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
3. Analysis of biodiversity change along gradients
3.1 Change in plant cover along gradients
Although plant cover has been shown in many studies to be sensitive to the increased grazing pressures near water points (Appendix 1), simple trends are often obscured by interactions among palatable and unpalatable species of herbs and shrubs. For example, "inverse gradients" may be formed where ground cover actually increases close to water because of increasing dominance by unpalatable plants; effects such as these are difficult to detect from ground-based measurements (Pickup et al. 1994). This may be partly why we found only weak trends in variation in ground cover with distance from water. In addition, none of our gradients showed signs of severe degradation. Even at the sites very close to water ground cover was still evident and by 2-3.5 km from water (sites 3 or 4) ground cover had generally recovered to levels very similar to those at the reference sites (Figures 1.3.3.1-8 and 3.1.1-8).
The cover trends that were significant, however, were generally in agreement with previous studies: as distance from water increased there was an overall tendency for the proportion of grass cover to increase and the proportion of bare ground to decrease (Table 3.1.1; Figures 3.1.1-3.1.8). These trends in cover were not significant for each individual gradient, but were nevertheless significant for the analyses of data combined across all gradients. Litter cover also increased with distance from water at two gradients (NSW and Qld mulga gradients) and forb cover showed a similar trend at three gradients (NSW mulga, WA chenopod/acacias, SA chenopod/myall).
There was little evidence of any marked seasonal differences in trends in ground cover. Plant cover tended to increase with distance from water regardless of season. Grass or forb cover showed a significant positive trend at all but two gradients; one of these – SA chenopod – was surveyed after a dry season and the other – Qld mulga – after an average season (see Section 1.2.4 for seasonal details and Table 3.1.1 for cover analyses). Similarly, the proportion of ground that was bare tended to decrease with distance from water, regardless of seasonal conditions prior to the surveys. Three of the gradients where this trend was significant were surveyed after drier than average seasons (NT mulga, NSW mulga and SA chenopod/myall) and two were surveyed after average seasons (Qld mulga, Qld gidgee/chenopod).
Trends in the cover of trees and shrubs were also weak (upper cover in Table 3.1.1). In addition, the direction of trend varied among even those gradients where a trend was significant. In some cases this may reflect circumstances peculiar to individual gradients. For example, the trends at the SA chenopod/myall gradient and the WA chenopod gradient were for cover of shrubs to decrease with increasing distance from water. This could relate to the potentially confounding influence of wildfires, the intensity of which is probably inversely proportional to distance from water (Section 1.3.4). At the WA chenopod gradient fire-killed chenopod shrubs appeared to become increasingly common as the distance from water increased (Section 1.3.2.8). At two of the gradients in acacia woodland (NSW mulga and Qld gidgee/chenopod) upper cover tended to increase with distance from water.
At the Qld gidgee/chenopod gradient, this may reflect a decline in landscape function closer to water. Landscape structure, diversity and function were assessed in a concurrent study at the two Queensland gradients surveyed for biodiversity (Ludwig et al. 1996). Results from this additional work are preliminary at this stage, but it appears that for both gradients, landscapes distant from water had attributes that indicate an ability to capture, store and utilise water and nutrient resources. Nearer the watering points, however, landscape patches were depleted; that is, groves of trees had "die-back", soil surfaces were rilled, and the herbaceous band that usually occurred above tree groves was reduced (Ludwig et al. 1996). This could indicate early stages of land degradation along these gradients. Landscape analyses at the other six gradients would show whether these trends are general.
Figure 3.1.4: Variation in the major categories of ground cover at the QLD gidgee/chenopod gradient.
Figure 3.1.5. Variation in the major categories of ground cover at the WA chenopod/acacias gradient.
3.2 Change in species richness along gradients
Changes in the number of species of plants and animals at each site on a gradient are shown in Tables 3.2.1-3.2.10. The distance of each site from water varied for each gradient depending on the type of grazing animal and local characteristics of a paddock (Table 1.3.1.1), so each site is described by an ordinal number of increasing distance from water. The total species richness for a gradient is shown in the column 'All sites'.
