P and S receiver functions observed at stations in Europe INTRODUCTION We study the teleseismic P and S receiver functions at European seismic stations. From the IRIS and ORFEUS databeses, we visually select 3002 teleseismic, moderate-to-large magnitude (i.e.M w ≥ 5.7) events recorded by 255 broadband European stations with high signal-to-noise ratio within the years Corrected for the instrument response, the observed seismograms are cut to length of 80 s (20 s pre-event, 60 s post-event) in the case of P receiver functions (PRF) and to a length of 72 s (60 s pre-event, 12 s post-event) in the case of S receiver function (SRF). We rotate the station/event reference frame (Z/R/T) into the ray coordinate reference frame (L/Q/T or P/SV/SH) where the rotation angle is estimated from a covariance analysis. We use a conventional technique to obtain the receiver functions (Langston 1979; Vinnik 1977). All the components (L, Q and T) are deconvolved by the L component (M component) where the deconvolved Q component (L component) produces PRF (SRF). The receiver function amplitudes are normalized by the zero-lag time amplitude of the deconvolved Q component (L component) of PRF (SRF) (Ammon 1991). The M component obtained by covariance analysis (Farra and Vinnik 2000) represents S-wave polarization incident beneath the station. We use a time domain deconvolution approach (Menke 1984; Sheean et all.1995) to isolate the receiver function. In order to secure the quality of measurements we suppress multiples from deep interfaces, remove outliers and average over many measurements. The outliend analysis procedure yields 1701 L- component SRF and 2103 Q-component PRF. The azimuthal coverage of both PRF and SRF waveforms binned at 20 0 backazimuth intervals is appropriate to analyze the observed receiver functions in terms of both azimuthal and radial anisotrophy beneath most of the seismic stations. CONCLUSIONS The Sp conversions preceding the direct S phase and also the crustal multiples are distinctly observed on the waveforms whereas the Ps phases are partly masked by the crustal multiples. Our solutions indicate that a positive velocity gradient characterizes the upper crust under stations at Europe. S receiver function observations have been shown to have significant power for the investigation of upper mantle velocity structure. We include the S receiver function (SRF) in the inversion procedure, along with the traditional approach that only utilizes P receiver fuctions (PRF).This new approach is shown to resolve the underground velocity structure better than the traditional approach, especially at the upper mantle depths.S receiver functions inherently lack high frequency resolution power due to strong attenuation through mantle propagation.They are also noisy because of interference from adverse phases.P receiver functions are of higher frequency, but are susceptible to complication from interfering crustal multiples.We apply the proposed approach to actual recordings using the seismic stations located in Europe. Similar to the theorical experiments the actual data show that S receiver functions are more advantageous at resolving upper mantle velocity structure. ACKNOWLEDGEMENTS This work is supported by The Scientific and Technological Research Council of Turkey (TUBITAK) (project number 109Y345). REFERENCES Ö. ÇAKIR, A. ERDURAN, E. KIRKAYA, Y. A. KUTLU, M. ERDURAN Nevsehir University, Department of Geophysics, 50300, Nevsehir, Turkey ( CALCULATION OF RECEIVER FUNCTION In this section, we explain the parameters used to describe anisotropic velocity structure and briefly present how to calculate synthetic seismogrsms and receiver functions of seismic body waves propagating into a stratified anisotropic medium. The upper mantle and the crust are assumed to have seismic anisotropy of hexagonal symmetry. The seismic velocity perturbations arising from weak hexagonal anisotropy are written as (Backus 1965). ( α 2 – α 0 2 ) / α 0 2 = A + B cos 2 η + C cos4 η for P wave (β 2 – β 0 2 )/ β 0 2 = D + Ecos2 η ρ is the density for S wave, Where α and β are P – and S – wave anisotropic velocities, respectively and η is the angle between the hexagonal-symmetry axis and the propagation direction of the seismic wave. Parameters with subscript 0 denote the isotropic reference velocities of P and S waves. Dimensionless parameters (B, C,and E) denote the anisotropic velocitiy perturbations, and A and D are the isotropic velocitiy perturbations. ρ = 0.32 α (gr/cm 3 ) d(km) β 0 (km/s)  ABCDE  0 0  0 The layer parameters of the anisotropic model down to 500 km depth are listed in Table 1. d shows layer thickness in km,  is Poisson’s ratio and Φ and Ψ are the azimuth and tilt angle of the symmetrry axis. Other parameters are defined on the left column. Table 1 Anisotropic model structure Figure 1: Theoretical PRF and SRF traces corresponding to the model in