Spatial Distribution Characteristics of PM2.5 in Urban Block Based on Remote Sensing Technology and Machine Learning: Taking Binhu New District of Hefei City as an Example

doi:10.16258/j.cnki.1674-5906.2024.09.010

Ecology and Environment ›› 2024, Vol. 33 ›› Issue (9): 1426-1437.DOI: 10.16258/j.cnki.1674-5906.2024.09.010

• Research Article [Environmental Science] • Previous Articles Next Articles

Spatial Distribution Characteristics of PM_2.5 in Urban Block Based on Remote Sensing Technology and Machine Learning: Taking Binhu New District of Hefei City as an Example

WANG Wei¹^,²^,^*(), XIA Yuxuan³

1. College of Architecture and Art, Hefei University of Technology, Hefei 230601, P. R. China
2. Anhui Provincial Key Laboratory of Digitalized Conservation and Innovative Revitalization of Ancient Huizhou Villages, Hefei 230601, P. R. China
3. Hefei City University, Hefei 238076, P. R. China

Received:2023-12-11 Online:2024-09-18 Published:2024-10-18
Contact: WANG Wei

基于遥感技术和机器学习的城市街区PM_2.5空间分布特征研究——以合肥市滨湖新区为例

王薇¹^,²^,^*(), 夏宇轩³

1.合肥工业大学建筑与艺术学院，安徽合肥 230601
2.徽州古村落数字化保护与传承创意安徽省重点实验室，安徽合肥 230601
3.合肥城市学院，安徽合肥 238076

通讯作者: 王薇
作者简介:王薇（1975年生），女，教授，博士，博士研究生导师，研究方向为建筑技术和人居环境。E-mail: vivi.gan@126.com
基金资助:
国家自然科学基金项目(52478016);安徽省自然科学基金项目(2308085ME182);徽州古村落数字化保护与传承创意安徽省重点实验室“自主创新专项”(PA2023GDSK0116);安徽省重点研究与开发计划项目(2023g07020003)

Abstract

Abstract:

In order to obtain the ground-level PM_2.5 concentration in Binhu New District of Hefei City, the ground wind speed was used to correct the wind speed of the MCD19A2 data of the MODIS sensor from July 31, 2022, to July 31, 2023, and the AOD-PM_2.5 inversion model was established with the same period of meteorological data and the PM_2.5 data from 10 state-controlled monitoring stations in Hefei City. The test results were most satisfactory when the BP network was set up to the 10th generation, and the Mean Square Error (MSE) value was 0.027. The regression values of R for the training, validation, and test sets were greater than 0.88, and the overall result was as high as 0.93. The model was built using the BP prediction toolbox and the goodness-of-fit was 0.807. Seven typical measuring points were selected in the Binhu New District to measure the daily average PM_2.5, ground wind speed, temperature, and humidity from July 12, 2023, to July 16, 2023. After combining the meteorological data into the inversion model, it was found that the estimated and measured value curves were highly consistent, which proves that the model can be used to reverse the ground PM_2.5 concentration in Binhu New District. The results of the estimation and analysis of the PM_2.5 concentration showed that: 1) the highest PM_2.5 concentrations were found on the southwest side of the Binhu New District and the lowest on the southeast side. 2) In terms of land use, PM_2.5 concentration was presented as industrial area>commercial area>residential area > ecological park. 3) Waste emissions in industrial zones led to an increase in PM_2.5. Optimizing the use of factory equipment and proper treatment of waste in the production process can reduce PM_2.5. 4) Tailpipe emissions, dust, and industrial exhaust in commercial and residential areas led to higher PM_2.5. Optimizing traffic organization and promoting the development of energy vehicles can reduce PM_2.5. 5) The development mode of industrial parks has a significant impact on the concentration of PM_2.5; accelerating the adjustment of industrial structure and promoting the innovation of industrial green technology can reduce the concentration of PM_2.5. 6) Urban green spaces had a significant regulatory effect on PM_2.5. Increasing the connectivity and aggregation of green spaces and strengthening the construction of large-scale green spaces can play a stronger role in PM_2.5 reduction. The AOD-PM_2.5 inversion model provides a reliable method for the study of the spatial distribution characteristics of PM_2.5 in urban neighborhoods, which has important practical significance.

