Embankments, Dams, and Slopes: Lessons From the New Orleans Levee Failures and Other Current Issues

A fully coupled hydromechanical model is introduced to analyze the macro-scale pore fluid flow and micro-scale solid phase deformation of saturated granular soils. The fluid motion is idealized using averaged Navier-Stokes equations, and the distinct element method is employed to model the assemblage of solid particles. The fluid-particle interactions are quantified using established relationships. Computational simulation is conducted to investigate the response of sandy deposits when subjected to a continuously increasing head difference causing water flow underneath a river levee. The conducted simulation provided information at the micro-scale level for the solid phase as well as at the macroscopic level for the pore-water flow. The proposed model is capable of evaluating the critical gradient causing piping under transient flow conditions as the hydraulic head difference gradually increases and the solid phase undergoes subsequent deformations. The conducted simulations captured the formation of sand boils and the corresponding exit gradients. The proposed model provides an effective tool to analyze saturated soil systems undergoing large deformation under extreme flow conditions such as those encountered during piping.

In a vast majority of cases, seepage barriers increase the reliability of dams. However, it is important to recognize that seepage barriers often drastically increase hydraulic gradients around the boundaries of the barrier, and through any windows or defects in the barrier. As a consequence, seepage barriers give rise to the potential for additional mechanisms of internal erosion and piping in the dam and the foundation. Mechanisms of erosion and piping that are uniquely related to seepage barriers are being investigated through review of measured performance and analytical studies, and are the subject of this paper. The information presented in this paper will be useful in 1) assessing the possibility of internal erosion and piping in dams with seepage barriers, 2) designing to minimize that possibility, and 3) assessing the risks associated with these mechanisms of erosion and piping.

The selection of mesh refinement and damping parameters, and the comparison of results to reasonable expectations, are challenging issues for the development of a numerical model in a dynamic analysis. Often it is possible to gain significant insight into these issues by conducting relatively simple analyses that are subordinate to those needed to address the primary purpose of the model. This case history presents a few such considerations for the seismic analysis of Clearwater Dam. The concerns addressed deal predominantly with frequency-related issues, but were conducted using software that performs analyses entirely in the time domain by an explicit solution scheme.

This paper utilizes the finite element method for obtaining the response of unsaturated infinite slopes to rainfall infiltration. Water flow is modeled using Richards' equation and the hydraulic properties of the soil are assumed to follow Brooks and Corey relationships. A representative two-dimensional periodic cell of a shallow infinite slope is considered. Periodic boundary conditions are then enforced at the two lateral sides of the cell to account for the lateral extension of the slope. The periodic domain is discretized into a relatively fine mesh to accurately capture the propagation of the wetting front. Transient solution for water flow in unsaturated soils is then obtained and a number of water flow characteristics are presented.

This paper presents the newly developed Wireless Shape-Acceleration Array (WSAA) sensor that measures 2D acceleration and deformation profiles. This sensor system constitutes a major step toward autonomous monitoring technology for soil and soil-structure systems. The Wireless Shape-Acceleration Array (WSAA) sensor employs micro-machined electromechanical sensors (MEMS), which have enabled gravity-based shape calculation along a sensorized substrate. The method is an extension of technologies that use fiber optic orientation sensing to calculate 3D polylines representing the shape of a sensor array. WSAA uses MEMS accelerometers in a pre-calibrated, geometrically constrained array to provide long-term stability previously unattainable with fiber optic methods. This sensor array is capable of measuring 2D soil acceleration and 2D permanent ground deformations, in the horizontal plane, to a depth of one hundred meters. Each sensor array is connected to a wireless earth station to enable real time monitoring of a wide range of soil and soil-structure systems as well as remote sensor configuration. This paper presents the evolving design of this new sensor array as well as close comparisons between the WSAA system and traditional instrumentation utilized in full scale laboratory tests at the National Research Institute for Earth Science and Disaster Prevention (NIED, Japan). The goal of validating and calibrating the measured accelerations and displacements of the WSAA system was achieved through these full scale tests.

