Atmospheric stability significantly influences the accumulation and dispersion of air pollutants in the near-surface atmosphere, yet few stability metrics have been applied as predictors in statistical PM2.5 concentration mapping practices. In this study, eleven stability metrics were derived from radiosonde soundings collected in eastern China for the time period of 2015-2018 and then applied as independent predictors to explore their potential in favoring the prediction of PM2.5. The statistical results show that the in situ PM2.5 concentration measurements correlated well with these stability metrics, especially at monthly and seasonal timescales. In contrast, correlations at the daily timescale differed markedly between stability metric and also varied with seasons. Nevertheless, the modeling results indicate that incorporating these stability metrics into the PM2.5 modeling framework rendered small contribution to PM2.5 prediction accuracy, yielding an increase of R2 by < 5% and a reduction of RMSE by < 1 μg/m3 on average. Compared with other stability indices, the inversion depth and intensity appeared to have relative larger benefiting potential. In general, our findings indicate that including these stability metrics would not result in significant contribution to the PM2.5 prediction accuracy in eastern China since their effects could be partially overwhelmed or offset by other predictors such as AOD and boundary layer height.
Keywords: Aerosol optical depth; Air pollution; Atmospheric stability; PM(2.5); Radiosonde data; Stagnant event.
Copyright © 2019 Elsevier Ltd. All rights reserved.