WeatherReal: A Benchmark Based on In-Situ Observations for Evaluating Weather Models

Weixin Jin, Jonathan Weyn, Haiyu Dong

Keywords: Weather, Benchmark, In-situ Observation

TL;DR: AI-based weather forecasting has progressed, yet often relies on reanalysis datasets; this paper introduces WeatherReal, a benchmark derived from global in-situ observations, enhancing model training and evaluation.

Confirmation Of Submission Requirements: I submit an abstract. It uses the template provided on the submission page and is no longer than 2 pages.

Submission Number: 121