Su3plus: An Enhanced Swin-Unet Architecture for Synthesizing Vertically Integrated Liquid from Multiple Meteorological Satellites

Abstract

Radar composite reflectivity is a crucial component of Earth observation data, playing a significant role in applications such as weather forecasting and climate disaster tracking. Due to the deployment challenges and limited coverage of meteorological radars, it is impossible to collect corresponding radar reflectivity in areas like mountains and oceans. In such cases, using deep learning methods to reconstruct radar reflectivity from meteorological satellite data, which has higher coverage, becomes an effective solution.  However, Earth observation data is complex and exhibits strong long-range dependencies. Such data characteristics require the ability to model long-distance dependencies, making the Transformer more suitable for this specific data. With the ultimate goal of producing high reconstruction quality, we propose in this paper a novel architecture, designated by SU3plus, based on the Swin Transformer with a multi-scale feature fusion mechanism, while leveraging multiple channels of satellite data. Moreover, we define an appropriate loss function by combining the weighted versions of standard metrics, to take into account the data distribution imbalance and improve the reconstruction performance.  Extensive experiments, carried out on the SEVIR storm dataset, confirm the effectiveness of the proposed approach compared to several state-of-the-art models.

Keywords: Meteorological satellite data, radar composite reflectivity, data reconstruction, neural network, fusion.