Teleseismic Body Wave Analysis of the 1988 Armenian Earthquake
Total Page:16
File Type:pdf, Size:1020Kb
GEOPHYSICALRESEARCH LETTERS, VOL. 16, NO. 12, PAGES1425-1428, DECEMBER1989 TELESEISMICBODY WAVE ANALYSIS OF THE 1988 ARMENIAN EARTHQUAKE J.F. Pachecol,2, C. H. Estabrookl,2, D. W. Simpson2 and J. L. N•b•lek3 Abstract.Long-period and broadband body waves from 14 50 ø digital seismicstations are usedto investigatethe rupture processof the December7, 1988 earthquakenear Spitak, i'......... " ' .--.GC Armenia, USSR. The inversionof thesedata gives the 40 ø followingcentroidal source parameters: strike 299 ø, dip 64ø, rake 151ø, depth6.3 km and seismicmoment 1.5x1019 Nm, .,!-•,".<• -, indicatingthat on averagethe earthquakehad a strike-slip 30 ø mechanismwith a substantialreverse component.The broadband waveforms, however, show significant complexity;they are bestfit with a sourcemodel that includes 20 ø 40 ø 60 ø three sub-events,very similar in size, but with distinctfocal mechanismsand locations. Rupture apparently initiated as a shallowreverse fault at a point of maximumbending on a right-lateralstrike-slip fault, andthen extended bilaterally, 41.0 first towards the southeast and then towards the west. This interpretationagrees with the aftershockdistribution and fault lineationsobserved on LANDSAT images. 40.9 Introduction 40.8 Focal parametersof the devastatingearthquake of December7, 1988near Spitak, Armenia, USSR, reported by 40.7 the National EarthquakeInformation Center (NEIC) are: 43.8 43.9 44.0 44.1 44.2 44.3 44.4 44.5 Ms=6.8, origin time 07h 41m 24.9s, latitude 40.987ø N, Fig. 1. Top figure shows NEIC location of the 1988 longitude44.185 ø E, and 10 km depth.Extensive casualties Armenianearthquake (star). Main plateboundaries and faults anddamage occurred at the townof Spitak,just southof the aredrawn on the map as solidand dashed lines. Shaded areas epicenter,and at the city of Leninakan,near the westernend areelevations higher than 2 km. Largearrow shows direction of the aftershockzone (Figure1). The rupturereached the of motionof the Arabianpeninsula with respectto Eurasia. surfacesouth of Spitakalong an approximately8-km-long, LC = Lesser Caucasus,GC = Greater Caucasus,AP = west-northwesttrending fault segment(Figure 1). The Arabianplate, BM = Bitlismassif, ZM = Zagrosmountains, displacementsalong the surfacetrace show reverse faulting T = Turkey.Bottom figure shows preliminary epicenters of steeply-dippingto the north with a large right-lateral aftershocksrecorded by theU.S. Geol.Survey and Lamont- component:the maximum measured slip was vertically 1.6 m DohertyGeol. Obs. (quality a andb locations).Epicenter of .andlaterally 0.5 m [Sharp,1989]. TP-: aftershock epicenters mainevent reported by NEIC is indicatedby the star, triangle [Simpsonet al., 1989] appearto divide into two groups, showsthe location from 19 localSoviet stations [Cisternas et probablydelineating at leasttwo fault segments(Figure 1): al., 1989].Symbol size is scaledto eventmagnitude, with the major segment,oriented roughly east-west,extends magnitudesranging between1 and 4. Dashedline indicates westward from the epicenter of the mainshock;a second mappedsurface rupture. Thin solidlines are main roads. segment extends from the mainshock epicenter to the southeast.The hypocentersextend to a depthof 15 km only in the west,but mostoccur at a depthshallower than 10 km. Bitlis massifand Zagros mountains to the southand along the The earthquakeoccurred along the southernedge of the Lesser and Greater Caucasusto the north. The region LesserCaucasus. In the Lesserand GreaterCaucasus, major between the Lesser Caucasus and the Bitlis massif is an structurestrend west-northwestand fault plane solutions elevatedvolcanic plateau with Quaternaryvolcanic edifices indicateprimarily reversefaulting [Jacksonand McKenzie, alignedin a north-southdirection, suggesting an almosteast- 1984]. However, two moderate events (Mw = 5.6 and 5.8) westorientation of the principaltensional axis [Kazminet al, that occurred in 1978 and 1986 within 50 km of the Armenian 1986].The deformationhere is characterizedby dispersed earthquakeepicenter, show strike-slipmechanisms with P- conjugatestrike-slip faults [e.g., $eng0ret al., 1985]. axesoriented north-south [Dziewonski et al., 1987a, 1987b]. Seismicitystudies of the convergencezone [Jackson and The regional tectonicsis characterizedby north-south McKenzie, 1984, 1988; Ambraseysand Adams, 1989] reveal compressionas the Arabianplate indentsthe Eurasianplate thatonly a smallpercentage of thedeformation takes place in with a convergencevelocity of 30 mm/yr acrossnorthern Iran earthquakes.Large (Ms > 7) historicalearthquakes in the (Figure 1). This collisionresults in the lateral escapeof the regionare scarce, although an earthquake catalog compiled by Turkishand Iranian