Grassland categories and services

To investigate where it is possible to integrate management for production and environmental goals, it is necessary to first identify the types of systems that must be considered. In broad terms, many grassland areas could be triaged into three very general categories:

• • Healthy or marginal: Those areas where grasslands are predominately in a native or naturalized state; where there are minimal or readily manageable environmental threats; where current levels of production are appropriate and can be maintained with current knowledge and policies, and without the need for dramatic interventions. The focus can be more on environmental management than on production.

• • Disturbed: Those areas where more intensive practices often apply; where the stable species base is reasonable, but could be improved, and more careful management is required to minimize nutrient and water leakage. Knowledge may be insufficient to achieve a good long-term balance between the environment and production; the area needs attention but is still retrievable within the current economic and political frameworks. Production values are emphasized over environmental issues, but significant environmental benefits can be obtained.

• • Dying: Very disturbed areas, often subject to intense use, where production may focus on survival rather than commercial output; where there may be an almost complete absence of stable plant species and a high risk of adverse effects on the neighbouring environment through dust storms and/or stream pollution. Such areas may not be realistically restored to an environmentally acceptable state unless significant transfer payments are made or higher commodity prices paid by other sectors of society, even though they may still produce significant quantities of livestock products. Where production remains a dominant goal, environmental management objectives may be directed more towards preventing adverse impacts on neighbours.

The third category is problematic and requires society to accept the rehabilitation costs in return for the environmental benefits gained. For example, this would include re-designing livestock enterprises to incorporate more intensive systems that are highly productive with linkages to cropping systems and in consequence reduce the need to over-use the other categories of grasslands, i.e. the environmental costs of these systems becomes a net benefit for the less degraded systems. Feedlots are another obvious case. The main environmental need in these systems may be simply to limit the external impacts such as nutrients leaching and odours from the system. This category may be where society needs to accept that it is impossible to eliminate some of these negative impacts and recognize them as part of providing food, fibre, fuel and medicines for ever-increasing human populations. It would not be practical to restore highly degraded grasslands to the diversity and amenity value of traditional meadows. To continue to provide food for the world's ever-increasing population will require acknowledgement of a cost in terms of some continued adverse impacts from the more intensively used grasslands. An extreme example of categorizing landscapes based on health is cities, which are enormously expensive ecological disasters.

Table 1. Comparison of managing grassland at points above and below the optimum to achieve 0·75 of the maximum productivity in animal gain per hectare (Fig. 6). Approximate levels of various factors are rated, along with the relative advantage/disadvantage of managing to point A or point B. The more ‘+’ symbols the greater the response; ‘?’ denotes uncertainty about response

Applying this concept to categorizing grassland management may be a way forward. In reality, this is already taking place by incorporating the more ‘pristine’ areas into national park networks and subsidizing farmers in some countries to manage such areas on their farms. In contrast, the more intensively used grasslands are being subjected to increased regulation and policing to prevent damage to surrounding environments. If this concept is applied to grassland research, then the area of highest priority may be the ‘disturbed’ category, i.e. those areas where the environmental values can be enhanced and where levels of production can either be sustained or even enhanced, cost-effectively. Since this ‘disturbed’ category is currently less regulated, the prospects of developing better practices that meet production and environmental objectives within a self-regulated policy framework are good. Unfortunately, many grassland policies focus on the extremes of the spectrum due to a sense of emergency to protect ‘healthy’ or pristine grasslands (Sutherland Reference Sutherland2002a, Reference Sutherlandb) and to rehabilitate ‘dying’ grasslands by imposing over-simplified strategies such as reducing livestock numbers or reseeding vegetation.

In the present paper, the major emphasis is placed upon the ‘disturbed’ category because the authors believe that is where current research is most likely to deliver substantial benefits in acceptable time-frames by empowering land managers with knowledge to modify their current practices to incorporate multiple-use goals. Only recently have ecologists acknowledged the conservation value of these areas with different disturbance regimes (Tscharntke et al. Reference Tscharntke, Klein, Kruess, Steffan-Dewenter and Thies2005). From a global perspective, ‘disturbed’ grasslands will continue to be a major source of livestock products in many parts of the world (Sincich Reference Sincich2002).

A consideration of the range of services provided by grasslands is useful in assessing management practices. Up until the latter part of the 20th century, the provision of food, fibre, fuel and medicines were considered to be the primary services from grasslands. However, in the last 30 years, animal and plant biodiversity has been increasingly valued per se, and also for its contribution to maintaining the nutrient, water and energy flows which is recognized as a significant ecosystem service that underpins the functionality of grassland ecosystems. Additional services emerging as crucial in the 21st century are the globally important roles of grasslands in improving air quality, sequestering carbon, supporting pollinating and symbiotic organisms and other processes that maintain or repair ecosystems and landscapes. Unfortunately, much of the research through the last century concentrated on grasslands as a production system rather than as an ecosystem and the multi-services that grasslands provide were often ignored. Where productivity must be sustained over the long term, the key issue is now how best to use grassland resources without causing deleterious effects to the environment, or decreasing incomes. To achieve this requires more integrated thinking about the processes operative in grassland ecosystems. Since decision-making typically devolves onto farmers, it is crucial to identify the factors that influence them.

