T33E-01 INVITED
Seismicity Along the South American Subduction Zone: Large Earthquakes, Tsunamis, and Subduction Zone Complexity
Earthquakes along the shallow South American subduction zone have variable rupture processes, including great (magnitude larger than Mw=8) events all along the margin as well as slow tsunami earthquakes in Peru. In addition, there is evidence of interaction between earthquake rupture processes and subducting plate complexity in this region, and I review here seismicity characteristics along the South American margin in light of subducting plate heterogeneity. Significant complexity exists on the subducting Nazca plate, including fracture zones and ridges such as the large Nazca Ridge. The Nazca Ridge enters the Peru-Chile trench at 15-16°S along Peru and appears to play a role in segmenting the seismic characteristics of the margin. Within the 20th century, large magnitude earthquakes have occurred in the region of the ridge subduction, as well as north and south of the feature, but no earthquake rupture has spanned the feature to adjacent regions, suggesting that the Nazca Ridge segments the margin. In addition, the Nazca fracture zone, a feature that subducts to the south of the Nazca Ridge, may have acted to slow down rupture during the Mw=8.4 Peru earthquake of 2001. Tsunami earthquakes, those events that produce large tsunami with respect to their size and typically have long rupture durations, occurred along the Peru section of the margin in 1960 and 1996. These may also be affected by subduction zone heterogeneity, as one mechanism for these tsunami earthquakes includes rupture through low velocity, near- trench interface materials. Presence of this low velocity material will influence other earthquake ruptures as well, and I will present results that suggest other slow events along the Peru section of the margin to the south of the 2 tsunami earthquakes.
T33E-02
The 2007 Pisco earthquake (Mw=8.0), Central Peru: Preliminary Field Investigations and Seismotectonic Context
This epicentral area of the 2007 Pisco earthquake marks a major transition in the characteristics of the Nazca subduction zone: 1) the megathrust dip angle is shallower (15-20) to the north than to the south (25-30); 2) megathrust earthquakes have distinctly smaller magnitudes and are more fragmented to the north; 3) the distance between the trench and the coastline changes abruptly from ~180km to the north to ~80km to the south. These variations are likely related to the oblique subduction of the Nazca ridge - a major bathymetric high - beneath the continental margin. The effect of the subduction of the ridge is obvious in the morphology and tectonics of the forearc, in particular, around the Paracas Peninsula where late Pliocene marine formations are uplifted and the forearc tectonic regime changes from compression to extension. The geometry of the coastline reflects the sweeping of ridge beneath the margin. The coastline also seems to relate to the mode of slip along the subduction interface: modeling of the available interseismic GPS data shows that the plate interface was locked at depth shallower than about 50km, with the downdip end of the Locked Fault Zone (LFZ) corresponding approximately to the coastline. However, the resolution of GPS data is not sufficient to test the idea that the coastline morphology mirrors in detail the variation of the downdip edge of the LFZ. We investigated evidence for uplift or subsidence along the coast and found that the coastline didn't experience any significant vertical displacement compared to the tide range (~40cm). The coastline approximately corresponds to a pivot line marking the transition from coastal uplift in the south to subsidence in the north, as the distance from the trench increases. This model is consistent with the co-seismic slip distribution inferred from waveform modeling, and with the distribution of aftershocks which suggests that the subduction interface ruptured mainly updip of the coastline. To place further constraints on the coseismic slip distribution, we have collected data on the spatial extent of Tsunami waves which hit the coast both south and north of the Paracas peninsula. Finally, our field surveys have also revealed evidence for active faulting of the forearc. In particular, the production of coseismic pressure ridges, with up to 50cm of vertical throw suggests that the east dipping Puente Huamani thrust fault system was reactivated over a distance of about 20km during this event. However, we didn't find evidence for reactivation of any of the normal faults on the Paracas Peninsula, although some had been reactivated by the 2006 Pisco earthquake (Mw6.5). Thus, the structure and deformation of the Peruvian forearc and coastline seems to contain important information on lateral variations of seismic and geodetic coupling along the subduction zone.
