Transformational adaptation: agriculture and climate change

Additional keywords: adaptation, agriculture, climate change, transformation.

Climate change adaptation is widely accepted as a necessity and opportunity but remains an ambiguous topic and task. Numerous definitions and types exist. An emerging focus is on ‘transformational’ change as an adaptive response (e.g. Howden et al. 2007; O’Brien 2011; Pelling 2011). This paper examines what transformational adaptation means in the context of agriculture. It defines transformational climate change adaptation in agriculture as major, purposeful action undertaken at the farm or supra-farm level in response to potential or actual climate change impacts and opportunities in the context of other drivers.

A recent comprehensive definition of climate change adaptation highlights the many questions that adaptation raises about what adapts, what is adapted to, how adaptation occurs and how effective it is:

• ‘Adaptation involves changes in social-ecological systems in response to actual and expected impacts of climate change in the context of interacting nonclimatic changes. Adaptation strategies and actions can range from short-term coping to longer-term, deeper transformations, aims to meet more than climate change goals alone, and may or may not succeed in moderating harm or exploiting beneficial opportunities’ (Moser and Ekstrom 2010).

In agriculture in Australia and elsewhere, research and action to date has been more in keeping with the Intergovernmental Panel on Climate Change definition, which frames adaptation as a question of ‘adjustment’ (Parry et al. 2007). Yet, as the definition above usefully highlights, adaptation also encompasses more radical change: ‘longer-term, deeper transformations’. Such transformational change is increasingly on the agenda both by necessity and design (e.g. Cork et al. 2010; O’Brien 2011; Pelling 2011; WGBU 2011). For example, the Stern report in the UK highlights that climate change presents the need and opportunity for ‘major, non-marginal’ change (Stern 2007: 1). In agriculture, the scale of the climate change challenge and the growing influence of outside interests mean that transformation is both especially likely and desired. Transformational adaptation in and of agriculture was recently recognised as a priority by the Australian Primary Industries National Adaptation Research Plan (http://piarn.org.au/about-piarn/national-adaptation-research-plan, accessed 30 March 2012) and is an emerging field of research (e.g. Walker et al. 2009; Park et al. 2012).

This paper provides an introductory overview of transformational adaptation (TA) to climate change in agriculture. Given the nascent stage of thinking on the topic and space limitations, it is necessarily conceptual and exploratory. It consists of three main parts. First, the question of what TA is in relation to agriculture is discussed, highlighting the multiple dimensions involved, the relative nature of the concept and the need to think about both in situ and migratory change in land use. Second, two historical examples of transformation in agriculture are reviewed to highlight agriculture’s experience with transformation and illustrate different types of transformation. Third, four inter-related issues of climate change adaptation of particular pertinence to TA of agriculture are discussed before conclusions and implications for further research are presented.

A range of biophysical and social drivers are contributing to the rising interest in TA. These include how climate change adaptation is intellectually framed. Two major frameworks focus on need and opportunity, respectively. In terms of need, adaptation is typically conceived as a process of ‘fitting to’ an altered set of external circumstances, reflecting the bio-physical origins of much of the early literature on adaptation. Illustrated by the growing literature on ‘limits’, ‘tipping points’ and ‘thresholds’ (e.g. Russill and Nyssa 2009; Kwadijk et al. 2010; McNeall et al. 2011), this interpretation of adaptation has reminded society of humans’ biological nature and the climatic, ecological and other natural limits we exist within. Such limits and the potential for major changes in socio-natural systems have been investigated over many years in numerous fields including disturbance ecology and disaster risk management. But it is the growing evidence of the potential for substantial and/or rapid changes in future climate as a result of anthropogenic climate change that has most powerfully underlined that transformational change in human systems may be necessary. As the projected degree and pace of climate change accelerates (Rahmstorf et al. 2007; New et al. 2011) alongside the exacerbation of other biophysical-based limits such as less land, water, fossil fuels, phosphorus and biodiversity and rising demand for agricultural products, the likely need for more systemic, powerful adaptation of agriculture is increasingly being realised (e.g. Alexandra and Riddington 2007; Howden et al. 2007; Campbell 2008; Brussaard et al. 2010; Cork 2010a; Horlings and Marsden 2011; Sounness 2011). The aim of such transformational change is to maintain sets of activities, products (sensu lato), values and processes in the sector or region. In particular, transformative adaptators either seek to: (1) pro-actively avoid, uncertain and severe challenges in order to avoid transformational change of a more involuntary, uncontrolled and negative character; or (2) be ‘first movers’ so as to take full advantage of what they see as emerging opportunities (Howden et al. 2007; Park et al. 2012).

