Computer climate models are the only way to predict future climate change and to understand the drivers of environmental change at the global scale. But these models do not always agree on regional changes, especially for monsoon regions. Part of this is because the gridcell resolution does not capture the full sub-grid dynamics of these systems and so paramerisations of for example convective storms are required (e.g. Bordoni et al 2025). It is increasingly being recognised that these parameterisations can be hugely improved through the addition of a stochastic or random component that is designed to represent small-scale features that occur in the real world but are weak or missing in gridded models (Palmer et al 2001; Wang et al 2016).
This can be partly understood by considering the famous butterfly effect – a tiny disturbance of the atmosphere in one region that has major knock-on effects elsewhere. This effect as stated cannot happen in global models because they lack the fine-scale disturbances. Stochastic parameterisations could go some way to resolving this and have already shown promise in improving the simulation characteristics of present-day rainfall (Wang et al 2016).
In this project you’ll build on these established schemes to add random (stochastic) perturbations to key parts of a climate model and use this to perform new simulations of past and future change. This work will allow us to assess the improvements relative to a standard climate model configuration and we’ll use deeper past changes (e.g. Hopcroft & Valdes, 2021) to gauge the response to a changed climate which is more relevant for testing the feedbacks that may come into play in a future warmer world.
This project does not offer a CASE studentship
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
In this project you will add a random or stochastic component to a version of the Met Office Unified Model (UM). In many different areas this has been shown to improve the upscale representation of the system because of the added perturbations which are meant to represent the small-scale features such as clouds. You will test the improved model against past evidence of climate change before also making new projections of near-future climate.
This project will be able to take advantage of the University of Birmingham’s state of the art BEAR high-performance computer which has over 40,000 cores. You will join an existing research group with significant expertise and experience of running coupled climate models and with a diverse supervisory team that can contribute expertise in paleoclimate modelling and data, climate dynamics and meteorology.
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.
DRs will receive training in running climate and Earth System models and in analysing and developing these codes. This will include attending the NCAS Unified Model users training course. Students are also able to attend MSc modules in atmospheric physics and dynamics and other courses run by the University of Birmingham’s BEAR covering high-performance computing, scientific R and/or python and more. Recent students have also attended a PhD summer school on climate change (e.g. Advanced Climate Dynamics course). You will also benefit from a long-term collaboration with the University of Southampton that includes joint PhD students and a recently-funded NERC research project.
Year 1: Develop a stochastic scheme within the climate model based on other recent examples (e.g. Wang et al 2016).
Year 2: Evaluate the new model against present day and past (paleo) climate observations (e.g. Hopcroft & Valdes, 2021).
Year 3: Test the model for future projections and assess how the processes and feedbacks differ.
Bordoni, S. et al. (2025). The futures of climate modeling, npj Climate and Atmos Sci 8(99).
Hopcroft, P. and Valdes, P. (2021). Paleoclimate-conditioning reveals a North Africa land-atmosphere tipping point, Proc Natl Acad Sci 118(45): e2108783118.
Palmer, T. (2001). A nonlinear dynamical perspective on model error: A proposal for nonlocal stochastic-dynamic parametrization in weather and climate prediction models, Q J Roy Met Soc 127(472): 279–304.
Wang, Y., Zhang, G. and Craig, G. (2016). Stochastic convective parameterization improving the simulation of tropical precipitation variability in the NCAR CAM5, Geophys Res Lett 43(12): 6612–6619.
For any queries about this project please feel free to contact Dr Peter Hocproft ([email protected]).
To apply to this project: