Situated between latitudes 6˚ and 8˚ North and longitudes 11˚ and 16˚ East, the Adamawa plateau in Cameroon (Figure 1) presents itself like a volcanic axis characterized by region of the fissured basaltic volcanism. The fault leaving Foumban joins this region and cuts across it nearly diagonal. It can also be noted that, this region is one of the zones in Cameroon where one can meet several geological accidents.
The basement of this region consists of a Precambrian magmatic gneiss complex (Figure 1) that recorded Pan-African granitization  . That basement is overlain by a sequence of basaltic to andesitic lavas that are largely of Tertiary age  . These lavas are essentially alkaline indicating an affinity to continental rifts  . The sedimentary formations here are mainly composed of conglomerates and marl of the Cretaceous Mbere and Djerem Troughs,   . These formations have undergone intense tectonic activities resulting in the displacement of basin structures which were frequently filled by volcanic material upwelling through deep fractures in the Adamawa region. Three major tectonic structures are associated with the Adamawa plateau: the Cameroon Volcanic Line (CVL), the Foumban Shear Zone (FSZ) and the South Adamawa Trough (Figure 1). The FSZ is part of the Central African Shear Zone (CASZ), which is a succession of major Pan-African faults, trending ENE-WSW and extending over about 2000 km from Cameroon to Sudan   . This shear zone in Central Cameroon is characterized by a wide band of mylonites that display a dextral sense of displacement   . In a pre-opening of the South Atlantic, these faults extend towards western Brazil through the Pernambuco fault  . The South Adamawa Trough corresponds to the Cretaceous Mbere and Djerem sedimentary basins, limited to the North by the CASZ and characterized by conglomerates and mylonites localized along the fault zone   . Two areas of seismicity exit, one of which is linked to Foumban fault and crosses the Adamawa Plateau while the other is associated with the north border of the Congo Craton. A majority of the earthquakes recorded in the Adamawa plateau region are generally in relation with the activity of the Volcanic Line of Cameroon  . The gravity analysis in the area by  highlighted a N70˚ trending anomaly coinciding with the strike of the fractures affecting the basement  used joint inversion of Rayleigh wave group velocities and receiver functions to study the structure of
Figure 1. Geologic map of study area (modified from  ). 1: High grade metamorphosed paleoproterozoic gneiss; 2: Syn-tectonic panafrican granite; 3: Principal undifferentiated panafrican Gneiss; 4: Panafrican Metasediments; 5: Envelope of cretaceous sediments; 6: Cenozoic volcanism; 7: Central cameroon fault line.
the crust beneath Cameroon. The lithosphere beneath Cameroon is characterised by a heterogeneous crust with a relatively constant thickness and a low velocity uppermost mantle at the edge of the Congo Craton  .
Previous geophysical research works studied the lithosphere without establishing its boundary and depth. This study will strive at examining the boundary of the lithosphere.
2. Data and Methodology
The data used for this study were recorded between January 2005 and February 2007 by the Cameroon Broadband Seismic Experiment, which consisted of 32 portable broad-band seismometers installed across the country (Figure 2). Each station (Table 1) was equipped with a broad-band seismometer (Guralp CMG-3T or Streckeisen STS-2), a 24-bit Reftek digitizer and a GPS (Global Positioning System) clock.
Figure 2. Colour elevation map showing seismic station locations and shear zones. The circled numbers refer to station codes, for example, 22 refers to station CM22 and the rectangle represents the study area.
Table 1. Characteristics of the different stations used to record the teleseismic events.
Data were recorded continuously at a rate of 40 samples per second. In the study area, two stations (red color) were installed in 2005 January and operated
during 2 years; the remaining four stations (yellow color) operated only during the second year of the experiment. The station spacing during the second year (2006) of operation was about 50 to 150 km.
Data from the Cameroon Broadband Seismic Experiment have been used to perform inversion of P-wave receiver functions. Receiver functions are chronological temporal series computed using the seismic components recorded at the large band station. They are generated by the time domain iterative deconvolution method of  , applied to seismograms rotated into vertical, radial and transverse components and can be used to image velocity contrasts across discontinuities.
2.1. Estimation of the Receiver Functions
Receiver functions were computed using data from teleseismic events (Table 2) that occurred at epicentral distances between 30˚ and 95˚ with magnitudes ≥5.5.
To compute the receiver functions, visual inspection is first applied in order to confirm the presence of the signal and if the different types of waves which appear on the three components of the seismogram can be identified (Figure 3).
Then the selected waveforms were decimated to 10 samples per second, windowed between 20 s and 140 s after the leading P arrival, de-trended, tapered and high pass filtered above 50 s to remove low-frequency, instrumental noise. Radial and transverse receiver functions were then obtained from the filtered traces by rotating the original horizontal components around the corresponding vertical component into the great circle path, and deconvolving the vertical component from the radial component through the iterative time domain deconvolution procedure of  , with 200 iterations using the Gaussian a = 2.5 corresponding at to the frequency 1.2 Hz because it helps to discriminate gradational transitions from sharp discontinuities in the receiver structure under the station  . The recovery percentage of the original radial waveform was evaluated from the rms misfit between the original radial waveform and the convolution of the radial receiver function with the original vertical component. Events that were recovered to less than 85 per cent were rejected. The remaining waveforms were visually inspected for coherence and stability (Figure 4).
Figure 3. Three components of seismogram recorded at station CM22 for event of 22/02/2016.
Figure 4. Example of the receiver function computed from station CM22 using teleseismes events. The horizontal axis represents the time in second (s) while the vertical axis represents the amplitude of P-wave in meter (m).
Table 2. Events with magnitude Mb ≥ 5.5 used for the study.
