Adversarial Observations in Weather Forecasting
This work addresses a critical security risk in weather forecasting that could cause large-scale disruptions and undermine public trust, highlighting a new threat as AI systems replace traditional methods.
The paper tackles the vulnerability of AI-based weather forecasting systems like GenCast by presenting a novel attack that manipulates forecasts and fabricates extreme events through subtle perturbations in weather observations, changing less than 0.1% of measurements.
AI-based systems, such as Google's GenCast, have recently redefined the state of the art in weather forecasting, offering more accurate and timely predictions of both everyday weather and extreme events. While these systems are on the verge of replacing traditional meteorological methods, they also introduce new vulnerabilities into the forecasting process. In this paper, we investigate this threat and present a novel attack on autoregressive diffusion models, such as those used in GenCast, capable of manipulating weather forecasts and fabricating extreme events, including hurricanes, heat waves, and intense rainfall. The attack introduces subtle perturbations into weather observations that are statistically indistinguishable from natural noise and change less than 0.1% of the measurements - comparable to tampering with data from a single meteorological satellite. As modern forecasting integrates data from nearly a hundred satellites and many other sources operated by different countries, our findings highlight a critical security risk with the potential to cause large-scale disruptions and undermine public trust in weather prediction.