Opportunities and Barriers for Water Co-Governance—A Critical Analysis of Seven Cases of Diffuse Water Pollution from Agriculture in Europe, Australia and North America

Abstract

Diffuse Water Pollution from Agriculture (DWPA) and its governance has received increased attention as a policy concern across the globe. Mitigation of DWPA is a complex problem that requires a mix of policy instruments and a multi-agency, broad societal response. In this paper, opportunities and barriers for developing co-governance, defined as collaborative societal involvement in the functions of government, and its suitability for mitigation of DWPA are reviewed using seven case studies in Europe (Poland, Denmark, Sweden, The Netherlands and UK), Australia (Murray-Darling Basin) and North America (State of Minnesota). An analytical framework for assessing opportunities and barriers of co-governance was developed and applied in this review. Results indicated that five key issues constitute both opportunities and barriers, and include: (i) pressure for change; (ii) connected governance structures and allocation of resources and funding; (iii) leadership and establishment of partnerships through capacity building; (iv) use and co-production of knowledge; and (v) time commitment to develop water co-governance.

1. Introduction

Diffuse water pollution from agriculture (DWPA) can consist of nutrients from fertilizers and manure, sediment, pesticides, and salinity from irrigation return flows. It constitutes a threat to water quality and aquatic ecosystems in many farmed regions of the world [1,2,3,4,5,6,7]. Information regarding DWPA is incomplete and asymmetrical; the polluters often have more information than the regulators [8]. This makes mitigation a difficult challenge for public policy and governments, requiring complex, multi-faceted solutions [9,10].

Sources of DWPA can include agricultural fields and feedlots, rural domestic wastewater treatment systems, private gardens, wildlife, golf courses and other sport facilities. This makes DWPA more difficult to monitor and regulate than visible point sources, such as discharge from factories and large-scale urban wastewater treatment plants. Consequently, when adopted as a sole policy measure, the costs of regulation and its monitoring and enforcement will be high and likely prohibitive [11,12,13]. Diffuse Water Pollution from Agriculture is conventionally managed through a combination of policy instruments [10,14]. These can include command and control regulation, such as nationally set limits for applied nitrogen per hectare; voluntary adoption of best management practices resulting in environmental and farm management improvements; and market-based incentives such as fertilizer taxes, “cap-and-trade” schemes for nutrient or given agro-chemical use in a given area, and payments for ecosystem service (PES) compensation for production and income foregone [15].

None of these policies are likely to be sufficient and cost-effective if used alone. Advocates of command and control regulation argue that it can internalize the cost of pollution and provide a continuing incentive to improve agricultural practices in ways that minimize use of the polluting input. This has proven to be effective in reducing excess use of nutrients in farming in some regions and countries [16], but, when used in isolation, a “command and control” approach may not be a cost-effective and sustainable long-term solution to secure clean water [17], especially if such regulation is not accepted as reasonable and legitimate by stakeholders, and is thus met with social and political resistance and by poor levels of stakeholder compliance [18].

Voluntary adoption of best management practices encouraged by farm advisory systems can elucidate opportunities to protect the environment, whilst also saving on farm production and labour costs. Examples include soil testing and precision application of fertilizers, along with proper manure management. High rates of adoption may be difficult to achieve without a complementary “stick” provided by (a threat of) regulation, and, if change in practice is constrained by farm income or availability of capital for investment, a complementary “carrot” provided by subsidies and other forms of financial support may be necessary.

Information needs, costs, difficulty of monitoring, and attribution to a source, limit the scope to directly tax DWPA emissions. Taxes on the polluting inputs, such as fertilizer, have been employed, but effectiveness may be constrained by inelastic demand and trade competitiveness at feasible price levels. Such policy may also risk pollution swapping, and if applied uniformly at a national scale, may exhibit a poor match with local variation in ecosystem capacity to absorb pollution [10]. “Cap-and-trade” type schemes for inputs use or emissions have potential too, but are demanding of information and institutional development. Lastly, PES style compensation requires at least a complementary baseline of simple and enforceable regulation to distinguish between environmental standards that farmers are expected to meet at their own cost and higher standards for which society is willing to pay. Budget constraints will necessitate spatial targeting, whilst the challenge of monitoring necessitates high levels of voluntary compliance.

The complexities of this policy agenda, the diverse range of actors involved, and the need for budgetary efficiencies, spatial targeting, and voluntary compliance have implications for governance [19,20]. To improve the effectiveness of DWPA policy design and implementation, there is a need for improved knowledge on, first, which governance forms, structures, mechanisms and processes help tailor and deliver the right combination of regulation, incentives and voluntarism in a given location; and, second, how these governance forms, structures, mechanisms and processes can be developed given prevailing institutional opportunities and barriers.

Governance refers to how a society and its economy are managed, including its institutions, organizations and policies [21]. Besides such generic usage, “governance” has been understood in political science as a shift from government by solely state actors and institutions to multi-level and polycentric forms, involving private and civil society actors, along with distribution of responsibility and authority (e.g., [22,23,24,25]). This can result in collaborative governance or “co-governance”, which we define as direct societal involvement in the core functions of government [26]. Co-governance can be characterized as multi-level and polycentric, with tasks delegated to the most suitable scale given existing responsibilities, local specificities and the scale of the issues concerned [10,27]. Thus, co-governance refers to the transition from hierarchical governance and exercise of power by the state, to more dispersed and relational power, often exercised in layered networks of governmental actors and stakeholders that enter into collaborative relationships with each other. The role of government becomes one of process facilitation, rather than command and control.

