Estimating winter wheat yield based on a light use efficiency model and wheat variety data

Satellite-based models are important tools for monitoring regional and global crop yields, because remotely sensed data is able to offer temporally and spatially continuous crop growth information over large areas. However, tracking inter-annual variability in crop yields remains a major challenge. Taking this challenge, a light use efficiency model (EC-LUE) is used to estimate winter wheat yield with 30-m spatial resolution Landsat data. Moreover, we integrate winter wheat variety data by considering different percentages of harvested organs to total crop organs. Using county-level statistical yield data for Kansas, United States, model validation shows that the EC-LUE model combined with wheat variety data can effectively capture the spatial variations of winter wheat yields. Specifically, the proposed method significantly improves model simulation performance for the inter-annual variation of yields during 2008-2017 and can explain 82% of the inter-annual yield variation. This method proposed in this study proves to be available for future applications of crop yield estimations on a larger scale. The great potential of incorporating satellite observations with crop variety data to monitor crop yield, and improve agricultural management is therefore indicated.