Weakly-Constrained 4D Var for Downscaling with Uncertainty using Data-Driven Surrogate Models
This addresses the problem of unreliable long-term weather predictions for meteorologists and climate scientists by providing a more stable and computationally efficient downscaling method, though it is incremental as it builds on existing data assimilation techniques.
The authors tackled the instability of data-driven weather forecasting models like FourCastNet over long lead times by integrating them into a weakly-constrained 4DVar data assimilation framework for dynamic downscaling, resulting in improved forecast accuracy and uncertainty quantification compared to ensemble Kalman filter and unstabilized FourCastNet using ERA5 data.
Dynamic downscaling typically involves using numerical weather prediction (NWP) solvers to refine coarse data to higher spatial resolutions. Data-driven models such as FourCastNet have emerged as a promising alternative to the traditional NWP models for forecasting. Once these models are trained, they are capable of delivering forecasts in a few seconds, thousands of times faster compared to classical NWP models. However, as the lead times, and, therefore, their forecast window, increase, these models show instability in that they tend to diverge from reality. In this paper, we propose to use data assimilation approaches to stabilize them when used for downscaling tasks. Data assimilation uses information from three different sources, namely an imperfect computational model based on partial differential equations (PDE), from noisy observations, and from an uncertainty-reflecting prior. In this work, when carrying out dynamic downscaling, we replace the computationally expensive PDE-based NWP models with FourCastNet in a ``weak-constrained 4DVar framework" that accounts for the implied model errors. We demonstrate the efficacy of this approach for a hurricane-tracking problem; moreover, the 4DVar framework naturally allows the expression and quantification of uncertainty. We demonstrate, using ERA5 data, that our approach performs better than the ensemble Kalman filter (EnKF) and the unstabilized FourCastNet model, both in terms of forecast accuracy and forecast uncertainty.