Assessing the Reliability and Uncertainty of a Remote Sensing Driven Crop Model - rsEPIC.

Input data uncertainty is one of the important sources of crop yield model errors. Model accuracy and reliability further degrade when applied in a data-poor region. Satellite based remote sensing, on the other hand, provides repeated real-time observations of crop growth status. We hypothesize that incorporation of remotely sensed crop parameter values with a physical crop model could improve the reliability and reduce the uncertainty in simulating crop yields at regional levels. Here we present a NASA/REASON Program funded case study that evaluated the performance of a remote sensing driven crop model, rsEPIC, in simulating corn yields in IL, USA and rice yields in Jiangsu, China