The University of Sheffield
Catchment Science Centre

Sewage risks to urban groundwater

B.N. Chisala and D.N. Lerner, 2008

Science Report SC030134, Environment Agency, Bristol. 36 pages. ISBN 978-1-84432-820-8

Urban groundwater is a largely underused resource and the potentially productive Triassic Sherwood Sandstone aquifer beneath the city of Nottingham in the UK is no exception (Lerner and Barrett, 1996; Davison et al., 2002). The perception that it is at risk of pollution from multiple industrial and residential sources and practices has, in the past, made it unattractive as a source of drinking water or indeed process water. However, in recent years, urban groundwater in the UK has been recognised as a potentially valuable resource because of the environmental problems associated with the rising urban water tables and heavy reliance upon rural aquifers. To achieve more sustainable use of our groundwater resource, a balance needs to be attained whereby pressures on rural groundwater are reduced and the continuing rise in urban groundwater levels is arrested. This in turn requires tools to be developed that can be used as aids in decision-making on matters pertaining to urban groundwater pollution risks.

Risk is defined as the expected outcome that is the sum of all the possible outcomes multiplied by their probability of occurrence. Risk analysis refers to methods that aim to develop a comprehensive understanding and awareness of the risk associated with a particular variable of interest (Hertz and Thomas, 1983). It is based on the source-pathway receptor concept. Lerner et al. (2000) state that a full risk analysis of groundwater combines the probabilities of (a) possible source term, that is types, quantities and frequencies of pollutant inputs, and (b) attenuation along the groundwater pathway, with the effects on receptors. The results of the risk analysis are presented in the form of probability distributions, weighing up the likelihood of alternative outcomes. Probability distributions are used in order to describe uncertainty and variability (Environment Agency, 2001a).

Significant risk of groundwater pollution can arise from high probability of occurrence combined with relatively mild effects, such as the widespread occurrence of chlorinated solvent pollution, or the converse. Sewer leakage is an example of the converse, with severe health consequences from microbiological pollution, but with low probabilities of occurrence. It is estimated that as many as 4 billion cases per year of water-borne diseases occur globally because of the consumption of polluted drinking water (WHO, 2000). In the UK, outbreaks of water-borne diseases associated with polluted groundwater sources have been documented (Lerner and Halliday, 1994; Bishop et al., 1998), but are rare, with only some 70 officially documented cases over the past 60 years or so (Blackwood et al., 2001). Nevertheless it is very possible that there are other cases which go unrecognised and unreported.

In contrast to our increasing knowledge about the prevalence of organic and inorganic pollutants in urban groundwater (Lerner and Cronin, 2004), we know little about the frequency and severity of sewer leakage into groundwater. In addition, we currently do not have tools to analyse the microbiological risks to urban groundwater from leaking sewers.

Lerner and Halliday (1994) carried out an early review of leakage from sewers to groundwater, work that was reaffirmed in a later study by Bishop et al. (1998). Both of these studies collected information on reported sewer leaks but did not quantify the risk. A recent exercise found widespread microbiological pollution in shallow groundwater in Nottingham (Barrett et al., 1999), and evidence of enteric bacteria and viruses tens of metres below the water table in both Nottingham and Birmingham (Cronin et al., 2003; Powell et al., 2003).

More recently BOS, which predicts the risk of pollution to groundwater in urban areas, was developed. BOS consists of (1) a Catchment Zone Probabilistic Model (CZPM) to find probabilistic borehole catchments from a groundwater flow model, (2) a land-use database, with information on all the industrial land parcels and their potential pollution for six dates in the 20th century, and (3) a probabilistic fate and transport model for the behaviour of organic pollutants in the unsaturated and saturated zones, linked together in a GIS with a single graphical user interface. BOS did not analyse microbiological risks. However, it provides an excellent framework that could be extended and used to assess microbiological risks to urban groundwater from leaking sewers. Detailed information on the development of BOS and its application in assessing the risk to urban groundwater from specific organic contaminants is available (Davison et al., 2002; Tait et al., 2004, 2007; Chisala et al., 2006).