Comparative reductions of norovirus, echovirus, adenovirus, Campylobacter jejuni and process indicator organisms during water filtration in alluvial sand.

Sand filtration is a cost-effective means of reducing microbial pathogens in drinking-water treatment. Our understanding of pathogen removal by sand filtration relies largely on studies of process microbial indicators, and comparative data from pathogens are sparse. In this study, we examined the reductions of norovirus, echovirus, adenovirus, bacteriophage MS2 and PRD1, Campylobacter jejuni, and Escherichia coli during water filtration through alluvial sand. Duplicate experiments were conducted using 2 sand columns (50 cm long, 10 cm diameter) and municipal tap water sourced from chlorine-free untreated groundwater (pH 8.0, 1.47 mM) at filtration rates of 1.1-1.3 m/day. The results were analysed using colloid filtration theory and the HYDRUS-1D 2-site attachment-detachment model. The average log reduction values (LRVs) of the normalised dimensionless peak concentrations (C/C) over 0.5 m were: MS2: 0.28; E. coli: 0.76; C. jejuni: 0.78; PRD1: 2.00; echovirus: 2.20; norovirus: 2.35; and adenovirus: 2.79. The relative reductions largely corresponded to the organisms' isoelectric points rather than their particle sizes or hydrophobicities. MS2 underestimated virus reductions by 1.7-2.5 log, and the LRVs, mass recoveries relative to bromide, collision efficiencies, and attachment and detachment rates differed mostly by ~1 order of magnitude. Conversely, PRD1 reductions were comparable with those of all 3 viruses tested, and its parameter values were mostly within the same orders of magnitude. E. coli seemed an adequate process indicator for C. jejuni with similar reductions. Comparative data describing pathogen and indicator reductions in alluvial sand have important implications for sand filter design, risk assessments of drinking-water supplies from riverbank filtration and the determination of safe setback distances for drinking-water supply wells.