Hurricane Katrina and the New Orleans Levee System

Extended Abstract

Hurricane Katrina and the New Orleans Levee System

Abstract – Hurricane Katrina was one of the worst natural disasters in North American History, resulting in 1600 fatalities and billions of dollars of direct economic loss in southern Louisiana. New Orleans, particularly, was the city most devastated by the catastrophe. This extended abstract will concisely discuss the calamities the city is most susceptible to, in addition to the various methods (Levee systems specifically) engineers utilized in an attempt to mitigate the repercussions. Furthermore, the paper will elaborate on what caused the failure and some recommendations on how to improve the flood and storm protection systems.

Index Terms – Levees, Hurricane Katrina, New Orleans, damage, flooding

Introduction – Before the hurricane hit New Orleans, it was apparent that it was not a hospitable place. The city was surrounded with swamps, the land was either at or below sea level and obviously flood prone. The hurricane protection system, however, was based on a 1965 hazard definition which mainly clarifies the severity of the disaster at the time. [1] Hurricane Katrina was rated as moderate and ‘‘not even close to the big one’’ (Grunwald, 2006), a Category III on the Saffir–Simpson scale, struck causing massive flooding and losses. [3] But what was the cause of this extreme suffrage?

Flood Protection in New Orleans

For flood protection, engineers constructed levees, floodwalls and pumping stations in the bank and much attention was traditionally directed at the levee system to protect the city against Mississippi River which could solitarily drown the city. Not surprisingly, however, the city was vulnerable to storms from locations other than the riverside too. Both nature and engineers can contribute to preventing the damage from such major storms. The barrier island beaches for example, can partially decrease the energy of the storm. The second defence line relied on engineers. [1] But was it an engineering failure?

Failure

The storm surge overtopped the levees in the east side and the counter-clockwise winds exacerbated the situation. The levees were not completed but the storm surge would even pass the design height by a close margin. In addition, according to ILIT (Independent Levee investigation Team), the levees were built from local drudge spoils which make them highly erodible. [1] Also the ring part of the levee was incomplete. Most importantly, as NIST reports, the horizontal size of the storm surge was the dominant cause of the damage. [4]

References

1. F.H. Griffis / Technology in Society 29 (2007), pp. 189–195
2. Journal of Geotechnical and Geoenvironmental Engineering © ASCE / May 2008, 134(5), pp. 668-680
3. Ocean Engineering 37 (2010), pp. 4–12
4. NIST TN 1476, p 35
5. J. Geotech. Geoenviron. Eng., 2008, 134(5): 701-717
6. J. Geotech. Geoenviron. Eng., 2008, 134(5): 718-739
7. J. Geotech. Geoenviron. Eng., 2008, 134(5): 740-761

Investigation

There are various methods to reproduce the conditions including constructing a full-scale model which is very expensive or build a small-scale levee model which is not accurate. However, one of the efficient ways for investigation that this paper will discuss is the centrifuge model which is economical and produces a relatively high accuracy. [2]

Phase 1: Plaquemines Parish

Fig 1. Orleans Parish flood extent (grey) with southern levees (dark red)

Plaquemines Parish (figure 1) is the place where the last portion of the Mississippi River pours into the Gulf of Mexico. The strip to the southeast is protected by a set of levees facing the Mississippi River and a parallel set facing away from the river that are responsible for protecting a number of small communities. These communities that are vividly in constant risk of flooding from Mississippi and also storms, consisted of unincorporated towns and villages. The levees were designed to have crest elevation of about +25 ft. MSL, supported by a set of back levees with crest elevation of around +18 ft. MSL protecting them from storm surges coming from the Gulf.

As the storm approached the coast, the water levels had risen to 20 ft. height, accompanied by constant storm winds which intensified the damage. This led to the overtopping and a dozen full levee breaches in addition to significant erosive damages due to storm-driven waves. (Regional hydrodynamic modelling). So, there are two important problems to be noted: first is the redundancy of constructing levees for small communities that are highly susceptible to flooding. And second is the problem of building the levees in transition, meaning many companies in various phases work on the project with different procedures and timelines resulting in unfinished projects. [5]

Phase 2: The East Flank

As the eye of the storm approached the North, the storm surge continued to expand the water levels within Lake Borgne, and the counter-clockwise winds caused a westward push of the water to the eastern coast. Furthermore, the storm-driven waves added to the hydrodynamic loading and pressure along the levee frontage in the eastern flank. The IPET calculations displays that the storm had maximum wave heights of up to 7 ft., periods of approximately 12–16 s which is significant. (IPET 2007, Vol. IV.)

