A modified frequency ratio method for landslide susceptibility assessment

Abstract

The frequency ratio method is one of the most widely adopted methods for landslide susceptibility assessment. However, due to the obligatory classifications of landslide-related factors with continuous factor values, the conventional frequency ratio method is complicated by a discontinuity problem of the frequency ratio values and a subjectivity problem. This paper has modified the conventional frequency ratio method and developed a handy geographical information system extension that implements the modified method. Through calculating the frequency ratios for every “identical normalized factor value” instead of for every “factor class,” the modified method radically increased the continuity of frequency ratio values and reduced the subjectivity accompanied by the classifications of factors. An automatic and quick assessment of landslide susceptibility becomes possible because the calculations of frequency ratios for different factors in the modified method are constrained by only two uniform parameters (precision and bin width). Two case studies were adopted to inspect the performances of the modified method. From a quantitative point of view, the modified method derives landslide susceptibility models having slightly larger AUC values than the conventional method. From a qualitative point of view, the modified method gives much more detailed variations of frequency ratio with factor value and, as a result, can reveal characteristic fluctuations of frequency ratio and can smoothen the spatial discontinuity of the landslide susceptibility map derived by the conventional method. In practice, this modified frequency ratio method is expected to benefit the landslide susceptibility assessment and get further evaluations in the meantime.