25.01.2015 Views

IASPEI - Picture Gallery

IASPEI - Picture Gallery

IASPEI - Picture Gallery

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

IUGG XXIV General Assembly July 2-13, 2007 Perugia, Italy<br />

(S) - <strong>IASPEI</strong> - International Association of Seismology and Physics of the Earth's<br />

Interior<br />

JSS013 Poster presentation 2250<br />

Finite difference seismic travel times and non-linear seismic tomography:<br />

recovery of slab-back-arc structures in the Mediterranean region<br />

Dr. Paola Serretti<br />

Andrea Morelli<br />

The main structural features in the upper mantle beneath the Africa-Eurasia collision belt are the<br />

seismically fast subducted lithosphere, and the slow back-arc basins. Seismic tomography has been able<br />

to image these features, with high relevance for tectonics and geodynamics, with increasing resolution.<br />

Although potentially affected by irregular spatial distribution, travel times of high-frequency P and S<br />

waves represent the data with best resolving power, and are often used in studies targeted at imaging<br />

the finer details of upper mantle structure. In strongly heterogeneous media, the seismic travel time<br />

inverse problem can be strongly non-linear, as ray paths in the real earth can significantly deviate from<br />

those computed in the reference model. We test recovery of upper mantle structures by nonlinear P-<br />

wave travel time tomography, trying to reconstruct a previously known, realistic wave speed structure.<br />

We use the real distribution of seismic sources and stations, and calculate P first arrival travel times by<br />

the accurate and efficient finite difference scheme of Podvin and Lecomte (1991), based on systematic<br />

application of Huygens principle, adapted to work in spherical geometry. We use different meshes for<br />

the crust (2 km spacing) and the mantle (6 km) to account for Moho depth with high precision. The<br />

same technique is used to compute synthetic data, and to calculate partial derivatives along the ray for<br />

the inversion. We use a perturbative iterative approach for the inversion, starting with a one<br />

dimensional prior model. Different strategies are tested and compared in the nonlinear inversion. We<br />

verify that realistic upper mantle structures strongly deflect seismic rays, and the correct paths can only<br />

be approached after a few iterations. Although linear inversion appears in fact able to delineate the<br />

main features quite well, general three-dimensional ray tracing -- capable of dealing with strong ray<br />

deflections -- and non-linear inversion show to be important to describe the finer details of the<br />

structure.<br />

Keywords: travel time, tomography, recovery test

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!