Evaluating the impact of underwater skimming on slow sand filter performance and operation

Conventional cleaning of slow sand filters (SSFs) requires the beds to be drained before a layer of media and the Schmutzdecke are removed, called 'dry skimming' (DS), which can result in significant downtime. An alternative is proposed whereby the filter is skimmed whilst still submerged, called 'underwater skimming' (UWS). Previous attempts to avoid draining the bed have led to concerns about the risks of UWS in terms of headloss development, particle penetration, and microbial water quality. In this study, pilot scale SSFs, cleaned by either DS or UWS, were operated concurrently, to compare and assess the potential risks of UWS in terms of filtrate quality, microbial removal, and recovery following skimming. While all filters exhibited effective turbidity removal (0.04-0.9 NTU turbidity from outlet), UWS filters had improved performance immediately after cleaning compared to DS in terms of recovery of filtrate microbial water quality. Specifically, total coliforms in the UWS filter outlets, in the first seven days post-skimming, ranged from 1 to 109 most probable number (MPN)/mL, compared to 1 to 1414 MPN/mL for DS filters. Both methods yielded satisfactory headloss recovery, indicative of limited particle penetration at depth, and effective cleaning. Exploring different sweetening flow rates during UWS revealed no observable differences in headloss, turbidity reduction, or microbial quality between the flow rates tested. Schmutzdecke microbial community was similar irrespective of cleaning method and was governed instead by seasonal changes and the ripening process. The improved microbial removal afforded by UWS provides a means of significantly improving water productivity and enabling options for much better utilisation of SSFs.