T12A-01 INVITED
On the feedback between forearc morphotectonics and megathrust earthquakes in subduction zones
An increasing number of observations suggest an intrinsic relationship between short- and long-term deformation processes in subduction zones. These include the global correlation between megathrust earthquake slip patterns with morphotectonic forearc features, the historical predominance of giant earthquakes (M > 9) along accretionary margins and the occurrence of (slow and shallow) tsunami earthquakes along erosive margins. To gain insight into the interplay between seismogenesis and tectonics in subduction settings we have developed a new modeling technique which joins analog and elastic dislocation approaches. Using elastoplastic wedges overlying a rate- and state-dependent interface, we demonstrate how analog earthquakes drive permanent wedge deformation consistent with the dynamic Coulomb wedge theory and how wedge deformation in turn controls basal "seismicity". During an experimental run, elastoplastic wedges evolve from those comparable to accretionary margins, characterized by plastic wedge shortening, to those mimicking erosive margins, characterized by minor plastic deformation. Permanent shortening localizes at the periphery of the "seismogenic" zone leading to a "morphotectonic" segmentation of the upper plate. Along with the evolving segmentation of the wedge, the magnitude- frequency relationship and recurrence distribution of analog earthquakes develop towards more periodic events of similar size (i.e. characteristic earthquakes). From the experiments we infer a positive feedback between short- and long-term deformation processes which tends to stabilize the spatiotemporal patterns of elastoplastic deformation in subduction settings. We suggest (1) that forearc anatomy reflects the distribution of seismic and aseismic slip at depth, (2) that morphotectonic segmentation assists the occurrence of more characteristic earthquakes, (3) that postseismic near-trench shortening relaxes coseismic compression by megathrust earthquakes and thus reduces tsunami earthquake risk in accretionary settings and (4) that permanent coastal shortening allows adjacent segments to fail more synchronized thus triggering much greater earthquakes in accretionary settings.
T12A-02
Crustal Structure of the Sunda-Banda Arc Transition: Linking Forearc Deformation and Lower Plate Variability
The Sunda-Banda arc transition, the easternmost portion of the Indonesian convergent margin, presents a
probably unique natural laboratory to study lower plate variability and related upper plate deformation in the
so-called 'subduction factory' for a deeper understanding of forearc evolution. In neighboring margin
segments, we can observe strong changes of the incoming plate (transition from an oceanic to a continental
lower plate, increasing plate age to the East, presence/absence of an oceanic plateau, variability in plate
roughness) as well as a wide range of corresponding forearc structures, including large sedimentary basins
and an accretionary prism/outer arc high of variable size and shape. During RV Sonne cruise SO190 in 2006
(SINDBAD: Seismic and Geoacoustic Investigations along the Sunda-Banda Arc Transition), we acquired a
combination of seismic wide-angle OBH/OBS refraction, multichannel streamer and gravity data in order to
study the seismic velocity structure of the subducting crust and the overriding island arc along four trench
normal corridors located between 113 °E and 121 °E. Additionally, a number of trench
parallel profiles were conducted which mainly focus on the internal structure of the large sedimentary basins
and which were also intended for further clarifying the type of underlying forearc crust and mantle
respectively.
We used a tomographic approach for refracted and reflected phases to obtain seismic velocity models which
again were used for prestack depth-migration of the MCS data (see poster of Lueschen et al. in this session).
In turn, we incorporated the highly resolved sedimentary portions as a priori structure in our
tomography. The results show the seismic velocity structure of the incoming plate, starting 100 km seaward
of the trench, and the adjoining forearc down to depths of 20-28 km, i.e. well into the upper mantle, and at
the same time fit the gravity data very well, using simple velocity-density relations. In the Argo abyssal plain,
the models show 8-9 km thick oceanic crust and slightly reduced velocities in the crust and uppermost mantle
within distances of 50 km seaward of the trench. Landward of the trench in the outer arc high, velocities do
not exceed 5.5 km/s down to the top of the subducting slab, which can be traced well beneath the forearc
down to 15 km depth. Offshore Lombok island, our models reveal the geometry of the Lombok basin as well
as the forearc Moho at 16-17 km depth. Reduced upper mantle velocities suggest a hydrated shallow mantle
wedge for this corridor. Further east offshore Sumba island, crustal-type velocities are found down to depths
of 25 km, which would point towards a much thicker forearc crust here. Our results give a detailed view into
the complex deformation in both the deeper and shallower portions of this convergent margin (see also
poster of Shulgin et al. in this session).
