where
is the skin depth in metres, 
is the resistivity of the homogeneous
medium through which the electromagnetic field is diffusing, 
is the angular
frequency of the signal and µ is magnetic permeability, usually assumed
to take its free space value everywhere. Although this expression is derived
assuming simple plane wave fields, it provides a useful guide to the attenuation
of more complicated dipole fields in the crust. The diffusive nature of
the fields means that sharp boundaries and anomalies with a scale length
less than a skin depth are not well resolved.
Our experiment involves the transmission of an electromagnetic signal at
discrete frequencies (in contrast to the time domain system of Edwards &
Chave, 1986) from a horizontal electric dipole (HED) source to an array
of remote sea bottom receivers which detect the horizontal electric field.
The frequency at which signals are transmitted must be carefully chosen,
taking into account the possible resistivity and scale of structures which
might be encountered. If the frequency is too low, the skin depth in the
seafloor is very long so the signal is not inductively attenuated between
the source and the receiver and cannot resolve crustal scale structure.
If the transmission frequency is too high, the skin depth is very small
and signals only penetrate the shallow part of the crust. All but the very
shortest range signals are attenuated to such an extent that they can no
longer be detected above the noise (e.g. Flossadottir & Constable, 1996).
Another consideration is the noise spectrum. At frequencies below approximately
0.3 Hz, microseism noise is significant (Webb & Cox, 1986), and
at lower frequencies still, ionospheric noise starts to leak through the
conductive ocean. The range of frequencies which can be usefully employed
for probing crustal structure to depths of a few kilometres or more is therefore
between a few tenths and a few tens of Hertz.