Conditional Diffusion Models for Global Precipitation Map Inpainting

TL;DR

This paper introduces a conditional diffusion model-based method for inpainting incomplete satellite precipitation maps, leveraging spatio-temporal data to produce more consistent global precipitation estimates.

Contribution

It formulates precipitation map completion as a video inpainting task and employs a 3D U-Net with a condition encoder, a novel approach for this application.

Findings

Outperforms conventional inpainting methods in spatio-temporal consistency.

Uses ERA5 data to train and pseudo-GSMaP masks for evaluation.

Demonstrates potential to enhance global precipitation monitoring.

Abstract

Incomplete satellite-based precipitation presents a significant challenge in global monitoring. For example, the Global Satellite Mapping of Precipitation (GSMaP) from JAXA suffers from substantial missing regions due to the orbital characteristics of satellites that have microwave sensors, and its current interpolation methods often result in spatial discontinuities. In this study, we formulate the completion of the precipitation map as a video inpainting task and propose a machine learning approach based on conditional diffusion models. Our method employs a 3D U-Net with a 3D condition encoder to reconstruct complete precipitation maps by leveraging spatio-temporal information from infrared images, latitude-longitude grids, and physical time inputs. Training was carried out on ERA5 hourly precipitation data from 2020 to 2023. We generated a pseudo-GSMaP dataset by randomly applying GSMaP masks to ERA maps. Performance was evaluated for the calendar year 2024, and our approach produces more spatio-temporally consistent inpainted precipitation maps compared to conventional methods. These results indicate the potential to improve global precipitation monitoring using the conditional diffusion models.