The purpose of the Floris Project The government considers it important that the public has a better understanding of the probability of their area being hit by a flood. The government also wants to have a clear view of the relatively weaker areas in flood protection. Other experts have, moreover, indicated that the protection against flooding may no longer be properly in proportion to the consequences of flooding. The Flood Risks and Safety in the Netherlands (Floris) Project was therefore initiated in 2001 at the request of the State Secretary of Transport, Public Works and Water Management. The purpose of the Floris project is to gain an understanding of the consequences and the probability of flooding in the Netherlands. The project was conducted by the Road and Hydraulic Engineering Institute of the Netherlands Public Works Department (Rijkswaterstaat), in close cooperation with the Water Boards and Provinces. The results were released in the summer of 2005. Essence of the method The Flood Risks and Safety in the Netherlands Project has resulted in the further development and application of a new method which can be used to calculate the consequences of flooding. Detailed calculations have been made of the number of victims and the economic losses resulting from various flooding scenarios for three dike rings. The consequences have been determined more globally for the remaining dike rings. A new method has also been applied for determining the probability of flooding. The essence of the method is that various 'failure mechanisms', as they are known, can initiate a flood: not only extremely high water levels, but also instability in a dike or failure to close a hydraulic structure on time. Any failure mechanism carries a risk of flooding. The probability of all failure mechanisms together determines the risk of flooding in a dike ring. This method has been used to determine the flood risks of 16 of the 53 dike rings. The 16 dike rings were selected to give a representative view of safety in the Netherlands related to flooding. The calculations also show where the relatively weak locations in the water defences lie. A great deal of data was needed to be able to apply the new methods, including information about the subsoil under the dikes and hydraulic structures. In some cases, this data is surrounded by many uncertainties. An essential element in the probability calculations is that the order of uncertainty is expressly included in the calculation. The greater the uncertainty, the greater the probability. Further research can, in some cases, reduce the uncertainty. In which event the probability of flooding will also turn out to be smaller. This research will take place in the next phase of the Floris Project. Only then can the probability of flooding be established on a sound basis. Consequences of flooding From the study it appeared that in the event of flooding there could be anything between a few dozen to several thousand victims. Most are likely to occur if the flooding is unexpected and evacuation is therefore no longer possible. It appears from the most likely flooding scenarios that floods will occur unexpectedly. The maximum economic damage in the event of flooding of a dike ring ranges from 160 million in Terschelling to almost 300 billion in the province of South Holland. These amounts have been roughly calculated and show the damage which would occur if the entire dike ring was to fill up with water. For three dike rings the average damage has also been calculated in detail. During this process it was analysed in various flood scenarios which part of the dike ring would be inundated and how much damage would be caused as a result. From these calculations it appears that in the most likely flood scenarios 'only' part of the dike ring would be flooded. Only in the rivers region would the dike ring almost always be completely flooded. The average damage in the province of South Holland amounted to approx. 6 billion. The global method can therefore lead to huge overestimation of the damage, particularly for the larger dike rings which are divided into compartments by obstacles. Probability of flooding The study shows that the probability of flooding in the 16 dike rings varies from 1/2500 per year in South Holland to more than 1/100 per year in a number of dike rings in the rivers region. These figures give only an indication of the actual probability of flooding and cannot yet be seen as absolute values. The method is not yet robust enough for that. The calculations do, however, provide the opportunity to analyse which failure mechanisms contribute most to the flooding probability and where the weakest locations are in a dike ring. In the 1950s the Delta Committee established that extremely high water levels constitute the greatest threat of flooding. This insight provided the basis for the present safety standards for water defences. From the results of the Floris project it appears that this assumption is now no longer universally applicable. The probability of flooding due to high water levels is sometimes small compared with the risk due to other failure mechanisms. In most dike rings the failure mechanism of 'piping' constitutes the greatest threat. Here the water forms channels under the dike, causing the dike to collapse. The large probability is probably partly due to the uncertainties surrounding the subsoil under the foundations of the water defences. Further investigation at the sites in question can show whether there actually is a relatively weak spot. But it is clear that piping is a real threat in the sandy and clay subsoil of the Netherlands. With each high water the Water Boards carefully check the water defences for signs of this phenomenon. They are also prepared, if signs of piping are found, to take emergency measures, such as covering the dike with textile and sandbags. The effects of this human intervention are otherwise not included in the calculation of the probability of flooding. The failure mechanism of 'not closing hydraulic structures' also led to a high probability of flooding in a number of dike rings. In almost all cases this was because the closing procedures were not properly defined. This risk can be quickly and easily reduced by having the procedures documented and through regular exercises. Further to the Floris project, several Water Boards have now taken these measures. Flood risks The risk of flooding in a dike ring is the flood damage multiplied by the probability of flooding. Based on a rough calculation of the maximum flood damage, the risk in the 16 dike rings ranges from 0.1 to 180 million per year. In the three dike rings where the potential damage has been calculated in detail, the risk of flooding ranges from 2 to 37 million per year. The flooding risk can be seen as the amount that should be set aside per year to be able, in the long term, to compensate for the damage caused by a flood. In dike rings along the rivers the risks of flooding are relatively great. This is partly because the flooding probabilities along the rivers are greater. In addition to this, the consequences are large because if there are floods, almost the entire dike ring will be inundated with water. Other dike rings in most cases will 'only' partially flood. Value of the figures and how they can be used The Floris project is just one step in a longer development pathway. For all 16 dike rings the flooding risks have now been identified at the first development level. The calculated value of the flooding probability gives an indication of the actual flooding probability, but cannot yet be considered as an absolute value. It is possible to identify the relatively weaker locations in each dike ring and their causes. For a number of these locations it will first be necessary to investigate whether the probability of failure is actually great, or if it is due to uncertainty in the data. The Floris project has reached the second development level for three dike rings. These are the dike rings where the consequences have been calculated at a detailed level. The results at this level of development are robust enough to be able to compare the flooding probabilities and the flood risks with other similar types of dike rings. As soon as the flood risks of all the dike rings in the rivers region are available at this level, this will create an overview of the consequences of a flood and the weakest links throughout the rivers region. Priorities can also be set for similar types of dike rings, along the coast or in tidal river areas. Development level three will be reached in the future when the flood risks for all dike rings have been soundly determined, with an acceptably small margin of error. The flood risks of dike rings throughout the Netherlands can then be compared with one another. It is necessary to reach this level to be able to make a cost/benefit analysis of investments to be made in providing flood protection and to be able to evaluate whether the present standards offer sufficient protection. The total risk of flooding in the Netherlands can then also be compared in absolute terms with other collective national risks. Conclusions and Recommendations All those involved share the view that the method used offers added value. The calculations provide the most realistic picture of the probability of flooding based on current understanding. The calculated probability of flooding, however, is not yet robust enough for these figures to be considered as absolute values. Further research and development of the method could help to make the method more robust in the future. For most applications the national picture of the flood risks needs to be completed. Therefore it is recommended that the method also be applied to the remaining 37 dike rings. To obtain a proper estimate of the consequences, the detailed method needs to be used for all dike rings. More attention also needs to be focused on providing cost/benefit analyses for dealing with relatively weak locations. The study should continue, preferably coordinated from one central point, to be able to compare all the results. The mechanism of piping plays a major role in the present flooding probabilities and deserves further investigation. The study should focus on a method of calculating the probability of piping, reducing the uncertainty in the data and ways of reducing the probability of piping. In so doing it is also important, of course, not to lose sight of other failure mechanisms.