The objective of this research is to clarify the structure of the slip surface in Miocene siliceous mudstones where rock slidesoccurred frequently. In this research, test pit observations, soil tests, surveys of the slip surface using pipe strain gauges andtopographical analyses were conducted at the Yachi landslide, alarge-scale active landslide in northeast Japan. As a result, itwas found that the slip surface is formed in thetuff layers with lowest residual strength sandwiched between themudstones. It was also shown that the slip surface is discontinuous at the locations of depressions and tension cracks. At the toepart, it is also discontinuous and heterogeneous with an imbricated structure. Consequently, the behavior of groundwater flow is also affected by such discontinuous structure of the slip surface. Inaddition to clarifying the structure of the slip surface wherelandslides occurred in siliceous mudstone, the results of thisstudy will also be helpful for slip surface investigation, analysis of groundwater infiltration in landslide areas and planning ofgroundwater drainage works.
This paper describes the mechanism of occurrence of a landslide with flow-type collapse at Hishikari town, Kagoshima Prefecture, Japan. The geological structure of this area is as follows: (1) Collapse soil is on the bedrock (weathering tuff breccia) and is about 3m thick. (2) The inclination of the bedrock is 27C. (3) There is transported cohesive soil along the boundary between the collapse soil and bedrock. When there is a friction-type sliding surface in transported cohesive soil, weak faces are composed of these two parts. When the total length of the sliding surface is ls and the length of the friction-type sliding surface is lf, then the "weak face ratio" Rf is expressed by Rf=lf/ls. The weak face ratio is generally about 0.05 for a stable slope, but increases when the debris on weak faces is repeatedly saturated by rainfall, and the ratio becomes unstable when it exceeds 0.2. When rainfall exceeds QEP, creep failure of the weak faces occurs, so QEP is given by QEP=100·hs, where QEP is an effective preceding rainfall, and hs is the depth of weak faces. When a landslide of this type occurs, sliding debris flows as groundwater flows into the sliding debris.
Assessment of landslide potential over a wide area is a crucial issue for mitigation of landslide disasters especially in the Hokuriku Region (Niigata, Toyama and Ishikawa Prefecture) , central Japan, where a large number of landslides are induced not only by intensive rainfall but also by snow melting. Hydrological data on dam inflow was analyzed in order to predict the occurrence of landslides induced by hydrological phenomena such as change of subsurface water flow and increase of groundwater pressure. Several hydrological indices to be analyzed their statistical characteristics were formulated on the basis of the hydrological observation results including dam inflow discharge from 13 selected dams on which few artificial influences are found. The relationship between those hydrological indices and landslide occurrences was examined using data set on landslides in the Hokuriku Region during 5 years. As a result the peak daily water head at the dam preceding the landslide occurrences were selected as the appropriate index for prediction of landslide occurrences because it shows the highest probability of detection (POD) and the lowest error rate with the least expected days exceeding the critical value. The selected index shows the value about 70% as the POD with 15 days exceeding the critical value per year on average. The index is useful as prediction criteria and applicable for initiating site patrolling and inspection.