Article pubs.acs.org/est
High Resolution Carbon Dioxide Emission Gridded Data for China Derived from Point Sources Jinnan Wang,*,†,‡ Bofeng Cai,*,† Lixiao Zhang,§ Dong Cao,† Lancui Liu,† Ying Zhou,† Zhansheng Zhang,† and Wenbo Xue† †
The Center for Climate Change and Environmental Policy, Chinese Academy for Environmental Planning, 8 Dayangfang, Beiyuan Road, Chaoyang District, Beijing 100012, China ‡ State Key Lab for Environmental Planning and Policy Stimulation, Beijing 100012, China § State Key Joint Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China S Supporting Information *
ABSTRACT: A high spatial resolution carbon dioxide (CO2) emission map of China is proving to be essential for China’s carbon cycle research and carbon reduction strategies given the current low quality of CO2 emission data and the inconsistencies in data quality between different regions. Ten km resolution CO2 emission gridded data has been built up for China based on point emission sources and other supporting data. The predominance of emissions from industrial point sources (84% of total emissions) in China supports the use of bottom-up methodology. The resultant emission map is informative and proved to be more spatially accurate than the EDGAR data. Spatial distribution of CO2 emissions in China is highly unbalanced and has positive spatial autocorrelation. The spatial pattern is mainly influenced by key cities and key regions, i.e., the Jing-Jin-Ji region, the Yangtze River delta region, and the Pearl River delta region. The emission map indicated that the supervision of 1% of total land could enable the management of about 70% of emissions in China.
■
INTRODUCTION Spatial distribution of carbon dioxide (CO2) emissions at fine resolution is critically needed for carbon cycle research and carbon reduction strategies. Carbon cycle research and climate change modeling require more accurate and finely resolved maps of CO2 emissions.1−3 High resolution data sets promote research on the spatial analysis of CO2 emissions and on the driving forces of emissions, e.g., demography and industrial clustering. A high resolution emissions map is also a key requirement for the emerging flux inversion approach that utilizes remote sensing data from satellites, e.g., the Japanese Greenhouse Gases Observing SATellite (GOSAT).4 The gridded data can help policy makers to better prioritize the emission control area and facilitates the meticulous and accurate management of carbon emission and mitigation. Considering the vast territory but low quality of CO2 emission data,5 highly resolved CO2 emission maps mean much more for China, which became the largest emitter in the world in 2007.6 The incompleteness of energy statistics for © 2014 American Chemical Society
cities and smaller spatial units as well as inconsistencies in data quality between areas5 substantially impeded CO2 emissions inventory and thus the carbon mitigation and low carbon development of China. High spatial resolution data for carbon emission in China is not only essential for research but is also valuable for regional burden sharing of the national reduction target (e.g., 17% reduction of CO2 emissions per unit of GDP during the 12th Five-Year-Plan period, 2011−2015), the ongoing regional emission trading system, and the future merged national market. Extensive work has been conducted to develop CO2 emissions grids at different resolutions at the global, national, and regional level.1,4,7−12 Most of the previously constructed emissions grids were population-based or largely relied on Received: Revised: Accepted: Published: 7085
December 4, 2013 May 12, 2014 May 19, 2014 May 19, 2014 dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
production, lime production, iron and steel production, glass production, and ammonia production.
