실제 대기경계층 내에 놓인 언덕 지형 주위 유동은 단순 모델을 적용한 풍동 실험이나 수치해석 결과와 는 큰 차이가 발생한다. 승학산은 풍향각에 따른 협곡, 매우 가파른 언덕 및 급격한 언덕을 지나는 유동의 후류 특성 등에 대한 여러 가지 지형적 특징을 지니고 있다. 이와 같은 유동 특성을 분석하기 위해 50m 높이의 기상 타워를 설치하여 30m, 40m, 50m 에서의 풍속 및 풍향을 각각 10분 평균으로 측정하였다. 경계층 풍속 분포 측정 결과, 급격한 언덕을 가진 풍향각에서는 큰 구배를 가지는 풍속 분포가 측정되었다. 특정 풍향각에 대하여 난류강도 분포가 협곡과 가파른 언덕에서 큰 값을 관찰할 수 있었으며, 프로파일 법으로 계산된 표면조도 역시 지형적인 특성으로 인한 경계층 풍속 분포를 효과적으로 나타내었다. 반면 시간적으로 분류된 대기안정성이 유동에 끼치는 영향은 복잡한 지형적 특성으로 인해 열유동 현상이 크게 나타나지 않는 것을 확인할 수 있었다.

자연지형에 의한 기류 변화를 CFD 모델과 풍동실험을 통하여 비교 연구하였다. 해안에 인접한 남부 산악지 형을 대상으로 풍동실험과 몇 종류의 CFD 모델 시뮬레이션을 수행하였다. 사용한 CFD 모델은 2 가지로, 하나는 LES 난류모델을 사용하는 가상경계기법을 활용한 유한체적법(Finite Volume Method, FVM) 코드와또하나는 4 가지 RANS 난류모델이 선택사양으로 제공되는 상업용 CFD 모델이다. 지형에 의해 풍속이 증가되는 지점과 감소되는 지점에서 수 직풍속분포를 직접 비교하고 상관도를 구하였다. CFD 모델과 풍동실험과의 상관도(R)는 0.890~0.965로 매우 높게 나 타났으며, CFD 모델중 VBM LES CFD 모델은상관도 0.965로나타나풍동실험을가장잘모사하는것으로분석되었다.

The mean wind speed and turbulence intensity profiles in the atmospheric boundary layer were extracted from a LIDAR remote sensing campaign in order to apply for CFD validation. After considering the semi-steady state field data requirements to be used for CFD validation, a neutral atmosphere campaign period, in which the main wind direction and the power-law exponent of the wind profile were constantly maintained, was chosen. The campaign site at the Pohang Accelerator Laboratory, surrounded by 40~50m high hills, with an apartment district spread beyond the hills, is to be classified as a semi-complex terrain. Nevertheless, wind speed profiles measured up to 100m above the ground fitted well into a theoretical-experimental logarithmic-law equation. The LIDAR remote-sensing data of the sub-layer of the atmospheric boundary layer has been proven to be superior to the data obtained by conventional extrapolation of the wind profile with 2 or 3 anemometer measurements.

In order to clarify the relation between sea breeze penetration and Planetary Boundary Layer development in southeastern part of the Korean Peninsula, several numerical assessments were carried out using atmospheric numerical model WRF(Weather Research and Forecasting). Compared with onset time of sea breeze at eastern coast area(Uljin), the time at southern coast region(Masan) with complex costal line tend to delay for several hours. The penetration patterns of sea breeze between two coastal regions are some different due to the shape of their coastal line and back ground topography. Intensified valley wind due to high topography of lee side of Uljin can help penetration of sea breeze at early time. So penetration of sea breeze at early time often prevent PBL to develop at Uljin and lower PBL height last for a day time. But because of late penetration of sea breeze at Masna, PBL Height dramatically decrease after 1500LST. The distribution of front genesis function based on the heat and momentum variation are explained obviously the sea breeze penetration patterns and agreed well with the PBL height distribution.

A system coupled the prognostic WRF mesoscale model and CALMET diagnostic model has been employed for predicting high-resolution wind field over complex coastal area. WRF has three nested grids down to 1km during two days from 24 August 2007 to 26 August 2007. CALMET simulation is performed using both initial meteorological field from WRF coarsest results and surface boundary condition that is Shuttle Radar Topography Mission (SRTM) 90m topography and Environmental Geographic Information System (EGIS) 30m landuse during same periods above. Four Automatic Weather System (AWS) and a Sonic Detection And Ranging (SODAR) are used to verify modeled wind fields. Horizontal wind fields in CM_100m is not only more complex but better simulated than WRF_1km results at Backwoon and Geumho in which there are shown stagnation, blocking effects and orographically driven winds. Being increased in horizontal grid spacing, CM_100m is well matched with vertically wind profile compared SODAR. This also mentions the importance of high-resolution surface boundary conditions when horizontal grid spacing is increased to produce detailed wind fields over complex terrain features.

In order to reduce the uncertainties and improve the air flow field, objective analysis using observational data is chosen as a method that enhances the reality of meteorology. To improve the meteorological components, the radius influence and nudging coefficient of the objective analysis should perform a adequate value on complex area for the objective analysis technique which related to data reliability and error suppression. Several numerical experiments have been undertaken in order to clarify the impacts of the radius influence and nudging coefficient of the objective analysis on meteorological environments. By analyzing practical urban ground conditions, we revealed that there were large differences in the meteorological differences in each case. In order to understand the quantitative impact of each run, the Statistical analysis by estimated by MM5 revealed the differences by the synoptic conditions. The strengthening of the synoptic wind condition tends to be well estimated when using quite a wide radius influence and a small nudging coefficient. On the other hand, the weakening of the synoptic wind is opposite.

The objective of this work is the air quality modeling according to the scenarios of emission on complex terrain. The prognostic meteorological fields and air quality field over complex areas of Seoul, Korea are generated by the PSU/NCAR mesoscale model (MM5) and the Third Generation Community Multi-scale Air Quality Modeling System (Models - 3/CMAQ), respectively. The emission source was driven from the Clean Air Policy Support System of the Korea National institute of Environmental Research (CAPSS), which is a 1 km x 1 km grid in South Korea during 2003. In comparison of air quality fields, the simulated averaged PM10, NO2, and O3 concentration on complex terrain in control case were decreased as compared with base case. Particularly PM10 revealed most substantial localized differences by (18 ～ 24 μg/m3). The reduction rate of PM10, NO2, and O3 is respectively 18.88, 13.34 and 4.17%.

A sea/land breeze circulation system and a regional scale circulation system are formed at a region which has complex terrain around coastal area and affect to the dispersion and advection of air pollutants. Therefore, it is important that atmospheric circulation model should be well designed for the simulation of regional dispersion of air pollutants. For this, Local Circulation Model, LCM which has an ability of high resolution is used. To verify the propriety of a LCM, we compared the simulation result of LCM with an exact solution of a linear theory over a simple topography. Since they presented almost the same value and pattern of a vertical velocity at the level of 1 ㎞, we had a reliance of a LCM. For the prediction of dispersion and advection of air pollutants, the wind field should be calculated with high accuracy. A numerical simulation using LCM will provide more accurate results over a complex terrain around coastal area.