Research

Subduction Zones

Subduction is a key element of the plate tectonic mechanism, introduced by Alfred Wegener in 1912, and which is the foundation of modern geosciences. Seismic, tsunamigenic and volcanic risks associated to subduction zones have, at all times, motivated the scientific community to better understand these complex tectonic features. Despite the fact that a lot effort was put into studying these zones, several questions remain unanswered. Here are some of them that excited my curiosity:

• Which parameters control the occurrence of mega-thrust earthquakes? Is it linked to the age of the subducted lithosphere or to the convergence rate at the trench?
• What controls the propagation of ruptures during mega-thrust earthquakes? Have fractured zones or subducted ridge an influence on that propagation?

Trying to answer these questions is a difficult task. During my thesis, supervised by Wayne C. Crawford and Valérie Ballu, I focused my work on the central part of the Vanuatu subduction zone. This region is particularly interesting because of the proximity of the Islands of Santo and Malekula to the trench (<35 km), which allows a high-resolution seismic study using onshore stations. Also this zone is characterized by the subduction of bathymetric highs such as the d’Entrecasteaux ridge and the West Torres Plateau under the North Fiji Microplate that influence considerably the seismicity and the geodynamics in the region.

My work implied defining an accurate earthquake catalog for the region in order to be able to define the main tectonic features in the region. The catalog helped us to define the geometry of the subduction interface, the downdip limit of the seismogenic zone and also to highlight a shallow double seismogenic zone subparallel to the subduction interface.

To better understand the subduction mechanism we also performed 2D mechanical simulations using a finite difference software, Pylith. The idea was to simulate the surface displacement and compare it to real data recorded by onshore GPS stations. By integrating our slab geometry we were able to explain the surface motions with negligible residuals. Based on this successful experiment, we derived our model to the Lesser Antilles subduction zone. This project is currently conducted by Valérie Ballu and Pierre Sakic at LIENSs.

Automatic Picking Procedures

Earthquake hypocenter locations are needed to map existing faults and to document their activity, both of which are of prime importance in defining hazards and forecasting events. The number of seismic stations around the world is rapidly growing, which should provide much more detailed information about seismically active regions, but only if seismic events can be accurately and uniformly picked on each station. Automatic picking procedures (APP) are needed to handle the larger datasets and they must be reliable, precise, and capable of distinguishing different phase onsets and adapting to different site and/or instrument characteristics. Compared with manual picking, APP save time and should be more consistent, because manual picks can differ between analysts. In seismology, the most commonly used event detection algorithm is the short-term average/long- term average (STA/LTA) detector proposed by Allen (1982), which is based on the ratio of the two averages calculated on sliding windows over the trace. This algorithm is rapid and remains useful for detecting events in continuous databases, but it generally gives significantly different results from manual picking. In Baillard et al, 2014 we developed a new accurate and adaptive picking procedure based on the Kurtosis. This PSPicker has proven its efficiency in different case scenarios: microseismicity in the mid-atlantic ridge recorded by a 4 OBS network; rockfalls events in the Dolomieu crater (La Réunion Island); seismicity in the Vanuatu subduction zone.