Conductivity‐depth imaging of airborne electromagnetic step‐response data

Publisher Website
Google Scholar
Abstract
An adaptation of the Macnae‐Lamontagne method allows transform of airborne step‐response electromagnetic (EM) data to a conductivity‐depth image. The algorithm is based on a nonlinear transformation of the amplitude of the measured response at each delay time to an apparent mirror image depth. Using matrix algebra, the set of mirror image depth‐delay time data pairs can then be used to derive a conductivity section. Data can be efficiently processed on a personal computer at rates faster or comparable to the rate required for collection. Stable conductivity fitting as a function of depth is obtained by damping the matrix inversion by specification of the first‐ and second‐derivative smoothness weights of the fitted conductivity‐depth sounding. Damping parameters may be either fixed or varied along the profile; their choice can be constrained by geologic control. Stability of the process is enhanced by accounting for the transmitter and receiver tilts. The mirror image depth‐delay time data can also be used directly with simple regression to obtain the best‐fitting thin‐sheet and half‐space models. With one novel assumption, the thin‐sheet model can be converted to a thick‐sheet overburden model without prespecification of either its conductivity or thickness. Depending on the geology, these simple models may prove quite useful. The conductivity imaging algorithm has been applied to a test data set collected with the SPECTREM system. The stability and speed of the imaging process were confirmed and have demonstrated airborne EM sounding to depths well over 400 m in an area with quite conductive sediments. Comparing the results with a better resolved image obtained from ground UTEM data shows that the airborne data can adequately define the geometry of the uppermost conductor encountered in the section. The geophysical results are consistent with geologic control and measurements of resistivity obtained from well logs.

This publication has 0 references indexed in Scilit: