Résumé
The role of subducting oceanic features on the seismogenic behavior of subduction zones has been increasingly addressed over the past years, although their exact relationship remains unclear. Do features like seamounts, fracture zones or submarine ridges act as barriers, prohibiting ruptures to propagate, or do they initiate megathrust earthquakes instead? With this question in mind, we aim to better understand the influence of subduction interface roughness on the location of an earthquake's hypocenter, rupture area and seismic asperity. Following the work on compiling a dual-wavelength subduction interface roughness (SubRough) database, we used this roughness proxy for a global comparison with large subduction earthquakes (M (sub W) > 7.5), which occurred since 1900 (SubQuake, new catalogue). We made a quantitative comparison between the earthquake data on the landward side of the trench and the roughness proxy on the seaward side, taking into account the most appropriate direction of roughness extrapolation. Main results show that areas with low roughness at long wavelengths (i.e. 80-100 km) are more prone to host large- to mega-earthquakes. In addition to this natural data study, we perform analogue experiments, which allow us to investigate the role subducting oceanic features play over the course of multiple seismic cycles. The experimental setup consists of a gelatin wedge and an underthrusting rigid aluminum plate (i.e. the analogues of the overriding and downgoing plates, respectively). By adding scaled 3D-printed topographic features (e.g. seamounts) on the downgoing plate, we are able to accurately monitor the initiation and propagation of ruptures with respect to the subducting features. Here we show the results of our natural data study, some preliminary results of the analogue models and our first conclusions on how the subduction interface roughness may influence the seismogenic potential of an area.