Forecasting Extreme Day and Night Heat in Paris

2025-08-18

AI Summary

This study addresses the short-term forecasting of extreme daytime and nighttime high temperatures (90th-percentile air temperature) in Paris. We propose a novel framework integrating quantile gradient boosting trees with conformal prediction. Leveraging observational data from the Montsouris meteorological station, the model incorporates 14-day lagged meteorological variables—including temperature, humidity, wind speed, and five other atmospheric features—to construct separate quantile regression models for daytime and nighttime, targeting the Q(0.90) quantile. Our key contributions are threefold: (i) the first statistically rigorous uncertainty quantification specifically tailored to diurnal extremes; (ii) explicit differentiation of daytime versus nighttime heatwave dynamics; and (iii) finite-sample coverage guarantees for prediction intervals via conformal inference. Experimental results demonstrate substantial improvements in both accuracy and reliability of two-week extreme temperature forecasts. The framework delivers a verifiable, operationally deployable decision-support tool for urban heat-risk early warning.

Technology Category

Machine Learning: Calibration & Uncertainty QuantificationReasoning under Uncertainty: Relational Probabilistic ModelsConstraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingSystems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applicationsGraph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs

Abstract

As a demonstration of concept, quantile gradient boosting is used to forecast diurnal and nocturnal Q(.90) air temperatures for Paris, France during late the spring and summer months of 2020. The data are provided by the Paris-Montsouris weather station. Q(.90) values are estimated because the 90th percentile requires that the temperatures be relatively rare and extreme. Predictors include seven routinely collected indicators of weather conditions, lagged by 14 days; the temperature forecasts are produced two weeks in advance. Conformal prediction regions capture forecasting uncertainty with provably valid properties. For both diurnal and nocturnal temperatures, forecasting accuracy is promising, and sound measures of uncertainty are provided.

Problem

• Forecasting extreme daytime and nighttime heat in Paris
• Using quantile gradient boosting for temperature prediction
• Providing accurate uncertainty measures for extreme temperatures

Innovation

• Quantile gradient boosting for extreme temperature forecasting
• 14-day lagged weather indicators as predictors
• Conformal prediction regions for uncertainty measurement