基于地理加权回归模型的新疆地区PM_(2.5)遥感估算
|
摘要
为了研究2016年新疆地区近地面PM_(2.5)浓度的时空分布,利用MODIS/Terra 10 km AOD数据和气象数据,结合卫星过境时地面监测的PM_(2.5)数据,采用地理加权回归的方法,构建了新疆地区PM_(2.5)遥感反演模型.结果表明:基于地理加权回归模型反演所得地面PM_(2.5)浓度相关系数R2为0.87,优于多元线性回归模型的0.78.在时间上,新疆地区PM_(2.5)浓度2016年1月份最高,为132.07μg/m3,3月份次之,达到113.22μg/m3,9月份最低,为28.55μg/m3;在空间上,总体来看2016年新疆喀什、和田以及库尔勒北部地区PM_(2.5)浓度较高,采暖季乌鲁木齐市附近PM_(2.5)浓度明显升高.
In order to study the temporal and spatial distribution of PM_(2.5) concentration near the ground in Xinjiang in2016,a PM_(2.5) remote sensing inversion model of Xinjiang was constructed by geographically weighted regression methodusing MODIS/Terra 10 km AOD data and meteorological data,combined with PM_(2.5) data from ground monitoring duringsatellite transit. The results show that the ground PM_(2.5) concentration correlation coefficient inverting by the geographic weightedregression model is 0.87, which is higher than the multiple linear regression method of 0.78. In terms of time,the highestconcentration of PM_(2.5) in January was 132.07 μg/m3,followed by March, the concentration of PM_(2.5) reached 113.22 μg/m3,and the lowest concentration of PM_(2.5) in September was 28.55 μg/m3;In terms of space,overall,the concentration of PM_(2.5) in Kashi and Hotan and northern regions of Korla in Xinjiang is higher in 2016,and the concentration of PM_(2.5) near Urumqiis significantly higher in the heating season.
In order to study the temporal and spatial distribution of PM_(2.5) concentration near the ground in Xinjiang in2016,a PM_(2.5) remote sensing inversion model of Xinjiang was constructed by geographically weighted regression methodusing MODIS/Terra 10 km AOD data and meteorological data,combined with PM_(2.5) data from ground monitoring duringsatellite transit. The results show that the ground PM_(2.5) concentration correlation coefficient inverting by the geographic weightedregression model is 0.87, which is higher than the multiple linear regression method of 0.78. In terms of time,the highestconcentration of PM_(2.5) in January was 132.07 μg/m3,followed by March, the concentration of PM_(2.5) reached 113.22 μg/m3,and the lowest concentration of PM_(2.5) in September was 28.55 μg/m3;In terms of space,overall,the concentration of PM_(2.5) in Kashi and Hotan and northern regions of Korla in Xinjiang is higher in 2016,and the concentration of PM_(2.5) near Urumqiis significantly higher in the heating season.
引文
[1]吴健生,王茜,李嘉诚,等. PM2.5浓度空间分异模拟模型对比:以京津冀地区为例[J].环境科学,2017,38(6):2191-2201.WU J S,WANG X,LI J C,et al. Comparison of models on spatialvariation of PM2.5concentration:A case of Beijing-Tianjin-Hebei region[J]. Environmental Science,2017,38(6):2191-2201(in Chinese).
[2] CHE H,XIA X,ZHU J,et al. Column aerosol optical properties andaerosol radiative forcing during a serious haze-fog month over NorthChina Plain in 2013 based on ground-based sun photometer measure-ments[J]. Atmospheric Chemistry&Physics,2014,14(4):2125-2138.
[3]武鹏飞,赵银军,卢远,等.不同时间尺度大气PM2.5和PM10卫星影像反演[J].广西师范学院学报(自然科学版),2016,33(3):80-88.WU P F,ZHAO Y J,LU Y,et al. Air particulate matter PM10and PM2.5satellite imageries inversion under different time scales[J]. Journal ofGuangxi Normal University(Natural Science),2016,33(3):80-88(inChinese).
[4] ENGEL-COX J A,HOLLOMAN C H,COUTANT B W,et al. Qualita-tive and quantitative evaluation of MODIS satellite sensor data forregional and urban scale air quality[J]. Atmospheric Environment,2004,38(16):2495-2509.
