A comparative study of convolutional neural network models for wind field downscaling

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.

Artificial Intelligence Revolutionises Weather Forecast, Climate Monitoring and Decadal Prediction

“…A possible solution is the use of DL techniques as described in [63]. Examples of the use of DL for the downscaling of wind fields were given in [64,65]. An example of the use of DL for the downscaling of temperature was given in [66].…”

Spatio-Temporal Downscaling of Climate Data Using Convolutional and Error-Predicting Neural Networks

“…A difference to the conventional U-Net is that we experimented with additional data channels that provide more information to resolve ambiguities in the inverse map f −1 . The effectiveness of additional data channels was already demonstrated by Vandal et al (2017) and Höhlein et al (2020) for downscaling. These additional channels include latitude and longitude such that the network can learn regional weather patterns, and altitude to include dependencies on the topography.…”

“…In computer vision terms, this approach can be considered a multi-view super-resolution problem, whereas we investigate the more challenging single-image super-resolution. Höhlein et al (2020) studied multiple architectures for spatial downscaling of wind velocity data, including a U-Net based architecture and deep networks with residual learning. The latter resulted in the best performance on the wind velocity fields that they studied.…”

Modeling of a Generic Edge Computing Application Design

“…In this sense, some enhancements of CNN have been proposed for specific duties in recent times, to obtain greater accuracy in predicting visual recognition in data science, such as subpixel displacement measures [36], defect identification in high-speed trains [37], correlating image-like data out of quantum systems [38], modeling wind field downscaling [39], designing a zero knowledge proof scheme [40], classifying satellite image time series [41], working with ensembles [42], dealing with osteoporosis diagnoses [43], screening and staging diabetic retinopathy [44], analyzing cloud particles [45], inspecting diffraction data [46], or examining x-ray images [47].…”