Atmospheric Chemistry Reaction in Turbulent Flow and Its Application to Chemical-Transport Model in Urban Environments

Turbulence mixes initially segregated chemical species in the air and hence affects the rate of chemical reactions. The significance of such interaction between turbulence and chemical reactions can be evaluated from the Damköhler number (Da), the ratio between the turbulent and the chemical timescales. For fast chemistry with chemical timescale shorter than the turbulent timescale (Da ≥ 1), this chemical-turbulence interaction can significantly affect the actual rate of chemical reaction in the flow; while for slow chemistry with Da < 1, the effect is expected to be insignificant. Such interaction is however unresolved in operational chemical-transport models. Low-resolution chemical-transport models dilute and mix precursors instantly in a coarse grid, and hence overestimate for example ozone production. This treatment is particularly unsuitable for urban environments where the pollutant emissions are strong and highly localized. In order to quantify the inaccuracy of such an assumption in an urban environment, the effect of chemical-turbulence interaction is investigated by means of direct numerical simulations of two initially segregated chemicals with a second-order chemistry scheme in a range of reaction rates in the atmospheric boundary layer, focusing on strong emission fluxes and heterogeneous emissions. In a top-bottom entrainment-emission configuration, the inefficiency of turbulence mixing of the reactants can cause a 15-35% reduction in reaction rate from the imposed value for fast chemistry conditions (Da = 1 -10). With heterogeneous surface emissions, the segregation of the two emitted reactants causes a further reduction in reaction rate to 95% of the imposed rate under fast-chemistry and strong-emission conditions. Such reduction in reaction rate due to the inefficient mixing of chemicals by turbulence is highly dependent on the instantaneous Da and the ratio between the corresponding Da of the two species. The results from our simulations are then degraded to lower resolution to mimic the calculations from a chemical-transport model. It is revealed that even at a horizontal resolution of 1 km the model can induce a spurious increase in reaction rate by 4-13%. We then extend our investigation to a simple NO_X-O₃ chemistry scheme, and apply our knowledge from the direct numerical simulations to a real-case scenario in the WRF-Chem model over the region of Hong Kong, focusing on its impact on ozone calculation.