Polder Flood Simulations for Greater

New Orleans
Hurricane Katrina August 2005

July 2007
 Polder Flood Simulations for Greater
New Orleans
Hurricane Katrina August 2005

M. Kok, M. Aalberts, B. Maaskant and L. de Wit

July 30, 2007

Flood simulations Greater New Orleans July 2007

Preface
The objective of the project is to discern the contributions of various water inputs that
combined led to the flooding of the three polders that comprise Greater New Orleans
(GNO) during hurricane Katrina August 2005. State-of-the art computer models are used
to assess these contributions. Like all computer models simulations, the output quality is
dependent upon the input parameters and accuracy of the base grid. Our team has
endeavored to use the best available data in this simulation.

The project team consists of M. Kok Ph.D. MSc. (project manager), B. Maaskant MSc,
M. Aalberts BSc (all from the Delft University of Technology, Netherlands) and L. de
Wit MSc (Svasek B.V., Netherlands). The project was independently reviewed by
Professor J.K.Vrijling MSc (Delft University of Technology).

Delft July 30, 2007.

i
Flood simulations Greater New Orleans July 2007

Contents
1 Introduction ..................................................................................................................... 1
1.1 Hurricane Katrina...................................................................................................... 1
1.2 Objectives ................................................................................................................. 1
1.3 General modelling approach ..................................................................................... 1
1.4 Definitions................................................................................................................. 3
2 Input data and Model setup .............................................................................................. 4
2.1 Assumptions.............................................................................................................. 4
2.2 Data ........................................................................................................................... 4
2.3 General Model setup ................................................................................................. 6
3 Orleans Metro .................................................................................................................. 9
3.1 Calibration of the model ........................................................................................... 9
3.2 What-if scenarios .................................................................................................... 17
4 New Orleans East ........................................................................................................... 26
4.1 Calibration of the model ......................................................................................... 26
4.2 What-if scenarios .................................................................................................... 32
5 Saint Bernard ................................................................................................................. 37
5.1 Calibration of the model ......................................................................................... 37
5.2 What-if scenarios .................................................................................................... 43
5 Conclusions .................................................................................................................... 51
6 References ...................................................................................................................... 52

ii
Flood simulations Greater New Orleans July 2007

1 Introduction
1.1 Hurricane Katrina
On the morning of August 29, 2005, Hurricane Katrina struck southeast Louisiana and
triggered what would become one of the worst disasters ever to befall an American city.
The storm overtopped levees and floodwalls throughout southeast Louisiana and also
caused the levees and floodwalls in the New Orleans area to fail or breach in more than
50 locations. Water rushed into New Orleans and flooded over 80 percent of the city -
more than ten feet deep in some neighborhoods [9].

Flooding of the city of New Orleans was caused by several sources (or causes): breaches,
overtopping (of the levees and flood control structures) and rainfall. The question to be
addressed in this report is to determine the relative contribution of the flooding sources
for selected sites within each polder or bowl.

1.2 Objectives
The objectives of this research project include:
1. The use of computer models to generate flood simulations for the three polders or
bowls (Orleans Metro bowl, New Orleans East bowl, and St Bernard bowl) that
comprise the greater New Orleans area, utilizing the best available input data;
2. The assessment of the relative contribution of the main causes of flooding
(breaches, overtopping, and rainfall) in the three polders utilizing water depth
graphs (hydrographs) for selected locations.

1.3 General modelling approach

The flooding of New Orleans is simulated with the dedicated software package Sobek-
1D2D [7]. Sobek-1D2D was developed by WL-Delft Hydraulics and is used worldwide
in flood simulations. It has proved to be a reliable and stable program to simulate floods.
It is the standard model used in the Netherlands for flood simulation [6]. Flooding is
simulated with a combination of 1D (one-dimensional) and 2D (two-dimensional)
elements; canals and boundaries are modeled as 1D elements and the flooding itself is
modeled as a 2D process. Sobek-1D2D uses the robust numerical Delft Scheme to model
the water motion (equations of Saint Venant) in the 1D and 2D elements. The results
show detailed spatial information about water levels, velocities and rise rates over time
and arrival times at desired locations.

The flooding of New Orleans is simulated in the three bowls (figure 1.1 - Orleans Metro
bowl, New Orleans East bowl and St Bernard bowl) with three separate and independent
model grids and simulations utilizing the Sobek1D2D computer flooding model.

1
Flood simulations Greater New Orleans July 2007

New
Orleans
East

Orleans
Metro Bowl St. Bernard

Figure 1.1 The three bowls that comprise Greater New Orleans.

In general, the following approach was used. First, a grid data base model of the three
polders was built. Two pumping stations in the Orleans Metro bowl are included in the
model, because they were (partly) functioning during Katrina. Next, calculations with the
three models were made using all sources of the flooding, and the results calibrated using
available data of the flooding (mainly eye witness reports). The next step was the creation
of ‘what-if scenarios’, for example what would have happened without the breaches, and
so on. The following sources of the flooding are taken into account for the three bowls:
- Rainfall and pumping stations
- Overtopping of the levees and flood control structures
- Breaches of levees and/or floodwalls.
We have chosen to present the results at specific locations (sites). However, the model
results are available for all sites. The impact of all scenarios was assessed with the
Delft1D2D model, with one exception. The impact of the scenario “only rainfall and
100% pumps available” was assessed using a spreadsheet model, which includes rainfall
and the capacities of the pumping stations. The Delft1D2D model is able to include the
sewage system and the pumping stations, but this could not be accomplished within the
deadline as indicated by counsel.

2
Flood simulations Greater New Orleans July 2007

1.4 Definitions
The following definitions are used in this report:

Water level Water height above or below NAVD88

Water depth Water height or elevation above ground level

Rise rate The height over which the water rises in one hour

Arrival time Point in time where the first water arrives at a

certain location

Overtopping Water that flows over a levee or flood control

structure due to an elevated water level which is
higher than the crest level

Digital Elevation Model (DEM) Model in which the considered area is divided in
squares to which the elevation, with reference to
NAVD88, is assigned

IPET Interagency Performance Evaluation Task Force

NOAA National Oceanic & Atmospheric Administration

ILIT Independent Levee Investigation Team

Grid Computational layer in which the considered area is

divided in squares, for each square the water motion
is calculated

Polder or bowl An area that is below sea level and is enclosed by

an uninterrupted chain of water retaining structures,
levees and/or high grounds

MRGO Mississippi River Gulf Outlet

IHNC Inner Harbor Navigation Canal

GIWW Gulf Intracoastal Water Way

NAVD88 NAVD88 2004.65

3
Flood simulations Greater New Orleans July 2007

2 Input data and Model setup

2.1 Assumptions
In the simulation of Katrina flooding of New Orleans the most important processes of the
flooding are incorporated. Some assumptions however remained unavoidable. The
assumptions used in the simulation are:
• Overtopping by waves is not taken into account directly, but when erosion
occurred due to wave overtopping this is taken into account in the crest elevation.
The water volume from wave overtopping is small compared with the overflow
volume from overtopping when the still water level is higher than the crest level.
• Sewers, ditches and other small waterways are not represented in the model.
These waterways may have resulted in faster initial spreading of the water, but
have negligible influence on the final water levels inside the three bowls.
• Buildings and other structures are not incorporated. They may have blocked water
flow and changed flow directions. This influence is local, the overall influence of
buildings and structures is taken into account in the bottom roughness (friction to
flow) used in the model.
• No distinction is made between house blocks and streets in the elevation and
roughness used in the model. The results show a general picture. At locations of
houses flow velocities will be overestimated; at locations of streets the velocities
will be underestimated.
• The interaction between surface water and groundwater is neglected. It is assumed
that the soil is fully saturated, which seems to be a good assumption for the
situation just before Katrina.
• The direct impact of wind is not taken into account.
• Under seepage below or through levees or flood control structures is not taken
into account. All evidence is that under seepage volumes are negligible in
comparison to the final water volumes within each bowl.

