Proposal of a method for wide‐area assessment of the landslide volume and the volume of driftwood generated immediately after a mudslides
Hiromi AKITA
Abstract
This study reorganized cases of estimated area of mudslide scars that have been extracted and published since 2020, and considered the setting of the ΔNDVI threshold and the setting of the minimum extraction area to be excluded during processing. Based on this, the aim was to develop a method to calculate the estimated volume of sediment and driftwood from the area of the estimated mudslide scars. When the extraction results were examined for ΔNDVI threshold cases from 0.1 to 0.5 in increments of 0.05, the lower the ΔNDVI threshold was set, the larger the extraction area became. The average ΔNDVI threshold value for the case with the maximum F measure could be calculated for each satellite data, and it was considered that 0.19 for Sentine l‐2 and 0.23 for Planet were applicable. When the minimum extraction area for each satellite data was sorted, Sentine l‐2 thresholds of 0.25 and 0.2 were able to extract actual sediment runoff of 49m2, and 0.15 was able to extract 42m2, while Planet thresholds of 0.25 and 0.2 were able to extract 23m2, and 0.15 was able to extract 5m2. Although the data was somewhat insufficient this time, it was estimated that it was possible to extract at least the size above the spatial resolution. Finally, the estimated volume of sediment and driftwood from the landslide area calculated by this research method was converted to per unit area, and the relationship with the standardized 48‐hour accumulated rainfall was arranged. Overall, as the 48‐hour accumulated rainfall increases, the estimated volume of sediment and driftwood per unit area also increases.
Key words
disaster response, information transmission, optical satellite data, change detection, mudslide scar