Deep mass redistribution prior to the Mw 8.8 Maule earthquake (Chile, 2010) revealed by GRACE satellite gravity

<p align="justify"><span lang="en-US">Subduction zones are places of intense seismic activity where the largest ruptures occur. </span><span lang="en-US">Studies</span><span lang="en-US"> on the causal mechanisms of </span><span lang="en-US">subduction </span><span lang="en-US">earthquakes </span><span lang="en-US">generally </span><span lang="en-US">focus on the accumulation of tectonic </span><span lang="en-US">stress and strain at </span><span lang="en-US">the shallow </span><span lang="en-US">plates</span><span lang="en-US"> interface, </span><span lang="en-US">which can be documented from</span><span lang="en-US"> surface </span><span lang="en-US">displacements and seismic activity</span><span lang="en-US">. </span><span lang="en-US">The control exerted by deeper sudbuction processes is however not well understood. It can be addressed from time-varying satellite gravity data, that provide</span><span lang="en-US"> a new and unique means of studying mass redistribution</span><span lang="en-US">s</span><span lang="en-US"> at intermediate spatial and temporal scales throughout the volume around plate boundaries, and in particular </span><span lang="en-US">at depht.</span></p> <p align="justify">Here we use gravity gradients from GRACE geoid <span lang="en-US">to </span><span lang="en-US">probe slow deep mass variations and their possible interactions with intraplate seismicity along the Chilean margin. We </span><span lang="en-US">work</span><span lang="en-US"> with three different GRACE geoid models </span><span lang="en-US">(GRGS, CSR,ITSG) </span><span lang="en-US">from 2003 to 2014, over a large region surrounding the rupture zone of the Mw 8.8 2010 Maule earthquake. From these </span><span lang="en-US">data</span><span lang="en-US"> we reconstruct </span><span lang="en-US">the Earth&#8217;s </span><span lang="en-US">gravity gradients at different spatial scales </span><span lang="en-US">in order</span><span lang="en-US"> to better separate signals associated with mass sources of different sizes, shapes or orientations in the GRACE geoids. </span>Our analysis reveals an anomalous gravity gradient signal north-east of the epicentral zone, which amplitude progressively increases during the months preceding the earthquake. This signal is consistently detected in all 3 GRACE solutions and we show that it cannot be explained by a water mass redistribution nor artefacts. Instead, it could be explained by an extension of the plunging Nazca plate near 150 km depth along the subduction direction. The migration of the gravity signal laterally and from the depths to the surface <span lang="en-US">from a weakly coupled </span><span lang="en-US">zone in the North</span><span lang="en-US"> to a strongly coupled zone </span><span lang="en-US">in the South </span>suggests that the Mw 8.8 earthquake may have originated the propagation of this deep slab deformation towards the surface. <span lang="en-US">Our results highlight the importance of time series of satellite observations of the Earth&#8217;s gravity field, to detect </span><span lang="en-US">and characterize</span><span lang="en-US"> mass redistributions at depth of major plate boundaries </span><span lang="en-US">at timescales of month to years. </span></p> <p lang="en-US" align="justify">&#160;</p>