No description available.

The SST bias in the tropical Atlantic exists in the early to latest generation of coupled general circulation models. The maximum bias is not on the equator but at 16°S, the cause of which has not been thoroughly studied. Newly released CMIP5 models provide a useful tool to investigate the contributions of different physical processes to the SST bias in this area in the coupled system. We tested three existing mechanisms and found that: 1) there is no significant relationship between the SST bias and surface heat flux bias; 2) deficient coastal upwelling is a contributing but not the sole source of the bias; and 3) the SST bias is correlated with temperature biases in the upstream equatorial region. The Angola-Benguela front is displaced southward by more than 10° in latitude in many CIMP5 models. Due to the huge temperature contrasts on two sides of the front, such a frontal displacement generates a very strong SST bias. The correlation between the SST bias and frontal location error in this region is significant at the 99% level, demonstrating that the SST bias in coupled GCMs is attributable to the models’ inability to reproduce a realistic position of the front and the consequent erroneous advection by the southward Angola current. This is due to both errors in the simulated surface wind field and systematic errors in ocean models. Ocean reanalysis datasets and a high-resolution regional model simulation suffer a similar pattern of SST biases. Although they produce a more realistic ocean circulation than coarser resolution simulations and alleviate some of the severe SST bias near the front, a warm bias overlies on a northward current to the south of the front, which actually comes from the north of the front through a subsurface passage. We identify a strong subsurface temperature bias caused by a too-deep and diffused simulated thermocline along the coast of Angola, originating from the equatorial thermocline, advected by the Angola Current and an undercurrent beneath the Benguela current, and then brought to the surface by the coastal upwelling along the Benguela coast, contributing to the warm SST bias south of the front.

Planet Earth is undergoing a rapid rate of ecosystem conversion and degradation and one of the major challenges of current environmental science is to contribute to the management and conservation of biodiversity through the development of tools for assessing environmental change. The main goal of this doctoral dissertation is to contribute to the scientific literature on remote sensing tools for monitoring tropical dry forests, which is one of most important global change frontiers. This thesis is composed of five chapters which have the goals of covering the following specific goals: 1) To estimate the extent and geographic distribution of the neotropical dry forest. 2) To evaluate the potential use of satellite-detected fires as deforestation predictors in tropical dry forest and 3) To evaluate the potential of remote sensing techniques to detect edge effects in tropical dry forest. Before assessing the main goals of the thesis, in chapter two, Integrating Remote Sensing and Biodiversity research, we stress out the necessity of integrated assessments using multiple spatial and spectral resolution sensors over a wide array of ecosystems in order to find relevant ecosystem properties that would be sensitive to species richness. Chapter three, Extent and Conservation of tropical dry forests in the Americas, describes a regional scale mapping effort using coarse-scale imagery (MODIS 500-m) of the extent and geographical distribution of tropical dry forests that introduces several innovations to previous assessments. Based on these techniques, the total current extent of tropical dry forest in the Americas is 519,597 Km2. I also found that 66% of the ecosystem has been already converted to other land uses while only 4.5 % of is under protected areas. Chapter four, MODIS Active fires and deforestation in tropical dry forest landscapes, we show correlations patterns between the number of MODIS Active Fires and forest cover change in four tropical dry forest landscapes in Latin America. At the Santa Cruz site (Bolivia), correlations were strong and significant while at Chamela Site (Mexico) and the Mata Seca site (Brazil) correlations were moderate but significant as well. In the Machango site (Venezuela), active fires showed no correlation to deforestation events. In general, our findings show that fires detected by the MODIS sensor may be used as predictors of deforestation in tropical dry forest ecosystems. Chapter five, Edge influence on canopy openness and understory microclimate in two Neotropical dry forest fragments, addresses one of the most characteristic features of fragmented tropical forests: the increase in disturbance near the edges of the fragment or what is known as edge effects. Results in gap fraction and Fraction of Intercepted Photosynthetically Active Radiation (FiPAR) show that edge influence at tropical dry forest sites extend to at least 300-m. Finally, Chapter Six, Remote sensing of edge effects in dry forest fragments using CHRIS/Proba Imagery, shows an assessment of changes in the fraction of intercepted photosynthetically active radiation (FiPAR) across four edge-to-interior transects in tropical dry forests fragments and their correlation to spectral vegetation indices (SVIs) computed from the hyperspectral and multiangular CHRIS sensor on board the Proba platform. Results show that the use of spectral vegetation indices for identifying and quantifying edge effects in tropical forests have the potential to improve modeling of forest disturbance in fragmented landscapes. The work contained in these five chapters address issues that are critical to the advancement of tropical dry forest monitoring. These studies contribute to the current scientific literature on the use and application of optical remote sensing tools, not only applicable in tropical dry forests, but for tropical forest conservation at the continental, regional and local level.

