Доклади на Българската академия на наукитеComptes rendus de l’Academie bulgare des Sciences

Tome 66, No 5, 2013

GEOPHYSIQUE

Sismologie

THE MOHO DEPTH AND CRUSTAL STRUCTUREBENEATH BULGARIA OBTAINED FROM RECEIVER

FUNCTION ANALYSIS

Gergana Georgieva, Svetlana Nikolova

(Submitted by Corresponding Member D. Solakov on January 17, 2013)

Abstract

Receiver function technique is applied to evaluate the Moho depth andVP /VS ration on the territory of Bulgaria. The method was applied to elevenstations of the Bulgarian National Seismological Network (BNSN) equippedwith broadband seismometers and the results are presented in this paper.

The results obtained show that the Moho depth varies between 30 and50 km. The crust is shallower in northeast of Bulgaria and goes deeper insouthwest direction. The crust is thicker in the Rhodopean massif, where itreaches more than 50 km. The VP /VS ratio in the study area varies between1.60 and 1.90.

Key words: receiver function analysis, Moho structure, VP /VS

Introduction. In the last decade, receiver function analysis [1–4] became areliable method for mapping of the crustal and upper mantle structure close tothe seismic stations. After the upgrade of the Bulgarian National SeismologicalNetwork (BNSN) to digital data acquisition and storage systems, the receiverfunction technique can be applied to study the discontinuities in the crust onthe territory of Bulgaria and especially the structure and the depth of the Mohoboundary.

Data. The upgrade of the National Operative Telemetric System for Seis-mological Information (NOTSSI) in Bulgaria was completed in the last quarter of2005 [5]. Currently BNSN consists of 14 stations and the recorded data are trans-mitted in real time to the Bulgarian National Data Centre for Automatic andInteractive Data Processing [6]. Data from 11 stations, equipped with broadband(BB) seismometers are used in this study.

Most of the events used for receiver function analysis are recorded in theperiod from January 2006 to July 2010. The analyzed seismic events are ofmagnitudes from 5.5 to 7.5, and with epicentral distances from 30◦ to 95◦. After

725

the removal of noisy traces from stacked seismograms, a set of about 200 tracesfor each station was generated. A good azimuthal coverage was achieved for eachstation.

Method. Receiver functions reflect the reaction of the Earth structuresbeneath seismic station to the coming seismic wave. S wave is generated when theP wave crosses the border between two layers with different velocities. ConvertedS wave has a velocity smaller than the velocity of the P wave and it is comingshort after the P wave in the seismic station. The depth of the discontinuitywhich generated the S waves could be restored if the time difference betweenboth phases is estimated.

The P wave coda from a teleseismic event contains P -to-S wave conversion.The converted S wave is usually weak and it is necessary to remove P waveto enable detection of the converted phase. This can be done by choosing alocal ray coordinate system LQT (P , SV and SH), where: the L component isin the direction of the P wave; the Q component is perpendicular to L and isselected so that the most of the Ps energy is in this direction; the T componentis perpendicular to the LQ plane. In the optimal conditions for the horizontallayered Earth, with the homogenous layers, no Ps energy should be observed inT direction. However, in the real data very often some energy can be observedon the T component. The reason is usually lateral anisotropy in the Earth crustbeneath the seismic station.

Deconvolution of the seismic traces is performed after their rotation intoLQT coordinate system. Deconvolution is applied to the SV component by theincident wave of P component. The resulting receiver functions are sorted byazimuth and staked together. A summed receiver function is computed for eachstation and it is used further for inversion and reconstruction of the velocitymodel.

The depth of the Moho discontinuity is estimated using the method devel-oped by Zhu and Kanamori [7]. The crustal thickness H and VP /VS ratio arecomputed for each event and the results are plotted together in H-κ domaindefined as

s(H,κ) = w1r(t1) + w2r(t2) + w3r(t3),

where t1, t2, t3 are arrival times of the P–S converted phase and following multi-plies (PpPs and PpSs+ PsPs), and w1, w2 and w3 are the weighing factors foreach phase selected, so that

∑wi = 1. Function s(H,κ) reaches its maximum

when all phases are stacked coherently.This method gives reliable results when the Moho discontinuity is flat and

there are no complex structures or strong anisotropy in the crust. The estimatederror of depth in this case is about ±0.5 km.

Results. We computed the receiver functions for 11 stations from the BNSN(Fig. 1) equipped with BB seismometers. Because the receiver function calcula-tions require initial velocity model, we combined two velocity models for the

726 G. Georgieva, S. Nikolova

Fig. 1. Bulgarian National Seismological Network: the triangles show the stations used inthe study; the stars show stations not used in the study; the circles show stations from Local

Seismological Network of Provadiya

purposes of the present study: the velocity model for the Earth crust down to100 km depth, as obtained by Raykova [8], and the IASPEI velocity model fordepth greater than 100 km. The velocity model by Raykova [8] has granularityof 1◦ × 1◦ and therefore for each station a specific initial model was applied forbetter representation of the local structure. We assume as representative theestimates of the thickness of the crust and VP /VS ratio if both values have prob-ability higher than 95%. These H and k correspond to coherently stacked phases(Figs 3c and 4c).

The results from 11 seismic stations can be divided into three groups based ontheir location: Northern Bulgaria (3 stations), Upper Thracian Plain (3 stations)and the region of Rila and the Rhodopes (5 stations).

We could reliably identify two discontinuities in stacked seismograms for thestations in Northern Bulgaria. The first one corresponds to the boundary be-tween sedimentary layer and consolidated crust. A thick sediment layer reachingabout 10 km was delineated in the receiver functions of stations MPE, PVL, SZH(Fig. 2a). The depth of the Moho discontinuity is estimated to about 29–31 km,and VP /VS is in the range of 1.80–1.87 in north and north-west direction fromMPE station. Figure 2 shows receiver functions sorted by azimuth from 0◦ to360◦ and stacked together from each event. In some azimuthal ranges, the phase

Compt. rend. Acad. bulg. Sci., 66, No 5, 2013 727

MPE_Q

sum

–10 0 10 20 30 40

–10 0 10 20 30 40

Delay time (s)

Ps PpPsPpSs

a)

MPE_T

–10 0 10 20 30 40

sum

Delay time (s)

–10 0 10 20 30 40

b)

Fig. 2. Receiver functions for Q (a) and T (b) components ofMPE station

728 G. Georgieva, S. Nikolova

–10 0 10 20 30 40JMB_T

Delay time (s)

–10 0 10 20 30 40

sum

JMB_Q

sum

Delay time (s)

–10 0 10 20 30 40

–10 0 10 20 30 40

Ps PpPsPpSs

a)

b) c)

Fig. 3. Receiver functions for Q (a) and T (b) components ofJMB station; (c) H-κ values for each event are represented as

isolines with different probability

from Moho is missing or comes at different time. For these azimuthal ranges,there is energy on the T component (Fig. 2b), which is indication for existence ofanisotropy and/or complex structures in the crust beneath the station. There-fore, the depth of Moho for these directions is not estimated (Fig. 2b). The Mohodepth beneath PVL station is about 31-32 km (VP /VS is in the interval 1.86–1.90)and the depth beneath SZH station is in the range of 32–34 km with velocity ratioin the range 1.79–1.87.

The results for the Upper Thracian Plain show flat and well defined crustalstructure. From all the stations used in the present study JMB is the only one,where no complex structures or anisotropy indications are observed (Fig. 3a).

8 Compt. rend. Acad. bulg. Sci., 66, No 5, 2013 729

MMB_Q

–10 0 10 20 30 40

sum

Delay time (s)–10 0 10 20 30 40

MMB_T

sum

–10 0 10 20 30 40

Delay time (s)

–10 0 10 20 30 40

Ps PpPs PpSs

a)

b) c)

Fig. 4. Receiver functions for Q (a) and T (b) components ofMMB station; (c) H-κ values for each event are represented as

isolines with different probability

The Moho discontinuity is flat and well defined for all back azimuths around thestation. In this case, distinct results can be obtained using Zhu and Kanamorimethod [7] (Fig. 3b). The depth beneath JMB is in the range of 30–32 km (VP /VSis in the range of 1.77–1.85). The Moho depth beneath PLD station is estimatedbetween 32 and 34 km and the VP /VS ratio is in the range of 1.76–1.83. PGBstation is situated in the highest part of Sredna Gora Mountain. The calculateddepth for Moho discontinuity is 37–39 km and VP /VS is in the range of 1.69–1.77.

The Rhodopean massif occupies the southwestern part of the study area. Theanalysis of the receiver function indicates the Moho depth of 28–33 km beneath theEastern Rhodopes (KDZ station) and the VP /VS ratio in the interval of 1.74–1.87.

730 G. Georgieva, S. Nikolova

The discontinuity sinks down in west direction and reaches about 36 km (34.5–37.5 km and VP /VS 1.73–1.82) in the central part of the Rhodopes (RZN station).The thickest Earth’s crust observed on the territory of Bulgaria is beneath MMBstation. The results indicate very complex crustal structure and the observedenergy on the T component (Fig. 4b) indicates strong anisotropy. The estimatedMoho depth here is about 50 km (49–50 km) (Fig. 4a). Two areas with probabilityhigher than 95% could be identified in Fig. 4c. The results for the first areacorrespond to a Moho depth within the range of 38–40 km and VP /VS between1.78 and 1.83. The second area indicates the range of 49–52 km for the Mohodepth and 1.60–1.65 for the velocity ratio. Previous geophysical studies [9,10]of the region define the depth of Moho at 50 km, and our results confirm theexistence of a well-defined boundary at this depth. The low VP /VS ratio is typicalof regions with felsic rocks, which is the case in the Western Rhodopes, east ofMMB station [11]. Most of the events (170 out of 220 events) used for constructionof the receiver function for MMB are located in the azimuthal range of 0◦–100◦,coinciding with the formation of the felsic rocks. Therefore, for MMB station weaccept that the Moho is at 49–52 km depth with 1.60–1.65 VP /VS . The estimatefor Moho depth beneath KKB station is between 27 and 30 km and the VP /VSratio is in the range of 1.84–1.92. The discontinuity is not well defined in the backazimuthal range between 300◦–45◦. The receiver function analysis is also sensitiveto vertical discontinuities in the medium; therefore a plaussible explanation forthis pattern could be the presence of Krupnik fault penetrating through the wholecrust [12]. The most complex patterns in the stacked receiver function plot areobserved for VTS station. Two peaks are identified as Ps on the summed trace.The Moho depth is shallower in the back azimuth range between 70◦–180◦ and theconverted phase from this direction comes 3.9 s after the main phase. For all otherback azimuths, the Ps phase comes 5 s after P . Furthermore, detailed analysisof these data needs to be performed as the current approach does not providereliable results for the depth of Moho discontinuity if structures are anisotropicand/or there are strong lateral inhomogeneities or faults.

