Biodiversity publications archive

The effects of artificial sources of water on rangeland biodiversity

Final report
Jill Landsberg, Craig D. James, Stephen R. Morton, Trevor J. Hobbs, Jacqui Stol, Alex Drew and Helen Tongway
CSIRO Division of Wildlife and Ecology
Biodiversity Convention and Strategy Section of the Biodiversity Group, Environment Australia, January 1997
ISBN 0 6422 7010 4

Appendix 1 - Provision of watering points in Australian rangelands: A literature review of effects on biota

Compiled by Craig James, Jill Landsberg and Stephen Morton. To be submitted for publication.

Summary

We review the effect on native flora and fauna of the provision of artificial sources of drinking water in arid and semi-arid zones of Australia. The water comes from underground sources from which it flows under pressure (artesian bore), or is pumped (sub-artesian and unconfined aquifers), or from stored surface runoff after rain. The water is provided on the surface at troughs, dams, bore-drains or in wetlands. Artificial water points are now so numerous over most of the arid and semi-arid rangelands that their spacing is rarely more than 10 km.

Water is primarily provided for domestic livestock to drink, but it is also drunk by native and feral mammalian herbivores. Thus, a major indirect effect of the provision of artificial water is grazing by large mammals. Recorded changes in vegetation in response to grazing at artificial water points are: (1) development of a zone of extreme degradation around the water (up to 0.5 km) where soil crust is broken, erosion is high, and forbs dominate after rain; (2) increase in the number of unpalatable perennial shrubs beyond the sacrifice zone, particularly in semi-arid woodland and arid shrubland habitats; and (3) decrease in abundance of palatable native perennial grasses due to selective grazing.

Direct effects of supplying artificial waters include: (1) the development of wetland habitats where artesian bores are allowed to flow freely, supporting vegetation similar to the natural wetlands in the same area and becoming significant habitat for both flora and fauna; (2) expansion of geographic range of many bird species which drink at water points; (3) expansion of range and increase in abundance of native mammals that require drinking water (e.g., kangaroos); and (4) the possible expansion of breeding ranges of many invertebrates that require water for some stage of their life cycle.

Indirect effects of artificial water on native fauna through grazing by mammalian herbivores are poorly documented. Nothing is published on the responses of reptiles or small mammals to grazing-induced changes in habitat noted above. Published effects of grazing on native fauna include: (1) speculative discussions that attribute the recent extinction of medium-sized native mammals to grazing; (2) the displacement of some ground-dwelling bird species from heavily grazed areas and a reduction in the range in many other species; and (3) changes in the distribution and abundance of invertebrates such as grasshoppers, ants and collembolans.

Other changes that effect flora and fauna are the use of artificial waters by introduced feral animals. Foxes and cats use water points as a focus for hunting and drinking activities. Horses, donkeys and goats require drinking water and have grazing impacts on vegetation which can be substantial where population sizes are large. Feral pigs inhabit some arid and semi-arid regions where artificial water is available and they disturb floodplain vegetation with their feeding activities.

Introduction

Water is the key to biological activity in arid and semi-arid zones. A lack of water causes inactivity or death more rapidly than a lack of other essential resources such as food. Prior to the arrival of Europeans, water was a rare resource in inland Australia. Water was readily available following rain, but as the country dried out, animals had to concentrate on widely spaced natural waterholes along drainage lines. Artificial supplies of water have now been provided over vast areas of arid and semi-arid Australia through the tapping of various forms of underground water, the pooling of surface run-off water in tanks and dams, and reticulation of water by piping.

This review summarises the literature on the effect of the provision of artificial water across most of the arid and semi-arid landscapes of Australia. We include discussions of the types of artificial water, the way they are available on the surface, and the direct and indirect effects of artificial water on the arid and semi-arid zone biota.

Distribution of artificial waters

Sources of artificial water

Most rain falling on the surface of the earth either soaks into the ground or runs off as surface flow; a small amount evaporates or is taken up rapidly by plants and animals. Water soaking into the ground moves down through porous sediments under force of gravity, below the zone of evapotranspiration to become groundwater. Groundwater is transported in zones of porous sub-surface sediments, termed aquifers. Run-off water flows along drainage lines eventually emptying into lakes, seas or, in the arid zone, flooding out into flat country and soaking into the ground or evaporating. Three sources of water are used to provide permanently-available drinking water in arid and semi-arid areas of Australia. These are explained below.

Unconfined aquifers

Where groundwater is bounded below by impervious strata but is not confined by an upper impervious layer, it is termed an unconfined or phreatic aquifer or a water table (Roberts 1978). Water in unconfined aquifers is at an equivalent pressure to the atmosphere and must be pumped to the surface, for example by windmill, solar or diesel pump. Water pumped to the surface from a bore in a sandy river bed is an example of tapping an unconfined aquifer. Unconfined aquifers are often shallow in depth and associated with drainage channels.

Artesian and sub-artesian aquifers

Figure A1.1: Artesian and sub-artesian water basins in Australia

Figure A1.1: Artesian and sub-artesian water basins in Australia

Where an aquifer is confined both above and below by impervious strata, the water is under pressure and it is called a confined aquifer (Roberts 1978). The Great Artesian Basin (GAB) of eastern Australia is a good example of a confined aquifer (Fig 1). In some cases the water in a confined aquifer is under such pressure that if there is a break in the upper confining layer, water is forced to the surface. A free flowing spring or bore from a confined aquifer is called an artesian well (Roberts 1978). The pressure in a confined aquifer is not always sufficient to force water to the surface. In these cases, water must be pumped the remainder of the way and a bore of this type is called a sub-artesian well (Roberts 1978). If too much water is extracted, or too many bores are drilled into a confined aquifer, the pressure may drop and artesian wells may become sub-artesian.

The earliest proof of the existence of the GAB was from the drilling of a bore hole between the Darling and Paroo Rivers in 1880 from which water flowed (Ward 1950). Subsequently, a successful bore hole was drilled at Anna Creek near mound springs in South Australia in 1881 and from there the number of bore holes drilled grew exponentially (Ward 1950). The distribution of all artesian basins across Australia is shown in Figure A1.1. Most of the bores in these artesian sources provide water that is too salty for human consumption but is suitable for domestic stock.

The GAB is the most significant artesian water source in Australia because of its size (1.7 million km² - Fig. A1.1) and the amount of water it provides. In 1970, there were around 3,000 artesian bores and more than 20,000 sub-artesian bores yielding 1,500 million litres per day (Fisher 1969; Habermehl and Seidel 1978). Water from artesian bores in the GAB is allowed to flow freely either along drainage lines forming bore drains or wetlands, or piped into troughs. Because of the large number of free-flowing bores, pressure of artesian flow has reduced substantially: from 1914 to 1928 there was a 40% drop in pressure in NSW (Ward 1950) and many bores in southern Queensland have ceased to flow (Habermehl 1980). In the 1950's it was considered extremely efficient to allow artesian water to flow in bore drains because more country could be watered from one bore (Ward 1950). Most of the water allowed to flow into bore drains evaporates or is lost in seepage (Habermehl 1980) so this method of distribution is now being reduced by bore capping and piping (Cowley and Rogers 1995).

Stored surface run-off

Some pastoral regions of Australia do not have access to water from aquifers and in these areas water is trapped from surface run-off in dams. Dams usually hold sufficient water to last from several months to years depending on their size, the amount of leakage and the number of stock watering. Water trapped in dams can also be piped around a property to provide water in troughs at distant locations. Surface dams are most useful in areas where rainfall is relatively reliable, for example, southern pastoral areas. In areas where rainfall is less reliable extended periods without rain result in the drying of dams.

Storage and supply of artificial water

Earth dams and open drainages

Free access by all animals to surface water is possible at dams and bore drains. In these cases feral animals such as pigs and goats, and native animals such as kangaroos, can benefit from the water. Dams can also provide habitat for water fowl, though mostly dams on pastoral properties do not develop adequate vegetation to support populations of waterfowl because they are ephemeral.

