本研究使用環保署71個測站及氣象局蘭嶼測站自1994年3月至2005年2月的監測資料，以日平均、日間平均、日極大值平均以及月最大值四種資料分類，分析台灣地區季節性臭氧的分佈與變化。 臭氧的資料校驗顯示台灣地區的年小時資料量有系統性的缺損發生，而在臭氧分佈的分析中，日平均與日間平均的結果較為一致，但是到月最大值時，北部高濃度臭氧出現的地點由日平均中的陽明山轉至新店以南的區域，時間則由春季移至夏季。在1994年與2004年全台季節性臭氧濃度的比較中，日平均的增加百分比最高，月最大值最低，四季中則以夏季的增加量最多。 臭氧的趨勢分析顯示日平均及日間平均中，全台多數地區為正趨勢；日極大值平均中，負趨勢的地區有增加的現象而且統計上具有顯著性的區域有減少的情況；月最大值的趨勢分佈中，南投地區於春、夏二季有每年1.2ppb達到90%信心水準的臭氧增加量，而雲嘉南以及台中彰化地區於秋、冬二季也有達到同樣顯著性及強度的臭氧增量。台灣地區中，唯一具有統計上顯著性的負趨勢發生在月最大值中的花蓮和新店，時間分別為春季及夏季。 臭氧頻率的趨勢分析中顯示中、高濃度臭氧增加而低濃度臭氧減少的情形。在依據頻率分佈的測站分類中，高汙染型測站與月最大值的臭氧分佈中濃度高於120ppb的區域相同，顯示以臭氧頻率的分佈來決定測站的分類是一個客觀的方法。背景測站中，蘭嶼測站與恆春測站有相同的氣象特性；高汙染型測站中，三重測站的頻率分佈集中於15ppb以下，占了總頻率的80%以上，顯示交通流量對於臭氧濃度的影響。 Long-term (Mar 1994-Feb 2005) near-surface ozone data from 72 sites have been analyzed to present the seasonal distribution and trends of ozone in Taiwan. In the data quality check, there are the systemic loss of data in some hours of day. The distribution and trends of seasonal ozone concentration are discussed by four classifications: (1) monthly mean; (2) monthly day-time( 08:00-20:00 ) average ; (3) monthly average of daily maxima; (4) maximum of month. The distribution of seasonal ozone concentration of monthly mean and monthly day-time average is similar. The distribution of maximum of month indicates summer maximum in Xin-Dian site. Comparison between average of seasonal ozone of entire Taiwan in 1994 and 2004 indicates increasing ozone. The enhanced amounts are largest in monthly mean, and smallest in maximum of month. It is also pointed out the largest increase of ozone in summer. The trend analyses of monthly mean and monthly day-time average show increasing ozone concentration in most areas. In monthly average of daily maxima, the areas with decreasing trend increase. The trends of maximum of month in Nan-Tou area in spring and summer reach 1.2ppb/yr and are up to 90% confidence levels, so do the area from Tai-Chung to Tai-Nan County. The decreasing ozone trends up to 90% confidence levels only occur in Hua-Lian in spring and in Xin-Dian in summer in maximum of month. Frequency analyses of seasonal ozone show the characteristic that the frequency of the low ozone concentration bins decrease while the higher increase. It was found the property of stations, such as high polluted or clean, could be determined by its frequency distribution. The classification of total stations shows a good agreement with the area of high polluted sites and the distribution of ozone concentration up to 120ppb in maximum of month. The result indicates it is a good way to identify the property of stations by using ozone frequency distribution. The correlation between ozone and NO shows the negative correlation of monthly average of daily maxima in most Taiwan areas, especially up to 0.6 in Nan-Tou in spring and in Tai-Chung and Zhang-Hua in winter. In maximum of month, there are positive correlation in Yi-Lan, the coast of Yun-Lin and Kao-Bing area in all seasons.
