Search

The slope failures are provoked by the several natural phenomena and artificial operations. The slope failures accompanied by the development in the hills increase and become the important social problem. In this paper, the influence of development in the hills is explained, and the main causes of the many occurrences of the slope failure had to do with the level of the investigation method and the civil engineering works. Another problem was no perfection of the restriction of the law. When the developments of the hills are projected, the main civil engineering works are the developments of the building land and the construction of the road and dams. The investigation methods of the slope in these developments are discussed and the points of problem are explained. After the development has been perfected, too, the observations of the all kinds of instruments are found to be significant.When the occurrence of the landslide occures rapidly, the methods of many kinds of the emergency disaster prevention must be planned, and the effect of the system of the watch was proved in many cases.

The author has engaged in the research on erosion, sediment transport and sedimentation for forty years since his graduation from Kyoto University in 1952 to the time of his retirement from Kyoto University in 1992.Firstly, he stressed the importance of field observation by taking an example from the results by Hodaka Sedimentary Observatory of disaster prevention research institute, Kyoto University which has continued the field study on sediment erosion and sediment transport since 1965. The authors found the existence of step-pool morphology which has an important influence on sediment transport process in mountain creek. This finding stimulated the fundamental investigation on step-pool problems and led to successful results.Secondly, he reviewed his study on the mechanism of sediment problems. His most interest in sediment research has been toward to the mechanism of the phenomena. He conducted the investigation of several sediment problem in mountain region and alluvial rivers. Among those some of the studies relating sediment transport, river bed variation and debris flow are reviewed in this paper.Lastly, he mentioned the phenomena relating to the catastrophic sediment disasters by taking some examples.

The drift wood flowing down the river in the midst of a flood is one of the main causes that increases the damages of disaster.When the drift wood is dammed up by the bridge piers, the water level in the river is increased and sometimes bank breach occurs due to overtopping of water.In such a case, drift wood flows into protected low-land with flooding water and destroys the houses or gives the severe damages to them.After the flood, drift wood is usually mounded on streets by jamming between the houses and mounded on farms.These phenomena not only damage to crops and houses but also disturb the restoration of the disaster.In this study, experiments for damming um of the drift wood in one and two dimensional simplified flow fields are carried out. The dam up factor of the drift wood is derived from one dimenasional experimental results.By using this factor, experimental results are comparatively well explained. It is also found that in two dimensional experiments, because the incidence angle of the drift wood to the fore-front row of house blocks becomes large owing to the lateral spread of water, the drift wood is more easily caught on houses than in one dimensional experiments.

A mountainous rock cliff collapsed on to the National heighway of Route 327 in Miyazaki Prefecture, Japan, on 31 May 1990. Fortunately no human life was lost owing to the previous holding up of traffic. While the creep of the slope continued to be monitored, whole process of the failure was visually recorded. The dip of stratum is opposite to the slope, whose rock is schalstein of Paleogene Miocene.Covering the slope with mortar was going on after cutting the slope as improvement work of the highway.The relationship between the rupture time and the strain rate which was obtained both from extensometers and from the visual records showed that the creep process of the rock slope to the failure can be explained with Saito's theory (1968) on the tertiary creep, in which as time goes on, strain rate increases along the curve of the Saito's formula (1966) on secondary creep. Though the motion of the final stage for 20 seconds just before the climax of the failure has taken a similar process in general as the law in the Saito's tertiary creep theory, this final creep seemed to have slightly changed its mechanism. The creep accelerated almost continuously, but rock fall occurred intermittently, alternating active and inactive periods of rock fall. Its frequency had increased since early time, but the main portion of volumetric accumulation of fallen debris had focused on the climax time of the failure. After the climax, secondary failures of rather smaller scale continued to occur intermittently for several hours.

In most rivers in Japan, tolerable degradation occurs because of the sediment trap by the reservoirs upstream, and the severe gravel mining which was done during the period of rapid increase in economy.As a result, disasters during flood have been increased due to local scouring around river dykes and other structures.Therefore, it is required to predict and mitigate the river bed variation not only for it' s mean value but also for local scouring, especially for meander channel.The authors investigate experimentally the channel deformation, bed variation and velocity distribution in an experimental meandering channel with flood plain of alternating bars, which was constructed by modeling the channel of the Naka river.More over, the mitigation of local scouring by groins was investigated.An analytical approach was also presented for predicting the river bed variation and for estimating the effect of groins for the mitigation of local scouring.The analitical approach was found to be applicable for the experiment.

Toppled rock blocks from the steep cliff along the Echizen Coast, Fukui Prefecture, hit apart of the shelter (rock-shed) below, and 15 persons in a micro-bus just passing there were killedby the breakage of the shelter on July 16th, 1989. The rock falls foregoing the toppling at leastfor one hour were reported by two eye-witnesses who had watched the full-process of the rock wallcollapse. Their statements give cleary the mode of the collapse starting with the tilting (toppl-ing) of a part of the joint block composing the partly over-hanged sea cliff. The joint block tiltedabout 30 degrees was then disintegrated, falling vertically on the earth surface. Distribution ofthe scattered blocks after the fall suggests the rotation of the blocks around an inclined axis atthe base of the cliff. The rotation is also confirmed by the fabric analysis of the distributed blocks.A part of the disintegrated blocks hit the rock-shed roof at its edge. Intensity of the impactforce given to the soil-covered roof is estimated to be 1100 ton-force, assuming the volume andsize of the individual block deduced from the survey after the fall.

