Boundary work to what end? Analysing the acid mine drainage case in Gauteng, South Africa (2024)

Abstract

1. Introduction

Complex and controversial environmental policy problems, also known as unstructured problems, require urgent policy interventions, which made it even more challenging due to high-decision stakes, scientific uncertainties, and value disputes that typically characterize them (Funtowicz and Ravetz 1993; Turnhout, Hisschemöller, and Eijsackers 2008). Due to the urgency required, decision-makers who develop policy interventions are forced to depend on best available yet incomplete knowledge and on carefully managed negotiations at the science–policy interface (Funtowicz and Ravetz 1993).

For this article, we study such an unstructured problem in the form of the acid mine drainage (AMD) policy controversy, set in the Witwatersrand’s gold mining areas of the greater Johannesburg region of South Africa. Once producing 50 per cent of the world’s gold output, the 1990s saw the cessation of much of greater Johannesburg’s gold mining activities (Bobbins 2015). This ultimately resulted in massive acidic mine water discharges from an abandoned mine shaft on the West Rand from August 2002 onwards, presenting a significant threat to environmental, animal, and human health (Bobbins 2015). Experts played a pivotal role in addressing the AMD problem, characterized by considerable disagreements and uncertainties about its immediate and potential impacts and the most appropriate treatment and management options (Funke etal. 2021). For this article, we use Weible (2008) definition of the term ‘experts’, which include policy analysts, scientists, consultants, and researchers in government and non-governmental organizations (NGOs).

The main aim of this paper is to evaluate whether the boundary arrangements that were set up at the science–policy interface to develop solutions to the AMD problem resulted in successful policy implementation. To do so, we analyse the boundary work that took place between experts and policy-makers. Boundary work is defined as the context-specific, strategic, social interaction processes that serve to demarcate and coordinate between scientific and policy domains in the context of a specific policy problem (Halffman 2003; Hoppe, Wesselink, and Cairns 2013). Boundary work is increasingly being recognized as a valuable approach for understanding the relationship between actors at the science–policy interface, tasked with informing and promoting action in response to complex policy problems (Van Heland, Crona, and Fidelman 2014).

For our analysis, we draw on key conceptual elements of several boundary work frameworks, as a combination of the analytical strengths of multiple frameworks has greater explanatory potential than any single framework could (Cairney 2013).

First, we ask the question ‘what is the work that needs to be done?’ To answer this question, we analyse the boundary arrangements that were set up at the AMD science–policy interface (Guston 2001) in terms of four core functions: convening, collaboration, translation, and mediation (Cash etal. 2003).

Second, we ask ‘who and what does the work?’ Here, we apply Halffman’s (2003) ‘texts, objects, and people’ framework to study the micro-level boundary work that took place within the boundary arrangements identified from the dual perspectives of demarcation and coordination.

Third, we consider ‘what are the success conditions of this work?’ by evaluating the influence and usability of the knowledge produced during the boundary work for decision-making. We do so by applying Cash etal.’s (2003) criteria of salience, credibility, and legitimacy.

Fourth, we ask the question ‘which enabling and constraining factors have influenced the success of the work that was done, and with what result?’ Here, following Van Kerkhoff and Pilbeam (2017), who also conduct a boundary work analysis in a developing country context, we recognize the value of taking into account various enabling and constraining factors, characterizing the socio-political context within which the AMD problem is situated. In particular, we evaluate how these factors influenced the policy implementation of the short-term and long-term solutions that were developed to address the problem.

We have selected concepts from the work of these particular frameworks as they present well-known and highly cited contributions to the boundary work tradition. We also argue that our synthesized framework is particularly useful as it consists of relevant conceptual elements that enable us to present a natural temporal progression from high-level boundary arrangements to context-specific boundary work and finally to influence and implementation. This enables us to emphasize the temporal dimension of boundary work (Schut, Van Paassen, and Leeuwis 2013; Van Heland, Crona, and Fidelman 2014) rather than being restricted to an interpretation that is overly static.

Our focus on both influence and implementation addresses a gap in boundary work literature that has been identified by White etal. (2010). These authors argue that the research should place greater emphasis on the ultimate outcomes of boundary work rather than only focusing on the methods and institutional forms employed. McNie (2007) contributes to this argument by stating that the ultimate metric of what constitutes useful knowledge is whether this knowledge is used to design and implement sustainable policy solutions.

To explain whether policy implementation was successful, we reflect on the influence of contextual enabling and constraining factors on boundary work, which constitutes a meta-level analysis. Incorporating this level of analysis addresses the criticism that most conceptualizations and applications of boundary work tend to limit their focus to micro-level boundary work analyses at the science–policy interface rather than taking crucial contextual factors into account (Hoppe 2010a; Dilling and Lemos 2011; Koch 2018). Understanding contextual factors is particularly relevant in developing country contexts, which are typically marked by strong power dynamics that, for example, accord greater importance to objective science vis-à-vis other forms of knowledge (Van Kerkhoff and Pilbeam 2017).

From an empirical perspective, we contribute a boundary theory analysis of an interesting South African case study which has not yet been studied from this perspective. This is particularly relevant given Koch’s (2018) argument that the existing body of research on science and policy interactions is strongly biased towards Western conceptualizations, ideals, and experiences, which are often used as recommendations for developing country contexts, even though they may not be entirely applicable. Therefore, various scholars have highlighted the need to address the current gap in literature on the science–policy interface in developing country contexts through empirical research contributions (Jones, Jones, and Walsh 2008; Hoppe, Wesselink, and Cairns 2013; Clark etal. 2016). With this article, we just make such a contribution.