While the above is an important interpretation and driver of transformational change in agriculture, it does not capture the full situation. Reflecting the coupled nature of social-ecological systems such as agriculture, adaptation needs to be viewed not just as ‘fitting to’ biophysical limits of the sort discussed above, but ‘fitting with’: a process of co-evolving with one’s system (Collins and Ison 2009a, 2009b; Ison 2010). This alternative framework – which in itself represents a mental adaptation – is particularly suitable for perceiving and conceiving of transformational change for three reasons. First, it better acknowledges ongoing ‘background’ change in the environment we are part of. Here, transformational change is a recognisable shift in the type of change occurring rather than the introduction of change to a stable setting. Second, in this frame, humans have a greater degree of agency over what can be influenced and changed, including the climate itself. A co-evolutionary perspective allows for a proactive as well as reactive approach to adaptation and broadens the scope of what can be changed, reframing accepted ‘inevitabilities’ such as infrastructure limitations, economic arrangements, agricultural policies or mental frameworks as a question of choice. This reframing helps move the debate from a negative perspective (cost-avoidance) into a potentially positive perspective (increased environmental or economic sustainability and opportunity) (Howden et al. 2007). As well as reducing harm, TA may be pursued as a way of creating new and better systems, consistent with the move in adaptation circles towards low-regrets strategies (where benefits exist regardless of whether projected impacts eventuate) and win-win options (in which all stakeholders benefit) (UNDP 2009). Adaptation in general is increasingly understood as a response to multiple drivers and a source of multiple benefits (e.g. Pittock 2009) and this is especially true of TA (Pelling 2011; Park et al. 2012).

A target of low-regrets strategies is to address the long-standing ecological, social and economic vulnerabilities that climate change exposes, such as soil degradation, biodiversity losses, dependence on declining oil reserves, low financial equity, and an aging farmer population (Green et al. 2003; Anderies et al. 2006; Walker and Salt 2006; Campbell 2008; Fresco 2009; PHAA 2009; Robinson 2009; Cribb 2010; Fedoroff et al. 2010). Research on agriculture’s ‘social licence’ to operate suggests that there is a related growing pressure on agriculture to demonstrate it is a legitimate and preferred user of a land, water, biodiversity and atmospheric base progressively strained by climate change (Martin and Williams 2011) where there are alternative land uses such as for biofuel production, carbon sequestration, biodiversity conservation and other ecosystem services. This includes political and social pressure on agriculture to reduce greenhouse gas emissions. This brings us to the fourth advantage of a co-evolutionary interpretation of adaptation, which is that, like others, farmers need to adapt to the full suite of indirect and direct impacts of climate change, including the full range of its social, cultural, political, financial and physical effects, as well as other concurrent situations and threats. As a response to and source of significant flow-on effects, transformational change requires a perspective that accommodates this breadth and dynamism, as a co-evolutionary perspective does.

Given that TA is being and increasingly will be demanded of agriculture in the face of climate change, work is needed to clarify the concept. This requires accepting its heuristic, subjective and relative character. Different perspectives on what counts as transformational change exist and are not reducible to one another. This is for two reasons. First of all, there are many dimensions over which TA to climate change can be conceived (Table 1). While a large spatial scale is one possible dimension, TA can also be thought of differently in relation to smaller scales, from an individual to a family to a community, or from a single farm business or property, to a sector or region.

