What Can Utilities Expect from New Lead Fifth-Liter Sampling Based on Historic First-Draw Data?

The United States Environmental Protection Agency recently released their most sweeping overhaul to the Lead and Copper Rule in three decades. One of the most significant changes is requiring a fifth-liter (L5) sample at homes with lead service lines (LSLs) rather than the original first-liter (L1) sample for a demonstration of compliance with water lead level (WLL) limits. We analyzed sequential sampling data from three large water systems and compliance data from Michigan utilities-which base compliance on the 90th percentile of the greater of L1 and L5 samples-to evaluate whether L5 WLLs better represent water in contact with LSLs and to explore regulatory impacts of including L5 samples in compliance monitoring. The sequential sampling data demonstrated that it is impossible to use a single sample volume within a sequential profile to universally capture the volume of water in an LSL. While L5 is not always a reliable indicator of water in contact with an LSL, Michigan compliance data showed that the L5 sample is more likely to be from an LSL and can identify utilities that benefit from an improved corrosion control treatment. Michigan compliance data indicate that it is likely that L5 sampling will result in more systems having a higher 90th percentile WLL and that a high proportion of the systems likely to exceed regulatory action levels based on L5 samples can be identified through a retrospective analysis of historic L1 data. The impact of the switch to L5 sampling on the effectiveness of corrosion control treatment over time has yet to be determined.

Analysis of building plumbing system flushing practices and communications.

Drinking water distribution system contamination incidents can prompt public agencies and drinking water utilities to issue do-not-drink and do-not-use advisories. After the contaminant is cleared from distribution mains, consumers are often directed to flush their plumbing. However, little validated guidance and few evaluated communications strategies are available on using flushing to decontaminate building water systems. Additionally, limited data support the effectiveness of current practices and recommendations. In this study, expert elicitation was used to assess existing flushing guidance and develop validated flushing guidance and communications for single-family residences. The resulting guidance recommends progressively opening all cold-water taps from the closest to point of entry to the furthest and allowing the water to run for at least 20 minutes. Hot-water taps should be opened progressively and run for at least 75 minutes. The guidance language and format conformed to grade-level and readability scores within recommended health communication ranges. The readability of eight other flushing guidance documents was also evaluated for contamination incidents from 2008-2015. Seven were written at a 10th-12th grade level, above the 6th-7th grade level recommended for health communications.

Seasonal Variations in Lead Release to Potable Water.

The influence of temperature on the solubility of representative lead solids present in drinking-water systems and the lead release to potable water was examined. Temperature had surprisingly little effect on the dissolution of cerrusite, hydrocerussite, chloropyromorphite, lead orthophosphate, and lead oxide solids; however, in the presence of natural organic matter, lead oxide dissolution was 36 times greater (36 versus 1277 ppb) at 20 °C compared to 4 °C due to accelerated reductive dissolution. The solubility of plumbonacrite was three times higher at 20 °C compared to 4 °C (260 versus 92 ppb). In full-scale pipe rigs using harvested lead service lines in Washington, DC and Providence, RI, dissolved lead release increased by as much as 2-3 times, and particulate lead increased 2-6 times in the summer versus winter. In four of the eight homes sampled in Providence, RI, dissolved lead levels were three times higher during the summer compared to the winter, and five homes had copper levels that were 2.5-15 times greater in the winter. These studies demonstrate a need to better understand how lead service line scales vary because patterns of release and temperature dependency sometimes vary markedly, even within the same distribution system.

Inherent variability in lead and copper collected during standardized sampling.

Variability in the concentration of lead and copper sampled at consumers' taps poses challenges to assessing consumer health threats and the effectiveness of corrosion control. To examine the minimum variability that is practically achievable, standardized rigs with three lead and copper containing plumbing materials (leaded brass, copper tube with lead solder, and a lead copper connection) were deployed at five utilities and sampled with regimented protocols. Variability represented by relative standard deviation (RSD) in lead release was high in all cases. The brass had the lowest variability in lead release (RSD = 31 %) followed by copper-solder (RSD = 49%) and lead-copper (RSD = 80%). This high inherent variability is due to semi-random detachment of particulate lead to water, and represents a modern reality of water lead problems that should be explicitly acknowledged and considered in all aspects of exposure, public education, and monitoring.

Distribution System Water Quality Affects Responses of Opportunistic Pathogen Gene Markers in Household Water Heaters.

