Clear-Sky Direct Aerosol Radiative Forcing Uncertainty Associated with Aerosol Vertical Distribution Based on CMIP6 models

• Published in: Journal of climate Vol. 35; no. 10; pp. 3021 - 3035
• Main Authors: Zhang, Lu, Li, Jing, Jiang, Zhongjing, Dong, Yueming, Ying, Tong, Zhang, Zhenyu
• Format: Journal Article
• Language: English
• Published: Boston American Meteorological Society 15.05.2022
• Subjects: Aerosol layers; Aerosol optical depth; Aerosol vertical distribution; Aerosols; Albedo; Anthropogenic factors; Clear sky; Climate models; Distribution; Energy balance; Global aerosols; Intercomparison; Mass; Meteorological satellites; Modelling; Normalizing; Optical analysis; Optical thickness; Perturbation; Radiative forcing; Radiative transfer; Uncertainty; Vertical distribution; Vertical profiles
• ISSN: 0894-8755; 1520-0442
• DOI: 10.1175/JCLI-D-21-0480.1

Abstract The direct perturbation of anthropogenic aerosols on Earth’s energy balance [i.e., direct aerosol radiative forcing (DARF)] remains uncertain in climate models. In this study, we investigate the uncertainty of DARF associated with aerosol vertical distribution, using simulation results from 14 global models within phase 6 of the Coupled Model Intercomparison Project (CMIP6). The column mass loading for each aerosol species is first normalized to the multimodel average for each model, which is called the mass-normalization process. The unified radiative transfer model and aerosol optical parameter are used, so that the differences in the calculated DARF are solely attributed to the difference in aerosol vertical profiles. The global mean DARF values in 2014 with respect to 1850 before and after mass normalization are −0.77 ± 0.52 and −0.81 ± 0.12 W m −2 respectively, assuming external mixing, which indicates that the intermodel difference in aerosol vertical distribution accounts for ∼20% of the total DARF uncertainty. We further conduct two separate experiments by normalizing aerosol optical depth (AOD) and aerosol single scattering albedo (SSA) profiles, respectively, and find that the vertical distribution of SSA results in larger DARF uncertainty (0.17 W m −2 ) than that of AOD (0.10 W m −2 ). Finally, compared with CALIPSO observation, CMIP6 models tend to produce higher aerosol layers. The bias in modeled aerosol profile with respect to CALIPSO leads to stronger DARF, especially for land regions.