Australian Natural Resources Atlas

Natural Resource Topics

Rangelands — tracking changes 2001

Australian Collaborative Rangeland Information System
National Land and Water Resources Audit, 2001
ISBN: 0 642 37114 8

Lake Eyre Dragon. Photo: Drusilla Patkin

Appendix 1. State Monitoring Activities

Rangeland monitoring, range condition assessment, resource inventory and other activities

The following section is a summary of the rangeland monitoring, range condition assessment, resource inventory and other activities for each rangeland State and the Northern Territory. More detail can be found in Anderson et al. (in press), Gould et al. (2001), Green et al. (2001), Karfs et al. (2001), and Watson et al. (2001).

New South Wales

Rangeland Assessment Program

The Department of Land and Water Conservation manages the Rangeland Assessment Program, the most advanced of the range monitoring programs in Australia. Almost 340 ground-based sites within seven range types have been recorded annually since the early 1990s (Table A1). Its emphasis is on reporting changes in both trend and condition of vegetation and soil attributes at a regional scale. Attributes assessed at each site include:

The density of perennial chenopods and percent canopy cover of trees and shrubs are measured in selected range types.

There is no operational remote sensing program to monitor changes in New South Wales rangelands, although this is intended by site design. Sites are read in conjunction with over 300 participating landholders. This increases landholder access to a range of related rangeland services provided by Department of Land and Water Conservation.

The Rangeland Assessment Program has collected annual data from 1989 to 2001 for most sites. Information products from the program have been produced on a site, range type and regional basis.

Table A1. Representation of Rangeland Assessment Program sites by bioregion in western New South Wales.

Range type

 Bioregion Total
Mulga Lands Murray Darling Depression Riverina Broken Hill Complex Cobar Peneplain Darling Riverine Plains Channel Country Simpson Strzelecki Dunefields
Saltbush 1 5 40 14 - - 3 2 65
Sandplain 39 9 - - - - - - 48
Bluebush - 36 7 6 - 1 - - 50
Granite - 1 - - 24 - - - 25
Rosewood/Belah - 48 2 - - - - - 50
Hard Red - - - - 45 4 - - 49
Northern Floodplain - - - - - 51 - - 51
Total 338

Species diversity of vascular plants

Total species diversity (Figure A1 & A2) recorded at all sites over the period 1989 to 2000 is approximately 1147 species.

Figure A1 Total flora, and tree and shrub species diversity of bioregions in western New South Wales.
Figure A1 Total flora, and tree and shrub species diversity of bioregions in western New South Wales.
Figure A2 Chenopod bush density (1990 to 2000) in three bioregions in western New South Wales.
Figure A2 Chenopod bush density (1990 to 2000) in three bioregions in western New South Wales.

For most bioregions, biomass levels increased in the 1993 to 1994 period then declined in the 1995 to 1998 period with large increases in biomass observed in 2000 (Figure A3, A4 & A5). While the amount of fluctuation of biomass is generally low for those bioregions with low overall average biomass levels, the Cobar Peneplain shows an enormous increase in the 1999 to 2000 period due to favourable seasonal conditions.

Figure A3 Biomass levels for bioregions in western New South Wales.
Figure A3 Biomass levels for bioregions in western New South Wales.
Figure A4 Change in biomass after subtracting the average biomass trend for each bioregion.

Deviations are from the mean or zero line: 1990 was an exceptional year for Riverina and 2000 was an exceptional year for Darling Riverina Plains. The severe drought that the south west experienced in the 1995 to 1998 period is indicated by the exceptionally low biomass for Riverina and Murray Darling Depression.

Figure A4 Change in biomass after subtracting the average biomass trend for each bioregion.

Ivanhoe, central western New South Wales

Approximately 20 sites in the Bluebush range type have been in existence in the Hay West district since 1991. At each of these sites a systematic assessment (Green et al. 1994) are made of vegetation and soil parameters. Pasture species composition, frequency and biomass are measured along with perennial chenopod community composition and density. Overall increase can probably be partially attributed to rainfall (Figure A6).

Additional data (e.g. rainfall, stocking rates and management actions likely to influence the dynamics of the site) were collected at each site. Information recorded on management actions included total grazing pressure control (e.g. kangaroo/goat control programs and rabbit ripping operations). A coordinated rabbit-ripping program began in the early 1990s in the Ivanhoe area, with ripping operations taking place on several properties. Rangeland Assessment Program sites were established on these properties in the summer of 1994/95. Ripping operations took place within paddocks where sites were established and this management action was recorded as a likely contributor to influence site data.

Figure A6 Correlation between the average number of species found at all Hay West sites and ripped areas, (Bluebush range type) and rainfall (1991-1999).
Figure A6 Correlation between the average number of species found at all Hay West sites and ripped areas, (Bluebush range type) and rainfall (1991-1999).

The average number of species found at all Hay West Bluebush sites and at those areas of rabbit ripping operations was generally the same over the 1991 to 1999 period (23.8 and 22.5 species respectively). While the maximum average number of species recorded at all sites versus sites in ripped areas was almost the same (29.37 and 29.25 respectively), the average minimum number recorded was quite different (16.1 versus 13.7 respectively). The average number of species in those areas of rabbit ripping increased by almost 10% (1995-1999) after the commencement of ripping in 1994/95. Further analysis is needed to determine the influence of other factors such as rainfall effectiveness during this period.

