Crop productivity is highly variable in time and space due to multiple interactions among genetics, soils, weather and management practices. The artificial intelligence (AI) and modeling team at UGA IIPA works to better understand and dissect the main drivers of crop yield and quality that includes digital soil mapping, weather forecasting and envirotyping at both the within-field and regional scales. For that, geospatial crop performance data and management and genetics information are matched with publicly available soils and weather big data. These complex data sets are then used to train a suite of machine learning and AI algorithms including tree ensembles, neural and deep networks, time series prediction, Bayesian, and mechanistic soil-plant-atmosphere models to untangle the main drivers of crop response.
These models are used to understand cropping systems productivity and resilience, and to explore alternative scenarios for within-field and regional crop management, as well as genetics placement. Complex big data and analytical engines deployment are made possible by the Georgia Advanced Computing Resource Center located at the UGA-Athens campus, a state-of-the-art high performance computer cluster available to IIPA members.