大氣長程傳送對台灣背景臭氧之影響：空氣胞逆軌跡線之群集分析

Effects of Air Long-range Transport on Background Ozone Concentration in Taiwan: The Cluster Analysis of Air Parcel Back Trajectories

彭眾恩(Chong-En Peng)

指導教授 ： 柳中明

國立臺灣大學/理學院/大氣科學研究所/碩士(2009年)

https://doi.org/10.6342/NTU.2009.02493

摘要

蘭嶼背景大氣測站(22°02‘N，121°33’E，324公尺高)自1995年開始進行臭氧之監測。柳(1999)曾指出，臭氧為最能反應季節性氣團變動之大氣化學物質：冬季乾冷北風攜帶高濃度臭氧，而夏季海洋暖濕南風伴隨低濃度臭氧。為了延續先前之議題，進一步了解臭氧和空氣來源的關係，本研究利用群集分析，配合蘭嶼臭氧夜間觀測資料，期求客觀量化分析台灣地區空氣來源分布及隨時間之變化，尋找1995-2007年間變化特徵，並觀察與臭氧濃度間之關聯性。

群集分析(Cluster analysis)為多變數分析方法，其目的是希望利用客觀計量，將事物根據某些屬性歸在各群體中，使同一集群內的事物都具有高度同質性(Homogeneity)。我們利用群集分析中K-means法進行改良來進行分群，最終得到六主要族群：A：從西北方來之族群 (佔全部來源之15%)，B：從北方來之族群 (17%)，C：從北北東方來之族群 (13%)，D：從東北方來之族群 (6%)，E：從東方來之族群 (23%)，F：從西南方來之族群 (21%)。

時間分佈上，依季節區分，冬季主要以大陸性族群A、B為主，夏季則以海洋性族群E、F為主，族群C、D則在秋季較盛行。比對空氣胞族群和蘭嶼臭氧含量，發現族群A、B臭氧含量較高，平均超過45ppbv，族群C、D略低，平均達為43及36ppbv；族群E、F之臭氧含量則明顯較低，平均分別僅22、23ppbv。

在臭氧值年際變化上，大陸性族群臭氧含量以2002為分界，前半段為上升，後半段為下降；海洋性族群則呈逐年上升趨勢。若依臭氧值高低將各族群再分成三類次族群，可發現高臭氧次族群其路徑較偏向排放源。藉由冬季兩個案模擬，也可發現偏向陸路到達台灣之個案，由於其路徑後期亦受排放源影響，因此抵達蘭嶼時之臭氧值較經由海路路徑的個案高。

關鍵字

長程傳送 ； 背景臭氧 ； 群集分析

並列摘要

Lanyu background station (22°02 N, 121°33 E, height:324m) has monitored ozone since 1995. Liu et al.(1999) observed 4-year Lanyu data, indicated that ozone was the best gas to response to seasonal air masses change. In order to continue his research about air sources and ozone background concentration of Taiwan, we used cluster analysis to separate all air parcel trajectories from 1995 to 2007 into groups, to find out the spatial and temporal characteristics of air parcels of Taiwan. We also compared back trajectories groups with background ozone concentration in Lanyu to find out ozone characteristics and long-term change trends of groups.

Cluster Analysis is a multivariate analysis technique that seeks to organize information about variables so that relatively homogeneous groups can be formed. In all methods of clustering, we used K-means method for clustering air parcels to Taiwan, and finally got 6 groups: group A from northwest (accounting for 15% of all trajectories), group B from north (17%), group C from north-northeast (13%), group D from northeast (6%), group E from east (23%), and group F from southwest (21%).

From the analysis of seasonal ozone distribution, groups from continental (A、B) prevailed in winter, group from ocean(E、F) prevail in summer, and group C and D prevailed in autumn. Comparing air parcel group with Lanyu ozone, we found that group A and B had larger mean ozone concentration, which was higher than 45 ppbv; the mean ozone concentration of group C and group D was 43ppbv and 36ppbv, respectively; group E and F was evidently lower mean ozone concentration than other groups, only 22 ppbv and 23 ppbv.

In terms of ozone yearly trend, there were two phases of continental group: before 2002, ozone concentration increased; after 2002, ozone concentration decreased. Oceanic group had lasting increase trend. According to ozone concentration, we separated each group into three sub-groups. The mean trajectory of high ozone sub-group was closer to emission zone than other two sub-groups. Using case study, we found because the ozone values of trajectories whose routes were near to emission zone still affected by emission when parcels approached Lanyu, ozone concentration of near emission air parcels was higher than air parcels whose routes far from emission zone.

並列關鍵字

long-range transport ； background ozone ； cluster analysis

參考文獻

吳承翰，2002：亞洲沙塵暴之模擬，國立中央大學，大氣物理研究所碩士論文。

王作台、胡志文、林松錦、柳中明，1994：氣象環流對背影氣成份傳送影響：第一部分：軌跡方法之建立及個案分析結果。大氣科學，22，57-76。

柳中明，1992：台灣背景大氣研究：(1)中美合作「氣候與空氣品質台灣站(CATS)研究發展與西太平洋地區觀測(PEM-west)研究計劃及背景大氣中甲烷(CH4)測量分析」。NSC810421M0023224z。

柳中明，張修武，劉紹臣，1999：蘭嶼背景大氣監測初析。大氣科學，27，99-130。

Abdalmogith , S.S., and R.M. Harrison, 2005: The use of trajectory cluster analysis to examine the long-range transport of secondary inorganic aerosol in the UK, Atmospheric Environment 39, pp. 6686–6695.