To reiterate a critical point made earlier: the MSS dataset was the primary source for detecting landcover disturbance. However, ancillary datasets were equally important in the interpretation, quantification, and enhancement of that disturbance assessment. In this section, these ancillary datasets are listed and discussed in order of importance.
Tenure is a key factor in any assessment of landuse and landcover because tenure is a licence that prescribes or proscribes various landuses, and thus various types of landcover disturbance. Moreover, the tenure type is a strong guide to future landuse and landcover change. For example, there is far more certainty about the future landuse activities on land held under public tenure than that for private tenure. Thus, tenure can indicate actual and potential landuse as well as its consequences for landcover disturbance.
The basic dataset used was that compiled by AUSLIG (1994). This digital dataset contains an initial 16 categories that were aggregated into a smaller set of 6 to indicate the probability of future landcover disturbance by grouping by differing levels of tenure protection.
The protection is principally against the threat of clearing but also includes grazing. We created classes of tenure protection that range from 0 with freehold tenure to 5 with existing dedicated conservation reserve tenure; Table 4 and the Tenure image. This assigned rank was used to rate and score the tenure protection for each landcover type.
Table 4: The aggregation of the AUSLIG tenures into six groups that offer different levels of protection against future landcover disturbance.
Type Grouping Area (km2) % Protection
____________________________________________________________________
1 Freehold 1591592 21 0
Mining leasehold
2 Pastoral leasehold 3257354 42 1
Defence tenure
3 Forestry 247467 3 2
Water Supply
Mixed, Other, Multi-
4 Aboriginal freehold / 1110656 14 3
leasehold national park
Aboriginal leasehold
national park
5 Unallocated crown land 961057 12 4
(unused)
6 Dedicated nature 523875 7 5
conservation
The size and distribution of the six tenures that reflect protection against future landcover disturbance. The protection rankings and colour codes are: 0, red; 1, brown; 2, purple; 3, light blue; 4, blue; 5, green.
After landuse-driven landcover disturbance, the next most significant contributor to the process of biotic erosion is habitat invasion by feral (exotic) animals; eg Newsome (1994), Short and Smith (1994), Short and Turner (1994).
By comparison, the significance of habitat invasion by feral plants is much less well understood or documented on a continental scale; eg Humphries et al (1994). Thus, exotic plants were not included in this project.
It is difficult to overestimate the significance of the impact of feral animals on the flora and fauna of this continent. Much has been made of the relentless influence two predators, the feral cat and fox, on the small-sized fauna, particularly in the ELZ. However, the rabbit cannot be ignored in the overall assessment of the driving forces of animal extinction. Its arrival and establishment in the more arid parts of the ELZ led to the destruction of entire landscapes as captured in this image of the Strzelecki Track in South Australia.
Continental datasets of density distribution for a range of pest mammals have been compiled; Wilson et al (1992). From this larger set, four were selected as being the most influential in the process of biotic impoverishment on continental scales; the rabbit, cat, fox and pig. Other feral species (camel, goat, horse) are not as widespread or as influential as the chosen species.
The feral density datasets are of relative rather than absolute, and overall their quality is relatively poor. Nevertheless, they are unique and were used to assess the spatial patterns of intensity of each feral species' current threat to biodiversity.
The density distributions of the four feral animal species used in the project. Clockwise from the top left the distributions are rabbit, cat, fox, and pig. The density codes are None (white); low (yellow); moderate (blue); high (red).
The best available datasets of fire season and fire frequency for the continent known to the authors were digitised and used to assess the current level of landuse disturbance from this facet of landuse.
The continental distribution of fire season. The categories and codes are: Winter-Spring (purple), Spring (blue); Spring-Summer (light blue); Summer (brown); Summer-Autumn (red). This dataset was derived from that published by Luke and McArthur (1978).
The continental distribution of fire frequency aggregated into just three classes: 1-3 years (red); 5-10 years (blue); > 10 years (brown). This dataset was derived from that published by Walker (1981)
These two datasets were used to express their influence on the process of biotic erosion. In doing so we question the conventional wisdom: not that fire has always been a part of the environment of all biota on this continent, but whether the fire regimes which we see today differ significantly from pre-European times. While wildfires occur from lightning ignition, particularly in the hummock grasslands of Central and Western Australia, and are constrained by climatically controlled fuel production and curing, we assert that the predominant source of ignition is increasingly from landuse; that is deliberate ignition.
The ancillary datasets discussed above were used at 0.01 x 0.01° (approximately 1 x 1 km) resolution; a much finer scale than at which any of them were compiled. This is not as serious a problem as is the veracity of the data itself. Nevertheless, these data represent the best continentally consistent sets available and they were used to derive the best continentally consistent assessments of landcover disturbance possible now.
