Numerical Simulation Reflecting Buildings in Area Damaged by Debris flow

. More than 80% of average annual precipitation in South Korea occurs between June and September owing to heavy rainfall and typhoons in summer, and its land is vulnerable to mountain disasters (landslides and debris flow) as 63% of it is mountainous areas. In this study, an area damaged by debris flow in Wondeok-eup, Samcheok-si, Gangwon-do, Korea under the influence of Typhoon Mitag in 2019 was surveyed and numerical modeling was performed. Topographic data were created using the 5m grid DEM derived through the field survey data and GIS technique as well as the building data of the damaged area, and debris flow modeling was performed using the Hyper KANAKO model. A comparison with the inundation trace map showed that the simulation results based on topographic data that reflected buildings exhibited similar flow patterns and characteristics to the actual damage.


Introduction
Global climate change is causing larger natural disasters. South Korea is vulnerable to mountain disasters, such as landslides and debris flow, due to heavy rainfall and typhoon-induced rainfall in summer each year because mountainous areas represent 63% of its land. Representative damage cases include the mountain disasters (landslides and debris flow) caused by Typhoon Rusa in 2002, Typhoon Maemi in 2003, localized heavy rainfall in 2011, and Typhoon Mitag in 2019, which caused human casualties and property damage. Studies have been conducted worldwide to predict and reproduce areas vulnerable to mountain disasters using GIS and numerical models to reduce such damage.  [8,9,10], and RAMMS(WSL-SLF). In this study, topographic data that reflected buildings in an area damaged by debris flow were created using the Hyper KANAKO model, a numerical model developed in Japan, and the flow characteristics of debris flow were analyzed by applying the data to the model.  [6]. It is based on the governing equations of the Takahashi (Takahashi et al., 2001;Takahashi, 2007) [11,12] model to analyze debris flow.
The model uses eq. 1 that obtains the erosion/deposition rate of debris flow by calculating temporal and spatial (x and y directions) changes in flow depth for the total volume of debris flow, and the continuity equation of eq. 2 that obtains the erosion/deposition rate of debris flow by calculating temporal and spatial (x and y directions) changes in flow depth according to the volume concentration of the deposited sediment.
where i is erosion/deposition velocity(if i < 0, deposition and if i >= 0, erosion), h is flow depth, t is time, u is xaxis flow velocity, v is y-axis direction flow velocity, C is sediment concentration by volume in the debris flow, C* is sediment concentration by volume in the movable bed layer.
The momentum equations of debris flow are expressed by eq. 3 and eq. 4. Spatial changes in flow velocity in the x and y directions can be expressed by the values obtained by subtracting the resistance force at the boundary from the driving force of debris flow. In addition, the riverbed change equation is given by eq. 5.
where is acceleration due to gravity, H is flow elevation (H = h + z), z is bed elevation, ρ is density of the fluid, τx and τy are riverbed shearing stresses in the x-and y-axis directions, respectively.

Application of Numerical Model
In this study, a 5m-resolution DEM was created from the 1:5,000 digital topographic map provided by the National Geographic Information Institute using ArcGIS to analyze the flow characteristics of debris flow, and it was superimposed onto building information. The binary conversion software included in the model was used to convert it into the topographic data required by the model. Total discharge, peak discharge, and sediment concentration, which are the input parameters of the model, were calculated using the empirical formulas presented by the National Institute for Land and Infrastructure Management (NILIM, 2016) [13] in Japan.

Target area
Galnam-ri, Wondeok-eup, Samcheok-si, Gangwon-do is an area close to the East Sea of Korea. It has a high terrain to the west and low terrain to the east as there is a high mountain range to the west. The watershed has an area of 0.84 km 2 , a maximum elevation of 316 m, a minimum elevation of 11.6 m, an average elevation of 163 m, and an average slope of 17˚. A debris flow disaster occurred in the area due to rainfall of up to 110 mm/hr under the influence of Typhoon Mitag in October 2019. 55 houses were buried or inundated and resulted in 111 victims in the downstream area( figure 1).   Fig. 1. Location of the study area.

Topographic data
For the topographic data of the area damaged by debris flow, a 5m grid DEM was created using the contour lines of the 1:5,000 digital topographic map. Since building information was in the Shp file format, it was converted into Raster in the same 5m grid size as DEM and superimposed upon DEM(figure 2). The creation of topographic data through a digital topographic map involves difficulty in implementing the characteristics of the coast, drainage channels, and covered rivers. Therefore, it is deemed necessary to create topographic data using measuring equipment such as drones and LiDAR for more accurate modeling.

Input parameter
To perform numerical simulation on debris flow, information on input parameters, i.e., sediment yield, sediment concentration, total discharge, and peak discharge, is required. The sediment yield, the width, depth, and length of the valley (Table 1) were measured through a field survey and the equation proposed by Ikeye (1981) [14] (eq. 6) was used. The sediment yield was applied to the empirical formulas presented by NILIM to calculate the sediment concentration (eq. 7), total discharge (eq. 8), and peak discharge (eq. 9). The values in a previous study were entered for the unit weight and internal friction angle of soil and the unit weight of water (Nakase, 2004) [15]. Table 2 shows the input parameters for the Hyper KANAKO model.      figure 5).

Conclusions
In this study, numerical simulation of debris flow was performed using the Hyper KANAKO model among debris flow numerical models, and flow characteristics in an area damaged by debris flow were analyzed under two topographic conditions. As for the input parameters of the model, sediment yield, peak discharge, and sediment concentration were used based on the field survey data. When the flow of the actual debris flow was compared with the damage estimated from the inundation trace map, it was found that topographic data that reflected buildings showed similar flow patterns and maximum flow depths. In future research, it is necessary to determine the validity of the results by creating precise topographic data and calculating input parameters using LiDAR and drones, and the applicability of the modeling results will be presented by simulating several damaged areas.