A surface distortion decomposition for vector CSAMT data over one-dimensional earth

Single source CSAMT data, also known as vector CSAMT data, has been routinely collected in recent years and processed using the conventional 2-D induction model which does not incorporate galvanic distortion effects, although the importance of these has been widely recognized. The small number of data available in vector CSAMT though, does not allow models that simultaneously incorporate both 2-D induction and local galvanic distortion. Either the 2-D inductive model with no distortion or alternatively a 1-D model with up to two distortion parameters can be used. Two 1-D distortion models were tested that are specialized cases of the general Groom and Bailey's (1989) MT decomposition method, (A) one which assumes the distortion parameter splitting (s) can be made zero choosing an appropriate coordinate system, and (B) other in terms of polarization ellipses that assumes the distortion parameter shear (e) can be zeroed by rotating the principal axes to the measurement axes. The former produces the distortion parameters twist and shear while the later generates twist and splitting. Computer algorithms were implemented and a data set was used to test the models and later compared with the classical 2-D model. Except for the magnitudes which differ by a scaling factor, the apparent resistivities computed with the three different approaches show very similar electrical distributions but the phases show discrepancies. The distortion parameters distribution of the distortion model (A) for the used data set has a complicated pattern and are frequency dependent. Distortion parameters obtained using model (B) define better the conductive structure and are approximately constant in a reasonable frequency interval. The results of the test suggest that the zero-split assumption is likely not valid while the zero-shear approach is a more realistic one, although further testing is necessary to confirm this.