A Coupled Data Assimilation Method for Visibility and Relative Humidity CAOP v1.0 module within the WRFDA v4.3 framework: Improving both PM_2.5 and Humidity Forecasts

Abstract. High precision prediction of near-surface PM_2.5 mass concentration is critical for effective air quality management and mitigating health risks. While Data assimilation (DA) of the aerosol mass concentration or the aerosol optical properties (AOP) observation optimizes the initial conditions of an atmospheric chemistry prediction model and improves aerosol prediction skill, the potential of assimilating surface visibility remains constrained by the overlooked regulatory effect of relative humidity (RH) on visibility variance. To address this, we developed a coupled DA method of AOP and RH (CAOP v1.0) within the WRFDA (version 4.3) framework. Four parallel experiments were conducted across mainland China in December 2020: a control experiment and three DA experiments. Results demonstrate that assimilating visibility alone (VisOnly) reduced RMSE by 21 % for visibility and 9 % for PM_2.5, validating the CAOP module. Furthermore, DA of both visibility and RH – whether performed simultaneously but independently in the VisRH_Indp experiment or via the coupled scheme in the VisRH_Copl experiment – effectively corrected the PM_2.5 analysis bias originating from background RH deviations. The coupled scheme (VisRH_Copl) achieved superior performance, lowering PM_2.5 RMSE by 17 % compared to 16 % in the independent run (VisRH_Indp). Crucially, the CAOP scheme more efficiently leveraged visibility data to constrain the RH field, improving the RH forecast correlation coefficient by 25 % (versus 23 % in the VisRH_Indp). This study confirms that visibility observations substantially enhanced RH analysis and demonstrated that the coupled DA method outperforms independent DA approaches for joint PM2.5 and RH forecasting.