Integrating GIS Based Fuzzy Set Theory in Multicriteria Evaluation Methods for Landslide Susceptibility Mapping

Abstract

In this study we employ two GIS Multicriteria Decision Analysis (MCDA) methods for Landslide Susceptibility Mapping (LSM): Analytical Hierarchy Process (AHP) and Fuzzy Analytical Hierarchy Process (FAHP). First, a classic AHP method is used to carry out a LSM. AHP applies crisp pairwise comparisons and standardization techniques. In a second stage, fuzzy set theory is integrated in the AHP process, both for weighting the criteria and for the criteria standardization. The Triangular Fuzzy Numbers (TFNs) method is used to compare and reduce the error that is inherent in criteria ranking. This is achieved with Fuzzy Membership Functions and by assessing all criteria individually. For both approaches landslide susceptibility maps are produced and results are validated based on information on existing landslides. A particular measure used in the validation exercise is the ‘Receiver Operating Characteristics’curve. Results of the validation exercise with known landslides indicate that the FAHP clearly performs best. The results of this study help to identify strategies of choosing an appropriate method for LSM studies and, respectively, to assess the reliability of the resulting maps.