The present study emphasizes that no differences exist in the dynamics of grainmovement in debris flow and general bed-loads, so far as bed shear stress is keptgreater enough for active movement of sediment.The structure of shear stress in the sediment layer under consideration isdiscussed in terms of a theory of debris flows which has been developed by theauthors. This brings about the differential equations for velocity and particleconcentration in the moving layer of sediment. These equations are solvedanalytically, assuming a linear profile of sediment concentration, and also analysednumerically. No significant differences are found between the results obtained bythe two; the approximate solution and the numerical solution.A bed load equation, which should be called a "sediment load formula", isdeveloped by using the formulas for the velocity profile of sediment layer and theflow depth of sediment layer. According to the formula, non-dimensional bed loadrate is proportional to r*5/2 (r* is the non-dimensional shear stress). This equation isverified in terms of flume data.

Observational studies on the runoff, sediment yield and the sediment transporthave been carried out in an experimental basin, the Ashiaraidani creek (6.5 km2), which is a tributary of the Jintsu River.(1) The relation between sediment yield and rainfall, slope grade are examined, to clarify the processes of sediment yield in various slopes, taking seasonalvariations as well as site conditions into accounts. (2) The relation betweensediment runoff and sediment storage in the steep channel with pools wereinvestigated in the Hirudani creek (0.85 km2). (3) The occurrence of debris flow wasinvestigated with attention focused on rainfall intensity and local rainfall in theShiramizudani creek (2.1 km2).

In a debris flow which is composed of widely distributed sediment mixture, verysmall fraction and other larger particles behave differently. Namely, the very smallone is suspended in the interstitial fluid among larger particles which are sustainedby the action of mutual encounters. Concentration of such fine particles isimportant, because increased density of the interstitial highly sediment laden fluidworks to increase capacity for transporting larger particles. In this paper, thecriteria to divide fine and coarse particles is given through the discussion on theturbulence and suspended sediment concentration in the interstitial fluid. Largerparticles whose mobility is helped by the buoyancy due to increased density ofsuspending fluid segregate themselves in the manner to concentrate largest particlesagainst the forefront and smaller ones to the rear. A computer simulation methodto predict such phenomena as well as to render possible to predict debris flowhydrograph, solid concentration distribution in the flow, etc. is proposed andvalidated by the laboratory experiments.

The purpose of this study is to get an experimental basis for the mechanism of mass move-ments, and to obtain exact knowledge on the influences on friction angle by shearing velocities.High-speed, high-stress ring shear tests on sandy and clayey materials have been carried out toexamine whether there is any change in the frictional behavior of these materials at high shearingvelocities of 0-l50 cm/sec and high normal stresses of 0-3.8 kgf/cm2. Sample used for tests wereglass beads, sandy soils in the university campus, the Toyoura standard (seashore) sands andbentonite clays. All samples were dry. Although results on glass beads was independent ofeither velocity or normal stress over the entire range of both shearing velocity and stress, changesin frictional coefficients were observed on the Toyoura sands and bentonite clays. In the Toyourasand, not only the fracture of sand particles, but the change in grain size distribution during shearseems to be the major reason. As for bentonite clays, fracture and griding of clay peds duringshear and may be effective for the variation of friction coefficient.

Horizontal velocity profile, concentration of gravel and gravel-size distribution were ob-tained from the visual data of debris flow at the field observation site of the Kamikamihori valleyon the eastern slope of Mount Yakedake, in order to make clear the flow characteristics and thesize characteristics of the gravel in the flow. Three typical surface-velocity distribution, one witha plug flow transversely in the central part of the flow, the smooth distribution with a horizontalshear, and the highly turbulent one, were found in debris flows. The mobility factor, which wasdefined as the ratio of the surface velocity, to the friction velocity of the flow, was inversely cor-relative with the content of gravel. In case of a successive occurrence of debris flow surge, con-centration of gravel in the frontal part of the surge had a tendency to decrease one after another.There was a close correlation between the size-distribution quantities of gravel and the flowscale, while the diameter of the largest boulder in each flow was nearly equal to its maximumflow depth.

Prediction of the hydrograph of the debris flow which occurs in an actual watershed neces-sitates the analysis of its formation on varied slope under irregular bed thickness, width, ma-terial, seepage water, and surface water supply conditions. In this paper, authors present thesystem of equations to predict the process. This system is not only able to estimate the hydro-graph of the debris flow at arbitrary position in the channel, but also obtain the change of theconcentration of particles in the flow separating them into coarse and fine components. Applica-tion of this system both to the laboratory experiments and the actual huge mud flow generatedby the eruption of Mt. Nevado Del Ruiz, Colombia, proved that it is a promising method.