2. Situating our analysis within the boundary work tradition

Science and policy are often portrayed as being separated by a divide that needs to be bridged so that expert knowledge can contribute effectively to managing complex environmental problems (Turnhout, Hisschemöller, and Eijsackers 2008). To better understand interactions that take place at the science–policy interface, authors contributing to social constructivist studies of science have depicted the boundaries at this interface as the outcomes of context-specific, strategic, social processes, known as boundary work (Koch 2018).

The term boundary work was coined by Gieryn (1983), who originally focused on activities involved in the demarcation of science from non-science. An example of such an activity is scientists drawing boundaries around their own ‘scientific’ activities to distinguish them from other intellectual activities (e.g. in the economic, legal, or health domains) (Halffman 2003).

Soon afterwards, researchers started applying the concept of boundary work in the context of the science–policy interface (Koch 2018) by examining boundary work as activities that distinguish science from policy but at the same time negotiate the conditions for interaction between the two domains (Halffman 2003). When working together, experts and policy-makers continuously create, negotiate, and maintain boundaries to demarcate each other’s responsibilities and functions, thereby shaping and reshaping the science–policy interface (Jasanoff 1994; Halffman 2003). To illustrate, Ascher, Steelman, and Healy (2010) identify the politics of knowledge as strongly contributing to the boundaries between environmental science and politics and argued that these need to be reimagined, especially through the integration of local knowledge and public preferences into decision-making. In addition, Braun and Kropp (2010) highlight the importance of reflexivity in scientific governance, which is often hampered by various limits and ambiguities as the approach competes with more control-dominated approaches.

Several seminal studies on the topic of boundary work (Jasanoff 1994; Guston 1999) have interrogated the processes and rules of boundary-making and the actors and organizations involved therein (Koch 2018). Subsequently, boundary work has developed into a comprehensive approach that incorporates multiple concepts, for example, ‘boundary objects’ (Star and Griesemer 1989), ‘boundary-ordering devices’ (Shackley and Wynne 1996), ‘boundary arrangements’ (Halffman 2003; Hoppe 2010b), and ‘boundary organizations’ (Guston 2001).

Several meta-analyses related to the topic of boundary work have also been conducted. For instance, Bremer and Meisch (2017) emphasize the need for greater reflexivity by researchers using the concept of co-production and highlighted the opportunity for combining different perspectives on this concept. Relatedly, Jensen-Ryan and German (2019) find that prioritizing the alignment of different knowledge types may promote the successful scientific integration into policy-oriented outcomes more than depending too much on formal boundary organizations.

In addition to Western theoretical conceptualizations of boundary work, boundary work–related concepts have also been (1) applied to various developing country contexts or (2) modified to improve their explanatory potential in such contexts.

For example, such applications include a focus on successful knowledge co-production and boundary work processes around freshwater conservation (Nel etal. 2015) and the finding that multiple boundary arrangements can co-emerge and co-exist at different research–stakeholder interfaces (Schut, Van Paassen, and Leeuwis 2013). Several applications also focus on the usability of the knowledge produced during boundary work processes with specific reference to Cash etal.’s (2003) criteria of salience, credibility, and legitimacy. For example, these include a focus on the importance of clear communication, context, and social equity (Brauman etal. 2013) and the possibility of increasing the credibility and legitimacy of controversial climate issues by linking them to other high-priority policy issues (Hoppe, Wesselink, and Cairns 2013).

Several authors have also proposed modifications of Western boundary work–related concepts. Examples include Ziervogel, Archer van Garderen, and Price’s (2016) argument for the replacement of the science–policy interface with the knowledge–policy interface (e.g. to include local knowledge) in the context of adaptation policy and planning and Clark etal.’s (2016) identification and call for support for distinct kinds of boundary work that have contributed to the success of sustainable food security research and policy programmes in different countries. Developing country–focused authors have also added an emphasis on the socio-political context within which boundary work takes place. Importantly, Van Kerkhoff and Pilbeam (2017) propose the concept of knowledge governance as bringing together ‘targeted inquiry into the socio-political context in which environmental science is situated, alongside analysis of specific interventions that change knowledge-to-action relationships’. Related to a focus on the socio-political context, Koch (2018) lists and analyses certain enabling factors, such as supportive legislation, which have contributed to the successful boundary work efforts of the South African National Biodiversity Institute. In contrast, other authors have focused on challenges or constraining factors, including resource inequalities (Van Heland, Crona, and Fidelman 2014) and tensions (such as the scientization of politics) characterizing the science–policy interface (Jones, Jones, and Walsh 2008). Taking the focus on challenges one step further, Dinesh etal. (2021) emphasize rather than downplaying the fail factors that can influence the outcomes of the science–policy engagement efforts, such as the lack of absorptive capacity in institutions.

3. Theoretical approach

This section presents the different components of our synthesized boundary work framework. The different components are presented inFig.1 and are summarized inTable1.

3.1 Boundary arrangements

Our first conceptual element that we use to answer the question ‘what work needs to be done?’ is that of boundary arrangements and the four core functions that such arrangements are expected to perform. Boundary arrangements refer to a wide range of collaborative institutional configurations that are situated at the interface between expert and/or stakeholder networks and public sector or policy organizations (Guston 2001). These arrangements act as bridges to manage the relationship between the science and policy domains while supporting the development of evidence-based, relevant, and effective policies and/or boundary objects (Hoppe, Wesselink, and Cairns 2013).

