Intelligent Systems Division

New ISI project receives $13 million award to model and predict impact of human activities on water and food

Forecasting how natural processes and human activities affect one another can help address major societal and environmental challenges. In many areas of the globe, climate affects water resources and therefore food availability, with significant economic and social implications. 

“When you try to understand problems or conflicts in a region, such as civil war or mass migration, there is often a history of human activity impacting natural resources,” says Yolanda Gil, the principal investigator of a new $13 million Defense Research Projects Award (DARPA) program called MINT (for Model INTegration).

“The question is how well can we understand and predict how each region will respond to certain climate dynamics or natural resource depletion, or see what’s going to happen in the future.”

Today, such analyses require significant effort to integrate different models from separate disciplines, including climate, hydrology, agriculture, economics, and social sciences. The process can take more than two years, by which time there could already be problems stirring in the region.

By developing a sophisticated modeling environment, the MINT project aims to significantly reduce the time needed to develop new and accurate integrated models to analyze disparate data, and ultimately generate a solution or response to an issue.

“Using an integrated model, we can see, for example, how different levels of rainfall can effect an event like mass migration, and to what degree,” says Gil.

“This provides actionable insights for policymakers. You cannot control the climate but you can control other things, for example, policies for how much water to store in reservoirs and how much agriculture and water consumption to allow.”

The project’s co-principal investigators are ISI researchers Ewa Deelman, Rafael Ferreira da Silva and Craig Knoblock, along with academics from Virginia Tech, the Pennsylvania State University, the University of Minnesota and the University of Colorado.

The team’s areas of expertise include artificial intelligence, model coupling, ontologies, knowledge graphs, semantic workflows, high-performance computing, hydrology, climate, agriculture and economics.

Over the course of the next four years, the researchers will integrate models of human activities and how these impact natural resources to create predictions for specific regions based on real-life scenarios, such as climate change, deforestation and crop cultivation.