(Pb) Contamination of Self-Supply Groundwater ... - ACS Publications

Jan 21, 2015 - Department of Civil and Environmental Engineering, University of South Florida, Tampa, Florida 33620, United States. ‡. Department of...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/est

Lead (Pb) Contamination of Self-Supply Groundwater Systems in Coastal Madagascar and Predictions of Blood Lead Levels in Exposed Children D. Brad Akers,† Michael F. MacCarthy,†,‡ Jeffrey A. Cunningham,*,† Jonathan Annis,§ and James R. Mihelcic† †

Department of Civil and Environmental Engineering, University of South Florida, Tampa, Florida 33620, United States Department of Environmental Engineering, Mercer University, Macon, Georgia 31207, United States § USAID/WASHplus program, CARE International, Washington, D.C. 20009, United States ‡

S Supporting Information *

ABSTRACT: Thousands of households in coastal Madagascar rely on locally manufactured pitcher-pump systems to provide water for drinking, cooking, and household use. These pumps typically include components made from lead (Pb). In this study, concentrations of Pb in water were monitored at 18 household pitcher pumps in the city of Tamatave over three sampling campaigns. Concentrations of Pb frequently exceeded the World Health Organization’s provisional guideline for drinking water of 10 μg/L. Under firstdraw conditions (i.e., after a pump had been inactive for 1 h), 67% of samples analyzed were in excess of 10 μg/L Pb, with a median concentration of 13 μg/ L. However, flushing the pump systems before collecting water resulted in a statistically significant (p < 0.0001) decrease in Pb concentrations: 35% of samples collected after flushing exceeded 10 μg/L, with a median concentration of 9 μg/L. Based on measured Pb concentrations, a biokinetic model estimates that anywhere from 15% to 70% of children living in households with pitcher pumps may be at risk for elevated blood lead levels (>5 μg/dL). Measured Pb concentrations in water were not correlated at statistically significant levels with pump-system age, well depth, system manufacturer, or season of sample collection; only the contact time (i.e., flushed or first-draw condition) was observed to correlate significantly with Pb concentrations. In two of the 18 systems, Pb valve weights were replaced with iron, which decreased the observed Pb concentrations in the water by 57−89% in one pump and by 89−96% in the other. Both systems produced samples exclusively below 10 μg/L after substitution. Therefore, relatively straightforward operational changes on the part of the pump-system manufacturers and pump users might reduce Pb exposure, thereby helping to ensure the continued sustainability of pitcher pumps in Madagascar.

■

INTRODUCTION

population) within 28 000 households currently use pitcher pumps in Tamatave.1 The pitcher-pump system consists of a manual suction pump with two check valves, which work to lift water from a shallow well (see Figure 1). In Madagascar, some of the components of pitcher-pump systems typically contain lead (Figure 1), and therefore present the risk of leaching lead (Pb) into pumped water. Specifically, local manufacturers in Tamatave typically fabricate check valves using leather seals and Pb weights, formed using lead extracted from old car batteries. Additionally, well screens are often made of brass, which may contain Pb, and screens are joined to the well casing using lead−tin (Pb− Sn) solder. As water comes into contact with these components, Pb may contaminate the supply via “electrochemical, geochemical, and hydraulic [physical] mechanisms.”10

In eastern Madagascar, access to publically provided potable water is unavailable to many citizens, or is prohibitively expensive.1,2 Furthermore, the capacity of the public utility to increase the number of water connections is limited, and water supply is sometimes unreliable.3−6 Therefore, many people in Madagascar make use of decentralized, household-level water supply (i.e., self-supply).1 Self-supply is defined as the development of household or small-scale water supplies through self-investment, often with low-cost technologies.7,8 Many households in coastal areas of Madagascar, particularly in the city of Tamatave (also known as Toamasina), have installed a locally produced hand pump for some or all of their water needs. In particular, the local population has adopted the simple, inexpensive technology of the pitcher pump.9 Pitcherpump systems have played a role in meeting the domestic water needs of Tamatave residents since the early 1960s and have proven a sustainable option for many households. In fact, approximately 170 000 people (over 50% of the city’s © XXXX American Chemical Society

Received: September 15, 2014 Revised: January 16, 2015 Accepted: January 21, 2015

