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

1. Introduction and methods

1.1 Water, grazing and biodiversity in the rangelands

Water is essential for the survival of all organisms. In semi-arid and arid regions its importance increases as its availability decreases. Much of inland Australia is semi-arid or arid and, prior to European settlement, permanent surface water was an extremely scarce resource (Appendix 1). The plant and animal communities that had evolved in this environment were, for the most part, adapted to survive on water that fell as rain. Plants harvested and harboured water stored in the soil, and relied on run-off and infiltration from infrequent falls of rain to replenish their supplies. The animals that persisted year-round obtained their water from their food, and were conservative and efficient in their water-use. Other animals were probably more transient in their use of the country, spreading out when green feed was available and moving on, or contracting back to the few permanent waterholes, when it dried out.

Most large mammals, including sheep and cattle, cannot survive without drinking water. Providing adequate supplies of water for livestock was a necessary prerequisite for pastoral expansion into the dry parts of Australia. Thus, wells, bores, dams and troughs proliferated throughout the more productive regions of the dry inland. Today, few areas of potentially productive rangeland are further than 10 km from an artificial source of water (Appendix 2) and most areas are much closer (James et al. 1996).

James et al. (Appendix 1) reviewed the published literature on what is currently known about how this proliferation of water has affected the native biodiversity of semi-arid and arid Australia. They found that some native animals have benefited from the widespread provision of water. Many bird species and some native mammals that need to drink frequently (e.g., grey kangaroos) have increased their geographic range and abundance, and others have benefited from the creation of artificial wetlands around free-flowing artesian bores. Apart from some native animals, many feral animals have also benefited, including goats, pigs, horses, donkeys and foxes. Feral camels also require drinking water, but they drink less frequently than the other large herbivores and can range more widely. Rabbits are the only widespread and abundant feral mammal that can persist without drinking water, although to do so they must have access to green or succulent feed. Feral and native predators, such as foxes and dingoes, focus on water points for hunting and drinking. Scavengers (including birds of prey) congregate to feed on dead livestock and other grazing animals that die around water points, particularly during drought.

Palatable plants and ground cover have clearly been affected by the sustained and widespread increase in grazing pressure that has developed around artificial sources of water. Numerous studies have documented marked changes in plant cover around watering points, with decreases in the proportion of palatable perennial grasses and forbs, and increases in the proportion of bare ground or unpalatable plants, particularly shrubs. Increased rates of soil erosion have also been shown, particularly within 2-3 km of water points where livestock traffic is high and changes in vegetation and ground cover are most marked.

Apart from declines in abundance of palatable plants, changes in ground cover, and increases in the abundance of some water-reliant birds and mammals, surprisingly little is known about how other native biota may have been affected by the provision of artificial water sources across most of the Australian rangelands. Even for plant communities, most of our current knowledge is restricted to changes within a few kilometres of water points. Although there have been speculations about negative impacts on biodiversity arising from changes in habitat and interactions among introduced and native species, there are few empirical data available against which to assess these speculations.

The purpose of the study presented here was to collect and analyse sufficient empirical data to determine the effect of the provision of artificial sources of water on native species (perennial and ephemeral plants, vertebrate and invertebrate animals), in two of Australia's most productive and extensive rangeland biomes. Two interim reports (James et al. 1995a; Landsberg et al. 1996) presented the results of the field component of surveys undertaken in 1994 and 1995, respectively. This final report integrates the results presented in each of the two interim reports, and provides interpretations and recommendations arising from the results.

1.2 Study design

The study was based around field surveys designed to sample biota along gradients in grazing intensity extending out from artificial water points. The rationale for adopting this approach is described in a background paper (James et al. 1995b). Each gradient extended from a water point to a reference area that was remote from all waters. The distance to the reference area was chosen to lie outside the normal grazing range of the livestock using the water. Our target distances for reference sites were 15 km for cattle and 9 km for sheep; these distances were also considered to be close to the normal limit of the water-centred grazing activities of kangaroos and most feral animals (Landsberg and Gillieson 1996; Appendix 1; Appendix 2).

For each gradient, we chose five sites in addition to the reference site. Each site was located progressively closer to water, at progressively closer spacings (Table 1.2.1). We arrayed the sites in this way because our review of the literature had led us to expect most changes to occur within 2-3 kms of the water point (Appendix 1). All six sites were chosen to lie in similar parts of the landscape, within the same landsystem, the same paddock, the same directional quadrant out from the water and as close as possible to the pre-determined distances from the water point. The positions of sites that met the selection criteria were determined from maps and air photos, and were then located using GPS navigation receivers, to avoid any other bias in their selection.

