The BC87 data exhibit distortions due to three-dimensional structures
at almost all scale sizes from that of the electrode array (<100 m) to
that of the 150 x 50 km plutonic Nelson batholith. These distortion e
ffects must be identified and removed, as far as possible, prior to in
terpretation. A galvanic model of these distortions is shown to be val
id for most of the frequency range of observation, but 3D induction is
significant at certain sites for some bands. A first-order regression
was fit to the decomposition recovered long period E-polarization app
arent resistivities to correct for the remnant local site ''static shi
fts''. We model the distortion- and level-corrected data using both 2D
forward trial-and-error fitting and Occam2 smooth inversion. The crus
tal section of the resulting models correlates well with some geologic
al features. However, the long period phase difference between the E-
and B-polarization lasting for more than a decade is difficult to inte
rpret without recourse to introduction of a layer with differing elect
rical conductivity in the two orthogonal horizontal directions, i.e.,
an anisotropic layer. There is greater than an order of magnitude diff
erence in the conductivities, with the higher value in conductivity be
ing in a direction which is commensurate with the Juan de Fuca plate-p
ush direction.