Variation of the NMVOC speciation in the solvent sector and the sensitivity of modelled tropospheric ozone

Abstract

Non-methane volatile organic compounds (NMVOCs) are detrimental to human health owing to the toxicity of many of the NMVOC species, as well as their role in the formation of secondary air pollutants such as tropospheric ozone (O3) and secondary organic aerosol. The speciation and amount of NMVOCs emitted into the troposphere are represented in emission inventories (EIs) for input to chemical transport models that predict air pollutant levels. Much of the information in EIs pertaining to speciation of NMVOCs is likely outdated, but before taking on the task of providing an up-to-date and highly speciated EI, a better understanding of the sensitivity of models to the change in NMVOC input would be highly beneficial. According to the EIs, the solvent sector is the most important sector for NMVOC emissions. Here, the sensitivity of modelled tropospheric O3 to NMVOC emission inventory speciation was investigated by comparing the maximum potential difference in O3 produced using a variety of reported solvent sector EI speciations in an idealized study using a box model. The sensitivity was tested using three chemical mechanisms that describe O3 production chemistry, typically employed for different types of modelling scales - point (MCM v3.2), regional (RADM2), and global (MOZART-4). In the box model simulations, a maximum difference of 15 ppbv (ca. 22% of the mean O3 mixing ratio of 69 ppbv) between the different EI speciations of the solvent sector was calculated. In comparison, for the same EI speciation, but comparing the three different mechanisms, a maximum difference of 6.7 ppbv was observed. Relationships were found between the relative contribution of NMVOC compound classes (alkanes and oxygenated species) in the speciations to the amount of Ox produced in the box model. These results indicate that modelled tropospheric O3 is sensitive to the speciation of NMVOCs as specified by emission inventories, suggesting that detailed updates to the EI speciation information would be warranted. Furthermore, modelled tropospheric O3 was also sensitive to the choice of chemical mechanism and further evaluation of both of these sensitivities in more realistic chemical-transport models is needed.