Search

Expressions are derived for the magnetic field directions produced by randomly varying Gauss coefficients. This approach was pioneered by Constable and Parker (1988), and followed by Kono and Tanaka (1995) and Kono and Hiroi (1996). The present treatment is a sequel to the latter two. The basic assumption is that the magnitude of the axial dipole component is much larger than all the other terms: equatorial dipoles or nondipole components. It is also assumed that, in a sufficiently long time interval, Gauss coefficients, g l m and h l m , behave as independent random variables that have the first and second moments defined. General expressions are obtained for the field directions (inclination I and declination D). Paleosecular variation (PSV) of field directions is formulated as latitude dependences of statistical parameters (the means and variances) of inclination and declination. It is possible to derive a family of PSV models by specifying the statistical nature of individual Gauss coefficients, i.e., assigning means and variances to them. In this paper, we describe a homogeneous background model (HBM), which consists of a large axial dipole component (g 1 0 ) and other harmonics that have zero mean and variances decreasing exponentially with the degree of the harmonic. This model is too homogeneous to reproduce observed latitudinal variation of secular variation, but is quite useful to identify the harmonics that have effect that is larger than produced by the HBM. Comparison with paleomagnetic data obtained from volcanic rocks of 0-5 Ma ages (Quidelleur et al., 1994) shows that it is necessary that the mean of g 2 0 must be about 5% of the mean of g 1 0 , and that g 2 1 (or h 2 1 ) should have about three to four times as large an effect as suggested by the HBM. The former conclusion is consistent with the result of Constable and Parker (1988), while the latter conclusion substantiates the similar conclusion obtained by Kono and Tanaka (1995) by the analysis of the scatter of the virtual geomagnetic poles (VGPs).

Five stalagmitic speleothems from two caves in Sichuan province China, have yielded data on the palaeosecular variation (PSV) of the past 8.9 ka. The uranium-thorium disequilibrium method has been utilised in order to estimate the age of each record. Due to the high detrital load contributing some allogenic thorium, age corrections had to be applied to most sub-samples. The directional agreement between coeval central and lateral sub-samples where growth layers are horizontal and steeply dipping respectively suggests that inclination errors are absent in these speleothems. Correlation between features of declination and inclination in the individual records, even after time-averaging, are reasonable indicating a local consistency in recorded PSV and adding to the reliability of the palaeomagnetic and age data. Comparisons between our speleothem directions and the directions of the modern field drifted westward past our site at 0.13° yr -1 suggests westward drifting of non-dipole sources during the past 2.8 ka. Using directional data from two speleothems with the longest growth periods a composite curve covering the past 8.9 ka has been constructed. The resulting VGP path again indicates predominantly westward drift. The declination and inclination features of these speleothems also compare reasonably with features of PSV in contemporaneous records from China and Japan.

A recent (2000 BP) lava flow from central Mexico has been sampled along a vertical profile with 55 cores covering a total flow thickness of 6.6 m. A wide range of physical and magnetic parameters have been studied to characterise the samples: Curie temperature and saturation magnetisation as intrinsic properties; density, magnetic susceptibility and remanence intensity as bulk properties; Konigsberger Q-factor and hysteresis parameters as coercivity parameters. All parameters vary smoothly over the profile, most probably due to grain size variation of the magnetic minerals present in the samples. Optical observations indicate that the main opaque minerals are deuterically oxidised titanomagnetites (C3-C5) and ilmenites (R2-R3), which increase in size away from the edges of the flow. Paleointensity (PI) was determined using the double heating Thellier-Thellier method with pTRM checks. According to reliability parameters (f-, g-, and q-factor) the obtained PIs are of reasonable to good quality. PI shows marked variation with vertical position in the flow, across a range of about 25 to 125 μT, with most samples having a PI between 50 and 100 μT. The flow-mean PI of 72 μT is higher than the present day field, consistent with global data for this time-period. No obvious correlation could be found between PI and any other measured parameter. The variation of PI with vertical position in the flow may show some systematic behaviour. It is important, therefore, to sample a flow both horizontally and vertically in order to obtain a reliable paleointensity.

The transformation of isotropically distributed geomagnetic poles to local site directions introduces slight apparent inclination shallowing if the directional average is compared to the geocentric axial dipole direction. This effect depends on site latitude and pole dispersions. For typical dispersions of poles, 063 between 10° and 20°, the average inclination will appear too shallow by about 1° to 2° for site latitudes of 10° to 60° North or South. On the other hand, when some of the scatter might be due to isotropic scatter in directions, averaging all the data in polar space, will introduce a steepening effect. For most paleomagnetic studies this intrinsic inclination bias will be small in comparison with other uncertainties, typically of the order of 5° to 10°. However, for integrated high resolution studies of specific aspects of the geomagnetic field this methodological ambiguity might bias the results, and we caution against averaging paleomagnetic data only in directional space.

One-dimensional transient response of the inner magnetosphere at the magnetic equator is investigated using two models of altitude distribution of the Alfven speed VA. The present paper concentrates on the transfer function of the system under consideration and its poles, which govern the transient response of the system. The poles, which are mathematical counterparts of the cavity resonances, appear owing to the inhomogeneity of VA and their locations depend on the altitude distribution of VA as well as the position of external source (or outer boundary of the inner magnetosphere). Even if there exists no strong Alfven velocity gradient at the outer boundary, an observable cavity-mode oscillation in the Pi2 range can be excited because of the existence of a strong gradient of the plasmapause within the inner magnetosphere. However, the existence of a strong gradient at the outer boundary brings about a long-lived nature of the cavity-mode oscillation as well as calls some new poles into existence. While the surface of the solid earth forms the inner boundary at which the almost perfect reflection of wave takes place, the ionosphere is of secondary importance as a reflector of wave. The existence of the solid earth plays an essential role in the observability of the compressional oscillation arising from the cavity resonance all over the inner magnetosphere. The real part of each pole has a negative value, meaning that the cavity-mode oscillation decays with a damping factor of absolute value of the real part of the pole. Such a damping is primarily due to the leakage of energy through the outer boundary of the inner magnetosphere.

