crops (after Aarts et al. 1996) Crop ryegrass lucerne silage maize fodder beets

optimal water suply droughtedconditions 1994

1995

1994

1996

2.86 2.16 6.02 4.55

2.99 1.42 6.29 4.44

2.78 2.10 5.49 4.55

2.69 1.66 6.33 4.72

ferent. When water is not sufficiently available during cob initiation, cob density is low resulting in limited sink capacity that through its feedback to assimilation leads to a reduction of growth far more than proportional to reduced water supply (Artlipp et al., 1995). When a particular resource is in short supply, its efficiency increases. The radiation use efficiency (Equation 1) and the water use efficiency (Equation 2) increase when light respectively water are in minimum supply. So a high resource efficiency is associated with the short supply of the resource. Therefore the best strategy for a high yield is to increase the availability of the resource. Increasing water use under rainfed conditions consequently is better achieved by optimizing water uptake than the water use efficiency. In Mediterranean conditions this can be achieved by matching the crop cycle with the rainy season which has a lower evapotranspirative demand, and by allowing a high proportion of water to be spent in transpiration with minimum losses to evaporation, drainage and runoff. This has led to earlier sowing dates, certain management practices and to selection of cold-resistant resistant cultivars. The effects of these strategies on yield, evapotranspiration (ET) and water use efficiency (WUE = Yield/ ET) are shown for field-grown sunflower in southern Spain (Gimeno et al., 1989) and for simulated rainfed faba beans in central Spain (Diaz-Ambrona et al., 1996) (Table 3). Sunflower can be sown early in winter in warmer regions, while in the case of colder areas in central Spain, sowing dates should be earlier in order to allow for crop establishment when soils are still warm, go through the winter in a cold resistant growth stage, and rapidly cover the soil in April. Under terminal water deficits or terminal drought,

196

Table 3 Yield, evapotranspiration (ET) and water use efficiency (WUE expressed as yield in kg per ha per mm rain fallen during the growing season) of winter- and spring-sown sunflower and faba beans in southern Europe under rainfed conditions Crop

Sowing date

Yield (kg ha -1)

Seasonal ET (ram)

WUE (kg ha -1 mm -1)

Sunflower

December 15 March 15 October 31 November 30 February 1

3050 640 3491 2408 1025

481 372 272 269 259

6.34 1.72 12.80 8.95 3.96

Faba beans

to apply the nitrogen at adequate quantities when the crop needs it. The efficiency of the use of an external resource is highest when all other resources are applied at their optimal level. Therefore it is essential that water does not limit growth if optimal use of nitrogen is required. Seligman and van Keulen (1981) described a model PAPRAN (Production of Arid Pastures limited by Rainfal And Nitrogen) including a quantitative description of the water and nitrogen balance in the soil, their availability to the crop (a mixture of annuals) and their effect on crop growth. The model was validated with data collected between 1971 and 1981 in Israel. Results of simulation studies, in which the interactions between water and nitrogen availability were examined, showed that the response to these two factors in terms of dry matter production could not be described simply by Liebig's 'law of the

an earlier sowing date may increase water use during the vegetative growth in detriment to the grain filling period. It would be necessary to match the length of the phenological stages to the pattern of water availability, especially under terminal drought.

4. Nitrogen uptake and nitrogen use efficiency Nitrogen fertilization is of importance for a number of reasons. As yield is much related to the amount of nitrogen available to the crop, and to avoid the risk of sub-optimal fertilization, farmers tend to over supply the crop. Excess nitrogen (from previous seasons, from mineralization and from application) may lead to excess nitrogen remaining in the soil at harvest, susceptible to leaching in the winter. These tendencies make it increasingly important

• 0 kg • 30kg. o 60kg o 90kg 6 1 2 0 kg 150 kg

primary production ( kg ha -I )

6000 -

N ha "1 . .... . . ... .. .

A

m e a n annual r a i n f a l l , 132mm •264mm primary production o 396mm ( kg ha " ! ) • 52 7 m m

6000

2 4 kg h a " m m "

/

sooo -

i

/ /

~ /

: ~

sooo

o

///logo 4000

-

~

~

4000

r-"---" ///.~,//s

3000"

2000 1000

0

3000

~..-"'~'~'--"

2000~r/

// 100

1000 ! ~ • ~ . _ o . _ _ _ . ,

200

300 400 500 600

m e a n annual r a i n f a l l

(ram)

[

0

..

i

30

i _

60

nitrogen

-_ I

90

,

I

120

• i

150

opplicOtaon ( kg hO"1 )

Fig. 3. Simulated response of mean primary production over a 21-year period to mean annual rainfall at different levels of nitrogen nutrition (A) and to nitrogen nutrition at different mean annual rainfall levels (B). In both cases soil depth is 1.5 m.

197

minimum' (Fig. 3). The maximum efficiency of rainfall utilization (kg dry matter ha -1 mm -~ rainfall) was 24 at high nitrogen availability, and declined to about 10 in the situation were nitrogen was the main limiting factor. Maximum nitrogen use efficiency (kg dry matter/kg N) was 40, at high rainfall, which is close to the biological maximum for C3species, whereas at low rainfall (hence where water is the main limiting factor) the value was only 5. In the intermediate range of nitrogen and water availabilities, the two factors appeared to be limiting at different times during the growing season, and the effect shows an interaction between the two factors. However, in all cases it is clear that the efficiency of utilization of one resource increases when the supply of another resource is closer to its optimum level. Crop nitrogen content increases with crop age and the amount of dry matter accumulated although its concentration decreases. The required (by the crop) nitrogen supply from the soil (from fertilizer and from mineralization), at any moment, depends on the amount of nitrogen already present in the crop, the expected amount still to be taken up until harvest, and the availability of mineral (mainly nitrate from mineralization of organic matter) nitrogen in the soil. Lack of information on the crop nitrogen content when sampling the soil is a bottleneck in the development of nitrogen fertilization expert systems. Apart from massive destructive sampling, presently no alternative methods are available to assess the crop nitrogen content. Reflectance and gas exchange characteristics may yield information on single leaves but research needs to find an answer to the question LAI

Brussels sprouts

/

rZ=O'913

4+

=

how to non-destructively determine the crop nitrogen status. Study of crop nitrogen uptake using crop ecological principles and methodologies, systems analysis and modeling in relation to genetic and environmental conditions is needed to implement expert systems aimed at optimizing nitrogen fertilization. Split dose application of nitrogen is becoming increasingly popular in crop production. Part of the total expected nitrogen requirement of the crop is then applied before or at planting and another part is applied later during the crop cycle. An advantage of split application, compared to one single application before planting is that this practice reduces the environmental risk of losses through leaching, volatilization and immobilization and it offers the financial possibility to reduce the total amount to be applied. Reduced nitrogen application aimed at adjusting application rates to crop needs reduces emission of nitrate and ammonia to the environment and enhances the environmental friendliness of crop production. Basic research is needed to develop methodologies for the assessment of crop nitrogen contents at the same moment that the soil is sampled for the amount of mineral nitrogen still present in the soil. Based on these two data a more sound advice for supplemental nitrogen fertilization can be given than based on soil mineral nitrogen only. The central hypothesis is the general applicability of the relationship between crop nitrogen content and its leaf area index (LAD. The leaf area index of the crop at the moment of sampling seems to be related to the total amount of LAI

" f

|

°y"

Leeks r2=0"975

4

•

I i Ig

2

2

0

" 0

0 100

200

N-uptake (kg/ha)

300

0

100

200

300

N-uptake (kg/ha)

Fig. 4. Relationship between leaf' area index (LAD and total amount of nitogcn taken up by the crop in leeks (Booij et al., 1996. Different symbols represent different N-fertilizer levels.)

198

The suitable temperature range for a crop The suitable period for crop, potential yield calculated from intercepted solar radiation Water limited length of the available growing season Attainable total dry matter yield and harvest index Ideotype identification with length of the cycle matching that of the season Nitrogen requirement of the crop calculated from water limited yields Split nitrogen application depending on availability in soil (sampling) and crop (sensing) Fig. 5. nitrogen taken up by the crop (Fig. 4). This recent finding may offer elegant possibilities to non-destructively asses the amount of nitrogen present in the crop. The expected total uptake leads to the amount that is still required. The crop requirement should be present in the soil or be applied. Two methods of LAI (thus crop nitrogen content) assessment are presently under investigation. The first is through modelling as it is known that the relative leaf extension rate as a function of temperature has a conservative value. Emergence date and the initial amount of leaf area at emergence are then crucial data. A second way is through non-destructive measuring with the aid of infrared reflectance, The second method probably will approach the LAI-value better than the first method.

5. Conclusions

In this paper it is illustrated that to optimally use solar radiation, water and nitrogen in a given environment, the following conditions must be met. The genotype used should have a length of a growth cycle (from emergence to senescence) that matches the length of the growing season. The season firstly is determined (limited) by temperature and secondly it may be further reduced when water is a limiting factor in (partly) rainfed conditions. Once the season has been defined, water limited yields can be calculated and the proper degree of cultivar lateness can be defined. Subsequently the crop nitrogen need is

calculated and the nitrogen supply at any moment is synchronized with future need until harvest. The procedure is schematically represented in Fig. 5. The approach described in this paper, i.e. matching of supply and demand of water and nitrogen of genotypes that have a length of a growth cycle that matches that of the growing season, is inspired by and based on research papers and current research. It is expected that it will assist in the need in modern agriculture to substitute inputs by knowledge such as to increase the efficient use of scarce or potentially environmentally harmful substances.

References

Aarts, H.F.M., Grashoff, C. and Smid, H.G., 1996. Evaluation of perennial ryegrass, lucerne, silage maize and fodder beets under drought. Grassland and land use systemsG. Parente, J. Frame and S. Orsi (Eds.). Proceedings of the 16th meeting of the European Grassland Federation: 363-366. Artlipp, T.S., Madison, J.T. and Settler, T.L., 1995. Water deficit in developing endosperm of maize; cell division and nuclear DNA edoreduplication. Plant Cell Environ., 18, 1034-1040. Booij, R, Kreuzer, A.D.H., Smit, A.L. and van der Weft, A.K., 1996. Effect of nitrogen availability on light interception, dry matter production and nitrogen uptake of Brussels sprouts and leeks. Neth. J. Agric. Sci., 44: 3-19. De Wit, C.T., 1992. Resource use efficiency in agriculture. Agric. Syst., 40: 125-151. Diaz-Ambrona, C.H., Conde, J.R., Hoyos, P. and Minguez, M.I., 1997. Simulation of water consumption in a cereal-legume rotation in a Mediterranean environment. International Congress of Agricultural Engineering. Madrid, September 1996. (in press.)

199 Gimeno, V., Fern~indez-Martinex, J.M. and Fereres, E., 1989. Winter planting as a means of drought escape in sunflower. Field Crops Res, 22: 307-316. Haverkort, A.J., Boerma, M., Velema, R. and Van de Waart, M., 1992. The influence of drought and cyst nematodes on potato growth. 4. Effects on crop growth under field conditions of four cultivars differing in tolerance. Neth. J. Plant Pathol., 98: 179-191. Haverkort, A.J. and Goudriaan, J., 1994. Perspectives of improved tolerance of drought in crops. Asp. Appl. Biol., 38: 79-92. Haverkort, A.J. and Kooman, P.L., 1997. The use of systems analysis and modelling of growth and development in potato ideotyping under conditions affecting yields. Euphytica (in press).

Haverkort, A.J., van de Waart, M. and Bodlaender, K.B.A., 1990. The effect of early drought stress on numbers of tubers and stolons of potato in controlled and field conditions. Potato Res., 33: 89-96. Loomis, R.S. and Connor, D.J., 1992. Crop Ecology: productivity and management in agricultural systems. Cambridge university press, Cambridge, 538 pp. Seligman, N.G. and van Keulen, H., 1981. PAPRAN: A simulation model of annual pasture production limited by rainfall and nitrogen. In: M.J. Frissel and J.A. van Veen, Eds. Simulation of nitrogen behaviour of soil-plant systems. Pudoc, Wageningen pp. 99-121.


© 1997 ElsevierScience B. It. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geij'n (Editors)

201

Soil-plant nitrogen dynamics" what concepts are required? E.A. Stockdale a'*, J.L. Gaunt a, J. V o s b aSoil Science Department, IACR-Rothamsted, Harpenden, Herts., AL5 2JQ, UK bDepartment of Agronomy, WageningenAgricultural University, P.O. Box 341, 6700 AH Wageningen, The Netherlands


Abstract

Soil-plant N dynamics lie at the heart of some of the questions being asked of researchers by farmers, environmentalists and policy makers. Our aim in this paper is to highlight areas in which research is needed to address these questions. Although we have a general understanding of many processes, fundamental understanding of the processes of the soil-plant system is not complete. Improved understanding of crop and soil processes should lead to the continuous improvement of simulation models, which are able to integrate the complex effects of management and environmental factors. However, farmers cannot wait for the achievement of perfect models, but need researchers to put their current knowledge to use. We suggest that for both crops and soils, diagnostic measurements be used in conjunction with the best of, or a combination of, current models. Development work should be carded out with all possible speed to draw suitable models and diagnostics together. We stress the importance of conducting research to understand and improve N efficiency at a range of scales and indicate the need for the involvement of related disciplines, such as statistics, to allow the development of robust guidelines and methodologies for up- and down-scaling measurements and models. Above all, it is essential that our understanding of the processes of soil-plant dynamics continues to underpin the development of strategies for dynamic optimisation and improve simulation models that are used in fertiliser recommendation systems. © 1997 Elsevier Science B.V. Keywords: Simulation models; Diagnostics; Dynamic optimisation; Scaling

1. Introduction

Research into soil-plant nitrogen (N) dynamics has been carried out from the very beginnings of agricultural investigation (Russell, 1966). The soil-plant N cycle is composed of a large number of complex and interacting processes which transform and transport nitrogen in, out and throughout the soil-plant system (Fig. l). Within the disciplines of both plant and soil science, countless experiments have been performed * Corresponding author. Tel.: +44 1582 763133, x 2665; fax: +44 1582 760981; e-mail: [email protected]

in laboratory and field to elucidate the mechanisms of N cycle processes and their interactions with one another and the environment, as examination of the literature rapidly shows. Recently mathematical modelling has begun to integrate our understanding of the soil-plant N cycle and the soil, plant, environmental and management factors which govern it. Models draw together current knowledge and hypotheses about biological systems, their subsystems and interactions and are able to integrate scattered data into a coherent whole (Jenkinson, 1990). However, the complexity of the cycle and the large number of interacting factors which control it means that, even for


202

temperate agricultural soils, our models do not closely approach reality (de Willigen, 1991). Soil-plant N dynamics lie at the heart of some of the questions being asked of researchers by farmers, environmentalists and policy makers. Farmers seek to apply economic optimum rates of fertiliser, considering the costs of application and the effect on crop quality as well as yield (Neeteson and Wadman, 1987; Vos, 1995). Farmers therefore seek answers to questions about components of their system, e.g., the amount of N released for a following crop when a ley is ploughed or the fertiliser value of their manure. Environmental concerns are focused on nitrogen losses from soils which may pollute the environment. Leaching is the major route by which nitrate enters ground and surface waters, while denitrification and nitrification are significant sources of N20, an important greenhouse gas (Royal Commission on Environmental Pollution, 1996). Improved efficiency of N use at a field and farm scale, both increasing crop yield and quality and reducing losses, is dependent upon dynamic optimisation to match supply of N and the N requirements of the crop at a field scale. This optimisation requires measurement and prediction of soil N supply, crop uptake and their variability (Vos and Marshall, 1994). Policy makers seeking to improve resource use in agriculture or reduce emissions of a pollutant on a national or regional scale usually express their questions to researchers at a catchment or landscape scale. Although their questions may be similar to those of farmers or environmentalists, at such scales a diverse range of managed and natural ecosystems are included and optimisation must include a consideration of the balance and interaction between these ecosystems, as well as their individual efficiencies (Tamm, 1991). Soil-plant N cycle processes assume changing levels of importance as scale changes and different measures are required to influence the N flows (Vos, 1996). Aggregation is used to simplify the pools measured or estimated (Robertson, 1982) but the methodologies available for up-scaling plot and field measurements to a catchment or regional scale are primitive. Our aim in this paper is not to provide a comprehensive review of the literature on soil-plant N dynamics, but to highlight areas in which research is needed to address the questions of farmers, environ-

mentalists and policy makers. Our question is: What concepts are required to complete our understanding of soil-plant N dynamics and which principles can be applied practically to improve management of the N cycle now? We put forward some promising research concepts in four main areas: soil N supply; crop N uptake, N losses from the soil-plant system and the application of data and models at a range of scales.

2. Soil N supply Recently emphasis has been placed on the measurement of inorganic N (usually nitrate) in soil before planting or at a specific time during the crop growing season to assess soil N supply (the 'Nmin' approach). In climates without extensive overwinter leaching, the use of a pre-planting test for soil nitrate is widely used; in the western United States before 1982, 16 states were using such a test (Keeney, 1982). A presidedress nitrate test has been developed for maize (Magdoff et al., 1984), which allows a period of nitrate accumulation under field conditions before sampling. Similar approaches are used in many countries in continental Europe and these were summarised by Mengel (1991). Snapshot measurements of soil inorganic N can only give a partial prediction of soil N supply since the bulk of soil N is found in organic forms. In a long term monitoring scheme, the average amount of mineral N measured in topsoil during autumn in the UK was only 76 kg ha -z compared to 7000 kg N ha -~ found in the soil organic matter (Shepherd et al., 1996). After comparing 14 models of the soil-plant cycle, de Willigen (1991) concluded that the biological processes controlling N supply were not well simulated, reflecting our incomplete understanding of the processes of mineralisation, immobilisation and their controls. Mineralisation is the process by which ammonium is released by soil micro-organisms as they utilise soil organic materials as an energy source, while immobilisation of ammonium and nitrate by micro-organisms is determined by the demands of protein synthesis. These processes reflect the properties of the substrate being mineralised and its interaction with the environment. The balance between them largely controls N

203

supply from soil. Measurement of mineralisation has until recently been limited to determination of net rates, which are the integration of a number of soil N processes. Relationships with environmental and management factors are difficult if not impossible to determine, since the factors act on other interacting processes as well as directly on mineralisation and immobilisation (Jarvis et al., 1996). Measurement of gross mineralisation rates by applying 15N and considedng isotope dilution theory (Kirkham and Bartholomew, 1954) enables mineralisation to be distinguished from immobilisation (e.g., Wessel and Tietema, 1992). Gross mineralisation rates may be able to be further resolved into the supply from a number of organic matter pools (Barraclough, 1995). The application of this technique under laboratory and field conditions provides the opportunity to establish a clearer mechanistic understanding of how the composition of soil organic matter and plant residues influence nitrogen mineralisation. Where pool dilution experiments are combined with direct tracer methods, determining the fate of N as well as rates of mineralisation, then with careful interpretation, we will be able to rapidly increase our understanding of soilplant N dynamics. Soil organic matter is not homogenous and is composed of a continuum of materials stabilised against mineralisation to varying degrees (Skjemstad et al., 1988). Chemical fractionation techniques have been used to define organic matter structures and have shown the presence of a wide range of functional groups (Stevenson, 1982). However, the fractions obtained by destructive chemical techniques have not been clearly related to soil N supply. Laboratory incubations of soil under various conditions have been used widely to indicate the potential of soils to supply N. These biological incubations and the quicker and more precise chemical extractions are believed to release available N pools preferentially. New methods continue to be developed and both approaches have been reviewed at length (Harmsen and van Schreven, 1955; Keeney, 1982; Jarvis et al., 1996). The limitations of N availability indices have been recognised; they were not expected to integrate the numerous inter-related soil, plant, environment and management factors which control N release and plant growth (Bremner, 1965), but simply to p r o vide extra information for assessment of soil N sup-

ply. N availability indices have not been used widely as part of fertiliser recommendation systems (Keeney, 1982). Today chemical and biological indices are thought to provide only a relative indication of N availability among soils differing in management (Bundy and Meisinger, 1994) and as such are used with other indicators to assess soil quality (FrancoVizcafno, 1997). Applying solid state J3C-NMR spectroscopy to whole soils (Wilson, 1987) indicates, in some cases, that carbohydrate and aliphatic compounds are more labile than other classes such as aromatics (Kinesch et al., 1995). However, Randall et al. (1995) concluded that for whole soils the proportions of different chemical species remain remarkably constant in response to different long-term management practices. The importance of the physical location of organic matter within the soil matrix, as well as its chemical composition, in influencing resistance to decomposition has been acknowledged widely (Oades et al., 1988). Soil organic matter physically separated into sand, silt and clay size fractions has been shown to decline in soils at different rates (Dalai and Meyer, 1986). Densiometric techniques have been developed to separate labile organic matter, usually after a partial or complete disruption of soil aggregates (Christensen, 1992). These labile fractions have been correlated to both soil microbial biomass and nitrogen mineralisation (Janzen et al., 1992; Hassink, 1992). The division of soil organic matter into pools of material, which behave similarly, is at the heart of many simulation models of soil carbon and nitrogen dynamics. However, a fundamental, and widely recognised deficiency of current models is that these pools are assumed to have biological significance, but are impossible to measure (Christensen, 1996; Jarvis et al., 1996). The Slow and Passive pools of organic matter of the Century model (Parton et al., 1987) and the Resistant, Decomposable and Inert organic matter pools of the Rothamsted Carbon Turnover Model (Jenkinson, 1990), are examples of these. Physical fractions obtained after destruction of soil structure prior to sieving at 53 #m (Cambardella and Elliot, 1992) have been related to the protected organic matter (POM) fraction of the Century model (Parton et al., 1987). The development of non-invasive and physical methods of dividing soil organic matter into pools may allow functional pools in models to be defined

204

and measured. However, the relationship of physical fractions to mineralisation still needs to be more clearly established. It is important to realise that the division of organic matter into a number of pools represents a simplification of the continuum of soil organic matter, both in terms of chemical characteristics and physical location, which exists in soils. The achievement of measurable model pools is not therefore the ultimate aim. Relationships between soil structure, management and N turnover are included rarely and inadequately in models (Elliott et al., 1996) and improved prediction of N supply from soil organic matter and crop residues will depend on understanding and quantifying all these effects. Improved understanding and prediction of mineralisation and immobilisation alone will not enable us to predict soil N supply, since mineralisation is only one of the processes of the soil N cycle. Models of the soil N cycle (e.g., ANIMO, Rijtema and Kroes, 1991; DAISY, Hansen et al., 1991; SUNDIAL, Smith et al., 1996) explicitly incorporate descriptions of the processes of soil N turnover (Fig. 2) and enable the effects of seasonal variations and timing of manage-

ment practices to be simulated. The integration of an improved understanding of mineralisation into a complete model of the soil N cycle is necessary to improve predictions of soil N supply. Using rapid and accurate field measurements of soil mineral N, along with estimates of mineralisation and denitrification, fertiliser inputs to grassland have been reduced by 30% whilst maintaining production levels (Titchen and Scholefield, 1994). Such a combination of field measurement and predictive modelling seems to be the way forward to achieve practical prediction of soil N supply for farmers. However, models are not yet widely used in fertiliser recommendation systems and it is unclear whether field measurements of soil inorganic N, N availability indices or organic matter pool sizes could be used to tune the models to specific field situations or adjust model recommendations during the growing season.

3. Crop N uptake Opportunities exist to modify the physiological

SOIL

PLANT

v

v

f Reduced •~,organic N Senesence//Glutamine /. synthetase

Volatilization Fertilizer Dry and wet Absorption deposition

1 1

//Nitrite reductase ~' / (dark and light)

l

~

Nitratereductase

~(dark

Absorption \

and light)

,

Denitrification

T lN;ol

Fertilizer Dry and wet deposition

t

Nitrification

1

] M,,,'~mmobilizatioJ ~ineraization

[!iiJ}iiiiiiii}~Up.,take

Volatilization Root/soil interface

OrganicN

1

V Leaching

Fig. 1. Simplified nitrogen cycle showing pools and processes in plant and soil.

205 efficiency of plants' use of NO3 (Harper, 1994). Nitrate reductase has a key role in nitrate metabolism and may be a point of metabolic limitation (Eichelberger et al., 1989; Campbell, 1990). Research is also being carded out to elucidate nitrate transport mechanisms (McClure et al., 1990) and to identify and clone nitrate uptake genes (Tsay et al., 1993). The gene encoding high affinity nitrate uptake in higher plants has not yet been identified. This paper does not address this research in detail. However, increased plant uptake efficiency of NO3 could significantly affect N losses as well as crop growth and yield. Efficient use of nitrogen in crop production requires an understanding of the relationship between nitrogen application, soil N supply, nitrogen uptake and that between uptake rates and growth rates. The so-called 'three quadrant diagrams' (e.g., de Wit, 1992) have been used to analyse crop nitrogen response retrospectively. The first quadrant is a plot of crop yield versus N applied, i.e., the overall or 'agronomic response'. In quadrants 2 and 3 the relationship between nitrogen uptake and nitrogen applied, determined primarily by soil processes, and the relationship between nitrogen uptake and yield are presented. Such plots are relatively easily obtained but are much more enlightening than the simple standard response curve, presenting hypotheses about the relationships between nitrogen application, soil N supply, nitrogen uptake and yields

Atmosphere I I Fertiliser

I

IOrganic Manure

Biomass Ammonium

Debris

Nitrate

Stubble & Straw Crop

Denitrificationl I Leachina I IHarvestl I Volatilisation

Fig. 2. Processes included in the Rothamsted Nitrogen Turnover model, a central part of SUNDIAL, from inputs through transformations to outputs. Nitrogen turnover is considered as a set of simple transformationprocesses (arrows)between discrete N compartments (boxes), where mineralisationis only one of the complex links.

with different application timings or placements (Black, 1993). A major limitation to three quadrant diagrams is that they do not provide any information about the dynamics of crop or soil processes. The relationship between N uptake rate and growth rate is described by the physiological efficiency of N use for a crop. Ingestad and Agren (1992) demonstrated that during exponential growth, the relative growth rate is proportional to the relative N uptake rate when this is constant. Nitrogen concentration in the plant is then also stable and controlled by the ratio between the relative uptake and relative growth rates. This approach allows the determination of physiological plant responses to N applications at a constant 'relative addition rate'. Although such an approach is important for increasing our understanding of the controls on plant growth, it is difficult to apply in the field. Linear rather than exponential growth occurs following canopy closure, nitrogen supply is depleted continually during growth or is topped up at arbitrary intervals and internal redistribution of N within the plant affects the simple relationship between uptake and growth. Based on numerous experiments, Greenwood et al. (1986, 1990) described a fairly fixed relation between crop mass and the critical N concentration required to ensure maximal growth. C3 and C4 species show different curves (Greenwood et al., 1990). Independent analyses showed that the 'Greenwood dilution curves' hold approximately in wheat (Justes et al., 1994) and potato (Vos, 1995). The concept of critical nitrogen concentration could therefore be used as a diagnostic tool against which to compare the nitrogen status of a crop (Justes et al., 1994). Total N requirements to meet yield targets without N deficiency can be calculated for the whole season or for growth periods. In practice, stresses other than nitrogen may reduce plant production and therefore modify the function describing the critical nitrogen concentration; this seems to invalidate the concept physiologically. Therefore the critical N concentration may be better related to development stage than crop mass. Unfortunately this would mean that critical (and minimum and maximum) concentrations for each of the development stages of crops would need to be determined. The required soil N supply for any period depends not only on crop N demand, but also on the crop uptake efficiency. A measure of that efficiency is the

206

apparent nitrogen fertiliser recovery, which is defined by: (N uptake of fertilised crop - N uptake of unfertilised crop)x Fertiliser applied. Apparent nitrogen recoveries usually range from 0.4 to 0.7 with 0.8 as an upper limit (Greenwood and Draycott, 1989). Recoveries are low where conditions favour losses of nitrate from the soil by denitrification or leaching and where carbon is available net microbial immobilisation will occur at least temporarily. Root exploitation of the soil volume is never 100% effective and this will also cause recoveries to be less than unity. However, the factors and processes determining the apparent recovery and its maximum value are not entirely understood. The efficiency of uptake is rarely accounted for in fertiliser recommendations systems, except implicitly. Computer simulation models are used as tools to integrate knowledge and explore the behaviour of the soil-plant system. Dynamic simulation models have been developed for a number of crops, e.g., CERES for maize, wheat and other cereals (Jones and Kiniry, 1996); ORYZA for rice (ten Berge et al., 1994) and CROPSYST suitable for a range of crops (Stockle and Nelson, 1996). However, as yet, models are rarely used to guide practical fertiliser decisions. Most models are able to adequately simulate the crop N uptake in response to variable N supply through the season. However, the effects of variable N supply on developmental processes, particularly changes in leaf area index, are not well understood and cannot be modelled using the concept of critical N concentrations (Greenwood et al., 1990). The distribution of nitrogen between component plant parts, such as leaves, stems and storage organs, as well as the final harvest index, is relatively constant irrespective of the average nitrogen concentration of the plant (e.g., Biemond and Vos, 1992). A fairly fixed distribution of nitrogen is therefore used in many empirical models of crop growth. However, in dynamic models the internal nitrogen distribution needs to be explained as a result of the underlying processes, e.g., turnover of proteins and sink-activity of competing organs. Plant N can be crudely divided into nitrate and reduced nitrogen pools, where the latter can be subdivided into metabolically active nitrogen and inactive fractions. The inactive fraction consists of N in storage proteins (seeds, tubers, etc.) and nitrogen contained in structural elements of

leaves, stems, roots and storage organs. As development proceeds, a larger amount of nitrogen is allocated at comparatively low concentrations in nonactive pools and this causes the decline in the critical nitrogen concentration with increase in crop mass (Caloin and Yu, 1984). Practical agriculture cannot wait until all details of responses of crops to nitrogen are quantified. Fertilisation guidelines and recommendation systems developed empirically have been in use for decades and are constantly being modified (e.g., Anon, 1994). Usually a fertiliser recommendation is given only once in the season, before or just after planting. Such a recommendation is necessarily inexact, since both the soil N supply and crop N requirement are not known in advance and depend on the future weather pattern. Therefore, wherever possible, the growing season should be divided into shorter periods for which recommendations are made, but which create opportunities to adjust and top up applications of fertiliser during the course of the season. Such 'dynamic optimisation' approaches are already in use in high value crops and where fertiliser applications are routinely split, e.g., the pre-sidedress nitrate test for maize (Magdoff et al., 1984). At the beginning of each of the growing periods, a decision is taken on the nitrogen nutrition, using quantitative data on the nitrogen status of crop and soil and estimates of the requirement of the crop and the supply from net mineralisation. Precise, robust and accurate field methods are required to collect the data needed for decision support. Comparison of tissue N concentrations with critical N concentrations is impractical as a diagnostic tool, not least due to the time involved in conducting analyses for total tissue N. Nitrate concentration in plant tissues can be measured using field diagnostic kits (Nitsch and Varis, 1991) and has been suggested as a useful diagnostic indicator of N status (Barraclough, 1993; Elliott et al., 1993). Tissue testing has not been widely adopted by growers, since critical nitrate levels can vary greatly from site to site (Beringer and Hass, 1979) and can change rapidly with time (Vos and Bom, 1993). Tissue nitrate testing is also relatively time consuming while the seemingly simple operations require good analytical skills. Field monitoring of crop N status is also possible

207

using hand-held chlorophyll meters (Fox et al., 1994; Peltonen et al., 1995) and near infra-red reflectometry (Young et al., 1993). Chlorophyll concentrations have been related to tissue N concentrations in different crops, e.g., rice (Peng et al., 1993) and potato (Vos and Bom, 1993). Future research should establish guidelines to help interpret chlorophyll values and to convert these into estimates of N status and future crop N requirement. Most fertiliser response trials have been carded out with single dose treatments but dynamic optimisation of N will require an understanding of the ability of crop to capture and utilise N under varied N management. The pattern of N uptake in potato crops can be drastically influenced by the dose and timing of split applications (Vos, 1995). For wheat, delaying N fertiliser applications beyond Feekes growth stage 7 can lower the potential yield response (Lutcher and Mahler, 1988). However, for most crops there is little information about which development stage N applications will continue to influence crop growth and quality of harvested components and in what form N should be applied to achieve maximal effect.

4. N losses

The focus of environmental interest is loss processes. Matching soil N supply and crop uptake minimises the amount of mineral N in the soil solution at any time and thereby losses. Thus the research concepts outlined in previous sections will also address environmental concerns to reduce N losses. The main rate limiting process controlling the availability and loss of the mobile nitrate ion may be nitrification, the process by which ammonium is converted to nitrate, rather than mineralisation. Despite a reasonable knowledge of the ecology of the bacteria involved (Prosser, 1986), nitrification remains a poorly defined process in many soils. In temperate tilled agricultural soils, nitrification rates are usually limited by mineralisation (Harmsen and van Schreven, 1955). However, in grassland soils significant quantities of ammonium may accumulate where swards are grazed or farm wastes are applied (Jarvis and Barraclough, 1991). The balance between mineralisation and nitrification is changed under changing environmental conditions and management

(Willison et al., 1997). While the factors controlling nitrification are known at a microsite scale, more work needs to be done to understand the controlling factors operating at larger scales. The use of ~SN dilution techniques to measure gross nitrification rates in conjunction with other processes may be key in this area. The root of the problem of nitrate leaching is 'untimely nitrate', i.e., nitrate which remains in the soil after crop uptake has ceased and therefore is vulnerable to loss in drainage. Some leaching loss of N seems inevitable however efficiently N is taken up by the crop, as plant processes have higher threshold temperatures for activity than mineralisation processes occurring in the soil (Vos, 1992). Even where dynamic optimisation is used, situations arise where large amounts of nitrate accumulate in the soil profile, for example, where fertiliser recovery is unexpectedly low due to drought or disease or following crops which leave large amounts of rapidly mineralised residues. In such situations, agricultural rotations may be modified to include catch or cover crops, which are able to absorb residual N and reduce losses of N by leaching (Martinez and Guiraud, 1990; Jensen, 1991). While such practices are usually effective in their year of application, some work has shown that the N is mineralised in subsequent years, and the leaching loss is delayed, rather than prevented entirely. Nitrous oxide is produced in soils by the microbially mediated processes of nitrification and denitrification; the capacity of soils to produce N20 increases as N availability increases. There have been a number of studies of N20 fluxes from soils, and the range of emission factors is considerable, generally emissions from natural ecosystems and uncultivated lands range between 0.1 and 9.1 kg NEO-N ha-! y-i. Emission factors from agricultural lands are more variable and generally higher (Bouwman, 1990). In agronomic terms the losses of N as N20 are probably negligible, however soils are a large source of N20 which contributes approximately 5% to radiative forcing. There is considerable uncertainty associated with estimates of N20 losses from soils because the factors controlling denitrification and nitrification are complex and still relatively poorly understood and emissions are temporally and spatially variable, typical coefficients of variation are in excess of 100%. Volatilisation of ammonia is also a possible route

208 5. Scaling

for N loss and ammonia deposition can cause pollution in natural ecosystems. The factors controlling NH3 emissions from soils and short and long range transport of NH3 between ecosystems are not well understood. However, injection or rapid incorporation of ammonium or urea fertilisers and manure, slurry or sewage sludge which reduce losses of NH3 significantly are amongst management practices recommended to UK farmers (Anon, 1992) and prescribed to Dutch farmers.

-'././///..(~_~

At different scales, processes assume changing levels of importance and different measures are required to influence the flows of input, output and losses (Vos, 1996). The driving variables for soil and crop processes are independent at a given scale, but as scale increases in time or space these may interact and new independent driving variables must be defined (Elliott and Paustian, 1996). As the tem-

,e

FIELD - water, nitrification, organic matter decomposition

, t i o ~ ~~.. ~ . . ' - ' ~ matter, physical disruption

.'-'... ~. ~~ .~..~..""

.

ORGANISM - oxygen, nitrate, carbon Fig. 3. Soil N cycle processes can be studied at different scales. However, the driving variables for crop and soil processes may change; this is illustrated for denitrification (adapted from Groffman et al., 1988; Jarvis et al., 1996).

209

poral and spatial scale of investigation increases, the primary factors controlling processes at the cellular level are affected by many physical and biological factors. At larger scales these secondary factors become increasingly significant and the process can be simplified and expressed as a function of the secondary variables (Groffman et al., 1988). This is illustrated for the process of denitrification in Fig. 3. As geographic size increases, a range of ecosystems may be included within system boundaries and fluxes between ecosystems and processes occurring at depth in soil may become important. Up-scaling research from pots or small plots to field, farm or landscape scales means that by necessity fewer measurements are possible in either time or space, and the uncertainty of estimates is increased. It is usually simply not practical to divide a field into a multitude of small plots and make replicate measurements on each, never mind at the landscape scale. An understanding of the soil-plant processes studied and their controlling factors allows us to define the system at an appropriate scale with regard to the goal of the study and to qualify the sub-systems processes and pools according to their importance to the system, e.g., discarding measurements of water erosion in flat or gently sloping areas. It is also crucial that research is seen within a hierarchical framework, where the research scale, can be related to scales above and below. N budgets have been calculated at a range of scales from global (Jenkinson, 1990; Isermann, 1993) national (Vos, 1996 (Netherlands); Royal Society, 1983 (UK)), catchment (Roberts, 1987; Lord, 1992), farm (Barry et al., 1993; Granstedt, 1992) field (Vinten et al., 1992) to small plot (Jenkinson and Parry, 1989; van Faassen and Lebbink, 1994). The underlying assumption of a nutrient budget is that of mass balance, i.e., N inputs to the system minus N outputs equal the change in N storage within the system. Although the nature and amount of inputs and outputs vary among farming systems, regions and even between fields, the mass balance concept provides a framework that can be applied systematically across a diversity of systems and scales (Committee on Long Range Soil and Water Conservation, 1993). The detail with which budgets are compiled varies; Stinner et al. (1984) included inputs, storage and processing by insects and micro-organisms in their budget. Farm budgets may be simple farm gate balances (Nguyen

et al., 1995) or include details of nutrient transfers between fields and within the farming system (Guiking et al., 1994). Careful definition of the system boundaries and compartments is essential (Meisinger and Randall, 1991). The construction of budgets at increasing scales necessitates aggregation of small compartments, where the degree of simplification is a function of lack of information or understanding of certain processes and the time and effort available for fine resolution, and it is made at the expense of both detailed insight and precision (Robertson, 1982). Where much has to be estimated rather than calculated, there is a danger of producing very speculative analyses (Dodgshon, 1994). However, budgets are often constructed with unjustified certainty, when specific values are assigned to fluxes or pools for which a wide range of values would be as accurate, and with inadequate documentation so that they are unverifiable (Robertson, 1982). N budgets may allow key processes or sites of high potential loss to be identified rapidly (Barry et al., 1993), allow the assessment of the efficiencies of nutrient use in the system to be determined (Karlovsky, 1981; Fowler et al., 1993) or add to our understanding of the importance of the processes and their interactions. It is relatively easy to compile N budgets at a range of scales, so that at its most detailed a regional budget is simply a compilation of the individual catchment budgets for the important systems of a region (Robertson, 1982). However, simple input/output balances can never deal adequately with the complexity of the N flows in the soil-plant system and a fundamental understanding of soil and plant processes is also required. In theory, N flows determined within a nested hierarchical classification similar to that used for the compilation of budgets (Lanyon and Beegle, 1989) can be brought together at any level to match information requirements. However, it is very difficult to use small-scale process measurements to estimate fluxes at field and landscape scales, e.g., for N20 fluxes (Tiedje et al., 1989). The simplest approach to up-scaling is to use the mean of a simple randomised sample to estimate the mean of the whole area or time period; this can be coupled with determination of the standard deviation to give a 95% confidence level, rather than a single value for the estimate of the mean. The size of the sample is important in attaining a representative mean

210

for the property (Goovaerts and Chiang, 1993). Where the area can be divided into a number of sub-units, e.g., soil series or land uses, stratified simple random samples will give a more representative estimate of the mean or confidence interval for the whole area, without necessarily increasing the total number of samples taken (Barnett, 1991). The cost and time necessary for numerous systematic measurements may hinder the use of these simple approaches; in this case measurement of an easily determined concomitant variable or combination of such variables along with a regression approach may provide useful estimates (e.g., Parton et al., 1987). This approach can be used to determine the probability level associated with the prediction, since the range of variation in any parameter is as important as the estimated mean value (Arrouays et al., 1995). Models can also be used to generate the required process or pool size estimates from other easily obtained data; Post et al. (1996) modelled global C dynamics by dividing the earth into half-degree cells and deriving the input values needed to run a carbon turnover model by using a model of net primary production which needed only simple inputs. Geostatistical algorithms, such as kriging, can also be used to estimate values for soil properties where sufficient data are available to calculate the variogram (McBratney et al., 1982; Voltz and Webster, 1990). Where other data are also known, e.g., soil mapping, co-located indicator co-kriging or full indicator co-kriging have been shown to improve the definition of cobalt and copper deficient areas (Goovaerts and Joumel, 1995). There is increased interest in such methods not just for up-scaling but also to allow the spatial distribution of soil properties to be mapped within fields in response to the increased use of yield-mapping (Stafford et al., 1996). However, our understanding of the factors causing spatial variability in crop yield and their interactions is relatively poor at the field scale (Parkin, 1993). The impacts of spatial variability in soils on plant growth and losses is not well understood and the temporal persistence of spatial patterns, e.g., caused by previous grazing is not known. The technology necessary to implement spatial management within fields, precision farming, is far in advance of the basic scientific knowledge necessary to underpin management recommendations. There is increased demand from policy makers that

models be linked, tested under a range of environments and applied on larger scales, both in space and time. Up- and down-scaling of models is not easy; Leffelaar (1990) itemised the criteria to be considered when determining an appropriate scale at which to perform simulations. Changing the scale of a model may result in changes in the scope of the model, the heterogeneity of input values and the data requirements of the model (Smith, 1995). There is need for closer involvement of related disciplines, such as statistics, to allow the development of robust methodologies and guidelines which enable us to scale up and scale down N cycle processes and models (Gaunt et al., 1997), so that the data collected for N cycle processes at the cellular, small plot or catchment scales for example can be applied at a range of scales.

