ALBERT

Description: The PUNA (Plateau Untersuchung Nw Argentinien) seismograph network was deployed across the Andes at ∼23.5°S. The array was centered in the backarc, atop the Puna high plateau in NW Argentina and was in operation for approximately 100 days between late August and late November 1997. Most stations were equipped with short-period1-Hz 3-component seismometers and PDAS data loggers recording continuously 100 sps. The deployment was part of the collaborative research center „Deformation processes in the Andes - SFB267”. Waveform data are available from the GEOFON data centre, under network code ZB under CC-BY 4.0 license.

Description: The focus of this study is the high- resolution localization of more than 800 earthquakes in the Northern Chilean Salar Grande region at about 21S within the Andean Costal Cordillera. The events have been recorded by a temporary local network in 2010. We find, that seismicity is not only related to the Nazca slab but also occurs widely scattered within the overlying continental crust. Our highly resolved locations with typical uncertainties below 200 meters image two distinct seismogenic zones at the top and deeper within the mantle of the Nazca slab, as well as the prominent Atacama Fault Zone. The latter could eventually penetrate the entire crust, possibly joining the subduction interface at a depth of about 40 kilometers. In our further investigation, we have applied a waveform cross-correlation approach by which we were able to identify clusters of similar events with respect to location and source mechanism. Within these clusters we took advantage of waveform similarity to further decrease location uncertainties. Most of the crustal seismicity clusters locate on a subvertical planar structure beneath the surface traces of the Atacama Fault Zone, which extends from close to the surface down to the slab. This could indicate that seismicity in the forearc is not only caused by subduction- related deformation, but also by fluid processes. The irregular spatial distribution of the Nazca slab related clusters may be a consequence of topographic variations within the downgoing slab.

Description: Subduction zones around the world show the common pattern of a Double Seismicity Zone, where seismicity is organized in the form of two sub-parallel planes, one at the plate contact and the other one, 10 to 30 km below, in the mantle of the oceanic lithosphere (Lower Seismicity Zone, LSZ). A commonly held hypothesis states that dehydration processes and the associated mineral reactions promote the earthquakes of the LSZ. Fluids filling a porespace strongly alter the petropyhsical properties of a rock. Especially the seismic P- to S-wave velocity ratio (Vp/Vs) has been shown to be sensitive to the presence of fluid-filled porosity. It transforms uniquely to Poisson’s ratio. To test the mineral–dehydration-hypothesis, we use local earthquake data to measure Vp/Vs in the oceanic mantle of the subducting Nazca slab at 21 ◦ S. We determine it as the slope of the de-meaned differential P- vs. S-wave arrivaltimes of a dense seismicity cluster in the LSZ. This measurement yields a value for Vp/Vs of 2.10 ± 0.09, i.e. a Poisson’s ratio of ∼ 0.35. This value clearly exceeds the range of Vp/Vs values expected for oceanic mantle rocks in their purely solid form at ∼ 50km depth. We follow a poroelastic approach to model the rock’s elastic properties, including Vp/Vs, as a function of porosity and porespace-geometry. This results in a porespace model for the target volume having a vein-like porosity occu- pying only a minor volume fraction. Porosity is in the order of 0.1%. These findings are in very good agreement with field surveys and laboratory experiments of mantle dehydration. The pore-geometry is close to the geometrical percolation threshold, where long-ranged interconnectivity statistically emerges, suggesting good draining capa- bilities. Indeed, porosity is soft so that the amount of porosity and, consequently, permeability is very sensitive to local fluid pressure. We conclude that in the oceanic mantle of the subducting Nazca slab, mineral dehydration reactions are contin- uously releasing water into a transient, dynamically evolving vein-system. Permeability is most probably high enough to drain the rock at the rate of metamorphic fluid production.