Key words: remote sensing technology, PM_2.5, aerosol, urban block, revised wind speed, spatial distribution, machine learning

摘要：

为获取合肥市滨湖新区地面PM_2.5浓度，利用地面风速对2022年7月31日－2023年7月31日MODIS传感器MCD19A2数据进行风速订正，并与同期气象数据和合肥市10个国控监测站点的PM_2.5数据建立AOD-PM_2.5反演模型。结果表明：BP网络训练到第10代时训练结果最为理想，均方误差MSE值约为0.027。训练集、验证集、测试集和总体结果回归值r均在0.88以上，总体结果高达0.93。通过BP预测工具箱建立模型，拟合优度达到0.807。选择滨湖新区7个典型测点，测量2023年7月12－16日每日平均PM_2.5浓度、地面风速和温湿度，结合气象数据代入反演模型对比发现，估算值和实测值曲线高度吻合，证明该模型可用于反演滨湖新区地面PM_2.5浓度。对PM_2.5浓度进行估算和分析，结果表明：1）滨湖新区西南侧PM_2.5浓度最高，东南侧最低；2）功能区PM_2.5浓度呈现为工业区>商业区>居住区>生态园区；3）工业园区废弃物的排放会导致PM_2.5浓度升高，对工厂设备的优化使用以及合理处理生产过程中的废弃物可降低PM_2.5浓度；4）商业区和居住区中的尾气排放、扬尘、工业排风等均会导致PM_2.5浓度升高，优化交通组织，推动新能源汽车的发展可降低PM_2.5浓度；5）产业园区的发展模式对城市PM_2.5浓度有显著影响，加快产业结构调整，推动产业绿色技术创新可降低PM_2.5浓度；6）城市绿地空间对PM_2.5有显著调节作用，增加绿地空间连通度、聚集度，加强大型绿地建设可发挥较强的PM_2.5消减作用。该研究所建立的AOD-PM_2.5反演模型旨在为城市街区PM_2.5空间分布特征研究提供可靠方法，具有重要的实践意义。

关键词: 遥感技术, PM_2.5, 气溶胶, 城市街区, 风速订正, 空间分布, 机器学习

CLC Number:

WANG Wei, XIA Yuxuan. Spatial Distribution Characteristics of PM_2.5 in Urban Block Based on Remote Sensing Technology and Machine Learning: Taking Binhu New District of Hefei City as an Example[J]. Ecology and Environment, 2024, 33(9): 1426-1437.

王薇, 夏宇轩. 基于遥感技术和机器学习的城市街区PM_2.5空间分布特征研究——以合肥市滨湖新区为例[J]. 生态环境学报, 2024, 33(9): 1426-1437.

Figures/Tables 14

Figure 1 Technology roadmap for AOD inversion of PM2.5 concentration

Figure 2 MATLAB modeling flow chart

Figure 3 Distribution of state-controlled monitoring stations in Hefei City

Figure 4 Step-by-step diagram of MCD19A2 data preprocessing

Figure 5 TIF image of the study area

Table 1 Table of collinearity test results for independent variables

变量		未标准化系数		标准化系数	检验值	系统因素	共线性统计
模型		B	标准错误	Beta	t	显著性	容差	VIF
1	(常量)	31.4	4.35		7.22	<0.001
	AOD	58.0	4.58	0.481	12.6	<0.001	0.770	1.30
	温度	−1.00	0.080	−0.467	−12.4	<0.001	0.792	1.26
	湿度	0.097	0.080	0.046	1.20	0.229	0.767	1.30
	经向风速	−0.184	0.432	−0.014	−0.427	0.670	0.970	1.03
	纬向风速	2.23	0.412	0.201	5.42	<0.001	0.812	1.23

