Using Artificial Neural Networks to study landslides and Rockfall Hazards in Maswar, Amran District
This study aims to a Spatial develop an Artificial Neural Network (ANN)-based model for predicting landslide and rockfall hazards and to produce a susceptibility map for these phenomena in Maswar District, Amran Governorate, Yemen, utilizing Geographic Information Systems (GIS) techniques. Twelve geological, geomorphological, hydrological, climatic, environmental, and anthropogenic factors influencing hazard occurrence were evaluated. Spatial datasets were collected, preprocessed, and standardized into a uniform digital format for model building. The dataset, comprising historical records and field observations of landslide and rockfall events, was partitioned into training, validation, and testing subsets. The ANN model incorporated two hidden layers to maximize predictive performance, and its outputs were exported to the ArcGIS environment for spatial susceptibility mapping.
The resulting map classifies the study area into five susceptibility levels: very low, low, moderate, high, and very high. Approximately (30.44%) of the district falls within the high to very high susceptibility categories. Model validation yielded an Area Under the Curve (AUC) of (0.68) for the Receiver Operating Characteristic (ROC) curve, indicating satisfactory predictive capability. The model achieved an overall accuracy of (0.72), a precision of (0.714), and a recall of (0.50), demonstrating operational effectiveness. The generated susceptibility map exhibits strong spatial correspondence with previously published hazard maps.
The findings indicate that the produced map provides reliable hazard modeling. This research underscores the value of integrating machine learning with GIS to support spatial planning and natural hazard risk management.
Metrics
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.