The authors carried out some observation on the runoff, sediment yield and transport in asmall mountain watershed, called Hirudani Experimental Watershed, which is a tributary of theJintsu River in Gifu Prefecture, Japan. This paper presents some results of the observations onrunoff characteristics, on rate of sediment yield and transport by petrologic classification of gravels, settled and transported, and of the survey on the change of river bed before and after floods.The following interesting fact are pointed Out after studing the observations and surveyresults.1) The characteristics of transport materials reflect the process of the runoff over the sedimentsource areas.2) The process of the sediment transport in a mountain watershed is revealed throughthe distribution of the grain size and petrologic character of gravels in the transport materials.3) The river bed variations as well as the sediment yield should be detected for estimationof the sediment discharge from a basin.

The effect of vertical acceleration on slope stability is analysed and found to be insignificant inassessing the value of critical acceleration. A simple relationship between static factor of safetyF and horizontal critical acceleration K is derived for an infinite slab- and a simplified slice model.The critical acceleration for a slope failed in 1978 Izu-Oshima earthquake and in rainfalls precedingand following the event is investigated and the effect of water content of surface soil on the criticalacceleration is evaluated. The speed of a slide in the same earthquake is analysed and thecoefficient of friction of 11° to 15° is deduced on the basis of energy-conservation law.

The project has been conducted in order to develop the method for predicting the runoutdistance of a soil block and a debris flow. The authours have presented a friction law in theforegoing papers.The friction law, which is a function of particle concentration, particle diameter and flow depth, is tested by flume experiments concerning liquified soil block and debri flow. The motion ofdebris flow is evaluated well by the formula.Flume experiments are carried out to investigate the process of debris flow deposition.The experimental results show that the running out distance of debris flow and the final shape ofdebris fan greately depend on particle size as well as slope gradient.Two numerical methods to calculate the shape of debris fan, which are based on Eulerianand Lagrangian coordinate systems respectively, are proposed. The final shapes simulated by thesemodels coincide with those from experiments.

The Totsukawa hazard occurred in August 1889 was triggered by heavy rainfall, and it isfamous at the report of the hazard published two years after the occurrence. The report consistsof 11 volumes, and the each contains the detailed description on characteristics of weather con-dition and landslides associated with the statistical tables. Reexamination of the record fromscientific point of view was carried out this time.The weather condition which brought this hazard was confirmed by the early weather maps.A small typhoon travelled the Shikoku and Chugoku districts from south to north slowly andcaused heavy rainfall at Kii Penninsula. Though measurement of precipitation had not beencarried out at that time in Nara Prefecture, it is speculated that the maximum daily precipitationwas at least about 1000 mm, by analyses of the data along the coast at that day and of the dis-tribution pattern of precipitation accompanied by recent typhoons which followed almost thesame course as in 1889.It is shown from quantitative consideration that the frequency of landslides is reciprocallyproportional to the logarithmic area of them. Data concerning the depth of landslides was col-lected too, and the regression line for area-depth relationship was obtained, by which estimationof the volume of landslides was carried out. The largest volume of landslides in this area haveorder of magnitudes over l0 7 m3. The large landslides often show a particular plan shape, namely, a rectangle considerablly close to a square.Sites of the largest landslides which damed up the river below are limited on the slopes facingnorth or northwest in this area. In relation to geologic structure, they have been occurred on dip-slope of layered mesozoic system. It is pointed out too that the situation like minor syncline onthe major fold flank, fault crossing the slope, or under-cutting of the slope by river has had theslide occur easier. Conclusively, it should be emphasized, in comparison to the other similarcase of landslides occurred in the terrain where Mesozoic layered rocks are distributed, that heavyrain is more important than earthquake as the trigger of large-scale mass-wasting of this type.

A big debris flow of 36 million m3 was caused by the Nagano Seibu earthquake at Mt.Ontake, 1983, in Japan. This debris flow seemed to be apparently different from the ordi-nary debris flow, because it looked to be not saturated.Then, the oppinion which this flow was a dry avalanche supported by pore air pressurewas presented in Newspapers & TV. Others insisted that it was an ordinary debris flow be-cause there can be no cause of such a high pore air pressure.This paper proposed the mechanism of unsaturated debris flow in use of the idea of un-drained shear of the saturated torrent deposit, which is analogus to the hydro-planing phe-nomenon in a rainy high way.

Explanation of landslide mechanism is described from extensometer's and internal strainmeter's records. There are both a creeping landslide and a brittle landslide. Landslide surfaceseems to shift from one place to another. Analyzing the onset time of these records makes clearthat rupture occures in middle landslide slope and is propagated bilateral along a slope at a fewmeters per an hour or unilateral along a slope at dozens of meters per an hour. Extensometer'srecord put between head cliff makes clear that rise time in landslide is about ten and odd hours.

