Seismic and aseismic slip on the Central Peru megathrust

2021

Slow slip phenomena deep in subduction zones reveal cyclic processes downdip of locked megathrusts. Here we analyze seismicity within a subducting oceanic slab under Kii Peninsula, Japan, spanning nearly 50 major deep slow slip and tremor episodes over 17 years. Changes in rate, b-values, and stress orientations of inslab seismicity are temporally associated with the slow slip episodes. Furthermore, although stress orientations in the slab below these slow slips may rotate slightly, inslab orientations 20 to 50 km updip from there rotate significantly, suggesting previously-unrecognized transient slow slip occurs on the plate interface updip. We infer that fluid migrates from slab to interface, promoting episodes of slow slip, which break mineral seals, letting fluid migrate 10’s of km further updip along the interface where it promotes transient slow slips. The proposed methodology, based primarily on inslab seismicity, may help monitor plate boundary conditions and slow slip pheno...

Nature Geoscience, 2021

Most destructive tsunamis are caused by seismic slip on the shallow part of offshore megathrusts. The likelihood of this behaviour is partly determined by the interseismic slip rate deficit, which is often assumed to be low based on frictional studies of shallow fault material. Here we present a new method for inferring the slip rate deficit from geodetic data that accounts for the stress shadow cast by frictionally locked patches, and show that this approach greatly improves our offshore resolution. We apply this technique to the Cascadia and Japan Trench megathrusts and find that wherever locked patches are present, the shallow fault generally has a slip rate deficit between 80 and 100% of the plate convergence rate, irrespective of its frictional properties. This finding rules out areas of low kinematic coupling at the trench considered by previous studies. If these areas of the shallow fault can slip seismically, global tsunami hazard could be higher than currently recognized. Our method identifies critical locations where seafloor observations could yield information about frictional properties of these faults in order to better understand their slip behaviour. Megathrust faults at convergent tectonic margins produce devastating great earthquakes and tsunamis. Understanding their potential rupture behavior, particularly in the shallow offshore part of the fault where most destructive tsunamis are generated 1 , is therefore a critical task for geoscientists forecasting seismic and tsunami inundation hazards 2. Earthquakes on megathrusts release stresses that accumulate due to plate convergence during the interseismic period, when the frictionally unstable portion of the fault is kinematically coupled, i.e. not slipping. This gradual accumulation of elastic strain can be detected geodetically 3-12. The strain budget available to drive seismic slip depends on whether the fault is able to slip (or creep) interseismically. On megathrusts, interseismic creep near the trench would indicate an absence of frictional locking there, and would result in a lower strain budget for coseismic slip and thus a lower tsunami hazard. On the other hand, if these faults are not creeping during the interseismic period, we can conclude that there is stored strain that may be released during or shortly after large earthquakes, increasing the possibility of tsunami generation.

Journal of Geophysical Research, 2007

Science Advances

At subduction zones, transient aseismic slip occurs either as afterslip following a large earthquake or as episodic slow slip events during the interseismic period. Afterslip and slow slip events are usually considered as distinct processes occurring on separate fault areas governed by different frictional properties. Continuous GPS (Global Positioning System) measurements following the 2016 M w (moment magnitude) 7.8 Ecuador earthquake reveal that large and rapid afterslip developed at discrete areas of the megathrust that had previously hosted slow slip events. Regardless of whether they were locked or not before the earthquake, these areas appear to persistently release stress by aseismic slip throughout the earthquake cycle and outline the seismic rupture, an observation potentially leading to a better anticipation of future large earthquakes.

Journal of Geophysical Research: Solid Earth

In February 2014 a M w = 7.0 slow slip event (SSE) took place beneath the Nicoya Peninsula, Costa Rica. This event occurred 17 months after the 5 September 2012, M w = 7.6, earthquake and along the same subduction zone segment, during a period when significant postseismic deformation was ongoing. A second SSE occurred in the middle of 2015, 21 months after the 2014 SSE and 38 months after the earthquake. The recurrence interval for Nicoya SSEs was unchanged by the earthquake. However, the spatial distribution of slip for the 2014 event differed significantly from previous events, having only deep (~40 km) slip, compared to previous events, which had both deep and shallow slip. The 2015 SSE marked a return to the combination of deep plus shallow slip of preearthquake SSEs. However, slip magnitude in 2015 was nearly twice as large (M w = 7.2) as preearthquake SSEs. We employ Coulomb Failure Stress change modeling in order to explain these changes. Stress changes associated with the earthquake and afterslip were highest near the shallow portion of the megathrust, where preearthquake SSEs had significant slip. Lower stress change occurred on the deeper parts of the plate interface, perhaps explaining why the deep (~40 km) region for SSEs remained unchanged. The large amount of shallow slip in the 2015 SSE may reflect lack of shallow slip in the prior SSE. These observations highlight the variability of aseismic strain release rates throughout the earthquake cycle. Plain Language Summary We analyzed small signals in continuous GPS time series. By averaging many GPS measurements over a day, we are able to get very precise measurements of the motion of the ground. We found two events in the Nicoya Peninsula of Costa Rica where the GPS changed direction and began moving toward the oceanic trench in the opposite direction of subduction plate motion. These events are called slow slip events and have been found in other regions such as Cascadia, Alaska, Japan, and New Zealand. In Nicoya, a large earthquake of magnitude 7.6 on the Richter scale occurred in 2012. The two slow slip events occurred in 2014 and 2015. We explored the relationship between the earthquake and the slow slip events and looked to see if the earthquake changed the behavior of the slow slip events. We found that the slow slip events have a regular timing before and after the earthquake, but the behavior of the slow slip events since the earthquake is different with slip taking place along different portions of the plate interface then was previously seen.

Journal of Geophysical Research, 2006

Tectonophysics, 2010

A long-standing goal of subduction zone earthquake studies is to determine whether or not there are physical processes that control seismogenesis and the along-strike segmentation of the megathrust. Studies of individual earthquakes and global compilations of earthquakes find favorable comparison between coseismic interplate slip distributions and several different long-lived forearc characteristics, such as bathymetry, coastline morphology, crustal structure, and interplate frictional properties, but no single explanation seems to govern the location and slip distribution of all earthquakes. One possible reason for the lack of a unifying explanation is that the inferred earthquake parameters, most importantly the slip distribution, calculated in some areas were inaccurate, blurring correlation between earthquake and physical parameters. In this paper, we seek to test this possibility by comparing accurate slip distributions constrained by multiple datasets along several segments of a single subduction zone with the various physical properties that have been proposed to control or correlate with seismogenesis. We examine the rupture area and slip distribution of 6 recent and historical large (M w N 7) earthquakes on the Peru-northern Chile subduction zone. This analysis includes a new slip distribution of the 14 November 2007 M w = 7.7 earthquake offshore Tocopilla, Chile constrained by teleseismic body wave and InSAR data. In studying the 6 events, we find that no single mechanism can explain the location or extent of rupture of all earthquakes, but analysis of the forearc gravity field and its gradients shows correlation with many of the observed slip patterns, as suggested by previous studies. Additionally, large-scale morphological features including the Nazca Ridge, Arica Bend, Mejillones Peninsula, and transverse crustal fault systems serve as boundaries between distinct earthquake segments.

Characterizing the time evolution of slip over different phases of the seismic cycle is crucial to a better understanding of the factors controlling the occurrence of large earthquakes. In this study, we take advantage of InSAR data and 3.5 years of continuous GPS (cGPS) measurements to determine inter-, co-and postseismic slip distributions in the region of the 2007, M w 8.0 Pisco, earthquake, Peru, using the same fault geometry and inversion method. Our interseismic model, based on pre-2007 campaign GPS data, suggests that the 2007 Pisco seismic slip occurred in a region strongly coupled before the earthquake while afterslip occurred in low coupled regions. Large afterslip occurred in the peripheral area of coseismic rupture in agreement with the notion that afterslip is mainly induced by co-seismic stress changes. The temporal evolution of the region of maximum afterslip, characterized by a relaxation time of about 2.3 years, is located in the region where the Nazca ridge is subducting, consistent with rate-strengthening friction promoting aseismic slip. We estimate a return period for the Pisco earthquake of about 230 years with an estimated aseismic slip that might account for about 50% of the slip budget in this region over the 0-50 km seismogenic

Journal of Geophysical Research: Solid Earth, 2013

We document a one week long slow-slip event (SSE) with an equivalent moment 27 magnitude of 6.0-6.3 which occurred in August 2010 below La Plata Island (Ecuador), south 28 of the rupture area of the Mw=8.8 1906 megathrust earthquake. GPS data reveal that the SSE 29 occurred at a depth of about 10km, within the downdip part of a shallow (<15km), isolated, 30 locked patch along the subduction interface. The availability of both broad-band seismometer 31 and continuous geodetic station located at the La Plata Island, 10km above the SSE, enables a 32 careful analysis of the relationships between slow and rapid processes of stress release along 33 the subduction interface. During the slow slip sequence, the seismic data shows a sharp 34 increase of the local seismicity, with more than 650 earthquakes detected, among which 50 35 have a moment magnitude between 1.8 and 4.1. However, the cumulative moment released 36 through earthquakes accounts at most for 0.2% of the total moment release estimated from 37 GPS displacements. Most of the largest earthquakes are located along or very close to the 38 subduction interface with focal mechanism consistent with the relative plate motion. While 39 the earthquake sizes show a classical distribution (Gutenberg-Richter law with a b-value close 40 to 1), the space-time occurrence presents a specific pattern. First, the largest earthquakes 41 appear to occur randomly during the slow slip sequence, which further evidence that the 42 seismicity is driven by the stress fluctuations related to aseismic slip. Moreover, the seismicity 43 observed during the SSE consists in individual events and families of repeating earthquakes. 44

Detailed new paleoseismic field investigations at two sites on the Talas-Fergana fault, a poorly known strike-slip structure that transects the Tien Shan mountain range, document late Holocene slip rates of 11–16 mm a À1. This prominent structure is distinctive in striking obliquely NW-SE across the Tien Shan, which is otherwise dominated by contractional structures striking generally E-W. Moreover, a satellite-based Global Positioning System network spanning the Tien Shan orogen records active N-S contraction rates of ~20 mm a À1 , but limits slip on the Talas-Fergana fault to <2 mm a À1. This profound mismatch between long-term geologic and short-term geodetic slip rates, which may suggest temporal variability in slip, highlights the importance of field-based investigations as a complement to remotely sensed data, particularly in evaluating models of lithosphere behavior and earthquake probabilities on presently locked faults such as the Talas-Fergana. Plain Language Summary The 700 km long Talas-Fergana fault is one of several large faults linked to the collision between the Indian and Eurasian tectonic plates and cuts obliquely across the Tien Shan mountain range north of the Himalayas. Satellite-based Global Positioning System data show that the Tien Shan range is being rapidly compressed tectonically, at the rate of about 20 mm per year, yet indicate a slip rate on the Talas-Fergana of less than 2 mm per year. However, our study employs intensive field geologic techniques to determine that the slip rate on this fault has averaged 11–16 mm per year over recent millennia. With increasing reliance on satellite-based measurements investigating such a profound mismatch is important for reliable earthquake hazard assessments on the Talas-Fergana fault, which appears locked and is at an unknown point in its earthquake cycle. Reconciling the two data sets suggests significant variability in slip rate on the Talas-Fergana through time, a conclusion that may have additional implications for models of lithosphere behavior.