A test of landscape function theory in the semi-arid shrublands of Western Australia

Australia’s rangelands encompass approximately 80% of the continent and generate significant wealth through a range of industries. The rangelands comprise four major ecosystem types, these are: grasslands, shrublands, woodlands and savanna. The ecological legacy of early pastoral development in most of Australia‟s semi-arid shrublands is largely one of degradation and desertification (Wilcox and McKinnon, 1974; Curry et al. 1994; McKeon et al. 2004; Mabbutt et al. 1963; Pringle and Tinley, 2001). Since the 1980‟s, there has been a slow and general shift by the pastoral industry towards sustainable stocking rates (Watson et al. 2007; Pringle and Tinley, 2001).To implement grazing systems that better align stocking rates with carrying capacity in the semi-arid shrublands, pastoralists require a much more advanced understanding of patch patterning and ecological processes at a paddock scale. This understanding of theory could improve the management decisions made by pastoralists and other land managers (e.g. mining environmental officers, carbon offset developers, conservation park rangers, Indigenous communities) and assist them in their immediate challenge of cost-effective rehabilitation of degraded areas. Landscape function theory was largely developed for this purpose.Landscape function theory was developed as a way to assess and interpret patch patterns and ecological processes that occur at a range of spatial scales (Ludwig et al. 1997). Landscape function theory is used to explain the concept of „functional heterogeneity‟ which is an information-rich phenomenon that has enabled the development of cross-scale metrics. Landscape function theory is based on four primary principles, these are: 1. Patchiness can be characterised by patch size, orientation, spacing and soil surface condition. 2. Natural landscapes have a characteristic spatial self-organisation, often expressed as patchiness. 3. Deviations from the „characteristics‟ or „natural‟ patchiness are seen as degrees of dysfunctionality and there is a long continuum from highly functional to highly dysfunctional patches. 4. Restoration or replacement of missing or ineffective processes in the landscape will improve soil surface conditions and soil habitat quality.Landscape function theory and the associated landscape function analysis (LFA) methodology have become an accepted standard for the ecological assessment of rangeland environments. However, there have been a limited number of studies in Australia that have thoroughly examined the four principles that underpin the landscape function theory. Further interrogation of the principles that underpin the theory has the potential to enhance its utility and validate its assumptions. Therefore, the aim of this thesis was to test the four principles of landscape function theory in a semi-arid shrubland environment.The overarching hypothesis of this thesis was that clear spatial patch patterns occur at a range of scales within the case study paddocks and these patterns determine the ecological functionality and resilience of the area. Specifically, I sought to find evidence of distinct patch-interpatch patterns and associations at a range of scales in the lower Murchison region of Western Australia (WA) and examined the impact that grazing and seasonal conditions can have on this natural patchiness. The study involved four components: 1. Empirical quantification of patch-scale heterogeneity and investigation into the existence of a gradient of functionality at a patch-scale. 2. Measurement of patch patterning at a paddock scale and an evaluation of sequences and associations of patch-interpatches in the context of their contribution to landscape functionality. 3. Investigation of the impact of grazing and seasonal conditions on perennial grass populations and the potential impacts of this on patch patterning and ecological processes. 4. Investigation of livestock grazing behaviour and its implications for patch patterning and ecological processes.Significant variation was detected (P45%) (P70% more rapid) (P85% of the area) compared to the other land-types. In contrast, Acacia sandplains had a higher proportion of the ecologically functional patches compared to granitic shrublands. Depending on the land-type, it was concluded that a functional ecosystem will generally have balanced proportions of particular patch-interpatch subclasses as this is likely to lead to the efficient capture and cycling of water and nutrients. In contrast, a degraded ecosystem will have higher proportions of ecologically dysfunctional patch-interpatch subclasses. A number of significant spatial associations and repeating sequences of individual patch-interpatch subclasses were found to occur in the study area. In the case study paddocks, distance-from-water did not have any significant effect on the relative proportions of the different patch-interpatch subclasses.Native perennial grasses support perennial shrub patches in maintaining ecological stability and landscape processes in semi-arid shrublands. Perennial grasses were monitored over a 12 month period to determine whether summer rest from Merino sheep grazing had an impact on perennial grass populations. The change in the number of perennial grasses over the 12 month period was variable ranging from a 36.7% increase to mortality rates as high as 80.2%. There was evidence to suggest that the grasses in the exclosure treatments (i.e. no grazing) may have been conditioned by previous grazing events in such a way that made them highly susceptible to extended periods of moisture stress. Land-type, distance-from-water and stocking rate did not have a significant impact on the change in the number of perennial grasses over the 12 month period. Resting perennial grasses during the summer may be of benefit; however, given the high mortality rates in most of the paddocks, it was clear that seasonal conditions had a substantial overriding influence on the outcome.Five major factors influenced sheep grazing behaviour in the study paddock; these include: land-type, greenness cover, time of day, air temperature and distance-from-water. The GPS-collared sheep demonstrated a clear grazing preference for the higher value land-types (i.e. alluvial plains and hardpan washplains). The results also indicated that the sheep were selectively grazing areas with high greenness cover (as expressed by Normalised Difference Vegetation Index [NDVI]). The results highlight the importance of understanding the selective grazing characteristics of Merino sheep when designing paddock layouts and developing management plans. This is because repetitive, selective grazing of high-value land-types has the potential to modify the natural patch patterning and ecological processes.This study found empirical evidence which substantiates landscape function theory and concludes that it has significant application at a range of spatial scales in the semi-arid shrublands. The results explain why landscape function theory is the current standard for the assessment and interpretation of rangeland environments in Australia. Landscape function theory is the result of an evolutionary synthesis of earlier work, and it continues to be progressively refined as our knowledge of the complex patterns and ecological processes of the rangelands is improved.Based on the findings of this study, I consider that there are three areas of further inquiry which could assist land managers to practically apply landscape function theory and contribute towards the restoration of degraded Australian rangelands, these are: determination of the „functional‟ composition of patch-interpatches subclasses for a range of land-types in the Australian rangelands that could be used as benchmarks for land managers; further empirical quantification of the ecological processes that drive landscape function to assist in simulation modeling and the estimation of cost / benefit of rehabilitation efforts; and the potential role that grazing management can play in the restoration of dysfunctional patch-interpatch patterns and ecological processes in the semi-arid shrublands