Black carbon mixing state impacts on aerosol activation investigations using particle-resolved model simulations

Black carbon-containing aerosol particles (called BC particles hereafter) are a major particle type in the atmosphere. Their source is the incomplete combustion of carbon containing material, which means that except for natural biomass burning all sources of BC particles are anthropogenic. They absorb solar radiation directly, warming the atmospheric layer where they reside. Because of the absorbing nature of BC particles and their relatively short life time compared to long-lived greenhouse gases (e.g. carbon dioxide), the reduction of BC emission has been suggested as a short-term global warming mitigation strategy. However, BC particles can also act as CCN, form cloud droplets and hence contribute to a cooling impact on climate. Both the optical properties and the CCN properties of BC particles depend on the BC mixing state, i.e. which other species are present within one BC-containing particle. The BC mixing state, in turn, evolves by so-called aerosol aging processes such as condensation of atmospheric gaseous components and coagulation with other aerosols. The climate effects of BC particles are therefore closely related to the BC mixing state and its evolution. Uncertainties regarding the representation of these processes in models are a major source of uncertainties in current climate model predictions. This dissertation focuses on studying the impact of mixing state of BC particles on cloud microphysical quantities during the early stages of cloud formation. With the recently-developed particle-resolved model PartMC-MOSAIC, the mixing state and other physico-chemical properties of individual aerosol particles can be tracked as the particles undergo aerosol aging processes. As part of this dissertation this model framework was extended, and a particle-resolved cloud parcel model was developed. This simulates the particle growth due to condensation of water vapor for each particle in a given population. The cloud parcel simulations are initialized with aerosol populations from urban plume simulations with PartMC-MOSAIC. This new particle-resolved cloud parcel model was then used for quantifying the errors in computing cloud microphysical quantities due to simplified model representations of aerosol particles. A library of scenarios was designed by varying BC emission rates, background aerosol number concentrations and gas emission rates for the urban plume scenarios, and varying the cooling rate in the cloud parcel scenarios to explore how the errors depend on environmental conditions. In this analysis we focused on four cloud microphysical quantities, namely activation fraction, fN, BC nucleation-scavenged mass fraction, fBC, effective radius, (r)eff , and dispersion, epsilon, of the cloud droplet spectrum. The errors due to simplified mixing state representation for these quantities were largest when the population contained subpopulations of hydrophobic particles and hygroscopic particles. The errors in activation fraction fN was within 45%, while the error in fBC reached up to 1300% for these conditions. The errors in (r)eff and epsilon were within 12% and 62% respectively. An estimation of cloud short wave albedo revealed that the errors in effective radius led to a range of error in albedo between -0.0055 and 0.03. For all quantities, the errors increased for the scenarios with reduced gas emission rates. For fBC decreasing errors were observed with increasing cooling rates, while for the other variables a dependence on cooling rate was not evident. In addition to the development of modeling tools and the error quantification framework, this dissertation provided process analysis of how the change in aerosol population affected cloud droplet number concentration. A metric was developed to attribute the difference in cloud droplet number concentration, Nd obtained (at the same cloud parcel cooling rate) between two particular environmental scenarios to the difference in aerosol population (called plume effect) and to the difference in water vapor competition (called parcel effect). For most of the scenarios presented, the plume effect dominated the parcel effect in changing Nd. Further, we investigated the kinetic limitation to cloud droplets growth. With the use of the particle-resolved cloud parcel model, the types of kinetic limitation mechanisms affecting individual cloud droplets were revealed. Besides, applying the recently developed concept of chemical composition diversity by Riemer & West 2013, the relationship between chemical composition diversity of the aerosol population and the relative spectral dispersion of the cloud droplet spectrum were investigated. It is found that there is no strong association between chemically diverse aerosol populations and cloud droplet spectra with large dispersion. Instead cloud droplet number concentration and cooling rate are the key parameters controlling dispersion