Troughs and controlled access

When water is extracted from a bore by pumping, or made available by pumping water to a paddock along pipelines, it is often delivered to stock in troughs so as to control evaporation and wastage. An increasing proportion of water from artesian bores is also controlled in troughs. Water level is usually maintained by float valves and there is rarely overflow. Troughs provide opportunities for feral and native animals to drink, as do dams and bore drains. They are also the focus for predators that take prey coming to the trough for drink.

Wetlands

Artesian bores in Queensland and SA were often left to run freely because there was a perception that the artesian source of water was unlimited. On the fringes of the GAB in SA a number of these free-flowing artesian bores have developed extensive wetlands that are of substantial biological importance because native vegetation, invertebrates and vertebrates have colonised them (Badman 1987).

Density of artificial water

Artificial water points are found at high densities over very large areas of the arid and semi-arid zones of Australia. Landsberg and Gillieson (1996) mapped all named water points from the Master Names File for arid and semi-arid zones and found that only the desert regions (Simpson, Tanami, Gibson, Great Sandy and Great Victoria deserts) had substantial areas that were more than 10 km from a named water point. They also found that on average, only 29% of water points were named (ie, would be entered in the Master Names File), suggesting that average distances between water points were substantially less than 10 km over much of the grazed arid and semi-arid zones.

Morrisey (1984) reported artificial water points at an average density of one per 4000 ha in sheep-dominated regions, and one per 10 000 ha in cattle-dominated regions of the mulga rangelands in Western Australia (a total area of 700 000 km²). These figures equate to an average distance between waters of 7 km and 11.2 km for sheep and cattle lands, respectively. In the mulga woodlands of south-west Queensland, the density of artificial waters varies from around 850 to over 1000 bores and earth tanks per 1:250 000 map sheet (Ross Blick, pers. comm.). Assuming a uniform spacing between water points, this represents one water point per 17 km², or an average spacing of around 5 km between water points. In addition to bores and earth tanks, there are 1000 to >2000 km of open bore drain shown on each 1:250 000 map sheet (Ross Blick, pers. comm.).

Artificial supplies of water have become so common and reliable in the arid and semi-arid zones that the term drought has taken on a functionally different meaning. Water is available ad libitum to most mobile vertebrates. Thus, most droughts no longer result in a shortage of drinking water. Now, grazing animals do not abandon land that was once naturally waterless, but continue to graze the perennial vegetation that survives during long dry periods. But because the vegetation does not grow very much without rainfall, perennial grasses and palatable shrubs can be removed over large areas. The implications of this situation are discussed in the following sections.

A brief history of pastoral expansion in Australia

The main reason for the establishment of large numbers of artificial water sources in arid Australia was to enable the expansion of a rangeland pastoral industry into the arid interior, where feed was apparently abundant but natural sources of surface water were few.

Eastern Australia

In eastern Australia, rangeland pastoralism spread mainly into the semi-arid shrub woodlands and chenopod shrublands. The semi-arid shrub woodlands consist of the mulga (Acacia aneura) and poplar box (Eucalyptus populnea)-dominated lands of western NSW and southern Queensland. The chenopod shrublands comprise the Atriplex and Maireana-dominated shrublands of southern and western NSW, and southern SA and WA.

Grazing leases were established over most of eastern Australia by the mid-1800s. Pastoral activities were focussed on the permanent and semi-permanent waters of the major water ways including the Murray, Darling, Paroo, Lachlan, Bogan, Warrego, Bulloo, Cooper, Diamantina and Georgina Rivers (Condon 1983). During seasons of high rainfall animals could walk away from the rivers to graze back-country but most of the grazing pressure was on the riparian habitats associated with the river frontages. Even with this uneven distribution of grazing pressure, notable changes to the landscape had taken place as a result of pastoralism by 1876 (Harrington et al. 1979).

The discovery of artesian water in 1880, and the development of machinery to excavate dams, led to the creation of a few artificial water points in the 1880s (Noble and Tongway 1983). But water points were widely spread so much of the country could not effectively be grazed. By today's standards, large numbers of sheep watered at each dam (Pickard 1990). As a result, stocking rates in the effective area around water points were much greater than could be sustained. This is the scenario that led to the build-up of extraordinarily large stock numbers by the late 1800s.

In NSW, stock numbers peaked at 19 million in the 1890s and crashed to 3.5 million with the drought in 1901-02 (Newman and Condon 1969; Williams and Oxley 1979). Land degradation was extensive and the Royal Commission of 1901 (Anonymous 1901) records the situation in grim detail. It was not until the 1950s that very large numbers of water points were established because a series of favourable seasons and high wool prices gave pastoralists surplus cash to invest in capital improvements (Noble and Tongway 1983). During this period, property sizes were reduced and fencing of paddocks reduced average flock sizes from the thousands seen in the early part of the century, to a few hundred. Smaller flocks placed less stress on individual watering points and some degree of land rehabilitation was begun (Condon 1983).

South Australia

The extent and severity of land degradation in SA arid and semi-arid pastoral areas was as severe as for any other rangeland area (Ratcliffe 1936). Widespread permanent artificial water points in the form of dams initially, and bores following the discovery of artesian water (from 1880s onward), became the focus for enormous flocks of sheep (up to 10,000 animals) during dry periods (Newman and Condon 1969). Such high numbers of sheep at one watering point stripped all vegetation for kilometres around. Severe drought from 1864 to 1869 combined with heavy grazing pressure killed much of the perennial saltbush in SA. A government Commission (Anonymous 1867) into the state of the pastoral leases at the time summarised evidence of widespread death of saltbush, loss of 7-15 cm of top soil, and massive stock losses. By the 1930s, perennial shrubs had been severely depleted and carrying capacity had decreased (Newman and Condon 1969).

Western Australia

The rangelands of WA developed later than for the eastern half of the continent. Pastoralism began in the late 1800s using rivers for water, and river frontages for grazing (Saunders and Curry 1990). With the use of bore-water sheep numbers increased from 1.6 million to 4.8 million in the period 1900-1930, and fell to around 3 million by 1960 (Newman and Condon 1969). Once again droughts from 1935-1941 resulted in a lack of growth of vegetation, and what was available to graze was removed by stock that were kept alive by artificial water. Thus, the decline in stock numbers was due to a lack of food – the combined effect of little rainfall and overgrazing. Detailed descriptions of the development of the pastoral industry in the Meekatharra area are given by Mabbutt et al. (1963).

Northern Territory

As with Western Australia, development of a pastoral industry in central Australia lagged behind eastern and southern states, but a similar pattern of development ensued: stock roamed freely around permanent natural waters with a dependence contingent upon the amount and time since recent rain. With artificial water and fencing, cattle numbers grew to a peak of approximately 360, 000 in 1958 and fell to 120,000 around the time of the drought in 1964 (Newman and Condon 1969). A second phase of high cattle numbers followed good rains in the mid-1970s but they again declined during drier conditions that followed (Friedel et al. 1990).

Impacts of artificial water

We identify six main effects of artificial supplies of water in arid and semi-arid zones: (1) they are a focus for activities of domestic stock, leading to changes in native flora and fauna caused by grazing and trampling; (2) they provide wetland habitats for various animal and plant species; (3) they maintain large populations of kangaroos; (4) they provide drinking water for other native animals; (5) they provide a focus for hunting and drinking by predators; and (6) they are a focus for drinking, grazing and trampling activities of feral grazing animals. These effects are documented with relevant literature and discussed in the following sections.