population gridded data as the main spatial proxy for the spatial distribution of CO2 emissions.7,13 However, population density may not be a good spatial proxy for CO2 emissions. For example, coal-fired power plants and cement factories are not always located in places with high population density. Nightlight data has recently been utilized to develop global, national, or regional CO2 emission maps.4,9,14,15 The nightlight data provide a global spatial distribution of the persistent lights on the earth’s surface. However, the light emissions from most of the large CO2 point emission sources, e.g., coal-fired power plants, are not as strong as light from urban commercial centers that are indirect CO2 emitters through electricity consumption. Therefore, an emissions map directly built bottom up by point emission sources will be much more accurate and reliable than maps compiled by top-down approaches based on spatial proxies. Point emission sources are valued most for the global 0.1° × 0.1° (10 km × 10 km) CO2 emissions in the EDGAR data set16 and in the U.S. Carbon Dioxide Information Analysis Center (CDIAC) data set17 as well as in the global 1 km emissions map by Oda and Maksyutov.4 The Vulcan project built up fine scale CO2 emission spatial distributions (10 km × 10 km) for the contiguous U.S. mainly with direct monitoring data of major point sources and supplementary spatial proxy data.1−3 Research on fine resolution CO2 emissions has advanced in Europe (EDGAR data set) and the U.S. (Vulcan) but lags behind in China. The work of CO2 emission mapping of China was first moved forward by Zhao et al.18−20 with the resolution of 0.25° × 0.25° (roughly 25 × 25 km). However, this work only used detailed location information on power plants as well as large cement and iron and steel plants for spatial distribution. Lack of available data in China is the major constraint for the bottom-up development of emission grids with fine resolution. The First China Pollution Source Census (FCPSC) provides a valuable opportunity for such work. The FCPSC surveyed and obtained information from 1.58 million industrial enterprises in 2007, including fuel consumption details at facility level as well as geographic coordinates (latitude and longitude). The survey scope of the FCPSC covers all the officially registered enterprises in China. It is more comprehensive than China’s national official statistics, which exclude small enterprises with annual business revenue less than 5 million CNY. There are very few enterprises (illegal enterprises) that exist outside of the official registration system. However, the primary energy consumption of these illegal enterprises is negligible. The most significant contribution of the FCPSC is that it is most likely the first China nationwide survey of all types of energy (energy category complied with the China Energy Statistical Yearbook) consumption, distinguished by combustion fuel and industrial feedstock, for facilities with spatial coordinate information. The main objective of this study is to compile a 10 km resolution CO2 emissions map via the point emissions sources from the FCPSC and to analyze the spatial characteristics of the CO2 emissions of China.
E=
∑ Mfuel × Ffuel + Ep
(1)
E is total CO2 emissions, Mfuel is energy use of a specific fuel, Ffuel is the CO2 emission factor for a specific fuel, and Ep represents CO2 emissions from industrial processes. The literature21 provides the emission factors for the GHG inventory in the Initial National Communication on Climate Change of China and contains detailed emission factors for different industries in different regions in terms of different energy types. Our emission factors were mainly from this literature. Data Sources. The activity data are organized and listed following the categories of the CO2 emissions inventory table in the Second National Communication on Climate Change of China.22 Data sources and their incoming resolution are shown in Table 1. Table 1. Data Sources and Incoming Resolutiona CO2 emissions sources energy activities
sectors industry
urban residential energy consumption transport agriculture/rural household energy consumption industrial processes
spatial resolution points county/ district provinces provinces points
data sources FCPSC data set FCPSC data set ref 23 ref 24 FCPSC data set
a
There were 31 provinces (excluding Taiwan, Hong Kong, and Macao) in China as of 2007, with an average area of 309,677 km2. There were 2,838 counties/districts (excluding counties/districts from Taiwan, Hong Kong, and Macao) in China as of 2007, with an average area of 3,383 km2.
The CO2 emissions (both from combustion of fuels and industrial processes) of each enterprise are calculated based on surveyed data according to eq 1. The CO2 emissions from industrial enterprises (point emission sources) accounted for 84% of total CO2 emissions, with 71% from energy combustion and 13% from industrial process. Four key energy-intensive sectors, i.e., thermal power plants, iron and steel production plants, cement production plants, and lime production plants, held 24,993 enterprises and contributed 56% to total CO2 emissions based on our accounting. These enterprises were large emission sources, and errors in their geographical coordinates will substantially influence the accuracy of spatially gridded data. In order to improve the accuracy of the point emissions data from these energy intensive sectors, we checked the accuracy of the position of enterprises by comparing their administrative properties with coordinate information. The administrative properties provide the county/district location for each of the enterprises. When the recorded coordinates of an enterprise conflict with their administrative properties, the coordinate of the geometric center of the recorded county/ district will be assigned as the geographical location for that enterprise. Roughly 94.3% of the key sector enterprises showed agreement between their geographical coordinates and administrative properties. The inaccurate coordinates were checked and corrected. Some large emitters were even verified and adjusted using the high resolution images of Google Earth.
