Table 3. Information products for impacts on biophysical resources.

Product 5. Climate variability

Rainfall and its effectiveness are major drivers of processes and functions in rangelands (e.g. drought can lead to wind erosion and requires change in animal populations and land management decisions).

Rainfall in Australia's rangelands varies from one year to the next (Figure 19), from season to season and spatially (Figure 20). Climate variability can lead to degradation:

• Years with above average rainfall may support increased stock numbers. Retention of high stocking rates in subsequent 'normal' or drought periods may lead to degradation.
• Years with below-average or little effective rainfall may lead to grazing pressures that degrade the rangeland resource.
• Years with above average rainfall may lead to changed rangeland condition (regeneration of desirable vegetation, opportunities for burning) and rangeland management problems (woody weed infestation, increases in feral and native herbivores).

Climate science has documented historical climate variability associated with global fluctuations of sea surface temperatures and atmospheric variables. The time-scales of these fluctuations range from annual to about every ten years (e.g. in rangelands of eastern Australia extreme droughts and floods have been associated with the interaction of the El Niño-Southern Oscillation phenomenon [Figure 21] and the Inter-decadal Pacific Oscillation). Oscillations in sea surface temperatures of which these are examples also provide some explanation of the likely causes of historical sequences of dry or wet years. This new understanding of the influence of climate in rangelands is leading to the development of improved climate forecasting systems and monitoring. Information provision will ensure better management decisions.

The Bureau of Meteorology routinely provides seasonal climate forecasts for both rainfall and temperature (Figure 22). Climate forecasting work is also under way in Queensland and Western Australian agencies, and will take much of the guess work out of assessments for exceptional circumstances.

Climate forecasts are described as chances and probabilities because the chaotic nature of the climate system means that precise predictions are not possible. Rainfall and other climate-based forecasts can be used in rangeland management as direct tools in forward planning and as inputs to models of pasture availability (e.g. Aussie GRASS). Longer-term forecast information is becoming available from coupled atmosphere-ocean (and other) models (e.g. see http://www.bom.gov.au/climate, http://www.bom.gov.au/silo and http://www.longpaddock.qld.gov.au/). Information products displaying and modelling climatic variability are routinely collated and made available by the Bureau of Meteorology.

Product 6. Predicting pasture availability

Aussie GRASS is a simulation model incorporating the complex interaction of climate, soils, vegetation, fire, animal numbers and management responses and is used to:

• simulate grass production;
• provide both monitoring and forecasts of potential grass growth and cover;
• estimate historical and projected grazing pressure;
• assess risks of degradation; and
• compare current conditions and opportunities for improved stock management with past situations.

Aussie GRASS information products display and model availability of herbage biomass (Figures 23 & 24). They are routinely collated and made publicly available by the Queensland Department of Natural Resources and Mines on behalf of a partnership across State and Territory agencies.

Product 7. Seasonal characteristics and influence on vegetation

Rangeland condition needs to be interpreted in a seasonal context. An increase in photosynthetic activity or greenness after rainfall is an indicator of season quality. Change in greenness is estimated using the Normalised Difference Vegetation Index. An increase in Normalised Difference Vegetation Index in response to vegetation growth will depend on the amount, structure and composition of vegetation present in an area. Comparing each area to itself over time gives a good indication of relative changes in herbage. A relative rating of season quality can be mapped by comparing a particular year with all years recorded (data starts in 1991/92). This provides a context for the interpretation of finer scale Landsat-derived assessments and interpolation of data collected from ground monitoring plots. It also provides information on the scale and extent of wet or dry periods.

Advantages of Normalised Difference Vegetation Index data over rainfall data include the ability to:

• provide estimates on 1 km² areas, rather than interpolated rainfall data from a limited number of stations, and;
• estimate the response of vegetation to climate including rainfall and evaporation rates, rather than simply estimating the rainfall amount.

The Normalised Difference Vegetation Index provides an estimate of maximum and minimum greenness in any given year (Figures 25a & 25b). The difference between maximum and minimum in any given year is called the flush (Figure 25c).

Figure 25a. Maximum greenness (January 2000 to December 2000) as estimated by the Normalised Difference Vegetation Index.

Figure 25b. Minimum greenness (June 1999 to July 2000) as estimated by the Normalised Difference Vegetation Index.

Product 8. Total grazing density

Products collated by the Audit for total grazing density include historical and current estimates of domestic stock (sheep and cattle), kangaroos and some feral animals (goats and rabbits). They will help understand the pressures on rangeland flora and habitat and allow trends to be determined.

