P波随时间衰减:以两条加利福尼亚地震破裂带为例(英文)
|
摘要
地震预测是地震学研究一个基本目标,而理解地震周期是地震预测的重要环节.遗憾的是,由于地震断层带不但结构复杂,大地震前后短时间内的震源过程更是随时间迅速变化,这些复杂因素使得常规研究方法不再适用.一类特殊的微震成群地发生在主震断层几乎同一地点上,震源系数几乎相同,近年来地震学界通称其为震群(earthquakemultiplet).本文论证如何将这类特殊的微震作为一组人工震源来阐述美国西部两次大地震(LomaPrieta和MorganHill)的震后弛豫过程.在加利福尼亚LomaPrieta(1989年9月18日,M7.0)及MorganHill(1984年4月24日,M6.2)附近的主震破裂带中就发生了此类震群.其中有两组各包含11个重复地震的震群,分别位于这两个破裂带中,其波形资料质量高,波形及震级变化极小,因而可与实验室条件下的人工震源相比.这些震群地震资料可直接用于计算P波沿波程衰减(即t)的时间性变化,其方法的简单性及结果的精确性都与实验室水平相当.我们的结果显示,每个震群周围分别存在一个异常体,在其各自主震之后的10个月中,该异常体中的P波衰减迅速上升并降落.衰减峰值发生在LomaPrieta地震后2～3周,其震群深度为10.2km.此峰值在MorganHill地震后5～6个月发生,其震群中心深度为2.6km.这两个事例中,t波动峰值都超过此前获得的加利福尼亚大致同一地区t绝对值的100%[1].此结果显示了因同震裂隙张开和震后弛豫过程(如流体移动、裂隙复原及岩石压缩),以及孔隙度和流体饱和度在主震源体部分区域发生的变化.衰减达峰值的时间之显著差异则提示我们,裂隙闭合及伴随的流体移动是负载敏感的,深度越大达峰值越快,反之亦然.如将此方法应用于主震前多重地震的资料,将有望更好地理解最终导致主震断层带大破裂的震前微破裂的物理过程.
Within the mainshock rupture zones near Loma Prieta (18 September 1989, M70) and Morgan Hill (24 April 1984, M62), in California, small clusters of nearly identical microearthquakes, called earthquake multiplets, have occurred. A pair of these multiplets, each consisting of 11 repeating events, has produced high-quality waveform data for direct measurements of temporal changes in P-wave path attenuation, or t~*, under near-laboratory conditions. Our results reveal the existence, near each cluster, of an anomalous body within which the attenuation exhibited rapid rise and fall in the first ten months following the respective mainshock. The peak attenuation occurred 2～3 weeks after the Loma Prieta earthquake, where the small cluster of multiplet earthquakes was 102 km deep. The peak occurred 5～6 months after the Morgan Hill earthquake, where the cluster of multiplet earthquakes was centred at 26 km depth. In both cases, the peak amplitude of the t~* fluctuation exceeded 100% of the absolute t~* obtained previously for the same general region in California~([1]). The results point to changes in porosity and fluid saturation taking place within parts of the mainshock source volumes, stemming from coseismic crack opening and postseismic relaxation processes such as fluid migration, crack healing and rock compaction. That the attenuation peaked at very different times is a clue that the crack closure and the accompanying fluid migration are overburden sensitive, being more rapid at greater depths and vice versa.
Within the mainshock rupture zones near Loma Prieta (18 September 1989, M70) and Morgan Hill (24 April 1984, M62), in California, small clusters of nearly identical microearthquakes, called earthquake multiplets, have occurred. A pair of these multiplets, each consisting of 11 repeating events, has produced high-quality waveform data for direct measurements of temporal changes in P-wave path attenuation, or t~*, under near-laboratory conditions. Our results reveal the existence, near each cluster, of an anomalous body within which the attenuation exhibited rapid rise and fall in the first ten months following the respective mainshock. The peak attenuation occurred 2～3 weeks after the Loma Prieta earthquake, where the small cluster of multiplet earthquakes was 102 km deep. The peak occurred 5～6 months after the Morgan Hill earthquake, where the cluster of multiplet earthquakes was centred at 26 km depth. In both cases, the peak amplitude of the t~* fluctuation exceeded 100% of the absolute t~* obtained previously for the same general region in California~([1]). The results point to changes in porosity and fluid saturation taking place within parts of the mainshock source volumes, stemming from coseismic crack opening and postseismic relaxation processes such as fluid migration, crack healing and rock compaction. That the attenuation peaked at very different times is a clue that the crack closure and the accompanying fluid migration are overburden sensitive, being more rapid at greater depths and vice versa.
