aiqbench analysis
The paper applies U-Net and GAN architectures to downscale ERA5 reanalysis fields to IMERG satellite precipitation over West Africa, with ML as the core method rather than a side tool. Its most useful finding for ML readers is that deterministic networks come out too dry and lose most observed variance at small scales, which is why the authors turned to a GAN. The permutation-importance result, which ranks 500 hPa specific humidity first, also gives a physically interpretable check on what the model is learning.
Statistical Downscaling of Daily Precipitation from ERA5 Atmospheric Fields to IMERG over West Africa
arXiv:2610.16816v1 Announce Type: cross Abstract: Global reanalyses are too coarse to resolve the convective systems that produce most West African monsoon rainfall. In this work, we use U-Net based architectures to downscale daily ERA5 fields (0.25{\deg}) to IMERG V07 satellite precipitation (0.1{\deg}) over West Africa for the June to October seasons of 2000 to 2024, using geopotential, temperature, humidity and wind at three pressure levels together with elevation and land fraction. Adding ERA5 precipitation as input slightly improves performance for this particular challenging region. We find that all the deterministic networks investigated for this work are too dry and lose almost all observed variance below a few hundred kilometers, which motivated our use of a GAN-based alternative. We also investigated permutation importance for the U-Net that ranks 500 hPa specific humidity first, consistent with the sensitivity of deep convection to mid-tropospheric moisture.
Abstract:Global reanalyses are too coarse to resolve the convective systems that produce most West African monsoon rainfall. In this work, we use U-Net based architectures to downscale daily ERA5 fields (0.25°) to IMERG V07 satellite precipitation (0.1°) over West Africa for the June to October seasons of 2000 to 2024, using geopotential, temperature, humidity and wind at three pressure levels together with elevation and land fraction. Adding ERA5 precipitation as input slightly improves performance for this particular challenging region. We find that all the deterministic networks investigated for this work are too dry and lose almost all observed variance below a few hundred kilometers, which motivated our use of a GAN-based alternative. We also investigated permutation importance for the U-Net that ranks 500 hPa specific humidity first, consistent with the sensitivity of deep convection to mid-tropospheric moisture.
Atmospheric and Oceanic Physics (physics.ao-ph); Machine Learning (cs.LG)
arXiv:2610.16816 [physics.ao-ph]
(or arXiv:2610.16816v1 [physics.ao-ph] for this version)
https://doi.org/10.48550/arXiv.2610.16816 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 236 words
Related coverage
ORBIT-2 based Weather and Climate Downscaling and Downscaled Global Forecasts on AMD Instinct
AMD · Jun 8, 2026
Precipitation Downscaling Using Foundation Model-Conditioned Diffusion
arXiv · Aug 27, 2026
SwAIther-Precip: Lead-Time-Aware Bias Correction Enables Kilometer-Scale Downscaling of Global AI Precipitation Forecasts over Switzerland
arXiv · Jun 9, 2026
Regional Climate Model Emulation with Diffusion Approaches: What is the Added Value of Generative Machine Learning?
arXiv · Jun 15, 2026