Evidence for slow slip events preceeding the M8, April 1rst, 2014 Pisagua Earthquake (Chile), from an underground, long base hydrostatic tiltmeter.

Abstract:

The two component, hydrostatic, long-base tiltmeter (~50 m long ) installed at 150 in depth in the Santa Rosa mine near the city of Iquique (northern Chile) was operational during the 3 months of preseismic activation for the M=8, 2014, April 1rst Pisagua Earthquake, 90 km from its epicenter. These instruments have a resolution of 10^-9 rad and a stability better than 10^-8rad/month (Boudin & al., 2008, 2013). We present here the associated records and their interpretation in terms of a possible succession of moderate slow slip events before the mainshock. The instrument was installed in July 2012 in order to detect low amplitude strain transients in the northern chile subduction gap, as part of the IPOC project. Removal of earth tide and correlation with temperature and atmospheric pressure were investigated to reduce the noise level. After a first 10 months recording phase, interruptured by temporary mining works, the recording restarted mid-December 2013. After stabilization, the westward tilt rate took a stable value, about 7.10-7 rad/yr, corresponding to the interseismic loading After the 5-7 january earthquake sequence, it jumped to a higher, stable value (phase A) until early February. The 2-4 February, the tilt rate changed to a smaller value, and stayed stable (phase B) until the 16 of March, M=6.8 major foreshock. The latter, with the following foreshocks, produced large tilt steps, but the tilt rate between these events kept stable (phase C), close to that of phase A. Finally, the north component shows a fast tilt rate change starting the 5 of March. The coincidence betwen the start of seismic swarms and the change in tilt rate lead us to associate the latter to the activation of moderate slow slip events. Incidentally, we also show that the large slow slip source concentrated in the epicentral area proposed by Ruiz et al. (2014) from their GPS analysis is not supported by our tilt records.