Designing ensembles for efficient climate model emulation and uncertainty quantification

computationally expensive climate model simulations, statistical emulators can be employed to represent the underlying simulator behaviour at a fraction of the computational cost. They also provide a principled measure of uncertainty for their predictions. This quantified uncertainty can be utilised within active learning frameworks to guide future simulations. For STEM graduates, this field offers an exciting opportunity to apply cutting-edge mathematics and computational tools to global climate resilience.

PhD project

Climate change is one of the greatest global challenges of our time, making the design of adaptation policies increasingly crucial. Traditionally, climate projections rely on a limited number of socio-economic scenarios to model the future. In this PhD, we will explore an alternative approach that represents potential emission scenarios using mathematical curves rather than static socio-economic pathways. This will allow us to build statistical emulators, such as Gaussian processes, capturing both parameter and emission uncertainties to deliver robust, probabilistic climate predictions. However, constructing these emulators typically requires large ensembles of computationally expensive climate model simulations. To address this challenge, a key objective of this research is to develop efficient, sequential experimental designs that minimise computational costs while maximising information gain. Overcoming this computational bottleneck forms a core part of the UNRISK project’s mission to revolutionise risk management and decision-making under uncertainty.

As a PhD researcher on this project, you will bridge the gap between advanced data science and climate policy. You will use machine learning, statistical modelling, and sequential experimental design to build efficient climate emulators and translate these probabilistic outputs into actionable decision tools. This research will deliver a transformative framework for climate risk assessment. By providing probabilistic climate predictions, your outcomes will allow policymakers and stakeholders to evaluate risks across an infinite spectrum of future emission pathways. Joining this project offers an exceptional opportunity to work within the vibrant UNRISK research group, developing highly sought-after skills in machine learning and uncertainty quantification while collaborating closely with leading experts across the network to ensure your research has direct, real-world impact.

Applicant Profile

This PhD is suited for students with a strong background in mathematics, statistics, data science, or a highly quantitative discipline (such as physics or theoretical computer science) who want to apply their advanced computational and analytical skills to global climate challenges. We are looking for a motivated researcher who is excited about bridging the gap between statistical theory and real-world environmental policy. Experience or a keen interest in Gaussian processes, uncertainty quantification, active learning, and climate model simulations is highly desirable.

Funding Notes

This project is part of the UNRISK CDT, which offers 15-18 fully-funded NERC studentships, covering full university tuition fees; a personal stipend at standard UKRI rates; £6000 individual research and training costs; £5000 (per student) of cohort-level training; and a ‘Flexible Fund’ for special projects.

International applicants will need to cover costs related to applying for a student visa and the international health surcharge (IHS).

Applications are open to UK and international applicants. The number of awards for international applicants is limited by UKRI rules.

More information is available on the UNRISK website.