Do farmers aim to maximize product or profit?

The productivity of grasslands and the environmental consequences are often influenced in the first instance by the attitudes of farmers. Do they have short- or long-term goals? Do they want social status and/or a profitable business? Do they see themselves as nurturing the landscape or simply exploiting a resource; sometimes out of necessity? In the past, management has been treated simply as a technical issue, but the reality is that farming is done in a social context and profit is only one of many goals. Farmers often seek maximum productivity even if it is not profitable. In many cultures, farmers aim to maximize the number of animals/hectare (as a measure of wealth) or the size of their animals (for show or as a demonstration of success) or the total animal product/hectare (almost irrespective of cost). Production focused solely on animal number or performance inevitably leads to high grazing pressures and grassland degradation.

Using profit as the main driver of production does not always guarantee success and may also result in declining grassland condition. This occurs because many financial analyses do not include all the direct and indirect costs associated with production, but rely on simplified gross margins which often give a false estimate of the actual profit. To date, the costs of any environmental damage or rehabilitation programmes are rarely included in farm budgets. Moving grassland-based livestock production to a sustainable basis requires a change in culture within both the industry and the broader community, where environmental and social values added from an action are as significant in assessing its worth to the nation as the economic value it brings. Ways of integrating these objectives are still being incorporated into the concept we know as the ‘triple bottom line’ (TBL) or ‘sustainability’ reporting that is becoming an increasing common activity of large companies and small businesses alike to build trust and demonstrate accountability (Higgins Reference Higgins2001).

Vanclay (Reference Vanclay2004) proposes 27 principles that govern farmer decision-making behaviour; a subset relevant to the present paper is:

• • ‘It is hard to be green when you are in the red’; environmental management costs money.

• • ‘Doing the right thing’ is a strong motivational factor; what is ‘right’ is a social construct.

• • Farmers do not distinguish environmental issues from other farm management issues; they are all part of managing the farm.

• • Sustainability means staying on the farm; the concept of sustainability often has a strong social component – passing on the farm to the next generation is a central issue. That requires retaining the resource base and land tenure from one generation to the next.

• • Non-adoption of ‘good’ practices may reflect the fact that past recommendations were not good; can we guarantee that new recommendations will work any better?

• • Farmers construct their own knowledge; practices need to be tailored to their frameworks.

• • Farmers need to feel valued, particularly if society wants them to adopt more sustainable practices that may not make more money.

Applying these principles to better recommendations for grassland management is a difficult task. They illustrate the need to consider the social context within which farming occurs and the ways that managing for the environment and production need to align.

Environment and society

The environment is rated as a significant issue in developed countries and is becoming more important in others due to the realization that environmental problems have worldwide consequences such as global warming. This increasing community concern provides further impetus for farmers to voluntarily incorporate environmental values into their management objectives. However, while the general importance of the environment in developed countries is evident in surveys, management practices to address these issues are often based on studies that fail to take into account the effects and implications for the human sector (Batisse Reference Batisse1986), despite the fact that human impact is important, not only as a factor of degradation, but also as a factor of environmental recovery. The challenge is to find approaches to environmental management that give people the quality of life they seek while protecting the environmental systems that underpin society's well-being. Resolving this dilemma is difficult as society's ‘wants’ are often based on ignorance about (a) how food production occurs, (b) ecological and environmental processes, (c) what would happen if you do, or do not, manage the land, (d) the view that farming is either idyllic or too hard and (e) the view of the countryside as ‘the bits you drive through between the cities’ or ‘the bit I go to for recreation’ and want to look ‘nice’.

In Australia, for example, members of the public rank the environment third behind family and friends as an issue of importance, ahead of leisure (which includes sport) service to others, work, religion or politics (Young Reference Young2003). The general picture is that the public is concerned about the need for a clean environment and many are willing to personally pay for environmental services, whereas others see the government and industry as responsible for funding farmers to implement environmental programmes. This suggests that agriculture has measured societal support provided it develops more environmentally friendly management practices. Similar attitudes apply in Europe, where policies of the European Union (EU) and the Common Agricultural Policy strongly support environmental management (Hall et al. Reference Hall, McVittie and Moran2004). However, given that farmers do not necessarily distinguish between environment and farm issues (Vanclay Reference Vanclay2004), better environmental practices on farms need to be developed as part of production systems. The large sectors of society that live in cities need to appreciate that environmental impacts in grassland systems are more difficult to control or reverse than is the case in urban areas. Reconnecting urban dwellers to the countryside through awareness and education programmes, if appropriately done, could help resolve some environmental dilemmas.

Productivity, profits and the environment

Productivity is considered by many to be the engine that drives overall well-being with higher rates of growth in productivity most benefiting society. The need to simultaneously maintain productivity, profitability and the environment is becoming the major interaction that farmers seek to resolve. Intensification has been a major driver of productivity increases from grassland systems. However, arguably environmental risks and costs increase with intensification, as does the level of disturbance and utilization of grasslands, leading to loss of soil resources which increases costs of replacing nutrients and species and in managing nutrient runoff.

