Presence and significance of backstops in the overriding plate during the megathrust earthquake cycle

<p>Seismological and geodetic observations indicate that similar physical processes are active at different subduction margins and provide information about the deformation at the different stages of the earthquake cycle. We analyze geodetic observations along sections of the South American subduction zone during the inter-seismic stage. Results show that overriding plates shorten from the trench to a “backstop”, where horizontal inter-seismic velocities become close to zero. In most, but not all regions, the backstop location from trench-perpendicular GPS velocities agrees with that from trench-parallel velocities. The distance of the backstop from the trench varies along the western South America margin. Backstop locations shows some correlation with gradients in the effective elastic thickness of the overriding plate. An apparently conflicting observation is that co-seismic and early post-seismic GPS-displacements during the 2010 Maule earthquake extended well beyond the backstop into eastern South America. Similarly conflicting observations were made in the overriding plate of the 2004 Sumatra earthquake and the 2011 Tohoku earthquake.</p><p>We use cyclic 3D numerical models with dynamically driven co-seismic and afterslip to test the hypothesis that lateral contrasts in the thickness and/or elasticity of the overriding plate explain the observations. The model setup allows us to explore the sensitivity of geodetically observable surface motion to the mechanical structure of the subduction system during all parts of the earthquake cycle. We conclude that the observations can be explained by a lateral contrast. Such contrast restricts inter-seismic horizontal velocities in the region between the trench and the backstop, controlling their gradient, while allowing deformation due to coseismic slip and afterslip to reach well into the far field. One particularly interesting finding from our models is that stress accumulation in the overriding plate is controlled by the distance to the backstop.</p>

Setting up of the Indian Tsunami Early Warning System (ITEWS): National and International Interactions for the Success of ITEWS

<p>The 26 December 2004 Sumatra earthquake of Mw 9.2 and the resultant tsunami that claimed over 2,50,000 human lives is probably the most destructive natural disaster of the 21<sup>st</sup> Century so far. Although the science of tsunami warning had advanced sufficiently by that time, with several tsunami warning centers operating in various oceans, no such system existed for the Indian Ocean. Here we present the discussions and interactions held in India and globally to convince setting up of ITEWS. False tsunami alarms subsequent to 26 December 2004 earthquake had developed a sense of scientific disbelief in the public and to a certain extent in Government of India. We demonstrated to the national and international community that there are only two stretches of faults that could host tsunamigenic earthquakes as far as the India Ocean is concerned. These are: 1) a stretch of some 4000 km of a fault segment extending from Sumatra to Andaman Islands and 2) an area of about 500 km radius off the Makaran Coast in the Arabian Sea. And if we cover these two areas with ocean bottom pressure recorders, the problem of false alarms would be reduced to a large- extant. This plan was finally agreed to and necessary financial, logistic and technical support was made available. The setting up of the ITEWS started in middle 2005 and was completed in August 2007. It has performed very efficiently since then. Over the past ~ 8 years, it monitored ~ 500 M ≥ 6.5 and provided advisories. As against the requirement placed by IOC of issuing an advisory in 10 to 15 minutes time, ITEWS has been doing it in ~ 8 minutes. Since its inception in 2007, no false alarm has been issued and it is rated among the best in the world.</p><p>IOC has designated ITEWS as the Regional Tsunami advisory Provider (TSP) Indian Ocean Regional Tsunami Center.</p>

Spatial Distribution of Viscoelastic Relaxation Following a Subduction Earthquake

<p>Viscoelastic relaxation is generally considered as the dominant process of the long-term post-seismic deformation, while viscoelastic characteristic relaxation time represents the time scale of deformation caused by viscoelastic relaxation effect after the earthquake. The subduction earthquakes which occurred at the boundary of the ocean and continental plates often release greater stress, and the stress relaxation of mantle materials is more significant due to the response to viscoelasticity. Satellite gravity mission GRACE (gravity recovery and climate experience) is able to observe the corresponding co-seismic and post-seismic gravity changes. Therefore, in this study, we use the monthly gravity field model data of GRACE RL06 to study the post-seismic gravity changes of 2011 Tohoku earthquake and 2004 Sumatra earthquake. After removing the influence of sea level changes, GIA changes and GLDAS on the seasonal precipitation changes in the land area, as well as the sea water correction, we get the post-seismic deformation only related to the deformation of the solid earth. Then we use the attenuation function to fit each grid value and obtain the spatial distribution of viscoelastic characteristic relaxation time after rejecting the afterslip from the total post-seismic deformation. Thus，we can capture  the viscous structure in the subduction area.</p>