Evaluating alternate response models of endangered box gum grassy woodland following livestock exclusion
(Part One)
The following article (part one) is the first of three parts, which make up my Master of Philosophy Thesis. Part One forms an overview of the thesis followed by some background reading. Part Two and Three detail two different studies looking at woodland responses to livestock grazing exclusion.
Abstract
Temperate grassy ecosystems are one of the most threatened terrestrial ecosystems in Australia. Since European settlement grassy ecosystems have been cleared and modified by farming practices, which continue today. Many restoration projects have been initiated using fencing as a first-step or stand-alone management tool in a bid to preserve and restore species composition and ecosystem function. Despite this restoration activity, few studies evaluate how sites respond following livestock exclusion and results are equivocal. This study aims to fill important knowledge gaps on the effectiveness of fencing as a tool in restoring temperate grassy woodlands of NSW.
This space-for-time comparative study uses cross fence-line contrasts of paired fenced and unfenced plots spanning a chronosequence of 1-17 years of livestock exclusion to evaluate how well empirical models predict vegetation response. Species richness and abundances were fitted as a function of fencing age using simple linear regression models. Support for the ‘null model’ was found. Under this model the grazed system shows limited recovery after fencing to exclude livestock grazing and no consistent trend was found in species richness, abundance or composition with fencing age.
The same paired sites were explicitly matched to nearby reference woodlands to ascertain whether livestock exclusion initiates convergence towards reference woodland states. Livestock exclusion resulted in weak compositional convergence towards reference woodlands, but this was not explained by livestock exclusion time. Topsoil chemistry properties also did not converge towards reference states with livestock-exclusion-time. Degree of degradation of initial state did not explain the degree of compositional divergence of fenced from unfenced plots. Initial state also did not explain soil condition divergence of fenced from unfenced and convergence of fenced towards reference states.
The findings highlight the importance of site selection. Thus, knowing site history and management practices when selecting degraded woodlands as restoration sites is crucial. Merely relying on time after fencing to restore degraded woodlands appears inadequate, as fencing alone appears to be insufficient to restore woodland communities floristically and environmentally for at least the first 17 years of exclusion. To improve fencing outcomes supplementary management such as re-introducing propagules of lost species, weed control and soil amelioration need to be employed additionally to livestock exclusion.
Introduction
Despite the extensive work documenting how communities and ecosystems change in response to disturbance, there is no agreed general conceptual framework concerning ecosystem development (Suding and Hobbs 2009). Over time, models to explain community dynamics have evolved from linear succession (Clements 1936) to non-linear, state and transition frameworks (Westoby et al. 1989) and community assembly theory (Keddy 1992).
Early equilibrium models suggest that grazing disturbance counterbalances natural successional process towards a climax community (Westoby et al. 1989). They predict that after removal of the degrading disturbance (i.e. livestock grazing), the system returns to its climax condition via successional processes (Westoby et al. 1989; Whalley 1994). These equilibrium models may be expected to be good predictors of change for ecosystems that have evolved with the disturbing pressure (i.e. heavy-hooved herbivores in African savannahs (Milchunas et al. 1988)). The evolution of these environments has been closely shaped by herbivore grazing and is considered essential in maintaining species diversity (McNaughton 1985). However, equilibrium models may not perform well for ecosystems that are exposed to new drivers (i.e. that they have not evolved with), such as large ungulate herbivores introduced into Australian grassy ecosystems.
State and Transition Models (STM) accommodate non-equilibrium dynamics. Originating and primarily applied in rangelands (Westoby et al. 1989), STMs use discrete stable states with transitions between these states to describe changes in vegetation and abiotic conditions. The transitions may be conditional on external processes or events. STM concepts build on equilibrium models by acknowledging the potential for multiple alternative trajectories. STMs allow for the possibility that removal of herbivores from overgrazed rangelands may not trigger recovery to a ‘climax’ state. One such model was constructed by Spooner & Allcock (2006) for Australian temperate grassy ecosystems. It predicts that once ecosystems cross particular ecological thresholds (i.e. defined by losses of propagules and increased nutrients and introduction/invasion of exotic species), a return to the pre-disturbance state is unlikely to occur unassisted. Thus, if overgrazing on a landscape level outlasts propagule dormancy, extirpation of grazing-sensitive plants that once characterised the intact state may be lost from a region.
Community assembly theory is an emerging advance in restoration ecology that suggests the composition of any local community (particularly after disturbance) is a product of filtering an available species pool (Funk et al. 2008; Keddy 1992). The theory is based on a series of selective filters that exclude plant species traits that are incompatible with local conditions. Filters include dispersal, abiotic and biotic factors. Dispersal is a function of species vagility. That is, species that can actually disperse to the site from a regional species gene pool. If a newcomer establishes at a site, then the species’ tolerance to local physical conditions and interactions with competitors, predators and pathogens will determine its persistence. Plant traits that navigate these filters ultimately compose the extant community.
Despite the effects of cattle grazing on grassy ecosystems in temperate Australia being well documented, less is known about the potential of ecosystems to recover after a long history of grazing (Lunt et al. 2007b). Seed availability, germination and survival of regenerating plants may determine ecosystem composition after fencing. Despite this knowledge, there is a current lack of understanding on how these determinants are affected by site conditions and management (Vesk and Dorrough 2006). This is largely because the knowledge acquired from such studies are site specific (Vesk and Dorrough 2006), contain different biota and potential barriers (Lunt and Spooner 2005), result in outcomes that are difficult to predict (Yates and Hobbs 1997b) and thus, cannot be extrapolated over different environs.
