Coastal Waves, Surge and Inundation in the Gulf of Mexico

Effect of wind input parameterizations on hurricane wave estimation

Description: The effect of wind input parameterizations on hurricane wave estimation in the SWAN wave model was evaluated for Hurricane Ike. The default/recommended setting for the wind input parameterization overestimated the maximum significant wave heights by about 2 m in the deep Gulf of Mexico when compared with observations from moorings. The overestimation could be relieved either by adjusting the maximum value of the surface drag coefficient or by substituting a high-wind-speed bulk formula for the default low-to-moderate one used in SWAN. Because of the dissipative effects of the shallow coastal areas, the overestimation of waves in deep water has limited impact on the waves in near-shore waters. Thus, previous wave model results using a low-to-moderate wind speed bulk formula may still be reliable in waters shallower than 20 m while overestimating significant wave heights in deeper waters for high wind speed conditions such as hurricanes.

Publications: Huang, Y., R.H. Weisberg, L. Zheng, M. Zijlema, 2013. Gulf of Mexico hurricane wave simulations using SWAN: Bulk formula-based drag coefficient sensitivity for Hurricane Ike, J Geophysical Research Oceans, 118(8):3916-3938, DOI: 10.1002/jgrc.20283

Hurricane Ike landfall

Description: Hurricane Ike made landfall near Galveston, TX, as a moderate intensity storm. However, its large wind field in conjunction with the broad Louisiana-Texas shelf and large-scale concave coastal geometry generated waves and surge that impacted over 1000 km of coastline. Ike's complex and varied wave and surge response included: the development of a storm surge “forerunner” 24 h prior to the storm's landfall due to strong shore-parallel, wind-driven currents and the associated across-shelf, geostrophic setup; the resulting early rise of water in coastal bays and lakes facilitating inland surge penetration and inundation; the shore-normal wind-driven peak surge; the southward propagation of a free wave along the Texas shelf; and the appearance of resonant and reflected waves on the adjacent continental shelf. Preexisting and rapidly deployed instrumentation provided the most comprehensive hurricane response data set ever recorded. More than 91 wave parameter time histories, 523 water level time histories, and 206 high water marks were collected in deep water, in the near shore, and up to 65km inland. A comprehensive skill assessment demonstrated the ability of ADCIRC+SWAN to capture the principal aspects of the observed storm response

Publications: M.E. Hope, J.J. Westerink, A.B. Kennedy, P.C. Kerr, J.C. Dietrich, C. Dawson, C.J. Bender, J.M. Smith, R.E. Jensen, M. Zijlema, L.H. Holthuijsen, R.A. Luettich, Jr., M.D. Powell, V.J. Cardone, A.T. Cox, H. Pourtaheri, H.J. Roberts, J.H. Atkinson, S. Tanaka, H.J. Westerink, L.G. Westerink, 2013. Hindcast and validation of Hurricane Ike (2008) waves, forerunner, and storm surge, J. Geophysical Research Oceans, 118(9): 4424-4460, DOI:10.1002/jgrc.20314.

ADCIRC+SWAN Sensitivity

Description: The sensitivity of ADCIRC+SWAN solutions of tides, surge, and waves during hurricane Ike was evaluated for differing grid resolution, topographical detail, bottom friction, wave-current interaction, and nonlinear advection at basin, shelf, wetland, and coastal channel scales. Grid resolution requirements were found to be less stringent in the open ocean, however, coarse resolution or the absence of intratidal zones decreased solution accuracy along protected near shore and inland coastal areas due to decreased frictional attenuation. Diurnal tidal amplitudes were more sensitive to the presence of intratidal zones and coarse mesh resolution than semidiurnal tides. The bottom friction parameterization had little effect on tidal skill, however, it had a significant impact on the strength of the alongshore current generated during hurricane Ike and the magnitude of the resulting geostrophic setup (forerunner). Nonlinear advection increased the geostrophic set up by 15–20 cm and increased resonant shelf waves by 20–30 cm. Wave radiation stress added 20–40 cm to water levels at coastal stations.

