Hydrogeological characterization of fissured aquifers in the crystalline basement zone of Dimbokro, Côte d'Ivoire: integrated approach by remote sensing and statistical analysis

 Corresponding author: Kouakou Noumh Dickens Atchérémi Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution License 4.0. Hydrogeological characterization of fissured aquifers in the crystalline basement zone of Dimbokro, Côte d'Ivoire: integrated approach by remote sensing and statistical analysis Kouakou Noumh Dickens Atchérémi 1, *, Zahibo Oscar Onétié 2, Yao Muller YAO 3, Mahaman Bachir SALEY 4, Kan Jean KOUAME 4 and Patrice Jean Roger JOURDA 4 1 Department of Coastal Sciences, Marine Sciences Training and Research Unit, University of San Pedro, BP 1800 San Pedro, Côte d'Ivoire. 2 Department of Geosciences, Biological Sciences Training and Research Unit, Peleforo Gon Coulibaly University, Korhogo, Côte d'Ivoire. 3 Department of Geospatial Science and Technology, UFR Computer Science and Digital Sciences, Virtual University of Côte d'Ivoire. 4 University Center for Research and Application in Remote Sensing (CURAT), Félix Houphouët-Boigny University, Abidjan, Côte d'Ivoire. GSC Advanced Research and Reviews, 2025, 25(01), 172-183 Publication history: Received on 11 September 2025; revised on 22 October 2025; accepted on 25 October 2025 Article DOI: https://doi.org/10.30574/gscarr.2025.25.1.0314 Abstract Water availability in the crystalline basement formations of Dimbokro, Côte d'Ivoire, is limited by low rock porosity and the complexity of fissured aquifers. The present study aims to characterize the underlying surface weathering and fissure aquifers through an integrated approach combining hydrogeology, multispectral remote sensing and geostatistical analysis. The data comes from 110 boreholes and 47 field stations, coupled with the analysis of lineaments extracted from Landsat 7 ETM+ images. The results reveal a strong heterogeneity of alteration thicknesses (30-60 m optimal for productivity) and average flow rates (3.42-5.28 m³/h), while satellite imagery can detect the dominant fracturing directions (N0-10° and N90-100°) more efficiently compared to aerial photography. The analysis of fracturing-induced permeability shows a trimodal spatial distribution, confirming the importance of major fractures in hydraulic connectivity. These results provide an operational framework for the optimized drilling and sustainable management of fissured aquifers in crystalline basement. Keywords: Basement Aquifers; Remote Sensing; Induced Permeability; Statistical Analysis; Dimbokro 1. Introduction The availability and management of groundwater is a major challenge for socio-economic development in Côte d'Ivoire, particularly in rural areas resting on crystalline foundations. These geological formations, covering nearly 97% of the territory, have low porosity and limited water storage, which makes the exploitation of fissured aquifers complex and prone to a high rate of borehole failure [1,2]. Increasing urbanization, agricultural pressure and changing climatic conditions are changing the recharge and distribution of water resources, increasing the vulnerability of local communities [3]. Fissured aquifers in the crystalline basement are controlled by fracture density, orientation, and connectivity, parameters that determine the transmissivity and permeability of the formations [4,5]. Traditional methods of GSC Advanced Research and Reviews, 2025, 25(01), 172-183 173 prospecting and hydraulic planning remain limited in the face of the complexity of these fracture networks, and do not always allow to identify preferential flow corridors or to spatially characterize hydrogeological properties. In this context, integrated approaches combining field data, multispectral remote sensing and statistical analysis appear to be robust tools to improve the understanding of fissured aquifers. The structural analysis of lineaments from Landsat 8 images, combined with the statistical analysis of hydrogeological parameters (weathering