Although there has been an overall increase in average frequency of desirable pasture species at all sites in the Hay and Buronga districts (Figure A7), the increase has been much greater on those sites in the rabbit-ripped areas. Most notably an increase can be seen in the 1995 reading. This increase is just over three hundred percent from the 1991 year. The average frequency of desirable pasture species has remained elevated after the 1994/95 increase. Pasture species that were determined to be desirable included those species of high palatability and generally of a perennial nature. They included grass species (e.g. Danthonia, Enneapogon, Eragrostis, Sporobolus) and annual and perennial saltbush and bluebush. Many of these species have shown a decrease in abundance with increasing grazing pressure.

Figure A7 Change in the average number of desirable pasture species (1991-1999). Change is shown as a proportion of those desirable species recorded at the first reading in 1991.
Figure A7 Change in the average number of desirable pasture species (1991-1999). Change is shown as a proportion of those desirable species recorded at the first reading in 1991.

Although an overall increase in perennial chenopod seedlings (a seedling or juvenile is recorded as being less than ten centimetres in height; the amount of woody thickening determines if it is a seedling or juvenile) has occurred, a particularly rapid increase occurred between 1996 and 1999 (Figure A8). Bluebush seedlings would be particularly vulnerable to rabbit grazing even in low densities. The drop in the 1999 reading is probably due to limited germination in the 1998 to 1999 period due to dry conditions.

Figure A8 Change in the average frequency of perennial chenopod seedlings and juveniles (1991-1999). Species include bluebush (Maireana spp.).
Figure A8 Change in the average frequency of perennial chenopod seedlings and juveniles (1991-1999). Species include bluebush (Maireana spp.).

Riverina Plain

During the period 1990 to 2000 an ecological phenomenon called 'dieback' affected large areas of saltbush on the Riverina Plain. Site 8007 shows a typical ecological response to this perturbation (Figure A9). In 1990/91 the saltbush was healthy with a density of approximately 11 300 plants per hectare. In 1992 the saltbush community was hit by 'dieback' coinciding with massive defoliation from a plague of caterpillars. In 1993 saltbush plants were still struggling on with very few leaves. At the same time a large germination of poverty bushes had occurred. By 1994 the poverty bush density had replaced the struggling saltbush that declined to a low of 475 plants per hectare. Saltbush numbers then built up steadily from 1995 to 1999 with a good germination in 1999 coinciding with favourable seasonal conditions. During this phase the poverty bush has reached full maturity and has senesced reaching a low level again in 1998. It also responded to favourable seasonal conditions in 1999 but with a lower capacity to do so. A four year cycle can be observed in the poverty bush from germination to senescence.

Figure A9 Dieback on the Riverina Plain.
1990 1990 1990 1992 Figure A9 Dieback on the Riverina Plain.
1990 1994 1990 1996 1990 1998 1990 2000

Queensland

Monitoring in Queensland's rangelands occurs in such programs as Transect Recording and Processing System (TRAPS); QGRAZE and Grass Check run by the Department of Primary Industries; and Statewide Landcover and Trees Study (SLATS) and Aussie GRASS run by the Department of Natural Resources and Mines.

Transect Recording and Processing System

The Transect Recording and Processing System (TRAPS) has been progressively establishing 150 sites in woodland communities since 1982 (Figure A10). Attributes assessed along permanent transects within the site area of 1 ha include:

The herbaceous component is recorded in quadrats (see QGRAZE p. 140). All Transect Recording and Processing System sites have been re-recorded at least twice. Transect Recording and Processing System data has been used to show that woody plant 'thickening' is a real and significant process occurring in Queensland's grazed woodlands (Figure A12).

Figure A10 Transect Recording and Processing System and QGRAZE site distribution in Queensland.
Figure A10 Transect Recording and Processing System and QGRAZE site distribution in Queensland.
Figure A11 Graphic generated by the Transect Recording and Processing System processing package showing the reduction in area (canopy cover) along four recording lines of the shrub Carissa ovata following a fire.
Figure A11 Graphic generated by the Transect Recording and Processing System processing package showing the reduction in area (canopy cover) along four recording lines of the shrub Carissa ovata following a fire.
Figure A12 Transect Recording and Processing System site in narrow-leaved ironbark woodland in central Queensland (1983, 1995 & 1999) showing a thickening of the shrubby understorey.
Figure A12 Transect Recording and Processing System site in narrow-leaved ironbark woodland in central Queensland (1983) showing a thickening of the shrubby understorey. Figure A12 Transect Recording and Processing System site in narrow-leaved ironbark woodland in central Queensland (1995) showing a thickening of the shrubby understorey.
Figure A12 Transect Recording and Processing System site in narrow-leaved ironbark woodland in central Queensland (1999) showing a thickening of the shrubby understorey.

QGRAZE

Three hundred and fifty QGRAZE sites have been established since 1991. They are located mainly in black speargrass mesic woodlands (44%), Aristida/Bothriochloa xeric woodlands (25%), brigalow (9%), spinifex (4%), channel pastures (3%), bluegrass/browntop (3%), gidyea (3%) and Mitchell grass (3%). Herbaceous species frequency, frequency and size of woody species and amount of cover are measured in quadrats (Figure A13, Table A2). Pasture yield (based on photographic standards), soil surface condition and tree basal areas are assessed over the general site area of 4 ha (Figure A14). Although it is planned to re-record sites at least once every five years this has only occurred in about 70% of the sites. Mechanisms for reporting are being developed around the attributes of ground cover (a), perennial grass species (b), palatable, productive and perennial grasses (3P grasses) (c), and exotic species (d). These are being tested as surrogates for landscape function (a+b), grazing condition (a+c) and landscape health/biodiversity (a+b+c).

