ABSTRACT
The performance of the Decision Support System for Agrometeorology Transfer (DSSAT) v4.6 model was assessed for two major rice-growing varieties (IR 36 and Swarna) over selected stations, viz. Balaghat, Jabalpur, Narsinghpur, and Seoni in Madhya Pradesh, India. Drought years in these regions were identified using India Meteorological Department (IMD) rainfall data sets during monsoon season of June, July, August, and September (JJAS), 1990–2011 and different sowing dates were chosen, respectively. It was found that the DSSAT model predicted crop yield higher than the observed yield. Furthermore, a bias correction and detrended yield anomaly analysis was carried out; it suggested that the model followed an observed pattern during most of the drought or deficit phases. Additionally, a sensitivity analysis was carried out to examine the model sensitivity 208with respect to variations in daily rainfall and genetic coefficients of the crop, affecting crop yield. Remote sensing data have also been used to compare the model, which estimates that daily soil moisture is closer to European Space Agency (ESA)–derived soil moisture. The model-simulated evapotranspiration (ET) reflects a minor gap with the remote sensing observations. The relationship of weak phases of rainfall over central India to real-time multivariate (RMM) indices of Madden–Julian oscillation (MJO) has been examined. RMM-6, RMM-7, RMM-1, and RMM-2 describe the drought conditions over central India. However, the frequency of drought occurrence over Madhya Pradesh is more during RMM-7 phase.
