The field campaigns conducted allowed for the sampling of materials. The first campaign consisted of identifying the different vulnerable areas. Several scars were identified in this area, but we focused on the most vulnerable zones (Figure 1). Approximately 35 vulnerable sites were identified (32 areas at risk of landslides with a percentage of 91.43% and 3 areas at risk of rockfalls compared to a percentage of 8.57%). These vulnerable sites are shown in Figure 2, which was extracted using Google Earth. Based on criteria such as the morphology of the different sites (steep and abrupt slopes) and the recurrence of minor landslides, six representative sites were selected to collect samples of disturbed and undisturbed soils for geotechnical analyses.

Figure 1. Overview of some vulnerable sites on the SWEBP. (a) Risk of rotational landslide; (b) rockfall risk2-Laboratory methods.

Figure 2. Mass movement density map of SWEBP (extracted from Google Earth).

For this study, 2 Rock samples were collected in December 2023 from the fresh and healthiest possible rock outcrops using a sledgehammer, with sampling points plotted on a topographic map using a GPS. The geographical coordinates of its sample are as follows: Gneiss (N5˚02'34.39''; E10˚09'50.9''; Altitude 1370 m); granite (N5˚02'34.3''; E10˚09'51''; Altitude 1253 m). These samples were taken from the steep and unstable slopes of the most representative vulnerable sites.

Undisturbed soil samples were extracted on August 28, 29, 30 and 31, 2024 from the detachment niches at various vulnerable sites identified using a cylindrical corer and labelled in the order of collection. These samples were carefully preserved, waxed, and labelled in the order of collection for laboratory testing. Using 10 cm diameter and 25 cm high pipes, we collected undisturbed soil samples from each site (vulnerability zones). These samples were waxed using melted candle wax and labelled in the order of collection at the different sites for laboratory testing. During this extraction process, the collected soil was considered a disturbed sample. These collected disturbed soil samples were stored in plastic bags for laboratory testing. These samples represent the 35 vulnerable sites because they exhibit a high level of vulnerability.

The detachment niches allowed for the description of the soil profiles. The boundaries between the different horizons were determined by certain specific characteristics that appeared along the profile and soil levels. Colors were determined using the Munsell soil chart. Texture was determined in situ through manual tests, during which a moistened sample was rubbed between the fingers.

3.1.1. Lithology and Petrographic Analyses

Microscopic observation and petrographic description of the collected rocks were carried out using a polarizing microscope. It initially involved producing a rock sugar sample on which a glass slide was placed and then cut. These thin sections were prepared in the laboratory of the IGMR (Institute of Geological and Mining Research) in Nkolbisson, Yaounde, and were subsequently observed using a polarizing microscope at the laboratory of the Department of Earth Sciences of the University of Dschang. The mineralogical study was carried out using a petrographic microscope on thin sections, following two complementary observation modes: plane-polarized light (PPL) and cross-polarized light (XPL). These two techniques make it possible to characterize the optical and mineralogical properties of the studied samples. Observation in PPL is performed using a single polarizer placed beneath the thin section. While observation in XPL is carried out using two polarizers crossed at 90˚; this mode allows the determination of the fine optical properties of minerals.

3.1.2. Geotechnical Parameters of the Soil

In this study, physical and mechanical soil identification tests were carried out on six samples (E₁, E₂, E₃, E₄, E₅, and E₆) taken from the different vulnerable sites (Fopwanga, Fondjanti, Balouk, Komako, Ndoko, and Koba). Laboratory experiments and calculation methods allowed us to determine the geotechnical parameters of the soil. Geotechnical analyses including physical identification tests (water content, specific gravity test, Atterberg limits, particle size analysis, and porosity) and mechanical tests (box shear test). These geotechnical parameters include, among others:

Particle size analysis: This test allows for the dimensional distribution (particle size class) by weight of the grains in a soil material. The particle size data obtained will allow for the plotting of curves showing the variation of particle size classes as a function of the percentages of sieved material and the mesh size.Atterberg limits: These allow the determination of water content corresponding to the three states of materials: solid, plastic, and liquid. They are determined using the Casagrande apparatus. This involves the liquid limit (ω_L) and the plastic limit (ω_P), in addition to the plasticity index (P_I) and consistency index (C_I).

After the Plasticity Index (P_I) was obtained, values were compared to those in Table 2 which gives the relationship between P_I and soils and helps to determine the level of vulnerability.

Table 2. Degree of soil plasticity as a function of the plasticity index.

Specific weight test or actual density: This is the ratio of the weight of the solid particles of a material to the volume they occupy (it is the absolute density of an anhydrous material).

The actual density is obtained using the following relationship:

Water content by weight: This test determines the amount of water in a soil measured by steam drying. The water content is therefore obtained from these different masses.