Figure A13 Photo pair showing ground cover change at a QGRAZE site in a narrow-leaved ironbark woodland, in the black speargrass native pasture community, approximately 150 km west of Charters Towers. Photo 1 is from 1994 during dry/drought years. Photo 2 is from 1999 following good summer rains.
Figure A13 - Photo 1 is from 1994 during dry/drought years. Figure A13 - Photo 2 is from 1999 following good summer rains.
Table A2 QGRAZE output showing changes in plant species frequency and cover on three geological units in Dalrymple Shire in North Queensland (data from 12 QGRAZE sites) (1995 to 1999/2000). The periods between 1991 and 1997 were drought years.
Year Geological units Average
Basalt Granodiorite Tertiary sediment
Pasture species
Aristida spp. 1 4.50 7.67 39.00 17.06
2 2.33 3.00 31.00 17.06
Bothriochloa ewartiana 1 30.33 3.33 t 11.22
2 53.33 t 5.00 19.44
Bothriochloa pertusa 1 t 57.33 1.00 19.44
2 24.00 51.00 1.00 25.33
Chrysopogon fallax 1 17.67 18.67 41.33 25.89
2 15.67 2.50 29.67 15.94
Forbs 1 40.33 65.00 25.33 43.56
2 25.00 4.33 31.67 20.33
Heteropogon contortus 1 6.33 23.50 12.67 14.17
2 16.33 12.00 23.00 17.11
Native legumes 1 1.50 7.00 t 2.83
2 1.50 t 2.33 1.28
Panicum spp. 1 t 4.50 1.50 2.00
2 3.67 t 10.50 4.72
Sedges 1 t 13.33 42.33 18.56
2 2.50 1.00 36.67 13.39
Sida spp. 1 21.00 9.33 10.00 13.44
2 2.00 t 1.67 1.22
Cover 1 35.06 41.45 38.26 38.26
2 83.14 72.57 56.71 70.81

Year 1 is 1995. Year 2 is either 1999 or 2000
t = traces

Figure A14 Rangeland officers estimating tree basal area using the Bitterlich Stick method, at a QGRAZE site, Cloncurry, north-west Queensland.
Figure A14 Rangeland officers estimating tree basal area using the Bitterlich Stick method, at a QGRAZE site, Cloncurry, north-west Queensland.

Grass Check

Land managers are encouraged to implement their own monitoring and adaptive grazing management through the Grass Check (Grazier Rangeland Assessment for Self-Sustainability) program by:

Statewide Landcover and Trees Study

The Statewide Landcover and Trees Study initiative has used Landsat thematic mapper images to develop a monitoring system to report regularly the extent, condition and trend of Queensland's vegetation cover and land use, and provide estimates of greenhouse gas emissions in the land use and forestry sectors.

Aussie GRASS

The Australian Grassland and Rangeland Assessment by Spatial Simulation project (Aussie GRASS) is a collaborative project led by the Department of Natural Resources and Mines. It uses advanced simulation modelling techniques to assess the condition of Australia's rangelands. The Aussie GRASS model operates across the continent on a 5 km grid basis. The model uses inputs of daily rainfall and climate, soil functional characteristics, vegetation characteristics, tree density, and grazing pressure for each grid cell. The model simulates the processes of runoff, infiltration, deep drainage, evapotranspiration, pasture growth and senescence, litter decay, and consumption of biomass by grazing animals. Model output is often presented in percentile format, that allows users to assess current seasonal conditions relative to historical conditions (Figures 23 & 24-p. 63).

South Australia

The rangelands program also runs a number of projects specifically related to its core function of carrying out monitoring and inventory of the land resources of South Australia's rangelands. These include a long-term historic photo relocation project whereby old pastoral inspection photographs are being systematically re-located where possible and archived. The more interesting photo-sequences are being formalised into photo points as part of the current pastoral monitoring program. The photographs included here show a typical 35-year sequence on Lilydale station in the Northeast Pastoral district. Chenopod shrubs, particularly blackbush, are now colonising what was a previously bare and degraded blackoak community.

Rangeland Monitoring Program

Primary Industries and Resources SA has a comprehensive, integrated program of:

Over the last 10 years South Australia has assessed resource condition and established a baseline monitoring system over all pastoral leases in the State. Each of the 219 pastoral properties on 328 leases and covering 409 000 kmē of the rangelands now has:

Approximately 5500 photo point monitoring sites have been established, 20 000 Land Condition Index sample points have been assessed and 4500 individual paddocks have been assessed and assigned a priority for land management action and further inspection. This work and the scientific standards upon which it is based are driven by requirements under the Pastoral Land Management and Conservation Act 1989 (SA) .

The focus of future activities by the Pastoral Board and Primary Industries and Resources SA is on reporting at both lease and district scales. Lease-scale reports remain confidential to the Pastoral Board and the lessee. Although little aggregated information is currently available, a pilot project to develop district-based reporting using the photo point monitoring site network is being undertaken across the Kingoonya and Gawler Ranges Districts.

There is no firm schedule for reassessment of leases or systematic reassessment of photo point monitoring sites at present. In the short term, 10% of photo points in the Kingoonya and Gawler Ranges District will be revisited through the District Monitoring Project. The Pastoral Inspection Program will also revisit a selection of sites as well as following up on land management issues identified during the Pastoral Assessment Program.