A

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

through relatively uncomplicated changes in pump operation (e.g., flushing the system before drawing water for consumption) or manufacture (e.g., replacing lead weights in check valves with an alternative metal such as iron). Therefore, the objectives of this paper are to (1) conduct a survey of Pb concentrations in water drawn from pitcher pumps at a set of households in Tamatave, Madagascar; (2) determine if Pb concentrations in pumped water decrease after flushing the systems; (3) perform an analysis of the correlation between Pb concentrations and pump-system characteristics such as age, depth to water table, manufacturer, season, and/or water quality; (4) assess whether replacing Pb check valves with iron check valves decreases Pb concentrations; (5) make a preliminary estimate of the blood lead levels (BLLs) that Malagasy children may experience as a result of exposure to Pb in their household water. The findings of this analysis may have implications for a large number of people. Beyond the tens of thousands of pitcherpump users in Tamatave, the pitcher-pump technology is continuing to spread to other parts of Madagascar, particularly in coastal regions where the groundwater is shallow. Furthermore, the market for pitcher-pump systems in Tamatave appears to be the most robust example of a lowcost self-supply hand-pump market in sub-Saharan Africa;1 therefore, this technology has a great potential for transfer to other African nations, where access to shallow groundwater via hand pumps is already known to be important.20−23 Also, pitcher pumps (or similar “No. 6 pumps”) are known to be in widespread use throughout parts of Asia. To the best of our knowledge, Pb is not used in the fabrication of these pumps, but the high degree of participation of individual manufacturers in markets found in many parts of the world (e.g., as noted by Robinson and Paul24) provides the potential for variability in the materials of construction, particularly for replacement parts that are obtained locally. Overall, then, it is difficult to estimate the total number of people affected by Pb in pitcher pumps, but the number is potentially quite high already, and could grow as the pitcher-pump technology spreads throughout Madagascar, Africa, and beyond.

Figure 1. Diagram of a pitcher pump, reproduced from Mihelcic et al.,9 with permission from ASCE; artwork by Linda Phillips. Insets show photographs of components that are potential sources of lead: (a) pure Pb valve weight; (b) leather valve providing a sliding seal; (c) brass well screen; (d) lead−tin solder.

The negative health effects of Pb exposure are far-reaching and difficult to quantify, and threaten long-term health even at very low concentrations and exposure levels.10−14 In water, there is no lower guideline or limit for the safe consumption of Pb.15,16 In the United States, the Lead and Copper Rule of the Safe Drinking Water Act sets an action level of 15 μg/L for Pb, a threshold above which centralized water supply systems must implement corrosion control or modify infrastructure. However, the action level of 15 μg/L is primarily based on practical considerations of measurement and implementation, not on health. The health-based maximum contaminant level goal (MCLG) for Pb in drinking water specified by the U.S. EPA under the same Act is zero.10,17,18 Based on practical limitations in measuring low concentrations of Pb, and in preventing lowlevel Pb corrosion, the World Health Organization (WHO) has set a provisional guideline of 10 μg/L in drinking water.15 In this paper, the WHO provisional guideline is employed as a threshold level to indicate the importance of Pb leaching from pitcher-pump systems. In a survey completed at 53 households in Tamatave, approximately 67% of respondents reported some use of pitcher-pump systems for potable water (i.e., cooking and drinking).19 However, to the best of the authors’ knowledge, there has only been one published study1 that considered Pb concentrations in household water accessed by pitcher pumps in Madagascar. Those preliminary water quality data, collected in one sampling campaign, showed water from four of 10 pumps to be above the WHO provisional guideline of 10 μg/ L.1 Therefore, it is not yet known if Pb contamination from pump components represents a significant threat to public health in Madagascar. However, based on the number of people who rely on pitcher-pump systems as their primary water source, there is potential for significant effects on public health if Pb concentrations prove to be unacceptably high. This risk justifies further assessment of the magnitude of this threat. Furthermore, if measurements reveal that pumped water contains unacceptable concentrations of Pb, it would be beneficial to determine if Pb concentrations can be reduced

■

MATERIALS AND METHODS Location and Site Selection. The first author served for 2 years in Madagascar as a Peace Corps volunteer as part of a Master’s International graduate program, and was therefore familiar with the local language and culture upon commencement of this investigation.19,25 Tamatave (see map in the Supporting Information) was selected as the study area for its long history and current scale of pitcher-pump use.1 A modified snowball method was employed to initially locate participating households. 26 Details are provided in the Supporting Information. In particular, households were selected to provide a range of pump ages and well depths, and all pump systems included in the sampling campaigns were fabricated and installed by one of six area manufacturers. A survey about pitcher pump use was conducted at 53 households, from which 18 were selected for water quality sampling. Relevant information about the 18 selected households (e.g., pump age, depth to water table) is provided in the Supporting Information. The survey also established that many people use pitcher-pump water for direct consumption via drinking and cooking.19 Sampling Campaigns and Sample Collection. This study comprised three sampling campaigns, conducted during B

