Question 3 of 10: Planning, Acquiring, Processing and Interpreting a Magnetotelluric (MT) Survey

Question 3: Planning, Acquiring, Processing and Interpreting a Magnetotelluric (MT) Survey (16.66% of paper)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

MT uses natural, passively-occurring electromagnetic energy (from lightning and solar-wind/ionospheric interactions) as its source, so it needs no transmitter and can reach far greater depths of investigation than any controlled-source EM or DC resistivity method — this is exactly what makes it the method of choice for deep, large-scale targets.

Situation where MT is useful. Geothermal exploration is a classic application: a geothermal reservoir typically has a hot, saline, and often clay-altered (hence conductive) cap and reservoir zone overlying more resistive basement, at depths of 1–5 km that are well beyond the practical reach of DC resistivity or most controlled-source EM systems, but squarely within MT's depth range.

Planning. Station spacing and array geometry follow the expected target depth and lateral extent — for a geothermal system a few kilometres across, station spacing of roughly 250–1000 m along one or more profile lines (or a 2-D grid, if a 3-D model is intended) crossing the interpreted structure is typical. Frequency range needed is set by the required depth of investigation via the skin-depth relation $\delta\approx503\sqrt{\rho/f}$ (m, with $\rho$ in Ω·m and $f$ in Hz) — lower frequencies penetrate deeper, so a broadband instrument recording from roughly 10²–10&sup4; Hz (near-surface resolution) down to 10²–10&sup4; s period (deep basement resolution) is chosen to bracket the target depth.

Acquisition. At each station, two orthogonal horizontal electric-field dipoles (Ex, Ey, typically 50–100 m electrode spacing) and three orthogonal magnetic-field coils (Hx, Hy, Hz) are deployed and logged simultaneously for a period long enough to capture the lowest frequencies needed (hours to days for deep targets). A remote reference station, recording simultaneously at a site far enough away to share the same natural source field but be free of local cultural EM noise, is used to suppress correlated noise during processing — essential near infrastructure such as power lines, pipelines or rail, which are common noise sources in geothermal-prospect terrain.

Processing. Time-series E and H data are Fourier-transformed and combined (with remote-reference and robust statistical processing to reject noisy time windows) into the MT impedance tensor $Z$ as a function of frequency, from which apparent resistivity $\rho_a(f)=\frac{1}{\omega\mu_0}|Z(f)|^2$ and phase $\phi(f)$ curves are derived at each station, along with tensor decomposition (e.g. Groom–Bailey) to recover the regional strike and remove galvanic twist/shear distortion. Decomposition cannot determine the frequency-independent site gain (static shift), so that is corrected separately — typically with a co-located TDEM sounding or by statistical/spatial averaging of the high-frequency apparent resistivity across neighbouring stations.

Interpretation. Apparent resistivity/phase curves from each station are inverted (1-D layered inversion for a first pass, 2-D or 3-D smooth-model inversion for the full profile/grid) to produce a resistivity cross-section or volume. For the geothermal case, the interpreter looks specifically for a shallow, low-resistivity clay-cap layer overlying a moderately resistive reservoir zone and resistive basement — the clay-cap conductor is itself often the primary exploration target, since its top and thickness closely track the top of the hydrothermal alteration zone that bounds the reservoir.