6. Conclusions

Improved N use efficiency is a common goal of the farming community and environmental lobby, albeit for different reasons. Empirical relationships have been established between crop N uptake and plant growth and to describe the partitioning of N within crops. Our understanding of the dynamics of these processes and our ability to simulate them is increasing, but as yet remains incomplete. Similarly the factors controlling many soil processes, e.g., mineralisation, nitrification, denitrification are complex and relatively poorly understood. There are many areas in which fundamental understanding of the processes of the soil-plant system needs to be increased. N mineralisation is a key soil process; we would like to highlight the importance of integrative studies using ~SN both as a tracer within the plant-soil system and to quantify the effects of the chemical and physical protection of organic matter on gross transformations. Increased understanding of the metabolic limitations and genetic basis of N use and uptake in crops will also lead to improvement of the physiological efficiency of N use. Improved understanding of crop and soil processes at a fundamental level should lead to the development of continuously improving simulation models, which are able to integrate the complex effects of management and environmental factors. We see the develop-

211

ment of models which contain measurable pools of soil organic matter as a key step forward. Models of crop growth, the soil N cycle and plant-soil models have been developed. However, these are little used in current fertiliser and farm management recommendation systems. Farmers cannot wait for our understanding of plant-soil dynamics to be perfect or for the achievement of perfect models, but need researchers to put their current knowledge to use. New diagnostic approaches for field crops such as the use of near infrared reflectance and chlorophyll meters show promise but research is needed to clearly establish the relationships between the indicators and plant N uptake. There is also the possibility that measurements of soil mineral N or some N availability indices might be used as soil diagnostics. We suggest that for both crops and soils, diagnostic measurements be used in conjunction with the best of, or a combination of, current models. Diagnostics could be used to increase the field-specific nature of recommendations or to adjust model recommendations during the growing season. This would enable a greater use of dynamic optimisation strategies in the field. At present such a vision seems a distant possibility, but development work should be carried out with all possible speed to draw suitable models and diagnostics together. Fine tuning fertiliser recommendations using dynamic optimisation strategies also requires research to establish the crops ability to capture and utilise N under varied management regimes. Research leading to improvement in N efficiency, as described above, will also impact directly on N losses from agricultural systems. However, we stress the importance of conducting research to understand and improve N efficiency at a range of scales to address the needs not only of farmers, but also environmentalists and policy makers. Careful definitions of the system under study are important and an awareness of the relationship of the system under study to the level above and below is crucial. The methodologies used for up- and down-scaling process rates and pool sizes are relatively primitive. The demand that models should also be readily up- and down-scaled has led to the need for the involvement of related disciplines, such as statistics, to allow the development of robust guidelines and methodologies. The mass balance approach is a powerful tool to compare

and evaluate systems with regard to their efficiency or potential for loss. However, we need to take care when presenting budgets; the compilation of a budget is only a first step and rarely indicates understanding of any system. We also need an increased understanding of the importance of spatial variability on crop N uptake and loss processes, if we are to provide a sound scientific basis for the development of precision farming. Perhaps most important of all is that research is not carried out in isolation and that our increasing understanding of the fundamental processes of soil-plant dynamics continues to underpin the development of strategies for dynamic optimisation, to allow the identification of and improve simulation models that are used as fertiliser recommendation systems. In this way we will be able to provide answers to the questions of farmers, environmentalists and policy makers.

Acknowledgements For all their help, advice and hard work: David S Powlson, Toby Willison, Pete Barraclough. IACR receives grant-aided support from the UK Biotechnology and Biological Sciences Research Council and the authors also acknowledge funding from the UK Ministry of Agriculture, Fisheries and Food.

References Anon. 1992.Code of Good Agricultural Practice for the Protection of Air. HMSO, London, 74 pp. Anon. 1994.Fertiliser Recommendations(sixth ed.). MAFFReference Book 209. HMSO, London, 193 pp. Arrouays, D., Vion, I. and Kicin, J.L., 1995. Spatial analysis and modelling of topsoil carbon storage in temperate forest humic loamy soils of France. Soil Sci., 159:191-198. Barnett, V., 1991.SampleSurvey.Principles and Methods. Edward Arnold, London, 105 pp. Barraclough, D., 1995. ZSNisotope dilution techniques to study soil nitrogen transformations and plant uptake. Fertilizer Res., 42: 185-192. Barraclough, P.B., 1993. Nutrient storage pool concentrations in plants as diagnostic indicators of nutrient sufficiency. Plant Soil, 155/156: 175-178. Barry, D.A.J., Goorahoo, D. and Goss, M.J., 1993. Estimation of nitrate concentrations in groundwater using a whole farm budget. J. Environ. Qual., 22: 767-775. Ten Berge, H.F.M., Wopereis, M.C.S., Riethoven, J.J.M., Thiya-

212

garajan, T.M. and Sivasamy, R., 1994. The Oryza-0 model applied to optimize nitrogen use in rice. In: H.F.M. ten Berge, M.C.S. Wopereis and J.C. Shin (Editors), Nitrogen Economy of Irrigated Rice: Field and Simulation Studies. SARP Research Proceedings, April 1994, pp. 235-253. Beringer, V.H. and Hass, G., 1979. Brauchbarkeit der Pflanzenanalyse zur Bemessung sp~iter N-gaben zu Winterweizen. Landwirtsch. Forsch., 32: 384-394. Biemond, H. and Vos, J., 1992. Effects of nitrogen on the development and growth of the potato plant. 2. The partitioning of dry matter, nitrogen and nitrate. Ann. Bot., 70: 37-45. Black, C.A., 1993. Soil Fertility Evaluation and Control. Lewis, Boca Raton, FL, 746 pp. Bouwman, A.F., 1990. Exchange of greenhouse gases between terrestrial ecosystems and the atmosphere. In: A.F. Bouwman (Editor), Soils and the Greenhouse Effect. Wiley, Chichester, pp. 61-127. Bremner, 1965. Nitrogen availability indices. In" C.A. Black et al. (Editors), Methods of Soil Analysis. Part 2. Agronomy Series 9. ASA, Madison, WI, pp. 1324-1345. Bundy, L.G. and Meisinger, J.J., 1994. Nitrogen availability indices. In: R.W. Weaver et al. (Editors). Methods of Soil Analysis. Part 2. SSSA Book Series 5. SSSA, Madison, WI, pp. 951984. Caloin, M. and Yu, O., 1984. Analysis of the time course of change in nitrogen content in Dactylis glomerata L. using a model of plant growth. Ann. Bot., 54: 69-76. Cambardella, C.A. and Elliot, E.T., 1992. Particulate soil organic matter changes across a grassland cultivation sequence. Soil Sci. Soc. Am. J., 56: 777-783. Campbell, W.H., 1990. Purification, characterization and immunochemistry of higher plant nitrate reductase. In: Y.P. Abrol (Editor), Nitrogen in Higher Plants. Research Studies Press, Taunton, pp. 65-91. Christensen, B.T., 1992. Physical fractionation of soil and organic matter in primary particle size and density separates. Adv. Soil Sci., 20: 1-90. Christensen, B.T., 1996. Matching measurable soil organic matter fractions with conceptual pools in simulation models of carbon turnover: revision of model structure. In: D.S. Powlson, P. Smith and J.U. Smith (Editors), Evaluation of Soil Organic Matter Models Using Existing Long-Term Datasets. Series 1 Global Environmental Change, NATO ASI Series, Vol. 38, pp. 143-161. Committee on Long Range Soil and Water Conservation (CLSWC), 1993. Nitrogen and phosphorus mass balances: methods and interpretation. In: Soil and Water Quality: An Agenda for Agriculture. National Research Council National Academy Press Washington, DC, pp. 431-477. Dalai, R.C. and Meyer, R.J., 1986. Long-term trends in fertility of soils under continuous cultivation and cereal cropping in Southern Queensland. III. Distribution and kinetics of soil organic carbon in particle-size fractions. Aust. J. Soil Res., 24: 293-300. Dodgshon, R.A., 1994. Budgeting for survival: nutrient flow and traditional Highland farming. In: S. Foster and T.C. Smout, The History of Soils and Field Systems. Scottish Cultural Press, Aberdeen, pp. 83-93. Eichelberger, K.D., Lamber, R.J., Below, F.E. and Hageman, R.H.,

1989. Divergent phenotypic recurrent selection for nitrate reductase activity in maize. II Efficient use of fertiliser nitrogen. Crop Sci., 29:1398-1402. Elliott, D.E., Pelham, S.D. and Reuter, D.J., 1993. Synchronising diagnosis and correction of nitrogen deficiency in barley grown in semi-arid environments. Plant Soil, 155/156: 363-366. Elliott, E.T. and Paustian, K., 1996. Why site networks? In: D.S. Powlson, P. Smith and J.U. Smith (Editors), Evaluation of Soil Organic Matter Models Using Existing Long-Term Datasets. Series l Global Environmental Change, NATO AS! Series, Vol. 38, pp. 27-36. Elliott, E.T., Paustian, K and Frey, S.D. 1996. Modeling the measurable or measuring the modelable: a hierarchical approach to isolating meaningful soil organic matter fractions. In: D.S. Powlson, P. Smith and J.U. Smith (Editors), Evaluation of Soil Organic Matter Models Using Existing Long-Term Datasets. Series l Global Environmental Change, NATO ASI Series, Vol. 38, pp. 161-180. Fowler, S.C., Watson, C.W. and Wilman, D., 1993. N, P, and K on organic farms: herbage and cereal production, purchases and sales. J. Agric. Sci. (Cambridge), 120: 353-360. Fox, R.H., Piekielek, W.P. and MacNeal, K.M., 1994. Using a chlorophyll meter to predict nitrogen fertilizer needs of winter wheat. Commun. Soil Sci. Plant Anal., 25:17 l - 18 I. Franco-Vizcafno, E., 1997. Comparative soil quality in maize rotations with high or low residue diversity. Biol. Fertil. Soils, 24: 32-38. Gaunt, J.L., Riley, J., Stein, A. and Penning de Vries, F.W.T., 1997. Requirements for effective modelling strategies. Agric. Systems, 54: 153-168. Goovaerts, P. and Chiang, C.N., 1993. Temporal persistence of spatial patterns for mineralizable nitrogen and selected soil properties. Soil Sci. Soc. Am. J., 57: 372-381. Goovaerts, P. and Journel, A.G., 1995. Integrating soil map information in modelling the spatial variation of continuous soil properties. Eur. J. Soil Sci., 46: 397-414. Granstedt, A., 1992. Case studies on the flow and supply of nitrogen in alternative farming in Sweden I Skilleby Farm 1981-87. Biol. Agric. Hortic., 9: 15-63. Greenwood, D.J. and Draycott, A., 1989. Quantitative relationships for growth and N content of different vegetable crops grown with and without ample fertilizer-N on the same soil. Fertilizer Res., 18: 175-188. Greenwood, D.J., Neeteson, J.J., and Draycott, A., 1986. Quantitative relationships for the dependence of growth rate of arable crops on their nitrogen content, dry weight and aerial environment. Plant Soil, 91: 281-301. Greenwood, D.J., Lemaire, G., Gosse, G., Cruz, P., Draycott, A. and Neeteson, J.J., 1990. Decline in percentage N of C3 and C4 crops with increasing plant mass. Ann. Bot., 74: 397407. Groffman, P.M., Tiedje, J.M., Robertson, G.P. and Christensen, S., 1988. Denitrification at different spatial and temporal scales: proximal and distal controls. In: J.R. Wilson (Editor), Advances in Nitrogen Cycling in Agricultural Ecosystems. Proceedings of a Symposium, Brisbane, Australia. CAB International, Wallingford, pp. 174-192.

213

Guiking, F.C.T., Jansen, D.M. and Fresco, L.O., 1994. The use of simplified nutrient balances at farm level to determine boundary conditions for sustainable production. In: J.K. Syers and D.L. Rimmer (Editors), Soil Science and Sustainable Land Management in the Tropics. Proceedings of a British Society of Soil Science Meeting, Newcastle, 1993. CAB International, Wallingford, pp. 248-257. Hansen, S., Jensen, H.E., Nielsen, N.E. and S vendsen, H., 1991. Simulation of the nitrogen dynamics and biomass production in winter wheat using the Danish simulation model DAISY. Fertilizer Res., 27: 245-259. Harmsen, G.W. and van Schreven, D.A., 1955. Mineralization of organic nitrogen in soil. Adv. Agron., 7: 299-398. Harper, J.E., 1994. Nitrogen metabolism. In: K.J. Boote, J.M. Bennett, T.R. Sinclair and G.M. Paulsen (Editors), Physiology and Determination of Crop Yield. ASA, CSSA, SSSA. Madison, WI, pp. 285-302. Hassink, J., 1992. Effects of soil texture and structure on carbon and nitrogen mineralization in grassland soils. Biol. Fertil. Soils, 14: 126-134. Ingestad, T. and ,~gren, G.I., 1992. Theories and methods on plant nutrition and growth. Physiol. Plant., 84: 177-184. Isermann, K., 1993. Territorial, continental and global aspects of C, N, P and S emissions from agricultural ecosystems. In: R. Wollast, F.T. Mackenzie and L. Chou (Editors), Interactions of C, N, P and S Biogeochemical Cycles and Global Change. SpringerVerlag Berlin, pp. 79-121. Janzen, H., Campbell, C., Brandt, S., Lafond, G. and TownleySmith, L. 1992. Light-fraction organic matter in soils from long term crop rotations. Soil Sci. Soc. Am. J., 56: 17991806. Jarvis, S.C. and Barraclough, D., 1991. Variation in mineral nitrogen content under grazed grassland swards. Plant Soil, 138: 177-188. Jarvis, S.C., Stockdale, E.A., Shepherd, M.A. and Powlson, D.S., 1996. Nitrogen mineralization in temperate agricultural soils: processes and measurement. Adv. Agron., 57: 188-237. Jenkinson, D.S. and Parry, L.C., 1989. The nitrogen cycle in the Broadbalk wheat experiment: a model for the turnover of nitrogen through the soil microbial biomass. Soil Biol. Biochem., 21: 535-541. Jenkinson, D.S., 1990. The turnover of organic carbon and nitrogen in soil. Philos. Trans. R. Soc., London Ser. B, 329: 361-368. Jensen, E.S., 1991. Nitrogen accumulation and residual effects of nitrogen catch crops. Acta Agric. Scand., 41: 333-344. Jones, C.A. and Kiniry, J.R. 1996. Ceres-Maize: A Simulation Model of Maize Growth and Development. Texas A and M University Press, College Station, TX. Justes, E., Mary, B., Beynard, J.-M., Machet, J.-M. and ThelierHuche, L., 1994. Determination of a critical nitrogen dilution curve for winter wheat crops. Ann. Bot., 74: 397-407. Karlovsky, J., 1981. Cycling of nutrients and their utilisation by plants in agricultural ecosystems. Agro-Ecosystems, 7: 127144. Keeney, D.R., 1982. Nitrogen Availability Indices. In: A.L. Page, R.H. Miller and D.R. Keeney (Editors), Methods of Soil Analysis. Part 2. ASA, Madison, WI, pp. 711-734.

Kinesch, P., Powlson, D.S. and Randall, E.W., 1995. t3C-NMR studies of SOM in whole soils. Part 2 A case study of some Rothamsted soils. Eur. J. Soil Sci., 46: 139-146. Kirkham, D. and Bartholomew, W.V., 1954. Equations for following nutrient transformations in soil, utilizing tracer data. Soil Sci. Soc. Am. proc., 19:189-192. Lanyon, L.E. and Beegle, D.B., 1989. The role of on farm nutrient balance assessments in an integrated approach to nutrient management. J. Soil Water Conserv., 44: 164-168. Leffelaar, P.A., 1990. On scale problems in modelling: an example from soil ecology. In: R. Rabbinge, J. Goudriaan, H. van Keulen, F.W.T. Penning deVries and H.H. van Laar (Editors), Theoretical Production Ecology: Reflections and Prospects. Simulation Monographs No. 34, Pudoc, Wageningen. Lord, E.I., 1992. Modelling of nitrate leaching: nitrate sensitive areas. Aspects Appl. Biol. (Nitrate and Farming Systems), 30: 19-28. Lutcher, L.K. and Mahler, R.L., 1988. Sources and timing of spring topdress nitrogen on winter wheat in Idaho. Agron. J., 80: 648654. McBratney, A.B., Webster, R., McLaren, R.G. and Spiers, R.B., 1982. Regional variation of extractable copper and cobalt in the topsoil of south-east Scotland. Agronomie, 2: 969-982. McCiure, P.R., Kochian, L.V., Spanswick, R.M. and Schaff, J.E., 1990. Evidence for co-transport of nitrate and protons in maize roots. I. Effects of nitrate on the membrane potential. Plant Physiol., 93:28 i-289. Magdoff, F.R., Ross, D. and Amadon, J. 1984. A soil test for nitrogen availability to com. Soil Sci. Soc. Am. J., 48: 13011304. Martinez, J. and Guiraud, G., 1990. A lysimeter study of the effects of a rye-grass catch crop, during a winter wheat/maize rotation, on nitrate leaching and the following crop. 3. Soil Sci., 41: 516. Meisinger, J.J. and Randall, G.W., 1991. Estimating nitrogen budgets for soil-crop systems. In: R.F. Follett, D.R. Keeney and R.M. Cruse (Editors), Managing Nitrogen for Groundwater Quality and Farm Profitability. SSSA, Madison WI, pp. 85124. Mengel, K., 1991. Available nitrogen in soils and its determination by the 'Nmin' method and by electroultrafiltration (EUF). Fertilizer Res., 28: 251-262. Neeteson, J.J. and W adman, W.P., 1987. Assessment of economically optimum application rates of fertilizer N on the basis of response curves. Fertilizer Res., 12: 37-52. Nguyen, M.L., Haynes, R.J. and Goh, K.M., 1995. Nutrient budgets and status in three pairs of conventional and alternative mixed cropping farms in Canterbury, New Zealand. Agric. Ecosystems Environ., 52: 149-162. Nitsch, A. and E. Vails, 1991. Nitrate estimates using the Nitracheck Test for precise N-fertilization during plant growth and, after harvest, for quality testing of potato tubers. Potato Res., 34: 95-105. Oades, J.M., Waters, A.G., Vassallo, A.M., Wilson, M.A. and Jones, G.P., 1988. Influence of management on the composition of organic matter in a red-brown earth as shown by ~3C nuclear magnetic resonance. Aust. J. Soil Res., 26: 289-299.

214

Parkin, T.B., 1993. Spatial variability of microbial processes in soil - a review. J. Environ. Qual., 22:409-417. Parton, W.J., Schimel, D.S., Cole, C.V. and Ojima, D.S., 1987. Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J., 51: 11731179. Peltonen, J., Virtanen, A. and Haggren, E., 1995. Using a chlorophyll meter to optimise nitrogen fertiliser application for intensively managed small-grain cereals. J. Agron. Crop Sci., 174: 309- 318. Peng, S., Garcia, F.V., Laza, R.C. and Cassman, K.G., 1993. Adjustment for specific leaf weight improves chlorophyll meter's estimate of rice leaf nitrogen concentration. Agron. J., 85: 987-990. Post, W.M., King, A.W. and Wullschleger, S.D., 1996. Soil organic matter models and global estimates of soil carbon. In: D.S. Powlson, P. Smith and J.U. Smith (Editors), Evaluation of Soil Organic Matter Models using Existing Long-Term Datasets. Series 1 Global Environmental Change. NATO ASI Series, Voi. 38, pp. 201-224. Prosser, J., 1986. Nitrification. Special Publication of Society of General Microbiology. IRL Press, Oxford, 217 pp. Randall, E.W., Mahieu, N., Powlson, D.S. and Christensen, B.T., 1995. Fertilization effects on organic matter in physically fractionated soils as studied by ~3C-NMR: results from two long term field experiments. Eur. J. Soil Sci., 46: 557-565. Rijtema, P.E. and Kroes, J.G., 1991. Some results of nitrogen simulations with the model ANIMO. Fertilizer Res., 27: 189198. Roberts, G., 1987. Nitrogen inputs and outputs in a small agricultural catchment in the Eastern part of the United Kingdom. Soil Use Mgmt., 3: 148-154. Robertson, G.P., 1982. Regional nitrogen budgets: approaches and problems. Plant Soil, 67: 73-79. Royal Commission on Environmental Pollution, 1996. Sustainable Use of Soil, 19th Report. HMSO, London, 260 pp. Royal Society, 1983. The Nitrogen Cycle of the United Kingdom. Report of a Royal Society Study Group. The Royal Society, London, 576 pp. Russell, E.W. 1966. A History of Agricultural Science in Great Britain. Allen and Unwin, London, 493 pp. Shepherd, M.A., Stockdale, E.A., Powlson, D.S. and Jarvis, S.C., 1996. The influence of organic N mineralization on the management of agricultural systems in the UK. Soil Use Mgmt., 12: 7685. Skjemstad, J.O., Vallis, I. and Myers, R.J.K., 1988. Decomposition of soil organic nitrogen. In: J.R. Wilson (Editor), Advances in Nitrogen Cycling in Agricultural Ecosystems. Proceedings of a Symposium, Brisbane, Australia. CAB International, Wallingford, pp. 145-155. Smith, J.U., 1995. Models and scale: up- and down-scaling. In: A. Stein, F.W.T. Penning de Vries and P.J. Schotman (Editors), Models in Action. AB-DLO, Haren, pp. 25-41. Smith, J.U., Bradbury, N.J. and Addiscott, T.M., 1996. SUNDIAL: a PC based system for simulating nitrogen dynamics in arable land. Agron. J., 88: 38-43. Stafford, J.V., Ambler, B., Lark, R.M. and Catt, J., 1996. Mapping

and interpreting the yield variation in cereal crops. Comput. Electron. Agric., 14: 101-119. Stevenson, F.J., 1982. Organic forms of nitrogen. In: F.J. Stevenson (Editor), Nitrogen in Agricultural Soils. ASA, CSSA, SSSA, Madison, WI, pp. 67-122. Stinner, B.R., Crossley, Jr., D.A., Odum, E.P. and Todd, R.L., 1984. Nutrient budgets and internal cycling of N, P, K, Ca and Mg in conventional tillage, no-tillage and old-field ecosystems in the Georgia piedmont, Ecology, 65: 354-369. Stockle, C.O. and Nelson, R. 1996. CROPSYST. Biological Systems Engineering Department. Washington State University, Pullman, WA. Tamm, C.O., 1991. Nitrogen in Terrestrial Ecosystems. Questions of Productivity, Vegetational Changes and Ecosystem Stability. Ecological Studies Vol. 81, Springer-Vedag. Berlin. 115 pp. Tiedje, J.M., Simkins, S. and Groffman, P.M., 1989. Perspectives on measurement of denitrification in the field including recommended protocols for acetylene-based methods. Plant Soil, 115: 261-284. Titchen, N.M. and Scholefield, D., 1994. Tactical fertilizer application on commercial dairy farms. Institute of Grassland and Environmental Research Annual Report 1993, 72 pp. Tsay, Y.F., Schroeder, J.I., Feldman, K.A. and Crawford, N.M., 1993. The herbicide sensitivity gene CHL1 of Arabidopsis encodes a nitrate inducible nitrate transporter. Cell, 72: 705713. van Faassen, H.G. and Lebbink, G., 1994. Organic matter and nitrogen dynamics in conventional versus integrated arable farming. Agric. Ecosystems Environ., 51: 209-226. Vinten, A.J.A., Vivian, B.J. and Howard, R.S., 1992. The effect of nitrogen fertilizer on the nitrogen cycle of two upland arable soils of contrasting textures. Proc. Fert. Soc., 329: 1-48. Voltz, M. and Webster, R., 1990. A comparison of kriging, cubic splines and classification for predicting soil properties from sample information. J. Soil Sci., 41: 473-490. Vos, J., 1992. Growth and nitrate accumulation of catch crops. In: E. Francois, K Pithan and N. Bartiaux-Till (Editors), Nitrogen Cycling in Cool and Wet Regions of Europe - COST 814. Proceedings of a Workshop at Gembloux, Belgium. Commission of the European Communities, Brussels, pp. 103109. Vos, J., 1995. Nitrogen and the growth of potato crops. In: A.J. Haverkort and D.K.L. MacKerron (Editors), Potato Ecology and Modelling of Crops under Conditions Limiting Growth. Current Issues in Production Ecology, Vol. 3. Kluwer, Dordrecht, pp. 115-128. Vos, J., 1996. Nitrogen cycle related to crop production in cool and wet climates. In: R. Samuelsen, B. Solsheim, K. Pithan and E. Watten Melvaer (Editors), Nitrogen Cycling in Cool and Wet Regions of Europe - COST 814. Proceedings of a Workshop held at Tromso, Norway, 1995. European Commission, Brussels, pp. 3-14. Vos, J. and Born, M., 1993. Hand-held chlorophyll meter: a promising tool to assess the nitrogen status of potato foliage. Potato Res., 36: 301-308. Vos, J. and Marshall, B., 1994. Nitrogen and potato production: strategies to reduce nitrogen leaching. In: Proceedings of the

215

12th Triennial Conference of the European Association for Potato Research, Pads, France, 1993 (INRA, France), pp. 101-110. Wessel, W.W. and Tietema, A., 1992. Calculating gross N transformation rates of ~SNpool dilution experiments with acid forest litter: analytical and numerical approaches. Soil Biol. Biochem., 24:931-942. de Willigen, P., 1991. Nitrogen turnover in the soil-crop system; comparison of fourteen models. Fertilizer Res., 27: 141149. Willison, T.W., Tlustos, P., Murphy, D.V., Baker, J.C., Goulding,

K.W.T. and Powlson, D.S., 1997. Short-term effects of nitrogen on methane oxidation in soils. Eur. J. Soil Sci., in press. Wilson, M.A., 1987. NMR Techniques and Applications in Geochemistry and Soil Chemistry. Pergamon Press, Oxford. de Wit, C.T., 1992. Resource use efficiency in agriculture. Agric. Systems, 40:125-15 I. Young, M.W., Davies, H.V. and MacKerron, D.K.L 1993. Comparison of techniques for nitrogen analysis in potato crops. In: M.A.C. Fragoso and Beusichem M.L. van (Editors), Optimization of Plant Nutrition. Developments in Plant and Soil Sciences, Vol. 53. Kluwer, Dordrecht, pp. 7-11.


t~ 1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geij'n (Editors)

217

Modeling crop nitrogen requirements" a critical analysis C.O. Stockle a'*, P. Debaeke b aBiological Systems Engineering Department, Washington State University, Pullman, WA 99164-6120, USA blNRA, Station d'Agronomie, BP 27, 31326 Castanet Tolosan cedex, France Accepted 2 June 1997

Abstract Four approaches to simulate N requirements of wheat were testedusing data collected at the Auzeville experiment station of INRA, near Toulouse, France, from an experiment providing a wide range of soil N available for crop uptake. In these approaches, crop N requirements are expressed in terms of characteristic plant N concentration curves (maximum, critical, and minimum), representing expected concentration for a given crop N status throughout the growing season. Modeling approaches were evaluated for their ability to discriminate between N-limited and non-limited wheat plots as well as to properly represent the upper and lower limits of observed plant N concentrations. Best results were obtained using the growth dilution concept to represent the characteristic curves, while others based on temperature sums, growth stages, or fraction of the growth cycle were less satisfactory. Simulation of crop growth and N uptake based on N requirements estimated using the growth dilution concept resulted in a relationship between biomass at harvest and N uptake that correctly described an upper boundary for all observed data points. However, simulated and observed crop N uptake on a plot by plot basis resulted in low agreement. This was attributed to uncertainty in the measurement of initial soil N and crop N uptake, and the effect of other growth reducing factors (e.g. diseases) and possibly physical and/or chemical restrictions to field N uptake normally not accounted for by crop growth models. © 1997 Elsevier Science B.V. Keywords: Growth; Model; Simulation; Wheat

I. Introduction Simulation models are increasingly used for the assessment of crop productivity and the impact on the environment that may result from given combinations of weather, soil, crop characteristics, and water and N management. For this purpose, the proper simulation of crop N requirements is important, both in terms of amount and distribution throughout the growing season. Several approaches have been proposed to simulate * Corresponding author. Tel.: +l 509 3353564; fax: +l 509 3352722; e-mail: [email protected]

crop N requirements. We selected four that are representative and able to model wheat N requirements, corresponding to those included in the following crop growth models: AFRCWHEAT2 (Porter, 1993), Daisy (Hansen et al., 1991), EPIC (Williams et al., 1989), and CropSyst (Stockle and Nelson, 1996). In these approaches, crop N requirements are expressed in terms of characteristic plant N concentration curves, which represent the expected concentration for a given crop N status throughout the growing season. The most complete approaches define three such curves: a maximum (Nmax), a critical (N¢,t), and a minimum (Nmi,) plant N concentration. Plant growth is not limited by N if plant concentration


218

is at or above Ncrit, while Nmax establishes the maximum crop N uptake. Below Ncnt, plant growth is reduced, stopping completely when plant N concentration reaches Nmin. It should be noticed that these definitions do not consider the quality of the harvested crop. Plant N concentration is not constant but decreases with time, and so do the three characteristic curves. To describe this process, models express these curves as a function of crop growth stage (AFRCWHEAT2), the fraction of the growth cycle (EPIC), or thermal time (Daisy). On the other hand, research has shown that Nc~it decreases with increasing shoot biomass according to an allometric equation (Salette and Lemaire, 1981; Greenwood et al., 1990), usually referred to as the growth dilution law. This concept has been tested with field data and shown able to discriminate between well-supplied and N-deficient crops (e.g. Justes et al., 1994; P1Enet, 1995). Furthermore, single allometric equations for C3 and Ca crops have been

6 5 ~4

proposed (Greenwood et al., 1990). Similar functional relationships with biomass may be used for Nmax and Nmi, (Justes et al., 1994). An implementation based on this concept was introduced to the CropSyst model. It must be noted, however, that adequate simulation of crop N requirements may not guarantee the ability of models to simulate crop growth in response to soil available N and N management, which is the ultimate objective in the application of these models. In most models, N uptake depends on crop N requirements (as needed to maintain Nmax) but also on the attainable N uptake as given by N concentration, moisture, and root distribution in the soil profile (Jones and Kiniry, 1986; Williams et al., 1989; Hansen et al., 1991; Porter, 1993; Stockle et al., 1994). The attainable N uptake may or may not satisfy crop N requirements. The objectives of this study were (1) to test four approaches to model crop N requirements using data collected for winter wheat at Auzeville, southern France, during the growing season of 1993, and (2) 6

k•

AFRCWHEAT2

5

"'. o

~

EPIC

~4

Z 3

o°°°

i,,1

Z3

~

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

1

o 0

,

v

,

,

20

40

60

80

o 100

0.2

o.o

Growth stages

0.4

0.6

0.8

1.0

Fraction of the cycle

DAISY MODEL

5

~4

CROPSYST

~-~--~

~4

;,.

z 3 N

z3

f= " r ' ~ O

--~2

"

0

500

•

n •

•°-.j~.

o

1000

1500

Degree Days

2000

2500

0

•

oDD o

o

....

3

......

6

9

12

Aboveground biomass (Mg/ha)

Fig. 1. Comparisonof four models for estimating characteristic plant N concentration curves with data from N-limited and non-limitedplots of winter wheat cultivar Soisson grown at Auzeville, France in 1992-1993. (Symbols: open square, N limited; closed square, N non-limited; Lines: dashed, Nman;solid, Nc,it;dotted, Nmin).

219

compare observed and simulated biomass production and crop N uptake associated with crop N requirements, estimated using the best of the four approaches from objective 1, under a wide range of available nitrogen.

curves are defined. A generic exponential function of the fraction of the growing season (Fgs) is used to calculate Nc~it. The value of Fgs is determined as the ratio of the current temperature sum to the total temperature sum at maturity. For wheat, the function is Ncrit - 100 (0.01 + 0.05 exp(-2.67 Fgs)). Nmi, is calculated as half of No,it.

2. Approaches to model crop N requirements 2.4. CropSyst A description of four approaches to determine the characteristic concentration curves used to model crop N requirements throughout a growing season is given below. Some of these include a separate set of curves for shoots and roots. For this study, only shoot concentrations are of interest.

2.1. AFRCWHEA~ In this wheat model, Nmax and Nmin curves are defined. Nmax is set at 4.5% from emergence to initiation of the terminal spikelet, from which point it falls to a value of 0.5% by the end of grain filling. Nmin starts at 2.5%, increases slightly until the double ridge stage, and then falls to 0.25% by the end of grain filling. The details of the change of Nmax and Nmin as a function of developmental stage are given graphically by Porter (1993), and are reproduced here in Fig. 1 using the phenological scale proposed by Zadoks et al. (1974).

2.2. Daisy In this wheat model, Nmax, Ncrit , and Nmi, curves are defined. Characteristic N concentrations are given as a function of air temperature sum (base 0°C) from emergence. All curves have a constant concentration up to a temperature sum of 100°C - days, with values of 5%, 3%, and 2% for Nmax, Nc~it, and Nmin, respectively. After a temperature sum of 1100°C- days, these values are constant at values of 1.2%, 1.0%, and 0.7%, respectively. All concentrations decay exponentially for temperature sums between these two boundaries. These three curves are given graphically by Hansen et al. (1991), and reproduced in Fig. 1.

2.3. EPIC In this genetic crop growth model, Neat and

Nmi n

Nmax, Ncrit, and Nmin curves are defined in this genetic crop growth model. Plant N concentration (N%), in percent, is assumed to be related with biomass accumulation (B) as follows: N% = a B -b, where a and b are fitted parameters (Salette and Lemaire, 1981; Greenwood et al., 1990; Justes et al., 1994; Pl6net, 1995). The value of Nmax during early growth is required as input parameter. Then, the characteristic N concentration curves are given by: Nmax= min (Nmax, amax B-°'35), Ncrit = min (0.7 Nmax, acrit B-°'35), and Nmin = rain (0.4 Nmax, aminB-0"35), where B is biomass for an unstressed crop (t/ha), amax= Nmax/ (2-0"35), acrit= 0.7 Nmax/(1.5-°'35), and ami n "" 0.4 Nmax/(O.5-°'35). This implementation, based on data from the references given above, is a generalization that works well for both C3 and C4 species. For wheat, a value of 5.0% was used for Nmaxduring early growth (see Fig. 1). The growth dilution law seems to work well up to flowering (Justes et al., 1994). After this point, the CropSyst implementation reduces linearly the three characteristic curves so as to meet specified (input parameters) values of Nmax, Ncrit, and Nmi, at maturity.

3. Methods Experimental data collected for winter wheat during 1993 growing season at the INRA station in Auzeville, southwestern France, were used to compare the four modeling approaches (Debaeke et al., 1996). The soil was a deep silty-clay loam, with an organic matter content of 1.6% (0-30 cm). The cultivar Soissons was grown in 16 plots (8 preceding crops x2 input levels). Preceding crops were faba bean, maize, pea, rapeseed, sorghum, soybean, sunflower, and wheat. Input levels differed by sowing date (high-input sown on 30 November 1992 and low-input on 16 December

220

1992) and N fertilization. N fertilizer rate was calculated using the French balance sheet method (R6my, 1981), adapted for southern France, using yield goals of 5 (low-input, N1) and 8 t/ha (high-input, N3). Briefly, this method calculates the N fertilizer dose as a function of soil availability and crop requirements for a yield objective, and includes a correction factor to increase the N dose when a limitation to N uptake by soil structure or soil moisture is expected. No supplementary irrigation was required in 1993. Each wheat plot was divided into two sections: (1) a central area (410 m2), receiving N fertilizer according to the yield objective (Nx treatment); (2) an unfertilized lateral area (60 m2), which was kept free of N fertilizer to assess the soil N contribution (NO treatment), yielding a total of 32 treatment combinations. Mineral N was determined before N fertilizer supply, in December 1992, from 5 soil cores per plot taken to a depth of 1.20 m in 0.2-m increments. Above-ground dry matter and N concentration were measured at tillering, stem elongation, shooting, anthesis and maturity. At each sampling date, the plants in five 0.25-m 2 quadrats were collected for biomass and N determination. At maturity, dry matter and N concentration were measured separately for grain, culm and chaff using the Kjeldahl method. Growth stages were monitored regularly and calculations of thermal time were done from emergence. Data were analyzed to discriminate between N-limited and non-limited plots using the method described by Justes et al. (1994). As a clarification, this method of discrimination is not based on the growth dilution model for Ncrit given above, and its use does not affect the predictive ability of this model. The Ncrit curves calculated using the four modeling approaches introduced above were evaluated for their ability to separate N-limited and non-limited plots throughout the growing season. Also Nmax and Nmi, curves for each approach were evaluated for their ability to define the upper and lower limits of observed plant N concentrations. To examine the biomass production and crop N uptake associated with crop N requirements, crop growth simulations using the best approach for estimating crop N requirements (objective 1) were performed for the 32 treatment combinations and compared to the experimental observations. Modeling of N uptake and associated crop growth is concep-

tually similar in most crop growth models, including the four introduced above. The implementation in the CropSyst model was used for this evaluation. Details of concepts and equations used to model crop growth as affected by water and nitrogen availability are given elsewhere (Stockle et al., 1994; Stockle and Nelson, 1996). Table 1 shows the crop input parameters used to implement simulations of winter wheat growth. These parameters are typical for the wheat cultivar Soissons grown in Auzeville. Phenology was adjusted as observed in the experimental plots. A few calibrated crop parameters (Table 1) were set based on plots not included in the evaluation reported herein.

4. Results and discussion

4.1. Evaluation of four approaches to model crop nitrogen requirements Fig. 1 compares the performance of the four approaches evaluated. The lower curve (Nmin) in the Table 1 Summary of crop parameters for CropSyst simulations Parameters Degree - days emergence (°C - days) D e g r e e - days begin flowering ( ° C - days) Degree - days peak LAI (°C - days) D e g r e e - days begin grain filling (°C - days) Degree - days maturity (°C - days) Base temperature (°C) Cutoff temperature (°C) Maximum root depth (m) Maximum LAI Specific leaf area (mE/kg) Stem/leaf partition coefficient Leaf duration (°C - days) Solar radiation extinction coefficient ET crop coefficient Maximum water uptake rate (mm/day) Critical canopy water potential (kPa) Wilting canopy water potential (kPa) Biomass-transpiration coefficient (Pa) Radiation-use efficiency (g/MJ) Maximum harvest index, HI

Obs Obs Obs Obs Obs Man Man Obs Obs Obs Obs Obs Man Cal Man Man Man Cal Man Obs

150 1600 1550 1780 2350 0 25 1.5 7 22 2.2 1200 0.46 1.2 10 -1300 -2000 4.5 3 0.45

Parameters were set as observed experimentally (Obs), extracted from the CropSyst manual (Man), or set by calibration (Cal).

221

AFRCWHEAT2 model effectively includes the observed minimum concentrations, with a few data points lying below the curve. However, the upper limit curve (Nmax) has problems up to stage 35 (midshooting), and again at about stage 90 (maturity), with many data points (even for N-limited conditions) above the Nmaxcurve. This will result in underprediction of the N requirement of wheat, thus affecting the simulation of crop growth in relation to available N. The performance of the approach in the EPIC model also presents problems. The Nm~,curve appears adequate, but the Ncnt curve has problems separating N-limited from non-limited data points, particularly before a fraction of the cycle of 0.4. Because crop N stress is simulated once plant N concentrations are below Ncr~t,this will lead to misrepresentation of stress during early growth (before completion of 40% of the growing cycle), a critical period of wheat growth and development. In addition, because a curve for Nmaxis not specified, crop N requirements are underpredicted, which may lead to improper simulation of crop growth in relation to available N, and will misrepresent the potential for N extraction. The implementation in the Daisy model, which includes three characteristic curves, has the worst performance of all approaches tested. Throughout most of the growing season, described in terms of temperature sums, most data points from plots not limited by N and a significant amount of N-limited data points are above the Nmax curve. Under these conditions, wheat N requirements will be severely underpredicted. The Ncnt curve does not discriminate between N-limited and non-limited data points. The Nmincurve, however, represents well the minimum concentrations observed. Given this performance, this approach should have significant limitations in predicting crop N requirements and crop response to N availability for these Auzeville data. The method based on growth dilution (CropSyst) is able to describe well the three characteristic curves that define crop N requirements and crop response to N availability. The Nerit curve discriminated well between N-limited and non-limited data points. The Nmax and Nmin curves effectively served as envelope curves to define the maximum and minimum limits for most observed data points with few exceptions. It seems that methods based on the growth dilution concept (Greenwood et al., 1990) should be preferred to

simulate crop N requirements. This approach has also the advantage of being easier to implement across different crop species and cultivars than the crop specific relations of some of the other models, and it is likely to be more transferable among growth environments.

