A case study of heavy PM2.5 secondary formation by N2O5 nocturnal chemistry in Seoul, Korea in January 2018: Model performance and error analysis

Abstract

Heterogeneous hydrolysis of dinitrogen pentoxide (N2O5) plays an important role in nighttime nitrate (NO3-) formation in urban areas, and sometimes influences the occurrence of heavy PM2.5 pollution the next day in the Seoul Metropolitan Area (SMA), Korea. Here, we discuss the heavy PM2.5 wintertime episode of January 13-15, 2018, which was mainly induced by nighttime N2O5 heterogeneous reaction in the SMA. In our case, we confirmed that nighttime N2O5 hydrolysis is the most critical factor in the rapid formation of aerosol nitrate at high levels during the night, which prevailed in the morning of the next day. Our Integrated Process Rate (IPR) analysis showed that nighttime nitrate production in the episode was almost solely attributable to N2O5 chemistry, with hourly mean production rates of 0.8 & PLUSMN; 0.4 mu g/m(3) per hour in SMA, which is comparable to the daytime nitrate photochemical production rates of 0.9 & PLUSMN; 0.5 mu g/m(3) per hour. We also carried out a series of assessment of N2O5-driven nitrate formation sensitivity, and relevant errors were quantified by applying different N2O5 uptake coefficients in the WRF-CMAQ model. The potential errors of nighttime-average nitrate concentrations induced by N2O5 uptake process were assessed from a linear perspective for the planetary boundary layer (PBL) variances caused by four different PBL parameterization schemes: YSU, ACM2, MYJ, and QNSE. The potential error ranges by N2O5 uptake process were analyzed to be 2.3 to 3.5 mu g/m(3) (~10% relative to the nighttime-average), while biases of PBL simulations from 4 parameterization schemes were 2.3 & PLUSMN; 1.0 mu g/m(3), showing similar ranges in our episode. Although N2O5-driven heavy PM2.5 episodes do not occur often in SMA, our findings suggest the importance of N2O5 chemistry in vigorous wintertime nitrate formation and operational prediction errors of such PM2.5 episodes, under the premise of enhanced PBL simulation capabilities.