T33E-03
The Seismogenic Zone in Southern Chile: Insights from high resolution receiver function analysis and seismic Tomography
Subduction zones, the expression of convergent plate boundaries, generate the world's largest and most destructive earthquakes. The Southern Chilean subduction zone is an ideal natural laboratory to study the processes involved in generating these devastat- ing earthquakes and is one of the main aims of the international and interdisciplinary research initiative TIPTEQ (from The Incoming Plate to megaThrust EarthQuake pro- cesses). High resolution images, using different techniques as well as different physi- cal parameters, form the base for identifying the processes involved. Here we present new data from teleseismic receiver function analysis and 3D seismic tomography to study in detail the down-dip end of the seismogenic zone in the nucleation area of the 1960 magnitude Mw=9.5 Valdivia, Chile, earthquake. Within the project TIPTEQ two dense amphibious passive seismic networks have been installed between Nov. 2004 and Oct. 2005, both covering the entire forearc from the trench to the active volcanic front. The Northern array was located between 37° and 39° South including the epicentre of the 1960 Chile earthquake. It consisted out of 120 continuously recording, three component stations on land and 10 continuously recording Ocean Bottom Seismometers/Hydrophones (OBS/H) at sea. The Southern array was located between 41.5° and 43.5° South roughly in the middle of the rupture zone of the Valdivia earthquake. It consisted out of 20 continuously recording three component stations on land and 20 continuously recording Ocean Bottom Seismometers/Hydrophones (OBS/H) at sea. Several hundreds of micro earthquakes could be located using manual picked P- and S-wave arrivals. Joint 2D/3D inversions for earthquake location, P-wave velocity and vp/vs-ratio were carried out and give a detailed image of the structure as well as a snapshot of the seismicity distribution in both study regions. The subducting Nazca plate can be clearly identified in both regions dipping at a similar angle of about 10-15° in the seismogenic zone and increasing to 30° at greater depth suggesting that the age of the subducting plate does not control the subduction angle. For the Northern region 3D migrated receiver function images depict clearly the subducting Nazca plate (oceanic moho) as well as the continental moho intersecting the subducting plate at about 45 km depth. The seismicity is predominantly localized in a sharp band at the seismogenic interface down to a depth of approximately 35-40 km. At greater depth the seismicity pattern is becoming more diffuse and migrating towards the subducting oceanic moho imaged by the receiver function analysis. This transition between inter-plate and intra-plate seismicity also falls into the area determined for the hypocenter of the 1960 Valdivia earthquake suggesting that this transition is one of the crucial controlling parameters for earthquake nucleation and the down-dip limit of the seismogenic zone.
T33E-04
Slab-Forearc Density Structure and Rigidity Controlling the Seismogenic Behaviour Along the Peru-Chile Subduction Zone
The rupture area and recurrence time of historical earthquakes along the Peru-Chile subduction zone and seismicity recorded by modern networks show a distinctive spatiotemporal distribution defining a characteristic segmentation of the seismogenic zone. It is unclear what factors control this segmentation. Knowledge about this topic is urgent to understand the processes generating devastating subduction earthquakes and to improve its hazard assessment. We are studying this problem for the Peru-Chile subduction zone from two perspectives. First, we applied a wavelet-based spectral isostatic analysis of topography and gravity to compute a high resolution map of the flexural rigidity along the subduction zone. This parameter is a function of the thermo- mechanical structure of both converging plates and the frictional properties of the subduction channel between them. Spatial variations on this map show correlation with the seismogenic segmentation, suggesting that rigidity and the associated physical factors play a fundamental role for the seismogenic behaviour along the margin. Second, we used an existing 3D density model to derive a map of vertical stress acting on the subducting slab below the forearc. This stress is a function of the thickness and density structure of the forearc resulting from long-term geological processes, and is the main component of the normal stress that regulates the magnitude of shear stresses to be released during earthquakes. The spatial variations of vertical stresses show significant correlations with the seismogenic segmentation, implying that the geologically-inherited density structure of the forearc is an important parameter for sustaining a time-persistent seismic segmentation. Of particular interest is the analysis of the giant (Mw 9.5) 1960 Valdivia earthquake, which nucleated in a region of high rigidity and high vertical stress and propagated southward into a region of very low rigidity and vertical stress. This could have important consequence for our understanding of the physical processes and factors interacting along seismogenic zones in general and the Peru-Chile trench in particular