Table 1.2.1 Target distances of sites from water

Livestock   Site 1    Site 2    Site 3    Site 4    Site 5    Site 6           

sheep       0.5 km    1.0 km    2.0 km    3.5 km    6.0 km    9.0 km           
cattle      0.5 km    1.5 km    3.5 km    6.0 km    9.0 km    15.0 km          

1.3 The gradients

1.3.1 Locations

We surveyed eight gradients that were widely separated across the chenopod and acacia shrublands of central and southern Australia (Fig. The dispersion of these sites across the continent was deliberate. We wanted to ensure that we covered a range of the climatic, geomorphic and vegetational variation that occurs in each of these major vegetation types. The gradients were not selected to be statistical replicates within each vegetation type; rather, each gradient was intended to provide an independent test of the effect of artificial water. This approach had potential costs and benefits: the major potential cost (apart from logistics) was that the gradients might be so different from each other that no general result for the effect of artificial water could be drawn out. The major potential benefit was that if a general and consistent result were found, it was more likely to be applicable to the entire area under study by logical interpolation of the result.

Figure Location of the grazing gradients

Figure Location of the grazing gradients.

Potential regions where gradients might be found were identified from a GIS analysis showing the distribution of named water points across Australia (Appendix 2). Likely regions were then examined in more detail using topographic and thematic maps and extensive consultation with local experts. Potential gradients in the most likely regions were then visited, to determine if there were any confounding influences not apparent from maps or local advice. Many potential gradients were unsuitable because it was not possible to find a full array of appropriately-spaced sites that were also comparable in landscape position and the landscape elements present. Other potential gradients were found to be strongly influenced by natural waterholes that sometimes held water for many months after rains. At several other potential gradients, we found that new water points had been recently installed to service the areas we had identified as remote from water. This had occurred independently of our interest as part of normal property development. Only approximately 1 in 20 of the potential gradients we visited was suitable for the study.

Three of the selected gradients were dominated by mulga (Acacia aneura); one in the Northern Territory, one in Queensland and one in New South Wales. Two gradients (one in South Australia and one in Western Australia) were predominantly chenopod shrub-steppe dominated by species of Maireana, Atriplex and Sclerolaena. The remaining three gradients were of varying character: in Queensland, a gidgee (Acacia cambagei) woodland over a low chenopod understorey; in South Australia, a predominantly chenopod shrubland with patches of western myall (A. papyrocarpa), and in Western Australia, a predominantly chenopod shrubland with bands of mixed acacias (A. tetragonophylla, A. sclerosperma, A. ramulosa).

All but one of the gradients (NT mulga) were in paddocks stocked primarily with sheep. Thus, most of the reference sites were 10 km or less from water. There was some variation among gradients in the actual distances of sites from water, because of the need to locate sites in those parts of the landscape that contained the elements being sampled for that gradient (Table

Table Actual distances (km) of sites from artificial sources of water

Gradient          Site 1  Site 2  Site 3   Site 4  Site 5   Site 6

NT mulga              0.5     2.0     3.5    5.9     9.0     15.1   
NSW mulga             0.7     1.0     2.0    3.5     6.1     8.9    
Qld mulga             0.5     1.1     2.0    3.7     6.0   7.0-8.9# 
Qld gidgee/           0.7     1.2     1.8    3.7     5.5     8.9    
WA chenopod/          0.6     0.9     2.1    3.6     6.0     8.3    
SA chenopod/ myall    0.5     1.0     2.0    3.6     6.0     10.2   
SA chenopod           0.5     1.1     2.2    3.6     6.4     8.2    
WA chenopod           0.5     1.1     2.0    3.8     6.4     8.9    

#Site 6 at the Qld mulga gradient was within the sphere of influence of two water points. The water point to the south-east, which was the one influencing sites 1-5 along this gradient, was 8.9 km distant. There was, however, another water point 7.0 km to the east of site 6. Few animals travelled from that water point to site 6, because it entailed crossing a stony plateau; thus site 6 was actually 8.9 km from the nearest readily accessible water point.

