- 1Department of Earth Sciences, University of Oxford, Oxford, United Kingdom (manuel.mojicaboada@st-annes.ox.ac.uk)
- 2Department of Earth Sciences, University College London, London, United Kingdom
- 3Institut Terre et Environnement de Strasbourg, Université de Strasbourg, Strasbourg, France
The transport and release of fluids play a fundamental role in subduction zone settings, including for the genesis of seismicity and arc volcanism. Understanding how fluids escape from the slab and move into and through the overlying mantle wedge is key for understanding these processes. The presence and distribution of fluids within these regions has been primarily investigated through estimates of the Vp/Vs ratio, which serves as a key indicator of fluid and magma content.
Local earthquake travel-time tomography (LET) has been extensively used to image Vp/Vs structure in subduction zones, providing critical insights into the morphology of the subducted slab and the properties of the overlying mantle wedge. LET typically involves the inversion of seismic data to simultaneously determine P- and S-wave velocity models alongside hypocentre locations. Differences in data coverage and quality lead both to the need for different regularisation in the Vp and Vs inversions and to mismatched resolution between the two velocity models. Together, these effects make a simple division to compute Vp/Vs inappropriate and result in unreliable and uninterpretable estimates of the Vp/Vs ratio. Therefore, several widely used algorithms incorporate a direct inversion for the Vp/Vs ratio, but these typically assume identical ray paths for P and S waves. The requirement to have both S- and P- wave arrivals also means that valuable data are discarded. Moreover, many LET schemes provide limited information on model uncertainty and resolution, complicating the assessment of model reliability.
In this work, we address these issues by utilising the Backus-Gilbert based SOLA method (Zaroli, 2016) to obtain robust and consistent Vp/Vs models in local earthquake tomography. The SOLA method provides direct control over model resolution and has recently been applied to obtain multiple physical parameters with the same local resolution. Here, we present the methodology and preliminary work towards implementing SOLA with LET, with the goal of improving constraints on fluid distribution in subduction zones. As an initial step towards this, we use data from the Lesser Antilles subduction zone (Bie et al, 2022) to obtain Vp and Vs models within a linearised framework. We will present preliminary findings for the Vp/Vs ratio including uncertainty and resolution information.
How to cite: Mojica Boada, M., Koelemeijer, P., Hicks, S., Zaroli, C., and Serra, E.: Towards imaging fluids in subduction zones through Backus-Gilbert inference of Vp/Vs structure: initial application to the Lesser Antilles, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-14860, https://doi.org/10.5194/egusphere-egu26-14860, 2026.