The magpie goose (Anseranas semipalmata) is a spectacular and unique waterbird from tropical north Australia and southern New Guinea. Due to recent human persecution, the species has been eliminated from most of its former strongholds in southern and south-eastern Australia – reduced to small conservation-dependent populations through habitat loss, exploitation and drought. Yet, genuine conservation opportunity still exists in northern Australia, in the country's Northern Territory in particular, to maintain viable populations through evidence-based management of wetlands that support the waterbirds and mitigation of the threats posed by global change. Much has been achieved over the last 50 years to understand the ecology and life history of magpie geese, but little has been done to understand important population-level interactions with wetland habitat and the likely outcomes under climate warming, wetland loss to sea level rise, altered competitive interactions among wetland plants, increased frequency and severity of epizootics, and synergies with over-hunting. My review of pathogens and parasites likely to cause morbidity and mass mortality in magpie geese shows that bacterial diseases such as avian cholera and botulism, as well as pathogenic avian influenza viruses, pose the most serious threats. Bacterial diseases in particular are more likely to occur under warmer and wetter conditions, and geese are susceptible to these given large aggregations at favoured nesting and feeding sites. I use a metapopulation model to demonstrate that increased frequency and severity of epizootics will likely force extirpation of geese under current harvest rates across the Northern Territory. Magpie geese are also vulnerable to climate change through dependency on a favoured food plant – the water chestnut (Eleocharis dulcis). As a result of a two-year field programme, I was able to show how birds seasonally migrate and aggregate in response to the availability of this resource and gain body condition following predation on the root tubers of the plants. My geospatial modelling of ocean level inundation of wetlands that support E. dulcis show marginal habitat loss under 1.4 m of sea level rise, and large-scale losses under multi-metre sea level rise, but the current resolution of GIS data do not account for fine-scale saline water intrusion through channel or eroded levees. The population models constructed predict that magpie geese are broadly resilient to change where harvest is tightly regulated, but current harvest rates are unlikely to be sustainable. Given the importance of maintaining viable, connected subpopulations large enough to maintain genetic diversity, and because of the value of magpie geese to Aboriginal Australians as a food source, ongoing monitoring of geese population trends will be essential. Wetland management options include erecting buffers to stop or slow down saltwater intrusion resulting from sea level rise, and implementing a system to monitor annual indigenous harvest. Temporary restrictions on harvest may be necessary following mass mortality events such as epizootics or droughts. Only the careful management of wetlands that support current geese populations, and close monitoring of populations will ensure continued sustainable harvest of geese under global change. Given the cultural and biological significance of this species and the north Australian wetlands that support it, this is a conservation resource we cannot afford to squander. / Thesis (Ph.D.) -- University of Adelaide, School of Earth and Environmental Science, 2009