2.2 Data
All input data such as topography, surge hydrographs, rainfall, levee and floodwall crest
heights, size and depth of breaches, levee breach times and pump station operation was
acquired from US experts. The interior bowl or flood model was built on these data. The
project team has, as far as possible, reviewed whether the data is consistent with other
available data in the various reports of IPET, Team Louisiana and ILIT (see reference
list). We know that there are uncertainties in the data, but we used, according to our
knowledge, the best available data.

4
Flood simulations Greater New Orleans July 2007

Data used in the flood simulation:

- Digital elevation model (DEM) available for the whole area with 15ft resolution,
the resolution around levee crests is one feet, as supplied by Mr. Chad Morris, a
professional surveyor.
- Surge hydrographs (graphs of water level variation over time due to hurricane
Katrina) were obtained for ten pertinent locations adjacent to the polders from Dr
Paul Kemp, a storm surge expert and oceanographer. See figure 2.1 for the
locations and table 2.1 for the period covered by the time series.
- Information on the breach locations and sizes was obtained from Geographical
Information System (GIS) data delivered by Mr. Chad Morris. Timelines for the
breaches were obtained from the Team Louisiana Report [3].
- Time series of rainfall are delivered by Dr. Lee E. Branscome (CCM,
Climatological Consulting Corporation) for the six areas numbered 1 to 6 in
figure 2.2. For the three bowls three different time series are used as input in the
model. The time series of Orleans Metro is equal to time series one, the time
series of New Orleans East consists of the (area) weighted average of time series
two and three and the time series of St. Bernard consists of the (area) weighted
average of time series four, five and six.
- Eye witness reports from the IPET and Team Louisiana reports.

In the following sections the input data on the breaches for the three bowls will be
discussed in detail.

b c
a f
g
h
d
e
i

Figure 2.1 Locations of available surge hydrographs used in this modeling study.

5
Flood simulations Greater New Orleans July 2007

Table 2.1 Start and end point in time for the surge hydrographs.
Location Start End
a 17th Street Canal 2005-08-25 07:30 2005-08-30 05:001
b London Avenue Canal 2005-08-25 07:30 2005-08-30 05:001
c Airport 2005-08-25 07:30 2005-08-30 05:001
d IHNC 2005-08-25 07:30 2005-08-30 13:001
e IHNC near lock 2005-08-25 07:30 2005-08-30 05:001
f GIWW B. Sauvage 2005-08-25 07:30 2005-08-30 13:001
g GIWW Paris Road 2005-08-25 07:30 2005-08-30 13:001
h MRGO B. Bienvenue 2005-08-25 07:30 2005-08-30 13:001
i MRGO B. Dupre 2005-08-25 07:30 2005-08-30 13:001
j MRGO Shell Beach 2005-08-25 07:30 2005-08-30 13:001

1 4
5

6
Figure 2.2 The six subdivisions of greater New Orleans as used to input rainfall data.

2.3 General Model setup

The Sobek-1D2D numerical model we used operates on the metric system of
measurement. We then convert this data to the imperial system for purposes of this report.
Because there is no interaction between the three bowls St. Bernard, New Orleans Central
and New Orleans East, they are modeled separately.

1
The surge level at 09-03-2005 0:00 is 2.5 ft (0.762 m). All simulations end before this enddate.
Simulations between the end of the surge hydrograph data and the enddate were linearly interpolated.

6
Flood simulations Greater New Orleans July 2007

Grid
The two dimensional (2D) area in the model consists of a grid with computational cells
with 50m (164 ft) sides. In this way each bowl consists of approximately 80,000 cells.
The elevation of each cell is based on the DEM data. Levee crest the heights, given the
sometimes marked variations over short distances, are carefully inserted into the model.
Sobek-1D2D calculates the water levels and velocities in every grid cell.

Simulation period
All simulations start at 28th August 2005 0:00am and continue for at least two days.
There are no simulations after 2nd September 2005 12:00pm.

Overtopping
Around every polder the water bodies are modeled using the available surge hydrographs
as presented in figure 2.3. As soon as the water level gets higher than the crest height of a
part of the levee or flood control structure, overtopping begins and water flows into the
polder. At breaches a gap in the levee or flood control structure is modeled.

Rainfall
The time series used for rainfall are shown in figure 2.4. The model assumes that the
rainfall occurs uniformly over each subsection (figure 2.2).

Water levels
20,00
18,00
1) 17th Street
16,00 4) IHNC
14,00 9) B. Dupre
Elevation (ft)

12,00
10,00
8,00
6,00
4,00
2,00
0,00
8-25-05 0:00 8-26-05 0:00 8-27-05 0:00 8-28-05 0:00 8-29-05 0:00 8-30-05 0:00 8-31-05 0:00

Time
th
Figure 2.3 Surge hydrographs for 17 street, IHNC and B. Dupre

7
Flood simulations Greater New Orleans July 2007

Rainfall used in model

2.5
Rainfal Orleans
Metro Bowl
Rainfall New
2
Orleans East
Rainfal St.
Rainfall [inch/hr]

Bernard
1.5

0.5

0
29/08/05 0:00 29/08/05 6:00 29/08/05 12:00 29/08/05 18:00 30/08/05 0:00
Date

Figure 2.4 Rainfall time series used in model

8
Flood simulations Greater New Orleans July 2007

3 Orleans Metro
3.1 Calibration of the model
Figure 3.1 shows the Digital Elevation model (DEM) of the Orleans Metro bowl. The
arrows and numbers in squares correspond to the breaches given in table 3.1 and the
arrows and numbers in circles correspond to the pumps which where available during
Katrina (table 3.2). These breaches and pumps in combination with the rainfall and water
levels as presented in section 2.3 are the basis of the model for the Orleans Metro bowl.

3
1
4
5
2
8
10 6

9 7

Figure 3.1 DEM Orleans Metro bowl, including breach locations (numbered squares) and pump stations
(numbered circles). Note how this polder actually consists of several small bowls separated by natural
ridges and/or railway or road embankments.

9
Flood simulations Greater New Orleans July 2007

Table 3.1: Information on the Breaches in Orleans Metro bowl

Time breach
Nr. Description Width Elevation height
development
6:00 - 7:002
Sill elevation -3.0 ft
1 17th street 458 ft (139.6 m)1 Major breach
(-0.91 m)1
at 9:003
Sill elevation +9.5 ft
300 ft (91.4 m)
(+2.90 m)
2 Orleans canal 7:30 - 8:00
Sill elevation +9.1 ft
70 ft (21.3 m)
(+2.77 m)
Sill elevation -3.5 ft
3 London West 400 ft (121.9 m)1 7:00 - 9:004
(- 1.07 m)1
Sill elevation -20.0 ft
4 London East 231 ft (70.4 m)1 7:00 - 9:003
(-6.10 m)1
Sill elevation +5.0 ft
5 Railroad 37 ft (11.3 m)2 4:30 - 5:002
(+1.55 m)2
Sill elevation +1.0 ft
6 IHNC West floodwall 148 ft (45.1 m)1 7:00 - 8:002
(+0.30 m)1
Sill elevation +1.0 ft
7 IHNC West levee 200 ft (61.0 m)1 5:00 - 6:005
(+0.30 m)1
Over a length of
approximately Differences in
Overtopping from
8 12000 ft. Position sections, based on 7:00 - 7:306
IHNC
and width based GIS data1
on GIS data1

Table 3.2: Information on the pumping stations in Orleans Metro bowl7

Rated Pumped Projected
Capacity
Nr. A) East Bank (E-3) Capacity During Capacity as
(m3/s)
(cfs) Katrina? of 1 June ‘06
9 3) OP 19 – PS 19 3650 103.4 YES 71%

10 10) OP I 10 – PS I 10 860 24.4 YES 100%

1
Source: GIS data with positions, widths and sill elevation heights of the breaches [delivered by Chad
Morris 28-6-2007]
2
Source: Team Louisiana report table 3, page 54 [4].
3
Source: IPET volume IV: The Storm, page 169 [1].
4
Source: IPET volume V: The Performance Levees and Floodwalls, page 40 [2]
5
Source: IPET volume V: The Performance Levees and Floodwalls, page 53 [2]
6
Source: Independent Levee Investigation Team (ILIT), page 7-6 [5]
7
Source: IPET volume VI: Performance Interior Drainage and Pumping, page 31 [3]

10
Flood simulations Greater New Orleans July 2007

Calibration results

Figure 3.2 shows the progress of the simulated flood for the Orleans Metro bowl over
time. The first figure is a snapshot August 29th at 5:00am; it shows only minor rainfall
and a small breach in the IHNC at the location where the railroad crosses the levee. In the
next hour the breach at the 17th street starts. This is consistent with the IPET, Team
Louisiana and the ILIT reports.
The water levels near the 17th street breach were lowered because of blocking debris at
the entrance of the 17th street canal.

According to IPET [1, page 181] site 11 started to flood at 09:00 - 10:30am, and the
results of this flood simulation model do confirm this IPET observation.

An eyewitness [1, page169] near the 17th Street Canal breach heard loud noises at
09:00am. Other eyewitness reports suggest that the breach occurred between 9:10am and
9:25am, so for the model we assume a 9:00am breach.

IPET reports [1, page 171] that no flooding was observed at site 14 until about
12:30pm on Monday afternoon. Our flood simulation model suggests that flooding
started at approximately 12:00pm. This variance is within the ‘error band’ of eyewitness
reports who would have been dealing with rain and poor visibility, wind, and the general
distraction of the noise and circumstances of a hurricane passing overhead.

Site 19 starts flooding after 1:00pm, where eyewitness reports in IPET [1, page 173] state
that the flood started at 1:00pm at this location. Again this variance is within the ‘error
band’ of eyewitness reports who would have been dealing with rain and poor visibility,
wind, and the general distraction of the noise and circumstances of a hurricane passing
overhead. Thus we feel there are really only minor differences between eyewitnesses and
the model results, which adds to our confidence in the results of the model.

11
Flood simulations Greater New Orleans July 2007

05.00 h 2005-08-29 06.00 h 2005-08-29

06.30 h 2005-08-29 07.00 h 2005-08-29

07.30 h 2005-08-29 08.00 h 2005-08-29

12
Flood simulations Greater New Orleans July 2007

11

08.30 h 2005-08-29 09.00 h 2005-08-29

11

0930 h 2005-08-29 10.00 h 2005-08-29

14 14

11.00 h 2005-08-29 12.00 h 2005-08-29

13
Flood simulations Greater New Orleans July 2007

19

13.00 h 2005-08-29 15.00 h 2005-08-29

17.00 h 2005-08-29 19.00 h 2005-08-29

21.00 h 2005-08-29 23.00 h 2005-08-29

Figure 3.2 Progressing snapshots of the computed water depth inside Orleans Metro bowl

14
Flood simulations Greater New Orleans July 2007

The maximum water depth that occurred in the Orleans Metro bowl is shown in figure
3.3. Figure 3.4 shows the comparison between hydrographs computed by our modeling
and hydrographs presented by IPET for the Orleans Metro bowl. IPET apparently based
its hydrographs on eye witness accounts, stopped clock times and time/date stamped
digital photographs.

There are differences between the model results and the IPET eye witness reports at some
locations, but at other locations the results are in agreement. It is important to note that
eye witness reports are subjective, and there will always be some uncertainty in eye
witness reports given the chaos during a strong hurricane passage, driving rain and wind
and the vagaries of the human memory. Additionally, during such extreme events, input
data such as the exact water level in the 17th Street before and after the breach, just aren’t
recoded and generally hard to determine with out direct measuring equipment at the
location that can function through a strong storm. Therefore, the differences between the
model results and the calibration data are fully acceptable. For the purpose of this study
we believe that the model results are very credible.

Figure 3.3 Maximum computed flood depths for Orleans Metro bowl as a consequence of the hurricane
Katrina flooding event.

15
Flood simulations Greater New Orleans July 2007

IPET versus computed Hydrographs

4,0

3,0

2,0
Floodwater Level (ft, NAVD88)

1,0

0,0

-1,0

-2,0

-3,0
8-29-05 0:00 8-29-05 6:00 8-29-05 12:00 8-29-05 18:00 8-30-05 0:00 8-30-05 6:00 8-30-05 12:00

17 East IV-168 London W. IV-175 Time

Computed 17th Computed London W

IPET versus computed Hydrographs

4,0

3,0

2,0
Floodwater Level (ft, NAVD88)

1,0

0,0

-1,0

-2,0

-3,0
8-29-05 0:00 8-29-05 6:00 8-29-05 12:00 8-29-05 18:00 8-30-05 0:00 8-30-05 6:00 8-30-05 12:00
Time
London E. IV-179 IHNC W. IV-184
Computed London E Computed IHNC

Figure 3.4 Comparison between our computed hydrographs and those presented by IPET, Orleans Metro
bowl

16
Flood simulations Greater New Orleans July 2007

3.2 What-if scenarios

In this paragraph several “what-if scenarios” for polder Orleans Metro will be presented.
These scenarios will provide more information on the contribution to the flood by the
different causes of the flood, like overtopping versus rainfall versus breaching and so on.

Eight sites were chosen in the Metro bowl, some being at very low elevation in the center
of bowls, while others were chosen close to internal ridges, i.e., higher ground (figure
3.5). In this way the reader will hopefully better understand the dynamics of the internal
flooding of the various New Orleans bowls.

3
1 2

4
8

7
6

Figure 3.5 Locations utilized for scenario development Orleans Metro bowl.

For each of these eight sites within the Orleans Metro bowl, five different flooding
scenarios were modeled, namely:
1) A simulation with all sources of flooding water including rain, pumping,
overtopping and breaches. In the following figures the graph or line
representing this scenario is labeled ‘All causes’.
2) A simulation without breaches and overtopping but including rainfall
and the available pumps operating as they did during hurricane Katrina.
In the following figures the graph or line representing this scenario is
labeled ‘Rain minus available pumps’.
3) A simulation with no breaches, without overtopping, but including
rainfall with the pumps running at 100% efficiency. In the following

17
Flood simulations Greater New Orleans July 2007

figures the graph or line representing this scenario is labeled ‘Rain

100% pumps’.
4) A simulation with only overtopping and including rainfall with the
available pumps operating as they did during hurricane Katrina; without
breaches. In the following figures the graph or line representing this
scenario is labeled ‘Overtopping’.
5) A simulation with all sources of flood water but without the
contribution of the 17th Street breach. In the following figures the graph
or line representing this scenario is labeled ‘No breach 17th street’.

Description of the scenarios

Scenario 1 (all causes) is the best fit to reality during Katrina, see section 3.1. This
scenario takes into account breaches, overtopping, rainfall and the two major pumping
stations which were available during Katrina; pumping stations OP 19 and OP I-10 (see
table 3.2).

Scenario 2 (rain minus available pumps) represents a scenario with rainfall and only the
available pumping stations (OP 19 and OP I-10) during Katrina. The pumping stations
have been taking into account by deducting the pump capacity per hour from the rainfall
data. This run assumes that once the rain quits, the pumps stop pumping.

Scenario 3 (rain and 100% pumps) represents a scenario with rainfall and the pumping
stations in the Metro bowl. It is assumed that they are all working at full capacity, and
this represents a lower bound estimate of water levels due to rain. The pumping stations
have been taking into account by deducting the pump capacity per hour from the rainfall
data. After the rain has stopped, the influence of the pumping stations is estimated with a
spreadsheet calculation. This estimate is uncertain, but it will represent a credible
estimate of the order of magnitude.

Scenario 4 (overtopping) takes into account overtopping, rainfall and the pumping
stations which were available during Katrina; pumping stations OP 19 and OP I-10. The
pumping stations have been taking into account by deducting the pump capacity per hour
from the rainfall data. This run assumes that once the rain quits, the pumps stop pumping.

