Erzincan manyetotelürik verilerinin değerlendirilmesi (profil-a)

Abstract

13 mart 1992 Erzincan depreminden sonraki süre içinde, yurt içinden ve yurt dışından çok sayıda araştırmacı bölgede incelemelerde bulunmuş ve sismolojik kayıtlar almışlardır. Depremin ana şok ve art sarsıntılarının incelenmesi ve bölgenin Jeofizik özellilerinin belirlenmesi amaçlı İ.T.Ü. Araştırma Fonu projesi (proje no:362) desteğinde sekiz noktada manyetotelürik (MT) ölçüler alınmıştır. Kuzeyden güneye doğru sırası ile K.Köşkünler, Akyazı, Mahmutlu ve Konakbaşı olmak üzere dört istasyona ait MT verilerin değerlendirilmesi tezin konusunu oluşturmaktadır. Özellikle manyetotelürik'de modelleme, düz ve ters çözüm konuları irdelenmiş ve değişik modellerin MT tepkileri hesaplanarak, özdirençleri frekansla değişen (karmaşık özdirenç CR) tabakalar içeren ortamların MT tepkileri hesaplanmıştır. Manyetotelürik'te İP etkisi olarak adlandırılan bu durumu en iyi tanımlayan model olarak Cole-Cole dispersiyon modeli alınmıştır. MT verilerinin spektral analizinden elde edilen verilerden yararlanarak, her bir ölçü noktası için verilerin uyumluluk ve yönbağımlılık analizi yapılarak çarpıklık etkisi hesaplanmıştır. Erzincan'a ait görünür özdirenç verilerinin değerlendirilmesinden, yüzeyde kalınlığı 850 m ile 3.8 km arasında değişen sedimanter tabaka olmak üzere, yer altının derinlere doğru iletken-dirençii ardalanması şeklinde uzanan tabakalardan oluştuğu ortaya çıkmıştır. Alüvyonun altında ise kabuk, iletken ara tabakalar içerecek şekilde 35 km'ye kadar uzanmaktadır. 15 km civarlarında bulunan ilk iletken tabakanın üst kabuk sınırındaki geçiş bölgesine karşılık geldiği düşünülmektedir. İkinci önemli iletken bölge ise 30-35 km civarlarında bulunmuştur.

