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 (continued)
The provision of artificial waters has clearly changed the abundance of many species of native biota. These changes were not readily apparent in total species richness at the gradients, but were very obvious in the composition of species persisting at different distances from water points. Numbers of plant and animal species occurring close to water points were similar to the numbers occurring at water-remote reference sites, but the identities of the species were very different in the two locations. Close to water there was a preponderance of increaser species, which, for plants and birds at least, were often widespread and/or cosmopolitan species. At sites further away from water, the mix of species was quite different. These sites were characterised by a high proportion of decreaser species which all tended to become more abundant at sites remote from water. Many were locally rare and some may be locally restricted to water-remote sites only. There is an urgent need for regional surveys to clarify the status of these species.
The highest proportions of decreaser species were found among plants growing in the understorey, averaging 38% of the plant species surveyed at each gradient (Table 188.8.131.52). The relative proportions of decreaser species were a little lower among the other plant groups and animal taxa, but, on average, represented 15-26% of the overstorey plants, seedbank plants, birds, reptiles and ants surveyed at each gradient. These trends were consistent for gradients in both acacia woodlands and chenopod shrublands, and during dry and average seasons. Exotic species were most prominent among the increaser plant species in the seedbank and growing in the understorey, where they averaged 3% of the flora at each gradient (Table 184.108.40.206). Exotic weeds were particularly prominent at the gradients in chenopod shrublands. No exotic bird species were detected but many of the increaser birds were native species whose ranges have expanded since water was provided.
(Calculated from the summary data in Tables 220.127.116.11-6).
Group or taxon No. species % Increasers % Decreasers Understorey 90±11 26±4 38±8 plants Overstorey plants 23±5 10±5 15±7 Seedbank plants 77±8 33±4 23±7 Birds 30±5 17±7 23±8 Reptiles 16±1 18±8 22±7 Ants 76±8 24±8 26±7
Table 18.104.22.168: Relative contributions of native and exotic plant species to the response groups in Table 22.214.171.124.
(There were no exotics among the animal species.)
Group or taxon % Increasers % Decreasers Native Exotic Native Exotic Understorey 23±4 3±1 38±7 0.6±0.2 plants Overstorey 100 0 100 0 plants Seedbank plants 31±4 3±1 22±6 0.6±0.4
The trends are of concern because, although the proportions of increaser and decreaser species are roughly comparable, the proportions of rangeland area providing suitable habitat for them are not. Most of the pastoral rangelands now lie within the sphere of influence of artificial waters and are therefore likely to become increasingly unsuitable for the persistence of decreaser species.
Our predominant criterion in selecting gradients for study was that they must include a reference site that was remote from all water points. The rarity of such sites imposed a major constraint on the number of potential gradients available for study. Although we were as rigorous as possible in choosing gradients that met other selection criteria, several gradients were affected by potentially confounding influences (Section 1.3.4).
Rather than detracting from the reliability of the trends we measured along the gradients, these imperfections indicate the overriding influence of distance from water, which appears to dominate above other, more localised, factors. The potentially confounding factors present at each gradient were different. The one factor that all gradients had in common was a consistent arrangement of sites away from water, in landscapes that were as similar as we could find to each gradient's reference site. Despite their local differences, all gradients showed remarkably similar and consistent relationships between species composition and distance from water.
It is not possible from our study to determine the proximate cause(s) of these changes. For those increaser bird species that require drinking water, the provision of water has probably been directly beneficial. Some increaser plant species, particularly exotic weeds, may have directly benefited from the introduction of livestock which aid the dispersal of their seed. Bird species that eat carrion may also have directly benefited from a concentration of animal carcasses close to water, and some decreaser bird species may have been disadvantaged by competition with some of the increasers. However, most of the indirect changes associated with the provision of water arise from the impact of grazing by large herbivores that focus their activities around sources of drinking water. Some decreaser plant species, particularly very palatable ones, have probably been directly disadvantaged by being grazed. Resulting changes in plant composition and architecture affect the type of food and habitat available for other biota; and changes in ground cover and soil properties affect flows of water and nutrients across the landscape, thereby initiating further change.
Despite the complexity of interactions and proximate causes, the consistency of the trends we observed among many plant groups and animal taxa along most of the gradients away from water points indicates that the provision of those water points is likely to be the ultimate cause of the trends.
For the gradients in acacia woodlands there were seasonal differences in the numbers of species seen, with substantially more plant and bird species being recorded at gradients that had experienced average, rather than below average rainfall during the preceding season. There was no indication, however, that changes in abundance of species with season had any ameliorating effect on the proportion of species identified as decreasers. For those gradients that showed seasonal trends in numbers of plant species the proportions of decreasers were substantially higher, rather than lower, at gradients that had experienced better seasons.
