Recent advances in deep learning, agentic AI, and related machine learning methods are expanding the frontiers of climate and Earth system science. Techniques such as diffusion models, normalizing flows, variational autoencoders, LSTMs, vision transformers, foundation models, physics-informed neural networks, graph neural networks, and AI agents are creating new opportunities to understand, simulate, and predict complex environmental processes. This session will showcase applications to critical climate challenges, including extreme flood modeling, wildfire progression, drought detection, Earth system emulation, and climate data analysis. It will also highlight how agentic AI can support multistep scientific workflows, integrate diverse data sources, assist with hypothesis and scenario analysis, and accelerate discovery. By combining advanced neural architectures, generative AI, autonomous agents, and hybrid physics-AI methods, researchers are improving predictive accuracy, computational efficiency, and scientific insight. Attendees will gain practical perspective on how these emerging tools can strengthen climate resilience and decision-making.

• A - Atmospheric Sciences
• IN - Informatics