황사감시기상탑 자료를 이용한 황사발원 관련 기상조건의 특성 연구

Abstract

Asian dust events are typically observed on the Korean Peninsula every spring between the months of March and May. In the 10 years (2002–2011) since the Korea Meteorological Administration (KMA) started issuing special reports on these outbreaks, Asian dust was observed on a total of 93 occasions, including 22 advisory-level events and 15 warning-level events. In particular, warning-level outbreaks have been observed an average of 1–2 times each year since the system has been in place, with 80% of the occurrences originating in the Gobi Desert and Inner Mongolia. The events observed on the Korean Peninsula in the past decade were mostly (67 events, accounting for 72% of the total occurrences) concentrated in spring, with 28 occurrences in March, 25 occurrences in April, and 14 occurrences in May. Out of the 7 events that affected the Korean Peninsula in spring 2011, this study has singled out three advisory-level events, and analyzed meteorological observations and turbulence and dust concentration data collected at the Asian dust monitoring towers in Erdene and Nomgon to identify the formation of spring dust storms originating in Mongolia and the characteristics of associated meteorological conditions. The findings were compared with output from UM-ADAM2, which addresses limitations of surface meteorological data and is employed in the analysis of horizontal spatial distribution and Asian dust forecast, with a view to a quantitative interpretation of the forecast output; this correlation analysis was based on PM10 concentration, wind speed, and friction velocity data from UM-ADAM2 and the monitoring towers. The study also examined the development and movement of air pressure systems in addition to the intensity and properties of aerosol observed over the Korean Peninsula, and found the following: First, meteorological variables such as PM10 concentrations (collected at 3 m), wind speed (at 16 m), friction velocity (at 16 m), temperature (at 16m and 2 m), relative humidity (at 16 m), and soil temperature (at 5 cm) taken from the monitoring towers in Erdene and Nomgon in Mongolia were analyzed to identify the meteorological conditions conducive to Asian dust formation. Somewhat elevated temperatures above 10°C were recorded at 16 m and 2 m immediately before the outbreaks, but plummeted as soon as the dust storms had formed. Likewise, the soil temperature at 5 cm exceeded 10°C before the outbreaks, and dropped down to below 5°C afterwards, showing a trend similar to the temperatures at 16 m and 2 m. Relative humidity was low (less than 40%) immediately preceding the outbreaks, and rose to 50% afterwards. Second, correlation analysis of measurements and model output found that the modeled values more or less trended like the observations in the case of PM10. Modeled wind speed approximated the observation trends, with some underestimation, while modeled friction velocity exhibited the same trend as in the observations, with a measure of overestimation in those cases where there were dust outbreaks. Third, statistical analysis of the observations and model output found that the simulated PM10 values in all cases other than those for Nomgon in Case 2 were generally higher than the actual measurements, with the model showing a deviation of approximately 58 ㎍m-3. The correlation was relatively low overall, between 0.09 and 0.28, although Case 2 (Erdene) and Case 3 (Nomgon) recorded high values of 0.87 and 0.61 respectively. PM10 concentrations for the three cases (spring 2011) showed a low correlation of 0.37, and a mean bias of 32.2 ㎍m-3 in the model relative to the observations, suggesting that model error must be taken into account in quantitatively interpreting forecasted Asian dust amount at the point of origin. As for wind speed, the observations only showed a mean bias of about 3.9 ms-1 relative to model output; the correlation was relatively high, recording 0.87 for Erdene in Case 2, 0.85 for Erdene in Case 3, and 0.79 for all three cases, suggesting that wind speed measurements may be useful in the quantitative forecast of Asian dust. Friction velocity showed little (a mean bias of 0.1 ms-1) deviation between the modeled and observed values, a high correlation of 0.5–0.8, and an even higher correlation of 0.65 for all three cases; like wind speed, friction velocity data may be potentially useful in the quantitative prediction of Asian dust. Fourth, the synoptic weather charts for the three cases revealed the outbreaks are to be attributed to strong wind due to a strengthened pressure gradient force in central Mongolia. The dust was transported to the Korean Peninsula by the northwestern air current formed by the expansion of a continental high pressure system located behind a low pressure system. The 300 hPa weather charts show an atmospheric pressure configuration conducive to dust storm formation, with a jet stream positioned in southern Mongolia and northern China that resulted in a developed surface low pressure system, which in turn created updrafts favoring dust formation. Fifth, the study considered NCEP/NCAR reanalysis data for 850 hPa and 500 hPa geopotential heights and spatial distribution of temperature fields for the three cases. The 850 hPa charts show a pressure gradient and a thermal gradient crossing over Inner Mongolia, producing strong baroclinic instability, and corresponding strong jet streams from the low pressure system over this area; behind this low pressure system, a high pressure system developed gradually, causing in turn the development of a cold advection in front, and traveled to northern Manchuria. On the other hand, the 500 hPa charts show a high pressure system developing behind an upper-level trough over Inner Mongolia and the pressure gradient force causing a strong wind zone to form in front of the high pressure system. The subsequent eastward movement of this trough facilitated the development of a high pressure system in the rear, so that dust was carried by the resulting northwestern current to the Korean Peninsula. Relative humidity was 10% or lower at 1,000 hPa and 850 hPa over the source areas in southern Mongolia and northern China; the air was fairly dry both on the surface and in the vertical profile, suggesting that the atmospheric conditions over Mongolia and China stayed dry for the duration of all three outbreaks. As for wind vector and speed, southern Mongolia recorded a strong wind zone with wind speed in excess of 16 ms-1 at 850 hPa, and 40 ms-1 at 300 hPa. The strong wind zone at 300 hPa facilitated the development of a low pressure system at ground level, which in turn further strengthened surface updrafts—a synoptic pattern conducive to dust formation, and wind vectors at 300 hPa were found to approximate the direction of the jet stream in the 300 hPa weather charts. Sixth, the backward trajectories for 500 m, 1,000 m, and 1,500 m showed northwestern currents moving from Inner Mongolia to the Korean Peninsula via Bohai Bay. The direction of the currents was found to match wind vectors in synoptic weather charts. Seventh, PM10 concentrations showed a spike after the 97 outbreak of Asian dust in Korea while PM2.5 and PM1.0 levels were low. The volume concentration of particles in the coarse particle mode was found to increase along with the mass concentration of PM10, indicating the infiltration of naturally-formed primary particulate matter such as Asian dust. The intensity trends of Asian dust (PM10) as measured at 28 locations on the Korean Peninsula were found to match those points when increased mass and volume concentrations were recorded at 8 locations using particle counters. High concentrations of aerosol were found at all observation points during the three outbreaks. The surface conditions and meteorological parameters at the point of Asian dust formation may have a substantial impact on the quantity of dust generated. However, the determination of meteorological conditions and the forecast of Asian dust are a challenge in the case of Mongolia, which lacks the necessary observing equipment. Therefore, if the network comprising the monitoring towers used in this study is expanded to include the source area, data obtained from the towers in real time will enable well-informed research on Asian dust, and help us better understand the mechanism underlying its formation and produce accurate prediction of outbreaks.