Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4740880 | Journal of Applied Geophysics | 2010 | 9 Pages |
The magnitudes of the initial amplitude of the magnetic resonance sounding (MRS) signals from an aquifer located in a layered electrically conductive earth, are nonlinear functions of water content distribution. Occam's inversion method is adapted to the nonlinear inversion problem. In the case of an electrically conductive medium, the Jacobian matrix is analytically evaluated at the beginning of the inversion. And the uniqueness of the inversion can be partially solved by imposing the flattest and smoothest model constraints on the optimization problem. Synthetic MRS signals from resistive and conductive earth, as well as field data, have been inverted by Occam's method. The results indicate that with the help of Occam's inversion, a true model can be obtained from an initial model of homogeneous water content. Furthermore, for noise-free MRS signals, both the flattest and smoothest models reveal correct water content distributions. When signals are contaminated by noises, the case is different; and the smoothest model might have a lower water content distributing in a larger range than that of the true model, while which might be obtained by utilizing the flattest model Occam's inversion.