In order to understand the characteristics of lower-latitude Pi 2 geomagnetic pulsations along a magnetic meridian, we have analyzed 1 year's data from the Adelaide (geomagnetic latitude Φ = -46.72°), Birdsville (-37.08°), Weipa (-23.06°), Guam (9.02°), Chichijima (20.65°), Kagoshima (25.23°), and Moshiri (37.76°) stations along the 210° magnetic meridian. The characteristics ofthe lower-latitude (|Φ| < 50°) Pi 2 pulsations are summarized as follows: (1) 80% ofH- and D-component Pi 2 events show very similar wave forms and the same frequency component; (2) not only H-component Pi 2 events but also D-component Pi 2 events are observed within ±5 s; (3) H-component Pi 2 amplitudes are of almost the same order, but D-component Pi 2 pulsations have a minimum near the magnetic equator and increase exponentially from lower to higher latitudes; (4) the observed amplitudes of both H- and D-component Pi 2 pulsations are larger in the summer hemisphere than in the winter hemisphere. Results 1 and 2 show that lower-latitude Pi 2 pulsations are not propagating-mode but global-mode oscillations in the inner magnetosphere. Result 4 cannot be explained by the ordinary screening effect in the ionosphere but gives a clue to the unresolved generation mechanism of lower-latitude Pi 2 pulsations.

Two basic but novel equations directly describing the generation of shear Alfven and compressional waves in the inner magnetosphere filled with a cold magnetized plasma are derived. The shear Alfveen wave is characterized by the field-aligned current and the compressional wave by the compressional component of the magnetic field. Such a generation arises from the effects of inhomogeneous Alfven speed and curvilinear field line. Around the magnetic equator, if the Alfven speed is inversely proportional to a power of the geocentric distance, these effects have magnitudes of the same order and their signs are identical. Considered in the present study is a situation that the earthward propagating compressional wave is launched from a large scale oscillating current wedge centered at midnight and symmetric about the magnetic equator. Then, it is found that the field-aligned current excited around the equator by the compressional wave has opposite senses in direction in the northern and southern hemispheres, in the pre- and post-midnight sectors as well as just inside the plasmapause and in its surrounding regions. As a result of the excitation of shear Alfven wave, two types of oscillations appear on a field line: One is a forced oscillation and the other is an eigenoscillation. Although a modulation of the compressional wave may be caused locally (or microscopically) around the equator by the eigenoscillation of field line, the modulation can be globally (or macroscopically) neglected. So far as the propagation along the source longitude (source-earth line) around the equator is concerned, the coupling between compressional and shear Alfven waves can be almost neglected and so one-dimensional response of the inner magnetosphere around the equator plays a significant role in the compressional oscillation.

The method of deconvolution of post-depositional DRM (detrital remanent magnetization) is applied to the paleomagnetic directions of continuous sedimentary sections from two sites in the Boso Peninsula, Japan. These sections have different accumulation rates and cover a time span of about 1000 years that includes the Matuyama-Brunhes reversal. Since the remanences are thought to be of the post-depositional DRM origin, a new deconvolution method with an exponential fixing function is applied. The deconvolution restores relative intensities of three components in the original fields from the remanence directions. As a result, the half-fixing depth is estimated to be 21 cm and the three-dimensional plot of the relative intensity is given. In the diagram of X and Z components, the polarity change in the geocentric axial dipole field with some additional field is recognized during the main reversal, while there seems to be no correlation between X and Y. After the deconvolution, the VGP path during the main reversal does not show longitudinal confinement, but does show longitudinal drift near the equator.

In the spring of 1994, a survey was carried out to redetermine the position of the North Magnetic Pole (NMP). Observations of declination, inclination, and total intensity were made at eight sites around the expected location of the pole. All observations were corrected for transient variations using data from a variometer set up in the survey area, and data from magnetic observatories at Mould Bay and Resolute Bay. The NMP position was determined by performing a spherical cap harmonic analysis of the corrected data. The computed position of 78.3°N, 104.0°W for 1994.0 was 150 km northwest of the position previously determined in 1984, and showed that the NMP was moving more rapidly than anticipated. The motion of the NMP has changed from a uniform drift of about 9 km/yr, prior to mid-1971, to a uniform acceleration of approximately 0.34 km/yr2, after mid-1971. This change may be related to the widely-reported geomagnetic jerk that occurred in 1969/70.

The magnetic and gravity anomaly maps covering the Izu-Ogasawara arc are compiled using the MGD77 data measured by the Hydrographic Department of Japan (JHD). A magnetic basement map is also obtained by applying the three-dimensional inversion method. Characteristic features of the crust of the Izu-Ogasawara arc are made apparent from the view point of the magnetic, gravity and topographic features. Especially, the distribution of the magnetic bodies with low ρ/J ratio, which belong to the granitic rock of magnetite series, becomes apparent in the forearc area of the northern part of the Izu-Ogasawara arc. The magnetic anomalies of the Nisi-Sitito ridge are well explained by the normal magnetization of topographic highs. This feature suggests the N-S arrangement of the arc when it was formed. The bulk magnetization per unit area (km2) along the Quaternary volcanic front of the southern arc is about 2.4 times as large as that of the northern arc. These features can be explained by the narrow and stable location of the volcanic front in the southern arc, which is related with the steep gradient of the subducting slab. The magnetic edge effect of the Moho boundary can be recognized in the western margin of the northern arc. However, this feature becomes vague in the southern arc. A tectonic model to explain the origin of the different characteristics between the northern and southern arcs is presented by assuming the long and stable state of the steep dip angle of the subducting slab in the southern arc under the N-S arrangement of the arc. The plausible origin of NE- and NNE-trending en-echelon ridge and trough morphology is also given in consistent with the tectonic framework mentioned above.

