the impact of artificial watering points on rangeland

2008

WaterSmart Pastoral Product ion™ Project L i terature Reviews

AL HowesCA McAlpine

Working Paper

15

The Working Paper Ser ies

The impact of ar t i f ic ia l water ing points on rangeland biodiversi ty:

A review

The impact of artif icial watering points on rangeland biodiversity:

A review

AL Howes

CA McAlpine

2008

WaterSmart Pastoral Production™ project participants

Gail StevensonManager National Landcare Program www.daff.gov.au

Colleen JamesProject Officer WaterSmart Pastoral Production™

www.dcq.org.au

Andy BubbProject Manager www.primaryindustry.nt.gov.au

John GavinGeneral Manager www.saalnrm.sa.gov.au

Derek WhiteGreat Artesian Basin Coordinating CommitteeAustralian Government Department of the Environment, Water, Heritage and the Arts

www.gabcc.org.au

Information contained in this publication may be copied or reproduced for study, research, information or educational purposes, subject to inclusion of an acknowledgement of the source.

ISBN: 1 74158 053 6 (Web copy)ISSN: 1833-7309 (Web copy)

CitationHowes AL and McAlpine CA 2008, The impact of artificial watering points on rangeland biodiversity: A review, DKCRC Working Paper 15, The WaterSmart™ Literature Reviews, Desert Knowledge CRC, Alice Springs.

The Desert Knowledge Cooperative Research Centre is an unincorporated joint venture with 28 partners whose mission is to develop and disseminate an understanding of sustainable living in remote desert environments, deliver enduring regional economies and livelihoods based on Desert Knowledge, and create the networks to market this knowledge in other desert lands.

AcknowledgementsThe Desert Knowledge CRC receives funding through the Australian Government Cooperative Research Centres Programme; the views expressed herein do not necessarily represent the views of Desert Knowledge CRC or its Participants.

For additional information please contactDesert Knowledge CRCPublications OfficerPO Box 3971Alice Springs NT 0871AustraliaTelephone +61 8 8959 6000 Fax +61 8 8959 6048www.desertknowledgecrc.com.au

© Desert Knowledge CRC 2008

Executive summary

This report describes the effects artificial watering points have on rangeland systems and their biota in Australia. One of the recent concerns regarding the large-scale spread of artificial watering points has been the effect that the increase in grazing pressure surrounding these focal points has on rangeland biodiversity. The loss of biodiversity in these regions has been attributed to the decline in landscape function as a result of human induced disturbances. As grazing pressure increases around these focal points, landscape function declines considerably, with a loss in vegetation cover, an increase in erosion, and a decrease in nutrient cycling.

Literature on the effects of grazing pressure surrounding artificial water sources was reviewed according to three main topics of interest.

1) The patterns displayed within the piosphere:

These patterns can be highly variable depending on grazing behaviour and forage conditions, the presence of fences, soil type, the salinity of water and climatic variables. A number of monitoring indices have been suggested; however, as there appears to be no universal pattern to the piosphere, researchers have been unable to agree on monitoring systems.

2) The degradation process, including soil compaction and erosion, and the effects on fauna and flora composition:

Soil erosion increases significantly within the first 2–3 km from water as a result of heavy traffic and vegetation stripping. A significant loss of functionality in this region alters the vegetation dynamics. This includes a decline in abundance and richness of palatable forage species. Feral herbivore species have increased in abundance, contributing to the pressure exerted within these regions. Native fauna species have mostly suffered declines as a result of increased pressure and habitat alterations.

3) Biodiversity monitoring:

This section covers more recent monitoring practices including remote sensing techniques and the use of Bayesian belief network models. These methods can monitor large areas and account for both temporal and spatial variation. Bayesian belief models can incorporate both empirical data and that of expert opinion, and can predict the response of a system to human disturbance. Both these models can be incorporated to produce a powerful monitoring tool.

Desert Knowledge CRC Working Paper 15: The WaterSmart™ Literature Reviews

Contents Executive summary 1 List of figures 3 List of tables 3 1. Introduction 4

1.1 Background and history 4 1.2 Aims 5 1.3 Chapter summary 5

2. Key concepts 6 2.1 Disturbance 6 2.2 The piosphere 6 2.3 Landscape function 6 2.4 Functional integrity 6 2.5 Resistance 6 2.6 Resilience 6 2.7 Degradation 7

3. Patterns and indices 8 3.1 The variability in piosphere patterns 8 3.2 Monitoring indices 9 3.3 Chapter summary 12

4. Degradation processes 13 4.1 Soil compaction and erosion 13 4.2 Loss of landscape functionality 13 4.3 Effects of other herbivores 15 4.4 Chapter summary 15

5. Floristics of the piosphere 16 5.1 The major impacts 16 5.2 Variability in responses 16 5.3 Regional variability 17 5.4 Chapter summary 19

6. Fauna of the piosphere 20 6.1 Native versus feral animals 20 6.2 Invertebrates 21 6.3 Chapter summary 21

7. Ways forward 22 7.1 Biodiversity monitoring 22 7.2 Monitoring with remote sensing 22 7.3 Monitoring with expert knowledge 24 7.4 An example of the use of Bayesian belief network models: the Julia Creek dunnart project 25 7.5 Chapter summary 27

8. Conclusions 28 9. Appendix 1 29 10. References 34

2 Desert Knowledge CRC The impact of artificial watering points on rangeland biodiversity: A review


The impact of artificial watering points on rangeland biodiversity: A review Desert Knowledge CRC 3

List of figures

Figure 1: Sections of the aerial photograph assembly used to study track patterns 9 Figure 2: Stylised patterns of cover change with increasing distance from water related to grazing 10 Figure 3: Influence of distance from water on track density for six paddocks in the north-eastern

goldfields of Western Australia 10 Figure 4: The linear distance of landscape patches (#/100m of line) with distance from water (km) 11 Figure 5: Attributes of recent cattle grazing in relation to distance from water 11 Figure 6: The sigmoid relationship between a grazing sensitive vegetation or soil measure and distance

from a watering point 12 Figures 7A and B: Photographs displaying the differences in functionality at two sites in Crows Nest,

QLD 14 Figure 8: Relative grazing intensity of cattle on plant communities around bores 17 Figure 9: Parameters used to describe the Band 4–Band 5 data space and to derive the PD54 index 24 Figure 10: The compiled network for the probability of dunnart occurrence in the Mitchell grasslands

26

List of tables

Table 1. Area and extent of land degradation for land systems grouped into vegetation types 18


1. Introduction

1.1 Background and history

Over 70% of Australia’s mainland consists of arid and semi arid rangelands (James et al. 1999). This includes a diverse group of relatively undisturbed ecosystems such as tropical savannas, woodlands, shrublands and grasslands that harbour much of the country’s most distinctive fauna and flora (James et al. 1999, Woinarski and Fisher 2003, Smyth and James 2004). With the increase of human-induced disturbances following the settlement of pastoral grazing lands, management of rangeland biodiversity has become a real concern. Thirteen percent of Australia’s rangeland is degraded, 16% is affected by erosion, 17% is affected by woody weeds, and 25% of land cover is significantly disturbed (Smyth & James 2004). Pastoralism, altered fire regimes, the spread of pests and exotic species, and changes in water availability have been identified as the main contributors to this degradation (Woinarski and Fisher 2003). One of the more recent and pressing concerns regards the spread of artificial watering points. How the concentration of livestock around water points affects the biodiversity and sustainability of rangelands has been debated for years, with the development of many indices to monitor the process. No monitoring scheme has yet proven the most accurate and effective in determining the impact of grazing pressure around water points.

Prior to European settlement, the biota of Australia’s arid and semi arid rangelands was adapted to long periods of dry conditions. Advantage was taken of annual rainfall and the formation of temporary water sources, with significant fluctuations in abundance and distribution following seasonal rains. Flora richness and abundance usually responds well to rainfall, with the return of annual short lived forbs (Phelps 1999). Fauna species display similar correlations to the presence of rain, irrupting in abundance and extending in distribution.

The grazing potential of rangelands was recognised early in the 19th century for their ability to support sustainable and hardy pasture species, and grazing leases were released over much of eastern and southern Australia. Early grazing was dependent on water availability, and was predominately concentrated around permanent and semi-permanent water sources of the major waterways (James et al. 1999). Following the discovery of the Great Artesian Basin, large numbers of artificial water sources in the form of troughs, dams and bores were established by the 1950s (James et al. 1999). This extended the available grazing land to water-dependent livestock, and stocking rates intensified significantly. Today, artificial water sources are found at high densities throughout Australia’s grazing rangelands, with an average distance between points of less than 10 km (James et al. 1999).

The uneven distribution of cattle and grazing intensity has been recognised as a dominant problem since the 1920s (Lamon 1927, 1977 cited in Low et al. 1980) and a key threat to rangeland biodiversity (Pringle and Landsberg 2004). An important facet of biodiversity is the array of services it provides to the ecosystem. This includes gaseous compositions, amelioration of climates, fertility and stability of soils, disposal of wastes, cycling of nutrients and the natural control of pathogens and parasitic organisms (West 1993). This delicately balanced system supports an array of fauna and flora compositions and assists in the ability to deal with catastrophic events. Any disruption or alteration to this system will unbalance the production of the entire ecosystem, reducing its resistance to dramatic events. The loss of biodiversity can influence both the quality and quantity of ecosystems, while creating economic disruption (West 1993). Australia’s rangelands are heavily used as grazing lands, with strong economic inflows. Cattle and calf production generate $4.4 billion per year, while wool, sheep and lamb production generate $1 billion each year (Australian Collaborative Rangeland Information System 2001). Any alteration in the functional ability of these systems may reduce rangeland productivity, causing economic declines of a short-term or long-term nature depending on the scale of disruption.


1.2 Aims

The aim of this review is to synthesise information on the effects of artificial watering points on rangeland systems and their biota within Australia. Literature on the effects of grazing pressure surrounding water sources will be reviewed according to three main topics of interest. These are:

• 1) the patterns displayed immediately surrounding the water source and the indices used to measure these patterns

• 2) the degradation processes including the effects on fauna and flora composition • 3) monitoring possibilities using remote sensing techniques and Bayesian models.

1.3 Chapter summary

• A large proportion of Australia consists of arid and semi-arid rangelands where the sustainability of biodiversity has become a real concern.

• Water points in rangelands have reached high densities since the discovery of the Great Artesian Basin.

• Today, the average distance between water points is less than 10 km. • With the provision of water, livestock densities and distribution have increased significantly. • Grazing pressure is a key threat to the loss of rangeland biodiversity, influencing the quality and the

quantity of the ecosystem.


2. Key concepts

2.1 Disturbance

A disturbance, as defined by White and Pickett (1985), is any relatively discrete event in time that disrupts ecosystem community or population structure and changes resources, substrate variability, or the physical environment. Disturbances can be highly variable and include differences in scale, type, frequency, intensity and duration (McAlpine 1997). The state of the vegetation mosaic is often indicative of this variation in disturbance.

2.2 The piosphere

Osborn et al (1932) was the first in Australia to document the radial symmetry in grazing intensity that develops around watering points, and in 1969 Lange termed this grazing pattern ‘the piosphere’ (James et al. 1999). The central idea of the piosphere is that within large areas of pasture, the impact of livestock distribution in relation to water is patterned. Although the original term was based on the radial patterns extending from the focal water point, it has since been recognised that these patterns can be highly variable and depend not only on the behaviour of the animals, but also on the number of animals with or without free movement (Andrew 1988). This report will document the pattern variation observed within the piosphere and the factors that produce these responses.

2.3 Landscape function

This evaluates the biophysical processes that take place within a landscape. Healthy landscapes maintain the ability to retain water, soils, nutrients and organic matter. Vegetated areas act as obstructions to water flows and tend to support a richer and more productive flora composition (Ludwig 2004).

