Quantitative Climate Science Seminar | Carlos Pereira Frontado, Envisioning Future Technologies for Numerical Weather Prediction

Weather forecasting is undergoing a revolutionary transformation. At Environment and Climate Change Canada (ECCC), the Global Environmental Multiscale Model (GEM) has powered operational forecasts for the past few decades, but today's advances in artificial intelligence are fundamentally reshaping how we predict the atmosphere. As we begin this revolution, a natural path forward is to blend new data-driven insights with more established numerical methods for solving partial differential equations. At ECCC, breakthrough innovations like spectral nudging marry AI with classical numerics to boost accuracy, while we are simultaneously developing novel numerical schemes and a physics-informed AI framework, PARADIS, to push low-cost predictive skill even further. In this talk, we will present our hybrid forecasting vision, showcasing our work on high-order accurate numerical techniques and AI to deliver faster, more reliable weather forecasts that could help to save lives and protect our economy.