基于支持向量机的京津冀城市群热环境时空形态模拟
|
摘要
城市群热环境作为区域生态重要组成部分,已成为近年来的研究热点。而如何选择针对城市群这种复杂地地貌特征的热环境量化工具一直是亟待解决的技术难点,基于此本研究提出了一种解决多样本、非线性、非平稳及高维函数拟合的计算方法,并建立了基于支持向量机(SVM)的京津冀城市群热环境曲面模型来揭示城市群热环境的时空形态变化。研究结果表明:①SVM模型在刻画多核心、多种土地利用类型城市群热环境的空间分布方面具有理论与实践可行性,能够根据热环境的整体空间布局通过高斯核函数进行局部优化差值,最大限度减少缺省值对模型拟合结果的影响。相比于对照方法可以模拟出更高精度的复杂地貌特征城市群热岛空间分布格局;②在SVM模型曲面拟合的过程中,拟合精度和拟合时间是衡量拟合结果的重要指标,而原始影像的分辨率则是影响该指标的决定性因素;③2003-2013年区域内北京市与天津市的城市热岛效应变化最为明显,热岛面积分别增加7091 km2与4196 km2,空间上呈现出逐年接近连片发展趋势,热岛重心移动轨迹具有明显的时空分异性。北京城市热岛特征为东南部地区异速增长,西部地区缓慢增长;天津城市热岛特征为以城市中心为圆心向周围扩展。本研究进一步丰富了城市群热环境评测的定量方法,可以在实践上对城市群的城市规划、城市建设、环境保护和区域可持续发展等提供定量化、可视化的决策支持。
As an important part of regional ecology, the thermal environment of urban agglomeration has become a research hot topic in recent years. How to choose the thermal environment quantification mehod for the complex geomorphological features of urban agglomeration has been a difficult technical problem to be solved.Based on this, This study proposes a solution to multi-sample, nonlinear, non-stationary and high-dimensional function fitting. The calculation method is established, and the thermal environment surface model of BeijingTianjin-Hebei urban agglomeration based on support vector machine(SVM) is established to reveal the temporal and spatial morphological changes of the thermal environment of urban agglomeration. The results show:(1)that the SVM model has theoretical and practical feasibility in describing the spatial distribution of the thermal environment of urban agglomerations with multi-core and multi-land-use types. It can optimize the differences locally through the Gaussian Kernel Function according to the overall spatial distribution of the thermal environment,and minimize the impact of default values on the fitting results of the model. Comparing with the control method, the spatial distribution pattern of heat island in urban agglomerations with complex geomorphologic features can be simulated with higher accuracy.(2) In the process of fitting the surface of SVM model, accuracy and the time of fitting are important indexes to measure the results, and original image resolution is the decisive influencing factor.(3) In 2003-2013, the most obviously change regions of urban heat island effect are Beijing and Tianjin. The heat island area of the two cities increased by 7091 km square and 4196 km square, respectively. The spatial trend was developing continually year by year, and the trajectory of gravity center of the heat island had obvious spatial and temporal variations. Beijing's urban heat island is characterized by uneven growth in the southeast and slow growth in the west, while Tianjin's urban heat island is characterized by the expansion of city center to the surrounding areas. This study further enriches the quantitative methods of urban thermal environment assessment, and can provide quantitative and visual decision supports for urban agglomeration planning, urban construction, environmental protection and regional sustainable development practically.
As an important part of regional ecology, the thermal environment of urban agglomeration has become a research hot topic in recent years. How to choose the thermal environment quantification mehod for the complex geomorphological features of urban agglomeration has been a difficult technical problem to be solved.Based on this, This study proposes a solution to multi-sample, nonlinear, non-stationary and high-dimensional function fitting. The calculation method is established, and the thermal environment surface model of BeijingTianjin-Hebei urban agglomeration based on support vector machine(SVM) is established to reveal the temporal and spatial morphological changes of the thermal environment of urban agglomeration. The results show:(1)that the SVM model has theoretical and practical feasibility in describing the spatial distribution of the thermal environment of urban agglomerations with multi-core and multi-land-use types. It can optimize the differences locally through the Gaussian Kernel Function according to the overall spatial distribution of the thermal environment,and minimize the impact of default values on the fitting results of the model. Comparing with the control method, the spatial distribution pattern of heat island in urban agglomerations with complex geomorphologic features can be simulated with higher accuracy.(2) In the process of fitting the surface of SVM model, accuracy and the time of fitting are important indexes to measure the results, and original image resolution is the decisive influencing factor.(3) In 2003-2013, the most obviously change regions of urban heat island effect are Beijing and Tianjin. The heat island area of the two cities increased by 7091 km square and 4196 km square, respectively. The spatial trend was developing continually year by year, and the trajectory of gravity center of the heat island had obvious spatial and temporal variations. Beijing's urban heat island is characterized by uneven growth in the southeast and slow growth in the west, while Tianjin's urban heat island is characterized by the expansion of city center to the surrounding areas. This study further enriches the quantitative methods of urban thermal environment assessment, and can provide quantitative and visual decision supports for urban agglomeration planning, urban construction, environmental protection and regional sustainable development practically.
引文
[1]刘伟东,尤焕苓,孙丹.1971-2010年京津冀大城市热岛效应多时间尺度分析[J].气象,2016,42(5):598-606.[Liu W D,You H L,Sun D.Multi-time scale analysis of megacities heat island effect in Beijing-Tianjin-Hebei region from 1971 to 2010[J].Meteorological Monthly,2016,42(5):598-606.]
[2]陈彬辉,冯瑶,袁建国,等.基于MODIS地表温度的京津冀地区城市热岛时空差异研究[J].北京大学学报(自然科学版),2016,52(6):1134-1140.[Chen B H,Fen Y,Yuan JG,et al.Spatiotemporal difference of urban heat island in Jing-Jin-Ji area based on MODIS land surface temperature[J].Acta Scientiarum Naturalium Universitatis Pekinensis,2016,52(6):1134-1140.]
[3]金鑫,金彦香,杨登兴.SWAT模型在土地利用/覆被变化剧烈地区的改进与应用[J].地球信息科学学报,2018,20(8):1064-1073.[Jin X,Jin Y X,Yang D X.Improved SWAT and its application to a region with severe land use/land cover changes[J]Journal of Geo-information Science,2018,20(8):1064-1073.]
[4]Camila M,Nakata-Osaki,Léa Cristina L S,et al.Thistool for heat island simulation:A GIS extension model to calculate urban heat island intensity based on urban geometry[J].Computers,Environment and Urban Systems,2018,67:157-168.
[5]彭保发,石忆邵,王贺封,等.城市热岛效应的影响机理及其作用规律——以上海市为例[J].地理学报,2013,68(11):1461-1471.[Peng B F,Shi Y S,Wang H F,et al.The impacting mechanism and laws of function of urban heat islands effect:A case study of Shanghai[J].Acta Geographica Sinica,2013,68(11):1461-1471.]
[6]Chen A L,Zhao X F,Yao L,et al.Application of a new integrated landscape index to predict potential urban heat islands[J].Ecological Indicators,2016,69:828-835.
[7]Jose L F R,Filho A J Pereira,Karam H A.Estimation of long term low resolution surface urban heat island intensities for tropical cities using MODIS remote sensing data[J].Urban Climate,2016,17:32-66.
[8]侯浩然,丁凤,黎勤生.近20年来福州城市热环境变化遥感分析[J].地球信息科学学报,2018,20(3):385-395.[Hou H R,Ding F,Li Q S.Remote sensing analysis of changes of urban thermal environment of Fuzhou city in China in the past 20 years[J].Journal of Geo-information Science,2018,20(3):385-395.]
[9]Sun Q,Wu Z,Tan J.The relationship between land surface temperature and land use/land cover in Guangzhou,China[J].Environmental Earth Sciences,2012,65(6):1687-1694.
[10]陈云浩,史培军,李晓兵,等.城市空间热环境的遥感研究——热场结构及其演变的分形测量[J].测绘学报,2002,31(4):322-326.[Chen Y H,Shi P J,Li X B,et al.Research on urban spatial thermal environment using remote sensing image[J].Acta Geodaetica et Cartographica Sinica,2002,31(4):322-326.]
[11]张心怡,刘敏,孟飞.基于RS和GIS的上海城建用地扩展研究[J].长江流域资源与环境,2006,15(1):29-33.[Zhang XY,Liu M,Meng F.Expansion of urban construction land in Shanghai City based on RS and GIS.Resources and Environment in the Yangtze Basin,2006,15(1):29-33.]
[12]王福海,陈丹.建设用地扩展对城市热环境影响研究[J].中国人口·资源与环境,2016,26(S1):194-197.[Wang FH,Chen D.Temporal-spatial variation of the thermal environment impact research by distribution characteristics of construction land[J].Chinese Journal of Population Resources and Environment,2016,26(S1):194-197.]
[13]张丽娟,于洋,粟练灵,等.1960-2010年黑龙江省土地利用变化对气温的影响[J].地理科学,2017,37(6):952-959.[Zhang L J,Yu Y,Li LL,et al.Effects of land use change on air temperature of Heilongjiang Province in 1960-2010[J].Chinese Geographical Science,2017,37(6):952-959.]
[14]Liu X,Wang C L,Jing Y S,et al.Annual variations of temperature on four urban underlying surfaces and fitting analysis[J].Journal of Tropical Meteorology,2013,19(2):138-144.
[15]Pramod N Chivate,Andrei G Jablokow.Review of surface representations and fitting for reverse engineering[J].Computer Integrated Manufacturing Systems,1995,8(3):193-204.
[16]占文凤,陈云浩,周纪等.支持向量机的北京城市热岛模拟:热岛强度空间格局曲面模拟及其应用[J].测绘学报,2011,40(1):96-103.[Zhan W F,Chen Y H,Zhou J,et al.Spatial simulation of urban heat island intensity based on the support vector machine technique:A case study in Beijing[J].Acta Geodaetica et Cartographica Sinica,2011,40(1):96-103.]
[17]Vapnik V.The nature of statistical learning theory[M]New York:Springer-Verlag,1995.
[18]丁世飞,齐丙娟,谭红艳.支持向量机理论与算法研究综述[J].电子科技大学学报,2011,40(1):2-10.[Ding S F,Qi B J,Tan H Y.An overview on theory and algorithm of support vector machines[J].Journal of University of Electronic Science and Technology of China,2011,40(1):2-10.]
[19]乔治,田光进.北京市热环境时空分异与区划[J].遥感学报,2014,18(3):715-734.[Qiao Z,Tian G J.Spatiotemporal diversity and regionalization of the urban thermal environment in Beijing[J].Journal of Remote Sensing,2014,18(3):715-734.]
[20]郭兵,陶和平,刘斌涛,等.基于GIS、Delaunay三角形的复杂地形区经济重心获取新方法[J].地理与地理信息科学,2012,28(2):55-59.[Guo B,Tao H P,Liu B,et al.Research on acquisition method of economic gravity center based on GIS and Delaunay in complicated terrain area[J].Geography and Geo-information Science,2012,28(2):55-59.]
[21]王劲峰,葛咏,李连发,等.地理学时空数据分析方法[J].地理学报,2014,69(9):1326-1345.[Wang J F,Ge Y,Li L F,et al.Spatiotemporal data analysis in geography[J].Acta Geologica Sinica,2014,69(9):1326-1345.]
[2]陈彬辉,冯瑶,袁建国,等.基于MODIS地表温度的京津冀地区城市热岛时空差异研究[J].北京大学学报(自然科学版),2016,52(6):1134-1140.[Chen B H,Fen Y,Yuan JG,et al.Spatiotemporal difference of urban heat island in Jing-Jin-Ji area based on MODIS land surface temperature[J].Acta Scientiarum Naturalium Universitatis Pekinensis,2016,52(6):1134-1140.]
[3]金鑫,金彦香,杨登兴.SWAT模型在土地利用/覆被变化剧烈地区的改进与应用[J].地球信息科学学报,2018,20(8):1064-1073.[Jin X,Jin Y X,Yang D X.Improved SWAT and its application to a region with severe land use/land cover changes[J]Journal of Geo-information Science,2018,20(8):1064-1073.]
[4]Camila M,Nakata-Osaki,Léa Cristina L S,et al.Thistool for heat island simulation:A GIS extension model to calculate urban heat island intensity based on urban geometry[J].Computers,Environment and Urban Systems,2018,67:157-168.
[5]彭保发,石忆邵,王贺封,等.城市热岛效应的影响机理及其作用规律——以上海市为例[J].地理学报,2013,68(11):1461-1471.[Peng B F,Shi Y S,Wang H F,et al.The impacting mechanism and laws of function of urban heat islands effect:A case study of Shanghai[J].Acta Geographica Sinica,2013,68(11):1461-1471.]
[6]Chen A L,Zhao X F,Yao L,et al.Application of a new integrated landscape index to predict potential urban heat islands[J].Ecological Indicators,2016,69:828-835.
[7]Jose L F R,Filho A J Pereira,Karam H A.Estimation of long term low resolution surface urban heat island intensities for tropical cities using MODIS remote sensing data[J].Urban Climate,2016,17:32-66.
[8]侯浩然,丁凤,黎勤生.近20年来福州城市热环境变化遥感分析[J].地球信息科学学报,2018,20(3):385-395.[Hou H R,Ding F,Li Q S.Remote sensing analysis of changes of urban thermal environment of Fuzhou city in China in the past 20 years[J].Journal of Geo-information Science,2018,20(3):385-395.]
[9]Sun Q,Wu Z,Tan J.The relationship between land surface temperature and land use/land cover in Guangzhou,China[J].Environmental Earth Sciences,2012,65(6):1687-1694.
[10]陈云浩,史培军,李晓兵,等.城市空间热环境的遥感研究——热场结构及其演变的分形测量[J].测绘学报,2002,31(4):322-326.[Chen Y H,Shi P J,Li X B,et al.Research on urban spatial thermal environment using remote sensing image[J].Acta Geodaetica et Cartographica Sinica,2002,31(4):322-326.]
[11]张心怡,刘敏,孟飞.基于RS和GIS的上海城建用地扩展研究[J].长江流域资源与环境,2006,15(1):29-33.[Zhang XY,Liu M,Meng F.Expansion of urban construction land in Shanghai City based on RS and GIS.Resources and Environment in the Yangtze Basin,2006,15(1):29-33.]
[12]王福海,陈丹.建设用地扩展对城市热环境影响研究[J].中国人口·资源与环境,2016,26(S1):194-197.[Wang FH,Chen D.Temporal-spatial variation of the thermal environment impact research by distribution characteristics of construction land[J].Chinese Journal of Population Resources and Environment,2016,26(S1):194-197.]
[13]张丽娟,于洋,粟练灵,等.1960-2010年黑龙江省土地利用变化对气温的影响[J].地理科学,2017,37(6):952-959.[Zhang L J,Yu Y,Li LL,et al.Effects of land use change on air temperature of Heilongjiang Province in 1960-2010[J].Chinese Geographical Science,2017,37(6):952-959.]
[14]Liu X,Wang C L,Jing Y S,et al.Annual variations of temperature on four urban underlying surfaces and fitting analysis[J].Journal of Tropical Meteorology,2013,19(2):138-144.
[15]Pramod N Chivate,Andrei G Jablokow.Review of surface representations and fitting for reverse engineering[J].Computer Integrated Manufacturing Systems,1995,8(3):193-204.
[16]占文凤,陈云浩,周纪等.支持向量机的北京城市热岛模拟:热岛强度空间格局曲面模拟及其应用[J].测绘学报,2011,40(1):96-103.[Zhan W F,Chen Y H,Zhou J,et al.Spatial simulation of urban heat island intensity based on the support vector machine technique:A case study in Beijing[J].Acta Geodaetica et Cartographica Sinica,2011,40(1):96-103.]
[17]Vapnik V.The nature of statistical learning theory[M]New York:Springer-Verlag,1995.
[18]丁世飞,齐丙娟,谭红艳.支持向量机理论与算法研究综述[J].电子科技大学学报,2011,40(1):2-10.[Ding S F,Qi B J,Tan H Y.An overview on theory and algorithm of support vector machines[J].Journal of University of Electronic Science and Technology of China,2011,40(1):2-10.]
[19]乔治,田光进.北京市热环境时空分异与区划[J].遥感学报,2014,18(3):715-734.[Qiao Z,Tian G J.Spatiotemporal diversity and regionalization of the urban thermal environment in Beijing[J].Journal of Remote Sensing,2014,18(3):715-734.]
[20]郭兵,陶和平,刘斌涛,等.基于GIS、Delaunay三角形的复杂地形区经济重心获取新方法[J].地理与地理信息科学,2012,28(2):55-59.[Guo B,Tao H P,Liu B,et al.Research on acquisition method of economic gravity center based on GIS and Delaunay in complicated terrain area[J].Geography and Geo-information Science,2012,28(2):55-59.]
[21]王劲峰,葛咏,李连发,等.地理学时空数据分析方法[J].地理学报,2014,69(9):1326-1345.[Wang J F,Ge Y,Li L F,et al.Spatiotemporal data analysis in geography[J].Acta Geologica Sinica,2014,69(9):1326-1345.]