Exploring biases in brown carbon model representation with in-situ flight observations

Abstract. The inclusion of brown carbon (BrC), the subset of organic aerosols (OAs) that absorb light in the near UV to visible wavelength range, in climate models is necessary to represent an abundant atmospheric constituent and to improve OA radiative forcing estimates. However, the large degree in variability of laboratory and field measurements of BrC properties makes model representation difficult. We utilized in-situ observations of BrC absorption from the DC3, SEAC⁴RS, ATom, WE-CAN, and FIREX-AQ flight campaigns to evaluate the GISS ModelE Earth system model BrC scheme. We focused on measurements influenced by biomass burning (BB) and the region of temperate and boreal North America. Average vertical profile comparisons of measured versus simulated absorption, corrected for model biases in carbon monoxide and black carbon (commonly co-emitted with BrC) concentrations, revealed a systematic underestimation in modelled BrC absorption. To explore possible causes of this bias, we evaluated the model's BrC-to-OA relationship. Sensitivity tests were run to determine if parameter changes could improve model performance and therefore substantiate the potential causes of bias identified. Increasing model organic aerosol-to-organic carbon (OA-to-OC) mass ratio for BB and aged OAs greatly improved alignment to measured OA-to-OC but decreased OA burden and increased BrC bias in the upper troposphere. Decreasing wet removal of BrC appeared to partially address the bias in aged air masses and fire plumes, while other potential BrC-specific biases, like missing secondary sources, could not be substantiated. Based on this, we highlight BrC processes that require further research and future directions for model development.