More than Floods and Droughts: Understanding Emergent Water Risks in South African Fruit Production Networks

Water is indispensable to agricultural production and trade. Using a global production network (GPN) framework, this article explores the variegated ways water risks manifest in agricultural value chains. Moving beyond extreme events, such as floods and droughts, we explore how the emergence and convergence of political, economic and ecological perspectives during the past 25 years has led to more nuanced understandings of water risks. By doing so, we contribute to ongoing efforts to move the environmental dimension more centrally into GPN literature, emphasising how globalised production and trade enmesh with environmental governance regimes. Our argument draws on empirical evidence from the South African fruit industry, which illustrates how water risks have a physical, reputational and regulatory/political dimension. These water risks are actor and context specific, manifest relationally to each other, and largely originate beyond the individual farm gate. They are often underlying and ongoing rather than one-off extreme events. These insights develop our understanding of how environmental risks are manifested and managed in GPNs.

This article draws on qualitative research conducted with the export-oriented fruit industry in South Africa's Western Cape throughout 2017.This included 76 semi-structured interviews with commercial farmers, industry and government stakeholders, as well as representatives from NGOs, consultancies and research institutions.
This introduction is followed by Section 2, which reviews the literature on GPNs and outlines our proposed water risk framework.In Section 3, we contextualise the study area and present our empirical evidence, detailing how fruit producers perceive physical, reputational and regulatory/political water risks.Section 4 ends the manuscript with a brief discussion, examining how water risks are actor and context specific, relational, predominantly exogenous, and tend to be underlying and ongoing rather than extreme events.This offers significant insight not only into how water risks affect the participation of producers in GPNs, but also how GPNs' actors embed in wider environmental governance regimes.

| WATER RISKS IN GLOBAL PRODUCTION NETWORKS
Our analysis engages with the concept of environmental risk as developed in GPN 2.0 (Coe & Yeung, 2015).GPN 2.0 builds on the GPN framework developed in the early 2000s within economic geography in dialogue with, and as a critique of global value chain (GVC) theories (Henderson et al., 2002).This criticism centred on the linearity of the GVC approach, which does not sufficiently take the spatial embeddedness of such chains into account.In contrast, GPN positioned itself as an explicitly relational and geographical framework.GPNs are understood as value chains that are also embedded in multi-scalar social, political and institutional contexts, while being shaped by economic and non-economic actors (Coe et al., 2008).
A GPN framework considers the horizontal value chain dynamics set by the market while also understanding how agricultural GPNs embed in local regulatory, socio-political and ecological landscapes.Through the concept of environmental risk, GPN 2.0 (Coe & Yeung, 2015) provides an interesting opening to centre the environment within its analysis.This is in line with a growing engagement with the environmental dimension in wider GPN/GVC literature, for example, environmental upgrading (e.g., De Marchi et al., 2013) or voluntary sustainability standards (Krauss & Krishnan, 2021).This work remains predominantly "vertical," investigating modes of inter-firm governance, for example, how stakeholders along the value chain environmentally upgrade or downgrade.There remains a lack of engagement with how GPNs interact with the governance of the natural resource base that is key to the production of agricultural goods, such as soil and water (briefly mentioned in Bolwig et al., 2010).
The understanding of environmental risk in GPN, however, is poorly developed (see Franz et al., 2018 for another critique), describing only extreme events (e.g., tsunamis).Characterised as a hazard or disaster it invokes a "one-off" conceptualisation of risk occurrence, negating the possibility of continuous and underlying environmental risks.This understanding of the environmental component is less nuanced than GPN theorists' contributions that explore social and economic dimensions (e.g., Barrientos et al., 2011).
Drawing on NGO reports (Reig et al., 2013), private sector actors (Pegram, 2010) and academia (Baleta & Winter, 2017), we develop a more refined understanding of environmental risk in agricultural GPNs.We propose a water risks framework that incorporates the variegated factors that threaten agricultural GPN actors' water security.As outlined in Table 1, we group these factors into the three categories of physical, reputational and regulatory/political water risks.This risks framework provides a route for exploring and developing the environmental dimensions within GPN literature.The notion of risk provides a useful focus as it accommodates the multiple scales that both water and GPNs traverse (Bakker, 2012).It also offers a common language between GPN and environmental governance and management literatures (Franz et al., 2018).Taking the Western Cape fruit industry as an example, we show the varied ways real or perceived water risks manifest for fruit producers.

