Predicting Cyanobacteria Abundance and Microcystin Detection in 125,000 On-Network US Lakes
On this page:
The presence of cyanobacterial harmful algal blooms (cyanoHABs) in freshwater lakes across the United States poses a serious threat to ecosystems and human health. National-scale datasets allow for engagement with broad spatial patterns such as watershed and waterbody specific characteristics that can increase or diminish the risk of cyanoHABs. With spatially explicit regression models, we use lake morphology, watershed nutrient input, land cover, and geographic and meteorological data to predict cyanobacteria abundance and the probability of the cyanotoxin microcystin in approximately 124,500 US lakes. Approximately 62.2% of lakes were predicted to exceed a high threshold of 100,000 cyanobacterial cells/mL at least once during the summer, and 14.8% had a high probability of detectable microcystin during this same period. Many of these predictions were expected based on anthropogenic nutrient inputs. Inconsistent with our expectations, 33.5% of the high cyanobacteria predictions and 6.5% of likely microcystin detections occurred in watersheds with low nutrient inputs. Within these subsets of lakes, we investigated other drivers influencing cyanoHABs beyond anthropogenic nutrient inputs and the model covariates. Lakes with low nutrient inputs and high cyanoHAB risk were observed in watersheds with higher drainage ratios (watershed:lake area) than other lakes. In addition, lakes with higher area:depth ratios (lake surface area compared to depth) had higher cyanoHABs risk in both high and low nutrient groups, providing additional insight into the increased risk associated with shallower lakes. This research can help determine how watershed and lake characteristics, in addition to nutrient supply, may inform algal bloom mitigation efforts.