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

2.1 Biodiversity: the variety of life

The term biodiversity was coined to describe the complex variety of life that characterises this planet. Biodiversity encompasses differences at all levels of biological organisation. It includes differences between individuals within a species; between species themselves; between communities, landscapes and ecosystems. Given that all-encompassing nature of the term, it is apparent that the conservation of this diversity is not a trivial task and that it will not be possible to develop management strategies that explicitly consider the needs of the biota at all of these levels of complexity. Management of biological diversity in production landscapes must therefore be based on procedures that consider particular levels of this hierarchy of biological diversity but which also have a high probability of encompassing the other levels. In the current study, attention was primarily directed to the consideration of species and communities at a landscape scale. Owing to the absence of information about the range of variation within species it is necessary to assume that strategies which will retain communities of species will also retain the variability that occurs between the individuals of those species.

2.2 Biotic impoverishment: the impact of agriculture

Because biodiversity incorporates differences at all levels of biological organisation, from individuals to ecosystems, and substantial change in a landscape will bring about a change in biodiversity. Such changes are of particular concern when patterns of landuse cause the loss of species or changes in their distribution.

In agricultural regions throughout Australia trajectories of change are consistently towards declining natural diversity (Saunders 1989; Department of Environment, Sport and Territories 1996). This decline in diversity results in changes to the rates and pathways of ecosystem processes that are essential for conserving biodiversity and sustaining agriculture (Saunders et al. 1991; Mooney et al. 1995).

2.2.1 Changes in biota

Twenty four plant species are known to have been lost from the wheatbelt as a whole and the area now has one of the highest numbers of rare and/or endangered plant species in Australia (Briggs & Leigh 1996). Of the 348 plant species listed as rare or endangered in the wheatbelt, only 79 occur in designated nature reserves. Of the remainder, 135 species occur in road verges and 53 are found on privately owned remnants (Hopper et al. 1990).

Of the 43 mammal species that occurred in the region prior to European settlement, only 12 species are now moderately common or abundant (Kitchener et al. 1980a). Mammal declines are continuing with several species disappearing or declining in abundance in the Kellerberrin area over the last 15 years (Hobbs et al. 1993). While loss of habitat has significantly reduced the population densities of these species, feral predators have also played a major role in the process of decline and extinction (Burbidge & McKenzie 1989). It appears that bird species may be following a similar trajectory of decline, albeit at a slower rate (Saunders 1989; Saunders & Ingram 1995). Thirty one species have decreased in range and/or abundance in the wheatbelt over the past 90 years. Of these, 15 species no longer occur in the Kellerberrin district (Saunders & Curry 1990; Hobbs et al. 1993). Lizards appear to have been less severely affected with no obvious widespread loss (Kitchener et al. 1980b). While less information is available for frogs, it would be expected that many species would be deleteriously affected by increasing salinity both in the soil and in water courses.

While there is no evidence that changing patterns of land use have resulted in the extinction of any invertebrate species, it is clear that they have caused substantial changes to local invertebrate diversity (Springett 1976; Scougall et al. 1993). The absence of evidence of invertebrate extinctions due to land use practices does not mean that invertebrate species are secure in agricultural landscapes. Rather, it simply reflects the absence of appropriate data for detecting such extinctions if they have occurred.

2.2.2 Changes to ecosystem processes

Changes to ecosystem functions have resulted from changes in physical processes due to altered land-cover, as well as from changes in the diversity of species that participate in the movement of nutrients, water and energy throughout the system.

Physical changes

Research conducted in a range of agricultural regions throughout the world indicate that extensive land clearing can modify patterns of radiation and alter fluxes of wind, water and nutrients across landscapes (Baudry 1989; Risser 1990; Saunders et al. 1991; Hobbs 1992a). Radiation fluxes tend to be more variable in agricultural fields than they are in remnant vegetation both throughout the day and between seasons, resulting in greater local temperature extremes. Where natural bushland and agricultural fields meet, the higher radiation from the bare ground affects the microclimate at the remnant edge causing changes in plant composition and structure (Lovejoy et al. 1986; Palik & Murphy 1990). Altered radiation levels in fields are also likely to modify soil micro-habitats for soil-dwelling invertebrates. Increased exposure to wind affects conditions at the edge of remnants, with increased frequency of tree blow-downs, increased evapotranspiration (Lovejoy et al. 1986) and increased transfer of weed seeds, nutrients and agricultural chemicals from farmland into vegetation remnants. Similarly, increased surface water flow brings nutrients, chemicals and weed seeds into remnants (Muir 1979; Cale & Hobbs 1991).