Examples of boundary arrangements include expert advisory committees, research management agencies, and advisory boards (Halffman and Hoppe 2005).

The four functions which boundary arrangements are expected to perform include a convening function (bringing stakeholders together to facilitate dialogue and exchange), a collaboration function (enabling collaboration to promote the production of knowledge and boundary objects), a translation function (to facilitate communication by translating between the different languages of the science and policy domains), and a mediating function (to ensure equal treatment of the different participants and their perspectives) (Cash etal. 2003).

In this article, we analyse the boundary arrangements that were set up to develop a short-term intervention and long-term solution to the AMD problem according to the four expected functions and whether successful policy implementation was achieved.

3.2 Boundary work at the science–policy interface

Our second set of conceptual elements that we utilize to answer the question ‘who and what does the work?’ is Halffman’s boundary device–focused framework, which focuses on the contributions of textual, material, and social boundary devices to demarcate and coordinate interactions at the science–policy interface (Halffman 2003).

Boundary texts refer to the discourses, concepts, literary devices, or language used to define the respective roles of experts and policy-makers at the science–policy interface (Halffman 2003) and serve the dual purpose of dividing these two domains and bringing them closer together (Gieryn 1983; Halffman 2003). For the AMD case, we analyse some of the key texts that were produced at the science–policy interface, notably the reports to recommend a short-term intervention and long-term solution to the AMD problem, in terms of both demarcation and coordination.

Boundary objects are conceptual, material, or physical boundary tools, techniques, or practices that are used to demarcate boundaries (Halffman 2003) but at the same time allow communication across different actors or social worlds (Halffman 2003). This is because they are not only ambiguous and flexible enough for different actors to interpret them differently (Star and Griesemer 1989; Shackley and Wynne 1996) but also rigorous enough to ensure a common identity (Cash etal. 2003). Boundary objects can be not only museums, maps, or specimens (Star and Griesemer 1989) but also more narrowly conceptualized as the material elements that form part of boundary arrangements, for example, the locations of meeting rooms (Halffman 2003). As a boundary object in the latter sense, we discuss the short-term water treatment facilities that were established in the Western, Central, and Eastern Basins under the management of the Trans-Caledon Tunnel Authority (TCTA) in the context of demarcation and coordination at the science–policy interface.

Boundary workers (Guston 1999) are actors from both policy and scientific domains who facilitate interactions at the science–policy interface (Jasanoff 1994; Guston 2001) as they possess the relevant communication skills, network ties, and reputation that enable them to access both sides (Koch 2018). Boundary workers can be individuals (Halffman 2003) or belong to boundary organizations or groups, such as committees, networks, and institutions. For this article, we analyse the roles of the boundary workers who participated in the boundary arrangements that were set up to develop recommendations for the short-term intervention and the long-term solution to the AMD problem in the context of demarcation and coordination.

3.3 Knowledge effectiveness in influencing decision-making

Our third analytical component that we use to answer the question ‘what are the success conditions of this work?’ is Cash etal.’s (2003) renowned ‘credibility, salience, and legitimacy framework’, which serves to evaluate the influence and usability of the knowledge produced during the boundary work for decision-making.

Cash etal.’s (2003) framework is widely regarded as the basis for evaluating the effectiveness of boundary work in resource governance case studies (Fudge and Hiruy 2019). According to this framework, for such knowledge to be influential and usable, it must be considered sufficiently credible, salient, and legitimate among various stakeholders (i.e.on both sides of the science–policy boundary) and across multiple institutional scales (Cash etal. 2003). The relationship between the three critical attributes of credibility, legitimacy, and salience is often characterized by trade-offs and synergies (Cash, Borck, and Patt 2006; Van der Molen, Swart, and Van der Windt 2018).

First, the knowledge produced during boundary work can be deemed sufficiently credible if it is widely perceived as authoritative, believable, and trustworthy (Cash etal. 2003). Credibility is also influenced by the perceived objectivity and track record of existing boundary arrangements and the reputation of those participating in these arrangements (Sarkki etal. 2015). The traditional standard for establishing the credibility of knowledge production processes is scientific peer review, but collaborative processes between different stakeholders are increasingly considered as evidence of credibility (McNie 2007; Pfaeffle etal. 2022).

Second, knowledge produced by boundary work is deemed sufficiently salient if it is considered relevant and important relative to other sources of knowledge (Cash etal. 2003) to address current or pending policy problems (Sarkki etal. 2015). This also includes being intelligible and relevant to the specific demands, values, and beliefs of the stakeholders involved (McNie 2007; Sarkki etal. 2015). Salient information should be informed by ecological, temporal, spatial, administrative, legislative, and political considerations (McNie 2007).

Third, such knowledge is deemed sufficiently legitimate if it is based on inclusive and fair processes, is transparent and objective, and is respectful of different values and beliefs (Cash etal. 2003). Additional important criteria for legitimacy include interdisciplinarity, objectivity, and transparent communication (McNie 2007).

For this article, we investigate whether and how the boundary arrangements that were set up to develop recommendations for the short-term intervention and the long-term solution to the AMD problem as well as the resulting outputs showed evidence of credibility, salience, and legitimacy.

