Type of supply water utilities
The alternative water sources used by the population of the Lefock watershed are wells, springs, boreholes, and surface waters. These water collection points belong to a complex environment: from a geological point of view, the rocks are granites, gneisses, basalts, and ignimbrites. The soils are hydromorphic in the lowlands and ferrallitic on the hillsides and tops. Several anthropogenic activities (pit latrine, domestic wastewater, agricultural plot, domestic breeding, wild garbage dump, and washing clothes) around these supply points (about 5–30 m radius) constitute a risk of contamination for the water, a phenomenon that is common to suburban communities in Cameroon (Tanawa et al. 2002; Djuikom et al. 2009)
The environmental characteristics (within 10m) of the alternative water points sampled show that the water points used by the people in the Lefock Watershed are frequently close to many potential sources of contamination. This is a recurrent situation in small towns in sub-Saharan Africa, where the urban environment develops in an anarchic manner with spontaneous settlements and constitutes a significant contamination risk for the resources exploited by the population (Frédéric et al. 2015; Akakuru et al. 2021).
Focusing on community contamination sources, it revealed that 64% (14/22) of the points are potentially impacted by wastewater, while 68% (15/22) should be under influence of agricultural practices and 9% (2/22) at the proximity of wild garbage dump. Regarding individual practices, it appears that pit latrines, domestic breeding, and washing clothes are found near 50% (11/22), 9% (2/22), and 9% (2/22) of the water sampled points respectively. It is worth noting that some activities are logically associated with one type of water, i.g. washing clothes is found in near-surface waters.
Heterogeneity of water contaminations and link with the surrounding environment
The correlation matrix (Table 2) between the contamination indicator parameters shows 03 observations. Firstly, the existence of a positive and strong correlation of E. coli with faecal streptococci (0.661) and faecal coliforms (0.630). E.coli are positively but weakly correlated with total coliforms (0.318). This suggests that the bacteriological contamination of the water resource is not only faecal. Then, it shows the existence of a positive and average correlation between NO3− and NH4+ (0.549). The correlation between NO3 and PO43− is positive but weak (0.322). This shows that fertilizers are not solely responsible for the nitrate pollution observed in 3 samples analyzed (E1, P2, and P9; Table 1). Finaly, A negative and weak correlation between bacteria content (faecal coliforms, E. coli, faecal and total streptococci) and nitrogen (respectively − 0.387; -0.198; -0.243 and − 0.253 with NO3− and − 0.174; -0.01; -0.005 and 0.051 with NH4+).
Table 2
Pearson correlation matrix between contamination indicator parameters
|
PO43−
|
NO3−
|
NH4+
|
Log FC
|
Log E. Coli
|
Log TC
|
Log FS
|
PO43−
|
1
|
|
|
|
|
|
|
NO3−
|
0,322
|
1
|
|
|
|
|
|
NH4+
|
-0,085
|
0,549
|
1
|
|
|
|
|
Log FC
|
-0,306
|
-0,387
|
-0,174
|
1
|
|
|
|
Log E. Coli
|
-0,170
|
-0,198
|
-0,010
|
0,630
|
1
|
|
|
Log TC
|
-0,139
|
-0,243
|
-0,005
|
0,760
|
0,381
|
1
|
|
Log FS
|
-0,252
|
-0,253
|
0,051
|
0,766
|
0,664
|
0,718
|
1
|
In bold, significant values; log FC: log faecal coliforms; log TC: log total coliforms; log FS: log faecal streptococci |
Similar observations have been made for groundwater in Iganga district in Uganda (Barrett et al., 1999); Niamey in Niger (Ousmane et al. 2006); Douala in Cameroon (Eneke Takem et al. 2009) and Yaoundé in Cameroon (Tabué Youmbi et al. 2013) and suggests that contamination comes from multiple and distinct sources. This feature appears therefore, consistent with the environmental description (within a 10 m radius) of the water supply sites highlighting multiple sources of contamination around these sites.
To assess this possible multi-dimensional contamination impact, a Multiple Factor Analysis (MFA) was carried out on all the data, i.e. 22 individuals (sampling sites) and 18 variables represented by four groups of variables:
- Two quantitative groups (active variables), namely: chemical contamination indicators (NO3−, NH4+ and PO43−) and bacteriological contamination indicators (faecal coliforms, E. Coli, total coliforms and faecal streptococci);
- Two qualitative groups (supplementary variables), namely: physiographic factors (lithology, nature of soil, topographic level, land-use class, and water supply structure) and anthropogenic factors (pit latrines, domestic breeding, domestic wastewater, agricultural activities, illegal dumping of rubbish, cleaning of clothes)
According to the Kaiser criterion, two principal components were extracted, accounting for 65.90% of the total variance in the dataset from all 07 dimensions produced. The projection of the variables on the Dim 1 * Dim 2 factorial plane favors the grouping of the variables into three groups (Fig. 3a):
-The group constituted by the indicator parameters of bacteriological contamination (faecal coliforms, faecal streptococci, E. coli, and total coliforms) which is positively correlated with dimensions 1 and 2;
-The group of nitrogenous indicators of chemical contamination (NH4 and NO3) which is strongly and negatively correlated with Dimension 1, but positively correlated with Dimension 2;
-A group consisting of PO4 is negatively correlated with dimensions 1 and 2.
