The study area is located in southeastern Côte d’Ivoire, between latitudes 5˚10' N and 5˚33' N and longitudes 3˚12' W and 3˚50' W, covering approximately 1864 km² and including the departments of Alépé, Bonoua, Aboisso, and Adiaké (Figure 1).

Figure 1. Geographic location of the study area.

It overlies the Continental Terminal aquifer of Bonoua, a strategic groundwater resource exploited through wells and boreholes for drinking water supply and agro-industrial activities. Dating from the Mio-Pliocene, the aquifer consists of sands, sandstones, clays, and conglomerates with high permeability and represents one of the major groundwater reservoirs in Côte d’Ivoire ([23]).

The region experiences a transitional equatorial climate with two rainy and two dry seasons, and annual rainfall varies from 800 to 2000 mm. The original vegetation (dense evergreen forests and littoral psammo-hygrophilous formations) has been largely modified by agricultural expansion, including industrial crops (rubber, oil palm, pineapple) and subsistence crops (cassava, yam, and plantain) ([31]). The hydrographic network includes the Comoé, Bia, and La Mé rivers, as well as the Soumié and Toumanguié tributaries, which feed the Aby, Potou, and Ébrié lagoons. The Comoé River, with an average discharge of approximately 300 m³/s, plays a central role in regional hydrological dynamics ([13]).

A water sampling campaign was carried out in July 2014, during the rainy season, targeting 30 wells. The selection of wells was based on several criteria, including accessibility, proximity to agricultural activities, and the presence of potential pollution sources such as latrines or septic tanks. Each water sample was collected using a dedicated dipper previously sterilized by rinsing to avoid contamination. Samples were transferred into pre-cleaned polyethylene bottles and stored under appropriate conditions. Sampling coverage was limited by the presence of non-functional wells and access difficulties caused by road conditions. Additional site information (including UTM coordinates and elevation) was recorded for each sampled well (Table 1).

Table 1. Location and geographic coordinates of sampled wells in the Bonoua area (Continental Terminal aquifer, southeastern Côte d’Ivoire).

Samples were preserved on dry ice and subsequently stored at +4˚C before laboratory analysis at the CIAPOL laboratory within 24 hours of collection. During the field campaign, several physicochemical parameters were measured in situ, including pH, redox potential, temperature, electrical conductivity, salinity, total dissolved solids, dissolved oxygen, and turbidity. These measurements were performed using a portable multiparameter probe (HANNA HI9828). Laboratory analyses focused on major ions (Ca²⁺, Mg²⁺, Na⁺, K⁺, Cl⁻, SO 4 2 − , HCO 3 − ), heavy metals (Pb²⁺, Cd²⁺, Fe²⁺, Mn²⁺, Al³⁺, Cu²⁺, Zn²⁺), and nutrient species ( NO 3 − , NO 2 − , NH 4 + , PO 4 3 − ), along with selected physical parameters (suspended solids, hydrotimetric titre). All analyses were conducted following standardized protocols ([4]; [30]) (Table 2).

The dataset, although collected in 2014, remains relevant for the present study. Since the sampling period, land use, industrial activities, and population density in the Bonoua region have remained relatively stable. No major environmental or anthropogenic changes have occurred that could significantly affect groundwater quality. In the context of limited recent data in West Africa, these historical records provide a reliable basis for evaluating water quality and supporting sustainable resource management.

Table 2. Analytical methods used for the determination of chemical parameters.

The assessment of well water quality was conducted using an integrated analytical approach combining several tools:

comparison of measured concentrations with World Health Organization ([47]) guideline values;calculation of the Water Quality Index (WQI) for a synthetic evaluation of water suitability;estimation of non-carcinogenic health risks using Hazard Quotient (HQ) and Hazard Index (HI); andapplication of Principal Component Analysis (PCA) to identify potential sources of contamination.

2.3.1. Comparison with WHO Drinking Water Standards

The concentrations of the analyzed physicochemical parameters (physical, major ions, trace metals, and nitrogen compounds) were compared with the [47] drinking water quality standards—this comparison aimed to evaluate the suitability of well water for human consumption.

2.3.2. Calculation of Water Quality Index (WQI)

The WQI was calculated using the weighted arithmetic method of [9], commonly applied in hydrogeochemical studies ([32]; [41]). This approach integrates multiple water quality parameters into a single index that reflects the overall potability status. The relative quality (Qi) of each parameter was determined using Equation (1):

where:

Vactuel: measured concentration of the parameter;Vstandard: WHO guideline value ([44]);Videal: optimal value (7 for pH, 14 mg/L for DO, and 0 for all other parameters) ([27]; [41]; [2]).

The relative weight (Wi) is defined by Equation (2):

Where Si is the WHO standard for the parameter and K is a normalization constant.

The overall WQI was then calculated using Equation (3):

In this study, nineteen parameters were incorporated into the analysis (pH, EC, TDS, Cl⁻, NO 3 − , SO 4 2 − , NH 4 + , Ca²⁺, Mg²⁺, Na⁺, K⁺, HCO 3 − , Fe²⁺, Zn²⁺, Cu²⁺, Al³⁺, Mn²⁺, Pb²⁺, Cd²⁺) because of their potential impact on human health ([44]; [32]; [41]), and their relevance to groundwater quality assessment in tropical regions affected by both geogenic and anthropogenic inputs.

The classification of WQI values followed the criteria proposed by [41] in Table 3.

Table 3. Classification of water quality based on the weighted arithmetic WQI method.

2.3.3. Health Risk Assessment

The health risk assessment related to the ingestion of contaminated water was performed following USEPA guidelines ([42]; [44]; [45]) to quantify non-carcinogenic effects from chronic exposure. The main indicators are the Chronic Daily Intake (CDI), Hazard Quotient (HQ), and Hazard Index (HI). This approach differentiates between adults and children, accounting for distinct ingestion behaviors and physiological characteristics.

1)Chronic Daily Intake

The CDI expresses the amount of contaminant ingested over a prolonged period and is calculated using Equation (4):

where:

C: contaminant concentration (mg/L);IR: ingestion rate (L/day);EF: exposure frequency (days/year);ED: exposure duration (years);BW: body weight (kg);AT: averaging time (days), calculated as ED × 365 for non-carcinogenic risk.

The parameter values used, obtained from [44], are given in Table 4.

Table 4. Typical exposure parameters used for health risk assessment ([44]).

2)Hazard Quotient (HQ)

The Hazard Quotient (HQ) measures the non-carcinogenic risk of a specific contaminant and is calculated using Equation (5):

where RfD is the reference dose (mg/kg/day), defined as the exposure level below which no adverse effects are expected.

The interpretation criteria are as follows:

HQ < 1: negligible risk;HQ ≥ 1: potential health risk.

Only contaminants with available RfD values (Table 5) were considered, namely Cu²⁺, Zn²⁺, Cd²⁺, Mn²⁺, Al³⁺, Fe²⁺, NO 3 − , NO 2 − , and NH 4 + . Pb²⁺ and Cu²⁺ were excluded due to concentrations below detection limits.

Table 5. Reference doses (RfD) of the parameters considered.

RfD: reference dose expressed in milligrams per kilogram of body weight per day.

3)Hazard Index (HI)

The Hazard Index (HI) represents the overall non-carcinogenic risk from simultaneous exposure to multiple contaminants and is computed as the sum of individual HQs, Equation (6):

The interpretation is based on the same principles as above:

HI < 1: no combined effect expected;HI ≥ 1: potential for synergistic or cumulative toxic effect.