The aim of this paper is to contribute to the understanding of alternative governance arrangements for management of DWPA. It presents new knowledge on the effectiveness of co-governance arrangements using an analysis of seven international case studies. It focuses on the identification of key enabling factors (opportunities) and constraining factors (barriers) for the development of effective co-governance arrangements for the mitigation of DWPA.

First, we present the method and cases. Then, we discuss in more detail co-governance for DWPA mitigation and develop a framework for analysing the case studies. Finally, we present the result of applying this framework to the seven cases, and draw several conclusions.

2. Materials and Methods—Case Studies on Water Co-Governance

In this study, a multiple case study design was applied [28], involving seven DWPA and co-governance arrangements (

Table 1. Selected case studies.

Case Study	Location	Diffuse Water Pollution from Agriculture
Norsminde catchment	Denmark	Nitrogen in surface waters
Tullstorp stream	Sweden	Nutrients in surface waters
Province of Overijssel	The Netherlands	Nitrogen in groundwater
The Catchment Based Approach	UK	Sediment, nutrients and pesticides
Kocinka catchment	Poland	Nutrients in surface waters
State of Minnesota	USA	Nutrients in surface waters
Murray-Darling Basin	Australia	Salinity in surface and groundwater

). Much of the literature on co-governance discusses single-case studies [29], but here a diverse set of case studies were selected to represent heterogeneity in the nature and severity of DWPA, the bio-physical, socio-economic and governance contexts, and governance approaches (see Appendix A for further description of the individual cases).

The seven cases were selected at an international workshop on water co-governance held in Copenhagen in 2016. The workshop (International workshop on Water co-governance—opportunities and barriers for effective decentralized decision-making in nonpoint source regulation of agriculture, held in Copenhagen, Denmark 28–29 November 2016) was initiated, co-organized and co-developed by the first author. At the workshop, 16 participants working with governance and DWPA theoretically and in practice were invited to present and share experiences regarding opportunities, barriers and outcomes from diverse co-governance frameworks (see Appendix E for a list of participants attending the workshop). The cases were selected by a case selection strategy based on having both most similar and typical cases on DWPA (DK, Sweden, NL and UK) and the most different and diverse cases on DWPA (USA, Australia, and Poland) [30]. By analysing both most similar and most different cases, the strategy was to be able to capture both the opportunities and barriers for improved water co-governance under different institutional arrangements. The aim of the workshop was to gather qualitative insights on opportunities and barriers for effective mitigation of DWPA. Researchers and practitioners were invited to present examples of how mitigation of DWPA was approached from a governance perspective. The workshop agenda progressed from empirical descriptive presentations of the seven case studies to the development of an analytical framework for comparative analysis (see Section 3 and Appendix B, Appendix C and Appendix D).

After the workshop, an electronic survey was prepared and distributed to the participating researchers to obtain further details about the cases. The case study experts scored each case for the key indicators that were developed as part of the analytical framework, using Likert-type scales [31]. This was supplemented by observational data and a narrative report of key processes and outcomes for each case study (see Appendix A for the narrative report summary and Appendix D for an overview of the survey structure).

2.1. Analytical Framework

To identify opportunities and barriers for effective co-governance for mitigation of DWPA, an analytical framework was developed at the aforementioned workshop. This included the specification of indicators of effective co-governance. The use of such indicators can help to simplify and describe the case and communicate the findings in a detailed and structured way [32].

The analysis framework developed consists of six indicators: two contextual indicators ((i) pressure for change; and (ii) institutional arrangements (the degree of delegation of decision making by task)); two process indicators (iii) the ability to develop and/or use innovative solutions; and (iv) the use of local knowledge); and two outcome indicators (v) goal achievement; and (vi) cost-effectiveness) (Figure 1). This framework focuses on the nature and performance of co-governance currently observable in each case study. It is acknowledged that it does not itself address wider contextual factors including socio-economic and socio-political facilitators or constraints, although key issues are identified as concisely as possible in Appendix A.

2.2. Pressure for Change

The pressure for change indicator expresses the degree to which DWPA is recognized as a societal and political concern and a cause of environmental, social and economic costs. In studies completed in regions where farmers do not believe that agriculture contributes to water pollution, there is a low level of perceived pressure for change amongst the polluters, which may contribute to low uptake of DWPA mitigation measures [33]. In other examples, wider non-agricultural stakeholders put DWPA on the political agenda, resulting in an increased pressure for change from those experiencing the impacts of pollution (e.g., [34]).

A low level of pressure is unlikely to result in more than incremental change to existing regulation, policies and procedures, whilst a greater pressure for change is likely to result in widespread media coverage and increased societal awareness. This can prompt a political response, including emergency and longer-term measures.

Pressure for change is ranked from low, moderate, major to severe (see Appendix C). It is hypothesized that there may be a relationship between the level of pressure for change and the degree to which co-governance has developed in each case.

2.3. Institutional Arrangements

For this indicator or rather group of indicators, institutional arrangements are analysed in terms of the allocation of decision-making tasks relevant to DWPA mitigation; i.e., “who is responsible for which tasks”. The tasks are expressed in terms of steps in the policy cycle (policy formation, policy implementation and policy evaluation), and the allocation is expressed in terms of scale: local, regional or central (see Appendix B). The aim is to explore how responsibilities for DWPA mitigation are shared or not shared between actors and levels of government, whether delegation occurs at the level most suitable to account for existing responsibilities and local specificities, and to be most effective for decision making for the issues concerned.

2.4. Process Indicators

Many indicators of successful co-governance processes have been identified in the literature and are sometimes also described as intermediary outcomes [35]. Process outcomes do not relate to a direct change in mitigation of DWPA at the point of time at which they are evaluated but can be essential to attain the expected outcome achievements [36]. The two process indicators chosen for this study can help describe how co-governance has been developed and what the cases are doing to address the DWPA problem. The two process indicators are the following:

(1)

The ability to develop and/or use innovative solutions in mitigating DWPA. In other words, how effectively do the stakeholders collaborate in developing and adopting solutions?

(2)

The use of local knowledge. This refers to: (a) the capacity for knowledge sharing, i.e., openness/accessibility of public sector data (including that of municipalities and water companies); and (b) the effectiveness of using data and knowledge (combining the use of public sector knowledge/data and local/private knowledge/data).

These indicators were ranked on a scale from none, weak, moderate, good to excellent (see Appendix C).

2.5. Outcome Indicators

The ultimate aims of DWPA mitigation are ecological and economic improvement. For these, many indicators can be found [35]. In this study, effectiveness is defined in terms of achieving two outcomes:

(1)

Goal achievement is assessed as the ability of governance arrangements to significantly contribute to achieve statutory local or national goals for DWPA mitigation and ecological improvement [35].

(2)

Cost-effectiveness is assessed as: (a) the ability of case governance arrangements to ensure that desired outcomes are being delivered affordably with the use of available resources; and (b) the ability of case governance arrangements to deliver desired outcomes at the lowest cost [35].

These indicators were ranked on a scale from none, weak, moderate, good to excellent (see Appendix C).

2.6. Opportunities and Barriers for Water Co-Governance

Based on the analysis of the contextual, process and outcome indicators, key issues are identified in the different cases that create either opportunities or barriers for co-governance for DWPA mitigation. Further data to inform this are drawn from the workshop presentations, workshop discussion and narrative template completed by case experts (see Appendix A, Appendix B, Appendix C, Appendix D and Appendix E).

3. Results

3.1. Pressure for Change

For the European cases, the assessment of pressure for change ranges from low to major. In the USA case, the ranking is moderate, and for the Australian case severe (see

Table 2. Expert rankings of the “pressure for change” (for mitigation of DWPA).

Case Study	Pressure for Change
Norsminde catchment (Denmark)	Moderate to Major
Tullstorp stream (Sweden)	Low to Moderate
Province of Overijssel (The Netherlands)	Moderate to Major
Catchment based approach (United Kingdom)	Major
Kocinka catchment (Poland)	Low
State of Minnesota (USA)	Moderate
Murray-Darling Basin (Australia)	Severe

).

These assessments are made at the catchment/case study scale and relate to the perceived pressure for change from stakeholders. For example, in the Polish case based on the Kocinka catchment, there is a low level of pressure for change from farmers and rural communities. The area is rural and relatively economically underdeveloped, and, although farming is a significant source of water pollution, it is not highly intensive so there is little recognition amongst the farming community of the need for change. However, pressure for change, at national level and amongst stakeholders in government, is much higher (Appendix A.5), given the need to transpose and implement relevant EU directives and international concern regarding Poland’s contribution to nutrient pollution of the Baltic Sea.

In the Swedish case, pressure for change was low when the Tullstorp stream association was initiated in 2009, as the national response to DWPA. The EU Water Framework Directive (WFD) was at early stage of development, but local stakeholders were motivated to prepare for the introduction of new legislation, suggesting a moderate pressure for change from farmers and the rural community.

In the Danish and Dutch cases, the pressure for change in each locality was assessed as moderate, but the pressure for change at the national level was major. This is because mitigation of DWPA is a high profile, politicized and controversial issue within both Danish and Dutch society, given the trade-offs between the profitability of intensive agriculture and better water quality. In the 1980s, severe pressure for change was put on policy-makers to deal with DWPA in Denmark, which initiated a range of action plans, that were almost all implemented with a top-down approach [16].

The UK case concerns the national level rather than a local catchment situation, and the pressure for change has been assessed as “major”. The launch of the Catchment-Based Approach (CaBA) by government was part of the response to widespread pressures for improved implementation of the WFD, which led to environmental NGOs threatening to use a judicial review (Appendix A.4). At a local level, stakeholders including farmers recognized an impending threat of stricter and more vigorously enforced water protection regulation, but also an opportunity for more integrated land and water management strategies.

In the USA case, also larger in scale (the state of Minnesota), a co-governance framework for water quality was developed because of pressure in the form of stricter federal government requirements (USEPA), which were triggered by eutrophication problems in the Mississippi River contributing to the Gulf of Mexico hypoxic zone [37]. A critical factor was also the approval by state voters of an increase in sales tax for investment in cleaner water and other environmental improvements.

