Spatially Coherent Postprocessing of Cloud Cover Ensemble Forecasts

Y Dai,Stephan Hemri

doi:10.1175/mwr-d-21-0046.1

Abstract

AbstractStatistical postprocessing is commonly applied to reduce location and dispersion errors of probabilistic forecasts provided by numerical weather prediction (NWP) models. If postprocessed forecast scenarios are required, the combination of ensemble model output statistics (EMOS) for univariate postprocessing with ensemble copula coupling (ECC) or the Schaake shuffle (ScS) to retain the dependence structure of the raw ensemble is a state-of-the-art approach. However, modern machine learning methods may lead to both a better univariate skill and more realistic forecast scenarios. In this study, we postprocess multimodel ensemble forecasts of cloud cover over Switzerland provided by COSMO-E and ECMWF-IFS using (i) EMOS + ECC, (ii) EMOS + ScS, (iii) dense neural networks (dense NN) + ECC, (iv) dense NN + ScS, and (v) conditional generative adversarial networks (cGAN). The different methods are verified using EUMETSAT satellite data. Dense NN shows the best univariate skill, but cGAN performed only slightly worse. Furthermore, cGAN generates realistic forecast scenario maps, while not relying on a dependence template like ECC or ScS, which is particularly favorable in the case of complex topography.

Highlights

In recent years, neural networks have become increasingly powerful, and their rise has influenced weather forecasting, especially the area of postprocessing (Shi et al 2017; Rasp and Lerch 2018; Bremnes 2019)
For conditional generative adversarial networks (cGAN) a best and a worst ensemble member are selected from an arbitrary sample of 21 cGAN forecast realizations
Both global EMOS (gEMOS) and dense NN generate some clouds, but ensemble copula coupling (ECC) does not provide any member with a spatial distribution of clouds similar to the observed field

Summary

Introduction

Neural networks have become increasingly powerful, and their rise has influenced weather forecasting, especially the area of postprocessing (Shi et al 2017; Rasp and Lerch 2018; Bremnes 2019). We will focus on postprocessing of cloud cover forecasts. The amount of cloud cover has a large influence on the intensity of solar radiation (Matuszko 2012), and has implications for both human health (Juzeniene et al 2011) and energy supply (Palz 2013). Being able to accurately forecast cloud cover could have benefits for observational astronomy (Ye and Chen 2013), and the presence or absence of clouds often highly affects how the weather is perceived. Postprocessing of cloud cover forecasts using classical statistical methods, namely, multinomial and proportional odds.

Methods

Results

Discussion

Conclusion