The meanings of vegetation condition. David Keith1 and Emma Gorrod1,2 (NSW Department of Environment & Conservation, PO Box 1967, Hurstville, NSW 2220. Tel.: (02) 9585 6498; Fax: (02) 9585 6606; E-mail: email@example.com; 2Botany Department, University of Melbourne, Parkville Vic. 3052, Australia).
Introduction. ‘Vegetation condition’ is a concept that has rapidly gained currency in recent land management policy within Australia. Recent applications of the concept have been developed to assist decision-making for incentive payments, clearing approvals and offset actions in the management of native vegetation. In general terms, ‘condition’ means state of being or health. When applied in biology at the scale of individuals, it refers to fitness – the ability of an individual to survive and reproduce (Begon et al. 1986). Coops et al. (2004) provide an example of such an application in which the health of eucalypt trees is assessed using a spectral reflectance algorithm. At higher levels of organization (communities, ecosystems, landscapes), the ecological meaning of ‘condition’ is less clear. We argue that confusion is at least in part due to historical legacies that have led to independent development and use of the term in several related, but different, ecological concepts. Like ‘rarity’, ‘condition’ has gained currency as one term with many meanings (Harper 1981), which is implicit within much policy and management discourse about vegetation condition (Thackway et al. 2005). In this note, we identify three main meanings of vegetation condition that are derived from the concepts of aesthetics, production and biodiversity. We briefly review these meanings with examples and discuss their interrelationships. We conclude that confusion about what ‘good condition’ means for contemporary management of native vegetation can be reduced with: (i) more explicit explanations accompanying usage of the term; (ii) further development of the condition concept to deal with the multidimensional properties of biodiversity; and (iii) explicit links to more robust models of vegetation dynamics than those currently implicit within assessments.
Aesthetics, production and biodiversity. Aesthetic values derive from human perceptions about what makes ‘a good patch of bush’. These perceptions are, by definition, subjective and likely to vary between people and with landscape context. Nevertheless, scientific approaches to evaluation of natural aesthetic values have been developed for such purposes as heritage listing (UNESCO 2005) and wilderness identification (Lesslie & Maslen 1995). The concept of naturalness is central to both of these applications (Kirkpatrick & Haney 1980), and an area with high naturalness values may be considered aesthetically to be in ‘good condition’. An assessment methodology proposed by Lesslie et al. (1988) considers two forms of naturalness that are inversely related to the impact of human activities. One of these (aesthetic naturalness) is dependent on the presence of modern human structures, such as fences, tracks, dams, buildings, etc., and the other (biophysical naturalness) is dependent on the history and intensity of human land uses or their effects, such as logging, grazing, weed invasion, etc. These attributes may be mapped using distance functions, remote sensing, tenure data and historical information, which may be synthesized in GIS applications to calculate naturalness values across a landscape (e.g. Lesslie et al. 1988). Although these assessment methodologies were developed for landscapes containing large natural areas, some of the principles are also applicable to finer scales. Relevant indicators of human disturbance, such as cut stumps, fences, tracks, soil compaction, etc., may be used to assess the naturalness of remnant patches. More generally, comparative approaches may be used to assess the aesthetic values of native vegetation by comparing particular sites with reference examples or ‘benchmarks’ that define one or more levels of aesthetic value.
Production values derive from the ability of native vegetation to deliver resources for human consumption. A patch of native vegetation may be considered to be in ‘good condition’ if it maintains a high capacity to produce resources such as livestock, timber, water, and other ‘ecosystem services’ (Heal 2000). The scientific basis for assessing these values derives from a long tradition in agriculture, particularly for the assessment of pasture condition. Landscape Function Analysis (Ludwig & Tongway 1992) provides an example of condition assessment with a strong theoretical basis and well-developed practical methodologies. The approach was originally developed in arid and semiarid rangelands used as native pastures for livestock production. Its central concept is based on the capacity of land to retain resources that are essential for plant growth, particularly water, soils and their nutrients. Assessments are focused on surface features (such as cover of perennial plants, soil lichens, leaf litter and woody debris) that interrupt, divert or absorb runoff and transported materials, and hence reduce loss of resources. Sites and landscapes with an abundance of such features are in ‘good condition’, able to sustain production of fodder and are relatively robust to degradation during drought or heavy rain. Landscape function, hence condition, may be degraded by overgrazing, which reduces the abundance of critical surface features and the capacity of the land to retain resources and sustain livestock. Assessments of rangeland condition therefore provide important guidance for sustainable management of livestock grazing. Methodologies are available for site-level assessments (Ludwig & Tongway 1992) and landscape scales using satellite image analysis and modelling of grazing halos around watering points (Bastin & Ludwig 2006).
