Earth system models improve risk decisions with better uncertainty measures
Decision-Oriented Uncertainty Quantification for Risk Control in Earth System Spatiotemporal Foundation Models
Machine Learning
Summary
Predicting Earth events like storms or floods is tricky, and just guessing the future isn't enough for decisions like warnings or resource planning. The authors developed a way to not only predict what might happen but also measure how risky different decisions are based on those predictions. This helps ensure that warnings or actions are more reliable and reduce costly mistakes. Their approach improved decision quality and reduced errors compared to other methods.
What this means in practice
- •For emergency response teams: Select more reliable warnings and resource allocations for extreme weather events using calibrated risk predictions.
- •For energy grid operators: Improve renewable energy dispatch decisions by incorporating decision-aware uncertainty in weather forecasts.
Authors
Ji Lu, Huiran Duan, Bo Zhao, Xianglong Wang, Yiru Fang, Kuo Yang, Xiaoqin Feng, Jianping Gou
Abstract
Earth system modeling is shifting from task-specific predictors toward foundation models with general spatiotemporal representation capabilities. Although these models can jointly encode dynamic Earth fields, external forcings, and static geographic context for multistep forecasting, accurate point predictions or statistically calibrated intervals alone are insufficient for high-impact applications such as extremeweather warning, flood control, renewable-energy dispatch, and emergency resource allocation. What matters in practice is whether predictive uncertainty can be translated into reliable decision risk under specific actions, loss functions, and risk preferences. We propose a decision-oriented uncertainty quantification framework for Earth system spatiotemporal foundation models. The framework produces predictive distributions of future states and uses a decision risk adapter to map forecast samples, decision context, and utility functions into action-conditional risks. A utility-aware calibration module further enforces reliability at the downstream decision-loss level rather than only at the forecast-value level. Calibrated risks are then used to select warning, dispatch, inspection, or resource-allocation actions. Compared with the strongest baseline, the proposed method reduces decision regret by 18.7%, lowers the missed-event rate from 14.2% to 9.1%, and improves expected utility by 11.6%, while maintaining 90.4% predictive coverage and reducing decision calibration error from 0.083 to 0.047. These results suggest that decision-oriented uncertainty quantification can improve the robustness and operational value of Earth system foundation models in risk-sensitive applications.