1. Importance of high-resolution modeling for storm surge, hurricane waves, coastal water levels, and currents in Puerto Rico and the U.S. Virgin Islands
Juan Gonzalez-Lopez1, Joannes Westerink2, Andre van der Westhuysen3
1Caribbean Coastal Ocean Observing System (CariCOOS), 2University of Notre Dame, 3NWS/NCEP/Environmental Modeling Center
ABSTRACT
As part of the SURA/IOOS COMT II, the ADCIRC and ADCIRC+SWAN models have been used for the simulation of tides, currents, water levels, storm surge, and hurricane waves under conditions of purely-tidal forcing, tide-atmospheric forcing, and hurricane wind forcing. A high resolution unstructured mesh, with a spatial resolution of at least 100 m along all the coastlines of Puerto Rico and the Virgin Islands, has been used for these simulations. This mesh provides a very accurate representation of the irregular coastlines of the region, the steep bathymetric gradients, and coastal features such as reef systems and shoals which affect the coastal dynamics and the breaking of waves during both hurricanes and winter swell events. Results of both ADCIRC and ADCIRC+SWAN simulations show very good agreement with observations of tidal harmonic constituents, water levels, storm surge, and wave parameters. In addition, the results show that the high resolution of the model allows for the generation of non-linear effects such as high- frequency water level oscillations and seiching modes. The coupling of the tide and atmospheric forcing during non-cyclone conditions has shown that atmospheric forcing is necessary for the generation of coastal currents in Puerto Rico and the U.S. Virgin Islands. Blocking and dissipative effects of the reef and shelf geometries generate very localized wave breaking and water level distribution and patterns during hurricane forcing, which would be of great benefit for NOAA if incorporated into the storm surge modeling systems. The inclusion of these simulations and the validation data in the COMT II database will allow for the distribution and easy accessibility of results by the scientific community and government agencies, particularly through the visualization infrastructure being built as part of the COMT.
High Resolution Mesh
Storm Surge and Hurricane Waves
PRVI215 computational domain and bathymetry.
PRVI215 bathymetry in the Puerto Rico and Virgin Islands and detail of bathymetry in the eastern coast of Puerto Rico.
Tide-Atmosphere Interaction
Observed and modeled wind speed, direction, and atmospheric pressure at San Juan during Hurricane Georges.
Parametric (top) and WRF-ARW (bottom) modeled winds for Hurricane Georges. WRF-ARW has a spatial resolution of 2 km. WRF-ARW winds provided by Dr. Luis Aponte from UPRM.
Water levels and seiching
Effect of reefs on surge and waves
Red: Observed, Blue: ADCIRC with tidal and atmospheric forcing, Black: ADCIRC with tidal forcing only
The top panel shows the water levels at Magueyes Island (left, southwest of PR) and the mean absolute error for all tide gauges in Puerto Rico and the U.S. Virgin Islands during January-November Lower panel shows the spectral energy spectrum as observed in Magueyes Island (left) and the computed spectrum from ADCIRC with tidal and atmospheric forcing (right). ADCIRC with tidal and atmospheric forcing is able to generate the observed 50-minute seiching period at Magueyes Island.
Coastal currents
Water levels (left panel) and significant wave heights (right panel) during Hurricane Georges using parametric wind forcing. In the second and third rows notice the effect of the reef chain on the northeast coast of Puerto Rico. This reef chain behaves as an obstacle that serves to create a “pool” of higher water levels between Vieques and Culebra Islands and the northeast coast of PR. In addition this reef chain and the shallow banks in this area help to dissipate the big hurricane waves.
Runtime sensitivity to SWAN accuracy constraints
Top panel: maximum speed for ADCIRC with tidal and atmospheric forcing (left) and tidal forcing only (right). Notice how without atmospheric forcing the current along the North and South of Puerto Rico are not generated as the horizontal tidal gradient is not enough to generate current by itself. Lower panel shows the ADCIRC currents with tidal and atmospheric forcing in the eastern coast of Puerto Rico.
ADCIRC+SWAN runtimes for Hurricane Georges using different SWAN configurations (left) and significant wave heights at San Juan (right) for each of these configurations. The results show that the greatest impact on the computed significant wave height comes from the percentage of nodes that need to satisfy the accuracy constraints and not the maximum number of iterations. This results in being able to get accurate reports while greatly increasing the runtime speed.
San Juan depth-averaged currents: Observed (red), ADCIRC with tidal and atmospheric forcing (blue) and ADCIRC with tidal forcing only (black).
Mean absolute error of depth-averaged current speeds at San Juan (blue) and St. John (red). Filled symbols are ADCIRC with tidal and atmospheric forcing and empty are tidal forcing only.