4.2. Simulating nitrogen uptake and crop growth in relation to crop N requirements After establishing the good performance of the growth dilution concept for the simulation of crop N requirements, an analysis of the nitrogen uptake and crop growth associated with such requirements was performed. Fig. 2 shows a comparison of measured (symbols) and simulated (lines) evolution of biomass and plant N concentration for selected plots. Plot labels are composed of a plot identifier (first two characters) and a N treatment identifier (last two characters). As introduced above, NO and N3 correspond to no- and high-nitrogen treatments, respectively. For plots F8-N3 and F8-N0, following pea, there was an excellent agreement between observed and simulated values (good soil structure, large amount of initial N). The same was the case for F6-N3, after early-harvested sorghum. For plots F6-N0 and C5-N3, the agreement was good for plant N concentration, and also for biomass, except for the last (C5N3) or two last (F6-N0) observations. In the case of C5-N3, the plant N concentration was overpredicted for the measurement previous to the last, but the agreement between simulated and observed biomass was good. This resulted in a slight overprediction of simulated N uptake and a final simulated biomass larger than observed. A similar situation was found for plot F6-NO, somewhat enhanced by simultaneous overprediction of biomass and plant N concentration before the last observation. For plot C l-N0, although measured and simulated plant N concentrations agreed well, simulated crop N uptake was insufficient to support the observed biomass production. Fig. 3 shows the comparison between observed and simulated aboveground N uptake for all the treatment combinations. The linear regression between simulated and observed N uptake has an interception of zero and slope of 0.98, very close to the 1:1 line of perfect agreement, but with a weak correlation

222 coefficient (r = 0.822). The agreement between simulated and observed N uptake is low for reliable model applications. One source of uncertainty is the varia-

bility of the five samples used to determine field N uptake. The average coefficient of variation was 15.4%, with a minimum of 5.3% and a maximum of

25 20

6

20

10 gt~

•i

s 0

Plot FS-N3 0

,

,

,

,

|

0

50

100

150

200

250

25

25

20

20

,

,

!

!

i

0

50

100

150

200

,0 250

"'

6

ot F6-N0 ,lw

3~

2Z Plot F6-N3 ,

,

,

,

!

0

50

100

150

200

i

0

ZS0

!

|

0

50

-

|

|

,

100

150

200

'

6

25 20

O

250

5~

20

Plot C1-N0

0 IN

,ha

Plot C5-N3 O

.m

3e,

•

.

!

$

O 2

-

Z m

0 0

50

100

150

200

Days after emergence

250

!

!

i

i

,

0

50

100

150

200

0

250

Days after emergence

Fig. 2. Comparison of observed (symbols) and simulated (lines) evolution of biomass and plant N concentration for selected plots (triangle, plant N concentration; circle, biomass; error bars represent + 1SD). The first two characters in the plot labels correspond to plot identification, and the last two characters to N treatment (see text).

223

300

~

250

0

•

•

o 200 .,~

15o Z

00

•-

50

I

I

I

I

I

I

0

50

100

150

200

250

300

Observed N U p t a k e (kg/ha) Fig. 3. Comparison between observed and simulated N uptake for 32 N treatment combinations on winter wheat cultivar Soisson grown at Auzeville, France in 1992-1993.

29.1%. Plotting the 32 means versus all the samples resulted in a linear regression explaining 88.5% of the observed variability. Other sources of variability exist, as discussed below. Fig. 4 compares observed and simulated N uptake as a function of soil N availability. Available soil N was determined as the sum of the initial soil N to a depth of 120 cm, plus fertilization and mineralization during the growing season. Initial soil N fluctuated from 50 to 215 kg/ha, and fertilization from 0 to 200 kg/ha. Soil N mineralization was estimated as 40 kg/ha from simulations, a magnitude in agreement with the N balance sheet method introduced above. Thus, a wide range of available soil N is represented in the 32 treatment combinations included in this study, fluctuating from 90 to 370 kg/ha. For this wide range, the simulated N uptake is strongly linearly correlated with soil N availability (r - 0.998), as expected for a system where no constraints to uptake and growth other than available soil N are assumed (soil moisture and rooting depth were not limiting factors). For the measured N uptake, the correlation is weaker (r - 0.84), with large variability around the regression line. For example, for a soil N availability of about 325 kg/ha, observed N uptake fluctuated from 146 to 257 kg/ha.

Other agronomic factors may have affected crop growth in some of the experimental plots (e.g., rust infection under low-input management) reducing N uptake in relation to N available. Sampling size (5 cores/470-m 2 plots) to determine initial soil N content, a quantity with significant spatial variability, may have been insufficient to obtain adequate measurements. The differences in uptake for similar N availability levels, however, appear high to be explained by this kind of reasoning only. It is possible that physical and/or chemical restrictions to N uptake in the field, other than those normally accounted for by crop growth models, may have also played a role. For example, soil structure could have been affected by wet soil conditions during harvest of some of the preceding crops as well as during sowing of the winter wheat. Fig. 5 shows observed and simulated biomass as a function of the corresponding observed or simulated aboveground N uptake. Simulated data points describe a N uptake/biomass production function that is an upper envelop for the observed data points, the latter showing large variability. This indicates that the characteristic plant N concentration curves were effective in defining crop N requirements and limitations to growth associated with plant N concentration. 31111

m

o o

~2500

~

200-

oo Z

o

1500

0 0

I

o o 0

o O

O~o ~

o

o

~

i~i tt

O~'

100-

0

0

II

~

o

~

"

0

50-

0

•

'

so

o

I"

i

I

I

100

1so

200

2so

.... I

300

I

3so

400

Observed Soil N Availability (kg/ha) Fig. 4. Observed (open circles) and simulated (closed circles) aboveground N uptake as a function of the observed soil N availability (initial soil N + N fertilization + net mineralization).

224

20000-

0 ~IIP~ e 0

• ¢I1

o • o @ d3

o~o~

o

o

OOo o

12000

o~ o 0 •~

i~0 o

8000

4000

0

I

I

I

I

I

50

100

150

200

250

Aboveground

300

N Uptake (kg/ha)

Fig. 5. Observed (open circles) and simulated (closed circles) biomass at harvest as a function of corresponding observed and simulated aboveground N uptake.

Data points falling below the simulated envelop must have resulted from growth reducing factors not accounted for by the model, which would also explain the differences between simulation and measurements shown in Fig. 2. No pattern of relationship between these data points and preceding crop or N input level was found.

5. C o n c l u s i o n s

The use of the growth dilution concept provided a good base to estimate characteristic plant N concentration curves (maximum, critical, and minimum concentrations) throughout the growth cycle. Using the implementation of this concept in the CropSyst model, the simulated critical curve discriminated well between N-limited and non-limited experimental wheat plots growing under a wide range of soil N availability. The simulated maximum and minimum curves effectively served as envelope to define the upper and lower N concentration limits for most observed data. Other approaches, based on empirical relationships with the fraction of the growth cycle, growth stages, or temperature sums, were less satisfactory.

The simulated relationship between biomass production and N uptake correctly described an upper boundary for all observed data points. Points falling below the curve represented limitation to growth other than nitrogen. Despite adequate simulation of crop N requirements, simulated and observed crop N uptake on a plot by plot basis resulted in low agreement. Simulated crop N uptake was highly correlated with soil N availability, while the correlation for the observed data was weaker. Uncertainty in the measurement of initial soil N and crop N uptake, other growth reducing factors (e.g., diseases) not accounted for by the model, and possibly physical and/or chemical restrictions to field N uptake not normally included in crop growth models may have limited the ability to simulate crop growth in response to N availability.

References

Debaeke, P., Aussenac, T., Fabre, J.L., Hilaire, A., Pujoi, B. and Thuries, L. 1996. Grain nitrogen content of winter bread wheat (Triticum aestivum L.) as related to crop managementand to the previous crop. Eur. J. Agron., 5: 273-286. Greenwood, D.J., Lemaire, G., Gosse, G., Cruz, P., Draycott, A. and Neeteson, J.J. 1990. Decline in percentage N of C3 and C4 crops with increasing plant mass. Ann. Bot., 66: 425-436. Hansen, S., Jensen, H.E., Nielsen, N.E. and Svendsen, H. 1991. Simulation of nitrogen dynamics and biomass production in winter wheat using the Danish simulation model Daisy. Fert. Res., 27: 245-259. Jones, C.A. and Kiniry, J.R. 1986. CERES-Maize, a Simulation Model of Maize Growth and Development. Texas A and M University Press, Texas, TX, 193 pp. Justes, E., Mary, B., Meynard, J.M., Machet, J.M. and ThelierHuch6, L. 1994. Determination of a critical nitrogen dilution curve for winter wheat crops. Ann. Bot., 74: 397-407. Pl6net, D. 1995. Fonctionnement des cultures de mai's sous contrainte azot6e. Doctoral Thesis, Acad6mie de Nancy-Metz, lnstitut National Polytechnique de Lorraine, France, 247 pp. Porter, J.R. 1993. AFRCWHEAT2: a model of the growth and development of wheat incorporating responses to water and nitrogen. Eur. J. Agron., 2: 69-82. R6my, J.C. 1981. Etat actuel et perspectives de la mise en oeuvre des techniques de pr6vision de la fumure azot6e. C. R. Acad. Agric. Fr., 67: 859-874. Salette, J. and Lemaire, G. 1981. Sur la variation de ia teneur en azote des gramin6es fourrag~respendant leur croissance: formulation d'une loi de dilution. C. R. Acad. Sci. Paris Ser. III, 292: 875-878. Stockle, C.O., Martin, S. and Campbell, G.S. 1994. CropSyst, a

225

cropping systems model: water/nitrogen budgets and crop yield. Agric. Syst., 46: 335-359. Stockle, C.O. and Nelson, R. 1996. CropSyst User's Manual. Biological Systems Engineering Department, Washington State University, Pullman, WA, 186 pp. Williams, J.R., Jones, C.A., Kiniry, J.R. and Spanel, D.A.

1989. The EPIC crop growth model. Trans. ASAE, 32: 497511. Zadoks, J.C., Chang, T.T. and Konzak, T.T. 1974. A decimal code for the growth stages of cereals. Weed Res., 14: 415421.



227

Maize production in a grass mulch system - seasonal patterns of indicators of the nitrogen status of maize B. Feil a'*, S.V. Garibay b, H.U. Ammon b, P. Stamp a alnstitute of Plant Sciences, ETHZ, Universitdltstrasse2, CH-8092 Zurich, Switzerland bEidgenOssische Forschungsanstaltfiir Agrari~kologie und Landbau, Reckenholzstrasse 161, CH-8046 Ziirich, Switzerland Accepted 13 April 1997

Abstract In hilly regions with high precipitation, planting maize (Zea mays L.) into living mulches of winter cover crops may alleviate some of the problems (erosion, runoff of agrochemicals, nitrate leaching) associated with conventional maize cropping. Nitrogen was found to be more yield-limiting for maize planted into a grass sod than for traditionally grown maize. The aim of the present study was to follow the seasonal patterns of some indicators of the N status (whole-shoot concentrations of N and nitrate as well as leaf greenness, as measured by the SPAD 502 chlorophyll meter) of silage maize in three cropping systems. The experiments were conducted in the midlands of Switzerland, where annual precipitation is high (>1000 mm), for 3 years. The cropping systems were PLOUGH (maize sown into the bare autumn-ploughed soil), GRASS/ HERB (maize planted into a living, subsequently herbicidally killed Italian ryegrass [Lolium multiflorum Lam.] sod), and GRASS/MECH (similar to GRASS/HERB, but the growth of the ryegrass sod was mechanically regulated). Planting was done with a strip tillage seeder. There were two N treatments: 110 and 250 kg N/ha (inclusive of the mineral N content of the soil from 0 to 90 cm depth). With 110 kg N/ha, the following differences between the cultural systems were found: (i) as early as the 3rd leaf stage, PLOUGH maize exhibited higher whole-shoot concentrations of nitrate than sod-planted maize; (ii) between the 3rd and 6th leaf stages, chlorophyll meter measurements revealed that the uppermost unfolded leaves of PLOUGH maize were markedly greener than those of maize on the mulch seeding plots; (iii) from the 6th leaf stage onwards, PLOUGH maize showed a higher N concentration than did maize grown in grass swards. Under 250 kg N/ha, the differences between the cultivation methods were much less pronounced. It is concluded that efforts to optimize the environmentally friendly living mulch systems should focus on reducing the competition between maize and the cover crop for N. © 1997 Elsevier Science B.V.

Keywords: Living mulch; Maize; N status; SPAD 502; Strip tillage

1. Introduction The Swiss midlands are hilly and the annual precipitation is high (>1000 mm). Most of the area devoted to maize production is moldboard ploughed *Corresponding author. Tel.: +41 1 6324746; fax: +41 1 6321143; e-mail: [email protected]

in autumn; the fields remain unplanted in winter, and the maize is sown into a finely pulverized seedbed. Sloping cropland treated in this manner is susceptible to erosion. Further environmental problems associated with conventional maize cropping practices are surface runoff of herbicides (Rtlttimann et al., 1995) and plant nutrients (Prasuhn and Braun, 1994) and nitrate leaching into the ground water.


228

Research conducted at various locations in the Swiss midlands and in the Jura range demonstrated that sowing maize in winter cover crop residues, killed by frost or herbicides, in conjunction with minimum tillage is a very effective means of controlling soil erosion and runoff of agrochemicals (Rtittimann et al., 1995). Under these cropping systems, however, maize produced lower silage yields than under plough tillage (Rtiegg et al., 1997). Another approach to alleviate the environmental problems associated with conventional maize production is to plant the maize into a living winter cover crop sward (Box et al., 1980; Echtenkamp and Moomaw, 1989; Klocke et al., 1989). A number of advantages of maize production in living mulches of Italian ryegrass have been reported. In contrast to non-winterhardy cover crops, Italian ryegrass provides a considerable hay or haylage yield in spring (Garibay et al., 1997). After the harvest of maize, the cover crop sward can be used for autumn pasture. It was found that mechanically stunted grass strips between the maize rows harbored many predatory insects and spiders; the maize was markedly less infested by maize smut (Ustilago maydis), aphids (Rhopalosiphon maidis), and European corn borers (Ostrinia nubilalis) (Bigler et al., 1995a,b). J~iggi et al. (1995) observed that the earth worm (Lumbricus terrestris) biomass was significantly higher than under plough tillage. Planting maize into living mulches of Italian ryegrass reduces the need for herbicides and may help to prevent the development of herbicide-resistant weed populations (Ammon et al., 1995). After the maize harvest, the grass strips regrow and remove mineral N from the soil, thus reducing the hazards of soil erosion and nitrate leaching during the cool season. Garibay et al. (1997) reported that maize planted into living mulches of Italian ryegrass has high N requirements to reach maximum yield, thus suggesting that cover crop and main crop compete for N. It is unclear when differences in N status between sodplanted maize and maize grown conventionally occur. Tissue concentrations of N (Cerrato and Blackmer, 1991; Binford et al., 1992) and nitrate (Iversen et al., 1985; McClenahan and Killorn, 1988) as well as leaf chlorophyll contents (Dwyer et al., 1991; Blackmer and Schepers, 1994) have been proposed as indicators of the N status of maize. To determine the periods during which differences in N status between

sod-planted and plough tilled maize are detectable, we monitored the seasonal patterns of the concentrations of N and nitrate in the tops and the chlorophyll contents of the uppermost fully expanded leaves.


2.1. Experimental site and climatic conditions Field experiments were conducted in 1990/91, 1991/92, and 1992/93 on a farm near Zurich (47030 ' N, 8030 ' E, 424 m asl). The soil was an Eutric Cambisol (FAO classification) with 480 g/kg sand, 360 g/ kg silt, 160 g/kg clay, and 21 g/kg organic matter in the top soil (0-30 cm). Cropping system precipitation (precipitation during the 14 month period from August to September) was 1004 mm, 1990/91; 779 mm, 1991/92; and 1099 mm, 1992/93. During the growth cycle of maize, precipitation totaled 268 mm, 1991; 303 mm, 1992; and 489 mm, 1993; the rainfall distribution during the maize growing season in the various years is given in Table 1. The mean daily temperature during the growth cycle of maize ranged from 16.6°C in 1993 to 18.0°C in 1991. The levels of soil P and K (Garibay et al., 1997) were sufficient for optimum plant growth. The experimental fields had been planted previously with winter wheat in 1990 and 1992 or winter rye in 1991.

2.2. The maize cropping systems The cover crop stands (Italian ryegrass) were established in the late summer of the year preceding the planting of maize. Three maize cropping systems Table 1 Monthly rainfall (mm) during the maize growing seasons and 40 year average Year Month

1991

May 101 June 174 July 56 August 17 September 68

1992

1993

40 year average

22 112 86 74 41

80 89 182 92 89

90 113 108 120 80

229

were 75 cm apart. In the GRASS/HERB and GRASS/ MECH systems, only the tilled strips were sprayed with 9 l/ha Primafit A ® (190 ml/l metolachlor plus 95 ml/l atrazine plus 95 ml/l pendimethalin) for weed control. In contrast, in the PLOUGH system, the entire area was treated with 9 l/ha Primafit A ®. The grass strips between the maize rows in the GRASS/HERB system were killed by applying 30 g/ha of the herbicide Titus ® (25 g/kg rimsulfuron) at the 1st and 2nd leaf stages of maize. Grass growth on the GRASS/MECH plots was suppressed with a one row 80 kg heavy mulching machine at the 1st, 3rd, and 6th leaf stages of the maize crop. More information on the cropping systems is given in Garibay et al. (1997).

were compared: (i) PLOUGH (the ryegrass stands were broken up with a moldboard plough in autumn/ winter; the maize was sown into the bare soil), (ii) GRASS/HERB (the maize was planted into a living ryegrass sod; the grass strips between the maize rows were killed by post-emergence applications of a herbicide), and (iii) GRASS/MECH (the maize was sown into a living ryegrass sod whose growth was suppressed mechanically). Planting was done with a one-pass minimum strip tillage seeder which operated at a depth of 15 cm and tilled 30 cm wide bands of soil. 2.3. Cover crop and maize management

Italian ryegrass (Lolium multiflorum Lam. cv. Lipo) was drilled at a rate of 20 kg seed/ha in August on the entire experimental area. Thirty kilograms N/ha (ammonium nitrate) was broadcast approximately 1 month after sowing. The ryegrass was clipped and removed from the plots in October. One-third of the plots was moldboard ploughed in autumn/winter. The grass on the remaining plots was clipped at ground level and removed a second time in spring, shortly before sowing of maize. Maize (Zea mays L. cv. Atlet; FAO 250) was planted at 100000 seeds/ha in May. The final plant densities were 9.6 (PLOUGH; GRASS/HERB), and 9.4 (GRASS/MECH) plants/m 2. The seeding depth was about 5 cm. The maize rows

2.4. Fertilization

There were two N treatments (referred to subsequently as N110 and N250): 110 kg N/ha (mineral N content of the soil plus fertilizer N applied at sowing) and 250 kg N/ha (N110 plus fertilizer N applied at the 4th and 6th leaf stages of maize) (Table 2). Rates of 40 kg (1991), 26 kg (1992), and 41 kg (1993) N/ha were applied as ammonium nitrate to all plots with the planter; the fertilizer was mixed with the soil and evenly distributed in the tilled strips by the rotary hoe. The remainder of the N was applied by hand immediately after planting; the N fertilizer

Table 2 Mineral N content of the soil (0-90 cm depth) shortly before maize planting, and rates and timings of N application PLOUGH i 991

NIIO Maize planting c N250 Maize planting c 4th leaf stage d 6th leaf stage d

GRASS a 1992

1993

1991

1992

1993

Soil mineral N (kg N/ha) b 70 84 Fertilization (kg N/ha)

69

23

23

I1

40

26

41

87

87

99

40 70 70

26 70 70

41 70 70

87 70 70

87 70 70

99 70 70

aGRASS/MECH and GRASS/HERB. bMeasured according to Wehrmann and Scharpf (1979), modified by Walther (1983). CUnder PLOUGH, the N fertilizer was incorporated in the tilled strips; the GRASS plots were treated in the same manner, but the remaining N fertilizer was placed right in the maize rows. dplaced right in the maize rows.

230

was placed right in the maize rows and not incorporated.

2.5. Plant sampling, plant analyses, and SPAD measurements Maize was harvested when 50% of the plants on the PLOUGH plots had reached the 3rd, 6th, and 9th leaf stages (fully expanded leaves), at 50% pollen shedding, and at silage maturity ( - 320 g dry matter/kg fresh weight). On average, these sampling dates corresponded to 204, 364, 510, 648, and 1176°Cd after maize planting (8°C base temperature). Ryegrass was sampled on the same dates as the maize and, in addition, at the 1st leaf stage of maize. The grass was cut at ground level. The sampling area was 3 m 2 except at silage maturity when it was 21 m 2. Aliquots of the maize and grass samples were dried at 65°C for 72 h and ground to pass through a 0.75-mm screen. Plant material collected in 1991 was digested with hot sulfuric acid at 150°C and afterwards at 420°C for 90 min; the analysis for ammonium was done with an autoanalyzer (Autoanalyzer II, Technicon Industrial Systems). The concentration of N in the remaining samples was assessed with a micro-Kjeldahl system (Kjeltec Auto 1030, Tecator). For the determination of nitrate, 50 mg plant material was incubated with 0.5 ml of 80% ethanol for 10 min at 60°C. After adding 5 ml of bidistilled water, samples were placed in a shaking-bath at 60°C for 50 min. The tubes were centrifuged at 3000 rev./min for 10 min; the supernatant Table 3 Dry matter yields (t/ha) of silage maize under three cropping systems and two N treatments. Data are means across three cropping years Cropping system

N treatment

Yield

PLOUGH GRASS/HERB GRASS/MECH PLOUGH GRASS/HERB GRASS/MECH

NIl0 NI I0 Nll0 N250 N250 N250

16.3 11.2 7.6 17.8 16.6 15.1 0.5

SE F-test Cropping system N treatment Cropping system x N treatment Significant at *P = 0.05 and **P = 0.01.

was analyzed for nitrate with an autoanalyzer (Evolution II, Alliance Instruments). To estimate the leaf chlorophyll content, the SPAD 502 chlorophyll meter from Minolta was used. Meter readings were taken on ten representative plants from each of the two center rows on the uppermost fully expanded leaf, midway between the base and tip and the leaf margin and midrib of the lamina.

2.6. Experimental design and statistical analysis The experiment was laid out as a randomized complete block design with four replicates. In the analysis of variance, years were treated as random effects (Gomez and Gomez, 1984).

3. Results

3.1. Maize silage yield The analysis of variance indicated a significant (P = 0.05) year x cropping system x N rate interaction, but the ranking of the cropping systems was similar in the various cropping years (Garibay et al., 1997). This paper, therefore, presents yield means across the years only. With N ll0, the GRASS/ MECH system produced only 47% and the GRASS/ HERB system only 69% of the total aboveground dry matter of maize produced under PLOUGH (Table 3). As compared with the N 110 treatment, the N250 treatment resulted in yield gains for all cropping systems, but the increment was smallest for the PLOUGH system. Nevertheless, PLOUGH was still the most productive cropping method; maize grown under GRASS/HERB and GRASS/MECH produced 95% and 85% of the dry matter produced under PLOUGH. The cropping system x N rate effect on dry matter yield was highly significant (P = 0.01).

3.2. Seasonal trends of some indicators of the N status of maize The N concentration in the tops was initially similar under all maize production systems (Fig. l a). From the 6th leaf stage onwards, however, the N concentration was higher in PLOUGH maize than in sodplanted maize. Irrespective of N level, GRASS/

231

0

--,

!

,

I

'

I

50

'

I

'

8.0

Ill

•

s.o

b

t

-

a 40

.|

4.0

30

2.0 Z

10 0.0 0

t.l

0

.

l

3OO 6 0 o

,

l

9oo

,

I

1200

Growing degree days (*C d)

0

300 600 900 1200

Growing degree days (*C d)

Fig. 1. Concentrations of N (a) and nitrate (b) in the tops of silage maize under three cropping systems, two N treatments, and five developmental stages (3rd, 6th, and 9th leaf stages, pollen shedding, silage maturity). Data are means across two (nitrate concentration at the 3rd leaf stage) or three cropping years. Bars are SE.

HERB maize tended to have a higher N concentration than maize under GRASS/MECH, but the differences were smaller with N250 than with N110. At the 3rd leaf stage, the N contents of the maize tops were highest under PLOUGH and lowest under GRASS/HERB (Table 4). From the 6th leaf stage onwards, however, GRASS/HERB maize contained more N than did GRASS/MECH maize; PLOUGH maize had still the highest N content. The N treatment did not affect the ranking of the cropping systems, but the differences among the cropping systems were more pronounced under N 110 than under N250.

At the 3rd and 6th leaf stages, nitrate levels were by far the highest for PLOUGH maize (Fig. l b). With N II0, the nitrate concentrations decreased rapidly, independent of cropping system. From the 9th leaf stage onwards, maize on the GRASS/HERB and GRASS/MECH plots contained only trace amounts of nitrate, whereas in PLOUGH maize, small quantities of nitrate were still detected at pollen shedding. Under N250, maize in the PLOUGH system had the highest nitrate concentration throughout the cropping period. While differences between GRASS/HERB and GRASS/MECH maize were still small at the 3rd leaf stage, nitrate concentrations tended to be higher for GRASS/HERB maize than for GRASS/MECH maize at the 6th and 9th leaf stages. In interpreting the changes in SPAD readings over time (Fig. 2), it must be taken into account that the measurements were made on different leaves. Maize development was only slightly affected by the cropping systems; it is possible, therefore, to compare the cropping methods on a given observation date. With N110, PLOUGH maize had the greenest leaves from the third (1991 and 1993) or from the first measurement (1992) onwards. In 1991 and 1993, differences between the mulch seeding systems occurred when the plants had reached or just passed the 6th leaf stage, while in 1992, the curves started to diverge only after pollen shedding. From these developmental stages onwards, the SPAD readings were slightly (1992) or clearly (1991 and 1993) higher under the

Table 4 Nitrogen content of the tops of silage maize (kghaa) under three cropping systems, two N treatments, and five stages of development. Data are means across three cropping years Cropping system

N treatment

3rd leaf stage

6th leaf stage

9th leaf stage

Pollen shedding

Silage maturity

PLOUGH GRASS/HERB GRASS/MECH PLOUGH GRASS/HERB GRASS/MECH SE F-test Cropping system N treatment Cropping system x N treatment

N110 N110 N110 N250 N250 N250

4.3 3.6 4.1 4.1 3.7 3.9 0.2

35.2 22.9 18.9 38.0 24.7 22.7 0.9

89.5 47.9 29.1 107.8 73.7 61.1 1.5

110.9 56.7 36.7 156.0 123.7 108.5 3.4

157.9 84.4 49.3 206.4 174.7 158.0 5.9

ns

*

*

**

**

ns

ns ns

(*) *

** *

* *

ns

Significant at (*)P = 0.10, *P = 0.05 and **P = 0.01. ns not significant.

232

•

I

'

I

Nl10 ~ ~

50

'

i

•

i

'

-?

i

'

I

N250 ~~~

'

i

'

competed with the maize crop after the first herbicide application.

-

I

1-

~

40

•

1991

~ 3o

4. Discussion

or) 20

_

-~. , •

I i

, •

I I

, '

I i

, '

,'

1 1

Ii

,•

Ii

', ~

I

,

II

5O w

~ 40

1992

O.

50 c

GRASS/HERB GRASS/MECH

"'0-_

20

~

~ -- N250~~

~I~LO ~,

-

-

40

1993

n 20 , 300

600

I 900

I 1200

,

I

,

300

Growing degree days

I 600

(*C

,

I 900

,

I 1200

d)

Fig. 2. Seasonal trends in chlorophyll meter readings taken on the uppermost fully expanded leaves of silage maize under three cropping systems and two N treatments. Arrows indicate the 3rd, 6th, and 9th leaf stages, pollen shedding, and silage maturity. GRASS/HERB system than under the GRASS/MECH system. With N250, the variations between the maize production systems were, generally, fairly small and inconsistent over time.

3.3. Nitrogen contents of the grass strips Immediately before the grass strips were mulched, i.e. at the 1st, 3rd, and 6th leaf stages of the maize, the N contents of the grass strips in the GRASS/MECH system varied from about 20 to 30 kg/ha (Table 5). After the last mowing, the grass swards recovered slowly, and the grass N contents remained almost constant (N110) or even decreased (N250) during grain filling. At the 1st leaf stage, i.e. prior to the first herbicide application in the GRASS/HERB system, grass N contents were the same in the GRASS/ MECH and GRASS/HERB systems. Even though the herbicide treatments were carded out between the first and second samplings, the grass in the GRASS/HERB system accumulated an additional 8 kg N/ha (mean of N I I 0 and N250) between the 1st and 3rd leaf stages (data not shown), indicating that the cover crop still

The present data demonstrate that differences in the N status of PLOUGH maize and sod-planted maize already occur at very early stages of maize development, even though more fertilizer N (on average 56 kg/ha) was invested in the GRASS/HERB and GRASS/MECH systems to offset the low mineral N contents of the soil immediately prior to maize sowing (Table 2). Furthermore, it is noteworthy that the extra N applied to the GRASS/HERB and GRASS/MECH plots was placed in the maize rows, suggesting that the spatial distribution of N was more favorable for maize in the mulch seeding systems than for PLOUGH maize. Nevertheless, as early as the 3rd leaf stage, low nitrate concentrations indicated the beginning N deficiency of sod-planted maize (Fig. l b). It was not until the 6th leaf stage that differences in the wholeplant N concentration between the cropping systems became detectable (Fig. I a). Even though concentrations of N tend to increase with increasing availability of N (Binford et al., 1992), the tissue N concentration is not a perfect indicator of the N supply to the plants. First, the tissue test for N has limited ability to detect excesses of available N in the soil (Binford et al., 1992). Second, it may be that differences in the concentration of N merely reflect differences in the developmental stage of the plants because the concentration of N declines with increasing age, i.e. increasing plant Table 5 Nitrogen content of the aboveground biomass of the grass strips (kg/ha) in the GRASS/MECH system under two N treatments. Corrections were made to account for the fact that the tilled strips were not grassed over. Data are means across 3 years Developmental stage

N 110

N250

SE

F-test~

Ist leaf stage 3rd leaf stage 6th leaf stage 9th leaf stage Pollen shedding Silage maturity

i 7.5 20.3 27.3 9.8 17.1 18.1

20.0 20.7 30.0 ! 2.6 19.4 9.4

1.1 0.9 2.2 1.1 1.9 4.8

ns ns ns ns ns ns

ans not significant.

233

mass (Greenwood et al., 1990). Third, under severe N deficiency, improving the availability of N to the crop may cause slight decreases in the tissue N concentration. This phenomenon is usually explained by nutrient dilution brought about by the higher production of plant dry matter (Wikstr6m, 1994). Consequently, the observation that, from the 6th stage onwards, PLOUGH maize showed the highest whole-plant concentration of N does not prove that more N was available to PLOUGH maize than to sod-planted maize. However, PLOUGH maize also had the highest N shoot content (Table 4), indicating that less N was available to sod-planted maize than to PLOUGH maize. The observation that, at the 3rd leaf stage, the tops of GRASS/HERB maize contained less N than those of GRASS/MECH maize may be traced to herbicide stress in the GRASS/HERB system. Before the 6th leaf stage was reached, plants treated with N 110 showed another symptom of N stress: the uppermost fully expanded leaves of sod-planted maize were slightly paler than those of conventionally produced maize (Fig. 2). Low SPAD readings are indicative of low concentrations of chlorophyll (Dwyer et al., 1991) and N (Schepers et al., 1992; Blackmer et al., 1994; Dwyer et al., 1995). In fact, from the 6th leaf stage onwards, the whole-shoot concentration of N was lower for the plants in the mulch seeding systems than for those on the PLOUGH plots (Fig. l a). Thus, the SPAD instrument detected a reduced N supply to the plants at a relatively early stage. The determination of the plant N status with a chlorophyll meter is simple, fast, and non-destructive, suggesting that the SPAD 502 instrument is a useful tool for testing tissues and may aid in the management of N fertilizer. However, the meter's costs may be prohibitive to many farmers. Furthermore, at least in the present study, the SPAD technique was often unable to detect the differences in N status between the cropping systems when the N supply was high. This may be due to the fact that SPAD readings reach a plateau at high leaf N concentrations (Dwyer et al., 1995). It is possible that the SPAD method provides better results if the measurements are made on older leaves (Piekielek and Fox, 1992), i.e. leaves with relatively low N concentrations. While tissue testing may be very helpful in detecting cropping system effects on the N status of maize in a given environment, for a given cultivar, and at a given stage

of development, the applicability of tissue analyses in the management of N fertilizer is limited by the lack of generally valid critical values (Cerrato and Blackmet, 1991). Binford et al. (1992) concluded that a tissue test based on concentrations of N in young plants is not a viable alternative to soil nitrate tests. Unfortunately, soil testing for mineral N in maize cropping systems which use living mulches seems to be impractical for two reasons. In living mulch systems, N fertilizer placement in or near the maize rows is a prerequisite for attaining acceptable maize yields (Garibay et al., 1997). Banding of N fertilizer leads to an extremely uneven distribution of mineral N in the soil, and the results of soil tests for mineral N, therefore, are strongly dependent on the sampling location (Garibay, 1996). Furthermore, it is unclear to which extent mineral N dectected in the soil is available to the main crop. Various factors may have contributed to the relatively poor N status of sod-planted maize. First, it is possible that more soil N was mineralized on the autumn-ploughed land than on the strip-tilled plots where roughly two-thirds of the topsoil remained undisturbed (Dowdell and Cannell, 1975; Powlson, 1980). Second, it is likely that some N was microbially immobilized in the tilled strips. Data presented by Jensen (1991) indicate that the C/N ratio in the roots of well developed Italian ryegrass stands is fairly high (60-70). Roots were reported to make up more than 60% of the total dry matter of the residues (stubble plus roots) of Italian ryegrass (Renius et al., 1992), suggesting that the average C/N ratio in the grass residues is above the critical value for net N immobilization. The third and likely the most important factor is that the grass strips removed N from the soil (Table 5). The grass N content was almost unaffected by the level of N supply. This can be attributed, in part, to the fact that the N fertilizer was placed in the maize rows. Furthermore, due to its larger canopy (data not shown), maize under N250 was more competitive with the grass strips than it was under N110. The stagnation (N110) of and decline (N250) in grass N after pollen shedding is probably due to the fact that the grass was strongly shaded by the fully developed maize canopy during the second half of the growing season. In the GRASS/MECH system, at least 80 kg mineral N/ha was taken up by the grass (summed

234

grass N contents on the mulching dates plus maximum grass N content later in the season) and converted to organic N (Table 5). The bulk of grass N returned to the soil surface after mulching. It is assumed that some grass N was mineralized and soon reassimilated by the regrowing grass strips, while the rest remained organically bound. There is no indication in Fig. l a,b and Fig. 2 that the decaying grass residues provided significant amounts of N to the maize crop late in the season. In evaluating the ecological value of the mulch seeding systems tested, it must be considered that N bound in the grass debris may be mineralized after tillage operations, thus increasing the nitrate leaching hazard during periods when the soil is uncropped or when the N uptake capacity of the crop is limited. It is concluded that efforts to optimize the environmentally sound living mulch systems should focus on reducing the competition between maize and the cover crop for N. An earlier and/or a more effective suppression of the cover crop sod would help to improve the availability of N to the main crop. Another approach is the use of legumes instead of Italian ryegrass as a cover crop. Legumes have relatively low C/N ratios (McKenney et al., 1993) suggesting that, in the tilled strips, N would be released rather than immobilized. Legume cover crops are less vigorous than Italian ryegrass and, therefore, are weaker competitors for N. Furthermore, as compared to Italian ryegrass, a more rapid decomposition of the debris in the mechanically stunted cover crop strips and, thus, a faster release of N may be expected. A negative aspect of leguminous winter cover crops is, however, that they provide only low dry matter yields in autumn and spring.

Acknowledgements This project was funded by the Swiss Federal Government and carried out within the scope of the European program COST 814.

References Ammon, H.-U., Scherrer, C. and Mayor, J.-P., 1995. Unkrautentwicklung und Bodenbedeckung. Agrarforsch., 2: 369-372.

Bigler, F., Waldburger, M. and Frei, G., 1995a. Vier Maisanbauverfahren 1990 bis 1993: Krankheiten und Schadlinge. Agrarforsch., 2: 380-382. Bigler, F., Waldburger, M. and Frei, G. 1995b. Vier Maisanbauverfahren 1990 bis 1993: lnsekten und Spinnen als NUtzlinge. Agrarforsch., 2: 383-386. Binford, G.D., Blackmer, A.M. and Cerrato, M.E., 1992. Nitrogen concentration of young corn plants as an indicator of nitrogen availability. Agron. J., 84, 219-223. Blackmer, T.M. and Schepers, J.S., 1994. Techniques for monitoring crop nitrogen status in corn. Commun. Soil Sci. Plant Anal., 25. 1791-1800. Blackmer, T.M., Schepers, J.S. and Varvel, G.E., 1994. Light reflectance compared with other nitrogen stress measurements in corn leaves. Agron. J., 86: 934-938. Box, Jr., J.E., Wilkinson, S.R., Dawson, R.N. and Kozachyn, J., 1980. Soil water effects on no-till corn production in strip and completely killed mulches. Agron. J., 72: 797-802. Cerrato, M.E. and Blackmer, A.M., 1991. Relationships between leaf nitrogen concentrations and the nitrogen status of corn. J. Prod. Agric., 4: 525-531. Dowdell, R.J. and Cannell, R.Q., 1975. Effect of ploughing and direct drilling on soil nitrate content. J. Soil Sci., 26: 53-61. Dwyer, L.M., Tollenaar, M. and Houwing, L., 1991. A nondestructive method to monitor leaf greenness in corn. Can. J. Plant Sci., 7 I: 505-509. Dwyer, L.M., Anderson, A.M., Ma, B.L., Stewart, D.W., Tollenaar, M. and Gregorich, E., 1995. Quantifying the nonlinearity in chlorophyll meter response to corn leaf nitrogen concentration. Can. J. Plant Sci., 75:179-182. Echtenkamp, G.W. and Moomaw, R.S., 1989. No-till corn production in a living mulch system. Weed Technol., 3: 261-266. Garibay, S.V., 1996. Maize Production in Living Mulches in a Humid Temperate Climate. Ph D thesis, ETH Zurich, Switzerland. Garibay, S.V., Stamp, P., Ammon, H.-U. and Feil, B., 1997. Yield and quality components of silage maize in killed and live cover crop sods. Eur. J. Agron., 6:179-190. Gomez, K.A. and Gomez, A.A., 1984. Statistical procedures for agricultural research. 2nd edn., John Wiley and Sons, New York. Greenwood, D.J., Lemaire, G., Gosse, G., Cruz, P., Draycott, A. and Neeteson, J.J., 1990. Decline in percentage N of C3 and C4 crops with increasing plant mass. Ann. Bot., 66: 425-436. Iversen, K.V., Fox, R.H. and Piekielek, W.P., 1985. The relationships of nitrate concentrations in young corn stalks to soil nitrogen availability and grain yields. Agron. J., 77: 927-932. J~iggi, W., Oberholzer, H.-R. and Waldburger, M., 1995. Vier Maisanbauverfahren 1990 bis 1993: Auswirkungen auf das Bodenleben. Agrarforsch., 2: 361-364. Jensen, E.S., 1991. Nitrogen accumulation and residual effects of nitrogen catch crops. Acta Agric. Scand., 41: 333-344. Klocke, N.L., Nichols, J.T., Grabouski, P.H. and Todd, R., 1989. Intercropping corn in perennial cool-season grass on irrigated sandy soil. J. Prod. Agric., 2: 42-46. McClenahan, E.J. and Killorn, R., 1988. Relationship between basal corn stem nitrate N content at V6 growth stage and grain yield. J. Prod. Agric., 1: 322-326.

235

McKenney, D.J., Wang, S.W., Drury, C.F. and Findlay, W.l., 1993. Denitrification and mineralization in soil amended with legume, grass, and corn residues. Soil Sci. Soc. Am. J., 57: 1013-1020. Piekielek, W.P. and Fox, R.H., 1992. Use of a chlorophyll meter to predict sidedress nitrogen requirements for maize. Agron. J., 84: 59-65. Powlson, D.S., 1980. Effect of cultivation on the mineralization of nitrogen in the soil. Plant Soil, 57: 151-153. Prasuhn, V. and Braun, M., 1994. Absch~itzung der Phosphor- und Stickstoffverluste aus diffusen Quellen in die Gewiisser des Kantons Bern. Schriftenreihe der FAC Liebefeld 17. EidgenGssische Forschungsanstalt fur Agrikulturchemie und Umwelthygiene, CH-3097 Liebefeld-Bern. Renius, W., Liitke Entrup, E. and Liitke Entrup, N., 1992. Zwischenfruchtbau zur Futtergewinnung und Grtindiingung. 3rd Edition. DLG-Vedag, Frankfurt (Main). Riiegg, W.T., Richner, W., Stamp, P., and Feil, B., 1997. Growth

and productivity of minimum tillage maize planted in winter cover crop residues. Eur. J. Agron. (in press). Riittimann, M., Schaub, D., Prasuhn, V. and Rtiegg, W., 1995. Measurement of runoff and soil erosion on regularly cultivated fields in Switzerland - some critical considerations. Catena, 25: 127-139. Schepers, J.S., Francis, D.D., Vigil, M. and Below, F.E., 1992. Comparison of corn leaf nitrogen concentration and chlorophyll meter readings. Commun. Soil Sci. Plant Anal., 23:2173-2187. Walther, U., 1983. Die heutigen Bodenuntersuchungsmethoden im Dienste einer optimalen Pflanzenern~rung. Die Griine, 7: 315. Wehrmann, J. and Scharpf, H.C., 1979. Der Mineralstickstoffgehait des Bodens als Ma/~stab fur den Stickstoffdtingerbedarf (NminMethode). Plant Soil, 52: 109-126. WikstrGm, F., 1994. A theoretical explanation of the Piper-Steenbjerg effect. Plant, Cell Environ., 17: 1053-1060.