Changes in vegetation cover have affected the hydrological balance in the catchment causing a rise in saline water tables which impacts initially on vegetation remnants in the lower valley floors but, as water tables continue to rise, vegetation located higher up the slopes is increasingly being threatened.

Biotic changes

Considerable attention has been paid to changes in ecosystem function associated with changes in biodiversity (Hobbs 1992b; Lambeck 1992; Main 1992; Mooney et al. 1995). Much of the debate revolves around the question of whether greater levels of diversity result in greater ecosystem stability or resilience in the face of environmental variability. MacArthur (1957) argued that a greater number of species provides more pathways for energy to reach a consumer. Consequently, the loss of any one pathway would be less likely to have a detrimental impact on that consumer. Elton (1958) also contended that the increasing complexity of an ecosystem that resulted from greater species diversity should generate greater stability. However, Cody (1986) found that high diversity in Mediterranean-climate regions tends to be associated with a greater probability of rarity and a consequent greater risk of extinction. However, the loss of a rare species from a diverse system is likely to have less impact on ecosystem processes than would the loss of a more abundant species from a less diverse system. This is because a more diverse system will have a greater probability of containing other species which play a similar functional role and hence can compensate for the loss of a particular species. Such compensation may not occur if a 'keystone' species is lost. This is a species which has an impact on ecosystem processes that is disproportionate to its representation in the community (Power & Mills 1995). In such circumstances, a small change in diversity can have a significant impact on ecosystem function.

The impact of changes in biodiversity on function will therefore depend upon a number of factors including the number of species in the system, the relative abundance of those species, the functional role played by the species lost or gained and the relationships between these and other species in the system (Mooney et al. 1995).

2.3 Protecting biological diversity: the need for clear objectives

Before any attempt is made to address conservation issues in landscapes used for production it is essential that the objective of the exercise is clearly identified. Two broad approaches to nature conservation can be considered:

General enhancement, which attempts to maximise the number of indigenous species retained or, alternatively, to minimise the number lost within constraints imposed by other land use objectives.

Strategic enhancement, which aims to ensure the persistence of particular species, groups of species, or all species that currently occur in a landscape. This type of approach could be extended to include the reintroduction of species that have been lost from the landscape being managed.

The objective of the first type of approach – to maximise the number of species retained, or to minimise the number lost – is an open-ended objective. It identifies a general trajectory along which to proceed, but does not specify targets which can be used to assess success or failure. This type of approach does not consider which species will be conserved or lost but simply aims to increase the probability that any given species will persist. It is a 'general enhancement' objective, in that it aims to 'make things better or 'minimise the impact' in an unspecified manner within the constraints of other land uses.

Approaches based on retaining or reintroducing specified components of the biota can be considered to be 'strategic' because they require specification of the landscape elements and management regimes that are required to meet a specific objective. They are more rigorous because they have quantifiable outcomes by which we can judge the effectiveness of our actions. If our objective is to retain all of the biota or a particular component of that biota, then the loss of any species, or of particular designated species would constitute failure. Similarly, if our goal is to reintroduce particular species that have been lost from an area, the failure of those species to become established is a clear indication that the objective has not been met.

This distinction between these two broad types of objectives is an important one because the management strategy adopted will differ considerably depending on the objective chosen. A general enhancement strategy must rely on ecological principles, while a strategic enhancement approach depends on having information about the requirements of the species to be retained or restored. These alternative strategies will result in different recommendations about the spatial and compositional characteristics of a landscape and will have significantly different conservation outcomes. The different types of objectives associated with these different approaches are therefore considered below in greater detail and the design and management implications of each are examined.

2.4 General enhancement

General enhancement strategies attempt to identify ways in which a landscape can be improved in order to reduce the probability of species being lost. Such strategies tend to draw on general ecological relationships between landscape or habitat characteristics and attributes of biological communities. These attributes typically include species richness or diversity, guild richness, biomass, or trophic structure. These types of relationships have received considerable attention at a theoretical level. For example, species-area models suggest that habitats of different sizes will support different numbers of species (Munroe 1953; MacArthur & Wilson 1963); patch dynamic theories (Cody 1975; Roth 1976; Pickett & White 1985) imply that areas having greater horizontal diversity (patchiness) both within and between habitats will support greater numbers of species; niche theory (Hutchinson 1958; MacArthur & MacArthur 1961; Wiens 1989) similarly predicts that areas having greater vertical and horizontal complexity will provide more niches and will therefore support more species; the intermediate disturbance hypothesis (Grime 1973; Connell 1978) implies that different intensities or frequencies of disturbance will produce communities with different levels of species richness; and island biogeographic theory (MacArthur & Wilson 1967) and metapopulation theory (Levins 1970; Gilpin & Hanski 1991) predict that more fragmented or isolated habitats will support fewer species than habitats that are more connected, and that populations in more fragmented and isolated habitats are less likely to persist through time.