In the longer term, an assessment is required every 14 years after the initial assessment or when the lessee accepts a new or extended term of the lease, as provided for under the Act. Thus properties that have not accepted a new or extended lease will require assessment in 2004.

There are two techniques for assessing resource condition depending on whether the land is under cattle or sheep grazing.

Grazing Gradient Assessment

The Grazing Gradient Assessment method (Pickup et al. 1994) is used on a proportion of the northern cattle leases and allows grazing effects on vegetation to be separated from those due to rainfall and local landscape variability by examining patterns of cover change with increasing distance from water. Vegetation cover tends to increase with distance from watering points as grazing intensity decreases, producing a grazing gradient (Figure 11, p. 43). Where this cover gradient persists after high rainfall, it indicates a degree of land degradation.

The grazing gradient method is an effective tool for describing present rangeland condition. A major benefit is the ability to separate grazing effects from seasonal change in vegetation cover.

Grazing gradient analyses were conducted in areas used for extensive grazing by domestic cattle in northern South Australia for the regional assessment and reporting of landscape function (vegetation cover). Areas where vegetation cover was below optimal levels under grazing land use were identified. It also provided a benchmark from which future changes could be measured.

There is considerable potential to integrate ground-based monitoring with Grazing Gradient Assessment to better understand the actual changes occurring to the soil and vegetation. This could occur by applying the grazing gradient method at property and paddock scales. Ground-based monitoring sites could be located at distances from water where change in vegetation is indicated by the grazing gradient plots. Such sites could collect data describing landscape function (Ludwig et al. 1997). The pattern of vegetation response in grazing gradient plots would allow determination of how much the landscape function has been changed. Areas at risk could then be identified and targeted. Pastoral administrators would be able to shift to a more predictive approach that would prevent damage from inappropriate grazing.

Land Condition Index

On most of the leases in the sheep production areas south of the dingo-proof fence, the Land Condition Index (Lange et al. 1994) is used as the primary assessment of land condition (Table A3).

The Land Condition Index is based on the condition rating of 80-100 sample sites within each lease. Assessments are made into one of three classes:

These classes are precisely specified for each component of each pasture type within a district. Under the Pastoral Land Management and Conservation Act 1989 (SA) , the optimal condition for the land is one that maintains the native plant and animal life. This is important, since it suggests that the Land Condition Index is more closely related to the maintenance of biodiversity than to pastoral production or landscape function. In practice, the maintenance of native species, pastoral production values and landscape function are closely related for many of the pasture types.

The disturbance categories are mostly based on the presence, absence and abundance of perennial plant species, the level of grazing and browsing of palatable species, and some consideration of soil surface condition. The condition classes provide an inherent assessment of the likelihood of the vegetation community to return to something like undisturbed condition (e.g. sites in Class 1 condition are unlikely to recover, while those in Class 2 have the potential to recover under sufficiently benign management).

A weighted average condition index is determined for each lease by multiplying the percentage of sample points for each condition rating by the rating. This gives a value for each lease of between 100 (all sample points severely disturbed) and 300 (all sample points assessed as low disturbance).

Table A3 Percentage of condition rating observations for pasture type communities within Gawler Ranges Soil Conservation District leases. Such tables provide data on the relative condition of different vegetation types and also indicate the relative proportion of each vegetation type sampled within each district. They do not necessarily represent the relative areas of each vegetation type within the district.
Vegetation community Land Condition Index rating (%)
High disturbance Moderate disturbance Low disturbance Total
Chenopod shrublands
1(a) Treeless plains 0.30 0.64 5.56 6.51
1(c) Samphire/saltlake 0.61 1.14 6.48 8.22
1(d) Calcareous plains 1.88 9.11 16.17 27.16
1(e) Arcoona tableland 0.30 0.30 4.68 5.29
1(g) Blackbush watercourses 2.08 6.78 9.52 18.38
1(h) Gawler Range alluvial valleys 5.23 5.20 1.41 11.85
Total 10.41 23.17 43.83 77.41
Low woodlands
2(a) Mulga grasslands-sandy 0.36 1.30 0.55 2.21
2(d) Mallee/blackoak 3.99 5.34 6.17 15.5
2(e) Native pine on dunes 0.33 1.14 0.28 1.74
2(f) Mallee/spinifex on dunes 0.00 0.39 1.16 1.555
Total 4.68 8.17 8.17 21.02
Hummock grasslands
4(a) Gawler Range rhyolite hills 0.61 0.58 0.36 1.55
Total 0.61 0.58 0.36 1.55
Total 15.70 31.92 52.36 100.00

Northern Territory

The Department of Lands and Planning and Environment undertakes all land resource assessment and mapping, pastoral lease monitoring and infrastructure mapping, regional land monitoring, and lease inspection activities. Monitoring of the pastoral estate involves permanent photo sites, soil and vegetation measurements, and remote sensing.

Monitoring of the pastoral estate is a two-tiered system, called Tier 1 and Tier 2.

Tier 1 is operational on a property by property basis, while Tier 2 is regionally based with current activity in the Victoria River District, Sturt Plateau, the Barkly Tablelands, the Mary River catchment and some of the Alice Springs District. </ p>

Tier 1

The Tier 1 system consists of permanent photo sites located in each of the major paddocks (or grazing areas) on pastoral leases. Pastoral Officers record soil and vegetation observations though lessees are encouraged to photograph sites annually and record their own observations.

Regional Rangeland Monitoring Program (Tier 2)

The Regional Rangeland Monitoring Program (Tier 2) uses two different (but related) satellite-based methods to account for variation. Landscape Cover Change Analysis is used in the tropical savannas in the north and Grazing Gradient is used in the arid interior to the south.

Remote sensing is seen as the major means of broadscale monitoring, augmented by ground-based sites. This system is the most sophisticated system in Australia and includes an explicit focus on landscape processes rather than vegetation species composition.

Land administration is focused on conservation of the base resource for pastoral purposes. Lease conditions in the Pastoral Land Act 1992 (NT) do not include the need to manage for biodiversity. Nor is there an explicit requirement to monitor biodiversity. However, monitoring of landscape processes has relevance to maintenance of biodiversity.

Grazing gradient: a single watering point

Grazing gradients at individual water points can be examined to provide information at the basic management unit level (Figure A15 - temporarily unavailable in web version). The comparison of two seasonal gradients demonstrates not only the extent of use, but also the subsequent response that occurred following rain.

Grazing gradient: landscape types at the paddock scale

Grazing gradients for one landscape type in a paddock were calculated from Landsat TM-derived data for December 1994 and March 1995 (Figure A16). Each gradient shows average cover levels present at increasing distance from water before (red line) and after (green line) a moderate rainfall event. The Y axis is scaled in units of PD54 cover between 0 and 254: values below 60 equate with bare ground while values above 240 represent 100% cover. The horizontal blue line indicates the expected cover level if vegetation close to water had fully recovered from the effects of grazing following the rainfall received.

Dams, or surface catchments, constructed adjacent to tree-lined creeks are the usual closest water source in this floodplain unit. Tree canopies thus contribute to the relatively high cover levels immediately adjacent to water. Average cover levels decline with increasing distance from water as progressively more of the low-cover floodplain (i.e. ephemeral herb land unit) is included in the calculation of average cover levels. Cover levels increase beyond about 4 km from water due to the presence of perennial tussock grasses on riparian areas and patches of timber on the floodplains.

The majority of this floodplain unit appears to be grazed fairly uniformly. There is no distinct pattern of increasing average cover with distance from water until the area at 4 km from water is reached. If there were areas further than 6 km from water, we would expect average cover levels to continue to increase before levelling off. The wet-period data indicate that there was considerable growth of herbage across the floodplain unit following the January 1995 rains. This suggests that the floodplain unit has a considerable potential to recover from grazing following substantial summer rain (even though maximum cover levels are not uniformly restored at all distances from water across the vegetation unit).

Figure A16 Grazing gradients for ephemeral herb land before and after a moderate rainfall event.
Figure A16 Grazing gradients for ephemeral herb land before and after a moderate rainfall event. Figure A16 - Ephemeral herb land in dry conditions.

Ephemeral herb land in dry conditions

Figure A16 - Ephemeral herb land in wet conditions.

Ephemeral herb land in wet conditions

Grazing gradient: resilience of vegetation at the paddock scale

Response of vegetation cover to rainfall can be assessed on a pixel by pixel basis for determining whether it is above or below what might be expected given little or no grazing impact. Below-expected response often results from desertification and can indicate areas with reduced pasture productivity. Above-expected response indicates a resilient landscape that:

In this type of assessment landscape features are numbered image and described (Figure A17).

Following initial assessment and in the longer term, further images can be produced in a repeatable manner following future good rains. The output can be used to assist in judging the range of trend under continued, or changed, grazing management.

Figure A17 Vegetation response in a paddock following 350 mm of rainfall in March 1989.
Figure A17 Vegetation response in a paddock following 350 mm of rainfall in March 1989.

Much of the paddock had an average to above-average vegetation response. The best response was in a broad watercourse where alluvial soils and water running off higher ground combined to produce ideal conditions for annual plant growth. Below-average vegetation response occurred on eroded areas surrounding a long-established watering point and in a heavily shrubbed watercourse.

Management options include piping water to lightly grazed country in the eastern part of the paddock and closing access to waters associated with below-average plant response except when required for mustering.

Grazing gradient: regional summaries

Analysis of 26 land systems covering 38 000 kmē indicated that those land systems containing a high proportion of palatable forage had the most persistent wet-period grazing gradients and were the most affected by grazing. Each land system was analysed separately and ranged in area from 67 kmē to 3137 kmē. Assessment of one of these land systems is represented in Figure A18 showing average cover levels present at increasing distance from water before (red line) and after (green line) a moderate rainfall event. The Y axis is scaled in units of pd54 cover between 0 and 254: values below 60 equate with bare ground while values above 240 represent 100% cover. The horizontal blue (or black) line indicates the expected cover level if vegetation close to water had fully recovered from the effects of grazing following the rainfall.

This land system is made up of alluvial plains and other major creeks in the northern part of the region. The major vegetation type is an open woodland over annual and perennial forage species though smaller areas of perennial grassland on clay loam soils also occur. An aerial photograph shows a typical landscape. The graphs show a moderate cover level. Strong wet and dry period normal grazing gradients persist to 7 km from water and although cover levels increased after rain, there is minimal recovery from the effects of grazing closer to water.

Management techniques that encourage the regeneration of palatable forage species, including perennial grasses, would seem necessary in some areas to allow vegetation closer to water to respond to rainfall in a similar manner to that on more distant areas.

Figure A18 Assessment of a land system showing average cover levels present at increasing distance from water before and after a moderate rainfall event.
Figure A18 Assessment of a land system showing average cover levels present at increasing distance from water before and after a moderate rainfall event. Figure A18 Assessment of a land system showing average cover levels present at increasing distance from water before and after a moderate rainfall event. Figure A18 Assessment of a land system showing average cover levels present at increasing distance from water before and after a moderate rainfall event.

Land Cover Change Analysis: long-term time trace of monitoring sites in the Victoria River District.

The time-trace of monitoring sites allows comparison of monitoring data in the context of long-term change in the landscape (Figure A19). The time trace of two monitoring sites, one in good condition (V9-green line) and one in poor condition (V1-red line) are compared over a fifteen-year period from 1983 to 1997. Bureau of Meteorology rainfall data is displayed as an indication of seasonality. Lower mean reflectance corresponds to lower vegetation cover. Dissimilarity between time traces is evident as the trace for site V1 has greater fluctuations, generally following seasonal trends, compared to site V9, that has been relatively stable.

Successive good seasons from 1993 to 1997 (rainfall data) are reflected by upward trends in both time traces. Time traces also respond to specific events (e.g. Site V9 was affected by fire in 1988 and exhibited a downward trend; in 1994 heavy grazing left site V1 virtually bare, that was expressed as a downward trend and low mean reflectance. The time trace analysis suggests Site V1 has been in a poor state dominated by annuals and herbaceous cover for a considerable period, in contrast to site V9).

Figure A19 Long-term time trace of monitoring sites in the Victoria River District.
Figure A19 Long-term time trace of monitoring sites in the Victoria River District.
Figure A19 Poor condition site (VI)

Figure A19 Poor condition site (VI)

Source: Karfs 1999.

Land Cover Change Analysis: multiple temporal satellite sequences at the paddock scale

Trend summary imagery can be used at the paddock scale and over varying time sequences (Figure A20). By comparing the same areas over time, changes to cover can be detected.

Colours indicate good vegetative cover (cyan and green), good cover with decreasing trend (yellow) and poor cover (blue and red). Watering points are shown as solid blue circles, fencelines in white and roads in red. Continuous high cover indices detected over the ten-year period indicate good condition grassland dominated by perennials, that ground truthing has confirmed. Drainage lines having a dense tree canopy display low cover indices (blue and red), due to absorption in Landsat MSS band 2. This response is similar to exposed, dark coloured soil on the adjacent plains, requiring stratification of these systems. Recovery is evident in pastures near the bore in the lower left, following grazing impact in 1987.

In this example, the spatial extents of under utilised perennial pastures provide landholders with baseline information on land condition to help develop management options. Most of the area has been in good condition for some time. The few areas adjacent to riparian corridors (that often provide refuge for stock and native wildlife) having lower cover values may indicate preferential patch grazing. By developing a management regime with more even utilisation of pastures over the broad landscape, the focus on other areas may be reduced.

Land Cover Change Analysis: regional summaries

Analysing three land types separately then combining them into a contiguous coverage created a regional trend product from 1992 to 1997 (Figure A21). The area analysed is 66 550 kmē of 126 000 kmē within four mosaic Landsat scenes.

Light green represents areas where cover increased and dark green represents areas with stable cover over the period. Areas shown in red represent a decreasing trend in cover. In this example, fire scars have not been removed and much of the red is attributed to burnt country. Fire history maps overlain on this image would aid in identifying areas affected by fire.

Clearly these data show that over most of the region cover has increased or remained stable. This regional trend can be attributed to an exceptional run of good seasons from 1993 to 1997. It is also consistent with the interpretation of ground data collected at monitoring sites over the same period.

Regional information products are beneficial to statutory agencies to report on differences within their jurisdiction. There is also considerable potential for examining the cover trends to make comparisons of current and past management with regard to seasonal variation over large tracks of rangelands.
Figure A20 Examples of trend summary imagery at the paddock scale over three time sequences.
Figure A20 Examples of trend summary imagery at the paddock scale over three time sequences.
Figure A21 A regional trend product from 1992 to 1997.
Figure A21 A regional trend product from 1992 to 1997.

Western Australia

Range inventory and monitoring activities

A range of activities are being used in Western Australia to track trend over time and report on condition of the pastoral rangelands. Elements of all the activities described below can be collated and used for Australia wide reporting on a commissioned basis.

Regional scale resource inventory and range condition survey

The Departments of Land Administration and Agriculture conduct a joint Range Survey program that maps rangelands to the land system level and combines soil and vegetation condition assessments into an assessment of range condition at the lease and land system scale. Severely degraded and eroded areas are also mapped. About 87% of the pastoral rangelands have been included in these surveys. In the Kimberley, land systems were mapped by CSIRO. Range condition was assessed by the Department of Agriculture in the mid-1970s (in the west Kimberley) and throughout the Kimberley Region in the early 1990s, but on a lease by lease rather than regional basis. South of the Kimberley, the Regional Surveys are one-off assessments and no repeat surveys have been scheduled. However, over 6000 inventory and condition sites were sampled during the Range Surveys and the potential remains to re-visit these sites if required and resourced.

Regional scale range monitoring

The Western Australian Rangeland Monitoring System (WARMS) consists of almost 1600 fixed sites, located on representative areas of specific pasture/vegetation communities. While most leases, of viable size, have at least one Western Australian Rangeland Monitoring System site and on average there are about three sites per lease, the Western Australian Rangeland Monitoring System is designed to report at the vegetation type or regional/district scale, not at the lease scale.

