Report No FR/D0006



Dec 1991


A comprehensive investigation has been made of the fate of the human enteric bacterium, Escherichia coli, in the Clyde Estuary. The results are based on carefully controlled laboratory work, field work and mathematical modelling carried out using desk-top computers. The findings are of general applicability to estuaries in the United Kingdom and in regions of similar climate.

The laboratory studies examined the effect of temperature, salinity, suspended solids concentration, and nutrient concentration on the inactivation (in the dark) of E. coli as determined by the membrane filtration technique for bacterial estimation. It was found that the presence of more than 5 mg l-1 inorganic solid particles substantially reduced the inactivation rates (or "die-off" rates) for E. coli under conditions of high salinity (27 ppt) and temperature (20C). In general higher salinity was linked to higher inactivation rates, although in the presence of solids a salinity of 27 ppt was more lethal than one of 32 ppt. In the absence of solids, inactivation rates were lower at lower temperatures but the opposite was true (ie higher at lower temperatures) when solids were added. Little significant difference was observed in the effect on inactivation of inorganic versus sewage-derived (mainly organic) solids. It was found that inactivation rates of E. coli were significantly reduced at high concentrations of bacteria (50,000 per 100 ml) relative to lower concentrations (5,000 per 100 ml). The explanation for these results could be that the lethal effect of salt on the E. coli is reduced by the bacteria attaching themselves to solids or clustering together via processes which are more effective at higher temperatures than at lower temperatures. However, the addition of nutrients tended to reduce inactivation rates at high solids concentration (>>25 mg l-1) but increase them at lower concentrations (12.5 mg l-1).

A laboratory study was also made of the removal of bacteria by attachment to solids and subsequent sedimentation. At high salinities (27-32 ppt), between 20% and 30% of bacteria were removed, but below 20 ppt, less than 10% were removed.

The mathematical model was constructed for an 11 km stretch of the Clyde Estuary from the tidal weir in Glasgow to Rothesay Dock. A one-dimensional, cross-sectionally averaged scheme was used, solving the hydrodynamic equations using the Preissman implicit finite difference method. The concentrations of salt, bacteria, and sediments were solved using a Galerkin finite element method using the same grid. Inactivation rates for the model were taken from laboratory work and from published results for the effect of sunlight. An extensive programme of field-work was undertaken for spring and neap tides to calibrate and validate the computer models. Agreement between predicted and measured water levels and velocities was excellent. Agreement for concentrations was less good, but adequate for reasonably accurate predictions of bacterial concentrations to be made using the computer model.

Copies of the report are available from FWR, price 15.00, less 20% to FWR Members.