Page 82 - Geothermal Energy Systems Exploration, Development, and Utilization

58 2 Exploration Methods
is noise in any of the field components, the coherence will be reduced. When
coherence drops below reasonable values (0.85–0.90), it is common practice to
discard the apparent resistivities that are calculated (Ritter, Junge, and Dawes,
1977).
This problem of noise is usually addressed with reference to data of another
site situated beyond the sphere of influence of the artificial signal (Gamble et al.,
1979; Clarke et al., 1983), often referred to as a remote reference site.Inareas where
uncorrelated noise has been a problem in obtaining MT soundings, this procedure
has resulted in significant improvements in the quality of the data, provided the
electromagnetic noise at the sites is not correlated. Remote reference stations are
therefore often situated on islands, where influence of cultural noise is low, for
example, the stationary reference site of the Japanese and South Korean Geological
Services on the Japanese island of Jeju, the island of Capraia in the Aegean Sea
off the west coast of Italy, used for Italian sites in Tuscany, such as Larderello and
Travale (Manzella et al., 2010) or the island of R¨ ugen in the Baltic Sea used for the
MT survey in Gross Sch¨ onebeck, Germany, within the I-GET project (Mu˜ noz et al.,
2010). Over a portion of the frequency range where noise is a particular problem
(from 0.1 to 10 Hz, the so-called dead band, where signal is particularly low), the
multiple-station approach has permitted data to be obtained where previously it
had been impossible. Experience taught that a combination of local and very far
remote sites – to face both local high frequency noise and far, planar noise sources
– proved tobethe most effectivesolution.
As might be expected, the spectral analysis of long data series, combined with
the need for extensive tensor rotation and testing of the spectral values, results in a
volume of processing that is as time consuming and as costly as the acquisition of
the data. Rapid analysis in the field is necessary as the MT method does not always
provide useful results, even after measurements have been made with reliable
equipment and for a long time. If the natural electromagnetic field strength is
unusually weak during a recording period, or if there is some phenomenon which
precludes an effective analysis of the fields, it may be necessary to repeat the
measurements at a more favorable time. When the analysis is done in the field,
decisions about reoccupying stations and siting additional stations can be made in
a timely manner that will reduce overall operating costs. After data processing, the
impedance components are scaled to obtain the apparent resistivity, ρ a , similar to
that used in DC resistivity techniques, and phase, φ, for given frequencies. The
processed data can then be used for interpretation. Apparent resistivity is defined
as the resistivity of the homogeneous earth, which would produce the measured
response at each frequency. Two data curves are defined both for resistivity and
phase, which are referred to the two pairs of orthogonal electric and magnetic
field horizontal components. Usually they are termed xy and yx,since they refer
to E x /H y and E y /H x in a Cartesian system. For a layered earth (Figure 2.7), the
apparent resistivity at high frequency is equal to the true resistivity of the surface
layer, where at lower frequencies it asymptotically approaches the resistivity of the
bottom layer.