Potential Supervisors and Topics for DPhil Research

Supervisors: Professor Jim Hall and Professor Yee Whye Teh

Management of natural resources, and building resilience to the impacts of climate change, depends critically on variability and extremes in the weather. Decisions as diverse as insurance contracts, designing renewable energy infrastructure and safeguarding food and water security, all depend upon quantifying the complexity of the weather. As infrastructure, food and trade systems span the entire planet, and are sensitive to the weather on a range of timescales, the information that is required to stress-test and optimise these systems is far from simple – it entails many variables (e.g. temperature, precipitation, wind, sunshine etc.) that vary in space and time.

So far, our capacity to learn and simulate all these interactions has been limited: today’s physics-based models of weather and climate have inevitable inaccuracies, whilst statistical methods are limited in the number of variables that can be estimated and their capacity to simulate realistic patterns of weather and climate change. Generative AI provides the transformative opportunity to learn realistic patterns from observations and physics-based computer models, and then generate simulations of synthetic weather that reproduce the statistical properties of the observations. Once a generative AI has been trained, the computational expense of simulating many realistic weather events is low, so vast numbers of events can be generated that span the range of possible spatial and temporal variations, enabling rigorous quantification of the likelihoods and characteristics of many complex and rare weather events, for use in risk analysis, stress-testing, and long-term planning. 

The aim of the proposed research is to develop a generative AI system to simulate synthetic extreme weather events over very large spatial (i.e. up to global) and temporal (i.e. up to multi-year) scales. 

Generative methods for multi-variate weather extremes are beginning to appear (Bhatia et al., 2021; Boulaguiem et al., 2022), with deep generative models, such as generative adversarial networks (GANs), showing particular promise (Bhatia et al., 2021; Girard et al., 2025; Lhaut et al., 2026; Wilson et al., 2022). GANs were the state of the art for image generation for several years, making them a natural fit for generating visually realistic spatial patterns. More recently, diffusion models have achieved significant advances in the perceptual quality of images and videos they can generate (Ho et al 2020, Song et al 2020), even though they tend to under-represent the tails of the distribution (Dhariwal & Nichol 2022, Ho et al 2022). 

Alison Peard, a researcher in the OPSIS group, recently published a GAN that simulates wind speed, precipitation, and atmospheric pressure during storms in the Bay of Bengal (Peard et al., 2026), overcoming some of the limitations encountered by previous researchers, to generate spatially coherent multi-hazard event ensembles that preserve both the marginal and joint distributions of the training data. This is a significant step which is attracting considerable attention, but greater challenges remain. Above all, neither this nor other generative models have yet been demonstrated for simulating the temporal dynamics as well as the multi-variate spatial patterns of extreme events. In many applications, it is very important to know how extreme events build up and evolve through time. Extending large-scale multi-variate generative models, which are already high dimensional, into the time domain represents a considerable challenge. 

This DPhil project will explore methods for generating more diverse and complex weather and climate extremes over space and time:

Time profiles conditional on the peak: A first strategy is to generate the temporal profiles for events, conditioned on the peak intensity of the extreme event. For example, take a map of peak windspeed along a hurricane track (e.g. as generated by HazGAN) and temporally downscale it to produce a time-dependent simulation of the hurricane event. This could be achieved by training on normalised events and learning key temporal parameters, such as rate of rise and decay, peak duration, and post-landfall weakening. These approaches lend themselves to relatively short-lived events with well-structured temporal evolution.

Generative video modelling: The next, more ambitious, step is to simultaneously model the whole spatial-temporal domain, harnessing developments in generative video modelling and treating the extreme event as a 4D video-like tensor of latitude, longitude, time and weather channels (such as wind, rain, pressure, soil moisture, or flood depth). This approach has recently had some success in tropical cyclone forecasting (Ren et al., 2025) and is required for prolonged (multi-week and beyond) events such as droughts. Video models can be orders of magnitude more computationally expensive than image models, with long spatial-temporal climate simulations approaching available compute limits; we will therefore explore more lightweight approaches such as latent video diffusion.

The project will involve adaptation and application of Machine Learning and Generative AI algorithms. It will suit students with a strong background in statistics, physics, mathematics or engineering. Students should be able to demonstrate aptitude for dealing with big datasets and simulation models, and enthusiasm to address real-world problems of great policy significance.

This project would be appropriate for students who have been accepted onto the ILESLA or Intelligent Earth Doctoral Training Partnership programmes. Successful UK applicants will be eligible for full or part funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford's several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment or the relevant DTP programme. 

