Douala IV is located within the Douala sedimentary basin with coordinates at latitudes 04˚03' - 04˚07' North of the equator and longitudes 09˚37' - 09˚41' East of the Greenwich meridian with an average elevation of not more than 10 m above sea level (Figure 1).

The climate is the wet equatorial type and the rainy season extends from April to November, followed by a short dry season typically occurring between December and March. Its annual total rainfall is generally high with values of up to 4000 mm in August and lower values in the month of February. It has, on the average, 180 rainy days a year. Temperatures are high averaging between 24˚C and 27˚C. The hydro graphical network within the zone of study is very dense and shows a dendritic pattern. The Wouri River constitutes one of the drainage systems of the Douala metropolis and flows all year round through the Akwa, Bonaberi and part of the Bassa zones. This river is fed by many intermittent streams that are normally loaded with most of the solid and liquid wastes channeled from industries, households and waste dumps.

The area is a sedimentary basin of Cretaceous to Tertiary age having a total surface area of 7000 km² and has a maximum width of 60 km. Reference [12] noted that the Douala basin consists of the Cretaceous Mungo River Formation, stratigraphically overlain by the Tertiary Mpundu Formation. The Douala

metropolis lies directly on the Mio-Pliocene to Recent alluvial sediments which constitute the Wouri Formation of the Douala basin. It is made up of unconsolidated fine to coarse-grained sand and gravel mixed with silt and clay in various proportions. The alluviums are composed predominantly of quartz and kaolinite with a general thickness that ranges between 50 and 60 m.

Zonal and azonal soils can be distinguished. The sandy ferralitic zonal soils are light coloured, freely drained with top sandy soils and sandy clay sub-soils. These soils are formed in sandy sediments that are rich in quartz with fairly high humus contents. They have a good friability and low water retention capacity; essentially, they result from the decomposition of sedimentary rocks as in [13] . Young azonal (alluvial) soils are found here and have not been sufficiently subjected to the effects of soil forming processes as in [14] . These are immature soils that are good for agriculture.

Four coastal zone routes in the Douala IV district of Bonassama were selected to check SWI into open wells between November 2014 to May 2015. The locations and altitudes of selected sample sites were determined on the field using a Garmin GPS 60 CX. This enabled the location of the coordinates to produce sample location maps (Figure 1). The routes and 16 of the sampling stations with their location are given in Table 1.

Water was drawn from the open wells using drawing buckets tied with ropes.

Each water sample was collected into a collector after thorough rinsing with the water to be sampled. Water from the collector was filled into 500 mL capacity plastic bottles and preserved to avoid evaporation. The samples were stored in ice-chest containers at about 4˚C prior to laboratory analysis to exclude microbial activity and unwanted chemical activity. Temperature, electrical conductivity (EC), dissolved oxygen (DO), pH salinity and TDS were measured directly on the field using a DO probe and an EC/pH/temperature metre.

Chemical parameters were determined as per standard APHA methods as in [15] . The chemical analyses were performed in the Laboratory for Waste Water Research of the Faculty of Science in the University of Yaounde 1, Cameroon. The total hardness of calcium (Ca²⁺) and magnesium (Mg²⁺) were determined by using the EDTA complexometric titration method; chloride (Cl⁻) by the argentometric method; nitrate, sulphate and phosphate measurements were done using ion chromatography on a Dionex ICS-900. Reference [16] , indicates that the charge balance calculated should be within the acceptable limit of ±5%. Sodium (Na⁺) and potassium (K⁺) were analysed by the flame photometric method; salinity was measured using HACH salinity metre. Bicarbonate () was measured through titration by adding 0.02-M HCl solution to the sample while stirring until the endpoint marked by a pH of 4.5, was obtained.

To get a clear grasp of the spatial chemical composition of water resources in the Douala IV area, water samples were plotted on a Piper diagram using AQUACHEM software as in [17] . The Wilcox formula was used to determine the % Na as in [18] . The physico-chemical properties of ground water resources in Douala IV were further evaluated graphically by plotting EC and sodium absorption ration (SAR) on the United States Salinity (USSL) diagram as in [19] .

