Modeling of Wheat Yields Using Multi‐temporal Terra/MODIS Satellite Data

Abstract Two approaches viz., spectral crop growth profile and biomass production model(s) are used to assess crop development, model crop yields and productivity in the highly input intensive region of western Uttar Pradesh (UP). Multi‐date WiFS, fifteen MODIS images (level 3) having 8‐day composites of surface reflectance and some standard meteorological measurements are used as inputs. A Badhwar model was fitted to wheat spectral profiles and subsequently various spectral profile parameters were generated. Monteith's model was also used to estimate above ground biomass based on estimates of absorbed photosynthetically active radiation (APAR) derived from temporal MODIS satellite plus point meteorological data and crop stage specific varying light use efficiency (F,). The conversion of above ground biomass to yield was accomplished by harvest index derived from randomly planned crop cutting experiments. These spectral profile parameters showed distinct variation between north‐western and rest of the tehsils in western Uttar Pradesh (UP). A significantly positive correlation was found between tehsil‐wise yield and some spectral profile parameters (α, β and Tmax). In case of Monteith's model, the validation with tehsil‐wise BES (Bureau of Economic Survey) estimates showed that wheat yield can be estimated for approximately 60% of cases with relative deviation less than 5%. The model indicates a better performance in the northwestern parts having homogeneity of wheat fields and better irrigation infra‐structure. Future work should rely on integration of high and low resolution images to estimates field scale wheat yields in western UP.