| WATER RISKS IN SOUTH AFRICAN FRUIT PRODUCTION NETWORKS
The changing profile of water risks for Western Cape fruit growers is linked both to the specific South African context as well as more generic dynamics of global change.South Africa's history of colonialism and apartheid significantly shapes the present-day emergence of water risks for growers.Apartheid was a system of institutionalised racial segregation that justified social, political and economic discrimination against non-whites until 1994 (Gumede, 2015).Determining access to and control over water played a key role during apartheid because vast stretches of fertile and well watered land were allocated to the White minority, while the Black majority were confined to small areas of arid and unforgiving land (Jegede & Shikwambane, 2021).The Western Cape horticulture industry developed out of these dynamics of discrimination and dispossession and remains, today, dominated by White South Africans (Hall, 2004).Consequently, water allocation in the Western Cape -who gets how much water and why -remains unequal and contested, with farmers being given an annual allocation based on their crop types and land area.
Global dynamics that contribute to the emergence of new water risks in South Africa include neoliberal market restructuring and trade liberalisation, which inserted many agricultural producers into global production networks in the late 1990s (Bek et al., 2017).South Africa's democratisation and the subsequent deregulation of its horticulture industry coincided with the rising hegemony of European supermarkets within global trade.During the 1990s, traditional structures of state support and regulation were replaced with novel forms of governance, such as audit-based private standards for quality assurance and sustainability criteria (McKenna & Murray, 2002).In parallel, climate change is progressively manifesting materially, with changes in temperature and rainfall distribution felt acutely in water-intensive production systems (Distefano & Kelly, 2017).While producers need to adapt to and mitigate these effects, wider societal concerns about the ecological cost of intensive agricultural production systems are rising (Williams & Murray, 2018).The convergence of these dynamics has changed the ways farmers manage water and what information about water use needs to be disclosed to buyers and consumers, contributing to the emergence of new water risk profiles.
Producing fruit is water intensive, requiring reliable sources of good quality water.South Africa is one of the 30 driest countries in the world (DWS, 2018) with climate change induced water scarcity projected to cause declines in agricultural productivity of up to 50% (Davis-Reddy & Vincent, 2017).Besides limited physical water availability, fruit production in the Western Cape is embedded in a challenging political context resulting from the complex legacies of apartheid (Debbané, 2013).Using the framework developed above, this section presents an analysis of how variegated water risks manifest for producers in the Western Cape.

| Physical water risks
The Western Cape is characterised by a Mediterranean climate, meaning it rains predominantly in winter while summers are hot and dry (Midgley et al., 2016).Fruit, however, is grown in summer from October to March.This leads to an inverse relationship between water availability and need.As a result, farmers rely on water storage infrastructure (dams), as explained by one respondent: "[W]e are a winter rainfall area … and we get limited summer rainfall, so we have to store everything we use.You can't rely on any summer rainfall for irrigation."Building such dams is costly and can be administratively complicated.When winter rains are below average during consecutive years, as happened during the 2015−2018 drought, farmers are left to manage with only a fraction of their water allocation over the forthcoming summer months.
Climate change aggravates such dynamics.The Western Cape is one of South Africa's most at-risk regions with projected future changes in temperature from +1.5 to +3°C (Midgley et al., 2016), while South Africa as a whole is already witnessing mean annual temperature increases above the global average (Talanow et al., 2021).Higher mean annual temperatures lead to increased evaporation and decrease the water balance, while increasing demand for irrigation water.
As clarified by one producer: "In theory, if it's getting hotter, then we've got to put more water down."To maintain current production, total demand for irrigation water is projected to increase by 10% until 2050 (Midgley et al., 2016).Combined with population growth and increasing urbanisation, this will lead to growing competition over water (see Section 3.2).Reduced water availability can reduce a river's dilution capacity, impacting water quality.As a result, water may become so polluted it cannot be used to irrigate crops, instead becoming a source of pre-harvest microbiological contamination (Ijabadeniyi & Buys, 2012).Water quality concerns are tightly linked to food safety and can evolve into reputational water risks (see Section 3.2).For fruit producers, "food safety is non-negotiable," as underlined by a fresh produce buyer for a large retailer.GlobalGAP is the minimum certification that all producers must fulfil to export from South Africa, and within which conformity to water quality standards is a "major must" (GlobalGAP, 2018).Therefore, as pointed out by one interviewee, for export fruit farmers: "One of the biggest threats that you can have … is when your water becomes so dirty that you cannot export anymore." Much pollution originates from failing municipal wastewater treatment plants or unserviced informal settlements.As a result, stormwater, greywater, blackwater and solid waste can mix in a toxic cocktail that drains into rivers (Fell, 2017).The expansion of informal settlements, however, is often intimately linked to a growth of the fruit industry, driven by increased demand from global markets and enabled by the mobilisation of additional water resources (Visser & Ferrer, 2015).An NGO representative explained this relationship: [Increased water storage] enabled [farmers] to grow a lot more fruit, and that drew in a lot more labour.The labour was seasonal … came with their families and so informal settlements developed … And those informal settlements are upstream from [farmers'] water intake.And then all the sanitation, rubbish and problems came down from there and are flowing into their water, which then has to go onto very high valued crops for the international market.
Municipalities in South Africa often lack capacity, are chronically understaffed and cash strapped (Debbané, 2013), so they struggle to maintain infrastructure and services for rapidly growing communities.As a result, farms are increasingly facing risks of E. coli contamination, which poses a risk to food safety and human populations (Iwu et al., 2021) and requires greater surveillance of on-farm practices via private standards such as GlobalGAP.We can conclude that the fruit industry is not only exposed to water quality risks but also plays a part in driving their formation, linking the dynamics of the global economy to local water resources management.

