Joint inversion of magnetotelluric and seismic data based on random resistivity and velocity distributions

Speaker: Zhang Luolei
We develop a model construction method based on common gridded model with random resistivity and velocity distributions to meet the needs of joint inversion and the complicated model with large variations of physical properties. The forward calculation for magnetotelluric (MT) field using finite element method and the seismic travel-time computing using improved ray-tracing method, are both suitable based on this kind of unified gridded media with random resistivity and velocity distributions. Incorporating with the improved very fast simulated annealing algorithm, we fulfill the synchronous joint inversion of MT and seismic data based on this kind of common gridded model. The test of the complicated model with uneven terrain and uncommon interface and large variations of resistivity and velocity, shows the effects of joint inversion in accurately determining the resistivity and velocity structures. It improves the limitation of former studies confined in simple models of joint inversion. Moreover, the joint inversion of the observed data in an area with low-quality surveyed seismic data, proves it is practicable in improving inversion accuracy with constrained prior information.