Prospects for sustainable use of the pastoral areas of Australia’s southern rangelands: a synthesis

Keywords: grazing management, total grazing pressure, emerging land uses, infrastructure, invasive native scrub, biodiversity conservation, policy.

The southern rangelands of Australia, defined by Hacker et al. (2019a) as those semiarid and arid areas that lie inland of the wheat–sheep zone and south of the summer-dominant rainfall zone (Fig. 1), occupy 4.3 million square kilometres or ~57% of mainland Australia¹. Long-established industries in this region have included extensive grazing, practised over 54% of the area (ABARES 2016; Fig. 1), mining and tourism, with estimated annual values of A$1.4 billion, A$66.0 billion and A$1.9 billion respectively². The region includes several important towns, but, overall, is very sparsely populated, accounting for less than the 1.7% of the Australian population (approx. 26 million), and 16.9% of the total population of Indigenous Australians, that has been estimated to reside in the whole of the Australian rangelands (derived from Foran et al. 2019, table 2).

Fig. 1.  Approximate boundaries of the southern rangelands. The northern boundary approximates the southern edge of the summer-dominant rainfall zone as defined by the Bureau of Meteorology. Elsewhere the inland boundary of the wheat-sheep (or mixed farming) zone as defined by the Australian Bureau of Agricultural and Resource Economics Sciences (ABARES) provides a reasonable approximation. (Source: Hacker et al. 2019a).

Other land uses include conservation estate (national parks and a range of other areas protected or managed for conservation, 29%), traditional Indigenous uses (14%) and a range of minor uses including defence land, cropping and horticulture, plantation forestry, mining, and services/infrastructure/utilities (ABARES 2016). Land held by Indigenous people under national, state and territory land rights legislation, and determined Native Title (exclusive possession), will be more extensive than the figure for ‘traditional Indigenous uses’ given above because Foran et al. (2019) estimated the extent of the former at 33.6% for the rangelands as a whole. There is probably some overlap, in this case, with those areas classified here (Fig. 1) as used for grazing, or included in the conservation estate. Over 8 million hectares of land are currently managed by private conservation organisations (Australian Wildlife Conservancy 2021; Bush Heritage Australia 2021), representing a small but growing land use that probably also overlaps land used for grazing as defined here because much was originally, or still is, held under pastoral lease.

Discussion of sustainable land use in this region has generally occurred in the context of pastoral production. However, in recent decades, societal preference for non-provisioning ecological services has focussed attention on issues such as conservation of biodiversity, Indigenous and non-Indigenous cultural heritage, landscape aesthetics, and space in a crowded world (Holmes 2002), which apply equally to non-grazed areas. This broader discussion is reflected in the systemic challenges within Australian rangelands identified by Foran et al. (2019), including (1) undermining of the social licence to operate by those not managing resources appropriately, (2) providing opportunities for a growing and more youthful Indigenous population, (3) managing opportunities and threats associated with improved technologies, (4) governance resulting in human and financial capital leakage, and (5) improving human capacity and capability. In a similar vein, Nielsen et al. (2020) considered the key challenges facing Australia’s rangelands to be related to (1) supporting local communities, (2) managing natural capital, (3) climate variability and change, (4) traditional knowledge, (5) governance, and (6) research and development. These challenges will define the environment in which sustainable land use will need to be achieved.

Low profit margins for rangeland livestock production, identified by Briske et al. (2020) as one of several challenges for rangelands used for livestock production throughout the developed world, have seen extensive parts of the grazed southern rangelands (e.g. the Western Division of New South Wales, south-western Queensland and the Gascoyne–Murchison region of Western Australia) subject to government interventions in recent decades aimed at improving economic, social and environmental outcomes. Some of these interventions achieved a measure of success (e.g. URS 2004, 2015) but, generally, they failed to resolve the structural issues affecting the pastoral industry at regional scales.

Recent decades, however, have also seen an evolutionary transformation of the pastoral industry with a contraction of wool production and expansion of beef, goat and sheep meat production, the latter being associated with the widespread adoption of new breeds that pose specific issues for sustainable land use (Alemseged and Hacker 2014; Hacker and Alemseged 2014). These changes have been largely driven by the social and economic condition of the industry, reflected in issues such as reduced relative profitability of wool production and ability to control predation of sheep flocks by dingoes and wild dogs (see definitions in Table 1) (Forsyth et al. 2014), limited labour availability on pastoral properties, and, in places, the declining state of pastoral infrastructure. Sheep production has largely contracted to areas protected by the 5600 km ‘dog-proof fence’ in South Australia, New South Wales and Queensland (Hacker 2010). In some places, novel industries have emerged alongside pastoralism (e.g. carbon sequestration in native forests), and more novelty has been foreshadowed (e.g. renewable energy production; Garnaut 2019), which may complement ongoing changes within the pastoral industry itself.

Table 1.  Definitions of terms used in the paper

In this paper, we consider the prospects for achieving sustainable land use across the southern rangelands of Australia, particularly within the pastoral areas, but recognising that some emerging land uses could be adopted more broadly. We are concerned to promote land use and management that accommodate changing societal preferences today, while maintaining or enhancing the capacity of future generations to satisfy their preferences. We provide an overview of the environment of the southern rangelands, the pastoral industry, emerging land uses, and the historical impacts of pastoralism. We then propose principles and guidelines to address the key issues for pastoral land use, and consider the prospects for sustainability more generally based either on pastoralism or alternative land uses.

Vegetation and soil characteristics of the southern rangelands are summarised in Table 2 (a and b, respectively). Compared with areas that remain ungrazed, areas used for pastoralism contain more extensive areas of soils with red duplex, gradational (both calcareous and non-calcareous), and uniform (both fine, cracking and medium textured) profiles, and much smaller areas of soils with uniform, coarse textured profiles. The corresponding differences in vegetation result in greater areas of Acacia (predominantly mulga, Acacia aneura) woodlands and shrublands, chenopod shrublands (characterised by the genera Atriplex and Maireana), tussock grasslands (with Astrebla spp. in the north-east and a range of C₃ species elsewhere), and Eucalyptus forests and woodlands (particularly in the eastern semiarid zone) within the pastoral rangelands, and smaller areas of hummock grasslands (predominantly Triodia spp) and mallee communities (dominated by particular growth forms of Eucalyptus). This differentiation between pastoral and non-pastoral areas reflects the relative grazing potential of the broad vegetation types identified.

Table 2.  Soil and vegetation characteristics of the southern rangelands as defined in Fig. 1
Data for soils are derived from the Australian Soil Resource Information System https://www.asris.csiro.au/themes/Atlas.html#Atlas_Downloads; vegetation data are from Australia – Present Major Vegetation Groups – NVIS Version 5.1 http://environment.gov.au/fed/catalog/search/resource/details.page?uuid = %7B991C36C0-3FEA-4469-8C30-BB56CC2C7772%7D

Average annual rainfall (AAR) varies from <200 mm to ~500 mm across the southern rangelands, with the higher-rainfall environments occurring on the eastern margin (Fig. 1). Inter-annual variability typically varies inversely with the average, but is not higher than in several other desert regions (van Etten 2009). Nevertheless, low and variable rainfall is a fundamental determinant of land use throughout the southern rangelands. Additionally, rainfall throughout the region is not distinctly seasonal, which creates difficulties for pastoral management. Even though winter rainfall is generally more effective for plant growth, forage can be produced at any time of year and stocking decisions can, thus, be more difficult than in environments with predictable seasonal growth patterns.

Over the long term, climate change can be expected, with high confidence, to result in increasing temperatures, increasing frequency of hot days and warm spells, and an increased intensity of extreme rainfall events throughout the region. Time spent in drought is also likely to increase throughout the region, although the confidence attached to this prediction is lower. A decline in the annual rainfall is expected with high confidence in the south of the region, but rainfall trends in the north are uncertain (CSIRO and Bureau of Meteorology 2021). A shift in seasonal rainfall distribution, associated with the southward and coastward expansion of the tropical summer rainfall zone could also be expected on the basis of trends since 2000 (White 2016). More extreme temperatures can be expected to directly affect livestock production (Hansen 2009; Johnson 2018) and, indeed, the future habitability of some parts of the region (Gergis 2018, pp. 146, 185). Changes in climatic variables over the period 1991–2007 have been simulated to result in reduced forage production over much of the eastern and southern parts of the region, but with increased production in much of the Western Australian portion, compared with a base line period of 1961–1990 (McKeon et al. 2009). Importantly, these changes in forage production reflect an amplification of the corresponding rainfall changes so that any rainfall decline is likely to result in a disproportionately large reduction in forage availability (McKeon et al. 2009). Effects of aridity on soil health indices can be of magnitude and sign similar to any grazing effects, suggesting that predicted increases in aridity are likely to reduce soil health potentially as much as changes in grazing intensity, again with negative consequences for pastoral productivity under a changed climate (Eldridge et al. 2017).

The heterogeneity of the region with respect to soils, vegetation and climate is reflected in a continuum of primary and potential secondary productivity. In very broad terms, one end is defined by the ‘mulga zone’ of Western Australia, with AAR generally <200 mm and the widespread occurrence of shallow soils (often <30 cm) underlain by a siliceous hardpan. These soils are widespread in the arid zone of Western Australia (and hence are a substantial component of the region overall) but of very restricted occurrence elsewhere (Hacker 1987). Also located at this end of the continuum are the Nullarbor Plain, with equally low, or lower, rainfall and extensive areas of shallow soil over limestone, and the drier parts of South Australia. At the other end are the higher-rainfall environments on the eastern margins of the region, in south-western Queensland and western New South Wales, with AAR ranging from ~375 to 500 mm where mulga, the signature tree species of the southern rangelands, can form low-open forests or even open forests in the most favourable environments (Nix and Austin 1973). This variability in secondary production potential, reflected in stocking rates ranging from <0.1 dry sheep equivalents (DSE)/ha in the more arid parts to >0.6 DSE/ha on the wetter eastern margins, is a fundamental determinant of the prospects for sustainable land use across the region.

