Seismotectonic characterization of the Turkana Depression from deep-learning-based seismicity relocation

6 months

Context and issues

The Turkana Depression and the Chew Bahir rift form a 200-km-wide lowland connecting the South Ethiopian Rift to the Kenya Rift. In contrast to its neighbor rifts, the depression is marked by a relatively low volcanic activity, an anomalously thin crust ( 10–20 km) relative to the flanking Ethiopian and Kenyan plateaus ( 40 km), and a mosaic of juxtaposed, N-S trending basins bounded by both new structures and reactivated pre-existing faults. This combination of thin crust, limited magmatic input, and structurally complex basin architecture makes the Turkana Depression an atypical segment of the East African Rift System, and a key natural laboratory for examining how continental extension evolves in the absence of a dominant magmatic activity.

Rift evolution and strain localization are driven by the interplay of several mechanisms, the most critical ones being the tectonically induced stress field, the volume of extruded magma, any preexisting lithospheric heterogeneity or feedbacks with surface processes. In a magma-poor, structurally inherited setting such as the Turkana depression, a few questions remain unanswered : (i) to what extent is the reactivation of pre-existing structure controlling the extension ? ; (ii) does the thin crust of the depression reflect a large-wavelength-zone of lithospheric weakness inherited from prior rifting episodes ? ; and (iii) how is strain transferred and partitioned between adjacent basins across the SER–Turkana–Kenya corridor ?

At the scale of the seismic cycle, it remains unclear how extensional deformation is accommodated : through seismic slip, aseismic creep, or transient magmatic processes such as dyke intrusion ? Despite geodetically-derived extension rates below 5 mm/yr, several Mw > 6 earthquakes have been instrumented in the region. Are some faults partially locked and potentially prone to moderate-to-large earthquakes generation ? How much of the accumulated strain energy budget is seismically dissipated ?

Objectives

We aim to reprocess continuous waveform data from the TRAILS experiment (Turkana Rift Array to Investigate Lithospheric Structure ; networks Y1 – Ebinger, 2018, and 6R – Bastow, 2019 ; 34 broadband stations, NW Kenya and SW Ethiopia, Jan. 2019–June 2021) using modern deep-learning earthquake detection workflows to build a substantially denser, higher-resolution earthquake catalog than manually picked published catalogs. We will use this new catalog to resolve fine-scale fault structure, characterize deformation style, and discuss the mechanical controls on strain localization across the depression. In particular, we will image fault geometry and segmentation via precise relocation. If the quality of the catalog allows, we will also search for earthquake swarms (indicative of transient deformation) or for repeating earthquakes (recurrent rupture of the same asperity), and characterize larger events (moment tensor, stress drop). The work performed during this internship will likely result in a publication in a solid earth or tectonics international journal.

Methodology

• Phase picking : automated P/S phase detection using PhaseNet .
• Phase association : event association via GaMMA or PyOcto (SeisBench framework), using a homogeneous or 1D regional velocity model.
• Absolute location : NonLinLoc (probabilistic, non-linear location) ; NonLinLoc SSST to be evaluated as an alternative.
• Relative relocation : double-difference relocation using GrowClust and/or HypoDD to sharpen catalog spatial resolution and delineate fault-plane geometries.
• Downstream analysis (contingent on catalog quality) : moment tensor inversion for selected events, spatial/temporal clustering analysis, b-value mapping, identification of seismic swarms and repeating earthquakes, stress-drop estimation.

Skills gained

• Earthquake source seismology basic knowledge (relocation, moment tensor, stress drop)
• Scientific python programming
• AI-based signal processing (PhaseNet, GaMMA/PyOcto, SeisBench)
• Statistical seismology basics (b-value, swarms, repeaters, clustering)
• Physical interpretation of rupture processes in a rift-tectonic context

Profile sought

Master's student (M2) or equivalent (diplôme d'ingénieur·e, ...) in Earth sciences or Physics or Signal processing, with an interest in seismology and active tectonics. Solid background in Python programming and basic signal processing required, so that the candidate is autonomous. Prior experience with seismological tools and/or machine learning applied to geophysical data is a plus, but not required.

How to apply

Send a CV and a cover letter to thea.ragon univ-grenoble-alpes.fr and caroline.chalumeau univ-grenoble-alpes.fr. If your academic path has been unconventional, feel free to briefly explain how it shaped your interest and motivation for this internship. Applications will be reviewed on a rolling, first-come-first-served basis, and we will respond to candidates as promptly as possible. Shortlisted applicants will be invited to a remote interview, which may be followed at a later date by a short technical test.

References

If you don't have access to one of these references, you can send an email to Théa to request a PDF version.

• Cynthia Ebinger. (2023). Crust and mantle structure and the expression of extension in the Turkana Depression of Kenya and Ethiopia (p. 1000000 MB) [SEED data]. International Federation of Digital Seismograph Networks. https://doi.org/10.7914/SN/Y1_2018
• Ian Bastow. (2023). Turkana Rift Arrays to Investigate Lithospheric Strains—UK component [SEED data]. International Federation of Digital Seismograph Networks. https://doi.org/10.7914/SN/6R_2019
• Ian Bastow, Derek Keir, Atalay Ayele, Christopher Ogden, Birhanu Abera, & Sisay Alemayehu Angere. (2024). Northern Lake Abaya Broadband Network [SEED data]. International Federation of Digital Seismograph Networks. https://doi.org/10.7914/SN/1Q_2019
• Musila, M., Ebinger, C. J., Bastow, I. D., Sullivan, G., Oliva, S. J., Knappe, E., Perry, M., Kounoudis, R., Ogden, C. S., Bendick, R., Mwangi, S., Mariita, N., Kianji, G., Kraus, E., & Illsley-Kemp, F. (2023). Active Deformation Constraints on the Nubia-Somalia Plate Boundary Through Heterogenous Lithosphere of the Turkana Depression. Geochemistry, Geophysics, Geosystems, 24(9), e2023GC010982. https://doi.org/10.1029/2023GC010982
• Nutz, A., Ragon, T., & Schuster, M. (2022). Cenozoic tectono-sedimentary evolution of the northern Turkana Depression (East African Rift System) and its significance for continental rifts. Earth and Planetary Science Letters, 578, 117285. https://doi.org/10.1016/j.epsl.2021.117285
• Sullivan, G., Ebinger, C. J., Musila, M., Perry, M., Kraus, E. R., Bastow, I., & Bendick, B. (2024). Kinematics of rift linkage between the Eastern and Ethiopian rifts in the Turkana Depression, Africa. Basin Research, 36(5), e12900. https://doi.org/10.1111/bre.12900