Three-year (2015-2017) winter-time air pollution in China: ambient PM2.5 concentration changes derived from retrospective analysis
Abstract
To better characterize the anthropogenic emission relevant aerosol species, the GSI-WRF/Chem data assimilation system was updated from the GOCART aerosol scheme to MOSAIC-4BIN scheme. Three year (2015-2017) winter-time (January) surface PM2.5 observations from 1600+ sites were assimilated hourly using the updated 3DVAR system in the assimilation experiment CONC_DA. Parallel control experiment that did not employ DA (NO_DA) was also performed. Both experiments were verified against the surface PM2.5 observations, MODIS 550-nm AOD and also sulfate-nitrate-ammonia species at a single site in Beijing. In NO_DA experiment that 2010_MEIC emissions were used, modeled PM2.5 are severely overestimated in Sichuan Basin (SB), Central China (CC), YRD (Yangzi River Delta), and PRD (Pearl River Delta) which indicated the emissions for 2010 are not appropriate for 2015-2017, as strict emission control strategies were implemented in recent years. Meanwhile, underestimations in Northeastern China (NEC) and Xin Jiang (XJ) were also observed. The assimilation experiments significantly reduced the high biases of surface PM2.5 in SB, CC, YRD, and PRD, and also low biases in NEC. However the improvement of the low biases in XJ is relatively small due to the large difference between the observations and the model background in the DA process, likely indicating that the emissions in the model are seriously underestimated in this region. Assimilating surface PM2.5 also significantly changed the column AOD and resulted in closer agreement with MODIS data. Comparisons of the sulfate-nitrate-ammonia species at Beijing showed that the assimilation of total PM2.5 mass concentrations also helped to improve the species partitioning through the statistics of background error covariance.
The observations and the reanalysis data from assimilation experiment were used to investigate the year-to-year changes. As the differences of the reanalysis data among years reflect combining effects of meteorology and emission and the differences of modeling result from control experiment (with same emissions) among years reflect the separate effect of meteorology, the important roles of emission and meteorology in driving the changes in the three years can be analyzed. The analysis indicated that meteorology played different roles in 2016 and 2017: the higher pressure system, lower temperature and higher PBLH in 2016 are favorable for pollution dispersion (compared with 2015) while the situation is almost the opposite in 2017 (compared with 2016) that leads to the increasing PM2.5 from 2016 to 2017 although emission control strategy were implemented in both years. There are still large uncertainties in this approach as the inaccurate emission input in the model brings large biases in the analysis. In another study, we will estimate the SO2 emissions by using an Ensemble Kalman Filter for the same period aiming to investigate further on this aspect.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.A33I3275C
- Keywords:
-
- 0305 Aerosols and particles;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0322 Constituent sources and sinks;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0345 Pollution: urban and regional;
- ATMOSPHERIC COMPOSITION AND STRUCTUREDE: 0368 Troposphere: constituent transport and chemistry;
- ATMOSPHERIC COMPOSITION AND STRUCTURE