Scenario 5 (no breach 17th Street) takes into account all breaches without the flood
contribution from the 17th Street breach. It includes also overtopping, rainfall and the
pumping stations which were available during Katrina; pumping stations OP 19 and OP
I-10. The pumping stations have been taking into account by deducting the pump
capacity per hour from the rainfall data. This run assumes that once the rain quits, the
pumps stop pumping.

18
Flood simulations Greater New Orleans July 2007

For each location or site given in figure 3.5, the five simulations are plotted below. Note
that given the dynamics of the flooding event, the scenarios “Rain minus available
pumps” and “Overtopping” are sometimes identical and one curve may mask the other in
the figures.

Water level location 1

4,00

3,00

2,00

1,00

0,00
1. All causes
Waterlevel (ft)

-1,00 2. Rain minus available pumps

3. Rain 100% pumps

-2,00 4. Overtopping

5. No breach 17 th street
-3,00

-4,00

-5,00

-6,00

-7,00

-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00

Time

Figure 3.6 Water levels location 1 Orleans Metro. The scenarios ‘Rain minus available pumps’ and
‘Overtopping’ are identical

This site or location (1) is approximately halfway between the 17th Street and Orleans
Avenue canals. Here rainfall collects and floods areas lower than minus 5 ft NAVD88.
As one would expect the overtopping at the east side levee (at IHNC) has no influence in
this area.

Scenario 3, rain with pumps working at 100% capacity while it rains, is a lower bound or
bracket for the influence of rain on the internal flood levels, and reveals that areas of the
bowl adjacent to site (1) lower than -5.9 ft NAVD88, would have flooded during the
storm from rainfall, even if the pumps had worked at maximum capacity. However, if
there would have been no overtopping or breaching then this area would have been
pumped dry once the rainfall ceased.
If the breach at the 17th Street canal had not occurred, the other breaches would have
filled this area, but somewhat later.

19
Flood simulations Greater New Orleans July 2007

Thus in this section of the Orleans Metro bowl during Katrina with the pumps operating
at about 50% capacity and being shut down early in the morning because the pump
houses flooding after the breaches occurred, 1.8 feet of the 10 feet of standing water was
attributable to the rainfall, six inches of the initial flooding being attributable to the
pumps not operating at full capacity and at least 7.5 feet (75%) due to the breaches.

Water level location 2

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumps

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.7 Water levels location 2 Orleans Metro. The scenarios ‘Rain minus available pumps’, ‘Rain
100% pumps’ and ‘Overtopping’ are identical

Site 2 is located on higher ground between the Orleans Ave and London Avenue canals.
Rainfall does not accumulate at site 2 because of its relatively high elevation (-0.5 ft
NAVD88). As one would expect the overtopping at IHNC has no influence at this
location.

Thus at this location of the Orleans Metro bowl during Katina there was no flooding
attributable to the rain or from the overtopping along the INHC. All the flooding (100%)
came from the breaches.

20
Flood simulations Greater New Orleans July 2007

Water level location 3

4,00

3,00

2,00

1,00

0,00
1. All causes
Waterlevel (ft)

-1,00 2. Rain minus available pumps

3. Rain 100% pumps

-2,00 4. Overtopping

5. No breach 17 th street
-3,00

-4,00

-5,00

-6,00

-7,00

-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.8 Internal flooding levels location 3, Orleans Metro bowl. The scenarios ‘Rain minus available
pumps’ and ‘Overtopping’ are almost identical

Site 3 is located between the London Avenue canal and the INHC. At this site rainfall
only accumulates in areas deeper than -5 ft NAVD88. If the pumps had operated at 100%
capacity during the storm rainfall flooding would have been restricted to areas lower
lying that minus 5.5 ft NAVD88

The influence of the 17th Street breach is less than at location 1 & 2 and only amounts to
approximately six inches of the initial flooding or 5%.

There is minor overtopping south of Interstate 10 from the INHC as can be discerned on
figure 3.8, the overtopping line lies just above the rainfall minus the pumps operation line.

21
Flood simulations Greater New Orleans July 2007

Water level location 4

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumps

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.9 Water levels location 4 Orleans Metro

Location 4 is within a bowl within the Orleans Metro bowl bounded by the Mississippi
River levee, the INHC and the Metairie-Gentilly Ridge

At location 4, rainfall influences areas lower than -1 ft NAVD88. The influence of the
17th Street breach is less then at location 1 & 2 and only marginal.
In this part of the Central bowl overtopping is more important than elsewhere in the
polder. Areas lower than 0 ft NAVD88 would have flooded by this scenario.

Pump OP I-19 situated near the IHNC was available during Katrina. If the pumps had
been operating at full capacity while it rained the flooding would have been about one
feet lower, assuming the pumps shutdown once the rain ended. However, it is important
to know what the flooding would have been if there had been no breaches, no
overtopping and the pumps had continued to work after the rain ceased and dewatered the
bowl. This is given in the scenario “Rain 100% pumps”.

22
Flood simulations Greater New Orleans July 2007

Water level location 5

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumps

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.10 Water levels location 5 Orleans Metro, the scenarios ‘Rain minus available pumps’ and
‘Overtopping’ are identical

Location 5 is located in a sub-bowl south of the Metairie-Gentilly Ridge. At location 5

rainfall influences area’s lower than -0.5 ft NAVD88. Overtopping along the INHC has
no influence on flooding at location 5, while the London Avenue breaches have some
important influence on the water level.

Location 6 is south of location 5 in the same sub bowl but is higher land. Rainfall
accounts for about nine inches of water and based on the internal flooding model most of
the flooding comes from the 17th Street Canal breach.

23
Flood simulations Greater New Orleans July 2007

Water level location 6

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumps

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.11 Internal flood water levels location 6 Orleans Metro bowl, the scenarios ‘Rain minus available
pumps’ and ‘Overtopping’ are identical

Pump OP 2 situated in the area of location 5 was not available during Katrina. However,
if the pump had worked and the only flood water had been rainfall then pump OP 2
would have dewatered this part of the Metro bowl in about 15 hours.

Water level location 7

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumpsl

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.12 Water levels location 7 Orleans Metro, the scenarios ‘ Rain minus available pumps’, ‘Rain
100% pumps’ and ‘Overtopping’ are identical

24
Flood simulations Greater New Orleans July 2007

Location 7 is along the western edge of the Metro bowl located on relatively high land
compared to say sites 5 and 6. The elevation level of Location 7 is 0.8 ft NAVD88. This
site experiences a short period of standing rain water about three inches deep at the
maximum and floods predominately from the 17th Street Canal breach. It’s interesting to
note that the water from the 17th Street breach took almost 24 hours to reach this site and
flooded thereafter at a rate of about one foot per day until the water level in the city
attained equilibrium with that in Lake Pontchartrain in the early hours of the 2nd
September 2005.

Location 8 is on the Metairie ridge (figure 3.13), and at 3.1 ft NAVD88 is some of the
highest land in the bowl. The model does indicate that for a period of 10 to 12 hours
water did flow over this site, up to about six inches in depth as the flood waters from the
17th Street breach moved south towards sites 5, 6 and 7.

Water level location 8

4,00

3,00

2,00

1,00

0,00
Waterlevel (ft)

-1,00

-2,00

-3,00

-4,00

-5,00
1. All causes

-6,00 2. Rain minus available pumpsl

3. Rain 100% pumps

-7,00 4. Overtopping

5. No breach 17 th street
-8,00
29-08-05 0:00 29-08-05 12:00 30-08-05 0:00 30-08-05 12:00 31-08-05 0:00 31-08-05 12:00 1-09-05 0:00 1-09-05 12:00 2-09-05 0:00 2-09-05 12:00 3-09-05 0:00
Time

Figure 3.13 Water levels location 8 New Orleans Metro bowl, all the lines are (almost) identical

25
Flood simulations Greater New Orleans July 2007

4 New Orleans East

4.1 Calibration of the model
Figure 4.1 displays the Digital Elevation model (DEM) for the New Orleans East bowl.
The star indicates the approximate location of the internal flooding hydrograph used to
check the model results. The arrows and numbers correspond to the breaches given in
table 4.1. These breaches in combination with the rainfall and water levels as given in
section 2.3 and are the basis of the model inputs, along with the DEM, for the New
Orleans East bowl.

Maxent
levee 2

3 Railroad
A 9
*
5 6
1
4 78

Figure 4.1 DEM New Orleans East bowl including breach (numbers) and the internal flooding hydrograph
location (A). Note that this bowl is divided east-west by a railroad embankment and the north-south Maxent
levee

26
Flood simulations Greater New Orleans July 2007

Table 4.1: Information on the Breaches in New Orleans East

Time breach
Nr. Description Width Elevation height
development
Sill elevation +4.5 ft
1 IHNC East Railroad 61 ft (18.6 m)1 4:30 - 4:452
(+1.37 m)1
Sill elevation +5.5 ft
2 Railroad East 24 ft (7.3 m)1 5:00 - 5:15
(+1.67 m)1
Sill elevation +9.8 ft
3 Airport 33 ft (10 m)1 7:00 - 7:30
(+3.00 m)1
IHNC East South of Sill elevation +4.9 ft
4 25 ft (7.5 m) 5:00 - 5:15
RR (+1.50 m)1
Back levee, East of Sill elevation +3.5 ft
5 112 ft (34 m)1 5:30 - 6:002/3
Michaud Canal 1 (+1.07 m)1
Back levee, East of Sill elevation +8.4 ft
6 34 ft (10.4 m)1 5:30 - 6:002/3
Michaud Canal 2 (+2.56 m)1
Citrus Back levee near Sill elevation +11 ft
7 406 ft (123.5 m)1 7:00 - 7:304
Elaine St. West (+3.35 m)1
Citrus Back levee near Sill elevation +8.5 ft
8 615 ft (187.5 m)1 7:00 - 7:304
Elaine St. East (+2.59 m)1
Approximately
6000 ft in length Differences in
extending from B. sections extending
Overtopping New
Sauvage towards from B. Sauvage
9 Orleans Back levee 5:30 - 6:002/3
the Michaud towards Michaud
(erosion crest)
Canal. Position Canal; Erosion levels
and width based based on GIS data1
on GIS data1

1
Source: GIS data with positions, widths and sill elevation heights of the breaches [delivered by Chad
Morris 28-6-2007]
2
Source: Team Louisiana report table 3, page 54-55 and timeline page 93-94 [4]
3
Source: Independent Levee Investigation Team (ILIT), page 7-5 [5]
4
Source: Independent Levee Investigation Team (ILIT), page 7-6 [5]

27
 Flood simulations Greater New Orleans July 2007

Calibration results

The image series below shows the progression of the simulated internal flood over time
(figure 4.2). The first image is a snapshot from August 29th at 5:00am; it shows only
minor rainfall and a small breach in the IHNC at the location where the railroad goes
through the levee. In the next hour the Back Levee begins to be overtopped and starts to
erode. This is consistent with the IPET, Team Louisiana and the ILIT reports, all of
which stated that the overtopping and erosion of the New Orleans East back levee started
between 5:00 and 6:00am. At 7:00am the water levels in the GIWW and IHNC start to
overtop the Citrus Back and IHNC levees. In the Team Louisiana and ILIT report this
overtopping initiated at approximately 7:00am, while the IPET report states that the
overtopping started earlier at 6:00am.

In the IPET report an eyewitness states that the water was flowing over the east-west
railroad line from south to north between 6:30 and 7:00am1. In the simulation the water
crosses the railroad line between 7:30 and 7:45am. One of two eyewitnesses2 situated at
the west side of the New Orleans East bowl states that between 6:30 and 7:30am the
water started to rise rapidly. The other eyewitness states that the flood started between
7:30 and 8:00am. The simulated flood shows that the water started to rise at those
locations between 7:30 and 8:00am.
According to eyewitness reports in the northwestern corner of the New Orleans East
bowl3 the flood started between 8:00 and 10:00am. In the simulation the flooding starts at
these locations between 7:45 and 8:30am. Again it’s important to note that eye witness
reports are very subjective, and there will always be some uncertainties given the chaos
during a strong hurricane passage; low visibility due to driving rain and wind; and the
vagaries of the human memory especially in highly stressful conditions.

Water depth (m) Water depth (m)

2005-08-29 2005-08-29
05:00h 05:30h

1
Source: IPET volume IV: The Storm, page 191-192 eyewitness at location 3 [1]
2
Source: IPET volume IV: The Storm, page 191-192 eyewitness locations 5 and 6 [1]
3
Source: IPET volume IV: The Storm, page 191-192 eyewitness locations 7 to 11 [1]

28
 Flood simulations Greater New Orleans July 2007

Water depth (m) Water depth (m)

2005-08-29 2005-08-29
06:00h 06:30h

Water depth (m)

Water depth (m)

2005-08-29 2005-08-29
07:00h 07:30h

Water depth (m) Water depth (m)

2005-08-29 2005-08-29
08:00h 08:30h

29
 Flood simulations Greater New Orleans July 2007

Water depth (m) Water depth (m)

2005-08-29 2005-08-29
09:00h 09:30h

Water depth (m) Water depth (m)

2005-08-29 2005-08-29
10:00h 12:00h
Figure 4.2 Progressing snapshots of the computed water depth inside New Orleans East

30
Flood simulations Greater New Orleans July 2007

Figure 4.3 Maximum computed internal flood depths as a consequence of the hurricane Katrina related
flooding, New Orleans East bowl.

Figure 4.3 shows the calculated maximum internal flood depths in the New Orleans East
bowl as a consequence of the hurricane Katrina related flooding.

In the western part of the bowl IPET constructed a hydrograph on basis of eyewitness
account and surveyed elevations at location 'A' in figure 4.1. This hydrograph is
presented in figure 4.4. The simulated water level based on our modeling is also shown in
this figure. Between 11:00am and 12:00pm the water levels diverge, in the simulation the
second peak in the water level is caused by the flood water coming from the overtopping
of the New Orleans East Back levee. The model results thus reveal some of the
complexities of the flooding event not discernable from measurements collected some
time after the event.

31
Flood simulations Greater New Orleans July 2007

IPET & Computed Flood Hydrographs New Orleans East

3

2
IPET hydrograph

1 Computed
Interior Water level (ft, NAVD88)

-1

-2

-3

-4

-5

-6

-7
7:00 8:00 9:00 10:00 11:00 12:00
Time on 8-29-05

Figure 4.4 Comparison model simulated and IPET flood hydrographs, New Orleans East bowl.

As discussed previously, in extreme events such as Katrina it is difficult if not impossible

to get direct measurement of some input data such as water levels at the time of a breach,
especially right at the breach location. However, the differences between the computer
model results and the calibration data are fully acceptable. For the purpose of this study
the model results are very credible.

4.2 What-if scenarios

In this paragraph several “what-if scenarios” for the New Orleans East bowl will be
presented. These scenarios will provide more information on the contribution to the flood
by the different causes of the flood, like overtopping versus rainfall versus breaching and
so on.

Five (5) different scenarios were modeled for the New Orleans East bowl, namely:
1) Scenario all causes and represented by the curve ‘All Causes’,
2) Scenario without breaches (inclusive overtopping and rainfall) and
represented by the curve or line ‘No Breaches’,
3) Scenario only overtopping (without breaches and without rainfall) and
represented by the curve ‘Only Overtopping’
4) Scenario only rainfall without pumps (without breaches and without
overtopping) represented by the curve marked ‘Rain no pumps’, and
5) Scenario only rainfall with 100% pump capacity (without breaches and
without overtopping) represented by the curve ‘Rain 100% pumps’.

32
Flood simulations Greater New Orleans July 2007

Scenario 1 is equal to the Katrina event. Scenario 2 is the Katrina event without any
breaches, only overtopping of the pre-Katrina levees/flood walls and rainfall contribute to
the flooding. In scenario 3 only overtopping of the pre-Katrina levees/flood walls is
considered. Scenarios 4 and 5 show the upper and lower bound or brackets of the water
levels in the situation if flooding had only being due to rainfall and there had been no
breaches and/or overtopping. Scenario 4 shows the water levels with only rainfall, when
no water is pumped out (upper bound of the water levels due to rainfall). Scenario 5
shows the water levels with only rainfall, when all pumps in New Orleans East function
at 100% of their rated pump capacity according to IPET [3]. During Katrina the pumps
did not function in New Orleans East [3].

