Abstract. On June 2, 1994, a large subduction thrust earthquake (Ms 7.2) produced a devastating tsunami on the island of Java. This earthquake had a number of unusual characteristics. It was the first recorded large thrust earthquake on the Java subduction zone. All of the aftershock mechanisms exhibit normal faulting; no mechanisms are similar to the main shock. Also, the large tsunami and the relatively low energy radiated by the main shock have led to suggestions that this earthquake might have involved slow, shallow rupture near the trench, similar to the 1992 Nicaragua earthquake. We first relocate the main shock and the aftershocks. We then invert long-period surface waves and broadband body waves to determine the depth and spatial distribution of the main shock slip. A dip of 12 ø, hypocenter depth of 16 km and moment of 3.5x102ø N rn (Mw 7.6) give the best fit to the combined seismic data and are consistent with the plate interface geometry. The source spectrum obtained from both body and surface waves has a single comer frequency (between 10 and 20 mHz) implying a stress drop of--0.3 MPa. The main energy release was preceded by a small subevent lasting -o12 s. The main slip occurred at --20 km depth, downdip and to the NW of the hypocenter. This area of slip is collocated with a prominent high in the bathymetry that has been identified as a subducting seamount. We interpret the Java earthquake as slip over this subducting seamount, which is a locked patch in an otherwise decoupled subduction zone. We find no evidence for slow, shallow rupture. No thrust aftershocks are expected if the entire locked zone slipped during the main shock, but extension of the subducting plate behind the seamount would promote normal faulting as observed. It seems probable that such a source model could also explain the size and timing of the observed tsunami.

[1] Two large (M w 7.9) earthquakes occurred on 4 and 18 June 2000, south of Sumatra, beneath the Indian Ocean. Both earthquakes were predominantly left-lateral strike-slip on vertical N-S trending faults that we interpret to be reactivated fracture zones. The 4 June Enggano earthquake occurred at the edge of the rupture area of the 1833 subduction earthquake. The first strike-slip subevent within the subducting plate triggered a thrust subevent on the plate interface, which comprised at least 35% of the total moment and ruptured SE away from the 1833 earthquake. The 18 June earthquake in the Wharton Basin is one of the largest shallow strike-slip faulting earthquakes ever recorded. A small second subevent with reverse slip is required to fit the body waves. The orientation of both subevents in our preferred model is consistent with the current stress field in the region. Both the June 2000 earthquakes are consistent with recent models of distributed deformation in the India-Australia composite plate. The occurrence of the Enggano earthquake implies that the stress field within the Indian plate continues to a depth of 50 km in the subducting slab. The purely strike-slip source model of the Wharton Basin earthquake obtained by Robinson et al. [2001] matches the P waves very poorly and fits the S waves no better than our preferred model. The strike-slip subevents of both earthquakes had few aftershocks and higher stress drops than the subduction thrust subevent of the Enggano earthquake. This difference is consistent with previous observations of oceanic and subduction earthquakes.

SUMMARY The discrimination between chemical and thermal heterogeneity in the Earth's mantle remains one of the most important and challenging questions to be answered by observational and theoretical geophysics. To answer this question requires a thorough knowledge of the ratio between compressional and shear velocity anomalies. We describe results of a joint inversion for compressional and shear velocity in the mantle using a large and diverse data set consisting of traveltimes, complete waveforms and surface wave dispersion measurements. A horizontal tessellation consisting of 362 spherical splines is used to parametrize the model, which is approximately equivalent to a spherical harmonic of degree 18 in resolution. The model contains peak variations (from PREM) of up to ±7 per cent in S velocity and ±2.5 per cent in P velocity in the upper mantle. These variations decrease to ±1.5 and ±0.6 per cent, respectively, at 1000 km depth and reach ±2.5 and ±1.0 per cent, respectively, in the D″ region. The rms ratio of S to P velocity perturbations is fairly constant between 2.0 and 2.5 in the lower mantle, but a local minimum in this ratio occurs at a depth of approximately 1700 km. Resolution tests show that the recovery of P and S velocity is not geographically uniform, but also show that this amplitude ratio is well resolved between 670 and 2700 km depth. We also observe a persistent negative correlation between bulk‐sound and shear velocity throughout the same region of the mantle. In addition, the model contains a minimum correlation between P and S velocity between 670 and ∼1100 km. This feature is supported by both the favourable outcome of resolution experiments and the poor fit provided by the starting model from the inversion (in which the P and S velocities are perfectly correlated) to our data set of P‐wave traveltimes. The power spectra of both P and S velocity heterogeneity are similar, although we note a slightly larger dominance of degree two in the spectra for P velocity in the mid‐mantle where resolution is highest.