This study demonstrates simulation of soil development and hillslope stability analysis on a geographic information system for prediction of location and volume of shallow landslides by rainfall in a granite hilly watershed near Kyoto, central Japan. Soil production rates were estimated by analysis of terrestrial cosmogenic ¹⁰Be in quartz in saprolite beneath soil layer, and then empirical coefficient of soil transport by soil creep was determined by relationship between soil thickness and topographic curvature in a nose, to calculate soil accumulation onto head hollow. Soil thickness reaches a steadystate on nose at mostly thinner than 0.5 m, and increases with time in hollow upto ~1.2 m for a few hundreds of years. We employed a slope stability model taking into account the effect of cohesive shear strength enhancement by tree roots. Return period of shallow landslide to remove soil in a hollow, i.e., lifetime of soil layer, is estimated to be 700-800 years, which is comparable to the timescale of forest regeneration. This approach helps understand the evolution of an eco-hydro-geomorphological system in temperate humid mountainous landscapes.

A distribution survey and a pull-out test of root systems in a Japanese cypress (Chamaecyparis obtusa) plantation were carried out to clarify the distribution of the collapse prevention force of the root system of surrounding trees. Based on the survey results and the relative values of the pull-out test resistance for each tree species, an estimation formula was created for the distribution of the collapse prevention force of a root system that consists of only three factors: tree species; coordinates; and diameter at breast height. This formula was consistent with values measured for various forests.

Stream flow within watersheds underlain by Paleozoic sedimentary rocks in Japan, tend to be characterized by sparse baseflow and significant peak flow. This flow regime can be attributed to subsurface flow processes within the weathered bedrock layers, which are well indurated and highly jointed. However, during storm runoff, these weathered layers were not considered to have a major effect on water movement due to their low permeability. This study analyses ground structure and moisture fluctuation at a steep hill slope that has no surface channels. The hill slope was located within a topographic hollow underlain by Paleozoic sedimentary rocks, in the Minami-dani catchment, within the Tatsunokuchi-yama Experimental Watershed, Japan. Borehole cores indicated that macroscopic water movement (through clayey material and joints) from the soil to the underlying relatively fresh bedrock basically conformed to Darcy’s Law. Therefore, it was assumed that hydrologic anisotropy, due to alternating sandstoneclay slate formation, could be ignored, and the weathered bed rock layers reaching the relatively fresh bedrock comprised the vadose zone, despite being permeable to 10⁻⁸-10⁻⁷ m/s. Therefore, at the mid-slope, the vertical extent of the vadose zone was greater than 18 m. This zone is characterized by a constant high water content, and low effective porosity. Therefore, rapid, vertical propagation of rainfall pulses, and large fluctuations of groundwater flow, were inferred for the mid and lower slopes. Occasionally, this also affected the upper slope depending upon the wetting degree of the hill slope vadose zone. During storm events, rapid and simultaneous responses were observed in groundwater stage and runoff. Direct runoff ratios became more than 50% when groundwater was observed in a 10 m deep well at the upper slope. At the mid-slope, when groundwater rose higher than 8.5 m in an 18 m deep well, the permeability of the weathered bedrock increased to 10⁻⁵ m/s (during earlier period of the recession limb of the groundwater). This permeability value is 100-1000 times greater than that during drier conditions, and is comparable to that of the surface soil layer. Therefore, the results from this study suggest that thick, weathered bedrock layers are functional on rainfall-runoff response during significant storm events.

To understand runoff processes in steep mountains, it is important to clarify the groundwater dynamics in the bedrock layer. Recent studies have found that bedrock groundwater is a dominant source of runoff from headwater catchments. To increase our understanding of this factor, our research group examined the spatial variability of baseflow characteristics based on hydrological and hydrochemical observations of a second-order catchment (2.3 ha) and seven nested zero-order catchments (0.1-0.5 ha) included in the second-order catchment underlain by granite bedrock. We also investigated bedrock groundwater using densely nested bedrock wells within each catchment. We discovered a catchment with a higher baseflow rate and SiO₂ concentration relative to the other catchments. The SiO₂ concentrations of the bedrock groundwater were variable and affected by the depth of the flowpath in the bedrock layer. The distribution of the bedrock groundwater table indicated that bedrock groundwater flowed to a neighboring catchment across the catchment boundary. The differences in baseflow among the zero-order catchments arose from the mixture of relatively shallow and deep bedrock groundwater. This study demonstrated that the bedrock groundwater flow across the surface divide produces spatial variability in baseflow, even in a small headwater catchment within homogeneous granite geology.

