The base of the mantle, the D'' layer, plays an important role in the global dynamics of the Earth as a thermal and chemical boundary layer. We must know its structure to understand the mantle convection, the birth of hot spots, and the structure and stability of superplumes. During the last 20 years, the imaging of this layer as improved, and it now appears as very complex. The main seismological structures such as the seismic velocity discontinuities on top of D'', the Ultra Low Velocity Zones (ULVZ) and the seismic anisotropy are known only in a few places, and they appears to depend on the geodynamic and mineralogic contexts. The seismic observations also suggest a large amount of velocity heterogeneities at small scales which are still a matter of debate. The main reason for that is the very “smooth” image of D'' layer provided by actual tomographic models. So, a high resolution seismic tomography of this region is a major goal to understand the nature and the influence of D'' layer on the mantle dynamics, in particular the link between the seismic structure and the perovskite/postperovskite phase transition.

The Goal of this PhD Thesis is to perform a new high resolution seismic tomography of the lower mantle in order to understand its dynamics (in relation with the geodynamics team of the laboratory). For that, some theoretical and methodology developments dealing with 3D sensitivity kernels of seismic waves are necessary to improve the resolution of tomographic models.

Up to now, different methods (based on mode summation, ray theory...) have been developped to compute the sensitivity kernels of body waves of the mantle (P, PP, S, SS, PcP, ScS...) and the core (PKP). However the waves the most sensitive the seismic structure of D'' layer are the ones with grazing incidence at the base of the mantle, like Pdiff, Sdiff et PKPab. These waves cannot be properly described by the ray theory. The mode summation method allows to take into account the diffraction effects and to compute the sensitivity kernels for periods larger than 10s. However the PKPab and Pdiff waves have a significant amount of energy at higher frequencies. Our team began to develop a new method of sensitivity kernel computation coupling the Born and Langer approximations. Next step is the extension of this method to PKP phases. This approach remains however very expensive in terms of computation time. The wavelet decomposition method actually developed in collaboration with Li Zhao allows a much more effective computation. The candidate will extend this method to the cases of PKP and Pdiff phases at high frequencies.

The use of sensitivity kernels is also strongly dependent of the seismological data processing, because it is necessary to know precisely the frequency content of the waves for which the travel time and amplitude anomalies are measured. Our team has developed a non-linear waveform inversion method which allows the simultaneous determination of travel times, amplitudes and average waveform approximating the source time function of the event. The data base actually contains travel times, amplitudes and waveforms for most of the teleseismic body waves. In order to reach the goal of global mantle tomography, the data base should be increased by the processing diffracted waves at the core-mantle boundary (Pdiff, Sdiff). This step requires to modify the data analysis method to take into account the diffraction effects. The imaging of both

This PhD thesis enters in the framework of a whole research team implying different persons with different skills inside the seismo-tectonic team. This research topic has already be funded by ANR, University and INSU (CNRS) projects. It will bring to the candidate new skills on various methods of data processing, tomography and theoretical seismology, as well as mantle dynamics. This PhD thesis is funded by priority reasearch topic of the french ministery of research.

The candidate must possess a master degree of earth sciences or physics. Knowledge of geophysics is favoured.

Deadline for application is 22^th June 2009. Foreign candidates will have to translate their Master degree diploma in a French embassy.