Just as importantly, high utilization rates and levels of productivity may not always be the most profitable. This is illustrated with the relative economics of sheep for wool production on grasslands in central New South Wales, Australia, using a gross-margin model (Vere et al. Reference Vere, Campbell and Kemp1993). A comparison was carried out between resowing a degraded pasture with introduced species (Phalaris spp. and Trifolium subterraneum) and managing the grassland to a more productive state with a combination of fertilizer, strategic rest and conservative stocking rate (Fig. 1).

Fig. 1. Modelled patterns in stocking rates (DSE: dry sheep equivalents) for grasslands managed to a more productive state or (ideal or typical) resown for wool production. The alternative (typical) resown pattern is more common in the region. Estimates are for an average area in central New South Wales (D. R. Kemp, unpublished, using a gross margin model developed by Vere et al. Reference Vere, Campbell and Kemp1993).

In each case, it was assumed that maximum stocking rates would be reached by year three in order to optimize the proportion of perennial grasses and improve the likely longevity of the pasture (Michalk et al. Reference Michalk, Liang, Feng, Kemp, Allsopp, Palmer, Milton, Kerley, Kirkham, Hurt and Brown2003b). The alternate pathway shown in Fig. 1 (dotted line) is likely to be more typical of farms in the region as farmers report that the productivity of sown pastures declines significantly over time (Reeve et al. Reference Reeve, Kaine, Lees and Barclay2000). High levels of grazing pressure/utilization result in the loss of more productive species, except in areas where the natural fertility and rainfall is high. The gross margin analysis of these pathways over a 10-year period for a range of conditions typical of the region (Fig. 2) demonstrated that there was little difference in the net profitability of the two main pathways unless higher stocking rates were possible – which applies in the higher rainfall, higher fertility areas of the region.

However, if the ‘typical’ (more common) scenario was considered (Fig. 1) the economics of resowing pastures were worse than the ‘ideal’ and not much different to ‘managed’ as lower mean stocking rates apply (Fig. 2). This leads to the general conclusion that the more appropriate pathways for development and improved environmental outcomes (as there was less disturbance to systems, less erosion risk and many native species remain) arise from conservative grazing management designed to retain the useful natural perennial species that are often lost rather than using reseeding to increase perenniality (Dowling et al. Reference Dowling, Jones, Kemp, Michalk, Gomide, Mattos and Carneiro da Silva2001). The ‘managed’ scenario involves less capital outlays and lower risks. Farmers are adopting that strategy. Within the 5 million ha of the region relevant to the analysis presented here, less than 0·01 of the area is being resown in any one year. The key point is that focus needs to be on net farm profits over the long term rather than biological output or gross income. Much of the literature seems to ignore these criteria. A similar outcome is evident in the hill country of the United Kingdom, where many areas are no longer being resown and sown pastures are restricted to the more productive parts of the landscape.

Fig. 2. Net present value of wool production for alternative grassland management strategies over a 10 year period (D. R. Kemp, unpublished data, using a gross margin model developed by Vere et al. Reference Vere, Campbell and Kemp1993). The ‘managed’ grassland is stocked at 0·80 of the ideal sown as per Fig. 1.

Exceptions to these generalizations do occur, such as the intensively grazed perennial ryegrass, white clover pastures of New Zealand, which do appear to survive a high level of utilization at least for part of the year, without any need to resow. For example, pastures producing 15 t dry matter/ha/yr carrying 25 dry sheep equivalent (DSE)/ha, have an annual utilization of c. 0·6.

A shift in mindset: revisiting concepts for productive versus sustainable grassland systems

There are two general views of farming systems conceptualized as ‘productivism’ and ‘post-productivism’ (Wilson Reference Wilson2001). The productivism view is that of a ‘factory’ where resources are put in one end of a pipe and products extracted at the other with increased production and short-term profitability used as the yardsticks of success (Kemp et al. Reference Kemp, Charry, Whiteley and Gardner2002). In the 19th century, J. von Liebig developed the concept of the ‘law of the minimum’, arguing that one factor is usually limiting production at any one point in space and/or time. Scientists have often focused on this approach and developed intensive farming methods and biochemical inputs to overcome production constraints (Ilbery & Bowler Reference Ilbery, Bowler and Ilbery1998). Society strongly supported this approach to ‘feed the world’, particularly through harsh times of food security, but now in times of plenty has become overly critical and rejects any sense of community responsibility. In this mindset the focus is restricted to a limited number of components with some consideration of efficiencies, but no solutions for unplanned or adverse environmental effects as these are considered externalities not necessarily to be managed within agricultural systems (Weiner Reference Weiner2003). Often in the search for higher productivity, limiting factors are applied to excess and any ‘leakage’ is seen only as inefficiency in primary production. High nitrate in water supplies is a legacy of this productivism approach. Nevertheless, von Liebig's ‘law’ helped sustain the ‘factory’ view up until the 1980s, when there was a change in focus from agriculture as ‘production’ to agriculture as ‘ecosystem manipulation’ (Weiner Reference Weiner2003), a conceptual shift required to accommodate increasing demands by society for more sustainable agriculture, including grazing systems for grasslands. The problem with the factory model is that while the view of production in terms of inputs and outputs is dynamic, its view of externalities such as N or glyphosate in runoff is static; hence the model is not very good at addressing these factors (Weiner Reference Weiner2003).

