Journal of Risk Research 7 (7–8), 745–757 (October–December 2004)

Thinking the unthinkable – the end of the Dutch river dike system? Exploring a new safety concept for the river management BERT ENSERINK* Faculty of Technology, Policy and Management, Delft University of Technology, PO Box 5015, 2600 GA Delft, The Netherlands

Abstract Since 1100 the Dutch relied on their continuously expanding extensive dike system for keeping dry feet and dry goods. But how durable and safe is this traditional dike concept, now the sea-level is rising and rainfall patterns seem to change? Can they continue to raise their dikes in the future or should they give more room to the river? In the Dutch mind higher dikes are saver. In practice, however, higher dikes may lead to higher risks as the consequences of failure rise. What is wrong with the risk perception of the Dutch citizen? How should risks be communicated to the public, how should a public debate on safety regimes be organized? In a study project a new safety concept in which the Dutch rivers got a free flow was explored and communicated. In an effort to reframe the issue a transition path, scenarios, impact studies and ex ante evaluations of this new safety management regime were made. A new safety paradigm seems to be taking shape. Context scenarios show under what circumstances frequent flooding can be made acceptable to the Dutch citizen and the inhabitants of the Dutch polders. KEY WORDS:

risk communication; safety management; river management; flooding; scenario analysis

1. Introduction The Dutch foster their heritage of eight centuries of dike construction as primary defence against flooding (TeBrake, 1985). This history led to the seemingly undisputed dominance of this way of coping with high waters in the Netherlands, while elsewhere in the world other approaches have become dominant like acceptation, adaptation, evacuation, insurance and zoning strategies (Hekal, 2000; DWW, 2001:9; Liu and Chan, 2001). Such a dominant perspective on coping with a problem can be called ‘framing’. A frame is a perspective that is used to make sense of an amorphous complex situation and provides guideposts for knowing and acting (Rein and Schön, 1993). Dikes have become the dominant frame for defences against the ‘waterwolf’ in the Netherlands as only remnants of artificial mounts like the ‘vlietbergen’ in Sealand and ‘terpen’ and ‘wierden’ in the northern provinces are the silent witnesses of older coping mechanisms. The Ministry of Transport, Public Works and Water Management’s Directorate General of Public Works and Water Management’s (further: Rijkswaterstaat) is responsible for *E-mail: [email protected] Journal of Risk Research ISSN 1366-9877 print/ISSN 1466-4461 online © 2004 Taylor & Francis Ltd http://www.tandf.co.uk/journals DOI: 10.1080/13669870210166185

746

Enserink

protection against floods. Traditionally it approaches issues of safety and risk in a probabilistic way, which is dominant in civil engineering. This approach rests explicitly on the rational paradigm of policy making and analysis and implicitly assumes that nature has to abide for engineering skills. As a consequence of this frame river safety is seen as a technical, engineering problem in which mathematical modelling and three-dimensional dynamic models combined with cost/benefit analysis provide the right answers. The agency’s engineers use probabilistic arguments that use past knowledge to predict future effects, in which risk is perceived as the probability of the realisation of an adverse event (Keeney, 1980; Zhou, 1995; Rogers 2001:5). This is, however, only one way of framing the complex problem of river safety and river management. The general public as well as many experts themselves find statistics and probabilities hard to understand (Camerer and Kunreuther, 1989; Huff, 1993; Hoek, 1995). Analogous to Slovic et al.’s (1978) a statistical chance of 1/1250 of a discharge above the design discharge is hard to grasp. Its translation into living through a major river flooding during your – statistically 80 years – lifetime with a chance of 8% is somewhat more informative as it compares to the chance of getting involves in a car accident or getting specific diseases during your lifetime. But public understanding of risks is not rational; opinions and perceptions of risks and hazards change rapidly after incidents (Rosenthal et al., 1998; Poortvliet, 1999). Two occasions of extreme high waters – not even flooding – in the 1990s brought about a major change in the public judgement of the risks of flooding and spurred the national government for an emergency plan for major dike improvements. Rijkswaterstaat’s probabilistic approach often competes with a second framework: the deterministic approach, which is more popular at the Ministry of the Interior and Kingdom Relations. This deterministic approach uses ‘scenarios’ – defined as descriptions of all possible events, as a basis for estimating the consequences of occurring failures in technical installations. In this approach, possible events and coincidences or successions of events lead to an accident or disaster. It is the standard approach to safety estimates in the chemical process industry, where it is usually also described as ‘probabilistic’. Risk here is framed as a distribution of possible events, scenarios, with associated effects and probabilities. The attention for extent of the adverse effects is what distinguishes the two approaches. The focus in the latter is on the product of the probability of events and the magnitude of specific consequences (Lowrance, 1976). With respect to these adverse events or scenarios Rogers (2001) describes ‘these scenarios move from a situation in which there is limited uncertainty in which there is sufficient scientific basis to establish the probabilistic relationship between hazard and harm to a situation in which the uncertainty is tending to ignorance’. Framing the river safety problem in probabilities of flooding, number of casualties, costs and financial losses turns out to be problematic too, as there is substantial uncertainty on the structural weakness of dykes as well as on the socio-economic and psychological impact of flooding. Rogers (2001) says about the use of scenarios in risk communication ‘the most common of such scenarios is the “worst case” scenario’. Consequently not the chance at failure but the scope and impact of the consequences of such failure and the social and political acceptability of these risks dominate political debates. Although the risks of nuclear energy are very different from the risks of flooding, the nuclear debate provides a clear example of this change of scope towards the consequences of failure. ‘Small chances – huge effects’ (Dijk and Smit, 1976), for instance, was a very influential report in the debates on expanding the capacity for nuclear power production in the Netherlands during the

Thinking the unthinkable

747

1970s and early 1980s. That small chances do come true was painfully illustrated by the Harrisburg accident and Tsjernobyl nuclear disaster. Moreover, Tsjernobyl illustrated that probabilistics did not take into account the ‘human factor’: operators willingly ignoring safety regulations, which in the end led to the melt down of the nuclear core. Consequently, the impacts, rather than chances became dominant in the nuclear energy debate throughout Europe. In policy making the probabilistic and deterministic frames often collide as was beautifully illustrated in the debate on the safety of the Westerschelde tunnel, a 7 km long tunnel with his deepest point 60 metres below sea-level, currently under construction in the Netherlands (Enserink, 2001). In this case this collision led to over $100Mln for additional safety measures, especially additional interconnections between the two major tubes, of which nobody really knows whether they contribute to the safety of the user. From a probabilistic perspective, chances at failure rise by making additional interconnections especially in the construction phase, while from a deterministic point of view the consequences of eventual failure and accidents will be less. The tunnel safety debate showed that a discussion between holders of different frames soon leads to a dialogue of the deaf (Bras-Klapwijk, 1999). Such a dialogue of the deaf has been going on for years over the issue of river dike improvement where the debate focused on safety levels, costs and the cost of smart solutions that would save natural and cultural values (Eeten, 1999). This will be returned to later. Why is communication on risk and safety so hard? Risk communication, as Renn (1998) states, serves three main objectives:

• • •

to make sure that all receivers of a risk message are able and capable of understanding the meaning of the message; to persuade receivers to change and modify their attitudes; to provide the conditions for a two-way communication process.

Neither the probabilistic nor the deterministic frame discussed above is easily communicated to the public, nor persuasive or debatable. The technicians and safety experts using these frameworks do not occupy themselves with two-way communication; they are statisticians. From a communication perspective, however, the probabilistic approach is too rational, too technical to be understood by laymen: the meaning of the message is not clear. Moreover, although deterministic worst-case scenarios do come true (Tsjernobyl, recent tunnel fires in Europe, and 11th of September), they are too gruesome to be believed. Moreover, when it comes to decisions the lure of concrete short-term gains can override the vague risks of long-term disaster as accusations of the Italian company Panaviation for selling unapproved used spare parts for aircraft of January 2002 illustrated (Volkskrant, 2002). The rational or even mathematical heuristics of both frames, worthwhile and necessary in themselves, do not take into account the human factor, the mistakes, miscommunication and misconduct, nor do they provide conditions for two-way communication on risk and safety. These rational approaches of safety and risks do not relate well to social acceptability and practical social acceptance. A different approach or at least another way of risk communication with the general public is required; a different discourse (Fischer and Forester, 1993) and reframing of the issue at hand is needed. Scenarios – not Rogers (2001) descriptions of possible disastrous events, but policy analytical scenarios depicting possible future states of the system

748

Enserink

itself and its environment (Beroggi, 1997), might be a good way for (re-) framing ideas and knowledge on safety and river management issues. Moreover, as Kunreuther et al. (2001:117) conclude: ‘If perception is to accurately reflect the probability of relevant risks, the public needs substantial context information to allow appropriate evaluation of this risk’. Kunreuther’s contextual scenarios provided individuals with enough context to draw on their own experience and developed risk perception through insurance mortgages in order to judge low probability adverse events in chemical factories. In this article it is described how more extensive policy scenarios and context scenarios were used to reframe the river dike safety issue, thus giving substantial context and allowing for two-way communication.