237

Nitrogen transformations after the spreading of pig slurry on bare soil and ryegrass using 15N-labelled ammonium T. Morvan a'*, Ph. Leterme a, G.G. Arsene b, B. Mary c aUnit~ INRA d'Agronomie, ENSAR, 65 rue de Saint-Brieuc, F35042 Rennes, France bUSAMVB Timisoara, Colea Arodulini, 19, Timisoara 1900, Romania CUnit~ 1NRA d'Agronomie, rue Fernand Christ, BP 101, F02004 Laon Cedex, France

Accepted 13 April 1997

Abstract A short field experiment (27 days) was carded out in summer 1995, to study the effect of an actively growing grass sward on nitrogen transformations of a pig slurry. The ammonium fraction of the slurry was labelled with (~5NH4)2SO4. The slurry was spread manually on microplots in mid-June, at the rate of 3 l/m2, on a cut ryegrass sward, and compared with bare soil. Absorption of 15N-labelled NH4 by the grass occurred very rapidly, attaining 41% after 13 days and showing no further change at 27 days. The gaseous losses, mainly through volatilization of ammonia, were considerable. ~SN recovery in soil and plant material on day 27 was 42.5% (+1.2) on the bare soil, versus 57.4% (+3.1) on the ryegrass. The grass sward significantly reduced: (i) volatilization, as shown by the difference of 14.9% in ~SNrecovery, on the 6th day; (ii) immobilization, which was 25% (+2.2) on day 27 on bare soil and 16.4% (-1-2.9) in the presence of ryegrass. 15N-labelled inorganic nitrogen was completely depleted beneath the ryegrass, 27 days after application, whereas ammonium was depleted and the nitrate was equal to 16.4% (-I-1.6) of the applied NH4 on the bare soil. It is clearly apparent that the ammonia from the slurry is more efficiently used when applied to an actively growing sward, rather than to bare soil, even though a significant portion of the plant is involved in internal recycling. © 1997 Elsevier Science B.V. Keywords: Slurry; 15N; Grassland; Volatilization; Immobilization

1. Introduction Slurries, because of their high ammonium content, provide nitrogen which is quickly available for crops. The availability of the ammonium fraction is determined by gaseous loss and microbial immobilization. The first occurs mainly through the volatilization of

*Corresponding author. Tel.: +33 99 287231; fax: +33 99 287230; e-mail: [email protected]

ammonia, and is highly variable, (20-70% of total ammonium nitrogen (TAN) applied) (Lauer et al., 1976; Beauchamp et al., 1982; Pain et al., 1989; G6nermont, 1996), whereas microbial immobilization represents 15-35% of the TAN (Morvan et al., 1996). Slurry incorporation slightly reduces ammonia volatilization, but is not always possible, especially as it may lead to plant injury, for example after late winter applications on wheat or rapeseed. Furthermore, high nitrogen utilization efficiencies have been obtained for slurry ammonium

Reprinted from the European Journal o f Agronomy 7 (1997) 181-188

238

after surface spreading of diluted pig slurries in actively growing wheat and were correlated with low levels of volatilization. A reduction in volatilization brought about by living plants has been reported and might be expected from: (i) absorption by plant leaves of ammonia volatilized from the underlying soil, as reported by Denmead et al., 1976; (ii) absorption of ammonium through the roots and (iii) microclimatic effects due to the canopy (Faurie and Bardin, 1979). Contradictory results have however been reported by other authors (Thompson et al., 1990), which could be explained by the greater surface area resulting from slurry retention on the leaves. What effect do plants have upon nitrogen immobilization? We might expect high rates of microbial immobilization under a grass sward, because of the large amounts of root exudate. Short-duration experiments, involving the measurement of actual rates of NH4 uptake by plants and microbes, have shown that microbial immobilization may be five times higher than plant uptake (Jackson et al., 1989). Ledgard et al. (1989), studying the partitioning of ~SN-labelled ammonium applied to grass-clover pasture confirmed that microbial immobilization was a significant component of 15N balance, but also observed that more fertilizer was immobilized when plant growth was slow, due to lower temperatures. Immobilization is known to depend on the amount and persistence of ammonia in the soil. Thus, we may suppose that an actively growing plant able to rapidly absorb significant amounts of inorganic nitrogen, will reduce microbial immobilization. The few studies describing the effect of plants on nitrogen transformations after slurry addition sometimes give contradictory conclusions about volatilization, and rarely provide a complete description of soil-plant behaviour and competition, either because only one process was studied, or because the time scale of the experiment was too short, or too long. The aim of the present work was to obtain a better understanding of the effect of an actively growing plant on nitrogen transformations of a pig slurry ammonia pool, using labelled ammonium. A short field experiment was therefore carried out from midJune to mid-July, in order to ensure that climatic conditions were favourable to plant growth.

Table 1 Physical and chemical properties of the soil, and slurry composition Soil properties Particle size distribution (%) Clay Silt Sand Total N (%) pH KCI Bulk density of the soil layers 0-10 cm 10-20 cm Slurry composition N-NH4 (g/l) (after addition of ammonium sulfate) Total N (Kjeldhal) (g/l) pH Dry matter (%)

14.4 72.5 13.1 0.13 6.2 1.53 1.50 4.02 6.12 7.36 1.4


2.1. Site and design The experiment was conducted at Le Rheu Experimental Station (INRA), in western France, on a loamy soil. Some of the chemical and physical properties of this soil are summarized in Table 1. The fate of the ammonia fraction of a pig slurry was studied using 15N-labelled NH4. Two treatments were compared: surface spreading on a ryegrass sward, and on bare soil. Daily temperatures were relatively high, varying from 15°C to 25°C (the mean air temperature was 20.3°C over the 27 days), and were favourable both to plant growth and to ammonia volatilization and nitrogen biotransformations, such as nitrification, mineralization and immobilization. The amount of rainfall and change in soil moisture in the soil surface layer are shown in Fig. 1; 2 x 10 mm were applied by irrigation, the day before the start of the experiment, and on day 3, to prevent the soil from drying out. The soil moisture therefore varied from 60 to 100% of the field capacity over the first 13 days and was favourable for microbial activity and plant growth. No significant differences were noticed between the two treatments during this period. The plots were 4.70 m x 2.30 m and laid out in a

239

0,25

30

~'25-

0,2

g 820

is

0

E

I

', ', ) ', : : : ', : : I : I'1 : I ;l:m,,-,,-,-: 1 3 5 7 9 11 13 15 17 19 21 23

, I ;

Time

I

i ', , 0

25

27

(days)

I m Rainfall + Irrigation - - Soil moisture BS* (%) 1 --,- Soil moisture R* (%)

nium at the rate of 188 g/ha), followed by removal of the aerial residues, just before spreading. The characteristics of the soil surface (structure, bulk density) and soil moisture, controlling initial infiltration of the slurry, and consequently the intensity of ammonia volatilization, were therefore similar in each treatment. Chemical weed control probably had no effect on the biological activity of the soil as the kinetics and level of immobilization of ZSN-labelled nitrogen were similar to those observed after a comparable slurry application to a soil that was not subjected to chemical weed control (unpublished data).

Fig. 1. Rainfall plus irrigation, and change in soil moisture on the bare soil (BS) and under ryegrass (R).

2.2. Soil and plant sampling

randomized block design with three replications. Each plot was divided into five microplots of 0.60 m 2, corresponding to the sampling area for a given date of measurement. Microplots were separated from each other by a 30-cm discard area and from the boundaries by 50-cm guard strips. The ammonium fraction of a pig slurry, (composition in Table 1), was enriched with 15N using a solution of (15NH4)2SO4 10% atom excess, which was added and thoroughly mixed to the slurry. The initial atom excess of the ammonium pool (1.089%) was determined just after spreading, by sampling the soil surface. The enriched slurry was applied at the rate of 3 l/m 2 on June 20th 1995, between 1500 h and 1645 h, in warm, sunny conditions. The amounts of ammonium and organic nitrogen applied were equal to 121 and 63 kg N/ha respectively. A watering can with a distribution bar was used to obtain as even a distribution as possible, and the plots were divided into microplots each receiving the same quantity of slurry. No run-off was observed after spreading, despite the high soil compaction (bulk density of the 0-10 cm soil layer: 1.53 g/cm3). The slurry was applied: (i) to a 2-year-old cut ryegrass (Lolium perenne) sward, which had not been fertilized with nitrogen since the date of sowing. The nitrogen content in plant material harvested at a height of 0.5 cm was therefore very low (1,12%) at the beginning of the trial. The grass sward was cut one week before the experiment, and was 10-12 cm high on the day of application; (ii) to bare soil, obtained by chemical destruction of the grass (glifosate ammo-

The microplots were sampled 1, 3, 6, 13 and 27 days after slurry spreading. The aerial parts were cut just above the soil, on a square area of 0.25 m 2. The plant material was washed free of soil and slurry, the volume of washing water was measured, and the plants and water sampled for inorganic nitrogen and 15N analysis. Soil samples were taken from the 0-10 cm and 1020 cm depths, to determine root biomass, organic and inorganic soil nitrogen. A 60 mm diameter probe was used for root sampling, and a 20 mm diameter probe for soil sampling. For soil nitrogen analysis, samples were obtained from each microplot by mixing 27 cores from each soil layer, and passing the whole sample through an 8-mm mesh sieve. Soil samples for roots were obtained by mixing six cores taken with the 60-mm diameter probe. The roots were separated from the soil by washing on a 2-mm mesh sieve; herbage residues were removed by hand. Above-ground parts and roots were dried at 60°C, finely ground to powder and put into tin containers for total nitrogen content determinations and ~5N atom excess analysis.

2.3. Analytical procedures Inorganic nitrogen in the soil was determined in a KC1 extract (600 ml 1 M KCI/300 g fresh soil, shaken for 30 min, then filtered through a Whatmann 42 filter), using the fractionated steam distillation with MgO for ammonium and Dewarda's alloy for nitrate analysis (Drouineau and Gouny, 1947). Organic plus clay-fixed nitrogen and ~SN was measured on a sample

240

of moist soil, after removal of the inorganic nitrogen, as described by Recous et al., 1988; the sample was dried at 60°C, and finely ground. Atom excess was determined on subsamples of plant, soil, and dried solutions resulting from the steam distillations, using a total combustion technique linked to a VG SIRA 9 mass spectrometer (Recous et al., 1988).

20

I

15

Z

_=

® 10 JD M Z u~

¥

Bare soil

s

0 3. R e s u l t s

I

Ryegrass]

5

10

15

20

25

30

35

Time (days)

3.1. Inorganic nitrogen dynamics Inorganic nitrogen occurred mainly in the surface soil layer (0-10 cm). The amounts of 15NHn-labelled ammonium observed in the 10-20 cm soil layer did not exceed 0.5% of the applied ammonium. Small amounts of 15N-labelled nitrate, not exceeding 1.3% of the applied nitrogen, were measured in the lower layer of the bare soil, on days 13 and 27. The fate of the ammonium nitrogen is shown in Fig. 2: the curves represent the change in the amount of ammonium in the 0 - 2 0 cm soil layer, and ammonia deposition on the leaves in the ryegrass treatment. The amount of ammonium deposited on the plant leaves was relatively large on day 1: this suggests that the initial deposition was probably considerable, if comparable with the great decrease of the ammonium pool in the soil observed during the first day. Ammonium amounts decreased sharply during the

Fig. 3. The proportions of the ~SN-labelled NH4 present as ~SNlabelled NO3 in the 0-20 cm soil layer. (Vertical bars indicate the standard deviation of the three replicates; SD not shown are smaller than symbol size.) first 6 days. The kinetics of 15N-labelled NH4 were similar in each treatment; the plant tended to stimulate the depletion of ammonium, particularly between days 6 and 13. The change in 15N-labelled NO3 in the soil is shown in Fig. 3. Some nitrification of the slurry ammonium had occurred by the end of the experiment, because the ammonium pool had been depleted by that time, in both treatments. ~SN-labelled NO3 content steadily increased until the end of the experiment in the bare soil, attaining only 16% of the nitrogen applied, but it was completely depleted beneath the sward. We did not observe a true latent period at the beginning of the trial as is often reported (Le Pham et al., 1984).

3.2. Dry matter and nitrogen absorption dynamics ,-=- Bare soil i--~- Ryegrass ! OepositJon

80 ~ z

~

4o 20 0 0

5

10

15

20

25

30

Time (days)

Fig. 2. The proportions of 15N-labelled NH4 in the 0-20 cm soil layer, and deposition on ryegrassleaves, after application of 121 kg N-NH4/ha from a pig slurry. (Vertical bars indicate the standard deviation of the three replicates; SD not shown are smaller than symbol size.)

The pattern of 15N absorption by the whole plant is shown in Fig. 4; nitrogen absorption occurred rapidly: the nitrogen utilization efficiency had already reached 11.5% the day after spreading and had attained 41% by day 13 and remained stable till day 27. Because the soil moisture, temperature and soil ammonia content were similar in both treatments (Figs. 1 and 2) during the first 6 days, it can be assumed that nitrification occurred at the same rate in both treatments, during this period. It then follows from the comparison of 15N-labelled nitrate dynamics and ~SN inorganic nitrogen absorption that absorption was mainly due to the uptake of ammonium or ammonia. The two routes of assimilation, through the roots and leaves, were probably efficient (see Section 4),

241

iod corresponding to the amount of nitrogen taken up by the grass.

35 30 25 z

it)

3.4. 15N balance

2O 15 It

10 5

Roots_J

0 ..

i

I

0

5

10

I

t

15

20

25

30

Time (days)

Fig. 4. The proportions of the ISN-labelled NH4 present in the above-ground parts and roots following the pig slurry application. (Vertical bars indicate the standard deviation of the three replicates; SD not shown are smaller than symbol size.)

and equal quantities can be found in aerial plant parts and roots on days 1 and 3. Plant dry matter and nitrogen content increased simultaneously during the experiment. The aboveground parts of the ryegrass rose from 4.1 to 6.8 t dry matter/ha, representing a mean growth rate of 0.1 t dry matter/ha per day. The nitrogen content of the aerial parts rose from 1.12 to 1.74%. The proportion of tSN-labelled nitrogen measured in the whole plant was 15.8% of the amount of JSN-labelled NH4 applied, on day 1, and reached a maximum value of 33.4% on day 13.

3.3. Immobilization The pattern of immobilization is shown in Fig. 5. In both treatments, ~5N immobilization rose sharply during the first 3 days; on bare soil, it exceeded the value in the ryegrass with 6% as early as the first day, and continued to increase, whereas it remained steady in the ryegrass sward. The difference observed on day 1 might be related to the significant deposition of ammonia and organic carbon from the slurry on the leaves (ammonia deposition representing 7% of the total applied ammonium, on day l) (Fig. 2), which thus reduced the amount of ammonium and organic carbon available in the soil. On the other hand, the increased margin between days 6 and 13 can be explained by the competitive effect of the plants which absorbed the inorganic ~5N-labelled nitrogen, the depletion of the ammonium pool during this per-

Fig. 6 shows the change in non-recovery of the ~SN, corresponding to gaseous losses. Denitrification was probably low during this experiment, in view of the climatic conditions and soil moisture. Ammonia volatilization probably accounts for most of the gaseous losses, all the more so as the kinetics of the ~SN nonrecovery exhibit a typical pattern of volatilization, as described by Sommer and Olesen (1991) and by Jarvis and Pain (1990). In fact, cumulative volatilization can generally be described by an asymptotic exponential curve, and is well fitted to the following equation: I.'-" Vmax ( l - e -kt) (Moal, 1995). The optimal values of the parameters of this equation were calculated for each treatment, using the Nlin procedure in the SAS package (SAS, 1988). The losses were the same on day l, in both treatments, attaining 40% of the applied nitrogen. This suggests that there was no significant microclimatic effect of the canopy, or that differences compensated. Volatilization was significant on the bare soil between days l and 3 (+l 9% according to the adjusted curve), but almost stopped after day 1 on the ryegrass treatment (+2% according to the adjusted curve). This halting of volatilization in the grass sward was unexpected, as (i) the amounts of ammoniacal nitrogen measured in the surface layer were still high on days l, 3 and 6, and were only a few kg N/ha less than the amounts measured on bare soil, (ii) volatilization after 35 u

30

I

•

i- OlV, o 15

.

I

1

14- a m ~ , I

-

-.0-

Ryegrass

0 0

5

10

15

20

25

30

Time (days) Fig. 5. Proportions of 15N-labelled NH4 immobilized in the two treatments. (Vertical bars indicate the standard deviation of the three replicates.)

242

80

60 z

¥

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

I

40

20

0 0

5

10

15

20

25

30

Time (days) [ =

Bare soil

:

Ryegrass ...... Bare soil adjust*

Ryegrass adjust* j

Fig. 6. Proportions of 15N-labelled NH4 not recovered, corresponding to gaseous losses, during the days following pig slurry application. Values were adjusted using the equation: y = a (1- e -k') (Vertical bars indicate the standard deviation of the three replicates.)

slurry spreading generally takes place over a period of 6 days (Jarvis and Pain, 1990; Moal, 1995) or even longer (Grnermont, 1996). The difference in ~SN balance between the two treatments attained a maximum from the 6th day onwards, and remained steady until the end of the experiment. The high rate of volatilization was the result of the climatic conditions, high ammonium content of the slurry, and surface spreading.

4. Discussion

The difference of 14.5% in nitrogen recovery shows that the grass sward significantly reduced volatilization. Similar results were obtained by Whitehead and Raistrick (1992) who used a direct method to quantify ammonia volatilization after spreading livestock urine; they observed a difference of 16% between cumulative volatilization over 8 days, on ryegrass, compared with bare soil. We observed, in agreement with these authors, that the effect of the canopy on volatilization was negligible during the first day, but substantial during the next 2-3 days. This effect could also be due to the rapid growth and great increase in nitrogen content during the 13 days following spreading. The low dry matter content of the slurry might also have reduced retention of the slurry on the canopy, despite the significant deposition of ammonium observed 24 h after the application. An

application of cattle slurry would probably have led to much higher gaseous losses from the canopy (Thompson et al., 1990), because of the higher dry matter content of cattle slurry. We also observed that volatilization halted after day 1 on the grass sward, despite the presence of significant amounts of ammoniacal nitrogen in the surface soil layer, between days 1 and 6. It is therefore almost certain that significant amounts of ammonia were given off at the soil surface in this treatment, the flux being slightly lower than that produced on bare soil, due to the lower temperatures of the soil surface (shade of sward) and slightly lower amounts of ammonium under grassland, on days 1 and 3. The total halting of net volatilization in the ryegrass treatment can only therefore be explained by the plant's absorption of the ammonia given off at the soil surface. It is consistent with the findings of Denmead et al. (1976) that ammonium concentration measured under a grass-clover pasture was greatest near the ground surface and exponentially decreased with height above the ground. The sink effect of the grass sward towards ammonium would in our experimental conditions thus be due both to absorption by the roots and direct assimilation of ammonia by the leaves (Faurie and Bardin, 1979, Jarvis and Pain, 1990). This is also in good agreement with the results of Porter et al. (1972) and Hutchinson et al. (1972) who showed that plant leaves from different species can absorb significant amounts of ammonia from the air. Lockyer and Whitehead (1986) and Whitehead and Lockyer (1987), measuring the uptake of gaseous ammonia by the leaves of Italian ryegrass exposed in chambers to different contents of ammonia in the air, observed that the amount of ammonia absorbed increased linearly with ammonia air concentration. In our experiment, the very low initial inorganic content of the soil probably contributed to the stimulation of ammonia absorption by the canopy, but contents comparable to those that we observed are frequently measured under grassland, outside the periods of fertilizer application; the situation under study was therefore not exceptional for this criterion. We also observed that the rate of immobilization was lower beneath the grass sward (Fig. 5). The driving force of immobilization is the amount of decomposable carbon (Recous et al., 1990). In our case, the carbon that could be assimilated by microorganisms

243

came (i) from the slurry which was supplied in equal quantities in both treatments, (ii) from dying leaves and roots and from root exudates in the case of ryegrass, and dead roots in the bare soil treatment, and (iii) from native organic matter. Mineral nitrogen is often a limiting factor of decomposition, which explains the very high stimulation of immobilization in the days immediately following an application, even without the addition of organic carbon; in this case, however, the measured levels of immobilization remain moderate, in the order of several mg N/kg soil, and are much greater following the addition of a carbon substrate (Recous et al., 1990). In this experiment, several factors could explain the differences between the two treatments: (i) the amounts of ammoniacal nitrogen measured at the soil surface during the first 6 days were not limiting for immobilization; the lower level of immobilization under ryegrass was therefore due to a lower availability of C-substrate, attributable in part to the deposition of organic matter from the slurry on the sward and the presence in bare soil of dead roots that had not yet been decomposed at the time of slurry application; (ii) the halting of immobilization of ~SN labelled nitrogen after the 6th day in the ryegrass treatment (whereas this continued on the bare soil) coincided in contrast with the total disappearance of mineral nitrogen under this treatment, due to plant absorption. Immobilization over 27 days accounted for 25% of the ~SN labelled nitrogen applied to the bare soil, and 16% in the ryegrass treatment. These rates of immobilization are relatively low, given that: (i) high levels of immobilization are expected under grassland (attaining 40-60% of 15Nrecovery, according to Jackson et al. (1989)) and that; (ii) soluble C was provided by the slurry. According to earlier findings (Ledgard et al., 1989; Guiraud et al., 1992), this low rate of immobilization might be due to low persistence of the ammonium pool, in our experimental conditions. Thus, this experiment shows that on a short time scale an actively growing grass sward can significantly modify partitioning of the ammonium pool following the application of a pig slurry. The amounts of 15N-labelled NH4 absorbed by the plants, attaining at the end of the experiment 41% of nitrogen applied for the whole plant, and 29% for the aerial parts, should be compared with the 16.4% of inorganic nitrogen available on the bare soil.

A study of the factors determining the infiltration and deposition of the slurry and the foliar assimilation of gaseous ammonia is required, as these factors govern the amounts of nitrogen volatilized, and thus have a considerable effect on nitrogen efficiency.

Acknowledgements We thank B. Blaise, Y. Fauvel for valuable assistance in the field and laboratory, R. Aubr6e for assistance in the field sampling, and O. Delfosse for 15N analysis.

References Beauchamp, E.G., Kidd, G.E. and Thurtell, G., 1982. Ammonia volatilization from liquid dairy cattle manure in the field. Can. J. Soil Sci., 62: ! 1-19. Denmead, O.T., Freney, J.R. and Simpson, J.R., 1976. A closed ammonia cycle within a plant canopy. Soil Biol. Biochem., 8: 161-164. Drouineau, G. and Gouny, P., 1947. Contribution /~ r6tude du dosage de I'azote nitrique par la m6thode Devarda. Ann. Agron., 17: 154-164. Faurie, G. and Bardin, R., 1979. La volatilisation de l'ammoniac. II. Influence des facteurs climatiques et du couvert v :6g~tal. Ann. Agron., 30:401-414. G6nermont, S., 1996. Mod61isation de la Volatilisation d'Ammoniac apr~s l~pandage de Lisier sur Parcelle Agricole. Thesis, University Paul Sabatier, Toulouse. Guiraud, G., Marol, C. and Fardeau, J.C., 1992. Balance and immobilization of (15NH4)2SO4 in a soil after the addition of Didin as a nitrification inhibitor. Biol. Fert. Soils, 14: 23-29. Hutchinson, G.L., Millington, R.J. and Peters, D.B., 1972. Atmospheric ammonia: absorption by plant leaves. Science, 175:771772. Jackson, L.E., Schimel, J.P. and Firestone, M.K., 1989. Short-term partitioning of ammonium and nitrate between plants and microbes in an annual grassland. Soil Biol. Biochem., 21: 409-415. Jarvis, S.C. and Pain, B.F., 1990. Ammonia volatilisation from agricultural land. Proc. Fert. Soc., 298: 3-35. Lauer, D.A., Bouldin, D.R. and Klausner, S.D., 1976. Ammonia volatilization from dairy manure spread on the soil surface. J. Environ. Qual., 5: 134-141. Ledgard, S.F., Brier, G.J. and Sarathchandra, S.U., 1989. Plant uptake and microbial immobilization of 15N-labelled ammonium applied to grass-clover pasture - Influence of simulated winter temperature and time of application. Soil Biol. Biochem., 21: 667-670. Le Pham, M., Lambert, R. and Laudelout, H., 1984. Estimation de ia valeur fertilisante azot :6e du lisier par simulation num6rique. Agronomie, 4: 63-74.

244

Lockyer, D.R. and Whitehead, D.C., 1986. The uptake of gaseous ammonia by the leaves of Italian ryegrass. J. Exp. Bot., 37: 919927. Moal, J.F., 1995. Volatilisation de l' Azote Ammoniacal des Lisiers apr~.s }~pandage: Quantification et l~tude des Facteurs d'Influence. Cemagref Dicova 229 pp. Morvan, T., Leterme, P. and Mary, B., 1996. Quantification par le marquage isotopique 15N des flux d'azote cons6cutifs ~ un 6pandage d'automne de lisier de porc sur triticale. Agronomie, 16: 541-552. Pain, B.F., Phillips, V.R., Clarkson, C.R. and Klarenbeek, J.V., 1989. Loss of nitrogen through ammonia volatilization during and following the application of pig or cattle slurry to grassland. J. Sci. Food Agric., 47: 1-12. Porter, L.K., Viers, F.G. and Hutchinson, G.L., 1972. Air containing nitrogen-15 ammonia: foliar absorption by corn seedlings. Science, 175: 759-761. Recous, S., Fresneau, C., Faurie, G. and Mary, B., 1988. The fate of labelled ~SN urea and ammonium nitrate applied to a winter wheat crop. Plant Soil, 112: 205-214. Recous, S., Mary, B. and Faurie, G., 1990. Microbial immobiliza-

tion of ammonium and nitrate in cultivated soils. Soil Biol. Biochem., 7: 913-922. SAS, 1988. SAS/STAT~ User's Guide, Release 6.03 Edition. SAS Institute Inc. Cary, NC, 1028 pp. Sommer, S.G. and Olesen, J.E., 199 I. Effects of dry matter content and temperature on ammonia loss from surface-applied cattle slurry. J Environ. Qual. 20: 679-683. Thompson, R.B., Pain, B.F. and Lockyer, D.R., 1990. Ammonia volatilization from cattle slurry following surface application to grassland. I. Influence of mechanical separation, changes in chemical composition during volatilization and the presence of the grass sward. Plant Soil, 125:109-117. Whitehead, D.C. and Lockyer, D.R., 1987. The influence of the concentration of gaseous ammonia on its uptake by the leaves of Italian ryegrass, with and without an adequate supply of nitrogen to the roots. J. Exp. Bot., 38: 818-827. Whitehead, D.C. and Raistrick, N., 1992. Effects of plant material on ammonia volatilization from simulated livestock urine applied to soil. Biol. Fert. Soils, 13: 92-95.

~) 1997 ElsevierScience B.V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van Ittersum and S.C. van de Geijn (Editors)

245

Size and density fractionation of soil organic matter and the physical capacity of soils to protect organic matter Jan Hassink a'*, Andrew P. Whitmore, Jaromir Kubfit b aResearch Institutefor Agrobiology and Soil Fertility (AB-DLO), P.O. Box 129, 9750 AC Haren, The Netherlands bResearch Institute of Crop Production, Drnovskd 507, 16106 Praha 6, Ruzyne, Czech Republic Accepted 16 May 1997

Abstract Soil organic matter (SOM) has important chemical, physical and biological functions in the soil. It is difficult to predict the dynamics of SOM because it is very heterogeneous and because its behaviour is affected by soil texture. In this study we used a new size and density fractionation to isolate SOM fractions that differ in stability and we estimated the amount of SOM that can be preserved in different soils. An investigation was carded out into (1) how fast size and density fractions of soil organic matter respond to changes in C input, (2) whether the capacity of soils to preserve C by its association with clay and silt particles is limited and related to soil texture and (3) whether the long term dynamics of soil C can be described with a simple model that makes the assumption that the net rate of decomposition of soil C does not simply depend on soil texture, but on the degree to which the protective capacity of the soil is already occupied. Light and intermediate fractions of the macroorganic matter (> 150/zm) respond much faster to changes in C input than smaller size fractions. This shows that the light and intermediate macroorganic matter fractions can be used as early indicators of effects of soil management on changes in SOM. There was a close positive relationship between the proportion of particles 6 crops for IAFS and >8 crops for EAFS); characterising the crops in their potential role in the M C R in biological, physical and chemical terms, as is done in Table 3.

(B) Drawing up an M C R based on (1) and simultaneously fulfilling a multi-functional set of demands: • •

•

filling the first rotation block with crop no. 1.; filling subsequent blocks while preserving biological soil fertility by limiting the share per crop species to _ 1 ha. To obtain a prototype farming system with sufficient agro-ecological identity, the fields as sub-units have to be of a minimum size. 3. Field length/width < 4. Round or square fields contribute optimally to the agro-ecological identity of a farming system. Therefore, a maximum is

303

required based on 4 (IAFS) or 6 (EAFS) rotation blocks, at least (crop rotation in time). 5. Adjacency of subsequent blocks = 0. Harmful semi-soilborne species are to be prevented from following their host crop by a crop rotation without any adjacency of subsequent blocks to ensure

to be set to the length/width ratio of fields, to limit the loss in identity. 4. Crop rotation blocks > 4 (IAFS) or >6 (EAFS). The shorter the crop rotation, the greater the biotic stress on the crops and the need for external inputs to control that stress. Therefore, crop rotation is

PAR

OPI wheat

NDW

onion Relative shortfall of achieved (a) to desired (d) results*

PAB

NAR

PSD

I N R ~

.........

1992 1996

EEP i

f

decreased

PSDN /

increased

f J J .

SRI

.

.

.

.

.

.

.

.

remained n o t yet tested

KAB

/ /

*relative shortfall = (a-d)/d

NS

QPI potato HHW FDI

KAR QPI carrot

Parameters (in order of increasing shortfalls of achieved to desired results)

Desired results

Achieved results

EEP = Exposure Environment to Pesticides INR -, Infrastructure for Nature and Recreation NAR = N Available Reserves NDW = N Drain Water PAR = P Available Reserves QPl • Quality Production Index (wheat) QPI = Quality Production Index (onion) PAB - P Annual Balance PSD • Plant Spedes Diversity PSDN = Plant Spedes Distribution KAB - K Annual Balance QPI = Quality Production Index (potato) KAR = K Available Reserves QPI = Quality Production Index (carrot) FOI - Rower Density Index HHW = Hours Hand Weeding NS - Net Surplus SRI • Soilcompaction RiseIndex

0 (air, water, soil) • 0.05 < 70 kg/ha (0-10Ocm) < 11.2 NO3-N mg/I 20 < Pw-count < 30 • 0.9 (average) • 0.9 (average) 0.B < PAB < 1.2 • 50 species/INR > 20 species/INR-section (100 m) 0.6 < KAB < 0.8 • 0.9 (average) 14 < K-count < 20 • 0.9 (average) • 10 flowers/m/month INR < 500 hours/farm • 0 guilders/ha ?

0 0.052 60 9.91 24.7 0.95 0.89 1.24 42 16 1.21 0.68 25.7 0.49 2.6 1306 ? ?

,

.

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

Main causes of shortfall 1996

MCR, ENM MCR, ENM ENM slow response slow response ENM MCR, ENM ENM MCR, ENM slow response MCR

Methods to be improved in: Ready Accept- Manage- Effectfor use ability ability iveness

ENM ENM

MCR MCR ENM

ENM ENM

MCR ENM

ENM

MCR MCR

Fig. 5. State of the art for EAFS in Flevoland (NL 2) 1992-1995, as an example of Part 6 of a prototype's identity card in the I/EAFS-network (the prototype is all-round if achieved results match the desired results).

304

crops are not just moved to an adjacent field from year to year (crop rotation in space). 6. Share of cereals < 0.5 (IAFS) or 5% of I/EAFS area. To bridge the gap between 2 growing seasons, airborne and semi-soilborne beneficials need an appropriate ecological infrastructure of at least 5% of the farm area. By laying out a prototype, it can be tested. By testing it will appear to what extent the desired results for any parameter have been achieved. If a shortfall appears between achieved and desired results, the prototype should be improved in the parameter in question, by adjusting the major or minor methods involved according to the theoretical prototype. The state of the art in step 4 for EAFS in NL 2 clearly shows which of the 16 parameters still have to be improved before the prototype is 'all round' (Fig. 5). It also proves that our prototyping is effective, considering the clear progress from 1992 to 1996. Improving a prototype is a matter of relating the shortfalls between achieved and desired results to the methods and improving them in a targeted way. Such shortfalls between achieved and desired results may arise from one or more of the following 4 causes: the method(s) in question is not ready for use, or not manageable by the farmer, or not acceptable to the farmer or not effective. In positive terms, step 4 (testing and improving) has been finalised if the prototype in general and the methods in particular fulfil these 4 consecutive criteria. Consequently, improving the prototype implies the following procedure (Outline 6).

Outline 6. Procedure to improve prototypes of I/ EAFS 1. Establishing which parameters have shortfalls between achieved and desired results. 2. Establishing from the theoretical prototype which methods are involved.

3. Establishing which criteria are not yet fulfilled by these methods: • • • •

ready for use; manageable by the farmers; acceptable to the farmers; effective.

4. Establishing targeted improvements to meet the successive criteria. 5. Laying out and retesting. The 4 criteria will have already received much attention before the prototype is laid out for the first time, especially in the case of testing and improving a prototype on pilot farms. Commercial farmers want to be sure a prototype is feasible and all its methods are safe? Nevertheless, on-farm testing will certainly bring to light various shortcomings of individual methods, which should be improved based on the set of 4 criteria. One major reason why a method may not appear ready for use, is unexpected occurrence of factors which interfere to such an extent that the method needs to be revised to include these factors and their effects. As a result, methods will gradually evolve from simple and subjective to comprehensive and objective.

Examples • •

management factors such as choice of crops and varieties, machines, fertilisers, pesticides; agro-ecological factors such as pests, diseases, weeds, and physical and chemical soil status.

Even if ready for use, a method may still not appear to be manageable to the farmers, for several reasons.

Examples • • •

planning or operations too complicated; too laborious to fit into the labour film; too specific to be carried out with the usual machinery.

Even if ready for use and manageable, a method may still not appear to be acceptable to the farmers, for several reasons.

305

Examples • •

too high costs and/or too few benefits, at least in the short term; too little confidence in utility and/or effectiveness.

Even if ready for use, manageable and acceptable, a method may still not appear to be effective to achieve the desired result in a certain parameter. This conclusion may be premature, in case of parameters with a slow response. Apart from this, the major reason why a method indeed may not be effective is that the theoretical prototype is too simple or distorted considering the method and the parameter in question.

Examples the method needs support by another method; the method has only a minor influence, another method should be established as the major method. Because most parameters are under control of more than one method, and many parameters have a slow response, effectiveness is the most difficult and also the most time-consuming of all 4 criteria to establish. Testing and improving a prototype will take at least 4 years for I/EAFS and 6 years for EAFS, corresponding with one run of the prototype as a complete crop rotation on each field, before reliable responses of abiota (soil, groundwater) and biota (crops, flora and fauna) are obtained. The effectiveness of the methods and the overall prototype can only be established on the basis of these reliable responses in terms of the multi-objective parameters. Theoretically, the number of years needed for step 4 would be the sum of the years needed to fulfil the first 3 criteria and the years needed to fulfil the fourth criterion. In practice, however, biota and abiota start developing a response from the first year the prototype is laid out, provided the prototype is well designed and will not change dramatically in subsequent years. As a result, the adaptation of abiota and biota will mostly occur simultaneously with the testing and improving by farmers and researchers, so step 4 could be done in a minimum of 4-6 years. However, this does not imply that all parameters will have

achieved a steady state. For example, it may take decades before possible excessive reserves of soil P have been diminished or depleted organic matter reserves have been replenished to desired ranges. Nevertheless, if the shortfalls between achieved and desired results are incontrovertably decreasing from year to year, you may speak about reliable responses proving the effectiveness of the prototype. As a result, the final step of dissemination can be envisaged with confidence.

2.5. Dissemination (step 5) If the first 4 steps of prototyping have been made on a single experimental farm, the prototype cannot just be handed over to the extension service for wide-scale dissemination! It is because such a prototype does not cover region-specific ranges in soil, climate and management conditions, which are crucial for its manageability, acceptability and effectiveness. Therefore, prototyping on an experimental farm always needs a follow-up with pilot farms to elaborate a range of variants of the prototype. Consequently, prototyping in interaction with pilot farms saves a lot of time and money and is always preferable. In addition, a group of capable and motivated farmers provides an indispensable technological and social base for an innovation project, which should include dissemination throughout the region. For this purpose we have developed a model of interactive prototyping with pilot farms (Fig. 6). Since it has appeared to work quite satisfactorily in our EAFS project, we have proposed it as a standard to the teams in the I/EAFSnetwork. In the case of interactive prototyping with 10-15 pilot farms, step 4 can be finalised with 10-15 variants of the prototype covering the regional ranges of soil, climate and management. This provides for an excellent base for wider dissemination throughout the region. The initial group of pilot farms can act as demonstration farms and the farmers can be involved in training and guiding farmers willing to convert. To disseminate the prototype in wider circles, regional extension services should be trained to participate and gradually take over the innovation project. The interaction model (Fig. 6) can be maintained to convert groups of farms in a programme of at least 4 years.

306

PILOT FARMS

various farms situations

RESEARCH TEAM

)

(~

theoretical prototype

Fig. 6. Interactive prototyping: designing, testing and improving a prototype by interaction of pilot farms and research team. 3. Discussion

All over the world, agriculture is still being intensified, causing destabilisation of agro-ecosystems and environmental pollution. In developing countries, it is understandable for various reasons, especially in those countries where food production can hardly keep pace with population increase. In industrialised countries, it is absurd when one considers the growing surpluses of agricultural products, the decreasing income and employment in most rural areas and the growing concern of the consumers about the quality of their food. Fortunately, there is also a growing awareness that these immense problems cannot be solved one by one on an ad-hoc basis, but that a more comprehensive and sustainable approach of agriculture is needed. As a result, several new approaches have been proposed, such as sustainable (Allen and van Dusen, 1988; Edwards et al., 1990), integrated (Vereijken and Royle, 1989) and alternative agriculture (National

Research Council of USA, 1989). However, their use is limited because they are hardly defined in measurable terms, elaborated into concrete farming systems and tested for feasibility. Therefore, the current methodical prototyping has been developed to enable agronomists to design, test and improve more sustainable farming systems in interaction with pilot farms. The methodical prototyping of I/EAFS presented here has its roots in two global organisations. The concept of IAFS is based on the work of the crop protectionists, assembled in the International Organisation for Biological and Integrated Control (IOBC). Initially, most working groups of IOBC aimed at the control of single pest species. However, this brought about various problems, such as insufficient cost effectiveness and the emergence of new pests. Therefore, they developed a wider scope and aimed at integrated protection of crops. During the last decade, the scope has further been widened to IAFS, comprising

307

the entire crop rotation and also considering soil cultivation and fertilisation (Anonymous, 1977; Vereijken et al., 1986; El Titi et al., 1993). The concept of EAFS has been developed by the teams in the I~AFS network, searching for a more consistent and sustainable elaboration of IAFS. They have been inspired by the concept of organic farming, as defined by the standards and guidelines of the International Federation of Organic Agriculture Movements (IFOAM) (Geier, 1991). The great advantage of the organic concept is that it offers a market model of shared responsibility by producers and consumers for a sustainable and multifunctional management of the rural areas as agro-ecosystems. It is expressed by the principle of premium prices for the added ecological value of the products under label. This provides for the necessary economic base for the consistent and sustainable elaboration of IAFS, to be called EAFS. However, EAFS should go further than is demanded by the IFOAM guidelines for organic farming in sustainable and multifunctional management of the environment, nature/landscape and health/well-being of people and farm animals. The methodical prototyping of I/EAFS presented here starts at the stage where most fanning systems research stops, and that is the stage of analysis and diagnosis (Gibbon, 1994). However, the I~AFS teams of the EU network, strong in agronomy and ecology, may improve their start by building on a more comprehensive and profound rural and farming systems analysis from research teams, strong in sociology and economy such as those led by Van der Ploeg (1995) and Sevilla Guzman and ISEC Team (1994). The methodical prototyping of I/ EAFS presented here is still provisionally elaborated considering the interaction with pilot farmers in general and the last step (5) of dissemination in particular. In this respect, the I/EAFS teams could also benefit from the expertise developed by teams, such as those led by R61ing (1994). With this enlargement and reinforcement of their capacity, the teams of the I/EAFS network have excellent chances to succeed where up to now most farming systems researchers failed (Gibbon, 1994). Their work comes further than the step of diagnosis and analysis, and includes the subsequent steps of design, layout for testing and improving, and eventually dissemination. Initial results are encouraging. Most teams are pro-

gressively achieving the desired results, although the effectiveness of prototyping can still be improved in various ways (Vereijken, 1996). Nevertheless, the clear progress we achieved in our EAFS prototype for Flevoland (Fig. 5) may be considered as an example of the effectiveness of the prototyping in the I/ EAFS network. The entire methodical approach to prototyping F EAFS will be available in a manual at the end of the current EU concerted action. References Allen, P. and van Dusen, D. (Editors), 1988. Global perspectives on agroecology and sustainable agricultural systems. Proceedings of the Sixth International Conference of International Federation Organic Agriculture Movements. Agroecology Program, University of California, Santa Cruz, 721 pp. Anonymous, 1977. An approach towards integrated agricultural production through integrated plant protection. IOBC/WPRS Bulletin no. 4, 163 pp. Edwards, C.A., Lal, R., Madden, P., Miller, R.H. and House, G. (Editors), 1990. Sustainable Agricultural Systems. Soil and Water Conservation Society, Iowa, 696 pp. El Titi, A., Boiler, E.F. and Gendrier, J.P., 1993. Integrated production, principles and technical guidelines. Publication of the Commission: IP-Guidelines and Endorsement. IOBC/WPRS Bulletin no. 16, 96 pp. ISBN 92-9067-048-0. Geier, B. (Editor), 1991. IFOAM Basic Standards of Organic Agriculture and Food Processing, 20 pp. Oecozentrum Imsbach, D66696 Tholey-Theley, Germany. Gibbon, D., 1994. Farming systems research/extension: background concepts, experience and networking. In: J.B. Dent and M.J. McGregor (Editors), Rural and Farming Systems Analysis. European Perspectives. Proceedings of the First European Convention on Farming Systems Research and Extension, Edinburgh 1993, pp. 3-19. AB International. ISBN 0851989144. National Research Council of USA, 1989. Alternative agriculture. Report of the Committee on the Role of Alternative Farming Methods in Modern Production Agriculture. NRC, Washington, DC, 448 pp. van der Ploeg, J.D., 1995. From structural development to structural involution: impact of new development in Dutch agriculture. In: J.D. van der Ploeg and G. van Dijk (Editors), Beyond Modernization, the Impact of Endogenous Rural Development. Van Gorcum, Assen, The Netherlands, pp. 109-147. R61ing, N., 1994. Interaction between extension services and farmer decision making: new issues and sustainable farming. In: J.B. Dent and M.J. McGregor (Editors), Rural and Farming Systems Analysis. European perspectives. Proceedings of the first European Convention on Farming Systems Research and Extension, Edinburgh, 1993. AB International, pp. 280-291 (ISBN 0851989144). Sevilla Guzman, E. and ISEC Team, 1994. The role of farming

308 systems research/extension in guiding low input systems towards sustainability: an agro-ecological approach for Andalusia. In: J.B. Dent and M.J. McGregor (Editors), Rural and Farming Systems Analysis. European Perspectives. Proceedings of the First European Convention on Farming Systems Research and Extension, Edinburgh, 1993, AB International, pp. 305-319 (ISBN 0851989144). Vereijken, P. and Royle, D.J. (Editors), 1989. Current Status of Integrated Arable Farming Systems Research in Western Europe. IOBC/WPRS Bulletin 1989/XII/5, Wageningen, 76 pp. Vereijken, P., 1992. A methodic way to more sustainable farming systems, Neth. J. Agric. Sci., 40: 209-223. Vereijken, P., 1994. Designing prototypes. Progress Report 1 of the Research Network on Integrated and Ecological Arable Farming

Systems for EU and Associated Countries. AB-DLO, Wageningen, 90 pp. Vereijken, P., 1995. Designing and testing prototypes. Progress Report 2 of the Research Network on Integrated and Ecological Arable Farming Systems for EU and Associated Countries, ABDLO, Wageningen, 90 pp. Vereijken, P., 1996. Testing and improving prototypes. Progress Report 3 of the Research Network on Integrated and Ecological Arable Farming Systems for EU and Associated Countries. ABDLO, Wageningen, 69 pp. Vereijken, P., C.A. Edwards, A El Titi, A. Fougeroux and M. Way, 1986. Report of the Study Group: Management of Farming Systems for Integrated Control. IOBC/WPRS Bulletin no. 9 (ISBN 92-9057-001-0).

