近地表臭氧 (O3) 是一種空氣污染物，過去的研究已證實，長期或高劑量暴露於O3會對人體健康造成負面影響。O3是一種二次污染物，交通排放為其前驅物主要來源之一，根據臺灣民眾的生活習慣，每日有兩個主要通勤時段，分別在上午7至9時以及下午4至6時，在這些時段中，繁忙的交通行為造成大量前驅物被排放，此時亦為民眾主要暴露於室外空氣污染的時段。就研究者所知，目前仍缺乏在精細空間尺度下通勤時段的O3濃度分布研究，為釐清通勤時段室外空氣污染時空分布，本研究以O3為例，基於地理人工智慧 (Geo-AI)，首先將O3觀測值和可能的潛在影響因子整理至空間變數資料庫，經過變數篩選步驟取得重要變數組後，透過集成學習整合克利金空間內插法、土地利用迴歸法和機器學習演算法發展模型，在50 m × 50 m網格解析度下，模擬並推估1994年至2019年間臺灣本島上午與下午通勤時段O3濃度。
本研究結果顯示上午和下午通勤時段之最終模型都具有高模型解釋力，兩個模型R2皆達0.911，透過80%/20%資料分割驗證、十折交叉驗證與外部時間驗證，皆證明了我們的模型推估的穩健性、準確性以及外推能力，與傳統土地利用迴歸模型相比，推估效能得到大幅提升。在本研究中，探討了篩選出的重要變數組與O3濃度的關聯，而由模型繪製的空間推估圖視覺化地展示了空氣污染情形，協助讀者更好地理解通勤時段中O3濃度模擬的重要解釋變數以及O3濃度時空分布。整體而言，本研究建立模擬模型以釐清過去長期之通勤時段O3濃度分布，為空氣品質管制、廢氣排放相關政策和環境醫學領域等提供參考。

Near-surface ozone (O3) is a worldwide air pollutant. Previous studies had demonstrated that long-term or high-dose exposure to O3 can have negative effects on human health. O3 is a secondary pollutant, and traffic emission is one of the main sources of its precursors. According to the commuting habits of Taiwanese people, there are two main periods of commuting each day, namely from 7 to 9 am and from 4 to 6 pm. During these periods, heavy traffic generates a substantial amount of pollutant precursors, leading to increased outdoor air pollution exposure for the general public. To elucidate the temporal and spatial distribution of O3 during commuting, this study developed models based on geographic artificial intelligence (Geo-AI). After organizing ozone and potential influencing factors into the database, important variables were obtained through variable screening, and machine learning and ensemble learning were utilized to develop a model to simulate the O3 concentration of Taiwan's main island during morning and dusk commuting periods with high spatial resolution. The results shown that the final models have high model explanatory power with R2 of both 0.911. The robustness, accuracy and extrapolation ability of the model were validated through data splitting validation, 10-fold cross-validation and external time validation. Moreover, the relationship between selected variables and O3 concentrations were discussed. And spatial estimation maps were produced, helping to clarify the simulation of O3 concentrations during commuting.

中華民國交通部中央氣象局。中華民國111年日出日沒時刻表。檢自：https://www.cwb.gov.tw/Data/astronomy/2022/sundat/01taipei.pdf (Feb. 6, 2023)
中華民國交通部機動車輛數簡易資料查詢網頁。機動車輛登記。檢自：https://stat.motc.gov.tw/mocdb/stmain.jsp?sys=100&funid=a3301 (Feb. 6, 2023)
中華民國行政院全球資訊網。國情簡介，土地與人民，人民 (民國111年9月1日)。檢自：https://www.ey.gov.tw/state/99B2E89521FC31E1/835a4dc2-2c2d-4ee0-9a36-a0629a5de9f0 (Feb. 6, 2023a)
中華民國行政院全球資訊網。國情簡介，土地與人民，土地 (民國111年12月14日)。檢自：https://www.ey.gov.tw/state/4447F4A951A1EC45/094b1d53-de8d-4393-bde6-ab092969cce4 (Feb. 6, 2023b)
中華民國行政院環保署空氣品質監測網。空氣品質與日常生活。檢自：https://airtw.epa.gov.tw/CHT/Encyclopedia/pedia08/pedia8_8.aspx (Feb. 4, 2023a)
中華民國行政院環保署空氣品質監測網。臭氧及氮氧化物濃度日變化。檢自：https://airtw.epa.gov.tw/cht/Encyclopedia/pedia09/pedia9_2.aspx (Feb. 4, 2023b)
臺灣電力公司。火力發電概況。檢自：https://www.taipower.com.tw/tc/page.aspx?mid=202&cid=129&cchk=675cea43-9c45-4ae1-80c6-4f18b3b38d8e (Feb. 8, 2023)
Ainsworth, E. A., Yendrek, C. R., Sitch, S., Collins, W. J., & Emberson, L. D. (2012). The effects of tropospheric ozone on net primary productivity and implications for climate change. Annual review of plant biology, 63, 637-661.