Biodiversity values in part relate to the capacity of native vegetation to sustain local populations of native plants and animals (as well as their genetic diversity and ecological interactions). ‘Good condition’ in this sense is related to high carrying capacity from population theory or high ‘habitat quality’ (Begon et al. 1986). However, unlike these species-specific parameters, the concept of condition for biodiversity applies to all species that may be reasonably expected to use a site (Parkes et al. 2003). Because the habitat requirements of all species are not known, assessment methods are based on surrogates that represent habitat features for some better-known groups of species (e.g. density of large trees), as well as features that, conversely, are indicators of habitat degradation for certain groups of species (e.g. weed abundance). Available site assessment methods (e.g. Gibbons & Freudenberger 2006) differ in their weightings for different attributes, as well as the mathematical details of index calculation. These variations could influence the performance of alternative methods for different biotic groups and habitats. Recent approaches incorporate reference or benchmark values in the assessment to standardize assessments across different types of habitat. Landscape-scale approaches for assessing the biodiversity component of vegetation condition based on remote sensing and spatial modelling are currently under development (e.g. Newell et al. 2006).
Confusion in the usage of ‘condition’. Each of the three main facets of vegetation condition share commonalities. For example, in forest and woodland vegetation, large trees contribute to aesthetic value through visual splendour; they contribute to production value if they represent a timber resource and also to biodiversity value by providing habitat for a variety of organisms. An abundance of large trees should therefore be an indicator of good condition across a range of value types. On the other hand, each facet of condition has unique elements that are not shared by others. Tree hollows of a particular size and configuration confer biodiversity value through provision of habitat for particular tree-dwelling vertebrates, but might be negatively related to production value if they render the timber unusable. Thus, any given patch in a landscape may have any combination of high and low values for each of the three components (Fig. 1).
Confusion may emerge when different meanings of condition are applied to the values represented in Figure 1. For example, some may interpret a site as being in ‘good condition’ only if it contains high values for each of the aesthetic, production and biodiversity components. Others may interpret a site as being in ‘good condition’ if it contains high values for any one of the three components. Still others may interpret a site as being in ‘good condition’ if it contains high values for one particular component (e.g. aesthetic), irrespective of its values for other components (production and biodiversity). Further scope for confusion emerges when one considers that aesthetic, production and biodiversity values are each comprised of a number of components that may not be well correlated with one another. These simple examples illustrate what Regan et al. (2002) defined as ambiguity – a form of uncertainty in which a term may have more than one meaning. We contend that ambiguity, along with underspecificity, context-dependence, vagueness and other forms of linguistic uncertainty (Regan et al. 2002), are pervasive in both the assessment and planning dialogue for vegetation condition. If the condition concept is to be used effectively in decision support, linguistic uncertainty needs to be reduced by articulating explicitly the scope, context and meaning of ‘condition’ in each application.
Beyond semantics: Conceptual difficulties and further development. Although more disciplined usage would benefit practical applications of the term, more fundamental problems stem from the efficacy of ‘condition’ as a model of vegetation state and change. At the heart of its use in policy and management is the presumption that vegetation condition represents a generalized measure of ecosystem function and habitat suitability for native biota, which varies along a dynamic continuum of degradation and restoration. The underlying model of change is implicitly Clementsian (Begon et al. 1986), and has been subject to little scientific scrutiny.
One of the main difficulties in applying a concept of vegetation condition stems from its univariate property – condition varies on a single scale from ‘good’ to ‘poor’. The univariate property of condition does not sit well with its common interpretation, which demands generalization across multiple values that may not be well correlated with one another. This is especially evident for biodiversity, which comprises many species with habitat requirements that may be poorly correlated or inversely correlated with one another – what may be good habitat for one species may be poor for another. The possible reduction of suitable habitat (open grassland) for the endangered Plains Wanderer (Pedionomus torquatus) in response to exclusion of grazing required to encourage regeneration of Myall (Acacia pendula) Woodland, also endangered, is a case in point (NPWS 2002). In general, the nature of these correlations within habitat types is poorly understood, but is crucial to understanding the dynamics of degradation, regeneration and restoration in vegetation.
Further theoretical development of the condition concept is needed to overcome these practical difficulties. In particular, vegetation condition for biodiversity is in need of a theory of dynamics that unifies its components and defines ecological processes that mediate vegetation change. It is noteworthy that applications of the condition concept that have a strong theoretical basis, including the Eucalypt Canopy Condition Index (Coops et al. 2004) and Landscape Function Analysis (Ludwig & Tongway 1992), deal with established causal links between indicators and a univariate response (leaf productivity and soil productivity, respectively). Existing models of succession and diversity (Connell & Slatyer 1977; Westoby et al. 1989; Silvertown 2004) provide some guidance for future development of the condition concept and represent a considerable advance on simplistic Clementsian thinking that underpins current policies aimed at improving biodiversity values across landscapes.