Environmental and social scientists rely on observational data to accurately model natural processes and human activities around the globe. Since these processes and activities do not occur in isolation, it is required to forecast how they affect each other by integrating different models. Model integration across disciplines requires resolving semantic, spatio-temporal, and execution mismatches, both at a data and software level. Today, these issues are largely resolved by hand and may take years of human effort to complete.

In this talk I will describe the main challenges we have faced when integrating hydrology, agronomy, climate and economic models; and how we are addressing these challenges with MINT, a Model INTegration framework that incorporates extensive knowledge about models and data. MINT includes 1) new principle-based ontology generation tools for modeling variables, used to describe models and data; 2) A novel workflow system that selects relevant models from a curated registry and uses abductive reasoning to hypothesize new models and data transformation steps; 3) A new data discovery and integration framework that learns to extract information from both online sources and remote sensing data and transforms the data into the format required by the models; 4) A novel framework for multi-modal scalable workflow execution.