This study used a high-resolution terrestrial gravity field model SGG-UGM-2, as the Bouguer gravity anomaly. Combining observations from the Gravity Field and Steady-State Ocean Circulation Explorer (GOCE), the Gravity Recovery and Climate Experiment (GRACE) normal equation, marine gravity data derived from satellite altimetry, and continental gravity data derived from EGM2008, this model is built up to degree 2190 and order 2159. It is derived from the theory behind the Ellipsoidal Harmonic Analysis and Coefficient Transformation (EHA-CT) method for calculating spherical harmonic coefficients from averaged grid areas and point gravity anomalies on the ellipsoid ( Liang et al., 2020 ; Liang et al., 2018 ). The Bouguer anomaly map of the SGG-UGM-2 model is shown in Figure 3 , reflecting the effect of all density heterogeneities beneath the topographic surface.

Various methods were applied to the Bouguer Anomaly of the SGG-UGM-2 model, in order to highlight the main structural features and the delineation of the various characteristic lineaments. The Bouguer anomaly is caused by geological structures of different dimensions and/or located at different depths. In an attempt to isolate the long-wave phenomena responsible for the regional anomaly, we perform a regional/residual separation using a polynomial fit ( Radhakrishna Murthy & Krishnamacharyulu, 1990 ). This procedure highlights the role of shallow geological structures.

To better exploit these data and bring out new elements in relation to the information provided by the residual map, we applied a series of mathematical treatments. These involve filtering the gravimetric signal in the frequency domain. The vertical gradient was applied to the residual map to highlight shallow lithologies. It is obtained using the following expression:

$\frac{\partial F}{\partial z}$(1)

where F is the gravity field.

According to Dobrin & Savit (1988) and Telford et al. (1990) , this transformation should attenuate, if not eliminate, the regional component that distorts and sometimes masks the relationship between shallow basement geology and anomaly shape. The vertical gradient acts as an amplifier for high frequencies, i.e. for anomalies of small extension, at least in one direction.

According to Nabighian (1984) , the analytical signal is the three-dimensional (3D) vector, where the absolute value of this signal is defined as the square root of the sum of the vertical and the two horizontal derivatives of the magnetic field. If F is the gravimetric field, then the absolute value of this signal is given by:

$AS=\sqrt{{\left(\frac{\partial F}{\partial x}\right)}^2+{\left(\frac{\partial F}{\partial y}\right)}^2+{\left(\frac{\partial F}{\partial z}\right)}^2}$(2)

where $\frac{\partial F}{\partial x},\frac{\partial F}{\partial y}$ and $\frac{\partial F}{\partial z}$ represent the gravity data gradients F.

Calculating the horizontal gradient from the residual anomaly is very useful for locating geological contacts, as the boundary between two blocks of different densities corresponds to the maximum of this gradient ( Blakely, 1995 ; Blakely & Simpson, 1986 ; Cordell & Grauch, 1985 ; Khattach et al., 2004 ). The THG formula is given by Cordell & Grauch (1985) :

$THG=\sqrt{{\left(\frac{\partial F}{\partial x}\right)}^2+{\left(\frac{\partial F}{\partial y}\right)}^2}$(3)

where $\frac{\partial F}{\partial x}$ and $\frac{\partial F}{\partial y}$ represent the gravity data gradients F.

Coupled with upward continuation, the horizontal gradient enables faults to be located and their dip determined ( Archibald et al., 1999 ; Khattach et al., 2004 ). Linear contacts correspond to faults, while circular contacts are the limits of diapirs or intrusive bodies.

The Euler deconvolution method, applied to potential field data, enables us not only to locate contacts in the horizontal plane, but also to estimate their depth ( Nabighian & Hansen, 2001 ). It is based on a mathematical procedure represented by the Euler homogeneity equation ( Thompson, 1982 ):

$\frac{\partial M}{\partial x}\left(x-x_0\right)+\frac{\partial M}{\partial y}\left(y-y_0\right)+\frac{\partial M}{\partial z}\left(z-z_0\right)=SI\left(B-M\right)$(4)

where (x₀, y₀, z₀) is the position of the source the effect is detected in (x, y, z).

Thompson (1982) and Reid et al. (1990) point out that the choice of structural index appears to be very important; for a number of structures, they have established a structural index (N) that can take values from 0 to 3, corresponding to whole numbers for certain simple structures. Thus, they consider that an N = 1 index is best suited for thin veins, dykes and faults with low vertical rejection, and an N = 0 index for faults with high rejection, and an N = 0.5 index for intermediate cases.

Analysis of the Bouguer map in the study area enables us to identify the main features of the anomalies observed and to establish the links between these anomalies and the geological and tectonic contexts of the region. The Bouguer anomaly map ( Figure 3 ) shows the existence of several anomalies of different shapes and signs, with values ranging from −82 to −6 mGal. These different gravity anomalies are separated by zones of strong gradients, indicating the presence of density discontinuities, generally materializing tectonic faults. A visual analysis of this map ( Figure 3 ) shows that the study area is characterized by elongated anomalies following dominant directions SSW-NNE to North-South (N-S) and secondary directions South-West to North-East (SW-NE), East to West (E-W), South-East to North-West (SE-NW) to South-South-East to North-North-West (SSE-NNW). Figure 3 shows, on the one hand, areas with dense anomaly sources (above-average anomalies of −42 mGal), and on the other, areas with less dense anomaly sources (below-average anomalies). Superimposing these anomalies on the major geological and structural features of the study area reveals the lack of correlation between the various gravity anomalies and the surface geology, suggesting structural complexity.

Implementing the separation technique introduced by Radhakrishna Murthy and Krishnamacharyulu (1990) , the residual anomaly map ( Figure 4 ) is derived from the Bouguer anomalies by subtraction of a second-order regional polynomial surface. The selection of the regional second-order surface was based on a careful examination of terrestrial gravity data, with the aim of effectively distinguishing deep and shallow structural anomalies in the South-Western region of Cameroon.

The residual anomaly map ( Figure 4 ) highlights anomalies related to shallower crustal structures than most of those observed on the Bouguer map. From a general point of view, we note that the residual anomaly map ( Figure 4 ) shows the same characteristics as those of the Bouguer anomaly. Like the Bouguer anomaly map, the residual anomaly map shows areas of positive and negative anomalies, separated by zones of significant gradients. The map shows anomalies with amplitudes ranging from −28 to +45 mGal. From west to east, the study area is characterized by:

The residual anomaly map shows several zones of strong gravimetric gradients that may correspond to contacts or discontinuities (faults, flexures, etc.). To study these zones of particular interest for structural characterization of the study area, we applied the analysis methods described above.

The purpose of the first vertical derivative is to highlight anomalies associated with shallow structures at the expense of those associated with deep structures. Illustrated in this study by Figure 5 , the first derivative map shows a lateral separation of anomalies and an amplification of the gravimetric effect of superficial

density contrasts to the detriment of deep density contrasts. A phenomenon of decoalescence can be observed here, with certain long-wavelength anomalies identified on the residual anomaly map being decomposed into two or more other short-wavelength anomalies. This phenomenon is clearly visible at the western edge of the study area, where the large anomaly that used to occupy this zone has broken up into short-wavelength anomalies of varying shape and size. A similar phenomenon was observed in the Ebolowa zone, at the western edge of the study area. The large negative anomaly that was present in this area has disappeared to the detriment of new anomalies of more restricted extension associated with superficial sources, which clearly confirms that it is of deep-seated origin. The new anomalies highlighted on the first derivative map are probably due to the presence of tectonic faults, intrusions and very significant superficial geological contacts in the subsurface of the study area. The presence on this map of the vertical gradient oriented ENE-SSW observed between the localities of Kribi and Edea could materialize the Kribi-Campo Fault observed on the geological map ( Figure 2 ).