1.3.2 Descriptions The NT mulga gradient

The gradient occurs on a large cattle-grazing property in the southern Northern Territory, near the Kulgera Roadhouse. The mean annual rainfall for the neighbouring station is 196 mm. The mean monthly maximum temperature recorded at the nearby town of Finke ranges from 37.8° C in January to 20.2° C in July.

The paddock containing the gradient is 41,100 ha in size with two active water points: one in the south, and one in the north. In the western section of the paddock, two dams have been used for temporary water storage. Recently (October 1994), water from the southern bore has been piped to one of the dams in the western half creating a third permanent water supply for the paddock. The grazing gradient that was surveyed originates from the southern bore and continues 15 km to a reference area near the south-western corner of the paddock. The new water point was not in use at the time of the survey, but as its use intensifies the reference value of the sites further from water will inevitably be down-graded.

The gradient crosses four recognised land systems: predominantly Karee and Simpson with lesser systems of Lindavale and Outounya (Perry et al. 1962; Shaw and Bastin 1988). The sites occur on the plains and broad drainage floors of the Karee land system with a mixture of red sandy loam and red earth soils. The vegetation of the gradient is dominated by groves of mulga (Acacia aneura) trees with a silver cassia (Senna artemisioides) understorey. The dominant ground cover species include bandicoot grass (Monochather paradoxa) and woollybutt grass (Eragrostis eriopoda).

This gradient was surveyed twice: a full survey in October 1994 and a re-survey of reptiles and invertebrates only, in November 1994. This was done because the dry conditions and cool weather experienced in the October survey resulted in so few captures that statistical analysis of these taxonomic groups would have been unreliable. During the October survey, overnight temperatures were 8-14° and daily maximums were 28-34°. During the November survey temperatures were high initially (daily max. > 40°), but a cool change on the third day brought light rain, and overcast weather for the remainder of the survey. The NSW mulga gradient

This gradient lies in a 5,000 ha paddock in the north-east corner of a pastoral station in north-west New South Wales, abutting the Queensland border. Median rainfall for the district is around 290 mm per year, and mean monthly maximum temperatures range from around 35°C in January to 19°C in July.

The paddock normally runs about 600 sheep and 80 cattle; thus the stocking rate is about 1 sheep-equivalent per 4 hectares. There are no feral goats but there is an unknown density of kangaroos contributing to the total grazing pressure. The paddock is permanently watered by two sub-artesian wells in the south-east and south-west corners respectively; occasionally animals are also allowed access to an earth tank along its western fence. The reference area is located in the north-east corner of the paddock, 9.5 km from the south-east well which is the nearest source of permanent water to which grazing animals have access. There is an additional earth tank in the centre of the paddock (about 2 km west of site 4) but it rarely holds water for more than a few weeks. There is also a water point on the station in Queensland, immediately to the north, about 2 km from our reference area. This water is not accessible to grazing animals, because the properties are separated by the State Border fence which is a well-maintained barrier against passage by dingoes and large grazing animals. However, the water point in Queensland is accessible to birds using the reference area.

The gradient occurs in a sandplain land system, consisting of extensive slightly undulating sandplain with a broad dendritic drainage pattern and relief to 5 m. Soils are deep sandy to sandy loam red earths. Vegetation consists of distinct groves of dense mulga (Acacia aneura) with some poplar box (Eucalyptus populnea) becoming continuous along the drainage tracts. This is interspersed with open areas of perennial grasses dominated by woollybutt (Eragrostis eriopoda), and areas of woody shrubs, particularly hopbush (Dodonaea viscosa) and punty bush (Senna artemisioides); isolated ephemeral swamps ringed by poplar box are also widespread (Soil Conservation Service of NSW 1979).

The weather was warm and sunny during the period of the survey, with little cloud and winds that were usually light to moderate. A gusty westerly wind caused local areas of raised dust on day 5. Temperatures ranged from 10-15° C overnight to 25-35° C during the day. The Qld mulga gradient

The gradient occurs on a large pastoral property north-west of the township of Quilpie, in south-west Queensland. The regional climate is arid sub-tropical, with a median annual rainfall of 285 mm, and mean monthly maximum and minimum temperatures of 37°C and 23°C in summer and 20°C and 6°C in winter (extrapolated from data for Quilpie and Windorah; Bureau of Meteorology 1988). Landforms, soils and major vegetation associations along the gradient are typical of mulga-dominated areas in this region (Neldner 1992). Predominant landforms are residual tablelands and dissected plateaus surrounded by level to gently undulating plains; geology is mainly unconsolidated weathered sediments and weathered tertiary land surfaces; and soils are predominantly loamy, sandy or gravelly red earths. Vegetation is dominated by mulga low woodland, frequently occurring in groves.

