U.S. and Ukraine corn yield, remote sensing, climate reanalysis, and soil properties transfer learning datasets

Contents This upload consists of six files: dataset-us-source-domain.csv: The dataset used for pre-training after merging, cleaning, and missing values imputation. It contains data from counties in 13 states in the U.S. corn belt (Illinois, Indiana, Iowa, Kansas, Kentucky, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, and Wisconsin) over 2010–2024, where each observation in the dataset corresponds to a county-year pair. dataset-ukraine-target-domain.csv: The dataset used for fine-tuning after merging, cleaning, and missing values imputation. It contains data from all 24 oblasts in Ukraine over 2015–2024, where each observation in the dataset corresponds to an oblast-year pair. gee-us-monthly.js: The JavaScript code used to export the monthly variables for U.S. from Google Earth Engine [1] via the Code Editor. gee-us-static.js: The JavaScript code used to export the static variables for U.S. from Google Earth Engine [1] via the Code Editor. gee-ukraine-monthly.js: The JavaScript code used to export the monthly variables for Ukraine from Google Earth Engine [1] via the Code Editor. gee-ukraine-static.js: The JavaScript code used to export the static variables for Ukraine from Google Earth Engine [1] via the Code Editor. Variables in the datasets Monthly variables were averaged for each county/oblast and aggregated on a monthly basis from March through September in Google Earth Engine [1] for each observation year. To filter features to pixels classified as corn cropland in the U.S. corn belt, a Cropland Data Layer (CDL) mask from USDA NASS [2] was applied, while the MODIS Land Cover Type Yearly [3] was used to filter pixels classified as cropland in Ukraine. All rights reserved by the original sources. Variable Aggregation Column name Source Enhanced Vegetation Index (EVI) Monthly mean and 95th percentile EVI_mean, EVI_p95 Moderate Resolution Imaging Spectroradiometer (MODIS) [4] Land Surface Water Index (LSWI) Monthly mean and minimum LSWI_mean, LSWI_min Moderate Resolution Imaging Spectroradiometer (MODIS) [5] Leaf Area Index (LAI) Monthly mean and maximum LAI_mean, LAI_max Moderate Resolution Imaging Spectroradiometer (MODIS) [6] Land Surface Temperature (LST) Monthly mean LST_mean Moderate Resolution Imaging Spectroradiometer (MODIS) [7] Growing Degree Days (GDD) Monthly sum GDD_sum ERA5-Land [8] (derived from daily air temperature) Extreme Degree Days (EDD) Monthly sum EDD_sum ERA5-Land [8] (derived from daily air temperature) Precipitation Monthly sum P_sum ERA5-Land [8] Vapor Pressure Deficit (VPD) Monthly mean and maximum VPD_mean, VPD_max ERA5-Land [8] Solar radiation Monthly sum SRAD_sum ERA5-Land [8] Soil moisture Monthly mean SM_mean ERA5-Land [8] Soil sand content Static sand_mean OpenLandMap [9] Soil clay content Static clay_mean OpenLandMap [10] Soil pH Static ph_mean OpenLandMap [11] Soil organic carbon content Static soc_mean OpenLandMap [12] Latitude and longitude Static lat, lon Google Earth Engine [1] U.S. corn belt corn yield Annual yield USDA NASS QuickStats [13] Ukraine corn yield Annual yield State Statistics Service of Ukraine (Ukrstat) [14]