The influence of physical process representations on regional and global-scale climate model output University of Leeds in United Kingdom

Background

The influence of aerosol particles on Earth’s energy balance, remains a major uncertainty in climate science. Even with increased model complexity, resolution, and data availability, aerosol and cloud interactions remain difficult to constrain, leaving gaps in understanding. This uncertainty masks the amount of warming from greenhouse gasses and limits how precisely we can simulate future climate change.

Perturbed parameter ensembles (PPEs), a machine learning approach, are a powerful tool to explore model uncertainties. Dr Regayre and Prof. Carslaw have used PPEs of UK climate models to reveal which physical processes cause model uncertainty at the regional and global scales. Additionally, they have used these methods to inform strategies for improving physical processes representations, to further reduce the model uncertainty. Using the next generation of regional and global-scale climate models, this project will provide critical insight into next steps for model development.

PhD opportunity

This PhD project will examine how representations of physical process in climate models affect our understanding of climate change. This research will work with UKESM2, the UK’s flgship Earth System Model, to a) quantify and understand how improvements in representations of cloud microphysics and aerosol processes influence model uncertainty, b) identify which physical processes cause the most uncertainty in how aerosol particles affect regional and global-scale climate, and c) evaluate the impact this uncertainty might have on our trust in future climate projections. The project will benefit from training in model-to-observation comparison methods provided by colleagues at CICERO (Center for International Climate Research) in Norway.

Throughout the project, the student will steer research questions, making it possible to specialize in areas that match their interests. One research stream within the project could focus on ensemble comparisons, contrasting causes of uncertainty in large ensembles of UKESM2 with ensembles of other climate models. These comparisons could identify some of the reasons climate models differ so much in how they represent climate change. Analyzing the extent to which physical process representations in UKESM2 align with, or differ from, representations in other models will help pinpoint model development priorities.

Another project challenge could involve exploring how different physical process representations in regional and global-scale climate models affect model output. High-resolution regional models capture process-level information about aerosol particle and cloud microphysics, that influence how clouds respond to pollution. This additional detail complicates interpretation of global-scale climate models that have simpler representations of these physical processes. Understanding how regional-scale model uncertainties relate to global-scale uncertainties will allow us to better understand the impacts of man-made climate change.