There are two different types of Western Australian Rangeland Monitoring System sites. Throughout the Kimberley, the Pilbara and some areas south of the Pilbara, grassland sites are used. Shrubland sites are used throughout the area south of the Kimberley, particularly south of the Pilbara. Attributes related to perennial vegetation dynamics and landscape function are recorded on both grassland and shrubland sites. On grassland sites, the frequency of all perennial species is assessed in quadrats and an estimate is made of crown cover of woody perennials. On shrubland sites, the demography and maximum crown dimensions of all shrubs are recorded using complete census techniques. On both grassland and shrubland sites standard Landscape Function Analysis techniques developed by CSIRO are used to assess soil surface and other attributes related to landscape function. Standard photographs are taken at both types of sites. Grassland sites are reassessed on a three-yearly cycle, shrubland sites on a six-yearly cycle. The system in its current form was begun in 1992, although many old monitoring sites were incorporated and some data and photo records go back to the 1970s.

Figure A22 The location of WARMS sites, coloured by year of assessment. The greyed areas denote pastoral leases and the khaki line work denotes Interim Biogeographic Regionalisation of Australia (version 4) boundaries. The heavy green line in the southwest shows the demarcation between rangelands and the south-west wheatbelt.
Figure A22 The location of WARMS sites, coloured by year of assessment. The greyed areas denote pastoral leases and the khaki line work denotes Interim Biogeographic Regionalisation of Australia (version 4) boundaries. The heavy green line in the southwest shows the demarcation between rangelands and the south-west wheatbelt.

Lease level range condition assessment

The Department of Agriculture, on behalf of the Pastoral Lands Board, maintains a regular program of individual lease inspection, on a maximum cycle of six years. Range Condition Assessment uses similar techniques to the Range Survey program and over time will enable regional assessments to be made of condition, and change in condition, at the sublease scale. While reports for individual leases are confidential to the lessee and the Pastoral Lands Board, outputs could be aggregated on a regional/district or vegetation type level.

Lease level monitoring by pastoralists

Many pastoral managers maintain their own systems of range monitoring, using photo or other techniques. While government does not formally maintain data, many pastoralists' sites overlap with Western Australian Rangeland Monitoring System sites and their photos are assessed by pastoral lease inspectors during their lease inspection activities. At least 3000 of these sites have been installed. Recently, the Department of Agriculture and the Department of Conservation and Land Management have begun supporting the monitoring by pastoralists of a range of attributes related to pastoral productivity and nature conservation. This monitoring is a component of the Ecosystem Management Unit Project.

Other monitoring activities

While not currently part of a formal system, the Western Australian Rangeland Monitoring System activity maintains photos and/or data records for about 3800 other monitoring sites on pastoral leases. Many of these overlap with individual lessee sites. These sites were put in for a range of purposes. However, the majority were installed using techniques similar to Western Australian Rangeland Monitoring System but pre-date the current Western Australian Rangeland Monitoring System activity. Although these photos and records are used in an ad hoc manner (e.g. by pastoral lease inspectors), the potential remains for their more formal use either within a better resourced program, or in more intensive studies of specific areas.

The Department of Agriculture helped develop the Land Cover Change Analysis technique of remote sensing as used by the Northern Territory Department of Lands Planning and Environment. While there is no scheduled program of State-wide coverage in Western Australia, analysis is complete for much of the east Kimberley and for several areas in the Pilbara and the shrublands. The technique is now being field tested on a district-level scale in the East Kimberley. The outputs are used for a range of purposes, including the identification of leases, or parts of leases, where the change in perennial cover over time is poor in relation to neighbouring areas. Changes in cover since 1982 can be produced for a range of stratifications within those areas for which processed imagery exists.

Range Survey Program

Almost 75 000 traverse assessments of range condition were made at 1 km intervals during the Range Survey Program. These allow a snapshot to be produced of range condition at the regional scale (Table A4). Range Condition Assessment during pastoral lease inspections will enable contemporary updates of these regional assessments for an Australia-wide information system. Summaries can also be produced for other stratifications, such as vegetation type.

Table A4 Summary of condition assessment outputs from the Range Survey Program, a joint initiative of the Departments of Land Administration and of Agriculture.
Region surveyed (and year commenced) Total area (kmē) No. of traverse assessments Severely degraded and eroded area (as mapped) Resource condition classes (% of traverse assessments)
kmē % Good Fair Poor
Gascoyne (1969) 63,400 2,426 1,205* 1.9* 32 53 15
West Kimberley (1972) 89,600 4,532 2,000* 2.2* 20 50 30
Eastern Nullarbor (1974) 47,400 1,273 0 0 50 10 40
Ashburton (1976) 93,600 8,608 534 0.6 50 34 16
Carnarvon Basin (1980) 74,500 10,952 647 0.9 45 32 23
Murchison (1985) 88,360 13,441 1,560 1.8 21 37 42
Roebourne Plains (1987) 10,216 1,172 233 2.3 51 27 22
North-eastern Goldfields (1988) 100,570 10,470 452 0.4 39 32 29
Sandstone-Yalgoo-Paynes Find (1992) 94,710 9,435 145 0.2 45 32 23
Pilbara (1995) 181,736 12,518 322 0.2 77 11 12
All areas surveyed 843,576 74,827 7,098 0.8 46 30 24

* Not mapped, estimate only

Western Australian Rangeland Monitoring System

The Western Australian Rangeland Monitoring System set of grassland sites was installed during 1994 to 1996, and a complete reassessment was finished in 1999. Shrubland sites were installed from 1993 to 1999. The first round of shrubland reassessments began in 2000 and is due to finish in 2004.

