1. Introduction
Burkina Faso is a landlocked country located at about 500 km from the sea (Atlantic Ocean towards the southwest). Most of the water resources come from rainfall, which generates runoff and groundwater recharge [
1]. However, the region has experienced a change in its rainfall pattern since the late 1960s. This change is proven by a period of declining cumulative and daily rainfall [
2]. During the 1960s, 1970s and 1980s, the annual number of rainy days decreased significantly in several localities in the region. This reduction averaged 14 days. Annual rainfall then increased over the period 1981–2015. During this period, the country recorded an average rainfall increase of about 31.7% in the Sahelian zone and 22.6% in the Sudano–Sahelian zone. Despite this general trend towards increased rainfall, drought has persisted since 2005 in some localities. This persistence reflects a significant spatial variability in regional rainfall [
2]. The decrease in rainfall, together with demographic pressure on the land use, have considerably affected groundwater recharge [
2]. Indeed, these natural and anthropogenic changes lead to an increasing depletion on groundwater recharge [
3]. This is coupled with overexploitation of groundwater (more than 60,000 boreholes in 2015) caused by high demand from various users [
1] and this is likely to contribute to the lowering of the water table and the disruption of its balance. In the Cascades district of Burkina Faso (
Figure 1), current water supply through groundwater abstraction is increasingly characterized by a high failure rate, sometimes reaching 20% [
4]. One of the reasons for this high failure rate is the poor knowledge capitalization [
5], especially on aquifer recharge, which determines the availability of groundwater resources over time. The Tamassari basin, which is the subject of this study, belongs to this region where groundwater is a primary resource for the socioeconomic development of the population (highly agricultural population) [
6]. Therefore, a better knowledge of the recharge in this area becomes an imperative for a better sustainable management of this groundwater resource.
A wide variety of approaches to assessing groundwater recharge exist in different contexts around the world, including hydraulic, isotopic, thermal and numerical methods [
7,
8]. However, the recharge values vary greatly from one method to another in the same area [
9]. Apart from the difference in values depending on the method used to assess recharge, some gaps are due to the fact that spatial variability is not properly assessed because of the low density of the measurement networks reaching 6 km for boreholes or piezometers and up to 50 km for rainfall stations. The spatial extent of the study areas linked to the size of the underlying aquifers requires intensive field exploration for an accurate characterization of the geomorphological variability (lithology, fractures and faults, soil). However, these patterns can greatly influence the potential recharge of groundwater. In this respect, remote sensing and GIS techniques have recently attracted the attention of many researchers [
10,
11,
12,
13]. In fact, approaches based on spatial analysis techniques improve recharge estimation and provide qualitative assessments of its spatial distribution [
14].
The methodology is essentially based on the description, classification and integration of factors influencing recharge [
10,
11,
12,
13,
14,
15,
16]. For example, in Côte d’Ivoire localized in west Africa, [
15] used remote sensing and GIS to identify potential recharge areas of fractured aquifers in the N’zo basin. The results indicate that the potential high recharge areas represent about 20% of the total area of the basin. In north Africa, authors of [
16] used the same approaches to map potential recharge areas in the Haouz plain of Morocco. They classified the study area into three descriptive levels of recharge with a range of values from 3.5 to 19%. Furthermore, the recharge potential map using remote sensing and GIS needs to be validated by field data. Thus, isotopic and hydrochemical methods and numerical tools, in particular the calibration of a hydrological model (piezometry, water balance) with available observations can be used as a validation approach [
15]. However, the application of numerical methods is very limited due to the existence of many uncertainties related to the conceptualization of the natural environment and the data used in the calibration [
17]. The hydrochemical method was applied by [
16] in Morocco to validate the recharge map produced from spatial techniques. The hydrochemical tracer used is chloride (Cl-) and the procedure is based on the link between the concentration of chlorides in groundwater and rainwater with annual rainfall. The result of the recharge rate obtained by this approach in the form of a map made it possible to assess with some precision the potential recharge areas of the Houz plain. In the Red Delta plain of Vietnam, the recharge potential map made by spatial techniques was verified by the analysis of the radioactive isotope of the water molecule, tritium (
3H) [
18]. The different recharge zones were delineated with good agreement with the direct estimation of groundwater recharge by radioactive analysis. The objective of this study is to determine the recharge potential of the aquifers in the Tamassari basin by spatial techniques and its validation using groundwater chloride (Cl
−) and tritium (
3H) contents. This approach should allow us to judge the relevance of the spatial modelling making use of remote sensing products and GIS integrating data in the assessment of recharge areas based on hydrochemical and isotopic data for the validation in the context of the Cascade region.
2. Study Area
With an area of 194 km
2, the Tamassari basin is located in the southwest of Burkina Faso in the Cascades district, about 515 km from the capital Ouagadougou (
Figure 1). It is located between longitudes 5°22′30″ W and 5°15′00″ W and latitudes 10°50′00″ N and 10°50′00″ N.
The Cascades district has a South Sudanese climate marked by two main seasons, a wet season (April to October) with rainfall that can exceed 1400 mm and a dry season (November to March). The average interannual temperatures are between 17 °C and 36 °C [
4]. However, it emerges that the rains are often violent and accompanied by devastating winds, making it difficult setting up the crops, causing soil erosion [
4]. Potential evapotranspiration values remain very high throughout the year. They are above 100 mm per month. The highest values are observed between February and March, when they reach 200 mm. The lowest values are in July, August and September, when potential evapotranspiration is compensated by rainfall [
19]. This situation shows that most of the rain that falls in the study area is consumed by evapotranspiration. In the district, the Fcover indicator shows a progressive degradation of vegetation cover during the period 2007–2014 [
20]. The Fcover or fraction of vegetation cover is an indicator that evaluates the proportion of vegetation covering a land surface. In fact, there has been a reduction of 3155 km
2 or 16.57% in the cover ratio class between 80 and 100, which constitutes 23.38% of the total area of the district (19,030 km
2). At the same time, there was an increase of 787 km
2 or 4.14% in the coverage class between 0 and 20, which represents 5.09% of the total area of the district (
Figure 2). Concerning the relief, the Tamassari basin has two (02) topographic units, namely the plateaus and the plains, which are crossed by important hydrographic networks. Generally, the dominant soils are ferruginous soils with little leaching and leaching on sandy, sandy clay and clayey materials. Ferralitic soils on sandy-clay are also found in the south and strips of poorly evolved soils on gravel in the central and northern parts of the basin. Hydromorphic soils are poorly represented in the north-eastern part of the basin [
21]. These soils are generally poor in fertilising elements, especially nitrogen and phosphorus, with a poor structure. Their physical and hydrodynamic properties are therefore unfavourable for infiltration and water retention [
22].
From a geological point of view, the study area belongs to the lower siliceous sandstone of Proterozoic age. These are sedimentary formations made up of conglomerate, more or less coarse sandstone with silty layers. These sediments belong to the Taoudéni sedimentary basin, which lies unconformably on the low-permeability basement rock. The domain constituted by the basement rock is characterised by andesites and gneisses dating from the Paleoproterozoic age. On the hydrogeological level, the Taoudéni sedimentary basin constitutes a primordial sandstone aquifer with an enormous capacity to renew its groundwater reserves thanks to the favourable climatological conditions of this region [
9,
23,
24]. However, the piezometric level between 1995 and 2007 showed a downward trend in all seasons despite a slight peak in 2000–2001 following a good rainy season [
25] (
Figure 3).
The groundwater flow in this sandstone aquifer is towards the basement rock aquifer [
23]. Concerning the basement rock, they form discontinuous aquifers, i.e., fractured aquifers. These fractured basement rock aquifers are sometimes underlying and associated with weathered zones. These weathered zones are 10 m to 80 m thickness and represent the upper part of the basement rock aquifer. The water table in this upper part may be between 5 and 30 m below the ground surface. For fractured basement rock aquifers, their thickness can reach 10 to 80 m and the water table can be between 20 and 60 m below the ground surface [
26]. Recharge by rainwater infiltration is generally low [
26]. The basin has only one main and temporary river which rises in the commune of Kankalaba, at an altitude of 600 m, and flows from north to south for 28.15 km before joining Loumana downstream at 250 m altitude (
Figure 4).
