School of Earth Sciences - Theses
Permanent URI for this collection
Search Results
1 - 7 of 7
-
ItemNo Preview AvailableAntarctic sea ice and its interactions with high latitude weather and climate( 1998)Antarctic sea ice plays a major role in the earth system by greatly influencing the high latitude exchanges of heat, moisture and momentum between the ocean and atmosphere, as well as profoundly effecting the salt budget of the ocean, and thus the production of Antarctic Bottom Water, one of the driving mechanisms of worldwide oceanic circulation. With such considerable and far reaching impact, it is important to document its climatology, understand its variability and quantify its influence. Climatologies and trends of the Southern Ocean sea ice pack are presented using the most recent satellite observations available from the Defense Meteorological Program’s (DMSP) Special Sensor Microwave Imager (SSM/I). The analysis of these data show that Antarctic sea ice is highly variable in both time and space. Statistically significant increases in the sea ice extent, open water and ice areas have been determined from the SSM/I data for the 9 year period 1987 to 1996, a result which differs from the Scanning Multichannel Microwave Radiometer (SMMR) observations (1978-1987). The increasing trend in the SSM/I observations can be attributed to the large increases in sea ice observed in 1994-1995, as confirmed by an analysis of data from the ERS-1 satellite. The mean season length during these years has remained relatively unchanged. Regional trends, both in the sea ice concentration and in season length, showed vast spatial inhomogeneity. SSM/I data displayed increasing season length in the central Weddell Sea, Bellingshausen Sea and Balleny Islands regions, with decreasing length in the Amundsen Sea, eastern Ross Sea and in the coastal areas off Wilkes Land. Similar trends are observed in the seasonal sea ice concentration. In most cases, these trends are opposite to those observed in the SMMR data, which may be linked to the shift observed in the Amundsen Sea low after 1990. Comparisons with historical data would suggest that no large scale anomalous change has occurred in the Antarctic sea ice limits over the course of human observation. Furthermore, the degree of variability suggests great care is needed in interpreting large scale changes in sea ice conditions, and hence atmospheric or oceanic change, from locally observed anomalies. Case studies of the effect of individual cyclones upon the sea ice concentration show small but definite modification of the ice conditions. To further diagnose aspects of the thermodynamic and dynamic forcing upon the Antarctic pack, detailed analysis of the sea ice concentration variability has been conducted using spectral techniques, and the spectra have been compared to those of the European Centre for Medium Range Weather Forecasts (ECMWF) temperature and wind data. In all cases, and with the seasonal cycle removed, the sea ice concentration shows a bias towards longer timescales of variability than either the wind stress or surface air temperature. This “red shift” in its frequency spectrum is strongest with the wind stress, and weakest with the temperature. For longer period waves, this may be due to the formation of new ice by surface cooling or the moderation of melting by the cold surface water, whereas for shorter period waves, where wind stress dominates temperature and ice concentration respectively, time is required for winds to draw in warmer or cooler air, as well as to overcome the ice masses inertia and keel friction to open or close leads. Strong intraseasonal variability of the sea ice concentration is observed in the 20-25 day period, reflecting similar timescales of the temperature variability, as well as that of the energetic eddies of the Antarctic circumpolar current. Examination of the latitudinal variation of the sea ice concentration, temperature and wind stress spectra showed not only the importance of the north-south temperature gradient in influencing the variability, but also the seasonal changes in the semi annual oscillation of the circumpolar trough. Regional spectra showed clear differences between location, and reflected the influences of the atmosphere and ocean upon the sea ice pack. This is clearly shown in the Weddell Polynya region and off East Antarctica, with high variability in the synoptic timescales, and in the western Ross Sea where changes occur in timescales of greater than 20 days. In order to determine if satellite derived, real time sea ice concentration and distribution would be of benefit to operational numerical weather prediction (NWP) schemes, the effect of sea ice concentration change upon the atmosphere in synoptic timescales was examined using a general circulation model in conjunction with the Australian Bureau of Meteorology’s GASP analyses. Experiments were conducted with a typical July sea ice concentration and distribution, as well as slab concentrations of 0, 10, 25, 50, 80 and 100%. Results from 5-day numerical weather forecasts show that the central pressure, structure and tracks of individual cyclones are sensitive to the ‘switch on’ of different sea ice conditions. Composites of all forecasts made with each concentration showed considerable, and mostly statistically significant, anomalies in the surface temperatures and turbulent heat fluxes over the sea ice. The magnitudes of these changes varied monotonically with the area of open water. The largest changes were simulated closest to the coast for all concentrations except for the typical July sea ice run, which displayed maxima over the outer pack. Significant westerly anomalies were induced over the ice in all cases, as were reductions in mean sea level pressure. The July sea ice runs displayed a distribution of the mean sea level pressure anomaly different from all others, with maxima occurring in the central to outer pack. All other forecasts displayed maxima at the coast. The results suggest that sea ice concentration does induce anomalies in the atmospheric parameters in timescales of less than five days. Further, the use of a realistic distribution of sea ice concentration produces results distinct from the constant concentration forecasts. Hence it is suggested that real time Antarctic sea ice data may be of considerable benefit to numerical weather prediction models.
