AN EFFECTIVE RIDGE-REGRESSION PROCEDURE FOR RESISTIVITY DATA INVERSION

The inversion of geoelectrical sounding data for the subsurface resistivity structure is a difficult task due to the nonlinear and nonunique nature of the problem. An effective algorithm for inverting resistivity data by ridge regression is described. It uses a line search procedure in a two-stage least-squares minimization process resulting in a stable and rapidly convergent algorithm. A FORTRAN-77 routine RIDGE based on the algorithm, its implementation and a driver program for magnetotelluric inversion MTINV, are discussed. The procedure is applied to magnetotelluric (MT) and transient electromagnetic (TEM) data to demonstrate its effectiveness and general applicability.