Application of WRF/Chem Model over East Asia: Model Evaluation and Aerosol-Meteorology Feedbacks

In recent years, the model of Weather Research and Forecasting coupled with Chemistry (WRF/Chem) has been rapidly developed as a new generation of regional air quality modeling system worldwide. During the past three decades, East Asia has experienced continuous rapid economic and population growth, industrialization and urbanization, which have caused significant degradation of air quality. In addition, East Asia has special topography and geographical location which are responsible for various climatic conditions in terms of temperature (like the heat low), pressure (like the Mongolian low pressure and subtropical high pressure), airflow (like summer monsoon and Tibetan Plateau monsoon) and rainfall (like Meiyu front). Thus, East Asia area has been considered as a supreme testbed for testing and improving the performance of WRF/Chem. In this work, WRF/Chem version 3.3.1 is applied to four months: January, April, July, and October in 2001, each representing a season, over East Asia to evaluate the model performance on meteorological predictions (including temperature at 2 meters (T2), water vapor mixing ratios at 2 meters (Q2), wind speeds at 10 meters (WS10), and precipitation) and the concentrations of major air pollutants (including PM₁₀, PM_2.5, CO, NO, NO₂, O₃ and SO₂). The surface meteorological variables are evaluated using hourly global surface observational data from the National Climatic Data Center (NCDC). The concentrations of chemical species are evaluated using surface measurements over mainland China (derived from the Air Pollution Index (API) in 42 main cities), Hong Kong (about 9 sampling sites), Taiwan (about 65 sampling sites), and Japan (over 2000 sampling sites). Aerosol optical depth (AOD) is evaluated using satellite data from Terra Moderate-resolution Imaging Spectroradiometer (MODIS), and column abundance of CO, NO₂, and O₃ is evaluated using the Measurements of Pollution in the Troposphere (MOPITT), Global Ozone Monitoring (GOME), and Total Ozone Mapping Spectrometer (TOMS). The objectives of this study are to evaluate the model performance over East Asia, and analyze the anthropogenic aerosol feedbacks. The results indicated that, for the surface meteorological variables, the T2 and Q2 are simulated reasonably well by WRF/Chem; however, WS10 is overestimated with Normalized Mean Biases (NMBs) of 47% to 81%. In addition, daily total precipitation is poorly estimated (correlation coefficients (rs) of 0.04 - 0.19 and NMBs of -5% to 4%). Column AOD is reasonably well simulated with rs of 0.52 - 0.71 in three months (January, July, and October) and NMBs of -13% to 3%, but moderately underpredicted in April (with a rs of 0.67 and an NMB of 40%). The simulated spatial distributions of PM₁₀ show that the northwestern and northern China are the two regions with the largest dust aerosol sources in East Asia, which is consistent with the API-derived PM₁₀ data (rs of 0.54 - 0.81 and NMBs of -25% to -11%). The model reproduces well the seasonality of PM₁₀ concentrations over mainland China, with the highest concentrations in winter, followed by spring, fall and summer. However, over Japan and Taiwan, PM10 concentrations are moderately-to-significantly underpredicted with NMBs from -67% to -28%, probably caused by the underprediction of the dust transport or emissions of local natural and anthropogenic sources. The simulated column mass of CO and NO₂ show good correlations with satellite data (with rs of 0.59 - 0.78); however, in most of months (except for April CO), column CO and NO₂ mass are overpredicted with NMBs of -4% - 29% and 34% - 71%, respectively, likely due to the uncertainties of the satellite data retrieval or air pollutant emissions. Unlike the column results, the surface concentrations of CO and NO₂ are underpredicted in nearly all months (except for July NO₂ over Japan) (with NMBs of -82% to -28% for CO and -88% to 5% for NO₂) over Japan, Taiwan and Hong Kong, likely due to the underestimation of the anthropogenic emissions or limited capabilities of the model on simulating some meteorological variables which significantly affect air quality (such as the planetary boundary layer height (PBLH), wind speed/direction (WS/WD), and vertical transport). O₃ surface mixing ratios are overpredicted by WRF/Chem model (NMBs of from 3 -40%, and rs of 0.33 to 0.50), likely due to the uncertainties of predictions of its precursors (such as NO_x and VOCs). WRF/Chem model also underpredicts the surface SO₂ mixing ratios over Japan and Taiwan with NMB of -50% to -35%. An examination of anthropogenic aerosol feedbacks shows that, in July, anthropogenic aerosols cause about a decrease of 9.54 W m^-2 (4%) in the domain-wide mean surface shortwave radiation, and an increase of about 3.86×10⁶ m^-3 (157%) in cloud droplet number concentrations (CDNC), indicating their strong direct and indirect effects on radiation and cloud formation over East Asia. These results have important implications to assess the effectiveness of the emission control effort since 2001 and the development of integrated emission control strategies that are effective for both air quality control and climate mitigation.