Yoshinobu Hoso

Crustal deformations in Kii peninsula associated with the SE off the Kii peninsula earthquake sequence of September 5, 2004 derived from dense GPS observations

We present crustal deformations in southern Kii peninsula, southwest Japan, before and after the SE off the Kii peninsula earthquake sequence of September 5, 2004, which were obtained from repeated surveys of dense GPS network. Disaster Prevention Research Institute (hereafter DPRI), Kyoto University made a reoccupation of their dense GPS network network that was occupied in February-March, 2004, right after the event to obtain crustal deformations associated with this event. Since we do not have the data just before the event, we apply an inversion of displacements of GEONET sites during the period from February-March to August, and interpolate displacements at DPRI’s sites using the inverted fault model. Southward movements of about 2 cm are obtained in the middle of Kii peninsula, while displacements are rotated to southwest on the west side. These characteristics of horizontal displacement field cannot be explained only by a pure thrust faulting, but additional sources such as right-lateral faulting on a NW-SE trending fault are required.

Seepage into a mountain tunnel and rain infiltration

Abstract To clarify the processes governing recharge of rainwater into an unconfined aquifer in a mountain basin, the seepage rate of water into a mountain tunnel was monitored continuously between May 1988 and December 1991 in Yura, Wakayama, Western Japan. The electrical conductivity and ionic composition of the seepage water was also measured regularly from June 1989. The mountain is composed of fractured sedimentary rocks such as sandstones and cherts. Seepage appeared only at specific locations on the roof of the tunnel. Seepage was measured through overlying sandstones of 10 m depth and cherts of 5 m depth. Time variations in the discharge suggested that seepage is formed both by rapid flow and basic flow components which correspond to the so-called fissure flow and matrix flow. This was confirmed by analysis of time variations in the concentration of chemical species in the seepage water. Fissure flow contributes the initial increment of seepage discharge, immediately after the occurrence of rainfall, and its velocity for sandstone can be represented by a function only of water content (θ), αθ n . α is constant and larger in winter than in summer, but n remains constant, at approximately two. Infiltration time in fissures through the base of the cherts is negligible, compared with that through the overlying soil layer. Whereas matrix flow through sandstones persists through the year, for cherts it disappears within a week after rainfall events and its decay is dependent on the seasonal magnitude of evapotranspiration. The behavior of the matrix flow for sandstones can be analyzed through a kinematic wave model, as can the fissure flow.

GPS observation of the first month of postseismic crustal deformation associated with the 2003 Tokachi-oki earthquake (MJMA 8.0), off southeastern Hokkaido, Japan

To investigate the postseismic crustal deformation associated with the Tokachi-oki earthquake (MJMA = 8.0) of 26 September 2003 in Japan Standard Time (JST), off southeastern Hokkaido, Japan, we newly established thirty GPS sites just after the mainshock in the eastern part of Hokkaido. Rapid data analysis for one month after the mainshock clearly indicated postseismic displacements only in the horizontal components. Observed maximum horizontal displacement was 6.6 cm from 28 September to 24 October, 2003. Absence of the vertical suggests that afterslip occurred in and around the coseismic fault rather than at downdip extension. Time series of coordinates are characterized by logarithmic decay functions with 4–11 days relaxation times. This suggests that postseismic deformation was due to afterslip on the fault following the large earthquake.

Postseismic displacements following the 2007 Noto peninsula earthquake detected by dense GPS observation

We have been conducting dense GPS observation in and around the epicentral region of the 2007 Noto peninsula earthquake since March 25, 2007, in order to detect postseismic displacements. Continuous observation has been underway at 12 sites to fill the gap of GEONET. Preliminary analysis of data up to early May shows that initial postseismic displacement rapidly decayed within 20 days after the occurrence of the mainshock. Horizontal displacements do not exceed 20 mm even at sites above the aftershock zone for this period. We also found a maximum uplift of about 20 mm there. Inversion of postseismic displacements with the variable slip model suggests a nearly right-lateral afterslip of less than 5 cm on the shallow portion of the source fault. Fitting a theoretical function to a time series of coordinate changes also suggests that the observed postseismic displacements might have been generated by afterslip.

Using short- and long-term transients in seepage discharge and chemistry in a mountain tunnel to quantify fracture and matrix water fluxes.

Abstract Infiltration of rain-water into a fractured sedimentary rock mountain is explored through continuous observations of discharge rate, Q , and electrical conductivity, EC, of seepage water into a mountain tunnel. Also concentrations, C CO 2 , of carbon dioxide gas near the tunnel ceiling, and the chemistry of the seeping water are examined. Earthquake events occurred in the period of the seepage observation and influenced characteristics of the time trends in Q and EC. This provided a mechanism for the identification of rapid flow (fissure flow) and slow flow (matrix flow) in the infiltration components in the fractured rock base. Also, a cycling of discharge water from the matrix via the fissures and back into the matrix was expected to occur. C CO 2 increased due to rainfall events, and its response was with a phase-shift to increased Q . For a heavy rainfall event, the increase in Q was mainly caused by the occurrence of fissure flow, and as soon as Q began to decrease moderately after a rapid decrease from a peak value, C CO 2 showed a peak value. The C CO 2 peak seemed to coincide with increased matrix flow. Wetting in the rock matrix was assumed to behave as a shock wave. For a light rainfall event, where only matrix flow is likely to occur in the fractured rock base, Q increases were delayed in comparison to C CO 2 increases. The variations in C CO 2 due to rainfall events appeared to relate to the movement of the matrix wetting front, when high moisture contents were apparent. The wetting front was inferred to be pushing void-airs with high concentrations of CO 2 gas towards the tunnel. High CO 2 concentrations were assumed to be formed near the ground surface via dissolution of organic matter and respiration of plant roots. The chemistry of seepage water observed at two close locations is seen to differ distinctly. Time-variations in EC for one location (A1) are consistent with those for C CO 2 , while for the other location (A3) this was not the case. The variations are due to dominant anions in the seepage water; HCO 3 − for A1 and SO 4 2− for A3. These occur via dissolution of CaCO 3 and CaSO 4 into infiltrating water, and CO 2 gas plays an important role in the former process. The time trends and integrated interpretation of the seepage volumes, chemistry of seepage water, and the concentration of CO 2 gas are shown to be useful indicators for understanding rainwater-infiltration process in the fractured rock mountain, and for separation of the seepage into fissure flow and matrix flow components.

Precipitation, Groundwater and Ground Deformation

Strainmeters and tiltmeters installed on or at a shallow part under the ground surface record ground deformations caused by rainfall. Such deformations are detected to a depth of several ten meters with these instruments, and accordingly they are main noises to observations of crustal movements and/or earth tides with these instruments.

Comparison of Crustal Deformations Observed with GPS and Strainmeters/Tiltmeters

Daily solutions from GPS observations can give uniform tectonic deformation and relative motions of stations more than 100km apart to each other. More precise wet delay correction is necessary in order to monitor local or the order of 10km deformations. On the other hand observation with strainmeters and tiltmeters show irregular changes which are probably associated with local stress accumulation or irregular concentration, but it is not easy to distinguish the irregular change caused by tectonic stress from that caused by meteorological or ground water disturbances. Combined observations with GPS and high sensitive strainmeters/tiltmeters are important to make clear the nature of observed irregular changes, and further to get information relating to the occurrence of earthquakes of M7 class and accordingly to approach the long-term earthquake prediction.