Table 1.  Possible characteristics of transformational adaptation (TA) across different dimensions of change at the farm and regional scales

Second, across any one dimension, the level of change that counts as ‘transformational’ is subjective and relative. It is relative in two ways. First it is a question of the difference between the deliberate versus background change. Climate change adaptation is essentially about ‘persistence through change’, which raises the questions of what persists (subject to pre-existing trajectories or variability) and what changes (in a desired, deliberate or obvious way). In TA, this ratio between persistence and change is low; more of the system is changed than is continued as is. Second, this ratio is relative to less transformational forms of change. Adaptation as adjustment (incremental adaptation) infers that the ideal ‘ratio’ of what remains relatively constant to what is deliberately changed is high. Most of the existing system remains on its pre-existing trajectory. Park et al. (2012) note that the central aim of incremental adaptations is to maintain ‘the essence and integrity of an incumbent system or process at a given scale’ (p. 5).

This second element of relativity is emphasised in the diagram by Howden et al. (2010) (Fig. 1) that brought the concept of TA to prominence in the agricultural sector. It presents ‘transformational adaptation’ as the more radical end of a spectrum of change that begins with incremental adaptation (changes in practices and technologies within an existing system such as planting times or row spacing) and extends through systems adaptation (changes to an existing system, such as major change in the focus between livestock and cropping, new crop types or the adoption of precision agriculture), mapped against an increasing degree of climate change.

Fig. 1.  Levels of adaptation in relation to benefits from adaptation actions and degree of climate change, with illustrative examples (from Howden et al. 2010).

The third point to note about TA is that it does not exist in a vacuum: the three types of adaptation in Fig. 1 overlap conceptually and in practice. The adaptation being pursued in any situation depends in part on the type of decision being faced. For some decisions (e.g. the amount of fertiliser to apply to a crop), reducing the decision horizon will likely always be the more rational way of proceeding. For other decisions (e.g. where to purchase a property), TA will be needed. Here, incremental adaptation alone may act as a blockage for necessary change by increasing investment in the existing system or locale and narrowing down alternatives for change: what the resilience, transition and policy literatures refer to as ‘lock in trap’, ‘incrementalism’ and ‘negative resilience’ (e.g. Handmer and Dovers 1996; Allison and Hobbs 2004; Anderies et al. 2006).

While the need for some degree of TA will likely become more indispensable and involuntary as climate change increases, it is not always necessary or desirable to wait for such pressures to accumulate. TA is already being pursued earlier to capture various benefits, address pre-existing structural issues, pre-empt possible problems or to accommodate the long lead times on associated decisions and actions (Howden et al. 2010; Park et al. 2012). TA will also always need to be accompanied by more incremental forms of change to allow for adjustments to fine-tune systems to ongoing change. The earlier decision makers act the more uncertainty and variability needs to be accommodated. This requires making decisions ‘robust’ across a range of possible future conditions future (Wilby and Dessai 2010; Stafford Smith et al. 2011) rather than focusing too narrowly on one climatic signal such as dryness. It also requires building in capacity for monitoring, learning and adjustment (Howden et al. 2007). If robustness and flexibility are taken seriously, deliberately acting early on some dimensions to alter the system (e.g. addressing pre-existing vulnerabilities, or adaptation to climate change) is likely to allow for a greater range of options and more influence over circumstances than non-action (Smit et al. 1996). How the different types of adaptation relate to each other in practice is a topic of ongoing research (Park et al. 2012). Depending on how TA is defined, incremental adaptations may accumulate on one farm or across a region to create or trigger a transformational change. Park et al. (2012) provide examples from the wine industry about how incremental and TA by actors operating at different scales interact, with transformational change at either the farm scale or sectoral scale enabling incremental adaptation at the opposite scale.

Further insight into TA in agriculture can be gained by considering two illustrative and potentially influential cases:

1. Changes in goal (for example resulting in a major change in land use and/or employment through trying to do more or something different); and/or

2. Changes in location (of an agricultural activity and/or agriculturalists) (Fig. 2).

Both strategies are directed at reducing adaptors’ vulnerability to climate change impacts (broadly defined), where vulnerability is understood as a combination of sensitivity and exposure to impacts and capacity to adapt (Smit et al. 1996). The main difference between them is their relative focus on sensitivity and exposure. Where the anticipated change is negative, an in situ change in land use or occupation primarily aims to reduce the adaptors’ sensitivity to impacts by shifting to a less climate-sensitive mode of operating, while a spatial relocation primarily aims to reduce the adaptors’ exposure to impacts by seeking out a new area anticipated to be more amenable to the continuance of their original activity or occupation. These transformational processes of replacement and displacement may result in the maintenance of agricultural activity, even if there is a shift from food- to non-food-based agriculture in some places. They may also involve a shift out of agriculture, particularly when combined. In contrast, if the perception of climate change is net positive, then the change in land use or location could aim to increase the sensitivity of the operation to capitalise on the anticipated benefits.

