STUDY OF CRUSTAL AND UPPER MANTLE STRUCTURE ALONG SELECTED PATHS IN IRAN USING SURFACE WAVE DISPERSION

Abstract

Seismological studies in Iran have been motivated by the need to reduce earthquake hazards. A significant event in this regard was the setting up of three WWSSN stations at Tabriz, Shiraz, and Mashhad. We investigate in this study the dispersion characteristics of Rayleigh waves in the Zagros, Central Iran, and Alborz regions by computing phase speeds using seismograms for the three pairs of stations, namely Tabriz-Shiraz, Shiraz-Mashhad, and Mashhad-Tabriz respectively. Starting from an initial list of 3500 earthquakes data for 9, 8, and 11, earthquakes were found usable for Zagros, Central. Iran, and Alborz region respectively. Fundamental mode Rayleigh waves on all the seismograms were digitized. Calibration pulses on the respective seismograms were also digitized for correction of instrumental phase shifts. Dziewonski and Hales's (1972) cross-correlogram variant of Aki's two station method was used to compute phase speed from these digitized signals. The algorithm was tested extensively through a set of 16 carefully designed experiments on synthetic data. A new method was proposed and tested to take instrumental phase shift into account. The phase speed dispersion curves obtained using extended Rayleigh wave trains in the group speed range of 4.0 to 2.9 km/s were not smooth in spite of all the precautions probably because of noise on seismograms. Smoother phase speed dispersion curves were obtained, through trial and error, using Rayleigh wave train segments corresponding to narrower ranges of group speeds. The latter data were interpreted for crust and upper mantle structures. It is estimated that the maximum error in a phase speed value for a given period i?i any particular case was less than 10.12 km/s . The inversion of phase speed results for crust and upper mantle structure was carried out using the singular value decomposition (SVD) variant of the generalized inverse method. A computer program was written for the purpose independently and tested thoroughly. Rayleigh wave phase speed data for the fundamental mode in the. period range of about 20 to 70 seconds were considered for inversion for the Zagros, Central Iran, and Alborz regions. Limited data in the period range of 70 to 230 seconds were provided by the inversely dispersed waves for the Alborz region. Layered models consisting of 1, 2, and 18 layers for the crust and a uniform half space corresponding to upper mantle were considered with Rayleigh wave data in the 20 to 70 seconds period range. Five and six layered models, with 2 to 3 layers representing the crust, were considered for the inversely dispersed Rayleigh waves. Shear wave speeds and layer thicknesses were determined in all cases through inversion keeping

Id compressional wave speeds and densities of layers constant as initially prescribed. The number of converged model considered for each region was in the range of 60 to 70. The inverted models based on 20 to 70 seconds period Rayleigh waves showed that the crustal thicknesses varied in Zagros, Central Iran, and Alborz regions though shear wave speeds in different layers were broadly similar in the three regions. Crustal thicknesses of 45 km, 39 km, and 35 km were obtained for Zagros, Central Iran and Alborz regions respectively. A low shear wave speed zone in upper mantle of Alborz region is inferred from inversely dispersed Rayleigh waves. The non-uniqueness of the interpretation is acknowledged. Comparison of our results with those by a few other worker using much less data indicate broad similarities and some differences. Inferences regarding crustal thicknesses in Zagros and Alborz regions are supported by limited gravity data and seismological studies using Iranian earthquake data. The 45 km thick crust for Zagros region is consistent with crustal thickening under lateral compression due to northeastward subduction of the Arabian plate beneath Iran.