Design and operational considerations in response to Legionella occurrence in Las Vegas Valley groundwater

Abstract

Legionella occurrence monitoring is not required by United States Environmental Protection Agency (USEPA) drinking water regulations, and few occurrence studies exist for Legionella in source water or distribution systems. Legionella occurrence was monitored in Las Vegas Valley (Las Vegas, Nevada, USA) drinking water sources, including non-treated surface water, seasonal groundwater (61 wells, before and after chlorination), finished water (after treatment at water treatment facilities), and chlorinated distribution system water (at 9 reservoirs and 75 sample locations throughout the network). Legionella pneumophila was detected at least once at each of the wells sampled before chlorination, with an overall positivity rate of 38% (343/908). During well start-up (time<2 hours; turbidity>3 NTU), L. pneumophila concentrations averaged 2,792±353 MPN/100 mL, with a median of 105 MPN/100 mL, and range of <1 to 90,490 MPN/100 mL across 61 seasonally operated (typically April-October) groundwater wells. After initial flushing (turbidity<3 NTU), the average concentration decreased by more than two orders of magnitude to 24±3 MPN/100 mL but ranged from <1 to >2,273 MPN/100 mL. This trend indicates that stagnation (up to 391 days) contributed to greater initial concentrations, and flushing alone was incapable of complete L. pneumophila elimination. L. pneumophila concentration was significantly, positively correlated with total aqueous adenosine triphosphate (ATP) (p<0.00001, r=0.41-0.71), turbidity (p<0.00001, r=0.27-0.51), orthophosphate (p=0.35-0.076, r=0.51-0.59), and pump depth (p=0.032, r=0.40). During a full-scale assessment of chlorination (Ct=0.7 to 10.5 mg-min/L; T=26.6-28.1°C), substantial reduction of Legionella spp. (up to 2.5 logs) was observed; although, detectable concentrations were still measured. Extrapolating from a Chick-Watson model (log inactivation=0.28*(Ct); R²=0.87) constructed from the full-scale chlorination results, 3- and 4-log inactivation in Las Vegas Valley groundwater would require 10.8 and 14.3 mg-min/L, respectively; at least 3-log inactivation was required to bring Legionella spp. to below detection at the studied well. Chlorine exposure (Ct=0.1 to 10.9 mg-min/L) at most wells discharging directly to the distribution system was insufficient to fully inactivate Legionella spp. After discussing these findings with the state regulatory agency, direct-to-distribution wells (38 of 61 wells) remained out of operation; the distribution system, wells, and reservoirs were monitored for Legionella and chlorine residual, and additional treatment scenarios were identified for further evaluation. Legionella was either not detected or was well controlled in surface water, finished effluent from the drinking water treatment plant, chlorinated reservoirs, and the chlorinated distribution system. This study emphasizes the importance of utility-driven, non-regulatory research in order to protect public health and also identifies the need for greater occurrence monitoring and guidance for Legionella in groundwater supplies.

Keywords: Chlorination; Disinfection; Distribution system; Flushing; Groundwater; Legionella; Monitoring; Occurrence; Pathogens; Utility; Well.