-
ItemAn attribution study of southeast Australian wildfire risk( 2016)Extreme weather and climate-related events often have a serious impact on our economy, environment and society. This is particularly true in Australia where recurring heat waves, floods, droughts and wildfires have resulted in the loss of life, property and livelihoods. The 2009 'Black Saturday' wildfires in southeast Australia provides a tragic example, having resulted in the death of 173 people and the destruction of over 2000 homes. While there are a number of recorded attribution studies for Australian temperature and precipitation-related events, no such study exists for fire weather. This thesis presents a new climate modelling system for regional climate simulation and the attribution of weather and climate extremes over Australia and New Zealand. The system, known as weather@home Australia-New Zealand, uses public volunteers' home computers to run a moderate-resolution global atmospheric model with a nested regional model over the Australasian region. By harnessing the aggregated computing power of home computers, weather@home is able to generate an unprecedented number of simulations of possible weather under various climate scenarios. This combination of large ensemble size with high spatial resolution allows a range of extreme events to be examined over Australia and New Zealand with well-constrained estimates of sampling uncertainty. The model is seen to be capable of resolving many climate features that are important for the Australian and New Zealand regions, including the influence of El Nino-Southern Oscillation (ENSO) on driving natural climate variability. Using the new weather@home modelling framework, this thesis presents the first known attribution study of southeast Australian fire weather. By applying the McArthur Forest Fire Danger Index to large ensembles of regional climate model simulations generated for factual and counterfactual climate scenarios, this thesis reveals that anthropogenic climate change increased the likelihood of elevated wildfire risk over southeast Australia during the 2008-2009 fire season. Furthermore, the influence of anthropogenic climate change on wildfire risk is found to be greater in spring than summer. Through a series of further modelling experiments, this thesis also demonstrates a novel approach for separating the influence of ENSO and anthropogenic climate change within the context of an attribution study. Across southeast Australia, the increase in wildfire risk due to a change in ENSO phase (from La Nina to El Nino conditions) was identified to be much greater than the increase attributed to anthropogenic climate change. This was largely due to the strong increase in drought factor, and decrease in relative humidity, from La Nina to El Nino conditions.
-
ItemThe response of nonlinear feedbacks associated with the Indian Ocean Dipole to global warming( 2015)The Indian Ocean Dipole (IOD) is an interannual mode of variability in the tropical Indian Ocean which has the potential to severely impact the surrounding countries. A prominent feature of the IOD is its positive skewness, where positive IOD events tend to be larger than negative events resulting in stronger impacts. Models from the Coupled Model Intercomparison Project phase 5 (CMIP5) show a threefold increase in the frequency of extreme events by 2100. However, the contribution of feedbacks associated with the IOD to this skewness and how they respond to increasing greenhouse gases are not well understood. The nonlinearity of four feedbacks (the Bjerknes feedback, sea surface temperature (SST)-cloud-radiation feedback, wind-evaporation-SST feedback, and nonlinear dynamic heating) was investigated and their responses to a warmer climate examined using CMIP5 models. A single model was first used to examine the nonlinearity of the four feedbacks, allowing the cause and effect of each feedback process to be determined. Following this, the analysis was applied to a multi-model ensemble to determine the cause of positive IOD skewness. The role of nonlinear dynamic heating in increasing the frequency of extreme events was then examined. This provides a better understanding of IOD dynamics and future behaviour. The positive Bjerknes feedback controls IOD skewness through the thermocline feedback. In a warmer climate, models show that the skewness of the IOD weakens and the asymmetry of the thermocline feedback displays a significant relationship with this reduced skewness. This decreased asymmetry occurs due to the mean state change of the tropical Indian Ocean where the Walker circulation exhibits weakening. The weaker mean westerly winds allow the climatological thermocline to shoal in the east, reducing the asymmetry of the thermocline feedback and thereby IOD skewness. The other feedbacks do not show as strong a relationship