Newsletter No 2 June 2011

Floods in the Rhine basin

Aline te Linde


On the 12th of May 2011 Aline te Linde defended her PhD entitled “Rhine at Risk? Impact of climate change on low-probability floods in the Rhine basin and the effectiveness of flood management measures”. Her conclusions were dramatized in the media: “Dutch dikes should be raised by more than 3 m”. Below, you will find a more accurate description of her results.

Floods in the Rhine basin

In the past decades, the number of fatalities and the economic loss caused by floods have worldwide increased considerably. The river Rhine in North-West Europe also has a long history of floods that have caused casualties and severe damage. Assessing the effect of climate change on flood risk, and maintaining and planning flood management measures are urgent issues for the riparian countries within the Rhine basin. Since safety levels along the Rhine are relatively high (the dikes are designed to withstand floods with occurrence probabilities of 1/200 to 1/1250 per year), the estimation of the size and duration of flood peaks at these low probabilities is very relevant to water managers. Many methodological challenges remain on simulating the effect of climate change and flood management measures on runoff, particularly when estimating the probability of occurrence of extreme flood peaks. In addition, no future basin-wide flood damage estimates exists to reflect socio-economic and land-use change.

The thesis investigated the combined effect of climate change and socio-economic projections on flood risk in the Rhine basin using improved simulation methods. The focus is on the estimation of low-probability flood events and anticipating to the impacts of these events through the development of cross-boundary flood management measures.

Improved simulation method

Carefully coupling of different methods and models improved the simulation of low-probability flood events (i.e. 1/200 per year, or less). A weather generator was used to create long series of resampled meteorological, which considerably decreased the statistical uncertainty within extreme value analysis. To estimate the impact of climate change on river discharge, different downscaling methods were applied to convert data from Global Climate Models to the required regional scale for hydrological modeling in the Rhine basin. Furthermore, a basin-wide flood damage model was developed to estimate current and future potential flood losses. These damage estimates can be combined with flood probability to estimate the annual expected loss (i.e. flood risk) in the Rhine basin.

Increase in flood-peak probability due to climate change

The simulation results indicate a substantial increase in flood-peak probability in 2050 throughout the Rhine basin, when compared to the control climate period of 1961-1995. At gauging station Lobith, at the German-Dutch border, the probability of flooding (currently the safety level is set at 1/1250 per year) is expected to be three to five times as high in 2050 as it is today.

Increase in basin-wide flood risk
It appears the annual expected loss may increase between 54% and 230% in 2030 compared to 2000, of which ~ three-quarters of the increase can be attributed to climate change, and the remaining part to socio-economic developments. Results show that the area with the highest flood risk is located in the Lower Rhine in Nordrhein- Westfalen in Germany, and not in the Netherlands as previously believed (Figure 1). This is mainly related to the high safety standards in the Netherlands.

Upstream flooding beneficial for the Netherlands
In the current situation, safety levels of dikes and other flood defense structures decrease in upstream direction along the Rhine. Consequently, flooding is more likely upstream in Germany and France, compared to the Netherlands. It has been shown that these upstream floods have a profound decreasing effect on discharge peaks at Lobith. Hence, they are favorable for reducing flood risk in the downstream areas of the Netherlands. However, flooded water in the German Lower Rhine might enter the Netherlands via other routes than the main Rhine channel, for example, from Emmerich to Doetinchem, or from Kleve to Nijmegen.


Figure 1. Potential damage (a) and flood risk (b), aggregated to seven regions along
the Rhine.

Most measures not effective in future climate scenarios

Various flood management measures in the Rhine basin have already been developed according to the Action Plan on Floods (APF) that was initiated in the 1990s by the International Commission for the Protection of the Rhine (ICPR). The effectiveness of the currently implemented and proposed measures in the APF, as well as several additional measures, such as extra retention areas, was evaluated. These measures, though, reduce peak water levels only by several centimeters and, as result, seem inadequate to cope with the increased flood levels that are expected in future climate scenarios. According to the results of this thesis, the only measure that can prevent the Rhine from flooding at current safety levels in 2050, is drastic, basin-wide, dike raise on the assumption that these dikes cannot fail. The necessary dike raise at Lobith is estimated at 1.30 m, and varies along the Rhine to a maximum of 3.30 m at the Middle Rhine in Germany.

Some thoughts on uncertainties and an advice towards flood risk management

However, Aline te Linde does not advocate a strategy of further dike raise in the future. She explains that although her research has been able to decrease the statistical uncertainty in estimations of low-probability flood peak events, a residual uncertainty remains for two reasons. First, due to the relatively short time series of available discharge measurements, estimating extreme events at high safety levels requires either statistical or hydrological modeling far outside the calibrated range of those models. Second, the impact of upstream flooding and climate change adds to the unpredictability of extremes.

Increasing safety levels to reduce the probability of flooding, which is currently considered in the Netherlands, will force engineers to even further stretch their models. It will increase uncertainty in the estimation of the accompanying peak discharge, no matter how far models have improved. Water managers should be more aware of this. An effective climate change adaptation policy on flood risk management should embrace inherent uncertainties, and not only address flood defense measures, but also consider a wide range of other adaptation options on damage reduction.