In August 2005 Hurricane Katrina swept across the Mississippi Delta bringing widespread devastation to the US city of New Orleans. The levees designed to contain floodwaters were breached in many places and 80% of the central districts were flooded.
In the wake of the storm, the US Army Corps of Engineers (USACE) was mobilised to provide immediate support and investigate the causes and events leading up to the disaster.
USACE has an important civilian as well as military role and it was in this capacity that it was tasked with providing immediate support and overseeing reconstruction works. High-Point Rendel director of group strategy Scott Steedman has been a consultant to USACE for many years and was invited to join the investigation task force, known as IPET, to co-lead the team assessing levee performance.
Steedman discussed the history, growth and layout of New Orleans – particularly its hurricane protection infrastructure. The city is built on the Mississippi Delta, which is continually settling as delta sediments consolidate. A series of canals and levees, which originally crossed an area of swampy marsh, link the city with Lake Pontchartrain to the north.
The river itself is canalised between large flood protection levees up to 6.1m higher than the city in some areas. Over the years, pumping has lowered the local water table and New Orleans has expanded northwards on to the previously swampy marshlands and towards the sea, leaving it more vulnerable to flooding.
The entire hurricane protection system, including levees, floodwalls and pumping stations, has been in a near-continuous state of upgrading since 1965 and the then-current programme was not scheduled to be complete until about 2015. During Hurricane Katrina, levees were overtopped and breached in many places, contributing heavily to the flooding and damage to the city.
Hurricane Katrina was portrayed in the press as a storm of unprecedented scale – an event that would rouse the US to the reality of climate change. However, Steedman said Katrina's size was within the normal range of expected storms in the area and that flooding of the city was neither unexpected nor unprecedented.
Within the last half century New Orleans had been flooded on at least three notable occasions – 1947, 1956 and 1965. Evidence suggests it was the storm path taken by Katrina, not simply the hurricane's intensity, that controlled the degree of damage inflicted on the city this time. Using IPET simulations, Steedman showed how the storm path brought a surge of water up the east side of the Delta, which then impacted on the Gulf Coast and shoreline around New Orleans.
He compared the trajectory of Hurricane Katrina with a USACE map prepared in 1964 showing the worst case theoretical trajectories that could affect the city. Katrina's path is very similar to one of the critical design cases.
Steedman explained that the floodwater came from a combination of sources. In many areas over 305mm of rain fell during the hurricane. In addition, seawater came through open floodgates, flowed backwards through pumping systems and overtopped and breached levees and floodwalls.
Steedman showed the results of inundation modelling that attempted to recreate the timeline of the flooding and high water levels.
The team carrying out the modelling incorporated all sources of water in detail, taking the complex drainage systems of a city into account using drains, streets, canals, pumping systems and known levee breaches. This enabled it to postulate what would have happened if the levees had not failed. The team concluded that, in the central area at least, about 60% of the flooding was the result of levee breaches. It was therefore of great importance to determine what had happened to the levees.
Steedman reminded the meeting that the city centre levees were built in an area previously occupied by swampy marsh and typically clay structures founded on marsh peat, underlain by either sand or clay depending on the location.
In recent decades, initiatives to raise the level of protection had used floodwalls, known as I or T walls. The use of walls may have been to avoid additional land take as since their original construction, in many areas, houses had been built very close to the downstream shoulder. The unpropped I walls were installed through the crest of the levees and sometimes into the underlying stratum.
Steedman presented the results of centrifuge modelling, showing the possible failure mechanism for the different levees as the water level was raised. Video stills of the model showed development of the failure as the unpropped wall allowed a water-filled crack to form on the canal side.
Depending on the geometry and foundation, the hydraulic force was sufficient in the model to lead either to a translational failure mechanism in the underlying stratum or a rotational failure towards the downstream side of the levee.
