A key target of the Millennium Development Goals was to halve the proportion of the population without access to safe drinking water by 2015. In an effort to increase access, piped water supply systems have quickly expanded in developing countries such as Mexico. Between 1990 and 2015 the percentage of Mexico’s population with access to “improved” drinking water increased from 82% to 96% [1
]. “Improved” denotes the construction of the water system (e.g., piped water, protected well). However, studies suggest that these “improved” drinking water systems are not analogous to clean, safe drinking water systems when assessed instead by water quality criteria [2
A major issue with “improved” drinking water systems in developing nations is the lack of consistent water supply. Previous studies have shown that the changes in hydraulic pressure due to intermittent water supply (IWS) lead to microbial contamination that can cause dangerous gastrointestinal illnesses [5
]. For example, Kumpel et al. found that samples from taps in Hubli-Dharwad, India, supplied intermittently were more likely to be contaminated by both total coliforms and Escherichia coli
than taps supplied continuously [5
]. Similarly, Erickson et al. demonstrated in Panama that IWS created hydraulic conditions which increased risk for contamination. Despite these impacts, water quality generally met relevant standards [8
]. In addition to these changes in hydraulic pressure due to IWS, household practices to cope with IWS can introduce microbial contamination at the point of use [10
]. Many homes in Mexico store water either in rooftop tanks or in underground cisterns in order to maintain access to water when there is a lapse in the supply. This practice may affect the risk of diarrheal disease [11
This study focused on Mexico’s second largest city, where the practice of using cisterns and rooftop tanks is nearly ubiquitous (Figure 1
). The Metropolitan Zone of Guadalajara (ZMG) encompasses the municipalities of San Pedro Tlaquepaque, Tonalá, Zapopan, Tlajomulco de Zúñiga, El Salto, and Guadalajara. The region is the second most heavily populated in Mexico, with a population that surpassed five million in 2017 [13
], and is expected to reach seven million by 2025 [14
]. Rapid population growth and over-exploitation of the water supply have resulted in a severe water crisis [15
]. The primary water source, Lake Chapala, provides around 60% of the region’s water [16
]. Wells and the Río Calderón serve as secondary sources. Water in the ZMG is provided by the Sistema Intermunicipal de los Servicios de Agua Potable y Alcantarillado (SIAPA). Despite high public concern about the quality of drinking water from these sources, access to independent studies on water quality and data from SIAPA have been limited. The objective of this study was to describe possible issues connected with IWS in the ZMG through a two-fold approach. The first part of the study analyzed biological, chemical and physical parameters linked to microbial contamination. The second part collected information on the public’s perception of water quality.
2. Materials and Methods
Sampling took place in July 2018 during the rainy season. Publicly available maps did not fully delineate the source of water in regions where water service has recently been added. Therefore, an AquaPro™ AP-2 conductivity meter (HM digital, Redondo Beach, CA, USA) was used at each house to determine the water source (Standard Methods No. 2510 [17
]). As conductivity of the three water sources (Lake Chapala, wells, and Río Calderón) is quite different, we used that parameter to differentiate the sources. We determined the conductivity in water from houses where the source was known, then used those values to determine the source where it was unknown. Residential housing stock located on public streets in the ZMG supplied by SIAPA is highly uniform. Therefore, homes used in the study are typical of homes in the ZMG. Figure 2
illustrates the typical water supply infrastructure in the region. Rooftop tanks are automatically filled from the public water supply with system pressure via a float valve. Samples were taken from a tap supplied directly from the mains (number 2 in Figure 2
) and from a tap supplied directly from the storage tank (number 6 in Figure 2
). Initial chlorine residual was measured at 51 houses. Samples were taken for coliform bacteria and E. coli
at 10 of the 51 houses. At these 10 houses, a repeat chlorine residual sample was taken 120 h after the initial sample. A map was created to illustrate the chlorine residual values using QGIS.
To limit ancillary variables, samples taken for bacteriological testing were from areas within the ZMG serviced by the main water provider, SIAPA, and supplied from the main water source, Lake Chapala. The ten homes were chosen based upon willingness to participate, dependency on a rooftop tank, and proximity to the other homes such that all ten homes were accessible by vehicle within a six-hour round-trip. This six-hour maximum holding time between collection and incubation is recommended by the World Health Organization [18
]. Within each neighborhood, houses were approached until a person agreed to participate, or a local contact was able to facilitate access. Each home was tested twice at the same time of day. Samples were collected in 125 mL sterile screw-top polystyrene bottles containing sodium thiosulfate. Samples were transported on ice according to Standard Methods [17
]. The Hach DR900™ (method 8021) was used to test for residual chlorine, the IDEXX Colilert-18™ Most Probable Number (MPN) test was used to test for total coliform bacteria, and the IDEXX Colilert-18™ test for fluorescence was used to test for E. coli
. Analysis of results were compared to the standards stated in the revised Official Mexican Standard NOM-127-SSA1-1994 [19
Prior studies have recognized the numerous factors that make testing for coliform bacteria in low-resourced settings challenging [20
]. One particular challenge is meeting standards on controlled temperatures during transport and incubation. No incubators were available, nor were they obtainable in the region for the duration of the project and within the budgetary constraints. Due to these limitations, an incubator was constructed from a reptile heating lamp, a Styrofoam container, and aluminum foil. A cloth sheet was placed between the lamp and the container to prevent bright light from reaching the samples. The makeshift laboratory and incubator were sterilized with alcohol. The incubator’s ability to maintain a temperature of 35 °C was evaluated 24 h prior to sample incubation. During sample incubation, the temperature of the water (in vials containing distilled water that were placed in the center and the periphery of the incubator) was measured every 3 h over 28 h. Recorded temperatures ranged from 33.5 °C to 36.5 °C. Bacteriologic sample results were analyzed after 24 h and again after 28 h.