Data collation activities required to complete the analysis were:

• collation of Australian Bureau of Statistics historical domestic stock information from 1956 to present for statistical local areas (available through the Atlas and Data Library);
• collation of historical and simulated data on macropod and feral animal numbers from the 1950s to present day (data on decadal time-steps available in the Atlas and Data Library);

Stock density

The only national, regular coverage of stock density is available from the Australian Bureau of Statistics agricultural census and survey data. Stock density has been compiled annually by statistical local area since 1956 (except for South Australia where data were only available as Hundreds and Counties prior to 1983) (Appendix 2). Stock included in the final database are beef bulls, beef heifers, beef calves, dairy cattle, rams, ewes, wethers, lambs and horses. The reliability of these data has been questioned (e.g. Mortiss 1995) with suggestions that the figures are likely to be underestimates.

Beef cattle density increased in Queensland, New South Wales, South Australia and the Northern Territory in the mid- to late-1970s; in this period, sheep density fell in all States. Sheep density peaked again in the early 1990s. Cattle density increased by 50% across Australia from 1956 to 1999 while sheep density fell to half of what it was in the 1950s (Figures 29, 30 & 31).

Figure 30. Cattle density for Australia's rangelands by statistical local area (1950s, 1960s, 1970s, 1980s, 1990s).

Figure 31. Sheep density for Australia's rangelands by statistical local area (1950s, 1960s, 1970s, 1980s, 1990s).

Generally sheep are found south of the dingo-proof fence which runs from Yalata near the Great Australian Bight, north to Coober Peedy, across to Tibooburra in New South Wales and across Queensland.

Figure 32. Kangaroo (Macropus rufus, M. fuliginosus and M. giganteus) density for Australia's rangelands (1950s, 1960s, 1970s, 1980s, 1990s).

The modelled data suggest that kangaroo numbers are erratic and coincide with rainfall and available feed.

Figure 33. Total density of feral goats and rabbits for Australia's rangelands (1950s, 1960s, 1970s, 1980s, 1990s).*

* Feral goat and rabbit maps have been compiled using data from a variety of sources, in conjunction with simulations and extrapolation of data based on factors such as the assessment of seasonal conditions as measured by rainfall and pasture growth. The scarcity of both time-series and spatial data on goats and rabbits means that, although they are the best available, the maps are indicative only.

Goats were introduced to Australia in 1788. By 1993 an estimated 2.6 million feral goats were spread across the country (Figure 34). They prefer high protein feed and green annual plants when available. They will eat shrubs and trees in dry conditions and will eat a wider variety of plants than sheep and cattle. Goats have a dramatic impact on ecosystems that have evolved without browsing animals. Control is difficult due to their high mobility and high reproduction rate.

Only limited data on the distribution and density of goats across Australia's rangelands were available before the mid-1970s. Available evidence and reports indicate that goats were widely distributed as domestic herds may subsequently become feral. From the mid-1970s, statistics and maps were produced for each of the States where feral goats were found.

Product 9. Fire extent, timing and frequency

Fire has shaped much of the vegetation and ecology of the rangelands and is an integral part of rangeland management. The frequency of fires used by Indigenous people to hunt and manage vegetation sometimes changed vegetation types (e.g. open savanna replaced open forest). European settlement and grazing have led to a generally lower frequency of burning and less fuel in the understorey. In semi-arid areas, woody weeds (both native and exotic) have become a major problem that needs to be controlled by fire and/or grazing.

The Tropical Savannas Management Cooperative Research Centre-among other research institutions-has researched the effects of fire on ecosystems and biodiversity (http://www.savanna.ntu.edu.au/). Frequent fires and fires late in the dry season are the most damaging to ecosystems and biodiversity. Australia-wide fire monitoring-beginning in 1997 and ending in 1999-provided information on location, timing and frequency of fires. The total area burnt from 1998 to 2000 represents 13% of the continent. Fires were started by lightning strikes and deliberately (e.g. control burns for hazard reduction, pasture management, Indigenous cultural reasons or for biodiversity objectives).

Remote sensing allows managers to view large areas of the rangelands and to track fire activity in real time, enhancing the ability to manage the effects of fire and assess impacts over time. The Western Australian Department of Land Administration has been conducting real-time fire monitoring of the Kimberley region since 1993 using NOAA-AVHRR satellite thermal signals from night images (Figure 36). These images provide 'hot spot' base data for verification with maps of visible burnt areas and ground truthing.

Annual assessments of fire-affected areas need to be continued as part of a rangeland monitoring program.