At five locations in the bowl, indicated in figure 4.5, water level changes have been
measured for each of the five scenarios.

4
▲

▲
2 ▲
3
▲
5
▲
1

Figure 4.5 Water level locations scenario runs New Orleans East

33
Flood simulations Greater New Orleans July 2007

Figures 4.6-4.10 show the changes in water levels on the five locations. The difference
between the black solid and orange dashed line is caused by the breaches and erosion that
occurred during the storm. The difference between the black line and the green dashed-
dot line is caused by the breaches and rainfall. The difference between the green dashed-
dot line and the orange dashed line is the water that comes into the polder due to rainfall.

The scenarios show that overtopping and breaches are the main causes of flood water in
New Orleans East. Along the east side (Willow Brook & Six Flags - sites 4 and 5) the
flooding is dominated by the breaches in the Back levee. The inner levee (Maxent Levee)
does create two compartments within the polder. Along the west side of the New Orleans
East bowl overtopping and breaching appear to contribute almost equal amounts to the
flooding, with overtopping dominating.

Rainfall has a smaller influence on the final water level in this bowl than overtopping or
the breaches. The lower bound approach for the contribution of rainfall flooding
(scenario 5 - only rain with 100% pump capacity) shows no substantial rain water at these
five locations. Also in the upper bound approach (scenario 4 - only rain without pumps)
there is not much rainwater at these locations.

Water levels at location 1

6
1. All Causes
5
2. No Breaches

4 3. Only Overtopping

4. Rain no pumps
3
5. Rain 100% pumps
2
Water level (ft)

-1

-2

-3

-4

-5

-6

-7

-8
29-08-05 0:00 29-08-05 3:00 29-08-05 6:00 29-08-05 9:00 29-08-05 12:00 29-08-05 15:00 29-08-05 18:00 29-08-05 21:00 30-08-05 0:00

Time

Figure 4.6 Water levels location 1 New Orleans East

34
Flood simulations Greater New Orleans July 2007

Water levels at location 2

6
1. All Causes
5
2. No Breaches

4 3. Only overtopping

4. Rain no pumps
3
5. Rain 100% pumps
2
Water level (ft)

-1

-2

-3

-4

-5

-6

-7

-8
29-08-05 0:00 29-08-05 3:00 29-08-05 6:00 29-08-05 9:00 29-08-05 12:00 29-08-05 15:00 29-08-05 18:00 29-08-05 21:00 30-08-05 0:00

Time

Figure 4.7 Water levels location 2 New Orleans East

Water levels at location 3

6
1. All Causes
5
2. No Breaches

4 3. Only overtopping

4. Rain no pumps
3
5. Rain 100% pumps
2
Water level (ft)

-1

-2

-3

-4

-5

-6

-7

-8
29-08-05 0:00 29-08-05 3:00 29-08-05 6:00 29-08-05 9:00 29-08-05 12:00 29-08-05 15:00 29-08-05 18:00 29-08-05 21:00 30-08-05 0:00

Time

Figure 4.8 Water levels location 3 New Orleans East, the blue dashed line lies under the blue solid line

35
Flood simulations Greater New Orleans July 2007

Water levels at location 4

6
1. All Causes
5
2. No Breaches

4 3. Only overtopping

4. Rain no pumps
3
5. Rain 100% pumps
2
Water level (ft)

-1

-2

-3

-4

-5

-6

-7

-8
29-08-05 0:00 29-08-05 3:00 29-08-05 6:00 29-08-05 9:00 29-08-05 12:00 29-08-05 15:00 29-08-05 18:00 29-08-05 21:00 30-08-05 0:00

Time

Figure 4.9 Water levels location 4 New Orleans East

Water levels at location 5

10
1. All Causes
9
2. No Breaches
8
3. Only overtopping
7 4. Rain no pumps

6 5. Rain 100% pumps

5
Water level (ft)

-1

-2

-3

-4

-5

-6
29-08-05 0:00 29-08-05 3:00 29-08-05 6:00 29-08-05 9:00 29-08-05 12:00 29-08-05 15:00 29-08-05 18:00 29-08-05 21:00 30-08-05 0:00

Time

Figure 4.10 Water levels location 5 New Orleans East

36
Flood simulations Greater New Orleans July 2007

5 Saint Bernard
5.1 Calibration of the model
Figure 5.1 shows the Digital Elevation model (DEM) for the St. Bernard bowl. The three
stars indicate the approximate locations of the internal flooding hydrographs used to
check the model results. The arrows and numbers correspond to the breaches given in
table 5.1. These breaches in combination with the rainfall and surge hydrographs outside
the bowl, as presented in section 2.3, are the basis of the model inputs, along with the
DEM, for the New Orleans East bowl. This bowl is divided into two parts by an internal
levee, the 40-Arpent levee. To the east there is a large wetland, 30,000 acre plus basin or
bowl and to the west the inhabited part of this parish. The crest height of the 40-Arpent
levee between the wetlands and populated areas is 6.5 ft.

1
2 *
A *B 5
*C
4

Figure 5.1 DEM St. Bernard, including breach (numbers) and internal hydrograph locations (letters)

37
Flood simulations Greater New Orleans July 2007

Table 5.1 Information on the Breaches in St. Bernard

Nr. Description Width Elevation height Time breach
development
1 IHNC North 221 ft (67 m)1 Sill elevation +1 ft 7:00 - 7:302
(+0.3 m)1
2 IHNC South 808 ft (274 m)1 Sill elevation +1 ft 7:00 - 7:302
(+0.3 m)1
3 MRGO Bayou 214 ft (65 m)1 Sill elevation -6 ft 5:00 - 8:303
Bienvenue (-1.8 m)1
4 MRGO Bayou 54 ft (17 m)1 Sill elevation -6 ft 5:00 - 8:303
Dupree (-1.8 m)1
5 Overtopping Various sections Various sections 5:00 - 8:303
MRGO levee along MRGO along MRGO levee;
(erosion crest) levee; position erosion levels based
and width based on GIS data1
on GIS data1

Calibration results
The time progressive image snapshot series shows how this bowl flooded during
hurricane Katrina. The first image in figure 5.2 is a snapshot from August 29th at 5:00am;
it shows only minor rainfall. Between 5:00 and 8:30am the erosion of the MRGO levee
takes place, and after 6:30am substantial volumes of surge water from the MRGO flow
into the wetlands bowl between the MRGO and 40-Arpent levee. In the simulation the
40-Arpent levee starts to overflow at 8:00am and at 8:30am the first water enters
Chalmette from a northeastern direction as a consequence of overtopping the 40-Arpent
levee. IPET in their report state this overtopping starts at about 8:20am4, which is in
agreement with Team Louisiana’s 8:30am5. IPET also reports that Chalmette flooded
from the northeast, strong agreement with the simulation results.

At 7:00am the North and South breaches from the IHNC channel into the Lower Ninth
Ward start to develop and at 7:30am the first flooding of the Lower 9th Ward is seen in
the model time-step images of the internal flooding. This timing is in agreement with
both the ILIT and Team Louisiana reports. IPET has another opinion about the North
breach and states that this breach started earlier, but both ILIT and Team Louisiana
contradict this conclusion. IPET mentions that surge water started to enter the Lower 9
Ward before 5:30am6 and Team Louisiana around 6:00am. Potential sources of this early
internal flooding include overtopping of the IHNC floodwall, the two breaches in the
IHNC floodwall, and flow from the overtopping of the MRGO/GIWW levees.

