I am interested in the response of the ocean to climate change. I use machine learning tools, alongside state-of-the-art ocean models and observations, to quantify thermodynamic and circulation changes in the Southern and global ocean.
Gaussian Mixture Modelling-based clusters obtained from the temperature and salinity distribution in the Antarctic shelf region, from the ACCESS-OM2-01 ocean model.
I work on applying unsupervised machine learning methods to oceanographic applications. I use classification algorithms like Binary Space Partitioning and Gaussian Mixture Modelling to explore changes in key oceanic regimes.
I also use Neural Networks and Transformer methods to infill missing ocean observations, and develop long-term projections of the ocean state over time.
Snapshot of sea surface temperature in a global 1/6th-degree NumericalEarth.jl ocean.
I am a developer of the Oceananigans.jl and NumericalEarth.jl ocean models. These models are GPU-powered, promising up to 50x speed up from traditional CPU-powered ocean models.
I assess past and future changes to the Southern Ocean, both in terms of thermohaline properties, and circulation dynamics.
I analyse high-resolution simulations of the Southern Ocean alongside observations to understand future changes in the region. This includes assessing the strength of overturning circulation, Antarctic Circumpolar Current changes, and missing salinity observations.