The Australian case is also large scale. In the Murray-Darling Basin, salinity can render water unusable for drinking and reduce agricultural yields and thereby farmer income. (Australia is an ancient dry continent and salt has accumulated in the landscape over millennia. While having economic benefits, the massive increase in irrigation started mobilizing the salt and river salinity started increasing. By the early 1980s, salinity levels at Morgan in South Australia exceeded 800 EC units (500 mg/L) for 18 months on end. In time, salinity levels have reached peaks of 1400 EC units, threatening the future of drinking water and irrigated horticulture. Ideally, drinking water should have less than 800 EC (WHO aesthetic limit), and horticultural crops are impacted above 500 EC. The Salinity and Drainage Strategy set a target to reduce the salinity levels at Morgan below a 90% percentile of 800 EC units for a benchmark climatic period). The fact that salinity was affecting the profitability of farms played a vital role that helped initiate co-governance. The upstream states of New South Wales and Victoria wish to manage water tables and discharge saline drainage water, but the downstream state of South Australia cannot tolerate high salinity levels. These pressures for change were therefore assessed as severe.

3.2. Institutional Arrangements

In all of the European cases, with exception to the Swedish study, the central government has the task and statutory responsibility of choosing goals and setting targets for mitigating DWPA (see

Table 3. Analysis of task allocation, showing an overview of responsibility for “who makes decisions” (who is responsible for which tasks) in the different steps in decision-making (a)–(j).

Cases
Tasks— “who is responsible for which tasks”: C = Central; R = Regional; L = local	Norsminde (DK)	Tullstorp (SE)	Overijssel (NL)	CaBA (UK)	Kocinka (PL)	Minnesota (USA)	MDB c (AUS)
(a) Choosing goal	C	R	C	C	C	C + L (State a)	C d + R e
(b) Setting targets	C	R	L (R)	C + L	C	C (State)	C + R
(c) Allocation of resources for action plans	C	C	R	C + L	C	C (State)	C + R
(d) Prepare action plans	C	R	L	R + L	C + R	C (State)	L f
(e) Approve action plans	C	C + R	R	C + R	C	C (State)	C + R
(f) Implementation	L	L	L	R + L	R + L	R (LGU b) + L	R + L
(g) Receipt of resources for action plans	L	R + L	L	R + L	R + L	R (LGU)	R
(h) Monitoring and control	C	C + R + L	R	R + L	C + R	C (State)	R
(i) Data collection	C	C	R + L	R + L	C + R	C (State) + R (LGU)	R + L
(j) Data use	C + L	R	R + L	C + R + L	C + R	C (State)	C + R + L

^a State is a state agency for Minnesota, ^b Local Government Unit (LGU) is the county or catchment scale, ^c Murray-Darling Basin (MDB), ^d Central refers to the river basin authority (the Murray-Darling Basin Commission), ^e Regional was applied to State Governments (in the Australian federal system), ^f Local refers to the community groups that developed the local Salinity Action Plans.

). In Sweden, the goals related to the implementation of the WFD and are chosen at the regional level.

In the USA case, the tasks of choosing goals and setting targets are shared between the central level and local government. To meet Federal requirements, the Minnesota Pollution Control Agency (MPCA) must complete Total Maximum Daily Loads (TMDLs) for impaired State waters. Impairment goals are set based on these TMDL reports, but the methods for prioritizing and addressing the impairments are defined by each county’s Soil and Water Conservation agency, a Local Government Unit. Although the initial goals are set by the State in the form of Total Maximum Daily Loads, the local government establishes what the specific, more local goals will be and then they are shared with the MPCA, at the State level.

In the Australian case, choosing goals and setting the target for salinity management is a shared responsibility of the central and local level. The goals, targets and strategic framework of the Salinity and Drainage Strategy were developed by the then Murray-Darling Basin Commission (MDBC), replaced by the Murray-Darling Authority, in 2007. The MDBC was established as an unincorporated joint venture bringing together officials from three States and the Federal Government; however, the MDBC consulted with affected communities in developing the targets. In this way, the regional Salinity Action Plans (SAP) were developed by communities (bottom up) to achieve the target. The development of the Basin wide strategy was achieved by an effective partnership between the Federal and State governments (central level). The regional communities supported the strategy and developed their own partnerships with local stakeholders [38].

For tasks beyond setting of goals and targets (listed as (c)–(j) in Table 3), there is more diversity in the allocation of responsibility across the cases. The cases from Sweden, the Netherlands, the UK, Poland and Australia tend to exhibit a greater degree of decentralisation for tasks related to action planning and implementation.

3.2.1. Development and Use of Innovative Solutions and Use of Local Knowledge and Data

The Australian and USA cases stand out as scoring highest for the process and outcome indicators (see

Table 4. Processes and outcome indicators for effectiveness.

Effectiveness (None, Weak, Moderate, Good, Excellent)	Norsminde (DK)	Tullstorp (SE)	Overijssel (NL)	CaBA (UK)	Kocinka (PL)	Minnesota (USA)	MDB (AUS)
Development and use of innovative solutions	Weak	Moderate	Good	Excellent	Weak	Good	Good
Use of local knowledge and data 1—Knowledge sharing—openness/accessibility of public sector scientific data (including that of municipalities and water companies)	None	Good	Excellent	Moderate	Weak	Good	Good
Use of local knowledge and data 2—Co-production of knowledge—combined use of scientific data and local knowledge/data	Moderate	None	Good	Good	Weak	Good	Good
Goal achievement	Good	Excellent	Good/Weak 1	Moderate	None	Moderate	Excellent
Cost effectiveness 1—desired outcomes being delivered affordably with the use of available resources	Moderate	Good	Good/Weak 1	Good	None	Good	Good
Cost effectiveness 2—desired outcomes being delivered at a lower cost than by alternative means as far as can be determined or judged	Weak	Weak	Good/Weak 1	Excellent	None	Good	Good

¹ Good at farm scale for participating farmers, weak at the scale of the recharge area considering that only approximately 20% of the agricultural area participates in program activity.

). The Netherlands and UK cases also exhibit some success by this measure.

An important part of the success in the Australian case has been locally developed innovative solutions, involving mapping and surveying regional soil salinity levels, in addition to re-allocating irrigation water away from the more saline soils (two highest soil salinity categories). This approach required a huge shift in community thinking, but because the decision was made by the local community, it was implemented. All data were shared across the area affected by the Basin Salinity and Drainage Strategy. Community groups preparing the plans had access to a high level of advice and knowledge from experts in the government agencies and access to all the available data. At the same time, the strategy and framework for the development of the SAP required co-production using both local knowledge of the local community group developing the SAP, and the scientific knowledge of State experts (e.g., hydrogeological), informed by the strategic framework provided by the Basin Salinity and Drainage Strategy.

In the Minnesota case, knowledge sharing and co-production of knowledge are continuing to improve with the process that has been deployed through the water quality framework. The One Watershed, One Plan (1W1P) (see Appendix A.6) approach has enhanced the local commitment to addressing the problem. There is an understanding that there is not an easy fix when it comes to water quality improvement. Rather, there are multi-faceted barriers that require everyone to collaborate and work together to address the problem holistically.

In the Dutch case, the leading example of innovation and use of local knowledge is that mitigation measures are identified and implemented together with the farmers at the farm level, based on the so-called Annual Nutrient Cycle Assessment (ANCA), with data from the farm. In the UK, Catchment Partnerships created under the CaBA are demonstrating success in stakeholder engagement, partnership working, leverage of multiple funding sources and used of shared data and decision support tools for local level planning [39].

The rankings for the Danish local case are moderate overall. Here, the preparation of action plans and the selection of measures are decided at the central government level and implementation has to follow national guidelines. New policies are typically implemented at national level without prior pilot projects. Altogether, the conditions for introducing innovative solutions are not good. The local catchment council has been excellent in implementing drainage solutions and wetlands by themselves; however, there is no coordination with official authorities.

3.2.2. Goal Achievement and Cost-Effectiveness

The development of the regional SAP in the Australian case has been successful in that the salinity target was reached in 2010 through the introduction of salinity debits and credits, water trading rules and salinity levies. At present, salinity levels are continuing to decline in the river basin. The formation of a river basin authority and the development of a single salinity strategy for the southern MDB allowed all the salt mitigation measures to be ranked according to cost-effectiveness (salt removed from the river/$). The strategy thereby assured that the most cost-effective options were selected and implemented. If the four States involved (Victoria, New South Wales, South Australia and Queensland) had made SAPs individually, higher cost solutions would have been selected because each State would have only been capable of investing within its own borders.

In the USA case, there is moderate evidence that the 1W1P process adopted in Minnesota will lead to well-coordinated implementation. All local and State agencies join together to inform the process, but ultimately the implementation plan is driven by the priorities of local stakeholders to refine the previously identified catchment-based strategies into an actionable path forward. The success can be explained by the programs for federal and State implementation funding being utilized by Local Government Units (LGUs) and landowners at the catchment scale. This suggests that the local implementation has been a defining factor for the Minnesota program. A key goal in this case has been to engage more people in the science available and the issues at hand, which can be seen by more programs currently being initiated to improve public outreach.

The Swedish case is an example of a bottom-up process initiated and driven by the landowners themselves. The goals were set by the landowners and are in some ways far from co-governance in that no authorities are directly involved in the project, as opposed to the USA and Australian cases. However, in the Swedish case, the landowners have achieved their own goal for implementation of measures and have restored 10 km of the stream and created 35 wetlands in the catchment.

In the Danish case, the top-down centralized regulation has been successful in goal achievement and ensured that farmers in Denmark have reduced nitrate leaching from the root zone by almost 50% between 1987 and 2007, through command and control measures [16]. However, additional reductions of 30–50% will be required to meet the WFD objectives of good ecological status. It has been suggested that to achieve these environmental objectives, agri-environmental measures should be spatially targeted to fields where nature is less effective at removing or retaining nitrogen [40,41].

The Dutch case shows evidence of both a decrease in the nitrate concentration of groundwater and an improvement in the economic performance of farms. The nitrate concentration decreased from approximately 90 to 75 mg NO₃/L in the period 2011 to 2015, and the economic result of the implemented measures increased approximately €97 per hectare. This shows that outcomes can be delivered at a lower cost than by alternative means, and that the combination of top-down and bottom-up structures in the Netherlands makes goal achievement feasible and cost effective at the farm level. The success was achieved by setting clear targets at the farm level in combination with use of the farm management tool ANCA for monitoring. However, due to the voluntary nature of the project, only 20% of the farmers in the agricultural area participated. Hence, the achievement of regional goals at the scale of the recharge area has been limited, and at the national level, objectives for complying with the WFD objectives have not yet been met.

