HAL Id: hal-02734782
https://hal.inrae.fr/hal-02734782
Submitted on 2 Jun 2020HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
Calibration of the 3D Hi-sAFe agroforestry model for
hybrid walnut
Kévin Wolz, Christian Dupraz, Isabelle Lecomte, Marie Gosme, Francesco
Reyes
To cite this version:
Kévin Wolz, Christian Dupraz, Isabelle Lecomte, Marie Gosme, Francesco Reyes. Calibration of the 3D Hi-sAFe agroforestry model for hybrid walnut. 4. World Congress on Agroforestry, May 2019, Montpellier, France. 933 p., 2019, Book of abstracts. 4th World Congress on Agroforestry. �hal-02734782�
845
Agroforestry 2019- Poster L24 Modelling
4th World Congress on Agroforestry
Strengthening links between science, society and policy 20-22 May 2019Le Corum, Montpellier, France Book of Abstracts L24.P.17
Calibration of the 3D Hi-sAFe agroforestry model for hybrid walnut
Wolz K. J. (wolzkevin@gmail.com), Dupraz C., Lecomte I., Gosme M., Reyes F. UMR-SYSTEM, INRA, Montpellier, France
Hi-sAFe is a 3D biophysical model designed to explore the interactions between tres and crops in agroforestry systems. Utilization of any tree species within Hi-sAFe requires parame-terization and calibration of the model for that tree species. We carried out these three steps for hybrid walnut (Juglans regia x nigra). Data used for parameterization and calibration came from the literature and long-term experimental plots at the Restinclières Estate near Montpel-lier in Southern France (Inurreta-Aguirre et al. 2018). The three plots included two alley crop-ping systems of hybrid walnut and durum wheat (Triticum durum) and a pure forestry system of hybrid walnut with a natural ground cover. A total of 21 Hi-sAFe tree inputs were unable to be parameterized directly and were included in the model calibration. Optimal parameter values were searched for using a multi-objective evolutionary algorithm (MOEA) similar to the NSGA-II algorithm (Deb et al. 2002). The root mean square errors of the measured vs. modeled diameter at breast height (DBH) in each plot were used as objective functions to minimize. The final Pareto optimal front of the MOEA contained a diversity of tree parameter sets. Accuracy of modeled crop relative yield in the A2 plot was used as a final criterion to select the best pa-rameter set. These calibrated Hi-sAFe papa-rameters for hybrid walnut can be used to confidently explore hybrid walnut agroforestry systems across pedoclimatic and management conditions.
Fig. 1 Measured vs. modeled DBH of hybrid walnut in two agroforestry systems (A2 and A3) and one pure forestry system (A4) in Southern France. Boxplots show the range of measured data for
individual trees each year. Lines show Hi-sAFe model predications after model calibration.
Keywords: alley cropping, silvoarable, model, genetic algorithm, optimization.
References:
1. Deb K, Agrawal S, Pratap A, Meyarivan T (2002) IEEE Trans Evolutionary Computation 6:182– 197.
2. Inurreta-Aguirre HD, Lauri P-É, Dupraz C, Gosme M (2018) Agrofor Syst 1–14. doi: 10.1007/