A univariate analysis of the physico-chemical data of ground and surface water in Douala IV is presented in Table 2. Water sources characteristics in the study area are not uniform implying that the chemical composition varies over a wide range in physical composition. The average ground water temperature ranged from 27.7˚C to 30.1˚C which is slightly higher than the present day, mean atmospheric temperature of 27.62˚C. Ground water usually maintains a fairly constant temperature which, for superficial aquifers, is normally close to the

mean annual air temperature. In this case, the mean of 27.6˚C was slightly higher than the mean atmospheric temperature of 26.6˚C. This could accelerate the rate of chemical reactions in such systems.

The dissolved oxygen ranged from 0.95 to 2.68 mg/L for ground water and 3.97 mg/L for surface water indicating a low degree of solubility. All water samples showed DO of below 5 mg/L. The measurement of DO is critical to the scientific understanding of the potential for chemical and biochemical processes in groundwater. Concentrations below 5 mg/L may adversely affect the functioning and survival of biological communities and below 2 mg/L may lead to the death of most fish. The measurement of DO can be used to indicate the degree of pollution by organic matter, the destruction of organic substances and the level of self-purification of the water.

The probable low solubility values could be as a result of higher water temperatures and organic matter pollution. Reference [20] highlighted this fact indicating that low DO could have severe adverse effects on the geochemistry and ecology of different systems.

The entire aquifer system is acidic because all samples (100%) showed pH values less than 7. The pH of ground water ranged from 4.7 to 6.2 (Table 2) indicating that the waters are highly acidic to slightly neutral. Similarly, pH value of 6.7 for surface waters indicate that they were slightly acidic. The presence of azonal and zonal soils and organic litter from mangroves in the area may have contributed to the acidic nature of the well waters. This could probably result from the decomposition of soil organic matter as in [21] . Additionally, poor sewage disposal methods in this area compounded by the intrusion of brackish water into the area could account for the pH of the water. Reference [22] , shows that these processes involve ground water reactions with lateritic soils and humic acids from the decomposition of organic matter in the infiltration zones leading to acidification from dissolved carbon dioxide. Related findings as in [23] attest to this. Reference [24] indicated that acidity in ground water could also result from the reduction in the acid buffering capacity due to soil acidification.

The mean TDS of ground water samples ranged from 70.33 - 837.75 mg/L while that for SW stood at 3703.33 mg/L (Table 2). Most of the ground waters except for two points of the coastal transect, are fresh while the surface water are not fresh following the definitions as in [25] who indicated that TDS values < 1000 mg/L represent fresh water. The EC readings ranged from 136.4 - 1661.5 and 1660 - 10500 for well water and surface water respectively. EC values were distinctly higher at stations close to the coast. The salinity data for this period ranged from 0.008 to 0.825 for IT3 and CT respectively while that for surface water ranged from 0.8 - 6 mg/L.

Generally, the findings have shown that salinity decreased with an increase in distance from the coastal transects to inner transects as expected. The highest salinity values were for the coastal transects at SW1 and SW2.

The huge variation in TDS of surface water and ground waters of 70.33 - 837.7 mg/L and 3703.33 mg/L respectively can be ascribed to the fact that these waters were close to the sea where mangroves could be found, probably indicating an intrusion of salt water. High TDS in water may produce bad taste, odour and colour and may also induce unfavourable physiological reactions in the consumer as in [26] . This could probably account for the bad odour and taste of ground water resources as confirmed by some residents of the study area.

The leading cation was Ca²⁺ which had an average percentage concentration of 62.4%, 50.6%, 44.6% and 33.3% as shown in Figure 2 at IT2, IT1, CT, SW and IT3 respectively. This was followed by Na⁺ (55.6%, 33.19%, 32.2%, 21.6% and 16%) in IT3, IT1, CT, IT2 and SW respectively. Mg²⁺ was the third and K⁺ the last in percentage concentration with values of 48.6% and 2.1% in SW and IT1 respectively in that order (Figure 2).

