臺北地區土地利用回歸模式評估氮氧化物個人暴露量之研究

Using Land Use Regression Models to Estimate Individual Exposure to NOx and NO2 in Taipei Metropolis

朱郁琇(Yu-Hsiu Chu)

指導教授 ： 詹長權

國立臺灣大學/公共衛生學院/職業醫學與工業衛生研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.02876

摘要

背景－高度都市化的臺北地區不僅地狹人稠，工商業和交通運輸蓬勃發展下，伴隨而來的空氣汙染問題也逐漸受到重視，過去大多數研究在臺北地區進行世代暴露評估時往往受限於經費人力，僅能使用環保署空氣品質監測站數據作為居民空氣汙染物暴露程度的指標。近年來，歐美地區相關研究透過土地利用回歸模式方法論在有限的研究資源下發展出能夠有效推估都市內部空氣汙染物受到地理環境影響的濃度分布情況。因此，本研究在臺北地區發展土地利用回歸模式，探討土地利用與交通排放對於氮氧化物濃度空間分布的貢獻，並將之應用於新生兒暴露評估與不同暴露評估方法比較。

方法－本研究於2009-2010年在臺北地區依據人口密度分布選取了水平方向40個家戶採樣點進行氮氧化物細部監測網絡，包括18個交通背景監測點和22個都市背景監測點，同時亦在21個高低樓層家戶採樣點進行氮氧化物在垂直方向上的濃度分佈探討。氮氧化物係以被動式採樣器進行為期三季每次採集14天的濃度監測，並將採樣結果結合土地利用資料透過逐步回歸分析法發展土地利用回歸模式。此外，本研究透過三種評估方法進行新生兒世代的氮氧化物個人暴露量評估，包含一般克利金模式、最近測站法、土地利用回歸模式，並透過繪圖呈現三種評估方法的空間解析度探討。

結果－本研究於水平方向的氮氧化物採樣結果，以連續測站校正後的NOx和NO2的年平均濃度在18個交通背景監測點的部分各別為55.4 ppb ± 13.4 ppb和28.7 ppb ± 6.4 ppb；而22個都市背景監測點則為39.4 ppb ± 11.1 ppb和23.8 ppb ± 5.9 ppb，且交通監測點的氮氧化物濃度皆顯著高於都市背景監測點。垂直方向方向部分，高樓層採樣點的NOx和NO2濃度為31.14 ± 15.29 ppb和18.57 ± 7.59 ppb；而低樓層的NOx和NO2濃度為41.74 ± 16.12 ppb和24.80 ± 7.04 ppb，達到統計上的顯著差異，並可以高低樓層比值0.74作為土地利用回歸模式校正參考用。另一方面，以逐步回歸分析結合土地利用資料和水平採樣結果成功建立出氮氧化物土地利用回歸模式，結果顯示透過6個土地利用變項可以解釋臺北地區內NOx濃度77%的空間分布變異，而透過4個土地利用變項即可解釋NO2濃度70%的空間變異程度。新生兒世代暴露評估結果和繪圖方法比較結果顯示，土地利用回歸模式相較其他方法可以獲得更高的空間解析度。

結論－本研究成功地在臺北地區建立氮氧化物的土地利用回歸模式，透過4-6個土地利用相關變項，即可有效預估70%-77%的氮氧化物空間分布變異，相較其他評估方法更可以達到更高的空間解析度。

關鍵字

土地利用回歸模式 ； 暴露評估 ； 地理資訊系統 ； 交通排放 ； 垂直分布 ； 氮氧化物 ； 二氧化氮

並列摘要

Background－This study developed land use regression models to assess NOx and NO2 ambient concentrations in order to minimize the exposure misclassification and achieve better individual exposure assessment for the future epidemiology studies.

Method－A population density based sampling strategy was used to select 40 locations as NOx and NO2 monitoring sites in the central region of Taipei Metropolis, including 18 traffic sites and 22 urban sites. One additional background site was selected for annual average adjustment. NOx and NO2 concentrations were monitored simultaneously by using Ogawa passive samplers for 2-week period in three seasons during 2009 and 2010. With collected land use information containing land-cover and traffic related data, a series of potential predictor variables in concentric circle buffers with several radii ranging from 25 to 5000m of each monitoring site were obtained by using GIS system. Based on stochastic modeling techniques, we combined the average concentrations and values of predictor variables in stepwise procedure to develop the final LUR models for NOx and NO2.

Results－The adjusted annual mean concentrations were 46.6±14.5 ppb for NOx and 26.0±6.5 ppb for NO2. Vertical variation (High/Low ratio =0.74) was discovered for the modification of LUR models. There were 6 variables for the final model of NOx (adjusted R2=0.77), including the major road length within 25m, urban green area within 300m, urban green area in 300-5000m buffer, major road length in 50-500m buffer, major road length in 25-50m buffer, and natural area within 500m. And there were 4 predictors for the NO2 model (adjusted R2=0.70), including the natural area within 500m, major road length within 25m, industry-residential area within 500m, and urban green area within 100m. One birth cohort was indentified for exposure assessment by three approaches including Ordinary Kriging, nearest station method, and developed LUR models. LUR models showed moderate spatial variation and higher resolution than other methods.

Conclusion－This study indicates that the ambient concentration of NOx and NO2 can be well estimated by LUR models with 4-6 variables in Taipei Metropolis, and LUR approach provides better spatial resolutions than interpolation method or regulatory stations.

並列關鍵字

Land use regression ； Exposure assessment ； Geographic information system (GIS) ； traffic emission ； vertical distribution ； Nitrogen oxides (NOx) ； nitrogen dioxide (NO2)

參考文獻

[3] Ott, W.andR. Eliassen. "Survay technique for determining representativeness of urban air monitoring stations with respect to carbon-monoxide." Journal of the Air Pollution Control Association 23(8): 685-690. (1973)

[4] Chan, C. C.andJ. S. Hwang. "Site representativeness of urban air monitoring stations." Journal of the Air & Waste Management Association 46(8): 755-760. (1996)

[5] Williams, R.,P. Jones, et al. "The influence of human and environmental exposure factors on personal NO2 exposures." Journal of Exposure Science and Environmental Epidemiology 22(2): 109-115. (2012)

[8] Chan, C. C.,Y. Yanagisawa, et al. "Personal and indoor outdoor nitrogen-dioxide exposure assessments of 23 homes in Taiwan." Toxicology and Industrial Health 6(1): 173-182. (1990)

[9] Lewne, M.,J. Cyrys, et al. "Spatial variation in nitrogen dioxide in three European areas." Science of the Total Environment 332(1-3): 217-230. (2004)