Table 1 are shown with respect to backazimuth. Black color traces are anisotropic while red color traces are isotropic obtained with B=E=0. DATA AND METHOD The receiver functions can be obtained from the deconvolution of either P waves (i.e. P receiver functions) or S waves (i.e. S receiver functions). The P receiver functions primarily emphasize the P-to-S (or Ps) conversions at the interfaces beneath the station whereas the S receiver functions primarily emphasize the S-to-P (or Sp) conversions. The receiver functions constrain the velocity discontinuities and travel times whereas the surface waves describe the average velocity in the medium, i.e. both data sets complement each other. The P receiver functions, which are more frequently utilized, may have certain disadvantages because of the reverberations (or multiples) that can mask the Ps conversions (e.g. Wittlinger et al. 2004). This masking effect could be particularly important for the Ps conversions from the interfaces in the upper mantle (or mantle lithosphere) arriving almost simultaneously with the crustal multiples. If the sedimentation beneath the station is significant, then those strong sedimentary multiples can even mask the Moho Ps conversion. Figure 2 : Global distrubition of the events used in the P and S receiver function study for station KEV. The 935 events for the P-receiver functions are shown as blue circles and 935 events for the S receiver functions are shown as red circles. The green triangle indicate the location of station KEV. Contours are shown for every 30 0 distance from the centre of station KEV. For our receiver function analysis, we use 3002 teleseismic, moderate-to-large magnitude (i.e.M w ≥ 5.7) events recorded by 255 broadband European stations within the years For the P-wave and S-wave receiver function analysis, epicentral distances ranging from to were used. The azimuthal coverage of both PRF and SRF waveforms binned at 20 0 backazimuth intervals is appropriate to analyze the observed receiver functions in terms of both azimuthal and radial anisotrophy beneath most of the seismic stations. Figure 3: P and S receiver functions are shown. In the upper panels are shown P-wave receiver functions (PRF) of Q and T components at 40 0 backazimuth for station ARSA. In the lower panels are shown S-wave receiver functions (SRF) of L and T components at 60 0 backazimuth for station AQU. Various color receiver functions other than green are discarded because of falling outside the error bounds defined by ± standard deviation. Figure 4: In the upper panels are shown P-wave receiver functions (PRF) of Q and T components at backazimuth for station ARSA. In the lower panels are shown S-wave receiver functions (SRF) of L and T components at backazimuth for station AQU. These and similar RFs are discarded because of uncertainties in the averaging process. Figure 5: Backazimuth dependent PRF results obtained for station ESK are shown. Red traces show the Q component and blue traces show the T component. The green color traces correspond to the joint inversion of Q-component P receiver functions and Rayleigh fundamental mode phase velocities (e.g. Çakır and Erduran 2011). Figure 5: Backazimuth dependent S RF results obtained for station ESK are shown. Red traces show the L component and blue traces show the T component. The green color traces correspond to the theoretical L-component SRFs due to the model structure after the joint inversion of P receiver functions and Rayleigh fundamental mode phase velocities. EGU General Assembly 2012 Ammon,C.,(1991). The isolation of receiver effects from teleseismic P waveforms. Bul. Seismol. Soc. Am. 81, Backus,G. E.,(1965). Possible forms of seismic anisotropy of the upper most mantle under oceans,.J. Geophys. Res. 70, Çakır, Ö.,Erduran, M.,(2011).On the P and S receiver functions used for inversing the one- dimensional upper mantle shear-wave. Surv. Geophys 32, Langston, C. A.,(1979).Structure under Mount Rainier, Washington, inferred from teleseismic body waves.J. Geophys. Res. 84(B9), Menke, W.,(1984). Geophysical Data Analysis:Discrete Inverse Theory, Academic Press, New York. Nagaya, M., et al.(2008).Receiver Functions of Seismic Waves in Layerd Anisotropic Media: Application to the Estimate of Seismic Anisotropy, BSSA,98, 6, Sheehan, A., et.al.(1995). Crustal thickness variations across the Colorado Rocky Mountains from teleseismic receiver functions, J. Geophys. Res.100,20,391-20,404. Snoke. J.A.,(2009),Traveltime Tables for iasp91 and ak135, Seismological Research Letters, 80(2), Vinnik, L.P.,(1977). Detection of waves converted from P to SV in the mantle. Phsy.Earth Planet Inter. 15, Wittlinger G. et al.(2004).Lithospheric and upper mantle stratifications beneath Tibet: new insights from Sp conversions. Geophys Res.Lett.31, L doi: /2004GL α 0= β0β0 α 0, β 0 are the isotropic reference velocities of P and S waves