The choices involved in such changes highlight the importance of willingness to adapt as well as capacity to adapt (Moser and Ekstrom 2010). The strategies contrast those farmers who are strongly attached to their land and will go to great lengths (including working off-farm, planting it with Blue Gums etc.) to maintain ownership of it, and those who are strongly attached to their farming occupation (and perhaps particular enterprise types such as livestock or cropping) but do not care deeply about where they practice it. In this and other ways, TA intersects with identity issues within farming (e.g. see Marshall 2010), including the different ideals of agricultural masculinity dominant in different settings (e.g. Ferrell 2011).

Changes in goal and location presuppose and intersect with other adaptations and broader change processes, both in the landscape and socio-economically, technically and politically. This includes adaptations at the sub-farm level, which are defined here as system changes and not considered in detail. That said, a farmer purchasing an additional property in a new area in order to avoid projected climate impacts in the original area (i.e. partially relocating) could count as transformation, as would substituting a significant amount (not necessarily all) of their agricultural work with off-farm employment. Evidence is emerging of both of these adaptation strategies at work, some in explicit response to climate change (e.g. Park et al. 2012; Rickards 2012). Like more definitive changes in goal or location, these adaptation processes at the farm level may aggregate over time and space to create large-scale change. Such a high level view is now taken to describe major transformations within agriculture in the past.

Not only are agriculturalists familiar with the idea of and need for climate adaptation (Kandlikar and Risbey 2000; Howden and Stokes 2009), but so too are they familiar with major transformations (Lobao and Meyer 2001; Barr 2009). These past and ongoing transformations boost the slim evidence base for transformational climate change adaptation by providing temporal and spatial analogues. As for climate change adaptation in general (Ford et al. 2010; McLeman and Hunter 2010), these can be used to gain insights into change processes and identify their possible characteristics and constraints.

At a broad scale, an analogue of the first form of transformational change in agriculture identified above – a change in goal – is provided by the emergence of different eras of agriculture. At a very coarse scale, there was first a shift from traditional agriculture to ‘modern’ science and market-based agriculture, substituting at least at the sectoral level subsistence-based goals and values with ones associated with productivity and professionalism (Rickards 2006). More recently, ‘productivist’ agriculture has been replaced in part by ‘multifunctional’ or ‘post-productivist’ agriculture in Britain and the European Union through alterations to the Common Agricultural Policy arising from changing societal values (Potter and Tilzey 2005; Wilson 2007, 2008). Seen to a far lesser extent in Australia, this shift represents the formal addition of goals associated with land stewardship, biodiversity conservation and ecosystem services to those to do with productivity (Cocklin et al. 2006). This insertion of environmental sustainability goals and policies into the agricultural arena is likely to increase under climate change, especially in relation to the mitigation and sequestration imperative that agriculture increasingly seems expected to respond to (Bebbington et al. 2011). More generally, the farm business has been decoupled to a degree from farm production per se as the goal of increasing productivity has been replaced by that of profitability. This has led to a range of diversifications of farm income, in the most extreme cases involving a complete replacement of farm production income and an exit from farming (in most cases involving a concomitant shift in location for the farm members). As growing numbers of farmers have taken the latter option, farmer numbers have declined significantly in Australia and elsewhere. The resultant ‘Great Transformation’ of agriculture, as Lobao and Meyer (2001) call it in relation to the United States, provides a further analogue of potential TA under climate change, not just as a generic transformational change but as one that is likely to specifically feature among climate change adaptations.

