Java Tomography System (JaTS), a Seismic Tomography Software Using Fresnel Volumes, a Fast Marching Eikonal Solver and a Probabilistic Reconstruction Method: Conclusive Synthetic Test Cases

Abstract.

Problems related to landscape management, natural hazards and civil engineering involve subsurface structures that can be delineated by geophysical imaging. Seismic tomography can accurately characterize a medium according to its velocity variations. Traditional seismic travel time tomography based on ray-tracing methods assumes that the waves’ frequency is infinite. Therefore, only the medium located along the ray path has an impact on the wave propagation. In subsurface tomography, the infinite frequency assumption does not hold, as targets have about the same size as the wavelength. The seismic waves’ propagation is affected not only by the medium along the shortest travel time path but also by the medium located in its vicinity. In this study, Fresnel volumes are used to determine the medium affecting the wave propagation given the seismic waves’ frequency. The choice of the travel time computation and reconstruction methods determines the overall efficiency and soundness of the tomography process. In this research, a second order Fast Marching eikonal solver is used for computing travel times. The Fast Marching Method is an original approach that propagates a monotonously expanding wave front in a medium. It is fast, reliable and easy to implement in both 2D and 3D. An innovative probabilistic approach enables the iterative reconstruction process based upon Fresnel volumes. This study compares the performances of JaTS, our java Fresnel volume tomography software to those of Sardine, a ray-tracing tomography software, over an unfavourable synthetic case.