ABSTRACT

The NSF Convergence Accelerator supports use-inspired, team-based, multidisciplinary efforts that address challenges of national importance and will produce deliverables of value to society in the near future.

This project on Water System Data Pooling for Climate Vulnerability Assessment and Warning System addresses a major gap in the resiliency of America's Water Supply, viz., resiliency to climate variability and change, especially focusing on the vulnerability of thousands of smaller utilities in the United States that may lack the financial wherewithal and technical capacity to analyze these risks and assess their impact on operations. This project establishes a convergence research agenda by bringing together experts in water systems, climate science, AI technologies, emulation models and software development for the conceptual design, development, and sharing of Artificial Intelligence (AI) and Machine Learning (ML) models to quantify America's water supply risk at the level of water utilities and their regulatory state and federal agencies. The aggregated data sources and scalable models for climate and water risk analyses will be made available and accessible to all communities interested in this information. The project employs AI-based techniques to facilitate the exploration of climate observations, climate model simulations and corresponding water system response to help create breakthroughs in our understanding of water supply risk. The models developed will assist in the strategic planning and operations of water systems in the face of an increasing frequency of floods and droughts under climate change and aging infrastructure conditions?factors that constitute significant risks to the nation?s safe supply of water.

A cloud-based, multi-scale AI-enabled modeling, and model and data sharing, platform will be developed to support user-centric analyses for the water supply industry. The platform provides multiscale modeling for feature identification, spatiotemporal modeling and forecasting, functional dependence, inverse problems and transfer learning. Physics-based models as well as AI models will be explored in this context. A diverse set of data sources will be used, including national-scale water data will along with utility-collected data. The outputs will be responsive to identified user needs and will become a community data and modeling resource. A deep collaboration with industry partners via the Columbia Water Center?s America?s Water initiative and through the University of Massachusetts?s Water Innovation Network for Sustainable Small Systems guides this process. A broad range of organizations and their constituents will be engaged via webinars and on-site training and demonstrations about the platform. A particular focus of this effort is on organizations and representatives of underrepresented communities that are especially vulnerable to climate driven disruption. Educational materials will be targeted towards users from such communities with the goal of developing additional trained individuals locally who could support the use and interpretation of ML tools for water risk analysis in a local system context. Outreach activities will be especially targeted to the smaller water utilities who may be resource constrained and can, thus, benefit from the shared platform that will be created.

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.

Please report errors in award information by writing to: awardsearch@nsf.gov.