Competition-funded PhDs at University of Dundee

Institute of Mathematics and its Applications; Dundee; On-site; GBP 80,000 - 100,000; Full time

Job summary

A research institute is seeking candidates for a project focused on developing AI models to optimize peatland regeneration in Scotland's Flow Country. Successful candidates will co-design tools with stakeholders, build emulators, and create decision-making frameworks. Ideal applicants will have a strong quantitative background in mathematics, statistics, or related fields, with a keen interest in machine learning and environmental applications. Applications are open until 9 January 2026. Contact Dr. Eric Hall for inquiries.

Qualifications

• Candidates should have a strong quantitative background.
• Experience in machine learning for PDE/simulation models is desirable.
• Interest in climate or environmental applications is preferred.

Responsibilities

• Co-design a Regenerative Scorecard with stakeholders.
• Build and validate “peat-aware” emulators of the JULES‑PEAT model.
• Create a decision layer for ranking rewetting options under uncertainty.

Skills

Quantitative analysis

Machine learning

Uncertainty quantification

Education

Strong quantitative background (e.g. mathematics, statistics, data science, physics)

Job description

CRAFT – Co-designed Regenerative AI Flow Country Toolkit: credible, fast decision support for peatland rewetting

Summary:

The project will develop AI surrogate models to optimise peatland regeneration in Scotland’s Flow Country across multiple objectives, including carbon sequestration, biodiversity, wildfire risk, flood attenuation, and community benefits.

You will:

• Co-design a Regenerative Scorecard with stakeholders (industry, government, NGOs, community groups) to capture local priorities and acceptable trade-offs.
• Build and validate “peat‑aware” emulators of the JULES‑PEAT land‑surface model using approaches such as sparse multi‑output Gaussian processes and/or operator‑learning architectures (e.g. Fourier Neural Operators, DeepONets).
• Create a decision layer that ranks rewetting options under uncertainty, exposing uncertainty bands and scenario narratives and delivering a Regenerative Planning Toolbox for rapid scenario exploration.

How to apply:

This competition‑funded project is being recruited as part of the Leverhulme Doctoral Programme for Regenerative Innovation (Regnr8‑i) at the University of Dundee and is open to UK and international applicants. It is ideal for candidates with a strong quantitative background (e.g. mathematics, statistics, data science, physics, or related fields) and an interest in uncertainty quantification, machine learning for PDE/simulation models, and climate or environmental applications.

The application deadline for this competition is 9 January 2026. Please email the principal supervisor, Dr Eric Hall ([e-mail] ), with any enquiries about the project as early as possible before the deadline.

Image Credit: Wetland in the Flow Country, Scotland, UK by Andrew Tryon CC BY-SA 2.0)