The rainfall conditions and geographical features on slope failure are analyzed by using thefield data.The rainfall condition for the occurrence of slope failure or debris flow is clearly demonstratdon a plane of maximum hourly precipitation and daily or total precipitation by grouping thegeological conditions. Not only the condition for the occurrence of slope failure but also the densityare determined by the rainfall intensity and total precipitation, of which relation is different fromplace by place due to the difference of the resisting force.A lot of information is obtained concerning the geographical features on slope failure.The geographical condition for the occurrence of slope failure is clearly demonstrated in thewatershed area and slope angle. The slope failure is apt to occur at the slope steeper than 30and at the place where the slope angle changes downward from steep to mild.

This paper concerns the motion of a soil block released by a slope failure. The momentumequation of a soil block sliding on slopes of various gradients is discussed first. Both the dynamicsolid friction and the fluid friction play an important role in the equation. The two parameters, therefore, are investigated; The dynamic friction coefficient Mk depends solely on static frictionMs, and their ratio Mk/Ms, takes a constant value. The theoretical discussion and experimentalresults on the fluid friction coefficient show the possibility that it can be defined inherently fromthe surface properties of the slope and materials of a soil block.The solution of the momentum equation are given to the sliding veloicity Thd running distanceof a soil block, and compared well with experiment and field data. The results of this study willbe very useful for the delineation of hazardous area due to slope failures.

A method was given that predicts the number of shallow landslides which occurin a mountainous basin with complex topography in case of a severe rain storm ofarbitrary intensity and time sequences. The method is based on the analysis ofinfiltrated flow, surface flow and the stability against slide in the top soil layers.The method was proved feasible by comparing to the actual data. Each slide unitwas allotted to a certain time and space, and if the route from the slide to thevalley bottom was steeper than 20 the slid earth block was supplied to the streamat the end of the route and eroded by the stream flow simultaneously calculated.The system of equations to analyse the sediment transport and channel bedvariation was used to estimate the sediment yield from the basin due to suchprocesses. The quantity of sediment yield changes with the distribution pattern ofthe landslides in the basin, but the width of variation was proved small incomparison to that change should be due to variation of other factors.

So-called open dams such as grid or slit type pass the sediment transported during normalfloods and store the materials due to debris flows which have big destructive forces. Moreover, there may be a possibility that the deposit is eroded gradually with flood and the storage capacityis restored. Therefore, the storage capacity can be effectively used for debris flow control.This paper examins the hydraulic function of grid type dam by experiments, and clarifies thatthe following conditions should be fulfilled for the effective uses, l/dmax=l.5～2l/db>2where, l is the span of grid mesh and dmax, db are respectively the maximum diameters of debrisflow material and of bed load sediment.

Because the junction of the stream channels in a basin is the most evidentchange in the boundary conditions for a debris flow on its way to run out to thedebouchment; the behavior and the hydrograph may change drastically there. Somecharacteristic behaviors and the methods to predict them at the junction area arediscussed in this paper. First, a system of equations to predict the locus oftravelling of the tributary debris flow front in the main channel with or withoutstream flow is presented and compared to the experimental data. Second, thebehaviors of the debris flows at junction area are classified to the categories; thatstops and deposits within the area (Type 1), that leaves some deposits andcontinues to run down (Type 2) and that passes through without any deposition(Type 3). The conditions to identify each type are given and the debris damformation is suggested. Third, for the Type 1 deposition a simple method to knowthe shape of debris cone which will be formed at the junction is proposed. Finally, for the case of Type 2 and 3 the method to calculate the deformation ofhydrograph via the junction area is presented and compared with the experimentallyobtained hydrograph.

Observations of debris flows at the Kamikamihori valley on the eastern slope of Mt. Yake-dake in the Japan Alps show quantitatively that many large boulders would concentrate at thefrontal part of debris flows, and that the size of the largest boulder is nearly equal to the max-imum flow depth.According to a well-circulated hypothesis, this focusing phenomena is due to a combinationof an inverse grading process and higher velocity at the upper layers of a flow. Two types of ex-periments showed that the inverse grading process may occur due to the effects of kinetic sievingand dispersive pressure depending on experimental conditions. However these effects were unex-pectedly small. Furthermore, the effect of upward motion of the largest boulder whose diameteris nearly equal to the flow depth is rather minor.Results of model experiments of debris flow using particle mixtures of three sizes with wateron a sloping channel showed not only a clear focusing of larger particles to a flow front, but alsoa lack of incorporation of large particles into the rear part of the flow. A test comparing thevelocity of particles of three different sizes in a bore water flow showed that larger particleswould soon attain a higher velocity and approached the bore front faster than did smaller parti-cles.A theoretical motion analysis of a boulder in a bore flow on a steep valley slope of upperreach showed that the larger a boulder is, the higher velocity exceeding the bore-front velocitythe boulder attains, so that a larger boulder can catch up with the bore front faster.

