Observations of Crustal Movements by Newly-Designed Horizontal Pendulum and Water-Tube Tiltmeters with Electromagnetic Transducers (3) --Time Variations of Tidal Admittance--

Abstract

In order to detect time variations in the amplitude and phase of earth tides, in relation to possible changes in crustal rigidity around earthquake source regions both the least squares and the time domain methods have been applied to tilt records. For the former method, it is effective for increasing the reliability of the results to analyze the observed and theoretical tides in the same way and to compare the results obtained for the M_2 constituent. For the latter method, it is effective to apply a band-pass filter which has dominant responses only for semidiurnal frequencies, removing meteorological influences with diurnal frequencies. For both of the methods, it is desirable to use the filtered data over one-month's duration to obtain the tidal admittance (the amplitude ratio and phase delay time) with good accuracies. Using these two methods, the amplitude ratios and phase differences of tidal tilts have been obtained from the records of three-component water-tube tiltmeters installed at the Kamitakara station (36°17'N, 137°20'E, H=800m) in the northwestern Chubu region, Japan, during the period from September 1977 to November 1979. The results show that time variations in the amplitude ratios in three different directions (S45°W, N45°W and E) do not exceed ±4% at their maximum values and ±2% for the RMS standard deviations. These variations are estimated to be nearly equal to the error limit of sensitivity measurement. It may be reasonable to conclude from the above results that elastic properties around the observation station have not changed significantly during the period. This conclusion may be consistent with the fact that seismic activity around this region has not been very high. In addition to the above, time variations of the tidal admittance from the water-tube tiltmeters are compared with those three-component strainmeters for the same period. No close correlations could be detected between the two types of observations. However, the tidal admittance for three-component tiltmeters and those for three-component strainmeters show consistent variations, respectively, over the entire period, although their absolute levels are different. The effects of precipitation and water discharge in the observation vault on the time variations of tidal admittance can not be clearly identified, while there are close correlations between these amounts and secular variations of ground tilts and strains. Several possibilities for these variations are discussed.