Patchiness of soil and vegetation is a distinguishing feature of semiarid and arid environments in which vegetation is too sparse to allow complete coverage of the ground surface. This patchiness creates, and in part results from, runon–runoff mosaics at small scales that are considered fundamental to the functioning of these landscapes. Water and nutrients are concentrated in vegetated patches that permit both the persistence of perennial vegetation and a higher level of net primary production than would be possible if water were uniformly distributed over the landscape (Noy-Meir 1973).

The ‘trigger–transfer–reserve–pulse’ framework describes this mode of landscape function (Ludwig and Tongway 1997). In terms of this framework, both ‘functional’ and ‘dysfunctional’ landscapes are described. In the former, the patch mosaic is sufficiently fine-grained to promote a tortuous pattern of water flow, resulting in most resources being retained within the local system, whereas in dysfunctional landscapes, where the vegetation pattern has been degraded, the flow is less inhibited, resulting is a loss of resources from the local system (Ludwig and Tongway 1997). Over much of the southern rangelands, maintenance or restoration of functional landscapes, in these terms, will be fundamental to pastoral productivity. Examples are provided by Holm et al. (2003), Holm et al. (2005) and Bean et al. (2015). Exceptions may include areas where deep sandy soils provide sufficient water storage to support perennial vegetation without the need for local redistribution, because of the ‘inverse texture effect’ (Noy-Meir 1973), or where landscapes inherently lack a capacity for local concentration of resources, particularly water (e.g. owing to heavy stone cover), and may be subject to seasonal boom–bust cycles that are largely independent of management.

Land used for pastoralism in the southern rangelands is administered by State and Territory authorities. Although some areas of freehold title exist, particularly in Queensland, most land is held under various forms of pastoral lease, which may be either ‘in perpetuity’ or for specified terms, depending on the jurisdiction. Leases typically require lessees to maintain the condition of the land, but failure to do so is often difficult to establish and available sanctions have generally been applied only in clearly unambiguous cases. Pastoralists operating with a freehold title have similar obligations but embodied in different legislation. The State- or Territory-based land administration system is associated with differences among jurisdictions in matters such as lease conditions, monitoring and compliance processes, and biosecurity arrangements relating to pest species and animal health.

Indigenous people now exercise total or partial management control over extensive areas through ownership of pastoral leases, and various forms of land rights, Native Title and Indigenous Land Use Agreements. Their aspirations for land use and their approach to management may differ from that of non-indigenous landowners, embracing Indigenous knowledge, culture and values (Ridges et al. 2020).

Apart from the relatively small areas that are held under freehold title, particularly in Queensland, almost all of the land used for pastoral production is held under various forms of pastoral lease. Although the southern rangelands have not seen the move to large corporate entities that characterise the northern pastoral industry, there has been a considerable corporatisation of parts of the region, resulting from the purchase of pastoral leases by mining companies. This trend has been particularly apparent in Western Australia (van Etten 2013). Poor pastoral profitability, particularly of those leases affected by ‘invasive native scrub’ (INS) in the eastern parts of the region, or other forms of land degradation, has resulted in a rapid turnover of ownership, use of leases for purposes other than pastoral production and lessees who are either absent or engaged in off-property work (URS 2013, 2015).

Nevertheless, land values have increased substantially across the region in recent decades, much more than could be justified by any increase in pastoral productivity, probably reflecting changes in land values elsewhere and potentially creating serious debt-servicing difficulties for recent entrants to the pastoral industry (Bastin and the ACRIS Management Committee 2008). In this respect, the situation in the southern rangelands is comparable to the ‘bubble’ produced by rising land values in the pastoral industry of northern Australia (Holmes 2015).

In the eastern parts of the region, particularly in Queensland and New South Wales, closer settlement policies, partially aimed at providing land for returned soldiers after World War 1, resulted in the break-up of large leases and the establishment of many smaller leases, which ultimately proved non-viable. Small lease size has also contributed to land degradation because pastoralists on smaller holdings have tended to adopt higher stocking rates (Table 1) for which the short-term economic benefits are perceived to outweigh the long-term cost of land degradation (MacLeod 1990; Passmore and Brown 1992). The legacy of these policies still persists, even though lease amalgamation, or the combination of several leases into single businesses, has long been a feature of the pastoral industry in these areas, and an objective of government-funded programs (e.g. WEST 2000 Plus in New South Wales and the South-West Strategy in Queensland) in recent decades.

In Western Australia, such closer settlement policies were not adopted but the low productivity of the resource base, noted above, and the effect of well documented land degradation on pastoral productivity (e.g. Payne et al. 1987, 1988) have produced a parlous economic situation for the pastoral industry in this part of the southern rangelands. Novelly and Warburton (2012) found that more than half of the 292 pastoral leases in the southern rangelands of this state were non-viable on the basis of their inherent productivity and the ‘capacity of the rangeland resources to be managed in an ecologically sustainable manner’. They considered that it was difficult to envisage a future for a substantial portion of leases without further rangeland degradation. However, URS (2013) found that structural adjustment to account for this situation is occurring and that across various subregions of the southern rangelands in Western Australia, the proportion of leases entirely dependent on pastoral activity ranged from 6% to 64% and that non-grazing activity of various types accounted for 30–68% of income. They envisaged a ‘new era’ for Western Australian rangelands in which businesses would be multi-faceted and the conduct of pastoral enterprises would not necessarily be a prerequisite for lease-holding of land within the region. Similar sentiments had previously been expressed by Fargher et al. (2003) and Hunt (2003) in relation to the rangelands nationally. The nature of this transition in the Western Australian portion of the southern rangelands has been summarised by van Etten (2013), noting that in some subregions, extensive contiguous tracts now exist with no or reduced domestic stocking, which can contribute to broad scale conservation management and restoration objectives. Indeed, some former pastoral leases have been specifically purchased by public or private agencies for this purpose.

Rangelands and the industries and communities dependent on them constitute human-environmental, or socio-ecological, systems. Reynolds et al. (2007) argued that in drylands, such systems are characterised by a unique set of features (a syndrome) that include climatic variability and unpredictability, sparse populations, distant markets and remote governance. They proposed the Dryland Development Paradigm, comprising five basic principles, to analyse changes in these systems. Local environmental knowledge is seen as fundamental in mediating the interactions between the human and environmental subsystems, which mediates their co-evolution. Stafford-Smith et al. (2007) found that these principles were all reflected in the degradation episodes in the Australian rangelands described by McKeon et al. (2004) but considered that strictly local environmental knowledge alone may be insufficient to ensure sustainable management. They considered that learning systems based on a wider community are required which combine local knowledge, formal research and institutional support. Strictly local learning is impeded by variability in both environment and social subsystems, and the extended time frames over which ‘slow variables’ (such as climate change, multi-decadal climate cycles and changes in landscape function) operate. This feature of the southern rangelands is an important part of the rationale for our attempt to distil management principles from the knowledge held outside local pastoral communities. However, application of these principles to achieve sustainable land use across the region is likely to be influenced by the extent to which governance arrangements for natural resource management reflect the principles of polycentricity and subsidiarity embedded in the concept of adaptive governance advocated by Marshall and Stafford-Smith (2010).

Gammage (2012) has collated numerous references to the open, park-like appearance of extensive areas of southern Australia at the time of European settlement, attributed to Aboriginal burning practices. Such practices were probably responsible for the preservation of extensive areas of open woodland in vegetation types dominated by Acacia and Eucalyptus species in the semiarid eastern parts of the southern rangelands, defined by Harrington et al. (1984a, 1984b) as ‘semiarid woodlands’ (although a contrary view has been expressed by Silcock and Fensham 2019).

The encroachment of INS in the semiarid woodlands under traditional pastoral management renders these environments well suited to the development of an alternative land use that provides an economic return from carbon sequestration (Moore et al. 2001). This potential is best developed in south-western Queensland, particularly the ‘eastern sector’ of the mulga lands recognised by Nix and Austin (1973), and in north-western New South Wales. Considerable areas within the semiarid woodlands have been devoted to carbon sequestration projects in recent years under either the ‘avoided deforestation’ (AD) or ‘human induced regeneration’ (HIR) methodologies of the Commonwealth Government’s Emissions Reduction Fund. The former (AD), involves landholders foregoing the right to clear native vegetation previously approved under a certified vegetation management plan; HIR projects involve the promotion of even-aged stands of native forests on land where development of forest cover has previously been actively suppressed. Cockfield et al. (2017) noted that at least 115 such projects, covering 1.8 million hectares, had been contracted to that time in the Cobar Peneplain and Mulgalands regions of New South Wales alone. The government investment in these projects was estimated at approximately AU$590 million. The current distribution of abatement projects (ERF 2020) indicates that comparable areas had probably been contracted in south-western Queensland. Since payments for carbon sequestered are generally loaded towards the early years of HIR projects, the value of carbon sequestration over the medium term is likely to be significant relative to the value of grazing (estimated above at AU$1.4 billion in 2016–17 for the southern rangelands in total; see footnote²).

