Through a collaboration with former colleagues (Menichelli et al., Geophysical Research Letters, 2023) the role of subducting geometrical features in tuning interseismic coupling and seismogenic behaviour is explained for a complex rheological plate boundary. We show using four analogue models with (i) a flat interface, (ii) high and (iii) low friction seamount, (iv) a low friction patch) that the presence of a seamount reduces recurrence time between earthquakes, interseismic coupling, and fault strength. This suggests it acts as a barrier; 80% of the ruptures concentrate in flat regions that surround the seamount and only smaller magnitude earthquakes nucleate above it (Fig. 1). The low-friction zone, which mimics the fluid accumulation or the establishment of fracture systems in natural cases, seems to be the most efficient in arresting rupture propagation.
In a collaboration with MSc student Maaike Fonteijn 2D seismotectonic numerical models are extended with seamounts of different geometries and lithologies on an oceanic plate, which were pushed to subduct under a Southern Chilean margin (Fonteijn, van Rijsingen, van Dinther, Tectonophysics, in prep.). We find three different styles of seamount subduction as a function of pore fluid pressure ratio and cohesion, which are accretion at the trench, subduction with flattening and stretching at depth, and intact subduction. A pore fluid pressure ratio ≤ 0.80 in combination with a cohesion ≥ 20 MPa is required for seamounts to subduct intact and reach seismogenic depths. The subduction of seamounts is accommodated by widespread permanent deformation around the seamount, through faulting of the fore-arc.
During this process, stresses in the overriding plate are altered and, in general, normal stresses in the wake of the seamount are reduced (up to 15 MPa), promoting stable slip and affecting earthquake ruptures. Results furthermore show that subducting seamounts have a profound effect on the observed seismicity. Most important are segmentation of the megathrust, concentration of seismic events on the seamount’s plateau, reduction of the amount of seismic events updip and downdip of the seamount, an increasing amount of splay faults, a shift of hypocenters to the seamount’s leading flank and increased interseismic coupling above the seamount. This implies that the seamount itself is a seismic asperity and its leading flank an area prone to earthquake nucleation, while the wake of the seamount is a barrier to earthquake propagation (Fig. 2). This dual behavior is able to explain seemingly contradictory observations that are made in the literature. Worth noting is that despite the barrier effect of a seamount’s wake, large earthquakes of up to Mw 8.5 are still likely to occur. Finally, if seamounts are instantaneously imposed rather than subducted, seismicity still largely occurs at the same location as for a smooth interface and fewer splay faults are activated (Fig. 2). This demonstrates the importance of simulating self-consistent subduction over millions of years for processes occurring on time scales of seconds. The absence of subduction of the seamount over long times in the analogue models may also be an explanation for the discrepancy between observations of asperity and barrier behaviour with respect to the analogue models.
