Full Session with Abstracts
505563-2 - Primary Factors Influencing the Physical Vulnerability of Single-Family Residences to Extreme Winds in Hurricane Harvey
Friday, April 26
1:30 PM - 3:00 PM
Location: Celebration 7-8
The physical vulnerability of single-family residences to hurricane wind and surge loads is a function of many factors, whose influences and interactions are not well understood. Empirical data of building performance in past hurricane events provide important opportunities to evaluate such factors in the real world, and gain insights that drive the development or calibration of engineering-based models. This paper describes an evaluation of post-hurricane landfall data collected after Hurricane Harvey made landfall near Rockport, TX on August 25, 2018. The dataset consists of a unique compilation of data and metadata regarding building performance, including building attributes, structural design characteristics, damage observations, local ground surface roughness estimates and peak gust wind speed estimates for over 1,250 individual single-family homes impacted by Hurricane Harvey. Damage was estimated from ground-based photographs taken by the survey teams in door-to-door assessments, supplemented by 360 degree orthogonal views of the post-hurricane condition provided to the team by the impacted counties. The methodology allowed for more precise evaluation of damage at the building component level in the form of damage ratios (percentage of building component that was damaged by the hurricane, e.g., 40% of roof cover), which were then aggregated into overall damage ratings for each building. Building attributes included the assessed value of the home, orientation of the home, the year built, number of stories, elevation above ground level, roof shape, and presence and orientation of garage doors. Information regarding major retrofits (e.g., year the roof cover was reinstalled and to what building code) was sourced from the Texas Department of Insurance Windstorm Insurance Association where available. Upwind surface roughness was quantified using the Multi-Resolution Land Characteristics (MRLC) Consortium 2011 National Land Cover Database (NLCD). An ordered logistic regression model is used to explore the significance, influence and interactions of these attributes relative to the observed wind speeds and the surrounding upwind terrain characteristics.
Ultimately, this paper will present an overview of the data collection, curation, assimilation and archival methodologies used by the team following Hurricane Harvey. The key contributors to the physical vulnerability (as indicated by the observed damage) of single-family residences to hurricane wind and surge will be identified. Following the historic 2017 hurricane season, the audience will likely be keenly interested in learning from the post-hurricane building assessments that followed. This study will provide a comprehensive look at a variety of factors influencing building performance that practicing engineers may not be be fully aware of. Conference attendees will also likely have unique perspectives and observations to share with the authors, and thus we anticipate that this paper and session will facilitate interactive dialogs between academia and practicing engineers regarding the physical vulnerability of structures to hurricane impacts.