thickness, transmissivity, permeability) and integration into a GIS, makes it possible to produce synthetic thematic maps on the availability, accessibility and exploitability of resources [6]. The present study thus aims to quantitatively characterize the fissured aquifers in crystalline basement at Dimbokro, distinguishing the superficial alteration aquifer from the underlying fissure aquifer, in order to propose an operational methodological framework for the selection of drilling sites. It contributes to the sustainable management of groundwater resources and local water planning, while strengthening the understanding of hydrogeological dynamics in vulnerable rural areas. 2. Materials and methods 2.1. Study area The study area (3°30'-4°30' W; 6°30'-8°00' N), covers about 4,200 km², is located in the forest-savannah transition zone (central-eastern Côte d'Ivoire) (Figure 1). It offers sufficient lithological and demographic homogeneity to analyze fissured aquifers in crystalline basement, in a representative framework allowing the development of transposable models and the optimization of catchment infrastructures. Figure 1 Presentation of the study area 2.2. Materials One hundred and ten (110) boreholes drilled over three periods (33 in 1990, 42 in 2000, 32 in 2010) made it possible to quantify hydrogeological, cartographic and satellite data in an integrated approach and to establish regional statistical profiles. The geological and topographic maps at 1:200,000 provided a structural framework for interpreting lineaments and correlating formations and boreholes. Landsat 7 ETM+ scenes facilitated manual lineament extraction GSC Advanced Research and Reviews, 2025, 25(01), 172-183 174 and fine spatial analysis. The numerical processing was carried out with MapInfo 12, ArcGIS 10.8, Envi 5.3 and the R Software version 4.4.3 and the synthesis and visualization with Excel and Paint. 2.3. Methods 2.3.1. Conceptual framework The study adopts an integrated approach combining hydrogeology, remote sensing, statistical analysis and geomatics to characterize fissured aquifers according to the bi-layer model ([7]): surface alteration aquifer and underlying fissure aquifer. 2.3.2. Approach strategy The approach is based on a statistical analysis of boreholes [8] and a structural analysis by multispectral remote sensing [9], as well as an assessment of fracture-induced permeabilities [10]. 2.3.3. Hydrogeological data acquisition and processing The 110 holes, selected by updated stratified sampling, provided UTM coordinates, depth, alteration thickness, normalized flows, piezometric levels and lithology, validated by pumping tests, Dixon test, electrical soundings. The structural analysis mobilized Landsat 7 ETM+ satellite images in the dry season (< 5% clouds), pre-processed (atmospheric correction, calibration, georeferencing), analyzed by PCR and enhancement (histogram equalization, directional filtering, Fourier transform) for lineament extraction (>500 m, sinuosity < 1.2) and for structural uniformity [11,12]. 2.3.4. Determination of fracture-induced permeabilities The determination of fracture-induced permeabilities is based on 47 field stations, with measurements of orientation (±2°), aperture (±0.1 mm), surface finish, filling [13] and fracture density [14]. Equivalent permeabilities were calculated according to [15], integrating structural anisotropy via a directional model weighting orientations and openings [16], then validated by correlation with pumping tests and analysis of the R² determination coefficient [17] (α = 0.05). 2.3.5. Statistical analysis and data processing Statistical parameters (mean, median, mode, standard deviation, CV, interquartile range) were calculated for all variables, with Shapiro-Wilk (No. < 50) and Kolmogorov-Smirnov (No. ≥ 50) normality tests. Log-normal variables were transformed and outliers were processed according to Tukey. Missing data (< 5%) were imputed by MICE and uncertainties (instrumentation, interpolation, sampling) propagated via Monte Carlo simulations (n = 1000, [18]). 3. Results 3.1. Hydrogeological characteristics of fissured aquifers 3.1.1. Relationship between alteration thickness and total depth of boreholes The observation in Figure 2 reveals a dispersion of the data and a strong heterogeneity in the first 40 m which have low flows (< 3 m³/h), creating an area of uncertainty. The 30-60 m interval is an optimal compromise between storage and permeability, while above 60 m, compaction and porosity reduction limit productivity (~1.5 m³/h). Thus, for the installation of boreholes, it is therefore necessary to favor alteration zones of 30-60 m. GSC Advanced Research and Reviews, 2025, 25(01), 172-183 175 Figure 2 Distribution of flow rates according to weathering thicknesses. Tables 1 and 2 below show the gradual increase in the average depth of drilling and the thickness of the altered zones. Table 1 Decadal variation in depth (m) of boreholes Parameter Average ± SD ARE 95% CI Min Max CV (%) Depth 1990 47.23 ± 17.53 2,96 [41,21; 53,25] 15 88,4 37,1 Depth 2000 52,67 ± 18,66 2,88 [46,85; 58,48] 15 100,4 35,4 Depth 2010 58,46 ± 20,24 3,52 [51,28; 65,64] 15 103,2 34,6 Table 2 Decadal variation in the thickness (m) of altered zones Parameter Average ± SD IS IC 95 % Min Max CV (%) Thickness 1990 25,86 ± 10,30 1,74 [22,33; 29,40] 5,0 51,2 39,8 Thickness 2000 27,95 ± 11,94 1,84 [24,23; 31,67] 5,0 50,2 42,7 Thickness 2010 29,40 ± 14,13 2,46 [24,39; 34,41] 5,7 60,2 48,1 SD = Standard deviation, ES = Standard error, CI = Confidence interval, CV = Coefficient of variation Notes: The data present the evolution of average flow, borehole depth and altered zones over three key periods, as well as the success rate of catchments, reflecting the regional variability of fissured aquifers. 3.2. Relationship between average flow rates and borehole success rates The additional observation in Table 3 highlights a decrease in the success rate of drilling (85.7 → 60.6%). Despite this difficulty in identifying the most productive areas, average flows are increasing (Table 4), although their dispersion remains significant. Table 3 Decadal variation in success rates and average borehole flows Parameter Average ± SD IS IC 95 % Min Max CV (%) Success 1990 85.70 ± 5.90 5.90 [74.10; 97.30] – – 6.9 Success 2000 78.60 ± 6.30 6.30 [66.20; 91.00] – – 8.1 Success 2010 60.60 ± 8.50 8.50 [43.90; 77.30] – – 14.0 GSC Advanced Research and Reviews, 2025, 25(01), 172-183 176 Table 4 Decadal variation in borehole flow (m³/h) Parameter Average ± SD IS IC 95% Min Max CV (%) Flow rate (m³/h) 1990 3,42 ± 2,20 0,37 [2,66; 4,18] - - 64,3 Flow rate (m³/h) 2000 4,24 ± 2,17 0,33 [3,56; 4,92] - - 51,2 Flow rate (m³/h) 2010 5,28 ± 3,23 0,56 [4,14; 6,42] - - 61,2 Notes: The data present the evolution of average flow, borehole depth and altered zones over three key periods, as well as the success rate of catchments, reflecting the regional variability of fissured aquifers. 3.3. Remote sensing analysis and structural interpretation 3.3.1. Structuring by principal component analysis and geological structures by satellite imagery The Bartlett sphericity test (P < 0.000) justifies the PCA, whose first three components concentrate 99.93% of the information. PC1 (99.91%) captures most of the radiometric information and highlights the NW–SE fractures, illustrating the major geological structuring (Figure 3A). PC2 and PC3, explaining 0.45% and 0.17% of the variance, reveal subtle contrasts: PC2 highlights bare soils and inhabited sites via the near-infrared (Figure 3B), and PC3 allows the fine discrimination of lineaments by albedo variations (Figure 3C). 3.4. Correlation between pedo-qualitative indicators through principal component analysis 3.4.1. Statistical characterization of fracturing directions Table 5 shows a predominance of the N0-10° (13.8%) and N90-100° (14.7%) directions among the detected lineaments. Satellite imagery exceeds aerial photography (+37%) for the detection of dominant structures, with low coefficients of variation (21.7-22.6%), attesting to the stability of directional measurements. Figure 3 Three main components of the PCA highlighting (A) the NW-SE fractures, (B) the bare soils and inhabited sites, and (C) the NW-SE linear fractures, with N'dolioko, N'guessankro and Hiandian geographical landmarks GSC Advanced Research and Reviews, 2025, 25(01), 172-183 177 Table 5 Descriptive statistics of fracturing directions by data source Parameter Average ± SD IS IC 95 % Min Max CV (%) Aerial photography Direction N0-10° (%) 11,6 ± 2,8 0,4 [10,8; 12,4] 8,2 15,3 24,1 Direction N80-90° (%) 8,3 ± 1,9 0,3 [7,7; 8,9] 5,4 11,2 22,9 Direction N90-100° (%) 12,3 ± 3,1 0,5 [11,3; 13,3] 7,8 17,1 25,2 Other Directorates (%) 10,7 ± 2,4 0,4 [9,9; 11,5] 6,8 14,9 22,4 Satellite imagery Direction N0-10° (%) 15,7 ± 3,4 0,5 [14,7; 16,7] 10,2 21,8 21,7 Direction N80-90° (%) 11,7 ± 2,6 0,4 [10,9; 12,5] 7,9 16,4 22,2 Direction N90-100° (%) 16,8 ± 3,8 0,6 [15,6; 18,0] 11,3 23,7 22,6 Other Directorates (%) 12,9 ± 2,9 0,4 [12,1; 13,7] 8,6 18,2 22,5 SD = Standard deviation, ES = Standard error, CI = Confidence interval, CV = Coefficient of variation, n = Sample size. All statistical analyses were performed with a significance level α = 0.05 3.5. Modelling the density-length relationship of fractures The linear regression model (Y = 1.57X + 6.92, R² = 0.944 ± 0.012) highlights a stable predictive relationship between the geometric parameters of fracturing, confirmed by a low variability of the parameters (CV < 2% for the slope), underlining the robustness of the spatial distribution of fractures (Table 6). Table 6 Fracture Length-Number Correlation Model Parameters Parameter Value ± AND IS IC 95 % Min Max CV (%) Slope (β₁) 1,57 ± 0,023 0,003 [1,52; 1,61] 1,48 1,65 1,5 Ordinate (β₀) 6,92 ± 1,46 0,21 [4,07; 9,78] 3,82 10,15 21,1 Coefficient R² 0,944 ± 0,012 0,002 [0,932; 0,955] 0,918 0,967 1,3 Standard error 7,89 ± 1,23 0,17 [7,55; 8,23] 5,67 10,45 15,6 Cumulative length (km) 42,5 ± 22,8 1,02 [40,5; 44,5] 5,0 80,0 53,6 Number of fractures 73,6 ± 38,4 1,72 [70,2; 77,0] 15,0 132,0 52,2 SD = Standard deviation, ES = Standard error, CI = Confidence interval, CV = Coefficient of variation, n = Sample size. All statistical analyses were performed with a significance level α = 0.05 3.6. Benchmarking remote sensing performance The comparative analysis (Table 7) shows that satellite imagery outperforms aerial photography in the characterization of fracturing networks at the regional scale, with a detection rate of the dominant directions of 44.2% compared to 32.2% (+37%) and a higher detection density (19.7 vs. 14.8 lineaments/km²), despite a slightly lower directional accuracy (3.1° vs. 2.3°) GSC Advanced Research and Reviews, 2025, 25(01), 172-183 178 Table 7 Benchmarking of performance by remote sensing method Parameter Average ± SD IS IC 95 % Min Max CV (%) Aerial Photography Detection Detection rate of dominant directions (%) 32,2 ± 5,4 0,8 [30,6; 33,8] 22,1 42,6 16,8 Density lineaments (nb/km²) 14,8 ± 3,7 0,5 [13,8; 15,8] 8,9 21,4 25,0 Directional Accuracy (°) 2,3 ± 0,6 0,08 [2,1; 2,5] 1,2 3,8 26,1 Satellite imagery detection Detection rate of dominant directions (%) 44,2 ± 6,8 0,9 [42,4; 46,0] 31,5 57,3 15,4 Density lineaments (nb/km²) 19,7 ± 4,9 0,7 [18,3; 21,1] 11,2 28,6 24,9 Directional Accuracy (°) 3,1 ± 0,8 0,11 [2,9; 3,3] 1,8 4,7 25,8 SD = Standard deviation, ES = Standard error, CI = Confidence interval, CV = Coefficient of variation, n = Sample size. All statistical analyses were performed with a significance level α = 0.05 3.7. Validation of geostructural hypotheses The observation in Table 8 statistically confirms the relevance of the hypotheses formulated on the basis of regional geological knowledge and previous field observations. All hypotheses formulated are statistically validated with confidence levels above 99%. The confirmation of the predominance of the N0-10° and N90-100° directions (28.5% ± 4.2%) and the robust correlation model (R² > 0.8) validates the methodological approach adopted. The high regional coherence index (0.92 ± 0.06) attests to the geological representativeness of the results