After Water content (ω) was obtained, values were compared to those in Table 3 which gives the relationship between ω and soils and helps to determine the level of vulnerability.

Dimensionless parameters: These indicate the proportions of the different phases of a soil. We define:Porosity (n) indicates the extent of voids, that is, whether the soil is loose or compact. It is defined as the ratio of void volume to total volume.

The void ratio (e), whose meaning is analogous to that of porosity, is defined by the following relationship:

Thecompactness denoted (C)

The box shear test: Its purpose is to measure the failure characteristics of a saturated fine soil sample. It provides two mechanical parameters: cohesion and the angle of internal friction. The test measures the upper limits of shear stress in a soil sample and predicts slope failure due to shear stress.

Table 3. Classification of soils in terms of water content ([6]).

3.1.3. Physical Description of Soil Properties

The detachment niches allowed us to describe typical soil profiles. Two soil profiles were described. We determined the boundaries between the different horizons through specific characteristics that appeared along the profile and soil levels. Soils colours were determined using the Munsell soil chart. Texture was determined in situ using tactile tests, which involved rubbing a pre-moistened sample between the fingers to assess the relative proportions of the different grain sizes.

The methodology consists of the extraction, processing and structural interpretation from the Landsat-8 OLI/TIRS satellite images of 27/02/2017 using an approach combining digital preprocessing, contrast enhancement and structural interpretation, thus allowing the extraction of lineaments in order to create a map of structural lineaments using ArcGIS software, which will thus allow the detection of geological structures (faults, ruptures, and tectonic directions) influencing the triggering of mass movements. The spectral bands used were selected for their ability to discriminate morphostructural contrasts: the visible (b2, b3, b4), near-infrared (b5), and mid-infrared (b6 and b7) bands. These combinations improve the distinction between structural discontinuities likely to correspond to lineaments. Directional statistical analysis of the lineaments obtained using RockWorks software allowed us to construct a lineament direction rosette. Subsequently, the lineaments were correlated with the geology of the area to demonstrate the degree of fracturing of the local rocks.

3.2.1. Preprocessing Applied

Before lineament extraction, several preprocessing operations were performed to improve the visual and spectral quality of the images:

Radiometric and atmospheric correction: The images were converted to surface reflectance to reduce the effects of atmospheric conditions, lighting, and atmospheric scattering.Contrast enhancement: Linear stretching of the histogram was applied to accentuate the contrast between homogeneous surfaces and linear structures.Principal Component Analysis (PCA): After contrast enhancement, Principal Component Analysis (PCA) was applied to the selected spectral bands to reduce information redundancy and improve the visualization of geological structures. PCA transforms the initial multispectral bands into a new set of uncorrelated images called principal components, ranked according to the variance explained. The first components generally contain most of the useful spectral information. In this study, the principal components from the Visible, NIR (Near Infrared), and SWIR (Short-wave Infrared) bands were examined visually and statistically. The components exhibiting the strongest structural contrasts (PC1, PC2, and PC3) were selected for structural analysis and lineament detection.Directional spatial filtering: Edge detection filters were applied to highlight topographic and textural breaks associated with fractures:7 × 7 Sobel filter: directional filters oriented along several azimuths (N–S, E–W, NE–SW, and NW–SE)Laplacian filter (Prewitt and Yesou)

These processing techniques accentuate radiometric gradients corresponding to geological discontinuities.

3.2.2. Lineament Identification and Extraction

Lineaments were identified through visual interpretation. The criteria used for their identification were: spatial continuity; alignment of radiometric contrasts; the presence of topographic or morphological breaks; alignment of valleys, drains, or escarpments; and directional consistency with the regional structural context. The identified structures were vectorized in the ArcMap GIS environment to create a lineament database suitable for statistical and structural analyses.

3.2.3. Minimum Length Selected

To avoid including minor, insignificant artifacts and structures, a minimum length of 200 m was selected for the preservation of extracted lineaments. This threshold was chosen taking into account: the spatial resolution of the Landsat images (30 m); the desired level of generalization; and the need to prioritize major structures with tectonic significance. Segments shorter than this were removed from the final dataset.

3.2.4. Elimination of Non-Geological Linear Structures

A validation phase was conducted to exclude linear structures of anthropogenic origin that could be confused with geological lineaments. The removed elements include, in particular: roads and tracks; agricultural or field boundaries; and urban infrastructure. This elimination was carried out by:

Cross-referencing with topographic maps and land cover data;Comparison with road networks;Morphological and contextual analysis, as anthropogenic structures generally exhibit excessive straightness, regular geometry, and a spatial organization different from natural geological discontinuities.

