UTILIZATION OF AIRBORNE LIDAR DATA FOR LANDSLIDE MAPPING IN FORESTED TERRAIN: STATUS AND CHALLENGES K.A. Razak 1,2, *, M.W. Straatsma 1 , C.J.Van Westen 1 , J.P. Malet 3 1 International Institute for Geo-Information Science and Earth Observation, ITC P.O. Box 6, 7500 AA Enschede, The Netherlands – razak@itc.nl; straatsma@itc.nl; westen@itc.nl 2 Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia 3 CNRS – University of Strasbourg, School and Observatory of Earth Sciences, Strasbourg, France - jeanphilippe.malet@eost.u-strasbg.fr ABSTRACT: With the advancement in sensor technology and point clouds processing algorithms, the capability of Airborne LIDAR data for landslide mapping in forested terrain is much promising. Most of the available air-based platform products either from an aerial-photograph, satellite imagery or synthetic aperture radar are not appropriate for interpretation of landslide features which are covered by dense vegetation. Mapping the landslides in such environment requires additional information that can help the interpreters to recognise landslide characteristics and later on be used to distinguish landslide activities. In this research, we present the status and challenges of utilizing the high-resolution airborne LIDAR data for landslide mapping in forested terrain. European dataset in Barcelonnette, Southern French Alps has been used. Amongst the important issues in this research is bare earth extraction. In order to preserve the important geomorphological feature, bare earth extraction based on hierarchical robust filtering is presented with a suitable filter parameterization. Finally, high-resolution Digital Terrain Model (DTM) is used to calculate the local topography roughness and characterize the landslide morphology. KEY WORDS: Airborne LIDAR, bare earth extraction, landslide mapping * Corresponding author. 1. INTRODUCTION 1.1 Research Motivation The frequency and destructive capacity of natural disasters are increasing around the world; landslides in particular are becoming important geomorphic agents which shape the surface of the earth. Evidence of these changes in the forms of mountainous areas, hills, valleys, rivers and streams, combined with preparatory and triggering factors on unstable terrain, shows that landslides can shift to be seen as a natural disaster. Moreover, landslides can attract widespread attention when there are many casualties and properties are damaged. It is important that these destructive events are well-documented and predicted in order to decrease the number of the casualties and major damages in different geographic and socioeconomic regions. In order to better understand the future behavior of these events, hazard assessment and knowledge on past events are essential. A comprehensive landslide mapping is a basic requirement and important for quantifying both landslide hazard and risk. For sparsely vegetated areas, there are many mapping techniques available for making a landslide inventory map. For instance, through field- based observation of geomorphology (Rupke et al. 1988; Weber and Associates, 1990; Wills et al., 2002), image interpretation on single aerial photograph (Donati and Turrini, 2002; He at al. 2003) or multi-temporal aerial photographs (Kaab, 2002); (Van Westen and Lulie Getahun, 2003) and of high resolution satellite images (Gupta and Saha, 2001; Lin et al. 2002) and satellite- based or ground-based synthetic aperture radar (Singhroy et al., 1998) is possible. However, these mapping techniques are labour intensive, time-consuming and are not suitable for the rugged terrain. These methods were unsuccessful in areas that are covered by dense vegetation or that is rapidly re-vegetated. The occurrence of landslides in these environments requires new mapping techniques and Airborne Light Detection and Ranging (LIDAR) is a promising technique for such applications In this paper, current landslide mapping techniques that has been used in forested terrain will be elaborated. Special attention on status and challenges of airborne LIDAR data is given. High-resolution airborne LIDAR data is utilized in order to produce a complete bare earth that preserves important landslide features. Hence, a local topography roughness that calculated based on LIDAR- derived DTM is also presented.