學術產出-Theses
Article View/Open
Publication Export
-
題名 使用衛星數據監測台灣空氣汙染物遠程輸送，優化PM2.5預測
Using Satellite Data on Remote Transportation of Air Pollutants for PM2.5 Prediction in Taiwan作者 喬治·威廉·基比里奇
Kibirige, George william貢獻者 陳孟彰<br>劉昭麟
Meng Chang Chen<br>Chao-Lin Liu
喬治·威廉·基比里奇
George william Kibirige關鍵詞 PM2.5
空氣污染
臺灣
PM2.5
Air Pollution
Composite Neural Network
Taiwan日期 2023 上傳時間 2-Jan-2024 15:22:16 (UTC+8) 摘要 精確的PM2.5預測是對抗空氣污染的重要一環，有助於政府制定環境政策。透過多角度大氣校正（MAIAC）算法處理的衛星遙感氣膠光學深度（AOD）資料，我們能夠觀察遙遠地區之間污染物的傳輸情況。在本論文中，我們提出了一種結合神經網絡模型的方法，稱為遠程傳輸污染（RTP）模型，可更準確預測當地的PM2.5濃度。本論文所提出的RTP模型結合了多個深度學習元件，並從異質特徵中學習。此外，我們還提出的分類算法，來檢測從AOD數據中識別出的兩個參考站點的遠程傳輸污染事件（RTPEs）。我們在台灣的北部地區和南部中部地區進行了大量模擬，以評估其在實際數據上的表現。對於北部地區，提出的RTP模型在+4小時至+24小時、+28小時至+48小時和+52小時至+72小時的時間範圍內，相對不考慮RTPEs的基本模型，提高了17%至30%、23%至26%和18%至22%的準確度，同時也優於只考慮RTPEs的最先進模型，提高了12%至22%、12%至14%和10%至11%的準確度 其他影響PM2.5濃度的特徵，陸海風在台灣南部和中部地區扮演著關鍵角色。我們在提出的RTP模型中使用了大面積海洋風特徵，以捕捉陸海風對PM2.5的影響。我們對RTP模型的不同組成特徵個別進行研究，結果顯示不管再提出特徵跟模型改進，我們提出的架構在整體性能方面都有明顯改進。我們還進行了月份分析，證明了在台灣南部和中部地區，陸海風經常發生並主導PM2.5濃度。
Accurate PM2.5 prediction is part of the fight against air pollution that helps governments to manage environmental policy. Satellite Remote sensing aerosol optical depth (AOD) pro- cessed by The Multi-Angle Implementation of Atmospheric Correlation (MAIAC) algorithm allows us to observe the transportation of remote pollutants between regions. In this dissertation we propose a composite neural network model, the Remote Trans- ported Pollutants (RTP) model, for such long-range pollutant transportation that predicts more accurate local PM2.5 concentrations using such satellite data. The proposed RTP model integrates several deep learning components and learns from the heterogeneous features of various domains. We also detected remote transportation pollution events (RTPEs) at two reference sites from the AOD data using proposed classification algorithm. Extensive experi- ments using real-world data were conducted in two regions, northern region and southern and central region of Taiwan. For northern region the proposed RTP model outperforms the base model that does not account for RTPEs by 17%-30%, 23%-26% and 18%-22% and state-of-the-art models that account for RTPEs by 12%-22%, 12%-14%, and 10%-11% at +4h to +24h, +28h to +48 hours, and +52h to +72h hours respectively. Including other features that affect PM2.5, Land-sea breeze plays an important role in increasing the level of PM2.5 in southern and central region of Taiwan. We used extra feature of large interpolation ocean wind in our proposed RTP model to capture the impact of land-sea breeze. We investigated the model outputs of different components of the RTP and concluded that the proposed composite neural network architecture yields significant improvements in the overall performance compared to each component and the other state-of-the-art model. We performed monthly analysis which also demonstrated the superiority of the proposed architecture for stations where land-sea breezes frequently occur in the southern and central region of Taiwan in the months when land-sea breeze dominates the accumulation of PM2.5.參考文獻 [1] Jai Shanker Pandey, Rakesh Kumar, and Sukumar Devotta. Health risks of NO2, SPM and SO2 in Delhi (India). Atmospheric Environment, 39(36):6868–6874, 2005. [2] Mihye Lee, Itai Kloog, Alexandra Chudnovsky, Alexei Lyapustin, Yujie Wang, Steven Melly, Brent Coull, Petros Koutrakis, and Joel Schwartz. Spatiotemporal prediction of fine particulate matter using high-resolution satellite images in the Southeastern US 2003–2011. Journal of exposure science & environmental epidemiology, 26(4):377–384, 2016. [3] Ssu-Ting Liu, Chu-Yung Liao, Cheng-Yu Kuo, and Hsien-Wen Kuo. The effects of pm2.5 from asian dust storms on emergency room visits for cardiovascular and respiratory diseases. International journal of environmental research and public health, 14(4):428, 2017. [4] Zhihao Song, Bin Chen, and Jianping Huang. Combining himawari-8 aod and deep forest model to obtain city-level distribution of pm2.5 in china. Environmental Pollution, 297:118826, 2022. [5] Mark D Gibson, Soumita Kundu, and Mysore Satish. Dispersion model evaluation of pm2.5, nox and so2 from point and major line sources in nova scotia, canada using aermod gaussian plume air dispersion model. Atmospheric Pollution Research, 4(2):157–167, 2013. [6] Ziyue Chen, Danlu Chen, Chuanfeng Zhao, Mei-po Kwan, Jun Cai, Yan Zhuang, Bo Zhao, Xiaoyan Wang, Bin Chen, Jing Yang, et al. Influence of meteorological conditions on pm2.5 concentrations across china: A review of methodology and mechanism. Environment international, 139:105558, 2020. [7] Wei-Ting Hung, Cheng-Hsuan Sarah Lu, Sheng-Hsiang Wang, Sheng-Po Chen, Fujung Tsai, and Charles C-K Chou. Investigation of long-range transported pm2.5 events over northern taiwan during 2005–2015 winter seasons. Atmospheric Environment, 217:116920, 2019. [8] Qiangqiang Xu, Xiaoling Chen, Shangbo Yang, Linling Tang, and Jiadan Dong. Spatiotemporal relationship between himawari-8 hourly columnar aerosol optical depth (aod) and ground-level pm2.5 mass concentration in mainland china. Science of the Total Environment, 765:144241, 2021. [9] Jianping Guo, Feng Xia, Yong Zhang, Huan Liu, Jing Li, Mengyun Lou, Jing He, Yan Yan, Fu Wang, Min Min, et al. Impact of diurnal variability and meteorological factors on the pm2.5-aod relationship: Implications for pm2.5 remote sensing. Environmental Pollution, 221:94–104, 2017. [10] Alexei Lyapustin, Yujie Wang, Sergey Korkin, and Dong Huang. MODIS Collection 6 MAIAC algorithm. Atmospheric Measurement Techniques, 11(10), 2018. [11] Minghui Tao, Jun Wang, Rong Li, Lili Wang, Lunche Wang, Zifeng Wang, Jinhua Tao, Huizheng Che, and Liangfu Chen. Performance of modis high-resolution maiac aerosol algorithm in china: Characterization and limitation. Atmospheric environment, 213:159–169, 2019. [12] Xin Su, Lunche Wang, Ming Zhang, Wenmin Qin, and Muhammad Bilal. A highprecision aerosol retrieval algorithm (hipara) for advanced himawari imager (ahi) data: Development and verification. Remote Sensing of Environment, 253:112221, 2021. [13] Dongming Qin, Jian Yu, Guojian Zou, Ruihan Yong, Qin Zhao, and Bo Zhang. A novel combined prediction scheme based on CNN and LSTM for urban PM2.5 concentration. IEEE Access, 7:20050–20059, 2019. [14] Yu Zheng, Xiuwen Yi, Ming Li, Ruiyuan Li, Zhangqing Shan, Eric Chang, and Tianrui Li. Forecasting fine-grained air quality based on big data. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 2267–2276, 2015. [15] Ming-Chuan Yang and Meng Chang Chen. PM2.5 Forecasting Using Pre-trained Components. In 2018 IEEE International Conference on Big Data (Big Data), pages 4488–4491. IEEE, 2018. [16] Xuhui Meng and George Em Karniadakis. A composite neural network that learns from multi-fidelity data: Application to function approximation and inverse PDE problems. Journal of Computational Physics, 401:109020, 2020. [17] Xiuwen Yi, Junbo Zhang, Zhaoyuan Wang, Tianrui Li, and Yu Zheng. Deep distributed fusion network for air quality prediction. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 965–973, 2018. [18] Ming-Chuan Yang and Meng Chang Chen. Composite neural network: Theory and application to pm2.5 prediction. IEEE Transactions on Knowledge and Data Engineering, 2021. [19] Tan Chen, Shulin Deng, and Manchun Li. Spatial patterns of satellite-retrieved PM2.5 and long-term exposure assessment of China from 1998 to 2016. International Journal of Environmental Research and Public Health, 15(12):2785, 2018. [20] Qian Di, Itai Kloog, Petros Koutrakis, Alexei Lyapustin, YujieWang, and Joel Schwartz. Assessing PM2.5 exposures with high spatiotemporal resolution across the continental United States. Environmental science & technology, 50(9):4712–4721, 2016. [21] Itai Kloog, Petros Koutrakis, Brent A Coull, Hyung Joo Lee, and Joel Schwartz. Assessing temporally and spatially resolved PM2.5 exposures for epidemiological studies using satellite aerosol optical depth measurements. Atmospheric environment, 45(35):6267–6275, 2011. [22] A Chudnovsky, A Lyapustin, Y Wang, C Tang, J Schwartz, and P Koutrakis. High resolution aerosol data from MODIS satellite for urban air quality studies. Open Geosciences, 6(1):17–26, 2014. [23] Ming-Tung Chuang, Chung-Te Lee, and Hui-Chun Hsu. Quantifying PM2.5 from long-range transport and local pollution in Taiwan during winter monsoon: An efficient estimation method. Journal of environmental management, 227:10–22, 2018. [24] Chuan-Yao Lin, Shaw Chen Liu, Charles CK Chou, Tsun Hsien Liu, Chung-Te Lee, Chung-Shin Yuan, Chein-Jung Shiu, Chea-Yuan Young, et al. Long-range transport of Asian dust and air pollutants to Taiwan. Terr. Atmos. Ocean. Sci, 15(5):759–784, 2004. [25] Ming-Tung Chuang, Joshua S Fu, Carey J Jang, Chang-Chuan Chan, Pei-Cheng Ni, and Chung-Te Lee. Simulation of long-range transport aerosols from the Asian Continent to Taiwan by a Southward Asian high-pressure system. Science of the total environment, 406(1-2):168–179, 2008. [26] Tu-Fu Chen, Ken-Hui Chang, and Chang-You Tsai. Modeling direct and indirect effect of long range transport on atmospheric PM2.5 levels. Atmospheric Environment, 89:1–9, 2014. [27] C-Y Lin, Z Wang, W-N Chen, S-Y Chang, CCK Chou, N Sugimoto, and X Zhao. Long-range transport of Asian dust and air pollutants to Taiwan: observed evidence and model simulation. 2007. [28] Xingjian Shi, Zhourong Chen, Hao Wang, Dit-Yan Yeung, Wai-Kin Wong, and Wangchun Woo. Convolutional lstm network: A machine learning approach for precipitation nowcasting. Advances in neural information processing systems, 28, 2015. [29] Binxu Zhai and Jianguo Chen. 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, 2018. [30] Qingyang Xiao, Howard H Chang, Guannan Geng, and Yang Liu. An ensemble machine-learning model to predict historical pm2.5 concentrations in china from satellite data. Environmental science & technology, 52(22):13260–13269, 2018. [31] Alexandra Shtein, Itai Kloog, Joel Schwartz, Camillo Silibello, Paola Michelozzi, Claudio Gariazzo, Giovanni Viegi, Francesco Forastiere, Arnon Karnieli, Allan C Just, et al. Estimating daily PM2.5 and PM10 over Italy using an ensemble model. Environmental Science & Technology, 54(1):120–128, 2019. [32] Bin Chen, Zhihao Song, Feng Pan, and Yue Huang. Obtaining vertical distribution of pm2.5 from caliop data and machine learning algorithms. Science of The Total Environment, 805:150338, 2022. [33] Zhihao Song, Bin Chen, Yue Huang, Li Dong, and Tingting Yang. Estimation of pm 2.5 concentration in china using linear hybrid machine learning model. Atmospheric Measurement Techniques, 14(8):5333–5347, 2021. [34] Pei-Yi Wong, Hsiao-Yun Lee, Yu-Cheng Chen, Yu-Ting Zeng, Yinq-Rong Chern, Nai-Tzu Chen, Shih-Chun Candice Lung, Huey-Jen Su, and Chih-Da Wu. Using a land use regression model with machine learning to estimate ground level pm2.5. Environmental Pollution, 277:116846, 2021. [35] Zhao-Yue Chen, Tian-Hao Zhang, Rong