The economic activity is essentially based on the primary sector, particularly agriculture and livestock farming, which employs 91.7% [
27] of the active population and contributes with more than 48% to local wealth creation. However, the climatic hazards affecting this sector make it difficult to implement agricultural activities, including livestock farming [
4].
5. Discussion
Remote sensing and GIS techniques proved their efficiency in the assessment of the recharge potential of aquifers in the study area. Five (5) classes of recharge degree were revealed. The very low degree of recharge observed in the localities of Léra and Kangoura is the result of the horizontal concentration of water near the surface in the absence of well-defined channels [
48]. However, it is important to note that the degree of low recharge obtained is related to rainfall without the contribution of runoff and runoff concentrations. In fact, in artificial and natural water reservoirs (ponds), knowledge of the ability of the soil to recharge the water table is important. In the event of major floods, the height of the water concentrated in the reservoir should be assimilated to the local rainfall. This will advantageously increase the groundwater recharge potential compared to rainfall only as a source of water input. If this water level is, for example, 3 m high, a contribution to recharge of 3.5 times the rainfall (for an average rainfall of 850 mm) could be expected. Thus, in the Sahelian strip, water retention sites are par excellence to those that recharge two to three times the amount of rainfall if human exfiltration and evapotranspiration are average [
30,
49]. Low recharge areas are related to basement rocks (gneiss, andesite). Although these formations are intensely affected by the Eburnean orogeny [
24,
50], the high thickness of weathering and the type of soil (clayey) that develops on these rocks would considerably reduce rainwater recharge. This observation is in agreement with the work of [
14], showing that the clay texture of soils would prevent precipitation water from infiltrating. The average degree of recharge in soils and rock outcrops may be the result of a deficit in soil moisture in these land use units. Where the soil surface is indurated (rock, soil), runoff is high, and the soil surface takes time to become saturated. This means that the transfer of rainwater to depth is not fast enough. The degree of average to high recharge could be related to the geological nature of the sedimentary formations. These formations are made up of fine to medium quartzite to past conglomeratic in places. This would promote good recharge by percolation of precipitation water towards the water table. High recharge areas result from the rapid and deep recharge of rainwater through preferential pathways such as fractures or fissures and along major rivers. This fracture drainage phenomenon plays a fundamental role in water transport within fractured aquifers [
51].
Overall, the medium and high recharge zones characterize the sedimentary formations of the Tamassari basin (Taoudéni basin) while the very low and low recharge zones occupy the basement domain. This shows that the Taoudéni sedimentary basin has a recharge potential with respect to the basement rock. This could be justified by the fact that the sandstone aquifer system is both porous and fractured, unlike the basement rock aquifer, whose recharge depends strongly on the presence of open fractures. These results go in the same direction as the work of [
52], stipulating that the Taoudéni basin constitutes a powerful sandstone aquifer system. In addition, the temporal analysis of the piezometry over the entire Taoudéni basin revealed a general trend towards the rise in the piezometric level [
53]. This result further confirms the good recharge areas mapped in the Tamassari basin. Indeed, the high piezometric levels in unconfined aquifers are due to the recharge of the water arriving there during rainfall events [
30].
The good correlation between tritium contents and the different recharge areas shows the importance of using tritium to study groundwater recharge. Indeed, the presence of tritium in water indicates a current recharge. This isotope has been used efficiently in many studies to characterise recharge. We can mention the work of [
9] in 2003 in the western part of Burkina Faso, where tritium contents were highlighted to characterise groundwater recharge. The results indicated three types of water: a first group of waters with a tritium content < 2 TU, i.e., old; a second group of waters with a content of 2 to 4 TU corresponding to a mixture of old and young waters and a third group of recent waters with a content of more than 4 TU. Similar studies were conducted by authors of [
43] in 2012 to investigate the groundwater of the Dargol basin (Liptako–Niger). The results also showed several types of water. These results show that tritium is an excellent indicator of recent groundwater.
In contrast to chloride contents, there is no correlation with recharge areas. This could be related to the thick weathering and the type of soil (clay) that develops in the study area. Indeed, the fine clay layers retain the chlorides coming from the rainwater in the unsaturated area, thus preventing them from entering the groundwater. This would explain the fact that most boreholes show chloride contents below the detection limit (0.02 mg/L). Therefore, the use of chlorides for recharge studies is limited in the study area.