| Reputational water risks
Reputational water risks occur because of the real or perceived negative social and ecological impacts of producers' water use (Hepworth & Orr, 2013).This can affect the brand value and market share of individual firms within fruit value chains.With the rise of social media and the 24-hour news cycle, scrutiny of firms has particularly increased (Lund-Thomsen & Lindgreen, 2018).Such reputational water risks do not need to be accompanied by legal proceedings or real environmental impacts to have long-lasting implications for companies (Hepworth & Orr, 2013).In the Western Cape, fruit producers are confronted with both downstream and upstream reputational water risks.
Downstream reputational water risks relate to buyers' and retailers' perceptions of water practices in the Western Cape fruit industry.Poor water quality, in particular, can lead to downstream reputational water risks if the media reports cases of polluted irrigation water.Because of the link between water quality and food safety, lead firms, who often carry the brand value, are sensitive to such media reports which can cause unease among consumers in the global North (Hepworth & Orr, 2013).For producers, therefore, the translation of water quality issues into a reputational water risk is serious.For example, in 2014, a local newspaper (news24, 2014) reported on high levels of E. coli in the Palmiet River, which flows through a major fruit production cluster in the Western Cape.While most farmers receive their water through closed pipes from a dam above human activity and therefore remained unaffected, a local farmer described the resulting reputational risk: It was totally misreported in the newspaper.And we nearly lost very big export orders.This is all below the dam [the E. coli].And so yes, it can affect about 8 or 10 farmers out of 160.The other 150 farmers get water pumped out of the dam, which is pristine.But you know what newspapers write, interpretation goes … The explanation highlights how reputational water risks can take on a life of their own, not necessarily linked to real material environmental impacts.Microbiological contamination of irrigation water is, therefore, a constant water risk to fruit producers that needs perpetual management.Downstream reputational water risks can have long-lasting economic impacts characterised by loss of market share and financial costs (Hepworth & Orr, 2013).
In contrast, upstream reputational water risks refer to the negative perceptions held by surrounding communities of fruit producers' water use.This can threaten producers' social or political licence to operate (Baleta & Winter, 2017).In the Western Cape, on average, 68% of available water is used for irrigated agriculture, making it the largest water user (Midgley et al., 2016).In situations of water stress, competition over water increases and having the lion-share of water allocations can become a reputational risk.Such negative imagery can lead to added regulation of agricultural water use, which is a particular worry for fruit farmers.These concerns must be understood against the political economy backdrop of the Western Cape, which combines the pressures of population growth, increased urbanisation, and the need to redress apartheid-era injustices in access to water resources (Sinclair-Smith & Winter, 2018).As explained by one producer: So, in the context of more and more people flocking to the cities, the problem isn't going to get easier for agriculture, it's going to get worse.As I say, water is at a very critical path.Without water, we can't have a business, and without water, people can't survive.
Alongside climate change, these political-economic dynamics are set to intensify pressures on water resources and increase existing water deficits.In such a future, the largest water users -that is, the fruit industry -will come under increased scrutiny.Maintaining a good reputation as responsible water users becomes crucial in this scenario, as articulated by one farmer: "we've got to be seen to do the right thing."

