Geographical and diurnal features of amine-enhanced boundary layer nucleation

Abstract

Amines have recently been found to be an important ingredient in the nucleation and initial growth of atmospheric aerosols; however, global estimates of the spatial and temporal extent of amine-enhanced nucleation are currently missing. We utilize two recently published laboratory data sets of amine-sulfuric acid nucleation to evaluate the accuracy of previously published nucleation parameterizations and to produce a new amine-enhanced new particle formation (NPF) parameterization that better reproduces the laboratory observations at atmospherically relevant sulfuric acid concentrations. We implement and compare the amine-enhanced NPF parameterizations and a kinetic nucleation parameterization within the global aerosol-climate model ECHAM-HAMMOZ and find that the spatial features of amine-enhanced and kinetic NPF are clearly different. Amine-enhanced NPF is limited to areas near the source regions of amine due to its short gas phase residence time of 6.9 h, whereas kinetic nucleation (which depends only on sulfuric acid concentration) produces particles more uniformly across the globe due to long-range transport of SO2. The notably stronger land-sea contrast in amine-enhanced nucleation simulations is in line with relatively rare atmospheric observations of NPF over open oceans. However, when the uptake of gas phase amine molecules to aerosol particles is limited according to previously published estimates (0.2% of collisions leading to uptake), the amine-enhanced NPF parameterization predicts in some regions unrealistically high NPF rates (~1000 cm-3 s-1) compared to typical observations. Our results indicate that amine-enhanced nucleation may be an important particle formation mechanism near amine source regions but also highlights the need for more tightly defined constraints on the spatial and temporal distribution of amine emissions, gas-to-particle partitioning mechanisms of amines, and condensation and coagulation sinks in global models.