T12A-03 INVITED
Transition of accretionary wedge structures around the up-dip limit of the seismogenic subduction zone
Accretive margins are generally divided into three segments: outer and inner wedges and their transition zone. These wedges reflect different aspects of wedge taper, internal deformation, and basal plate boundary fault. The outer wedge is characterized by narrow critical taper, internal deformation by in-sequence-fold and thrust and aseismic decollement. The inner wedge represents a stable narrow taper, weakly deformed internal structure with extensional deformation and seismic plate boundary fault along its base. The transition zone between the two wedges shows large critical taper with steep surface slope, internal structure of out-of- sequence thrusts, and step-down of decollement into the sediment-oceaninc basement interface or switch to a shallower new plate boundary fault bounding the underthrust sediments and hanging wall accretionary prism. These common aspects might be related to the lithification of both accrreted and underthrust sediments and the resultant switch of the plate boundary fault. Deformation and lithification process recorded in exhumed on-land melange of accretionary complexes suggest that the switch of the plate boundary fault occurs around the up-dip limit of seismogenic subduction zone.
T12A-04
Interseismic Deformation and the Mechanical Behavior of Megathrusts: Transient Postseismic Creep, Stress Shadows, and Megathrust Rheology
Geodetic observations of interseismic deformation in subduction zones give a glimpse of the mechanical behavior of megathrust fault-zones (MFZ) as strain is accumulated and released on asperities. Hence, geodetic observations have the potential to constrain not only the geometry of asperities, but also a range of plausible MFZ rheologies. Under the classic elastic back-slip model proposed by Savage [1983], regions of the MFZ creeping at or near the plate convergence rate are assumed to be relatively uncoupled, while regions that are not creeping are presumed to be mechanically coupled, and thus accumulating strain. However, transient creep is frequently observed following large megathrust earthquakes, and for some MFZ rheologies, the postseismic creep can be quite complicated. Large amounts of postseismic creep also results in low creep rates late in a seismic cycle, resulting in regions larger than the asperities that may appear mechanically coupled in a back-slip model. The decrease of creep rates later in the interseismic period are due to stress shadows around the asperities, and the nature of stress shadowing depends on both the geometry of asperities and the mechanical properties of the MFZ. We have developed a 3D model of interseismic creep that is due to repeated earthquakes. This model contains either a viscous or frictional MFZ rheology. In this presentation, we explore the effects of asperity geometries, earthquake history, and the MFZ rheology on the deformation throughout the seismic cycle. We also address constraints on MFZ rheology by simultaneously modeling postseismic and interseismic deformation.
T12A-05
Early Forefront Exhumation as a Consequence of Continental Subduction: Insights From the Northern Apennines of Italy
In the Northern Apennines apatite thermochronology has been used to unravel the exhumation history associated with accretion and retreat. The data indicate that exhumation of the internal part of the chain started at 10-13 Ma, while the core of the Apennines was still thickening and began to be exhumed only later at 8 Ma. The frontal part of the chain was too shallowly buried to completely reset the fission-track system. Here we present new structural and thermochronological data from a late Oligocene-early Miocene unit involved in the shallow part of the plate boundary. This unit is coeval with the initiation of the foredeep indicating ongoing continental subduction, which followed the deactivation of the accretionary prism, occurred in the middle Eocene, and the switch to subduction dominated by tectonic erosion. The studied unit is a tectonic mélange composed by Late Cretaceous to late Oligocene blocks, and crops out 20 km inland from the inferred front of the fossil Apennine subduction complex. Stratigraphic and structural data indicate that the mélange originated by removal of the frontal part of the subduction complex and successive incorporation in the subduction channel. Apatite fission-tracks analysis has been performed on samples from representative mélange blocks. All the samples were exhumed starting at 22-17 Ma, while the rest of the chain shows exhumation times of 12-4 Ma. Hence the exhumation of the mélange occurred at the front of the subduction complex and after short time period from its incorporation in the subduction channel. Our data show that during the early stages of continent-continent subduction two simultaneous and apparently competing mechanisms were particularly active: frontal and basal tectonic erosion leading to the development of a subduction channel, and shallow underplating leading to early exhumation of portions of the channel closely inboard. As both tectonic erosion and internal thickening, the last resulting in underplating, coexist during wedge development, they account for space variations in strain localization and material transfer.