Figure 6 Plot of model accuracy validation results

Figure 7 BP prediction result chart

Figure 8 Distribution of measuring points in Binhu New District, Hefei City

Figure 9 Estimated versus measured values graph

Figure 10 TIF image of the estimation area

Figure 11 Directions to the estimated points of the Binhu New District

Figure 12 Distribution of PM2.5 Concentration at the Estimation Points in Binhu New District

Figure 13 Distribution area map of PM2.5 concentration at estimated points in Binhu New District

References 39

[1]	GUPTA P, CHRISTOPHER S A, WANG J, et al., 2006. Satellite remote sensing of particulate matter and air quality assessment over global cities[J]. Atmospheric Environment, 40(30): 5880-5892.
[2]	KUMAR N, CHU A D, FOSTER A D, et al., 2011. Satellite remote sensing for developing time and space resolved estimates of ambient particulate in Cleveland, OH[J]. Aerosol Science and Technology, 45(9): 1090-1108. DOI PMID
[3]	LIU J J, WENG F Z, LI Z Q, 2019. Satellite-based PM_2.5 estimation directly from reflectance at the top of the atmosphere using a machine learning algorithm[J]. Atmospheric Environment, 208: 113-122.
[4]	TIAN J, CHEN D M, 2010. Spatial and temporal sensitivity of correlating MODIS aerosol optical depth with ground-based fine particulate matter (PM_2.5) across southern Ontario[J]. Canadian Journal of Remote Sensing, 36(2): 119-128.
[5]	WANG J, CHRISTOPHER S A, 2003. Intercomparison between satellite-derived aerosol optical thickness and PM_2.5 mass: Implications for air quality studies[J]. Geophysical Research Letters, 30(21): 2095.
[6]	YANG Y, LUO L W, SONG C, et al., 2018. Spatiotemporal assessment of PM_2.5 related economic losses from health impacts during 2014-2016 in China[J]. International Journal of Environmental Research and Public Health, 15(6): 1278.
[7]	安徽省生态环境厅, 2023. 合肥市国控点逐小时PM_2.5数据源[DB/OL]. (2023-07-31) [2024-05-08]. https://sthjt.ah.gov.cn.
	Department of Ecology and Environment of Anhui Province, 2023. Hourly PM_2.5 data source of Hefei state control point[DB/OL]. (2023-07-31) [2024-05-08]. https://sthjt.ah.gov.cn.
[8]	曹惠明, 孟祥亮, 孔梅, 等, 2022. 山东省大气气溶胶时空分布特征遥感研究[J]. 环境污染与防治, 44(8): 1046-1053.
	CAO H M, MENG X L, KONG M, et al., 2022. Remote sensing research on spatiotemporal distribution characteristics of atmospheric aerosol in Shandong Province[J]. Environmental Pollution & Control, 44(8): 1046-1053.
[9]	陈辉, 厉青, 王中挺, 等, 2014. 利用MODIS资料监测京津冀地区近地面PM_2.5方法研究[J]. 气象与环境学报, 30(5): 27-37.
	CHEN H, LI Q, WANG Z T, et al., 2014. Study on monitoring surface PM_2.5concentration in Jing-Jin-Ji regions using MODIS data[J]. Journal of Meteorology and Environment, 30(5): 27-37.