The related concept of ‘grasslands as crops’, developed from the early work of Stapledon and co-workers in Wales, proved an excellent stimulus to better management of grasslands, but is it still useful for addressing the problems facing us in the 21st century? The crop analogy has merit for good husbandry, but as most crops aim to be monocultures it conflicts with the reality of many grasslands where multi-species arrays need to be studied and managed at the community to ecosystem levels. Although N-fertilized perennial grass systems are still important in many regions, systems that incorporate legumes to deliver better environmental outcomes are now used in the grassland phase to restore soil resources to crop systems. If grasslands are treated simply as crops, this may encourage over-use of resources and poorer environmental outcomes (Scholefield et al. Reference Scholefield, Tyson, Garwood, Armstrong, Hawkins and Stone1993; Farruggia et al. Reference Farruggia, Gastal and Scholefield2004).

A second view of farming systems (coming more from managing natural grasslands) is an ecosystem model where it is acknowledged that the system is sustained by maintaining species and the cycling of water, nutrient and energy flows (Kemp et al. Reference Kemp, Michalk and Charry2001, Reference Kemp, Charry, Whiteley and Gardner2002; Weiner Reference Weiner2003). In this ‘systems’ scenario, it is acknowledged that the products harvested from grassland ecosystems are significantly less than the internal flows of energy and nutrients, and that the ecosystem processes need to be maintained in order to sustain production. Because more of the factors affecting production (including many environmental factors) are considered to be part of a system, the ecosystem view aims to cycle nutrients internally and minimize the leakage, as much as possible, of more judiciously applied inputs. A focus on overcoming single limiting factors is then inappropriate because the farm as a managed ecosystem is not self-delimiting (Weiner Reference Weiner2003). This approach fosters a more conservative, optimizing net profit strategy. Examples of shifts towards ‘best’ practice that emphasize environmentally friendly forms of production are already evident in new sustainable management practices implemented by farmer self-help groups such as Landcare Australia (Campbell Reference Campbell1994).

A switch from a ‘factory’ to ‘ecosystem’ view of farming has implications in how we develop management strategies for grassland systems and evaluate their impact. A key consideration is that within an ecosystem it would be impractical to manage each component to an optimum. Rather, there is a range of values around the optimum that do not result in great losses in productivity or ecosystem function; hence managing within a range is acceptable, e.g. animal gain/hectare (Kemp Reference Kemp1991). Farmers can manage more successfully over a range than continually chasing optimum or maximum values, thus corrective actions only need to be initiated when near the limit of that range, rather than continuously. This goal of maximum sustainable yield by using an optimal range which maximizes long-term profit or economic survival (as may be the case in developing countries where farmers live close to subsistence) illustrates that the ecological view forces us to define our objective on a different time scale than the factory approach, which is typically short-term (Weiner Reference Weiner2003).

Linkages between the environment and grassland production

The real difficulty of the ecosystem view is assigning value to the production and environmental components of the grassland system, which are often measured in different and non-interchangeable currencies. Weiner (Reference Weiner2003) asks, for example, how much money would compensate a community for groundwater polluted by nitrate? Or how much is a rare species worth? Understanding the linkages between grazing, productivity and environmental changes is paramount for grassland management where contrasting perspectives of economics and public concerns are emphasized.

While some of the close natural linkages between grazing, grassland productivity and gross environmental effects are well known (e.g. normal seasonal cycles, the extremes of droughts and floods; Retzer Reference Retzer2006), some more subtle effects and how the resource base is sustained are ill defined. Grazing can affect the soil water balance, depending upon the grazing intensity (Smoliak et al. Reference Smoliak, Dormaar and Johnston1972; Naeth & Chanasyk Reference Naeth and Chanasyk1995), the residual herbage mass (Hughes et al. Reference Hughes, Packer, Michalk, Dowling, King, Brisbane, Millar, Priest, Kemp and Koen2006) and the proportion of bare ground (Naeth et al. Reference Naeth, Chanasyk, Rothwell and Bailey1991). Reductions in soil water through grazing are attributed to soil compaction and sealing by animal treading, which increases surface runoff and leads to increased nutrient losses and reduced infiltration rates (Elliott & Carlson Reference Elliot and Carlson2004). These grazing impacts on grasslands can be modified by management with more conservative utilization strategies (e.g. maybe down to 0·75 of the biological optima), effectively increasing the perennial grass composition, which in turn reduces runoff, increases water transpired (reducing salinity and acidity risks) and reduces the need to reseed to restore ground cover (Michalk et al. Reference Michalk, Liang, Feng, Kemp, Allsopp, Palmer, Milton, Kerley, Kirkham, Hurt and Brown2003b). These components may have been influenced more by the standing herbage mass than directly by grassland growth rates or livestock productivity. The challenge for farmers is to determine the grazing pressure where adverse environmental impacts are low, yet productivity is satisfactory (Fig. 1). These interactions will be considered from an ecosystem perspective in terms of the management of water, nutrient and energy flows, and the diversity of species required to sustain the ecosystem function.

