The world is heating up and rainfall patterns are shifting. Yet our ability to prepare is severely hindered because the hydrological predictions of advanced Earth system models (ESMs) come with large uncertainties, especially over equatorial regions (e.g. Chadwick et al 2017). Newer model generations, including higher resolution configurations, have shown only marginal improvements and the out-of-sample problem complicates AI approaches. However, observations of rainfall during warm intervals of the geological past can provide invaluable context (e.g. Tierney et al 2020) either to discern between ESMs or improve them. The mid-Pliocene (~3.2 million years ago) when atmospheric carbon dioxide levels last rose to today’s level is a key testbed. ESM simulations of the Pliocene reconstruct a wetter North Africa with potential vegetation expansion across the Sahara.
Our confidence in this result is undermined because the same ESMs, completely fail to reproduce well documented cyclical wetting and greening of the Sahara in response to precession (insolation), most recently during the mid-Holocene (6 to 9 thousand years ago: Claussen et al. 2017). A recent breakthrough in ’paleo-conditioning’ has allowed one ESM to reproduce these transitions under a range of Pleistocene conditions (Hopcroft & Valdes, 2021; Armstrong et al 2023), but this same configuration fails to green under Pliocene background conditions, contradicting well-resolved records (e.g. Crocker et al 2022).
This project will address this major discrepancy by integrating new paleo- records and ensemble simulations with the flagship UKESM for a selection of key timeslices comparing the late Pleistocene and Pliocene.
This project is not suitable for CASE funding
Each host has a slightly different application process.
Find out how to apply for this studentship.
All applications must include the CENTA application form.
Choose your application route
Newly assembled palaeo–records and UKESM simulations will allow for a separation of the orbitally- versus CO2-driven rainfall changes and you will use this to identify robust spatial and seasonal signals that can serve as key benchmarks for EMSs. The project will then expand on recent tuning methods (e.g. Hopcroft et al. 2021) to develop a Bayesian and/or machine-learning approach to efficiently tune the parameters in the land-only component (JULES) and/or in UKESM. The aim is a model that it is able to reproduce past events much more accurately. This model configuration will then be used to repeat future scenarios and test the impact of learning from the geological past on future predictions.
DRs will be awarded CENTA Training Credits (CTCs) for participation in CENTA-provided and ‘free choice’ external training. One CTC can be earned per 3 hours training, and DRs must accrue 100 CTCs across the three and a half years of their PhD.
During this project the student will have training in analysing and evaluating climate change simulations and in the associated computer coding skills. Prior experience of computing or coding would be helpful but is not necessary. The project would ideally suit physics, maths, engineering, geography, environmental sciences or meteorology graduates, but motivated applicants with any relevant background are strongly encouraged.
The student will have the opportunity to visit collaborators in Southampton, the Met Office and elsewhere, and to present their research at conferences both within the UK and internationally.
This project will bring together researchers at the Universities of Birmingham (expertise in palaeoclimate modelling and atmospheric dynamics), Southampton (expertise in paleoclimate of North Africa and paleorecords), Dortmund, Germany (expertise in statistical modelling and model tuning) along the Met Office (expertise in model development and land-surface processes). This will expand on an existing Southampton-Birmingham NERC project. You will join a broader network of existing PhD students and post-docs working on related topics across institutions and will benefit from Birmingham’s membership of the Met Office Academic Partnership.
Year 1: setup and test UKESM under a range of palaeo boundary conditions and evaluate against available well-resolve records.
Year 2: Setup and run an ensemble of UKESM runs and develop existing tuning methodology to constrain key uncertain parameters related to clouds, convection and dynamic vegetation.
Year 3: Develop a final tuned configuration and re-run key simulations for the past and future.
Armstrong et al. (2023). North African Humid Periods over the past 800000 years, Nature Communications, 14, 5549.
Chadwick et al. (2016). Large rainfall changes consistently projected over substantial areas of tropical land. Nature Climate Change, 6:177–182.
Claussen, et al. (2017). Theory and Modeling of the African Humid Period and the Green Sahara. Oxford Research Encyclopaedia of Climate Science, 1–38.
Crocker et al. (2022). Astronomically controlled aridity in the Sahara since at least 11 million years ago. Nature Geoscience, 15:671–676. doi: 10.1038/s41561-022-00990-7.
Hopcroft et al. (2021). Using the mid-Holocene ’greening’ of the Sahara to narrow acceptable ranges on climate model parameters. Geophys Res Lett, 48(6):e2020GL092043.
Hopcroft & Valdes (2021). Paleoclimate-conditioning reveals a North Africa land-atmosphere tipping point. PNAS, 118(45): e2108783118.
Mitsunaga et al. (2025). Fundamentally unchanged northwestern African rainfall regimes across the Plio-Pleistocene transition. Science Advances, 11(eads3149).
Tierney, et al. (2020). Past climates inform our future. Science, 370 (6517).
Please contact Peter Hopcroft ([email protected]) or Ruth Geen ([email protected]) with any queries about this project.
To apply to this project:
Applications must be submitted by 23:59 GMT on Wednesday 7th January 2026.