I’m an atmospheric scientist passionate about using AI to predict typhoons and extreme weather. My work focuses on combining AI with traditional weather models to give communities earlier, more accurate storm warnings – especially in coastal areas most vulnerable to climate change. What excites me about this Forum is how it bridges groundbreaking science with real-world problem solving. I’d love to discuss with mathematicians how to make AI predictions more trustworthy when dealing with chaotic weather systems. At the same time, sharing my experience in translating complex forecasts into simple risk alerts could help others improve science communication. I also want to explore how we can use AI creatively – like making storm prediction maps more visual and intuitive. This directly supports the Lee Shau Kee Foundation’s goal of using knowledge to empower communities.

The Hong Kong University of Science and Technology

Extreme wind and precipitation events cause significant societal disruptions in the South China coastal region, primarily triggered by tropical cyclones (TCs) or mesoscale storms. The atmospheric circulation processes across large, meso-, and small scales that influence these high-impact weather events may be modified by climate change, potentially altering TC characteristics. Furthermore, climate-driven variations in atmospheric chemical composition could exacerbate extreme weather phenomena, particularly precipitation intensity.
However, quantifying the sensitivity of TCs and extreme precipitation to climate change and air quality remains challenging, largely due to the limited spatial detail in coarse-resolution global model simulations. To address this, I employ a deep learning-enhanced framework that integrates high-resolution numerical weather modeling with climate projections. This approach enables detailed simulation of extreme events and evaluation of their climate change responses.
Concurrently, my research quantifies chemical composition changes under clean versus polluted emission scenarios, specifically investigating how air quality variations modulate extreme precipitation mechanisms. Through this dual focus on dynamic and chemical drivers, the study aims to improve predictive understanding of compound climate risks.