A model study investigating the sensitivity of aerosol forcing to the volatilities of semi-volatile organic compounds

Abstract

Secondary organic aerosol (SOA) constitutes an important component of atmospheric particulate matter, with a substantial influence on air quality, human health and the global climate. The volatility basis set (VBS) framework has provided a valuable tool for better simulating the formation and evolution of SOA where SOA precursors are grouped by their volatility. This is done in order to avoid the computational cost of simulating possibly hundreds of atmospheric organic species involved in SOA formation. The accuracy of this framework relies upon the accuracy of the volatility distribution of the oxidation products of volatile organic compounds (VOCs) used to represent SOA formation. However, the volatility distribution of SOA-forming vapours remains inadequately constrained within global climate models, leading to uncertainties in the predicted aerosol mass loads and climate impacts. This study presents the results from simulations using a process-scale particle growth model and a global climate model, illustrating how uncertainties in the volatility distribution of biogenic SOA precursor gases affect the simulated cloud condensation nuclei (CCN). We primarily focused on the volatility of oxidation products derived from monoterpenes as they represent the dominant class of VOCs emitted by boreal trees. Our findings reveal that the particle growth rate and their survival to CCN sizes, as simulated by the process-scale model, are highly sensitive to uncertainties in the volatilities of condensing organic vapours. Interestingly, we note that this high sensitivity is less pronounce in global-scale model simulations as the CCN concentration and cloud droplet number concentration (CDNC) simulated in the global model remain insensitive to a 1-order-of-magnitude shift in the volatility distribution of organics. However, a notable difference arises in the SOA mass concentration as a result of volatility shifts in the global model. Specifically, a 1-order-of-magnitude decrease in volatility corresponds to an approximate 13 % increase in SOA mass concentration, while a 1-order-of-magnitude increase results in a 9 % decrease in SOA mass concentration over the boreal region. SOA mass and CCN concentrations are found to be more sensitive to the uncertainties associated with the volatility of semi-volatile compounds, with saturation concentrations of 10−1 µg m−3 or higher, than the low-volatility compounds. This finding underscores the importance of having a higher resolution in the semi-volatile bins, especially in global models, to accurately capture SOA formation. Furthermore, the study highlights the importance of a better representation of saturation concentration values for volatility bins when employing a reduced number of bins in a global-scale model. A comparative analysis between a finely resolved nine-bin VBS setup and a simpler three-bin VBS setup highlights the significance of these choices. The study also indicates that radiative forcing attributed to changes in SOA over the boreal forest region is notably more sensitive to the volatility distribution of semi-volatile compounds than low-volatility compounds. In the three-bin VBS setup, a 10-fold decrease in the volatility of the highest-volatility bin results in a shortwave instantaneous radiative forcing (IRFari) of −0.2 ± 0.10 W m−2 and an effective radiative forcing (ERF) of +0.8 ± 2.24 W m−2, while a 10-fold increase in volatility leads to an IRFari of +0.05 ± 0.04 W m−2 and an ERF of +0.45 ± 2.3 W m−2 over the boreal forest region. These findings underscore the critical need for a more accurate representation of semi-volatile compounds within global-scale models to effectively capture the aerosol loads and the subsequent climate effects.