Search for a command to run...
Long-term, spatially comprehensive meteorological observations are essential for climate research, yet ground-based station networks worldwide suffer from temporal discontinuities that limit their utility. We present RusWeather-GF, a gap-filled daily temperature and precipitation dataset for 593 Russian weather stations spanning 1980–2023. The dataset addresses critical limitations of existing Russian climate data through validated multi-method gap-filling achieving 100% temporal completeness, integration of high-resolution FABDEM v1.2 topographic descriptors, and extension 13 years beyond previous publicly available compilations. All 145,122 temperature gaps (1.62% of observations) and 161,534 precipitation gaps (1.80% of observations) were filled using an adaptive approach: inverse distance weighting for short gaps (≤7 days), Random Forest regression incorporating temporal, spatial, and topographic predictors for medium gaps (8–30 days), and station-specific climatology for extended gaps (>30 days). Comprehensive validation demonstrates preservation of temporal autocorrelation structure (Δ < 0.01), spatial consistency ( r > 0.999), and statistical distributions. Cross-validation using 59 stratified stations demonstrates temperature RMSE of 5.02 °C (R 2 = 0.9) and precipitation MAE of 1.79 mm with negligible bias, confirming robust performance across Arctic to temperate climate zones. The dataset comprises 8,893,613 daily records with station-level metadata including coordinates, elevation, slope, and terrain roughness. RusWeather-GF enables diverse applications including climate trend analysis, hydrological modeling, agricultural studies, and validation of gridded products and reanalysis datasets across Russia's climatically and topographically diverse territory. The dataset is publicly available via Zenodo ( https://doi.org/10.5281/zenodo.17789545 ) under CC BY 4.0 license. • Daily temperature and precipitation data restored for 593 Russian weather stations (1980–2023). • 306,656 gaps filled using three methods depending on interruption duration. • Validation confirms preservation of temporal autocorrelation and spatial patterns ( r > 0.999). • High-resolution FABDEM v1.2 topographic data integrated for all stations. • 8.9 million daily records spanning 44 years openly accessible via Zenodo with permanent DOI.