Submitted to: Water Research
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/23/2013
Publication Date: 2/1/2013
Publication URL: http://handle.nal.usda.gov/10113/56595
Citation: Blaustein, R., Pachepsky, Y.A., Hill, R., Shelton, D.R., Whelan, G. 2013. E. coli survival in waters: temperature dependence. Water Research. 47:569-578. Interpretive Summary: E. coli are a common indicator of fecal contamination in surface waters. In order to properly interpret E. coli concentration data, it is important to know how rapidly they die in water. Die-off rates of E. coli in water are known to depend on temperature. We assembled the largest known database of experimental data on E. coli die-off in natural waters, and identified typical temperature dependencies in different types of waters using the Q10 model; the Q10 model is commonly used to express the dependencies of biological processes on temperature. We also established die-off patterns and the ranges of die-off parameters encountered in different types of waters. Results of this work will be useful for environmental engineering and decision support work to control the microbial water quality in that they provide information on typical values and variability ranges for temperature –dependent E. coli die-off.
Technical Abstract: Knowing the survival rates of water-borne Escherichia coli is important for evaluating microbial contamination and in making appropriate management decisions. E. coli survival rates are dependent on temperature; this dependency is routinely expressed using an analog of the Q10 model. This suggestion was made 34 years ago based on 20 survival curves taken from published literature, and has not been revisited since then. The objective of this study was to re-evaluate the accuracy of the Q10 equation utilizing data accumulated since 1978. We assembled a database consisting of 450 E. coli survival datasets from 70 peer-reviewed papers. We then focused on the 170 curves taken from experiments that were performed in the laboratory under dark conditions to exclude the effects of sunlight and other field factors that would cause additional variability in results. All datasets were tabulated dependencies “log concentration vs. time.” There were three major patterns of inactivation. About half of the datasets had a section of fast log-linear inactivation followed by a section of slow log-linear inactivation; about a quarter of the datasets had a lag period followed by log-linear inactivation; and the remaining quarter were approximately linear throughout. First-order inactivation rate constants were calculated from the linear sections of all survival curves and the data grouped by water sources, including waters of agricultural origin, pristine water sources, groundwater and wells, lakes and reservoirs, rivers and streams, estuaries and sea water, and wastewater. Dependencies of E. coli inactivation rates on temperature varied among the different water sources. There was a significant difference in inactivation rate values at the reference temperature between rivers and agricultural waters, rivers and seawater, rivers and lakes, and wastewater and lakes. At specific sites, the Q10 equation was more accurate in rivers and coastal waters than in lakes. The value of the Q10 coefficient appeared to be site-specific. Results of this work indicate possible sources of uncertainty to be accounted for in watershed-scale microbial water quality modeling.