Case Study: 17th Street Canal Breach Closure Procedures

On August 29, 2005 Hurricane Katrina resulted in several breaches in the levees and flood walls protecting New Orleans. Of the 20 breaches, the 17th Street Canal breach caused much of the city flooding. In this paper, the results of studies on a 1:50 scale hydraulic model of this breach based on the Froude similitude relationships are presented. It is assumed in the model that the bed is fixed and the levee below the flood wall remains intact during breach closure. This was the case in the 17th Street Canal breach. Because of the many uncertainties in the values of various variables, a range of conditions were run on the model in an attempt to bracket the results for the flooding depths and the initial failed attempts to close the breach. Then, various possible methods for breach closure were investigated utilizing the procedures developed for cofferdam closure for river diversion, e.g., toe dumping, transverse dumping, single- and multibarrier embankments, etc. Closures of the breach and the closure of the canal at the Old Hammond Highway Bridge were investigated. Results from the case study show that some of these methods could have been utilized for closing the Katrina breaches. However, special care should be exercised when extending them for breach closure at other sites.     

Active Flood Hazard Mitigation.?I: Bidirectional Wave Control

Active mitigation of unimodal flood waves is achieved by selective lateral flow withdrawal. This is shown to create depression waves that reduce the impact of hazardous flood waves in rivers. Lateral outflow is induced by a deliberate levee breach or through an emergency side channel spillway generating bidirectional wave action in the main channel. An adjoint sensitivity method founded on 1D shallow-water equations is used to identify the optimal locations and timing for the lateral outflow. The efficiency of the adjoint sensitivity method allows flood hazards to be mitigated in real time by rapidly providing the flow sensitivity to changes in lateral outflow. The spatial and temporal coordinates associated with the peak values of the sensitivities are examined as indicators of optimal locations and timing for active control. The sensitivities are found to be good indicators of optimal control points even in cases where the magnitude and duration of the lateral outflow exceeds the small perturbation assumption.     

Active Flood Hazard Mitigation.?II: Omnidirectional Wave Control

Active mitigation of unimodal flood waves is achieved by selective boundary flow withdrawal. This is shown to create omnidirectional depression waves that can reduce the impact of hazardous flood waves in wide rivers, harbors, and reservoirs. Boundary outflow is induced by a deliberate levee breach or through an emergency side channel spillway generating disturbances that reflect on the banks of the channel further complicating the wave pattern. An adjoint sensitivity method based on the 2D shallow-water equations is presented to aid in the mitigation of an extreme flooding event by identifying optimal locations and times for the selective withdrawal of flood waters. The efficiency of the method allows adaptive flood control to proceed in real time. It is shown that wave reflections from solid boundaries create a complex pattern of sensitivity waves and multiple options for control. The adjoint sensitivity results become less accurate as the magnitude and duration of the perturbation become large. However, the adjoint sensitivities provide reliable information for identifying optimal locations and times for a selective withdrawal of large magnitude and a fair indication of the spatial dependence of the objective function sensitivity to large changes in flow.     

Stability of I-Walls in New Orleans during Hurricane Katrina

Failures of I-walls during Hurricane Katrina were responsible for many breaches in the flood protection system in New Orleans. Six breaches were examined in detail by Task Group 7 of the Interagency Performance Evaluation Taskforce. Four of these failures and breaches, which occurred before the water levels reached the top of the wall, were not caused by overtopping erosion. The failure of the I-wall at the 17th Street Canal resulted from shear through the weak foundation clay. The south failure of the London Avenue I-wall was caused by subsurface erosion, which carried massive amounts of sand inland, and removed support for the wall, leading to catastrophic instability. At the north breach on London Avenue, the failure was caused by high pore pressures, combined with a lower friction angle in the loose sand, which resulted in gross instability of the I-wall under the water pressure load from the storm surge. Looking back, with the benefit of 20-20 hindsight, these stability and erosion failures can be explained in terms of modern soil mechanics, exploration techniques, laboratory test procedures, and analysis methods. An important factor in all of the cases investigated was development of a gap behind the wall as the water rose against the wall and caused it to deflect. Formation of the gap increased the load on the wall, because the water pressures in the gap were higher than the earth pressures that had acted on the wall before the gap formed. Where the foundation soil was clay, formation of a gap eliminated the shearing resistance of the soil on the flood side of the wall, because the slip surface stopped at the gap. Where the foundation soil was sand, formation of the gap opened a direct hydraulic connection between the water in the canal and the sand beneath the levee. This hydraulic short circuit made seepage conditions worse, and erosion due to underseepage more likely. It also increased the uplift pressures on the base of the levee and marsh layer landward of the levee, reducing stability. Because gap formation has such important effects on I-wall stability, and because gaps behind I-walls were found in many locations after the storm surge receded, the presence of the gap should always be assumed in I-wall design studies.     