The gradient lies within a 8,633 ha paddock that was developed late last century and has only ever been stocked with sheep. The current manager usually stocks it with around 1,000 sheep. We saw moderate numbers of kangaroos, particularly Euros (Macropus robustus erubescens), but very few rabbits and no feral goats. We saw sign of a small number of feral horses, particularly at site 5. There are four water points in the paddock, with three concentrated in the eastern half. The study gradient was located in the western part of the paddock, running in a north-westerly direction away from the fourth water point, a bore located near the middle of the southern boundary fence.

The study sites were all located on broad alluvial plains above drainage lines that were sometimes channelled and sometimes very diffuse. Mulga groves were the dominant structural component of the vegetation at the sites, with more open intergroves having cover that varied from bare earth to mixed grasses and forbs, to shrubs, particularly several subspecies of the cassia Senna artemisioides. Upslope, the plains quickly became steeper with very stony surfaces, and mulga was generally replaced by bastard mulga (Acacia stowardii). The landscape strata sampled were the mulga groves and adjacent open inter-groves.

The weather at the time of the survey was warm and sunny, with light cloud and light to moderate winds. The average diurnal temperature range was 19°C to 35° C. The Qld gidgee/chenopod gradient

This gradient occurs on a large pastoral property south-west of the township of Quilpie, about 130 km south of the mulga gradient described above. Its climate is similar to that gradient's, although rainfall may be a little higher (median annual 308 mm at Quilpie; Bureau of Meteorology 1988). However its landscapes are markedly different. They are dominated by dissected residuals with mantled pediments sloping down to undulating plains with brown and grey cracking clay soils overlain by cobbles. The vegetation of the mantled pediments is predominantly low open-woodlands of gidgee (Acacia cambagei) (Neldner 1992). In the study area the gidgee occurs in discontinuous groves lining dry stream channels that drain away from the dissected residuals and across the mantled, undulating pediments that surround them.

The gradient lies within a very large paddock of 45,369 ha that was fenced late last century. Although cattle may have been stocked in the paddock in the past, it has been stocked with sheep for at least the last 15 years. The current manager usually stocks it with around 3,000 sheep. The only other grazing mammals we saw were occasional kangaroos (both Euros, Macropus robustus erubescens, and Red Kangaroos, M. rufus); we saw no rabbits, goats or other feral animals.

There are five artificial sources of water in the paddock, and one moderately permanent natural waterhole. The study gradient was located in the eastern part of the paddock, running just inside (0.5-1 km) and parallel to the north-south fence. The reference site was 9 km south of the permanent natural waterhole in the north-east corner of the paddock and 8.5 km north of a bore and associated earth tank in the south-east corner. The other sites were located in increasing proximity to the southern bore.

The study sites were all located on alluvial plains or terraces associated with the drainage lines flowing away from a series of dissected mesas and plateaux. The drainage varied from weakly defined and meandering to well-defined and developed channels. The gidgee occurred in patches along the drainage lines and in small groves, often associated with gilgais. Much of the surrounding country was dominated by low succulent shrub-steppe and open chenopod herbland with many species of Maireana, Atriplex, Sclerolaena and Sclerostegia, and a mixture of other herbs in which daisies (Asteraceae family) were particularly common. Upslope, the plains quickly became steep and stony, supporting a wide variety of shrubs in which various species of Acacia and Eremophila were prominent. The survey was stratified around landscape strata consisting of gidgee groves or patches and the adjacent open chenopod plains.

The weather at the time of the survey was generally mild and sunny, with light to moderate winds and a temperature range from around 15°C to 35°C. There was one night and morning of cool cloudy weather with light showers in the middle of the survey period. The WA chenopod/acacias gradient

This gradient is located about 60 km north of the town of Carnarvon in north-western Western Australia. The regional climate is arid sub-tropical, with a predominantly winter rainfall pattern. The median annual rainfall for Carnarvon is 203 mm. Mean monthly maximum temperatures at Carnarvon range from 22.0°C in July to 32.7°C in February, with minima of 11.0°C and 23.3°C in the same months. However, because the study gradient is further inland than Carnarvon it probably experiences greater temperature extremes.

The gradient is located in a 6,000 ha paddock that is bordered on the western side by a large, normally dry salt lake. It is usually stocked discontinuously with around 400-600 sheep. Feral goats also contribute to grazing pressure, although only a few were observed during the survey period. The single water point, a well, lies in the south-eastern corner of the paddock. The gradient extends from this water point north-westwards 8.3 km to the reference site. A large salt lake in the adjacent paddock, is about 5 km west of the reference. Many foxes were observed on this gradient during the survey period.

The landscape strata sampled for the survey comprised the two major landscape elements in the paddock. The first consists of chenopod plain with duplex soils supporting mainly saltbushes (Atriplex bunburyana, A. vesicaria) and Gascoyne bluebush (Maireana polypterygia). The second consists of sandy dunes with dead finish (Acacia tetragonophylla), A. sclerosperma, horse mulga (A. ramulosa) and scattered Eremophila spp. The ground cover is dominated by several species of daisies, introduced buffel grass (Cenchrus ciliaris) and Tetragonia eremaea.

The weather at the time of the survey was mild, overcast with a few showers for the first three days, then warm and windy for the remainder of the survey. Daily temperatures ranged from 14° C to 36° C. The SA chenopod/myall gradient

The gradient is situated in a large paddock on the eastern edge of Lake Gairdner, 170 km north-west of Port Augusta, South Australia. The climate is characterised by hot, dry summers and cool, wet winters. The average annual rainfall for the station is 185 mm. The mean monthly maximum temperature for the nearest meteorological recording station at Woomera is 34.2°C in January and 16.6°C in July.

The paddock is 12,030 ha in size and serviced by a near-permanent water supply in the south-eastern corner, with temporary water supplies held in a number of small claypans for a few weeks after rain. Lake Gairdner itself is too saline to be a source of drinking water. The gradient runs from the water to the reference site, 10.2 km to the north-west. The paddock is usually stocked with around 1,300 sheep, although the station has some cattle. Goats are a significant component of grazing pressure in the region and were abundant around the water point. Western Grey Kangaroos (Macropus fuliginosus) and Red Kangaroos (M. rufus) were also frequently seen.

The gradient is situated within the northern edge of the Mahanewo environmental association (Laut et al. 1977). The landscape consists of undulating sandy plains interspersed with low hills. Soils are a mixture of red calcareous earths and loams. The vegetation is dominated by patches of low chenopod shrublands (Maireana spp., Atriplex spp.), and low woodlands of western myall (Acacia papyrocarpa) with a cassia (Senna spp.) understorey. The major herbaceous species include rough speargrass (Stipa scabra), oat grasses (Enneapogon spp.) and white paper daisy (Rhodanthe floribunda). The vegetation at sites 5 and 6 was more open and grassy than vegetation at sites 1, 2 and 3. This structural variation was probably induced by fires, which tend to burn more intensely further from water, where grass biomass, and therefore fuel loads, tend to be higher.

Weather during the survey was variable. The first few days were warm, followed by two days of cool weather (resulting in reduced reptile activity), and the final day was 40° with a sand storm that filled many of the pit-traps at site 1 with drifting sand. The SA chenopod gradient

The gradient is situated in a paddock which abuts the western edge of Lake Torrens in South Australia. The climate in this area is characterised by hot, dry summers and cool, wet winters. The average annual rainfall for the station is 182 mm. The mean monthly maximum temperature for the nearest meteorological recording station at Woomera is 34.2°C in January and 16.6°C in July.

The paddock is 9,290 ha in size and serviced by artificial and ephemeral natural water points. An artificial water point (a tank) on a hill in the south-western corner is the only permanent water, and is supplied with water pumped from a dam in a creek 15 km away. There is also a dam on a small creek in the central western area of the paddock which is filled by winter rains but dries during summer. Lake Torrens is too saline to be a source of drinking water. The gradient runs from the permanent tank, 6 km eastward towards Lake Torrens, then 4 km north to the reference area which is 8.2 km from permanent water. The paddock is usually stocked with 700-900 sheep. Red Kangaroos (Macropus rufus) were also numerous in the paddock, particularly near Lake Torrens.

