Changes in Escherichia coli to enteric protozoa ratios in rivers: Implications for risk-based assessment of drinking water treatment requirements
Graphical abstract
Introduction
Concentrations of reference enteric protozoa (Cryptosporidium, Giardia) in source water must be adequately estimated to define health-based protozoa treatment requirements for drinking water safety (WHO 2016). However, data are not always available because of the difficulty and cost associated with analyzing waterborne protozoa in environmental samples. In these situations, fecal indicator bacteria (FIB) such as Escherichia coli (E. coli) are commonly used as indicators of pathogen occurrence. Although FIB monitoring data sets can provide important information on fluctuations of fecal contamination in source water, it is essential to keep in mind that this indicator has limitations for predicting concentrations of enteric protozoa (Wu et al. 2011, Health Canada 2017). Meteorological and environmental factors can have different effects on the fate and transport of indicators and pathogens in water. Moreover, indicators can originate from other sources than pathogens.
Monitoring of protozoa in raw water from drinking water treatment plants (DWTPs) is recommended in Canada (Health Canada 2017). Still, protozoa monitoring is not mandatory in most Canadian provinces (Government of Manitoba 2007, Gouvernement du Québec 2012). In the United States and Alberta, Canada, protozoa monitoring is mandatory for large community water supplies and small community water supplies when E. coli concentrations are low (USEPA 2010, Government of Alberta 2012). Most of these regulations rely on the assumption that drinking water sources exposed to high E. coli concentrations have a greater probability of protozoa occurrence, independently of their concentration (Payment and Locas 2011). However, a quantitative relationship between concentrations of E. coli and protozoa needs to be established to define health-based minimum treatment requirements for protozoa using quantitative microbial risk assessment (QMRA).
To determine whether E. coli data can be used to support the implementation of health-based treatment requirements, E. coli to protozoa ratios can be evaluated over a given period at multiple DWTPs supplied by different types of drinking water sources. A meta-analysis of E. coli/Cryptosporidium ratios in primary sources of fecal contamination suggested that E. coli is generally a good indicator for predicting Cryptosporidium occurrence for urban pollution sources (raw and treated wastewater) but not for agricultural runoff (Lalancette et al. 2014). In this study, E. coli/Cryptosporidium ratios were also evaluated at drinking water intakes from 13 DWTPs in Quebec, Canada, using data from Payment et al. (2000). As estimated in their meta-analysis, ratios at drinking water intakes were lower for sources influenced by agricultural runoff than those influenced by municipal sewage. The present study was designed to validate these findings using recent data collected at 27 surface DWTPs supplied by rivers dominated by urban, agricultural, or wildlife sources of fecal pollution in Quebec, Canada. The mathematical model of Lalancette et al. (2014) was also extended to evaluate the uncertainty associated with arithmetic mean E. coli/Cryptosporidium and E. coli/Giardia ratios.
The objectives of this study were (i) to derive the arithmetic mean ratio of two correlated lognormal distributions and use this model to characterize E. coli/Cryptosporidium and E. coli/Giardia ratios in source water for 27 DWTPs supplied by rivers; (ii) to investigate the influence of river types and seasons on the variation in the magnitude of the mean ratios; and (iii) to evaluate whether E. coli trigger levels provide valuable information for defining health-based treatment requirements for pathogenic protozoa at these DWTPs.
Section snippets
Classification of sites
Source water supplies were anonymized and classified according to their annual mean flow rate (Table 1). Group A, Group B, and Group C represent rivers with mean flow rates of less than 20 m3 s−1, between 20 and 100 m3 s−1, and larger than 100 m3 s−1, respectively. The main land cover type for each catchment was determined visually with Google Earth.
