CAO Zheng, WEN Ya, SONG Song, HUNG Chak Ho, SUN Hui, 2021. Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China. Chinese Geographical Science, 31(3): 444−458 doi:  10.1007/s11769-021-1203-y
Citation: CAO Zheng, WEN Ya, SONG Song, HUNG Chak Ho, SUN Hui, 2021. Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China. Chinese Geographical Science, 31(3): 444−458 doi:  10.1007/s11769-021-1203-y

Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China

doi: 10.1007/s11769-021-1203-y
Funds:  Under the auspices of National Natural Science Foundation of China (No. 41901219, 41671430, 41801326), Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (No. GML2019ZD0301)
More Information
  • Spatiotemporal variations of anthropogenic heat flux (AHF) is reported to be associated with global warming. However, confined to the low spatial resolution of energy consumption statistical data, details of AHF was not well descripted. To obtain high spatial resolution data of AHF, Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light time-series product and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite monthly normalized difference vegetation index (NDVI) product were applied to construct the human settlement index. Based on the spatial regression relationship between human settlement index and energy consumption data. A 1-km resolution dataset of AHF of 12 selected cities in the eastern China was obtained. Ordinary least-squares (OLS) model was applied to detect the mechanism of spatial patterns of AHF. Results showed that industrial emission in selected cities of the eastern China was accountable for 63% of the total emission. AHF emission in megacities, such as Tianjin, Jinan, Qingdao, and Hangzhou, was most significant. AHF increasing speed in most areas in the chosen cities was quite low. High growth or extremely high growth of AHF were located in central downtown areas. In Beijing, Shanghai, Guangzhou, Jinan, Hangzhou, Changzhou, Zhaoqing, and Jiangmen, a single kernel of AHF was observed. Potential influencing factors showed that precipitation, temperature, elevation, normalized different vegetation index, gross domestic product, and urbanization level were positive with AHF. Overall, this investigation implied that urbanization level and economic development level might dominate the increasing of AHF and the spatial heterogeneousness of AHF. Higher urbanization level or economic development level resulted in high increasing speeds of AHF. These findings provide a novel way to reconstruct of AHF and scientific supports for energy management strategy development.
  • 加载中
  • [1] Afshari A, Schuch F, Marpu P, 2018. Estimation of the traffic related anthropogenic heat release using BTEX measurements: a case study in Abu Dhabi. Urban Climate, 24: 311–325. doi:  10.1016/j.uclim.2017.02.001
    [2] Anselin L, 1995. Local indicators of spatial association—LISA. Geographical Analysis, 27(2): 93–115. doi:  10.1111/j.1538-4632.1995.tb00338.x
    [3] Best M J, Grimmond C S B, 2016. Investigation of the impact of anthropogenic heat flux within an urban land surface model and PILPS-urban. Theoretical and Applied Climatology, 126(1–2): 51–60. doi:  10.1007/s00704-015-1554-3
    [4] Buchwitz M, Schneising O, Burrows J P et al., 2007. First direct observation of the atmospheric CO2 year-to-year increase from space. Atmospheric Chemistry and Physics, 7(16): 4249–4256. doi:  10.5194/acp-7-4249-2007
    [5] Cao Z, Wu Z F, Liu L et al., 2019. Assessing the relationship between anthropogenic heat release warming and building characteristics in Guangzhou: a sustainable development perspective. Science of the Total Environment, 695: 133759. doi:  10.1016/j.scitotenv.2019.133759
    [6] Chadburn S E, Burke E J, Cox P M et al., 2017. An observation-based constraint on permafrost loss as a function of global warming. Nature Climate Change, 7: 340–344. doi:  10.1038/nclimate3262
    [7] Chen B, Dong L, Shi G Y et al., 2014. Anthropogenic heat release: estimation of global distribution and possible climate effect. Journal of the Meteorological Society of Japan, 92A: 157–165. doi:  10.2151/jmsj.2014-A10
    [8] Chen S S, Hu D Y, 2017. Parameterizing anthropogenic heat flux with an energy-consumption inventory and multi-source remote sensing data. Remote Sensing, 9(11): 1165. doi:  10.3390/rs9111165
    [9] Doan V Q, Kusaka H, Nguyen T M, 2019. Roles of past, present, and future land use and anthropogenic heat release changes on urban heat island effects in Hanoi, Vietnam: Numerical experiments with a regional climate model. Sustainable Cities and Society, 47: 101479. doi:  10.1016/j.scs.2019.101479
    [10] Dong Y, Varquez A C G, Kanda M, 2017. Global anthropogenic heat flux database with high spatial resolution. Atmospheric Environment, 150: 276–294. doi:  10.1016/j.atmosenv.2016.11.040
    [11] Falasca S, Catalano F, Moroni M, 2016. Numerical study of the daytime planetary boundary layer over an idealized urban area: influence of surface properties, anthropogenic heat flux, and geostrophic wind intensity. Journal of Applied Meteorology and Climatology, 55(4): 1021–1039. doi:  10.1175/JAMC-D-15-0135.1
    [12] Feng J M, Wang Y L, Ma Z G et al., 2012. Simulating the regional impacts of urbanization and anthropogenic heat release on climate across China. Journal of Climate, 25(20): 7187–7203. doi:  10.1175/JCLI-D-11-00333.1
    [13] Fry G, Sarlöv-Herlin I, 1997. The ecological and amenity functions of woodland edges in the agricultural landscape; a basis for design and management. Landscape and Urban Planning, 37(1–2): 45–55. doi:  10.1016/S0169-2046(96)00369-6
    [14] He C Y, Ma Q, Li T et al., 2012. Spatiotemporal dynamics of electric power consumption in Chinese Mainland from 1995 to 2008 modeled using DMSP/OLS stable nighttime lights data. Journal of Geographical Sciences, 22(5): 125–136. doi:  10.1007/s11442-012-0916-3
    [15] Holst C C, Tam C Y, Chan J C L, 2016. Sensitivity of urban rainfall to anthropogenic heat flux: a numerical experiment. Geophysical Research Letters, 43(5): 2240–2248. doi:  10.1002/2015GL067628
    [16] Hutchinson M F, McKenney D W, Lawrence K et al., 2009. Development and testing of canada-wide interpolated spatial models of daily minimum-maximum temperature and precipitation for 1961–2003. Journal of Applied Meteorology and Climatology, 48(4): 725–741. doi:  10.1175/2008JAMC1979.1
    [17] Iamarino M, Beevers S, Grimmond C S B, 2012. High‐resolution (space, time) anthropogenic heat emissions: London 1970–2025. International Journal of Climatology, 32(11): 1754–1767. doi:  10.1002/joc.2390
    [18] Ichinose T, Shimodozono K, Hanaki K, 1999. Impact of anthropogenic heat on urban climate in Tokyo. Atmospheric Environment, 33(24–25): 3897–3909. doi:  10.1016/S1352-2310(99)00132-6
    [19] Kato S, Yamaguchi Y, 2005. Analysis of urban heat-island effect using ASTER and ETM+ Data: Separation of anthropogenic heat discharge and natural heat radiation from sensible heat flux. Remote Sensing of Environment, 99(1–2): 44–54. doi:  10.1016/j.rse.2005.04.026
