Greater New Orleans is home to 1.2 million people. Located between the Gulf Coast, the Mississippi and a large inland lake, almost half of its territory lies below sea level. Every year hurricanes build up in the Gulf of Mexico and sweep inland, often causing devastation and flooding. The Mississippi river is also prone to flooding. The city’s flood defences are on a vast scale, comprising 350 miles of levees and barriers, but their weaknesses were exposed when Hurricane Katrina struck in August 2005.
Katrina was a slow moving storm with massive amounts of rainfall and an unexpectedly severe storm surge in the Gulf of Mexico. Multiple breaches of the levees and floodwall defences resulted in the flooding of some 80 per cent of the city. The floodwaters penetrated up to six miles inland, caused over 1,000 deaths and destroyed or severely damaged over 200,000 homes and businesses. More than 800,000 people fled the area.
In 2006, Royal HaskoningDHV was commissioned to advise the United States Army Corps of Engineers (USACE) on the redesign and rebuilding of the entire Hurricane Storm and Damage Risk Reduction System. Since then they have provided expertise in hydraulic modelling and design, flood forecasting, GIS-based mapping and risk management.
Tasks have included establishing baseline data, analysing flood defence performance, and developing and applying innovative modelling and data management tools. An American company, Haskoning Inc., was set up with a New Orleans office staffed by experts from the Netherlands and the UK.
The client’s overall objective was to provide a “100 year level of protection”, or more accurately to protect the flood prone areas against all except the worst ‘one per cent’ of hurricane events. This apparently simple definition raises some interesting questions, not least what this means in practice and what happens when one of the worst ‘one per cent’ hurricanes actually occurs.
Mathijs van Ledden explains: “The Federal Emergency Management Agency (FEMA) guidance said that the levee height should be one foot above the height of the ‘one per cent’ wave, the extra foot being added to allow for uncertainty in the design variables. However this approach did not take account of different levels of uncertainty and degrees of risk. For example, if two locations had the same expected wave height for a given storm, but the forecast for one location was more uncertain for the other, it would make sense to provide additional cover for the risks at the second location.