2.4 Functional integrity

A landscape with greater vegetation cover and richness often indicates good functional integrity and subsequently greater fauna and flora biodiversity (Ludwig 2004). Functional integrity measures the intactness of soil, native vegetation patterns and the processes that maintain these patterns (Ludwig 2004).

2.5 Resistance

In this report, resistance refers to the ability of flora species to survive grazing or browsing activity. Grazing resistance can be related to plant anatomy, chemistry or physiology and can be divided into avoidance or tolerance responses. Avoidance responses reduce the impact of grazing pressure by the species declining significantly in heavily grazed locations. Tolerance responses promote active growth following extensive grazing periods. The most grazing resistant plants are grasses, followed by forbs, deciduous shrubs and trees and evergreen shrubs and trees (Lyons & Hanselka nd).

2.6 Resilience

Resilience of pasture is an ability to respond to an outside pressure, shown by a strong tendency to return to its original state once the pressure has been lifted. The response changes are not fundamental to the functional ability of the system (Walker & Noy-Meir 1982).




2.7 Degradation

Land degradation refers to the undesirable changes in plant composition, soil and land surface characteristics. Land degradation in response to grazing practices in Australia’s rangelands has accelerated, with serious concerns regarding soil erosion and significant changes in vegetation composition (Bastin & James 2002). Substantial degradation can be observed with increasing proximity to water.


3. Patterns and indices

3.1 The variability in piosphere patterns

Livestock are limited in the distance they can travel due to their dependence on water (Foran 1980, James et al. 1999). Grazing impact is therefore significantly greater surrounding watering points within rangeland communities. It was originally perceived that in relatively uniform landscapes, such as chenopod shrublands, a radial grazing pattern representative of livestock movement would be observed (Lange 1969, Andrew 1988). Based on this idea of radial symmetry, the sacrifice zone immediately surrounding the water source was termed the piosphere (Figure 1). It has since been recognised that these grazing patterns can be highly variable and piospheres in the traditional sense are usually ‘warped’ by the presence of fences, grazing preferences, the salinity of water, climatic variables, and the availability of permanent and temporal water sources (Pringle & Landsberg 2004, James et al. 1999, Bastin & James 2002). These radial patterns become distorted into star shapes with axes of concentration extending into those landscape types that are preferred by the grazing animal (Pickup et al. 1998).

Differences exist in grazing behaviour between livestock species. This contributes to the uneven grazing pressure that spans the sacrifice zone. Sheep graze head first into prevailing winds and high temperatures reduce sheep grazing to just 3 km from a water source (Lange 1969). This pattern of behaviour increases the grazing pressure closer to the water source and in the direction of prevailing winds (Orr 1980). Wet conditions and lower temperatures permit foraging away from water points for extended periods of time. During this time sheep are dependent on ephemeral water and forage moisture (James et al. 1999) and grazing pressure extends further out from the water source.

In favourable forage conditions, cattle will graze just 3–4 km from a water source, and display a preference for riparian vegetative communities by grazing the same patch for hours (Hodder & Low 1978, James et al. 1999). Forage preferences may be altered in response to rainfall, and in drier conditions grazing distances may extend to 10 km (Low 1972 cited in Low et al. 1980). Water points act as the primary controlling factor in the distribution of cattle, with a ‘faithfulness’ displayed to a particular water point (Low et al. 1980, Hodder & Low 1978). This means repeated visits to the one water point, with the carving of distinctive tracks to and from that water source. Saline water and the presence of fences may also limit the distances livestock move from water. Saline conditions would increase the number of return visits, while the presence of fences reduces the area available for grazing, thus increasing the pressure observed within the piosphere. Each of these factors causes substantial variation within the piosphere (James et al. 1999) making monitoring processes difficult.



Figure 1: Sections of the aerial photograph assembly used to study track patterns in a sheep bluebush-myall piosphere on Lincoln Gap Station, South Australia. (Left) The tank and trough upon which the pattern is centered (the black strip is Highway 1 to Western Australia); (center) track patterns among myall with large canopy; and (right) bluebush with small canopy. Scale = 1: 650. (Lange 1969).

3.2 Monitoring indices

As there appears to be no universal response to the impact of grazing on biodiversity, researchers have been unable to settle on adequate management solutions and monitoring schemes. The theory of non-equilibrium behaviour displayed within the rangeland system explains the variations observed in vegetative responses to rainfall and grazing influences on a spatial scale (DeAngelis and Waterhouse 1987 cited in Pickup et al. 1998). This theory considers the resilience and resistance of vegetative species and how they vary according to grazing pressure. Some species may be more tolerant to grazing pressures, and respond well to rainfall, while others may be restricted to water remote locations (Figure 2). This variable behaviour makes it difficult to determine whether land is continuing to degrade, remaining stable or improving (Pickup et al. 1998).




Figure 2: Stylised patterns of cover change with increasing distance from water related to grazing The bottom curve illustrates cover levels in a dry period while the upper curve show cover responses following rain for undamaged (resilient) and degraded (non-resilient) landscapes. Each trend line represents a grazing gradient. ( Bastin & James 2002).

A number of indices have been used in an attempt to describe the grazing gradient that extends from the water point. These include density of herbivore tracks (Figure 3) (Pringle & Landsberg 2004), responses in vegetative cover (Figure 4) (Li & Reynolds 1993, McGarigal & Marks 1995, Ludwig et al. 1999, Bastin & James 2002), density of faecal matter (Figure 5)(Ludwig et al. 1999), and percent cover (Figure 6) (Graetz & Ludwig 1978).

Figure 3: Influence of distance from water on track density for six paddocks in the north-eastern goldfields of Western Australia (Pringle & Landsberg 2004)


Figure 4: The linear distance of landscape patches (#/100m of line) with distance from water (km) For: (a) eucalypt savannas, Kidman Springs and (b) Mitchell grasslands, Mount Sanford, Northern Territory. Also illustrated is the cover of patches (%) for: (c) Kidman Springs and (d) Mount Sanford, and the total obstruction widths for patches (m/ 100m) for: (e) Kidman Springs and (f) Mount Sanford. (Ludwig et al. 1999).

Figure 5: Attributes of recent cattle grazing in relation to distance from water Graphs show dung-pats on (c ) Kidman Springs and (d) Mount Sanford sites, Northern Territory (Ludwig et al. 1999).



Figure 6: The sigmoid relationship between a grazing sensitive vegetation or soil measure and distance from a watering point The sigmoid curve can be described by a logistic equation with parameters for an upper asymptote K, a lower asymptote zero, and a slope b. The curve is symmetrical about an inflection point GF. GF is the point of symmetry and points EG and FP, which are points of maximum slope change (Graetz and Ludwig 1978).

Each of these indices provides a correlation between grazing pressure and loss of functionality. Most of these correspond to the amount of time spent in a location and are unable to account for temporal variation. They do not accurately reflect the direct and present damage to the current system (Andrew 1988). For example, track density may be indicative of a past stocking rate and not attributable to the current grazing practices.

Vegetative indices such as the contagion index, interspersion index, and the inflexion index measure the response in vegetation cover and are more accurate in determining spatial and temporal variation. The contagion index (Li & Reynolds 1993) measures the clumping or dispersion (patchiness) of the landscape. The interspersion index (McGarigal & Marks 1995) indicates the intermixing of patches, while the inflexion index measures the extent to which vegetation is restored following good rainfall and describes the extent of grazing damage ( Bastin & James 2002). The longer a system takes to return to its natural state, the greater the damage.

3.3 Chapter summary

• Livestock movement is dependent on water, thus grazing impacts are often greatest around water sources.

• The grazing patterns observed within the piosphere can be highly variable and are dependent on the differences in livestock grazing and drinking behaviour, water condition, and the presence of fences.

• A number of indices have been used to describe the grazing gradient, providing a correlation between grazing pressure and loss of functionality. Most of these indices are unable to account for the temporal variation in vegetation responses.

• Vegetative indices may be more accurate in monitoring grazing pressures.



4. Degradation processes

4.1 Soil compaction and erosion

Heavy stock trafficking areas within the first 2–3 km of a water source severely affects the soil stability. Soil erosion increases in these areas as result of vegetation stripping and heavy trampling (Landsberg et al. 1997, James et al. 1999), while livestock tracks intensify the natural drainage patterns and exacerbate the natural erosion process (Pringle & Landsberg 2004). Most nitrogen available to plants in rangeland soils, is held within the top 10 cm of the soil due to the breaking down of organic matter. Nitrogen-fixing algae are found in the cryptogamic crusts. Heavy trampling and vegetation removal breaks this crust, creating an unstable soil surface that is vulnerable to wind and water erosion. Regeneration of vegetation is also hindered by the removal of this nutrient rich soil layer (James et al. 1999).

Soil erosion in rangelands can seriously affect the productivity of the system. Sheet erosion involves the uniform removal of soil over large areas as a result of wind or water (Carey & Silbum 2006). This form of erosion is not always obvious, and results in the loss of productive soil. In some cases, the topsoil may be totally removed leaving a scalded surface layer.

4.2 Loss of landscape functionality

Erosion results in the loss of productive and nutrient rich topsoils. These scalded areas are more susceptible to higher temperatures, have lower porosity and less microbial activity (Carey & Silbum 2006) and often result in a subsequent decline in functional integrity. Higher rates of run-off occur in these areas, reducing the infiltration and nutrient cycling of the system (Figure 7) (Ludwig 2004). This loss in soil moisture and nutrients reduces the resilience of a system and its ability to deal with catastrophic events. The loss in functional integrity has been identified as one of the likely causes of extinctions and drastic declines in native flora, small mammals and granivorous birds (Woinarski 1999 cited in Ludwig 2004). This occurs as a result of the loss or alteration of habitat, shelter, and availability of food. With the removal of vegetation, fauna in these scalded areas may become more exposed to feral and native predators.


Figure 7. A Figure 7. B

Figures 7A and B: Photographs displaying the differences in functionality at two sites in Crows Nest, QLD Figure 7A displays a landscape of high functionality: a) dominated by trees and large perennial grass tussocks, b) complex litter on the landscape surface, and c) soil under a tussock with crumb structure and open fabric. Figure 7B displays a location with much poorer functionality with: a) trees cleared to the creek line, b) greatly reduced litter and ground cover and c) soil with a massive crusted structure(Ludwig 2004).



4.3 Effects of other herbivores

The primary reason for the establishment of large numbers of artificial watering points in arid rangelands has been to support domestic livestock. An adverse affect of this has been the increase in water dependent herbivores other than livestock. The increase in the distribution and abundance of feral animals such as goats, pigs, camels, horses and donkeys, rabbits and cats is well documented (Landsberg et al. 1997, James et al. 1999). Macropods have also benefited greatly from the provision of water, and abundance of many species has increased dramatically. These animals are not restricted by fences, and grazing pressure exerted by kangaroos often matches that of livestock (Pringle & Landsberg 2004, McAlpine et al. 1999). The impact of these water dependent herbivores on the piosphere depends on individual grazing preferences, drinking behaviour and the animals’ ability to easily reach water supplies (presence of fences, the closing of water points, Finlayson troughs etc). These variations in grazing preferences and behaviour also contribute to the uneven grazing pressure that spans the sacrifice zone.

Competition for pasture between some herbivores has become a concern, and flora species that may otherwise escape domestic livestock pressure now incur heavy grazing pressure exerted by other herbivores. Kangaroos display strong preferences for particular vegetation types, with little overlap with that of domestic livestock (Landsberg & Stol 1996), and grazing by these animals can greatly retard vegetation growth even in the absence of other livestock (Hacker & Freudenberger 1997).

Rabbits have a profound effect on vegetation structure and significantly degrade the vegetation within the piosphere. Although rabbits obtain moisture from their diet and don’t depend on water sources, perennial grasses are often replaced by annuals as a result of intensive rabbit grazing and erosion caused by their burrowing (Hall & Lee 1980).

