AEROS: Aerosol model robustness and sensitivity study for improved climate and air quality predictio
The AEROS project is a collaboration between Leeds (Ken Carslaw) and Oxford (Philip Stier) with partners at NILU, AEROCOM and the Met Office. AEROS runs from 2009–2011.
Although aerosol models have developed considerably over the last 15 years, aerosol has persistently been assessed as the largest radiative forcing uncertainty (IPCC, 2007). International model intercomparisons have been very successful in quantifying aerosol model diversity. However, the attribution of differences to specific processes has been limited, so it has not been possible to identify and then reduce the key sources of model uncertainty.
In AEROS we will tackle the persistent aerosol model uncertainty in a new way. We will use 4 advanced 3-D regional and global aerosol models and a range of statistical tools and uncertainty analysis techniques to quantify the sources of model uncertainty at the process level and separate the uncertainties due to parameters and model structures. Model microphysical and radiative properties will be assessed at many levels of detail and spatial scale using a synthesis of aircraft, ground, remote-sensed and satellite observations from the most intensive aerosol field campaign ever conducted over Europe. The outcome will be a new understanding of model performance versus complexity, greatly improved confidence in aerosol model predictions, and an objective strategy for their future development.
The overall aim of AEROS is to take the quantitative uncertainty assessment of aerosol models to a new level and thereby substantially improve confidence in climate and air quality model predictions.
The specific objectives are to:
1. Develop techniques for sensitivity and uncertainty analysis of complex aerosol models and quantify the most important factors controlling prediction diversity and biases against observations.
A comprehensive uncertainty analysis will be unique in this field and will identify causes of error – a procedure lacking in previous model assessments. The techniques we develop will be transferred to other model assessments such as AeroCom.
2. Improve global and regional aerosol models through identification and attribution of deficiencies.
The establishment of multi-criteria, multi-parameter skill scores will advance aerosol model benchmarking and provide guidance for assessments such as AeroCom and IPCC.
3. Define an appropriate level of model complexity for a wide range of model structures and quantify the impact of structural choices on model predictions.
The outcomes will enable future model development to be prioritised. The project is a close collaboration with the Met Office so our findings will directly inform development of the UK Unified Model used for climate prediction, Earth System research and, in future, operational air quality prediction.
To meet these objectives we will:
4. Evaluate four 3-D aerosol microphysics models on a range of scales against synthesised aircraft, ground and remote observations from the EUCAARI and EMEP intensive campaigns. New techniques of multi-criteria assessment will be used to handle the large number of possible metrics.
The aim of such a comprehensive synthesis is to extract the maximum information for model improvement and minimise the risk of error compensation.