Agroforestry Systems 43: 1–4, 1999.  1999 Kluwer Academic Publishers. Printed in the Netherlands.

Preface In the last fifty years European agricultural productivity has increased dramatically, and even greater productivity may be expected in the future as a result of further technical innovations. However, although this has resulted in tremendous success concerning food production, it has also led to excessive costs, and to conflicts with trading partners due to subsidised dumping of European Union surpluses on the world market, mainly to the detriment of developing countries. In addition, environmental problems are arising due to current intensive production methods.1 A number of dynamic farmers, with the advice of research scientists, have started to innovate, by updating traditional agroforestry techniques in order to diversify their production, reduce costs, and improve their environment. But in late 1997, a local representative of the French agricultural administration suddenly decided that intercrops in agroforestry plots could no longer be considered as legal, and would therefore no longer benefit from the current agricultural crops support scheme. This resulted in the interruption of a major agroforestry experiment, hopefully only temporarily. Indeed, up to now the French law officially recognises only forest land or agricultural land, but a mixture of both is still inconceivable for legal and fiscal reasons. The European common agricultural policy has not yet caught up with the agroforestry innovations which have been implemented in real farms in the past decade. However, there is an increasing awareness of the importance of the role of trees as environmental agents. Temperate examples of this include their use to reduce pollution of water courses in riparian buffer zones in the US, to control rising water tables and salinity in Australia or to provide wildlife habitat by the re-introduction of woody perennials into agricultural land in Europe. The impact of this trend on the task of the researcher and the modeller is to increase the complexity of what is required to encompass the series of environmental as well as productive functions. While there has been considerable progress over the last decade in both research and modelling, particularly in understanding interactions amongst productive components of agroforestry systems, in most countries and environmental contexts there is still much work to be done before reliable predictions of trade-offs between productive functions and environmental impacts can be made. The scientific study of agroforestry in the past has been mostly concerned with tropical regions, focusing on the diagnosis of land use problems and the development of potential solutions to these problems that involve the use of trees. Recently, however, the understanding of agroforestry systems, that involve greater diversity and complexity than conventional agricultural and forest systems, has become a research interest world-wide, posing fundamental

2 questions in ecological science and demanding integration across disciplines. Despite great differences in social, economic and political situations throughout the world, fundamental aspects concerning the way agroforestry practices modify basic ecological and sociological processes are common to many land-use systems. In order to collate, evaluate and disseminate the basic knowledge gained in recent years, and to bridge the gap between temperate and tropical agroforestry research, a workshop was organised in Montpellier Agropolis International Centre (23–29 June 1997), France, under the auspices of IUFRO research group 1.15.00 (Agroforestry), and more specifically by working parties 1.15.03 (fundamental research and modelling in agroforestry) and 1.15.01 (temperate agroforestry). It seemed relevant for a meeting on agroforestry to be held in Montpellier, one of the most ancient university towns in Europe, where one of the world’s first botanical gardens was established in 1593. The Institut de Botanique provided some of the pioneers in tropical agroforestry research. Despite the above-mentioned administrative barriers, the Languedoc-Roussillon region hosts over a hundred hectares of on-farm modern agroforestry experimental plots, and two main agricultural research and development institutes involved in agroforestry (despite institutional and hierarchical barriers) are present in Montpellier, INRA focusing mostly on temperate regions, and CIRAD devoted to tropical aspects. The two main objectives of the workshop were firstly, to bring together what is understood about the processes underlying agroforestry practice, and, secondly, to provide a forum to explore relevant models and modelling approaches. The workshop was also able to play a role in examining the agroforestry systems encountered in temperate and Mediterranean areas, including both traditional and more innovative agroforestry practices. It gathered over 140 scientists from 30 countries world-wide. Ninety four contributions were presented, concerning the following aspects: ecological interactions amongst components, environmental impact, economics and policy, modelling. A book of abstracts was published.2 This special issue of the journal Agroforestry Systems comprises a selection of original contributions presented at the workshop, all of them refereed in accordance with the practice of the journal. They represent most of the key issues raised during the meeting. Significant progress has been made in the last few years in developing a quantitative understanding of interactions amongst the major productive components of agroforestry practices, in terms of their resource capture and ameliorative effects on soil and microclimate. Key advances centre on being able to deal with trees as individuals rather than as stands, both in terms of measurement and modelling. While more has been achieved in this respect for tree crowns and atmospheric processes, than for root systems and the soil, new techniques and a recent concentration of research funding have led to

