Quality of Life and Management of Living Resources Silvoarable Agroforestry For Europe (SAFE)
European Research contract QLK5-CT-2001-00560
SAFE PROJECT FINAL PROGRESS REPORT August 2004-January 2005
May 2005
Volume 4 :Appendices
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 1
Quality of Life and Management of Living Resources Silvoarable Agroforestry For Europe (SAFE)
European Research contract QLK5-CT-2001-00560
CAPTION OF THE COVER PICTURE
In March 2005, Jérome Feracci, a cereal farmer near the Béziers town in France, established a 25 ha silvoarable plot with walnut trees. He was convinced to adopt agroforestry after visiting some SAFE consortium experiments such a the Restinclières farm near Montpellier.
Such plantations are one of the most evident consequence of the SAFE project, and will be landmarks for the future development of agroforestry across Europe.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 2 Appendices TABLE OF CONTENT
This annex includes annexes to contractor reports and some scientific papers produced by the SAFE consortium during the last year of the project
ANNEX 1. LIST OF THE SAFE PUBLICATIONS ...... 8 Papers published or accepted ...... 8 Papers submitted ...... 10 Papers in préparation ...... 10 Extension papers ...... 11 Internal reports ...... 11 Posters ...... 12 Communications to congresses...... 12
ANNEX 2. ECONOMICS OF SILVOARABLE SYSTEMS USING A NOVEL APPROACH : THE LAND EQUIVALENT RATIO BASED GENERATOR ...... 13 Introduction ...... 14 The LER approach...... 14 LER-biomass and LER-product...... 14 The LER-based generator...... 15 A wise hypothesis for the agroforest tree growth ...... 16 Maximum expectable LER in function of species and final density...... 17 Impact of the TGA on the LER results...... 19 Data references and main hypothesis...... 20 The forestry references ...... 20 Reference data in agriculture ...... 21 The profitability threshold yield...... 22 Main management features of the agroforestry systems ...... 23 Economic hypothesis...... 25 Main results ...... 27 Labour impact for one silvoarable hectare ...... 27 Prediction of yield evolution...... 27 Cash flow impact ...... 29 Profitability of a silvoarable investment ...... 32 Main conclusions ...... 39 Bibliography ...... 39 Annex ...... 40
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 3 Annex 1: Detailed description of the LERbased-Generator ...... 40 Annex 2: Labour, revenues and costs in the 3 types of farms ...... 46 Annex 3: Economic data relative to monocropped or intercropped walnut, wild cherry and poplar in the 3 farms...... 47
ANNEX 3. WHAT LAND STATUS FOR AGROFORESTRY PLOTS IN FRANCE ? (IN FRENCH) 49 Définition de l’agroforesterie ...... 50 Problématique liée au statut ...... 51 Les solutions possibles...... 52 Un forfait spécial agroforesterie...... 53 Un forfait distinct au prorata ...... 53 Analyse des solutions présentées ...... 54 D’un point de vue administratif ...... 54 Conséquence pour le calcul de l’impôt foncier...... 55 Conséquence pour le calcul de l’impôt sur le revenu...... 56 Conséquence pour le calcul de l’imposition sur le patrimoine ...... 56 Conséquence pour la gestion de l’exploitation...... 56 Solution proposée...... 57 Faut-il modifier la loi ? ...... 58 Les surfaces boisées et le code rural...... 58 Le bail agroforestier...... 59 Appendix : rapport du Bureau des Etudes Fiscales du 8 oct. 2004 ...... 60
ANNEX 4. QUELLE PLACE POUR LES ARBRES HORS FORET DANS LA NOUVELLE PAC ? ...... 63 Eligibilité des parcelles arborées aux paiements compensatoires...... 69 Place de l’arbre dans l’historique des réformes de la PAC ...... 69 Conclusions au niveau européen ...... 73 Le régime d’application en France ...... 75 Les arbres dans le deuxième pilier de la PAC...... 82 Le Règlement de Développement Rural ...... 82 Les aides disponibles en France...... 85 Bibliographie ...... 93 Annexe 1: Article TransRural Initiative de déc 2004 ...... 94 Annexe 2: Propositions du Groupe Réglementations – SAFE ...... 95 Annexe 3: Lettre de Luc Guyau APCA au MAAPAR – PAC...... 97 Proposition de modification du projet de règlement européen concernant le soutien au développement rural (proposition du 14/07/04) ...... 98 Annexe 4: Chapitre 10 de la circulaire forêt de protection du 7 mai 01 ...... 102
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 4 Annexe 5: Texte MAE Habitats Agroforestiers...... 105 Annexe 6: Lettre de Luc Guyau APCA au MAAPAR – MAE ...... 114
ANNEX 5. THE DISTRIBUTION OF SILVO-ARABLE SYSTEMS IN WESTERN EUROPE AND THEIR ECOLOGICAL CHARACTERISTICS ...... 116 Introduction ...... 116 A European Stratification System for Resource Assessment ...... 116 A Worked Example of the application of strata in Atlantic Europe ...... 117 Case Studies in Southern Europe ...... 118 Ecological Considerations of Silvo-arable ...... 119 Future Work ...... 120 Conclusions...... 121
ANNEX 6. SILVOARABLE AGRICULTURE IN EUROPE – PAST, PRESENT AND FUTURE 122 Abstract...... 122 Introduction ...... 123
Historical perspective...... 125 Data Collection ...... 127 Systems...... 128
Olive tree associations...... 130
Fruit tree associations ...... 131
Timber tree associations...... 136
Oak tree associations ...... 138
Fodder tree associations ...... 143
Conclusions ...... 150 Acknowledgements...... 151 References ...... 151
ANNEX 7. THE DEVELOPMENT AND USE OF A FRAMEWORK FOR CHARACTERISING COMPUTER MODELS OF SILVOARABLE ECONOMICS...163 Abstract...... 164 Introduction ...... 165 Materials and methods ...... 166 Development of a framework for characterising models ...... 166 Use of the framework...... 168 Results and discussion ...... 168
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 5 Model background ...... 168 Systems modelled ...... 170 Objective of economic analysis ...... 171 Viewpoint of the economic analysis ...... 174 Spatial scale ...... 175 Generation and use of biophysical data...... 176 Model platform and interface...... 177 Input requirements and outputs generated ...... 178 Conclusions...... 180 References ...... 181
ANNEX 8. FINE ROOT DISTRIBUTION IN DEHESAS OF CENTRAL-WESTERN SPAIN 192 Abstract...... 193 Introduction ...... 194 Material and Methods...... 195 Study Area...... 195 Soil cores: Root Length Density ...... 195 Road Cuts: Maximum tree rooting depth and horizontal spread...... 196 Results ...... 197 Linear relationships for root length density estimation ...... 197 Vertical profiles of root length density ...... 197 Lateral root distribution...... 198 Effect of soil management on root distribution ...... 198 Tree roots in road cuts...... 198 Discussion ...... 199 Herbaceous plants root system...... 199 Holm-oak root profiles ...... 199 Lateral root distribution...... 200 Combined root system: implication on competition for soil resources ...... 200 Conclusions...... 201 References ...... 201
ANNEX 9. THE DEVELOPMENT AND APPLICATION OF BIO-ECONOMIC MODELLING FOR SILVOARABLE SYSTEMS IN EUROPE ...... 210 Keywords ...... 210 Abstract...... 210 Introduction ...... 211 Method...... 211
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 6 Identification of landscape test sites...... 212 Characterisation of landscape test sites...... 212 Selection and management of tree and crop species ...... 213 Biophysical modelling ...... 213 Plot-scale economic modelling...... 214 Parameterisation and use of Farm-SAFE ...... 216 Results and discussion ...... 219 Biophysical production in arable and forestry systems ...... 219 Biophysical production in silvoarable systems ...... 220 Land equivalent ratios ...... 220 Plot-scale economic results...... 221 Farm-scale feasibility...... 223 Summary and recommendations...... 224 Conclusion...... 226 Acknowledgements...... 227 References ...... 227 Tables...... 231 Captions for Figures ...... 237 Figures ...... 238
ANNEX 10. YIELD-SAFE: A PARAMETER-SPARSE PROCESS-BASED DYNAMIC MODEL FOR PREDICTING RESOURCE CAPTURE, GROWTH AND PRODUCTION IN AGROFORESTRY SYSTEMS...... 247 ABSTRACT...... 248 INTRODUCTION...... 249 MATERIALS & METHODS...... 253 MODEL DESCRIPTION ...... 253 POPLAR VALIDATION DATA...... 266 MODEL CALIBRATION FOR POPLAR AND INTERCROPS ...... 267 MODEL VALIDATION FOR POPLAR AGROFORESTRY SYSTEMS ...... 269 RESULTS...... 270 AGROFORESTRY EXPERIMENTS WITH POPLAR ...... 270 MODEL CALIBRATION...... 271 MODEL VALIDATION ...... 273 SENSITIVITY ANALYSIS...... 273 DISCUSSION ...... 275 ACKNOWLEDGEMENT ...... 277 REFERENCES...... 277
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 7 ANNEX 1. List of the SAFE publications
The full manuscripts of the 5 papers in bold are included in this annex. Most of the other papers are available on the SAFE web site. Only published or accepted papers are available in full on the public pages. The other papers are available in the private pages of the web site (password controlled access)
PAPERS PUBLISHED OR ACCEPTED
Chifflot, V., Bertoni, G., Cabanettes, A. & Gavaland, A. (2004) Beneficial effects of intercropping on the growth and nitrogen status of wild cherry and hybrid walnut trees. Agroforestry Systems (in press).
Cubera E., Montero M.J., Moreno G. 2004. Effect of land use on the soil water dynamic in dehesas of Central-Western Spain. In: Sustainability of Agrosilvopastoral Systems - Dehesa, Montados -. S. Schnabel and A. Gonçalves (eds.). Advances in GeoEcology, 37, Catena Verlag, Reiskirchen (in press).
De Filippi R., Reisner Y., Herzog F., Dupraz C., Gavaland A. 2004. Modelling the potential distribution of Agroforestry systems in Europe, using GIS EnviroInfo 2004, CERN, p 423-426.
Dupraz C., 2005. From silvopastoral to silvoarable systems in Europe: sharing concepts, unifying policies. In Silvopastoralism and Sustainable Land Management. Mosquera-Losada R., Riguerio, A., McAdam J., Eds, CAB International, 432 pages.
Dupraz C., 2006. Entre agronomie et écologie : vers la gestion d’écosystèmes cultivés. Cahier d’étude DEMETER - Economie et Stratégies agricoles, Paris, pagination en cours, 16 pages
Dupraz C., Capillon A. 2006. L’agroforesterie : une voie de diversification écologique de l’agriculture européenne? Cahier d’étude DEMETER - Economie et Stratégies agricoles, Paris, pagination en cours, 11 pages
Dupraz C., Liagre F., Manchon O., Lawson G., 2004. Implications of legal and policy regulations on rural development: the challenge of silvoarable agroforestry in Europe. In : Meeting the challenge : Silvicultural Research in a changing world. IUFRO World Series Volume 15, Parotta et al, (Eds.), 34-36.
Eichhorn E.P., Paris P., Herzog F., Incoll L.D., Liagre F., Mantzanas K., Mayus M., Moreno Marcos C., Dupraz C., Pilbeam DJ., 2005. Silvoarable agriculture in Europe – past, present and future. Agroforestry Systems, in press
Graves, A. R., Burgess, P.J., Liagre, F., Dupraz C. & Terreaux, J.-P. (2005). Development and use of a framework for characterising computer models of silvoarable economics. Agroforestry Systems, 65:53–65
Keesman, K.J. and R. Stappers. 2004. Nonlinear set-membership estimation: A support vector machine approach. Journal of Inverse and Ill-Posed Problems, Volume 12, No. 1, pp. 27-42.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 8 Montero M.J., Obrador J.J., Cubera E., Moreno G. 2004. The Role of Dehesa land use on Tree Water Status in Central-Western Spain. In: Sustainability of Agrosilvopastoral Systems - Dehesa, Montados -. S. Schnabel and A. Gonçalves (eds.). Advances in GeoEcology, 37, Catena Verlag, Reiskirchen (in press).
Montero, M.J., Moreno, G. 2004. Light availability for understory pasture in Holm-oak dehesas. In: Silvopastoralism and Sustainable land management" published by CAB INTERNATIONAL (in press). 4 pages.
Moreno G., Obrador J.J., Cubera E., Dupraz C., 2005. Root distribution in dehesas of Central-Western Spain. Plant and Soil, accepted for publication
Moreno, G., Obrador, J. J., Garcia, E., Cubera, E., Montero, M. J., Pulido, F. & Dupraz, C. (2005) Competitive and Facilitative interactions in dehesas of C-W Spain. In Press, special issue of Agroforestry Systems.
Moreno, G., Obrador, J., García, E., Cubera, E., Montero, M.J., Pulido, F. 2004. Consequences of dehesa management on the tree-understory interactions. In: Silvopastoralism and Sustainable land management" published by CAB INTERNATIONAL (in press). 4 pages.
Obrador, J.J., Moreno, G. 2004. Soil nutrient status and forage yield at varying distances from trees in four dehesas in Extremadura, Spain. In: Silvopastoralism and Sustainable land management" published by CAB INTERNATIONAL (in press). 4 pages.
Obrador-Olán J.J., García-López E., Moreno G. 2004. Consequences of dehesa land use on nutritional status of vegetation in Central-Western spain. In: Sustainability of Agrosilvopastoral Systems - Dehesa, Montados -. S. Schnabel and A. Gonçalves (eds.). Advances in GeoEcology, 37, Catena Verlag, Reiskirchen (in press).
Palma, J., Graves, A., Bregt, A., Bunce, R., Burgess P., Garcia, M., Herzog, F., Mohren, G., Moreno, G. and Reisner, Y. (2004). Integrating soil erosion and profitability in the assessment of silvoarable agroforestry at the landscape scale. In Proceedings of the Sixth of the International Farming Systems Association (IFSA) European Symposium on Farming and Rural Systems at Vila Real 4-7 April 2004. 817-827. The Proceedings are available at: http://home.utad.pt/~des/ifsa/index.htm
Paris P., Pisanelli A., Tadaro L., Olimpieri G. Cannata F. 2004. Growth and water relations of walnut trees (Juglans regia L.) on a mesic site in central Italy Agroforestry system (in press).
Parveaud C.E., Sabatier S.A., Dauzat J., Auclair D. 2003. Influence of morphometric characteristics of the Hybrid Walnut tree crown (Juglans nigra x Juglans regia) on its radiative balance. In: Hu B.G., Jaeger M. (ed.) Plant Growth Modeling and Applications. Tsinghua Univ. Press / Springer , Beijing (China). pp. 296-304.
Pulido, F.J., García, E., Obrador, J.J., Montero, M.J. 2004. Effects of management on acorn production and viability in holm oak dehesas. In: Silvopastoralism and Sustainable land management" published by CAB INTERNATIONAL (in press). 4 pages.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 9 PAPERS SUBMITTED
Graves A.R., Burgess P.J., Palma J.H.N., Herzog F., Moreno G., Bertomeu ., Dupraz C., Liagre F., Keesman K., van der Werf W., 2005. The development and application of bio-economic modelling for silvoarable systems in Europe. Submitted to Ecological Engineering.
Keesman, K.J., W. v.d. Werf and H. v. Keulen, 2005. Mathematical Production Ecology: Analysis of a Silvo-arable Agro-forestry System. Submitted to Bull. Math. Biol.
Lambs, L., Muller, E., Chifflot, V. & Gavaland, A. Sap flow measurements of wild cherry trees (Prunus avium) in an agroforestry system during a dry summer, South- west of France. Submitted to Annals of Forest Science, 15th November 2004.
Reisner, Y.; Herzog, F. and De Filippi, R. (2005): Target regions for silvoarable Agroforestry in Europe. Submitted to Ecological Engineering. van der Werf W., Keesman K., Burgess P., Graves A., Pilbeam D., Incoll L.D., Metselaar K., Mayus M., Stappers R., van Keulen H., Palma J., Dupraz C., 2005. Yield-SAFE: a parameter-sparse process-based dynamic model for predicting resource capture, growth and production in agroforestry systems. Submitted to Ecological Engineering.
PAPERS IN PREPARATION
Dufour L., Dupraz C., (2005). Effect of tree competition on durum wheat yield in a Mediterranean agroforestry system. En preparation pour EJA
Dupraz C, Vincent G., Lecomte I., Van Noordwijk M. (2006) Modelling 3D interactions of trees and crops with the Hi-SAFE model. En preparation pour Forest Ecology and Management
Lusiana B., Noordwijk M V., Dupraz C. and de Willigen P., (2006) A process-based algorithm for sharing nutrient and water uptake between plants rooted in the same volume of soil II. Nutrients in static root systems
Moreno G., Obrador J.J., García E., Cubera E., Montero M.J., Pulido F. and C. Dupraz. *Competitive vs Facilitative interactions determined by land use in oak Dehesas. In preparation for Agroforestry Systems
Mulia R., Dupraz C., (2005). The growth behaviour of plant root system, including negative-geotropism, in homogeneous and heterogeneous soil resource condition
Mulia R., Dupraz C., (2005). Unusual 3D fine root distributions of two deciduous tree species observed in Southern France: what consequences for root dynamics modelling? In preparation for Plant and Soil
Mulia R., Dupraz C., van Noordwijk M. (2005) A 3D model with voxel automata to simulate plant root growth in heterogeneous soil condition. I. Modelling concepts
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 10 Noordwijk M. v, Mulia R., Dupraz C., Lusiana B. (2006) A process-based algorithm for sharing nutrient and water uptake between plants rooted in the same volume of soil III. Growing root systems
Noordwijk M. v., Lusiana B., Dupraz C.,, Radersma S., Ozier-Lafontaine H., de Willigen P., (2006). A process-based algorithm for sharing nutrient and water uptake between plants rooted in the same volume of soil I. Water in static root systems
Palma, J.; Bunce, R.; De Filippi, R.; Herzog, F.; Van Keulen, H.; Mayus M., Reisner, Y. (2005): Assessing the environmental effects of agroforestry at the landscape scale. Ecological Engineering. In prep.
EXTENSION PAPERS
Dupraz C., Liagre F., (2006). Agroforesterie pratique. Editions France Agricole, en préparation, 250 pages environ
Gerardo MORENO. 2004. El árbol en el medio agricola. Foresta (in press).
INTERNAL REPORTS
Chavet M., E. Hallot, C. Pichery, H. Pruvost, J. P. Terreaux, 2004, Agroforestry: Towards economic land equivalent ratio, Zürich, 2 - 3 novembre 2004
F. Herzog and C. Dupraz, 2005. Agroforestry in Europe – Learning from Tropical Agroforestry
Graves, A. R., Burgess, P.J., Liagre, F., Terreaux, J.-P. & Dupraz, C. (2003). The development of a model of arable, silvoarable and forestry economics. Unpublished draft paper. Silsoe, Bedfordshire: Cranfield University.
Lambs L., E. Muller, V. Chifflot et C. Dupraz 2005 Consommation en eau et ressources hydriques pour des peupliers en agroforesterie JEF
Pasturel, P. (2004). Light and water use in a poplar silvoarable system. Unpublished MSc by Research thesis. Silsoe: Bedfordshire: Cranfield University. 143 pp.
Terreaux J.P., M. Chavet, 2002, Problèmes économiques liés à l'agroforesterie, Cabinet Chavet, Paris, 85 pages.
Terreaux J.P., M. Chavet, 2004, An intertemporal approach of Land Equivalent Ratio for agroforestry plots, Lameta, DT 2004-15, 18 p.
Terreaux J.P., M. Chavet, T.H. Thomas, 2003, Silvoarable agroforestry: some economic problems, Powerpoint presentation, Orvieto, Italy, 14-18 octobre 2003, 30 p.
Terreaux JP, Michel Chavet, Anil Graves, Christian Dupraz, Paul Burgess and Fabien Liagre Evaluating agroforestry investments
Yoda K. Dupraz C., Dauzat J., 2005. Comparison of daily and seasonal variations of radii among trunk, branch and root in Juglans nigra L. x regia L.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 11 POSTERS
Graves A., P. Burgess, F. Liagre, C. Dupraz, J.-Ph. Terreaux, 2004, The development of an economic model of arable, agroforestry and forestry systems, 1st World Congress of Agroforestry, Orlando, Florida, 27 June – 02 July, Poster.
COMMUNICATIONS TO CONGRESSES
Agostini, F., Pilbeam, D.J., Incoll, L. D. & Lecomte, I. (2004). Data management for Decision Support Systems (DSS) in agroforestry. Poster presented (by Dr P Burgess) at 1st World Congress of Agroforestry, Orlando, Florida, USA, 27th June – 2nd July 2004. Abstract published http://conference.ifas.ufl.edu/wca/Abstracts2.pdf
Auclair D., Laurans M. , Chopard J., Leroy C. Parveaud C.E. 2004. Three- dimensional tree architectural analysis and modelling to study biophysical interactions. In: First World Congress on Agroforestry, Orlando, FL (USA), 27/06- 02/07 2004. Poster presentation.
Chifflot V., Bertoni G., Gavaland A., Cabanettes A. and Dupraz C., 2004. Improving growth and nutritional status of highly valuable broadleaf species through intercropping. 1st World Congress of Agroforestry, Orlando, Florida, USA, 27 June to 02 july 2004.
Chifflot V., Bertoni G., Gavaland A., Cabanettes A. and Dupraz C., 2004. Improving growth and nutritional status Improving growth and nutritional status of highly valuable broadleaf species through of highly valuable broadleaf species through intercropping. Poster presented at the first world congress of Agroforestry, Orlando, June 2004.
Cubera E., Moreno G., Solla A., 2004. TDR-measurement for the study of the seasonal variations of soil moisture on quercus ilex dehesas. In: Proceedings: Workshop on Water Use of Woody Crops: techniques, issues, modelling and applications on water management. Ílhavo (Aveiro, Portugal) -May 2004. 2 pages.
Dupraz C., Vincent G., Lecomte I., Mulia R, Jackson N., Mayus M., Van Noorwijk M., 2004. Integrating tree-crop dynamic interactions with the Hi-SAFE model. Communication presented at the first world congress of Agroforestry, Orlando, June 2004.
Graves A., P. Burgess, F. Liagre, J.-Ph. Terreaux, C. Dupraz, 2004, A comparison of computer-based models of silvoarable economics, 1st World Congress of Agroforestry, Orlando, Florida, 27 June – 02 July.
Guido Bongi, Pierluigi Paris (2004) Leaflet heterogeneity in Juglans regia: an un- adverted bias in assimilation models. International Walnut Congress, Sorrento, Italy. October 2004
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 12 ANNEX 2. Economics of silvoarable systems using a novel approach : the Land Equivalent Ratio based generator
T, Borrell1, C. Dupraz1 , and F, Liagre2
1Institut National de la Recherche Agronomique, Montpellier, France
2Assemblée Permanente des Chambres d’Agriculture, Paris, France
Mars 2005
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 13
INTRODUCTION
During the fourth year, Thomas Borrell and Fabien Liagre, in collaboration with Christian Dupraz - INRA, have realised all the technical and economical simulations for the French context. APCA participated to the redaction of the French chapter for the deliverable 7.2 - Plot economics of European silvoarable systems report – leaded by the partner Chavet.
Face to the late we observed in the results we should have received from yield-safe, it was decided to go on our simulations using the Ler-Safe data to feed Farm-Safe.
All the results of our simulations have been presented to all the Chambers of Agriculture which have provided the economical data in the 3 regional meetings we named before.
THE LER APPROACH
LER-biomass and LER-product
The Land Equivalent Ratio indicates the area of monocultures needed to produce as much as one intercropped hectare (Vandermeer, 1989). It is calculated as the sum of relative areas (RA), i.e. productions ratios: for each product, the intercrop production divided by the monoculture production. In most of the agroforestry cases, there are 2 RAs: the crop RA and the tree RA. For instance, a tree RA of 0.7 means that an agroforestry plot produces as much timber as a forestry plot of 0,7 ha. A LER of 1.3 thus indicates than intercropping produces 30% more than monocropping.
However, it can be calculated either with total biomass or only with commercial products, particularly in the case of timber production: the higher rate of thinning in forestry than in agroforestry implies different tree relative yields whether it is calculated with or without thinned trees.
This distinction leads to two different indicators: the LER-products, calculated with the commercial products (bole of timber of the felled trees, grain of the cereals, etc.), and the LER-biomass, calculated with the total biomass produced on the plot (for their detailed way of calculation, see Dupraz et al., 2005). Although the likely range of values for the LER-products is still to be defined with experimental plots and models, we already know that the expected values of LER-biomass are likely to be comprised between 1 and 1.4. Indeed, a value below 1 is biologically unrealistic considering that if one of the intercrops dominates too much the other, it shall perform as in a monoculture plot and thus produce as much biomass of the same area of monoculture production. A value above 1.4 seems too much optimistic with regards to present experimental results and bibliographical documentation (Dupraz et al., 2005).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 14 The LER-based generator
For this study, we used a constrained generator of data: forest, arable and agroforest time-series are generated in accordance with an expected LER-biomass (see Annex 1: Detailed description of the LERbased-Generator).
The tree RA-biomass is defined according to the densities in forestry and agroforestry and to the expected increase in tree growth rate at low density. The crop RA-biomass is then deduced in order to reach the predetermined LER-biomass. The LER is only divided in a crop component and a tree component (timber); it is thus impossible to generate data-sets for a third component (fruits for example), such as for a traditional orchard or double purpose walnuts.
The arable and forest reference data and the values of these two RAs permit to generate all the time-series under constraint :
o the arable time-series is the repetition of the reference yields in accordance with the rotation;
o the forest time-series is generated in function of the reference volume of timber per ha at felling;
o the two agroforest times-series (one for the intercrop, one for the trees) are generated so that the constraint fixed by the RAs is respected (sum of productions for the intercrop, volume of timber per ha at felling for the trees).
Amongst the hypothesis made in this generator, we assume that :
o the agroforest trees are felled at the same time as the forest trees, but their higher growth induces bigger individual pieces of timber; In any case, the unit volume in agroforestry doesn’t exceed 20 % of the forestry volume one.
o there is no difference in the partition of biomass for the intercrop and a classical arable crop: the crop RA-products is thus equal to the crop RA- biomass, which shall both be called “crop RA”;
o the intercrops cannot offer higher yields than the arable crops without any tree (consequently the value of the crop RA cannot be superior to the maximum intercropping area: 1 – the proportion of area occupied by the tree strips); we made therefore the hypothesis that the trees don’t affect positively the crop yield which could be discuss on a long term period (soil erosion and fertility, wind effect, etc.).
o the width of the intercropped alley can be reduced by successive steps when the yield decreases (less productive areas are given up), in order to preserve economically acceptable yields as long as possible. When it cannot be reduced anymore (at a minimum width), the intercrop is suppressed when it is no more profitable (profitability threshold yield).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 15 A wise hypothesis for the agroforest tree growth
In most of the cases, agroforest designs are at lower density than forest designs. As a consequence, trees grow quicker. We assume that the growth rate increases when the density decreases, until to reach a critical final density where the genetic potential is fully expressed. Below this density, we assume that trees don’t grow more, even if they are completely isolated.
At this critical density, we assume that trees grow at a rate driven by a coefficient: the individual tree timber volume growth acceleration in low density AF conditions, or Tree Growth Acceleration (TGA). At critical density, the volume of an agroforest individual piece of timber at felling, VAF, is thus calculated as:
VmaxAF =TGA x VF
where VF = volume of a forest individual piece of timber at felling (in the forestry reference data which is used).
VmaxAF is thus the maximum volume of an individual piece of timber.
Unfortunately, the critical densities and the likely range of values for TGA are not well documented. Thus these parameters had to be fixed by expert knowledge.
In order to realise wise simulations, we assumed a quite low value of TGA: 1,2 for the three species (Table I).
Species Final density in VF TGA Critical VmaxAF forestry density
Walnut 100 trees/ha 1 1,2 50 trees/ha 1,2 m3/tree m3/tree
Wild 150 trees/ha 0,8 1,2 60 trees/ha 0,96 cherry m3/tree m3/tree
Poplar 200 trees/ha 1,5 1,2 100 trees/ha 1,8 m3/tree m3/tree
VF is the volume of timber of an individual forest tree. TGA is individual Tree timber volume Growth Acceleration in agroforestry at densities lower or equal to the critical density. The critical density is the highest density at which the maximum volume of an individual piece of timber is reached. VmaxAF is the maximum timber volume of an agroforest tree, reached at densities lower or equal to the critical density.
Table I: reference values in forestry and values of TGA, the critical density and VmaxAF for each of three tree species
Nevertheless, some unpublished experimental results are in favour of higher values for TGA: at M. Jollet’s farm (Les Eduts, Charentes Maritimes, France), INRA’s measurements of the forest and agroforest trees at the middle of the revolution indicate a TGA of 2 for black walnuts, at 80 trees/ha (Gavaland, pers. com.). But another thinning will soon accelerate the growth of the forest trees, and then this estimated TGA is likely to decrease.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 16 There is thus an important difference between our hypothesis and what we could expect (Figure 1).
2 with TGA = 2
with TGA = 1,2
1,5 timber (m3/ tree)
1 individual piece of
0,5 0 20406080100 Final density (trees/ha)
Figure 1: Volume of the individual walnut timber volume in function of the final density and of the value of TGA.
As an economic consequence of such a wise hypothesis, the volume of timber at felling is less important, thus the revenue of the tree component might be under-estimated.
Maximum expectable LER in function of species and final density
As the production of agroforest timber is determined in accordance with the densities, the critical density and TGA, the tree RA-products and the tree RA-biomass are fixed: it is impossible to tune them without modifying one of these previous parameters. Then the range of variation of the LER (biomass or products) corresponds to the crop RA:
• As a LER-biomass inferior to 1 is biologically unrealistic, the minimum value of the crop RA is equal to 1 – tree RA-biomass.
• As we assume lower or equal yields, the maximum value of the crop RA must be inferior or equal to the maximum intercropping area. In our optimistic assumptions, at highest densities (tree lines every 10 m), we assumed a crop RY at ¾ of the maximum intercropping area.
• A likely value would be the mean of these two extreme values.
As the proportion of land required by the trees strips rises with the density, the maximum crop RA decreases when the density gets higher. A first conclusion is that we obtain acceptable RA with densities which correspond to distances between the tree lines included between 24 to 40 m.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 17 1,1
1
0,9
0,8
0,7 LER max 0,6 LER-optimist LER-medium 0,5 LER-pessimist 0,4
0,3
0,2 No tree 0 20 40 60 80 100 120 140 Distance between the trees lines (m)
Figure 2: Range of values of the crop RA with walnut, wild cherry and poplar, according the distance of the tree lines and depending on how optimistic the dynamic of the LER is. In forestry, the realisation of many thinnings means that a lot of biomass is synthesised in addition of the trees which shall be conserved until the last fall. As we assume that the volume of the agroforest trees is maximum 20% bigger than the one of the forest trees, the production of woody biomass is small compared to the one of a forestry plot. Thus the ratio of woody biomass, i.e. the tree RA-biomass, is low. At low density, even an optimistic value of the crop RA is insufficient to compensate such a low tree RA. Consequently, high LER-biomass cannot be reached for all densities, in particular for species with a high rate of thinning in forestry such as wild cherry (Figure 3).
However, very satisfactory LER-products can be reached even with these species.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 18 1,6
1,4
1,2
Optimist Crop RA 1 Medium Crop RA LER - Products Pessimist Crop RA 0,8
0,6 0 20 40 60 80 100 120 140 Tree final density (trees/ha)
Figure 3: Expectable LER-products for walnut, cherry and poplar, depending on how optimistic the dynamic of the intercrop is
Impact of the TGA on the LER results
In our simulations, we used a TGA of 1.20. We were cautious in our predictions if we consider some experimental plots (such in Restinclières in France) or private site (Farm of Claude Jollet in Charente Maritime) where we observed some TGA which reach 2. If we had taken this value of 2, the tree RA would have increased between 15 to 30 % in comparison with what we obtained with 1.20.
1
0,9 y = 0,4191x + 0,1123 R2 = 0,9995 0,8
120 trees/ha - products 0,7 y = 0,2745x + 0,0955 2 R = 0,9987 120 trees/ha - biomass 0,6
0,5 50 trees/ha - products y = 0,2655x + 0,0018 2 TREE RA R = 0,9989 0,4 50 trees/ha - biomass
y = 0,1773x + 0,005 0,3 R2 = 0,9965
0,2
0,1
0 0,811,21,41,61,822,2 Tree Growth Acceleration
Figure 4: Influence of the Tree Growth Acceleration on the tree RA (biomass and products), according to the tree density.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 19 DATA REFERENCES AND MAIN HYPOTHESIS
The forestry references
The revolution duration, timber production and production techniques (initial density, prunings, thinnings, sward maintenance, and final density) were determined by expert knowledge, in accordance with available documentation.
The densities correspond to the schedule of conditions of the French circular “Forêts de production” and to forestry organisms’ advises.
density Revolution duration Mean annual production Individual (trees/ha) (years) (m3/ha/year) piece of Volume timber at at felling Good Land Bad Good Land felling (m3/ha) Bad initial final land unit land land unit (m3/tree) land unit unit medium unit unit medium
Walnut 1 200 100 100 46 53 60 2,17 1,89 1,67
Wild cherry 0,8 800 150 120 50 55 60 2,40 2,18 2
Poplar 1,5 200 200 300 19 22 25 15,79 13,64 12
Table 1: Densities, revolution duration and mean annual and total productions for walnut, wild cherry and poplar. With walnut, 2 thinnings of 50 trees/ha are realised at 1/3 and 2/3 of the revolution; with wild cherry, 3 thinnings are realised at 1/3 (400 trees/ha), half (200 trees/ha) and 2/3 (50 trees/ha) of the revolution. Supports for afforestation on agricultural land vary in function of the region and of the tree species: as the poplar revolution is shorter, the Compensation Payment for Agricultural Loss (PCPR) is available for 7 years instead of 10.
Type of farm Ly-Lc Hy-Hc Hy-Lc (Poitou- (Franche- (Centre) (region) Charentes) Comté)
Walnut and wild cherry
Establishment grant (4 first years) 50% of the costs 50% of the costs 0
PCPR farmer (10 first years) 240 €/ha 300 €/ha 0
Poplar
Establishment grant (4 first years) 50% of the costs 50% of the costs 0
PCPR farmer (7 first years) 240 €/ha 300 €/ha 0
Table 2 : Regional supports for afforestation on agricultural land for walnut, wild cherry and poplar (year 2003). The PCPR is the Compensation Payment for the loss of agricultural income. Franche-Comté is a particular region. More than 50% of the area are already woodlands, thus afforestation is not encouraged: there is no support available for new forestry plantation.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 20 Everywhere in France, newly afforested plots benefit from an exemption from land tax: for 10 years with poplar, 50 years with walnut and wild cherry. In our simulations, this land tax is comprised between 30 €/ha (Centre) and 39 €/ha (Poitou-Charentes).
Reference data in agriculture
All arable data come from the Farm observatory ROSACE, a tool produced by APCA. Thanks to this typology of farms made by the regional Chamber of Agriculture, several types of farms are defined and described, each one corresponding to the mean of 5 to 10 farms selected by the Chambers experts. Each year, the economical inputs are re-calculated (yield, net margin, farm costs, labour and CAP payment). In addition, all the technical orientations and strategies of the farm are also described.
We selected 3 types of farm, which we shall now designate with 4 initials:
• Hy-Lc: High yields and Low fixed costs
• Hy-Hc: High yields but High fixed costs
• Ly-Lc: Low yields and Low fixed costs
For each of them, the ROSACE typology indicates:
• The cropping area of the farm, distinguishing tenant farming and property;
• The crop rotation in function of the quality of the soil (up to 3 Land Units: best, medium, worst);
• The mean yields, attributed to the medium Land Unit (for the best and worst Land Units, we respectively assumed an increase and a decrease of 10% of the mean yields);
• The variable costs, assignable fixed costs and fixed costs and labour.
• The prices of the products and sub-products (straws of the wheat) and the CAP payments of the farm Single Farm Payment, SFP).
To elaborate the selection of each type of farms, various partners from the Chambers of Agriculture have participated: Camille Laborie, who is in charge of ROSACE in APCA, Anne-Marie Meudre (Franche Comté), Catherine Micheluzzi (Poitou- Charentes) and Benoît Tassin (Centre).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 21 Cropping area of the farm (ha) Gross Total Fixed Mean Net Typical Margi area Net costs Crop yield Marg
rotation(s) n (ha) Margin (€/farm) (t/ha) (€/ha) (€/ha) (€/farm) Total tenant farming property
8 (a) wheat wheat (straw 983 776 42,0 wheat 2 t/ha) oilseed 72 8 80 36805 18045 or oilseed 4 918 711 30,0
Hy-Lc (b) wheat oilseed set – 323 116 8,0 aside
6,5 wheat (straw 818 566 42,3 2 t/ha) wheat wheat oilseed 3,2 674 422 14,1 sunflower 56,4 37,6 94 28785 17153 wheat sunflow Ly-Lc oilseed 2,5 799 547 28,2 sunflower er set – 318 66 9,4 aside
(a) wheat 6,7 wheat wheat (straw 794 479 87,8 wheat 2 t/ha) wheat wheat oilseed 3,5 728 413 14,1 97,5 32,5 130 40370 12031 maize
Hy-Hc or maize 7,5 555 240 28,2 (b) wheat wheat set – 313 -2 9,4 oilseed aside
Table 3 : Main economic data and total net margin (€/farm) for every type of farm. Rotation (a) corresponds to the best land units, rotation (b) to the worst. Set aside is realised on 10% of the total farm area.
The Net Margin is equal to the Gross Margin minus the fixed assignable costs (land tax and machinery costs). The Total Net Margin is equal to the Net Margin minus the fixed costs (rent of land, amortisation and maintenance of the buildings, social contributions, banking costs). Labour costs are not taken into account.
The profitability threshold yield
With the development of the trees, the crop yield decreases progressively. Below a certain level, the crop is not more profitable, above al near the tree area. For each crop of the three types of farm, the threshold yield was first determined according to
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 22 the price of the product, the CAP payment and the variable costs, assignable fixed costs and a part of the fixed costs1. As the results, in proportion of the mean yield of each crop, were roughly the same in the three farms, we fixed this proportion in order to facilitate the extrapolation to other types of farm.
Mean yield Profitability Mean yield in Profitability threshold crop in the farm threshold yield Hy-Hc yield in Hy-Hc
Winter wheat 100 % 50 % 6,7 t/ha 3,35 t/ha
Maize 100 % 70 % 7,5 t/ha 5,25 t/ha
Oilseed rape 100 % 60 % 3,5 t/ha 2,1 t/ha
Table 4: Profitability threshold yield in proportion of the mean yield in the farm and example for the farm Hy-Hc
The threshold yield is the same in every farm, whatever the land unit is. Thus it shall be reached more quickly in the worst land unit than in the best land unit.
Main management features of the agroforestry systems
For each type of farm, we simulated the introduction of 2 agroforestry designs in the 3 land units (best, medium, and worst):
• Plantation at 50 trees/ha, on 40 m spaced tree-lines;
• Plantation at 120 trees/ha, on 22 m spaced tree-lines.
The tree strip is 2 m wide. The width of the intercropped alley is respectively of 38 m and 20 m, thus the maximum crop area represents 95% of the initial area at 50 trees/ha and 91% at 113 trees/ha.
With walnut and wild cherry, an early thinning is realised when the timber volume reaches 0,1 m3 (around the years 10-13), therefore the final densities are different from the poplars’ one (see Table 5).
1 If the crop is abandoned on a part of the cropping area, we assume that the fixed costs should decrease a little; thus they must be taken into account in the calculation of this threshold yield.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 23 Agroforestry Forestry Tree Density Density Tree RA- Timber Timber RA- (trees/ha) Productio (trees/ha) Productio products volume volume biomass (m3/ha) (m3/ha) (m3/tree) (m3/tree) initial final initial final
50 40 1,2 48 0,36 0,48 Walnut 200 100 1 100 120 80 1,08 86 0,66 0,86
50 40 0,96 38 0,22 0,32 Wild 800 150 0,8 120 cherry 120 80 0,92 74 0,42 0,62
50 50 1,8 90 0,3 0,3 Poplar 200 200 1,5 300 120 113 1,76 199 0,66 0,66
Table 5: Initial and final densities, volume of an individual piece of timber and production in forestry and in the simulated agroforestry systems; tree Relative Area (RA)-biomass and tree RA-products
The crops Relative Areas (RA) have been fixed for 3 hypothesis: optimistic, probable and pessimistic.
The pessimistic hypothesis means that the LER-biomass is equal to 1. Therefore, the crop RA is equal to: (1 – tree RA-biomass).
The optimistic crop RA is determined according to 2 constraints:
• The crop RA must be inferior to the maximum intercropping area
• We also assumed to fix a ceiling for the LER-biomass of 1.4. Thus the crop RA is equal to: (1.4 – tree RA-biomass). This ceiling of 1.4 was reached with walnut and poplar at 120 trees/ha, so the crop RA seems quite low with regards to the maximum intercropping area.
We assumed a probable crop RA as the arithmetic average of the 2 previous values (pessimistic and optimistic) (see Table 6 and Table 7).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 24 Width Initial Width of the Maximum between Pessimistic Probable Optimistic density intercropped intercropping tree lines crop RA crop RA crop RA (trees/ha) alley (m) area (m)
50 40 38 0,95 0,64 0,79 0,94 Walnut 120 22 20 0,91 0,34 0,54 0,74
50 40 38 0,95 0,78 0,85 0,93 Wild cherry 120 22 20 0,91 0,57 0,72 0,87
50 40 38 0,95 0,7 0,8 0,9 Poplar 120 22 20 0,91 0,34 0,54 0,74
Table 6: Crop RA in function of the tree species, density and optimism level. Bold values are those which depend on the ceiling of 1.4 for the LER-biomass.
Width Width of LER-biomass reached with LER-products reached with Initial between the the the density tree intercrop (trees/ha) lines ped alley Pessim. probabl Optimist Pessim. probabl Optimist (m) (m) crop RA crop RA crop RA crop RA crop RA crop RA
Walnut 50 40 38 1 1,15 1,3 1,12 1,27 1,42
120 22 20 1 1,2 1,4 1,2 1,4 1,6
Wild 50 40 38 1 1,07 1,15 1,1 1,17 1,25 cherry 120 22 20 1 1,15 1,3 1,19 1,34 1,49
Poplar 50 40 38 1 1,1 1,2 1 1,1 1,2
120 22 20 1 1,2 1,4 1 1,2 1,4
Table 7 : LER-biomass and LER-products in function of the tree species, density and hypothesis of optimism for the intercrop bold values are those which depend on the ceiling of 1.4 for the LER-biomass.
Economic hypothesis
CAP payments
In agriculture, the crops area benefits from the Single Farm Payment (SFP): it was calculated on the basis of the historical references of each farm, in accordance with the way France decided to implement the new CAP in 2006.
In the basic scenario, we assumed that the intercrops are eligible to the SFP proportionally to the area of the plot that they occupy. It is the present situation in France. The rights corresponding to the tree area could be transferable to another
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 25 eligible area which doesn’t benefit from a payment right. In our simulations, we did not attribute them to new plots, considering therefore that these rights were lost for the farmer.
Tree grants
In our basic scenario, agroforest trees benefit from the same establishment payments as the forest trees: 50% of the costs of the 4 first years in Poitou-Charentes and Centre. It corresponds to the present situation, permitted by the circular “Forêts de protection” which relies on the line i of the French National Rural Development Programme. However an agroforest plot cannot benefit from neither the PCPR nor the exemption of land tax.
In France, an agro-environmental measure called “agroforest habitats” can be contracted under certain conditions, but it still faces administrative difficulties and is not available in most of the departments, thus it was not taken into account in our simulations.
Costs and prices
Some key points have to be underlined:
• The cost of sward maintenance is higher in forestry than in agroforestry. In forestry, at the beginning of the revolution, sward maintenance is realised thanks to two grindings instead of one for the maintenance of the tree strip in agroforestry.
• The farmer makes all operations himself, except the marking out and plantation of the young trees. Both of these operations are charged 15 €/h. The timber prices correspond to standing trees, thus neither the harvesting cost is taken into account.
• In a cash flow approach, the basic scenario doesn’t include the labour cost for the farmer. While in a farming management scenario, we consider an hourly cost of 7,62 €/h (minimum salary in France). In this last approach, it’s therefore possible to evaluate the efficiency of the farmer labour.
As it seems impossible to anticipate the future evolution of prices and costs, we assumed constant values. For instance, a rise or a drop of timber value would respectively increase or decrease the tree revenue.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 26 MAIN RESULTS
Labour impact for one silvoarable hectare
temps de travail temps de travail (h/ha/an) (h/ha/an) 16 16 agriculture agriculture 14 cultures intercalaires 14 cultures intercalaires arbres arbres 12 12
10 10
8 8
6 6
4 4
2 2
0 0 1 6 11 16 21 26 31 36 41 46 1 6 11 16 21 26 31 36 41 46 année année
Case 1: Plantation of 120 trees/ha Case 2: Plantation of 50 trees/ha
Figure 5: Labour evolution in the management of a silvoarable plot during the tree rotation, separating the crop from the tree labour.
An essential condition for adopting agroforestry from the farmers’ point of view is that they don’t want to devote more time to a new system. If the farmer planted more trees (case 1), he would need 1 to 1.5 days each year to maintain the trees. But in the second half of the rotation, the labour decreases progressively due to the fact that trees don’t need more special maintenance and that the intercrop activity is reduced. If he plants fewer trees, the impact during the first years is poor. With the small density, the intercrop activity is longer, because the crop yield is not so affected by the trees. The labour requested in the second half of the rotation is therefore lower but very near from the initial scenario.
Prediction of yield evolution
Crop yield evolution
Predicting the crop yield during the second half of the rotation is a perilous venture. If we know the behaviour of the intercrop during the first half thanks to experimental measures on existing plots, we asked the bio-economics model to predict the yield evolution. In our simulation, as we said, we used the LER-Safe prediction. We made the essential hypothesis that the LER must be include between 1 and 1.4. This condition helps us to determine a possible range of crop yield evolution, from the pessimist one to the optimist one (see Figure 6).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 27 Tree plantation Tree Harvest
100 100 Optimist 80 80
60 60 Pessimist
40 40
Crop yield (%) Crop yield Intercrop Yield
20 Pure Crop 20
0 0 0 Time
Figure 6: Evolution of the relative intercrop yield according to optimist or pessimist view about the tree competition. Case of one ha of wild cherry with an initial density of 120 trees/ha for a final density of 80 trees/ha.
In this example of a plantation of wild cherry at 80 trees/ha (final density), which means a distance between the trees rows of 25m, the crop yield represent more than 90% during the first half of the tree rotation. According to the interaction level, the crop yield varies between 30 and 75 % of the pure crop yield of reference the year before harvesting the trees.
The crop yield depends on different parameters:
• The parameters due to some initial choices: the crop nature (a sunflower will be more affected by the shadow of the adult trees than a cereal), the density of the plantation and the distance between the lines, choice of the land unit (a deeper soil will be more adapted),...
• The parameters depending on the capacity of the farmer: well pruned trees, tree root maintenance (root cutting), …
In our economical scenarios, we have tested the different level of interaction.
Tree yield evolution
As for the crop yield estimation, we put forward the hypothesis of different level of timber productivity. But for our simulations, we only use one prediction of timber production. To validate our approach, we use a very cautious estimation of production (see Figure 7). Our results can therefore be considered as the minimum result we can get from our hypothesis.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 28 140 140 Interval 120 120 Basic 100 scenario 100 Pure Optimist 80 plantation 80
60 60
40 40 Pessimist Standing volume (m3/ ha) volume Standing 20 20
0 0 5 1015202530354045 Time
Figure 7: Range of timber volume evolution for an initial plantation of 120 wild cherry. The figure indicates of the cautious hypothesis of standing volume we used for our simulations (77 m3 for 80 final trees).
Cash flow impact
To evaluate the impact of the project on the cash flow, we must distinguish first the investment cost and then the evolution of the annual cash flow depending of the crop yield evolution and the possible over cost to crop between the trees in comparison with a pure crop system.
Initial investment
The poor number of trees to plant in an agroforestry system reduces considerably the investment cost if we compare with a current afforestation cost on agricultural land. The tree cost is nonetheless higher. The owner will choose a better quality of the trees and will have to protect each one with a strong protection: each tree has a possible future value and demands a special attention.
The total cost of a plantation (without subsidy) varies between 500 and 1000 euros/ha according to the tree specie (the walnut plantation being the most expensive). This cost represents between 20 to 60 % of the average cost in the case of common land afforestation (see Figure 8).
Afforestation
1 233 €/ha 120 trees/ha Poplar 695 €/ha 367 €/ha 50 trees/ha
1 518 €/ha Wild Cherry 469 €/ha 267 €/ha
1 633 €/ha
Walnut 1 034 €/ha 517 €/ha
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 29 Figure 8: Comparison of the investment in agroforestry and forestry scenario, WITHOUT subsidy.
In France, it’s current to get a subsidy of 40 to 70% to cover the investment cost and the maintenance cost of the trees during the 4 first years (except in Franche Comté).
Since 2004, the French Government decided to suspend all economic aids to the land afforestation, excepted for agroforestry. In our simulations, we decided to conserve this aid, to be able to compare between the two options (see Figure 9).
Afforestation
617 €/ha 120 trees/ha Poplar 348 €/ha 184 €/ha 50 trees/ha
759 €/ha
Wild Cherry 235 €/ha 134 €/ha
817 €/ha
Walnut 517 €/ha 259 €/ha
Figure 9: Comparison of the investment in agroforestry and forestry scenario, WITH subsidy.
Cash flow evolution
Evolution of the cash flow at the plot scale
The cash flow evolution will depend of the crop yield evolution and the LER level we have selected and the final density. For example, in the Figure 10, we’ve illustrated the cash evolution for two different densities but for a medium LER level.
100 100
90 50 trees/ha 90 80 to 90 % 80 70 to 85 % 80
70 70 120 trees/ha 60 60
50 30 to 60 % 50
40 40
30 30 % Annual Gross Margin Gross Annual % 20 Silvoarable Gross Margin 20 Agricultural Gross Margin 10 10
0 0 Time Trees Harvesting Plantation
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 30 Figure 10: Evolution of the annual cash flow for a probable scenario with wild cherry (LER=1.07 for a density of 50 trees/ha and 1.15 for a density of 120 trees).
Being cautious in our forecast, we notice nonetheless that at half of the rotation, the gross margin still represent 80 to 90 % of the agricultural gross margin. We must underline that in our simulations, we’ve considered that the crop payment area is reduced progressively by the tree area. In case of the silvoarable area was eligible in its totality, the impact on the cash would be sensible, above all in some regions where man get poor crop yield and where the crop payment is essential in the gross margin calculation (Franche Comté for example).
Let’s also underline the fact that in the INRA experimental plots, the LER reaches more 1.3 than 1.15 that we have chosen in our simulation with an initial density of 120 trees/ha.
Influence of the CAP payment policy
Inside the first pillar policy, the situation of the agroforestry plots could be different depending of each country member. In fact, at a European level, the agroforestry plot could be eligible to the Compensatory Payment. We compare here the possibility to get the payment on the whole area (Request of the Safe consortium) or only on the intercrop area (French situation).
The impact of the eligibility given to the whole surface on the profitability is not so important. In all our simulations, the profitability increases by 3% in the best option for agroforestry. The impact is more at a cash flow level, when the crop gross margin is low. That’s typically the case for the farms where:
• The crop component is lower than the payment component in the gross margin calculation (Mediterranean area or farm with high cost of production)
• The yield is decreasing faster in the silvoarable scenario (high density of plantation or strong impact of the trees on the crop RA) (see Figure 11)
100%
80%
60% agriculture Payment on intercrop area - 50 trees/ha 40% Payment 100% - 50 trees/ha Payment on intercrop area - 120 trees/ha Payment 100% - 120 trees/ha
% of the Arable Gross Margin 20%
0% 2% 12% 22% 32% 42% 52% 62% 72% 82% 92% Time (Tree rotation)
Figure 11: Influence of the different CAP payment policies in agroforestry on the annual cash flow evolution.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 31 Evolution of the cash flow at the farm scale
At the farm scale, one of the first questions of the farmer is about the importance of the area to plant. Does he have to plant on a big area? In several plots or in a single plot? There is no only one answer. According to the strategy of the farmer, a large range of scenarios is available. The choice will depend to the cash flow context and to know if the farmer can support a strong investment or not, and above all if he aims to decrease progressively his crop activity or not. The labour availability is also a strong parameter to decide which area to plant. According to our simulation and experimental experience, we often recommend not planting more than 10 % of the cropping area. In that case, the impact on the farm gross margin is less than 3 % in average on the first half of the tree rotation. A gradual plantation will allow a reduction of the cash flow impact (see Figure 12).
435 % 191% 178% 180% 183% 175 Farm with 8% silvoarable area 175 Farm with 8% silvoarable area Farm with 100 % of cropping area Farm with 100 % of cropping area 150 150
125 125
100 100
75 75 % of Farm Gross Margin without AF % of Farm Gross Margin without AF 50 50 0% 20% 40% 60% 80% 100% 120% 0% 20% 40% 60% 80% 100% 120% % of the tree rotation % of the first tree rotation a. Case of a single plantation b. Case of a gradual plantation
Figure 12: Comparison of the cash flow evolution when the farmer plants 8 % of his cropping area (50/50 Walnut/Wild cherry). We compare the option where the farmer would plant the silvoarable area in once time or if he decides to plant 2 % every 5 years during 20 years.
A gradual plantation will also allow a soft distribution of the timber income in the time from the moment where the owner begins to harvest the first mature trees (case b). From this moment, the timber income is regular. In our example, he can harvest the trees every 5 years. In this context, the farm gross margin increase by 15 %. According to the importance of the plantation and of the species he planted, a farmer could increase his farm income between 10 to 100%. Of course, it can suppose a long term to wait for the farmer before the first tree harvest…
Profitability of a silvoarable investment
Comparing a silvoarable scenario with agricultural scenario
For our simulations, we have selected 3 kinds of farms:
• Farm with good crop yields and few fixed costs.
• Farm with medium crop yields with few fixed costs.
• Farm with medium crop yields and high fixed costs.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 32 For each farm, corresponding to each region of the LTS of the WP8, we have run different scenarios according to:
• the tree density: 120 versus 50 for the initial density which corresponds to a final density of 80/40.
• the LER level: optimist, probable and pessimist
• the land unit: good/medium
• the 3 tree species: poplar, walnut and wild cherry
108 scenarios have been run in total (36 scenarios / LTS). The Figure 13 shows a synthesis of the Agricultural Values for all these scenarios we have calculated for each specie according to the level of LER.
Walnut Wild Cherry Poplar 100% Agricultural 80% Value Index > 1,35 60% 1,20 - 1,35 1,05 - 1,20 40% 0,95 - 1,04 20% < 0,95 % of realised simulations realised of %
0%
st le st t i b i is tim a tim m p b p o ro optimist robable o p pessimist p pessimist probablepessi Scenario for intercrop productivity
Figure 13: Profitability of the silvoarable scenarios according to the tree specie and the LER level.
A first interesting result is that the silvoarable scenarios are at least as profitable as the agricultural scenario.
Walnut timber is actually the most expensive timber on the market. For a same duration of rotation, the best results have been logically obtained with the walnut than the wild cherry. The period of harvesting time is a key parameter in the profitability calculation (see Figure 14).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 33 2,00 1,80 1,60 1,40 1,20 1,00 0,80 0,60 0,40 0,20 0,00 VeryTrès well bien prunedBien Well prunedformé Mal Badly formé pruned formé (50 ans) (60 ans) (4040 years ans) 50 years 60 years
Figure 14: Influence of the maintenance quality on the profitability.
A late in the pruning dates can put the harvesting date back by 10 or 20 years, above all for some sensitive specie such as the hybrid walnut. In this example, a late of 20 years means a reduction of 60% of the profitability in comparison of the agricultural profitability.
Influence of the TGA on the Agricultural Value
The value of the Tree Growth Acceleration has a strong impact on the profitability of the silvoarable scenarios. This impact is stronger for the scenario with higher densities of plantation. In the following figure, we noticed that the scenario with a density of 120 ha react much quicker than a scenario with 50 trees.
Again, in our simulations, we used a TGA of 1,20 which could be considered as a pessimist approach with what we observe in the reality. For example, in the Jollet's case, the agricultural value would have been increased by 10 to 15 % (see Figure 15).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 34 Jollet TGA 1,10
1,05
Hypothesis simulation 1,00
120 wild cherry/ha 0,95 Indice of Agricultural Value 50 wild cherry/ha
0,90 0,8 1 1,2 1,4 1,6 1,8 2 Tree Growth Acceleration
Figure 15: Influence of the TGA on the Agricultural Value
Which density to plant to get the best profitability?
A common question from the farmers is about the number of trees to plant. The farmers often want to maintain a correct crop yield during the whole rotation but trying in the same time to get the best investment for timber. Other decides to plant more trees with the aim to decrease the agricultural activity, even till to suppress the intercrop. We didn’t take this case in this study.
For each specie, Walnut, Wild cherry and Poplar, according to our production hypothesis, we simulated the impact of the density to the LER but also to the Agricultural Value (see Figure 16).
1,6
1,4 LER_optimist LER_medium 1,2 LER_pessimist
Val-agri_optimist 1 Val-agri_medium Val-agri_pessimist 0,8
0,6 0 20 40 60 80 100 120 140 WILD CHERRY - Finale Density (trees/ha)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 35 1,6
1,4 LER_optimist LER_medium 1,2 LER_pessimist Val-agri_optimist 1 Val-agri_medium Val-agri_pessimist 0,8
0,6 0 20406080100 WALNUT - Final Density (trees/ha)
1,6
1,4 LER_optimist
1,2 LER_medium LER_pessimist 1 Val-agri_optimist
Val-agri_medium 0,8 Val-agri_pessimist
0,6
0,4 0 50 100 150 200 POPLAR - Final density (trees/ha)
Figure 16: Influence of the tree density on the LER value and the Agricultural value for wild cherry, walnut and poplar.
We observe that for each specie, the best density to get the optimum LER is higher than the best density to get the optimum Agricultural Value. For the species with a poor Tree RA (Walnut and wild cherry), the range of density are similar (see Table 8). The best density would vary between 80 to 120 trees/ha to get the highest LER, while the farmer will get the best profitability with a density included between 60 and 90 trees/ha. Of course, with a higher TGA, this range would increase.
Result Wild Cherry Walnut Poplar
LER 80 - 120 80 - 120 130 - 200
Agricultural Value 60 - 90 60 - 90 100 - 130
Table 8: Range of density to get the optimum LER and Agricultural Value results for each specie (trees/ha – final density).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 36 For the poplar, the optimum densities are higher than for the 2 others species. This result is due to the fact that the biomass produced by the silvoarable poplar is similar to the biomass produced by the forestry poplar. The Tree RA is therefore higher for a given density compared to other species which demand more important fellings.
What could influence these results? As we already said, the TGA level could strongly influence these results, giving priority to higher densities. The policy schedule and the price level of the crop and tree component will be therefore the most important parameters. In the case of the walnut, the choice of a density of 75 trees/ha is a wise option. Below, the farm doesn’t want to take any risk at a long term period, above he bets more on the trees.
Comparing a silvoarable scenario with a forestry scenario
We compare also the case where the farmer was hesitating between a forestry investment rather than a silvoarable investment from a profitability point of view (see Figure 17).
Agricultural Value Index 1,50 of a silvoarable scenario of a pure plantation scenario
1,00
1,55
1,21 0,50 1,04 1,00 0,89
0,48
0,00 Poplar Walnut Wild Cherry
Figure 17: Comparison of the profitability of the silvoarable and afforestation scenario with the agricultural scenario. Silvoarable plantation of 120 wild cherry by ha characterized by a LER of 1,15.
In this example, we explore the case of a probable LER of 1.15 in the silvoarable option. In almost all our simulations, the silvoarable options are more profitable than the forestry option. The forestry option may be more profitable in the case where the crop margin is very poor, above all if it’s possible to plant some valuable species such as walnut for example.
It’s also interesting to notice that for the poplar, the silvoarable option could be a possibility to stimulate the poplar market. In France, the poplar area is currently decreasing because of the price fall of the timber (less than 45 €/m3). Agroforestry could therefore be a possible strategy to reduce the market risks.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 37 Property holdings evaluation in agroforestry
According to his age, a land owner who plants trees, will not necessary benefit from the harvest… But, as a farmer told us, a farmer has three possibilities of income: the sale of his products, the stock variation and the possibility to make a capital gain. In this last option, a silvoarable plot is a capital which could be evaluated if necessary (inheritance, expropriation, etc). The land evaluation in agroforestry is the combination of the agricultural land evaluation and the future value of the trees (see Figure 18).
16 000 € No commercial value with commercial value 14 000 €
12 000 €
10 000 €
8 000 €
Euros by ha 6 000 €
4 000 €
2 000 €
0 € 10 20 30 40 Age of the trees (years) Agriculture agroforestry
Figure 18: Evolution of the monetary value of the silvoarable land according to the age of the trees. In agroforestry, this value is the sum of the agricultural value plus the timber future value. If the young trees could have a future value, for example at 10 years old, they don’t necessary have a commercial value in the sense that the landowner can not expect some income if he cut them.