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

This U.S.-EPA-approved method is satisfactory for measuring dissolved Pb, but would require acid digestion and long holding times to obtain a measurement for total Pb (i.e., both dissolved and particulate Pb).30,31 It has been demonstrated that particulate Pb (e.g., flakes from solder, brass, or Pb pipes) can be significant under some circumstances.32 However, because flow rates from the pumps are low, and because pitcher-pump systems contain few fittings (Figure 1) and no piped distribution system, it may be reasonable to assume that most Pb is in the dissolved (not particulate) form. Thus, the measurements obtained via the field procedure must be considered lower bounds of the total Pb concentrations, but are likely to be reasonable estimates of those total concentrations. Prior to field sampling, the ASV instrument was tested in the laboratory at the University of South Florida in Tampa, Florida. Groundwater with varying pH and varying concentrations of dissolved oxygen was spiked with known concentrations of dissolved Pb, and then analyzed via ASV as described above. This initial testing validated the instrument’s precalibration, validated the sample pretreatment method, and showed that typical variations in pH and dissolved oxygen did not affect the accuracy of the instrument.19 The water quality parameters of pH, alkalinity, and hardness were recorded in the second and third sampling campaigns with the following testing strips: Fisher Scientific, LLC Whatman pH indicator papers for pH range 6.0−8.1; Fisher Scientific, LLC Insta-Test Alkal Test Strips for alkalinity as carbonate for range 0−180 mg/L; Fisher Scientific, LLC Insta-Test Hard Test Strips for calcium hardness for low range of 0−180 mg/L. Water temperature was typically measured with a thermometer. When available, a Hydrolab Quanta Water Quality Monitoring System (Hydrolab Corporation, Austin, TX), capable of making multiple instantaneous in situ measurements, was used in sampling campaigns 1 and 3 for temperature, conductance, and dissolved oxygen.33 Data Analysis. For each sampling campaign, the median Pb concentration, mean Pb concentration, and standard deviation were calculated for both flushed samples and first-draw samples. To compare Pb concentrations in flushed and first-draw samples, the Wilcoxon signed-rank test (valid for skewed distributions) was used to compare the medians of the concentration distributions.34 To determine if measured Pb concentration is correlated with other factors, Spearman’s correlation was determined for the interrelationship of each variable. The six variables included in this multivariate analysis are Pb concentration (in μg/L), pump age (in years), manufacturer (assigned an integer value of 1 through 6), depth to well screen (in meters below ground surface), season of sampling campaign (assigned an integer value of 1 through 3), and contact time (i.e., flushed or firstdraw conditions, assigned an integer value of 1 or 2). Fourteen pump systems were included in the multivariate analysis; these were the 15 pumps that were sampled in all three campaigns (see the Sampling Campaigns and Sample Collection section above), minus the pump that underwent renovation during the first sampling campaign. A total of 84 samples were used in the multivariate analysis (fourteen pump systems, three sampling campaigns per pump, two samples from each pump during each campaign). Pb Equilibration Time and Effect of Component Substitution. The final field component of this investigation involved two multipurpose case studies. Two households were

different seasons (Dec 2012, Mar−Apr 2013, Jul 2013) to determine if season has a significant effect on Pb concentration. Fifteen of the original 18 household pumps were sampled during all three sampling campaigns. One of those 15 underwent renovation (valve weight replacement, described subsequently) during the first sampling campaign (Dec 2012), and a second pump underwent a similar renovation during the third sampling campaign (Jul 2013). Three of the original 18 households were sampled during the first two campaigns, but not during the third, because they had consistently shown low concentrations of Pb ( 10 μg/L) long after installation (e.g., >20 years). This apparent lack of a “breaking in” period might be due to the frequency with which valves are repaired or replaced in the pitcher pumps in Madagascar. Though some systems may have been installed over 20 years ago, valves are commonly repaired or replaced, as frequently as multiple times per year.1 During valve replacement or repair, the Pb valve weights are reused and affixed to new leather seals (Figure 1), but require some manipulation (e.g., hammering), which may disturb the system for short periods of time. This situation is somewhat analogous to partial lead service line replacement10 in the United States. In a recent study of 32 U.S. homes, measurements of high Pb concentration were often associated with sites having known disturbances to lead service lines.45 Analogously, regular disturbance of Pb valves in pitcher pumps might explain why older pump systems continue to leach Pb at appreciable levels even after years of use. The particular system manufacturer also shows no significant correlation with Pb concentration (p > 0.40). However, it should be noted that more than 50 small businesses that produce pitcher pumps are thought to operate in Tamatave.1 It is known that at least one manufacturer employs a different fabrication process than the common one studied in this paper, specifically using iron in place of the Pb valve weights shown in Figure 1.19 Therefore, given a larger sample size (currently n = 6 manufacturers), this variable could prove significant. Finally, Table 2 shows that the season of sampling does not significantly affect Pb concentrations, as is corroborated with Table 1, which provides similar profiles for Pb leaching in all three seasonal sampling campaigns. Average temperature ranges in April, July, and December are 22−28, 18−24, and 23−29 °C, respectively; average monthly precipitation is 399, 302, and 262 mm.46 Weather in Tamatave may be too similar throughout the year to change groundwater conditions enough to affect Pb leaching from the pumps. A secondary feature of Table 2 is that the correlations between all remaining variables are included. Within these other potential correlations of variables, only C[4,2] = −0.63, the relationship between pump age and well screen depth, is deemed to be statistically significant (p < 0.001). Overall, only the relationship between Pb concentration and contact time and the relationship between pump age and well depth showed correlations significant above a 90% confidence level. The negative correlation between pump age and well depth

conditions in Madagascar is no trivial undertaking, and is therefore considered outside the scope of the current analysis. Therefore, estimates of BLLs provided here by the IEUBK model should be considered preliminary and may be subject to further refinement in the future, if it is possible to obtain better estimates of exposures to Malagasy children.