At low latitudes, the phase of H component of the ground magnetic perturbation can be expected to be independent of the latitude for low-frequency geomagnetic pulsations. Because of such a spatial phase structure of the H component, it is found that the D component has a 180° out-of-phase relation between the northern and southern hemispheres and its amplitude increases monotonously from the magnetic equator to higher latitudes. It is also shown that the sense of rotation of the magnetic perturbation vector in the H-D plane is clockwise and counterclockwise for the eastward and westward propagations, respectively, in the northern hemisphere and vice versa in the southern hemisphere. On the other hand, since the spatial phase structure of the H component far away from the magnetic equator becomes more complicated as the latitude increases, the correlation between signal propagation sense and horizontal polarization may deteriorate at high latitues. For the amplitudes of H and D components due to the compressional component of the magnetic field incident upon the ionosphere their upper bounds are also given.

Measurements of F region vertical drift, Vz made at Kodaikanal (dip 4°N) on 133 days during March 1991-February 1992 under non-equatorial spread-F(ESF) conditions, are analysed for the characteristics of short-period (5-33 min) variations in Vz in the evening hours (1630-0130 IST). The spectral content of the fluctuations in Vz is found to depend on local time. The amplitude of the fluctuations, quantified in terms of the variance (σ2), increases during the period of postsunset enhancement of upward vertical drift and decreases thereafter reaching a minimum during 2200-0000 IST. In consonance with this, while functuations in Vz in the entire band 5-33 min manifest in the postsunset hours, the longer-period (13-33 min) components dominate thereafter (2200-0000 IST). The amplitude of the fluctuations and its day-to-day variability are higher in local summer than in local winter. On a day-to-day basis, the amplitude of Vz functuations increases with peak height in the interval 1756-2004 IST corresponding to the epoch of the evening enhancement in Vz. The results support the view that the short-period variations in F region vertical drift near the dip equator during the evening hours could be due to electric fields associated with the commonly occurring atmospheric gravity waves.

About 100 year variations are found in the field direction of the Matuyama-Brunhes transition. The analysed data are composed of the palaeomagnetic directions from the marine sediment in the Boso Peninsula, whichpreciselyrecordedthe Matuyama-Brunhes transition with a very high accumulation rate (3.7 mm/yr). The section is tentatively divided into two subsections from the directional change, where the main reversal and the rebound occurred in the early stage (subsection I) and subsequently the field directions show moderate changes in the late stage (subsection II). Applying the Bayesian smoothing method to the directions spanning about 4100 years, the inclination, declination and angular changes in the transition field imply short-period components of the order of 102 years. After removing long-term trends, the maximum entropy method was employed for the detection of short-period variations. The results of the spectral analyses suggest the 60-80 year and the 120-150 year periodicity in the subsections of I and II, respectively. The main reversal and the subsequent rebound seem to have been accomplished for short intervals of about 200 and 300 years, respectively. These rapid variations of the order of 102 years are suggested to be a substantial feature of the transition field.

Two lava sections in southwestern Iceland were revisited ; sectoin IG of Kristjansson et al. (1988) of Brunhes chron age and section K of Wilson et al. (1972) of early Matuyama chron age. Paleodirections obtained from 22 lavas at the former section do not seem to have averaged out the secular variation completely because of their small angular standard deviation, and this suggests rapid accumulation of the lavas. The latter section consists of 29 lavas and contains the Gauss to Matuyama transition near its base. The time interval during which the lavas at this section were erupted also seems to be short, judging from the small directional dispersion. Thelliers' paleointensity method was also applied to the samples from this section. A characteristic low paleointensity was obtained from the transitional lava flow. However, low paleointensities were also obtained from the lavas which erupted some time after the transition, and which record directions close to those of an axial geomagnetic dipole.

The magnetic anomaly map of Japan and adjoining areas was produced by compiling many data sets collected by air-borne and ship-borne surveys. In order to tie between adjoining areas, IGRF removal, upward continuation, and a linear shift were applied to each data set. Linear features are detected from the map as a special reference to suggest possible crustal structures. Two significant features in the linear map are found. The one is the existence of lineations crossing the trench. The lineations are caused by oceanic magnetic layers that lie beneath the overriding plate. The distance from trench to the end of lineation is variant. The variation of distance depends on the amplitude of anomaly, the structure of the topmost part of lithosphere, the dip angle of subducting plate, and the Curie point depth. The second is the occurrence of two types of lineations found in the back-arc spreading regions described as, (1) lineation& running parallel with the arc trend on a rifting or spreading stage, and (2) lineations across the arc trend in the young back-arc rift region. They are related to igneous rocks produced by volcanic activities of back-arc rifting and spreading. Consequently, the two features express igneous rocks either (1) parallel with, or (2) across the arc trend. They are probably associated with two different spreading stages of the back-arc basin.

Magnetic measurements on 265 samples from an 8 m long deep-sea sediment gravity core, 450 km SW of Dakar, West Africa, are presented. The core is normally magnetized, thus being of Brunhes age (younger than 730 ka). Cyclic variations in the NRM intensity, the susceptibility, and especially 10 major and 20 minor cycles in the Q-ratio (Koenigsberger ratio) are discussed in relation to climatic variations. The possible effects of secondary magnetizations are not known, but because of the regular pattern in the Q-ratio, it is suggested that the periodicities in this ratio correspond to Milankovitch cycles with periods of 21 ka and 41 ka years, and therefore that the Q-ratio may be used as an easily-obtained time scale, and hence as a stratigraphical tool, in these kinds of deep-sea sediments.

Here we present a detailed record of the Matuyama/Brunhes (M/B) polarity transition recorded in a loess section at Xifeng (36°N, 108°E), in the central part of the Chinese loess plateau. The transition zone, which occurs in loess unit L8, has a thickness of 3.8 m and was sampled in full. At this point in the section the loess accumulation rate has been estimated at 10 cm/ka, and with the size of sub-sample used provides a potential time resolution of 240 years. Magnetite has been identified as the main magnetic mineral with a small contribution from maghaemite or cation deficient magnetite. The remanence of all samples was investigated using thermal demagnetisation from room temperature up to 650°C. The results indicate a duration of 32, 000 years for the decay and recovery of the main dipole intensity, and 20, 000 years for the period of major directional change. During the transition the field intensity falls to an average of 30% of its value before and after the transition, although there are strong variations about this weak value which accompany a sequence of 4 reversed/normal/reversed fluctuations of the local field vector. These fluctuations immediately preceed the final reversed-normal transition, after which the field intensity recovers to its pre-transition value. Calculated VGP paths of most stages are continuous and confined to preferred longitude bands which include America, the Atlantic Ocean or Africa. Field behaviour during the transition period suggests that non-dipolar terms are dominant while the dipole field is weak, with the dipole decay and recovery associated with the increased influence of low order zonal terms of the geomagnetic field.

