doi:10.1016/j.pepi.2006.05.007
Copyright © 2006 Elsevier B.V. All rights reserved.
Texture of mantle lithosphere along the Dead Sea Rift: Recently imposed or inherited?
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
Vadim Levina, 
, 
, Alissa Henzaa, Jeffrey Parkb and Arthur Rodgersc
aDepartment of Geological Sciences, Rutgers University, Piscataway, NJ, United States
bDepartment of Geology and Geophysics, Yale University, New Haven, CT, United States
cEarth and Environmental Science Directorate, Lawrence Livermore National Laboratory, Livermore, CA 94551, United States
Received 20 June 2005;
revised 12 December 2005;
accepted 8 May 2006.
Available online 14 August 2006.
Abstract
Seismic anisotropy, a property linked to the texture of the mantle rock, should be distributed with depth along the trace of the Dead Sea Rift (DSR), owing to a combination of present day and ancient tectonics. Using data from four permanent and one temporary seismic observatories we evaluate birefringence (splitting) of 91 teleseismic core-refracted shear waves, primarily SKS phases. We find significant levels of birefringence in the bulk of observed phases. We also find that birefringence parameters (fast directions and delays) vary as a function of source–receiver geometry. Notably, the pattern of this directional variation in birefringence is quite similar at all sites we have examined. We interpret observed birefringence in SKS phases in terms of one- and two-layer models. Single-layer models for all stations exhibit a fast polarization oriented 12–19° east of north, with anisotropy sufficient to generate 1.3-s time delay. We find strong evidence for at least two distinct anisotropic layers. For the two layer models, the upper layers resemble the single-layer models, showing near-north fast polarizations and delays on the order of 1 s. Three out of four sites show fast polarizations in the lower layer that strike 50–80° CW from north with time delays 0.3–0.6 s. One site, at the northern end of the DSR, displays a higher degree of anisotropy in the lower layer, and a more northerly fast polarization. Overall, the lower layers at all sites appear to be consistent with the deformation caused by plate motion relative to the asthenosphere. The fabric in the upper layer is sub-parallel to the present-day transcurrent motion on the DSR, but also matches the typical orientation of lithospheric seismic anisotropy in the Arabian shield. Our overall conclusion is that the impact of the DSR on the rock fabric of the mantle lithosphere is probably quite weak.
Keywords: Anisotropy, Lithosphere, Upper mantle, Body waves
Fig. 2. Examples of data used in the study. Two SKS phases recorded by station JER (Jerusalem) illustrate the single-event measurement procedure. (Left) Radial (solid) and transverse (grey) components of recorded waveforms. Data are bandpass filtered between 0.02 and 0.2 Hz. Event labels and back-azimuth values are noted on the plot (middle) Waveforms rotated into fast (grey) and slow (solid) components show close similarity of pulse shapes. Note different values for fast direction. (Right) Enlarged sections show 
0.5 s of time lag between fast and slow components for both observations.
Fig. 3. A topographic relief map of the study area showing measurements of shear wave birefringence arranged according to the geometry of observations. Individual measurements are plotted as bars aligned with the azimuth of fast polarization in the shear wave, and scaled by the delay value. Centers of the bars correspond to map projections of their rays’ piercing points at 150 km depth.
Fig. 4. Results of group inversions for one-layer (left) and two-layer (right) structures. (Left column) Misfit value surfaces for one-layer group inversions. Global minima of the surfaces are marked by white stars, and corresponding values are noted on the plot. (Right column) Plots of best (smallest) misfit values for combinations of fast directions in two layers. Values less then those of a global minimum for a corresponding one-layer model are shaded. Green stars mark the smallest value of misfit found in a search of two-layer model parameters (see Table 1 for specific values). Lines show the fast direction value of the one-layer model. Noted on the plots are ranges of values for two-layer models that fit data better then corresponding one-layer solutions. See text for a description of how these models are selected.
 |
Fig. 5. Parameters of final one-layer (left) and two-layer (right) models of seismic anisotropy derived for individual sites are shown as bars aligned with fast polarization within a layer and scaled with the amount of delay this layer contributes. For one-layer models, values of the model with the smallest misfit are shown. For two-layer models, two bars are shown per layer to delimit a range of values that yield similar solutions. Parameters of the lower layer are depicted by wider darker bars. Open arrows illustrate various models of plate motion in the region. Plate motion model values are taken from the UNAVCO plate motion calculator at http://sps.unavco.org/crustal_motion/dxdt/model. Absolute plate motion (model NUVEL1A in no-net-rotation frame) is shown on the right panel. Relative plate motion between Africa and Arabia is shown by red arrows on the left panel (values from Global Strain Rate Map project, Kreemer et al., 2003). The direction of relative motion across the DSR from continuous GPS observations (Prawirodirdjo and Bock, 2004) is shown by green arrows. This later estimate takes into account independent motion of the Sinai block. Red arrows on both plots show motion on the order of 30 mm/yr, while green arrows (GPS-derived motion) show direction of motion on the order of 1 mm/yr.
Fig. A1. Values of birefringence estimates for individual phases plotted as a function of event back-azimuth. Observations for each station are given distinct symbols: EIL (square); JER (circle); HIT (upright triangle); KSDI/MRNI combination (inverted triangle). Errors are from the cross-correlation algorithm of Levin et al., 1999. Fast polarization values are shown in (a) and delay values are shown in (b). Note the smooth and rapid change in value of fast polarization as a function of back-azimuth, an effect that arises from complex anisotropic structure at depth.
Table 1.
Results of the inversion for one and two layer models
Values for two layer models are those of the model with a smallest misfit value in the entire search. In the two-layer models layer 2 is on top. See Appendix A for the discussion of relative quality of these solutions.
Table A1.
Seismic stations used in the study
Table A2.
Earthquakes used in this study
Table A3.
Splitting measurements using a cross-correlation algorithm of Levin et al. (1999)
Polarization of the particle motion “corrected” for the effect of birefringence is given in the fourth column. Stars denote events used in group inversions, except for site HIT where all observations were used for the group solution. Error estimates for individual measurements are derived using the shape of the cross-correlation surface, see Levin et al. (1999) for description.
Table A4.
Data fit measures for one- and two-layered models (E1 and E2, respectively), degrees of freedom (M) and F-test critical values
Corresponding author. Tel.: +1 732 445 5415; fax: +1 732 445 3374.
Abstract
Purchase PDF (809 K)
E-mail Article
Add to my Quick Links
Cited By in Scopus (3)
Purchase PDF (1736 K)
