The 1995 Hyogo-ken Nanbu earthquake (M_JMA7.3) occurred on January 17, 1995. To reveal the detailed stress field after the mainshock, we relocate hypocenters of aftershocks (M ≥ 2.3) and determine their focal mechanisms using seismic records obtained by GROUPS-95, a temporary dense seismic network in and around the aftershock region. Along the mainshock rupture zone, some aftershocks are nearly N-S compression or normal fault type events, which is inconsistent with the regional stress field of approximately horizontal E-W compression. We call these aftershocks atypical, defined as events which have focal mechanisms with P-axis directions more than 45°from the regional stress field. The atypical aftershocks amount to about 17% of the total. No temporal variations in aftershock mechanism are found in the analysis period. Their spatial distribution is compared with the slip and stress distribution of the mainshock, as well as the P- and S-wave velocity structure. Most of them are located at the upper boundary between the slipped and unslipped zones of the mainshock. We suggest that the atypical aftershocks are caused by the disturbance of the local tectonic stress field due to the heterogeneous coseismic slip.

View full abstract

The inversion of electron density from total electron content (TEC) measurements of GPS radio occultation is investigated by means of simulated data from the International Reference Ionosphere IRI-2001 and observations by the GPS/MET satellite experiment. In both cases a meridional slice of electron density is derived for northern summer solstice June/July 1995 from the midnight to the noon sector of the Earth's ionosphere. By means of the simulated occultation data a new 2-D recovery method is tested considering electron density variations along the ray path through a non-spherical ionosphere. This method is as fast as the Abel inversion. The relative retrieval error is less than a few percent around and beyond the F₂-layer peak. The resolution (in latitude) of the 2-D recovery method is significantly better than those of the Abel inversion which assumes spherical symmetry of the ionosphere along the ray path. After this simulation test, the 2-D recovery method and the Abel inversion are applied to the noon-midnight GPS/MET data in June/July 1995, near to solar minimum. The advantages of the 2-D recovery method are in case of the GPS/MET observations questionable. This could be due to data gaps and TEC errors disturbing the 2-D recovery method more than the robust Abel inversion. Finally meridional slices are derived for other local times by the Abel inversion. Because of missing data and uncertainty of possible retrieval errors the discussion of the diurnal and global variations of the F region is confined to the most significant features. Clear departures between GPS/MET and IRI are found for the polar winter ionosphere and the nighttime topside ionosphere at low latitudes. Generally the GPS/MET observations and the IRI predictions agree well for most local times and latitude regions.

View full abstract

The dependence of maximum variations of H-component negative bays at magnetic conjugate stations on a sign of the IMF B_Y-component during the substorms of small and moderate intensity with |ΔH|max < 750 nT developed at B_Z < 0 and |B_Z /B_Y | < 1 has been studied. By data from Kotzebue (Φ = 64.5°; Λ = 249.7°) and Macquarie Isl. (Φ = -64.5°; Λ = 247.8°) observatories in the eastern hemisphere and Leirvogur (Φ = 65.3°; Λ = 68.2°) and Syowa (Φ = -66.1°; Λ = 71.0°) stations in the western hemisphere it is shown that at B_Y < 0 the absolute values of negative bays |ΔH|max are greater in the southern hemisphere and at B_Y > 0, on the contrary, in the northern hemisphere.

View full abstract

We present a case study of substorm-time magnetic field perturbations in the high-altitude polar magnetosphere, corresponding to the polar lobe, using data from the POLAR spacecraft together with ground-based CANOPUS observations and WIND solar-wind observations. The substorm of this paper had a clear growth phase and one clear major expansion onset, enabling an unmistakable comparison of the substorm timings and those of the magnetic field perturbations observed by POLAR. During the growth phase, the magnetic field strength (B_total) increased and the flaring angle of the field line decreased at POLAR: They are ascribed to the pileup of magnetic field lines over the polar magnetosphere and to the braking effect of the ionosphere on the tailward motion of the piled-up field lines, respectively. For -28 min after the expansion onset, B_total at POLAR did not decrease toward its presubstorm value, in contrast to past tail-lobe observations where B_total decreased during the expansion phase. This absence of the field-decrease signature in the polar lobe for -28 min, reported for the first time in this paper, could be ascribed to dipolarization/compression of the inner magnetosphere during the expansion phase, cancelling the field-decreasing effect.

View full abstract

Lunar Radar Sounder (LRS) is a spaceborne HF radar system and is a science mission of Japanese lunar exploration project, SELENE, which is scheduled to be launched in 2005. The primary objective of LRS is to investigate the geologic structure of lunar subsurface from orbit. Computer simulations of LRS observation of lunar mare region have been carried out by utilizing a newly developed simulation code, the Kirchhoff-approximation Sounding Simulation (KiSS) code. The purpose of the simulations is to understand the nature of reflection/refraction of HF wave at the lunar surface as well as at the lunar subsurface boundary, and to confirm that the lunar subsurface structure can be investigated from orbit by means of an HF radar. Gaussian random rough surfaces are employed to represent the surface feature of a lunar mare region. From simulation results, we have found that the power flux of both surface nadir echo and subsurface nadir echo vary little if roughness of either/both surface or/and subsurface boundary interface changes. However, their intensity of surface off-nadir backscattering echo varies following a power law of (kσ0) ², where k is the wave number of LRS transmission pulse, and σ₀ is the RMS height of the surface. Thus slight roughness of the surface causes significant increase of the power flux of surface offnadir backscattering echo, which easily masks weak subsurface echoes. These observations have been understood qualitatively by geometrical optics approximation and quantitatively by examining the Stratton's integral formula in an analytic way. Computer simulations have revealed that subsurface echoes are received even if they are completely masked by surface off-nadir backscattering echo. To distinguish those subsurface echoes from strong surface backscattering echo, the data stacking technique has been proved to be effective on reducing surface backscattering echoes due to their random nature.

View full abstract

Subsurface sounding of lunar highland region by Lunar Radar Sounder of SELENE project has been investigated based on a computer simulation technique which enables us to analyze sounder echo signals from a modeled lunar surface. Using a numerically generated surface feature of a highland region which is characterized by impact craters, whole sequence of LRS observations, and data analyses have been simulated. The study was carried out on those simulated LRS observation data in order to establish a data analysis methodology of subsurface sounding in highland region. The established data analysis methodology consists of a set of data analyses, and gives a guide line to determine the detectability of subsurface signal. The points of the methodology are (1) the subsurface echo signal is recognized as a straight linear pattern in B-scan display, and (2) data stack technique is introduced to reduce surface off-nadir echoes, however, (3) the data stack should be done within the range of optimal data stack, and, (4) the limit of optimal data stack is determined from the behavior of the surface nadir echo.

View full abstract

Usu volcano, one of the most active volcanoes in Japan, erupted at regular intervals of ca. 30 years during the 1900s. The 2000 eruption took place on March 31 following 23 years of dormancy. An increase in seismic activity and remarkable ground deformation preceded the eruption. Many small-scale craters were successively formed in two areas at the foot of the volcano and ash was ejected intermittently from them. Ten samples out of the 2000 ash collected around the volcano during half a year were analyzed. They were depleted in SiO₂ but enriched in FeO compared with the essential ejecta of the five major historic eruptions and all were highly altered as a whole. XRD analysis revealed smectite as a major clay mineral in the ash and kaolin as a minor one. The clay mineral assemblage and the trend of change in chemical composition of the 2000 ash are quite similar to those of the 1977-1978 ash. No temporal changes in chemical composition and clay mineral assemblage of the 2000 ash were detected. The percentages of the LOI and the relative proportions of SiO₂ in the ashes and the essential ejecta showed a linear correlation. These results indicate that the constituents of the 2000 ash were alteration products of the volcaniclastics of the historic eruptions in the subsurface.

View full abstract