In the UK, the pilot and evaluation phase of the CaBA was concluded in March 2013. Over time, it is expected that the CaBA approach will mature as a mechanism for ensuring that there is strong local support, consensus, effective coordination, and efficient channelling of existing finances, new funding, and additional resources to deliver local aspirations for the water environment. The CaBA partnerships drive cost-effective practical delivery on the ground, resulting in multiple benefits, including improvements to water quality, enhanced biodiversity, reduced flood risk, resilience to climate change, and greater community engagement with the local river.

A contrasting case to the northwestern EU, Australian and USA cases is the Polish case. Poland has been successful in terms of achieving goals related to improving the water quality status of particular water bodies and has made progress in terms of harmonizing existing water legislation and its water governance structure with EU requirements. However, Poland’s performance in engaging local stakeholders has largely fallen short of the objectives set in the WFD, consistent with the results summarized with the Kocinka case.

Although there is some diversity, the rankings in Table 2, Table 3 and Table 4 tend to suggest that the European cases, which all operate under EU Directives related to DWPA (the WFD and the Nitrate Directive), are less developed in terms of their co-governance processes and outcomes for mitigation of DWPA than the Australian and USA cases. In the next section, key issues for these differences are discussed.

4. Discussion

The case study results suggest that the contextual, process and outcomes indicators for co-governance of DWPA mitigation are positively related. The two cases that score highest on the contextual indicators, the USA and Australian cases, also score highest on the process and outcome indicators. In this section, we discuss the opportunities and barriers for effective co-governance.

4.1. Pressure for Change

Pressure for change to mitigate DWPA can come from citizen’s expectations concerning water quality; from government in response to either citizen expectations or demands from higher-level government; and from farmers if DWPA harms their crops or to avert strict government regulation. If high enough, any pressure for change is an opportunity for effective co-governance, as shown in the USA and Australian cases. What may explain the shift in governance in the Australian case in comparison to the European cases, is that the pressure for change was severe and came from both political enforcement and societal concerns (see Table 2). The severe salinity issue in the MDB has been a major factor in achieving joint action and community support. A clear community understanding of the threat created political capital for reform, allowing governments and community groups to make some tough decisions. The fact that rising salinity directly impacted the profitability of the farms was also a key factor in persuading farmers and their communities to accept difficult decisions. Mitigation of DWPA can be difficult because the costs are individual and often upstream, while the benefits are often collective and downstream. Salinity has more immediate consequences for crops and drinking water supply than, for example, nitrate pollution, which can make mobilization of stakeholders, resources and change to mitigate salinity less difficult than it may be to mitigate nitrate pollution.

When the stakes are high and clear, as in the Australian case, this can create an opportunity for action and allocation of resources. This may be contrasted with the Polish case concerning nitrates, where pressure for change was lower due to lack of information and resources.

Pressure for change may also be primarily “top-down” and from government, as in the Swedish, Danish, Dutch and UK cases. Nutrient pollution can be very damaging to water quality, but not necessarily an issue for farmers unless national or international (EU) regulation limits the use of fertilizers and manures. Political and societal concerns may play a role as well, either as an opportunity or an obstacle.

For example, in the Dutch case, parliament decided that WFD-implementation should not lead to additional costs for the agricultural sector (Parliamentary Papers 2002, 27 625 Water Policy, Amendment Van der Vlies No. 92). This implies that in specific vulnerable areas, where the existing general rules on the use of manure and pesticides are insufficient, the WFD objectives may not be reached [42]. Meanwhile, the agricultural sector and especially younger farmers, have taken initiatives to introduce more sustainable farming practices on a voluntary basis.

The situation in the Danish case is similar. The newest N-reduction target for Norsminde by 2016 is about 70%. Studies have shown that such reductions can only be achieved by taking 50% of the agricultural land out of production [43]. The present Danish Minister of Environment and Food has stated that even a 10% reduction of agricultural area will not happen during his term, showing the difficulties for change when there is low pressure.

4.2. Connected Governance Structures and Allocation of Resources and Funding

A second key opportunity found in this study is the need for establishing connected governance structures (multi-level governance) in the institutional arrangements of mitigating DWPA. Many water systems are rather large (e.g., the Murray-Darling Basin) and may require a coordinated approach to maximize effectiveness and manage conflicts.

In the Australian and USA cases, we see that this issue is addressed by creating multi-level governance partnerships.

In the Australian case, many of the decision-making processes are shared between the central (river basin authority), regional (State governments) and local level (community groups). The community groups were formed by the State governments to develop the SAP. Since the Murray River flows through three States, the central level had to be involved as well. The Salinity and Drainage Strategy provides an example of balancing top down governance establishing targets and directing investment with bottom up governance reflecting the practicalities of addressing local problems.

In the USA case, the 1W1P vision had the purpose to align local water planning with State strategies. Although the initial goals are set by the State (i.e., meeting a specific water quality goal) the local government establishes what the specific more “local” goals will be and then they are shared at the State level. This kind of framework is trying to provide more ownership at the local catchment so that there is more collaboration and communication between all parties at the local scale as they work to meet the State goals. This results in a connected governance structures and decentralization of responsibility down to the most effective scale.

The main finding of the Australian, USA and UK cases is that targets and goals can be set at the central level if this is done in coordination with the local actors that are responsible for local implementation.