| Regulatory and political water risks
Under apartheid, water law and policy aimed to serve the benefits of the ruling White minority (Madigele, 2018).Many fruit farms in the study area were established under this system of privilege.As explained by one producer: "In the old South Africa… we had a bit of a nanny state when it came to farming.Everything was taken care of."This included subsidies and technical support from the government to develop irrigation schemes and gain access to water resources, giving producers a high degree of security in their water allocation.When South Africa transitioned to multi-racial democracy in 1994, the new government withdrew much of this support, translating into a loss of privilege and influence over water policy.One farmer explained how many fruit producers feel that now "we are on our own," receiving much less support and "certainly no sympathy from the government."The change in political orientation, therefore, has been perceived by some Western Cape fruit growers as undermining their water security for productive activities.
This perception of water risks is compounded by a lack of human and financial resources in the South African water sector (SAWC, 2017).For fruit producers, this poses a regulatory water risk because there are often serious backlogs in issuing water licences and permits for the construction of on-farm storage, as explained by one producer: "[I]t took me three years to get my water registered properly, and I've got a very simple registration." Public sector resourcing and resultant capacity gaps are deeply problematic in the South African context.While legislation is robust on paper, process delivery and enforcement, whether in relation to labour standards, biodiversity management or water management, is slow and often ineffective (Alford et al., 2021;Visser & Ferrer, 2015).The long waiting times and delays in obtaining water licences can stall production, posing a regulatory water risk that can have serious financial implications.For export-oriented producers, this backlog is problematic when they need to provide proof of water licences and permits during audits for standards that grant them access to overseas markets.One fruit producer explained this: Yes, because for the audits, they always ask me about the registration of the dams.… The dams we can't [register]…you get to a certain point and from thereon you can't go further.
In other situations, the combination of loss of privilege and the public sector capacity gap has led to irregular water use by farmers, as explained by one respondent: You want to dig a borehole, you need to get permission, but if there's nobody in that office, or nobody answers the phone … you eventually say, I can do what I like.So, you put the borehole down, which allows you to convert more virgin land to agriculture, and there's another law that says that you need a permit to do that, but again, there's nobody in the office.A tremendous amount of stuff happened, I'll call it illegally because I think the government had a duty to have someone in the office, but the farmer had a duty to go so far.
Such rogue water use further puts pressure on already stretched water resources.Overall, many fruit producers have "got more arable land available than we have water," as explained by one interviewee; hence, water remains the limiting factor for fruit production in the Western Cape.

| CONCLUSION
Building on GPN 2.0, we develop a water risk framework that we back up with empirical data from the Western Cape horticulture industry.We present evidence of emergent water risks that have become apparent not only because of ecological features but also due to their convergence with political and economic dynamics over the past 25 years.Such dynamics include the growing material effects of climate change, democratisation, neoliberal market restructuring and trade liberalisation, and the rise of ethical consumerism.We show how these processes converge and touch down in the specific local context of the Western Cape.Consequently, we argue that water risks should not merely be understood as extreme physical events but must be recognised as more subtle ongoing and underlying processes.This is a pertinent area of study, as water risks become increasingly important in global terms.It is significant for policy and practice because it shifts our perspective on how, and by whom, emerging water risks need to be addressed.Having categorised water risks as physical, reputational and regulatory/political, we demonstrate how such underlying and ongoing water risks are relational, often reinforcing each other.While water risks predominantly originate beyond the farm gate, they are actor and context specific, meaning that here they manifest as they do for growers because of the specificities of the Western Cape's political economy and ecological dynamics, as well as the export orientation of the horticultural industry.
The manuscript presents a discrete effort to move the environmental dimension more centrally into the study of GPNs.It shows how the production (and ultimately consumption) of fruit enmeshes not only with global governance regimes (i.e., private standards) but also with place-specific environmental governance regimes.Importantly, thinking through environmental risks in this way can open up possibilities for analysing how such risks are mitigated, by whom, and with what sort of outcomes (Lanari et al., 2021).Presumably, lead firms and other GPN actors have an interest in mitigating risks that threaten their ability to produce agricultural goods.Such risk-mitigating strategies will undoubtedly impact local environmental governance regimes and thus other water users, largely located outside of GPNs.Because water is a flow resource that does not respect any boundaries and connects a range of geographical, institutional and political scales (Norman et al., 2015), it is paramount that any risk-mitigating strategies within the value chain do not negatively impact those water users located outside of it.This points to the imperative of looking beyond the linear value chain when evaluating the impacts of globalised agricultural production, distribution and consumption on sustainability in the context of fast-moving environmental change.

ACKNOWLEDGEMENTS
Nora Lanari would like to acknowledge Coventry University for providing a PhD studentship which enabled this research to be undertaken.She would also like to acknowledge Swiss National Science Foundation's Early Postdoc Mobility fellowship, during which this article was written.David Bek would like to acknowledge a UKRI QR Strategic Priorities Fund grant provided via Coventry University which supported research underpinning this paper.Both authors would like to thank the respondents for agreeing to take part in the research and Zaitun Rosenberg and Mireille Lewarne for assisting in the conduct of the research.We would also like to thank helpful comments by the anonymous reviewers and the editor.Open access publishing facilitated by Victoria University of Wellington, as part of the Wiley -Victoria University of Wellington agreement via the Council of Australian University Librarians.

DATA AVAILABILITY STATEMENT
Research data are not shared.ORCID Nora Lanari https://orcid.org/0000-0002-1636-7072David Bek https://orcid.org/0000-0003-1503-3609 Different water risks in agricultural production networks (source: adapted from Baleta & Winter 2017; Hepworth & Orr 2013; Pegram 2010) T A B L E 1Physical water risksRelates to a shortage of water (scarcity), too much water (flooding), or water that is unfit for use (pollution), and impacts of climate change on water supply and demand Disruption of production; production cost may escalate; lower investor confidence