■

RESULTS Summary of Pb Concentrations in Pitcher-Pump Systems. Table 1 summarizes Pb concentrations measured at the 18 household pumps over the three sampling campaigns. (Data from pumps where Pb valve weights had been replaced with iron weights are not included in Table 1, so N < 18 for the April 2013 sampling campaign.) Concentrations above the WHO provisional guideline of 10 μg/L were observed in 51% of field measurements (considering both flushed and first-draw samples). In fact, 15 of the 18 pumps sampled produced at least one sample with a measured Pb concentration above this provisional guideline. Table 1 provides a preliminary profile of Pb concentrations and shows that pitcher-pump system users might be consuming contaminated water. However, Table 1 also shows that flushing of pitcher-pump systems prior to drawing a sample generally decreases the Pb concentration. In each of the three sampling campaigns, the percentage of pumps producing water above 10 μg/L was greater than 60% under first-draw conditions, but decreased to less than 50% after flushing. Considering all first-draw and flushed samples, 67% and 35%, respectively, exceeded 10 μg/L. The Wilcoxon signed-ranked test34 indicates that the median value of 13 μg/L for first-draw samples is significantly greater than that of 9 μg/L for flushed samples (p < 0.0001). Therefore, pump users could likely reduce their Pb exposure if water drawn after prolonged pump inactivity is employed exclusively for nonpotable uses such as personal hygiene or washing clothes. Analysis of Factors Affecting Pb Concentrations in Pitcher-Pump Systems. Table 2 shows the values for Spearman’s correlation, or measure of association between variables,34 for the six variables considered in this study (Pb concentration, pump age, manufacturer, depth to well screen, season of sampling campaign, and contact time). The left-most column shows that contact time (i.e., whether the sample was flushed or first-draw) is the only variable with a statistically significant correlation to Pb concentrations (p < 0.01). The next-highest correlation with Pb concentration comes from reported well depths (C[4,1] = 0.17), but the value is not statistically significant (p > 0.10). Therefore, it is concluded that well depth does not significantly affect Pb concentrations in the area of Tamatave under consideration. E

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

the benefits of flushing a pump before use. In one instance, after only 6 h of pump inactivity, a water sample read at >100 μg/L, the upper limit of detection for the ASV unit. This scenario provides an example of the concentrations that could be encountered under realistic first-draw conditions (e.g., early in the morning after a night of inactivity). Also from Figure 2, it is estimated that equilibrium between water and leaded components is reached in approximately 4−12 h. Results here are generally consistent with those of Lytle and Schock,27 who found that it can take days of stagnation time for water to fully equilibrate with lead pipe, but that Pb concentrations often begin to level off in the range of 6−16 h, consistent with the specifications of the Lead and Copper Rule.17 One might expect low concentrations (close to zero) from a freshly flushed pump. However, Figure 2 shows that there were measurable levels of Pb even in samples drawn immediately after flushing the pump (contact time = 0 h). Deviations from zero in flushed samples are hypothesized to come from rapid dissolution of Pb-containing components, especially from the Pb valve weights. As water moves through the system (Figure 1), it picks up some initial amount of the metal. For Pump #1, the second phase of data collection (i.e., after valve replacement) was conducted several months after the initial time-release study in a separate sampling campaign due to time and material constraints. For both pilot studies, replacing Pb valve weights with iron weights decreased Pb concentrations in the supplied water. For the first pump, the Pb concentration was 37−100+ μg/L before valve replacement, and 3−4 μg/L after replacement with iron valve weights. In the second pump, the Pb concentration was 7−24 μg/L before valve replacement, and 2−8 μg/L after replacement with iron valve weights. Both retrofitted pumps exhibited concentrations below the WHO provisional guideline consistently for up to 12+ h of contact time (i.e., pump inactivity). Additionally, Pump #1, whose valves were replaced in December 2012 following the first sampling campaign, continued to show low levels of Pb in the final sampling campaign (July 2013), with measurements of 2 μg/L under both flushed and first-draw pumping conditions. These results show that retrofitting existing pumps and/or altering the manufacturing process for new ones might mitigate or avoid issues with Pb contamination. The results in Figure 2 also suggest that most of the soluble Pb present in pitcher-pump systems is drawn from the nominally pure Pb valve weights, at least in these two households; if Pb were leaching predominantly from the well screen or the solder, then replacing the weights would not be expected to result in such a dramatic and sustained decrease in Pb concentration. Though having less overall surface area than either the solder or the well screen, the valve weights have much more Pb available at the water interface, and thereby provide much more leachable Pb to the system. Estimation of BLLs. Figure 3 shows distributions of blood lead levels (BLLs) in Malagasy children estimated by the IEUBK model, based on the four different aqueous Pb concentrations used as inputs. Figure 3 also includes results of a “baseline” scenario in which the water is uncontaminated (0 μg/L). Even though a single value of Pb concentration is used as an input, the output from the IEUBK model is a distribution of expected BLLs, not a single value, because not all children respond identically to a given Pb exposure.39 Figure 3 represents the predicted distributions of BLLs as cumulative distribution functions. Thus, it can be seen, for instance, that if the concentration of Pb in household water is 23.5 μg/L,

