Earth Planets Space, Vol. 51 (No. 10), pp. 1135-1143, 1999
Xavier Garcia1, Juanjo Ledo1,2, and Pilar Queralt2
1Geological Survey of Canada, 615 Booth Street, Ottawa, Ontario, K1A 0E9, Canada
2Departament de Geodinàmica i Geofísica, Universitat de Barcelona, Spain
(Received November 9, 1998; Revised May 13, 1999; Accepted May 17, 1999)
Abstract: In the last two Magnetotelluric Data Interpretation Workshops (MT-DIW) the participants were asked to model the Kayabe magnetotelluric dataset, a dense (100 m) grid of thirteen lines, with thirteen stations in each line. Bahr's phase-sensitive skew and the Groom and Bailey decomposition were used to select those lines for which the data could be considered two-dimensional. For these lines we used a 2D inversion algorithm to obtain a series of resistivity models for the earth. Finally, we constructed a 3D model using the 2D models and critically examined the validity and practicality of this approach based on 3D model study. We found that in the Kayabe dataset case the common practice of using 2D models to depict 3D models, can only be used to create a starting model for 3D interpretation. The sequential 2D models as a representation of a 3D body is unacceptable in terms of fit to the observed data. We question the validity of some of the conductivity structures in the 2D models, as they can be mere artifacts created by the algorithm to match 3D effects.