Larry J. Ruff

Depth of seismic coupling along subduction zones

Underthrusting at subduction zones can cause large earthquakes at shallow depths but it is always accommodated by aseismic deformation below a certain depth. The maximum depth of the seismically coupled zone (or seismogenic zone) is a transition from unstable to stable sliding along the plate interface. We have determined the depth of this stability transition for the circurn-Pacific subduction zones of: Honshu, Kuriles, Kamchatka, Aleutians, Alaska, Mexico, and Chile. These subduction zones have experienced great interplate earthquakes and the aftershock regions are well-located. Depth estimates of interplate events that are located at the downdip edge of the aftershock regions are used to determine the maximum depth of seismic coupling. For an average P wave velocity of 6.7 km s−1 above the plate interface, we find that for most subduction zones the stability transition occurs at 40 ± 5 km depth. There are, however, several exceptions. At the Hokkaido trench junction, where the Japan trench and the Kurile trench intersect, seismic coupling is deep and extends down to 52–55 km. Deep coupling was also found in the Coquimbo region in central Chile. The Mexico subduction zone has shallow coupling: the transition occurs at 20–30 km depth. Previous studies of micro-earthquakes in Honshu, Hokkaido, the Aleutians, and Alaska show that earthquakes within the upper plate extend no deeper than the downdip edge of the coupled zone that we find. Given our measurements of seismic coupling depth, we then explore the mechanism that may determine coupling depth. The concept of critical temperature has been used to explain the depth of seismic coupling in other tectonic environments, thus we first test whether a critical temperature can explain our results. Temperatures at the plate interface are dependent on many variables; but two that are poorly determined are shear stress and radiogenic heat generation. Shear stress has been constrained by inversion of heat flow data. Assuming a crustal radiogenic heat production rate of 3.1 exp−z/8.5 μWm−3 and a constant coefficient of friction, we find two critical temperatures of about 400 ° C and 550 ° C. The lower critical temperature may be characteristic of regions with a relatively thick continental crust and the higher temperature of regions with a relatively thin continental crust. On the other hand, one single critical temperature of about 250 ° C can explain the coupling depths if shear stresses are constant with depth.

Seismicity and the subduction process

There is considerable variation between subduction zones in the largest characteristic earthquake within each zone. Assuming that coupling between downgoing and upper plates is directly related to characteristic earthquake size, we have tested for correlations between variation in coupling and other physical features of subduction zones: the lateral extent and penetration depth of Benioff zones, age of subducting lithosphere, convergence rate, and back-arc spreading. Using linear multivariate regression, coupling is correlated with two variables: convergence rate and lithosphere age. Secondary correlations within the data set are penetration depth versus lithosphere age, and lateral extent versus convergence rate. An important additional correlation is that back-arc spreading is found to be associated with subduction zones where coupling is low (those characterised by small earthquakes). Taken together, the observed correlations suggest a simple qualitative model where convergence rate and lithosphere age determine the horizontal and sinking rates, respectively, of slabs: these parameters influence the seismic coupling in the subduction zone. In the limit of a fast sinking rate and slow convergence rate, back-arc spreading occurs and thereby appears to be a passive process.

Seismic coupling and uncoupling at subduction zones

Seismic coupling has been used as a qualitative measure of the “interaction” between the two plates at subduction zones. Kanamori (1971) introduced seismic coupling after noting that the characteristic size of earthquakes varies systematically for the northern Pacific subduction zones. A quantitative global comparison of many subduction zones reveals a strong correlation of earthquake size with two other variables: age of the subducting lithosphere and convergence rate. The largest earthquakes occur in zones with young lithosphere and fast convergence rates, while zones with old lithosphere and slow rates are relatively aseismic for large earthquakes. Results from a study of the rupture process of three great earthquakes indicate that maximum earthquake size is directly related to the asperity distribution on the fault plane (asperities are strong regions that resist the motion between the two plates). The zones with the largest earthquakes have very large asperities, while the zones with smaller earthquakes have small scattered asperities. This observation can be translated into a simple model of seismic coupling, where the horizontal compressive stress between the two plates is proportional to the ratio of the summed asperity area to the total area of the contact surface. While the variation in asperity size is used to establish a connection between earthquake size and tectonic stress, it also implies that plate age and rate affect the asperity distribution. Plate age and rate can control asperity distribution directly by use of the horizontal compressive stress associated with the “preferred trajectory” (i.e. the vertical and horizontal velocities of subducting slabs are determined by the plate age and convergence velocity). Indirect influences are many, including oceanic plate topography and the amount of subducted sediments. All subduction zones are apparently uncoupled below a depth of about 40 km, and we propose that the basalt to eclogite phase change in the down-going oceanic crust may be largely responsible. This phase change should start at a depth of 30–35 km, and could at least partially uncouple the plates by superplastic deformation throughout the oceanic crust during the phase change.

The rupture process and asperity distribution of three great earthquakes from long-period diffracted P-waves

Moment tensor source mechanisms for the largest twenty-nine shallow earthquakes from 1981 (M_s ≥ 6.5) have been determined using long-period Rayleigh-wave spectra and P-wave first-motions, recorded by the International Deployment of Accelerograms (IDA), Global Digital Seismic Network (GDSN), and World-Wide Standardised Seismic Network (WWSSN). These are compared with the centroid moment tensor mechanisms presented by Dziewonski and Woodhouse, which were obtained in a contrasting manner. In most cases, the source is well represented by a single double-couple mechanisms, with good agreement of scalar moment and orientation between the two studies. Apparent mismatches for other cases may be interpreted in terms of source complexity. We emphasise the objectivity of comparing moment tensor elements; apparent discrepancies in the moment and orientation of double-couple fault plane solutions are usually due to differences in the Mxz and Myz moment tensor elements, which are poorly constrained by long-period data.

Do Trench Sediments Affect Great Earthquake Occurrence in Subduction Zones

Seismic energy release is dominated by the underthrusting earthquakes in subduction zones, and this energy release is further concentrated in a few subduction zones. While some subduction zones are characterized by the occurrence of great earthquakes, others are relatively aseismic. This variation in maximum earthquake size between subduction zones is one of the most important features of global seismicity. Previous work has shown that the variation in maximum earthquake size is correlated with the variation in two other subduction zone properties: age of the subducting lithosphere and convergence rate. These two properties do not explain all the variance in maximum earthquake size. I propose that a third subduction zone property, “trench sediments”, explains part of the remaining variance in maximum earthquake size. Subduction zones are divided into two groups: (1) those with excess trench sediments, and (2) those with horst and graben structure at the trench. Thirteen of the 19 largest subduction zone events, including the three largest, occur in zones with excess trench sediments. About half the zones with excess trench sediments are characterized by great earthquake occurrence. Most of the other zones with excess trench sediments but without great earthquakes are predicted to have small earthquakes by the age-rate correlation. Two notable exceptions are the Oregon-Washington and Middle America zones. Overall, the presence of excess trench sediments appears to enhance great earthquake occurrence. One speculative physical mechanism that connects trench sediments and earthquake size is that excess trench sediments are associated with the subduction of a coherent sedimentary layer, which at elevated temperature and pressure, forms a homogeneous and strong contact zone between the plates.

Seismic coupling along the Chilean Subduction Zone

Underthrusting at subduction zones can cause large earthquakes at shallow depths but is accommodated by aseismic creep below a certain depth. The maximum depth of the seismically coupled zone (or seismogenic zone) is a transition from unstable to stable sliding. We have determined the maximum depth of the coupled zone and its variability along the Chilean subduction zone. The maximum depth of seismic coupling is defined by the depth of large (M > 6) underthrusting earthquakes that have occurred at the downdip edge of the coupled plate interface. Earthquake depth is determined with omnilinear waveform inversion of long-period P waves. The statistical uncertainty in the depth is estimated using bootstrapping. Omnilinear inversion formally accounts for the scaling incompatibility between the P waveforms and decreases the uncertainty in the depth estimate. We have found the depths of 27 earthquakes in the time period from 1961 to 1987. Seismic coupling in Chile extends down to 48–53 km. There is a resolvable change in the maximum depth of coupling around 28°S. The region immediately north of this latitude has a coupled zone that extends no deeper than 36–41 km, while south of this latitude coupling extends down to 48–53 km. This transition is not simply related to the oceanic lithospheric age and plate convergence rate but coincides with a change in several geophysical and geological phenomena. With respect to shallower (north) and deeper (south) coupling there is a change from active volcanos to no volcanos, the dip of the deep slab changes from steep to shallow, the thickness of trench sediments changes from thin to thick, and the depth of the oceanic basement changes from deeper to shallower at 28°S. Due to the multiplicity of changes at 28°S there is no clear unique interpretation of what physical mechanism controls the variability in the maximum depth of the seismically coupled zone in Chile.

How good are our best models? Jackknifing, bootstrapping, and earthquake depth

Many techniques have been developed to extract a model from data. In general, these techniques are based on minimization of the misfit between measured data and predicted “data.” The model is connected to the predicted “data” by a physical theory. To know how good the model is, one must evaluate model variance. Since the data variance, or alternatively the misfit, is generally nonzero, model variance is generally nonzero. In many cases, the model is a linear function of the data, and model variance can be estimated by formally mapping the data variance to model space [e.g., Menke, 1984].

