Deutsch
 
Datenschutzhinweis Impressum
  DetailsucheBrowse

Datensatz

DATENSATZ AKTIONENEXPORT

Freigegeben

Konferenzbeitrag

Electromagnetic induction studies with long-periodic geomagnetic variations in Europe - I. Theory and methods of data analysis

Urheber*innen

Schmucker,  Ulrich
27. Kolloquium, 2017, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung, External Organizations;

Externe Ressourcen
Volltexte (frei zugänglich)

EMTF_2017_Breklum_34b_Schmucker.pdf
(Verlagsversion), 21MB

Ergänzendes Material (frei zugänglich)
Es sind keine frei zugänglichen Ergänzenden Materialien verfügbar
Zitation

Schmucker, U. (2018): Electromagnetic induction studies with long-periodic geomagnetic variations in Europe - I. Theory and methods of data analysis. - In: Becken, M., Hölz, S. (Eds.), - Protokoll über das 27. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung: Breklum, 25.-29. September 2017, 27. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung (Breklum 2017), 113-185.


Zitierlink: https://gfzpublic.gfz-potsdam.de/pubman/item/item_3208964
Zusammenfassung
The theoretical foundations of the various methods of magneto-variational sounding (MVS) are developed from first principles. Because only time variations of the Earth’s magnetic field are involved, these methods respond exclusively to the tangential-electric (TE) mode of the electromagnetic field on and above ground, presuming that the inducing source field is also in this mode. With increasing period MVS results become less and less sensitive to lateral resistivity contrasts at shallow depth, which is demonstrated. It sets these methods apart from magneto-telluric soundings (MTS) subject to persistent surface effects due to anomalous electric field variations in the tangential-magnetic (TM) mode. This work concentrates on the analysis of daily variations and associated activity-related variations, yielding response estimates for periods between three hours and two days. The relevant depth range of penetration extends from 250km to 750km and includes the transition from a resistive upper mantle (# 100Wm) to a conducting deeper mantle ( 1Wm). The ultimate purpose of this study is to obtain information about the degree of lateral uniformity in resistivity beneath Europe. Among MVS methods the gradient method is the most versatile one, relating the vertical component of geomagnetic variations to the spatial derivatives of their horizontal components. The connecting transfer function is the C-response. Within certain limits, which are specified, the method can be applied without concern about the spatial structure of the inducing source field, which is tested with response estimates for variations from two different sources: Quasi-periodic daily variations and transient storm-time variations. All calculations are carried out in spherical coordinates, Alternative MVS methods based on global presentations of the horizontal components by one or more spherical harmonics are of restricted applicability. Tests show that the gradient method gives the best results. A new generalised version of this method removes the constraint about one-dimensionality. It combines gradient sounding with geomagnetic depth sounding (GDS), provided the source field is of sufficient spatial complexity to rule out representation by a single spherical harmonic. The resulting multivariate relation involves up to five transfer functions, including a tensor C-response in close relation to the tensor impedance for the TE mode in the electric field. Input variables are three of the four spatial derivatives of the horizontal components and these components themselves. They are derived from polynomials fitted to the horizontal components in a network of observing sites. An eigen-value analysis ascertains that the performed fit of 2-dimensional second degree polynomials is a numerically stable process. A second eigen-value analysis concerns the inversion of the spectral matrix in multivariate regressions, indicating that regularisation is required, when all five transfer functions are to be found. Eigen-solutions identify the two GDS transfer functions together with the Berdichevsky-average of the tensor C-response as the best resolvable combination. Two kinds of errors are derived for (robust) estimates of transfer functions, distribution-dependent errors and jack-knife errors. For univariate regressions both errors are of comparable size, but for trivariate regressions the former turn out to be twice as large. The data base for exemplary soundings are two years of hourly mean values (1964-65) at 35 European observatories.