引文
[1]FletcherJB,BoatwrightJ.SourceparametersofLomaPrieta aftershocksandwavepropagationcharacteristicsalongtheSan Franciscopeninsulafromajointinversionofdigitalseismograms. Bull.Seismol.Soc.Am.,1991,81(5):1783～1812
[2]BaischS,BokelmannGHR.Seismicwaveformattributesbefore andaftertheLomaPrietaearthquake:Scatteringchangenearthe earthquakeandtemporalrecovery.J.Geophys.Res.,2001,106 (B8):16323～16337
[3]ChunKY,HendersonGA,LiuJS.TemporalchangesinPwave attenuationintheLomaPrietarupturezone.J.Geophys.Res., 2004,109:B02317
[4]GellerRJ,MuellerCS.Foursimilarearthquakesincentral California.Geophys.Res.Lett.,1980,7:821～824
[5]NadeauRM,FoxallW,McEvillyTV.Clusteringandperiodic recurrenceofmicroearthquakesontheSanAndreasfaultat Parkfield,California.Science,1995,267:503～507
[6]SchaffD,BerozaGC.Coseismicandpostseismicvelocitychanges measuredbyrepeatingearthquakes.J.Geophys.Res.,2004, 109:B10302
[7]WaldhauserF,EllsworthWL.Adouble differenceearthquake locationalgorithm:MethodandapplicationtotheNorthernHayward Fault,California.Bull.Seismol.Soc.Am.,2000,90(6): 1353～1368
[8]RietbrockA.Pwaveattenuationinthefaultareaofthe1995Kobe earthquake.J.Geophys.Res.,2001,106(B3):4141～4154
[9]ScherbaumF.Combinedinversionforthethree dimensionalQ structureandsourceparametersusingmicroearthquakespectra.J. Geophys.Res.,1990,95(B8):12423～12438
[10]AndersonJC,HoughS.AmodelfortheshapeoftheFourier amplitudespectrumofaccelerationathighfrequencies.Bull. Seismol.Soc.Am.,1984,74(5):1969～1994
[11]ZhuT,ChunK Y.WestGF.High frequencyPwaveattenuation determinationusingmultiple windowspectralanalysismethod.Bull. Seismol.Soc.Am.,1989,79(4):1054～1069
[12]LeesJM,LindleyGT.Three dimensionalattenuationtomographyat LomaPrieta:Inversionoft forQ.J.Geophys.Res.,1994,99 (B4):6843～6863
[13]HoughSE,SeeberL,Lerner LamA,etal.EmpiricalGreen’s analysisofLomaPrietaaftershocks.Bull.Seismol.Soc.Am., 1991,81(5):1737～1753
[14]DodgeAD,BerozaGC.Sourcearrayanalysisofcodawavesnear the1989LomaPrieta,California,mainshock:Implicationsforthe mechanismofcoseismicvelocitychanges.J.Geophys.Res., 1997,102(B11):24437～24458
[15]BoslWJ,NurA.Aftershocksandporefluiddiffusionfollowingthe 1992Landersearthquake.J.Geophys.Res.,2002,107(B12): 2366
[16]BerozaGC,ColeAT,EllsworthWL.Stabilityofcodawave attenuationduringtheLomaPrieta,California,earthquake sequence.J.Geophys.Res.,1995,100(B3):3977～3987
[17]SegallP,RiceJR.Dilatancy,compaction,andslipinstabilityofa fluid infiltratedfault.J.Geophys.Res.,1995,100(B11): 22155～22171
[18]ByerleeJ.Modelforepisodicflowofhigh pressurewaterinfault zonesbeforeearthquakes.Geology,1993,21:303～306
[19]WinklerKW,NurA.Porefluidsandseismicattenuationinrocks. Geophys.Res.Lett.,1979,6:1～4
[20]WinklerKW,NurA.Seismicattenuation:Effectsofporefluids andfrictionalsliding.Geophysics,1982,47(1):1～15
[21]AkiK,RichardsPG.QuantitativeSeismology:Theoryand Methods.NewYork:W.H.Freeman,1980.819～825