3.4 The influence of enabling and constraining factors on policy implementation

Our fourth conceptual element that we use to answer the question ‘which enabling and constraining factors have influenced the success of the work that was done, and with what result?’ is a focus on the enabling (e.g. Koch 2018) and constraining factors (e.g. Dinesh etal. 2021; Jones, Jones, and Walsh 2008; Van Heland, Crona, and Fidelman 2014) that characterize the socio-political context within which boundary work takes place (Van Kerkhoff and Pilbeam 2017). We identify these factors for the AMD case and investigate their influence on the implementation of the short-term intervention and the long-term solution that were developed. We do not enter our analysis with a specific list of hypothetical enabling or constraining to be tested but instead allow these to emerge from our analysis.

4. Data and methods

Data collection for this article consisted of a literature search and nineteen semi-structured interviews with key actors involved in the AMD policy problem. We searched for literature focusing on AMD in South Africa between August 2002 and August 2023 on the Google, Google Scholar, SACat, EBSCOhost, and Scopus databases. Our sources included media reports, popular articles, transcripts of a South African investigative television programme, technical reports, academic articles, policy documents, official South African government communications, and the transcripts of Parliamentary Portfolio Committee hearings.

The nineteen semi-structured interviews complemented the data assimilated from the literature search as they contained respondents’ first-hand narrative accounts of their perceptions of various key dynamics that have shaped the AMD policy domain. Such accounts constituted a key input to our analysis of the boundary work that took place at the AMD science–policy interface. The interview questions focused on the roles of experts in developments in the AMD policy domain; actors’ collaboration with experts; the existence and degree of influence of different proposed technical solutions to the AMD problem; different beliefs about the AMD issue; the nature and extent of coalition formation, activity, and cross-coalition interaction; and actors’ views on the government’s response to the AMD problem. The interviews were conducted face-to-face or telephonically and were held with government officials from two government departments: a representative of a state-owned water treatment implementation agency and a representative of a mining company, academics, scientists, consultants, and environmental activists. All interviews were voice-recorded and transcribed.

The literature and interview data were consolidated and analysed using the cross-sectional code and retrieve method, which involves identifying and describing key themes and sub-themes (Spencer, Ritchi, and O’Connor 2003). The themes identified were boundary arrangements and their functioning, boundary work (texts, objects, and people), evidence of knowledge effectiveness in terms of influencing decision-making (salience, credibility, and legitimacy), and the influence of enabling and constraining factors on policy implementation. We subsequently used these themes and the relevant sub-themes and discussions under each to populate our analytical framework, as presented inSection5.2.

We adopted a semi-inductive approach for writing this article by consciously embarking on an analysis through the lens of our boundary work–focused framework but at the same time allowing the findings to emerge during our data analysis process (Creswell 1994).

5. Analysis

5.1 Case study background

The setting of our case study is the Witwatersrand’s gold mining areas of the greater Johannesburg region of South Africa(Fig.2). While operational, the once prolific gold mines in this area pumped water to the surface to enable them to access underground gold deposits, but when dwindling gold reserves resulted in the cessation of much of the gold mining operations in the late 1990s, the underground voids started refilling with groundwater. This rising groundwater interacts with exposed sulphite-bearing minerals in the rock formations to form AMD, characterized by a low pH (comparable to that of battery acid) and a very high concentration of dissolved metals and sulphate salts [TCTA (Trans-Caledon Tunnel Authority) 2021; Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010].

AMD subsequently began discharging onto the surface from a mine shaft of an abandoned gold mine close to Krugersdorp on the West Rand in August 2002 at a substantial volume of 15–20 Ml per day [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. The acidic discharge immediately started polluting various streams and tributaries flowing into the Vaal River, which ultimately feeds into the Vaal Dam, the main source of water supply to South Africa’s Gauteng province. Given its high levels of acidity, salinity, and heavy metal and radionuclide content, the polluted mine water began posing a severe environmental threat to the receiving water environment and surrounding ecosystems and a health threat to humans, animals, and plants [TCTA (Trans-Caledon Tunnel Authority 2021); Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010].

The Department of Water Affairs and Forestry (DWAF), which had been tasked with responding to the AMD problem, tried to bring the AMD discharge on the West Rand under control but only managed to implement ad hoc pump and treat methods together with some of the mining companies between 2005 and 2010 (Bobbins 2015). Although DWAF issued a series of directives to the Rand Uranium mining company, which offers mining services and produces both uranium and gold (EMIS 2013), to pump water from affected mine shafts, these were successfully appealed in courts of law and could ultimately not be enforced (Bobbins 2015).

In 2010, once underground gold mining on the Witwatersrand had almost completely ceased, the realization dawned that underground water levels were also rising in the Central and Eastern Basins, but with much uncertainty about the dates, volumes, locations, and potential impacts of the inevitable uncontrolled AMD discharge [DWA (Department of Water Affairs) 2013a]. Some of the impacts that were predicted at the time included the corrosion of the foundations of certain buildings in Johannesburg’s city centre [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010], which is located on the top of the Central Basin, and damage to electricity cables situated in underground tunnels that could ‘plunge the city into darkness’ (Bega 2010). It is here where our analysis begins (seeTable2 for a summary).

