Observation-constrained black carbon size and mixing state improve global simulations of aerosol optical properties
Abstract. Black carbon (BC) is among the strongest warming aerosols and exerts substantial climatic impacts, yet global atmospheric models employ oversimplified representations of BC mixing state and microphysical properties, which introduce substantial biases and uncertainties in simulated BC optical properties. Here we build an explicit chain from process-level optical simulations to observation-constrained model scheme development. Particle-resolved optical sensitivity experiments spanning wavelength, refractive index, core size, coating thickness rank the sensitivity of the BC mass absorption coefficient (MAC) to each microphysical attribute. The particle-to-BC diameter ratio (Dp/DBC) is identified as the dominant controllable attribute that a model can diagnose and that observations can constrain, and therefore as the natural lever for scheme improvement. Accordingly, two observation-constrained optical schemes are implemented in the Community Atmosphere Model version 6. The DBC-constrained scheme applies constraints on the volume-equivalent diameters of freshly emitted BC from fossil-fuel and biomass-burning sources. This adjustment reduces the global mean Dp/DBC from 6.3 to 2.1, consistent with field observations. Validated against AERONET retrievals, this scheme lowers the global normalized mean bias (NMB) of the single-scattering coalbedo from 36 % to −3 %, with regional reductions of 16 % ~ 70 % over the BC source regions. The fBC-constrained scheme incorporates regionally observed number fractions of BC-containing particles and reduces NMB of simulated MAC by 13 % ~ 68 % over the representative regions. The two schemes offer a practical, observation-anchored route to more accurate BC optics and less uncertain assessments of aerosol absorption.