Frontiers in Built Environment (Aug 2020)
An Imputation of First-Floor Elevation Data for the Avoided Loss Analysis of Flood-Mitigated Single-Family Homes in Louisiana, United States
Abstract
PurposeStatistical data imputation methods are important in a wide range of scientific research; however, in construction management research, they are not used widely. Specifically, in research for building loss studies due to extreme hazard events, data are frequently missing, inaccessible, spurious, or expensive to collect. First-floor elevation (FFE) data are vital in building flood loss analysis, so the lack of high-quality FFE data before and/or after elevating structures represents a major barrier to understanding avoided loss (AL) in flood mitigation projects. While a few guidelines exist to estimate FFE, the guidelines lack information on estimation of FFE for mitigated and non-mitigated buildings. Existing techniques tend to rely on recommendations by professional engineers that have not been evaluated for statistical fit in elevated homes in Louisiana, United States.MethodsThis Louisiana-based case study statistically evaluates the effectiveness of existing guidelines on building elevation data. Furthermore, it provides a state-of-the-art methodology to impute missing FFE data statistically for buildings in mitigated and unmitigated conditions without relying on building foundation-type data, which itself is commonly needed but often missing in previous building mitigation AL studies.FindingsResults here suggest that existing guidelines for FFE estimation match reported FFE only moderately well in flood-mitigated residences in Louisiana. Moreover, an update and inclusion of foundation-type data in the guidelines would improve FFE estimates for Louisiana homes. Among the imputation methods by multiple linear regression, random forest, and generalized additive models (GAM) overlay, the GAM model performs most effectively based on the accuracy in data imputation for missing FFE data. These results will assist builders, developers, and communities in their quest to enhance resilience to the ever-increasing flood hazard.
Keywords