    [20] Koralegedara S B, Lin C Y, Sheng Y F et al., 2016. Estimation of anthropogenic heat emissions in urban Taiwan and their spatial patterns. Environmental Pollution, 215: 84–95. doi:  10.1016/j.envpol.2016.04.055
    [21] Lafferty K D, 2009. The ecology of climate change and infectious diseases. Ecology, 90(4): 888–900. doi:  10.1890/08-0079.1
    [22] Lindberg F, Grimmond C S B, Yogeswaran N et al., 2013. Impact of city changes and weather on anthropogenic heat flux in Europe 1995–2015. Urban Climate, 4: 1–15. doi:  10.1016/j.uclim.2013.03.002
    [23] Liu X, Hu G, Chen Y et al., 2018. High-resolution multi-temporal mapping of global urban land using Landsat images based on the Google Earth Engine Platform. Remote Sensing of Environment, 209: 227–239. doi:  10.1016/j.rse.2018.02.055
    [24] Lu Y, Wang Q G, Zhang Y Y et al., 2016. An estimate of anthropogenic heat emissions in China. International Journal of Climatology, 36(3): 1134–1142. doi:  10.1002/joc.4407
    [25] Ma S Y, Ma Z X, 2011. The efficiency of Chinese energy consumption and GDP. Energy Procedia, 5: 1727–1731. doi:  10.1016/j.egypro.2011.03.294
    [26] McMichael A J, Woodruff R E, Hales S, 2006. Climate change and human health: present and future risks. The Lancet, 367(9513): 859–869. doi:  10.1016/S0140-6736(06)68079-3
    [27] Nie W S, Sun T, Ni G H, 2014. Spatiotemporal characteristics of anthropogenic heat in an urban environment: a case study of Tsinghua Campus. Building and Environment, 82: 675–686. doi:  10.1016/j.buildenv.2014.10.011
    [28] Oke T R, 1982. The energetic basis of the urban heat island. Quarterly Journal of the Royal Meteorological Society, 108(455): 1–24. doi:  10.1002/qj.49710845502
    [29] Panetta A M, Stanton M L, Harte J, 2018. Climate warming drives local extinction: evidence from observation and experimentation. Science Advances, 4(2): eaaq1819. doi:  10.1126/sciadv.aaq1819
    [30] Patz J A, Campbell-Lendrum D, Holloway T et al., 2005. Impact of regional climate change on human health. Nature, 438(7066): 310–317. doi:  10.1038/nature04188
    [31] Peng J, Chen S, Lü H L et al., 2016. Spatiotemporal patterns of remotely sensed PM2.5 concentration in China from 1999 to 2011. Remote Sensing of Environment, 174: 109–121. doi:  10.1016/j.rse.2015.12.008
    [32] Peng J, Tian L, Liu Y X et al., 2017. Ecosystem services response to urbanization in metropolitan areas: Thresholds identification. Science of the Total Environment, 607–60: 706–714. doi:  10.1016/j.scitotenv.2017.06.218
    [33] Pigeon G, Legain D, Durand P et al., 2007. Anthropogenic heat release in an old European agglomeration (Toulouse, France). International Journal of Climatology, 27(14): 1969–1981. doi:  10.1002/joc.1530
    [34] Price D T, McKenney D W, Nalder I A et al., 2000. A comparison of two statistical methods for spatial interpolation of canadian monthly mean climate data. Agricultural and Forest Meteorology, 101(2–3): 81–94. doi:  10.1016/S0168-1923(99)00169-0
    [35] Sailor D J, 2011. A review of methods for estimating anthropogenic heat and moisture emissions in the urban environment. International Journal Of Climatology, 31(2): 189–199. doi:  10.1002/joc.2106
    [36] Su S L, Xiao R, Jiang Z L et al., 2012. Characterizing landscape pattern and ecosystem service value changes for urbanization impacts at an eco-regional scale. Applied Geography, 34: 295–305. doi:  10.1016/j.apgeog.2011.12.001
    [37] Taha H, 1997. Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic heat. Energy and Buildings, 25: 99–103. doi:  10.1016/S0378-7788(96)00999-1
    [38] Wang J, Huang B, Fu D J et al., 2015. Spatiotemporal variation in surface urban heat island intensity and associated determinants across major Chinese cities. Remote Sensing, 7(4): 3670–3689. doi:  10.3390/rs70403670
    [39] Wang J, Zhou W Q, Jiao M et al., 2020. Significant effects of ecological context on urban trees’ cooling efficiency. ISPRS Journal of Photogrammetry and Remote Sensing, 159: 78–89. doi:  10.1016/j.isprsjprs.2019.11.001
    [40] Wang S S, Hu D Y, Chen S S et al., 2019. A partition modeling for anthropogenic heat flux mapping in China. Remote Sensing, 11(9): 1132. doi:  10.3390/rs11091132
    [41] Wang S S, Hu D Y, Yu C et al., 2020. Mapping China’s time-series anthropogenic heat flux with inventory method and multi-source remotely sensed data. Science of the Total Environment, 734: 139457. doi:  10.1016/j.scitotenv.2020.139457
    [42] Wang Y C, Bian Z F, Qin K et al., 2019. A modified building energy model coupled with urban parameterization for estimating anthropogenic heat in urban areas. Energy and Buildings, 202: 109377. doi:  10.1016/j.enbuild.2019.109377
    [43] Wong M S, Yang J X, Nichol J et al., 2015. Modeling of anthropogenic heat flux using HJ-1B Chinese small satellite image: a study of heterogeneous urbanized areas in Hong Kong. IEEE Geoscience and Remote Sensing Letters, 12(7): 1466–1470. doi:  10.1109/LGRS.2015.2409111
    [44] Yu B L, Shi K F, Hu Y J et al., 2015. Poverty evaluation using NPP-VIIRS nighttime light composite data at the county level in China. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 8(3): 1217–1229. doi:  10.1109/JSTARS.2015.2399416
    [45] Yu B L, Deng S Q, Liu G et al., 2018. Nighttime light images reveal spatial-temporal dynamics of global anthropogenic resources accumulation above ground. Environmental Science & Technology, 52(20): 11520–11527. doi:  10.1021/acs.est.8b02838
    [46] Zhang N, Wang X M, Chen Y et al., 2016. Numerical simulations on influence of urban land cover expansion and anthropogenic heat release on urban meteorological environment in Pearl River Delta. Theoretical and Applied Climatology, 126(3–4): 469–479. doi:  10.1007/s00704-015-1601-0
    [47] Zhang X M, Estoque R C, Murayama Y, 2017. An urban heat island study in Nanchang City, China based on land surface temperature and social-ecological variables. Sustainable Cities and Society, 32: 557–568. doi:  10.1016/j.scs.2017.05.005
    [48] Zhang Y S, Balzter H, Wu X C, 2013. Spatial–temporal patterns of urban anthropogenic heat discharge in Fuzhou, China, observed from sensible heat flux using Landsat TM/ETM+ data. International Journal of Remote Sensing, 34(4): 1459–1477. doi:  10.1080/01431161.2012.718465
    [49] Zhang Z X, Ren X, 2011. Causal relationships between energy consumption and economic growth. Energy Procedia, 5: 2065–2071. doi:  10.1016/j.egypro.2011.03.356
    [50] Zheng Yulan, Miao Shiguang, Zhang Qi et al., 2015. Improvements of building energy model and anthropogenic heat release from cooling system. Plateau Meteorology, 34(3): 786–796. (in Chinese)
    [51] Zhou Y Y, Weng Q H, Gurney K R et al., 2012. Estimation of the relationship between remotely sensed anthropogenic heat discharge and building energy use. ISPRS Journal of Photogrammetry and Remote Sensing, 67: 65–72. doi:  10.1016/j.isprsjprs.2011.10.007
  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Figures(7)  / Tables(5)