In this example of a wild cherry plantation, the capital evaluation may represent between twice and four time the agricultural land value according to the age of the trees. In the case of a walnut plantation, it may represent till 7 times this value 10 years before the tree harvesting.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 38
Photo 1: In this plot of 4 ha, the wild cherries are 30 years old. The value of the standing volume is estimated to 4 000 €/ha, which represents the same value of the agricultural land. But the future value of this plantation is much higher and overpass the 10 000 €/ha.
Main conclusions
To invest in agroforestry represents a light investment in money and labour comparing with some new systems of diversification. In our simulations, the profitability reaches 10 to 50 % with walnut, and -5 to +15 % with wild cherry and poplar, comparing with the agricultural scenario.
A regular calendar of plantation on a few surfaces is a good option for the farmer (labour and cash flow impact). 10 % represents between 2 and 3 % of reduction of the farm gross margin. But in the balanced period, the income increases by more than 15% (mixed plantation of walnut and wild cherry trees). The gross margin could double if the farmer plants progressively his whole cropping area. But in that case, it means a stronger impact on the initial cash flow and demands a consequent labour...
If the best bio-physical option is to plant between 80 to 120 trees by hectare (130 to 200 for the poplars), the best economical option is to plant a lower density around 60 to 90 trees by hectare (100 to 130 for the poplar). This means a distance between the trees lines varying between 24 to 36 m.
All our simulations haven’t taken into account the environmental benefits such the carbon sequestration, or the impact on the nitrogen pollution. These aspects could be calculated and to be summed to the whole profitability of the silvoarable systems.
BIBLIOGRAPHY
Borrell, T. (2004) De l’importance des interactions arbres-cultures sur les performances économiques de l’agroforesterie tempérée. Mémoire de Diplôme d’Agronomie Approfondie, ENSAM-INRA, Montpellier. 98 p + annexes
Boulet-Gercourt, B. (1997) Le merisier. IDF, 2ème édition. 128 pp.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 39 Coulon F, Dupraz C., Liagre F., Pointereau P. (2000) Etude des pratiques agroforestières associant des arbres fruitiers de haute tige à des cultures et pâtures, Rapport au ministère de l’environnement, 199 p, Solagro/INRA, Fr
CRPF (1997) Boiser une Terre Agricole. 28 pp.
Dupraz C., Lagacherie M., Liagre F., Boutland A., (1995). Perspectives de diversification des exploitations agricoles de la région Midi-Pyrénées par l’agroforesterie. Rapport de fin d’étude commandité par le Conseil Régional Midi-Pyrénées, Inra-lepse éditeur, Montpellier, 253 pp.
Dupraz C., Lagacherie M., Liagre F., Cabannes B., (1996). Des systèmes agroforestiers pour le Languedoc-Roussillon. Impact sur les exploitations agricoles et aspects environnementaux. Inra-Lepse éditeur, Montpellier, 418 pp.
Dupraz, C., Liagre, F. & Borrell, T. (2005) The Land Equivalent Ratio of a silvoarable agroforestry system. In preparation.
Graves, A.R., Burgess, P.J., Liagre, F., Dupraz, C. & Terreaux, J.-P. (in preparation) The developmentof an economic model of arable, agroforestry and forestry systems. To be published soon in Agroforestry Systems.
IDF (1997) Les noyer à bois. 3ème édition, Février 1997. 132 pp.
Liagre F., (1993). Les pratiques de cultures intercalaires dans la noyeraie fruitière du Dauphiné. Mémoire de Mastère en Sciences Forestières, ENGREF, Montpellier, 80 pp
Segouin O., Valadon A., (1997) Enquête sur les boisements récents de peupliers en Lot-et- Garonne, Analyse de pratiques agroforestières ; les cultures intercalaires. Cemagref, Nogent-sur Vernisson, 45 pp.
Souleres, G. (1992) les milieux de la populiculture, IDF, 310 pp.
Terreaux, J.-P. & Chavet, M. (2002) Problèmes économiques liés à l’agroforesterie : éléments qualitatifs et quantitatifs. Silvoarable Agroforestry For Europe (SAFE) ; Cabinet Michel Chavet, Paris – UMR Lameta, Montpellier.
Vandermeer, J. (1989) The Ecology of Intercropping, Cambridge University Press, 225 pp.
ANNEX
Annex 1: Detailed description of the LERbased-Generator
Principle
Farm-SAFE does not have any biophysical module, the time-series must be generated independently: pure crop and intercrop yields, timber production in forestry and agroforestry. We used a generator constrained by the LER-biomass: depending on a previously fixed value and on a quite low number of parameters, these times- series are produced. The starting and final points are known, the evolution between them is drawn thanks to a logistic equation.
A key characteristic is that the climatic variability is not taken into account. It would have necessitate to define the impact of variables (temperature, water, light, etc...)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 40 which are not implicated in this type of constrained prediction. Nevertheless, we assume that except in very particular cases, this variability does not have any impact on the economic results: on a whole revolution (20 to 60 years), “bad weather” years are compensated by “good weather” ones, as we are not interested in year-by-year results but final profitability and global evolution of financial results. Because of discounting, climate would only have a strong effect if “bad weather” years were concentrated in a specific part of the revolution, which is very unlikely to happen.
Notations
We use the words “forestry” and “forest trees” for all types of pure trees plantation, even when the initial density is very low, such as for walnut.
A distinction is made, in forestry and agroforestry, between the trees which are cut at thinnings and the trees which are maintained until last felling: the firsts are called “thinned trees”, the others “felled trees”.
We call “timber” the bole of the tree, which has the highest commercial value. The same word is used for the thinned trees, even if the bole is often too small to be sold as good timber.
VF is the individual forest tree timber volume at forestry reference density.
VAF is the individual agroforest tree timber volume, depending on the density.
VmaxAF is the maximum individual agroforest tree timber volume.
Parameters
• Parameters per tree species
- DC, the critical agroforestry density, i.e. the density at which the tree growth potential is attained: the individual agroforest tree timber volume is equal to VmaxAF, the trees cannot be bigger, even at lower densities.
- TGA, the coefficient of individual Tree timber volume Growth Acceleration in low density AF conditions, or Tree Growth Acceleration; e.g. 1,2 indicates that the individual agroforest tree timber volume at a lower or equal density than DC will be 20 % bigger than the one of a forest tree, due to the positive impact of both low density and intercropping (exceeds of nitrogen, less competition than the perennial vegetation classically established between forest trees, etc…).
- Timber To Biomass in forestry, e.g. the timber contribution to the total biomass of a young forest tree (TTByoung-F) and of a felled forest tree (TTBfell- F);
- Timber To Biomass of a felled agroforest tree (TTBfell-AF);
- maximum value for the forestry ratio: biomass of all the thinned trees/biomass of all the felled trees;
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 41 - individual tree timber volume of the agroforest tree at thinning.
• Parameters of the logistic curves
- curvature and inflexion for the individual forest tree timber volume, for the individual agroforest tree timber volume;
- curvature and inflexion for the height of the forest trees, of the agroforest trees;
- curvature for the intercrop yields.
• Parameters used only as Farm-SAFE entries2
- final tree height (same in forestry and agroforestry);
- maximum bole height (same in forestry and agroforestry);
- fixed value of the ratio: firewood volume/timber volume in forestry, in agroforestry.
Entries
- arable rotation, reference yield and threshold yield for profitability for every crop;
- tree species, revolution duration (60 years maximum);
- forestry: reference production, initial density, years of thinnings (maximum 5) and numbers of thinned trees;
- agroforestry: initial density, number of trees cut in the unique thinning, plot design (distance between tree lines, initial width of the intercropped alley, width of the intercropped alley reduction step);
- LER-biomass aimed.
The generation of data-sets
• The first step is the generation of the time-series of the monocropping systems:
The time-series of pure agricultural yields are produced simply by repeating the reference yields as many times as necessary to last the duration of the revolution.
The time series of the timber production of the felled trees are generated with the following logistic equation:
2 These three parameters are not used in the generation of tree production data-sets (timber), but they are needed as entries for Farm-SAFE (tree height and production of firewood).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 42 Y initial − Y final Y = + Y final t courbure t 1 + T inf lexion
where Yt is the value of Y at year ;
Y initial and Y final are the initial and final values of Y;
T inflexion is the inflexion date (end of linear growth).
Y initial is equal to zero and Y final to any value, as the curve is then distended in order to go through a point {X’;Y’}: X’ is the date of fell of the trees and Y’ is the reference production (in m3/ha of timber at felling).
For forestry, 3 other time series are generated :
- The timber production of the thinned trees3, as it is considered that they can be smaller than the felled trees. The same logistic equation is used, with an Y’ calculated in function of 2 constraints:
(i) thinned tree timber volume ≤ felled tree timber volume.
(ii) the parameter “maximum value for the ratio: biomass of all the thinned trees/biomass of all the felled trees”
- The biomass of both the felled and the thinned trees, thanks to the ratio Timber To Biomass. We assume that if TTBF may vary with time t, it is a linear variation:
TTBfell−F −TTByoung−F TTB F ()t = × t +TTByoung−F T
where T is the revolution duration.
The biomass of the felled trees and of the thinned trees is thus calculated by dividing their respective timber time series by TTBF(t).
• The Relative Areas calculated in function of the Tree Growth Acceleration:
The coefficient of individual Tree timber volume Growth Acceleration in low density AF conditions, or Tree Growth Acceleration (TGA), permits the calculation of the agroforest tree timber volume:
- At D ≤ DC, VAF = VmaxAF
- At D = DF, VAF = VF
3 There is only one time series for all the thinnings: late-thinned trees have the same rate of growth as early-thinned trees.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 43 - D > DF is not possible
VF −V max AF V max AF −VF - At DC < D < DF, VAF = VF +D× − DF × DF − DC DF − DC
Where: D, DC and DF are the actual density, the critical density and the forest density
VAF, VF and VmaxAF are the actual agroforest tree timber volume, the forest tree timber volume and the maximum agroforest tree timber volume, with VmaxAF = TGA x VF
On the basis of the final densities in forestry and agroforestry, the forestry RA- products can then be deduced.
With TTBfell-AF, we easily know the agroforest trees biomass at f elling.
With regards to the low initial density in agroforestry, we assume that the thinned trees grow as well as the felled trees: the thinning is early enough to avoid a strong effect of competition, thus the individual thinned tree timber volume is the same as the one of a felled tree at that time. And as the number of thinned trees is low and the thinning quite early, the volume of thinned biomass is poor enough to permit us to consider a fixed TTBAF in time. Thus the volume of thinned biomass in agroforestry is calculated by dividing the thinned timber production by TTBfell-AF.
The forestry RA-biomass can then be calculated.
The arable RA-biomass is deduced in function of the aimed LER-biomass. It is equal to the arable RA-products, as we assume that the proportion of grain in the biomass of the crop is the same in agriculture and in agroforestry.
• The generation of the agroforestry data-sets:
The agroforestry timber time-series are generated with the same logistic equation: Y’ is then the forestry reference production multiplied by the forestry RA-products.
Until the thinning, the volume of timber of the thinned trees is taken into account simply by adding the equivalent number of trees with the same individual tree timber volume.
The intercrop time-series are also generated with this logistic equation, with an Yfinal equal to 0: one time-series per crop of the rotation (maize, wheat, oilseed, etc...). For each crop, the curve is adjusted in function of the threshold yield and the width of the intercropped alley reduction step: as the yield per total ha decreases with time due to tree growth and increasing light competition, we assume that the cropped area is reduced by successive steps (see Figure 19). A reduction of the width of the intercropped alley happens every time the yield per cropped ha passes under an economically defined threshold. The last reduction corresponds to the suppression of the intercrop.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 44 w
Figure 19: chronological schema of the intercropped area in the alley in function of tree growth.
The width of the intercropped strip (w) is reduced at t1 and t2 in order to increase the mean yield per cropped ha. Its next reduction at t3 corresponds to the suppression of the intercrop. In the generator, up to 6 reductions can be made.
The intercrop curves are adjusted by modifying their inflexion date so that the sum of the intercrop productions is equal to the crop reference yield multiplied by the arable RA-products.
The intercrop time-series are then mixed according to the arable rotation to obtain a single time-series.
• The same tree height curve time series in forestry and agroforestry
A last time-series is generated for both forest and agroforest trees : their height growth. It is not used in the timber volume calculation, but this time-series is needed in Farm-SAFE for its “autoprune” function.
We use the Boltzmann logistic equation :
Yinitial −Yfinal Y(t) = +Yfinal −Y()t = 0 t −Tinf lexion 1+e curvature
As for the timber time-series, Y initial is equal to zero and Y final to any value, as the curve is distended in order to go through a point {X’;Y’} : X’ is the date of fell of the trees and Y’ is the aimed height.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 45
Annex 2: Labour, revenues and costs in the 3 types of farms
Assignabl Single Farm Variable Price e fixed Annual labour Payment costs Crop costs (h/ha) (€/t) (€/ha) (€/ha) (€/ha)
110 wheat 6, (straw 30 €/t) 343 300 207
oilseed 5,5 215 360 302 207
set aside 1,5 – 338 15 207 y-Lc H 110 wheat 6 (straw 30 €/t) 345 302 252
oilseed 5,5 220 361 391 252
sunflower 5,5 280 361 262 252
set aside 1,5 – 345 27 252 y-Lc L 102,10 wheat 7 (straw 30 €/t) 328 278 315
oilseed 5,5 220 348 390 315
maize 7 85,4 348 434 315
set aside 1,5 – 328 15 315 y-Hc H
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 46
Annex 3: Economic data relative to monocropped or intercropped walnut, wild cherry and poplar in the 3 farms
Tree timber standing value
Standing value (€/m3)
Walnut Wild cherry Poplar Tree timber volume (m3/tree) thinned trees felled trees thinned trees felled trees felled trees
0.03 00000 0.04 10 10 10 10 7 0.05 10 10 10 10 7 0.06 10 10 10 10 7 0.07 10 20 10 10 7 0.08 10 20 10 10 7 0.09 10 30 10 10 7 0.1 10 40 10 10 7 0.11 10 60 10 10 7 0.12 10 80 10 10 8 0.13 10 100 10 10 8 0.14 10 126 10 10 8 0.15 20 135 10 10 8 0.16 20 144 10 10 8 0.17 20 153 10 12 8 0.18 20 162 10 15 8 0.19 20 171 10 20 13 0.2 30 180 10 40 15 0.3 30 270 15 55 20 0.4 40 360 20 75 24 0.5 50 450 22 150 28 0.6 100 540 25 250 32 0.7 190 630 35 275 35 0.8 220 720 45 300 37 0.9 300 810 55 325 39 1 400 900 65 350 41 1.1 500 925 75 360 43 1.2 600 950 85 370 45 1.3 700 1000 95 380 46 1.4 800 1000 105 380 47 1.5 900 1000 115 380 48 1.6 1000 1000 125 380 49 1.7 1000 1000 135 380 50 1.8 1000 1000 145 380 51 1.9 1000 1000 165 380 52 2 1100 1100 175 380 53 3 1200 1200 200 380 55 4 1300 1300 200 380 55
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 47 Establishment costs
Cost of Labour Labour for Labour Labour for Labour for Labour for Cost of individual for ground for full planting tree localised plant tree marking preparation weeding trees protection weeding protection out
(€/tree) (€/tree) (h/ha) (h/ha) (h/ha) (min/tree) (min/tree) (min/tree) agroforest walnut 6.00 1.50 4.00 0.50 7 2 2 0.50 forest walnut 4.00 0.50 6.50 1.50 4 2 1 0.50 agroforest wild cherry 1.00 1.50 4.00 0.50 7 2 2 0.53 forest wild cherry 0.50 0.50 6.50 1.50 4 2 1 0.53 agroforest poplar 4.00 0.50 12.00 0.50 7 2 2 0.50 forest poplar 4.00 0.50 16.00 1.50 4 2 1 0.50 Maintenance and pruning costs
Labour for Materials for Annual Annual cost Height at Minutes per Minutes per Weeding annual grass annual grass Height at Removal of labour for of herbicide first tree at first tree at last period sward sward last prune pruning weeding for weeding prune prune prune maintenance maintenance
(years) (min/tree) (€/tree) (h/ha) (€/ha) (m) (min/tree) (m) (min/tree) (min/tree) agroforest walnut 1 - 3 0.50 0.14 2.0 30 1.00 1.00 4 7.00 4.00 forest walnut 1 - 3 0.50 0.14 4.0 90 1.00 0.20 4 7.00 4.00 agroforest wild cherry 1 - 3 0.50 0.14 2.0 30 1.00 1.00 6 6.40 4.00 forest wild cherry 1 - 3 0.50 0.14 4.0 90 1.00 0.18 6 6.40 4.00 agroforest poplar 1 - 3 0.53 0.14 2.0 30 1.50 1.00 8 10.00 4.00 forest poplar 1 - 3 0.53 0.14 4.0 90 1.50 1.00 8 10.00 4.00 Labour for thinning and felling
Thinnings Clear felling Marking up & Removal of tree Labour Removal of tree labour (min/tree) (min/tree) (min/tree) (min/tree) agroforest walnut 7542 forest walnut 7542 agroforest wild cherry 7542 forest wild cherry 7542 agroforest poplar 7542 forest poplar 7542 Administrative costs
In agroforestry, the land tax is the same as in an agricultural plot. It was thus applied to the tree strips.
Land tax Insurance (€/ha) (€/ha) Hy-Lc AF plot (agricultural tax) 44 20 (Centre) forestry plot 30 20 Hy-Hc (Franche- AF plot (agricultural tax) 52 20 Comté) forestry plot 36 20 Ly-Lc (Poitou- AF plot (agricultural tax) 58 20 Charentes) forestry plot 39 20
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 48
ANNEX 3. What land status for agroforestry plots in France ? (in French)
Fiche de synthèse dans le cadre de la Loi d’Orientation Agricole
Assemblée Permanente des Chambres d’Agriculture
Fabien Liagre - Mars 2005
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 49
Sommaire
1 DÉFINITION DE L’AGROFORESTERIE...... 50
2 PROBLÉMATIQUE LIÉE AU STATUT ...... 51
3 LES SOLUTIONS POSSIBLES ...... 52 Un forfait spécial agroforesterie 53 Un forfait distinct au prorata 53
4 ANALYSE DES SOLUTIONS PRÉSENTÉES ...... 54 D’un point de vue administratif 54 Conséquence pour le calcul de l’impôt foncier 55 Conséquence pour le calcul de l’impôt sur le revenu 56 Conséquence pour le calcul de l’imposition sur le patrimoine 56 Conséquence pour la gestion de l’exploitation 56
5 SOLUTION PROPOSÉE ...... 57
6 FAUT-IL MODIFIER LA LOI ? ...... 58 Les surfaces boisées et le code rural 58 Le bail agroforestier 59
7 ANNEXE : RAPPORT DU BUREAU DES ETUDES FISCALES ...... 60
DEFINITION DE L’AGROFORESTERIE
L’agroforesterie consiste à associer étroitement des arbres à faible densité avec une culture ou une pâture sur une même surface. Deux types d’agroforesterie sont envisageables :
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 50
L’agroforesterie sur terres agricoles. Des arbres sont plantés ou maintenus dans une parcelle agricole. Sous forme traditionnelle, on peut citer les pratiques de prés-vergers, de cultures intercalaires dans les noyeraies du Dauphiné et du Périgord. Dans le cadre du programme Européen SAFE, il s’agissait justement d’étudier les relations arbres / cultures dans ce type
de schéma agroforestier.
L’agroforesterie sur terres forestières. Il s’agit ici d’aménager des espaces boisés afin de rendre possible une production agricole, souvent de nature fourragère, tout en conservant une production de nature sylvicole. Traditionnellement, il s’agit des systèmes de sylvopastoralisme ou de pré-bois. Mais on peut également citer la production de myrtilles dans les Vosges ou de vanille à la Réunion.
De plus en plus de projets agroforestiers voient le jour en France. Ainsi, plus de 1500 ha sont à l’étude pour 2005/06. Et il est vraisemblable que l’on doublera cette surface avant fin 2006. Cette évolution constatée sur l’ensemble du territoire national suscite souvent des questions quant à la nature fiscale de ce type de parcelle, que ce soit de la part des porteurs de projets mais également des services fiscaux, un peu désorientés face à des pratiques innovantes. L’objet de cette note est de donc de faire un état des lieux de la situation statutaire et de proposer des solutions, éventuellement dans le cadre de la Loi d’Orientation Agricole.
PROBLEMATIQUE LIEE AU STATUT
Dans le cadre des Boisements de Terres Agricoles ou BTA, à partir du moment où l’on réalise une plantation d’arbres sur une parcelle agricole, on modifie complètement le statut de la parcelle qui devient forestière. Dans ce cas, après le boisement, une notification du propriétaire est envoyée aux services cadastraux pour signifier le changement de nature de l’occupation du sol. Le changement de statut est donc essentiellement déclaratif plus que technique à ce niveau. Le propriétaire y a tout intérêt car la législation actuelle lui permet de bénéficier d’une exonération de l’impôt foncier pendant une période comprise entre 30 et 70 ans. Cette exonération lui ouvre également la porte à des réductions des prélèvements fiscaux de ses revenus agricoles.
En agroforesterie, le propriétaire plante nettement moins d’arbres à l’hectare que pour une plantation forestière (sauf pour le peuplier et le noyer où les densités peuvent parfois être similaires). Entre les arbres, un agriculteur (le propriétaire ou un fermier) cultive pendant 50 % à 100 % de la durée de vie de la plantation. Dans ce
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 51
cas, le statut de la parcelle est plus difficile à définir car elle relève de deux activités ou occupations différentes : agricole et forestière.
En agroforesterie, la proportion de surface occupée par les arbres varie dans le temps au fur et à mesure du développement des arbres (surface en hausse) et du rythme des éclaircies (surface en baisse). Plus la densité initiale sera forte (entre 100 et 200 arbres/ha par exemple), et plus l’impact des arbres sur la culture sera fort avec le temps. En absence d’éclaircie forte, et bien que cette éventualité n’est pas une fin en soi dans les principes de l’agroforesterie, l’agriculteur pourrait décider de supprimer la culture si celle-ci n’était plus rentable (sauf éventuellement en la remplaçant par une prairie ce qui permettrait de conserver une activité considérée comme agricole). Le statut proposé devra tenir compte de cette spécificité et être assez souple pour pouvoir s’adapter aux différentes évolutions de la parcelle agroforestière. Une erreur serait sans doute de figer le statut de la parcelle en faveur d’une des deux composantes, arbre ou culture.
100 100
80 ARBRES 80 ARBRES
60 60
40 40 PATURE CULTURES 20 20 % de la surface de la parcelle % de la surface de la parcelle 0 0 0 25 50 75 100 0255075100 % de la durée de vie des arbres % de la durée de vie des arbres
Exemple d’évolution des surfaces agricoles et arborées dans un habitat agroforestier au cours de la vie des arbres pour une densité de 120 arbres à l’hectare. Avec une densité plus faible, de l’ordre de 50 arbres à l’hectare, la proportion de la culture est bien sûr plus importante. En fin de cycle le taux d’occupation de la culture dépasse généralement les 60 % de la surface totale au sol.
D’autre part, il convient de souligner qu’en agroforesterie, les cultures présentes sont complètement indépendantes : les cultures agricoles d’une part et les arbres d’autre part. Ces cultures sont généralement imposées sur une base à l’hectare. On ne peut donc confondre les productions issues d’une parcelle agroforestière avec les productions d’une seule culture à vocation multiple (exemple : arbre fourrager produisant à la fois du bois et du fourrage).
LES SOLUTIONS POSSIBLES
La première question est de savoir s’il faut créer un nouveau statut spécifique à l’agroforesterie ou si l’on peut s’accommoder d’une combinaison de l’existant.
La deuxième question serait : si l’on s’accommode d’une combinaison de l’existant, faut-il pour autant modifier la loi française pour prendre en compte cette solution ?
Les solutions proposées ci-après sont le fruit de réflexion d’un groupe de professionnels, notamment dans le cadre du programme européen SAFE ainsi que des propositions émises par le Bureau des Etudes Fiscales du MAAPAR.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 52
Ces propositions tiennent compte des spécificités des systèmes agroforestiers énoncées au paragraphe précédent.
Un forfait spécial agroforesterie
L’agroforesterie est ici considérée comme une culture spécialisée. Une nouvelle classe est créée au niveau cadastral et la base de calcul du bénéfice forfaitaire est constituée par un bénéfice moyen à l’hectare ou selon les quantités de production présentes.
Ce statut existe dans le département de l’Isère où les parcelles nouvellement plantées de noyers bénéficient du statut de Terres Plantées. Ce statut couvre ainsi le cas des parcelles comportant des arbres improductifs avec des cultures intercalaires
Un forfait distinct au prorata
Le rapport des services du Bureau des Etudes fiscales du MAAPAR indique que « dès lors que sur une même parcelle coexistent deux cultures indépendantes l’une de l’autre, il pourrait être envisagé d’avoir deux forfaits distincts, l’un forestier, l’autre propre à la culture elle-même, chacun au prorata de la surface occupée par les différentes exploitations. »
Sur une même parcelle agroforestière co-existent deux cultures indépendantes l’une de l’autre en terme de surface et d’imposition fiscale. Le forfait de chaque culture est déterminé à l’hectare.
A Vézénobres, dans le Gard, suite à une réflexion départementale, les services fiscaux ont attribué un statut fixé au prorata des surfaces occupées par chacune des composantes, culture et peuplier, dans le cas d’une parcelle agroforestière.
Dans cette solution, deux forfaits distincts sont donc appliqués au prorata des surfaces de chaque culture. Mais, comme nous l’avons vu en introduction, la surface respective de chaque culture évolue dans le temps. Deux prorata sont donc envisageables : l’un variable et l’autre fixe.
• Un prorata variable : le bénéfice de chaque production est ajusté à la surface respective. L’ajustement est réalisé annuellement ou par périodeUn prorata
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 53
fixe : le bénéfice de chaque production est ajusté à une surface moyenne occupée sur toute la durée de l’association. L’ajustement est définitif et forfaitaireANALYSE DES SOLUTIONS PRESENTEES
Les 3 solutions proposées, le forfait spécial, le prorata variable et le prorata fixe sont comparées selon diffèrents niveaux :
• Au niveau administratif d’abord, afin de juger de leur facilité d’application
• Au niveau de l’imposition du foncier, du revenu et du patrimoine
• Au niveau des dispositifs réglementaires nationaux
La proposition retenue devra tenir compte de quelques principes de base. Dans la mesure du possible, il faudra :
• Proposer des modifications simples à mettre en œuvre et cohérentes, peu onéreuses à l’Etat ou équilibrées dans le rapport Coût / Economie réalisé pour l’Etat.
• Laisser la possibilité d’aménagements ultérieurs des textes officiels pour encourager la pratique de l’agroforesterie. L’objectif n’est pas ici de rechercher immédiatement des mesures de soutien ou d’exonération de certaines taxes. On se tiendra à quelques notions fondamentales qui seront évaluées voir réajustées au fur et à mesure de leurs applications. Car il semble impossible de prévoir toutes les implications juridiques de telle ou telle solution.
D’un point de vue administratif
Forfait Spécial Prorata variable Prorata fixe
Simplicité des calculs à Prise en compte de la Prise en compte de la l’échelle de la parcelle production forestière production forestière Système équitable, Facilité de gestion des adaptable avec l’évolution déclarations des surfaces de chaque production
Risques de distorsions Suivi plus complexe par Prorata définitif et entre les départements les services des impôts forfaitaire. Augmentation du nombre de classes cadastrales Augmentation du nombre de comptes existant dans le cadre du forfait collectif
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 54
Conséquence pour le calcul de l’impôt foncier
Forfait Spécial Prorata variable Prorata fixe
Risque d’impôt foncier Impôt calculé au plus juste Dans une parcelle élevé par rapport aux surfaces forestière, la surface Après la coupe des arbres, occupées par les occupée par la culture est la parcelle redevient productions associées plus importante en début agricole. Si le propriétaire Possibilité d’exonération de rotation. L’application ne replante pas, le statut partie arbre à l’image de d’un forfait fixe, calculé sur doit être revu au niveau du ce qui est réalisé en BTA la moyenne d’occupation cadastre. des surfaces de chacune Le calcul des surfaces des deux productions, a demande à être réalisé comme conséquence de périodiquement. Cette favoriser l’agriculteur en réactualisation pourrait début de rotation, et de le être réalisée par période défavoriser en fin de conséquente de 5 à 10 rotation. Par exemple, ans par exemple. pour un forfait Après la coupe des arbres, correspondant à une la parcelle redevient combinaison de 50/50, agricole. Si la partie c’est avantageux pour forestière était importante, l’agriculteur qui cultive sur voire à 100% en cas 90 % de la parcelle les d’arrêt des cultures, premières années. Mais existe-t-il un risque de nettement moins, avant la procédure complexe pour coupe des arbres où il repasser en statut continue de payer un agricole ? impôt foncier agricole évalué à 50 % alors qu’il peut ne cultiver que 20 à 30 % par exemple. Possibilité d’exonération de la partie arbre à l’image de ce qui est réalisé en BTA Le calcul du forfait peut s’avérer complexe, car il va dépendre de l’essence, de la densité et de l’écartement entre les lignes d’arbres. De plus, se pose le problème des plantations à essences multiples : comment calculer un forfait fixe avec des essences à rotation différentes plantées dans une même parcelle?
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 55
Conséquence pour le calcul de l’impôt sur le revenu
Forfait Spécial Prorata variable Prorata fixe
Si mode d’imposition au Si mode d’imposition au Si mode d’imposition au réel normal, situation réel : situation inchangée. réel normal : situation inchangée. inchangée. L’imposition au Si mode d’imposition au forfait est facilitée par le forfait, le calcul est réalisé Si mode d’imposition au forfait spécial en fonction des surfaces forfait, l’imposition sur la agroforesterie. respectives déclarées. partie agricole sous- évaluée les premières Néanmoins, on retrouve L’imposition de l’activité années et surévaluée les les mêmes conséquences forestière est réalisée au dernières années. que pour le scénario forfait selon la surface prorata fixe. déclarée L’imposition forfaitaire sur la partie boisée est Possibilité de bénéficier de surévaluée les premières dégrèvements fiscaux sur années et sous-évaluée le revenu forestier à les dernières années l’image de ce qui est réalisé en BTA Possibilité de bénéficier de dégrèvements fiscaux sur
le revenu forestier à l’image de ce qui est réalisé en BTA
Conséquence pour le calcul de l’imposition sur le patrimoine
Le calcul de l’imposition sur le patrimoine (ISF, transmission) est réalisé sur les valeurs déclarées par le contribuable.
Chaque année, il évalue la partie vénale de la partie arbre et de la partie agricole.
Il existe des abattements au prorata de la partie agricole et forestière dans le cadre du calcul des droits de mutation. Les abattements sont plus avantageux pour la partie forêt que la partie agricole.
Le statut cadastral doit être de nature forestière pour prétendre à ces dispositifs. Un statut spécial pour une parcelle agroforestière devra être pris en compte au niveau du dispositif réglementaire. Ce qui n’est pas le cas dans le cas des statuts au prorata, où les dispositifs existant s’appliqueraient au prorata de la surface forestière. Néanmoins, dans le cas du statut au prorata fixe, les abattements fiscaux seraient plus élevés les premières années, compte tenu de l’importance de la surface forestière réelle, et moins favorable les dernières années.
Conséquence pour la gestion de l’exploitation
Deux exemples sont ici cités pour montrer les implications possibles du choix du statut sur des aspects de réglementations courantes concernant les exploitations agricoles ou forestières.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 56
o Le calcul des cotisations MSA
Le statut de la parcelle intervient pour le calcul de cotisations sociales des agriculteurs au forfait uniquement. L’influence du statut cadastral est forte dans le cas du forfait au prorata fixe. Tout comme pour l’imposition forfaitaire, les premières années le montant des charges sociales sera plus faible compte tenu de la surface agricole plus importante que celle retenue dans le calcul du forfait, et inversement en fin de rotation.
o Statut foncier et sinistres naturels
Les indemnités versées aux propriétaires ou agriculteurs lors de sinistres exceptionnels sont liées à la nature des sols. Ainsi, les indemnités versées suite aux tempêtes de Noël 99 ont été conditionnées au statut forestier des parcelles. De même, les indemnités agricoles suite à un gel ou à une forte grêle sont-elles versées en fonction du statut agricole des parcelles. Ici aussi, le système le plus équitable en terme de surface est le statut au prorata variable qui est le seul statut permettant de distinguer réellement les surfaces agricoles et forestières au moment du sinistre.
SOLUTION PROPOSEE
Seules 3 solutions ont été comparées dans ce document. Nous avons écartés notamment les autres possibilités qui auraient été le choix d’un statut purement agricole, voire purement forestier.
Au vu des éléments donnés dans ce document, il semble que la solution la plus favorable au cas de l’agroforesterie réside dans le choix d’un statut au prorata des surfaces occupées par la culture agricole d’une part et forestière d’autre part. Afin d’éviter une réactualisation annuelle de ce statut, qui serait sans doute lourd à gérer pour l’administration fiscale, pour les propriétaires ainsi que pour les agriculteurs dans la gestion de leur exploitation, cette actualisation pourrait être réalisée par période. Pour les essences de moyenne ou longue rotation, supérieure à 30 ans, on pourrait envisager la mise en place d’un forfait décennal, sur la base déclarative du propriétaire. Pour des essences à plus courte rotation, comme le peuplier, l’actualisation pourrait être réalisée tous les 5 ans.
Afin de faciliter d’éventuel contrôles fiscaux, il est proposé que la méthode de calcul des surfaces soit la plus simple possible. On conseille de calculer la surface agricole en fonction de la surface réellement semée ou pâturée. Un contrôle de surface par photographie aérienne est à déconseiller car les surfaces d’emprise des arbres vue d’altitude sont largement agrandies par rapport à leur surface d’emprise réelle.
Un statut évolutif s’adapte bien aux différentes conséquences fiscales et réglementaires exposées dans les paragraphes précédents. Cette option ne signifie pas de coût particulier pour l’administration. Elle rend possible également de soutenir fiscalement l’agroforesterie sur les mêmes bases que les boisements de terres agricoles mais uniquement sur les surfaces occupées par les arbres.
Le principe de cette solution est repris en partie dans les conclusions du document du Bureau des Etudes Fiscales du Ministère de l’Agriculture du 8 oct. 2004. En
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 57
conclusion de ce document, il est indiqué que « le régime fiscal de l’agroforesterie pourrait être étudié dans le cadre du projet de Loi d’Orientation Agricole. »
FAUT-IL MODIFIER LA LOI ?
L’objectif de ce document n’est pas de répondre à cette question mais d’apporter des éléments de réflexion dans cette perspective.
Les surfaces boisées et le code rural
Dans l’article du code rural, on peut lire les articles suivants se rapportant aux surfaces boisées hors forêt, lorsque qu’une démarche de protection des zones boisées est engagée:
Article R126-37
L'emprise et l'indication des parcelles cadastrales sur lesquelles sont situés les boisements linéaires, haies, plantations d'alignement ou vergers de hautes tiges, dont la protection est prononcée, doivent être matérialisées sur un plan parcellaire annexé à l'arrêté préfectoral prononçant la protection ou sur le plan des aménagements fonciers prévu à l'article L. 121-21. L'arrêté précise les éléments techniques visés à l'article ci-dessus.
Article R126-38
Les boisements linéaires, haies ou plantations d'alignement nouvellement protégés doivent être portés à la connaissance de l'administration des impôts dans les formes et délais définis à l'article 1406 du code général des impôts. Les emprises ainsi créées, matérialisées dans les conditions prévues à l'article ci-dessus, seront considérées comme nature de culture se rapportant au groupe des bois.
A ma connaissance, il s’agit du seul texte indiquant clairement la possibilité de distinguer sur une parcelle agricole une surface de nature forestière. Néanmoins, ces articles s’appliquent lors d’un processus de protection de formations arborées hors forêt. En dehors de cette perspective, il ne semble pas que cette éventualité soit considérée, que ce soit au niveau du code rural comme du code forestier.
Il semble donc souhaitable de débattre de la possibilité d’inclure un paragraphe décrivant clairement la situation des parcelles agricoles arborées - qui citerait notamment le cas de l’agroforesterie - où l’on pourrait distinguer les surfaces cadastrales agricoles d’une part et forestières d’autre part.
Il serait alors spécifié :
o le processus des déclarations à effectuer (date de déclaration et durée de la période entre deux déclarations)
o le fait que chacune des classes relève des mêmes régimes fiscaux que dans le cas des parcelles avec cultures pures.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 58
Le bail agroforestier
L’agroforesterie mais également l’agriculture moderne ne correspond pas au concept de bail rural tel que l’on entend aujourd’hui.
En effet, un fermier peut aujourd’hui laisser sa parcelle louée en jachère ou sans attendre de production et toucher des primes PAC. La notion d’exploitation agricole du bien loué en bon père de famille tel que cela était définie après guerre, n’est plus valable aujourd’hui.
De même, dans le cas de l’agroforesterie, il serait souhaitable de modifier les textes actuels pour rendre possible la création d’un bail agroforestier, où l’agriculteur pourrait cultiver les surfaces agricoles sans toucher aux arbres du propriétaire. De même, on pourrait imaginer que le fermier se lance en agroforesterie (avec des essences à courte rotation par exemple), sans que cela ne remette en cause le contrat signé avec le propriétaire.
Une analyse plus approfondie des textes actuels et des réformes proposées permettrait d’inclure le cas particulier de l’agroforesterie dans le cadre de Loi d’Orientation Agricole.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 59
APPENDIX : RAPPORT DU BUREAU DES ETUDES FISCALES DU 8 OCT. 2004
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 60
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 61
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 62
ANNEX 4. Quelle place pour les arbres hors forêt dans la nouvelle PAC ?
Synthèse et propositions pour les réglementations européennes et françaises…
Assemblée Permanente des Chambres d’Agriculture
Fabien Liagre - Mars 2005
Résumé des propositions
La réforme des accords de Luxembourg se met en place progressivement dans chacun des Etats membres. En France, le gouvernement prépare le prochain règlement d’application de la PAC en vue de 2006, année de mise en place du découplage des aides.
Parallèlement, la Commission Européenne, après validation du Parlement Européen et du conseil des ministres de l’Agriculture en début d’année, finalisera le prochain Règlement de Développement Rural qui interviendra sur la période 2007-2013.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 63
Chaque Etat membre débutera alors les travaux en vue de l’élaboration de son Programme de Développement Rural National.
Dans le cadre de ces évolutions réglementaires, les dispositions en faveur des formations arborées hors forêt évoluent, certes avec une progression variable selon les Etats membres mais avec certitude. Pour la première fois, une mesure en faveur de l’agroforesterie est citée dans le projet de RDR. Jamais une mesure en faveur des arbres ruraux n’avait été aussi clairement énoncée et appuyée dans un règlement européen. Dans cette évolution, la France fait figure de précurseur grâce aux différentes mesures réglementaires prises dès 2001 en faveur de l’agroforesterie notamment.
L’APCA a souhaité faire un bilan de l’existant, fruit de 3 années de recherche dans le cadre du programme européen SAFE, afin de clarifier la situation réglementaire et d’améliorer la lisibilité et la portée des mesures proposées.
Ce travail de réflexion amène aux principales conclusions suivantes qui sont développées dans le document :
Au niveau européen
1. Compte tenu de la valeur agro-environnementale des formations arborées hors forêt et de la complexité administrative que suscite leur prise en compte dans le calcul des surfaces éligibles au paiement compensatoire du 1er pilier, il est proposé de rendre admissible au paiement unique les parcelles arborées dans la totalité de leur surface. De même, il est proposé que les surfaces occupées par les arbres puissent être éligibles au paiement unique. Ces deux propositions demandent une clarification de la définition d’une parcelle arborée ainsi que la modification des règlements concernant les accords du Luxembourg.
2. Sur le modèle de l’article 41 du projet de RDR du 14 juillet 2004, il est demandé une homogénéisation des mesures en faveur des arbres hors forêts. Sur le même principe d’éligibilité de l’agroforesterie, une mesure universelle en faveur de l’arbre rural simplifierait l’adoption de mesure de soutien dans chacun des pays membre.
Au niveau français
Concernant les aides du premier pilier, il faut distinguer le cas des aides couplées des aides découplées.
Les propositions pour le régime des aides découplées (DPU) sont :
3. Dans le cadre de l’élaboration de la prochaine circulaire d’application de la PAC concernant les paiements compensatoires en 2006, la France pourrait dès à présent autoriser l’éligibilité totale des parcelles arborées dans le cadre du découplage des aides, pour des raisons agroenvironnementales. Des critères techniques, faciles à contrôler, sont proposés.
4. L’obtention d’une éligibilité totale permettrait de simplifier les procédures administratives tant au niveau de l’instruction des dossiers qu’au niveau du
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 64
contrôle. En effet, les textes font appel aux normes locales pour déterminer les méthodes de calcul des surfaces d’emprise des arbres en agroforesterie. Hors aucune norme départementale n’existe en matière d’agroforesterie moderne et très peu de normes existent pour les autres formations arborées.
5. En cas d’éligibilité des parcelles arborées au paiement unique, il conviendrait alors de supprimer l’éligibilité des parcelles agroforestières à la Prime de Compensation à la Perte de Revenu agricole.
Dans le cadre des aides couplées à la production (aides SCOP) :
6. Il est proposé de calculer la surface éligible en fonction de la surface réellement occupée par la culture intercalaire, et non de la surface obtenue après déduction de la surface d’emprise des arbres. En effet, dans le cadre des systèmes de contrôle par vue aérienne, du fait de l’angle de vue, la surface d’emprise des arbres dépasse la surface d’emprise réelle des arbres. L’impact des arbres est donc beacoup plus fort que dans la réalité ce qui est très pénalisant pour l’agriculteur.
Dans le cadre du second pilier :
7. Face à la demande grandissante et aux résultats probants de la recherche en la matière, il est proposé de modifier la circulaire d’application de la mesure en faveur de l’agroforesterie afin de lever l’obligation de suivi d’un organisme à titre expérimental. Une clarification du cahier des charges permettrait également aux administrations locales d’instruire plus facilement les dossiers.
8. Enfin, afin de mettre à disposition un outil efficace aux collectivités territoriales souhaitant appuyer des actions agroforestières, il est proposé de revoir le cahier des charges de la MAE Habitats Agroforestiers afin de la rendre plus cohérente vis-à-vis des autres mesures. D’autre part, une évolution de cette MAE en MAE universelle en faveur de l’arbre est souhaitée afin de simplifier la procédure des Contrats d’Agriculture Durable. Son action porterait alors exclusivement sur des actions où le caractère environnemental serait renforcé.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 65
Sommaire
1 INTRODUCTION ...... 69
2 ELIGIBILITÉ DES PARCELLES ARBORÉES AUX PAIEMENTS COMPENSATOIRES ...... 69 Place de l’arbre dans l’historique des réformes de la PAC 69 L’agenda 2000 69 La réforme des accords du Luxembourg 70 Les règles de base ...... 70
Les règlements d’application...... 71
Conclusions au niveau européen 73 Une situation ambiguë 73 Le principe de subsidiarité des Etats membre prévaut ...... 73
Un système de contrôle des surfaces inadapté ? ...... 74
Quelle situation pour les nouveaux membres de l’UE ? ...... 74
Les propositions 74 Le régime d’application en France 75 Evolution de l’historique de l’application des règlements européens 75 Avant la réforme des accords du Luxembourg...... 75
Après la réforme des accords du Luxembourg ...... 76
Conclusions 77 Les agriculteurs agroforestiers pénalisés...... 77
Conséquences possibles pour des parcelles arborées existantes ...... 77
Une gestion des droits qui incite à l’arrachage...... 77
Des normes locales quasi-inexistantes ...... 78
Propositions au niveau national 78 Améliorer la définition des normes usuelles...... 78
Simplifier le contrôle des surfaces arborées ...... 78
Admissibilité et éligibilité des parcelles agroforestières ...... 79
Cas des aides découplées ...... 79
Simplifier l’approche administrative...... 79
Une éligibilité totale soumise à conditions...... 79
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 66
Conséquences à l’échelle de l’exploitation et au niveau national ...... 80
Cas des aides couplées ...... 81
Rendre cohérent l’articulation entre 1er et 2ème pilier ...... 82
Intégrer les surfaces arborées dans le couvert environnemental ...... 82
3 LES ARBRES DANS LE DEUXIÈME PILIER DE LA PAC ...... 82 Le Règlement de Développement Rural 82 L’apparition d’une mesure Agroforesterie à l’horizon 2007 82 Une mesure incomplète ? 85 Les aides disponibles en France 85 Le cas des formations arborées hors agroforesterie 85 Le cas des systèmes agroforestiers 87 Les aides à la plantation...... 87
Les textes officiels ...... 87
Une mesure encourageante mais difficile d’application ...... 88
La compensation à la perte de revenu ...... 89
Les textes officiels ...... 89
Faut-il maintenir cette disposition ?...... 90
La MAE Habitats Agroforestiers ...... 91
Présentation de la mesure...... 91
Quel avenir pour la MAE Habitats Agroforestiers ?...... 91
Faut-il revoir le cahier des charges ? ...... 92
Quelles conséquences avec l’application du prochain RDR ?...... 92
Une reconnaissance facilitée de l’agroforesterie...... 92
Homogénéiser le soutien aux formations arborées hors forêt ? ...... 92
4 BIBLIOGRAPHIE ...... 93
5 ANNEXES ...... ERREUR ! SIGNET NON DEFINI. Annexe 1: Article TransRural Initiative de déc 2004 94 Annexe 2: Propositions du Groupe Réglementations – SAFE 95 Annexe 3: Lettre de Luc Guyau APCA au MAAPAR – PAC 97 Annexe 4: Chapitre 10 de la circulaire forêt de protection du 7 mai 01 102 Annexe 5: Texte MAE Habitats Agroforestiers 105
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 67
Annexe 6: Lettre de Luc Guyau APCA au MAAPAR – MAE 114
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 68
Introduction
Ce document a pour objectif de décrire la situation réglementaire des parcelles agricoles arborées vis-à-vis de la Politique Agricole Commune, de prévenir des possibles complications de leur prise en compte et de faire des propositions au niveau européen et national.
Afin de bien cerner les enjeux actuels dans le cadre de la réforme de PAC en cours, une analyse de la situation avant 2005 sera effectuée. Une deuxième partie abordera alors la nouvelle situation générée par la réforme et les questions qu’elle suscite pour le cas des parcelles arborées.
Nous tenterons de prendre en compte les différentes formes arborées existantes : arbres isolés, alignement, haies et agroforesterie. Les pré-vergers et les noyeraies du Dauphiné avec cultures intercalaires ne sont pas oubliés, en tant que forme traditionnelle d’agroforesterie en France.
ELIGIBILITE DES PARCELLES ARBOREES AUX PAIEMENTS COMPENSATOIRES
Place de l’arbre dans l’historique des réformes de la PAC
Avant d’aborder la prise en compte des arbres hors forêt en France, il est nécessaire de faire un rapide tour d’horizon des principaux textes européens qui régissent la PAC. Dans un deuxième temps, on détaillera la position française lors de l’application des règles de la PAC.
L’agenda 2000
Lors de la réforme de l’agenda 2000, les modalités d’application du SIGC ou Système intégré de Gestion et de Contrôle des surfaces éligibles étaient alors régies par le règlement CE No 2419/2001 relatif à certains régimes d'aides communautaires établis par le règlement (CEE) no 3508/92. Cette réforme de l’agenda 2000 désolidarisait le paiement compensatoire de la production pour l’attribuer en fonction de la surface occupée par les cultures. Aujourd’hui ces deux règlements ne sont plus en vigueur mais le règlement 2419/2001 reste souvent une référence des règlements suivants concernant l’éligibilité des surfaces arborées.
Dans le règlement 2419/2001, l’éligibilité des parcelles arborées est abordée à l’article 5.
Article 5 a) une parcelle portant à la fois des arbres et une culture prévue à l'article 1er du règlement (CEE) no 3508/92 est considérée comme une parcelle agricole à condition que la culture susvisée puisse être effectuée dans des conditions comparables à celles des parcelles non arborées de la même région;
Extrait art 5 – Règ. 2419/2001
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 69
Les surfaces en haies peuvent être considérées comme éligible sur une largeur de 2 mètres en fonction des choix de chaque Etat membre correspondant aux pratiques traditionnelles. Cet aspect est abordé dans l’article 22.
Article 22
Dans les régions où certaines caractéristiques, en particulier les haies, les fossés et les murs, font traditionnellement partie des bonnes pratiques agricoles en matière de culture ou d'utilisation, les États membres peuvent considérer que la superficie correspondante fait partie de la superficie totale utilisée, pour autant qu'elle ne dépasse pas une largeur totale à déterminer par les États membres. Cette largeur doit correspondre à une largeur traditionnelle dans la région en question et ne doit pas excéder deux mètres.
Extrait art 22 – Règ. 2419/2001
La réforme des accords du Luxembourg
Les règles de base
Dans le cadre de la réforme des accords du Luxembourg en 2003, la Commission met en place un paiement unique par exploitation pour les agriculteurs de l'UE, indépendant de la production. Des éléments de couplage limités sont toutefois maintenus pour éviter l'abandon de la production dans certains pays, comme la France par exemple.
A ce niveau, il convient de distinguer les notions d’admissibilité et d’éligibilité. Dans le cadre du régime de Droit à Paiement Unique ou DPU, une première étable est de déterminer les surfaces admissibles au DPU. En effet, la surface adminissible par exploitation peut être supérieure à la surface éligible dans le cadre de l’aide à la production ou l’aide SCOP. Sur ces surfaces, les cultures pratiquées doivent être éligibles, c'est-à-dire ouvrant droits à paiement.
Le règlement (CE) No 1782/2003 du 29 septembre 2003 établit les règles communes pour le soutien direct. A l’article 44, il est spécifié que les surfaces comportant des éléments permanents ne peuvent être comptés dans les surfaces ouvrant droits à paiement.
Aux fins du paragraphe 2, point b), du présent article, on entend par «superficie fourragère» la superficie de l'exploitation disponible pendant toute l'année civile, conformément à l'article 5 du règlement (CE) no 2419/2001 (1) de la Commission, pour l'élevage d'animaux, y compris les superficies utilisées en commun et les superficies soumises à une culture mixte. Ne sont pas comptés dans cette superficie:
— les bâtiments, les bois, les étangs, les chemins,
— les superficies utilisées pour d'autres cultures admissibles au bénéfice d'une aide communautaire, pour des cultures permanentes ou pour des cultures horticoles,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 70
— les superficies bénéficiant du régime de soutien aux agriculteurs produisant certaines grandes cultures, qui sont utilisées dans le cadre du régime d'aide concernant les fourrages séchés ou soumises à un programme national ou communautaire de gel des terres.
Extrait Article 44 – Règ. 1782/2003
A remarquer qu’il est indiqué que les cultures mixtes peuvent être comptabilisées dans les surfaces éligibles. A noter également que les bois pâturés ne pourraient pas être éligibles.
A l’article 51, il est spécifié que l’agriculteur perd ses droits à paiements pour les surfaces mises en cultures permanentes.
Article 51
Utilisation agricole des terres
Les agriculteurs peuvent utiliser les parcelles déclarées conformément à l'article 44, paragraphe 3, pour toute activité agricole à l'exception des cultures permanentes et de la production de produits visés à l'article 1er, paragraphe 2, du règlement (CE) no 2200/96 du Conseil du 28 octobre 1996 portant organisation commune des marchés dans le secteur des fruits et légumes (1) et à l'article 1er, paragraphe 2, du règlement (CE) no 2201/96 du Conseil du 28 octobre 1996 portant organisation commune des marchés dans le secteur des produits transformés à base de fruits et légumes (2) ainsi que de pommes de terre autres que celles qui sont destinées à la fabrication de fécule pour lesquelles l'aide est octroyée au titre de l'article 93 du présent règlement.
Extrait art 51 – Règ. 1782/2003
Enfin, certaines surfaces boisées comme les boisements de terres agricoles peuvent toutefois être comptabilisées en surfaces ouvrant droits à gel (art 54).
Les règlements d’application
La mise en œuvre de la réforme est régie par 3 règlements d’application :
• Le premier règlement concerne les dispositions relatives à la conditionnalité, aux contrôles et à la modulation. Ce règlement 796/2004 abroge notamment le règlement 2419/2001.
• Le deuxième règlement introduit le paiement unique par exploitation et le découplage de la production (règlement 795/2004)
• Le troisième règlement porte sur les secteurs d'aides qui demeureront spécifiques à certaines productions (2237/2003)
Le premier principe de l’article 8 du règlement d’application 796/2004 spécifie qu’ « une parcelle boisée est considérée comme une parcelle agricole aux fins du régime d’aide « surfaces » sous réserve que les activités agricoles visées à
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 71
l’article 51 du règlement (CE) n° 1782/2003 ou, le cas échéant, que la production envisagée puissent se dérouler comme elles se dérouleraient sur des parcelles non boisées situées dans la même zone. » Ce principe, qui reprend l’article 5 du règlement 2419/2001 ouvre la porte à l’éligibilité des parcelles agroforestières mais ne spécifie en aucune manière si cette éligibilité peut être totale ou partielle (au prorata de la surface agricole par exemple). A noter que de la notion de parcelle arborée évolue en parcelle boisée.
L’article 8 reste cependant vague sur la distinction entre parcelle boisée et parcelle agricole. Aucune proportion de surface ou de production n’est indiquée. En fait, il faut se référer au document de travail AGRI/2254/2003 pour avoir une première définition d’une parcelle boisée qui ne serait plus considérée comme agricole et éligible. En effet, dans ce document, il est précisé qu’une surface arborée est considérée comme inéligible si le nombre d’arbres dépasse 50 arbres par hectare. Au-delà de ce seuil, la parcelle devient inéligible au titre du Paiement Unique sauf dérogation pour des motifs agro-environnementaux. Néanmoins, cette définition ne s’applique qu’aux parcelles fourragères ou prairies. Rien n’est spécifié pour les parcelles avec grandes cultures. Il faut également souligner que ce document de travail se rapporte au règlement 2419/2001. Ce règlement qui a été remplacé par le Règ. 796/2004 concernait la période de l’Agenda 2000 et non celle de la réforme des accords du Luxembourg et du régime de paiement unique.
A noter également dans ce document, qu’une surface est considérée comme fourragère si plus de 50 % de la surface est considérée comme jachère.
Les superficies couvertes d'arbres – en particulier d'arbres avec utilisation potentielle uniquement pour la production de bois – à l'intérieur d'une parcelle agricole d'une densité supérieure à 50 arbres/ha doivent être considérées comme inéligibles. Des exceptions peuvent être envisagées pour les classes d'arbres de cultures mixtes d'arbres fruitiers et autres. Les exceptions éventuelles doivent être définies à l'avance par les États membres.
Nonobstant une communication spécifique de l'État membre à la Commission, lorsque des caractéristiques pouvant aller jusqu'à 4 m de large (des murs, des fossés, des haies) servent de limites entre les parcelles agricoles et font traditionnellement partie des bonnes pratiques agricoles (par exemple murs de terrasse, fossés de drainage), ces caractéristiques sont néanmoins considérées comme étant incluses et une largeur de 2 m peut être attribuée à chaque parcelle agricole adjacente.
Extrait du document de travail AGRI/2254/2003
Cas des paiements aux vergers
Dans le cadre du règlement 2237/2003, une disposition nouvelle permet un paiement à la surface pour les vergers de production de fruits à coques. Il est spécifié à l’art. 19, que le verger éligible ne doit pas être entrecoupé de cultures. La surface éligible doit être supérieur à 0.1 ha. Le nombre d'arbres producteurs de fruits à coque par hectare de verger ne peut être inférieur à:
• 125 pour les noisetiers,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 72
• 50 pour les amandiers,
• 50 pour les noyers,
• 50 pour pistachiers,
• 30 pour les caroubiers.
Ces normes de densité peuvent être revues à la hausse par tout Etat membre.
L’article 19 qui spécifie que le verger ne doit pas être entrecoupé de cultures prête à confusion. Devons-nous penser que les cultures intercalaires non aidées doivent être exclues du verger sous peine de voir annuler le paiement pour fruits à coque ? Cette disposition semble aller à l’encontre de l’article 8 du règ. 796/2004.
Enfin, le document de travail AGRI/2254/2003 spécifiait que chaque Etat membre pouvaient accorder l’éligibilité aux prés-vergers pour le paiement unique. Si l’on s’en tient au principe de non cumul des paiements sur une même surface, un agriculteur pourrait avoir droit au paiement unique à condition de ne pas solliciter le paiement fruit à coques. Par contre, il semblerait qu’il n’ait pas droit au paiement fruit à coque dès qu’il y a une présence soit de culture, soit de pâture, aidées ou non.
Aparté sur les normes locales
Dans les différentes réglementations européennes, il est régulièrement fait mention que les pratiques doivent être réalisées selon les normes locales.
Ainsi, les produits cultivés sur des superficies doivent être entièrement ensemencées et cultivées conformément aux normes locales. De même que la présence des arbres en milieu cultivé est reconnu et ne modifie pas l’éligibilité des parcelles si ces pratiques relèvent des normes locales. Dans l’article 30 du règlement 796/2004, il est indiqué que la superficie totale d’une parcelle peut être prise en compte à condition qu’elle soit entièrement utilisée selon les normes usuelles de l’Etat membre ou de la région concernée. Dans les autres cas, c’est la superficie réellement utilisée qui est prise en compte.
Si pour certains systèmes traditionnels (bocage, agroforesterie traditionnelle, etc.), il est possible de recourir à certaines normes, quoique parfois très difficilement, concernant les systèmes modernes d’agroforesterie, il n’existe aucune norme locale disponible.
Conclusions au niveau européen
Une situation ambiguë
Le principe de subsidiarité des Etats membre prévaut
Les textes européens laisse toute liberté aux Etats membres de définir eux-mêmes le niveau d’éligibilité des parcelles arborées. Il est en effet possible pour des raisons environnementales de rendre admissibles la totalité d’une surface agroforestière ainsi que de toutes formations arborées hors forêts. Néanmoins, par principe, la règle
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 73
de base est de déduire la surface d’une culture ou élément permanent des surfaces ouvrant droits à prime.
D’autre part, le seuil d’arbres par hectare en dessous duquel la parcelle reste agricole n’est pas clairement identifié au niveau européen ce qui prête à confusion dans la recherche d’une définition claire d’une parcelle agricole.
Un système de contrôle des surfaces inadapté ?
Dans l’article 20 du règlement 1782/2003 la Commission encourage le recours aux techniques de couverture d'ortho imagerie aérienne ou spatiale, ce qui suscite quelques interrogations sur la répercussion que peuvent avoir ces techniques sur le calcul des surfaces arborées. En effet, par photo aérienne ou satellitaire, l’erreur de parallaxe est majeure et fausse totalement l’estimation de la surface. Une haie ou des arbres isolés vus de biais ont une surface plus importante que dans la réalité.
Il conviendrait de définir une méthode standardisée de calcul de la surface occupée par les arbres dans une parcelle donnée pour faciliter le contrôle, car ce point donne lieu à des litiges. Dès à présent, des agriculteurs songent sérieusement à rabattre la hauteur de leurs haies, voire à les supprimer, afin d’en réduire l’impact dans les calculs des surfaces éligibles...
Quelle situation pour les nouveaux membres de l’UE ?
Les agriculteurs des pays de l’Est sont confrontés à leurs premières déclarations de surface. Comme lors de l’Agenda 2000, les surfaces comportant des arbres doivent être déduites des surfaces éligibles déclarées. Face au montant des primes en jeu - une exploitation polonaise de 15-20 ha peut recevoir 6000 euros soit l’équivalent d’une année de salaire (Annexe 1), les agriculteurs sont poussés à enlever haies, bosquets et bois des parcelles pourtant inclus dans leur gestion agronomique des surfaces. Si aucune décision n’est prise concernant l’éligibilité des parcelles agricoles arborées, nous assisterons à une forte diminution du nombre d’arbres telle que celle observée en Europe Occidentale ces 20 dernières années…
Les propositions
Certains aspects réglementaires mériteraient d’être éclaircis ou ajoutés :
1. Le manque de définition claire pour distinguer une parcelle arborée agricole admissible au DPU d’une parcelle boisée admissible (les 2 termes sont employés) prête à confusion. Il conviendrait de réactualiser la définition donnée pour les surfaces fourragères boisées dans le document de travail AGRI/2254/2003, lui-même basé sur des règlements qui ne sont plus en vigueur. La définition pourrait être introduite dans le règlement 795/2004 à l’article 2 à la suite des autres définitions.
2. Compte tenu du système de calcul des droits à primes basé sur l’historique des aides reçues, il est proposé de rendre éligible les surfaces occupées par les arbres dans leur intégralité. Un seuil minimum d’arbres à l’hectare pourra être proposé par défaut aux Etats membres qui pourront le revoir à la hausse. Ce seuil serait conforme à la définition de la parcelle arborée proposée ci- dessus. Pour introduire l’idée de l’éligibilité des parcelles arborées, il est alors
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 74
proposé de modifier le paragraphe 2 du règlement Art44 du règlement 1782/2003 afin d’intégrer l’éligibilité de la parcelle arborée dans la définition de l’hectare admissible. Dans un deuxième temps, il est nécessaire de modifier l’article 51 de ce même règlement afin d’indiquer que les parcelles agricoles arborées restent éligibles.
3. Quelles normes locales lorsque l’on est face à une pratique innovante telles que les parcelles agrisylvicoles ? Il serait judicieux qu’au niveau européen, un avenant réglementaire précise éventuellement ces nouvelles normes. Ainsi dans l’article 30 du règlement 796/2004, il pourrait être introduit une précision permettant la prise en compte de la totalité de la surface agroforestière.
Dans ces propositions, nous suggérons d’apporter des rectificatifs aux règlements 1782/2003 ainsi que 795/2004 et 796/2004. Il pourrait également être envisagé de réactualiser le document de travail AGRI/2254/2003 en conformité avec les nouveaux règlements. La définition de la parcelle arborée serait précisée ainsi que son éligibilité aux droits à paiement unique.
Ces propositions reprennent les propositions formulées par le groupe de travail sur les réglementations du programme européen SAFE (Annexe 2).
Le régime d’application en France
Chaque Etat membre a défini des règlements nationaux d’application des règlements européens.
Evolution de l’historique de l’application des règlements européens
Avant la réforme des accords du Luxembourg
Avant 2001, les surfaces arborées excluaient toute possibilité de paiement compensatoire aux cultures intercalaires, excepté dans le cas des jeunes plantations si ces pratiques correspondaient à des normes locales. Seuls quelques départements ou quelques pratiques isolées bénéficiaient de ce régime particulier.
Dans les départements de la Drôme et de l’Isère, les surfaces de cultures intercalaires entre les jeunes noyers étaient éligibles sous certaines conditions. Un nuciculteur percevait un paiement forfaitaire correspondant à 60 % du paiement correspondant à la totalité de la surface de la parcelle pendant 7 ans. A noter que dans le Périgord, pour le même type de pratique, aucun paiement compensatoire n’était possible.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 75
Dans certains départements, les cultures intercalaires entre les peupliers étaient éligibles pour une durée de 3 ans au prorata de la surface semée, bien que la parcelle relevait du statut foncier de la peupleraie.
En 2000, suite à la mobilisation du monde professionnel agricole et forestier, le gouvernement modifia les modalités de déclaration de surface et de paiements à la surface. Ces modifications intervinrent dans la circulaire DPEI-C2001-4008 du 8 mars 2001. Les cultures intercalaires sont devenus éligibles, quelque soit l’âge des arbres et le lieu géographique.
A la page 14 de cette circulaire, il est spécifié :
Lorsque la culture est pratiquée sur une parcelle arborée, la superficie déclarée pour la culture doit être corrigée proportionnellement au nombre d'arbres, leur emprise étant calculée selon les normes usuelles de votre département. En tout état de cause, la culture arable pour laquelle le bénéfice d'un paiement à la surface est demandé devra pouvoir être effectuée dans des conditions comparables à celles des parcelles non arborées dans la même région.
Des paiements à la surface au titre des cultures arables peuvent être demandés pour des surfaces éligibles nouvellement plantées en jeunes arbres après déduction de l'emprise (que vous établirez forfaitairement et annuellement) des jeunes arbres. Les parcelles doivent porter des cultures éligibles pratiquées selon les usages reconnus localement.
Extrait circ. DPEI-C2001-4008
Le premier paragraphe s’inspire directement de l’article 5 du règlement 2419/2001 de la Commission Européenne. Il ajoute néanmoins deux conditions importantes :
• La surface de la culture éligible doit être réduite de la surface d’emprise des arbres.
• La méthode de calcul de la surface d’emprise des arbres doit suivre les normes usuelles du département
Après la réforme des accords du Luxembourg
Le même principe de calcul des surfaces ouvrant droits à primes est appliqué. Il s’appuie notamment sur l’article 51 du règlement 1782/2003 qui stipule que toute
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 76
surface correspondant à une production permanente doit être soustraite de la surface éligible.
La circulaire DPEI/SPM/SDCPV/MGA/C2004 – N° 4021 du 25 mars 2004 décrit ce principe d’application et la définition des règles à suivre sur les parcelles arborées (paragraphe 2.3.2).
La circulaire reprend également en annexe 9 « les normes locales » et demande à chaque département de prendre les arrêtés définissant ces normes locales.
Suite à la réforme de la PAC, une cellule d’information officielle a été créée au niveau du MAAPAR. Suite à une question concernant l’éligibilité du maïs intercalaire entre des peupliers, la note d’Information n°29 aux DDAF, DRAF et DDSV, rédigée par la Direction des Politiques Economique et Internationale précise :
« Si on peut considérer que la parcelle reste agricole (pour cela, la culture intercalaire doit pouvoir être effectuée dans des conditions comparables à celles des parcelles non arborées dans la même région), elle sera admissible. Dans ce cas, la surface déclarée comme «admissible» devra être corrigée proportionnellement au nombre d’arbres, leur emprise étant calculé selon les normes usuelles du département. »
Extrait note d’information°29 - DPEI
Conclusions
Les agriculteurs agroforestiers pénalisés
Bien que la possibilité était laissée à tout Etat membre de rendre éligible les surfaces arborées pour des motifs environnementaux, la France a décidé de réduire la surface éligible en fonction de la surface d’emprise des arbres, ce qui pénalise fortement les agriculteurs souhaitant conserver ou planter des arbres ruraux. Cette décision va à l’encontre du principe même des Bonnes Pratiques Agricoles. Seules, les surfaces des haies entretenues peuvent être considérées éligibles à condition qu’elles correspondent aux normes locales et que leur largeur soit comprise entre 2 et 4 m.
Conséquences possibles pour des parcelles arborées existantes
Le calcul des droits à paiement est réalisé sur un historique des aides reçues. Lors de cet historique, la surface d’emprise des arbres a normalement été déduite de la surface primable. Dans le cadre de la réforme, la situation ne devrait pas changer. SAUF si l’on considère que les arbres se développant, leur surface d’emprise continue de s’accroître… De ce fait, lors du contrôle, l’agriculteur pourrait voir réduire ses surfaces éligibles.
Une gestion des droits qui incite à l’arrachage
Si les surfaces d’emprise des arbres sont inéligibles, le propriétaire peut être incité à les arracher pour percevoir davantage de droits à prime. En effet, compte tenu des nouvelles possibilités de récupération de droits par le jeu de l’offre et la demande, un propriétaire peut acheter des droits s’il récupère des surfaces éligibles.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 77
De même, en cas de perte de surface éligible suite à des aménagements de territoires (expropriation), afin de ne pas perdre les droits, l’agriculteur peut également être tenté de récupérer des surfaces sur son exploitation afin de conserver ses droits.
Des normes locales quasi-inexistantes
La notion de normes locales est très ambiguë du fait que très peu de départements ont pris les arrêtés préfectoraux correspondants. Ainsi, une récente enquête effectuée par l’association Solagro montre que :
• 1 département sur 6 n’a pas pris (ou pas encore) d’arrêté préfectoral concernant les usages locaux.
• Dans les arrêtés existants, il existe une grande disparité et imprécision concernant les méthodes de prise en compte des haies et de leur entretien.
• Le cas des arbres isolés et des bosquets sont rarement abordés.
• En tant que système innovant, les parcelles agroforestières associant arbres non fruitiers et cultures intercalaires ne sont jamais pris en compte.
Propositions au niveau national
Améliorer la définition des normes usuelles
En absence de normes de calcul de la surface d’emprise des arbres, le calcul de la surface éligible est délicat à mettre en œuvre et reste à préciser dans la plupart des départements français. Une norme nationale apporterait une simplification administrative dans le calcul des droits à paiement.
On peut suggérer que l’emprise des arbres corresponde à la surface de la parcelle qui n’est pas occupée par la culture (surface complémentaire). L’avantage de cette solution est qu’elle facilite le contrôle et les méthodes de calcul. Elle est valable quelle que soit la dimension des arbres ou leur disposition. Elle peut cependant prêter à discussion dans le cas de prairies pâturées, l’emprise des arbres disséminés devenant négligeable puisque tout est pâturé y compris sous les arbres.
Une autre solution serait que la parcelle soit éligible dans sa totalité, ce qui évite le recours à des normes locales de calcul.
Simplifier le contrôle des surfaces arborées
En instaurant un système de contrôle des surfaces par photos aériennes, il est vraisemblable que l’on assiste à une diminution du nombre d’arbres hors forêt ou que les arbres de hautes tiges disparaissent des haies. Si l’on peut aisément compter le nombre d’arbres isolés par photos aériennes, la surface d’emprise correspondant à la projection des houppiers sera plus importante que dans la réalité. On peut reprendre ici la précédente idée de calculer la surface éligible en fonction de la surface occupée par la culture. Il convient lors du contrôle de s’assurer que les cultures soient effectivement menées dans des conditions normales. De même, la
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 78
deuxième option citée, à savoir l’éligibilité totale de la parcelle arborée, aurait le mérite de simplifier les procédures de contrôle.
Admissibilité et éligibilité des parcelles agroforestières
La France a décidé d’opter pour un régime mixte en maintenant une aide couplée à la production (25%) et une aide découplée, en droit à paiement unique (75%).
Cas des aides découplées
Simplifier l’approche administrative
Compte tenu que l’agroforesterie répond :
• Aux 4 objectifs fixés par les bonnes conditions agricoles et environnementales, à savoir :
o protection contre l’érosion des sols grâce au maillage des lignes d’arbres enherbées,
o maintien de la matière organique sous le double effet de l’enherbement et de la décomposition du feuillage et des racines annuelles
o maintien de la structure des sols
o niveau minimum d’entretien, assuré par les animaux dans les zones sylvopastorales.
• Aux enjeux définis par les directives européennes sur l’environnement, en particulier les directives concernant la préservation de la qualité de l’eau (directive 91/676) et la directive sur le bien-être des animaux (directive 98/58),
L’APCA a proposé au MAAPAR que la totalité de la parcelle agroforestière soit admissible aux aides découplées (Annexe 3).
L’admissibilité de la parcelle arborée et l’éligibilité des surfaces comportant des arbres présente le mérite séduisant de simplifier les calculs des surfaces éligibles ainsi que les procédures de contrôle.
Une éligibilité totale soumise à conditions
Pour être déclarée éligible dans sa totalité, la parcelle arborée devra respecter les normes usuelles qui la distinguent de la parcelle forestière. La parcelle doit être majoritairement agricole (culture ou pâture) et la densité d’arbres doit être inférieure à 200 arbres par ha. La surface agricole au sol doit représenter plus de 50 %.
Les arbres double-fin, cultivés pour le bois et pour leur production fruitière, sont éligibles à condition que la hauteur de bille soit supérieure à 2 m et nette de tout point de greffage sur cette hauteur. Conformément à la réglementation, l’exploitant ne pourra cumuler différentes aides sur cette surface :
o soit, il opte pour la déclaration de la surface dans le cadre du DPU,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 79
o soit, il opte pour une déclaration de surface en verger. Dans ce dernier cas, la parcelle n’est plus éligible aux droits à prime mais peut prétendre aux aides vergers (ex aides aux fruitiers à coque).
La densité de 200 arbres/ha peut être discutée. Elle correspond en fait au cas d’une plantation de jeunes arbres sur lesquels il sera pratiqué une éclaircie afin de sélectionner les plus beaux. Sachant que dans tous les cas, la surface au sol doit être majoritairement agricole, on ne pourra conserver un grand nombre d’arbres adulte.
Une alternative serait d’imposer un critère additionnel sur le diamètre des arbres. Le seuil serait calculé en fonction de la taille des arbres présents. La parcelle resterait éligible si elle comporte moins de:
o 50 arbres de diamètre supérieur à 30 cm (les "gros")
o 100 arbres de diamètre supérieur à 15 cm (les "moyens")
Les arbres de petits (diamètre < 15 cm) ne comptent pas dans le critère.
Les diamètres sont pris à 130 cm du sol selon les normes dendrométriques classiques.
En cas de parcelle hétérogène (avec mélange de gros arbres et d’arbres moyens), une règle simple peut être proposée : 1 gros = 2 moyens. Ainsi, le seuil est égal au total des gros arbres auquel on rajoute la moitié du total des arbres moyens. Ce nombre doit être inférieur à 50.
En cas de dépassement du seuil, une surface forfaitaire de 100 m² est retirée pour chaque gros arbre excédentaire. Ainsi, une parcelle perd toute éligibilité à partir du moment où elle atteint 150 gros arbres/ha. Et elle garde 50% d'éligibilité pour 100 gros arbres/ha.
Dans tous les cas de figure, la solution proposée devra simplifier les procédures de calcul et de contrôle. Cette deuxième solution, si elle présente le mérite d’être rigoureuse, peut s’avérer difficile dans son application sur le terrain. Il convient d’en discuter avec les services de contrôle (ONIC et CNASEA).
Conséquences à l’échelle de l’exploitation et au niveau national
Pour apprécier l’impact de l’admissibilité des surfaces arborées, il faut considérer le cas des parcelles arborées existantes des nouvelles parcelles plantées.
o Cas d’une parcelle arborée existante avant la période historique de calcul
Le montant de l’aide unique ne changera pas. Lors de la période ayant servi au calcul de l’aide, les surfaces des arbres ont été déduites des aides à la productions ou aides SCOP. Par contre, l’agriculteur va récupérer une surface admissible ouvrant droit à paiement unique correspondant la surface des arbres de cette période. S’il a la possibilité, il peut donc racheter des droits disponibles.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 80
Au niveau national, il n’y aura pas d’impact puisque le nombre de droits est fixe. Dans les régions de bocage ou de pré-vergers, il y aura une grande surface admissible, auparavant non éligible aux aides SCOP, qui va générer une forte demande de la part des agriculteurs pour acheter des droits. Le montant des droits devrait augmenter dans ces régions du fait du jeu de l’offre et la demande.
o Cas d’une parcelle nouvellement plantée
Dans ce cas, la plantation des arbres ne modifie pas le nombre de droits de l’exploitation qui reste le même qu’avant la plantation.
Cas des aides couplées
Contrairement aux aides découplées, il n’est pas demandé ici de rendre éligible la totalité des surfaces occupées par les arbres.
La raison est simple et relève de la gestion nationale des aides :
o Dans le cas des parcelles arborées existantes, les exploitants pourraient prétendre à des aides couplées qu’il n’avait pas ou qu’ils avaient perdues lors de la plantation (css moins fréquent). Dans certaines régions, ces surfaces occupées par les arbres (arbres isolés, haies, pré-vergers) peuvent être très importantes. L’ensemble des ces surfaces bénéficieraient d’aides couplées. Au niveau national, sachant que le montant total des aides couplées reste identique, on assistera alors à un transfert des aides des zones céréalières non arborées (comme la Beauce ou la Picardie) vers ces régions plus arborées. Cette décision, très politique, aurait peu de chance d’aboutir car les groupements céréaliers ne seraient sans doute pas disposés à partager les aides couplées dont ils bénéficiaient depuis le départ.
o Dans le cas des jeunes plantations, l’idée pourrait être de laisser l’éligibilité totale. Mais, d’une part, cette situation serait alors inéquitable avec les exploitations possédant déjà des surfaces arborées non éligibles. D’autre part, cela compliquerait la gestion administrative des aides. En effet, si cette solution parait simple aujourd’hui, qu’en sera-t-il dans 5 ans, où lors des contrôles, il faudra distinguer les arbres de plus de 5 ans, non éligibles, des moins de 5 ans éligibles… Dans le cas des plantations avec différents âges de plantation, on induit des conditions qui vont rendre difficile tout contrôle ou simplement toute instruction de dossiers d’aide.
Nous proposons toutefois une amélioration de la situation actuelle. En effet, les calculs des surfaces éligibles aux aides couplées sont difficiles à mettre en œuvre comme nous l’avons vu précédemment. Il est en effet très difficile de calculer les surfaces d’emprise des arbres en absence de normes locales clairement définies. De plus, les nouvelles méthodes de contrôle par photos aériennes déforment les surfaces d’emprise.
Il est donc proposé, dans le cadre du calcul de la surface éligible à l’aide couplée, que soit pris en compte la surface réellement occupée par la culture, et non la surface agricole de laquelle on déduit la surface d’emprise des arbres.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 81
Rendre cohérent l’articulation entre 1er et 2ème pilier
Actuellement, les agriculteurs doivent soustraire les surfaces d’emprise des arbres isolés. Dans le cadre du deuxième pilier, et notamment des CAD, ils peuvent prétendre à une compensation dans certains départements calculée au prorata du nombre d’arbres ou de la surface arborée. Ce que l’agriculteur perd d’un côté, il le regagne de l’autre, mais auquel il faut rajouter un surcoût représenté par le temps passé, l’instruction des dossier tant par l’intéressé que par l’administration. En accordant l’éligibilité totale des surfaces arborées, on améliore l’efficacité du système d’aide. Il serait logique dans ces conditions de simplifier les procédures de compensation du deuxième pilier, ce qui constitue également une avancée vers une simplification administrative. On ne conserverait alors que les MAE avec une réelle justification environnementale. Ainsi, il faudra sans doute revoir le contenu de la MAE Habitats Agroforestiers afin de la réserver qu’à des territoires à forts enjeux environnementaux (zone de captage par exemple). Son cahier des charges pourrait revu et le montant de la compensation diminué.
Intégrer les surfaces arborées dans le couvert environnemental
Parmi les dispositions que la France a prises au titre de la conditionnalité des aides, figure l’obligation d’implanter des bandes enherbées le long des cours d’eau, puis au delà sous forme de couvert environnemental, jusqu’à 3% des terres arables.
Au delà de l’obligation d’implantation des bandes enherbées le long des cours d’eau, les surfaces au sol des arbres hors forêt doivent pouvoir être considérées comme couvert environnemental au titre de l’obligation de 3 %.
LES ARBRES DANS LE DEUXIEME PILIER DE LA PAC
Le Règlement de Développement Rural
L’apparition d’une mesure Agroforesterie à l’horizon 2007
Les aides à la plantation et à l’entretien des arbres hors forêt proviennent de différentes mesures du Règlement de Développement Rural (2000-2006), essentiellement les lignes f (MAE), h et i (mesures forestières).
Chaque pays, en fonction de ses priorités, opte pour un choix de mesures dans le cadre du Programme de Développement Rural National (PDRN). Nous ne ferons pas de tour d’horizon des mesures en faveur des arbres hors forêt qui ont été retenues dans chacun des pays membres, ce qui serait beaucoup trop long compte tenu de la diversité des situations. Néanmoins, il convient de signaler que les dispositions en faveur des haies et alignements d’arbres étaient envisagées, ce n’était pas le cas pour les systèmes agroforestiers. Cette situation va certainement évoluer lors du prochain RDR.
En effet, le projet de RDR portant sur la période 2007-2013 dont l’approbation par le Parlement Européen et le Conseil des Ministres de l’Agriculture est prévue en mai 2005, intègre le soutien à l’agroforesterie comme nouvelle mesure.
L’apparition de cette mesure est le résultat du succès de la Recherche Développement en France notamment dans le cadre du programme européen SAFE.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 82
Ainsi, on peut lire dans le Règlement du Conseil concernant le soutien au développement rural par le Fonds européen agricole pour le développement rural (FEADER) du 14 juillet 2004, une description des enjeux à la création d’une mesure de soutien à l’agroforesterie.
Considérant (38)
Les systèmes agro-forestiers ont une valeur élevée du point de vue écologique et social puisqu'ils combinent des systèmes d'agriculture extensive et des systèmes sylvicoles, qui ont pour objectif la production de bois et d'autres produits sylvicoles de grande qualité. Il y a lieu de favoriser leur mise en place.
Considérant 38 du projet de RDR du 14/07/04
Cette considération amène à la proposition d’une mesure en agroforesterie. Celle mesure se situe dans l’axe 2 du RDR au titre de l’aménagement de l’espace. Cet axe 2 comporte deux sous-sections :
1. Les mesures axées sur l’utilisation durable des terres agricoles
2. Les mesures axées sur l’utilisation durable des terres sylvicoles
La mesure de soutien à l’agroforesterie sur terre agricole est incluse dans les mesures concernant les terres sylvicoles (sous-section 2).
La liste des mesures est spécifiée dans l’article 34. On remarquera que la mesure agroforesterie se situe juste après la mesure de boisement de terres agricoles, dans la même sous-section.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 83
L’aide prévue au titre de la présente section concerne les mesures suivantes : a) Mesures axées sur l’utilisation durable des terres agricoles grâce à : i) des paiements destinés aux exploitants agricoles pour les handicaps naturels en zone de montagne; ii) des paiements aux exploitants agricoles situés dans des zones présentant des handicaps, autres que ceux des zones de montagne; iii) des paiements NATURA 2000; iv) des paiements agroenvironnementaux et en faveur du bien-être animal; v) un soutien aux investissements non productifs. b) Mesures axées sur l’utilisation durable des terres sylvicoles grâce à : i) un soutien au premier boisement de terres agricoles; ii) un soutien à la première installation de systèmes agro-forestiers sur des terres agricoles; iii) un soutien au premier boisement de terres non agricoles; iv) des paiements NATURA 2000; v) des paiements environnementaux forestiers; vi) un soutien à la restauration du potentiel de production sylvicole et à l'introduction de mesures de prévention; vii) un soutien aux investissements non productifs. Art 34 du projet de RDR du 14/07/04
Chaque mesure est présentée dans un article spécifique. La mesure Agroforesterie fait l’objet de l’article 41.
Première installation de systèmes agroforestiers sur des terres agricoles
1. Le soutien prévu à l’article 34, point b) ii), est accordée aux exploitants agricoles qui mettent en place des systèmes agroforestiers combinant des systèmes d’agriculture extensive et des systèmes de sylviculture.
L’aide couvre les coûts d'installation.
2. Par «systèmes agro-forestiers», on entend les systèmes d’utilisation des terres qui combinent la croissance d’arbres et l’agriculture sur les mêmes terres.
3. Les sapins de Noël et les espèces à croissance rapide cultivées à court terme ne sont pas admissibles au bénéfice de cette aide.
4. Le soutien est limité aux plafonds fixés à l'annexe I.
Article 41 du projet de RDR du 14/07/04
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 84
Pour la première fois dans un document officiel de la Commission Européenne, le terme agroforesterie apparaît. Notons qu’aucune mesure n’est spécifique aux autres formations arborées hors forêt, telles que haies, alignement ou bosquet. En fait, les mesures de soutien prises au niveau des PDRN pourront dépendre d’une mesure de type agroenvironnementale ou en faveur du bien-être des animaux (article 37). Cette mesure se rapporte aux mesures sur terres agricoles (cf. art 34).
Une mesure incomplète ?
L’incorporation de la mesure en faveur de l’installation de systèmes agroforestiers parmi les mesures sur terres sylvicoles suscite une certaine ambiguïté. Alors que le considérant 38 reconnaît la vocation mixte des parcelles agroforestières, la mesure Agroforesterie proposée aurait pu être incluse parmi les mesures concernant les terres agricoles et non sylvicoles.
En fait, deux mesures en faveur de l’agroforesterie auraient pu être proposées comme le suggère l’APCA dans la lettre au MAAPAR (voir Annexe 3) :
o Une mesure de soutien à l’agroforesterie sur terres agricoles classée dans les mesures sur terres agricoles (1ère sous-section) : il s’agit ici d’un soutien à l’investissement des pré-vergers hautes tiges et des parcelles agrisylvicoles.
o Une mesure de soutien à l’agroforesterie sur terres forestières classée dans les mesures sur terres sylvicoles (2ème sous-section) : il s’agit ici d’un soutien à la mise en place de systèmes agroforestiers dans des bois ou forêts existantes dans un objectif de production fourragère (sylvopastoralisme) ou de production agricole associée aux arbres (fruits forestiers, champignon, …).
Enfin, étant donné le manque d’aide clairement exprimée concernant les autres formations arborées hors forêt, peut-être serait-il judicieux à l’avenir d’envisager une mesure unique en faveur de l’arbre agricole dont l’application serait universelle, se rapportant aussi bien aux haies, pré-vergers qu’aux systèmes agrisylvicoles.
Les aides disponibles en France
Les aides à la mise en place et à l’entretien des formations arborées ne sont pas soumises aux mêmes lignes budgétaires en fonction de leurs caractéristiques.
Le cas des formations arborées hors agroforesterie
Les mesures d’aide à la plantation et l’entretien des haies et arbres isolés relèvent généralement des MAE qui sont retenues dans les Contrats d’Agriculture Durable. Il n’existe pas de cahier des charges national sur ce type de mesure. Selon les départements, des enjeux prioritaires sont fixés par territoire. L’adoption d’une MAE en faveur des haies n’est donc pas systématique et de nombreux départements n’offrent aucune possibilité de soutien à la création de haie.
Néanmoins, il existe également différentes possibilités de financement sur des fonds nationaux forestiers, grâce à la circulaire concernant les conditions de financement des projets d’investissements forestiers ou d’actions forestières à caractère protecteur, environnemental et social du 7 mai 2001 (circ. DERF/SDF/C2001-3010). Dans cette circulaire, que nous appellerons par la suite circulaire Forêt de Protection,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 85
il est spécifié que les haies et alignements peuvent bénéficier de subventions si la largeur des plantations est supérieure à 10 m, en accord avec l’article R.126-36 du Code Rural. Cet article mentionne en fait les caractéristiques des haies et alignements pouvant être protégés par arrêté préfectoral. On peut se poser la question de la légitimité de ce renvoi à cet article, dont l’objectif n’est peut-être pas celui des personnes sollicitant une demande d’aide via la circulaire de protection. Le planteur d’une haie ne recherche peut-être pas forcément sa mise en protection par décision préfectorale.
Article R126-36
Les boisements linéaires, haies et plantations d'alignement susceptibles d'être protégés en application de l'article L. 126-6 du code rural: a) Sont constitués d'espèces ligneuses buissonnantes et de haute tige figurant sur une liste fixée par arrêté du ministre chargé des forêts. Ils sont structurés selon des modalités fixées par ce même arrêté; b) Doivent avoir une surface minimale de 500 mètres carrés. La surface des haies est égale au produit de leur longueur par une largeur forfaitaire, fixée à cinq mètres pour les haies constituées d'espèces buissonnantes et à dix mètres pour les haies d'arbres de haute tige. Extrait de l’article R126-36 du code rural
Par le biais de la circulaire forêt de protection, il est ainsi possible de financer des plantations dans le cadre de projets dont les enjeux sont purement environnementaux (protection des ressources en eau, aménagement du paysage, aménagement de brise-vent, etc.…).
Il convient de souligner que le financement de cette mesure dépend de la ligne i du PDRN et non de la ligne h qui a été suspendue par le MAAPAR jusqu’en 2006. Il est donc possible de solliciter une aide à la plantation pour des haies ou des bosquets contrairement au cas des Boisements de Terres Agricoles conventionnels qui dépendent de la ligne h.
Pour les opérations de boisement ou de reboisement, les prescriptions de la circulaire DERF/SDF/C2000/3021 du 18 août 2000 s’appliquent aux opérations de protection des ressources en eau et des sols, sous les conditions de surface définies ci-dessous.
La surface minimale d’un projet de boisement ou de reboisement susceptible d’être aidé dans le cadre de la protection de l’eau et des sols est de 1 ha d’un seul tenant pour les bosquets et les boqueteaux. Les alignements et les bandes boisées devront couvrir une surface minimum de 500 mètres carrés soit une longueur minimale de 50 mètres (l’article R.126-36 du code rural, relatif aux boisements linéaires, haies et plantations d’alignement susceptibles d’être protégés, fixe en effet une largeur minimale de 10 mètres pour ces structures)
Extrait circulaire DERF/SDF/C2001-3010 du 7 mai 2001 – Chap. 7
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 86
Le respect de ces critères entraîne automatiquement l’inéligibilité de ces surfaces arborées dans le cadre du droit à prime du premier pilier puisque la largeur se situe au dessus du seuil de 4 mètres. On constate donc que peu d’agriculteurs ne souhaitent financer la plantation des haies via cette circulaire et font plutôt appel, lorsque cela est possible à des fonds issus des CAD, plus souples en matière de cahier des charges.
Néanmoins, ces plantations arborées sont éligibles à la prime de compensation à la perte de revenu agricole (PCPR). Cette éligibilité est définie dans la circulaire du 8 août 2001 à la page 10.
- Haies, bosquets, boisements linéaires, plantations truffières
Les plantations qui visent d’autres fins que la production de bois à titre principal, telles que les haies, bosquets, boisements linéaires, plantations truffières, réalisés dans des conditions ouvrant droit au soutien financier de l'Etat (notamment circulaire DERF/SDF/C2000-3010 du 7 mai 2001) ou de collectivités territoriales, sont également éligibles à la prime.
Le montant de la prime sera fixé par calcul de la surface équivalente avec une largeur forfaitaire de 10 m par rang de plantation.
Extrait de la Circulaire DERF/SDF/C2001-3020 - DEPSE/C2001-7034 du 08 août 2001
Par contre, cette mesure est financée grâce à la ligne h du PDRN qui est donc suspendue jusqu’en 2006.
Le cas des systèmes agroforestiers
Les aides à la plantation
Les textes officiels
Bien que le RDR correspondant à la période 2000-2006 ne prévoyait pas de mesure spécifique en faveur de l’agroforesterie, le gouvernement français a fait figure de précurseur en imaginant une aide à la plantation, basée sur le même principe que les aides aux boisements des terres agricoles. Cette disposition allait fortement influencer la création de l’article 41 du prochain RDR.
Ainsi dans la même circulaire forêt de protection citée au paragraphe précédent (circulaire DERF/SDF/C2001-3010 du 7 mai 2001), après une description de l’intérêt l’agroforesterie, une mesure spécifique de soutien est spécifiée dans le chapitre 10 (Annexe 4).
Elle autorise l’application d’une subvention comprise entre 20 et 50% du montant des travaux à la plantation et des 3 premiers entretiens (établi selon un forfait régional ou sur devis estimatif). Ce montant peut-être majoré jusqu’à 20% en fonction du zonage et s’il s’agit d’un projet collectif.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 87
- la plantation d’arbres, à titre expérimental, capables de donner du bois de qualité, dans des parcelles agricoles, dans le cadre d’un projet agroforestier formalisé à l’échelle de l’exploitation agricole, et suivi par un organisme de recherche (INRA, Cemagref, AFOCEL) ou de développement (IDF, CRPF, chambre d’agriculture…). Nota : les caractéristiques de ces expérimentations liées à l’agroforesterie, incluant l’engagement écrit du bénéficiaire de l’aide concernant les soins apportés aux arbres (protections contre les animaux, si besoin est, entretiens, tailles de formation et élagages pendant 15 ans) sont adressées au Cemagref de Nogent sur Vernisson (45) par le DDAF du département d’implantation. Cinq à dix ans après la clôture financière de l’opération, la DDAF adresse à la direction en charge de la politique forestière, à la DRAF et au Cemagref, un rapport technique sur les résultats de ces expérimentations.
Extrait circulaire DERF/SDF/C2001-3010 du 7 mai 2001 – Chap. 10
Une mesure encourageante mais difficile d’application
Pour la première fois en France, le terme « Agroforesterie » est cité dans un texte officiel. Pour la plupart des acteurs agroforestiers, il s’agit d’un grand pas en avant dans la reconnaissance de cette pratique.
Tout agriculteur ou propriétaire en France peut donc solliciter une demande de financement pour la mise en place d’un projet agroforestier.
Mais deux dispositions dans la mesure d’aide ont considérablement freiné son application dans les différents dossiers qui ont été présentés dans les départements français :
o La dérogation à titre expérimental : de nombreuses DDAF se replient derrière cette disposition soit pour refuser le financement, soit, et ce qui est légitime, en demandant un suivi par un des organismes cités. La demande de suivi est souvent liée à l’obligation d’une convention de suivi entre l’organisme et le porteur du projet. Or, rien n’est spécifié sur le contenu du suivi ni sur les modalités de financement qui pourrait finalement incomber au propriétaire.
o Le financement de cette mesure est accordé au titre des mesures environnementales. La vocation de production des parcelles agroforestières n’est pas reconnue dans ce schéma de financement.
D’autre part, le manque d’un cahier des charges explicite dans la circulaire concernant les plantations agroforestières, soulève souvent des questions de la part des techniciens encadrant le projet : le projet doit-il être d’un seul tenant ? Y a-t-il une surface minimale par projet et par îlot ? Peut-on autoriser le mélange pied à pied d’autorisation plus contraignante administrativement dans le cas des boisements forestiers conventionnel ? Peut-on autoriser plus de 4 essences objectifs ? Doit-on forcément se référer à la liste des essences de la circulaire forêt de production DERF/SDF/C2000-3021 du 18 août 2000 pour un projet agroforestier ? Une question également soulevée par les techniciens concerne la surface à prendre en compte dans une parcelle agroforestière en cas de contrôle : la surface totale ou la surface occupée par les arbres ?
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 88
En fait, il est dit clairement que tout projet doit respecter les directives définies par la circulaire DERF/SDF 2000/3021 du 18 août 2000 sauf pour les conditions de surfaces minimales. Celles-ci sont définies pour les haies (50 m de longueur pour 10 m de large) ou pour les bosquets (1 ha) mais par pour les surfaces agroforestières.
Les directives techniques applicables aux BTA qu’il est demandé de suivre vont même à l’encontre de ce qui est généralement proposé en agroforesterie. Ainsi, un mélange pied à pied des essences est tout à fait possible en agroforesterie. D’autre part, un projet agroforestier n’a pas forcément vocation à former un massif continu avec un bois ou une forêt adjacente. Enfin, en agroforesterie, il est possible d’élargir le choix des essences à planter et la liste proposée dans la circulaire DERF/SDF/C2000-3021 est plutôt contraignante pour les porteurs de projets agroforestiers.
Devant les résultats acquis par la recherche prouvant la faisabilité technique et économique de ce type de système d’une part, et la forte demande des agriculteurs en France d’autre part, l’APCA souhaite l’annulation de l’obligation de suivi de la part d’un organisme professionnel. D’autre part, afin de régler certains litiges techniques, il serait souhaitable de clarifier davantage le cahier des charges de l’agroforesterie et de ne pas le calquer sur celui des BTA dont le concept et les objectifs sont très différents.
Enfin, il convient de souligner que le financement de cette mesure dépend comme pour les autres formations hors forêt de la ligne i du PDRN et non de la ligne h qui a été suspendue par le MAAPAR jusqu’en 2006. Il est donc possible de solliciter une aide à la plantation contrairement au cas des Boisements de Terres Agricoles conventionnels qui dépendent de la ligne h.
La compensation à la perte de revenu
Les textes officiels
Parallèlement à l’adoption d’une mesure de soutien à l’investissement, le gouvernement français a également modifié la mesure de compensation de revenu agricole pour boisement sur terres agricoles afin d’inclure une disposition en faveur des plantations agroforestières. Cette modification, comme pour l’éligibilité des haies à la PCPR précédemment abordée, intervient dans la circulaire DERF/SDF/C2001- 3020 - DEPSE/C2001-7034 du 08 août 2001 à la page 11.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 89
- Boisements agroforestiers
La circulaire DERF/SDF/C2000-3010 du 7 mai 2001 précise les modalités de financement par l’Etat et/ou les collectivités territoriales de projets de boisements agroforestiers. Ces derniers sont éligibles à la prime de compensation de perte de revenu. Par ailleurs, la circulaire DPEI/SPM/C2001-4008 du 8 mars 2001 prévoit que des primes à la surface au titre des cultures arables peuvent être versées pour des terres en partie plantées d’arbres, dont l’emprise est déduite des surfaces agricoles éligibles.
Les surfaces éligibles au boisement agroforestier correspondent, dans le respect des itinéraires techniques définis régionalement, à la somme des surfaces boisées, celles-ci ne bénéficiant pas d’un paiement à la surface agricole.
Le demandeur s’assurera annuellement pour chaque parcelle cultivée en agroforesterie que la surface boisée déclarée, cumulée avec la surface agricole déclarée, n’est pas supérieure à la surface totale de la parcelle.
Comme pour les surfaces agricoles éligibles aux paiements à la surface, les surfaces conduites en agroforesterie feront chaque année l’objet d’une déclaration de surface éligible à la prime de compensation de perte de revenu découlant du boisement de terres agricoles. La déclaration sera adressée chaque année par le bénéficiaire à la DDAF avant le 30 avril.
En raison de la variation annuelle de la surface couverte par le boisement, les projets agroforestiers seront intégrés dans l’analyse de risque en vue de la sélection des dossiers à contrôler sur place.
Extrait de la Circulaire DERF/SDF/C2001-3020 - DEPSE/C2001-7034 du 08 août 2001
Le montant de la prime pour les agriculteurs (prime A) est le double de celui de la prime des propriétaires non agriculteur (prime B). Il est compris entre 100 et 350 euros pour la prime A, entre 50 et 175 euros pour la prime B.
Le seuil financier minimum pour la constitution d’un dossier de demande de prime annuelle est fixé à 100 euros, sauf pour les projets de plantation de peupliers et noyers éligibles aux aides à l’investissement de l’Etat. Pour ces dernières essences, la surface minimale par projet est de 1 ha avec 0.5 ha par îlot pour le noyer.
Faut-il maintenir cette disposition ?
Avant toute chose, comme pour les haies et bosquets ou les BTA, cette mesure est financée grâce à la ligne h du PDRN qui est donc suspendue jusqu’en 2006. Aujourd’hui cette aide est donc inaccessible à tout porteur de projet y compris agroforestier.
Tout comme pour les haies, la compensation auquel a droit l’agriculteur est relativement faible compte tenu de la surface en jeu.
En agroforesterie, les surfaces occupées par les arbres représentent entre 5 et 10 % de la surface de la parcelle. Si la prime est de 100 €/ha (montant minimum) cela
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 90
représente entre 5 et 10 € de compensation par ha agroforestier. Si la prime est de 350 €/ha (montant maximum), cela représente entre 17.5 et 35 € par ha agroforestier. Pour atteindre les seuils minimum de présentation d’un dossier qui est de 100 €/dossier, il faut donc réaliser un projet:
o De 10 à 20 ha lorsque la prime est de 100 €
o De 3 à 6 ha lorsque la prime est de 350 €
L’intérêt de présenter un dossier de sollicitude de la PCPR (si celle-ci venait à être rétablie), dépend donc du type de projet et du montant de la prime à l’hectare boisé.
Mais, au-delà de ces aspects financiers, si l’on considère l’agroforesterie comme une pratique agricole, et surtout si l’on attribuait l’éligibilité aux paiements compensatoires de ces formations arborées, l’obtention de la PCPR ne se justifie plus. Pour des raisons de simplifications administratives, il est souhaitable de ne pas modifier l’éligibilité de la parcelle plutôt que de réduire le paiement compensatoire pour solliciter ensuite un paiement de compensation à la perte de revenu…
La MAE Habitats Agroforestiers
Présentation de la mesure
En novembre 2001, le comité STAR de Bruxelles approuve la MAE Habitats Agroforestiers, comportant deux volets : les MAE 2201 et 2202.
« Cette mesure consiste, pour l'agriculteur volontaire, à créer et/ou entretenir des habitats agroforestiers dans des parcelles où les activités agricoles - cultures ou élevage - sont pratiquées en présence d'arbres espacés disséminés sur l’ensemble de la parcelle. »
Comme toute MAE, la mesure compense un surcoût pour l’agriculteur que représente l’adoption d’une mesure environnementale (voir Annexe 5: Texte MAE Habitats Agroforestiers).
La MAE Habitats Agroforestiers distingue le soutien en faveur des jeunes plantations (volet 2201) du soutien en faveur de l’entretien d’habitats de plus de 5 ans (volet 2202).
Quel avenir pour la MAE Habitats Agroforestiers ?
La MAE Habitats Agroforestiers est une MAE reconnue comme mesure nationale dans le cadre des Contrats Territoriaux d’Exploitation (CTE). Tout comme les 5 autres mesures nationales de 2001, tout porteur de CTE pouvait adopter la MAE agroforesterie, quelque soit sa position géographique. Le cahier des charges était identique sur tout le territoire national.
Suite à l’adoption des Contrats d’Agriculture Durable et à l’abandon des CTE, la MAE Agroforesterie tout comme deux autres MAE ont été écartées des mesures nationales. Son application n’est aujourd’hui possible que si les régions l’inscrivent parmi les mesures utilisables dans les CAD départementaux. Cette éventualité a rarement été le cas vu le faible nombre de mesures possibles pouvant être retenues
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 91
au niveau de chaque département. L’APCA a d’ailleurs signalé cette difficulté et demandé au Ministre de l’Agriculture de réintégrer la MAE Agroforesterie comme mesure prioritaire au niveau national. Cette demande a pour l’instant était refusée (voir Annexe 6: Lettre de Luc Guyau APCA au MAAPAR – MAE).
Il est vraisemblable que sans décision allant dans le sens de la demande de l’APCA, le nombre de bénéficiaires de cette mesure restera extrêmement faible…
Faut-il revoir le cahier des charges ?
Dans l’hypothèse où les parcelles agroforestières pourraient bénéficier de l’éligibilité au paiement compensatoire, il serait envisageable de modifier la MAE existante afin de l’affecter uniquement à des systèmes agroforestiers à objectifs clairement environnementaux.
Dans cet objectif, il serait judicieux de revoir le contenu de la MAE Habitats Agroforestiers afin de la réserver qu’à des territoires à forts enjeux environnementaux (zone de captage par exemple). Son cahier des charges pourrait être revu en proposant par exemple des conditions d’enherbement au pied des arbres sur une largeur qu’il faudrait déterminer. Le montant de la compensation serait également à revoir. A priori, il serait légitime de soustraire du montant existant un montant au moins égal au montant du paiement compensatoire pour la surface occupée par les arbres…
Quelles conséquences avec l’application du prochain RDR ?
Une reconnaissance facilitée de l’agroforesterie
L’application de la mesure Agroforesterie de l’article 41 du RDR devrait favoriser une meilleure clarté dans le soutien à l’agroforesterie en France. Les projets agroforestiers pourront être éligible à une subvention à l’investissement sans être considérés comme environnementaux ni expérimentaux.
Homogénéiser le soutien aux formations arborées hors forêt ?
Lors de l’élaboration de la circulaire Forêt de Protection, le gouvernement français avait fait un premier pas pour tenter d’homogénéiser les mécanismes de financement en faveur des formations arbores hors forêt. Mais l’application des différents cahiers des cahiers des charges entraient souvent en collision avec le cahier des charges des règlements d’application de du premier pilier. D’autre part, les sources de financement pour certaines formations arborées, comme les haies dépendaient de deux sources de financement possibles (circulaire protection et CAD).
Lors des premiers travaux concernant le prochain PDRN, il semble que nous allons évoluer vers une programmation à deux volets : un volet national avec des mesures d’application national et un volet déconcentré où les mesures seraient d’application régionale. Une politique de soutien intéressante aux arbres hors forêt pourrait être l’établissement d’une seule mesure, de type nationale, en faveur de l’arbre rural. Une subvention individuelle à l’arbre, avec un plafond délimité par des conditions de densité à l’hectare permettrait sans aucun doute de simplifier les méthodes de soutien actuel. En tout état de cause, cette mesure devra être nationale. Il semble
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 92
essentiel que tout agriculteur puisse recevoir un soutien pour un projet agroforestier ou de plantation de haies, quelque soit le lieu géographique.
BIBLIOGRAPHIE
Clos, Blanchard, (2001) Rapport au Ministère de l’Agriculture « CTE et relation en agriculture et Forêt », 6 p
Coulon F, Dupraz C., Liagre F., Pointereau P. (2000) Etude des pratiques agroforestières associant des arbres fruitiers de haute tige à des cultures et pâtures, Rapport au ministère de l’environnement, 199 p, Solagro/INRA, Fr
Dupraz C., Lagacherie M., Liagre F., Boutland A., (1995). Perspectives de diversification des exploitations agricoles de la région Midi-Pyrénées par l’agroforesterie. Rapport de fin d’étude commandité par le Conseil Régional Midi- Pyrénées, Inra-lepse éditeur, Montpellier, 253 pp.
Dupraz C., Lagacherie M., Liagre F., Cabannes B., (1996). Des systèmes agroforestiers pour le Languedoc-Roussillon. Impact sur les exploitations agricoles et aspects environnementaux. Inra-Lepse éditeur, Montpellier, 418 pp.
Grousset E, Pointereau P, (2005) Rapport sur la prise en compte de l’arbre champêtre dans les soutiens européens, 25 p, SOLAGRO, Fr
Liagre F., (1993). Les pratiques de cultures intercalaires dans la noyeraie fruitière du Dauphiné. Mémoire de Mastère en Sciences Forestières, ENGREF, Montpellier, 80 pp
Mémorandum Agroforestier (1999) 6 p.
Rapports de Recherche du programme européen SAFE Partner 9 : APCA (2002- 2005)
Rapports de Recherche du programme européen SAFE Work Package 9 (2002- 2005)
Roux V., (1996). Les formations boisées hors forêt: aspects juridiques et fiscaux. APCA et Ministère de l’Agriculture et de la Pêche, éditeurs, Paris, 144 pp.
SCAFR, (1999) Mise en place d’un statut spécifique pour les parcelles agroforestières. Rapport final, Ministère de l’Agriculture, DERF, Paris, avril 1999.
Segouin O., Valadon A., (1997) Enquête sur les boisements récents de peupliers en Lot-et-Garonne, Analyse de pratiques agroforestières ; les cultures intercalaires. Cemagref, Nogent-sur Vernisson, 45 pp.
SRFB Languedoc-Roussillon, (1998). Recherche d’un statut pour les parcelles agroforestières. Rapport final du groupe de travail sur l’agroforesterie, 8pp + annex.
Site du programme SAFE : www.montpellier.inra.fr/safe
Appendices
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 93
ANNEXE 1: ARTICLE TRANSRURAL INITIATIVE DE DEC 2004
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 94
ANNEXE 2: PROPOSITIONS DU GROUPE REGLEMENTATIONS – SAFE
Changes in the CAP due to be introduced on 1 January 2005 have major implications for future uptake of agroforestry and the following lobbying statement was agreed. Safe members should translate and send to appropriate authorities in their countries.
The CAP mid-term review approved on 26th June 03 led to Regulation 1782/03 (29th September 03) defining the conditions for the Single Payment Scheme (SPS). This includes a provision that areas of 'woodland' should be excluded from the area of the farm eligible for SPS. Recommendation: this definition of woodland should be clarified to ensure that it does not lead to the removal of trees from farmed landscapes, and accompanying landscape and environmental damage.
Guidance Document (AGRI/2254/2003) recommends that the threshold of 'woodland' is > 50 stems per ha. The specific wording is 'areas of trees - particularly trees with a potential use only for wood production - inside an agricultural parcel with density of more than 50 trees/ha should, as a general rule, be considered as ineligible
Article 5 of Regulation 2419/01 indicates that: 'a parcel that both contains trees and is used for crop production covered by Article 1 of Regulation (EEC) No 3508/92 shall be considered an agricultural parcel provided that the production envisaged can be carried out in a similar way as on parcels without trees in the same area'. This article is the basis of the dispensation in AGRI/2254/2003. It is important that agroforestry remains classified as ‘agriculture’. Recommendation: wording of Article 5 of Regulation 2419/01 should be retained in any replacement Regulation.
There are internationally accepted definitions of ‘forest’ or ‘forest land’ used by the UN-ECE/FAO and the UNFCCC which use threshold values of crown cover, tree height at maturity, minimum area and bounding areas. However ‘woodland’ as used in EU Regulation (1782/03) is less well defined. Recommendation: 50 trees per ha is an acceptable definition of ‘woodland’ for the purposes of 1782/03, but should be clarified to say ’50 trees per ha of more than 15cm diameter at breast height’.
Crop and pastoral production can maintain acceptable production beneath well- pruned trees at densities greater than 50 trees per ha. Recommendation: a) for silvoarable systems - SPS can be paid for the cultivated proportion of a parcel provided that at least 50% of the parcel is cultivated; b) for silvopastoral systems – SPS payments can be maintained provided that more than 50% of the non-shaded pasture production is maintained.
Some countries declare parcels as x% covered by one ‘activity’ or y% covered by another ‘activity’ or even ‘owner’. Recommendation: the EU should make clear to all EU countries that they have the flexibility to allow multiple activities within
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 95
parcels in their national IACS systems (e.g. ‘forestry’ and ‘cropping’ in the same parcel)
A farmer will loose SPS payments if he introduces a non cereal, pasture or fodder crop, including ‘perennial crops’. Recommendation: it should be made clear that that trees planted at agroforestry spacings do not constitute a ‘perennial crop’.
Farmers obtaining the SPS are obliged to demonstrate that they maintain the farm in ‘good agricultural and environmental condition’. Annex IV of Regulation1782/03 indicates that one condition is ‘avoiding encroachment of unwanted vegetation on agricultural land’. Recommendation: national definitions of ‘good agricultural and environmental condition’ could include the phrase ‘well-managed agroforestry is recognized as a mechanism of improving landscape and environmental diversity’.
Regulation 2237/03 Chapter 5 indicates that payments to nut trees orchards will NOT be made if these are intercropped. Recommendation: the annual per nut-tree payment should only be removed if the intercrops are subsidized as part of the Single Payment Scheme.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 96
ANNEXE 3: LETTRE DE LUC GUYAU APCA AU MAAPAR – PAC
Monsieur Hervé GAYMARD Ministre de l’Agriculture, de l’Alimentation, de la Pêche et des Affaires Rurales 78, rue de Varenne 75007 PARIS
Paris, le 16 novembre 2004
Monsieur le Ministre,
L’agroforesterie connaît un écho grandissant auprès des agriculteurs et des collectivités territoriales. Cet intérêt croissant se traduira, en 2005, par 1000 ha supplémentaires de parcelles forestières sur terres agricoles.
L’agroforesterie représente, en effet, un atout au sein de l’exploitation agricole. Elle en enrichit la valeur patrimoniale, améliore les performances agro-environnementales du système d’exploitation et permet à l’agriculteur de maintenir son revenu réel, tout en investissant pour l’avenir.
L’agroforesterie constitue, donc, une voie de développement qui répond à des exigences économiques et environnementales, mais aussi sociétales, du fait de son empreinte dans le territoire et de son impact paysager.
A cet égard, nous nous félicitons que le projet de règlement européen du 14 juillet 2004 ait prévu de soutenir l’agroforesterie, se basant sur les actions de recherche menées conjointement par l’INRA et les Chambres d’Agriculture. Toutefois, la rédaction de l’article 34 introduit une ambiguïté en limitant la mesure du soutien à l’agroforesterie sur les terres forestières exclusivement. Nous vous proposons un amendement afin de clarifier le contenu en distinguant les deux types de système agroforestier : forestier et agricole.
Par ailleurs, les avantages environnementaux tirés de l’agroforesterie sont multiples (biodiversité, paysage, protection climatique, des sols et des eaux), il est donc indispensable que la totalité de la parcelle agroforestière soit admissible aux aides découplées, et que les rangées d’arbres de parcelles agroforestières soient considérées comme couvert environnemental au titre de l’obligation de 3 % dans les bonnes conditions agricoles et environnementales.
La France a acquis une avance reconnue au niveau européen. Il serait donc regrettable que la déclinaison française de l’Accord de Luxembourg et que le nouveau règlement développement rural viennent annihiler l’investissement et les efforts entrepris depuis plusieurs années.
En espérant que ces propositions retiendront toute votre attention, je vous prie d’agréer, Monsieur le Ministre, l’expression de ma haute considération.
Luc GUYAU
PJ : annexes pour propositions d’amendement
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 97 APCA-SDPAR
PROPOSITION DE MODIFICATION DU PROJET DE REGLEMENT EUROPEEN CONCERNANT LE SOUTIEN AU DEVELOPPEMENT RURAL (PROPOSITION DU 14/07/04)
L’axe 2 du RDR au titre de l’aménagement de l’espace comporte deux sous-sections :
1. Les mesures axées sur l’utilisation durable des terres agricoles
2. Les mesures axées sur l’utilisation durable des terres sylvicoles
La sous-section 2 comprend la mesure de soutien à l’agroforesterie sur terre agricole.
Cette rédaction est ambiguë. En effet, cette mesure qui soutient l’agroforesterie sur terre agricole se situe dans la sous-section des mesures forestières. Afin d’éviter toute confusion, il est proposé de distinguer les deux types de systèmes agroforestiers et d’intégrer une mesure de soutien à l’agroforesterie sur terre agricole dans la première sous-section et une mesure de soutien à l’agroforesterie sur terre forestière dans la deuxième sous-section.
Cette distinction demande une modification des articles 34 et 41 ainsi que l’introduction d’une nouvelle mesure donnant lieu à un nouvel article.
Modification de l’article 34
Il est ajouté un point vi à l’article 34 a). La rédaction de l’article 34 a) serait la suivante (modifications proposées en gras) :
Article 34
L’aide prévue au titre de la présente section concerne les mesures suivantes:
a) Mesures axées sur l’utilisation durable des terres agricoles grâce à :
i) des paiements destinés aux exploitants agricoles pour les handicaps naturels en zone de montagne;
ii) des paiements aux exploitants agricoles situés dans des zones présentant des handicaps, autres que ceux des zones de montagne;
iii) des paiements NATURA 2000;
iv) des paiements agroenvironnementaux et en faveur du bien-être animal;
v) un soutien aux investissements non productifs.
vi) un soutien à la première installation de systèmes agroforestiers sur des terres agricoles.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 98
b) Mesures axées sur l’utilisation durable des terres sylvicoles grâce à :
i) un soutien au premier boisement de terres agricoles;
ii) un soutien à la première installation de systèmes agro-forestiers sur des terres
forestières;
iii) un soutien au premier boisement de terres non agricoles;
iv) des paiements NATURA 2000;
v) des paiements environnementaux forestiers;
vi) un soutien à la restauration du potentiel de production sylvicole et à l'introduction de mesures de prévention;
vii) un soutien aux investissements non productifs.
Le point a – vi) donne lieu à nouvel article.
Proposition d’article pour l’agroforesterie sur terres agricoles
Dans la sous-section 1 (Conditions relatives aux mesures en faveur d’une utilisation durable des terres agricoles), on ajoute un nouvel article rédigé comme suit :
Article 39
Première installation de systèmes agroforestiers sur des terres agricoles
1. Le soutien prévu à l’article 34, point a) vi), est accordée aux exploitants agricoles qui mettent en place des systèmes agroforestiers combinant des systèmes d’agriculture extensive et des systèmes de sylviculture.
L’aide couvre les coûts d'installation.
2. Par «systèmes agro-forestiers», on entend les systèmes d’utilisation des terres qui combinent la croissance d’arbres et l’agriculture sur les mêmes terres.
3. Les sapins de Noël et les espèces à croissance rapide cultivées à court terme ne sont pas admissibles au bénéfice de cette aide.
4. Le soutien est limité aux plafonds fixés à l'annexe I.
Modification de l’article 41 (qui devient 42)
L’article 41 de la sous-section 2 (Conditions relatives aux mesures en faveur d’une utilisation durable des terres sylvicoles), concerne la mesure en faveur de l’agroforesterie sur terres forestières. Il convient d’adapter le contenu actuel de l’article.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 99
Proposition de rédaction :
Article 42
Installation de systèmes agroforestiers sur des terres sylvicoles
1.Le soutien prévu à l’article 34, point b) ii), est accordée aux exploitants agricoles qui mettent en place des systèmes agroforestiers combinant des systèmes d’agriculture extensive et des systèmes de sylviculture.
L’aide couvre les coûts de l’aménagement.
2.Par «systèmes agroforestiers», on entend les systèmes d’utilisation des terres qui combinent la croissance d’arbres et l’agriculture sur les mêmes terres.
3.Les sapins de Noël et les espèces à croissance rapide cultivées à court terme ne sont pas admissibles au bénéfice de cette aide.
4.Le soutien est limité aux plafonds fixés à l'annexe I.
Propositions concernant l’agroforesterie dans le cadre de l’application des accords du Luxembourg
1. Agroforesterie et DPU
Le document de travail AGRI/2254/2003 recommande que le seuil pris en compte pour caractériser une parcelle arborée soit de 50 tiges par ha. Au-delà, la parcelle devient inéligible au titre du PU sauf dérogation pour des motifs agro- environnementaux.
Il est également spécifié dans le règlement 1782/03 que l’agriculteur perd ses droits à paiements pour les surfaces mises en cultures pérennes (article 51).
Néanmoins, le premier principe de l’article 8 du règlement d’application 796/2004 spécifie qu’ « une parcelle boisée est considérée comme une parcelle agricole aux fins du régime d’aide « surfaces » sous réserve que les activités agricoles visées à l’article 51 du règlement (CE) n° 1782/2003 ou, le cas échéant, que la production envisagée puissent se dérouler comme elles se dérouleraient sur des parcelles non boisées situées dans la même zone. »
Proposition
Compte tenu que l’agroforesterie répond aux :
- 4 objectifs fixés par les bonnes conditions agricoles et environnementales, à savoir :
protection contre l’érosion des sols grâce au maillage des lignes d’arbres enherbées,
maintien de la matière organique sous le double effet de l’enherbement et de la décomposition du feuillage et des racines annuelles
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 100
maintien de la structure des sols
niveau minimum d’entretien, assuré par les animaux dans les zones sylvopastorales.
- enjeux définis par les directives européennes sur l’environnement, en particulier les directives concernant la préservation de la qualité de l’eau (directive 91/676) et la directive sur le bien-être des animaux (directive 98/58),
Il est proposé que la totalité de la parcelle agroforestière soit admissible aux aides découplées.
Pour cela, la parcelle agroforestière devra respecter les normes usuelles en agroforesterie qui la distinguent de la parcelle forestière, la parcelle doit être majoritairement agricole (culture ou pâture) et la densité d’arbres doit être comprise entre 50 et 200 arbres par ha.
Les arbres double-fin, cultivés pour le bois et pour leur production fruitière, sont éligibles à condition que la hauteur de bille soit supérieure à 2 m et nette de tout point de greffage sur cette hauteur. Conformément à la réglementation, l’exploitant ne pourra cumuler différentes aides sur cette surface :
soit, il opte pour la déclaration de la surface dans le cadre du DPU,
soit, il opte pour une déclaration de surface en verger. Dans ce dernier cas, la parcelle n’est plus éligible aux droits à prime mais peut prétendre aux aides vergers (ex aides aux fruitiers à coque).
2. Agroforesterie - BCAE
Parmi les dispositions que la France a prises au titre de la conditionnalité des aides, figure l’obligation d’implanter des bandes enherbées le long des cours d’eau, puis au delà sous forme de couvert environnemental, jusqu’à 3% des terres arables.
Proposition
Au delà de l’obligation d’implantation des bandes enherbées le long des cours d’eau, les rangées d’arbres de parcelles agroforestières doivent pouvoir être considérées comme couvert environnemental au titre de l’obligation de 3 %.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 101
ANNEXE 4: CHAPITRE 10 DE LA CIRCULAIRE FORET DE PROTECTION DU 7 MAI 01
10 - CREATION OU RESTAURATION DES FORMATIONS ARBOREES HORS FORET
La politique forestière, tout en privilégiant clairement les investissements en forêt, a été progressivement amenée à s’intéresser aux formations arborées hors forêts, telles que les haies, les bosquets et boqueteaux, ainsi qu’à participer à des expérimentations qui peuvent préfigurer d’une nouvelle association entre l’agriculture et la forêt, comme l’agroforesterie. Dans les zones faiblement boisées, de telles formations arborées hors forêt peuvent en effet contribuer à préserver ou restaurer la diversité biologique, à structurer le paysage, à fixer les sols, tout en jouant un rôle de production de bois d’œuvre (pour des essences précieuses) et de feu pour les propriétaires. Elles peuvent donc, sous certaines conditions, bénéficier des aides aux investissements forestiers à caractère protecteur, environnemental et social.
Nota : les dispositions prévues au III (Aides directes) de la circulaire DERF/SDEF/N°3016 du 27 septembre 1995 sont abrogées.
10.1 CONDITIONS GENERALES D’ELIGIBILITE
10.1.1 OPERATIONS ELIGIBLES
Au titre du PDRN le cofinancement est assuré par la mesure i.1, sont éligibles :
- les opérations de plantation destinées à créer de nouvelles haies arborées, selon des critères techniques fixés au niveau régional, sur proposition des préfets de département, et s’inscrivant dans des usages locaux traditionnels, en particulier les haies brise-vent destinées à limiter l’évapotranspiration, ainsi que le renforcement du réseau de boisement linéaire ;
- les opérations de boisement ou reboisement, dans les zones faiblement boisées, de bosquets ou boqueteaux présentant un fort intérêt au titre de la diversité biologique et des paysages, compatibles avec une politique raisonnée d’occupation de l’espace rural, et répondant à des critères techniques fixés au niveau régional sur proposition des préfets de département ;
- la plantation d’arbres, à titre expérimental, capables de donner du bois de qualité, dans des parcelles agricoles, dans le cadre d’un projet agroforestier formalisé à l’échelle de l’exploitation agricole, et suivi par un organisme de recherche (INRA, Cemagref, AFOCEL) ou de développement (IDF, CRPF, chambre d’agriculture…).
Nota : les caractéristiques de ces expérimentations liées à l’agroforesterie, incluant l’engagement écrit du bénéficiaire de l’aide concernant les soins apportés aux arbres (protections contre les animaux, si besoin est, entretiens, tailles de formation et élagages pendant 15 ans) sont adressées au Cemagref de Nogent sur Vernisson (45) par le DDAF du département d’implantation. Cinq à dix ans après la clôture financière de l’opération, la DDAF adresse à la direction en charge de la politique forestière, à la DRAF et au Cemagref, un rapport technique sur les résultats de ces expérimentations.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 102
Les travaux éligibles dans le cadre de la mesure i.1 sont :
- élimination de la végétation préexistante
- préparation du sol
- fourniture et mise en place de graines et plants d’une espèce ou d’une provenance génétique adaptée à la station en conformité avec la réglementation sur le matériel forestier de reproduction en vigueur
- les trois premiers entretiens
- les travaux annexes indispensables (fossés, protection contre le gibier, les insectes ravageurs et les champignons pathogènes) dans la limite des plafonds fixés au niveau régional
- maîtrise d’œuvre des travaux et leur suivi par un expert forestier ou un homme de l’art agréé, avec un montant maximal de 10% du coût total des travaux
- desserte interne au chantier et son raccordement sur une voirie opérationnelle
- étude préalable d’impact environnemental ou d’insertion paysagère pour un montant maximal de 10% du coût total des travaux.
10.1.2 CONDITIONS D’OCTROI DES AIDES
Outre le niveau minimum d’investissement financier requis pour rendre recevable une demande d’aide (1000 Euros) et le respect des directives définies par la circulaire DERF/SDF 2000/3021 du 18 août 2000, en dehors des conditions de surface, les opérations devront couvrir une surface minimum de 500 mètres carrés soit, pour les haies arborées, une longueur minimale de 50 mètres (l’article R.126-36 du code rural, relatif aux boisements linéaires, haies et plantations d’alignement susceptibles d’être protégés, fixe en effet une largeur minimale de 10 mètres pour ces structures). Pour les bosquets et boqueteaux la surface minimale éligible à une aide est de 1 ha d’un seul tenant. Sont considérées comme contiguës les formations arborées séparées par un chemin public ou privé ou par un ruisseau. Les formations arborées protégées en application de l’article L.126-6 du code rural ou de l’article L.130.1 du code de l’urbanisme ou d’une décision préfectorale sont prioritaires à l’octroi des aides.
En application du principe d’exclusion, l’obtention des aides sera uniquement envisagée si les autres possibilités de financement de l’Etat ne peuvent être retenues.
10.1.3 CONDITIONS RELATIVES AUX PEUPLEMENTS
Les opérations de boisement, reboisement ou reconstitution de formations dégradées devront prévoir l’utilisation d’espèces traditionnelles convenant au type de formation souhaitée. Les essences utilisées seront adaptées au sol et au climat de la zone concernée. Les espèces végétales qui ont un comportement envahissant sont à proscrire. Pour la strate arborée des formations arborées faisant l’objet d’une
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 103
aide à l’investissement, les essences objectif sont celles définies par la circulaire DERF/SDF 2000/3021 du 18 août 2000.
10.2 CONDITIONS PARTICULIERES DEFINIES AU PLAN REGIONAL
Hormis le cas des expérimentations, les conditions techniques et financières de mise en oeuvre de ces opérations sont arrêtés par le préfet de région, après consultation de la commission régionale de la forêt et des produits forestiers. Elles sont en cohérence avec les priorités et les programmes d’actions définis par les orientations régionales forestières. Les orientations définies à cet égard par la circulaire précitée du 27 septembre 1995 restent valables.
Dans le cas des haies, il appartient au préfet de région de définir la liste des essences accessoires et des essences d’accompagnement qui seront retenues au niveau régional sur propositions des préfets de départements. Cette liste sera déterminée à partir de l’annexe 1 de la circulaire DERF/SDEF n° 3016 du 27 septembre 1995.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 104
ANNEXE 5: TEXTE MAE HABITATS AGROFORESTIERS
MESURE TYPE NATIONALE
n° 2201 et 2202
CREATION (2201) ET GESTION (2202) D'HABITATS AGROFORESTIERS
I - PRINCIPE
Cette mesure consiste, pour l'agriculteur volontaire, à créer et/ou entretenir des habitats agroforestiers dans des parcelles où les activités agricoles - cultures ou élevage - sont pratiquées en présence d'arbres espacés disséminés sur l’ensemble de la parcelle.
II - AVANTAGES ESCOMPTES POUR L'ENVIRONNEMENT
Selon les zones d'implantation, les essences présentes et les activités agricoles auxquelles sont associés les arbres, les bénéfices escomptés pour l'environnement sont de divers ordres. Tous contribuent à améliorer le caractère durable du système de production agricole en jouant sur la complémentarité des arbres et des cultures, obtenue par un choix judicieux des associations et une gestion technique appropriée. Les différents avantages environnementaux relèvent de 7 catégories :
• Protection des sols
-protection physique contre l’érosion hydrique (amélioration de la macro-porosité du sol par le système racinaire des arbres, permettant une meilleure infiltrabilité ; ralentissement des écoulements de surface par les alignements d’arbres) et éolienne (ralentissement du vent par le maillage d’arbres) ;
-amélioration de la qualité des sols (enrichissement en matière organique par le turn- over racinaire des arbres et l’incorporation de leur litière)
-récupération d’éléments nutritifs minéraux en profondeur par le système racinaire profond des arbres (pompe à nutriments) ;
-stimulation de l’activité des micro organismes du sol : les extrêmes climatiques sont modérés par l’ombrage du houppier des arbres ;
• Protection des eaux
-réduction des risques de pollution diffuse des nappes et rivières par interception des lixiviats, notamment de l’azote, par les racines des arbres, soit sous la zone d’enracinement des cultures, soit dans les écoulements hypodermiques (parcelles en pente) ;
-effet brise-vent et humidificateur de l’air des arbres limitant l’évapotranspiration, donc les besoins en irrigations de la culture;
• Stimulation de la biodiversité
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 105
-Maintien ou reconstitution d’une large biodiversité, par les refuges et les milieux de lisière variés que les arbres procurent au sein d’agrosystèmes cultivés ou pâturés intensifs, en particulier pour les groupes d’espèces suivants :
-végétation au sol sous l’emprise des arbres ;
-bryophytes, lichens, épiphytes : les arbres constituent des milieux souvent obligatoires pour de nombreuses espèces devenues rares ;
-oiseaux (perchoirs pour oiseaux chasseurs, lieux de nidification, refuges contre les prédateurs, protection climatique) ;
-chiroptères (arbres repères pour les déplacements nocturnes) ;
-petits mammifères (rongeurs, insectivores, et leurs prédateurs) ;
-insectes : plus de la moitié de la faune d’insectes est inféodée aux arbres, et de nombreuses espèces ont régressé suite à la généralisation de traitements insecticides à large spectre sur les cultures ; les arbres sont des refuges où de nombreuses espèces peuvent échapper aux traitements et peuvent être des réservoirs d’auxiliaires pour la lutte biologique contre les ravageurs des cultures.
-Gibier : les arbres isolés offrent des refuges en milieu cultivé.
• Fixation du carbone
Fixation à long terme du carbone dans les arbres, sans baisse significative du stock de carbone des sols de la parcelle (important dans le cas de prairies qui peuvent perdre une part de leur carbone organique lors d’un boisement en plein par exemple).
• Bien-être animal
Les arbres offrent une protection contre le soleil, le vent, la pluie, réduisant les dépenses énergétiques corporelles des animaux.
• Qualité des paysages
-création et maintien de paysages semi-arborés, ouverts, pittoresques et sécurisants
-création d’îlots verts en zones de grandes cultures intensives ;
• Protection contre les incendies
Les habitats agroforestiers sont incombustibles par nature : pas de strate basse combustible, arbres espacés, culture intercalaire ou pâture entretenue. Dans les zones méridionales, ces habitats peuvent contribuer efficacement à l’entretien des coupures vertes, avec des cultures intercalaires de vigne par exemple.
III - CONDITIONS D'ÉLIGIBILITÉ
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 106
Pour la création d’habitats agroforestiers, les surfaces pour lesquelles l’aide est demandée doivent être situées:
en priorité, dans des zones peu arborées (moins de 5% de la SAU de l’exploitation occupée par des arbres hors forêt);
dans des zones où les bénéfices agri-environnementaux attendus de la création ou l'entretien d'habitats agroforestiers sont confirmés par le service environnemental de la DDAF;
dans des zones de forte déprise agricole lorsque la présence d'habitats agroforestiers permet de maintenir une activité agricole sur des surfaces menacées de se boiser naturellement en générant des risques naturels (fermeture du paysage, incendie, etc.)
Pour la gestion d'habitats existants, les surfaces doivent répondre aux conditions techniques définissant un habitat agroforestier.
IV - ENGAGEMENTS DU CONTRACTANT
Le contractant qui crée un habitat agroforestier s'engage pour une durée de cinq ans à:
•choisir dans la liste annexée des essences d’arbres adaptées aux conditions pédoclimatiques de la parcelle et compatibles avec les pratiques agricoles (engins, animaux), notamment par un port arboré de hauteur suffisante (2 mètres au moins de hauteur de tronc sans branche); ce choix doit être validé par la DDAF.
•ne pas planter des espèces envahissantes;
•se conformer à la réglementation en vigueur pour les essences dont la plantation est encadrée;
•respecter un espacement de 10 à 40 mètres entre les lignes d’arbres, de 4 mètres minimum entre les arbres sur la ligne de plantation ;
•planter entre 50 et 200 arbres/ha. Pour le cas particulier des peupliers et des noyers (à bois ou double fin), la densité de plantation sera comprise entre 50 et 100 arbres/ha.
•planter une surface minimale de 0,5 ha;
•planter au cours de la première année du contrat au moins le nombre d'arbres pour lequel l'aide a été accordée;
•Conduire les arbres de manière à obtenir à terme un arbre adulte avec un tronc sans branches de 2 m de hauteur au moins (arbres de basse ou moyenne tige exclus)
•pratiquer une culture ou une pâture intercalaire entre les arbres;
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 107
•remplacer les plants n'ayant pas pris pour maintenir au moins 50 arbres/ha;
•protéger les troncs en fonction des contraintes agricoles (petits animaux, gros animaux, gibier, machines);
•entretenir les arbres pour maintenir la compatibilité avec les pratiques agricoles dans les conditions habituelles à la région (élagage du tronc, entretien du sol de la bande des arbres);
Il s'engage en outre à respecter les bonnes pratiques agricoles définies dans le PDRN sur l'ensemble de l'exploitation.
Le contractant qui gère un habitat agroforestier existant s'engage pour cinq ans à:
•maintenir un tronc sans branches d’au moins deux mètres de hauteur ;
•pratiquer une culture ou une pâture intercalaire entre les arbres;
•maintenir au moins 50 arbres/ha (regarnis possibles à tout moment, y compris la première année pour atteindre le seuil de 50 arbres/ha à partir d’un habitat de trop faible densité);
•respecter un espacement de 10 à 40 mètres entre les lignes d’arbres, de 4 mètres minimum entre les arbres sur la ligne;
•protéger les troncs en fonction des contraintes agricoles (petits animaux, gros animaux, gibier, machines);
•entretenir les arbres pour maintenir la compatibilité avec les pratiques agricoles dans les conditions habituelles de la région (allégement du houppier, entretien du sol de la bande des arbres).
Il s'engage en outre à respecter les bonnes pratiques agricoles définies dans le PDRN sur l'ensemble de l'exploitation.
V - MONTANT DE L’AIDE
Pour la création d'habitats agroforestiers, le montant de l’aide est de :
Avec culture intercalaire
Aide de base : 240€/ha/an Action 2201 A Aide si CTE : 288€/ha/an
Marge Natura 2000 : 20%
Avec pâturage de petits animaux
Action 2201 B Aide de base : 250€/ha/an
Aide si CTE : 300€/ha/an
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 108
Marge Natura 2000 : 20%
Avec pâturage de gros animaux
Aide de base : 362€/ha/an Action 2201 C Aide si CTE : 434€/ha/an
Marge Natura 2000 : 20%
Le montant de l’aide est calculé pour la mise en place de l’habitat en première année et pour 4 années de soins aux arbres, pour une densité de 100 arbres/ha. L’agriculteur peut planter à ses frais des arbres supplémentaires, dans le respect des fourchettes techniques définies ci-dessus (maximum de 200 arbres, de 100 arbres pour les peupliers et les noyers).
Inversement, si la densité plantée est inférieure à 100 arbres/ha, dans le respect des fourchettes techniques, l’aide sera calculée au prorata du nombre réel d’arbres plantés (minimum 50 arbres/ha).
Pour la gestion des habitats agroforestiers, le montant de l'aide est indépendant du nombre d'arbres, à condition que l'habitat comprenne au moins 50 arbres. En revanche, le montant de l'aide est fonction de la nature de la pratique agricole intercalaire ainsi que de l'âge des arbres :
Avec culture Avec pâturage Avec pâturage intercalaire petits animaux gros animaux
Age des arbres <20 >20 <20 >20 <20 >20
Action numéro 2202A 2202B 2202 2202D 2202E 2202F C
Aide de base €/ha/an 102 140 95 114 114 114
Aide si CTE €/ha/an 122 168 114 137 137 137
Marge Natura 2000 20 % 20 % 20 % 20 % 20 % 20 %
Si des aides des collectivités territoriales ou professionnelles sont disponibles pour la création ou la gestion d'habitats agroforestiers, le montant de l’aide au titre de la MAE sera calculé en déduisant des valeurs précédentes le montant des autres aides accordées.
VI - JUSTIFICATIONS DU MONTANT DES AIDES DE LA MESURE "CREATION ET GESTION D'HABITATS AGROFORESTIERS "
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 109
L’aide est calculée sur la base de la création d’habitats de 100 arbres/ha ou de l’entretien d’habitats de 50 arbres adultes par hectare.
CREATION D'HABITATS AGROFORESTIERS
L’introduction de 100 arbres par hectare dans une parcelle agricole cultivée ou pâturée entraîne des surcoûts pour l’activité agricole qui sont les suivants :
augmentation des frais de culture en intercalaire: 750F/ha/an (soit 2,5 heures de travail mécanisé par hectare .) correspondant au ralentissement des opérations mécanisées par les arbres (manœuvres en bout de parcelle, passage entre les arbres, précautions pour préserver les racines, précautions lors des traitements aux cultures, épandage de précision des intrants).
augmentation des frais de conduite des pâtures en intercalaire : 250 F/ha/an (soit 1 heure de travail mécanisé par hectare) correspondant au ralentissement des opérations mécanisées de fauchage des refus, épandage des amendements organiques, contrôle des infestations de plantes nuisibles sous les arbres qui risquent de gagner sur le pâturage, entretien des zones de piétinement par les animaux près des arbres. Cette estimation est un minimum correspondant aux terrains plats. Un temps plus long est nécessaire dans les terrains en pente.
gestion de l’emprise au sol des arbres pour favoriser la biodiversité sans nuire aux pratiques agricoles intercalaires : 320 F/ha/an quelle que soit la pratique agricole intercalaire ;
achat, pose et entretien des protections des plants d’arbres contre les activités agricoles :
Le coût des fournitures : les prix des manchons à l'unité sont respectivement de 10F pour les cultures (manchon de 120 cm, piquet de 150 cm); 30F pour les pâturages de petits animaux (manchon de 170 cm, piquet de 150 cm, contre-piquet de blocage); 60F pour les pâturages de gros animaux (manchon de 230 cm; 2 piquets de 250 cm, spirale de barbelé) .
La mise en place des protections : en 1 heure on pose respectivement 30 protections (culture), 15 (pâturage de petits animaux) et 8 (pâturage de gros animaux), y compris le temps de distribution du matériel auprès de chaque arbre. Pour un coût de l’heure de travail de 150F, le coût de pose d’une protection est de 5 F avec culture, 10 F avec pâturage de petits animaux et 20 F avec pâturage de grands animaux.
L’entretien des protections : il faut, au cours des années qui suivent la plantation, retendre les attaches, redresser les manchons inclinés, remplacer les manchons détériorés. On estime ce coût à 2, 3 et 4 F/manchon et par an pour les trois types de protection respectivement, soit 160, 240 et 320 F/ha/an (coût de 4 entretiens annuels répartis sur 5 années). Cela correspond à un temps de travail de 1h20, 2h et 2h40/hectare/an selon les trois types de protection.
conduite des arbres : pour assurer le bon développement de l'arbre, il faut chaque année, pendant les cinq premières années, entretenir le tronc dans le
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 110
manchon (dépose, soins à l'arbre, repose du manchon). On estime que 3 heures de travail par hectare et par an sont un minimum. On propose une aide de 18 F/arbre sur 5 ans, soit 360F/ha/an.
Création d’habitats agroforestiers Action 2201A Action 2201B Action 2201C
Avec pâturage Avec pâturage Avec culture Différents types de coûts de petits de gros intercalaire animaux animaux
Augmentation des frais de culture en 750 250 250 intercalaire
Gestion de l’emprise des arbres au sol 320 320 320
Achat des protections4 200 600 1200
Mise en place des protections1 100 200 400
Entretien des protections5 160 240 320
Formation des arbres compatible avec 360 360 360 la pratique intercalaire
Total des surcoûts Aide si CTE 1890 1970 2850 (F/ha/an)
Aide de base 1575 1642 2375 (F/ha/an)
Note : Les coûts de plantation (achat des plants, préparation du sol) et le manque à gagner résultant de la diminution de surface cultivée ne sont pas pris en compte dans le calcul de l’aide. Le cas échéant, ces aspects pourront bénéficier de mesures prévues à cet effet.
GESTION D'HABITATS EXISTANTS
Surcoûts générés pour l’activité agricole par la présence de 50 arbres/ha âgés de plus de 5 ans. Les coûts sont calculés pour 50 arbres par hectare, les coûts générés par un nombre plus élevé d’arbres sont à la charge de l’exploitant.
augmentation des frais de culture en pratique intercalaire quel que soit l‘âge des arbres : temps de travail accru des opérations mécanisées (manœuvres en bout de parcelle, passage entre les arbres, précautions pour préserver les racines, précautions lors des traitements aux cultures, épandage de précision des intrants, fauchage des refus liés aux arbres). Lorsque les arbres atteignent une taille
4 Coût en première année réparti sur 5 ans
5 Coût réparti sur 5 ans de 4 opérations annuelles d’entretien.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 111
importante l'agriculteur peut avoir à modifier son équipement en machines. Le temps de travail est le même que dans le cas de la création d'un habitat mais la surface réellement cultivée est plus faible, ce qui ramène le surcoût réel à 450F/ha/an . Pour les pâtures, les surcoûts sont les mêmes que pour la création d'habitat (250 F/ha/an), le contrôle des espèces nuisibles dans les pâtures étant encore plus délicat avec des arbres de forte taille.
Gestion de l’emprise au sol des arbres pour favoriser la biodiversité sans nuire aux pratiques agricoles intercalaire, quel que soit l’âge des arbres : 250F/ha/an (gyrobroyage de la végétation naturelle de la bande des arbres);
protection des troncs des arbres âgés de moins de 20 ans contre les activités agricoles, calculés sur la base de 50 arbres protégés par hectare :
Pour le pâturage, le remplacement des manchons rigides par des filets à large maille est nécessaire dès lors que le tronc de l'arbre atteint une taille presque égale au diamètre du manchon. La protection des troncs est indispensable tant que les arbres ne sont pas autodéfensables. L’amortissement et l’entretien des filets est estimé à 5 F/arbre/an avec de petits animaux et 8F/arbre/an avec de grands animaux, soit des coûts respectifs de 250 et 400 F/ha/an.
Pour les cultures, la protection des arbres consiste à faire des traitements cicatrisants pour les blessures occasionnées aux troncs par le passage des machines (elle est calculée sur la base de 2F/arbre/an, soit 100 F/ha/an, ce qui ne couvre pas la totalité des dépenses prévisibles);
entretien des houppiers d’arbres âgés de plus de 20 ans : on estime que deux interventions de 8mn par an et par arbre sont nécessaires en cinq ans pour relever la base de la couronne des arbres. Les branches des arbres isolés, même élagués à 4 mètres, ploient vers le sol et finissent souvent par toucher le sol. Dans ce cas, le passage des machines devient impossible. Il faut donc régulièrement écimer les branches retombantes pour éviter d'avoir une trop grande surface inaccessible pour les travaux. Les travaux nécessaires pour une bonne conduite des travaux agricoles sont: élagage latéral, allégement des branches basses, émondage, éclaircissage du houppier. Cela représente un coût de 400 F/ha/an pour 50 arbres.
Gestion d’habitats agroforestiers
Avec culture Avec pâturage Avec pâturage intercalaire Différents types de coûts petits animaux gros animaux
Age des arbres <20 >20 <20 >20 <20 >20
Action numéro 2202A 2202B 2202C 2202D 2202E 2202F
Augmentation des frais de culture en 450 450 250 250 250 250 intercalaire
Gestion de l’emprise des arbres au 250 250 250 250 250 250 sol
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 112
Protection des troncs (arbres de 100 250 400 moins de 20 ans)
Entretien des houppiers (arbres de 400 400 400 plus de 20 ans)
Total des surcoûts Aide si CTE 800 1100 750 900 900 900 (F/ha/an)
Aide de base (F/ha/an) 667 917 625 750 750 750
Note : le manque à gagner résultant de la diminution de surface cultivée ou pâturée à cause des arbres n'est pas pris en compte dans le calcul de l’aide. Le cas échéant, il pourra être pris en compte au titre d'autres mesures prévues à cet effet.
VII - SUIVI DE LA MESURE
L'application de la mesure fera l'objet d'un suivi pendant trois années (2002 - 2004). Au terme de cette période, un bilan sera établi sur les points suivants;
1. Le nombre d'hectares d'habitats créés et d'habitats gérés qui bénéficient de la mesure, dans les trois catégories prévues (cultures, pâturage de petits animaux, pâturage de gros animaux);
2. Pour chacune des catégories, une typologie des associations activité agricole/essences présentes;
3. La part de la surface agricole utile de l'exploitation occupée par les habitats agroforestiers;
4. Les caractéristiques techniques des habitats (densité, espacement des lignes, espacement des arbres sur les lignes, habitats monospécifiques ou plurispécifiques, etc.);
5. L'évolution des pratiques et techniques agricoles liée à la présence des arbres, y compris les innovations techniques introduites par les exploitants;
6. Les incidences sur l'économie de l'exploitation.
Les éléments nécessaires au bilan seront recueillis au niveau départemental et transmis à la DERF qui en fera la synthèse.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 113
ANNEXE 6: LETTRE DE LUC GUYAU APCA AU MAAPAR – MAE
Monsieur Hervé GAYMARD Ministre de l’Agriculture, de l’Alimentation, de la Pêche et des Affaires Rurales 78, rue de Varenne 75007 PARIS
Paris, le 23 avril 2003
Monsieur le Ministre,
Je me permets d’attirer votre attention sur les conséquences dommageables du nouveau dispositif des CAD sur l’agroforesterie.
Cette pratique qui consiste à associer sur une même surface une production agricole (culture ou élevage) et une production sylvicole (arbres plantés à faible densité) intéresse les agriculteurs à plusieurs titres. Elle permet : de constituer un patrimoine de valeur par la plantation d’arbres ayant vocation à produire des bois de qualité, tout en maintenant un revenu agricole ; de préserver l’environnement. En effet, la présence régulière, sur des parcelles agricoles, d’arbres à faible densité améliore la structure des sols et freine l’érosion ; de concilier la constitution d’un capital bois sans abandon de l’activité agricole d’origine, présentant ainsi une alternative au boisement des terres agricoles. Les parcelles ainsi plantées sont de plus facilement réversibles ; de créer des paysages originaux très appréciés.
L’agroforesterie reçoit, de plus, un accueil très favorable auprès de la Commission Européenne et la France est en pointe sur ce dossier.
Dans ce contexte, je voulais vous faire part de mon étonnement à la lecture de la première circulaire, DEPSE/SDEA/C 2003-7007 du 12 mars 2003 sur les modalités d’élaboration des contrats types dans le cadre des CAD. En effet, celle-ci ne fait aucune mention quant au mode d’application de la MAE “Habitats agroforestiers”.
Approuvée par Bruxelles en 2001, cette MAE qui permet de soutenir la création et l’entretien de parcelles agroforestières avait, ensuite, été inscrite au PDRN en tant que mesure nationale.
Mais cette mesure a été retirée de la liste des mesures nationales. Pourtant, l’agroforesterie intéresse toute exploitation, sans véritable restriction géographique. Son application ne sera donc possible que si les régions l’inscrivent parmi les mesures utilisables dans les CAD locaux. Ceci est peu probable, compte tenu de la liste réduite des mesures retenues pour chaque CAD et de la nouveauté de cette pratique.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 114
Je vous propose, donc, de réintégrer la MAE “Habitats agroforestiers” en tant qu’action agroenvironnementale d’application nationale dans le contrat type départemental aux côtés des mesures de conversion à l’agriculture biologique et de préservation des races menacées.
En espérant que vous pourrez tenir compte de notre proposition, je vous prie d’agréer, Monsieur le Ministre, l’expression de ma haute considération.
Luc GUYAU
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 115
ANNEX 5. The distribution of silvo-arable systems in western Europe and their ecological characteristics
Attachment 1 to WP8 Report
R G H Bunce
Alterra Green World Research
P 0 Box 47, 6700 AA Wageningen
The Netherlands
INTRODUCTION
The inherent nature of silvo-arable systems are that it is necessary not only to know the characteristics of the tree cover, but also the crop or ground cover beneath them. The term is therefore a combination of land cover and land use, in the general use of these terms. lt follows that remote sensed images can only give an indication of their occurrence on the ground, although local knowledge can assist in interpretation. Thus the “open sclerophyllous forest“ class of the CORINE land cover map can be used to indicate dehesas or montados but cannot give information as to whether there is fallow land, crops or grass between, or beneath, the trees. Aerial photo interpretation can give more detail e.g. on the density of the trees, but again cannot determine what is growing on the ground. Field visits are therefore necessary not only to determine the presence of silvo-arable systems but also to obtain measurements of their extent and characteristics.
Expert local knowledge can be used to describe the principal characteristics of silvo- arable systems and an overview of their extent. However, whilst this approach gives a good overview of local conditions, it cannot provide objective overall estimates because the relationship of the local area to the whole domain is not known. The present chapter therefore provides a worked example of a procedure that could be applied to the whole of Europe and its applications to Atlantic Europe. Some case studies are then described for southern Europe before suggesting a possible future approach for obtaining estimates.
A EUROPEAN STRATIFICATION SYSTEM FOR RESOURCE ASSESSMENT
The need for detailed field survey on the one hand and an accompanying policy requirement for strategie estimates has long been recognized in landscape ecology. Such apparently opposing needs make it essential to use sampling and then to have a system of relating the samples to the whole population - comparable to socio- economic surveys of voting intention or opinion polls. Such an approach was initiated at a regional level in the mid 70s by Bunce (1975) with Sheail & Bunce (2003) describing its eventual development at a European scale. The approach is based on the regression principle of ecological parameters being related to environmental factors. At a regional and European scale, altitude and climate data can be recorded and classified using modern statistical methods into relatively homogenous classes which can then be used as strata for sampling. The methodology hus been utilized in
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 116
GB to assess ecological resources since 1978 (Haines-Young et al. 2000) and at a national level in Spain for habitat change using aerial photographs in the SISPARES project. European strata have now been produced (Mucher et al. 2004 and Metzger et al. (in press)) and were used in the worked example for Altantic Europe described in the next section.
A 1km square sampling unit was used in GB as representing a scale suitable for field sampling but also enabling all units to be classified at a national level. Elena- Rossello (1997) developed an approach at different scales according to the heterogeneity of the topography but modern statistical packages have enabled the European strata to be constructed in one analysis at the 1km square scale. In GB dispersed random 1km squares were drawn from die environmental strata and surveyed in the fleld for land cover, habitats, vegetation and soil. National estimates of extent were then made using standard statistical procedures. Once such representative field samples are available, they can be used for modeling exercises for example, for assessing the potential for wood energy plantations in GB (MitchelI et al. 1993) by estimating potential yield and value of trees and comparing it with current agriculture, using a similar procedure to the one adopted in SAFE. Bunce et al. (1987) showed how such a procedure could be used in Spain and Jones et al. (1995) how the framework could be used to model potential changes in agricuitural enterprises.
A WORKED EXAMPLE OF THE APPLICATION OF STRATA IN ATLANTIC EUROPE
A good example of resource assessment required by a policy customer, involving a comparable requirement for fleld survey and strategic estimates is provided by the Veteran tree survey of Atlantic Europe as described by Smith & Bunce (2003, 2004). This project was initiated because of controversy concerning the extent of veteran, i.e. those over 150 years old in GB, as compared with elsewhere in Atlantic Europe. The customer, English Nature needed such figures to establish an appropriate policy for maintaining the resource. For practical reasons, the extent of the survey was restricted to the Atlantic Zone, as described by Mucher et al. (2003) which is hierarchically drawn from linked strata of the full survey of 84classes, with other regions such as Alpine (south) and Mediterranean (north) not being sampled, although they could be subsequently included using the same procedure. 31 sites were taken at random from these classes with three 1km squares surveyed at each using a standard list of habitats based on the GB Countryside Survey (Haines-Young et al. 2000) recorded. Further details were added for wood pastures, which are silvo- pastoral systems and their current states such as whether they were still in use or abandoned. Details of the veteran trees were recorded as described by Smith & Bunce (2004) and the strata used to obtain estimates of tree resource and their distribution in the landscape. The results showed that the majority of veteran trees were outside GB - which was the opposite view of what was believed before the survey. However, GB did have more of the largest size category, trees which were mainly in parklands, rather than field boundaries. Whilst these results are not important to SAFE they do show how the stratification system can be used and that the results do not always reflect expert judgement.
The area of wood pastures (silvo-pastural) recorded was very low but there were some records. There was no silvo-arable in the sample, although one plantation of
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 117
Walnuts was recorded in northern France but did not have a crop between the trees. Effectively therefore, silvo-arable systems are absent in Atlantic Europe, although they are known to occur in experimental sites and other isolated cases.
In addition 11 1km squres taken at random in Asturias, north-west Spain were surveyed using the same methodology, but were not included in the analyses. There was no silvo-arable and only a few small areas of silvopastural. In the Picos de Europa, north-west Spain 30 1km squares were surveyed as described by Bunce et al (1996), no silvo-arable was present but orchards of walnut and apples over grass, sometimes cut for hay. During work over almost 20 years in this region‚ no more silvo-arable has been seen.
This project therefore demonstrates the procedure of applying the European stratification to estimate a land use resource.
CASE STUDIES IN SOUTHERN EUROPE
Following the SAFE Meeting in Toulouse, it was decided to extend the sampling to Southern France and Spain. However sufficient time and resources were only available to carry out some individual site surveys, which are effectively case studies. These are described below.
1. Castelnandry (Southern France). Three 1km squares were surveyed using the same procedure as the veteran tree survey. No silvo-arable systems were recorded, although one square had several hectares of walnut plantations.
2. Port Vendres (Southern France). Three 1km squares were surveyed, with many vineyards but no silvo-arable.
3. Cadalso los Vidrios, Navaluenga and Almorox, Gredos mountains (Central Spain). Three 1km squares were surveyed, with vineyards and fruit trees, but no silvo-arable, although one field had a line of lives among the cereals. There were also about 20 hectares of dehesa but with grassland beneath the trees. To the south of one of the actual samples, there were extensive areas of dehesa with crops between and beneath the trees. Further south there were also extensive areas of cereal dehesas with varable tree densities.
Other site visits were made during the course of excursions elsewhere in Spain and Portugal, and whilst they cannot be used in any quantitative way they are illustrative of the occurrence of silvo-arable in southem Europe.
1. Matute, central Delamada mountains, Rioja (Northern Spain). Two areas of about 5 hectares: (a) walnut and cherry over vegetables; (b) poplar over wheat and vegetables. Both areas were in narrow strips beside small rivers and were irrigated. Above and below the two sites there were extensive plantations of Poplar, all of which were intensively managed with mainly bare ground between the trees. All the plantations were in narrow corridors beside the river and occupied only a small part of the total landscape, which was otherwise cereal fields or Mediterranean scrub and forest.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 118
2. Ezcaray, west Delamada mountains, Rioja. Where the river from the mountains joined the main valley, there was about 20 hectares of Walnut about 20 years old in rows over a wheat crop. Once in the lowland plains there was only intensive cereals. Above the silvo-arable site there were a series of Poplar plantations extending along the river until at higher altitudes the steep valley sides were covered in Beech forest.
Because of the intensive management of the Poplar, there was little ground vegetation that there is likely to be few ecological benefits from the vegetation point of view. The silvo-arable sites were also intensively cultivated so were comparable. The benefits are therefore from additional diversity in the landscape and for cover for birds and insects. The scale of the small river valleys means that there is likely to be limited effect on restricting erosion. In both areas the silvo-arable sites were within an existing network of plantations, whereas new sites in cereal dominated landscapes would have a greater impact both in visual and ecological terms.
Both these sites were so restricted in area that they would be unlikely to be picked up by dispersed random samples. This is in contrast to the next site which is comparable to Extramadura in the extense of dehesa with crops.
3. Monticola, (south-east Portugal). This is an area in south-east Portugal with extensive montado, an open forest landscape with various densities of scattered trees. Anna Keersma has studied this area using aerial photographs and with these categories of tree cover - over 30% canopy cover; 10-30% and less than 10%. Although there is no difference between the proportion of arable/fallow in the different categories, there is over 1000 ha of the three categories under crop or fallow covering over 25% of the land surface.There is also about 10% of the land under crops with the rest being mainly different types of forest and scrub. As in Extramadura therefore a high proportion of the landscape is an active silvo-arable system. Pineda (2003) gives a figure of 6-8 M Ha of dehesa in Spain but figures are not available to separate that into silvo-arable and silvo-pastural. If the Momticola case was representative, then about 25 % could be silvo-arable, which from general observation could be realistic, then there could be over 2 M Ha in Spain alone. The SAFE case studies in Spain, Italy and Greece will provide more details.
In general terms however it can be definitely stated tht silvo-arable systems cover large areas in Mediterranean Europe but only a few patches are present in the Atlantic Zone.
ECOLOGICAL CONSIDERATIONS OF SILVO-ARABLE
The visits to the Silsoe and Montpellier sites, together with experience of Walnut and Poplar plantations elsewhere, indicates that a well managed silvo-arable site is comparable in its ecology to a crop monoculture with additional trees. In areas less intensively managed along the lines of trees, there were residual patches of weed species, both annual and perennial which could contribute to biodiversity both in
terms of fauna and flora - especialiy if the latter contained some of the rarer arable weeds. Limited
information can be gained from most existing poplar plantations without crops, as these are usually managed in a different way. There is also a major difference
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 119
between new poplar plantations within an existing matrix of older patches, which adds to the present network, as opposed to new silvo-arable patches in an otherwise intensively formed landscape. These are likely to have a much greater, and perhaps beneficial, contribution to both biodiversity and landscape complexity. In particular, they may act as stepping stones and refugia for anirnals, especially birds, which are moving through otherwise ceral dominated land. The current states of the landscape is therefore important in determining the potential contribution of new silvo-arable patches.
The Situation is very different in dehesas and montados. Here there is already an existing functioning network of mature silvo-arable systems with well established high-levels of biodiversity in both flora and fauna. This is recognized in many texts e.g. Pineda (2003) and Gomez-Sal (2003) and by die establishment of agri- environment schemes to maintain such systems in Extremadura.
FUTURE WORK
Characterisation
In many parts of southern Europe, lines of trees, whether vines, olives, nut trees or fruit trees, are important elements in the landscape. lt is essential, as emphasized by Haines-Young (2000) and Sheail & Bunce (2003) that consistent definitions are required for any objective baseline survey. The procedure developed in the BioHab project (a framework for linking Biodiversity and Habitats) is suitable for this and is described on www.biohab.alterra.nl. As far as silvo-arable features are concerned, three habitats are involved, Crops (woody), Crops (annual), Annuals (fallow) and Forest. These would link directly to the typologies developed in SAFE for silvo-arable systems. Rules are provided for determining all patches over 400m2 and at least 40m long, in order to include both linear and aerial features, with records being made at a 1km square, as described above.
Distribution
The results described above indicate that there is little point in further sampling in Atlantic Europe because of the scarcity of silvo-arable at the current time. Rather, efforts should be concentrated on southern Europe in order to obtain better estimates of the resource in this part of Europe. Screening exercises using aerial photographs in conjunction with the Environmental Stratification System could be used to target any field surveys in the most efficient way. Such procedures have already been developed to target key habitats in GB. This system would enable silvo-arable to be fitted into other land uses as has been tried and tested over the last 25 years in GB.
Ecological Characteristics
Some general characteristics of the likely ecological features of silvo-arable systems have been described, but more detailed studies are required to quantify these. There is likely to be major differences between new sites in monocultures compared with those within an existing matrix or within a landscape with other patches of woodland and scrub already present. Such a study would be very interesting in landscape ecological terms and could provide additional benefits for silvo-arable schemes.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 120
Existing sites within the SAFE network could be used but monitoring new sites from their inception would also likely to be of considerable value.
CONCLUSIONS
The sample survey already carried out for Atlantic Europe shows that silvo-arable systems are very rare. In contrast, in southern Europe, existing knowledge indicates that sivo-arable systems are widespread, although these have to be quantified and separated from silvo-pastoral systems as these are confused in many current data sets. Existing methodology could be used to fill this gap but significant financial support would be required - although if student labour was used this would be achievable.
Finally, there are likely to be significant ecological benefits associated with new silvo- arable sites. especially in monotonous landscapes dominated by arable crops.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 121
ANNEX 6. Silvoarable agriculture in Europe – past, present and future
M.P. Eichhorn1,*, P. Paris2, F. Herzog3, L.D. Incoll1, F. Liagre4, K. Mantzanas5, M. Mayus4, G. Moreno Marcos6 C. Dupraz4 and D.J. Pilbeam1
1School of Biology, University of Leeds, Leeds LS2 9JT, UK; 2 Istituto per l’Agroselvicoltura Villa Paolina, Via G. Marconi 2, 05010 Porano (TR), Italy; 3 Eidgenössische Forschungsanstalt für Agrarökologie und Landbau (FAL), Reckenholzstr, 191CH-8046 Zürich, Switzerland; 4 INRA Montpellier, UMR Systèmes de Culture Méditerranéens et Tropicaux, 2 Place Viala, 34060 MONTPELLIER Cedex, France; 5 Laboratory of Rangeland Ecology, Aristotle University, 54006 Thessaloniki, Greece; 6 Centro Universitario Plasencia, Forestry School , Avd. Virgen del Puerto 2, 10600 Plasencia – Cáceres, Spain
*Author for correspondence: e-mail: [email protected]
Key words: dehesa, pré-vergers, Streuobst, Hauberg Waldsystem, piantata, orchards
ABSTRACT
Combinations of trees and crops have formed key elements of the landscape of Europe throughout historical times, and many such systems continue to operate in the present day. In many cases they represent traditional systems in decline, and a number of formerly widespread practices have already become extinct or exist only in a threatened state. The causes for this include both practical and economic considerations. The agricultural subsidy regime within the European Union is presently unfavourable with regard to silvoarable practices, and this has been a major factor in their recent decline.
The silvoarable systems of Europe can be split into two zones – northern Europe and the Mediterranean. The latter contains not only a greater area of silvoarable cultivation, but also a greater diversity due to the broader range of commercial tree and crop species available. In general the systems of northern Europe are limited by light, whilst those in the Mediterranean are limited by water availability.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 122
Innovative new combinations have been developed, some of which have the potential for future expansion. An appreciation of the legacy of previous systems of agroforestry is necessary when developing novel approaches or seeking solutions to present problems.
Mixed systems of agriculture present an opportunity for the future of European rural development, and have the potential to contribute towards the enhancement of biodiversity and increased sustainability of agriculture, whilst also preserving landscapes that are both culturally important and aesthetically pleasing. The adoption of a consistent definition of silvoarable agriculture within Europe is recommended.
INTRODUCTION
Silvoarable agroforestry consists of widely spaced trees inter-sown with annual crops. Such systems have traditionally formed key elements of the European landscape, and have the potential to make a positive contribution towards sustainable agriculture in Europe in the future.
Trees have traditionally served three purposes in the agrarian economy – the production of fruits, fodder and wood (for fuel, litter or timber). In addition, they have amenity value, providing shade and shelter to labourers and livestock, and they combat erosion by wind and water. When grown in combination with crops, trees are known to compete for key resources, and hence the modern convention is to separate forestry and agriculture into discrete activities. To focus upon the deleterious effects of trees upon associated crops is however overly simplistic and ignores a range of effects that are both positive and negative in their influence on arable productivity (Jose et al., 2004).
Trees compete with surrounding crops for soil water, which can inhibit arable production in dry climates. Despite this, trees also intercept driving rain and aid the condensation of water droplets from fog and dew (Grove & Rackham,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 123
2001). By acting as windbreaks, they slow air movement and can reduce evaporative stress on crops (Hawke & Wedderburn, 1994; Jose et al., 2004). The deeper rooting systems of trees are thought to impede drainage of rainfall from thin soils, and via the process of hydraulic lift they may draw water from lower soil levels and release it into the upper soil where it benefits shallow- rooted plants (Burgess et al., 1998; Dawson, 1993; Jose et al., 2004; van Noordwijk et al., 1996). Similarly, although the shade cast by trees may limit the growth of crops, there is a benefit to reduced irradiance (and hence transpiration) in arid areas, especially when growing sensitive vegetable sub- crops, and in some cases there may be a moderate yield benefit to shading (Lin et al., 1999). In colder climes the canopy insulates against ground frosts.
Although competition for nutrients may occur, the deeper rooting systems of trees also bring up nutrients from lower soil layers, and reduce the leaching of topsoil. These nutrients are then recycled via leaf litter and root turnover and increase the overall resource-use efficiency of the system (Jose et al., 2004; van Noordwijk et al., 1996). Litter can itself act as a buffer against wind and water erosion and as such increases the sustainability of agriculture by protecting topsoil when crops are absent. Trees may also attract sheltering livestock, which are therefore more likely to deposit manure beneath them (Grove & Rackham, 2001). Scattered trees in croplands and pastures are likely to improve soil structural characteristics beneath the tree canopy.
In order to minimise the potential for negative interactions between trees and crops, it is necessary to carefully select combinations of trees and associated arable crops that have positive interactions (facilitation). The most efficient and sustainable systems are those which are able optimise the use of spatial, temporal and physical resources by avoiding competition between components (Jose et al., 2004). Research has indicated that mixtures of crops can in some circumstances be more productive than monocultures, especially if the trees can
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 124
obtain resources that would otherwise be unavailable to the crops (Cannell et al., 1996), and may reduce the need for agrochemicals (Vandermeer, 1989). By developing an understanding of enduring traditional systems and practices we may gain insight into the potential future applications of silvoarable techniques and their advantages.
Historical perspective
When the original forests of Europe were cleared, trees with high value were retained in the landscape. These included various fruit trees in the Rosaceae, oaks (Quercus spp.) and beech (Fagus spp.) for their production of acorns and mast as animal forage, and ash (Fraxinus spp.) from which lopped branches were used as fodder (Dupraz & Newman, 1997). These formed elements of early agroforestry systems and were continually replaced throughout history as they did not obstruct manual cultivation techniques.
The earliest evidence for planned agroforestry in Europe dates back as far as the Copper Age (c. 2,500 BC). Stevenson & Harrison (1992) identified a change in the composition of pollen cores collected from south-western Spain, with mixed oak and pine forests being replaced by scattered oaks and herbaceous vegetation. A large proportion of the identifiable pollen was from weeds of cultivation. They define this shift as the beginnings of the dehesa, a land-use system characterised by intermittent cultivation, grazing and burning. Some have argued that in fact the transition occurred during a period of climatic change when the region was becoming more arid (Grove & Rackham, 2001). A similar and concurrent change in Italy has been identified as a possible shift towards wood-pasture land usage (Potter, 1979).
The earliest stages of agriculture involved systems of shifting cultivation, with intercalated agricultural and forestry land use. As civilisation progressed towards more stable patterns of agriculture woodland grazing and silvopastoral
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 125
systems were abundant, and there was a continuous transfer of fertility from woods to cultivated lands via manure. The maintenance of soil fertility was based upon a strict connectivity between agriculture, husbandry and forestry. Eckert (1995) estimated that in the Neidlingen valley (Baden-Wurttemberg, Germany) up until 1500, 75% of the nitrogen and 90% of the phosphorus required for arable fertilisation came from woodlands in the form of fodder, litter or wood for domestic fires. This input was vital to maintaining the sustainability of agriculture.
A further reason for the maintenance of trees in the landscape was the production of fruit for human consumption. Fruit was an essential part of the diet, being a crucial source of many vitamins (Herzog, 1998a), and culturally important for the production of alcohol. Many economically valuable tree species are dual-purpose, producing an annual fruit crop and an ultimate timber end product (e.g. cherry, walnut), often in addition to litter and fuel wood.
In the Middle Ages, with the introduction of sustainable crop rotations, soil fertility became less dependent upon woods and trees. This shift in the role of trees was accelerated during the 19th century by the introduction of chemical fertilisers and the mechanisation of agriculture. Nowadays forestry, agriculture and husbandry are discrete activities with few chemical and energetic relationships. Nevertheless, many historical agroforestry practices have been retained, and continue to be maintained in a traditional fashion.
At present, information on the status of agroforestry in general, and of silvoarable systems in particular, is quite poor in Europe. This is due to a bias towards single crop systems in both research activities and institutional interests. Throughout the last century there has been a marked decline in the use of silvoarable agroforestry systems across Western Europe. In many countries this decline can be attributed to the same basic causes:
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 126
• Scattered trees in arable landscapes impeded mechanised agriculture and have been deliberately removed or were damaged by machinery.
• The post-war drive for increased yields led to a focus on maximising productivity through mono-cultural systems.
• A reduction of manpower in agriculture limited the commercial viability of labour-intensive systems, e.g. full stature fruit tree orchards.
• Consolidation of fragmented land holdings into larger single farms and fields removed boundary trees and reduced the scope for landscape diversity
• The subsidy regime of the Common Agricultural Policy (CAP) indirectly led to a reduction in crop associations by favouring single crop systems.
• Wooded areas were for many years ineligible for direct subsidy payments, and in many cases trees were grubbed out to increase subsidy income.
• A stricter quality norm applied to dessert fruit (EEC regulation 1641/71) tended to standardise their production in intensively managed orchards.
Traditional silvoarable systems have gradually been abandoned in marginal agricultural areas, while on more productive soils they have been replaced by crop monocultures. There is no motivation provided to farmers to maintain silvoarable systems, and they have often been perceived as an obstacle to modernisation via mechanisation.
DATA COLLECTION
This paper collates contributions from seven European countries (France, Greece, Germany, Italy, Spain, the Netherlands and the United Kingdom).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 127
These are the result of information collection from bibliographical and other sources (e.g. personal contacts and internet sites). In addition, inventories of silvoarable systems from local to national levels were conducted within the Silvoarable Agroforestry For Europe (SAFE) project during the first year of its activity (2001-02). In Germany, Greece and the Netherlands these represented the first attempts to quantify the existing silvoarable systems in that country and are therefore limited to what could be achieved within a single year.
Member countries faced consistent difficulties in their attempts to document both the types and extent of silvoarable practices within their borders. These included a lack of official statistical data, largely due to a failure to distinguish between silvoarable agroforestry and conventional forestry plantations within land use surveys. The existing literature on silvoarable agroforestry is largely confined to local journals and magazines, and is inaccessible to conventional literature searches. There were also logistical difficulties in locating and contacting individual farmers to verify reports.
SYSTEMS
The combinations of trees and crops employed by European farmers are immensely varied. In this review we will concern ourselves only with those systems that are currently extensive, have been in the recent past (i.e. the last century), or have clear potential for commercial expansion in the future. The major extant systems within Europe are summarised in Table 1. Mixed systems of agriculture remain common in garden plots, and very small fields with trees on the boundaries are effectively silvoarable. These, however, are generally too small and inconsistent in their composition to be considered coherent systems in their own right. They represent the needs of individual farmers to maximise returns from a small area, and tend to be composed of fruit trees and vegetables for domestic consumption rather than economic returns.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 128
The climate of northern Europe imposes greater constraints on silvoarable productivity than in tropical or Mediterranean systems. Lower photon flux densities (PFD) at higher latitudes make it increasingly difficult to support an economically viable ground crop beneath tree canopy cover. During the later stages of tree development, canopy closure prohibits the growth of many crops unless the tree rows are widely spaced, which in turn increases the pruning requirements of the trees. In linear systems, once the tree height exceeds the width of the row the system is often no longer suitable for alley cropping.
The incentives for silvoarable systems in northern Europe therefore need to be clearly defined, as it is likely that the total economic output of a mixed system will be less than might be achieved with a single crop. They are invariably planted, and have not arisen directly from semi-natural vegetation (as with some Mediterranean systems). In this sense they differ from landscape-level systems, where the economic assessment is based upon the advantages of maintaining pre-existing trees. In almost all cases the principle advantages of silvoarable systems are in yield diversification and the production of a short-term return on land while planted trees are still small. Their potential roles in agricultural sustainability and maintenance of biodiversity are an area of active research (Gordon & Newman, 1997).
In Mediterranean Europe the silvoarable systems present in northern Europe are supplemented by a number of additional types. This is due to the greater diversity of economically valuable trees in the region, along with a natural tendency towards savannah-type vegetation in arid areas, since the relative size of the root system required to support the above-ground parts of the tree is greater, causing the trees to be naturally dispersed. In contrast to northern Europe the limiting factor in most systems is water rather than light. The main additional tree species present in Mediterranean regions are olive (Olea
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 129
europaea var. europaea L.), evergreen oaks (Quercus spp.), carob (Ceratonia siliqua L.) and a broader range of fruit trees.
Over large portions of Mediterranean Europe, silvoarable agroforestry remains a common system of land use. In contrast to northern Europe this almost exclusively results from the maintenance of traditional systems that have persisted for thousands of years rather than the development of novel or modernised systems. In many Mediterranean areas agricultural land is still divided into fragmented smallholdings, unlike northern Europe, where for the most part these have been consolidated into larger, more efficient units of land. Small fields result in more boundaries, and therefore a greater number of trees remain in the landscape. These are seldom of a single dominant species, and are often spread throughout fields with no planned pattern or density.
In other areas, particularly the olive groves of central Italy, or the dehesas of SW Spain and Portugal, it is the trees themselves that define the landscape, and they form a consistent component within a variety of arable or pastoral land uses. A great diversity of tree/crop associations therefore exists, and it is likely that all possible permutations occur, albeit only on small scales, where they may be planted according to the specific needs of local farmers. The various systems are discussed here in terms of the key tree products, although often they are mixed in function.
Olive tree associations
Olives form a continuous landscape element in many parts of southern Europe, with diverse crops sown between the stems. This practice is thought to date back to pre-Roman times, when wheat was cultivated between rows of olive trees on alternate years, since this was known to increase their yield in the following year and thereby splitting the grove increased overall productivity (Lelle & Gold, 1994). In Greece olives cover an estimated 650,000 ha in total
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 130
with intercropping of cereals, vegetables and fodder crops. Olive trees are typically planted in rows, although they may also be irregularly scattered when groves have been thinned. Oaks, carob, walnut (Juglans regia L.), almond (Prunus dulcis (Miller) D. A. Webb) and other fruit trees often form a minor mixed component.
In the central Italian regions of Umbria and Lazio the silvoarable system formed by olive trees is the most abundant in the country, covering some 79,000 ha. As in Greece, they are commonly intercropped with cereals or fodder legumes. The olives form a consistent component of the landscape in contiguous silvopastoral and horticultural fields, either as scattered trees or in rows with 5-10 m between stems. A similar landscape survives in some parts of Spain. Grape vines are sometimes grown along the rows of trees as part of a formerly extensive system known as piantata (see below).
Fruit tree associations
Silvoarable systems based upon fruit production have covered extensive tracts in central Europe as recently as the last century. The pré-vergers are areas of low-density fruit tree plantations which double as grazing land, and are abundant in northeast France. The fruit trees are often dual-purpose and produce a timber end product, especially walnut, pear (Pyrus communis L.) and apple (Malus domestica Borkh.). Some of these plantations are intercropped during the early years of tree growth, especially walnut plantations in the regions of Dauphiné and Périgord, covering some 15,000 ha. Around 4,000 ha may be silvoarable at any given time (Liagre, 1993a). Typically crops are grown between 5 and 15 years into an approximately 30-year cycle (Liagre, 1993b; Mary et al., 1998) with a variety of crops including maize and other cereals, sorghum, soybean, oil-seed rape, sunflower, tobacco, alfalfa, lavender and bush fruits (Ribes spp.). In Dauphiné around 20% of walnut orchards are intercropped (80% of those below 10 years of age) (Dupraz & Newman, 1997).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 131
Walnuts are commonly found irregularly scattered amongst cereal crops throughout France, a practice dating back at least 300 years in Burgundy (Dupraz & Newman, 1997).
A comparable but less regimented style of silvoarable orchard is the central European system of Streuobst, defined as ‘tall trees of different types and varieties of fruit, belonging to different age groups, which are dispersed on croplands, meadows and pastures in a rather irregular association’ (translated from Lucke et al., 1992). Streuobst was formerly a widespread land use system, and was typically practised in areas with highly productive arable land. The system was sub-divided into silvoarable (Streuobstäcker) and silvopastoral (Streuobstwiesen) forms. Streuobstäcker generally consisted of two rows of fruit trees, intercropped close to the tree trunks, with relatively low branches to facilitate fruit harvest. The most common fruit trees were apple, pear, plum (Prunus domestica L.) and mazard cherry (P. avium L.) planted at a density of 20-100 stems ha-1 (Herzog, 1998b).
Initially, during the 16th and 17th centuries, German national and regional policies encouraged the planting of fruit trees and creation of Streuobst, which was maintained throughout the following centuries. A fruit tree survey of 1938 recorded approximately 800,000 ha of active Streuobst (Herzog, 1998a). This area declined dramatically during the second half of the century (Rösler, 1996) due to replacement by intensively-managed orchards with narrow grassed alleys between rows of dwarf fruit trees (Figure 1). These permit greater mechanisation but exclude intercrops. Streuobst eradication programmes were originally subsidised under the CAP in favour of more standardised means of production. Those that remain operate at a loss due to high manual labour costs.
In more recent years there have been subsidised schemes for protection on local to national scales. Nevertheless, the majority of sites maintain Streuobstwiesen
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 132
rather than the arable Streuobstäcker, which is presently limited to a few very small fields (< 0.5 ha). Similar orchards formerly existed in the Netherlands, albeit on a smaller scale than the German Streuobst, but the remaining fragments are maintained solely as a cultural heritage.
One of the reasons for the development of the pré-vergers in the 16th century, and the former abundance of Streuobst in Central Europe during the 18th century, was that long term fruit production could be combined with annual income from the arable crops in the system (Herzog & Oetmann, 2000). Reasons for continued preservation may include landscape restoration, combating erosion and nature conservation or recreation, but in such cases they are not managed for economic returns. A recent fashion for hobby fruit production has led to small plots being maintained by individuals as a leisure pursuit.
A similar orchard system was reinvented and subsequently lost in the United Kingdom. During the early 20th century it was common practice to grow crops between saplings of fruit trees in orchards, particularly apple (not for production of alcohol) and cherry, especially in the Kent region (Hoare, 1928). Soft fruits (e.g. blackcurrant, gooseberry, raspberry, strawberry) or vegetables (e.g. asparagus) were the most common intercrops (Roach, 1985). Although at their peak such orchards covered c.110, 000 ha in 1951-55, rotations were lengthy (50-100 years) and the area intercropped at any one time represented only a small proportion of this (Roach, 1985).
There is a necessary distinction to be made between the comparable Streuobst and pré-verger systems of northern Europe and those that are present further south on similar sites but which also incorporate grape vines. They differ from the majority of silvoarable systems in that the trees are no longer the focal element and chief economic resource. Mixed vineyards have a venerable and
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 133
well-documented history in the European landscape, with trees incorporated as living mechanical supports for the vines and to increase the economic return from the land through diversification. Meiggs (1982) collates historical records from the classical period of a wide variety of trees intercalated with grape rows, and the diverse functions they served. He found no evidence of the planting of trees for timber alone, and assumes that timber was sourced from trees grown between and amongst other crops. Ancient authorities on agriculture such as Cato (234-149 BC) did not recommend using any land specifically for the growth of trees, but advocated their inclusion within grape systems.
In flat fertile areas of Italy (e.g. the Po Valley), poplars (Populus spp.) and fruit trees (Rosaceae spp.) were used as support for the vines, organised into rows and intercropped (piantate), a system dating back to the Etruscan period (699- 464 BC). In the most productive regions, such as Campania, the rows of vines could reach as high as 10 m. In hilly regions the vines tended to be supported by smaller stature trees including ash (Fraxinus spp.), maple (Acer spp.) and mulberries (Morus spp.). Special consideration was given to those crops least likely to compete strongly for water during the dry season (May-October) such as wheat or fodder legumes including clovers (Trifolium spp.) and vetches (Vicia spp.). Fragments of the system remain in areas of Sicily.
In southern France the formerly prevalent system referred to as Joualle was composed of rows of grapevine with peach, walnut and olive trees inserted. In some cases the trees were used as support for the vine (hautain). In order to maximise returns from the land, the gaps between rows would often be sown with annual crops (usually cereals). This system has greatly declined due to mechanisation, which makes the manual harvesting of such narrow crop rows uneconomical, and the French national agricultural policy of separating agriculture from forestry. A similar system continues to operate in Greece, with olive, walnut, various oak species and wild pear incorporated amongst the vines,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 134
and in Sicily intercropped vineyards still cover some 153,000 ha. In Spain the intercropping of vineyards continues only in restricted areas. In general, specialised intensive vineyards have replaced the traditional mixed form, and the system persists only as scattered fragments.
In northern Spain intercropped small orchards (less than 0.5 ha) combining fruit trees with vegetables remain abundant, but do not cover a substantial land area, estimated at 6,200 ha in total (INE, 2002). Throughout the Mediterranean region small orchards of walnut, almond, peach (Prunus persica (L.) Batsch), apricot (Prunus armeniaca L.) and olive are intercropped with vegetables and cereals.
Small-scale silvoarable plots still exist in eastern Germany (e.g. Magdeburger Börde), usually for household consumption rather than as commercial systems. Commonly cherry trees (Prunus avium L.) are undercropped with turnip, but also with alfalfa, potatoes, oats and formerly asparagus. Its continued survival in the former DDR can be attributed to the need to maximise returns from the small amount of private land assigned to each farmer following expropriation, and the high fertility of land in the region.
In the Languedoc-Roussillon province of southern France a modern intensive agroforestry system combines peach trees with intercropped vegetables. The system is highly profitable and efficient in light use as the vegetables are able to grow through winter and spring before the trees come into leaf, although it requires irrigation.
The greatest expanse and diversity of fruit-producing silvoarable systems is found in Greece. There is substantial regional variation in the dominant fruit tree species, although in all areas a mixture occurs. In northern and central areas pear (Pyrus spp.) dominates, with cereals, vegetables or tobacco cropped between them. Walnut is preferred in montane areas, grown also for timber, while mulberry (Morus spp.) is favoured in Thrace. Silvoarable combinations
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 135
with figs (Ficus spp., also grown for fodder) occur in Crete and the Aegean Islands. Cereals (wheat and barley) are the most common intercrops in all systems.
Walnut is a major component of silvoarable systems in Italy, where again it is used as a dual-purpose fruit and timber tree in montane regions. In Campania it is intercropped with vegetables, and often mixed with hazel (Corylus avellana L.) grown for wood and nuts and as a trainer tree to improve the form of the walnut trunks. On the fertile volcanic soils of the region tree growth is fast, and a low plantation density (50 stems ha-1) permits crop cultivation for a number of years. The system is very much in decline, due to competition from foreign imports (especially from California), the greater profitability of vegetables when planted alone, and the high value of land for development in a densely populated region.
Timber tree associations
The increased demand for high-quality timber in Europe, coupled with the reduced availability of tropical hardwoods, has led to the development and expansion of a number of silvoarable systems designed specifically for the production of high-grade timber. A number of experimental approaches have been adopted in different countries, with great potential for increased application. In contrast to fruit trees, it is thought that there is no critical stage for diameter growth of timber trees, and therefore they may be more easily incorporated into silvoarable systems without deleterious competition from crops (Dupraz, 1994).
Silvoarable systems combining hybrid poplars grown for timber with cereal crops were pioneered in northern Italy and have since been adopted throughout northern Europe. The practice continues in the Po Valley region on flat fertile soils. Maize, soybean and cereals are grown between tree rows during the first
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 136
two years of a ten-year cultivation cycle, although with intensification on fertile soils this may be reduced to as little as seven years. Lapietra et al. (1991) estimated that 4.1% of the total poplar area is intercropped at any given time. The system is presently in decline in Italy due to the European grant policy of subsidising tree planting on arable land, but which did not permit concurrent intercropping.