Managing water with grasslands

Sustaining clean and safe drinking water and irrigation supplies to meet the needs of the global population is a major issue for the 21st century. Water demand already exceeds supply in more than 80 countries which together account for 0·4 of the world's population (Swain Reference Swain2001). In Africa alone, more than 300 million people live under severe water restrictions (Bennett Reference Bennett2000). Grasslands are vitally important for the provision of water for agriculture, industry and domestic use.

The way grasslands are managed can have major impacts on the water cycle, ecosystem function and biomass production. In Australia, solutions to these problems lie in increasing the longevity of plants in the landscape (Singh et al. Reference Singh, Bird and Saul2003). Perennial-dominated grasslands fulfil two roles in water management – they efficiently utilize soil water, due to deeper root systems and longer growing seasons than annual species, and they can protect the soil surface from soil and nutrient loss, particularly during high intensity short-duration storms. Unfortunately there is often a decline in the perennial component in grasslands, resulting in more variable production over time, higher incidence of weed invasions, soil acidification, acute soil erosion and increasing salinity. This decline arguably reflects grazing pressures that are not sustainable.

Harvesting water

Clean runoff and aquifer recharge from grassland catchments is crucial to maintaining water supplies for agricultural, industrial, domestic and environmental use, food security and to progress many societies beyond subsistence (Raju Reference Raju1998; van Wesemael et al. Reference van Wesemael, Poesen, Benet, Barrionuevo and Puigdefabregas1998). In addition, groundwater supports major ecosystems by providing base flow to rivers and perennial flow to wetlands of international wildlife importance. Grassland management practices can vary the proportion of surface runoff to aquifer recharge (Bergkamp Reference Bergkamp1998). This was evidenced in the recent Australian drought, where over-grazing was controlled and less water (from the limited rain that did fall) flowed into dams and rivers. Maintaining herbage mass >2 t DM/ha in upland recharge areas (Fig. 3) effectively controls runoff (Packer et al. Reference Packer, Michalk, Brisbane, Dowling, Millar, King, Kemp, Priest, Unkovich and O'Leary2003). Excessive clearing and high grazing pressures result in less water being retained in the landscape, higher runoff in average to wetter seasons, reduced soil water recharge and reduced dry season river flows.

Fig. 3. Relationship between herbage mass for a range of pasture types (naturalized and sown) and maximum daily runoff. Data from Packer et al. (Reference Packer, Michalk, Brisbane, Dowling, Millar, King, Kemp, Priest, Unkovich and O'Leary2003).

Changes in vegetation structure can have differential impacts on water management. On the Edwards Plateau in Texas, for example, woody weed (mainly Juniperus, Quercus and Prosopsis spp.) invasion of savanna grassland, due to reduced fire frequency and overgrazing (Scholes & Archer Reference Scholes and Archer1997), has reduced recharge to the Edwards Aquifer because of the higher evapotranspiration losses from shrub cover (interception and transpiration of woody vegetation is higher, e.g. >0·70 of precipitation for juniper (Eddleman Reference Eddleman1983) compared with <0·18 for herbaceous vegetation (Thurow et al. Reference Thurow, Blackburn, Warren and Taylor1987)). Wu & Tiessen (Reference Wu and Tiessen2002) estimated that the hydrological impact of a shrub cover of c. 0·2 would decrease water yield to the aquifer by 35%. Invasion of Australian Acacia spp. in South Africa (Dye & Jarmain Reference Dye and Jarmain2004) had a similar negative impact on aquifer recharge, as have reafforestation projects. Restoration of forest cover in Spain (Bellot et al. Reference Bellot, Bonet, Sanchez and Chirino2001) and India (Sikka et al. Reference Sikka, Samra, Sharda, Samraj and Lakshmanan2003) reduced both runoff by c. 25%, and aquifer recharge by c. 40%.

In environments where grasslands have been a significant part of ecosystems, learning to manage them appropriately would be preferable to altering the ecosystem. Recharging aquifers can be advantageous for human health as there is a reduced risk of pathogens such as Cryptosporidium and Giardia that often contaminate surface waters (David & Mazumder Reference Davies and Mazumder2003). As a minimum, domestic water supplies need to come from water collected from catchments with high grass cover. A high quality natural habitat can significantly improve water quality and filtration costs (Sutherland Reference Sutherland2002a, Reference Sutherlandb).

