Observations and modelling of glyoxal in the tropical Atlantic marine boundary layer

Abstract

In situ field measurements of glyoxal at the surface in the tropical marine boundary layer have been made with a temporal resolution of a few minutes during two 4-week campaigns in June–July and August–September 2014 at the Cape Verde Atmospheric Observatory (CVAO; 16∘52′ N, 24∘52′ W). Using laser-induced phosphorescence spectroscopy with an instrumental detection limit of ∼1 pptv (1 h averaging), volume mixing ratios up to ∼10 pptv were observed, with 24 h averaged mixing ratios of 4.9 and 6.3 pptv observed during the first and second campaigns, respectively. Some diel behaviour was observed, but this was not marked. A box model using the detailed Master Chemical Mechanism (version 3.2) and constrained with detailed observations of a suite of species co-measured at the observatory was used to calculate glyoxal mixing ratios. There is a general model underestimation of the glyoxal observations during both campaigns, with mean midday (11:00–13:00) observed-to-modelled ratios for glyoxal of 3.2 and 4.2 for the two campaigns, respectively, and higher ratios at night. A rate of production analysis shows the dominant sources of glyoxal in this environment to be the reactions of OH with glycolaldehyde and acetylene, with a significant contribution from the reaction of OH with the peroxide HC(O)CH2OOH, which itself derives from OH oxidation of acetaldehyde. Increased mixing ratios of acetaldehyde, which is unconstrained and potentially underestimated in the base model, can significantly improve the agreement between the observed and modelled glyoxal during the day. Mean midday observed-to-modelled glyoxal ratios decreased to 1.3 and 1.8 for campaigns 1 and 2, respectively, on constraint to a fixed acetaldehyde mixing ratio of 200 pptv, which is consistent with recent airborne measurements near CVAO. However, a significant model under-prediction remains at night. The model showed limited sensitivity to changes in deposition rates of model intermediates and the uptake of glyoxal onto aerosol compared with sensitivity to uncertainties in chemical precursors. The midday (11:00–13:00) mean modelled glyoxal mixing ratio decreased by factors of 0.87 and 0.90 on doubling the deposition rates of model intermediates and aerosol uptake of glyoxal, respectively, and increased by factors of 1.10 and 1.06 on halving the deposition rates of model intermediates and aerosol uptake of glyoxal, respectively. Although measured levels of monoterpenes at the site (total of ∼1 pptv) do not significantly influence the model calculated levels of glyoxal, transport of air from a source region with high monoterpene emissions to the site has the potential to give elevated mixing ratios of glyoxal from monoterpene oxidation products, but the values are highly sensitive to the deposition rates of these oxidised intermediates. A source of glyoxal derived from production in the ocean surface organic microlayer cannot be ruled out on the basis of this work and may be significant at night.