Characteristics of supercritical flows around curved sections of open channels are discussedin this paper. Laboratory experiments on 45° bend of clear water flow and high concentra-tion flow are accomplished and the applicability of the theories so far is checked. Lenau'sanalysis which is based on the energy, continuity, and irrotational equations better fits thanKnapp's wave concept which is based on the momentum, continuity, and the wave angleequations. There are no evident differences in the characteristics of the flow between clearwater and highly concentrated water. Amplification of the roll waves along the outer wallis rather well estimated by applying Lenau's formula. Among some available methods forthe reduction of superelevation in curves and in the downstream tangents, the method ofmultiple vertical curved vanes across the channel are discussed.

The study concerns the slide motion and its transformation into debris flow for a soil massproduced by a slope failure. First, slide motions are investigated by flume experiments with para-meters; inclination angle, supply discharge, soil characters and geometric shape. Simultaneously, mass and momentum equations are formulated for a soil mass sliding on a slope, and the theo-retical solutions are obtained for the sliding velocity and location of a soil mass. Second, themechanism is discussed for the transformation of a soil mass from slide motion into debris flow.A theoretical method is obtained to predict the critical slide distance which dicide the location wherethe motion of soil mass is transformed into debris flow. Formulas are compared with flumeexperiments for a slide motion and its transformation into debris flow.

The authors carried out some observations on the runoff of rain, sediment yield andtransport in a mountain watershed, called Ashiaraidani (7.2km2), which is a tributary ofthe Jintsu River in Gifu Prefecture, Japan.In this paper, the process and mechanism of sediment transport on the stream channelwere analyzed by petorological and grain-size classification of transported gravels for oneyear cycle at Hirudani experimental watershed (0.85 km2). The mechanism of occurrence and behavior of mud-debris flow were partially revealedthrough the field observation on variations of river bed, and the water balance of mud-debris flow at Ashiaraidani experimental watershed (6.5 km2).

Stability of the landslide block especially with an elongated plan shape and an arcuatecross profile is discussed theoretically at first by means of variation calculus. This kind oflandslide is treated as the transversal two-dimensional problem. The cross profile should beso that makes the slide block most unstable. The profile in homogeneous materials is thusthe arc or the ellipse which is the stationary curve obtained as the solution of iso-perimetricproblem in variation. This analysis introduces a new and reasonable method to evaluate thecohesive strength and the coefficient of friction of soil mass, based on the cross-profilemorphology of land slide block. Especially, the cohesive strength estimated in this way is ahalf as large as some value by the previous method applied to longitudinal profile. Applica-tion of this method to some actual cases is attempted and discussed.

Direct observation of debris flow and various surveys concerning the phenomena accom-panied with debris flow were carried out on the eastern slope of Mt. Yakedake in 1978. Inexecution of the former, a synthetic observation system equipped with many instrumentationswas set in operation which had been devised and put to practical use till the last year.The latter mainly consists of three types of survey. In the source area of valley Kamikamihori, measurement of slope erosion and rill profile was frequently carried out besides the mete-orological and hydrological observations. In the channel reach, some measurements were alsofrequently held concerning with microgeomorphological processes in the valley bed.Special surveys on debris flow was started at the alluvial fan this year in order to studythe physical character of the debris flow in retarding and stopping stages and to investigatethe distinguishing feature of debris deposition caused with decrease in slope angle andseparation of water from debris. Recording of moving state of the debris flow with cameras, microtopographical survey on erosional and depositional effects, and particle size analysis ofdeposit were carried out near the intersection point at the fan.As the result, we have obtained the following knowledges; (a) correlation between rainfallvector and wind direction or wind velocity, (b) relationship between the locality of rainfalldistribution and occurrence of debris flow, ? relationship between rainfall intensity andcoefficient of runoff or infiltration capacity, (d) lateral inclination of the head part of debrisflow, (e) exact measurement on surface velocity of debris flow by use of spatial filterspeedometer, (f) frequency characteristics of the ground vibration attended with debris flow, (g) seasonal changes in the valley bed processes, (h) characteristic moving state as acollective fluid of debris flow at the fan, (I) comparison of erosional and depositonal featuresbetween debris flow and sand flow, and (j) relationship between particle size distribution ofdebris and depositional conditions.

Debris flow control becomes a major problem to prevent and to mitigate sedimenthazards these ten years. A large number of concrete dams, called sabo dams, have beenconstructed to reduce or to control sediment runoff from mountain regions. Most of themare filled up soon by sediment in the areas of active sediment yield, decreasing theirsediment storage capacities. Concrete dams, consequently, could not have controlled debrisflows successfully if they lose their storage capacities. Instead of them, grid dams, whichare some kinds of open dams, may be attractive as a sediment control work.In the present study, two types of grid dam, zigzag and square grid dams, are chosenfor considerations, to investigate the mechanisms of debris flow control. Flume tests aremade with attention focused on the relation among pass rates of sediment through griddam, sediment particle sizes and openning lengths of grids. A mathematical model whichgives a boundary condition at the dam is presented to evaluate the pass rate of sedimentthrough the dams.Numerical analyses are made with Asida and Egashira's governing equations of debrisflow for the outflowing sediment rate through the dams as well as the deposition of sedimentfrom the debris flows in the upstream reach of the dams. These simulated results comparewell with the data from flume tests.

