Session EM4: Natural EM Source Effects EM4-1 STUDY OF THE ELECTRIC FIELD HARMONICS IN THE SEA OF JAPAN

Transcription

1 Session EM4: Natural EM Source Effects EM4-1 STUDY OF THE ELECTRIC FIELD HARMONICS IN THE SEA OF JAPAN Daria Abramova (Institute of Terrestrial Magnetism, Ionosphere and Radio Wave Propagation RAS, Russia, Leonid Vanyan, Nick Palshin (both at Shirshov Institute of Oceanology, Russia) Eugenie Kharin (World Data Center, Russia) In 1996 Japan Sea Cable (JASC) was retired from telecommunications services and its ownership was transferred to the scientific community. The cable voltage across the Sea of Japan was monitoring several years, so to now much experimental data were accumulated. Five harmonics of solar daily variations on quite days were extracted reliably analyzing the uninterrupted voltage records of the submarine cable during nine month of It has been noted that up to 90% of field energy is concentrated in the periodic part of the measured field spectrum. Furthermore, the cable measurements as they had been made at the bottom possess a quite favorable feature: short-period variations are removed by a seawater layer what makes easier solar daily variation and its harmonic extraction. Magnetotelluric impedance and apparent resistivity for main solar daily harmonics were calculated using the submarine cable voltage and simultaneous magnetic field variation records at the Kakioka observatory (Japan). It makes clear that the impedance value, calculated from solar daily harmonics, is weaker influenced by the land sea conductivity difference than one what had been obtained for this region from continuum spectra of the magnetotelluric variations generated by the magnetospheric ring current. This result indicates the importance of the space structure of the external field inducing geomagnetic variations in forming the electrical currents system, induced in the 3D inhomogeneous Earth. The simultaneous influence of both the solar daily variation source configuration and the conductivity inhomogeneity of Far East region was number estimated with the modified P.Weidelt s algorithm for two types of source. The first was the plane wave, describing approximately field of the magnetospheric ring current variations, and the second was the ionospheric current curl, what could represent the solar daily variation. The result analysis shows the evident difference of the electrical field for these two types of source. In the case of plane wave the significant decrease of the field is observed in the Sea of Japan in comparison to the normal ocean value, as for the solar daily variation model the electrical field there increases the normal ocean value. In general the behavior of calculated electrical field confirms the observed difference between the values of apparent conductivity, calculated from solar daily harmonics, and those what have been obtained from continuum spectra of the magnetotelluric variations. This effect could been explained from the configuration of the shallow geoelectrical inhomogeneities in Far East region where Sea of Japan is situated between well-conductive ocean and low conductive continent. Geomagnetic field induced by the plane wave in the Sea of Japan is distorted by the coast effect. It seems to be similar to H-polarization and so the electrical field is decreased here. The electrical field, induced by the solar daily variation, on the contrary, looks like E-polarization and proves to be more intensive: its values in the Sea of Japan are between the normal ocean and the normal continental largeness.

2 EM4-2 SOURCE EFFECT MODELLING USING THE FENNOSCANDIAN 3D EARTH MODEL AND EQUIVALENT CURRENT SYSTEMS DERIVED FROM BEAR DATA Martin Engels (Uppsala University, Sweden, Antti Pulkkinen, Ari Viljanen (Finnish Meteorological Institute), the BEAR Working Group Traditional source effect model studies are using either simplified earth model (e.g. 1D) or simplified source model (e.g. line or band current models as electrojet approximations). Often, source models are prototypes of typical source field configurations but static in space. The aim of source field modelling should be both, a rather realistic 3D earth model and real dynamic 3D sources, realized here as follows: 1. Recently, a digital 3D crustal conductivity model based on all available data of the whole Fennoscandian Shield and surrounding areas has been compiled (Korja et al., 2002). This a- priori Earth model is used for volume 3D modelling using the X3D code by Avdeev, Kuvshinov and Pankratov. 2. Equivalent current systems have been derived for the whole recording period of the BEAR experiment (Baltic Electromagnetic Array Research) from about 70 magnetometer sites by Pulkkinen (note his separate paper, this workshop), using the method of elementary current systems developed by Amm. These equivalent current systems describe the same electromagnetic field on ground as the physical 3D current systems in the ionospheremagnetosphere system. The 3D modelling code by Avdeev, Kuvshinov and Pankratov allows to implement these equivalent current systems via a dense array of electrical dipoles. Thus real dynamic current systems can be used as excitation for the 3D earth model. Selected BEAR data events of polar substorm activity are analyzed by transforming the equivalent current systems from time into frequency domain. About 12 harmonics representing the frequency range of interest are used for 3D modelling. Comparisons of electromagnetic substorm response functions with those obtained by plane wave excitation (for the same Fennoscandian Earth model) demonstrate realistic source effects.