References:

• Bhatia, S., et al. 2021 Exgan: Adversarial generation of extreme samples, pp. 6750–6758.
• Boulaguiem, Y., et al. 2022.  Modeling and simulating spatial extremes by combining extreme value theory with generative adversarial networks. Environmental Data Science 1, e5.
• Girard, S., et al. 2025.  HTGAN: heavy-tail GAN for multivariate dependent extremes via latent-dimensional control. International Journal of Computer Mathematics, 1–22.
• Lhaut, S., et al. 2026.  Simulation of multivariate extremes: A Wasserstein–Aitchison GAN approach. Extremes, 1–38.
• Peard, A., et al. 2026.  Simulating spatial multi-hazards with generative deep learning. Nat. Hazards Earth Syst. Sci. 26(4), 1663–1683.
• Ren, Z., et al.  2025.  Improving tropical cyclone forecasting with video diffusion models. arXiv preprint arXiv:2501.16003.
• Wilson, T., et al. 2022  DeepGPD: A deep learning approach for modeling geospatio-temporal extreme events, pp. 4245–4253.

Supervisors: Professor Jim Hall, Dr Yue Li and Tom Russell

Spatial economic development co-evolves with the provision of infrastructure – the development of businesses and housing needs to be served by utilities, whilst the provision of new infrastructure, such as new transport capacity, can stimulate economic development. These co-evolving processes have long been the subject of economic geography (Fujita et al., 1999, Marrewijk et al., 2009) which has developed theoretical models, including Land Use and Transport Interaction (LUTI) models (Wegener and Fürst, 1999) and Spatial Computable General Equilibrium (SCGE) models (Bröcker, 1998). These models have seen some application in practice, in particular for evaluation of the wider economic benefits of major transport investments such as the HS2 railway and the third runway at Heathrow Airport. However, these applied models are complex to parameterise and tend to be implemented by consultants for specific projects. 

A more empirical strand of work uses granular spatial data on businesses (e.g. employment, wages, output) and housing to interpret patterns of economic activity. Data-driven analysis has, for example, sought to understand pattens of unequal productivity in the UK. A few carefully constructed econometric studies have managed to identify significant relationships between transport infrastructure and spatial economic development, notwithstanding the multiple endogeneities (Dyckerhoff et al., 2025, Rietveld, 1994). 

The questions about economic development and infrastructure have intensified at a time of sluggish economic growth and productivity improvements in the UK and several other advanced economies. The accessibility and commuter journey times in several British cities are inferior to their European counterparts. Yet resources for infrastructure investment are constrained. There is localised evidence that development projects (e.g. business investment and housing) are being constrained because of inadequate capacity or connection delays for power and water utilities. At a time of rapid transformation in the electric power sector, and increasing climate impacts on water availability, traditional business-as-usual planning assumptions no longer apply. 

Given these urgent policy questions, there is renewed interest in spatial analysis of infrastructure needs and the contribution that infrastructure may make to regional economic growth. This is coming at a time when there is increasing spatial data to measure economic activity and mobility, from a variety of sources including mobile phone location data, smart card and station utilisation data, active travel and micro-mobility apps (uber, e-scooters, bikeshares), flooding zone warning apps (revealing inaccessibility), POI data (Google places, OpenStreetMap). 

This project will open up a new phase of research in spatial analysis of economic development and infrastructure, to answer the questions: 

1. What spatial data are most useful for predicting infrastructure needs? 
   1. Do these data also provide evidence regarding the relationship between infrastructure investment and economic development?

The research will entail a combination of data collection and model development. A wide range of public and private datasets will be considered, for example on real estate development, rents and vacancies. The spatial and temporal resolution of datasets may vary, so will need to be reconciled. Fortunately, within the OPSIS group we have extensive processed data on the transport network and mobility; we have expertise in the analysis of spatial datasets and close links with UK government organisations working on these problems. 

Model development could entail projections of economic activity, based on trends and covariates, and/or the development of models of transport flows based on analysis of origins, destinations and accessibility. Analysing how economic activity may be constrained by accessibility is complex, given the possibility of multiple transport modes and the various motives for mobility. The analysis will build on previous research within the OPSIS group (Li et al., 2026) and elsewhere (Bansal and Graham, 2023, Liu et al., 2026), which had developed several different metrics of accessibility to assess the transport-related constraints on economic activity. 