The paddock lies within the Andamooka environmental association (Laut et al. 1977). The landscape consists of undulating gibber-covered slopes and hills interspersed with gilgai and drainage depressions. Most of the gibber landscapes consist of crusty red duplex soils with alkaline powdery clay loams in the gilgai depressions. Bladder saltbush (Atriplex vesicaria) dominates the landscape with many other chenopod species (Sclerostegia tenuis, Sclerolaena spp. and Dissocarpus spp.) and Euphorbia tannensis contributing to the low shrub cover. The herbage layer contains a variety of daisies and burrs (Asteraceae family), woollybutt grass (Eragrostis eriopoda) and the introduced Arabian grass (Schismus barbatus).

Weather during the survey was consistently fine and warm to hot. Day-time temperatures were around 30-35°C, and night-time temperatures around 12-20°C. The WA chenopod gradient

This gradient is located on the Nullarbor Plain of Western Australia, approximately 50 km north of the Cocklebiddy roadhouse. The regional climate is arid temperate, with a predominantly winter rainfall pattern. The median annual rainfall for the nearby township of Madura is 280 mm. Mean monthly temperatures range from maxima of around 18°C in July and 32°C in January, and minima of around 5°C and 15°C in the same months. The extensive, mostly flat Nullarbor Plain is underlain by flat-bedded limestones on which have developed karst landscapes and shallow calcareous soils that support a chenopod low shrub-steppe.

The lease for the property that includes the gradient was taken up recently, in the early 1960s. The gradient lies in a 14,250 ha paddock which has been stocked with sheep since 1964. Current numbers are around 2,000. Rabbits were common in the paddock, but all sites were located to avoid areas with rabbit warrens or other obvious sign of rabbits. Red Kangaroos (M. rufus) were also seen, but at low densities. The paddock has a single water point, a tank filled from a bore, located on the central southern boundary to the paddock. The gradient extends north-eastwards from this water point 8.9 km to the reference site.

The paddock was burnt by wildfire in 1974, following good summer rains that had led to a build up of grass biomass. Most chenopod bushes appeared to have survived the fire, but fire scars and bush architecture indicate that the fire may have been hotter at sites 4, 5 and 6 that at sites 1, 2 and 3.

The paddock's gently undulating landscape is devoid of trees and is dominated by pearl bluebush (Maireana sedifolia) on the slightly higher ground and bladder saltbush (Atriplex vesicaria) in the abundant shallow depressions. The shallow depressions and the surrounding plain were the landscape strata recognised for the survey. Common wallaby-grass (Danthonia caespitosa), spear grasses (Stipa spp.) and annual forbs dominate the herbaceous layer across most of the landscape.

Day-time weather during the survey was mild and generally sunny but with some overcast periods in the afternoon when sea breezes penetrated inland. Because of the sea breezes, afternoons were windy and cool. Day-time temperatures were around 25-30°C, and night-time temperatures around 7-15°C.

1.3.3 Illustrations

Figure The NT mulga gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The NT mulga gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The NSW mulga gradient showing sites 1 (upper), 3 (middle) and 6 (lower)

Figure The NSW mulga gradient showing sites 1 (upper), 3 (middle) and 6 (lower)

Figure The Qld mulga gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The Qld mulga gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The Qld gidgee/chenopod gradient showing sites 1 (upper), 4 and 6 (lower)

Figure The Qld gidgee/chenopod gradient showing sites 1 (upper), 4 and 6 (lower)

Figure The WA chenopod/acacias gradient showing sites 1 (upper), 4 and 6 (lower)

Figure The WA chenopod/acacias gradient showing sites 1 (upper), 4 and 6 (lower):

Figure The SA chenopod/myall gradient showing sites 1 (upper), 3 and 6 (lower)

Figure The SA chenopod/myall gradient showing sites 1 (upper), 3 and 6 (lower)

Figure The SA chenopod gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The SA chenopod gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The WA chenopod gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

Figure The WA chenopod gradient showing sites 1 (upper), 4 (middle) and 6 (lower)

1.3.4 Potentially confounding influences

Sites along each gradient were selected to be as similar as possible to the reference site in all respects except their distance from water. There were so few reference sites to choose from, however, that the matching of sites was less than ideal at some gradients. Thus there were a number of variables in addition to distance from water that may have contributed to differences in biodiversity between sites along each gradient (Table Fortunately, these potentially confounding influences were all relatively localised and did not vary consistently between gradients.