The horizontal gradient map of the Bouguer residual anomalies shown here in Figure 6 highlights zones of horizontal gradients of varying amplitude, shape

and orientation. The diversity of anomalies observed on the map reflects the complexity of the geological structures in the area under investigation. Gradient zones can be easily distinguished in areas such as Aka, Bipindi, Kribi, Ebolowa, Eseka and Ndokbassaben. The trends of these high-amplitude horizontal gradients are approximately NNE, NNW to NS in some places. These gradients probably correspond to contact/fault-type structures (high-gradient zone) and intrusive structures. The presence of these lineaments, although widely spaced due to the resolution limitations of the SGUGM-2 global gravity model data, serves as evidence that the subsurface of the study area is significantly influenced by geodynamic phenomena.

To highlight the geological contacts associated with the faults or fractures suspected on the previous maps, we applied the multi-scale horizontal gradient analysis method described above to the Bouguer residual anomalies. The map below ( Figure 7 ) is an overlay of the horizontal gradient maxima maps of the

Bouguer residual anomalies extended upwards at 2 km, 4 km, 6 km, 8 km and 10 km. These maxima highlight areas of abrupt density change, interpreted either by faults or geological contact, as intrusive information ( Jilali & Khattach, 2023 ; Souga Kassia et al., 2020 ). The behavior and directions of superimposed maxima provide information on the orientation and dip of localized lineaments ( Figure 7 ). The degree of importance (at depth) of a fault is determined by the persistence of the presence of local maxima at increasingly higher extension altitudes. In the context of this study, it would appear at first glance that the maxima of the various extension levels are superimposed over almost the entire map, which implies that most of the structures in the study area are vertical structures, highlighting (1) linear contacts corresponding to faults or (2) circular contacts corresponding to the horizontal contours of the boundaries of intrusive bodies or diapirs. These include:

The map of local maxima of horizontal gradients calculated at different altitudes ( Figure 7 ) has enabled us to construct a synthetic structural map highlighting accidents (faults numbered 1 to 17) and the contours of intrusive formations in the bedrock (1 to 4) ( Figure 8 ). This map is a valuable tool for hydrogeological and mining research, providing essential information on flow directions, drainage axes and recharge zones. Statistical analysis reveals the presence of 21 gravitational structural features, ranging in size from 0.21 km to 62.81 km, with an average of 17.12 km. The rosette of mapped fracture directions highlights the main preferential directions: NW-SE, NNE-SSW, ENE-WSW and secondary directions: NNW-SSE, NE-SW, NS and E-W. This interpretation validates the interpretations already made on the residual and derived maps of the study area. These different structural directions all belong to the Pan-African domain. The deep basement faults observed in this area concur with the findings of Shandini et al. (2010) and Basseka et al. (2011) , who confirmed the CC/NEFB collision and identified NE-SW and NW-SE lineaments within the area. The ENE-WSW orientation, characteristic of Pan-African tectonics on a regional scale, has also been noted in the study area by Feumoe et al. (2012) and Noutchogwe et al. (2010) . This is also the direction of major Pan-African structures in Cameroon, such as the Cameroon Shear Center and the Sanaga Fault. Analysis of gravity data revealed numerous NNE-SSW to NS fractures and faults, acting as local connections with the Kribi-Campo fault associated with the Kribi Shear Zone. This fault is recognized as the southern extension of the

Eseka-Dja fault by Angue et al. (2011) . The EW, NS, NE-SW and NW-SE trending faults identified in this investigation, align with faults observed in previous aeromagnetic studies conducted in southern Cameroon by Feumoe et al. (2012) and Ndougsa-Mbarga et al. (2012) . The presence of these lineaments, although widely spaced due to data resolution limitations of the SGG-UGM-2 global gravity model, serves as evidence that the subsurface of the study area is significantly influenced by geodynamic phenomena ( Figure 9 ).

In this work, to improve the estimation of depth-to-basement solutions, Euler’s 3D deconvolution technique was performed on Bouguer data with the aim of finding the depth-to-basement of lineaments. The procedure is carried out with a 10 × 10 moving window, a depth tolerance of 15% and structural indices 0.5 to effectively locate contact boundaries, faults and dykes. The Euler solution map reveals new deep contacts and clearly defines that the solution for depth ranges from 2.7 km to 11.2 km. Contact depths in the zone are not uniform, suggesting that not all lineaments have the same origin. These Euler depths seem to detect the edges of geological formations, as well as fractures or fault patterns ( Figure 10 ).