indicates that newer pumps are generally installed at greater depths. One possible explanation for this observation is that urbanization could be causing newly arriving families to settle in places where the water table is lower than in older, well established neighborhoods. It is reasonable to hypothesize that some aspect of water quality may control Pb leaching behavior, which might explain variations seen in Pb concentrations (e.g., 2−44 μg/L) in similarly fabricated pumps under the same conditions. Water quality parameters including temperature, pH, conductance, hardness, alkalinity, and dissolved oxygen were measured when possible, and are presented in the Supporting Information. However, none of the measured parameters show a statistically significant relationship with Pb concentration.19 Therefore, a more complete analysis is necessary to make a more conclusive assessment of how the groundwater chemistry affects Pb leaching; previous work by other authors27,47−51 indicates that pH, alkalinity, and other water quality parameters control dissolved lead release in complex ways that we may not be able to discern based on the data gathered so far and the preliminary statistical tests performed here. Time-Release Characterization of Pb Leaching and Pilot Studies for Component Substitution. Two households with pitcher-pump systems exhibiting elevated Pb levels in water were selected during different sampling campaigns to participate in time-release studies to record changes in Pb levels after 0−13 h of pump inactivity. Following the time-release studies, those two households received pump renovations, including the replacement of Pb valve weights with iron valve weights. Then, time-release experiments were repeated with the new valve weights. Pb concentrations measured in samples drawn after various periods of pump inactivity are plotted in Figure 2. From this figure, it is apparent that leaded components leach into water and equilibrate over time as a pitcher-pump system sits idle. This observation strengthens those made via Table 1 regarding

Figure 2. Time-release characterization of Pb leaching and pilot studies of iron-for-Pb valve replacement. Pumps #1 (top) and #2 (bottom) were sampled at various periods of pump inactivity both before and after valve component replacement. F

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

consider, for instance, dietary uptake due to cooking rice (a staple of the Malagasy diet) in Pb-contaminated water. However, consideration of such factors is not trivial and is therefore beyond the scope of the current analysis.

■

DISCUSSION AND IMPLICATIONS The pitcher-pump system has proven to be a locally valued and sustainable self-supply option for household water provision in Tamatave and other coastal areas of Madagascar. To guarantee the continued sustainability of this technology, safe water quality should be ensured in addition to the realized benefits of ample water quantity. Based on the successful history of the pitcher pump as a sustainable self-supply water technology, and on the need to ensure its future success and sustainability in Madagascar, the results of this study have at least three important implications. First, pump users should be educated in the practice of flushing pumps before collection for direct consumption, especially after long periods of pump inactivity such as firstdraw times in mornings. Flushed water likely remains suitable for nonconsumptive use. Inorganic Pb present in water is not absorbed through the skin,52−55 so flushed water could be used for household cleaning, gardening, washing clothes, and bathing and personal hygiene. Furthermore, although this study did not examine whether pump inactivity leads to microbiological growth, it is possible that flushing the well prior to drawing water for drinking may provide an additional benefit of reducing microbiological contamination. Second, pump manufacturers should be educated and trained on the importance of using iron in lieu of Pb as valve weights, both in the manufacture of new pumps and in the modification of existing systems. The current use of iron in the tube well, pump head, and piston arm does not lead to unacceptable aesthetics of the water, so it is not expected that iron valve weights would pose an aesthetic problem, either. Furthermore, iron is reported to be a less expensive material in Tamatave than the Pb recovered from car batteries. One pump manufacturer known to use iron exclusively for valve weight fabrication has indicated that the decision to use iron and discontinue the use of Pb was purely economic, which is an indicator of the potential for economic incentives to drive this improvement to pump construction.19 This manufacturer has made use of scrap iron for several years and has not experienced problems with dissatisfied customers. Both health-related and economic benefits of substituting iron for lead should be promoted throughout Tamatave and other emerging pitcher-pump markets in Madagascar. We emphasize that even simple mitigation measures such as replacing a relatively small part may have a significant effect on the BLLs in children. Third and finally, blood lead levels (BLLs) should be measured in a representative number of children under the age of 5 from households in Tamatave using pitcher pumps as their primary water source. Results should be used to relate Pb exposures in drinking and cooking water to a health indicator. In absence of a formal sampling campaign to establish BLLs of children in Tamatave, this paper has made use of a biokinetic model to predict BLLs under different Pb exposure scenarios relevant to pitcher pump users. The preliminary model predictions suggest than anywhere from 15% to 70% of children might be at risk of elevated BLL. Therefore, actual measurements of BLLs are warranted to provide a more precise

Figure 3. Estimated distributions of blood lead levels (BLLs) in Malagasy children aged birth to 5 years, as predicted by the IEUBK model. Four scenarios are considered, corresponding to different concentrations of Pb in drinking water, as indicated in the figure legend. A “baseline” scenario of uncontaminated water is also considered for comparison.

approximately 60% of children are estimated to have a BLL below 5 μg/dL; thus about 40% of children are estimated to have an elevated BLL above 5 μg/dL, at which point negative health outcomes are expected. From Figure 3, it can be seen that even if the water does not contain any lead (“baseline” scenario of 0 μg/L), the IEUBK model predicts that about 10% of children may experience an elevated BLL (>5 μg/dL). This is due to assumed Pb exposure through other routes (e.g., inhalation, diet). Accounting for Pb in the drinking water then increases the predicted risk of elevated BLL. At a low Pb concentration of 3.95 μg/L (corresponding to the 10th percentile of measured household concentrations), the percentage of children predicted to experience an elevated BLL increases to about 15%, a little bit higher than the baseline case. Not surprisingly, the occurrence of elevated BLL increases as the concentration of Pb in the household water increases. Scenarios based upon the median Pb concentration and the 90th percentile of Pb concentration result in estimates of, respectively, 25% and 40% of children with elevated BLL. For the worst-case scenario of 54.5 μg/L (corresponding to households in which the concentration reaches 100 μg/L during overnight pump inactivity), the IEUBK model predicts that approximately 70% of children may experience elevated BLL. For all scenarios considered, even the worst-case scenario, essentially all children are expected to have BLLs below 20 μg/dL. The model estimates suggest that Pb exposure due to leaded components in pitcher pumps could be detrimental to the health of children in Tamatave, because anywhere from 15% to 70% of children using these pumps may be at risk of elevated BLL, as compared to an estimated 10% of children who are not exposed to contaminated water. This reinforces the importance of the factors considered previously (i.e., that two effective ways of reducing Pb exposure may be to flush the pumps prior to collecting water for drinking and to replace Pb valve weights with iron valve weights). If these measures are able to decrease a household’s average Pb concentration from, say, 23.5 μg/L (90th percentile) to 3.95 μg/L (10th percentile), it may reduce the risk of elevated BLL from about 40% to about 15%. The model estimates in Figure 3 were generated with IEUBK default values for uptake from diet, air, and soil and dust. Therefore, these estimates should be considered preliminary. It is possible that the estimates presented here may be too low, because the default values of the IEUBK model do not G

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology

(7) Sutton, S. Self Supply: A Fresh Approach to Water for Rural Populations; Technical Report for World Bank Water and Sanitation Program - Africa; RWSN/WSP/DFID: Nairobi, Kenya, 2004. (8) Sutton, S. Preliminary Desk Study of Potential for Self Supply in Sub-Saharan Africa; Technical Report for Rural Water Supply Network; RWSN/WaterAid: St Gallen, Switzerland, 2004. (9) Mihelcic, J. R.; Myre, E. A.; Fry, L. M.; Phillips, L. D.; Barkdoll, B. D. Field Guide in Environmental Engineering for Development Workers: Water, Sanitation, Indoor Air; American Society of Civil Engineers (ASCE) Press: Reston, VA, 2009. (10) Triantafyllidou, S.; Edwards, M. Lead (Pb) in tap water and in blood: Implications for lead exposure in the United States. Crit. Rev. Environ. Sci. Technol. 2012, 42, 1297−1352. (11) Lanphear, B. P.; Dietrich, K.; Auinger, P.; Cox, C. Cognitive deficits associated with blood lead concentrations < 10 μg/dL in US children and adolescents. Public Health Rep. 2000, 115, 521−529. (12) Lanphear, B. P.; Hornung, R.; Koury, J.; Yolton, K.; Baghurst, P.; Bellinger, D. C.; Canfield, R. L.; Dietrich, K. N.; Bornschein, R.; Greene, T.; Rothenberg, S. J.; Needleman, H. L.; Schnaas, L.; Wasserman, G.; Graziano, J.; Roberts, R. Low-level environmental lead exposure and children’s intellectual function: An international pooled analysis. Environ. Health Perspect. 2005, 113 (7), 894−899. (13) Needleman, H. Lead poisoning. Annu. Rev. Med. 2004, 55, 209− 222. (14) Rossi, E. Low level environmental lead exposure: A continuing challenge. Clin. Biochem. Rev. 2008, 29, 63−70. (15) World Health Organization (WHO). Guidelines for DrinkingWater Quality, Fourth ed.; WHO Press: Geneva, Switzerland, 2011. (16) World Health Organization (WHO). Lead in Drinking-Water: Background Document for Development of WHO Guidelines for DrinkingWater Quality; WHO Press: Geneva, Switzerland, 2011. (17) Drinking Water Regulations − Maximum Contaminant Level Goals and National Primary Drinking Water Regulations for Lead and Copper, Final Rule; U.S. Environmental Protection Agency: Washington DC; 56 Federal Register 26460−26564: June 7, 1991. (18) Lead and Copper Rule: A Quick Reference Guide; U.S. Environmental Protection Agency: Washington, DC, 2008. (19) Akers, D. B. Lead (Pb) contamination of water drawn from pitcher pumps in eastern Madagascar. Master’s Thesis, University of South Florida, Tampa, FL, 2014. (20) Carruthers, R. M.; Greenbaum, D.; Peart, R. J.; Herbert, R. Geophysical investigations of photolineaments in southeast Zimbabwe. Q. J. Eng. Geol. 1991, 24, 437−451. (21) Kuma, J. S. Is groundwater in the Tarkwa gold mining district of Ghana potable? Environ. Geol. 2004, 45, 391−400. (22) Abo-Amer, A. E.; Soltan, E. S. M.; Abu-Gharbia, M. A. Molecular approach and bacterial quality of drinking water of urban and rural communities in Egypt. Acta Microbiol. Immunol. Hung. 2008, 55 (3), 311−326. (23) MacDonald, A. M.; Calow, R. C.; MacDonald, D. M. J.; Darling, W. G.; Dochartaigh, B. E. O. What impact will climate change have on rural groundwater supplies in Africa? Hydrol. Sci. J. 2009, 54 (4), 690− 703. (24) Robinson, A.; Paul, A. Developing Private Sector Supply Chains to Deliver Rural Water Technology: The Growth of Private Sector Participation in Rural Water Supply and Sanitation in Bangladesh; Technical Report for Water and Sanitation Program; WSP: Washington, DC, 2000. (25) Mihelcic, J. R.; Phillips, L. D.; Watkins, D. W. Integrating a global perspective into engineering education and research: Engineering international sustainable development. Environ. Eng. Sci. 2006, 23 (3), 426−438. (26) Everitt, B. S.; Skrondal, A. The Cambridge Dictionary of Statistics, Fourth ed.; Cambridge University Press: New York, NY, 2010. (27) Lytle, D. A.; Schock, M. R. Impact of stagnation time on metal dissolution from plumbing materials in drinking water. J. Water Supply: Res. Technol.AQUA 2000, 49 (5), 243−257.

