Doctor of Philosophy (PhD)


Engineering Science

Document Type



Accurate economic loss assessment for natural hazards is vital for planning, mitigation, and actuarial purposes. The widespread and costly nature of flood hazards, with the economically disadvantaged disproportionately victimized, makes flood risk assessment and mitigation particularly important. Recent studies have made advancements in assessing flood risk at individual building level. But these methods require solving an integral equation every time the analysis is run, which is computationally intensive. In addition, analyses to date do not consider insurance (i.e., coverage, deductible) information and thus do not estimate the portion of risk that homeowner and insurer bear. Although existing research does consider the freeboard mitigation option for a building, the role of other types of hazard mitigations, especially in coastal areas, in reducing flood risk reduction is not well explored.

This dissertation develops advanced analytical methods to quantify homeowner’s flood risk considering insurance information and hazard mitigation options. The flood hazard is characterized using a two-parameter (i.e., location and scale parameter) Gumbel extreme value distribution, with flood risk quantified in terms of average annual loss (AAL) at the individual building level. Monte Carlo simulation is used to estimate the AAL value as the area under the loss-exceedance probability curve. A synthetic data set of AAL values is generated for buildings located in Special Flood Hazard Area (SFHA – areas with at one-percent or greater chance of experiencing flood in a given year). Using the data set, generalized equations are developed to estimate AAL of buildings that are located in the SFHA using flood hazard parameter. Then, a machine learning model is trained on a synthetic data set of AAL values that considers flood hazard parameters, building replacement values, and insurance parameters (i.e., coverage and deductible) to apportion the flood loss between homeowner and insurer. The model predicts the apportionment factor that is used in determining the homeowner AAL. Finally, flood hazard mitigation in coastal areas by the reduction of wave and surge is quantified as the reduction of base flood elevation (BFE), and subsequently the reduction in homeowner’s AAL is estimated.

The result of this research will present generalized AAL equations to estimate flood AAL at individual building level, a machine learning model to predict the apportionment factor to determine the homeowner’s AAL, and the reduction of AAL for different hazard mitigation designs (wave and surge reductions scenarios). This comprehensive methodology will help community officials and government agencies to gain more insights about homeowner flood risk and assist researchers in quantifying flood risk for individual buildings and conducting benefit-cost analysis to recommend the optimal hazard mitigation option for coastal areas.



Committee Chair

Friedland, Carol J.

Available for download on Thursday, July 09, 2026

Included in

Risk Analysis Commons