Historical records of large slope failures in Japan and foreign countries were collected andanalyzed to investigate occurrence frequency of the large slope failures and some dynamical andgeometrical features of debris deposits of the slope failures.It is shown that large slope failures have occurred in Japan much more frequently than inWestern countries. The relationship between the volume of slided debris and the equivalentcoefficient of friction for the debris motion is common to Japan and Western countries. Thefeatures of deposits spreading of slided debris showed a similar relationship between deposit areaand volume, but different shape over a wide range of debris volume with various types of slopefailures.

Severe landslides at the south-west part of Rokko-san mountains in 1938, 1961 and 1967are discussed and a time series of rain fall character which concerned the landslides is analyzed.From statistical analysis of the landslides and topography in 1967, it is shown that the numberof landslides per unit area increases till the critical angel (31) as the average slope angle in thearea increases. It is also shown that the number of landslides increases till the number of 1st orderstreams and the drainage density become the critical points, 105/km2 and 15 km/km2, respectivelyand it decreases above the values.The pattern of distribution in the number or the total area of landslides does not change in theseveral decades from the data of landslides in 1938, 1961 and 1967.Annual landslide rainfall is introduced as one of the rainfall characters which is the annualsummation of excess daily rainfall which is the excess quantity over 100 mm/day. The result ofharmonic analysis for the rainfall shows that the big extrema of the harmonic function are corre-sponding to the ages when severe landslides happened and the values also correspond to the scalesof the severe landslides.

Observational studies on the runoff, sediment yield and the mechanics of sediment trans-port have been carried out in the experimental basins, the Ashiaraidani(6.5 km2) and the Taka-haragawa(472 km2), which are tributary of the Jintsu River in Gifu Prefecture, Japan. In thisbasin all the aspects of sediment transport which appear in mountain ravines, namely, bed load, suspended load, wash load, and debris flow can be seen under various rainfall and dischargeconditions.In this paper, sediment production on bare slopes and gully walls and the behavior of thedebris fiow the stream channel and as it passes through a crib dam are discussed.

To estimate chemical weathering rate of granitic rocks, natural waters in a very smallbasin in the Rokko mountains have been collected weekly over a six-year period (1980-1985). The XRD indicates that kaolinite is the main product of plagioclase alteration inthe weathered granitic rocks. The water compositions are plotted in the field of stabilityof kaolinite in a phase diagram for the system Na-SiO2-Al2O3-H2O. This fits the XRDevidence that the main product phase is kaolinite, and it indicates that solution/kaoliniteequilibrium may play a role in controlling the water compositions. The amount of kaoliniteproduced by CO2-bearing water reacting with granitic rocks is estimated as 35.1 g/m2/year, and the total amount of weathering rocks is estimated as 55.3 g/m2/year.

The variation with time in the concentrations of dissolved chemical species and the dischargeof ground waters in the Kushibayashi area, Shiga Prefecture were investigated from Oct. 1972 toNov. 1973 in order to obtain the basic data for estimate of the amounts of clay minerals produced byweathering processes through the chemical compositions of ground waters in this area.The authors have examined the relationships among the chemical species contents of the waters, and between the chemical species contents and discharge of the waters. And the following have beenattained: (1) A clear negative relationship is observed between sulfate and bicarbonate contents inthe waters drainaged from the layer about 5 m depth under ground surface, (2) a good correlationis observed between discharge and bicarbonate and sodium contents, respectively, and (3) the trans-ported amount of bicarbonate ions increases with an increase in discharge but suddenly the increasingrate of transported amount changes after a definite discharge and the slpoe in the graph trends tobecome steep in the order 1 m, 5 m, and 10 m depths under ground surface.A hypothesis on the mechanism of this phenomenon is also presented.

To date, numerom researches have been pursued on the liquefaction of saturated loose sand inconnection with the investigation of the possible disastrous effects of earthquakes upon soft ground.In this report, the author attempts to devote a discussion on the following items: (1) liquefactionof quick soils, e.g. quick clay and true quick sand, (2) liquefaction of saturated sand due to cyclicloading and (3) evaluation of liquefaction potential of a deposit. It is further attempted to esti-mate the liquefaction potential at Osaka area for some design earthquake, and the results areshown in Fig. 8.

Now the landslide surface and the mechanism of a landslide are commonly discussed, based on surveys by means of the internal strain meter. On case of a shattered zone typeof landslide ares, however, the data gathered in this method have a strain value of unknownorigin and have many things unsolved. This time, therefore, the observation of undergrounddeformation was made with a pipe strain meter of ordinary type and an internal strain meterof insertion type at a shattered zone type of landslide areas; Yakuaji, Kuki snd Nishiikawa.It has proved helpful for finding where the landslide surface exists.