estimate of the extent to which leaded components in pitcher pumps represent a risk to public health.

■

ASSOCIATED CONTENT

S Supporting Information *

Map of Madagascar indicating the location of the study; description of the modified snowball method employed to select participating households; pump-system data and waterquality data for the 18 household pumps included in the study; estimation of the volume of water required to flush the pump systems; data for the time-release and component-substitution case studies; preliminary analysis of the correlation between pH and Pb concentration. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*J. A. Cunningham. Address: Department of Civil and Environmental Engineering, University of South Florida, 4202 East Fowler Avenue, ENB 118 Tampa, FL 33620. E-mail: [email protected]. Tel.: +1-813-974-9540. Fax: +1-813-9742957. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This material is based upon work supported by the National Science Foundation (NSF) under grants DUE 0965743 and DUE 1200682. The study was conducted as part of a partnership between the University of South Florida (USF) and RANO HamPivoatra (Water for Progress), a USAIDfunded project in Madagascar implemented by a consortium led by the nongovernmental organizations Catholic Relief Services (CRS) and CARE. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of NSF, RANO HamPivoatra, USAID, CRS, or CARE. The authors thank Austin Atkins of USF for his assistance in validating the anodic stripping voltammetry (ASV) procedures described herein, Meghan Wahlstrom-Ramler for her assistance with collection of field samples, and Onnie Razafikalo for her assistance with various field activities in Tamatave. The authors thank three anonymous reviewers for their constructive comments on an earlier version of the paper.

■

REFERENCES

(1) MacCarthy, M. F.; Annis, J.; Mihelcic, J. R. Unsubsidized selfsupply in eastern Madagascar. Water Altern. 2013, 6 (3), 424−438. (2) WHO/UNICEF. Joint Monitoring Programme for Water Supply and Sanitation. Progress on Drinking Water and Sanitation: 2013 Update; UNICEF and World Health Organization: New York, NY, 2012. (3) African Development Fund. Madagascar, Rural Drinking Water Supply and Sanitation Programme, Appraisal Report; African Development Bank Group: Abidjan, Ivory Coast, 2005. (4) WaterAid. Madagascar: Where Local Commune Administrations Urgently Need More Staff and Resources to Deliver Increases in Sustainable Access to Water and Sanitation; WaterAid: London, United Kingdom, 2005. (5) Madagascar: Water and Sanitation Profile; U.S. Agency for International Development: Washington, DC, 2010. (6) Annis, J.; Razafinjato, E. Public-private partnerships in Madagascar: Increasing the sustainability of piped water-supply systems in rural towns. Waterlines 2012, 31 (3), 184−196. H