In addition to the water quality tests, a brief survey on water access and perception was conducted using Survey123™ (ESRI, Redlands, CA, USA), a mobile-GIS application. Permission for sampling was obtained and the survey was conducted by the local non-profit organization Instituto de Investigaciones Tecnológicas del Agua (IITAAC) in Spanish. They conducted the survey during sample collection. Homeowners’ verbal consent was obtained to collect samples after explaining the purpose of the study and that no one was obliged to answer any questions. No personal identifiers were obtained. Participants were asked about their water storage devices, how often they cleaned them, how they used the piped water, the intermittency of the supply, and the perceived aesthetic quality of the water. A total of 61 surveys were completed; the chlorine residual samples were taken from 51 of these homes.
The results of this pilot study suggest that there may be serious issues concerning the water quality in the ZMG that merit further investigation. Many households reported fluctuations in service from daily cuts to annual losses of service. Participants stated that these events were often associated with unpleasant colors and odors. The decline in water quality reported to occur yearly in April and May is consistent with the marked decline in water quality that occurred during the end of the dry season in 2016, as reported by COFEPRIS, Mexico’s Federal Commission for the Protection against Sanitary Risk (Figure 6
]. On average, the dry season begins in October and ends as the rainy season begins in May [24
]. Although such aesthetic problems are not always harmful, they can be related to contamination. The high chlorine residual in some areas was reported as alarming to residents who complained that their water smelled overwhelmingly of chlorine. We observed a broad range in chlorine residual (both high and low), as well as fluctuations from one day to the next, at the same time and location. These results suggest that disinfection practices in the system may need improvement to supply water that is both palatable and safe.
The effect of the region’s intermittent water supply may have an impact on health because of the household storage of water in rooftop tanks. Half of the tanks tested positive for total coliform bacteria. Although the tanks tested negative for E. coli
, Colilert™ does not test for all pathogenic serotypes of E. coli
. The World Health Organization describes coliform bacteria as a good test for the assessment of disinfection practices but not necessarily an indicator of health risk [25
]. E. coli
is considered the best indicator of fecal contamination. However, both of these bacterial tests are limited in their ability to reveal the presence of other microbes such as viruses and protozoa. Our study indicated that water storage systems, and thus the intermittent water supply that causes these devices to be necessary, may be a contributing factor to the exposure to microbial contaminants. More comprehensive bacterial testing, including speciation of coliform bacteria, would shed light on the extent of the microbial contamination in rooftop tanks in the ZMG. Furthermore, the survey showed that the public had low confidence in the quality of water supplied by SIAPA. The lack of perceived access to clean drinking water may have serious health implications if the lack of perceived access to clean water encourages the consumption of sugary drinks [26
]. Such behaviors are associated with serious adverse health outcomes, such as diabetes, obesity, cardiovascular disease, and tooth decay [27
As a pilot study, there were limitations in our ability to assess the overall microbial quality and public perception of water in the ZMG. The difficulty in randomizing participation in the study and the limited sample size mean that the heterogeneity in water quality and perception in the ZMG were not fully captured. During the survey, some respondents indicated that they did not use the water for cooking but were observed using the water to boil food. Furthermore, only adults were interviewed. In at least one home, children drank the water without a parent’s permission, suggesting that the number of people who are exposed to the water may be higher. It can be difficult to maintain proper conditions during the transportation of samples from the field in warm, rainy climates and improvements in field testing conditions could improve results. Finally, the incubator’s design could be improved to limit samples’ exposure to light. Limited access to equipment is a common issue in testing water quality in low-resourced settings. While temperatures were closely monitored throughout transport and incubation, future work to address bacterial contamination may be improved if access to laboratory incubation or testing is made available for a follow-up study. A broader incubating temperature range has been shown to not significantly impact the performance of the Colilert™ test to detect E. coli
, but has been shown to impact detection of total coliforms [20
]. Regardless, these results demonstrate the real need to further evaluate the water quality that is supplied to over five million people.