為瞭解台灣地區空氣品質狀況，行政院環保署自民國八十二年起開始設置地面空氣品質測站，並依地形、氣象條件及空氣流通等條件，將台灣地區分為七大空品區，分別為北部空品區、竹苗空品區、中部空品區、宜蘭空品區、花東空品區、雲嘉南空品區及高屏空品區。本研究之研究目標為解析南台灣空氣品質污染物之長期趨勢，因此選定區域為雲嘉南空品區與高屏空品區。本研究之研究方法為利用時間序列的長期趨勢特性，將週期性因子與氣象因子分離出來，便可直接探討南部地區空氣污染物長期變化趨勢，並探討不同因子的加入對於長期趨勢變化之影響。
本研究將各測站空氣品質污染物之長期趨勢分為四組加以迴歸討論，分別為Holland原公式迴歸、原公式加入循環因子之迴歸、原公式經氣象因子修正後之迴歸與原公式加入循環因子並經氣象因子修正後之迴歸，比較各組修正前後之系數差異與趨勢變化。為確認迴歸值已完整解析實測數據，本研究以模擬值與實測值之殘差值做常態分佈檢定，並以p-value為判別值，倘若殘差值為常態分佈且平均值為零，則認定迴歸式之因子完整解析實測值大部分趨勢變化因子，即可進一步分析其長期趨勢變化特性；反之，則判定殘差值中仍有未知趨勢存在，其長期趨勢特性僅供參考。
研究結果發現，各測站之長期趨勢不盡相同；整體來看則可發現主要空品區內之長期趨勢特性類似。循環變數的加入使得SO2、碳氫化合物等污染物種之殘差值較接近常態分佈，顯示景氣循環為該物種趨勢變化因子之一。氣象因子的加入對於長期趨勢之影響較大，並有出現趨勢逆轉的現象。一般來說，經由氣象因子的修正後可得到較具代表性之長期趨勢，故本研究主要趨勢探討以經氣象因子修正並加入循環因子之長期趨勢為主。南台灣各地區之臭氧長期趨勢以0.3%~4.8%逐年增加，NO2+O3之長期趨勢呈現主要都會地區為逐年降低，其餘地區則為逐年增加，然而由其餘污染物種之長期趨勢可知都會地區SO2、NO、NMHC等為逐年降低，顯示鄉村地區之臭氧管制策略成果較不明顯。
南台灣臭氧事件日長期趨勢可發現，歷年臭氧事件日發生時段有延後現象，此現象出現於雲林、台南、高雄縣市較為明顯。由臭氧事件日逐時特性可知，南台灣地區臭氧事件時段為十五時以前與十六時以後之特性明顯不同，靠海之測站之尖峰時段差距較大，靠山之測站則為尖峰時段濃度差距較大。

In order to understand air quality in Taiwan, Taiwan EPA has constructed air quality monitoring network since 1993. Based on geographical and meteorological conditions and the nature of air contaminants, seven Air Quality Zones are divided by EPA—northern Taiwan, Jhu-Miao area, central Taiwan, Yun-Chia-Nan area, Kao-Ping area, Hua-Dong area, and Yilan. By analysis of temporal long-term trend of pollutant, meterological and economical factors were separated from data base in this study, and then the impact caused by each factor or both is also discussed.
Holland model, ME-Regression model(with Economical-adjusted), MM-Regression model(with Meteorological-adjusted), MME-Regression model(with Meteorological-&Economical-adjusted), are used to estimate the long-term trend of daily mean concentrations. The trend for these four methods was cauculated and compared. For checking the difference between measured and calculated data, residual analysis was used to test. If residual values is presented by normal distribution and average is zero, model fits the measured data and analysis of long-term pollutant will be discussed. However, if model cannot fit, it reveals that unknown trend exists.
The results show that the ME-Regression model (with economical-adjusted) is more suitable than the Holland model for estimating the long-term trend of SO2 and NMHC. It means that the economical factor is one of the important factors for regression model. The meteorological factor exerts large impacts on ozone concentrations, and may mask the long-term trend of ozone concentrations resulting from precursor emissions. The trend without meteorological-adjusted and economical-adjusted was analyzed in this study. Regression results indicated that most ambient O3 concentration increased by more than 2.2%. The long-term trends of O3+NO2, SO2, NOx, and NMHC, concentration revealed downward trend at urban sites, but upward trend in the other sites.
Special characteristics were investigated by analysis of O3 episodes in south Taiwan. Time of O3 episodes were delay in south Taiwan. This study separates O3 episodes into two groups: O3 episodes before 15 o’clock and O3 episodes after 16 o’clock. The results show that characteristics of these two groups were different greatly. Time between peak concentration is larger in shore sites, and difference between two peak concentration is larger in sites near mountains.