In the present paper, the sliding soil mass of a generalized earth slope is divided into a number of vertical slices considering the interfacial forces between two consecutive slices, and using the limit equilibrium analysis for these slices under the influence of static forces along with the pseudo-static seismic forces to obtain the dynamic factor of safety of the slopes. The effects of variation of different parameters like slope angle (β), soil friction angle (φ) and seismic acceleration coefficients both in the horizontal and vertical directions (k_h and k_v) on the dynamic factor of safety (DFS) are presented. A detailed analysis shows that the DFS increases with increase in the soil friction angle. The DFS decreases with the increase in slope angle under seismic conditions and it also drastically decreases with increase in both the horizontal and vertical seismic accelerations acting on the slope which finally leads to the instability of the slope when the factor of safety gets reduced below 1.0. Finally the present results are compared to the existing solutions available in literature and found to give the minimum values of dynamic factor of safety using the present approach for seismic slope stability analysis.

An unstable slope with a slip surface in residual clay was instrumented with slope inclinometer tubes and piezometers to assist in subsurface characterization and to locate the failure surface. Shear strengths in the residual clays were determined by torsional ring shear tests. Slope stability analyses indicate a slip surface that is generally consistent with surface observations and the instrumentation. Drilled piers socketed into shale bedrock were used to stabilize the slope. Pier displacements and loads for the structural design of the piers were estimated by a finite element analysis using a linear elastic beam-on-elastic-foundation model. The piers were installed in two parallel, staggered rows and inclinometers were installed in one pier and in the slope between piers. Post-construction slope movements and pier deflections were monitored. Numerical analysis was used to calculate the shear forces and moments that develop in the piers, and the results are compared to the predictions used in pier design. The shear forces and bending moments determined from the post-construction monitoring compare favorably with the predictions.

This paper describes a landslide that involves a single family residence near Seattle, Washington in the United States. Remediation of the slide initially involved a heavily anchored, soldier pile and timber lagging wall that would be extremely costly for the homeowner. An alternative remediation using smaller drilled-in-place piles with timber lagging held in-place with several drilled tie-back anchors and lightweight geofoam block backfill was selected. The smaller wall size and cost was feasible because lightweight block geofoam was used as the backfill material. The static and seismic design of the wall, the wall construction, and the cost savings to the homeowner are discussed in this paper.

This project involves construction of a 5.4 mile new railroad alignment located in Indiana County, Pennsylvania. The project required the removal of 1.4 million cubic yards of soil and sedimentary rock and resulted in cut slopes as deep as 150 ft and side-hill fill slopes as high as 45 ft. Subsurface investigation indicated that the excavations would expose rock strata that exhibit differential weathering characteristics and could potentially re-activate pre-existing shear planes identified by slickensided surfaces in the test borings. In order to reduce the excavation quantity and right-of-way acquisition, a slope protection system consisting of rock anchors, reinforcing steel mesh, shotcrete, and sub-horizontal drains was developed. This innovative system allowed the design team to utilize steeper rock cut slopes, reducing right-of-way and construction costs. Six inclinometers and seven monitoring wells were installed to monitor the three major cut slopes during and after construction. Inclinometer data indicated movement along distinct shear planes that asymptotically approached zero following the completion of excavation and blasting activities. Construction quality assurance involved 14 verification tests and 31 proof tests for the rock anchors as well as laboratory testing of the shotcrete test panels.

Soil samples were collected from different depths ranging from the ground surface to a depth of 15 m at two different drainage wells in a landslide area. Specimens represent the residual soil near the surface, desiccated soil/rock sample at a depth range of ground water table fluctuation, less weathered rock from the depth below ground water table, and the remolded soil from the sliding zone. Drained simple shear test, drained ring shear tests, Atterberg's limits, particle size analysis, X-ray diffraction, slake durability test, alternate drying and wetting (slaking) test, and X-ray fluorescence tests were conducted. Test result showed insignificant variations of index properties, mineral composition and chemical composition with depth. Because of insignificant differences in mineralogical composition, fully softened shear strength and residual shear strengths did not vary much with depth. The research results suggested that physical weathering by slaking and disintegration are significant to the depth of ground water fluctuation and such weathering reduces shear strength (specifically the peak shear strength) significantly.

This paper provides an overview of the Southeast Louisiana Flood and Hurricane Protection System that was in place at the time of Hurricane Katrina. Both geography and components of the system are described. A brief description of the development of the storm and the major damage caused is also discussed. Almost immediately after the realization that various components of the system had failed, the U.S. Army Corps of Engineers (USACE) responded through an intensive mode of emergency operations. The entire system was surveyed by air to understand the condition of the system and to assess the extent of the damage. The survey was followed by rescue operations and planning for closure of the breaches and "unwatering" of the flooded areas.

