Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Doctorat → Australie → 2001 → Biogeography and conservation of Mitchell grasslands of Northern Australia

Charles Darwin University (2001)

Biogeography and conservation of Mitchell grasslands of Northern Australia

Fisher, Alaric

Titre : Biogeography and conservation of Mitchell grasslands of Northern Australia

Auteur : Fisher, Alaric

Université de soutenance : Charles Darwin University

Grade : Doctor of Philosophy (PhD) 2001

Résumé partiel
Maintenance of biodiversity in many Australian landscapes will be best achieved by the integration of nature conservation and sustainable land management practices, in addition to formal systems of conservation reserves. Implementation of off-reserve conservation management requires an adequate understanding of ecological values and processes and the biodiversity costs and benefits of prevailing land use practices. This study is concerned with the Astrebla (Mitchell grass) tussock grasslands and related sparse woodlands and shrublands that occur on heavy clay soils in arid and semi-arid northern Australia. Covering an area of approximately 420 OOOkm2, they are one of the most important of the Australian rangelands and virtually their entire extent is given over to grazing of cattle and sheep on native pastures. Consequently, Mitchell grasslands are poorly represented in the national reserve system and there is a lack of ecological knowledge, particularly regarding the fauna, to inform decisions about conservation management. The aim of this study was to address this knowledge gap, in order to determine the most effective approach to conservation management in this environment. Biogeographic context To assess the distinctiveness of Mitchell grassland communities at a sub-continental scale, biogeographic regionalisations of the Northern Territory were developed by classification of 0.5° x 0.5° cells according to their plant and vertebrate composition. The cells corresponding to Mitchell grasslands on the Barkly Tableland formed a distinct zone in regionalisations for vascular plants, all vertebrates taxa and some vertebrate groups (birds, landbirds and reptiles), even at coarse levels of classification (5-6 groups) within the Northern Territory. The area of Mitchell grasslands in the western Northern Territory was less sharply defined as a biogeographic zone. Large geocoded databases for plant and vertebrate records in the Northern Territory were also analysed to identify species with a high fidelity to Mitchell grassland communities. 61 plant species and 17 vertebrate species had more than 50% of their records from Mitchell grassland communities. One feature of the Mitchell grasslands is a number of endemic reptile species, including one discovered during this study. Vegetation and fauna A systematic biological survey of Mitchell grassland communities in the Northern Territory was undertaken to investigate the distribution and environmental relations of the vegetation and fauna. Vascular plants, vertebrates and ants were sampled at 107 sites from 12 locations, selected to sample the environmental variation and geographic range of Mitchell grassland communities in the Northern Territory. For comparative purposes, a small number of sites were also sampled in woodlands on non-clay soils. Classification and ordination were used to examine patterns of species composition and generalised linear modelling was used to relate summary richness and abundance variables, as well as the abundance of individual species, to environmental variation. The biological survey confirmed that Mitchell grassland communities support a distinctive flora, vertebrate fauna and ant fauna which, rather than comprising merely a subset of a widely distributed semi-arid zone species, has a significant component of relatively habitat-specific species. For each taxon, there were a small set of species that were frequent and abundant in the study region and a large proportion of species that were recorded very infrequently. The vertebrate and ant faunas are notable for a very low local and regional species richness. The ant fauna also differed markedly from that reported from other Australian rangelands in terms of biogeographic profile and functional group composition, with an unusually high proportion of ant abundance made up of Opportunist species. There were some commonalities between the distribution patterns of plant, vertebrate and . ant species within the Mitchell grasslands communities. At a regional scale, there was a gradual turnover in species composition in response to a latitudinal climate gradient, with the intrusion of Torresian species into the northern, high-rainfall areas, and Eyrean species into the southern margins of the biome. At a local scale, species composition was influenced by variation in substrate characteristics. Gravel rises that occur patchily throughout the clay plain support a distinct assemblage of species, as do woodlands and shrublands associated with run-on areas. Vertebrate richness and composition is also strongly determined by vegetation structure. Within each major taxon, the relative abundance of many species was also related to the relative greenness (NDVI) score for the sample site, which was developed as a measure of the influence of temporal variation in seasonal rainfall. A notable feature of the distribution patterns of plants and ants was the strong ’location effect’, whereby compositional similarity was greatest between sites within a location. For ants and plants, compositional similarity between sites also decreased with geographic distance, independently of quantified environmental differences. A minimum-set selection algorithm was used to investigate the implications of the distribution patterns of plants, ants and vertebrates for conservation management. The majority of Mitchell grassland species could be represented in a small number of selected sites. However, because there were a high proportion of singleton or very infrequent species, a large number of sites were required to achieve reservation targets of 90-100% of species. Reservation of a high proportion of species is more likely to be achieved by a number of small reserves with a wide geographic distribution than one or few large reserves. No single taxon (ants, plants, vertebrates) can be safely used as a surrogate for the other taxa in designing the most efficient reserve network, as there is only limited assemblage fidelity between taxa. Waterpoint and grazing distribution In many arid rangelands, the distribution of grazing activity is determined by the location of waterpoints. Measurement of cattle dung, footprints and grazing at the sample sites were combined into an index of recent cattle activity, which was strongly related to distance from water. Artificial waterpoints are regularly distributed throughout the extent of the Mitchell grassland communities in the Northern Territory. Spatial analysis showed that 52% of the total area of these communities on the Barkly Tableland are within 4km of water and 97.4% within 10km of water


Version intégrale (17,2 Mb)

Page publiée le 18 novembre 2019