New Orleans Levee System Performance during Hurricane Katrina: 17th Street Canal and Orleans Canal North

Centrifuge modeling of the 17th Street Canal and Orleans Canal North levees was performed in this study. During hurricane Katrina the levees on the 17th Street Canal failed, leading to breaches in the outfall canal in the city. Two mechanisms were observed in the centrifuge modeling that could cause a breach. First, a water-filled crack formed in front of the floodwall as the water in the canal rose above the top of the levee. The levees on the 17th Street Canal, which were supported on clay foundations, failed when this cracking led to a translational (sliding) failure in the clay layer commencing at the toe of the floodwall. The levees at Orleans Canal North, where failure did not occur, were also modeled to demonstrate that the model tests could successfully simulate failure and nonfailure conditions. The centrifuge model tests identified the importance of the crack formation in relation to the stability of the floodwall. These tests also confirmed that levee geometry, floodwall depth of penetration, and the underlying soil profile were all critical to the performance of the system under flood loading.     

New Orleans and Hurricane Katrina. I: Introduction, Overview, and the East Flank

The failure of the New Orleans regional flood protection systems, and the resultant catastrophic flooding of much of New Orleans during Hurricane Katrina, represents the most costly failure of an engineered system in U.S. history. This paper presents an overview of the principal events that unfolded during this catastrophic hurricane, and then a more detailed look at the early stages of the event as the storm first drove onshore and then began to pass to the east of the main populated areas. The emphasis in this paper is on geotechnical lessons and it also includes broader lessons with regard to the design, implementation, operation, and maintenance of major flood protection systems. This paper focuses principally on the early stages of this disaster, including the initial inundation of Plaquemines Parish along the lower reaches of the Mississippi River as Katrina made landfall, and the subsequent additional early levee breaches and erosion along the eastern flanks of the regional flood protection systems fronting Lake Borgne that resulted in the flooding of the two large protected basins of New Orleans East and St. Bernard Parish. Significant lessons learned include (1) the need for realistic assessment of risk exposure as an element of flood protection policy; (2) the importance of considering erodibility of embankment and foundation soils in levee design and construction; (3) the importance of considering all potential failure modes; and (4) the problems inherent in the construction of major regional systems over extended periods of multiple decades. These are important lessons, as they are applicable to other regional flood protection systems in other areas of the United States, and throughout much of the world.     

Embankment Dam Breach Parameters and Their Uncertainties

Potential flood hazards that would be created by breached embankment dams need to be evaluated to select spillway design floods and to prepare emergency action plans. The breaches are often modeled simply, usually in the shape of a trapezoid that is defined by its final height, base width or average width, and side slopes, along with the time needed for the opening to form completely. Data collected from 74 embankment dam failures were used to develop mathematical expressions for the expected values of the final width and side slope of a trapezoidal breach along with its formation time. Information is provided that allows variances of the predicted quantities to be calculated as well. The findings of the statistical analysis were then applied in a Monte Carlo simulation to estimate the degree of uncertainty of predicted peak flows and water levels downstream from breached embankment dams.     

New Orleans and Hurricane Katrina. III: The 17th Street Drainage Canal

The failure of the levee and floodwall section on the east bank of the 17th Street drainage canal was one of the most catastrophic breaches that occurred during Hurricane Katrina. It produced a breach that rapidly scoured a flow pathway below sea level, so that after the storm surge had largely subsided, floodwaters still continued to stream in through this breach for the next two and a half days. This particular failure contributed massively to the overall flooding of the Metropolitan Orleans East Bank protected basin. Slightly more than half of the loss of life, and a similar fraction of the overall damages, occurred in this heavily populated basin. There are a number of important geotechnical and geoforensic lessons associated with this failure. Accordingly, this paper is dedicated solely to investigating this single failure. Geological and geotechnical details, such as a thin layer of sensitive clay that was laid down by a previous hurricane, proper strength characterization of soils at and beyond the toe of the levee, and recognition of a water-filled gap on the inboard side of the sheet pile cutoff wall are judged to be among the most critical factors in understanding this failure. The lessons learned from this study are of importance for similar flood protection systems throughout other regions of the United States and the world.     