The major anions were sulphate, chloride, nitrate, bicarbonate and phosphate representing 64.29% (IT2), 63.19% (CT), 34.03% (IT3), 8.41% (IT3) and 1.69% (IT1) respectively (Figure 3). The low K⁺ content was probably due to its low content in rocks of the study area.

Nitrate concentrations indicate that 6.3% of sample (one sample) exceeded 10 mg/L (WHO limit for nitrates in drinking water). The probable sources of nitrates could have been domestic detergents (application of fertilizers and domestic waste waters).

The Spearman correlation analysis (Table 3) indicates the association between key ions in water resources within the area of study. The results highlight a strong regression statistical analysis with R² values of 1 for EC-Salinity, EC-TDS, EC- Mg²⁺, Salinity-TDS, Salinity-Mg²⁺, TDS-Mg²⁺. Strong positive correlations of T˚C- DO, , , EC-Cl⁻, , salinity-Cl⁻, , TDS-CL⁻ of 0.9 suggesting that the different aquifer chemistry is controlled by these parameters as in [27] . A strong positive correlation of 0.9 exists between, , Mg²⁺-Cl⁻. Another minor correlations of, , of 0.7, 0.6 and 0.7 respectively which further control the aquifer chemistry of the study area was observed.

The correlations between and of 0.7 and 0.9 respectively could suggest poor sanitation conditions in terms of proximity of toilets to wells and the use of detergents.

The multivariate analysis approach (MAA) has been used in the last years for the assessment of water quality. This is because water quality assessment often requires that several variables be studied concurrently. The use of MAA to link physical and chemical data will enable an overview of the potential cause-effect associated with concentrations of major ions and physical parameters of multi-variable sets as in [28] . MAA values identify the sources and or controlling factors of SWI into the Douala IV coastal ground water as in [29] . It is a multivariate statistical technique to explore variable distributions. The Factor analysis was performed on the correlation matrix that is the variables were auto-scaled (standardized) so as to be treated with equal importance. All analyses were performed using the PCA option of the factor procedure, followed by the basic set-up for factor analysis procedure from the XLSTAT 2015 statistical software.

The present study chose to interpret a variable or group of variables as those associated with a particular factor where loadings were 0.4 or greater, corresponding to an associated explained variance over 65%. This approximates Comrey’s (1973) cut-off of 0.55 for a good association between an original variable and a factor and takes into account discontinuities in the magnitudes of loadings approximating these original variables as in [30] .

The results of varimax with the Kaiser normalisation rotated factor loading for water resource chemistry data in Douala IV are shown in Table 4 and Figure 4 (biplot) respectively which highlight the controlling factors determining the water chemistry of the study area. These resulting factors are vital because they

pinpoint on the controlling processes that lead to the different water facies obtained in this area. Four factor loadings with a 100% variance explained in the dataset are accounted for from Table 4. The loading of variables on each factor and the corresponding percentage of data variance are explained by each factor.

The first principal factor is predorminant and accounts for 61.6% of the total variance with high positive loadings of pH, EC, salinity, TDS, , Ca²⁺, Mg²⁺ and Cl⁻ and high negative loadings of K⁺, Na⁺, ,. This factor can be termed salinity-dissolved solute factor. The first factor determines the main processes that control ground water chemistry. The association of variables (EC, pH, salinity, TDS, , Ca²⁺, Mg⁺ and Cl⁻) in factor 1 (61.6%) of total variance indicates that these variables will impact significantly on the resulting water facies that would be obtained. This is in agreement with the observations as in [24] who indicated that the major ions, pH and TDS are sensitive to variations in EC. The fact that, Ca²⁺ and Cl⁻ ions are under factor one, shows their predominance in the determination of the water types.

Reference [31] , noted that a high negative loading of variables in a factor shows that they go contrary to the process that generated the negative high loadings. Positive loading indicates the interrelationship of variables that determine the water types of the study area by linking their different concentrations.