The extent to which carbon sequestration will form a basis for changed land use in other parts of the southern rangelands is questionable. Garnaut (2019, p. 150) suggested a sequestration potential of ~250 million tonnes of CO₂-equivalent per annum in the semiarid rangelands. Outside the areas discussed above, this potential is likely to be realised mainly on the fringes of the wheat–sheep zone in South Australia and Western Australia, as these boundaries contract under a warming and drying climate. Even though landholder interest and project initiation have been considerable outside these fringe areas, especially in the mulga zone of Western Australia (Baumber et al. 2020), it is doubtful that these more arid parts of the southern rangelands will be capable of meeting the definition of forest cover required by the HIR methodology. The absence of approved plans for vegetation clearing in the more arid parts of the region will, likewise, prevent the implementation of carbon sequestration projects under the AD methodology.

Exclusion of both fire and grazing is generally required to maximise long-term (110 years) carbon sequestration in HIR projects, although over the 25-year commitment of many projects, carbon sequestration would be little affected by light grazing (20% utilisation of available forage; Howden et al. 2001). Even over longer terms, carbon sequestration might be little affected by grazing that is managed to ensure that germination and establishment of new cohorts of mulga or other species is not inhibited.

Payment for ecosystem services through a range of market-based instruments is an emerging alternative land use that can motivate land managers to conserve or manage land to achieve environmental or public benefits (Higgins et al. 2014; Ansell et al. 2016). Examples of these schemes in the southern rangelands include the ‘Pastoral Stewardship Incentive Pilot’ in South Australia (Wilson and Freebairn 2017) and the Enterprise Based Conservation project conducted under the West 2000 Plus Commonwealth–State Rural Partnership Program in western NSW (URS 2015). In the latter scheme, pastoralists could choose to receive financial incentives either for managing particular areas of land for conservation objectives, or for maintaining ground cover as close as possible to agreed targets (Hacker et al. 2010).

Whereas these schemes have applied to only a small proportion of grazing land in the southern rangelands, the growing recognition of the importance of biodiversity conservation and other components of natural capital in agricultural landscapes is likely to see markets for ecosystem services increase in the future. However, for such mechanisms to be effective in delivering public benefits, they will need either to be assured of long-term funding through incorporation into government policy and funding arrangements, or be able to operate profitably in the commercial market as proposed by Wilson et al. (2020) in relation to custodianship of wildlife on private land to support both diversified incomes for landholders and conservation. They will also require credible, landscape-scale monitoring methodologies.

Garnaut (2019) has drawn attention to the major economic opportunities provided by Australia’s high quality solar and wind energy resources, derived from extensive land areas, favourable climatic regimes and limited competitive land uses. The southern rangelands would, thus, appear to be ideally suited to industries based on renewable energy generation. Although distance from demand sources may favour less remote areas for development of small and medium scale projects utilising legacy grids, the transmission technology available to mega-scale projects should mean that potential for large renewable energy farms should exist throughout much of the region (Boulaire 2021).

As in many cases of settlers moving into new environments, the capacity of the southern rangelands to support pastoral production was often initially overestimated, an assessment sometimes exacerbated by ‘long periods of good seasons’ (Fyfe 1940, p. 20) in the early years of settlement before a major drought. This, combined with the limited availability of water supplies before the advent of heavy machinery for dam construction, and polythene pipelines for water distribution, often resulted in excessive numbers of animals being carried on individual watering points and rapid and severe degradation of the surrounding area (e.g. Osborn et al. 1932, cited in Andrew and Lange 1986; Nicholson 2017, pp. 35–36). In addition, the desire of governments in some states to ensure that pastoral leaseholds were taken up by genuine settlers and that pastoral development proceed as rapidly as possible, combined with optimistic assessments of carrying capacity, led to the imposition of lease conditions that no doubt contributed to land degradation.

Since climatic conditions in the southern rangelands do not require stock to be removed on a regular basis (e.g. due to very cold winters), the grazing system is predominantly sedentary. Grazing has traditionally continued year-long, with stocking rate being the major variable under management control, and stock being removed from their paddock only for annual operations such as shearing, or when pastures could no longer support them, rather than for pasture management purposes. This traditional system of management is arguably incompatible with the ecology of some parts of the southern rangelands.

The incursion of INS over extensive areas of the semiarid woodlands has been noted above. Contributing factors are generally considered to have been the changed fire regime resulting from the cessation of Indigenous burning, the European tendency to supress wildfire, consumption of the grassy fuel by livestock, and the weakening or removal of the competition for establishing shrub seedlings previously provided by perennial grasses (Harrington 1991). This latter factor is probably of increased importance following the reduction of rabbits, which ‘consume almost all seedlings of these [semiarid zone tree] species’; Wilson et al. 1992, p. 10) by biological control methods from the 1950s. Witt et al. (2009) demonstrated considerable variation in multi-decadal woody cover change among sites, land types and rainfall zones in south-western Queensland, so the generality of this model may be questioned. Nevertheless, since traditional pastoral management provides limited opportunities to rest country to build adequate fuel loads for management burning, to implement burns over extensive areas when seasonal conditions are favourable, or to maintain perennial grasses, it is arguably incompatible in many situations with maintenance of the productive grassy state that was attractive to early pastoralists. We would argue that this incompatibility remains despite any contribution to INS encroachment that may have resulted from increased atmospheric CO₂ concentrations (Polley et al. 1997; Ward 2010; Donohue et al. 2013), whose correlation with global woody plant increases was considered by Archer et al. (1995) not to represent cause and effect.

Where palatable perennial shrubs allow livestock to be retained under dry conditions, there may be another mismatch between ecology and traditional pastoral management. Under these circumstances, excessive pressure may be placed on perennial grasses, especially where these were originally sparse as, for example, on the extensive areas of shallow hardpan soils in the mulga shrublands of Western Australia. Beeton et al. (2005, cited by Silcock and Fensham 2019), considered drought feeding of palatable mulga scrub to be a major factor contributing to degradation of mulga communities in the eastern part of the region. Long-lived perennial grasses were probably important for animal production, given the nutritional importance of green leaf (Freudenberger et al. 1999), but their retention has proved difficult when browse permits the prolonged retention of livestock under adverse seasonal conditions.

In chenopod shrublands, there is a natural ecology that is arguably more compatible with sedentary grazing than those situations described above. In these environments, inter-shrub growth is composed of annual herbs and annual or short-lived perennial grasses, the species mix depending on the seasonal incidence of rainfall. Since this ephemeral growth is preferred by livestock when available (Pahl 2019a) there is an inherent capacity in these communities to provide some natural rest from grazing for the major perennial elements. For this reason, Wilson (1979) considered that ‘controlled continuous grazing’, at stocking rates that avoid complete defoliation of the shrubs, was the best management option for saltbush (Atriplex vesicaria) communities in south-western New South Wales. Nevertheless, continuous grazing even at conservative stocking rates leads to some long-term depletion of the perennial shrubs (Andrew and Lange 1986; Watson et al. 1997).

Although modern pastoral management is often superior to traditional practice, incorporating for example resting of paddocks and containment feeding under drought conditions, the land and the pastoral industry still suffer the legacy of the incompatibilities described above.

Six major degradation events since 1890, in the sense of depletion of perennial vegetation and soil erosion under conditions of drought and high grazing pressure, have been described by McKeon et al. (2004) in the southern rangelands. Incursion of INS into previously ‘open’ areas in the semiarid woodlands, associated with a period of high rainfall, is described as a seventh major ‘degradation’ event.

Noble et al. (1996) summarised the various reports on the condition of the grazed southern rangelands available at that time. The proportion of regional survey areas in Western Australia, South Australia and Queensland considered to be in ‘Poor’ or ‘C’ (degraded; Tothill and Gillies 1992) condition classes ranged from 20% to 42%. However, the extent of degradation was highly variable among survey regions and vegetation types, and methodologies differed among surveys, so that generalisation is difficult. Survey results from the Western Division of New South Wales reported the extent of particular types of degradation rather than condition classes. Encroachment of INS was considered ‘moderate’ or ‘severe’ over 37.6% of the division, whereas ‘moderate’ or ‘severe’ sheet/rill erosion, gully erosion and scalding were reported for 0.1%, 8.9% and 18.6% of the region respectively.

A common finding of regional surveys is that the level of degradation tends to be greatest in land types or vegetation communities with the highest potential for pastoral production (Noble et al. 1996). The effect of degradation on pastoral productivity at landscape scale could thus be greater than suggested by the figures above. Several studies in the southern rangelands have found that loss of palatable shrubs, associated with enhanced growth of ephemeral species, does not reduce animal production under reasonable seasonal conditions at commercially realistic stocking rates (Leigh et al. 1968; Wilson and Leigh 1970; Graetz 1986; Wilson and MacLeod 1991; Holm et al. 2005), but a disproportionate effect of degradation of the more productive land is likely to be manifest under poor seasonal conditions.

Across the southern rangelands in New South Wales, South Australia, Northern Territory and Western Australia indicators of both landscape function and critical stock forage were stable or increased for a majority of sites from the early 1990s to 2005 (Bastin and the ACRIS Management Committee 2008, figs 3–7 and 3–12 respectively). When these gross changes were adjusted for seasonal conditions, up to 20% of sites (depending on bioregion) showed an improvement in landscape function despite poor seasonal conditions, whereas a somewhat smaller percentage of sites showed a decrease in landscape function under favourable seasonal conditions (fig. 3.8 in Bastin and the ACRIS Management Committee 2008). A similar picture emerged for the various indicators of critical stock forage (fig. 3–13 in Bastin and the ACRIS Management Committee 2008). These counterintuitive trends, observed on a minority of sites, suggest that both beneficial and non-beneficial impacts of management are reflected in the data.