New Orleans and Hurricane Katrina. II: The Central Region and the Lower Ninth Ward

The failure of the New Orleans regional flood protection systems, and the resultant catastrophic flooding of much of New Orleans during Hurricane Katrina, represents the most costly failure of an engineered system in U.S. history. This paper presents an overview of the principal events that unfolded in the central portion of the New Orleans metropolitan region during this hurricane, and addresses the levee failures and breaches that occurred along the east–west trending section of the shared Gulf Intracoastal Waterway/Mississippi River Gulf Outlet channel, and along the Inner Harbor Navigation Channel, that affected the New Orleans East, the St. Bernard Parish, and the Lower Ninth Ward protected basins. The emphasis in this paper is on geotechnical lessons, and also broader lessons with regard to the design, implementation, operation, and maintenance of major flood protection systems. Significant lessons learned here in the central region include: (1) the need for regional-scale flood protection systems to perform as systems, with the various components meshing well together in a mutually complementary manner; (2) the importance of considering all potential failure modes in the engineering design and evaluation of these complex systems; and (3) the problems inherent in the construction of major regional systems over extended periods of multiple decades. These are important lessons, as they are applicable to other regional flood protection systems in other areas of the United States, and throughout much of the world.     

50-Year Flood in the Lower Illinois River: Sensitivity of Spillway and Levee Failure Option Parameters in the UNET Model

The one-dimensional unsteady-state hydraulic model (UNET) was developed by the U.S. Army Corps of Engineers and has been used for flood protection in some large rivers. In its use for the real-time management of levee and drainage districts for flood protection, such as on the Mississippi River Basin, the required number of simulations for combinations of scenarios would present a great challenge for modelers to provide data for operational-mode decision making. The goal of this paper is to identify the most sensitive parameters for the spillway, simple levee failure, and complicated levee failure options in UNET so that simulations for real-time operation can focus on the most sensitive parameters without sacrificing much modeling accuracy. The sensitivity analysis was performed for the 50-year frequency flood in the lower Illinois River during which the Thompson Lake Levee and Drainage District or the Emiquon Area were used as temporary flood storage by using a state-of-the-art variance-based sensitivity analysis method—extended Fourier amplitude sensitivity test (FAST). The first-order and total-effect sensitivity indexes computed from the FAST method were used to evaluate the affect of the parameters to flood-peak reduction. As a result, the least and most sensitive variables for the spillway, simple levee failure, and complicated levee failure options in UNET were identified, and this knowledge helps significantly reduce the number of simulations for management scenarios by focusing on only the most sensitive parameters.     

Reducing Erosion of Earthen Levees Using Engineered Flood Wall Surface

Erosion was one of the major causes for the failure of New Orleans levee system during Hurricane Katrina. Protection of flood walls from erosion failure can be achieved in many different ways. This study consisted of experimental research to develop engineered flood wall surfaces that can reduce the erosion energy of the plunging water before the water hits the levee surface, so that the flood protection system becomes more resilient. Test results were compared to hydrodynamics simulation results by using FLOW3D. The results revealed that the erosion resistance of levees can be substantially reduced by providing protective structures at the surface of the flood walls. An effectively designed protective structure could reduce the erosion depth as much as 40% and extend the erosion time as much as 400%.     

Effect of Channel Restoration on Flood Wave Attenuation

Stream channel restoration can increase flow storage and energy dissipation of passing flood waves. Elements of restoration design that can enhance attenuation include remeandering, which reduces channel slope and increases channel length relative to the floodplain; restoring channel-floodplain connectivity; and revegetating banks and the floodplain. Reestablishment of floodplain hydraulic function is increasingly a goal of restoration programs, yet the approximate magnitude of possible change to attenuation due to reach-scale restoration remains poorly quantified. We examined the efficacy of channel restoration on flood attenuation using restored reaches and synthetic reaches representing median dimensions of channel restoration projects in North Carolina (e.g., ～ 1?km in length). We applied an industry standard dynamic flood routing model (UNET in HEC-RAS) to route floods in impaired and restored reach models. Floods routed through field-based reach models either exhibited very small increases in attenuation, largely due to assumed increases in floodplain roughness, or a decrease in attenuation. Analysis demonstrated that attenuation of peak discharge is overall most sensitive to channel and valley slope, channel and floodplain roughness, and channel and valley length in decreasing order, but is dependent on flood magnitude. Restoration most impacted floods of intermediate magnitude (between 2- and 50-year return interval), particularly those confined to the channel under the impaired morphology but able to access the floodplain under the restored morphology. Restoration may rehabilitate a channel’s ability to attenuate small to intermediate floods by augmenting flood access to the floodplain, changing channel geometry, and enhancing channel and floodplain roughness over time. However, our study shows that the predominantly small scale of current channel restoration will provide minimally quantifiable enhancement to flood attenuation.     

Geologic Conditions Underlying the 2005 17th Street Canal Levee Failure in New Orleans

A careful program of subsurface sampling and cone penetration test soundings was employed to characterize the geologic conditions beneath the failed portion of the 17th Street Canal levee in New Orleans, where a 150?m long section of the levee and floodwall translated up to ～ 16?m when flood waters rose to 1–2?m of the wall’s crest on August 29, 2005, during Hurricane Katrina. The subsurface conditions are characterized by discrete layers of fill placed upon the historic cypress swamp, which is underlain by a deeper, prehistoric cypress swamp. These swamp deposits were consolidated beneath the levee, and in the area of the 2005 failure, the swamp materials infilled a natural depression believed to be an old slough, which dipped below the sheetpile tips for a distance of about 50?m, which corresponds to where the breach appears to have initiated. Detailed examination of the recovered soils suggest that recent hurricanes periodically inundated the swamps with saline and/or brackish water, which cause a mass dieoff of swamp vegetation and flocculation of suspended clays, due to the sudden increase in salinity. These conditions promote deposition of discontinuous clay seams beneath layers of organics, which are then covered by fresh water swamp deposits. This sequence is repeated, like a series of tree rings, throughout the swamp deposits. The cypress swamp deposits lying beneath the levee also exhibit high hydraulic conductivity. These materials contain corky wood, and recovered samples often exhibited densities less than water. Nine of the post-Katrina borings recovered intact samples of a basal rupture surface comprised of organic silty clay exhibited near zero residual shear strength after shearing 80 to 100 mm.     

Effect of Flood Recession on Scouring at Bed Sills

The effect of the flood recession time on the local scour depth at bed sills in gravel deposits is examined. Experiments were carried out to study the development of scour holes under time-varying hydraulic conditions with no upstream sediment feed. Triangular-shaped hydrographs, having recession times up to three times the duration of the rising limb, were used. Traditionally, the peak water discharge in any flood event is used as a design value in estimating the final depth of scour formed by a flood. This approach is overly conservative when the flow hydrograph is steep, i.e., during the occurrence of flash floods. The actual reduction of the scour depth from this estimated value is dependent on both the characteristics of the flood event and the characteristics of the stream. The results show that the maximum potential scour depth can be achieved only for hydrographs with long recession times, while the rate of this process can be estimated as a function of the ratio between a characteristic flood time and the steady-state temporal scale of scour development. A method is proposed for the prediction of the scouring process under unsteady flows in terms of two dimensionless temporal parameters. Results obtained for clear-water boundary conditions can be extended to sediment-supply tests if specific supply input conditions hold.     

Transient Head Development due to Flood Induced Seepage under Levees

The purpose of this study was to predict the uplift force during floods on confining layers that overlay extensive horizontal confined aquifers that intersect a large river in response to the water level changes that occur with time in a flooding river. Transient flow of water through the confined aquifer was described by a diffusion type of equation with a boundary condition at the river in which the river head varied with time. The transient head distribution developed from the unsteady flow model applied to the aquifer was compared with the hydraulic head distributions obtained from U.S. Army Corps of Engineers steady-state flow model and a finite-element seepage model. This study concluded that the transient flow model has the potential to analyze time lag in head development, and to predict the seepage condition and heaving potential at various times and distances landside of a levee during a flood cycle, but additional case histories are needed to justify widespread use of the model.     

Institutional arrangements for flood hazard management in Malaysia: an evaluation using the criteria approach

Institutional aspects of flood hazards significantly affect their outcomes in Malaysia. Institutional arrangements to deal with floods include: legislative activity, organisational structures, attitudes and sub-culture, and policies and instruments. When assessed in terms of four specific criteria, institutional aspects of flood hazards are found to be largely inadequate. Disaster reduction programmes are over-dependent on a reactive approach based largely on technology and not even aimed at floods specifically. Structural flood reduction measures are the predominant management tool and, although the importance of non-structural measures is recognised, thus far they have been under-employed. Current laws and regulations with regard to flood management are also insufficient and both the financial and human resources of flood hazard organisations are generally found to be wanting. Finally, economic efficiency, equity and public accountability issues are not adequately addressed by institutional arrangements for flood hazards.     