A record of the secular variation of the geomagnetic field for the last 11000 years has been obtained from the remanent magnetization of seven wide-diameter (20 cm) cores of marine and lacustrine sediment from central to south-west Japan. Magnetization directions of five cores, possessing high-amplitude variations, exhibit good correlation. Two cores show low- amplitude variations, but they can be correlated well with other records after deconvolution to remove filtering effects of the post-depositional magnetization process. A composite secular variation curve was constructed by stacking the field direction records. Time constraints were obtained from radiocarbon ages of shell and wood fractions or tephrochronology. The secular variation curve agrees well with the archeomagnetic record after 1400 yrBP, with a slight difference in the time range 1400-2000 yrBP. The secular variation for the last 11000 years shows an elongated distribution of VGP's. The azimuth of the elongation, about 40°E, is consistent with that of VGP's from Japanese volcanic and sedimentary rocks during Brunhes epoch. The angular standard deviation is 14.5° (upper limit = 15.4°, lower limit = 13.7°). This estimate is slightly larger than that observed globally during Brunhes epoch and the position of the average VGP deviates 8.5° from the geographic north pole. These VGP analyses suggest a stationary nondipole source. The secular variation records from five northern hemisphere sites between 135°E and 95°W in longitude possess a single prominent feature: an extreme easterly swing in declination. A plot of age versus longitude of the swing at each site shows a clear westward drift, at a rate of about 0.13° /yr. The field vector around the swing represents clockwise looping at all the sites. This suggests that the swing is caused by a large non-dipole source which drifts at least from 135°E to 95°W in longitude.

A detailed record of the Matuyama-Jaramillo (reversed-normal) transition has been obtained from a loess section near the city of Lanzhou, in the Gansu province of China. At the position in the section where the reversal occurs, the accumulation rate has been estimated at 28 cm per thousand years. A continuous section of 6 m was collected for investigation using both thermal and alternating field demagnetisation. Results indicate that the reversal in direction had a duration of approximately 5000 years while the field strength was much reduced for a period of at least 9000 years. Secular variation in both direction and strength of the field is important throughout the sequence, being particularly strong while the field strength was low. As a consequence, the stability of the field leading up to the reversal in direction is poor, with a number of apparent short duration “reversals” due to a relative enhancement of the non-dipole field. Fourier analysis of the record shows an important periodicity close to 750 years in the field strength and periodicities in the horizontal and vertical field components which can be correlated with more recent secular variation. During the reversal, the VGP path is restricted in longitude, initially following the coast of Africa until it reaches the equator. It then swings west to the Caribbean before moving up through North America to complete the reversal. The second part of this path is within one of the preferred longitude bands recently highlighted by LAJ et al. (1991).

The Olduvai termination record recovered from remanent magnetization of the Red Loam has demonstrated the possibility of recording and preserving the geomagnetic polarity transition in eolian deposits and displayed some characteristics different from those of the Olduvai termination records available. The paleofield direction undergoes three phases of change during the polarity transition. In the first phase, only the declination changes, resulting in the VGP's moving along a meridional band, bounded by 30°E-150°W and 50°E-130°W, in high northern latitudes. The following phase is characterized by swift change in both declination and inclination. The VGPs move to the far hemisphere, migrate southerly, and come back to the near hemisphere at moderate southern latitude.

The core-mantle boundary below the Pacific is one of the most interesting regions for the study of the Earth's core because of the phenomena which seem to be characteristic to this region. The geomagnetic phenomeno is one of them. Two features are widely believed to exist in the geomagnetic field in this region. One is absence of the non-dipole field, and the other is its small secular variation, which might be related to special structures of the core-mantle boundary region below the Pacific. This paper re-examines historical data of declination, and suggests that an intense focus of the non-dipole field existed in the North Pacific in the 17th century, and concludes that the absence of the non-dipole field as seen at present has not been a permanent feature of the magnetic field in the Pacific, implying that the Pacific region is no special region in this sense.

Based on that the order of magnitude of the anisotropy eigenvalue ratios of thermal remanent magnetization (TRM) is expected to be the square of the principal susceptibility ratios in anisotropic multidomain magnetic assemblages, a concrete correction formula is derived in this paper, which can be used to correct the TRM deviation induced by the anisotropy of magnetic susceptibility (AMS) away from the ambient field, and successfully applied to the practical paleomagnetic researches in Yueshan diorite (China) and Flamanville granite (France). An interesting discovery is that TRM direction is deflected away from that of the paleofield in each site. In contrast, the overall mean TRM direction is quite a good estimation of the paleofield direction, because the major axis of susceptibility ellipsoid are in the radial distribution.

Observatory data for temporal variations of the geomagnetic field in central Mexico are analyzed for the period since 1923 AD. Annual mean values of the intensity have decreased from 46, 076 nT to 42, 832 nT at a linear-regression rate of about 48 nT per year. The geomagnetic field direction has been varying from easterly to westerly declinations and towards steeper inclinations. During an early interval from 1923 to 1938, changes were mainly in inclination, this was followed by a brief period with low directional change from about 1938 to 1942, and since 1942, changes have been mainly in declination. Comparison of the observatory data with IGRF models for the period 1945 to 1990 shows that IGRF provides an excellent estimation of the geomagnetic parameters in this low latitude region (19.7°N). In particular this comparison study supports the existence in central Mexico of a relatively small secular variation as implied by the IGRF and observatory data.

A review of indications of the return currents of the equatorial electrojet (EEJ) and associated phenomena has been presented. The need for good fitting of the observed dip equatorial magnetic variation profiles had compelled the inclusion of westward currents on the flanks of the dip equator in practically all the cases. The three sources of the return currents: the curvature of the field lines through the ionospheric dynamo region, the divergence of east-west currents close to the magnetic dip equator, and the local neutral winds varying with height combine constructively, each peaking around 5° dip latitude, to provide for the return currents of practically all the eastward electrojet current. The return currents flow on the flanks of the dip equator from about 3° and terminate at a dip distance L1 in the interval 7°≤L1≤17° with a mean of L1=10±3°, sometimes extending further but not beyond the Sq focus. At a location about 5.2±0.8° dip latitude, they reach a peak of about -27±3 per cent of the peak of the eastward equatorial electrojet. The negative correlation between the width and the intensity of the equatorial electrojet has been confirmed with data derived from physical model, indicating its origin in the westward return currents. The ionospheric current system so far detected by rockets is essentially in two layers. It is suggested that the intense lower layer including the return currents around 5° dip latitude, be associated with the equatorial electrojet; and that the weak upper layer that maintains more steady altitude extent everywhere be associated with the worldwide part of the Sq currents.

A paleomagnetic study was carried out on three gabbroic intrusives in the Avalonian terrane of southeastern Cape Breton Island. Thirty five samples (96 specimens) from seven sites were collected for this purpose. Thermal and/or alternating field treatments were performed. Four components of magnetization were detected. One component coincides with the present Earth's field direction. A second component (A) oriented south-southeast with shallow to intermediate inclination (165, +41) is interpreted as representing a Siluro-Devonian overprint. A third component (B) yields a mean direction to the southeast and an intermediate inclination (147, -54). This component is not observed in North American rock units of Devonian age or younger; it is thus bracketed in the Early Devonian-Late Hadrynian interval. Based on magnetic properties, mineralogical and geological grounds, an Ordovician-Cambrian age of acquisition is attributed to this component of magnetization. A fourth component characterized by a low unblocking temperature and median destructive field has a mean direction to the northeast and an intermediate inclination (54, +47). This uncommon direction is inferred to represent an incompletely isolated magnetization. Compaged to low southern paleolatitudes for the Late Hadrynian period of cratonic North America, the paleolatitudes of the Avalonian terrane are slightly higher but this is not really significant. A sea at the most 1000km wide may have separated North America and Avalon at that time. A comparison of the paleolatitude of Avalon and those of Late Hadrynian Atlantic Ocean borderlands indicates that Avalon could have belonged to North Africa at that time. The Avalonian apparent polar wander path in the Devonian-Late Hadrynian interval still remains uncharted. The collision of Avalon with North America occurred most probably at Middle Devonian time.

Precipitation of low energy protons (-1MeV) at low equatorial altitude has been investigated by the Phoenix-1 experiment on board the S81-1 mission. Results of the analysis of data received by the monitor telescope and their geophysical interpretation are presented. Seven months long observation during geomagnetic average conditions in May through early December of 1982 shows that the peak precipitation of protons occurs along the line of minimum magnetic field strength. The full width at half maximum (FWHM) is nearly 13°. The particle flux shows a strong altitude dependence, and below 220km it shows double peak structures. Spectral index of the precipitation flux has been obtained from the data of previous observations. Further, peak precipitation flux in 1982 is almost three times as large as peak precipitation flux in 1969. Charge exchange mechanism favors a stronger source than AP8MAX model. Particle flux does not depend on geomagnetic longitudes. Finally, investigations into the presence of Z≥2 particles show that they are virtually absent in the energy range considered.

155 samples from 170 collected at 36 archaeological sites dating between the 15th and 1st century B. C. defines the variation in declination (±20°), inclination (±10°) and ancient intensity (±15-20μT) in Georgia. These new observations are consistent with earlier observations and, when combined with them, these errors are reduced to ±5°, ±4°, and ±5-6μT. The virtual dipole path shows an anticlockwise rotation during this period.

The paper describes the results of daytime VHF scintillations at a number of stations within the electrojet region in India. Slow and weak scintillations (about 1dB peak to peak fluctuations) are observed almost on 90% of occasions during the midday hours at the station on the magnetic equator. The average occurrence of Es-q type of scintillation decreases very rapidly with increasing dip latitude. The occurrence of Es-q type of scintillations is precisely coincident with the period when the equatorial electrojet defined by ΔH (Trivandrum minus Alibag) is positive. These observations clarify the apparent discrepancy of earlier observations of daytime VHF radio scintillations at Huancayo and Ootacamund.

Magnetization process of dirt-snow containing fine-grained magnetic minerals was investigated in order to understand the acquisition mechanism of natural remanent magnetization (NRM) by dirt-ice layers in Antarctica. Dirt-snow acquired NRM from the direction of geomagnetic field (GMF) direction at rooms -10°C and -20°C, and inclination becomes shallower than the GMF after 50 days.The following two types of NRM acquisition mechanisms in the GMF are proposed; (1) magnetic grains align along the GMF when they drop onto the surfaces of underlying snow due to vaporization of snow; (2) the magnetic grains are rotated to align in the GMF direction on the snow surfaces by the torque with the GMF and the NRM. The compaction of the natural snow at formation of depth hoar works to flatten the NRM direction.Probably the Antarctic dirt-ice has acquired NRM during the periods of accumulation of snow and magnetic grains and formation of the depth hoar. The NRM of dirt-snow and -ice are usuful for understanding the glacial dynamics and paleomagnetism of the Antarctic ice sheet, although many problems still remain.

Archaeomagnetic measurements on a kiln thought to be operating at the time of the Pompeii disaster, 24th August 79 A. D., indicates a normal field direction with an archaeointensity value of 61 microtesla. The low error and precise date makes this an important reference point to which other less accurately dated results can be compared.

The magnetic field distribution at the earth’s surface present an important temporal variation, named secular variation (SV). The secular variation is one of the important characteristics of the geomagnetism, and the extraordinary rapidity of the phenomenon, at geological scale time, allow to suppose that their origin should be a manifestation of fluid motion in the outer core of the earth (HONKURA and MATSUSHIMA, 1988). Several attempts have been made to estimate near the core, surface motion from magnetic field data (GIRE et al., 1984; LE MOUEL et al., 1985; VOORHIES, 1986), but the sources of this mechanism are still partially unknown. In central Atlantic, the secular variation of vertical Z component (and so inclination I component) of geomagnetic field is particularly important. There are here the most intense cell of SV (dZ(t)=-210 nT/year), like a “magnetic hurricane” (CAIN, 1969), which affect strongly the west African and Brasilian magnetic field. This cell, westward drifting, according to general trend of the drift of the geomagnetic field, carry away the magnetic equator (or dip equator), locus at the earth surface where the Z vertical component is zero. This westward drift is a little understood phenomenon whose nature and extend is debated. In this paper, we study in West Africa the drift speed of the equator since the beginning of century. We study the differences between the observed and IGRFcomputed equator in the recent past. We compare so the magnetic drift in West Africa and Brasil.

Monthly and semimonthly harmonic analyses of H are performed and some evidence of a diurnal variation of the lunar tide strength is found. Contamination of the data by magnetic disturbance can be identified by examining the phase progression in local time of the semimonthly component. This contamination is also found in the work of some other authors. It is suggested that some effects previously described as “lunar” are due to other causes.

Noontime absorption on 1.8, 2.0, 2.2 and 2.5MHz at Trivandrum (magnetic dip -0.6°S) can be expressed in terms of solar X-ray and 10.7cm (F10.7) fluxes and the noontime solar zenith angle by an empirical relation. The components of absorption related to X-ray, F10.7 and the component independent of these two as obtained from the empirical relation have been examined on the basis of theoretical estimates of contributions to absorption from EUV, X-ray and Lyman-α solar fluxes. The solar activity independent component in the empirical relation is found to agree with the contribution from the solar activity independent component of solar Lyman-α radiation. From the theoretical analysis it has been shown that the wavelength range of solar X-ray effect on absorption for which 1-8A integrated X-ray flux will be a good representative index is 1A to 30-35A.

A boring core of 41m in length and 3.2cm in diameter has been obtained from the Takase granite dated to between 51Ma and 41Ma by the K-Ar dating method. We preliminarily measured the remanent magnetization and examined magnetic minerals in the pilot samples of the core. Then, a section of the core from 32m to 41m in depth was cut into the form of cylinders 1.8cm high and 1.8cm in diameter. From the remanent magnetization of about 350 cylindrical samples obtained, an apparent magnetic polarity change was found to take place at short intervals at approximately 35.8m to 37.4m in depth. The interval between the normal polarity and reversed polarity states is about 1.6m in length, however, it is unknown whether it provides a true interval perpendicular to isothermal planes. On the other hand, the variation of remanent intensity during this polarity transition had no connection with a change in inclination of the remanent magnetization, but it was significantly concordant with the variation of intensity of the saturation magnetization.

The statistical analyses of occurrence characteristics of the plasma bubbles and blobs are made by using more than 1000 cases of observation data from the swept frequency impedance probe aboard the Hinotori satellite. The results show that formations of the plasma bubbles have the following characteristics. 1) The spatial distribution of the occurrence of the plasma bubbles is in the equatorial region within ±30° of the magnetic dip latitude centered around the magnetic equator, 2) the occurrence of the plasma bubbles is limited in the nighttime; favorable periods of the plasma bubble formation depend on the type of the plasma bubbles; multiple plasma bubbles (MPB) have tendency to be generated in pre-midnight, while quasi periodic plasma bubbles (QPB) show their maximum occurrence in the post-midnight period. Solitary plasma bubbles (SPB) occur, however, rather independently to the local time in so far as in the night period, 3) the magnetic activity also controls the occurrence of the plasma bubbles; the occurrence of MPB phenomena shifts the peak period to the midnight side with increasing magnetic activity and the occurrence of QPB phenomena which covers the period from midnight to the morning side shows the expansion of their occurrence into late morning period with increasing magnetic activity, 4) correlation of the occurrence of the plasma bubbles to the solar radiation flux represented by F10.7 solar radio flux is evident for MPB and QPB while there is no relation between F10.7 and the occurrence of SPB. Occurrence of the plasma blobs has a complementary nature with that of the plasma bubbles. The occurrence region of the plasma blobs is limited in the edge parts of the plasma bubble occurrence region being limited in the nighttime. The occurrence of the plasma blobs decreases with increasing magnetic activity, while there is a strong anti-correlation of the occurrence of the plasma blobs to the solar radiation (F10.7). The relation of occurrence of the plasma bubbles and blobs to the development of the equatorial anomaly and the asymmetrical distributions of the background electron density suggests the importance of the generalized Rayleigh-Taylor instability including the effects of the electric field and the neutral wind in addition to the gravitational force for the generation of the plasma bubbles and blobs.

A two-dipole model is used to simulate the behaviour of the geomagnetic field during two successive polarity transitions (R→N followed by N→R) reported by VALET and LAJ (1981) from a 6Ma old marine clay sequence in western Crete. The model consists of an axial geocentric dipole and an offset minor dipole located at the northern core-mantle boundary. During the transitions the geocentric dipole decays exponentially through zero and then recovers its opposite polarity state. The axial offset dipole has two components: a time-varying and a standing one. In the middle part of the reversal process the standing component controls the transitional field and is responsible for the steep negative inclinations observed in both transitions.The advantage of the two-dipole model compared to some other transition models is that it is capable of describing not only the inclination and intensity transitions but also the properties of the field during intervals of stable polarity, including the departure of inclinations from the axial geocentric dipole value, the symmetry of the older (R→N) reversal and the asymmetry of the younger (N→R) reversal. In a global perspective the model predicts that the shapes of intensity and inclination curves during the transitions depend on the observer's site latitude and that the onset and termination parts of these curves may differ. Because the model is axial it cannot produce the observed declination transitions. In spite of this the model produces VGP-latitude vs. stratigraphic height curves and field intensity vs. VGP-latitude curves consistent with results derived from Hoffman's flooding model.

Observations of Pc waves in the magnetosphere are reviewed with emphasis on the discussion of satellite-ground correlations. Gross features of wave occurrence in the Pc 3–5 frequency range are shown to be well summarized by taking into account the polarization characteristics of the waves with respect to the ambient magnetic field. In the Pc4–5 frequency range, azimuthally polarized-transverse waves, which occur predominantly on the morning side of the magnetosphere, show a good correlation with ground Pc events. Radially polarized waves dominate in the afternoon and dusk sectors and are less correlated with ground pulsations than azimuthal waves. Compressional waves, such as stormtime Pc5 observed at synchronous altitude in the afternoon and compressional Pc 5 observed with HEOS 1 in the dusk, have not yet been identified on the ground. Only compressional waves in space, which belong to the radial class and show a strong ground correlation, are giant pulsations as observed in the morning sector on the ground. Although the problem remains to be studied further, the difference in degree of ground-satellite correlations of Pc waves seems to reflect the spatial extent of respective wave phenomena. As for most of ground-correlated Pc waves, observations appear to yield evidence that a rotation of the wave ellipse orientation occurs between the magnetosphere and the ground as predicted by theory.

A power-spectrum analysis of the Geomagnetic Inclination data time-series for the last 9000 years as obtained from studies of sediment cores from Lake Kasumigaura, Japan, revealed highly significant periodicities of about 900, 1900 and 3200 years and less significant periodicities of about 375, 450 and 650 years.

The Akiyoshi limestone group is an accreted reef complex originated on oceanic basaltic seamounts. The direction value of the natural remanent magnetization (NRM) of the Akiyoshi limestone is not adequate to deduce where the limestone was deposited, because the NRM seems to be secondary (a remagnetized value). The mean in situ direction, namely the remagnetized direction, of the limestone is 49.90° in declination and 61.1° in inclination (Fisher statistic parameters: k=90.0 and α95=4.1°) for fifteen sites with the precision parameter k>10.0. This mean direction is almost the same as the in situ directions of welded tuffs from four formations distributed around the Akiyoshi Plateau, which have been already reported by previous study. These welded tuffs have fission track ages between 85 and 94Ma. The remagnetized direction is also rather similar to the well-known paleomagnetic direction for Southwest Japan before about 15Ma. These correspon-dences suggest that the Akiyoshi limestone and the welded tuffs were remagnetized by some event after about 90Ma and before 15Ma. Results of thermal demagnetization and observation in thin sections of the limestone suggest that dominant magnetic particle is magnetite, and that the limestone may not have suffered high temperature (>500°C) alteration. This implies the presence of a low temperature chemical event.

Cretaceous and early Tertiary rocks have been collected, by many workers including the present authors, from about two hundred sites in the Japanese Islands and South Korea for paleomagnetic investigations. The area mean direction of magnetization for the Oshima Peninsula and Okushiri Island of Hokkaido was 30.0° in declination and 58.4° in inclination. It differs distinctly from the area mean direction for granitic rocks from the Kitakami and Abukuma mountains. The area mean direction of magnetization obtained from the northern Kitakami belt was extremely westerly (D=-81.1°, I=22.3°) and that from the southern Kitakami belt was northwesterly (D=-44.4°, I=51.7°). The area mean from the Abukuma belt was also northwesterly (D=-31.9°, I=54.6°). The site mean directions of magnetization for granitic rocks in the Uetsu Province were not significantly grouped in the area. In Southwest Honshu, the area mean direction of the normal samples from various kinds of Cretaceous rocks was 50.2° in declination and 60.5° in inclination and that of the reversed ones was -114.80° and -55.8°, respectively. The normal and reversed magnetizations of early Tertiary rocks have shown similar directions to those of the Cretaceous rocks. The area mean of the normal samples was D=37.6°, I=52.3° and that of the reversed ones was D=-131.4°, I=-44.8°, respectively. The site mean directions of magnetization for Cretaceous rocks in South Korea appear to be slightly northeasterly at each sampling site. The remarkable discrepancy between these area mean directions of magnetization is interpreted as being due to regional tilting movements which occurred in the Japanese Islands since the Cretaceous.

A reversal of the vertical F region drifts over the magnetic equator during the nighttime hours from the normal downward direction to upward direction is shown to cause the sudden generation of the Range type spread F configuration on the ionograms after a delay of about one hour. This phenomenon is suggested due to a sudden reversal of the equatorial horizontal electric field associated with the sudden reversal of the Interplanetary Magnetic Field from southward direction.

The motions of the Earth's magnetic dip poles and geomagnetic poles over the past 400 years are investigated. In addition, relations are sought between geomagnetic and virtual geomagnetic poles over historic times.Since 1750, the secular motion of the dip poles along their paths has been counterclockwise. During this century this secular motion has been directed in a northwest direction with an average velocity of -11km per year for the north dip pole and -10km per year for the south dip pole. In 1980, the north dip pole was located at 77.3°N and 101.8°W in the Canadian Archipelago, and the south dip pole at 65.6°S and 139.4°E just off mainland Antarctica.In addition to its secular motion, each dip pole undergoes a diurnal motion caused by ionospheric current systems. This motion, which is clockwise in the north and anticlockwise in the south, may result in a displacement from the unperturbed position of 65km and 30km respectively, on an average disturbed day.The average velocity of the north geomagnetic pole during this century is -1km per year in a northwest direction. It is about a magnitude less for the virtual geomagnetic pole. Both the geomagnetic and virtual geomagnetic pole paths display a clockwise rotation.There is a great deal of uncertainty in our attempt to use virtual geomagnetic poles to extend the geomagnetic pole path back in time to -7500 B. C. Much of this uncertainty is due to the sparse distribution of archeomagnetic measurements at a given epoch.

