Hazard-consistent selection of storm scenarios for long-term storm surge hazard estimation across large geographic regions

The recent destructive hurricane seasons have increased the importance of storm surge hazard estimation for discussing coastal community resilience.The hazard is generally represented by surge inundation probabilities for each location (hazard curves) or surge levels corresponding to specific exceedance probabilities across all locations (hazard maps) within the coastal domain of interest.It is assessed through a Monte Carlo simulation (MCS), utilizing an ensemble of storm scenarios and their occurrence rates that represent the regional climatology.The impact prediction of each storm scenario within the MCS is performed by a hydrodynamicbased numerical model with a high computational cost.To address the challenge, reducing the number of storm scenarios considered for the hazard estimation is desirable.This is especially true in pre-disaster planning settings when investigating the impact of adaptation measures on coastal resilience.Since the entire process needs to be repeated for each measure, the number of storm scenarios needs to be kept moderately low.This study investigates the selection of fewer storms from the full ensemble within this setting, which can produce consistent hazard products with some chosen ones obtained from the full ensemble.Beyond the identification, the occurrence rates of selected storms are also updated.The final subset can be subsequently utilized as a representative ensemble for the regional hazard estimation.The storm selection is established as a two-stage optimization: the inner-loop adjusts the occurrence rates for a candidate subset; the outer-loop searches for the best subset with the desired number of storms.This work extends the implementation to extended coastal regions with many locations of interest.In this case, it is computationally intractable to consider all the locations in the domain, and for this reason, a subset of representative locations is chosen through cluster analysis and considered for storm selection.Additionally, the correlation in the hazard description is better integrated into the storm selection by modifying the objective function for the outer-loop optimization.