Seismic vulnerability assessment of residential buildings using logistic regression and geographic information system (GIS) in Pleret Sub District (Yogyakarta, Indonesia)
Background: The Southeast of Yogyakarta City has had the heaviest damages to buildings in the 2006 of Yogyakarta Earthquake disaster. A moderate to strong earthquake of 6.3 Mw shook the 20 km southeast part of the Yogyakarta City early in the morning at 5:54 local time. On top of extensive damage in Yogyakarta and Central Java, more than 5700 people perished; 37,927 people were injured in the collapse of more than 240,396 residential buildings. Furthermore, the earthquake also affected the infrastructure and local economic activities. The total damages and losses because of the earthquake was 29.1 trillion rupiahs or equal to approximately 3.1 million US dollar. Two main factors that caused the severe damages were a dense population and the lack of seismic design of residential buildings. After reconstruction and rehabilitation, the area where the study was conducted grew into a densely populated area. This urbanistic change is feared to be potentially the lead to a great disaster if an earthquake occurs again. Thus, a comprehensive study about building vulnerability is absolutely needed in study area. Therefore, the main objective of this study has been the provision of a probabilistic model of seismic building vulnerability based on the damage data of the last big earthquake. By considering the relationship between building characteristics, site conditions, and the damage level based on probabilistic analysis, this study can offer a better understanding of earthquake damage estimation for residential building in Java.
Results: The main findings of this study were as follows: The most vulnerable building type is the reinforced masonry structure with clay tile roof, it is located between 8.1-10 km of the epicentre and it is built on young Merapi volcanic deposits. On the contrary, the safest building type is the houses which has characteristics of reinforced masonry structure, asbestos or zinc roof type, and being located in Semilir Formation. The results showed that the building damage probability provided a high accuracy of prediction about 75.81%.
Conclusions: The results explain the prediction of building vulnerability based on the building damaged of the Yogyakarta earthquake 2006. This study is suitable for preliminary study at the region scale. Thus, the site investigation still needs to be conducted for the future research to determine the safety and vulnerability of residential building.