Abstract

Keywords

Introduction

Natural resource management (NRM) issues are characterised by multiple stakeholders and multiple goals. Because of this they cut across traditional boundaries; not only administrative and geophysical boundaries, but also disciplinary boundaries. One example of such NRM issues is catchment management. Catchment management requires the balancing of many different goals and interests; diverse land uses (agriculture, forestry, recreation, residential) and consideration of their different demands on the diverse resources described by a catchment (water, open space, high value agricultural land). Addressing these issues requires work across boundaries, and has led to the emergence of professionals whose practice is specifically about spanning boundaries. Both research and extension have become increasingly concerned with working across boundaries in the field of catchment management. In the next section we will outline some of the changes that have occurred in research and extension practice in recent years, before describing briefly a case study of extension and research working together in catchment management. We then turn our attention to the uncertainties inherent in complex systems that undermine ‘practice change’ directed towards better catchment management. The characteristic ways in which research and extension experience uncertainty provide insights into how the two professions might work together more effectively.

Extension, research and brokering practice

Extension is a profession concerned with communication, information exchange and promotion of learning in order to build capacity and change practice. While there is no single agreed definition of extension, a number of authors have described the range of types of extension that are undertaken. Black (2000) identified four main types of extension: linear ‘top-down’ transfer of technology; participatory ‘bottom-up’ approaches (also termed ‘group empowerment’); one-to-one advice or information exchange; and formal or structured education and training. More recently, Coutts and Roberts (2003) described five interacting models of extension in use in contemporary Australian extension practice. Significantly, ‘top down’ transfer of technology was not included. This highlights the many changes that have occurred in extension practice since the days when ‘transfer of technology’ was the dominant model. Black’s other three types of extension were identified, along with two additional models (technological development and information access). The ‘transfer of technology’ approach has been heavily criticised for the way in which it describes knowledge flowing in one direction only; from science (at the top) to farmers (at the bottom). Many of the key points of this critique are summarised in Vanclay’s (2004) social principles for agricultural extension. In particular, within contemporary models of extension, alternative knowledge systems to science (such as farmers’ local, experiential knowledge) are recognised as valid sources of ‘truth’. The diverse goals of different stakeholders are also recognised as valid; the goals implied within technological innovations are no longer prioritised to the extent they once were. A technical rationality (prioritising scientific knowledge) is replaced by communicative rationality (prioritising exchange between different knowledge systems; see Eshuis and Stuiver (2005) and Attwater et al. (2005) for examples of the prioritisation of knowledge exchange). Paine (2005) describes the unique role of extension as that of a ‘mediating practice’; a practice whose characteristic role is brokering exchange between different knowledge systems. A diverse range of practitioners play this extension role. These include not only government agency extension agents and their private sector counterparts working as farm consultants (mediating between farmers and research), but also players such as fertiliser sales agronomists, Landcare facilitators and committees, government agency assessors of farmer applications for environmental rehabilitation works funding, and industry funded NRM coordinators (who mediate between farmers and commercial, policy or community interest imperatives).

This shift from a technical to a communicative rationality fundamentally alters the relationship between extension and research. Extension practitioners (and society more generally) no longer look to science as the only source of reliable knowledge about the world. Science is attempting to redefine itself in a world where its knowledge claims are increasingly contested. As Lane et al. (2004) point out, an approach grounded in communicative rationality requires scientists to ‘qualify their unwavering faith in the power of science to help us with wicked and contested environmental decisions’. This is not an easy change to make; it challenges the central tenets of modernist science. However, a new model of research is emerging in some domains, where knowledge claims developed by science are tested and transformed through interaction with and challenges from other knowledge systems (Gibbons et al. 1994). Participatory (action) research methodologies include multiple forms of knowledge within research teams; notably farmers, extension professionals and biophysical researchers (Ridley et al. 2003). Different scientific disciplines are also starting to work together. Social and biophysical researchers are working together in many agricultural research organisations (Roughley 2005). The epistemological diversity within different branches of the same discipline is being recognised in the formation of transdisciplinary teams (Horlick-Jones and Sime 2004; Müller et al. 2005). Van Kerkhoff (2005) describes this new model of research as one of integration, ‘a move away from individualistic, discipline-driven science to utility-focused research that connects research activity across a number of boundaries’. In this paper we will suggest that some forms of research are playing a brokering role between knowledge types; similar to the brokering role that has previously been claimed as the unique role of extension. This is prompting a renegotiation of roles between research and extension. We will suggest that, although research seems to be encroaching on the brokering role of extension, there are actually characteristic differences between the way in which research and extension play this role that suggest possibilities for them to work together in a complementary rather than competitive way.

