Biodiversity publications archive

Landscape planning for biodiversity conservation in agricultural regions: A case study from the Wheatbelt of Western Australia

Biodiversity Technical Paper, No. 2
Robert J. Lambeck, CSIRO Division of Wildlife and Ecology
Commonwealth of Australia, 1999
ISBN 0 6422 1423 9

Chapter 2 - Retaining biodiversity in agricultural landscapes (continued)

2.6 Reintroductions

The second of the strategic enhancement objectives identified in section 2.3 aims not only to retain the current species in the landscape but also to restore species that previously occurred in the region but are no longer present.

The reintroduction of species which have disappeared as a result of habitat removal and fragmentation would ideally be considered in the context of a regional planning approach. Information about the habitat and resource requirements of the species to be reintroduced will rarely be available for the site where the reintroduction is to occur and therefore needs to be acquired from places where the target species still exists. However, it cannot be assumed that the local site characteristics where a species currently occurs will also be adequate for that species if duplicated elsewhere. For example, a species which has been lost from the central wheatbelt is the rufous treecreeper, a medium sized, insectivorous, woodland bird. Examination of its habitat requirements in a location where it does still occur reveals that it occupies a wide range of woodland patches ranging from large pristine reserves through to small patches of degraded woodland. Simple observations of habitat use would suggest that small patches of woodland are adequate for this species. However, there are many such patches in the Kellerberrin area and yet the species has disappeared.

It appears that this species is able to occupy these small patches only because of the regional context in which it is found. The presence of large pristine reserves nearby probably provides a source of new recruits to these smaller remnants. In fact, these small remnants may act as sinks into which individuals disperse, but from which there is no recruitment. It is also possible that rufous tree-creepers may not be secure in the areas where they currently occur. Their presence may simply reflect a slower rate of decline due to the less fragmented landscape than was the case in Kellerberrin.

The important point to be drawn from this example is that it is not possible to simply duplicate, in another location, the local habitat patterns of an area where the target species still persists and expect them to deliver the same conservation value. Any translocation of a species between locations will require a detailed knowledge of the ecology of that species, which takes into account not only the local habitat requirements, but also the regional context in which that species occurs. Such an analysis was beyond the scope of the current project but is worthy of further investigation if restoration objectives are to be considered seriously. A list of species known to have disappeared from the Kellerberrin area is provided in Appendix 4.

2.7 Mixed strategies in the face of partial knowledge

Ideally, any strategy for conserving the biota in a region would be based on an understanding of the requirements of the flora and fauna. In reality, such complete knowledge will never be available, resulting in a need to develop strategies based on a combination of quantitative data, expert opinion and general principles. The dilemma facing managers is that our major concerns are for those elements of the biota that are at greatest risk. By and large, these tend to be less common species which are less amenable to rigorous scientific investigation. As a consequence we are forced to rely on expert opinion. While this opinion is often based on many years of accumulated experience and anecdotal observations by a number of people, its reliability cannot be easily assessed. General principles on the other hand provide even less guidance for ensuring that particular species will persist.

If we are to adopt objectives which aim to achieve more than simply improving on our current position in some unspecified manner, we will need to develop ways of efficiently acquiring the essential knowledge required for managing particular landscapes. However, in situations where there is an urgent need to act, or where the resources required for effective management are unlikely to become available, it will be necessary to increasingly rely on expert opinion and general principles. It must be recognised, however, that the greater the emphasis on general principles, the less likely we will be to meet the needs of particular species that are at risk in our agricultural landscapes.

2.8 Summary of design and management recommendations for Wallatin Creek

Because the focal species approach presented here is a novel one, much of the data that are required for such an analysis were not available for the study area. The approach requires information about the most sensitive and vulnerable species in the landscape. This information is difficult to obtain because such species are usually uncommon. Where the required information was not available or insufficiently detailed it was necessary to draw upon the informed opinion of professional ecologists and knowledgeable locals. Where even this level of knowledge was unavailable it was necessary to rely on general ecological principles.

By applying the focal-species procedure it was possible to identify area-limited and dispersal-limited focal species whose requirements were used to develop estimates of the spatial characteristics of habitat patches and the maximum acceptable distances between patches. The recommended minimum area for the dominant patch types are greater than 23 ha for woodland, 25 ha for shrubland/mallee, and 25 ha for heathland.

Information about corridor characteristics were more difficult to prescribe, but it is clear that for many small vertebrates and less mobile invertebrates, this linear vegetation must act as habitat in which population processes must persist, rather than as conduits through which individuals move. On this basis it was possible to present some tentative recommendations about the preferred lengths and widths of linear vegetation. The distance between remnants should not exceed two kilometres. Linear vegetation linking occupied habitat patches of dispersal-limited birds to the nearest suitable patch should be approximately 50m wide. Other linear vegetation, if it is to provide habitat for smaller animals and be resilient to weed invasion should be greater than 30m wide for heath communities and greater than 60m wide for more open vegetation types. It was not possible, in the time available for this project, to use the focal species approach to specify the compositional attributes of linear vegetation. Consequently, it was necessary to draw on general principles which suggest that the type of species planted should be of local provenance and that they should be planted on the appropriate soil type. In order to meet the habitat needs of some of the vulnerable invertebrates and small vertebrates, areas reconstructed with open vegetation types such as woodland should also contain clumps of denser understorey vegetation.