A distinctive feature of this analysis will be the spatial scale of assessment which will be for all of Great Britain (or possibly England). The aim will be to identify hotspots of growth in infrastructure needs, and locations where growth is constrained by lack of infrastructure. Ideally the research will provide quantified evidence of the growth-inducing potential of infrastructure investments across Britain. 

To achieve these objectives, this project will suit students with any quantified background, including economics, engineering, mathematics and physical science. Ideal candidates should be able to demonstrate their capabilities in geospatial data discovery and analysis, computer modelling, and enthusiasm to address interdisciplinary problems of direct policy relevance in the UK. The project would be suitable for support from the ESRC Grand Union Doctoral Training Partnership. Successful UK applicants may be eligible for full or partial funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford’s several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Bansal, P., & Graham, D. J. (2023). Congestion in cities: Can road capacity expansions provide a solution?. Transportation Research Part A: Policy and Practice, 174, 103726.
• Bröcker, J. (1998). “Operational Spatial Computable General Equilibrium Modeling.” The Annals of Regional Science, 32, 367–387.
• Dyckerhoff, H., Hörcher, D., & Graham, D. J. (2025). Heterogeneous radial commuting bottlenecks with agglomeration economies: Methods and calibration for Bogotá. Transportation Research Part A: Policy and Practice, 192, 104331.
• Fujita, M., Krugman, P. and Venables, A.J., 1999. The spatial economy: Cities, regions, and international trade. The MIT press.
• Li, Y., Pant, R., Russell, T., Thomas, F., Hall, J. W., Oldham, P., ... & Young, P. J. (2026). Stress-testing road network resilience using counterfactual flood events (1953–2024) in Great Britain. Transportation Research Part D: Transport and Environment, 155, 105292.
• Liu, X., Hickman, R., Cao, M., & Tao, Y. (2026). Can transport infrastructure really facilitate economic activity concentration? A spatial and temporal spillover perspective. Transportation Research Part A: Policy and Practice, 204, 104831.
• Brakman, S., Garretsen, H and Van Marrewijk, C., (2009) The new introduction to geographical economics, Cambridge University Press
• Rietveld, P. (1994). Spatial economic impacts of transport infrastructure supply. Transportation Research Part A: Policy and Practice, 28(4), 329-341.
• Wegener, M. & Fürst, F. (1999). Land-Use Transport Interaction: State of the Art. Dortmund: Institut für Raumplanung, Universität Dortmund. 

Supervisors: Professor Jim Hall and Dr Raghav Pant

Rapid urbanisation, population growth and economic development in Africa means that demand for infrastructure, including energy, transport and water, are rapidly increasing. The pace of infrastructure development in Africa is significant, but infrastructure needs still far exceed the resources that are available for infrastructure investment. This intensifies the pressure to ensure that scarce resources are allocated efficiently. 

Meanwhile, the impacts of climate change on infrastructure in Africa are intensifying. Flooding and lack of drainage capacity impact roads and transport networks in both urban and rural areas (Douglas et al., 2008; Espinet and Rozenberg, 2018). Renewable and fossil fuelled energy supplies and transmission networks are impacted by storms (Thomas et al., 2026), whilst water supplies are threatened by droughts (Ilyas et al., 2014). 

The most cost-effective time to enhance the resilience of infrastructure is when it is being planned, designed and built. Yet too often, the impacts of climate change are still not being factored into infrastructure plans and investments. This may be due to lack of tools to map climate risks, or inadequate methodology to quantify the benefits of infrastructure resilience. These benefits go beyond avoided damages to assets, and include avoided business disruptions and supply chain losses (Koks et al., 2019; Pant et al., 2022).   

Fortunately, there have been rapid advances in the availability of data and spatial tools for climate risk analysis and resilience appraisal. Within the OPSIS group, we have developed the African Transport Systems Database (AfTS-Db) (Colombo et al., 2025) which is a complete geospatial dataset of multi-modal transport networks across the entire continent. We have modelled shipping trade and future needs for port expansion across Africa (Verschuur et al., 2026). We are also developing a global geospatial mapping of sectoral economic activity (Schlosser et al., 2026).

This DPhil project will develop a systematic and comprehensive methodology for: 

1. Predicting future infrastructure investment needs and their locations and connectivity of existing networks, across the continent of Africa. 
2. Analysing climate-related risks to these future investments. 
3. Prioritizing adaptations of new infrastructure investments, based on their benefits for climate risk reduction. 

