Introduction

The challenges posed by climate-related extreme events to river basins are manifold, especially since water resource issues interact with a wide range of environmental and socioeconomic sectors including health, public safety, agriculture, biodiversity, industry, navigation, and tourism. According to the Intergovernmental Panel on Climate Change (IPCC 2007), an increase in the surface temperature of water and changes in the hydrological cycle could result in changing rainfall patterns. Some areas may experience intense rainfall resulting in heavy floods, while other areas may witness less rainfall, and also frequent droughts.

On a global scale, the number of disasters caused by weather-related phenomena such as storms, floods, and droughts has more than doubled over the past decade, from 175 in 1996 to 391 in 2005 (IFRC 2008). The same trend (Fig. 1) is being observed in the case-studies under investigation, which is Rivierenland in the Netherlands, the Alentejo region (including the Alqueva reservoir) of the lower Guadiana River Basin in Portugal, the Hungarian part of the Upper Tisza, and the Ukrainian part of the Upper Tisza (also called the Zacarpathian Tisza).

Fig. 1
figure 1

Number of reported flood and drought disasters in the past decades in Rivierenland, Alentejo, Hungarian part of Upper Tisza, and Ukrainian part of Upper Tisza. Based on data from EM-DAT: The OFDA/CRED International Disaster Database 2008

Full size image

Comparing data for the past decade (1997–2006) with data for the previous decade (1987–1996), the number of reported flood and drought disastersFootnote 1 has increased (see Fig. 1). Over the same period, the average total number of people reported affectedFootnote 2 per decade rose by 1200%, from approximately 38 thousand to 480 thousand. Meanwhile, the total cost of reported damageFootnote 3 doubled, from US$ 1.2 billion to US$ 2.4 billion (2006 prices), although this could partially be explained by a higher density or vulnerability of infrastructure and areas being used by humans. Moreover, more detailed documentation of smaller disasters partially explains the earlier mentioned increases. It shows nevertheless that flood and drought problems become more visible and cannot be ignored. Especially since climate change models project more frequent and intense summer droughts across many parts of Europe, particularly in the southern part (Goodess et al. 2007; Fowler et al. 2007). This may be further exacerbated because of an increasing demand for water as a result of elevated temperatures.

The increasing frequency and intensity of floods and droughts is also being confirmed by the latest report of the Intergovernmental Panel on Climate Change (IPCC, Fourth Assessment Report 2007), which argues that “there are multiple lines of evidence that climate change is happening now, and the impacts are being seen now”. Given the expected increase of climate-related extreme events, water management capabilities in the case-studies, and on a global scale, will be tested to their limits by the effects of climate change. This requires innovative and adaptive ways of managing water, which can be referred to as “Adaptive and Integrated Water Management” (AIWM). This paper addresses the role of AIWM in coping with the impacts of climate change on floods and droughts in four case-studies in three European river basins. The explorative character of this paper intends to identify general patterns in the characteristics of AIWM and assumes that regimes with a higher level of AIWM consider and implement more advanced and a more diverse set of structural and nonstructural measures.

The selected case-studies are all confronted with floods and droughts. Even the most drought-prone case-study, the Alentejo Region in Portugal, was confronted with serious flooding events (notably in 1997 and 2006). Next to this similarity of being confronted with floods and droughts, the management regimes in all four case-studies are subject to a similar institutional setting on the European level [e.g. Water Framework Directive (Directive 2000/60/EC), Common Agricultural Policy, etcetera], except for the Ukrainian part of the Tisza. However, the Ukraine shows strong incentives to enter the EU community and thus the EU acquis communautaire is used as key reference for the development of its water management principles. It was nevertheless decided to select two case-studies in the Tisza Basin, because the international Tisza river basin is situated at the current borderline of the EU (e.g. Hungary and Ukraine) and thus the national and regional water management might follow different frameworks for the one river.

Despite the earlier mentioned similarities between the case-studies, it is expectedFootnote 4 that there are (substantial) differences between the water management regimes of these case-studies, more specifically in their level of AIWM.Footnote 5 Therefore, this paper intends to identify differences and similarities in AIWM, to detect general patterns, and to compare measures being taken to cope with floods and droughts.

The characteristics (see next paragraphs) of AIWM are to be regarded as working hypotheses, since the change toward more adaptive management regimes is yet slow and empirical data and practical experience thus limited—in particular regarding the interdependence of elements of the management regime. The strong interdependence of the factors stabilizing current management regimes is also one possible reason for this lack of innovation. One cannot, for example, move easily from top–down to participatory management practices without changing the whole approach to information and risk management. Hence, research is urgently needed to better understand the interdependence of key elements of water management regimes and the dynamics of transition processes in order to be able to compare and evaluate alternative management regimes and to implement and support transition processes if required.

This article also addresses the question whether a higher level of AIWM is showing a different response in coping with floods and droughts than case-studies with a lower level of AIWM. This will be done by looking at their adaptation strategies and their planned or implemented physical interventions in the river basin.

