Stress-Testing Dynamical and Generative Downscaling Using Subseasonal Extreme Precipitation Forecasts
Abstract
Coarse spatial resolution limits the ability of subseasonal prediction models to resolve extreme precipitation. Downscaling with either dynamical or deep generative models can overcome this issue, but the comparative performance of these models for extremes across different atmospheric regimes remains poorly understood. In this work, we evaluate the Weather Research and Forecasting (WRF) model against a diffusion-based generative model by downscaling two physically distinct, extreme precipitation events up to lead times of 3 weeks. For a fair comparison with WRF, which can downscale boundary conditions from different driving models without model-specific training, the diffusion model is trained in an unpaired fashion. Both approaches improve upon the raw European Centre for Medium-Range Weather Forecasts forecasts, in comparison to fused rain gauge-radar observations in Switzerland (CombiPrecip), but exhibit regime-dependent strengths. WRF achieves the highest probabilistic skill for a multicell, non-stationary event. Conversely, the diffusion model is more consistent across different performance metrics for the two events, outperforming WRF in a more stationary supercell event. These results demonstrate that explicit dynamical modeling can add value for specific precipitation events for subseasonal lead times, and that generative downscaling adds value more broadly in different situations.
Related
- SwAIther-Precip: Lead-Time-Aware Bias Correction Enables Kilometer-Scale Downscaling of Global AI Precipitation Forecasts over Switzerland
- Ensemble reliability and the signal‐to‐noise paradox in ECMWF subseasonal forecasts
- Statistical Downscaling of Daily Precipitation from ERA5 Atmospheric Fields to IMERG over West Africa
- Spatio-temporal precipitation downscaling at hourly resolution based on km-scale regional simulations
- Evaluating Cross-region Generalization for Wavelet-Diffusion Precipitation Downscaling
BibTeX
@misc{lima2026stresstesting,
title = {{Stress-Testing Dynamical and Generative Downscaling Using Subseasonal Extreme Precipitation Forecasts}},
author = {Mauricio Lima and Marika Koukoula and Romain Pilon and Monika Feldmann and Erwan Koch and Daniela I. V. Domeisen and Tom Beucler},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.11696},
eprint = {2609.11696},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.11696},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Stress-Testing Dynamical and Generative Downscaling Using Subseasonal Extreme Precipitation Forecasts
AU - Mauricio Lima
AU - Marika Koukoula
AU - Romain Pilon
AU - Monika Feldmann
AU - Erwan Koch
AU - Daniela I. V. Domeisen
AU - Tom Beucler
PY - 2026
DA - 2026/09/10
PB - arXiv
DO - 10.48550/arxiv.2609.11696
UR - https://arxiv.org/abs/2609.11696
AB - Coarse spatial resolution limits the ability of subseasonal prediction models to resolve extreme precipitation. Downscaling with either dynamical or deep generative models can overcome this issue, but the comparative performance of these models for extremes across different atmospheric regimes remains poorly understood. In this work, we evaluate the Weather Research and Forecasting (WRF) model against a diffusion-based generative model by downscaling two physically distinct, extreme precipitation events up to lead times of 3 weeks. For a fair comparison with WRF, which can downscale boundary conditions from different driving models without model-specific training, the diffusion model is trained in an unpaired fashion. Both approaches improve upon the raw European Centre for Medium-Range Weather Forecasts forecasts, in comparison to fused rain gauge-radar observations in Switzerland (CombiPrecip), but exhibit regime-dependent strengths. WRF achieves the highest probabilistic skill for a multicell, non-stationary event. Conversely, the diffusion model is more consistent across different performance metrics for the two events, outperforming WRF in a more stationary supercell event. These results demonstrate that explicit dynamical modeling can add value for specific precipitation events for subseasonal lead times, and that generative downscaling adds value more broadly in different situations.
ER -