In this study we used seismic data available from the permanent stations in the NW Iran to study seismic activities and crustal structure beneath the Tabriz region. The region of study is situated in the seismically active NW tectonic province of Iran, where detailed seismic information on the crustal structure and seismicity is lacking. The availability of seismic data collected since 1996 by Tabriz permanent short-period seismological network provided us with an unprecedented opportunity to study the seismicity and crustal seismic structure of this region. Our seismic data consisted of waveforms of local events collected by Tabriz seismological network operated by the Institute of Geophysics of Tehran University (IGTU) and two broad-band permanent seismic stations from the Iranian National Seismic Network operated by the International Institute of Earthquake Engineering and Seismology (IIEES). We manually picked the P and S arrival times on all collected waveforms. The events were initially located using a 1-D model used by the Tabriz network for routine daily event location. Of a total number of 728 events of local magnitudes greater than 2 recorded between 1996 and 2006 by the seismic stations, 361 events of azimuthal gaps of less than 180 degrees and time residual of less than 1.0 second were selected for the subsequent analyses.
  We then simultaneously inverted the arrival times of the local events for one dimensional velocity-depth models and event relocation. By considering the non-uniqueness of the inversion process, we inverted the data using a sufficient number of initial models. Finally, the models showing consistency with each other were selected. Some standard tests were performed to make sure that the final models were in agreement with the observed data. These tests revealed that the results of inversion were stable.
  The final best 1-D model is a 5-layer model to a depth of 35 km. Since we used local shallow events recorded at a small-aperture local network, the models are not valid for the depths deeper than 35 km.
   The final 1-D model gave P and S velocities smaller than the global average values for the orogens. This was because the study area was an active orogenic belt. On the other hand, the Vp/Vs ratio (which was found to be equal to 1.745 using the Wadati Method) of our 1-D models showed values comparable to the average values for a typical continental crust. This presumably indicates that sources of partial melts, if exist beneath the region, are most likely residing in the lower crust.
We also examined the event distribution in depth in order to delineate the geometry of the active faults. The majority of events occurred at the north of the North Tabriz fault. Therefore, it seems to be a vertical fault with a gentle northward dip and the events frequently occur in a depth range of 10 to 25 km (within the crystalline basement). And also some hidden faults seem to exist, which are not traced on the surface geological maps.
)