Variation in landscape position could have confounded gradient trends if sites had been consistently arrayed along a catenary sequence (for example, if the trend from site 6 to site 1 was consistently downhill or uphill) but any such trend was deliberately avoided during site selection. At the two gradients where there was evidence of fire the possibility of a confounding gradient in fire intensity could not be avoided, because fuel load, grass biomass, grazing intensity and distance from water are inevitably linked. Nevertheless there was no evidence of any fire damage at six of the eight gradients. At three of the eight gradients the reference sites were moderately close to large salt lakes. The lakes were not potable, however, and were probably too far from any of the sites to influence most biota. (The existence of the reference sites may partly be due to their proximity to the salt lakes, which may have made them difficult for access.) The dingo fence reference affected only one of the eight gradients.

Table The main potentially confounding influences at each of the gradients

Influence           Gradients most    Sites most affected                  

Position in         NT mulga          Site 3 higher in landscape           

                    Qld mulga         Site 6 higher in landscape           

                    Qld mulga         Slight catenary slope from site 2    
                                      to site 1                            

                    Qld gidg/ chen    Sites 4 and 5 higher in landscape    

                    SA chen/ myall    Slight catenary slope from site 2    
                                      to site 1                            

Fire scar           SA chen/ myall    Increase with distance from water    

                    WA chenopod       As above                             

Large salt lake     WA chen/ acacias  Closest to site 6 (5 km)             

                    SA chenopod       Closest to site 6 (2 km)             

                    SA chen/ myall    Closest to site 6 (7 km)             

Temporary water     Qld mulga         Most sites affected by widespread    
pondage                               clay pans and/ or stream channels    
                                      that may hold water for several      
                                      months after substantial rains.      

                    Qld gidg/ chen    As above                             

                    WA chen/ acac     As above                             

                    SA chen/ myall    As above                             

"Dingo fence"       NSW mulga         Water inaccessible to large grazing  
reference                             animals from site 6, because of      
                                      dingo fence barrier, but birds       
                                      cross the fence for water            

The most widespread potentially confounding influence was proximity to temporary sources of water, which affected half the gradients. After substantial rainfalls these temporary waters may allow grazing animals to spread out and make transient use of pastures in their vicinity. Grazing around temporary waters must inevitably be infrequent and relatively light, however, since rains also increase the availability of green feed and thus lessen the reliance of animals on drinking water. When feed and water dry out the animals must contract back to more permanent sources of water. This was probably the natural pattern of usage of the country by kangaroos before artificial sources of water were provided. To quote Francis Ratcliffe:

"It is not as though the land were a desert: it must have been far otherwise when the white man first saw the country in its unravished state. It was merely waterless. ...probably the greater part of the animal life moved into it only when the flush of feed was on. " (Ratcliffe 1951, p201).

In any case, where temporary waters occurred they were widespread and likely to influence most parts of the affected gradients to a similar extent. Distance from permanent water was the only variable for which the pattern of change between sites along a gradient was consistent for all eight gradients. None of the other influences showed a consistent pattern of variation at more than three gradients and most were restricted to a few sites at one or two gradients.

1.3.5 Timing and seasonal conditions prior to surveys

The gradients were all surveyed in late spring, when animals were expected to be most active and many plants were in flower. This limited the number of gradients that could be sampled in any one season; four were surveyed in 1994 and four in 1995 (Table

Seasonal conditions prior to the surveys were estimated by comparing the monthly rainfall figures from a property near each gradient with long-term rainfall statistics for the rainfall district in which the gradient lies (Bureau of Meteorology 1986). Following Bureau of Meteorology terminology, seasons were declared "average" if monthly rainfall values over the 6 months prior to a survey generally fell within in the decile range 4-7 for the district, "below average" if they were generally in decile ranges below 4, and "above average" if they were generally in decile ranges above 7.

For all four gradients surveyed in 1994, the rainfall during the preceding 6 months had been generally below average, in contrast to the four gradients surveyed in 1995, which had generally received average levels of rainfall during the 6 months preceding their surveys (Table

Table Timing and seasonal conditions for the surveys of each gradient.

Gradient                  Date of survey               Seasonal conditions       

NT mulga                  5-14 Oct & 11-18 Nov 1994    below average             
NSW mulga                 2-9 Nov 1994                 below average             
Qld mulga                 13-20 Oct 1995               average                   
Qld gidgee/ chenopod      22-29 Oct 1995               average                   
WA chenopod/ acacias      5-12 Nov 1995                average                   
SA chenopod/ myall        25 Oct -1 Nov 1994           below average             
SA chenopod               20-27 Nov 1994               below average             
WA chenopod               17-25 Nov 1995               average