More recent survey data from the West Australian Rangeland Monitoring System showed that in the upper and lower southern rangelands of that state, respectively, vegetation cover for 15% and 43% of the area of the most productive pasture types had decreased or not improved between 2009 and 2019 (Department of Primary Industries and Regional Development 2019). The average density of pastorally desirable shrubs had decreased across both subregions, following widespread increases in density, canopy area and species richness, under generally favourable seasonal conditions, reported earlier by Watson et al. (2007). Continuing decline in vegetation condition was considered highly likely if grazing pressure was maintained or seasonal conditions deteriorated (Department of Primary Industries and Regional Development 2019).

Trends in range condition indicators are today variable in space and time. They reflect overwhelmingly the influence of seasonal conditions, but with management, nevertheless, able to exercise a discernible effect, either by allowing seasonal responses to be expressed or by inducing counterintuitive trends. It is not possible to discern any overall directional trend, and broad generalisations about the state of ‘the rangelands’, or their trend, should be avoided. Furthermore, regional assessments such as those reported above may be insensitive to degradation processes operating in restricted parts of the landscape (e.g. Pringle and Tinley 2003; Pringle et al. 2006), which will be important for individual land managers.

Since European settlement, drought, habitat degradation by introduced herbivores and predation by feral cats and foxes has resulted in the extinction of 11 medium-sized arid-zone mammal species and a dramatic reduction in the range of many others (Morton 1990; Lunney 2001; Silcock and Fensham 2019). Woinarski and Fisher (2003) suggested that most mammal losses from the semiarid rangelands occurred between the 1850s and 1880s. In the 19th and 20th centuries, bounties were paid as an incentive for population control of native fauna including dingos, kangaroos, wallabies, bandicoots, bettongs, emus, eagles and kites (Hrdina 1997). The campaign against the dingo was successful in reducing their populations across large tracts of the arid zone (e.g. Glen and Short 2000).

Rangeland birds, reptiles and amphibians have been less affected than mammals. Although range and abundance of many arid zone bird species, particularly ground-nesters and those associated with riparian habitats, have declined since colonisation (Reid and Fleming 1992), Morton (1990) reported that no species has become extinct. Saunders and Curry (1990) suggested that there is little to indicate that the present-day status of any bird species in the Australian arid zone is changing rapidly. Increasing, decreasing and stable trends have been identified among 49 rangeland bird species over the period 1999–2006 (Cunningham et al. 2007). Causal factors were unknown but seasonal variation appeared to be involved for many species. No desert-dwelling reptile or amphibian has become extinct (Morton 1990) and there is no evidence that the abundance of any reptiles has declined (Wilson and Swan 2017).

Despite this, the abundance of native fauna has been reported to be greater under light than heavy grazing intensity (James 2003; Read and Cunningham 2010), and Haby and Brandle (2018) found that the gradual passive recovery of small mammal and reptile assemblages in the Flinders Ranges in South Australia was facilitated by the removal of livestock in open A. aneura woodland in fair condition, but not in degraded chenopod shrubland. In contrast, no significant difference in animal richness or abundance was reported with differences in grazing intensity in a comprehensive review of Australian literature by Eldridge et al. (2016).

The impact of grazing, or pastoral development generally, on biodiversity could be expected to be non-uniform, associated with gradients of grazing pressure around watering points. Although localised effects on biodiversity may be observed, the proliferation of artificial waters across the southern rangelands (Hacker and McLeod 2003; Watson et al. 2005, cited by Bastin and ACRIS Management Committee 2008, pp. 56–57) means that few areas today would be unaffected. Studies of the response of animal species to these gradients indicate that some increase in abundance with increasing grazing pressure, others decrease and others are unresponsive (Landsberg et al. 1997).

Although some plant species have been shown to be negatively affected by pastoralism and, in some instances, have been eliminated from heavily grazed areas (Landsberg et al. 2003), most species remain widespread and common and, with the exception of some species endemic to artesian mound springs subject to substantial drawdown (Silcock and Fensham 2019; Powell et al. 2015), few have declined so as to be considered rare at the landscape scale (Silcock et al. 2014; Silcock and Fensham 2019). Overall, the drier parts of Australia have been much less modified by European settlement than have more arable regions, and the adaptation of the flora to drought has probably conferred some resilience to exotic herbivores (Silcock et al. 2014).

Nevertheless, severe grazing pressure in the past has caused transitions that are still apparent in landscapes today. Some shrub species, for example, have increased where populations of grazing-sensitive species such as Atriplex vesicaria have been reduced by over-grazing (Tiver and Andrew 1997), sometimes to the point of local extinction because of a lack of recruitment (Hunt 2001). Under current land management, recruitment of numerous dryland shrubs and trees is limited or even prevented so that some of the more restricted species are at risk of extinction as older plants senesce (Tiver and Andrew 1997; Denham and Auld 2004).

Grazing gradient studies have shown, as for faunal taxa, that some plant species increase in abundance with increasing grazing pressure, others decrease, and some appear to be unresponsive (Landsberg et al. 1997, 2003). Díaz et al. (2007) reported an increase in the proportion of annual, prostrate and unpalatable species with increasing grazing intensity. At low grazing intensities, impacts on plant diversity are minimal (Fensham 1998; Fensham et al. 2014), although changes in composition are often observed (Fensham et al. 2010).

A comprehensive picture of responses of rangeland ecosystems to plant invasions is not available (Grice 2006). Of 17 grass species identified by van Klinken and Friedel (2017) as ‘high-impact’ environmental weeds in Australia, only Cenchrus ciliaris (buffel grass) and C. setiger (birdwood grass), introduced as pasture species suitable for semiarid environments, have made significant inroads in the southern rangelands (Miller et al. 2010). Buffel grass has most commonly colonised badly degraded and eroded areas in the more fertile parts of the landscape and is considered ‘naturalised’ in most Australian rangelands, except South Australia where it remains a ‘declared’ plant (Landscape South Australia Act 2019). These species place native ecosystems at risk through competition (e.g. Franks 2002) and elevated fire frequency and intensity (Miller et al. 2010). Weeds of National Significance in the southern rangelands include various opuntioid cacti (Austrocylindropuntia, Cylindropuntia and Opuntia species), which are widespread in both western and eastern parts of the region (Hosking et al.1988; Western Australian Herbarium 1998). Ward’s weed (Carrichtera annua) has invaded extensive areas of the southern rangelands since its accidental introduction to South Australia in the early 1900s, particularly on calcareous soils in areas of winter-dominant rainfall (Cooke et al. 2011).

Impacts of pastoralism on the ‘weak points’ in spatially and temporally variable rangeland systems (Stafford-Smith and McAllister 2008) may well have disrupted evolutionary processes in ways that are not yet well defined. Nevertheless, care should be taken not to accept uncritically the conventional narratives about rangeland degradation (Silcock and Fensham 2019). There is no evidence that any plant species has become extinct as a result of pastoral settlement (Silcock et al. 2014; Powell et al. 2015), and most faunal taxa appear to have been little affected, in stark contrast to the ‘catastrophic’ record of mammal extinctions in the arid and semiarid zones attributable primarily to introduced predators (cats and foxes) rather than pastoralism (Silcock and Fensham 2019).

In this section, we review what we consider to be the key issues that must be addressed to achieve sustainable pastoral land use in the southern Australian rangelands and formulate this understanding, following the format of Hunt et al. (2014), in terms of principles and guidelines for sustainable pastoral management (Table 3).

Table 3.  Principles and guidelines for sustainable pastoral management in the southern rangelands

While of little value for day-to-day stocking decisions, knowledge of the estimated carrying capacity (see Table 1) of a property in its current condition is fundamental to sustainable management because this establishes the centroid around which actual stocking rates can be expected to fluctuate. It also provides a measure of productive potential, both absolute and relative to other properties, which will be important in dealing with financial institutions. Carrying capacity estimates for all pastoral leases in the southern rangelands have been compiled by land administration authorities on the basis of a range of methods (e.g. Condon 1968; Condon et al. 1969; Johnston et al. 1996a, 1996b; Hamilton et al. 2008), but often do not reflect the current condition of the property. In practice, it may be necessary for landholders to develop their own estimate of carrying capacity on the basis of historical stock records, a critical assessment of the condition of the land that has resulted from historical management, and any other local knowledge available. Because carrying capacity needs to be expressed in terms of some equivalence scheme (Table 1), information is required concerning the equivalence ratings for the species, domestic and non-domestic, that graze on the property (Guidelines 1.1 and 1.2).

Despite the acknowledged importance of stocking rate for animal production, the classical relationships defined by Jones and Sandland (1974) are often difficult to detect in large semiarid and arid zone paddocks because of spatial and temporal variability of forage supply (Roshier and Barchia 1993; Ash and Stafford-Smith 1996; Roshier and Nichol 1998; Freudenberger et al. 1999), and often the presence of substantial numbers of non-domestic herbivores (Hacker et al. 2019a). Numerous studies in the southern rangelands (e.g. Wilson and Leigh 1970; Wilson 1991a; Roe and Allen 1993; Roshier and Barchia 1993; Freudenberger et al. 1999; Holm et al. 2005) have demonstrated the overriding importance of seasonal variation in determining both animal and vegetation responses. Generally, the order of influence for almost any response variable is therefore as follows: (1) seasonal conditions, (2) stocking rate and (3) management system.