Focus for grazing and trampling

Livestock radiate out from a source of drinking water to graze within a paddock. Depending on the salinity of forage, the salinity of the water and environmental conditions (e.g., temperature, cloud cover), stock may need to drink from twice per day to once every three days. Hence, the most obvious effects of artificial water are those associated with grazing and trampling by stock. Other grazing mammals such as goats and kangaroos contribute to grazing and trampling effects but domestic stock are generally the most significant source of these effects.

An important part of this review is to assess the disturbance created by stock at artificial waters and the effect of this disturbance on biodiversity. Unfortunately there are very few studies that deal with the direct effect of stock on biodiversity in the context of artificial water. There are a number of published studies on the effect of grazing on various plant and animal groups, but these are usually from studies in more mesic environments, or from studies that contrast a grazed with an ungrazed area. Much of this literature has been reviewed by Fleischner (1994) and James et al. (1995a) and we have not here attempted to include a comprehensive review of all literature on the effect of grazing on plants and animals. Instead, we have tried to give an overview of the direct and indirect effects of domestic stock which use artificial water points as a focus for activities.

Distances stock move from water

The distance from water stock will travel to feed is driven largely by vegetation condition and temperature. For cattle in central Australia, mesic periods when forage is of high quality results in cattle movements of only 4 km from water (Hodder and Low 1978). Under dry conditions, when feed is sparse or of poor-quality, movements up to 10 km result (Hodder and Low 1978), and in very poor quality habitat, or during winter, cattle may occasionally move over 20 km from water (Low et al. 1978; Table A1.1).

Table A1.1: Recorded distances domestic stock move from permanent artificial water points in arid and semi-arid Australian rangelands
Stock Distance (km) Habitat, location Source
Sheep 2.4 on saline feed Chenopod shrubland, southern SA Osborn et al. 1932.
7 in winter Atriplex-dominated shrubland (saline), southern Australia Squires 1976, 1978
3 in summer or when drinking twice/day (ie. salty water) Squires 1976, 1978
3 in summer Chenopod shrubland, southern SA Lange 1969
6.8 Merino Experimental situation Squires and Wilson 1971
8.8 Border Leicester Squires and Wilson 1971
5 km in hot weather Casuarina-Heterodendron woodland, southern NSW Lynch 1974
Cattle 4 in good season; >8 in dry season Mulga woodland, central Australia Hodder and Low 1978
3-4 km average in most seasons (max 7-8) Mulga woodland, Kunoth paddock central Australia Low et al. 1978
7-11 in most seasons (max 14.5-24 when feed sparse) Short-grass shrubland, central Australia (poor quality feed) Low et al. 1978

Sheep grazing is similarly constrained in southern rangelands. That is, high temperatures in summer necessitate frequent drinks, while wet conditions combined with low temperatures may allow sheep to forage away from permanent watering points for weeks at a time, relying on ephemeral water and the moisture content of the forage (Osborn et al. 1932; Wilson 1978). Under hot conditions, the foraging range of sheep may be reduced to as little as 3 km, particularly if the water or food is high in salts or minerals (Squires 1976, 1978; Lange 1969; Lynch 1974; Table A1.1). Chenopod shrubs are high in minerals and animals eating this vegetation must drink larger quantities of water to flush the salts from the body (Wilson and Graetz 1979).

The piosphere

Osborn et al. (1932 et al.) were the first in Australia to recognise the radial symmetry in grazing intensity that develops around a water point. Osborn et al. (1932 et al.) used this radial symmetry to examine the effects of grazing on vegetation along transects radiating from water. Subsequently, Lange (1969) coined the term piosphere for this water-focussed grazing pattern. Valentine (1947) also drew attention to the graduated use of forage away from an artificial water point in a black gramma grassland in the Chihuahuan Desert, North America. Valentine proposed that overstocking of pastures in the south-western desert areas of the United States up until that time had been the result of the failure to account for non-uniform use of forage in a paddock.

Grazing impact is greatest close to a water point and decreases with distance from the water for two reasons: (1) the area available to graze increases with distance from the focus point resulting in a reduction in density of stock; and (2) stock have to return to water regularly to drink so they are limited in how far they can travel to graze before having to return to drink. As well as grazing effects, there are also effects from trampling and dust associated with the movement of animals close to the water point (Andrew and Lange 1986a). Trampling is most obvious within 100 m of the water point and this zone is often called the "sacrifice zone".

In uniform landscapes such as chenopod shrublands, partial radial symmetry of sheep develops around artificial water points (Lange 1969). In Lange's (1969) study, track density was so high that he suggested that no shrub would have gone unattended by sheep. But radial symmetry of a piosphere is rarely uniform. Piospheres are "warped" by the constriction of fencelines (Lange 1969), preference for particular types of vegetation which are eaten disproportionately to the distance from water (Low et al. 1973 et al.; Lynch 1977; Graetz 1978; Fatchen and Lange 1979; Lange 1985), prevailing wind direction (Orr 1979 [in Foran 1980]; Andrew and Lange 1986a), and availability of shade which may force stock to move further than necessary to graze (Orr 1979 [in Foran 1980]).

Recently, techniques have been developed to assess vegetation-cover change from satellite data, and hence to assess the effect of grazing over large areas relatively inexpensively (Pickup et al. 1994 et al.). These techniques involve using reflectance values in red and green wavelengths to generate a PD54 index (Pickup et al. 1993). Differences between PD54 taken at wet and dry periods can indicate the response of different areas to rainfall, and hence the typical growth response to rain of vegetation in a particular habitat or place on the landscape. From this baseline it is possible to compare the vegetation response of land close to artificial water points with land far from water to indicate the effect of grazing. Verification of the satellite-based techniques was reported by Bastin et al. (1993 et al.) who found high values of correlation of ground-based vegetation-cover assessment with satellite vegetation-cover assessment.

Similar techniques were used by Hanan et al. (1991 et al.) to examine piosphere effects around bores in Senegal, in the Sahel region of Africa. They found no consistent patterns in primary production with increasing distance from water points during the wet season and concluded that piosphere effects on vegetation, if present, were overridden by variation due to local topography, soil and rainfall patterns.

Effects on soil surface, infiltration and nutrient status

Soils in arid environments worldwide are generally nitrogen and phosphorous deficient and soils in arid Australia are even more deficient than those in other countries (Charley and Cowling 1968). Most nitrogen that is available to plants is held in the top 10 cm of soil as a result of breakdown of organic matter (Charley and Cowling 1968) and nitrogen fixing algae in cryptogamic crusts (Mayland and MacIntosh 1966). (However, the importance of cryptogamic crusts in providing nitrogen has been questioned (Snyder and Wullstein 1973).) Heavy traffic by stock breaks up the surface cryptogam crust (Crisp 1975) which has two consequences: (1) the nitrogen-fixing action of the cryptogams is disrupted; and (2) the soil surface is loosened allowing wind and water erosion to remove surface layers. Charley and Cowling (1968) suggested that the removal of this vital surface layer prevented regeneration of degraded chenopod shrublands.

Denudation of vegetation and pulverisation of the soil crust under heavy stocking rates leads to soil erosion by wind and water. The extent of such erosion in Australian rangelands has been documented extensively over the last century of grazing (e.g., Anonymous 1901; Noble and Tongway 1983). The problem of denudation by removal of surface vegetation and subsequent erosion, is a major problem in many other countries. In the Sahelian region of Africa, the drilling of bores at regular intervals has led to extensive land degradation (Rapp 1976; Glantz 1977). Traditionally, overgrazing during the dry season resulted in areas of denudation up to about 30 km from a well, but wells were spaced at large intervals so most of the land was ungrazed. With bores now at less than 30 km intervals most of the land area is denuded by the end of the dry season (Glantz 1977). Rapp (1976) reports that cattle grazing around artificial water in northern Sudan has resulted in the denudation of vegetation and soil compaction over areas of up to 100 km diameter.

Compaction of the soil surface due to stock traffic is well documented for non-arid regions, but there remains little evidence of widespread compaction in arid and semi-arid zone rangelands (Lee 1977). Compaction appears to be restricted to heavy stock-traffic areas such as around water points and along tracks (Crisp 1975; Lee 1977). Compaction of soil along sheep tracks reduces infiltration (Tunstall and Webb 1981; Noble and Tongway 1983) thus altering soil water balance (Marshall 1974) but the direct impact of altered soil-water balance on biota does not appear to have been examined.

The accumulation of dung and urine from grazing animals is directly proportional to the distance from permanent water. This correlation has been shown for a number of systems including sheep in Australian chenopod shrublands (Lange 1969; Lange and Willcocks 1978; Lange 1985; Andrew and Lange 1986b), kangaroos in semi-arid woodlands (Gibson 1995), cattle in the Chihuahuan desert (Fusco et al. 1995) and large mammals in Africa (Weir 1971; Thrash et al. 1995). A supposed benefit of accumulation of dung and urine close to the water point from grazing mammals is the nutritive input to soils (Weir 1971; Perkins and Thomas 1993a). However, Noy-Meir and Harpaz (1976 [in Warren and Maizels 1977]) argue that dung and urine denitrify when concentrated around water points and therefore do not get returned to the soil in a form useable by plants for primary production.

Changes to vegetation cover and composition
Chenopod shrublands

Chenopod shrublands have a relatively floristically simple perennial vegetation which has been studied extensively. In reviewing changes to the chenopod shrublands, Wilson (1990) states that the major impact of grazing has been the removal of dominant Atriplex or Maireana shrubs close to water points, to be replaced by annual chenopod or forb grassland. However, there has been very little examination of the effect of grazing on the diversity of annual plants in chenopod shrublands.

Osborn et al. (1932) recorded changes in the density of Atriplex vesicaria along 8 km transects radiating from a water point. They divided the transect into four zones of use: Zone A around water points where the grazing and trampling effect was devastating; Zone B where sheep grazed most of the time; Zone C at 3-6 km where sheep grazed rarely; and Zone D at 6-8 km which was virtually ungrazed. Osborn et al. found that there was no difference in the density of shrubs among Zones B, C, or D, and that the plants in Zone B were more vigorous than those in C or D. They concluded that moderately heavy stocking was good for the vigour of the shrubland.

Most studies are consistent with the results obtained by Osborn et al. (1932). There is a zone of 0-400 m where vegetation has been dramatically altered and is dominated by annual herbs after rain (Fatchen 1978; Graetz 1978; Graetz and Ludwig 1978; Andrew and Lange 1986a). From 400 m to about 1 km there is often a zone of bush encroachment. Major piosphere effects are only detectable up to about 2 km (Graetz and Ludwig 1978). Barker (1979) documented an increase in Maireana pyramidata close to the water point (suggested to have occurred because of the accumulation of dung and urine), a decrease in populations of Atriplex vesicaria near water, and no change in M. sedifolia populations over a 2 km gradient from water. Other species that increased around a water point were Dissocarpus paradoxus (Bassia paradoxa), Scleroleana (Bassia) patenticuspis, Atriplex eardleyae, A. spongiosa and M. brevifolia. With the exception of M. brevifolia, all are annual species and represent a stage of degradation noted by Wilson (1990).

Graetz (1978) also found lowered density of Atriplex vesicaria shrubs within 250 m of the water point, but none of the changes over the rest of the paddock could be correlated with expected piosphere effects. Biomass of shrubs was reduced close to the water point, but as with shrub density, biomass did not change in a predictable way with increasing distance from water up to 4 km. Graetz also found a five-fold shift in the sex ratio of A. vesicaria. Selective grazing of female shrubs and selective grazing of habitat patches seemed to account for these results.

Graetz and Ludwig (1978) modelled the relationship between cover of chenopod shrubs and distance from water up to 3 km in pastures of southern NSW and SA. They found that sigmoid equations best described changes in ground cover. Shrub cover was low in the sacrifice zone within 400 m of the water point, increased rapidly around 1 km, and reached a stable maximum percentage cover 2-3 km from water.

Semi-arid shrub woodlands

There has been a major shift in the composition of understorey vegetation in the semi-arid woodlands of eastern Australia. These woodlands were originally dominated by an understorey of perennial grasses (e.g., Eragrostis, Monochather and Thyridolepis) and now are dominated by native shrubs, particularly species of Eremophila, Senna (Cassia) and Dodonaea (Moore 1973; Harrington et al. 1979; Hodgkinson and Harrington 1985). This shift is undoubtedly linked to increased grazing pressure associated with pastoralism, but there is still some debate about the main mechanism.

Grazing in semi-arid shrub woodlands has also changed the abundance and distribution of grasses. Many species are only affected by very heavy grazing (e.g., Monochather paradoxa and Eragrostis eriopoda), and some species are sensitive to grazing over a range of intensities (e.g., Thyridolepis mitchelliana and Aristida jerichoensis; Hodgkinson 1991, 1992).

Changes in vegetation away from bore drains also show similar trends to those identified above for point water sources (Cowley and Rogers 1995): a rapid increase in ground cover, and a decline in the proportion of annuals and unpalatable species with increasing distance from the drain. In surveying transects up to 3 km from boredrains in southern Queensland, Cowley (1994) showed that species in the family Chenopodiaceae tended to decline in relative abundance, species in the family Amaranthaceae to be most abundant at an intermediate distance, and the introduced species, buffel grass (Cenchrus ciliaris), to increase in relative abundance with increasing distance.

Arid shrub woodlands

As with the semi-arid woodlands of eastern Australia, the most dramatic change in vegetation in arid shrub woodlands is the increase in perennial native shrubs such as Acacia, Eremophila and Senna (Cassia) (Friedel 1981; Friedel et al. 1990 et al.). Foran (1980) investigated changes in vegetation condition at different distances from artificial watering points up to 6 km, in two habitats in central Australia: open woodland habitats with scattered shrubs, trees and nutritious grasses which fringe rocky ranges and are relatively productive; and mulga annual habitats that have an overstorey of mulga (Acacia aneura) and a groundcover of annual grasses and forbs but are less nutritious than those of the open woodland habitats. Foran found that open woodland habitats were degraded in quality within 2 km of the water point whereas mulga annual habitats did not change significantly. This result is consistent with the preference of cattle in central Australia for the open woodland habitat (Low et al. 1980 et al. ), and consistent with results from studies in chenopod habitats cited above.

Studies outside Australia

The effect of grazing by domestic stock on plant communities has received considerable attention, particularly in North America. Most studies have examined the effect of a fixed grazing intensity, mostly in non-arid environments (Collins and Barber 1985; Crawley 1983; Facelli 1988; Miller 1982; Shmida and Wilson 1985; Sousa 1984; review by Fleischner 1994; Fusco et al. 1995 et al.). Two general trends emerge from these studies: (1) grazing at moderate densities leads to higher within-habitat species richness compared with grazing at low or high densities (Andresen et al. 1990 et al.; Archer et al. 1987 et al.; Chaneton and Facelli 1991; Wilcox et al. 1987 et al.); and (2) very heavy grazing results in a decline in the number of species, a reduction in abundance of the remaining species and dominance by a few species (O'Connor 1991; Pandey and Singh 1991).

Recent studies that examined changes in species abundance with increasing distance from a permanent artificial water sources showed results similar to those from Australian studies (Perkins and Thomas 1993a, 1993b and Zumer-Linder 1976 in Africa; Fusco et al. 1995 in North America). Grazing by cattle virtually removed palatable species (perennial grasses) close to the water point, with substantial disruption to the soil surface, and unpalatable species dominated. With increasing distance from the water, abundance of palatable grasses increased and unpalatable species decreased.

