Models for Decision Making
Part of the 21st Century Decision Making and Energy, Environment and Sustainable Development research clusters at UCL STEaPP
Funded by a UKRI Future Leaders Fellowship (MR/V024426/2), this project bridges mathematical and social sciences and seeks to understand and improve the way that mathematical models are used to support decision making and public policy. Here are a few examples of the kinds of things I have worked on:
Mathematical methods for model-based science. When “all models are wrong”, we need to ensure that analysis methods take this into account. I am interested in forecast evaluation, model evaluation for extrapolatory contexts, decision-making under deep uncertainty (DMDU), expert judgement, ensemble weighting, data assimilation, emulation for policy support (with Steven Sperazza), combining statistical and dynamical perspectives and “Escaping from Model Land“. I convened a 2024 international conference on Mathematical Models for 21st Century Decisions at the ICMS which has sparked various follow up activities including a policy brief. This work is also relevant for the implementation of AI and machine-learning-based models as decision support tools, and autonomous systems.
Climate modelling strategies and climate information for decision support. Because there are so many climate decision questions, we need many different kinds of climate information. With Dr Marina Baldissera Pacchetti, David Stainforth and others, I am exploring how we can redesign climate science to more effectively support the information needs of society. Nadia Leonova is looking at multihazard models for disaster risk reduction.
Supporting quality climate information in the finance/insurance sector. With Dr Marina Baldissera Pacchetti, Conor Lamb and others, I am investigating how the finance and insurance sector actually use models and the ways in which organisations, central banks and regulators can ensure that they are basing decisions on credible and trustworthy information. With the Willis Research Network and the London Mathematical Laboratory, I apply these ideas to catastrophe model evaluation.
Anticipatory humanitarian action. I have ongoing collaborations in the humanitarian sector and have supported development of anticipatory action for heatwave and cyclone. I am now particularly interested in financing mechanisms, basis risk, and the interaction of the humanitarian sector with insurance and reinsurance facilities. In 2025 I joined the Start Ready Governance Committee which provides technical and financial oversight to the Start Ready risk pool.
Interdisciplinary connections. Since my work crosses multiple application areas and disciplinary boundaries, I have an interesting perspective on the development and use of methods in different contexts. For example, I was invited to join the Lancet Commission on Strengthening the Use of Epidemiological Modelling to contribute this independent perspective on model use and evaluation (despite not having a background in health science). I have convened cross-disciplinary discussions and roundtables for public sector organisations to connect with relevant science.
Science advice to public bodies. As a member of the DESNZ Science and Technology Advisory Council, I offer independent expert advice and challenge to the department and help them to make the best use of up to date science to inform policy and strategy. I also sit on the Met Office Hadley Centre Science Review and Advisory Group, advising on the quality, policy relevance, and strategic direction of the world-leading Hadley Centre Climate Programme.
Ethics, values and responsible modelling. Philosophical aspects of values in science, particularly as they become embedded in mathematical models, have been of interest to me for a long time. This theme is built into all of the above work. I am supporting various workshops on aspects of responsible modelling. Anna Hardisty is looking at values and uncertainty in the construction of an early warning system for North Atlantic climate tipping points.