obtained. Table 8 Statistical validation of geostructural hypotheses Parameter Observedvalue± AND IS IC 95 % Threshold p-value Dominant directions N0-10° + N90100° (%) 28,5 ± 4,2 0,6 [27,3; 29,7] >25,0 <0.001 R² 0,944 ± 0,012 0,002 [0,932; 0,955] >0,800 <0.001 Cross-source correlation 0,89 ± 0,04 0,006 [0,87; 0,91] >0,750 <0.001 Secondary direction N80-90° (%) 10,1 ± 2,1 0,3 [9,5; 10,7] 5,0-15,0 <0.01 Regional coherence (index) 0,92 ± 0,06 0,008 [0,90; 0,94] >0,800 <0.001 3.8. Temporal evolution of fracturing detection The diachronic analysis (Table 9) shows a gradual improvement in detection capabilities thanks to technological advances in remote sensing sensors, with a detection density increasing from 47.23 ± 17.53 in 1990 to 58.46 ± 20.24 in 2010 (+23.8%) and a coefficient of variation decreasing from 37.1% to 34.6%, reflecting better homogeneity. The tenyear improvement rate of 11.8% ± 2.9% illustrates the steady progress in space imagery. Table 9 Data quality settings by acquisition period Parameter Average ± SD IS IC 95 % Min Max CV (%) Fracking density 1990 47,23 ± 17,53 2,96 [41,21; 53,25] 15,0 88,4 37,1 Fracking density 2000 52,67 ± 18,66 2,88 [46,85; 58,48] 15,0 100,4 35,4 Fracturing density 2010 58,46 ± 20,24 3,52 [51,28; 65,64] 15,0 103,2 34,6 Ten-year evolution 5,62 ± 1,36 0,28 [5,05; 6,19] 2,1 8,9 24,2 Rate of improvement (%) 11,8 ± 2,9 0,6 [10,6; 13,0] 7,2 17,4 24,6 SD = Standard deviation, ES = Standard error, CI = Confidence interval, CV = Coefficient of variation, n = Sample size All statistical analyses were performed with a significance level α = 0.05 GSC Advanced Research and Reviews, 2025, 25(01), 172-183 179 3.9. Fracturing-induced permeability: quantification of hydrodynamic efficiency 3.9.1. Statistical characterization of hydrodynamic properties The spatial distribution of permeability (Table 10) reveals the intrinsic heterogeneity of fractured environments and quantifies the hydraulic efficiency of discontinuity networks identified by remote sensing. Table 10 Descriptive statistics of the permeability of fissured aquifers (K, m/d) Parameter Average ± SD IS IC 95% Min Max CV (%) Kmax (×10⁻⁴ m/s) 8,74 ± 3,67 0,54 [7,69, 9,79] 1,50 19,7 42,0 Kmin (×10⁻⁴ m/s) 5,18 ± 2,48 0,36 [4,47, 5,89] 1,13 14,5 47,9 Kmoy (×10⁻⁴ m/s) 6,92 ± 2,80 0,41 [6,12, 7,72] 1,31 17,1 40,5 Mini and Max correspond to the minimum and maximum values observed, respectively. The standard error (SE) is calculated according to ES = ET/√n, with n = 47 stations. The 95% confidence interval can be estimated by mean ± 1.96 × ES. The coefficient of variation (CV) is calculated as (SD/Mean) × 100. The values are expressed in the original units of each parameter. 3.10. Distribution of permeability classes and hydrogeological zonation The frequency analysis of the permeabilities reveals a trimodal stratification defining hydraulically contrasting domains: the high values (Kmax: 8-12 × 10⁻⁴ m/s) correspond to the preferential flows along the major fractures (N010° and N90-100°), the low values (Kmin: 3-6 × 10⁻⁴ m/s) to the basal drainage zones, and the intermediate values (Kmoy: 4-8 × 10⁻⁴ m/s) reflect the effective regional permeability. This hierarchical organization underlines the role of major fractures in flow dynamics and that of secondary fracturing in regional hydraulic connectivity. 3.11. Temporal evolution of hydrodynamic performance The evolution of specific flows between 1990 and 2010 shows a significant improvement in hydraulic performance (p < 0.001). The median decreased from 2.1 m³/h in 1990 to 2.8 m³/h in 2000 and then to 3.6 m³/h in 2010, with a gradual reduction in interquartile variability. Post-hoc tests (Tukey) confirm marked differences between decades (1990b ≠ 2000ab ≠ 2010a). This increase reflects the positive impact of the optimization of implantation techniques, based on the prior structural analysis of fracturing (Figures 4 and 5). GSC Advanced Research and Reviews, 2025, 25(01), 172-183 180 Figure 4 Distribution of induced permeabilities Figure 5 Evolution of operating flows (1990-2010) a ab b