plates and major thrustzones, along the Ambraseysand Adams[1989] showsseveral devastating earthquakeswith magnitudesless than 7. Long-periodand broadbandbody waves recordedat 14 1Deparm•entof Geological Sciences of ColumbiaUniversity. digital stationsfrom the Global Digital SeismicNetwork 2Lamont-DohertyGeological Observatory. (GDSN) and the IncorporatedResearch Institutions for 3CollegeofOceanography, Oregon State University. Seismology(IRIS) networkare used here to studythe rupture processof theArmenian earthquake. Copyright 1989 by the American Geophysical Union. Data andData Analysis The analysis of the mainshock follows procedures Paper number 89GL03366. describedby Ngb61ek [1984]. We use the P, SH and SV 0094-8276/89/89GL-03366 $03 . 00 wavesrecorded by GDSN and IRIS stations,taking only 1425 1426 Pachecoet al.: TeleseismicAnalysis of the ArmenianEarthquake TABLE1. Crustal Structure Used in Computing COL 5' TOL 285' SLR 196' CHTO 98' TheoreticalSeismograms Thickness Vp Vs Density km km/s km/s g/cm3 Source 1 4.00 2.14 2.35 5 5.60 3.23 2.70 30 6.50 3.75 2.85 half-space 8.10 4.68 3.30 Receiver half-space 6.00 3.46 2.75 sec stationsin a distancerange between30 ø and 83ø from the epicenterto avoid strong,regionally variable, upper mantle Fig. 2. Observed(solid) and theoretical(dashed) broadband arrivalsand core phases.The crustalstructure (Table 1) we seismogramsfor different source models from several assumedin computingtheoretical seismograms is similarto azimuthallydistributed stations. Shown are seismogramsfor thatused by Simpsonet al. [1989] to locatethe aftershocks. Alaska (COL); Spain (TOL); Thailand (CHTO); and South The seismogramsare invertedin a least-squaressense for the Africa (SLR). The numbernext to stationname is its azimuth source model parameters. For long-period data and from the source. 1S shows the effect of first subevent (see intermediatemagnitude earthquakes such as the Armenian Table 2 for source parameters); 2S shows effects of earthquake,a singlepoint source model usually provides an subevents1 and 2; 3S shows combined effect of 3 subevents adequatesource description. In thatcase, the inversion yields (our final model); and BB-PS for broadbandpoint source the average(centroidal) parameters including the fault plane model. solution (strike, dip, and rake of the slip direction), depth, seismicmoment and time function. If requiredby the data, individualpoint sourcesinto our model,each having its own additionalpoint sourcesseparated in spaceand time can be mechanism,time functionand depth. The two point sources introduced,revealing more detail aboutthe ruptureprocess. thatdescribe the later part of the ruptureare allowedto have In this study,we startedwith the long-periodseismograms to arbitrarytime delay and locationwith respectto the first. In determinethe averageproperties of the source,and then the subsequentdiscussion, we will refer to the three point refinedthe model usingthe broadbandrecords. sourcesas subevents,although their contributionsin mostof thewaveforms overlap considerably. Inversion Results The parametersof the three subeventsare summarizedin Table 2 andthe matchesto the completedata set are shownin The results of a series of inversions we performed are Figure 3. The initial event exhibitsprimarily reversemotion summarizedin Table 2. The long-periodmodel (model LP- with a small right-lateral strike-slip component. Its PS) indicatesthat, on average,the Armenianearthquake had a mechanismis in good agreementwith the mappedsurface strike-slip mechanism with a large reverse-faulting rupture.The nodal plane correspondingto the fault plane, component.The mean (centroid) depth is about 6 km, the based on field observations,strikes 299 ø and dips 45 ø seismicmoment is 1.4x1019Nm, and the sourceduration is northward.The mean depthof this part of the ruptureis 3.1 about18 s. When we invert the broadbanddata using a single km and the seismicmoment is 5.9x1018Nm. The rupture point source,we obtain essentiallyidentical results (model builds up gradually and lasts for almost 9 s. The second BB-PS), however, the model does not provide adequate subevent with a moment of 7.5x10 •8 Nm locates 17 km at an matchto the data(Figure 2), indicatingthat the sourceprocess azimuth of 153ø relative to the first subevent (nucleation was more complicated.The main featuresof the data that the point). This subeventruptured a nearly pure right-lateral singlepoint source model cannot adequately reproduce are the strike-slipfault striking 309ø and dipping 90 ø, using the amplitudes in the early portion of the waveforms (e.g., aftershockdistribution (Figure 1) as a guidein choosingthe compareBB-PS and 3S for COL in Figure 2) and the large fault plane.The meandepth is 7.2 km andthe sourceduration amplitudearrival about 10 s after the first motion (e.g., SLR