Disastrous levee failures continued to happen in the southeast corner of the New Orleans East protected basin, and along the northeast edge of the St. Bernard/Lower Ninth Ward protected basin (figure 2). The erosive damage of the levees exacerbated the situation with the levee frontage in northeast of St. Bernard Parish protected basin being eroded disastrously for the lengths of many miles. As a consequence, considerable amount of floodwater passed through the frontages, breached the levee and engulfed both the New Orleans East and the St. Bernard protected
basins.

These two sections shared three similar, fatal characteristics. Both section had long sections of levees that had been constructed out of hydraulic fill materials dredged from excavations of the neighboring shipping channels. These soils laid down at high water contents and could not be compressed together. Second, both sections were “exposed”, meaning their geometrical design was facing the direct storm surges and severe wind-driven waves. And third, there were no mature vegetation on the outward sides to mitigate the repercussions or at least they were not effective.

In addition, two concrete lock structures had been built along the northeastern frontage of the St. Bernard Parish protected basin, one at Bayou Dupre and the other at Bayou Bienvenue.

The embankment soils at this site were mainly comprised of lightweight shell-sand fill made up of a mixture of mollusk shells and sands (and lightweight shells basically) which was employed to reduce embankment load. But what hadn’t been considered was the relatively lower specific gravity of the Mollusk (mainly containing Calcium) in comparison to other mineral soils, making them highly erodible.

As a result of all the reasoning mentioned above, the levees in these two sections could not resist the water pressure for long and easily breached. In contradiction, levees nearby that were mostly constructed with clayey soils, were compressed more easily and despite having been under water pressure for 2 hours in some cases, the erosive damage was considered minimal. ILIT 2006. This problem had been observed by the USACE and made multiple requests for Congressional authorization and funding to armor the levees. But for cost-saving issues they did not hear back before the hurricane hit the city. (ILIT 2006; U.S. Senate Committee on Homeland Security and Governmental Affairs 2006)

Fig 2. Overtopping and through-passage of floodwaters along the east flank

Erosion testing and performance

As shown in figure 3, the graph represents the rate of change of erosion rate (mm/h) per velocity (m/s). The samples from failed sections where the soil was lightly compacted are in the upper left part of the graph (high to very high erodibility) and the sections with more compressibility are located in the lower right part of the graph. (Low to very low erodibility)
Fig 3. Erodibility rate per velocity graph from samples gathered from New Orleans levee embankment. Failed sections are displayed with closed symbols and non-breached (Including highly distressed) sections with open symbols. (ILIT 2006)

                                                         $\frac{\dot{Z}}{u}=\alpha {\left(\frac{\tau –{\tau }_c}{\rho u^2}\right)}^m$

          (1)

The graph can be sketched by knowing a few numbers that equation (1) depend on: $\dot{\mathrm{{\rm Z}}}$

= erosion rate (m/s), u = water velocity (m/s)

, $\tau – {\tau }_c=\mathrm{the\; net\; shear\; stress\; in\; the\; horizontal\; direction }\left(\mathrm{Pa}\right),\mathit{ \rho }=\mathrm{mass\; density\; of\; water}$

(kg/m³), and m and α are the parameters that characterize the soil being eroded.

At the end, both ILIT and IPET confirmed that the two main causes of the failure along the frontages were the overtopping and the subsequent erosive damage.

Phase 3: Storm Surge and the Central Region

By this time, water levels at Lake Borgne had significantly been increased. This resulted in waters in the lake pushing westwards, passing through the east-west trending Mississippi River Gulf Outlet (MRGO)/Gulf Intracoastal Waterway (GIWW). The problem was that the south end in the “T” intersection was closed off by a navigational lock, which caused the water level to swell even more within the trending Inner Harbor Navigation Channel IHNC and then flowed into the Lake Pontchartrain in the north.

Failures at Transitions

In this phase, most of the failures occurred in transitions including 21 breaches and 10 partial failures. There are multiple reasons for dividing flood protection system projects. The most common ones are funding issues and penetrations (utility, roadway, pipe etc.)

Fig. 4 portrays an example of a transition of a full height levee section to a half-height earthen embankment topped by a sheetpile supported by a floodwall. The left levee section is approximately 1 ft. higher than the neighboring section. In this case, the two sections were connected by a transition section consisting of a simple sheetpile wall section which had a lower height in comparison to the other two.

Fig 4. A breach caused by a transition between

the two levee sections

And the overtopping probably started from this lower point. This overtopping flow eroded a ditch behind the back side of the sheetpile wall, and as a result the week supported sheetpile wall could not resist the lateral force of the floodwaters.

Failure at Penetrations