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 2 - Continental analysis of the distribution of water points in arid and semi-arid Australia

Undertaken by Jill Landsberg, in collaboration with David Gillieson, University College, Australian Defence Force Academy, University of New South Wales, Campbell ACT 2601

Introduction

Most large mammals require regular access to drinking water, and in arid environments its availability determines where they graze. Thus sources of water become foci of grazing activity, resulting in a zone of attenuated impact (a piosphere) around each water point (Lange, 1969). The proliferation of artificial water points across much of Australia's rangelands means that piospheres associated with large grazing mammals (domestic, feral and native) have become widespread. Areas that lie beyond the piospheres are potential reference areas for regional biodiversity patterns that may have existed before the spread of this grazing impact. In order to locate such areas, we conducted a continental-scale GIS analysis using buffers constructed around the watering points which were named on maps in that acacia woodlands and chenopod shrublands, which are two of Australia's major rangeland biomes.

Methods

Water Points

Water points were overlain on an image of the Australian continent using a commercial GIS software package (IDRISI 4.1) and the Master Names File, a digital product from the Australian Surveying and Land Information Group (AUSLIG). The Master Names File comprises a listing of latitude and longitude of all named features appearing on the national series of AUSLIG 1:250,000 and 1:100,000 scale topographic maps. We subsetted from the file those feature names referring to point water sources (bores, dams, tanks, rockholes, wells, waterholes), converted the data to a raster format with a cell size of 5 km and rectified the map to a Transverse Mercator projection with a RMS error of 2.6 km. Many of the water points shown on map sheets are not named. To determine the proportion of mapped water points represented in the Master Names File we conducted counts of all named and unnamed point waters shown on a sample of 10 topographic map sheets. The sheets were chosen alphabetically from our collection, taking the first three rangeland sheets in Queensland, South Australia and New South Wales, and an additional one chosen at random.

Piospheres

The distribution of piospheres was investigated by constructing buffers with 10 km radii around the named water points. This distance was a conservative compromise. Although the main grazing impact probably occurs within 5 km of water for sheep and within 10 km for cattle, sheep may walk up to 10 km and cattle for 20 km to reach preferred plant communities (Arnold & Dudzinski, 1978). Less is known about the distances covered by kangaroos and goats, which are responsible for more than half the total grazing pressure in some southern rangeland regions (Dawson et al. 1975; Landsberg et al., 1992), but 20 km is probably a very conservative estimate of the distance these animals may range from water. The main feral grazing animals of northern Australia (horses, donkeys and camels) can also range at least this far (Wilson et al., 1992). Thus 10 km is a reasonable average piosphere radius, well within the maximum range of most of these species.

Vegetation

We constructed vegetation overlays from the digital version of the 1:5,000,000 AUSMAP "Present Vegetation", which is a compilation of Australia's vegetation in the 1980s (AUSLIG 1986). From this we extracted a subset of data representing rangelands in which chenopods or acacias are floristically prominent; this subset included the major rangeland biomes of central and southern Australia (Harrington et al. 1984).

Results

Our analysis showed that large areas of the major deserts and the arid Nullarbor Plain are remote from named water points, but much of the rest of the rangelands lie within 10 km of a named water point (Fig. A2.1). Even across the deserts, stock routes show as paths clearly marked by overlapping or aligned piospheres. Large areas of the more intensively settled parts of eastern Australia are also remote from named water points, but, unlike the deserts, these regions are well provided with other sources of water including numerous small, unnamed farm dams and linear water courses (e.g. creeks, rivers, bore drains) which are not available as geocoded digital data. Thus we excluded these more humid regions from subsequent analyses.

Nearly all the water points marked on our sample of rangeland maps were artificial. Many of the mapped waterholes may have been natural, but waterholes constituted only 3% of the total number of water points shown on our sample of maps (Table A2.1). Although there was considerable variation among map sheets, only 29% of all the water points shown on the maps were named. Thus the map in Figure A2.1 severely underestimates the actual frequency of water points and, by extension, the potential area of Australian rangeland that might be subject to grazing impacts associated with the provision of water points.

Of the half million km² of chenopod lands we analysed, only 25% is beyond the influence of named water points. The greatest proportion lies on the remote Nullarbor Plain; only about 8% of the remaining productive chenopod lands are potentially remote from water (Table A2.2).

Of the 2.6 million km² of acacia lands we analysed, around 50% lie more than 10 km from the nearest named water point, but much of this is desert; only 26% of the more productive acacia lands are remote from named water points. Furthermore, many of these areas, particularly in the eastern rangelands, lie well within the influence of unnamed, linear bore drains (Orr & Holmes, 1984), which we could not include in this analysis (Table A2.3).

Discussion

Our analysis is only an approximate guide to the distribution of potential grazing impacts in the rangelands, for two reasons. The first relates to our data source which, being limited to mapped named water points, considerably underestimates the total number of water points that actually exist. Addition of mapped but unnamed water points would increase the density of water points more than three fold; addition of unmapped water points would almost certainly extend the area of potential piosphere influence further still, as would addition of linear features such as bore drains.

