Landslide susceptibility mapping of Tegucigalpa, Honduras

Abstract

Landslides pose a significant threat to Tegucigalpa, Honduras, a city characterized by steep terrain, heavy rainfall, and rapid urban expansion. This study advances regional hazard management by developing a detailed landslide susceptibility map using a logistic regression model that integrates 12 morphometric predictors derived from a high-resolution 15-m Digital Elevation Model (DEM). Key predictors—such as slope, plan curvature (Planc), topographic wetness index (TWI), and relative slope position (RSP)—demonstrated strong explanatory power, with slope emerging as the dominant factor driving susceptibility (Estimate = 5.09, p < 0.001). Additional variables, including terrain roughness index (TRI) and flow accumulation (FlowAcc), enriched the model by accounting for surface irregularity and hydrological influences. The model achieved an Akaike Information Criterion (AIC) value of 744,092 and a moderate Area Under the Curve (AUC) of 0.65 based on the Receiver Operating Characteristic (ROC) curve, indicating reasonable predictive performance. AIC was preferred over measures like Spearman’s rank correlation as it better balances model fit and complexity, making it suitable for logistic regression. The susceptibility map classified 30 % of the metropolitan area as high or very high risk, with landslides primarily concentrated along steep escarpments and drainage networks. These findings highlight the compounded risks posed by natural and anthropogenic factors, including deforestation, urban development, and poor drainage systems. This research underscores the critical importance of integrating morphometric analysis into landslide risk assessments, providing actionable insights for mitigation planning in Tegucigalpa. By addressing key challenges faced by urban centers in developing countries, the study offers a transferable methodology for mapping landslide susceptibility globally and contributes to the broader discourse on disaster risk reduction in vulnerable regions. The study concludes that morphometric-based logistic modeling is an effective and transferable approach for improving landslide risk management in urban areas of the Global South.