13:45 〜 14:00
[SEM14-01] Magnetotelluric Rotational Invariant Impedances and Galvanic Distortion: A Review and Its Application.
★Invited Papers
キーワード：Magnetotellurics, Galvanic distortion, Rotational Invariant
Galvanic distortion is a well-known and challenging problem in magnetotellurics method (MT). The distortion inherently deteriorates the interpretation of MT data. In contrast to the advancement to MT data acquisition, inversion software, and interpretation, approaches to solve galvanic distortion remain questionable. Not only the solution, its definition are also arguable. In general, the galvanic distortion is an alteration of the regional electric field by local heterogeneities, and mathematically an under-determined problem. However, the geometry of the structures which is said to be regional or local is not static but depends on geology and MT array design, e.g., array size, site spacing, and selected period range. The galvanic distortion problem itself is complicated and can then be subjective. Since the problem was recognized, several attempts have been developed to analyze or solve the galvanic distortion. Each approach has its own pros and cons and limitations.
In this talk, I will begin with the definition of galvanic distortion, its mathematical model, and numerical examples. Then, a review of the effect of galvanic distortion on the rotational invariant impedance, determinant (det) and sum-of-the-squared-element (ssq), will be given. It is shown that the ssq impedance is less affected by the galvanic distortion in comparison to the det impedance, which is traditionally used to derive 1-D regional structure.
Their implications are: (1) An approach to reliably estimate the 1-D regional structure using the average ssq impedance; (2) The local distortion indicator to detect the geometric distortion, the distortion that changes the geometry of the impedance tensor; (3) A statistical approach to estimate site gain, scaling of the impedance tensor magnitude. A concept of local distortion indicator is crucial. It may help reduce the damage caused by solving the problem without knowing its presence. The site gain is claimed to be a non-determinable parameter without other independent geophysical information. The proposed approach has shown that estimating site gain is statistically viable.
To yield the reliable 3-D inversion results, the phase tensor-based inversion is suggested. However, the phase tensor itself does not contain information of magnitude. The choice of initial/prior models is, therefore, important. The presented method to estimate the 1-D regional structure using the average ssq impedance is a promising candidate. Still, the inverted models must be interpreted with caution as the phase tensor may include the inductive effect from the distorters. In addition to the previous works, an application to the field data, collected from Kanchanaburi, Thailand, will be presented.
In this talk, I will begin with the definition of galvanic distortion, its mathematical model, and numerical examples. Then, a review of the effect of galvanic distortion on the rotational invariant impedance, determinant (det) and sum-of-the-squared-element (ssq), will be given. It is shown that the ssq impedance is less affected by the galvanic distortion in comparison to the det impedance, which is traditionally used to derive 1-D regional structure.
Their implications are: (1) An approach to reliably estimate the 1-D regional structure using the average ssq impedance; (2) The local distortion indicator to detect the geometric distortion, the distortion that changes the geometry of the impedance tensor; (3) A statistical approach to estimate site gain, scaling of the impedance tensor magnitude. A concept of local distortion indicator is crucial. It may help reduce the damage caused by solving the problem without knowing its presence. The site gain is claimed to be a non-determinable parameter without other independent geophysical information. The proposed approach has shown that estimating site gain is statistically viable.
To yield the reliable 3-D inversion results, the phase tensor-based inversion is suggested. However, the phase tensor itself does not contain information of magnitude. The choice of initial/prior models is, therefore, important. The presented method to estimate the 1-D regional structure using the average ssq impedance is a promising candidate. Still, the inverted models must be interpreted with caution as the phase tensor may include the inductive effect from the distorters. In addition to the previous works, an application to the field data, collected from Kanchanaburi, Thailand, will be presented.