Assessment of the impact of pit latrines on groundwater contamination in Hopley Settlement, Harare, Zimbabwe

A study was conducted to assess the water quality of the groundwater sources and possible impacts of pit latrines on the groundwater for selected boreholes and wells. The City of Harare ’ s peri-urban settlement of Hopley predominantly uses pit latrines for excreta disposal. This puts groundwater at risk to contamination thereby threatening human health. Pit latrine density around groundwater points was assessed using a Geographical Information System (GIS). The pit latrine density ranged from 0 to 5 latrines in a 15 m radius to 3 – 63 latrines in a 100 m radius. From the analysis of the water samples, it was observed that on average, only 63% and 48% of samples met drinking water quality standards set by the World Health Organization guidelines and Standards Association of Zimbabwe limits. Principal component analysis (PCA) showed that only three components had an eigenvalue of over 1 that explained 76.9% of the total cumulative variance of the observed variable. From the PCA, key parameters in groundwater contamination were nitrates, electrical conductivity, chlorides, ammonia, and thermotolerant coliforms. The spatial variation of the selected water quality parameters suggests that water points at the lowest end of the settlement had the poorest water quality. The point-of-use treatment is recommended for wells. A study was conducted


INTRODUCTION
It is estimated that 70% of the people in the Southern African Development Community (SADC) region rely on groundwater as a major source of drinking water and, in most cases, consumed without receiving treatment to improve quality (Rosa & Clasen ). Peri-urban areas are often characterised by heavily compromised groundwater, with excess levels of nitrate, chloride and microbial pathogens (Xu & Usher ). There is a concern that chemical and microbial contaminants in pit latrines can leach into groundwater sources thereby threatening human health (Dzwairo et al. ; Graham & Polizzotto ). Previous studies on the impacts of pit latrines on groundwater quality have demonstrated deterioration in groundwater quality (Haruna et al. ; Graham & Polizzotto ). Thus, the protection of groundwater sources from pollution by pit latrines is critical. There is strong evidence that access to improved sanitation and safe drinking water can reduce diarrhoea morbidity and mortality and soil-transmitted hel- The majority of houses in Hopley Settlement are temporary to semi-permanent shelters that range from plastic shacks to unapproved structures built with moulded and partially burnt earth bricks (Nyama ). A few houses have been built using approved plans and materials, and a large part of these structures was constructed by the then Ministry of Local Government and Urban Development (Nyama ).

Background on water supply and sanitation in Harare and Hopley
The problems of water and sanitation in Harare have been attributed to the rapid population growth, inadequate rehabilitation, and the maintenance of water and wastewater infrastructure, expensive technologies and the poor institutional framework (Nhapi ). Harare experiences huge water losses reported being around 60% (Ndunguru & Hoko ). This has resulted in erratic water supply and failure by the council to supply water regularly to all areas, with some areas not getting water over 10 years. These problems have led residents to resort to alternative sources including unsafe ones. The residents of Hopley have built structures on unserviced land; hence, development is taking place without adequate water and sanitation support infrastructure (Chirisa et al. ). The residents of Hopley Settlement rely on communal boreholes and standpipes that cannot meet the water demands; hence the use of shallow wells as an alternative source of drinking water.
The World Bank () reported that the existing wastewater treatment facilities in the Greater Harare are not able to treat the existing volumes of wastewater generated. Harare has not been able to expand the sewerage infrastructure, and the most affected areas in terms of infrastructure

Selection of study area and sites
The peri-urban settlement of Hopley was selected as the study area as it is one of the largest peri-urban settlements with serious water supply and sanitation challenges with a potential to trigger another cholera outbreak in Harare. A total of 11 sampling sites were selected, and these included

Methods of data collection
Water sampling and analytical techniques

Mapping of a pit latrine and latrine density assessment
The location of pit latrines in the study area was estab-

Methods of data analysis and interpretation
Spatial variation of groundwater source contaminants Spatial analysis was carried out in a GIS environment. In GIS, spatial interpolation of the groundwater points was applied using the Inverse Distance Weighting interpolation method to create raster surfaces for all the selected water parameters in this study (Nas ). Triangulated irregular network interpolation was used in GIS to create water-

Groundwater quality
Tables 1 and 2 show the results of the mean levels and Student's t-test for the selected groundwater parameters. pH The mean pH of all 44 samples (N ¼ 44) was 6.6 and was acceptable according to the WHO () guidelines and SAZ () drinking water standards. The spatial variation in Figure 2 shows that only sampling points W2, W4, W6,  and W7 had mean concentrations (average of four samples) of 6.15, 6.13, 6.38, and 6.40, respectively, that were less than 6.5, and the lateral distance from the nearest pit latrine was Sample B1 had a reddish brown colour emanating from rusting borehole casing pipes. The acceptability of the water for drinking at water points B1 and W1 was reduced. The lateral distance of the water point from the nearest pit latrine had no effect on the turbidity levels found in the samples.