Composition and mass flux of sediment entering the world's subduction zones: Implications for global sediment budgets, great earthquakes, and volcanism

Lithologic data compiled from Deep Sea Drilling Project and Ocean Drilling Program sites, when combined with orthogonal convergence rates at convergent plate boundaries, permit quantification of the mass flux of sediment into subduction zones. We have made such calculations for each major sediment component — terrigenous grains, calcium carbonate, opal, and water — for twelve trench systems. Results show that 1.4 × 1015 g/yr of sediment and 0.9 × 1015 g/yr of water enter the trenches in the oceanic sedimentary layer. Most of the entering sediment, 1.1 × 1015 g/yr, is terrigenous; the remainder is more carbonate than opal. For most of geologic time an order of magnitude more sediment enters the ocean than leaves it via subduction. The global sedimentary cycle need be in balance only over an entire Wilson cycle. Comparison of sediment fluxes into trenches with the magnitude of large earthquakes and with the composition of bulk volcanic rock shows no correlation.

The 1957 great Aleutian earthquake

The 9 March 1957 Aleutian earthquake has been estimated as the third largest earthquake this century and has the longest aftershock zone of any earthquake ever recorded—1200 km. However, due to a lack of high-quality seismic data, the actual source parameters for this earthquake have been poorly determined. We have examined all the available waveform data to determine the seismic moment, rupture area, and slip distribution. These data include body, surface and tsunami waves. Using body waves, we have estimated the duration of significant moment release as 4 min. From surface wave analysis, we have determined that significant moment release occurred only in the western half of the aftershock zone and that the best estimate for the seismic moment is 50–100×1020 Nm. Using the tsunami waveforms, we estimated the source area of the 1957 tsunami by backward propagation. The tsunami source area is smaller than the aftershock zone and is about 850 km long. This does not include the Unalaska Island area in the eastern end of the aftershock zone, making this area a possible seismic gap and a possible site of a future large or great earthquake. We also inverted the tsunami waveforms for the slip distribution. Slip on the 1957 rupture zone was highest in the western half near the epicenter. Little slip occurred in the eastern half. The moment is estimated as 88×1020 Nm, orMw=8.6, making it the seventh largest earthquake during the period 1900 to 1993. We also compare the 1957 earthquake to the 1986 Andreanof Islands earthquake, which occurred within a segment of the 1957 rupture area. The 1986 earthquake represents a rerupturing of the major 1957 asperity.

Great earthquakes and subduction along the Peru trench

Subduction along the Peru trench, between 9 and 15° S. involves both large interplate underthrusting earthquakes and intraplate normal-fault earthquakes. The four largest earthquakes along the Peru trench are, from north to south, the 1970 (M~ 7.9) intraplate normal-fault earthquake, and the interplate underthrusting earthquakes in 1966 (M~ = 8.0), 1940 (M = 8) and 1974 (M~ = 8.0). We have studied the rupture process of these earthquakes and can locate spatial concentrations of moment release through directivity analysis of source-time functions deconvolved from long-period P-wave seismograms. The 1966 earthquake has a source duration of 45 s with most of the moment release concentrated near the epicenter. Two intraplate normal-fault events occurred in 1963 (M~ = 6.7 and 7.0), at the down-dip edge of the 1966 dominant asperity. The 1940 earthquake is an underthrusting event with a simple source time function of 30 s duration that represents the rupture of a single asperity near the epicenter. The 1974 earthquake has a source duration of 45—50s and two pulses of moment release. This earthquake has a bilateral rupture with the first pulse of moment release located northwest of the epicenter and the second pulse of moment release located southeast of the epicenter. Both pulses of moment release occur on the northern half of the aftershock area. The 1970 earthquake is one of the largest intraplate normal-fault earthquakes to occur in a subduction zone and has a moment release comparable with many large underthrusting events. The aftershocks for the 1970 earthquake form two distinct clusters, the smaller cluster near the epicenter has focal mechanisms characterized by down-dip tension but the second aftershock cluster, located 80 km southeast of the epicenter, has focal mechanisms characterized by down-dip compression. The P-waves for the main shock can be modeled as two sources with different focal mechanisms and depths similar to the two clusters of aftershocks. The first event has a down-dip tensional focal mechanism and is followed 40 s later by a distinct second event located 80 km southeast of the epicenter with a down-dip compressional focal mechanism and a somewhat shallower depth than the first event. The observable directivity indicates that the second source is located at the second cluster of aftershocks that have down-dip compressional focal mechanisms. The occurrence of both down-dip tensional and compressional focal mechanisms may be explained by extreme ‘unbending’ stresses associated with the anomalous slab geometry. The unusually large size of the 1970 earthquake may also be related to the subduction of the Mendana fracture zone. The historic earthquake record along the Peru trench indicates that the previous event in 1746 was much larger than any of the three underthrusting earthquakes this century. The 1746 earthquake may have ruptured the entire segment in a multiple asperity earthquake. Thus, the mode of rupture along the Peru coast has changed between successive earthquake cycles.