Article Metrics

Article views(42) PDF downloads(21) Cited by()

Proportional views
Related

Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China

doi: 10.1007/s11769-021-1203-y
Funds:  Under the auspices of National Natural Science Foundation of China (No. 41901219, 41671430, 41801326), Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (No. GML2019ZD0301)

Abstract: Spatiotemporal variations of anthropogenic heat flux (AHF) is reported to be associated with global warming. However, confined to the low spatial resolution of energy consumption statistical data, details of AHF was not well descripted. To obtain high spatial resolution data of AHF, Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light time-series product and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite monthly normalized difference vegetation index (NDVI) product were applied to construct the human settlement index. Based on the spatial regression relationship between human settlement index and energy consumption data. A 1-km resolution dataset of AHF of 12 selected cities in the eastern China was obtained. Ordinary least-squares (OLS) model was applied to detect the mechanism of spatial patterns of AHF. Results showed that industrial emission in selected cities of the eastern China was accountable for 63% of the total emission. AHF emission in megacities, such as Tianjin, Jinan, Qingdao, and Hangzhou, was most significant. AHF increasing speed in most areas in the chosen cities was quite low. High growth or extremely high growth of AHF were located in central downtown areas. In Beijing, Shanghai, Guangzhou, Jinan, Hangzhou, Changzhou, Zhaoqing, and Jiangmen, a single kernel of AHF was observed. Potential influencing factors showed that precipitation, temperature, elevation, normalized different vegetation index, gross domestic product, and urbanization level were positive with AHF. Overall, this investigation implied that urbanization level and economic development level might dominate the increasing of AHF and the spatial heterogeneousness of AHF. Higher urbanization level or economic development level resulted in high increasing speeds of AHF. These findings provide a novel way to reconstruct of AHF and scientific supports for energy management strategy development.

CAO Zheng, WEN Ya, SONG Song, HUNG Chak Ho, SUN Hui, 2021. Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China. Chinese Geographical Science, 31(3): 444−458 doi:  10.1007/s11769-021-1203-y
Citation: CAO Zheng, WEN Ya, SONG Song, HUNG Chak Ho, SUN Hui, 2021. Spatiotemporal Variations and Controls on Anthropogenic Heat Fluxes in 12 Selected Cities in the Eastern China. Chinese Geographical Science, 31(3): 444−458 doi:  10.1007/s11769-021-1203-y
    • The occurrence of global warming is indisputable, as evidenced by numerous observations and simulation studies (Patz et al., 2005; McMichael et al., 2006; Buchwitz et al., 2007; Lafferty, 2009; Chadburn et al., 2017; Panetta et al., 2018). As one of the major driving forces of global warming, anthropogenic heat flux (AHF) is believed to partially aggravate existing global warming (Best and Grimmond, 2016; Falasca et al., 2016; Holst et al., 2016). AHF refers to the energy emissions from human activities per unit area and time (Iamarino et al., 2012; Chen and Hu, 2017). As the AHF increases, several negative impacts can occur in the local ecological environment. For example, AHF contributes to more than 50% to the urban heat island effect, which has resulted in a temperature rise of more than 0.5℃ in Tokyo, Beijing, and Guangzhou, among other cities (Ichinose et al., 1999; Feng et al., 2012; Chen et al., 2014). Therefore, quantitatively assessing the environmental influence of AHF is important for providing scientifically supported suggestions for urban design and energy saving measures.

      To effectively assess the environmental influences of anthropogenic heat release (AHR), a series of AHF data are needed. In recent years, significant effort has been made to construct spatial distribution models of AHF (Pigeon et al., 2007; Zhou et al., 2012; Wong et al., 2015; Afshari et al., 2018). Four main approaches have been applied, namely: 1) inventory approaches, which gather energy consumption data and distribute them at a coarse resolution (Taha, 1997; Sailor, 2011); 2) methods based on the energy balance equation, in which the AHF is calculated based on the energy balance equation (Oke, 1982; Kato and Yamaguchi, 2005); 3) numerical simulations, which use data related to energy consumption, population, and traffic (Zheng et al., 2015; Wang Y C et al., 2019); and 4) statistical analysis methods, which develop linear or nonlinear relationships between AHF panel data and Defense Meteorological Program/Operational Linescan System (DMSP/OLS) data (Wang S S et al., 2019). However, the disadvantages of each method should be addressed. Inventory approach results rely on utility energy consumption data or energy consumption survey data. The resolution of results from this approach is very coarse, which is less than 1 km at city-wide or larger scales. Unlike that of the inventory method, the AHF data from the energy balance equation are indirect and of high resolution. The sum of latent, sensible, and stored thermal energy excluding net radiation is the AHF. This method is usually conducted with remote sensing technology. However, the remote sensing data were captured at one moment and the application of empirical parameters lead to uncertainty, so the representativeness of the results is questionable. In contrast to the above two methods, numerical simulation is a dynamic process. A building energy model was integrated with a meteorological model, such as the Town Energy Budget model and Weather Research and Forecasting model. However, a large amount of computing resources and other auxiliary data consumption was required during the numerical simulation. The statistical analysis method is easy to conduct; however, the resolution of the results relies on the original data. With the development of nighttime remote sensing data, high resolution AHF data can be obtained using the statistical analysis method. However, the results from different methods are inconsistent and not comparable. For example, the AHF for 2012 in Hong Kong of China ranged from 0 to 1000 W/m2 (Wong et al., 2015). A similar study showed that the AHF for 2013 in Hong Kong of China was less than 493 W/m2 (Dong et al, 2017). In addition to systematic errors, a single year analysis might account for this phenomenon. Therefore, multiyear analysis of AHF should be addressed. Above all, the advantages and disadvantages are both observed. Compared with the other three estimation methods, the statistical analysis method is easy to conduct the available of long-term nighttime satellite data and the absence of empirical parameter confirm the reliability of calculation results. Moreover, with the development of nighttime satellite resolution, statistical analysis method is capable to describe AHF more precisely.

      Most of the AHF analysis studies focus on the spatiotemporal variation characteristics of AHF. However, only a few efforts have been made to investigate the factors influencing AHF or to quantitatively assess their influence levels, and these tend to only focus on a single city. For example, Lindberg et al. (2013) determined that changing energy use and variability in air temperatures can result in large changes in the magnitude of AHF. In contrast, in Koralegedara’s experiment in Taiwan of China, urban land use and population density (Pop) influenced the spatial distribution of AHF, but not the air temperature and energy use (Koralegedara et al., 2016). Thus, the conclusions from different works differ. Other than local geographical characteristics, this might be caused by the lack of a comparative study for different developing cities. Therefore, two scientific questions were asked in this study. First, are the spatiotemporal variation characteristics of the AHF different in different developing cities? Second, do the domain influencing factors vary among different developing cities?

      The eastern China is the most developed area in China. The population and gross domestic product (GDP) account for 38.5% and 53.4% of China’s total, respectively (http://www.stats.gov.cn/tjsj/ndsj/). Rapid development has resulted in significant AHF emissions, thereby providing a significant experimental field for detecting the factors that influence AHF. Therefore, we addressed the above scientific questions by conducting a comparative investigation targeting 12 selected cities at different development levels. To illustrate the spatiotemporal variation and driving forces of anthropogenic heat flux, industrial, residential and transportation emission sections are included in this work. First, a spatial reconstruction model was developed to obtain gridded AHF data. Second, the spatiotemporal variation characteristics of different developing cities were determined. Third, the explanation levels of the chosen factors for the AHF values in different developing cities were examined using the ordinary least-squares (OLS) model. The findings from this work will help the city governors develop proper strategies to alleviate the influence of anthropogenic heat flux on climate change. Further, the goal to develop an eco-healthy city will be obtained.

    • We chose 12 cities in the eastern China as samples for developing the spatial distribution data of AHF (Fig. 1). To investigate the spatiotemporal variation in the AHF in different developing cities, 12 cities were grouped into three categories, namely super cities (Beijing, Shanghai, Guangzhou, and Shenzhen) with a population larger than 10 million, mega cities (Tianjin, Jinan, Qingdao, and Hangzhou) with a population ranging from 5 million to 10 million, and big cities (Changzhou, Jinhua, Zhaoqing, and Jiangmen) with a population of less than 5 million. The selection of the above 12 cities is based on three reasons: 1) the availability of data. AHF calculation is based on annual energy consumption data, multiyear energy consumption data of the selected 12 cities is available and easy to obtain. 2) the representativeness of chosen samples. All the chosen cities are located in the eastern China, which is the most urbanized and developed area in China. AHF emission in this area varied significantly since Reform and Opening Up. Investigation of AHF emission of the chosen cities can help governors in other areas, which are experiencing the similar developing process as the chosen cities, to construct scientific strategies to support the sustainable development. 3) different development levels. All the 12 cities have different development levels, which is the reflection of urbanization and economic development. Selecting these 12 cities can provide suggestion for cities and regions, which have similar development levels, to take advantages of AHF emissions as well as avoid the disadvantages of AHF emissions.

      Figure 1.  Location of the 12 selected cities in the eastern China

    • Data used in this study including three categories: energy consumption data, remote sensing data and socio-ecological data. Three major energy emission sources were considered, including industrial, residential, and transportations. Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light time-series product and Moderate Resolution Imaging Spectroradiometer (MODIS) satellite monthly normalized difference vegetation index (NDVI) product were major remote sensing data in this study (Su et al., 2012; Yu et al., 2015; 2018). Socio-ecological data was chosen as indicators to investigate its potential influence on anthropogenic heat flux. Seven socio-ecological data were chosen, including elevation, monthly mean temperature, monthly total precipitation, percentage of green space, gross domestic product (GDP), population density, and impervious surface (Price et al., 2000; Hutchinson et al., 2009). In analysis section, we applied 1 km × 1 km grid net to extract data information to conduct relative research in Section 4.4. Socioecological factors contain two parts: social economic factors and ecological environment factors. In this study, socio ecological factors are defined as the ecological or non-ecological factors that influence the amount, growth, spatial distribution of AHF. Selection of socio-ecological factors is based on one major criteria: the potential association between potential influencing factors and AHF. For example, annual mean temperature and mean annual precipitation are associated with the utility of air-conditioning. Urbanization level and population density are directly associated with energy consumptions, which is the cause of AHF. To validate the accuracy of our spatial reconstruction model, AHF data in 2012 obtained from Wang (Wang S S et al., 2020) was used to validate. The resolution of this data is 500 m. Details of data were illustrated in Table 1.

