Guided Diffusion Sampling for Precipitation Forecast Interventions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ueyama, Ayumu, Kawamoto, Kazuhiko, Kera, Hiroshi
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917494298509312
author Ueyama, Ayumu
Kawamoto, Kazuhiko
Kera, Hiroshi
author_facet Ueyama, Ayumu
Kawamoto, Kazuhiko
Kera, Hiroshi
contents Extreme precipitation causes severe societal and economic damage, and weather control has long been discussed as a potential mitigation strategy. However, to the best of our knowledge, perturbation-based interventions for weather control using data-driven weather forecasting models have not yet been explored. While adversarial attacks also generate perturbations that alter forecasts, they aim to exploit model artifacts and do not account for physical plausibility. In this paper, we propose a gradient-based guidance framework for precipitation-reduction interventions through diffusion sampling in diffusion-based weather forecasting models. Instead of directly perturbing atmospheric states, our method steers the diffusion sampling trajectory, enabling precipitation reduction while maintaining consistency with the atmospheric distribution. To assess physical plausibility, we evaluate from three perspectives: (i) vertical and variable-wise perturbation profiles, (ii) latent-space trajectory deviation, and (iii) cross-model transferability. Experiments on extreme precipitation events from WeatherBench2 demonstrate that our method achieves effective precipitation reduction while yielding more physically plausible interventions than adversarial perturbations.
format Preprint
id arxiv_https___arxiv_org_abs_2605_14317
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Guided Diffusion Sampling for Precipitation Forecast Interventions
Ueyama, Ayumu
Kawamoto, Kazuhiko
Kera, Hiroshi
Machine Learning
Atmospheric and Oceanic Physics
Extreme precipitation causes severe societal and economic damage, and weather control has long been discussed as a potential mitigation strategy. However, to the best of our knowledge, perturbation-based interventions for weather control using data-driven weather forecasting models have not yet been explored. While adversarial attacks also generate perturbations that alter forecasts, they aim to exploit model artifacts and do not account for physical plausibility. In this paper, we propose a gradient-based guidance framework for precipitation-reduction interventions through diffusion sampling in diffusion-based weather forecasting models. Instead of directly perturbing atmospheric states, our method steers the diffusion sampling trajectory, enabling precipitation reduction while maintaining consistency with the atmospheric distribution. To assess physical plausibility, we evaluate from three perspectives: (i) vertical and variable-wise perturbation profiles, (ii) latent-space trajectory deviation, and (iii) cross-model transferability. Experiments on extreme precipitation events from WeatherBench2 demonstrate that our method achieves effective precipitation reduction while yielding more physically plausible interventions than adversarial perturbations.
title Guided Diffusion Sampling for Precipitation Forecast Interventions
topic Machine Learning
Atmospheric and Oceanic Physics
url https://arxiv.org/abs/2605.14317