The impact of Louisiana's levees and wetlands on Katrina's storm surge

Katrina generated a U.S.-record storm surge which impacted a wide region from Grand Isle, LA, to Mobile Bay, AL, and killed about 1350 people with hundreds still missing. Sensitivity experiments by WorldWinds Inc. using the ADvanced CIRCulation (ADCIRC) storm surge model show a large hurricane produces water elevations 20-40% higher than a small hurricane with the same intensity and with considerably more widespread inundation (Valenti, Fitzpatrick, and Li 2005). The previous U.S.-record surge was Hurricane Camille (1969) which impacted the same region, an intense but smaller Category 5 hurricane. Camille's hurricane-force winds extended 60 miles from the storm center, while Katrina's extended 120 miles. Camille's tropical storm-force winds reached 180 miles outward, while Katrina was 230 miles. While this large wind field mostly caused the surge, an open question remained on the whether the Mississippi River levees or Louisiana's eroded wetlands also contributed to the record surge.

The Sea, Lake, and Overland Surges from Hurricanes (SLOSH) model (Jelesnianski et al. 1992) is used to simulate the storm surge generated by Hurricane Katrina along the MS / LA coast. SLOSH is a hydrodynamic model which solves the depth- integrated equations of motion and continuity equation. The model is forced by a two-dimensional parametric wind equation based on storm's track, radius of maximum winds (RMW), and pressure deficit (environmental pressure minus central pressure). SLOSH incorporates bathymetry, topography, channels, and barriers (levees) on a stretched finite difference regional grid system. The computational cells are designed such that resolution increases toward the coastline on a telescoping polar grid mesh. The input follows that of Myers and Malkin (1961), which assumes a balance between the centrifugal force, friction, Coriolis force, and pressure gradient force parallel and perpendicular to a wind trajectory.

Additional tests examines the possible impacts of the Mississippi River levee system in enhancing the storm surge. In simulations with no river levees, land elevation is reduced to 4 feet to mimic the natural ridge along the river system and allow surge overflow into Barataria Basin. This "removal" of the man-made levee system (which are 10-15 ft tall) extends northward to near Chalmette where secondary levee systems exist.

All simulations are conducted for the period of 18 UTC 26 August 2005 to 18 UTC 29 August 2005.

The control run simulation (with river levees) shows the surge gradually rising, then quickly inundating Plaquemines Parish followed by St. Bernard Parish between 09 UTC and 12 UTC. The surge later peaks along the Mississippi coast between 14 and 16 UTC with the record surge.

The impact of the Mississippi River levee system at 13 UTC is contrasted to a simulation with only a natural river system south of Chalmette. SLOSH shows the east of the river where water is either trapped by the levees or has restricted overflow. The Barataria Basin west of the Mississippi River has higher surge.

Maximum water elevation values were computed for each grid cell. Simulations with SLOSH show that the river levees impeded the surge from spreading westward into Barataria Bay, accumulating the water in areas east of the levee system as the storm moved northward. This resulted in faster inundation in Chalmette and the Ninth Ward by 1-5 hours, and a higher surge of 3-7 feet. In contrast, the levees had little impact on the timing or height of the surge on the Mississippi coast. It is possible that, without the river levees, hard-hit Chalmette and the 9th Ward may have experienced significantly less flooding. However SLOSH only accounts for overtopping of the Chalmette levees, not breaches. The levees north of Chalmette experienced numerous breaches from a combination of overtopping, wave action, and poor soil levee composition. The experiments were repeated with ADCIRC, and show similar results. However, ADCIRC does not show differences inside the Chalmette levee system.

An additional issue involves the impact of the Louisiana wetlands and the Mississippi River. It is widely believed that wetland erosion has increased storm surge vulnerability in southeast Louisiana. SLOSH model runs were performed to examine the impact of wetland loss in the last 65 years. Specifically, we investigated: 1) a hurricane moving over the less-eroded marsh of Louisiana in 1930; and 2) a weaker hurricane due to more marshland. Results show a surge increase of 1-3 feet higher due to wetland erosion and subsidence.