The performance of landslides frequency-area distribution analyses using a newly developed fully automatic tool

Abstract

Frequency-Area Distribution (FAD) analyses were introduced to landslides research since the early 2000’s. This technique is a powerful tool that allows assessing the completeness of landslide inventory maps (LIM), used to build both landslides susceptibility and landslides hazard assessment models. However, FAD analyses are not commonly used in such studies despite the significant potential of the technique. The long processing steps needed to generate FAD curves, which involve logarithmic binning and iterative model fitting using various statistical tools, constitutes an energy and time-consuming task that pushes many researchers away from using the technique. In fact, no fully automatic tool capable of generating FAD curves and models exists as of July 2023. Therefore, we attempt to provide a fully automatic computer program capable of binning, fitting FAD curves and assessing their goodness of fit to theoretical models in a fully automatic, one step process. An example is provided using real data from Taounate province, Northern Morocco, so as to demonstrate the ability of the tool to deal with exhaustive datasets. In addition, the Kolmogorov-Smirnov, goodness of fit test is added to provide an objective assessment of the data fitting, which constitutes a better alternative to the subjective visual assessment that most landslides researchers rely on. To sum up, we believe that this tool will help popularize the FAD technique, which will consequently improve the accuracy and objectivity of landslides risk and hazard assessment disciplines.