Feral goats have a wide dietary tolerance and contribute between 3 and 30% of the total grazing pressure in south-western Queensland (Thompson et al. 2002). Australia’s rangelands support 0.5–2 million goats in New South Wales, Queensland, South Australia and Western Australia (Pople et al. 1996 cited in Edwards et al. 2004). Goats inflict considerable damage to perennial vegetation while competing with domestic livestock for water and pasture (Edwards et al. 2004). They are also capable of grazing unstable, rocky areas that may otherwise provide a refuge for grazing sensitive flora (James et al. 1999). Camels, horses, donkeys and pigs also compete with livestock for forage, contribute to soil erosion, and foul and damage waterholes (Edwards et al. 2004).

4.4 Chapter summary

• Heavy livestock traffic severely affects the stability of the piosphere, particularly within in the first 2–3km from the water source.

• The removal and breakdown of the surface crust increases the erosion produced by wind and water. This removal of the nutrient rich topsoil severely affects the functional integrity of the area, and reduces the biodiversity.

• Livestock compete with other water dependent feral and native herbivores for both water and forage, increasing the grazing pressure exerted within this region.

• Macropods increase significantly in number with the provision of water and affect the grazing gradient significantly.

• The impact these animals have on the grazing gradient depends on grazing preferences, drinking behaviour and the ability to easily reach water.


5. Floristics of the piosphere

5.1 The major impacts

One of the most documented approaches to studying the effects of grazing in rangelands has been to examine the changes in flora species composition, abundance and community structure along the grazing gradient, varying from a few hundred meters to several kilometres around an artificial water source (Foran 1980, Fusco et al. 1995, Landsberg et al. 1997, Ludwig et al. 1999, Bastin & James 2002, Landsberg et al. 2003, Ludwig 2004). A number of general patterns have emerged:

• The piosphere is often degraded as a result of heavy livestock use, leading to a decline in landscape functionality and an increase in dung and urine within these zones.

• This alters the vegetative dynamics of the piosphere considerably. This zone may be bare during dry conditions, and support short–lived, unpalatable and ‘trample resistant’ species following rain (James et al. 1999).

• There is often a zone of unpalatable woody weeds beyond the denuded area surrounding water.

Martin and Ward (1970, in Ludwig et al. 1999) found that this zone dominated the region within 500m from water. Substantial changes in species composition and vegetative cover occur within this zone, with a decline in abundance and richness of palatable forage species (Figure 8). These ‘decreaser’ species are mostly native species, and significantly outnumber those that increase within the piosphere (Landsberg et al. 2003).

Short term results of this change in vegetation composition include the loss or alteration of flora species, while long term changes include the irreversible alteration of habitat from arid zones into mallee savanna habitats and the restriction of grazing intolerant species to water remote locations ( Bastin & James 2002)

5.2 Variability in responses

The degree of pressure exerted on rangeland vegetation varies greatly with pasture type, distance and direction from water, grazing behaviour, and seasonal variation. If grazing were uniform we would observe zones of different vegetation conditions forming annuli around the water points (Foran 1980).

It has been documented that grazing at moderate levels improves species richness and pasture productivity (Orr 1975). Heavy grazing, however, severely impedes the regenerative ability of native pasture species and results in a decline in species richness (James et al. 1999, Orr 1975). Some species may display more of a gradual decline in their abundance, while others may be removed from the area altogether, only to be found in water remote locations ( Bastin & James 2002). These areas may become valuable regions of plant diversity supporting grazing sensitive species (Landsberg et al. 2003).

A number of indices have been proposed to describe the variation in flora responses to grazing pressures over both temporal and spatial scales. These include the contagion index (Li & Reynolds 1993), the interspersion index (McGarigal & Marks 1995) and the inflexion index (Bastin & James 2002). Systems with a higher proportion of palatable forage are most intensively stocked as a result of this productivity. These systems usually incur higher index values and are most adversely affected by grazing ( Bastin & James 2002). Alternatively, a higher contagion index for areas of lower productively indicates less patchiness and less intermixing, which signifies non- degraded areas (Bastin & James 2002). The interspersion index for both the land system and vegetation type are also lower for the non-degraded areas compared with the total grazing area. These results indicate homogeneity within the non-degraded locations, attributable to a higher proportion of pastorally less-productive land systems or vegetation types. These systems display minimal grazing impact and return to complete recovery following rainfall.




Figure 8: Relative grazing intensity of cattle on plant communities around bores A: Sandy Bore, Napperby Station, Alice Springs during periods when majority of cattle were grazing within 3 km, 3–6 km and >6 km from bore. B: Charley and Currenjackie bores, Amburla Station, Alice Springs when cattle were grazing 0–3 km, 3–6 km and >6 km from bore (Hodder and Low 1978).

5.3 Regional variability

Australia’s arid rangelands consist of four main types of vegetation: spinifex on gibber plain, salt bush steppe, Mitchell grass tussocks and mulga shrublands. The latter two have been severely affected by grazing practices (Davies 1977). The most pastorally productive vegetation types include open woodland plains, alluvial fans and plains and floodouts, and are often the most degraded in terms of poor vegetation recovery following rainfall (Table 1). Mulga shrublands, hills and foothills, and salt lakes are less pastorally productive and display almost full recovery following rainfall ( Bastin & James 2002).


Table 1. Area and extent of land degradation for land systems grouped into vegetation types Vegetation types are arranged in order of decreasing pastoral productivity (based on stratification of land systems used by Bastin et al. 1993). The ‘distance to inflexion point’ indicates that distance from water beyond which complete recovery in wet period vegetation cover occurred. %CPL (percentage of Cover Production Loss) values were assigned to the same categories of land degradation. ( Bastin & James 2002).

Past. Prod. Vegetation type Area (km2) Dist. to infl. point (km)

%CPL value Land degradation cat.

Very High alluvial fans & plains (open woodland) 1215.1 6.6 17 severe floodouts 920.8 6.2 16 severe cracking clay plain 54.3 5.0 13 high High plains with open woodland 325.0 5.6 50 severe calcareous slopes & plains 4487.8 3.4 3 slight granitic plains– open woodland 1343.8 5.0 8 moderate sattbush & myall 384.5 4.6 17 severe gidyea woodland 1129.8 4.6 5 moderate Moderate clayey stony slopes - Chandlers I.s. 2486.6 2.4 1 slight foothills, footslopes, colluvial fans & plains 942.8 2.6 2 slight southern mulga type 1946.6 7.2 16 severe Low sand dunes & plains 14901.8 1.0 inverse salt lakes 354.0 1.4 1 slight hills & ridges 464.8 1.6 1 slight mulga shrubland 3395.9 1.8 0 nil mulga terraces 144.3 3.2 5 moderate Total 34497.9

Biomass and palatable shrub species have shown to decrease rapidly within the zone of 0–400m from water in chenopod shrublands (James et al. 1999) and the dominant perennial shrub species (Atriplex and Maireana spp) are replaced by annual chenopod subshrubs, forbs or annual grasses (Osborn et al. 1932, Graetz & Ludwig 1978). Species richness in these areas does not seem to vary consistently with distance from water (James et al. 1999).

A major shift in the composition of understorey vegetation has been identified within Acacia dominated woodlands and shrublands. The palatable perennial grasses (Thyridolepis mitchelliana and Themeda australis) are replaced by unpalatable species such as Aristida species. Heavy and sustained grazing of pasture may deplete the seed bank, which leads to local extinctions (Hodgkinson 1992 cited in James et al. 1999).

Unpalatable native shrubs such as Eremophila, Senna, Dodonaea and Acacia species have begun to dominate acacia woodlands (James et al. 1999). These shrubs remove valuable nutrients and moisture from the soils, preventing grasses from re-establishing. Vegetation composition and cover along the gradient undergo major changes, showing a decrease in ground cover with increasing distance to water (Foran 1980, James et al. 1999).

Annual grasses appear to dominate the zones closer to water within the Mitchell grasslands. The perennial grasses (Astrebla spp) increase with distance from water and tolerate moderate to light grazing intensity, displaying a large degree of resilience (Orr 1980). The lightly grazed sites or those exclosed to grazing support a greater number of perennial species, higher than average percentage of ground cover and a greater cover pattern of diversity than sites closer to water (Orr 1980).



5.4 Chapter summary

• A number of patterns have been identified in vegetative responses to grazing pressure within the piosphere: the piosphere is degraded with a loss of functionality; the vegetation dynamics are altered, with bare ground during dry conditions followed by short-lived unpalatable forage following rain; and a zone of unpalatable woody weeds dominates beyond the denuded zone. This area sees a decline in abundance and richness of palatable forage.

• There is a large degree of variation observed in vegetative responses as a result of pasture type, distance and direction from water, grazing behaviour and seasonal variation.

• The contagion index, interspersion index and inflexion point have been used to measure grazing pressure with distance to water. Regions with higher productivity incur higher index results.

• Regional variation in vegetation responses occurs as a result of varying tolerance levels to grazing pressures.

• The most pastorally productive systems are woodland plains, alluvial fans and plains and floodouts. Systems that are most severely affected by grazing practices are mulga shrublands and Mitchell grasslands.



6. Fauna of the piosphere

6.1 Native versus feral animals

The increase in feral species surrounding water sources has been well documented. These animals increase in abundance and distribution with the provision of water, often competing with livestock for forage and water supplies. What is unknown is the response of native fauna species to the presence of artificial water points.

Australia’s rangelands are unique systems that support an abundance of native fauna species, each adapted to long periods of dry conditions. Reptile and mammal species find shelter in vegetation and cracks produced in soils to avoid extreme temperatures, and often irrupt in abundance and distribution following seasonal rainfall. Natural water sources and drainage lines (billabongs, clay pans, duricrusts) support an increase in biodiversity, and encourage the return of migratory birdlife (Ford 2001, Pringle & Landsberg 2004).

Although no reptile species are known to have become extinct in the arid and semi-arid zones of Australia, a number are listed as endangered or threatened and are disadvantaged by pastoral activities. The scincid species Morethia boulengeri and Ctenotus regius are associated with leaf litter in western New South Wales and where heavy grazing removes the ground cover, these species may be locally absent (James unpublished data). A weak trend of increasing reptile species richness with increasing distance from water in the Mitchell grasslands was found by Fisher (1996); however other authors (Smith et al. 1996) have found no effect on species richness (James et al. 1999).

Most authors report declines in the geographic range of birds in response to grazing pressure. In particular, ground dwelling birds of riparian and chenopod habitats have declined greatly. Granivorous birds are largely listed as threatened bird species, with declines observed in the tropical and sub-tropical savannas and arid zones (Reid & Fleming 1992). Franklin (1999), reports that granivorous birds within tropical savannas are in a state of considerable flux. The paradise parrot (Psephotus pulcherrimus) has suffered significant range contraction and is probably extinct, while the golden- shouldered parrot (Psephotus chrysopterigus) has declined and is now considered endangered. The Gouldian finch (Erythrura gouldiae) is also endangered. Families most affected by grazing practices include pigeons, doves, finches and parrots (Franklin 1999).

While many native species have shown significant declines, some have also benefited as a result of available water. Temporary and replenished natural water sources support an increase in biodiversity, including water dependent and migratory birds (Ford 2001). The presence of water has significantly increased the abundance and distribution of emus (Dromaius novaehollandiae), which now exploit food sources from previously limited regions. Bourke’s parrot (Neopsephotus bourkii), once rare, is now common throughout the mulga zone. The crested pigeon (Ocyphaps lophotes), zebra finch (Taeniopygia guttata), and galahs (Cacatua roseicapilla) have also increased in abundance (Davies 1977, Noble et al. 1998). This increase in abundance and distribution is thought to be attributable to water and the clearing of land for grazing, which spreads seed sources.

Mammal species of Australia’s rangelands have also declined, with numerous extinctions among medium sized (35–5500g) animals in response to altered vegetative dynamics and grazing pressures (Davies 1977, James et al. 1999, Ludwig et al. 1999). The long haired rat (Rattus villosissimus) and Forrest’s Mouse (Leggadina forresti) of the Mitchell grasslands are good examples of species that irrupt in abundance in response to the onset of rains (Mifsud 1999). Both of these species are now listed as ‘vulnerable’ as a result of heavy grazing practices, the loss of suitable habitat through erosion and vegetation damage, and the introduction of feral species.

With the change in vegetation structure and the increase in water dependent herbivores, piospheres have become hunting and scavenging grounds for feral cats (Felis catus), foxes (Vulpes vulpes) and



dingoes (Canis lupus) (James et al. 1999). Cats are capable of prey switching, and don’t display any strong correlations with the abundance of a particular food source. Typical densities for feral cats in rangelands extend from 0.1 to 0.6 animals/km2, although densities as high as 6.3/km2 have been recorded in the Mitchell grass downs (Edwards et al. 2004). There is no current published literature surrounding how these animals use water points and the types of prey most affected. It is thought, however, that cats may focus their hunting effort around water points where birds are an easy target. The diet of foxes is mostly mammalian carrion, but birds, reptiles and amphibians may also be targeted around water. Dingoes prey on kangaroos around water, while carcasses of sheep and cattle during drought periods help maintain dingo and fox populations (James et al. 1999).

More research is required to determine the effects artificial watering points and the increase in grazing pressure have on Australia’s native fauna. Many species have displayed drastic declines; others have indicated a tolerance, while some have benefited from the provision of water and the spread of seeds. It is unclear how these responses influence rangeland biodiversity.

6.2 Invertebrates

The few studies that have investigated invertebrate responses to increased grazing pressure are highly variable and depend on taxonomic level, season of sampling, nature of changes and vegetation composition (James et al. 1999). Ants are good indicators of taxa disturbance, as abundance and species richness appear to alter under varying grazing pressures. Their usefulness in monitoring grazing pressures, however, is still in its infancy (James et al. 1999). Spiders also display significant differences in abundance and richness between areas with varying grazing intensities; however responses also vary among families. Termite richness has shown no direct response to grazing, and termite mounds either increase or decrease with low to moderate levels of grazing (Watson et al. 1973, Braithwaite et al. 1988, both cited in James et al. 1999).

A study conducted by Ludwig et al (1999) found that grasshopper density decreased with decreasing distance to water. The study suggests strong linkages between landscape function, biodiversity and impacts of cattle grazing and trampling. The decline in grasshopper density most likely corresponded with that of a decline in flora cover and richness, indicating a loss in landscape function. These interactions displayed between plants, animal, soil and the landscape processes need to be better understood to achieve a balance between rangeland production and conservation (West 1993, Ludwig et al. 1999).

6.3 Chapter summary

• A large degree of uncertainty still surrounds the response of native fauna to the presence of water points.

• Reptile species are threatened and disadvantaged by pastoral activities. • Most authors report declines in the geographic range of birds in response to grazing pressure. In

particular, granivorous birds have suffered great declines. • Other native bird species have benefited with the provision of water and the spreading of seed

sources through pastoral activities. • Most medium-sized mammals have declined with numerous extinctions observed in rangelands. • The loss of habitat and vegetation, and the introduction of feral animals have had a significant

impact on these animals. • Little is known regarding the invertebrate community and its response to grazing pressures and

results appear to be highly variable. • A better understanding of fauna responses to the presence of water points is required to achieve a

balance between rangeland production and conservation.



7. Ways forward

7.1 Biodiversity monitoring

Regular observations and monitoring are an integral part of management and may contribute to the understanding of how an ecosystem responds to a particular pressure, or can be used to analyse previous decisions and outcomes (Australian Collaborative Rangeland Information System 2001). For land managers, accurate information is required to make economic and ecologically viable decisions. In order to do this, managers need to be able to distinguish between the distribution of pressure, grazing induced change, seasonal fluctuations and other influencing variables (Pringle & Landsberg 2004). Monitoring processes established over the last 30 years have not always considered how these variables relate to production, biodiversity, society, economics and culture. These monitoring processes have generally consisted of ground-based data collection with a focus on pasture response to grazing domestic stock. Until now monitoring of biodiversity has received very little consideration (Australian Collaborative Rangeland Information System 2001, Smyth et al. 2003).

Recent discussions surrounding the suitability of current monitoring practices have acknowledged the need for adequate biodiversity programs that foster long term administration and public support (Smyth et al. 2003). The primary purpose of monitoring is to detect the magnitude and direction of change over long periods of time, not to establish causality. To monitor the sustainability and health of rangeland biodiversity, a number of values require consideration. These include composition (species, population, community and genetic), structure (habitat and landscape), function (functional processes, resilience, integrity and sensitivity to threats) and social perception (Smyth et al. 2003).

One method for assessing the relative condition of a system involves the use of biodiversity indicators. These indicators are a set of biotic, environmental, pressure and landscape attributes that best signal the effects of human-exerted pressures on biodiversity within the rangeland. Biotic response attributes include actual or derived measurements of biotic entities that are collected in the field using ground-based sampling techniques. They can be used to indicate the state or condition of biodiversity (species richness, the abundance of targeted species). Environmental attributes are biophysical measures of the environment (climate, topography, soil properties, vegetation or habitat characteristics) that can be measured in the field or remotely. These usually measure variables that drive biotic responses or can be used to derive landscape measures. Pressure attributes provide a measure of a threatening process caused by human activities that affect biota. The variables measured are likely to interactively affect environmental and biotic response attributes (e.g. extent and rate of clearing, change in human population density, abundance of foxes). Landscape attributes use remote-sensing techniques (satellite imagery, airborne photo/videography, GIS mapping) and measure the environmental and pressure attributes of ecosystems at multiple spatial scales. These may be used as surrogate measures of biodiversity at broad scales (leakiness index, mapping of artificial water points, habitat complexity scores) (Smyth et al. 2003). A number of indicators may be required to accurately monitor the many dimensions of biodiversity including the trends observed in the variety of ecosystems, landscapes, flora and fauna of interest.

7.2 Monitoring with remote sensing

Vegetation indices, currently used to monitor the decline in vegetation and land functionality as a result of grazing pressure, presume that the relationships between an index and vegetation cover remain constant over time (Pickup et al. 1993). This is rarely the case and suitable indices are needed to recognise both spatial and temporal variability. Pickup (1989) described how grazing induced degradation can be separated from natural variability in the landscape, and short term change due to rainfall, using spatial and temporal patterns of relative cover coupled with grazing distribution models. Pickup et al (1998) also argue that these consistent patterns of change over long periods in response to rainfall are more indicative of the true condition of the landscape, than the immediate response to major rainfall events.



A number of authors have proposed the use of Landsat Multispectural Scanner (MSS) as a means of identifying historical cover change within arid and semi-arid rangelands (Pickup et al. 1993, Karfs et al. 2000). Landsat satellite imagery provides information on vegetation responses to rainfall and allows an assessment of landscape function. Images may be selected based on climate or season history and can be used to provide a historical measure of vegetation cover and change over time (Australian Collaborative Rangeland Information System 2001).

The Audit Rangeland Implementation Project (Karfs et al. 2000) is one example of how remote sensing techniques may be implemented with the use of land resource property infrastructure, detailed monitoring sites, ancillary data, and with summaries of vegetation change derived from multi-scale Landsat Satellite data. These methods may be used to implement benchmarks and long-term operational rangeland monitoring systems. The method was trialed in Australia’s Tropical Savannas a project titled ‘Landscape Cover Change Analysis’ (LCCA), and holds the potential for further development within other rangelands systems (Karfs et al. 2000). The method is based on statistical brightness and trends through time of indices related to vegetation cover, and is used to infer land condition and to identify change within the landscape (Karfs et al. 2000). Additional outputs include temporal satellite datasets, basic stratification of land resources, identification of problem areas and their history, data on vegetation cover trends, coordinated monitoring databases derived from satellite and ground-based monitoring, increased knowledge of landscape processes and their responses to variable seasons, fire, and grazing impact and the consequences on biodiversity (Karfs et al. 2000).

Pickup et al (1993) also discussed the use of three basic approaches to vegetation cover estimation from Landsat MSS – mixture models, calibrated cover radiance relationships, and generalised vegetation indices. Mixture models separate green vegetation components, but are unsuitable for the dry components often found in arid and semi-arid rangelands. Cover radiance is derived empirically and provides absolute values for the percentage of vegetation cover. This can be costly and time consuming as data are collected through ground sampling (Pickup et al. 1998). Thirdly, generated vegetation indices provide information on the relative amount of cover present and relative changes that occur over time, rather than absolute values. The use of this data, however, is based on the assumptions that these correlations remain stable over time (Pickup et al. 1993), which rarely occurs.

One approach to estimating vegetation cover changes over time has been suggested by Graetz and others (Graetz & Gentle 1982; and Graetz et al. 1983, 1986, all cited by Pickup et al. 1998). They present the theory of the soil ‘cover line’ that extends through shadow, litter and shrubs at low radiance values, compared with bright bare soil at higher values. An increase in vegetation cover (unless green) shifts data towards the soil line (Pickup et al. 1993). While this may be adequate for single rangeland types, comparisons are unable to be made between rangelands, due to variations in soil radiance and vegetation brightness. The PD54 index (Pickup et al. 1993) is derived from the radiance properties of soil and vegetation that form a triangular structure in the four-dimensional Landsat MSS data space. The data space limits are described as the slope of the soil line (indicating the upper limit of the soil band), its offset, and the perpendicular distance to the vegetation line (Figure 9). This index is capable of denoting different levels of cover, depending on the soil background characteristics, making it one of the better indices for rangelands (Pickup et al.1993).



Figure 9: Parameters used to describe the Band 4–Band 5 data space and to derive the PD54 index (Pickup et al. 1993)

Remote sensing is a good monitoring tool as image databases consist of objective and consistently processed data, it allows monitoring over a range of scales and environments, it can be integrated with ground data to detect change and identify function in rangelands, it is cost effective, and can be applied across large areas (Australian Collaborative Rangeland Information System 2001).

7.3 Monitoring with expert knowledge

To develop an adequate biodiversity monitoring system, it is necessary to identify the core factors that may be driving the change, identify the changes occurring and what significance these changes hold for the biodiversity values, and include people with relevant expertise (Smyth et al. 2003). With the use of hierarchical models it is possible to infer the transformation of a rangeland system as a result of human or climatic alterations, while considering the uncertainty associated with this change (Gelfand et al. 2006).

The uses of Bayesian belief network models are proving to be successful tools in ecological decision making and management practices. Bayesian models are compact hierarchical models with the ability to represent all of the environmental variables and human induced disturbances that may be present within a particular system at a particular point in time. The model provides the ability to collate all forms of data and information (both empirical and hypothesised from expert opinion) and formulates the most appropriate management strategy based on this available information. It determines the degrees of confidence we have in various possible conclusions, based on an obtained body of evidence. The primary advantage of the model is the ability to predict a response to a particular event and the ability to rerun the model with different alternatives and new assumptions (Taylor 2003).

A Bayesian belief network is composed of three core elements:

1) A set of nodes representing the variables within the system, each with a finite set of mutually exclusive states that can be either discrete or continuous. 2) A set of links representing the causal relationship between these nodes. Links follow the direction from cause to effect.




3) A set of probabilities for each node specifying the belief that a node will be in a particular state, according to the states of those nodes that affect it directly (its parents). This relationship is displayed in a conditional probability table and accounts for the uncertainty surrounding the problem under investigation (Cain 2001).

