PHD PROJECTS
AAPP PhD Projects and Scholarships
A number of PhD research projects related to the research program of the AAPP will be advertised below and on the University of Tasmania Graduate Research website. Applicants will be able to apply for Stipend Scholarships and fee waivers from the University of Tasmania or from other sources. If successful, applicants will also receive a top-up scholarship of $6,000 per annum for 3.5 years. This scholarship is funded from the Australian Government as part of the Antarctic Science Collaboration Initiative program through the Australian Antarctic Program Partnership (AAPP).
If you are interested in undertaking a PhD with the AAPP, please check this page frequently for opportunities or contact any of our researchers directly.
Ice shelf deep learning
Project 3: Ice shelves
Abstract
Deep learning with artificial neural networks has evolved rapidly and become a widely applied tool for the study of Earth surface processes. In glaciology, neural networks have been used to emulate physically advanced glacier models, speeding up the computational simulation by several orders of magnitude. This efficiency gain makes deep learning a promising new tool for assessment of ice shelves, which are the floating extension of the Antarctic Ice Sheet.
This project will use machine learning to understand how ice shelves in Antarctica behave and interact with the ocean and climate. The student will simulate ice shelves using the Instructed Glacier Model which emulates the physics of ice shelves. Upon configuration and training, the emulator will be used to investigate ice shelves across Antarctica with the aim of identifying which ones are fragile and which ones are stable over the coming decades and century.
Specifically, the research will:
- Use a physics-informed emulator to simulate the flow of ice shelves in East Antarctica
- Explore the emulator model's ability to reproduce ice shelves as observed
- Assess ice shelf stability of East Antarctic ice shelves under different environmental forcing scenarios
With a modern approach and broad scope, the research will advance our understanding of ice shelves, which have disappeared almost entirely in Greenland (where climate is warmer) but still play a major role for the stability of the Antarctic Ice Sheet.
Primary Supervisor
Poul Christoffersen
Modelling cloud-aerosol interaction
Project 1: Atmosphere
Abstract
The project will seek to better understand aerosol-cloud interaction in pristine clouds at Kennaook/Cape Grim using both field campaign data and a regional atmospheric model
Kennaook/Cape Grim, located on the northwest coast of Tasmania, is considered an excellent location to observe 'baseline' air, i.e. pristine Southern Ocean airmasses (Gras & Keywood, 2017). Such pristine airmasses provide an excellent opportunity to understand atmospheric processes away from the influence of human activity, which can help us understand how our weather and climate is changing.
Climate and weather models still produce large significant Southern Ocean radiation biases that have been attributed to the poor simulation of aerosol-cloud processes in such pristine environments (Fiddes et al., 2022; Mallet et al., 2023). One of the primary reasons for this problem is the limited availability of observations in the area, leading to a lack of understanding of crucial cloud formation processes specific to the Southern Ocean (McFarquhar et al., 2020).
Beginning in April 2024, the Cloud and Precipitation Experiment at Kennaook (Cape-K) campaign will begin, through to September 2025 (Mace et al., 2023). This campaign will be complemented by a comparison voyage of the RV Investigator and a potential flight campaign. These extensive and co-ordinated field campaigns, targeting aerosol-cloud interaction, provide a rare opportunity for processes understanding, model evaluation and development across a range of seasons.
Objectives
1. What are the microphysical and aerosol properties of different cloud types in pristine Southern Ocean airmasses?
2. Can a high-resolution model represent these microphysical and aerosol properties?
3. Can we improve the representation of these cloud properties by improving the aerosol and cloud microphysics in the model, or with machine learning?
4. What is the seasonal dependence of cloud properties with respect to aerosol sources and can we replicate this in an atmospheric model?
Primary Supervisor
Sonya Fiddes