2. The Dutch dikes Floods and inundation threaten large part of the world’s population as most people are living in coastal plains and along rivers. The Netherlands is a densely populated river delta area and the land lies just above or just beneath sea level. The country has always been threatened by flooding both from the sea and from the rivers that flow through the country, of which Rhine, Meuse, and Scheldt are the biggest ones. Since the Stormramp (storm disaster) of 1953, when the North Sea flooded large parts of the provinces Zeeland and Zuid-Holland, much attention has been paid to the defence against the sea. All big sea arms in Zeeland, except for the Westerscheldt, were closed off and storm surge barriers were constructed along the main rivers. Dutch civil engineers became renowned for their know how and technical capabilities all over the world. During that 40 years period relatively little money was invested in maintenance and strengthening of the Dutch river dikes. When during the 1970s and 1980s the focus of attention slowly shifted towards strengthening the river dikes, the same Rijkswaterstaat and waterboard engineers that had been cheered at for their robust defences against the sea, met heavy resistance when they tried to build the same type of large defences along the rivers. The majority of the population did not accept their energetic approach, which included the removal of complete historical villages and destruction of nature preservation areas in order to build their massive dikes. Fierce debates started on the preservation of the so-called LNC values (Landscape, Nature and Cultural values) of the riverland and on the costs of ‘smart’ dikes that would preserve these values. This debate finally came to a halt in 1993 with the advice of the so-called Committee Boertien and after extensive studies (Walker et al., 1993). Boertien showed that little extra investment would save most of the LNC values, thus reaching a typical Dutch compromise. However, the delay of the required improvements caused by these discussions had led to a poor condition of the riverdike system in the central river zone and raised the risk of dike breach. This became apparent during high waters in 1993 and 1995. Eventually, during the 1995 long-drawn high water period, the dikes became soaked and instable. Because of the threat of acute dike-breech several low-lying polders had to be evacuated. Nearly 250 000 people had to leave their homes (CUR/TAW, 1995; Rosenthal et al., 1998). In the end the dikes did not collapse but this event led to the immediate start of a major dike-strengthening programme. Historically the Dutch river dikes protected the low areas and polders from flooding. But by constructing dikes the river was prevented from changing course and forced to deposit its sediments in the area between the dikes. Consequently this dike bound

Thinking the unthinkable

749 sediment deposition

dike river-foreland winterbed

riverbed ditch

polder subsidence

Fig. 1.

Typical Dutch river system.

floodplain or foreland was filled with sand and clay, the riverbed-level rose and in the end came higher than the surrounding land (see Fig. 1). At the same time heavy drainage and intensive farming led to subsidence of the peat and clay substratum of the neighbouring polders. Nowadays the dikes can be more than 8 metres high. In case of dike breech and flooding inundation depths in the upper part of the Rhine delta will be between 4 and 6 metres (Walker et al., 1993:33), up to 8 metres in the lowest parts of some of the polders in the downstream area. As there is no natural outflow of surplus water from these low spots it might take months to drain them again. Generally speaking dike breech of river dikes and inundation of polders in a rich and well-developed country like the Netherlands will not lead to many casualties as high waters can be predicted and people and cattle will be evacuated in time. Nevertheless Dutch society does not seem to be a very disaster resilient community (Mileti, 1999; McEntire, 2001). The economical damage, loss of production capacity and production goods is what counts here and amounts to billions of Euros (Walker et al., 1993:42; Rosenthal, 1998:144; Kuijper, 2000). Next people will suffer from the disruption of their well-organized social life and some will suffer from emotional shock, as many Dutch believe they have mastered nature (CUR/TAW, 1995).

3. Traditional risk safety management Rijkswaterstaat traditionally works along probabilistic lines, as the latter is believed to prevent over-dimensioning and needless expenses. An example of this probabilistic approach is the fault tree of the Dutch flood defence system as depicted by Hekal (2000) given in Fig. 2.

