Year-Round Analysis of Multiphase Sulfate Production in Aerosol Particles in East Asia

Abstract

Missing sulfate production pathways have been implicated as the cause of model underestimates of sulfate during haze events in East Asia. We add multiphase oxidation of SO₂ in aerosol particles by H₂O₂, O₃, NO₂, HCHO, and O₂, catalyzed by transition metal ions (TMIs), to the GEOS-Chem model and evaluate the model with (1) year-round ground-based observations in Seoul, South Korea, (2) airborne observations from the KORUS-AQ field campaign, and (3) fall and winter ground-based observations in Beijing, China. Multiphase chemistry contributes 14% to 90% to total sulfate production depending on the location and season and increases model daily average sulfate by 2 to 3 μg m^-3, with maximum daily increases up to 12 μg m^-3. From winter to summer, oxidation pathways shift, with the largest fraction of multiphase sulfate production occurring during spring and summer due to oxidation by H₂O₂. Multiphase oxidation of SO₂ by the H₂O₂ pathway reduces gas-phase H₂O₂ concentrations by -40% in spring, which improves model agreement with H₂O₂ airborne observations. Oxidation pathways also shift between cities, in particular the contribution from the TMI and NO₂ pathways, which are more important in Beijing than in Seoul. This is due to higher levels of transition metals, and a larger impact of an overly shallow mixed layer in Beijing compared to Seoul. The implementation of multiphase aerosol chemistry in GEOS-Chem here allows for the use of this chemistry in other models that can address boundary layer errors, including WRF-GC and CESM-GC. The analysis presented here shows that this chemistry is important to the simulation of sulfate year-round, not only during haze events, and is unique in showing coupled gas- and aerosol-phase impacts of multiphase chemistry.

Keywords: KORUS-AQ; aerosol; east Asia; haze; model; sulfate.

1

Comparison of the model and observed daily average total Fe (a) and total Mn (b) at Olympic Park in Seoul during KORUS-AQ and in Beijing during November and December 2016. The observed and modeled concentrations of each metal are shown in the inset. The Mn concentrations from the original calculation of 50× less than natural Fe described in the text are given in pink.

2

Calculation of the reactive uptake probability using the equations in Table and eq for T = 298K, ALW = 10 μg m^–3, r _a = 1.5 × 10^–5 cm, A = 12 × 10^–5 cm^–2 cm^–3, SO₂ = 20 ppb, Fe = 200 ng m^–3, Mn = 10 ng m^–3. Dashed colored lines represent the impact of the ionic strength (I = 25M). Dashed vertical lines represent the range of model pH during KORUS-AQ.

3

Timeseries of sulfate measured by an aerosol mass spectrometer (HR-ToF-MS) at the Korea Institute of Science and Technology (KIST) in Seoul, South Korea, split into five periods: (a) January: 1/1/2016 to 1/22/2016, (b) May–June: 5/1/2016 to 6/10/2016, (c) August: 7/25/2016 to 8/31/2016, (d) October: 9/27/2016 to 10/31/16, and (e) November–December: 1/01/16 to 12/30/2016. Sulfate was increased by 1.2 to account for differences between PM₁ and PM_2.5. The model sensitivity with aerosol multiphase chemistry (yellow) is described in Sections . Model sulfate is the sum of inorganic sulfate and hydroxymethane sulfonate (HMS). Panels (f) through (j) show the contribution of average modeled sulfate production pathways to the multiphase chemistry simulation for each time period.

4

Mean vertical profiles of sulfate for the descents over Olympic Park for the KORUS-AQ meteorological periods: (a) dynamic period (5/1/16–5/16/16), (b) stagnant period (5/17/16–5/22/16), (c) haze period (5/25/16–5/31/16), and blocking period (6/1/16–6/7/16). The observations (black) and model (red) are binned to the nearest 0.5 km below 3 km. The model sensitivities with parameterized sulfate production (blue) and with multiphase chemistry (yellow) are described in Sections . Panel (c) has a wider x-axis than the other panels (0–20 μg m^–3) to allow for better visualization of the other meteorological periods. Observed sulfate was increased by 1.2 to account for differences between PM₁ and PM_2.5. Model sulfate is the sum of inorganic sulfate and hydroxymethane sulfonate (HMS).

5

Mean model production pathways for sulfate for East Asia (111 °E to 127 °E, 20.5 °N to 41 °N) (a) and Seoul (b) during the haze period (5/25/16–5/31/16). Mean vertical profiles of (c) hydrogen peroxide (H₂O₂) and (d) sulfur dioxide (SO₂) for the descents over Olympic Park for the KORUS-AQ haze period. The observations (black) and model (red) are binned to the nearest 0.25 km below 2 km. The model sensitivities with parameterization of sulfate production (blue) and with multiphase chemistry (yellow) are described in Sections .

6

Daily average timeseries of sulfate measured by TOF-ACSM at Peking University in Beijing for (a) September: 9/3/16–10/4/16 and (b) December: 12/4–12/26/16. Panel (a) includes both PM₁ and PM_2.5 sulfate and panel (b) has PM₁ sulfate. The model sensitivity with explicit multiphase aerosol chemistry (yellow) is described in Sections . Panels (c) and (d) show the contribution of modeled sulfate production pathways to the multiphase chemistry simulation for each time period.

7

Mean diel variability in Beijing of nitrate in September (a) and December (b) and sulfate in September (c) and December (d) for the data shown in Figure . Model sensitivities are described in the text.