T33E-05 INVITED
Subduction Erosion Processes Along the Northwestern Margin of South America
Subduction erosion is one of the dominant processes that shape convergent margins. Mechanisms favoring subduction erosion occur at both highly- and weakly-coupled margins. Multibeam bathymetry and MCS data collected along the Ecuador-SW Colombia trench show an erosional margin fronted by a narrow wedge of imbricated slope sediment. Ubiquitous arcuate slump scarps on the relatively steep inner trench slope denote frequent slope instabilities along a margin that consists of a trenchward-tilted oceanic basement and fore-arc basin, overlain by slope sediments. PSDM seismic reflection sections across three segments of the Ecuador-SW Colombia margin show that physical conditions enabling basal erosion at the plate interface vary along the margin. In southern Ecuador, seaward of the Gulf of Guayaquil, margin extensional deformation suggests a low- friction plate interface. There, the shallow segment of the subduction channel (SC) is roofed by a strong reflector, and dominated by high excess pore pressure that peaks to 40 MPa. Such pore pressure implies that a permeability barrier prevent fluids from migrating upward. Therefore, breaking the permeability barrier, likely during a megathrust slip, would release over-pressured fluids and allow basal erosion by hydrofracturation In central Ecuador, where the Carnegie Ridge enters subduction, margin extensional deformation indicates a low- friction plate interface. The deeper section of the margin basement gradually thins seaward and disappears ~13 km from the trench indicating basal erosion. PSDM sections image a 3D patchiness across the plate interface implying rapid variations in mechanical coupling and erosion processes. Basement weakening, which results from over pressured fluids in the SC, is marked by enhanced reflectivity at the base of the upper plate. Moreover, at the basement apex, basement breakup is caused by superposition of compressional and extensional fault systems. In northern Ecuador, the seaward section of the 2 km-thick fore-arc basin is sharply tilted trenchward at the inner-trench slope break, thus reflecting subduction erosion of the outer margin wedge. Compressive deformation suggests relatively high-friction plate interface. A strong reflector does not roof the thick, presumably water-rich subduction channel, suggesting that over-pressured fluids are not confined in the SC, but pervasively invade the overlaying outer wedge. Moreover, a crustal splay fault associated with low rock velocities is interpreted as a major conduit for fluid flow. The overall low velocity of outer wedge rocks suggest that they are altered by fluids, and therefore considered weak and easy to break up. Furthermore, diffuse shearing in relation with slip hardening is thought to occur along the upper segment of the megathrust. Therefore, faulting, rock alteration, hydrofracturation and diffuse shearing along the interplate fault would favor basal erosion of the outer wedge.
T33E-06
Morphotectonic, Morphometric, and Critical Taper Analysis of the Offshore Peruvian Continental Slope – Implications for Wedge Mechanics and Submarine Landscape Evolution
We use new swath bathymetry data acquired during the RV Sonne cruise GEOPECO and complement them with swath data from adjacent regions to analyse the morphotectonics of the Peruvian convergent margin. The Nazca plate is not covered with sediments and therefore has a rough surface along the entire Peruvian trench. The styles of roughness differ significantly along the margin with linear morphological features trending in various directions, most of them oblique to the trench and roughness magnitudes of a few to several hundred meters. Oblique convergence and resulting strain partitioning cause a transtensional stress regime in the outer fore-arc. The lower slope is locally very rough and at the verge of failure throughout the entire Peruvian margin, as a result of subduction erosion causing the lower slope to oversteepen. Critical taper analysis is applied to the forearc wedge along several transects. Using curvature attributes to quantitatively examine the morphology in the Yaquina and Mendaña areas revealed that the latter shows a larger local roughness both seaward and landward of the trench however, the amplitude of morphological roughness is larger in the Yaquina area. We identified a 250 km2 large slump on the Lima middle slope. Morphometric dating suggests an age of 74,500 years within 50% error and incision rates on the upper slope are between 0.1 and 0.3 mm per year suggesting that landscape evolution on the Peruvian submarine continental slope is similarly slow than that in the Atacama desert.
