Using models to estimate microcystin concentrations in Ohio recreational and source waters
Cyanobacterial harmful algal blooms (cyanoHABs) and associated toxins, such as microcystin, are a major global water-quality issue. In Lake Erie and inland lakes in Ohio, elevated microcystin concentrations have caused water-resource managers to issue recreational water-quality advisories, and detections of microcystin in source waters have caused drinking-water plant managers to increase monitoring and adjust treatment methods (Ohio Environmental Protection Agency, 2015).
Models are needed to estimate toxin concentrations so that swimmers and boaters can be warned and water-treatment plants can take measures to avoid or appropriately treat toxin-laden waters. Satellite data have been used to predict and detect cyanoHABs in lakes and estuaries; however, this technology currently is not capable of directly detecting cyanobacterial toxins (Lunetta et al., 2015). Although satellite data are useful on a broad scale and as an initial warning system for cyanoHAB patterns of occurrence, site-specific models are needed to estimate the serious public health concern from toxin concentrations at a drinking-water intake or beach.
Predicting toxin concentrations is complicated and site specific because of the many factors affecting toxin production. In a previous study (Francy and others, 2015; Francy and others, 2016), data were collected weekly to monthly at Ohio recreational sites to identify factors that could be used to develop two types of models to estimate microcystin concentrations: (1) Real-time models that include easily- or continuously-measured factors and available environmental data that do not require a sample be collected and (2) comprehensive models that use results from samples collected and analyzed in a laboratory along with real-time factors. For real-time models, statistically significant correlations were found between microcystin concentrations and factors such as phycocyanin, turbidity, pH, streamflow from a nearby river, and water level change over 24 hours. Continuous water-quality measurements over multiple days showed the highest correlations to microcystin concentrations. For comprehensive models, statistically significant correlations to microcystin concentrations were found for nutrient constituents, cyanobacterial genes (measured by use of molecular methods), and measures of cyanobacterial biovolume or abundance.
To develop accurate site-specific models to estimate toxin concentrations, data need to be collected more frequently than weekly and for several consecutive days each week before, during, and after the cyanoHAB season (Francy and others, 2015). In addition, model results need to be available to water-resource managers and the public. An example of a real-time notification system is the Ohio Nowcast, where the probability of exceeding an E. coli threshold concentration is available to the public (http://ny.water.usgs.gov/maps/ohnowcast/); this type of system could be developed for cyanoHABs.
Goal:
The goals of this project are to develop models for real-time and comprehensive predictions of microcystin concentrations at two beach sites, building on the knowledge gained in the previous study, and to evaluate factors suitable for model development in source waters for drinking-water plants and at additional recreational sites.
Approach:
Sampling (May–October, 2016 and 2017)
Samples are collected by local agencies and collaborators (Graham and others, 2008) at least weekly at 11 sites in Ohio:
- Three beaches sites—MBSP Lake Erie, Harsha Campers beach, and Harsha Main beach
- One boater recreational site—Put-in-Bay
- Seven drinking water plants—Oregon, Carroll, Ottawa County, Marblehead, Alliance, Cadiz, and Clermont County
Real-time data
- Data on physical parameters (specific conductance, dissolved oxygen, pH, temperature, chlorophyll, and phycocyanin) from hand-held equipment and from the nearest continuous monitor in an existing network of monitors (http://habs.glos.us/map/).
- Environmental data for locations in close proximity to each sampling site from the National Oceanic and Atmospheric Administration (NOAA) and other sources. These include data for rainfall and wind speed and direction, water levels, streamflow, satellite imagery (cell counts), and solar radiation.
Data for comprehensive predictions
Samples will be preserved on ice and shipped to laboratories for:
- Microcystin by ELISA
- Dissolved and total nutrients
- Cyanobacterial genes by quantitative polymerase chain reaction (qPCR) for:
- General cyanobacteria
- General Microcystis, Dolichospermum, and Planktothrix
- mcyE toxin genes for Microsystis, Dolichospermum, and Planktothrix
- Genes for production of taste and odor compounds (inland lake sites)
(Rinta-Kanto and others, 2005; Doblin and others, 2007; Rantala and others, 2004 and 2006; Sipari and others, 2010) - Phytoplankton abundance and community structure by microscopy (subset of samples)
Data analysis
Data will be examined by use of time-series plots and Spearman’s correlation coefficients. The USGS will work with other agencies to check and qualify continuous monitor data per USGS protocols (Wagner et al., 2006), and create 3-day, 7-day, and 14-day moving averages to use in correlation analysis and model development. Models will be developed using Virtual Beach, a software program developed by USEPA as a decision support tool to construct site-specific multiple-linear regression models to predict pathogen concentrations at recreational beaches (U.S. Environmental Protection Agency, 2015). Virtual Beach can be used for other dependent variables, such as microcystin concentrations.
