Winter photochemistry in Beijing: Observation and model simulation of OH and HO2 radicals at an urban site.

A field campaign was conducted from November to December 2017 at the campus of Peking University (PKU) to investigate the formation mechanism of the winter air pollution in Beijing with the measurement of hydroxyl and hydroperoxyl radical (OH and HO2) with the support from comprehensive observation of trace gases compounds. The extent of air pollution depends on meteorological conditions. The daily maximum OH radical concentrations are on average 2.0 × 106 cm-3 and 1.5 × 106 cm-3 during the clean and polluted episodes, respectively. The daily maximum HO2 radical concentrations are on average 0.4 × 108 cm-3 and 0.3 × 108 cm-3 during the clean and polluted episodes, respectively (diurnal averaged for one hour bin). A box model based on RACM2-LIM1 mechanism can reproduce the OH concentrations but underestimate the HO2 concentrations by 50% during the clean episode. The OH and HO2 concentrations are underestimated by 50% and 12 folds during the polluted episode, respectively. Strong dependence on nitric oxide (NO) concentration is found for both observed and modeled HO2 concentrations, with the modeled HO2 decreasing more rapidly than observed HO2, leading to severe HO2 underestimation at higher NO concentrations. The OH reactivity is calculated from measured and modeled species and inorganic compounds (carbon monoxide (CO), NO, and nitrogen dioxide (NO2)) make up 69%-76% of the calculated OH reactivity. The photochemical oxidation rate denoted by the OH loss rate increases by 3 times from the clean to polluted episodes, indicating the strong oxidation capacity in polluted conditions. The comparison between measurements at PKU site and a suburban site from one previous study shows that chemical conditions are similar in both urban and suburban areas. Hence, the strong oxidation capacity and its potential contribution to the pollution bursts are relatively homogeneous over the whole Beijing city and its surrounding areas.