3 EM4-3 LOCAL SQ FIELD BEHAVIOUR AROUND JAPAN Masahiro Ichiki (IFREE, Japan Marine Science and Technology Center) Hisashi Utada (Earthquake Research Institute, University of Tokyo) Hisayoshi Shimizu (Earthquake Research Institute, University of Tokyo) Rikitake et al. (1956) analyzed magnetic variation in Japan during a geomagnetically quiet period between 1952 and 1955, and pointed out that there is a phase progress in the vertical component of Sq field at Kakioka Magnetic Observatory relative to those at other observatories. From this observation, they proposed a qualitative model in which the phase shift can be attributed to anomalous distribution of the mantle electrical conductivity. However, quantitative evaluation of this model has not been made because of the difficulty in 3-dimensional numerical modeling until very recently when Kuvshinov et al. (1999) indicated that most anomalous features in Sq field variations can be explained by the induction in the oceans. We try to further investigate this problem from a different viewpoint. Rikitake (1956) showed that there is no seasonal dependence of the phase shift. This was considered as one of the reasons for the interpretation that this feature is not of external but internal origin. To examine this more in details, we analyzed the geomagnetic data observed at 14 stations in and around Japan in 1997 when geomagnetic activities were most calm in recent years. From a preliminary investigation, we found that the relative phase shift has seasonal dependence. This result strongly indicated that the phase shift is not simply caused by the conductivity structure but probably characteristics of external field variations as well. In order to examine quantitatively whether the seasonal dependence is ascribed to the external field, we attempted to separate the Sq field into external and internal parts. Two methods, spherical harmonic analysis (SHA) and spherical cap harmonic analysis (SCH) were adopted for this purpose. In the SHA, geomagnetic data sampling with 1-minute interval of the world data center (WDC) in 1997 were used in addition to those at 14 stations in and around Japan. However, the result was not reliable enough to examine the possible presence of the seasonal dependence because of the nonuniformity of global distribution of the observatories. On the other hand, the SCH by using 14 stations data has clearly delineated seasonal variations in the external Sq field. However, the analysis also showed that the spatial pattern of the internal field is different from that of that of external field, which may suggest the induction by laterally heterogeneous conductivity in the Earth. We conclude that the seasonal dependence should be taken into account whn we study the conductivity structure by using Sq field variations.

4 EM4-4 PHYSICAL MODELLING OF ELECTROMAGNETIC FIELD OF AURORAL ELECTROJET Valentyna Kobzova (Carpathian Branch of Subbotin Institute of Geophysics, National Academy of Sciences of Ukraine), Borys Ladanivskyy, (Carpathian Branch of Subbotin Institute of Geophysics, National Academy of Sciences of Ukraine, Valery Korepanov (Lviv Centre of Institute of Space Research, National Academy of Sciences of Ukraine) It is known that at polar latitudes the field of geoelectromagnetic variations can strongly differ from the plane one. This means that the limits of application of the classic theory of MT method have to be outlined and the estimation of the error that may be caused by the deviation of the original source field from the plane-homogeneous one must be given for these regions. To investigate the electromagnetic field of source imitating the real 3D ionospheric electrojet the physical modeling method was used. The Earth s deep geoelectric cut was modeled in a electrolytic tank in the form of horizontally-homogeneous three-layered medium consisting of NaCl salt solution, ceramic plates and metalic basement. Above the liquid surface the U-like system of wires has been fixed imitating auroral electrojet with length about 1000 km, width 150 km, height above the Earth s surface about 125 km. Magnetotelluric sounding plots and spatial structure of electromagnetic field components of this auroral source scale model were obtained. The comparison of the results with those obtained in a plane homogeneous field and in the field of vertical magnetic dipole located at some height above the tank display the following regularities. In the electrojet model field some specific domains are outlined where the apparent resistivity plots are of the same type: over small area under the center of electrojet the plots of apparent resistivity are identical with those of the magnetic dipole, over a great territory under electrojet the results of MT soundings are very close to those obtained in the field of a planehomogeneous wave, substantial deviations from plane homogeneous source are observed only under the edges of electrojet Thus, in the auroral zone under the electrojet the results of MT soundings in the regions of periods of sec. are quite close to those obtained in a planehomogeneous field. The source effects can be significant and not always unique mainly in the regions under the "edges" of current spreading. The main error that can occur at interpretation of the results of deep MT sounding in the auroral zone is the underreading of depth values to the highly conductive mantle. This work was supported by INTAS grant.

