|UK Students, EU Students, International Students
|£18,622 per annum
|2nd November 2023
|9th January 2024
About the Partnership
This project is one of a number that are in competition for funding from the NERC Great Western Four+ Doctoral Training Partnership (GW4+ DTP). The GW4+ DTP consists of the Great Western Four alliance of the University of Bath, University of Bristol, Cardiff University and the University of Exeter plus five Research Organisation partners: British Antarctic Survey, British Geological Survey, Centre for Ecology and Hydrology, the Natural History Museum and Plymouth Marine Laboratory. The partnership aims to provide a broad training in earth and environmental sciences, designed to train tomorrow’s leaders in earth and environmental science.
Climate change significantly affects the health of our oceans. Directly influenced ocean health indicators (OHIs) include net primary productivity, oxygen, and pH/acidity, with major impacts on food security, livelihoods and economy, as well as providing crucial feedback to climate. The most direct way to understand how these OHIs respond to the increase in carbon dioxide in the atmosphere and the warming planet is by running multi-decadal simulations (both future projections and hindcasts) of Earth System Models. However, such models are too complex and computationally prohibitive to explore large and complex landscapes of emission scenarios for ‘what-if’ analyses to inform policy and decision-making. With rapid developments in machine learning (ML) in the last decade there is an emerging research interest to develop ML emulators trained on complex model climate simulations that can accurately reproduce the results of complex marine ecosystem models with order of magnitude less computation. Such ML emulators could provide valuable tools for researchers, policy makers and environmental managers. The emulators could particularly benefit users from developing countries who lack access to high-performance computational resources and established data archives.
Project Aims and Methods
This project will test the hypothesis that machine learning (ML) models can be used to accurately reproduce the results of sophisticated mechanistic marine ecosystem models at a fraction of the computational cost. Our aim is to create ML based emulators that can do the job of the mechanistic models faithfully (see Fig.1) and make these available to the research and policy communities as an open-source tool. More specifically, our central hypothesis is that ML models trained on the outputs of CMIP6 Earth system models will be able to reveal crucial patterns in the responses of net primary production, oxygen, and acidity to changes in atmospheric carbon dioxide concentrations, temperature, and other key physical variables.
The student will be encouraged to discuss the detailed research directions, to better match the interests and maximise the project outcomes.
Type / Role: