Correlations between gravity anomaly, preseismic locking and coseismic slip during the 2014 Mw8.2 Pisagua/Iquique earthquake and implications for the frictional structure of the Central Andean megathrust

Abstract:

Gravity anomalies at subduction zones and the degree of megathrust locking inverted from geodetic data have been individually correlated with coseismic asperities ruptured by great earthquakes. The time-scale of both observables varies from millions of years in the case of gravity to years for geodetic locking. Since the relationship between these different proxies of plate coupling has not been established, it is unclear whether megathrust asperities are permanent or transient features, and therefore the combined use of these proxies for seismic hazard assessment remains problematic. Here we show that the area ruptured by the April 1^st 2014 Mw8.2 Pisagua/Iquique earthquake at the Central Andean subduction zone coincides with an anomalous region of the northern Chile seismic gap where the degree of preseismic fault locking was much larger than what could be expected from the value of its gravity anomaly. Preseismic locking and coseismic slip were inverted from GPS observations using a finite element approach with a realistic slab geometry derived from a 3D density model. The same model was used to compute the slab gravity effect that was subtracted from a global geopotential model (EGM2008) in order to obtain a residual gravity anomaly due to lateral variations of forearc density structure. We suppose that locking and gravity must be positively correlated as could be inferred if regions of low/high forearc density would induce small/large vertical stresses on the megathrust. We found that the location of the main coseismic slip patch almost perfectly matches a region of anomalously high locking compared with its gravity anomaly. This result favors a conceptual model where regions of low gravity anomaly are caused by a low-density forearc that imposes a low normal stress loading the megathrust, which promotes a conditionally-stable frictional regime. Under this regime the megathrust creeps aseismically during certain time after a great earthquake but can eventually lock and start building up elastic strain during years or decades before its coseismic release. This model challenges previous interpretations of gravity anomalies at subduction zones and helps understanding controls of seismogenic processes for the Central Andean megathrust (and elsewhere) with implications for its short-term seismic hazard