GPS-derived interseismic coupling on the subduction and seismic hazards in the Atacama region, Chile

Abstract

The Atacama region (between 29 • S and 25 • S) is located in the North-Central area of Chile, a tectonically complex transition area between North and Central Chile. Deformation in Atacama is due mainly to elastic loading on the subduction interface but also to diffuse shortening in the Sierras Pampeanas, Argentina. The seismicity of the subduction is complex in this region: seismic swarms often occur, moderate (M w ∼ 6) to large (M w ∼ 7) earthquakes occur repeatedly and finally, megathrust earthquakes of magnitudes significantly larger than 8 occur once in a while, the last one being in 1922-almost a century ago. We use new GPS data we collected in the Atacama region between 2008 and 2012 to complete and densify existing data we acquired since 2004 in North-Central Chile. These new data allow to quantify the motion of the Andean sliver and assess the kinematic coupling on the subduction interface at these latitudes. We find that only 7 per cent of the whole convergence motion is taken up by an eastward rotation of the rigid sliver. A large part of the remaining 93 per cent (approximately 6 cm yr −1 ) gives way to accumulation of elastic deformation in the upper plate, due to locking on the plate interface. This accumulation shows important along-strike and along-dip variations, interpreted in terms of variable coupling which we correlate with seismicity. We identify two areas of low coupling near the 'La Serena' (30 • S) and 'Baranquilla' (27.5 • S) bays. Both are correlated with the subduction of singular bathymetric features and seem to stop the propagation of large seismic ruptures. These zones are also seismic swarm prone areas, which seem to occur rather on their edges. These low coupling areas separate two seismic segments where coupling is high: the Atacama segment (∼100 km long between 29 • S and 28 • S) and the Chañaral segment (∼200 km long between 27 • S and 25 • S). Should they rupture alone, these segments are sufficiently coupled and apparently since long enough, to produce M w ∼ 8 events. However, a collective failure of both segments could generate a megathrust earthquake of magnitude close to 8.5, similar to the 1819 and 1922 complex events, which produced important tsunamis. Such giant events may occur in the area once a century.

Rupture Process During the 2015 Illapel, Chile Earthquake: Zigzag-Along-Dip Rupture Episodes

“…Even considering this imaging artifact in the synthetic test, the strength of the observed signals at around the terminus of the second rupture episode (60-90 s) is still weaker than that of the strong highfrequency signals observed at the initiation of the second rupture episode at the down-dip edge of the slip, and this artifact does not significantly affect the validity of the discussion. Slip distribution inferred from the waveform inversion analysis correlates well with the highly locked region in the plate-coupling models by MOR-ENO et al (2010) andMÉTOIS et al (2014). The locations of the 1997-1998 swarms coincide with the northern and northeastern edges of the slip area (Fig.…”

“…This tendency is contrary to what is observed during the 2015 Gorkha, Nepal, earthquake, where the rupture front velocity abruptly decelerates and it generates strong high-frequency radiation as a stopping phase (e.g., MADARIAGA 1977) at the end of the rupture sequence (YAGI and OKUWAKI 2015). The northern edge of the 2015 Illapel source region coincides with the transition zone between high and low coupled region (MORENO et al 2010;MÉTOIS et al 2014) and the southern edge of the northern part of the 1997-1998 swarms (Fig. 1).…”

“…MORENO et al (2010), using the global positioning system (GPS) data from 1996-2008, found that the plate interface in the middle part of the CoquimboIllapel region is strongly locked (almost 100 % coupled), making it the potential site of a future earthquake. MÉTOIS et al (2014) also used the GPS data to invert a degree of interplate coupling and showed that the middle to southern part of the Coquimbo-Illapel region is almost 100 % coupled at 15-25 km depth, while the coupling at northern part is relatively low, which corresponds to the northern part of the 1997-1998 swarm region. Considering the swarm activity and the plate-locking distribution along the Coquimbo-Illapel region, the 2015 Illapel earthquake occurred where the stress build-up has progressed non-uniformly in both space and time since the 1943 Illapel earthquake.…”

Subduction zone megathrust earthquakes

“…However, the largest documented historical event occurred along the coast from Guerrero to Tehuantepec in 1787 with M ~ 8.6 (Suárez and Albini, 2009) (Iquique,M ~8.8), and this remains a major seismic gap. The 1922 rupture zone along Atacama likely has substantial strain accumulation relative to surrounding regions, as geodetic measurements indicate that the megathrust is pervasively locked in the region (Métois et al, 2014).…”

Rise of the central Andean coast by earthquakes straddling the Moho

“…5b). Both coastal uplift and subsidence occur during interseismic plate locking 15,39,40 . At ∼21 • S, the coast emerges at ∼3 mm yr −1 , as suggested by geodesy 15 and a back-slip model of full plate locking down to 35 km depth (Fig.…”