Satellite data for the provision of early, area-wide and continuous information on crop yield estimates for agricultural statistics and policy advice.

Increasing climate variability and adverse weather patterns are posing unprecedented challenges to the agricultural sector, with negative impacts on farms, the entire value chain and consumer prices. When damaging events occur on a national and regional scale, federal and state governments provide support in the form of aid payments to mitigate farm losses and safeguard livelihoods. The results of the german crop yield statistics serve in these cases as a basis for decision-making. Reliable, regionalised in-season forecasts and the determination of the actual harvested yields of certain major crops would fill information gaps, support decision-making by federal and state authorities, and significantly improve planning in agricultural practice. However, the current estimation procedure is not spatially small scaled and, in some cases, suffers regionally from the partly declining participation of voluntary reporters. The SatErnte project aims to use new data sources and innovative methods from earth observation, data science and yield modelling to further develop existing methods of official agricultural statistics, to reduce current shortcomings and to expand and improve the content, timeliness and comparability of official statistics. To this end, SatErnte is investigating two complementary methodological approaches, machine learning-based and process-based yield forecasting, for the provision of intra-seasonal, regionalised yield forecasts using Copernicus Sentinel satellite data. SatErnte focuses on selected major field crops that are winter barley, winter wheat, winter oilseed rape, spring barley.