Zhang, Zhong-Min Zhu, Jun Yang, Ping-Yan Chen, Chun-Quan Ou, and Yuming Guo. Extreme gradient boosting model to estimate pm2.5 concentrations with missing-filled satellite data in china. Atmospheric Environment, 202:180–189, 2019. [36] Hossein Bagheri. A machine learning-based framework for high resolution mapping of pm2.5 in tehran, iran, using maiac aod data. Advances in Space Research, 69(9):3333–3349, 2022. [37] Padmavati Kulkarni, V Sreekanth, Adithi R Upadhya, and Hrishikesh Chandra Gautam. Which model to choose? performance comparison of statistical and machine learning models in predicting pm2.5 from high-resolution satellite aerosol optical depth. Atmospheric Environment, page 119164, 2022. [38] Gong Cheng, Ceyuan Yang, Xiwen Yao, Lei Guo, and Junwei Han. When deep learning meets metric learning: Remote sensing image scene classification via learning discriminative CNNs. IEEE transactions on geoscience and remote sensing, 56(5):2811–2821, 2018. [39] Wen-Shuai Hu, Heng-Chao Li, Lei Pan, Wei Li, Ran Tao, and Qian Du. Feature extraction and classification based on spatial-spectral convlstm neural network for hyperspectral images. arXiv preprint arXiv:1905.03577, 2019. [40] Casper Kaae Sønderby, Lasse Espeholt, Jonathan Heek, Mostafa Dehghani, Avital Oliver, Tim Salimans, Shreya Agrawal, Jason Hickey, and Nal Kalchbrenner. Metnet: A Neural Weather Model for Precipitation Forecasting. arXiv preprint arXiv:2003.12140, 2020. [41] SHI Xingjian, Zhourong Chen, Hao Wang, Dit-Yan Yeung, Wai-Kin Wong, and Wangchun Woo. Convolutional LSTM network: A machine learning approach for precipitation nowcasting. In Advances in neural information processing systems, pages 802–810, 2015. [42] Sepp Hochreiter and Jürgen Schmidhuber. Long short-term memory. Neural computation, 9(8):1735–1780, 1997. [43] Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167, 2015. [44] Nina Zumel, John Mount, and Jim Porzak. Practical data science with R. Manning Shelter Island, NY, 2014. [45] NASA EOSDIS. MODIS/Terra+Aqua Land Aerosol Optical Depth. https://ladsweb. modaps.eosdis.nasa.gov/, 2007. [Online; accessed 2-December-2017]. [46] Taiwan EPA. Taiwan Air Quality Monitoring Network. https://opendata.epa.gov.tw, 2002. [Online; accessed 7-December-2017]. [47] Taiwan CWB. Weather Forecast. http://opendata.cwb.gov.tw/index, 1971. [Online; accessed 9-December-2017]. [48] NCEP(GFS - FNL). GLobal Weather Forecast. https://rda.ucar.edu/datasets/, 2000. [Online; accessed 12-December-2018]. [49] Chuan-Yao Lin, Shaw C Liu, Charles C-K Chou, Saint-Jer Huang, Chung-Ming Liu, Ching-Huei Kuo, and Chea-Yuan Young. Long-range transport of aerosols and their impact on the air quality of taiwan. Atmospheric Environment, 39(33):6066–6076, 2005. [50] Zhipeng Luo, Jianqiang Huang, Ke Hu, Xue Li, and Peng Zhang. Accuair: Winning solution to air quality prediction for KDD Cup 2018. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 1842–1850, 2019. 描述 博士
國立政治大學
社群網路與人智計算國際研究生博士學位學程(TIGP)
104761506資料來源 http://thesis.lib.nccu.edu.tw/record/#G0104761506 資料類型 thesis dc.contributor.advisor 陳孟彰<br>劉昭麟 zh_TW dc.contributor.advisor Meng Chang Chen<br>Chao-Lin Liu en_US dc.contributor.author (Authors) 喬治·威廉·基比里奇 zh_TW dc.contributor.author (Authors) George william Kibirige en_US dc.creator (作者) 喬治·威廉·基比里奇 zh_TW dc.creator (作者) Kibirige, George william en_US dc.date (日期) 2023 en_US dc.date.accessioned 2-Jan-2024 15:22:16 (UTC+8) - dc.date.available 2-Jan-2024 15:22:16 (UTC+8) - dc.date.issued (上傳時間) 2-Jan-2024 15:22:16 (UTC+8) - dc.identifier (Other Identifiers) G0104761506 en_US dc.identifier.uri (URI) https://nccur.lib.nccu.edu.tw/handle/140.119/149030 - dc.description (描述) 博士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 社群網路與人智計算國際研究生博士學位學程(TIGP) zh_TW dc.description (描述) 104761506 zh_TW dc.description.abstract (摘要) 精確的PM2.5預測是對抗空氣污染的重要一環，有助於政府制定環境政策。