Managing nutrient loss in grasslands

Managing carbon in grasslands

The energy in organic carbon compounds within grasslands is central to their productivity and environmental values. Estimates suggest that grasslands account for >0·1 of the total biosphere C store (Jones & Donnelly Reference Jones and Donnelly2004), >0·9 of which is sequestered below ground (Schuman et al. Reference Schuman, Reeder, Manley, Hart and Manley1999). As atmospheric CO₂ concentrations increase, identifying opportunities and promoting practices that enhance C stocks is a major grassland management objective. This is a formidable challenge because grasslands are complex ecosystems subjected to a wide range of environmental and management conditions. Nevertheless, under existing management conditions most temperate grasslands are considered to be C sinks (Jones & Donnelly Reference Jones and Donnelly2004), although their potential to sequester C is determined by their productivity. In the moderately productive Central Plains of the United States, for example, Gebhart et al. (Reference Gebhart, Johnson, Mayeux and Polley1994) measured C accumulation of 1·1 t C/ha/yr, whereas in the more arid shortgrass prairie of Colorado, Burke et al. (Reference Burke, Lauenroth and Coffin1995) measured accumulation of only 0·03 t C/ha/yr. Further, the same grassland may sequester C at different rates depending on its condition, with high rates sequestered during the ‘restoration’ phase relative to grasslands in a ‘maintenance’ phase (Schlesinger Reference Schlesinger1990; Fisher et al. Reference Fisher, Rao, Ayarza, Lascano, Sanz, Thomas and Vera1994).

The amount of soil organic matter retained by grassland soils is strongly influenced by management (Conant et al. Reference Conant, Paustian and Elliott2001). Land use change, especially the conversion of large areas of grassland to crop production (Vitousek et al. Reference Vitousek, Mooney, Lubchenco and Melillo1997; Scholes & Noble Reference Scholes and Noble2001), has contributed significantly to elevated atmospheric CO₂ levels. Estimates suggest that the combination of deforestation, grassland conversion and land management practices accounts for about 0·25 of C released from soil organic matter. Cultivation of the grassland steppe in China, for example, decreased soil organic carbon by 25% in 8 years, rendering the soils highly susceptible to erosion (Wu & Tiessen Reference Wu and Tiessen2002). The challenge for grassland management is to reverse CO₂ losses by storing more C in plant and soil organic matter to help minimize adverse environmental impacts (Vleeshouwers & Verhagen Reference Vleeshouwers and Verhagen2002), as well as to aid water and nutrient management. Farming systems where pasture phases are wedged between cropping cycles pose a continuing problem to accumulation of C.

In general, those permanent grasslands most depleted by poor management have the greatest potential for soil C increase (Jones & Donnelly Reference Jones and Donnelly2004) through the application of management practices such as moderate fertilizer application, oversowing legumes and moderate stocking density that increase forage production (Conant et al. Reference Conant, Paustian and Elliott2001; Reeder & Schuman Reference Reeder and Schuman2002). These practices not only increase biomass, but replace annual species which lack the dense fibrous root systems conducive to soil organic matter formation and accumulation with deep-rooted perennial species with higher carbon sequestration capacity (Jones & Donnelly Reference Jones and Donnelly2004). Soil C response to changed grazing management is less certain, particularly in the long term, with both light-to-moderate (LeCain et al. Reference LeCain, Morgan, Schuman, Reeder and Hart2002; Reeder & Schuman Reference Reeder and Schuman2002; Wright et al. Reference Wright, Hons and Rouquette2004) and heavy grazing (Schuman et al. Reference Schuman, Reeder, Manley, Hart and Manley1999) increasing soil C, depending on grassland type and condition.

Carbon mitigation options from ruminant livestock include: increasing production per animal, modifying diet, improving management of bedding materials and manure heaps and reducing overall livestock numbers (Monteny et al. Reference Monteny, Bannink and Chadwick2006). Of these mitigation strategies, only a reduction in the number of animals can be currently registered as a reduction in the Intergovernmental Panel on Climate Change inventory (IPCC 1997). These management options may involve important trade-offs such as a change from intensive livestock production reliant on annual fodder and cereal crops to grass-fed systems in order to reduce greenhouse gas emissions (e.g. CH₄) and increase soil organic matter (Soussana et al. Reference Soussana, Loiseau, Vuichard, Ceschia, Balesdent, Chevallier and Arrouays2004). Selecting livestock for reduced CH₄ emissions may provide further long-term solutions (Lassey et al. Reference Lassey, Ulyatt, Martin, Walker and Shelton1997) if it can be demonstrated that individual variation is under genetic control (Pinares-Patino et al. Reference Pinares-Patino, Ulyatt, Lassey, Barry and Holmes2003).

Grasslands are not part of the current Kyoto accounting protocol (2007–2012) but the size of the world's grassland resource makes it imperative to take into account this potential sink. Developing payments for carbon sequestration in grassland ecosystems could mean the difference between profit and loss for many livestock systems. However, this will require a commercially realistic verification system to be developed and accepted. Grasslands not only capture and store energy, but are often significant consumers of energy and from an environmental perspective these terms need to show a positive balance. However, it is clear that the energy balance becomes more negative as livestock production is intensified and manufactured inputs increase, especially N fertilizer (Kelm et al. Reference Kelm, Wachendorf, Trott, Volkers and Taube2004) herbicides and fossil fuels. This can be offset to some extent by the use of legumes and better management of farm manure. Ultimately, accounting procedures will need to consider the net energy balance if society is to make effective judgements about the sustainability of grassland practices.