© 1997 ElsevierScience B.V. ,411rights reserved Perspectives for ,4gronomy - ,4dopting Ecological Principles and Managing Resource Use M.K. van Ittersum and S.C. van de Geijn (Editors)

309

The Log irden project" development of an ecological and an integrated arable farming system C.A. Helander * Rural Economy and Agricultural Society of Skaraborg, P.O. Box 124, S-532 22 Skara, Sweden Abstract

On the research-farm Loghrden in south-west Sweden a farming system project is carried out according to the methodology for farming systems research elaborated by a European research network in an EU Concerted action. The main emphasis is on development of an Ecological Arable Farming System (EAFS) and of an Integrated Arable Farming System OAFS). The design of the project does not give priority to comparisons between the different systems, but a Conventional Arable Farming System (CAFS) is also used in the project as a reference. The aim of the project is to achieve a long-term persistent, sustainable and productive food supply in combination with a minimum of negative impacts on the abiotic environment. In the Ecological Arable Farming System (EAFS) there were lower yield levels than expected, probably mainly due to nitrogen deficiency during especially the early part of the growing season. The economic result in EAFS, as net surplus, was dramatically improved in 1995 compared to previous years. This is due to the new situation when Sweden joined the EU, with high arable area payments, together with a large demand for ecological products on the market that gave higher prices of these products. In the Integrated Arable Farming System OAFS) the total use of pesticides was reduced by 70% during 1995, but the use of herbicides was above the desired level of 50% compared with the Conventional Arable Farming System (CAFS). The average yields in IAFS were similar to those in the conventional system. Despite this, the economic result for IAFS was lower, the reason being higher machinery costs, extra costs for seeds for undersowing (white clover), and more expensive weed control in the undersown crops. Also the yields of triticale (not grown in the other systems) were low due to winter damage. Keywords: Arable farming; Ecological; Integrated farming systems; Sustainable agriculture

1. Introduction

The predominant goal for agriculture has been that of high production but is now gradually changing for other objectives. Overproduction in combination with an increased awareness of negative impacts of conventional agriculture on environment, nature and the landscape, in addition to a questioning of the long-term sustainability of the present production systems, raised the need for new production methods (Ebbersten, 1990). For instance, the use of fertilisers and pesticides has increased substantially * Tel: +46 511 13160; Fax: +46 511 186 31; E-mail: [email protected]

during the last 30 years in most countries, to a great extent without corresponding increases in yields. This means that the utilisation efficiency has decreased simultaneously with increasing inputs. Such aspects partly require a new research methodology as traditional research in agriculture has generally been concentrated on one or two production factors at a time, without consideration of system effects. The aim of the project described in this paper is to consider at the farming system as a whole, the main objective being to achieve a long-term persistent, sustainable and productive food supply with a minimum of negative impacts on the abiotic environment. Additional objectives are the minimum input of external energy by means of maximum use of

310

bioenergy produced on the farm (fuel from rapeseed) and to minimise the use of other external inputs, such as nitrogen fertilisers. New objectives, such as quality of the abiotic environment, landscape and nature values, agronomic sustainability and animal welfare, are integrated with traditional agricultural goals. For this, a research-farm, Logfirden, in the southwest of Sweden was chosen. The farming system project at Logfirden is connected to a working group within IOBC (International Organization of Biological and Integrated Control of Noxious Animals and Plants) (El Titi et al., 1993) and also to a European research network in an EU Concerted action (Vereijken, 1994; 1995; 1997). This European network includes the project leaders from most of the ongoing farming system research projects in western Europe, such as the Nagele-project in the Netherlands (Vereijken, 1992), the former Lautenbach-project in Germany (El Titi, 1990) and the LIFE project in England (Jordan and Hutcheon, 1994).

Experimental F a r m ( S-1 )

Log~den

[-]

EAFS (22.0ha)

D

IAFS (2S.0ha)

I

CAFS ( 12.0 ha)

I-VIIICrop rotation blocks a-b

Rotations with 75-50% cereals

Ecological infi'astruetum .)

N

It

(n b)


A large-scale farming system research project started in 1991 at the Logfirden research farm, Grfistorp, (58*20' N, 12"38' E) Sweden. The total area for the experiment covers 60 ha of arable land, the size of each field being between 2.5 and 4.0 ha, see map of the farm, Fig. 1. The soil on the farm is a fertile, very heavy clay soil (40-50% clay), with an organic matter content between 2 and 3%. The soil structure is rather poor, due to compaction with heavy machinery and a spring-cereal dominated crop rotation.

~-- 300m

Fig. 1. Map of Logfirden showing the design of the farming system research project.

The weed seed bank is rather small, with Matricaria inodora and Stellaria media as dominating species. The average annual rainfall is about 600 mm. The main emphasis is on development of an Ecological Arable Farming System (EAFS) and an Integrated Arable Farming System (IAFS). The design of the project does not give priority to comparisons

Table 1 The Multifunctional Crop Rotations (MCR) at Logfirden. The rotation in the ecological system was changed from 1996, but this table shows the rotation used in '91-'95 Year

Conventional

Ecological

Integrateda)

Integrated b)

1 2 3 4 5 6 7 8

peas w-wheat oats w-wheat s-rape w-wheat oats w-wheat

peas w-wheat fieldbeans oat vetch w-wheat set-aside rye

peas w-wheat (undersown) oats w-wheat s-rape w-wheat (undersown) oats triticale

peas w-wheat (undersown) set-aside (grass/lucerne) set-aside (grass/lucerne) w-rape w-wheat (undersown) oats triticale

311

Eli

NS, EE QPI~ ~ (IEP 8~, ......

SCI SR SSC

Q

farmingmethods 0.arckr6fdm~m) parameters

PAR, PAB

major links

NAR OMAB

Fig. 2. Theoretical EAFS prototype of Log~rden. For explanation of codes, see text.

PI

Ell

NS, EE

QPI ~uc~ 1

I o--L

PAB PAR

SR

ss_.g__c

Q

farmingmethocb .o

parameters majorlinks

OMAB NAR

Fig. 3. Theoretical IAFS prototype of Log~rden. For explanation of codes, see text.

312 Table 2 Average yields during 1993-95 for crops grown in two or all three of the farming systems Crop

W-wheat Oats S-rape

Integrated

Conventional

Ecological

kg ha -1

kg ha -1

E/C

kg ha -1

I/C

6940 6450 2070

42001) 36801)

0.612) 0.582 )

7080 6350 2350

1.02 0.98 1.15

l) Only results from 1993-94. 2) Average for conventional 1993-94, compared to the same years for ecological.

between the different systems. The Ecological Arable Farming System (EAFS) implies that no chemical fertilisers or agrochemicals are being used. The yields are expected to be lower, 70-80% of conventional farming, and have to be compensated through higher prices. The Integrated Arable Farming System (IAFS) is a system that emphasises reduction of inputs. The yields are expected to be slightly lower than in a conventional system, but the economic result for the farmer is not expected to be reduced. In both these systems the methods described below are being used to develop a long-term persistent, sustainable and productive farming system. In IAFS the yearly planning and the decisions in field are made by the project leader together with the farm manager. The Conventional Arable Farming System (CAFS) is used as a reference and reflects a common type of farming in the region. In CAFS the yearly planning and the decisions in field are made by a local agricultural advisor together with the farm manager. The project at Log~trden follows the methodology for farming systems research elaborated by the European research network on integrated and ecological arable farming systems (Nilsson, 1994; Vereijken, 1994; 1995; 1997). The methods used follow a European shortlist (Vereijken, 1994). They are used in the following order: 1.

2.

Multifunctional Crop Rotation (MCR): the major method to preserve soil fertility in biological, physical and chemical terms and to sustain quality production with a minimum of inputs (pesticides, fertilisers, support energy and labour). Integrated/Ecological Nutrient Management (INM/ENM): supports MCR by maintaining agronomically desired and ecologically accept-

3.

4.

5.

6.

able nutrient reserves in the soil and contributes, together with MSC (see below), to maintain an appropriate content of organic matter. Minimum Soil Cultivation (MSC), only in IAFS: supports MCR by incorporating crop residues, controlling weeds and restoring physical soil fertility, while maintaining sufficient soil cover as a basis for avoiding nutrient losses, shelter for natural enemies and for landscape/nature values. Ecological Infrastructure Management (EIM): supports MCR by providing airborne and semi-soilborne beneficials a place to overwinter and recover and disperse in spring. In addition, EIM should achieve different nature and landscape objectives. Integrated Crop Protection (ICP)~ only in IAFS" supports MCR and EIM by selectively controlling remaining harmful species with minimal exposure to the environment of pesticides. Farm Structure Optimisation (FSO): the method to make a farming system economically optimal by determining the minimum amounts of land, labour and capital needed.

The standardised design of these methods is described by Vereijken (1995; 1997), where also the methods and parameters used at the Log~rden project are described. The three different parts of the farm, Ecological (EAFS), Integrated (IAFS) and Conventional (CAFS), have different crop rotations (see Table l) using the Multifunctional Crop Rotation (MCR) concept (Vereijken, 1994). The IAFS has been divided into two different rotations, results presented in this paper being from rotation b. The integrated system is based on non-ploughing tillage practices. In

313

Table 3 Soil Cover Index (SCI). SCI is the extent to which the fields on a farm is covered by crops overall the year. Desired result: SCI > 0.8 Year

Conventional (CAFS)

Ecological (EAFS)

Integrated (IAFS)

1993 1994 1995

0.67 0.54 0.79

0.69 0.61 0.76

0.89 0.80 0.81

the ecological and conventional system ploughing is carried out almost every year. The systems are evaluated using 11 (EAFS) or 12 (IAFS) multiobjective parameters from the European shortlist (Vereijken, 1994). Furthermore, some local parameters are operated. All agronomic data recorded are processed by means of a yearly analysis program to quantify the parameters that are used in the evaluation. The different parameters are linked to one or more of the methods used for development of the Ecological Arable Farming System (EAFS) and the Integrated Arable Farming System (IAFS). Figs. 2 and 3 presents the theoretical prototypes for EAFS and IAFS showing the major and minor methods to be followed to achieve the desired result for each parameter. Results from the following 6 multi-objective parameters are presented in this paper" 1. Soil Cover Index (SCI) 2. Nitrogen-Available Reserves (NAR) 3. Nitrogen Utilisation (local) 4. Energy Efficiency (EE) 5. Pesticide Index (PI), (only in IAFS) 6. Net Surplus (NS)

Desired result: 80% soil cover Desired result: < 30 kg N ha -1 Desired result: > 70% efficiency

parameters, but the results from these analyses are still under evaluation.

3. Results

The results are presented below in terms of parameters and related methods. The project is still, after five years, in a development stage and therefore the results only demonstrate the progress so far. It is not possible to make a direct comparison between the different systems due to the layout of the experiment. The average yields of crops grown in all three systems (winterwheat and oats) and for springrape grown in CAFS and IAFS are presented in Table 2. The average yields of "ecological" crops and of "integrated" crops are compared with the yield of the "conventional" crops as relative numbers, E/C and I/C. These numbers can only be seen as an indication of the level of difference. After a few more years (at least a full crop rotation of eight years) the systems should have reached some kind of equilibrium and then the average yields from the different systems should become more reliable.

Desired result: > 6 (output/input) Desired result: < 50% of CAFS Desired result: > 0 SEK

Five other parameters are also used, but results are not presented in this paper. These parameters are: Phosphor-Available Reserves (PAR), Phosphor-Annual Balance (PAn), Organic Matter-Annual Balance (OMAB), Quality Production Index (QPI) and Environmental Exposure to Pesticides (EEP). Two more parameters that are planned to be used are Soil Respiration (SR) and Soil Structure and Compaction (SSC). An extensive analytical programme is carried out every year as a basis for these

3.1. Soil Cover Index (SCI) The desired result for this parameter is to have crops covering the ground for more than 80% of the year. A high SCI is very important for reducing the risks of nutrient leaching and soil compaction. It is also important for building up the organic matter and for a high biological activity in the topsoil. The IAFS achieved the target each year but not the EAFS nor the CAFS although the EAFS had marginally higher values than the CAFS (Table 3).

3.2. Nitrogen-Available Reserves (NAR) NAR is the major parameter used to evaluate the

314

Table 4 Nitrogen-Available Reserves (NAR) in kg N ha -1 (0-90 cm), after the indicated crop, at start of the leaching period. Desired level: 30 kg N ha -1 Year

Conventional

Ecological

NminSoil

Crop

NminSoil

Crop

NminSoil

Crop

1992 1993 1994 1995

40 58 64 35

w-wheat s-rape oats w-wheat

76 51 42 41

green manure rye field beans w-wheat

56 31 43 37

s-rape w-wheat oats triticale

Integrated and Ecological Nutrient Management (I/ ENM). The goal is to achieve a NminSoil level below 30 kg N ha -1 (in the 0-90 cm soil layer) at the start of the leaching period (Nov-Dec). This level is probably difficult, but possible to reach. So far the values measured in all three systems have been above the desired level, see Table 4.

Integrated

utilisation of nitrogen was achieved in the ecological system with no chemical fertilisers, followed by the conventional system. A nitrogen balance sheet for 1995 is also presented in Table 5, where an average for all inputs and outputs in each system is calculated.

3.4. Energy Efficiency (EE) 3.3. Nitrogen utilisation A local parameter used in 1995 to evaluate the Integrated/Ecological Nutrient Management (I/ ENM) is a Nitrogen Utilisation calculation. The nitrogen uptake in harvested plant products was divided by all inputs of nitrogen, i.e. fertilisers, manure and nitrogen from nitrogen-fixing organisms. The results are shown in Table 5. The best

EE is a parameter which is planned to be used for the evaluation of the energy efficiency in the farming system. The method of calculating energy efficiency is still under discussion. Calculations presented in Table 6 are made by means of a computer model where the total energy content in the harvested plant products (in kWh/ha) is divided by the total input of energy (in kWh/ha) into the cropping system. The

Table 5 Nitrogen balance sheet and Nitrogen Utilisation in 1995. Nitrogen balances (kg N/ha) are calculated for inputs and outputs as an average for all harvested crops in each system. Nitrogen Utilisation is calculated as nitrogen in harvested plant products divided by all inputs of nitrogen Nitrogen balance sheet Inputs: N-fertilisers Seeds N-fixating crops 1) Deposition from air Manure 2) Total input Outputs: Crops Straw (off field)

Total output Nitrogen Utilisation

Conventional

Ecological

Integrated

132 4 14 12

5 9 14 15

76 3 14 -

162

43

93

108 14 122 75%

33 1 34 79%

50 6 56 60%

1) In 1995 the pea-crop was completely damaged due to very wet conditions in May and June, no pea-crop was harvested. 2) The conventional system were given 40 ton/ha of slurry in 3 of the 4 fields. This is more than a normal supply for a single year.

315 Table 6 Energy Efficiency (EE) for 1995. EE is calculated by dividing the energy in harvested plant products by the total energy input into the cropping system. Desired level: EE > 6.0

Average for all harvested crops Best crop

Conventional

Ecological

Integrated

4.7 6.0 (w-wheat)

1.9 5.6 (s-wheat)

3.6 7.4 (w-wheat)

total input includes energy used to produce machinery, diesel, electricity, fuel oil, nitrogen, pesticides and seed all calculated in kWh/ha. The calculations in Table 6 are fully based on the actual use on Loghrden experimental farm in 1995. The target value is EE > 6 (output of energy divided by input of energy) as average for all harvested crops. This level was not reached in any of the systems in 1995.

3.5. Ecological Infrastructure Index (Eli) An Ecological Infrastructure Management (EIM) is very important for creating improved biodiversity which provides useful support for a sustainable farming system. Eli is the main parameter for measuring the level of ecological infrastructure. At Log~rden, 6% of the arable land in both the ecological and the integrated system has been used for ecological infrastructures such as hedges, green belts and pathways. In the conventional system no part of the arable land is used for ecological infrastructure.

3.6. Pesticide Index ( PI) A Pesticide Index (PI) is used as a parameter to evaluate the need for pesticide treatments in the In-

tegrated Arable Farming System (IAFS) compared to the Conventional Arable Farming System (CAFS). No pesticides are used in the ecological system and therefore this index is not used in EAFS. The calculations presented in Table 7 are based on the use of pesticides in the integrated system compared to the use in the conventional system. The actually used dosage and number of treatments for each type of pesticide (insecticides, fungicides and herbicides) is recalculated to an average for the whole crop rotation in each system (for instance: one fungicide treatment with half the recommended dose on all fields will give the PI=0.5 for fungicides). The total index calculated for Log~rden is well below the desired level of 50% for the use of insecticides and fungicides. However, the use of herbicides is above the desired level.

3. 7. Net Surplus (NS) NS is a parameter for evaluation of the economic efficiency of the whole farming system. Table 8 shows the NS for all three systems during 1993-95. NS is calculated for each year's conditions. It is important to notice that Sweden joined EU in 1995. This resulted in completely different conditions, with for instance, an arable payment and an extra payment for ecological farming.

Table 7 Pesticide Index (PI) for 1995. PI is based on the actually used dosage and number of treatments for each type of pesticide in each system compared to recommended dose for each pesticide. The use in the integrated system is divided by the use in the conventional system. Desired level for Pl IAFS/CAFS: < 0.5 Type of pesticide

Pl for CAFS

PI for IAFS

PI for IAFS/CAFS

Insecticides Fungicides Herbicides Mean of all pesticides

1.18 0.82 0.82 0.94

0.09 0.11 0.60 0.26

0.08 0.14 0.72 0.30

316

Table 8 Net Surplus (NS) in SEK ha -1. NS is gross revenues minus all costs, including payments for all labour hours and for the land used. Desired result: NS > 0 Year

Conventional

Ecological

Integrated

1993 1994 19951)

130 -280 880

-2660 -2840 980

- 160 -2140 -40

1) EU/CAP

chanical weed control was quite successful. The number of weeds was low and did not cause any major problems. The economic result in EAFS, as net surplus, was dramatically improved in 1995 due to the new situation when Sweden joined the EU with high arable area payments. A large demand for ecological products on the market gave high prices of these products.

4.2. Nitrogen management 4. Discussion

The project is still, after five years, in a development stage and therefore the results only demonstrate the progress so far. It is very important to realise that some of the positive effects, for instance a higher nitrogen mineralisation capacity, which are expected from a change into an ecological or integrated farming system, are still to come.

4.1. Ecological Arable Farming System (EAFS) In the Ecological Arable Farming System the yields were lower than expected, probably caused by nitrogen deficiency during especially the early part of the growing season. The low level of plant-available nitrogen was, at least partly, related to a very compact soil and thereby low microbial activity. Compact soil may also have caused losses of nitrogen due to denitrification under wet conditions (as during June of 1995). The results of the Energy Efficiency (EE) calculations very clearly show the importance of crop yield levels. The yields in 1995 were very poor for many crops, especially in the ecological system but also in the integrated. Table 6 shows that the desired level of EE >6.0 can be achieved for some individual crops, but that the target level is not reached for the average of the whole crop rotation. The result for Soil Cover Index (SCI) in the EAFS has not reached the desired level (Table 3). This is one of the reasons for making a change in this rotation: vetch and winterwheat have been substituted by set-aside (an undersown green manure crop) and winterrape. Weeds, insects and diseases constituted a smaller problem than expected. The use of me-

A low level of available nitrogen before the winter period will reduce the risk for losses during this period when the crops take up no or very little nitrogen. So far, the NAR values measured in all three systems have been above the desired level. The best possibilities to reduce NAR include a higher precision in the application of nitrogen and more restricted soil tillage after harvest, and especially not to stimulate nitrogen mineralisation in the autumn by means of early soil tillage. Another factor of importance, to be better used, is to bind nitrogen organically during winter, by means of winter crops.

4.3. Integrated Arable Farming System (IAFS) The general use of external inputs (pesticides and chemical fertilisers) in conventional farming in Sweden is very low compared to that in many other western European countries. This means that there is not much scope for reduction of, for instance, pesticide use. In spite of that, the total use of pesticides was reduced by 70% during 1995 in the Integrated Arable Farming System (Table 7). The use of herbicides is still above the desired level of 50% compared to the conventional farming system. To improve this, more mechanical weed control will be carried out. Mechanical weed control proved to be quite successful in the ecological system since the number of weeds was low and did not cause any major problems. The reduction in the use of nitrogen fertilisers and in the use of fuel for soil preparation also has been substantial. The average yields in IAFS were similar to those in the conventional system. Despite this, the economic results for IAFS were lower, the reason being higher machinery costs, extra costs for seeds for undersowing (white clover), and

317

more expensive weed control in the undersown crops. Also the yields of triticale (not grown in the other systems) were low due to winter damage. To improve the economic result for IAFS we are looking into better alternatives for weed control in the undersown crops and also more winter-hardy varieties of triticale. The full effect from the undersown white clover being a higher nitrogen mineralisation capacity is hopefully still to come. The economic results for IAFS in other comparable projects such as the Nagele project and the LIFE project show positive results compared to conventional systems. The fact that we in general have a low use of external inputs in conventional farming in Sweden is perhaps one reason for not getting that positive results in the Log~rden project, when we compare the integrated with the conventional system.

Acknowledgments The arable farming system research project at Logarden has been financed by the Swedish Board of Agriculture and the Agricultural Society of Skaraborg.

References Ebbersten, S., 1990. Lantbruksvetenskap - en omvirldsanalys inf6r 2000-talet med sarskild h/insyn till agronom-, hortonom- och landskapsarkitektutbildningarna. SLU/F6rvaltning nr 16, Sveriges Lantbruksuniversitet, Uppsala. El Tiff, A., 1990. Farming System Research at Lautenbach, Germany. Schweiz. Landwirtsch. Forsch., 29(4): 23%247. El Tiff, A., Boiler, E.F. and Gendrier, J.P., 1993. Integrated Production, Principles and Technical Guidelines. IOBC/WPRS Bulletin OILB/SROP, Vol. 16 (1), 97 p. Jordan, V.W.L. and Hutcheon, J.A., 1994. Economic viability of less-intensive farming systems designed to meet current and future policy requirements: 5 year summary of the LIFE project. Asp. Appl. Biol., 40: 61-68. Nilsson, C., 1994. Integrated farming systems research at Alnarp. Proceedings NJF symposium 'Integrated systems in agriculture', 1-3 December 1993 in Norway: pp. 65-70. Vereijken, P., 1992. A methodic way to more sustainable farming systems. Neth. J. Agric. Sci., 40: 209-223. Vereijken, P., 1994. 1. Designing Prototypes, Progress Reports of Research Network on Integrated and Ecological Arable Farming Systems for EU and associated countries. AB-DLO, Wageningen, 87 p. Vereijken, P., 1995. 2. Designing and Testing Prototypes, Progress Reports of Research Network on Integrated and Ecological Arable Farming Systems for EU and associated countries. AB-DLO, Wageningen, 90 p. Vereijken, P., 1997. A methodical way of prototyping integrated and ecological arable farming system (I/EAFS) in interaction with pilst farms. In: M.K. van lttersum and S.C. van de Geijn (eds.), Proceedings of the 4th ESA Congress, Elsevier, Amsterdam, the Netherlands, pp.


© 1997 Elsevier Science B. II". All rights reserved Perspectives for Agronomy- Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

319

Integrated crop protection and environment exposure to pesticides: methods to reduce use and impact of pesticides in arable farming F.G. Wijnands* Applied Researchfor Arable Farming and Field Production of Vegetables, P.O. Box 430, NL 8200 AK, Lelystad, Netherlands

Accepted 14 July 1997

Abstract

Prototypes of Integrated Farming Systems for arable farming are being developed in the Netherlands based on a coherent methodology elaborated in an European Union concerted action. The role of crop protection in Integrated systems is, additional to all other methods, to efficiently control the remaining harmful species, with minimal use of well selected pesticides. The overall aim of more sustainable farming systems is to reduce the exposure of the environment to pesticides in order to prevent short- and long term effects on all species over all the biosphere. An innovative approach to quantify this exposure of the environment to pesticides, based on molecular-chemical properties of the pesticides, is presented. The results of prototyping on an experimental farm in the Netherlands shows that not only drastic reductions in pesticide use are possible but that subsequent careful selection of pesticides can also lead to minimal environmental impact. © 1997 Elsevier Science B.V. Keywords: Arable farming; Environment; Integrated crop protection; Integrated farming; Pesticides; Prevention; Farming systems research; Indicators; Pesticide risk evaluation

I. Introduction

The use of pesticides in current arable farming systems is extremely high due to the almost exclusive choice for pesticides to correct structural problems in farm management such as insufficient crop rotation, susceptible varieties and high nitrogen inputs. The high pesticide use is only one symptom, however a major one, of the shortcomings of current farming in the European Union. Current farming is associated with a complex of environmental, agronomic and ecological problems. In reaction to these problems, Integrated Farming *Tel.: +31 320 29Ill1; fax: +31 320230479.

Systems have been developed as a coherent new vision on agriculture alongside other concepts such as ecological farming. Over the last 15 years these systems that integrate potentially conflicting objectives concerning economy, environment and agronomy are being developed on experimental farms all over Europe (Vereijken and Royle, 1989; Vereijken, 1994); in the last 5 years increasingly also in cooperation with commercial farms (Vereijken 1995, 1997). The methodology of designing, testing, improving and disseminating Integrated and Ecological Farming Systems for arable farming is elaborated in a 4 year European Union Concerted Action involving the leading research teams in Europe. This methodology is


320

called prototyping and comprises five steps (Vereijken, 1994, 1995, 1997). After the objectives have been set (1) and transformed into a suitable set of multi-objective parameters (2), appropriate farming methods (comprehensive strategies built on different techniques) that sufficiently integrate the potentially conflicting objectives need to be developed or redesigned (3). Top priority is given to the design of a multifunctional crop rotation. Then nutrient management strategies need to be designed, followed by the design of soil cultivation strategies and the lay-out of an ecological infrastructure on the farm. All these methods are aimed at sustaining quality production with minimum external inputs and environmental hazards. In a theoretical prototype parameters and methods are linked as last check (4) before the testing in practice may start (5). Testing and improving the prototype in general and the method in particular continues until the objectives as quantified in the set of parameters have been achieved. This can either be done on experimental farms or on pilot farms. Dissemination of the results including implementation in practice concludes this approach. This paper is based on prototyping research on the Nagele experimental farm (Wijnands and Vereijken, 1992) in the Netherlands and elaborates the role of the farming method Integrated Crop Protection (ICP) in Integrated systems. This method is complementary to the methods that consider crop rotation, nutrient management, soil cultivation and ecological infrastructure, that were mentioned before. It will be shown how ICP can reduce the input of pesticides drastically. A new concept of quantifying the environmental burden due to pesticide use will be elaborated. This concept is called Environment Exposure to Pesticides (EEP). Minimising the latter is the basic aim for more sustainable farming systems in order to prevent short- and long term adverse effects on all species over all the biosphere. Results of the Nagele farm will demonstrate the perspective of this concept.

designed for three specific regions in the Netherlands and laid out on experimental farms with region-specific crop rotations and cropping systems (Wijnands and Vereijken, 1992). From 1990 to 1993 the tested prototypes were evaluated on commercial farms in a national pilot farm network (Wijnands, 1992; Wijnands et al., 1995). The Integrated prototype for the Central Clay area will serve here as example. The small farm size (2550 ha) in this region encourages farmers to grow cash crops in short rotations needing heavy inputs. Potato is the most profitable crop, followed by sugar beet and vegetables such as onion and cabbage. Cereals are financially less attractive but are needed as break crops. Most rotations are for only 3 or 4 years. Consequently, beet and potato cyst nematodes (Heterodera spp and Globodera spp) cause serious problems, forcing farmers to fumigate soil regularly as a curative or preventive measure. The Integrated prototype for the Central Clay area has been developed since 1979 on the 'Development of Farming Systems' experimental farm at Nagele (central clay region). The farm size is 72 ha and the soil is heavy sandy marine clay (24% clay). Three farming systems were studied until 1991: Integrated, Conventional (rotations see Table 1) and Ecological. In 1991 the experimental layout was drastically Table 1 Crop rotations of the different systems and periods at the experimental farm at Nagele Integrated and Conventional 1986-1990

Integrated and Experimental 1991

Year

Crop

Year

Crop

l

i

One-half ware, one-half seed potato Sugar beet

3

One-halfware, one-half seed potato One-halfdry pea, one-quarter carrot, one-quarter onion Sugar beet

4

Winter wheat

2

2 3

2. Material and methods 4

2.1. Prototypes of integrated systems in the Netherlands Integrated prototypes for arable farming were

Experimental, advanced integrated.

One-half carrot, one-half onion (experimental: one-half carrot, one-half chicory) Winter wheat (Experimental: one-half winter wheat, one-half sugar barley)

321

revised. Because of the promising results of the Integrated prototype (Wijnands and Vereijken, 1992; economically the Integrated system was competitive with the Conventional reference system) and the subsequent progress in policy (Ministry of Agriculture, Nature Management and Fisheries, 1990, 1991), the Conventional reference system was no longer needed. It was therefore replaced by a demonstration-Integrated prototype meeting the policy aims of 2000. Subsequently a new Integrated prototype (Experimental) was designed, aimed at further reductions in inputs of pesticides and nutrients (rotations; see Table 1). Concerning the farming methods that are used in the Integrated system the following specifications can be given. Concerning the Multifunctional Crop Rotation: a potato cropping frequency of 1:4 is considered as an acceptable compromise between a more sound rotation (1:5 or 1:6) and more profitable short rotations (1:3) with more biotic stress and therefore requiring more inputs. The Integrated Nutrient Management strategy applied is based on the environmental safe and agronomic efficient use of manure as a basic source of nutrients and organic matter and is aimed at minimum losses. For more details on Ecological Infrastructure Management and the economic aspects see Wijnands (1994). The ICP strategy that was followed will be elaborated in Section 2.2.

2.2. Integrated crop protection The role of crop protection in an Integrated system is, additional to all the other methods, to efficiently control the residual harmful species, with minimal use of well selected pesticides. ICP focuses on the real problems, namely the residual ones, after all other methods are designed and optimised. Consequently this means that in the design of the multi-objective methods Multifunctional Crop Rotation and Integrated Nutrient Management, all crop protection aspects are taken into consideration. This concerns for instance the choice of the (inter)crops, their frequency and sequence as well as the spatial aspects of the crop rotation (Vereijken, 1994). Moreover a well balanced Ecological Infrastructure Management should enhance the stability of the system. When designing an Integrated Nutrient Management strategy, the interaction between weeds, pests and dis-

eases and soil fertility, and the plant nutritional status are taken into account. The ultimate objective of the Integrated and Ecological systems with respect to pesticides is the same: zero use and zero negative impact on environment and ecology. However whilst an Ecological system radically abandons pesticide use and consequently produces under label for higher prices on special markets, Integrated systems still use pesticides since production for the world market does not allow to radically abandon pesticides. However also for Integrated systems the target can only be zero use of pesticides. This is a major challenge for agronomy and crop protection science. First of all the use of pesticides in an Integrated approach can be minimised by putting maximum emphasis on prevention (resistant varieties, cultural measures such as adapted sowing date and row spacing). Whenever a disease, pest or weed population occurs a correct interpretation of the need for control (guided control systems, thresholds, signalising systems, etc.) can prevent unjustified use of pesticides. Secondly all available non-chemical control measures (mechanical weed control, physical and biological control) should optimally be integrated in effective and manageable control strategies. Pesticides are only necessary in very specific cases. They always have to be integrated in crop- and location specific control strategies. Application methods are preferred that lead to a minimum use, such as seed treatment and row- or spotwise application. The latter techniques require careful integration of chemical and mechanical techniques when applied in weed control. Appropriate dosages and when possible a curative approach (field- and year specific) further reduce the input. The residual required pesticide use then requires a careful selection of pesticides to avoid disturbance of non-target organisms (selectivity) and to minimise the exposure of the environment to pesticides.

3. Pesticides

3.1. Behaviour and impact Current agriculture depends to a great extent on pesticides. It is estimated that world wide some 2.5

322 million t of pesticides are applied annually in agricultural crops. Pesticides can be described as the only group of toxic chemicals which are intentionally dispersed in the environment (The Pesticides Trust UK, information leaflet). Only a fraction of the pesticides gets in contact with its target organisms (directly or indirectly). Pimentel (1995) estimates that in the case of pests, only 0.4% of the pesticide contacts its target pest. Inevitably a large part of the applied pesticides become part of the abiotic environment. Pesticides may volatilise into the air, runoff or leach into surface- and groundwater, be taken up by plants or soil organisms or remain in the soil, depending on pesticide properties, climatic and crop conditions, soil type and 'infrastructure' (slope of fields, nearness of surface water, hydrology, etc.). The environment thus gets exposed to a certain pesticide load. The combination of pesticide properties and 'environmental' conditions determines the 'persistence' of the compounds (adsorption, degradation, photolysis, etc.). Pesticide behaviour in soil (persistence and leaching to groundwater) has been studied extensively and is relatively well known. The total seasonal losses in runoff rarely exceed 5-10% of the total amount applied (Leonard, 1990). The fraction removed by leaching is probably less than 5-10% (Taylor and Spencer, 1990) however both runoff and leaching have a very significant impact on water quality causing world wide serious concern over the past three decades. Volatilisation is the major cause of pesticide loss. Volatilisation losses up to 80-90%, within a few days after application, have been reported (Taylor and Spencer, 1990). A recent study in the Netherlands (in the framework of the evaluation of the crop protection policy) estimates that some 50% of the total pesticide use volatilises (Multi-Year Crop Protection Plan, 1996). The fate of pesticides in the atmosphere is relatively unknown. However by atmospheric transport and deposition (global distillation) many pesticides may be distributed all over the earth (Gregor and Gummer, 1989; Atlas and Schauffler, 1990; Schomburg and Glotfelty, 1991; Simonich and Hites, 1995). Pesticides unavoidably cause ecological effects, since no pesticide is specifically toxic to only one species. Consequently the presence of pesticides in the abiotic environment is potentially a threat for all involved biota (non-target). The magnitude and dif-

ferentiation of this threat is only very partially known and quantified. Pesticide toxicity for humans and some mammals is relatively well known. Much less is known about the effects on other biota, the so-called ecotoxicity. A proper evaluation of the ecotoxicity of a substance is virtually impossible since it involves thousands of different species that react differently when exposed to a certain substance. It not only involves direct toxicity but also mid- and long term effects on, for instance fertility, vitality and population dynamics. This knowledge calls for a radical strategy. A preventive strategy that aims at minimising any potential effect of pesticides on biota. Therefore the exposure of the environment to pesticides should be minimised. This should be reached by minimising the pesticide requirements of farming systems (e.g. by ICP, see Section 2.2) and consequently careful selection of pesticides taking into account the extent to which the environment gets exposed to pesticides. The use of pesticides is currently often quantified as number of treatments, as kg active ingredients or as a relative number expressing the ratio used dose/recommended full field dose. These parameters only quantify use and cropping technique. In Section 3.2 the quantification of pesticide properties in terms of potential presence in the environment will be elaborated. 3.2. Environment exposure to pesticides

EEP is quantified by taking into account the active ingredient properties and the amount used. EEP-air = active ingredient (kg/ha) x vapour pressure (VP at 20-25°C) (Pa). EEP-soil = active ingredient (kg/ha) x50% degradation time (DT50) (days). EEP-groundwater = EEP-soil (kg days/ha) x mobility of the pesticide (-). Mobility = Kom; g o m -" partitioning coefficient of the pesticide over dry matter and water fraction of the soil/organic matter fraction of the soil. The properties of active ingredients of pesticides, i.e. VP, DT50 and Kom, are known under standardised conditions, since this is required for the approval procedures (Linders et al., 1994). For instance the ratio DT50/Kom in the Netherlands is used in model studies to establish the leaching risk as part of the approval procedures.

323

These rather simple calculations do not take into account any division of the compounds over the three compartments of the abiotic environment nor do they relate to the period of the year and the crop conditions (soil cover) during application. EEP quantifies the maximum risk of environment exposure to pesticides and can be used to evaluate pesticide use or to select pesticides. Of course any additional knowledge of ecological effects should be taken into consideration. EEP can be quantified per pesticide, but also be summarised as EEP per crop (sum of EEP per pesticide) or EEP per farm (weighted average of EEP per crop with respect to area). Comparative surveys of available pesticides provide the basis for rational pesticide choice. Evaluation of pesticide use implies quantification of the EEP-water, -air and -soil per pesticide, per crop and per farm. Pesticides then can be ranked by calculating their relative contribution to the EEP per farm (Table 7). This provides a rational basis for targeted improvement in EEP. EEP targets should be achieved by: (1) substitution of the highest ranking compounds by non-chemical measures or lower ranked pesticides, or (2) reducing the used amount by a more appropriate dose or by bandspray or spotwise treatments.

4. Results The results of crop protection in terms of pesticide use and EEP of the Nagele experimental farm over the period 1986-1990 are presented, including an outlook

Table 2 Average marketable crop yields (t/ha) in the Integrated and Conventional farming systemat Nagele in different periods Crop

Ware potato Seed potato Sugar beet Winter wheat Pea Winter carrot Sown onion

1986-1990

1992-1995

Conventional Integrated

Integrated

54.4 33.9 64.3 7.6 5.0 52.2 40.6

53.1 32.0 59.2 9.7 69.3 45.3

54.6 34.3 60.2 6.7 4.7 52.2 31.0

into the period 1992-1996. The crop rotation in both periods is given in Table 1. Yields of the crops in the Conventional and Integrated systems were in general similar, with exception of winter wheat and sown onion (Table 2). For ware potato (cultivar value), seed potato (cultivar value), sown onion (quality) and sugar beet (sugar content and extractability) higher product prices were achieved in the Integrated system. Prices of produce for the other crops were similar in both systems. The costs of pesticides and fertilisers were lower and the costs of seeds and tubers were higher in the Integrated system. The final gross margin per crop was higher in the Integrated system, still with exception of winter wheat and sown onion. At farm level, costs of machinery and labour were slightly higher in the Integrated system, but these did not fully remove the financial advantage of the higher gross margins. As a result the net surplus of the Integrated system was slightly higher than that of the Conventional system (Wijnands and Vereijken, 1992). More details about the physical and financial results of the various systems have been reported by Bos et al. (1992). The yields of the 1992-1995 period show considerable improvement in yields of wheat (cultivar and N-management), onion (N management) and carrot (cultivar and Nmanagement) in the Integrated system. For all crops, the yield levels are now similar to the average 'conventional' yields of the region (Conventional no longer available at the experimental farm). 4.1. Pe s tic id e use

Table 3 specifies the number of ICP interventions. Compared to the Conventional system the annual input of pesticides in kg active ingredients/ha in the Integrated system was reduced by 65%, excluding nematicides and by 90% if nematicides are included (Table 4). The largest reduction in active ingredient use was realised by substituting dichloropropene (DCP), a soil fumigant used to control potato cyst nematodes in the Conventional system, by non-chemical measures such as the use of appropriate cultivars based on detailed monitoring techniques. Herbicide input in the Integrated system was largely replaced by mechanical control and by band spraying or low dose techniques (Tables 3 and 4).

