Volume XLII-4/W16
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W16, 441–449, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W16-441-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W16, 441–449, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W16-441-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  01 Oct 2019

01 Oct 2019

SPATIAL PREDICTION MODELS FOR LANDSLIDE ACTIVITY MAPPING USING VEGETATION ANOMALIES

M. R. Mohd Salleh1, Z. Ismail1, S. A. Mohd Ariff1, M. Z. Abd Rahman1, M. F. Abdul Khanan1, M. A. Asmadi1, and K. A. Razak2 M. R. Mohd Salleh et al.
  • 1Department of Geoinformation, Faculty of Built Environment and Surveying, Universiti Teknologi Malaysia, Johor, Malaysia
  • 2UTM Razak School of Engineering and Advanced Technology, Universiti Teknologi Malaysia, Kuala Lumpur, Malaysia

Keywords: Tropical rain forest, Landslide activity, Vegetation anomalies, Remote Sensing

Abstract. An area that located in Kundasang which in Ranau district in Sabah, Malaysia that lies along the bank of Kundasang valley was chosen for comparing the reliability of frequency ratio (FR) and weight of evidence (WoE) methods for landslide activity probability mapping by using related vegetation anomalies indicator. The locations of 47 and 189 of active and dormant landslides respectively were identified using 4 raster layers (topographic openness, hillshade, colour composite and high resolution orthophoto). Each landslide activites were randomly divided into two groups as training (70%) and testing (30%) datasets. Tree height irregularities, DVI, NDVI, SAVI, and OSAVI were considered as landslide bio-indicator. The landslide activity probability maps were prepared using the FR and WoE method. The generated maps were validated by calculating the success and prediction rates from area under receiver operating characteristics (ROC) curve. The results of WoE method were relatively reliable (AUC > 0.8) for dormant landslide while only about 40% of active landslide have been predicted accurately. Similar trend yielded for FR method where least accuracy for active landslide prediction.