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 6 8 12 8 9 9 19
NSW mulga 10 12 12 11 11 12 19
Qld mulga 14 12 11 11 12 17 24
Qld gidgee/ 8 5 15 20 15 14 28
chenopod
WA chenopod/ 20 25 25 22 27 24 50
acacias
SA chenopod/ myall 12 15 13 11 8 11 25
SA chenopod 2 3 2 3 5 1 6
WA chenopod 8 6 7 4 9 8 13
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 8 12 12 12 15 17 28
NSW mulga 11 15 11 10 15 13 28
Qld mulga 27 22 27 16 31 23 50
Qld gidgee/ 25 21 18 19 29 32 48
chenopod
WA chenopod/ 11 10 13 14 9 14 22
acacias
SA chenopod/ myall 13 17 27 24 28 23 47
SA chenopod 10 10 8 4 6 7 16
WA chenopod 7 9 10 8 10 8 17
Note that these numbers do not include large mammal species that were seen but not sampled in pit-traps; Mus musculus was the only exotic mammal species sampled.
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 1 2 2 2 2 2 2
NSW mulga 0 0 0 0 0 0 0
Qld mulga 1 0 2 (1) 1 1 1 2 (1)
Qld gidgee/ 2 (1) 1 1 (1) 1 1 1 2 (1)
chenopod
WA chenopod/ 2 2 2 2 1 0 3
acacias
SA chenopod/ myall 1 (1) 0 1 (1) 1 1 1 2 (1)
SA chenopod 1 2 2 2 4 3 6
WA chenopod 1 (1) 0 0 1 (1) 1 (1) 1 (1) 1 (1)
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 4 7 10 6 7 8 16
NSW mulga 7 5 6 6 8 6 14
Qld mulga 5 10 5 6 7 7 15
Qld gidgee/ 3 3 11 8 8 3 15
chenopod
WA chenopod/ 6 11 9 9 14 7 21
acacias
SA chenopod/ myall 8 6 8 10 9 12 23
SA chenopod 9 6 8 4 5 6 13
WA chenopod 9 7 7 7 9 5 14
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 32 34 40 51 39 33 81
NSW mulga 36 46 62 52 52 63 101
Qld mulga 52 45 39 46 57 50 94
Qld gidgee/ 21 30 25 32 42 33 69
chenopod
WA chenopod/ 56 57 40 41 47 41 96
acacias
SA chenopod/ myall 23 45 54 51 34 34 88
SA chenopod 18 22 12 16 11 14 35
WA chenopod 23 25 29 31 27 26 50
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 2 4 1 3 2 0 10
NSW mulga 5 9 6 10 9 9 27
Qld mulga 13 12 9 11 13 9 41
Qld gidgee/ 5 5 5 10 7 5 25
chenopod
WA chenopod/ 3 2 5 4 2 1 13
acacias
SA chenopod/ myall 6 8 3 5 6 4 24
SA chenopod 3 3 0 2 0 3 7
WA chenopod 4 1 3 8 7 4 18
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 3 4 2 5 6 5 7
NSW mulga 5 5 5 5 5 7 8
SA chenopod/ 2 4 5 5 8 5 10
myall
SA chenopod 3 3 1 3 3 2 4
Gradient Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 All
sites
NT mulga 8 7 4 5 4 5 15
NSW mulga 4 5 4 3 4 3 8
SA chenopod/ 6 13 11 8 9 3 27
myall
SA chenopod 4 4 6 4 5 3 8
Examination of Tables 3.2.1-3.2.10 indicates that there is no strong or consistent trend for species richness to either increase or decrease with increasing distance from an artificial water point. A statistical examination of trends in species richness with distance from water was conducted with least-squares regression of linear and second-order polynomial functions. Of the 69 regressions tested for the plant groups and animal taxa surveyed at each gradient, only five were significant at P < 0.05 and these were all in different groups or taxa or at different gradients (Table 3.2.1.11). A further 40 regressions were tested for groups and taxa combined into higher order amalgamations, but only five yielded P-values < 0.05 (Table 3.2.11). The few significant relationships detected between species richness and distance were generally better described by second-order polynomials than linear relationships. The parameters for the 10 significant regressions are shown in Table 3.2.12.
Linear (italicised type) and second-order polynomial functions (non-italicised type) were tested and the values shown are for the function returning the highest r² (ie, explaining the most variance). P-values <0.05 are shown in bold face.
