A 3-year program of research on great earthquakes and the subduction process and on the properties of large shocks along transform faults is proposed. Under an existing grant from NSF a synthesis of data from seismology and marine geophysics indicates that the interplate thrust boundary from depths of 0 to 15 km at subduction zones moves mainly aseismically. The transition to slip in large earthquakes at greater depth occurs at the transition along the plate boundary from young, weak sediments to stronger rocks of what is called the backstop. That transition also controls the morphology of the accretionary wedge- forearc basin. We propose to map the downdip width, W, of rupture zones since it is likely to be a critical parameter governing the maximum size of thrust events and the average repeat time of large shocks along given parts of subduction zones. Knowledge of W should allow us to ascertain if the plate boundaries off Washington, Oregon, New Zealand and the southern Lesser Antilles - all regions of young thick sediments - are capable of generating large or great thrust events. A catalog of seismic moments for large shocks worldwide is being prepared that will be used to study the distribution of seismic and aseismic slip along and among various plate boundaries. We will examine the transition form seismic to aseismic slip at depth in subduction zones and factors that cause that depth to vary from 25 to 70 km. Accurate hypocentral locations, focal mechanisms and knots of seismic activity will be used to search for major asperities along thrust boundaries. We think that offsets downdip from a planar geometry at duplex and ramp structures can act as major asperities. Such features are well known to structural geologists but their effect on the mechanics of large earthquakes does not appear to have been investigated. Work under existing and prior grants led to the identification of a seismic gap off central Chile as having a high probability of rupturing from 1982 to 2000. A large and damaging event occurred there in l985. The proposed research should have major social benefits by the identification of repeat times and maximum sizes of large earthquakes and specific sites where precursory phenomena are likely to occur.