324 Table 3 Number of interventions for crop protection (n/crop) in the Conventional and Integrated farming system (1986-1990) at Nagele Weeds

Conventional Integrated

Pest/diseases

Mechanical

Thermal

Chemical

Total

Chemical

0.9 2.0

0.2

2.6 1.5

3.5 3.7

4.2 1.9

Per herbicide application, the amount of active ingredient used was 2 5 - 5 0 % less. The labour demand increased because the band spraying and mechanical interventions took more time than full field herbicide spraying. In seed and ware potato crops it was not necessary to use herbicides for weed control. Growing winter wheat at wider inter-row spacing (26 cm) enabled herbicides to be replaced by mechanical control. Fungicide input was reduced by using resistant cultivars, moderate nitrogen supply, control thresholds and decision support systems. The largest reduction at farm level was achieved in potato (Table 4 and Table 5). Fungicide input in onion was largely reduced by supervised control based on monitoring initial infestation by Botrytis squamosa and weather conditions (Table 4). Growth regulators were only used in sown onions to inhibit sprouting during storage. Insecticide input was minimal due to low insect pressure and the use of control thresholds, reduced dose techniques and band spraying. Fig. 1 shows the further decrease in pesticide use in

Total

7.7 5.6

the most recent period (1992-1996). The reduction in fungicide use is mainly based on the substitution of the 'old' compounds used to control Phytophthora infestans by a new low active ingredient compound called fluazinam. The reduction in herbicide use is based on a further increased and optimised use of mechanical techniques and appropriate dosage herbicide systems Compared to the Conventional system of 1 9 8 6 1990 the reduction percentage increased. In absolute terms the level of pesticide use is very low (Fig. 2). Does this also mean that the environmental impact of the Integrated system is much lower with respect to pesticides?

4.2. Environment exposure to pesticides Active ingredient input and EEP were quantified for the Conventional and Integrated system in 1988, and to demonstrate the progress that was made also for the Integrated system in 1992 (Table 6). From the Conventional 1988 system (representative for 1986-

Table 4 Annual input of pesticides (kg active ingredients/ha) in the Integrated and Conventional farming system (1986-1990) at Nagele Herbicides

Fungicides

Insecticides

Growth regulator Nematicides

Total

lnte- Conven- I n t e - Conven- I n t e - Conven- I n t e - Conven- l n t e - Conven- I n t e - Convengrated tional grated tional grated tional grated tional grated tional grated tional 0.1 a Ware potato Seed potato 2.0a Sugar beet 1.3 Winter wheat 1.2 Pea 2.1 Winter carrot 1.4 Sown onion 2.7 System average 1.4

2.5 a

4.5a 3.8 3.7 3.3 3.5 9.0 4.0

8.9 4.3 0.0 0.3 0.6 0.0 2.5 2.0

19.6 13.9 0.0 2.3 1.1 0.7 8.6 5.5

0.0 0.3 0.1 0.0 0.2 1.3 0.0 0.2

0.4 0.8 0.3 0.1 0.4 3.7 0.1 0.5

. 0.0 . . 1.8 0.1

.

.

.

. .

. .

0.6 . . 2.3 0.3

-

104.6 135.8

-

-

-

29.7

9.0 6.6 1.4 1.6 2.9 2.7 7.0 3.7

124.1 155.1 4.1 6.7 4.8 7.9 20.0 40.0

325 l 86-90 / 192-96 j X 92-961

Table 5 Number of interventions (n/crop) and fungicide input (kg active ingredients/ha) for Phytophthora infestans control in the Integrated and Conventional farming system (1986-1990) at Nagele Ware potato

Seed potato

Ii

100 90 80

Integrated Conventional Integrated Conventional

70 60

Interventions 6.3 Active 8.9 ingredients

11.4 19.5

2.6 4.2

5.4 12.0

50 40 30

1990) to the Integrated 1988 system the input of pesticides was strongly reduced over all crops (Table 6). In 1992 (representative for 1992-1996) the Integrated system reduced the pesticide input even further, again over all crops. The main cause of the drastic decline in EEP-air, water and-soil going from the Conventional 1988 system to the Integrated 1988 system is that DCP, the soil fumigant used to control potato cyst nematodes, has been replaced by non-chemical measures. DCP is extremely volatile and used in high dosages (80-110 kg active ingredients/ha). The major change going from the Integrated 1988 system to the Integrated 1992 system again occurs in the potato crop. The dithiocarbamates and fentin acetate, fungicides against late potato blight (Phytophthora infestans), were replaced by fluazinam, a new low dosage compound. From the Conventional 1988 system to the Integrated 1992 system the EEP-air, -water and -soil

0 fungicides

==herbicides

C 86-90

1 86-90

1 92-96

X 92-96

Fig. I. Annual input of pesticides (kg active ingredients/ha) in different systems and periods (I, Integrated; C, Conventional; X, Experimental) at Nagele.

20 10 0 ,Q_

t"

~

~

t~

m

¢-

..,

Fig. 2. Reduction (%) in input of pesticides (kg active ingredients/ ha) of the Integrated (I) (1986-1990) and the I and Experimental (X) farming system (1992-1996) in comparison to the Conventional farming system (1986-1990) at Nagele.

were reduced by >>99, 96 and 98%, respectively. The active ingredient use was reduced by 'only' 95% (including nematicides). The basis for the reduction in EEP is ICP. However the beneficial effect of selecting pesticides based on EEP is large as is apparent from the foregoing, especially in EEP-air. Table 7 presents, as an example the pesticides used in the Integrated 1992 system ranked according to their share in the farm average EEP-soil. Fluazinam is present at position 1 and 3 on the list, in ware-and seed potato (32%), respectively. It is used full field, to prevent late blight in potatoes, however in a low dosage system based on the higher resistance of the cultivars cropped in the Integrated system. Then herbicides used in band-spray appropriate dosage systems in sugar beet are present at position 2, 5, 11 and 14 (23%). To reduce the EEP further, harrow treatments might replace the last one or two low dose herbicide applications in sugar beet. Other compounds, that would decrease EEP are not available yet. Then glyphosate is used to control spot-wise perennial weeds in different crops (6, 9, 10; 12%). Pirimicarb and propiconazole are used in winter wheat (10%). However, the crop growth stage during application will largely prevent these compounds to reach the soil. Other compounds on the list contribute only marginally to the EEP-soil per farm.

326 Table 6 Pesticide use (kg/ha, active ingredients) and Environment Exposure to Pesticides (EEP)-air, -water and -soil by crop for the Conventional and Integrated farming system at Nagele in different years Active ingredients

Ware potato Seed potato Winter wheat Sugar beet Sown onion Average c

EEP-aira

1988

1992

1988

Conven- Integtional rated

lntegrated

Conventional

196.8 187.7 5.1 2.1 23.4 60.3

7.2 3.5 1.0 1.8 3.8 2.9

156 152 141 1.1 7.4 86

9.1 6.4 2.6 0.4 9.8 4.5

EEP-waterb

EEP-soilb

1992

1988

1992

1988

Integrated

Integrated

Conven- lntegtional rated

Integrated

Conventional

lntegrated

Integrated

0.9 0.9 1.5 0.1 3.8 1.3

0.5 0.3 0.4 0.4 0.7 0.4

3493 3218 162 103 722 1170

82 66 10 55 47 46

196.6 187.5 5.0 3.0 23.4 60.5

9.7 6.4 2.6 0.4 9.8 4.6

2.0 1.4 0.4 1.6 2.1 1.4

376 185 25 15 401 ! 29

1992

aln log (106 x EEP-air); bsee text; ¢cropping plan.

5. Discussion

The presented approach clearly distinguishes three phases in pesticide use" (1) the use (characterised by number of applications and kg active ingredients/ha);

(2) the exposure of the environment to pesticides (quantified by the EEP) and; (3) the effects on biota. The presented approach focuses on the first and second step. The governmental approval procedures for pesticides guarantee that the 'worst' pesticides

Table 7 Ranking of pesticides based on share in farm level EEP-soil for the Integrated farming system (1992) at Nagele Product

Type a

Active ingredients

Method b

Crop

Share farm level %¢

Cumulative share in farm level

1 2 3 4 5

Shirlan Goltix Shirlan Pirimor Betanal Progress

F H F I H

FF RT FF FF RT

Ware potato Sugar beet Seed potato Seed potato Sugar beet

22 11 10 7 7

22 33 43 50 57

6 7 8 9 10 11

Roundup Royal MH-30 Tilt Roundup Roudnup Betanal Tandem

H GR F H H H

SPOTW FF FF SPOTW SPOTW RT

Sugar beet Sowed onions Spring barley Sowed onions Chicory Sugar beet

7 5 3 3 3 3

64 69 72 75 78 81

12 13 14 15

Chloor-IPC Kerb Pyramin Flow Chloor-IPC

H H H H

Fluazinam Metamitron Fluazinam Pirimicarb Ethofumesate + desmedipham + fenmedipham G lyphosate Maleine-hydrazide Propiconazole Glyphosate Glyphosate Ethofumesate + fenmedipham Chlorpropham Propyzamide Chloridazon Chlorpropham

RT RT FF RT

Sowed onions Chicory Sugar beet Chicory

3 2 2 2

84 86 88 90

Ranking

a GR, growth regulator; F, fungicide; H, herbicide; I, insecticide. bFF, full field; RT, row treatment; SPOTW, spotwise. ¢Share in farm level in % = [(EEP by pesticide x crop share in farm area)/(EEP by farm)] xl00; Crop share in farm area = area of one crop/ total area of farm.

327

from an environmental and ecological point of view are not approved. ICP enables a minimum use of pesticides. EEP is a useful instrument to select then within the range of approved pesticides. EEP in combination with ICP enables a quantitative approach to a stepwise, targeted reduction in pesticide use and environmental impact. Van der Werf (1996) reviewed different approaches to evaluate pesticide impact on environment and biota. The reviewed methods show considerable differences in the parameters that are considered to asses environmental impact. From this review that includes the proposed approach in this article it is clear that EEP is the only approach that takes volatilisation of pesticides into account. It is also the only one that 'on purpose' does not consider effect on biota, since an overall comprehensive assessment is virtually impossible. Overall quantitative scores of 'ecosafety' therefore may easily lead to unjustified classification of a pesticide as being safe. The more radical approach of EEP enables a basic approach towards prevention. The volatilisation losses are obviously the largest ones and thus have to be taken into account when trying to reduce environmental impact. It can be argued, whether the parameter VP is the best to account for volatilisation losses. Henry's law constant (Kh, the ratio of the VP to the water solubility) might also be an appropriate criterion for the volatility of a pesticide. However in a recent study in the Netherlands VP was also chosen in the model calculations as most simple and accurate prediction parameter (Multi-Year Crop Protection Plan, 1996). Both DT50 and Komare in fact exponentially related to the leaching risk. Nevertheless by using the straight forward ratio DT50/Kom the risk of leaching is estimated properly. The presented approach might be extended to consider all involved processes, however the risk of loosing the simple and user-friendly character of EEP should be taken seriously. The Integrated prototype as designed, tested and improved on the Nagele experimental farm for the Central Clay conditions proved to have good perspectives in terms of minimising pesticide input and environment exposure to pesticides. ICP and pesticide selection based on EEP proved to be effective. Since the economic perspectives of the Integrated prototype are equal to those of the current 'conventional' farm-

ing systems, large scale implementation in practice inevitably should be the next step

Acknowledgements The author cordially thanks Dr. Boesten, Ir. Spoorenberg and particularly Dr. Vereijken for the support in developing the concept of Environment Exposure to Pesticides.

References Atlas, E.A. and Schauffler, S., 1990. Concentration and variation of trace organic compounds in the north pacific atmosphere. In: D.A. Kurtz (Editor), Long Range Transports of Pesticides. Lewis, Chelsea, MI, pp. 161 - 183. Bos, A., Janssens, S.R.M. and Krikke, A.T., 1992. Analysis of economic results. In: H.H. Cheng (Editor), More Sustainable Farming Systems for Arable Farming. Themaboekje hr. 14, PAV, Lelystad, pp. 126-181 (in Dutch). Gregor, D.J. and Gummer, W.D., 1989. Evidence of atmospheric transport and deposition of organochiorine pesticides and polychlorinated biphenyl's in Canadian arctic snow. Environ. Sci. Technol., 23: 561-565. Leonard, R.A., 1990. Movement of pesticides into surface waters. In: Pesticides in the Soil Environment. Soil Science Society of America Book Series, No. 2, Madison, WI, pp. 303-349. Linders, J.B.M.J., Jansma, J.W., Mensink, B.J.W.G. and Ottermann, K., 1994. Pesticides: Benefaction or Pandora's Box, A Synopsis of the Environmental Aspects of 243 Pesticides. Report no. 6791014. National Institute of Public Health and Environmental Protection, Bilthoven, 201 pp. Ministry of Agriculture, Nature Management and Fisheries, 1990. Agriculture Structure Memorandum. Government decision (in Dutch). Ministry of Agriculture, Nature Management and Fisheries. DS, The Hague (essentials available in English). Ministry of Agriculture, Nature Management and Fisheries, 1991. Multi-Year Crop Protection Plan. Government decision. (In Dutch). Ministry of Agriculture, Nature Management and Fisheries. SDU, The Hague (essentials available in English). Multi-Year Crop Protection Plan, 1996. Multi-Year Crop Protection Plan. Evaluation emission 1995, background document. IKC-L. Ede, 127 pp. plus annexes (in Dutch). Pimentel, D., 1995. Amounts of pesticides reaching target pests: environmental impacts and ethics. J. Agric. Environ. Ethics, 8: 17-29. Schomburg, C.J. and Glotfelty, D.E., 1991. Pesticide occurrence and distribution in fog collected near Monterey, California. Environ. Sci. Technol., 25: 155-160. Simonich, S.L. and Hites, R.A., 1995. Global distribution of organochlorine compounds. Science, 269:185 l - 1854. Taylor, A.W. and Spencer, W.F., 1990. Volatilisation and vapor transport processes. In: H.H. Cheng (Editor), Pesticides in the

328

Soil Environment. Soil Science Society of America Book Series, No. 2, Madison, WI, pp. 213-269. The Pesticides Trust UK, information leaflet. Eurolink Business Centre, 49 EFFRA Road, London, SW2 IB2, UK. Vereijken, P., 1994. 1. Designing Prototypes. Progress Reports of Research Network on Integrated and Ecological Arable Farming Systems for EU- and Associated Countries (concerted action AIR3-CT927705). AB-DLO, Wageningen, 87 pp. Vereijken, P., 1995.2. Designing and Testing Prototypes. Progress Reports of Research Network on Integrated and Ecological Arable Farming Systems for EU- and Associated Countries (concerted action AIR3-CT927705). AB-DLO, Wageningen, 76 pp. Vereijken, P., 1997. A methodical way of prototyping integrated and ecological arable farming systems (I/EAFS) in interaction with pilot farms. In: M.K. van lttersum and S.C. van de Geijn (Editors), Proceedings of the 4 th ESA Congress, Elsevier, Amsterdam. Vereijken, P. and D.J. Royle, 1989. Current status of Integrated arable farming systems research in Western Europe. IOBC/ WPRS Bull., XII: 5.

Werf van der, H.G.M., 1996. Assessing the impact of pesticides on the environment. Agric. Ecosyst. Environ., 60: 81-96. Wijnands, F.G., 1992. Introduction and evaluation of integrated arable farming in practice. Neth. J. Agric. Sci., 40: 239-250. Wijnands, F.G., 1994. Focus on IAFS prototyping in Nagele experimental farm (NL1). In: P. Vereijken (Editor) 1. Designing Prototypes. Progress Reports of Research Network on Integrated and Ecological Arable Farming Systems for EU- and Associated Countries (concerted action AIR3-CT927705). AB-DLO, Wageningen, pp. 79-84. Wijnands, F.G. and Vereijken, P., 1992. Region-wise development of prototypes of Integrated arable farming and outdoor horticulture. Neth. J. Agric. Sci., 40: 225-238. Wijnands, F.G., van Asperen, P., van Dongen, G.J.M., Janssens, S.R.M., Schrrder, J.J. and van Bon, K.B., 1995. Innovatiebedrijven ge'fntegreerde akkerbouw, beknopt overzicht technische en economische resultaten. PAGV-verslag nr. 196. PAV, Lelystad, 126 pp. (in Dutch with English summary and tables and figures).

1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van lttersum and S.C. van de Geijn (Editors)

329

Use of agro-ecological indicators for the evaluation of farming systems C. Bockstaller a'*, P. Girardin b, H.M.G. van der

Werf b

aAssociation pour la Relance Agronomique en Alsace (ARAA), Laboratoire d'Agronamie, 68021 Colmar Cedex, France blNRA, Laboratoire d'Agronomie, BP 507, 68021 Colmar Cedex, France


Abstract

For the development of integrated arable farming systems (IAFS), tools are needed to evaluate the achievement of agronomic and environmental objectives, in order to optimize the systems. A set of agro-ecological indicators (AEI) is proposed. These indicators estimate the impact of cultivation practices on the agrosystem and its environment. AEI are aimed, first of all, at being used as decision aid tools, to help farmers to adapt their cultivation practices to IAFS requirements, from one cropping year to the next. So far, seven indicators have been elaborated for the evaluation of farming systems: crop diversity, crop succession, pesticide, nitrogen, phosphorus, organic matter and irrigation. The calculation method for the organic matter and pesticide indicators is presented. Possibilities for use of the AEI at the farm and field level, for farmers and decision makers are given with data from a network of 17 commercial arable farms. The elaboration of a single aggregated indicator is discussed. © 1997 Elsevier Science B.V. Keywords: Integrated arable farming; Environmental assessment; Indicator; Organic matter; Pesticide; Agro-ecological indicators; Environmental impact; Decision aid tool; Soil fertility; Hdnin-Dupuis model; Fuzzy logic; Farm network; Multi-criteria method

I. Introduction

Integrated arable farming systems (IAFS), based on the concepts of Integrated Agriculture or Integrated Production (El Titi et al., 1993), are generating an increasing interest as an alternative to conventional intensive farming systems. The results from several pioneer IAFS projects in the 80s and early 90s, concerning profitability as well as environmental and agronomic effects, are promising (Holland et al., 1994). However, for the development of IAFS, tools

* Corresponding author. E-mail: [email protected]

are needed to evaluate the achievement of the objectives (Girardin and Spiertz, 1993), in order to optimize the system (Vereijken, 1992). This kind of evaluation concerning especially the environmental and agronomic objectives needs new methods taking into account specific criteria. Economic criteria used in modern market-oriented agriculture such as the yield or gross margin are no longer sufficient for a global evaluation of agricultural practices, which should include an assessment of their environmental impact. The most obvious approach to environmental impact evaluation is based on direct measurements at the field level. Such a solution is possible on an experimental farm, but its extension to commercial


330

farms for routine measurements poses practical problems, because measurements are costly and often time consuming (Sharpley, 1995). Simulation models may be used for impact evaluation, but comprehensive models required for a multi-objective evaluation are not available. Another problem of many models is that they are not adapted for use at farm level, requiring too many input data. In other cases models are not validated for a broad range of conditions (Hansen, 1996). Indicators are an alternative when it is not possible to carry out direct measurements. The term 'indicator' has been defined as a variable which supplies information on other variables which are difficult to access (Gras et al., 1989). Indicators help to understand and to interpret a complex system by: (1) synthesizing data; (2) showing the current state; (3) demonstrating the achievement or not of objectives; (4) communicating the current status to users for management decisions (Mitchell et al., 1995). The parameters proposed by Vereijken et al. (1995) in their methodological framework for prototyping IAFS are either based on direct measurements (e.g. nitrate in ground water) or are calculated from data available on the farm, (e.g. environment exposure to pesticides; EEP). As mentioned before, parameters based on measurements present practical problems for use on a broad scale. The second type of parameter corresponds to what we call indicators. However, the expression of these parameters in physical units as proposed by Vereijken et al. (1995) may not be easily understood by farmers for some parameters (e.g. EEP air expressed in kg/ha per year). As an alternative to indicators expressed in physical units, we propose a set of agro-ecological indicators (AEI) to evaluate the degree of achievement of the IAFS objectives by farming systems. Unlike the parameters proposed by Vereijken et al. (1995), the AEI are calculated with data available on the farm, and thus do not require specific field measurements. Neither are AEI expressed in physical units, but on a scale between 0 and 10, to make them easily understandable by farmers. The IAFS objectives will be restricted to the environmental effects of farming systems (e.g. nitrogen losses) and to some agronomic effects (e.g. on soil fertility). In this way, AEI are estimators of the impact of cultivation practices on the agrosystem and its environment. However, as

mentioned by Mitchell et al. (1995), the assessment of environmental impacts should affect management decisions. We therefore consider AEI to be, first of all, decision aid tools to help farmers adapt their cultivation practices to IAFS requirements, from one cropping year to the next. However, AEI may also be used by decision makers to monitor or to evaluate their agri-environmental policies. The purpose of this paper is to present the principles of elaboration of the AEI, illustrated by the calculation methods of two indicators, and to discuss the possibilities for their use.

2. Elaboration of the AEI

An indicator assesses the impact of a cultural practice on one or several objectives (Table 1). We defined a list of environmental and agronomic objectives in relation to the principles of IAFS (El Titi et al., 1993). Some of them are close to the objectives defined by Vereijken et al. (1994). We did not include economic and social objectives. From the list of objectives a set of indicators was defined and connected to the objectives (Table 1). These indicators are identified as 'key' components for the management of a farming system. Crop rotation (or cropping sequence) and cultivation practices within each crop make up the two parts of an arable farming system. Cultivation practices are assessed by the AEI listed in Table 1. We added the crop sequence indicator (Bockstaller and Girardin, 1996) to that list of AEI. The proposed crop sequence indicator assesses whether the cropping sequence satisfies the requirement of IAFS, for example concerning the use of natural regulation mechanisms, or of nitrogen supply from the previous crop (El Titi et al., 1993). The crop sequence indicator, contrary to the other AEI, is not aimed at being used as an estimator of an environmental impact but as a decision aid tool along with the other AEI, in order to help farmers adapt their farming system to the principles of IAFS. For simplicity of presentation, AEI are constructed in such a way that they take a value between 0 and 10. The value 7 represents the achievement of a minimum level for IAFS requirements (e.g. the organic matter content is maintained at a satisfactory level). A value below 7 indicates that the IAFS requirements are not

331

Table 1 List of agro-ecological indicators in association with their environmental and agronomic objectives Agro-ecological indicators Objectives

Nitrogen Phosphorus Pesticide Irrigation Organic Energy Crop matter diversity

Soil Soil structure cover

Ecological structures

Protection of'.

Ground water X quality Surface water quality Air quality Soil quality Non-renewable resource Biodiversity Landscape quality

X (x)

x x

X

x x (x)

x

x

x

x

(x)

x x

x

x x

Each mark represents an association, (X) means that the objective is not fully assessed by the indicator for the moment. met (e.g. the soil organic matter content decreases) and a value above 7 indicates that the farmer does better than the minimum IAFS requirements (e.g. the soil organic matter content increases). The reference level, corresponding to a value of 7, is based on scientific knowledge or expert judgment. This level may be adapted to local conditions by local experts. For example: maintaining the soil organic matter at a satisfactory level is an IAFS principl e admitted by experts. This level should be quantified according to local conditions (soil type, climate, etc.). The AEI are calculated with data available on the farm (cultivation practices recorded by the farmer, soil analyses, permanent characteristics such as field size, slope, etc.). All indicators, except for the crop diversity indicator, are calculated at the field level and then weighted by the field size to obtain a mean farm value. The time scale of calculation is generally the period between the harvest of the preceding crop and the harvest of the crop in the current year. The calculation algorithm of an AEI is based on available scientific knowledge or on expert judgement. In some cases (e.g. organic matter indicator, see Section 3.1) the indicator is based on a simple model for which the required data are available on the farm. The nitrogen and phosphorus indicators compare the farmer's practices with a sequence of practices corresponding to the minimum requirements for IAFS. A sequence of farmer's practices in accor-

dance with this reference sequence yields an indicator value of 7. Practices deviating from the reference sequence may lead to loss of nitrogen (or phosphorus) to the environment and will yield a sum of plus or minus marks, which is added to the value of 7. These plus or minus marks are based on the equivalence: one point is equal to 30 kg N/ha or 30 kg P2Ofl ha. Thus, in the case of the nitrogen and phosphorus indicators, or when a simulation model is used, the evaluation is quantitative. For other indicators such as crop diversity, crop sequence and pesticide, the evaluation is qualitative because quantitative data and adapted models are not available. So far, seven indicators have been elaborated for the evaluation of farming systems: nitrogen, phosphorus, pesticide, irrigation, organic matter, crop diversity and crop sequence. Indicators for: energy, soil structure, soil cover and ecological structures are planned. Following their elaboration, the indicators should be evaluated in order to improve them, if necessary. Three tests are proposed. The sensitivity test aims to observe the behaviour of the indicator when the value of its input variables is varied. Two examples will be given below. The purpose of the probability test is to study the soundness of the indicator as an estimator of environmental impact. The test consists of the establishment of a relationship between the values taken by the indicator and those taken by an observed or mea-

332

sured criterion which reflects the environmental impact. The usefulness test should be carried out in order to see whether the tools developed are helpful for the target users. These two latter tests will not be elaborated in this paper.

3. Calculation algorithm of two indicators

O t~

.u

3.1. The organic matter indicator The organic matter indicator (lMo) evaluates the effect of farmers' practices on the evolution of soil organic matter (SOM) in order to help farmers adapt their cultural practices to maintain the SOM at a satisfactory level. The calculation of the indicator as given in Eq. (1) is based on the comparison of the organic matter (OM) inputs by manure and crop residues with recommended levels of inputs:

IMo = 7Ax/A a

(1)

where A x is the mean OM input of the 4 preceding cropping years, and AR is the recommended level of OM input. The indicator ranges from 0 to 10. If the indicator is less than 7, the organic matter inputs are not sufficient, if it is equal to or above 7, the recommended levels of input are reached or exceeded. The recommended levels of inputs are expected to maintain a satisfying level of SOM in the long term. They were obtained by running the H6nin-Dupuis model (Mary and Gu6rif, 1994) for several classes of clay and limestone contents in the soil (Table 2). The H6nin-Dupuis model is a monocompartment model of the evolution of SOM. The calculation of OM inputs by crops and manure are based on data of Boiffin et al. (1989). Examples of OM inputs are given in Table 3. The indicator was Table 2 Recommended level of OM inputs (AR) (kg OM/ha) Clay content (%) Limestone content

15-20

20-25

25-30

30-35

0-5% 5-15% > 15%

1085 945 840

945 840 735

910 770 700

840 735 630

6 5 4 i

3 2

1 0 17% 28% 17% 28% 17% Clay Clay Clay Clay Clay 0% 0% 10% 10% 20% CaO CaO CaO CaO CaO

28% Clay 20% CaO

Fig. 1. Sensitivity of the organic matter indicator to soil characteristics (clay and limestone contents), to the crop succession (crop residues incorporated, without manure) and to yield. The indicator was calculated with data from Tables 2 and 3. Crop succession 1: grain maize monoculture (yield, respectively, 11 t/ha and 9 t/ha); crop succession 2: grain maize (11 t/ha)/winter wheat (8 t/ha); crop succession 3: sugar beet (70 t/ha)/winter wheat (8 t/ha)/grain maize (10 t/ha)/winter wheat (8 t/ha). 0, crop suc. 1 (yield 1 ! t.ha-J); O, crop suc. 1 (yield 9 t.ha-~); ~, crop suc. 2; • crop suc. 3; - - - , recommended value.

calculated for a range of clay and limestone contents in soil, for several rotations and yield levels, with data from Tables 2 and 3. Fig. 1 shows the sensitivity of the indicator to these variables.

3.2. The pesticide indicator (Ipest) The environmental impact of a pesticide largely depends on: (a) the amount applied, (b) its rate of degradation, (c) its partitioning to the air, the surface water and the groundwater, (d) its toxicity to the species in those environmental compartments (van der Weft, 1996). Several methods have been proposed to estimate pesticide environmental impact (Levitan et al., 1995; van der Weft, 1996). None of these methods aggregates the four criteria mentioned above into a single output value. The pesticide indicator (Ip~t) we propose is based on an expert system using a collection of fuzzy mem-

333

value depends on a set of 16 decision rules, five of which are given below as an example:

bership functions and decision rules. This technique is robust when uncertain or imprecise data is used. It also allows the aggregation of knowledge which is expressed in every-day language (Bouchon-Meunier, 1993). In a first step, /pest is calculated for each single application of an active ingredient. The value of/pest depends on four modules: P (presence, reflecting amount and persistence), Rgro (risk of groundwater contamination), Rsur (risk of surface water contamination) and Rair (volatilization risk). The value of each of these modules depends on two to four input variables according to fuzzy decision rules which will not be presented here. For all modules the membership to a fuzzy set F (favorable) and a fuzzy set U (unfavorable) has been defined. The value of P depends on the rate applied and its soil degradation half-life. The value of Rgro depends on the leaching potential of the pesticide, the site of application (in the soil, on the soil, or on the crop), the month of application and its toxicity to man. The value of Rsur depends on the run-off risk of the field, the site of application and the toxicity of the pesticide to aquatic organisms. The value of Rair depends on the volatility of the pesticide and its site of application. For the air the toxicity is not taken into account because an appropriate variable is not available. The four modules (P, Rgro, Rsur and Rair) can be either considered individually or aggregated in a single value, for instance by summation, multiplication or a combination of both. We present here a mode of aggregation using decision rules to calculate Ipe~t. For each application of an active ingredient,/pest can take values between 0 (no risk of environmental impact) and 1 (maximum risk of environmental impact). Its

(a) If P is F and Rgro is F and Rsur is F and Rair is F then/pest is 0.0; (b) If P is F and Rgro is F and Rsur is F and Rair is U then/pest is 0.1; (c) If P is F and Rgro is U and Rsur is U and Rair is U then/pest is 0.5; (d) If P is U and Rgro is F and Rsur is F and Rair is F then/pest is 0.5; (e) If P is U and Rgro is U and Rsur is U and Rair is U then/pest is 1.0. Volatilization risk is given less weight than groundwater and surface water contamination risks (rules a, b and c). This weighting is an example of expert judgment. We valued the pollution of air as less 'important' than the pollution of surface water or groundwater. The presence of a pesticide is considered as an environmental risk, even when the risk for each of the environmental compartments is nil (rule d). Examples of calculation for a range of active ingredient applications are given in Table 4. An analysis of the sensitivity of/pest to variation of input variables is presented in Fig. 2. Each input variable was varied from favorable (0%) to unfavorable (100%), while the other input variables are kept at their median value. In a second step the indicator (IpEsr) is calculated for all pesticide applications on a crop within a year. At this level/PEST takes values between 0 (highest risk) and 10 (no risk): •PEST

10 - k ,Y_,/pesti

""

where k is a constant depending on the crop and Ipesti

Table 3 Examples of OM inputs (kg dry OM/ha) obtained with data from Boiffin et al. (1989). Crop yield (t/ha) Crop

6

7

8

9

10

I1

Wheat (straw incorporated) Wheat (straw exported) Grain maize Sugar beet ( l e a v e s incorporated)

470 250 .

550 290 -

620 330 650

700 370 740 .

780 420 820

. . 900

.

.

.

.

.

12 . .

60 . .

980

. . 420

70

80

500

570

. .

OM inputs include straw and root contribution to the steady SOM. They are derived from root, stem and leaves dry matter. These latter values are obtained from the crop yield (expressed at normalized humidity: 16% for wheat and maize, 85% for sugar beet).

334

Table 4 The values of P, Rgro, R.... Rair and/pest for a number of pesticides applied at their recommended rate in a field with medium run-off risk Pesticide (active ingredient) Rimsulfuron Cyfluthrin 2,4-D Parathion EPTC Carbofuran Glyphosate Alachlor Isoproturon Atrazine Lindane

Application

Sitea (%)

Month

Rate

P/S P/S P/S P/S S S P/S P/S P/S P/S S

June July April Aug. April April April April Jan. April April

0.015 0.040 0.600 0.300 3.600 0.600 4.300 2.400 1.800 1.500 1.350

(50) (100) (50) (100)

(100) (0) (10) (0)

P (kg/ha)

Rgro

Rsur

Rair

]pest

0.00 0.07 0.12 0.23 0.50 0.35 0.66 0.52 0.51 0.63 0.88

0.00 0.00 0.20 0.00 0.00 0.65 0.00 0.23 0.31 0.84 0.46

0.18 0.00 0.19 0.00 0.00 0.00 0.00 0.47 0.48 0.34 0.00

0.00 0.98 0.00 0.48 0.00 0.00 0.00 0.00 0.00 0.00 0.00

0.02 0.11 0.13 0.15 0.25 0.30 0.37 0.39 0.41 0.55 0.55

ap/s, applied on plant/soil; S, applied in the soil. The percentage (in parentheses) indicates the crop cover of soil at the time of application.

are first of all decision aid tools for farmers. The type of presentation shown in Fig. 3 is inspired by Magnollay (1993). Vereijken (1997) also used this presentation. It gives an overview of the AEI values at the farm level, allows the comparison with r e c o m m e n d e d values and shows the weak and strong points of an arable fanning system. The farmer knows which cultivation practices he should improve according to IAFS requirements. For most of the indicators, results at the field level are available, so they can be used to help the farmers to take into account the differences between the fields

is the value of the indicator (ranging between 0 and 1) for a single application of an active ingredient i. The constant k is chosen such that a value of 7 for/PEST is obtained when a crop protection program satisfies the m i n i m u m r e c o m m e n d a t i o n s for IAFS.

4. Using agro-ecological indicators The AEI were calculated in 1994 and 1995 with data from a network of 17 commercial arable farms of the Rhine plain in France and Germany. The AEI

0.60 [

•

0.55 i

Soil half-life

--÷-- Soil mobility (GUS)

,-, 0.50

:,-,:$-:---.$

--- - Amount applied

.2 0.45 . . . - - Month

.,-~

.~ o.4o - -. - : : t t ' . ' .

" ll"

e

0.35

Runoffrisk

-- o-- Crop cover

0.30 0.25 [ 0.20 /

0

= ,

10

20

~

!

t

,

,

,

~

;

30

40

50

60

70

80

90

100

Transition interval of each input variable (%)

Toxicity-human

-- a - - Volatility =

Toxicity-aquatic

Fig. 2. Analysis of the sensitivity of the pesticide indicator (Ipe~t)to variation of input variables. Each input variable is varied from favorable (0%) to unfavorable (100%), while the other input variables are kept at their median value.

335

Crop diversity IO

I0

IO

Organic matter

Crop sequence . . . . . . Farm vaalue in 1994

:t

---

Farm value in 1995 Recommended value

Phosphorus Io

lo Nitrogen

Fig. 3. Example of use of the agro-ecological indicators at the farm level (90 ha: grain maize, sugar beet, winter rape, winter wheat). as shown by Fig. 4 in case of the organic matter indicator. This type of presentation helps the farmer to adapt his management to the specific conditions of each field (soil type, crops, etc.). Another use for farmers would be the possibility to follow up the evolution of his cultivation practices over several years or to analyze the cropping history. These indicators can also be used by decision makers (politicians, environmentalists, etc.) to follow up the evolution of cultural practices and the influence of an agri-environmental policy on a sample of farms in a given area (e.g. water catchment). Results can be

presented as in Fig. 5, which shows the organic matter indicator. In this example, the differences between the 2 years are minor, because farmers did not change their rotations and organic fertilization management. We used the database Microsoft Access 2.0® to implement the calculation of the AEI. The software is user-friendly, and with the help of a handbook or a short training it can be used by farmers or farmeradvisers. Farmers may need help from their advisers to enter a few parameters (nitrogen soil mineralization, recommended nitrogen and phosphorus fertilizations, etc.).

10 ~. • ""

.~

R e c o m m e n d e d value

9 8 7 6 5 4 3

1 0 2

9

14

14

18

19

25

25

26

26

28

Field number Fig. 4. Example of calculation at the field level for the organic matter indicator.

31

336 Recommended value

10 9 O 8 ~3 7 "k. "

6

5 4 3

1994 m 1995

1 0 1

2 3 4 5 6

7

8 9 1011121314151617 Farm number

Fig. 5. Value of the organic matter indicator in 1994 and 1995 for the 17 farms of the network (1-13: France; 14-17: Germany).

5. Discussion The type of evaluation using AEI yields results which are easy to understand. This is an important feature for decision aid tools. Obviously, when given to farmers or to decision makers, such results should be followed by an interpretation and advice about ways to improve poor indicator values. In this way the AEI will be really used as decision aid tools. The use of a set of indicators may seem an analytical approach to describe a complex system. However, the approach is holistic in the sense that we deal with all cultivation practices within a farming system by means of a set of indicators. Furthermore, one AEI is generally related to several environmental compartments (e.g. in the case of the pesticide indicator) and often does not deal with a single environmental problem. It thus helps farmers to take into account the whole agrosystem and the adjacent ecosystem. On the other hand, because these tools are intended to be used to help farmers, there was a necessity to deal with elements reflecting the farmers' reality. Each indicator is related to a set of practices which are interrelated. The presentation of several AEI, as in Fig. 3, brings

up the question of the importance of each indicator. No comparison is possible between the results of the different indicators because one unit does not have the same meaning in each indicator. This problem is especially acute if there is a need to classify farms or farming systems by means of a set of indicators. A single aggregated indicator resulting from the aggregation of the set of AEI might be used in this case. This aggregation will involve weighting each AEI according to the user and his objectives, which involves a certain degree of subjectivity. Some authors (e.g. Hansen, 1996) do not accept this subjectivity. Another problem of a single aggregated indicator is the compensation which can occur between the values of its components. For instance, low nitrate leaching risk cannot balance a higher risk of pesticide volatilization. Such compensation has no scientific basis and it is not acceptable if global environmental impact is assessed. A single major environmental risk is sufficient to put in question the sustainability of the system. The use of multi-criteria methods (Simos, 1990) can be an alternative approach to aggregate the information supplied by a set of indicators. This approach

337

supplies a solution to the problem of compensation, and to rank or to classify actions (in our case farming systems). In this case, techniques of operational research are used. The researchers in this discipline generally accept the subjectivity inherent to decision making (Roy, 1992). In fact, there are no scientific results or rules to decide which impact is more important (e.g. between risk for soil quality or for water quality). At this stage of our work, we did not introduce any weighting of the indicators for the reasons mentioned above. Neither did Vereijken (1997) in his figure presenting the results of the set of parameters. The main objective of our AEI is to help farmers improve their management. In the perspective of a global approach as in IAFS, we consider that the value of each indicator should be improved and attain at least the reference value of 7.

6. Conclusion The approach presented in this paper is based on a set of indicators which are, first of all, decision aid tools to help farmers adapt their cultivation practices to IAFS requirements. The results of the AEI are expressed in a simple and straightforward fashion, which is an important quality for decision aid tools. Their calculation is based on data available on the farm and has been implemented in an user-friendly software tool, so that farmers or farmer-advisers can calculate the AEI. For other production systems with livestock or perennial crops, our indicators can be adapted and other specific indicators should be developed. The use of a set of indicators raises the question of the relevance of a single aggregated indicator. Several possibilities exist to aggregate the individual indicators. All these methods present a certain degree of subjectivity, which is inherent to human decision making.

Acknowledgements This work is sponsored by the EU (Interreg programme), the Land Baden-Wtirttemberg (Germany) and the Alsace Region (France) as part of the ITADA programme (C. Bockstaller), the 'Agriculture

Demain' programme of the French Research Ministry (C. Bockstaller), and by a EU Research Training Fellowship (H.M.G. van der Werf). The technical assistance of C. Zimmer (programming of the software) is gratefully acknowledged.

References Bockstaller, C. and Girardin, P., 1996. The crop sequence indicator; a tool to evaluate crop rotations in relation to the requirement of Integrated Arable Farming Systems. Aspects of Applied Biology 47, Rotations and Cropping Systems, Association of Applied Biology, Warwick, UK, pp. 405-408. Boiffin, J., K61i Zagbahi, J. and Sebillote, M., 1989. Syst~mes de culture et statut organique des sols dans le Noyonnais: un essai d'application du module de H6nin et Dupuis. In: M. Sebillote (Editor), Fertilit6 et Syst~mes de Production. lnstitut National de la Recherche Agronomique, Paris, France, pp. 234-258. Bouchon-Meunier, B., 1993. La Logique Floue. Presses Universitaires de France, Paris, France, 128 pp. E! Titi, A., Boiler, E.F. and Gendrier, J.P., 1993. Integrated production. Principles and technical guidelines. IOBC/WPRS Bull., 16: 13-38. Girardin, P. and Spiertz, J.H.J., 1993. Integrated agriculture in western Europe: Researchers' experience and limitations. J. Sustainable Agric., 3: 155-170. Gras, R., Benoit, M., Deffontaines, J.P., Duru, M., Lafarge, M., Langlet, A. and Osty, P.L., 1989. Le fait technique en agronomie. Activit~ Agricole, Concepts et M6thodes d'l~tude. Institut National de la Recherche Agronomique, L'Hamarttan, Paris, France, 184 pp. Hansen, J.W., 1996. Is agricultural sustainability a useful concept? Agr. Syst., 50:117-143. Holland, J.M., Frampton, G.K., t~ilgi, T. and Wratten, S.D., 1994. Arable acronyms analysed - a review of integrated arable farming systems research in western Europe. Ann. Appl. Biol., 125: 399-438. Levitan, L., Mervin, I. and Kovach, J., 1995. Assessing the relative environmental impacts of agricultural pesticides: the quest for a holistic method. Agric. Ecosystems Environ. 55: 153-168. Magnollay, F., 1993. R6seau PI: des progr~s mesurables. Revue Suisse Agric., 25: 361-363. Mary, B. and Gu6rif, J., 1994. lnt~r~ts et limites des modules de pr6vision de l'6volution des mati~,res organiques et de i'azote du sol. Cahiers Agric., 4: 247-257. Mitchell, G., May, A. and McDonald, A., 1995. PICABUE: a methodological framework for the development of indicators of sustainable development. Int. J. Sustain. Dev. World Ecol., 2: 104-123. Roy, B., 1992. Science de la d6cision ou science de I'aide ~ la d6cision? Rev. Int. Syst6mique, 6: 980-1472. Sharpley, A.N., 1995. Dependence of runoff phosphorus on extractable soil phosphorus. J. Environ. Qual., 24: 920-926. Simos, J., 1990. Evaluer I'Impact sur l'Environnement. Presses

338

Polytechniques et Universitaires Romandes, Lausanne, Swizerland, 261 pp. Van der Werf, H.M.G., 1996. Assessing the impact of pesticides on the environment. Agric. Ecosystems Environ., 60: 81-96. Vereijken, P., 1992. A methodic way to more sustainable farming systems. Neth. J. Agr. Sci., 3: 209-223. Vereijken, P., 1997. A methodical way of prototyping integrated and ecological arable farming systems (I/EAFS) in interaction with pilot farms. In: M.K. van Ittersum and S.C. van de Geijn (Editors), Proceedings of the 4th ESA Congress, Elsevier, Amsterdam, pp. 56-60.