This paper is focusing on conditions and processes in the management of a specific part of the river basin at sub national level (e.g. water boards), but being embedded in a wider context (e.g. institutional setting at different levels). The sub national level is conceived as the level where all elements of a water management regime are at play. At the same time, this level is influencing, or is being influenced by, higher and lower levels. This central position also allows for assessing the outcomes of a water management regime at the operational level, since the management on the sub-basin level (e.g. water board or regional water authority) is influenced by international or national regulation, while implementing at the operational/local level.

A calibrated approach (standardized questionnaires and interviews, expert judgment, and reinterpretation of outcomes by means of relevant literature) was used to compare the state of affairs in water management in the selected case-studies.

Adaptive and integrated water management

Given the expected increase of climate-related extreme events, water governance capabilities in the case-studies, and on a global scale, will be tested to their limits by the effects of climate change. This requires innovative and adaptive ways of managing water, which can be referred to as “Adaptive and Integrated Water Management” (AIWM).

To deal with existing and new complexities, water resources management must be able to respond to changes in the natural and social environment and to anticipate associated uncertainties (Folke et al. 2005; Pahl-Wost et al. 2007). Adaptation to climate change and management of related risks should therefore be built into water resources management plans and programmes. Adaptive and integrated management is considered to be an appropriate approach for doing so. Adaptive and integrated management can be defined as a structured process for improving systemic management policies and practices by learning from the outcomes of implemented management strategies (Pahl-Wostl et al. 2007a). By re-evaluating goals, objectives and means how to achieve them as new information and insights become available, adaptive management is more responsive to changing conditions of and demands on ecosystems when compared to traditional approaches to water resource management.

AIWM requires different capabilities than traditional forms of water management, particularly when it comes to creating forms of collaboration between water managers and stakeholders, the relation between science and policy, the importance of participatory learning processes, dealing with uncertainty, and assessing a wide variety of possible measures and future scenario’s. It requires many instances of social learning to implement and sustain innovative management approaches (Pahl-Wostl et al. 2007). Folke et al. (2005) already pointed out that social learning is needed to build up experience for coping with uncertainty and change. They emphasize that “knowledge generation in itself is not sufficient for building adaptive capacity in social-ecological systems to meet the challenge of navigating nature’s dynamics” and conclude that “learning how to sustain social-ecological systems in a world of continuous change needs an institutional and social context within which to develop and act”. Knowledge and the ability to act upon new insights are continuously enacted in social processes. The social network of stakeholders is an invaluable asset for dealing with change.

These considerations highlight that one important element of adaptive water management is the governance structure. Adaptive governance can be understood as the synthesis of collaborative management and adaptive management (Wailand 2006). It can be defined as the totality of interactions, by private and public actors, to achieve adaptation and to enhance the capacity of processes, institutional arrangements and actors to adapt to future environmental changes (Huitema et al. 2009). Adaptive governance depends on adaptive institutions (Pahl-Wostl 2002) that are able to cope with complexity and uncertainty and to face new challenges such as climate change.

However, in technology-dominated water management practice, design and structure of governance regimes has not played a prominent role. Adaptive and integrated water resources management implies a real paradigm shift in water management from what can be described as a prediction and control to a management as learning approach. Such change aims at increasing the adaptive capacity of river basins at different scales and implies a change in the whole water management regime (Pahl-Wostl 2007; Adger et al. 2005). Some structural requirements for a water management regime to be adaptive are summarized in Table 1. Two different regimes characterized by two different management paradigms—management as control versus management as learning—are contrasted as the extreme, opposing ends of six axes.

Table 1 Different regimes and their characteristics (From: Pahl-Wostl et al. 2007a)
Full size table

Interdependence of the regime characteristics and responsiveness to floods and droughts

The key objective of our research is to see whether there is a link between regime characteristics and responsiveness to floods and droughts (as an output of the regime at play). For this purpose, we needed to develop two different and independent analytical frameworks:

  • A framework for assessing the characteristics of a water management regime

  • A framework for assessing responsiveness to floods and droughts

In the next paragraphs, we will provide the conceptual background for both frameworks and related development of variables and indicators.

A framework for assessing the characteristics of a water management regime

For this purpose, we first needed to develop a normative framework of how an adaptive and integrated management regime looks like, in order to develop indicators for assessing the regime characteristics. For this normative framework, we have used the working hypotheses on the characteristics of AIWM being presented in Table 1. These working hypotheses have been further developed into a methodology for evaluating the level of Adaptive and Integrated Water Management, which resulted in an analytical framework for assessing regime characteristics, consisting of nine different dimensions of variables:

  1. 1.

    Agency

  2. 2.

    Awareness Raising & Education

  3. 3.

    Type of governance

  4. 4.

    Cooperation structures

  5. 5.

    Policy development & implementation

  6. 6.

    Information management & sharing

  7. 7.

    Finances and cost recovery

  8. 8.

    Risk management

  9. 9.