Where management objectives aim simply to maximise the number of species present, these general principles can provide a useful guide to the attributes that we should retain in existing habitat or incorporate into revegetation strategies. If such an objective is to be employed, emphasis should be placed on maximising the number of locally indigenous species, rather than simply maximising species numbers per se.

Vertical heterogeneity

Much of the literature relating structural diversity to species numbers is based on studies of bird communities. MacArthur and MacArthur (1961) and Recher (1969) identified strong relationships between foliage height diversity and bird-species diversity in both America and Australia. While other studies have revealed weaker relationships, or in some cases no relationships between these variables (eg. Emlen 1977; Rice et al. 1983), they often found a correlation between bird-species diversity and some other aspect of vegetation structure (Wiens 1989). Mammal species also respond strongly to differences in vegetation structure. Rosenzweig and Winakur (1969) and August (1983) found a positive correlation between habitat complexity (number of vertical strata) and the total number of mammal species in tropical forests. In Australia, the diversity and abundance of small mammals and arboreal marsupials are also influenced by vegetation structure (Barnett et al. 1978; Laidlaw & Wilson 1989; Lindenmayer et al. 1994a, 1994b). While these relationships do not hold under all circumstances (see for example, Bond et al. 1978), the majority of observations suggest that greater structural complexity is unlikely to have a detrimental effect on species diversity and in most cases will enhance species numbers through the provision of a wider variety of habitats. On the basis of these observations, attempts should be made to ensure that the appropriate number of structural layers are maintained in vegetation remnants or are incorporated into reconstructed habitat. Depending on the vegetation types typical of the area, these layers may include some form of ground cover, understorey shrubs, taller middle-story shrubs or smaller trees, and an upper canopy of taller tree species. Guidance as to appropriate levels of structural diversity could be acquired by examining non-degraded vegetation remnants in the area being managed.

Horizontal heterogeneity

The importance of heterogeneity in general and patchiness in particular, has received considerable attention in the ecological literature (eg. Pickett & White 1985; Kolasa & Pickett 1991). Patchiness in a landscape not only provides a greater range of habitats which can be occupied by a more diverse fauna, but also provides multiple representation of equivalent patch types. In the event of a disturbance which impacts differently on different patches of the same habitat, the biota in the less affected patches can provide source populations to recolonise the more affected patches (Grubb 1977). Such patchiness must be retained in agricultural landscapes and should be incorporated into reconstructed habitat. Guidance as to the appropriate scale of this patchiness for the area being managed can be obtained by observing the characteristics of existing remnants or by examining patterns in the distribution of soils and landform types.


Two important aspects of configuration include patch proximity and connectivity between patches. Both of these parameters affect the movement of individuals between patches and therefore influence the probability of local extinction of populations and the likelihood of recolonisation if local extinction does occur. This is particularly important in highly fragmented landscapes where sub-populations are subject to extinction as a result of natural catastrophes, human impacts, or stochastic fluctuations in population sizes (Fahrig & Merriam 1985; Newmark 1985; Fahrig & Paloheimo 1988; Pulliam 1988). In general, patches that are closer together and are connected by habitat which allows movement of individuals are more likely to sustain populations of their constituent biota than are more isolated patches (Forman & Godron 1986; Saunders & Hobbs 1991). In general, the distance between patches should be minimised and where possible, the width and structural complexity of vegetation linking remnants should be maximised.


There has been lengthy debate over whether a single large area is better than several small ones for maintaining biological diversity (see for example Margules et al. 1982; Simberloff & Abele 1982; Shaffer & Samson 1985; Margules 1987). This debate centers around the question of whether it is area per se which results in increased species number, or whether increasing area is correlated with greater environmental variability which in turn contributes to greater habitat diversity and hence to higher species diversity. Whatever the cause of this increasing diversity with increasing area, the relationship is not a linear one. The rate at which species accumulate with increasing area tends to decline beyond a particular size. In general, larger areas are likely to contain more species than smaller ones. However, the number of species will also depend upon the patchiness of those areas. This suggests that larger areas of monocultures are likely to contain fewer species than the same area with a number of patch types. It may also be possible for smaller areas made up of a range of patch types to contain more species than a bigger area containing fewer patches. In regions where levels of local endemism are high, it is possible that small areas may contain unique species that do not occur in other larger areas. When protecting or enhancing vegetation for conservation, it is best to aim for bigger rather than smaller areas but also to incorporate heterogeneity into the design where possible. Small areas should be included if they are known to contain species that are poorly represented elsewhere.


