Bulletin of the Seismological Society of America, Vol. 94, No. 6, pp. 2380–2399, December 2004 E Plate-Tectonic Analysis of Shallow Seismicity: Apparent Boundary Width, Beta, Corner Magnitude, Coupled Lithosphere Thickness, and Coupling in Seven Tectonic Settings by Peter Bird and Yan Y. Kagan Abstract A new plate model is used to analyze the mean seismicities of seven types of plate boundary (CRB, continental rift boundary; CTF, continental transform fault; CCB, continental convergent boundary; OSR, oceanic spreading ridge; OTF, oceanic transform fault; OCB, oceanic convergent boundary; SUB, subduction zone). We compare the platelike (nonorogen) regions of model PB2002 (Bird, 2003) with the centroid moment tensor (CMT) catalog to select apparent boundary half-widths and then assign 95% of shallow earthquakes to one of these settings. A tapered Gutenberg-Richter model of the frequency/moment relation is fit to the subcatalog for each setting by maximum likelihood. Best-fitting b values range from 0.53 to 0.92, but all 95% confidence ranges are consistent with a common value of 0.61– 0.66. To better determine some corner magnitudes we expand the subcatalogs by (1) inclusion of orogens and (2) inclusion of years 1900–1975 from the catalog of Pacheco and Sykes (1992). Combining both earthquake statistics and the plate- tectonic constraint on moment rate, corner magnitudes include the following: CRB, 0.48ם 0.52ם 0.21ם 0.47ם 0.76ם -Cou . 0.46מand SUB, 9.58 ; 0.22מOCB, 8.04 0.39מCCB,; 8.46 0.21מCTF,; 8.01 ;0.26מ7.64 7.0ם ,CTF 1.4מpled lithosphere thicknesses are found to be the following: CRB,;3.0 ?ם 0.13ם ?ם 11ם ;for strike slip 0.21מfor normal faulting and 0.40 0.09מOSR, 0.13 11מCCB,; 18 ; 4מ8.6 1.1ם 1.1ם ?ם ,at low, medium, and high velocities; OCB 0.5מand 1.6 , 0.5מ1.8 , 7.3מOTF,13 ?ם 13.7ם In general, high coupling of subduction and continental . 10.8מand SUB, 18.0 ; 2.3מ3.8 plate boundaries suggests that here all seismic gaps are dangerous unless proven to be creeping. In general, low coupling within oceanic lithosphere suggests a different model of isolated seismic fault patches surrounded by large seismic gaps that may be permanent. Online Material: Global seismic subcatalogs of shallow earthquakes. Introduction Since the introduction of plate-tectonic theory four de- (e.g., Cascadia and South Chile trenches) and many centu- cades ago, it has been widely expected to provide a basis for ries in plate interiors. quantitative prediction of long-term-average seismicity and The plate model of Bird (2003) was specifically con- seismic hazard. This promise has not been fully realized be- structed to address the first two problems. Instead of 14 cause of several problems: (1) Early models with 12–14 plates, it has 52, allowing most of the complexity of the plates gave seriously oversimplified kinematic predictions in Pacific margin to be described. It explicitly excludes the some of the most seismically active areas (e.g., the southwest most complex regions as “orogens.” This new global model Pacific). (2) Some regions are so complex that it is doubtful allows the third problem to be addressed by use of the er- whether any rigid-plate model can adequately describe them godic assumption: for studies of globally uncorrelated be- (e.g., Tibet Plateau and North Fiji Basin). (3) It is not re- havior, data collected widely in space can substitute for local solved whether seismic coupling is nearly perfect, or highly data collected over long times. By merging the behavior of variable, within the cold frictional seismogenic layer of the all subduction zones or all continental transform faults lithosphere. This question is difficult to answer in any re- worldwide for a century, we may have enough information gional study because the instrumental seismic record only to extract their average seismicity properties with confidence covers one century, whereas there are good indications of limits that are small enough to be useful. seismic cycles lasting five centuries in some plate boundaries The model of Bird (2003) also explicitly classifies each 2380 Plate-Tectonic Analysis of Shallow Seismicity 2381 short plate-boundary step (interval) as belonging to one of www.seismology.harvard.edu/projects/CMT/. This catalog seven classes: continental rift boundary (CRB), continental begins 1 January 1977, and through the cutoff date for this transform fault (CTF), continental convergent boundary study (30 September 2002) it contains 15,015 events with (CCB), oceanic spreading ridge (OSR), oceanic transform centroid depth Յ70 km. Nuclear tests and other explosions fault (OTF), oceanic convergent boundary (OCB), and sub- are excluded from CMT. The threshold moment for com- ഡ 17 ן ס duction zone (SUB). We first consider the reasons why earth- pleteness is Mt 3.5 10 Nm(mt 5.66) overall; for quake epicenters do not lie precisely on plate boundaries and specific years and/or regions, the threshold is sometimes adopt a rule for selecting the apparent seismic half-widths lower (Kagan, 2003). Each solution includes a moment ten- of plate-boundaries. Then, we use a probabilistic approach sor; we make use of the orientations of the principal axes to match shallow earthquakes with particular plate-boundary and the scalar moment but do not consider possible non- segments by searching for the best match between earth- double-couple source components. Ekstro¨m and Nettles quake properties. Then, we collect the earthquakes into eight (1997) extended the CMT catalog by finding moment tensors subcatalogs, corresponding to seven plate-boundary types of 108 large events (including 98 shallow events) in 1976. 