5 EM4-5 Study of Auroral Source Field Effects on the Slave Craton, NW Canada Pamela Lezaeta, Alan Chave, and Rob Evans (Woods Hole Oceanographic Institution, Alan G. Jones (Geological Survey of Canada) This paper presents an investigation of source field effects of the electric and magnetic field components, from nineteen long period magnetotelluric (MT) sites collected on the floors of lakes throughout the Slave craton from 1998 to After preliminary bounded influence data processing, high quality MT and magnetic Z/B responses were obtained for periods ranging between 80 and 6,000 s, whilst lower quality data was obtained for the longer period band 5,000-25,000 s or more. Strong source field effects may be influencing these long period data, since the Slave craton is located in the midst the auroral zone. We first investigate qualitatively the spatial and temporal variations of the electric and magnetic fields by calculating power spectra for the field components and selecting daily and monthly time section lengths. The analysis shows that during high disturbance levels (given by the Kp geomagnetic index), the spectral amplitudes change drastically over daily and monthly section lengths, even at periods as short as 200 s. We also observed also seasonal differences between the daytime and nighttime variations, which suggest that the geomagnetic activity during the years 1999 and 2000 was localized in the auroral zone, since this could not be inferred from the global geomagnetic indices. With the aim of quantifying the wavelengths of the source fields, especially for the summer time, frequency domain principal component and maximum likelihood wavenumber analyses of the electromagnetic data were performed. Principal component analysis separates the spatial structure of the fields into a simple set of mutually orthogonal modes of variation, and wavenumber information can be derived from the first principal component under basic assumptions about the noise distribution. Maximum likelihood wavenumber spectra determine the dominant wavelength present in the source. These analyses are used to estimate the influence of short wavelength auroral sources on the magnetotelluric response and devise data analysis strategies to minimize their influence.

6 EM4-6 IONOSPHERIC DRIVERS OF LARGE INDUCTION EFFECTS ON TECHNOLOGICAL SYSTEMS ON THE GROUND Pulkkinen Antti (Finnish Meteorological Institute, Thomson Alan (British Geological Survey) Clarke Ellen (British Geological Survey) McKay Allan (University of Edinburgh) Viljanen Ari (Finnish Meteorological Institute) Geomagnetically induced currents (GIC) flowing in technological systems on the ground are a manifestation of space weather. Due to the proximity of very dynamic ionospheric current systems, GIC are of special interest at high latitudes where they are known to cause harm e.g. for normal operation of power transmission systems and buried pipelines. Despite numerous studies on GIC, there still exists no well established picture of the detailed structure of the ionospheric currents driving large GIC. Although some rough estimations of large-scale electrojet intensities and structures during GIC events have been carried out, no rigorous study of the ionospheric source currents has been made so far. In this study, two intense geomagnetic storm events on April 6-7, 2000 and November 24, 2001 are investigated. During these events, large GIC were measured in technological systems both in Finland and in Great Britain, providing a basis for a detailed GIC study over quite a large spatial scale. By using these GIC data and geomagnetic data from north European magnetometer networks, the ionospheric drivers of the large GIC during the event were identified and analyzed. Although most of the very largest GIC during the storms were clearly related to substorm intensifications, there were no common characteristics in substorm behaviour that could be associated with all large GIC. For example, both very localized ionospheric currents structures and relatively large scale, propagating structures were observed during large GIC. Only cases of coherently behaving, large scale GIC were associated with the sudden commencement at the beginning of the event and with geomagnetic pulsations at the end of the event. The typical duration of large GIC was quite short, varying between 2-10 minutes.