7.4 An example of the use of Bayesian belief network models: the Julia Creek dunnart project

The Julia Creek dunnart (Sminthopsis douglasi) is an endangered marsupial confined entirely to the Mitchell Grass Downs and Gulf Plains bioregions of north-western Queensland (Lees 1999). The distinctive cracking clay soils and the abundance of adequate ground cover produced by the Mitchell grasses form an essential component of this species habitat preference. It is thought that the extensive cracks formed in dry, ashy, clay soils provide protection and shelter from extreme temperatures and predation by native and feral animals. As these soils become waterlogged with the onset of heavy rains, new grasses and forbs continue to provide this animal with shelter (Woolley 1992).

A number of threatening processes have contributed to the decline of this species. These include climatic factors, inappropriate grazing and land use pressures, and the impact of feral predators (Lundie-Jenkins & Payne 2002). The invasion of prickly acacia (Acacia nilotica) is also considered a significant threat as this weed competes with native vegetation, often resulting in a decline in cover and abundance of native species.

In a research project to identify and map the core areas of habitat critical to the survival of S. douglasi, a Bayesian belief network model was developed to represent the Mitchell grass system. This model was later linked to GIS data obtained for the region. Information relating to the key correlates within the Mitchell grass system was obtained from both empirical data and expert opinions. A number of experts were interviewed and surveyed and responses were used to build the model. The modeling program Netica (Norsys) was selected as the developing tool for the model. Netica has the ability to perform standard belief updating which solves the network by finding the marginal posterior probability for each node. A posterior probability is the likelihood that some parameter will be in a particular state, such as ground cover being low, given the input parameters, conditional probabilities and the rules governing how the probabilities combine (Marcot et al. 2001).

The states for each of the nodes along with the conditional probability tables were completed based on expert opinion. A number of government bodies provided GIS data for the main components of the system under investigation. This included land tenure, presence of water points, prickly acacia abundance, and soil and vegetation composition. This information was used to complete a map of the study location within the Mitchell grasslands.

Figure 10 shows the Bayesian belief network model for the Julia Creek dunnart. Although this model targets a specific species, it can be used across a range of fauna types, or may be used to display changes observed in floristic composition with distance from water points. The model can be altered on a case-by-case basis and can incorporate time scales in its output. Although some level of uncertainty surrounds the generated probabilities, these models are useful in expressing the likelihood of risk where uncertainty or bias may exist in expert judgment and where deficiencies exist in empirical data.

The information produced in this Bayesian output was used to map the likely distribution of the dunnart within the Mitchell grasslands. The Bayesian model was linked with the GIS data to produce maps of this rangeland system and the response of S. douglasi to the disturbances present within these locations.


Figure 10: The compiled network for the probability of dunnart occurrence in the Mitchell grasslands Each node displays the probability of it falling into a particular state. For example, the probability that grazing pressure is low is 52.9% and that feral animal predation is medium is 62.8%. Based on these probabilities, the likelihood of that S. douglasi would occur in any locality is 24%.



7.5 Chapter summary

• Monitoring and regular observations play an integral part in management practices. • Managers need to be able to distinguish between pressure, grazing induced change, seasonal

fluctuations and other variables. • To date, the monitoring of rangeland biodiversity has received little consideration. • To adequately monitor this, the composition, structure, function and social perception of the

ecosystem needs to be considered. • Biodiversity indicators assist in the assessment of relative ecosystem condition. These include

biotic, environmental, pressure and landscape attributes. • Suitable indicators are needed that recognise spatial and temporal variation. Landsat MSS

techniques have been proposed as a means of identifying historical change. • Various approaches to the assessment of vegetation cover have been built on this technique,

evaluating cover radiance, vegetation mixture and the development of various vegetation indices. • The PD54 index was derived from the radiance properties of soil and vegetation and is used to

indicate the different levels of cover depending on the soil characteristics. • Expert knowledge is also helpful in the absence of adequate empirical data. • Bayesian belief network models are useful in ecological decision making and management

practices. They can predict the response of a system to a certain pressure, based on prior observations.

• An example of the use of this model is given for the Julia Creek dunnart in the Mitchell grasslands.




8. Conclusions

The establishment of artificial water sources since pastoral development in Australia’s rangelands has resulted in large-scale loss of biodiversity and landscape functionality. The number and spacing of artificial water sources across the arid and semi-arid zones dramatically alters the landscape. The piosphere is the zone that immediately surrounds the water source and provides a measure of increased erosion, loss of palatable forage and an increase in unpalatable woody weeds. These negative impacts decline with distance from the water source. Native and feral animal species have increased in geographic range and abundance as a result of water availability and contribute to the grazing pressure exerted in these regions.

A number of factors determine the degradation patterns observed within the piosphere. These include the species grazing preferences, drinking behaviour, ability of free movement, resistance and resilience of vegetation species, and seasonal fluctuations.

Although some native bird species have appeared to benefit from the provision of water and the spreading of seed sources in these arid lands as a result of grazing practices, most native fauna species have displayed significant declines in abundance and distribution. The loss of landscape functionality has been attributed to the large scale loss of biodiversity within rangelands.

It is difficult to monitor change in rangeland systems by human induced activity, due to the large areas involved and the limited availability of suitable empirical data ( Bastin & James 2002). A number of monitoring indices have been proposed to measure this decline in functionality moving along the grazing gradient; however, most of these are unable to account for spatial and temporal variation. The development of remote sensing techniques (Pickup et al. 1993, Karfs et al. 2000, Bastin & James 2002) has provided the ability to monitor changes in vegetation cover across vast sections of rangeland and over varying intervals of time. Bayesian belief network models may also provide a management tool when suitable empirical data is lacking. This model incorporates both empirical data and expert opinion and provides the ability to predict the response of a particular system to human induced disturbances.

The concern regarding the spread of artificial watering points and the subsequent spread of grazing pressure is relatively new to Australia’s rangelands. Little empirical data is available regarding the loss of biodiversity within these regions, and although a number of negative speculations have been made, surprisingly little is known about the effects on biota (Landsberg et al. 1997). To date there is no agreed method for determining the effects grazing pressure has on biodiversity, nor the best approach to managing this. This review discussed the current knowledge about the effects water points have on rangeland biodiversity. It also identified the large gaps in knowledge surrounding suitable monitoring techniques. We briefly discussed the usefulness of remote sensing and hierarchical modeling which may hold potential for future research.

9. A

ppen

dix

1

Revie

w of

stud

ies co

nduc

ted

into

the e

ffect

s of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngela

nd d

egra

datio

n, an

d fa

una a

nd fl

ora c

ompo

sitio

n.

Ma

jor f

indi

ngs

Reso

urce

St

udy c

ondu

cted

La

nd d

egra

datio

n Fl

ora

Faun

a Ba

stin a

nd Ja

mes

2002

Th

is ca

se st

udy c

over

s 16 p

astor

al lea

ses

betw

een t

he T

ropic

of C

apric

orn a

nd th

e Sou

th Au

strali

an bo

rder

. Lan

d ass

essm

ent w

as

base

d on L

ands

at MS

S an

d spa

ns 2

perio

ds of

ma

jor ve

getat

ion gr

owth

at the

regio

nal s

cale

throu

gh th

e 198

0s.

The P

D54 i

ndex

was

used

to ca

lculat

e mu

ltitem

pora

l cov

er. T

his in

dex i

s der

ived f

rom

radia

nce p

rope

rties o

f soil

and v

egeta

tion.

Graz

ing gr

adien

t ana

lysis

and %

of C

over

Pr

oduc

tion L

oss w

as us

ed to

deter

mine

the

exten

t of la

nd de

grad

ation

. Indic

es us

ed w

ere

the co

ntagio

n ind

ex, in

tersp

ersio

n ind

ex an

d the

infle

xion p

oint.

Most

pasto

rally

usefu

l land

syste

ms ar

e wi

thin g

razin

g ran

ge of

wate

r, wh

ile

signif

icant

area

s of c

ountr

y with

low

pasto

ral

value

are p

rotec

ted by

dista

nce f

rom

water

. A

very

small

prop

ortio

n of to

tal ar

ea of

ve

getat

ion ty

pes w

ith hi

gh pa

stora

l pr

oduc

tivity

is no

n-de

grad

ed. M

ore o

f the

highly

prod

uctiv

e cou

ntry i

s clos

e to w

ater

and i

s mor

e inte

nsive

ly gr

azed

. Non

-de

grad

ed ar

eas a

re m

ore h

omog

eneo

us as

the

y con

tain a

high

er pr

opor

tion o

f pas

torall

y les

s pro

ducti

ve la

nd sy

stems

. Lan

d de

grad

ation

sepa

rates

patch

es an

d red

uces

the

amou

nt of

inter

mixin

g with

patch

es of

oth

er sy

stems

.

Most

pasto

rally

prod

uctiv

e sys

tems (

open

woo

dland

pla

ins, a

lluvia

l fans

and p

lains

and f

loodo

uts) h

ave t

he

highe

st %

CPL

and a

re th

e mos

t deg

rade

d in t

erms

of

lack o

f veg

etatio

n rec

over

y foll

owing

rainf

all. P

astor

ally

less p

rodu

ctive

mulg

a shr

ublan

ds, h

ills, fo

othills

and

salt l

akes

, sho

w co

mplet

e veg

etatio

n rec

over

y foll

owing

ra

infall

. Sy

stems

conta

ining

a hig

her p

ropo

rtion o

f pala

table

forag

e hav

e high

er in

dex v

alues

and a

re m

ost

adve

rsely

affec

ted by

graz

ing. A

high

er co

ntagio

n ind

ex

in no

n-de

grad

ed ar

eas s

ugge

sts le

ss pa

tchine

ss an

d les

s inte

rmixi

ng.

Inter

sper

sion v

alues

are l

ower

for n

on-d

egra

ded a

reas

of

all ve

getat

ion ty

pes.

This

indica

tes re

duce

d int

ermi

xing o

f patc

hes.

Th

e gra

zing g

radie

nt me

asur

es an

incre

ase i

n ve

getat

ion co

ver w

ith in

creas

ed di

stanc

e fro

m wa

ter.

Graz

ing in

toler

ant s

pecie

s are

usua

lly w

ater r

emote

. 5–

10%

of th

e biot

a are

graz

ing in

toler

ant o

f eve

n ligh

t gr

azing

. Mos

t veg

etatio

n typ

es ha

ve le

ss th

an 10

% of

tot

al ar

ea be

yond

10km

from

wate

r poin

t. Gra

zing

distrib

ution

is un

even

in bo

th tim

e and

spac

e.

Rece

nt re

sear

ch in

dicate

s tha

t ran

gelan

d fau

na an

d flor

a exh

ibit a

rang

e of

resp

onse

s to t

he ef

fects

of wa

ter po

ints.

Fora

n 198

0 Th

e effe

ct of

distan

ce fr

om w

ater p

oints

on

rang

eland

cond

ition w

as ex

amine

d for

2 ra

ngela

nd ty

pes i

n cen

tral A

ustra

lia. T

hese

we

re op

en w

oodla

nd co

mmun

ities a

nd m

ulga

annu

al co

mmun

ities.

Rang

eland

cond

ition w

as

asse

ssed

with

STA

RC m

ethod

whic

h co

mpar

es sp

ecies

comp

ositio

n to t

hat o

f a

benc

hmar

k reli

ct sit

e.

A sig

nifica

nt de

creas

e in c

ondit

ion w

ith

distan

ce fr

om w

ater w

as ob

serve

d for

both

rang

eland

type

s. Th

e pios

pher

e effe

ct ca

n be

attrib

utable

to m

any f

actor

s othe

r tha

n gr

azing

and s

tockin

g inte

nsity

, e.g.

wind

dir

ectio

n, sh

ade a

vaila

bility

, type

of

rang

eland

, pala

tabilit

y, cli

matic

effec

ts,

distan

ce, s

hape

of w

ater p

oint, e

tc.