[10]	地球数据, 2023. 高精度数据源[DB/OL]. (2023-07-31) [2024-05-08]. https://search.earthdata.nasa.gov.
	EARTHDATA, 2023. High precision data source[DB/OL]. (2023-07-31) [2024-05-08]. https://search.earthdata.nasa.gov.
[11]	郭新彪, 魏红英, 2013. 大气PM_2.5对健康影响的研究进展[J]. 科学通报, 58(13): 1171-1177.
	GUO X B, WEI H Y, 2013. Progress on the health effects of ambient PM_2.5 pollution[J]. Chinese Science Bulletin, 58(13): 1171-1177.
[12]	韩桂兰, 李雪姣, 2022. 青海海东工业园区PM_2.5浓度预测及分析[J]. 佳木斯大学学报(自然科学版), 40(3): 118-121.
	HAN G L, LI X J, 2022. Predicted and analysis of PM_2.5 concentration in Qinghai Haidong industrial park[J]. Journal of Jiamusi University (Natural Science Edition), 40(3): 118-121.
[13]	韩阳, 康凌, 宋宇, 2021. 2000-2019年东北三省气溶胶光学厚度的时空分布特征[J]. 北京大学学报(自然科学版), 57(6): 1027-1034.
	HAN Y, KANG L, SONG Y, 2021. Spatial-temporal distribution of aerosol optical depth over northeastern China during 2000-2019[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 57(6): 1027-1034.
[14]	黄慧芬, 2020. 合肥市城市空间形态演变多尺度研究 (1949-2015年)[D]. 合肥: 合肥工业大学: 1-110.
	HUANG H F, 2020. Evolution of urban spatial form in Hefei Multi-scale Research (1949-2015)[D]. Hefei: Hefei University of Technology: 1-110.
[15]	金囝囡, 杨兴川, 晏星, 等, 2021. 京津冀及周边MAIAC AOD和PM_2.5质量浓度特征及相关性分析[J]. 环境科学, 42(6): 2604-2615.
	JIN N N, YANG X C, YAN X, et al., 2021. MAIAC AOD and PM_2.5 mass concentrations characteristics and correlation analysis in Beijing-Tianjin-Hebei and surrounding areas[J]. Environmental Science, 42(6): 2604-2615.
[16]	刘喆, 赵威伦, 田晓青, 等, 2022. 利用葵花8号卫星资料反演中国东部地区地面PM_2.5浓度[J]. 北京大学学报(自然科学版), 58(3): 443-452.
	LIU Z, ZHAO W L, TIAN X Q, et al., 2022. Retrieval of ground PM_2.5 concentrations in eastern China using data from Himawari-8 satellite[J]. Acta Scientiarum Naturalium Universitatis Pekinensis, 58(3): 443-452.
[17]	林少凯, 王恺, 詹小海, 等, 2022. 福州市工业区和商业居住区大气PM_2.5中多环芳烃特征的变化及其健康风险[J]. 环境与职业医学, 39(11): 1277-1283.
	LIN S K, WANG K, ZHAN X H, et al., 2022. Variations and health risk of polycyclic aromatic hydrocarbons in ambient PM_2.5 in industrialand residential areas in Fuzhou city[J]. Journal of Environmental and Occupational Medicine, 39(11): 1277-1283.
[18]	李涛, 薛领, 赵威, 2023. 产业结构调整对城市PM_2.5的影响机理[J]. 经济地理, 43(5): 180-192. DOI
	LI T, XUE L, ZHAO W, 2023. Influencing mechanism of industrial structure adjustment on urban PM_2.5[J]. Economic Geography, 43(5): 180-192.
[19]	李同文, 孙越乔, 杨晨雪, 等, 2015. 融合卫星遥感与地面测站的区域PM_2.5反演[J]. 测绘地理信息, 40(3): 6-9.
	LI T W, SUN Y Q, YANG C X, et al., 2015. Retrieving PM_2.5 using satellite remote sensing and ground station[J]. Journal of Geomatics, 40(3): 6-9.
[20]	李小东, 张玉鹏, 李德, 2018. 合肥瑶海区旧城更新改造规划路径探索[J]. 规划师, 34(S1): 44-49.
