A unique GPS velocity field that spans the entire Southeast Asia region is presented. It is based on 10 years (1994–2004) of GPS data at more than 100 sites in Indonesia, Malaysia, Thailand, Myanmar, the Philippines, and Vietnam. The majority of the horizontal velocity vectors have a demonstrated global accuracy of ∼1 mm/yr (at 95% confidence level). The results have been used to (better) characterize the Sundaland block boundaries and to derive a new geokinematic model for the region. The rotation pole of the undeformed core of the Sundaland block is located at 49.0°N–94.2°E, with a clockwise rotation rate of 0.34°/Myr. With respect to both geodetically and geophysically defined Eurasia plate models, Sundaland moves eastward at a velocity of 6 ± 1 to 10 ± 1 mm/yr from south to north, respectively. Contrary to previous studies, Sundaland is shown to move independently with respect to South China, the eastern part of Java, the island of Sulawesi, and the northern tip of Borneo. The Red River fault in South China and Vietnam is still active and accommodates a strike‐slip motion of ∼2 mm/yr. Although Sundaland internal deformation is general very small (less than 7 nanostrain/yr), important accumulation of elastic deformation occurs along its boundaries with fast‐moving neighboring plates. In particular in northern Sumatra and Malaysia, inland‐pointing trench‐perpendicular residual velocities were detected prior to the megathrust earthquake of 26 December 2004. Earlier studies in Sumatra already showed this but underestimated the extent of the deformation zone, which reaches more than 600 km away from the trench. This study shows that only a regional Southeast Asia network spanning thousands of kilometers can provide a reference frame solid enough to analyze intraplate and interplate deformation in detail.

Data collected at approximately 60 Global Positioning System (GPS) sites in southeast Asia show the crustal deformation caused by the 26 December 2004 Sumatra-Andaman earthquake at an unprecedented large scale. Small but significant co-seismic jumps are clearly detected more than 3,000 km from the earthquake epicentre. The nearest sites, still more than 400 km away, show displacements of 10 cm or more. Here we show that the rupture plane for this earthquake must have been at least 1,000 km long and that non-homogeneous slip is required to fit the large displacement gradients revealed by the GPS measurements. Our kinematic analysis of the GPS recordings indicates that the centroid of released deformation is located at least 200 km north of the seismological epicentre. It also provides evidence that the rupture propagated northward sufficiently fast for stations in northern Thailand to have reached their final positions less than 10 min after the earthquake, hence ruling out the hypothesis of a silent slow aseismic rupture.

Global Positioning System (GPS) measurements made in Sulawesi, Indonesia, from 1992 to 1999 detected coseismic and transient postseismic deformation related to the 1 January 1996, Mw = 7.9 earthquake on the North Sulawesi (Minahassa) trench. These motions are superimposed on the long‐term secular motion (40 mm/yr) of the left‐lateral Palu fault in central Sulawesi and continued for about 1.5–2 years. Following the earthquake, a string of earthquakes (of magnitude >6) migrated along the Minahassa trench, from west to east. Subsequently, two earthquakes of magnitude >6 occurred on or near the Palu fault migrating toward the south. Modeling the increase in Coulomb stress generated by the successive earthquakes agrees with the hypothesis of interacting events. An unclamping effect, possibly due to fluid migration in the Palu area, is also suggested by the stress computations and the detected (GPS) displacements.

No abstract

No abstract