4) have been carefully selected at each station. We considered a time window of 110s, starting 10 s before the P-onset arrival time. Firstly, to broaden the response of short-period instruments into a more useful teleseismic frequency band, the instrument response is denconvolved from the original records.
ZNE components are then rotated into the local LQT ray-based coordinate system (using theoretical back azimuth and incidence angle), in which L shows the direction of P wave incident to the surface; Q is perpendicular to L and T is perpendicular to both L and Q forming the third axis of the right-hand LQT system. To isolate the P-to-S conversions on the Q component, the L component is deconvolved from the Q component. A band-pass filter of 2-10s is applied to the P receiver functions (PRFs). They are stacked after move out correction for reference slowness of 6.4 s/°. P-RFs are sorted by increasing back azimuth. A notable feature, which can be observed underneath all stations, is the presence of a significant sedimentary layer at about 0.2-1.2s delay time. The middle crustal layer at about 2.3-4.0s delay time can be also seen beneath all stations. The most coherent conversion is however the conversion at the Moho boundary arriving between 5.6-6.6 s delay time.
This conversion can be clearly followed in the individual traces as well as in the stacked traces. The minimum arrival time of the Moho converted phase (5.6s) is observed beneath the station MOK located in the northeastern part of the area. Even though, the largest arrival time (6.6 s) is seen beneath the station PAR located in the northwestern part of the region. Moho depths are obtained by using an average crustal P wave velocity of 6.3 km/s and a Vp/Vs ratio of 1.73. This keeps our depth values independent from possible errors of preliminary shear wave velocity models. However, we estimate the deviation to this model to be less than 5%. Therefore, this procedure results in a ±2 km error in the Moho depth determination. For station KAZ, we couldn’t obtain P-RF because of low signal-to-noise ratio. We have used the arrival times of crustal multiples for determination of crustal thickness (H) and Vp/Vs ratio. This was done using Zhu and Kanamori method, which performs a grid search through the H and Vp/Vs space and searches for the largest amplitudes at the predicted times of direct conversions and multiples. We used weight factors of 0.5, 0.25, 0.25 for the Moho conversion and multiples, respectively.
The average Moho depth is 49.5 km and varies between 46.0 km and 56.5km. We observed that Moho depth decrease from north to south. The thinnest crust was found beneath MOK station whereas the deepest crust was observed beneath PAR station. The Shiraz region crust has an average Vp/Vs ratio of 1.73, with higher ratio of 1.74 in MOK station and lower ratio of 1.70 in SHI station. 2D migrated PRF section depth-distance obtained from a profile perpendicular to the strike of Zagros (NE-SW), shows the Moho boundary varies from 45 km to more than 50 km.]]>