Fig. 2.

Fault tree due to floods.

750

Enserink

1:1250 1:2000 1:4000 1:10000 High grounds

Fig. 3.

Safety levels in the Netherlands.

Within this rational framework different safety levels are assigned to various regions in the Netherlands, which are determined in the ‘Wet op de waterkering’, the Law on flood prevention of 1996. The dike rings in the central river area have a safety level of 1/1250, while in the western part of the country these levels rise through 1/2000 to 1/10.000 for the dike ring ‘Central Holland’ where the large cities are located (see Fig. 3). In practice this implies that the dikes in the central river area are dimensioned in such a way that water levels exceeding this so-called design level will occur only once in 1250 years, at least as far as our linear regression analysis shows, which are based on measurements over a 100 year period. The higher safety levels in the western parts of the country coincide with the economical interests and high population density, as well as with the fact that inundation by salt seawater will result in higher damage levels. In the Dutch central river area 665 km of dikes protect the low lands and polder areas from flooding. In determining the height of these river dikes probabilities are the dominant frame and the so-called design discharge (‘maatgevende afvoer’) is the determining statistical factor. This design discharge is the discharge, which according to statistics occurs once every 1250 years, which is the accepted flooding risk for the Dutch central river area. It is the basis for the design water levels (‘maatgevende hoogwaterstanden’) on which the height of the dikes is based (Walker et al., 1993). The current discharge of 15 000 m3/s was derived from linear regression of a relatively short time series of measurements between 1901 and 1990 in Lobith where the Rhine enters the country (Silva et al., 2000). The two recent highs of 1993 and 1995, although statistically occurring with a possibility of once every 80 years, showed the weakness of regression analysis as predictive method. New regression analysis after these events showed that design discharges should be expected to be much higher and actual safety levels in the ventral river area were at 1/850 rather than 1/1250. It is generally expected that in 2002 the design water levels will be upgraded, as the design discharge will rise from 15 000 m3/s to 16 000 m3/s. Expectations about sea-level rise and climate change are expected to lead in the long-term to design discharges of 18 000 m3/s in 2050 and 20 000 m3/s in 2100.

Thinking the unthinkable

751

4. Thinking about new safety concepts The high waters of 1993 and 1995 forced the national government to immediate action. A so-called ‘Deltaplan Grote Rivieren’ was drawn up and enacted. This plan concerned the immediate execution of dike improvement programs in order to reach a safety level of 1/1250 in the year 2001. Execution of this programme went on schedule (CIW, 2000: 31,32). Immediately after this decision, in 1996, discussions started on the durability of this Deltaplan and the river dike concept and the formal long-term policy objective became ‘Ruimte voor de Rivier’. This ‘Room for the River’ programme had the explicit objective to maintain and, if possible, to expand the capacity of the riverbed as well as the prevention of damage in case of high water levels (Silva et al., 2000). Moreover in 1998 the 12th Conference of Rhine Ministers adopted the ‘Action Plan on Flood Defense’ earmarking up to 12 billion Euros to durable flood prevention measures to be implemented in the next 20 years (Worm and Villeneuve, 1999). The discussions on durable river management were held against the background of the expectation that due to climate change the future might bring more frequent and (much) higher peaks in river discharges. Room for the River aimed at creating more storage capacity through technical and managerial measures in the area between the dikes, like deepening the riverbed, digging by-passes, lowering the forelands, replacing dikes for widening the floodplain and creating retention capacity (Silva et al., 2000). Room for the River is now being executed and has a firm social base. The implementation of this policy programme is planned to be ready by 2015. The discussion on durable river management started off a discussion within the research unit of Rijkswaterstaat, DWW on the durability of the traditional (dike) safety concept in the very long term. They argued that if confronted with an expected sea-level rise of 1 metre, changing rain patterns leading to discharge peaks and continued subsidence of polder-grounds by 50 centimetres, all in less than 100 years from now, dikes might no longer guarantee safety. Moreover, higher dikes mean bigger effects when things do go wrong (see: Fig. 4). The growing inundation depths in combination with bigger losses through the accumulation of goods (caused by economic growth in combination with a (false) sense of security) lead to higher risks. Continued river dike improvement might then not be a robust and durable concept. A programme was started in search of a new safety concept that would not depend on dikes alone. A paradigm shift might be needed. Starting point should be that water is the determining factor or guiding principle (Smit, 1989; Elzen et al., 1990). In their Project Plan DWW (2000) writes: ‘New possibilities should be explored as to how water can find it’s way more naturally; even outside existing dikes. Consequently how people would organize their society at those places should be explored where this is possible: at high places, in the zone between wet and dry areas and even on the water. This new safety

Fig. 4. Higher Dikes, bigger risks.