The second potential source of inaccuracy relates to the radius of the buffers we erected around the named water points, and here also our analysis is likely to underestimate the true distribution of grazing pressure. In addition to those water-dependent large mammals that can range further than 10 km from water, feral rabbits also contribute significantly to the total grazing pressure on Australian rangelands. However, rabbits do not generally require drinking water (Wilson et al., 1992), so their impact is widespread throughout their range.

Prescriptions for better management of Australia's rangelands have long advocated increasing the number and distribution of water points, in order to avoid localised degradation by spreading the impact of grazing by livestock over a wider area (Harrington et al., 1984). Unfortunately, artificial water points are now so widespread in Australia's semi-arid and arid rangelands that vast areas, previously beyond the reach of large grazing animals, may now be exposed to sustained grazing pressure. It remains to be seen what impact this may have had on the native biota of the rangelands, but our results show that potential reference areas for determining pre-grazing patterns of biodiversity have become extremely rare.

Acknowledgments

This work was undertaken in collaboration with Dr David Gillieson, University College, Australian Defence Force Academy, University of New South Wales, Campbell ACT 2601, and was partly funded by a CSIRO/ UNSW Collaborative Research Grant. Paul Walker, CSIRO Division of Wildlife and Ecology and Robin Grau, Human Geography, Australian National University, assisted with data translation.

Figure A2.1: Distribution of all water points named on the 1:250,000 and 1:100,000 topographic maps covering mainland Australia.

Figure A2.1: Distribution of all water points named on the 1:250,000 and 1:100,000 topographic maps covering mainland Australia.
Table A2.1: Water points marked on a sample of 1: 250 000 map sheets covering rangeland areas
Map sheet name Tanks & dams Bores Waterholes Wells Total
Named Unnamed Named Unnamed Named Unnamed Named Unnamed Named Unnamed
Anabranch (NSW) 41 191 7 0 0 0 0 0 48 191
Balranald (NSW) 195 755 8 22 5 1 0 0 208 778
Angledool (NSW) 10 1168 33 0 1 0 0 0 44 1168
Betoota (Qld) 8 7 6 0 96 64 2 0 112 71
Augathella (Qld) 43 430 240 60 0 0 2 0 285 490
Barrolka (Qld) 2 8 37 0 97 100 9 0 145 108
Alberga (SA) 6 2 40 0 0 0 47 0 93 2
Abminga (SA) 29 2 22 0 35 3 16 0 102 5
Andamooka (SA) 90 33 16 2 0 0 34 14 140 49
Thargomindah (Qld) 5 60 3 3 39 101 3 0 50 164
Total 429 2656 412 87 273 269 113 14 1227 3026
Mean per map (±std.dev.) 43±60 266±402 41±71 9±19 27±37 27±44 11±16 1±4 123±77 303±391
Proportion of total 14% 86% 83% 17% 50% 50% 89% 11% 29% 71%
Table A2.2: Area of chenopod lands and the proportion of different vegetation categories that occur more than 10 km from a named water point
Category Description Area analysed (km²) Total area >10 km from a named water point Subset area >10 km from a named water point
Wood/chShrb Low woodlands or tall shrublands with a chenopod shrub lower stratum 37 540 13% 6%
Wood/chHerb Low open woodlands over herbfields dominated by persistent chenopods (usually Atriplex or Sclerolaena spp.) 5 690 1% 1%
chShrb Low shrublands dominated by chenopods with grass or forb ground cover 385 346 29% 8%
samphire Low shrublands dominated by chenopods (usually samphires) with no significant ground cover 14 938 32% 32%
chHerb Herbfields dominated by persistent chenopods (usually Atriplex or Sclerolaena spp.) 56 335 12% 12%
All All the above vegetation classes 499 849 26% 11%
ex samphire All the above classes except samphire 484 911 25% 8%

Areas are totals calculated within the analysis window. The subset area excludes the Nullarbor Plain.

Table A2.3: Area of acacia lands and the proportion of different vegetation categories that occur more than 10 km from a named water point
Category Description Area analysed (km²) Total area >10 km from a named water point Subset area >10 km from a named water point
ForW/aShrb Forests or woodlands with a lower stratum of acacia trees or shrubs 85 924 57% 57%
aWood/Shrb Low open forests or woodlands dominated by acacias with a shrubby lower stratum 137 690 52% 3%n
aWood/GorF Low open forests or woodlands with grass or forb ground cover 226 408 32% 32%
aShrb/Shrb Shrublands dominated by acacias with a shrubby lower stratum 384 203 20% 13%wt
aShrb/GorF Shrublands dominated by acacias with grass or forb ground cover 630 012 32% 32%
aShrb/Spin Shrublands dominated by acacias with spinifex ground cover 1 110 220 73% 73%
All All the above vegetation classes 2 574 458 50% 46%
ex aShrb/Spin All the above classes except the shrublands with spinifex ground cover 1 464 238 32% 26%

Areas are totals calculated within the analysis window. The subset area excludes the Nullarbor Plain (superscript n) and an area of wattle thicket (superscript wt), which consists of mixed acacia shrubs with a heathy shrub lower stratum, and occurs in the most southerly area of shrubland in Western Australia.