Improved global modelling of HO<sub>x</sub> recycling in isoprene oxidation: evaluation against the GABRIEL and INTEX-A aircraft campaign measurements

Abstract

Abstract. Stimulated by recent important developments regarding the oxidation chemistry of isoprene, this study evaluates and quantifies the impacts of different mechanism updates on the boundary layer concentrations of OH and HO 2 radicals using the IMAGESv2 global chemistry transport model. The model results for HO x , isoprene, NO, and ozone are evaluated against air-based observations from the GABRIEL campaign, conducted over the Guyanas in October 2005, and from the INTEX-A campaign over the Eastern US in summer 2004. The version 2 of the Mainz Isoprene Mechanism (MIM2, Taraborrelli et al., 2009) used as reference mechanism in our simulations, has been modified to test (i) the artificial OH recycling proposed by , (ii) the epoxide formation mechanism proposed by Paulot et al. (2009b), and finally (iii) the HO x regeneration of the Leuven Isoprene Mechanism (LIM0) proposed by Peeters et al. (2009);Peeters and Müller (2010). The simulations show that the LIM0 scheme holds by far the largest potential impact on HO x concentrations over densely vegetated areas in the Tropics as well as at mid-latitudes. Strong increases, by up to a factor of 4 in the modelled OH concentrations, and by a factor of 2.5-3 in the HO 2 abundances are estimated through the LIM0 mechanism compared to the traditional isoprene degradation schemes. Comparatively much smaller OH increases (<25%) are associated with the implementation of the mechanism of Paulot et al. (2009b); moreover, the global production of epoxides is strongly suppressed (by a factor of 4) when the LIM0 scheme is combined Correspondence to: T. Stavrakou (jenny@aeronomie.be) with this mechanism. Hydroperoxy-aldehydes (HPALDs) are found to be major first-generation products in the oxidation of isoprene by OH, with a combined globally averaged yield of 50-60%. The use of the LIM0 chemistry in the global model allows for reconciling the model with the observed concentrations at a satisfactory level, compared to the other tested mechanisms, as the observed averaged mixing ratios of both OH and HO 2 in the boundary layer can be reproduced to within 30%. In spite of the remaining uncertainties in the theoretically-predicted rates of critical radical reactions leading to the formation of HPALDs, and even more in the subsequent degradation of these new compounds, the current findings make a strong case for the newly proposed chemical scheme. Experimental confirmation and quantification is urgently needed for the formation of HPALDs and for their fast OH-generating photolysis.

Peroxy radical isomerization in the oxidation of isoprene

“…3. While not the dominant process suggested previously, 7,22,23,32 peroxy radical isomerization is an important process for isoprene oxidation in the atmosphere even at the slow rate reported here. HPALD oxidation pathway(s), the resulting product(s), and the oxidation thereof remain unknown.…”

“…[22][23][24] A direct comparison to the theoretical rate is not, however, straightforward as we only observe the end product of two important processes, namely: (a) interconversion between the b-OH-OO, Z-d-OH-OO, and E-d-OH-OO peroxy radicals (shown for 1-OH isomers in R9); and (b) the 1,6-H-shift isomerization reactions from the two Z-d-OH-OO peroxy radicals (R5 and R6):…”

Photolysis, OH reactivity and ozone reactivity of a proxy for isoprene-derived hydroperoxyenals (HPALDs)

“…21 On the other hand, recent work has also called into question the accuracy of OH concentrations observed via laser-induced fluorescence in biogenic-dominated regions. 48 As noted and accounted for in earlier theoretical work 16,20 and recently highlighted, 21 the influence of C 5 -HPALD chemistry on the radical budget depends on the relative rates of C 5 -HPALD photolysis (which generates OH) and OH reaction (which removes C 5 -HPALDs and, possibly, also OH). Fig.…”

“…Global models 20,22 have demonstrated dramatic increases in OH-by factors of 3 or more in some regions-upon incorporating the ISO 2 isomerization and C 5 -HPALD chemistry as originally proposed. 21,22 This mechanism also affects the distribution of oxidized VOC, leading Archibald et al 22 to suggest that calculated isomerization rates should be reduced by an order of magnitude.…”

“…Photolysis of the PACALD is assumed to be even more efficient, leading to further production of OH and HO 2 . Incorporation of this chemistry into numerical simulations improves model-measurement agreement for OH in some cases 20,21 but can create additional problems, such as misrepresentation of other oxidized VOC 22 and over-prediction of HO 2 . 11 Ultimately, the impact of this mechanism on HO x abundance depends on both the rate of C 5 -HPALD production from ISO 2 isomerization and the photolysis and OH reaction rates of the C 5 -HPALD products.…”