Spatial arrangement of well and latrine and their influence on water quality in clayey soil – a study in low-income peri-urban neighborhoods in Lichinga, Mozambique


 In this study, the influence of the spatial arrangement of shallow wells and pit latrines on water quality was evaluated in clayey soil during dry and rainy seasons, using 123 randomly selected wells. The distance between well and the nearest latrine was measured and the location of the well in the yard was characterized. The colony forming units (CFU/100 mL) of fecal coliforms were quantified, and pH, electrical conductivity (CE) and turbidity were measured. 100% of the wells were located less than 25 m from the latrine, 74.8% were located in the middle of the yard. In the dry season, 42.4% of the samples presented up to 12 CFU/100 mL, and in the rainy season, 84.4% presented up to 139 CFU/100 mL. 56.9% had pH values between 6.5 and 8.5, and 63.4% presented EC values between 50 and 571 μS/cm. In the dry season, 40.7% of the samples had values below 5 NTU, and 59.3% up to 50 NTU. In the rainy season, 86.6% had values between 6 and 300 NTU. Pearson's correlation between all variables was weak. The wells are susceptible to high fecal contamination, although the clayey soil seems to mitigate the expected high levels of microbial contamination.

Similarly, soils with high particle size, rocky, sandy, or swampy soils, and the rapid reload of aquifers during heavy rains increase contamination of well water through fecal origin contaminants (Caldwell ; Howard et al. ; Pujari et al. ). Wells located in low relief areas have been reported to have a higher burden of fecal contamination and were more strongly implicated in cases of diarrhea than those located in high relief areas (Uprety et al. ). The literature suggests that microorganisms of greater longevity and/or smaller sizes are more likely to reach the well while they are still viable and/or travel a greater lateral distance from the latrine to the well (Taylor et al. ; Verheyen et al. ; Graham & Polizzotto ).
The access of microorganisms of fecal origin to the wells can occur in several ways, for example: (1) they can be leached from the soil surface, accessing the well through its upper opening; (2) pass through the soil pores, moving from the soil surface or the latrine wall to the well water; (3) water contamination in the well can, also, directly occur through water from contaminated aquifers Nevertheless, it is plausible to say that the level of contamination of the soil surface influences the rate of contamination of the wells by different microorganisms, including those of fecal origin. Thus, it is safe to say, that the higher the microbial load on the soil surface next to the well and fewer barriers that prevent these microorganisms from accessing the well water, the higher is the frequency of contamination and the microbial load in the well water. This thought is corroborated by the longitudinal study of Chambers et al. (), who showed that the surface of sidewalks and puddles near roads had a high relative density of fecal coliforms.
These authors demonstrated that, in each footprint, about 10% of the bacteria on the soles of shoes used for walking were moved to previously sterilized floor surfaces. This study aimed to evaluate the spatial arrangement of well and latrine and their influence on water quality in clayey soil, in low-income peri-urban neighborhoods (Lucheringo and Estação) in Lichinga city. Lichinga city was chosen because it is located in an area with clay soils (MAE ). These low-income suburban neighborhoods were selected because the wells and latrines in these areas are dug by hand, most of them without interior lining, and are the main source of water and for treatment of human feces.
A significant association between family income and the proportion of positive samples for total coliforms for well water in rural areas was found (Smith et al. ). These authors' findings suggest that low-income families tend to build wells that are more susceptible to fecal contamination than higher-income families. This is particularly true because low-income families in suburban neighborhoods generally live in smaller yards that do not allow safe spacing between the well and the latrines, which are hand-dug, shallow, and without an interior lining, increasing the possibility of water and feces to directly come in contact with the soil Considering all of this, the wells in the Lucheringo and Estação neighborhoods were suitable for the purpose of this study, as the characteristics of the wells and latrines in these neighborhoods allow us to hypothesize the worse possible scenario of well water contamination by latrines. This clearly allows understanding the influences of the lateral distancing and the location of the wells in the yard for the water quality of shallow wells on clay soil.

Study area
The study was carried out in the district of Lichinga, at Administrative Post number 2, Chiuaula, specifically in the Lucheringo and Estação neighborhoods. Lichinga district includes suburban neighborhoods and the city of Lichinga, which is the capital of the province of Niassa, located in northern Mozambique. Lichinga district is a plateau region, above 1,000 m altitude, and has a territorial exten- In the entire district, only 1% of the families have access to safely managed drinking water sources inside or outside their yards. About 64% use water from wells and, of that percentage, 75% use unimproved water sources, and about 29.7% use surface water from ponds or rivers (INE ).

Water sampling
The study was longitudinal and covered 123 shallow domestic wells. All the selected wells for the study were dug by hand and are cylindrical, not exceeding a depth of 20 m and without an internal lining. Stratified probabilistic sampling was used to split the two neighborhoods into four plots, according to the number of houses ( Figure 1).
Within each plot, the wells were randomly selected, observing an interval of 25 houses organized in a straight line to separate the sampling points.
The water was not analyzed directly in the well. Samples of 500 mL of water were collected in PET bottles of mineral

Field data collection
Before collecting each water sample, the lateral distance between the well and the nearest latrine was measured with a tape measure. The distances between the well and each latrine belonging to the entire adjacent lot of latrines were considered, and the shortest distance was included in the data.
The location of the well in the yard was visually determined. The well located in the central part, or in any other part of the yard close to where people usually walk, was considered to be in the middle of the yard and was coded with the number 1; the opposite was considered to be at the edge of the yard and was coded with number 0.