Case study methodology

Catchment management is one of the areas in which new brokering forms of research are emerging, because catchments are characterised by complexity and thus there is a need for multiple forms of knowledge even just to deal with their different biophysical characteristics (eg ecology, hydrology, soil chemistry). This paper will draw on a case study of research involvement in catchment management in the Tarago catchment in order to highlight some of the characteristics of a new brokering science practice and identify some of the implications that this may have for extension. Case study research is particularly appropriate where the investigator has little control over the set of events in which they are interested (Yin 1994); as is the case with an investigation of science practice in catchment management. The case study presented in this paper was not initially designed to examine the relationship between research and extension, but to look at ways in which research groups can better contribute to catchment management. However this case study provides an opportunity to examine the research – extension relationship in the arena of NRM for a number of reasons. Firstly, a research group from the Department of Primary Industries has been intensively involved with the management of this catchment for the past five years. This provides a tangible example of how integrative research is being practiced and how it is negotiating relationships with extension. Secondly, the Tarago catchment supplies water to the Tarago reservoir, which is owned by Melbourne Water for the purpose of metropolitan drinking water supply. The reservoir is currently used only to supply small towns adjacent to the reservoir, however it is intended that the reservoir will come back on line to Melbourne’s water supply in 2011. Because of the larger volumes of water and associated higher risks when water is supplied to a major metropolitan area, this provides a strong imperative for different stakeholders in the catchment to work together to improve water quality and necessitates exchanges between different groups (including research and extension).

Analysis of this case study draws on diverse sources of data, primarily participant observation of the everyday work of the research group conducted over the past three years by the primary author of this paper as part of her PhD studies. This ‘everyday work’ has included activities such as committee meetings and field days in the catchment, a program of trialling nutrient management decision support system (DSS) within the catchment, and team meetings. Informal discussions between the authors (Gourley and Melland are also members of the research team described in the case study) have also provided a source of data as well as occasions for reflection on the experience of negotiating research and extension roles in catchment management. This case study has been undertaken as action research, and Kincheloe’s (2005) description of the methodological bricolage is particularly apt in describing how data collection and analysis is approached in this methodology:

We actively construct our research methods from the tools at hand rather than passively receiving the "correct," universally applicable methodologies.

As involved participants in the management of the Tarago catchment, data collection and analysis was integrated with other activities and adapted over time.

Brokering practice in the Tarago catchment

Although this case study describes the involvement of a research group in catchment management, we argue that many of the implications of the paper relate to extension. The following description of the work of the research group in the Tarago catchment demonstrates some of the similarities and overlaps between research and extension practice. The research group comprises approximately 20 scientific, technical and administrative staff. Group members work on a range of projects that address issues around integrating productivity and environmental objectives in grazing systems. The group focuses particularly on nutrient management, and group members have expertise in soil science, hydrology and farming systems as well as skills in the development of tools to assist farmer decision making. Activities that the group members have undertaken in the catchment include participating in the development, implementation and monitoring of a catchment management plan, presenting at field days, and contributing to the design and delivery of a whole farm planning course. The group also trialled two nutrient management decision support systems with farmers in the Tarago catchment (Love et al. 2005). This involved group members visiting farmers and using these tools to make best practice recommendations for particular farms. We would suggest that there is not only a superficial similarity between research and extension related to the types of activities that are undertaken (field days and farm visits) but also that research is playing a knowledge brokering role similar to that played by extension. Face to face interactions between farmers and science involve exchanges between different knowledge systems; as scientists and farmers discuss for example which areas of the farm are most as risk of contributing sediment and nutrients to waterways. Similarly, the ongoing development of DSS tools requires researchers to draw on and integrate scientific knowledge from diverse disciplines about hydrological, chemical and physical relationships.

