|Funding for:||UK Students, EU Students, International Students|
|Funding amount:||From £17,668 per annum|
|Placed On:||4th May 2023|
|Closes:||31st May 2023|
Project Title: Machine Learning approaches to estuarine and coastal ecosystem health
Centre for Doctoral Training in Environmental Intelligence, Streatham Campus, Exeter
The University of Exeter’s Centre for Doctoral Training in Environmental Intelligence, in partnership with the Plymouth Marine Lab, is inviting applications for a fully-funded PhD studentship to commence in September 2023. The successful applicant will join the UKRI CDT in Environmental Intelligence, and will be included in CDT cohort building and training activities. The successful applicant will work on the below project under the supervision of Peter Challenor and Daniel Williamson (University of Exeter), with additional supervision and support from Plymouth Marine Lab.
Advances in hydrodynamic and biogeochemical modelling are dramatically increasing our ability to model marine coastal areas at unprecedented detail. Such operational models can provide real-time information about coastal conditions like currents, suspended sediment concentrations and biogeochemical parameters such as chlorophyll-a concentrations, plankton biomass and nutrients. These models, while highly sophisticated, can still suffer from structural uncertainty, parameter uncertainty or initial condition uncertainty which can limit their uptake.
Additionally, their high computational cost also restricts their use in scenario and/or ensemble mode. On the other hand, our ability to efficiently monitor the vast number of ecosystem processes is still very limited. The ability to combine scarce data with “imperfect” models promises to improve our understanding of marine ecosystems and will play a significant role in our ability to monitor our coasts to fulfil environmental regulations and UK’s implementation of climate legislation to reduce carbon emissions.
We propose to develop new approaches of data blending that are operational feasible and suitable for exploring a wide range of policy and climate scenarios as well as contribute to regular monitoring of the health of our coast. The candidate will develop light weight emulators for the Tamar Estuary to act as an interface for an eventual Digital Twin of the biogeochemistry of the area. The emulators will be designed to predict from potentially observable inputs (remote sensing, distributed network of in-water sensors) key indicators of estuarine conditions, such as nutrients, biological production, CO2 air-sea fluxes and bottom oxygen concentrations.
Furthermore, the emulators will determine the requirements on the satellite (low-resolution) as well as in situ observing networks (variables and locations) that enable skilled prediction of those environmental indicators.
Finally, these emulators should be capable of addressing what-if type scenarios related to:
Type / Role: