MetMamba: Regional Weather Forecasting with Spatial-Temporal Mamba Model
This work addresses the need for improved regional weather forecasting models, offering incremental advancements in backbone architecture and limited area modeling for meteorology applications.
The paper tackled the problem of regional weather forecasting by developing MetMamba, a deep learning model based on the Mamba state-space model, which achieved notable performance gains over traditional attention-based and neural operator backbones, and demonstrated the feasibility of limited area modeling through coupled training with a global host model.
Deep Learning based Weather Prediction (DLWP) models have been improving rapidly over the last few years, surpassing state of the art numerical weather forecasts by significant margins. While much of the optimization effort is focused on training curriculum to extend forecast range in the global context, two aspects remains less explored: limited area modeling and better backbones for weather forecasting. We show in this paper that MetMamba, a DLWP model built on a state-of-the-art state-space model, Mamba, offers notable performance gains and unique advantages over other popular backbones using traditional attention mechanisms and neural operators. We also demonstrate the feasibility of deep learning based limited area modeling via coupled training with a global host model.