Geologically, SWEBP is composed of gneisses, granites, conglomerates, lava flows and acidic lavas (phonolites, trachytes, rhyolites) (Figure 3) which are intensly fractured and occur as slab-like outcrops.

Figure 3. Geological map of the study area modified from the geological map of [8].

4.1.1. Biotite Mylonitized Gneiss

It outcrops in slabs (Figure 4(a)). The rock is characterized by a coarse-grained structure (Figure 4(b)) composed mainly of alkali feldspar (orthoclase), quartz, and biotite; it has a heterogranular augen mylonitic texture (Figure 4(c)). Under the microscope, this rock shows more or less crushed minerals.

Alkali feldspars (35%): These are primarily composed of orthoclase, characterized by Carlsbad twinning, and occur as xenomorphic crystals ranging in size from 0.1 mm to 1.2 mm. In PPL, they exhibit first-order grey polarization. They are highly altered. Some are peocilitic, containing rounded quartz crystals as inclusions.Quartz (30%): It occurs as subhedral and xenomorphic grains ranging in size from 0.1 mm to 0.8 mm (Figure 4(f)). In XPL, it displays low first-order interference colors and all grains exhibit undulating extinction. Some very small quartz crystals are found as inclusions within the alkali feldspars.Biotite (20%): Occurs as elongated brown flakes, ranging in size from 0.1 to 0.3 mm (Figure 4(f)). Under XPL, it displays moderate second-order interference colors and parallel extinction.Plagioclase (15%): Occurs as sub-hedral crystals with a lamellar shape. In XPL, it shows polysynthetic twinning (Figure 4(e)). These xenomorphic crystals range in size from 0.1 mm to 0.4 mm and also exhibit some edge reactions upon contact with quartz.

Figure 4. Photographs and photomicrographs of gneiss. (a) Gneiss outcrop; (b) Hand specimen of Gneiss; (c), (d) (e) and (f) photomicrographs of gneiss (Pl: plagioclase, Qtz: quartz, Bt: biotite, Fks: Alkali feldspars).

4.1.2. Granite

This rock is also fractured throughout. It outcrops in slabs (Figure 5(a)). It is fine-grained (Figure 5(b)). Under polarizing light, the rock displays biotite, alkali feldspars, quartz, opaque minerals, and plagioclase.

Biotite (35%): Biotite occurs as elongated brown flakes, ranging from 0.1 to 1.5 mm. Some areas show larger biotite crystals (Figure 5(d)). The biotite crystals are intimately bound to the feldspars with straight contacts showing no edge reaction; some are found as inclusions (Figure 5(e)). In PPL, it exhibits pleochroism ranging from light brown to dark brown. Cleavage and a few rare quartz inclusions (Figure 5(c)) are present.Alkali feldspars (30%): Alkali feldspars are subhedral to xenomorphic with sizes varying between 0.1 mm and 0.6 mm (Figure 5(e)). They polarize in the first-order grays. The beginning of alteration of alkali feldspars is observed (Figure 5(c)).Quartz (20%): Quartz occurs as xenomorphic grains ranging in size from 0.1 mm to 0.6 mm (Figure 5(f)). The contact between quartz and other minerals is rectilinear, rarely curvilinear. All its grains exhibit undulating extinction. Some very small quartz crystals are found as inclusions within feldspars.Plagioclase (15%):Plagioclase occurs as polysynthetic twinned crystals. They are xenomorphic, ranging in size from 0.1 mm to 0.6 mm (Figure 5(e)), and are sometimes intercalated among untwined crystals.

Figure 5. Photographs and photomicrographs of granite. (a) Granite outcrop; (b) Hand specimen of Granite; (c), (d), (e) and (f) photomicrographs (Pl: plagioclase, Qtz: quartz, Bt: biotite, Fks: Alkali feldspars).

The mineralogical assemblage of granite (Bt + Fks + Qtz + Pl) (Table 4) largely determines its geomechanical behaviour. That of biotite-bearing mylonitized gneiss (Fks + Qtz + Bt + Pl) exhibits anisotropy due to the presence of biotite. These mineralogical assemblages are crucial for slope stability because the progressive decomposition of feldspars and biotite into clay products leads to a decrease in intergranular cohesion, an increase in porosity and permeability, creating conditions favorable to slope destabilization and the initiation of mass movements.

Table 4. Paragenesis and mineralogical associations of the rocks in the study area.

Laboratory tests on samples of disturbed and undisturbed soils yielded the physical and mechanical parameters recorded in Tables 5-7 respectively. These include, among others, specific weight or actual density, natural water content (ωNat), dry and wet density, particle size analysis (Figure 6), Atterberg limits (Figure 7), and straight-line shear in the box, in addition to other calculated physical parameters.