3 increasing resolution below-ground. There remains a requirement to include effects of the various non-productive biological components of agroforestry practices such as weeds, pests and diseases and secondly, to encompass a greater productive biodiversity. There is also a pressing requirement to provide reliable procedures for scaling up from measurements made on research plots and a field scale to understand and predict processes at farm, landscape and regional levels. With respect to modelling, it is clear that the development of process-based biophysical models has been effective in synthesising what is known and has led to explanations for why particular agroforestry practices have been successful in some environments but not in others. These developments are, however, fairly specific to the species combinations and management regimes studied, and concerns remain about the extent to which results of research and modelling in agroforestry can be extrapolated to other situations. In New Zealand, empirical models, based on a well-defined production system that operates over a wide geographical area, have been reliably used for some time to support farmer decision-making. Apart from this example, there is little hard evidence that process-based models are being effectively used to facilitate decisions either at the scale of the farmer or the policy-maker seeking to influence farm decisions. This exposes a critical need to validate processbased models and initiate interaction between models and their intended users. Bringing together the increasing complexity of agroforestry models (in terms of the number of components considered and their degree of disaggregation) and the opportunities presented in information technology, there is now a welldeveloped interest in modelling environments that improve the efficiency of translating research results into model formulations and may assist in the standardisation of both measurement protocols and model representations. In conclusion, it appears that agroforestry research and modelling approaches are gradually evolving to cope with the greater complexity involved in dealing with biodiverse systems, demanded by changing land use and environmental imperatives. Significant challenges exist in ensuring that increasingly sophisticated models are made readily accessible to people who make decisions about farming and land use policy. Temperate countries, which have up to recent times considered agroforestry as an archaic practice, are now exhibiting a renewed interest. Their research organisations are liable to play an important role in the understanding of fundamental processes related to such complex systems, which should benefit the rest of the world. This exchange of knowledge has helped make progress towards the understanding of key factors such as the interactions between plants, in particular belowground, as well as biophysical or socio-economic modelling and the integration of models. Much more is to be expected from current agroforestry experiments, if they are allowed to run until their completion. . . .

4 Acknowledgements Thanks are due to the scientists who anonymously reviewed the manuscripts selected for this special issue. By their comments and suggestions, they have greatly contributed to the scientific quality of many of these papers. The discussion leaders, particularly Fergus Sinclair (co-ordinator of IUFRO research group 1.15.00), have contributed some of the ideas expressed in this preface. The organisation of the workshop has been made possible thanks to financial contributions by INRA, CIRAD, the European Commission (DG VI), Région Languedoc-Roussillon, Agropolis, Conseil Général de l’Hérault, IUFRO-SPDC, Ministère Français des Affaires Étrangères, Ministère Français de la culture, TUBEX Company.

Notes 1. Ground for choices – four perspectives for the rural areas in the European Community. Netherlands scientific council for government policy, report 42, 1992, 144 pp. 2. Agroforestry for sustainable land-use. Book of abstracts, international workshop, Montpellier. Auclair D (ed), INRA-CIRAD, Montpellier, 1997, 452 pp. Available at CIRAD, service des publications, BP 5035, 34032 Montpellier, France.

D. Auclair1 and C. Dupraz2 (guest editors) 1 INRA-CIRAD, Montpellier, France 2 INRA, LEPSE, Montpellier, France