For plants, the ability of species to re-establish at sites from which they have been lost depends on their persistence as seeds in the soil. Although our seedbank studies indicate a tendency toward a moderately high proportion of increaser species in the seedbank, this is probably because colonising species tend to dominate the readily germinable portion of the seedbank (Section 3.3.3). Of more concern was the proportion of decreaser species detected in the soil seedbank, including some species that also showed decreaser trends in the field and others that were rarely detected in field surveys. Trends such as this suggest that at least some of the changes in plant species composition may be very difficult to reverse.
Judging by the levels of ground cover along our gradients, none of them was grazed sufficiently to cause widespread degradation. Most of the obvious impacts of grazing on ground cover were confined to sites very close to water and were no longer evident at distances of 2-3 km from water. In contrast, the changes in biodiversity that we measured were evident along the whole length of the gradients. Even at reference sites 8-15 km from water the abundance of decreaser species was still continuing to rise.
Many studies of changes in biodiversity in response to disturbance have focussed on species richness and considerable importance has been attached to how "rich" in species different sites are (e.g. many of the studies reviewed in Appendix 1). While the total number of species may be an appropriate focus for some activities, it is clearly inadequate for investigating the impact of disturbance on natural biotic assemblages. For example, while the species richness of some of our gradients has been enhanced by the addition of exotic plant species, their natural biodiversity has been diminished. From the point of view of maintaining natural biodiversity in the face of exogenous disturbance, what matters is the degree of intactness, or integrity, of the biota of interest (Majer and Beeston 1996; McIntyre and Lavorel 1994). Using this approach the focus shifts from measuring total numbers of species to assessing change in the composition of species relative to the pre-disturbance condition. Thus the lack of responsiveness of species richness to distance from water along our gradients (Section 3.2) does not tell us anything about the degree of intactness of communities at the different sites. Changes in species composition (Section 3.3) are much more informative and much more substantial.
The extent to which changes in species composition affect species richness depends on how different groups of species respond to the disturbance, and how many species are in each group. Based on responses to grazing, soil disturbance and environment, McIntyre and Lavorel (1994) identified three broad groups of plant species in temperate Australian grasslands. These were (1) intolerant species – native species intolerant of severe disturbances and which constitute the species rich component of the vegetation; (2) tolerant species – exotic and native taxa occurring at both disturbed and undisturbed habitats; and (3) disturbance specialists – predominantly exotic species correlated with high levels of disturbance. These broad groups appear to roughly correspond to our decreaser, not determined and increaser groups, with the exception that exotic species play a less prominent role. McIntyre and Lavorel (1994) suggest that the total number of species in a community depends on the balance between intolerant and disturbance-specialist species. Our results indicate that the relationships may be a little more complicated than that, since the numbers of decreaser species at a gradient frequently outnumbered the number of increaser species yet species numbers usually remained constant.
Balance is another important consideration when measuring the status of biodiversity, in terms of both the composition of species and the habitat available for them. Along our gradients the proportions of decreaser and increaser species tended to be in balance, but the distribution of waters in the pastoral rangelands indicates that the area of habitat suitable for them is likely to be severely biased against decreasers. Our gradients extended much further from water than is usual in the pastoral rangelands. For the mulga woodlands of southwest Queensland most land is likely to be within 2.5 km of a source of water (interpreted from Blick pers. comm., reported in Appendix 1). The situation is similar in the mulga woodlands of Western Australia, where most of the sheep grazing lands are likely to be within 3.5 km of a water point and most of the cattle lands within 5.5 km (interpreted from Morrisey 1984, reported in Appendix 1). Our reference sites were more than twice as far from water as this, but even at our reference sites the abundance of decreaser species had not levelled off (Sections 3.3.1-2), suggesting that water was still exerting an influence. The proportion of pastoral rangeland likely to be as far from water as our reference sites is clearly very small, but it is in this small portion that the decreaser species are most likely to be able to persist.
This also demonstrates the importance of considering change in biodiversity relative to the "best-possible" reference condition. If we had surveyed biota along truncated gradients we may have failed to detect the shift in balance between increaser and decreaser species that was apparent along our extended gradients. If we had restricted our surveys to the three sites closest to water it is probable that we would have detected a preponderance of increaser species and unaffected species, and that we may have failed to detect many of the 15-38% of species per gradient that were decreasers.
Monitoring the complex impacts of the disturbances associated with the provision of artificial waters requires useful indicators of change. Indicators that are apparent at high levels of spatial or taxonomic organisation have the most general utility. Of the variables we measured, landscape pattern and plant cover were the broadest in scale. Unfortunately, plant cover as we assessed it using standard field techniques, was not very sensitive to grazing gradients out from water. Plant cover assessed from satellite imagery appears to be sensitive to grazing gradients in at least some environments (Pickup et al. 1994). Investigating its applicability in the context of changes in biodiversity would therefore appear to be worthwhile. On-ground measures of landscape pattern, as measured on the two Queensland gradients (Ludwig et al. 1996), appear to have potential as indicators of grazing disturbance, but further work is required to determine the most appropriate parameters for monitoring landscape pattern at gradient scales, and to determine their utility in a wider variety of regions.