Table 5. Physical parameters of soils developed in the SWEBP.

Porosity (n) is very high (47.37 - 51.21) (Table 5).The water content (ω) of the materials varies between 24% and 31%, which is very close to or even lower than the plastic limit ω_P (31.4% - 38.7%) (Table 5). According to [6] (Table 2) the soils in the study area are of average to bad quality.Atterberg limits (Table 5): This refers to the liquid limit (ω_L) and the plasticity limit (ω_P), to which are added the indices of plasticity (P_I), liquidity (L_I), consistency (C_I), and handling (I_W).The plasticity indices are between 29.9% and 35% (Table 3) indicating very plastic soils that are likely to swell. According to the table of degrees of plasticity of a soil as a function of the plasticity index, (Table 2) we can say that the soils on the SWEBP are plastic soils.The most important particle size class in the SWEBP is that of clays and silts. Thus, these soils belong to the clay particle size class because the percentage of particles passing through the 0.08 µm sieve is greater than 35% (70% - 77%) (Table 6).The specific weight (γ_S) of the soils in the study area is average, being slightly less than 2.65 g/cm³, which is the average density of mineral solid particles (Table 5). These specific weight values are very close to each other in the six vulnerability zones.

Figure 6. Particle size distribution curves of soils developed in the SWEBP.

The soils of the SWEBP are subdivided into two classes (Figure 6): granular soils (gravel, coarse sand, fine sand) with a diameter greater than 0.02 mm and fine soils (clays and silts) with a diameter less than 0.02 mm. Table 4 presents this distribution of the different granular classes.For shear, the samples underwent straight-line shear box tests or direct shear tests. The cohesion values Cu (bar) vary between 0.22 and 0.45 bars while the values of the angle of internal friction φu (˚) are high vary between (22 - 26) (Table 7).

Table 6. Grain size distribution of the different classes of soils developed in the SWEBP.

Figure 7. Atterberg limits of the different soil samples developed in the SWEBP.

Table 7. Mechanical parameters of soils developed in the SWEBP.

1)Representative soil profile of Komako: It is located halfway up the slope, on a road cutting in the locality of Komako, and is situated at an altitude of 862 m. It has a height of 150 cm and consists of two soil horizons with irregular boundaries. From top to bottom we have:

Horizon A or organic horizon: It is approximately 40 cm thick. This horizon is characterized by the presence of organic matter, a low degree of root penetration, and very high levels of biological activity. This horizon has a clayey-sandy texture and a loose structure. It has a reddish-black colour (10YR4/4) which corresponds to dark yellow brown (Figure 8(a) and Figure 8(b)).Horizon BC or weathering horizon: Approximately 110 cm thick, it has a very low degree of root penetration and the presence of biological activity (porosity), including boulders. This horizon has a sandy-clayey texture and a compact structure. It has a brown colour (10YR4/3) which corresponds to brown (Figure 8(a) and Figure 8(b)).

Figure 8. Soil profiles of Komako.

2)Representative Soil Profile of Koba: It is located mid-slope, on a road cutting in the Koba area. It has a height of 190 cm. This profile consists of two soil horizons with irregular boundaries. From top to bottom, we have:

Horizon A, or organic horizon:It is approximately 30 cm thick. This horizon is characterized by the presence of organic matter (porosity), a low degree of root penetration, and very high levels of biological activity. This horizon has a clayey-sandy texture and a loose structure. It has a light brown color (2.5Y5/4), corresponding to light olive brown (Figure 9(a) and Figure 9(b)).Horizon BC, or weathering horizon: Approximately 160 cm thick, it exhibits a very low degree of root penetration and an absence of biological activity, with the presence of small rocks. This horizon has a sandy-clayey texture and a compact structure. It has a brown colour (2.5Y5/4), corresponding to light olive brown (Figure 9(a) and Figure 9(b)).

The lineament map (Figure10(a)) obtained shows numerous fractures (99 lineaments), indicating that it represents areas of ground instability and therefore areas susceptible to mass movements. Statistical directional analysis of the lineaments obtained allowed the construction of a direction rosette (Figure 10(b)) of the lineaments in RockWorks to determine the dominant orientations representing highly fractured areas. The main direction of the fracture zones is shown in yellow, oriented NE-SW, and a secondary direction is shown in green, oriented ENE-WSW.

Figure 9. Soil profiles of Koba.

Figure 10. Lineaments and rose diagram of SWEPB. (a) Lineaments map, (b) Rose diagram.

The superposition of the lineaments with the geological map (Figure 11) of the area shows the high degree of fracturing of the SWEPB rocks.

Figure 11. Lineament map associated with the geology of the SWEPB.