New approaches to quantifying the uncertainty in aerosol radiative forcing of climate

Kirsty Pringle and Ken Carslaw (University of Leeds) and Nicolas Bellouin (Met Office CASE partner)

Enquiries: Kirsty Pringle K.Pringle(at)leeds.ac.uk or Prof Ken Carslaw lecksc(at)ds.leeds.ac.uk

Background

The uncertainty in climate model predictions limits our understanding of climate change and reduces the level of confidence with which results can be conveyed to policy makers and government. The Intergovernmental Panel on Climate Change (IPCC) reports that there are very large uncertainties in the radiative forcing due to aerosols, with the upper range being large enough to cancel the warming effect of greenhouse gases (see Figure 1, below).

However, the error bar on the diagram below shows only model diversity (the range of predictions of different models) and thus does not reflect the true uncertainty. This leaves certain key questions that are yet to be answered:

1. Why are the aerosol models so diverse?
2. How much would the range shrink if we were able to improve specific aspects of the models? 
3. What is the likelihood of the radiative forcing (and hence climate change) lying in a given range?

This is the kind of information that scientists, policy makers and the public increasingly want to get from climate scientists.  To address this issue, a new field within climate science is emerging which uses climate models combined with new statistical techniques to gather robust uncertainty information to be included in future assessment reports.

Gathering this uncertainty information is easy if the experiment can be performed many hundreds of times.  But doing this with a complex climate model is not so easy because the model takes a very long time to run.  New techniques of model emulation now allow us to extract the same key statistical information from far fewer model simulations. We recently demonstrated the use of emulation of a complex global aerosol model in Lee et al. (2011). In this PhD these new techniques will be applied to produce true error bars on model predictions of radiative forcing in a manner that is useful to both policy makers and government.

Research aims:

1)      To identify the factors responsible for radiative forcing uncertainty

What factors contribute most to the uncertainty in the direct and indirect aerosol radiative forcing? Are there some parts of the world where uncertainties are particularly large? If so, why? Where should we devote resources to improve the model confidence? 

2)      To quantify the importance of different pollutant sources to radiative forcing

How does the uncertainty in radiative forcing vary among the main pollution sources (e.g. shipping, industry, power production, agriculture)?  Each of these sources emits in different regions of the atmosphere and produces different types of aerosol, so the uncertainties in attributed forcing are likely to vary substantially. The results will allow the attribution of a climate sensitivity to each emission sector which will enable the development of more targeted emission reduction strategies. 

3)      To understand the role of natural emissions in radiative forcing

How do the uncertainties in anthropogenic forcing compare to those due to natural aerosols from e.g. forests and oceans?  The natural aerosol “baseline” is important as it controls the magnitude of anthropogenic forcing, but it is usually neglected as a model uncertainty.

Approach

The research will be based on the Met Office Hadley Centre General Environment Model (HadGEM) combined with the UK Chemistry and Aerosol Model. The uncertainty analysis will be performed using Gaussian emulation. We recently completed a study on parametric uncertainty in the model, and in this project the student will build on the expertise gained by the group during this work and apply similar techniques to the quantification of radiative forcing uncertainty.

Suitable candidates

 This PhD will be very suitable for students with strong quantitative skills. This background will allow the student to develop the statistical techniques and apply them to the complex models. It would also be possible to pursue this PhD with more of an interest in the implications of the uncertainty for climate prediction – using the statistical tools more as a black box. Students with a background in environmental sciences or any other quantitative science would therefore be suitable.

Role of the Met Office CASE partner

This project has been approved as a "potential CASE studentship" by the Met Office. Three Met Office CASE studentships will be awarded at the School of Earth and Environment in the 2012 competition, and will be assigned to those students whose applications are ranked highest.

The Met Office are developers of the UK Earth System Model HadGEM. Leeds has had a long cooperation with the Met Office in the development of the aerosol component of the model. The outcome of this project will be of great interest to the Met Office because it has the potential to alter their climate projections and affect the way they develop the model in future. If CASE is awarded, the student will have the opportunity to spend significant periods of time at the Met Office working with specialists in Earth system models.

Where will this PhD take me?

Uncertainty analysis is an emerging and very important area of climate science. Policy makers want reliable estimates of uncertainty and likelihood. This PhD will provide excellent training in this emerging research area and equip students with the skills needed for the next phase of climate prediction.

Supervisors and the research group at Leeds

Prof Ken Carslaw is the Principal Investigator of the aerosol part of the UKCA project and the Global Model of Aerosol Processes (researchpages.net/glomap). His group has published on a wide range of aerosol issues including atmospheric dust, marine aerosol, aerosol feedbacks, stratospheric aerosol, geoengineering, cloud processes, volcanic emissions and Arctic aerosol. Dr Kirsty Pringle is a leading developer of the GLOMAP model and has worked on climate models at the Met Office and at the Max Planck Institute in Germany.

References

L. A. Lee, K. S. Carslaw, K. Pringle, G. W. Mann, and D. V. Spracklen, Emulation of a complex global aerosol model to quantify sensitivity to uncertain parameters, Atmos. Chem. Phys. Discuss., 11, 20433-20485, 2011 www.atmos-chem-phys-discuss.net/11/20433/2011/

Forster, P., V. Ramaswamy, P. Artaxo, T. Berntsen, R. Betts, D.W. Fahey, J. Haywood, J. Lean, D.C. Lowe, G. Myhre, J. Nganga, R. Prinn, G. Raga, M. Schulz and R. Van Dorland, 2007: Changes in Atmospheric Constituents and in Radiative Forcing. In: Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change