Watch an AI Weather Model Learn (and Unlearn) Tropical Cyclones

TL;DR

This paper investigates how a Spherical Fourier Neural Operator learns and unlearns tropical cyclones during training, revealing insights into the model's learning dynamics and environmental influences on extreme weather event modeling.

Contribution

It introduces a method to analyze training dynamics of AI weather models for tropical cyclones, highlighting unlearning patterns linked to environmental factors.

Findings

SFNO learns and then unlearns storm intensity information.

Unlearning is associated with moist environmental conditions.

Insights into model learning dynamics for extreme weather events.

Abstract

In a changing climate, artificial intelligence (AI) weather models have the potential to provide cheaper, faster, and more accurate forecasts of high-impact weather events. To realize this potential and gauge trustworthiness, there is a need for more research on how models learn extreme events and how that learning might be improved. Here, we investigate how a Spherical Fourier Neural Operator (SFNO) learns tropical cyclones (TCs) by saving every checkpoint from training and analyzing storm specific metrics. We find evidence that for some storms the SFNO learns information about TC intensity that it loses later in training. This unlearning pattern is associated with anomalously moist environments and may be due to the model unlearning the relationship between moisture and TC intensity. This work provides a first example of leveraging task-specific training dynamics to further our understanding of how AI weather models learn extreme events.