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Article

Environmental Science & Technology (28) Brezonik, P. L.; Brauner, P. A.; Stumm, W. Trace metal analysis by anodic stripping voltammetry: Effect of sorption by natural and model organic compounds. Water Res. 1967, 10, 605−612. (29) Palintest. Palintest® SA1100 Scanning Analyzer Manual; Palintest USA: Erlanger, KY, 2009. (30) Palintest. Method 1001: Lead in Drinking Water by Differential Pulse Anodic Stripping Voltammetry; Palintest USA: Erlanger, KY, 1999. (31) Analytical Methods Approved for Drinking Water Compliance Monitoring of Inorganic Contaminants and Other Inorganic Constituents; U.S. Environmental Protection Agency: Washington, DC, 2009. (32) Triantafyllidou, S.; Nguyen, C. K.; Zhang, Y.; Edwards, M. A. Lead (Pb) quantification in potable water samples: Implications for regulatory compliance and assessment of human exposure. Environ. Monit. Assess. 2013, 185 (2), 1355−1365. (33) Hydrolab®. Quanta ® Water Quality Monitoring System, Operating Manual, Revision C; Hydrolab Corporation®: Austin, TX, 2002. (34) Gibbons, J. D.; Chakraborti, S. Nonparametric Statistical Inference, Third ed.; Marcel Dekker: New York, NY, 1992. (35) Moya, J.; Bearer, C. F.; Etzel, R. A. Children’s behavior and physiology and how it affects exposure to environmental contaminants. Pediatrics 2004, 113 (4), 996−1006. (36) CDC Response to Advisory Committee on Childhood Lead Poisoning Prevention Recommendations in “Low Level Lead Exposure Harms Children: A Renewed Call of Primary Prevention”; Centers for Disease Control and Prevention: Atlanta, GA, 2012. (37) Fewtrell, L.; Kaufmann, F.; Prüss-Ü stün, A. Lead: Assessing the Environmental Burden of Disease at National and Local Levels; Environmental Burden of Disease, Series No. 2; The World Health Organization: Geneva, Switzerland, 2003. (38) Rossi, E. Low-level environmental lead exposure: A continuing challenge. Clinical Clin. Biochem. Rev. (Ultimo, Aust.) 2008, 29, 63−70. (39) User’s Guide for the Integrated Exposure Uptake Biokinetic Model for Lead in Children (IEUBK) for Windows®; EPA-540-K-01-005; U.S. Environmental Protection Agency: Washington, DC, 2007. (40) Triantafyllidou, S.; Le, T.; Gallagher, D.; Edwards, M. Reduced risk estimations after remediation of lead (Pb) in drinking water at two US school districts. Sci. Total Environ. 2014, 466−467, 1011−1021. (41) Triantafyllidou, S.; Gallagher, D.; Edwards, M. Assessing risk with increasingly stringent public health goals: The case of water lead and blood lead in children. J. Water Health 2014, 12 (1), 57−68. (42) Deshommes, E.; Prévost, M.; Levallois, P.; Lemieux, F.; Nour, S. Application of lead monitoring results to predict 0−7 year old children’s exposure at the tap. Water Res. 2013, 47, 2409−2420. (43) Maas, R. P.; Patch, S. C.; Pope, J.; Thornton, L. Lead-leaching characteristics of submersible residential water pumps. J. Environ. Health 1998, 60, 8−13. (44) Fact Sheet: Update on Lead Leaching from Submersible Well Pumps and Private Drinking Water Systems; U.S. Environmental Protection Agency: Washinton, DC, 1995. (45) Del Toral, M. A.; Porter, A.; Schock, M. R. Detection and evaluation of elevated lead release from service lines: A field study. Environ. Sci. Technol. 2013, 47, 9300−9307. (46) BBC Weather: Toamasina, Average Conditions. http://www. bbc.com/weather/1053384 (accessed June 1, 2014). (47) Schock, M. R. Response of lead solubility to dissolved carbonate in drinking water. J. Am. Water Works Assoc. 1980, 72 (12), 695−704. (48) Schock, M. R. Understanding corrosion control strategies for lead. J. Am. Water Works Assoc. 1989, 81 (7), 88−100. (49) Dodrill, D. M.; Edwards, M. Corrosion control on the basis of utility experience. J. Am. Water Works Assoc. 1995, 87 (7), 74−85. (50) Edwards, M.; Triantafyllidou, S. Chloride-to-sulfate mass ratio and lead leaching to water. J. Am. Water Works Assoc. 2007, 99 (7), 96−109. (51) Schock, M. R.; Lytle, D. A. Internal corrosion and deposition control. In Water Quality & Treatment, 6th ed.; Edzwald, J. K., Ed.; American Water Works Association: Denver, CO, 2011; pp 20.1− 20.103.

(52) Stauber, J. L.; Florence, T. M.; Gulson, B. L.; Dale, L. S. Percutaneous absorption of inorganic lead compounds. Sci. Total Environ. 1994, 145, 55−70. (53) Filon, F. L.; Boeniger, M.; Maina, G.; Adami, G.; Spinelli, P.; Damian, A. Skin absorption of inorganic lead (PbO) and the effect of skin cleansers. J. Occup. Environ. Med. 2006, 48 (7), 692−699. (54) Abadin, H.; Ashizawa, A.; Stevens, Y. W.; Llados, F.; Diamond, G.; Sage, G.; Citra, M.; Quinones, A.; Bosch, S. J.; Swarts, S. G. Toxicological Profile for Lead; Agency for Toxic Substances and Disease Registry: Atlanta, GA, 2007. (55) Lead Tips: Sources of Lead in Water; Centers for Disease Control and Prevention: Atlanta, GA, 2010.

I

DOI: 10.1021/es504517r Environ. Sci. Technol. XXXX, XXX, XXX−XXX