SynWeather: Weather Observation Data Synthesis across Multiple Regions and Variables via a General Diffusion Transformer

TL;DR

SynWeather introduces a comprehensive dataset and a probabilistic diffusion transformer model for unified multi-region and multi-variable weather data synthesis, overcoming limitations of existing deterministic approaches and over-smoothing issues.

Contribution

We present SynWeather, the first dataset for multi-region, multi-variable weather synthesis, and SynWeatherDiff, a diffusion transformer model that improves probabilistic weather data generation.

Findings

SynWeather dataset covers four major regions and key weather variables.

SynWeatherDiff outperforms existing models in weather data synthesis tasks.

The approach effectively addresses over-smoothing in probabilistic weather modeling.

Abstract

With the advancement of meteorological instruments, abundant data has become available. Current approaches are typically focus on single-variable, single-region tasks and primarily rely on deterministic modeling. This limits unified synthesis across variables and regions, overlooks cross-variable complementarity and often leads to over-smoothed results. To address above challenges, we introduce SynWeather, the first dataset designed for Unified Multi-region and Multi-variable Weather Observation Data Synthesis. SynWeather covers four representative regions: the Continental United States, Europe, East Asia, and Tropical Cyclone regions, as well as provides high-resolution observations of key weather variables, including Composite Radar Reflectivity, Hourly Precipitation, Visible Light, and Microwave Brightness Temperature. In addition, we introduce SynWeatherDiff, a general and probabilistic weather synthesis model built upon the Diffusion Transformer framework to address the over-smoothed problem. Experiments on the SynWeather dataset demonstrate the effectiveness of our network compared with both task-specific and general models.