Description: The ratio of seismic P- to S-wave velocities (the Vp/Vs ratio) of a given rock volume is a sensitive proxy for the detection of fluids and melts. In subduction regimes it has often been inferred from seismic tomography and been used, e.g., to detect pathways of ascending melt above the seismogenic zone, where tomographic methods have their highest resolution. We present Vp/Vs ratios that were computed using only seismic arrival time observations following the approach of Lin and Shearer (2007). This approach has its highest sensitivity in the source volume of a set of nearby seismic events and is hence particularly well suited to directly probe the plate interface. We present data from a temporary local network of short period seismometers that was in operation in the forearc of the Central Andean subduction zone at 21 ◦ S between 2005 and 2012. From this database we were able to localize 3253 seismic events (Ml ∼ 0.5–4) with high precision, yielding a detailed image of the seismicity distribution in this region. Seismicity is pervasive within the entire crust of the South American continental plate and exhibits three distinct bands in the subducting slab, the lowermost one being located in the lithospheric mantle of the subducting plate. The highest concentration of seismic events is found in the contact zone between the continental and the oceanic lithosphere at depths between 30 and 50 km. We group seismic events into approximately 100 subsets of nearby events that origin from the same geo- logical structure. For about half of these subsets we are able to extract a reliable local Vp/Vs ratio. In the middle continental crust, Vp/Vs ratios show slightly enhanced values ( ∼ 1.75). In the lower continental crust towards the plate interface they tend to increase from this value updip and decrease downdip. At the plate interface itself, we observe higher Vp/Vs ratios (〉1.8) at shallower depths (between 20 and 40 km). Downdip (40–60 km depth) Vp/Vs ratios decrease to rather typical values ( ∼ 1.75). The same trend is observed in the lowermost band of mantle seismicity in the subducting slab. Below 80 km depth, where mineral transitions toward the eclogite facies are expected to occur, Vp/Vs ratios tend to be low (〈1.75). The consistently high Vp/Vs ratios in the shallow part of the subducting slab hint at the presence of fluids in the porespace of the subducting lithosphere there. In the deeper part, downdip variations of Vp/Vs may be attributed to mineral phase transitions due to the changing P-T-conditions along the subduction pathway.

Description: This paper presents new insights into the South American subduction zone from reprocessed seismic images. We applied a 3D Kirchhoff prestack depth migration scheme to data sets containing different narrow frequency ranges in order to extract additional details from seismic reflection images. This approach accounts for the effects of scattering on the seismic image, especially for structures below a heterogeneous overburden. The reflection image in such environments will differ significantly when focusing on different frequency ranges due to the frequency dependence of scattering that is likely to be present. The narrow frequency range images uncover reflectors in one frequency range that are masked in another range. Furthermore, the images enable for instance the characterization of the medium in terms of scatterer concentration and thus improve the structural interpretation. The analysis of these images might help to distinguish between small-scale structures in the high-frequency band and large-scale structures in the low-frequency band. We call this imaging approach Reflection Image Spectroscopy (RIS). We applied the RIS approach to the ANCORP'96 data set, an onshore deep seismic reflection profile across the South American Central Andes. The narrow frequency range images revealed additional details that we interpret as features directly linked to fluid migration processes in the subduction zone. Furthermore, structural details of the oceanic crust and the overlying mantle and crust are revealed. From the narrow-frequency range images we interpret that the top of the so called Nazca reflector at 70 km depth marks the upper limit of the hydrated mantle wedge, whereas the bottom of the reflector represents the top of the subducted oceanic crust. The compilation of our results with local earthquake data confirms this interpretation. Similar features observed in another deep seismic profile (PRECORP'95) support this interpretation, too. Furthermore, the RIS images show a highly reflective heterogeneous zone between the Nazca reflector at 70 km depth and a prominent mid-crustal Bright Spot (Quebrada Blanca Bright Spot) at about 30 km depth. We associate this zone with a complex network of ascending fluids or partial melts, initiated by ascending fluids released from the subducted oceanic plate. This observation links the Quebrada Bright Spot directly to the dehydrating oceanic plate.

