To improve the accuracy of air quality prediction, a refined emission inventory for Guangdong Province and updated model-ready emissions for regions outside GD have been integrated into the comprehensive air quality modelling system (WRF-SMOKE-CMAQ).

Annual emissions of SO₂, NOx, VOC, PM_2.5, PM₁₀, and NH₃ in the Pearl River Delta (PRD) region, as well as pollutant contributions in 2015 and 2017, base years of previous and updated emission inventory, have been analyzed. Statics reveal that SO₂, PM_2.5, PM₁₀, VOC and NH₃ emissions have a reduction over PRD, while NOx has an increase (14.4%). Non-road mobile sources become the largest contributor to SO₂ and NOx in 2017, instead of stationary combustion in 2015. Industrial processes and dust remain the largest contributor to PM_2.5 and PM₁₀,...[ Read more ]

To improve the accuracy of air quality prediction, a refined emission inventory for Guangdong Province and updated model-ready emissions for regions outside GD have been integrated into the comprehensive air quality modelling system (WRF-SMOKE-CMAQ).

Annual emissions of SO₂, NOx, VOC, PM_2.5, PM₁₀, and NH₃ in the Pearl River Delta (PRD) region, as well as pollutant contributions in 2015 and 2017, base years of previous and updated emission inventory, have been analyzed. Statics reveal that SO₂, PM_2.5, PM₁₀, VOC and NH₃ emissions have a reduction over PRD, while NOx has an increase (14.4%). Non-road mobile sources become the largest contributor to SO₂ and NOx in 2017, instead of stationary combustion in 2015. Industrial processes and dust remain the largest contributor to PM_2.5 and PM₁₀, respectively. Same as VOC and NH₃, which are dominated by one primary source in both two years, solvent use and agriculture, respectively.

Emission inventories and air quality model-ready emissions have been updated from a base year on 2015 to 2017 for all datasets. Four representative months, January, April, July, and October in 2019, have been selected to run with two case scenarios: one using previous emission settings (2015 EI for PRD, 2015 MEIC for non-PRD and 2015 MEIC for outside GD) and the other one using refined emissions (2017 EI for PRD, 2017 EI for non-PRD and 2017 MEIC for outside GD).

Model performance of SO₂, NO₂, O₃, PM_2.5, and PM₁₀ in January, April, July, and October 2019 has then been evaluated. SO₂ and NO₂ have shown significant improvements when using refined emission inventory over GD, especially in removing extremely high peaks. However, O₃, PM_2.5, and PM₁₀ performance are relatively not good. R of O₃ increases over non-PRD, but increases in RMSE, NME of O₃, PM_2.5, and PM₁₀ indicates deviations are larger between new simulation and observation. PM_2.5 and PM₁₀ have been underpredicted over 30%, while ozone is overestimated over 20%. For HK, though its emission inventory has not been modified, pollutant patterns are affected by the regional impact. The changes for each pollutant are the same as it in PRD. Though the loss of PM meets the uncertainty range of raw EI (-43% for PM_2.5 and -45% for PM₁₀), further adjustments need to be made based on the uncertainty of each emission source.