2.3.4. Principal Component Analysis (ACP)

Principal Component Analysis (PCA) was applied to the hydrochemical dataset to identify latent structures, reduce variable redundancy, and detect potential sources of contamination. This multivariate approach is particularly effective in analyzing complex environmental systems where multiple parameters interact simultaneously ([22]; [34]). The data were standardized (mean-centered and scaled to unit variance) to eliminate the influence of differing measurement units. The adequacy of PCA was tested using the Kaiser-Meyer-Olkin (KMO) index and Bartlett’s sphericity test. Principal components were retained according to Kaiser’s criterion ([17]), selecting only those with eigenvalues greater than 1. Interpretation was based on the correlation matrix, total variance explained, and component loadings. The analysis included all physicochemical parameters contributing to the determination of WQI and HI, namely Cl⁻, HCO 3 − , SO 4 2 − , K⁺, Na⁺, Ca²⁺, Mg²⁺, NO 3 − , NO 2 − , NH 4 + , EC, Cd²⁺, Mn²⁺, Al³⁺, and Fe²⁺. All statistical analyses were performed using SPSS version 29.0, a widely used software package for environmental multivariate analysis.

The statistical summary of the physicochemical parameters, major ions, nutrients, and heavy metals measured in situ and in the laboratory is presented in Table 6. These results are expressed through standard statistical descriptors (minimum, mean, maximum, median, standard deviation, and coefficient of variation (CV)).

Overall, the measured parameters exhibit wide variability, reflecting the influence of both natural and anthropogenic factors on groundwater composition within the Continental Terminal aquifer of Bonoua.

Table 6.Physicochemical characteristics of well water from the Continental Terminal aquifer of Bonoua during the rainy season (July 2014), compared with the ([47]) guideline values.

ND = not detected; SD = standard deviation; CV = coefficient of variation.

3.1.1. Physical Parameters

The water temperature ranged from 25.10˚C to 28.70˚C, with a mean of 27.09˚C, which is slightly above the WHO guideline of 25˚C, consistent with typical tropical aquifer conditions. Suspended solids (SS) varied from 17.25 mg/L to 52.78 mg/L (mean: 27.92 mg/L), suggesting moderate turbidity associated with surface infiltration. Turbidity displayed substantial variability (CV = 124.27%), remaining below the WHO limit of 5 NTU for most wells, although some reached up to 11.70 NTU.

The oxidation-reduction potential (Eh) ranged from 39 mV to +122 mV (mean: 63.51 mV), indicating heterogeneous redox conditions across wells, with both reducing and oxidizing environments. Such variability often reflects differences in organic matter content, recharge rate, and microbial activity within the aquifer.

3.1.2. Physicochemical Parameters

pH values (4.04 - 6.87; mean = 5.16) indicate pronounced acidity in most wells, below the WHO acceptable range (6.5 - 8.5). This acidity may result from the oxidation of organic matter or the leaching of lateritic soils rich in iron and aluminium oxides. Electrical conductivity (EC) (28.81 - 420.65 µS/cm; mean = 161.85 µS/cm) and total dissolved solids (TDS) (10 - 220 mg/L; mean = 86 mg/L) are well below WHO limits (1500 µS/cm and 1000 mg/L, respectively), confirming low mineralization and limited ionic enrichment. Salinity remained low (mean = 0.08), and total hardness (THT) averaged 3.09˚F, classifying the water as very soft. Dissolved oxygen averaged 4.14 mg/L (slightly below the 5 mg/L guideline), suggesting occasional oxygen depletion due to microbial activity in shallow wells.

3.1.3. Majors Ions

Major ion concentrations were generally low (Table 6). Mean concentrations were 14.79 mg/L for Cl⁻, 21.19 mg/L for HCO 3 − , and 7.27 mg/L for SO 4 2 − , all well below WHO limits. Cation concentrations also remained moderate: K⁺ = 6.62 mg/L, Na⁺ = 6.14 mg/L, Ca²⁺ = 8.09 mg/L, and Mg²⁺ = 2.56 mg/L.

The high coefficients of variation (34% - 79%) for most ions indicate localized inputs, possibly linked to variations in lithology or anthropogenic activities (fertilizers, domestic wastewater). The dominance of bicarbonate and chloride suggests mixed sources, with groundwater influenced by both mineral weathering and infiltration of surface runoff.