Description: The coda of passive seismic recordings is often rich in arrivals that are coherent across several stations. If reflections can be extracted, then they may be used for seismic reflection subsurface imaging. With the objective to image the upper crust of the North Chilean Precordillera (Central Andes; approximate location 21°S 69°W), we developed a workflow to process passive seismic data into subsurface reflection images. We analysed the waveform recordings of several hundred microseismic events using signal processing and imaging techniques adapted from active (controlled source) seismic imaging as used in the oil industry. Key processing steps involved precise arrival time picking and hypocentre determination, removing signal amplitude variations due to varying source radiation patterns, identification and separation of reflections from coherent noise, and transformation of the processed waveforms into images of the subsurface reflectivity. When designing our microseismic reflection imaging workflow, we took advantage of the fact that the passive seismic recording geometry with the hypocentres located at depth and the receivers positioned at the surface resembles a reverse vertical-seismic profiling experiment. The resultant P- and S-wave reflection images reveal several reflective features, such as an approximate 15° westward dipping reflector over the 5–25 km depth range that largely coincides with a distinct seismicity boundary. We interpret the imaged interface as the brittle-ductile transition zone boundary, possibly enhanced by a tectonic shear zone. For the area of the North Chilean Precordillera, the deduced microseismic reflection sections with horizontal extensions of about 50 km represent the first high-resolution images of the shallow crust, which could not be obtained from previous active-source seismic-reflection data.

Description: We describe results of an active-source seismology experiment across the Chilean subduction zone at 38.2◦S. The seismic sections clearly show the subducted Nazca plate with varying reflectivity. Below the coast the plate interface occurs at 25 km depth as the sharp lower boundary of a 2–5 km thick, highly reflective region, which we interpret as the subduction channel, that is, a zone of subducted material with a velocity gradient with respect to the upper and lower plate. Further downdip along the seismogenic coupling zone the reflectivity decreases in the area of the presumed 1960 Valdivia hypocentre. The plate interface itself can be traced further down to depths of 50–60 km below the Central Valley. We observe strong reflectivity at the plate interface as well as in the continental mantle wedge. The sections also show a segmented forearc crust in the overriding South American plate. Major features in the accretionary wedge, such as the Lanalhue fault zone, can be identified. At the eastern end of the profile a bright west-dipping reflector lies perpendicular to the plate interface and may be linked to the volcanic arc.

Description: We present a dataset of seismicity from a temporary local network that was installed in the Iquique segment of the northern Chilean subduction zone. The segment has experienced its last activation 135 years ago and is hence expected to be in the late interseismic phase of the earthquake cycle. The dataset exhibits great details and fine structures of the fore-arc subduction system in general and the plate-interface in detail. We performed a state-of-the-art relocation procedure that features a waveform-based correction of arrivaltime pick uncertainties, the incorporation of an independently obtained velocity model, and the application of source-specific station terms to reduce effects of inconsistencies in the latter. This yielded locations of nearly 5,500 events with a mean RMS-misfits as low as 30ms. We find a high downdip-variability in seismic activity along the plate interface. This includes well-defined, platy shaped patches of enhanced seismicity at depths around 35 and 45km, respectively, and a sudden downdip end of seismicity near the tip of the continental mantle wedge. Seismicity at the plate interface correlates tightly with the previously obtained reflectivity image of the down-going slab from the ANCORP’96 experiment. More details are revealed in the direction perpendicular to the slab. Whilst seismicity is highest within a few kilometer thick layer directly at the plate contact, an overlying region of reduced seismicity separates this from the also abundant seismicity within the deforming continental crust of the overriding plate. Apart from seismicity at the plate interface, we also find a second band deeper inside the slab near the oceanic Moho and a well defined third band approximately 15km below the oceanic Moho inside the oceanic lithospheric mantle that streches from a depth of only 40km offshore to 90km near 69.0°W.

Description: We obtained high-precision locations for 5250 earthquakes in the Iquique segment of the northern Chilean subduction zone from two temporary local seismic networks around 21°S. A double seismic zone in the downgoing Nazca slab can be clearly identified. One band of seismicity is located at the plate interface and a second one 20–25 km deeper in the oceanic mantle. It can be traced updip to uncommonly shallow levels of 50 km. A combined interpretation of seismicity and reflectivity along the seismic ANCORP’96 experiment suggests the prevalence of fluid processes in the subducted oceanic crust as well as in the uppermost 20 km of the mantle. Crustal seismicity is pervasive below the Coastal Cordillera. Beneath the Precordillera, the lower bound of crustal seismicity delineates a sharp west-dipping boundary down to 20 km depth, consistent with earlier findings indicating a rheological boundary.