752

Enserink

concept is robust because it really guarantees safety. Living and working are no longer threatened by water at higher places.’ [translation by author] The DWW project was named Rivierenland what translates into River and Land, expressing their intimate dependence and refers to the change of quality that might be required. In their Rivierenland Courant – a leaflet on the results of the project, issued March 2001, they explain that ‘the concept leads to so many space for the water that even the highest discharges will lead only to a small increase of the general water level in the river basin area. Part of the land will be permanently drowned, but at other places dry spots can be found depending on the height.’

5. Scenarios By anticipating on possible changes one can deal with the dynamics in the problem setting to get robust or flexible solutions (Riet, 2002). When exploring long-term changes historical analyses as well as scenarios can be good instruments. Historical analysis of the Dutch river system showed that in unrestricted conditions the meandering rivers changed their flow every 325 years in average, filling up the lower areas with their sediments until the course of the river changed again (DWW, 2001). This knowledge was used as a basis for assessing the geomorphologic and hydrologic situation after disbanding the dikes, which in turn served as a basis for the design of an ‘image’ of the central river area after implementation of this radical new safety regime. This ‘image’ can be considered to be a policy scenario as it describes a normatively wanted future situation. Scenarios can be defined as a rich and detailed portrait of a plausible future world or as a future state of a system (Beroggi, 1997:16). A scenario is not a prediction or a specific forecast per se; rather, it is a plausible description of what might occur. Plausibility is what distinguishes scenarios from mere fantasy (Enserink, 2000). Although several classifications of scenarios do exist, in this case it is most helpful to differentiate between policy scenarios, which relate to the future state of the system and context or environmental scenarios relating to the impact of the changing environment on the system. This distinction is depicted in Fig. 5 where the conceptual model of systems analysis is shown. The goal of generating scenarios is to understand the mix of strategic decisions that are of maximum benefit in the face of various uncertainties and challenges posed by the external environment. Awareness raising and discussion on problems and possible solutions can be another goal of scenario making and was indeed formulated as one of the objectives of the RivierenLand research project. Several scenario and ‘imaging’ studies were made. The aforementioned distinction between strategic and context scenarios is relevant

Fig. 5.

Framework for delineation of system and environment.

Thinking the unthinkable

753

for discussing the different kinds of scenarios or ‘images’ of the future that were designed for RivierenLand and their role in developing and communicating this new safety concept.

6. Images of the RivierenLand safety concept The essential characteristic of RivierenLand is the free flow of river water and the adaptation of society to this situation. Several partial studies were made to create an impression of the situation in the central river area after 200 or 300 years with respect to landscape, hydrology, geomorphology and human occupation. These studies resulted in an ‘image’ or scenario in which in the higher middle and eastern parts of the Netherlands meandering rivers flow through natural areas, which to the west develop into a delta area with lakes at the west-side, which once were polders. Once a year almost the entire area will be flooded, but the fluctuation of the water level will be restricted and inundation depths in most parts rather limited. The inhabitants will have adapted to living on the water; they live in floating or elevated houses or houses built on stilts and mounts (DWW, 2001:18,19). The image is split up in four different (geomorphologically based) types of land use and urbanization: (1)

The Waarden, relatively low areas containing main transport axes, natural areas, wetland crops and extensively populated mounts. In the eastern, upstream part occasional flooding is possible but inundation depths will be limited. To the west the riverbanks will be shallower and large low-lying wet basins will dominate the landscape. (2) The Old Riverbed, high and dry grounds, winding ribbons with high natural, cultural and historical values and concentrations of occupation. (3) The Randstad Lake area, the low lying western polders will be permanently inundated as they are beneath sea level. In this lagoon natural values will prevail while recreational activities, floating villages and waterfront living will develop, (4) The New Biesbosch, a large nature reserve with tidal influence. These four areas together form the normative ‘image’ or policy scenario of RivierenLand. It depicts as concrete as possible how a radically different but robust and durable safety concept would affect the hydrological and natural circumstances and consequently the organization of society in the area. The ‘image’ was created to be an intriguing and tempting metaphor of what the future could look like.

7. Context scenarios Another scenario study was made which concentrated at the assessment of robust implementation strategies – strategies that are compatible with multiple futures (Meijer and Ruigh-Van der Ploeg, 2001). For this exercise a more systematic design process was used. Although numerous methods have been developed to create scenarios, many methods recognize the need to understand the system under study and to identify the trends, issues and events that are critical to the system. In this study the Schwartz (1991: 226–34) approach, on its turn based on the RAND methodology for systematic scenario development was followed. An important element in this approach, apart from the identification of the focal issue or decision, is the identification of the key forces and trends in the

754

Fig. 6.

Enserink

Scenario–logic context scenarios RivierenLand.

environment. This determination is followed by a ranking exercise in which the ‘driving forces’ are ordered by importance and uncertainty. As these context scenarios are used to assess the robustness of policies they should be based on those factors that are uncertain and hard to influence by the policy makers themselves. A one-day workshop was organized in which a core group of the project staff participated. The participants designed a rudimentary policy plan for a possible transition trajectory as well as a scenario–logic. The axes of this scenario–logic are depicted in Fig. 6. After reaching consensus on the scenario–logic four possible futures were detailed in which the transition from dikes as dominant safety concept to ‘no-dikes’ would have to come about:

• • • •

‘Country of isles’: men abides, space saving and individualistic society, where people live in small towns on natural heights in balance with theire natural environment. ‘City land’; nature abides, space saving, individualistic society, where the wealthy people live in high density luxurious urban centers, well defended against high waters and seemingly ignorant of their natural environment. Less endowed people live at places with low quality and high risk of flooding. ‘Villapark’: nature abides, space consuming, individualistic society, an English garden landscape where wealthy people live in well protected secluded estates, well protected against high waters. The former population concentrations have changed into deprived living and industrial production areas for less-endowed people. ‘Parkland’; nature abides, space consuming, collectivist society, where safety and welfare are a collective responsibility and scarce resources like space and clean air are evenly distributed among all inhabitants. It is a well-organized and orchestrated village society.

After detailing these four context scenarios the transition plan (Rotmans et al., 2001) was assessed upon its robustness in these four very different plausible future environments. The evaluation made clear that in all four scenarios the initiative for starting up a

Thinking the unthinkable

755

transition trajectory towards a new safety and river management regime should come from the national government (Enserink et al., 2000). This is the only forum that can initiate and might be willing to fund research with these kind of long-term horizons and which can stimulate debate on the outcomes of these projects. In time the initiative in the transition process should shift to other, private and semi-governmental parties. Surprisingly the ex ante evaluation of a possible transition showed that chances for success were better in more liberal and individualistic societies in which government recedes from collective responsibilities and regulation is restricted. When central collective and shared responsibility for river dike maintenance and safety issues are lacking, regions with insufficient social and financial base will have weak defenses against high waters, which makes a transition more likely (Enserink et al., 2000). Another revelation was that currently popular romantic images of ‘living near water’ and ‘living on water’ and the expected gain of natural areas and nature preservations prove to be ineffective for gaining support. In most scenarios the future citizens focus on the attractive densely populated urban spheres rather than on nature preservations.

8. Round up In the paragraph on risk communication and frame reflection it was argued that neither the traditional probabilistic approach, nor the deterministic approaches of risk were suited for communication and awareness raising on safety issues, nor for engaging in debates on river dike safety or river basin management. A different discourse was needed to make sure that all receivers of the risk message do understand its meaning and to persuade them to change their attitudes. The ‘image’ or policy scenario depicted by the project group showed the prospect of a new society adapted to a new river management regime leading to an inherently robust safe situation. In Summer 2001 this ‘image’ has been presented to a public concerned with water management for the first time. The conceptual idea was well received although some sceptical comments could be heard too (Kamps, 2001). The policy scenario will now be presented to the general public through dedicated newspapers and conferences in the months and years to come and research will continue. The scenarios in the RivierenLand case presented a new way of framing a safety issue. People present at the workshop had to abide their traditional way of thinking about safety and risk and were confronted with an ‘image’ of a possible future in which sustainable security was reached through adaptation of society to a new situation. The rich and detailed stories of plausible future worlds led to intensive discussions on handling the risks of flooding differently and on the impact on society and its adaptation to these new circumstances. The abstract concepts of probability and risk could be disposed off and attention shifted to assessing the impact on society of sustainable, durable safety concepts. The context scenarios helped the project group to assess the strengths and weaknesses of their implementation plan. Can this controversial plan be made public; will there be much upheaval and resistance or will people go along with the concept and debate on the details? The context scenarios allowed discussion on these issues and gave insight in the hurdles that will have to be overcome. Strategic insights were won on how to start discussion on the reliability and long-term durability of the traditional riverdike safety concept.