‘British’ agriculture in its increasingly scientific form was exported to Australia with colonialism in the eighteenth and nineteenth centuries. Indeed, expansion of agricultural land was a key tool of British colonialisation of the ‘empty landscape’ of Terra Nullius (Anderson 2003). This relocation or expansion of sedentary agriculture provides an analogue of the second form of transformational change in agriculture identified above and can illustrate significant lessons about such changes in location. First, in contrast to the frontier mentality that accompanied the clearing of land for agriculture, we now know that rather than inserting agriculture into a vacuum, the process was one of replacing existing land uses, people and values, including indigenous ‘fire stick’ farming (Jones 1969). The same holds for any relocation of agriculture today: it will always be in replacement of another set of attributes, values and goals, and thus necessitates careful consideration of costs, trade-offs and power issues. In this way, the processes of displacement and replacement are intricately linked. They are also linked to the extent that no agricultural system can be ‘imported’ to a new location wholesale without significant modifications. Pointing to the interaction of the TA process of relocation with incremental and systems adaptations, farmers and extension agents are acutely aware of the need to adjust generic advice to the local biophysical context (Pannell et al. 2006). While the point of relocating agriculture under climate change is to better match an existing agricultural system to a local climate elsewhere, this match will not be perfect, the climate will change in the new location also, and numerous other biophysical factors will be different from the original location (Park et al. 2012), requiring in situ changes to the agricultural system in its new location, as illustrated in Fig. 2 above. Understanding how the ‘new’ agricultural system interacts with the local environment is essential for identifying, avoiding and managing negative externalities such as habitat loss and water pollution, which is another critical lesson that has emerged from the importation of British agriculture into the Australian context.

Fig. 2.  Schematic diagram of the two main forms of transformational adaptation in agriculture: (a) a major in situ change; (b) a change in spatial location, with minor accompanying in situ changes.

The above discussion of some major past transformational changes in agriculture illustrate that such change can involve serious challenges, risks and benefits. It also reinforces that climate change and adaptation are not happening in isolation; there is already a complex collection of change processes in play. Climate change impacts, adaptation efforts, and the impacts of adaptation itself will interact with these processes, and will likely exacerbate the complexity and uncertainty involved.

In this section, we turn to the climate change adaptation literature to present four general climate change adaptation issues of particular pertinence to TA in agriculture. Bridging literatures in this way is important for facilitating interdisciplinary, social learning of the sort needed to respond to climate change (Collins and Ison 2009a; Ison 2010). These issues – costs, maladaptation, adaptive capacity, and role of government – are discussed in turn below.

Common to all change, adaptation to climate change can involve some significant costs. Five are considered here. First are transaction costs, which is the toll on resources (mental, emotional, physical, financial, social etc) that the process of change exacts. Transaction costs are largely independent of the positive value or rationality of the change; a point that agricultural extension has come to realise following decades of frustration at farmers’ seemingly illogical recalcitrance (Pannell et al. 2006). For a farmer, even learning about and considering the applicability of a potential change takes precious time and energy (Pannell et al. 2006). Climate change adaptations can involve particularly significant transaction costs, in part because of the uncertainty surrounding what adaptation is needed in response to what potential changes, what the options are and what is most appropriate: the contingency cost can be high. Transaction costs and risks can also be high because the expected benefits of action may not be realised for a long time, or even ever, and the actions can contest with existing norms and values (Howden et al. 2008; Orlove 2009). In the case of TA, the retarding influence of uncertainty is greatly exacerbated as the longer time scale, larger spatial and social scale, and greater magnitude of action increases the degree of uncertainty and the deliberation needed. As Stafford Smith et al. (2011) discuss, adaptation decisions with long life times (long lasting effects) usually also have long lead times.

The second type of cost to note is opportunity cost. The path dependency of adaptation decisions and especially TA decisions means that this form of cost is also particularly significant in the context of climate change. Path dependency refers to the way in which decisions taken now – particularly transformative ones – may constrain the options available at a later point (Inderberg and Eikeland 2009). As the uncertainty surrounding conditions under climate change demands a flexible or robust approach in which options are kept open (Hallegatte 2009; Wilby and Dessai 2010), this path dependency is especially problematic and requires that strenuous efforts are made to reduce ‘lock in’ (Allison and Hobbs 2004) and other opportunity costs. TA also highlights the potential for positive change and wider benefits under the guise of adaptation. A further opportunity cost is that which emerges when action is not taken to seize such potential. This appears to be a key motivation for current transformative climate change adaptations in the wine, peanut and rice industries (Park et al. 2012).

