A Study of Macroseismic Intensity Prediction using Regression of Philippine Earthquakes

Earthquake is the second most common natural disaster occurrence in the Philippines over the past years that poses the greatest threat to Metropolitan Manila that could considerably affect the nation's economy. In this study, a predictive relationship between the macroseismic intensity, ground motion parameters, and earthquake event characteristics for the Philippine region was derived as a basis for a pro-active approach in earthquake preparedness response planning for a more disaster-resilient country. A 40 light-to-large earthquake events in the Philippines had been collected between 2018 and 2020 from the USGS Earthquake Hazard Program and data pre-processing produced a total of 198 unique earthquake records which had been used in the model development. In multiple linear regression, it was observed that the coefficient of determination (R-squared) measure of the intensity relationship yielded 87.5 percent for Peak Ground Acceleration (PGA) and 87.1 percent for Peak Ground Velocity (PGV). Meanwhile, using random forest ensemble technique, an increase in the prediction accuracy of the model was observed with an R-squared score equal to 92.2 percent for PGA and 93.6 percent for PGV. Furthermore, the regression models presented and performed in this study proved to be a suitable method for macroseismic intensity prediction for the Philippines even with limited available earthquake event data and could be a basis for earthquake preparedness response planning for a more disaster-resilient country. Moreover, as suggested in previous studies, macroseismic intensity prediction should not be adopted in other regions due to regional dependencies characteristics of an earthquake that vary in different countries.