CAIG: Hybrid Artificial Intelligence-Geophysics Modeling for Understanding Carbon-Water Ecosystem Services Interactions

Carbon and water cycling jointly sustain many ecosystem services critical for the wellbeing of society. A fundamental challenge is predicting how these intertwined processes vary from local observations to landscape and regional dynamics. This challenge is especially acute for carbon-water ecosystem services, where tradeoffs and synergies vary across space, shift over time, and respond in complex ways to environmental changes. By addressing this challenge, this project improves understanding of how carbon and water cycles couple across scales and how interactions among land, water, and atmosphere shape landscape behavior. Better predictions of carbon–water ecosystem dynamics enhance economic resilience by reducing environmental risks and improving resource management. Key outcomes include open-access, multiscale datasets, modeling tools and infrastructure for assessing carbon-water ecosystem services and tradeoffs across scales; an interactive gamification decision-support platform; stakeholder engagement, and training for interdisciplinary scholars in Artificial Intelligence (AI) and Geosciences. The project investigates physics-based models with data-driven AI approaches to robustly predict carbon-water ecosystem services across space and time. The project combines field measurements, sensor networks, remote sensing, and modeling outputs of various spatial and temporal magnitudes with an agentic, physics-informed AI architecture. The framework uses adaptive AI agents that learn from high-resolution data and process-based models. This design enables scalable and emergent system behavior from local interactions while improving predictive capacity. This agentic structure integrates with gamification to support complex landscape and resource management. The framework is applied in Florida coastal plains, with pronounced carbon-water coupling, long-term data, and established system knowledge. By integrating AI with Geosciences, this project improves understanding of Earth system dynamics and supports better environmental decision-making. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.