Opportunity costs can be accepted consciously or unconsciously. The latter unintended consequences are the third form of cost of relevance to climate change adaptation: those costs arising out of the unexpected, often long-term and long distance risks created by our actions, epitomised by climate change itself (Beck 2006) that can be systematically hidden from view if a narrow conceptualisation of value is used in economic framing (Costanza 2001). As Eriksen et al. (2011) note, ‘not every response to climate change is a good one’ (p. 7). For example, a climate adaptation that increases greenhouse gas emissions would be difficult to justify as being internally consistent (Howden et al. 2007). TA is particularly vulnerable to creating unintended consequences because the complexity of change involved is beyond the limits of our scientific capacity to predict or trace, as illustrated in the cases of transformational change in agriculture above. That said, avoidance of the unintended consequences of inadequate action is also a major motivator for TA (see Handmer and Dovers 1996).

The fourth type of cost is that which arises out of the incompleteness of adaptation, as adaptation will never fully compensate for climate change losses where the change is negative (Easterling et al. 2007). There will therefore always be residual losses and in some cases enhanced gains. Some of these losses may be intended. Economists focussed on cost-efficiency, for example, advocate for a priori acceptance of a certain level of damage (Fankhauser 2010). The rationale is that there is an economically optimal level of adaptation, represented by where the cost of adaptation action intersects with the projected cost of climate change impacts at a particular point in time (the latter being the fifth type of cost to note). Although seriously marred by the ignorance within and static nature of the predictions on which the calculations are based (Hulme et al. 2009), this approach has popularised the notion of ‘over-adaptation’: the idea that it is possible to act too soon or too vigorously. TA is especially vulnerable to criticism of over-adaptation, given that the long lead time and opportunity-focus of much such adaptation can require acting in advance of a necessitating degree of climate change (Stafford Smith et al. 2011). To the extent that TA involves preventative action, it is further vulnerable to this criticism, as the very successfulness of prevention can diminish public awareness of the reality of the problem at which such action is directed, leading to assertions that such actions are ‘over the top’ (Zaalberg et al. 2009). Nevertheless, there are already examples of farmers making transformative changes in response to the climate trends that are already occurring (Park et al. 2012). Ongoing evaluation of the success of these will hopefully inform the discussions in relation to over-adaptation. Hindsight will also inform us whether the costs incurred by reactive, incremental adaptation justified an earlier, staged approach to more systemic and transformational change.

The upshot of the above discussion is that where climate change is negative, adaptation is likely to reduce the size of the losses rather than remove them. Decision makers therefore need to address the types, level and distribution of losses they want to aim for. While TA carries the risk of significant losses, it also has a unique potential to avert potentially greater losses. Rising calls for radical change and action (Bailey and Wilson 2009) indicate that, as discussed by (Jones 2010) and implied by Fig. 1 above, across sectors people are increasingly willing to take a risk in order to try to reduce the extent to which we are at risk under climate change. Agriculture is no exception.

Whether society will succeed in reducing the extent to which we are at risk from climate change is an open question. Adaptation efforts that fail in this way, or involve excessive costs in the process, are a form of maladaptation. In addition to how effective and efficient adaptation actions are, ‘good adaptation’ is increasingly assessed, at least in theory, according to other more searching criteria, notably: equity (distributional justice, including future generations); social legitimacy (procedural justice); and sustainability (including inter-species justice) (Adger et al. 2005; Barnett and O’Neill 2010; O’Brien et al. 2009; Orlove 2009; Eriksen and Brown 2011) as well as risk of increased greenhouse gas emissions which would operate as a positive feedback to climate change (Howden et al. 2007).