This distribution of the contamination indicators in space (Dim 1 × Dim 2) confirms a multiple and heterogeneous origin of the contamination of the drinking water resource taken by the alternative points in the semi-urban watershed of Lefock. The detachment of PO43− from the other chemical contamination indicator parameters shows a different source of this element. The highest PO43− content is obtained in sample P1 (Table 1), this sample also has a NO3− concentration above the WHO potability limit and would indicate punctual chemical contamination in the Lefock watershed.
Hierarchical clustering favors grouping of alternative supply sites in the factorial map according to the nature of the contamination (Fig. 3b, Table 3):
Table 3
Distribution of contamination indicator parameters within the clusters formed in the factorial plane dim 1 * dim 2 (according to Table 1 and Fig. 3b)
Cluster
|
Range
|
PO4− (mg/l)
|
NO3− (mg/l)
|
NH4+ (mg/l)
|
Faecal Coliforms (CFU/100 ml)
|
E. coli (CFU/100 ml)
|
Total coliforms (CFU/100 ml)
|
Faecal Streptococci (CFU/100 ml)
|
Cluster 1
(Nitrate pollution)
|
Min
|
0,74
|
56
|
5,6
|
20
|
0
|
1680
|
2
|
Max
|
2,26
|
84
|
140
|
290
|
160
|
7200
|
196
|
Cluster 2
|
Min
|
0,5
|
0
|
0
|
20
|
0
|
190
|
0
|
Max
|
1
|
28
|
72,8
|
570
|
280
|
10800
|
260
|
Cluster 3
highest bacterial content
|
Min
|
0,74
|
0
|
0
|
1300
|
60
|
8200
|
140
|
Max
|
0,85
|
28
|
0,003
|
10800
|
9000
|
42530
|
4320
|
WHO standards
|
|
-
|
50
|
0,5
|
20
|
0
|
-
|
20
|
- Cluster 1 (E1, P1 and P2) is made up of sites with poor chemical quality (nitrate content greater than 50 mg/L);
- Cluster 2 (E2, P3, P5, P6, P8, P10, P11, F1, F2, S1, S2, S3, and S4) is made up of sites with levels of elements indicative of chemical and bacteriological contamination that tend towards the average of the data set analysed;
- Cluster 3 (E3, E4, E5, P4, P7 and P9) is made up of sites with the highest bacterial content.
This spatially heterogeneous distribution of sites according to the type of contamination underlines once again the heterogeneous character of the contamination of the water resource in the Lefock watershed. This heterogeneity of contamination type needs therefore, to be investigated as it reveals that further management should account for multi-dimensional aspect of risks and vulnerability.
Table 4 shows that the clusters constituted in the plan are also a function of the nearby anthropic context and confirms the influence of the activities carried out around the water supply points on the water quality. Cluster 3 groups the most polluted samples (Table 3). This cluster is the only one to present within it the presence of the 06 human activities referenced. This tends to confirm the role of the multiplicity of sources in the contamination occurrence.
Main environmental factors controlling water contamination and vulnerability in Lefock watershed
In order to clarify which type of factors mainly constrain the chemical and bacteriological contaminations, we therefore projected the physiographic and anthropogenic factors (supplementary variables) on the two-dimensional (Dim 1 * Dim 2) plane (Fig. 4a). This shows that on Dimension 1 that Bacteriological and chemical indicators are tightly associated (loadings of 0.63 and 0.65), suggesting that both types of contamination occur preferentially together. Dimension 2 is mainly influenced by chemical contamination indicators (0.65) that however, remain distant from Bacteriological indicators closer to both Natural and anthropogenic factors (loadings respectively of 0.24, 0.27, and 0.12). This suggests that while occurring preferentially together, bacteriological contamination is more related to a combination of anthropogenic and physiographic conditions, while chemical contamination is relatively more sensitive to surrounding physiographic patterns.
Considering these complex patterns, each type of contamination and their most probable drivers are discussed in this article.
