Sampled Building Dataset for Building Vulnerability Classifiction and Tsunami Fragility Analysis
Description
This dataset contains sampled buildings used to develop multi-event tsunami fragility curves. It includes records from three major tsunami events: the 2011 Great East Japan Earthquake and Tsunami (Natori and Rikuzentakata, Japan), the 2018 Sulawesi Tsunami (Palu, Indonesia), and the 2018 Sunda Strait Tsunami (Banten and Lampung, Indonesia). For each building, hazard intensity (inundation depth), structural attributes (construction material and number of storeys), surrounding characteristics (coastal vegetation and coastal protection), and the computed vulnerability index (VI) with recorded damage states are provided. Data are organised into three sheets corresponding to the vulnerability classes: V1 (less vulnerable), V2 (vulnerable), and V3 (highly vulnerable).
Files
Steps to reproduce
Data collection: 1. Building data are extracted from published field survey reports and databases for three tsunami events: 2011 Great East Japan Earthquake and Tsunami (Natori and Rikuzentakata city, Japan), 2018 Sulawesi Tsunami (Palu, Indonesia), and 2018 Sunda Strait Tsunami (Banten and Lampung provinces, Indonesia). 2. Suitable buildings are first screened based on availability of inundation depth, construction material (Cm), number of storeys (Bs), and damage states data. Buildings with unknown data, and located more than approximately 2km from the coastline are excluded. 3. Width of coastal vegetation, Cv (between shoreline in the direction of tsunami wave and building wall facing the shoreline) is measured based on satellite imageries pre-disaster. 4. Coastal protection structures are identified based on post-tsunami field survey data, publications, and/or visual assessment from satellite imageries pre-disaster. 5. Observed tsunami damage states (DS0–DS4) are harmonised across datasets according to the modified classification scheme provided in the Methods section. Indicator scoring: 6. Each building attribute is converted into a normalized value i.e. Construction material = Cm, Building storeys = Bs, Coastal vegetation = Cv, Coastal defence = Cd following the weightages described in the Methods section of the article. Vulnerability Index (VI) calculation: 7. The VI for each building is calculated as the weighted sum of the normalised indicators using the refined BVI equation stated in the Findings section of the article. 8. Buildings are categorized into three vulnerability classes (V1 = less vulnerable, V2 = vulnerable, V3 = highly vulnerable) using the Jenks Natural Breaks classification, as detailed in the paper. R Studio software is used to simplify the categorization process.