Vereijken, P., Wijnands, F., Stol, W. and Visser, R., 1994. Progress Report 1. Designing Prototypes. Progress repots of the research network on integrated and ecological arable farming systems for EU and associated countries (Concerted action AIR 3 CT920755) AB-DLO, Wageningen, The Netherlands, 87 pp. Vereijken, P., Wijnands, F. and Stol, W., 1995. Progress Report 2. Designing and Testing Prototypes. Progress repots of the research network on integrated and ecological arable farming systems for EU and associated countries (Concerted action AIR 3 -CT920755) AB-DLO, Wageningen, The Netherlands, 90 pp.


339

Model-based explorations to support development of sustainable farming systems" case studies from France and the Netherlands W.A.H. Rossing a'*, J.M. Meynard b, M.K.

van Ittersum a

aDepartment of Theoretical Production Ecology, WageningenAgricultural University, P.O. Box 430, 6700 AK Wageningen, The Netherlands bUnit~, d'Agronomie INRA-INA PG, F-78850 Thiverval-Grignon, France Accepted 13 June 1997

Abstract Sustainable land use requires development of agricultural production systems that, in addition to economic objectives, contribute to objectives in areas such as environment, health and well-being, rural scenery and nature. Since these objectives are at least partially conflicting, development of sustainable farming systems is characterized by negotiation about acceptable compromises among objectives. Four phases can be distinguished in the course of farming systems development: diagnosis, design, testing and improvement, and dissemination. During the last decade an approach coined 'prototyping' has emerged as a promising method for empirical farming systems development in Western Europe. Limitations of the approach include: (1) the limited number of systems that can be evaluated, resulting in a lack of perspective on conflicts among objectives, and (2) the expertise-based nature of rules used during systems design which unduly narrows the range of available options and obscures understanding of systems behaviour. In the paper, explorative studies based on transparent models of agronomy and management are put forward to supplement empirical prototyping and to remedy its shortcomings. To illustrate the potential of model-based explorations, two case studies are presented. The first case study deals with diagnosis and design of wheatbased rotations in the Paris Basin of France, aimed at alleviating tactical problems of poor resource-use efficiency within the constraints imposed by existing crop rotations. The second case study addresses design of sustainable bulb-based farming systems in the Netherlands with the purpose of investigating strategic options at crop rotation and farm level to resolve conflicts between economic and environmental objectives. In the discussion, methodological elements of model-based explorations and interaction with stakeholders are addressed, and opportunities for enhanced development of sustainable farming systems are identified. © 1997 Elsevier Science B.V. Keywords: Sustainable agriculture; Farming systems; Cropping systems; Prototyping; Model-based learning; Participatory research

I. Introduction I n m a n y parts of Europe, arable farmers have been very successful in increasing yields per unit area dur*Corresponding author. Tel.: +31 317 484766; fax: +31 317 484892; e-mail: [email protected]

ing the last decades. However, the production techniques that were utilized have resulted in negative side effects" emissions of pesticides and plant nutrients, (in)organic waste, high energy consumption. Public concern is reflected in a suite of national and international policy statements that call for more sustainable agricultural farming systems.

Reprinted from the European Journal of AgronOmy 7 (1997) 271-283

340

For operational purposes, sustainability can be defined as a combination of socio-economic, ecological and agro-technical objectives of agricultural production (WRR, 1995). The weight attributed to these objectives is value-driven and varies among interest groups. Because the objectives are usually at least partially conflicting, development of sustainable farming systems is characterized by negotiation about acceptable compromises among objectives by various stakeholders. Actors include farmers, agricultural industry, consumers and the public sector. Agricultural research contributes to this process by developing methodology to demonstrate consequences of alternative options. During the last decade, a promising empirical methodology for developing sustainable farming systems has been elaborated, coined 'prototyping' (Vereijken, 1994, 1997). Prototyping involves application-oriented design and testing of farming systems in collaboration with commercial farmers or at experimental farms, according to a methodical approach. Four phases can be distinguished: diagnosis, design, testing and improvement and dissemination. In the diagnostic phase the objectives of agricultural production and the value-driven weights attached to them by various interest groups are established, and problems caused by the current system design are identified. The diagnostic phase should result in a strategic alliance of stakeholders with a common motivation to design alternative ways of agricultural production. In the next phase, farmers and researchers set out to design new production systems that better meet the objectives. The result of the design phase is a number of promising theoretical prototypes of sustainable production systems. Following implementation of these theoretical prototypes, monitoring their performance in terms of objectives provides the basis for iterative prototype improvement. This phase of testing and improvement, executed on experimental or commercial farms, results in practical prototypes which have demonstrated acceptable performance in all objectives. Capitalizing upon the insights gained during the first three phases, a larger farmer audience can be addressed during the final phase, which is aimed at dissemination of the prototypes within the farming community. Prototyping suffers from two shortcomings. Only few theoretical prototypes can be tested, resulting in

a lack of information on trade-off among objectives, and systems design is based on expertise summarized in simple rules which unduly narrows the range of available options and obscures understanding of systems behaviour. Model-based explorations can remedy the limitations of empirical prototyping. Models are used because they represent devices for combining detailed information on system components and creating system designs that meet objectives of the various actors involved in farming systems development. The models are used in an explorative, as opposed to predictive, fashion. Rather than aiming at predicting which farming systems are plausible, explorations focus on designs which are possible in relation to the objectives of the actors involved. As a consequence, results of explorative studies are presented as options rather than recipes. While different techniques may be used, including continuous simulation, rule-based simulation or optimization techniques, explorative models are mechanistic and integrate components to create designs at next higher aggregation levels. The mechanistic approach enables elucidation of causes of calculated system behaviour based on insight in component behaviour and provides the opportunity to enhance understanding of systems behaviour. To optimally utilize their potential, model-based explorations should be conducted at different levels of aggregation of agricultural systems. Objectives and constraints at the various levels differ, resulting in conflicting options between levels. For instance, sustainability from a regional perspective may lead to farming systems that are economically not viable. The time horizon adopted in a study determines the type of options available, longer time horizons leading to designs with more futuristic farm strategies, alternative crops or production techniques. This paper focuses on the contribution of modelbased explorations at the level of field, crop rotation and farm to design of sustainable farming systems. Studies aimed at exploring options at regional and global levels are described by de Wit et al. (1988), Veeneklaas (1990), van de Ven (1994), Rabbinge et al. (1994) and Penning de Vries et al. (1995). In the next section, two case studies are described, with emphasis on the role of model-based explorations in different phases of farming systems development. The first case study illustrates the role of explorations dur-

341

ing diagnosis and design of wheat management systems at field and crop rotation levels, with a time horizon of 1-5 years. The second case study exemplifies model-based design at the crop rotation and farm level, with a time horizon of 10-15 years. In the discussion, methodological elements related to model-based explorations and their delivery are discussed and opportunities for enhanced support of development of sustainable fanning systems are identified.

2. Explorations at the level of field, crop rotation and farm: two case studies

2.1. Opportunities for improving wheat-based systems in the Paris Basin of France In the Paris Basin, farmers often have more than half their farm land under cereals (mainly wheat and barley), in rotation with sugar beet, potato, oilseed rape, fodder peas and sometimes vegetables for industry. High input levels of pesticides and nutrients constitute a threat to environmental sustainability of these farming systems. A series of studies were carried out at the aggregation levels of field and crop rotation, aimed at improving wheat management in the short term, i.e. with a time horizon of 1-5 years. A combination of model-based explorations and empirical research was used to diagnose current practices in wheat cultivation and design improved systems. In the explorations, regression models and rule-based simulation models were used, their nature and role differing among phases.

2.1.1. Diagnosis Causes of variation in input efficiency and yield between wheat crops grown on similar soils and as part of similar crop rotations were diagnosed. Data on actual wheat management were obtained through surveys. In each surveyed field, data were collected on crop (total biomass, yield components, crop nutritional status, crop health), soil (available macro-nutrients, structure of the arable layer) and environmental conditions (temperature, radiation and precipitation). Furthermore, information was gathered on wholefarm management, i.e. crop and soil management practices (timing of operations, labour requirement

and equipment used) for all crops in the rotation (Meynard, 1985a; Dor6 et al., 1997). In the Picardie region, the actual yield ranged from 4 to 9 t grain/ha. Variation in yield appeared to be more related to variations in ear and grain density than to variation in weight per grain (Meynard, 1985a). To identify the factors limiting each of these yield components, a yield gap analysis was carried out. Potential size of yield components were explored using a set of linked empirical regression models that related the size of a yield component in the absence of limiting or reducing factors, to critical crop variables. For instance, potential ear density was estimated by the regression model of Masle (1985) in which biomass of the shoot at the start of stem elongation and variety are used as input variables, while potential grain density was described by the regression relation of Boiffin et al. (1981) that uses shoot biomass at the start of flowering as input. Next, the yield gap, defined as the ratio of actual level and potential level of a yield component, was correlated to soil and environmental variables. Correlation and categorical analyses revealed significant association between yield gap and two factors: compacted soil structure and belated applications of nitrogen fertiliser that caused nitrogen deficiency during the first part of stem elongation (Meynard and David, 1992). The same factors decreased nitrogen fertiliser efficiency. For example, uptake of N on compacted soils was reduced by about 30-50 kg/ha compared to soils without compaction, which increased risk of leaching of nitrogen. Belated fertilizer application and soil compaction appeared correlated to aspects of whole-farm work organisation. Sowing of sugar beet was demonstrated to receive priority over simultaneously required fertilizer application in wheat. Ploughing to alleviate soil compaction, generally caused in crops preceding wheat, competed for labour with sugar beet harvesting, and was replaced by less time consuming, but also less effective, shallow tillage. The scale of these problems varied among farms depending on the relative importance of the competing crops and the labour-power and equipment available (Meynard, 1985a; Aubry, 1995). Yield reduction caused by aphids and diseases was found to be positively correlated to early sowing (before mid October), high nitrogen fertilizer input (about 200 kg/ha), high seed rates (about 300 grains/

342 m 2 for early sown crops) and the level of susceptibility of the cultivar, except when fungicides and aphicides had been applied twice (Meynard, 1991). Survey data were summarized in a multiple regression model that predicted yield loss relative to yield of a 'healthy' crop that had received two pesticide applications (Chevallier-G6rard et al., 1994). Input variables included attainable yield, i.e. crop yield in absence of pests and diseases, year, location, sowing date, varietal resistance spectrum and preceding crop. The model was used to explore the economic returns on pesticide application for different combinations of target yield, sowing dates and cultivar resistance, using weather data from 1978 to 1991. The results demonstrated that pesticide application was justified in the vast majority of years for a susceptible cultivar such as Th6s6e that was sown early with a high yield target. In contrast, positive returns on pesticide input were obtained in four out of 14 years only when sowing occurred after mid November and a more resistant variety such as Renan was selected in combination with a lower target yield.

2.1.2. Design The results of the diagnostic phase stimulated the design of low-input wheat management systems that were environmentally friendly and provided economic margins that were at least equal to those of

intensive systems. Three major constraints of prevailing systems should be overcome: soil compaction, belated fertilizer application and disease risk. Working hypothesis during design of these new systems was that by adopting a target yield below the level in conventional systems, it would pay off to sow later, to reduce sowing rates and nitrogen input, and to adopt varieties that, although lower yielding than popular varieties, were more disease resistant. As a consequence, risks of lodging and infestation by diseases and aphids were expected to be lower and costs of growth regulators and pesticides would decrease, while at the same time less labour would be required (Meynard, 1985a). A range of alternative systems was assessed by model-based explorations. In the study, the set of linked regression models that were used during the diagnostic phase was extended to account for the entire growing season. For instance, Shinosaki and Kira's model, modified by Willey and Heath (1969) and Meynard (1985b), was added to calculate aerial biomass at the start of stem elongation using plant density and sowing date as inputs. This model's output was input to calculation of ear density according to the model of Masle (1985). The balance sheet approach (R6my and H6bert, 1977) was used to calculate the effect of nitrogen fertilizer application on yield. The complete set of models is described by Meynard (1985a). Results of the study indicated that

Table 1 Comparison of characteristics of a prototypeintegrated wheat managementsystem and a conventional intensive systemfor the Picardie region in the Paris Basin of France Aspect

Integrated system

Intensive system

Target yield range (t/ha) N requirement (kg/ha) Sowing rate (grain/m2) Sowing before 25 October Sowing after 11 November Nitrogen fertilizer First application: Date Rate (kg/ha) Second application: Date Rate (kg/ha) Growth regulator Fungicides

6.5-7.5 195

8.0-9.0 240

180 250

260 490 15 February + 10 days

40

70

Beginning of stem elongation According to balance sheet method: crop N-requirement minus estimated soil supply No CCC at start of stem elongation According to damage threshold Two fixed applications: at heading and 4 weeks before

Prototype design is based on an explorative model-based study by Meynard (1985a).

343

~~

• Climatic database

• Cropping history Soil type

(20 to 30 years)

• Set of decision rules]

J

i

I

[

DECISIONSIMULATOR

I

i

Crop management (year i, i= I to30) i=l+ 1

Yield, protein content, soil mineral N at harvest, gross margin (year i)

• Frequency distribution of yield, protein content, soil mineral N at harvest, gross margin

• Risks linked to the particular set of decision rules ~l

illl

i

Fig. 1. General Day-out of D6cibi6, a software tool for interactive design of wheat management systems (after Aubry et al., 1992).

the 'integrated system' described in Table 1 would provide the largest returns. Widespread application of these design principles to the diversity of constraints of specific farmers, has been facilitated by the development of the interactive software tools 'D6cibl6' (Aubry et al., 1992; Chevallier-G6rard et al., 1994) and 'Otelo' (Attonaty et al., 1993). D6cibl6 simulates the effects of crop management on wheat yield, gross margin, protein content, and soil mineral nitrogen at harvest for specific fields characterized by cropping history, soil type and weather (Fig. 1). Crop management is described by a set of decision rules, representative for a farmer or proposed by researchers and extensionists. In the decision rules environmental and agronomic conditions are related to actions. For example, a possible rule would be: 'if the wheat crop is in development stage 30 and calculated trafficability of the soil is sufficient, then apply nitrogen dressing calculated according to the balance sheet method'. These generic decision rules are made specific for a particular crop in a particular year by a 'decision simulator'. Simulated crop management is combined with modules of crop growth and development in a 'crop simulator' which contains the agronomic relations described earlier.

These modules can easily be tested and adapted to new varieties and different areas. When run with up to 30 years of historical weather data, D6cibl6 enables exploration of different sets of decision rules, thus providing information to the farmer for selection of the set that is most desirable for his specific objectives regarding grain quality, work organization, or economic and environmental goals. While D6cibl6 is used to explore options at the crop and field level, opportunities for improvement of work organization at the crop rotation and farm level can be explored with the interactive software tool Otelo (Attonaty et al., 1993). Otelo is a rule-based system which simulates consequences of work prioritization on dates of tillage, sowing, harvest, etc. for a given farm. Otelo enables the farmer to explore different ways to reduce competition among activities and assess the possible contribution of changing machinery, manpower, or cropping plan. During exploration, the farmer specifies his decision rules in the same format as described for D6cibl6, for all crops and activities on the different fields of the farm. These decision rules are input to Otelo. The farmer then simulates the dates of the various operations using weather data from the last two or three years. The comparison between simulated and actual dates is a validation of the representation of the farmer's decision rules. Such validation determines the quality of the ensuing explorations, and increases the farmer's confidence in the model. Similar to D6cibl~, risks associated with the farmer's decision rules can be assessed by running Otelo with weather data for a period up to 30 years. Risk may be expressed as probability of exceedence of a threshold value and can be estimated from simulated frequency distributions for sowing date, lateness in fertilization, or soil compaction at sowing or harvest. Various sets of decision rules, cropping patterns or equipment can be compared to those of the farmer. Both in Otelo and in D6cibl6, the agronomist and the farmer iteratively identify the best organization pattern, integrating the characteristics of the farm. Combination of surveys and model-based explorations during diagnosis resulted in a perspective on bottlenecks in existing wheat-based systems at the level of wheat crop and crop rotation. In the design phase, models were used to assess alternative solutions with respect to objectives pertaining to eco-

344

nomic returns, environment and labour availability. Empirical evaluation of the performance of the integrated system that emerged as most promising (Table 1) will be addressed in Section 3.1.

2.2. Opportunities for improving flower bulb based farming systems in the Netherlands Current systems of flower bulb production in the Netherlands use considerable amounts of nutrients and pesticides per unit area. High prices of product and land, relatively low input prices and a defensive attitude among growers towards environmental issues are among the causes for these high input levels. Legislation is aimed at reducing negative environmental side-effects, particularly addressing pesticides and nutrients. To support design of environmentally more acceptable production systems by an association of growers and environmentalists, an explorative study was carried out (Rossing et al., 1997). In the study, fragmented agronomic information was synthesized in a database and a linear programming optimization model was used to explore technical options for flower bulb production with a time horizon of 10-15 years. The choice of time horizon was reflected in the choice of farm sizes (in terms of labour and area, both treated as exogenous variables) and in the choice of production techniques. The study focused on farms located on coarse sandy soils in the west of the Netherlands, allowing rent of land for bulb production on heavier soils further away from the farm. In accordance with the operational definition of sustainability proposed by WRR (1995), a distinction was made between value-driven objectives and factdriven agronomic information. One economic and two environmental objectives were formulated in interaction with the association of stakeholders. The economic objective was represented by maximization of farm gross margin. The environmental objectives were minimization of pesticide input expressed in kg active ingredient (a.i.) averaged over the cropped area and minimization of nitrogen surplus calculated as nitrogen not taken up by the crop and not transferred to a subsequent crop, averaged over the cropped area. Important value-driven constraints that were formulated comprised farm size, the possibility to rent additional land free of soil-borne pests and diseases, and the variety of crops that could be grown.

Agronomic information was synthesized to define management systems for four bulb crops, i.e. tulip, narcissus, hyacinth and lily, and for one break crop, i.e. winter wheat, which has positive effects on soil structure and soil health. Crop management systems were characterized by soil type and soil health, cropping frequency, crop protection regime and nutrient regime. The characteristics were chosen such that a wide array of crop production techniques could be defined that varied distinctively in terms of the objectives of flower bulb production. In addition, inter-crop management systems were defined, such as soil fumigation, inundation, and prevention of wind erosion with straw. Congruent with the time horizon of 1015 years, attention was focused on production techniques still in an experimental stage and techniques derived from other crops, rather than on current practices only. For all specified crop and inter-crop management systems inputs and outputs were formulated using empirical information, expertise and production ecological theory (Rabbinge, 1993; de Koning et al., 1995; van Ittersum and Rabbinge, 1997). Crop and inter-crop management systems were combined to rotations in a multiple goal linear pro-

x t X

60-

X

X

X

X

c.-

m 4010(

io_- 20-

x 60

°T'-"--r-. 80

100

~ 120

b 140

x 160

t

180

Nitrogen surplus (kg N ha-l)

Fig. 2. Calculated maximum farm gross margin (index, see Table 2) associated with combinations of farm-based average pesticide input (kg active ingredient/ha) and farm-based average nitrogen surplus (kg N/ha) for a farm with 15 ha sandy soils, three full time labour equivalent, and optional rent of land. Optional crops on sand: tulip, hyacinth, narcissus, lily and winter wheat; on clay: tulip and narcissus. Points of equal farm gross margin are connected (iso-lines). Each combination of pesticide input and nitrogen surplus for which maximumfarm gross margin is calculated is indicated as x. Arrows indicate development paths (see text and Tables 2 and 3). (Rossinget al., 1997;reprinted with permissionof Kluwer Academic).

345

tulip-lily-wheat to narcissus-wheat and farm gross margin becomes negative. In all steps, the rented land, free of soil-borne pests and diseases remains approximately 11 ha. On this rented area tulip is grown with a relatively moderate pesticide input of 12 kg (a.i.)/ha. In contrast, the results for the development path for nitrogen surplus reduction show that with the defined techniques, N-surplus reduction is only possible at the expense of a considerable reduction in income. A decrease in N-surplus of 30% beyond the levels anticipated for 2000 is associated with a 40% decrease in farm gross margin (Table 3). In the cropping sequence lily is replaced by narcissus, which has a much lower gross margin but higher N-efficiency. Experiences on two experimental farms and current trends in the sector support the conclusion that reducing pesticide use affects farm income less than N-surplus reduction. Remedy may be sought in development of new technologies, aiming at more precise application of nutrients in time and space, or in re-evaluation of strategic choices, such as the current use of alluvial sandy soils for growing the bulk of nutrient-inefficient flower bulb crops. The sensitivity of results to farm size, range of crops, prices and assumptions on input-out-

gramming approach (de Wit et al., 1988; Schans, 1996) to allow evaluation of objectives. By maximising farm gross margin at increasingly tighter constraints on the environmental objectives, the tradeoff between market and environment was explored. The reference situation represents a production system which just meets the (anticipated) governmental targets with respect to pesticide input and nitrogen surplus for the year 2000. Two development paths were assessed, representing gradually reduced pesticide use and N-surplus, respectively (Fig. 2). The development path for pesticide reduction (Table 2) shows that in the first step a substantial reduction in pesticide input may be achieved with relatively little loss of farm gross margin. This is achieved mainly by substituting soil fumigation by inundation and adoption of new low-dosage fungicides in tulip production. No changes in cropping sequence or area rented land occur. Further reduction in pesticide input (step 2) is most economically accomplished by abolishing the use of mineral oil for virus control in lily. The associated yield loss in lily causes a reduction in farm gross margin. Again, no changes in cropping frequency occur. The third step, zero pesticide input, causes major changes: the rotation changes from

Table 2 Exploration of flowerbulb production systems under increasingly tighter constraints on pesticide input (kg active ingredient/ha) for a farm of 15 ha sandy soils, three full time labour equivalent, and possibility to grow tulip, narcissus, lily, hyacinth and winter wheat

Objectives Farm gross margin (indexed) Average pesticide input (kg a.i./ha) Average nitrogen surplus (kg N/ha) Production techniques Fraction area per crop (%) Tulip Lily Narcissus Winter wheat Pesticide input per crop (kg a.i./ha) Tulip Lily Narcissus Winter wheat Area fumigated (ha) Area rented (ha)

Referencea

Step I b

Step 2

Step 3

100 50 140

97 30 140

77 10 140

-4 0 140

33 33

33 33

33 33

-

-

-

33

33

33

18 86

12 78

12 18

-

-

-

-

50 50 0 0

0

0

0

0

1.2 11

0 11

0 10

0 11

aThe reference farming system, equivalent to point A in Fig. 1, just meets the anticipated governmental targets regarding pesticide input and

nitrogen surplus for the year 2000. The associated farm gross margin has index value 100. Zero gross margin has index value 0. bStep 1 results in point E in Fig. 1, step 2 in point F, step 3 in point G.

346 Table 3 Exploration of flowerbulb production systemsunder increasingly tighter constraints on nitrogen surplus for a farm of 15 ha sandy soils, three full time labour equivalent, and possibility to grow tulip, narcissus, lily, hyacinth and winter wheat Referencea

Step Ib

Step 2

Step 3

100 50 140

62 50 100

38 9 90

2 8 55

33 33

33 16 16 33 11

50 50 11

50 50 12

Objectives Farm gross margin (indexed) Average pesticide input (kg a.i./ha) Average nitrogen surplus (kg N/ha)

Production techniques Fraction area per crop (%) Tulip Lily Narcissus Winter wheat Area rented (ha)

33 11

aThe reference farming systemjust meets the anticipated governmentaltargets regarding pesticide input and nitrogen surplus for the year 2000. The associated farm gross margin has index value 100. Zero gross margin has index value 0. bStep 1 results in point B in Fig. 1, step 2 in point C, step 3 in point D.

put relations is reported elsewhere (Rossing et al., 1997). The approach of separating objectives and bio-physical options was much appreciated by the association of growers and environmentalists and resulted in bridging the gap between the two parties involved. The existing polarization appeared to be caused by divergent views on objectives, rather than by disagreement on bio-technical relations. The perspective on the trade-offs among all objectives focused the discussion on preferred development pathways. While the a priori outlook of growers was especially focused on tactical decision making, the study increased awareness of the importance of strategic choices over tactical choices (Rabbinge and Zadoks, 1989). In particular, the importance of introducing a soil health restoring break crop in the rotation, such as winter wheat, and renting healthy land proved to be important strategic options for mitigating the decrease in farm gross margin associated with less pesticide input and lower nitrogen surplus. Based on the results of the study, participating farmers actively promoted research on ecology of soil-borne pests at their experimental station in response to the lack of knowledge that had become apparent during the explorations. Despite uncertainty in a number of the agronomic relations, the results were deemed sufficiently robust for testing and improvement on commercial farms. A major project was formulated and is anticipated to start in 1998. The project envisages continuous train-

ing of selected farmers and extensionists and efforts are currently focused on adapting and extending the exploratory design tools for this educational purpose.

3. Discussion

3.1. Methodological aspects of model-based explorations In the introduction of this paper, model-based explorations were put forward to supplement empirical prototyping and to remedy its shortcomings: the limited number of production systems that can be evaluated and the rules of thumb used during the design process. The case studies demonstrated the capacity of models to explore large numbers of alternative production systems and to enhance understanding of systems behaviour due to the transparency of model components. The case studies differed with respect to model types and aggregation levels. In the Dutch case study, input-output relations stored in databases that were linked to a linear programming model were used to address strategic changes in flower bulb production systems needed to resolve conflicts between economic and environmental objectives. In the French case study, a combination of regression models and rule-based simulation models were used for tactical exploration of wheat management systems aimed at adjustment of bottle-

347

necks within the constraints imposed by existing crop rotations. In the regression models agronomic information was summarized to assess potential production during diagnosis or target production levels during design. The rule-based models were used to mimic farm management decisions, both of a particular farmer and in an explorative sense. In principle, results of these tactical explorations provide inputoutput relations for strategic optimization studies. Establishing such link constitutes an important research area to be developed, as it would improve the coherence and consistency of explorative studies at different time and spatial scales. The case studies indicate that relevant answers require explorative studies at different aggregation levels. Only by combining the opportunities at the crop rotation level with those at the crop level, soil and nutrient management in wheat could be improved without sacrifices in other crops caused by labour constraints. For flower bulbs, a study at the sectoral level would be desirable to explore the implications of various options identified at the farm level, because prices are largely determined by the production volume re alised in the Netherlands. The case study in the Netherlands demonstrated the usefulness of sensitivity analysis to reveal gaps in knowledge relevant to the problem. The consequences of uncertainty in agronomic knowledge were revealed by varying single parameters or parameters in a single relation and assessing the resulting change in model output in the conventional way (cf. Janssen, 1994). However, sensitivity in linear programming models represents a special case, because, typically, uncertainty in agronomic knowledge may have little effect on the realization of objectives, but leads to very different optimal production systems (Scheele, 1992; Hijmans and van Ittersum, 1996). New mathematical techniques are needed to reveal the range of production systems that results in similar levels of satisfaction of objectives. By definition, models are simplified representations of reality, targeted at capturing the essential elements of system dynamics. To be relevant, statements based on model calculations should be accompanied by an indication of their quality. Quality assessment may address model components at the field level, such as a single regression relation or an input-output relation, and compare it to reality. Concepts for model valida-

tion or evaluation at the field level have been described by various authors (e.g. Teng, 1981; Rossing, 1991). A similar approach to quality assessment of models at the level of crop rotations or farms is less useful because the large number of uncontrollable variables impedes classical experimental design. Quality may then be interpreted as the degree to which a model-based systems design that emerged as potentially successful, performs better than an existing system. In such output-oriented assessment of model quality, causality requires attention: the model results must be better for the fight reasons. Approaches to output-oriented assessment of model quality that were developed in the French case study include (1) cropping system experiments based on decision rules, and (2) monitoring of farms that have adopted the prototype systems. In a cropping system experiment alternative wheat management systems are evaluated on-farm for their effects on a set of objectives. Each management systems consists of a specific set of decision rules that emerged as promising from the design phase and was further refined by discussion among farmers, advisers and researchers (Meynard et al., 1996). Variation in production situations among farms is taken into account by executing the experiment in a network of farms. This approach was adopted by Meynard ( 1985a, 1991) for evaluating the simulated design for integrated wheat management (Table 1). During 4 years, 28 on-farm experiments were executed in which this integrated system was compared to the conventional system. Outputs of the integrated system that were compared to the conventional system comprised mean gross margin, yield variability, and risk of nitrate leaching. The integrated system appeared better than the conventional system from both the economic and environmental point of view (Table 4). An important spin-off of farmers being responsible for execution of the experiments in their fields, was the increased credibility of the results to the farming community. In a survey of farms that had adopted integrated wheat production systems, Aubry (1995) showed that farmers modified the underlying decision rules to simplify their decision-making tasks. Farmers classified their large number of different wheat fields in groups which could be treated in similar ways and monitored crop development and diseases in only one field. Further improvement of the relevance of

348

Table 4 Performance of a prototype integrated wheat management system and a conventional intensive system in 28 on-farm experiments during 4 years, on different soil types and with different previous crops in the Picardie region in the Paris Basin of France Aspect Actual yield Mean (t/ha) Highest mean yield (cases out of 28 experiments) Gross margin Highest mean gross margin (cases out of 28 experiments) Standard deviation of yield (t/ha) N-recovery Highest mean N-recovery (cases out of 28 experiments) Fungicide treatment a Average number

Integrated system

7.5 4

Intensive system

8.0 24

21 1.10

7 1.14

22

6

1.4

2.0

Agronomic details of the systems are described in Table !. Soil tillage, sowing date and herbicide application were farmer-specific. In the systems, the same variety was used. aln additional experiments with a more disease resistant variety the average number of fungicide treatments was 0.7.

systems design may be expected when during design the set of fields on the farm sown to a given crop, i.e. a decision-making level intermediate between the crop and the crop rotation, is taken into account. 3.2. Interaction with stakeholders

As was argued in the introduction to this paper, development of sustainable farming systems implies negotiations about change and has an important social dimension. Contributions by model-based explorations to this social aspect can be assessed in terms of 'product', 'process' and 'instrument'. The contribution of explorative studies may be assessed in terms of their envisaged product, i.e. change in perceptions and/or change in actions of actors in the agricultural knowledge chain. Such impact assessment may help to improve design and delivery of explorative studies, but few retrospective studies of this sort have been carried out and reported (Sebillotte, 1996). The funding body informally evaluated the case study on flower bulb systems by asking a journalist to interview the participants. Counterintuitive results of the study were reported with respect to the importance of introducing a soil health restoring break crop in the rotation, and with respect to renting healthy land. In contrast to these strategic options, the a priori attention of the participants had been focused on improving management of pests, diseases and nutrients. The study prompted a range of

activities by the association of farmers and environmentalists. After completion of the study, the association continued discussions, involved other parties from the flower bulb industry, and formulated a proposal for testing and improvement of prototype systems of integrated flower bulb farming that was widely supported. Apart from its outcome, the process, i.e. the execution of an explorative study in itself may contribute to development of sustainable farming systems by stimulating discussions among stakeholders based on scientific information. In both case studies, participants considered communication and reflection on sustainability to be improved as a result of the explorative approach in diagnosis and design. Essential elements in this respect are the clear separation of value-driven objectives and fact-driven options, the quantitative nature of model results that enabled discussion on acceptable trade-offs among objectives, and the transparency of underlying information which improved understanding of system behaviour. Effective contribution to the process of design by stakeholders necessitates research planning in which communication between researchers and stakeholders is explicitly taken into account. For instance, during the flower bulb case study researchers interacted with a delegation of the association once every 6 weeks to discuss general progress, and the association organized two workshops to formulate the value-driven objectives that were used to evaluate options during

349

the study. These frequent social interactions may well have paved the way for the projects 'product' described in the previous paragraph. The instruments, i.e. the models that were used during the explorations served as computer-aided learning tools for stakeholders (Leeuwis, 1993; Cerf et al., 1994; Papy, 1996). Characteristically, models were used to answer 'what-if' questions, with researchers acting as intermediates between model and stakeholders (in the flower bulb case) or process facilitators (in the wheat case). Fast response to 'whatif' questions is necessary to maintain process momentum, requiring further improvement of the linear programming model concerning user-friendliness, flexibility and methodology for sensitivity analysis. While in the flower bulb case the model itself was meant t o b e used by researchers only, the tools Otelo and D6cibl6 were designed for interactive application by farmers and extensionists to a specific farm. Two aspects were found to influence the success of the application (Mousset et al., 1997). Firstly, the users should be able to gain confidence in the tool. Farmers were usually able to recall their agronomic decisions of the preceding three years, which enabled assessment of model quality by comparison of simulation results with actual data. Secondly, the amount of time spent in creating input for the model should be commensurate to the expected value of information of model output. Especially for Otelo, the amount of farm-specific input data is considerable, and the tool is used only for complex situations, requiring innovative solutions.

3.3. Model-based explorations and empirical prototyping The case studies in this paper demonstrated how model-based explorations supplement empirical prototyping during the first two phases of sustainable farming systems development: diagnosis and design. The diagnostic surveys of the prototyping approach were complemented by modeling studies exploring production potential. Their combination enabled identification of constraints in current practices and assessment of opportunities for improvement. Opportunities were elaborated during the model-based design phase to reveal trade-offs among objectives. Although not substantiated by the case studies, sup-

plementation of prototyping by model-based explorations may also be expected during the last two phases of development of sustainable farming systems" testing and improvement, and dissemination. Promising options emerging from the design phase are put to the test empirically. During testing and improvement of these prototypes, explorations can reveal yield gaps or trends in slow processes such as soil organic matter turn-over. At the start of dissemination, empirical prototyping has resulted in prototype systems that have proven their value in practice, while essential elements of production systems are synthesized in models at different levels of aggregation to facilitate extrapolation to new conditions. During all phases model-based explorations may indicate gaps in knowledge of researchers, extensionists and farmers, thus contributing to learning about the system by all actors. The pivotal role of learning appears prominently in farmer-oriented projects that have been successful in stimulating more judicious use of resources (Zadoks, 1989; Kenmore, 1991; Sebillotte, 1996). Two important characteristics of learning are (1) cyclic iteration of experimentation, action, observation and reflection, and (2) repeated switching between aggregation levels, time periods, knowledge domains and farm types. Model-based explorations in interplay with prototyping have the potential of contributing to such non-formal education. To realize this potential requires a research approach in which equal attention is devoted to creating and synthesizing relevant agronomic knowledge and to creating settings in which learning can take place (Chatelin et al., 1994; Okali et al., 1994; R61ing, 1996; Somers, 1997).

References

Attonaty, J.M., Chatelin, M.H.,Poussin, J.C. and Soler, L.G., 1993. Advice and decision support systems in agriculture: new issues. Farm Level Information Systems, Woudschoten, Zeist, The Netherlands. Aubry, C., 1995. Gestion de la solel d'une culture dans l'exploitation agricole. Cas du BW d'Hiver en Grande Culture dans la R6gion Picarde. Doctoral Thesis, Institut National Agronomique Paris-Grignon, 277 pp. Aubry, C., Chatelin, M.H., Poussin, J.C., Attonaty,J.M., Meynard, J.M., G6rard, C. and Robert, D., 1992. D6cibl6: a decision support systemfor wheat management.In: Book of Abstracts,4~.me Congr~s d'lnformatique agricole, Versailles, SAF, Paris.

350

Boiffin, J., Caneill, J., Meynard, J.M. and Sebillotte, M., 1981. Elaboration du rendement et fertilisation azot6e du bl6 d'hiver en Champagne Crayeuse. I. Protocole et mdthode d'6tude d'un probl6me technique r6gional. Agronomie, I: 549-558. Cerf, M., Papy, F., Aubry, C. and Meynard, J.M., 1994. Agronomic theory and decision tools. In: J. Brossier, L. de Bonreval and E. Landais (Editors), Systems Studies in Agriculture and Rural Development. INRA, Paris, pp. 343-356. Chatelin, M.H., Mousset, J. and Papy, F., 1994. In: B.H. Jacobsen, D.E. Pedersen, J. Christensen and S. Rasmunsen (Editors), Farmers' Decision-Making - A Description Approach. Proc. from the 38th EAAE seminar, Institute of Agricultural Economics and the Royal Veterinary and Agricultural University, Copenhagen, pp. 369-381. Chevallier-G6rard, C., Denis J.B. and Meynard J.M., 1994. Perte de rendement due aux maladies cryptogamiques sur b16 tendre d'hiver. Construction et validation d'un mod61e de l'effet du syst6me de culture. Agronomie, 14: 305-318. de Koning, G.H.J., van Keulen, H., Rabbinge, R. and Janssen, H., 1995. Determination of input and output coefficients of cropping systems in the European community. Agric. Syst., 48: 485502. de Wit, C.T., van Keulen, H., Seligman, N.G. and Spharim, I., 1988. Application of interactive multiple goal programming techniques for analysis and planning of regional agricultural development. Agric. Syst., 26:211-230. Dor6, T., Sebillotte, M. and Meynard, J.M., 1997. A diagnostic method for assessing regional variations in crop yield. Agric. Syst., 54: 169-188. Hijmans, R.J. and van lttersum, M.K., 1996. Aggregation of spatial units in linear programming models to explore land use options. Neth. J. Agric. Sci., 44: 145-162. Janssen, P.H.M., 1994. Assessing sensitivities and uncertainties in models: a critical evaluation. In: J. Grasman and G. van Straten (Editors), Predictability and Non-Linear Modelling in Natural Sciences and Economics. Kluwer Academic, Dordrecht, pp. 344-361. Kenmore, P.E., 1991. How rice farmers clean up the environment, conserve biodiversity, raise more food, make higher profits Indonesia's IPM - a model for Asia. FAO Inter-Country Programme for Integrated Pest Control in Rice in South and Southeast Asia, Manila, Philippines, 56 pp. Leeuwis, C., 1993. Of computer, myths and modelling. The social construction of diversity, knowledge, information and communication technologies in Dutch horticulture and agricultural extension. Wageningen Studies in Sociology, Pudoc, Wageningen, Vol. 36, 468 pp. Masle, J., 1985. Competition among tillers in winter wheat: consequences for growth and development of the crop. In: W. Day and R.K. Atkin (Editors), Wheat Growth and Modelling. Plenum Press, New York, pp. 33-54. Meynard, J.M., 1985a. Construction d'itin6raires techniques pour la conduite du b16 d'hiver. Doctoral Thesis, Institut National Agronomique Paris-Grignon. Paris, 258 pp. Meynard, J.M., 1985b. Les besions en azote du b16 d'hiver jusqu'au d6but de la montaison. Agronomie, 5: 579-589. Meynard, J.M., 1991. Pesticides et itin6raires techniques. In: P.