To achieve these aims will involve developing methodology for projecting future investment needs, based on spatial demographic and economic trends (e.g. urbanisation) and evaluating service delivery gaps in current infrastructure provision. A rationale for prioritization of interventions will be developed. Their potential exposure to a range of climate-related hazards will be analysed, to calculate baseline risks without adaptation. A range of resilience adaptations will be explored and prioritized.

The research will respond to urgent needs for transparent methods to prioritize resilience interventions in future infrastructure in Africa, based on openly available datasets. There is considerable interest in such methods, from organisations including the World Bank and African Development Bank as well as national governments. The research will develop novel methodology and datasets, whilst also addressing a growing development need. 

The project will suit students with a strong quantified background (e.g. engineering, economics, physics) but also a good appreciation of the wider societal context of infrastructure service provision and climate risks.

Candidates for this project from an engineering, mathematics or physical sciences background would be eligible to apply for funding from Oxford University's EPSRC Doctoral Training Partnership. Successful UK applicants will be eligible for full or part funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford's several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Colombo, S., Pant, R., Young, M., Thomas, F., Russell, T., Verschuur, J. and Hall, J.W. The African Transport Systems Database - a geospatial database of multi-modal connected networks, Scientific Data, 13(2026): 166. DOI: 10.1038/s41597-025-06483-7
• Douglas, Ian, Kurshid Alam, Maryanne Maghenda, Yasmin Mcdonnell, Louise McLean, and Jack Campbell. "Unjust waters: climate change, flooding and the urban poor in Africa." Environment and urbanization 20, no. 1 (2008): 187-205.
• Espinet, Xavier, and Julie Rozenberg. Prioritization of climate change adaptation interventions in a road network combining spatial socio-economic data, network criticality analysis, and flood risk assessments. Transportation Research Record 2672.2 (2018): 44-53.
• Koks, Elco, Raghav Pant, Scott Thacker, and Jim W. Hall. Understanding Business Disruption and Economic Losses Due to Electricity Failures and Flooding. International Journal of Disaster Risk Science 10, no. 4 (2019): 421-438.
• Masih, Ilyas, Shreedhar Maskey, F. E. F. Mussá, and Patricia Trambauer. A review of droughts on the African continent: a geospatial and long-term perspective. Hydrology and earth system sciences 18, no. 9 (2014): 3635-3649.
• Pant, Raghav. Advances in climate adaptation modeling of infrastructure networks.In Climate Adaptation Modelling, pp. 159-167. Cham: Springer International Publishing, 2022.
• Verschuur, J., Fernandez-Perez, A., Martinez, L., Koks, E. and Hall, J.W. Climate risks to port infrastructure for future global trade, Nature Climate Change, in review.

Supervisors: Professor Jim Hall and Dr Yue Li

As climate change intensifies, traditional centralised infrastructure systems are increasingly vulnerable to disruptions from extreme weather events and long-term environmental shifts (Adapt, G. C. A., 2019). This project aims to explore decentralised planning as a transformative strategy that offers flexibility and resilience in infrastructure development (Hoffmann et al., 2020). By focusing on multiple sectors, such as water, energy, transportation, and telecommunications, the project seeks to identify how decentralised solutions can be integrated into existing systems to reduce the dependency on large-scale centralised infrastructures while creating more locally tailored, responsive solutions that bolster climate resilience and enhance sustainability. 

A key area of investigation will be how decentralised technologies can address climate challenges specific to individual sectors. For instance, in the water sector, decentralised water technologies such as rainwater harvesting, greywater recycling and water reuse, can contribute to water conservation and reduce the pressure on centralised water systems, ensuring safe access to all (including places sitting far and small islands, etc) during droughts and floods (Li et al., 2021); Similarly, distributed renewable generation and storage can transform energy system into active distribution networks, reducing reliance on large power grids and improving resilience during power outages (Wu et al., 2021). The project will also explore how cross-sector decentralised planning can enhance overall system resilience, particularly considering failure cascading effects (Pant et al., 2020). For example, decentralised energy solutions, such as solar microgrids, could support critical transportation and telecommunication infrastructure during climate hazard events, ensuring the continuity of services under adverse conditions.

In addition to enhancing resilience, decentralised solutions present opportunities for achieving sustainability goals. However, the adoption of these technologies must be done thoughtfully considering the potential increase in energy consumption (Li et al, 2021). Meanwhile, this project will explore how decentralised infrastructure planning can be optimised to balance access and equity, ensuring that vulnerable and underserved populations are not disproportionately affected by the transition to more decentralised systems. Moreover, it will address the trade-offs between different scales of decentralisation and assess how to optimise the design of technologies (such as type, location, and capacity) and integration with existing infrastructures across these different scales (Li et al, 2022). 

