Geophysical Tomography Group

From Wikipedia, the free encyclopedia

The group's logo reflected its name, but also a famous formula in least-squares optimization: transpose of G times G.
The group's logo reflected its name, but also a famous formula in least-squares optimization: transpose of G times G.

The Geophysical Tomography Group was part of the Institut de Physique du Globe de Paris (Jussieu Campus). During the years 1985-2000 it made pioneering developments in the domain of nonlinear fitting of seismic waveforms. Many of the past members of the group (created by professor Albert Tarantola) are now leading individuals in Academy, Industry and within Governmental Agencies.

Contents

[edit] Context

By the year 1985, the imaging methods used by the oil industry were based in signal-processing concepts, and it seemed that modern methods, based on careful waveform modeling and waveform fitting optimization were desirable. After a couple of theoretical founding papers by Albert Tarantola, the team tried to made real-life demonstrations that the methods of Inverse Theory could be applied to the imaging problem typical of seismic exploration. This group is largely credited for having opened some research avenues that are now being explored by many other geophysical teams.

[edit] Members

Around Professor Albert Tarantola, the Geophysical Tomography Group was formed. The members, along the years, were Alexandre Nercessian, Odile Gauhtier, Antonio Pica, Eric Lewin, Philippe Lognonné, Luc Ikelle, Zhang Xiankang, Wafik Beydoun, Manuela Mendes, Denis Trézéguet, Roel Snieder, Hakim Jannane, Edward Crase, Cao Di, Zvi Koren, Evgeny Landa, Mark Noble, Bruno Riollet, Jonas Lindgren, Gunter Röth, Heiner Igel, Satish Singh, M.Y. Xie, Peter Mora, Stefan Nielsen, Bertrand Maillot, Roland Martin, Dmitri Pissanrenko, Marwan Charara, Christophe Barnes, Miguel Bosch, Dominique Rodrigues, Sophie-Adelaïde Magnier, Frédéric Donzé, Emmanuelle Juvé, Agathe Girard, Tom Houlder, Lianjie Huang, Emmanuel Dormy, Wojciech Debski, Dimitri Komatitsch, Albane Saintenoy, Emmanuel Chaljub, and Hughes Djikpéssé. Most of these members are now leading individuals in Academy, Industry and within Governmental Agencies. Inside the Institut de Physique du Globe de Paris, this group was famous for hard working, but also for the friendly atmosphere it helped develop. Tuesday's "pizza seminars" were famous, and usually had a large attendance.

[edit] Scientific findings

First ever fully elastic nonlinear waveform fitting of exploration data (vertical seismic profile). From Charara et al., 2000.
First ever fully elastic nonlinear waveform fitting of exploration data (vertical seismic profile). From Charara et al., 2000.

The papers published by the Geophysical Tomography Group demonstrated that fully nonlinear seismic waveform fitting is possible for seismic exploration data, using elastic modeling and gradient-based optimization techniques (adjoint methods). These findings have been largely ignored by the oil industry, but the methods developed by the Group are now being used for whole Earth tomography. The group's work in Monte Carlo method for waveform fitting was, perhaps, ahead of its time, and is only slowly being integrated in present-day developments.

[edit] Influence

The influence of this team in the present-day scientific landscape is easy to measure. For instance, between the papers cited by the University of Utah for the Basics of Waveform Tomography" more than half of them were produced by this group. Quite often, the work of the group is described as pioneering (document from ETH-Zürich) (document from Stanford University). One of the founding papers of the group (Inversion of seismic reflection data in the acoustic approximation), that opened the way to full waveform nonlinear fitting, was ony published (in the prestigious journal Geophysics) after a difficult battle with the Editors. Yet, a citation analysis conducted by the Society of Exploration Geophysicists showed that this paper was the most cited of all papers published by Geophysics that year.