Kim, N.D.; Stewart, C.; Cousins, W.J. 2016 Evolution in water quality over one year in emergency rainwater collection tanks installed in the Wellington region, New Zealand. Lower Hutt, N.Z.: GNS Science. GNS Science report 2015/59 vi, 98 p.
Abstract: Six 200-litre emergency water tanks were installed at properties across the Wellington region. Five of these tanks were allowed to fill naturally with rainwater and the sixth was filled with town supply tap water as a control. Sampling from these tanks was carried out fortnightly for one year, and samples analysed for E. coli, pH, conductivity, a range of major and trace elements, and organic compounds. The purpose of this study was to provide an evidence base on microbial and chemical contamination in tank water to assess health hazards for householders in an emergency situation. In the rain-fed tanks, the overall rate of E.coli detections in all samples was 17.7%, reducing to 12.3% if marginal detections were excluded. This is a low prevalence compared to other studies, both from New Zealand and overseas. The observation that tank-waters with the highest levels of Zn had no (0%) clear E. coli detections has led us to propose a zinc sterilisation hypothesis whereby the observed levels of zinc (in the mg/L range) may be causing biocidal effects on microbiota within the tanks. The median concentration of lead in all rain-fed tank samples was almost 40 times higher than in a background rainwater sample, and 20 times higher than in Wellington town supply tap water. Sixty-nine percent of these samples exceeded the maximum acceptable value (MAV) of 10 µg/L set by the New Zealand Drinking Water Standards 2005 (NZDWS), which is a high level of exceedance compared to other published studies. Two of the five tanks were particularly high in lead, exceeding the MAV in 100% and 96% of samples. These two tanks were fed by the oldest roof catchment systems (based on original painted roof cladding) in the study. However, the mechanism for generating high levels of lead within tanks may be less related to sources of lead (such as lead-head nails and flashing) and more related to within-tank processes. The older roof systems contributed less zinc to tanks, and in line with the zinc sterilisation hypothesis, may have become more readily acidified which may in turn have released adsorbed lead from tank sediment or tank walls. Lead cannot be removed from drinking water by boiling. Since the MAV for lead is designed with lifetime exposure in mind, results indicate unsuitability of such rain-fed tank water for routine drinking. However comparison to the drinking water MAV does not provide any index of potential risks of short-term consumption in an emergency context. This indicates a need for a more detailed health-risk assessment relating to risks of consuming water containing lead at the upper end of the observed range (40 µg/L) over a 20–50 day period when municipal water supply may be interrupted. As for lead, zinc concentrations are highly enriched in rain-fed tanks, with the median zinc level in all rain-fed tank samples enriched over a background rainfall sample by factor of just over 60. Zinc is typically a minor (µg/L level) constituent of most natural wat ers, but is a major constituent (mg/L level) of the rain-fed tank water and is the second-most abundant cation after sodium. Variations between tanks are readily explained with reference to roof cladding materials, with the highest levels found in systems with relatively new unpainted galvanised steel cladding. We propose that the high zinc concentrations observed in rain-fed tanks may be exerting a biocidal effect on microbiota within the tanks, leading to other observed features of our data set. (auth)