General conclusions

Most studies to document changes in species composition, abundance or community structure of vegetation along a gradient of grazing intensity away from artificial water have done so on gradients up to 4 or 6 km in length, for sheep and cattle respectively. Within 2 km of artificial water some generalisations are apparent. When heavy grazing removes competition from palatable species, or those sensitive to trampling damage, "increaser" species establish. These are typically species with "annual" life histories that flourish after rain, or unpalatable perennial shrubs. While these studies have highlighted substantial changes over 2-3 km, there is still no indication of how plant communities that are frequently grazed differ from those that are rarely grazed (e.g., at >10 km from water).

Changes in vegetation also have an impact on soil properties, including soil fauna, with consequences for the scale of redistribution of water and nutrients leading to accelerated erosion (Noble and Tongway 1983). Because stocking rates in rangelands are generally low, these indirect effects on soil are likely to outweigh more direct effects such as trampling, except in high-use areas such as near watering points.

Changes to vertebrate animals

Studies of the impact of grazing by domestic stock on native animals have been conducted on all major terrestrial animal groups. Most of these studies have been undertaken in North America and relatively few are from Australian rangelands. Published studies have usually compared populations of reptiles, birds, small mammals and invertebrates in grazed versus ungrazed plots without replication. Some studies have compared plots with a range of densities of stock. Most studies on birds have been in riparian fringes and in North America these habitats appear to be particularly sensitive to cattle grazing (Fleischner 1994). Vertebrate groups are dealt with systematically below.

Reptiles

Published studies of the effect of grazing by domestic stock on reptiles have been reported only from North American deserts (Busack and Bury 1974; Jones 1981; Bock et al. 1990 et al.) and have all concluded that grazing significantly decreases abundance of reptiles, but only sometimes affects species richness.

In Australian rangelands, there is a general perception that grazing has severely disadvantaged reptile populations. There are no published data to support this perception. No reptile species are recorded to have become extinct in the arid and semi-arid zones (Sadlier and Pressey 1994), and few of the species currently listed as endangered or threatened are found in arid and semi-arid pastoral zones (Cogger et al. 1993 et al.). However, some species may be disadvantaged by pastoral activities. For example, the scincid species Morethia boulengeri and Ctenotus regius are strongly associated with shrub leaf-litter in western NSW (Henle 1989) and where overgrazing removes such accumulations, these species are rare or absent (C. D. James unpubl. data).

Birds

Many species of birds have become more abundant, increased their geographic range in the rangelands, or have expanded into rangelands as a result of the provision of permanent drinking water at bores, dams and tanks. Fisher et al. (1972 et al.) listed those birds that are dependent on free drinking water and those that are not. Species that are dependent on free water could only inhabit arid areas around permanent natural water, and disperse over larger areas following good falls of rain. Many of these species have expanded their range and can inhabit areas where artificial water has been provided so long as food resources are sufficient. Two examples of water dependent birds that have become more abundant and widespread are emus (Davies 1969, 1972, 1977) and Bourke's parrots (Ford 1961; Davies 1977).

Species that have been identified to have changed in abundance or range in arid and semi-arid pastoral country are identified in Table A1.2. In an extensive review of changes in the status of birds in arid Australia Reid and Fleming (1992) concluded that 19 species are at risk nationally, 50 species have declined, and 45 species have increased in range or abundance. Birds of riparian and chenopod habitats have shown the greatest decline. Reid and Fleming (1992) attribute the cause of the declines to overgrazing of these habitats, because canopy-dwelling species have been less affected than ground-dwelling species. Curry and Hacker (1990) suggest that 20 species have been advantaged in the chenopod shrublands of southern WA because of artificial water.

Table A1.2: Changes in abundance and distribution of native species of birds in arid and semi-arid rangelands attributed to the effects of the provision of artificial water or pastoralism.
Response Species Habitat Source
Extinct in many regions Black Bittern Aquatic vegetation 5,6
Lewin's Rail Aquatic vegetation 5,6
Little Wattlebird Heath 5,6
Malleefowl Mallee 2,3,4,6
Night Parrot Spinifex, Chenopod 2,4
Regent Honeyeater Temperate forest 5,6
Scarlet-chested Parrot Spinifex, Acacia scrub, Mallee 2,4
Thick-billed Grasswren Chenopod 2,4,5,6
Lowered abundance and/or reduced range Alexandra's Parrot Spinifex, Acacia scrub 2
Barking Owl Woodland 4
Black-breasted Buzzard Grassland, Spinifex, Acacia scrub, Chen. 4,5,6
Black-chinned Honeyeater Spinifex, Acacia scrub 4
Brolga Grassland, Aquatic vegetation 4
Bush Thick-knee Grassland, Chenopod 4,5,6
Channel-billed Cuckoo Not given 4
Chestnut Quail-thrush Spinifex, Acacia scrub, Mallee 4
Chestnut-breasted Quail-thrush Spinifex, Acacia scrub, Chenopod 4
Chiming Wedgebill Spinifex, Acacia scrub, Chenopod, Mallee 4
Cinnamon Quail-thrush Acacia scrub, Chenopod 4
Dusky Grasswren Spinifex 4
Eastern Grass Owl Grassland 4
Eyrean Grasswren Grassland 4
Flock Bronzewing Grassland, Spinifex, Chenopod 2,4,5,6
Grey Currawong Heath, Mallee 3,4
Grey Grasswren Grassland, Chenopod 4
Grey Honeyeater Acacia scrub 4
Hall's Babbler Acacia scrub 4
Letter-winged Kite Grassland, Acacia scrub 4,5,6
Little Button-quail Grassland, Spinifex, Acacia scrub 3,4
Magpie Goose Grassland, Aquatic vegetation 4,5,6
Masked Owl Not given 5,6
Nullarbor Quail-thrush Chenopod 2
Peregrine Falcon Heath, Spinifex 4
Pheasant Coucal Grassland, Heath 2,4
Pied Honeyeater Spinifex, Acacia scrub, Mallee 4
Pink Cockatoo Grassland, Acacia scrub, Mallee 4
Plains-wanderer Grassland 1,5,6
Purple-crowned Lorikeet Mallee 5,6
Red-browed Pardalote Spinifex, Acacia scrub 4
Red-winged Parrot Acacia scrub 4
Redthroat Spinifex, Acacia scrub, Chenopod 4
Rufous Field-wren Heath, Chenopod, Mallee 2,3,4
Squatter Pigeon Acacia scrub 5,6
Striated Grasswren Spinifex, Mallee 5,6
Tawny-crowned Honeyeater Heath, Mallee 5,6
White-fronted Chat Grassland, Heath, Chenopod 3
White-winged Fairy-wren Spinifex 2,3,4
Increased abundance and/or enlarged range Australian Crake Aquatic vegetation 2
Australian Kestrel Grassland, Heath, Spinifex, Acacia scrub 4
Australian Magpie Acacia scrub, Mallee 4
Australian Magpie Lark Grassland, Acacia scrub, Chenopod, Mallee 2,3,4
Australian Pratincole Grassland 4
Australian Raven Grassland, Heath, Spinifex, Acacia scrub 4
Banded Lapwing Grassland, Chenopod 2,3,4
Banded Whiteface Spinifex, Acacia scrub, Chenopod 2,3,4,6
Bar-shouldered Dove Acacia scrub 5,6
Black Kite Grassland, Spinifex 4
Black-faced Woodswallow Spinifex, Acacia scrub, Chenopod 3,4
Black-shouldered Kite Grassland 4
Bourke's Parrot Acacia scrub 2,3,4,6
Brown Honeyeater Heath 4
Brown Songlark Grassland, Spinifex, Acacia scrub 4
Cattle Egret Grassland 5,6
Common Bronzewing Heath, Acacia scrub, Chenopod, Mallee 2,4
Crested Pigeon Grassland, Chenopod, Mallee 2,3,4
Fairy Martin Grassland 4
Galah Grassland, Heath, Spinifex, Acacia scrub 2,3,4
Grey Butcherbird Grassland 4
Grey-crowned Babbler Acacia scrub 2,3,4
Inland Dotterel Grassland, Chenopod 2,3,4
Little Corella Acacia scrub 4,6
Little Crow Grassland, Spinifex, Acacia scrub, Chen. 3,4
Little Egret Aquatic vegetation 5,6
Mallee Ringneck Acacia scrub, Mallee 2,4
Peaceful Dove Temperate forest 4
Pied Butcherbird Acacia scrub, Chenopod, Mallee 3,4
Red-rumped Parrot Acacia scrub 4
Richard's Pipit Grassland, Spinifex, Chenopod 4
Rufous Songlark Grassland, Spinifex, Acacia scrub 4
Southern Whiteface Spinifex, Acacia scrub, Mallee 4
Spiny-cheeked Honeyeater Spinifex, Acacia scrub, Mallee 3,4
Spotless Crake Grassland, Aquatic vegetation 2
Striated Pardalote Mallee 4
Superb Fairy-wren Grassland, Heath 6
Torresian Crow Acacia scrub, Chenopod, Mallee 3,4
Welcome Swallow Grassland, Heath, Acacia scrub 2,3,4
Western Bowerbird Acacia scrub 2,3,4
White-backed Swallow Grassland, Acacia scrub, Chenopod, Mallee 4
White-breasted Woodswallow Not given 4
White-plumed Honeyeater Not given 4
Willie Wagtail Grassland, Heath, Acacia scrub, Chenopod 4
Yellow-rumped Thornbill Grassland, Spinifex, Acacia scrub, Chen. 2,3,4
Yellow-throated Miner Heath, Mallee 4
Zebra Finch Grassland, Spinifex, Acacia scrub, Chenopod 2,3,4
Abundance has increased in some areas and decreased in others Emu Grassland, Heath, Spinifex, Acacia scrub 2,3,4,6
Plumed Whistling-Duck Grassland 4,5,6
Red-tailed Black-Cockatoo Not given 3,4
Spinifex Pigeon Spinifex, Acacia scrub 2,4,6
Wedge-tailed Eagle Grassland, Heath, Spinifex, Acacia scrub 2,3,4