Although the presence of a water supply reservoir within the catchment and its likely use for metropolitan water supply increases the imperatives for management, improving water quality is neither the only nor the most important goal of many catchment stakeholders (although all agree that water quality is important). Approximately two thirds of the catchment’s 11,400ha area is State Forest, while the remaining third is agricultural land. A small proportion of land is also owned by Melbourne Water for the purpose of protection of the water supply (notably a buffer strip around the edge of the reservoir). Stakeholders include (among others) people involved in logging operations, trail bike riders and seed collectors; dairy farmers, hobby farmers and residents; local businesses and keen anglers; and the research group. Falkenmark and Folke (2002), discuss the ‘evident’ and ‘hidden’ functions of water and mention that multiple functionality makes goal conflicts likely. Many goals are related to ‘hidden’ functions of water (for example transport of nutrients). Through these ‘hidden’ functions, land management goals are connected to water quality. These land management goals are in themselves varied. Some land managers in the catchment are concerned with obtaining a commercial return from agricultural activities; others have significant off-farm income and their land management priority is enjoyment and self-fulfilment. Other stakeholders have concerns seemingly even further removed from water quality. Under current project funding structures, an important priority for the research group is to develop interesting research proposals that will attract funding from their investors (whether these investigations take place in Tarago or not, and whether or not they contribute to improved water quality in the Tarago reservoir). Perhaps we could say that a ‘hidden’ function of water here is convincing funding bodies to support particular research projects. Despite the driving importance of improved water quality, many goals and priorities need to be negotiated in improving catchment management in the Tarago catchment. In this context, brokering is not only needed between different knowledge systems but also between different priorities. By using the Tarago catchment as a site for trialling of DSS tools, the research group brokers the demands of their own research projects (for trial sites) and of stakeholders in the Tarago catchment (for specific best practice advice). The DSS developed by the group are not merely descriptive models of complex systems; they are explicitly normative models. They contain assumptions about priorities and goals. Assessing the validity of these models involves not only checking whether they accurately represent the complex (biophysical) system, but whether the normative assumptions they make are acceptable and agreed to by different stakeholders. These tools are a means of brokering the priorities of farmers and water managers.

Despite the important brokering roles played by the team, they continue to define their role as research (not extension). In a meeting to evaluate and reflect on the experience of trialling decision support systems in the Tarago catchment, a significant portion of the discussion centred around defining in what way the group’s activities were in fact research. Group members identified the way in which they were learning about farm systems and refining their models (decision support systems) as characteristic of research. However we would suggest that this is not in fact characteristic of research; extension practitioners also learn about farm systems as they interact with farmers, and adapt their conceptual models (of, for example, nutrient movement) to cope better with particular situations. So what then is different about research and extension (since both researchers and extensionists persist in saying they are different, even if they have trouble defining exactly how they are different)? What is characteristic of research in this situation is the way in which it deals with the uncertainties inherent in catchment management. This is discussed further below.

Practice change and uncertainty

‘Practice change’ has emerged as a coordinating concept for research and extension working on complex natural resource management issues. Both extension and research groups are expected to create ‘practice change’ through their programs of activities by brokering exchanges between different types of knowledge. The emphasis is on change in practice; the assumption being that change in practice leads to change in NRM outcomes. This may be a large and problematic assumption, particularly in the case of complex catchment scale issues. Some of the critical uncertainties faced by scientists working with ‘practice change’ in the Tarago catchment are outlined below. The first uncertainty is overt; it is discussed by researchers as something problematic in their attempts at practice change. The second uncertainty is more hidden; it tends to be downplayed and ignored if it is discussed at all, and only emerges as important from observations of the actual activities of practice change undertaken by the research group. Dealing with ‘practice change’ enables practitioners to side-step some of the critical uncertainties they have about making changes in land management practice in order to improve NRM outcomes. In order to move forward with the way we manage natural resources, we need to look at the problems hidden by ‘practice change’ rather than simply ignoring them.