Some formal studies in the southern rangelands, and much anecdotal evidence, indicate that ‘conservative’ stocking, which includes both moderate stocking rates and early response to developing drought conditions, gives the best outcomes both ecologically and economically in this highly variable environment (Hacker et al. undated; Morrissey and O’Connor 1989; Woods 1992; Buxton and Stafford-Smith 1996; Stone 2004; Stafford-Smith and McAllister 2008). Over the long term, benefits of conservative stocking are achieved through higher production per head, improved reproduction rates, reduced production costs particularly in dry years, and opportunities to improve both the genetic potential of the herd/flock and land condition. Similar conclusions have been reached by studies outside the southern rangelands in Australia (e.g. Rickert 1996; Landsberg et al. 1998) and overseas (e.g. Holechek et al. 1999; Higgins et al. 2007; Fynn et al. 2017).

A method of calculating conservative stocking rates, proposed by Hacker and Smith (2007), uses a benchmark (BM) value of DSE days per ha/100 mm of average annual rainfall (DDH/100 mm_(BM)), derived from long-term average annual pasture growth and a ‘safe’ utilisation factor derived from Johnston et al. (1996a). In practice, the numerator of this ratio, a measure of carrying capacity, may need to be derived more subjectively, as noted in the discussion above. Once established, the BM can be used proactively to predict future conservative stocking rates on the basis of seasonal climate forecasts for the next 3 months (Guideline 1.3). However, careful monitoring of utilisation levels achieved in the field is still required. Alternatively, use of a grazing chart that records the stocking history of a paddock, and thus allows the effective stocking rate to be subjectively related to the impact on measures of land condition (e.g. utilisation of preferred species or ground cover), could be used to estimate conservative stocking rates retrospectively.

Even at low levels of grazing, impacts on ecosystem function, structure and composition are largely negative, and more pronounced in dry, lower-productivity environments (Eldridge et al. 2016, 2017). Accepting that some impact of pastoral use is inevitable, such findings reinforce statements made elsewhere in relation to Australian rangelands that ‘getting the stocking rate right’ is the major management issue (Stafford-Smith et al. 2007) and that stocking rate is ‘the most critical aspect of livestock grazing’ because of its profound effects on livestock production, financial performance and land condition (Hunt et al. 2014). We concur with these statements in relation to the southern rangelands while acknowledging that management system, i.e. the pattern of grazing and resting imposed on particular paddocks, or the property overall, can have important additional effects (see below).

In large paddocks at low stocking rates, livestock do not graze the landscape uniformly (Fuls 1992; Barnes et al. 2008). Selective grazing under the traditional continuous grazing model can result in over-utilisation and a loss of perennial grasses and other desirable species (Norton 1998), thereby increasing the risk of land degradation and associated negative impacts on production.

Grazing systems that incorporate periods of rest between grazing events have been widely promoted as improving both production and environmental sustainability (e.g. Savory 1983; McCosker 2000; Teague et al. 2008). Numerous systems have been developed, defined by the number of paddocks and number of herds/flocks, their dependence on vegetation dynamics, and the timing, duration and frequency of grazing and rest periods (di Virgilio et al. 2019). Included are systems designated as seasonal, deferred, rotational, short duration, holistic and cell grazing. In these systems, grazing smaller areas at higher stock densities can increase the uniformity of grazing by reducing selective and patch grazing, and increasing grazing pressure on species that would not otherwise be utilised (Norton 1998, 2003). In addition, rest allows vigour of palatable species to be maintained as they can recover between grazing events (Norton 1998). However, excessive subdivision, and rapid movement of animals among paddocks, may also deny their access to the functional heterogeneity that can be beneficial to health and production (Fynn et al. 2017). In similar vein, Fuhlendorf et al. (2017) contended that grazing management that aims to promote uniform, moderate grazing across the entire landscape fails to promote the spatial and temporal heterogeneity that enhances biodiversity in agricultural landscapes. These authors put the case for large continuous tracts of rangeland, not closely subdivided, to enable disturbance processes to interact with inherent heterogeneity, so as to form multi-scaled mosaics capable of providing multiple goods and services.

The production and sustainability benefits of rotational grazing strategies have long been debated (e.g. Briske et al. 2008, 2011; Teague et al. 2013). A global meta-analysis of 176 studies incorporating periods of planned rest (McDonald et al. 2019a) reported increased ground cover and animal production per hectare under grazing systems that include rest, with benefits generally increasing with longer periods of rest relative to grazing. However, incorporation of rest did not result in significant differences in biomass, plant diversity or animal production per head. In contrast, another recent global review of grazing strategies in rangelands did not find statistical evidence of improved sustainability under rotational grazing strategies (di Virgilio et al. 2019). Similarly, reviews by Hawkins (2017), Briske et al. (2008) and Holechek et al. (2000) also reported no benefits of holistic or rotational grazing for production or sustainability.

However, producers often claim considerable benefit from non-continuous grazing systems. Differences between producers’ experience and the scientific evidence outlined above are in part attributable to the short duration and small spatial scales at which experimental studies are typically conducted, along with confounding by stocking rates and inflexible, experimental grazing treatments (Teague et al. 2013). In particular, more uniform utilisation of the landscape owing to the subdivision that usually accompanies rotational grazing can allow higher stocking rates to be maintained (Norton 1998, 2003), realising benefits for commercial producers that are not evident at experimental scales.

Within the southern rangelands, few studies have compared rotational and continuously grazed systems and all have been conducted in the semiarid rangelands of New South Wales. Total ground cover was consistently reported to be greater under rotational grazing (Alemseged et al. 2011; Waters et al. 2017; McDonald et al. 2018). In addition, rotational grazing was reported to result in greater total standing dry matter (Alemseged et al. 2011), greater plant diversity but lower invertebrate diversity (Waters et al. 2017), and comparable or greater plant diversity and richness (McDonald et al. 2019b). No differences in soil organic carbon, landscape function, soil properties, plant evenness and turnover, or plant functional diversity were reported (Alemseged et al. 2011; Orgill et al. 2017; Waters et al. 2017; McDonald et al. 2018, 2019b). Studies in other lower-rainfall regions of Australia have also reported a range of ecological benefits from rotational grazing (Sanderman et al. 2015; Bowman et al. 2009; Sanjari et al. 2008, 2016). However, in southern New South Wales, Tupper (1978) found few differences in plant and animal responses between deferred and continuous grazing systems.

It seems axiomatic that non-continuous grazing systems that involve calendar-based movements of livestock are not suited to the highly variable rainfall environment of the southern rangelands. At the same time, rainfall variability in the region is such that continuous grazing (at appropriate stocking rates) could be sustained without ecological damage for periods considerably longer than the graze periods typical of formal rotational grazing systems. In this situation, a tactical approach to grazing should be preferable to either a specific grazing system or continuous stocking, and has been advocated by Hacker and Hodgkinson (1996), Campbell and Hacker (2000) and Fisher et al. (2005) (Principle 1). Further, the demonstration by Windh et al. (2020) that fluctuations in market conditions can outweigh production difference among grazing systems in terms of net returns cautions against commitment to any rigid grazing management philosophy.

Four key steps to tactical grazing management, as recommended by Campbell and Hacker (2000), include the following: (1) setting a management objective for each management unit (broadly, either maintenance or restoration of resource productivity, although biodiversity issues should also be considered), which may entail changing pasture composition to favour palatable perennials or reduce weeds, or increasing plant density or ground cover (Guideline 1.4); (2) determining a strategy, essentially a statement of management principles, to achieve the desired objective in each paddock, including threshold levels of pasture utilisation and ground cover, or more specific guidelines for timing, duration and frequency of rest; specification of the conditions for management burning could also be included, as could the use of landscape rehabilitation techniques such as water ponding (Guideline 1.4); (3) implementing the strategy on a day-to-day basis (i.e. tactically) in response to changing seasonal conditions and animal requirements; and (4) monitoring results to ensure objectives are being achieved, or to revise them if necessary (Guideline 1.5).

Grazing conducted according to this framework recognises that the way in which the landscape has been ‘conditioned’ will determine its response to any particular event. The levels of ground cover and soil seed pools, for example, will determine the response to a high rainfall event, and the response to a period of water deficit will be determined by the level of defoliation (Hacker et al. 2006a). It represents a ‘continuous’ rather than ‘event-driven’ model of grazing management, which Watson et al. (1996) considered more appropriate for managers seeking to maintain or restore productive landscapes, and will have the effect of promoting favourable responses to seasonal opportunities while minimising the impact of seasonal hazards.

Two types of monitoring are implied by this approach, namely, short-term monitoring of critical thresholds or other guidelines to inform tactical responses, and long-term monitoring of variables that reflect the objectives (Guideline 1.5). Bowman et al. (2009) provided an example of the successful application of this basic approach. The short- and long-term monitoring requirements would be assisted by the development of technology-based monitoring capabilities (Nielsen et al. 2020).

Implementation of tactical grazing will be assisted by the identification of local ‘trigger points’, i.e. calendar dates beyond which stocking decisions should not be delayed in the hope of improved seasonal conditions, or additional growth to support increased stocking (Guideline 1.6). Such trigger points have been identified for numerous locations in western New South Wales by Hacker et al. (2006b), but elsewhere will need to be devised from local knowledge and rainfall records.

In practice it may not always be feasible to implement the strategy in each management unit at all times. All stocking decisions have implications for the economic welfare of the business, and the seasonal or market risk to which the business is exposed, as well as the ecological condition of the land. Balancing these requirements may mean that the ‘ideal’ management response specified by the strategy is not always feasible or chosen. However, using the tactical grazing approach will allow managers to make informed decisions on how best to move towards their management goals (Campbell and Hacker 2000), and better appreciate the longer-term implications of short-term management decisions. The tactical grazing concept thus entails a risk management framework within which graziers will find their own point of balance in an informed way (Principle 1).

