Abstract
A slow sand filter system was developed which was suitable both for experimental purposes and full scale community water supply. The performance of the system was enhanced by the incorporation of pretreatments such as gravel prefiltration, sub-sand prefiltration and direct protection of filters by the incorporation of synthetic fabric layers. The system was used to examine aspects of the biological and physico-chemical nature of Slow Sand Filtration with particular reference to the removal of Rotavirus and a variety of bacterial and viral indicators. The results of the developmental phase and experiments conducted with operational filters and in the laboratory confirmed the importance of biological mechanisms, in particular adsorption to biological surfaces present in the upper horizons of slow sand filters. It was observed that both the efficiency and pattern of removal of Rotavirus in Slow Sand Filtration were similar to those observed for faecal indicator bacteria, bacteriophage and turbidity ie colloidal clay and other particles. Thus it was concluded that negatively charged particulate colloidal entities appear to behave similarly despite differences in size and surface characteristics and that Rotavirus is no more or less likely to pass through slow sand filters than any other particle. Thus, a well operated slow sand filter may be expected to achieve a reduction in Rotavirus densities of 2 log10 orders of magnitude. The principal mechanisms of removal of Rotavirus in the upper horizons would appear to be transportation by diffusion and adsorption to biofilms and biomass. Microbial predation is not likely to play a dominant role in the removal of Rotavirus from the aqueous phase but may be important in inactivation. A detailed examination of the incidence, prevalence and seasonal occurrence of rotaviral and other diarrhoeas in less developed countries led to the observation that unlike some of the bacterial pathogens, transmission of Rotavirus does not have a strong association with water quality. Hence. The case for low level waterborne transmission and thereby maintenance of endemicity in the community remains entirely theoretical. Moreover, bearing in mind the likelihood that slow sand filters will be operated sub-optimally in many cases, a moderate reduction in Rotavirus densities eg 1 - 2 log10 units may not be significant in terms of diarrhoeal disease risk, particularly for children and especially in the rural sector. In view of these observations and other factors eg the low infective dose of the virus and the undoubted pre-eminence of person-to-person spread, it was concluded that as a single process, Slow Sand Filtration is probably not capable of significantly reducing the incidence of rotaviral diarrhoea in less developed countries.