透過多角度大氣校正（MAIAC）算法處理的衛星遙感氣膠光學深度（AOD）資料，我們能夠觀察遙遠地區之間污染物的傳輸情況。在本論文中，我們提出了一種結合神經網絡模型的方法，稱為遠程傳輸污染（RTP）模型，可更準確預測當地的PM2.5濃度。本論文所提出的RTP模型結合了多個深度學習元件，並從異質特徵中學習。此外，我們還提出的分類算法，來檢測從AOD數據中識別出的兩個參考站點的遠程傳輸污染事件（RTPEs）。我們在台灣的北部地區和南部中部地區進行了大量模擬，以評估其在實際數據上的表現。對於北部地區，提出的RTP模型在+4小時至+24小時、+28小時至+48小時和+52小時至+72小時的時間範圍內，相對不考慮RTPEs的基本模型，提高了17%至30%、23%至26%和18%至22%的準確度，同時也優於只考慮RTPEs的最先進模型，提高了12%至22%、12%至14%和10%至11%的準確度 其他影響PM2.5濃度的特徵，陸海風在台灣南部和中部地區扮演著關鍵角色。我們在提出的RTP模型中使用了大面積海洋風特徵，以捕捉陸海風對PM2.5的影響。我們對RTP模型的不同組成特徵個別進行研究，結果顯示不管再提出特徵跟模型改進，我們提出的架構在整體性能方面都有明顯改進。我們還進行了月份分析，證明了在台灣南部和中部地區，陸海風經常發生並主導PM2.5濃度。 zh_TW dc.description.abstract (摘要) Accurate PM2.5 prediction is part of the fight against air pollution that helps governments to manage environmental policy. Satellite Remote sensing aerosol optical depth (AOD) pro- cessed by The Multi-Angle Implementation of Atmospheric Correlation (MAIAC) algorithm allows us to observe the transportation of remote pollutants between regions. In this dissertation we propose a composite neural network model, the Remote Trans- ported Pollutants (RTP) model, for such long-range pollutant transportation that predicts more accurate local PM2.5 concentrations using such satellite data. The proposed RTP model integrates several deep learning components and learns from the heterogeneous features of various domains. We also detected remote transportation pollution events (RTPEs) at two reference sites from the AOD data using proposed classification algorithm. Extensive experi- ments using real-world data were conducted in two regions, northern region and southern and central region of Taiwan. For northern region the proposed RTP model outperforms the base model that does not account for RTPEs by 17%-30%, 23%-26% and 18%-22% and state-of-the-art models that account for RTPEs by 12%-22%, 12%-14%, and 10%-11% at +4h to +24h, +28h to +48 hours, and +52h to +72h hours respectively. Including other features that affect PM2.5, Land-sea breeze plays an important role in increasing the level of PM2.5 in southern and central region of Taiwan. We used extra feature of large interpolation ocean wind in our proposed RTP model to capture the impact of land-sea breeze. We investigated the model outputs of different components of the RTP and concluded that the proposed composite neural network architecture yields significant improvements in the overall performance compared to each component and the other state-of-the-art model. We performed monthly analysis which also demonstrated the superiority of the proposed architecture for stations where land-sea breezes frequently occur in the southern and central region of Taiwan in the months when land-sea breeze dominates the accumulation of PM2.5. en_US dc.description.tableofcontents CHAPTER 1 Introduction 1 CHAPTER 2 Literature Review 9 CHAPTER 3 Proposed Models 13 CHAPTER 4 Classification of Remote Pollutants 29 CHAPTER 5 Datasets and Data Preprocessing 35 CHAPTER 6 Evaluation and Experimental Results 43 CHAPTER 7 Conclusions and Future Works 71 References 75 Appendix Publications 81 zh_TW dc.format.extent 18037161 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0104761506 en_US dc.subject (關鍵詞) PM2.5 zh_TW dc.subject (關鍵詞) 空氣污染 zh_TW dc.subject (關鍵詞) 臺灣 zh_TW dc.subject (關鍵詞) PM2.5 en_US dc.subject (關鍵詞) Air Pollution en_US dc.subject (關鍵詞) Composite Neural Network en_US dc.subject (關鍵詞) Taiwan en_US dc.title (題名) 使用衛星數據監測台灣空氣汙染物遠程輸送，優化PM2.5預測 zh_TW dc.title (題名) Using