Postprocessing of gridded precipitation forecasts using conditional generative adversarial networks and quantile regression

Probabilistic predictions of precipitation call for rather sophisticated postprocessing approaches due to its low predictability, high spatio-temporal variability and highly positive skewness. Moreover, the large number of zeros makes the generation of physically realistic postprocessed forecast scenarios using standard approaches like ensemble copula coupling (ECC) rather difficult. In addition to classical statistical approaches, recently, machine learning based methods gained increasing popularity in the field of postprocessing of probabilistic weather forecasts.

In this study, we compare conditional generative adversarial network (cGAN) based postprocessing of daily precipitation with a quantile regression based approach. In principle, an appropriately trained cGAN model should be able to generate postprocessed forecast scenarios that improve forecast skill and cannot be distinguished from observed data in terms of spatial structure. While we use ECC to generate physically realistic forecast scenarios from quantile regression, cGAN does not need any additional ECC steps. For training and verification, we use COSMO-E ensemble forecasts with a grid resolution of about 2 km over Switzerland and the corresponding CombiPrecip observations, which are a gridded blend of radar and gauge observations. Preliminary results suggest that it is possible to generate realistic looking forecast scenarios using cGAN, but up to now, we have not been able to increase forecast skill. On the other hand, quantile regression seems to increase forecast skill at the expense of relying on an additional ECC step to generate forecast scenarios.