Back to search results

Flood risk reduction through natural retention and physical protection: Flood protection design levels –a risk-based approach

At present, flood protection design and the return period the protection should be built to is typically based on land use. Agricultural land has the lowest protection level, industrial land may have a higher one, populated areas of villages, towns and cities are protected usually to a 100-year return-period flood, and certain special areas may be protected to the level of a very rare event, such as a 1000- or 10,000-year return period event. However, these decisions do not usually reflect a risk-based approach, but instead are applied universally. Yet often it is more sensible to look at the protection requirements from the perspective of undesirable or unacceptable consequences – one might recall the tsunami that overwhelmed a nuclear power plant inJapan in 2011, or the impact that flooding affecting chemical industries or large industrial estates would have. Dresden provides a positive example of a risk-based approach that was used to prevent the city center flooding again in 2013, after the main railway station and historic buildings flooded in 2002. These examples demonstrate how important a more consistent approach to risk-based decisions can be. When designing protection, it is important to consider the following points. These underscore the difficulties of factoring in variables, quantifying key factors, and/or planning ahead from an economic standpoint: • The overload case: modern protection structures such as levees must not fail in an event that slightly exceeds, for example, the 100-year return period event, even if technically they were designed for 100-year return period events – they must retain a minimum of protection even if they are ‘overloaded’ and leave enough time and provide options to consider alternatives for defense or repairs and intervention. They cannot simply be allowed to collapse ‘catastrophically.’ • Uncertainty: it is difficult to deal with uncertainties, especially when constructing a device that will function up to a ‘hard’ threshold. In order to reduce the probability of failure in a changing and dynamic environment, various assumptions must allow for uncertainties. These include, among other things: the uncertainty inherent in flood statistics used to calculate a 100-year return period (or other protection level) event; the uncertainty that the data available to produce these statistics (usually a few data points for extreme events during a record of the last 50 to 100 years for larger rivers) is representative of the realm of possible events (often it is not); the uncertainty that conditions in the future will remain basically the same, or that future conditions can be anticipated in current protection level design; and finally the uncertainty that the protection system’s condition will still meet the design level despite degradation and use over its lifetime. • Changing risk: are the assumptions still risk-based in tomorrow’s world, or will a changing risk landscape – or changing requirements and technical feasibility of flood protection – mean that other forms of protection replace the ones in existence? We may need other forms of protection based on the assumption that risk will change in tomorrow’s world.

PERC Reports