Evidently, there is a need for more research to fill knowledge gaps on the variability of vegetation response across environmental gradients. This is particularly important for threatened ecosystems, such as temperate grassy woodlands which span multiple climates of south-eastern Australia.
Among the most important knowledge gaps are whether livestock exclusion initiates woodland restoration, if so how rapidly the responses occur, whether responses are conditional on management history and environmental factors, and how well alternative models predict the responses. The aim of this study is to improve current understanding of temperate grassy woodland regeneration on these points. Specifically, the following hypotheses were tested:
- livestock exclusion promotes colonisation by native plant species and decline of exotic plants species’;
- species composition and soil conditions of fenced sites diverge from those of unfenced sites over time-since-livestock-exclusion;
- species composition and soil conditions of fenced sites converge with those of reference sites over time-since-livestock-exclusion; and
- responses to fencing depend on time-since-exclusion of livestock grazing, degree of degradation of the initial state and/or environmental conditions.
Part Two describes a livestock exclusion study undertaken in New South Wales (NSW) to evaluate four alternative models derived from the literature on the response Australian grassy ecosystems to grazing. Table 1 summarises findings from prior studies on livestock exclusion. Although productivity and time have been shown to be poor predictors of woodland recovery after livestock grazing, the Recovery Model (model 1), broadly describes the response of low productivity woodlands and the Competitive Exclusion Model (model 2) broadly describes recovery of high productivity woodlands (Part Three). Other key variables are likely to impact recovery, such as the intensity and duration of grazing disturbance, which will influence the degradation state of woodlands prior to fencing (e.g., Prober et al. 2011). Disturbance metrics can be estimated by stocking rates and time under livestock grazing, but these data are not always available. Therefore, these variables were not included in the response models.
Part Three explores the impact of grazing disturbance on recovery by including the initial state, approximated by the state of the unfenced plot of each (fenced-unfenced) pair, as an explanatory variable in the linear mixed models.
Table 1 Recovery of Australian eucalypt woodlands and grasslands after fencing to exclude grazing livestock. Woodlands had been grazed for hundreds of years before fencing. These data were used to define the response models tested in Chapter 2. For a full description of the models please refer to Chapter 2. SFT = space for time.
*Woodlands are grouped by productivity, which is determined by mean annual rainfall (MAR) in millimetres and soil nutrient availability. The productivity of my study sites is towards the higher end of the spectrum represented here (i.e. high-nutrient soils and MAR varies from 600 to 800 mm).
The intent of this thesis was driven by my own curiosity as a botanist. As consultants, we are expected to make judgements on the regeneration potential of degraded ecosystems. There is very little empirical data that can be used to predict vegetation response following relief from livestock grazing pressures, and what does exist is extremely site-specific. This research aims to increase the success and reduce costs associated with woodland rehabilitation. It aims to inform site selection and design of management strategies by identifying the types of sites associated with the lowest risks of regeneration failure and by diagnosing potential ecological barriers responsible for slow woodland regeneration or undesirable trajectories. Ultimately, the outcomes could also serve to guide government policy on the management and conservation of temperate woodlands.
Read More: Part Two
Literature Cited
Clements, F.E., 1936. Nature and structure of the climax. Journal of Ecology 24 (1), 252-284.
Funk, J.L., Cleland, E.E., Suding, K.N., Zavaleta, E.S., 2008. Restoration through reassembly: plant traits and invasion resistance. Trends in Ecology & Evolution 23 (12), 695-703.
Keddy, P.A., 1992. Assembly and response rules: two goals for predictive community ecology. Journal of Vegetation Science 3 (2), 157-164.
Lunt, I.D., Jansen, A., Binns, D.L., Kenny, S.A., 2007. Long-term effects of exclusion of grazing stock on degraded herbaceous plant communities in a riparian Eucalyptus camaldulensis forest in south-eastern Australia. Austral Ecology 32 (8), 937-949.
Lunt, I.D., Spooner, P.G., 2005. Using historical ecology to understand patterns of biodiversity in fragmented agricultural landscapes. Journal of Biogeography 32 (11), 1859-1873.
McNaughton, S.J., 1985. Ecology of a Grazing Ecosystem: The Serengeti. Ecological Monographs 55 (3), 260-294.
Milchunas, D.G., Sala, O.E., Laurenroth, W.K., 1988. A generalized model of the effects of grazing by large herbivores on grassland community structure. The American Naturalist 132 (1), 87-106.
Suding, K., Hobbs, R., 2009. Models of ecosystem dynamics as frameworks for restoration ecology, In New models for ecosystem dynamics and restoration. (eds K. Suding, R. Hobbs), pp. 3-22. Isand Press, Washington, DC, USA.
Vesk, P.A., Dorrough, J.W., 2006. Getting trees on farms the easy way? Lessons from a model of eucalypt regeneration on pastures. Australian Journal of Botany 54 (6), 509-519.
Westoby, M., Walker, B., Meir, I., 1989. Opportunistic Management for Rangelands Not at Equilibrium. Journal of Range Management 42 (4).
Whalley, R.D.B., 1994. State and transition models for rangelands. 1. Successional theory and vegetation change. Tropical Grasslands 28 195–205.
Yates, C.J., Hobbs, R.J., 1997. Woodland restoration in the Western Australian wheatbelt: a conceptual framework using a state and transition model. Restoration Ecology 5 (1), 28-35.