      Data nameDescriptionYearResolutionSource
      Energy Energy consumption data of industry, transportation and resident 2011–2014 City level http://www.stats.gov.cn/tjsj/ndsj/
      DMSP/OLS Nighttime light data (F18) 2010–2013 1 km http://www.ngdc.noaa.gov
      MODIS NDVI Monthly NDVI data 2010–2013 1 km http://www.gscloud.cn/
      Elevation DEM 2003 90 m http://www.gscloud.cn/
      Temperature Monthly mean temperature 2010–2013 1 km http://www.resdc.cn/
      Precipitation Monthly total precipitation 2010–2013 1 km http://www.resdc.cn/
      Percentage of Green Space Percentage of green space 2010 30 m http://www.globallandcover.com
      Population density Population density 2011–2014 100 m https://www.worldpop.org/
      Gross domestic product Gross domestic product 2011–2014 City level http://www.stats.gov.cn/tjsj/ndsj/
      Impervious surface Impervious surface 2015 30 m Liu et al., 2018
      AHF data in 2012 Anthropogenic heat flux emission 2012 500 m http://data.tpdc.ac.cn/zh-hans/
      Notes: DMSP/OLS is Defense Meteorological Satellite Program/Operational Linescan System, MODIS NDVI is Moderate Resolution Imaging Spectroradiometer (MODIS) satellite monthly normalized difference vegetation index (NDVI) product, DEM is digital elevation model, AHF is anthropogenic heat flux

      Table 1.  Information of data used in this study

    • Based on top-down analysis approach, AHF panel data between 2010 and 2013 were obtained. The detailed calculation procedures can be found in Cao’s work (Cao et al., 2019). The panel results of AHF ignored the spatial heterogeneousness within administrative unit. Therefore, to illustrate the spatial heterogeneousness of AHF, spatial reconstruction method was applied. Human Settlement Index (HSI) was applied (Peng et al., 2017) (Fig. 2). It is expressed as follows:

      Figure 2.  Flowchart of grid AHF deriving method. AHF is anthropogenic heat flux, DMSP/OLS is Defense Meteorological Satellite Program/Operational Linescan System (DMSP/OLS) nighttime light time-series product, MODIS NDVI is Moderate Resolution Imaging Spectroradiometer (MODIS) satellite monthly normalized difference vegetation index (NDVI) product, NTL is nighttime light time-series product, HSI is human settlement index

      $$ HSI = \frac{{\left({1 - NDV{I_{{\rm{max}}}}} \right) + NT{L_{{\rm{nor}}}}}}{{\left({1 - NT{L_{{\rm{nor}}}}} \right) + NDV{I_{{\rm{max}}}} + NT{L_{{\rm{nor}}}} \times NDV{I_{{\rm{max}}}}}} $$ (1)

      where NTLnor is the normalization results of nighttime light data, NDVImax denotes the maximum value of monthly NDVI data. Spatial reconstruction model was developed based on the linear regression relationship between HSI and AHF panel data.

    • The correlation coefficient (R) and root mean square error (RMSE) were applied to assess the accuracy of the spatial reconstruction model. These two indicators are calculated as follows:

      $$ R=\frac{\displaystyle\sum\nolimits_{i=1}^{m}({X}_{ip}-\overline {{X}_{p}})({X}_{ie}-\overline {{X}_{e}})}{\sqrt{\displaystyle\sum\nolimits_{i=1}^{m}{({X}_{ip}-\overline {{X}_{p}})}^{2}\displaystyle\sum\nolimits_{i=1}^{m}{({X}_{ie}-\overline {{X}_{e}})}^{2}}} $$ (2)

      where $ {x}_{ip} $ and $ \overline {{x}_{p}} $ are the AHF of grid i and the mean value obtained from previous research, respectively, $ {x}_{ie} $ and $ \overline {{x}_{e}} $ are the AHF of grid i and mean value of AHF in this research, respectively, and m is the count of the grid.

      $$ RMSE=\sqrt{\frac{\displaystyle\sum\nolimits_{i=1}^{m}{({X}_{ie}-{X}_{ip})}^{2}}{m}} $$ (3)
    • AHFs, which had similar spatial characteristics tent to be close. To detect different spatial patterns of AHF, Anselin Local Moran’s I was applied to investigate the spatial distribution of AHF in this study (Anselin, 1995). It was calculated as follows:

      $$ {I}_{i}=\frac{{x}_{i}-\overline {x}}{{s}^{2}}\sum\nolimits_{j}{c}_{ij}({x}_{j}-\overline {x}) $$ (4)

      where $ {x}_{i} $ is the AHF of grid i, $ {c}_{ij} $ represents the weights of the relationship between grid i and grid j, and $ \overline {x} $ and $ {s}^{2} $ represent the mean value and variance in the AHF value, respectively.

      Four categories of spatial clustering patterns were detected using Anselin Local Moran’s I: High-High, Low-Low, High-Low, and Low-High. In this study, High-High clustering pattern represented the hotspots of AHF, which was defined as the areas with AHF emission levels higher than the average. Low-Low clustering pattern represented the cold spots of AHF, which was defined as the areas with AHF emission level lower than the average and its surroundings. High-Low clustering pattern represented the islands of AHF, which was defined as the areas with an obvious high AHF emission whereas low AHF emission of its surrounding. Low-High clustering pattern represented the atoll of AHF which was defined as the areas with significantly low AHF emission level whereas relatively high AHF emission of its surroundings.

      Moran’s I was performed on ArcGIS 10.2 platform. Grids with 1-km resolution were built to extract the information of AHF. Then spatial pattern of AHF was measured using Local Moran’s I and four categories of spatial pattern results were obtained (Table 2).

      CategoryAbbreviationI valueAHF value of grid i and its neighbors
      High-High HH Positive Both larger than the average value
      Low-Low LL Positive Both smaller than the average value
      High-Low HL Negative AHF of grid i is larger than average, its neighbors are smaller than the average
      Low-High LH Negative AHF of grid i is smaller than average, its neighbors are larger than the average

      Table 2.  Classification of four spatial clustering patterns

    • A linear slope index was applied to evaluate the temporal trends of AHF between 2010 and 2013. This method is suggested to be effective and is calculated as follows (He et al., 2012):

      $$ Slope=\frac{n\times \displaystyle\sum\nolimits_{i=1}^{n}i\times {AHF}_{j}-\left(\displaystyle\sum\nolimits_{i=1}^{n}i\right)\left(\displaystyle\sum\nolimits_{i=1}^{n}{AHF}_{j}\right)}{n\times \displaystyle\sum\nolimits_{i=1}^{n}{i}^{2}-{\left(\displaystyle\sum\nolimits_{i=1}^{n}i\right)}^{2}} $$ (5)

      where $ {AHF}_{j} $ denotes the AHF value of grid j, n is the time span, and i is the time unit. The slope result was classified into five categories, namely stable, low growth, moderate growth, high growth, and extremely high growth. Stable is defined as areas where the slope is less than 0.3 W/(m2∙yr). Low growth is defined as the area where the slope is larger than 0.3 W/(m2∙yr) and less than the sum of the mean value and 0.5 times the standard deviation. Moderate growth is defined as the area where the slope is larger than the sum of the mean value and 0.5 times the standard deviation and less than the sum of the mean value and the standard deviation. High growth is defined as the area where the slope is larger than the sum of the mean value and the standard deviation and less than the sum of the mean value and 1.5 times the standard deviation. The remaining area is defined as extremely high growth (Peng et al., 2016).

    • We used the OLS model to investigate the potential spatial driving forces of AHF. OLS regression is the most common statistical analysis method. The assumption of the OLS model is that the error is independent. In this study, it is expressed as follows:

      $$ y = \alpha \times x + \delta $$ (6)

      where y is the column vector of the derived annual mean AHF value, x refers to the selected potential influencing factor, $ \alpha $ refers to the regression coefficients, and $ \delta $ is a vector of independent random errors. Moran’s I test was applied to determine whether there was spatial autocorrelation. If obvious spatial autocorrelation is observed in $ \delta $, this indicates that some key explanation factors are missing. Therefore, independent factors should be rearranged.

    • The AHR from different sources between 2010 and 2013 is shown in Figs. 3A3C. AHR mostly originated from industry, followed by residential usage and transportation. These three sources accounted for 62.5%, 23.1%, and 14.4% of the total AHR, respectively. The maximum emission from industrial sources was observed in Shanghai, with a mean value of 1.76 × 1018 J. The maximum emissions from both transportation and residential sources were observed in Beijing, with mean values of 5.2 × 1017 J and 3.9 × 1017 J, respectively. Minimum emissions from all three sources were observed in Zhaoqing, with mean values of 4.9 × 1015 J (residential emissions), 2.0 × 1016 J (transportation emissions), and 3.4 × 1016 J (industrial emissions).

      Figure 3.  Anthropogenic heat release from industrial (A), residential (B) and transportation sources (C) and total emissions in super cities (Beijing, Shanghai, Guangzhou, Shenzhen), mega cities (Tianjin, Qingdao, Jinan, Hangzhou), and big cities (Changzhou, Jinhua, Jiangmen, Zhaoqing) during 2010 and 2013 (D)

      We compared the total anthropogenic heat emissions in three city groups (super cities, mega cities, and big cities) (Fig. 3D). The total anthropogenic heat was mostly released from super cities, followed by mega cities and big cities. The total AHR from these three city groups accounted for 63.4% (super cities), 31.9% (mega cities), and 5.7% (big cities).

    • Mean value of R2 of spatial reconstruction model in 2010, 2011, 2012 and 2013 were 0.79, 0.82, 0.81 and 0.82, respectively. However, the performance of spatial reconstruction model should be further validate without only relying on R2. As the AHF released by human activities cannot be fully measured or monitored by the monitoring stations, the estimated results obtained in this study cannot be verified according to field station data. Therefore, conclusions from previous similar studies were applied as reference data for validating the performance of the model. In this research, we firstly validate the accuracy of AHF spatial distribution, then the amount of AHF emission was validated.

      The spatial distribution of the AHF in 2013 was verified using points of interest (POI), which were categorized into four major classes, namely living areas, commercial areas, transportation areas, and natural environment areas. We overlaid the AHF results and POI data and found that all high AHF areas were located near airports or bus stations (Fig. 4). Fig. 4 shows that airports, bus stations, and railway stations were all located in areas with high AHF. These are typically locations of high energy consumption and high human activities, which lead to high AHF emissions. Moreover, we selected regional landmarks, rivers, lakes and other different places to further validate the spatial distribution of AHF obtained from our research. The results showed that regional landmarks have high AHF emissions, such as Canton Tower with the AHF emission of 16.00 W/m2, The Oriental Pearl Radio & TV Tower with the AHF emission of 30.30 W/m2. While rivers or mountain areas have low AHF emissions. For example, Thousand-island Lake have an AHF emission of 1.49 W/m2. This result indicated the accuracy of the spatial reconstruction model to describe details of the spatial distribution of AHF.

      Figure 4.  Spatial distribution of the anthropogenic heat flux (AHF) model in selected cities of eastern China in 2013. High AHF emission locations including airports, railway stations etc.

      In addition, a Chinese AHF map of 2012 obtained from Wang was applied to calculate the correlation coefficient (R) and RMSE to examine the accuracy of the modeled AHF (Table 3) (Wang S S et al., 2020). All calculations of R were larger than 0.51 W/m2 and RMSE were all less than 2.78 W/m2. Thereby indicating the close spatial association between the modeled AHF and that from the previous study.

      IndexBeijingShanghaiGuangzhouShenzhenTianjinQingdaoJinanHangzhouChangzhouJinhuaZhaoqingJiangmen
      R0.56*0.61*0.67*0.60*0.51*0.52*0.57*0.66*0.52*0.61*0.55*0.51*
      RMSE / (W/m2)2.201.031.001.332.412.632.351.212.511.112.312.78
      Notes: R is the correlation coefficient, RMSE is the root mean square error, and * indicates P < 0.05

      Table 3.  Validation of the anthropogenic heat flux model in 2012

      To further test our modeling results, we compared the AHF value range with that from other related studies. AHF data obtained from Wang’s work was used as a validation data, which is proved to have accurate spatial distribution of AHF (Wang S S et al., 2020). Table 4 shows the peak value of AHF from previous research. The comparison showed that both our research and the previous studies implied high AHF emissions in super cities and megacities. The modeled AHF result of our research was all larger than that of Wang’s work (Wang S S et al., 2020). Differences between our study and the previous studies were largest in the megacities, which were ranging from 0.72 to 2.87 W/m2. While differences were smaller in the super cities and big cities, which were ranging from 0.22 to 2.17 W/m2. The utility of different approaches and different indexes may account for the differences between our results and that from the previous. In Wang’s work (Wang S S et al., 2020), normalized nighttime light adjusted urban index (VANUI) is used instead of HSI. HSI is built based on nighttime light remote sensing data, which is highly associated with human activities. Therefore, better relationship between HSI and AHF exists in central urban area, while this relationship may not be well observed in low human activity area with large green space cover. For example, differences between our research and Wang’s work are less obvious in super cities.

      CityWang et al. (2020b)Our research (2013)
      Beijing35.0837.04
      Shanghai50.8251.81
      Guangzhou29.8131.71
      Shenzhen31.8633.92
      Tianjin40.2141.16
      Qingdao41.6043.77
      Jinan30.9331.65
      Hangzhou49.7952.66
      Changzhou43.2644.90
      Jinhua37.9938.21
      Zhaoqing24.3824.87
      Jiangmen25.7826.01

      Table 4.  Comparison of the anthropogenic heat flux between our research and previous studies / (W/m2)

    • Table 5 reveals the areal percentage of different AHF growth between 2010 and 2013. Stable growth accounted for most of the different AHF growth types, which ranged from 44.2% of Shanghai to 92.4% of Zhaoqing. The changes in AHF in super cities became more evident, with a mean high growth and mean extremely high growth areal percentage of 5.83%, followed by that in mega cities and big cities. For single cities, the changes in the AHF in Beijing and Shanghai were the most evident.

      CityStableLowModerateHighExtremely high
      Beijing69.86.39.86.67.5
      Shanghai44.231.210.86.47.4
      Guangzhou74.910.65.42.96.2
      Shenzhen61.121.77.54.55.2
      Tianjin71.411.17.64.95.0
      Qingdao81.44.83.92.77.2
      Jinan86.20.33.32.77.5
      Hangzhou80.58.02.52.46.6
      Changzhou62.913.19.96.27.9
      Jinhua83.21.93.83.27.9
      Zhaoqing92.45.30.40.51.4
      Jiangmen89.74.01.91.03.4

      Table 5.  Areal percentage of each anthropogenic heat flux growth type between 2010 and 2013 / %

      Spatial characteristics of anthropogenic heat flux growth types revealed that most areas of the chosen cities have stable growth types. Extremely-high growth types were all located in the central downtown areas of chosen cities. High growth types were located around the extremely-high types. Low growth types were scattered within the cities. Therefore, Fast growth of AHF have circle layer characteristics, which indicated the AHF growth speed decreases from the central downtown to the rural area (Fig. 5).

      Figure 5.  Spatial distribution of different AHF growth types in chosen cities of easatern China during 2010 and 2013

    • Fig. 6 illustrates the spatial pattern of multiyear mean AHF in the chosen cities. Hot spots and cold spots were detected. Hot spots indicated high value clustering of AHF, while cold spots indicated low value clustering of AHF. The areal percentages of hot spots in the chosen cities were 29.9% (Shanghai), 25.9% (Changzhou), 23.5% (Beijing), 23.3% (Tianjin), 21.5% (Guangzhou), 19.6% (Shenzhen), 16.6% (Hangzhou), 16.2% (Jinan), 15.9% (Jinhua), 14.7% (Qingdao), 9.1% (Jiangmen), and 4.8% (Zhaoqing). The areal percentages of cold spots in the chosen cities were 48.7% (Guangzhou), 48.4% (Jiangmen), 48.2% (Beijing), 45.0% (Jinhua), 43.6% (Changzhou), 43.5% (Hangzhou), 43.5% (Tianjin), 42.4% (Jinan), 40.7% (Qingdao), 30.5% (Shenzhen), 28.9% (Shanghai), and16.7% (Zhaoqing).

      Figure 6.  Spatial pattern of multiyear mean anthropogenic heat flux (AHF) values in chosen cities in the eastern China

      According to the spatial pattern of AHF, two types of clustering patterns were categorized, namely: 1) single kernel and scattered satellite kernels and 2) double kernel and scattered satellite kernels. Beijing, Shanghai, Guangzhou, Jinan, Hangzhou, Changzhou, Zhaoqing, and Jiangmen had a single kernel located in the center of downtown and scattered satellite kernels located in urban fringes. Unlike the above cities, Shenzhen, Qingdao, and Jinhua had double kernels and scattered kernels.

    • Fig. 7 illustrates the potential relationships between socio-ecological factors and spatial distributions of AHF. As shown in Fig. 6, the R2 values were all above 0.55, which meant that the selected variables could explain more than 70% of the distribution of AHF. Moreover, the significance of the explanation varied in different developing cities. The distribution of AHF was better explained in mega cities and super cities, with mean explanation levels of 88.0% and 83.0%, respectively.

      Figure 7.  Coefficient of between AHF and potential influencing factors in super cities (A), megacities (B) and big cities (C) during 2010 and 2013. T, P, DEM, PG, Pop, GDP and Urb denotes to annual mean temperature, annual total precipitation, digital elevation model, percentage of green space, population density, gross domestic product, and urbanization level

      Different variables showed different relationships with the spatial distribution of AHF. The coefficient results showed that the spatial distribution of AHF was normally positively correlated with T, Pop, GDP, and the urbanization level (Urb), whereas it was negatively correlated with precipitation and the NDVI. However, in Shanghai, the spatial distribution of the AHF was negatively correlated with Pop and GDP. To explain this, we examined the energy consumption efficiency. According to an investigation conducted by Ma (Ma, 2011), Zhang and Ren (2011), energy consumption efficiency increases with developments in technology. For example, the development of electric car technology and exhaust purification technology lead to less consumption of gasoline and expanding the lifetime of gasoline resulting in less emission of AHF. Consequently, more advanced technology and reasonable industrial structures can lead to increases in green GDP with less AHF emissions. In Shanghai, such tertiary industries have accounted for more than 57.0% of the GDP increase since 2010. These types of industries require less energy, while resulting in increased GDP.

      Among the chosen factors, socioeconomic factors played more significant roles than natural ecological factors. For example, the mean values of the coefficients of temperature and precipitation were 4.1 and –3.6, respectively. The mean values of the coefficients of GDP and Urb were 5.4 and 10.0, respectively. Moreover, the roles of potential influencing factors in different developing cities varied geographically. Taking the GDP and Urb as examples, every change in GDP and Urb resulted in 3.5 (GDP; super cities), 10.1 (Urb; super cities), 9.9 (GDP; mega cities), 25.0 (Urb; mega cities), 6.8 (GDP; big cities), and 12.2 (Urb; big cities) unit increases in the AHF in super cities, mega cities, and big cities.

    • As one of the major driving forces of global warming, the contribution of AHF to global warming has been widely acknowledged (Ichinose et al., 1999; Zhang et al., 2016; Doan et al., 2019). However, owing to the lack of multiyear AHF data, stable modeling methods, and comparative studies, the investigations of its impact are inconsistent (Cao et al., 2019). Therefore, our study used a spatial reconstruction model to quantitatively assess the multiyear spatiotemporal variations among different developing cities. The potential influencing factors of the AHF were also detected for different developing cities. This study will also help to gain insights into the mechanisms of the potential relationships between socioecological factors and variations in AHF. The theoretical and management implications from this investigation will be useful for aiding urban planners and governments in constructing energy-saving and comfortable neighborhoods.

