Incorporating the missing heterogeneous oxidation of S(IV) by NO into the WRF-Chem model. Sulfate production is not sensitive to increase in SO emission. The newly added reaction reproduces sulfate concentrations well during winter haze. We implemented the online coupled WRF-Chem model to reproduce the 2013 January haze event in North China, and evaluated simulated meteorological and chemical fields using multiple observations. The comparisons suggest that temperature and relative humidity (RH) were simulated well (mean biases are -0.2K and 2.7%, respectively), but wind speeds were overestimated (mean bias is 0.5 m?s ). At the Beijing station, sulfur dioxide (SO ) concentrations were overpredicted and sulfate concentrations were largely underpredicted, which may result from uncertainties in SO emissions and missing heterogeneous oxidation in current model. We conducted three parallel experiments to examine the impacts of doubling SO emissions and incorporating heterogeneous oxidation of dissolved SO by nitrogen dioxide (NO ) on sulfate formation during winter haze. The results suggest that doubling SO emissions do not significantly affect sulfate concentrations, but adding heterogeneous oxidation of dissolved SO by NO substantially improve simulations of sulfate and other inorganic aerosols. Although the enhanced SO to sulfate conversion in the HetS (heterogeneous oxidation by NO ) case reduces SO concentrations, it is still largely overestimated by the model, indicating the overestimations of SO concentrations in the North China Plain (NCP) are mostly due to errors in SO emission inventory.