Evaluation of WRF-CMAQ performance for PM10 simulation in Seoul, South Korea based on back-trajectory analysis
Abstract
For an operational forecast of particulate matters with diameters ≤ 10 μm (PM10), the Korean government has used the air quality forecasting system based on the Weather Research and Forecasting version 3.1 and the United States Environmental Prediction Agency Models-3/Community Multiscale Air Quality version 4.6. For evaluating the performance of the model, we have compared model-simulated PM10 concentrations with observations during the cold seasons (October through March) of 2015-2018 in Seoul, South Korea. In general, 70% of the model simulations are hit within 10 µg m-3 error range while the model frequently overestimates the moderate-PM10 days (30 < PM10 ≤ 60 µg m-3) and underestimates the high-PM10 days (60 < PM10 ≤ 100 µg m-3). As PM10 concentrations in Korea are largely influenced by the transboundary transports over East Asia, we examined the model simulations based on 72-hour back-trajectories from Seoul point (37.51°N, 126.99°E) at 500 m altitude. The 410 days in the 30-100 µg m-3 range are grouped into three major wind flows of westerly (163 days), northerly (99 days), and northwesterly (148 days) with similar hit rates of 72%, 67%, and 71%, respectively. On the other hand, missed forecasts are concentrated to certain groups as overestimation in the moderate-PM10 days are most frequent for the westerly group (32 days) while underestimation in the high-PM10 days are frequent for the northerly (13 days) and the northwesterly group (13 days). To further investigate the cause for these biases, the model inputs of meteorological initial analysis field and estimates of emission amount are analyzed for each group. Overestimation in the moderate-PM10 days of the westerly group is associated with a typical synoptic pattern of the calm anticyclonic system suggesting more stagnation of PM10 especially when emission amount is overestimated in eastern China where the wind blows from. Underestimation in the high-PM10 days of the northerly and the northwesterly groups are related to climatological patterns or windy strong cyclonic system suggesting more scattering of PM10 in addition to the small inflow from underestimated emission amount in northeastern China. Therefore, individual consideration of each back-trajectory group along with more accurate estimates of emission amount may help improve the model performance.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFM.A13Q3157H
- Keywords:
-
- 0345 Pollution: urban and regional;
- ATMOSPHERIC COMPOSITION AND STRUCTURE;
- 0365 Troposphere: composition and chemistry;
- ATMOSPHERIC COMPOSITION AND STRUCTURE;
- 3305 Climate change and variability;
- ATMOSPHERIC PROCESSES;
- 3364 Synoptic-scale meteorology;
- ATMOSPHERIC PROCESSES