Uncertainty 1: Will these ‘best practices’, if implemented, actually have any effect on movement of nutrients off farms, and ultimately on water quality in the Tarago reservoir?

This uncertainty relates to the natural variation within catchment systems and to the validity of the system models that are used to define best practice. Science actually has extensive methods for testing validity; experimental methods (including replication, controls etc) and methods of statistical analysis exist in order to enable researchers to empirically test their models. However, in the case of complex system models, testing is not always possible. It may be impossible (or prohibitively expensive) to collect the data required to validate the model. It is not possible to make controlled changes to catchment systems and measure the results. ‘Controls’ do not exist for such complex systems. Where empirical testing of models is not feasible, the next best option is usually comparison with other models of the same system. Even where such models are available, doubt remains about how well they represent the ‘real’ world, and whether or not this form of validation is circular (ie just because the two models agree doesn’t mean they are both correct; it might just mean that they share the same errors). Lahsen (2005) suggests that knowledge users might provide an alternative avenue for assessing the validity of model outputs. This approach has proven to be useful for testing farm scale models of nutrient loss. Applying models on particular farms and getting feedback from farmers about how well the outputs of these models ‘fitted’ with their knowledge of the farm system was a useful validation tool for researchers. While some of these approaches are helpful in resolving uncertainty about farm scale validity, perplexing questions remain about the cross-scale validity of nutrient movement models. Researchers are acutely aware that they don’t really know what effect best practices on farms will have at a catchment scale. The group has been seeking research funding to carry out detailed hydrological modelling of the catchment in order to look at how farm scale changes link to catchment outcomes. However, their funding bids have so far been unsuccessful. The users of models showing ‘cross-scale’ effects would most likely be policy makers; for example those who decide the priorities for changing practice in particular parts of the landscape. However the group has so far not been able to make connections with policy makers (or others) with extensive experience in looking at catchment scale impacts of farm scale changes in order to attempt user validation of their models.

None of these approaches have been able to resolve the uncertainty about the link between practice and outcomes for scientists engaged in practice change. Therefore the group is left in a position of uncertainty about how effective their recommended best practices are, and they begin to ask themselves why farmers should make changes suggested when there is little proof that the changes will actually be effective. However unresolved this problem may be, it is one which scientists are acutely aware of and take account of in their work. Although this uncertainty is more troublesome to scientists than the following uncertainty, it is actually less problematic. It is better to ‘know we don’t know’ than to ‘not know we don’t know’.

Uncertainty 2: What do all these different stakeholders really want to achieve? How much will this ‘best practice’ cost, how much time will it take, and what do we do when those who bear the costs don’t get the benefits? Given all these extra considerations, what is actually best practice for this particular farm?

Part of the reason why models of complex systems are extremely difficult to test is because there cannot be only one model of a complex system (Batty and Torrens 2005). Models imply particular viewpoints about the purpose of the system. Because complex systems are characterised by multiple purposes, researchers are often unsure about the criteria for a ‘best’ practice. Argyris and Schon (1974: 15) identify this dilemma when they write that

we act within a field of governing variables, all of which are affected by our behavior, all of which we strive to keep within an acceptable range. Instead of actions being related to ends on a one-to-one basis, any given action may affect many variables; all of them are ends in the sense that all behavior is shaped so as to keep all variables within an acceptable range.

