Undergraduate opportunities at UNSW

Evaluating Australian Climate Responses to the 2026-27 El Niño

Supervisors: Dr Linyuan Sun and A/Prof Andrea Taschetto

A potentially strong El Niño is currently developing in the tropical Pacific, providing a timely opportunity to evaluate ENSO-based climate predictions using real-world observations. Meanwhile, the climate community has recognized that the traditional Oceanic Niño Index (ONI) is increasingly contaminated by the long-term warming trend, motivating the recent adoption of the Relative Oceanic Niño Index (RONI), which removes the tropical mean warming signal to better isolate ENSO-related interannual variability. This project will compare ONI- and RONI-based reconstructions of South Pacific teleconnections and Australian climate responses using the updated data. In particular, the developing 2026-27 El Niño will serve as a timely out-of-sample test of whether RONI provides a more physically robust predictor of regional climate anomalies in a warming climate.

Requirements: Essential programming skills (e.g. Python), and basic knowleage in climate sciences is welcomed

Machine Learning-Based Equation Discovery in Land-Atmosphere Modelling

Supervisors: Dr Sanaa Hobeichi, Dr Ulrike Bende-Michl (BoM), and Prof Gab Abramowitz

Machine learning-based equation discovery has gained traction in many fields but remains largely underexplored in land surface modelling, where it offers the potential to reveal alternative formulations that may complement or improve existing representations of land–atmosphere fluxes. This project will explore equation discovery, in particular the Sparse Identification of Nonlinear Dynamics (SINDy) approach, to select a parsimonious combination of physically inspired candidate terms derived from meteorological forcings measured at FLUXNET sites and infer an interpretable functional equation for land–atmosphere fluxes. The project will explore both the capabilities and limitations of the ML-based equation discovery approach and its potential to inform improvements in land surface model representations of the examined fluxes.

Requirements: The selected student needs to have experience with Python, High-Performance Computing, and GitHub to be considered for this project.

How well can we simulate land surface processes across Australia?

Supervisors: Mat Lipson, Anna Ukkola

Many impacts of climate change are felt through land surface processes which also modulate the weather and climate processes around us. Land surface models that represent these processes are an integral part of climate models and inform us about many policy-relevant questions including future water resources, carbon uptake and ecosystem processes. The Australian climate research community currently uses two land surface models, CABLE and JULES.

This project will evaluate new high resolution simulations from the two models across Australia to understand their relative strengths and weaknesses in simulating Australia water, carbon and energy cycles. The specific focus can be tailored to suit the student’s interests but could include evaluating long-term hydrological processes (e.g. evaporation and runoff), carbon fluxes, urban processes or extremes such as drought. The project will contribute to a wider collaboration across several universities, the Bureau of Meteorology and ACCESS-NRI.

Requirements: Familiarity with scientific programming (R, Python or similar)

Building a tool to visualise data size vs performance

Supervisors: Sam Green, Sanaa Hobeichi

Build a tool that benchmarks and visualises how common scientific array operations scale with dataset size and chunking (NumPy vs xarray eager vs xarray+Dask), producing interactive “scaling curves” and “chunking maps” that teach users how performance changes.

Requirements: A student who is comfortable writing Python, curious about performance, willing to run controlled experiments, and interested in understanding how scientific code behaves as data sizes grow.

Evaluating the benefits of bias correction for regional climate downscaling

Supervisors: Jason Evans, Ulrike Bende-Michl, Christian Stassen, Leena Khadke

Global climate models (GCMs) are essential for projecting long-term climate trends, but their coarse spatial resolution limits their ability to capture regional climate variability and extremes. To address this, high-resolution regional downscaling has been carried out over Australia, in collaboration with the Bureau of Meteorology, CSIRO, the NSW Department of Climate Change, Energy, the Environment and Water (DCCEEW), the University of New South Wales, and the University of Queensland. The first phase of project investigated the added value of the four regional downscaled simulations compared to global CMIP6 models. Now, in the second phase, the student will focus on whether bias correction will influence the outcomes of the added value datasets. The student will target specific questions, 1) Does bias correction improve representation of extremes or only the mean climate? 2) Does bias correction alter spatial patterns? 3) Does bias correction affect trend signals? Through this work, the student will develop skills in analysing high-resolution climate datasets, interpreting model outputs, and working with scientific programming tools commonly used in climate science. 

Requirements: Basic proficiency in Python or another scientific programming language is required. Familiarity with climate data formats (e.g., NetCDF) or experience working with large datasets is beneficial but not essential