Devolution of agri-environmental policy for enhancing nature and achieving net zero
Background
Land-use policy must reconcile food production with net zero, biodiversity recovery and water-quality goals, often on the same land. Environmental decision-support research combines spatial models of carbon, water and biodiversity with optimisation, economics and uncertainty quantification to identify where land-use change can deliver the greatest benefits and how incentives can encourage it. Recent advances allow linked environmental outcomes to be modelled at field or parcel scale, bringing quantitative evidence much closer to real land-management decisions. Yet most approaches still treat decisions as occurring at a single level: either a land manager chooses what to do where, or a national policymaker sets incentives and constraints. In practice, decisions are distributed across national, regional and local institutions. New methods are needed to connect these scales while retaining spatial detail, multiple environmental objectives and explicit treatment of uncertainty.
PhD project
Achieving net zero, nature recovery and improved water quality will require major changes in how land is managed, but responsibility for delivering these outcomes is increasingly distributed across different levels of decision making. This PhD will ask how national environmental objectives and budgets should be translated into regional and local decisions, and how those decisions should adapt when their environmental and economic consequences are uncertain.
The student will build on high-resolution systems models that predict the carbon, biodiversity and water-quality consequences of land-use change at field-parcel scale. A central challenge is to connect decisions across scales: how should a national budget be allocated between regions with different opportunities and risks, and how should regional decision makers then set incentives, constraints or local targets so that their combined actions deliver national goals? Because direct optimisation using detailed environmental models is computationally prohibitive, the student will develop new machine-learning and uncertainty-quantification methods, together with decision-support and optimisation tools, to make these nested policy questions tractable.
The project sits directly at the interface of data science and decision science central to UNRISK, with uncertainty quantification a key component of the decision problem. The student will work with policymakers to identify realistic governance structures, constraints and policy levers, and test the resulting tools on live land-use policy questions. Co-supervision by an environmental economist will bring expertise in incentives, behaviour and policy design. The work offers scope for methodological advances in machine learning and decision-making under uncertainty, alongside developing tools with direct relevance to government and other organisations managing land across multiple spatial and organisational scales.
Applicant Profile
This project will be suitable for students with a strong quantitative background in statistics, computer science or economics and who want to apply these skills to solve real world decision problems and see those solutions impact the world.