Data dissemination and reports
Discrete data will be entered by site identification number into the USGS database for water-quality data, available to the public through NWISWeb (http://waterdata.usgs.gov/oh/nwis/qw/). Two reports will be written when the project is complete.
References
Doblin, M.A., Coyne, K.J., Rinta-Kanto, J.M., Wilhelm, S.W., and Dobbs, F.C., 2007, Dynamics and short-term survival of toxic cyanobacteria species in ballast water from NOBOB vessels transiting the Great lakes—implications for HAB invasions: Harmful Algae, vol. 6., p. 519-530.
Francy, D.S., Graham, J.L., Stelzer, E.A., Ecker, C.D., Brady, A.M.G., Struffolino, P., and Loftin, K.A., 2015, Water quality, cyanobacteria, and environmental factors and their relations to microcystin concentrations for use in predictive models at Ohio Lake Erie and inland lake recreational sites, 2013–14: U.S. Geological Survey Scientific Investigations Report 2015–5120, 58 p., http://dx.doi.org/10.3133/sir20155120.
Francy, D.S., Brady, A.M.G., Ecker, C.D., Graham, J.L., Stelzer E.A., Struffolino, P., Dwyer, D.F., and Loftin K.A., 2016, Estimating microcystin concentrations at recreational sites in western Lake Erie and Ohio: Harmful Algae, v. 58, p. 23-34.
Lunetta, R.S., Schaeffer, B.A., Stumpf, R.P., Deith, D., Jacobs, S.A., Murphy, M.S. 2015. Evaluation of cyanobacteria cell count detection derived from MERIS imagery across the eastern USA. Remote Sensing of Envir. 157, 24‒34.
Ohio Environmental Protection Agency, 2015, State of Ohio—harmful algal bloom response strategy for recreational waters, accessed September 2014 at http://epa.ohio.gov/habalgae.aspx
Rantala, A., Fewer, D.P., Hisbergues, M., Rouhiainen, L., Vaitomaa, J., Borner, T., Sivonen, K., 2004, Phylogenetic evidence for the early evolution of microcystin synthesis: Proceedings of the National Academy of Science, v. 101, no. 2, p. 568-573.
Rantala, A., Rajaniemi-Wacklin, P., Lyra, C., Lepisto, L., Rintala, J., Mankiewicz-Boczek, J., and Sivonen, K., 2006, Detection of microcystin-producing cyanobacteria in Finnish lakes with genus-specific microcystin synthetase gene E (mcyE) PCR and associations with environmental factors: Appl. Environ. Microbiol. 72, 6101–6110.
Rinta-Kanto, J.M., Ouellette, A.J.A., Boyer, G.L., Twiss, M.R., Bridgeman, T.B., and Wilhelm, S.W., 2005, Quantification of toxic Microcystis spp. during the 2003 and 2004 blooms in western Lake Erie using quantitative real-time PCR: Environ. Sci. Technol. 39, 4198-4205.
Sipari, H., Rantala-Ylinen, A., Jokela, J., Oksanen, I., and Sivonen, K., 2010, Development of a chip assay and quantitative PCR for detecting microcystin synthetase E gene expression: Appl. Environ. Microbiol. 76, 3797–3805.
Wagner, R.J., Boulger, R.W., Jr., Oblinger, C.J., and Smith, B.A., 2006, Guidelines and standard procedures for continuous water-quality monitors—Station operation, record computation, and data reporting: U.S. Geological Survey Techniques and Methods 1–D3, 51 p. + 8 attachments; accessed April 10, 2006, at http://pubs.water.usgs.gov/tm1d3
U.S. Environmental Protection Agency, Center for Exposure Assessment Modeling, 2015, Exposure Assessment Models—Virtual Beach, accessed April 2015 at http://www2.epa.gov/exposure-assessment-models/virtual-beach-vb
Cyanobacterial harmful algal blooms (cyanoHABs) and associated toxins, such as microcystin, are a major global water-quality issue. In Lake Erie and inland lakes in Ohio, elevated microcystin concentrations have caused water-resource managers to issue recreational water-quality advisories, and detections of microcystin in source waters have caused drinking-water plant managers to increase monitoring and adjust treatment methods (Ohio Environmental Protection Agency, 2015).