Altman, D.G., Bland, J.M.,1995. Statistics notes: the normal distribution. BMJ,310,298.

Frischer, T., Studnicka, M., Gartner, C., Tauber, E., Horak, F., Veiter, A., Spengler, J., Kuhr, J., Urbanek, R., 1999. Lung function growth and ambient ozone. A three-year population study in school children. American Journal of Respiratory and Critical Care Medicine 160, 390-396.

Goyal, P., Chan, A. T., Jaiswal, N., 2006. Statistical models for the prediction of respirable suspended particulate matter in urban cities. Atmospheric Environment 40, 2068-2077.

Holland, D. M., Principe, P. P. and Sickles, J. E., II., 1999. Trends in atmospheric sulfur and nitrogen species in the eastern United States for 1989-1995. Atmospheric Environment 33,37-49.

Lin, Y. C., Lan, Y. Y., Tsuang B., Engling G., 2008. Long-term spatial distributions and trends of ambient CO concentrations in the central Taiwan Basin, Atmospheric Environment 42, 4320-4331.

Lu, H., Chang, T., 2005. Meterologically adjusted trends of daily maximum ozone concentrations in Taipei, Taiwan. Atmospheric Environment 39, 6491-6501.

Lin, C., Wu, Y.-L., Lai, C.-H., Lin, P.-H., Lai, H.-C., Lin, P.-L., 2004. Experimental investigation of ozone accumulation overnight during a wintertime ozone episode in south Taiwan, Atmospheric Environment 38, 4267-4278.

Leeuw, F. A.A.M., 2000. Trends in ground level ozone concentrations in the European Union. Environmental Science & Policy 3, 189-199.

Shi, J. P., Harrison, R., 1997. Regression modelling of hourly NOx and NO2 concentration in urban air in London. Atmospheric Environment 31 (24), 4081-4094.

Tsai, D.-M., Wu, Y.-L., 2006. Effects of highway networks on ambient ozone concentrations-A case study in southern Taiwan, Atmospheric Environment 40, 4004-4015.

Vingarzan, R., 2004. A review of surface ozone background levels and trends. Atmospheric Environment 38, 3431-3442.

Willmott, C. J., Ackleson, S. G., Davis, R. E., Feddema, J. J., Klink, K. M., Legates, D. R., O’Donnell, J., Rowe, C. M.,1985. Statistic for the Evaluation and Comparison of Models. Journal of Geophysical Research 90,8995-9005.

行政院環境保護署，1999，「空氣品質地區別與時空特性分析－中南部」，EPA-88-FA31-03-0009。

行政院環境保護署，1997，「都會區臭氧污染趨勢分析及防治之研究」，EPA-86-FA42-09-02。

柯忠佑，2007，「南台灣不同光化指標一致性研究」，國立成功大學環境工程研究所論文。

倪國敦，2004，「高高屏地區臭氧趨勢分析與氣象因子相關性之探討」，國立中山大學環境工程研究所碩士論文。

廖坤泉，2008，「大高雄地區主要空氣污染物經氣象因子修正之長期趨勢分析及事件日測站污染型態統計檢定」，國立中山大學環境工程研究所碩士論文。

簡智祥，2007，「高屏地區冬季奈米微粒分佈、來源與成長特性」，國立成功大學環境工程研究所碩士論文。