Depth Variation of the Lithosphere beneath Garoua Rift Region (Cameroon Volcanic Line) Studied from Teleseismic P-Waves

Teleseismic events have been selected from a database of earthquakes with three components which were recorded between February 2005 and January 2007 by five seismic stations across the Garoua rift region which constitutes a part of the Cameroon Volcanic Line (CVL). The iterative time deconvolution performed by [1] applied on these teleseismic events, permitted us to obtain P-receiver functions. The latter were subsequently inverted in order to obtain S-wave velocity models with respect to depth which were then associated to the synthetic receiver functions. This made it possible to explain the behavior of the wave and the medium through which they traveled. The main results obtained indicate that: (1) The lithosphere appears to be thin in its crustal part with a mean Moho depth of 28 km and S wave velocity of 3.7 km/s. (2) In its mantle part, the lithosphere is thick in nature having a thickness that varies between 42 km and 67.2 km. The greatest depth is noticed towards the center located around Garoua while the least depth corresponds to a location around Yagoua in the North. The Low velocity zone which makes it possible to determine the depth of the lithosphere was seen to have a thickness which varies between 42 km


Introduction
The Garoua rift region is the continuation of the Northern section of Cameroon Volcanic Line and the part of the large Benue trough which, is situated between the latitude 8˚ and 11˚ North and longitudes 13˚ and 16˚ East. It is bounded to the west by the Federal Republic of Nigeria, to the North by the Mandara Mounts and to the south by the Adamawa plateau ( Figure 1).
Structural and geological studies by [2] and [3] show that the Garoua basin is an E-W to N120 trending trough infilled by Middle to Upper Cretaceous marine sandstones. These sediments have also been described by [4] and [5]. The Garoua sandstone series overlap an approximately E-W trending trough called the Tcheboa trough, similar to the Figuil, Hama-koussou and Mayo-Oulo basins [6] [7]. The whole region of the Garoua basin presents outcrops of sandstone and intrusive granites, which form the basement complex below the sediments, and intrusive diorites along the Poli-Lere axis [8]. Some hypovolcanic dykes are also found within the Garoua sandstones. The basaltic lavas ( Figure 1) found here are similar to those of the Cameroon Volcanic Line [9]. The basin is limited by normal faults which outcrop on its northern and southern borders [3] [8] [10]. Some geophysical works have been carried out in the Garoua rift region and particularly in the Garoua basin area. The regional structural setting of the Garoua Basin is characterized by three major normal faults striking mainly in the NW-SE to NNE-SSW direction [12]. The continental crust underneath the basin (about 24 km) is thinner than the normal crust, but may be a little thicker to the east. This thinning of the crust is due to extensional regional stress and the uplift of the Asthenosphere is as a consequence and this result from an isostatic compensation. This leads to an average sedimentary pile thickness of about 6 km from results obtained by [11] and [12]. This thinning is probably due to the extensional process of basin formation in the Cretaceous. The Moho is found to be uplifted in the basin, and would be the result of this extension and the associated thermal and isostatic compensation. [ age of the lithospheric structure in Cameroon using the stack method of the receiver functions resulting from the surface waves [25] or focused mainly on the mantle [23] and [24] in other hand. In this present work, we will come out a 1-D velocity model of S-waves of the lithosphere from the inversion of the P-waves receiver functions having the properties to propagate beyond the crust. This shear wave velocity models could also provide information regarding its structure.

Data
The data used in this study consist of the earthquakes recorded by 32 portable broad-band seismometers installed across the country between January 2005 and February 2007 by the Cameroon Broadband Seismic Experiment. The data were collected by a team of geoscientist from the University of Yaounde I and the Institute of Geological and Mining Research-Cameroon in partnership with researchers from Pen State University in the United State of America. The seismic network comprised 5 broadband stations (Table 1) [26]. From the earthquake data recorded, the teleseismic events that occurred at epicentral distances between 30˚

Computation of the Receiver Functions
The treatment of the teleseismic is based on the receiver function method. The To complete the estimation of the receiver functions, 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 [26]. Events that were recovered to less than 85 percent were rejected. The remaining waveforms were visually inspected for coherence and stability ( Figure 5). Open Journal of Earthquake Research

