Recent climate change caused by global warming has affected the environment in Korea, resulting in increased precipitation and frequency and magnitudes of typhoons. From 1994-2003, a mean of 2.3 days/year experienced heavy rainfall (≥80 mm), in contrast to a mean of 1.6 days/year from 1954-1963. Of the ten typhoons that have resulted in the maximum daily rainfall over the last ten decades, seven occurred between 1990 and 2007. The maximum instantaneous wind velocity of typhoons has greatly increased from 20 m/sec (in the 1970s) to 40 m/sec (in the 2000s). Furthermore, increases in the occurrence and scale of forest fires and landslides, as well as increased infrastructure and land use, contribute to sediment disasters. Accordingly, environmental changes and human-induced factors have resulted in significant increases in the magnitude and frequency of natural disasters, especially in the Gangwon region on the east coast of Korea. Therefore, in 2004 the Korea Forest Service instituted an integrated and environmentally-friendly system for forest management, which has helped prevent sediment disasters. Multiple prevention strategies are also required in addition to these systemic changes to forest structure management, including control of hill-slope erosion and torrent erosion, debris flow mitigation, water storage and slit dams, grade-stabilization structures and forest improvement, and watershed management.
The friction coefficients of debris flows over a rigid bed from several previous experiments were compiled in a preliminary investigation on the classification of phase transitions in debris flows. The collected friction coefficients were compared to the theoretical values of the friction coefficients in the relationship with the relative flow depth on the basis of sediment particle size (h/d) under various conditions. The friction coefficients of debris flows with h/d values less than 20 agreed closely with the theoretical value for boulder debris flows derived from the constitutive equations, while the friction coefficients with h/d values in the range 1000-10,000 agreed roughly with the theoretical value for turbulent water flows. The friction coefficients with h/d values of 30-300 exceeded the theoretical value for both debris and turbulent water flows. These intermediate debris flows were observed in experiments involving turbulent mud flows. However, a review of these experiments revealed that they may have included debris flows in which the turbulent structure was not well developed, and could be considered as debris flows in transition from laminar to turbulent flows. In some of the transitional debris flows, an interface dividing the flow structure into an upper turbulent-flow layer and a lower debris-flow layer was observed as reported for sediment-laden flows. The friction coefficient for transitional debris flows was modeled considering the shift of this interface. The model was able to explain the value for transitional debris flows, inferring that phase transition in debris flows from laminar to turbulent flows is induced by the shift of the interface.
Debris flows often cause substantial losses of human life as well as economic losses. Damage can be estimated using numerical simulation models that describe the debris flow process. Some models can be used to determine the possible effects of sabo dams and have been practically employed to plan sabo dam arrangement. However, the existing simulation systems currently do not have efficient user interfaces, making it difficult for non-experts in debris flow simulations to run simulations without the aid of specialists. We developed a system that produces one- and two-dimensional debris flow simulations and is equipped with a graphical user interface. The system is based on an integration model and employs one-dimensional simulations of gully areas and two-dimensional simulations of alluvial fan areas; it then considers mutual influences in boundary areas between gullies and alluvial fans. Data can be input using a mouse and viewed on the monitor, and users can see real-time visualized images of a debris flow during a simulation. The interface enables users to run a debris-flow simulation without expert knowledge of the model, enabling better solutions for sabo engineering.
Landslides caused by rain often take the form of earth flows when they reach their final stage in river valleys. Conventional steel grid- or slit-type sabo dams were originally developed for debris flows containing boulders at their leading edge and do not contain earth flows as effectively as they contain the flows for which they were designed. If the volume behind these types of dams is not quickly filled by a debris flow, subsequent earth flows pass through the dam, because these subsequent flows are composed of less coarse rock and gravel. Therefore, we designed and built several model sabo dams with closely-spaced horizontal units. We then conducted hydraulic experiments focusing on the effects of these closely-spaced horizontal units, with the objective of ensuring that they capture materials in the later phases of debris flows. Experimental results indicated that this new design of sabo dam is better at capturing finer sediments in both debris flows and earth flows.
Throughout Japan, almost all areas near hillsides or mountain slopes are threatened by landslides caused by heavy rainfall during the rainy and typhoon seasons. To mitigate potential disasters, many researchers have investigated landslides caused by heavy rainfall and have developed simulation models to predict landslide occurrence. However, because landslide mechanisms are complicated and involve many factors such as rainfall, surface and subsurface geomorphology, and soil physical properties, accurate prediction of landslides remains difficult by conventional simulation methods. Over the last several decades, many researchers have reported the existence of numerous preferential flow pathways for subsurface runoff, such as soil pipes or macro-pores in the soil layer. Simple infiltration analysis ignores those preferential flow pathways and therefore does not accurately simulate actual soil water flow, making it difficult to predict landslide occurrence accurately. For the last five years, we have investigated several landslide sites and have used experimental and modeling approaches to examine landslide occurrence. The results confirm that to improve precision of landslide prediction, simulation models should incorporate some important factors that affect landslide occurrence. This paper reports these issues as well as the results of the landslide investigations, and discusses the methods required to predict and mitigate landslide disasters.