Recognizing that coastal Louisiana is super susceptible to storm surges, researchers from LSU started providing emergency managers with storm surge forecasts generated by the ADvanced CIRCulation hydrodynamic model (ADCIRC) in 2002. During Katrina LSU issued five surge forecasts over the two days prior to landfall. Model output was made available in near-real-time (www.hurricane.lsu.edu/floodprediction). One advisory was run post landfall. Hurricane Katrina made landfall in Louisiana as a moderately fast-moving Category 3 hurricane 29^th of August, 2005. Of the populated areas that constitute Greater New Orleans (GNO), 80 percent of Orleans Parish, 99 percent of St. Bernard Parish, and approximately 40 percent of Jefferson Parish were flooded, in some cases for weeks. This flooding cost the lives of more than 1,500 residents. Over 100,000 families were rendered homeless. High water marks (HWMs) acquired outside the flood protection levees permitted assessment of the accuracy of the LSU ADCIRC forecasts. The surge forecast associated with the most intense storm prediction pre-landfall produced the best overall match to observed values, resulting in a predicted GNO mean peak elevation of 3.65 m, 0.18 m lower than the observed (3.83 m). The RMS error overall was 0.60 m. The post-storm simulation produced a mean elevation (3.75 m) only 0.08 m lower than the observed mean, but with a larger RMS error (0.72 m) than the best pre-landfall. Surge forecasts show where levees are challenged and can also provide an estimate of the volume of water that would have been introduced if the levees and floodwalls had not failed. Overtopping predicted by ADCIRC would have delivered far less water than what actually entered the city, ranging from 11 percent in St. Bernard to 37 percent in Orleans East. Just under 60 percent of the nearly 23,000 ha that were submerged would not have experienced standing water. On average, breaches are estimated to have transmitted nearly 84 percent of the observed GNO flood volume. Because breaching played such a dominant role, processes that led to the initiation and expansion of breaches greatly influenced the volume of water introduced into each polder.

Almost immediately after the realization that a number of New Orleans area levees and floodwalls had failed during the onslaught of the storm surge created by Hurricane Katrina, ASCE mobilized to respond to this disaster. It was clear that something had gone terribly wrong with the engineered infrastructure designed to protect the city and surrounding areas from just this type of catastrophe. ASCE and its membership, as well as other organizations, responded immediately with a pledge to assist with support by whatever means were most suitable. One very important mission was to collect perishable data and other evidence to assist in determining the causes of the levee failures. Groups from ASCE's Geo-Institute (G-I) and Coasts, Oceans, Ports & Rivers Institute (COPRI), together with a group sponsored by the National Science Foundation (NSF), banned together in organizing a reconnaissance team, self-titled the Levee Assessment Team (LAT). The LAT joined a U.S Army Corps of Engineers (USACE) reconnaissance team from the Vicksburg, Mississippi Engineering Research and Development Center (ERDC) in the field to make initial assessments of the failures that had occurred and collect perishable data to help provide answers to the causes of those failures. This initial, preliminary study was an important starting point from which a number of detailed investigations have followed. The collection of perishable data was an invaluable source of information that was quickly disappearing as recovery operations proceeded. The independent teams spent more than two weeks from late September to early October 2005 visiting significant failure sites, performing forensic field inspections and documenting observations. While specifics of failure mechanics and detailed source points would take much more in-depth studies, the data collected in this early field reconnaissance allowed many preliminary evaluations and hypotheses to be drawn. While some difficulties were initially encountered in obtaining adequate background information, this unprecedented involvement of independent assessment experts for the failure of a major federal project proved to be an important positive step forward in practice and policy for engineers responding in the wake of disaster.