Sources: 1 - Harrington et al. 1988; 2 - Curry and Hacker 1990; 3 - Saunders and Curry 1990; 4 - Reid and Fleming 1992; 5 - Smith and Smith 1994; 6 - Smith et al. 1994.

Studies of the effect of grazing on birds in North American arid regions also show varied results. The abundance of some groups is greater in the presence of grazing (Bock and Webb 1984; Baker and Guthery 1990) and lower in others (Bock et al. 1984; Knopf et al. 1988). Other studies have found an overall decrease in species richness associated with grazing (Taylor 1986), no differences between grazed and ungrazed plots (Medin 1986), or higher species richness associated with grazed plots (Medin and Clary 1990). The results appear to depend on the taxonomic group involved and the habitat preference of the group. Most studies acknowledge and demonstrate that changes in bird assemblages are correlated with changes in the architecture of vegetation as a result of grazing (e.g., Taylor 1986).

Mammals

Most species of small mammals (ie., excluding kangaroos) have fared extremely poorly in semi-arid and arid pastoral lands. Throughout the arid, western half of the continent, 42% (32 species) of the mammal species that were present at the time of European settlement are now extinct (Curry and Hacker 1990). Some of these species appear to have been driven to extinction primarily by the invasion of rabbits, and the presence of foxes and cats. Other species may also have been disadvantaged by domestic stock.

A number of reviews of past and present mammalian fauna in arid and semi-arid zones have been published and are summarised in Table A1.3. All of these studies emphasise that extinctions have been numerous and that most are for mammals of "intermediate" body size (Critical Weight Range [CRW] species of Burbidge and McKenzie 1989). These reviews also show that extinctions have been equally as common in non-pastoral land as pastoral. For example, the region on which Boscacci et al. (1987) report has been grazed by sheep only since the 1950s and many of the mammals that occupied this region declined before pastoralism began.

Table A1.3: Published literature on changes in the species richness of mammalian fauna pre-European settlement to present day
Region Past number¹ % lost² Source
Arid zone 46 spp, marsupials 30% Aslin 1983
Arid zone (central and western) Rodents 44% Morton and Baynes 1985
  Polyprotodont marsupials 41%  
Nullarbor plain 23 spp. ground mammals 43% Boscacci et al. 1987
Simpson desert ? 31% Gibson and Cole 1988
Arid WA (pastoral and desert zones in Fig 1 of Burbidge and McKenzie) 51 spp, pastoral, non-volant 27.50% Burbidge and McKenzie 1989
  41 spp, desert, non-volant 39%  
  63 spp, all arid zone, non-volant 28.60%  
Arid north-western SA ? 53.50% Copley et al. 1989
Arid north-western Victoria ? 27.30% Robertson et al. 1989
Arid western half of continent 70 spp, all ground mammals 56% rodents Curry and Hacker 1990
    59% marsupials  
Uluru National Park 76 spp, all ground mammals 25% Baynes and Baird 1992
  12 spp, bats 8.30%  
Arid and semi-arid western NSW 71 spp, all mammals 38% Dickman et al. 1993

¹ Presumed species richness of fauna prior to European settlement from early records and/or sub-fossil deposits.
² Refers to probable extinction from the region under study.

Grazing by domestic stock in arid and semi-arid regions of North America has not been associated with the extinction of large numbers of native mammals as it has been in Australia. The abundance of small, ground-dwelling mammals in North America is generally greater on ungrazed plots than on grazed plots (Hanley and Page 1981; Bock et al. 1984; Bowland and Perrin 1989; Medin and Clary 1989; Putman et al. 1989; Heske and Campbell 1991). However, in one study of a seasonally grazed riparian habitat, small mammal abundance increased under grazing (Medin and Clary 1990). As with birds, some species responded positively to grazing while other species did not (e.g., Bock et al. 1984). Species richness was either not affected or was lowered by grazing (Medin and Clary 1989; Putman et al. 1989; Medin and Clary 1990).

Changes to above-ground invertebrates

Very few studies have been conducted on invertebrates in arid and semi-arid areas and those that have indicate responses to grazing are complex. The most obvious generalisation is that abundance, and species composition and richness respond to the season of sampling and changes in the architecture and species composition of vegetation that result from grazing.

Within the grasshoppers, abundance of some taxa increases in overgrazed rangeland resulting in pestiferous outbreaks (Smith 1940; Capinera and Sechrist 1982; Denny 1983). Other taxa decrease or remain at the same abundance (Capinera and Sechrist 1982). In an ungrazed exclosure in a gramma (Bouteloua sp.) grassland of southern Arizona, Jepson-Innes and Bock (1989) found grasshoppers to be 3.7 times more abundant than on nearby grazed sites during summer but the situation reversed in autumn. The difference in abundance was mainly due to the high abundance of species of the subfamily Gomphocerinae which are predominantly grass-feeding. In a more detailed study of species compositional changes of grasshoppers in relation to vegetation architecture altered by grazing, Quinn and Walgenbach (1990) found complex changes in abundance and composition. They suggest that this result was expected because of the high degree of feeding specificity among species of grasshoppers. Overall, there was a correlation between species richness of grasshoppers and the species richness of vegetation (Quinn and Walgenbach 1990).

Grazing may be responsible for the widespread decline in abundance of at least one species of grasshopper in eastern Australia. The morabine grasshopper (Keyaris scurra) used to occupy temperate grasslands now grazed by sheep, and is presently restricted to small remnants of the original vegetation, such cemetery reserves (Key 1978).

In a study of dung-feeding scarabaeid beetles in prairie grasslands in Nebraska, Jameson (1989) found few differences in the species richness or species composition, but some changes in abundance of species on grazed versus ungrazed sites. These differences were minor considering that greater amounts of dung (potential egg-laying sites for the beetles) and less ground cover (facilitating easier movement and searching for dung) were available on grazed sites.

