Spatiotemporal Data Fusion for Precipitation Nowcasting

TL;DR

This paper introduces a novel data fusion pipeline that combines radar and satellite data using computer vision techniques, including a new inpainting algorithm, to improve global precipitation nowcasting.

Contribution

It presents a new data fusion method with a novel inpainting algorithm for integrating radar and satellite data in precipitation prediction.

Findings

Enhanced precipitation nowcasting accuracy

Effective fusion of radar and satellite data

Novel inpainting algorithm improves data quality

Abstract

Precipitation nowcasting using neural networks and ground-based radars has become one of the key components of modern weather prediction services, but it is limited to the regions covered by ground-based radars. Truly global precipitation nowcasting requires fusion of radar and satellite observations. We propose the data fusion pipeline based on computer vision techniques, including novel inpainting algorithm with soft masking.