Tactical grazing as described above has similarities to holistic management or cell grazing, especially if herds or flocks are combined into large units and most of the landscape is unstocked at any time. Both, for example, stress the setting of management objectives and eschew calendar-based movements. However, unlike these approaches, tactical grazing does not stress the importance of animal impact or herd effect, aimed at breaking up capped soil surfaces and facilitating local infiltration of rainfall, and the intensive subdivision that it often associated with such emphasis. As outlined previously, local redistribution of resources is considered an essential feature of functional arid and semiarid ecosystems, and although animal impact may have application in some specific situations, it is not considered a fundamental management principle (Hacker 1993).

In the southern rangelands, domestic livestock graze alongside populations of native and feral herbivores which contribute substantially to the total grazing pressure (TGP). The non-domestic herbivores found almost ubiquitously throughout the region include kangaroos, ‘unmanaged goats’ (see Table 1) and rabbits (although the latter are not widespread in the arid zone of Western Australia), and, in some areas, feral pigs, donkeys, horses and dromedary camels. The widespread provision of more or less permanent stock waters, suppression of the major predator, the dingo (particularly within the ‘dog-proof’ fence), and, in some areas, favourable modification of vegetation by livestock has contributed to the increase in both native and feral animals in the southern rangelands (Hacker et al. 2019a).

At times, it has been shown that less than half the (vertebrate) herbivory in the southern rangelands (defined in this case as areas below the Tropic of Capricorn) is managed by pastoralists (Waters et al. 2018); these authors estimated that, in total, 28.93 million DSE were currently grazing the region, of which 15.57 million (approx. 54%) were macropods and goats, and 13.36 million (approx. 46%) were livestock. These figures are consistent with a national survey of landholders and natural resource management service providers who agreed that, on average, 40–50% of the total demand for forage is due to non-domestic animals, and that these levels are at least double what would be desirable from pastoral and land management perspectives (Atkinson et al. 2019). Waters et al. (2018) estimated that kangaroos represent 83% of the unmanaged grazing pressure, or ~44% of TGP.

Effective management of TGP is thus critical to sustainable management of the southern rangelands (Guideline 1.7). Where unmanaged grazing pressure has been controlled and/or long periods of rest are included in the grazing system, increases in perennial ground cover of 9–15% have been reported (Waters et al. 2017). Management of kangaroos is of central importance (a) because they represent a major component of the non-domestic grazing pressure, (b) commercial harvesting, the means of control that has the highest social acceptability, is ineffective due to the loss of markets and the actions of activist groups, (c) the task is largely beyond the capacity of individual landholders and (d) the same constraints do not apply to other non-domestic species (Hacker et al. 2019b). However, because of their unique status as protected wildlife, whose control by either commercial or non-commercial harvesting is regulated by State and Commonwealth governments, management of kangaroos is largely beyond the control (and, as noted above, capacity) of landholders. It currently represents a case of market failure because landholders acting in their own interest cannot be confidently expected to deliver outcomes consistent with public expectations (Hacker et al. 2019b).

Many pastoralists cite the tendency of kangaroos to graze preferentially in destocked paddocks (e.g. Wilson 1991b; Norbury and Norbury 1993) as a reason for maintaining continuous grazing practices, and their incursion into destocked paddocks can indeed limit the success of rangeland regeneration programs (Norbury et al. 1993). Some suggested approaches to this problem include establishing facilities such as self-mustering yards at water points to ensure that stock water supplies can be turned off when paddocks are destocked, resting areas as large as possible in an attempt to dilute the effect of any concentration, given the observation by Norbury and Norbury (1993) that kangaroo incursions into destocked paddocks are apparently from local rather than remote populations, and leaving a few sheep in a paddock to deter the influx of kangaroos (Hacker and McLeod 2003). However, neither the ‘dilution’ effect nor the effect of a low sheep stocking rate has been tested experimentally. Watson et al. (1988) demonstrated that kangaroo grazing was concentrated in the most lightly stocked paddocks in their grazing trial, particularly under poor seasonal conditions, suggesting that the latter option may not be realistic, but their study did not include comparison with an unstocked area. Closure of watering points in destocked paddocks should be practiced routinely and, although it may not necessarily lead to a reduction in kangaroo grazing pressure from the resident population, it may help deter increases due to incursion (Freudenberger and Hacker 1997). Temporary exclusion of kangaroos from sheep watering points by means of a low-lying electrified wire (the Finlayson trough, Norbury 1992) has potential to promote short-term concentration of kangaroos around the water, which may facilitate local harvesting or culling as an aid to paddock resting (Hacker and Freudenberger 1997).

Management of other non-domestic species presents fewer issues for landholders, provided the control methods are humane (Sinclair et al. 2019a, 2019b; Guideline 1.7). Some have been subject to ongoing or periodic control programs that are in part publicly funded (e.g. rabbits; camels, Hart and Edwards 2016); trapping/mustering and sale of unmanaged goats is both socially acceptable and can represent a significant source of income for landholders in some parts of the region (Khairo et al. 2011; Guideline 1.7).

Since management of TGP is such a fundamental issue for pastoralists in the southern rangelands, it is important that the equivalence of the various species on a relevant scale be understood. Estimates of the DSE ratings of most species (in various physiological states) are readily available, and are not repeated here, and although these are usually based on maintenance energy requirements, there is generally a reasonable correspondence between ratings on this scale and forage intake, and therefore contribution to TGP. However, for kangaroos, there has been considerable controversy about the appropriate rating. Pahl (2019b) showed that a rating of 1 DSE for a 50 kg kangaroo is more appropriate for assessing their contribution to total forage consumption than is the rating of 0.5, on the basis of basal metabolic rate or energy expended in grazing, widely accepted in recent years (Guideline 1.1).

Waters et al. (2019, appendix 2) described both ‘exclusion fencing’ and ‘TGP fencing’. The former is aimed primarily at the exclusion of wild dogs, although with simultaneous impact on kangaroos, and is predominantly used to surround large areas or several properties in a ‘cluster’; the latter is aimed at partial exclusion of unmanaged goats, kangaroos and wild dogs and is used for boundary or internal fences on individual properties. Interest in the establishment of both types of fence, and other fence designs aimed at deterring goats and kangaroos, has increased considerably in recent years and is indicative of the serious impact that wild dogs and non-domestic herbivores are perceived to have on pastoral enterprises, particularly sheep operations. Interest in these developments has also been stimulated by the availability of incentive funding from State and Commonwealth governments and, for cluster fences, establishment of formal agreements to ensure fence maintenance. Waters et al. (2019) cited increases in gross margin of up to 345% attributable to cluster fencing, but acknowledged that there is uncertainty about the biophysical and economic impacts of these investments. Given the range of fencing designs available, and establishment costs of up to A$15 000 km⁻¹, it is important that careful consideration be given to the likely cost-effectiveness of new investment (Principle 2; Guideline 2.1).

Where new internal fences are established, the advice generally given by extension agencies is to fence to land type as much as possible (Guideline 2.2). Given the scale of pastoral operations in the southern rangelands, this concept would probably not entirely deprive animals of the benefits of functional heterogeneity noted above, and is supported by the common observation that small areas of preferred land types fenced with large areas of less preferred types are often severely degraded.

The size of paddocks is the subject of considerable debate but there is little objective guidance that is specific to the region. In the chenopod shrublands of South Australia, one model of development based on subdivision into paddocks of ~2000 ha, each with a permanent watering point and grazed (continuously) by flocks of 250–350 sheep, has provided profitable pastoralism and well preserved shrublands (Lange et al. 1984), although even here some long-term loss of shrubs has occurred (Andrew and Lange 1986). Smaller paddocks could be expected to provide more uniform utilisation of the landscape, and this mechanism probably accounts for much of the increase in carrying capacity claimed for holistic management, with its intensive subdivision, in commercial situations (Norton 2003). Some pastoralists would advocate that paddocks should simply be small enough to allow mustering and stock handling to be completed in one day.

A major trend in recent decades has been the development of self-mustering facilities, or trap yards, at watering points, initially in Western Australia, but subsequently throughout the eastern states (see e.g. Connelly et al. 2000). These facilities offer not only a cost-effective means of mustering livestock where surface water is unavailable, but also a means of trapping unmanaged goats and restricting access to water by kangaroos when paddocks are destocked (Guideline 2.3). Although closure of the traps should ideally exclude kangaroos from water entirely, individuals do sometimes gain entry and present an animal welfare issue if they cannot exit; this issue needs to be addressed, either by trap design or management. As with fencing, the implementation of trap yards has been substantially assisted in recent years by the availability of incentive funding from regional natural resource management organisations.

Throughout much of the grazed southern rangelands, there is little need for the development of additional livestock watering points, apart from what may be required to allow subdivision for more intensive grazing systems. The proliferation of artificial waters in the region has been noted above so that few areas are beyond the grazing range of livestock and grazing distribution is not seriously restricted by limited water supplies (Fensham and Fairfax 2008), even in those areas where saline water limits the grazing range of animals. The situation is such that a reduction in the number of watering points, creating water-remote areas, could be desirable from the perspective of biodiversity conservation (Guideline 4.1).

Self-herding, whereby attractants and positive reinforcement are used to encourage animals to move around landscapes without the need for extensive subdivisional fencing, has recently been employed by some pastoralists in the southern rangelands to manage livestock distribution and provide rest to areas as needed (Revell et al. 2015). Unfortunately, the hoped-for developments in virtual fencing (Anderson 2007) have not materialised to date on a commercial scale.