	LI X D, ZHANG Y P, LI D, 2018. Old city renewal planning in Yaohai district, Hefei[J]. Planners, 34(S1): 44-49.
[21]	王飞, 吴开亚, 王慧慧, 等, 2014. 合肥市PM_2.5与用地类型特征比较及建设性方案研究[J]. 环境科学与管理, 39(10): 73-79.
	WANG F, WU K Y, WANG H H, et al., 2014. Correlations between PM_2.5 with land use types in Hefei and constructive solutions[J]. Environmental Science and Management, 39(10): 73-79.
[22]	王薇, 程歆玥, 胡春, 等, 2021. 城市街道峡谷PM_2.5时空分布特征与空气质量评价——以合肥市长淮街道为例[J]. 生态环境学报, 30(11): 2157-2164. DOI
	WANG W, CHENG X Y, HU C, et al., 2021. Spatio-temporal distribution characteristics of PM_2.5 and air quality evaluation in urban street canyons:Take Changhuai street in Hefei as an example[J]. Ecology and Environmental Sciences, 30(11): 2157-2164.
[23]	王勇, 任栋, 刘严萍, 等, 2020. 基于小波变换与回归分析的融合GNSS水汽、风速和PM₁₀要素的PM_2.5浓度模型[J]. 系统工程理论与实践, 40(3): 761-770. DOI
	WANG Y, REN D, LIU Y P, et al., 2020. PM_2.5 concentration model of GNSS precipitable water vapor, wind speed and PM₁₀ based on wavelet transform and regression analysis[J]. Systems Engineering-Theory & Practice, 40(3): 761-770.
[24]	王跃思, 张军科, 王莉莉, 等, 2014. 京津冀区域大气霾污染研究意义、现状及展望[J]. 地球科学进展, 29(3): 388-396. DOI
	WANG Y S, ZHANG J K, WANG L L, et al., 2014. Researching significance, status and expectation of haze in Beijing-Tianjin-Hebei region[J]. Advances in Earth Science, 29(3): 388-396. DOI
[25]	王珏, 徐海韵, 岳峰, 等, 2023. 北京市中心城区绿地空间布局与PM_2.5的关联性研究[J]. 北京建筑大学学报, 39(1): 44-52.
	WANG Y, XU H Y, YUE F, et al., 2023. Research on the correlation between green space layout and PM_2.5 in the central urban area of Beijing[J]. Journal of Beijing University of Civil Engineering and Architecture, 39(1): 44-52.
[26]	王子峰, 曾巧林, 陈良富, 等, 2019. 利用卫星遥感数据估算PM_2.5浓度的应用研究进展[J]. 环境监控与预警, 11(5): 33-38.
	WANG Z F, ZENG Q L, CHEN L F, et al., 2019. Research progress of methodology and applications of PM_2.5 estimation using satellite remote sensing[J]. Environmental Monitoring and Forewarning, 11(5): 33-38.
[27]	韦小丽, 2022. 基于多源数据协同融合的高分辨率AOD反演与PM_2.5浓度估算研究[D]. 上海: 华东师范大学: 1-137.
	WEI X L, 2022. Leveraging multi-sensor data fusion to generate high resolution AOD and PM_2.5 concentration maps[D]. Shanghai: East China Normal University: 1-137.
[28]	翁天信, 2019. 基于神经网络的机器学习方法[J]. 通讯世界, 26(2): 61-62.
	WENG T X, 2019. Neural network based machine learning methods[J]. Telecom World, 26(2): 61-62.
[29]	吴迪, 高枞亭, 李建平, 等, 2023. 东北地区PM_2.5质量浓度遥感估算与时空分布特征[J]. 地理科学, 43(10): 1869-1878. DOI
	WU D, GAO Z T, LI J P, et al., 2023. Remote sensing estimation and spatial-temporal distribution of PM_2.5 concentration in Northeast China[J]. Scientia Geographica Sinica, 43(10): 1869-1878.
[30]	吴健生, 王茜, 2017. 基于AOD数据反演地面PM_2.5浓度研究进展[J]. 环境科学与技术, 40(8): 68-76.