An explanation why water co-governance has not been so developed in the other European countries is that the institutional arrangements in many of the European countries are top-down and centralized (see Table 3). The Dutch case shows that if the co-governance initiative is purely voluntary and no central authority is part of the process, there is a low rate of participation and a risk of free riding behaviour (Appendix A.3). In the case study area WFD targets (50 mg NO₃/L) were translated into targets for each participating farmer, to be realized by improving agricultural management. Plans are set-up and agreed on with the individual farmers. However, because of the voluntary character of the project, the plans have no formal status and there are no sanctions if they are not implemented.

In addition, the Danish and Swedish cases show that purely farmer driven, bottom up initiatives are not enough in themselves (Appendix A.1 and Appendix A.2). There will be a lack of integration in the formal regulatory framework, and central or regional goal setting is necessary for the creation of legitimacy and continuity in developing solutions for mitigating DWPA. Co-governance is by definition a process of interaction in which either the State “invites” stakeholders to participate in its core activities [26] or the stakeholders mobilize the State to engage in joint solutions of DWPA problems. Therefore, if there are no joint action and connected governance structures initiated either by the government or by stakeholders, no genuine or effective water co-governance is likely to develop.

Another key finding, related to the need for connected governance structures, is that hydrological boundaries, e.g., the watershed, and administrative boundaries often do not coincide. It is then important that the collaboration across administrative boundaries takes place. In the Australian case, for instance, implementation of cost-effective measures was made possible because the involved authorities cooperated across administrative borders. In the first stage, the most cost-effective salt mitigation schemes were in South Australia, which resulted in Victoria and New South Wales investing their funds in South Australia.

Another key finding from the Australian and USA cases is that delegation of authority to lower level actors requires allocation of funding and freedom to spend the funds according to locally chosen priorities. Resources and funding are essential to coordinate and host meetings, prepare reports and supporting documentation and implement communication and outreach efforts.

In the Australian case, the salinity issue required fundamental reforms. The River Murray Commission, which had been established in 1915 to manage water flows in the Murray River, had no role in water quality and its powers did not extend beyond the main river. In 1987, the Murray-Darling Basin Commission was created and the MDB Ministerial Council formed, which was assigned the key decision-making role. The first task of this new Commission was to develop the Salinity and Drainage Strategy. The strategy was successful and based on a combination of centrally driven governance and community consultation supported by local Salinity Action Plans, which were developed from the ground up by community groups empowered by regional (State) governments.

4.3. Establishment of Partnerships and Strong Leadership

The Australian, USA, UK and in part the Netherlands cases indicate that a key enabling factor for effective water co-governance is creating an institutional and administrative framework within which stakeholders with different interests can discuss and agree to cooperate and coordinate their actions. Therefore, partnerships are important and may function where other governing structures do not [21,29].

These cases support the findings from the literature that co-governance happens when affected stakeholders are engaged in an equal partnership and the stakeholders share their perspectives on the development and implementation of policies [44,45]. This implies that decision-making powers are shared with stakeholders affected by the policy (polycentric governance) [46].

The Polish case provides a contrasting example for the importance of partnership. Poland is characterized by a weak civil society with low levels of trust [47,48]. Low levels of trust are indicative of an unwillingness to participate and collaborate. The lack of a participatory culture in Poland is a major barrier to realizing water co-governance. A key lesson from the Polish case is that, due to lack of trust, there is a lack of support for delegating power from the central level to local level government and stakeholders. This can be explained with narrow perceptions of defining participatory approaches by the national government, revealing a reluctance towards devolving power to local levels of government.

In the Swedish case, the project manager was active in starting the project and provided the necessary leadership, perhaps because governmental did not. By having another person facilitate the process, the manager had the time and resources to search for additional funding and deal with administrative issues. Landowners would probably have been overwhelmed by these tasks if done alone. Trust in the “leader” is essential for establishing good relationship, but because the project manager was a farmer. All the landowners had trust in this person and in the first stage they all signed the agreement sending a positive signal to other landowners. Another important lesson from the Swedish case is that to establish partnerships, meetings also have to serve as a social get-together with the chance of exchanging information and knowledge. The question is whether the lessons from the Swedish case also apply elsewhere in Sweden and can be used to achieve national goals and integrate national regulatory frameworks.

One of the main reasons why there is currently no co-governance in the Polish Kocinka catchment is the missing need for action, along with the lack of ownership and leadership for starting such an initiative. In the USA case, a key issue was that a successful program requires clear communication regarding roles and expectations.

An important lesson from the Dutch case was that individual support and coaching by agricultural advisors was a major factor for implementation and participation in the project. This coaching was carried out by senior experts interested in the farmer and his management and providing advice based on farm specific conditions and practical experience of an experimental farm (De Marke) in the Netherlands. Without this support and coaching, the willingness of farmers to change and participate would have been substantially lower, reducing the chance of meeting the standards.

4.4. Innovative Solutions Through the Use and Co-Production of Knowledge

The co-governance structure in the Australian and USA cases has led to positive environmental outcomes by deploying new innovative solutions for water co-governance. For example, in the Australian case, the Nyah to the South Australian border SAP, contained a system of trading rules and levies to direct irrigation water away from high salinity impact zones to low impact zones. Additionally, stakeholders mapped the soil salinity levels in the Tragowell Plains SAP, which included a large data collection exercise by the local community and resulted in irrigation water being transferred away from the two most salinized soil categories. Both innovations were developed by the local community and required effective interchange of knowledge between the State and Local SAP, effectively co-producing scientific (e.g., hydrogeological) and local knowledge of the community group developing the SAP [49].