As noted above, thickening of native woody vegetation, formerly called ‘woody weeds’ and currently ‘INS’, has been a feature of the eastern part of the southern rangelands, particularly the Mulgalands of south-western Queensland and north-western New South Wales, and the Cobar Peneplain in western New South Wales. Elsewhere in the region, changes in the composition and density of mid-storey shrub species have been observed under pastoral use (e.g. Watson and Novelly 2012, in the arid zone of Western Australia), or degraded rangelands have been invaded by woody exotic species (e.g. Acacia nilotica in the north-east of the region). However, our interest here is in the impact of INS in the higher-rainfall environments in the east of the region where the issue was reported as early as the 1870s (Noble undated, p. 8) and has continued to be of major concern to landholders. The nature and management of this problem have been the subjects of extensive research from the 1960s to the 1990s by CSIRO and state agencies in New South Wales and Queensland (Noble undated, pp. 6–7). The major outcomes, and guides to practical management, are summarised in Noble (undated), CWCMA and WCMA (2010) and WLLS (2019).

The extensive experience of both researchers and landholders has shown that there is no one-off solution to encroachment of INS (Principle 3). An overall approach will necessarily involve control of TGP, grazing management that maintains competitive perennial grasses under varying seasonal conditions, and treatment of regrowth. Monitoring the presence of INS seedlings when seasonal conditions are conducive to establishment will be an essential feature of this process (Guideline 3.1). Planning to address all of these components should be part of both short- and medium–long-term property management planning (Principle 3). Because open areas provide the highest economic return, and the cost of maintaining them is much lower than the cost of restoring encroached areas, such maintenance should always be the first priority of INS management (Guideline 3.2). This priority is emphasised by the fact that only small increases in shrub cover are sufficient to reduce forage production, which declines rapidly as shrub cover increases (Guideline 3.2). Remaining areas should be prioritised for treatment based on the expected costs and benefits, and the ease of incorporating INS management into overall property management (Guideline 3.3).

Management of INS at property scale will involve production of an appropriate property plan on which the areas designated for treatment can be identified and prioritised, determining whether a paddock should be retained for grazing or used to obtain income via carbon credits (Guideline 3.4), and development of an appropriate post-treatment strategy (e.g. in relation to TGP and grazing management) to ensure that the anticipated benefits of any investment can be realised. Because it is difficult to provide accurate information regarding the cost of an INS control program in a specific situation, a simple net present value approach has been proposed (WLLS 2019) to estimate the economic benefit of a successful treatment program and, thus, an estimate of the maximum amount that could be reasonably invested (Guideline 3.3).

It is notable that although management burning is by far the cheapest control option available (WLLS 2019), and is effective against seedlings of all invasive species, even though some are tolerant of fire as adults (Hodgkinson and Harington 1985), it has not been widely adopted. Reasons include the reluctance of graziers to use the high levels of biomass produced only on decadal time scales for fuel rather than forage, the risk of erosion of bare soil surfaces, concerns (largely unfounded) that fire may stimulate the germination of more shrubs or ‘bake’ the soil, fears of losing control of a fire, and the limited manpower now available in many areas to undertake a well planned burn. The extent to which Indigenous traditional knowledge could be incorporated into the use of fire for INS management (see Nielsen et al. 2020) seems not to have been substantially explored. Notably though, Purvis (1986) provided an example from central Australia of the effective use of small-scale, opportunistic burning over the long-term for INS management, effectively restoring Aboriginal practices.

While shrub encroachment is generally viewed as negatively affecting pastoral production, there is also a growing body of evidence that some ecosystem services are not necessarily affected, and may be enhanced, by encroachment so that no single state (including grassland without shrubs) will maximise all services (Eldridge and Soliveres 2014). These authors emphasised that the effects of shrubs on ecosystems are strongly scale-, species- and environment-dependent. They considered that overgrazing is the primary determinant of ecosystem degradation rather than encroachment per se, although from a pastoral perspective, we would argue that encroachment can be both a symptom of land degradation (when it results in part from the reduction of perennial grass competition) and a cause (when seasonal conditions permit the encroachment of shrubs into otherwise healthy grassland). However, we would not dispute their view that in some situations at least, money would be better invested in projects to maximise carbon sequestration, enhance biodiversity, or improve soil fertility and conservation rather than in shrub removal.

There is a growing awareness among pastoralists and industry organisations of the requirement for demonstrable environmental stewardship and animal welfare to underpin the social licence and sustainability of the pastoral industry (Waters et al. 2019), and that landholders’ duty of care extends beyond the maintenance of pastoral values (Principle 4). Similarly, as the need for global food security increases, livestock grazing management that promotes biodiversity values may offer an alternative approach for broad-scale conservation while achieving dual production and ecological outcomes (Dorrough et al. 2004). In global terms, much of Australia’s rangeland, both grazed and ungrazed, offers a chance to maintain a vast functioning environment that has been forfeited in many overseas rangeland areas (Woinarski and Fisher 2003).

Silcock and Fensham (2019), in their extensive review of the impacts of 160 years of pastoralism, concluded that conservation of plant biodiversity seems largely compatible with commercial pastoralism, at least in inland eastern Australia. This is consistent with other studies that have suggested that careful grazing management can achieve dual conservation and production outcomes and play a role in increasing biodiversity in agricultural landscapes (Curry and Hacker 1990; Dorrough et al. 2004; Fensham et al. 2011, 2014; Savory 2013; Silcock and Fensham 2013; McDonald et al. 2019b).

Domestic livestock (sheep, cattle) are sometimes considered to have a greater impact on biodiversity than other native and introduced species. Tiver and Andrew (1997), for example, concluded that, at existing levels of herbivory, sheep were the most important vertebrate herbivore affecting regeneration of 18 tree and shrub species in South Australia, and that the impacts of non-domestic vertebrates (rabbits, goats and kangaroos) were much less important. Similarly, Eldridge et al. (2018) reported little effect of kangaroo grazing on biodiversity, even at high densities (Guideline 1.4). However, ‘competition and land degradation by unmanaged goats’ is listed as a key threatening process under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999. Goats compete with native fauna, such as the yellow-footed rock wallaby, for food, water and shelter, and contribute to other degradation processes (Department of Sustainability, Environment, Water, Population and Communities 2011), so that goat population management should be undertaken, particularly within sensitive areas (Guideline 4.2).

Where specific habitats that are important for biodiversity occur on pastoral leases, landholders arguably have a greater duty of care to exercise stewardship of them (Guideline 4.2). This might apply, for example, to mound springs, yellow footed rock wallaby habitat, or other specific areas that Morton et al. (1995) considered should be established as ‘excised management units’ to protect non-pastoral values. Such stewardship need not impose substantial additional costs on landholders. Removal of unmanaged goats, for example, can be an important source of income (Khairo et al. 2011, 2013).

Given the ‘catastrophic’ record of mammal extinctions in the arid and semiarid zones attributable primarily to introduced cats and foxes, as noted above, attempts to control this impact should be beneficial for biodiversity conservation in pastoral landscapes. Control of foxes, in particular, can also be expected to provide an economic benefit in areas where predation on domestic flocks (particularly lambing ewes) is significant (Jones et al. (2006); Guideline 4.3). There has been considerable interest in recent years in the potential to control these exotic predators and facilitate biodiversity conservation by positive management of dingoes as a top predator. Moseby et al. (2012) provided evidence that dingoes can suppress both cats and foxes, particularly the latter, within a securely fenced ‘dingo paddock,’ consistent with much anecdotal evidence. However, reviews of the potential for positive management of dingoes to promote biodiversity conservation warn strongly that evidence to support this approach is inconclusive, and that adverse consequences, both ecological and economic, could result (Allen et al. 2012, 2013; Fleming et al. 2012).

Reviews of the impact of alternative grazing systems on biodiversity provide no clear evidence for the advantage of one system over another (e.g. Dorrough et al. 2004; McDonald et al. 2019a; Waters et al. 2019). Rotational grazing systems are believed to favour biodiversity because they reduce selective grazing and allow for recovery of palatable species between grazing events (Norton 1998; Teague et al. 2008). However, few studies examining biodiversity under contrasting grazing systems have been undertaken in the southern rangelands. In western New South Wales, biodiversity under rotational grazing systems has been reported to be comparable to ungrazed areas, and at times greater than in continuously grazed areas (Waters et al. 2017; McDonald et al. 2019a, 2020).

We have argued above that grazing should be managed in a tactical grazing framework, which recognises that seasonal and/or economic conditions may not always permit grazing to be managed according to the strategy originally defined to achieve the objective for a particular management unit. When compromises are required, the implications for biodiversity outcomes, as well as short- and long-term production outcomes, should be considered (Guideline 4.4).

The impact of grazing on biodiversity will vary depending on location, herbivore type, grazing intensity and grazing system. While some authors have considered the impact of grazing on biodiversity in the southern rangelands to be generally negative (Eldridge et al. 2016; Eldridge and Delgado-Baquerizo 2017), we would argue on the basis of the above discussion that well managed grazing can be compatible with conservation of biodiversity in the region. Further, since markets for carbon sequestration, biodiversity conservation and other ecosystem services are now either established or are emerging, landholders should have the opportunity to obtain financial reward from management that delivers public goods beyond what is reasonably required by their duty of care (Guideline 4.5).

Questions about the viability of pastoralism in the southern rangelands have long been raised, reflected in efforts to classify bioregions, in terms of their suitability for pastoralism, using axes of ‘likely pastoral benefit’ and ‘likely risk to public values’ (Stafford-Smith et al. 2000), to assess the quality of pastoral management (Pickup and Stafford-Smith 1993), and to encourage administrative provisions to assure conservation of non-pastoral values (Morton et al. 1995). We have also noted the operation of several publicly funded programs intended to achieve a re-vitalisation of the pastoral industry in the region, and their limited success. Several other substantial projects have undertaken a consultative approach to defining a desirable future for various parts of the region, but without any marked, lasting influence (e.g. DAWA 2002; Project 21C in New South Wales).