In France the practice became fashionable during the 18th century, and continues to be practised in well-irrigated alluvial regions throughout the country, covering some c. 6,000 ha. Typically cereals are intercropped for the first three years. In northern Greece cereals, vegetables or fodder crops may be grown among the trees. High levels of irrigation are usually required, which precludes the use of this strategy in more arid regions. Fertilisation is also necessary, along with intensive weed control and pruning of the trees. Similar systems in France are not managed with the same intensity.
In the United Kingdom during the 1950s Bryant & May Forestry Ltd. managed large-scale plantations of hybrid poplar in southern England for the manufacture of matches (Beaton, 1987; Dupraz & Newman, 1997). Alleys were cropped with cereals for eight years, with an under-sown grass/clover mixture in the final year. The plantations were then used for grazing cattle until year 20, when canopy closure prevented the formation of pasture. The poplars were harvested at 25 years old. The availability of cheap Scandinavian lumber and the crisis in cereal prices caused these plantations to be abandoned in the 1970s. In recent years similar trial plots combining hybrid poplars with various crops have been established in the Netherlands (Edelenbosch & Dik, 1995) and the United Kingdom (Beaton et al., 1999; Incoll et al., 2002).
Other linear combinations of timber trees and crops exist with a limited distribution, but none have been so widely adopted. Silvoarable methods have
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 137
been incorporated into systems that were formerly the preserve of pure forestry. The wider spacing between rows of trees in silvoarable systems increases rates of growth and can, with an appropriate pruning regime, increase the value of the timber through enhanced form and increased timber length.
In France, a farm in Aude combines the leguminous timber tree black locust (Robinia pseudacacia L.) with cereals on 20 ha of land, with the aim of maintaining soil fertility while reducing the need for fertilisation. In the UK a commercial silvoarable system for the production of furniture timber includes five tree species (Fraxinus excelsior L., Juglans nigra L., Prunus avium, Quercus rober L. and Acer pseudoplatanus L.) with alley cropping of cereals or pulses. The tree rows contain additional specimen trees for early transplanting to urban parks, gardens and streets, negating the need for row thinning and improving the overall efficiency of the system.
In the Netherlands innovative combinations of trees grown for high-grade timber and ground-level flower production have been attempted. The Stichting Robinia Foundation in Lelystad, an organisation promoting sustainable timber production, runs a small demonstration plot (0.5 ha) with several tree species (Catalpa bignonioides Walt., Alnus glutinosa (L.) Gaertner, Prunus avium and Gleditsia triacanthos L.) intercropped with hyacinth (Liliaceae) for flowers and bulbs. In Fryslân the Boslandbouw Foundation has experimented with cedar (Cedrus spp.) intercropped with flowering quince (Chaenomeles spp.) for flower and fruit production, and a number of further plots. The potential of these systems for wider commercial application has yet to be established, although they have great aesthetic appeal.
Oak tree associations
In certain regions of Europe the landscape is defined by the presence of scattered oaks, forming contiguous arable and pastoral associations. This is
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 138
most characteristic of the dehesas of SW Spain and Portugal, a system of land use that may have been practised for as long as 4,500 years (Stevenson & Harrison, 1992). The dehesa is the dominant agroforestry system in Spain, and probably the largest such system in Europe. Estimates of present extent vary (Carruthers, 1993), largely due to inconsistencies in its definition, with operational dehesas in the strictest sense thought to cover more than two million hectares in SW Spain and 0.2 million hectares in Portugal, where it is known as montado (Joffre et al., 1988)(Figure 2). Similar systems occur in northern Greece, and cover much of Crete and Sardinia, but have almost entirely disappeared on the Italian mainland (Grove & Rackham, 2001).
Although some linear dehesas exist, in the majority of cases the trees are scattered at relatively low densities (10-40 stems ha-1). The shapes of the trees confirm that they have developed in an open environment, suggesting that the savannah is at least partly natural, and that is has not developed by subtraction of trees from a pre-existing forest, although mixed oak and pine forests are thought to have once dominated the region (Stevenson & Harrison, 1992). The constituent trees have been actively selected for sweet acorn production, and consist mostly of evergreen oak (Quercus ilex L.), but also cork oak (Q. suber L.) and Pyrenean oak (Q. pyrenaica Willd.).
The ground beneath the trees is sown with cereals, fodder crops or sunflower, or is used as wood-pasture. The lengths of the rotations vary from 2 to 12 years depending on the maturity of the system. The system is therefore sometimes referred to as being agro-silvo-pastoral, since it combines a range of different practices, with arable cultivation shifting somewhat irregularly over successive years. Pigs, sheep, grain, acorns and fuel wood are the main products (Grove & Rackham, 2001) although cork can be a valuable commodity where suitable trees are found.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 139
Within the dehesas only a small proportion is cropped in any given year, between 10.3% (MAPA, 1985) and 16% (Escribano & Pulido, 1998). Moreover, crops are only harvested from 30% of the cultivated land, with the remainder being grazed directly by cattle or used as fodder, supplemented by acorns from the trees (Escribano & Pulido, 1998). In most cases the principle aim of cropping is not to produce a commercial product but to control encroachment of shrubs and to improve the soil and pasture.
The dehesa system exists primarily for the production of the fine hams that are a speciality of the region, since pigs are known to fatten better in savannah than woodland. This is because savannah oaks produce more acorns than woodland trees, combined with the availability of understorey grasses and herbs for grazing, which are beneficial for their protein content. The economic importance of acorns in dehesa is far greater than the relative value of pannage in northern European woodlands, and in 1957 it was estimated that acorns comprised one sixth of the value of all forest products in Spain (Balabanian, 1984; Parsons, 1962). Acorns were formerly a common human food (belotas), and not only in times of famine, although acorn-bread is now seldom baked (Grove & Rackham, 2001).
The minimum size of an operational dehesa estate is thought to be around 400 ha (Grove & Rackham, 2001). Management is seasonally labour-intensive, with branch lopping and acorn gathering being physically demanding tasks. Increased labour costs in Europe therefore threaten this way of life. Cork cutting is lucrative employment, but highly seasonal, and labourers are assisted by European grants for the remainder of the year.
Lopped branches from the trees are used for firewood or charcoal production. The trees tend to be lopped in a distinctive pattern, particular to different
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 140
regions, which determines their crown shape. Although valuable trees in many other regards, savannah oaks tend to produce low quality timber.
A subtle change in dehesa management practices has also occurred over the last 30 years (Romero Candau, 1981). The traditional rotation was five years of cereal cropping followed by five to ten years of pasture management. Ploughing prevented the establishment of inedible shrubs (such as Cistus spp., Erica spp., Arbutus unendo L.) and maintained high quality pasture for succeeding years. The crisis in cereal production in the 1970s led to this practice being abandoned for some 20 years, following which severe bush encroachment has reduced the value of the pasture (Dupraz & Newman, 1997) and created a substantial uncontrolled fire hazard. In other areas, localised overgrazing has led to the loss of the most favoured shrub species such as Medicago arborea L. and Colutea arborescens L.
Dehesas are in a less threatened state than many traditional land use systems, due in part to protective legislation. Approximately a million hectares were lost between 1950 and 1980 when EU subsidy made cereal cropping more profitable. Only water shortages restricted the ability of irrigation schemes to convert more dehesa into purely arable cropland. Since 1984 state law has forbidden substitution of oak woodland in Extremadura. Nevertheless, there is some evidence of a more recent decline from 2.3m ha to 1.7m ha between 1985 and 1998 (Miguel et al., 2000), although the reliability of these data is compromised by the lack of a firm definition of dehesas and an absence of systematically collected information. A more subtle change in structure may be occurring through alterations in tree density; a decline of 23% between 1951 and 1981 has been documented (Miguel et al., 2000). Almost no oaks have been planted for the last century, which has led to concerns over the long-term regeneration of the system (Grove & Rackham, 2001).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 141
Figure 3 illustrates the decline of intercropped open woodland in Spain, which largely refers to traditional dehesa systems (in approximately 90% of cases). It must be noted that open woodland is not in decline in Spain, with new woods developing on abandoned agricultural land in marginal areas. Instead it is the actual practice of arable cultivation that has been greatly reduced.
Silvoarable oak systems are widespread in Greece, with cereals grown commercially between scattered trees at densities ranging from 10 to 100 stems ha-1. A variety of oak species (Quercus pubescens Willd., Q. petraea (Mattuschka) Liebl., Q. cerris L. and Q. trojana Webb in Loudon) are found in northern and central areas, associated with cereals, tobacco, sunflower and fodder crops. In southern and western areas Valonia oak (Q. macrolepis Kotschy) is most abundant, usually with intercropped cereals.
Similar associations persist in marginal areas of central and southern Italy and Sardinia, with scattered oaks at densities between 7-250 stems ha-1. The system is referred to as seminativo pascolo arborato. Various oak species are involved, most commonly Q. pubescens and Q. cerris, although cork oak occurs in Sardinia, where the system is similar to dehesa. Wild pear sometimes forms a minor element. The trees were formerly used for fuel wood production, but increasingly they are not managed and are retained purely as landscape elements, or to reduce erosion. A rotation of wheat and clover is grown beneath the trees, with oats being a less common intercrop.
In northern Europe the Hauberg Waldsystem of west-central Germany was an ancient agroforestry system, practised for around 2,500 years. The system combined the growth of trees for fuel wood with crop production and grazing on a long rotation. The dominant tree species were oak and birch (Betula spp.). According to historical records, trees in the Hauberg system were cut for firewood and charcoal every 16 to 20 years, with the stumps left in the ground
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 142
to allow re-growth. After tree felling any remaining ground vegetation was removed or burnt, and for the following years a crop was sown between the stumps. Typically this was either buckwheat in June or rye in autumn as an over-wintering crop. After about nine years of silvoarable farming, when the trees had reached a certain size and could tolerate the presence of livestock, the area was grazed with sheep and cattle. This silvopastoral land use continued until the next felling.
The Hauberg was a collective system of land management, farmed by the entire village community. At the end of the 19th century the Hauberg co-operations began conversion of low forest into high forest as timber wood became more valuable than firewood. This transformation occurred at a slower rate than may have been expected, and in 2000 between 6,000 and 7,000 ha of traditional low forest still survived. The reasons for its maintenance include the need for protection against erosion and drainage control, preservation of biodiversity, and the maintenance of a landscape recognised as being historically important and culturally unique. It is however no longer a commercially viable land-use option since there is a limited market for the wood products, and intercropping is excessively labour intensive.
Fodder tree associations
In several of the examples given above, most notably the dehesa system, the trees provide an important source of fodder leaves. There are also many cases, both historical and contemporary, of trees being managed exclusively for their leaves as a source of valuable nutriment for livestock during seasons when ground vegetation for grazing is sparse (Dupraz & Newman, 1997; Lachaux et al., 1988). This applies especially to the more arid areas of the Mediterranean.
In Greece, where the growing season for grasses is short, leaves are shredded from deciduous oaks and dried to feed sheep throughout the rest of the year. In
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 143
Crete a distinctive form of pollarding creates ‘goat pollards’, low platforms with dense shrubby regrowth that goats clamber amongst and forage upon. These trees only reach full stature if grazing is withheld for several years, and in practice they remain at ground level in perpetuity.
Cato (234-149 BC) was an advocate of the maintenance of trees as a source of fodder alone. Leaves of elm (Ulmus spp.) were considered best, followed by poplar: ‘If you have poplar leaves mix them with the elm to make the latter last longer; and failing elms, feed oak and fig leaves’. All four continue to be used to the present day. Columella (1st century AD) regarded poplar, elm and ash as providing the best fodder (Meiggs, 1982). In ancient times these trees were most likely planted within vineyards in mixed systems as described above.
Crete and the Aegean Islands contain silvoarable combinations with figs, with carob also favoured in Crete, where the pods provide an important source of stored fodder. In Sicily, carob is grown over some 20,000 ha and intercropped with cereals and fodder legumes. The pods are also utilised as raw materials in the food processing industry.
Non-native trees have more recently been considered as fodder sources in the Mediterranean. The use of honey locust (Gleditsia triacanthos), a leguminous tree native to America, has been promoted for many years (Wilson, 1993). A number of other species have been considered, including Amorpha fruticosa, Robinia pseudacacia, Colutea arborescens, Corinilla emerus, Medicago arborea and Morus latifolia (Dupraz & Newman, 1997). Species characteristic of dry regions, such as Acacia spp. and Atriplex spp., have also been trialled in various locations (Correal, 1987; Dupraz & Newman, 1997).
Present status of agroforestry in Europe
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 144
There are a number of obstacles faced by those farmers and institutions who are current or potential practitioners of silvoarable agroforestry, and who may benefit from increased knowledge and awareness of its potential applications. There is a lack of received knowledge on former agroforestry systems which have now largely disappeared, and a focus on single crop systems within research institutions reduces the advice and training available to farmers wishing to manage trees in an agricultural environment. In terms of agricultural subsidies, EEC Regulation 2080/92 provides grant funding for tree planting on arable land, but does not permit intercropping. The current political climate is therefore generally unfavourable and mixed agriculture requires more powerful promotion at the regional level.
In this context, a recently proposed European Council Regulation on support for rural development by the European Agricultural Fund for Rural Development (EAFRD) states that "Agri-forestry systems have a high ecological and social value by combining extensive agriculture and forestry systems, aimed at the production of high-quality wood and other forest products. Their establishment should be supported" (EAFRD, 2004). Should this recommendation be approved, the prospects for the adoption of silvoarable methods within Europe would improve greatly.
In three of the northern European countries included in this review (Germany, the Netherlands, and the United Kingdom) silvoarable agroforestry no longer has a significant role in the agrarian economy, and multifunctional land use generally persists only on a small scale. France alone retains silvoarable systems of any notable importance. The Hauberg Waldsystem, formerly practiced extensively in west-central Germany, has almost completely vanished. The intercropped fruit orchards of central Europe, Streuobst and pré-vergers, are much diminished in extent and largely silvopastoral. Other ancient types of
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 145
silvoarable agroforestry are often very poorly documented and little is known about their former content or extent.
Attitudes towards mixed agriculture among governmental bodies have altered in recent years. In France, a census of silvoarable practices, commissioned by the Environment Ministry, was conducted in 2000 by the SOLAGRO Association and INRA, Montpellier (Coulon et al., 2001). An informal network of interested parties has formed to lobby for the reform of agricultural and forestry laws to favour agroforestry systems and has succeeded in changing the application of subsidies. Since 2002, intercrops are eligible for CAP subsidies (DPEI/SPM/C2001-4008, 8th March 2001), agroforestry plantations receive forestry subsidies (DPEI/SDF/C2001-3010, 7th May 2001) and the area planted with trees is eligible for the European PCPR subsidy for lost arable income (DERF/SDF/C2001-3020, 8th August 2001). Agroforestry is therefore currently strongly favoured by the regulations within France.
Current silvoarable practices in France are relatively well-documented and there are movements to preserve existing practices and encourage novel approaches. As an example of a system operated as a going concern, Claude Jollet (Charentes Maritimes) maintains 56 ha of walnut and wild cherry trees (c. 25 years old), intercropped with barley and sunflower. It is estimated that by 2005 there will be more than 80 similar silvoarable projects in France.
In the Netherlands the centralised system of agrarian research and the established network of farmers’ associations have encouraged innovation and enabled effective dissemination of research findings, although at present most systems operate only on very small scales. In the UK a number of trial plots have been set up under the aegis of the Farm Woodland Forum, an organisation actively promoting agroforestry options.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 146
Silvoarable agriculture remains of great importance in many regions of the Mediterranean. The systems are usually relicts of formerly extensive agricultural practices now restricted to more marginal areas, especially olive grove intercropping and dehesas. Despite their former importance, and the threats posed to them by the expansion of modern intensive agriculture, they have been relatively poorly studied, both in terms of their economic importance and their role in preserving local and regional biodiversity.
The greatest diversity of silvoarable systems is found in Greece, where a large variety of combinations of trees and crops exists. These are generally characterised by a small plot area, with a number of different tree species present, often dispersed throughout the field and at the margins but without a fixed pattern or spacing. A range of understorey crops with different management are often planted side by side. This poses an obstacle to the strict categorisation of systems. At present in Greece there is no regional or national policy to improve silvoarable systems and make them economically viable. A particular problem here is that the typically short lengths of land tenancies do not encourage farmers to initiate novel long-term management practices such as agroforestry. The trees belong to the landowner rather than the tenant.
Silvoarable agroforestry remains widespread in Italy, although it suffered a decline in the latter part of the last century. In many cases, silvoarable systems survive in regions where the terrain and climate have impeded intensification (e.g. in Umbria, a relatively dry, hilly region)(Bertolotto et al., 1995). Recent interest in silvoarable techniques has been stimulated by the demand for high- quality local timber for domestic furniture manufacture. Many of the systems described above have been in existence for centuries in Italian agriculture, but have substantially declined.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 147
In Spain, regular national agricultural census data throughout the last 50 years enable a clear picture of extant silvoarable systems and the trends in their distribution to be identified, although the census categories are often broad and do not specify tree or crop species within associations. For example, irregularly structured dehesas are recorded in the same category as linear poplar/maize silvoarable systems. Despite the continued survival of the dehesas, silvoarable systems are a minority land use relative to conventional arable cropland, and are mostly found on more marginal arable soils.
During the latter half of the 20th century the most pronounced was the reduced intercropping of fruit trees, which fell by 97% between 1962 and 1999. During the same period intercropped olive systems were reduced by 94%. This trend has continued still further (Table 2). Despite some inconsistencies in classifications between years, the general pattern of severe decline is evident. Factors specific to Spain include the migration of people from marginal agricultural land (causing silvoarable fields to revert to woodland), the consolidation of fragmented land holdings into larger single farms and irrigation projects that reduce the need for shade trees among crops.
Future prospects for silvoarable agriculture
A modern focus on sustainable agriculture and the conservation of nature and landscapes in Europe has increased the interest in silvoarable systems, and encouraged the establishment of research projects. Multifunctional land use has been identified as a potential means of increasing the biological species richness of farmland through increased habitat diversity as well as protecting against erosion and reducing the need for agrochemical input (Jose et al., 2004; Vandermeer, 1989).
There is a pressing need in Europe for a local source of high quality hardwoods to replace tropical sources (Smith, 1990). Europe is now almost self-sufficient
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 148
in other wood and wood-based products, largely derived from sustainable forests in Scandinavia, but high-grade timber retains a high market value. Deciduous broad-leaved species are preferred to conifers, principally to satisfy market needs, but also for environmental, cultural and aesthetic reasons (Dupraz & Newman, 1997). The economic potential for timber silvoarable systems is difficult to assess due to a lack of research and great variability in the local incentives provided to farmers for tree planting.
Timber trees are thought to have greater potential than fruit trees in silvoarable systems, as there are no constraints posed by fruit harvesting that might limit the choice of intercrops. In addition, fruit trees are sensitive to competition during the earlier stages of growth, whereas timber trees are more resilient and there is no critical period for determining diameter growth rates (Dupraz, 1994; Dupraz & Newman, 1997). Market demands for a standardised form of fruit also favour their production in intensively managed and dedicated orchards.
There has been an increase in recent years in the use of trees purely for fodder production. This has been researched in the Rougier des Camarès area in southern France as a means of combating erosion in fields that were previously sown purely with annual fodder crops (Dupraz & Newman, 1997). Trees may have a valuable role to play in maintaining the integrity of soils and combating erosion in other areas of Europe.
The search for alternative energy sources has led to silvoarable systems being considered as a source of bio-fuels (Hall, 1997; Herzog, 1994), or else they may have a potential role in the reduction of atmospheric CO2 (Herzog, 1994). Poplar short-rotation coppice (SRC) would seem to be the most viable option (Newman et al., 1991), although willow (Salix spp.) is also being investigated. Such ‘carbon-neutral’ fuel sources have been highlighted as potential alternatives to fossil fuels for energy production at local levels (Newman et al.,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 149
1991). An experimental intercropping has been attempted at Long Ashton Research Station, Bristol, UK (Nichols et al., 2000), but there is no historical precedent for combined coppice and arable production.
Although there has been a recent surge in research interest in silvoarable agroforestry, all experimental findings on novel systems are of necessity preliminary, since modern scientific research has yet to cover the lifespan of a cohort of trees in a plot.
It is also necessary to be cautious in claiming environmental benefits for silvoarable systems in general, especially in the context of increased sustainability of agriculture. Considerable experience accumulated in the tropics has shown that the management of intercropped systems is often intensive. The high cost of manual labour in Europe is likely to lead to a greater reliance on agrochemical input, especially when unfavourable combinations of trees and crops are employed. The combined peach and vegetable systems of southern France, which require intensive fertilisation and irrigation, are an illustration of this.
Conclusions
Despite the limitations of this review, it is clear that there are two distinct geographical and climatic zones with respect to European silvoarable agroforestry –northern Europe and the Mediterranean. The latter contains a broader range of systems, reflecting the higher diversity of commercial crops and plant resources. In general, the form and structure of systems in northern Europe are determined by light limitation, whereas in the Mediterranean water is the key resource.
In their review of agroforestry practices in temperate regions around the globe, Newman and Gordon conclude that the most successfully optimised systems are those for which there is a clearly defined market for a tree product (Newman &
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 150
Gordon, 1997). In assessing the prospects for the preservation of traditional silvoarable systems, and the scope for novel and innovative approaches to combinations of trees and crops, we should therefore focus upon the economic value of the trees.
Although extant silvoarable practices in Europe are mostly residual elements of formerly widespread systems, there is still a considerable diversity in existence. The precise quantification of silvoarable systems in Europe is difficult due to lack of documentation. The application of a consistent definition of silvoarable agroforestry in land use surveys and recognition of their unique characteristics would go some way towards an accurate appraisal of their present extent and importance in the landscape of Europe. Their productive role in the European countries studied is not yet fully understood and deserves more attention, especially in the context of the diversification of farm income and the development of sustainable farming systems, two issues of immense strategic importance to the future of European agriculture. There are economic, environmental and aesthetic reasons to encourage their adoption in all regions of the European Union.
ACKNOWLEDGEMENTS
This research was carried out as part of the SAFE (Silvoarable Agroforestry For Europe) collaborative research project. SAFE is funded by the EU under its Quality of Life programme; contract number QLKS-CT-2001-00560.
REFERENCES
Balabanian, O. (1984) Problemas agrícolas e reformas agrárias no Alto Alentejo e na Estremadura Espanhola, Lisboa.
Beaton, A. (1987) Poplars and agroforestry. Quarterly Journal of Forestry, 81, 225-233.
Beaton, A., Incoll, L.D., & Burgess, P.J. (1999) Silvoarable agroforestry. Scottish Forestry, 53, 28-32.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 151
Bertolotto, U., Pisanelli, A., & Cannata, F. (1995) Pratiche agroforestali nella regione Umbria. Monti e Boschi, 2, 5-11.
Burgess, S.O., Adams, M.A., Turner, N.C., & Ong, C.K. (1998) The redistribution of soil water by tree roots systems. Oecologia, 115, 306- 311.
Cannell, M.G.R., van Noordwijk, M., & Ong, C.K. (1996) The central agroforestry hypothesis: the trees must acquire resources that the crop would not otherwise acquire. Agroforestry Systems, 34, 27-31.
Carruthers, S.P. (1993) The dehesas of Spain - exemplars or anachronisms? Agroforestry Forum, 4, 43-52.
Correal, E. (1987) Trees and shrubs in the fodder and pastoral mediterranean systems. FAO European Network on Pastures and Forage Crop Production Bulletin, 5, 46-53.
Coulon, F., Dupraz, C., Liagre, F., & Pointereau, P. (2001). Étude des pratiques agroforestières associant des arbres fruitiers de haute tige à des cultures ou des pâtures. Solagro/Inra, Ministère de l'Aménagement et du Territoire et de l'Environnement.
Dawson, T.E. (1993) Hydraulic lift and water use by plants: implications for water balance, performance and plant-plant interactions. Oecologia, 95, 565-574.
Dupraz, C. (1994). Les associations d'arbres et de cultures intercalaires annuelles sous climat tempéré. In Revue Forestière Française, pp. 72- 83.
Dupraz, C. & Newman, S.M. (1997). Temperate Agroforestry: The European Way. In Temperate Agroforestry Systems (eds A.M. Gordon & S.M. Newman), pp. 181-236. CAB International.
EAFRD (2004). Proposal for a council regulation on support for rural development by the European Agricultural Fund for Rural Development (EAFRD), Rep. No. COM(2004)490 final. Commission of the European Communities, Brussels.
Eckert, G. (1995) Untersuchungen zur Geschichte der Landnutzung und zur Landschaftspflege auf brachgefallenen Wacholderheiden und Steinobstwiesen im Neidlinger Tal (Kreis Eßlingen) Verlag Ulrich E. Grauer, Stuttgart.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 152
Edelenbosch, N.E. & Dik, E.J. (1995). Mengteelt van populier met suikerbieten, snijmais en gras. Deel 1. Economische evaluatie van mengteelt van bomen met landbouwgewassen. Report IBN-DLO 181, Wageningen, the Netherlands.
Escribano, M. & Pulido, F. (1998). La dehesa de Extremadura: estructura económica y recursos naturales. Consejería de Agricultura y Comercio, Junta de Extremadura, Merida (Spain).
Gordon, A.M. & Newman, S.M. (1997) Temperate Agroforestry Systems CAB International, Wallingford, UK.
Grove, A.T. & Rackham, O. (2001) The nature of Mediterranean Europe: an ecological history. Yale University Press, New Haven, USA.
Hall, D.O. (1997) Biomass energy in industrialised countries - a view of the future. Forest Ecology and Management, 91, 17-45.
Hawke, M.F. & Wedderburn, M.E. (1994) Microclimatic changes under Pinus radiata agroforestry regimes in New Zealand. Agricultural and Forest Meteorology, 71, 133-145.
Herzog, F. (1994) Agroforestry and Land Use Change in Industrialised Nations. Schweizerische Zeitschrift für Forstwesen, 145, 761-763.
Herzog, F. (1998a) Agroforestry in temperate Europe: History, present importance and future development. In Mixed Farming Systems in Europe. Workshop Proceedings, Dronten, The Netherlands, 25-28 May 1998 (eds H. van Keulen, L.E. A. & v.L.H. H.), Vol. 2, pp. 47-52. A. P. Minderhoudhoeve-Series.
Herzog, F. (1998b) Streuobst: a traditional agroforestry system as a model for agroforestry development in temperate Europe. Agroforestry Systems, 42, 61-80.
Herzog, F. & Oetmann, A., eds. (2000) Communities of Interest and Agro-Ecosystem Restoration: Streuobst in Germany. Interactions between Agroecosystems and Rural Human Community. Advances in Agroecology. Springer, New York.
Hoare, A.H. (1928) The English Grass Orchard and The Principles of Fruit Growing. Ernest Benn, London.
Incoll, L.D., Rogers, D.G., Clements, R.O., Newman, S.M., Beaton, A., Burgess, P.J., Acworth, W.F., Wolfe, M.S., Russell, K., Jones, M., Sinclair, F.L., Teklehaimanot, Z., Jones, B.L., Watson, C.A., Sibbald,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 153
A.R., Hulbert, I.A.R., & McAdam, J.H. (2002) UK Agroforestry Forum: a survey of agroforestry in the British Isles, 2001. Agroforestry Forum Newsletter, 3, 2-63.
INE (2002). Censo Agrario de España 1999. Instituto Nacional de Estadística, Madrid.
Joffre, R., Vacher, J., De Los Llanos, C., & Long, G. (1988) The dehesa: an agrosilvopastoral system of the Mediterranean region with special reference to the Sierra Morena area of Spain. Agroforestry Systems, 6, 71-96.
Jose, S., Gillespie, A.R., & Pallardy, S.G. (2004) Interspecific interactions in temperate agroforestry. Agroforestry Systems, 61, 237- 255.
Lachaux, M., De Bonneval, L., & Delabraze, P. (1988) Pratiques anciennes et perspectives d'utilisation fourragère des arbres. Fourrages, 81-104.
Lapietra, G., Coaloa, D., & Chiarabaglio, P.M. (1991) Rapporto annuale sulla pioppicoltura 1990. Cellulosa e Carta, 3, 20-23.
Lelle, M.A. & Gold, M.A. (1994) Agroforestry systems for temperate climates: lessons from Roman Italy. Forest and Conservation History, 38, 118-126.
Liagre, F. (1993a) Les practiques de cultures intercalaires dans la noyeraie fruitière du Dauphiné. Mémoire de mastère en Science Forestières, ENGREF-INRA, Montpellier.
Liagre, F. (1993b) Les pratiques de cultures intercalaires dans le noyeraie fruitière du Dauphiné. MSc, ENGREF, Montpellier.
Lin, C.H., McGraw, R.L., George, M.F., & Garrett, H.E. (1999) Shade effects on forage crops with potential in temperate agroforestry practices. Agroforestry Systems, 44, 109-119.
Lucke, R., Silbereisen, R., & Herzberger, E. (1992) Obstbäume in der Landschaft Eugen Ulmer Verlag, Stuttgart.
MAPA (1985). Anuario de Estadística Agraria 1985. Ministerio de Agricultura, Pesca y Alimentación, Madrid.
Mary, F., Dupraz, C., Delannoy, E., & Liagre, F. (1998) Incorporating agroforestry practices in the management of walnut plantations in
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 154
Dauphiné, France: an analysis of farmers' motivations. Agroforestry Systems, 43, 243-256.
Meiggs, R. (1982) Trees and Timber in the Ancient Mediterranean World Oxford University Press, Oxford, UK.
Miguel, E., Pointereau, P., & Steiner, C. (2000). Los árboles en el espacio agrario. Importancia hidrológica y ecológica. Banco Santander Central Hispano, Madrid.
Newman, S.M. & Gordon, A.M. (1997). Temperate Agroforestry: Synthesis and Future Directions. In Temperate Agroforestry Systems (eds A.M. Gordon & S.M. Newman), pp. 251-266. CAB International, Wallingford, UK.
Newman, S.M., Park, J., Wainwright, J., Oliver, P., Acworth, J.M., & Hutton, N. (1991) Tree productivity, economics and light use efficiency of poplar silvoarable systems for energy. In Proceedings of the 6th European Conference on Biomass Energy Industry and Environment, Athens.
Nichols, A.R., Kendall, D.A., & Iles, D.R. (2000) The agronomic and environmental implications of a combined food and energy system. Aspects of Applied Biology, 58, 363-372.
Parsons, J.J. (1962) The acorn-hog economy of the oak woodlands of southwestern Spain. Geographical Review, 52, 210-235.
Potter, T.W. (1979) The changing landscape of South Etruria. Elek, London.
Roach, F.A. (1985) Cultivated Fruits of Britain. Their Origin and History. Blackwell, Oxford.
Romero Candau, L. (1981) La dehesa como forma de explotacion agraria. Problemas actuales. Estudios territoriales Ceotma, 3, 149-152.
Rösler, M. (1996) Marktwirtschaftliche Bedeutung des Streuobstbaus. In Auswirkung des EU-Streuobstforums im Rahmen der 2. Rhöner Apfelmesse., pp. 3-10. Rhöner Apfelbüro, Zella.
Smith, J. (1990). La Forêt et le Bois dans le Marché Unique. Etude du club de Bruxelles.
Stevenson, A.C. & Harrison, R.J. (1992) Ancient forests in Spain: a model for land-use and dry forest management in south-west Spain from
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 155
4000 BC to 1900 AD. Proceedings of the Prehistoric Society, 58, 227- 247. van Noordwijk, M., Lawson, G., Soumaré, A., Groot, J.J.R., & Hairiah, K. (1996). Root distribution of trees and crops: competition and/or complementarity. In Tree-Crop Interactions: A Physiological Approach (eds C.K. Ong & P. Huxley), pp. 319-364. CAB International, Wallingford, UK.
Vandermeer, J. (1989) The Ecology of Intercropping. Cambridge University Press, Cambridge.
Wilson, A. (1993) Silvopastoral agroforestry using honeylocust. In Proceedings of the Third North American Agroforestry Conference, pp. 265-269, Ames, Iowa, 16-18 August 1993.
26
24
22
20
18
16 Millions of trees of Millions
14
12
10
1900 1920 1940 1960 1980 2000 Year
Fig. 1. Number of fruit trees in Streuobst systems in Baden-Wurttemberg (in 1900 and 1912 trees in home gardens are included). Redrawn from Herzog & Oetmann (2001).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 156
Fig. 2. Distribution of dehesas within Spain and Portugal. Reproduced from Blanco-Castro et al. (1997) –.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 157
800000
700000
600000
500000
400000
300000
200000 Area of intercropped woodland (ha)
100000
1960 1970 1980 1990 2000 Year
Fig. 3. Area of intercropped woodlands (equivalent to planted dehesas) from 1962-2001 in Spain. Data from Anuario de Estadística Agraria (Annual of Agricultural Statistics) 1962-2001.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 158
Component Stems -1
System name trees Tree species Location Extent (ha) ha Layout Fruit Annual crops Timber Fodder
Firewood Olive systems
Olea europaea var. Olive groves Olive europaea Italy 20,000 25 - 100 S / L Y Y Y C - V - FL - FG Other fruit trees Rosaceae France 3,000 25 - 300 S / L Y Y C C - M - FL - GV - Greece 650,000 50 - 100 S / L Y Y GF Spain 15,030 50 - 100 L Y Y C
Orchard systems
Almond Prunus dulcis Sicily 18,000 50 - 100 S / L Y C - FL - FG
Mixed Rosaceae France 2,000 50 - 300 S / L Y V - GV - GF
Joualle Peach Prunus persica France 100 200 - 300 L Y Y GV Walnut Juglans regia Olea europaea var. Olive europaea
Mulberry Morus nigra N Greece 500 10 - 50 S / L Y M - FL - V
Crete, Aegean Fig Ficus carica Islands 10,200 10 - 50 S / L Y Y C
Common pear Pyrus communis N & C Greece 7,000 20 - 50 S / L Y Y C - T - V - GV
Mixed Rosaceae Spain 13,484 40 - 200 S / L Y C - M - V - BF
Timber trees
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 159
Component Stems -1
System name trees Tree species Location Extent (ha) ha Layout Fruit Annual crops Timber Fodder
Firewood
Poplar plantations Poplar Populus cv. N Italy 12,500 200 L Y C - M - S France 6,300 180 - 220 L Y C - M Greece Unknown 10 - 50 L Y M - FL - V
Walnut plantations Walnut Juglans regia C & S Italy 10,000 25 - 100 L Y Y Y C - V - FL Hazel Corylus avellana
Walnut Juglans nigra, J. regia France 15,000 * 80 - 120 L Y Y Y All crops Walnut Juglans nigra, J. regia Greece (montane) 7,600 10 - 25 S Y Y Y C - T - FL - GV
Oak systems
Quercie C & S Italy, Sardinia, camporili Oak Quercus spp. Sicily 180,000 10 - 100 S Y Y C - FL Pear Pyrus spp.
Les plantades Oak Quercus spp. France 100 50 - 150 S / L Y Y FG
Q. ithaburensis subsp. Valonia oak macrolepsis S & W Greece 29,600 10 - 50 S Y Y C
C - T - Sun - FL - Downy oak Q. pubescens N & C Greece 1,470,000 * 10 - 100 S Y Y GV Sessile oak Q. petraea Turkey oak Q. cerris Macedonian oak Q. trojana
Evergreen Dehesas oak Q. ilex subsp. ballota W & SW Spain 2,300,000 * 10 - 40 S Y Y C - Sun - FL
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 160
Component Stems -1
System name trees Tree species Location Extent (ha) ha Layout Fruit Annual crops Timber Fodder
Firewood Cork oak Q. suber Pyrenean oak Q. pyrenaica
Fodder trees
Carob Carob Ceratonia siliqua Sicily 20,000 100 - 125 S / L Y Y Y C - FL Crete 7,900 25 - 100 S / L Y V - GF
* only a proportion is intercropped in any single year.
Table 1. Extant silvoarable agricultural systems in Europe, their composition, present extent, structure and main economic products. Layout of stems is either scattered (S) or linear (L). Annual crops sown between the stems include maize (M), other cereals (C), vegetables (V), oil seed rape (OSR), soya (S), tobacco (T), sunflower (Sun), fodder legumes (FL), fodder grasses i.e. hay (FG), bush fruits such as Ribes spp. (BF), ground fruits (GF) and grape vines (GV).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 161
System 1962 1972 1982 1989 1999
Fruit + annual crop a 402,005 78,999 27,562 13,484
Vineyard + annual crop a 21,677 8,175 8,359
Olive + annual crop a 242,628 39,092 20,219 15,030
Woodland tree + annual crop b 685,893 478,375 433,000 357,000 213,100 c
a Data derived from National Agriculture Census (INE, 1963, 1975, 1985, 1991 and 2002). b Data derived from Annual Report of Agricultural Statistics (MAPA, 1985 and 2001). Refers to annual crops with presence of some mature woodland trees covering between 5 – 20 % of the surface (open woodland). This type of intercropped system refers mainly to dehesas (in more than 90% of cases). c MAPA (2001) gives a value of 600,000 ha in 1999, which highlights the lack of an adequate definition of agroforestry.
Table 2. Recent trends in the extent of silvoarable systems (in hectares) in Spain.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 162
ANNEX 7. The development and use of a framework for characterising computer models of silvoarable economics Accepted for publication in Agroforestry Systems
Running head (shortened title): Agroforestry economic models
A.R. Graves 1, P. J. Burgess 1, F. Liagre 2, J-P. Terreaux 3 & C. Dupraz 4
1 Cranfield University, Silsoe, Bedford MK45 4DT, UK;
2 Assemblée Permanente des Chambres d’Agriculture, 9 Avenue Georges V, 75008
Paris, France;
3 Cabinet d’expertises forestières, Chavet, Paris, France;
4 Institut National de la Recherche Agronomique, 2 Place Viala, 34060 Montpellier,
France
Full address for correspondence:
Mr A.R. Graves, Cranfield University, Silsoe, Bedford, MK45 4DT, U.K
Telephone number: +44-(0)1525-863107
Facsimile number: +44-(0)1525-863344
E-mail address: [email protected]
Key words: ARBUSTRA, Agroforestry Estate Model, Agroforestry Calculator, farm,
POPMOD, WaNuLCAS
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 163
The development and use of a framework for characterising computer models of silvoarable economics
ABSTRACT
A review of existing computer models of silvoarable economics was undertaken for a project, entitled ‘Silvoarable Agroforestry for Europe’ (SAFE), which aims to reduce uncertainty regarding the introduction and management of silvoarable systems in
Europe. Because the published literature describing and comparing such models is sparse, a framework was developed and then used to characterise five computer models: POPMOD, ARBUSTRA, the Agroforestry Estate Model, WaNuLCAS, and the Agroforestry Calculator. The key characteristics described for each model were: the background, the systems modelled, the objective of the economic analysis, the economic viewpoint, the spatial and temporal scales, the generation and use of biophysical data, the model platform and interface, and the input requirements and outputs. Each of the models could produce a partial budget of the profitability of a silvoarable, arable or forestry system at a one-hectare level using discounted cost- benefit analysis. Whilst the research models undertook the analysis from a viewpoint of a generic farmer, the models developed for decision-support also included appraisals from the perspectives of tenants, share-croppers and participants in a joint-venture. The two farm-scale models, ARBUSTRA and the Agroforestry Estate
Model, could also be used to examine the feasibility of silvoarable systems on an existing business, and to determine the effects of heterogeneous land types and phased planting. The framework allows users to identify the pertinent issues for selecting or developing a particular model.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 164
INTRODUCTION
Computer-based models of silvoarable economics (CMSEs) are coherent, numerical representations of the economic structure and processes of silvoarable enterprises, which can be manipulated to compare economic opportunities, determine feasibility, optimise management practices, or predict economic behaviour. The importance of agroforestry economics has often been stated (Singh et al., 1998, Nelson et al.,
1998; Chianu et al., 2002), but there is little published information describing CMSEs and their development. Biophysical issues continue to dominate agroforestry research, despite the observation that many agroforestry projects fail as a result of inadequate attention given to socio-economic factors (Mercer et al., 1998).
In the 1960s, the first crop simulation models were developed on mainframe computers to estimate light interception and photosynthesis (Bouman et al. 1996;
Loomis & Williams, 1962). At a similar time, the first forestry simulation models were also developed, using distance-dependent growth models to understand the effect of management practice on stand development (Fries, 1974). Biophysical simulations of agroforestry systems commenced in the 1980s, and included evaluations of the potential of agroforestry on grazing land in New Zealand (Arthur-Worsop, 1984), and the intercropping of crops with pine in North Carolina in the USA (McNeel & Stuart,
1984).
Early computer models of agroforestry economics tended to focus on silvopastoral systems and used forestry models to simulate the returns from trees (Arthur-Worsop,
1984; Cox et al., 1988). In China, computer “models” were used to optimise the intercropping of Paulownia with arable crops, according to economic, ecological and social objectives (Jiang et al., 1986; Qun, 1991). Etherington and Matthews (1984) describe a computer programme that was used for developing partial and whole-farm
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 165
budgets for land-use systems involving trees. Subsequently there has been the development of a range of computer models of silvoarable economics including
POPMOD (Thomas, 1991), ARBUSTRA (Liagre, 1997) and the Agroforestry Estate
Model (Knowles and Middlemiss, 1999)
Since 1992, the European Union has introduced a series of measures to promote the integration of trees within existing farm businesses. In 2001, a project entitled
‘Silvoarable Agroforestry for Europe’ (SAFE) was initiated by the European
Commission to reduce the uncertainties regarding silvoarable systems in Europe. An important objective of the project was the development and use of a computer-based model of silvoarable economics to compare the profitability of silvoarable, arable and forestry systems at a one-hectare scale and to determine the feasibility of silvoarable systems at the farm scale. The aim of this paper is to describe the development and use of a framework for characterising existing models of silvoarable economics. This can then be used to provide a coherent means of considering the key characteristics required in the new model.
MATERIALS AND METHODS
Development of a framework for characterising models
In order to provide a consistent approach for describing and comparing existing models of silvoarable economics, it was necessary to develop a framework for characterising those models. Because, to our knowledge, there is no existing framework for the categorisation of computer models of silvoarable economics, the framework was derived from criteria used for other categorisations of agroforestry systems, economic analyses and computer models.
The importance of providing adequate background to the model was evident from the register of ecological models described by Kassel University and GSF (2004) and the
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 166
review of bio-economic models by Brown (2000). The bibliographic framework for categorising of the economic analysis of agroforestry technologies, developed by
Swinkels and Scherr (1991), provide background in terms of the surname of the author, the language and the geographical location. In a review of crop-soil simulation models, Matthews (2002) grouped models according to their use in research, decision-support or education.
An important feature in many of the categorisations is a description of the systems modelled. This is included in the framework used by Swinkels and Scherr (1991) and
Brown (2000). Swinkels and Scherr (1991), and Antonopoulou (2003) in a review of decision-support systems for crop growth and management, also describe the crops that are modelled.
Brown (2000) notes the importance of understanding the objectives of the model, taking note of the mathematical approach used and defining the temporal and spatial scales. Swinkels and Scherr (1991) also describe six types of economic analysis including cost-benefit analysis and optimisation, and seven levels of analysis including the research plot, the farm, the project, and the region. As the models are computer-based, the provision of technical information on the software and a description of the input and output of data are also important (Kassel University and
GSF, 2004; Brown, 2000).
The final framework (Table 1) comprised nine major divisions. These were: (1) the model background, (2) the systems modelled, (3) the objective of the economic analysis, (4) the economic viewpoint, (5) the spatial scale, (6) the temporal scale, (7) the generation and use of biophysical data, (8) the model structure and interface, and
(9) the input requirements and the outputs generated.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 167
Table 1
USE OF THE FRAMEWORK
The framework described in Table 1 was used to characterise five contrasting
computer models of silvoarable economics. The first two models, which have been
used extensively by the authors, are POPMOD and ARBUSTRA. POPMOD was
developed by the Bio-economic Agroforestry Modelling Project at the University of
Wales, in Bangor, North Wales (BEAM project, 2002; Thomas, 1991). ARBUSTRA
was developed by the “Equipe de Recherches en Agroforesterie” within the Institute
National de la Recherche Agronomique (INRA) in Montpellier, France (Liagre, 1997).
The third model to be considered was the Agroforestry Estate Model developed by
Forest Research in New Zealand (Knowles and Middlemiss, 1999, Forest Research
2002). The fourth model was WaNuLCAS (Water Nutrients and Light Capture in
Agroforestry Systems), which was developed by the International Centre for
Research on Agroforestry (ICRAF) under the Southeast Asia Programme (van
Noordwijk and Lusiana, 1999, 2000, 2003). Although it is primarily a biophysical
model, it can undertake economic evaluations and its inclusion broadens the scope
of the comparison. The last model to be considered was the Agroforestry Calculator
(Department of Agriculture, 2002) developed by Campbell & White Associates Pty
Ltd in Australia under the “Decision-support for Adoption of Agroforestry Project”.
Although it was originally developed to focus on silvopastoral economics it can also
be used to evaluate silvoarable systems.
RESULTS AND DISCUSSION
Model background
The date of first mention in white or grey literature, the country of origin and the
language provide a context for each model. This is important as many of the inputs,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 168
such as grant payments, are “closed” or pre-defined (Thomas and Willis, 1997). This is the case in POPMOD and ARBUSTRA. A “closed” approach is useful if the model is used in scenarios for which it was developed, as the user receives guidance on the input data required. However it can create difficulties if the model is used in a different context. For example both ARBUSTRA and POPMOD require modification to model the current silvoarable grant system in Spain. The Agroforestry Estate
Model is an example of an “open” model which can be easily used for different economic scenarios. The language of the model is also important as this determines the number of users who understand and operate the model. For example
ARBUSTRA is available in French, whereas WaNuLCAS is available in English,
Bahasa Indonesian and Portuguese (Van Noordwijk and Lusiana, 2003).
Models are generally developed for the purpose of research, decision-support or education (Matthews 2002; Graves et al., 2002). As research tools, models of silvoarable economics can be used to compare systems and management practices.
This may help to identify knowledge gaps, generate and test hypotheses, and determine key parameters. As decision-support tools, models can be used to identify preferred enterprises or scenarios or to optimise resource-use amongst a given set of enterprises. In education, models can allow students to investigate the long-term interactions of silvoarable systems, without the time requirements or financial costs of real experiments.
Three of the selected models, POPMOD, ARBUSTRA and WaNuLCAS, were primarily developed as research tools. POPMOD was initially developed to compare agrosilvopastoral systems with poplar in England (Thomas, 1991) and to examine the effect of recently-introduced fast-growing poplar hybrids on the profitability of silvoarable systems (Willis et al. 1993). The aim of ARBUSTRA was to evaluate the
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 169
effect of introducing silvoarable systems to farms in the Midi-Pyrénées. A major objective of WaNuLCAS was to synthesise existing knowledge and hypotheses on above and below-ground resource-use by trees and crops at the “patch-scale” (van
Noordwijk and Lusiana, 1999).
The Agroforestry Estate Model and the Agroforestry Calculator were initially developed with the aim of decision-support. The Agroforestry Estate Model was developed to evaluate the physical and financial impact of agroforestry projects on existing farms in New Zealand, Australia, Canada and the United States (Knowles and Middlemiss, 1999). The Agroforestry Calculator was developed to give farmers and consultants an easy way of estimating the profitability of plot (one-hectare) scale agroforestry projects and comparing these with existing enterprises.
Once developed, a model can be used for other purposes. For example ARBUSTRA has also been used as a decision-support tool to advise farmers on the economic impact of silvoarable projects in France. Similarly POPMOD, the Agroforestry Estate
Model and WaNuLCAS have also been used in graduate education. Because decision-support and education models need to be readily understood by new users, they tend to be better “finished” and more clearly presented than research models.
Decision-support models will also tend to distinguish between the specific requirements of different types of user, such as owner occupiers and tenant farmers.
Systems modelled
Agroforestry systems can be described by their components (trees, crops, and animals), and their temporal (ranging from coincident to sequential) and spatial arrangement (Nair, 1985). Each of the five models was able to model agroforestry systems where trees and crops are grown simultaneously, while the Agroforestry
Estate Model, WaNuLCAS and ARBUSTRA are also able to model sequential
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 170
systems. POPMOD can be used to model one arable, silvoarable and forestry system concurrently, which makes rapid comparison between different systems possible; ARBUSTRA can model as many as six arable, silvoarable and forestry systems simultaneously. By contrast, WaNuLCAS can model only one system at a time, so that comparisons between forestry, silvoarable and arable systems need to be made by consecutive runs. The Agroforestry Estate Model and Agroforestry
Calculator model a “current scenario” and the impact of agroforestry on the “current scenario”.
Objective of economic analysis
The objectives for undertaking an economic analysis of a silvoarable system can include comparison, an assessment of feasibility, optimisation, and prediction of actual farmer behaviour. Each of these can also be subjected to uncertainty analysis.
Comparison
Each of the models can be used to compare the economic effect of different silvoarable, arable, or forestry systems, using a partial budget. When one arable system is compared with another, a partial budget is usually undertaken on the basis of the gross margin (revenue minus variable costs) on a per hectare basis. The variable costs are those costs, such as seed, fertiliser and sprays, which are specific to a system and vary in proportion to the area. Although labour and machinery costs may be regarded as ‘fixed’ over a short period of time, it is often possible to assign them to a specific system. As such they can be termed ‘assignable fixed costs’. In forestry, the costs of labour and machinery are typically included, and therefore for comparisons of arable and forestry systems, it is best to compare the ‘net margin’
(revenue minus variable and assignable fixed costs) (Willis et al., 1993).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 171
A second implication of comparing silvoarable, arable and forestry systems is the need to aggregate the economic evaluation over a defined period of time. Whereas an economic comparison of two arable crops is often undertaken on an annual basis, the economics of a forestry plantation is generally considered over a rotation which lasts many years. Because of inflation, the opportunity cost of money and the increased flexibility of having money available now, most people will ‘discount’ the value of future income. Discounting is a method that allows the user to directly compare money realised at different periods of time (Barnard and Nix 1979). By discounting the value of, say timber production in year 50, it is possible to calculate the present value of the timber, as if it was available in year 1. In each of the five models, a cash-flow is dynamically simulated over time, so that the user can compare annual and perennial systems over specified time horizons, using the net present value (NPV) at a selected discount rate.
Feasibility
A second potential objective for an economic model is to determine if a combination of systems is feasible within a specified context. The approach is similar to that used for comparison, but the combined effect of more than one system on a specified resource, such as timber output, labour or cash-flow, is also ascertained. Because the inputs and the outputs of forestry and silvoarable systems can be “lumpy”, farmers may decide to smooth input requirements and output flows by planting trees over a number of years (phased or multiple planting). Both the Agroforestry Estate
Model and ARBUSTRA are capable of determining the feasibility of silvoarable systems in a whole-farm context.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 172
Optimisation
A third potential objective for economic models is to inform the user how to optimise a given system, when faced with a particular or numerous objectives, within specified constraints (Mendoza et al., 1987). For example: determining the optimum area of agroforestry to maximise profit and output, given a limited budget. None of the models reviewed here were specifically designed for optimisation.
Prediction of actual farmer behaviour
Economic models can also be used to predict the actual response of farmers to changes in cost, prices or policy - for example, determining the likely change in the area of agroforestry following a change in government grants. In relation to forestry or arable systems, the possible approaches used have included positive mathematical programming (Judez et al. 2001) and positivistic mathematical programming (C. Yates, pers. comm. 2001). However none of the selected economic models use these approaches.
Uncertainty analysis
All of the reviewed models are deterministic in that each input is typically described by a mean and a given set of inputs leads to a uniquely definable outcome.
However, biophysical and economic predictions are subject to uncertainty and the use of single values can be misleading. One alternative approach is to develop a stochastic model, where the variability of selected inputs is defined, for example, by standard deviations or the random selection of historical data. The resulting output values can also be described in terms of a mean and standard deviation. A second alternative approach is use the model to undertake an analysis of the sensitivity of a specified output to changes in the value of an input. ARBUSTRA allows a sensitivity analysis of the effect of the discount rate, the price of wood, the agricultural gross
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 173
margin and tree subsidies on the NPV. The Agroforestry Calculator allows a sensitivity analysis of the effects of the tree returns and the response of the silvoarable crop to trees on the NPV. Sensitivity analysis can also be achieved using
POPMOD, the Agroforestry Estate Model and WaNuLCAS, but this requires the user to manually change each input, re-run the model and store the outputs. However even with these models there may be the potential to link them to additional software that allows the automation of such analyses.
Viewpoint of the economic analysis
An important model characteristic is the viewpoint of the economic analysis. Each of the five reviewed models primarily provides a micro-economic, farmer-based analysis of the modelled systems. From such a viewpoint government subsidies are considered as revenue and taxes can be considered as a cost.
Amongst the five sampled models, those models that have been used for decision- support tend to include a wider range of viewpoints than those developed for research. For example ARBUSTRA and the Agroforestry Calculator allow economic analyses from the perspective of a tenant or an owner involved in a share-cropping arrangement respectively. The Agroforestry Estate Model extends the possible viewpoints to include joint-venture agreements between a farmer and a landowner, and cutting right agreements, where the right to harvest the trees at the end of the rotation is bought before the trees are clear-felled.
An alternative viewpoint is the macro-economic perspective or that of society as a whole. This may require adjustment to some of the market prices to reflect the public value. Likewise taxes could be included as revenue and subsidies as costs. A consideration of the environmental costs and benefits of the system may also be included. The reviewed models are not specifically designed to provide a macro-
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 174
economic analysis, but there are options to alter prices and to set the level of grants
to zero. ARBUSTRA can also be used to determine the cost of a project “to society”
and the Agroforestry Calculator includes the economic effects of soil degradation.
Spatial scale
The spatial scale for the analysis of silvoarable economics may be at the one-
hectare, field, farm, or even the regional, national or global scale (Figure 1). The
smallest scale is typically “one-hectare” on which a single system (i.e. arable, forestry
or silvoarable) completely occupies a homogenous area. A “field-scale” analysis
differs from a “one-hectare-scale” analysis, in that the revenues and costs related to
headlands and field boundaries can be included. Whereas each of the five models
can generate results at the ‘one-hectare-scale’, ARBUSTRA is also capable of
incorporating the effects of headlands within the silvoarable calculations.
Figure 1
A “farm-scale” analysis is used to determine the effect of a combination of arable,
forestry or silvoarable systems on the resources of a specified farm business. Of the
models examined, only ARBUSTRA and the Agroforestry Estate Model are designed
for economic analysis at a farm-scale. Both models include fixed-costs, can model
several systems simultaneously, and allow an evaluation of the overall effect of
introducing new systems. Both also allow the user the option of defining more than
one possible planting date for silvoarable and forestry systems. An additional feature
of the ARBUSTRA farm-model is the inclusion of heterogeneous land types on an
individual farm. The inclusion of six areas of different fertility with arable and/or
silvoarable and/or forestry systems on each allows for variations in productivity
across the farm.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 175
Above the farm-scale, analyses could also be taken at a project-, catchment-,
community-, regional-, national- or international-scale (Swinkels and Scherr, 1991).
None of the models examined were specifically designed for such scales, but they
could be developed for such a use if integrated with spatial biophysical and economic
data within a Geographical Information System.
Temporal scale
The temporal scale relates both to the time-increment for the economic calculations
and the maximum time period that can be considered. Each of the five selected
models, except WaNuLCAS, run on an annual time-step; WaNuLCAS can be run on
an hourly, or daily time-step. The original version of POPMOD was designed to run
over a 30 year period, which was the typical rotation period for poplar in the UK. At
the other extreme, ARBUSTRA can run simulations over 120 years. In ARBUSTRA,
there is also a mathematical procedure for calculating an “infinite” NPV.
Generation and use of biophysical data
A distinguishing characteristic of computer models of silvoarable economics is the
method by which biophysical data are generated and used. The simplest structure is
a stand-alone economic model that operates without a biophysical module (Figure 2).
This is the structure used by ARBUSTRA and the Agroforestry Calculator. In
ARBUSTRA, the arable systems and the silvoarable crop are described solely in
terms of a gross margin entered by the user. In the Agroforestry Calculator only a
gross margin value is required for the “current enterprise”.
Figure 2
The second and third forms are effectively bio-economic models in that they include
both biophysical and economic modules. The biophysical module may feed
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 176
biophysical data to the economic component of the model in a one-way flow of information or may allow the economic component of the model to feed information back to the biophysical module in a two-way flow (Brown, 2000). In the simplest version of WaNuLCAS, the production data are fed to the economic module and there is no feedback. In POPMOD, there is a two-way flow of information. For example, if the silvoarable crop component is no longer profitable, then the economic module can instruct the biophysical module to stop “planting”. In turn, the cessation of arable cropping can affect the productivity of the tree component of the silvoarable system.
The biophysical component of a model can be characterised as empirical, functional, or mechanistic (Passioura, 1996). It is worth noting that all models reach empirical boundaries at some point and that these boundaries are effectively a matter of
“scale”. For example POPMOD uses an empirical biophysical model of annual crop yield and it derives timber output from yield tables for poplar at stated stand densities and yield classes. By contrast WaNuLCAS includes a number of mechanistic biophysical modules. Although this allows tree and crop yields to respond directly to changes in the climate and soil conditions, it does require additional inputs and functions.
Model platform and interface
The choice of the software platform for the model is dependent on availability and costs, the suitability for the task, the knowledge of the developers, and the ability to transfer the models between users. Both POPMOD and ARBUSTRA were developed as spreadsheet models in QUATTRO PRO (Borland International Inc,
Scots Valley, California). The Agroforestry Estate Model was originally developed in
Microsoft® Excel (Microsoft Corporation, Seattle, Washington, USA) and has recently
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 177
been re-written in Visual Basic. WaNuLCAS was developed in a graphical development environment called Stella (High Performance Systems Inc, Hanover,
New Hampshire, USA), but the biophysical and economic inputs and outputs can be developed or obtained in linked Microsoft® Excel spreadsheets. The Agroforestry
Calculator is a spreadsheet model developed in Microsoft® Excel.
The model interface can determine how easily a model can be operated by a new user. ARBUSTRA and the Agroforestry Estate Model, which have been used for decision-support, ease the user’s task by providing a graphical user interface (GUI) for data input and model navigation. In contrast POPMOD allows direct access to logically grouped input cells in the spreadsheet. Such transparency in the working of the model can be particularly useful to researchers and developers.
Input requirements and outputs generated
The ability of the user to interact with the model depends in part on the user’s familiarity with the modelling platform described in the preceding section. However it also depends on the input requirements, the availability of associated databases, and the provision of outputs.
Inputs required
One factor determining the level of inputs is whether the model operates at a one- hectare- or a farm-scale. POPMOD and the Agroforestry Calculator, which are one- hectare-scale models, generally require fewer inputs than farm-scale models such as
ARBUSTRA and Agroforestry Estate Model. The input requirements also increase if there are a wide range of options. For example, in both ARBUSTRA and
Agroforestry Estate Model there is the capacity to use different machine and labour costs for different silvicultural operations, i.e. silvicultural operations undertaken by the farmer may be costed at a different rate than work undertaken by contractors.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 178
A second factor determining the input requirements for an economic model is the requirement to provide or generate biophysical data. For example, the two stand- alone models, ARBUSTRA and the Agroforestry Calculator, describe timber production from a discrete value provided by the user for the timber yield in the year of harvest. By contrast, bio-economic models, such as POPMOD and the
Agroforestry Estate Model, require a continuous set of timber volume values for each year of the tree rotation. These data may come from field measurements, historical records or biophysical models. Of the five models, WaNuLCAS requires the most inputs, because of the many data required for mechanistic modelling.
Availability of databases
Although peripheral to the model itself, the provision of databases can be important in determining the ease with which the model can be used. Both WaNuLCAS and
Agroforestry Estate Model allow access to databases of inputs, and the Agroforestry
Calculator allows users to add or change databases within the model itself.
Outputs generated
Each of the five models calculates a Net Present Value for the modelled system.
Other values that are commonly calculated include the annuity value, the benefit: cost ratio, the payback period and a description of the cash-flow of the modelled systems. The generation of further economic outputs is partly determined by the spatial scale of the model. Hence the two farm-scale models, ARBUSTRA and
Agroforestry Estate Model, provide a farm Net Present Value and describe the amount of land occupied by different systems and the effect on farm labour requirements.
The production of physical outputs from the economic model is primarily determined by the relative importance of the biophysical component of the model. Hence bio-
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 179
economic models such as POPMOD and Agroforestry Calculator can provide annual timber yields. However, the greatest range of biophysical outputs, including the soil- water and nutrient status, is provided by WaNuLCAS.
CONCLUSIONS
A framework of analysis (Table 1) was developed and used to provide a consistent approach for describing and comparing the selected models. The results of a comparison of five models are summarised in Table 2. This was done to help inform the development of an economic model for a project to reduce the uncertainty of introducing and managing silvoarable systems in Europe.
Table 2
The first stage of the framework was to describe the background for each model, in terms of the country of origin and initial purpose. The models developed for research tended to provide a more generic analysis than those developed for decision-support, which included, for example, appraisals of share-cropping, tenant farming, lease- holding, joint-ventures and cutting right schemes.