Most

chan

ges i

n veg

etatio

n com

posit

ion to

ok pl

ace

withi

n 2 km

of th

e wate

r sou

rce. G

razin

g of p

astur

es is

se

ldom

unifo

rm an

d var

ies w

ith pa

sture

type

, dist

ance

an

d dire

ction

from

wate

r poin

t. Hod

der a

nd Lo

w (1

978)

fou

nd th

at ca

ttle w

ill gr

aze u

p to 3

–4 km

in lo

w to

mode

rate

forag

e con

dition

s. If g

razin

g was

unifo

rm w

e wo

uld se

e zon

es of

diffe

rent

vege

tation

type

s for

m an

nuli a

roun

d the

wate

rs, bu

t this

is no

t the c

ase.

Des

ert K

now

ledg

e C

RC

Th

e im

pact

of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngel

and

biod

iver

sity

: A re

view

29

Ma

jor f

indi

ngs

Reso

urce

St

udy c

ondu

cted

La

nd d

egra

datio

n Fl

ora

Faun

a Fu

sco e

t al. 1

995.

The l

ong t

erm

influe

nces

of liv

estoc

k gra

zing

were

stud

ied on

2 up

land C

hihua

huan

dese

rt ra

nges

in N

ew M

exico

. One

rang

e was

co

nser

vativ

ely st

ocke

d, wh

ile th

e othe

r was

sto

cked

mor

e hea

vily.

Linea

r Reg

ress

ion

Analy

sis w

as us

ed to

evalu

ate st

andin

g cro

p da

ta po

oled a

cross

year

s.

The e

ffects

of gr

azing

pres

sure

star

ts to

dimini

sh 10

00m

from

water

. The

rese

arch

int

o lon

g-ter

m us

e of li

vesto

ck ar

ound

the

piosp

here

is la

cking

. The

bree

d of c

attle

may

be im

porta

nt.

The d

ata co

llecte

d sho

w tha

t long

-term

conc

entra

tion o

f ca

ttle ca

uses

a de

cline

in pe

renn

ial gr

asse

s with

a de

creas

e in d

istan

ce fr

om w

ater.

Wee

ds an

d ann

ual

forbs

domi

nate

the zo

ne un

der 5

00m

from

water

. Ma

rtin an

d War

d (19

70) f

ound

that

regu

lating

acce

ss to

wa

ter po

ints c

ould

be ef

fectiv

e in i

ncre

asing

pere

nnial

gr

ass p

rodu

ction

, com

pare

d with

that

of co

ntinu

ous

graz

ing. R

otatin

g acc

ess t

o wate

r poin

ts do

ubled

the

yield

of pe

renn

ial gr

asse

s ove

r an 8

year

perio

d. Mo

st im

prov

emen

t occ

urre

d with

in 30

0m of

wate

r.

Hodd

er an

d Low

19

78

Cattle

wer

e obs

erve

d ove

r a pe

riod o

f 5 ye

ars

on 3

sites

near

Alic

e Spr

ings,

to de

termi

ne ho

w dif

feren

t plan

t com

munit

ies ar

e utili

sed.

The m

axim

um di

stanc

e catt

le we

re ob

serve

d fro

m wa

ter du

ring d

ry co

nditio

ns w

as 13

km.

The d

istan

ce ca

ttle gr

aze o

ut fro

m wa

ter

relat

ed to

the q

uality

and t

he qu

antity

of fo

od.

Cattle

wer

e rar

ely se

en be

yond

8 km

from

wa

ter. T

his de

pend

ence

on w

ater r

estric

ts the

for

age a

rea.

Extre

me dr

ough

t con

dition

s see

mo

veme

nt of

up to

20km

. Catt

le ap

pear

fai

thful

to a p

artic

ular w

aterin

g poin

t, only

ab

ando

ning i

t whe

n othe

r tem

pora

ry su

rface

wa

ter is

avail

able.

With

the a

bsen

ce of

surfa

ce w

ater,

cattle

graz

e clos

er to

wa

ter. T

hey m

ove o

ut as

feed

cond

itions

deter

iorate

. In

good

cond

itions

, catt

le dis

play a

mar

ked p

refer

ence

for

ripar

ian co

mmun

ities.

Land

sber

g and

Stol

19

96

The d

ensit

ies an

d dist

ributi

on of

shee

p, ka

ngar

oos a

nd fe

ral g

oats

were

asse

ssed

from

ex

tensiv

e dun

g sur

veys

follo

wing

dry,

mode

rate

and g

reen

seas

ons i

n thr

ee la

rge p

addo

cks i

n the

woo

ded r

ange

lands

of no

rth-w

ester

n NSW

.

Gr

azing

by ka

ngar

oos i

mped

es re

gene

ratio

n afte

r the

re

mova

l of li

vesto

ck. T

he to

tal gr

azing

pres

sure

de

pend

s not

only

on de

nsity

, but

on ho

w gr

azing

ac

tivitie

s are

distr

ibuted

amon

gst d

iffere

nt ve

getat

ion

types

. Ka

ngar

oos s

howe

d stro

ng se

lectiv

ity fo

r veg

etatio

n typ

es

and v

ery l

ittle o

verla

p with

goats

and s

heep

. The

y wer

e mu

ch m

ore s

electi

ve ab

out th

e env

ironm

ent in

whic

h the

y gr

aze.

Anim

al de

nsity

was

sign

ifican

tly re

lated

to

vege

tation

cove

r for

mos

t spe

cies i

n mos

t se

ason

s. Up

per c

over

(tre

es an

d shr

ubs)

was i

mpor

tant fo

r goa

ts, ot

her w

ise th

ere

were

no pr

efere

nces

. Go

ats co

ntribu

te to

as m

uch o

r mor

e tha

n sh

eep t

owar

ds th

e tota

l gra

zing p

ress

ure

durin

g the

dry s

easo

n. Ka

ngar

oos

contr

ibuted

the l

east.

Th

e dist

ributi

on of

goats

and k

anga

roos

was

co

rrelat

ed w

ith sh

eep d

istrib

ution

durin

g the

dr

y sea

son b

ut no

t at o

ther t

imes

. The

gr

eates

t com

petiti

on w

as be

twee

n she

ep

and g

oats.

Des

ert K

now

ledg

e C

RC

Th

e im

pact

of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngel

and

biod

iver

sity

: A re

view

30

Ma

jor f

indi

ngs

Reso

urce

St

udy c

ondu

cted

La

nd d

egra

datio

n Fl

ora

Faun

a La

ndsb

erg e

t al.

2003

St

udy m

easu

red t

he fr

eque

ncy o

f occ

urre

nce

of all

plan

t spe

cies a

t site

s alon

g wate

r cen

tred

graz

ing gr

adien

ts in

centr

al an

d sou

thern

Au

strali

a. Fo

ur gr

adien

ts we

re m

easu

red i

n ch

enop

od sh

rubla

nd an

d aca

cia w

oodla

nds.

Bare

grou

nd an

d gra

ss co

ver w

ere

signif

icantl

y affe

cted b

y dist

ance

to w

ater.

Bare

grou

nd an

d gra

ss co

ver d

ecre

ased

with

dis

tance

. For

b cov

er sh

owed

little

varia

tion

betw

een b

iome o

r dist

ance

zone

.

Some

spec

ies ar

e adv

antag

ed by

mod

erate

leve

ls of

distur

banc

e, wh

ile ot

hers

hold

no to

leran

ce. N

ative

sp

ecies

disp

lay m

ore o

f a de

cline

and a

re re

strict

ed to

mo

re w

ater r

emote

loca

tions

. The

decre

aser

spec

ies

outnu

mber

the i

ncre

aser

spec

ies. F

or ne

utral

spec

ies,

this m

ay re

pres

ent e

nviro

nmen

tal di

ffere

nces

. The

re

was a

larg

e deg

ree o

f spa

tial v

ariat

ion in

plan

t spe

cies

comp

ositio

n. Si

nglet

on sp

ecies

occu

r fur

ther a

way f

rom

water

. Few

area

s rem

ain fa

r eno

ugh a

way f

rom

water

to

supp

ort s

ensit

ive bi

ota. S

uch a

reas

are v

aluab

le ar

eas

of pla

nt div

ersit

y.

Low

et al.

1980

Di

stribu

tion o

f catt

le gr

azing

the r

ange

land

comm

unitie

s in c

entra

l Aus

tralia

was

de

termi

ned f

rom

aeria

l sur

veys

at fo

rtnigh

tly

inter

vals

over

a 4.5

year

perio

d. Ch

ange

s in

graz

ing in

tensit

y, as

indic

ated b

y den

sity o

f ca

ttle, w

ere e

xami

ned u

nder

diffe

rent

forag

e co

nditio

ns

Uneq

ual d

istrib

ution

of ca

ttle an

d gra

zing

inten

sity h

as be

en re

cogn

ised a

s a do

mina

nt pr

oblem

sinc

e 192

0 (Lo

w 19

72).

It is

nece

ssar

y to u

nder

stand

graz

ing

comm

unitie

s and

inten

sities

to in

terpr

et ch

ange

s in r

ange

land c

ondit

ion.

Low

(197

2) fo

und t

hat c

attle

alter

their

prefe

renc

es in

ac

cord

ance

with

rainf

all an

d for

age c

ondit

ions.

Wate

r po

ints a

ct as

the p

rimar

y fac

tor in

contr

olling

distr

ibutio

n of

cattle

. Occ

asion

al alt

erati

ons w

ere o

bser

ved i

n gr

azing

patte

rns d

ue to

envir

onme

ntal c

hang

es. T

he

longe

st an

d mos

t inten

sive g

razin

g per

iod oc

curs

an

hour

and a

half a

fter d

ay br

eak.

Cattle

graz

e fre

quen

tly

at the

same

patch

for m

any h

ours

and o

nly ch

ange

in

poor

er co

nditio

ns.

Ludw

ig et

al. 19

99

This

study

quan

tifies

the d

ensit

y, co

ver a

nd

obstr

uctio

n widt

h of v

egeta

tion p

atche

s, the

ro

ughn

ess o

f land

scap

e sur

faces

and t

he

diver

sities

of pl

ants

and g

rass

hopp

ers w

ith

distan

ce fr

om ca

ttle w

aterin

g poin

ts. D

istan

ce

from

water

was

used

as a

surro

gate

for a

grad

ient in

graz

ing pr

essu

re. F

ourte

en s

tudy

sites

wer

e loc

ated i

n the

Vict

oria

Rive

r Dist

rict

of no

rther

n Aus

tralia

A ro

ugh s

urfac

e ten

ds to

slow

wate

r run

-off

and i

ncre

ase t

he tim

e for

wate

r infilt

ratio

n an

d soil

wate

r stor

age.

Surfa

ce ro

ughn

ess

decli

nes n

ear w

ater.

Over

graz

ing ca

n cau

se

a los

s of la

ndsc

ape p

atche

s and

exce

ssive

so

il ero

sion.

In ar

id an

d sem

i-arid

en

viron

ments

, live

stock

conc

entra

te ar

ound

wa

ter so

urce

s, an

d the

n rad

iate o

ut to

feed.

A de

creas

e occ

urs i

n the

soil c

rusts

with

in the

se re

gions

. Den

sity o

f dun

g and

hoof

print

s are

high

est in

sacri

fice z

ones

.

A de

creas

e in p

alatab

le pla

nts w

as ob

serve

d clos

e to

water

. Dec

lines

wer

e obs

erve

d jus

t 200

m fro

m wa

ter.

Graz

ing gr

adien

t res

pons

es fo

r land

scap

e patc

hes d

o oc

cur w

ith un

palat

able

forbs

and g

rass

es in

creas

ing

near

wate

r. Di

versi

ty of

both

plant

and g

rass

hopp

er

spec

ies de

cline

d nea

r wate

r and

incre

ased

furth

er

away

, befo

re le

vellin

g off.