756

Enserink

Clearly these scenarios reframed the river dike safety debate and proved to work well for risk communication. In the workshops where they were presented they appealed to all participants and led to discussions on the concept of river dike safety rather than on statistics. The scenarios were persuasive and led to two-sided communication.

References Beroggi, Giampierro E. G. (1997) Decision Modeling in Policy Management, An Introduction to the Analytic Concepts, SEPA: Delft. Bras-Klapwijk, R. M. (1999) Adjusting Life Cycle Assessment Methodology for Use in Public Policy Discourse. PhD thesis Delft University of Technology. Camerer, C. and Kunreuther, H. (1989) Decision processes for low probability events: policy implications, Journal of Policy Analysis and Management, 8, 565–92. CIW (2000) Water in Beeld, Voortgangsrapportage over het waterbeheer in Nederland, Commissie Integraal Waterbeheer (Committee on Integral Watermanagement). CUR/TAW (1995) Syllabus Hoogwater 1995. Voordrachten Symposium Hoogwater. Civieltechnisch Centrum Uitvoering Research en Regelgeveing, Technische Adviescommissie voor de Waterkeringen. DWW (2000) Rivierenland. Verkennende studie gericht op de verre toekomst. Plan van Aanpak. Rijkswaterstaat Dienst Weg- en Waterbouwkunde. DWW (2001) Dijken voor de Eeuwigheid? Een troef achter de hand. Rapportage Project RivierenLand, Ministerie van Verkeer en Waterstaat, Directoraat-Generaal Rijkswaterstaat, Dienst Weg- en Waterbouwkunde, W-DWW-2001-017. Dijk, G. van and Smit, W. A. (1976) Kleine kansen – grote gevolgen of gevolgen van een groot ongeluk met een kerncentrale. THT Boerderijcahier 7601. Eeten, M. van, (1999) Dialogues of the Deaf, Defining New Agendas for Environmental Deadlocks. PhD thesis, Delft: Eburon. Elzen, B., Enserink, B. and Smit, W. A. (1990) Weapon innovation: networks and guiding principles, Science and Public Policy 17(3), 171–93. Enserink, B. (2001) Integral assessment – putting safety on the agenda for mitigation and preparedness, Safety Science 39, 93–105. Enserink, B. (2000) Buiding scenarios for the university, International Transactions in Operational Research 7(6), 569–84. Enserink, B., Bemmelen, I van and Kuiper, O. (2000) RivierenLand, Ex-ante evaluatie van het Transitieproces – een scenariostudie, Report to DWW, DG RWS, Min VenW, TBM, Delft: TU. Fischer, F. and Forester, J. (eds) (1993) The Argumentative Turn in Policy Analysis and Planning, Durham, NC and London: Duke University Press. Hekal, N. (2000) An Inventory of Measures and Solutions Coping with Floods around the World. MSc thesis, Delft: IHE. Hoek, A. W. van der (1995) Elfstedentocht en Deltaplan Grote Rivieren in: CUR/TAW, Syllabus Hoogwater 1995. Huff, D. (1953)(1993) How to Lie with Statistics, New York : Norton. Kamps, D. (2001) Evaluatie werkconferentie ‘Dijken voor de eeuwigheid’, Report to DWW, TBM, (internal report), Delft: TU. Keeney, R. L. (1980) Equity and public risk, Operations Research 28(3), 527–34. Kuijper, O. (2000) Rivierenland, Risico’s van het huidige waterbeleid, Report to DWW, DG RWS, Min VenW. TBM, (internal report), Delft: TU. Kunreuther, H., Novemsky, N. and Kahneman, D. (2001) Making low probabilities useful, The Journal of Risk and Uncertainty 23(2), 103–20. Liu, P.-S. and Chan, N. W. (2001) The Malaysian Flood Hazard Management Programme, paper on CD-Rom. Oslo, Norway: The International Emergency Management Society.