Residence times of groundwater and spring water were estimated using the concentrations of chlorofluorocarbons (CFCs) as tracers at four mountainous headwater catchments, and the mechanisms of determination for these residence times were discussed. It is mainly discussed that residence times are controlled by the topographic factors in many studies, although it is still a controversial subject. On the other hand, we did not find a clear relationship between the topographic factors and the residence times at our observation sites. Then, we considered the relationship between the residence times and the groundwater dynamics in each catchment, and found that the contribution of groundwater from deeper soil or bedrock layers and/or differences of bedrock geology are important controlling factors for the residence times. The detailed discussions about the mechanisms at small headwater scales will be important information to clarify the control factors of the residence times as one of the parameters of scaling issues of the hydrological processes.

Rainfall-runoff responses can be generally simulated by a single tank model with a one-to-one relationship between runoff rate and storage (called as QCS: quasisteady-state conversion system), whereas field observations in hillslope hydrology have demonstrated complex runoff processes with high heterogeneities. A contrast between the simple responses and the complex processes was addressed in this paper based on the mechanism of saturated-unsaturated flow. The following findings were obtained from simulation examples for a soil layer on a steep hillslope using the twodimensional Richards equation. The contrast was derived from a different influence of the spatial heterogeneities of soil-pore distribution on each of the unsaturated and saturated zones. In an unsaturated zone, the pressure head, volumetric soil water content, and unsaturated hydraulic conductivity in each local point simultaneously increased in response to an increase of the flow rate, even though the heterogeneities were involved. The soil-physical characteristics reflected a storage increase of the entire soil layer, contributing to an approximate applicability of QCS to the runoff responses, although an intervene dry zone with very small unsaturated hydraulic conductivity caused a hysteresis for the relationship between runoff rate and storage. In a saturated zone, however, an efficient drainage of groundwater through macropores was a key factor about the spatial heterogeneities, which played an important role in the runoff responses. This contributed well to applications of QCS through the reduction of groundwater-table rise. Overall, our analysis concluded that complex runoff processes in a hillslope soil layer with heterogeneous soil-physical properties might produce a QCS after the unsaturated hydraulic conductivity in the entire soil layer went up to a similar magnitude to the rainfall intensity (roughly from 1 to 100 mm h^－1) by an enough rainfall amount.

In order to evaluate the effects of slope gradient on unsaturated rainwater infiltration, this study conducted numerical experiments by solving the two-dimensional Richards’ equation for saturated-unsaturated soil water flow. We assumed slopes with a horizontal length of 10 m, a vertical soil thickness of 2 m, and slope gradients of 12.5° through 42.5°. The initial condition was attained by continuously supplying rainwater at a constant intensity corresponding to an annual effective rainfall. Then, rainfall intensity was increased to a level corresponding to a heavy storm event, and the dynamics of infiltrated rainwater were analyzed. The results indicated that, as the slope gradient increases, water infiltration becomes more diffusive in regions close to the soil surface. That is, in steeper slopes, water content increased more rapidly and required a longer time to reach a constant value than in gentler slopes. The diffusive infiltration phenomena in the steep slope are attributable to the horizontal water flux q_x, which contributed to rapid and delayed increases in water content immediately below and above the wetting front, respectively. These phenomena were observed in most parts of the steep slope, except for the downstream and upstream edges. At the downstream edge, the steep slope produced slower wetting front movements than the gentle slope. At the upstream edge, wetting front movements were faster in the steep slope than in the gentle slope. It is suggested that in the steep slopes the diffusive infiltration phenomena tend to contribute to more rapid increases in pore water pressures at the bottom of the soil profile.

Intersite comparisons might be a possible approach to clarify first-order control of rainfall-runoff response, since it has been difficult to derive general hydrologic principles from research studies of intensively studied one or two small basins. Here we completed intersite comparisons to discuss first-order control of rainfall-runoff response function in headwaters using recently developed database. We focused headwater catchments and collected 164 gauging station data in Japan. To characterize rainfall-runoff response function, we used Tank Model with four reservoirs and calibrated parameters for each gauging stations. Then, we conducted multiple linear regression analysis for clarifying roles of topography, geology and historical rainfall pattern. We found several roles of geology, historical rainfall and topography in rainfall-runoff response functions in mountainous catchment: (1) Rock types controlled rainfall-runoff response function, but roles of geological age were unclear. (2) Except for mean flowpath length, roles of topography were unclear. (3) Annual maximum daily rainfall affected rainfall-runoff response function. We confirmed the effectiveness of intersite comparison to understand first-order control of rainfall-runoff response function at headwater catchments.