Predicting the Distribution of Cyanobacteria and Microcystin in US Lakes from Lake, Watershed, and Geographic Characteristics

With increasing concerns about freshwater cyanobacteria blooms, there is a need to identify which waterbodies are at risk for developing these blooms, especially those that produce cyanotoxins. To address this concern, we developed spatial statistical models using the US National Lakes Assessment, a survey with over 3500 spring and summer observations of cyanobacteria abundance and microcystin concentration in lakes across the conterminous US. We combined these observations with other nationally available data to model which lake, watershed, and climate factors best explain the presence of harmful cyanobacterial blooms. We found that both higher cyanobacteria abundance and probability of microcystin detection was associated with higher farm nitrogen inputs to the watershed. In addition, cyanobacteria abundance was also positively associated with farm phosphorus inputs and nitrogen inputs to developed areas. Lakes with low stormflow relative to baseflow, shallower depth, and longer fetch tended to have higher cyanobacteria abundance and probability of microcystin detection. The probability of microcystin detection was negatively associated with forested land cover and higher in the eastern and central portions of the US compared to the west. Both models had spatial covariance, indicating there are similarities in lakes that are closer geographically. The models fit data well with squared correlation coefficient between observed and fitted values for the cyanobacteria abundance of 0.48 and an area under the ROC curve of 0.92 for the probability of microcystin detection. These models can be used to predict the cyanobacteria abundance and risk of microcystin detection in lakes across the country based on the covariates and spatial correlation. As a result, the models can help identify which lakes are more vulnerable to harmful cyanobacteria blooms and in need of additional monitoring or communicating with communities near high-risk lakes.

Impact/Purpose

Harmful cyanobacterial blooms can produce toxins which impair freshwater ecosystems used for drinking water, recreation, and habitat for aquatic biota. With increased awareness and reporting of cyanobacteria blooms, water managers need information about how to prioritize monitoring locations in surface waters that may be affected. This presentation highlights analyses that combine data from the National Lakes Assessments field assessments and a variety of watershed, climate, and lake characteristic to model the risk of toxic blooms in lakes across the US. This presentation will be to an audience at the National Water Quality Monitoring Conference which is attended by resource managers and research scientists in aquatic resources.

Citation

Handler, A., J. Compton, M. Dumelle, M. Reynolds, M. Weber, L. Jansen, R. Hill, M. Brehob, M. Pennino, AND R. Sabo. Predicting the Distribution of Cyanobacteria and Microcystin in US Lakes from Lake, Watershed, and Geographic Characteristics. National Water Quality Monitoring Meeting, Green Bay, WI, March 10 - 14, 2025.