The observations of long-period (1- to 5- sec) microtremors for elucidating ground characteristics are very scarce until now, though the importance of the short-period microtremors in the earthquake engineering is well recognized. This is perhaps due to the background that almost all the buildings were constructed not so high in the past. It is, however, a fashion that recent buildings are very large both in height and width. In this new situation the technique to know the underground conditions by means of the short-period microtremors has little use. Our present purpose is to try out whether the long-period microtremors are powerful for detecting deeper ground conditions.
In 1973 an observation of the long-period microtremors was carried out in Hachinohe City, Aomori prefecture. The apparatuses employed there are 1cps electromagnetic seismometers having specially designed integration circuits to lengthen the observation capability to longer periods, DC-amplifiers, and 4chs portable magnetic data recorders.
According to the geological survey the Hachinohe area is classified into three major blocks of which bed rocks locate very shallow, moderate, and deep in depths. So, after examining whether the observation reflects differences in the blocks we set up A-line as to intersect the three blocks and observed 10 and more points. In Tokachi-oki earthquake of 1968 a very large acceleration with 2.5sec in period was recorded at Hachinohe Branch Office of the Bureau of Port and Harbor, but no appropriate interpretation to this is achieved yet. Thus the line from Hachinohe Local Meteorological Observatory lying just above the bed rock to the strong motion seismograph site was surveyed as B-line.
First we calculated the Fourier spectra of the records by means of an analog-type spectrum analyzer. Then these were classified into three peculiar types of which peaks in spectrum differ significantly. And it was ascertained that the corresponding subareas show a fairly good agreement to the geologically known blocks.
A very systematic change in the predominant period was observed along the B-line. The shortest period was 0.7sec at the point close to the Meteorological Observatory and the largest period obtained, 2.5sec, was at the same site where the strong motion seismograph had been installed. The well-known law of quarter wave length was applied to the ground structure to estimate the expected predominant periods along the B-line. The coincidence between the calculated and observed periods was very satisfactory. This suggests we might have recorded the strong motions with the periods ranging from 0.7 to 2.5sec in the Tokachi-oki earthquake, if the seismographs had been set along this line.
