Post-processing & calibration★ Highlight

Neural networks for post-processing ensemble weather forecasts

Stephan Rasp, Sebastian Lerch

Monthly Weather Review 2018, 146, 3885-3900 · 23 May 2018 · Karlsruhe Institute of Technology (KIT)

Why it matters. Influential neural network ensemble post-processing with embeddings that beat EMOS benchmarks; became a standard baseline for ML post-processing.

In one sentence. Neural networks for ensemble post-processing outperform benchmark distributional regression for 2-m temperature at German stations, using embeddings and auxiliary predictors.

Abstract

Ensemble weather predictions require statistical post-processing of systematic errors to obtain reliable and accurate probabilistic forecasts. Traditionally, this is accomplished with distributional regression models in which the parameters of a predictive distribution are estimated from a training period. We propose a flexible alternative based on neural networks that can incorporate nonlinear relationships between arbitrary predictor variables and forecast distribution parameters that are automatically learned in a data-driven way rather than requiring pre-specified link functions. In a case study of 2-meter temperature forecasts at surface stations in Germany, the neural network approach significantly outperforms benchmark post-processing methods while being computationally more affordable. Key components to this improvement are the use of auxiliary predictor variables and station-specific information with the help of embeddings. Furthermore, the trained neural network can be used to gain insight into the importance of meteorological variables thereby challenging the notion of neural networks as uninterpretable black boxes. Our approach can easily be extended to other statistical post-processing and forecasting problems. We anticipate that recent advances in deep learning combined with the ever-increasing amounts of model and observation data will transform the post-processing of numerical weather forecasts in the coming decade.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.