with IOD skewness. The negative SST-cloud-radiation feedback shows stronger damping of positive IODs and this is detrimental for IOD skewness. Under greenhouse warming there appears to be a shift amongst most models, with weaker damping of positive IODs and stronger damping of negative IODs. This is unfavourable for the reduction in skewness that occurs under increasing greenhouse gases. The positive wind-evaporation-SST feedback is not well simulated in coupled models. Therefore the relationship between IOD skewness and the wind-evaporation-SST feedback is uncertain and there does not appear to be a significant change in the asymmetry of this feedback in a warmer climate. Nonlinear dynamic heating involves advection of heat by oceanic currents, reinforcing positive IODs but damping negative IOD events. Amongst the models analysed, this process does not appear to contribute significantly to IOD skewness; however it does play an important role in the generation of extreme events and the projected increase. Nonlinear zonal and vertical advection during moderate events becomes more extreme-like due to the mean state change of the tropical Indian Ocean. This means that conditions during future moderate positive IOD events are similar to present-day extreme events, facilitating the projected increase of extreme positive IODs in a warmer climate.
-
ItemHydrological extremes and consequences of climate change( 2013)In the design of infrastructure, risk has been – and often still is – assessed on the basis of long-term averages. Information on variation in hydrological extremes is required as the basis for informed decision-making, preparedness and possible adaptation. Long-term trends are fairly well understood for temperature but less well for precipitation. While climate models have become sophisticated tools for projecting future changes in our climate, their ability to replicate observed variations in precipitation is limited and it is therefore prudent to complement climate models through analysis of historical observations. Design rainfall is used as one of the required inputs for hydrological models in the design of structures such as dams and bridges. Design rainfall estimates are supplied in form of intensity-frequency-duration curves. Rainfall frequency analysis is almost invariably based on the assumption of a stationary climate. Sub-daily durations are of particular interest for urban applications. This thesis was strongly driven by the motivation to provide guidance to decision makers who have to account for non-stationarity in rainfall extremes. Non-stationarity in rainfall extremes comes about as a conflation of climate change and climate variability. Unlike for temperature extremes, rainfall extremes for Australia as a whole exhibit no clear increase or decrease in intensity over time but strong association with the El Niño-Southern Oscillation (ENSO). This has implications for the choice of suitable analysis techniques, e.g. sophisticated non-parametric techniques. Depending on the planning horizons both climate change and climate variability may have to be accounted for. The association of rainfall extremes with ENSO leads to an opportunity to develop statistical models to support decision-making on shorter time scales. Analysis of seasonality in frequency and magnitude of rainfall extremes revealed considerable variation across a set of sites in the southeast of Australia, implying different dominating rainfall-producing mechanisms and/or interactions with local topography. The strongest signal for an increase in extreme precipitation is found for short durations. Changes in rainfall extremes come about through a combination of changes in thermodynamical and dynamical variables. To assess large-scale changes in circulation, a classification technique (self-organising maps, SOM) was applied and synoptic types were identified. Rainfall extremes were then related to the synoptic type under which they occurred, to assess observed changes in the frequency of rainfall extremes. Rainfall extremes are typically preceded by conditions that are much wetter (both in absolute and relative terms) and warmer than the climatological average. These anomalies tend to be larger for shorter durations, and for rarer events. Given that increase in humidity exhibits strong regional variability and that it may be counteracted by changes in dynamics, it appears simplistic to state categorically that climate change will lead to an increase in extreme rainfall events and observed trends in rainfall extremes show a picture that is more complex. In summary, the combination of changes in thermodynamic and dynamic variables will define the change in frequency and intensity of rainfall extremes. The factors that are most relevant for the effect of climate change on rainfall extremes depend on geographical location.