    Effectiveness of (international) regulation

As a useful starting point for operationalization, we considered the River Basin Assessment framework developed by Raadgever et al. (2008). Raadgever et al. developed a framework including four regime elements (4–7). Based on relevant literature (see endnotes in Table 2), we have added the dimensions agency, governance, awareness raising and education, risk management, and effectiveness of (international) regulation to the four regime elements that Raadgever et al. (idem) used to describe a management regime in their article. Furthermore, we have developed variables and indicators for the added dimensions and adjusted some of the variables and indicators in the framework of Raadgever et al. (2008). By doing so, we have further developed the methodology for assessing and comparing governance regimes. The operationalization resulted in 33 variables; comprising 62 indicators (see Table 2, including related literature). In addition to the framework of Raadgever et al. (2008), we have provided a more detailed explanation and related literature for each indicator in order to prevent ambiguity in the interpretation of indicators.

Table 2 Overview of variables and indicators for AIWM
Full size table

The current large number of variables is inherent to the explorative character of this research, but also to the nature of complex (governance) systems. Hence, one cannot simply omit variables without eroding the comprehensiveness of our analysis of complex governance systems. In other words, with our current number of variables, we consciously intended to reduce the risk of glossing over potential key variables, especially since the empirical base for AIWM is yet rather weak, and we cannot a priori justify which variables should be in or out.

A framework for assessing responsiveness to floods and droughts

A second analytical framework, independent from the first one, has been developed for assessing the responsiveness to floods and droughts in river basins, being defined as physical adaptation measures within the context of flood protection and drought/low flow protection. The type of measures listed in this framework (see Tables 4 and 5) are drawn from two different reports. The first one is a report of the European Environment Agency called ‘Climate change and water adaptation issues’ (EEA Technical Report No 2/2007), the second report is called the ‘Bouwstenen Nota’ (RIZA 2007), which is a study describing all potential flood protection measures in Dutch Rivers. The current status of physical interventions in each case-study has been assessed by means of expert judgment (see methodology in the next paragraph).

In principle, the measures themselves might have been implemented very recently or decades ago. Moreover, it should be taken into account that these management regimes may be currently evolving as a response to climate-related extreme events. In other words, even when a new management regime has been established, it may not have achieved its projected outcomes yet. For this reason, the responsiveness of the management systems in the case-studies to extreme events as presented here should be viewed as a snapshot, especially since there normally is a time lag between policy development and implementation of measures. This time lag has been taken into account in this research by asking experts to indicate whether measures are (1) already implemented (=physically present), (2) being planned, or (3) necessary, but not being planned (yet). Nevertheless, the results presented in this research are not able to indicate: (a) whether the implemented measures are working properly or whether their quality and scale is sufficient to deal with the problems they are designed for; (b) if, how and when the planned measures will be really implemented and whether there will be complications during implementation (e.g. delays, insufficient funds, and objection by citizen groups.), and (c) measures which are necessary, but have not been planned yet, could be planned after all in the nearby future. Hence, it is important to monitor the responsiveness of the management regimes in these case-studies for a longer period and on a frequent basis.

Apart from the physical intervention in the water system, a regime’s performance could and should also be measured by its management interventions, such as awareness campaigns, measures for improving information management and exchange, cooperation structures, and insurance mechanisms. The assessment of management interventions is beyond the scope of this paper, but has been taken into account in succeeding research of EU NeWater.

Methodology

A calibrated approach using a standardized questionnaire for the elements of AIWM, and a questionnaire for physical interventions in the river basin (see Tables 4 and 5), expert judgment for both questionnaires, and reinterpretation of outcomes by means of relevant literature was used to compare the water management regimes in the selected case-studies.

By combining in-depth case-studies with more extensive and formal comparative analysis, we can to some extent use the strengths of one method to compensate for limitations inherent in the other. This explorative research has taken the research problem, rather than a favorite methodology, to determine the research approach, and both the quantitative and qualitative aspects have been and can be used in a consonant manner (de Leon 1998). As such it is possible to combine the qualities of the case-oriented approach with the qualities of the variable-oriented approach (Berg-Schlosser et al. 2008). In this research, expert judgment has been used as method for knowledge elicitation on regime elements and internal processes. The list of potential respondents for each case-study was developed in cooperation with the case-study teams of NeWater, with the objective of selecting a group of respondents with enough knowledge on the case-study under consideration and with the ability to answer the whole questionnaire, or at least a major part of it. Moreover, with the objective of including all relevant perspectives and experiences in the case-study, the respondents group was selected as a reflection of the most important stakeholders, policymakers, water practitioners, involved scientists, private sector, and civil society. This resulted in a consultation round involving a minimum of ten experts in each case-study, reflecting perspectives from different stakeholder groups (see Fig. 2). A complete list of respondents has been provided in Huntjens et al. (2007; updated 2009).

