Hindcasting of Compound Pluvial, Fluvial, and Coastal Flooding during Hurricane Harvey (2017) using Delft3D-FM

Abstract

Abstract Hurricane Harvey (2017) resulted in unprecedented damage from storm surge, and rainfall (pluvial) and riverine (fluvial) flooding in the Houston-Galveston area of the U.S. Gulf Coast. The objective of this study was to better quantify the impacts of compound flooding and to assess the relative contributions of storm surge, pluvial and fluvial flooding in a complex coastal environment using Hurricane Harvey as a case study. Although significant work has been done on Hurricane Harvey hindcasting, large-scale coupled modeling incorporating a multitude of land and ocean flood generation mechanisms is still at its early stage. Here we developed a comprehensive numerical modeling framework to simulate flood exents and levels during Hurricane Harvey using the open-source Delft3D Flexible Mesh, and validated results against observed water levels, waves, winds, hydrographs and high water marks. A nested mesh was developed to represent ocean and inland areas, enabling higher resolution for land regions of interest while balancing overall computational load. Results show that pluvial flooding dominated during Harvey, accounting for ~ 60–65% of flooding in the Houston/Galveston areas, attributed to widespread heavy rainfall being the dominant driving force. Widespread rainfall caused extensive pluvial flooding in watersheds and floodplains in West and South Bays ( ≤ ~ 1.5 m), upper Galveston Bay (Trinity River Basin, 2 ~ 3 m), and Harris County ( ≤ ~ 2.5 m). River runoff led the local flooding of ~ 1 to 2 m in the river basins. Significant surge levels were simulated northwest of main Bay (2 ~ 2.5 m) and Galveston Bay (1 ~ 2 m) areas and in several watersheds in West/East of Galveston Bay. Maximum flooding extent developed around August 29, 2017, which compared well to the flood depth data released by FEMA. Additional sensitivity studies suggest that increased compound flooding (e.g., 15% increase in combined pluvial and fluvial flooding) can lead to significantly more increase (0.3 ~ 0.5 m) in flood depths in low-lying regions. Nonlinear effects of compound flooding greater than individual components summed up. Results from this large-scale modeling analysis contribute to understanding of compound flooding risks in coastal urban areas, providing a useful basis for coastal risk management and hazard mitigation amid climate change. Our integrated framework is general and can be readily applied to other coastal compound flooding analyses.