Hurricane Ivan Storm Surge Re-analysis
 

Following Hurricane Ivan’s devastating landfall near the Florida-Alabama border in September 2004, Emerald Ocean Engineering LLC was tasked by Escambia County FL with designing the emergency berm on Perdido Key, FL, one of the hardest-hit areas stretches of shoreline. The design criteria for the berm was to provide protection for a 5-year return period storm event, but a flurry of storms between 1995 and 2004, two of which exceeded the 100-year surge elevation used by the Federal Emergency Management Agency’s (FEMA) to create the Flood Insurance Rate Maps for Florida, called  the currently-used statistics into question. EOE performed a preliminary updated extremal analysis that took all of the storms from 1886 through 2004, including Hurricanes Opal and Ivan, into account.

The original database of storm surge elevations was not saved by FEMA and is no longer available. Without these data, a straightforward calculation of the effect of hurricanes occurring after1995 on Perdido Key surge frequency was not possible. Fortunately, another source of surge data was available. In 1994, the US Army Corps of Engineers produced a database of surge elevations for the US Atlantic and Gulf of Mexico coasts (Scheffner et al, 1994). This report used the ADCIRC numerical model to generate nearshore surge elevations based upon NOAA’s extant HURDAT database of tropical storm meteorological parameters. That version of HURDAT included storms from 1886 through 1989. The nearest output station to Perdido Key is station 503, Lat 30.24580 N, Long 87.42510 W, or 3 nm offshore of the eastern end of the Project. The maximum predicted surge values for station 503 from this report were defined as the “old” surge subset. It includes 21 storms that occurred in the 92-year interval between 1893 and 1985

After publication of the Corps’ report, the HURDAT database was updated by NOAA to include additional historical, as well as the more recent, hurricanes. More recently, it has been updated annually. The 2004 update was used to develop the “new storm susbset” of storms that impacted Perdido Key. It includes 41 storms that occurred in the 151-year interval between 1853 and 2004 (including Hurricane Ivan). However, no modeled surge values were available for the 20 additional storms, so estimated values of maximum storm surge were utilized in the analyses.

The “old” and “new” surge data sets were input into the Empirical Simulation Technique (EST) computer model to produce storm surge frequency values (surge elevation by return period). Figure 1, shows the results. The surge for any return period is a statistic based upon 1000 different simulations of the historical events. In addition to the mean value, the 90 % non-exceedence values is provided, meaning 900 of the simulations resulted in elevations at or below this curve. These surge elevations include wind stress and barometric pressure effects only, but not astronomical tide nor dynamic wave setup. The effect of tide stage tends to average out to mean sea level (approximately 0 NAVD) so it has only a moderate effect. Wave set up can be a significant component of surge, however. To a first approximation, the 90% N.E. curve can be considered the worst case scenario, including nearshore wave set up and a landfall at high tide. From Figure 1,  it is clear that the maximum expected surge elevation for a 5- year event does not exceed + 4 ft NAVD at this site.  However, the 100-year event could exceed 16 ft.

Figure 1.
This plot shows results of an updated extremal analysis of Perdido Key FL. The revised surge level for the 100-year event is nearly 3 ft higher than the value in use by the Federal Emergency Management Agency when Ivan struck.

For comparison, the fifth curve on Figure 1 plots values produced for Perdido Key by Florida State University’s Beaches and Shores Resource Center (Dean and Chiu, 1986). The BSRC values began with the 10-year event, so the curve was extended linearly down to the 5-year event. Finally, FEMA’s current (effective Feb. 2000) 100-year still water and maximum wave crest elevations are shown. The BSRC and FEMA data include astronomical tide and nearshore wave set up as well as wind and pressure effects, and could be expected to be higher than predictions without these components. The fact that both are exceeded by the EST results indicates that a more thorough evaluation of the surge frequencies in Escambia County is warranted.