This mechanism, which was duplicated in numerical analyses, closely matches the field observations where the upstream flank of the levee was often left intact and the failed downstream flank was swept away. Steedman also said in areas close to failed levees, underlying sand or peat strata had been exposed by the failure and large volumes of sand or peat lumps were found washed into the local area. In the case of the 17th Street failure, the strength of the shallow peat layer appeared to be less significant than the strength of the underlying clay as the shear surfaces had developed in the clay rather than the peat.
The IPET report concludes that the water-filled crack mechanism was a key element contributing to the levee breaches in the city centre. Steedman said this mechanism was well documented in UK literature and guidance but did not appear to have been considered in the design process in New Orleans.
Outside the city
The IPET team also carried out a major investigation into levee breaches outside the central area, where scour and erosion appeared to have been a major contributor to the damage. The IPET report categorises the level of damage on tens of miles of levees and identified particular weak links, for example, much of the damage recorded during these inspections occurred at locations where soft flood defences met hard structures.
Steedman described how the IPET survey team had worked hard to establish the true height of the levees and flood defences. They found that in some cases levees had been designed and built based on datum levels not current at the time, and that with continuing subsidence large parts of the hurricane protection system were in fact 610mm or more lower than people had thought. However, Steedman said breaches in the city centre occurred before floodwaters reached the top of the floodwalls, so even if the defences had been at the correct elevation this might not have prevented the levee breaches.
Steedman concluded his presentation with a reference to the social impact of the disaster and the sad loss of culture and population of New Orleans. After the flooding, much of the population was relocated and social infrastructure was not replaced. It seems likely that many former residents will not return. He said: "How do you come back from that?"
In view of the need to learn lessons from failures in New Orleans, the discussion focused mainly on how to disseminate information within the industry, both internationally and locally.
Malcolm Bolton (University of Cambridge) said although the mode of failure seen in the New Orleans' levees was well understood, and is also incorporated into BS8002, this was apparently not taken on board by the levee designers.
He asked how the Institution of Civil Engineers (ICE) could assist engineers to inform themselves of work done in the past and in other countries without resorting to "bigger codes of practice". Steedman concurred and stressed the importance of continuous technical self-development and passing on wisdom to younger engineers. He said it was a shame that new knowledge takes so long to pass into current practice. He added that old structures are designed to old standards and it may be valuable to ensure important structures are reassessed in the light of current knowledge.
Independent seismic consultant Edmund Booth stressed the need to identify the risks and consequences of potential failures in ageing infrastructure. Steedman said he saw this as one issue that must be industry rather than ICE driven, but that it was an important lesson to take on board.
Another speaker said young engineers do not have time to familiarise themselves with state-of-the-art knowledge. He suggested engineers should be encouraged to visit libraries and speak to senior colleagues about their work. Steedman replied that the BGA should encourage groups to form expert panels and take a public role in the issues of the day to provide "a more intelligent insight into what's happening in the field today".
BGA chairman Hilary Skinner (Whitbybird) asked if, in Steedman's experience as ICE vice-president, the ICE could use its influence to affect government policy. Steedman replied that applying the lessons of Katrina to flood defences in the UK had not really been discussed. The Dutch had apparently been taking the results of the investigation on board, but in the UK the attitude appeared to be "it wouldn't happen here".
Independent consultant Angus Skinner said many areas of levee must have been on the point of failure after the disaster and asked if USACE was changing the levee's design as it reconstructed them. Steedman replied that its immediate action had been to construct gates at the ends of each canal to control the inflow of water in case of a repeat event. Levee heights were strengthened in other areas, but he emphasised that the same thing could happen again if a similar intensity hurricane were to follow the same trajectory into the city.
It was clear from the discussion that making information available via publishing alone is not enough to keep engineers informed of established knowledge because of the volume of existing guidance and data.
It is therefore valuable to encourage communication between engineers of different disciplines and levels of experience to discuss past and present projects, failures, successes and technical issues within project teams.
The key lesson to be learned from this meeting is to keep trying to resolve the gap between published wisdom and engineering practice.