Abstract
Abstract. In August 2009, Typhoon Morakot caused a record-breaking rainfall in Taiwan. The heavy rainfall triggered thousands of landslides, in particular in the central-southern part of the island. Large landslides can block rivers and can lead to the formation of landslide-dammed lakes. Cascading hazards like floods and debris flows after landslide dam breaches pose a high risk for people and infrastructure downstream. Thus, better knowledge about landslides that significantly impact the channel system and about the resulting landslide-dammed lakes are key elements for assessing the direct and indirect hazards caused by the moving mass. The main objectives of this study are 1) to develop an object-based image analysis (OBIA) approach for semi-automated detection of landslides that caused the formation of landslide-dammed lakes and 2) to monitor the evolution of landslide-dammed lakes based on Landsat imagery. For landslide and lake mapping, primarily spectral indices and contextual information were used. By integrating morphological and hydrological parameters derived from a digital elevation model (DEM) into the OBIA framework, we automatically identified landslide-dammed lakes, and the landslides that likely caused the formation of those lakes, due to the input of large amounts of debris into the channel system. The proposed approach can be adapted to other remote sensing platforms and can be used to monitor the evolution of landslide-dammed lakes and triggering landslides at regional scale after typhoon and heavy rainstorm events within an efficient time range after suitable remote sensing data has been provided.
Highlights
1.1 BackgroundLandslides play an important role in the evolution of landscapes (Guzzetti et al, 2012)
We compared the results to a reference dataset, created based on literature and visual interpretation, showing the location of dammed lakes and triggering landslides
Feature extraction relied on: a) surface parameters derived from a DEM, such as slope and curvature, b) hydrological parameters derived from a DEM, such as flow accumulation and basin, c) spectral indices, such as NDVI, NDWI and brightness(calculated separately for each year), and d) spectral information from Landsat 5 imagery for three successive years (2009/10/30, 2010/12/20, and 2011/02/06)
Summary
Landslides play an important role in the evolution of landscapes (Guzzetti et al, 2012). They are a significant and frequent hazard in many areas in the world, including Taiwan. The Xiaolin landslide, a large deep-seated landslide in southern-central Taiwan, which dammed the river leading to the formation of a landslidedammed lake (Wu et al, 2014). Earth Observation (EO) data provides a unique possibility to detect and monitor landslides and landslide-dammed lakes from space (Cigna, 2018; Delaney and Evans, 2015). The main objectives of this study are: 1) to semi-automatically detect landslides that caused landslide-dammed lakes, using object-based image analysis (OBIA), and 2) to monitor the evolution of landslide-dammed lakes using EO imagery
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