2.2. Inversion of the Receiver Functions
Receiver functions are traditionally inverted to obtain an S-wave velocity model that produces an estimation of shear velocity structure under a given seismic station. There is no guarantee that a unique inversion result will be obtained, as the method seeks to minimize the differences between observed and synthetic receiver functions. The inversion was performed using the method developed by   . The method is based on a linearized inversion procedure that minimizes a weighted combination of least squares norms for each data set, a model roughness norm and a vector-difference norm between inverted and pre-set model parameters. The velocity models obtained are consequently a compromise between fitting the observations, model simplicity and a priori constraints. The starting model used in this inversion consisted of an isotropic medium of constant velocity layers that increase in thickness with depth. The thicknesses of the first, second and third layers are, respectively, 45 km, 90 km and 80 km, while the thickness increases at each instant to 5 km between 0 and 45 km depth, to 10 km between 45 and 135 km and 20 km below a depth of 135 km and a linear shear wave velocity increase in the crust from 3.2 to 4.0 km/s and 4.0 to 4.7 km/s in the lithospheric mantle overlying a flattened PREM (Preliminary Reference Earth Model) model  for the mantle.
3. Results and Discussion
Table 3 shows that the synthetic receiver functions have very important percent of Signal Power Fit (>80%) apart of that corresponding to CM21 located at Tibati. This justifies a stability of the curves of current and initial model and it causes the wave to undergo a rare conversion. The times of Ps conversion (tPs) are practically uniform at close to 0.2 s.
Table 3. Interpretation of synthetic receiver functions.
Figure 5. Shear velocity profiles for each of six stations. The solid line shows the depth delimitation. (2.5 is the Gaussian filter; 91.15 is Percent of Signal Power Fit; 0.074 is the ray parameter 053 is the day of year).
Table 4. Interpretation of the inversion curves.
Table 4 shows from the velocity models that the mean velocity of the S waves is 3.7 km/s in the crust and more than 4 km/s in the lithospheric mantle. Discontinuities are localized; firstly at an average depth 35.1 km between for the crust and upper mantle (Moho) and secondly with variation from 73.8 km to 85 km at the boundary between the lithospheric and asthenosphere. The lithospheric mantle and the low velocity zone thicknesses varies between 39 km and 49.6 km and between and 73.8 km and 116.7 km respectively.
3.2.1. Comparison of the Synthetic Receiver Function by Localities
Comparing the different synthetics receivers functions, it is observed that the current (red colour) and initial (blue color) synthetic receiver functions are stable (Percent of Signal Power Fit is more than 80%) and the time of Ps conversion phase practically homogeneous at 0.2 s except for CM21 located at Tibati where they are not stable. This non stability would be due to the Foumban shear fault that crosses the Tibati locality.
3.2.2. Comparison with Previous Estimates
The uniformity of the time of Ps conversion at the level of the moho (Table 3) and the low amplitudes as seen on the synthetic receiver functions suggest a homogenous nature of the crust in the Adamawa plateau region Tokam et al. (2010). The small length interval of the secondary wave velocity Vs at the crust and lithospheric mantle expresses the fact that the lithospheric mantle is thin in the Adamawa plateau region.
Comparison with the previous estimations are showed in Table 5.
Table 5 shows that, estimates of average S-wave velocity Vs in the crust and upper mantle on one hand and the moho depth on the other hand are in very good agreement to previous estimates based on both gravity and seismic data in the region. A slight difference is noticed at the level of the depth of the moho and the lithospheric mantle boundary. This can be justified by the type of method or data used.
Table 5. Comparing with previous estimates.
3.2.3. Comparison the Lithospheric Mantle and the Thickness Low Velocity Zone (LVZ) of Different Localities
A comparison of the lithospheric mantle and low velocity zone thickness (LVZ) for localities in the Adamawa plateau region is shown in Figure 6.
These figures show that the lithospheric mantle and low velocity zone thicknesses are not uniform in this region but that they are variable which also entails a variation of the depth of the lithospheric mantle in this region.
The teleseismic events recorded between 2005 and 2007 from the six seismic stations installed in the Adamawa Plateau have been treated with the receiver function method to investigate the lithospheric mantle and the crust. It was found from this survey that: 1) the synthetic receiver functions obtained show the existence of Ps conversion at 4.7s at the moho; 2) the crust in this region is thick with an S wave velocity of 3.7km/s and an average depth of 35.1 km; 3) the lithospheric mantle has a thickness that varies between 39 km and 49.6 km for an average S wave velocity greater than 4 km/s; 4) In this region, there exist a low velocity zone which has a variable thickness ranging between 20 km and 43.9 km and the lithosphere-asthenosphere boundary also varies between 73.8 km and 85 km according the low velocity zone position. The results obtained in this work have been compared to others existing in this region. Some similarities have been noticed in some cases like in the depth of the crust, the velocity of the S waves in the crust and in the lithospheric mantle, and the existence of a Low Velocity Zone. The slight differences with other cases have to do with the depth of the lithospheric mantle. These differences can be justified by the type of method or data used. Nevertheless an alone station (CM21) situeted at Tibati locality does not product the good results especially with particularly the synthetic receiver function due to the Foumban shear fault that across that station. In this work, though the boundary of the lithosphere has been studied, it will never the less be important to carry out geodynamic studies to investigate the causes of the observed variations as well as the level of stability of this lithospheric boundary.
Figure 6. Comparison the lithospheric mantle thickness and the thickness Low velocity zone (LVZ) of different localities.
We wish to express our gratitude to the team and sponsor of the Cameroon broadband seismic experiment for the data that was colled and has been used in this work.