The importance of remnant shape depends, to some extent on the size of the patch or remnant being considered. While shape may be less important for large remnants, it can be a critical factor in determining the conservation value of smaller remnants. Long, thin remnants, such as corridors or riparian vegetation, have a much higher edge to interior ratio and hence are more exposed to deleterious edge effects (Diamond 1975; Wilson & Willis 1975). These include increased nest predation and brood parasitism in birds (AndrÚn & Anglestam 1988; Johnson & Temple 1990; Paton 1994), a shift from characteristic tree species to weedy generalists including the invasion of exotic weeds (Noss 1987; Hobbs & Atkins 1988; Scougall et al. 1993) and increased exposure to wind which impacts on vegetation condition (Moen 1974; Grace 1977). Linear habitat can also present problems for species that have to return to a particular location, such as a nest, when foraging. In linear habitat individuals may have to travel much greater distances to meet their food requirements than they would do if those resources were distributed within a more compact patch (Recher et al. 1987; Saunders 1990).

In general, compact areas will be better than linear areas for the provision of habitat and for protection against external impacts from adjoining land uses. In the absence of more precise guidelines, patches should have the least possible edge and linking vegetation should be as wide as is practically feasible (Wilson & Lindenmayer 1996).


An enhancement strategy can also contribute to the protection of remnant vegetation by providing a buffer from external impacts such as strong winds and fluxes of chemicals, fertilisers and weeds. This buffering role may be achieved by planting bands of shrubs and trees around the edges of existing remnants or along the edges of corridors and stream-side vegetation. Not only will these actions help to protect the vegetation but they may provide additional habitat for plants and animals. If the buffering vegetation is part of a commercial timber or fodder enterprise it must be recognised that the buffering role will diminish following harvesting as will the suitability of that vegetation for habitat. Vegetation that is not intended for harvest should be used preferentially in areas adjacent to remnant vegetation. Where ever possible, locally indigenous species should be used and particular care should be taken to ensure that any non-indigenous species used are not potential weeds.

Management within remnants

General enhancement principles can be useful for providing management guidelines where there is a simple relationship between the process to be managed and the conservation response. For example, if increasing levels of grazing have increasing impacts on biodiversity, then it is necessary to reduce grazing intensity in order to increase the conservation value of a remnant. However, many relationships are not so straightforward. Disturbances such as fire can be detrimental at both low and high frequencies or intensities, but may be beneficial at some intermediate level. In addition, the nature of the relationship may differ between ecosystems and even between communities and species within an ecosystem. A precautionary approach for managing a process such as fire, for which the outcomes are unpredictable, may be to ensure that any management regime is highly variable. This could be achieved by varying the inter-fire intervals, the intensity of sequential burns, the season in which fire is initiated, and the area over which fires are allowed to burn. However, without a knowledge of the responses of the biota to fire, there will be no clear basis for specifying the appropriate values for any of these variables. Effective management will therefore require a knowledge of how particular species or communities respond to these processes.

Species composition

Many of the principles described above relate primarily to the spatial and structural attributes of an ecosystem and could, to some extent, be met by using any particular combination of plant species provided that they provided spatially and structurally diverse habitat. However, the use of non-local species will reduce the chances of particular critical resource needs being met and will increase the risk of creating new problems through the introduction of potential weeds. Emphasis should therefore be placed on using locally indigenous plant species.

The limits of general enhancement

Because the above ecological principles provide only general guidelines, it is not possible to specify the magnitude of the response required, or to predict exactly what contribution those guidelines will make to the maintenance of biodiversity. They indicate that remnants should be bigger and more diverse, but cannot indicate how much bigger or how patchy. Similarly, they cannot tell us which species (or even how many species) will be retained or lost. They simply tell us that some designs or management practices are likely to result in the retention of more species than will others. While such an approach clearly has serious limitations, situations will arise where there is an urgent need to act and the requirements of the biota in the area to be managed are not known. Under these circumstances there will be little choice but to use such guidelines.


   Links to another web site
   Opens a pop-up window