地表參數化對中尺度氣象模擬及空氣品質模擬的影響

The effects of land surface representation on the simulations of mesoscale meteorological and photochemical air quality model

蔡哲慧(Che-Hui Tsai)

指導教授 ： 江旭程

淡江大學/工學院/水資源及環境工程學系博士班/博士(2009年)

https://doi.org/10.6846/TKU.2009.00303

摘要

臭氧事件日常發生於大尺度風速微弱、日照強烈時，在此氣象條件下，土地使用及覆蓋（Land use/ Land cover, LULC）資料庫和地表參數化對氣象模擬有顯著的影響。為了改進臭氧預測結果，氣象模式需要更好的地面表示方式。本研究目的是針對台灣臭氧事件日進行靈敏度分析，以了解LULC資料和地表參數化對氣象模擬和空氣污染物空間分佈的影響。

本研究選擇發生於2003年6月4日至6日的臭氧事件日進行探討，使用RAMS和CAM x模式進行模擬，並將都市能量模式（Town Energy Balance Model）併入RAMS模式，進行三個案例模擬：案例一使用AVHRR資料庫並關閉TEB模組；案例二使用MODIS資料庫並關閉TEB模組。此兩案例都以LEAF3模組計算都市影響。案例三使用MODIS資料庫，都市地區計算啟動TEB模組。

氣象模擬結果和8個地面氣象測站觀測資料進行驗證評估，結果顯示案例一模擬風速高估，地面溫度低估。MODIS案例風速模擬結果和觀測值接近；白天地面溫度模擬值已有改善，但夜晚地面溫度仍低估。案例三風速和地面溫度模擬值都有改善。整體評估後，啟動TEB模組可得到比較好的氣象模擬，氣象模擬結果亦發現都市的存在，使地面溫度上升1∼2℃，風速降低1∼2m/s。夜晚都市熱島效應的強度比白天大；都市的存在使對流、大氣亂流強度以及混合層高度增加。

將三個案例氣象模擬結果輸入空氣品質模式進行模擬，並和環保署空氣品質監測站觀測值進行驗證評估。模擬結果相當合理，仍以TEB案例表現最好。不論白天或夜晚，都市對臭氧、NO2濃度的空間和時間分佈都有重要的影響。

本研究亦針對2003年5月30日至31日高臭氧事件進行模擬，評估RAMS-TEB-CAMx模式預報臭氧的能力。模擬結果和臭氧、NO2觀測值進行驗證，結果尚稱合理。台灣北部地區模擬結果相當良好，但台灣中南部地區臭氧模擬值低估，NO2模擬值高估。海岸附近，大型點污染源會因NOx滴定作用而降低臭氧濃度，靠近山區臭氧的尖峰值偏低。模擬結果顯示模式有預報的能力，然而準確性不高，可能因預測風向偏差造成。

關鍵字

土地使用 ； 土地覆蓋 ； 都市熱島效應 ； 臭氧預報 ； 數值模擬

並列摘要

It is well known that ozone episodes occur under weak synoptic conditions with low wind speed and strong sunlight. During these periods, the Land use/ Land cover (LULC) dataset and surface parameterization scheme have significant influence on the meteorological simulation. To improve the performances of numerical ozone forecast, a better representation of land surface in the meteorological model is needed. The aims of this study is to carried out a numerical study to investigate the influence of LULC data and surface parameterization on the results of meteorological simulation and the calculated spatial distributions of atmospheric pollutants during ozone episodes in Taiwan.

An ozone episode occurred during the period June 4-6, 2003, was simulated using the RAMS and CAMx model. The town energy balance (TEB) urban canopy model was coupled to RAMS. Three numerical simulations were performed. The first simulation used AVHRR dataset and the TEB computation was turn off. The second simulation used MODIS dataset and the TEB computation was also turn off. The urban effects were represented by LEAF3 scheme for the first two runs. The third simulation used MODIS data and the TEB computation was turn on. The wind speeds computed by the first case were overestimated and the surface temperatures were underestimated when compared with the observations obtained from eight urban surface weather stations. In MODIS case, the surface wind speeds have better agreement when compared with observations. The results of surface temperature during day time were improved; however, they still underestimated during night time. As for the third case, the results for wind speeds and surface temperature were improved. Overall the meteorological simulations are superior when TEB is used. Numerical results indicate that urban can increases the surface temperature by 1~2 ℃ and decreases the wind speed by 1~2 m/s. The simulation indicates that the urban heat island effect is stronger at night than daytime. Urban can enhance the convection and atmospheric turbulence and increase mixing height.

The results of three meteorological simulations were used to drive the air quality model. The calculated concentrations of air pollutants were evaluated against the observations obtained by Taiwan EPA air quality monitoring network. Most of the results are very reasonable and the simulation using TEB is superior to the other cases. The present study shows that both nocturnal and diurnal urban effects have an important impact on the temporal and spatial distributions of ozone and nitrogen dioxide.

This study also examined the capability of applying RAMS-TEB-CAMx model to ozone forecasting during a 3-day episode of 29–31 May 2003. The results are fair when compared with observed O3 and NO2 concentrations. The performance of forecast in northern Taiwan is pretty good. However, the results of O3 concentrations were underestimated and NO2 concentrations were overestimated in middle and southern Taiwan. The large point sources located in the coastal zone will reduce the ozone concentration because of NOx titration effect. The peak ozone concentrations near the mountain range were underestimated. A possible reason of this imprecision is the error of wind direction caused by the complicated wind system in the simulation domain. Although the performance of this ozone forecast system is reasonable, there are numerous opportunities for improvements in the future.

並列關鍵字

land use ； land cover ； urban heat island ； ozone forecast ； numerical simulation

參考文獻

潘雲潔, 2008： 台灣地區不同土地使用形態對氣象模式模擬結果的影響, 淡江大學, 碩士論文

游沛蓁, 2008： 台灣地區人為廢熱對氣象模式模擬之影響, 淡江大學, 碩士論文

黃運貴, 2005： 運輸部門能源消費量及節能措施之研究, 台灣大學, 博士論文

曹嘉宏, 2007： 台灣土地利用型態對MM5模擬局部環流之影響, 中央大學, 碩士論文

Avissar R and Pielke RA, 1989： A Parameterization of Heterogeneous Land Surfaces for Atmospheric Numerical Models and Its Impact on Regional Meteorology. Monthly Weather Review. 117(10):2113–2136.