1
Source: GIS data with positions, widths and sill elevation heights of the breaches [delivered by Chad
Morris 2007-6-28]
2
Source: Team Louisiana report table 3, page 54-55 and timeline page 93-94 [4]
3
At 5:00 waves start to erode the MRGO levee [Source: see footnote 2], in the simulation erosion
continues until the peak water level in Lake Borne is reached at 8:30.
4
Source: IPET volume IV: The Storm, page 199 [1]
5
Source: Team Louisiana, page 64 [4]
6
Source: IPET volume IV: The Storm, page 200 [1]

38
 Flood simulations Greater New Orleans July 2007

In our model simulations no overtopping occurs before 7:00am, and the two breaches
start at 7:00am. Therefore no water, except rain water, prior to 7:00am is visible inside
the Lower 9th Ward in the simulation results. ILIT gives an alternative explanation for the
early water mentioned by eye witness reports in IPET. ILIT mentions under seepage,
water migrating from the channel, under pressure, beneath the IHNC floodwalls as a
possible explanation for the early flood water in the Lower 9th Ward. Under seepage
through a levee is not captured in the simulation.

2005-08-29 2005-08-29
05:00h 05:30h

2005-08-29 2005-08-29
06:00h 06:30h

39
 Flood simulations Greater New Orleans July 2007

2005-08-29 2005-08-29
07:00h 07:30h

2005-08-29 2005-08-29
08:00h 08:30h

2005-08-29 2005-08-29
09:00h 10:00h

40
 Flood simulations Greater New Orleans July 2007

2005-08-29 2005-08-29
11:00h 12:00h

Figure 5.2 Progressing snapshots of the computed internal flooding water depth inside the St. Bernard
bowl

Figure 5.2 shows the calculated maximum internal flooding depths in the St. Bernard
bowl as a consequence of the breaches and overtopping initiated by hurricane Katrina.
Figure 5.4 shows the comparison between our computed internal flood hydrographs and
those presented by IPET for the St. Bernard bowl. The locations of the hydrographs are
indicated in figure 5.1, location A corresponds with Lower 9th Ward NW., B with Lower
9th Ward E. and C with Chalmette. IPET based its hydrographs on eye witness accounts,
stopped clock times and time/date stamped digital photographs. Our computed flood
hydrographs are very similar to the data presented by IPET. The predicted peak flood
level is within a foot of IPET’s measurements. The peak in the model is a couple of hours
earlier and the water level starts to decrease earlier than in the IPET hydrographs. The
final water descent rate in the model is about equal to the IPET hydrograph.

As discussed previously, in extreme events such as Katrina is difficult if not impossible

to get direct measurement of some input data such as water levels at the time of a breach,
especially right at the breach location. However, the differences between the computer
model results and the calibration data are fully acceptable. For the purpose of this study
the model results are very credible.

41
Flood simulations Greater New Orleans July 2007

Figure 5.3 Maximum computed internal flood depths, St. Bernard bowl, as a consequence of hurricane
Katrina and the numerous levee failures.

IPET & Computed Flood Hydrographs from Lower 9th Ward and Chalmette

12

10

8
Interior Water Level (ft, NAVD88)

-2

-4

-6
8/29/05 0:00 8/29/05 12:00 8/30/05 0:00 8/30/05 12:00 8/31/05 0:00 8/31/05 12:00 9/1/05 0:00
Time
L.9 Ward NW. IV-194 L.9 Ward E. IV-194 Chalmette IV-194
Computed L.9 NW Computed L.9 E. Computed Chalmette

Figure 5.4 Comparison of our computed internal flood hydrographs and that presented by the IPET team.

42
Flood simulations Greater New Orleans July 2007

5.2 What-if scenarios

In this paragraph several “what-if scenarios” for Saint Bernard will be presented. These
scenarios will provide more information on the contribution to the flood by the different
causes of the flood, like overtopping versus rainfall versus breaching and so on.

For the St. Bernard bowl seven different internal flooding model scenarios were
developed:
1) Scenario all causes represented as the line ‘All causes’
2) Scenario without breaches (inclusive overtopping and rainfall) represented by
the line or curve ‘No Breaches’
3) Scenario without breaches at MRGO (inclusive breaches at IHNC,
overtopping and rainfall) represented by the curve ‘No Breaches MRGO’
4) Scenario without breaches at IHNC (inclusive breaches at MRGO,
overtopping and rainfall) represented by the line ‘No Breaches INHC’
5) Scenario only overtopping (without breaches and without rainfall) represented
by the curve ‘Only Overtopping’
6) Scenario only rainfall without pumps (without breaches and without
overtopping) represented by the line ‘Rain no pumps’
7) Scenario only rainfall with 100% pump capacity (without breaches and
without overtopping) represented by the curve ‘Rain 100% pumps’

Scenario 1 is equal to the Katrina event. Scenario 2 is the Katrina event without all the
breaches whether MRGO or INHC and flooding is due to overtopping of the pre-Katrina
levees/flood walls and rainfall. In scenario 3 the breaches at MRGO are left out. These
breaches are the eroded MRGO levee and the two breaches at Bayou Bienvenue and
Bayou Dupree. In scenario 4 the breaches at IHNC are left out, these breaches are the
North and South breach at the IHNC. Scenario 5 considers only overtopping of the pre-
Katrina levees/flood walls. Scenarios 6 and 7 show the upper and lower bound or limits
of the flooding if there had been no breaches and no overtopping. Scenario 6 shows the
water levels with only rainfall, when no water is pumped out (upper bound of the water
levels due to rainfall). Scenario 7 shows the water levels with only rainfall, when all
pumps in St. Bernard function at 100% of their rated pump capacity according to IPET
[3]. During Katrina the pumps did not function in St. Bernard [3].

At ten locations in the bowl, indicated in figure 5.5, internal flooding water level changes
have been computed for each of the 7 scenarios. The locations 1 to 6 are pair wise located
along lines perpendicular to the 40-Arpent levee. The northern location of each pair has a
lower surface elevation than the southern one, therefore the development in time of the
flooding (water level) for each scenario is different for both locations in the same pair.

43
Flood simulations Greater New Orleans July 2007

▲
1
▲2 ▲
3
▲
▲ 4 5
▲
▲ 6 7

▲ 8

▲ 9 ▲
10

Figure 5.5 Water level locations scenario runs St. Bernard bowl.

Figures 5.6-5.15 show the changes in water levels on the ten locations. The flooding that
is represented by the difference between the black solid (All Causes) and red dashed (No
Breaches) line is that caused by all the breaches and levee erosion that occurred during
the storm. The difference between the black (All Causes) and gray (No Breaches INHC)
line is the flooding that resulted from the two big the breaches along the IHNC into the
Lower Ninth ward. The contribution to the overall internal flooding resulting from the
breaches of the MRGO levee are depicted by the difference between the black (All
Causes) and purple dashed (NO Breaches MRGO) lines. Flooding due to rainfall, such a
there is, is depicted by the difference between the green dashed-dot line (Only
Overtopping) and the red dashed line (No Breaches).

The model results depicted in Figures 5.6 to 5.15 clearly show the real importance of
elevation in this polder in relation to the depth of flooding. Rainfall for example flows
into the lower lying areas, and the higher lying locations hardly show any flood level
increase with the two scenarios (scenario 6 and 7) that only consider rain. In the under
bound approach (scenario 7 - only rain with 100% pump capacity) only the relatively low
lying locations 1 and 7 show a small amount of rain water, being less than one foot. The
upper bound approach (scenario 6 - only rain without pumps) also does not show much
rainwater flooding. The scenario that considers only overtopping shows some flood water
at the lower western part of the Lower 9th Ward (location 1). The overtopping from the
IHNC is not enough to fill more of the St. Bernard bowl beyond the Lower Ninth ward.

The overtopping of the MRGO levee is not enough to fill the Wetlands basin between
MRGO and 40-Arpent levee up to the crest elevation of the 40-Arpent levee. This is
significant as the model shows that if the structures along the INHC and MRGO had not
breached, there would have been limited flooding in the St Bernard bowl.

44
Flood simulations Greater New Orleans July 2007

Therefore the main causes for the flooding of the populated area inside the St. Bernard
bowl are the breaches in the IHNC and MRGO.