Longitudinal and Latitudinal characteristics of the O+-H+ ion transition height in the topside ionosphere at the low latitudes during the sunspot minimum period (1975-76) were observed by a retarding potential analyzer aboard TAIYO.The results indicate that:1) A semi-empirical formula of the ion transition height (H in km) is expressed as a function of the solar zenith angle (χ in deg.) and the geomagnetic dip angle (I in deg.); H=591-87.1sin2 I, (χ≥90°), and H=591-87.1sin2 I+(217+473sin2 I) cos0.331χ, (χ≤90°).2) The ion transition heights around the South Atlantic geomagnetic anomaly are rather variable than those at another region in the nighttime. The region with variable ion transition height roughly coincides with that of anomalous intense count observed with a high energy measuring instrument (SXR) aboard TAIYO in the ionosphere.3) Discussions and comparisons with other experiments are given. The remarkable asymmetry of the ion transition heights between the southern and northern hemispheres is explained by the semi-empirical formula to some extent.

Equatorial Pc pulsations, having periods from 30 to 100sec and amplitude up to several nT, were studied by making use of the fluxgate magnetometer data obtained at Huancayo (0.8°N, 335.6°E in geomagnetic coordinates), Peru during the period from December 5, 1985 to January 5, 1986. Diurnal change of polarization parameters (orientation angle of the major axis and ellipticity) in the H-D plane were investigated. The following results were obtained.1) Pulsations are linearly polarized. The ellipticities lie between +0.3 and -0.3.2) orientation angle of the principal axis of variation vectors shows a diurnal variation. It deviates westward from the north for the local time 6-7h. The deviation angle of the major axis from the north occasionally amounts to 60 degrees. A small eastward deviation is also seen in the local time 15-17h. Around 17h a westward turning tendency is noted. During the other time of day, it remains in the north-south direction.It is suggested that the changes in the D component which bring about the deviation of the orientation angle from the north could locally be produced through an east-west non-uniformity of the ionospheric conductivity distribution.

Statistical analyses were made to clarify the latitude dependence of occurrence and other characteristics of Pc5 waves near L=6.6, using magnetic field data obtained with GOES 2 and 3 during 1978-1984. It is found that the occurrence probability of Pc5 waves with azimuthal polarization (A-class) is small at GOES 3 (4.7°N in magnetic latitude), as compared with that at GOES 2 (9.3-11.4°N). Most of A-class waves observed at GOES 2 are simultaneously detected near the subsatellite point of GOES 2. A typical amplitude ratio of ground Pc5 pulsations to A-class Pc5 waves at synchronous orbit is about 5 (50nT/10nT). In contrast to the A-class waves, compressional waves with radial polarization, observed mostly in the afternoon around 15h LT during magnetically disturbed conditions, have a larger occurrence probability at GOES 3 than at GOES 2. Amplitudes of the compressional Pc5 are larger at GOES 3 than at GOES 2, when the wave simultaneously occurs at both the satellites. The difference of Pc5 characteristics at GOES 2 and 3 suggests that the compressional Pc5 wave is confined to a latitudinally narrow region near the magnetic equator. The compressional Pc5 can be further divided into two types. One is of long duration and mostly occurs around noon in the recovery phase of magnetic storm. The other is of short duration and is observed on the afternoon-dusk side. The occurrence of short-duration events is well correlated with the individual substorm onset. A systematic delay of Pc5 occurrence behind the substorm onset is statistically obtained as a function of the local time of the satellite position. This delay suggests that the source region of the short-duration compressional Pc5 drifts westward with a speed of the gradient drift of 50-60keV protons.

The new technique of spherical cap harmonic analysis has enabled a model to be constructed of both the main geomagnetic field and its secular variation, that applies to Canada and its adjacent areas over a time range of 25 years. The technique permits the optimum utilization of data over only a portion of the earth's surface, while ensuring that the curl and divergence of the total magnetic field vector are identically zero. With this technique, proper consideration can be taken of altitude variation, a particularly important factor when dealing with satellite data. The effect of crustal anomalies on the determination of the secular variation (SV) model was eliminated by fitting SV independently of main field, and then integrating the SV model. Main field data at epoch 1985.0 were derived by updating Canadian aeromagnetic and Magsat satellite data by means of that integrated SV model. A spherical cap harmonic residual model was constructed from the updated main field data after subtraction of the 1980 International Geomagnetic Reference Field (IGRF 1980) extrapolated to 1985.0. Merging this residual model for 1985.0 and the integrated SV model provided a model with maximum spatial index of 16 for the main field terms at epoch (temporal degree 0) and of 4 for the integrated secular variation terms (temporal degrees 1 to 3), over a spherical cap of half angle 30°. The vector sum of the residual model and the IGRF 1980 at 1985.0 is called the Canadian Geomagnetic Reference Field 1985, and applies from 1960 to 1985 with extrapolation possible to 1990. This reference field has been used to produce the Magnetic Charts of Canada for epoch 1985.0.

Geomagnetic observations have intensively been carried out since 1976 in the Izu Peninsula, Japan, in an attempt to search for precursory and coseismic changes in the geomagnetic field. In order to enhance the significance of such changes, it is indispensable to clarify noises in the observed geomagnetic field. In this paper, noise problems in the Izu region are investigated, and it is particularly shown that rather systematic changes arise from the ocean tide effect. Since most of the observation sites in the Izu Peninsula are located close to the sea surrounding the peninsula, the total intensities observed at these sites must be affected by a magnetic field induced by movements of conducting sea-water under the presence of the geomagnetic field. A technique for reducing noises due to the oceanic tide is also investigated in this paper.

From the correlations of seven landmark distances, s, of the equatorial electrojet with its peak eastward current intensity J0 and its space gradient dJ0/ds, we find that the entire electrojet current system, with its magnetic field, tends to contract; consequently, the current and magnetic field foci, as well as contours of equal current intensity or of equal magnetic field, move towards the magnetic dip equator as the electrojet intensity increases. Furthermore, for a given increase in intensity, the greater the landmark distance the greater its contraction; and for movement of unit distance, the nearer the position to the dip equator the greater the increase in intensity required. Three causes which possibly contribute to the contraction are briefly discussed.