Yield Estimation of Rice Crop Using Semi-Physical Approach and Remotely Sensed Data

AbstractTimely and accurate crop yield prediction is vital for agricultural land management and policymaking. With the advent of satellite sensors, the availability of large proxy parameters, advanced computing mechanisms, and strong analytical capability create enormous scope to develop more accurate and reliable crop production estimates. In the present study, an effort has been made to use the multi-sensor, multi-resolution satellite data and crop biophysical parameters to estimate rice crop yield for the Saharanpur district of Uttar Pradesh. A semi-physical approach, also known as light-use efficiency or Production Efficiency Models, has been employed to estimate rice crop productivity using remote sensing and physiological concepts such as the Photosynthetically Active Radiation (PAR) and the fraction of PAR absorbed by the crop. Rice crop map generated using multi-date Sentinel-1 (20 m resolution) microwave data through Sentinel-1 Toolbox “SNAP”, the accuracy of crop map was assessed with geotagged ground truth. Pixel level estimates were developed and accumulated at the district level using an area-weighted approach. The estimated yield was compared with the actual estimate obtained from the District Agriculture Department. The result obtained was very encouraging and well within the range of ±10%. The study reconfirms that the approach can generate the yield at a lower pixel level aggregated at sub-districts or below administrate units, but percolation at a lower scale requires more ground-specific information such as intensive ground truth and other yield proxies.KeywordsMonteith efficiency modelKharif riceMicrowave dataYieldRadiation-use efficiencyNDVIHarvest index