A common feature of each model was the need to combine short-term, e.g. annual crops, and long-term, e.g. timber production, enterprises within the same partial budget. This was typically done on the basis of a net margin. In each of the models, the long-term nature of silvoarable and forestry systems was accounted for by discounting future costs and revenues to derive a net present value at a specified discount rate.
The principal objectives of the reviewed models were to compare the economic effect of alternative silvoarable systems relative to agriculture or forestry or to determine the feasibility of silvoarable enterprises in a farm context. Comparisons of the profitability
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 180
of the systems were made using partial budgets at a one-hectare scale.
Examinations of feasibility were made with farm-scale models which could include spatial heterogeneity within the farm, multiple planting schemes and the interactions between different systems.
The models intended for decision-support tended to have a clearer interface for new users than those developed for research. However those models which operate solely at a spreadsheet, such as POPMOD, are transparent and are relatively easy for a researcher to manipulate. The actual choice of a model may also depend on the cost and the availability of the software.
The full range of features described here may be difficult to satisfy in one model and compromises are often made in selecting or developing a model. The framework developed here allows model developers and users to identify the pertinent issues and prioritise what is most important for their needs.
Acknowledgements
This research was carried out as part of the SAFE (Silvoarable Agroforestry for
Europe) collaborative research project. SAFE is funded by the EU under its Quality of Life programme, contract number QLF5-CT-2001-00560, and the support is gratefully acknowledged. We also acknowledge the valuable comments on this paper from Leith Knowles and Lars Hansen.
REFERENCES
Antonopoulou E (2003) DSSs in major field crops: classification and performance. In: Harnos Z, Herdon M and Wiwczaroski TB (eds) Proceedings of the 4th Conference of the European Federation for Information Technology in Agriculture, Food and the Environment (EFITA). Volume 1 pp 103-113. University of Debrecen, Hungary. (Accessed 19 January 2004). http://www.date.hu/efita2003/centre/pdf/0110.pdf
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 181
Arthur-Worsop MJ (1984) An economic evaluation of agroforestry: the national viewpoint. In: Bilbrough GW (ed) Proceedings of a Technical Workshop on Agroforestry pp 61-70. New Zealand Forest Service.
Barnard CS and Nix J (1979) Farm Planning and Control. 2nd Edition. Cambridge University Press, Cambridge, UK.
BEAM project (2002) BEAM models – POPMOD (poplar model). BEAM project, University of Wales, Bangor, Wales. (Accessed 19 January 2004). http://www.bangor.ac.uk/~afs117/popmod.htm
Bouman BAM, van Keulen H and Rabbinge R (1996) The 'School of de Wit' crop growth simulation models: a pedigree and historical overview. Agricultural Systems 52: 171-198.
Brown DR (2000) A review of bio-economic models. Cornell African Food Security and Natural Resources Management Program, Cornell University, USA. 101 pp. (Accessed 19 January 2004) http://aem.cornell.edu/faculty_sites/cbb2/grad/Brown/BrownReviewofBioeconomicMo dels.pdf
Chianu JN, Akintola JO and Kormawa PM (2002) Profitability of cassava-maize production under different fallow systems and land-use intensities in the derived savanna of southwest Nigeria. Experimental Agriculture 38:51-63.
Cox O, McGregor M and Maclaren P (1988) Agroforestry components of the radiata pine stand model. In: Maclaren P (ed) Proceedings of the Agroforestry Symposium, Rotorua, 24-27 November 1986. pp 175-182. Forest Research Institute-Bulletin 139, New Zealand.
Department of Agriculture (2002). Agroforestry Calculator. Department of Agriculture, Western Australia. (Accessed 19 January 2004). http://agspsrv34.agric.wa.gov.au/environment/tools/trees/Agroforestry_Calculator.ht m
Etherington DM and Matthews PJ (1984). MULBUD User’s Manual. National Centre for Development Studies, Australia National University, Canberra. 115 pp.
Fries J (1974). (ed) Growth models for tree and stand simulation. Research Notes 30. Department of Forest Yield Research, Royal College of Forestry, Stockholm, Sweden. 379 pp.
Forest Research (2002). Agroforestry Estate Model. Forest Research, Rotorua, New Zealand. Accessed 19 January 2004. http://www.forestresearch.co.nz/default.asp
Graves AR, Hess T, Matthews RB, Stephens W and Mason T (2002) Crop simulation models as tools in education. Journal of Natural Resources and Life Sciences Education 31: 48-54
Jiang J, Zhu J, Wu L and Liu T (1986) Comprehensive benefit analysis and optimisation of Paulownia intercropping combinations. In: Hsuing W and Chandler
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 182
PF (eds) Agroforestry research and practice, pp 163-173. China Publishing House, Beijing.
Judez L, Chaya C, Martinez S and Gonzalez. AA (2001) Effects of measures envisaged in "Agenda 2000" on arable crop producers and beef and veal producers; an application of positive mathematical programming to representative farms of a Spanish region. Agricultural Systems 67: 121-138
Kassel University and GSF (2004). The Register of Ecological Models. (Accessed 19 January 2004). http://eco.wiz.uni-kassel.de/ecobas.html
Knowles L and Middlemiss P (1999) Evaluating Agroforestry Options: A Continuing Professional Development Course. Forest Research, Rotura, New Zealand.
Liagre F (1997) ARBUSTRA Manuel de l’utilisateur. User manual for ARBUSTRA, Centre Régional de la Propriété Forestière (CRPF) and l'Institut National de la recherche Agronomique (INRA) Montpellier, France. 71pp.
Loomis RS and Williams WA (1962) Maximum crop productivity: an estimate. Crop Science 3: 67-72.
Matthews RB (2002) Introduction. In: Matthews RB and Stephens W (eds). Crop- soil simulation models: applications in developing countries, pp 1-5. CAB International, Wallingford, UK.
McNeel JF and Stuart WB (1984) Feasibility of agri-silviculture for pine plantations in the south. Paper, American Society of Agricultural Engineers, No. 84-1609, 14 pp.
Mendoza, GQ, Campbell GE and Rolfe GL (1987). Multiple objective programming: an approach to planning and evaluation of agroforestry systems: Part 2 – an illustrative example and analysis. Agricultural Systems 23: 1-18.
Mercer DE, Miller RP, Nair PKR and Latt CR (1998) Socioeconomic research in agroforestry: progress, prospects, priorities. Agroforestry Systems 38: 177-193.
Nair PKR (1985) Classification of agroforestry systems. Agroforestry Systems 3: 97- 128.
Nelson RA, Cramb RA and Mamicpic MA, (1998) Erosion/productivity modelling of maize farming in the Philippine uplands. Part III: Economic analysis of alternative farming methods. Agricultural Systems 58:165-183.
Passioura JB (1996) Simulation models: science, snake oil, education, or engineering? Agronomy Journal 88: 690-694.
Qun H (1991) Economic evaluation of intercropping with paulownia – optimisation and choice of intercropping models. In: Zhaolua Z, Mantang C, Shiji W and Youxu J (eds) Agroforestry systems in China. pp. 163-173. The Chinese Academy of Forestry People’s Republic of China and International Development Research Centre, Canada.
Singh K, Rao BRR, Singh CP, Bhattacharya AK, Kaul PN, and Rajeswara Rao BR (1998) Production potential of aromatic crops in the alleys of Eucalyptus citriodora in
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 183
semi-arid tropical climate of south India. Journal of Medicinal and Aromatic Plant Sciences 20:749-752.
Swinkels RA and Scherr SJ (1991) (compilers) Economic analysis of agroforestry technologies: an annotated bibliography. International Council for Research in Agroforestry, Nairobi, Kenya. 215 pp
Thomas TH (1991) A spreadsheet approach to the economic modelling of agroforestry systems. Forest Ecology and Management 45: 207-235.
Thomas TH and Willis RW (1997) Linking bio-economics to biophysical agroforestry models. Agroforestry Forum 8(2): 40-42. van Noordwijk M and Lusiana B (1999). WaNuLCAS 1.0, a model of water, nutrient and light capture in agroforestry systems. Agroforestry Systems 45: 131-58. van Noordwijk M and Lusiana B (2000) WaNuLCAS 2.0 Background on a model of water, nutrient and light capture in agroforestry systems. International Centre for Research in Agroforestry (ICRAF), Bogor, Indonesia. van Noordwijk M and Lusiana B (2003) Welcome to the world of WaNuLCAS. A model of water nutrient and light capture in Agroforestry Systems. ICRAF South East Asia Programme, Bogor, Indonesia. (Accessed 19 January 2004). http://www.worldagroforestrycentre.org/sea/Products/AFModels/WaNulCAS/
Willis RW, Thomas TH and van Slycken J (1993) Poplar agroforestry: a re- evaluation of its economic potential on arable land in the United Kingdom. Forest Ecology and Management 57: 85-97.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 184
Table 1. The framework used to characterise computer models of silvoarable economics.
Model characteristic Examples of options
1. Background
1.1 First reference Date
1.2 Country of origin Country
1.3 Language English, Portuguese, French
1.4 Initial primary use Research, decision-support or educational
2. Systems modelled
2.1 Components of Agriculture, forestry, silvoarable, silvopastoral or system agrosilvopastoral
2.2 Number of Number of systems modelled simultaneously systems
3. Objectives of Comparison of two or more possible designs economic
analysis Feasibility: effect of combinations with specified resources
Optimisation: identification of ‘best’ design on basis of criteria
Prediction of actual behaviour
Uncertainty analysis
4. Viewpoint of Micro-economic (i.e. perspective of owner occupier or tenant) analysis
Macro-economic (i.e. perspective of society)
5. Spatial scale One-hectare: sub-field level analysis of a homogenous area
Field-scale: includes the effect of headlands
Farm-scale: fixed costs included
Catchment, regional, national or international-scale
6. Temporal scale
6.1 Time-step Hourly, daily or annual
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 185
6.2 Time period Maximum rotation considered
7. Generation and use of biophysical data
7.1 Biophysical- Stand-alone economic model economic links Bio-economic model with one- or two-way information flow
7.2 Nature of Empirical: single relationship models biophysical model Functional: use of several empirical relationships
Mechanistic: models which include growth processes
8 Platform and interface
8.1 Model platform Spreadsheet, programming language, graphical environment or database
8.2 Model interface Direct or a specifically-designed graphical user interface
9. Inputs and outputs
9.1 Inputs requirements Low, moderate or high
9.2 Databases Availability of linked databases
9.3 Outputs Biophysical, economic or both
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 186
Table 2. Characterisation of five computer models of silvoarable economics.
POPMOD ARBUSTRA Agroforestry WaNuLCAS Agroforestry Estate v 2.0 Calculator Model v 4.0
1. Background
First 1989 1995 1996 1996 1999 reference Country of origin UK France New Zealand Indonesia Australia
Langauge English French English English, English Indonesian, Portuguese
Initial primary Research Research Decision- Research Decision- use support support
2. Systems modelled
Systems Widely-spaced Forestry, Forestry, Forestry, Arable and poplar, arable arable and arable, arable and silvoarable and silvoarable livestock, silvoarable silvoarable silvoarable and silvopastoral
More than Yes Yes Yes No Yes one system
3. Objective of Comparison Comparison Comparison Comparison Comparison economic analysis and feasibility and feasibility
4. Economic Owner- Owner- Owner- Owner- Owner- viewpoint occupier occupier or occupier occupier occupier, tenant joint-ventures, leasehold or cutting rights share-cropping
5. Spatial scale
Scale One-hectare One-hectare, One-hectare One-hectare One-hectare field and farm and farm
Farm N/a Yes Yes N/a N/a heterogeneity
Multiple No Yes Yes No No planting
6. Temporal scale
Time-step Annual Annual Annual User-defined Annual
Time period 30 years 120 years User-defined User-defined 50 years
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 187
7. Generation and Empirical No biophysical Empirical Mechanistic No biophysical use of biophysical model, two- model, direct model, one- model, direct data way yield data for way model, one- yield data for information trees information way trees flow flow information flow
8. Platform/inte rface
Platform Spreadsheet Spreadsheet Visual Basic Graphical Spreadsheet development environment
Model Input into cells Graphical user Graphical user Graphical user Semi-graphical interface interface interface interface user interface
9. Inputs and outputs
Input Moderate High Moderate Very high Low requirements
Databases Tree data None Tree data Various Tree data
Outputs Mostly Mostly Biophysical Mostly Mostly economic economic and economic biophysical economic
Note: N/a = not applicable
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 188
Figure 1. Schematic representation of one-hectare, field, farm, and regional scales of modelling of silvoarable economics.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 189
Country, region, community or project scale A defined geographical area comprising more than one farm business.
Farm scale
An area managed as one business. Fixed costs and a range of enterprises and land types are typically
Field scale
The effect of headlands and boundaries may be included and the field may be given a specific
One-hectare scale
Homogenous area with economic data typically based on gross or net margins per hectare
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 190
Figure 2. Three possible types of information flow within computer models of silvoarable economics.
Stand alone Bio-economic model Bio-economic model economic model with one-way with two-way information flow information flow
Biophysical Biophysical module module
Economic Economic Economic model module module
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 191
ANNEX 8. Fine root distribution in Dehesas of Central- Western Spain
G. Moreno1, J.J. Obrador1, E. Cubera1, C. Dupraz2
1 I.T.Forestal, Centro Universitario, UEX. Plasencia 10600. Cáceres. Spain
2 INRA, UMR-SYSTEM, 2 Place Viala. 34060 MONTPELLIER Cedex, France.
1Corresponding author:
Gerardo MORENO MARCOS
I.T. Forestal, Universidad de Extremadura,
Avda. Virgen del Puerto, Plasencia, 10600 (CÁCERES, Spain)
Phone: + 34.927427000 Fax: + 34.927425209 Email: [email protected]
Running Title: Rooting system in dehesas
Nº of text pages: 17
Nº of tables: 1
Nº of figures: 5 (+ 1 page with figure captions)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 192
Fine root distribution in Dehesas of Central-Western Spain G Moreno1*, JJ Obrador1, E Cubera1, C Dupraz2 6 I.T.Forestal, Centro Universitario, UEX. Plasencia 10600. Cáceres. Spain 2 INRA, UMR-SYSTEM, 2 Place Viala. 34060 MONTPELLIER Cedex, France.
ABSTRACT
A Dehesa is a structurally complex agro-silvo-pastoral system where at least two strata of vegetation, trees and herbaceous plants coexist. We studied the root distribution of trees (Quercus ilex L.) and herbaceous plants, in order to evaluate tree and crops competition and complementarity in Dehesas of Central Western Spain. 72 soil cores of 10 cm diameter (one to two metre deep) were taken out around 13 trees. Seven trees were intercropped with Avena sativa L. and six trees were in a grazed pasture dominated by native grasses. Soil coring was performed at four distances from the tree trunks, from 2.5 (beneath canopy) till 20 m (out of the canopy). Root length density (RLD) of herbaceous plants and trees was measured using the soil core-break method. Additionally, we mapped tree roots in 51 profiles of 7 recently opened road cuts, located between 4 and 26 m of distance from the nearest tree. The depth of the road cuts varied between 2.5 and 5.5 m. Herbaceous plant roots were located mostly in the upper 30 cm, above a clayey, dense soil layer. RLD of herbaceous plants decreased exponentially with depth until 100 cm depth. Holm-oak showed a much lower RLD than herbs (on average 2.4 versus 23.7 km.m-3, respectively, in the first 10 cm of the soil depth). Tree RLD was surprisingly almost uniform with depth and distance to trees. We estimated a 5.2 m maximum depth and a 33 m maximum horizontal extension for tree roots. The huge surface of soil explored by tree roots (around 7 times the projection of the canopy) could allow trees to meet their water needs during the dry Mediterranean summers. The limited vertical overlap of the two root profiles suggests that competition for soil resources between trees and the herbaceous understorey in the Dehesa is probably not as strong as usually assumed.
Keywords: Agroforestry, grasses, open woodland, Quercus ilex, root length density, rooting system, core-break.
6 * Correspondence to: [email protected]
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 193
INTRODUCTION
Dehesas are multi-purpose open woodlands where at least two strata of vegetation coexist. They have been common in the Iberian Peninsula, at least since the middle ages (Montero et al., 1998). At present, Dehesas cover 3.1 million hectares in western Spain and Portugal, and they are considered as habitats to be preserved because of the high biological diversity they support (Díaz et al., 1997). However, in the last decades, a significant decrease in extension and tree density has been occurring as a consequence of increased mechanisation, changes in land use and death of trees in over-aged stands (Plieninger et al., 2003). A better knowledge of the role of trees on dehesa functioning and sustainability could contribute to improve its management and conservation.
Some pioneer studies on the effect of trees in dehesa functioning have shown the positive effects of trees on soil nutrient contents (Escudero, 1985), soil water storage capacity (Joffre and Rambal, 1988), water stress for the underlying herbaceous stratum (Joffre and Rambal, 1993), and pasture production (Puerto et al., 1987). Several authors have also shown the positive effect of tree clearance on the remaining trees physiological status (e.g., Infante et al. 1999 and Montero et al., 2004) and productivity (e.g., Diaz et al., 1997). The improved physiological status of the dehesa trees could be due to an increase in the available soil volume, and thus water and nutrients, for each individual tree. Joffre et al. (1999) stressed the need for better knowledge of the extension of the tree root system to understand the implications of soil water balance on the stability of the dehesa (tree-tree and tree- understorey interactions), and therefore to predict the consequences of long-term climatic and land use changes.
Much of the competition among plants takes place underground (Casper and Jackson 1997), and below-ground competition knowledge is a major difficulty for understanding simultaneous agroforestry systems (van Noordwijk et al., 1996). The use of a higher proportion of below ground resources can be achieved if deep networks of tree roots are able to capture water or nutrients draining or leaching below the rooting zone of the crops (van Noordwijk et al., 1996).
Understanding and predicting ecosystem functioning (e.g. nutrient cycling, carbon and water fluxes) requires an accurate assessment of plant rooting distribution (Jackson et al., 1996). A realistic map of root length density, both horizontal and vertical, is needed to model possible interactions (facilitation, competition and complementarity) between plants. However, many models (e.g. HyPAR: Mobbs et al., 1999; WaNuLCAS: Van Noordwijk and Lusiana, 2000) assume a simple shape of the tree rooting system: an exponential decrease with soil depth and distance from the tree trunk is often considered. This simplicity is explained by the difficulty and complexity of root studies, which have resulted in a lack of quantitative information on rooting systems (Smith et al., 1999; Jose et al., 2001).
In the last decades some new methods have been proposed to describe the root system: soil windows, minirhizotrons, soil cores washing, (Smit et al., 2000). However, all them have any limitation in term of collecting representative samples and of cost or time required (Smit et al., 2000). To address the issue of representativeness, some authors applied a combination of methods, e.g. trenches plus soil cores (e.g. Jones et al., 1998, Silva and Rego, 2003), in order to
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 194
characterise the horizontal and/or vertical root extension and to determine the root density. Additionally, to reduce the labour needs and speed up the process, van Noordwijk et al. (2000) proposed an indirect method to estimate the root length density by counting roots emerging from the horizontal planes of broken soil cores.
To our knowledge, only two studies have been carried out to characterise the root systems in dehesas (Barrera et al., 1987; Joffre et al., 1987), and both were limited to natural grasses roots, in the first 30 and 60 cm of soil, respectively. None of these two works dealt with the tree root distribution. The present study focus on the root distribution (fine root length density) of both tree (holm-oak) and herbaceous vegetation (cereal crop or natural grasses) considering both vertical and horizontal dimensions. Additionally, we have documented the effect of soil tillage on the root density of the trees. To achieve that we used two different methods, based on the study of soil cores and on the record of root maps in recently opened road cuts.
MATERIAL AND METHODS
Study Area
The study has been carried out in two dehesas (Cierra lobata? CL and xxx ST) of C- W Spain (39º 41’ N - 6º 13’ W; altitude: 380 m.a.s.l.), with an average tree density of 35 tree ha-1 (Quercus ilex L in both grazed and cropped plots. Average tree dimensions were 44.9 cm for DBH (diameter at breast height), 10.4 m for canopy width and 7.8 m for tree height. In grazed plots the main grasses were Lolium rigidum Gaudin, Plantago lanceolata L., Erodium sp L., Taraxacum obovatum (Willd.) DC. and Echium plantagineum L. These species were also abundant as weeds in the intercropped (oats) plots given that herbicides are not applied in dehesas crops. The aboveground biomass of the intercrop was low : 5.2 and 2.9 Mg of dry weight per ha in CL and ST respectively. The difference reflected the difference of the fertilisation schemes : 200 and 50 kg of NPK ha-1 in CL and ST farms, respectively.
The climate is semi-arid Mediterranean, with an annual rainfall of 579 mm, mean annual temperature of 16.2 ºC, and mean annual potential evapotranspiration of 864 mm. Climate is classified as subtropical Mediterranean, following the Papadakis classification (1966), with dry, warm and cold (with frost) periods of 4, 3 and 5 months, respectively. Soils are chromic Luvisols (FAO, 1998) in both farms, developed over tertiary sediments with abundant gravels and stones of quartzite, one or several very red argic horizons, with slight brown and silty-sand texture in the surface horizon and a very sandy layer in depth (below 100 cm). Soils also showed poor internal drainage, resulting in variegated colours and/or pseudo-gleic, low soil chemical fertility, and occasional CaCO3 accumulation between 1.5 and 2 m depth.
Soil cores: Root Length Density
Soil cores were taken with a stainless steel soil column cylinder with a cutting shoe and a removable cover (diameter 10 cm, length 1000 mm), inserted into the soil with a heavy electrical powered percussion hammer (Makita HM 1800, provided by Eijkelkamp, Giesbeek, The Netherlands). Between 23rd - 28th April 2002 thirty-six soil cores of 1 m-depth were extracted at 2, 5, 8 and 12 m distance from the tree trunk of 4 intercropped holm-oaks, at two orientations (three orientations in one tree). Between 10th - 20th May 2003, thirty-six soil cores of 2 m-depth (at maximum) were taken at 2, 5, 10 and 20 m of distance from the tree trunk of 3 intercropped holm-
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 195
oaks and 6 ‘intergrazed’ holm-oaks. Soil cores were covered by a gutter (two halves of a PVC tube) and transparent plastic to avoid damage during transport to the laboratory, where they were stored at 6°C in a cold chamber to keep the roots fresh until analysis.
Cylindrical soil cores were divided into 10 cm-length samples. Each sample was then broken by hand in two parts, and the number of tree and herbaceous plants fine roots (using a 2 mm diameter threshold for tree fine roots) were recorded in both sides of the sample parts. Decayed tree roots were excluded. Holm-oak roots were identified by their black cork, while grasses roots were white. Very new tree roots (growing tips) are also white, but much thicker than crop/grass roots. Differentiating crop and weed roots was not possible.
The method of soil core-break allowed us to estimate the Root Length Density (RLD) from the number of roots sticking out of two soil surfaces of a horizontally broken soil core (Van Noordwijk et al., 2000). A total of 65 randomly selected samples were washed each year to provide a direct calibration of root length versus counts (Van Noordwijk et al., 2000). To wash samples, different filters between 2 and 0.125 mm mesh size were used. This was done to avoid losing fine roots. All samples had tree fine roots, and only 46 had herbaceous roots. Roots obtained from the washing activity were laid on plastic paper and then photocopied and the length of the fine roots was measured manually for each soil core. Data are expressed as Root Length Density (km m-3 of soil) because root length is a better indicator of root system functions in terms of uptake of water and nutrients than root weight and root number (Jones et al., 1998).
Road Cuts: Maximum tree rooting depth and horizontal spread
We took advantage of current works in the road that cross the study area, to check the maximum distance and depth of the holm oak rooting system. Road cuts had been recently (3-4 months) opened. We counted the number of emerging tree roots (fine and coarse). Roots were counted in 51 profiles, located every 5 meters in 7 different road cuts. To identify the position of the root (depth and distance from the nearest tree), we used a metallic square of 50 cm size, divided into small squares of 100 cm2. The maximum depth of the profiles varied between 250 and 550 cm. The slope of the road cuts was measured to calculate the actual depth of roots.
We were able to confirm (by means of recent aerial photographs) in four of the road cuts that no trees had been removed during the works. In these cases we measured the distance to the nearest tree. For the other three road cuts we confirmed the previous existence of some trees where the road cuts had been created and data were only used to describe the root distribution in depth. We therefore have 51 profiles for the description of vertical root distribution, and only 34 for horizontal root distribution. Results are expressed as number of roots per m2. Schenk and Jackson (2002a) have shown that patterns of rooting profiles based on root length and root count do not differ. Thus, results of both methods (soil cores and road cuts) could be compared. Data Analysis
A simple linear regression was found to be fit for calibrating the soil core-break method (relationship between RLD and the number of counted roots), as suggested
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 196
by van Noordwijk et al., 2000. Root density has been regressed with depth (linear and non-linear regressions) or with both depth and distance (multiple regression) in order to describe rooting patterns of both trees and herbaceous plants.
Differences between treatments and position in RLD were assessed by analysis of variance (ANOVA). Two-way ANOVAs were applied to detect differences in mean values of RLD (dependent variable) between distance and depth (as independent variables) for both herbaceous plants and trees. Two-way ANOVAS were also applied to contrast the effect of soil management (cropped versus grazed) on RLD at different distances or depths. Results are expressed as F values (and degree of freedom) and significance level (p).
Depth of 50% and 95% cumulative root density (d50 and d95, respectively) were calculated according to the Gale and Grigal (1987) model. Following these authors, a d non-linear regression was used to fit the function fc = 1 – β to the profile of cumulative root fraction (fc), from the soil surface to depth d (cm). β is the fitted
“extinction coefficient”. Values of d50 and d95 were then calculated from d50 = Ln (0.5)
/ Ln (β) and from d95 = Ln (0.05) / Ln (β), respectively.
RESULTS
Linear relationships for root length density estimation
The calibration curves between Nroot and RLD were determined separately for herbaceous and tree fine roots (Figure 1). The relationships was better for herbaceous roots than for tree roots (R2 = 0.85 vs 0.42, respectively). This difference is partly explained because the range of Nroot and RLD was much lower for tree (0- 3800 roots m-2 and 0-8 km m-3) than for herbs (0-28000 roots m-2 and 0-44 km m-3). A further explanation is probably linked to the patchy pattern of tree roots : tree roots are less evenly distributed in the soil core sample volume, resulting in a less accurate prediction from the core-break count. Both regressions were however highly significant, with p < 0.001 (n = 46 and 65 for herbs and trees, respectively).
Vertical profiles of root length density Figure Herbaceous RLD was very high in the first cm of the soil (Fig. 2), decreasing very1 and -0.999 2 sharply, exponentially with depth: RLDkm.m-3 = 122.1 * Depthcm (R = 97.0; F1, 218 = 714; p <0.00001). At 40 cm depth, RLD was ten times lower than in the first 10 cm. The depth of 50% cumulative root length (d50) was found at 10.7 cm (Fig. 2 and Table 1). Results of a two-way ANOVA (depth and distance as independent variables) showed significant differences between consecutive depths till 60 cm (F9, 442 = 94.46; p < 0.001); then, differences were not significant. Below 90 cm, herbaceous plant roots were only found very occasionally, reaching a maximum rooting depth of 100 cm.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 197
By contrast, a linear but non-significant decrease in tree RLD was observed from 0 till 200 cm of soil (F9, 442 = 0.36; p < 0.952). At 2 metres depth the holm-oak RLD was still about half with respect to the uppermost soil layer (Fig. 2). From the regression 2 between tree RLD and depth (RLDkm.m-3 = 2.24 – 0.0056 x Depthcm; R = 0.56; F1, 18 = 22.9; p < 0.00015), the expected maximum rooting depth for holm-oak in this area was estimated at 400 cm, and the d50 value at 96.4 cm (Fig.2 and Table 1).
Lateral root distribution
RLD varied significantly with distance to the tree trunk, for both trees and herbaceousTable 1 plants (Figure 3a). Herbaceous plants RLD was significantly higher at 10 and 20 m of distance than at 2.5 and 5 m of distance (F3, 442 = 10.26; p < 0.0004). Tree RLD decreased smoothly with the distance to the tree. Significant differences were detected between 2.5 and 5m and between 5 and 10m (F3, 442 = 7.64; p < 0.001).
In spite of the differences with distance, the RLD profile shape did not vary with distance, neither for herbaceous plants nor for holm-oak (Fig. 3b). In fact, we did not find any significant interaction between both factors (Distance x Depth) either for herbaceous plants (F27, 442 = 1.16; p < 0.289) or for trees (F27, 442 = 0.80; p < 0.702). Only a slight increase in d50 value with the distance for herbaceous plant roots has been estimated (9.6, 9.3, 10.6 and 14.4 at 2, 5 10 and 20 m of distance); the opposite was observed for tree roots (67, 69, 57 and 55 cm at 2, 5 10 and 20 m of distance, considering only the roots of the first 200 cm of soil).
Effect of soil management on root distribution
Two main differences have been observed in the root distribution when two different Figure 3a soil management types (cropped or grazed) are compared (Fig. 4a). Root length and 3b density of herbaceous plants was much lower in grazed plots (native grasses) than in intercropped plots (oats + weeds) (F1, 249 = 5.55; p = 0.004), at any distance (not- significant interaction: F6, 249 = 0.062; p = 0.991). d50 value was also clearly deeper in intercropped plots (cm) than in grazed plots (13.9 and 7.4 cm, respectively; Table 1). The profile of RLD was similar for both cropped and grazed plots (Fig. 4b).
Tree RLD was however very similar in both types of management, cropped and grazed plots (Fig. 4a). The only difference was observed in the top soil, where intercropped trees had a lower RLD than grazed trees (Fig. 4b), although the differences were not statistically significant (F1, 104 = 1.18; p = 0.31). d50 value was only slightly deeper in intercropped plots (cm) than in grazed plots (69.1 and 64.6 cm, respectively, considering only the two first metres of depth).
Tree roots in road cuts Figure 4 This study confirmed results obtained with soil cores. The maximum depth where Table 2 roots were found was 450 cm, with a d50 value of 81.2 cm vs a d50 value of 96.4 found in the soil core study (Table 1).
A significant decrease in the number of tree roots with distance was also found (F3, 27 = 96.81; p = 3.0E-05). Significant differences were found between 0-10 m and 10-15 m, between 10-15 m and 15-20 m, but not between 15-20 m and >20 m. The maximum measured distance was 26 m, where roots were found even at 3 m of depth (data not shown).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 198
Considering the relationship between the number of roots and both distance and depth, the following equation was found: Nroot (m2) = 76.2– 0.14* Depth(cm) – 2 2.23*Distance(m); with R = 0.56 (F2, 155 = 36.24; p < 0.0001). According to the stepwise analysis, the amount of variability explained by both parameters (their contribution to R2) was 0.39 % and 0.18 %, for distance and depth, respectively. From this equation, 33 m was estimated as the maximum distance, and 520 cm as the maximum depth.
DISCUSSION
Herbaceous plants root system
Most of the herbaceous plant roots were located in the first centimetres of the soil, a common pattern for most of the herbaceous plants in the world (Jackson et al., 1996). Native grasses showed a very shallow root system, with d50 at 7.4 cm, and with 94% of the root length in the first 30 cm of soil. The maximum rooting depth was evidenced at about 80 cm. The oat crop had a deeper rooting pattern than native grasses, with ‘only’ 78% of the root length in the first 30 cm, and with a maximum rooting depth of 100 cm. Other authors have reported deeper root systems for temperate grassland and crops than those found in this study (e.g., Canadell et al., 1996 and Jackson et al., 1996; see Table 1). The apparent low capacity of oats and native grasses to go deep in our study area could be explained by the presence of a very clayey soil layer between 40-80 cm depth. Nevertheless, a very shallow root system for native grasses of dehesas of Quercus ilex has also been reported by Barrera et al. (1987) and Joffre et al. (1987).
The highest RLD of cropped plants as compared to natural pasture grasses was a surprise. This result does not coincide with the reported values by Jackson et al. (1996), who found that crops showed very low root density when compared to most other biomes (even a tenth part respect to temperate grassland). The denser and deeper root system in oat crops with respect to native grasses could induce an additional competition for soil resources (mainly water) with trees when compared to the pasture.
Holm-oak root profiles
The root profiles of mature holm-oaks were surprisingly almost uniform with depth. Most reported tree root profiles feature a significant decrease with depth, and this decrease often follows an exponential negative pattern (Jackson et al., 1996). Nevertheless, a linear or quasi-linear decrease of root density has been also reported by few authors (e.g., Kummerov and Mangan, 1981 and Schulze et al., 1996).
As a consequence, we put in evidence an unusual deep rooting system for Quercus ilex in this dehesa study, with a d50 value between 96,4 cm and 81,3 cm, and with a d95 value between 417 cm and 351 cm (with soil cores and road cuts methods respectively). Schenk and Jackson (2002a), after an exhaustive review of 475 root studies, concluded that most of the plants have at least 50% of the roots in the first 30 cm of the soil, even in the desert. For sclerophyllous Mediterranean plants, these authors reported mean values of 19 and 171 cm for d50 and d95 (Table 2). Rather shallow root systems, with most of the roots in the first 50 cm of the soil depth, have also been reported in the Iberian Peninsula for forests of Quercus ilex (Canadell and Rodá, 1991 and López et al., 2001), for Quercus coccifera (Cañellas and San Miguel,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 199
2000), for a sand dune shrub community (Martínez et al., 1998), and for Erica and Ulex species (Silva and Rego, 2003) .
A deep-root pattern is often found in water-limited situations, mainly for species with taproots in desert, savannah, tropical evergreen forest and sclerophyllous shrubland and forest (Canadell et al., 1996). We have found a maximum rooting depth for holm- oak at around 5 m, meanwhile Canadell et al. (1996) and Canadell and Rodá (1991) reported maximum rooting depth of only 3.7 and 1 m, respectively, for close forests of Quercus ilex. Regarding other evergreen Quercus species there are several references with maximum rooting depth between 5 and 10 m (see Canadell et al., 1996). These authors reported a mean value of 5.2 m for sclerophyllous shrubland and forest of the world.
Lateral root distribution
In semiarid conditions, the survival of trees facing severe drought conditions is only possible if the tree root system can extend beyond the influence of the tree canopy (Joffre and Rambal, 1999). Lateral root spread influences how many neighbours compete for resources available to plants in an ecosystem (Schenk and Jackson, 2002b). In the present study, the maximum lateral rooting (estimated at 33 m) was slightly larger than the average distance between trees (26 m). This pattern may be common in semiarid open woodland. Schenk and Jackson (2002b) pointed out that larger lateral root spreads were found in plants growing at low density in dry environments, where plants can explore the soil in interspaces between plants. These authors reported several cases of trees with maximum lateral root spread above 20 m.
An outstanding consequence of this result is that lateral roots can explore the whole inter-tree space, allowing full use of the soil volume by mature trees in dehesas. The surface of explored soil by roots was around 7 times the projected area of the canopy. The dynamics of soil water content in the same plots (Cubera et al., 2004) showed that soil water beyond the tree canopy was depleted throughout the summer, while no grasses were active, confirming that water was extracted by trees. Our results support thus the hypothesis that mature tree density in dehesas could be water-availability dependent (Joffre and Rambal, 1999).
Combined root system: implication on competition for soil resources
To reduce competition with crops/grasses for below-ground resources, tree should have a deep root system and little root proliferation near the top of the profile, thereby enabling the herbaceous plants to utilise resources from near the soil surface, while trees have sole access to deeper layers (Schroth 1995). We have shown such a pattern of spatial separation between herbaceous plants and tree root systems. Trees had a much deeper root system, with a rather low RLD in the upper layers of the soil, and herbaceous vegetation did not reach deep layers, where tree roots were still abundant.
This rooting pattern contributes to reducing below-ground competition, thereby probably falling into the general category of ‘niche separation’ (Casper and Jackson, 1997). Thus, although water limitation is an important feature in most dehesas (including our study area), this does not necessarily mean that competition for water is high. Many authors have shown that woody plants took up more water from deeper layers than herbaceous ones (e.g. Ehleringer et al., 1991 and Sala et al., 1989) avoiding thus a direct competition. In fact, this two-layer model appears to be most appropriate in drier regimes and in systems with
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 200
substantial winter precipitation (Schenk and Jackson, 2002b), as it is the case of Iberian dehesas. The possible existence of this water partitioning in dehesas should be addressed in future research, and is only relevant when the rainfall pattern allows deep soil layer to be systematically refilled during the cold season.
The very high herbaceous RLD in the first cm of soil could induce a strong competition for nutrients with trees, as a result of the fact that nutrients (mainly N) may be available only in the upper soil layers (Jackson et al., 1996). Finally, this extensive and deep rooting system of the trees may indicate a very good capability to avoid any nitrate leaching from the cropped area. Most leached nitrates can be captured by the tree roots, as the evergreen oak trees are active throughout the year.
CONCLUSIONS
If we follow the assumption that roots grow only as deeply and as far as needed to fulfil the plant resource requirements (Schenk and Jackson, 2002b), it is obvious that mature holm-oaks need a huge volume of soil to capture below-ground resources in oligotrophic soils, under a semi-arid climate with a long summer drought. As Joffre et al. (1999) have pointed out, dehesas have to cope with the high variability of the Mediterranean climate; the tree extensive root system undoubtedly must contribute both in adapting to natural conditions and in overcoming unpredictability.
Holm-oak RLD decreased very slowly with distance and depth. This rooting pattern should have important consequences in modelling the coexistence of tree and grass/crop. It is at odds with the commonly assumed pattern of tree fine roots distributions, that is often described with an exponential decrease with soil depth and distance from the tree trunk.
We have found a limited vertical overlap of tree and herbaceous understorey root systems, and this feature is probably a key to the stability and productivity of this agro-silvo-pastoral system. However, we documented the rooting pattern only in spring, and more information is still required on temporal dynamics of the fine roots of both holm-oak and herbaceous plants in the dehesa. Acknowledgements This study was supported by the European Union (SAFE project : Silvoarable Agroforestry For Europe QLK5-CT-2001-0560), by the Spanish Ministerio de Ciencia y Tecnología (MICASA project) and by the Consejería de Educación de Extremadura (CASA project). Elena Cubera has been awarded a grant by the Consejería de Educación de la Junta de Extremadura (Spain) and Jesús Obrador has been awarded a grant by ANUIES (México).
REFERENCES
Barrera I, Galindo P and Gómez J M 1987. Modelo de distribución de la biomasa radical en función de la profundidad. Anuario del CEBA de Salamanca 12, 313 - 323.
Canadell J, Jackson R B, Ehleringer J R, Mooney H A, Sala O E and Schulze E-D 1996 Maximum rooting depth of vegetation types at the global scale. Oecologia 108, 583- 595.
Canadell J and Rodá F 1991 Root biomass of Quercus ilex in a montane Mediterranean forest. Can. J. For. Res. 21, 1771-1778.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 201
Cañellas I and San Miguel A 2000 Biomass of root and shoot systems of Quercus coccifera shrublands in Eastern Spain. Ann. For. Sci. 57, 803-810.
Casper B B and Jackson B J (1997) Plant competition underground. Annu. Rev. Ecol. Syst. 28, 545 - 570.
Cubera E, Montero M J and Moreno G 2004 Effect of land use on soil water dynamics in dehesas of Central-Western Spain. In Advances in GeoEcology 37: Sustainability of Agrosilvopastoral systems –Dehesas, Montados-. Eds. S Schnabel and A Ferreira. pp. 109-123. Catena Verlag, Reiskirchen.
Díaz M., Campos P., Pulido F.J. 1997. The Spanish dehesas: a diversity in land-use
and wildlife. In: Pain D.J. and Pienkowski M.W. (eds) Farming and Birds in
Europe: The common Agricultural Policy and its Implications for Bird Conservation.
Academic press, Cambridge, UK, pp 178-209.
Ehleringer J R, Phillips S L, Schuste W F S and Sandquist D R 1991 Differential utilization of summer rains by desert plants: implications for competition and climate change. Oecologia 88, 430-434.
Escudero A 1985 Efectos de árboles aislados sobre las propiedades químicas del suelo. Rev. Ecol. Biol. Sol 22(2), 149-159.
FAO 1998 World reference base for soil resources. FAO, ISRIC and ISCC. Rome. pp. 109.
Gale M R and Grigal D F 1987 Vertical root distributions of northern tree species in relation to successional status. Can. J. For. Res. 17, 829-834.
Infante J M, Damesin C, Rambal S and Fernández-Alés R 1999 Modelling leaf gas exchange in holm-oak trees in southern Spain. Agr. For. Meteorol. 95, 203-223.
Jackson RB, Canadell J, Ehleringer J R, Mooney H A, Sala O E and Schulze E.-D 1996 A global analysis of root distributions for terrestrial biomes. Oecologia 108, 389-411.
Joffre R, Leiva Morales M J, Rambal S and Fernández Alés R 1987 Dynamique racinaire et extraction de l’eau du sol par des graminées pérennes et annuelles méditerranéennes. Acta Oecol. Oecol. Plant. 8(22), 181-194.
Joffre R and Rambal S 1988 Soil water improvement by trees in the rangelands of southern Spain. Oecol. Plant. 9, 405-422.
Joffre R and Rambal S 1993 How tree cover influences the water balance of Mediterranean rangelands. Ecology 74: 570-582.
Joffre R, Rambal S and Ratte J P 1999 The dehesa system of southern Spain and Portugal as a natural ecosystem mimic. Agroforest. Syst., 45, 57-79.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 202
Jose S, Gillespie A R, Seifert J R and Pope P E, 2001. Comparison of minirhizotron and soil core methods for quantifying root biomass in a temperate alley cropping system. Agroforest. Syst. 52, 161-168.
Jones M, Sinclair F L and Grime V L 1998 Effect of tree species and crown pruning on root length and soil water content in semi-arid agroforestry. Plant Soil 201, 197-207.
Kummerov J and Mangan R 1981 Root system in Quercus dumosa Nutt. dominated chaparral in southern California. Acta Oecol. 2, 177-188.
López B, Sabaté S and Gracia C 2001 Vertical distribution of fine root density, length density, area index and mean diameter in a Quercus ilex forest. Tree Physiol. 21, 555-560.
Martínez F, Merino O, Martín A, García Martín D and Merino J 1998 Belowground structure and production in a Mediterranean sand dune community. Plant Soil 201, 209-216.
Mobbs D C, Lawson G J, Friend A D, Crout N M J, Arah J R M, Hodnet M G 1999 Hypar Model for agroforestry system. Technical Manual. DFID Forestry Research Programme.
Montero G, San Miguel A and Cañellas I 1998 System of Mediterranean Silviculture “La Dehesa”. In Agricultura Sostenible. Eds. R M Jiménez Díaz and J Lamo de Espinos. pp 519-554. Mundi Prensa, Madrid.
Montero M J, Obrador J J, Cubera E and Moreno G 2004 The role of dehesa land use on tree water status in Central-Western Spain. In Advances in GeoEcology 37: Sustainability of Agrosilvopastoral systems –Dehesas, Montados-. Eds. S Schnabel and A Ferreira. pp. 125-136. Catena Verlag, Reiskirchen.
Papadakis J 1966 Climates of the World and Their Agricultural Potentialities. Ed. J Papadakis. Buenos Aires.
Plieninger T, Pulido F J and Konold W 2003 Effects of land-history on size structure of holm oak stands in Spanish dehesas: implications for conservation and restoration. Environ. Conerv. 30, 61-70.
Puerto A, García J A and García A. 1987. El sistema de ladera como elemento esclarecedor de algunos efectos del arbolado sobre el pasto. Anuario del CEBA de Salamanca 12, 297 - 312.
Sala O E, Golluscio R A, Laurenroth W K and Soriano A 1989 Resource partitioning between shrubs and grasses in the Patagonian steppe. Oecolgia 81, 501-505.
Schenk H J and Jackson R B 2002a The global biogeography of roots. Ecological Monographs 72(3), 311-328.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 203
Schenk H J and Jackson R B 2002b Rooting depths, lateral root spreads and below- ground/above-ground allometries of plants in water-limited ecosystems. J. Ecol. 90, 480-494.
Schroth G 1995 Tree root characteristics as criteria for species selection and systems design in agroforestry. Agrofor. Syst. 30, 125-143.
Schulze E-D, Mooney H A, Sala O E, Jobbagy E, Buchmann N, Bauer G, Canadell J, Jackson R B, Loreti J, Oesterheld M, and Ehleringer J R 1996 Rooting depth, water availability, and vegetation cover along an aridity gradient in Patagonia. Oecologia 108, 503-5111.
Silva J S and Rego F C 2003 Root distribution of a Mediterranean shrubland in Portugal. Plant Soil 255, 529-540.
Smith D M, Jackson N A, Roberts J M and Ong C K 1999 Root distribution in a Grevillea robusta-maize agroforestry system in semi-arid Kenya. Plant Soil 211, 191-205.
Smit A L, Bengough A G, Engels C, van Noordwijk M, Pellerin S and van de Geijn 2000 Root Methods: A Handbook. Springer Verlag. Berlin Heidelberg. pp. 510.
Van Noordwijk M, Brouwer G, Meijboom F, Do Rosario G, Oliveira M and Bengough A G, 2000. Trench Profile Techniques and Core Break Methods. In Root Methods: A Handbook. Eds. L Smit, A G Bengough, C Engels, M van Noordwijk, S Pellerin and van de Geijn. pp. 211-233. Springer-Verlag, Berlin Heidelberg.
Van Noordwijk M, Lawson G, Soumare A, Groot JJR and Hairiah A 1996 Root distribution of trees and crops: competition and/or complementarity. In Tree-crop interactions: A Physiological approach. Eds. C K Ong and P Huxley. Pp. 319-364. CAB International, Wallingford.
Van Noordwijk M and Lusiana B 2000 WaNuLCAS 2.0, Background on a model of water nutrient and light capture in agroforestry systems. International Centre for Research in Agroforestry (ICRAF), Bogor, Indonesia. 186 pp.
LEGEND of FIGURES
Figure 1. Linear regressions between the number of roots crossing a horizontal plane (Nroot) and root length density (RLD) for (a) herbaceous plants (oats and native grasses), and (b) Holm-oak (only fine-roots). Figure 2. Variation of root length densities with soil depth for holm-oak and herbaceous plants in dehesas developed over chromic Luvisols in CW Spain. The inset shows the cumulative fractional root distribution plotted against the soil depth. Figure 3. (A) Mean values of RLD of holm-oak and herbaceous plants (oats and native grasses) measured at different distances to the tree trunk in dehesas. (B) Distribution of RLD plotted against the distance to the tree trunk and the depth for both herbaceous plants and tree. Figure 4. Distribution of the tree (holm-oak) and herbaceous plants (oats and native grasses) RLD at differences distances (A) and depth (B) under two different types of dehesa management: cropped and grazed. Title: Fine Root distribution in dehesas of Central-Western Spain
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 204
Authors: G Moreno, JJ Obrador, E Cubera, C Dupraz FIGURE 1 50 RLD = 0.0012 Nh Herbaceous 2 40 R = 0.85 Plants ) -3 30
20 RLD (km m
10 A 0 0 5000 10000 15000 20000 25000 30000
Nroot (m-2)
10 RLD = 0.0015 Nh Holm-oak 8 R2 = 0.42 ) -3 6
4 RLD (km m
2 B 0 0 1000 2000 3000 4000 Nroot (m-2 )
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 205
Title: Fine Root distribution in dehesas of Central-Western Spain
Authors: G Moreno, JJ Obrador, E Cubera, C Dupraz FIGURE 2
Root lenght density, km m-3 0 5 10 15 20 25 Cumulative root fraction 0 0,0 0,2 0,4 0,6 0,8 1,0 0
25 50 50 100 150 75 200 Depth, cm 250
Depth,cm 100 300
350 125 400 150 175 200 Herbaceous plants Holm oak
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 206
Title: Fine Root distribution in dehesas of Central-Western Spain
Authors: G Moreno, JJ Obrador, E Cubera, C Dupraz FIGURE 3
8 Herbaceous plants A
) 6 Holm-oak -3 m
4
RLD (km . 2
0 2,5m 5m 10m 20m Distance to tree trunk
Distance, m 0 5 10 15 20 25 0
25
50
75
100 Depth, cm 125
150
175 B 200 RLD of Herbaceous plants RLD of Holm-oak
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 207
Title: Fine Root distribution in dehesas of Central-Western Spain
Authors: G Moreno, JJ Obrador, E Cubera, C Dupraz FIGURE 4
Herbaceous plants (Cropped) Herbaceous plants (Grazed) Holm-oak (Cropped) Holm-oak (Grazed) 16
-3 14 12 10 8 6 4
Root lenght density, km m 2 A 0 0 5 10 15 20 25 Distance
Root lenght density, km m-3 0 5 10 15 20 25 30 35 0
50
100 Depth, cm Depth, Herbaceous plants (Cropped) Herbaceous plants (Grazed) 150 Holm-oak (Cropped) Holm-oak (Grazed) B
200
Title: Fine Root distribution in dehesas of Central-Western Spain
Authors: G Moreno, JJ Obrador, E Cubera, C Dupraz
TABLE 1
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 208
Maximum d # d # Vegetation type β R2 50 95 rooting (m) (m) depth (m) Herbaceous plants 0.937 96.0 0.11 0.46 1 Oats + weeds 0.951 97.4 0.14 0.60 1 Native grasses (mostly 0.911 98.8 0.07 0.32 8 annual) Holm-oak (soil cores)* 0.993 98.9 0.96 4.16 4 Holm-oak (road cuts) 0.992 99.5 0.81 3.51 4.5 Mediterranean woody plants 0.19(a) 1.71(a) 5.2(c) Temperate grassland 0.943(b) 82.0 0.12(b) 0.51(b) 2.6(c) Crops 0.961(b) 94.3 0.17(b) 0.75(b) 2.1(c) # d50 and d95 indicate the depth in cm corresponding to 50 and 95%, respectively, of the cumulative root fraction. Both values are estimate from the Gage and Grigal (1987) model: Y = 1-βd, where Y is the cumulative root fraction from the surface to depth d (cm), and β is the fitted “extinction coefficient”.
* The equation RLDkm.m-3 = 2.24 – 0.0056 x Depthcm (see text) was applied to get values of RLD from 200 to 400 cm depth.
Table 1. Comparison of the root profiles of holm-oak and herbaceous vegetation in dehesas of Central-Western Spain with average values from recent comprehensive reviews. (a) Schenk and Jackson (2002a) averaged from 475 root studies. (b) Jackson et al. (1996) averaged data from many different species from all over the world. (c) Canadell et al. (1996) data averaged data from 82 species of temperate grassland.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 209
ANNEX 9. The development and application of bio- economic modelling for silvoarable systems in Europe
A.R. Graves1, P.J. Burgess1, J.H.N. Palma2, F. Herzog2
G. Moreno3, M. Bertomeu3, C. Dupraz4, F. Liagre5, K. Keesman6, W. van der Werf6
A. Koeffeman de Nooy7, J.P. van den Briel7
1Institute of Water and Environment, Cranfield University, Silsoe, UK
2Agroscope FAL Reckenholz, Zürich, Switzerland
3Universidad de Extremadura, Centro Universitario de Plasencia, Plasencia, Spain
4Institut National de la Recherche Agronomique, Montpellier, France
5Assemblée Permanente des Chambres d’Agriculture, Paris, France
6Wageningen University, P.O. Box 43, 6700 AA, Wageningen, the Netherlands
7Stafkantoor Gelders Particulier Grondbezit, Wageningen, the Netherlands
Corresponding author: A.R. Graves
Tel: +44 (0)1525 863107
Fax: +44 (0)1525 863001
KEYWORDS
Agroforestry, silvoarable, arable, forestry, modelling, biophysical, economics, Farm- SAFE, Yield-SAFE, temperate, walnut, poplar, wild cherry, oak, stone pine
ABSTRACT
The European Union has introduced measures to promote the integration of trees within farm businesses. Although silvoarable agroforestry is one method by which this can be achieved, the implications at a plot- and farm-scale are poorly understood. From 2001 to 2005, the Silvoarable Agroforestry for Europe project therefore developed computer-based tools to evaluate both the biophysical and economic performance of arable, forestry and silvoarable systems under different European conditions. A biophysical model called “Yield-SAFE”, based on light and water competition, was developed to predict long-term arable, forestry and silvoarable yields for given sets of climate and soil conditions. The output from this model was then used in a plot- and farm-scale economic model called “Farm-SAFE” to determine profitability and resource use. Both models were parameterised and used for selected regions of France, Spain and the Netherlands. The analysis in France suggests that walnut and poplar silvoarable systems could provide a profitable alternative to arable and forestry systems, while in Spain a modest restructuring of the amount and delivery of agricultural payments would increase the attractiveness of silvoarable systems of holm oak and stone pine. In the
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 210
Netherlands, low timber value and the opportunity cost of losing arable land for slurry manure application made both silvoarable and forestry systems uncompetitive with arable systems.
INTRODUCTION
Agroforestry is a form of multi-cropping involving at least one woody-perennial species and significant ecological and economic interactions. Agroforestry systems can be described by their components (crops, animals and trees) and their spatial (dispersed or zoned) and temporal (coincident to sequential) arrangement (Nair 1985, Sinclair 1999). Silvoarable agroforestry, defined as the practice of growing an arable crop between spatially-zoned trees (Dupraz and Newman, 1997, Burgess et al., 2004b), is a form of agroforestry that could be undertaken on mechanised arable farms in Europe.
The majority of research on agroforestry systems has been undertaken to evaluate their biophysical performance despite the observation that it is often socio-economic constraints that limit their adoption (Graves et al., 2004, Mercer et al., 1998). Since there are potentially many biophysical and socio-economic interactions between the tree and crop components of silvoarable systems (Dyack et al., 1999) there is a need to consider both the biophysical and socio-economic aspects together. However both the biophysical and the socio-economic analysis of such systems are constrained by lack of experimental data describing the effect of different permutations, for example spacing and different tree species. There are also problems in describing the socio-economic integration and the interaction between the short- and long-term components over the length of a tree rotation.
Computer simulations provide a means of systematically undertaking biophysical and economic analyses of silvoarable systems in the absence of empirical data. Various biophysical and economic models have been developed for monocultures of arable and forestry systems, but few have been developed for silvoarable agroforestry (Graves et al., 2005a). The current bio-economic models of silvoarable systems range from detailed biophysical models with limited economic analysis to economic models that use biophysical data from an external source (Graves et al., 2005a). Bio-economic models have been used to examine the profitability (Thomas 1991, Willis et al. 1993, Thomas and Willis, 1997; Burgess et al. 2000) and feasibility (Dupraz et al., 1995) of silvoarable systems in Europe. Profitability is normally assessed at a one-hectare scale and performance is compared to competing enterprises such as arable agriculture and forestry. Feasibility is often determined at a farm-scale to view how silvoarable agroforestry affects cash-flow and resource use. This paper describes the integrated use of a bio-physical and an economic model, at both a one-hectare- and farm-scale, to determine the potential profitability and feasibility of silvoarable agroforestry in Europe.
METHOD
The study focused on three countries (Spain, France and the Netherlands) with differing climates, tree and crop species, and levels of practical experience in implementing agroforestry. Potential sites for the uptake of silvoarable agroforestry, termed landscape test sites, were identified in each country using a geographical information system. Annual yields of trees and crops were derived using a bio-
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 211
physical model called “Yield-SAFE” (van der Werf et al., 2005) and profitability and feasibility were determined using an economic model called “Farm-SAFE” (Graves et al., 2005).
Identification of landscape test sites
An initial requirement was to identify landscape test sites where silvoarable systems could be used as an alternative to arable systems. A geographical information system (ArcGIS - ArcInfo© and ArcInfo WorkStation© 8.3) was used to select the three dominant environmental classes in each country from an environmental classification of Europe, based on a statistical analysis of climate and topography (Mücher et al. 2003). In Spain and France three classes were identified in each country, but in the Netherlands there was only one class. The location of arable land was derived using a land cover classification from the Pan-European Land Cover Monitoring (PELCOM) project (Mücher et al. 2000). From the combined dataset, three landscape test sites measuring 4 km x 4 km were selected for each environmental class to give nine, nine and three landscape test sites in Spain, France and the Netherlands respectively; two landscape test sites in France were later discarded for lack of associated data, bringing this total to seven landscape test sites. In Spain the sites ranged from Alcala la Real in Andalucia in the south to St Maria del Paramo in Castilla y Leon in the north. In France, the landscape test sites ran across central France from Champdeniers in Poitou Charente in the west to Champlitte in Franche Comté in the east (Table 1). In the Netherlands the landscape test sites were located in the central and eastern part of the country.
Characterisation of landscape test sites
To provide input data for the biophysical model, daily mean values of air temperature, total short-wave radiation, and rainfall were generated for each landscape test site using CLIGEN 5.2 (United States Department of Agriculture, 2005) with reference values from the nearest weather station (Global Data Systems, 2005). The annual values for mean air temperature and total radiation were highest in Spain (9.1-15.5oC and 5480-6600 MJ m-2) and lowest in the Netherlands (8.8-9.0oC and 3690-4830 MJ m-2) (Table 1). Mean annual total rainfall was lowest in Spain (320-530 mm) and highest in France (590-1080 mm). In Spain, much of the rainfall occurred in winter with minimal rainfall in the summer months. In France, this seasonality of rainfall, although greatly reduced, was still evident while in the Netherlands rainfall was generally consistent throughout the year.
Each landscape test site was also characterised in terms of soil depth and texture and using a classification of hydraulic properties of European soils (Wösten et al. 1999), the available soil water content was calculated using van Genuchten’s equation (1980). Further data layers for elevation and land cover were developed. Pre- existing electronic data were used where available or digitised from paper sources, for example, topographic maps. The radiation throughout each landscape test site was calculated with digital elevation models in the Digitales Gelände-Modell (DiGeM©) (Conrad, 2002), and the relative radiation obtained using the value from a flat un-shaded pixel as the reference for one hundred per cent. Finally, field visits were made to each site to confirm existing interpretation, improve existing data, and provide missing data.
(Table 1)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 212
A cluster analysis of the available soil water content and the percentage solar radiation within each landscape test site was then used to develop between one and four “land units” (Table 2). The land available for silvoarable agroforestry in each landscape test site was assumed to be equivalent to the land available for arable production and this was selected by excluding non-arable land. The area of a specialist cereal farm in each landscape test site was determined from regional data in the Farm Accountancy Data Network (FADN) (European Commission, 2005) in Spain, the Agricultural Economics Research Institute in the Netherlands (2005) and from the Réseau d’observation des systèmes d’exploitation (ROSACE) (Assemblée Permanente des Chambres d’Agriculture, 2005) in France. Where there were no data relating to a specialist cereal farm, farm size was related to the most frequently occurring farm types for the landscape test site region, in the case of Alcala la Real in Spain, an olive farm, and in the case of the Netherlands, pig, dairying and general field cropping farms. At each landscape test site, the proportion of the area of each land unit relative to the total area of the land units was used to represent the proportion of each land unit within a hypothetical farm.
(Table 2)
Selection and management of tree and crop species
Annual yields of trees and crops in arable, forestry and silvoarable systems were required for each land unit as inputs for the economic analysis. The tree and crop species for forestry and arable production were chosen to reflect the most the likely practice at each landscape test site. In France and the Netherlands, the trees were selected because they were timber trees; in Spain the choice of tree also reflected policy constraints and issues of ecological importance. The forestry systems selected for Spain comprised holm oak (Quercus ilex) and stone pine (Pinus pinea). In France wild cherry (Prunus avium), walnut (Juglans spp.), and poplar (Populus spp.) were chosen and walnut and poplar were selected in the Netherlands. The arable systems in Spain were based on wheat, sunflower and fallow. In Poitou Charentes and Centre in France, they were based on wheat and sunflower and in Franche Comté, on wheat, oilseed and grain maize; in the Netherlands on wheat and forage maize. The silvoarable systems integrated the forestry tree species and arable crop species and rotation for each land unit.
The management of the forestry systems at each landscape test site was based on local practice. In Spain, planting densities, thinning and pruning for oak were derived from Pulido et al., (2003) and for stone pine from Yagüe (1995) and Montero and Cañella (2000). In France management for forestry systems was developed from the Institut pour le Développement Forestier (1997), Souleres (1992), Boulet-Gercourt (1997) and the Centre Régional de la Propriété Forestière (1997) for walnut, wild cherry and poplar. In the Netherlands, the receipt of grants was conditional on an appropriate planting density, given by the Ministerie van Landbouw, Natuur en Voedselkwaliteit (2004), and thinning and pruning regimes were applied using the management rules in France. The management for the arable systems reflected local practice.
Biophysical modelling
The radiation, temperature, rainfall, soil depth and texture data for each land unit were used as inputs in a daily time-step bio-physical model of tree and crop production, based on competition for light and water (Yield-SAFE) (van der Werf et
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 213
al., 2005) and implemented in Microsoft Excel© by Burgess et al. (2004a) to predict annual tree and crop yields.
