Geodatabase-assisted storm surge modeling

• Main Author: Binselam, Sait Ahmet
• Format: Text
• Language: unknown
• Published: LSU Digital Commons 2013
• Subjects: Storm Surge Modeling; Hurricane; Genesis; Tropical Cyclone Track; Artificial Neural Network; Storm Surge; Geodatabase; Engineering Science and Materials; North Atlantic
• Online Access: https://digitalcommons.lsu.edu/gradschool_dissertations/2061; https://doi.org/10.31390/gradschool_dissertations.2061; https://digitalcommons.lsu.edu/context/gradschool_dissertations/article/3060/viewcontent/uc.pdf

Tropical cyclone-generated storm surge frequently causes catastrophic damage in communities along the Gulf of Mexico. The prediction of landfalling or hypothetical storm surge magnitudes in U.S. Gulf Coast regions remains problematic, in part, because of the dearth of historic event parameter data, including accurate records of storm surge magnitude (elevation) at locations along the coast from hurricanes. While detailed historical records exist that describe hurricane tracks, these data have rarely been correlated with the resulting storm surge, limiting our ability to make statistical inferences, which are needed to fully understand the vulnerability of the U.S. Gulf Coast to hurricane-induced storm surge hazards. This dissertation addresses the need for reliable statistical storm surge estimation by proposing a probabilistic geodatabase-assisted methodology to generate a storm surge surface based on hurricane location and intensity parameters on a single desktop computer. The proposed methodology draws from a statistically representative synthetic tropical cyclone dataset to estimate hurricane track patterns and storm surge elevations. The proposed methodology integrates four modules: tropical cyclone genesis, track propagation, storm surge estimation, and a geodatabase. Implementation of the developed methodology will provide a means to study and improve long-term tropical cyclone activity patterns and predictions. Specific contributions are made to the current state of the art through each of the four modules. In the genesis module, improved representative data from historical genesis populations are achieved through implementation of a stratified-Monte-Carlo sampling method to simulate genesis locations for the North Atlantic Basin, avoiding potential non-representative clustering of sampled genesis locations. In the track module, the improved synthetic genesis locations are used as the starting point for a track location and intensity methodology that incorporates storm strength parameters into the ...