The paper sets out a method for structural analysis of seismotectonic data using centroid moment tensors and associated hypocenters from the Global Centroid Moment Tensor project, here illustrated for aftershocks from the 2004 great Sumatran earthquake. We show that the Sumatran segments of the megathrust were subject to compression in a direction near to orthogonal with the margin trend, consistent with the effect of relative movement of the adjacent tectonic plates. In contrast, the crust above the Andaman Sea segments was subject to margin‐orthogonal extension, consistent with motion toward the gravitational potential well accumulated due to prior lateral (westward) rollback of the subducting edge of the northward moving Indian plate. Since this potential well is largely defined by topography, this episode of margin‐orthogonal extension is at least in part “gravity driven.” It did not last long. Within 15 months, an earthquake cluster across an Andaman Sea spreading segment showed a return to kinematics driven by relative plate motion. The transition can be explained if fluid activity temporarily reduced basal friction (or effective stress) but then led to healing so that the megathrust once again began to develop friction‐locked segments. The influence of slab rollback is in developing a gravitational potential well facing the megathrust, hence drawing the overriding crust toward it in the immediate postrupture phase while the megathrust is in a weakened state. Plate tectonics dominates during interseismic gaps, once the megathrust heals, and regains frictional resistance.

Abstract. The Martabe deposits in Sumatra, Indonesia formed in a shallow crustal epithermal environment (200–350 °C) associated with mafic intrusions, usually recognised in domes, adjacent to an active right-lateral wrench system. Ten samples containing alunite were collected for high-resolution geochronology, to determine if overprinting fluid systems could be recognised by dating alunite separates. The heating time for each step was chosen to ensure reasonable uniformity in terms of the incremental percentage of 39Ar gas release during each of many steps, allowing age spectra to be analysed using the method of asymptotes and limits. Several distinct growth events could be recognised. In addition, each sample was subjected to ultra-high-vacuum (UHV) furnace step-heating, and 39Ar diffusion experiments conducted at the same time as 40Ar/39Ar geochronology, to determine the argon retentivity of the mineral grains being analysed. The heating schedule ensured Arrhenius data uniformly populated the inverse temperature axis, with sufficient detail to allow the application of the Fundamental Asymmetry Principle (FAP) during analysis of the Arrhenius spectrum. Results show activation energies between 370–660 kJ/mol. Application of Dodson’s recursion determines that closure temperatures would range from 400–560 °C for a cooling rate of 20 °C/Ma, which is higher than any possible temperature to be expected in the natural system. This gives confidence that the ages represent growth during periods of active fluid movement and alteration, since the deposit formed at temperatures < 200 °C at a depth of < 2 km. We conclude that gold in the Purnama pit was the result of fluid rock interactions during very short-lived mineral growth episodes at ~ 2.25 and ~2.00 Ma.

Abstract. The Martabe deposits in Sumatra, Indonesia formed in a shallow crustal epithermal environment (200–350 °C) associated with mafic intrusions, usually recognised in domes, adjacent to an active right-lateral wrench system. Ten samples containing alunite were collected for high-resolution 40Ar/39Ar geochronology, to determine if overprinting fluid systems could be recognised. At the same time, ultra-high-vacuum (UHV) furnace step-heating 39Ar diffusion experiments were conducted, to determine the argon retentivity of the mineral grains being analysed. The heating schedule chosen to ensure Arrhenius data uniformly populated the inverse temperature axis, with sufficient detail to allow the application of the Fundamental Asymmetry Principle (FAP) during data analysis. The heating time for each step was chosen to ensure reasonable uniformity in terms of incremental percentage gas release during each step. Results show activation energies between 360–500 kJ/mol, with normalised frequency factor between 1.89e14s−1 and 8.62e18s−1. Closure temperatures range from 390–519 °C for a cooling rates of 20 °C/Ma, giving confidence that the ages represent growth during periods of active fluid movement and alteration. The Martabe deposit formed at temperatures

&lt;p&gt;Here we present the first 4D tectonic reconstruction that models the Vrancea slablet and incorporates the floated slab as a constraint on the magnitude of slab rollback during collapse of the Palaeo-Pannonian Basin. Seismic tomographic images provide insight into the geometry and tectonic history of subducted slabs. High velocity anomalies can be interpreted as &amp;#8216;cold&amp;#8217; lithosphere penetrating &amp;#8216;warmer&amp;#8217; lower velocity asthenosphere, and 3D models created using the &lt;em&gt;SKUA-GOCAD&lt;/em&gt; modelling software. Combined with information from the 3D distribution of earthquake hypocentres, we thereby obtain a simple approximation to slab geometry beneath the Vrancea region. The resultant DXF was imported into the &lt;em&gt;Pplates&lt;/em&gt; tectonic reconstruction software, and floated back to the Earth&amp;#8217;s surface. The method utilised assumes no significant deformation (stretching, buckling, folding, shortening) during or after subduction, so that the obtained geometry estimates the pre-subduction configuration. The resultant floated slab is then incorporated as a constraint on 2D + time tectonic reconstructions. We apply a double-saloon-door rollback model, which involves propagation of a slab tear along the mid-Hungarian lineament. Each saloon-door rolls back independently of the other and this leads to two epochs of extension. AlPaCa is &amp;#8216;pulled&amp;#8217; eastwards and rotated counter-clockwise as the western saloon-door rolls back. The Tisza-Dacia unit is then &amp;#8216;pulled&amp;#8217; eastward, and rotated, but in a clockwise sense as the eastern saloon-door rolls back. Once the subduction hinge reached the East European Platform, the slab was left hanging. Gravitational forces then drove slab-boudinage and detachment in a similar fashion as occurs today beneath the Hindu Kush. This model explains the large opposing-sense vertical-axis rotations that occurred during convergence of the AlPaCa and Tisza-Dacia terranes. The zipper fault model rotates the microplates without requiring large-scale thrusting. Interpretation of the Mid-Hungarian lineament as a zipper-fault system is also consistent with the geodynamic effects expected because of tearing in a subducting plate leading to a double-saloon-door rollback. The vertical extent of the slab is roughly 300 km, which only fills half of the basin, consistent with the double-saloon-door roll-back model interpretation.&lt;/p&gt;

No abstract