Ants have been shown to be good indicators of disturbance (e.g., Andersen 1995a) and therefore they may show response to grazing. Heske and Campbell (1991) found no difference in the abundance of ant nests in a comparison of grazed and exclosed rangeland in the northern Chihuahuan Desert. They also found that most of the 14 species were found in approximately equal numbers inside and outside the grazing exclosure, but that there was one species that was found only inside the exclosure, and one species found only outside the exclosure. However, a recent and more detailed study of changes in ant communities along a gradient of grazing disturbance from artificial water points in the semi-arid Chaco region of Argentina, found substantial changes in species composition (Bestelmeyer and Wiens, 1996). Species richness of all sample transects at site along the gradient was highest in the least disturbed (grazed) site furthest from water during the dry season, but highest at intermediate sites during the wet season. The functional make-up of the ant communities varied along the gradient: litter-inhabiting cryptic species and specialised predators were common in less disturbed sites, whereas opportunists and hot-climate specialists dominated highly disturbed sites.

Changes to below-ground invertebrates

Little is known about the effect of grazing on below-ground invertebrates. Grazing may affect ants and termites by removing food supplies such as grass and seeds, or by collapsing burrows in areas of high traffic. However, grazing has been shown to have no general effect on termite abundance or species composition in a woodland in Australia (Abensperg-Traun 1992). Herbivore dung that accumulates near water points may facilitate an increase in the abundance of termites because dung is a frequent food resource for termites (Weir 1971).

Other below-ground taxa to have been studied are Collembola and Cicadidae (Hemiptera). Surface-dwelling Collembola decreased in species richness under heavy grazing pressure but Collembola living deep in the soil did not react as strongly (Hutchinson and King 1970; King et al. 1976 et al.). Subterranean larval stages of cicadas were lower in abundance on degraded rangeland (Milton and Dean 1992). The cause attributed to this reduction was the removal of perennial plant cover by overgrazing.

Conclusions

A clear result from this review is that changes in abundance and community structure can be quite complex, particularly for invertebrates. We have limited ability to predict grazing impacts, even in North America where the evolutionary history of grazing is relatively long (Milchunas and Lauenroth 1993). The North American work shows no general trend toward declining or increasing abundance and species richness under domestic grazing. Whether the same is true for Australian rangelands, with their generally lower productivity and evolutionary history of lower intensity grazing, remains to be determined. The studies that have been done all suggest that responses to grazing depend on the life history characteristics of the biota considered, the climate, soils and habitat architecture of the environment, the spatial scale of examination (i.e., whether within-patch or between-patch diversity is examined), severity of grazing and the season of study.

Habitat for native flora and fauna

Some free-flowing artesian bores in SA and parts of Queensland have produced extensive wetlands which have developed into habitats of major significance for biota, particularly birds. Badman (1987) surveyed 171 bores in north-eastern SA and concluded that 22 of them were significant enough to warrant conservation effort. These man-made wetlands are superficially similar in to the natural ones at places like Dalhousie Springs, but often the unique combinations of locally endemic fauna, notably fish, hydrobiid snails, and other invertebrates, are lacking in the artificial wetlands. There is usually a central core of reeds (Typha spp. and Phragmites spp.) surrounded by mudflats, often with Halosarcia spp. Beyond that, there can be a fringe of shrubs of Acacia spp., chenopod shrubs, and Nitraria spp. Where water follows a creek line, a eucalypt woodland can be maintained.

Badman (1987) indicates that the man-made wetlands at artesian bores support a large number of species of birds that otherwise would not occur in the arid north-eastern corner of SA. If the wetlands were to be removed, only one species (brolga) may disappear from SA; the rest would persist in other areas or at lower densities.

Artificial water in the form of wetlands, dams or tanks may also provide habitat for a range of aquatic or semi-aquatic species. However, Weir (1971) suggests that permanent water supplied to claypans from bores in the Wankie National Park, Zimbabwe, has disadvantaged crustaceans, Conchostraca and Anstraca which were adapted to a wet-dry season flooding of claypans. Other taxa that rely on water for habitat or to complete a life cycle such as dragonflies (Odonata), may have increased their range or duration of occupancy in a region because of artificial water (Weir 1971); however this has not been investigated in Australia.

Maintenance of high abundance of kangaroos

In recent pre-European times, populations of kangaroos and wallabies which grazed native grasses and perennial vegetation probably declined during drought when surface water became restricted to refuge areas. Under these conditions, kangaroo populations may have been concentrated around riparian areas, expanding into the surrounding country only after good rainfalls when green forage and free water were available.

A number of authors have made mention of the prosperity of large kangaroos since the arrival of Europeans: large populations of red (Macropus rufus), eastern grey (M. giganteus) and western grey (M. fuliginosus) kangaroos, or wallaroos (M. robustus) are present in most pastoral regions (Main et al. 1959 et al. ; Frith 1964; Calaby 1971; Doohan 1971; Newsome 1971, 1975; Cunningham 1981; Denny 1985; Norbury 1992). It is impossible to judge the population sizes of kangaroos prior to the advent of the pastoral industry because the records in explorers diaries are unreliable (Noble and Tongway 1983). But, the widespread generality of abundant kangaroos in pastoral areas has led to the hypothesis that artificial water helps maintain greater populations than would be possible without the water.

It is clear that kangaroos are most numerous around artificial water sources, particularly during dry periods such as summer or drought (Newsome 1965, Norbury and Norbury 1993; Gibson 1995). Kangaroos require most water in summer for thermoregulation (Dawson et al. 1975 et al.) and removing sources of artificial water results in a decline in kangaroo abundance (Gibson 1995).

This facilitation of kangaroo populations by artificial water probably is most important for grey kangaroos and wallaroos, but of lesser importance for red kangaroos. Densities of red kangaroos vary greatly across the geographic ranges of the species and vary greatly in time within a region despite the provision of artificial water. Caughley et al. (1977) et al. estimated the abundance of red kangaroos in western NSW at 4.2 per km². During a drought in 1982, populations of red kangaroos in western NSW declined by 40% (Caughley et al. 1985 et al.). This decline was attributed to a shortage of food, not water. Caughley et al. (1980) et al. also sought to explain a 1-2 order-of-magnitude higher density of red kangaroos in northern NSW compared to adjacent habitats in South Australia and Queensland. They argued that lower densities of red kangaroos were due to predation by high populations of dingoes, which were being maintained in Queensland and SA by the availability of alternate prey, mainly rabbits.

Newsome (1965) also provides indirect evidence that suggests large populations of red kangaroos may be the result of lack of predation rather than the ad libitum availability of drinking water. During drought in central Australia, Newsome (1965) found that 83% of the red kangaroos were found within 2 km of grasslands where green pick was available. After rain, only 48% were so close to the same grasslands. The distribution of red kangaroos did not change so dramatically with respect to permanent water with 68% being near water during drought and 63% near water after rain. Thus, Newsome suggested that red kangaroos moved in response to the availability of nutritious food rather than to water. Newsome's (1965) study, along with others showed the changes in abundance of red kangaroos across the dingo fence (Caughley et al. 1980), and the limited need for red kangaroos to drink (Newsome 1964; Dawson et al. 1975).

Focus for drinking by other native animals

Many of the references cited above indicate native vertebrate species that use artificial water points for drink. Some kangaroos and birds such as crows, galahs, crested pigeons and parrots are good examples of species that drink at artificial waters and have expanded their geographic ranges or increased in abundance because of the provision of artificial water.

Williams and Wells (1986) investigated the effect of artificial water and grazing on populations of birds in a semi-arid mallee environment in SA. The presence of water facilitated larger populations of birds in all seasons except spring, but transects that were grazed had fewer birds than transects that were not grazed. Species richness was higher on transects with artificial water than transects without.

