Two Dimensional Joint Inversion Of Seismic Reflection Travel Time And Magnetotelluric Data

On the base of cross-gradient theory which is proposed by Gallardo and Meju，we have developed a2-D joint inversion algorithm of seismic reflection travel timedata and magnetotelluric (MT) data. Seismic forward using ray tracing method to findthe minimum walking paths, seismic inversion calculated using the least squaresmethod, adding the model regularization, increasing the constraints of the objectivefunction for the model disturbance, for that,we get more stable data of the model.Magnetotelluric （MT） forward using finite difference method, inversion usingGauss-Newton (G-N)method, the calculation process using a regularization term toconstraints on the inversion result to control the ill-posed problem, and make theconvergence faster. In this paper, using the central differences to discrete thecross-gradient equation makes the constraints on the model’s boundary more reliableand improves the exactness of inverse results.In this paper, we use the joint inversion algorithm, the core idea is: at eachiteration update, seismic velocity model is not only to make the data fit andregularization term minimum objective function, but also consider the last update ofthe magnetotelluric resistivity model to make it’s gradient intersection with seismicvelocity close to zero, while the magnetotelluric resistivity model in the next iterationupdate must also consider the seismic velocity model generated by the previousiteration, with this way,it iterates till the calculated sum of earthquake andmagnetotelluric reach our desired value.After the successful development of this algorithm FORTRAN code, wedesigned several theoretical model inversion, the calculations show that our inversionalgorithm with cross-gradient constraints can cause the velocity and resistivity toconstrain each other so that the results of inversion can get more reliable results.Comparing results show: the joint inversion is better than traditional separateinversion, which can reduce the multiple solution of geophysical inversion, Therefore,the2-D joint inversion algorithm of seismic reflection travel time data andmagnetotelluric (MT) data is verified to be validity and correctness.