Managing biodiversity

Intensified land use in agriculture and forestry, urbanization and climate change are recognized as the main causes of biodiversity loss (Tscharntke et al. Reference Tscharntke, Klein, Kruess, Steffan-Dewenter and Thies2005). However, the long-term sustainability of grassland ecosystems and the services they generate depends on conserving biodiversity at a landscape scale to ensure recovery from minor and major, small- and large-scale disturbances (Bengtsson et al. Reference Bengtsson, Angelstam, Elmqvist, Emanuelsson, Folke, Ihse, Moberg and Nystrom2003). Conservation biologists have been mainly concerned with the origins and maintenance of biodiversity, whereas agroecologists have been mainly concerned with its effects on ecosystem function and how these effects can be better managed in a production context (Vandermeer & Perfecto Reference Vandermeer and Perfecto1997). There is considerable debate about how agriculture and biodiversity can be managed alongside each other (Firbank Reference Firbank2005).

Biodiversity and production

Biodiversity is crucial to the productivity of grasslands, with more diverse communities being more productive because they make full use of the most limiting resources (Zhou et al. Reference Zhou, Sun, Huang, Gao and Han2006); however, an understanding of the numbers and types of species that optimize production is still to be determined. Many agronomists assume that grass plus clover equals a good pasture. This reflects the ‘factory’ approach where management is focused on a few species and their specific requirements, yet the potential value of many other plant species are ignored (Robertson & Swinton Reference Robertson and Swinton2005). Early work in the new field of biodiversity studies, using small plots, suggested that maximum productivity occurred when 6–10 plant species were present (Tilman Reference Tilman1996). At a small scale, sometimes fewer species were sufficient to maximize grassland production (Nicholas et al. Reference Nicholas, Kemp, Barker, Brock, Grant, Buchanan-Smith, Bailey and McCaughey1997). Recent work in small paddocks (<2 ha) in temperate Australia suggested that 10–20 species may be optimal for productivity (Kemp et al. Reference Kemp, King, Gilmour, Lodge, Murphy, Quigley and Sanford2003). The net primary productivity of those paddocks declined as species number increased from 20 to 50 (Fig. 4) while the variability in net primary production through the year increased substantially (Fig. 5).

Fig. 4. Mean site annual productivity in relation to site total plant species diversity for seven grazing sites across southern Australia (Kemp et al. Reference Kemp, King, Gilmour, Lodge, Murphy, Quigley and Sanford2003).

Bullock et al. (Reference Bullock, Pywell, Burke and Walker2001) showed hay yield to be higher (by up to 60%) in species-rich meadows (25–41 species) than species-poor meadows (6–17 species), but the yield of species-rich meadows was <60% of intensively fertilized grasslands. Arguably, the significance of biodiversity in grasslands may appear only at the landscape scale as the optimum number of species would be expected to increase as the area and number of different resource niches increase (Bengtsson et al. Reference Bengtsson, Engelhardt, Giller, Hobbie, Lawrence, Levine, Vilà, Wolters, Loreau, Naeem and Inchausti2002).

Since not all plant species are equal in their impact on the structure and function of plant communities, using plant functional type may provide a better definition of the relationships between diversity and productivity (Paruelo & Lauenroth Reference Paruelo and Lauenroth1996). The perennial grasses that dominate many grasslands (Tilman Reference Tilman1996) are often considered to be the main functional type driving ecosystem function (Kemp & Dowling Reference Kemp and Dowling2000). However, if the fertility is low they may be less competitive and minor species such as annual grasses and forbs become more frequent (Kemp et al. Reference Kemp, King, Gilmour, Lodge, Murphy, Quigley and Sanford2003); which may alter soil water dynamics (Hughes et al. Reference Hughes, Packer, Michalk, Dowling, King, Brisbane, Millar, Priest, Kemp and Koen2006) leading to dryland salinity and poor downstream water quality. The importance of plant functional type over species per se supports the species redundancy theory. Lolium perenne, for example, can be replaced by Festuca arundinacea (tall fescue) without any significant impact on grassland productivity, whereas the replacement of a perennial grass with a range of forbs or annual grasses may significantly alter productivity and ecosystem function. An implication for native grasslands is that management to conserve minor species that are part of a common plant functional type may not be important for productivity as natural grasslands are unlikely to be dominated by one or two species. Invariably other species invade under heavy grazing and variable climatic conditions, supporting the view that stable grasslands require many species.

Weed invasions, particularly by forbs, is perceived by many as the biggest issue confronting grassland managers. However, in more simple sown grassland the typical mixtures of grass and clover can be augmented with other forbs as ‘gap fillers’. The release of cultivars of Cichorium intybus (chicory) (Rumball Reference Rumball1986) and Plantago lanceolata (plantain) are a revival of an old practice (Foster Reference Foster1988) that fell out of favour during the 20th century. More diverse sown grasslands can reduce invasion of other species (van Ruijven et al. Reference van Ruijven, de Deyn and Berendse2003). Managing grasslands to enable existing species to utilize more resources to maintain higher mean levels of herbage mass can also restrict weed invasion (Badgery Reference Badgery2003; Meiners et al. Reference Meiners, Cadenasso and Pickett2004).