	WU J S, WANG X, 2017. Research progress of retrieval ground-level PM_2.5 concentration based on AOD data[J]. Environmental Science & Technology, 40(8): 68-76.
[31]	吴奕霄, 银燕, 顾雪松, 等, 2014. 南京北郊大气气溶胶的吸湿性观测研究[J]. 中国环境科学, 34(8): 1938-1949.
	WU Y X, YIN Y, GU X S, et al., 2014. An observational study of the hygroscopic properties of aerosols in north suburb of Nanjing[J]. China Environmental Science, 34(8): 1938-1949.
[32]	向娟, 陶明辉, 郭玲, 等, 2022. 基于卫星遥感的近地面PM_2.5浓度反演进展[J]. 遥感学报, 26(9): 1757-1776.
	XIANG J, TAO M H, GUO L, et al., 2022. Progress of near-surface PM_2.5 concentration retrieve based on satellite remote sensing[J]. National Remote Sensing Bulletin, 26(9): 1757-1776.
[33]	羲和能源大数据平台, 2023. 覆盖中国全境的超高精度数据源[DB/OL]. (2023-07-31) [2024-05-08]. https://www.xihe-energy.com.
	Xihe-energy, 2023. Ultra-high precision data source covering all of China[DB/OL]. (2023-07-31) [2024-05-08]. https://www.xihe-energy.com.
[34]	杨晓辉, 肖登攀, 王卫, 等, 2022. 基于遥感数据估算近地面PM_2.5浓度的研究进展[J]. 环境科学研究, 35(1): 40-50.
	YANG X H, XIAO D P, WANG W, et al., 2022. Research progress of ground-level PM_2.5 concentration estimation based on remote sensing data[J]. Research of Environmental Sciences, 35(1): 40-50.
[35]	杨颖川, 葛宝珠, 郝赛宇, 等, 2020. 基于能见度及AOD数据的北京市PM_2.5浓度的反演[J]. 气候与环境研究, 25(5): 521-530.
	YANG Y C, GE B Z, HAO S Y, et al., 2020. Lnversion of PM_2.5 concentration in Beijing based on visibility and AOD data[J]. Climatic and Environmental Research, 25(5): 521-530.
[36]	袁野, 2019. Matlab可视化与机器学习课程的案例教学实践[J]. 福建电脑, 35(7): 116-118.
	YUAN Y, 2019. Case teaching practice of matlab visualization and machine learning course[J]. Journal of Fujian Computer, 35(7): 116-118.
[37]	于志翔, 2023. 基于MODIS数据的新疆大气气溶胶光学厚度时空变化分析[J]. 沙漠与绿洲气象, 17(2): 128-136.
	YU Z X, 2023. Spatial and temporal variation of aerosol optical depth in Xinjiang derived from MODIS observations[J]. Desert and Oasis Meteorology, 17(2): 128-136.
[38]	张猛, 张博, 2020. 基于机器学习和卫星遥感的PM_2.5/10空间连续分布反演方法研究[J]. 地球环境学报, 11(4): 447-455.
	ZHANG M, ZHANG B, 2020. Research on the methods to retrieve continuous spatial distribution of PM_2.5/10 based on machine learning and satellite imagery[J]. Journal of Earth Environment, 11(4): 447-455.
[39]	赵青琳, 曾巧林, 罗彬, 等, 2022. 基于垂直订正和湿度订正方法估算川渝地区逐小时PM_2.5[J]. 遥感学报, 26(10): 1946-1962.
	ZHAO Q L, ZENG Q L, LUO B, et al., 2022. Estimation of hourly PM_2.5 concentration by using vertical and humidity correction methods in Sichuan-Chongqing[J]. National Remote Sensing Bulletin, 26(10): 1946-1962.

Spatial Distribution Characteristics of PM_2.5 in Urban Block Based on Remote Sensing Technology and Machine Learning: Taking Binhu New District of Hefei City as an Example

基于遥感技术和机器学习的城市街区PM_2.5空间分布特征研究——以合肥市滨湖新区为例

RichHTML

PDF