Risk assessment of toxic cyanobacterial blooms in recreational waters: A comparative study of monitoring methods.

Toxic cyanobacterial blooms impose a health risk to recreational users, and monitoring of cyanobacteria and associated toxins is required to assess this risk. Traditionally, monitoring for risk assessment is based on cyanobacterial biomass, which assumes that all cyanobacteria potentially produce toxins. While these methods may be cost effective, relatively fast, and more widely accessible, they often lead to an overestimation of the health risk induced by cyanotoxins. Monitoring methods that more directly target toxins, or toxin producing genes, may provide a better risk assessment, yet these methods may be more costly, usually take longer, or are not widely accessible. In this study, we compared six monitoring methods (fluorometry, microscopy, qPCR of 16S and mcyE, ELISA assays, and LC-MS/MS), of which the last three focussed on the most abundant cyanotoxin microcystins, across 11 lakes in the Netherlands during the bathing water season (May-October) of 2019. Results of all monitoring methods significantly correlated with LC-MS/MS obtained microcystin levels (the assumed 'golden standard'), with stronger correlations for methods targeting microcystins (ELISA) and microcystin genes (mcyE). The estimated risk levels differed substantially between methods, with 78 % and 56 % of alert level exceedances in the total number of collected samples for fluorometry and microscopy-based methods, respectively, while this was only 16 % and 6 % when the risk assessment was based on ELISA and LC-MS/MS obtained toxin concentrations, respectively. Integrating our results with earlier findings confirmed a strong association between microcystin concentration and the biovolume of potential microcystin-producing genera. Moreover, using an extended database consisting of 4265 observations from 461 locations across the Netherlands in the bathing water seasons of 2015 - 2019, we showed a strong association between fluorescence and the biovolume of potentially toxin-producing genera. Our results indicate that a two-tiered approach may be an effective risk assessment strategy, with first a biomass-based method (fluorometry, biovolume) until the first alert level is exceeded, after which the risk level can be confirmed or adjusted based on follow-up toxin or toxin gene analyses.