This project was the result of two separate activities: the interpretation of landcover disturbance from satellite data and the integration of these interpretations with ancillary data.
A satellite data-set was used because it was contemporary, objectively acquired, and digitally malleable.
The ancillary data-sets were used to interpret (landcover, tenure), enhance (feral animals and fire), and report (ILZ, ELZ, States and Territories) the interpreted satellite data.
The objective was to present an assessment of the landcover disturbance for the continent, and to interpret this assessment in terms of present and potential loss of biodiversity.
The acceptance of the findings presented here is largely determined by the credibility of the interpretation of landcover disturbance from satellite data because it underpins all else. The interpretation that was finally used was one of many that were exhaustively tested during the lifetime of this project.
Errors of commission and omission were unavoidable. We strove to minimise both. Nonetheless, uncertainty remains: most critically with the assessment of clearing. Where possible, we tested our findings against other estimates to indicate the direction and magnitude of error.
The results are presented in a pragmatic spirit of applied science - to present the best continental picture possible at this time. This is the same spirit and intent that motivated the publication of the Atlas of Australian Resources series; in particular, the most recent Vegetation volume; AUSLIG (1990). Australia is, as the result of considerable effort, a well-mapped continent. This raft of resource knowledge supports not only government policies but also public debate about those policies. We hope that this work will also contribute to a better understanding of the Australian continent.
Five significant points emerged at the end of this study. The first was that the level of thinning and clearing was high. Even allowing for an omission error as high as 20%, the area was much larger than has been generally appreciated. Most of this unrecognised clearing was in the traditional pastoral areas of Queensland and Northern NSW where areas of sown pastures have expanded at the expense of open eucalypt and acacia woodlands. Also, to a much smaller extent, we suspect that there has been a loss of large remnants within the core areas of the croplands. This speculation will only be tested by a continental study of the rate of change over the last few decades.
The aggregate tables of clearing and the graphics illustrating fragmentation of clearing do not convey the same impact as the view from space. The established cereal cropping areas of southern Australia present a particularly bleak panorama: especially those of Victoria, South Australia and Western Australia. They are extensively cleared. As habitats, they are irreversibly altered. Those remnants remaining are both small and isolated.
The second point was that the ELZ was, in contrast to the ILZ, much less disturbed by grazing landuse on a continental scale. This reduced level of disturbance owes more to the intrinsic resilience of the more extensive landcover types, the grasslands and low shrublands, than to the conservative management of pastoral landuse, past and present.
Considered together, the disturbance of clearing and grazing must have been, and will continue to be, a significant driver of the process of biotic erosion. The landcovers that have been most disturbed are the best country; those that enjoy the highest rainfall and therefore have the highest productivity. While we do not yet have the quantitative functional relationships to support this conclusion, we believe it must be so. Between them, the landuses of agriculture and pastoralism have 'taken the eyes of the land'. The case studies support this view.
The third point was the exacerbating impact of feral animals. Where landuse disturbance was high, so also was the density of feral animals.
The next point was not illustrated by data in this report but it bears heavily on any conclusions that can be drawn from it. Simply: Of the two drivers of future biotic erosion, the contribution of landuse, from clearing and poor grazing management is manageable while that from feral animals is not; or not yet, although concerted efforts are underway to achieve this goal.
Finally, the basic premise of this project - that there is a relationship between disturbance and biodiversity - remains qualitative and the limiting factor in translating this or any other measure of landcover disturbance into quantitative estimates of biotic erosion. While we hold that habitat destruction by clearing or grazing, and habitat invasion by feral species are the two most significant drivers of biotic erosion, we cannot translate the assessments we have compiled here into any action.
Why not?
Two key relationships are missing. These are for any landcover type, the relationship between species abundance and disturbance, and between the level of endemism and the level of disturbance threat.
These two relationships are considered sequentially. The first relationship, displays species abundance within a landcover type as a function of its use of landcover geography; ie in space and time (x, y, z, t). While this relationship is strictly a multidimensional surface, and not necessarily continuous, three hypothetical responses are presented. Whatever the shape of the curve, it represents a means of evaluating the current level of Threat. That is, given this general relationship and actual measures of the current state of landscape disturbance dimensioned in (x, y, z, t), then the species response to Threat, in terms of its abundance in time and space, can then be estimated.
An idealised framework for relating species abundance in time and space to the time and space dimensions of habitat disturbance. Three plausible response relationships are presented: a species declines in linear proportion to the degree of disturbance (1); a species shows resilience, then catastrophe (2); a species is highly sensitive to disturbance (3).
The second relationship is between threat to a landcover type and the level of endemism within that species complement. We presume that endemism is the ultimate measure of irreplaceability. Landcover types will be distributed within this two dimensional space as indicated by hypothetical examples. The purpose of mapping this relationship is to determine priority for action. Even the simplest partition of this space into two levels, low and high, would be more informative than our current understanding.
Hypothetical relationships between the overall level of species endemism for landcover types and the level of Threat to which each landcover type is exposed. By partitioning this space into high and low, priorities for management can be rationally determined. For example, landcover types 2 and 7 score the highest priority for conservation management.
In closing, we bring attention to the issue of the location and rate of landcover change; the dynamics of disturbance that was outside the scope of this project.
Obviously, clearing was an essential part of agricultural development. However, land degradation is an unnecessary consequence of poor pastoral management.
The expansion of these two agents of landcover disturbance has been driven and controlled by social forces, both local and international.
These forces are dynamic. Highly favourable commodity export prices for wool, grain and beef come and go: those for the first two are unlikely to ever be quite so favourable as in the past.
In the last two decades, changes in the values society attaches to all the resources that land has to offer have found political expression with most State Governments altering legislation, or its management expression, relating to clearing and to pastoral management. The motivation has been a long-term concern about sustainability, and the expensive experience that the past approaches were not perfect.
Given these changes, what are the dynamics of landcover change?
It is surprising, given the pivotal role of this question, that it cannot be accurately answered; see Biodiversity Unit (1995).
An aggregate, non-spatial summary from data held by the ABS has been compiled but not yet published by Bullen (F. Bullen, personal communication). This time series shows the great expansion of clearing post-World War II. This clearing was both for cereal crops (WA, SA, NSW and Qld) and for pastures (principally Qld).
Noting that these ABS figures report less than 50% of the area that we have found to be thinned and cleared, they also suggest that clearing is slowing. However, non-systematic sampling surveys using satellite data show the loss of uncleared remnants within the core agricultural land continues; Graetz et al (1992), Sivertsen (1994). In terms of biotic erosion, the loss of these remnants is far more critical than indicated by their small areas.
The historical time course of clearing for crops and pastures. Since the beginning of this century, the area cleared for pastures has exceeded that for crops culminating in a marked expansion from 1950 onwards.
Land management decisions at the level of the individual and of governments will continue to influence the rate and direction of the process of biotic erosion. The objective of this project was to provide a contemporary assessment of the landcover condition of the continent. This assessment in a small way will support that decision making.
