Local variations of landslide factors in Pulau Pinang, Malaysia

Abstract A landslide is one of the most notorious natural disasters, resulting in massive losses and significant damages. Thus, this paper aims to analyze spatial heterogeneity influencing factors which later inspect relationship between occurrences. total 988 landslides historical data eight were obtained from proper field validation maps provided by those concerned government, including distance roads, rivers, faults, slope angle, curvature, aspect, land use, lithology. Geographically Weighted Logistic Regression (GWLR) introduced carry out local analysis, angle aspect playing role landslides. The Akaike’s information criterion (AICc) GWLR 824.51 has a lower value than global regression represented as 906.09 revealing that best model. Other evaluation criteria such deviance, percent deviance explained (pdev), Bayesian (BIC) also validate significance results show degree variation study area coefficient values every factor are inconsistent, providing reference for managers formulate targeted decision-making measures. In meantime, urgent action sustain disaster suggested SDG 13 be taken earnestly avoid bigger impacts on both society environment.