A bivariate statistical technique with knowledge-based analytical hierarchy process for landslide susceptibility assessment in naryn river basin, Kyrgyzstan

Abstract

The model used statistical analysis for eight causal factors (slope, curvature, elevation, TWI, NDVI, NDSI, land cover, precipitation) of landslide occurrence. All factors were weighted to apply two-dimensional statistical method of a knowledge-based analytic hierarchical process with data extracted from the spatial database and then converted into a map. Final susceptibility maps showed a close agreement between the two models. The models predicted 72.1% and 69% of the empirical data used for the analysis respectively. These maps can be used to demonstrate the effectiveness of two-dimensional statistical model through the relationship between each factor with a resultant landslide susceptibility. The proposed model can be used to reproduce the relationship between each conditional factor without having to resort to multivariate statistics. The models are a powerful tool for assessing natural hazards, and to produce landslide probability maps for a better definition of risk zones.