Keywords:Black carbon, Wet scavenging coefficients, FLEXPART, Long-term measurements
Using long-term and harmonized measurements of black carbon (BC) and carbon monoxide (CO) from three representative background sites in East Asia (Baengnyeong and Gosan in South Korea, and Noto in Japan), the regional and seasonal differences in wet removal rates were investigated according to airmass transport pathways and meteorological conditions. To identify the airmass origin and accumulated precipitation along with trajectories (APT), backward trajectories at 500 m during 72 hours were calculated by the Hybrid Single Particle Lagrangian Integrated Trajectory 4 model. The European Center for Medium-Range Weather Forecasts (ECMWF) ERA5 hourly pressure and surface level meteorological data (0.25° × 0.25°) were used as input data to obtain accurate and detailed information of each endpoint of trajectories. The wet removal rates were more efficient in South Korea and Japan among the region and in fall and winter among the season. These differences depending on the regional and seasonal division could be partially explained by the inherent difference in the coating thickness of BC particles, according to the dominant emission sectors. Because the industrial sector (thin coated) is dominant in East and North China, in contrast, the transportation sector (thick coated), mainly emitted from diesel vehicles, has a high portion in South Korea and Japan. By the same token, wet removal rates in winter and summer showed the highest and lowest depending on the dominant emission sectors, such as house heating (thick coated) and industry, respectively. Next, we calculated transport efficiency (TE) from FLEXible PARTicle (FLEXPART) Lagrangian transport model (version 10.4) to investigate the representativeness of wet removal rate of model. Overall median TE from FLEXPART was overestimated compared to measurement, implying underestimation of wet scavenging coefficients. The median of estimated below cloud scavenging coefficients showed slightly overestimation than that calculated from FLEXPART as a factor by 1.7. On the other hand, the median of calculated in-cloud removal coefficients including sub-grid effect was highly underestimated than that of estimation by a factor of 5.1. From the analysis of artificial neuron networks, the convective available potential energy (CAPE), which is well known as an indicator of vertical instability, should be considered in in-cloud scavenging process to improve the better representative regional difference in BC wet scavenging over East Asia.