Modeling of Hyperconcentrated Sediment-Laden Floods in Lower Yellow River

This paper presents a rapid forecast model for simulating hyperconcentrated sediment-laden floods in the Lower Yellow River. The model is a hybrid of a conventional one-dimensional mathematical model for unsteady sediment-laden flow and an artificial neural networks model for encapsulation of numerical results. The former provides detailed river flood routing information under typical scenarios, whereas the latter extracts modeling outputs from the former and establishes a station-specific model for efficient flood forecasting. Three typical floods that occurred in the Lower Yellow River in 1977, 1982, and 1996 are simulated. Not only the hybrid model predictions are found to be in close agreement with measured data, but also the computational speed is significantly enhanced. It is found that sediment transport is of significance with regard to the flooding behavior of hyperconcentrated flows. Therefore, the model presented herein is of particular use for rivers with high sediment concentration.     

Development of the New Orleans Flood Protection System prior to Hurricane Katrina

The system of flood protection surrounding New Orleans and its adjoining parishes prior to Hurricane Katrina evolved over a period of 280?years. The earliest drainage works sought to elevate the river’s natural levees and excavate drainage canals leading towards Bayou St. John, the only natural break across the Metairie-Gentilly distributary ridge. An extensive zone of Cypress Swamps occupied the levee flank depression between the ridge and Lake Pontchartrain. 58?km of drainage canals were excavated across the natural levee backslope and through the swamp depressions bordering the lake between 1833 and 1878. These canals sought to drain the lower portions of the city, which suffered periodic outbreaks of yellow fever, which killed more than 100,000 people during the 19th century. The city has not suffered flooding from the Mississippi River since 1895, most damaging floods having emanated from hurricane surge off of Lake Pontchartrain. Since 1559, 177 hurricanes have struck the Louisiana coastline. A system of pump stations was constructed between 1895 and 1927, which pump water into the river, the lake, and adjacent bayous. The cypress swamps were replaced by the Lakeview and Gentilly residential districts, built after 1945. This old swamp zone has settled as much as 3+?m since 1895. After 1927 the Army Corps of Engineers assumed a leadership role in providing flood control infrastructure, supervising the Mississippi River & Tributaries Project in 1931–1972. In 1955 the Corps role was expanded to include the City of New Orleans, which included maintaining capacity and freeboard of the old drainage canals. After a series of lawsuits between 1961 and 1977, the Corps was forced to employ concrete flood walls along the subsiding drainage canals. These walls were constructed in the 1990s, though some transition elements remained incomplete when Hurricane Katrina struck in August 2005.     

Field Measurements and Simulation of Bridge Scour Depth Variations during Floods

An understanding of bridge scour mechanisms during floods in a fluvial river is very important for cost-effective bridge foundation design. Reliable bridge scour data for flood events are limited. In this study, field experiments were performed at the Si-Lo Bridge in the lower Cho-Shui River, the longest river in Taiwan, to collect scour-depth data using a sliding magnetic collar, a steel rod, and a numbered-brick column. By separating each scour component, a methodology for simulating the temporal variations of the total scour depth under unsteady flow conditions is proposed. The proposed total-scour model integrates three scour components, namely general scour, contraction scour, and local scour. The collected field data, comprising both general scour and total scour depths, are used to validate the applicability of the proposed model. Based on the peak flow discharges during floods, a comparison of the local scour depths calculated using several commonly used equilibrium local scour formulas indicates that most equations may overestimate the local scour depth.     

Case Study: Flood Mitigation of the Muda River, Malaysia

The 2003 flood of the Muda River reached 1,340?m3/s at Ladang Victoria and adversely impacted 45,000 people in Malaysia. A flood control remediation plan proposed a levee height based on a 50-year discharge of 1,815?m3/s obtained from hydrologic models. This design discharge falls outside the 95% confidence intervals of the flood frequency analysis based on field measurements. Instream sand and gravel mining operations also caused excessive riverbed degradation, which largely off sets apparent benefits for flood control. Pumping stations have been systematically required at irrigation canal intakes. Several bridge piers have also been severely undermined and emergency abutment protection works were needed in several places. Instream sand and gravel mining activities should be replaced with offstream mining in the future.