A Search for D" Layer Silent Earthquakes Using the SG Records at Canberra, Australia, and the Extensometer Records at Matsushiro, Japan


I. Kawasaki

Disaster Prevention Research Institute, Kyoto University, Japan

kawasaki@rcep.dpri.kyoto-u.ac.jp


Solid Earth geophysics has been significantly developed for decades. However, many of excitation sources of the global dynamics such as geomagnetic jerk and irregularity of intervals of leap seconds have remained unknown and should be strange phenomena that are difficult to be recognized by prevailing seismological records. Here, we seek for symptoms of silent earthquakes in the lowest mantle, using vertical acceleration records of the super conducting gravimeter (SGC) at Canberra, Australia, and horizontal strain records by the 100m-long quartz-tube extensometer (EXTM) at Matsushiro, central Japan. Their noise levels seems to be among the lowest in the world. For convenience, they are tentatively called D" layer silent earthquakes and are abbreviated as DPSE. Accordingly, upper mantle silent earthquakes are abbreviated as UMSE hereafter. DPSE modes are defined as Earth's eigenmodes that could be excited by DPSE. From theoretical computation with several assumptions, we expect : (1) fundamental modes 0S4(1546s)-0S14(448s) for SGC and 0S3(2134s)-0S12(502s) and 0T2(2637s)-0T5(737s) for EXTM. (2) quasi-Stoneley modes 1S4(853s)-1S14(336s), 2S3(804s)-2S4(725s) and 2S10(416s)-2S13(345s) for SGC and 1S8(556s), 2S4(725s)-2S8(488s) and 1T3(695s)-1T9(408s) for EXTM. Although many of DPSE modes are the same as UMSE modes, 1S6(657s)-1S14(336s) in case of SGC and 1T3(695s)-1T9(408s) in case of EXTM have larger spectral amplitudes for DPSE than those for UMSE, which are called DPSE-distinction modes in this study. After applying the following procedure (1) to (5), we computed FFT and MEM spectra of SGC and EXTM records for more than two years from March, 1998 to June, 2000 : (1) Continuous records are re-sampled with 10 s time interval. (2) The re-sampled records are divided into day-long segments with a mutual time lag of 0.5 day. (3) Taper of hanning window is operated. (4) Band pass filter of 200s to 1800 s is operated in order to remove Earth's tides. (5) Atmospheric pressure disturbance is corrected by linear regression The length of the records is 8192x10s and thus the frequency interval is 0.0122mHz. We define a spectral peak as that of a DPSE mode, if following two criteria are met : (1) The MEM spectral peak is located within 0.006 mHz of corresponding DPSE mode. (2) Its spectral amplitude is two times larger than the noise level defined by stacking of FFT spectra of records collected on quiet days. The average number of eigenmodes identified are around 3 for SGC records and 4-5 for EXTM records, respectively. In day-long segment records of 3/24-25/99, 4/2-3/99 and 5/07/99, around 10 eigenmodes for SGC records and around 15 for EXT records are identified, respectively. By assuming the Poisson distribution for the number of UMSE and DPSE modes identified in the day-long segment records, we found that the hypothesis that UMSE or DPSE occurred in those days can be statistically accepted. Although we thus can conclude that Mw8 class silent earthquakes occurred, we could not determine wether they are UMSE or DPSE by the statistical test. However, in 4/2-3/99, DPSE-distinction modes of 1S4, 1S5, 1S6 and 1S10 in SGC spectra and 1T4, 1T5 and 1T8 in EXTM spectra are clearly identified and thus we would like to conclude that DPSE occurred in 4/2-3/99. This anomalous excitation of UMSE and DPSE modes is not observed in the IRIS/IDA records, likely because their noise levers are higher than those of SGC and EXTM. On the other hand, although around 10 of UMSE and DPSE modes are identified in the spectra of atmospheric pressure records of 4/2-3/99, very few UMSE and DPSE modes are found in the records of 3/24-25/99 and 5/07/99. In this sense, our conclusion above is just tentative. Nonetheless, as a candidate of the unknown excitation source of the global dynamics, DPSE seems to be worthy of further investigation. This study is undertaken by DPSE search project team of I. Kawasaki, Y. Imanishi, K. Funo, M. Hara, Y. Sakai, S. Kashiwabara, O. Kamigaichi, and J. Funasaki.


Back to the list of abstracts.
Back to main page.