3.1.4. Nutrients

Nitrate ( NO 3 − ) concentrations (1.70 - 21.87 mg/L; mean = 14.33 mg/L) remained below the WHO threshold (50 mg/L), suggesting limited fertilizer leaching during the rainy season. Nitrite ( NO 2 − ) (mean = 0.04 mg/L) and ammonium ( NH 4 + ) (mean = 0.27 mg/L) showed very high CVs (> 120 %), implying sporadic contamination from local sanitation sources or decomposing organic matter.

Phosphate ( PO 4 3 − ) (mean = 0.18 mg/L) also remained within permissible limits (0.5 mg/L) but indicates occasional domestic or agricultural inputs. Overall, nutrient concentrations confirm moderate anthropogenic influence with spatial heterogeneity across wells.

3.1.5. Heavy Metals

Among the analyzed trace metals, several exceeded WHO guidelines (Table 6). Cadmium (Cd²⁺) (mean = 5.10 µg/L; max = 78 µg/L) surpassed the permissible limit (3 µg/L), indicating possible inputs from industrial or agricultural sources. Manganese (Mn²⁺) showed particularly high levels (mean = 350 µg/L; max = 998 µg/L), exceeding the 50 µg/L guideline by up to sevenfold. Iron (Fe²⁺) concentrations (mean = 203 µg/L; max = 1023 µg/L) often surpassed the 300 µg L⁻¹limit, consistent with natural leaching of lateritic soils. Aluminium (Al³⁺) (mean = 111.53 µg/L) occasionally exceeded the 200 µg/L threshold, while zinc (Zn²⁺) (mean = 93.9 µg/L) remained far below the 3000 µg/L limit.

Copper (Cu) and lead (Pb) were below the detection limit of 1 µg/L, indicating minimal industrial contamination. These results reveal that metal enrichment primarily arises from geochemical weathering and soil leaching, aggravated by acidic pH that promotes metal mobility. Due to their concentrations being below the detection limit, Pb and Cu were excluded from the health risk calculations to ensure methodological transparency.

The computed WQI values for the thirty (30) sampled wells (Table 7) exhibited substantial spatial variability, ranging from 7.08 (P1) to 758.40 (P10), with a mean of 136.55 ± 142.80 (CV = 104.57%), indicating pronounced heterogeneity in groundwater quality across the study area. The lowest WQI (7.08) recorded in well P1 falls within the “excellent” category, suggesting suitability for human consumption without treatment, whereas well P10 exhibited an “inferior” quality (758.40), reflecting significant contamination. The classification of wells according to WQI categories revealed that 10% were excellent (WQI < 25), 13.33% good (26 ≤ WQI ≤ 50), 20% medium (51 ≤ WQI ≤ 75), 16.67% poor (76 ≤ WQI ≤ 100), and 40% very poor (WQI > 100). Overall, 76.67% of wells recorded WQI values exceeding 50, implying that most groundwater samples do not fully meet WHO drinking-water standards without prior treatment.

Table 7. Calculated WQI values and water quality classification for the 30 sampled wells.

WQI = Water Quality Index. Classification based on the weighted arithmetic method (see Table 3).

The detailed HQ and HI values by parameter and by well are presented in Appendix A (Table A1 and Table A2). The summary of non-carcinogenic risk indices (HI) for children and adults is shown in Table 8, while Figure 2 provides a graphical representation of these indices on a linear scale. HI values range from 0.87 to 84.68 for children and from 0.37 to 36.29 for adults, with mean values of 19.50 and 8.49, respectively. These results indicate considerable variability in risk levels among wells, with children systematically exhibiting higher HI values than adults.

Table 8. Non-carcinogenic hazard index (HI) values for children and adults in groundwater samples from the Continental Terminal aquifer in Bonoua (Southeastern Côte d’Ivoire). HI values greater than 1 indicate potential non-carcinogenic health risks, while values below 1 suggest negligible risk levels.

