The relation between subsoil condition and the collapse rate of wooden houses has been examined in Yokohama City where was severely damaged by the Great Kanto Earthquake in 1923. The authors have made a landform classification map (Fig. 1), a geologic section in the coastal lowland (Fig. 2), a contour map showing the basal landform of the post-Würm deposits (Fig. 3) and a distribution map of the collapse rate of wooden houses (Fig. 4) in order to provide themselves with the basic data for analysing the relation. The following results have been obtained from the statistical and the seis-mologic engineering analyses on these figures and many bore hole records used to construct Figs. 2 and 3.
(1) The relation between the collapse rate (P %) of wooden houses and the thickness (X meter) of the post-Würm deposits excluding Basal Gravel (BG ; cf Fig. 2) is shown in Fig. 5. The regression line of
P=2.7 X-17.2
with the correlation coefficient of 0.76 has been derived from a method of least squares.
(2) After classifying the post-Würm deposits into seven parts of X
1 (uppermost deposits ; mainly filled up soils), X
2 (humus), X
3 (clayey deposits with N-value of 5 or less), X
4 (clayey deposits with N-value of 6 or more), X
5 (sandy deposits with the N-value of 10 or less), X
6 (sandy deposits with the N-value of 11 or more) and X
7 (sand and gravel), the collapse rate (Pc %) of wooden houses has been related to the thickness of each of these seven parts by means of multiple regression analysis. As a result, the equation with the multiple regression coefficient of 0.87,
Pc=-0.99-2.05 X
1+3.30 X
2+2.33 X
3+1.42 X
4+1.99 X
5+2.33 X
6-0.05 X
7has been obtained (Fig. 6).
(3) On the basis of the results of frequency response analysis, the relations among the period and the value of the maximum response magnification factor and the collapse rate of wooden houses have been examined. As a result, it has been clarified that the collapse rate of wooden houses is largely determined by the period of the maximum response magnification factor rather than the value of that (Figs. 9 and 10).
View full abstract