The escalating impacts of climate change, manifested through the heightened frequency and severity of extreme weather events around the world, underscore the critical need to advance weather forecast models.
The notable success of the Pangu-Weather model and GraphCast in medium-range weather forecasting has revolutionised operational weather forecasts, prompting a paradigm shift towards adopting deep learning-based methods for more accurate climate modelling and predictions. Climate scientists are calling for the utilisation of AI to propel climate modelling on a broader scale (Schneider et al., 2023). Amid these global developments, Hong Kong and the Greater Bay Area are facing their own pressing need to fortify their local forecasting capabilities.
The proposed ASI seeks to address the urgency of improving weather forecasts and climate predictions by providing a platform for exchanging ideas, fostering collaborations, and discussing the latest advancements in AI foundation models for weather and climate.
Key topics
Some of the key areas for discussion will be:
- The latest advancements in AI, such as:
- deep learning,
- generative AI, and
- foundation models for weather and climate studies,
- The fusion of AI and physics models and unified weather-climate modelling
- Models to be taught include AIFS, GraphCast, FuXi, FengWu, Pangu-Weather, and ClimaX, among others.
Who the Institute is for
Early career researchers in fields relevant to the topic as well as practitioners
Course director
Global STEM Professor, Hong Kong University of Science and Technology