The systems perspective associated with TA is crucial for attempting to trace the consequences of potential and actual adaptation actions across time, space and social settings in order to try to avoid maladaptation for one’s self or others. The high order change that TA involves is also posited as a way of avoiding maladaptation, given the existing problems of inequity and unsustainability that currently exist and potentially may be exacerbated by climate change (Leichenko and O’Brien 2008). It is also in response to concerns about the perceived inadequacy of current proposed adaptation and mitigation actions that calls for TA have been initiated. Calls for consideration of TA are escalating, whether based on the ongoing failure to initiate effective mitigation (e.g. Ayers and Huq 2009), a concern about the long-term ineffectiveness of specific adaptation approaches (e.g. Stokes and Howden 2010), or a desire to address pre-climate change vulnerability (e.g. O’Brien and Wolf 2010).

Of particular pertinence to agriculture is debate about how adaptive the existing responses to climate variability are in the long term. Responding to the impacts of climatic extremes and natural disasters is unavoidable and therefore a stepping stone to longer term adaptation (e.g. McKeon et al. 1993). Failure to acknowledge this can lead to unrealistic and callous expectations of those recovering from disasters (McKeon et al. 2004). Recovery processes can also involve many of the same approaches needed for climate change adaptation, particularly when the climatic extreme involved is aligned with changes in average conditions, as is the case for drought in Australia (Howden et al. 2010). Yet, research also indicates that responses to variability may in the longer term prove maladaptive. In a study of graziers in Queensland, for example, Marshall (2010) concludes that: ‘Enhanced strategies for coping with climate variability will provide a way for encouraging gradual, incremental adjustments for climate adaptation’ but may also lead to over-confidence among producers and ‘in fact, make them vulnerable to more extreme and frequent climate events predicted for the future (p. 40). Importantly, the consequences of these extreme events may be effectively irreversible and thus reactive, incremental adaptation based on historical experience of extremes and post-hoc learning will potential degrade future production options (McKeon et al. 2004). Similarly Handmer and Dovers (1996) stress that ‘change at the margins’ can create a false sense of security and delay needed transformational change and (Davies 2009) argues that perpetuating a state of mere ‘coping’ is no substitute for genuine adaptation. Overall, the relationship between adaptation to climate variability and climate change is complex, highly pertinent to agriculture, and in need of in-depth research.

TA requires a higher level of adaptive capacity than incremental or systems-scale adaptation due to the greater risks and complexity arising from the change. To the extent that it can involve collective, voluntary action (a common mode of change in agriculture), it also requires an especially high level of willingness to adapt, which is determined in part by the potential benefits to be gained and the perceived legitimacy of the adaptation process. In many cases TA also requires substantial resources.

The attributes required for transformation may be somewhat different from those required for incremental adaptation. Cork (2010b) lists the characteristics of transformability as: experimentation; support for change; trust (social capital); human, built and natural capital; and cross-scale awareness. These characteristics are important in incremental or systems adaptations as well as in change processes more generally, but it could be argued that they are more crucial in transformative climate adaptation. More generally, the five types of capital identified by Nelson et al. (2010a, 2010b) – human capital (e.g. health, education), social capital (within-group and between-group connectedness), financial capital (assets, cash), physical capital (infrastructure, technology) and natural capital (soil, water, biodiversity) – are needed for the perpetuation and improvement of rural livelihoods in general and adaptation to climate change in particular. Empirical research on farming families’ decision making experiencing drought and other pressures (Greenhill et al. 2009; King et al. 2009; Rickards 2012) suggests that the accumulative effects of climate change and other issues erode their net level of these ‘capitals’ or assets and may encourage maladaptive substitution between them (e.g. exploiting one’s human health and energy to substitute for a lack of financial capital by taking on more work). At the level of human capital, transformative learning research with producers suggests that critical reflection on the relationship between practices, the environment, roles, responsibilities and social norms can lead to the kind of transformative change needed for major adaptation to climate change (Tarnoczi 2011). A wealth of research also highlights the importance of social and human capital in enabling the sort of positive, enduring, collective action TA will involve (e.g. Meert et al. 2005; Gooch and Rigano 2010; Burton and Paragahawewa 2011).