Data sharing and co-production of knowledge have been identified as important process indicators for development and success of co-governance. In the Australian case, data were shared across the area affected by salinity. At the same time, the local SAP required knowledge co-production combining local knowledge and experience with the scientific knowledge of State experts, informed by the strategic framework of the Basin Salinity & Drainage Strategy.

Similarly, in Minnesota, community input through a public participation process helped to incorporate local knowledge into the strategies. The first round of 1W1P pilot projects were launched in 2014 with the intent to build on existing catchment efforts, using local plans together with state and local knowledge, implementing a systematic and science-based approach to catchment management.

In the Swedish case, data are locally collected and shared openly, even though this knowledge sharing takes place only at the local level. Scientific actors and their expert knowledge are not involved in choosing the measures or integration of this knowledge at the catchment level or with other water councils. This is again related to the question of available funding and lack of multi-level coordination. The landowners are open to see improvements in environmental conditions if this does not hamper their business model in the absence of a coherent top-down framework of regulation and funding mechanisms.

Rogers and Hall [21] showed in the analysis of principles for better governance that a collaborative process should be open and transparent. This requires that decision making is transparent so that both insiders and outsiders can follow the necessary steps.

In the Dutch case, measures are identified and implemented together with the farmers at the farm level, based on the ANCA with farm scale data. Farmers are using the ANCA data to improve management for optimizing nutrient use efficiency. This is in contrast to the Polish case, where knowledge sharing is limited to the information and data that are relevant to achieving policy deliverables. Knowledge is collected at the local level and shared with higher levels of government, with central government determining what knowledge is relevant. At present, there is a limited opportunity for co-production of knowledge as collaboration and public engagement are largely absent.

4.5. Developing Co-Governance Takes Time

An additional and clear observation that emerges from the more successful case studies is that the development of effective co-governance is an evolutionary process that requires time. The USA case shows that finger-pointing to fix DWPA problems does not work. Rather, these problems are multi-faceted and require everyone to collaborate and work together to address the problem holistically. There are no easy fixes and it takes time. In the USA case, time and patience were essential to the development of the partnerships in order to create a catchment-based management plan. Taking time at the beginning to coordinate and ensure clear and consistent communication may smooth the process later. Including stakeholders with a range of expertise, backgrounds, and perspectives into the partnership and valuing their input resulted in more trust, which should lead to more acceptance and better implementation of the resulting strategy.

In the Australian case, time for institutional development was also shown to be essential. It took a long time to reach a consensus over difficult issues in the MDB.

In the Dutch case, considerable time was needed to build trust, create awareness, and show (with monitoring results and farm data) what role the farmers have in the DWPA and how the measures implemented help in reducing the nitrate concentration.

5. Conclusions

Co-governance is not an end in itself, but a means to an end. In the context of this paper, its value is its potential to enhance both better social outcomes and better water quality by better implementation of DWPA policies [50]. Evidence for improvement can be difficult to find, given monitoring challenges and time lags. The seven cases applied in this study, have shown that actual environmental improvements can be achieved under co-governance arrangements. Moreover, this paper shows that five opportunities and barriers for water co-governance may exist and that experiences may be shared across catchments, regions and nations.

In our study of the seven cases on water co-governance in a DWPA context, we have identified a number of opportunities and barriers for co-governance:

• pressure for change;
• connected governance structures and allocation of resources and funding;
• establishment of partnerships and strong leadership through capacity building;
• use and co-production of knowledge; and
• developing co-governance takes time.

First, when dealing with a challenging problem such as DWPA mitigation, a high level of pressure for change (e.g., Australian case) can be an opportunity for co-governance development and a low level an obstacle (e.g., Polish case). However, social and political factors may play a role as well. This is shown in the Danish case, where societal pressure has pushed for reforms and regulation of the agricultural sector, which through the last 30 years have reduced the nitrogen leaching by 50%.

Second, effective co-governance requires institutional arrangements in the form of multi-level governance structures that connect and coordinate top-down policy frameworks and funding mechanisms with bottom-up planning and implementation, the latter being essential for knowledge co-production, innovation and supporting voluntarism. The current provisions of the EU WFD directive could prompt and support such institutional arrangements because the WFD directive is a “framework” directive. It sets ambitious goals but there is flexibility for member states on how these are achieved. In turn, member states can delegate that responsibility to a local level.

Third, establishment of partnerships and strong leadership through capacity building and knowledge sharing is essential for the development and support of connected governance structures. Member states need to invest in the capacity building necessary for the co-governance necessary to plan and manage change. The needs for horizontal integration and coordination are also important and could be further researched. This can facilitate planning alignment and complementarities in resource use and interventions. Relevant observation and literature also notes the importance of “bridging” or “boundary” organisations (e.g., [51]).

Last, we have shown in our analysis that developing effective co-governance frameworks for mitigating DWPA takes considerable time to deliver upon intended objectives. At the same, it is important that stakeholders are involved early in the policy phases to create trust and commitment. It is also important that there is continuity in policy, political support for institutional arrangements and funding mechanisms.

By focusing on a qualitative analysis of five opportunities and barriers in different contexts, we have identified new and transferable knowledge from our cases that expands the current knowledge of what makes multi-level co-governance work.

The five opportunities and obstacles are by no means exhaustive, but if they were addressed comprehensively and consistently they offer real promise of sustainable environmental outcomes in the world’s important water systems.