Archibald, A. T., Neu, J. L., Elshorbany, Y. F., Cooper, O. R., Young, P. J., Akiyoshi, H., ... & Zeng, G. (2020). Tropospheric Ozone Assessment ReportA critical review of changes in the tropospheric ozone burden and budget from 1850 to 2100. Elementa: Science of the Anthropocene, 8 (1).
Babaan, J., Hsu, F. T., Wong, P. Y., Chen, P. C., Guo, Y. L., Lung, S. C. C., ... & Wu, C. D. (2023). A Geo-AI-based Ensemble Mixed Spatial Prediction Model with Fine Spatial-Temporal Resolution for Estimating Daytime/Nighttime/Daily Average Ozone Concentrations Variations in Taiwan. Journal of Hazardous Materials, 130749.
Bell, M. L., McDermott, A., Zeger, S. L., Samet, J. M., & Dominici, F. (2004). Ozone and short-term mortality in 95 US urban communities, 1987-2000. Jama, 292(19), 2372-2378.
Breiman, L. (2001). Random forests. Machine learning, 45, 5-32.
Briggs, D. J., Collins, S., Elliott, P., Fischer, P., Kingham, S., Lebret, E., ... & Van Der Veen, A. (1997). Mapping urban air pollution using GIS: a regression-based approach. International Journal of Geographical Information Science, 11(7), 699-718.
Cai, J., Luo, J., Wang, S., & Yang, S. (2018). Feature selection in machine learning: A new perspective. Neurocomputing, 300, 70-79.
Chen, C. L., Tsuang, B. J., Tu, C. Y., Cheng, W. L., & Lin, M. D. (2002). Wintertime vertical profiles of air pollutants over a suburban area in central Taiwan. Atmospheric Environment, 36 (12), 2049-2059.
Chen, L., Bai, Z., Kong, S., Han, B., You, Y., Ding, X., ... & Liu, A. (2010). A land use regression for predicting NO2 and PM10 concentrations in different seasons in Tianjin region, China. Journal of Environmental Sciences, 22 (9), 1364-1373.
Chen, L., Liang, S., Li, X., Mao, J., Gao, S., Zhang, H., ... & Azzi, M. (2021). A hybrid approach to estimating long-term and short-term exposure levels of ozone at the national scale in China using land use regression and Bayesian maximum entropy. Science of The Total Environment, 752, 141780.
Chen, Y. Y., Sung, F. C., Chen, M. L., Mao, I. F., & Lu, C. Y. (2016). Indoor air quality in the metro system in north Taiwan. International journal of environmental research and public health, 13 (12), 1200.
Coates, J., Mar, K. A., Ojha, N., & Butler, T. M. (2016). The influence of temperature on ozone production under varying NOx conditions–a modelling study. Atmospheric Chemistry and Physics, 16 (18), 11601-11615.
Cooper, O. R., Parrish, D. D., Ziemke, J., Balashov, N. V., Cupeiro, M., Galbally, I. E., ... & Zbinden, R. M. (2014). Global distribution and trends of tropospheric ozone: An observation-based reviewGlobal distribution and trends of tropospheric ozone. Elementa: Science of the Anthropocene, 2.
Di, Q., Amini, H., Shi, L., Kloog, I., Silvern, R., Kelly, J., ... & Schwartz, J. (2019). Assessing NO2 concentration and model uncertainty with high spatiotemporal resolution across the contiguous United States using ensemble model averaging. Environmental science & technology, 54(3), 1372-1384.
Dimakopoulou, K., Douros, J., Samoli, E., Karakatsani, A., Rodopoulou, S., Papakosta, D., ... & Katsouyanni, K. (2020). Long-term exposure to ozone and children's respiratory health: Results from the RESPOZE study. Environmental research, 182, 109002.
Friedman, J. H. (2001). Greedy function approximation: a gradient boosting machine. Annals of statistics, 1189-1232.
Gaudel, A., Cooper, O. R., Ancellet, G., Barret, B., Boynard, A., Burrows, J. P., ... & Ziemke, J. (2018). Tropospheric Ozone Assessment Report: Present-day distribution and trends of tropospheric ozone relevant to climate and global atmospheric chemistry model evaluation. Elementa: science of the anthropocene, 6.
Geng, F., Tie, X., Guenther, A., Li, G., Cao, J., & Harley, P. (2011). Effect of isoprene emissions from major forests on ozone formation in the city of Shanghai, China. Atmospheric Chemistry and Physics, 11(20), 10449-10459.
Gilpin, L. H., Bau, D., Yuan, B. Z., Bajwa, A., Specter, M., & Kagal, L. (2018, October). Explaining explanations: An overview of interpretability of machine learning. In 2018 IEEE 5th International Conference on data science and advanced analytics (DSAA) (pp. 80-89). IEEE.
Hu, J., Chen, J., Ying, Q., & Zhang, H. (2016). One-year simulation of ozone and particulate matter in China using WRF/CMAQ modeling system. Atmospheric Chemistry and Physics, 16 (16), 10333-10350.
Jerrett, M., Burnett, R. T., Pope III, C. A., Ito, K., Thurston, G., Krewski, D., ... & Thun, M. (2009). Long-term ozone exposure and mortality. New England Journal of Medicine, 360 (11), 1085-1095.
Kang, Y., Choi, H., Im, J., Park, S., Shin, M., Song, C. K., & Kim, S. (2021). Estimation of surface-level NO2 and O3 concentrations using TROPOMI data and machine learning over East Asia. Environmental Pollution, 288, 117711.
Kavassalis, S. C., & Murphy, J. G. (2017). Understanding ozone‐meteorology correlations: A role for dry deposition. Geophysical Research Letters, 44(6), 2922-2931.
Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., ... & Liu, T. Y. (2017). Lightgbm: A highly efficient gradient boosting decision tree. Advances in neural information processing systems, 30.
Kerckhoffs, J., Wang, M., Meliefste, K., Malmqvist, E., Fischer, P., Janssen, N. A., ... & Hoek, G. (2015). A national fine spatial scale land-use regression model for ozone. Environmental research, 140, 440-448.
Klein, A., Ravetta, F., Thomas, J. L., Ancellet, G., Augustin, P., Wilson, R., ... & Pelon, J. (2019). Influence of vertical mixing and nighttime transport on surface ozone variability in the morning in Paris and the surrounding region. Atmospheric Environment, 197, 92-102.
Kleinman, L. I., Daum, P. H., Imre, D., Lee, Y. N., Nunnermacker, L. J., Springston, S. R., ... & Rudolph, J. (2002). Ozone production rate and hydrocarbon reactivity in 5 urban areas: A cause of high ozone concentration in Houston. Geophysical Research Letters, 29 (10), 105-1.
Li, H. C., Chiueh, P. T., Liu, S. P., & Huang, Y. Y. (2017). Assessment of different route choice on commuters’ exposure to air pollution in Taipei, Taiwan. Environmental Science and Pollution Research, 24, 3163-3171.
Li, Y., Liu, L., Wei, Q., Ren, X., & An, M. (2021). Highly selective separation of acetic acid and hydrochloric acid by alkylamide based on double hydrogen bond coupling mechanism. Separation and Purification Technology, 275, 119110.
Li, Z. (2022). Extracting spatial effects from machine learning model using local interpretation method: An example of SHAP and XGBoost. Computers, Environment and Urban Systems, 96, 101845.
Lin, H., Wang, M., Duan, Y., Fu, Q., Ji, W., Cui, H., ... & Hu, K. (2020). O3 sensitivity and contributions of different NMHC sources in O3 formation at urban and suburban sites in Shanghai. Atmosphere, 11(3), 295.
Lin, X., Trainer, M., & Liu, S. C. (1988). On the nonlinearity of the tropospheric ozone production. Journal of Geophysical Research: Atmospheres, 93 (D12), 15879-15888.
Liu, Y., & Zheng, Y. F. (2006). FS_SFS: A novel feature selection method for support vector machines. Pattern recognition, 39(7), 1333-1345.
Liu, Y., Wang, H., Jing, S., Gao, Y., Peng, Y., Lou, S., ... & An, J. (2019). Characteristics and sources of volatile organic compounds (VOCs) in Shanghai during summer: Implications of regional transport. Atmospheric Environment, 215, 116902.
Lu, X., Zhang, L., & Shen, L. (2019). Meteorology and climate influences on tropospheric ozone: a review of natural sources, chemistry, and transport patterns. Current Pollution Reports, 5, 238-260.
Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in neural information processing systems, 30.
Lyu, X. P., Guo, H., Wang, N., Simpson, I. J., Cheng, H. R., Zeng, L. W., ... & Blake, D. R. (2017). Modeling C1–C4 alkyl nitrate photochemistry and their impacts on O3 production in urban and suburban environments of Hong Kong. Journal of Geophysical Research: Atmospheres, 122(19), 10-539.
Ma, X., Longley, I., Gao, J., & Salmond, J. (2020). Assessing schoolchildren's exposure to air pollution during the daily commute-A systematic review. Science of the Total Environment, 737, 140389.
Maji, K. J., Ye, W. F., Arora, M., & Nagendra, S. S. (2019). Ozone pollution in Chinese cities: Assessment of seasonal variation, health effects and economic burden. Environmental pollution, 247, 792-801.
McCulloch, A. (1999). CFC and Halon replacements in the environment. Journal of Fluorine chemistry, 100(1-2), 163-173.
Michael, R., O’Lenick, C. R., Monaghan, A., Wilhelmi, O., Wiedinmyer, C., Hayden, M., & Estes, M. (2019). Application of geostatistical approaches to predict the spatio-temporal distribution of summer ozone in Houston, Texas. Journal of Exposure Science & Environmental Epidemiology, 29 (6), 806-820.
Miller, T. H., Gallidabino, M. D., MacRae, J. I., Hogstrand, C., Bury, N. R., Barron, L. P., ... & Owen, S. F. (2018). Machine learning for environmental toxicology: a call for integration and innovation.
Mills, G., Pleijel, H., Malley, C. S., Sinha, B., Cooper, O. R., Schultz, M. G., ... & Lewis, A. (2018). Tropospheric Ozone Assessment Report: Present-day tropospheric ozone distribution and trends relevant to vegetation. Elementa: Science of the Anthropocene, 6.
Mirowsky, J. E., Carraway, M. S., Dhingra, R., Tong, H., Neas, L., Diaz-Sanchez, D., ... & Devlin, R. B. (2017). Ozone exposure is associated with acute changes in inflammation, fibrinolysis, and endothelial cell function in coronary artery disease patients. Environmental Health, 16 (1), 1-10.
Placet, M., Mann, C. O., Gilbert, R. O., & Niefer, M. J. (2000). Emissions of ozone precursors from stationary sources:: a critical review. Atmospheric Environment, 34(12-14), 2183-2204.
Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2018). CatBoost: unbiased boosting with categorical features. Advances in neural information processing systems, 31.
Ren, X., Mi, Z., & Georgopoulos, P. G. (2020). Comparison of Machine Learning and Land Use Regression for fine scale spatiotemporal estimation of ambient air pollution: Modeling ozone concentrations across the contiguous United States. Environment international, 142, 105827.
Requia, W. J., Di, Q., Silvern, R., Kelly, J. T., Koutrakis, P., Mickley, L. J., ... & Schwartz, J. (2020). An ensemble learning approach for estimating high spatiotemporal resolution of ground-level ozone in the contiguous United States. Environmental science & technology, 54 (18), 11037-11047.
Ryan, P. H., & LeMasters, G. K. (2007). A review of land-use regression models for characterizing intraurban air pollution exposure. Inhalation toxicology, 19 (sup1), 127-133.
Sarwar, G., & Bhave, P. V. (2007). Modeling the effect of chlorine emissions on ozone levels over the eastern United States. Journal of applied meteorology and climatology, 46(7), 1009-1019.
Shahriar, S. A., Kayes, I., Hasan, K., Hasan, M., Islam, R., Awang, N. R., ... & Salam, M. A. (2021). Potential of Arima-ann, Arima-SVM, dt and catboost for atmospheric PM2. 5 forecasting in bangladesh. Atmosphere, 12 (1), 100.
Shapley, L. S. (1953). A value for n-person games.
Sharma, S., Sharma, P., & Khare, M. (2017). Photo-chemical transport modelling of tropospheric ozone: A review. Atmospheric Environment, 159, 34-54.
Shtein, A., Kloog, I., Schwartz, J., Silibello, C., Michelozzi, P., Gariazzo, C., ... & Stafoggia, M. (2019). Estimating daily PM2. 5 and PM10 over Italy using an ensemble model. Environmental science & technology, 54(1), 120-128.
Sillman, S. (1999). The relation between ozone, NOx and hydrocarbons in urban and polluted rural environments. Atmospheric Environment, 33 (12), 1821-1845.
