基於 PM2.5微型感測資料之中尺度空品區劃分研究—以臺灣五都為例

Determining Mesoscale Air Quality Management Areas using Micro PM2.5 Measurements in Taiwan

李雅君(Lee Ya-Chun)

指導教授 ： 陳伶志

國立臺灣師範大學/理學院/資訊工程學系/碩士(2018年)

https://doi.org/10.6345/THE.NTNU.DCSIE.024.2018.B02

摘要

空氣污染會提高罹患呼吸道疾病及死亡之風險，是目前全世界都關注的環境議題，其中PM2.5對人體的健康有極大的威脅而被重視，由於PM2.5的來源辨別困難，進而難以做到防治。然而在PM2.5的資料中能夠觀察到，感測器若有相似波動的數值，其背後的影響因素可能相同，因此本論文提出PM2.5的中尺度劃分研究，清楚的劃分空氣品質區能夠瞭解當地空氣品質特徵，也能幫助污染源判斷、PM2.5數值預測等其他研究。本研究使用臺灣PM2.5的開放資料，以哈爾小波轉換之分群策略，並利用動態時間變形技術提高分群準確度，不需要加入其他額外的地理環境資訊，也能夠清楚劃分中尺度空品區，分群結果符合當地環境且能做出合理解釋，能夠與該地區的生活型態、地形氣候有高相關性。

關鍵字

懸浮微粒 ； 微型感測 ； 分群分析 ； 時間序列 ； 哈爾小波轉換

並列摘要

Air pollution will increase the risk of respiratory diseases and death. It is an environmental issue that is of concern to the whole world. PM2.5 is a great threat to human health and is therefore been extensively studied. Due to the difficulty of identifying the source of PM2.5, effective prevention measures are hard to realized. However, it can be observed in the data of PM2.5 that if the sensor has similar fluctuation values, the influencing factors behind it may be the same. Therefore, this paper proposes a Mesoscale Air Quality Management Areas of PM2.5, in which local characteristics of air quality is studied in each area. Understanding local air quality characteristics can also help with pollution source identification, PM2.5 numerical prediction and other research. This study uses the open data of PM2.5 in Taiwan, the clustering technique of Haar wavelet transform and dynamic time warping techniques to improve the grouping accuracy. Without adding other additional geographical environment information, it can successfully partition the mesoscale air quality management area. The grouping results are in line with the local environment and can be reasonably explained, which can be highly correlated with the life style and topographic climate of the area.

並列關鍵字

PM2.5 ； micro measurements ； clustering analysis ； time series ； Haar wavelet transform

參考文獻

[1] 行政院環境保護署 (2017)，中華民國空氣品質監測報告105年年報。

[2] 行政院環境保護署(2018)，空氣品質指標，取自https://taqm.epa.gov.tw/taqm/tw/b0201.aspx

[3] He, Z., Kim, Y. J., Ogunjobi, K. O., & Hong, C. S. (2003). Characteristics of PM2. 5 species and long-range transport of air masses at Taean background station, South Korea. Atmospheric Environment, 37(2), 219-230.

[4] Perrone, M. R., Becagli, S., Orza, J. G., Vecchi, R., Dinoi, A., Udisti, R., & Cabello, M. (2013). The impact of long-range-transport on PM1 and PM2. 5 at a Central Mediterranean site. Atmospheric environment, 71, 176-186.

[5] Huang, D., Zhang, N., Yu, H., Zhou, H., Ma, Z., Hu, W., & Guo, J. (2015, August). Activation force-based air pollution observation station clustering. In Heterogeneous Networking for Quality, Reliability, Security and Robustness (QSHINE), 2015 11th International Conference on (pp. 115-121). IEEE.