3:30 PM - 5:00 PM
[HDS10-P08] Investigating the determinants of collapse depth of landslide using air-borne laser survey data
With the intensification of rainfall in recent years, there has been an increase in the number of serious landslide disasters. In particular, sediment yielded by slope failure has the potential to damage people's living areas, and residual collapsed sediment on slopes can cause secondary disasters such as debris flow. Clarifying and prediction of the amount of sediment volume of landslide is one of the most basic and important elements for planning countermeasures.
Since the development of air-borne laser surveying technology has made, it possible to compare the topography before and after a disaster, making it possible to accurately estimate not only collapse area, but also collapse depth and the amount of collapsed sediment. It has been shown that the amount of collapsed sediment increases exponentially with the area of collapse (e.g. Larsen et al., 2010). On the other hand, although the collapse depth tends to increase with the collapse area, there are variations in the values, and the factors that determine the collapse depth are not fully understood (e.g. Akita et al., 2023). In this study, the influence of pre-disaster slope topography on collapse depth was examined and discussed using air-borne laser survey data.
The study area was the Noborigawa River basin located in Minami-uonuma City, Niigata Prefecture. A digital elevation model (DEM) was developed by the Yuzawa Sabo Office, Hokuriku Regional Development Bureau, Ministry of Land, Infrastructure, Transport and Tourism. The surveys were conducted on October to December 2009 and July to December 2010 for the pre-disaster period and August to September, 2011 for the post-disaster period.
ArcGIS Pro (ver. 3.0.3), Excel and Python were used for analysis. First, the collapse sites were interpreted from CS 3D maps (Toda, 2014), contour lines at 1 m intervals, difference maps, created using aerial photographs and 1 m mesh DEM. To focus only on the initiation zone of the landslides in this study, the area where sediment was considered to have been deposited from aerial photographs, etc., was not included in the landslide polygon. The analysis targets were selected from the deciphered landslides on the condition that they were based on igneous rocks, were new landslides and were located on slopes with a catchment area of less than 10000 m2.
The slope angle and profile curvature of the pre-disaster slope were then used to classify the landslides. The mean slope angle θ within the failure site was calculated and the landslide was classified according to 0°≦θ<30°, 30°≦θ<35°, 35°≦θ<40° and 40°≦θ<60°. The mean profile curvature K within the landslide was then calculated and the collapsed area was classified for each K<-0.5, -0.5≦K≦0.5 and 0.5<K.
For assessing the collapse depth characteristics, the 90 % value, the mean value and the 20 % value of the collapse depth were computed for each landslide.
The analysis was carried out on 147 decoded landslides. The mean collapse depth for all landslides was 2.14 m and the mean collapse area was 273 m2. Around 60% of all landslides had a collapse area of less than 200 m2. By mean slope angle, the collapsed areas and mean collapse depths of 415.46 m2 and 2.69 m, respectively, were the largest for landslides belonging to 30°≦θ<35°. By mean profile curvature, the collapsed areas and mean collapse depths of 399.56 m2 and 2.39 m, respectively, were the largest for collapsed areas belonging to -0.5≦K≦0.5.
The results show that both collapse area and depth show different trends depending on the pre-disaster slope topography. Therefore, it is suggested that both slope angle and profile curvature are factors that influence the determination of collapse depth.
Since the development of air-borne laser surveying technology has made, it possible to compare the topography before and after a disaster, making it possible to accurately estimate not only collapse area, but also collapse depth and the amount of collapsed sediment. It has been shown that the amount of collapsed sediment increases exponentially with the area of collapse (e.g. Larsen et al., 2010). On the other hand, although the collapse depth tends to increase with the collapse area, there are variations in the values, and the factors that determine the collapse depth are not fully understood (e.g. Akita et al., 2023). In this study, the influence of pre-disaster slope topography on collapse depth was examined and discussed using air-borne laser survey data.
The study area was the Noborigawa River basin located in Minami-uonuma City, Niigata Prefecture. A digital elevation model (DEM) was developed by the Yuzawa Sabo Office, Hokuriku Regional Development Bureau, Ministry of Land, Infrastructure, Transport and Tourism. The surveys were conducted on October to December 2009 and July to December 2010 for the pre-disaster period and August to September, 2011 for the post-disaster period.
ArcGIS Pro (ver. 3.0.3), Excel and Python were used for analysis. First, the collapse sites were interpreted from CS 3D maps (Toda, 2014), contour lines at 1 m intervals, difference maps, created using aerial photographs and 1 m mesh DEM. To focus only on the initiation zone of the landslides in this study, the area where sediment was considered to have been deposited from aerial photographs, etc., was not included in the landslide polygon. The analysis targets were selected from the deciphered landslides on the condition that they were based on igneous rocks, were new landslides and were located on slopes with a catchment area of less than 10000 m2.
The slope angle and profile curvature of the pre-disaster slope were then used to classify the landslides. The mean slope angle θ within the failure site was calculated and the landslide was classified according to 0°≦θ<30°, 30°≦θ<35°, 35°≦θ<40° and 40°≦θ<60°. The mean profile curvature K within the landslide was then calculated and the collapsed area was classified for each K<-0.5, -0.5≦K≦0.5 and 0.5<K.
For assessing the collapse depth characteristics, the 90 % value, the mean value and the 20 % value of the collapse depth were computed for each landslide.
The analysis was carried out on 147 decoded landslides. The mean collapse depth for all landslides was 2.14 m and the mean collapse area was 273 m2. Around 60% of all landslides had a collapse area of less than 200 m2. By mean slope angle, the collapsed areas and mean collapse depths of 415.46 m2 and 2.69 m, respectively, were the largest for landslides belonging to 30°≦θ<35°. By mean profile curvature, the collapsed areas and mean collapse depths of 399.56 m2 and 2.39 m, respectively, were the largest for collapsed areas belonging to -0.5≦K≦0.5.
The results show that both collapse area and depth show different trends depending on the pre-disaster slope topography. Therefore, it is suggested that both slope angle and profile curvature are factors that influence the determination of collapse depth.