An interesting observation is that without the IHNC breaches almost the same volume of
flood water enters the populated area of the St. Bernard bowl compared with the scenario
off MRGO breaches plus overtopping plus rain. Without the IHNC breaches it takes a
little longer in the western part of the bowl to reach the peak water level, but the peak
itself is only marginally lower. In the eastern part of the bowl the influence of the IHNC
breaches on the development of the water level is even smaller. The IHNC breaches do
however influence the outflow of the water out of the Lower 9th Ward and Chalmette.
Floodwater ‘leaked’ out of the bowl through the INHC breaches for many days after the
storm had passed. Without these breaches the flood water can only flow over the 40-
Arpent levee into the wetlands, therefore after the peak the flood water level decreases
slower without the IHNC breaches.

On the other hand, the scenario with only the IHNC breaches, overtopping and rainfall
(without MRGO breaches), reveals that the St. Bernard bowl still suffers significant
flooding although flood depths are significantly less than the scenario that includes the
MRGO breaches.

45
Flood simulations Greater New Orleans July 2007

Water levels at location 1

13
12 All Causes
11 No Breaches
10
9 No Breaches MRGO
8 No Breaches IHNC
7
Water level (ft)

Only Overtopping
6
5 Rain no pumps
4 Rain 100% pumps
3
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Figure 5.6 Water levels location 1 St. Bernard

Water levels at location 2

13
12 All Causes
11 No Breaches
10
9 No Breaches MRGO
8 No Breaches IHNC
7
Water level (ft)

Only Overtopping
6
5 Rain no pumps
4 Rain 100% pumps
3
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Figure 5.7 Water levels location 2 St. Bernard, the red, green and 2 blue lines are lying (almost) under
each other

46
Flood simulations Greater New Orleans July 2007

Water levels at location 3

13
12 All Causes
11 No Breaches
10
9 No Breaches MRGO
8 No Breaches IHNC
7
Water level (ft)

Only Overtopping
6
5 Rain no pumps
4 Rain 100% pumps
3
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Figure 5.8 Water levels location 3 St. Bernard, the red line lies under the solid blue line, the dashed blue
line lies under the dashed blue line

Water levels at location 4

13
12 All Causes
11 No Breaches
10
No Breaches MRGO
9
8 No Breaches IHNC
7
Water level (ft)

Only Overtopping
6
5 Only Rainfall
4 Rain 100% pumps
3
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Figure 5.9 Water levels location 4 St. Bernard, the red line lies under the solid blue line, the dashed blue
line lies under the dashed green line

47
Flood simulations Greater New Orleans July 2007

Water levels at location 5

13
12 All Causes
11
No Breaches
10
9 No Breaches MRGO
8
No Breaches IHNC
Water level (ft)

7
6 Only Overtopping
5 Only Rainfall
4
3 Rain 100% pumps
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00
Time
Figure 5.10 Water levels location 5 St. Bernard, the red line lies under the solid blue line, the dashed blue
line lies under the dashed green line

Water levels at location 6

13
12
11
10
9
8
7
Water level (ft)

6 All Causes
5
4 No Breaches
3 No Breaches MRGO
2
1 No Breaches IHNC
0 Only Overtopping
-1
-2 Only Rainfall
-3 Rain 100% pumps
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time
Figure 5.11 Water levels location 6 St. Bernard, the red, purple, green and blue dashed lines lie under the
blue line.

48
Flood simulations Greater New Orleans July 2007

Water levels at location 7

13
12 All Causes
11
10 No Breaches
9 No Breaches MRGO
8
No Breaches IHNC
7
Water level (ft)

6 Only Overtopping
5
Only Rainfall
4
3 Rain 100% pumps
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time
Figure 5.12 Water levels location 7 St. Bernard, the red line lies under the solid blue line.

Water levels at location 8

13
12 All Causes
11
10 No Breaches
9 No Breaches MRGO
8
No Breaches IHNC
Water level (ft)

7
6 Only Overtopping
5
Only Rainfall
4
3 Rain 100% pumps
2
1
0
-1
-2
-3
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time
Figure 5.13 Water levels location 8 St. Bernard, the red, green and blue dashed lines lie under the blue
line.

49
Flood simulations Greater New Orleans July 2007

Water levels at location 9

13
12
11
10
9
8
Water level (ft)

7
6
All Causes
5
4 No Breaches
3 No Breaches MRGO
2
1 No Breaches IHNC
0 Only Overtopping
-1
-2 Only Rainfall
-3 Rain 100% pumps
-4
-5
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time
Figure 5.14 Water levels location 9 St. Bernard, the red, purple, green and blue dashed lines lie under the
blue line.

Water levels at location 10

13
12
11
10
9
8
7
Water level (ft)

6
5
4 All Causes
3
No Breaches
2
1 No Breaches MRGO
0 No Breaches IHNC
-1
-2 Only Overtopping
-3 Only Rainfall
-4
-5 Rain 100% pumps
29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 29/08/05 30/08/05
0:00 3:00 6:00 9:00 12:00 15:00 18:00 21:00 0:00

Time
Figure 5.15 Water levels location 10 St. Bernard, the red, purple, green and blue dashed lines lie under the
blue line.

50
Flood simulations Greater New Orleans July 2007

5 Conclusions
In this research project we built flood simulation models for the three bowls that
comprise Greater New Orleans. With this model, the Katrina flood in August 2007 has
been simulated, and we obtained good agreement between the model output and the eye
witness reports.

The model is used to assess the following causes of the flooding: rainfall (and the use of
the pumping stations), overtopping of the levees and flood control structures, and the
breaches. It is concluded that the contribution of these sources depend on the location,
and the results are shown for a number of locations. It can be concluded that rainfall and
overtopping contributed heavily to the flooding of Orleans East. In the bowls Orleans
Metro and Saint Bernard the contribution of rainfall and overtopping is relatively small
compared with the contribution of the breaches. We refer to the figures where the relative
contribution of the causes is shown.

Further refinements of the modeling may be possible as additional data becomes

available especially as relates to the operation of the pumps, how long they pumped for
and their capacities with changing water levels both within and without the polders. The
model results could aid in assess the economic damage and causes of loss of life related
to the various water inputs and contributions to the internal flooding of the polders during
hurricane Katrina.

51
Flood simulations Greater New Orleans July 2007

6 References
[1] Link, E. et al. [IPET] (26 March 2007) Performance Evaluation of the New
Orleans and SouthEast Louisiana Hurricane Protection System. Final Report of
the Interagency Performance Evaluation Task Force. Volume IV - The Storm
[2] Link, E. et al. [IPET] (1 June 2006) Performance Evaluation of the New Orleans
and SouthEast Louisiana Hurricane Protection System. Draft Final Report of the
Interagency Performance Evaluation Task Force. Volume V - The Performance -
Levees and Floodwalls
[3] Link, E. et al. [IPET] (26 March 2007) Performance Evaluation of the New
Orleans and SouthEast Louisiana Hurricane Protection System. Final Report of
the Interagency Performance Evaluation Task Force. Volume VI - The
Performance - Interior Drainage and Pumping
[4] Van Heerden, I. et al [Team Louisiana] (18 December 2006) The Failure of the
New Orleans Levee System during Hurricane Katrina.
[5] Seed, R. B. et al. [ILIT] (31 July 2006) Investigation of the Performance of the
New Orleans Flood Protection Systems in Hurricane Katrina on August 29, 2005
[6] Ministry of Public Works and Water management (September 2005). Flood risk
simulation in the Netherlands. RWS - DWW, Delft, September 2005.
[7] Delft Hydraulics, 2005. User manual Delft 1D2D. Delft.
[8] Maaskant, B., 2007, Research on the relationships between flood characteristics
and fatalities. Based on the flooding in New Orleans caused by hurricane Katrina.
Msc thesis, Delft University of Technology.
[9] American Society of Civil Engineers [ASCE], 2007. The New Orleans Hurricane
Protection System: What Went Wrong and Why. A report by the American
Society of Civil Engineers Hurricane Katrina External Review Panel.

52