A comparative study of convolutional neural network models for wind field downscaling. (3rd December 2020)

• Record Type: Journal Article
• Authors: Höhlein, Kevin; Kern, Michael; Hewson, Timothy; Westermann, Rüdiger
• Is Part Of: Meteorological applications. Volume 27:Number 6(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-03
• Subjects: convolutional neural network (CNN) -- deep learning -- statistical downscaling -- wind field simulation; Meteorology -- Periodicals; Meteorological services -- Periodicals; 551.5
• DOI: 10.1002/met.1961
• Languages: English
• ISSNs: 1350-4827

Abstract

Abstract: We analyze the applicability of convolutional neural network (CNN) architectures for downscaling of short‐range forecasts of near‐surface winds on extended spatial domains. Short‐range wind forecasts (at the 100 m level) from European Centre for Medium Range Weather Forecasts ERA5 reanalysis initial conditions at 31 km horizontal resolution are downscaled to mimic high resolution (HRES) (deterministic) short‐range forecasts at 9 km resolution. We evaluate the downscaling quality of four exemplary CNN architectures and compare these against a multilinear regression model. We conduct a qualitative and quantitative comparison of model predictions and examine whether the predictive skill of CNNs can be enhanced by incorporating additional atmospheric variables, such as geopotential height and forecast surface roughness, or static high‐resolution fields, like land–sea mask and topography. We further propose DeepRU, a novel U‐Net‐based CNN architecture, which is able to infer situation‐dependent wind structures that cannot be reconstructed by other models. Inferring a target 9 km resolution wind field from the low‐resolution input fields over the Alpine area takes less than 10 ms on our graphics processing unit target architecture, which compares favorably to an overhead in simulation time of minutes or hours between low‐ and high‐resolution forecast simulations. Abstract : In this study, we explore the use of convolutional neural network (CNN) models for statistical downscaling of low‐resolution wind forecast simulations to a higher spatial resolution. We compare different model architectures with respect to their predictive skills, and examine whether these skills can be enhanced by incorporating additional atmospheric variables. With DeepRU, we propose a novel CNN model for statistical downscaling that achieves superior downscaling quality. We finally discuss potential use cases of CNN‐based downscaling in future forecasting applications.