Andreas Rietbrock

Subduction and collision processes in the Central Andes constrained by converted seismic phases

The Central Andes are the Earths highest mountain belt formed by ocean–continent collision. Most of this uplift is thought to have occurred in the past 20 Myr, owing mainly to thickening of the continental crust, dominated by tectonic shortening. Here we use P-to-S (compressional-to-shear) converted teleseismic waves observed on several temporary networks in the Central Andes to image the deep structure associated with these tectonic processes. We find that the Moho (the Mohorovičić discontinuity—generally thought to separate crust from mantle) ranges from a depth of 75 km under the Altiplano plateau to 50 km beneath the 4-km-high Puna plateau. This relatively thin crust below such a high-elevation region indicates that thinning of the lithospheric mantle may have contributed to the uplift of the Puna plateau. We have also imaged the subducted crust of the Nazca oceanic plate down to 120 km depth, where it becomes invisible to converted teleseismic waves, probably owing to completion of the gabbro–eclogite transformation; this is direct evidence for the presence of kinetically delayed metamorphic reactions in subducting plates. Most of the intermediate-depth seismicity in the subducting plate stops at 120 km depth as well, suggesting a relation with this transformation. We see an intracrustal low-velocity zone, 10–20 km thick, below the entire Altiplano and Puna plateaux, which we interpret as a zone of continuing metamorphism and partial melting that decouples upper-crustal imbrication from lower-crustal thickening.

Complex patterns of fluid and melt transport in the central Andean subduction zone revealed by attenuation tomography

Abstract We present a high resolution 3-D model of P -wave attenuation ( Q p −1 ) for the central Andean subduction zone. Data from 1500, mostly intermediate depth (60–250 km) earthquakes recorded at three temporary seismic networks covering the forearc, arc, and backarc around 23°S were used for tomographic inversion. The forearc is characterised by uniformly high Q p values, indicating low temperature rocks, in accordance with low surface heat flow values. Prominent low Q p anomalies are found beneath the magmatic arc and the backarc in the crust and mantle. Continuous regions of low Q p connect earthquake clusters at 100 km and 200 km depth with zones of active volcanism in the arc and backarc. Fluids fluxed from the subducted oceanic lithosphere into the overlying mantle wedge, where they induce melting, explain our observations. We propose that low Q p regions indicate source and ascent pathways of metamorphic fluids and partial melts. Ascent of fluids and melts, as imaged by seismic Q p , are not vertical, as is often implicitely assumed. Instead, sources of fluids are located at different depth levels, and ascent paths are complex and exhibit significant variation within the study area. The largest Quaternary backarc volcano Cerro Tuzgle is fed by mantle melts which are imaged as a plume of low Q p material that reaches to the strong earthquake cluster at 200 km depth.

The Southern Andes between 36° and 40°S latitude: seismicity and average seismic velocities

Abstract The project ISSA 2000 (Integrated Seismological experiment in the Southern Andes) consists of a temporary seismological network and a seismic refraction profile. A network of 62 seismological stations was deployed across the Southern Andes at ∼38°S. Three hundred thirty-three local seismic events were observed in a 3-month period. P and S arrival times of a subset of high quality data were inverted simultaneously for 1-D velocity structure, hypocentral coordinates and station delays. Seismic refraction data along a transect at 39°S provide further constraints on the crustal structure. Low crustal velocities beneath the forearc may be either due to subducted trench sediments or serpentinized mantle material of the continental lithosphere. The continental Moho is not clearly observed in this region. Average velocities of the crust beneath the arc are higher than those beneath the forearc. Crustal thickness is about 40 km. Crustal seismicity concentrates in the forearc region along the Bio-Bio and Gastre fault zones. The area between these two prominent fault zones seems to be nearly devoid of crustal seismicity but shows highest uplift and topography in the forearc region. Benioff seismicity is observed down to 150 km depth resulting in the first accurate image of the Benioff zone in the Southern Andes. A maximum of seismicity at 60 km depth may be caused by dehydration embrittlement.

