Preprint · 2026 · Open access
A deep learning framework for downscaling meteorological fields in the European domain
Published 24 June 2026 · DOI 10.22541/essoar.15005151/v1
In brief
“A deep learning framework for downscaling meteorological fields in the European domain” is a 2026 preprint by Georgios Fragkoulidis and 4 co-authors. OpenAlex counts 1 citations as of 21 November 2026. It is open access under a CC BY-NC-ND licence: the full text is free to read.
- Citations
- 1
- Most cited in
- 2026 (1)
- Authors
- 5
Abstract
To overcome the high computational costs of dynamical models and the linear, pointwise assumptions of traditional statistical methods, deep learning techniques constitute promising pathways toward the high-resolution meteorological fields required for local impact assessments. The present study proposes a downscaling framework for four essential meteorological fields: 2-meter temperature, precipitation, surface solar radiation, and 10-meter wind speed. For each of them, it employs a multivariate Residual U-Net architecture, augmented with a macro-residual output engine and a dual-injection topography scheme. Trained on 39 years of 3-hourly CERRA reanalysis data, the four models map coarse 1° predictors to a high-resolution 0.05° grid over the topographically complex Greek domain. A systematic evaluation demonstrates that the new framework achieves domain-wide annual mean absolute errors of 0.38°C for 2-meter temperature, 1.29 mm/3h for precipitation, 13.89 W/m2 for surface solar radiation, and 0.54 m/s for 10-meter wind speed. Performance remains stable across the bulk of the climatological distributions of the four fields, though a considerable but expected underestimation emerges for intense precipitation. Spatial and intra-annual skill variations reflect the complexity in the geographical properties and physical processes shaping local meteorological conditions, rather than flaws in the model design. Downscaling accuracy exhibits a pronounced sensitivity to the prevailing weather over the region, while a learning curve analysis reveals that in some locations the downscaling skill of the temperature distribution tails requires more training years to converge compared to the skill of the full distribution. Overall, the proposed framework serves as a viable, computationally inexpensive alternative to both dynamical and traditional statistical downscaling for local climate assessments and impact modelling.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Fragkoulidis, G., Kyros, G., Vafeiadis, V., Kotroni, V., & Lagouvardos, K. (2026). A deep learning framework for downscaling meteorological fields in the European domain. https://doi.org/10.22541/essoar.15005151/v1
MLA 9
Fragkoulidis, Georgios, et al. “A deep learning framework for downscaling meteorological fields in the European domain.” 2026. https://doi.org/10.22541/essoar.15005151/v1.
Chicago
Fragkoulidis, Georgios, Giorgos Kyros, Vasileios Vafeiadis, Vassiliki Kotroni, and K. Lagouvardos. 2026. “A deep learning framework for downscaling meteorological fields in the European domain.” https://doi.org/10.22541/essoar.15005151/v1.
Harvard
Fragkoulidis, G. et al. (2026) ‘A deep learning framework for downscaling meteorological fields in the European domain’. doi:10.22541/essoar.15005151/v1.
IEEE
G. Fragkoulidis, G. Kyros, V. Vafeiadis, V. Kotroni, and K. Lagouvardos, “A deep learning framework for downscaling meteorological fields in the European domain,” 2026, doi: 10.22541/essoar.15005151/v1.
BibTeX
@misc{fragkoulidis2026deep,
title = {{A deep learning framework for downscaling meteorological fields in the European domain}},
author = {Fragkoulidis, Georgios and Kyros, Giorgos and Vafeiadis, Vasileios and Kotroni, Vassiliki and Lagouvardos, K.},
year = {2026},
doi = {10.22541/essoar.15005151/v1},
url = {https://doi.org/10.22541/essoar.15005151/v1}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Earth and Planetary Sciences › Atmospheric Science › Meteorological Phenomena and Simulations
- deep learning
- ResUNet
- precipitation
- surface solar radiation
- learning curve analysis
Work with this paper
Most-cited papers in Earth and Planetary Sciences
- The ERA5 global reanalysisHans Hersbach, Bill Bell, Paul Berrisford and 40 more · Quarterly Journal of the Royal Meteorological Society · 2020 · 35,424 citations
- The NCEP/NCAR 40-Year Reanalysis ProjectEugenia Kalnay, Masao Kanamitsu, Robert E. Kistler and 19 more · Bulletin of the American Meteorological Society · 1996 · 28,486 citations
- Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processesShen‐Su Sun, W. F. McDonough · Geological Society London Special Publications · 1989 · 26,603 citations
- The ERA‐Interim reanalysis: configuration and performance of the data assimilation systemDick P. Dee, Sakari M. Uppala, A. J. Simmons and 33 more · Quarterly Journal of the Royal Meteorological Society · 2011 · 25,245 citations
- PAST: PALEONTOLOGICAL STATISTICAL SOFTWARE PACKAGE FOR EDUCATION AND DATA ANALYSISØyvind Hammer, David A. T. Harper, Paul D. Ryan · 2001 · 17,871 citations
Questions about this paper
How do I cite this paper in APA style?
Fragkoulidis, G., Kyros, G., Vafeiadis, V., Kotroni, V., & Lagouvardos, K. (2026). A deep learning framework for downscaling meteorological fields in the European domain. https://doi.org/10.22541/essoar.15005151/v1
How many times has this paper been cited?
1 times, according to OpenAlex on 21 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access: sign in and open the full text from this page.
Who wrote this paper?
Georgios Fragkoulidis, Giorgos Kyros, Vasileios Vafeiadis, Vassiliki Kotroni, K. Lagouvardos.
What is the DOI of this paper?
10.22541/essoar.15005151/v1, permanently at https://doi.org/10.22541/essoar.15005151/v1. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 21 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.