    • AHF spatial reconstruction methods have been extensively investigated and documented in many previous studies. Three sources of AHF were taken into consideration. A linear regression model using AHF panel data and DMSP/OLS methods was applied in our study. However, the modeled AHF results differed from different investigations. This might have been due to the index used to build the spatial reconstruction model, the spatial resolution of the results, and the period of analysis. Except for the HSI, the normalized nighttime light (NTL) index and NTL adjusted urban index (VANUI) were also applied to obtain the spatial distribution of AHF. The normalized nighttime light index showed better performance in Jiangsu and Zhejiang. The VANUI showed better performance in the middle Yellow River region. Compared with the above two indexes, the HSI can be used to develop an AHF spatial reconstruction model for the entire coastal region of East China with high-quality performance (Wang S S et al., 2019). Spatial resolution is another potential factor affecting the AHF reconstruction model. The results using data of resources and environment satellites were usually higher than those from the nighttime light data. The resolution of the data of resources and environmental satellites was higher than 60 m, while that of the nighttime light data was 500 m or 1 km. A higher resolution can characterize the land fraction more precisely. Consequently, the energy exchange is more quantitatively described (Kato and Yamaguchi, 2005; Zhang et al., 2013). However, confined to the quality and quantity, data from resources and environment satellites are not suitable for large-scale AHF reconstruction.

      Therefore, we used the nighttime light data to conduct our research. Three major results were obtained. First, industrial AHF dominated the three sources, which accounted for more than 60% of the total AHF. The amount of AHF in super cities was larger than that in mega cities and big cities. Second, the AHF in most of the chosen cities was stable, and only 1.9% to 14.0% of the AHF increased significantly. Third, two types of spatial clustering patterns were detected, namely single kernel and scattered satellite kernels and double kernels and scattered satellite kernels. One uncertainty should be illustrated. Slope of trend analysis is the result of linear regression, whose accuracy depends on the amounts of samples. However, confined to the complexity of calculation and availability of AHF panel data, only four years’ data were included. Some abnormal value might led to the uncertainty. Therefore, overestimations or underestimations might be occurred. Thus, to avoid this uncertainty, more samples of data should be involved to construct long time series AHF data.

      Theoretically, the pattern of AHF was associated with the population and Urb, especially in China. Over the last four decades, rapid economic development has resulted in increases in population and individual wealth. Conversely, continued increases in population and individual wealth have accelerated energy consumption and commodity marketing. Modern fuels were applied in vehicle and construction materials. Eventually, the AHF continually increased.

    • Introducing urban ecological theory into urban design and energy saving is a popular topic (Fry and Sarlöv-Herlin, 1997; Zhang et al., 2017; Wang J et al., 2020). It is widely recognized that the influence of the AHF can be mitigated via urban planning (Nie et al., 2014; Cao et al., 2019). Our research indicated that AHF mitigation strategies should consider not only the local climatic conditions, but also the vegetation cover status and economic development structures. These results provide useful insights and implications not only for ecological research workers, but also for urban planners and energy officials.

      We found that the NDVI was strongly negatively correlated with the AHF distribution, especially in highly urbanized areas. For example, each increase in the NDVI in Beijing led to a greater than 8.7 unit decrease in the AHF in Beijing. The evaporation and shadow effects of tree leaves accounted for the negative relationships. This suggests that urban planners can mitigate the AHF by planting trees. A weaker negative relationship between the DEM and AHF was also found in some of the chosen cities. Air movement between elevated mountains and/or hills formed local turbulence that promoted the migration of the AHF, thereby resulting in less influence. Therefore, hills or mountains should be retained during the urbanization process not only for their scenic value, but also for their mitigation value with respect to AHF.

      Understanding the most influencing socioecological factors on the increase in AHF is helpful for proposing solutions for AHF mitigation and developing administrative strategies. Both positive and negative correlations were found. In China, the rapid urbanization process resulted in large demands for construction materials and high population mobility. Consequently, the AHF from industrial and residential applications increased. However, no clear positive relationships between AHF and GDP or Pop were detected in Shanghai. High economic levels do not correspond to high AHF. We referred to the economic construction in Shanghai, and found that high-technology industries, such as financial industries and communications industries, accounted for most of the GDP. Previous studies also implied that high-technology industries are more energy efficient than traditional industries (Ma and Ma, 2011). Moreover, this finding is also related to the Greening Economy Project issued by the State Council of China. The aim of this project is to develop low-cost energy, highly efficient energy consumption, and sustainable economic patterns. Low AHF is one of the achievements of this project. Moreover, we found similar conclusions in Lu’s work (Lu et al., 2016). This author implied that AHF is related to the intensity of human activities. Developing renewable energy and clean energy technologies can improve the efficiency of industrial energy use and household appliances. Thus, energy conservation and emission reduction can be achieved.

    • Previous studies have suggested that variations in AHF largely depend on temporal and spatial scales (Wang et al., 2015). However, owing to a lack of precise panel data, only the annual mean AHF (not the seasonal or even daily AHF) was derived. In addition, the AHF spatial reconstruction results might be impacted by the spatial resolution of the original data and/or the construction indicator(s). Moreover, the AHF will increase significantly as the spatial resolution becomes finer (Zhou et al., 2012).

      Therefore, long-term time-series AHF data with finer temporal and spatial resolution are urgently needed. To avoid the disadvantages of the HSI, which cannot reflect the energy consumption from large factories or lights inside houses, an index developed from multiple remote sensors should be discussed.

    • The HSI was applied to develop a spatial reconstruction model for AHF. Moran’s I and the temporal trend analysis method were then used to characterize the spatiotemporal variations of AHF. The mechanisms behind the spatial distribution of AHF in the coastal area of East China were investigated. The results showed that industrial emissions from megacities dominated the total AHF. The increase in AHF mainly was due to population growth, especially in megacities. Different spatial patterns in the coastal area of East China were also detected.

      Overall, this study provides an efficient and accurate method to obtain spatial details of multiyear AHF in one of the most rapidly developing areas worldwide. Spatiotemporal variation characteristics of AHF in different developing cities varies dramatically. AHF discharged from megacities accounted for the most parts of the total AHF and also is the major contribution to the increasing of AHF. Socio-ecological factors plays different roles in spatiotemporal variations of AHF. In most of the chosen cities, temperature, population density, GDP and urbanization levels are positive with the AHF, while precipitation, DEM and NDVI are negative with the AHF.

Reference (51)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return