Modeled coastal surge depth grids were generated for Hurricane Nate. These depth grids are intended for use in a high-level analysis of flood depth impacts to buildings for geospatial damage assessments.
Data Sources & Methodology
Area of Interest
The Area of Interest (AOI) for this dataset is based on the hurricane landfall location in southern Mississippi. This area extended from Hancock County, MS to Baldwin County, AL, as shown below. Coastal flooding depth grid products were produced for five counties in this area.
The elevation data used for this dataset comes from various sources:
- The Hancock, Harrison, and Jackson County, MS analyses were analyzed with NOAA Sea Level Rise Viewer DEMs. These models had been previously processed to remove false elevations over coastal waters, or water elevations less than zero.
- The Mobile County, AL flood depth analyses were performed using a 2014 NOAA Bare Earth LiDAR DEM obtain from the NOAA Digital Coast.
- Baldwin County, AL flood depth grids used data from the recent coastal FEMA storm surge study by Northwest Florida Water Management District. The data was comprised of 2001 Baldwin County wide LiDAR and post-Katrina LiDAR flown along the coastlines. The data was provided in tiles and was mosaicked together and converted to feet for this work.
- The Mississippi and Mobile County, AL DEM data for this project were projected to NAD83 UTM 16N and resampled to 7.62 meters (25 feet). Baldwin County, AL was sampled to 3 meters (9.84 feet). Vertical units of all DEMS are feet.
Hurricane Storm Surge Elevation Data Validation and Selection
Hurricane Nate coastal storm surge elevation data were derived from the National Hurricane Center (NHC) P-Surge advisory 16 10% probabilistic storm surge exceedance elevations relative to NAVD88. P-Surge storm surge data is developed from the NHC SLOSH forecasting model. These included ESRI-compatible polygon shapefile coverages. Additional P-Surge data were retrieved from the NHC, including the 10% and 50% probabilistic storm surge exceedance elevations relative to NAVD88 for advisory’s 14, 15, and 16 for comparison.
Hurricane Nate coastal storm surge elevation data were also obtained from the ADCIRC Surge Guidance System (ASGS) group, made up of the University of North Carolina at Chapel Hill, the University of Notre Dame, and the Coastal Emergency Risks Assessment (CERA) program at the Louisiana State University. CERA is based upon work supported by Louisiana Sea Grant at LSU and the U.S. Department of Homeland Security under Grant Award Number, DHS-14-ST-061-COE-001A. CERA storm surge data is developed from the ADCIRC/SWAN coupled modeling system. Native maximum storm surge elevations were in units of meters relative to NAVD88. Post-processing included conversion of units to feet using a factor of 3.2808, and removal of points without surge elevations. CERA products were evaluated; however, it was found that the NHC forecasts results provided slightly better validation against observed data taken together across all areas.
For each area of interest, storm surge modeling products from NHC and CERA from the three closest NHC advisory forecasts to the area of interest were evaluated for best fit to available peak surge observations from water level stations. Station elevations were pulled after peak surge occurred in each area of interest via a custom script from the NOAA CO-OPS stations and USGS gages. CERA data were retrieved from http://cera.cct.lsu.edu in NetCDF format and converted to an ESRI-compatible point shapefile. Next, the points were converted to an initial raster surface for the validation assessment. NHC forecast data were obtained in GRIB II format and then converted to ESRI raster format. Surge elevation values were extracted to the point locations of the valid station observations. The resulting data tables below were reduced to locations allowing cross-comparison between all model products in the evaluation. Finally, the differences between modeled and observed results were tabulated, and then RMSE and cumulative deviation from observed were calculated for each area of interest. Stations that broke during the storm and did not capture a peak water surface elevation were not used in the calculation of errors. The selection of model source for depth mapping was then made based on the source with best validation metrics. Summaries of the validation are provided below, see the table here. Results show that the NHC surge simulation elevations for Advisory 16 10% exceedance were markedly better than other NHC forecast advisories and CERA advisory 14, 15, and 16. NHC forecast advisory 16 10% exceedance results were selected based on their best fit to the observations.
Water surface creation
Point shapefiles created from the NHC P-Surge format were subdivided into smaller data-manageable sections of coastline. Surge elevations were converted from meters to feet using a factor of 3.2808. A water surface elevation (WSE) raster was then created for each AOI through an inverse distance weighting technique, making use of breaklines, where necessary, to separate ocean, back-bay, and estuarine flood elevations. The WSE was the visually quality controlled by a senior coastal scientist. Any anomalous elevations or poor transitions were identified and corrected by the analyst. Overall the elevations had good local consistency and negligible editing was necessary.
Flood extent creation
A “raw” floodplain was delineated from each WSE, and then automatically post-processed to remove insignificant small voids and disconnected flooded areas. The flood extent was visually inspected to remove additional areas shown as flooded, but without apparent hydraulic connection to flood sources. This review was not to the level afforded a flood insurance study floodplain, but to quickly remove readily identifiable areas to improve accuracy of the identified flooded areas, given the constraints of the rapid production timeline. A senior coastal scientist then performed a review and adjusted inclusion/exclusion, where necessary.
Depth grid creation
The depth grid was generated by subtracting the digital elevation model from the surge water surface elevation model at a cell size of 7.62 m (25 feet). The product was then extracted to the cleaned flood extent polygon to remove artifacts and hydraulically disconnected areas.
Access & Use Information
Public: This dataset is intended for public access and use.
Downloads & Resources