Applications of multi-scale remote sensing data to determine potential landslides in the Laonong watershed areas

Since its opening to traffic, the Southern Cross-Island Highway has been playing an important role in linking Southern and Eastern cities in Taiwan. Nevertheless, Southern Cross-Island Highway was firstly struck by the 1999 Ji-ji Earthquake, which resulted in collapses along the route, and then Typhoon Morakot, which caused damages of 22 bridges and a number of deep-seated landslides between Meishan of Kaohsiung and Siangyang on the East. A long-term road reconstruction and improvement project of Southern Cross-Island Highway was therefore initiated in 2009. In August 2021, the torrential rain triggered a deep-seated landslide located in the upstream of Yusui Stream and an enormous amount of debris was brought to the downstream and crashed Minbaklu Bridge between Meishan and Siangyang. Although the Directorate General of Highways cleared the route and built a temporary steel bridge for people to pass through, this incident has highlighted the fact that the road breaks when plum rain or torrential rain occurs. This project has adopted optical satellite imagery, aerial LiDAR and UAV, technologies that complement each other with their respective benefits and drawbacks. Aerial LiDAR can remove vegetation and present the real ground surface, enabling researchers to calculate the volume of landslide materials of Yusui Stream and Putanpunas Stream using LiDAR derived DTM (digital terrain model) constructed based on the images collected between 2016 and 2022. The three-dimensional terrain interpretation and landslide volume calculation results reveal that the landslide surface area had been continuously increased over the last six years due to abundant rainfall brought by typhoons and torrential rain, causing an enormous volume of debris falling into the main river channel and piled up at its confluence with Laonong River. Nevertheless, the interpretation can be hard in areas with small-scale shallow landslide due to smaller changes to the surface elevation. Optical satellite imagery before and after the sliding is therefore required to quantify the change of landslide volume, helping to determine potential landslide and accumulation areas even more effectively.