ÉVALUATION DE LA PERFORMANCE DU MODÈLE AQUACROP POUR LA PRÉVISION DES RENDEMENTS DU BLÉ TENDRE (TRITICUM ÆSTIVUM) AU MAROC (PROVINCE DE MEKNÈS)

Abstract

ABSTRACT Worldwide, agricultural lands located in arid and semi-arid areas such as Morocco are strongly dependent on weather for the rainfall agriculture. In fact, 80 percent of yield losses are explained by climatic variation which were mainly recorded as a crucial lack of rainfall. Heretofore, cereal yields forecasts are estimated based on meteorological variables and satellites (vegetation monitoring) by CGMS (Crop Growth Monitoring System - Morocco). Three variables, selected according to their level of performance, are plotted into a multiple regression equation to predict the cereal crop yield. This study was designed to estimate performance in grain of soft wheat (Triticum aestivum) using a different model : AquaCop developed by FAO (Food and Agriculture Organization - United Nations). Based on the transfer of water among soil, plant and the atmosphere, it requires a set of meteorological, phenological and soil data. The model was set up, calibrated and validated for prediction of yields, in the province of Meknes (North - West of the country) on a database between 1998-2016. Firstly, simulations were conducted across the all grid (space entity of 20.25 km²) contained in Meknes province, secondly aggregated at the provincial level. The grid results displayed coefficient of determination (R²) of 0.47 and a root mean square error (RMSE) of 0.82 T/ha. For the province, the results are R² = 0.74 et le RMSE = 0.56 T/ha Suggestions for improvement of the input data have been proposed in order to achieve more precise results. In addition, in order to improve the control on soft wheat cultures, other output software could also be calibrated and validated such as the growth of the biomass, the evolution of water content in the soil or even the root growth in the crop year.