Subduction Channel Evolution in Brittle Fore-Arc Wedges — a Combined Study with Scaled Sandbox Experiments, Seismological and Reflection Seismic Data and Geological Field Evidence

* Final gross prices may vary according to local VAT.

Get Access


With a series of scaled sandbox experiments, we investigated the mass-flux patterns at the interface of convergent plates, with emphasis on the upper (brittle) part of subduction channels. Analysis of the particle displacement field integrated over short time periods shows that both types of simulated subduction channels (accretive and tectonically erosive) are characterized by episodically active thrusts (roof thrusts) at the top and a continuously active basal detachment. The short-term material flux reveals a complex temporal and spatial variability in the active mass-transfer processes within the subduction channel, and is particularly influenced by the activity of fore-arc structures (e.g. reactivation of backthrusts or duplexes). In the subduction channel, the localization of deformation also shows temporal and spatial fluctuations, which range from periodic kinematic cycles to unpredictable, apparently chaotic behaviour involving the activation and reactivation of shear zones. However, the location of the roof thrusts and their reactivation pattern during the periodic cycles is indicative of either tectonically erosive or accretive mass-transfer modes. In contrast to the short-term observations, the longterm material flux integrated over one kinematic cycle exhibited diagnostic patterns for the location of sediment accretion and subduction erosion. The series of accretive experiments shows that the combination of several parameters (initial wedge thickness, absence/presence of upper-crustal structures, and depth-dependent softening of the top of the subduction channel) can cause the same bulk effect in the upper plate (i.e., the migration of the center of uplift as an indicator of the position of rearward accretion). This experimental result precludes the determination of controlling parameters in nature.

We demonstrate that comparison of the subduction channel-related structures detected in the sandbox experiments with the structures of the accretive south-central Chilean fore-arc (37–38° S) and the tectonically erosive north Chilean fore-arc (21–24° S) is possible, when the restrictions of the analogue experiments (strongly idealized set-up and simplified material behaviour) and observational methods for nature (observation time window, spatial resolution) are taken into account. We show that the differences between analogue models simulating accreting and tectonically eroding subduction channels can be applied to distinguish these end-member types of subduction channels in nature. In contrast to the data from the analogue simulations, the reflection seismic profiles and the seismological data only reveal the geometry of the currently active subduction channel, which might have fluctuated over time. The results imply that subduction channels in nature could have a particle-velocity pattern at least as complex as those seen in the analogue experiments.