QuantWeather: Quantile-Aware Probabilistic Forecasting for Subseasonal Precipitation
2026-05-11 • Computational Engineering, Finance, and Science
Computational Engineering, Finance, and Science
AI summaryⓘ
The authors point out that predicting rainfall a few weeks ahead is hard because weather is chaotic, and existing methods often use groups of models that aren’t always accurate. Instead of relying on these groups and fixing them later with costly adjustments, the authors created QuantWeather, a system that predicts rainfall with built-in uncertainty estimates in one step. Their system has two parts working together to make both precise predictions and measure how uncertain those predictions are. Tests showed QuantWeather is better at forecasting and uses fewer computer resources than traditional methods.
subseasonal forecastingprecipitation predictionuncertainty estimationensemble forecastingmodel calibrationprobabilistic forecastingstochastic samplingdeep learningdual-head architecturereforecast datasets
Authors
Lei Chen, Xinyu Su, Xiaohui Zhong, Hao Li
Abstract
Subseasonal precipitation forecasting is inherently uncertain due to chaotic atmospheric dynamics, making reliable uncertainty estimation essential for real-world applications. Existing approaches typically represent uncertainty through ensemble forecasts rather than directly modeling predictive distributions. However, due to systematic model biases, raw ensemble outputs are often not well calibrated and cannot be directly interpreted as reliable uncertainty estimates. As a result, operational systems rely on post-hoc calibration based on reforecast datasets, which are computationally expensive to generate and maintain. To address these limitations, we propose QuantWeather, an end-to-end probabilistic forecasting framework with a dual-head design. The probabilistic and deterministic heads are supervised with separate objectives and optimized jointly. The framework further supports stochastic sampling, enabling probabilistic outputs even with a single stochastic forward pass and allowing optional multi-sample aggregation. Extensive experiments show that QuantWeather demonstrates superior probabilistic forecasting skill while substantially reducing inference-time computational and storage costs.