Fig. 2
figure 2

Overview of the number of experts (per stakeholder group) consulted in each case-study

Full size image

By using standardized questionnaires, or using these questionnaires for standardized interviews, qualitative data were being collected in such a way that it was possible to compare weighted averages on each separate indicator. The weighted average has been calculated by multiplying each individual score by the weight which respondents assigned to it; the total sum of all respondents in one case-study was then divided by the total assigned weight (by adding up all weights assigned to this specific indicator). Furthermore, the level of inconsistency (standard deviation) for each variable has been calculated, next to ‘Independent Samples T Tests’, in order to test for significant differences between the case-studies.Footnote 6

The reason for developing standardized answering options in the questionnaire is that it supports a formal comparative analysis of the results. Furthermore, the questionnaire allows for assigning weights to each indicator. In this way, it is possible to aggregate multiple indicators, resulting in a score for one variable, or for aggregated variables, resulting in a score for one meta-variable (e.g. dimension of variables).Footnote 7

Not much work is available on comparative analyses of river basins including full range of a water management regime’s complexity (Myint 2005; Wolf 1997). Many studies on IWRM are descriptive and limited to recording success or failure of single cases. The initial comparisons in this research will help develop and test protocols (cf. Breitmeier et al. 1996) that open the way for efforts at broader generalizations about options for institutional designs and procedures with a special emphasis on assessing what does and does not work well with respect to adaptive and integrated water management.

Empirical results show general patterns in the characteristics of adaptive and integrated water management

Based on our formal comparative analyses of the water management regimes in the four case-studies, it is possible to identify general patterns in the dimensions of AIWM. It is important to acknowledge that management regimes with the majority of variables having a weighted average closely to 2 are considered regimes that follow a pattern of configuration more closely related to AIWM. Additionally, we assume that a necessary prerequisite for being adaptive is a high degree of integration (Pahl-Wostl and Sendzimir 2005). Our results show that the case of Rivierenland appears to be the most closely related to AIWM (see Fig. 3). Moreover, the empirical results clearly show that the dimensions of AIWM are interdependent, meaning that when variables show high averages in one dimension, the variables in other dimensions of the respective case-study equally show high averages. On the other hand, when variables in one dimension show low averages, the variables in other dimensions also show low averages. This interdependency is especially evident between variables in information management, cooperation structures, and conflict resolution (see Table 3), but also a number of indicators in other variables such as in risk management and policy development and implementation. Within these dimensions, the following very significant correlations (P ≥ 0.95) stand out:Footnote 8

Fig. 3
figure 3

Level of Adaptive and Integrated Water Management in coping with climate-related extreme events in the four case-studies (0 = non-adaptive and non-integrated, 2 = adaptive and integrated). From: Huntjens et al. (2007)

Full size image
Table 3 Overview of the results (weighted averages) of expert judgement on each indicator in the regime dimensions Type of Governance, Cooperation Structures, and Information Management (From: Huntjens et al. 2007)
Full size table

  • Information management (in particular joint/participative information production, consideration of uncertainties, and broad communication) shows very significant positive correlations (P ≥ 0.95) with cooperation structures (in particular vertical cooperation, vertical conflict resolution, and transboundary cooperation).

  • Risk perceptions (in particular participative risk perceptions and participatory decision-making on what are acceptable risks) show very significant positive correlations (P ≥ 0.95) with information management (all variables) and with cooperation structures (in particular vertical cooperation, vertical conflict resolution, and transboundary cooperation).

  • Consideration of possible measures shows very significant positive correlations (P ≥ 0.95) with bottom-up governance, stakeholder participation, implementation of policies, participative risk perceptions, and adaptive leadership.

Earlier mentioned results reveal many positive correlations between variables from different regime elements. This suggests a strong interdependence of the elements within a water management regime, and as such this interdependence is a stabilizing factor in current management regimes. For example, based on the very significant positive correlations between joint/participative information production, and vertical cooperation, transboundary cooperation, consideration of uncertainties, and broad communication, we can conclude that a lack of joint participative knowledge production is an important obstacle for vertical cooperation and conflict resolution, or vice versa. Cooperation and joint/participative knowledge production is even more important when dealing with uncertainty and change. Several authors (Olsson et al. 2006; Stubbs and Lemon 2001; Loeber et al. 2007) argue that linking different networks and creating opportunities for new interactions are critical factors for learning and nurturing integrated adaptive responses to change. Therefore, mechanisms which facilitate social learning such as coproduction (e.g. joint/participative knowledge production) and multiple stakeholder platforms are therefore being suggested as an important design principle for river basin management at different scales (Pahl-Wostl et al. 2007b). Such an institutional design principle would be an addition to the design principles developed by Ostrom (1990), which are based on long enduring local and single-use common pool resource communities. In contrast, river basins are open resources systems on a larger scale, with a complex constellation of multiple-uses, meaning that additional, or perhaps different, design principles might be relevant.

An important hypothesis in the concept of social learning is that information management and social (cooperation) structures are interlinked (Pahl-Wostl et al. 2007a), which corresponds to the socio-cognitive theory of information systems (Hemmingway 1998). Hemmingway pays attention to the impacts of presented information on learning and action, and the centrality of the selection and organization of information to the nature of organizational forms. Our research confirms that information management and social (cooperation) structures are interlinked in the management regimes under consideration. This interdependency can be described as the socio-cognitive dimension of water management regimes. We define the socio-cognitive dimension as the integrated cognitive and social properties of complex systems and related processes. These (informal or formal) learning environments are an emergent property of the interlinkage between information management and social cooperation structures and are perceived to be crucial for the adaptive governance of socio-ecological systems (see also Folke et al. 2005; Pahl-Wostl 2005).