Analysis of water samples
The microbiological analyses aimed to determine the number of colony forming units (CFU) of fecal coliforms, using method 10029, from the company Hach, as previously described in the literature (Adhikari et al. ), with modifications. Briefly, 100 mL of undiluted sample was filtered and the membrane was transferred to a plate containing an absorbent pad that was previously soaked in 2 mL of tryptose lauryl sulfate broth (LSB). Fecal coliforms CFU were checked after 24 hours of incubation at 44 C. Mineral water samples were used for the negative control. All the colonies were counted, and the results below the detection limit were considered 0 to find out the average. The turbidity was measured using a digital turbidity meter (HANNA brand HI 98703 Turbidimeter), based on the manufacturer's guidelines. The pH and electrical conductivity (EC) were measured using a HANNA-edge digital multifunctional meter, based on the manufacturer's guidelines.

Data analysis
An arithmetic average was calculated between the CFU values of the duplicates of each sample. The normality of data was verified by the Shapiro-Wilk test at 5% level of probability. The data were submitted to descriptive statistics (minimum, maximum, average, and standard deviation). The two-tailed paired t-test was used to determine the differences between the data from dry and rainy seasons. As well, Pearson's correlation was determined among the variables studied. Significant differences were considered once p < 0.05. The analyses were done based on BioEstat 5.0 and GraphPad prims 8.02 software.

RESULTS AND DISCUSSION
The results of the 123 shallow domestic wells randomly selected revealed that 58.5% of the wells were located at a lateral distance less than 15 m from the nearest latrines ( Figure 3(a)). Most of the domestic wells (74.8%) are located near the areas people usually walk, even when they are not fetching water.
The results (Figure 3 In general, it was noticed that a considerable number of the wells were built as distant from the latrine in the same yard as possible, but in most cases, it was noticed that they were close to the latrines in the neighboring yards. It is necessary to mention that the distances considered safe by the literature are impossible to be applied in the contexts of peri-urban settlements, as the yards are not large enough to allow the ideal distance between the well and the latrine. This means that the spatial coexistence between wells and latrines in peri-urban contexts will always be a problem; therefore, the adoption of water treatment methods at the point-of-use is extremely necessary.  The results of fecal coliforms are expressed as CFU/100 mL, electrical conductivity (EC) as μS/cm and turbidity (turbidity) as NTU.
Our results (Figure 3(b)) also show that most of the wells were dug in the middle of the yard, meaning that their opening (mouth) is close to the place where people usually walk. This fact is particularly worrying, because the more centralized the well, the greater the chance of it being near the latrine in the same yard. In this study, the size of the yards was not measured and this is a limitation.
Although the yards selected were not large enough consid- Around the well should be delimited by a fence, and the top of the well should be covered with a waterproof cover and sufficiently wide and oblique. It is necessary to encourage communities, especially in peri-urban contexts, to not consider water from domestic shallow wells as good for immediate consumption, water should be treated before its consumption, as previously recommended (Sobsey & World Health Organization ).
The numbers of CFU/100 mL of fecal coliforms found in the dry season and in the rainy season significantly differed from each other (p < 0.05). In 57.6% of water samples collected in the dry season, no fecal coliform CFU was detected, 35.6% of the samples had less than 9 CFU/ 100 mL and 6.8% had up to 12 CFU/100 mL (Figure 4(a)).
In the rainy season, fecal coliforms CFU were not detected in 15.6% of the samples, 40.6% of the samples had less than 9 CFU/100 mL, and 42.1% had up to 138 CFU/ 100 mL (Figure 4(b)).
The wells are at risk (Figure 4 Unlike the dry season, in the rainy season, the percentage of contaminated wells and the number of CFU counted per 100 mL were significantly high (Figure 3(b) and Table 1

).
It is interesting to note that although the number of CFU/100 mL in the water of many wells was high, especially in the rainy season, the density of coliforms was unexpectedly low, since the wells are highly susceptible to fecal contamination. This seems to result from the mitigation of the contamination by the clayey soil. This assumption is reinforced by the results in Table 2, which show a weak Pearson correlation between the variables studied.
We believe that the location of the wells in the yard close to where people often pass, as well as the manual lifting of water from the well using a rope that often comes into contact with the soil, are factors that also contribute to contamination of the water. Likewise, because the soil is not very permeable to water infiltration, on days of intense rain, water can flood in the backyards and latrines (which are shallower than wells), spreading fecal contaminants over the surface. Associating this factor with the factors previously mentioned (proximity to latrines, location of wells in the yard, moisture on the soil surface for longer, manual between 11 and 50 NTU (Figure 7(a)). In the rainy season,   12.5% of the wells showed turbidity values up to 5 NTU, 7.8% between 6 and 10 NTU and 79.7% between 15 and 300 NTU (Figure 7(b)).
The water turbidity values showed a wide temporal variation, being moderate in the dry season (Figure 7(a)) and extremely high in the rainy season (Figure 7(b)). The low level of water turbidity is partly explained by the fact that Mozambique has just two seasons, rainy and dry. Therefore, the dry season is very long and, consequently, a long time

CONCLUSIONS
This study aimed to evaluate the influence of the spatial arrangement of wells and latrines on water quality in clayey soil. The distance between the well and the latrine was measured, the location of the well in the yard was characterized, and the microbiological and physical-chemical quality of the water was analyzed during the dry and rainy season. This study confirmed the impact of latrines on fecal water contamination in shallow wells, in the context of low-income peri-urban neighborhoods, and it was found that the wells are highly vulnerable to microbial contamination of fecal origin. However, although microbial contamination is high, especially in the rainy season, the level of contamination was below the expectations, and this seems to result from the mitigation of the contamination by the clayey soil.