13B.5 Physically Constrained Deep Generative Approach to Precipitation Nowcasting

Thursday, 31 August 2023: 11:30 AM
Great Lakes A (Hyatt Regency Minneapolis)
Matej Choma, Prague, Czech republic; and M. Murín, M. Troller, J. Bartel, P. Šimánek, and M. Najman

For the past four years, we have been working on deep learning nowcasting of precipitation represented by radar reflectivity. Our choice of architecture quickly shifted towards recurrent convolutional neural networks, and our Meteopress Weather Net (MWNet) has achieved superior quantitative results against other neural networks and traditional non-ML extrapolation methods. We have tested it both on a historical dataset composed of selected precipitation situations and during operational trial use during the 2021 summer storm season in the Czech Republic. Its performance transfers well to other radar domains and non-radar spatiotemporal data, such as satellite imagery.

However, MWNet has yet to be suitable for operational use by forecasters. Focusing on the quantitative performance represented by Mean Squared Error resulted in blurry predictions that entirely miss the critical high-frequency features of weather measurements. Furthermore, forecasters are naturally suspicious of predictions not backed by laws of physics, no matter their accuracy.

This work presents PhyDGAN, a physically constrained recurrent neural network trained in the GAN framework. PhyDGAN, trained for precipitation prediction for up to an hour into the future on a radar reflectivity dataset, disentangles base physical dynamics, represented by the advection-diffusion partial differential equation, from other residual factors that are unknown given the limited 2D radar reflectivity input data. Among other things, these residual factors are responsible for modeling high-frequency features.

PhyDGAN solves both aforementioned challenges, producing highly detailed and partially explainable probabilistic nowcasts. We quantify these improvements and present an analysis of PhyDGAN's probabilistic properties and performance trade-offs associated with the proposed changes.

Finally, we explore latent features of PhyDGAN and the learned disentanglement of physical and residual dynamics, shedding light on the reasoning behind the predictions and creating a novel advection field output solely from radar reflectivity data. PhyDGAN is operationally used by forecasters in Meteopress.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner