The tank model is very simple as shown in Fig. 1. We can consider that it corresponds to the zonal structure of groundwater as shown in Fig. 2. In spite of its simple outlook, the behaviour of the tank model is not so simple. Corresponding to various types of input rainfall, it shows various types of response as shown in Fig. 5 by its nonlinear structure caused by the positions of side outlets which are set somewhat higher than the level of the bottom. The tank model shown in Fig. 1 is used to calculate the daily discharge from the daily precipitation for Japanese river basins. For the flood analysis, data of short time unit are necessary and an appropriate time unit is suggested to be given as T.U.=0.05√A, where T.U. is the time unit (hour) and A is the catchment area (km2). Table 1 shows some examples of appropriate time unit for various catchment areas. For the flood analysis the tank model with two tanks shown in Fig. 6 is applicable. In Japan, the tank model without soil moisture structure can give fairly good results because it is always very humid in Japan. However, for most river basins, the tank model with soil moisture structure shown in Fig. 7 must be applied. The assumed soil moisture structure is composed of two parts, the primary and the secondary soil moisture storages. When the primary soil moisture storage is not saturated, the water is absorbed from the lower tank and there is water transfer between the primary and the secondary soil moisture storages. These two kinds of water transfer are given as shown in Fig. 7c. In regions with long dry season, there is no tree on mountain area or trees have no leave in dry season and vegitation covering can be found on plains or along rivers. In such regions, mountain areas become dry during the dry season, because water moves to lower part of the basin by gravitation. To simulate such a basin, the basin is divided into zones each of which is simulated by the tank model. The tank model of 4X4 type shown in Fig. 8 is derived under such a consideration. During the dry season, zones become dry from mountain side and no evapotranspiration occurs in dry zones. In this model real evapotranspiration from the basin decreases with time corresponding to the dry condition of zones, i.e. areal real evapotranspiration of the basin decreases automatically. The tank model is considered as a black box model without physical meaning by most hydrologists. However, we can ask ourselves, if it is a mere black box, how can such a simple tank model successfully simulate river discharge from high flood to low base flow? There must be some physical meaning in the tank model. Very recently, we were able to find the phenomenon to prove the existence of two kinds of water storage corresponding to the top and the second tank of the tank model by analysing the record of crustal tilt meters affected by rainfall (Fig. 9). The crust is some sort of spring balance which weighs' and so measures the groundwater storage (Fig. 10).
Age of the Japan Sea has been controversial and its ambiguity makes an obstacle for the study of Cenozoic tectonic history of the Japanese Islands. Ages of the Japan and Yamato Basins of the Japan Sea were examined by three independent methods; stratigraphic consideration of the basin, age-depth relation, and age-heat flow relation. The stratigraphic constraints of the Yamato Basin are good by the presence of DSDP drilling holes in the central part and marginal part of the basin and bottom sampling at the outcrops of the lower sedimentary sequence. The age-depth relation of the open ocean is not applicable to the marginal seas, while the age-heat flow relation is valid in the marginal seas. Basement depths after sediment loading correction of the Japan Basin and the Yamato Basin were compared with those of other marginal basins in the Western Pacific whose ages have been well constrained. The age estimations of the Japan Basin by age-depth relation and age heat-flow relation show good coincidence. The age estimations of the Yamato Basin, however, show discrepancy among the three methods. The basement depth shows the age range of 6-0 Ma, the stratigraphic consideration shows that of around 10 Ma, and the heat flow data show estimated age older than 10 Ma. This discrepancy is due to thick accumulation of Neogene volcanonclastics (correlated to the Green Tuff Formation on Japanese Islands) which are represented by 3.5km/sec velocity layer on seismic refraction records and make an acoustic basement on seismic reflection record. As a preliminary conclusion, the age of both of the Japan Basin and the Yamato Basin is estimated to be roughly identical with the range of 30 Ma or older to 17 Ma or slightly younger. The identical age of the both basins contradicts the hypothesis of two stage of spreading of the Japan Sea; the older Japan Basin and the younger Yamato Basin. The presented age estimation is also against recently proposed clockwise rotation of the Southwest Japan at 15 Ma time with the duration less than 1 Ma which was documented by the paleomagnetic study and is believed to be a direct result of the opening of the Japan Sea. The discrepancy will be solved by a future more sophisticated understanding of tectonics of the Japan Sea.