Observations of radicals in the atmosphere: measurement validation and model comparisons

Comparisons of the levels of radicals observed during field campaigns to the results of detailed chemical box models serve as a vital tool to assess our understanding of the underlying chemical mechanisms involved in tropospheric oxidation. Recent measurements of OH radicals are significantly higher than those predicted by models in certain environments, especially those at low NOx influenced by high emissions of biogenic compounds. Other studies have suggested that fluorescence assay by gas expansion (FAGE) instruments may be susceptible to an unknown interference in the measurement of OH. The interference hypothesis can be tested through the implementation of an alternative method to determine the OH background signal, whereby OH is removed by the addition of a chemical scavenger prior to sampling by FAGE (known as OHchem). The more established method to determine the background is to move the laser excitation wavelength to a value where OH does not absorb (OHwave). The Leeds FAGE instrument was modified to facilitate OHchem by the construction of an inlet pre-injector (IPI), where OH is removed through reaction with propane. Following optimisation, the modified instrument was deployed at a coastal location in Norfolk, England during summer 2015 as part of the ICOZA (Integrated Chemistry of OZone in the Atmosphere) campaign, and in the highly polluted megacity, Beijing, in winter 2016 and summer 2017 as part of the AIRPRO (An Integrated Study of AIR Pollution PROcesses in Beijing) project. An automated analysis procedure was written in IGOR to facilitate data workup and to provide quality assurance and control, ensuring valid comparisons of the two OH measurements, and between observed and simulated radical concentrations. The IPI was characterised in terms of sensitivity (virtually identical to traditional FAGE sampling) and scavenging efficiency (>99% removal). For all three field campaigns, measurements of OH made using the alternative background technique were in very good agreement with the traditional method, with intercomparison slopes (OHwave vs OHchem) of 1.05–1.16, providing confidence in previous measurements of OH made using the Leeds FAGE instrument. However, a significant interference was observed at night during the ICOZA campaign, accounting for ~40% of the total OHwave signal on average, although the chemical identity of the species responsible could not be determined. The ICOZA measurements were compared to radical levels predicted using the explicit Master Chemical Mechanism. For a model constrained to HO2, OH concentrations were in agreement with FAGE observations to within instrumental uncertainty (~26%) during the daytime, for which the rate of OH production from the photolysis of HONO was equal to that from the reaction of O(1D) with water vapour. However, OH levels were underpredicted by approximately a factor of ~3 at night, which cannot be explained by OH measurement interferences alone. In contrast, HO2 observations were overestimated by ~40% during the daytime and significant concentrations were also observed at night (~2–3 × 107 molecule cm-3), which were underpredicted by up to an order of magnitude. The daytime HO2 discrepancy was most severe at low NO levels, with measurement-to-model ratios of <0.5 for NO mixing ratios below 0.1 ppbv. Total organic peroxy radical (RO2) concentrations were also measured during ICOZA, representing one of the first few FAGE datasets of RO2. Severe measurement-model discrepancies were found for both day and nighttime periods, with RO2 concentrations underpredicted by a factor of ~9 on average. In contrast to HO2, the model could capture daytime RO2 observations reasonably well at low NO but the discrepancy was most severe in the high NO regime, reaching a factor of ~20 for NO levels above 3 ppbv. This result is consistent with previous studies and suggests that our understanding of atmospheric oxidation chemistry under high NOx conditions is incomplete.