The second factor explains 21.5% variance indicating that the solubility of DO decreases with an increase in temperature; the reason why they could be found under the same factor. MAAs have been carried out as in [32] - [38] . Reference [39] highlighted two factors that were responsible for controlling hydro-chemi- cal processes of ground water in Togo. Their factors were related to the ions with dominant concentrations and therefore the main contributors to groundwater salinity in this study.

Factor 3 was observed to account for 9.9% of the variance while factor 4 accounted for 6.9% of the total variance and it is associated with high pH and Na⁺ loading and K⁺, and Ca²⁺ loadings respectively for factors 3 and 4. Their preferences in regulating ground water chemistry could therefore be minimal.

The factor analysis does not produce cause effect relationship but highlights the associations from which such cause-effect relationships can be inferred as in [31] . Factor analysis is not used in isolation but it is used with several graphical techniques to make a meaning from hydro-chemical analysis and to identify the water types in the area under investigation. This is why some equations of bivariate plots of EC and chloride and major ions and chloride were exploited together with the Piper diagram to better understand the underlying controlling processes of ground water resources in the study area as in [17] .

The association between key ion chemistry is not enough to deduce the cause- effect relationship of hydro-chemical facies. Reference [40] [41] [42] noted that an early indication of SWI into ground water may be a high EC because sea water has a higher EC value than fresh water.

Additionally, an increase in Cl⁻ concentration may be the first indication of SWI into an area devoid of other possible sources of saline contamination. Consequently, equations and R² value of EC and chloride and major ions and chloride were derived from their corresponding plots to attest to this.

Chloride/EC:

Sodium/chloride:

Calcium/chloride:

Magnesium/chloride:

Sulphate/chloride:

Equation (1) reveals a good correlation of electrical conductivity and the chloride concentration in Douala IV (R² = 0.7782). This indicates that a mixing between sea water and fresh water are taking place. From Equation (1), the mechanisms of mixing between sea water and fresh water are complex because otherwise, there would have been a perfectly linear correlation between electrical conductivity and the chloride concentration as has similarly been observed as in [43] .

From Equation (2), the relationship between sodium and chloride shows a depletion of sodium and a nearly linear relation of calcium and chloride as seen in Equation (3). However, magnesium and sulphate enrichment with chloride, at some points, suggest a strong water aquifer interaction related to direct cation exchange (Equations (4) and (5)). This sample must be from the CT and SW transects as they show a large deviation from the theoretical line of mixing as a result of their high sulphate and magnesium content.

The Piper diagram (Figure 5) indicates the different water types. Reference [44] , indicates that the Piper diagram is preferable because it give a better identification of the water samples and the dorminant geo-chemical processes in the water chemistry.

Figure 5, indicates that the water types are of Ca²⁺-2Cl⁻, Na⁺-Cl⁺, types. Five samples (IT3, 2 IT2 and 2 IT1) representing 26.32% of the total samples are of the Ca²⁺-Cl⁻ type. Two samples (10.53%) of the total samples are of the Na⁺-Cl⁻ type and are found in the IT3 transects and 12 samples (one in IT3, two in IT2, two in IT1, four in CT and three in SW) representing 63.2% of the total number of samples are of the facies water type.

The study shows that the abundance of cations was Ca²⁺ > Na⁺ > Mg²⁺ > K⁺ indicating that Ca²⁺ is the most abundant followed by Na⁺, Mg²⁺ and K⁺. The relative abundance of anions is in the order of. This indicates that the leading anion is followed by Cl⁻, , and. The likely source of Ca²⁺ and could be from the dissolution of gypsum and halite as in [45] which could be present in the catchment area. This could be so because a positive correlation matrix of (R² = 0.7) exists. Reference [46] , shows gypsum to be extensively distributed in the subsurface layer of aquifers and generally its dissolution produces calcium and sulphate as seen in Equation (6).

The correlation between (R² = 0.9) suggests a probably poor sanitation conditions in the area.

The majority of samples are depleted in Na⁺ and enriched in Mg²⁺ and (Equations (2), (4) and (5) respectively). This shows that Na⁺ usually contained in sea water has been replaced by Ca²⁺ on the one hand, but this could also result from the dissolution of gypsum on the other hand.