Best practices will not be acceptable if they deal only with one stated goal (for example water quality) and have unacceptable effects on other goals (such as agricultural production). Despite the uncertainties outlined above, research is actually quite good at defining ‘best practice’ where they have a particular goal in mind. The two DSS that were tested in the Tarago catchment are examples of models that attempt to define best practice in a quite nuanced way. They account for many of the contextual (biophysical) factors that make ‘best practice’ so different from one farm to another and even from one paddock to the next on the same farm. They are even beginning to some extent take into account the multiple goals of a farm system; such as reducing nutrient loss and improving productivity. The two DSS trialled in the Tarago catchment were designed to cope with one of these goals each. Yet it was possible to use them in conjunction with one another to provide some level of ‘integrated’ nutrient management advice that recognised both goals. There is still work to be done; outputs from the two models must currently be integrated by the manager and are not designed to easily work together (ie the tools produce two separate nutrient management maps that can be compared but do not modify one another).

DSS are explicitly normative models. They do not only describe systems, but identify how systems may be manipulated for certain goals. These goals are part of the assumptions behind the models; for example the assumption that reducing nutrient loss and improving productivity are both important. Because these goals are treated as assumptions, they tend to be fixed within models. Goals are actually highly diverse and fluid; for some farmers nutrient loss is not a priority, for others productivity may not be important. Even though researchers are aware that different stakeholders have different goals, and that different people have different priorities, they tend to ignore or discount the importance of these goals in the validity of their ‘best practice’ models. For example, the nutrient budgeting tool trialled in the Tarago catchment was designed for dairy farms. Built into this tool were a number of assumptions about target nutrient levels. These targets were based on the assumption that nutrient levels should be maintained at a level where pasture is produced in the most cost effective way for a dairy farmer. This assumption was revealed when the model was used on beef properties in the Tarago catchment. While it was possible to modify the model to take account of different goals (ie new look-up tables for nutrient levels or by manually entering target levels) this experience highlights the way in which goals must be fixed in these models. A key limitation of current DSS is that they do not adequately allow for customising of the assumptions within the model. By this we mean the first set of questions asked in the model should be about things like prioritising system goals (profitable, productive, low labour use, minimal environmental impact etc), or alternatively specifying the value assumptions that affect the applicability of the model. Since understanding goals (and their fluidity) tends to lie outside the area of interest of biophysical scientists developing best practices, values and goals tend to be fixed and forgotten. While they remain unaddressed, uncertainty surrounds whether or not a particular practice is actually ‘best’ for a particular situation.

Implications for extension

Recognising and managing uncertainties is essential within a brokering practice attempting to initiate practice change. We have identified above that research seems to have clearly identified one type of uncertainty (related to how well linked practices are likely to be to particular outcomes) and have developed a number of tools to cope with this uncertainty. However uncertainty related to goals is inadequately recognised and accounted for. We suggest that the situation in extension practice is actually the opposite. One extension officer, discussing effluent management, said that he was often asked what the ‘best’ effluent system was. His response was that it is impossible to say, because the best system is different for every farm depending on how the system is managed and how the rest of the farm is managed (for example how the yards are cleaned, and what happens to water from feeding pads and calving pads). This response suggests that he is quite comfortable with the diversity of farm systems and goals that affect judgements about what is ‘best’ practice for a particular farm. In fact it seems likely that there are a number of tools, methods or ways of working embedded in his practice that help him to identify diverse goals and take account of them. He is also probably quite good at picking out the hidden assumptions within effluent systems in order to see if they fit within a particular farm system. These are the types of skills and tools that research needs in order to address goal related uncertainty in their efforts to support practice change. On the other hand, the extension officer’s statement that it is impossible to define the perfect effluent system suggests that he could perhaps use some assistance from researchers who are defining ‘best’ practice in increasingly nuanced ways.

Conclusion: A new relationship between research and extension?

This analysis of the complementary needs and skills of extension and research suggests that by working together as brokers of practice change, both practices would be enhanced. Nevertheless extension and research should not attempt to merge as one ‘super-practice’. It seems likely to be more efficient for each practice to develop its own strengths. Research and extension are definitely not the same; although both are broker practices they do this quite differently. A discussion of the tools they use for dealing with uncertainty might be a good starting point for negotiating new relationships between the two practices. Practice change requires practitioners to cope with multiple types of uncertainty simultaneously. It is therefore important for the two practices to learn to work together. Definitions of best practice need to take more account of values and goals that affect what ‘best practice’ is (as well as the biophysical factors that influence the definition). Interpretations of what ‘best practice’ is for a particular farm need to make better use of the detailed definitions that already exist. The challenge will be to develop new institutional forms that support this relationship and the exchange of knowledge between research and extension.