Attenuation tomography in the western central Andes: A detailed insight into the structure of a magmatic arc

High-quality data from 1498 local earthquakes recorded by the PISCO ’94 (Proyecto de Investigation Sismologica de la Cordillera Occidental, 1994) and ANCORP ’96 (Andean Continental Research Project, 1996) temporary seismological networks allowed the detailed determination of the three-dimensional (3-D) attenuation structure (Qp−1) beneath the recent magmatic arc in the western central Andes (20° to 24°S). Assuming a frequency-independent Qp−1 in a frequency band between 1 and 30 Hz, whole path attenuation (t*) was estimated from the amplitude spectra of the P waves using spectral ratios and a spectral inversion technique. The damped least squares inversion (tomography) of the data reveals a complex attenuation structure. Crust and mantle of the forearc and subducting slab are generally characterized by low attenuation (Qp > 1000). Crust and mantle beneath the magmatic arc show elevated attenuation. The strongest anomaly of extremely low Qp is found in the crust between 22° and 23°S beneath the recent volcanic arc (Qp < 100). N-S variations can be observed: The western flank of the crustal attenuation anomaly follows the curved course of the volcanic front. North of 21°S the attenuation is less developed. In the northern part of the study area the low-Qp zone penetrates in the forearc mantle down to the subducting slab. In the south a deeper zone of high attenuation is resolved between 23° and 24°S directly above the subducting slab. Low Qp in the mantle correlates with earthquake clusters. The strong crustal attenuation is confined to the distribution of young ignimbrites and silicic volcanism and is interpreted as a thermally weakened zone with partial melts. The attenuation pattern in the upper mantle might reflect the variable extent of the asthenosphere and maps variations of subduction-related hydration processes in the mantle wedge from slab-derived fluids.

Seismic slip on a low angle normal fault in the Gulf of Corinth: Evidence from high‐resolution cluster analysis of microearthquakes

The Gulf of Corinth in Western Greece is one of the most active extensional zones in the Aegean region. It is still an open question whether extension can be actively accommodated on low angle faulting or if those faults as seen in geological records have been rotated. Whilst numerous fault plane solutions obtained from a dense temporary network deployed in the western part of the gulf in July-August of 1991 showed one of the nodal planes as a subhorizomal plane, slip on the high-angle conjugate plane is equally probable from the focal mechanism data (Rigo et al. 1996). Since part of the activity occurred in spatial clusters with similar focal mechanisms, we used a high resolution cluster analysis to determine the most likely active plane. Exploiting the waveform similarity of these events, relative onset times of P and S waves could be determined at subsample accuracy (less than 0.01 s). The cluster analyzed here contains 12 evems, among these 8 have a well con- strained normal faulting fault plane solution with a shal- low (12-20 o) north dipping plane and a steeply south dipping plane. A master evem relocation shows that the relocated 12 hypocenter cemroids are aligned along the low angle plane showing clear evidence for active low angle normal faulting.

P wave attenuation structure in the fault area of the 1995 Kobe earthquake

The three-dimensional (3-D) P wave attenuation structure in the Kobe epicentral area is determined using aftershock records from the 1995 Kobe (Hyogo-ken Nanbu) earthquake. Waveform data from both permanent seismic networks and portable stations, set up following the Kobe mainshock, were used. The observed P wave amplitude spectra were inverted for the spectral plateau value, the source corner frequency, and a frequency-independent t* operator assuming an ω2-type source model. The calculated t* operators were subsequently used for the computation of the 3-D absorption structure using a 3-D velocity model. Only high-quality amplitude spectra were used for the estimation of reliable attenuation parameters, resulting in a final data set of ∼4100 t* operators. Two regions of high absorption were found in the upper part of the crust. One is below the northern part of Awaji Island and reaches down to a depth of 5 km. The other region, northeast of Kobe city, is along the Arima-Takatsuki tectonic line and extends along the Rokko, Kashiodani, and Yamada Faults. It extends northward to the Hokusetsu-Sanchi area. In both regions a high density of fault lineations is mapped, and high Poisson ratios are known. This suggests that partially saturated cracks cause the observed high attenuation.

Structure of the seismogenic zone of the southcentral Chilean margin revealed by local earthquake traveltime tomography

We use traveltime data of local earthquakes and controlled sources observed by a large, temporary, amphibious seismic network to reveal the anatomy of the southcentral Chilean subduction zone (37–39°S) between the trench and the magmatic arc. At this location the giant 1960 earthquake (M = 9.5) nucleated and ruptured almost 1000 km of the subduction megathrust. For the three-dimensional tomographic inversion we used 17,148 P wave and 10,049 S wave arrival time readings from 439 local earthquakes and 94 shots. The resolution of the tomographic images was explored by analyzing the model resolution matrix and conducting extensive numerical tests. The downgoing lithosphere is delineated by high seismic P wave velocities. High vp/vs ratio in the subducting slab reflects hydrated oceanic crust and serpentinized uppermost oceanic mantle. The subducting oceanic crust can be traced down to a depth of 80 km, as indicated by a low velocity channel. The continental crust extends to approximately a 50-km depth near the intersection with the subducting plate. This suggests a wide contact zone between continental and oceanic crust of about 150 km, potentially supporting the development of large asperities. Eastward the crustal thickness decreases again to a minimum of about a 30-km depth. Relatively low vp/vs at the base of the forearc does not support a large-scale serpentinization of the mantle wedge. Offshore, low vp and high vp/vs reflect young, fluid-saturated sediments of forearc basins and the accretionary prism.

