Gro launches new Real-Time Assessment on China’s Food SecurityLearn More
Ukraine and Russia winter wheat models were trained on the province-level yield data, since the UKRSTAT and EMISS data sets had province-level data and no district-level data. As a result, the sample sizes were a lot smaller for those two countries. All the provinces in Ukraine were included in our Ukraine winter wheat yield model, whereas the Russia winter wheat yield model included only provinces on the European side of the country that accounted for over 90% of the winter wheat production in Russia. The geospatial variables chosen for the Ukraine winter wheat model were the normalized difference vegetation index (NDVI), evapotranspiration (ET), and GHCN weather station precipitation data. Those chosen for the Russia winter wheat model also included daily maximum and minimum temperature from the GHCN weather stations and soil moisture, in addition to NDVI and ET.