Guided Diffusion Sampling for Precipitation Forecast Interventions

TL;DR

This paper introduces a gradient-guided diffusion sampling method to reduce extreme precipitation in weather forecasts, ensuring physical plausibility and transferability across models.

Contribution

It presents a novel data-driven intervention technique that steers diffusion sampling trajectories for precipitation reduction, unlike previous perturbation methods.

Findings

Effective precipitation reduction demonstrated on WeatherBench2 data.

Interventions maintain physical plausibility across multiple evaluation criteria.

Method outperforms adversarial perturbations in physical consistency.

Abstract

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.