Application of risk assessment models for seismic risk map

• Main Author: Abd. Razak, Janatul Aziera
• Format: Thesis
• Language: English
• Published: 2022
• Subjects: Q Science (General); TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/100351/1/JanatulAzieraAbdPMJIIT2022.pdf; http://eprints.utm.my/id/eprint/100351/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:150998

Seismic risk evaluation at a high potential area such as the Ranau district in Sabah is very important. However, the current method of seismic risk analysis through one of its parameters, seismic vulnerability, is mostly focused on assessing the physical damage of structures of the affected area. Thus, this research aimed to develop a simple and novel seismic risk model, specifically for Ranau, by combining two parameters; vulnerability and hazard through Geographic Information Systems (GIS)-based analysis. The model was developed from hybrid models previously used individually for disaster-related analysis. The hybrid models experimented with were the Frequency Ratio-Index of Entropy (FR-IoE), (FR-IoE) with Analytical Hierarchical Process (AHP), (FR-IoE) with Logistic Regression (LR) and (FR-IoE) with Naïve Bayes (NB). The seismic vulnerability results computed from these hybrid models were validated using the areas under the curve (AUC) of the relative operating characteristic (ROC). It was found that the NB model showed the lowest reliability with the AUC values of 0.640 and 0.741 for its success rate and prediction rate, respectively. The AUC values for other models' success rates are 0.853, 0.856, and 0.869 for FR-IoE, (FR-IoE) AHP, and (FR-IoE) LR, respectively, while their prediction rates are 0.863, 0.906, and 0.844 for FR-IoE, (FR-IoE) AHP and (FR-IoE) LR, respectively. A seismic hazard analysis was performed to complete the seismic risk model computation. The determination of seismic hazard was done by evaluating the Peak Ground Acceleration (PGA), which was derived from Ground Motion Prediction Equation (GMPE). From the PGA computation, a non-linear regression model with an accuracy of R2 = 0.997 was obtained from constraint Campbell (1981) fitted GMPE, which was the best-fitted model compared to the other 5 GMPEs tested. Finally, the novel fitted GMPE for Ranau using the Campbell (1981) fitted GMPE were integrated with the seismic vulnerability obtained from the hybrid (FR-IoE) AHP model to derive the seismic risk information for Ranau, Sabah in the form of a seismic risk map. In the long run, the computed seismic risk map obtained from these findings can be implemented for disaster preparedness and mitigation purposes and is useful the future earthquake disaster.