The final issue to highlight in relation to TA in agriculture is the role of government. This is shaped by broader debates about the role of government both in relation to climate change mitigation and adaptation, agriculture and society more generally. In terms of the former, it is acknowledged that climate change can be framed as an unprecedented market failure (Stern 2007), and therefore legitimates government intervention. Yet, the form this intervention could most effectively take is less certain, especially in relation to adaptation which is often misconceived as a mere local-level, private-good concern that will happen autonomously (Nelson et al. 2011). Sensitivity to the private benefits of adaptation is enhanced in Australian agriculture, where since the 1970s there has been a move to more precisely distinguish private good from public good activities in agriculture in order to help reduce and justify government expenditure in the sector (Simpson and Dargusch 2010), including in relation to climate-related disasters (Botterill 2003). While this stance remains strongly influential, climate change and associated food security concerns are prompting growing calls to redress some of the consequent reduction in government investment in agricultural research, development and extension (RDE), both in Australia and beyond (e.g. Cribb 2010; PMSEIC 2010). The case for such investment seems particularly strong in support of transformational adaptation, with Pannell (2010) for example arguing that government RDE investment should be focussed on supporting ‘decisions that are larger and indivisible, such as land purchase or the decision to exit from agriculture’ (unpaginated). Nelson et al. (2011) argue that there is a key role of government in identifying and implementing policies that foster pro-active and anticipatory adaptation especially addressing market failures such as moral hazard and adverse selection of risk management approaches.

The appropriate type as well as quantum of agricultural RDE is needed to support TA in agriculture. Information is needed that provides flexible options for farmers and others to use, monitor and evaluate. To be effective, particularly at a transformational level, it is crucial that the underpinning research is cross scale, cross sectoral and integrated with mitigation (Nelson 2009), the latter being a major policy and societal shift that agriculture needs to adapt to. Research to support transformational change therefore needs to be interdisciplinary and engage with relevant insights from diverse areas, including futures thinking (Inayatullah 2008), innovation studies and socio-technical transitions studies (e.g. Geels and Schot 2007; Dolata 2009; Leggewie and Welzer 2010), rural and political geography (e.g. Gibson et al. 2010), resilience thinking (e.g. Cork 2010b; Gelcich et al. 2010), and the mainstream climate risk management literature, including that on migration and climate change adaptation (e.g. Black et al. 2011). As pointed out elsewhere, conventional insular agricultural research is increasingly inadequate in the face of growing complexity and uncertainty (see, for example, PMSEIC 2010). More than interdisciplinary, such research also needs to be transdisciplinary, which refers to the integration of non-academic knowledge through participatory processes and in developing systems of adaptive governance (Nelson et al. 2008). CSIRO, for example, is starting to work with farmers undertaking transformational changes to better understand and help assess the consequences of different transformation options, in part to avoid problems of the past and the issues of over- or maladaptation discussed above (Park et al. 2012). Policy makers, agricultural organisations, agribusiness, agricultural advisors and rural community groups are other important ‘end users’ and shapers of research into TA of and in agriculture. Effort is needed to bring these different perspectives together to discuss the cross-scale and cross-sectoral shifts that may eventuate under climate change and increase the social learning that is needed for appropriate adaptation.

TA is a crucial and increasingly relevant aspect of agriculture’s adaptation to climate change. It is also little studied and poorly understood. This paper has sought to provide an introductory overview of the topic in relation to agriculture. It has highlighted that it is a multidimensional, relative and critical concept, which can be understood more easily by focusing on two major types of change: changes in the goals and/or location, including the option for farmers of exiting the industry. Historical and ongoing examples of other transformational change processes in agriculture provide important insights into what TA under climate change may involve, and further research is required in this area. Further research is also needed to engage the agricultural research community with the broader learning emerging out of the mainstream climate change adaptation literature and other disciplines. TA reinforces the realisation that agricultural research can no longer remain insulated from off-farm, non-science or non-agricultural knowledge or processes. Just as support and guidance of TA requires that we understand how Australian agriculture is currently, and could be, positioned climatically, we need to understand its position within the landscape, rural communities, and broader social, political and cultural environment. It is only by building our capacity in adopting such a broad systems perspective that TA will avoid the maladaptations of the past and enable opportunities for positive change to be realised.