Two more stations – Provadiya (PRD) and Preselentsi (PSN) are locatedin Eastern Bulgaria, close to the Black Sea. The standard analysis of the datafrom these stations could not provide reliable results, due to the high microseismicnoise from the Black Sea. The sea microseisms have frequency maximum at about4–8 s [13], which is within the frequency range of the analyzed receiver functions,and therefore standard suppressing/filtering was affecting the data and results.Several filters were used to suppress the high microseismic noise from the BlackSea but another technique needs to be applied in order to distinguish the phasesfrom Moho.

Conclusions. The structure of the crust in Bulgaria is evaluated usingthe receiver function technique beneath 11 seismic stations on the territory ofBulgaria. The crustal thickness on the territory of Bulgaria varies between 30

Compt. rend. Acad. bulg. Sci., 66, No 5, 2013 731

and 50 km. The depth of the Moho boundary is about 30 km beneath stationsPVL, MPE, SZH, PLD and JMB. The crust is thickening beneath the Rhodopeanmassif and Pirin Mountain, where it reaches about 50 km. The Moho structure inSouthwestern Bulgaria is very complex and further more detailed analysis of datashould be performed. A sedimentary layer with thickness of 10 km was delineatedin Northern Bulgaria (PVL, MPE, SZH).

The VP /VS ratio in Northern Bulgaria varies between 1.80 and 1.90 anddecreases in south and southwest direction. It is between 1.76–1.85 in the UpperThracian Plain and for the Rhodopean massif the range is about 1.72–1.82. Thewestern part of the Rhodopean massif is characterized by low VP /VS (1.6–1.65)and could be explained with the presence of felsic rock formation.

REFERENCES

[1] Vinnik L. P. Phys. Earth Planet. Inter., 15, 1977, 39–45.[2] Kosarev G. L., L. I Makeyeva, L. P. Vinnik. Phys. Earth Planet. Inter., 47,

1987, 11–24.[3] Langston C. A. Bull Seis. Soc. Am., 67, 1977, 1029–1050.[4] Owens T., G. Zandt, S. R. Taylor. J. Geophys. Res., 89, 1984, 7783–7795.[5] Report of Project: Modernization of National Operative Telemetric System for

Seismological Information (NOTSSI), Decision SB-3/04.05.2005, 2005.[6] Haikin L., A. Kushnir. SYNAPSE Science Center, 2005, 122.[7] Zhu L., H. Kanamori. J. Geophys. Res., 105, 2000, 2969–2980.[8] Raykova R. B., S. Nikolova. Stud. Geophys. Geod., 51, 2007, 164–180.[9] Raykova R. B. Structure of the Earth’s Crust and Upper Mantle in Southeastern

Europe from Surface Waves, PhD Thesis, Archive Nat. Inst. Geolody, Geodesy andGeography, Bulgarian Academy of Sciences, 2004, 293 pp. (in Bulgarian).

[10] Dachev Hr. Earth’s Crust Structure in Bulgaria, Sofia, Technika, 1988, 334 pp.(in Bulgarian).

[11] Dobrzhinetskaya L., S. Faryad, S. Cuthbert. Ultrahigh Pressure Metamor-phism: 25 Years after the Discovery of Coesite and Diamond, Amsterdam, ElsevierInc., 2011, 299 pp.

[12] Dobrev N. Review of the Bulgarian Geological Society , 66, 2005, Nos 1–3, 115–124.

[13] Dimitrova L. Analysis and Automation of Seismological Observations of the Dig-ital System of the NOTSSI, PhD Thesis, Archive Nat. Inst. Geolody, Geodesy andGeography, 2010, 199 pp. (in Bulgarian).

Faculty of PhysicsSt. Kliment Ohridski University of Sofia

5, J. Bourchier Blvd1164 Sofia, Bulgaria

e-mail: ggeorgieva@phys.uni-sofia.bg

∗National Institute of Geolody,Geodesy and Geography

Bulgarian Academy of SciencesAcad. G. Bonchev Str., Bl. 3

1113 Sofia, Bulgariae-mail: sbnik@geophys.bas.bg

732 G. Georgieva, S. Nikolova