Assessment of N2O5 heterogeneous reaction mechanisms in the Community Multiscale Air Quality (CMAQ) model

Abstract

Nitrogen oxides have negative human health impacts and play a central role in the production of secondary air quality pollutants like ozone and PM2.5. While day-time NOx cycling depends on photolytic reactions, night-time heterogeneous chemistry, including the loss of N2O5 to particles and the production of ClNO2, regulates the nocturnal reservoirs and sinks of NOx. However, existing parameterizations of nocturnal NOx heterogeneous chemistry used in air quality models do not capture the variability in observed N2O5 removal or ClNO2 production rates. Here, we implemented new N2O5 uptake (𝛾(𝑁2𝑂5)) and ClNO2 yield (Φ(𝐶𝑙𝑁𝑂2)) parameterizations in the Community Multiscale Air Quality (CMAQ) model that account for the role of particulate organic matter in regulating N2O5 reactive uptake and the role of reactive solutes in suppressing ClNO2 production. We compared the performance of these chemically representative parameterizations against existing model parameterizations and measurements taken during the NSF/NCAR Wintertime Investigation of Transport, Emission, and Reactivity (WINTER) 2015 flight campaign. When using the new parameterizations for uptake and yield, we find that the coarse mode contributed modestly to N2O5 loss (17.2%) but significantly to ClNO2 production (60.3%), highlighting the role for coarse mode chemistry with the new parameterizations. The new 𝛾(𝑁2𝑂5) parameterization in the fine mode decreased average predicted 𝛾(𝑁2𝑂5) by an order of magnitude compared to the model default, increasing agreement between modeled N2O5 concentration and observations (𝑅𝑀𝑆𝐸=0.37 𝑝𝑝𝑏 compared to 0.43 𝑝𝑝𝑏 for the model default). Despite being more chemically representative, the new uptake parameterization was biased low due to underestimates in modelled particle organic-phase O:C ratio. The new Φ(𝐶𝑙𝑁𝑂2) parameterization decreased Φ(𝐶𝑙𝑁𝑂2) compared to the fine mode model default, resulting in further underestimation of the ClNO2 concentration (𝑁𝑀𝐵=−73.7% compared to −37.9% for the default case). The low bias in the new Φ(𝐶𝑙𝑁𝑂2) parameterization was due to underestimation of in Aitken and accumulation mode particulate chloride. We expect that under conditions where model particulate O:C and Cl- are better represented the new parameterizations will more accurately capture the mean state and variability in 𝛾(𝑁2𝑂5) and Φ(𝐶𝑙𝑁𝑂2).