[22]ZhuW,WongT F.Networkmodellingoftheevolutionof permeabilityanddilatancyincompactrock.J.Geophys.Res., 1999,104(B2):2963～2971
[2]BaischS,BokelmannGHR.Seismicwaveformattributesbefore andaftertheLomaPrietaearthquake:Scatteringchangenearthe earthquakeandtemporalrecovery.J.Geophys.Res.,2001,106 (B8):16323～16337
[3]ChunKY,HendersonGA,LiuJS.TemporalchangesinPwave attenuationintheLomaPrietarupturezone.J.Geophys.Res., 2004,109:B02317
[4]GellerRJ,MuellerCS.Foursimilarearthquakesincentral California.Geophys.Res.Lett.,1980,7:821～824
[5]NadeauRM,FoxallW,McEvillyTV.Clusteringandperiodic recurrenceofmicroearthquakesontheSanAndreasfaultat Parkfield,California.Science,1995,267:503～507
[6]SchaffD,BerozaGC.Coseismicandpostseismicvelocitychanges measuredbyrepeatingearthquakes.J.Geophys.Res.,2004, 109:B10302
[7]WaldhauserF,EllsworthWL.Adouble differenceearthquake locationalgorithm:MethodandapplicationtotheNorthernHayward Fault,California.Bull.Seismol.Soc.Am.,2000,90(6): 1353～1368
[8]RietbrockA.Pwaveattenuationinthefaultareaofthe1995Kobe earthquake.J.Geophys.Res.,2001,106(B3):4141～4154
[9]ScherbaumF.Combinedinversionforthethree dimensionalQ structureandsourceparametersusingmicroearthquakespectra.J. Geophys.Res.,1990,95(B8):12423～12438
[10]AndersonJC,HoughS.AmodelfortheshapeoftheFourier amplitudespectrumofaccelerationathighfrequencies.Bull. Seismol.Soc.Am.,1984,74(5):1969～1994
[11]ZhuT,ChunK Y.WestGF.High frequencyPwaveattenuation determinationusingmultiple windowspectralanalysismethod.Bull. Seismol.Soc.Am.,1989,79(4):1054～1069
[12]LeesJM,LindleyGT.Three dimensionalattenuationtomographyat LomaPrieta:Inversionoft forQ.J.Geophys.Res.,1994,99 (B4):6843～6863
[13]HoughSE,SeeberL,Lerner LamA,etal.EmpiricalGreen’s analysisofLomaPrietaaftershocks.Bull.Seismol.Soc.Am., 1991,81(5):1737～1753
[14]DodgeAD,BerozaGC.Sourcearrayanalysisofcodawavesnear the1989LomaPrieta,California,mainshock:Implicationsforthe mechanismofcoseismicvelocitychanges.J.Geophys.Res., 1997,102(B11):24437～24458
[15]BoslWJ,NurA.Aftershocksandporefluiddiffusionfollowingthe 1992Landersearthquake.J.Geophys.Res.,2002,107(B12): 2366
[16]BerozaGC,ColeAT,EllsworthWL.Stabilityofcodawave attenuationduringtheLomaPrieta,California,earthquake sequence.J.Geophys.Res.,1995,100(B3):3977～3987
[17]SegallP,RiceJR.Dilatancy,compaction,andslipinstabilityofa fluid infiltratedfault.J.Geophys.Res.,1995,100(B11): 22155～22171
[18]ByerleeJ.Modelforepisodicflowofhigh pressurewaterinfault zonesbeforeearthquakes.Geology,1993,21:303～306
[19]WinklerKW,NurA.Porefluidsandseismicattenuationinrocks. Geophys.Res.Lett.,1979,6:1～4
[20]WinklerKW,NurA.Seismicattenuation:Effectsofporefluids andfrictionalsliding.Geophysics,1982,47(1):1～15
[21]AkiK,RichardsPG.QuantitativeSeismology:Theoryand Methods.NewYork:W.H.Freeman,1980.819～825
[22]ZhuW,WongT F.Networkmodellingoftheevolutionof permeabilityanddilatancyincompactrock.J.Geophys.Res., 1999,104(B2):2963～2971
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心