5.2 Case study analysis

5.2.1 Boundary arrangements.

The Inter-Ministerial Committee on AMD was set up by the executive branch of government in July 2010 and included the Ministers of Mineral Resources, Water Affairs, Science and Technology and the Minister in the Presidency: National Planning Commission. This was followed by the appointment of the Expert Team of the Inter-Ministerial Committee on AMD (hereafter referred to as the Team of Experts), which was instructed to advise the Inter-Ministerial Committee on the AMD situation in the Witwatersrand Gold Fields and produce recommendations for a short-term intervention [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010].

The Team of Experts consisted of prominent experts from state-affiliated institutions with a narrow and specific set of expertise based in the natural sciences and was instructed to propose a short-term intervention to the AMD problem. These experts specialized in environmental geoscience, hydrogeology, geochemistry, geology, sustainability science, and water utilization. Additional advice was provided by experts in seismology and chemistry [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010] from state-affiliated institutions with a narrow and specific set of expertise based in the natural sciences and was instructed to propose a short-term intervention to the AMD problem. Additional expert contributions were provided by several government agencies, as well as two government departments: the Department of Water Affairs (DWA), having changed its name from DWAF in 2009, and the Department of Mineral Resources [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. This strongly suggests that the recommendations of the Team of Experts, although having been produced by an independent core team, were influenced by government guidance.

The AMD problem was initially treated as a structured problem, which lends itself to the involvement of a bureaucratic, scientific/professional, or closed epistemic community to address it (Hoppe, Wesselink, and Cairns 2013). The Team of Experts represented such a closed community and, after an intensive internal consultation process, released a report that proposed a combination of improved pumping facilities and neutralizing the acidic mine water through high-density sludge (HDS) treatment as a short-termintervention [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. HDS treatment neutralizes acidic mine water and separates this mixture into alkaline water (i.e.HDS effluent) and a salt and mineral-rich sludge waste product which has to be disposed of [INPA (International Network for Acid Prevention) 2014]. This HDS effluent water has a significantly lower salt and slightly lower metal load than untreated mine water but requires additional treatment [TCTA (Trans-Caledon Tunnel Authority) 2012]. This recommendation was made with the proviso that additional water treatment would be required and that the government would need to decide which method to adopt [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. Evidence of the collaborative function can be seen in the Team of Experts’ presentation of the recommendations for a short-term intervention to Cabinet in February 2011 and the fact that a government directive to implement HDS treatment in all three basins was issued soon afterwards [TCTA (Trans-Caledon Tunnel Authority) 2021].

In April 2011, the implementing agent appointed by DWA, the state-owned TCTA, began HDS treatment on the Western Basin (Bobbins 2015). However, this implementation process did not run smoothly, as reflected in the growing number of complaints from the greater stakeholder community active in the AMD policy domain. In a private sector-commissioned report to investigate the impacts of rising AMD levels on certain buildings in Johannesburg’s city centre, the MWRG (Mine Water Research Group) (2011), affiliated to North-West University, critiqued the methods, data, and overall credibility of the short-term intervention on the basis that it had been developed too hastily and exaggerated the possible impacts of future AMD discharges. Corroborating this view, the Federation for a Sustainable Environment, the most vocal and influential activist organization on the West Rand, argued in a letter threatening possible legal action against the TCTA that the ‘defective’ assumption underpinning the decision to recommend HDS as the short-term treatment intervention resulted from the absence of ‘consultation and input from experts in the field and public at large’ [FSE (Federation for a Sustainable Environment) 2012].

In response to these critiques and threats of legal action, DWA decided to adopt a more consultative approach for the feasibility study it was commissioning to determine options for a long-term solution to the AMD problem. The study, led by the consulting firm Aurecon at the head of a consortium of specialist implementing partners, followed a more extensive stakeholder engagement process and addressed a number of content-related concerns that had been expressed regarding the short-term intervention [DWA (Department of Water Affairs) 2013a]. The decision to run a more consultative process is indicative of the government’s realization that the AMD problem was unstructured rather than structured and that a more inclusive approach was needed to facilitate discussion and learning on various difficult elements that needed addressing (Funke etal. 2021).

The focus areas of the long-term feasibility study process comprised technical, legal, institutional, financial/economic, and environmental assessments, predominantly carried out in the form of desk research, alongside public communication and key stakeholder engagements. The key outputs were a feasibility report and an implementation strategy and action plan. The stakeholder engagements involved not only experts who had participated in the short-term intervention development process but also national, provincial, district, and local government, NGOs, organized business, mining, industry, labour, agriculture, affected mines, affected water utilities, community leaders, and academics as key stakeholders [DWA (Department of Water Affairs) 2013b]. These engagements, for example, included technical workshops, individual consultation meetings, focus group discussions, and Study Stakeholder Committee meetings (with broad stakeholder representation to discuss common issues) [DWA (Department of Water Affairs) 2013b] and took the form of soliciting information, data and technical expertise as well as inviting certain stakeholders to review reports.

The feasibility study process took place over 18 months and was concluded in July 2013 [DWS (Department of Water and Sanitation) 2018]. Upon completion of the study, Aurecon recommended the construction of reverse osmosis (RO) treatment plants as reference projects in the Central and Eastern Basins [DWA (Department of Water Affairs) 2013c] against which other potentially suitable long-term AMD treatment technologies would be tested in the Western Basin. Reverse osmosis removes a large majority of contaminants from water by pushing the water under pressure through a semi-permeable membrane (Puretec Water 2016).

