Date of Award
Master of Science in Ocean Engineering
Annette R. Grilli
This body of work consists of two manuscripts related to modeling ocean waves and erosion associated with the 100-year storm in southern Rhode Island.
Predicting the 100-year storm inundation along the Narragansett Bay shoreline
The lack of confidence in FEMA maps after Hurricane Sandy (2012) led to question the accuracy of the methodology used in FEMA Flood Insurance Rate Maps (FIRM). The present analysis presents a case study, re-computing the 100- year inundation maps in Washington County, RI, using a chain of two-dimensional models. Presented results focus on the West Passage of the Narragansett Bay, Rhode Island, USA, an area characterized by a complex shoreline. The selection of the initial and boundary conditions for ocean wave simulations are based on results of fully coupled 2-D surge and wave models, the ADvanced CIRCulation model (ADCIRC) and the phase-averaged STeady state spectral WAVE model (STWAVE), respectively, from the North Atlantic Coast Comprehensive Study (NACCS) performed by the U.S. Army Corps of Engineers. Waves are simulated using STWAVE over a series of high-resolution grids in near-shore and overland areas. The method is referred to as the NAST method. Results are mapped and compared with FEMA's results along transects. Cross-sections across the Base Flood Elevation (BFE) at the site of the FEMA 1-D transects are in relatively good agreement using both methodologies, with larger discrepancies shown in the northern section of the bay. Discrepancies are primarily due to a difference in the 100-year wind assumption and secondarily to the 100-year storm surge assumption, which, besides defining the depth and extension of the inundation, controls the wave impact over land. Mapping the results reinforces the importance of adopting a 2-D approach to fully represent the inundation, showing that the selected transects often miss the sites of the most extreme wave incursion and most destructive impacts. The analysis highlights the differences in methods and results but also suggests adopting, besides a 2-D approach, a scenario-based approach to the 100-year storm rather than a deterministic single map to assess the uncertainty associated with this event.
Modeling the erosion associated with the 100-year storm on the southern Rhode Island coast
The erosion due to a synthetic 100-year storm is modeled on the southern shore of Rhode Island using the process-based morphodynamic model XBeach (Roelvink et al., 2009). The study area includes barrier beaches, coastal ponds, and residential areas. The model is forced using a synthetic storm time series extracted from the North Atlantic Coast Comprehensive Study (NACCS; Cialone et al., 2015, Nadal-Caraballo et al., 2015) database, selected as a proxy 100-year storm.
An empirical method (Stockdon et al., 2006; Stockdon et al., 2012) is used to parameterize wave setup, swash, and runup to estimate the erosion impact regime (Sallenger, 2000), using NACCS water levels and offshore wave heights as inputs at seven cross-shore transects within the study area. As the storm is expected to evolve from the collision regime to the overwash and inundation regimes over time, the two-step XBeach model calibration process suggested by Nederhoff (2014) is followed. The first step consists of adjusting the facua wave skewness/asymmetry parameter to calibrate the collision regime. The second step involves testing the sensitivity to friction across the dune to calibrate the overwash regime. The sensitivity of the model to the facua parameter is tested by modeling a historical storm that stayed within the collision regime. Hurricane Irene (August, 2011) is selected and simulated erosion is compared to measured erosion at five cross-shore profiles monitored by the University of Rhode Island Graduate School of Oceanography (GSO). GSO performed measurements and data collection before and after Irene's impact at the site. For the 100-year storm, results of four simulation scenarios are presented: the facua parameter is tested for the default and maximum suggested values of 0.1 and 0.3, respectively, and the bottom friction is either set to a constant Manning's n of 0.02 or a variable Manning's n bed friction coefficient based on land cover.
The resulting eroded dunes for the four simulations are presented in 2-D maps as well as 1-D cross-sections at the seven cross-shore locations. Two additional transects to the GSO transects are included, located at the Federal Emergency Management Agency (FEMA) coastal transect sites that were used to create the 100-year Flood Insurance Rate Map (FIRM) for the region. Results are compared at these two sites with FEMAs dune erosion protocol. Additionally, the simulated dune is compared to the generalized barrier profile for a 100-year storm developed for the region by Oakley (2015).
Schambach, Lauren, "A Comparison of Wave and Erosion Modeling Methods for the 100-Year Storm in Southern Rhode Island" (2017). Open Access Master's Theses. Paper 978.