Interaction between forearc and oceanic plate at the south‐central Chilean margin as seen in local seismic data

We installed a dense, amphibious, temporary seismological network to study the seismicity and structure of the seismogenic zone in southern Chile between 37° and 39°S, the nucleation area of the great 1960 Chile earthquake. 213 local earthquakes with 14.754 onset times were used for a simultaneous inversion for the 1-D velocity model and precise earthquake locations. Relocated artificial shots suggest an accuracy of the earthquake hypocenter of about 1 km (horizontally) and 500 m (vertically). Crustal events along trench-parallel and transverse, deep-reaching faults reflect the interseismic transpressional deformation of the forearc crust due to the subduction of the Nazca plate. The transverse faults seems to accomplish differential lateral stresses between subduction zone segments. Many events situated in an internally structured, planar seismicity patch at 20 to 40 km depth near the coast indicate a stress concentration at the plates interface at 38°S which might in part be induced by the fragmented forearc structure.

Partial Melting in the Central Andean Crust: a Review of Geophysical, Petrophysical, and Petrologic Evidence

The thickened crust of the Central Andes is characterized by several first-order geophysical anomalies that seem to reflect the presence of partial melts. Magnetotelluric and geomagnetic deep-sounding studies in Northern Chile have revealed a high conductivity zone (HCZ) beneath the Altiplano Plateau and the Western Cordillera, which is extreme both in terms of its size and integrated conductivity of > 20000 Siemens. Furthermore, this region is characterized by an extremely high seismic attenuation and reduced seismic velocity. The interrelation between the different petrophysical observations, in combination with petrological and heat-flow density studies, strongly indicates a huge area of partially molten rocks that is possibly topped with a thin, saline fluid film. The average melt fraction is deduced to be ∼20 vol.%, which agrees with typical values deduced from eroded migmatites. Based on the distribution and geochemical composition of Pliocene to Quaternary silicic ignimbrites in this area, this zone is thought to be dominated by crustally-derived rhyodacite melts with minor andesitic contribution. An interconnected melt distribution — typical for migmatites - would satisfy both the magnetotelluric and seismic observations. The high melt fraction in this mid-crustal zone should lead to strong weakening, which may be a main cause for the development of the flat topography of the Altiplano Plateau.

Scattering attenuation and intrinsic absorption using uniform and depth dependent model – Application to full seismogram envelope recorded in Northern Chile

Two seismic wave attenuation factors, scatteringattenuation Qs-1 and intrinsicabsorption Qi-1 are measured using theMultiple Lapse Time Window (MLTW) analysis method forthree different frequency bands, 1–2, 2–4, and 4–8 Hz.Data from 54 temporally deployed seismic stationslocated in northern Chile are used. This methodcompares time integrated seismic wave energies withsynthetic coda wave envelopes for a multiple isotropicscattering model. In the present analysis, the waveenergy is assumed to decay with distance in proportionto1/GSF·exp(- (Qs-1+Qi-1)·ω r/v), where r, ω and v are the propagationdistance, angular frequency and S wave velocity,respectively, and GSF is the geometricalspreading factor. When spatial uniformity of Qs-1, Qi-1 and v isassumed, i.e. GSF = 4πr2, theestimates of the reciprocal of the extinction length,Le-1 (= (Qs-1+Qi-1)·ω/v), are 0.017,0.012 and 0.010 km-1, and those of the seismicalbedo, B0 (= Qs-1/ (Qs-1+Qi-1)), are 0.48, 0.40and 0.34 for 1–2, 2–4 and 4–8 Hz, respectively, whichindicates that scattering attenuation is comparable toor smaller than intrinsic absorption. When we assumea depth dependent velocity structure, we also findthat scattering attenuation is comparable to orsmaller than intrinsic absorption. However, since thequantitative estimates of scattering attenuationdepend on the assumed velocity structure (strength ofvelocity discontinuity and/or Moho depth), it isimportant to consider differences in velocitystructure models when comparing attenuation estimates.