Taxon NT mu NSW mu Qld mu Qld WA SA SA ch WA ch
gi/ch ch/ac ch/my
Understorey 0.50 0.23 0.84 0.88 0.21 0.62 0.61 0.35
plants
Overstorey 0.64 0.51 0.40 0.92 0.26 0.63 0.35 0.46
plants
Seedbank 0.11 0.14 - - 0.35 0.30 - 0.51
plants
Birds 0.89 0.09 0.13 0.66 0.08 0.76 0.84 0.29
Reptiles 0.17 0.20 0.04 0.54 0.36 0.75 0.60 0.42
Ants 0.65 0.42 0.37 0.77 0.58 0.24 0.52 0.70
Springtails 0.54 0.94 - - - 0.85 0.01 -
Beetles 0.49 0.41 0.11 0.71 0.63 0.15 0.50 0.60
G'hoppers & 0.77 0.43 - - - 0.56 0.42 -
crickets
All plants 0.01 0.03 0.96 0.78 0.03 0.67 0.54 0.53
All 0.66 0.35 0.15 0.59 0.78 0.80 0.67 0.49
vertebrates
All 0.56 0.55 0.19 0.84 0.64 0.37 0.56 0.82
invertebrates
All animals 0.78 0.63 0.22 0.85 0.58 0.47 0.89 0.90
All biota 0.71 0.50 0.60 0.97 0.18 0.47 0.95 0.72
Taxon Gradient Best-fit r² F-value P-value
function
Understorey Qld gi/ ch quadratic 0.88 10.80 0.040
plants
Overstorey Qld gi/ ch quadratic 0.92 18.20 0.020
plants
All plants Qld mulga quadratic 0.96 32.80 0.009
Birds NT mulga quadratic 0.89 11.50 0.040
Reptiles SA ch/my linear 0.75 12.00 0.026
Collembola NSW mulga quadratic 0.94 22.80 0.015
All animals WA chen quadratic 0.90 12.90 0.030
All animals SA chen quadratic 0.89 11.83 0.040
All biota Qld gi/ch quadratic 0.97 44.30 0.001
All biota SA chen quadratic 0.95 29.50 0.011
P-values of <0.05 are not appropriate for determining statistical (or biological) significance in tables containing many correlations that are partly inter-dependent (Rice 1989). In our case, 14 sets of data for species richness of different groups and taxa both separately and in various combination have been correlated against a single set of distances from water for each gradient. By chance alone, it is possible that one of the correlations would have a P-value < 0.05 and therefore be judged significant. That is, the chance of rejecting a true null-hypothesis has to be controlled by adjusting the table-wide level of significance (Rice 1989). This adjustment is a conservative test of statistical and biological significance. In our case, the sequential Bonferroni test (Rice 1989), where the minimum P-level was 0.05/14, resulted in one significant correlation: that of all species on the Qld gidgee/ chenopod gradient.
The shapes of the curves fitted to the species-richness trends were not consistent: there were U-shaped, increasing and hump-shaped curves. Graphs of the data for the combinations with the highest r²-values are presented in Figure 3.2.1. These examples also show the variety of response shapes. The lack of consistency in the shape of the curves with species richness, and the lack of significant results are far more noteworthy than the few significant correlations. In this regard, our results are consistent with the variety of responses reported in the literature (Appendix 1). It may be argued that different responses are to be expected across a range of studies in different environments, at different times, using different techniques, different spatial scales and measuring different plant groups and animal taxa. In our study some of this variation was controlled: we surveyed a range of groups and taxa using different techniques, but the surveys were undertaken concurrently, at the same sites, in similar environments. Thus, the different responses we detected for different biotic groups are unlikely to be artefacts of the survey method but are more likely to reflect a real diversity of responses among the different biotic groups.
A = Overstorey plants on the Qld gidgee/chenopod gradient; B = All plants on the Qld mulga gradient; C = all species of plants and animals on the Qld gidgee/chenopod gradient; and D = all species of plants and animals on the SA chenopod gradient. Statistical significance of the correlations is shown in Table 3.2.12
Before you download
Most publications are available as PDF files. Adobe Acrobat Reader is required to view PDF files.
If you are unable to access a publication, please contact us to organise a suitable alternative format.
Key
Links to another web site
Opens a pop-up window