The variation of Moho depth and VP/VS ratio beneath Isfahan region, using analysis of teleseismic receiver function

Authors

Abstract

Iran is located in a roughly triangular deforming region, consisting of relatively undeformed shield areas to the southwest (Arabia) and northeast and the more recently deformed, though currently inactive, southwest Afghanistan block in the east. The current geological and tectonic setting of Iran is due to the ongoing convergence between the Arabian and Eurasian Plates, which resulted in the formation of the Iranian plateau, mountain building, extensive deformation and seismicity. The deformation involves intracontinental shortening except where the Oman Sea subducts towards the north beneath the southern east of Iran. The edges of the deformation zone are well defined by the distribution of seismicity and the local topography. It is concentrated in the mountain belts along the
SW borders (Zagros), the southern shore of the Caspian Sea (Alborz) and along the NE (Kopeh Dagh) and eastern borders. These belts enclose a series of relatively aseismic and flat blocks.
The Isfahan Seismic Network belongs to IGUT (Institute of Geophysics, University of Tehran), consists of 5 stations, which are located in Isfehan province. The short-period seismographs (SS-1) are connected to the central recording station via telemetry. The recording is performed on an event-triggered basis. Teleseismic data between 2000 and 2007 have been used in this study. More than 200 teleseismic events with magnitudes greater than 5.5 at epicentral distances between 30? and 95? have been used for P receiver function analysis. Then we have been processed all of data by using the P receiver function and Zhu and Kanamori (2000) methods to calculate the Moho depth and VP/VS ratio beneath Isfahan region.
The teleseismic P receiver function method has become a popular technique to constrain crustal and upper mantle velocity discontinuities under a seismic station (e.g. Langston, 1977; Owens et al., 1984; Kind and Vinnik, 1988; Ammon, 1991; Kosarev et al., 1999; Yuan et al., 2000). Telesismic body waveforms recorded at a three-component seismic station contain a wealth of information on the earthquake source, the earth structure in the vicinity of both source and the receiver, and mantle propagation effects. The resulting receiver function is obtained by removing the effects of source and mantle path. The basic aspect of this method is that a few percent of the incident P wave energy from teleseismic events at significant and relatively sharp velocity discontinuities in the crust and upper mantle will be converted to S wave (Ps), and arrive at the station within the P wave coda directly after the direct P wave. Ps converted waves are best observed at epicentral distances between 30° and 95° and are contained largely on the horizontal components. The amplitude, arrival time, and polarity of the locally generated Ps phases are sensitive to the S-velocity structure beneath the recording station. By calculating the time difference in arrival of the converted Ps phase relative to the the direct P wave, the depth of the discontinuity can be estimated using a reference velocity model (in this paper, the IASP91 reference velocity model is used). After rotating the coordinate system into a local LQT (P-SV-SH) recording system, in which the L component is in the direction of the incident P wave, the Ps energy is mostly observed on the Q component perpendicular to the L component. The Q components (P receiver functions) contain Ps converted waves as well as related S type multiples. To obtain the P receiver function, the following steps are generally used.
Restitution, To utilize data recorded at different types of seismometers, the instrument responses have to be deconvolved.
Rotation, Firstly, the two horizontal components N and E are rotated to radial (R) and tangential (T) directions. Most of the energy of the direct P and Ps waves are dominating the Z and R components, respectively. To isolate the converted Ps wave from the direct P wave, the ZRT components are rotated into an LQT (P-SV-SH) ray-based coordinate system, in which the L component is in the direction of the incident P wave; the Q component is perpendicular to the L component and is positive away from the source; the T component is the third component of the LQT right hand system.
Deconvolution, To eliminate the influence of the source and ray path, an equalization procedure is applied by deconvolving the Q and T component seismograms with the P signal on the L component (Yuan et al., 2000, 2002). The resulting Q component data are named P receiver functions and are mainly composed of the P-to-S converted energy and contain information on the structure beneath a seismic station. The arrival time of the converted Ps phase in receiver functions depends on depth of the discontinuity, whereas the amplitude of the converted phase depends on the S-wave velocity contrast across the discontinuity.
Moveout correction (distance equalization), The converted Ps phases are usually weak and of low amplitude. In order to increase signal-to-noise, it is necessary to align and stack receiver functions from different epicentral distances at each station. However, successful alignment and constructive summation of conversion phases requires that the receiver functions be equalized in terms of their ray parameters.
Migration, To improve the spatial resolution and convert the delay times into depths, the Ps amplitudes on each receiver function can be back projected along the ray path onto the spatial locations of the conversion points to their true locations in a process similar to migrating in exploration seismology (Kosarev et al., 1999). The ray paths are calculated using a one dimensional global velocity model (IASP91) with assumption that conversions are produced from planar interfaces. Sometimes a spatial smoothing filter is used to improve the spatial correlation so that the space is gridded and back projected amplitudes originating from adjacent boxes are stacked to improve signal to noise ratio.
Estimation of crustal thickness and Vp/Vs ratio, The converted Ps phase and crustal multiples (PpPs, PpSs and PsPs) contain a wealth of information concerning the average crustal properties such as the Moho depth and the Vp/Vs ratio.
We compute P receiver functions (PRF) for all stations. We rotate the ZNE-component waveforms into the local LQT ray-based coordinate system and deconvolved the L component from the Q component to isolate the P-to-S conversions on the Q component. Individual and summed PRF for PIR station are presented in Fig 2(Up) and VP/VS ratio base on Zhu and Kanamori (2000) method is plotted in Fig 2(Bottem) as an example.
P receiver function analysis of recorded events between 2000 and 2007 by 5 short period stations from the Isfehan Seismic Network shows clear conversions from the crust mantle boundary beneath the Isfehan region and VP/VS ratio. We have been able to present clear images from the Moho at depths ranging from 38.5 to 43 km beneath the Isfahan region and VP/VS ratio ranging 1.71 to 1.79. The average Moho depth and Vp/Vs ratio are achieved 40 km and 1.74 which confirmed previous results obtained by other methods.

Keywords