Reducing the persistent uncertainty in how aerosols affect climate change

Aerosols from air pollution are the cause of the largest uncertainty in historical climate change. Changes in historical temperature are caused by a combination of greenhouse gas warming and aerosol-induced cooling. While the greenhouse effect is quite well quantified, the aerosol cooling effect (often known as global dimming) has been persistently uncertain in climate models for over two decades. This uncertainty is important because it hinders our ability to attribute historical climate change and to make reliable climate projections. The uncertainty in models has persisted despite huge increases in model complexity and available measurements to test the models.

In this project the student will use advanced sets of climate model simulations (ensembles) combined with extensive observations to tackle the uncertainty problem head-on. The ensembles, combined with model emulators, enable essentially millions of variants of the UK climate model to be created and analysed in a big data approach. This approach has been used in several previous projects, leading to a much clearer path to robustly reducing the uncertainty.

The project will use combinations of advanced high-resolution regional models and global climate models to determine the key aerosol and cloud processes that contribute most to the uncertainty. Observations from aircraft, surface sites and satellite will then be used to constrain the simulations. The ultimate aims are to 1) determine the “maximum feasible reduction” in uncertainty that can be achieved with a modern climate model and current observations; 2) produce a climate model that simulates the aerosol effects on climate with error bars that can be robustly traced to the driving processes and observational constraints. If this can be achieved, then the results will have a substantial impact on the reliability of climate projections.

The key research challenges of the project include: 1) Exploring the very large model datasets to diagnose model deficiencies, such as poorly represented processes; 2) Ways to exploit high-resolution regional model ensembles to improve the global climate model; 3) Understanding how uncertainties in aerosol effects vary regionally and on different time horizons.