Satellite Data on Remote Transportation of Air Pollutants for PM2.5 Prediction in Taiwan en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) [1] Jai Shanker Pandey, Rakesh Kumar, and Sukumar Devotta. Health risks of NO2, SPM and SO2 in Delhi (India). Atmospheric Environment, 39(36):6868–6874, 2005. [2] Mihye Lee, Itai Kloog, Alexandra Chudnovsky, Alexei Lyapustin, Yujie Wang, Steven Melly, Brent Coull, Petros Koutrakis, and Joel Schwartz. Spatiotemporal prediction of fine particulate matter using high-resolution satellite images in the Southeastern US 2003–2011. Journal of exposure science & environmental epidemiology, 26(4):377–384, 2016. [3] Ssu-Ting Liu, Chu-Yung Liao, Cheng-Yu Kuo, and Hsien-Wen Kuo. The effects of pm2.5 from asian dust storms on emergency room visits for cardiovascular and respiratory diseases. International journal of environmental research and public health, 14(4):428, 2017. [4] Zhihao Song, Bin Chen, and Jianping Huang. Combining himawari-8 aod and deep forest model to obtain city-level distribution of pm2.5 in china. Environmental Pollution, 297:118826, 2022. [5] Mark D Gibson, Soumita Kundu, and Mysore Satish. Dispersion model evaluation of pm2.5, nox and so2 from point and major line sources in nova scotia, canada using aermod gaussian plume air dispersion model. Atmospheric Pollution Research, 4(2):157–167, 2013. [6] Ziyue Chen, Danlu Chen, Chuanfeng Zhao, Mei-po Kwan, Jun Cai, Yan Zhuang, Bo Zhao, Xiaoyan Wang, Bin Chen, Jing Yang, et al. Influence of meteorological conditions on pm2.5 concentrations across china: A review of methodology and mechanism. Environment international, 139:105558, 2020. [7] Wei-Ting Hung, Cheng-Hsuan Sarah Lu, Sheng-Hsiang Wang, Sheng-Po Chen, Fujung Tsai, and Charles C-K Chou. Investigation of long-range transported pm2.5 events over northern taiwan during 2005–2015 winter seasons. Atmospheric Environment, 217:116920, 2019. [8] Qiangqiang Xu, Xiaoling Chen, Shangbo Yang, Linling Tang, and Jiadan Dong. Spatiotemporal relationship between himawari-8 hourly columnar aerosol optical depth (aod) and ground-level pm2.5 mass concentration in mainland china. Science of the Total Environment, 765:144241, 2021. [9] Jianping Guo, Feng Xia, Yong Zhang, Huan Liu, Jing Li, Mengyun Lou, Jing He, Yan Yan, Fu Wang, Min Min, et al. Impact of diurnal variability and meteorological factors on the pm2.5-aod relationship: Implications for pm2.5 remote sensing. Environmental Pollution, 221:94–104, 2017. [10] Alexei Lyapustin, Yujie Wang, Sergey Korkin, and Dong Huang. MODIS Collection 6 MAIAC algorithm. Atmospheric Measurement Techniques, 11(10), 2018. [11] Minghui Tao, Jun Wang, Rong Li, Lili Wang, Lunche Wang, Zifeng Wang, Jinhua Tao, Huizheng Che, and Liangfu Chen. Performance of modis high-resolution maiac aerosol algorithm in china: Characterization and limitation. Atmospheric environment, 213:159–169, 2019. [12] Xin Su, Lunche Wang, Ming Zhang, Wenmin Qin, and Muhammad Bilal. A highprecision aerosol retrieval algorithm (hipara) for advanced himawari imager (ahi) data: Development and verification. Remote Sensing of Environment, 253:112221, 2021. [13] Dongming Qin, Jian Yu, Guojian Zou, Ruihan Yong, Qin Zhao, and Bo Zhang. A novel combined prediction scheme based on CNN and LSTM for urban PM2.5 concentration. IEEE Access, 7:20050–20059, 2019. [14] Yu Zheng, Xiuwen Yi, Ming Li, Ruiyuan Li, Zhangqing Shan, Eric Chang, and Tianrui Li. Forecasting fine-grained air quality based on big data. In Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 2267–2276, 2015. [15] Ming-Chuan Yang and Meng Chang Chen. PM2.5 Forecasting Using Pre-trained Components. In 2018 IEEE International Conference on Big Data (Big Data), pages 4488–4491. IEEE, 2018. [16] Xuhui Meng and George Em Karniadakis. A composite neural network that learns from multi-fidelity data: Application to function approximation and inverse PDE problems. Journal of Computational Physics, 401:109020, 2020. [17] Xiuwen Yi, Junbo Zhang, Zhaoyuan Wang, Tianrui Li, and Yu Zheng. Deep distributed fusion network for air quality prediction. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 965–973, 2018. [18] Ming-Chuan Yang and Meng Chang Chen. Composite neural network: Theory and application to pm2.5 prediction. IEEE Transactions on Knowledge and Data Engineering, 2021. [19] Tan Chen, Shulin Deng, and Manchun Li. Spatial patterns of satellite-retrieved PM2.5 and long-term exposure assessment of China from 1998 to 2016. International Journal of Environmental Research and Public Health, 15(12):2785, 2018. [20] Qian Di, Itai Kloog, Petros Koutrakis, Alexei Lyapustin, YujieWang, and Joel Schwartz. Assessing PM2.5 exposures with high spatiotemporal resolution across the continental United States. Environmental science & technology, 50(9):4712–4721, 2016. [21] Itai Kloog, Petros Koutrakis, Brent A Coull, Hyung Joo Lee, and Joel Schwartz. Assessing temporally and spatially resolved PM2.5 exposures for epidemiological studies using satellite aerosol optical depth measurements. Atmospheric environment, 45(35):6267–6275, 2011. [22] A Chudnovsky, A Lyapustin, Y Wang, C Tang, J Schwartz, and P Koutrakis. High resolution aerosol data from MODIS satellite for urban air quality studies. Open Geosciences, 6(1):17–26, 2014. [23] Ming-Tung Chuang, Chung-Te Lee, and Hui-Chun Hsu. Quantifying PM2.5 from long-range transport and local pollution in Taiwan during winter monsoon: An efficient estimation method. Journal of environmental management, 227:10–22, 2018. [24] Chuan-Yao Lin, Shaw Chen Liu, Charles CK Chou, Tsun Hsien Liu, Chung-Te Lee, Chung-Shin Yuan, Chein-Jung Shiu, Chea-Yuan Young, et al. Long-range transport of Asian dust and air pollutants to Taiwan. Terr. Atmos. Ocean. Sci, 15(5):759–784, 2004. [25] Ming-Tung Chuang, Joshua S Fu, Carey J Jang, Chang-Chuan Chan, Pei-Cheng Ni, and Chung-Te Lee. Simulation of long-range transport aerosols from the Asian Continent to Taiwan by a Southward Asian high-pressure system. Science of the total environment, 406(1-2):168–179, 2008. [26] Tu-Fu Chen, Ken-Hui Chang, and Chang-You Tsai. Modeling direct and indirect effect of long range transport on atmospheric PM2.5 levels. Atmospheric Environment, 89:1–9, 2014. [27] C-Y Lin, Z Wang, W-N Chen, S-Y Chang, CCK Chou, N Sugimoto, and X Zhao. Long-range transport of Asian dust and air pollutants to Taiwan: observed evidence and model simulation. 2007. [28] Xingjian Shi, Zhourong Chen, Hao Wang, Dit-Yan Yeung, Wai-Kin Wong, and Wangchun Woo. Convolutional lstm network: A machine learning approach for precipitation nowcasting. Advances in neural information processing systems, 28, 2015. [29] Binxu Zhai and Jianguo Chen. 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, 2018. [30] Qingyang Xiao, Howard H Chang, Guannan Geng, and Yang Liu. An ensemble machine-learning model to predict historical pm2.5 concentrations in china from satellite data. Environmental science & technology, 52(22):13260–13269, 2018. [31] Alexandra Shtein, Itai Kloog, Joel Schwartz, Camillo Silibello, Paola Michelozzi, Claudio Gariazzo, Giovanni Viegi, Francesco Forastiere, Arnon Karnieli, Allan C Just, et al. Estimating daily PM2.5 and PM10 over Italy using an ensemble model. Environmental Science & Technology, 54(1):120–128, 2019. [32] Bin Chen, Zhihao Song, Feng Pan, and