The parameters used in Yield-SAFE to describe the growth of each tree and crop species were determined from published material and calibrations. An initial calibration for “potential” monoculture yields (Ittersum and Rabbinge, 1997) was undertaken against datasets of tree volume and crop yields under high yielding conditions in the Atlantic and Mediterranean zones, assuming within the model that light and temperature but not water, limited growth (Burgess et al. 2004a). Then at each landscape test site and assuming light, temperature, and water limited growth within the model, the values of three parameters (harvest index, water use efficiency and a management factor) were adjusted within acceptable boundaries so that output from the model over the duration of the tree component matched an “actual” monoculture tree and crop yield (Ittersum and Rabbinge, 1997). The tree and crop management defined previously for the monocultures and “reference” soil depth and texture were also used. The monoculture management and actual and reference values were determined for each landscape test site during workshops held in each country (Palma & Reisner, 2004; Reisner, 2004; Herzog, 2004).
In Spain, the actual timber volumes for oak and stone pine in all the landscape test sites in year 60 were assumed to be 0.22 m3 and 0.26 m3 tree-1 respectively, indicating slow growth. In France, wild cherry (1.04-1.06 m3 tree-1) and walnut (1.04 m3 tree-1) for the same rotation were comparatively fast-growing trees. Poplar was the fastest growing tree with actual yields of 1.46-1.51 m3 tree-1 after 20 years. In Spain, actual yields for wheat were comparatively low (1.62-3.71 t ha-1) compared to those in France (6.5-8.0 t ha-1) and the Netherlands (7.8 t ha-1). Actual sunflower yields were lower in Spain (0.60-1.09 t ha-1) than in France (2.3-2.5 t ha-1). Actual yields for oilseed (3.2-4.0 t ha-1) and grain maize (7.5-8.0 t ha-1) were assumed only for France and an actual yield for fodder maize (12 t ha-1) assumed only for the Netherlands.
Using the parameter set developed for actual yields and soils at each landscape test site, tree and crop yields for each land unit were predicted for monoculture forestry and arable systems and two silvoarable systems of 50 or 113 trees ha-1. From the biophysical yields, it was possible to estimate a land equivalent ratio (LER) for each system. LERs were initially defined for mixed cropping systems (Mead and Willey, 1980) and have been adapted for agroforestry systems (Ong 1996, Dupraz, 1998). The LER is “the ratio of the area under sole cropping to the area under the agroforestry system, at the same level of management that gives an equal amount of yield” (Ong, 1996) and is expressed as:
Tree silvoarable yield Crop silvoarable yield LER = + Equation 1 Tree monoculture yield Crop monoculture yield
Where more than one crop occurred in the rotation, a weighted ratio for each crop was used, depending on its proportion in the rotation.
Plot-scale economic modelling
The predicted annual yields of trees and crops were used as inputs for a plot- and farm-scale cost-benefit economic model called “Farm-SAFE” (Graves et al., 2005).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 214
In arable systems, profitability is typically compared on an annual and per unit area basis by adding the revenue generated (R) to the variable costs associated with generating that revenue (V) to give a gross margin (Gross margin = R − V ) (Nix, 1999; Ministry of Agriculture, Fisheries and Food, 1983). However, in tree-based systems, “assignable fixed costs” such as labour and machinery (A) are commonly included and can be derived per unit area. Therefore the arable, forestry, and silvoarable systems were compared using their net margin (Net margin = R − V − A ) (Willis et al. 1993; Burgess et al. 2000, Graves et al., in press). As the benefits and costs associated with tree-based systems occur over many years, discounted cost benefit analysis was used to define the “present” value of future costs and benefits from the arable, forestry and silvoarable systems using the approach defined by Faustmann (1849). The net “present” value (NPV; units: € ha-1) was expressed as:
t=T (R −V − A ) NPV = t t t Equation 2 ∑ t t=0 (1+ i)
Where: NPV was the net present value of the arable, forestry or silvoarable -1 enterprise (€ ha ), Rt was the revenue from the enterprise (including subsidies) in -1 -1 year t (€ ha ), Vt was the variable costs in year t (€ ha ), At was the assignable fixed costs in year t (€ ha-1), T was the time horizon (years), and i was the discount rate (discount rate = 4%).
In order to compare systems with different rotation lengths, an infinite net present value was calculated. This was the net present value defined over an infinite rotation, in which each replication had a rotation of n years. The infinite NPV was defined as:
(1+ i) n Infinite NPV =NPV Equation 3 (1+ i) n −1
The infinite net present value was also expressed as an equivalent annual value (EAV) using the following formula:
EAV = infinite NPV × i Equation 4
Assessing the feasibility of a given system involves determining how it modifies flows of farm resources. This is achieved by multiplying plot-scale flows of money, land, and labour by their area on the farm and aggregating the results, then substituting a given system with another system, and assessing the effect on farm resources with and without the substituted system. A maximum of four arable, four forestry and four silvoarable systems could be used to represent a single farm in the “Farm-SAFE” economic model. Economic feasibility was determined using the infinite NPV of the -1 farm ( iNPVfarm; units: € farm ). This combined the NPV of the different systems and -1 the NPV of “farm fixed costs” (Ft: units: € farm ) over the same period of time and was defined as:
l=4 t=T F (1+ i) n iNPV = NPV a + NPV a + NPV a − t Equation 5 farm ∑()a a f f s s ∑ t n l=1 t=0 (1+ i) (1+ i) −1
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 215
Where: l was one of four possible land units, NPVa, NPVf, and NPVs were the net present values (€ ha-1) of arable, forestry and silvoarable enterprises in each unit l; aa , a f , and, as were the area (ha) of arable, forestry, and silvoarable systems in each -1 unit l, Ft was the farm fixed cost in year t (€ farm ), T was the time horizon (years), i was the discount rate and n was the duration of the rotation (years).
Parameterisation and use of Farm-SAFE
The financial data for arable, forestry and silvoarable systems were collected on electronic templates for each landscape test site, using local and national statistics, and expert opinion.
Arable and crop component finance
The revenue (crop value and associated subsidy), and the variable and assignable fixed costs for each arable system are described fully by Graves et al. (2005b). However, for clarity some key values are described. The assumed value of the arable crops ranged from 85 € t-1 for grain maize to 280 € t-1 for sunflower; the assumed value of wheat grain ranged from 102 to 142 € t-1. Assumed variable costs tended to be lowest in Spain (45-189 € ha-1) and highest in the Netherlands (457-479 € ha-1), and assignable fixed costs such as machinery and labour followed a similar pattern. For the crop component of the silvoarable system, the variable and assignable fixed costs were applied according to the proportion of intercrop area in the system which was constant. Also, as intercrop yields decrease over time due to tree growth, it was assumed that cropping would only continue for as long as the intercrop net margin (calculated on a five year moving average to remove the effect of yield failure caused by poor weather) was profitable, after which it was assumed the intercrop area would be fallow.
Forestry and tree component finance
The financial data for forestry and the tree component of the silvoarable system comprised the revenue from timber and subsidies, and the costs of woodland establishment and management. These are summarised below, but explained fully in Graves et al. (2005a). The revenue from timber was calculated using relationships between the standing value of the tree and the average tree volume for each species in each country. In Spain, the value of oak (17 € m-3) and pine (8-19 € m-3) was low. By contrast, in France, the value of walnut (40-1300 € m-3), wild cherry (10-380 € m- 3), and poplar (7-55 € m-3) was relatively high; thinned timber, given a different per cubic metre price to clear-felled timber, was also relatively valuable. In the Netherlands, the perceived value of walnut (18-41 € m-3) was much lower than in France, but the value of poplar (19-97 € m-3) was slightly higher.
The costs associated with the forestry system and the tree component of the silvoarable systems were based on numerous sources. Costs varied between countries, tree species and regions and regarding the tree component of the silvoarable system, were not assumed to be proportional to the number of trees or the area of the tree component (except in the Netherlands), as was the assumption for the crop component. The cost of ground preparation was anticipated to be highest in Spain and the Netherlands and lowest in France. This was due to difficult soil conditions in Spain, where it was anticipated that tree pits would need to be prepared, requiring use of specialised machinery (including labour) at a contract rate
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 216
of 31 € hr-1. In the Netherlands, it was anticipated that labour and machinery would be provided by external enterprises at a cost of 22 € hr-1. In France, however, it was anticipated that the farmer would undertake the majority of operations at a cost of € 7.8 hr-1. The cost of planting materials was greatest for walnut (6 € tree-1) and poplar (4 € tree-1) in France and walnut (5 € tree-1) in the Netherlands. Oak (0.36 € ha-1) and pinus (0.76 € ha-1) in Spain were relatively inexpensive. Tree protection materials, such as spiral guards or fencing, were highest for walnut and cherry in France (1.5 € tree-1) and lowest for walnut and poplar in the Netherlands (0.29 € tree- 1). The time required for planting and protecting the trees was highest in Spain (2.7 min tree-1), than France (1.0-2.0 min tree-1), and lowest in the Netherlands (0.8 min tree-1). In France (15 € hr-1) and the Netherlands (22 € hr-1), it was anticipated that planting and protection would be carried out by externally contracted enterprises; in Spain it was anticipated that this would be done using locally available labour (7.8 € hr-1). The full establishment cost of forestry systems was greatest in the Netherlands (3420 € ha-1 for walnut; 1940 € ha-1 for poplar) and lowest in Spain (770 € ha-1) for oak systems at 400 trees ha-1. The full establishment cost for forestry systems of cherry (1510 € ha-1), walnut (1633 € ha-1), and poplar (1260 € ha-1) in France and high density oak (1470 € ha-1) and pine (1786 € ha-1) in Spain were between these extremes. The full establishment cost of the tree component in the silvoarable systems was lower for each species. For the 113 trees ha-1 systems, these ranged from 1200 € ha-1 for walnut in the Netherlands to 233 € ha-1 for oak in Spain; for the 50 trees ha-1 systems they ranged from 710 € ha-1 in the Netherlands to 120 € ha-1 for oak in Spain.
Significant maintenance costs included weeding, sward establishment, pruning and thinning. In Spain, it was anticipated that management would be minimal because of the low financial value of the oak and pine timber. The main costs in the forestry system were associated with weeding in the initial three years and establishing a grass sward in year 12. For the tree component of the silvoarable system, the only cost-bearing maintenance operation was assumed to be weeding in the initial five years. Both these operations were assumed to be externally contracted at a rate of 31 € h-1. Pruning and thinning were assumed to be free of cost as an established system exists whereby harvested oak and pine timber is given in lieu of payment to those who undertake the work. By contrast, management was much more intensive in France and in the Netherlands. In France, the control of undergrowth between trees was a significant cost for about the first quarter of a forestry rotation and for the duration of arable cropping in a silvoarable system. Other significant costs included pruning and an annual land tax that varied marginally between regions. In the Netherlands, the costs of establishing a grass sward in the first year (417 € ha-1 grass) and subsequent maintenance (136 € ha-1 grass a-1) were high. Pruning, especially for walnut, and thinning were also significant costs. In addition, it was assumed that an opportunity cost (a nitrate levy of 408 € ha-1a-1) was incurred when arable land was converted to forest, because the land could no longer be used to accept slurry manure. This was also applied on a pro-rata basis to the tree-strips in the silvoarable system.
Pre-2005 grant regime
The relative profitability of forestry, arable and agroforestry systems on farms in the European Union is significantly affected by the grant regime. In the pre-2005 grant scenario, it was assumed that direct payments on the arable system and crop
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 217
component of the silvoarable system would be dependent on the crop species and the portion of arable land in the system. These were greatest for maize (400 € ha-1) in the Netherlands and least for wheat (129 € ha-1) in Spain, but also varied with crop species and in France, with region. The pre-2005 payments on forestry and tree component of the silvoarable systems were established from local and national statistics and expert opinion. In Spain, farmers received a planting grant (849-1593 € ha-1) dependent on tree species, a compensation payment (225-325 € ha-1 a-1) for 20 years depending on location and previous land-use and a maintenance grant (180- 288 € ha-1 a-1) for five years, subject to appropriate management of the trees (Graves et al., 2005). In France, in Poitou Charentes and Centre, planting grants covered 50% of tree costs in the first four years and compensation payments (240-300 € ha-1 a-1) were available for walnut and cherry for ten years and for poplar for seven years. In Franche Comté, there were no grants or payments, due to existing and substantial areas of forest. In the Netherlands, a planting grant of 95% of costs was available up to a maximum of 1500 € ha-1, a compensation payment of 100 € ha-1 a-1 for five years and a maintenance payment of 545 € ha-1 a-1 for 18 years. For the tree component of the silvoarable system, all tree payments were forfeited in Spain and the Netherlands. In the Poitou Charentes and Centre regions of France, establishment grants were available at 50% of the tree costs in the first four years, but no tree payments were available in Franche Comté.
Post-2005 grant regime
In the post-2005 grant scenario, the changes anticipated for the Common Agriculture Policy were implemented. For the arable crop, the changes meant that the area payments could be fully decoupled from crop type, resulting in a single farm payment for as long as the land was cropped. The per hectare value of these payments were calculated to be lowest in Spain (116-330 € ha-1) and highest in France (329-353 € ha-1) and the Netherlands (353-586 € ha-1). In the post-2005 scenario for forestry, existing levels of payments applied, where they were in accordance with the rural development strategy of the European Union (2004). In France, there was therefore no change, but in Spain and the Netherlands, planting payments at each site were changed to 50% of tree costs in the first four years. The compensation payments and maintenance grants were reduced to 500 € ha-1 a-1 with a maximum duration of 10 years, unless they were already below these levels. In that case existing values were used.
Since the effect of these changes on silvoarable systems is still unclear, two extreme scenarios were developed for the post-2005 situation (Erreur ! Source du renvoi introuvable.). In scenario 1, the single farm payment was assumed for the percentage of crop area in the system with no tree payments. In scenario 2, the single farm payment was assumed for the whole system with 50% of the tree costs in the first four years covered by a planting grant.
(Table 3)
Farm-scale data
Only a brief description of the approach and data used in the farm-scale modelling is provided here. A more detailed description can be found in Graves et al. (2005a).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 218
Economic feasibility was assessed by multiplying the one-hectare results for each land unit by their area and adding farm fixed costs from the FADN and ROSACE for the hypothetical farms at each site (Equation 5). The quality of the land units was ranked assuming that higher average yields meant better land. Expert opinion was then used to determine which tree species and which crop rotation would be most suitable for each land unit. The infinite net present value of the farm was used to evaluate the economic effect of planting 10% of the farm with forestry or silvoarable systems in comparison with the status quo arable farm under the pre-2005 and post- 2005 grant regimes. Planting was assumed in year 1 and holm oak, stone pine, wild cherry and walnut were “harvested” to provide revenue in year 60. A rotation of 20 years was assumed for poplar, and by re-planting in years 21 and 41, three full rotations of poplar were completed in 60 years. It was assumed for poplar that the tree related grants in year 21 and 41 would be the same as for year 1.
RESULTS AND DISCUSSION
Biophysical production in arable and forestry systems
The predicted yield of the monoculture arable crops within a specific year on the 42 land units ranged from 0.2 t ha-1 for sunflower in Spain to 15.9 t ha-1 for maize in the Netherlands (Erreur ! Source du renvoi introuvable.). Although the greatest absolute variation in yield was associated with high yielding crops in the Netherlands and France, the relative variation in yields was greatest in Spain. For the forestry systems, the mean timber volume per tree ranged from 0.25 m3 for stone pine after 60 years, to 1.34 m3 for poplar after 20 years. The maximum recorded tree size was for poplar (1.59 m3) in France and the minimum for oak (0.23 m3) in Spain. The standard deviation suggested that absolute variation was greatest for wild cherry in France and poplar in the Netherlands. The coefficient of variation showed that the relative variation was greatest for wild cherry, oak and poplar in the Netherlands.
(Table 4)
Within each landscape test site, crop yield within a land unit could potentially vary with soil depth, soil type and radiation level. For each crop, except wheat in Spain, there was a significant positive correlation between predicted annual crop yields and soil depth (Erreur ! Source du renvoi introuvable.). The standard error of the estimate showed that in absolute terms, variation was greatest for wheat in Spain and France. However, in relative terms, the variation was greatest for wheat in Spain. Predicted timber yields were also positively correlated with soil depth for cherry, poplar and oak (Erreur ! Source du renvoi introuvable.). However, this correlation was only significant (P=0.05) in the case of wild cherry.
In each country, analysis of variance (analysis not summarised here) showed that there were significant differences (P=0.05) in soil texture and predicted crop yields, except in the case of oilseed in France. However, there were no significant difference in soil texture and predicted timber yields in any of the countries.
(Table 5)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 219
Biophysical production in silvoarable systems
The biophysical outputs from Yield-SAFE for the silvoarable systems (50 and 113 trees ha-1) showed a general decline in crop yields as the trees became larger and competed more effectively for light and water. Typical relations for four land units are shown in (Figure 2). Oak ((Figure 2)a) and stone pine (which showed similar growth over time to oak and is therefore not shown) grew slowly throughout the whole rotation. Hence relatively high crop yields were sustained for most of the tree rotation. The initial rate of timber formation by wild cherry ((Figure 2)b) was slow compared with walnut ((Figure 2)c) and poplar ((Figure 2)d), and crop yield reduction in the walnut and poplar systems was predicted to occur earlier than in the wild cherry systems.
The model was also used to predict difference in crop and tree yield at two tree densities (50 and 113 trees ha-1). As expected relative crop yields were greatest in the 50 tree ha-1 system and relative timber yields (m3 ha-1) were greatest in the 113 tree ha-1 system (Erreur ! Source du renvoi introuvable.).
(Figure 1)
(Figure 2)
In Spain, the relative yields of autumn-planted species, such as wheat, tended to be greater than for spring-planted crops, such as sunflower (Erreur ! Source du renvoi introuvable.a). As oak and stone pine are evergreen species, it was assumed that this was due to greater competition experienced by the spring-planted crop for water. In France, the difference in the relative yield of the autumn- (i.e. wheat and oilseed) and spring-planted (sunflower and grain maize) crops was larger than in Spain (Erreur ! Source du renvoi introuvable.b). This was probably due to reduced competition for light, because the tree species planted in France were deciduous and hence had no leaves for a large proportion of the growing period of the autumn- planted crops, whereas in Spain as the trees were evergreen and competition for light was similar for both the spring and autumn-planted crops. Under poplar in the Netherlands (Erreur ! Source du renvoi introuvable.c), similar effects regarding the difference between autumn-planted wheat and spring-planted forage maize were evident. These patterns were similar in both the low density and high density systems, but the relative yields of the crops were higher at 50 trees ha-1 than at 113 trees ha-1.
(Figure 3)
Land equivalent ratios
The predicted land equivalent ratios for timber (including thinnings) and crop yield (assuming a full rotation) of the silvoarable systems at both 113 and 50 trees ha-1, with a few exceptions, were between 1 and 1.4. Hence the Yield-SAFE model predicted that, under typical management, integrating crops and trees on the same land was more productive than growing them separately. The relationship between relative tree and crop yield suggested that the land equivalent ratio formed a convex arc with maximum values obtained when the trees and crops had similar relative yields and minimum values where either the tree or crop component was dominant (Erreur ! Source du renvoi introuvable.). At each landscape test site, the land
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 220
equivalent ratio at 113 trees ha-1 (Erreur ! Source du renvoi introuvable.a) was greater than that at 50 trees ha-1 (Erreur ! Source du renvoi introuvable.b), suggesting that in biophysical terms, 50 trees ha-1 was sub-optimal and more efficient use of resources in silvoarable systems could be achieved above this density.
The highest land equivalent ratios at both tree densities were associated with poplar, walnut and cherry systems in France ((Figure 4)
(Figure 5)a and (Figure 4)
(Figure 5)b). Oak and pine in Spain at both densities were associated with much lower land equivalent ratios. The reason for this is not clear; it may be that predicted growth of oak and pine was so slow that they were unable to make use of available resources at the densities used in the silvoarable systems. Alternatively, it may be that the crops competed more strongly for water than other trees of the same species. In either case, production benefits from oak and pine-based silvoarable systems in Spain appear to be limited unless tree densities can be increased without detriment to the relative yield of either component.
(Figure 4)
(Figure 5)
Plot-scale economic results
The annual time-series production data developed using Yield-SAFE and economic data for crop grants and crop revenue and costs, tree grants and tree revenue and costs for landscape test site were modelled in Farm-SAFE (Graves et al., 2005b). The economic performance of the arable, forestry and the silvoarable systems (113 trees ha-1 only) was compared using the equivalent annual value (EAV) (discount rate = 4%). The effects of zero grants, the pre-2005 grants and the post-2005 grants were also examined. As intercrop yields decreased over time due to tree growth, the crop rotation was optimised by ending intercrop production when the five-year moving-average of the intercrop net margin was zero.
Profitability with no grants
The equivalent annual values (at a discount rate of 4%) of the forestry systems with oak and stone pine in Spain, poplar and walnut in the Netherlands, and cherry in France were negative (Erreur ! Source du renvoi introuvable.). Only walnut, due to the high value of the timber, and poplar, due to the short rotation, in France was profitable. The low profitability of forestry in the Netherlands was partly due to the opportunity cost of slurry manure management, as the application allowance was assumed to be zero for forest land. The equivalent annual values (4% discount rate) of the arable system were positive in Alcala la Real, Cardenosa El Espinar, Fontiveros, Olmedo and St Maria del Paramo in Spain, in Poitou Charentes and Centre in France and at all sites in the Netherlands, but negative in Torrijos, Ocaña and St Maria del Campo and at most sites in Franche Comté (i.e. at Dampierre and Vitrey). Positive values were associated with sites of high productivity. In Franche Comté, relatively high assignable fixed costs explained the negative values.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 221
The equivalent annual values (4% discount rate) of the silvoarable systems in Spain were marginally below those for the arable system. By contrast, in France, values for the silvoarable systems with walnut, with poplar in Centre, and with wild cherry in Poitou Charentes and Franche Comté were higher than those for both arable agriculture and forestry. In the Netherlands, the values for the silvoarable system with poplar were marginally greater than the arable system, but the value for the silvoarable system with walnut was negative because of the long tree rotation and low value given to walnut timber.
(Figure 6)
The long-term cash value of pre-2005 and post-2005 grant regimes
Under the pre-2005 grant regime, the actual cash value (discount rate = 0%) of forest payments for the duration of the tree rotation was greatest in the Netherlands and lowest in France (Erreur ! Source du renvoi introuvable.). The assumed levels of arable compensation payments were marginally greater in the Netherlands than in France, and both were much greater than in Spain.
Within Spain, support for silvoarable agroforestry was lower than for forestry and the arable system because of ineligibility for tree grants and reduction of the arable compensation payments by twice the proportion of the canopy area of the trees. In France, in Poitou Charentes and Centre, arable payments were at least five-times the value of forestry payments and the value of silvoarable payments was marginally less than that for arable systems. In Champlitte, Dampierre and Vitrey in Franche Comté, there were no forestry payments and hence the greatest level of support was for arable systems. For poplar sites, payments for all systems were relatively low because of the 20- rather than the 60-year rotation. Support for walnut and poplar forestry in the Netherlands was identical because they were both temporary, production-based systems. Since arable payments were dependent on the length of the tree rotation, they were greater for walnut (Bentelo) rather than for poplar (Balkbrug and Scherpenzeel). In each case, the support for silvoarable systems was less than for forestry and arable systems, as no payments were received for the tree component.
The actual cash value of each system in the post-2005 payment scenario and the change, relative to the pre-2005 scenario was determined (Table 6). The greatest relative change was predicted for Spain, where forestry payments were greatly reduced, due to compensation being limited to 10 years, while for arable and particularly for silvoarable systems, payments were predicted to increase. The predicted value of the new single farm payment at Alcala la Real, and St Maria del Paramo and St Maria del Campo was greater than pre-2005 area payments, as support for non-arable activities on typical farms in these areas was assumed to be re-allocated on an area basis. The large relative increase of the cash value of payments in the silvoarable systems demonstrated the disadvantage of the system under pre-2005 regime. In France, there was no change for forestry, and only marginal changes for arable systems due to modulation under the single farm payment. For silvoarable systems, scenario 1 was similar to the pre-2005 regime but marginal benefits were evident under scenario 2. In the Netherlands, the major change was due to the reduction in the compensation payments associated with forestry from 18 to 10 years.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 222
(Table 6)
Equivalent annual value under pre-2005 grant regime
In the pre-2005 grant regime (Erreur ! Source du renvoi introuvable.), the equivalent annual values (4% discount rate) of forestry in Spain was generally higher than those for arable systems, except where crop yields were high. Because of the low level of government support, the equivalent annual value of the silvoarable systems was generally lower than for forestry and arable systems. In France, the equivalent annual value for the arable systems, tended to be greater than that for silvoarable agroforestry with wild cherry, which was much greater than that for wild cherry forestry. Hence, silvoarable agroforestry offered the most profitable means of establishing cherry trees at these sites. The predicted equivalent annual values of the silvoarable systems with poplar and walnut systems in France were higher than that for both forestry and arable systems. In the Netherlands, the conventional arable systems were the most profitable, followed by silvoarable agroforestry. Thus in the Netherlands, silvoarable systems also appeared to provide a more profitable means of establishing trees in the landscape.
(Figure 7)
Equivalent annual value under post-2005 grant regime
In the post-2005 (Erreur ! Source du renvoi introuvable.), compared to the pre- 2005 (Erreur ! Source du renvoi introuvable.), grant regime in Spain, the equivalent annual value of forestry was predicted to be reduced whilst it was predicted to increase for arable and silvoarable systems despite modulation (Figure 7). In France, the values for the equivalent annual value were generally similar to those under the pre-2005 regime. For silvoarable systems, the pessimistic scenario, scenario 1, resulted in marginal reductions, while the optimistic scenario, scenario 2, resulted in marginal increases. In the Netherlands, a substantial decrease in the equivalent annual value of forestry was predicted, whilst the change in the equivalent annual value of arable systems was marginal. For silvoarable agroforestry, little change was predicted for scenario 1, but a small and consistent increase was predicted for scenario 2.
The net effect of the above changes was most significant in Spain. Under the pre- 2005 scenario, forestry systems were consistently more profitable than silvoarable systems. Under the post-2005 scenario, silvoarable agroforestry was predicted to be more profitable than forestry in almost 50% of cases, although both systems were predicted to remain less profitable than arable agriculture. At sites in France and the Netherlands, the ranking of the systems in the post-2005 and pre-2005 regimes were similar.
(Figure 8)
Farm-scale feasibility
Under the pre-2005 grant-regime in Spain, it was not profitable to re-plant arable land with a silvoarable system. This was due to low timber volume and value, the lack of tree grants and the loss of arable area payments by twice the canopy area of the tree component. By contrast, establishing forestry on arable land was predicted generally
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 223
to increase farm profitability (Figure 9). In France, establishing silvoarable agroforestry was predicted to increase farm profitability when it involved walnut or poplar, and decrease it if it included wild cherry. In each case, silvoarable systems improved farm profitability relative to forestry on the same area of land. In the Netherlands, both forestry and silvoarable systems reduced farm profitability.
Under the post-2005 grant regime in Spain, replanting arable land with silvoarable systems continued to result in reduced farm profitability. Replanting arable land with forestry was predicted to increase farm profitability at five sites (Torrijos, Ocaña, Almonacid de Zorita, Olmedo and St Maria del Campo), and decrease it at the other four. In France, farm profitability increased following the establishment of silvoarable systems with walnut and poplar, and decreased with silvoarable systems using wild cherry; in both cases silvoarable systems were more profitable than forestry. In the Netherlands, there was no advantage to introducing silvoarable systems or forestry in comparison with the status quo.
An analysis of the frequency with which silvoarable systems increased profitability relative to the status quo (Erreur ! Source du renvoi introuvable.) showed that in Spain there were no cases where farm profitability was improved by establishing silvoarable systems. Instead government support favoured the establishment of forestry, and this was attractive in about 80% of cases under the pre-2005 grant regime. The post-2005 regime was predicted to reduce the relative profitability of forestry, but forestry still remained financially attractive on about 50% of the selected farms. In France, under the pre-2005 grant regime, silvoarable systems were predicted to increase farm profitability in approximately 50% of cases. This frequency remained similar under scenario 1 of the post-2005 grant regime, and increased to 80% under scenario 2. The proportion of farms where forestry was attractive (20%) was less than for silvoarable systems and was the same for both the pre-2005 or post 2005 regimes. In the Netherlands (not shown), the introduction of forestry and silvoarable systems always reduced farm profitability, under the pre- 2005 and post-2005 payment scenarios.
(Figure 9)
In Spain the use of silvoarable systems was preferable to forestry in 12% of cases under the pre-2005 grant regime and 50% of cases in scenarios 1 and 2 of the post- 2005 grant regime. In France and the Netherlands, farm profitability was always increased with the use of silvoarable rather than forestry systems. Hence, in Spain, forestry generally provided the most cost effective method of establishing trees under the pre-2005 regime, an advantage predicted to disappear under post-2005 regime. In France and the Netherlands, silvoarable systems with walnut, wild cherry, and poplar provided the most profitable means of establishing trees on farms irrespective of grant regime.
(Figure 10)
SUMMARY AND RECOMMENDATIONS
Using a geographical information system, a statistical analysis of climatic, topographic and land classification data was used to select 19 landscape test sites in Spain, France and the Netherlands. Within each site, land use, soil depth and texture, and elevation were digitised. Daily weather data were generated for each
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 224
site using a weather generator. Proportional differences in solar radiation and soil water holding capacity were calculated and used in cluster analysis to divide the arable land at each site into between one and four land units. A biophysical model called “Yield-SAFE” was developed and calibrated for potential yields of a range of tree and crop species. Typical forestry and arable systems and associated management regimes were determined for each land unit and Yield-SAFE was calibrated for actual tree and crop yields at each site. The calibrated model was then used to calculate daily values of tree and crop yields for a forestry, arable and agroforestry system at each land unit according to changes in solar radiation, soil depth and texture. Financial data for forestry, arable, and silvoarable production at each site were collected and four grant scenarios were described (no grants, a pre- 2005 scenario, and two possible post-2005 scenarios). The financial data was combined with the physical values in an economic model called “Farm-SAFE”, and the equivalent annual value (discount rate = 4%) at a plot-scale and the infinite net present value at a farm-scale were used to examine the profitability of different systems.
The Yield-SAFE biophysical model predicted lower timber yields and crop yields per hectare for silvoarable systems compared to the forestry and arable systems respectively (Erreur ! Source du renvoi introuvable.). However, the total productivity of the silvoarable system, as determined by a land equivalent ratio, was predicted to be between 100 and 140% of that for the monoculture systems (Erreur ! Source du renvoi introuvable. and (Figure 4) (Figure 5)). High land equivalent ratios were achieved with a tree stand density of 113 rather than 50 trees ha-1, suggesting that the high density system made fuller use of the available light and water resources. The highest ratios were obtained by integrating deciduous trees and autumn-planted crops, which were complementary in terms of light use (Erreur ! Source du renvoi introuvable.). The lowest ratios were obtained from evergreen tree species in Spain, where productivity was appeared to be constrained by the slow growth of the trees and low soil water availability (Erreur ! Source du renvoi introuvable.).
At a plot scale, the economic performance of the systems was compared in a zero grant scenario (Erreur ! Source du renvoi introuvable.). In Spain, arable systems were marginally more profitable than silvoarable systems with oak or stone pine, which in turn were more profitable than forestry systems with the same species. By contrast in France, silvoarable systems with walnut in each of three regions, poplar in one region, and wild cherry in two regions were more profitable than arable and forestry systems. In the Netherlands, silvoarable systems with poplar, but not walnut, were predicted to be more profitable than the described arable system. However, both the poplar and walnut silvoarable systems were more profitable than forestry.
Under pre-2005 grants (Erreur ! Source du renvoi introuvable.), support for silvoarable systems in Spain and the Netherlands was substantially lower than for arable and forestry systems. Hence, the profitability of silvoarable systems was always less than for arable or forestry systems. In France, support for silvoarable systems was marginally lower than for arable systems but significantly higher than for forestry systems. Hence it was predicted that silvoarable systems with poplar and walnut could be more profitable (at a 4% discount rate) than both forestry and arable
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 225
systems. Silvoarable systems with cherry although more profitable than forestry were predicted to be less profitable than arable systems. In the Netherlands, silvoarable systems were more profitable than forestry, but less profitable than arable systems.
Under two possible post-2005 grant regime (Erreur ! Source du renvoi introuvable.), the relative value of support for forestry in Spain was predicted to decrease, whilst for silvoarable and arable systems it was predicted to increase. In France and the Netherlands the relative value of support for silvoarable systems compared to arable and forestry systems remained similar to the pre-2005 regime for scenario 1, and increased marginally for scenario 2. Hence the profitability of silvoarable systems in Spain increased and frequently exceeded the profitability of forestry systems, but remained marginally less profitable than arable systems. In France and the Netherlands, little relative change in profitability between the systems was predicted.
At a farm-scale and under both pre-2005 and post-2005 grants in France (Erreur ! Source du renvoi introuvable.), planting arable land with silvoarable systems of walnut and poplar increased farm profitability, while silvoarable systems with cherry reduced farm profitability. In Spain and the Netherlands, silvoarable systems consistently reduced farm profitability in comparison with the arable status quo. However, in both France and the Netherlands, silvoarable systems were a more cost- effective way of establishing trees on the farm than forestry (Erreur ! Source du renvoi introuvable.). In Spain, under pre-2005 grants, silvoarable systems were a less cost-effective means of establishing trees than forestry. However, under post- 2005 grants, silvoarable systems were predicted to be a most profitable means of establishing trees in half the examined cases.
A number of recommendations regarding further research can be made. Predictions are subject to uncertainty and this could be examined using sensitivity analysis or stochastic modelling. Certain baseline data could also be re-examined. The recorded value of walnut timber in the Netherlands and France differed greatly, even though both countries are part of a free-trade zone. This strongly influenced the relative profitability of walnut systems in these countries. The assumption regarding prohibition of slurry manure application in the Netherlands in forests also had an important effect. If this is a true opportunity cost, the establishment of productive forests on farms is unlikely to be attractive, unless the opportunity cost is removed or payment schemes can account for it. Assumptions regarding beating-up, tree management and the extent of payments could also be re-assessed for Spain. Tree mortality is likely to be high due to difficult conditions and should be accounted for; the assumptions regarding pruning and thinning costs in Spain may be valid for traditional management of widely spaced trees in open woodlands (dehesas), but invalid for forestry and silvoarable systems, even if these are established within areas where dehesas predominate. Finally, the assumptions and value of post-2005 grants should be re-assessed when the changes are implemented.
CONCLUSION
The process used to model plot- and farm-scale economics of arable, silvoarable and forestry systems in three European countries has been described. This integrated the use of geographical information systems with a biophysical model of tree, crop
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 226
and integrated tree and crop growth, and an economic model developed during the SAFE project.
Under the economic conditions envisaged in the analysis, the most financially attractive silvoarable systems tended to have a land equivalent ratio that was substantially above one. Conditions that most favoured a high land equivalent ratio appeared to be the use of relatively high tree-densities to make full use of available resources, the use of deciduous trees and autumn-planted crops to make complementary use of light, and a high soil water availability to ensure that extra biomass production could be sustained. Conversely, it appeared that low ratios were associated with low tree density, evergreen trees, spring-planted crops, and low soil water availability.
Silvoarable agroforestry was most financially attractive where both components of the system were profitable as a monoculture since an unprofitable or relatively unprofitable component tended to reduce the profitability of the mixed system. In addition, the profitability of silvoarable agroforestry tended to be maximised if the profitability of the forestry and agricultural system were similar. Under the two proposed post-2005 grant regimes, it is predicted that silvoarable systems with walnut and poplar in France could provide a profitable alternative to arable or forestry systems. In Spain, it appeared that holm oak and stone pine could be integrated into arable systems without significantly reducing arable production for many years. Since these trees are of ecological and landscape importance, rather than productive importance, additional support in the form of an agri-environment payment would be justified. A moderate annual amount would be sufficient to overcome income losses caused by yield reductions and encourage establishment for non-productive benefits. In the Netherlands, the low value of timber and an assumed opportunity cost of losing arable land for slurry manure application made silvoarable and forestry systems relatively unattractive compared with arable systems.
ACKNOWLEDGEMENTS
This research was carried out as part of the SAFE (Silvoarable Agroforestry for Europe) collaborative research project. SAFE is funded by the EU under its Quality of Life programme, contract number QLK5-CT-2001-00560, and the support is gratefully acknowledged. We also acknowledge and are thankful for the involvement of Terry Thomas, Bob Bunce, Yvonne Reisner, Klaas Metselaar and Martina Mayus at key stages in the project.
REFERENCES
Assemblée Permanente des Chambres d’Agriculture (2005). Réseau d’observation des systèmes d’exploitation. http://paris.apca.chambagri.fr/apca/default.htm (Accessed 5 May 2005) Agricultural Economics Research Institute (2005). Database of agricultural statistics for the Netherlands. http://www.lei.dlo.nl/uk/ (Accessed 5 May 2005). Boulet-Gercourt, B. (1997) Le merisier. Institut pour le Développement Forestier, 2ème édition. 128 pp. Burgess, P.J., Graves, A., Metselaar, K., Stappers, R., Keesman, K., Palma, J., Mayus, M. & van der Werf, W. (2005). Parameterisation of the Yield-SAFE model
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 227
and its use to determine yields at the Landscape Test Sites. Unpublished report. Silsoe, Bedfordshire: Cranfield University. 53 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005) Burgess, P.J., Graves, A.R., Metselaar, K., Stappers, R., Keesman, K., Palma, J, Mayus, M., & van der Werf, W. (2004a). Description of the Plot-SAFE Version 0.3. Unpublished document. 15 September 2004. Silsoe, Bedfordshire: Cranfield University. 52 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005) Burgess, P.J., Seymour, I., Incoll, L.D., Corry, D.T., Hart, B. & Beaton, A. (2000). The application of silvoarable agroforestry in the UK. Aspects of Applied Biology 62:269-276. Burgess. P.J., Incoll, L.D., Corry, D.T., Beaton, A.. & Hart, B.J. (2004b). Poplar (Populus spp) growth and crop yields in a silvoarable experiment at three lowland sites in England. Agroforestry Systems 63: 157-169. Centre Régional de la Propriété Forestière (1997). Boiser une Terre Agricole. 28 pp. Commission of the European Union (2004). Proposal for a Council Regulation on support for rural development by the European Agricultural Fund for Rural Development (EAFRD). European Union Report No 2004.0161 (CNS). 68 pp. Conrad O (2002). DiGeM – Software for Digital Terrain Analysis. Accessed 4 April 2005. http://www.geogr.uni-goettingen.de/pg/saga/digem/index.html Dupraz C, Lagacherie M, Liagre F, Boutland A (1995). Perspectives de diversification des exploitation agricoles de la région Midi-Pyrénées par l’agroforesterie. Rapport de fin d’études commandité par le Conseil Régional Midi- Pyrénées. Institute National de la Recherche Agronomique, Montpellier. Contract AIR3 CT92-0134. 253 pp Dupraz C., 1998. Adequate design of control treatments in long term agroforestry experiments with multiple objectives. Agroforestry Systems, 43(1/3): 35-48. Dupraz C., Newman S., 1997. Temperate agroforestry : the European way. In : A. M. Gordon and S.M. Newman (editors), Temperate Agroforestry Systems, CAB International, Wallingford, UK, 181-236. European Commission (2005). FADN Public Database. Accessed 4 April 2005. http://europa.eu.int/comm/agriculture/rica/dwh/index_en.cfm Faustmann, M. (1849). Berechnung des Wertes Waldboden sowie noch nicht haubare Holzbestände für die Waldwirfschaft besitzen, Allgemeine Forst und Jagd- Zeitung, 25, 411-455. Global Data Systems (2005). Database of historical climate data compiled by Global Data Systems for the United States Department of Agriculture World Weather Board from World Meteorological Organisation climate reporting systems. http://hydrolab.arsusda.gov/nicks/nicks.htm (Accessed 5 May 2005). Graves AR, Matthews RB and Waldie K. (2004). Low external input technologies for livelihood improvement in subsistence agriculture. Advances in Agronomy 82: 473- 555 Graves, A.R., Burgess, P.J., Liagre, F., Dupraz, C. & Terreaux, J.-P. (2003). The development of a model of arable, silvoarable and forestry economics. Paper prepared for submission to Agroforestry Systems. Silsoe, Bedfordshire : Cranfield University. 30 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 228
Graves, A.R., Burgess, P.J., Liagre, F., Terreaux, J.P., & Dupraz, C. (2005a). Development and use of a framework for characterising computer models of silvoarable economics. Paper accepted by Agroforestry Systems. Graves, A.R., Burgess, P.J., Palma, J.H.N., Herzog, F., Moreno, G., Bertomeu, M., Dupraz, C. and Liagre, F. (2005b). Report on plot-economics of European silvoarable systems in target regions and the economic feasibility of silvoarable in target regions report. Silsoe, Bedfordshire: Cranfield University 11 March 2005. 42 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005) Herzog, F. (2004). Working visit report on upscaling for nine landscape test sites in Spain. Workshop at Plasencia, Spain 5-8 July 2004. Unpublished report Zurich : Agroscope FAL Reckenholz. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005) Institut pour le Développement Forestier (1997). Les noyer à bois. 3ème édition, Février 1997. 132 pp. Mead R and Willey RW (1980). The concept of a "Land Equivalent Ratio" and advantages in yields from intercropping. Expl. Agric. 16: 217-228 Mercer DE, Miller RP, Nair PKR and Latt CR (1998) Socioeconomic research in agroforestry: progress, prospects, priorities. Agroforestry Systems 38: 177-193 Ministerie van Landbouw, Natuur en Voedselkwaliteit (2004). Subsidieregeling Agrarisch Natuurbeheer. LASER vestging Roermond, Az Roermond, The Netherlands. 51 pp. Ministry of Agriculture, Fisheries and Food (1983). Definitions of terms used in agricultural business management. Alnwick, Northumberland: MAFF Publication Booklet 2269. 39 pp. Montero G and Cañella I (2000). Selvicultura de Pinus pinea L. Estado actual de los conociminetos en España. In: Simposio del pino piñonero (Pinus pinea l.). Valladolid, pp 21-38 Mücher, C.A., Bunce, R.G.H., Jongman, R.H.G., Klijn, J.A., Kooment, A.J.M. Metzger, M.J. and Wascher, D.M. (2003). Identification and characterisation of environments and landscapes in Europe. Alterra-rapport 832. Wageningen University, 119 pp. Mücher, CA (2000) PELCOM project. Final report submitted to the European Commission. Contract No ENV4-CT96-0315. 299 pp Nair PKR (1985) Classification of agroforestry systems. Agroforestry Systems 3: 97- 128 Nix, J. (2001). Farm Management Pocketbook. Ashford Kent: Wye College Press. 244 pp. Ong, C.K. (1996). A framework for quantifying the various effects of tree-crop interactions. In: Tree-Crop Interactions A Physiological Approach 1-23 Eds. C.K. Ong and P. Huxley. Wallingford: CAB International. Palma, J. & Reisner, Y. (2004). Work visit report on the upscaling of the seven landscape test sites in France. Unpublished report. Zurich: Agroscope FAL Reckenholz 15 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 229
Pulido, F.J. Campos, P and Montero, G. (2003). La gestión forestal de las dehesas. Historia, Ecológía, Selvicultura y economía. IPROCOR-Junta de Extremadura. Merida, Spain. Reisner Y (2004). Work visit report: upscaling for three landscape test sites in the Netherlands. 24-28 May 2004. Unpublished report. Zurich: Agroscope FAL Reckenholz. 9 pp. http://montpellier.inra.fr/safe/ (Accessed 5 May 2005) Sinclair FL (1999) A general classification of agroforestry practice. Agroforestry Systems, 46: 161-180. Souleres, G. (1992) les milieux de la populiculture, Institut pour le Développement Forestier, 310 pp. Thomas TH (1991) A spreadsheet approach to the economic modelling of agroforestry systems. Forest Ecology and Management 45: 207-235 Thomas TH and Willis RW (1997) Linking bio-economics to biophysical agroforestry models. Agroforestry Forum 8(2): 40-42 United States Department of Agriculture (2005). CLIGEN Weather Generator. United States Department of Agriculture Agricultural Research Service and Unites States Forest Service http://horizon.nserl.purdue.edu/Cligen/ (Accessed 5 May 2005). Van der Werf, W., Keesman, K., Burgess, P.J., Graves, A.R., Pilbeam, D., Incoll, L.D., Metselaar, K., Mayus, M., Stappers, R., Palma, J., Dupraz, C. & van Keulen, H. (2005) Yield-SAFE, a parameter sparse model for yield predictions, including uncertainty analysis, in European agro-forestry systems. Paper prepared for submission to Ecological Engineering. Van Genuchten, M. Th., (1980). A closed-form equation for predicting the hydraulic conductivity of unsaturated soils, Soil Science Society of America Journal, 44, 892- 898. Van Ittersum, MK and Rabbinge, R (1997). Concepts in production ecology for analysis and quantification of agricultural input-output combinations. Field Crops Research 52(3) 197-208 Willis, R.W., Thomas, T.H., and J. van Slycken, (1993) Poplar agroforestry: a re- evaluation of its economic potential on arable land in the United Kingdom. Forest Ecology and Management, 57, 85-97. Wösten, J.H.M., Lilly, A., Nemes, A., & Le Bas, C. (1999). Development and use of a database of hydraulic properties of European soils. Geoderma 90: 169-185. Yagüe, S. (1994). Producción y selvicultura del pino piñonero (Pinus pinea L.) en la provincia de Avila. Montes 36: 45-51.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 230
TABLES
Table 2. Summary of the latitude, longitude, altitude, mean air temperature, annual solar radiation receipt and annual rainfall at each site
Country and Site name Latitude Longitude Altitude Mean Solar Annual region temp radiation rainfall
(m) (°C) (MJ m-2) (mm)
Spain
Andalucia Alcala la real 37.36N 3.88W 1000 15.3 5490 355
Castilla La Torrijos 39.89N 4.39W 500 15.5 5560 348
Mancha Ocaña 39.94N 3.44W 700 14.7 5780 316
Almonacid de Zorita 40.23N 2.61W 900 12.6 6610 404
Castilla y Leon Cardenosa El Espinar 40.78N 4.53W 1000 12.0 5700 404
Fontiveros 40.86N 5.00W 900 12.0 6170 393
Olmedo 41.28N 4.80W 750 12.5 5480 410
St Maria del Campo 42.11N 3.91W 800 9.1 5630 530
St Maria del Paramo 42.44N 5.69W 800 10.2 6600 519
France
Poitou Charentes Champdeniers 46.41N 0.02E 200 11.0 4740 648
Centre Chateauroux 46.92N 1.65E 150 11.0 4750 587
Fussy 47.18N 2.47E 200 10.6 4800 626
Sancerre 47.30N 2.72E 400 10.7 4590 724
Franche Comté Champlitte 47.64N 5.58E 300 8.5 4940 773
Dampierre 47.61N 5.82E 300 10.0 5090 1072
Vitrey 47.81N 5.78E 400 9.5 4900 1084
The Netherlands
Balkbrug 52.57N 6.34E 0 8.9 4830 818
Bentelo 52.22N 6.67E 0 8.8 3690 729
Scherpenzeel 52.57N 6.34E 0 9.0 3710 801
Table 3. The total utilised agricultural area for each hypothetical farm and description of the 42 different land units, and the selected tree and crop species
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 231
Site Area Land Area Radiation Soil Soil Tree Crop rotation
of unit of (%) type depth
farm land (cm)
(ha) unit
(ha)
Spain
Alcala la real 73 LU1 58 97 M 140 Oak w/w/f
LU2 15 86 M 50 Oak w/w/f
Torrijos 63 LU1 10 101 M 140 Oak w/f
LU2 56 100 M 140 Oak w/w/f
Ocaña 66 LU1 66 100 M 140 Oak w/w/f
Almonacid de 66 LU1 59 97 M 140 Oak w/f
Zorita LU2 7 83 F 140 Oak s/s/s/s/s/w/f
Cardenosa El 58 LU1 23 93 M 140 Oak w/w/w/f
Espinar LU2 35 101 F 140 Oak w/w/w/f
Fontiveros 58 LU1 49 99 C 140 Oak w/w/w/w/f
LU2 9 98 C 140 Pine w/w/w/w/f
Olmedo 57 LU1 5 100 C 140 Pine w/s/f
LU2 34 100 M 140 Oak w/s/f
LU3 18 99 C 140 Oak w/s/f
St Maria del 58 LU1 44 99 C 140 Pine w/w/w/f
Campo LU2 14 99 M 140 Oak w/w/w/w/w/f
St Maria del 59 LU1 4 100 M 140 Oak w/w/w/s/f
Paramo LU2 34 100 M 140 Oak w/w/w/s/f
LU3 21 101 M 140 Oak w/w/w/s/f
France
Champdeniers 94 LU1 67 100 F 80 Cherry w/w/s/w/o/s
LU2 27 100 M 120 Walnut w/w/s/w/o/s
Chateauroux 152 LU1 32 102 F 80 Walnut w/w/o/w/o/s
LU3 86 102 M 120 Walnut w/w/o
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 232
LU2 23 102 F 40 Cherry w/w/o/w/o/s
LU4 11 100 F 40 Cherry w/w/o/w/o/s
Fussy 80 LU1 10 101 F 40 Cherry w/o
LU2 43 103 M 80 Poplar w/w/o
LU3 27 102 F 120 Cherry w/o
Sancerre 98 LU1 37 103 F 40 Cherry o/w/s/w/w/w/o
LU3 44 101 Vf 120 Cherry o/w/s/w/w/w/o
LU4 7 100 C 80 Cherry o/w/s/w
LU2 10 102 Vf 140 Poplar o/w/s/w/w/w/o
Champlitte 130 LU1 68 103 M 140 Cherry w/w/o
LU2 62 103 M-f 35 Walnut w/w/w/w/w/gm
Dampierre 130 LU1 64 98 M 140 Cherry w/w/gm
LU2 43 97 F 35 Cherry w/w/w/gm
LU3 23 95 Mf 60 Poplar w/gm
Vitrey 120 LU1 46 103 M 60 Cherry w/w/o
LU2 74 103 Mf 60 Poplar w/w/gm
The Netherlands
Balkbrugg 40 LU1 40 100 C 140 Poplar fm
Bentelo 40 LU1 40 100 C 140 Walnut w/w/fm
Scherpenzeel 10 LU1 10 100 C 140 Poplar fm
Note: Soil type: C: coarse; M: Medium; M-f: Medium-fine, F: Fine; V-f: Very fine
Crop type: w: wheat; f: fallow; o: oilseed; s: sunflower; gm: grain maize; fm: forage maize
Table 4 Four post-2005 grant scenarios assumed for silvoarable agroforestry
Arable payment Tree payment
Scenario 1 Percentage crop area in system None
Scenario 2 Total area of system Fifty percent costs in years 1-4
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 233
Table 5 Summary and description of yields for crops and trees in France, Spain and the Netherlands
Country Arable crop No of Mean Standard Range Coefficien value deviation t of s variation
(t ha-1) (t ha-1) (t ha-1) (%)
Spain Sunflower 120 0.8 0.4 0.2-1.7 52
Wheat 697 2.5 1.0 0.6-5.8 40
France Grain maize 61 6.3 1.2 2.9-9.8 20
Oilseed 260 3.2 0.4 1.9-4.3 13
Sunflower 106 1.7 0.4 0.7-2.6 26
Wheat 613 5.5 1.5 0.9-10.5 27 the Netherlands Forage maize 80 11.5 1.7 8.0-15.9 15
Wheat 20 7.9 1.2 5.9-11.1 16
Tree species (m3 ha-1) (m3 ha-1) (m3 ha-1) (%)
Spain Oak (60) 16 0.33 0.050 0.23-0.43 15
Pine (60) 3 0.25 0.005 0.25-0.26 2
France Cherry (60) 12 0.88 0.151 0.71-1.15 17
Poplar (60) 4 1.34 0.143 1.26-1.59 11
Walnut (60) 4 1.01 0.008 1.00-1.02 1 the Netherlands Poplar (20) 2 1.28 0.215 1.06-1.49 17
Walnut (60) 1 0.71 n/a 0.71 n/a
Note: values in brackets show length of rotation
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 234
Table 6 Relationship between a) crop yield and b) timber volume and soil depth for selected crop and tree species in Spain and France
Country Crop or tree No of Linear regression of crop yield (t ha-1) Correlation Significan species pairs or timber volume (m3 ha-1) against depth (d; coefficient t m) (P=0.05)
Spain Wheat 697 2.34 (± 1.02) + 0.19 d 0.02 No
France Wheat 613 3.69 (± 1.23) + 2.15 d 0.57 Yes
Grain maize 61 4.90 (± 0.92) + 1.82 d 0.67 Yes
Sunflower 106 1.06 (± 0.39) + 0.72 d 0.49 Yes
Oilseed 260 2.95 (± 0.42) + 0.32 d 0.26 Yes
Spain Oak 16 0.17 (± 0.04) + 0.12 d 0.53 No
France Cherry 12 0.64 (± 0.10) + 0.29 d 0.75 Yes
Walnut 4 1.02 (± 0.01) + 0.0028 d -0.12 No
Poplar 4 0.98 (± 0.04) + 0.42 d 0.97 No
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 235
Table 7 The predicted value of government support (€ ha-1), over a full tree-rotation (60 years for oak, pine walnut and cherry; 20 years for poplar), for forestry, arable and silvoarable systems in the pre-2005 grant regime, and the predicted change in that support in a post-2005 grant regime (scenario 1 and scenario 2)
Pre-2005 government support Predicted net change in support with the post-2005 grant regime
Land unit Rotation Forestry Arable Silvoarable Forestry Arable Silvoarable Silvoarable (a) scenario 1 scenario 2
Spain
Alcala 1 60 6860 5170 2010 -2940 8030 10010 11408
Alcala 2 60 6860 5170 2690 -2940 8030 9320 10728
Torrijos 1 60 9380 3870 1410 -4190 210 820 1256
Torrijos 2 60 9380 5170 1920 -4180 270 1790 2378
Ocaña 1 60 9380 5170 1770 -4190 350 2120 2864
Almonacid 1 60 9380 3870 1380 -4190 600 2010 2712
Almonacid 2 60 9370 8770 4080 -4180 -1030 2980 3886
Cardenosa 1 60 8860 5810 2900 -3380 -590 1850 2538
Cardenosa 2 60 8860 5810 2670 -3390 -590 2080 2768
Fontiveros 1 60 8850 6200 2940 -3380 950 3570 4430
Olmedo 2 60 8860 5160 2260 -3390 600 2990 3718
Olmedo 3 60 8860 6100 2520 -3380 -340 2720 3458
Campo 2 60 8860 6460 2610 -3380 1990 4160 6058
Paramo 1 60 8860 6760 3080 -3390 2500 5350 6402
Paramo 2 60 8860 6760 3080 -3390 2500 5350 6402
Paramo 3 60 8860 6760 3060 -3390 2500 5370 6422
Fontiveros 2 60 8000 6200 2060 -2960 950 4450 5335
Olmedo 1 60 8010 6100 1780 -2970 -340 3470 4223
Campo 1 60 8010 5810 1050 -2970 1790 2640 4263
France
Champdeniers 60 4440 21180 16130 0 1564 1 0 -390
Fussy 3 60 3840 21090 19590 0 -30 -430 1865
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 236
Sancerre 3 60 3840 20860 19380 0 -70 -460 1805
Fussy 1 60 3840 21090 19590 0 -30 -420 1865
Chateauroux 60 3840 21000 19510 0 1835 2 -50 -440
Chateauroux 60 3840 21000 19510 0 1835 4 -50 -440
Sancerre 4 60 3850 20940 19450 0 -150 -530 1735
Sancerre 1 60 3840 20860 19380 0 -70 -460 1805
Champlitte 1 60 0 20080 11880 0 -100 -60 2169
Dampierre 1 60 0 21040 15320 0 -1300 -1550 868
Vitrey 1 60 0 19840 14450 0 -60 -50 2759
Dampierre 2 60 0 20940 12700 0 -1200 -730 856
Champdeniers 60 4270 21180 16440 0 1567 2 0 -700
Chateauroux 60 3670 20800 19630 0 2027 3 150 -570
Chateauroux 60 3680 21000 19820 0 1837 1 -50 -750
Champlitte 2 60 0 19880 3920 0 100 20 3449
Sancerre 2 20 2720 6940 6850 0 -10 -540 610
Fussy 2 20 2720 6960 6870 0 60 -480 680
Dampierre 3 20 0 7080 3870 0 -500 -280 609
Vitrey 2 20 0 6600 3610 0 -10 -10 877
The Netherlands
Bentelo 60 11810 23000 5230 -1980 -1820 -410 2811
Balkbrug 20 11810 8000 3640 -3310 0 0 1026
Sherpenzeel 20 11810 8000 4370 -3640 0 0 1096
Note: Negative changes are shown in brackets
CAPTIONS FOR FIGURES
Figure 20 Predicted effects of tree species in a silvoarable system planted at a) 113 trees ha-1 and b) 50 trees ha-1 on the yield of the tree and the crop components relative to a monoculture (error bars show confidence intervals for mean values)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 237
Figure 21 Predicted relative crop yields and the timber volume for (a) an oak, b) a wild cherry, c) a walnut and a d) poplar silvoarable agroforestry systems (50 and 113 trees ha-1) for selected land units
Figure 22 Effect of crop species on the relative crop yield over a complete tree rotation, in (a) Spain and (b) France, under all tree species and in (c) the Netherlands under poplar, at 113 trees ha-1 and 50 trees ha-1 (error bars show the maximum and minimum values in each group)
Figure 23 Predicted land equivalent ratio in France, Spain and the Netherlands
Figure 24 Predicted land equivalent ratios for poplar, cherry walnut, oak and pine
Figure 25 Equivalent annual value (discount rate of 4%) without grants of the arable, forestry and silvoarable (113 trees ha-1) system in a) Spain, b) France and c) the Netherlands
Figure 26 Equivalent annual value (4% discount rate) of the arable, forestry and silvoarable (113 trees ha-1) system in a) Spain and b) France and c) the Netherlands, assuming the pre-2005 grant regime.