7 EM4-7 SEPARATION OF THE GEOMAGNETIC DISTURBANCE FIELD ON THE GROUND INTO EXTERNAL AND INTERNAL PARTS USING THE ELEMENTARY CURRENT SYSTEM METHOD Pulkkinen Antti (Finnish Meteorological Institute, Olaf Amm (Finnish Meteorological Institute) Ari Viljanen (Finnish Meteorological Institute) BEAR Working Group (Project leader: Toivo Korja, Geological Survey of Finland ) Traditionally the separation of the ground geomagnetic field variations into external and internal parts is carried out by applying methods familiar from the potential theory. However, these methods require a separate field interpolation and extrapolation, can be computationally slow, require a minimum wavelength to be specified to which the spatial resolution is limited globally. A novel method that utilizes elementary current systems can overcome these shortcomings. The basis of the method is the fact that inside a domain free of current flow, the magnetic field can be continued to any selected plane in terms of equivalent currents. Two layers of equivalent currents, each composed of superposition of spherical elementary systems, are placed to reproduce the ground magnetic field effect: One above the surface of the Earth representing the magnetic field of ionospheric origin, and one below it representing the magnetic field caused by induced currents in the Earth. The method can be applied for single time steps and the solution of the associated underdetermined linear system is found to be fast and reliable when using singular value decomposition. The applicability of the method is evaluated using synthetic magnetic data computed from different ionospheric current models and associated image currents placed below the surface of the Earth. Following these tests, the method is applied to the measurements of Baltic Electromagnetic Array Research (BEAR) (June - July 1998). External and internal components of the geomagnetic variations were computed for the entire measurement period. Also the adequacy of the sparser IMAGE magnetometer network for the 2D field separation was tested.

8 EM4-8 On the Magnetotelluric Source Field Effects in the Indian Equatorial Electrojet Region - an Experimental Study. Madhusudhan Rao, Sarma, S.V.S*., Nandini Nagarajan, Harinarayana T, Manoj,C., Naganjaneyulu, K., and Kumaraswamy, V.T.C. (National Geophysical Research Institute, Hyderabad-7, India) * Corresponding author: sreepati@hotmail.com Abstract To study the possible source field effects due to Equatorial Electrojet on MT measurements Magnetotelluric field investigation was carried out along a NS traverse across the Indian Equatorial electrojet zone. Wide band MT data (10 3 Hz 10-3 Hz) covering both day and night times have been acquired corresponding to the presence or otherwise of equatorial electrojet. The results indicate that the apparent resistivity for longer periods (T>100 sec) shows a decrease during daytime as compared to that during night time at a site very close to the centre of the electrojet in contrast to those sites well away from the centre of the jet which do not show any difference in the MT resistivity between day and night. The results have been discussed in the light of other experimental studies and theoretical models.

9 EM4-9 DETERMINATION OF THE REGIONAL 1D TRANSFER FUNCTION OF FENNOSCANDIA FROM THE RATIO OF INTERNAL AND EXTERNAL PARTS OF THE GEOMAGNETIC FIELD BASED ON BEAR DATA Maxim Smirnov, Laust B. Pedersen, Martin Engels (Uppsala University, Sweden, and the BEAR Working Group One of the main goals of the BEAR (Baltic Electromagnetic Array Research) project is to investigate the deep geoelectrical structure of the Fennoscandian (Baltic) Shield by means of simultaneous magnetotelluric observations. Several teams applying different robust codes processed the magnetotelluric data. It was shown that robust procedures supplemented by coherence sorting are capable to reduce the influence of the source distortions significantly. But the remaining part, which might be considered as a bias of the estimates, is still unknown. In order to estimate transfer functions for long periods it is worth while to consider other techniques, which explicitly make use of the source field inhomogeneities. Here we apply a method based on the definition of the inductive scale length C from the ratio of internal and external parts of any magnetic field component. The method requires a separation of the total field into internal and external parts, which was done recently for the BEAR data by Pulkkinen (Pulkkinen et al., 2002) by deriving equivalent current systems. In 1D case the C- function is expressed as follows: 1 1 Q C(ω, k) = with wavenumber k, angular frequency ω ι( ω, k) and potential ratioq( ω, k) = k 1+ Q ε( ω, k) between internal (ι ) and external (ε ) parts of any magnetic field component in frequencywavenumber domain. In order to implement the method first the spectral transformation from time to frequency domain of the internal (ι ) and external ( ε ) parts of a selected magnetic component is carried out for the whole area of investigation. Then a 2D FFT is applied to transform data from space to wavenumber domain. First we test the method on synthetic data calculated by a 3D modelling technique involving a non-plane wave dynamic source structure. Then the results of BEAR data field separation are used to estimate the C-function in frequency-wavenumber domain. Pulkkinen A., Olaf A., Viljanen A., the BEAR WG, Separation of the geomagnetic disturbance field on the ground into external and internal parts using the elementary current system method. This Workshop.

