Córdova et al. (2026) Precipitation downscaling for climate characterization with scarce data using transfer learning

Identification

• Journal: Environmental Modelling & Software
• Year: 2026
• Date: 2026-08-27
• Authors: Mario Córdova, Luis Timbe, Esteban P. Samaniego
• DOI: 10.1016/j.envsoft.2026.107139

Research Groups

• Departamento de Recursos Hídricos y Ciencias Ambientales, Universidad de Cuenca
• Facultad de Ingeniería, Universidad de Cuenca

Short Summary

This study presents a deep learning-based downscaling framework using transfer learning to improve the accessibility and quality of high-resolution climate information in data-scarce regions. The method produces accurate long-term precipitation estimates with only a few years of observations.

Objective

• Investigate the feasibility of transfer learning for downscaling precipitation data in regions with limited local data.

Study Configuration

• Spatial Scale: Regional (Andean region of Ecuador)
• Temporal Scale: Long-term (years to decades)

Methodology and Data

• Models used: Deep learning-based framework using transfer learning
• Data sources: Satellite, reanalysis products, and a few years of local observations

Main Results

• The proposed method produces accurate long-term precipitation estimates in the Andean region of Ecuador with only a few years of observations.
• Transfer learning reduces the need for extensive weather station networks.

Contributions

• This study demonstrates the potential of transfer learning to improve the accessibility and quality of high-resolution climate information in data-scarce, topographically complex regions.
• The proposed method is flexible and extendable to other climate variables and future climate projections.

Funding

• No funding information provided.

Citation

@article{Córdova2026Precipitation,
  author = {Córdova, David and Córdova, Mario and Timbe, Luis and Samaniego, Esteban P.},
  title = {Precipitation downscaling for climate characterization with scarce data using transfer learning},
  journal = {Environmental Modelling & Software},
  year = {2026},
  doi = {10.1016/j.envsoft.2026.107139},
  url = {https://doi.org/10.1016/j.envsoft.2026.107139}
}