To achieve these objectives, this project will suit students with any quantified background, including environmental science, engineering, mathematics and physical science. Ideal candidates should be able to demonstrate their capabilities in geospatial data discovery and analysis, computer modelling, and enthusiasm to address interdisciplinary problems with system thinking. Candidates for this project from an engineering or physical sciences background would be eligible to apply for funding from Oxford University’s EPSRC Doctoral Training Partnership. Successful UK applicants will be eligible for full or partial funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford’s several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Adapt, G. C. A. (2019). A Global Call for Leadership on Climate Resilience. Global Center on Adaptation and World Resources Institute.
• Hoffmann, S., Feldmann, U., Bach, P. M., Binz, C., Farrelly, M., Frantzeskaki, N., ... & Udert, K. M. (2020). A research agenda for the future of urban water management: exploring the potential of nongrid, small-grid, and hybrid solutions. Environmental science & technology, 54(9), 5312-5322.
• Li, Y., Mo, W., Derrible, S., & Lu, Z. (2022). Integration of multi-objective spatial optimization and data-driven interpretation to direct the citywide sustainable promotion of building-based decentralized water technologies. Water Research, 222, 118880.
• Li, Y., Khalkhali, M., Mo, W., & Lu, Z. (2021). Modeling spatial diffusion of decentralized water technologies and impacts on the urban water systems. Journal of Cleaner Production, 315, 128169. 
• Pant, R., Russell, T., Zorn, C., Oughton, E., & Hall, J. W. (2020). Resilience Study Research for NIC–Systems Analysis of Interdependent Network Vulnerabilities. Environmental Change Institute, Oxford University.
• Wu, R., & Sansavini, G. (2021). Active distribution networks or microgrids? Optimal design of resilient and flexible distribution grids with energy service provision. Sustainable Energy, Grids and Networks, 26, 100461. 

Supervisors: Professor Jim Hall, Professor Michael Obersteiner, and Dr Raghav Pant

There is growing recognition of the risk from wildfires (Jones et al., 2020), which is a threatening dimension of climate-related risks (Zscheischler et al., 2018). A number of wildfire risk assessment frameworks have been developed, though these tend to have been focussed at national and regional scales (Fiorucci, 2008, Scott, 2013, Thompson, 2011, 2016). Global wildfire models have been established and wildfire has been incorporated in the land surface schemes of Earth system models (Krause et al., 2014). There are also growing spatial datasets of wildfire observations, with accompanying statistical analysis (Parisien and Moritz, 2009). However, so far there has been limited analysis of the potential systemic impacts of wildfires and they ways in which they may compound with other climate-related hazards, including heat waves and droughts.

This DPhil project will take a spatial systems approach to fully appraising wildfire risks and the ways in which they may propagate through society and the economy. We are not aiming to generate a new wildfire model – we will use model simulations and projections from other sources, including the Inter-Sectoral Impact Model Intercomparison datasets and other sources. We will combine these with wildfire observations, to build up a statistical picture of wildfire hazard and supplement that with future projections. That will provide a wildfire hazard layer and associated uncertainties. The main focus of the research will be to properly characterise the impact of wildfires on communities, infrastructure and the economy. The direct damages from wildfires are documented, at least in countries that have high levels of insurance coverage. However, there is very little knowledge of the wider effects of wildfire on economic activities and how quickly communities recover. Wildfire damage to infrastructure can have widespread impacts on people and economic activities. Repairing damaged infrastructure can be a considerable burden on public finances. This study will explore these and other dimensions of wildfire impacts and develop datasets and models for damage assessment, including both direct and indirect damages. We aim to do this in a way which is scalable to global scales, so that we can generate much improved global estimates of the risks from wildfires and estimates of how these risks may change in the context of climate change, land use change, economic development and different wildfire management strategies.

The project will involve a combination of statistical analysis and geospatial modelling. It will suit students with a strong background in environmental sciences, engineering or another quantified subject. Students should be able to demonstrate aptitude for quantified geospatial analysis, and enthusiasm to address real-world problems of great policy significance.