The earlier mentioned positive correlations provide evidence that there is a strong interdependence of the elements within a water management regime. Such interdependences may constitute stabilization of currently evolving water management regimes toward AIWM regimes. One cannot, for example, move easily from top–down to participatory management practices without changing the whole approach to information management, policy development, risk management, cooperation structures, and styles of leadership. Additional limiting factors for moving toward adaptive and integrated water management regimes are a lack of funding and capacity in current water management regimes. Lack of capacity refers to a lack of adaptive leadership, limited number of water professionals who are familiar with AIWM approaches, and limited public participation and stakeholder participation.

Finding a balance between bottom-up and top–down governance in river basin management

The weighted averages on governance suggest that bottom-up governance plays not such a primary role in AIWM as earlier suggested. For example, the weighted average on governance in Rivierenland (see Table 4) indicates that there is much more top–down governance than could be expected from consensus-based decision-making, particularly since the Netherlands is well known for its consensus-based decision-making (also called the ‘Poldermodel’). The fact that Rivierenland also shows a substantial degree of top–down governance is reflected in the framework for flood management (PKB Room for Rivers) in the Netherlands, which was initiated by the national government (Berenschot 2007). Nevertheless, after initiation by the national government, the framework was being further developed, and this process was much more characterized by bottom-up governance (Berenschot 2007). This bottom-up process was, among others, reflected in the document called “Advice to the parliament as regard the PKB Room for Rivers by nine civil society organizations” (LIRR 2003).

Table 4 Overview of expert judgment on physical interventions as regards drought and low flow protection or mitigation measures in each case-study
Full size table

In July 2006, the Directorate-General Water in the Netherlands published a policy document (Waterkoers 2), which refers to transitions and five paradigm shifts, which it considers to be important for future water management. This report concludes that ‘public and civil society are still not adequately mobilised, despite substantial investments’. In other words, a balance between bottom-up governance and top–down governance has not been found yet, if such a balance is to be found at all.

It should be emphasized that even where bottom-up governance is a useful tool, it may be usefully supplemented by the state. For example, decentralization does not exclude public control, since public and governmental authority can also be located at lower levels—e.g. independent provinces. Carruthers and Stoner (1981) even argued that ‘common property resources require public control if economic efficiency is to result from their development’, and they concluded that public control is needed to avoid overexploitation of resources. On the other hand, Elinor Ostrom (1990) argues that centralized control is based on assumptions concerning the accuracy of information, monitoring capabilities, sanctioning reliability, and zero costs of administration. Without valid and reliable information, a central agency could make several errors, including setting the carrying capacity or the fine too high or too low, sanctioning water users who cooperate, or not sanctioning defectors.

Acknowledging both sides of the coin, Scharpf has pointed out the importance of “bargaining in the shadow of hierarchy” (Scharpf 1997). A variety of empirical studies exemplify this claim. For instance, Bardhan (2002) has found that the state can promote bottom-up, consensual decision-making by playing an active role in the mobilization of people in local processes, it can aid in neutralizing local oligarchs, can provide funds for local initiatives, provide technical and professional services to help local capacity building, guarantee quality standards, invest in larger infrastructure, and coordinate in externalities that span more than one local government. Healey et al. (2003) demonstrate that collaborative decision-making may be promoted by strong, strategically operating (local) governments. In a case-study of a practice in the area of sustainable agriculture, Hendriks and Grin (2007) have found that “Parliament may have been inactive, but it was certainly not insignificant.” It neither intervened, nor did it much to implement the recommendations, but still was important as “it provided the locus around which much discursive activity was centred, and this helped to legitimate the project.”

In summary, for large-scale, complex multiple-use systems, such as river basins, this research suggests that bottom-up governance and decentralization are not a straightforward solution to water management problems. There will probably always be the need for a certain degree of top–down governance (or centralization), where a central authority has the responsibility and resources for issues like facilitation of participatory processes, setting of standards, capacity building, conflict resolution, and cooperation across boundaries. All the case-studies in this research seem to be in a process of finding a balance between bottom-up and top–down governance.

Responsiveness to extreme droughts and low flow problems

Next to comparing the level of AIWM, and identifying general patterns in the characteristics of AIWM, it is important to compare whether there are differences between the management regimes as regards measures being taken to cope with floods and droughts. After all, a higher level of AIWM is expected to result in more advanced adaptation strategies and measures to cope with (the threat of) floods and droughts. This research has labeled this outcome of a water management regime as the responsiveness to flood and drought problems.

The responsiveness of the management regimes to severe drought events in the recent years is very limited in all case-studies (see Table 4). This lack of an adequate response in all case-studies is noteworthy, since droughts have already caused enormous adverse social, economic, and environmental effects in recent years (Della-Marta et al. 2007: 252; EM-DAT 2008), and it is expected that this will become even worse due to climate change (IPCC 2007). Climate change models project more frequent and intense summer droughts across many parts of Europe, particularly in the southern part (Goodess et al. 2007; EEA 2005). This may be further exacerbated because of an increasing demand for water as a result of elevated temperatures (EEA 2005).

This research shows that the management regime confronted with the most severe water scarcity—the Alentejo case-study—is showing a poor response, at least in terms of diversity in measures, when compared to other case-studies. Despite the fact that the Alentejo region has been confronted with drought events for many decades, out of the twelve possible measures in Table 4, only two measures for dealing with droughts have been implemented up until today: (1) increasing reservoir volumes (e.g. Alqueva Reservoir) and (2) crop rotation. The Alqueva Reservoir is an extensive multipurpose project that aims at a social and economic development of the Alentejo region in the south of Portugal, by means of setting up a strategic water reserve and providing a guarantee of water supply for irrigation (110000 ha), population, and industry (EDIA 2008).

Besides crop rotation and the large-scale Alqueva reservoir, only three additional measures have been planned in the Alentejo region (water transfers, securing minimum flows, and more efficient irrigation), while six measures are being mentioned as necessary, but for which no plans have been developed. Experts in the Alentejo region indicate that much faith has been put in the water transfers of the EMFA-project,Footnote 9 a project seen as a solution to all water shortages. However, these water transfers are expected to lead to serious negative environmental impacts, and even to violations of the EU Habitat Directive (Platform for Sustainable Alentejo 2005). In summary, large-scale infrastructure is not seen as a problem in the Alqueva case-study, and as such there is also no sign of a “paradigm shift” as regards the implementation of more small-scale and non-structural measures.

Extensive studies in the Netherlands (Droogtestudie 2005, update 2008) indicate that droughts and heat waves are becoming a serious problem for many sectors, such as water transport, agriculture, and nature. Even though the problem is recognized, the political climate in the Netherlands is not ready for a policy change with regard to extreme droughts and heat waves. This political climate is represented in the Dutch National Safety Reports (Ministry of Domestic Affairs, sub-report on extreme droughts and heat waves, 2006), which concludes that the situation with respect to mitigation and prevention of extreme droughts and heat waves is not as urgent when compared to other natural threats, such as pandemics (e.g. mouth-and-foot disease) and the threat of floods. Even though Dutch society had clear warning signals before, regarding the impacts of extreme droughts, it was not enough to put drought problems high on the public or political agenda. These clear warning signals included for example the fact that agricultural production in the Netherlands is reduced by 5 to 35% because of water shortages (National Safety Reports, Ministry of Domestic Affairs, Netherlands 2006). This means an average economical damage of 180 million Euros/year, and even 1800 million Euros/year in extreme years (idem). Damage to the Dutch water transport sector goes up to 800 million Euros in extreme years (Droogtestudie 2005). During severe droughts in the Netherlands in the summer of 2005, the Dutch Government desperately launched a national ad hoc campaign to urge people to use less drinking water and less electricity. Simply because there was not enough water for public and agricultural water supply, and water temperatures were too high for abstracting cooling water from the rivers, which is necessary for electricity production. Nevertheless, in Rivierenland, the response to current and expected drought problems is poor. Experts in Rivierenland have indicated that seven measures are necessary or effective for dealing with droughts, but none of them have been implemented so far. Only for a few subsystems, there are plans for increasing reservoir volumes, water transfers, securing minimum flows in dry periods, leakage reduction, and restriction of water uses, but none of these plans have been implemented yet.

Based on the conclusions of the Dutch National Safety Reports (2006), but also based on expert judgment in this research, it is justified to conclude that the political climate does not make the time right for change. Kingdon (1995) stresses the importance of timing for initiating policy changes and opening “policy windows”. He argues that significant changes are most likely when three independently operating “streams,” i.e., problems, solutions, and politics, come together at critical times. In other words, even though the problem of extreme droughts in the Netherlands is being recognized to some extent (Droogtestudie 2005, update 2008), there is no window of opportunity to initiate new policy measures, since this would require, in addition to recognition of the problem, that “a solution is available, and the political climate makes the time right for change, and the constraints do not prohibit actions” (Kingdon 1995).

As regard the Upper Tisza, agriculture and forestry have suffered from extensive droughts in successive years. However, the implemented and planned measures for drought and low flow protection are rather limited so far. The extreme drought events have at least initiated the development of the Hungarian Drought Strategy (2004), with support of UNCCD, although an integrated strategy at national level is not yet available (Pers. Comm. Dr. Z. Flachner, Hungarian Academy of Sciences). The present version of the national drought strategy (to be approved yet) states that drought can affect the whole of society and requires a systematic approach to mitigate its consequences. It envisages the establishment of a National Drought Committee and a Drought Fund (EEA 2006).

In the Zacarpathian Tisza, there is no literature and data available on the frequency and intensity of droughts and heat waves. If there would be a problem with severe drought events in this case-study, we can at least conclude that this problem is not being recognized up until today, or that the problem is not that severe in this case-study when compared to other case-studies. Nevertheless, the experts in this case-study indicate that nine (9) measures for drought and low flow protection are necessary, but none of them has been planned yet. As such we can conclude that this case-study is confronted with drought and low flow problems, but that the responsiveness of the management regime is almost zero.