Large scale landslides induced by earthquakes in the historical years are investigated torestore the landslide topography to the original state, to reveal the origin and the mechanism, andto estimate the extent of the damaged area. The landslides discussed here (landslides at Kiun-zan, Machii and Kokuzo-san) were induced by the inland earthquakes of a magnitude greater thanM.7. All of these landslides formed natural dams, which were afterwards breached and causedfloods in the downstream areas. It is estimated that the loss of human lives reached severalhundreds for each of these landslides.

Observational research of the runoff process, sediment yield and transport have been carriedout in an experimental basin, Ashiaraidani (6.5 km2), which is located in the Northern Japan AlpsRange. In this basin all aspects of sediment transportation which appear in the mountain ravines, namely, bed load, suspended load, wash load, and debris flow can be seen under various rainfalland discharge conditions.Erosion processes of the bare slopes and gully walls and transports of gravels on the steepchannel as contact load and accompanying stream channel variations are discussed in view of theobserved facts in the Ashiaraidani. The process and mechanism of sediment transport duringa flood on the steep channels with sequential falls and pools were considered by means of thepetorological and grain-size classification of transported gravels in reference to the geologicaldistribution in the basin, and the relationships between flow rate and sediment discharge in theHirudani basin (0.85 km2) are explained as a function of sediment storage in the pools.

Mechanisms of slope failure and debris flow are chosen for consideration. These twosubjects are discussed separetely in the present paper. For the slope failure, occurrence ofslump and slide volume are discussed, with attention forcused on the geometric characteristicof wavy, impermeable interface of two layer model as well as on the characteristic of seepageflow over the wavy interface. Stability analysis of surface layer shows that the slide volumeof each failure is closely related to the potential slide volume. The potential slide volume, which depends on wave keight and length of interface, seepage flow discharge, slope gradi-ents and so on, can be predicted by a stability analysis.For the subjects on debris flow, discussions are made on the mechanical structure ofuniform debris flow as well as on the governing eguations, including the deposition velocityof sediment from debris flow. Formulas, which are based on a fluid model of a quasi-Bingham type, are presented for the velocity profile and the friction law of debris flow, andcompared with flume data. Governing equations for the running-out process of debris floware derived. These consist of mass and momentum conservation laws, and the equationgoverning bed level. Erosion and deposition of sediment play an important role in therunning-out process, as does the solid-solid and fluid-fluid friction coefficients as well. Theformula of deposition rate presented here is tested by flume data. Numerical predictionsfor the running-out process obtained with these equations compare fairly well with flumedata.

The sled model has been used for interpretation of landslide motion for long. However, what is the coefficient of friction during motion has not been known. The author hasproposed the method to decide the coefficient of friction by the new high speed ring shearapparatus and the new formula to express the pore water pressure during motion. Thisnew sled model will be effective for the prediction of travel distance and its applicationfor the Ontake landslide could give a reasonable interpretation to its high mobility.As the next step for hazard mapping, the author has deduced the new simulation methodby the combination of soil mechanics and hydraulics, which has enabled the prediction oflandslide motion as area.

Time variations in chloride ion contents of thirteen hot springs and those of water temper-atures in one groundwater and two hot springs have been observed for a few years in order toclarify the relation between these data and earthquakes in the Sanin District. The earthquake(M=6.2) occurred on Oct. 31, 1983 at the central part of Tottori Pref.Anomalous changes in chlolide ion content began at seven hot springs about one year beforethe earthquake occurrence. It is quite interesting because a microearthquake quiescence appearedin the vicinity of the epicenter in this period. Coseismic change of water temperature was observedat one hot spring, too. It is also interesting that the change of the water temperature is approxi-mately exponential.These facts show that deep groundwater moved in relation to the earthquake and that it isuseful to observe the chemistry and water temperature of hot springs in order to know suchmovement of the deep groundwater.

The 18th typhoon in 1974 carried heavy rain with it, and a lot of mountains inShikoku were damaged by landslides. The writer has walked around to see what are theactual condition of landslides in central Shikoku, and obtained results as follows.1) The landslides have two types; a small scale of landslide and a debris flow, andthey are closely connected with the geological condition of the mountains around.2) From a geomorphological viewpoint, the landslides are largely seen in areas witha steep slope and a high valley density.3) The greater part of the small scale of landslides occurred at the zero oder vally.

