Influence of Shannon’s entropy on landslide-causing parameters for vulnerability study and zonation—a case study in Sikkim, India

Abstract

Landslide is a common hazard in the hilly regions, which causes heavy losses to life and properties every year. Since 1980, various researches and analyses have been carried out in the geographic information systems (GIS) environment to identify factors responsible for causing landslides. The important conditioning factors identified by the researchers are slope, geological, geomorphologic structures, and land use coupled with triggering factors like rainfall and a few of the anthropogenic activities. Almost all landslides vulnerability studies carried out so far used parameters of landslide events of the past as essential inputs and advanced methods like information value, regression analysis, fuzzy logic, etc. The present research is an attempt to investigate the landslide vulnerabilities in different slope areas with simple and realistic method of assignments of weights to the parameters based on experts’ opinion and generic logic, without using the parameters of past landslide events as inputs. The identified factors were assigned appropriate weights based on experts’ opinion and these weights were further balanced with respect to the Shannon’s entropy of their occurrences within the study area. The study area was finally classified into three zones namely least vulnerable zone, moderately vulnerable zone, and most vulnerable zone. When compared with the actual landslide history of the past, it was found that Shannon’s entropy applied zonation model matched to real landslide events with higher value of landslide density as compared to the model developed without Shannon’s entropy.