Livestock practices such as the timing and intensity of grazing are powerful management tools that affect both plant species composition and richness of grasslands (McIntyre et al. Reference McIntyre, Heard and Martin2003). Developing the pivotal strategies that maintain profitable livestock production enterprises and at the same time enhance local and regional biodiversity is a major challenge for farmers, land management agents and conservationists (Watkinson & Ormerod Reference Watkinson and Ormerod2001). Unfortunately, we still have only a very basic understanding of how to manage the impacts of different management practices on biodiversity (Dorrough et al. Reference Dorrough, Yen, Turner, Clark, Crosthwaite and Hirth2004), despite considerable research to tease out the linkages between biodiversity, productivity and ecosystem services (Naeem & Wright Reference Naeem and Wright2003) and that this point was being acknowledged by Stapledon 80 years ago (Stapledon Reference Stapledon1927). These impacts extend beyond the edible herbaceous strata, as studies on invertebrate communities in grazing systems indicate that more intensively managed grasslands often have lower invertebrate richness and abundance (Di Giulo et al. Reference Di Giulio, Edwards and Meister2001).

Bringing production and environmental goals together

The dilemma considered through the present paper is how to achieve production goals while attaining many of the environmental goals society now expects for grassland resources. The challenge is to develop a common framework that includes both production and environmental goals and from which guidelines can be provided for livestock producers. The main tool available to livestock producers is to vary stocking rates in relation to the feed requirements to optimize animal production (per head and per hectare), the available herbage mass and capture environmental benefits. As discussed earlier, many key environmental issues such as water and nutrient management are related to herbage mass and thus it provides an intermediary measure of the state of the system for both production and environmental criteria. A suitable framework is the relationships between stocking rate and animal production per head and per hectare (Fig. 6) developed by Jones & Sandland (Reference Jones and Sandland1974) ; for the purposes of the present discussion their normalized, linear function (for gain per head) is used. As pointed out by Connolly (Reference Connolly1976), that underlying relationship may also be curvilinear (e.g. quadratic).

Fig. 5. The relationship between species number and the coefficient of variability in seasonal production (from six weekly harvests over 3 years) for five grassland sites across southern Australia (Kemp et al. Reference Kemp, King, Gilmour, Lodge, Murphy, Quigley and Sanford2003).

Fig. 6. Relationships between relative stocking rate and relative animal production per head and per hectare based on Jones & Sandland (Reference Jones and Sandland1974). Points A and B are where production per hectare is 0·75 of the biological optimum. The dashed line is y=2−x and the solid line y=(2−x)·x.

The on-farm variability in forage resources, spatially and temporarily, means that it is arguably impractical ever to maintain the system at the biological optimum for production per hectare, i.e. the maximum point on the curve in Fig. 6 for any extended period. A realistic goal would be to aim for 0·75 of that maximum, which is also probably closer to the economic optimum for many systems. As the biological maximum is approached, the gain from each extra animal tends to zero, while the costs of maintaining that animal remains above zero. Economic analyses of optimal stocking rates tend to show a decline in the sustainable stocking rate as those analyses extend from a single year to decades (Randall Jones, personal communication). This is a result of the inclusion of more maintenance costs, including resowing, destocking, etc., over longer periods of time.

There are two instances where producers can achieve the 0·75 goal, represented by points A and B in Fig. 6. In the past, many producers arguably sought to stock pastures towards point B, probably due to being over-optimistic about seasonal conditions, aiming for a high utilization level, maximizing stock numbers for social reasons and maybe not appreciating which side of the biological optimum they were on. However, at point B, compared with A, the stocking rate is nearly three times greater and the growth rate of animals only one-third; this has significant implications for production efficiency and environmental impacts. Net profit would clearly be higher at point A. To achieve the higher animal growth rates at A it is a reasonable assumption that the quantity and quality of the herbage available is at least 0·75 of that required for maximum animal growth rates (this applies for the linear function in Fig. 6, but if a quadratic function is applied then the per head gain and hence forage available would be slightly larger – D. R. Kemp, unpublished). For sheep, this means 1·5 t (green) DM/ha, a level approaching high intake levels and at which water management is enhanced, biodiversity levels are reasonable and the desirable species are more likely to be maintained. The relative impacts of managing to points A or B are summarized in Table 1.

One of the major benefits gained from managing to A rather than B is a substantially greater profit. This qualitative rating identifies a host of reasons for managing to point A. The only uncertainty is forage quality, where the greater herbage mass that would result at A may have e.g. a lower legume content and older leaf age. These ratings apply where the majority of the animals produce saleable products.

A further implication of these relationships (Fig. 6) for improving the sustainability of grazing systems, which appears to have been rarely exploited, is in estimating the reductions in stocking rates required in adverse seasons, e.g. droughts, or cold winters. During adverse seasons when forage supply is limited, animal production declines towards maintenance, which means that the relationships move to the far right of these curves, sometimes below the axis where weight loss occurs. To then return to somewhere closer to point A would require a 75% reduction in animal numbers. Supplements could then be fed to animals to overcome such gross deficiencies, but to continue to maintain high stocking rates in such circumstances would result in a high grazing pressure and probably a loss in desirable plants. Overlaying such relationships with herbage mass, botanical composition and other performance measures that can inform on system sustainability would help resolve the boundaries within which livestock numbers should be managed.