| description abstract | Recent advancements in technology and infrastructure have greatly improved the capabilities of the natural hazards community to collect robust samples of building performance following hazard events and make them available to the research community for broad reuse purposes. Yet, there lacks standardized, open access data sets that combine reconnaissance data from multiple individual hazard events into unified, living testbeds that can grow through community participation. The objectives of this study are to (1) synthesize and present a unified, multievent windstorm performance data set (WiSPD) and (2) summarize common damage patterns observed in the WiSPD. The WiSPD currently consists of four hurricanes (occurring between 2017 and 2020) and four tornadoes (occurring between 2011 and 2020), all of which struck the United States. Each event’s building performance assessments were collected with similar methodologies and contain details such as location, physical address, basic building attributes, estimated 3 s gust wind speed, basic wind speeds, and component-level damage percentage with a precision of ±5%. In combination, the testbed reveals that roof cover damage dominates in windstorms, regardless of the year of construction or building code enforced. Additionally, tornadoes tend to produce higher damage rates than hurricanes, specifically in fenestration and roof structure damage. Fragility functions for hurricanes exhibit a nonmonotonic relationship between wind speed and damage, potentially evidencing the strong influence of other confounding variables. Ultimately, the unified data set promises to be a rich testbed for further knowledge discovery and model validation by the research community. | |
