National Wilderness Inventory
Australia: Our national stories
Australian Heritage Commission, 2003
ISBN 0 642 23561 9
6. Concluding Remarks (continued)
6.2 The Modelling Process
There are other important sets of assumptions built into the techniques devised for measuring indicator values. It is important for these assumptions to be recognised in making use of survey results, as should the more fundamental conceptual considerations mentioned above. There will always be limitations in the way real-world complexity is encompassed within a modelling process. In formulating procedures for the construction and calculation of wilderness indices, the principle was adopted that the modelling process should be as simple as possible, consistent with producing sensible and useful results. There were three main reasons for this.
The first reason was that the task of the wilderness quality indicators is to measure attributes that are inherently complex and difficult to define, even in conceptual terms. Notions of remoteness and naturalness are exceedingly difficult to deal with in a complete, precise and meaningful way. Because of this, there is likely to be little advantage gained from pursuing complex modelling techniques, when these are likely to have deficiencies of similar orders of magnitude to relatively simpler procedures.
Secondly, relatively complex modelling has the disadvantage that it can become difficult to understand and interpret. Given that any modelling procedure dealing with the attributes described above will have deficiencies and contentious aspects, it is far better for these to be relatively simple and exposed, and therefore more likely to be understood. Results produced can then be accepted, rejected, or interpreted with full knowledge of processes.
The final reason for adopting a more straightforward approach to modelling was that of primary data availability. For national and even regional surveying purposes only limited primary data sets were available. For instance, at the time NWI survey work commenced, a suitable national digital elevation model was not in existence. This meant that incorporating terrain into remoteness and apparent naturalness modelling was not feasible consistently across the continent. Likewise, important data sets for the Biophysical Naturalness assessment were not consistently available meaning that only very simple approaches to naturalness could be contemplated. Some of the more important assumptions built into the implementation of the NWI model are presented below.
- Remoteness was measured using Euclidean distance from selected landscape features. No attempt was made to introduce terrain factors such as those relating to elevation and surface roughness and other surface restrictions and impediments.
- View field was not taken into consideration with regard to the visibility of intrusions in the landscape.
- Distance decay functions and class limits used in standardising index values were essentially arbitrary. These were set in such a manner as to effectively display variation in parameters in landscapes subject to moderate levels of development (as established in Victoria). The limits applied in previous studies were also taken into account.
- The Biophysical Naturalness assessment procedure is based on the key assumption that the greater the intensity of land use activity, the greater degree of biophysical alteration the natural environment has undergone. It also assumes that land use activity within essentially natural landscapes are restricted to timber harvesting and livestock grazing. Such assumptions (and others detailed in this handbook) do not take into account many of the vast array of additional factors relevant, if not critical, to an assessment of the integrity of natural biophysical processes in the landscape.
None of these assumptions invalidates the implementation process, but they do have important implications for the interpretation of results. In particular, where there is additional local information that is relevant to a consideration of wilderness quality, then this local information should be taken into account when interpreting standard results. This is particularly important for local area applications where it is known that particular factors, such as feral animals, have a known and significant impact on biophysical conditions.