The requirements of resource-limited species suggest that habitat reconstruction should emphasise plant species that produce nectar over the summer-autumn period and that reconstructed habitat patches should contain areas of intact litter and areas in which the canopy is sufficiently closed to provide the more mesic conditions required by some vulnerable scorpion species.

Recommendations for managing other threatening processes in the catchment include (i) that feral predators be reduced to the lowest possible densities with an ultimate objective of eradication; (ii) that grazing by stock of remnant vegetation and reconstructed habitat be stopped (iii) that fire be used as a management tool only in situations where senescence of the dominant vegetation is apparent and (iv) that inter-fire periods should exceed 50 years unless additional evidence becomes available to suggest otherwise. It is important to recognise that these recommendations have been derived specifically for the Wallatin Catchment and should not be transferred to other locations without a critical assessment of their appropriateness.

2.9 Priorities for implementation

Sufficient resources will rarely be available to immediately implement the results of a planning exercise. Consequently, it is necessary to identify those actions which would contribute most to meeting the objectives of the exercise and hence should be preferentially implemented. A general principle for setting priorities for habitat reconstruction is to firstly protect areas of high conservation value and then to build outwards from these with the aim of providing adequate habitat and maximising the connectivity of the landscape.

Beyond this very general principle, it is best to base priorities firstly, on identifying areas which best represent the biological diversity and secondly, on ensuring that the biota is able to persist in the selected sites. Previous sections have addressed the issue of species retention. This section will focus on attempts to represent biological diversity in the landscape.

2.9.1 Representing biological diversity

If detailed surveys of the biota have not been conducted, priorities can simply be set on the basis of easily assessed variables such as the size and condition of the remnants and the range of patch types that they contain (Safstrom 1995). Such an approach will increase the probability of capturing the diversity of the region but there will always be the risk that important elements will be overlooked. Ideally the setting of priorities would be based on more detailed surveys of the flora and fauna in order to identify those remnants which most efficiently represent the full range of biological diversity in the area.

Priority setting should also take into account the broader regional context in which the planning is being conducted. Species which are common locally and hence do not appear to be a priority, may not occur elsewhere and hence are significant from a regional perspective. Conversely species that are locally rare may be common elsewhere. By having a regional context for planning it will be possible to better identify the conservation priorities and determine the most effective contribution that local people can make towards addressing these regional priorities.

The irregular distribution of conservation values in landscapes has underpinned a significant effort over the last decade to develop systematic methods to represent the full array of biological diversity in a region. These methods include simple scoring or ranking procedures (review by Margules & Usher 1981; Safstrom 1995), linear programming (Cocks & Baird 1989) and iterative procedures (Kirkpatrick 1983; Margules & Nicholls 1987; Margules et al. 1988; Margules 1989; Pressey & Nicholls 1989; Pressey et al. 1990; Pressey & Nicholls 1991; Ryti 1992; Brooker & Margules 1996). These procedures vary in their complexity and data requirements and the reader is referred to the references provided for more information about their capabilities.

Priority areas for the Wallatin Creek case study

Formal assessment of the representativeness of the biota has been conducted for plant species in both the Kellerberrin district and in the Wallatin Catchment (Brooker & Margules 1996). Additional surveys have been conducted for plant assemblages (Arnold & Weeldenburg 1991; Lambeck & Wallace 1993), birds (Cale & Lambeck unpublished data), reptiles (Smith unpublished data) and invertebrates (Abensperg-Traun et al. 1996). The invertebrate and reptile surveys have concentrated on selected vegetation assemblages and hence are not comprehensive. They do however enable us to assess the relative value of the sampled vegetation types for representing the greatest numbers of locally indigenous species.

Vegetation assemblages

Arnold and Weeldenburg (1991) estimated the representation of different landform types in remnants and in road verges. Landforms provide a rough surrogate for vegetation communities because each landform type tends to correspond with characteristic vegetation associations (Beard 1980). They found that the proportional representation of the different landform types in road verges was roughly equivalent to that in the landscape as a whole, but that the vegetation remnants over-represented rock outcrops and under-represented the low-lying fertile land units which supported salmon gum/gimlet woodlands and York gum woodlands. The high correspondence within road verge vegetation reflects the fact that the road network covers the whole region in a north-south, east-west grid and provides an effective sample of the whole range of vegetation diversity.

