Margarita et al. (2026) RusWeather-GF: A gap-filled daily weather dataset for Russia (1980–2023) with integrated topographic data

Identification

• Journal: Atmospheric Research
• Year: 2026
• Date: 2026-03-21
• Authors: Tkachenko Margarita, Fomin Dmitriy
• DOI: 10.1016/j.atmosres.2026.108935

Research Groups

• V.V. Dokuchaev Soil Science Institute, Moscow, Russia
• Saint Petersburg State University, Saint Petersburg, Russia

Short Summary

This study presents RusWeather-GF, a novel gap-filled daily temperature and precipitation dataset for 593 Russian weather stations from 1980 to 2023, achieving 100% temporal completeness through a validated multi-method approach and integrating high-resolution topographic data. The dataset addresses critical limitations of existing Russian climate data, extending coverage and enhancing utility for diverse climate research applications.

Objective

• To develop and validate a comprehensive, gap-filled daily temperature and precipitation dataset for Russia, addressing temporal discontinuities and extending the temporal coverage of existing publicly available climate data.

Study Configuration

• Spatial Scale: Russia, covering over 17 million square kilometers, utilizing data from 593 weather stations.
• Temporal Scale: Daily records spanning 44 years, from 1980 to 2023.

Methodology and Data

• Models used:
  • Inverse Distance Weighting (IDW) for short gaps (≤7 days).
  • Random Forest regression for medium gaps (8–30 days), incorporating temporal, spatial, and topographic predictors.
  • Station-specific climatology for extended gaps (>30 days).
• Data sources:
  • Ground-based station observations from the Roshydromet network, archived by the All-Russian Research Institute of Hydrometeorological Information - World Data Centre (RIHMI-WDC).
  • FABDEM v1.2 topographic descriptors (elevation, slope, terrain roughness).

Main Results

• The RusWeather-GF dataset provides 100% temporal completeness for daily temperature and precipitation records across 593 Russian weather stations (1980–2023).
• A total of 145,122 temperature gaps (1.62% of observations) and 161,534 precipitation gaps (1.80% of observations) were successfully filled.
• Validation demonstrated preservation of temporal autocorrelation structure (Δ < 0.01) and spatial consistency (r > 0.999).
• Cross-validation using 59 stratified stations yielded a Root Mean Square Error (RMSE) of 5.02 °C for temperature (R² = 0.9) and a Mean Absolute Error (MAE) of 1.79 mm for precipitation with negligible bias.
• The dataset comprises 8,893,613 daily records, including station-level metadata such as coordinates, elevation, slope, and terrain roughness.
• The dataset is publicly available via Zenodo (https://doi.org/10.5281/zenodo.17789545) under a CC BY 4.0 license.

Contributions

• Creation of the first 100% temporally complete, gap-filled daily temperature and precipitation dataset for Russia, significantly enhancing data utility for climate research.
• Extension of the dataset's temporal coverage by 13 years beyond previously publicly available compilations.
• Integration of high-resolution FABDEM v1.2 topographic descriptors, providing crucial contextual information for each station.
• Development and validation of an adaptive multi-method gap-filling approach tailored to different gap lengths, ensuring robust performance across diverse climate zones.
• Provides a valuable resource for climate trend analysis, hydrological modeling, agricultural studies, and validation of gridded products and reanalysis datasets across Russia's vast and diverse territory.

Funding

• Not specified in the provided text.

Citation

@article{Margarita2026RusWeatherGF,
  author = {Margarita, Tkachenko and Dmitriy, Fomin},
  title = {RusWeather-GF: A gap-filled daily weather dataset for Russia (1980–2023) with integrated topographic data},
  journal = {Atmospheric Research},
  year = {2026},
  doi = {10.1016/j.atmosres.2026.108935},
  url = {https://doi.org/10.1016/j.atmosres.2026.108935}
}