Sampling and microbial detection methods
Paired samples were collected monthly over approximately two consecutive years between 2011 and 2020 at each of the 27 DWTPs. For the detection of
Poisson counts in mixture distributions
A pragmatic way to account for non-detects is to assume that each observed count is Poisson distributed and that the unknown microbial concentration is described by a mixture distribution (Haas et al. 1999). Within a mixed Poisson modeling framework, the probability of finding organisms in a homogenous sample of volume collected from a suspension of mean concentration is given by a Poisson distribution with probability mass function:
Overdispersion relative to the
Results
Site-specific mean E. coli/protozoa ratios were typically 1.0 to 2.0-log lower at DWTPs supplied by small and medium rivers (Group A, Group B) in comparison with DWTPs supplied by large rivers (Group C) (Fig. 1). Ratios varied by approximately 4.0-log for both Cryptosporidium (103–107) and Giardia (100–104). Mean E. coli/Cryptosporidium ratios were generally 2.0 to 3.0-log higher than mean E. coli/Giardia ratios because Cryptosporidium concentrations were lower and more uncertain than Giardia
Discussion
Despite inherent uncertainties associated with quantifying reference pathogen concentrations from indicator data, fecal indicator bacteria trigger levels are still commonly used to define treatment requirements in drinking water safety regulations. The values and limitations of indicators for predicting reference pathogen concentrations need to be rigorously assessed to support the development and the revision of risk-based regulations. This study was undertaken to determine whether E. coli is
Conclusions
In this study, relationships between E. coli, Cryptosporidium and Giardia were quantified using paired E. coli and protozoa data collected in source water from 27 drinking water treatment plants (DWTPs). To evaluate the potential of E. coli for defining minimum treatment requirements, risk-based critical mean protozoa concentrations in source water were determined with a quantitative microbial risk assessment (QMRA) model developed using the methods and assumptions recommended by the World
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This work was funded by the NSERC Industrial Chair on Drinking Water, the Canadian Research Chair on Source Water Protection, NSERC Collaborative Research and Development Grant Project (CRDPJ-505651–16) and the Canada Foundation for Innovation. A part of the outcomes presented in this paper was based on research financed by the Dutch-Flemish Joint Research Programme for the Water Companies. We thank the technical staff of the biology and microbiology division at centre d'expertise en analyze
References (50)
- et al.
Tracking the contribution of multiple raw and treated wastewater discharges at an urban drinking water supply using near real-time monitoring of β-d-glucuronidase activity
Water Res.
(2019) - et al.
Respective contributions of point and non-point sources of E. coli and enterococci in a large urbanized watershed (the Seine river, France)
J. Environ. Manage.
(2007) - et al.
Quantification of fecal coliform inputs to aquatic systems through soil leaching
Water Res.
(2004) - et al.
Disease burden of selected gastrointestinal pathogens in Australia, 2010
Int. J. Infect. Dis.
(2014) - et al.
Changes in E. coli to Cryptosporidium ratio from various fecal pollution sources and drinking water intakes
Water Res.
(2014) - et al.
Towards a more accurate quantitative assessment of seasonal Cryptosporidium infection risks in surface waters using species and genotype information
Water Res.
(2016) - et al.
Correlated gamma variables in the analysis of microbial densities in water
Adv. Water Res.
(2005) - et al.
The effects of combined sewer overflow events on riverine sources of drinking water
Water Res.
(2016) - et al.
Understanding human infectious Cryptosporidium risk in drinking water supply catchments
Water Res.
(2018) - et al.
Can routine monitoring of E. coli fully account for peak event concentrations at drinking water intakes in agricultural and urban rivers?
Water Res.
(2020)
Characterization of drinking water treatment for virus risk assessment
Water Res.
Effects of fulvic acid and fulvic ions on Escherichia coli survival in river under repeated freeze-thaw cycles
Environ. Pollut.
Impact of Freeze–Thaw cycles on die-off of E. coli and intestinal enterococci in deer and dairy faeces: implications for landscape contamination of watercourses
Int. J. Environ. Res. Public Health
Lognormal distributions
Inference from simulations and monitoring convergence
Handbook Markov Chain Monte Carlo
Règlement et autres actes. Règlement modifiant le Règlement sur la qualité de l'eau potable (Loi sur la qualité de l'environnement)
Part 1: Standards for municipal waterworks of a total of 5 parts. Page 104 Standard and guidelines for municipal waterworks, wastewater and storm drainage systems
The drinking water safety Act
Drinking water quality standards regulation (C.C.S.M. c D101)
Quantitative microbial risk assessment
Quantifying public health risk in the WHO guidelines for drinking-water quality: a burden of disease approach
Markov chain Monte Carlo in practice: a roundtable discussion
The American Statistician
Effects of freeze–thaw events on the viability of Cryptosporidium parvum oocysts in soil
J. Parasitol.
Doing Bayesian data analysis: A tutorial with R, JAGS, and Stan
Cited by (3)
Surrogates of foodborne and waterborne protozoan parasites: A review
2023, Food and Waterborne ParasitologyAssociation between water source and chronic gastrointestinal diseases in Chinese: A cross-sectional and longitudinal study
2022, Frontiers in Public Health