Using modelled relationships and satellite observations to attribute modelled aerosol biases over biomass burning regions

Biomass burning (BB) is a major source of aerosols that remain the most uncertain components of the global radiative forcing. Current global models have great difficulty matching observed aerosol optical depth (AOD) over BB regions. A common solution to address modelled AOD biases is scaling BB emissions. Using the relationship from an ensemble of aerosol models and satellite observations, we show that the bias in aerosol modelling results primarily from incorrect lifetimes and underestimated mass extinction coefficients. In turn, these biases seem to be related to incorrect precipitation and underestimated particle sizes. We further show that boosting BB emissions to correct AOD biases over the source region causes an overestimation of AOD in the outflow from Africa by 48%, leading to a double warming effect compared with when biases are simultaneously addressed for both aforementioned factors. Such deviations are particularly concerning in a warming future with increasing emissions from fires. Error attribution based on modelled relationships and satellite observations suggests that errors in global models are more important and require more concerns than emission errors in creating the overall uncertainties for biomass burning aerosols.

Automated summary

Biomass burning is a major source of uncertainty in global radiative forcing, and modeling the aerosol optical depth (AOD) over biomass burning regions is challenging. This study finds that AOD biases in aerosol modeling are primarily caused by incorrect lifetimes and underestimated mass extinction coefficients, which in turn are related to incorrect precipitation and particle size underestimation. Increasing biomass burning emissions to correct AOD biases leads to overestimation of AOD in outflow from Africa, resulting in a double warming effect. Error attribution based on model relationships and satellite observations suggests that errors in global models are more important than emission errors in creating overall uncertainties for biomass burning aerosols.