We accept that there are parts of the southern rangelands the future of which does not lie in the continuation of pastoralism, either in the traditional form or with the implementation of more progressive approaches to grazing and natural resource management. However, we are not able, in this paper, to provide a detailed account of the likely distribution of these areas beyond the generalisation that they are most likely to occur at the lower end of the productivity spectrum identified previously, such as in the arid zone of Western Australia where many leases are considered non-viable, or in areas with a high conservation value (van Etten 2013). We anticipate that the emergence of alternative land uses, as discussed previously, will be an important factor in providing a market-based solution to the allocation of land resources within the region, although how this may function could vary among jurisdictions. However, we also acknowledge that the market alone may not be sufficient to ensure that land use is appropriately matched with land condition/capability. A role for government will remain, particularly at State and Territory levels, in ensuring that policy settings and financial assistance facilitate a rational, market-based allocation of land to the most appropriate use and, where market failure is evident, that last resort measures are available to remove land from the pastoral domain.

We would argue that, at one level, the pastoral industry overall in the southern rangelands has never been better placed to achieve sustainability than at the present time. This situation arises in part from the public investment made over the past two decades in a range of regional programs aimed at revitalising and restructuring the industry. Although these have, by no means, achieved all of their stated objectives, they have improved the potential for sustainable management by, for example, supporting investment in infrastructure that allows better control of both domestic and non-domestic herbivores, and assisting the development of some emerging industries such as on-farm tourism and rangeland goats (Table 1; e.g. URS 2015). Current technological developments also support economically efficient pastoral production in ways not previously available (e.g. drones for remote monitoring of livestock and infrastructure, ‘walk-over’ weighing, farm and grazing management software packages, Internet-based services for both market and seasonal data and forecasts), even if further developments are desirable to tailor the application of available technology to specific property situations. Furthermore, there is a growing awareness, as noted above, of the need for demonstrable environmental stewardship. We, therefore, consider that pastoralism conducted in accordance with the principles and guidelines outlined above has good prospects for ecological, economic and social sustainability in those areas where it continues. Nevertheless, the application of these principles/guidelines may still be constrained by many factors beyond the control of individual landholders, some of which are listed below.

The management of kangaroos is a key component of ecologically sustainable management in the region because they represent a significant component of TGP, their control by commercial harvesting has been adversely affected in recent years by loss of markets and stakeholder activism, and their control is largely beyond the capacity of individual landholders (Hacker et al. 2019b). These characteristics distinguish them from other non-domestic herbivores which landholders can more readily control. We consider that the market failure represented by the present kangaroo management arrangements, whereby neither the kangaroo industry nor landholders, acting in their own interests, can deliver outcomes for natural resource management and animal welfare that are socially acceptable, urgently requires a paradigm shift in kangaroo management. We are unable to argue in detail here the possible nature of this shift but note that several suggestions have been made in recent times, including the adoption of an active adaptive management approach by wildlife management authorities (McLeod and Hacker 2019) and the granting to landholders of a form of proprietorship or custodianship over kangaroos (Wilson and Edwards 2019; Wilson et al. 2020). While other non-domestic herbivores (e.g. camels, pigs and goats) represent a lesser, or more localised, challenge for management of TGP, we concur with Nielsen et al. (2020) that their management would be assisted by the establishment or further development of market-based approaches, and development of appropriate product branding.

The likely impacts of climate change on the region have been outlined above. Marshall (2015) considered that only 16% of beef producers in northern Australia had sufficient adaptive capacity to cope with climate change. We would expect that a similar situation would apply to the southern rangelands. Although a considerable range of financial instruments (e.g. Farm Management Deposits) and technical information is available to assist producers manage climatic extremes, the particular situation of the individual business, and the particular sequence of climatic events, may easily override the best attempts to manage according to the principles and guidelines given above. Climate change can thus be expected to increase the difficulty of achieving sustainable land use across the region. It also highlights the desirability of integrating rangeland enterprises with operations in less variable environments to improve the overall resilience of businesses.

Since trends in land condition are variable in time and space, feedback from monitoring programs to policy makers will need to be highly granular to be effective. Policies that seek to influence land condition will need to be guided by monitoring feedback at a scale relevant to regional management, and not by broad generalisations or perceptions. Nielsen et al. (2020, p. 364) noted a similar need for spatial and temporal flexibility in policy application, particularly in relation to drought. Failure of policy settings to recognise the particular requirements of local contexts could be regarded as a form of sovereign risk for individual landholders. It would represent a failure to implement the adaptive model of natural resource governance that Marshall and Stafford-Smith (2010) considered most appropriate for remote areas with highly variable environments, such as the southern rangelands, and may impede the mobilisation of local enthusiasms and skills in pursuit of sustainable pastoral land use across the region.

Resources available for R, D and E in the region have always been low and have declined substantially over recent decades (Hacker 2013), consistent with the declining importance of the southern rangelands in the overall pastoral economy, relative to the more productive northern savannas (Foran et al. 2019). It is unlikely that this situation will be reversed unless some of the initiatives proposed by Nielsen et al. (2020) come to fruition, such as establishment of an Outback Commission and/or an Outback Capital Trust Fund to divert a portion of royalties from mining in the rangelands to R, D and E. Although information gaps can be readily identified (e.g. effects of different grazing strategies/management practices on biodiversity, Dorrough et al. 2004; management requirements of grazing sensitive species, Silcock and Fensham 2019; effect of animal impact on the functioning of arid and semiarid ecosystems, Hacker 1993), the capacity to address these through critical research will be strictly limited. Less conventional forms of R, D and E will be required, based around approaches such as adaptive management, action learning and exploitation of natural experiments which involve stakeholders, including Indigenous people, in the design, implementation, interpretation and adoption of the outcomes. Establishment of ‘innovation hubs’ sensu Nielsen et al. (2020) would also be beneficial.

Carbon farming has considerable potential to contribute to increased socio-ecological resilience in the southern rangelands, because, in addition to the primary objective of carbon sequestration, it can also deliver co-benefits such as diversification of land uses and income streams, a range of consequential community benefits, and ecological benefits including increased biodiversity and habitat provision, improvements to soil health, structure and water holding capacity, management of erosion and salinity, and water quality (Baumber et al. 2020). However, these authors also noted the potential for negative impacts arising from policy uncertainty, loss of future land use flexibility, possible social divisions arising from carbon farming eligibility criteria, and landholder absenteeism. Additionally, ecological risks may be associated with possible exacerbation of INS encroachment, and simplification of diverse ecosystems. Given that the legislated focus of the Emissions Reduction Fund (ERF) is least-cost carbon abatement, it could produce perverse outcomes unless complementary incentive schemes are introduced to value co-benefits (e.g. the Queensland Government’s co-benefits standards for assessing carbon farming projects under its Land Restoration Fund), or eligibility rules for ERF projects are modified (Baumber et al. 2020).

We are not aware of any potential for perverse outcomes that would attend the use of land for provision of ecosystem services or renewable energy generation.

Hunt et al. (2014) expressed the requirements for grazing management of Australia’s northern rangelands in terms of four principles and 13 guidelines, compared with four principles and 19 guidelines discussed above and listed in Table 3. However, their principles and guidelines related specifically to the four topics of stocking rate, pasture resting, prescribed fire and paddock size/water distribution. The first three of these topics are subsumed, in our summation, under the philosophy of tactical grazing (Principle 1), whereas we place more explicit emphasis on development of infrastructure for management of TGP (Principle 2) and conservation of biodiversity within pastoral production systems (Principle 4). Although there is considerable overlap between the two summations, particularly at the guideline level, these differences reflect both the broad ecological contrasts between the two regions (e.g. the abundance of kangaroos and other non-domestic grazers in the southern rangelands, and the limited pastoral potential of much of this region), together with our explicit attempt to recognise the increasing societal preference for non-pastoral values.

Stakeholders in Australia’s rangelands aspire to ‘resilient and sustainable rangelands that provide cultural, societal, environmental and economic outcomes simultaneously’ (Nielsen et al. 2020). We have attempted to elucidate the more specific requirements of this vision in relation to the southern rangelands.

We acknowledge that achievement of this vision will almost certainly involve some continued withdrawal of land from pastoral use. Where pastoralism continues, we consider that sustainable land use can be achieved by application of four management principles and 19 associated management guidelines. These management principles are (1) manage grazing in a tactical risk management framework, (2) develop infrastructure to allow best management of both domestic and non-domestic grazing pressure, (3) incorporate INS management, where required, into overall property management on an ongoing basis and (4) manage grazing to enhance biodiversity conservation at landscape scale.

Due to the partial success of several publicly funded regional reconstruction programs, the pastoral industry is today probably better equipped than previously to implement these principles and guidelines. However, sustainable land use under pastoralism will still be subject to constraints not under the control of individual landholders and will require the development of appropriate policy settings, particularly in relation to kangaroo management, climate change, and governance arrangements, as well as innovative approaches to R, D and E.

The future pastoral industry will co-exist with the new industry of carbon sequestration. This will infuse large amounts of capital into those areas that are able to satisfy the requirements of the protocols, providing a range of socio-ecological benefits but with the potential for perverse outcomes unless policy initiatives appropriately value the co-benefits which the industry can provide. Industries that utilise the major renewable energy resources of the region, or are based on provision of ecosystem services, would seem to have considerable potential. They are considered unlikely to have adverse ecological consequences, although the former may be limited by the remoteness of much of the area.

The authors declare no conflict of interest.

Data used in the paper are not available in a public repository but may be obtained from the corresponding author.

This research did not receive any specific funding.