10 EM4-10 SOURCE EFFECTS IN THE BEAR ARRAY TRANSFER FUNCTIONS: DETERMINATION AND ELIMINATION Elena Sokolova, Ivan Varentsov, and the BEAR Working Group (GEMRI RAS, Troitsk, Russia; A set of impedance Z, tipper Wz and inter-site magnetic transfer functions was constructed on a base of the BEAR simultaneous array EM data in Fennoscandia with a robust estimator consequently used in the single site, remote reference and multi-rr modes (Varentsov et al., 2002). A special attention in this high-latitude area was paid to the non-stationary disturbing effects of sub-polar field excitation. The monitoring of temporal changes in the partial TF estimates in relation with source activity/inhomogeneity has shown strong TF distortions caused by intensive electrojet events normally characterised by a significant drop in the quality of EM field linear relations (Sokolova et al., 2002). Thus, the coherency-based data rejecting and weighting routines applied at the entrance of the robust procedure in the averaging of partial TF estimates sufficiently improve the elimination of source distortions. It holds true in Z and Wz data till 3-hour period in almost all of the BEAR sites, and till Sq-periods in impedance at a part of the array. Additional tests based on different selection of processed data sets (e.g. with only day or night or specific-level geomagnetic activity events) prove the conclusion. The question is open, whether the estimated TF data have some stationary bias (in the BEAR month time-scale) caused by the effective source model ( Vanyan et al., 2002)? Such a bias is not seen in the spatial structure of Z and Wz responses below 3-hour period as well as in the spatially averaged longer period impedance data compared with European GDS references. However, it is evident in strong northern trends in the amplitude components of the horizontal magnetic tensor M. We applied new approaches in the elimination of source distortions in M and other TF data, based on weighting strategies accounting magnetic field spatial gradients and magnetic field separation into internal and external parts (Pulkkinen et al., 2002). We also directly involve the internal field time series into the processing. This work was supported through the contract INTAS Varentsov Iv.M., Sokolova E.Yu. et al., Method to construct a system of transfer operators of natural EM field for the high-latitude simultaneous sounding array (the BEAR experiment). Physics Solid Earth (in print). Sokolova E.Yu., Varentsov Iv.M., the BEAR WG, Diagnosing and eliminating distortions of transfer operators of natural EM field caused by non-uniform excitation structure (the BEAR experiment). Physics Solid Earth (in print). Vanyan L.L., Palshin N.A., Levitin A.E., the BEAR WG, This Workshop. Pulkkinen A., Olaf A., Viljanen A., the BEAR WG, This Workshop.

11 EM4-11 TWO TYPES OF IONOSPHERIC SOURCES AT HIGH LATITUDES Leonid Vanyan, Nick Palshin (Shirshov Institute of Oceanology, Moscow, Anatoly Levitin (Institute of Terrestrial Magnetism and Radiowave Propagation, Troitsk) and the BEAR Working Group The analysis of the interplanetary magnetic field carried out with the imployment of the model of high-latitude ionosphere current systems associated with parameters of the solar wind showed that during the international BEAR experiment (June-July, 1998) the main ionospheric current source was often located in the polar cap far from the measuring sites. Thus, the theoretical estimations predicting strong distortions of the MT impedances caused by powerful electrojets in the auroral oval may contradict to data collected during continuous time interval. This conclusion is backed up by statistical and spectral analysis of the magnetic field variations measured along the meridional chain of the sites during the BEAR experiment. Therefore, instead of the models characterising the momentary state of auroral source, it is reasonable to use "effective" model characterising time-averaged state of the current systems. A flexible numerical technique for the modelling of the low-frequency EM field induced by arbitrary shaped ionospheric current systems was developed. Test numerical modelling for idealalized models of the "effective" auroral electrojet and polar cap current system showed that EM field for both models is much closer to the plain wave model response in contrast to the momentary dipole-type source. It explains qualitativeely the applicability of the plain wave source model for MT sounding in high-latitudesup to the period about s. The work presented was made possible through the following contracts: INTAS , RFBR