Candidates for this project from a natural sciences background would be eligible to apply for funding from Oxford University's NERC Doctoral Training Partnership. Successful UK applicants will be eligible for full or part funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford's several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Fiorucci, P., Gaetani, F. and Minciardi, R. 2008. Development and application of a system for dynamic wildfire risk assessment in Italy. Environmental Modelling & Software, 23: 690-702.
• Jones, M.W., Smith, A., Betts, R., Canadell, J.G., Prentice, I.C. and Le Quéré, C., 2020. Climate change increases the risk of wildfires. ScienceBrief Review, 116: 117.
• Krause, Andreas, et al. 2014. The sensitivity of global wildfires to simulated past, present, and future lightning frequency. Journal of Geophysical Research: Biogeosciences, 119.3 (2014): 312-322.
• Parisien, M-A, and Moritz, M.A. 2009. Environmental controls on the distribution of wildfire at multiple spatial scales. Ecological Monographs, 79.1 (2009): 127-154.
• Scott, J.H., Thompson, M.P. and Calkin, D.E. 2013. A wildfire risk assessment framework for land and resource management. General Technical Report RMRS-GTR-315. U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 83 p. doi: 10.2737/rmrs-gtr-315
• Thompson, M.P., Bowden, P., Brough, A., Scott, J.H., Gilbertson-Day, J., Taylor, A., Anderson, J. and Haas, J. R. 2016. Application of wildfire risk assessment results to wildfire response planning in the southern Sierra Nevada, California, USA. Forests 7, 64.
• Thompson, M.P., Calkin, D.E., Finney, M.A., Ager, A.A. and Gilbertson-Day, J.W. 2011 Integrated national-scale assessment of wildfire risk to human and ecological values. Stochastic Environmental Research and Risk Assessment, 25: 761-780.
• Zscheischler, J., Westra, S., Van Den Hurk, B.J., Seneviratne, S.I., Ward, P.J., Pitman, A., AghaKouchak, A., Bresch, D.N., Leonard, M., Wahl, T. and Zhang, X. 2018. Future climate risk from compound events. Nature Climate Change, 8(6): 469-477.

Supervisors: Professor Jim Hall , Dr Raghav Pant and Dr Edward Oughton

Infrastructure systems that deliver essential services to society (e.g. energy, water, transport and telecommunications) are increasingly regarded as being cyber-physical systems, as they are controlled by digital networks and depend upon software and digital communication systems. The risks to these systems have been widely studied, but from rather different perspectives. There has been extensive research, much of it by our group, on physical risks to infrastructure networks, with a focus on weather-related extremes (Koks et al., 2019, Lamb et al., 2019) but also including terrorist threats (Oughton et al., 2019). Meanwhile, there has been extensive research on questions of cyber security for infrastructure networks, for example relating to the security of the Internet of Things (IoT). Our aim in this project is to bring these perspectives together.

In the first instance the focus will be on modelling the networks of interdependent electricity and telecommunications systems. We have a fairly complete model of electricity transmission and distribution networks in Britain, and recently as part of research with the National Infrastructure Commission we coupled this with a representation of telecommunications networks in Britain.

The DPhil project will involve modelling of electricity and digital communications networks (including SCADA systems), which we will seek to validate with data on faults in the electricity and telecommunications networks. This will be used to model possible interdependent and cascading failures. The analysis will be used to identify how these interdependent networks can be made more resilient. For example, what is the potential benefit of increased connectivity or backup capacity within the network? We also wish to examine how technological trends (like electrification of transport and the proliferation of renewable energy supply technologies) could impact the resilience of infrastructure networks.

The project will therefore involve using and adapting existing simulation models and creating new models of infrastructure systems and development of methods for vulnerability analysis and optimisation. It will suit students from any quantified background, including engineering, mathematics and the physical sciences. Students should be able to demonstrate aptitude for computer modelling and enthusiasm to address real-world problems of great policy significance.

