analysing a large ensemble of 17 simulations with a global climate model from the
ESSENCE project. We looked only at the present-day climate here. Whether climate
change affects these joint probabilities is not known. For more, see KNMI scientific
The storm surge barrier near Hoek van Holland, the Maeslantkering, protects the densely
populated Rotterdam area when high sea levels occur due to wind storms, the tide
and pressure effects. It closes automatically when the water level is predicted to
exceed the mean sea level by more than 3m. A worst case scenario is that of simultaneous
occurrence of extreme storm surge and extreme discharge of the rivers Lek and Rhine.
In the event of an extreme discharge alone, the barrier should remain open to prevent
the damming of excess water. In the event of a surge, the barrier should be closed
to protect the densely populated Rotterdam area.
In present estimates of the probability of such a worst case scenario, it is assumed
that the storm surges and high discharges are independent events. But since both
rainfall events and winds storms are caused by the same weather systems – synoptic
low pressure systems often called depressions – this assumption of independency could
be challenged. Because both events are rare it is, however, difficult to establish
their dependency. For this reason we could not use observational time series, but
had to rely on the much longer time series generated by a climate model. We used
the control period 1950-1980 of all 17 members of the ESSENCE global climate model
ensemble, thus providing us with 500 years of simulated data for the present-day
Probability of the exceedance of the 99th percentile of NNW wind. After a 10-day
precipitation extreme the probability of exceedance is approximately 0.04 (red diamond),
which is a factor 4 higher than the probability of 0.01 (by construction) in the
full data set (blue diamond). The increase in probability is outside the range that
could be expected due to chance because of the much smaller sample size (blue bars).
What we found:
Events of (multi-day) extreme precipitation in the Rhine catchment area and storm
surges are not independent as is commonly assumed in computations of safety norms.
After a period with abundant rain in the Rhine catchment area the probability on
a storm surge from the NNW (north-north-west) direction could be 4 times larger than