Identification of rockfall source areas using the seed cell concept and bivariate susceptibility modelling

• DOI: 10.1007/s10064-021-02441-w
• Publication Date: 2021-09-16T16:06:39Z
• Authors: Aleksandar Toševski; Davor Pollak; Dario Perković
• Keywords: Rockfall; Orthophoto; Lithology; Land Cover

Abstract The objective of this research was to prepare a rockfall susceptibility map. Explorations were conducted in the Dubračina River basin (Croatia). input data included geological map, an orthophoto and 1-m digital terrain model (DTM). After talus inventory prepared, seed cell concept applied define source areas. contributing factors (predictors) rockfalls evaluated by chi-squared test. analysis confirmed following predictors: CORINE land cover, lithology, slope, aspect, distance from spring, road, fault, stream, rock-soil boundary. A matrix pairwise comparison predictor ratings used most significant factors. predictors that affected map share 86.3% slope (61.6%), lithology (13.4%), cover (6.2%), boundary (5.1%). Two maps prepared: one using all nine another four showed both good, with same areas under receiver operating characteristic (ROC) curves. prepared only can be considered better due its more precise spatial definition critical It concluded DTM provide enough reliable application bivariate statistical zonation method called “frequency ratio method” proven successful. This demonstrates useful speed up process area detections large regions.