Combining predictive distributions for the statistical post-processing of ensemble forecasts

Abstract

Statistical post-processing techniques are now widely used to correct systematic biases and errors in calibration of ensemble forecasts obtained from multiple runs of numerical weather prediction models. A standard approach is the ensemble model output statistics (EMOS) method, a distributional regression approach where the forecast distribution is given by a single parametric law with parameters depending on the ensemble members. Choosing an appropriate parametric family for the weather variable of interest is a critical, however, often non-trivial task, and has been the focus of much recent research. In this article, we assess the merits of combining predictive distributions from multiple EMOS models based on different parametric families. In four case studies with wind speed and precipitation forecasts from two ensemble prediction systems, we study whether state of the art forecast combination methods are able to improve forecast skill.

Neural Networks for Postprocessing Ensemble Weather Forecasts

“…Having to describe the distribution of the target variable in parametric techniques is a nontrivial task. For temperature, a Gaussian distribution is a good approximation but for other variables, such as wind speed or precipitation, finding a distribution that fits the data is a substantial challenge (e.g., Taillardat et al, 2016;Baran and Lerch, 2018). Ideally, a machine learning algorithm would learn to predict the full probability distribution rather than distribution parameters only.…”

Statistical post‐processing of dual‐resolution ensemble forecasts

“…As suggested by Gneiting and Ranjan (), to assess the statistical significance of the differences between the verification scores, we make use of the Diebold–Mariano (DM) test of equal predictive performance, as it allows us to account for the temporal dependencies in the forecast errors (Diebold and Mariano, ). Baran and Lerch () give more details about the DM test. Further, confidence intervals for mean score values and mean score differences are obtained with the help of 2,000 block bootstrap samples using the stationary bootstrap scheme with mean block length according to Politis and Romano ().…”

Aggregating expert-elicited data for prediction: A scoping review of statistical methods, experiments, and applications v1

“…To recalibrate, authors recommend transforming the aggregated forecast distribution. The Spread-adjusted Linear Pool (SLP) (Berrocal et al, 2007;Glahn et al, 2009;Kleiber et al, 2011) transforms each individual distribution before combining, the Beta Linear Pool (BLP) applies a beta transform to the final combined distribution Gneiting et al (2013); Ranjan and Gneiting (2010), and a more flexible infinite mixture version of the BLP Bassetti et al (2018), mixture of Normal densities Baran and Lerch (2018), and empirical cumulative distribution function Garratt et al (2019) also aim to recalibrate forecasts made from a combination of predictive densities.…”