Gras

shop

per r

ichne

ss an

d den

sity

decli

ned w

ith th

e dec

line i

n veg

etatio

n clo

ser t

o wate

r. Th

ese d

eclin

es su

gges

t str

ong l

inkag

es be

twee

n lan

dsca

pe pl

ant

functi

on, b

iodive

rsity

and i

mpac

ts of

cattle

graz

ing an

d tra

mplin

g. Fe

w stu

dies

on di

stanc

e fro

m wa

ter h

ave l

ooke

d at

the ef

fects

on bi

odive

rsity.

Des

ert K

now

ledg

e C

RC

Th

e im

pact

of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngel

and

biod

iver

sity

: A re

view

31

Ma

jor f

indi

ngs

Reso

urce

St

udy c

ondu

cted

La

nd d

egra

datio

n Fl

ora

Faun

a Lu

dwig

et al.

2004

Ex

ample

s tak

en fr

om pu

blish

ed lit

eratu

re

docu

ment

that a

t fine

r patc

h and

hill-s

lopes

sc

ales s

ever

al ind

icator

s of la

ndsc

ape

functi

onal

integ

rity ha

ve be

en id

entifi

ed, e

ach

relat

ed to

biod

iversi

ty.

Func

tiona

l integ

rity is

the i

ntactn

ess o

f soil

an

d nati

ve ve

getat

ion an

d the

proc

esse

s tha

t ma

intain

thes

e. Fu

nctio

nal in

tegrity

patte

rns

have

been

mod

ified t

hrou

gh gr

azing

, cle

aring

and f

ire. G

raze

d are

as di

splay

a ra

pid de

cline

in fu

nctio

nality

for s

urfac

e sta

bility

, infilt

ratio

n and

nutrie

nt cy

cling

. Fu

nctio

nal in

tegrity

varie

s with

soil t

ype,

graz

ing an

d dist

ance

to w

ater.

Plan

t rich

ness

and g

rass

hopp

er de

nsity

was

lowe

r nea

r wa

ter po

ints.

The i

ncre

ases

in sp

ecies

near

wate

r are

sh

orter

lived

herb

s. Na

tive s

pecie

s disp

lay m

ore o

f a

decli

ne w

ith gr

azing

pres

sure

. Und

ersto

rey p

lants

and

birds

are m

ore a

ccur

ate in

dicato

rs of

graz

ing in

tensit

y.

Biod

iversi

ty sh

ould

be st

rong

ly re

lated

to

integ

rity.

Pick

up et

al. 1

998

This

pape

r sho

ws ho

w tre

nds i

n ran

gelan

d co

nditio

ns ca

n be i

denti

fied f

rom

chan

ges o

ver

time i

n the

patte

rn of

vege

tation

grow

th ac

ross

gr

azing

grad

ients

of dif

fering

inten

sity.

Graz

ing

inten

sity w

as m

easu

red u

sing d

istan

ce fr

om

water

, and

vege

tation

grow

th wa

s der

ived f

rom

remo

tely s

ense

d veg

etatio

n ind

ex va

lues b

efore

an

d afte

r larg

e rain

falls.

A ve

getat

ion re

spon

se

was d

erive

d by c

ompa

ring a

reas

less

than

4km

away

from

wate

r with

benc

hmar

k are

as fu

rther

aw

ay. T

he ve

getat

ion in

dex P

D54 d

erive

d fro

m La

ndsa

t MSS

data

was u

sed t

o mea

sure

cove

r.

Anim

al mo

veme

nt to

and f

rom

water

ing po

ints

prod

uces

radia

l patt

erns

in un

iform

coun

try,

with

graz

ing im

pact

decre

asing

with

dista

nce

from

water

. In no

n-un

iform

coun

try th

ese

patte

rns a

re di

storte

d into

star

shap

es w

ith

axes

of co

ncen

tratio

n exte

nding

into

those

lan

dsca

pes m

ost p

refer

red b

y the

graz

ing

anim

al. V

egeta

tion r

espo

nse a

lso de

pend

s on

land c

ondit

ion. S

ome a

uthor

s hav

e arg

ued

that th

e app

eara

nce o

f the l

ands

cape

fol

lowing

rainf

all is

a go

od in

dicati

on of

its

state.

Pick

up et

al. s

ugge

st tha

t this

is sk

ewed

towa

rds u

nusu

al cli

matic

situa

tions

an

d cou

ld ma

sk un

derly

ing tr

ends

. Tre

nds i

n ra

nge c

ondit

ions s

hould

be de

fined

as a

cons

isten

t patt

ern o

f cha

nge i

n the

abilit

y of a

lan

dsca

pe to

resp

ond t

o rain

fall e

vents

over

on

e or t

wo de

cade

s.

Plan

t cov

er no

t only

chan

ges t

hrou

gh tim

e with

rainf

all,

but a

lso va

ries i

n spa

ce bo

th as

a fun

ction

of gr

azing

an

d as a

resu

lt of n

atura

l var

iabilit

y. Pl

ant g

rowt

h in a

rid

and s

emi-a

rid ar

eas o

ccur

s as a

serie

s of p

ulses

fol

lowed

by pe

riods

of de

cline

as ve

getat

ion is

co

nsum

ed by

graz

ing an

d los

t by n

atura

l dec

ay.

Chan

ges i

n puls

e mag

nitud

e thr

ough

time s

hould

ind

icate

chan

ges i

n lan

dsca

pe re

silien

ce an

d if th

ey va

ry sy

stema

ticall

y alon

g gra

zing g

radie

nts, m

easu

re gr

azing

ind

uced

degr

adati

on or

reco

very.

One

way

of

stand

ardis

ing th

e veg

etatio

n res

pons

e for

diffe

rent

rainf

alls m

ay be

to ex

amine

how

the ve

getat

ion

resp

onse

to ra

infall

chan

ges o

ver t

ime a

cross

the

grad

ient. A

s a la

ndsc

ape b

ecom

es de

grad

ed th

e are

a clo

sest

to wa

ter w

ould

show

the g

reate

st re

ducti

on in

re

spon

se to

rainf

all. U

sing t

he G

IS da

ta, th

e veg

etatio

n re

spon

se is

calcu

lated

by su

btrac

ting a

vege

tation

inde

x ma

p sho

wing

cove

r befo

re a

grow

th pu

lse an

d afte

r a

grow

th pu

lse.

Des

ert K

now

ledg

e C

RC

Th

e im

pact

of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngel

and

biod

iver

sity

: A re

view

32

Des

ert K

now

ledg

e C

RC

Th

e im

pact

of a

rtific

ial w

ater

ing

poin

ts o

n ra

ngel

and

biod

iver

sity

: A re

view

33

Ma

jor f

indi

ngs

Reso

urce

St

udy c

ondu

cted

La

nd d

egra

datio

n Fl

ora

Faun

a Pr

ingle

and

Land

sber

g 200

4 Us

ing re

sear

ch re

sults

from

the G

oldfie

lds of

W

ester

n Aus

tralia

, dist

ance

from

wate

r was

inc

orpo

rated

into

mode

ls of

graz

ing hi

story

at dif

feren

t site

s with

in the

padd

ock.

Stoc

k tra

ck de

nsitie

s dec

lined

with

dista

nce

from

water

sour

ces a

nd ar

e us

ed to

es

timate

graz

ing pr

essu

re. T

his m

ay al

so

refle

ct pa

st pa

stora

l acti

vity a

s it is

diffic

ult

to de

termi

ne re

cent

beha

viour

. Are

as of

hig

h tra

ck de

nsity

refle

ct pa

st he

avy u

sage

; it r

eflec

ts a l

onge

r tim

e fra

me an

d is e

asier

to

samp

le. T

rack

dens

ity de

cline

s rap

idly

with

distan

ce fr

om w

ater.

The l

ivesto

ck

paths

may

inten

sify t

he na

tural

drain

age

patte

rns a

nd ex

acer

bate

the er

osion

pr

oces

s. W

ater u

se de

pend

s on t

he an

imal,

ag

e, se

x, wa

ter sa

linity

, temp

eratu

res,

size

and s

hape

of pa

ddoc

k, av

ailab

ility o

f na

tural

water

sour

ces,

etc.

Heav

ily gr

azed

area

s are

desir

able

habit

at for

the

spre

ad of

exoti

c spe

cies.

Fire a

ppea

rs to

impr

ove

pastu

res,

but g

razin

g nee

ds to

be co

ntroll

ed.

Comp

etitio

n betw

een l

ivesto

ck an

d mac

ropo

ds oc

curs

arou

nd pr

eferre

d pas

tures

and w

ater p

oints.

Ma

cropo

ds ar

e not

conta

ined b

y fen

ces.

The l

oss o

f na

tural

water

sour

ces e

ncou

rage

s the

conc

entra

tion o

f fau

na ar

ound

perm

anen

t wate

r sou

rces.

Rabb

its do

no

t dep

end o

n wate

r but

do se

vere

ly aff

ect th

e ve

getat

ion. F

eral

anim

als ar

e attr

acted

to th

e reg

ion,

yet a

ll use

the l

ands

cape

diffe

rentl

y.

Ther

e app

ears

to be

no un

iversa

l re

spon

se in

biod

iversi

ty to

the im

pact

of

graz

ing. A

regio

nal a

ppro

ach i

s nee

ded.

Drain

age f

eatur

es (b

illabo

ngs,

clay

pans

, dur

icrus

ts, et

c) oft

en su

ppor

t the

most

biodiv

ersit

y. Na

ture a

nd

distrib

ution

of fe

ral a

nimals

vary

betw

een s

pecie

s and

regio

ns. G

oats

rely

on w

ater,

but u

se th

e cou

ntry

differ

ently

to sh

eep a

nd ca

ttle. R

ootin

g ac

tivitie

s by p

igs ca

use d

amag

e to t

he

water

sour

ce.

Read

1999

Th

e effe

ct of

2 puls

es of

heav

y catt

le gr

azing

on

chen

opod

shru

bland

plan

ts an

d inv

erteb

rates

was

asse

ssed

over

a 2-

year

tria

l in S

outh

Austr

alia.

Stud

y loc

ation

co

nsist

ed of

sand

dune

s veg

etated

with

larg

e sh

rubs

and s

mall t

rees

. Stoc

k gra

zing h

ad

been

abse

nt be

fore s

tudy.

Both

plant

and

inver

tebra

te co

mmun

ities w

ere s

tructu

red b

y su

b-ha

bitats

defin

ed by

edap

hic an

d wate

r dis

tributi

on fe

ature

s. Du

ng co

unts

were

used

to

meas

ure r

elativ

e gra

zing p

ress

ure.

Soil s

urfac

e on t

he du

nes a

nd ru

n-on

area

s we

re se

vere

ly dis

turbe

d, bu

t micr

obiot

ic cru

sts re

maine

d lar

gely

intac

t.

The c

over

of m

ost c

ommo

n plan

t spe

cies c

hang

ed

follow

ing gr

azing

. Atri

plex v

esica

ria an

d Mair

eana

as

trotri

cha (

two o

f the m

ost d

omina

te ve

getat

ion

spec

ies) d

eclin

ed si

gnific

antly

follo

wing

graz

ing. T

he

cove

r of s

horte

r-live

d unp

alatab

le ch

enop

ods

gene

rally

incre

ased

at gr

azed

sites

. A nu

mber

of pl

ant

spec

ies in

creas

ed in

the g

raze

d are

as, p

artic

ularly

fol

lowing

good

rains

. The

se in

clude

d a nu

mber

of

unpa

latab

le sp

ecies

.

Cattle

, rab

bit an

d par

ticula

rly ka

ngar

oo

dung

wer

e mos

t abu

ndan

t at r

un-o

n sit

es an

d lea

st ab

unda

nt at

run-

off si

tes.

