Machine learning methods for postprocessing ensemble forecasts of wind gusts: A systematic comparison

TL;DR

This paper systematically compares eight statistical and machine learning methods for probabilistic wind gust forecasting from ensemble predictions, demonstrating that neural networks with additional meteorological predictors significantly outperform traditional techniques.

Contribution

The study introduces a flexible neural network framework for wind gust postprocessing that outperforms existing methods and captures physical atmospheric relations.

Findings

All methods produce calibrated forecasts correcting systematic errors.

Adding meteorological predictors improves forecast skill.

Neural networks outperform benchmark methods and learn physical relations.

Abstract

Postprocessing ensemble weather predictions to correct systematic errors has become a standard practice in research and operations. However, only few recent studies have focused on ensemble postprocessing of wind gust forecasts, despite its importance for severe weather warnings. Here, we provide a comprehensive review and systematic comparison of eight statistical and machine learning methods for probabilistic wind gust forecasting via ensemble postprocessing, that can be divided in three groups: State of the art postprocessing techniques from statistics (ensemble model output statistics (EMOS), member-by-member postprocessing, isotonic distributional regression), established machine learning methods (gradient-boosting extended EMOS, quantile regression forests) and neural network-based approaches (distributional regression network, Bernstein quantile network, histogram estimation network). The methods are systematically compared using six years of data from a high-resolution, convection-permitting ensemble prediction system that was run operationally at the German weather service, and hourly observations at 175 surface weather stations in Germany. While all postprocessing methods yield calibrated forecasts and are able to correct the systematic errors of the raw ensemble predictions, incorporating information from additional meteorological predictor variables beyond wind gusts leads to significant improvements in forecast skill. In particular, we propose a flexible framework of locally adaptive neural networks with different probabilistic forecast types as output, which not only significantly outperform all benchmark postprocessing methods but also learn physically consistent relations associated with the diurnal cycle, especially the evening transition of the planetary boundary layer.