Based on the risk classification, most wells present potential health risks (HI > 1) for both children and adults, whereas only a few wells (P8, P9, P12, P16, P17, P20, P25, and P26) show negligible risk. The graphical comparison confirms this pattern, clearly highlighting the predominance of elevated HI values, particularly among children. The highest HI values, observed in wells P3, P4, P5, P11, P23, and P27, identify priority sites for monitoring and potential mitigation measures. Overall, 97% of the wells exceed the threshold value of 1 in children, and 73% in adults, revealing a significant non-carcinogenic health risk in the study area.

Figure 2. Variation of the non-carcinogenic hazard index (HI) for children and adults (linear scale). The red horizontal line represents the threshold limit (HI = 1) separating negligible from potential health risk levels.

The Kaiser-Meyer-Olkin (KMO) index (0.733) confirmed the adequacy of the dataset for multivariate analysis, while Bartlett’s test of sphericity (χ² = 689.187; df = 105; p < 0.001) indicated significant inter-variable correlations (Table 9).

These statistical results validated the application of Principal Component Analysis (PCA) to the dataset.

Table 9.Kaiser-Meyer-Olkin (KMO) index and Bartlett’s test of sphericity.

The correlation matrix (Table 10) showed very strong positive associations among the major ions (Cl⁻, SO 4 2 − , Na⁺, K⁺, HCO 3 − , Mg²⁺, and EC; r > 0.95), reflecting a common geogenic origin mainly related to mineral dissolution and ion exchange processes. Moderate correlations between NH 4 + , NO 2 − , and dissolved ions (r ≈ 0.30 - 0.50) indicate possible agricultural inputs. In contrast, NO 3 − exhibited negative correlations with most major ions (r ≈ −0.45 to −0.55), suggesting distinct anthropogenic contamination pathways, likely from domestic or agricultural effluents. Transition metals such as Mn²⁺, Fe²⁺, and Al³⁺ displayed moderate to strong correlations (up to 0.60), implying joint mobilization under reducing and acidic conditions.

Table 10.Pearson correlation matrix among physicochemical parameters.

Values represent Pearson correlation coefficients (r). All correlations are significant at p < 0.05.

According to Kaiser’s criterion (eigenvalues > 1), three principal components (F1-F3) were extracted, cumulatively explaining 79.29% of the total variance (Table 11).

Beyond the third component, eigenvalues dropped below unity, justifying the retention of three main factors. The component matrix (Table 12) details the loadings of each variable on these three factors:

1) Factor 1 (F1), explaining 54.06% of the total variance, was strongly associated with Cl⁻, HCO 3 − , SO 4 2 − , Na⁺, K⁺, Mg²⁺, Ca²⁺, and EC, and negatively correlated with NO 3 − . This factor represents natural mineralization and ion exchange processes, with minor anthropogenic input. The strong negative correlation between F1 and NO 3 − suggests a spatial separation between geogenic and anthropogenic influences: wells with high natural mineralization tend to exhibit low nitrate concentrations, whereas those with elevated nitrate levels are likely affected by human activities such as domestic wastewater infiltration and agricultural runoff. 2) Factor 2 (F2), accounting for 16.08% of the variance, showed high positive loadings for Mn²⁺, Fe²⁺, and Al³⁺, indicating metal mobilization from lateritic horizons under variable redox and pH conditions, and 3) Factor 3 (F3), explaining 9.15% of the variance, presented positive loadings for NH 4 + and Cd²⁺ and a negative loading for NO 3 − , corresponding to localized anthropogenic contamination from agricultural runoff and domestic wastewater.

Overall, the integration of results from Table 9, Table 10, Table 11, and Table 12 confirms that groundwater chemistry in the Bonoua aquifer is governed by three dominant processes: 1) natural mineral dissolution and ion exchange, 2) redox-controlled mobilization of metals, and 3) anthropogenic pollution linked to agricultural and domestic sources

Table 11. Total variance explained by principal component analysis (PCA).

Extraction method: Principal Component Analysis (PCA). Rotation not applied.

Table 12. Principal component matrix of hydrochemical parameters (PCA).

Extraction method: Principal Component Analysis (PCA). Rotation not applied.