Candidates for this project from an engineering of physical sciences background would be eligible to apply for funding from Oxford University's EPSRC Doctoral Training Partnership. Successful UK applicants will be eligible for full or part funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford's several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Koks, E., Pant, R., Thacker, S., Hall, J.W. 2019. Understanding business disruption and economic losses due to electricity failures and flooding, International Journal of Disaster Risk Science, 10: 421-438. doi:10.1007/s13753-019-00236-y
• Lamb, R., Garside, P., Pant, R. and Hall, J.W. 2019. A network-scale analysis of the risk of railway bridge failure from scour during flood events in Britain. Risk Analysis, 39(11): 2457-2478. doi: 10.1111/risa.13370
• Oughton, E., Ralph, D., Leverett, E., Pant, R., Thacker, S., Hall, J.W., Copic, J., Ruffle, S. and Tuveson, M. 2019. Stochastic counterfactual analysis for the vulnerability assessment of cyber-physical attacks on electricity distribution infrastructure networks, Risk Analysis, 39(9): 2012-2031. doi: 10.1111/risa.13291
• Thacker, S., Barr, S., Pant, R., Hall, J.W., and Alderson, D. 2018. Geographic hotspots of critical national infrastructure. Risk Analysis, 11(1): 22-33. doi: 10.1111/risa.12840.
• Thacker, S., Kelly, S., Pant, R. and Hall, J.W. 2018. Evaluating the benefits of adaptation of critical infrastructures to hydrometeorological risks. Risk Analysis, 38(1): 134-150. doi: 10.1111/risa.12839.
• Thacker, S., Hall, J.W. and Pant, R. 2018. Preserving key topological and structural features in the synthesis of multi-level electricity networks for modeling of resilience and risk. Journal of Infrastructure Systems, ASCE, 24(1): 04017043. doi: 10.1061/(ASCE)IS.1943-555X.0000404
• Thacker, S., Pant, R. and Hall, J.W. 2017. System-of-systems formulation and disruption analysis for multi-scale critical national infrastructures, Reliability Engineering and Systems Safety, 167: 30-41. doi: 10.1016/j.ress.2017.04.023

Supervisors: Professor Jim Hall and Dr Raghav Pant

Increasing frequency and severity of heatwaves is one of the most obvious consequences of climate change. This will impact people directly and will also have effects upon the built environment and infrastructure upon which we depend. There has been quite extensive research upon the impact of extreme heat on buildings and human comfort/health (e.g. Jenkins et al., 2014a), and analysis of specific instances of the impact of extreme heat on infrastructure systems, e.g. rail buckling (Mulholland and Feyen, 2021), road surfaces (Smoyer-Tomic et al., 2003), power transmission (Cadini et al, 2017) and underground metro systems (Jenkins at al., 2014b). However, there has not been a full global analysis of the risks of extreme temperatures for infrastructure networks. This is now becoming possible thanks to the assembly of global datasets of infrastructure assets and networks, assembled by the OPSIS team (Koks et al., 2019) and others. The research will combine (i) climatic information on temperature extremes (ii) geospatial analysis of the exposure of infrastructure assets and networks (iii) assessment of the vulnerability of these networks to temperature extremes, and (iv) quantification of the impacts of temperature-related infrastructure failures.

For the analysis of temperature extremes, particular attention will be paid to the spatial extent of heatwaves, given their potential to impact large areas at the same time. Statistical analysis of temperature extremes will use weather station observations and reanalysis data to assemble a global spatial catalogue of heatwave events. It may be possible to combine this analysis with climate attribution studies (Perkins-Kirkpatrick and Lewis, 2020) to understand the non-stationarity in the temperature record. Future spatial heatwave projections will be obtained from climate model outputs (e.g. CMIP6, CMIP7). To obtain a large ensemble of spatial heatwave events, it may be necessary to develop a statistical model of spatial extremes.

Analysis of the exposure of infrastructure assets will be based upon OPSIS’s ongoing activities to assemble global infrastructure asset and network data. A review of temperature vulnerability functions will cover mechanisms including: road/runway surface softening, railway line buckling, overheating of mechanical and electrical equipment, transmission line sagging, etc. Where appropriate these functions will be modified to reflect local conditions, e.g. local temperature rating/standards of equipment.

The research will also examine the impacts of heatwaves on infrastructure demand, notably on electricity demand for cooling. This can place additional demand upon electricity supplies and power networks which combine with other heatwave effects to result in systemic failures. Given satisfactory progress, it may also be possible to examine coincident risks e.g. wildfires and droughts, and their impacts on cooling water availability, hydropower and inland waterway navigation.

The research will quantify the impacts of heatwave-related failures in terms of the numbers of infrastructure users who are impacted. We will also seek to quantify the economic impacts of heatwave-related infrastructure failures. There have already been several studies of the impacts of heat on economic production (Burke et al., 2015) but none of these have sought to disentangle the specific effects of infrastructure failures in heatwave. The output will be a global analysis of hotpots of infrastructure heatwave vulnerability, taking into account the degree of local adaptation, and probabilistic quantification of the scale of potential disruptions at present and in future scenarios.

Finally, the research will examine the potential benefits of future adaptations, examining a range of adaptation options, which may be applied to existing infrastructure or incorporated when infrastructure is replaced or upgraded.

The project will involve a combination of geospatial analysis and spatial statistics. It will suit students with a strong background in engineer, physics or another quantified subject. Students should be able to demonstrate aptitude for computer modelling and geospatial analysis, and enthusiasm to address real-world problems of great policy significance.