T12A-06 INVITED
A seismological perspective on fluid transport in the Cascadia subduction zone
Cascadia represents an end-member in terms of thermal characteristics for subduction zones. Because it involves one of the youngest and warmest subducted plates on earth, it was predicted that the downgoing plate had to undergo dehydration at unusually shallow depths, yet until recently this hypothesis had not been thoroughly tested. Here, we review seismic imaging results from the past decade showing clearly that along the length of the subduction zone, the subducted crust loses the bulk of its water at depths < 40-50 km as its hydrated metabasalts progressively transform into dry eclogite under increasing pressure and temperature. This depth estimate is supported by geodynamic models, petrological data, and magnetotelluric results, and it contrasts with colder subduction zones where eclogitization is observed at average depths of ~100 km. These results raise interesting questions regarding other observations in the Cascadia subduction zone. First, Cascadia possesses a clearly defined volcanic arc that sits above the ~90 km depth-contour of the subduction interface, which means that some amount of water is entrained at depths greater than that where eclogitization occurs; second, the slab in central and southern Oregon is aseismic despite evidence that it is undergoing dehydration, which suggests that dehydration reactions may not always cause seismicity in the slab.
T12A-07
Western Boundary of the Seattle Uplift, Washington
Lidar topographic scarps and aeromagnetic anomalies define a northeast-striking, en echelon sequence of faults along the southeastern Olympic Peninsula of Washington, all active in Holocene time and possibly linked kinematically with the Seattle fault. The northeast-striking Saddle Mountain fault, the northernmost fault in the sequence, was first recognized in the early 1970s and is now well mapped in the Hoodsport area on the basis of lidar surveys, aerial photography, and trench excavations. Drowned trees and trench excavations demonstrate Holocene deformation on the Saddle Mountain fault approximately contemporaneous with the MW 7.5 Seattle fault earthquake 1100 years ago and with a wide variety of other fault and landslide activity observed over much of the eastern Olympic Peninsula and central and southern Puget Lowland. The northwest-striking Frigid Creek and Canyon River faults, lying 4 km and 27 km to the southwest of Saddle Mountain, respectively, also show evidence of late Holocene deformation. A detailed analysis of aeromagnetic data suggests that the Saddle Mountain fault extends at least 35 km, from 6 km southwest of Lake Cushman to the latitude of the Seattle fault as mapped east of Hood Canal. Regional aeromagnetic data also indicate that the Seattle fault may extend westward across Hood Canal and into the Olympic Mountains, where it terminates near the magnetically inferred northern end of the Saddle Mountain fault. The en echelon alignment of the Saddle Mountain, Frigid Creek, and Canyon River faults, all active in late Holocene time, reflects a >45-km-long, northeast-trending zone of deformation that may accommodate the northward shortening of Puget Lowland crust inboard of the Olympic massif. In this model, the Seattle fault and Saddle Mountain deformation zone form the northern and western boundaries of the northward advancing Seattle uplift.
T12A-08
Looking for serpentine, finding some, within Cascadia megathrust.
Anisotropic texture in earth materials forms as a response to stress applied to materials at depth, with the sense of anisotropy depending on the type of rocks, the pressure and temperature regimes, the presence of volatiles etc. In a subduction zone, the contact region between two converging tectonic plates is expected to contain highly deformed rocks, and thus we expect anisotropy of seismic properties there. We use receiver function analysis of data from long-running seismic stations in Cascadia to probe the lithosphere for the presence of anisotropic layers, which have very distinct signatures (systematic directionally dependent phases on both R and T components). At most sites in the area between the coast and the volcanic arc we are able to identify coherent SV-polarized phases consistent with P-S conversion from the top of the subducting Juan-de-Fuca plate, although timing of these phases does not always agree with expectations from the published models of the top of the subducting Juan-de-Fuca slab. We note that in all cases these phases display significant directional variation, sometimes even switching polarity. We also note that in many cases there is corresponding SH-polarized energy associated with these phases. The significant amplitude of SH components suggests P-S conversion in the presence of seismic anisotropy. We compare receiver function signatures from newly analyzed data sets to that from Corvallis, Oregon, where the plate contact region was previously shown to hold highly anisotropic rock with low seismic wave speeds and high Poisson's ratio, a combination suggesting the presence of serpentinite. We find close similarity at all sites located along the 40 km depth contour of the subduction zone (same as COR). Also, some sites further up-dip (closer to the coast) show evidence of concentrated anisotropy near the top of the subducting slab, albeit less consistent with the velocity profile expected from a strongly serpentinized layer. Of note is the diversity of anisotropic symmetry orientation that we infer from characteristic receiver function features at different sites. Assuming a relationship between the rock fabric that causes anisotropy and the deformation on or close to the Cascadia megathrust, our findings suggest local control of fabric development. The strength of the anisotropic signature at the 40-km depth-contour points to a significant fraction of antigorite serpentine in the layer wedged between the North American and Juan de Fuca plates in Cascadia subduction zone.