18 ן ഡ and plate interiors. We analyze each subcatalog by maxi- If their threshold magnitude is estimated as Mt 3 10 ഡ mum-likelihood methods to determine the parameters of its Nm(mt 6.28) for shallow events, then there are 59 shal- tapered Gutenberg-Richter frequency/moment relation (Ka- low earthquakes in the complete part. gan, 2002a). Finally, we integrate these distributions to es- There are at least three valuable catalogs of the earlier timate the long-term-average moment rate for each class and years of the twentieth century (1900–1975): Pacheco and compare them with plate tectonic model predictions of the Sykes (1992), Triep and Sykes (1997), and Engdahl and Vil- line integrals of relative velocity. From the ratios, we deter- lasen˜or (2002). The first catalog is most convenient for our mine coupled lithosphere thicknesses for each type of plate study because Pacheco and Sykes collected moment deter- boundary. minations from the literature, when available, and scaled 20- sec surface-wave magnitudes to estimated moments for other Data Sets events. Saturation of surface-wave magnitudes should not greatly affect this catalog, since most of the larger events The plate-tectonic model used in this study is named (all but two of the m Ͼ 8.0 events) have independently de- PB2002. Bird (2003) described its construction and format, termined moments. Where multiple-moment estimates were which included explicit mapping of boundary locations. In available, we accept their selection. In addition, the Pacheco addition to the 14 large plates of the NUVEL-1 model and Sykes (1992) catalog provides limited mechanism in- (DeMets et al., 1990), it includes 38 small plates docu- formation (generic type) for 41% of the events, which is mented with recent geodetic, bathymetric, and/or seismic useful in assigning events to the correct tectonic setting. It data. Euler poles are tabulated for estimation of present rela- lists 598 shallow events during 1900–1975 and is complete Ն tive velocities. Model PB2002 does not claim to achieve a for shallow earthquakes with surface-wave magnitude Ms complete description of kinematics on the Earth. Instead, it 7. One peculiarity of the Pacheco and Sykes (1992) catalog formally designates 13 “orogen” regions in which the plate is that it lists the two largest earthquakes of the twentieth model is known to be inaccurate, either due to very many century as occurring only 1 min apart at virtually the same small plates or to nonplatelike behavior. In most phases of location: 19:10 versus 19:11 on 22 May 1960, at 38.05/ and 9.49 ס this study, seismicity of these orogens will be excluded. 38.20Њ S, 73.34/73.50Њ W, 32 km depth, with m However, it will be considered in some cases where we seek 9.64, respectively. The first event apparently refers to the to maximize the sizes of subcatalogs of large earthquakes. low-frequency precursor detected by both Kanamori and Ci- Earthquake sizes are best characterized by scalar mo- par (1974) and Cifuentes and Silver (1989) 15–19 min be- ment M, but moment magnitude m is also a popular scale. fore the main shock. Such fault-creep events (also called In this project, we used the conversion “slow” or “silent” earthquakes) are rarely detected unless they generate tsunamis or precede famous main shocks. מ ס m (2/3)(log10 (M) 9.05) (1) Since it is not possible to include all fault-creep events down to the same moment threshold used for earthquakes, we pre- of Hanks and Kanamori (1979). There is confusion because fer to drop this one slow precursor event from the catalog. some later authors have used slightly different values of the We note that Engdahl and Villasen˜or (2002) treated this constant (Kagan, 2003). Magnitudes quoted in this article event in the same way. (as a convenience for readers) should be converted to scalar moments using (1) only. Apparent Boundary Half-Widths Our primary seismic catalog is the shallow-earthquake subset of the Harvard Centroid Moment Tensor (CMT) cat- To analyze earthquakes by plate setting, it is necessary alog, which was presented in a number of incremental pub- to associate most shallow earthquakes with nearby plate lications (e.g., Dziewonski et al., 1981; Ekstro¨m et al., 2003) boundaries, while classifying the remainder as intraplate and which is presently available in digital form at events.