Esri News -- ArcNews Fall 2002 Issue -- Rhine Flood Hazard Mitigated With the Help of GIS

The river Rhine has been one of Europe's most important waterways for centuries. Its borders are shared by no less than five countries--Austria, France, Germany, the Netherlands, and Switzerland. In addition, water from Belgium, Italy, and Luxembourg flows to the Rhine. Many important industrial production sites are concentrated along the river course, among them a lot of facilities producing hazardous chemicals. Approximately 10.5 million people live within the wider flood prone areas, which have undergone a tremendous man-made change. From its source in the Swiss Alps the Rhine flows through Lake Constance and defines the border between France and Germany before it finally flows into the Dutch North Sea. A long historical record of flooding is documented in the archives. Recent devastating floods in the 1990s caused considerable damage and are well remembered by the responsible agencies as well as the local population.

An extensive system of levees in France, Germany, and the Netherlands, built and maintained during the last few decades, provides for a certain level of protection. It was recognized, though, by all affected countries that the level of safety provided by this system may not be sufficient for extreme events when floodwater levels exceed the levee height or when levees fail.

The Rhine countries are organized in a multilateral committee, the International Committee for the Protection of the Rhine (ICPR), based in Koblenz, Germany. In 1998 the ICPR enacted the Action Plan on Flood Defense. It was agreed to reduce damage risks by 25 percent within a 25-year period. The action plan mentions four targets that should contribute to that objective:

Reducing damage risks up to 25 percent by 2020
Reducing extreme flood stages up to 70 cm by 2020
Improving awareness through the preparation of hazard and risk maps
Improving warning systems

Before any technical measures were taken, the committee commissioned a study to produce maps showing the extent and the inundation depths of a 10-year flood, a 100-year flood, and an extreme event. The maps should also depict the monetary values potentially at risk during an extreme flood event.

A team of hydrologists, economists, and GIS experts was commissioned to carry out the study and produce maps at a scale of 1:100,000 to be published as an atlas. The study was carried out in 2000 and 2001. The bound version of the atlas and a CD-ROM version with 41 double maps were published in March 2002.

Data Requirements

An important issue in the study was the data on which the hydrologic and economic models were to be based. The type of data needed was soon defined, but data availability, quality, model, format, scale, and coordinate systems varied greatly. In total, six different coordinate systems had to be integrated into a single reference system. For vector data these issues are relatively easy to handle; however, projecting raster digital elevation models and even scanned topographic maps proved to be a time-consuming effort. The data came from several subnational, national, and European agencies. Topographic agencies contributed digital elevation models at several resolutions. In some areas light detection and ranging (LIDAR) data provided by water agencies could be used. The European Union provided a 100-meter resolution land use data set derived from satellite imagery, and several national statistical agencies helped with socioeconomic and demographic data. The water agencies finally brought in the hydraulic data with water levels for several flood recurrence intervals along the entire course of the river at 100-meter intervals. The rest was experience brought in by the experts.

Hydrodynamic Modeling

GIS played a vital role throughout the entire project. It was used for data preparation, data unification, data visualization and, most important, for data modeling. The central tool was the ArcView extension FloodArea, developed by Esri Business Partner geomer GmbH, based in Heidelberg, Germany, and Ruiz-Rodriguez and Zeisler Water Engineers, Wiesbaden, Germany. The modeling solution geomer could present to the ICPR was familiar to the German Federal Institute of Hydrology and thus considered ideal for the purpose of the project. FloodArea allows for hydrodynamic modeling within the ArcView Spatial Analyst extension. It calculates the inundation depths to be expected in the floodplain from certain water levels in the river. Since there is a gradient in the river water level, it had to be considered; therefore, a simple intersection of an extrapolated river surface and the terrain itself was not sufficient. The flow paths also had to be calculated, which means the relevant water level for a point 10 km away from the river course is not necessarily the one with the shortest distance to the river, but the one from where the water actually comes.

Assumptions had to be made for an extreme event, since there is no such scientifically defined term. In a rather pragmatic way the committee defined extreme water levels for various sections of the Rhine based on the experience of past flood events. Furthermore, it was assumed that flood protection measures, levees in particular, would fail at the most undesirable locations. Different assumptions had to be made because of different flow regimes and river characteristics. The hydrodynamic modeling process is initiated by the water level, attached as an attribute to a GRID theme. A hydrodynamic approach (Manning and Strickler) is then used to distribute the water to neighboring raster cells until the water has leveled. The inundation depth is then saved as a new GRID theme.

Economic Modeling

The degree of potential damage is defined by the depth at which a certain object is underwater and the object's value. These relationships are called damage functions. They depend on the economic values of the administrative unit and can be derived from the experience of damages registered in past flood events. The economist for the project provided the required formulas. The values were estimated from the land use data set. The functions were then calculated and attached to the inundation raster data, thus providing for the possibility of a monetary risk attached to each raster cell. Summary statistics were calculated for various spatial entities, such as municipalities, second order political divisions, countries, and river sections, all easily made possible using ArcView Spatial Analyst tools.

Map Production

For cartographic purposes, both the inundation depth data and the monetary risk data were classified using the ArcView Spatial Analyst extension, subsequently vectorized and processed for printing. The atlas contains 41 double maps, one of each double map showing the inundation depth on opaque paper, the other one showing the monetary damage and the number of people at risk on a transparent sheet that can be overlaid.

With the new atlas, ICPR has a valuable tool for creating disaster awareness among the public and the responsible authorities. The action plan now has a sound base on which further measures can be determined. GIS proved to be an invaluable instrument during the entire process.