Abstract
For better estimation of hazardous areas affected by different types of simultaneous slope failures, this paper proposes an algorithm for improving representativeness of training data sets (i.e., past occurrences of slope failures) for producing susceptibility maps of slope failures. The spatial quantitative models generally elucidate the relationship between training data sets and causal factors (e.g. topography, surface geology, soil, vegetation, slope, aspect, drainage, etc.). The proposed algorithm resamples the matched pixels between original groups of past slope failures (i.e., surface slope failures, deep-seated slope failures, landslides) and classified three groups by cluster analysis for all pixels corresponding to those slope failures. For all cases of three types of slope failures, the improvement of success rates with respect to resampled training data sets was confirmed. Furthermore, the differences between produced susceptibility maps by using original and resampled training data sets are delineated on difference maps which are effective in evaluating hazardous areas affected by different types of simultaneous slope failures.
