Understanding Earth as a coupled system demands integrated, cross-timescale modeling and analysis capabilities. Priority science questions require unified representation across atmosphere, ocean, land surface, hydrology, cryosphere, sea ice, and human systems, with consistent treatment across reanalysis, prediction, and projection. Advances in satellite observations, computing architectures, and AI/ML methods, including foundation models, hybrid physics-ML approaches, and scalable data assimilation, open new pathways for integrating heterogeneous observations with process-based models, yet also amplify the need for coherent systems and clear transition pathways from research to applications. This session seeks contributions addressing: How do we bring disparate Earth system components into consistent frameworks from reanalysis through subseasonal prediction to climate projection? How do we leverage these advances to sharpen observational requirements, inform mission design, and accelerate uptake of new measurements? And how do we demonstrate societal value through systematic benchmarking, defined AI/ML integration targets, and structured approaches linking applications to observing system planning?

• A - Atmospheric Sciences
• EP - Earth and Planetary Surface Processes
• H - Hydrology
• OS - Ocean Sciences