The long-term solution recommended as part of the AMD process builds on the short-term intervention’s HDS treatment by treating acidic mine water to industrial or potable water standards [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. This second step in the AMD treatment process carries both social and political significance due to the potential of selling the fully treated water to potential off-takers (Bega 2021), which is enormously valuable in the water-scarce South African context [DWA (Department of Water Affairs) 2013a]. In the words of the then Minister of Water and Sanitation, Nomvula Mokonyane, when officially announcing the implementation of the long-term solution in May 2016, ‘a polluted resource once considered with contempt, now becomes a commodity contributing to security of the availability of water resources in the Vaal River System’ [DWS (Department of Water and Sanitation) 2016].

This period of the AMD policy response also saw evidence of the other three functions expected of boundary work arrangements. These include the convening function, shown by bringing together a substantial number of stakeholders; the mediating function, a role performed by the study leader Aurecon who acted as a policy broker to mediate differences of opinion between the stakeholders involved [DWA (Department of Water Affairs) 2013b; Respondent 4]; and the translation function. The latter is demonstrated by the newsletters, press releases, and study reports which DWA used to communicate the outcomes of the long-term solution feasibility study process [DWA (Department of Water Affairs) 2013b].

5.2.2. Boundary work at the AMD problem science–policy interface.

5.2.2.1 Texts.

In their report, the Team of Experts recommended several urgent actions that should be taken immediately to start addressing the impacts of AMD in the Western, Central, and Eastern Basins. These actions included technical responses, such as pumping, to ensure that water levels are maintained at or below the environmental critical level for each basin to prevent further discharge at the surface level. The environmental critical level is defined as the highest water level within the mine void where no AMD flows out of the mine workings to pollute surrounding groundwater or surface water systems [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. Another urgent action was the consideration of financing options for short- and long-term AMD treatment [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. Furthermore, according to one of its members, the Team of Experts worked exceedingly hard to produce the report and believed that their recommendations would strongly influence the government’s short-term AMD response: ‘certainly at the time, we all shared this possibly idealistic view, that if we said something had to be done then probably it would be done…’ (Respondent 4).

The recommendations on a short-term intervention produced by the Team of Experts clearly reflect its advisory role and demarcate it from the decision-making role of the Inter-Ministerial Committee on AMD, to whom this group of experts reported.

By contrast, an example of coordination at the AMD science–policy interface can be found in DWA’s report on the progress of the AMD long-term solution feasibility study [DWA (Department of Water Affairs) 2013c]. This study drew on several different kinds of expertise, and these knowledge contributions served to inform the development of a ‘feasible solution’ that is ‘technically sound, economically viable, environmentally sustainable, institutionally feasible and legally achievable’ [DWA (Department of Water Affairs) 2013c].

5.2.2.2 Objects.

The short-term HDS water treatment facilities performed a material role as objects at the science–policy boundary. In April 2011, the Minister of Water and Sanitation issued a directive to the TCTA to start with the implementation of these facilities, while the feasibility study for a long-term AMD solution was being undertaken [TCTA (Trans-Caledon Tunnel Authority) 2018]. The main aim of the AMD short-term intervention was to extract (through pumping) and neutralize the acidic mine water before conveying the treated water to nearby water sources. Another critical objective was to maintain the level of acidic mine water below the environmental critical level, as defined for each basin [TCTA (Trans-Caledon Tunnel Authority) 2021; Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010].

From the perspective of demarcation, the HDS treatment facilities were characterized by a division between technical expertise (e.g. the technical process of neutralizing acidic mine water before release into the environment) and policy (e.g. the determination of what constitutes acceptable levels of treated effluent and the implications of using this treatment approach vs other approaches). From a coordination perspective, the implementation of the HDS treatment facilities was the result of the short-term intervention development process and the uptake of the targeted recommendations to decision-makers that had been prepared by the Team of Experts.

5.2.2.3 People.

Several boundary workers from the science and policy domains participated in the Team of Experts that was tasked to recommend a short-term intervention to address the AMD problem [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010]. A particular type of demarcation took place as only certain scientific experts were invited to join the team. These experts were drawn from public sector research organizations and, according to a government official who was involved in the appointment process, had to be able to demonstrate advanced academic qualifications in their respective disciplines, a solid publication record, and good project management skills (Respondent 18). Consideration was also given to the individual experts’ reputation and whether someone was likely to contribute constructively to ‘answer the question that government was asking’ or likely to be ‘difficult and derail the process’ (Respondent 18).

Despite the demarcation between some experts being included in the Team of Experts and others being excluded, there was also evidence of coordination across the science–policy interface as the participating boundary workers successfully cooperated to recommend an approach for short-term AMD treatment that was subsequently implemented by the South African government [Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) 2010; Bobbins 2015].

This trend of coordination amongst boundary workers working across the science–policy interface continued during the long-term solution feasibility study process and was strengthened by the fact that this process was considerably more inclusive of different stakeholders and their views than the short-term intervention process had been (Bobbins 2015).

5.2.3 Knowledge effectiveness in influencing decision-making.

The South African government intended the appointment of public sector experts to the Team of Experts to yield a high level of credibility, with the expectation that the authoritative, believable, and trustworthy knowledge would result from this process, given the credentials and experience of the experts involved (Funke etal. 2021). However, this credibility was vehemently challenged by some who had been excluded from this process (Funke etal. 2021), indicating that the scientific authority and peer review are not enough to guarantee the credibility of a knowledge production process nor the resultant knowledge outputs.