Systematic observation of debris flow was carried out at valley Kamikamihori on the easternslope of Mt. Yakedake in 1979. A synthetic observation system equipped with many instru-mentations was set in operation which had been devised and put to practical use in these ten years.In the source area, topographic changes of slopes were surveyed besides the hydrologicalobservations. In the channel reach, measurements were frequently carried out to reveal thegeomorphological processes in the valley bed, besides the observation of flowing state of debrisflow by the system.Special effort was made to study the depositional process of the debris flows. In order toreveal the dynamical aspects of debris flow in flowing, retarding and stopping stages, many typesof visual recording systems were set in the whole area of the fan. For the investigation of thedepositional features, leveling survey and polar coordinate survey have been frequently carriedout before and after debris flows.With heavy rainfalls, this year, seven debris flows occurred, among which the one on Aug.22 was the largest in the last ten years. Following items are discussed on the basis of observedfacts; (a) relationship between rainfall intensity and runoff at the time when debris flow occurrs, (b) exact calculation of gully runoff by the kinematic-wave method, ? violent erosion in therill and gully, (d) strength and frequency characteristics of ground vibration due to debris flows, (e) lateral inclination of the head part of debris flow, (f) seasonal changes in the valley profile, (g) dependence of debris accumulation behind the dams on the individual scale of debris flow, (h) arrival of a large debris flow to gentle slope with an inclination of 1-2°, (I) development ofthe fan due to the repeated deposition of debris lobes, (j) erosional and depositional effects bydebris flow and sand flow over the fan, (k) a low infiltration loss of water in the debris flow atthe fan.

The study of the sedimentation system such as detachment, transportation anddeposition of sediments are necessary to develop a method for evaluating sedimenta-tion processes in a drainage basin. Setting up the experimental basin in the Ashia-raidani creek (7.2 km2) of the Jintsu river where various phenomenon of sedimentationare occurring, observational studies have been carried out since 1966 in order to in-vestigate sediment phenomena.In this study, four problems are discussed;(1) The relation between sediment yield and rainfall, slope grade and etc. areexamined, to clarify actual condition of sediment yield in various slopes. (2) Therelation between sediment runoff and sediment storage in the steep channel withpools were investigated in the Hirudani (0. 85 km2). (3) The occurrence of debrisflow were investigated with attention forcused on rainfall intensity and in-filtration in the Shiramizudani. (4) The numerical simulation of the debris flow con-trol by crib dam was made, and compared with observed data in the Shiramizudani(2.1 km2)

Efficacies of the newly devised methods to control debris flow once under wayare discussed experimentally and numerically. These methods are retarding andstoring the debris flow by the series of alternately projecting groynes, stopping thedebris flow front by removing water from it through the grating layed on the bedor on the channel wall, and changing the flow regime from inertial to macro-viscousrange, I. E., slow down the velocity by adding a kind of chemicals.First, a numerical simulation method which is useful to discuss the behaviour ofdebris flow around such structures is presented. Then, the processes of retardation, deformation and stoppage of debris flow by the structures are examined by both theflume experiments and the numerical simulations.Second, the characteristics of macro-viscous flow produced by adding water ab-sorbent into debris flow are observed experimentally and a new flow equation to ex-plain those characteristics is presented. After that, possibility and the problems tobe cleared for the practical use are pointed out.

Turbid water with a prescrived volume flowing into fresh water, a turbidity currentlike thermals is formed, which we call "turbid water mass".We carried out theoretical study on the behaviour of turbid water mass, and proposedphysical and mathematical model for the wave height, density and advancing velocity of thehead. Then, the parameters introduced in the model were investigated by experiments.Using these values of the parameters, applicability of the theory on the behaviour of thehead was discussed.

Many disasters have happened due to occurrence of so-called sediment flow, which is in manytimes originated from erosion of the debris flow deposit upstream. To understand the phenomena ofsediment flows, the author hydraulically defines the sediment flow and studies the mechanics of it.He proposes the formulae of sediment discharge, resistance to flow, and sediment concentrationsuitable for the sediment flow. Application of these formulae for erosion of a debris cone and re-deposit as an alluvial fan necessitates a simplified method to analyse the process. The author givesthe method and calculates the initial process successfully. The process of armour coating and thecharacteristics of rhythmic spacing of pools and riffles are also investigated.

Spatial distribution of debris materials of debris-flow lobes was studied in relation to the deposi-tional process of debris flow at Kamikamihori Fan.At three lobes with different depositional conditions, several survey pits were dug in each lobe, to make the analysis of debris size of deposits on the ground surface and in the underground part.As for the formative processes of lobes at least of late years, it might be said that the swollen lobes inthe upper fan were the massive deposits of debris flow which had settled almost conserving the originalstructure of inverse grading in motion, while the flat lobes in the lower fan were the sandy deposits, almost having lost the original structure in motion, with scattered gravels lacking larger boulderswhich had settled along the running course of the debris flow.At the distal fan, thick strata by debris flows and thin strata by floods were found clearly in thealternation of the deposits, while in the upper and mid fan, the debris materials are assumed to havebeen supplied only by debris flow.

The authors carried out some observations on the runoff of rain, sediment yield and transportin a mountain watershed, called Ashiaraidani (7.2 Km2), which is a tributary of the Jintsu Riverin Gifu Prefecture, Japan.In this paper, the process and mechanism of sediment transport on the stream channel inHirudani experimental watershed were analyzed by petorological and grain-size classification oftransported gravels. Colored gravels were also used as the tracer to identify the characteristics ofsediment particle movements.The mechanism of occurrence and behavior of debris flow were partially revealed throughthe field observations on variation of river bed, and the water balance of debris flows in Ashiaraidaniexperimental watershed (6.5 Km2).