Candidates for this project from an engineering of physical sciences background would be eligible to apply for funding from Oxford University's EPSRC Doctoral Training Partnership. Successful UK applicants will be eligible for full or part funding. Overseas applicants in need of financial support are encouraged to apply for one of Oxford's several doctoral scholarship schemes for UK or overseas students. Closing dates apply on these schemes and students are encouraged to apply early. Applications are made through the School of Geography and the Environment.

References:

• Burke, M., Hsiang, S.M. and Miguel, E. 2015. Global non-linear effect of temperature on economic production. Nature, 527(7577): 235-239.
• Jenkins, K., Hall, J.W., Glenis, V., Kilsby, C.G., McCarthy, M., Goodess, C., Smith, D., Malleson, N. and Birkin, M. Probabilistic spatial risk assessment of heat impacts and adaptations for London. Climatic Change, 124(1-2): 105-117.
• Jenkins, K., Gilby, M., Hall, J.W., Glenis, V., Kilsby, C.G. Implications of climate change for thermal discomfort on underground railways. Transportation Research Part D: Transport and Environment, 30: 1-9.
• Cadini, F., Agliardi, G.L. and Zio, E., 2017. A modeling and simulation framework for the reliability/availability assessment of a power transmission grid subject to cascading failures under extreme weather conditions. Applied energy, 185: 267-279.
• Koks, E.E., Rozenberg, J., Zorn, C., Tariverdi, M., Vousdoukas, M., Fraser, S.A., Hall, J.W., Hallegatte, S. 2019. A global multi-hazard risk analysis of road and railway infrastructure assets. Nature Communications, 10(1): 2677. doi: 10.1038/s41467-019-10442-3
• Mulholland, E. and Feyen, L. 2021. Increased risk of extreme heat to European roads and railways with global warming, Climate Risk Management, 34: 100365
• Perkins-Kirkpatrick, S.E. and Lewis, S.C., 2020. Increasing trends in regional heatwaves. Nature communications, 11(1): 3357.
• Smoyer-Tomic, K.E., Kuhn, R. and Hudson, A. 2003. Heat wave hazards: an overview of heat wave impacts in Canada. Natural hazards, 28: 465-486.

Supervisor: Professor David Thomas

The dry valleys of the Namibian Skeleton Coast are significant emitters of atmospheric mineral dust. Preliminary studies show that Holocene valley fills are the primary source of dust (Thomas et al., in press). There are however distinct spatial variations in both dust sources, revealed by remote sensing analysis, and the accumulation of fluvial valley deposits. The latter are a potentially significant source of Holocene palaeoclimatic data, but disparate studies to date reveal complex and sometimes conflicting records of climate change (Stone and Thomas, 2013). The aim of this project is to provide a robust, chronometrically controlled, record of extensive valley fill deposits and past fluvial system behaviour. This will involve examining the spatial variability in deposit characteristics and age that will allow potential dust emission sources and controls to be identified. The initial focus will be on the Huab Valley system, but this is extendable to other valleys as results emerge. The project will include fieldwork in the Skeleton Coast region that will examine and map sedimentary exposures and drill fill deposits, allowing sampling for sedimentary analyses including mineral properties, provenance work and age control through OSL dating. This project offers a unique opportunity to develop and apply field and laboratory skills, and will make a major contribution to southern African Quaternary research and investigations of atmospheric dust emissions, both major research themes in the School.

Supervisor: Dr Ben Caldecott

• Transition finance principles. Stakeholders, products, incentives, metrics and targets
• The risk of stranded assets in food infrastructure from the transition to sustainable diets
• Stranded assets in downstream SMEs in Thailand: assessing risks and resilience
• The Future of Engagement: An investigation into optimal engagement practices and the potential for alpha generation
• Are primary market transactions allocating capital to companies that are aligned to the Paris Agreement?
• Paris aligned portfolios and the real economy
• Shaping a resilient green bond market: non-financial disclosure, public issuance and climate risk management

Supervisors: Dr Ben Caldecott and Professor Doyne Farmer

• Climate risk scenario analysis and the role of central banks and supervisors

Supervisors: Dr Ben Caldecott and Professor Cameron Hepburn

• Motives beyond markets: State ownership, stranded assets, and decarbonisation in the power sector
• Quantifying barriers to the power generation sector’s low-carbon transition using machine learning and asset-level datasets.
• Technology, information, and the governance of environmental risk