Recurrence intervals for the closure of the Dutch Maeslant surge barrier

The Dutch Maeslant Barrier, a movable surge barrier in the mouth of the river Rhine, closes when there is a surge in the North Sea and the water level in the river at Rotterdam exceeds 3m above mean sea level. An important aspect of the failure probability is that the barrier might get damaged during a closure and that, within the time needed for repair, a second critical storm surge may occur. With an estimated closure frequency of once in 10 years, the question of how often the barrier has to be closed twice within one month arises.

Instead of tackling this problem by the application of statistical models on the (short) observational series, we solve the problem by combining the surge model WAQUA/DCSMv5 with the output of all seasonal forecasts of the European Centre of Medium-Range Weather Forecasting (ECMWF) in the period 1981–2015, whose combination cumulates in a pseudo-observational series of more than 6000 years.

We show that the Poisson process model leads to wrong results as it neglects the temporal correlations that are present on daily, weekly and monthly timescales.

By counting the number of double events over a threshold of 2.5m and assuming that the number of events is exponentially related to the threshold, it is found that two closures occur on average once in 150 years within a month, and once in 330 years within a week. The large uncertainty in these recurrence intervals of more than a factor of two is caused by the sensitivity of the results to the Gumbel parameters of the observed record, which are used for bias correction.

Sea level rise has a significant impact on the recurrence time for both single and double closures. The recurrence time of single closures doubles with every 18cm mean sea level rise (assuming that other influences remain unchanged) and double closures double with every 10cm rise. This implies a 3–14 times higher probability of a double closure for a 15–40cm sea level rise in 2050 (according to the KNMI climate scenarios).