This thesis considers ways to improve biodiversity conservation at the bioregional level in Australia through the use of geospatial science technologies and biological modelling techniques. Following a review of approaches to biodiversity conservation at the bioregional level, including the roles and potential of geospatial science technologies in this regard, I consider biodiversity modelling using a case study of the Burt Plain bioregion in central Australia that focuses on selected taxa, ecosystems and landscapes. The Burt Plain bioregion was chosen since it is one of 19 bioregions nationally that has been given a 'very high' priority status for biological survey, assessment and potential reservation of land for conservation purposes. The specific research objectives for the Burt Plain bioregion study were to: · describe the species composition, distribution and nature of the dominant vegetation communities within the bioregion; · characterise environmental niche of communities with respect to selected environmental and management variables - latitude, longitude, climate, land systems and land units, geology, hydrography, topography, and tenure; · analyse how well or otherwise taxa have been sampled (during previous ground surveys) with respect to geographical and environmental variables; Biodiversity conservation at the bioregional level · develop and compare quantitative habitat models of the potential distribution of selected species based on presence-only distributional data; and · examine the significance of radiometric data as a potential correlate and predictor of the distribution of those selected species. National conservation initiatives such as the bioregional approach and international initiatives such as the biosphere reserves program to support the planning and management of biodiversity conservation are discussed in chapter two. The scientific and related processes underpinning the development of bioregions and strategies across the Australian states and territories are then considered. An important finding arising from this review is the need to improve biological information, especially through systematic surveys and on-going monitoring of ecosystems and populations of species, at the bioregional level to inform land use allocation and management. This finding is consistent with one of the general aims of the thesis to improve the spatial modelling techniques available for bioregional assessment and biodiversity conservation. In chapter three I review the role and limitations of geospatial technologies currently employed for biodiversity conservation management. Current developments and applications of GIS and remote sensing to wildlife research, conservation gap analysis and conservation reserve design are considered. Geographic information systems (GIS) are now routinely used by ecologists to Biodiversity conservation at the bioregional level analyse spatial data. Although various forms of GIS have been available for 15 to 25 years, the biological applications of GIS have figured most prominently in the ecological literature only in the past 15 years. The use of computer-generated models to simulate environmental events can provide a greater understanding of ecosystems, and offers improved predictive powers to conservation and land managers. The decision support offered by computer-based modelling techniques appears likely to underpin conservation and management decisions much more into the future providing that adequate biological and other datasets are available for this purpose. Dominant vegetation communities and various environmental gradients were analysed to characterise environmental niches at the bioregional scale for the Burt Plain bioregion (Chapter 4) and more locally at the catchment scale for the Upper Todd River Catchment (Chapter 5). In Chapters four and five I describe in detail the land tenure and use, land systems, climate soil, geology, topography, hydrology, vegetation and biodiversity of the Burt Plain bioregion and Upper Todd River Catchment. The bioregion contains some ephemeral watercourses, which are generally in fair to good condition, but are afforded little protection from a range of threatening processes, including grazing and trampling by feral animals and livestock and weed infestation. The major river systems occurring in the bioregion include parts of the Plenty, Hanson, Sandover and Lander Rivers. In the Upper Todd River Catchment the major watercourses Biodiversity conservation at the bioregional level are the Todd River and Station Creek, which exit the area via two narrow gaps in the low rocky hills on the southern boundary of the bioregion. The dominant geology can be summarised as plains and low rocky ranges of Pre-Cambrian granites on red earths. The bioregion has approximately 200 - 250 mm of summer rainfall, with rainfall occurring on 20 - 30 days per year. There is a high variability and range of temperatures, with an annual mean temperature of approximately 22-23°C. In Chapter six I consider a range of species found within the Burt Plain bioregion using existing survey data and techniques that enables the prediction of the spatial distribution of taxa. Using GLM and GAM models, Black-footed Rock- Wallaby (Petrogale lateralis), Spinifex Hopping Mouse (Notomys alexis) and Spencers Frog (Limnodynastes spenceri) were chosen for a more in-depth analysis. Environmental variables correlated with the presence of each species are then described and prediction maps showing the probability or likelihood of the presence of the species within the bioregion developed. In Chapter seven I examine the utility of radiometric data for wildlife habitat modelling. Statistical relationships are tested between the concentrations of the elements uranium, thorium and potassium and terrain characteristics such as position in the landscape, slope and aspect as well as other climatic variables. Radiometric data were found to be useful for developing statistical predictive Biodiversity conservation at the bioregional level models of six species: Red Kangaroo (Macropus rufus), Desert Dunnart (Sminthopsis youngsoni), Rabbit (Orcytolagus cuniculus), Brown Honeyeater (Lichmera indistincta), Little Spotted Snake (Suda punctata) and Southern Boobook (Ninox novaeseelandiae). I suggest that the utility of radiometric data for wildlife habitat modelling would appear significant and should be explored further using alternative quantitative modelling techniques and presence/ absence records for target faunal species. Predictions of species distributions may be useful for prioritising land acquisitions for reservation as well as in the future design of biological surveys. The thesis concludes with a synthesis of the major research findings, discussion of the limitations of the datasets available for the study, perspectives on management issues in the Burt Plain bioregion, and possible future research directions. It is important that purposefully-designed biological survey research be undertaken across the bioregions of the arid zone of Australia to enhance basic understanding of biodiversity patterns and their relationships to environmental heterogeneity and site-landscape level processes. Geospatial modelling techniques can assist such biodiversity survey and evaluation and make their conduct more cost-efficient and the inferences drawn from subsequent data analyses more powerful. This knowledge is required to contribute to the emergent concepts and theory of ecosystem dynamics and associated biodiversity patterns in arid Australia and, most significantly, to enhance the conservation and management of the unique biological complement and systems found in this region.
| Identifer | oai:union.ndltd.org:ADTP/210336 |
| Date | January 2006 |
| Creators | Pert, Petina Lesley, petina.pert@bigpond.com |
| Publisher | RMIT University. Mathematical and Geospatial Sciences |
| Source Sets | Australiasian Digital Theses Program |
| Language | English |
| Detected Language | English |
| Rights | http://www.rmit.edu.au/help/disclaimer, Copyright Petina Lesley Pert |
Page generated in 0.0022 seconds