Evaluation of DSSAT model for potential yield prediction of potato under autumn cropping system (Case study: Jiroft, Iran)

Crop Simulation Models are advanced tools to estimate crop yield and optimizing of crop management practices. This study was conducted in order to evaluate DSSAT model under autumn cropping system condition in Jiroft, Iran, 2012-2014. The field experiment which repeated in two sequential years and three replications were performed as randomized complete block design in split plots. The main factor was planting date (17th, 23th, 29th September and 5th October) and its subfactor was potato cultivars (Sante, Satina and Boren). Data of first and second years of field experiment was applied respectively for calibration and evaluation of model. Data-base requirement for model was created from 1. climatic data includes: temperature, radiation, relative humidity, wind and precipitation; 2. crop data attained from field experiment, 3. soil data. Results of statistical evaluation of model showed the good fitness of simulated and actual yield performance. Simulated tuber yields were similar to their observed value with RMSE of 19% (1210 kg/ha) and high correlation between observed and simulated tuber yield (R2>0.9). Simulated biomass had 2673.5 kg/ha difference to its observed value that showed low estimation precision (RMSE>30%). This model couldn’t predict tuber initiation stage correctly (RMSE>30%). Therefore, it was concluded that DSSAT model predicted potato yield correctly but due to weak precision in estimation of phenological stage, had low assurance for use in autumn cropping system of potato under autumn cropping system in Jiroft region.