Publications: Kerr, P. C., R. C. Martyr, A. S. Donahue, M. E. Hope, J. J. Westerink, R. A. Luettich Jr., A. B. Kennedy, J. C. Dietrich, C. Dawson, and H. J. Westerink, 2013, U.S. IOOS coastal and ocean modeling testbed: Evaluation of tide, wave, and hurricane surge response sensitivities to mesh resolution and friction in the Gulf of Mexico, J. Geophysical Research Oceans, 118(9): 4633-4661, DOI:10.1002/jgrc.20305.

Unstructured surge-wave models

Description: Three unstructured, coupled surge-wave models, ADCIRC+SWAN, FVCOM+SWAN, and SELFE+WWM were compared using identical grids, forcing and parameterizations for hurricanes Ike and Rita. In addition, NWS's official operational forecast storm surge model, SLOSH, was used on both local (Galveston and Sabine Pass basins) and Gulf of Mexico (ETSS) grids. The three unstructured models yielded similar results for both hurricanes. These models all appeared to reproduce the important physical processes, showed minimal water level or wave height bias, and comparable variances versus observations. SLOSH using the Galveston Basin grid failed to capture the hurricane Ike forerunner and was therefore biased significantly low. SLOSH performed better on the Sabine Pass Basin grid for hurricane Rita, which did not elicit a significant forerunner. SLOSH on the ETSS grid showed minimal bias for either storm, although its accuracy was limited by the nearly 5 km resolution in near shore/onshore areas. In all cases, SLOSH deviations from observations were greater than those from the three unstructured grid models. The largest difference in model performance was observed in execution speed and scalability. The implicit time stepping scheme of SELFE+WWM performed well at small numbers of cores, but scaled poorly at larger numbers of cores. ADCIRC+SWAN had better scaling and absolute performance when more than 128 cores were used per run. SLOSH is not configured to utilize modern parallel computing architecture and rather is limited to running on a single core. Runtimes for ADCIRC on a single core were more than 10 times longer than for SLOSH on the ETSS grid, even after the SLOSH runtimes were normalize for the number of grid nodes. Thus, SLOSH remains more efficient for use in probabilistic forecasting based on large ensembles of model runs. However, SLOSH-based probabilistic forecasts should be assessed for accuracy (particularly high or low bias) by comparing select, individual SLOSH runs with similar runs using higher resolution unstructured grid models as presented herein.

Publications: Kerr, P. C., A.S. Donahue, J.J. Westerink, R. A. Luettich, Jr., L.Y. Zheng, R.H. Weisberg, Y. Huang, H.V. Wang, Y. Teng, D.R. Forrest, A. Roland, A.T. Haase, A.W. Kramer, A.A. Taylor, J.R. Rhome, J.C. Feyen, R.P. Signell, J.L. Hanson, M.E. Hope, R.M. Estes, R.A. Dominguez, R.P. Dunbar, L.N. Semeraro, H.J. Westerink, A.B. Kennedy, J.M. Smith, M.D. Powell, V.J. Cardone, A.T. Cox, 2013, U.S. IOOS coastal and ocean modeling testbed: Inter-model evaluation of tides, waves, and hurricane surge in the Gulf of Mexico, J. Geophys. Res. Oceans, 118(10): 5129-5172, DOI:10.1002/jgrc.20376.

Comparison of 2-D and 3-D model responses

Description: A comparison of two-dimensional (2-D), vertically integrated, and three-dimensional (3D) model responses was performed using FVCOM for hurricane Ike. Both 2-D and 3-D models were found to accurately predict the surge response although different bottom friction formulations are required by each type of model. Sensitivity studies indicated that hurricane storm surge in both 2-D and 3-D depends critically upon the bottom friction parameterization.

Publications: Zheng, L., R. H. Weisberg, Y. Huang, R. A. Luettich, J. J. Westerink, P. C. Kerr, A. S. Donahue, G. Crane, and L. Akli, 2013, Implications from the comparisons between two- and three-dimensional model simulations of the Hurricane Ike storm surge, J. Geophysical Research Oceans, 118(7): 3350–3369, DOI:10.1002/jgrc.20248.