Conditional Diffusion Models for Global Precipitation Map Inpainting
Earth and Space Science · 2026-09 · 13 (9)
DOI: 10.1029/2025ea004666 ↗Kurz gefasst
Fachartikel in der Zeitschrift Earth and Space Science, online erschienen am 24.09.2026. Verfasst von Daiko Kishikawa und 2 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Japan Aerospace Exploration Agency, Japan Society for the Promotion of Science.
Abstract (Originalsprache)
Incomplete satellite‐based precipitation presents a significant challenge in global monitoring. For example, the Global Satellite Mapping of Precipitation (GSMaP) from the Japan Aerospace Exploration Agency (JAXA) contains regions without valid microwave observations because Low Earth orbiting (LEO) satellites cannot observe the entire globe simultaneously, and its current interpolation methods often result in spatial discontinuities. In this study, we formulate the completion of the precipitation map as a video inpainting task and propose a machine learning approach based on conditional diffusion models. Our method employs a three‐dimensional (3D) U‐Net with a 3D condition encoder to reconstruct complete precipitation maps by leveraging spatio‐temporal information from infrared images, latitude‐longitude grids, and physical time inputs. Training was carried out on hourly precipitation data from the fifth‐generation European Center for Medium‐Range Weather Forecasts (ECMWF) reanalysis (ERA5) from 2020 to 2023. We generated a pseudo‐GSMaP data set by randomly applying GSMaP masks to the ERA5 maps. Performance was evaluated for the calendar year 2024, and our approach produces more spatio‐temporally consistent inpainted precipitation maps compared to conventional methods. These results indicate the potential to improve global precipitation monitoring using the conditional diffusion models.
Förderung
- Japan Aerospace Exploration Agency
- Japan Society for the Promotion of Science
Zitieren
APA
Kishikawa, D., Muto, Y., & Kotsuki, S. (2026). Conditional Diffusion Models for Global Precipitation Map Inpainting. Earth and Space Science, 13(9). https://doi.org/10.1029/2025ea004666
Vancouver
Kishikawa D, Muto Y, Kotsuki S. Conditional Diffusion Models for Global Precipitation Map Inpainting. Earth and Space Science. 2026;13(9). doi:10.1029/2025ea004666
BibTeX
@article{Kishikawa2026,
author = {Kishikawa, Daiko and Muto, Yuka and Kotsuki, Shunji},
title = {Conditional Diffusion Models for Global Precipitation Map Inpainting},
journal = {Earth and Space Science},
year = {2026},
volume = {13},
number = {9},
doi = {10.1029/2025ea004666}
}
EUVIMED