Figure 27 Equivalent annual value (4% discount rate) of a forestry, arable, and silvoarable (113 tree ha-1) system in a) Spain and b) France and c) the Netherlands, assuming the 2005 grant scenario 1 (error bars show the equivalent annual value for scenario 2)
Figure 28 Proportion of farms where the farm net present value was improved compared with the status quo by the introduction of silvoarable systems or forestry (Spain: n =17; France: n = 14)
Figure 29 Frequency with which silvoarable systems outperformed forestry (Spain: n = 17; France: n = 14; the Netherlands: n = 3)
FIGURES a) Relative yields for 113 trees ha-1 b) Relative yields for 50 trees ha-1
1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
Relative yield 0.2 0.2
0.0 0.0 Cherry Walnut Poplar Oak Pine Cherry Walnut Poplar Oak Pine
Relative tree yield Relative crop yield
Figure 1
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 238
a) Land unit 2, St Maria del Campo, Spain (oak; wheat/wheat/wheat/wheat/wheat/fallow)
2.0 1.0 /tree) 0.8 3 1.5 0.6 1.0 0.4 0.2 0.5 Relative crop yield 0.0 0.0 0204060 Timber volume (m 0204060 b) Land unit 1, Champdeniers, France (wild cherry; wheat/wheat/s/wheat/oilseed/sunflower)
2.0 1.0 /tree) 3 0.8 1.5
0.6 1.0 0.4 0.5 0.2 Relative crop yield 0.0 0.0 0204060 (m volume Timber 0 204060
c) Land unit 2, Champdeniers, France (walnut; wheat/wheat/s/wheat/oilseed/sunflower)
2.0 1.0 /tree) 3 1.5 0.8
0.6 1.0 0.4 0.5 0.2 Relative crop yield
0.0 Timber volume (m 0.0 0 102030405060 0204060
d) Land unit 1, Sherpenzeel, the Netherlands (poplar; forage maize)
2.0 1.0 /tree) 0.8 3
0.6 1.0 0.4
0.2 Relative crop yield
0.0 Timber volume (m 0.0 020 020 Time from tree planting (a) Time from tree planting (a)
Arable 50 trees/ha 113 trees/ha 50 trees/ha 113 trees/ha Forestry Figure 2
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 239
Figure 2
a) Spain b) France c) the Netherlands
1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2 Relative crop yield 0.0 0.0 0.0 Wheat Sunflower Wheat Oilseed Sunflower Grain Wheat Forage maiz e maize 113 trees 50 trees
Figure 3
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 240
a) 113 trees per hectare b) 50 trees per hectare
1.4 1.4 France 1.2 1.2
d Spain 1.0 Netherlands 1.0 0.8 0.8
0.6 0.6
0.4 0.4 Relative tree yiel tree Relative 0.2 0.2
0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.00.20.40.60.81.01.21.4 Relative crop yield Relative crop yield
Figure 4
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 241
a) 113 trees per hectare b) 50 trees per hectare
1.4 1.4 Poplar Cherry 1.2 1.2 Oak Pine 1.0 1.0 Walnut 0.8 0.8 0.6 0.6 0.4 0.4 0.2 Relative tree yield tree Relative 0.2 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 Relative crop yield
Relative crop yield
Figure 5
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 242
a) Spain )
-1 600 a Forestry -1 -1 400 Arable Silvoarable 200
0
-200 Alcala_1 Alcala_2 Ocaña_1 Campo_2 Campo_1 Torrijos_1 Torrijos_2 Olmedo_2 Olmedo_3 Paramo_1 Paramo_2 Paramo_3 Olmedo_1 Equivalent annual value (€ ha Almonacid_1 Almonacid_2 Fontiveros_1 Fontiveros_2 Cardenosa_1 Cardenosa_2 Oak Pine
b) France c) the Netherlands )
-1 600 600 a -1 -1 400 400
200 200
0 0
-200 -200 Vitrey_1 Vitrey_2 Fussy_1 Fussy_3 Fussy_2 Bentelo_1 Sancerre_1 Sancerre_3 Sancerre_4 Sancerre_2 Champlitte_1 Dampierre_1 Dampierre_2 Champlitte_2 Dampierre_3 Balkbrugg_1 Equivalent annual value (€ ha Chateauroux_2 Chateauroux_4 Chateauroux_1 Chateauroux_3 Champdeniers_1 Champdeniers_2 Scherpenzeel_1 Cherry Walnut Poplar Walnut Poplar
Figure 6
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 243
a) Spain )
-1 400 Forestry a -1 Arable 200 Silvoarable 0
-200 Alcala_1 Alcala_2 Ocaña_1 Campo_2 Campo_1 Torrijos_1 Torrijos_2 Olmedo_2 Olmedo_3 Paramo_1 Paramo_2 Paramo_3 Olmedo_1 Fontiveros_1 Fontiveros_2 Almonacid_1 Almonacid_2 Cardenosa_1 Cardenosa_2
Equivalent annual value (€ ha (€ value annual Equivalent Oak Pine
b) France c) the Netherlands )
-1 800 800 a 600 600 -1 400 400 200 200 0 0 -200 -200 -400 -400 -600 -600 -800 -800 Bentelo_1 Vitrey_1 Vitrey_2 Fussy_1 Fussy_3 Fussy_2 Balkbrugg_1 Sancerre_1 Sancerre_3 Sancerre_4 Sancerre_2 Dampierre_1 Dampierre_2 Dampierre_3 Scherpenzeel_1 Champlitte_1 Champlitte_2 Equivalent annual value (€ ha (€ value annual Equivalent Chateauroux_2 Chateauroux_4 Chateauroux_1 Chateauroux_3 WalnutPoplar Champdeniers_1 Champdeniers_2 Cherry Walnut Poplar
Figure 7
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 244
a) Spain )
-1 400 a Forestry -1 -1 200 Arable Silvoarable 0
-200 Alcala_1 Alcala_2 Ocaña_1 Campo_2 Campo_1 Torrijos_1 Torrijos_2 Olmedo_2 Olmedo_3 Paramo_1 Paramo_2 Paramo_3 Olmedo_1 Fontiveros_1 Fontiveros_2 Almonacid_1 Almonacid_2 Cardenosa_1 Cardenosa_2 Equivalent annual value (€ ha (€ value annual Equivalent Oak Pine
b) France c) the Netherlands )
-1 800
a 800
-1 -1 600 600 400 400 200 200 0 0 -200 -200 -400 -400 -600 -600 -800 -800 Vitrey_2 Vitrey_1 Bentelo_1 Fussy_2 Fussy_1 Fussy_3 Equivalent annual value (€ ha (€ value annual Equivalent Balkbrugg_1 Sancerre_2 Sancerre_1 Sancerre_3 Sancerre_4 Champlitte_2 Dampierre_3 Champlitte_1 Dampierre_1 Dampierre_2 Scherpenzeel_1 Chateauroux_1 Chateauroux_3 Chateauroux_2 Chateauroux_4 Champdeniers_2 Champdeniers_1 Walnut Poplar Cherry Walnut Poplar
Figure 8 ) 100 Silvoarable agroforestry Forestry
80
60
40
20
introducing silvoarable 0 agroforestry or forestry forestry or agroforestry Proportion of cases where cases where of Proportion Spain France Spain France Spain France increased farm profitability (% profitability farm increased
Pre-2005 Post-2005 scenario 1 Post-2005 scenario 2
Country and grant regime
Figure 9
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 245
100 80 60 40 20 0 than forestry (%) Spain Spain Spain France France France Proportion of cases where where cases Proportion of Netherlands Netherlands Netherlands agroforestry was more profitable 2004 grant scenario 2005 grant scenario 1.1 2005 grant scenario 2.2
Figure 10
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 246
ANNEX 10. Yield-SAFE: a parameter-sparse process-
based dynamic model for predicting resource capture,
growth and production in agroforestry systems
Wopke van der Werf1, Karel Keesman2, Paul Burgess3, Anil Graves3, David
Pilbeam4, L.D. Incoll4, Klaas Metselaar1,2, Martina Mayus1,7, Roel Stappers1,2,
Herman van Keulen5, João Palma6 & Christian Dupraz7
1Wageningen University, Group Crop & Weed Ecology, P.O. Box 430, 6700 AK Wageningen, The
Netherlands; 2Wageningen University, Systems & Control Group, P.O. Box 43, 6700 AA, Wageningen,
The Netherlands; 3Cranfield University, Silsoe, MK45 4DT, Bedfordshire, United Kingdom; 4School of
Biology, University of Leeds, Leeds LS2 9JT United Kingdom; 5Wageningen University, Group Plant
Production Systems, P.O. Box 430, 6700 AK Wageningen, The Netherlands; 6Agroscope FAL
Reckenholz. Swiss Federal Research Station for Agroecology and Agriculture, Reckenholzstrasse
191, CH-8046 Zurich, Switzerland; 7Institut National de Recherche Agronomique, UMR Systèmes de
Cultures Méditerranéens et Tropicaux, 2, Place Viala, 34060 Montpellier, France
Corresponding author:
Dr Wopke van der Werf, Wageningen University, Department of Plant Sciences,
Group Crop & Weed Ecology, P.O. Box 430, 6700 AK, Wageningen, The
Netherlands
Email: [email protected]
Tel.: +31 317 484 765
Fax: +31 317 484 892
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 247
ABSTRACT
1. Silvoarable agroforestry (SAF) is the cultivation of trees and arable crops on the same parcel of land. SAF may contribute to modern diversified land use objectives in
Europe, such as enhanced biodiversity and productivity, reduced leaching of nitrogen, protection against flooding and erosion, and attractiveness of the landscape. Long term yield predictions are needed to assess long term economic profitability of SAF.
2. A model for growth, resource sharing and productivity in agroforestry systems was developed to act as a tool in forecasts of yield, economic optimization of farming enterprises, and exploration of policy options for land use in Europe. The model is called Yield-SAFE; from “YIeld Estimator for Long term Design of Silvoarable
AgroForestry in Europe”. The model was developed with as few equations and parameters as possible to allow model parameterization under constrained availability of data from long term experiments.
3. The model consists of seven state equations expressing the temporal dynamics of:
(1) tree biomass; (2) tree leaf area; (3) number of shoots per tree; (4) crop biomass;
(5) crop leaf area index; (6) soil water content, and (7) heat sum. The main outputs of the model are the growth dynamics and final yields of trees and crops. Daily inputs are temperature, radiation and precipitation. Planting densities, initial biomasses of tree and crop species, and soil parameters must be specified.
4. A parameterization of Yield-SAFE is generated, using published yield tables for tree growth and output from the comprehensive crop simulation model STICS.
Analysis of tree and crop growth data from two poplar agroforestry stands in the
United Kingdom demonstrates the validity of the modelling concept and calibration
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 248
philosophy of Yield-SAFE. A sensitivity analysis is presented to elucidate which biological parameters most influence short and long term productivity and land equivalent ratio.
5. The conceptual model, elaborated in Yield-SAFE, in combination with the outlined procedure for model calibration, offers a valid tool for exploratory land use studies.
INTRODUCTION
Silvoarable agroforestry is the mixed cultivation of arable crops and trees on a single parcel of land. Interest in the introduction of trees in arable systems in Europe is increasing to diversify the farm landscape, promote biodiversity, enhance productivity, and benefit from the function of trees as windbreaks or as protection against nitrogen leaching, flooding and erosion. In recent years, European agricultural policy has sought to reduce arable surpluses and increase the number of trees planted on farms (Burgess et al., 2000). Unlike monoculture forestry, silvoarable agroforestry can provide an annual income. This is obtained from an arable intercrop which is grown for the initial or full duration of the tree rotation, depending in part on the tree density. In tropical countries, there are economic benefits from timber and non-timber tree products on arable land and the production of annual intercrops in plots planted with trees (Graves et al., 2004). Such diversification contributes to economic resilience to external fluctuations in markets. Tree-crop complementarity, leading to higher biomass production than in equivalent areas of arable or forestry monocultures (Droppelmann et al., 2000) lays a basis for higher economic returns.
To express the benefits of mixed cropping systems various characteristics have been proposed (Vandermeer, 1989). In the current analysis a choice has been made for the use of the Land Equivalent Ratio LER), first proposed by Mead and Willey (1980). LER
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 249
is defined as: ratio of the area needed under sole cropping to the area of intercropping at the same management level to obtain a particular yield. LER is calculated as the sum of the fractions of the yields of the intercrops relative to their sole crop yields:
II I LER =+++12... n [1] M12MMn where
I = yield of crop when intercropped
M = yield of crop as a monoculture
1 = one crop; 2 = another crop; n = nth crop
In agroforestry systems, which are characterized by differences in growing period of the component plant species of the mixture, many approaches are possible to calculate an integrated value of LER over multi-year periods. In this paper we calculate LER in two ways. The first method integrates productivity over the whole rotation, calculating LER as the sum of (1) average value of relative crop yield, compared to monocrop crop yields, and (2) cumulative timber production compared to the monoculture (Equation 2):
Sum of silvoarable crop yields Silvoarable timber volume LER =+ [2] rotation Sum of monoculture crop yields Monoculture timber volume
The second method of calculating LER produces an estimate for each year i in the tree rotation. This estimate is calculated as the sum of (1) the relative crop yield in year i (compared with monocrop crop yield in the same year) and (2) cumulative
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 250
timber production from year 1 through i (compared with monoculture tree growth from year 1 through i) (Equation 3).
Silvoarable crop yieldii Silvoarable timber volume LER i =+ [3] Monoculture crop yieldii Monoculture timber volume
Two of the key factors in determining adoption and maintenance of silvoarable systems are their profitability relative to alternative enterprises and their feasibility, in terms of the use of farm resources (Burgess et al., 2000; Graves et al., 2005b). The profitability of silvoarable systems, relative to pure arable agriculture and forestry, can be determined by comparing their net present value (NPV), calculated from cost-benefit analysis by discounting and aggregating future benefits and costs (Graves et al.,
2005a). The feasibility of the system, within a specific farm depends, among others, on the availability of and requirements for labour or finance. Fundamental to both assessments, is the need for biophysical data on yields of crops and trees in silvoarable as well as in arable and forestry systems. As empirical data on silvoarable systems are scarce, an alternative method is necessary to generate long-term time series of yields based on interactions of trees and crops in mixed systems. Such a method is the use of dynamic computer simulations that predict the effect of climate, tree and crop species, soil type and management choices on tree and crop production, economics and the environment.
The need for a minimal modelling approach. Key issues in the analysis of dynamic simulation models are stability, sensitivity of the output to parameter values, uncertainty propagation and identifiability. Identifiability analysis attempts to answer the question: can we estimate a unique value for specific parameter, given sufficient data? In general, identifiability decreases with increasing complexity of a model, because of the potential interactions between parameters. If, for a complex model,
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 251
poorly identifiable parameters are estimated from experimental data, errors in parameter estimates may become very large. As a consequence, uncertainty in model predictions will become large. Hence, from the viewpoint of restricting uncertainty in model predictions, a minimal modelling approach, allowing estimation of a maximum set of identifiable parameters, is preferred (Young, 1984; Ljung, 1987).
The need for a minimal modelling approach is high for agroforestry systems, because of the lack of quantitative long term data on the productivity of those systems.
Currently available biophysical models for agroforestry systems, such as WaNulCAS
(Van Noordwijk & Lusiana, 1999) and HyPAR (Mobbs et al. 1999) are highly complex and rich in parameters, and the above-mentioned drawbacks of complex models apply. As an alternative approach, a very parameter sparse, yet process-based model is proposed and presented here. The conceptual and algorithmic simplicity of this model, called YIELD-SAFE7, allow the application of powerful mathematical methods for parameter estimation, and the analysis of uncertainties in model predictions. The model can be easily adapted to different crops and environmental conditions by adjusting parameter values and input functions (Graves et al., this volume), and its code is compact enough to be included in agro-environmental modelling environments that aim at levels of aggregation above the field level
(Rabbinge & van Latesteijn, 1992; van Ittersum & Donatelli, 2001).
The ultimate goal of the YIELD-SAFE model is to predict dynamically site-specific long-term tree and crop yields under competitive conditions on the basis of historical or generated weather data, i.e. solar radiation, temperature and precipitation and relevant soil physical characteristics. Growth of trees and crops can essentially be
7 YIeld Estimator for Long term Design of Silvoarable AgroForestry in Europe
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 252
described as the conversion of primary resources, i.e. light, water and inorganic ions into useful organic material, and can therefore be described in terms of the availability of these resources and their utilization efficiency (Monteith, 1990). The objective of the current version of the model is to describe conditions where availability of plant nutrients is not a limiting factor for crop production, hence light and temperature as yield-determining factors and water as (possible) yield-limiting factor (van Ittersum and Rabbinge, 1997) are taken into account.
The objectives of this paper are:
- To describe and justify the conceptual background and equations of Yield-
SAFE;
- to provide the first calibration and validation of Yield-SAFE, using published
yield tables for poplar stands and two experimental data sets pertaining to the
growth of an agroforestry system with poplars and arable crops at two sites in
the United Kingdom.
- To provide a sensitivity analysis of Yield-SAFE.
MATERIALS & METHODS
MODEL DESCRIPTION
The objective of the YIELD-SAFE model is to describe the dynamics of competitive resource acquisition and the associated growth of the constituent components in an agroforestry stand with the minimum number of equations. Such an equation- and parameter-sparse approach is chosen because it provides the best chance that robust parameter values can be identified from experiments. Dynamic equations for the following state variables were identified as essential:
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 253
(1) biomass of tree
(2) leaf area of tree
(3) number of shoots of tree
(4) biomass of crop
(5) leaf area of crop
(6) heat sum
(7) available soil water
Biomasses of tree and crop are used to derive temporally-integrated timber volumes and crop yields. Leaf area of tree and crop are essential because they govern radiation capture, and thus the capacity for dry matter production and the associated water loss through transpiration. The number of shoots per tree is required because it governs the potential leaf area within a given year. By contrast the intra-annual leaf area dynamics (at the time scale of days to months) are primarily governed by the growth of leaf area per shoot. Available soil water is included to account for differential growth conditions across Europe with respect to the degree of water limitation, due to variation in precipitation, soil depth and water holding properties of soils. Finally, heat sum is integrated each season to define phenological development of the crops. Nutrient dynamics are not included, because of lack of information from existing agroforestry trials necessary to determine pertinent parameters. The model can be readily extended to include nutrient dynamics, e.g. by quantifying the minimum nutrient uptake required to produce calculated water-limited yields (cf. van Keulen & Wolf, 1988).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 254
Equations and associated parameters were developed as follows:
Potential tree growth
The potential growth rate of the woody biomass of the tree (Bt) is described as: dB Ifε ttt= [4] dt ρ where
Bt is the woody biomass of the tree (g dry matter per tree)
I is the global radiation, incoming to the forestry or agroforestry stand (MJ per m2 per day)
ft is the proportion of incoming radiation (I) intercepted by the trees
εt is the radiation use efficiency of the trees (g woody dry matter per MJ intercepted global radiation), and
2 ρ is the tree density (number of trees per m silvoarable area)
The variable t (italicized) is time (d), while the subscript t (in roman type) indicates parameters and variables for the tree.
The fraction of radiation intercepted by the trees in the agroforestry system is calculated as:
−kLtt fet =−1 [5] where
kt is the radiation extinction coefficient of the tree leaf canopy
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 255
2 2 Lt is the leaf area index of the tree stand (m tree leaf area per m silvoarable stand)
Water limited effective tree growth
Under water-limiting conditions, and accounting for biomass losses due to maintenance or attrition such as branch senescence and storm damage, Equation
[5] is modified into: dBIfwε tttt=−aB [6] dt ρ t
where wt expresses the relative effect of soil water potential on the tree growth rate.
This factor is calculated as:
pF≤= pF :w 1 ct pFPWP − pF pFcPWPt<≤ pF pF : w = [7] pFPWP− pF c pF>= pFPWP :w t 0 where pF is the soil water tension, defined as the negative log of the water potential in cm water. Hence as long as pF is below the critical value (pFc), there is no reduction, when pF is between the critical value and the permanent wilting point
( pFPWP ), the degree of reduction is proportional to the difference between current pF
and pFPWP as scaled by the difference between pFc and pFPWP , while the reduction is
100% when pF is greater than pFPWP (Fig. 1).
The product term aBt ensures that in due course, the growth rate of the tree will slow down until, ultimately, the tree will reach a maximum biomass. Outside the growing season, the rate of change of tree biomass is set to 0.
Water use by the tree
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 256
The amount of water that is used by the trees per unit area is calculated by multiplying the water-limited growth rate per tree by the tree density (ρ) and a
transpiration coefficient,γ t :
dB W = γρ t [8] ttdt where
3 2 Wt is the tree water use (m water per m silvoarable area per day)
3 γ t is the transpiration coefficient of the trees (m water per g woody dry matter)
Leaf area of the tree
The rate of increase in leaf area index of a tree leaf canopy ( Lt ) is calculated as: dL AA− t = ρN m [9] dt τ where
2 2 Lt is the leaf area index of the tree (m tree leaf area per m silvoarable area)
ρ is the density of trees (number of trees per m2 silvoarable area)
N is the number of shoots on a single tree (see below)
2 Am is the maximum leaf area per shoot on a tree (m )
A is the current leaf area per shoot on a tree (m2; see below)
τ is the time constant of the leaf unfolding process (day) as driven by re-allocation of reserve carbohydrates in the spring (Versteeg and van Keulen, 1986)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 257
The rationale for Equation 9 is that early leaf growth in trees is not an autonomous positive feedback process as in crop plants, governed by incident radiation interception, but a translocation and conversion process from reserve carbohydrates, stored at the end of the preceding season, to new leaf biomass. Hence, the dynamics are fundamentally different. The state variable N, the number of shoots on a tree, expresses the “memory” of the tree with respect to preceding year’s number of branches and storage of reserve carbohydrates. Leaves start to unfold at time tb, the date of bud burst and all leaf canopy is shed at the day of leaf fall (tf).
Number of shoots per tree
The number of shoots per tree is calculated on the basis of a saturating curvilinear
Monod function of tree biomass, according to:
Bt NN= m [10] Bt + K N where
Nm is the maximum number of shoots on a mature tree
KN is the biomass of a single tree at which the number of shoots is half the maximum
As KN is difficult to estimate from data, an expression for the growth of N was derived from which the parameter KN is eliminated.
From Equation 10 we derive:
NN− KB= m [11] N t N
After differentiation of Equation [10] and substitution of Equation [11] one obtains:
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 258
dNNNdB =−t 1 [12] ddttBNtm
with unknown initial condition N(t0) where t0 is the planting date of the trees. In practice, N(t0) is easier to estimate from experimental data than KN, hence this reformulation of the model. Furthermore, Equation 12 allows a straightforward adjustment of the tree growth in case of pruning.
Pruning and thinning
When pruning takes place, biomass and number of shoots are reduced by appropriate factors πB and πN, which can, in principle, be different. Thinning is effectuated by reducing tree density ρ by a thinning factor πρ.
Potential crop growth
Within each cropping season, crop biomass starts at an initial value of Bc(te) where te is the date of crop emergence. The subsequent potential growth rate of the crop is described as: dB c = I f ε [13] dt ccc where
2 Bc is the above-ground biomass of the crop (g dry matter per m silvoarable area)
2 Ic is the radiation, transmitted by the trees (MJ per m silvoarable area)
fc is the proportion of Ic intercepted by the crop
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 259
εc is the radiation use efficiency of the crop (g above-ground dry matter per MJ intercepted global radiation)
The radiation transmitted through the trees is calculated as:
Ict=−()1 fI [14] where
ft is the proportion of incoming global radiation intercepted by the tree crowns
I is global radiation, incoming to the agroforestry stand (MJ per m2 per day)
The fraction of radiation intercepted by the crop ( fc ) is calculated as:
L −k c c C fCc =−1 e [15] where
C is the proportion of the total area that is cropped (m2 cropped area per m2 silvoarable area)
kc is the radiation extinction coefficient of the crop
2 Lc is the leaf area index of the crop (m crop leaf area per silvoarable area)
Water-limited crop growth
Under water limiting conditions, Equation [13] is modified into: dB c = I fwε [16] dt ccc c
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 260
where wc expresses the reduction in crop growth rate, relative to the potential growth rate. This is calculated in the same way as the value for wt (Equation 7, Figure 1), but with crop specific parameter values for pFc and pFPWP.
Water use by the crop
Water use by the crop is calculated by multiplying the water-limited growth rate by a
transpiration coefficient, γ c :
[17]
dB W = γ c ccdt where
3 2 wc is the crop water uptake (m water per m silvoarable area per day)
3 γ c is the transpiration coefficient of the crop (m water per g above-ground dry matter. The value of γc can vary with crop type and the water vapour pressure deficit of air (VPD) (Loomis & Connor, 1992), but otherwise the value is relatively constant
(Monteith, 1990).
Leaf area of the crop
2 Change in leaf area index of the crop ( Lc ; m ) is calculated as: ddLB cc= σ P [18] ddtt where
σ is the specific leaf area of the crop (m2 leaf area per g leaf dry matter), and
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 261
P is the partitioning coefficient to leaves for the crop; i.e. the proportion of the daily increase in above-ground dry matter that is invested in growth of new leaves,
Leaf area starts at an initial value of Lc(te) where te is the date of emergence.
Leaf area growth is set to zero when the heat sum at harvest (Sh) is attained (see below).
Heat sum
The increase in cumulative temperature (heat sum) is calculated as: dS =−max[] 0,TTb [19] dt where
S is the heat sum since crop emergence (°C d)
T is daily average temperature (°C)
Tb is the base temperature for phenological development (°C)
The function max[ g]takes the maximum value of the arguments
Partitioning of dry matter to leaves in the crop
Partitioning of dry matter to leaves decreases linearly with crop development stage, according to:
SS≤=: PP 10 SS2 − SSS12<≤: PP = 0 [20] SS21− SS>=2 :0 P
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 262
where
P0 is the proportion of above-ground biomass initially partitioned to leaves
S1 is heat sum where partitioning of dry matter to leaves starts to decline
S2 is heat sum where the partitioning coefficient becomes zero.
Soil water dynamics
The model assumes a homogeneous soil of depth D (m) and volumetric water content θ, which is described by: d1θ =++−−−()R WFWWEirr gw c t act dtD
[21] where
θ is soil volumetric moisture content (m3 per m3)
R is precipitation (m3 per m2 silvoarable area per day)
3 2 Wirr is irrigation (m per m silvoarable area per day)
3 2 Fgw is drainage of soil water below the potential rooting zone (m per m silvoarable area per day)
3 2 Eact is actual soil evaporation (m per m silvoarable area per day)
Soil moisture characteristics are often described in terms of soil moisture tension, ψ, i.e. the force with which the soil matrix holds the water. For ease of notation, the tension is then expressed in terms of pF, where pF = log10(ψ), with ψ is expressed in
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 263
cm water tension. The relation between ψ and θ is given by the van Genuchten
(1980) equation:
m 1 θθ=+−PWP() θs θ PWP n [22] 1(+ αψ ) where
3 3 θs is soil water content at saturation (m per m )
θPWP is soil water content at permanent wilting point (the lower limit of plant-available water in m3 per m3)
α , m and n are soil-type specific parameters
ψ is soil water tension in cm water.
Precipitation and irrigation are introduced as forcing functions. Drainage flow to groundwater is dependent on the pF of the soil according to:
pF<= pFFC : FK gwδ s [23] pF≥= pFFC : F gw 0 where
pFFC is the pF value at field capacity, usually set to 2.3
Ks is soil hydraulic conductivity at saturation (m per day)
The factor δ is given by:
pF − FC δ =10 2 [24]
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 264
Where in practice, if soil water data is available, the factor 2 will be estimated between 1 and 4 depending on the water distribution in the soil, which depends on many factors, but especially on soil characteristics.
Evaporation from the soil surface (Eact) is calculated as:
EIw=η act s s [25] where
η is heat of vaporization (m3 water per MJ)
2 Is is the radiation incident on the soil (MJ per m per day)
ws is a factor accounting for the reduction in soil evaporation due to drying of the soil, and is calculated in the same way as the reduction factor for the tree (Equation 7;
Fig. 1)
Radiation incident on the soil (Is) is calculated as:
Istc=+−If () Cf(1 C ) [26]
Model implementation
The model has been implemented as a set of difference equations on several computer platforms including MatLab (Stappers et al., 2003) and Microsoft© Excel
(Burgess et al., 2004b). These references give further implementation details that are omitted here for clarity.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 265
POPLAR VALIDATION DATA
Two agroforestry experiments with poplar (Populus species) were carried out in the
United Kingdom. Full initial details of the experiments are provided by Burgess et al.
(2004a), but the key features are summarised here for clarity. The cooler and most northerly site is at the Leeds University Farms at Bramham near Tadcaster in West
Yorkshire (53°53’ N, 1°15’ W); the warmest site is at Silsoe in Bedfordshire (52°0’ N,
0°26’ W) in eastern England. Soils at the Leeds and Silsoe sites are sandy clay loam over limestone and clay over clay, respectively. At both sites the main experiment covered 2.5 ha and comprised three replicate blocks that included each combination of four poplar hybrids and three or four arable treatments.
Poplars were planted as unrooted sets in spring 1992 at a rectangular spacing of 10 m between tree rows (in a North-South orientation) and 6.4 m between trees within the rows. Part of the alleys between the tree rows were cropped yearly in the middle
8 m (leaving a 2 m uncultivated strip for the tree row), while another part of the alleys was left uncropped and weed free in subsequent years in order to obtain estimates of the yield of poplar in an agroforestry situation compared to a pure poplar stand at the same density. An area of the same field at least 15 m from the trees was used as an arable monocrop area. Starting in 1992, the rotation at Leeds comprised spring barley (Hordeum vulgare L.), peas (Pisum sativum L.), two crops of winter wheat
(Triticum aestivum L.), winter barley, spring mustard (Brassica alba L.), winter wheat, winter barley, two winter wheat crops, winter barley and winter oilseed rape (Brassica napus L.). At Silsoe, following poor crop yields in the initial three years, there were three winter wheat crops followed winter beans (Vicia faba L.), spring wheat, winter wheat, fallow, winter barley and spring beans. Crop management was the same for intercrop and monocrop. The poplar cultivars were Beaupré, Gibecq, Trichobel and
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 266
Robusta. From 1992 to 2000 the trees were regularly pruned, by removing the lower whorls of branches, in order to maintain a canopy depth equal to about half the tree height (Burgess et al., 2003).
Measurements
From 1992 onwards, the height of each tree in each arable treatment was measured after leaf fall. The diameters of the same trees were measured at breast height (1.3 m above the ground) each winter from 1994 onwards at Bramham and from 1995 onwards at Silsoe. Timber volume was estimated by first assuming the trunk is a perfect cylinder, with a volume calculated from height and diameter, and then multiplying this calculated volume by a form factor to account for taper of the trunk
(Burgess et al., 2004a). The form factor was derived from poplar yield tables, given in
Christie (1994).
Each year, grain, bean or pea yield within each poplar-hybrid x arable-treatment plot was determined by harvesting with a plot combine. Corresponding measurements were also taken within the monocropped control area.
MODEL CALIBRATION FOR POPLAR AND INTERCROPS
For the calibration of Yield-SAFE the following approach was used. First, the potential growth of monoculture stands of tree and crop species were fitted under specific climatic conditions in Europe, using yield tables for trees (e.g. Thomas et al.
1998) and validated model calculations for crops (Brisson et al., 2003). Potential growth is determined foremost by temperature (which drives developmental and phenological processes) and radiation (which drives photosynthesis) but is unaffected by water and nutrients as these are assumed to be non-limiting under the potential growth assumption (van Ittersum & Rabbinge, 1997). Second, given actual
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 267
monoculture yields of tree and crops as “reference” yields for a specific experimental site the model was fine-tuned by adjusting – within physiologically meaningful bounds
- the transpiration coefficient (γ) and harvest index (HI) and by introducing – if necessary – a management factor (between 0 and 1) that reduces the radiation use efficiency (ε). Hence, yield in agroforestry stands is predicted from the resulting model, which is – as described - calibrated to represent site-specific monoculture behaviour of trees and crops as affected by temperature and radiation driven growth potential in combination with site specific limitations due to water and nutrients, soil properties, and the local effects of weeds, pest, diseases and management shortcomings.
The calibration of model parameters for the potential growth of poplar trees was conducted using published yield tables for unthinned poplar (monoculture) stands with 8 x 8 spacing and a site class of 58 (Thomas et al., 1998). Because timber growth is expressed in terms of timber volume, it was necessary to convert the
3 -1 biomass yield into a timber volume. The timber volume of a tree (Vt; m tree ) was derived from:
HItimberB t Vt = [27] ρtimber
where HItimber is the proportion of the total woody-biomass partitioned to timber, and
-3 ρtimber is the density of the timber (g m ).
On the basis of practical identifiability analysis we decided to estimate the initial number of shoots, N(t0), and the radiation use efficiency, εt. Other parameter values were fixed at biologically plausible parameter values, based on literature (see results). Attempts to estimate additional parameters led to unreliable results and did
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 268
not improve the fit. A least-squares optimization algorithm was used to estimate both
N(t0) and εt.
As the crop data available related to harvested crop yield (Yc) rather than crop biomass, it was necessary to assume a crop harvest index (HIc).
YBccc= HI [28]
Simulation data from STICS (Brisson et al., 2003), given appropriate parameters for an Atlantic climate was used to provide potential growth curves for winter wheat.
In particular, the following parameters were adjusted: εc, S0 (heat sum after sowing when crop emerges), S1, S2, Sh (heat sum at harvest) and harvest index HIc. Again, a least-squares optimization was performed to identify the parameter values from the data.
MODEL VALIDATION FOR POPLAR AGROFORESTRY SYSTEMS
Given the calibrated parameters related to potential growth, in a second step only
three parameters: transpiration coefficient (γ t or γ c ), harvest index (HItimber or HIc) and a management factor were adjusted to fit actual yields (i.e. locally attained yields; van
Ittersum and Rabbinge, 1997) for the monoculture tree and crop systems at a specific site, in our case Silsoe (UK). The model was then used to predict tree and crop growth within a silvoarable system using these site-specific parameters, and these results were compared with experimental data collected over 12 years.
SENSITIVITY ANALYSIS
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 269
In this paper the objective of the sensitivity analysis is to investigate how different biophysical parameters influence the land equivalent ratio (LER). The model parameters were analyzed by systematically changing their nominal values by adding
±10%. The nominal values were obtained from the calibration of Yield-SAFE using the procedure described in the previous section. Then, after running the model with the perturbed parameters, the outputs were stored and the sensitivity was calculated from
∆y yp()()− yp = iM,, im [29] ∆−pppiiMim,,
where y(pi,M) and y(pi,m) denote the simulation model output (e.g. LER) when only the ith parameter is changed while keeping the others fixed at their nominal value. In order to avoid scale effects the relative sensitivity was calculated and used for analyses. The relative sensitivity, or elasticity (eLER), of LER for a specific parameter
pi, with nominal values pi and LER , is given by
∆LER pi eLER = . [30] ∆pi LER
This very simple type of sensitivity analysis provided a first indication of those parameters that dominate the output.
RESULTS
AGROFORESTRY EXPERIMENTS WITH POPLAR
During the first 12 years, the UK field experiments showed that poplar tree growth was reduced by the presence of arable crops, rather than a bare-fallow, between the rows of poplars (Fig. 2). The effect on timber volume per tree (or equivalently, per
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 270
hectare) was approximately minus 30% after 12 years of poplar growth, both in
Silsoe and in Leeds. Growth in Silsoe was marginally greater than in Leeds but the effect of crop competition on tree growth was similar at the two sites. During the initial nine years, the mean crop yield in the silvoarable system was 94% of the monoculture yield on a cropped area basis, and 75% on a total area basis, after allowing for the 20% of the area that was uncropped (Fig. 3). After the ninth year, relative crop yields started to decline substantially due to the cessation of pruning and the development of large tree canopies. A trend of the resulting LER is provided in Fig. 4, showing initially high values and a decline after nine years. Different ways of calculating LER, give different results. In Fig. 4, LER was calculated according to
Equation 3, that is by summing relative tree growth in SAF as shown in Fig. 1, and relative crop growth in SAF (Fig. 2). Initial calculations (results not shown) indicate that an annual LER, calculated as the sum of annual crop yields (normalized by comparison with monoculture) and the annual increment in timber volume (also normalized by comparison with monoculture) maintained stable values of the order
1.3-1.4 for any year in the experimental period.
MODEL CALIBRATION
Calibration, to represent these data, started with the calibration of the potential growth of a poplar forestry system under Atlantic growing conditions. The calibration was made on the basis of the development of timber volume for poplar with a site class of 58, assuming an unthinned stand of 8 m x 8 m (Thomas et al., 1998) and weather data from Orleans in France. The dynamic model parameters are described
-1 in Table 1, and the estimated model parameters for poplar were εt = 1.409 g MJ and
N(t0) = 0.6225 The timber volume calculated by the model was similar to that provided by the yield table (Fig. 5)
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 271
For the potential growth of e.g. winter wheat five parameters were obtained: εc = 1.34
-1 g MJ ; S0= 57 °Cd; S1= 456 °Cd; S2 = 464; Sh = 1312 °Cd and HIc = 0.51. HIc was derived directly from the simulation results; the other parameters by calibration.
Figure 5 presents the Yield-SAFE prediction of biomass growth in a monoculture wheat crop in Wageningen, using 1983/1984 Wageningen weather data, in comparison with the output from STICS.
The next stage was to calibrate the tree and crop components of the Yield-SAFE model for the specific conditions of Silsoe. For the tree component, the model was calibrated by assuming a timber volume per tree at the end of the tree rotation, in this case, of 30 years. At Silsoe, the increase in timber volume during the first 12 years matched that of the yield tables provided by Christie (1994) for an 8 m x 8 m poplar stand with a maximum mean annual increment of 13 m3 ha-1. Hence from the yield table, a reference timber volume of 2.41 m3 tree-1 was assumed for year 30. Using the Yield-SAFE model, and meteorological and soils data for Silsoe, the values of the transpiration coefficient and the harvest index were modified (Table 2) so that the model predicted a timber yield of 2.41 m3 tree-1 in year 30 (Fig. 7). The tree growth predicted by Yield-SAFE lags somewhat behind during early tree growth; this may partly be due to the assumption of a constant harvest index.
A continuous rotation of winter wheat was assumed for the crop component of the agroforestry system and a reference yield of 8.23 t ha-1 was derived from regional farm surveys. To obtain such a mean value over 30 years, it was necessary to modify the transpiration coefficient for the wheat to 0.316 m2 kg-1 (Table 2), which is within the plausible range for temperate conditions. It was not necessary to modify the harvest index. Thus, the model was calibrated to a site-specific reference yield using
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 272
eco-physiologically meaningful values for all the parameters. This is evidence that the model structure is eco-physiologically appropriate.
MODEL VALIDATION
The calibrated model was then run to calculate growth trajectories and yields (under water limitation) for crops and trees within a silvoarable system over a 30 year tree rotation. The predicted relative crop yields for the first twelve years (Fig. 7) generally matched the experimental results. This match between data and simulation results in the agroforestry situation provides further evidence for the validity for the modelling concept and calibration philosophy. Remember that the model was not fitted to any data from the agroforestry stand, but only to data from pure stands of crops or trees.
Thus, the rather good fit of the model to the yields in an actual agroforestry experiment provides evidence that it correctly captures the essence of the crop-tree interactions.
SENSITIVITY ANALYSIS
Using the Yield-SAFE model it was possible to predict the LER over a tree rotation of
30 years, using Equation 2. Assuming a continuous rotation of wheat the predicted
LER, at the end of the tree rotation of after 30 years, was 1.34. Perturbations of plus or minus 10% in the parameters used for this analysis resulted in values of LER ranging from 1.30 to 1.39 (Table 3). Thus, LER estimates by Yield-SAFE are moderately robust to parameter inaccuracies. The parameters kt, εt, N(t0) and Am had the greatest relative effect on LER (cf. Keesman et al., 2005). These tree parameters define to a large extent the shading of the tree on the crop.
A sensitivity analysis (Dennis & Schnabel, 1983) was also undertaken to determine how the elasticity of the LER to specific parameters changed during the tree rotation and with the light extinction coefficient. For this analysis, LER in a specific year was
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 273
determined using Equation 3. The default parameter sets, with varying values for the tree light extinction coefficient are given in Table 4. The results from both datasets matched the tree and crop growth during the first 12 years of the agroforestry stand, but resulted in a long term overestimation of tree growth, compared to yield tables of
Christie (1994). No water limitation was taken into account.
As a result of the different choice of nominal kt in the two parameter sets, different values are obtained for other parameters, notably those that affect the early growth of the tree: εt and the initial number of shoots, N(t0). The values of εt and N(t0) when kt was small (0.4) of 1.84 g MJ-1 and 1.32 respectively, were greater than the
-1 corresponding values of 1.09 g MJ and 1.075 when kt was large (0.8).
The elasticity analyses show that the most sensitive parameters were associated with the tree component of the model (Table 5). The importance of the tree parameters in determining the complementarity of resource use, as expressed by the value of LER, is also shown in a mathematical analysis by Keesman et al. (2005). Complementarity under potential growing conditions is entirely the result of the tree leaf canopy transmitting light that can be utilized by the crop component in the system. The maximum number of branches of the tree (Nm) has very low elasticity initially, but gains in importance as the trees grow. For mature trees, the maximum amount of shading by trees is determined in part by Nm; hence this parameter influences LER in a mature stand more than in a young stand.
The crop’s partitioning coefficient to leaves showed large sensitivity during the early years of the tree rotation. Surprisingly, some crop parameters attained greater relative importance to LER during the late years (20 and 25) of the tree rotation. For instance, in year 25, when the maximum number of shoots is (almost) achieved and
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 274
the shade is severe and the contribution of crop growth to LER small, the crop parameters light extinction (kc) coefficient and light use efficiency (εc) still become important. This is because, due to the large leaf area of the tree leaf canopy, changing the value of kt by a factor 0.002 (0.2 %) has only a small effect on the amount of light available for the crop. Given the shade condition, a small change (0.2
%) of the value of P, kc and εc (responsible for light interception and light use efficiency by the crop) has an impact on crop growth and LER. The effect is clearest at the greater nominal value of tree light interception (kt = 0.8; Table 5).
DISCUSSION
Compared to existing bio-physical agroforestry models (e.g. Mobbs et al., 1999; van
Noordwijk & Lusiana, 2000), the model proposed here is very simple. In support of this approach the following arguments can be given: a simpler model is often easier to parameterise and may produce more robust results; it is less work to build; and it is easier to explain and understand. This results in a shorter learning curve when the model is used in upscaling studies, and this may favour its inclusion in higher level studies, e.g. explorations of land use. Of course, a simple model may be underparameterised and unable to represent real situations using the few equations that were chosen as essential. We have not encountered data sets in which this is the case. This model was built with the philosophy that it could be extended when simulation of realistic situations required further detailing. This might be necessary, for instance, when agroforestry at different nutrient levels and nutrient limitation is simulated. However, the current set of parameters can represent many realistic situations without expanding the set of variables or equations, by simply adjusting values of parameters to specific conditions. For instance, the effect of nutrient limitation on growth rates can be captured in the value of the light efficiencies εc and
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 275
εt. Our philosophy with Yield-SAFE is that the model should keep its present simple structure until it is unable to represent real situations due to lack of structure or degrees of freedom. In this sense we follow Peters’ (1991) plea for simple, useful and predictive models in ecology.
In the current model version, the leaf area of the trees was assumed to spread out over the whole of the agroforested area, without explicitly accounting for clumping of tree leaf area in the tree crowns. Reasoning from existing literature on light distribution in crops (e.g. Goudriaan & van Laar, 1994) indicates that the extinction coefficient might change at low tree densities as the canopy is more heterogeneous.
Initial use of the model has suggested that it may be necessary to modify the light extinction coefficient in such situations. An alternative approach is to use detailed models on light distribution (e.g. Pronk et al., 2002) to estimate parameters for Yield-
SAFE. Likewise, detailed models for root distribution and activity in agroforestry might be used to parameterize Yield-SAFE functions for water capture by crops and trees.
During the same project an elaborate model was built for agroforestry system performance, based on details of resource use processes in agroforestry systems.
This model is called Hi-SAFE to indicate the high level of process detail contained in it. The applications of Hi-SAFE are more geared towards shorter time scales, and detailed questions regarding spatial configuration in agroforestry designs, whereas
Yield-SAFE focuses on issues of production and resource use in the longer term. For both models, parameter estimation is an issue. Yield-SAFE requires long term data on tree growth for parameter estimation and validation of model results. Such data are not yet available for agroforestry systems, but they may be come available in the future as the experiments that have been planted in the 1990s mature and accumulate timber. It is quite important that minimal data are collected in such
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 276
experiments to allow estimation of parameters of the model proposed here. In this respect it would be very helpful if records were taken of leaf area index and/or soil cover by the crop as well as the trees at different times during the season. Moreover, allometric relationships for widely-spaced trees are needed. At the present time, for studies on future land use, there is a pressing need for models that can be built with the limited information on agroforestry that is now available, as very few agroforestry systems have yet been planted in Europe. A simple model like Yield-SAFE can play a pivotal role in land use explorations by predicting production in agroforestry systems by integrating the vast information on forestry and arable systems, based on well proven eco-physiological principles, that – as this study shows – hold up as well in agroforestry as in agriculture and forestry.
ACKNOWLEDGEMENT
The research presented here was carried out as part of the collaborative research project SAFE: Silvoarable Agroforestry for Europe. Support for SAFE was provided by the Quality of Life Programme of the European Union (contract number QLK5-CT-
2001-00560). The silvoarable experiments in the UK were conducted with support from, what is now, the UK Department for the Environment, Food and Rural Affairs.
REFERENCES
Brisson, N., C. Gary, E. Justes, R. Roche, B. Mary, D. Ripoche, D. Zimmer, J. Sierra,
P. Bertuzzi, P. Burger, F. Bussière, Y.M. Cabidoche, P. Cellier, P. Debaeke, J.P.
Gaudillère, C. Hénault, F. Maraux, B. Seguin, H. Sinoquet, (2003). An overview of
the crop model STICS. European Journal of Agronomy 18, 309-332.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 277
Burgess, P.J., I. Seymour, L.D. Incoll, B. Hart & A. Beaton (2000). The application of
silvoarable agroforestry in the UK. Aspects of Applied Biology 62, 269-276.
Burgess, P.J., L.D. Incoll, B.J. Hart., A. Beaton, R.W. Piper, I. Seymour, F.H. Reynolds, C.
Wright, D.J. Pilbeam & A.R. Graves (2003). The impact of silvoarable agroforestry with
poplar on farm profitability and biological diversity. Final report to DEFRA Project Code
AF0105. Silsoe, Bedfordshire: Cranfield University. 63 pp.
Burgess, P.J., L.D. Incoll, D.T. Corry & B.J. Hart (2004a). Poplar (Populus spp) growth and
crop yields in a silvoarable experiment at three lowland sites in England. Agroforestry
Systems 63: 157-169.
Burgess, P.J., K. Metselaar, A.R. Graves, R. Stappers, K. Keesman, M. Mayus and W. van
der Werf (2004b). The SAFE-RESULT equations in Excel. Version 10. Technical report,
21 April 2004, Cranfield University, Silsoe, UK, 26 pp.
Christie, J.M. (1994) Provisional yield tables for poplar in Britain. Forestry Commission
Technical Paper 6. Forestry Commission, Edinburgh, 36 pp.
Dennis J. and Schnabel, R. (1996). Numerical methods for unconstrained optimization and
nonlinear equations. Society for Industrial and Applied Mathematics, Philadelphia, 378
pp.
Droppelmann, K.J., J.E. Ephrath & P.R. Berliner (2000). Tree/crop complementarity
in an arid zone runoff agroforestry system in northern Kenya. Agroforestry
Systems 50, 1-16.
Goudriaan, J. & H.H. van Laar (1994). Modelling potential crop growth processes.
Kluwer Academic Publishers, Dordrecht, The Netherlands, 238 pp.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 278
Graves, A.R., R.B. Matthews & K. Waldie (2004). Low external input technologies for
livelihood improvement in subsistence agriculture. Advances in Agronomy 82,
473-555.
Graves, A.R., P.J. Burgess, F. Liagre, J.-P. Terreaux & C. Dupraz (2005a).
Development and use of a framework for characterising computer models of
silvoarable economics. Agroforestry Systems, in press.
Graves, A.R., P.J. Burgess, J.H.N. Palma, F. Herzog, G. Moreno, M. Bertomeu, C.
Dupraz, F. Liagre, A. Koffeman and J. van den Briel (2005b). The development
and application of bio-economic modelling for silvoarable systems in Europe.
Submitted to this issue of Ecological Engineering
Keesman, K.J., W. van der Werf and H. van Keulen (2005). Mathematical Production
Ecology: Analysis of a Silvo-arable Agro-forestry System. Submitted to Bulletin
Mathematical Biology.
Ljung, L. (1987). System Identification: Theory for the User, 2nd ed. Prentice Hall, New
Jersey, 609 pp.
Loomis, R.S. & D.J. Connor (1992). Crop ecology; productivity and management in
agricultural systems. Cambridge University Press, 538 pp.
Mead, D. R., R.W. Willey (1980). The concept of a ‘land equivalent ratio’ and
advantages in yields from intercropping. Experimental Agriculture 16, 217-228.
Mobbs, D.C., G.J. Lawson, A.D. Friend, N.M.J. Crout, J.R.M. Arah & M.G. Hodnett
(1999). HyPAR Model for Agroforestry Systems. Technical Manual Model
Description for Version 3.0. DFID Forestry Research Programme R5652 Penicuik,
Edinburgh: Institute of Terrestrial Ecology, 113 pp.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 279
Monteith, J.L. (1990). Conservative behaviour in the response of crops to water and
light. In: Rabbinge, R., J. Goudriaan, H. van Keulen, F.W.T. Penning de Vries and
H.H. van Laar (Eds.), Theoretical Production Ecology: Reflections and Prospects.
Pudoc, Wageningen, pp. 3-16.
Peters, R.H. (1991). A Critique for Ecology. Cambridge University Press, 384 pp.
Powell, M. J. D. (1964). An efficient method for finding the minimum of a function of
several variables without calculating derivatives. Computer Journal. 7, 155 -162.
Pronk, A.A., J. Goudriaan, E. Stilma and H. Challa (2003). A simple method to
estimate radiation interception by nursery stock conifers: a case study of eastern
white cedar. Netherlands Journal of Agricultural Science 51, 279-295.
Rabbinge, R. and H.C. van Latesteijn (1992). Long-term options for land use in the
European community. Agricultural Systems 40, 195-210.
Stappers, R., K.J. Keesman and W. van der Werf (2003). The SAFE-RESULT
Equations: an Agro-Forestry Model. Technical Report, Oct. 2003, Wageningen
University, The Netherlands, 20 pp.
Thomas, T.H., P. Tabbush, M. Bulfin, T. Bradford, T. Kent, N. O’Dowd, P. Bonduelle,
J.M. Roda, A. Berthelot, D. Coaloa, P.M. Chiarabaglio, J. Bonany, F. Camps, J.
van Slycken, L. Meiresonne & R.M. Willis (1998). Poplars for Farmers. Final
Technical report. Appendix 3-2. AIR3-CT94-1753, European Commission DG12,
Brussels.
Vandermeer, J. (1989). The ecology of intercropping. Cambridge University Press,
Cambridge, MA, 237 pp.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 280
Van Genuchten, M.T. (1980). A closed-form equation for predicting the hydraulic
conductivity of unsaturated soils. Soil Science Society of America Journal 44,
892-898.
Van Ittersum, M.K. and R. Rabbinge (1997). Concepts in production ecology for
analysis and quantification of agricultural input-output combinations. Field Crops
Research 52, 197-208.
Van Ittersum, M.K. and M. Donatelli (Eds) (2001). Modelling cropping systems:
science, software and applications. Special Issue: European Journal of Agronomy
18(3-4), pp. 187-393.
Van Keulen, H. & J. Wolf (Eds) (1986). Modelling of agricultural production: weather,
soils and crops. Pudoc, Wageningen, 479 pp.
Van Noordwijk, M. & B. Lusiana (2000). WaNuLCAS version 2.0, Background on a
model of water, nutrient and light capture in agroforestry systems. International
Centre for Research in Agroforestry (ICRAF), Bogor, Indonesia, 186 pp.
Versteeg, M.N. and H. van Keulen (1986). Potential crop production prediction by
some simple calculation methods, as compared with computer simulations.
Agricultural Systems 19, 249-272.
Young, P.C. (1984). Recursive Estimation and Time Series Analysis. Springer
Verlag, Berlin, 300 pp.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 281
Table 1. Assumed and estimated tree dynamic model parameters for poplar.
Symbol Description Value Units
Assumed parameters
-3 ρt Timber density 410000 g m kt Light extinction coefficient 0.8 -
ρ Tree stand density 0.0156 m-2
Nm Maximum number of shoots per tree 10000 -
2 Am Maximum leaf area per shoot 0.05 m a Attrition rate of standing tree biomass 0.0001 d-1
τ Time constant of leaf area growth 10 d
HItimber Proportion of woody biomass partitioned to 0.5 - timber
Estimated parameters
-1 ε t Radiation use efficiency 1.409 g MJ
Initial number of shoots per tree 0.6225 Nt()0
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 282
Table 2. Reference yields and calibrated values for transpiration coefficient and harvest index for poplar and wheat at Silsoe. The calibrated management factor was
1 for both species.
Specie Time of Reference Reference Calibrated Calibrated s clear fell yield at crop yield transpiration harvest clear fell coefficient index
(year) (m3 tree-1) (t ha-1 a-1 (m3 kg-1) (%)
Poplar 30 2.41 - 0.420 48.6
Wheat - - 8.23 0.316 51.0
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 283
Table 3: The effect of a ± 10% change in selected parameters in the Yield-SAFE model on the
predicted tree and crop yields, and land equivalent ratios (LER) for a poplar silvoarable
system in year 30 (LER calculated with Equation 3).
Monoculture Silvoarable Sensitivity
Nominal Tree Crop Tree Crop LER (normalized value of yield yield yield yield main effects: parameter (m3 ha-1) (t ha-1) (m3 ha-1) (t ha-1 ) ∆LER/LER*p/∆p)
Reference 377 247 345 104 1.34
Tree parameters kt 0.8 334 na 302 120 1.39
408 na 377 92 1.30 -0.36
εt 1.4086 345 na 316 114 1.38
402 na 369 95 1.30 -0.28
Am 0.05 350 na 319 114 1.37
399 na 367 96 1.31 -0.24
N(t0) 0.6225 352 na 321 113 1.37
397 na 365 97 1.31 -0.23
γt 0.00042 409 na 375 106 1.35
350 na 320 102 1.33 -0.08 pFc 4 369 na 325 110 1.33
361 na 332 102 1.33 0.03
Nm 10000 374 na 342 105 1.34
379 na 347 103 1.33 -0.02
HItimber 0.486 340 na 311 104 1.34
340 na 311 104 1.34 0.00
Crop parameters
Sh 1312 na 237 345 101 1.34
na 262 344 109 1.33 -0.05 pFc 2.9 na 237 352 95 1.34
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 284
na 255 339 110 1.33 -0.02
εc 1.34 na 233 352 93 1.33
na 256 337 114 1.34 0.01
S0 57 na 247 345 104 1.34
na 246 346 103 1.34 0.00
HIc 0.51 na 222 345 94 1.34
na 272 345 114 1.34 0.00
γc 0.00032 na 269 349 110 1.34
na 228 341 98 1.34 0.00
na = not applicable
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 285
Table 4: Parameter setting and initial conditions (after calibration) for a sensitivity analysis of Yield-SAFE for a poplar agroforestry stand (156 trees ha-1) with continuous wheat.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 286
Parameter Parameter Component Symbol Unit set 1 set 2
-1 g MJ 1.84 1.4086 Tree εt - 0.4 0.8 kt m2 0.05 0.05 Am d 10 10 τ - 0 0.0001 a
-1 1.32 0.6225 N(t0) tree g tree-1 100 100 Bt(t0) m2 tree-1 Lt(t0) 0 0 tree-1 8000 10000 Nm Day of t 100 100 b year
Day of 265 300 t f year
g MJ-1 Crop εc 1.6 1.6 - 0.7 0.7 kc m2 g-1 0.02 0.02 σ - 0.8 0.8 P oC 0 0 T0 oC d 150 150 S0 oC d 160 160 S1 oC d 2350 2350 S2 oC d Sh 2950 2950 - 0.1 0.1 Lc(t0) g m-2 10 10 Bc(t0) Day of 280 280 t s year
Day of 235 235 t h year
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 287
Table 5. Ranking of elasticities of land equivalent ratios in years 2, 10 and 25 of a poplar-wheat agroforestry stand to biological parameters of tree and crop, for tree parameter scenario’s based on an assumed coefficient of light extinction kt of 0.4 and
0.8. Ranking per column, i.e. over all parameters in a given year, with the first rank
(1) for the most sensitive parameter.
Parameter set 2 (kt = Parameter set 1 (kt = 0.4) 0.8)
Component Year Year 10 Year 25 Parameter Year 2 Year 10 Year 25 2
Tree tb 1 1 1 1 2 1
Crop P 2 8 7 2 8 3
Tree kt 3 3 2 3 3 6
Tree Am 4 4 3 4 4 5
Tree N(t0) 5 5 9 5 5 9
Tree tf 6 2 5 6 1 7
Tree Bt(t0) 7 7 11 7 7 11
Tree εt 8 6 10 8 6 10
Tree τ 9 10 12 9 9 12
Crop kc 10 11 6 10 10 2
Crop εc 11 12 8 11 11 4
Crop S1 12 16 13 12 14 15
Crop S2 13 14 14 13 13 13
Crop Bc(t0) 14 13 15 14 15 14
Crop Lc(t0) 15 15 16 15 16 16
Crop σ 16 17 17 16 17 17
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 288
Tree Nm 17 9 4 17 12 8
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 289
1 ) c w (
0
Relative growthrate coefficient 012345 pF
Figure 1. Relationship between the reduction factor for the rate of crop growth ( wc ) and the pF of the soil (pFc = 2.9 and pFPWP = 4.2).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 290
0.4 ) 3 0.35 Silsoe monoculture Leeds monoculture 0.3 Silsoe agroforestry
0.25 Leeds agroforestry
0.2
0.15
0.1
0.05 Timber volume per tree (m
0 1992 1994 1996 1998 2000 2002 2004
Year
Fig. 2: Growth of poplar in agroforestry stand and monoculture Silsoe (UK) and
Leeds (UK), 1992-2003.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 291
1.0 0.8 0.6 0.4 Silsoe Leeds
(per total area) 0.2 Relative crop yield 0.0 1992 1994 1996 1998 2000 2002 Year
Fig. 3: Relative yield of crops in agroforestry stands at Silsoe (UK) and Leeds (UK),
1992-2003. Yield in the intercrop is expressed as a proportion of yield in monocrop.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 292
2 Silsoe 1.8 1.6 Leeds 1.4 1.2 1 LER 0.8 0.6 0.4 0.2 0 1992 1994 1996 1998 2000 2002 2004
Year
Fig 4. Evolution of the annual land equivalent ratio at Silsoe (UK) and Bramham
(UK), 1992-2003. Annual land Equivalent Ratio is calculated as the sum of crop
yield in any year and the cumulative tree growth up to the same year, both
normalized by their productions in monoculture (Equation 3).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 293
700
) 600 -1 500 ha 3 400 300 200
Volume (m Volume 100 0 0 2 4 6 8 101214161820 Time from tree planting (a)
Figure 5. Potential poplar growth in the Atlantic region, simulated with Yield-SAFE.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 294
) 30 -1 25 20 15 10 5
Biomass (t ha (t Biomass 0 0 100 200 300 400 Time from planting (d)
Yield-SAFE STICS
Figure 5. Total crop biomass predictions (wheat) from Yield-SAFE (dashed line) calibrated to outcomes from the comprehensive crop growth model STICS (drawn line). Weather data from Wageningen, 1984.
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 295
Yield-SAFE forestry 2.0 Silsoe forestry measured /tree) 3 Silsoe forestry predicted YC=13
1.0 Timber volume (m 0.0 0102030
Fig. 7: Calibration of Yield-SAFE: Model prediction of tree growth in a poplar agro- forestry stand, compared to yield tables (YC 13; Christie, 1994) and tree growth in the forestry treatment at Silsoe (1992-2003).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 296
1.0
0.8 Arable Silvoarable 0.6 Silsoe 0.4 Leeds
Relative cropyield 0.2
0.0 0102030 Time from tree planting (a)
Fig. 8: Validation of Yield-SAFE: model prediction of relative yield of continuous winter wheat, compared with monoculture wheat yield, in a poplar agroforestry stand (156 trees ha-1), compared to observed relative crop yields in Silsoe and Leeds agroforestry experiments, 1992-2004 (open symbols).
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 297
Appendix A. Variables and parameters in Yield-SAFE.
Symbol Units Meaning
State variables
-1 Bt g tree Dry mass of the trunk and branches of the tree
2 -2 Lt m m Leaf area index of trees, i.e. tree leaf area per area silvoarable system
Nt - Number of shoots per tree
-2 Bc g m Above-ground dry mass of the crop per area of the silvoarable system
2 -2 Lc m m Leaf area index of crop, i.e. crop leaf area per area of silvoarable system
θ m3 m-3 Volumetric water content of the soil
S °C d Heat sum since crop emergence
Tree parameters
-1 εt g MJ Radiation use efficiency of the trees, i.e. woody biomass produced per unit intercepted short-wave radiation kt - Light extinction coefficient of the trees
3 -1 γt m g Transpiration coefficient of the trees, i.e. water transpired per unit of woody dry matter produced
2 Am m Maximum leaf area of a single tree shoot
τ d Time constant of leaf area growth of a tree shoot a d-1 Relative rate of attrition of standing tree biomass
Crop parameters
-1 εc g MJ Radiation use efficiency of the crop, i.e. above-ground dry biomass production per unit of intercepted total short- wave radiation kc - Light extinction coefficient of the crop
3 -1 γc m g Transpiration coefficient of the crop; i.e. water transpired per unit of above-ground crop dry biomass
2 -1 σ m g Specific leaf area of crop; i.e. leaf area per mass of dry matter
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 298
matter
Sh °C d Heat sum at crop harvest
Tb °C Base temperature for crop phenological development
- P0 Initial partitioning factor to leaves
S1 °C d Heat sum at which partitioning to leaves starts to decrease
S2 °C d Heat sum at which partitioning to leaves ceases
Soil parameters
pFPWP - Log of soil water tension expressed as cm of water at permanent wilting point
pFFC - Log of soil water tension expressed as cm of water at field capacity m, n - Shape parameters of the van Genuchten equation describing the (θ, ψ) function
-1 Ks m d Soil hydraulic conductivity at field capacity
D m Depth of the soil compartment
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 299
Appendix A. Variables and parameters in Yield-SAFE (continued)
Symbol Units Meaning
Intermediate variables
P g g-1 Partitioning coefficient of above-ground dry matter to leaves
-2 Ic MJ m Radiation underneath the tree leaf canopy per area of silvoarable system
fc - Proportion of radiation incident on crop intercepted by crop
wc - Coefficient (0-1) expressing response of crop growth rate to water shortage
ft - Proportion of incident radiation intercepted by trees
wt - Coefficient (0-1) expressing response of tree growth rate to water shortage ws Coefficient (0-1) expressing response of soil evaporation to water shortage
ψ cm water Water tension of soil pF - Water tension of soil using a log scale in pF-units: log10(ψ)
δ - Parameter affecting drainage rate below root zone
Physical constants
η g MJ-1 1/heat of vaporization
Forcing functions
Ι MJ m-2 Daily total short wave radiation
Τ °C Daily mean temperature
R m3 m-2 Daily precipitation
Management functions
C m2 m-2 The cropped area expressed as a proportion of the total silvoarable area
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 300
ts DOY Crop sowing date (for each year in the tree cycle)
ρ trees m-2 Tree stand density
πt - Proportion of trees thinned (time-dependent)
πb - Proportion of tree biomass pruned (time-dependent)
πs - Proportion of tree shoots pruned (time-dependent)
Initial conditions
-1 Ν(t0) tree Number of shoots on a newly planted tree
Note: DOY is Day of Year
SAFE Final Progress Report – Volume 4 (Annexes) – May 2005 301