The magneîotelluric (MT) method had its origin in the 1950's. It's a way of determining the electrical conductivity distribution of the subsurface from measurements of natural electric and magnetic fields on the surface. Using the electromagnetic fields which are ranged at IQ"3^ hz frequency band it is posible to study the electrical substructure of the Earth's crust and upper mantle by measuring the tangential (horizontal) components of both the electrical and magnetic fields at an observation point. The theory for the magnetotelluric sounding proposed firstly by Cagniard is based on several assumption. These include: - Assamption that both the primary and secondary electromagnetic fields are planar. The fields do not change character at parallel interfaces. - Assumption that the medium is horizontally stratified. - Both the electrical and magnetic field are linearly polarized. This means that the direction of the electric or magnetic field does not change with time. - The oscillation of the electromagnetic field components are harmonic function of time. These assumptions permit us to computation of apparent resistivity very simply by using Cagniard's formula. The MT method depends on the penetration of electromagnetic energy into earth. The depth of investigation changes with the period of oscillations. With increasing period, the skin depth of EM wave increases and currents penetrate deeper. On the other hand when the frequency of the oscillation increases, the skin depth decreases and currents consentrate closer to the Earth's surface. The ratio of electric field intensity to the magnetic field intensity is called electrical impedance and under certain conditions this impedance is a function of the electrical properties of the medium. At high frequencies, because of the skin effect, the impedance carries information only about theproperties of the uppermost layer. At lower frequencies, the depth of penetration of the field increases and thus the magnetotelluric method can be used to carry out "sounding" at a single observation point. The experience with both the geomagnetic method and telluric method had some influence on the early stages of the developments of magnetotelluric sounding. The method can be used for deep sounding or for method for application to problems in geophysical exploration where the depth of investigation not to exceed 10-15 km in the usual case. In the case of a uniform half-space one can see that the modulus of the impedance decreases with increasing conductivity and with increasing period of oscillation. The electric field is shifted in phase -45° with respect to the magnetic field in a uniform half space and the phase is independent of the rezistivity. In the case of n-layered Earth it is assumed that there is an electromagnetic field present which does not depend on the coordinates in the horizontal plane, but is function only of z-coordinate and distribution of resistivities. Both the electric field, Ex, and the magnetic field, Hy, must be continuous across the boundaries between layers, since there are neither charge accumulations nor free poles allowed at these boundaries. The wave impedance at the bottom of the first layer must equal the wave impedance at the top of the second layer, and so on. The variations of electromagnetic fields which were recorded during 1992's summer in Erzincan are analyzed to obtain their spectra and apparent resistivity as a function of frequency are computed from the IX spectra. Since all geophysical interpretation is based on comparing observed earth response data with model data, interpretation of magnetotelluric data is also consist of matching the computed plots of apparent resistivity against frequency to curve calculated for simplified models. In section two, many of two and three-layer models and their theoric apparent resistivity and phase curves calculated. In the case of two-layered earth at high frequency the skin depths are small enough that no energy penetrates to the basements. Apparent resistivity is therefore asymptotic to p1 at high frequencies and the upper layer is not penetrated. When frequencies are low enoughthe apparent resistivity approaches p2. At very low and very high frequencies the phase retardation of the magnetic field is 45°, and does not depend on frequency. Deviation from 45° retardation occur only for intermediate frequencies. The inversion procedure is applied as follows. An estimate of th probable resistivity-depth model is made arbitrarily, and a theoretical curve, known as a "forward solution", is computed and compared with the field curve. The difference between the two curves represents the error in the arbitrary first estimate. Changing parameter describing the resistivity-depth function slightly the error reduced. This trials are continued until the error between theoretical curve and the experimental data is reduced to an acceptable value. Inversion of magnetotelluric sounding for subsurface resistivity disturbution is a nonlinear and nonunique problem. The goal of inversion is to determine a hyphotetica! Earth model whose responses are identical to the observed data. X An effective algorithm and program for inverting resistivity data verseus frequency by ridge regression (Marquart-Levenberg Method) proposed by Meju is used. Because of nonlinearity, problem generally is linearized using Taylor's expansion and standart least-squares method used iteratively to refine an initial guess model. The diffrences between the observed data and those calculated for model wanted to be minimum. For this purpose a damping process applied to the partial derivative matrix A. The parameter correction Am can be obtained by application of singular value decomposition (SVD) method. In this method, A matrix is factored into product of three other matrixes, the data space, parameter space eigenvectors and diagonal matrix containing the positive eigenvalues of A. The effect of complex resistivity (CR) into magnetotelluric diffision process, in other words, resistivity-frequency dispersion in rocks by magnetotelluric are investigated. The magnetotelluric method can identify the zones in the Earth's crust where the electric conductivity is abnormally high because of the presence of a geothermal system. In the frequency-domain the IP phenomenology in rocks consist of a process which is known in the specific literature as dispersion of resistivity against frequency. This means that the resistivity of a polarizable meteriai is not constant as the frequency of the field is changed. In geophysical applications, the Cole-Cole model is the most suitable representation of the IP characteristic dispersion function. XI Modeling of complex resistivity in magnetotelluric shows that the amount of distortion of the curves increases as the values of chargeability, time constant and frequency dependence increase. MT sounding apparent resistivity and phase theoretical curves were calculated for various n-layered earth models. In some of these models one layer taken as with complex resistivity. Calculated apparent resistivity was decreased with increasing frequency when the incident fields are polarized the layer of complex resistivity (CR). The effects of dispersion parameters (chargeability m, time constant x, and frequency constant c) on modelling of such earth models are investigated during calculations. Modelling technique by using forward solution is applied to the apparent resistivity depth sounding curves obtained from the magnetotelluric field data recorded in Erzincan Basin after 1 3rd march 1 992 earthquake. For the MT stations Küçük Köşkünler, Akyazı, Mahmutlu and Konakbaşı sites deep jeoelektrical models are described. Six or seven horizontally layered models are interpretated as earth structures for all stations. Automatic inversion technique proposed by Meju(1991) is finally trained on fitting field and modeled curves. Erzincan Plain is situated between 39 15 - 39 50 longitudes and 39°32 - 39°52 latitudes in Northeast Anatolia Region. Length of the plain is 45-50 kms. from northwest to southeast and width is 5-20 kms. During 1992's summer magnetotelluric data were collected in Erzincan plain at eight different stations. Four of these station lie on a profile approximately from North to East are studied in this thesis. Recorded analog signals are digitized at equal intervals by using a digitiser. For each of frequency band different sampling rates were used. The apparent resistivities as a function of frequency is obtained by processing of these field measurements. For these purpose an interactive program were used to obtain the apparent resistivity. Apparent resistivity XII values are smoothed by a seven point moving avarage. interpretated geoelectrical models of MT sounding curves in all stattions are indicates a conductive shallow layer (65-850 Qm) which has thickness about 850-3800 meters. It may be due to alluvion contaminated with clay, rough stone. Thickness of this layer is approximately 2000-850 meters at Konakbaşı and K.Köşkünler, 2.8-3.8 kms at Mahmutlu and Akyazı. A resistive layer (8900-9500 Qm) of thichness 6.8-7.5 km estimated beneath of the alluvion along the profile Konakbaşı-Küçük Köşkünler. Geoelectrical models results another highly conductive layer (6.5-38 Qm) at depth 12-15 km at Akyazı, K.Köşkünler, Konakbaşı and Mahmutlu. Deep of the geoelectrical structure of Erzincan basin is characterized by lower crust's inhomogenities respect to MT data