African stress pattern from formal inversion of focal mechanism data
Introduction
The 2nd and 3rd order patterns of intraplate stress field are becoming recognized in the continental plates with improved data interpretation and spatial coverage. The lack of data has always been a problem in East Africa and impeded adequate tectonic interpretation. The new focal mechanism data and their analysis presented here allow and facilitate a revisiting of the tectonic interpretations of the stress field patterns. The 2nd and 3rd order stress patterns obtained as the result of formal stress inversion of lustrated focal mechanism solutions reveal many interesting “complications” in the stress field that were not captured by large-scale numerical models of previous studies. The results presented here intend to lay the foundation for constraints on higher detailed, local, high-resolution numerical models that will help to differentiate between the various sources of the 2nd and 3rd order stress field.
In the absence of continental gravitational potential energy (PE) forces, the entire African plate would be expected to be in compression, due to the surrounding ridge push forces (Fig. 1). Within most of the African plate (except for the Mediterranean region and Western Africa), those PE forces are dominated by the effect of the East African Rift System (EARS).
The East African Rift System is an example of the relatively rare instance of active continental rifting. Passing for nearly 3000 km through the continent, the EARS separates the Nubian subplate to the west from the Somalian subplate to the east (Fig. 1). Beginning in the Afar triple junction, it crosses the Ethiopian highland, forms the Gregory Rift in Kenya and disperses in northern Tanzania after Lake Natron, forming the Eastern Rift Branch. Grimison and Chen (1988) proposed an extended Eastern Branch that connects to the Davie Ridge along the continental margin on the Indian Ocean coast of East-Africa to explain the seismic activity in the northern Mozambique Channel. The Western Rift Branch starts in southern Sudan and runs through the rift valley lakes — including Lakes Tanganyika and Malawi — to Mozambique. The southern end of the EARS is less clear and its connection to the Southwest Indian Ridge is still controversial (Horner-Johnson et al., 2005, Lemaux et al., 2002). Chu and Gordon (1999) used seafloor-spreading rates to determine the Euler pole of rotation between the Nubian and the Somalian subplate in the southern Mozambique Channel. Hence, the tectonic regime changes from extension north of the pole to compression south of it.
The opening kinematics of the EARS has been the focus of investigations since decades. In the absence of sufficient earthquake focal mechanism data, the kinematic models were first proposed on the basis of a geometrical interpretation of the large-scale tectonic structures interpreted from remote-sensing imagery (Chorowicz and Mukonki, 1980, Kazmin, 1980). These models were soon supplemented by paleostress inversion of geological fault-slip data collected along the major faults bordering the rift depressions (Tiercelin et al., 1988, Chorowicz, 1989). This approach suffers an insufficient timing control, as the fault-slip data have often been measured in basement rocks along the supposedly active faults. It was later shown that most of the measured data could belong to past tectonic events unrelated to the present rifting dynamics, hence highlighting the importance of stress field fluctuation through geological time (Strecker et al., 1990, Bosworth et al., 1992, Delvaux et al., 1992, Ring et al., 1992, Delvaux, 1993, Bosworth and Strecker, 1997). Such time fluctuation of stress field is also supported by independent observations from seismic profiling in the rift lakes (Morley et al., 1999) and other field-based studies (Le Gall et al., 2005, Nicholas et al., 2007). Recently, kinematic models based on GPS geodesy have been proposed (Fernandez et al., 2004, Calais et al., 2006, Stamps et al., 2008), but due to the small number of permanent GPS stations used, they are not yet able to resolve the plate motions with sufficient details. In the meantime, the knowledge of low magnitude focal mechanisms in regions off-side the main rift zone gives the opportunity to invert for the regional stress field (Barth and Wenzel, 2010). During the last 10 years, the number of available focal mechanisms available for East Africa has increased considerably due to a longer observation time, improvement of the seismic network, installation of local seismic networks and more detailed calculation procedures. As a result, it is now possible to image the second order — and locally the third order — pattern of tectonic stress for large regions of the EARS and adjacent parts of the African Continent, evidencing the lateral variability of the present-day stress field.
In this paper, first we compile all presently available well constrained single focal mechanism data up to September 2008 (336 single events and 1 composite event). We then group the data into 24 distinct regions (boxes) in function of the geographic proximity and the general tectonic structure and we perform a formal inversion in order to determine the present-day stress field. We compare the results obtained by two different methods (TENSOR program of Delvaux & Sperner, 2003) and SLICK method of Michael, 1984, Michael, 1987) for the same boxes and datasets. We also discuss the possible sources for the observed tectonic stress pattern.