[5] HE Q,HUANG B. Satellite-based mapping of daily high-resolutionground PM2.5,in China via space-time regression modeling[J]. RemoteSensing of Environment,2018,206:72-83.
[6]焦利民,张博恩,许刚,等.气溶胶光学厚度与PM2.5浓度相关关系的时空变异[J].干旱区资源与环境,2016,30(12):34-39.JIAO L M,ZHANG B N,XU G,et al. Spatiotemporal variation ofaerosol optical thickness and PM2.5concentration[J]. Journal of Arid LandResources and Environment,2016,30(12):34-39(in Chinese).
[7]田宏伟. 2006—2016年河南省气溶胶光学厚度时空分布[J].干旱气象,2018,36(1):104-109.TIAN H W. Spatial-temporal distribution characteristics of aerosoloptical depth in He′nan Province from 2006 to 2016[J]. Journal of AridMeteorology,2018,36(1):104-109(in Chinese).
[8] LI S,CHEN L,XIONG X,et al. Retrieval of the haze optical thicknessin North China plain using MODIS data[J]. Transactions on Geoscience&Remote Sensing,2013,51(5):2528-2540.
[9]吴健生,王茜.基于AOD数据反演地面PM2.5浓度研究进展[J].环境科学与技术,2017,40(8):68-76.WU J S,WANG X. Research progress of retrieval ground-level PM2.5concentration based on AOD data[J]. Environmental Science&Techno-logy,2017,40(8):68-76(in Chinese).
[10]张天棋.基于MODIS影像的中国地区气溶胶产品验证与PM2.5反演[D].成都:西南交通大学,2017.ZHANG T Q. Evaluation of the Aerosol Optical Depth Production andRetrieval of the PM2.5in China Using MODIS Images[D]. Chengdu:Southwest Jiaotong University,2017(in Chinese).
[11]胡蝶,张镭,沙莎,等.西北地区MODIS气溶胶产品的对比应用分析[J].干旱气象,2013,31(4):677-683.HU D,ZHANG L,SHA S,et al. Comparative application analysis ofMODIS aerosol products in Northwest China[J]. Journal of Arid Meteo-rology,2013,31(4):677-683(in Chinese).
[12]侯爱华,高伟,王中挺,等.利用高分一号卫星监测开封地区PM2.5[J].国土资源遥感,2017,29(4):161-165.HOU A H,GAO W,WANG Z T,et al. Monitoring of PM2.5in Kaifeng byHigh-altitude one satellite[J]. Remote Sensing for Land and Resources,2017,29(4):161-165(in Chinese).
[13]石灵芝,邓启红,路婵,等.基于BP人工神经网络的大气颗粒物PM10质量浓度预测[J].中南大学学报(自然科学版),2012,43(5):1969-1974.SHI L Z,DENG Q H,LU C,et al. Prediction of PM10mass concentra-tion of atmospheric particulate matter based on BP artificial neuralnetwork[J]. Journal of Central South University(Natural Science),2012,43(5):1969-1974(in Chinese).
[14] MA Z,LIU Y,ZHAO Q,et al. Satellite-derived high resolution PM2.5,concentrations in Yangtze River Delta Region of China using improvedlinear mixed effects model[J]. Atmospheric Environment,2016,133:156-164.
[15] FOTHERINGHAM A S,BRUNSDON C,CHARLTON M. Geographi-cally Weighted Regression:The Analysis of Spatially Varying Relation-ships[M]. Chichester:Wiley,2002.
[16] FOTHERINGHAM A S,CHARLTON M E,BRUNSDON C. Geographi-cally weighted regression:A natural evolution of the expansion methodfor spatial data analysis[J]. Environment&Planning A,1998,30(11):1905-1927.
[17] HU Z. Spatial analysis of MODIS aerosol optical depth,PM2.5,and chr-onic coronary heart disease[J]. International Journal of Health Geograph-ics,2009,8(1):27.
