Collection and analysis of geospatial information of sediment-related disaster (1/3)

ABSTRACT

Since the geography of Taiwan is unique and the island of Taiwan is located in a region where natural disasters such as earthquake, typhoon, and torrential rains occur frequently, actively collecting remote sensing data as well as geospatial information and effectively applying them to disaster prevention and rescue system to protect lives and properties are significant purposes of disaster prevention. The project of this year simultaneously integrates disaster-related data collection into the application of satellite images to carry out post-disaster landslide interpretations and complete the report of post-disaster image interpretations in terms of five natural disasters, including the earthquake on May 16, extremely heavy rains on June 13, Typhoon Nesat as well as Typhoon Haitang on July 28, and torrential rains on October 11, by using Landsat-8, SPOT-6, SPOT-7, and Sentinel-2 during the emergency operation period of the SWCB. Additionally, UAVs were operated to take aerial photos of primary soil and water conservation area to create digital elevation models (DEMs) and orthophotos of at least ten areas, such as Hungye Village located in Yanping Township, Taitung County.

In addition to continuously maintaining and updating relevant graphic information in the Soil and Sand Disaster Spatial Information System, this project also builds the back-end database with the aim of improving the efficiency of looking up information in the system. To integrate the great amount of geospatial information acquired by UAVs and quickly synthesize the results of aerial photos, this project invents a system for updating images from UAVs, and provides the function of website management and information inquiry system.

Furthermore, based on three designated townships with potential debris flow hazard torrents, including Sandimen Township, Maolin Township, and Liouguei Township, the project of this year completed the analysis and the verification of landslides caused by Typhoon Matmo in 2014 and Typhoon Soudelor in 2015. The results have verified that the accuracy rate of the analysis is between 77 to 81 percent. This project also applies the precipitation forecast of the Central Weather Bureau (CWB) and produces landslide hazard index (LHI) as references of landslide early warning for the SWCB. According to historical landslide data, this project simultaneously analyzes historical landslide situations of 1705 potential debris flow torrents in watershed areas. At last, the promotion of research results is included in this project—three promotional activities have been held, one research paper for an international conference and one journal paper have been published, and the posters have been made to increase international recognition of this work.

Keywords:sediment-related disasters, Geospatial Information System (GIS), satellite imagery, Unmanned Aerial Vehicle (UAV), landslide hazard analysis