MODELLING FAECAL COLIFORM CONCENTRATIONS IN STREAMS
Report No DWI0668

Jan 1995

II Executive Summary

This is the final report to DoE on contract PECD 7/7/385 "Modelling E.coli in streams" awarded to the Institute of Hydrology 1991-1994. The stated objectives of the project were "... to ascertain the key processes by which faecal coliforms are transported through river catchments; to suggest land-use impacts on stream faecal coliform concentrations; and to develop an integrated predictive model of faecal coliform concentrations from point and non-point sources". The work can be integrated with research in areas seeking to describe bacterial water quality and to assess the risks associated with exposure to waters containing faecal contamination.

The above objectives have been met by a combined approach including a literature survey, field experimentation, model development and validation using both field and existing databases, and the examination of national GIS databases in assessing land-use impacts.

1. Examination of the key influences on the survival of faecal coliforms in streams and rivers demonstrated that these include; light and turbidity, temperature and pH. Sunlight is probably the most important single driving variable with regard to faecal coliform die-off in streams and rivers. 90% reduction of an initial population in a few hours might be expected in bright sunlight, in darkness the organisms may persist for many days. The effect of solar radiation is reduced in turbid waters where light penetration is reduced and the organisms are shielded by an envelope of small particles. Temperature and pH play a lesser role in determining faecal coliform survival. In sewage contaminated or oxygen stressed waters faecal coliform survival is extended.

   Examination of the key processes of faecal coliform transport within catchments demonstrated how the significance of different processes and sources of faecal contamination change with location. In headwater areas the supply of organisms is dominated by non-point sources; organisms are transported from the catchment surface by a combination of surface run-off and non-matrix throughflow in the subsurface zone during rainfall events. Further downstream the emphasis changes, point sources and channel storage interactions becoming more significant to the supply of contaminative organisms.

2. Previous models for faecal coliform dynamics used a range of approaches. Multivariate statistical approaches relate the bacterial concentration to a number of driving variables using simple statistical relationships. Simple deterministic first order decay functions have been used for describing the exponential die-off of a bacterial population and in application to rivers have been combined with equations to describe fluid mixing processes and flow hydraulics. These models all lacked the necessary structure to describe the process of bacterial transport in rivers adequately. Only the model of Jenkins (1984) sought to describe the transfer of organisms to and from storage within the channel.

   The new model presented in Section 6 of this report uses a mass balance structure similar to that adopted by Jenkins and can successfully reproduce the faecal coliform time-series produced during the field experiments described in Section 5.

   The model structure and operation incorporates the following assumptions: 1. The channel-store is distributed across the entire channel. 2. The regions of storage respond sequentially to rises in flow. 3. Any given rise in flow will produce entrainment of organisms from the channel. 4. At any quasi-steady flow the active supply area of organisms will become depleted. 5. No further entrainment can occur once the flow recession commences. Further higher flows will still release organisms from storage.

   The model incorporates terms for the effects of environmental influences, derived from data in the literature, describing the effect of sunlight and turbidity. temperature and pH on faecal coliform survival in the water column.

   The model was successfully applied to a reach of the River Exe in Devon for the years 1990 and 1991. The model was seen to operate well for extended periods of data, the numbers of organisms in the channel store remained stable and were, in effect, self regulating. Seasonal effects were modelled with a simple cosine function accounting for die-off changes resulting from solar radiation and temperature, overcoming the need for data for these variables and reducing the number of parameters needed to calibrate the model. No previous model has given a satisfactory description of faecal coliform river dynamics; the model applied here not only gave a good fit to the observed data it also has scope for application to other water quality determinants.

3. Analysis of faecal coliform concentrations in 13 upland Welsh catchments and data from ADAS land use maps and derived from surveys of stocking practices and fertiliser use showed that catchments with higher proportions of improved agricultural land, with higher fertiliser use and livestock densities. produce higher geometric mean faecal coliform concentrations than forested catchments. This finding reflects the higher loadings of organisms from livestock in agricultural catchments.

   Examination of a further 12 catchments in England, Scotland and Wales representing a broad range of size, land-use and faecal contamination demonstrated behaviour consistent with the Welsh study. It was found that agricultural land classes and groupings of classes perhaps relating to the lowland nature of the catchment produce more faecal coliforms than more upland catchments with non-agricultural landuses. The results indicate the importance of near channel areas as delivery fronts from faecal coliform supply areas within the catchment.

   Further studies should examine the relationships between faecal coliform concentrations and the more recent ITE land use classification system which differentiates between grasslands used for pasture or rough grazing etc. Combined with the analysis of a greater number of sites and study of travel-time effects, this would represent valuable enhancement of the results already presented.

   The current version of the model is capable of simulating the changes in faecal coliform concentrations in a river network at both seasonal and storm event time scales. The scope of the model includes assisting in the assessment of changes in effluent discharges or land-use on, for example, the health risk posed to recreators by a given river reach; assessing loadings of faecal contaminants to the marine system and hence the impact on compliance of local bathing waters; and to assist drinking water abstractors predict the timing, duration and magnitude of events of peak bacterial concentration in order to prevent the intake of large loadings of faecal contaminants.

Copies of this report may be available as an Acrobat pdf download under the 'Find Completed Research' heading on the DWI website.