A new 1D $V_p$ and $V_s$ velocity model of the western Rift of Corinth, Greece, using a fully non-linear tomography algorithm

TL;DR

This study develops a highly accurate 1D velocity model of the western Rift of Corinth using a fully non-linear tomography algorithm, significantly improving earthquake location precision and revealing velocity and depth distribution features.

Contribution

It introduces a novel non-linear tomography approach combining velocity model exploration with earthquake location grid search, achieving more accurate velocity models and earthquake depths.

Findings

30% reduction in P and S residuals with the new model

Significant decrease in velocities down to 3 km depth

Focal depths are generally shallower with the new model

Abstract

The objective of this study is to propose a new updated accurate 1-D $V_p$ and $V_s$ velocity model of the western Rift of Corinth for precise absolute earthquake locations. The methodology used to obtain this new model associates two fully non-linear algorithms, a complete exploration of the $V_p$ and $V_s$ model space combined with a grid search method for earthquake locations to minimize the P and S arrival time residuals. We calculated the misfit function for approximately 2.10 6 velocity models. For the best model (minimum misfit) we observed a global decrease of 30 % for both the P and S residuals with respect to the commonly used velocity model. The main features of this new model compared to other studies is a significant decrease of $V_p$ and $V_s$ velocities from surface down to a depth of approximately 3 km and a variable $V_p/V_s$ ratio decreasing from 1.86 at the surface to 1.78 at a depth of 8 km. The major influence on the locations of events is a global decrease of focal depths that range from a few hundred metres for deep events to more than one kilometre for shallow earthquakes.