Citation:

Handler, A., J. Compton, R. Hill, S. Leibowitz, AND B. Schaeffer. Identifying lakes at risk of toxic cyanobacterial blooms using satellite imagery and field surveys across the United States. SCIENCE OF THE TOTAL ENVIRONMENT. Elsevier BV, AMSTERDAM, Netherlands, 869:161784, (2023). https://doi.org/10.1016/j.scitotenv.2023.161784

Impact/Purpose:

Harmful algal blooms caused by toxin-producing cyanobacteria are a threat to human health and the environment. Water managers need more tools to help prioritize monitoring resources and inform decisions about water body closures. This study develops an approach that combines satellite imagery with field data to predict which of the largest lakes in the contiguous US are at-risk for developing cyanobacterial harmful algal blooms (cyanoHABs). Satellite imagery from the Cyanobacteria Assessment Network (CyAN) collected 2008-2011 was combined with field data from the 2007 and 2012 National Lakes Assessments (NLA) to model the odds of exceeding lower and higher demonstration thresholds for microcystin toxin (0.2 and 1.0 µg/L), cyanobacteria abundance (20,000 and 100,000 cells/mL), and chlorophyll a (10 and 50 µg/L) in large lakes (>1.25 km2) across the US. Across models and lakes, a unit increase in bloom magnitude of 0.01 CI_cyano/km2 in the satellite data was associated with a 23-54% increase in the odds of exceeding the thresholds in the field. When applied to all 2,192 satellite-monitored lakes, the models identified at most 335 lakes that have a high probability (>75%) of exceeding the lower thresholds for cyanoHABs and at most 70 lakes exceeding the higher thresholds within the 2007-2012 period. When combined with contemporary field data, this approach can be used to help identify large lakes in a state or region that at-risk of exceeding thresholds for cyanoHABs. This capability enables water managers to prioritize more time- and resource-intensive field monitoring for the lakes that have the highest risk of cyanoHABs. The inclusion of microcystin in our models is an important feature as water managers are increasingly using cyanoHAB toxins to assess risk to human health. Our approach differs from previous efforts to relate satellite data to microcystin by developing an occurrence model of microcystin at a specified threshold across lakes rather than a continuous relationship within a particular lake or set of lakes. Future applications of our approach, combined with real-time satellite imagery and updated field sampling, could potentially inform issuance of water advisories.

Description:

Harmful algal blooms caused by cyanobacteria are a threat to global water resources and human health. Satellite remote sensing has vastly expanded spatial and temporal data on lake cyanobacteria, yet there is still acute need for tools that identify which waterbodies are at-risk for toxic cyanobacterial blooms. Algal toxins cannot be directly detected through imagery but monitoring toxins associated with cyanobacterial blooms is critical for assessing risk to the environment, animals, and people. The objective of this study is to address this need by developing an approach relating satellite imagery on cyanobacteria with field surveys to model the risk of toxic blooms among lakes. The Medium Resolution Imaging Spectrometer (MERIS) and United States (US) National Lakes Assessments are leveraged to model the probability among lakes of exceeding lower and higher demonstration thresholds for microcystin toxin, cyanobacteria, and chlorophyll a. By leveraging the large spatial variation among lakes using two national-scale data sources, rather than focusing on temporal variability, this approach avoids many of the previous challenges in relating satellite imagery to cyanotoxins. For every satellite-derived lake-level Cyanobacteria Index (CI_cyano) increase of 0.01 CI_cyano/km2, the odds of exceeding six bloom thresholds increased by 23–54 %. When the models were applied to the 2192 satellite monitored lakes in the US, the number of lakes identified with ≥75 % probability of exceeding the thresholds included as many as 335 lakes for the lower thresholds and 70 lakes for the higher thresholds, respectively. For microcystin, the models identified 162 and 70 lakes with ≥75 % probability of exceeding the lower (0.2 μg/L) and higher (1.0 μg/L) thresholds, respectively. This approach represents a critical advancement in using satellite imagery and field data to identify lakes at risk for developing toxic cyanobacteria blooms. Such models can help translate satellite data to aid water quality monitoring and management.

URLs/Downloads:

Record Details: