Kinematic Study of Pisagua Earthquake 2014 - Northern Chile: Analysis of the Frequency Content and its Impact on the Understanding of the Seismogenic Zone

Abstract:

Recent megathrust earthquakes that occurred in Chile (2010) and Japan (2011) revealed a segmentation with depth of the megathrust interface: short period radiations were emitted from the deeper portion of the seismogenic zone (35-55km depth), while large coseismic slip, associated with little short period radiation and responsible for the tsunami generation, occurred on the shallower part of the subduction interface.

On April 1, 2014, a Mw 8.2 earthquake occurred close to the city of Pisagua in the central part of the seismic gap of southern Peru - northern Chile. This earthquake occurred in a densely instrumented area in a joint effort between Chilean, French, German and USA institutions. It is therefore an excellent case study to better understand the seismic rupture process on the subduction seismogenic interface. The availability of collocated cGPS and strong motion data (IPOC network: Integrated Plate Boundary Observatory Chile) offers a unique opportunity to study the seismic source, compare the results derived from both type of data and study the variability of the seismic source with the frequency.

We perform a detailed comparison of the co-seismic movements registered by cGPS and three-component accelerograms. Then, we carry out a series of inversions to study the kinematic rupture associated with this earthquake. For this purpose, we use the accelerometer (9 stations) and high-frequency GPS (1 Hz, 13 stations) independently and jointly. We use a two-step inversion method proposed by Hernandez et al. (1999), where we use the contribution of GPS in obtaining the static displacement and then carry out the kinematic inversion using the method in frequency domain (Cotton & Campillo, 1995). Finally we study the dependency of the Mw8.2 Pisagua seismic source with the signal frequency. Data are filtered in different frequency bands and then frequency dependent inversions are conducted to explore the segmentation of the seismogenic zone in the area affected by the earthquake. In particular, we aim at testing whether a Mw8.2 earthquake can show a frequency dependency similar to the one observed for megathrust events, characterize this dependency and define segmentation along depth and along strike of the seismogenic zone, compare this segmentation with the one issued from GPS-derived interseismic coupling maps.