CEEES - Systematic Investigation of Building Code Efficacy in Reducing Windstorm Risk to Single Family Housing
Housing is highly vulnerable to climate risk and particularly windstorm risk, as evidenced by the mounting toll of billion dollar disasters in the US, questioning how codes and standards might better protect homes from climate-related risks. In response, this proposal leverages and expands a unique database of over 4600 detailed forensic evaluations of single family housing conducted by the Structural Extreme Events Reconnaissance (StEER) Network after seven major hurricanes and tornado outbreaks. Unsupervised machine learning will mine this data to establish performance metrics to be tracked across codes and standards and reveal the combinations of housing features driving observed performance (RQ 1.2). Upon linking these features to specific provisions in the codes governing at time of construction, the study will reveal not only the extent to which current codes are successfully achieving resilience outcomes (RQ 1.1) and how modern codes outperform older editions (RQ 1.3), but also provide systematic evidence of what code reforms or retrofitting programs should be prioritized by policymakers. As a result of the comparative power of the leveraged database, the study will also systematically document the effects of fragmented code adoption and enforcement across jurisdictions (RQ 2.1) through a comparative case study of three southeastern states along a continuum of regulatory compliance.
The study will yield systematic evidence to guide what code reforms or retrofitting programs should be prioritized by policymakers. Thus students will understand how we "learn from disasters" and more importantly conduct research that has the potential to change building codes and practices to reduce the toll on disasters on the United States. It is very rare for students to have access to data on how real-world, as-built structures perform and how to mine these large volumes of data to extract insights.