-
ItemDrivers of Southern Hemisphere climate change( 2013)The climate of the Southern Hemisphere (SH) has undergone significant changes over recent decades, with additional warming expected under future emission scenarios. However, while temperature increases are robust across models there is more uncertainty around changes in rainfall, atmospheric circulation and extremes, all of which have a large impact on human society and ecosystems. The objective of this thesis was to increase our confidence in future projections by considering the relative importance of various drivers of past and future climate change, with a focus on the Southern Hemisphere. These drivers include sea surface temperatures (SSTs; which may or may not be anthropogenically forced), greenhouse gases and stratospheric ozone. To achieve this objective a hierarchy of model experiments were utilised, from idealised experiments to multimodel datasets. Insights were gained by exploring both the consistencies and the spread across the model results. The main results were: 1) The internal variability of the climate system, such as the El Niño-Southern Oscillation (ENSO), impacts the emerging signals of anthropogenic climate change and characterisations of this noise were explored. Opposite phases of ENSO were found to drive marked contrasts in maximum temperature extremes, with ENSO fidelity crucial in simulating the observed relationships. These patterns are unlikely to change substantially under future climate change. Over the Australian continent, future warming leads to increases in warm temperature extremes and a propensity for longer dry spells interspersed with heavy rainfall events. In general, the magnitude of changes in both temperature and precipitation extremes indices scaled with the strength of emissions. 2) Coupled model simulations were able to reproduce the large-scale features of SH climate trends since 1950, if observed changes in anthropogenic forcings were included. However, atmospheric models driven by observed SSTs and anthropogenic forcings were unable to capture wintertime trends, suggesting either deficiencies in the modelling framework, SSTs or models themselves or that internal variability has been largely responsible for these trends. The improvement of the simulated trends in experiments with partial coupling suggests the modelling framework plays some role in this deficiency. 3) Future changes in the SH atmospheric circulation will be driven by the competing effects of greenhouse gases and stratospheric ozone recovery. Climate sensitivity was found to largely explain the difference in Southern Annular Mode (SAM) projections between two coupled climate models under identical greenhouse gas and stratospheric ozone forcing. This result extended to multimodel simulations under transient carbon dioxide (CO2) conditions in all seasons, with the stronger the warming the larger the trend in the SAM. Tropical upper tropospheric warming was found to be more relevant than polar stratospheric cooling to the model spread in SAM responses to CO2. 4) Idealised SST experiments in two models showed a consistent poleward shift of the SH wintertime westerly jet under Southern Ocean cooling. However, the austral winter response to increased tropical SSTs was found to be model dependent, with opposite latitudinal shifts in the SH westerly jet in the two models. This finding was linked to different tropical rainfall and convective atmospheric heating responses in the models to identical SST increases. These results highlight the reliability of current climate model simulations as well as some of their limitations. Potential deficiencies in forcing datasets, modelling frameworks and the simulation of internal variability were identified. Understanding and improving these deficiencies is crucial for interpreting recent observed change and understanding future projections, particularly at the regional scale.
-
ItemThe impacts of climate variability and change on severe thunderstorm environments in Australia( 2012)Severe thunderstorms present a relatively infrequent but significant threat to property and life in Australia during the spring and summer. These thunderstorms can produce hailstones over 2cm in diameter, winds in excess of 90kmh-1 and less frequently tornadoes. Any of these phenomena can result in localised high impact severe events. Recent examples of this potential are illustrated by damage caused by the 1999 Sydney Hailstorm, 2008 Gap Microburst and the 2010 and 2011 Melbourne Hailstorms. This risk makes the implications of a changing and variable climate on severe thunderstorms important to understand. Recent studies into the impacts of anthropogenic climate change on severe weather events, including thunderstorms, suggest a potential increasing trend in both frequency and intensity for Australia. While current convective parameterisations in both global and regional climate models limit direct assessments of future convection, the use of environmental parameters to estimate changes in severe thunderstorm environments has been successful in other geographical regions. This study seeks an answer to the question “Is the frequency and distribution of severe thunderstorm environments in Australia likely to change in the future?” A database of 1550 independent severe thunderstorm reports in Australia has been developed for the period March 2003 to April 2010. Severe thunderstorm reports are then used to identify relationships with their associated environments estimated using proximal soundings from a mesoscale numerical weather assimilation and prediction model (MesoLAPS). This proximity climatology of known severe thunderstorm environments has been successfully used to derive covariate discriminants that identify the potential of an environment to produce severe thunderstorms. These covariates use variables describing the potential for organised convection (deep-layer wind shear), and the potential for instability over the depth of the atmosphere (convective available potential energy). Applying these discriminants to a reanalysis dataset (ERA-Interim), a climatology of the frequency and spatial distribution of environments favourable to the development of severe and significant severe thunderstorms for Australia has been developed for warm seasons during the period 1979-2011. This climatology demonstrates that inter-annual variability in terms of both the frequency and spatial distribution of environments is influenced by El Niño- Southern Oscillation. La Niña conditions are typically associated with an increased frequency and an inland shift of favourable environments over eastern Australia, while El Niño typically results in fewer environments, particularly along the coastal fringe. Applying this climatology, the environments simulated by two climate models (CSIRO Mk3.6 and CCAM) for the 20-season period 1980-2000 are examined over Australia and tested against the reanalysis climatology. In particular, the ability of the models to resolve the intra-annual variability and both quantify and simulate the spatial distribution of convective variables are analysed, and are found to perform reasonably well, especially in the case of the higher resolution CCAM. Finally, future simulations of severe thunderstorm environments from high emissions projections for the period 2079-2099 are presented for both models. Comparing these simulations to the 20th century, a potential small increase in the frequency of severe thunderstorm environments appears likely for southeast and eastern Australia under a warming climate.
