Variation of an indicator of Escherichia coli persistence from surface waters of mixed-use watersheds, and relationship with environmental factors

¹ Agriculture and Agri-Food Canada, 1391 Sandford St, London, ON N5V 4T3, Canada
² Agriculture and Agri-Food Canada, 960 Carling Ave, Ottawa, ON K1A 0C6, Canada
³ Agriculture and Agri-Food Canada, 5403 1st Ave S, Lethbridge, AB T1J 4B1, Canada
⁴ Agriculture and Agri-Food Canada, 2560 boul. Hochelaga, Sainte-Foy, QC G1V 2J3, Canada
⁵ Institute for Resources, Environment and Sustainability, University of British Columbia, 429-2202 Main Mall, Vancouver, BC V6T 1Z4, Canada
⁶ Alberta Environment, 2938-11 St. NE, Calgary, AB T2E 7L7, Canada
⁷ Environment Canada, 867 Lakeshore Road, Burlington, ON L7R 4A6, Canada

^* Corresponding author: Ed.Topp@agr.gc.ca

Received: 18 August 2009
Accepted: 31 October 2010

Abstract

Escherichia coli is an indicator of fecal pollution used to mandate recreational and drinking water quality. Concentrations of culturable E. coli following contamination of surface water are determined by three factors: dilution; cell attachment to particulate material and settling or resuspension in the water column; and the net rate of change in viability. This study evaluated the variability in the latter parameter, and how predictive variation in death rate was of culturable population densities at the time of sampling. Water samples (N=232) with varying levels of E. coli contamination were collected from 46 discrete locations in four watersheds across Canada over a three-month period and enumerated for culturable E. coli by membrane filtration plate counting (T0_EC). Water samples were again enumerated following a laboratory 24 h holding period at 30°C in the dark, and the difference considered the death rate (Δ_EC). Relationships of T0_EC and Δ_EC with environmental and water chemistry factors were explored using step-wise multiple regression. The model predicting T0_EC indicated that stream order, total rainfall seven days in advance of sampling day, total phosphorus, and Δ_EC were the most significant contributors. The model predicting Δ_EC indicated that turbidity and NH₃+NH₄ were the most important contributors. A model suggests that the persistence factor is less important than dilution (i.e. stream order) in describing E. coli densities, followed by factors that influence the loading of E. coli into watersheds.