In the Kimberley, perennial grass frequency remained constant or increased on 69% of sites between the period 1994 to 1996 (installation) and 1997-1999 (first reassessment). Figure A23 provides an example output for three vegetation types. Overall, 10% of sites were judged to have improved, and 5% declined (Table A5). Across all sites, the average frequency of perennial grasses increased (Table A6), while the crown cover of woody perennials decreased on limestone grass and black soil plain vegetation groups during the same period.

Preparation of these summaries was based on stratifications that meet the needs of the Department of Agriculture and the Pastoral Lands Board. Other stratifications (e.g. using different combinations of species or on a district by district basis) can be used to prepare specific outputs for an Australia-wide system to meet commissioned needs.

Figure A23 Change in perennial grass frequency, and judgement of change, for three vegetation types (Mitchell grass, curly spinifex, southern ribbon grass) on Kimberley WARMS sites between installation (1994-1996) and reassessment (1997-1999).
Figure A23 Change in perennial grass frequency, and judgement of change, for three vegetation types (Mitchell grass, curly spinifex, southern ribbon grass) on Kimberley WARMS sites between installation (1994-1996) and reassessment (1997-1999).
Table A5 Summary of trend assessment for Western Australian Rangeland Monitoring System grassland sites in the Kimberley, by vegetation group and Land Conservation District over the period from 1994-1996 (installation) to 1997-1999 (reassessment). The analysis was based on perennial grass frequency, all species pooled.
Land Conservation District Vegetation group No. of sites Improved Static Declined
Broome Curly spinifex 19 2 16 1
Coastal vegetation 12 2 9 1
Northern ribbon grass 2 2
Southern ribbon grass 1 1
Soft spinifex 9 8 1
Subtotal 43 4 36 3
Derby West Kimberley Black soil plains 63 11 51 1
Curly spinifex 31 3 26 2
Frontage grass 11 1 9 1
Southern ribbon grass 59 3 52 4
Soft spinifex 1 1
Subtotal 165 18 138 9
Halls Creek - East Kimberley Black soil plains 46 4 41 1
Curly spinifex 6 6
Limestone grass 12 3 9
Northern ribbon grass 5 1 4
Southern ribbon grass 3 1 2
Soft spinifex 13 1 11 1
Subtotal 85 10 73 2
North Kimberley Black soil plains 4 4
Curly spinifex 13 1 12
Frontage grass 2 2
Northern ribbon grass 25 22 3
Southern ribbon grass 1 1
Subtotal 45 1 41 3
Total 340 33 290 17
Table A6 Average change in perennial grass frequency and average change in crown cover estimates (%) for all woody species>1 m, by vegetation groups for Kimberley Western Australian Rangeland Monitoring System sites assessed and reassessed between 1994 and 1999.
Vegetation group No. of sites Mean frequency Significant change
1994-1996 1997-1999
Average change in perennial grass frequency (%)
Black soil plains 113 74.4 80.7 *
Curly spinifex 69 83.7 85.7
Coastal vegetation 12 86.2 89.2
Frontage grass 13 70.0 75.4
Limestone grass 14 39.9 47.1
Northern ribbon grass 32 88.5 85.7
Southern ribbon grass 64 75.0 76.6
Soft spinifex 23 84.9 86.5
Average crown cover (%)
Black soil plains 113 1.8 1.4 *
Curly spinifex 69 13.2 13.8
Coastal vegetation 12 1.0 0.5
Frontage grass 13 7.9 9.3
Limestone grass 14 6.7 4.8 *
Northern ribbon grass 32 12.5 12.5
Southern ribbon grass 64 6.1 5.6
Soft spinifex 23 5.0 7.7

Significance was tested using the two tailed paired t-test.

not significant = P>0.05

Figure A24 Change in shrub populations from all 214WARMS shrubland sites reassessed to end of 2000.

After combining individual species into categories based on turnover rates, 351 turnover rate by site combinations were produced. Populations were filtered to include onyl those where either the initial or final number was greater than 20 individuals.

Figure A24 Change in shrub populations from all 214WARMS shrubland sites reassessed to end of 2000.

In the shrublands, most sites appear stable or are showing improvement, based on analysis of shrub dynamics. An increase in shrub populations, in most vegetation types, is seen as desirable. Furthermore, species with low turnover rates (essentially longer-lived species) are more reliable indicators of range health, being less subject to demographic fluctuations caused by recent seasonal conditions. During the period between installation (1993-1996) and reassessment (1999-2000), only 19% of shrub populations, when considered by turnover rate category, declined (Figure A24). Only a few sites showed catastrophic decline and many of these were in areas that had been uncharacteristically inundated by sequential cyclonic rains.

Outputs from the Land Cover Change Analysis technique are identical to those produced by the Northern Territory Department of Lands Planning and Environment (see Northern Territory section). These include time traces by vegetation group-either for individual leases or for entire land systems-as well as images showing change over time for specific areas and images, and which areas are performing better or worse than neighbouring areas (see Land Cover Change Analysis section for the Northern Territory).