The first-arrival traveltime tomography is currently a standard approach for imaging the near surface structures. But it cannot resolve complex situation such as hidden lowvelocity layers. Early arrival waveform inversion is a robust approach for dealing with complex structures, but it may take significant computational efforts. Furthermore, in practice, we found that the results from waveform inversion may not allow fitting the first arrival traveltimes anymore because of nonuniqueness of the

more »

... odel solutions. These theoretical and practical issues motivated us to develop joint first arrival traveltime and waveform inversion. We apply a regularized nonlinear conjugate gradients method to simultaneously invert both traveltime and waveform data. One of the difficulties for performing waveform inversion alone is the lack of effective preconditioning in nonlinear inversion. Therefore, it requires hundreds of iterations to converge to a desired minimum misfit. With the inclusion of traveltime data in joint inversion, however, the inverse matrix of traveltime sensitivity could serve as an effective preconditioner to waveform inversion. Thus, it helps significantly speeding up waveform inversion. We demonstrate the effectiveness of the joint inversion by applying to both synthetics and real data.