Predicting microcystin concentration action-level exceedances resulting from cyanobacterial blooms in selected lake sites in Ohio.

Cyanobacterial harmful algal blooms and the toxins they produce are a global water-quality problem. Monitoring and prediction tools are needed to quickly predict cyanotoxin action-level exceedances in recreational and drinking waters used by the public. To address this need, data were collected at eight locations in Ohio, USA, to identify factors significantly related to observed concentrations of microcystins (a freshwater cyanotoxin) that could be used in two types of site-specific regression models. Real-time models include easily or continuously-measured factors that do not require that a sample be collected; comprehensive models use a combination of discrete sample-based measurements and real-time factors. The study sites included two recreational sites and six water treatment plant sites. Real-time models commonly included variables such as phycocyanin, pH, specific conductance, and streamflow or gage height. Many real-time factors were averages over time periods antecedent to the time the microcystin sample was collected, including water-quality data compiled from continuous monitors. Comprehensive models were useful at some sites with lagged variables for cyanobacterial toxin genes, dissolved nutrients, and (or) nitrogen to phosphorus ratios. Because models can be used for management decisions, important measures of model performance were sensitivity, specificity, and accuracy of estimates above or below the microcystin concentration threshold standard or action level. Sensitivity is how well the predictive tool correctly predicts exceedance of a threshold, an important measure for water-resource managers. Sensitivities > 90% at four Lake Erie water treatment plants indicated that models with continuous monitor data were especially promising. The planned next steps are to collect more data to build larger site-specific datasets and validate models before they can be used for management decisions.

Estimating microcystin levels at recreational sites in western Lake Erie and Ohio.

Cyanobacterial harmful algal blooms (cyanoHABs) and associated toxins, such as microcystin, are a major global water-quality issue. Water-resource managers need tools to quickly predict when and where toxin-producing cyanoHABs will occur. This could be done by using site-specific models that estimate the potential for elevated toxin concentrations that cause public health concerns. With this study, samples were collected at three Ohio lakes to identify environmental and water-quality factors to develop linear-regression models to estimate microcystin levels. Measures of the algal community (phycocyanin, cyanobacterial biovolume, and cyanobacterial gene concentrations) and pH were most strongly correlated with microcystin concentrations. Cyanobacterial genes were quantified for general cyanobacteria, general Microcystis and Dolichospermum, and for microcystin synthetase (mcyE) for Microcystis, Dolichospermum, and Planktothrix. For phycocyanin, the relations were different between sites and were different between hand-held measurements on-site and nearby continuous monitor measurements for the same site. Continuous measurements of parameters such as phycocyanin, pH, and temperature over multiple days showed the highest correlations to microcystin concentrations. The development of models with high R2 values (0.81-0.90), sensitivities (92%), and specificities (100%) for estimating microcystin concentrations above or below the Ohio Recreational Public Health Advisory level of 6µgL-1 was demonstrated for one site; these statistics may change as more data are collected in subsequent years. This study showed that models could be developed for estimates of exceeding a microcystin threshold concentration at a recreational freshwater lake site, with potential to expand their use to provide relevant public health information to water resource managers and the public for both recreational and drinking waters.

Predictive models for Escherichia coli concentrations at inland lake beaches and relationship of model variables to pathogen detection.

Predictive models, based on environmental and water quality variables, have been used to improve the timeliness and accuracy of recreational water quality assessments, but their effectiveness has not been studied in inland waters. Sampling at eight inland recreational lakes in Ohio was done in order to investigate using predictive models for Escherichia coli and to understand the links between E. coli concentrations, predictive variables, and pathogens. Based upon results from 21 beach sites, models were developed for 13 sites, and the most predictive variables were rainfall, wind direction and speed, turbidity, and water temperature. Models were not developed at sites where the E. coli standard was seldom exceeded. Models were validated at nine sites during an independent year. At three sites, the model resulted in increased correct responses, sensitivities, and specificities compared to use of the previous day's E. coli concentration (the current method). Drought conditions during the validation year precluded being able to adequately assess model performance at most of the other sites. Cryptosporidium, adenovirus, eaeA (E. coli), ipaH (Shigella), and spvC (Salmonella) were found in at least 20% of samples collected for pathogens at five sites. The presence or absence of the three bacterial genes was related to some of the model variables but was not consistently related to E. coli concentrations. Predictive models were not effective at all inland lake sites; however, their use at two lakes with high swimmer densities will provide better estimates of public health risk than current methods and will be a valuable resource for beach managers and the public.

Comparison of filters for concentrating microbial indicators and pathogens in lake water samples.

Bacterial indicators are used to indicate increased health risk from pathogens and to make beach closure and advisory decisions; however, beaches are seldom monitored for the pathogens themselves. Studies of sources and types of pathogens at beaches are needed to improve estimates of swimming-associated health risks. It would be advantageous and cost-effective, especially for studies conducted on a regional scale, to use a method that can simultaneously filter and concentrate all classes of pathogens from the large volumes of water needed to detect pathogens. In seven recovery experiments, stock cultures of viruses and protozoa were seeded into 10-liter lake water samples, and concentrations of naturally occurring bacterial indicators were used to determine recoveries. For the five filtration methods tested, the highest median recoveries were as follows: glass wool for adenovirus (4.7%); NanoCeram for enterovirus (14.5%) and MS2 coliphage (84%); continuous-flow centrifugation (CFC) plus Virocap (CFC+ViroCap) for Escherichia coli (68.3%) and Cryptosporidium (54%); automatic ultrafiltration (UF) for norovirus GII (2.4%); and dead-end UF for Enterococcus faecalis (80.5%), avian influenza virus (0.02%), and Giardia (57%). In evaluating filter performance in terms of both recovery and variability, the automatic UF resulted in the highest recovery while maintaining low variability for all nine microorganisms. The automatic UF was used to demonstrate that filtration can be scaled up to field deployment and the collection of 200-liter lake water samples.