-
ItemEstimating uncertainties in future global warming using a simple climate model( 2011)This research has investigated the sources of uncertainty that apply to global–mean temperature change projections. Uncertainties in climate system processes have led to a wide range of projections for future temperature changes, which are compounded by the range of possible future greenhouse–gas emissions. For example, the 2007 Intergovernmental Panel on Climate Change Fourth Assessment Report reported that, by 2100, the global–mean temperature increase relative to 1990 is likely to be in the range 1.1°C to 6.4°C, a result that reflects uncertainties in both future emissions and the response of the climate system. However, such a wide range is not particularly helpful for policy and planning purposes, especially in the absence of probabilities. This research has investigated the reasons for this wide range, assessed the sources of uncertainty and developed a method for producing probabilistic temperature change projections. A simple climate model was selected as the research tool for this investigation, instead of a complex three–dimensional model. The model chosen was the latest version of MAGICC (Model for the Assessment of Greenhouse–gas Induced Climate Change), which represents many of the important processes that determine variations of the Earth’s climate, including radiative forcing, heat uptake in the ocean and the carbon cycle, albeit highly simplified and only for temperature changes at the global scale. One of the features of this research is the development of alternative approaches to constraining the model’s primary climate system and carbon cycle parameters. It was found that indices using land minus ocean and Northern Hemisphere minus Southern Hemisphere temperature anomalies are only weakly correlated with global–mean temperatures, and therefore provide additional independent information that can assist in better estimating some model parameters. A ratio of sea–surface temperature to ocean heat content was also found to have a low correlation to the sea– surface temperatures, creating an alternate measure for constraining ocean parameters. This ratio, as well as the vertical ocean temperature change profile, led to revised estimates for the ocean vertical diffusivity parameter, resulting in a new estimate that is nearly a quarter of the previously standard setting used with the Third and Fourth IPCC assessment report versions of MAGICC. In addition to constraining individual model parameters with targeted observational information, a Bayesian statistical technique, the Monte Carlo Metropolis–Hastings algorithm (MCMH), was applied to constraining groups of model parameters using historical observations. One advantage of the MCMH technique is that it addresses uncertainty that arises from observations, model structure and climate system response. This resulted in probability distributions for the key model parameters which then allowed the production of probabilistic temperature change projections. The carbon cycle was included in the MCMH process, leading to a successful calibration of MAGICC’s key carbon cycle parameters with observations for the first time. The MCMH technique was applied to a number of emissions scenarios, enabling probabilistic estimates to be made of global–mean temperature changes to the end of this century. These show reduced uncertainty ranges for future warming projections, with higher lower bounds for warming due to business–as–usual emissions as compared to the results reported in the IPCC’s Fourth Assessment Report. The upper bound for the likely range is also considerably reduced. For the highest emissions scenario, the SRES A1FI, there is a 50% probability of exceeding 2°C by 2042, with a 73% probability of exceeding 4°C by 2100. Analysis of stabilisation scenarios shows that limiting further increases in global–mean temperature to 2°C above pre-industrial requires massive reductions in anthropogenic greenhouse–gas emissions, to the extent that almost zero CO2 emissions are required by the end of this century. Even then, the temperature increase will peak around mid-century, with the upper bound of the likely range temperature change exceeding 2°C, which then entails the risk of irreversible changes to the climate system.