Acknowledgements

This research was funded by Dairy Australia, the University of Melbourne, and the Victorian Department of Primary Industries. Many thanks to all those who participated and collaborated in the activities described in this paper; particularly Ivor Awty, Barrie Bradshaw, Andrew Smith and landholders in the Tarago catchment.

References

Argyris, C and Schon, DA (1974). Theory in practice: increasing professional effectiveness. Jossey-Bass Publishers, San Francisco.

Attwater, R, Booth, S and Guthrie, A (2005). The role of contestable concepts in transdisciplinary management of water in the landscape. Systems Research and Behavioral Science 22(3), 185-192.

Batty, M and Torrens, PM (2005). Modelling and prediction in a complex world. Futures 37(7), 745-766.

Black, AW (2000). Extension Theory and Practice: a review. Australian Journal of Experimental Agriculture 40(4), 493-502.

Coutts, J and Roberts, K (2003). Extension Models and Best Practice in Extension. Proceedings of APEN National Forum, Hobart.

Eshuis, J and Stuiver, M (2005). Learning in context through conflict and alignment: Farmers and scientists in search of sustainable agriculture. Agriculture and Human Values 22(2), 137-148.

Falkenmark, M and Folke, C (2002). The ethics of socio-ecohydrological catchment management. Hydrology and Earth System Sciences 6(1), 1-9.

Gibbons, M, Limoges, C, Nowotny, H, Schwartzman, S, Scott, P and Trow, M (1994). The New Production of Knowledge : The Dynamics of Science and Research in Contemporary Societies. SAGE Publications, London.

Horlick-Jones, T and Sime, J (2004). Living on the border: knowledge, risk and transdisciplinarity. Futures 36, 441-456.

Kincheloe, JL (2005). On to the Next Level: Continuing the Conceptualization of the Bricolage. Qualitative Inquiry 11(3), 323-350.

Lahsen, M (2005). Seductive Simulations? Uncertainty Distribution Around Climate Models. Social Studies of Science 35(6), 895-922.

Lane, MB, McDonald, GT and Morrison, TH (2004). Decentralisation and Environmental Management in Australia: a Comment on the Prescriptions of The Wentworth Group. Australian Geographical Studies 42(1), 103-115.

Love, S, Melland, A, Gourley, C and Paine, M (2005). Decision support systems reveal new types of scientist - land manager collaboration. Proceedings of Environmental Research Event. Environmental Change: Making it happen, University of Tasmania, Hobart.

Müller, DB, Tjallingii, SP and Canters, KJ (2005). A transdisciplinary learning approach to foster convergence of design, science and deliberation in urban and regional planning. Systems Research and Behavioral Science 22(3), 193-208.

Paine, M (2005). Social research in innovation and change management. Proceedings of Knowledge Utilisation 05 Colloquium, University of Melbourne, Australia.

Ridley, AM, Paramore, TR, Beverly, CR, Dunin, FX and Froelich, VMC (2003). Developing environmental monitoring tools from sustainability indicators in the southern Riverina. Australian Journal of Experimental Agriculture 43, 271-284.

Roughley, A (2005). Knowing People. Reflections on Integrating Social Science 1978-2002. Canberra, Land and Water Australia: 124.

van Kerkhoff, L (2005). Integrated research: concepts of connection in environmental science and policy. Environmental Science & Policy 8(5), 452-463.

Vanclay, F (2004). Social Principles for agricultural extension to assist in the promotion of natural resource management. Australian Journal of Experimental Agriculture 44, 213-222.

Yin, RK (1994). Case Study Research. Design and methods. Sage Publications, Thousand Oaks.