Bye, C. Descoins and A. Deshayes (Editors), Phytosanitaires, Protection des Plantes, Biopesticides. INRA, Paris, pp. 85100. Meynard, J.M. and David, G., 1992. Diagnostic sur l'61aboration du rendement des cultures. Cahiers Agric., 1: 9-19. Meynard, J.M., Reau, R., Robert, D. and Saulas, P., 1996. Evaluation exp6rimentale des itin6raires techniques. In: G. Urbano et al. (Editors), Experimenter sur les Conduites de Culture. DERF! ACTA, Paris, pp. 63-72. Mousset, J., Aslah6, C., Billa, Ph., Boiffin, J., Chatelin, M.H., Chopplet, M., Franqois, M., Gandon, H., Groell, F., His, M., Hopquin, J.P., Klein, D., Papy, F., Qui6vreux, D. and Soler, L.G., 1997. Le conseil Agro-Equipement en Picardie: MECAGRO. Colloque 'Aide h la d6cision et choix de strat6gies dans les entreprises agricoles', Laon (France), 10-11/12/1996, INRA, pp. 195-206. Okali, C., Sumberg, J. and Farrington, J., 1994. Farmer Participatory Research. Rhethoric and Reality. Intermediate Technology Publications, London, 159 pp. Papy, F., 1996. Modelling the farm as a business to help farmers make their decisions: a summary review. In: J.P. Colin, E. Crawford and C. Fillonneau (Editors), Research Methodology for Agricultural Systems Analysis. In press. Penning de Vries, F.W.T., van Keulen, H. and Rabbinge, R., 1995. Natural resources and limits to food production in 2040. In: J. Bouma, A. Kuyvenhoven, B.A.M. Bouman and J.C. Luyten (Editors), Eco-Regional Approaches for Sustainable Land Use and Food Production. Kluwer Academic, Dordrecht, The Netherlands, pp. 65-87. Rabbinge, R., 1993. The ecological background of food production. In: D.J. Chadwick and J. Marsh (Editors), Crop Protection and Sustainable Agriculture. Ciba Foundation Symposium 177, Wiley, Chichester, pp. 2-29. Rabbinge, R., van Diepen, C.A., Dijsselbloem, J., de Koning, G.J.H., van Latesteijn, H.C., Woltjer, E. and van Zijl, J., 1994. 'Ground for choices': A scenario study on perspectives for rural areas in the European community. In: L.O. Fresco, L. Stroosnijder, J. Bouma and H. van Keulen (Editors), The Future of the Land: Mobilising and Integrating Knowledge for Land Use Options. Wiley, New York, pp. 95-121. Rabbinge, R. and Zadoks, J.C., 1989. Disease and pest control. In: J.C. Zadoks (Editor), Development of Farming Systems. Pudoc, Wageningen, pp. 32-39. R6my, J.C. and H~bert, J., 1977. Le devenir des engrais azot~s dans le sol. Comptes Rendus de l'Academie d'Agriculture de France, 63: 700-710. R61ing, N.G., 1996. Towards an interactive agricultural science. Eur. J. Agric. Educ. Ext., 2: 35-48. Rossing W.A.H., 1991. Simulation of damage in winter wheat caused by the grain aphid Sitobion avenae. 3. Calculation of damage at various attainable yield levels. Neth. J. Plant Pathol., 97: 87-103. Rossing, W.A.H., Jansma, J.E., de Ruijter, F.J. and Schans, J., 1997. Operationalizing sustainability: exploring options for environmentally friendly flower bulb production systems. Eur. J. Plant Pathol., 103: 217-234. Schans, J., 1996. MGOPT_CROP, a Multiple Goal Linear Pro-

351 gramming Model for Optimisation of Crop Rotations, version 1.0. AB-DLO, Wageningen, Report 69, 48 pp. Scheele, D., 1992. Formulation and characteristics of GOAL. Technical Working Document (W64), Netherlands Scientific Council for Government Policy, The Hague, The Netherlands, 64 pp. Sebillotte, M., 1996. Les Mondes de l'Agriculture: Une Recherche pour Demain. INRA, Pads, 258 pp. Somers, B.M., 1997. Learning for sustainable agriculture. In: M.K. van Ittersum and S.C. van de Geijn (Editors), Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use. Elsevier Science, Amsterdam, The Netherlands, pp. 353-359. Teng, P.S., 1981. Validation of computer models of plant disease epidemics: a review of philosophy and methodology. Z. Pflanzenkr. Pflanzenschutz, 88: 49-63. van de Ven, G.W.J., 1994. A mathematical approach for comparison of environmental and economic goals in dairy farming at the regional scale. In: L.'t Mannetje and J. Frame (Editors), Grassland and Society. Proceedings of the 15th General Meeting of the European Grassland Federation (EGF), Wageningen, 6-10 Juni 1994, Wageningen Press, Wageningen, pp. 453457. van Ittersum, M.K. and Rabbinge, R., 1997. Concepts in production

ecology for analysis and quantification of agricultural input-output combinations. Field Crops Res., 52: 197-208. Veeneklaas, F.R., 1990. Dovetailing technical and economic analysis. Doctoral Thesis, Erasmus University Rotterdam, The Netherlands, 159 pp. Vereijken, P., 1994. Designing prototypes. Progress report 1. Progress reports of research network on integrated and ecological arable farming systems for EU and associated countries. ABDLO, Wageningen, 87 pp. Vereijken, P., 1997. A methodical way of prototyping integrated and ecological arable farming systems (I/EAFS) in interaction with pilot farms. Eur. J. Agr. 7. 235-250. Willey, R.W. and Heath, S.B., 1969. The quantitative relationships between plant population and crop yield. Adv. Agron., 21: 281331. WRR, 1995. Sustained risks, a lasting phenomenon. Reports to the Government No. 44, Netherlands Scientific Council for Government Policy, The Hague, The Netherlands, 205 pp. Zadoks, J.C., 1989. EPIPRE, a computer-based decision support system for pest and disease control in wheat: Its development and implementation in Europe. In: K.J. Leonard and W.E. Fry (Editors), Plant Disease Epidemiology, Volume II. Macmillan, New York, pp. 3-29.

© 1997 Elsevier Science B. V. All rights reserved Perspectives for Agronomy - Adopting Ecological Principles and Managing Resource Use M.K. van Ittersum and S.C. van de Geijn (Editors)

353

Learning for sustainable agriculture B.M. Somers * Hoofdweg 3, 3233 LH Oostvoorne, The Netherlands Abstract

Policy makers are searching for ways of accelerating the rate of adoption of sustainable agriculture. However, sustainable agriculture is a complex innovation and it is still unclear what is the best way to stimulate farmers to make the change. Compared to other types of innovations, the change to sustainable agriculture imposes various risks on farmers. These risks, not only of an economic and technical nature, also have social and political aspects. In this article it is argued that the characteristics of sustainable farming systems, and the risks involved, imply a gradual learning process instead of a linear adoption process. Through adaptive learning, farmers can lower their levels of perceived risks. The possibility to try out modules of new farming systems and to combine general, scientific knowledge with specific, tacit knowledge form positive incentives for the introduction of sustainable farming systems. There is also a need for a more interactive development of such farming systems. Even more than is usual, researchers should be willing to learn from practical knowledge and should develop a feeling for the policy measures and risks under which farmers are working. Keywords: Adapative learning; Development of sustainable farming systems; Risk perception

1. Introduction

Designing sustainable farming systems is one thing, farmers practising sustainable agriculture on a large scale is another. In a scientific community where researchers have developed specialised fields of knowledge and skills it would be normal to leave the development of new methods to researchers and the introduction of it among farmers to extensionists. This route is a linear process in which sequential stages, from design to diffusion, must be completed. However, Rossing et al. (1995) point to the importance of interaction between several stages in the process of development and introduction of sustainable farming systems. They make a distinction between the following stages: a) designing (prototype); b) testing (including improvements); c) implementation on small scale; and d) implementation on large * Tel: +31 0181 484634; Fax: +31 0181 484071. E-mail: [email protected]

scale. According to Rossing et al. the large-scale implementation is furthered when stages b) and c) are closely intertwined. During these stages a close cooperation between research, extension and a group of well-motivated farmers is important. An efficient introduction of new farming systems on a large scale can take place when region-specific, practically tested knowledge is available in the agricultural community, and when this community is motivated and familiar with (elements of) the new prototype. Therefore, co-operation between different participants is a necessary condition for the diffusion of knowledge about sustainable farming systems. The innovation project "Integrated Arable Farming" (Anonymous, 1994) is a good example of the interaction mentioned above. Also other initiatives in the Netherlands (Somers and R61ing, 1993; Somers, 1995; Van Weperen, 1994) and outside the Netherlands (R61ing and Van de Fliert, 1994; Pretty, 1994) benefit from a close co-operation between farmers and organizations for research and extension. Yet,

354

despite the co-operation between research, extension and a group of well-motivated farmers, the large scale implementation of sustainable arable farming is hampered. When the actual performances are measured in terms of governmental norms about the use of chemicals, it is clear that much still has to be done. In this contribution I will elaborate on some factors that influence farmers' acceptance of sustainable methods and technologies. Usually knowledge, or a lack of knowledge, is mentioned as an important factor in the diffusion of innovations. Knowledge apparently reduces farmers' feelings of risks that are connected with new concepts. Therefore, I will start with recounting some characteristics of innovations that influence their acceptance by farmers. These characteristics refer to farmers' perception of risks. Subsequently I describe the possibilities of adaptive learning and the types of knowledge that are relevant in the case of sustainable farming systems. These specific characteristics of knowledge require changing roles of the actors involved in the development and introduction of alternative farming systems.

2. Types of risks involved with the shift to sustainable farming systems Sustainable farming systems can be conceived of as a set of novel technologies and farming methods that together form a complex innovation. What interests extensionists and policy makers is how to make this innovation acceptable to farmers. In extension science there is a vast amount of literature about the willingness of farmers to adopt new methods and technologies. Rogers (1983) made a compilation of the previous known research into this topic and created a checklist of factors that influence the process of diffusion of innovations. This checklist formed the basis for the studies of Somers and R6ling (1993) and of a workgroup of the Dutch Ministry of Agriculture (Somers, 1996a). In both studies, Rogers' checklist proved to be valuable for understanding the types of risks farmers encounter when they are confronted with complex innovations. Moreover, we became aware of the fact that the shift to sustainable agriculture meant that new requirements are imposed on farmers' pro-

fessional goals and skills. The complexity of the innovation and the diverging, and in short term conflicting goals, are both characteristic features of sustainable agriculture. Farmers are required to find a balance between economic and ecological goals instead of focusing on productivity levels. Besides, the public increasingly appeals to farmers where the care for landscape, tourism and nature is concerned. The implementation of sustainable farming systems is not just "another" innovation to be accepted, but reflects new professional goals in a context of new functions of the agricultural sector. In the studies of Somers and R61ing (1993) and the working group of the Ministry of Agriculture (Somers, 1996a), the following characteristics of innovations were found to influence their acceptance by farmers: The degree of certainty for the farmer that the innovation is advantageous for him: his trust that the innovation has a positive effect on his goals; The technical complexity of the innovation which can have repercussions on the economic performance of the farm; The degree to which the innovation is compatible with existing norms and values: the degree to which the farmer is confronted with social complexity; The possibility to try out the innovation: to try out a module of the total package, to try out the innovation on a part of the farm or to make a choice in the level of input; The effects of the innovation on current farm management: the degree to which the innovation alters the direction of farm management. Most of these factors have to do with the feeling of risk that hampers the acceptance of the innovation by farmers. Very often, risks perceived by farmers are connected to technical risks and negative economic consequences. Yet, we must be aware that the introduction of some innovations is also surrounded by social and political risks. This is well illustrated in the "Innovation Project Integrated Arable Farming" in the Netherlands. Thirty-four of the thirty-eight arable farmers who were involved in the stage of testing and improving the prototypes of sustainable

355

farming systems were questioned in 1994 about their experiences (Van Weperen, 1995). These thirty-four farmers mentioned the following three risk-items that caused most tension during their change to integrated agriculture" economic risks; risks connected to governmental policy; and risks in the social relationship with colleagues. The experience of risks has much to do with the fear of making wrong technical decisions which may adversely affect the financial situation. Governmental pressure of laws and regulations also cause tension, argued several of the interviewed farmers. Too strong pressure or sudden change of governmental regulations could cause a diminishing willingness among farmers to co-operate with the change towards a more sustainable agriculture. There comes a time when farmers think governmental regulations not attainable, particularly when the economic prospects are gloomy because of low prices for arable products. The interviewed farmers also mentioned the fear that the government would use the achievements of the project to oblige all Dutch arable farmers to attain extreme norms for chemical and nutrient input. Other studies also point to this sociopolitical tension that hampers the shift to sustainable agriculture (Van Weperen, 1994; Buurma, 1996). Moreover, the farmers who were farming in an integrated or organic way experienced the distrust of neighbours and colleagues, distrust that was aroused for instance by their higher tolerance for weeds and their diverging strategies for combating pests and diseases. In fact, the integrated and organic farmers were challenging the norms about "good-farming practice". Consequently, as will be shown later in this paper, new norms have to be developed. This process can take place in a group situation, for instance study groups. Thus, there are many different factors that influence the process of diffusion among farmers. Some of these factors point to the economic prospects and technical risks inherent in the innovation. Extensionists and policy makers are searching for ways of accelerating the rate of adoption of sustainable agriculture. The question is how this can be done, since farmers perceive risks in so many areas. There is the economic risk of setbacks in yields and inappropriate marketing strategies. There are unsolved problems in pest and disease management. About many aspects

there is still a lack of knowledge. Moreover, the actual and perceived risks are not confined to economic and technological aspects" there are also social and political risks. In the political arena the seriousness and sources of the environmental problems are constantly contested. The same holds true for political measures, norms and penalties to fight the problems. Forerunners in sustainable agriculture risk the distrust of colleagues who fear that their evidence of newly achieved low input agriculture will be used to create new political norms. The social and political risks that are inherent in a change to sustainable agriculture form strong impediments to its introduction.

3. Gaining knowledge Where innovations are perceived by farmers as risky, gaining knowledge can lower the level of perceived risks. The economist Bayes pointed to the importance of learning in the process of adopting an innovation (Leathers and Smale, 1991; Lindner and Gibbs, 1990). The so-called Bayesian learning model is an adaptive learning model. Crucial in it is the notion that the perceptions of farmers will change when they gain experience with a part of the innovation. This means that farmers must have the possibility to try out one or more modules of the system, to try out the new methods on a part of their farm, or have a choice concerning the level of input. By trying out, the farmer gains extra information with which he can adapt his original perceptions about the risks of applying the innovation. He also discovers whether the information given by researchers or extensionists is relevant for his own farm situation. The notion of reducing perceived risks through adaptive learning coincides with the ideas of Kolb (1984) regarding "learning by doing". His central idea is the interaction between cognitive processes and action. Kolb's theory is incorporated into theories about "learning organizations" that have become very popular in certain circles of management science. According to these management theories, organizations can improve their utilization of the knowledge that is available in the organization at all levels. Scientists in the field of extension can

356

also learn from these theories and support bottomup interactive learning processes as being more effective than top-down linear adoption processes. By shifting the perspective from a (top-down) adoption process to a (bottom-up) learning process, the roles of the participants also change. Applying this notion to the agricultural knowledge network, we will find that roles, attitudes and skills of the people involved change when learning processes and supporting learning processes become the foundation instead of innovations being imposed on farmers "from above".

4. Types of knowledge 4.1. General and tacit knowledge Through gaining knowledge, feelings of risks can be diminished, yet what types of knowledge can we think of? We can make a distinction between two types of knowledge: general knowledge and tacit knowledge. General knowledge is in principle accessible, though not always gratuitous for everybody (Somers, 1996b). General knowledge can be developed in scientific experiments, apart from the specific situations in which the knowledge will be applied. It can be incorporated in standard advice, for instance about the levels of fertilization or time schedules for spraying fungicides. General knowledge consists of facts and figures and it is transferable, orally or through written material. Tacit knowledge, on the contrary, can only be developed in specific enterprises or situations. Tacit knowledge has much to do with "knowing how" and very often is hard to verbalize. It is not readily transferable because it is formed by experience through the years. Tacit knowledge is stored in the heads of people, therefore it is often referred to as "human capital". This type of knowledge is only transferable by learning. Researchers and extensionists can teach the farmer much about indicators to determine critical levels and the nutritional needs of plants. These indicators and critical levels are usually developed on research stations and they are a form of general knowledge that the farmers can rely on. Yet, because of the varying circumstances on farms and, consequently the qualities of the specific knowledge needed, farm-

ers cannot depend solely on knowledge that is developed elsewhere. They not only need knowledge about indicators, but also the experience to observe and interprete these indicators in specific farming circumstances. Many of these interpretative skills are not transferable to farmers: much of it depends on tacit knowledge of the farmer about the precise qualities of his fields and his experiences with the growth of his crops in the past. Much of the knowledge that is required for sustainable agriculture is thus not only very specific, but also internal to the farm. The shift to sustainable agriculture implies that the farmer tries to replace chemical input by knowledge as much as possible. One field of knowledge consists of combating pests, herbs and insects. The farmer must have detailed knowledge and skills in order to observe and judge the threat of pests and diseases. He must know how to take adequate measures in advance in order to prevent the outbreak of pests and diseases. He needs figures in order to balance the pros and contras of different chemicals. In order to diminish his use of pesticides, herbicides and insecticides, the farmer will not follow general spraying schemes, but determine himself the time of spraying, based on the combination of general knowledge about critical levels and his past experiences with combating pests/herbs/insects in this specific crop on his own farm. In short, the farmer combines general and tacit knowledge. Another area of knowledge is the "fine-tuning" of minerals to match the needs of plants. In order to diminish mineral surpluses, the farmer will not follow standard advice about fertilization. Instead, he tries to adapt the administering of nutrients to the plants' needs by combining scientifically developed criteria and his own, farm-internal knowledge. 4.2. Partial and holistic knowledge Applying more sustainable methods has consequences for the whole farm management, both for the short and the long term. Organic farmers are aquainted with the fact that the decisions they take during raising a crop influences the successes and failures of subsequent crops. In an agriculture that depends largely on chemical input this relation is less decisive, because farmers can intervene during the

357

growing season. In integrated arable farming, where the aim is to become less dependent on chemicals, the farmer will prevent or postpone chemical interventions. This requires a special type of knowledge. Kropff (1996) elaborates on this different type of knowledge, for the case of weed control. Instead of short-term, partial knowledge, he focuses on knowledge about the functioning of crop-weed-systems, whereby strategic or long-term aspects must be given more emphasis than usual. In fact, Kropff is discussing holistic knowledge, about "managing coo-systems" instead of short-term combating weeds. However, much of the knowledge that is developed on research stations provides solutions for specific problems. Yet, the farmer who changes towards sustainable agriculture must have the skills to interpret this partial knowledge and estimate their long-term effects. Again, we see that a combination of types of knowledge is needed.

4.3. Farmers' experiments Above we discussed the different types of knowledge that are needed for a succesful turnover to sustainable arable farming. The availability of scientific knowledge is not sufficient in itself to induce a diminishing dependance on chemicals. The feeling of risk is not diminished by scientific knowledge alone. Farmers must develop the skills to connect the scientific knowledge with their own tacit knowledge about the specific characteristics of their land. In several projects in the Netherlands ~ we observe that farmers are experimenting with modules of integrated farming systems, such as trial methods for identifying the nutritional needs of plants and the availability of nutrients in the soil (Somers and R6l The four projects that were evaluated by Somers and R61ing (1993) were: a) "Integrated Arable Farming", a project organized by the Ministry of Agriculture, the Landbouwschap (the apex organization of the three farmers' organizations), formal research and extension institutions; b) "Agriculture and Environment", a project of the Province of Brabant, the Landbouwschap, the local farmers' organization and the local extension institute; c) "Farmers' Bread", a farmers' initiative in the Province of Zeeland, the local farmers' organization, environmental and consumer organizations, millers and bakers; d) "Organic Farming", counting 440 farmers who are organized in an ecological and a biological-dynamic organization and supported by a separate knowledge infrastructure.

ling, 1993). Experimenting with these methods creates confidence in the possibility to farm in a more sustainable way. It moreover creates the willingness to try out other parts of the system. Another observation is that factors other than just knowledge determine the application of sustainable technologies and methods, such as the availability of labour, the desires of the farmer and social and political circumstances (ibid, 1993). It is not only knowledge that counts, but also the situations in which the learning process takes place. This brings us to the changing roles of researchers when they support farmers' learning processes.

5. Towards an interactive development of sustainable farming systems

The introduction of a complex innovation such as sustainable farming systems requires that we have attention for the type of knowledge that is needed, the form in which the knowledge is delivered and the structure in which the knowledge development takes place. The furthering of sustainable agriculture seems to benefit from an intensive interaction between scientific knowledge that has been developed on experimental research stations and experimental knowledge that is developed on farms (Somers and R61ing, 1993). The fact that sustainable farming methods are perceived as risky, requires that standards which are developed elsewhere are tried out locally. On-farm research should get a more important place than it has at the present time. Moreover, the study of Somers and R61ing showed that farmers do not always ask for ready solutions, but also for suggestions and ideas that they can test in their specific farm situation. Groupwork is important, especially where it concerns careful thought about registrations with new methods and the determination of socially acceptable behaviour. New norms about "good"-farming practice can spring from a group process (Van Weperen, 1994). We can speak of "social" learning. Increasingly, social learning takes place in so-called "environmental co-operatives". Environmental co-operatives are local groups of farmers who search for ways to realize environmental goals that are specific for their own locality and for their type of farming. Often, they hold on-farm

358 experiments in order to gain knowledge about environment, nature and landscape. Their aims vary from achieving measurable values of nature to minimizing the input of fungicides. One of these groups is the working group "Soil Based Horticulture Under Glass", on which I will elaborate here. Soil-based horticulture under glass is especially threatened by policy measures such as the requirement to recirculate the drainage water (Somers, 1995). Technical solutions for this problem have not been found until now. Furthermore, over the years the agricultural research for soil-based horticulture has been minimized in favour of horticulture on artificial substratum. The underlying argument for this is the expectation that it is easier to control the input of nutrients and the dose of water in horticulture on artificial substratum. In other words: horticulture on artificial substratum was expected to be less environmentally harmful than soil-based horticulture under glass. Because of this choice, soil-based growers feel that they have lost room for manoeuvring. They think it important to quickly search for solutions and help develop sustainable methods and technologies under "practical circumstances". By means of experiments they started, together with researchers, to search for environmental parameters that are specific for their situation. They think it very urgent to gain knowledge on practical methods and technologies because strict policy measures are already enacted. In some cases they have requested delay of measures until their experiments will yield results. The experience of the working group "Soil Based Horticulture Under Glass" shows that researchers and growers both participate in a learning process. Growers learned to formulate their problems and wishes for research experiments in a very detailed manner. They also established a good relationship with the experimental research station in order to influence the research programmes. For the participating researchers, supporting learning processes implied that they had open minds for forms of knowledge other than just scientific. In general, an important researchers' skill is that he/she can learn systematically from the experiences of farmers and growers. Part of the research actually takes place on the firms of the horticulturists. Researchers also must develop a feeling for the policy restrictions and

risks with which agricultural entrepreneurs must deal. One of the conclusions of the above mentioned working group is that the close co-operation between growers and researchers altered their relationship and roles. The working group "Soil Based Horticulture Under Glass" gives a good example of an interactive development of sustainable farming methods. Yet, in general there are many hindrances for an approach like this. Many impediments to a better support of learning processes can be found on the institutional level (Somers, 1996a). For instance, the way the research institutes are financed cause a lack of flexibility in the research programmes. Also the resistance to cultural changes is found to be an impediment. The ability to support farmers in learning is not only a question of skills and training, but also of self entrepreneurship. An active learning attitude and the willingness to cope with uncertainties form a part of the profile of the modern farmer. Teachers, researchers and extensionists also must change their attitude in order to adequately support farmers' learning processes. However, in general this change meets much resistance (ibid, 1996a).

6. Conclusions

Both the projects that were studied by Somers and Rrling (1993) and the working group "Soil-Based Horticulture Under Glass" show that farmers are willing to contribute to a more sustainable agriculture when the necessary conditions are created that facilitate learning processes. Also other experiments, such as the growing amount of environmental cooperatives are illustrative. Some conclusions we can draw from these experiences are that: a) The introduction of sustainable farming systems is encouraged when the systems bear the possibility of learning by doing. The introduction of a total concept, a system, seems unattainable for a large group of farmers. Following the Bayesian theory about adaptive learning, the introduction of sustainable farming systems will benefit from a modular composition. A modular composition not only lowers the perceived risks, but also is attainable in a context of social and political tensions, b) Due to the social and political impediments, the introduction of sustainable farming

359 systems will benefit from "social" learning: groups of farmers setting their goals and finding ways to realize these together, c) A greater interaction between researchers, extensionists and farmers is needed for taking into account valuable practical experiences of farmers. This requires a more systematic apprehension by researchers of farmers' experiences.

References Anonymous, 1994. Telen met perspectief: Teeltstrategiei~n gericht op een duurzame akkerbouw. IKC/AGV/DLV, Lelystad. Buurma, J.S., 1996. Oorzaken van verschillen in middelenverbruik tussen bedrijven: vuurbestrijding in tulpen. Landbouw-Economisch lnstituut (LEI-DLO), Den Haag. Kolb, D.A., 1984. Experiential Learning. Prentice Hall Inc., Englewood Cliffs, New Jersey. Kropff, M.J., 1996. Strategisch balanceren: onkruidkunde als toegepaste plantenecologie. Inaugurele rede. Landbouwuniversiteit, Wageningen. Leathers, H.D. and Smale, M., 1991. A Baysian Approach to Explaining Sequential Adoption of Components of a Technological Package. Am. J. Agric. Econ., 73(3), 734-742. Lindner, R. and Gibbs, M., 1990. A test of Bayesian Learning from Farmer Trials of New Wheat Varieties. Aust. J. Agric. Econ., 34(1), 21-38.

Pretty, J., 1994. Alternative Systems of Inquiry for Sustainable Agriculture. IDS Bull., Vol.25(2), 37-49. Rogers, E.M., 1983. Diffusion of Innovations (third edition). The Free Press, New York. R61ing, N. and Van de Fliert, E., 1994. Transforming extension for sustainable agriculture: the case of Integrated Pest Management in rice in Indonesia. Agric. and Human Values, Vol 11 (2-3), 96-108. Rossing, W.A.H., Wijnands, F.G. and Krikke, A.T., 1995. Voortgaande vernieuwing in de landbouw: het samenspel van prototypering en toekomstverkenning. In Studiedag KLV, AB-DLO en PELUW, Wageningen 21 november 1995. Eds. A J Haverkort en P A van der Werff. Pp. 115-135. AB-DLO thema's, Wageningen. Somers, B.M. and R61ing, N.G., 1993. Kennisontwikkeling voor duurzame landbouw. NRLO (National Council for Agricultural Research), Den Haag. Somers, B.M. (ed.) 1995. Plan van Aanpak Werkgroep Telen in de grond. NTS (Dutch Federation of Horticulture Study Groups), Honselersdijk. Somers, B.M., 1996a. Zoeken en leren als invalshoek voor kennisprocessen. Spil 137-138/139-140, 34-38. Somers, B.M. (ed.) 1996b. Kennis op Bedrijfsniveau. LandbouwEconomisch Instituut (LEI-DLO), Den Haag. Van Weperen, W., 1994. Balancing the Minerals, Moving Boundaries: Mineral Balance Extension in Dutch Dairy Farming. Agricultural University, Wageningen. Van Weperen, W. (ed.) 1995. Het veranderingsproces: ervaringen van akkerbouwers bij het omschakelen naar een ge'integreerde bedrijfsvoering. Landbouwuniversiteit/IKCdPAGV, Wageningen/Lelystad.


361

Author Index Agtiera, F., 145 Ammon, H.U., 227 Arsene, G.G., 237 BaRs, G.R., 67 Bockstaller, C., 329 Bonciarelli, F., 179 Bonet Torrens, M., 123 Bonnet, A.-C., 135 Bosch Serra, A.D., 123 Breman, H., 39 Bryson, R.J., 77 Bullock, P., 29 Cabelguenne, M., 113 Chassin, P., 257 Clark, W.S., 77 Dauzat, J., 87 Debaeke, P., 113, 217 Delprat, L., 257 Dirks, B.O.M., 57 Domingo Oliv6, F., 123 Durand, J.-L., 135 Edmeades, G.O., 155 Elings, A., 155 Ellis, R.H., 67 Eroy, M.N., 87 Etchebest, S., 135

Hadley, P., 67 Hammer, G.L., 99 Hassink, J., 171,245 Haverkort, A.J., 191 Helander, C.A., 309 Hensen, A., 57 Jambert, C., 257 Janssen, B.H., 267 Keating, B.A., 99 Kessler, J.J., 39 Kub~it, J., 245 Langeveld, C.A., 57 Leakey, R.R.B., 19 Leterme, Ph., 237 Lin6res, M., 257 Mary, B., 237 Meinke, H., 99 Melines Pag6s, M.A., 123 Mengel, K., 277 Meynard, J.M., 339 Minguez, M.I., 191 Morison, J.I.L., 67 Morvan, T., 237

Neeteson, J.J., 171 Neumann, R., 49

Feil, B., 227

Orgaz, F., 145

Garibay, S.V., 227 Gastal, F., 135 Gatmt, J.L., 201 Ghesqui6re, M., 135 Girardin, P., 329

Paveley, N.D., 77 Porceddu, E., 3 Rabbinge, R., 3, 99 Rossing, W.A.H., 399

Sanchez, P.A., 19 Scott, R.K., 77 Segers, R., 57 Somers, B.M., 353 Stamp, P., 227 Stockdale, E.A., 201 Stockle, C.O., 113, 217 Struik, P.C., 179 Sylvester-Bradley, R., 77 van den Pol-van Dasselaar, A., 57 van der Weft, H.M.G., 329 van Keulen, H., 99, 191 van Ittersum, M.K., 339 Velthof, G.L., 57 Vereijken, P., 293 Villalobos, F.J., 145 Vos, J., 201 Wheeler, T.R., 67 White, J.W., 155 Whitmore, A.P., 245 Wijnands, F.G., 319 Yang, H.S., 267


Subject Index Acacia seyal 41 Africa, sub-Saharan 20, 39 Agricultural Systems 100 Agriculture broadened objectives v, 8, 10 education 4 history 3 infrastructure 26 land use 6, 9 policy, CAP 7 research and development 4, 10, 49, 339, 348 Agro-ecological characterization 29 indicators 329 zones 31 Agroforestry 24, 40, 46 Agroindustry 49 Agronomy broadened objectives v exploration 8, 12 history 3 objectives 8 reorientation vi, 6, 13 synthesis v Allium cepa L. 123 Ammonia volatilization 173, 241 Arable farming 191,268,294, 309, 319 integrated 329 irrigated 194 rainfed 194 AUDPC 82

Breeding 123, 155, 192 Calcium phosphates 281 Carbon dioxide elevated 69, 74 emission 57 Catch crop 179 Cattle 60, 172 Chemical crop protection 49 product innovation 50 synthesis 53 Chlorophyll meter 230 Climate change 33, 57, 67 Coconut 87 Cocos nurcifera 87 Cover crop 184, 229 Crop canopy 78, 81 development 70, 72, 220 disease 82, 182 growth 69, 103, 156, 221 growth models 34, 147, 157, 217 parameters 117 protection 49, 51,297, 312, 320, 345 residues 180, 269 requirements 31, 221 rotation 181,299, 310, 320, 341,344 thermal time 70, 231 yield 51, 71, 73, 116 Cropping systems 45, 99, 184, 227, 277, 339

Biological control 183

Dairy farms 58, 171 Databases, soil, climate 33 Decision aid tool 329, 334, 340 Denitrification 58, 173,258 Diagnostics 206, 341 Dissolved organic carbon (DOC) 257 Drainage 62 Drip irrigation 124 Drought resistance 137 Ear density 341 formation 164 prolificacy 164 Early vigour 146, 153 Ecological infrastructure 299, 312, 315 Eddy correlation 58 Energy efficiency 314 Environment 319 assessment 330 exposure to pesticides 297, 322 impact 331 risk assessment 31, 34 European research network 293 European Union 7, 293 Evapotranspiration 114, 118, 145, 150 Exploration, model based v, 155, 340, 346, 349 Extraction methods, cold water 258 Farm network 302, 330, 334

363

364 Farming system arable 293, 309, 319, 329, 339,353 case study 339 conventional arable 310, 320 ecological arable 294, 306, 311,319 flower bulb based 344 integrated 311,320 integrated arable 293, 306, 310, 319, 329, 353 organic arable 294, 306 sustainable 293, 309, 319, 329, 339, 353 Farms experimental 171, 310, 320 pilot 171,302, 330 Farmyard manure 249, 268, 278, 284, 302, 334 Fertilizer nitrogen 22, 229 phosphorus 23,277 recommendation 206 Festuca arundinacea 136 Festulolium 136 Flower bulbs 344 Fodder crops 195 Food production 10, 30, 35 security v, 19, 29, 32 Forage 136 Fuzzy logic 333 FYM 249, 268, 278,284, 302, 334 Gaseous losses 241 Geographical Information 34 Global warming 33, 57, 67 Glycine max, L. 116 Grain yield 71 Grass canopy structure 139 tiller density 138 Grassland 60, 172, 247 Greenhouse gases carbon dioxide 58, 60, 69, 74 emission estimates 58, 63 flux measurements 59

methane 58, 61 N20 58, 61,207, 257 warming potential 60 Groundwater table 58 Growing season 193 Harvest index 73, 151 Helianthus annuus L. 145 Herbicide 323 History 3 IAFS 329 Ideotyping 123, 135, 145, 161, 193 Immobilization 202, 241 Indicator agro-ecologica1297, 321,329 crop sequence 330 environmental 322, 331 environmental impact pesticides 297, 322, 332 nitrogen status 227 organic matter 332 pesticides 315,322 Integrated arable farming 293,306, 310, 319, 329, 353 crop protection 312, 321 farming systems 311, 319 pest management 51 Intercropping 44, 89 Interspecific hybridization 136 Irrigation 116, 136 Irvingia gobonesis 24 Isotopes ~513C203,259 15N 203, 241 Italian ryegrass 136, 228 LAI 79, 105, 158, 179 Land use diversification 23 intensification 23, 25 sustainable 32 transformation 19 Land Degradation 29 Leaf

area growth 105, 108, 136, 139 area index 79, 105, 158, 197 elongation rate 136, 139 form factor 78 green leaf area index 77 growth, day-night 136 healthy area duration 83, 85 senescence 164 water potential 136 Learning adaptive learning 337 computer-aided learning 348 social learning 357 Leguminous crop 184 Ley crop 184 Light interception 40, 42, 92, 163 sum 70, 84 Linear programming 344 Livestock 173,280 Lolium multiflorum 136, 229 Lolium perenne L. 136, 239 LUE see RUE Maize 258 hybrid 116, 156, 162, 227 open pollinated variety 156 Methane 58, 61 Mineralization 269 Model 100, 217, 339 ACCESS 34 ALMANAC 34 AFRCwheat2 73, 217 biomass 118 CERES 206 comparison 113, 219 complexity 113 CRIES 34 CropSyst 114, 181,206, 217 DAISY 217 D6cibl6 343 DSSAT 34 EUROSEM 34 EPIC 34, 217 exploration v, 155, 340, 346 H6nin-Depuis 332

365 LEACH.M 34 LINTUL 194 maize growth 157 mechanistic 340 N-cycle 204 MIR 93 Musc 93 OILCROP-SUN 147 ORYZA 206 Otelo 343 PAPRAN 196 regression 342 soil organic matter 157, 248, 268 soil-plant cycle 202 soil water 115 SUCROS 87 157 SUNDIAL 34, 205 synthesis v validation 116 winter wheat 73 WOFOST 34 yield 118 yield loss 82 Mulch seeding of mulch 229 living mulch 229 Multi criteria method 330, 336, 341,344 Mycorrhiza 283 Nature and landscape 295 Nematodes 182, 320 Nitrate in maize 229 leaching 173 Nitrogen 99, 101,103 15N 240 accumulation 173 atmospheric deposition 174 availability 40, 223, 314 available reserves 313 budget 171,209 competition for 252 content in plant parts 206, 221,231 diagnostics 206, 233 dynamics 43, 203, 240

fertilisation 58 immobilization 241 in dung and urine 173 in organic matter 173 in soil 229 limitations 99, 110 losses 173, 187, 207 management 172, 196, 316 mineralization 202 models 204, 219 nutrition index 136, 138 plant nitrogen use 105, 110 recovery 42, 241 recycling 43,203, 240 requirements 197, 219 soil nitrogen supply 202, 239 soil- plant transfer 202 status of maize 230 surplus 174, 345 uptake 105, 110, 173, 204, 223,240 use efficiency 196, 205, 341 utilisation 314 Nitrous oxide 58, 61,207, 257 NUE 196, 205, 341 Nutrient balance 21, 43, 185 catch crop 187 loss 41 management 299, 313 residual 184 Objectives 295, 313, 344 Onion 123 Optimalisation 344 Organic farming 294, 306 inputs 249, 296, 344 Organic matter 332 capacity to protect 246 dissolved organic carbon 257 dynamics 253,267 farmyard manure 250, 269 fractionation, density 248 fractionation, size 248 indicator 245, 336 inputs 249, 296, 344 physical protection 246

pools 203 roots and stubble 269 soil type 250 straw 269 Pasture 60 Peat soil drained 58, 62 pasture 58, 60 Perennial ryegrass 136, 140, 239 Pesticide risk evaluation 322 Pesticides 49, 315, 319, 332 pH liming 282 Phosphate adsorption 280 availability 185, 279 fertilizer 277, 284 fixation 280 inorganic 280 mobilization 286 reserves 278 rock phosphate 285 Policy enabling 20, 25 EU-CAP 7 GATT 7 Porous cup vacuum extraction 258 Potato 192, 320 Productivity 20 agricultural 5, 10, 12 labour 5 post-harvest losses 23 soil related constraints 21 Prototyping v, vi, 294, 310, 320, 339, 349 Prunus africana 24 Radiation incident 70, 80, 157, 193 intercepted 78, 84, 93, 161, 193 use efficiency 83, 105, 108, 161,194 Regulations 7, 177 Research and development agriculture 4, 10 chemical crop protection 49

366 participatory 339, 348 Resistance: systemic activated 54 Resource use efficiency vi, 6, 22, 161, 180 Risk and benefit 52, 354 Root competition 43 development 129 distribution 128 length-density 125 mass 148 sampling 125 uptake (N,P,K) 127 Root-shoot relation 127, 148 Root system tree 43 onion 123 RUE 83, 105, 108, 161,181,194 Sahel 20 Scaling, nitrogen budgets 209 Sclerocarya birrea 41 Season length 146, 149 Semi-arid 99 Shorea javanica 24 Simulation model 34, 40, 100, 149, 217,219, 341 Size and density fractionation 248 Slurry 238 Smallholder 22 Soil cultivation 312 depth 149 fertility 20, 46, 181,253,330 moisture 115 organic matter 42, 245 soil-bome fungi 182, 186 test for P 279 type 250 Soil cover index 313

Soil organic matter 44, 46, 203, canopy cover 45 damar 24 245, 332 domestication of indigenous content 269 24 dissolved organic carbon profitable crops 21, 24 258, 261 dynamics 268 Triticum aestivum L. 67, 77, 100, 217, 341 forest soil 257 from maize 265 Volatilization 208, 241,322 molecular weight 261 Solanum tuberosum L. 192 Water 99 Solar radiation 70, 157, 193 availability 40, 116, 156 Sorghum 116 balance 41, 45 Sorghum bicolor, L. 116 deficit 136 Soybean 116 evapotranspiration 145 Straw 269 extraction 43, 46, 106, 137 Sunflower 145 limitation 99, 106, 109 Sustainable loss 41 agriculture v, 179, 309 stress 116, 137 farming systems 186, 293, transport 115 306, 309, 344, 353 use efficiency 145, 150, 194 land use and management 35 Weed(ing) 183,298, 323 production systems v Wheat 99, 341 Sustainability models 100, 217, 341 objectives 340 spring wheat 100 Symbiotic nitrogen fixation 174 winter wheat 67, 77, 217, 341 System analysis 10 Systemic activated resistance 54 Woody plants 40, 45 Temperature 192 gradient 68 high temperature stress 74 sensitivity 73 thermal time 70, 117, 146 Tillage strip tillage of maize 229 minimum tillage 229 Tillering 74 Transpiration efficiency 101, 106, 110, 150 coefficient 116 Tree bush mango 24 crown 41

Yellow rust 77, 82 Yield actual yields 12 attainable yields 12, 191 components 104, 162, 341 gap analysis 341 grain yields 71, 82, 155 loss by disease 51, 82, 182, 341 potential yields 12, 191 seed 149 silage 230 Zea mays L. 155, 227,258


Printed in the United Kingdom by Lightning Source UK Ltd. 117018UKS00001B/26

IL!J!L !!11!11IIIJII 2852

Perspectives for Agronomy (Developments in Crop Science, Volume 25)

Advances in Agronomy, Volume 25

Advances in Agronomy, Volume 60

Advances in Agronomy, Volume 50

Advances in Agronomy, Volume 68

Advances in Agronomy, Volume 54

Advances in Agronomy, Volume 11

Advances in Agronomy, Volume 31

Advances in Agronomy, Volume 33

Advances in Agronomy, Volume 65

Advances in Agronomy, Volume 21

Advances in Agronomy, Volume 64

Advances in Agronomy, Volume 105

Advances in Agronomy, Volume 20

Advances in Agronomy, Volume 100

Advances in Agronomy, Volume 108

Advances in Agronomy, Volume 69

Advances in Agronomy, Volume 83

Advances in Agronomy, Volume 91

Advances in Agronomy, Volume 46

Advances in Agronomy, Volume 111

Advances in Agronomy, Volume 62

Advances in Agronomy, Volume 56

Advances in Agronomy, Volume 28

Advances in Agronomy, Volume 53

Advances in Agronomy, Volume 14

Advances in Agronomy, Volume 26

Advances in Agronomy, Volume 103

Advances in Agronomy, Volume 81

Advances in Agronomy, Volume 40

Advances in Agronomy, Volume 45

Perspectives for Agronomy (Developments in Crop Science, Volume 25)

Advances in Agronomy, Volume 25

Advances in Agronomy, Volume 60

Advances in Agronomy, Volume 50

Advances in Agronomy, Volume 68

Advances in Agronomy, Volume 54

Advances in Agronomy, Volume 11

Advances in Agronomy, Volume 31

Advances in Agronomy, Volume 33

Advances in Agronomy, Volume 65

Advances in Agronomy, Volume 21

Advances in Agronomy, Volume 64

Advances in Agronomy, Volume 105

Advances in Agronomy, Volume 20

Advances in Agronomy, Volume 100

Advances in Agronomy, Volume 108

Advances in Agronomy, Volume 69

Advances in Agronomy, Volume 83

Advances in Agronomy, Volume 91

Advances in Agronomy, Volume 46

Advances in Agronomy, Volume 111

Advances in Agronomy, Volume 62

Advances in Agronomy, Volume 56

Advances in Agronomy, Volume 28

Advances in Agronomy, Volume 53

Advances in Agronomy, Volume 14

Advances in Agronomy, Volume 26

Advances in Agronomy, Volume 103

Advances in Agronomy, Volume 81

Advances in Agronomy, Volume 40

Advances in Agronomy, Volume 45

Recommend Documents