The geologic history of the New Orleans area significantly influences the engineering properties of the foundation soils beneath the levees. Geologic and engineering data gathered from the different levee failures identifies a spatially complex geomorphic landscape, caused by Holocene sea level rise, development of different Mississippi River delta lobes, and the distributary channels associated with delta development. Overlying the Pleistocene surface beneath the New Orleans area are predominantly fine-grained, shallow water depositional environments and related sediments associated with bay sound (or estuarine) nearshore-gulf, sandy beach, lacustrine, interdistributary, and paludal (marsh and swamp) environments. These environments define the New Orleans area history during the Holocene, and comprise the levee foundation for the different failure areas. A barrier beach ridge is present in the subsurface along the southern shore of Lake Ponchartrain and blocked the filling of the lake with fluvial-deltaic sediments. This buried beach impacted the supply and texture of sediment being deposited by advancing distributary channels, and influenced the engineering properties of these soils. Marsh and swamp soils beneath the failure area at the 17th Street Canal are much thicker in comparison to those beneath the London Avenue Canal failures because of the influence of the beach complex, and are thickest in the Industrial Canal area. Additionally, man's activities in the New Orleans area during historic time contributed to the spatial complexity and affected the engineering properties of the foundation soils. These activities include construction of drainage and navigation canals, pumping ground water drainage, hydraulic filling of the Lake Ponchartrain lake front, and construction of levees to prevent river flooding. Man's activities, combined with the geologic setting and subsidence in this region are responsible for the unique landscape that was impacted by Hurricane Katrina.

Geotechnical engineering in Southeast Louisiana is far more complex than the "let's drive piles to rock" approach common in more geologically fortunate regions. This area is the delta of the Mississippi River, where Pleistocene-age erosion up to hundreds of meters deep has been refilled with normally- or even underconsolidated deposits, predominately clays. It thus shares geotechnical difficulties with similar deltaic areas around the world. The groundwater levels are high. Away from major rivers, the normal groundwater level is essentially at the ground surface. Near rivers or flooded canals, the potentiometric level can approach river/canal level, i.e., be well above the ground surface. Geologic subsidence is a long established phenomenon throughout the region, with subsidence rates of 0.5 to 1.5m per century. A big geotechnical problem here is evaluation of shear strength; sample disturbance in soft materials or those from great depths affects laboratory test results. Therefore, common practice today is to avoid unconfined compression tests for such materials in favor of triaxial or direct simple shear tests in the laboratory. In-situ vane shear testing has been used here for decades, but the cone penetrometer is becoming the preferred field strength determination procedure. Many projects such as levees, landfills, or large storage tanks are feasible only by using strength gain with consolidation techniques. This procedure was used by the Corps of Engineers to attain the levee heights they did achieve along the MRGO, Atchafalaya River, etc. Several landfills in the New Orleans area have reached heights twice those which would have caused failure without strength gain over time. Excavating can be difficult because of the combination of soft soils and high groundwater; hydrostatic bottom heave is often a concern. For levees adjacent to hurricane- or river-impacted water bodies, hydrostatic heave at the land side is a problem.

Twenty three samples were retrieved from eleven locations at the top of the levees around New Orleans. Thirteen were samples from Shelby tubes while ten were bag samples. These samples were tested in the EFA ort Erosion Function Apparatus. The results of such an erosion test is an erosion function which links the erosion rate in mm/hr to the water velocity in m/s. The results obtained show a large variation of erosion resistance among the soils tested and point to the need to evaluate the remaining levees. New ideas for such an evaluation and for improving the redundancy of such long linear systems are presented.

The most populous of the four major protected areas that suffered significant flooding during Hurricane Katrina was the main Orleans East Bank (downtown) protected area. Two sets of failures and breaches admitted floodwaters to this region during hurricane Katrina: a series of failures along the Inner Harbor Navigation Channel - IHNC (as discussed in the companion paper), and three major breaches on the drainage canals at the north end of this populous area. The IHNC frontage failures all failed to scour paths consistently below sea level, so flow through these ceased fairly rapidly as the hurricane induced storm surge subsided. The three breaches on the drainage canals, on the other hand, all rapidly scoured to well below sea level. Flow through these thus continued for three days, eventually filling the basin and accounting for approximately half of the overall loss of life and a similar fraction of the overall damages in this catastrophic event. Flow in through these breaches finally equilibrated with the still slightly elevated waters of Lake Borgne on Thursday, September 1. This paper presents a summary of the results of an investigation of the performance of the levees and floodwalls along these three critical drainage canals.

