ESTIMATED AGES OF LATE PLEISTOCENE MARINE TERRACES IN JAPAN, DEDUCED FROM UPLIFT RATE

Abstract

Marine terraces formed during the last interglacial maximum of 125 KA and following high stands of sea level at 100 KA, 80 KA and 60 KA, are recognized at many localities throughout the world (Miyoshi, 1983). In Japan one or two marine terraces (Terrace I and Terrace II in descending order for this study) are found at altitudes lower than the last interglacial maximum terrace (S terrace). No radiometric dating for these terraces, excluding Kikai Island (Konishi et al., 1974) and South Kanto (Machida, 1975) has been made.
The purpose of this paper is to estimate the age of Terrace I and Terrace II in Japan, by using regression equation gradients (Ota et al., 1968). The gradients are calculated by plotting heights of the S terrace against younger shorelines, assuming that the rate of uplift has been uniform since 125 KA. Fifteen areas are selected in this study (Fig. 1 and Table 1), where previous work indicates that Terrace I and/or Terrace II are well developed and heights of former shoreline have been measured.
In order to test the method described above, the relationship is examined between the heights of 125 KA terrace and those of younger terraces at New Guinea (Veeh and Chappell, 1974; Bloom et al., 1974), Barbados (Broecker et al., 1968; Mesolella et al., 1969) and New Hebrides (Gaven et al., 1980; Jouannic et al., 1980) where radiometric dating data are available (Figs. 6 and 7). It is clear that the relationship between 125 KA (in abscissa) and 100 KA, 80KA and 60 KA terraces (in ordinate) is a positive, linear correlation. The gradients of the regression equations are 0. 82, 0.63 and 0.44 respectively (Fig. 7), being a function of the age ratio between the older and younger terraces. It can therefore be concluded that the assumption of uniform uplift in these three areas since 125 KA is valid.
Figure 8 shows the relationship of S terrace heights and Terrace I and Terrace II heights by using the above method for five areas in Japan where a large number of height data are available. The gradients of regression equations for Terrace I range from 0.6 to 0.7, and those of Terrace II are 0.47 and 0.50. The gradient of regression equation for Terrace I is 0.69 and that of Terrace II is 0.47 in the fifteen areas in Japan (Fig. 9). These values agree well with the gradients of 80 KA terrace (0.63) and 60 KA terrace (0.44) from New Guinea, Barbados and New Hebrides. It is probable that Terraces I and II were formed at 80 KA and 60 KA respectively, when sea level was relatively high due to the glacial eustasy. Terraces I and II can thus be correlated with the Obaradai and Misaki surfaces (Machida, 1975) in South Kanto. The marine terrace sequence in Japan is thus 125 KA (S), 80 KA (I) and 60 KA (II) terraces in descending order. The marine terraces formed at 100 KA, which is correlated with Hikihashi surface of South Kanto, are not recognized in this study.
Terrace I is well developed, its marine sediments bury bedrocks with irregular relief and its facies suggest transgression. In contrast with this, Terrace II is only locally developed, its marine sediments are thinner and cover an erosional bedrock surface (Fig. 10). It is a commom tendency in many areas of the world that the 80 KA marine terrace is well developed and its sediments suggest transgression (Miyoshi, 1983). Some sea level curves (Fig. 2) show that paleo sea level after the last interglacial maximum was relatively high at 80 KA, in contrast to a lower maximum at 60 KA. This trend is consistent with the marine terrace sequence in Japan.