Factors relating to natural regeneration and tree community recovery after slash and burn agricultural practices were investigated in secondary seasonally dry tropical forests (SDTF) of varying land use intensity and post-disturbance management. Patterns and processes of regeneration were explored by collecting community composition data at secondary and primary sites, by conducting seedling experiments, and by investigating existing forest rehabilitation efforts in both Lao PDR (Laos) and Vietnam. In the forests surveyed in Laos, forest recovery in terms of the composition of primary forest tree species juveniles in fallow forests was high at most sites regardless of previous land use intensity. While community compositional studies indicated significant differences between primary and secondary communities, little or no significant differences were found in terms of primary tree species or family diversity, evenness, or dominance of regenerating juveniles. Significant differences appeared to be due to changes in the relative abundances of different species rather than the absence of primary forest obligate species in the secondary communities. This effect is expected to be due to the relatively common resprouting ability of SDTF species, the wide range of ecological conditions that these species can persist under, and the mosaic pattern of land-uses across the landscape. Evidence of environmental filtering affecting composition patterns in secondary communities was found for two plant traits; dispersal syndrome and plasticity of specific leaf area (SLA). Ecological strategies of primary SDTF tree species were investigated in two ways. Within the narrow range of species capable of regenerating within the primary forest environment, evidence of different life history strategies was found, indicated by significant correlations between continuous plant traits of seed size, fruit size, maximum tree height, SLA, leaf size and wood density. The correlations found largely mirrored patterns found in similar studies in Neotropical forests. Evidence of different seedling regeneration strategies was investigated by comparing species of two different successional preference groups in a seedling/light experiment. Seedlings of non-pioneer later successional species responded differently than species more associated with disturbed environments (long-lived pioneers) when exposed to both increasing absolute light treatments, and to light received under various sunfleck treatments. Earlier successional species demonstrated greater plasticity of SLA, leaf area ratio (LAR), stem elongation, and root mass ratio to increasing light. Regeneration under minimally managed monoculture and pair-wise plantations of both native and exotic species at the Laos field site was not significantly different when compared with nearby remnant primary forest in terms of diversity, evenness, and richness. Therefore, plantation overstoreys did not appear to be suppressing the regeneration of primary forest species. Seedlings grown in enrichment designs nearby had survival rates of approximately 50%, and diameter increments of less than 0.2 cm/yr after 6 to 9 years of growth. The performance of seedlings in such plantings is expected to be sufficient for conservation aims but is unlikely to encourage private investment for forestry purposes due to long expected rotation lengths. At Hai Van Pass in central Vietnam, the combination of a fast growing exotic species to capture a degraded site and ameliorate site conditions, followed by enrichment planting of native species was demonstrated to be a successful approach to reforesting degraded land, whilst funding itself through the sale of timber. The results indicate the high regeneration potential of the studied secondary forests, both in terms of observed patterns of regeneration in slash and burn fallow forests, and in terms of the general regenerative abilities of SDTF species. Decisions regarding future land-uses and secondary forest management should have consideration of this possible high level of recovery, and simple diversity sampling techniques should be included in any related processes to confirm the regeneration potential of a particular secondary forest.

No description available.

No description available.

No description available.

Ross River virus (RRV) is an Alphavirus distributed throughout Australia. It is transmitted by mosquitoes and is known to cause moderate to severe disease symptoms in humans. Along with other alphaviruses such as Sindbis virus and Chikungunya virus, RRV is known to cause arthritic symptoms, characterised by muscle and joint inflammation. Several investigations have established the role of macrophage cells and pro-inflammatory host factors in the development of RRV-induced disease. In this study, we attempted to determine differences between RRV passaged in mammalian and mosquito cells. There is strong evidence that arthropod-borne viruses are able to display enhanced infectivity when passaged into arthropod cell line. We showed that mosquito cell-derived RRV (mos-RRV) was able to replicate to higher titres than mammalian cell-derived RRV. We also showed that mos-RRV failed to induce Type I IFN-associated antiviral responses. The second aim of this study was to investigate the role of TNF-ᬠa pro-inflammatory cytokine implicated in arthritic diseases, in the development of RRV disease. We treated RRV-infected C57BL/6J mice with a commercially available TNF-ᠩnhibitor drug and monitored disease signs. We found that the TNF-ᠩnhibitor does not ameliorate RRV disease (RRVD) symptoms, and that it does not prevent muscle and joint inflammation. We analysed histological sections of muscle and joint tissue of Enbrel-treated and untreated, RRV-infected cells. We also determined and compared host cytokine expression profiles. Finally, we sought to determine the requirement for natural killer (NK) cells in RRV disease. NK cells have been detected in the synovium of RRV-infected patients since early studies, but their role in disease pathogenesis remains unclear. Using a NK-dysfunctional mouse (C57BL/6J-Lystbg), we showed that mice lacking a functional NK system are more susceptible to RRV disease than wildtype, C57BL/6J mice. We monitored disease symptoms following RRV infection and assessed muscle and joint inflammation in Lystbg and C57BL/6J mice. This thesis examines mechanisms of viral infection and immune evasion employed by RRV, as well as into the role of host cells and cytokines in RRVD pathogenesis disease mechanisms. We showed that a functional NK cell system is required for the regulation of RRV-induced muscle and joint inflammation. Our characterisation of the use of a commercial TNF-ᠩnhibitor in RRV-induced disease in mice may provide information on the role of TNF-ᠩn viral arthritis, and may help towards developing safe and effective treatment.

No description available.