Inversion of the Receiver Functions
In this work, the P-waves were inverted to obtain an S-wave velocity model that produces an estimation of shear velocity structure beneath 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 (or predicted) receiver functions. The inversion was performed using the Rftn96 program developed by [28] and [29]. 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 be-

Results
The different results coming from the inversion of the receiver function computed for the five seismic stations studied are shown in Figure 6 (a, b, c, d and e).
The different interpretations are presented in Table 3 and Table 4.
Regrouping the types and times of different conversions and multiple conversions coming from the curves of synthetic receiver functions from the Garoua rift stations, Table 3 shows that, a good agreement exists between the curves,   The velocity models obtained by inversion of the receiver function followed by the different values coming from Table 4

Comparison of the Synthetic Receiver Functions Results by Station and with Previous Estimates
Observing the different times of Ps phase and subsequent reverberations (PpPs and PpSs) obtained at each station (Table 4), in addition to their different amplitudes, one notes respectively an increase and a decrease from the station CM28 to station CM32. This leads to the deduction that the wave is converted and attenuates according along the South-North direction of the Garoua rift.
Previous estimates of the conversion time in the Garoua rift region have been carried out by [25] and the comparison with the new estimates is shown in Table 5.
Comparing the values of this study with those obtained by [25], one observes that, there is slight difference with respect to the different conversions. This difference can be explained by the method used in computing the same data.

Comparison of the Shear Wave Velocity Model Results by Locality and with Previous Estimates a) Comparison of the shear wave velocity model by locality
The shear velocity models obtained in each locality present a thin behavior in the crustal part of lithosphere. This behavior is supported by [11] [19] and [22].
The mantle part of the lithosphere is thick, with the depth of the lithosphere that varies beneath the Garoua rift region between 42 km and 67.2 km (Figure 7). It comes out again that, the deepest part of the lithosphere is located in the center of the rift (Garoua) and the section having the least depth is located at the North, more precisely in the Yagoua locality ( Figure 7).

b) Comparison of the shear velocity model results with previous estimates
Many studies have already been carried out on this behavior and the depth in this area by some geoscientist like [11] [22]. Comparisons with previous estimates of the shear wave velocities as well as with the thickness of the crust and lithosphere are done in Table 6.  [21]. The lithospheric depth has been determined by the existing of the low velocity zone started at the same depth. This low velocity zone has been confirmed by [23] that found a low-velocity anomaly beneath the CVL extending to at least 300 km plausibly related to a thermal perturbation.

Conclusion
Iterative deconvolution has been applied on teleseismic events recorded between 2005 and 2007 to obtain the receiver functions. These receiver functions have been inverted to study the lithosphere-Asthenosphere Boundary beneath the Garoua rift region. It was found from this study that: 1) The synthetic receiver functions associated to the shear velocity model obtained show the existence of a Ps phase and subsequence reverberations PpPs and PpSs at 3.7 s, 11 s and 17.6 s respectively. The amplitude of the subsequent reverberations decrease and finally disappears at station CM32 passing through stations CM29, CM30 and CM31.
The existence of the different phases and the low amplitude of the different phases lead to the deduction that, the wave has really undergone conversion and attenuation along the South-North direction of the Garoua rift region. 2) The lithosphere has a thin nature in its crustal part with the Moho located at a mean depth of 28 km according to the S-wave velocity of 3.7 km/s. The lithosphere is thick in its mantle part with its limit depth that varies between 42 km and 67.2 km of depth. The deepest part of the lithosphere in the Garoua rift region is situated towards the center that is located around Garoua while the least depth is situated around Yagoua to the North due to the existence of the first low velocity zone (LVZ) in the upper mantle whose thickness also varies between 42 km and 116.7 km. The different results obtained in this study have been compared to previous results existing in this region. Some similarities have been noticed in some cases like in the velocity of the S waves in the crust and the time of PpPs phase. The main differences with other results were noticed in the times of Ps and PpSs phases and Moho depth. These differences can be justified by the type of method or data used. In this study, though studies have been carried out on the boundary and the behavior of the lithosphere, it will never the less be important to carry out a study on the origin of the variation noticed in the depth of the lithosphere.