The island of Java in Indonesia is the most populated island in the world, with over 150 million
residents. Java is located in a complex tectonic setting and frequently experiences damaging earthquakes. In Java, many buildings are non-engineered and built with limited or no seismic design provisions. This has led to poor quality construction and materials being used, particularly in residential houses, with the housing sector often sustaining the highest economic losses in Indonesia. Despite this high hazard, high exposure, and high vulnerability there is limited seismic risk information available for Java. This is in part due to poor data availability to develop exposure and vulnerability models to verify them. This study has focussed on developing a village scale exposure model for Java from multiple sources and understanding the uncertainty in this process. This new exposure model harmonises multiple eras of census data, government reports, the
Global Human Settlement building height model, OpenStreetMap, post-disaster surveys and other published literature to develop an exposure model that includes residential, commercial, and industrial buildings, and schools. The model shows spatial variation in the seismic resilience of buildings based on the relative wealth, region (rural, urban or city) and population density. Model performance was then assessed through scenario damage analysis of several recent earthquakes in Java.
A reliable exposure model is a critical input to a Probabilistic Seismic Risk Assessment (PSRA) that would help Indonesia mitigate the threat earthquakes pose to communities and infrastructure. However, the scenario damage analyses highlighted two key issues that must be addressed before conducting a PSRA. Firstly, it identified that the ground motion models used in recent Indonesia PSHAs do not accurately represent the ground motion, particularly in areas with active volcanism. Secondly it identified that the performance of the exposure model varies across Java, but the reason for this variation is difficult to elucidate without a better understanding of the uncertainty in the exposure model. To assess this uncertainty, a suite of alternate mapping schema was developed for the exposure model. The result of this study is a detailed fine scale exposure model with uncertainty that can be used as an input for PSRA to support decision making in Java, Indonesia.
Presented at the 2024 18th World Conference on Earthquake Engineering, Milan, Italy