The short-term intervention process saw a relatively high level of salience from the perspective of the South African government, as the recommendations produced by the Team of Experts report were directly taken up and implemented in the form of HDS treatment (Funke etal. 2021). The choice of short-term intervention was strongly influenced by the seriousness which the government ascribed to the AMD problem in 2010. This is evident in the Team of Experts (Expert Team of the Inter-Ministerial Committee on Acid Mine Drainage) (2010), which explicitly refers to ‘the urgency of implementing intervention measures before problems become more critical’ and also mentions the Western, Central, and Eastern Basins’ ‘proximity to densely populated areas’. While, according to one of its members, the Team of Experts carefully considered various options before deciding on HDS, this treatment option was selected for various reasons. These included HDS being a useful first step for a multitude of potential subsequent long-term treatment options, its comparative simplicity and cost-effectiveness, and the fact that it was a tried and tested approach (Respondent 4).

However, the government’s views regarding the salience of the short-term intervention were not shared by those stakeholders challenging its underpinnings and the process by which it was developed (Funke etal. 2021). In addition, the Team of Experts report was limited to a largely technical assessment of the AMD situation and did not address other critical focus areas necessary for a comprehensive policy responsive, thereby further limiting its salience and legitimacy.

Although the South African government accorded legitimacy to the recommendations produced in the Team of Experts report by deciding to implement HDS as the short-term AMD treatment option, neither the report nor the process followed were regarded as inclusive, fair, transparent, or objective by those who had been excluded from it (Funke, Huitema, and Petersen 2022).

The feasibility study process exhibited substantially higher levels of credibility, salience, and legitimacy amongst all actors involved. The recommendations produced by this process were deemed credible enough by the South African government to strongly influence the official AMD long-term solution, which was announced in May 2016 (Funke etal. 2021). This process also saw a high level of credibility amongst other stakeholders in the AMD policy domain, because it was considerably more consultative than the process of developing the short-term intervention had been [DWA (Department of Water Affairs) 2013b; Funke etal. 2021]. This observation supports the view that collaborative stakeholder engagement processes are increasingly being considered as evidence of high levels of credibility (McNie 2007; Pfaeffle etal. 2022).

In addition, the recommendations resulting from the feasibility study were considered salient by all stakeholders involved as, due to the highly inclusive nature of the process followed, they constituted the primary input into the AMD long-term solution (Funke etal. 2021). The feasibility study also extended beyond the predominantly technical focus of the Team of Experts report and included several other critical aspects, including financial and legal, thereby more accurately reflecting the values, needs, and demands of the various stakeholders involved (Cash etal. 2003).

Finally, from the perspective of legitimacy, the feasibility study process can be described as more transparent, objective, inclusive, and fair (Cash etal. 2003) than the development of the short-term intervention as it included a much wider range of stakeholders in the consultations (Funke etal. 2021) and due to open communication around this process in a series of newsletters. Given the multi-dimensional nature of the study, the feasibility study can also be classified as being considerably more inter-disciplinary than the short-term intervention process had been (McNie 2007).

In summary, the AMD long-term solution feasibility study process and the knowledge developed as part of it saw a strong progression of credibility, salience, and legitimacy from predominantly being supported by the South African government towards securing buy-in from stakeholders across the AMD policy domain.

5.2.4 Influence of enabling and constraining factors on policy implementation.

In terms of the implementation of the AMD short-term intervention, by September 2018, short-term HDS treatment plants were running successfully in all three basins under the management of the TCTA [TCTA (Trans-Caledon Tunnel Authority) 2018]. By 2022, the AMD treatment plant in the Western Basin and Central Basins were meeting their AMD water treatment targets, while the Eastern Basin treatment plan was facing considerable operating difficulties and was awaiting the arrival of spare parts from Germany for repair [TCTA (Trans-Caledon Tunnel Authority) 2022].

In combination, the AMD treatment plants continued to contribute significantly to the prevention and reduction of pollution impacts through the abstraction and neutralization of AMD and subsequent release of the treated water into the Blesbokspruit and Klipspruit water courses, thereby resulting in the overall improvement of the condition of these streams [TCTA (Trans-Caledon Tunnel Authority) 2022]. However, the release of this partially treated water results in 362 tonnes of total dissolved solids or salts entering the Integrated Vaal River System on a daily basis (Bega 2021), thereby requiring dilution releases from the Vaal Dam [TCTA (Trans-Caledon Tunnel Authority) 2022]. The continued salt loading of the Integrated Vaal River System is threatening the acceptable levels of assurance of supply of the Vaal Dam, with negative social and economic consequences, particularly during periods of drought [DWA (Department of Water Affairs) 2013b].

While the implementation of the AMD short-term intervention seems to be going relatively smoothly, the same unfortunately cannot be said for the implementation of the AMD long-term intervention. In June 2021, the Mail and Guardian newspaper reported that the government had postponed its R10 billion long-term solution for AMD due to funding constraints and recalibration of the salinity and hydrology model for the Integrated Vaal River System (Bega 2021). At the time of writing this article, there had been no further communication in the public domain regarding the implementation of the long-term AMD solution.

Acknowledging the importance of the socio-political context within which the AMD issue is situated, we now discuss the enabling factors positively influencing the implementation of the short-term intervention and the constraining factors that had a negative influence on the continuing delays in the implementation of the long-term solution.

