FLORIDA HURRICANE ALLIANCE
PROJECT TITLE: Development of the Method and Geodatabase for Assessing Storm Surge Models
CONTRIBUTORS: Keqi Zhang, Chengyou Xiao, and Lixin Huang, International Hurricane Research Center
PERFORMANCE PERIOD: June 1, 2007 through November 30, 2007
Percentage of Work Completed: Round 1: 100 %; Round 2: 60 %
Project Proceeding on Schedule: [√] Yes [ ] No
Cost Status Round 1: [√] Cost Unchanged [ ] Under Budget [ ] Over Budget
Cost Status Round 2: [√] Cost Unchanged [ ] Under Budget [ ] Over Budget
Cost Status Round 3: [√] Cost Unchanged [ ] Under Budget [ ] Over Budget
Describe milestones achieved (publications, technology transferred, outreach actions, other results):
1. Standardizing input and output data of storm surge models
A consistent input and output data formats are critical to evaluate the results from various storm surge models. In
this period of research we continue to standardize the inputs and outputs for storm surge models using XML (Extensible
Markup Language) and NetCDF (Network Common Data Form). For example, the following is the snapshot of XML code
to describe the grid (using structured orthogonal grid as an example):
<QuadGrid IM="234" JM="458">
<Depth>grid.Depth</ Depth >
<Proj4>+proj=tmerc +lat_0=24.3+lon_0=-81 +k=0.999941 +x_0=200000 +y_0=0…</Proj4>
This piece of code exists as attributes in a NetCDF grid file. Note that the grid dimension, data, and grid projection are
clearly shown in NetCDF attributes. The standardized files also include storm track, H*Wind field, water level time series,
and high water output file. These files facilitate the model comparison by maintaining consistent input and output formats
from various models.
Fig. 1 Grid coverage of Florida Fig. 2 Hypothesis tracks for model testing
2. Refined the wetting and drying scheme to simulation storm surge flooding
Accurately simulating storm surge inland flooding is a challenging task because of the discontinuous nature of the
boundary between dry land and flooded areas and vice versa. A robust wetting and drying scheme in a storm surge model is
essential for simulating this process. We have developed a novel wetting and drying scheme in our Coastal and Estuarine
Storm Tide (CEST) model to better represent the storm surge flooding process by introducing a more reasonable velocity
estimation for water accumulation in dry cells (Xiao et al., 2006). This wetting and drying algorithm has been successfully
applied to the simulations of several historical landfall hurricanes along the U.S. and Gulf coasts.
However, the robustness of the algorithm has not been extensively tested against possible extreme cases, which is
essential for converting the CEST model into an operational model. Therefore, we first created a set of grids which cover
the entire Florida coast and the boundary of Lake Okeechobee (Fig. 1). Then, we employed a set of hypothetical category
5 hurricanes (Fig. 2) to test the robustness and the reliability of the wetting and drying scheme. For each grid, we examined
the effects of the variations in storm approach angles and landfall locations by altering storm tracks. The storm intensity is
designed to be Category 5 (wind speed >155 mph) at landfall. Fig. 3 and Fig. 4 are examples of the test results in Miami
and West Palm Beach, respectively. The results (highest water level) are shown in images which are overlain over Google
EarthTM for visualization. So far, the CEST model has been stable in testing using more than 50 tracks.
Fig. 3 Test in Miami Fig. 4 Test in West Palm Beach
3. Setup ADCIRC for model test
The ADCIRC (Advanced Circulation) model is a well known circulation model developed by the University of North
Carolina and the University of Notre Dame. It has been adopted by the US Army Corps of Engineers and other agencies
for tide and surge simulation. In order to compare the ADCIRC model with CEST and other models, we started to set up
the ADCRIC model for the Florida. Because the ADCIRC model uses a finite element scheme and covers a large area (Fig.
5), it is difficult to run ADCIRC using a single PC workstation. To speed up the simulation, we set up the ADCIRC model
in IHRC’s 16-node dual processor Linux cluster. We first simulated tides for the U.S. East and Gulf coasts. The model
domain covers the western U.S. Atlantic Ocean, Gulf of Mexico, and Caribbean Sea (Fig. 5). It took several hours for a 32
CPU cluster to simulate tides over one year. Model results are similar to the East Coast database provided by US Army
Corps of Engineers (Mukai, 2001). Second, we used input data provide by ADCIRC development group to test the storm
surge simulation of Hurricane Isabel in 2003. However, their data were derived by only considering storm surge simulation
without tide. We added a tide boundary to incorporate the interaction of surge and tide to further test the model. Snapshots
of water levels for Isabel without and with tide are shown in Fig. 6 and Fig. 7, respectively.
Fig. 5 ADCIRC grid
Fig. 6 Hurricane Isabel without tide Fig. 7 Hurricane Isabel with tide
Provide a schedule for the remainder of work to project completion:
• Continue to collect and convert storm surge field observation data, to evaluate the storm surge models, to collect
overland flooding data.
• Continue to test the CEST model over Florida.
• Continue to set up and test the ADCIRC model for the Florida; to refine the portion of the ADCRIC grid for the
Describe problems or circumstances affecting completion date, milestones, scope of work, and cost:
Xiao, C., Zhang, K. and Shen, J., 2006. CEST: A Three-Dimensional Coastal and Estuarine Storm Tide Model,
International Hurricane Research Center, Florida International University, Miami, Florida, 20 pp.
Mukai, A.Y., Westerink, J.J., Luettich, R.A. and Mark, D., 2002. Eastcoast 2001, A Tidal Constituent Database for Western
North Atlantic, Gulf of Mexico, and Caribbean Sea. ERDC/CHL TR-02-24, U.S. Army Corps of Engineers,
Vicksburg, Mississippi, 23 p.