LAUSR

Abstract Cyanobacteria blooms can cause public health concerns related to drinking water quality and water recreation. The rapidly changing global climate is anticipated to bring about an increased frequency of extreme weather events (e.g. stronger storms, more extensive droughts), which are expected to promote more frequent cyanobacteria blooms that persist for longer durations in freshwater. Land use planning, landscape management, and ecological engineering may present mitigation opportunities for decreasing the occurrence and intensity of current and future cyanobacteria blooms through improved nutrient management strategies thereby reducing eutrophication of watersheds. To examine the potential impacts of various land cover classes (and their relative density) on cyanobacteria bloom coverage, county-level data were obtained or generated from the National Land Cover Database and the national nutrient inputs to the land surface database. These data were paired with county-level estimates of cyanobacteria bloom area obtained by satellite-based MERIS (Medium Resolution Imaging Spectrometer). Multivariable zero-inflated beta regression models were constructed for the U.S. and five U.S. regions for assessing the relationships between the proportion of county area experiencing a cyanobacteria bloom, county land cover types, and nutrient loading. The land cover type associated with the greatest decreases in bloom area in the national model was deciduous forest ( p   0.001). Open water extent ( p  = 0.001) and nitrogen loading from manure ( p  = 0.002) and fertilizer ( p   0.001) were positively associated with the proportion of water characterized as experiencing a cyanobacteria bloom. A significant interaction ( p   0.001) was observed between cultivated crop coverage and open water extent. Overall, increasing cultivated crop coverage was associated with increasing proportions of cyanobacteria blooms. Low intensity, medium intensity, and high intensity development land uses were not associated with bloom coverage in the national model, although development land uses were positively associated in several regional models. Ultimately, there is evidence that county-level land cover and nutrient loading, notably N in the national model, can impact county-level cyanobacteria bloom coverage. Given regional model differences, additional remote sensing-based studies that examine watershed-based effects on cyanobacteria coverage are needed to establish watershed-specific associations. Studies that transcend county boundaries may provide greater utility than this correlational study for better characterizing land uses and mitigation measures that impact or could impact cyanobacteria bloom coverage in U.S. surface waters.