Comparative effectiveness of membrane bioreactors, conventional secondary treatment, and chlorine and UV disinfection to remove microorganisms from municipal wastewaters.

Log removals of bacterial indicators, coliphage, and enteric viruses were studied in three membrane bioreactor (MBR) activated-sludge and two conventional secondary activated-sludge municipal wastewater treatment plants during three recreational seasons (May-Oct.) when disinfection of effluents is required. In total, 73 regular samples were collected from key locations throughout treatment processes: post-preliminary, post-MBR, post-secondary, post-tertiary, and post-disinfection (UV or chlorine). Out of 19 post-preliminary samples, adenovirus by quantitative polymerase chain reaction (qPCR) was detected in all 19, enterovirus by quantitative reverse transcription polymerase chain reaction (qRT-PCR) was detected in 15, and norovirus GI by qRT-PCR was detected in 11. Norovirus GII and Hepatitis A virus were not detected in any samples, and rotavirus was detected in one sample but could not be quantified. Although culturable viruses were found in 12 out of 19 post-preliminary samples, they were not detected in any post-secondary, post-MBR, post-ultraviolet, or post-chlorine samples. Median log removals for all organisms were higher for MBR secondary treatment (3.02 to >6.73) than for conventional secondary (1.53-4.19) treatment. Ultraviolet disinfection after MBR treatment provided little additional log removal of any organism except for somatic coliphage (>2.18), whereas ultraviolet or chlorine disinfection after conventional secondary treatment provided significant log removals (above the analytical variability) of all bacterial indicators (1.18-3.89) and somatic and F-specific coliphage (0.71 and >2.98). Median log removals of adenovirus across disinfection were low in both MBR and conventional secondary plants (no removal detected and 0.24), and few removals of individual samples were near or above the analytical variability of 1.2 log genomic copies per liter. Based on qualitative examinations of plots showing reductions of organisms throughout treatment processes, somatic coliphage may best represent the removal of viruses across secondary treatment in both MBR and conventional secondary plants. F-specific coliphage and Escherichia coli may best represent the removal of viruses across the disinfection process in MBR facilities, but none of the indicators represented the removal of viruses across disinfection in conventional secondary plants.

Detection of human enteric viruses in stream water with RT-PCR and cell culture.

A multiplex RT-PCR method was used to measure virus occurrence at five stream water sites that span a range of hydroclimatic, water-quality, and land-use characteristics. The performance of the molecular method was evaluated in comparison with traditional cell culture and Escherichia coli membrane filtration assays. The study incorporated multiple quality controls and included a control for virus recovery during the sampling procedure as well as controls to detect potentially false-negative and false-positive data. Poliovirus recovery ranged from 16 to 65% and was variable, even in samples collected within the same stream. All five sites were positive for viruses by both molecular and cell culture-based virus assays. Enteroviruses, reoviruses, rotaviruses, and hepatitis A viruses were detected, but the use of the quality controls proved critical for interpretation of the molecular data. All sites showed evidence of faecal contamination, and culturable viruses were detected in four samples that would have met the US Environmental Protection Agency's recommended E. coli guideline for safe recreational water.

Effects of seeding procedures and water quality on recovery of Cryptosporidium oocysts from stream water by using U.S. Environmental Protection Agency Method 1623.

U.S. Environmental Protection Agency method 1623 is widely used to monitor source waters and drinking water supplies for Cryptosporidium oocysts. Matrix spikes, used to determine the effect of the environmental matrix on the method's recovery efficiency for the target organism, require the collection and analysis of two environmental samples, one for analysis of endemic oocysts and the other for analysis of recovery efficiency. A new product, ColorSeed, enables the analyst to determine recovery efficiency by using modified seeded oocysts that can be differentiated from endemic organisms in a single sample. Twenty-nine stream water samples and one untreated effluent sample from a cattle feedlot were collected in triplicate to compare modified seeding procedures to conventional seeding procedures that use viable, unmodified oocysts. Significant negative correlations were found between the average oocyst recovery and turbidity or suspended sediment; this was especially apparent in samples with turbidities greater than 100 nephelometric turbidity units and suspended sediment concentrations greater than 100 mg/liter. Cryptosporidium oocysts were found in 16.7% of the unseeded environmental samples, and concentrations, adjusted for recoveries, ranged from 4 to 80 oocysts per 10 liters. Determining recovery efficiency also provided data to calculate detection limits; these ranged from <2 to <215 oocysts per 10 liters. Recoveries of oocysts ranged from 2.0 to 61% for viable oocysts and from 3.0 to 59% for modified oocysts. The recoveries between the two seeding procedures were highly correlated (r = 0.802) and were not significantly different. Recoveries by using modified oocysts, therefore, were comparable to recoveries by using conventional seeding procedures.