Yue Huang. Obtaining vertical distribution of pm2.5 from caliop data and machine learning algorithms. Science of The Total Environment, 805:150338, 2022. [33] Zhihao Song, Bin Chen, Yue Huang, Li Dong, and Tingting Yang. Estimation of pm 2.5 concentration in china using linear hybrid machine learning model. Atmospheric Measurement Techniques, 14(8):5333–5347, 2021. [34] Pei-Yi Wong, Hsiao-Yun Lee, Yu-Cheng Chen, Yu-Ting Zeng, Yinq-Rong Chern, Nai-Tzu Chen, Shih-Chun Candice Lung, Huey-Jen Su, and Chih-Da Wu. Using a land use regression model with machine learning to estimate ground level pm2.5. Environmental Pollution, 277:116846, 2021. [35] Zhao-Yue Chen, Tian-Hao Zhang, Rong Zhang, Zhong-Min Zhu, Jun Yang, Ping-Yan Chen, Chun-Quan Ou, and Yuming Guo. Extreme gradient boosting model to estimate pm2.5 concentrations with missing-filled satellite data in china. Atmospheric Environment, 202:180–189, 2019. [36] Hossein Bagheri. A machine learning-based framework for high resolution mapping of pm2.5 in tehran, iran, using maiac aod data. Advances in Space Research, 69(9):3333–3349, 2022. [37] Padmavati Kulkarni, V Sreekanth, Adithi R Upadhya, and Hrishikesh Chandra Gautam. Which model to choose? performance comparison of statistical and machine learning models in predicting pm2.5 from high-resolution satellite aerosol optical depth. Atmospheric Environment, page 119164, 2022. [38] Gong Cheng, Ceyuan Yang, Xiwen Yao, Lei Guo, and Junwei Han. When deep learning meets metric learning: Remote sensing image scene classification via learning discriminative CNNs. IEEE transactions on geoscience and remote sensing, 56(5):2811–2821, 2018. [39] Wen-Shuai Hu, Heng-Chao Li, Lei Pan, Wei Li, Ran Tao, and Qian Du. Feature extraction and classification based on spatial-spectral convlstm neural network for hyperspectral images. arXiv preprint arXiv:1905.03577, 2019. [40] Casper Kaae Sønderby, Lasse Espeholt, Jonathan Heek, Mostafa Dehghani, Avital Oliver, Tim Salimans, Shreya Agrawal, Jason Hickey, and Nal Kalchbrenner. Metnet: A Neural Weather Model for Precipitation Forecasting. arXiv preprint arXiv:2003.12140, 2020. [41] SHI Xingjian, Zhourong Chen, Hao Wang, Dit-Yan Yeung, Wai-Kin Wong, and Wangchun Woo. Convolutional LSTM network: A machine learning approach for precipitation nowcasting. In Advances in neural information processing systems, pages 802–810, 2015. [42] Sepp Hochreiter and Jürgen Schmidhuber. Long short-term memory. Neural computation, 9(8):1735–1780, 1997. [43] Sergey Ioffe and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167, 2015. [44] Nina Zumel, John Mount, and Jim Porzak. Practical data science with R. Manning Shelter Island, NY, 2014. [45] NASA EOSDIS. MODIS/Terra+Aqua Land Aerosol Optical Depth. https://ladsweb. modaps.eosdis.nasa.gov/, 2007. [Online; accessed 2-December-2017]. [46] Taiwan EPA. Taiwan Air Quality Monitoring Network. https://opendata.epa.gov.tw, 2002. [Online; accessed 7-December-2017]. [47] Taiwan CWB. Weather Forecast. http://opendata.cwb.gov.tw/index, 1971. [Online; accessed 9-December-2017]. [48] NCEP(GFS - FNL). GLobal Weather Forecast. https://rda.ucar.edu/datasets/, 2000. [Online; accessed 12-December-2018]. [49] Chuan-Yao Lin, Shaw C Liu, Charles C-K Chou, Saint-Jer Huang, Chung-Ming Liu, Ching-Huei Kuo, and Chea-Yuan Young. Long-range transport of aerosols and their impact on the air quality of taiwan. Atmospheric Environment, 39(33):6066–6076, 2005. [50] Zhipeng Luo, Jianqiang Huang, Ke Hu, Xue Li, and Peng Zhang. Accuair: Winning solution to air quality prediction for KDD Cup 2018. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, pages 1842–1850, 2019. zh_TW