Models are needed to estimate toxin concentrations so that swimmers and boaters can be warned and water-treatment plants can take measures to avoid or appropriately treat toxin-laden waters. Satellite data have been used to predict and detect cyanoHABs in lakes and estuaries; however, this technology currently is not capable of directly detecting cyanobacterial toxins (Lunetta et al., 2015). Although satellite data are useful on a broad scale and as an initial warning system for cyanoHAB patterns of occurrence, site-specific models are needed to estimate the serious public health concern from toxin concentrations at a drinking-water intake or beach.
Predicting toxin concentrations is complicated and site specific because of the many factors affecting toxin production. In a previous study (Francy and others, 2015; Francy and others, 2016), data were collected weekly to monthly at Ohio recreational sites to identify factors that could be used to develop two types of models to estimate microcystin concentrations: (1) Real-time models that include easily- or continuously-measured factors and available environmental data that do not require a sample be collected and (2) comprehensive models that use results from samples collected and analyzed in a laboratory along with real-time factors. For real-time models, statistically significant correlations were found between microcystin concentrations and factors such as phycocyanin, turbidity, pH, streamflow from a nearby river, and water level change over 24 hours. Continuous water-quality measurements over multiple days showed the highest correlations to microcystin concentrations. For comprehensive models, statistically significant correlations to microcystin concentrations were found for nutrient constituents, cyanobacterial genes (measured by use of molecular methods), and measures of cyanobacterial biovolume or abundance.
To develop accurate site-specific models to estimate toxin concentrations, data need to be collected more frequently than weekly and for several consecutive days each week before, during, and after the cyanoHAB season (Francy and others, 2015). In addition, model results need to be available to water-resource managers and the public. An example of a real-time notification system is the Ohio Nowcast, where the probability of exceeding an E. coli threshold concentration is available to the public (http://ny.water.usgs.gov/maps/ohnowcast/); this type of system could be developed for cyanoHABs.
Goal:
The goals of this project are to develop models for real-time and comprehensive predictions of microcystin concentrations at two beach sites, building on the knowledge gained in the previous study, and to evaluate factors suitable for model development in source waters for drinking-water plants and at additional recreational sites.
Approach:
Sampling (May–October, 2016 and 2017)
Samples are collected by local agencies and collaborators (Graham and others, 2008) at least weekly at 11 sites in Ohio:
- Three beaches sites—MBSP Lake Erie, Harsha Campers beach, and Harsha Main beach
- One boater recreational site—Put-in-Bay
- Seven drinking water plants—Oregon, Carroll, Ottawa County, Marblehead, Alliance, Cadiz, and Clermont County
Real-time data
- Data on physical parameters (specific conductance, dissolved oxygen, pH, temperature, chlorophyll, and phycocyanin) from hand-held equipment and from the nearest continuous monitor in an existing network of monitors (http://habs.glos.us/map/).
- Environmental data for locations in close proximity to each sampling site from the National Oceanic and Atmospheric Administration (NOAA) and other sources. These include data for rainfall and wind speed and direction, water levels, streamflow, satellite imagery (cell counts), and solar radiation.
Data for comprehensive predictions
Samples will be preserved on ice and shipped to laboratories for:
- Microcystin by ELISA
- Dissolved and total nutrients
- Cyanobacterial genes by quantitative polymerase chain reaction (qPCR) for:
- General cyanobacteria
- General Microcystis, Dolichospermum, and Planktothrix
- mcyE toxin genes for Microsystis, Dolichospermum, and Planktothrix
- Genes for production of taste and odor compounds (inland lake sites)
(Rinta-Kanto and others, 2005; Doblin and others, 2007; Rantala and others, 2004 and 2006; Sipari and others, 2010) - Phytoplankton abundance and community structure by microscopy (subset of samples)
Data analysis
Data will be examined by use of time-series plots and Spearman’s correlation coefficients. The USGS will work with other agencies to check and qualify continuous monitor data per USGS protocols (Wagner et al., 2006), and create 3-day, 7-day, and 14-day moving averages to use in correlation analysis and model development. Models will be developed using Virtual Beach, a software program developed by USEPA as a decision support tool to construct site-specific multiple-linear regression models to predict pathogen concentrations at recreational beaches (U.S. Environmental Protection Agency, 2015). Virtual Beach can be used for other dependent variables, such as microcystin concentrations.