Illustrative distribution system operation and management practices shaped the occurrence and persistence of Legionella spp., nontuberculous mycobacteria (NTM), Pseudomonas aeruginosa, and two amoebae host (Acanthamoeba spp., Vermamoeba vermiformis) gene markers in the effluent of standardized simulated household water heaters (SWHs). The interplay between disinfectant type (chlorine or chloramine), water age (2.3-5.7 days) and materials (polyvinyl chloride (PVC), cement or iron) in upstream simulated distribution systems (SDSs) profoundly influenced levels of pathogen gene markers in corresponding SWH bulk waters. For example, Legionella spp. were 3-4 log higher in SWHs receiving water from chloraminated vs chlorinated SDSs, because of disinfectant decay from nitrification. By contrast, SWHs fed with chlorinated PVC SDS water not only harbored the lowest levels of all pathogen markers, but effluent from the chlorinated SWHs were even lower than influent levels in several instances (e.g., 2 log less Legionella spp. and NTM for PVC and 3-5 log less P. aeruginosa for cement). However, pathogen gene marker influent levels correlated positively to effluent levels in the SWHs (P < 0.05). Likewise, microbial community structures were similar between SWHs and the corresponding SDS feed waters. This study highlights the importance and challenges of distribution system management/operation to help control opportunistic pathogens.

Distribution system water age can create premise plumbing corrosion hotspots.

Cumulative changes in chemical and biological properties associated with higher "water age" in distribution systems may impact water corrosivity and regulatory compliance with lead and copper action levels. The purpose of this study was to examine the effects of water age and chemistry on corrosivity of various downstream premise plumbing pipe materials and configurations using a combination of controlled laboratory studies and a field survey. Examination of lead pipe, copper pipe with lead solder, and leaded brass materials in a replicated lab rig simulating premise plumbing stagnation events indicated that lead or copper release could increase as much as ∼440 % or decrease as much as 98 % relative to water treatment plant effluent. In field studies at five utilities, trends in lead and copper release were highly dependent on circumstance; for example, lead release increased with water age in 13 % of cases and decreased with water age in 33 % of conditions tested. Levels of copper in the distribution system were up to 50 % lower and as much as 30 % higher relative to levels at the treatment plant. In many cases, high-risks of elevated lead and copper did not co-occur, demonstrating that these contaminants will have to be sampled separately to identify "worst case" conditions for human exposure and monitoring.

Profile sampling to characterize particulate lead risks in potable water.

Traditional lead (Pb) profiling, or collecting sequential liters of water that flow from a consumer tap after a stagnation event, has recently received widespread use in understanding sources of Pb in drinking water and risks to consumer health, but has limitations in quantifying particulate Pb risks. A new profiling protocol was developed in which a series of traditional profiles are collected from the same tap at escalating flow rates. The results revealed marked differences in risks of Pb exposure from one consumer home to another as a function of flow rate, with homes grouped into four risk categories with differing flushing requirements and public education to protect consumers. On average, Pb concentrations detected in water at high flow without stagnation were at least three to four times higher than in first draw samples collected at low flow with stagnation, demonstrating a new "worst case" lead release scenario, contrary to the original regulatory assumption that stagnant, first draw samples contain the highest lead concentrations. Testing also revealed that in some cases water samples with visible particulates had much higher Pb than samples without visible particulates, and tests of different sample handling protocols confirmed that some EPA-allowed methods would not quantify as much as 99.9% of the Pb actually present (avg. 27% of Pb not quantified).

Effect of disinfectant, water age, and pipe materials on bacterial and eukaryotic community structure in drinking water biofilm.

Availability of safe, pathogen-free drinking water is vital to public health; however, it is impossible to deliver sterile drinking water to consumers. Recent microbiome research is bringing new understanding to the true extent and diversity of microbes that inhabit water distribution systems. The purpose of this study was to determine how water chemistry in main distribution lines shape the microbiome in drinking water biofilms and to explore potential associations between opportunistic pathogens and indigenous drinking water microbes. Effects of disinfectant (chloramines, chlorine), water age (2.3 days, 5.7 days), and pipe material (cement, iron, PVC) were compared in parallel triplicate simulated water distribution systems. Pyrosequencing was employed to characterize bacteria and terminal restriction fragment polymorphism was used to profile both bacteria and eukaryotes inhabiting pipe biofilms. Disinfectant and water age were both observed to be strong factors in shaping bacterial and eukaryotic community structures. Pipe material only influenced the bacterial community structure (ANOSIM test, P < 0.05). Interactive effects of disinfectant, pipe material, and water age on both bacteria and eukaryotes were noted. Disinfectant concentration had the strongest effect on bacteria, while dissolved oxygen appeared to be a major driver for eukaryotes (BEST test). Several correlations of similarity metrics among populations of bacteria, eukaryotes, and opportunistic pathogens, as well as one significant association between mycobacterial and proteobacterial operational taxonomic units, provides insight into means by which manipulating the microbiome may lead to new avenues for limiting the growth of opportunistic pathogens (e.g., Legionella) or other nuisance organisms (e.g., nitrifiers).

Effect of disinfectant, water age, and pipe material on occurrence and persistence of Legionella, mycobacteria, Pseudomonas aeruginosa, and two amoebas.

Opportunistic pathogens represent a unique challenge because they establish and grow within drinking water systems, yet the factors stimulating their proliferation are largely unknown. The purpose of this study was to examine the influence of pipe materials, disinfectant type, and water age on occurrence and persistence of three opportunistic pathogens (Legionella pneumophila, Mycobacterium avium, and Pseudomonas aeruginosa), broader genera (Legionella and mycobacteria), and two amoeba hosts (Acanthamoeba spp. and Hartmanella vermiformis). Triplicate simulated distribution systems (SDSs) compared iron, cement, and PVC pipe materials fed either chlorinated or chloraminated tap water and were sampled at water ages ranging from 1 day to 5.7 days. Quantitative polymerase chain reaction quantified gene copies of target microorganisms in both biofilm and bulk water. Legionella, mycobacteria, P. aeruginosa, and both amoebas naturally colonized the six SDSs, but L. pneumophila and M. avium were not detected. Disinfectant type and dose was observed to have the strongest influence on the microbiota. Disinfectant decay was noted with water age, particularly in chloraminated SDSs (due to nitrification), generally resulting in increased microbial detection frequencies and densities with water age. The influence of pipe material became apparent at water ages corresponding to low disinfectant residual. Each target microbe appeared to display a distinct response to disinfectant type, pipe materials, water age, and their interactions. Differences between the first and the second samplings (e.g., appearance of Legionella, reduction in P. aeruginosa and Acanthamoeba) suggest a temporally dynamic drinking water microbial community.

Effectiveness of Prevailing Flush Guidelines to Prevent Exposure to Lead in Tap Water.

Flushing tap water is promoted as a low cost approach to reducing water lead exposures. This study evaluated lead reduction when prevailing flush guidelines (30 s⁻2 min) are implemented in a city compliant with lead-associated water regulations (New Orleans, LA, USA). Water samples (n = 1497) collected from a convenience sample of 376 residential sites (2015⁻2017) were analyzed for lead. Samples were collected at (1) first draw (n = 375) and after incremental flushes of (2) 30⁻45 s (n = 375); (3) 2.5⁻3 min (n = 373), and (4) 5.5⁻6 min (n = 218). There was a small but significant increase in water lead after the 30 s flush (vs. first draw lead). There was no significant lead reduction until the 6 min flush (p < 0.05); but of these samples, 52% still had detectable lead (≥1 ppb). Older homes (pre-1950) and low occupancy sites had significantly higher water lead (p < 0.05). Each sample type had health-based standard exceedances in over 50% of sites sampled (max: 58 ppb). While flushing may be an effective short-term approach to remediate high lead, prevailing flush recommendations are an inconsistently effective exposure prevention measure that may inadvertently increase exposures. Public health messages should be modified to ensure appropriate application of flushing, while acknowledging its short-comings and practical limitations.

Redox gradients in distribution systems influence water quality, corrosion, and microbial ecology.

Simulated distribution systems (SDSs) defined the interplay between disinfectant type (free chlorine and chloramines), water age (1-10.2 days), and pipe material (PVC, iron and cement surfaces) on water chemistry, redox zones and infrastructure degradation. Redox gradients developed as a function of water age and pipe material affected the quality of water consumers would receive. Free chlorine was most stable in the presence of PVC while chloramine was most stable in the presence of cement. At a 3.6 day water age the residual in the chlorinated PVC SDS was more than 3.5 times higher than in the chlorinated iron or cement systems. In contrast, the residual in the chloraminated cement SDS was more than 10 times greater than in the chloraminated iron or PVC systems. Near the point of entry to the SDSs where disinfectant residuals were present, free chlorine tended to cause as much as 4 times more iron corrosion when compared to chloramines. Facultative denitrifying bacteria were ubiquitous, and caused complete loss of nitrogen at distal points in systems with iron, and these bacteria co-occurred with very severe pitting attack (1.6-1.9 mm/year) at high water age.