Improving soil moisture estimation by assimilating remotely sensed data into crop growth model for agricultural drought monitoring

Soil moisture is an effective variable for agricultural drought monitoring, and data assimilation is a useful tool to improve soil moisture estimates. In this study, we assimilated remotely sensed soil moisture (SM) and leaf area index (LAI) into DSSAT-CSM-Wheat crop growth model to estimate soil moisture. The results showed that compared to open-loop scenario, assimilating LAI independently could slightly improve soil moisture accuracy with reduction in average RMSE (root mean square error) by 4%, and the average RMSEs were decreased by 7% and 10% for SM and LAI+SM assimilations, respectively. The yield differences with observation were decreased by 379 kg/ha for LAI assimilation, 592 kg/ha for SM assimilation and 866 kg/ha for LAI+SM assimilation. Assimilating LAI and SM jointly received best performances in soil moisture and yield estimation. Hence, assimilating remotely sensed data into crop growth model provides a robust method to improve soil moisture for agricultural drought monitoring.