Responsiveness to (the risk of) floods

In contradiction to the poor response to drought and low flow problems, all management regimes, except in the Alentejo region, seem to be more effective in terms of flood protection measures (see Table 5). In the Alentejo region, respondents indicate that flood problem (and flood protection) is not an issue, although it was confronted with serious flooding events (notably in 1997 and 2006).

Table 5 Overview of expert judgment on physical intervention as regards flood protection and mitigation measures in each case-study
Full size table

Rivierenland shows the most advanced policy development and implementation regarding flood protection, although this policy was mainly the result of (the threat of) floods in 1993 and 1995. As such its policy is a response due to extreme events in the past. These extreme events created political momentum for developing new flood protection policy, including climate change (scenarios) and structural and non-structural measures for flood protection. Also in both Tisza case-studies, recent floods have increased political momentum for planning new flood protection measures, although this has started a decade later, mainly due to the floods of 2005.

The revised State Program for Flood Protection in the Ukraine has been planning new flood protection measures in the period 2002–2010, and after re-assessment of what is effective (which allows at the same time for reallocation of resources), there will be a new implementation plan for the period 2010–2015.

The Zacarpathian Tisza seems to be quite effective in dealing with floods, at least in terms of policy development and implementation. However, the revised State Program for Flood Protection does not include any climate change scenarios (Pers. Comm. Dr. A. Iaroshevitch, Ukrainian Center of Environmental and Water Projects). In other words, there is a serious risk that the implemented and planned flood protection measures are not effective enough for dealing with increased frequency and intensity of floods in the coming decades. On the other hand, in Hungary, there are a number of research programmes (Hungarian MoEW 2005; VAHAVA 1 and 2)Footnote 10 which are trying to define the scale of possible impacts of climate change. However, since there is no agreement yet on these climate change scenarios, it has not been included in current Hungarian water policy. Jolánkai (2005) concludes that, based on the analysis of climate change and precipitation scenarios, higher floods than observed so far may occur, needing upgraded flood-control strategies (with the meaning that presently contemplated strategies, such as the VTTFootnote 11 in Hungary, may not be sufficient to cope with floods).

The Dutch water policy (Ministry of Transport, Public Works and Water Management 2000) recognizes that in the coming years increasing water levels in the rivers and the accelerated rise in sea levels will mean that technical measures, such as raising dykes, will no longer be sufficient. The policy is to allow more space for water. In order to prevent floods, rivers are allowed to expand into side channels and wetland areas. Greater emphasis is also placed on managing water levels rather than keeping the water out. This paradigm shift can also be observed by the current implementation of non-structural measures such as flood insurance, flood zoning restrictions, land-use management, economic incentives, public information, and community education (Raadgever and Mostert 2005). Non-structural measures are intended to modify flood susceptibility and flood impact. Also, voluntary measures by land users (e.g. change of land-use) are now starting to come into practice.

Discussion and conclusions

Comparing the responsiveness of the management regimes to drought and low flow problems and their responsiveness to flood problems suggests that the former is considerably lower than the latter. Drawing on the discussion in the preceding two sections, part of the explanation may be that flood problems are more directly perceived and experienced by not only experts and policy makers, but also by the wider public, creating a ′policy window′ through public pressure (which may induce responses) and concern (which may help politicians to legitimize responses). This research shows that in a basin where one type of extreme is dominant—like droughts in the Alqueva (Portugal) and floods in Rivierenland (Netherlands)—the potential impacts of other extremes are somehow ignored or not perceived with the urgency they might deserve. Within this context, it is important to acknowledge the importance of a critical awareness threshold, prior to the opening of a policy window. We argue that such a critical awareness threshold is directly related to joint/participative information production, as a necessary prerequisite for reaching a critical awareness threshold concerning the impacts of climate change and the need to take anticipatory action. However, such a policy window will only appear if the other conditions, as suggested by Kingdon (1995), have been met. For example, based on the conclusions of the Dutch National Safety Reports (2006), but also based on expert judgment in this research, it is justified to conclude that the political climate in the Netherlands does not make the time right for change as regards the development of a new and innovative drought policy.

