A water balance model for a subsurface dam in the Quaternary limestone region in Okinawa, Japan, was developed. Conduit flow through karstic features such as caves and springs in the catchment area was represented by conduit-flow submodel. Fluctuations in groundwater level and conduit flow discharge were simulated, and the results showed that increased conduit flow discharge due to intensified rainfall can cause a decrease in water storage of the subsurface dam; hence occupancy of conduit flow discharge in the water budget should be considered in climate change impact evaluation. As well as improved model parameters and inputs of the water balance model, alternative use of distributed groundwater flow model would lead to better estimation of the amount of conduit flow discharge.

  We investigated temporal variations in the landslide prevention function of forest root networks on shallow landslides resulting from coniferous plantation growth, and the recovery period of the landslide prevention function of forest root networks on shallow landslides in broad-leaf forests on shallow landslide scars in an area of Shirasu-soil (Ito pyroclastic flow deposits) steep slopes in Kagoshima Prefecture, Japan. The study site has an area of 3.94 km². To study the temporal variations in the landslide prevention function resulting from coniferous plantation growth in the study site, we interpreted aerial photographs taken in 1965, 1967, 1970, 1974, 1978, 1989, 1994 and 2003, and detected temporal variations in the occurrence of shallow landslides. The numbers and areas of shallow landslides were greater for tree ages of 20 years or less than for tree ages of 20 years or more. The prevention function increased over time and was restored after 20 years. Conversely, on slopes covered with broad-leaf forests on shallow landslide scars the prevention function was restored after 50 years, because the process of natural forest recovery was slower in these areas. This finding indicates that the risk of landslides in broad-leaf forests persists for 30 years after the occurrence of shallow landslides, considerably longer than in actively managed coniferous plantations.

  In the Tedori River alluvial fan, the groundwater level significantly decreased immediately after the occurrence of high-turbidity water. This study was conducted to clarify the temporal and spatial characteristics of groundwater level decline due to high turbidity water by comparing it with those caused by heavy groundwater use for snow melting. Furthermore, an experiment that replicated the impact of high-turbidity water and farming operations, using 1/2000a Wagner pots, was conducted to grasp changes in paddy field percolation rate following the occurrence of high-turbidity water. The results revealed that the impact of groundwater use for snow melting on groundwater level decline was limited to areas with high-density building sites and only in the short term. In contrast, the impact of high-turbidity water lasted about two years over a wide area around the Tedori River from the fan top to the fan center. And the replicated experiment showed that the amount of paddy field percolation was suppressed by the inflow of high-turbidity water before midsummer drainage. However, the suppression of paddy field percolation by high-turbidity water could be improved after midsummer drainage, depending on how the drying process formed the cracks.

  This study aims to quantify the rainfall storage effect of paddy fields during a flood, using runoff analysis with discharge rate data from an outlet with a notch for outflow control. The study site is located in the low-lying paddy area in the Chikugo basin. To quantify the flood mitigation capacity of the paddy field, outflow rate from the outlet of the paddy field was estimated and rainfall runoff analysis was conducted during actual inland flooding periods. Using the Radar AMeDAS Rainfall as model input data, spatial distribution of the rain in the study site was analyzed. To study spatial distribution of flood mitigation capacity, the object area was divided into 36 sections mesh of 1 km² and rainfall storage effect was quantified for each section. By calculating outflow over the paddy field notch, time variation of rainfall storage discharge was identified and the relationship between precipitation and flood mitigation capacity of paddy field was clarified. Additionally, division of the drainage basin into the Radar AMeDAS Rainfall data led to an improvement in the analysis precision of flood mitigation in the whole target district.