The Inner Harbor Navigation Channel (IHNC) sits at the heart of the three main populated regions of New Orleans. As shown in Figure 1, a number of breaches of varying severity occurred along the shores of the IHNC during Hurricane Katrina, contributing to the flooding of all three of the most heavily populated protected areas flooded in this event. This paper presents the results of geotechnical and geo-forensic investigations performed to determine the causes of the various failures along the shorelines of the IHNC. This includes the two massive breaches at the west end of the Lower Ninth Ward, and a number of lesser breaches and partial breaches (distressed sections) along other sections of the canal. There are many lessons here, not just for the greater New Orleans region, but for the nation and for the Profession as well.

In August 2005, Hurricane Katrina made landfall just east of New Orleans and inflicted widespread damage on the Hurricane Protection System (HPS) for southeast Louisiana. The storm surge produced by Hurricane Katrina in some cases overwhelmed the HPS beyond its design, but in other cases levee failures occurred at water levels well below their design. The response to this disaster by the US Army Corps of Engineers (USACE) included forming an Interagency Performance Evaluation Taskforce (IPET) to study the response of the system and, among many lines of inquiry, to identify the causes of failure of levees and floodwalls. Beginning in September 2005, the IPET gathered forensic geotechnical data from failed portions of levees and floodwalls. These data were considered perishable and had to be gathered quickly due to levee rebuilding operations. The performance of the levee and floodwall system provided valuable lessons. These lessons consist of: the need for resilience of the HPS; the need for risk-based planning and design approach; and the deficiency in knowledge, technology, and expertise in the hurricane protection system arena. The rebuilding efforts and future assessments and designs of hurricane protection systems will incorporate the lessons learned.

The recent flooding and devastation of the greater New Orleans region during hurricane Katrina represented the most costly peace-time failure of an engineered system in North American history. Extensive investigations and analyses have been performed by several major teams in the wake of this disaster, and some very important lessons have been learned. Many of these have very direct and urgent applications to levee systems in other regions throughout the U.S., and the world. Lessons include the importance of proper evaluation of risk and hazard; so that appropriate decisions can be made regarding the levels of expense and effort that should be directed towards prevention of catastrophe, and the levels of post-disaster response capability that should be maintained as well. The making of appropriate decisions, given this information regarding risk levels, is then also important. Also of vital importance are numerous "engineering" lessons regarding analysis, design, construction and maintenance; hard-won lessons with applications to flood protection systems everywhere. We must now do everything possible to capitalize upon these; and to prevent a recurrence of this type of catastrophe in the future.

The coastal wetlands and estuaries of Louisiana are one of the world's great ecosystems. For millennia, the Mississippi River has supplied the coast with an immense resource of freshwater, nutrients, and sediment to build a vast expanse of marsh and swamp land. Millions of people rely directly or indirectly on the marshes for their livelihood and for protection against hurricanes and storms. In the last several decades humans have impacted this ecosystem in many ways, especially by controlling rivers so natural floods are no longer a part of wetland maintenance and creation, and by building channels that expose freshwater marshes to salt water at an unnatural rate. As a consequence the natural surge buffering offered by the wetlands has been severely reduced. Louisiana is thus very vulnerable to hurricanes and major flooding as 70 percent of the population lives on 36 percent of land at or below sea level. Hurricane Katrina made landfall in Louisiana as a moderately fast-moving Category 3 hurricane the early morning of 29th of August, 2005. In greater New Orleans the flooding cost the lives of more than 1,500 residents and over 100,000 families were rendered homeless, the great majority of whom had heeded evacuation orders. The 1927 Mississippi River flood prompted a complete redesign of the Nation's strategy for managing the lower Mississippi River leading to construction of the Mississippi River & Tributaries Project (MR&T) that has successfully prevented a repeat of the 1927 disaster. In contrast, many observers have noted similarities between the patterns of surge-induced flooding that occurred during Katrina and the previous storm of record, Hurricane Betsy in 1965. The greater New Orleans hurricane protection system was, in fact, intended to prevent a repeat of Betsy disaster, but this system failed during Katrina with catastrophic results. A hypothesis for the a long-term hurricane protection system is presented for coastal Louisiana that focuses on levee systems coupled with wetland creation, restoration, and maintenance and has as its central theme, delivering sediment and fresh water from the Mississippi. This `survival' plan would not be cheap; however, it will ensure the continued existence of coastal Louisiana, its inhabitants, and unique culture.