Investigating the impact of regional and global constraints on reducing the uncertainty in aerosol effective radiative forcing

Despite substantial developments in global aerosol modelling and improved in-situ and remote sensing observational methods, reducing the persistent uncertainty in aerosol Effective Radiative Forcing (ERF) has proven to be one of the most challenging issues in atmospheric science (Textor et al, 2006).

The A-CURE project aims to understand and quantify the uncertainties in how changes in anthropogenic aerosols affect climate. This is achieved through comprehensive sampling of the uncertainty within a single model and challenging it with multiple observation types. In this work we constrain aerosol-radiation interaction (ERFari) in the HadGEM-UKCA global climate model using ground-based and satellite observations. We use a perturbed parameter ensemble (PPE) of 26 climate model parameters, including processes and emissions related to aerosols and clouds (Yoshioka et al., 2019), to explore the potential observational constraints. 235 simulations of the present-day and pre-industrial atmospheres simultaneously sample the uncertainty, and an implausibility metric (Johnson et al., 2019, in prep.) is used in retaining plausible parameter combinations that will eventually improve ERF estimates.

We first constrain the emulated PPE output using high-temporal resolution aerosol optical depth (AOD) and aerosol absorption optical depth (AAOD) measured with the AERONET sun-photometers, over multiple regions. We also apply constraints on a global scale, using AERONET AOD and AAOD, MODIS monthly mean AOD and black carbon (BC) in-situ aircraft measurements. Some parameter constraints are consistent across regions or observation type, while for others the constraining method retains different part of the parameter space. This emphasizes the need of regionally perturbing the emissions and investigating the aerosol refractive index, emissions and scavenging in different parts of the globe. The mean constrained ERFari changed from -0.75±0.29 Wm^-2 to -0.63±0.18 Wm^-2 by AERONET AOD and to -0.82±0.29 Wm^-2 by MODIS AOD. The AERONET AAOD and aircraft BC have a constraint on ERFari of -0.72±0.29 Wm^-2.

Future developments include combined constraints and an improved PPE that uses the observed differences between regional constraints and different types of observable to correct important model-observation biases.