Calibration and Validation of CROPGRO-cowpea Model of DSSAT for Four Cowpea Varieties under Drought Stress

The Decision Support System for Agrotechnology Transfert (DSSAT) is a software comprising crop simulation models. The aim of this study was to: Calibrate the CROPGRO-cowpea model of DSSAT for four cowpea varieties, and Validate the model using a data set collected under drought stress For this purpose, experiments were carried out in 2020 and 2021 dry seasons in Burkina Faso at Kamboinsin in the Centre of Environmental and Agricultural Research and Training (CREAF). The model calibration was done using data including days to 50% flowering, days to maturity, above-ground biomass, and grain yield per hectare collected from an experiment without water and nutrient stress, and the GenCalc software was used for estimating the genetic coefficients of the varieties. Data from two drought stress treatments, such as D1: drought at seedling stage, and D2: drought at flowering stage, were used for validating the model. The results of the calibration showed that the model excellently simulated the days to flowering with a normalized root mean square error (nRMSE) of less than 10% and a high degree of agreement for all the varieties. The simulation of the days to physiological maturity was excellent for 50% of the varieties (nRMSE<10%) and good for the others (10 < nRMSE < 20%). The simulation of the grain yield per hectare ranged from excellent to good. Poor prediction of the above-ground biomass was attained for 75% of the varieties (nRMSE ≥ 40%), while the fair simulation was recorded for 25% (nRMSE=27%) during the model calibration. The statistics of the validation process showed an excellent simulation of the days to flowering (nRMSE= 1.85%; R2=0.98; d-index=0.99) and a good prediction of the days to maturity (nRMSE=13.82; R2=0.87; d-index=0.53) for all the varieties. Simulated above-ground biomass was in poor agreement with the observed values (nRMSE=106.62%; R2=0.92; d-index=0.36). Fair prediction of the grain yield by the model was achieved during the validation (nRMSE = 27%). From these results, it can be concluded that the DSSAT model can be considered an efficient tool for predicting cowpea phenology, growth, and yield in optimum conditions of development. However, in drought stress conditions, the sensitivity or tolerance status to the drought of the variety can reduce the accuracy of the grain yield prediction by the model.