Collaborative Research: CAIG: Physics Informed Graph Neural Network-based Snow Water Equivalent Forecasting

Understanding the dynamics of snow water equivalent (SWE) is vital for effective water resource management, especially in the Western United States where snowpack serves as a major water source. Accurate SWE forecasting is a significant challenge due to the complex interactions among snow physics, atmospheric conditions, and varied terrains. This project aims to revolutionize SWE prediction by integrating cutting-edge artificial intelligence (AI) techniques, specifically physics-informed neural networks (PINNs).

In addition to advancing scientific knowledge about snow water processes, this project is expected to have positive societal impacts, such as improved water resource management and informed decision-making in response to climate change. The project will also enable inclusivity and education by involving graduate students and underrepresented groups in AI research, fostering a diverse community of future experts in SWE forecasting research.

This project will employ an innovative approach that combines graph neural network models with physics-based constraints and partial differential equations. This integration will enable the creation of more accurate and reliable SWE forecasts by capturing the detailed processes of snow accumulation and melt. The GeoWeaver workflow management platform will be utilized for making advanced AI tools accessible to researchers and practitioners.

The project also includes a series of hackathon-style workshops providing students and snow researchers with hands-on experience in AI and SWE forecasting. Overall, the project seeks to democratize access to AI research workflows and tools for snow researchers, foster interdisciplinary collaboration, and support sustainable resource management, thereby enhancing our understanding of water resources and contributing to the broader discourse on climate change and water sustainability.

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.