Role of anthropogenic and natural drivers for chemical contamination
All three samples, which exceeded the WHO standard for nitrate (E1, P1, and P2; Table 1), were collected in urban areas, with the exception of sample E1, which was collected in surface water in a peri-urban area on the campus of the University of Dschang (the water collected flows from the agricultural and fish farming experimental plots of the University's Faculty of Agronomy and Agricultural Sciences). Samples P3 and P5 were also collected in the urban environment; in this samples, ammonium (NH4+) level well above the average of 0.2 mg/l expected in natural waters (WHO 2017). This shows that the urban environment is more likely to result in chemical contamination of the resource than the peri-urban environment. All of these sites (E1, P1, P2, P3, and P5) have two or three of the following four elements in their immediate environment (Table 2): agricultural plot, domestic wastewater, domestic livestock, and pit latrine. This shows that the urban environment is much more vulnerable to chemical contamination of the resource than the peri-urban environment. The load of pollutants produced is indeed much higher in urban areas due to the higher population density. This result is in agreement with the observation of Lapworth et al (2017) that in sub-Saharan urban environments, the quality status of groundwater resources is often very poor due to inadequate waste management and source protection. Yemeli et al (2021) obtained similar results and showed by geostatistical interpolation of alkaline and alkaline earth element concentrations in groundwater and surface water, the existence of a geochemical anomaly in the center of the city of Dschang (in Cameroon), which she attributed to the much higher anthropogenic pressures in this location. The low values obtained for nitrates can be explained by the fact that agricultural activities in the area are seasonal with the non-intensive use of fertilizers. Overall, the relative disconnection between surrounding physiographic and anthropogenic features (Fig. 4a) would be consistent with the fact that the chemical degradation of the water quality is more related to a general and diffuse infiltration and flowing of water impacted by anthropogenic (both agriculture and urban-related processes) activities.
Role of anthropogenic and natural drivers for bacteriological contaminations
Each sampled point is split in the factorial space into two partial points (Fig. 4b), one of which represents the partial projection of this point in relation to the concentration of chemical indicators (red colour) and the other, the partial projection of the same point in relation to the concentration of bacteriological indicators (green colour).
Considering that bacteriological indicators appear more tightly related with the assessed natural and anthropogenic factors in the surrounding environment, we evaluated how partial individuals (Fig. 4b) influenced by the group of bacteriological contamination indicator variables are oriented in the plane along the SW (least contaminated pole)-NE (most contaminated pole) direction.
In order to clarify which anthropogenic or natural factors would be more related to the bacteriological contamination, then we discriminated each factor to visualize the strongest influence over the bacteriological contamination.
Three anthropogenic factors were found to have the greatest influence on the bacterial concentration measured in the samples studied. They are located in the environment close to more than 50% of the catchment points that are bacteriologically the most contaminated (Table 4). These are the presence of pit latrine, domestic wastewater and agricultural plot. These results complement those reported by Temgoua et al 2010 who attributed the origin of the bacteria observed in the groundwater of Dschang to the proximity of pit latrines.
According to Akhtar et al. 2019, water supply points (wells, boreholes, springs, and surface water) are fed by rainfall through infiltration for groundwater and runoff for surface water carrying pollutants from the soil surface with limited attenuation of contamination. This process would be responsible for the bacteriological contamination of the waters collected by alternative means in the Lefock watershed.
From a physiographic point of view, the hydromorphic or ferrallitic nature of soils (water supply sites on hydromorphic soils are the most vulnerable to anthropogenic contamination) and the nature of the water supply (surface water is the most vulnerable to contamination, followed by wells. springs are the least vulnerable to contamination, while boreholes have a degree of vulnerability shared between springs and wells) are the elements of the nearby natural environment that best discriminate individuals in the factorial design (Fig. 4c).
According to Curmi et al. (1997); landscapes with a hydromorphic domain are characterized by an upstream, homogeneous, and highly permeable domain where water and solute transfers are mainly vertical; and a downstream, multilayered, low-permeability domain with two levels of water tables. This last domain features lateral transfers in the soil and of surface runoff, whose importance depends on the spatial and temporal extension of the saturated surfaces and their connectivity. This process could explain the high concentration of bacteria found in hydromorphic zones in the Lefock catchment.
On the other hand, Curmi et al. (1997) also show that in these areas denitrification is very high, which would justify the fact that these areas are not characterized by high nitrate concentrations although the sampling points are located close to nitrate-generating anthropogenic activities (see paragraph 4.1). Furthermore, the type of pit latrines used in the Lefock watershed is such that they reach the water table in hydromorphic areas. At this level, the groundwater table does not benefit from the vertical protection of the soil that exists in a ferralitic context.