Harmful algal blooms (HABs) in freshwaters are intensifying globally, posing serious threats to water quality, ecosystem health, and regional economies. While total nitrogen loads from land to freshwaters have long been recognized as a key driver of HABs, recent findings highlight that the chemical form of nitrogen is just as or more important than the total load of nitrogen as a control on HABs. Specifically, recent work suggests that dissolved organic nitrogen (DON) and its photochemical products are forms of nitrogen critical for sustaining toxin-producing cyanobacteria such as Microcystis.
As a Schmidt AI in Science Fellow, Chen Zhao will develop a biogeochemistry-informed artificial intelligence (AI) framework to test for relationships between DON and the severity and toxicity of HABs in Lake Erie by integrating process-based biogeochemical knowledge with machine learning techniques. Leveraging a decade of high-resolution water quality data in Lake Erie, his research will develop customized model architectures to incorporate known biogeochemical constraints into predictive AI models for forecasting HABs in the Great Lakes. The proposed approach bridges the gap between data-driven modeling and biogeochemical process understanding, offering a robust AI framework for water quality management in a changing climate.