11:45 AM - 12:00 PM
[SVC26-11] Measurement of topographic changes during the 1977-82 Usu volcanic activity using SfM-MVS analysis
Keywords:Usu volcano, SfM-MVS analysis, topographic changes, digital surface model, 1977-82 Usu volcanic activity
The 1977-82 Usu volcanic activity caused significant topographic changes around the summit and northern foot of Usu Volcano (Okada et al., 1981; Yokoyama et al., 1981; Katsui et al., 1985). The displacement around the summit has been studied in detail based on topographic interpretation using aerial photographs and 1/2,500-1/5,000 topographic maps. On the other hand, displacement around the northern foot of the fault has been investigated mainly by sidelength surveys and field fault displacement measurements, and spatial distribution studies have been limited to schematic ones (Moriya et al., 1982; Imagawa, 1984). In addition, the Higashimaruyama fault (Yamagishi et al., 1982; Kadomura et al., 1983), which shows horizontal displacement of about 60-80 m, occurred at the northeastern foot of Usu Volcano, but this fault displacement is not fully reflected in these research results.
SfM-MVS analysis was performed using aerial photographs taken during five periods from 1974 to 1985, and the spatial distribution of ground displacement was measured over the entire Usu volcano. Metashape 2.0 was used for SfM-MVS analysis. By aligning photos from all five periods together and obtaining ground reference and verification points, errors due to differences in the timing of photography were minimized. We acquired 17 ground control points and 7 verification points, and were careful not to acquire these points around the summit and northern foot of Usu Volcano, where large displacements had occurred. A point cloud, digital surface model (DSM), and orthomosaic were created for each period. The positional accuracy of the verification point was determined using the position in the orthogonal image obtained from the latest aerial photograph of the GSI maps as the true value. The maximum and minimum values of RMSE of the residual for each period were 0.48-1.15m for X, 0.47-0.96m for Y, and 0.81-2.28m for Z. DSM difference maps for each period were created and the amount of vertical displacement of the ground surface was determined. Topographic decipherment from the DSM and orthomosaic was used to determine the spatial distribution of the amount and direction of horizontal displacement of the land surface during (1) 1974-85 (entire period), (2) 1974-77 (early period), (3) 1977-79 (middle period), and (4) 1979-85 (late period).
In terms of horizontal displacement of the ground surface, displacement of the northeast-east foot was widely recognized during the early stage (②), but during the middle-late stage (③ and ④), the amount of displacement of the east foot became smaller with the formation of the East Maruyama Fault, and the displacement of the northeast foot became predominant. The amount of displacement at the northeast foot is outstanding. Fault displacement was sharpest in the area indicated by Kadomura et al. (1983), and its extension, the southwestern end, was continuous to the summit of the mountain to the Oo-Usu, and the northeastern end to the lake shore. Based on the spatial distribution of displacement, it is considered that a U-shaped block with a northeast opening at the summit of the mountain extended to the northeast foot slope with a width of about 1 km, and its end reached the lakeshore and below the lake surface. These findings are in harmony with the findings from field surveys and observations in previous studies, and this method is considered useful for investigating and analyzing the spatial distribution of land surface displacement before the spread of interferometric SAR analysis and aerial laser surveying (approx. 1940s-90s).
SfM-MVS analysis was performed using aerial photographs taken during five periods from 1974 to 1985, and the spatial distribution of ground displacement was measured over the entire Usu volcano. Metashape 2.0 was used for SfM-MVS analysis. By aligning photos from all five periods together and obtaining ground reference and verification points, errors due to differences in the timing of photography were minimized. We acquired 17 ground control points and 7 verification points, and were careful not to acquire these points around the summit and northern foot of Usu Volcano, where large displacements had occurred. A point cloud, digital surface model (DSM), and orthomosaic were created for each period. The positional accuracy of the verification point was determined using the position in the orthogonal image obtained from the latest aerial photograph of the GSI maps as the true value. The maximum and minimum values of RMSE of the residual for each period were 0.48-1.15m for X, 0.47-0.96m for Y, and 0.81-2.28m for Z. DSM difference maps for each period were created and the amount of vertical displacement of the ground surface was determined. Topographic decipherment from the DSM and orthomosaic was used to determine the spatial distribution of the amount and direction of horizontal displacement of the land surface during (1) 1974-85 (entire period), (2) 1974-77 (early period), (3) 1977-79 (middle period), and (4) 1979-85 (late period).
In terms of horizontal displacement of the ground surface, displacement of the northeast-east foot was widely recognized during the early stage (②), but during the middle-late stage (③ and ④), the amount of displacement of the east foot became smaller with the formation of the East Maruyama Fault, and the displacement of the northeast foot became predominant. The amount of displacement at the northeast foot is outstanding. Fault displacement was sharpest in the area indicated by Kadomura et al. (1983), and its extension, the southwestern end, was continuous to the summit of the mountain to the Oo-Usu, and the northeastern end to the lake shore. Based on the spatial distribution of displacement, it is considered that a U-shaped block with a northeast opening at the summit of the mountain extended to the northeast foot slope with a width of about 1 km, and its end reached the lakeshore and below the lake surface. These findings are in harmony with the findings from field surveys and observations in previous studies, and this method is considered useful for investigating and analyzing the spatial distribution of land surface displacement before the spread of interferometric SAR analysis and aerial laser surveying (approx. 1940s-90s).