Further assessment of the vegetation communities using Landsat TM imagery combined with air photo interpretation and ground survey (Lambeck & Wallace 1993) found that within the remnants that occurred in the district, only 30% of the remnant vegetation could be considered to be healthy examples of the dominant vegetation types in the region. Many of the remnants contained substantial areas of rock which were not suitable for agriculture and of the remainder, most are degraded to varying degrees by stock grazing, weed invasion, rubbish dumping, vehicle tracks, and gravel extraction. A comparison of the percentage of vegetation types in good condition with landform types in each remnant showed that woodland and sand-plain heath communities were under-represented and rock outcrops over-represented. Banksia woodlands were particularly poorly represented with only 1 small patch and a few scattered trees remaining in the local area. Of the other woodland types, Salmon gum and Gimlet woodlands are less well represented than are Wandoo and York gum. From these studies, it is clear that the remaining vegetation does not represent the original vegetation in terms of its distribution, proportions or condition and that priority should be directed towards protecting and enhancing the less well represented vegetation types.

Plant species

The study area contains 11 species listed as Declared Rare Flora by the Department of Conservation and Land Management (Mollemans et al. 1993). Of these, 6 species occur on private land or in road verges. All remnants in the catchment containing these species were nominated as priorities for protection and enhancement.

Further prioritisation of remnants was conducted by Brooker and Margules (1996) who ranked remnants in terms of their relative importance for representing plant diversity at both local and regional scales. In this study, plant species, together with a suite of environmental variables, were recorded at 125 sample sites. From these variables, a model was developed which predicted the plant species diversity in other remnants in the catchment based on their environmental characteristics. Figures 8 and 9 show the 10 most important remnants based on regional and local analyses. Comparison of these figures reveals that remnant priorities are scale dependent with some remnants considered equally important at both scales while others are only important when assessed at a local scale.

Figure 8: Map of the Kellerberrin region showing the top ten remnants for representing the plant species diversity of the region.

Figure 8: Map of the Kellerberrin region showing the top ten remnants for representing the plant species diversity of the region.

Source: Brooker & Margules (1996)


The only mammal species considered at risk in the catchment is the brush wallaby (Macropus irma) which is known only from Durokoppin Nature Reserve. The distribution and abundance of other mammal species in the catchment is not known. It is therefore not possible to rank remnants according to their importance for mammals, apart from considering Durokoppin Nature Reserve to be the highest priority.


Survey data from all remnants in the catchment (Lambeck & Cale unpublished data) were used to rank remnants according to the number of bird species that were considered a conservation priority in the region. Selection of priority species was based on records of species known to have declined in the region (Saunders 1989; Saunders & Ingram 1995) and which were considered at risk of further decline or local extinction in the absence of interventionary management. The remnants selected on the basis of presence/absence of priority species are illustrated in Figure 10.


No comprehensive survey has been conducted for reptiles in the study area. However, examination of lizard species richness in selected vegetation types indicates weak relationships with native plant species richness, vegetation structure and remnant area in Gimlet woodland, but no relationships with vegetation attributes or area in shrublands (Abensperg-Traun et al. 1996). On the basis of these limited data, the only criteria for prioritising remnants for reptiles would be on woodland area. However, the data were not considered to be sufficiently comprehensive to warrant such an assessment.

Figure 9: The top ten remnants in Wallatin Catchment ranked on the basis of their contribution towards representing the plant species diversity of the catchment.

Figure 9: The top ten remnants in Wallatin Catchment ranked on the basis of their contribution towards representing the plant species diversity of the catchment.

Source: Brooker & Margules (1996)


Invertebrate diversity has been sampled only in woodland and shrubland communities in the study area and the study considered species richness of only a selected subset of the arthropod fauna (beetles, ants, scorpions, isopods, cockroaches, termites, earwigs, hemipterans and butterflies (Abensperg-Traun et al. 1996)). This study indicated that vegetation characteristics such as structural diversity or plant species richness were poor predictors of invertebrate diversity. In the absence of such predictors, it is considered inappropriate to attempt to rank remnants in terms of their contribution to representing invertebrate diversity.

The above information indicates that the only available data appropriate for ranking remnants in terms of their relative importance for representing the biota are those for plants and birds. Comparison of the rankings for birds and plants indicates that the same remnants have different values from the perspective of each group. It should be expected therefore that similar surveys for reptiles, mammals or invertebrates would produce other combinations of sites for best representing the diversity of each group. As a consequence, priorities for protection and reconstruction in the Wallatin Catchment are firstly, remnants known to contain vulnerable plant and bird species and secondly, remnants that best represent various taxonomic groups.

It is important to bear in mind that, for the purpose of this exercise, all remnants are considered worthy of retention and the priority setting process is aimed primarily at identifying where protection or restoration actions should preferentially occur. Because a remnant is not identified as a priority remnant does not imply that it is not important in the landscape and hence does not need to be protected.

Ranking remnants in terms of their conservation value is clearly only the first step in a conservation plan. There is no point in simply identifying important remnants if they are degrading and losing the values that they currently contain. The focal species approach described previously should subsequently be applied to ensure that the remnants are able to retain their biota.

Figure 10: The Wallatin Creek Catchment illustrating the top 10 remnants for representing the most vulnerable bird species.

Figure 10: The Wallatin Creek Catchment illustrating the top 10 remnants for representing the most vulnerable bird species.

Note: Many of these remnants extend beyond the catchment boundary and remnants 4 & 8 lie outside of the catchment but are owned by land-holders within the catchment.