The first enabling factor was the strong political will and sense of urgency present within the South African government, especially between 2010 and 2012, to respond to the AMD problem. This translated into the recommendations of the Team of Experts report almost immediately being converted into the implementation of HDS treatment facilities in the three affected basins (Bobbins 2015).

The second enabling factor was the high degree of credibility, salience, and legitimacy which the South African government accorded to HDS as a tried and tested option vis-à-vis alternative responses to the AMD problem. The strong belief in the merits of HDS treatment as a short-term intervention gave the government the impetus to act swiftly and decisively despite the limited degree of credibility and legitimacy which the short-term intervention held amongst various influential members of the broader stakeholder community.

The third enabling factor was the fact that financial resources were made available to implement the short-term intervention and, importantly, that the South African government was willing and able to shoulder the full (and rather significant) cost of this treatment [DWA (Department of Water Affairs) 2013b].

Despite high levels of credibility, salience, and legitimacy characterizing the feasibility study process to develop the long-term AMD solution and associated outputs, as well as the Minister of Water and Sanitation’s official endorsem*nt of this solution in May 2016, two constraining factors in particular explain the persistent delays in its implementation.

The first constraining factor is a decline in political will which has gone hand in hand with the drop in the AMD issue’s status as a national emergency. When the AMD long-term solution was announced in May 2016, South Africa found itself firmly in the grip of the worst drought in 30 years (BBC 2015). This factor likely influenced the government’s impetus to augment the water supply to the densely populated Gauteng province by adding an additional desalination treatment stage to the current AMD neutralization process.

By 2018, however, the drought had largely subsided in Gauteng province, and the critical status of the AMD issue started waning as it was soon replaced by a string of other crises, including revelations of large-scale corruption in the public sector (Commission of Inquiry into State Capture 2018) and the crippling electricity supply crisis that continues to severely curtail South Africa’s prospects for economic growth (Ziady 2023). It remains to be seen at what point the issue of safeguarding Gauteng’s water security, closely linked to implementing the long-term AMD solution, will be deemed sufficiently urgent again to resurrect the required political will and associated policy response. Our observation that the AMD issue only managed to preserve its political prominence for a limited period of time mirrors a trend that has been observed in the USA, whereby domestic problems, often of crucial importance to society, suddenly gain considerable importance in the public eye but only remain prominent for a short time, even if they are still unresolved when their prominence starts to fade (Downs 1972).

The second constraining factor to the implementation of the AMD long-term solution pertains to a lack of funding (Bega 2021). This can largely be attributed to South Africa’s overall constrained economic position, which is evident in the forecast for a meagre 0.3 per cent in economic growth in 2023 (Ndaba 2023) and a 32.6 per cent official unemployment rate, exacerbated by factors such as the rolling power cuts that the country currently experiences on an almost daily basis (Gumbi 2023). In 2016, Minister Mokonyane announced mechanisms to recover the costs of the implementation of the AMD long-term solution, namely, an environmental levy imposed on the mining sector equating to 67 per cent cost recovery and 33 per cent cost recovery from other water users [DWS (Department of Water and Sanitation) 2016]. It is, however, likely that given the country’s precarious economic situation, there is a reluctance to increase the burden on an already overstretched middle class or the struggling yet politically influential mining sector.

6. Conclusion

In conclusion, we reflect on the most significant finding of our analysis and what it suggests about the science–policy interface in South Africa. We also reflect on how best to analyse this interface not only in South Africa but also in other developing countries. As demonstrated in our analysis, the AMD issue was first approached as a structured problem, leading to a closed, technocratic process to develop the short-term intervention. This process produced a report that was considered salient by decision-makers but, due to its exclusionary nature, lacked the necessary buy-in from the broader stakeholder community. The closed short-term intervention development process was followed by a much more open and deliberative process, which, according to Hoppe, Wesselink, and Cairns’s (2013) expectations, should have resulted in a much more stable and sustainable policy outcome. In fact, the long-term AMD solution was expected not only to resolve the continuing, albeit reduced, pollution of the Integrated Vaal River System resulting from the partial HDS treatment method but also to provide the substantial contribution of augmenting Gauteng province’s fragile water supply. Yet, despite expectations to the contrary, only the technocratic short-term intervention is running relatively successfully in the Central, Western, and Eastern Basins and, while by no means perfect or complete, has at least resulted in the reduction in pollution impacts to the surrounding groundwater and surface water sources [TCTA (Trans-Caledon Tunnel Authority) 2022].

What does this finding reveal about the AMD science–policy interface and how we should analyse it and those of other contentious environmental issues in developing country contexts? Regarding the interface itself, we see the critical importance of sufficient political will and financial resources as enabling factors when it comes to successful policy implementation. In the presence of these factors, the AMD short-term intervention was successfully implemented; however, in their absence, the promise of the implementation of the long-term AMD solution is likely to remain unfulfilled. On the topic of analysis, our article has shown the critical importance of adding a meta-analytical layer by taking into account relevant contextual factors. Adding such a layer can help us understand why, even if everything goes according to the plan at the micro-level, this does not guarantee successful policy implementation. We believe that adding a meta-analytical layer to boundary work analyses is especially critical for developing country contexts, often characterized by complex political and economic environments, but could benefit boundary work analyses in all country contexts.

Conflict of interest statement.

None declared.

Funding

None declared.

Data availability

The data underlying this article cannot be shared publicly due to privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.

References

© The Author(s) 2023. Published by Oxford University Press.