Section snippets
Data compilation
Source mechanisms in East-Africa were examined by several studies using both first motion analysis (e.g. Fairhead and Girdler, 1971) and waveform inversion. The Global CMT Project — formerly Harvard-CMT (Dziewonski et al., 1987) — routinely determines focal mechanisms by moment tensor inversion of both long period body- and surface-waves. The global level of completeness for CMT-solutions is approximately MW ~ 5.5 (Arvidsson and Ekström, 1998), whereas it is lowered to MW ~ 5.1 for East-Africa,
Zonation and Box definition
Since the focal mechanism data are not consistent in terms of stress regime and stress orientation over the entire East-Africa and thus cannot be inverted altogether (Barth, 2007), we divide the region into sub-areas (boxes) to study regional changes in stress orientation. For this purpose the zonation of the Global Seismic Hazard Assessment Program (GSHAP) is applied as a starting point. These worldwide zonations take into account the recent and historic seismicity and hence define areas with
Tectonic stress inversion
To study the recent stress field for East Africa we perform formal stress inversions of the given focal mechanisms following two different techniques: the TENSOR program (Delvaux and Sperner, 2003) and the SLICK method (Michael, 1984, Michael, 1987). Both attempts rely on two major assumptions for the study region: (a) the stress field is uniform and invariant in space and time, and (b) earthquake slip d occurs in the direction of maximum shear stress τ (Wallace–Bott hypothesis, Bott, 1959).
Second and third-order stress field
According to Heidbach et al. (2010-this issue), the tectonic stress field can be classified as a function of the spatial scale of investigation: the 1st order stress field is of continental scale and induced by plate boundaries, 2nd order is of intraplate origin such as continental rifting, isostatic compensation, topography and deglaciation, and the 3rd order corresponds to the detailed stress pattern at the scale of less than 100 km, generated by local density and strength contrasts,
Sensitivity of the interpretation and results to the box boundaries
During the delicate step of data selection through the definition of the box boundaries, both methods showed the same order of variation. The choice of the box boundaries was made in order to restrict the coverage area of the subset to a minimum, with the greatest density of similar focal mechanisms. The principal difficulties were the determination of the boundaries of boxes 6a and 6b, as the stress field changes rapidly in the Tanzanian sector of the Gregory Rift and there is a spatial mixing
Conclusion
We show that using focal mechanism data of 332 earthquakes in the African plate, it is possible to resolve the first and second order stress field of the East African Rift by formal stress inversion. For some distinct regions we even obtain information on the local third order stress pattern. Both techniques used; the TENSOR method (Delvaux and Sperner, 2003) and the SLICK method (Michael, 1984/1987), show very similar results of the SHmax orientation. Only for boxes that contain a low number
Acknowledgements
Damien Delvaux is funded by the SPP Belgian Science Policy under Action 1 program, this work being performed under project TectoSediCongo. We thank the Royal Museum for Central Africa for active support all along this work. The manuscript benefited from the constructive comments of an anonymous reviewer as from the encouragements of the Editor, Mike Sandiford.
References (93)
Active compressional tectonics in Central Africa and implications for plate tectonic models: evidence from fault mechanism studies on the 1998 earthquakes in the Congo Basin
J. Afr. Earth Sci.
(2002)- et al.
New constraints on the intraplate stress field of the Amurian plate deduced from low magnitude earthquake focal mechanisms
Tectonophysics
(2010) - et al.
Stress changes in the Afro-Arabian rift system during the Miocene to Recent period
Transfer and transform fault zones in continental rifts: examples in the Afro-Arabian rift system. Implications of crust breaking
J. Afr. Earth Sci.
(1989)- et al.
Paleostress reconstructions and geodynamics of the Baikal region, Central Asia. Part II: Cenozoic rifting
- et al.
Late Quaternary tectonic activity and lake level fluctuation in the Rukwa rift basin, East Africa
J. Afr. Earth Sci.
(1998) - et al.
Strike-slip faulting in continental rifts: examples from Sabukia, East Africa (1928), and other regions
Tectonophysics
(1991) - et al.
Global seismicity of 1977: centroid-moment tensor solutions for 471 earthquakes
Phys. Earth Planet. Inter.
(1987) - et al.
Regional geodynamic implications of the May–July 1990 earthquake sequence in southern Sudan
Tectonophysics
(1992) - et al.
Distribution of crustal extension and regional basin architecture of the Albertine rift system, East Africa
Mar. Pet. Geol.
(2000)