Dune

base

sites

typic

ally s

uppo

rted a

hig

her a

bund

ance

of in

verte

brate

s. Ru

n-off

sites

wer

e gen

erall

y imp

over

ished

in

inver

tebra

tes in

comp

ariso

n with

run-

on

sites

.


10. References

Andrew MH 1988, ‘Grazing impact in relation to livestock watering points’, Trends in Ecology and Evolution, Vol. 3, pp. 336–339.

Australian Collaborative Rangeland Information System 2001, Rangelands – Tracking Changes, National Land and Water Resources Audit, Canberra.

Bastin GN and James C 2002, Assessment of Biodiversity Condition in a Rangelands Environment Using Remote Sensing, CSIRO Centre for Arid Zone Research, Alice Springs.

Cain J 2001, Planning improvements in natural resources management: Guidelines for using Bayesian networks to support the planning and management of development programmes in the water sector and beyond, Centre for Ecology and Hydrology, Crowmarsh Gifford, Wallingford, Oxon, UK.

Carey B and Silbum M 2006, Erosion control in grazing lands, Fact Sheet, Queensland Department of Natural Resources and Water.

Davies SJJF 1977, ‘Man’s Activities and Birds’ Distribution in the Arid Zone’, Emu, vol. 77, no. 4, pp 169–172.

Edwards GP, Pople AR, Saalfeld K and Caley P 2004, ‘Introduced mammals in Australian Rangelands: Future threats and the role of monitoring programmes in management strategies’, Austral Ecology, vol. 29, pp. 40–50.

Foran B 1980, ‘Change in range condition with distance from watering point and its implication for field survey’, The Australian Rangeland Journal vol. 2, pp. 59–66.

Ford F 2001, Mitchell grass downs (AA0707), in WWFund, editor. Terrestrial Ecoregions. http://www.worldwildlife.org/wildworld/profiles/terrestrial/aa/aa0707_full.html.

Franklin DC 1999, ‘Evidence of disarray amongst granivorous bird assemblages in the savannas of northern Australian, a region of sparse human settlement’, Biological Conservation, vol. 90, pp. 53–68.

Fusco M, Holechek J, Tembo A, Daniel A and Cardenas M 1995, ‘Grazing influences on watering point vegetation in the Chihuahuan desert,’ The Journal of Rangeland Management, vol. 48, pp. 32–38.

Gelfand AE, Silander JA, Wu S, Latimer A, Lewis PO, Rebelo AG and Holder M 2006, ‘Explaining species distribution patterns through hierarchical modeling’, Bayesian Analysis, vol. 1, pp. 41–92.

Graetz RD and Ludwig JA 1978, ‘A method for the analysis of piosphere data applicable to range assessment’, The Australian Rangeland Journal, vol. 1, pp. 126–136.

Hacker RB and Freudenberger D 1997, ‘The effect of short-term exclosure of watering points on the behaviour and harvesting efficiency of grey and red kangaroos’, The Rangeland Journal, Vol. 19, pp. 145–157.

Hall TJ and Lee GR 1980, ‘Response of an Astrebla spp grassland to heavy grazing by cattle and light grazing by sheep in north-west Queensland’, The Australian Rangeland Journal, Vol. 2, pp. 83–93.



Hodder RM and Low WA 1978, ‘Grazing distribution of free-ranging cattle at three sites in the Alice Springs District, Central Australia’, The Australian Rangeland Journal, Vol. 1, pp. 95–105.

James C, Landsberg J and Morton S 1999, ‘Provision of watering points in the Australian arid zone: a review of effects on biota’, Journal of Arid Environments, Vol. 41, pp. 87–121.

Karfs K, Applegate R, Fisher R, Lynch D, Mullin D, Novelly P, Peel L, Richardson K, Thomas P and Wallace J 2000, Regional Land Condition and Trend Assessment in Tropical Savannas; The Audit Rangeland Implementation Project, National Land and Water Resources Audit Final Report, Canberra.

Landsberg J, James C, Morton S, Hobbs T, Stol J, Drew A and Tongway H 1997, The Effects of Artificial Sources of Water on Rangeland Biodiversity, Final report to the Biodiversity Convention and Strategy Section of the Biodiversity Group, Environment Australia, Canberra.

Landsberg J, James C, Morton S, Muller WJ and Stol J 2003, ‘Abundance and composition of plant species along grazing gradients in Australian rangelands’, Journal of Applied Ecology, Vol. 40, 1008–1024.

Landsberg J and Stol J 1996, ‘Spatial distribution of sheep, feral goats and kangaroos in woody rangeland paddocks’, The Rangeland Journal, Vol. 18, pp. 270–292.

Lange RT 1969, ‘The piosphere: sheep track and dung patterns’, Journal of Rangeland Management, Vol. 22, pp. 396–400.

Lees N 1999, Sminthopsis Douglasi, Environmental Management, Forest Resources, Department of Natural Resources, Canberra.

Li H and Reynolds F 1993, ‘A new contagion index to quantify spatial patterns of landscapes’, Landscape Ecology, Vol. 8, pp. 155–162.

Low WA, Müller WJ and Dudzinski ML 1980, ‘Grazing intensity of cattle on a complex of rangeland communities in central Australia’, Australian Rangeland Journal, Vol 2, 76–82.

Low WA 1972, ‘Community preference by free-ranging shorthorns in the Alice Springs district’, Proc. Aust. Soc. Anim. Prod. 9, 281–386.

Ludwig JA 2004, ‘Monitoring ecological indicators of rangeland functional integrity and their relation to biodiversity at local to regional scales’, Austral Ecology, Vol. 29, pp. 108–200.

Ludwig JA, Eager RW, Williams RJ and Lowe LM 1999, ‘Declines in vegetation patches, plant diversity, and grasshopper diversity near cattle watering-points in the Victoria River District, Northern Australia’, The Rangeland Journal, Vol. 21, pp. 135–149.

Lundie-Jenkins, G and Payne, A 2002, Recovery plan for the Julia Creek dunnart (Sminthopsis douglasi) 2000-2004, Queensland Parks and Wildlife Service, Brisbane.

Lyons RK and Hanselka CW nd, ‘Grazing and Browsing: How are Plants Affected’, Range Detect Series, The Texas A&M University System. [Online] Available at http://rangeweb.tamu.edu/extension/rangedetect/b6114_grz&brw.pdf

Marcot BG, Holthausen RS, Raphael MG, Rowland MM and Wisdom MJ 2001, ‘Using Bayesian belief networks to evaluate fish and wildlife population viability under land management alternatives from an environmental impact statement’, Forest Ecology and Management, Vol. 153, pp. 29–42.


http://rangeweb.tamu.edu/extension/rangedetect/b6114_grz&brw.pdf


McAlpine CA 1997, ‘Ecology and Spatial Modelling of Kangaroo Populations in Mosaic Landscapes of Queensland’s Semi-Arid Rangelands’, PhD Thesis, The University of Queensland, Brisbane.

McAlpine CA, Grigg GC, Mott JJ and Sharma P 1999, ‘Influence of landscape structure on kangaroo abundance in a disturbed semi-arid woodland of Queensland’, The Rangeland Journal, Vol. 21, pp.104–135.

McGarigal K and Marks BJ 1995, FRAGSTATS: spatial pattern analysis program for quantifying landscape structure, Oregon State University, Oregon.

Mifsud G 1999, ‘Ecology of the Julia Creek dunnart, Sminthopsis douglasi, (Marsupialia: Dasyuridae)’, Masters Thesis. La Trobe University, Bundoora.

Noble JC, Greene RS and Muller WJ 1998, ‘Herbage production following rainfall redistribution in semi-arid mulga (Acacia anuera) woodland in western New South Wales’, The Rangeland Journal, Vol. 20, pp. 206–226.

Orr DM 1975, ‘A Review of Astrebla (Mitchell grass) pastures in Australia’, Tropical Grasslands, Vol. 9, pp. 21–36.

Orr DM 1980, ‘Effects of sheep grazing Astrebla grassland in Central Western Queensland. II. Effects of seasonal rainfall’, Australian Journal of Agricultural Research, Vol. 31, pp. 807–820.

Osborn TG, Wood JC and Paltridge TB 1932, ‘On the growth and reaction to grazing of the perennial saltbush (Atriplex vesicarium), An ecological study of thebiotic factor’, Proceedings of the Linnean Society of New South Wales, Vol. 57, pp. 377–402.

Phelps D 1999, ‘Mitchell grass Australia’s own grass’, Queensland DPI&F Note [Online] Available at www.dpi.qld.gov.au/sheep.

Pickup G 1989, ‘New land degradation survey techniques for arid Australia: problems and prospects’, Australian Rangeland Journal, Vol 11, pp. 74–82.

Pickup G, Bastin GN and Chewings VH 1998, ‘Identifying trends in land degradation in non- equilibrium rangelands’, The Journal of Applied Ecology, Vol. 35, pp. 365–377.

Pickup G, Chewings VH and Nelson DJ 1993, ‘Estimating changes in vegetation cover over time in arid rangelands using Landsat MSS data’, Remote Sensing of Environment, Vol. 43, pp. 243–263.

Pringle H and Landsberg J 2004, ‘Predicting the distribution of livestock grazing pressure in rangelands’, Austral Ecology, Vol. 29, pp. 31–39.

Reid J and Fleming M 1992, ‘The conservation status of birds in arid Australia’, The Rangeland Journal, Vol. 14, no. 2, pp. 65–91.

Smith GT, Arnold GW, Sarre S, Abensperg-Traun M & Steven DE 1996, ‘The effect of habitat fragmentation and livestock grazing on animal communities in remnants of gimlet Eucalyptus salubris woodland in the Western Australian wheatbelt. II. Lizards’, Journal of Applied Ecology, 33, 1302–1310.

Smyth A 2003, Key conceptual, biophysical and statistical issues for designing biodiversity monitoring framework for the rangelands, Centre for Arid Zone Research, CSIRO Sustainable Ecosystems, Alice Springs.


http://www.dpi.qld.gov.au/sheep



Smyth A, James C and Whiteman G 2003, Biodiversity Monitoring in the Rangelands: A way forward, Centre for Arid Zone Research, CSIRO Sustainable Ecosystems, Alice Springs.

Smyth AK and James C 2004, ‘Characteristics of Australia’s rangelands and key design issues for monitoring biodiversity’, Austral Ecology, Vol. 29, pp. 3–16.

Taylor KJ 2003, ‘Bayesian belief networks: a conceptual approach to assessing risk to habitat’, Masters Thesis, Utah State University, Logan, Utah.

Thompson J, Reithmuller J, Kelly D, Miller E and Scanlan JC 2002, ‘Feral goats in south-western Queensland; a permanent component of grazing lands’, The Rangleand Journal, Vol. 24, pp. 268–288.

Walker BH and Noy-Meir I 1982, ‘Aspects of stability and resilience of savanna ecosystems’, pp. 556–590 in Huntley BJ and Walker BH, Ecology of Tropical Savannas, Springer Verlag.

West N 1993, ‘Biodiversity of rangelands’, The Journal of Rangeland Management, Vol. 46, pp. 2–13.

White PS and Pickett STA 1985, ‘Natural disturbance and patch dynamics: an introduction’, In: Pickett STA and White PS, Eds, The ecology of natural disturbance and patch dynamics, Academic Press, New York, pp. 3–16.

Woinarski JCZ and Fisher A 2003, ‘Conservation and the maintenance of biodiversity in the rangelands, The Rangeland Journal, Vol. 25, pp. 157–171.

Woolley PA 1992, ‘New records of the Julia Creek dunnart, Sminthopsis douglasi (Marsupialia: Dasyuridae)’, Wildlife Research, Vol. 19, pp. 779–783.

DKCRC Partners

the impact of artificial watering points on rangeland

Documents