The microbiology of urban UK £oodwaters and a quantitative microbial risk assessment of £ooding and gastrointestinal illness L. Fewtrell 1 , D. Kay 2 , J. Watkins 3 , C. Davies 3 and C. Francis 3 1 Centre for Research into Environment and Health, Coppice House, Quakers Coppice, Crewe, Cheshire, UK 2 CCCR and CREH, Hydrology and River Basin Dynamics Research Group, IGES, Aberystwyth University, Aberystwyth, UK 3 CREH Analytical Ltd, Hoyland House, Horsforth, Leeds, UK Correspondence Lorna Fewtrell, Centre for Research into Environment and Health, Coppice House, Quakers Coppice, Crewe, Cheshire CW1 6FA, UK Tel: 144 12 70 25 05 83 Fax: 144 12 70 58 97 61 Email: lorna@creh.demon.co.uk DOI:10.1111/j.1753-318X.2011.01092.x Key words Flooding; health; impact; quality; risk assessment. Abstract A quantitative microbial risk assessment of flooding and gastrointestinal illness was carried out using a combination of floodwater quality data from the existing literature on pathogen concentrations in the various floodwater components, the use of microcosms to determine pathogen die-off rates in floodwaters and opportunistic sampling of actual UK floodwaters. Two flood-related scenarios were examined in a hypothetical population to provide an assessment of the likely cases of gastrointestinal illness resulting from contact with floodwater. The results of the study suggest that significant numbers of people are at risk of illness (especially from viral gastroenteritis) during the clean-up process rather than during the inundation and withdrawal phases. Additional sampling during flood events of water and sediment would strengthen the empirical policy evidence base required for estimating the severity and range of health impacts using the risk assessment methodology developed in this paper. Introduction Urban flooding may result in a number of human health impacts (reviewed by Ohl and Tapsell, 2000; Hajat et al., 2003; Tapsell and Tunstall, 2003) both during the immediate flood event (drowning) and in the long term (anxiety and mental health problems). In addition, exposure to microbial pathogenic contaminants present in floodwater may cause infection following contact with floodwater. The media often show a great deal of interest in the ‘quality’ of floodwaters and sediments and speculate on the likely impact of illness within the affected communities (e.g. Anon, 2007). However, there is relatively little empirical evidence describing the pathogen concentrations to which people are exposed during flooding. A number of component flood flows may contribute to the microbial quality of urban floodwater, including:  foul flow (including sewage and combined sewer over- flows);  urban nonfoul pollution (including roof run-off); and  rivers (which may include a rural diffuse pollution component). Some limited and opportunistic studies have identified microbial contaminants, including bacterial pathogens, in- testinal parasites and enteric viruses, in the various compo- nents of floodwater (Table 1). The pathogens identified in Table 1 largely reflect those microorganisms for which routine tests are available and the absence of an entry for a specific pathogen may mean that the pathogen has not been looked for in that component. While it is acknowledged that the microbiological quality of floodwaters will be event-specific, a three-pronged ap- proach was adopted to determine the pathogen concentra- tions needed to inform the quantitative microbial risk assessment (QMRA) for gastrointestinal illness associated with flooding in the United Kingdom. This involved carry- ing out a review of existing data on pathogen concentrations in floodwater components (Fewtrell et al., 2008; Fewtrell et al., 2010), the use of microcosms to determine microbial die-off rates in floodwaters and opportunistic sampling of actual floodwaters. Methodology Simulated floodwater microcosms In order to investigate pathogen concentrations and die-off during flood conditions, samples of untreated wastewater or wastewater mixed with river water were allowed to stand for intervals of up to 21 days at ambient indoor temperature and in semidarkness. Containers were removed at predeter- mined intervals and the supernatant (i.e. overlying liquid) and sediment were analysed for indicator bacteria (faecal coliforms, Escherichia coli and enterococci), pathogenic J Flood Risk Management 4 (2011) 77–87 c  2011 The Authors Journal of Flood Risk Management c  2011 The Chartered Institution of Water and Environmental Management