Thinking the unthinkable

757

Lowrance, W. W. (1976) Of Acceptable Risk, Science and the Determination of Safety, Los Altos, California: William Kaufman, Inc. McEntire, D. E. (2001) The Future of Emergency Management: The Search for a Paradigm and Policy Guide, paper on CD-Rom. Oslo, Norway: The International Emergency Management Society. Meijer, M. H., and Ploeg, M.R.-van der (2001) Critical factors in strategic planning for infrastructure design, in Weijnen, M. P. C, Bauer, J. M., Chamoux, J.-P., Ehrenfeld, J. R. and Jones, D. N. (eds) Walking a Thin Line in Infrastructures, Balancing Short Term Goals and Long Term Nature, Proceedings of the Third Annual Symposium Delft Interfaculty Research Center Design and Management of Infrastructures, pp.153–78. DUP Science: Delft. Mileti, D. S. (1999) Disasters by Design: A Reassessment of Natural Hazards in the United States, Washington DC: Joseph Henry Press. Poortvliet, A. van (1999) Risks, Disasters, and Management; a Comparative Study of Three Passenger Transport Systems, Delft: TU. Renn, O. (1998) The role of risk communication and public dialogue for improving risk management, Risk Decision and Policy 3(1), 5–30. Rein, M. and Schön, D. (1993) Reframing policy discourse in F. Fischer and J. Forester (eds) The Argumentative Turn in Policy Analysis and Planning, Durham, NC and London: Duke University Press, pp. 145–66. Riet, O. A. W. T. van der (2002) What’s the Problem in Infrastructure Planning, Delft: TU. Rogers, M. D. (2001) Scientific and technological uncertainty, the precautionary principle, scenarios and risk management, Journal of Risk Research 4(1), 1–15. Rosenthal, U., Duin, M. van, Bezuyen, M., Vreeze-Verhoef, M.-L. de and Hart, P. ‘t (1998) Flood management in the netherlands, in U. Rosenthal and P. ‘t Hart (eds), Flood Response and Crisis Management in Western Europe. A Comparative Analysis, Berlin: Springer. Rotmans, J., Kemp, R. and va Asselt, M. (2001) More evolution than revolution. Transition management in public policy, Foresight 3(1), 15–31. Schwartz, P. (1991) The Art of the Long View: Planning for the Future in an Uncertain World, New York: Doubleday. Silva, W., Klijn, F. and Dijkman, J. (2000) Ruimte voor Rijntakken. Wat het onderzoek ons heeft geleerd. RIZA-nota 2000.026; WL-rapport R3294. Slovic, P., Fischhoff, B. and Liechtensteinn, S. (1978) Accident probabilities and seat belt usage: a psychological perspective, Accident Analysis and Prevention 10, 281–85. Smit, W. (!989) Defence technology assessment and the control of emerging technologies, in M. ter Borg, and W. A. Smit (eds) Non-provocative Defence as a Principle of Arms Reduction and its Implications for Assessing Defence Technologies, pp.61–76. Amsterdam: Free University Press. TeBrake, W. H. (1985) Medieval frontier. Culture and ecology, in M. V. Melosi (ed.) Number Seven: Environmental History Series, College Station: Texas A&M University Press. Volkskrant, de. (2002) Verdachte firma Panaviation had geen rol in vliegtuigcrash Genua’, de Volkskrant digitaal archief, 31 januari. Walker, W., Abrahamse, A., Bolten, J., den Braber, M., Garber, S., Kahan, J., Kok, M. and van der Riet, O. (1993) Investigating Basic Principles of River Dike Improvement, Safety Analysis, Cost Estimation, and Impact Assessment, Santa Monica, CA: RAND European-American Center for Policy Analysis. Worm, J. and de Villeneuve, C. H. V. (1999) Flood/discharge management; Rhine Action Plan on flood defence, in Mostert (ed.) River Basin Management, Proceedings of the International Workshop, The Hague, 27–29 October 1999, pp. 257–61. Zhou, H.-M. (1995) Towards an Operational Risk Assessment in Flood Alleviation – theory, operationalization and application. PhD thesis, Delft: Delft University Press.