Among the potential biotic indicators surveyed along the gradients, understorey plants growing in the field, and birds, were both relatively efficient to sample (Section 2.3.2) and were also sensitive to the disturbance being investigated (Section 3.3.2). The most consistent response was apparent among understorey plants. This is likely to be a relatively general finding where grazing is the disturbance being investigated and the understorey flora is rich in species. Plants also offer advantages over animals as indicators, because they do not move away or hide. In practice, locally rare plant species can be easily overlooked in broad-scale surveys. The two-stage field survey method we used (quadrats plus timed searches; Section 126.96.36.199) was particularly effective for detecting locally rare species. Birds could be surveyed much more quickly, but the numbers of species of birds detected were much lower than the numbers of plant species, which may be why it was not possible to detect increaser and decreaser response groups for the birds at all gradients.
To be useful as indicators of change in other plant or animal groups or taxa, targets for survey must not only be amenable to sampling at reasonable levels of efficiency, they must also co-vary with other groups or taxa that are less easily monitored. Thus although understorey plants and birds may both be relatively efficient to sample, they are probably responding to different influences associated with water (plants to grazing, birds to changes in habitat and the presence of water). Therefore, changes in their abundance will not necessarily co-vary and both groups should be monitored to detect their range of biotic responses.
Similarly, changes in plant and bird communities will not necessarily co-vary with changes in other biotic groups, particularly those that have very different life history strategies or environmental requirements. Ground-active invertebrates, for example, are more likely to be influenced by changes in ground cover and food availability than by changes in the composition of understorey plant species per se. Invertebrates are invariably time-consuming to survey effectively, because of their small size, abundance and often cryptic habits, and because of a relative paucity of taxonomic or behavioural information (New 1995). However, invertebrates make up such a high proportion of the world's biota (at least 80% of all eucaryote species; World Conservation Monitoring Centre 1992) that they cannot be ignored when considering effects of disturbance on biodiversity.
Of the invertebrate taxa we surveyed, ants showed the strongest response to distance from water, probably because they were the most efficient to sample. They were sufficiently abundant and diverse to allow collection of large samples, and sufficiently easy to recognise to allow accurate sorting to species level by an experienced technician. They appeared to be sensitive to distance of sites from water, while showing some variation from the trends shown by vertebrates and plants. It is possible, though we do not have sufficient data to say, that other invertebrate groups could show similar patterns. If ants and any other invertebrates do co-vary, ants would have potential as indicators of change in those other groups.
It would be very helpful for understanding mechanisms and predicting changes to characterise the biological attributes of increaser and decreaser species. Such functional classifications offer the potential for distilling from long lists of increaser and decreaser species, the core descriptors of the kinds of organisms that are most responsive to change. This has already been attempted for identifying functional groups of plant species according to their sensitivity to grazing (Friedel et al. 1988; McIntyre and Lavorel 1994) but so far groupings have only been apparent at very general levels of palatability and life history strategies. Some progress has also been made toward classifying ant species along similar lines (Andersen 1995b), but again, levels of classification are still very general.
We did not identify any species recognised as nationally endangered among our decreaser response groups. Some of them may be becoming rare across much of their former range, but others may still be locally abundant in parts of their range other than our gradients. Without more detailed and systematic surveys at a regional scale, we have no way of knowing. Even without this information, our results indicate that populations of many species are in decline, representing an early warning signal that the status quo is a risky strategy for future conservation of biodiversity.
Under the status quo, very few areas have been left a long way away from water in the pastoral rangelands. This is a relatively recent state of affairs; the provision of water has only occurred since pastoral expansion. Our continental analysis of the distribution of named water points indicated how widespread artificial water points have become in the pastoral rangelands of inland Australia (Appendix 2). Artificial sources of water also persist in many conservation reserves which were originally developed as pastoral enterprises. Our results suggest that high densities of water points may be responsible for the decline of 15-38% of species in the most diverse and abundant groups of flora and fauna across much of the chenopod and acacia rangelands (Section 3.3). Therefore, there is a need for regional conservation planning to address the imbalance that has developed in the proportion of pastoral rangeland under the influence of artificial sources of water compared with the proportion remote from it.
There is also a positive side to our results: some 17-24% of animal species at a locality may be advantaged by the provision of water. (The proportions of advantaged species are a little higher among plants, but some of these are exotic weeds.) For both animals and plants from 36 to 75% of species may be unaffected. Apparently, rangeland grazing is compatible with the persistence of a moderately large proportion of native species.
For a significant proportion of species, however, there is a consistent trend of decreasing abundance with proximity to water and its associated grazing pressures. Co-existence of many of these species with the current levels and extent of grazing may not be possible. Under the status quo, some of these species may be at risk of extinction, at least on a local scale. Should regional surveys confirm that the trends are widespread, our concerns about the risk of extinction would increase. To preserve viable populations of all species at regional scales, some parts of each region need to be maintained, fostered or developed at distances far enough from water to provide secure habitat for decreaser species.