In contrast to the poor responses to drought and low flow problems, a higher level of AIWM seems to result in more adequate responses to flood problems, at least in terms of flood protection or mitigation measures. The response of more adaptive and integrated regimes is characterized by a higher diversity in the type of measures being implemented or planned, including more attention for non-traditional measures, such as green by-passes and natural retention of flood water. The higher diversity in response measures indicates that horizons of possibility are being expanded, which corresponds to the concept of triple loop learning developed by Hargrove (2002:118). Case-studies showing a lower level of AIWM, for example the Lower Guadiana in Portugal, merely improve performance (e.g. increase reservoir volumes) without changing guiding assumptions or without taking entirely alternative actions into account. This involves doing things better without necessarily examining or challenging underlying beliefs and assumptions (Kahane 2004). These types of actions correspond to single loop learning (Hargrove 2002). In other words, our research indicates that case-studies showing a higher level of AIWM seem to have higher levels of learning in terms of its physical interventions. However, this observation needs to be further investigated by scrutinizing the policy responses of the management regimes in terms of different levels of policy learning (Grin and Loeber 2007). Nevertheless, this first observation is in line with our working hypothesis that AIWM requires many instances of social learning to implement and sustain innovative approaches (Pahl-Wost et al. 2007c). Additionally, our research shows that uncertainty related to climate change is often considered as an excuse for not taking action. An exception is case-study Rivierenland where the Room for Rivers-policy is taking into account climate change scenarios, besides other types of uncertainty, such as political and technological uncertainties (Huntjens et al. 2007, update 2009). Hence, we argue that management regimes having the intention to constructively deal with uncertainty are doing this by the active involvement of a diverse group of stakeholders in policy-making and by means of joint/participative information production. Moreover, we have also seen that some uncertainties might be overcome by conducting policy experiments, such as the management experiment near Avelingen (in the Netherlands) for testing alternative designs for the processes of developing integrated flood management plans (Huntjens et al. 2007, update 2009).

In terms of the current status of physical interventions in the river basin, the responsiveness of all case-studies to deal with extreme drought events is poor, and we have not found conclusive evidence that a higher level of AIWM creates a better response to drought and low flow problems. However, it is important to acknowledge that a regime with a higher level of AIWM may not have achieved its projected outcomes yet, since there normally is time lag between policy development and factual implementation. Our research provides evidence that drought response/adaptation seems to be slower than responsiveness related to floods. This might be explained by different risk perceptions (Green et al. 2007; Neuvel 2004; De Hollander and Hanemaaijer 2003; Lijklema 2001) and differences in the availability of solutions. We also suggest that these differences might be explained by the nature of the problem itself, whereas flood management is determined by safety concerns, drought management is determined by water scarcity and related problems in the allocation of water resources. Moreover, we argue that the threat of floods is often perceived as more threatening and acute than the threat of droughts, since the latter is spread out of longer time periods and consequences are often felt indirectly. Moreover, in the Netherlands, the risk of drought problems and water scarcity caused by climate change is not acknowledged by all stakeholders, as indicated by the experts in this research, but also by Neuvel (2004). Combining the poor response to drought and low flows problems with the enormous social, economic, and environmental effects in recent years, it is justified to support Milligan’s statement that droughts and heat waves are the developed world’s hidden disaster (Milligan 2004), and in this case the hidden disaster of Europe. We argue that a regime element such as joint/participative information production is crucial for identifying hidden disasters and for reaching a critical awareness threshold for initiating a policy window. From this perspective, it is expected that a regime with a higher level of AIWM is more responsive to the threat of new or hidden disasters than a less adaptive regime. This seems to be confirmed by case-study Rivierenland, since they have been confronted with serious droughts only recently, but the first steps are being taken as regards developing a nationwide adaptation strategy for this specific problem, expected to be published in 2012 (Pers. Comm. Luit-Jan Dijkhuis, Directorate-General Water, Netherlands). However, the current political climate might be a limiting factor for consolidating a comprehensive strategy.

Besides the earlier explanations for the different responses to flood and drought problems, we also argue that looking at physical interventions provides an incomplete picture of the capabilities or outputs of a water management regime. Hence, drawing conclusions merely based on the physical response would be premature, and we argue that a management regime with a higher level of AIWM might be more productive in terms of management interventions than in terms of physical interventions. In other words, our research provides valuable insights as regards the research design for analyzing complex multi-level governance systems. One of the limitations of the research design in this paper is the narrow definition of the outputs of a management regime, being defined as physical interventions in the river basin. Apart from the physical intervention in the water system, a regime’s performance could and should also be measured by its ‘non-physical’ measures, such as awareness campaigns, measures for improving information management and exchange, cooperation structures, and insurance mechanisms. In principle, these are measures which target at all the elements of a water management regime, and even could target at (changing) the formal institutional setting (e.g. European Directives and national or regional water laws). This means that there could be many nonlinear feedback loops within the regime itself, and the regime is in that sense creating its own enabling environment. Especially in regimes with a higher level of AIWM, the formal institutional setting is being altered by the demand for governance as regards (new developments in) water-related problems, such as the impacts of climate change. Examples include the Dutch National Water Agreement (Bestuursakkoord Water 2002), leading toward the start of implementing the Room for Rivers-policy (PKB Ruimte voor de Rivier 2006), and the Hungarian National Drought Strategy (2004). Therefore, in succeeding research of the EU Newater project, we will focus on climate change adaptation strategies, being defined as an output of the management regime. These strategies will be assessed by looking at physical AND management interventions. Additionally, the outputs of a management regime will be evaluated in terms of different levels of policy learning (Hall 1988; Bennett and Howlett 1992; Sanderson 2002; Huitema and Meijerink 2007; Grin and Loeber 2007). We believe that such a research design, thanks to lessons learned in this research, is better equipped for analyzing the adaptive capacities of water management regimes.