To the Editor —

Hallegatte et al.1 present current and future average annual losses or economic risk for various major coastal cities at a global level. But to get a better understanding of these risk levels, sophisticated flood risk analyses at a local, city level will be needed.

For their analysis, Hallegatte et al. estimated the return period of flooding on the basis of available information and their own expert assessments. For many cities they used a return period of flooding equal to the defence standard of the region. However, in many countries (such as the Netherlands2) a substantial part of the defences are less safe than the standards require. Existing defence standards generally refer to the frequency of failure due to overtopping or overflow. But other geotechnical failure mechanisms, such as those caused by the instability and seepage that occurred in New Orleans3, can lead to failure when water levels are below the crest of the defences, thereby increasing risk. In addition, various different flood events have different return periods and impacts. For most cities in deltas — examples include Rotterdam, Jakarta and Shanghai — both river and coastal floods can be a threat.

Owing to these factors, the actual protection levels could differ by more than a factor of 10 from the protection standard4, and the effect on risk will be similar. For a realistic analysis of expected losses, more advanced approaches to flood risk analysis will be needed. These methods must take into account the various defence types, failure mechanisms and flood scenarios5. Future risks will also depend on subsidence and sea-level rise, and on how cities develop in flood-prone areas6.

Most major cities in the developing world and the USA still use relatively low defence standards with 100-year return periods, and Hallegatte et al. rightly say that the estimated risk levels will necessitate a higher demand for safety for many cities. Cost–benefit studies for the Netherlands7 and New Orleans8 suggest that optimal protection standards for urbanized areas should generally be around 1,000-year return periods or even higher. Only a few fast-growing cities, such as Shanghai, have already adopted these higher defence standards.

Future investments in adaptation to sea-level rise and modification of standards will be high, but are generally small relative to gross domestic product (GDP) and the financial value of damages and risks. For the Netherlands, yearly investments in adaptation of the defences for sea-level rise are estimated to be 0.12% of the current GDP in the year 2025 and 0.14% of GDP if the standards are also heightened9. The key question is whether the authorities in the fast-growing delta cities at risk are willing and able to make the much-needed investments in adaptation to sea-level rise, better protection standards and other forms of risk reduction. Various hard (dykes, floodwalls and storm surge barriers) and soft (wetlands and nourishments) measures are available to effectively reduce the flood risks in coastal cities10.