The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLII-3/W8
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-3/W8, 25–30, 2019
https://doi.org/10.5194/isprs-archives-XLII-3-W8-25-2019
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-3/W8, 25–30, 2019
https://doi.org/10.5194/isprs-archives-XLII-3-W8-25-2019

  20 Aug 2019

20 Aug 2019

UTILIZATION OF FINE RESOLUTION SATELLITE DATA FOR LANDSLIDE SUSCEPTIBILITY MODELLING: A CASE STUDY OF KASHMIR EARTHQUAKE INDUCED LANDSLIDES

M. Z. Ali1, H.-J. Chu1, S. Ullah2, M. Shafique3, and A. Ali2 M. Z. Ali et al.
  • 1National Cheng Kung University, Tainan, Taiwan
  • 2Institute of Space Technology, Islamabad, Pakistan
  • 3National Centre of Excellence in Geology, University of Peshawar, Peshawar, Pakistan

Keywords: Landslide Inventory, Support Vector Machine, Influential Factors, Logistic regression, Landslide Susceptibility map

Abstract. The 2005 Kashmir earthquake has triggered thousands of landslides which devastated most of the livelihood and other infrastructure in the area. Landslide inventory and subsequently landslide susceptibility mapping is one of the main prerequisite for taking mitigation measure against landslide effects. This study has focused on developing most updated and realistic landslide inventory and Susceptibility mapping. The high resolution data of Worldveiw-2 having spatial resolution of 0.4 m is used for landslide inventory. Support Vector Machine (SVM) classifier was used for landslide inventory developing. Total 51460 number of landslides were classified using semi-automatic technique with covering area of 265 Km2, smallest landslide mapped is covering area of 2.01 m2 and the maximum covered area of single landslide is 3.01 Km2. Nine influential causative factors are used for landslide susceptibility mapping. Those causative factors include slope, aspect, profile curvature, elevation, distance from fault lines, distance from streams and geology. Logistic regression model was used for the Landslides susceptibility modelling. From model the highest coefficient was assigned to geology which shows that the geology has higher influence in the area. For landslide susceptibility mapping the 70 % of the data was used and 30% is used for the validation of the model. The prediction accuracy of the model in this study is 92 % using validation data. This landslide susceptibility map can be used for land use planning and also for the mitigation measure during any disaster.