Artificial water from bores is also a focus for mammals in arid and semi-arid regions of southern Africa. Thrash et al. (1995) et al. studied the distribution and abundance of herbivorous mammals around bore holes in the Kruger National Park, Republic South Africa, in the dry season. They found that most animals were found within 1 km of the water point. In contrast, van der Walt (1986) concluded that only springbok and blue wildebeests showed a preference for habitat around water in the Kalahari-Gemsbok National Park, and mainly during winter.

Focus for hunting/drinking by predators

As a direct result of native species using artificial water as sources of drink, predators converge on water to drink and hunt. The predators commonly seen at water points are cats, dingoes and foxes.

Cats

There appears to no published information documenting how frequently or intensively cats predate animals at water points in arid environments. However, it seems to be common anecdotal knowledge that they do. Examples of this anecdotal knowledge are: (1) cats hide close to troughs and capture birds which come to drink; and (2) 90% of the cats collected during a drought, for a current project on the ecology of cats in central Australia, were collected around artificial water points (G. Edwards, pers. comm.).

Cats probably do not need to drink except during very hot, dry periods. Hence, their presence around water points may be primarily related to the abundance of prey. There is no published information on the types of prey most frequently captured at artificial water points, but birds are a likely target (Catling 1988; Jones and Coman 1981; G. Edwards pers. comm.). Anecdotal reports suggest that cats catch a range of water-dependent birds that come to drink. One comparative study of the diets of cats in different environments (Jones and Coman 1981) shows that birds are more frequently eaten by cats living in an arid environment than they are by cats in a mesic environment. This may be because cats focus their activity on artificial water points where birds are common.

Cats are generally thought to be disadvantaged by the presence of dingoes, which do frequent water points, because dingoes kill cats. However, cats may also be advantaged by the presence of dingoes: dry conditions in central Australia in 1994 forced dingoes, cats and kangaroos onto water points where dingoes killed kangaroos and cats scavenged on the remains of the kangaroos (G. Edwards, pers. comm.).

Little is known of the direct effect of cats on biodiversity, although cats have been implicated in the extinction of many medium-sized native mammals (Morton 1990; Smith and Quin, in press ). This role is probably as one of the predators that delivers a coup de grace to small populations already in decline because of other disturbances in their environment.

Dingoes and foxes

Dingoes and foxes may have been advantaged by the provision of artificial water. Both species need to drink regularly in hot weather (Newsome and Coman 1989) and are therefore tied to sources of free water during dry periods. Thus, populations are probably greater, and maintained over a larger area, than would be possible without artificial water (Newsome and Corbett 1977). Dingoes are major regulators of kangaroo populations in eastern Australia (Caughley et al. 1987), and as noted above, dingoes prey on kangaroos around water points. Death of cattle and sheep around water points during drought also helps maintain populations of dingoes and foxes. Foxes, as with cats, are implicated in the extinction of many medium-sized native mammals (Morton 1990; Smith and Quin, in press).

Focus for foraging/drinking by feral animals

Camels, horses, donkeys, pigs, goats and rabbits all use artificial water for drink. The impact of horses, donkeys and camels is relatively localised in extent and most pronounced during droughts when permanent waters become the focus for large numbers of domestic and feral animals.

Camels

Camels live in desert country and are independent of drinking water during periods when forage is green (McKnight 1969 [cited in Low 1974]). Camels only use artificial water during extreme droughts (Newsome and Corbett 1977). During these times, large aggregation can cause considerable damage to vegetation but at other times, their low density does not appear to result in concentrated impact. The effect of camels on arid and semi-arid biota is not known.

Horses and donkeys

Horses and donkeys require drinking water regularly. They can travel further than cattle to water and can move to areas where feed is more plentiful between drinks (Low 1974; Berman and Jarman 1987). In central Australia, horses will graze floodout and riparian habitat after rain when palatable plants are available. As the forage dries, horses move into rocky country to obtain nutritious feed (Berman and Jarman 1987). Donkeys are also most abundant in rocky country (Giles 1978). While horses prefer more nutritious feed than cattle (Berman and Jarman 1987) the general effect of grazing by horses and donkeys is likely to be similar to that of cattle.

Pigs

Pigs utilize wetland habitats around free flowing bores or natural wetland areas often associated with earth dams. They are the only feral species to require man-made wetland areas for a habitat; all other feral mammals use artificial waters primarily for drinking. They eat green herbage or roots close to wetland areas and water. Populations densities have been recorded as high as 80/km² in the Macquarie Marshes, and 20/km² near a stock watering point in northwest NSW (Giles 1978). With such large population densities they can do considerable damage to vegetation while foraging. Their effect on the diversity of native biota is unknown but they can be significant predators of native animals.

Goats

Goats can be as numerous as sheep in areas such as northern NSW and southern WA (Freudenberger 1993; Landsberg and Stol 1996). They use artificial water in a similar way to sheep. Direct and indirect changes to vegetation and fauna, attributable to grazing by sheep, can also be attributed to grazing by goats, although goats differ slightly in their vegetation preferences (Landsberg and Stol 1996).

Rabbits

Most of the time rabbits get the water they need from forage, but occasionally they are observed drinking at artificial water points (Cooke 1982). This seems to happen when the forage is too dry or too scarce to provide the necessary water. Cooke (1982) suggested that such circumstances arose because the rabbits had eaten-out the succulent vegetation and high mortality occurred because of malnutrition.

Concluding comments

Artificial water has primarily been provided to facilitate the production of domestic stock. The number and spacing of artificial water points across the arid and semi-arid pastoral rangelands of Australia have irrevocably altered the essential character of the landscape. Native wild animal species that rely on drinking water, or water as a habitat for part of their life cycle, are now able to persist in areas that were previously uninhabitable most of the time, resulting in larger and more widespread populations of these species than would otherwise be possible. In some cases, this increase in abundance of a species may have a significant effect on other species. For example, Yellow-throated miners may have become more abundant and their aggressive nature, if similar to that of Noisy miners, may be resulting in the local displacement of some small birds (Grey 1996). Similar competitive or aggressive displacement interactions may be occurring among other species of birds, or other taxa (e.g., dragonflies), but there have been no studies of this topic.

Apart from specific direct effects, another major effect of artificial sources of water is the maintenance of high levels of grazing pressure over very large areas. Plants and animals that are not directly affected by, or dependent on, water are probably affected by the presence of large numbers of grazing animals and the results of their grazing. Given the high numbers and density of artificial water points across the arid and semi-arid rangelands, the influence of grazing is likely to be widespread. Despite this widespread influence, there is a dearth of direct evidence of the effect of grazing or grazing animals on the abundance or diversity of biota other than pasture plants (see James et al. 1995a). This dearth of information prompted us to begin studies in 1993 on the effect of artificial waters and their associated grazing on biodiversity.

The results of this work indicate that species composition does change along gradients of increasing distance from water (and decreasing grazing intensity) and that between 15% and 38% of plant and animal species reach their greatest abundance at water-remote sites (ie., greater then 9 kms from water; James et al. 1995a; Landsberg et al. 1996). Although there has been no comprehensive assessment of water point density for Australian rangelands, it is clear that areas more than 5 kms or 8 kms from water in rangelands occupied predominantly by sheep and cattle, respectively, are rare (R. Blick, pers. comm.; Landsberg and Gillieson 1996). These findings led James et al. (1996) to suggest that water point density is one of the root-causes of changes in the distribution and abundance of native biota in the rangelands. Water point density and the grazing associated with water points need to be immediately incorporated into strategies for on- and off-reserve conservation of biodiversity in arid and semi-arid zones.

Acknowledgments

We are grateful to Marita Thompson and Inge Newman for their help in assembling the references for this review. Valuable comments on the manuscript were received from David Freudenberger.