Spatial and temporal assessment of the risk associated to bacteria in recreational waters of a large South American Reservoir

Abstract

The characterization of risk due to recreational exposure to water needs to assess the concentration of pathogens in the water and the degree of contact with those pathogens. In this study we assessed the risk associated to E. coli and cyanobacteria in a large South American Reservoir heavily used for recreation, by i) gathering field environmental data from two water agencies (six sites, summers 2011-2015), ii) generating satellite data at landscape scale (750 km2, summers 2011-2017) and running a health survey related to water exposure (summer 2017). Field data showed that cyanobacteria abundances recurrently surpassed the moderate and high-risk categories across sites and year analyzed, and a significant positive link between cyanobacteria abundance and microcystin concentration. Nevertheless, microcystin concentrations were in 90% of cases mostly within the low to moderate risk categories. Mean E. coli concentrations during 2011-2015 were within the high-risk category in 30% of the sites, but in 2017, sites identified as low risk had high-risk. The latter underscores the high risk posed by E. coli in the reservoir. Cyanobacteria (cell abundance and microcystin) and coliform bacteria abundances were unrelated, suggesting different responses to environmental or anthropogenic triggers. Satellite data evidenced that the highest risk related to cyanobacteria abundance occurred in the dendritic areas of the Argentinean side of the reservoir, areas which currently remain unmonitored by water agencies. Satellite monitoring bridged the limited spatial and temporal coverage of field samplings for cyanobacteria abundance (yet not for toxicity nor E. coli abundances) and rendered a risk map at landscape scale, which can be used by water agencies to effectively monitor and manage cyanobacteria blooms, and to-coupled with exposure variables-assess health risks related to cyanobacteria. The health survey identified few numbers of suspected patients with symptoms and who bathed in the Salto Grande reservoir. At the time of exposure, sites in the environment evidenced high bacteria concentration (mostly E. coli and to a lesser extent cyanobacteria) denoting situations where aspects of the biophysical environment affect human health. More studies and integration among environmental and health disciplines are needed to assess the impacts of these water born bacteria in human health. Finally, we further assessed how well cyanobacteria quantitative proxies monitored in the field explained the outcome of a qualitative risk communication system-the cyano-traffic-light-which is ongoing since 2011. We obtained a significant predictive model only for cyanobacteria abundance, yet with low predictive value. This probably occurred because the variables used to build each cyano-traffic-light category (cyanobacteria abundance, toxicity and chlorophyll-a, scums) were monitored with different frequencies, and because at least two of these variables needed to surpass the threshold of each category to be allocated into a risk category. Based upon our results we propose several modifications to the current cyano-traffic-light, that believe would better reflect what happens in the field and protect human health: i) include E. coli concentration and satellite estimated cyanobacteria abundance (mostly in areas not covered by field monitoring), ii) relax the thresholds for cyanobacteria abundance and toxicity, and iv) base each risk category upon the surpassing of one of either E. coli, cyanobacteria abundance, microcystin.