[18] HU X,WALLER L A,ALHAMDAN M Z,et al. Estimating ground-level PM2.5concentrations in the southeastern U.S. using geographicallyweighted regression[J]. Environmental Research,2013,121(2):1-10.
[19]陈辉,厉青,张玉环,等.基于地理加权模型的我国冬季PM2.5遥感估算方法研究[J].环境科学学报,2016,36(6):2142-2151.CHEN H,LI Q,ZHANG Y H,et al. Estimations of PM2.5concentrationsbased on the method of geographically weighted regression[J]. Acta Sci-entiae Circumstantiae,2016,36(6):2142-2151(in Chinese).
[20]王敬,毕晓辉,冯银厂,等.乌鲁木齐市重污染期间PM2.5污染特征与来源解析[J].环境科学研究,2014,27(2):113-119.WANG J,BI X H,FENG Y C,et al. Characteristics and sources ofPM2.5pollution during heavy pollution in urumqi city[J]. EnvironmentalScience Research,2014,27(2):113-119(in Chinese).
[21]新疆维吾尔自治区环境保护厅.新疆维吾尔自治区2015—2017年环境状况公报[EB/OL]. http://www.xjepb.gov.cn/xjepb/_296/_304/in-dex.html,2016-06-05/2018-06-04.Xinjiang department of environmental protection. Bulletin on the Stateof the Environment of Xinjiang Uygur Autonomous Region from 2015 to2017[EB/OL]. http://www.xjepb.gov.cn/xjepb/_296/_304/index.html,2016-06-05/2018-06-04(in Chinese).
[22]黄观,刘伟,刘志红,等.乌鲁木齐市MODIS气溶胶光学厚度与PM10浓度关系模型研究[J].环境科学学报,2016,36(2):649-657.HUANG G,LIU W,LIU Z H,et al. Relationship between MODISaerosol optical depth and PM10ground concentration in Urumqi[J]. ActaScientiae Circumstantiae,2016,36(2):649-657(in Chinese).
[23]李成才,毛节泰,刘启汉,等.利用MODIS光学厚度遥感产品研究北京及周边地区的大气污染[J].大气科学,2003,27(5):869-880.LI C C,MAO J T,LIU Q H,et al. Study on air pollution in Beijing andsurrounding areas using MODIS optical thickness remote sensing prod-ucts[J]. Journal of Atmospheric Sciences,2003,27(5):869-880(inChinese).
[24]孙晓雷,甘伟,林燕,等. MODIS 3 km气溶胶光学厚度产品检验及其环境空气质量指示[J].环境科学学报,2015,35(6):1657-1666.SUN X L,GAN W,LIN Y,et al. MODIS 3 km aerosol optical thick-ness product inspection and its environmental air quality indication[J].Chinese Journal of Environmental Science,2015,35(6):1657-1666(in Chinese).
[25] HSU N C,BETTENHAUSEN C,SAYER A M,et al. Enhanced deepblue aerosol retrieval algorithm:The second generation[J]. Journal ofGeophysical Research Atmospheres,2013,118(16):9296-9315.
[26] SAYER A M,HSU N C,BETTENHAUSEN C,et al. Validation anduncertainty estimates for MODIS Collection 6“Deep Blue”aerosol data[J].Journal of Geophysical Research Atmospheres,2013,118(14):7864-7872.
[27] CLEVELAND W S. Robust locally weighted and smoothing scatter plots[J].Journal of the American Statistical Association,1979,74(368):829-836.
[28] CHU Y,LIU Y,LI X,et al. A Review on predicting ground PM2.5con-centration using satellite aerosol optical depth[J]. Atmosphere,2016,7(10):1-25.
[29] WU J,LI J,PENG J,et al. Applying land use regression model to esti-mate spatial variation of PM2.5,in Beijing,China[J]. EnvironmentalScience&Pollution Research International,2015,22(9):7045-7061.
[30] ABERNETHY R C,ALLEN R W,MCKENDRY I G,et al. A land useregression model for ultrafine particles in Vancouver,Canada[J]. Envi-ronmental Science&Technology,2013,47(10):5217-5225.
[31]谷超.新疆环境空气中PM2.5/PM10比值特征分析[J].干旱环境监测,2017,31(2):70-74.GU C. Analysis of PM2.5/PM10ratio in Xinjiang ambient air[J]. Arid En-vironmental Monitoring,2017,31(2):70-74.
[32]张志伟,杨发相,吴吉龙,等.新疆沙漠空间分布格局与类型结构[J].干旱区研究,2014,31(4):763-770.ZHANG Z W,YANG F X,WU J L,et al. Spatial distribution patternsand type structure of the deserts in Xinjiang[J]. Arid Zone Research,2014,31(4):763-770(in Chinese).
[33]张欣,季晓燕.库尔勒市空气污染现状与成因分析[J].干旱环境监测,2016,30(2):54-57.ZHANG X,JI X Y. Present presentation status and countermeasures ofatmospheric pollution in Korla[J]. Arid Environmental Monitoring,2016,30(2):54-57(in Chinese).
[34]宋健侃.浅析沙尘天气对和田市大气污染的影响及对策[J].干旱环境监测,2003,17(4):227-229.SONG J K. Impact and countermeasure of sand dust weather on air pol-lution in Hetan City[J]. Arid Environmental Monitoring,2003,17(4):227-229(in Chinese).
[35]木尼拉·阿不都木太力甫,玉米提·哈力克,塔依尔江·艾山,等.乌鲁木齐市主要大气污染物浓度变化特征研究[J].生态环境学报,2018,27(3):533-541.MUNILA A B D M T L P,YUMITI H L K,TAYIERJIANG A S,et al.Variation characteristics of the main atmospheric pollutant concentra-tion in Urumqi[J]. Ecology and Environmental Sciences,2018,27(3):533-541(in Chinese).
[2] CHE H,XIA X,ZHU J,et al. Column aerosol optical properties andaerosol radiative forcing during a serious haze-fog month over NorthChina Plain in 2013 based on ground-based sun photometer measure-ments[J]. Atmospheric Chemistry&Physics,2014,14(4):2125-2138.
[3]武鹏飞,赵银军,卢远,等.不同时间尺度大气PM2.5和PM10卫星影像反演[J].广西师范学院学报(自然科学版),2016,33(3):80-88.WU P F,ZHAO Y J,LU Y,et al. Air particulate matter PM10and PM2.5satellite imageries inversion under different time scales[J]. Journal ofGuangxi Normal University(Natural Science),2016,33(3):80-88(inChinese).
[4] ENGEL-COX J A,HOLLOMAN C H,COUTANT B W,et al. Qualita-tive and quantitative evaluation of MODIS satellite sensor data forregional and urban scale air quality[J]. Atmospheric Environment,2004,38(16):2495-2509.
[5] HE Q,HUANG B. Satellite-based mapping of daily high-resolutionground PM2.5,in China via space-time regression modeling[J]. RemoteSensing of Environment,2018,206:72-83.
[6]焦利民,张博恩,许刚,等.气溶胶光学厚度与PM2.5浓度相关关系的时空变异[J].干旱区资源与环境,2016,30(12):34-39.JIAO L M,ZHANG B N,XU G,et al. Spatiotemporal variation ofaerosol optical thickness and PM2.5concentration[J]. Journal of Arid LandResources and Environment,2016,30(12):34-39(in Chinese).
[7]田宏伟. 2006—2016年河南省气溶胶光学厚度时空分布[J].干旱气象,2018,36(1):104-109.TIAN H W. Spatial-temporal distribution characteristics of aerosoloptical depth in He′nan Province from 2006 to 2016[J]. Journal of AridMeteorology,2018,36(1):104-109(in Chinese).
[8] LI S,CHEN L,XIONG X,et al. Retrieval of the haze optical thicknessin North China plain using MODIS data[J]. Transactions on Geoscience&Remote Sensing,2013,51(5):2528-2540.
[9]吴健生,王茜.基于AOD数据反演地面PM2.5浓度研究进展[J].环境科学与技术,2017,40(8):68-76.WU J S,WANG X. Research progress of retrieval ground-level PM2.5concentration based on AOD data[J]. Environmental Science&Techno-logy,2017,40(8):68-76(in Chinese).