Dissolved oxygen
Only sampling point W3 had, on average of four samples, DO of 5.4 mg/L that was recommended in drinking water in    Figure 6 shows the spatial variations for nitrates.

Electrical conductivity
Based on the WHO () drinking water guidelines, only samples B1, W1, W2, W3, and W8 had EC at unacceptable mean levels in drinking water, while samples B1, W1, and W2 had levels unacceptable in drinking water in terms of SAZ. The WHO () guideline is more stringent in the  EC limit in drinking water than the SAZ () drinking water standard. Figure 7 shows the spatial variation for EC.

Ammonia
The mean value for ammonia (average four samples) for water samples W1, W2, and W3 exceeded the WHO () guideline of 0.2 mg/L. SAZ () drinking water standards do not have a specified value for ammonia. The spatial distribution for ammonia is shown in Figure 8. The relatively short distance of 3.5 m and 13.0 m of the nearest pit latrine to W1 and W2, respectively, and the direction of the groundwater flow shown in Figure 9 might account for the relatively high concentration levels for ammonia at the water points.

Thermotolerant coliforms
The results showed a mean (N ¼ 44) TTC count of 82 cfu/100 mL that was greater than the permissible count of 0 cfu/100 mL in drinking water. On average of four samples, only water point B1 had TTC levels of 0 cfu/100 mL. Figure 10 shows the spatial distribution of TTC. The colour schemes assigned (blue: 0 cfu/100 mL (no risk areas); orange: 1-100 cfu/100 mL (low to intermediate risk areas); red: >100 cfu/100 mL (high to very high risk areas)) were to reflect the different levels in the spatial distribution of TTC on the water points studied based on WHO () segregation for risks to pathogenic contamination of drinking water.
Water points B1, B2 and W7 had a spatial distribution that suggested no risk to water contamination by pathogenic organisms. Water points B3, W3, and W5 had a spatial distribution of TTC that suggests low to intermediate risk.
Levels above 100 cfu/100 mL were found at water points W1, W2, W6, and W8, thereby putting these sources at high to very high risk to water contamination by pathogenic organisms harmful to human health (WHO ).  It was generally observed that shallow wells at W1 and W2 were not safe for drinking water since they had five out of eight parameters not meeting the recommended levels.
Water points W3, W4, W6, and W7 had four out of the eight parameters that were at levels not safe in drinking water. The short lateral distance of the water points from the nearest pit latrine strongly suggests the poor water quality at these wells. However, water points W5 and W8 had better quality, i.e. three out of eight parameters at levels not safe in drinking water, and the water quality was not affected by the short lateral distance of 10 m and 15 m from the nearest pit latrine. Boreholes B1 and B2 also had three out of the eight parameters at levels not recommended in drinking water. Borehole B3 had the best water quality with only two out of eight parameters at levels unacceptable in drinking water. All samples, except B1, were contaminated with TTC, while seven out of the 11 sampling points had nitrates at levels unsafe in drinking water in terms of the WHO guidelines. The expectation will be a decrease in water quality in wet weather conditions. Therefore, nitrates posed an immediate health risk to consumers. The results also suggest the risk of groundwater contamination by pathogenic organisms due to elevated TTC counts in the water samples.

Determination of key parameters
The data were tested for suitability for PCA through the correlation matrix and Bartlett's test of sphericity. Table 4 shows the Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy. The data were considered suitable for   The result of PCA in Table 6 shows that of the eight components, only three were extracted based on Chatfield & Collin's () assumption that components with an eigenvalue of less than 1 should be eliminated. The extracted three components were rotated according to the varimax rotation in order to make interpretation easier (Table 7).
Based on the component loadings, the variables were grouped accordingly with their designated components as follows: variable. Also based on the scree plot shown in Figure 11, three components that had an eigenvalue greater than 1 were extracted. Therefore, the key parameters in groundwater contamination were nitrates, EC, chlorides, ammonia, and TTC that accounted for 41.7% of the total variance.   Impact of pit latrine density on groundwater quality The pit latrine density was correlated with groundwater levels of pH, turbidity, DO, chlorides, nitrates, EC, ammonia, and TTC (Table 9). The results show that an increase in the number of pit latrines from 15 m to 100 m radius from the groundwater point showed a strong positive linear correlation with levels of nitrate, TTC, ammonia, chloride, and EC, while turbidity had an inverse relationship. DO and pH showed no relationship with increasing pit latrine density.
Nitrate, TTC, ammonia, chloride, and EC were related to an increase in pit latrine density. The results showed that there was a strong association of nitrate, TTC, ammonia, chloride, and EC levels to high pit latrine density that suggested groundwater contamination by pit latrines.

CONCLUSIONS AND RECOMMENDATIONS
The results show that, on average, 63% of all the groundwater samples were acceptable for drinking water in