Comparative evaluation of deep learning and statistical models for wind speed downscaling
Wind Engineering · 13.08.2026
DOI: 10.1177/0309524x261477588 ↗Kurz gefasst
Fachartikel in der Zeitschrift Wind Engineering, online erschienen am 13.08.2026. Verfasst von Juhi Kumari und 2 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben.
Abstract (Originalsprache)
Accurate wind speed downscaling is essential for meteorological applications and reliable station-scale wind estimation. This study presents a systematic comparison of recurrent deep learning architectures, including RNN, GRU, LSTM, Bidirectional LSTM, Sequence-to-Sequence LSTM, and Stacked LSTM, against conventional statistical models for wind speed downscaling using paired IMD station observations and NASA MERRA-2 reanalysis data for Pune, India. All models were evaluated under identical preprocessing and training protocols to ensure a fair comparison. Among the standard recurrent architectures, the Stacked LSTM achieved the best predictive performance. An Attention-Enhanced LSTM was subsequently evaluated as an extension to investigate the benefits of incorporating a soft-attention mechanism. It achieved the highest overall accuracy, with an RMSE of 0.5034 and an MAE of 0.3650, demonstrating additional performance gains over the standard recurrent architectures. The findings provide a robust benchmark for deep learning-based wind speed downscaling and future atmospheric prediction studies.
Zitieren
APA
Kumari, J., Wadhvani, R., & Shukla, S. (2026). Comparative evaluation of deep learning and statistical models for wind speed downscaling. Wind Engineering. https://doi.org/10.1177/0309524x261477588
Vancouver
Kumari J, Wadhvani R, Shukla S. Comparative evaluation of deep learning and statistical models for wind speed downscaling. Wind Engineering. 2026. doi:10.1177/0309524x261477588
BibTeX
@article{Kumari2026,
author = {Kumari, Juhi and Wadhvani, Rajesh and Shukla, Sanyam},
title = {Comparative evaluation of deep learning and statistical models for wind speed downscaling},
journal = {Wind Engineering},
year = {2026},
doi = {10.1177/0309524x261477588}
}
EUVIMED