Data dissemination and reports
Discrete data will be entered by site identification number into the USGS database for water-quality data, available to the public through NWISWeb (http://waterdata.usgs.gov/oh/nwis/qw/). Two reports will be written when the project is complete.
References
Doblin, M.A., Coyne, K.J., Rinta-Kanto, J.M., Wilhelm, S.W., and Dobbs, F.C., 2007, Dynamics and short-term survival of toxic cyanobacteria species in ballast water from NOBOB vessels transiting the Great lakes—implications for HAB invasions: Harmful Algae, vol. 6., p. 519-530.
Francy, D.S., Graham, J.L., Stelzer, E.A., Ecker, C.D., Brady, A.M.G., Struffolino, P., and Loftin, K.A., 2015, Water quality, cyanobacteria, and environmental factors and their relations to microcystin concentrations for use in predictive models at Ohio Lake Erie and inland lake recreational sites, 2013–14: U.S. Geological Survey Scientific Investigations Report 2015–5120, 58 p., http://dx.doi.org/10.3133/sir20155120.
Francy, D.S., Brady, A.M.G., Ecker, C.D., Graham, J.L., Stelzer E.A., Struffolino, P., Dwyer, D.F., and Loftin K.A., 2016, Estimating microcystin concentrations at recreational sites in western Lake Erie and Ohio: Harmful Algae, v. 58, p. 23-34.
Lunetta, R.S., Schaeffer, B.A., Stumpf, R.P., Deith, D., Jacobs, S.A., Murphy, M.S. 2015. Evaluation of cyanobacteria cell count detection derived from MERIS imagery across the eastern USA. Remote Sensing of Envir. 157, 24‒34.
Ohio Environmental Protection Agency, 2015, State of Ohio—harmful algal bloom response strategy for recreational waters, accessed September 2014 at http://epa.ohio.gov/habalgae.aspx
Rantala, A., Fewer, D.P., Hisbergues, M., Rouhiainen, L., Vaitomaa, J., Borner, T., Sivonen, K., 2004, Phylogenetic evidence for the early evolution of microcystin synthesis: Proceedings of the National Academy of Science, v. 101, no. 2, p. 568-573.
Rantala, A., Rajaniemi-Wacklin, P., Lyra, C., Lepisto, L., Rintala, J., Mankiewicz-Boczek, J., and Sivonen, K., 2006, Detection of microcystin-producing cyanobacteria in Finnish lakes with genus-specific microcystin synthetase gene E (mcyE) PCR and associations with environmental factors: Appl. Environ. Microbiol. 72, 6101–6110.
Rinta-Kanto, J.M., Ouellette, A.J.A., Boyer, G.L., Twiss, M.R., Bridgeman, T.B., and Wilhelm, S.W., 2005, Quantification of toxic Microcystis spp. during the 2003 and 2004 blooms in western Lake Erie using quantitative real-time PCR: Environ. Sci. Technol. 39, 4198-4205.
Sipari, H., Rantala-Ylinen, A., Jokela, J., Oksanen, I., and Sivonen, K., 2010, Development of a chip assay and quantitative PCR for detecting microcystin synthetase E gene expression: Appl. Environ. Microbiol. 76, 3797–3805.
Wagner, R.J., Boulger, R.W., Jr., Oblinger, C.J., and Smith, B.A., 2006, Guidelines and standard procedures for continuous water-quality monitors—Station operation, record computation, and data reporting: U.S. Geological Survey Techniques and Methods 1–D3, 51 p. + 8 attachments; accessed April 10, 2006, at http://pubs.water.usgs.gov/tm1d3
U.S. Environmental Protection Agency, Center for Exposure Assessment Modeling, 2015, Exposure Assessment Models—Virtual Beach, accessed April 2015 at http://www2.epa.gov/exposure-assessment-models/virtual-beach-vb