[10]张天棋.基于MODIS影像的中国地区气溶胶产品验证与PM2.5反演[D].成都:西南交通大学,2017.ZHANG T Q. Evaluation of the Aerosol Optical Depth Production andRetrieval of the PM2.5in China Using MODIS Images[D]. Chengdu:Southwest Jiaotong University,2017(in Chinese).
[11]胡蝶,张镭,沙莎,等.西北地区MODIS气溶胶产品的对比应用分析[J].干旱气象,2013,31(4):677-683.HU D,ZHANG L,SHA S,et al. Comparative application analysis ofMODIS aerosol products in Northwest China[J]. Journal of Arid Meteo-rology,2013,31(4):677-683(in Chinese).
[12]侯爱华,高伟,王中挺,等.利用高分一号卫星监测开封地区PM2.5[J].国土资源遥感,2017,29(4):161-165.HOU A H,GAO W,WANG Z T,et al. Monitoring of PM2.5in Kaifeng byHigh-altitude one satellite[J]. Remote Sensing for Land and Resources,2017,29(4):161-165(in Chinese).
[13]石灵芝,邓启红,路婵,等.基于BP人工神经网络的大气颗粒物PM10质量浓度预测[J].中南大学学报(自然科学版),2012,43(5):1969-1974.SHI L Z,DENG Q H,LU C,et al. Prediction of PM10mass concentra-tion of atmospheric particulate matter based on BP artificial neuralnetwork[J]. Journal of Central South University(Natural Science),2012,43(5):1969-1974(in Chinese).
[14] MA Z,LIU Y,ZHAO Q,et al. Satellite-derived high resolution PM2.5,concentrations in Yangtze River Delta Region of China using improvedlinear mixed effects model[J]. Atmospheric Environment,2016,133:156-164.
[15] FOTHERINGHAM A S,BRUNSDON C,CHARLTON M. Geographi-cally Weighted Regression:The Analysis of Spatially Varying Relation-ships[M]. Chichester:Wiley,2002.
[16] FOTHERINGHAM A S,CHARLTON M E,BRUNSDON C. Geographi-cally weighted regression:A natural evolution of the expansion methodfor spatial data analysis[J]. Environment&Planning A,1998,30(11):1905-1927.
[17] HU Z. Spatial analysis of MODIS aerosol optical depth,PM2.5,and chr-onic coronary heart disease[J]. International Journal of Health Geograph-ics,2009,8(1):27.
[18] HU X,WALLER L A,ALHAMDAN M Z,et al. Estimating ground-level PM2.5concentrations in the southeastern U.S. using geographicallyweighted regression[J]. Environmental Research,2013,121(2):1-10.
[19]陈辉,厉青,张玉环,等.基于地理加权模型的我国冬季PM2.5遥感估算方法研究[J].环境科学学报,2016,36(6):2142-2151.CHEN H,LI Q,ZHANG Y H,et al. Estimations of PM2.5concentrationsbased on the method of geographically weighted regression[J]. Acta Sci-entiae Circumstantiae,2016,36(6):2142-2151(in Chinese).
[20]王敬,毕晓辉,冯银厂,等.乌鲁木齐市重污染期间PM2.5污染特征与来源解析[J].环境科学研究,2014,27(2):113-119.WANG J,BI X H,FENG Y C,et al. Characteristics and sources ofPM2.5pollution during heavy pollution in urumqi city[J]. EnvironmentalScience Research,2014,27(2):113-119(in Chinese).
[21]新疆维吾尔自治区环境保护厅.新疆维吾尔自治区2015—2017年环境状况公报[EB/OL]. http://www.xjepb.gov.cn/xjepb/_296/_304/in-dex.html,2016-06-05/2018-06-04.Xinjiang department of environmental protection. Bulletin on the Stateof the Environment of Xinjiang Uygur Autonomous Region from 2015 to2017[EB/OL]. http://www.xjepb.gov.cn/xjepb/_296/_304/index.html,2016-06-05/2018-06-04(in Chinese).
[22]黄观,刘伟,刘志红,等.乌鲁木齐市MODIS气溶胶光学厚度与PM10浓度关系模型研究[J].环境科学学报,2016,36(2):649-657.HUANG G,LIU W,LIU Z H,et al. Relationship between MODISaerosol optical depth and PM10ground concentration in Urumqi[J]. ActaScientiae Circumstantiae,2016,36(2):649-657(in Chinese).
[23]李成才,毛节泰,刘启汉,等.利用MODIS光学厚度遥感产品研究北京及周边地区的大气污染[J].大气科学,2003,27(5):869-880.LI C C,MAO J T,LIU Q H,et al. Study on air pollution in Beijing andsurrounding areas using MODIS optical thickness remote sensing prod-ucts[J]. Journal of Atmospheric Sciences,2003,27(5):869-880(inChinese).
[24]孙晓雷,甘伟,林燕,等. MODIS 3 km气溶胶光学厚度产品检验及其环境空气质量指示[J].环境科学学报,2015,35(6):1657-1666.SUN X L,GAN W,LIN Y,et al. MODIS 3 km aerosol optical thick-ness product inspection and its environmental air quality indication[J].Chinese Journal of Environmental Science,2015,35(6):1657-1666(in Chinese).
[25] HSU N C,BETTENHAUSEN C,SAYER A M,et al. Enhanced deepblue aerosol retrieval algorithm:The second generation[J]. Journal ofGeophysical Research Atmospheres,2013,118(16):9296-9315.
[26] SAYER A M,HSU N C,BETTENHAUSEN C,et al. Validation anduncertainty estimates for MODIS Collection 6“Deep Blue”aerosol data[J].Journal of Geophysical Research Atmospheres,2013,118(14):7864-7872.
[27] CLEVELAND W S. Robust locally weighted and smoothing scatter plots[J].Journal of the American Statistical Association,1979,74(368):829-836.
[28] CHU Y,LIU Y,LI X,et al. A Review on predicting ground PM2.5con-centration using satellite aerosol optical depth[J]. Atmosphere,2016,7(10):1-25.
[29] WU J,LI J,PENG J,et al. Applying land use regression model to esti-mate spatial variation of PM2.5,in Beijing,China[J]. EnvironmentalScience&Pollution Research International,2015,22(9):7045-7061.
[30] ABERNETHY R C,ALLEN R W,MCKENDRY I G,et al. A land useregression model for ultrafine particles in Vancouver,Canada[J]. Envi-ronmental Science&Technology,2013,47(10):5217-5225.
[31]谷超.新疆环境空气中PM2.5/PM10比值特征分析[J].干旱环境监测,2017,31(2):70-74.GU C. Analysis of PM2.5/PM10ratio in Xinjiang ambient air[J]. Arid En-vironmental Monitoring,2017,31(2):70-74.
[32]张志伟,杨发相,吴吉龙,等.新疆沙漠空间分布格局与类型结构[J].干旱区研究,2014,31(4):763-770.ZHANG Z W,YANG F X,WU J L,et al. Spatial distribution patternsand type structure of the deserts in Xinjiang[J]. Arid Zone Research,2014,31(4):763-770(in Chinese).
[33]张欣,季晓燕.库尔勒市空气污染现状与成因分析[J].干旱环境监测,2016,30(2):54-57.ZHANG X,JI X Y. Present presentation status and countermeasures ofatmospheric pollution in Korla[J]. Arid Environmental Monitoring,2016,30(2):54-57(in Chinese).
[34]宋健侃.浅析沙尘天气对和田市大气污染的影响及对策[J].干旱环境监测,2003,17(4):227-229.SONG J K. Impact and countermeasure of sand dust weather on air pol-lution in Hetan City[J]. Arid Environmental Monitoring,2003,17(4):227-229(in Chinese).
[35]木尼拉·阿不都木太力甫,玉米提·哈力克,塔依尔江·艾山,等.乌鲁木齐市主要大气污染物浓度变化特征研究[J].生态环境学报,2018,27(3):533-541.MUNILA A B D M T L P,YUMITI H L K,TAYIERJIANG A S,et al.Variation characteristics of the main atmospheric pollutant concentra-tion in Urumqi[J]. Ecology and Environmental Sciences,2018,27(3):533-541(in Chinese).