S U M M A R YWe invert surface-wave and geodetic data for the spatio-temporal complexity of slip during the M,=8.1 Chile 1995 event by simulated annealing. This quasi-global inversion method allows for a wide exploration of model space, and retains the nonlinearity of the source tomography problem. Complex source spectra are obtained from 5 to 45 mHz from first-and second-orbit fundamental-mode Rayleigh waves using an empirical Green's function cross-correlation technique. Coseismic displacement vectors were measured at 10 GPS sites near Antofagasta. They are part of a French-Chilean experiment which monitors the Northern Chile seismic gap. The spectra, together with the geodetic data, are inverted for the moment distribution on a 2-D dipping fault, under the physical constraints of slip positivity and causality. Marginal a posteriori distributions of the model parameters are obtained from several independently inverted solutions. In general, features of the slip model are well resolved. Data are well fitted by a purely unilateral southward rupture with a nearly uniform velocity around 2.5-3.0 km s-l, and a total duration of 65 s. Several regions of moment release were imaged, one near the hypocentre, a major one 80 km south of it and a minor one 160 km south of it. The major patch of moment release seemed to have propagated to relatively shallow depths near the trench, 100 km SSW of the epicentre. The region of major slip is located updip of the 1987, M,=7.5 earthquake, suggesting a causal relationship. Most of the slip occurred updip of the hypocentre (36 km), but the entire coupled plate interface (20-40 km) ruptured during the Chile 1995 event.

The spatio‐temporal complexity of moment release of the 1992 Nicaragua slow earthquake (Mw=7.7) is estimated by simulated annealing. Contrary to linearized inversions, Monte Carlo techniques do not make assumptions about a priori features of the model, in particular, large variations in rupture velocities can be imaged. Relative spectra between the main shock and an empirical Green's function were obtained from 4–44 mHz using first and second orbit fundamental mode Rayleigh waves. The spectra were inverted for the moment distribution on a 2D horizontal fault, under the physical constraints of slip positivity and causality. The non‐linear inversion revealed two main regions of slip. One centered 20 km to the NW of the epicenter, the other 160 km to the SE, but delayed by 70 s. Rupture velocities in the slipping patches were found to be low, around 1 km/s. Large amount of slip occurred trench ward of the epicenter, accounting for the large tsunami excitation. Analysis of the slip distribution suggests that the slow and smooth nature of the 1992 Nicaragua event may be related to the frictional properties at the plate interface.

Low-frequency spectral anomalies have indicated that some large earthquakes are preceded by extended episodes of smooth moment release, but the reality of these slow precursors has been debated because they have not been directly observed in the time domain. High-gain seismograms from the 14 March 1994 Romanche Transform event (moment magnitude M w 7.0) show a precursory ramp with a moment of 7 × 10 18 newton meters beginning about 100 seconds before the arrival of the high-frequency P waves. This precursor was the initial phase of a slow component of slip that released nearly half of the total moment of the earthquake. Such behavior may be typical for large earthquakes on the oceanic ridge-transform system.

Abstract. We investigate the spatio-temporal complexity of than a few tens of cm at various sites in the Pacific, the Marmoment release of the February 21, 1996 Peru earthquake quesas experienced locally water waves on the order of 2 m (M w 7.5). We use a non-linear source tomographic technique, [Heinrich et al., 1998]. We investigate the source process of based on simulated annealing, to invert surface wave source the 1996 Peru earthquake using two different methods: a) spectra for the slip distribution on a gently dipping fault plane. source tomography based on surface wave data that provide us The spectra (5-65 mHz) are obtained using an empirical with an appropriate low-frequency resolution of the source Green's function (EGF) method applied to first and second or-process (5-65 mHz), and an enhanced sensitivity to source dibit fundamental mode Rayleigh waves. Spectra are well fit by rectivity effects; b) inversion of broadband body waves to exa 110 km bilateral rupture, subparallel to the trench, updip of tend our analysis to higher frequencies (10-500 mHz). Our the hypocenter, with a total moment of 2.0x102ø Nm. The source tomographic model serves as a basis to calculate the non-linear inversion reveals a 30x30 km 2 major slip patch south the onset, where rupture velocities attain 1.5-2.0 km/s. Inversions of teleseismic broadband P and SH waves (10-500 mHz) indicate little or no directivity, consistent with the surface wave data. The average source time functions for both the surface and body wave data are similar in shape and in duration (-50 s). Part of the slip during the 1996 Peru event occurred in a region of reduced background seismicity, as was the case for the 1992 Nicaragua slow event, suggesting that the seismogenic potential of low seismicity regions near the trench should be globally reassessed.

Abstract. The February 1996 Peruvian earthquake generated a larger tsunami than expected from its surface magnitude. This discrepancy as well as the long rupture duration indicate that this event is a 'tsunami earthquake'. The associated tsunami was strong locally with runup heights of 1 to 5 meters along a coastline of 400 km. It is shown that this tsunami can be modeled for a seismic moment of 2.102ø Nm, using a rigidity of 2.10 •ø N/m 2. The tsunami propagation is modeled solv-