Product 10. Land tenure

An understanding of land tenure and how it has changed over time provides a basis for evaluating land use impacts (Figure 37). Land management and administration have been integral to the Australian landscape since its first human occupation.

• Indigenous people managed the land using fire and selective harvesting and developed complex systems for the administration of land through tribal lore and the 'dreaming' (the Indigenous system of beliefs, morals, family and the afterlife).
• The arrival of Europeans saw development of land management and administration systems that were thought to be the most appropriate at the time but lacked understanding of ecological factors that interplay on the Australian landscape (Childs 2000).

The hierarchical 'property rights' system has resulted in land held under a variety of tenures (e.g. freehold, Crown leasehold and unallocated Crown land) each with differing land use covenants. Lease tenures have evolved prescribing dominant and sometimes exclusive land use types (Holmes 2000) (Table 4).

Covenants on pastoral leasehold land generally limit use to grazing activity, with access to the public provided certain conditions are met. However, other stakeholders (e.g. those involved in mining, tourism and agriculture) may also have a vested interest in the land. The case for flexibility is strongest on marginal lands where the economic returns from pastoralism are lower (Holmes 2000). A system of flexible use would require increased responsibility from governments to ensure coordinated administration, and for land administration to be an extension of public policy.

The Audit collated changes in land tenure for the 1950s to 1999 as context information to assess trends in rangelands use and management.

Key Audit findings are:

• Land set aside for nature conservation purposes has increased more than fifteen-fold since the 1950s from 29 100 km² to 441 200 km² (7.8% of the total rangelands area) (Figure 38).
• In the 1950s, land reserved for Indigenous use and benefit (covering a variety of titles but no Indigenous groups actually owned land) was 347 200 km². In 1999, Indigenous-held land and land reserved for Indigenous use and benefit was 925 200 km² (16% of the total rangelands area and an increase of about 2.5 times) (Figure 39).
• Total freehold and leasehold land has remained substantially the same-approximately 57% of the total rangelands area. The majority of these lands are leasehold. Nature conservation and Native Title holdings come principally from unallocated lands (Figure 40).

The full data set is available in Appendix 3.

Product 11. Introduced plants and animals

Weeds

More than 3000 exotic plant species cause billions of dollars worth of damage each year to Australia's productive capacity and natural resources (National Weeds Strategy Executive Committee 2000). A detailed breakdown of costs for the rangelands is not available. Invasive weeds displace native species and some are unpalatable or poisonous to livestock. Production from rangelands has a marketing edge in that pastoralists are able to be certified organic because historically agricultural chemicals have not been used. Conflict arises over weed control since chemical control over large areas affects organic status of graziers.

The National Weeds Strategy is concerned with managing priority weeds that pose threats to primary industries, land management, human and animal welfare, biodiversity, and conservation values. It has listed 20 weeds of national significance; a full list is available on the National Weeds Strategy website (http://www.weeds.org.au ). Four species that affect rangelands are athel pine, mesquite, prickly acacia and parkinsonia.

Feral animals

The major introduced species affecting rangelands-goat, rabbit, pig, buffalo, donkey, camel, horse, cat, fox and cane toad-now make up over 10% of Australia's fauna. Impacts on production include competition with livestock for food and shelter, predation on stock, land degradation (especially in localised areas of high feral population), and spread of diseases. Impacts on biodiversity include predation, competition for food and shelter, and displacement of native species.

The National Feral Animal Control Program aims to reduce the damage to agriculture and the environment caused by feral animals. It is administered by the Bureau of Rural Sciences and the Biodiversity Group of Environment Australia.

Domestic pigs were first introduced into Australia in 1788 to provide food for early settlers. Feral pigs are Australia's most popular game animal and the associated meat industry is worth $10 m to $20 m annually. Pigs have a varied diet: they prefer tender green vegetation, fruit and grain, but also eat rodents, lizards, frogs and insects. Pigs often prefer wetter areas and cause most damage to habitat in wetlands, marshes and watercourses (Figure 46). They are partially responsible for spreading the seeds of exotic plant species (e.g. Mimosa pigra).

Product 12. Native vegetation clearing

Change in the extent of native vegetation indicates loss of habitat and is a key part of biodiversity monitoring and assessment (see Audit project: Developing an adaptive framework for monitoring biodiversity in rangelands available on the Atlas).

Australia-wide change in the extent of native vegetation has not yet been compiled. The Australian Greenhouse Office has Landsat data (as part of the National Carbon Accounting System) that will provide an Australia-wide assessment of vegetation change. Linking these data to the Audit's National Vegetation Information System will provide information on types and loss of native vegetation.