Thus, despite the fact that the number of boreholes does not cover the study area homogeneously, this study shows an overall trend in recharge, which is low for the basement formations and high for most of the sedimentary zone. This can be explained on the one hand by the favourable lithology (high permeability) of the sedimentary formations and on the other hand by the high density of fracturing observed in this sector, which is therefore favourable to infiltration.
This study has shown that remote sensing and GIS have enormous advantages in the spatialization of potential recharge areas in the Tamassari basin. Indeed, the use of a recent satellite image (Landsat 8 OLI of the year 2021) for land use mapping and corresponding to the date of groundwater abstraction, allowed an easy update of the recharge map. Furthermore, this study revealed that land use is the second most important factor (W = 0.298) in determining recharge areas. As for the significant influence of land use on the recharge, which has been the subject of certain studies [
54,
55], it must be approached according to a temporal approach covering the last 10 years [
55] and based on remote sensing to produce land cover classes and produce recharge map with varying land use patterns.
Despite this importance, this factor is conditioned by the type of soil, which itself is dependent on the nature of the underlying rock. Moreover, the spatial database developed in the framework of this study can be used in the modelling of aquifers such as the hydrogeological model of groundwater flow in sedimentary aquifers using the HUF module of Modflow2000 [
56] or any other model that can be interfaced with a GIS for a better follow-up by scientists as well as managers.
Mapping the potential recharge of aquifers using remote sensing and GIS techniques in Tamassari basin reveals some limits. The main difficulty lies in the definition of class limits and weights, which are assigned to the various factors entering into the realization of the GIS [
15]. The choice of class limits is most often made according to the operator’s ability to discern and his sense of judgment [
57]. As is the case with methods based on the subjective appreciation of class limits, the final result of recharge mapping can be greatly influenced by the choices of these limits [
12]. Therefore, a sensitivity study is required to determine the limits of the method and the influence of the choice of class limits on the result. The incorporation of probabilistic uncertainty into the AHP technique is considered a powerful approach to quantifying the sensitivity and uncertainty of the proposed model. The beta-PERT distribution has been widely used to model expert judgements and provide a close fit to normal distributions with little data [
58,
59]. This technique uses the most likely, minimum and maximum values of expert estimates to generate a probability distribution that measures the level of confidence in the AHP decision [
60]. Moreover, the use of certain types of remote sensing data such as lineaments, which are considered as expressions of fractures and geological accidents is much controversial in the literature [
15]. The number of lineaments observed generally depends on the methods of interpretation and the experience of the operator [
15]. However, in our case, there is a good correlation between the lineaments and the geological features in the field.
6. Conclusions
The objective of this study was to determine the recharge potential of the aquifers of the Tamassari basin by spatial techniques and its validation using chloride (Cl−) and tritium (3H) contents of groundwater. The methodological approach allowed reliable results to be obtained. Indeed, the validation of the thematic maps produced in this study by the field work ensured the reliability of the model input parameters. Similarly, the AHP method used allowed the relative importance of the factors to be assessed in relation to each other, leading to the determination of consistent weights for each factor. Moreover, the use of a recent satellite image (Landsat 8 OLI of the year 2021) for land use mapping and corresponding to the date of groundwater sampling, allowed an easy update of the recharge map. The novelty of this study is the use of a correlation matrix between the recharge areas identified by remote sensing and GIS techniques and the known tritium and chloride water contents of existing boreholes in the basin.
Thus, the recharge potential map obtained shows a generally good recharge trend for the sedimentary formations. The areas with low recharge are located in the basement zone. This map also allows for the identification of areas vulnerable to anthropogenic pollution, and therefore constitutes a decision support tool that should be made available to decisionmakers for a better spatiotemporal optimisation of the quantity and quality of groundwater resources for the wellbeing of the surrounding population of the Tamassari basin. Furthermore, this study has shown the importance of using isotopes, especially tritium, to identify recharge areas in the region. In terms of perspectives, these results could be refined by a piezometric study by installing piezometers on the different recharge zones identified by the model.