T33E-07
Surface cracks as a long-term record of Andean plate boundary segmentation
Meter-scale surface cracks throughout the northern Chilean and southern Peruvian forearcs provide a long-term record of seismic segmentation along the Andean plate boundary. The cracks, mapped on high-resolution satellite imagery, show strong preferred orientations over large regions and the mean strikes of cracks vary systematically as a function of position along the margin. The spatial scale of this variation suggests that stress fields operating with similar dimensions, namely those produced by strong subduction zone earthquakes, are primarily responsible for crack evolution. The orientations of cracks are consistent with the static and dynamic coseismic stress fields calculated for several recent and historical earthquakes on distinct segments of the subduction interface. Field observations indicate that the cracks have experienced multiple episodes of opening and proximal age evidence suggests that they represent deformation as old as several hundred thousand years. We invert the crack orientation data to solve for plausible slip distributions on the Iquique, Chile segment of the margin (19°--23° S), which last ruptured in a M~8--9 event in 1877. We find that concentrations of coseismic slip resolved by the inversion coincide spatially with negative gravity anomalies, consistent with recent studies correlating subduction zone earthquake slip with forearc structure. These results suggest that distinct seismic segments or asperities on the subduction interface define characteristic earthquakes with rupture dimensions and magnitudes that are similar over many seismic cycles.
T33E-08
Climatic controls on drainage basin topography â€" a synopsis of the western Andean flanks between 15.5 S and 41.5 S lat
Topography in tectonically active mountain ranges is determined by the interplay between tectonics and climate. Due to the complexity of natural systems it is difficult to evaluate tectonic versus climatic contributions to the long- term landscape evolution. Previous studies suggest that rainfall and its variability strongly influence the morphology of river profiles and mountain ranges. However, it is still controversially discussed how drainage basins reflect tectonic and climatic processes. The Andean Cordillera provides a unique natural setting for studying the relationship between climate, tectonics, and topography. The Andes host various climatic zones with pronounced differences in rainfall regimes. In the central to southern western Andes, climate ranges from hyperarid in the Atacama Desert, 22 to 23°S lat, with a mean annual rainfall of ~ 5 mm/yr to year-round humid conditions south of Valdivia, ~ 40°S lat, with more than 2500 mm/yr. This zonation is controlled by hemisphere-scale atmospheric circulation patterns. With the exception of a northward shift of the Southern Hemisphere Westerlies during glacials the overall precipitation pattern has remained stable on the west coast of South America. The shelf width is reasonably constant along the margin. Uplift rate and lithology vary non-systematically and do not correlate with climatic parameters. Here, we present an analysis of 120 drainage basins along the watershed of the western Andean flank between 15.5 S and 41.5 S lat, using SRTMV3-90m data and a high-resolution rainfall dataset (TRMM 5x5 km). The basins comprise drainage areas of 1 to ~ 30 x 103 km2 and were split into subsets according to position and size. For each basin, we extracted 21 geometry, relief, and climate parameters in order to unravel the determinants of drainage-basin morphology. Our data shows that river-profile concavity and slope, hypsometric integral, basin maximum and mean elevation decrease with increasing rainfall and descending snowline. Interestingly, our results document that local relief (calculated over a 4.5-km-radius) reaches a maximum of ~ 750 m in a zone between ~ 30° to 35°S lat, which is characterized by a low-frequency, high-magnitude rainfall regime in the transition between arid to semi-arid climate. During glacials this region was affected by the Westerlies and influenced by localized glacial erosion. Relief generation in this region might be enhanced by landslides, flooding events, and debris flows, which quickly dislocate and transport large amounts of material out of the higher-elevated parts of drainage basins. High local relief appears to be preserved due to the absence of continuous rainfall and associated diffusive hillslope processes, preventing sediment production, valley infill, and the coeval smoothing of interfluves. Contrasting, the southern regions between 35° to 40°S lat, receiving higher rainfall amounts, show a lower local relief of ~ 200 m. This might be controlled by a combination of more efficient erosion, sediment production, and accumulation. Our results suggests that the transition zone between 30° and 35°S lat constitutes a transient in landscape evolution which still reflects past climatic conditions. The interaction between sparse, highly episodic rainfall and long periods of aridity appears to significantly facilitate relief preservation and maybe generation. Contrasting, continuous rainfall and the dominance of fluvial processes appear to efficiently smooth relief.