Stafoggia, M., Johansson, C., Glantz, P., Renzi, M., Shtein, A., de Hoogh, K., ... & Bellander, T. (2020). A random forest approach to estimate daily particulate matter, nitrogen dioxide, and ozone at fine spatial resolution in Sweden. Atmosphere, 11 (3), 239.
Tong, N. Y., Leung, D. Y., & Liu, C. H. (2011). A review on ozone evolution and its relationship with boundary layer characteristics in urban environments. Water, Air, & Soil Pollution, 214, 13-36.
Tseng, P. H., Lin, D. Y., & Chien, S. (2014). Investigating the impact of highway electronic toll collection to the external cost: A case study in Taiwan. Technological Forecasting and Social Change, 86, 265-272.
Wang, T., Xue, L., Brimblecombe, P., Lam, Y. F., Li, L., & Zhang, L. (2017). Ozone pollution in China: A review of concentrations, meteorological influences, chemical precursors, and effects. Science of the Total Environment, 575, 1582-1596.
Watson, G. L., Telesca, D., Reid, C. E., Pfister, G. G., & Jerrett, M. (2019). Machine learning models accurately predict ozone exposure during wildfire events. Environmental Pollution, 254, 112792.
Wei, J., Li, Z., Li, K., Dickerson, R. R., Pinker, R. T., Wang, J., ... & Cribb, M. (2022). Full-coverage mapping and spatiotemporal variations of ground-level ozone (O3) pollution from 2013 to 2020 across China. Remote Sensing of Environment, 270, 112775.
Wolf, K., Cyrys, J., Harciníková, T., Gu, J., Kusch, T., Hampel, R., ... & Peters, A. (2017). Land use regression modeling of ultrafine particles, ozone, nitrogen oxides and markers of particulate matter pollution in Augsburg, Germany. Science of the Total Environment, 579, 1531-1540.
Wolpert, D. H. (1992). Stacked generalization. Neural networks, 5(2), 241-259.
Wong, D. W., Yuan, L., & Perlin, S. A. (2004). Comparison of spatial interpolation methods for the estimation of air quality data. Journal of Exposure Science & Environmental Epidemiology, 14 (5), 404-415.
Wong, P. Y., Hsu, C. Y., Wu, J. Y., Teo, T. A., Huang, J. W., Guo, H. R., ... & Spengler, J. D. (2021). Incorporating land-use regression into machine learning algorithms in estimating the spatial-temporal variation of carbon monoxide in Taiwan. Environmental Modelling & Software, 139, 104996.
Wu, C. D., Zeng, Y. T., & Lung, S. C. C. (2018). A hybrid kriging/land-use regression model to assess PM2. 5 spatial-temporal variability. Science of the Total Environment, 645, 1456-1464.
Wu, S., Mickley, L. J., Leibensperger, E. M., Jacob, D. J., Rind, D., & Streets, D. G. (2008). Effects of 2000–2050 global change on ozone air quality in the United States. Journal of Geophysical Research: Atmospheres, 113 (D6).
Xu, J., Wang, A., Schmidt, N., Adams, M., & Hatzopoulou, M. (2020). A gradient boost approach for predicting near-road ultrafine particle concentrations using detailed traffic characterization. Environmental Pollution, 265, 114777.
Yin, P., Chen, R., Wang, L., Meng, X., Liu, C., Niu, Y., ... & Kan, H. (2017). Ambient ozone pollution and daily mortality: a nationwide study in 272 Chinese cities. Environmental health perspectives, 125 (11), 117006.
Zamani Joharestani, M., Cao, C., Ni, X., Bashir, B., & Talebiesfandarani, S. (2019). PM2. 5 prediction based on random forest, XGBoost, and deep learning using multisource remote sensing data. Atmosphere, 10 (7), 373.
Zhai, B., & Chen, J. (2018). Development of a stacked ensemble model for forecasting and analyzing daily average PM2. 5 concentrations in Beijing, China. Science of the Total Environment, 635, 644-658.
Zhan, Y., Luo, Y., Deng, X., Grieneisen, M. L., Zhang, M., & Di, B. (2018). Spatiotemporal prediction of daily ambient ozone levels across China using random forest for human exposure assessment. Environmental Pollution, 233, 464-473.
Zhang, L., Jacob, D. J., Downey, N. V., Wood, D. A., Blewitt, D., Carouge, C. C., ... & Wang, Y. (2011). Improved estimate of the policy-relevant background ozone in the United States using the GEOS-Chem global model with 1/2× 2/3 horizontal resolution over North America. Atmospheric Environment, 45 (37), 6769-6776.