Riverbank filtration has been shown to be effective for removing viable Cryptosporidium parvum oocysts. Drinking water systems that employ riverbank filtration may receive additional treatment credits beyond that which they can obtain using traditional engineering approaches. In order to develop guidance for removal effectiveness, screening level predictive modeling by colloid filtration theory combined with advection and dispersion modeling is potentially useful. Currently, only few studies have measured basic effective colloid filtration parameters for Cryptosporidium oocysts with naturally occurring riverbank sediments. In the focus of this study we conducted flow column experiments in triplicate and measured effective attachment rate coefficients for sandy river sediments of the Southern Great Plains which are low in organic matter. We found that for sediment sampled from these high-energy rivers there was no apparent dependency of C. parvum removal with carbon content, bacterial colony forming units, or with gross texture properties of the sands. The differences in particle size distribution for the sediments suggested that straining did not play a role in removal efficiency. First-order colloid attachment rate coefficients followed lognormal distribution functions. The coefficients also appeared to be unrelated to the differences in particle size distributions of the sediments, bacterial counts, or levels of total carbon or total organic carbon. Using Monte Carlo analyses, the lowest observed 5th percentile was 8.0 x 10(-6) min(-1) and the highest observed 95th percentile was 1.6 x 10(-3). Total log(10) removals ranged from 23 to 200 m(-1). These results have application for screening level colloid filtration modeling of riverbank filtration in these systems.
Published by Elsevier Ltd.