■
METHODS Calculation of CO2 Emissions. CO2 emissions were calculated by summing the CO2 emissions from the combustion of fuels and from industrial processes, as shown in eq 1. Carbon emissions from land use, land-use change, and forestry are not considered in this paper. For CO2 emissions from the industrial processes, we took into account cement 7086
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
Figure 1. CO2 emissions of enterprises in the 4 key energy-intensive sectors. Four separate maps with high resolution can be found in the Supporting Information (Figure S1−Figure S4).
explicit by allocating the CO2 emissions in each county/district proportionally based on 10 km population grids. A population grid data set is valuable for the spatial distribution of CO2 emissions. The Data Sharing Infrastructure of Earth System Science of China has already produced a 1-km Chinese population grid in 2003 that was widely adopted for research work and proved to be reliable.25−28 To meet the requirements of our study, we updated the population grid with data from 2007 and up-scaled to 10-km resolution. (5) The transport CO2 emissions included emissions from road, railway, water, and aviation, so it could not be downscaled by proxies like transportation network. Therefore, we allocated the CO2 emissions of every province proportionally to the 10 km population grid cells in the corresponding province. (6) We summed the CO2 emissions from the industrial sector, agriculture/rural household sector, urban residential sector, and transport sector in every grid to complete the overall gridded CO2 emissions of China (Figure 2). The CO2 emissions from Taiwan, Hong Kong, and Macao are not included in our calculation.
The resultant CO2 emission distribution of plants in the 4 key energy-intensive sectors is shown in Figure 1. The transport energy use and CO2 emissions were directly adopted from our previous work.23 Spatial Mapping. CO2 emission gridded data were built in a stepwise manner (Figure 2). (1) We first created fishnet grids of China at 10-km resolution using the Krasovsky 1940 Albers Projected Coordinate System. (2) For the point emission sources in the industrial sector, we create dots based on their coordinates with values of CO2 emissions (from both energy consumption and industrial process) and sum the CO2 emissions of dots that were within each cell of the fishnet grids. (3) For CO2 emissions from agriculture/rural household energy consumption, the CO2 emissions were calculated at a province level based on the data from China Energy Statistical Yearbook and were then allocated evenly to grid cells in the corresponding provinces. (4) The urban residential energy consumption consisted of energy use from hotels, restaurants, hospitals, schools, and household energy use (heating and/or cooling and cooking) from the FCPSC. The residential energy consumption was surveyed at the county/district level. The urban household energy use was determined by sampling conducted in every town, and then the average level was multiplied by the population of the districts/counties. The CO2 emissions in urban residential sector were made spatially
■
RESULTS AND DISCUSSION Figure 3 shows the derived 10 km gridded CO2 emissions map. The aggregated CO2 emissions from fuel combustion is 6608 million tons of CO2, about 4.62% higher than the IEA’s 7087
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
Figure 2. Flowchart of the development of the 10 km gridded CO2 emissions of China in 2007.
estimation (6316 million tons)29 and 8.12% higher than CDIAC’s estimation (6112 million tons).17 The CO2 emissions from the cement production in our accounting are 510 million tons, which are only 75% of the estimation of CDIAC,17 but only 2.47% higher than the estimation (IPCC clinker-based) of Ke et al.30The possible reason for the estimation discrepancy between us and CDIAC may be that the CDIAC overestimated the clinker production of China in 2007. The overall spatial pattern of CO2 emissions in China could be roughly divided into two parts, east and west. The CO2 emissions of China increased from west to east. The emissions in the eastern region are obviously higher than that in the western region. Within the eastern region, the CO2 emissions in and around key cities, e.g., Shanghai, Beijing, Tianjin, Guangzhou, Zhengzhou, Chengdu, Chongqing, Wuhan, and Shenyang, were obviously higher than other regions. Zooming into relatively large scale, the high emission grids (red and purple colors in Figure 3) tend to be concentrated along the three main rivers of China, i.e., Yangtze River, Yellow River, and Pearl River. The main reason for this phenomenon is
that human settlements and cities frequently develop along rivers. Plains along rivers are the best place for human settlements and activities and are responsible for a large proportion of CO2 emissions. The Jing-Jin-Ji region (mostly along the Yellow River and the Haihe River), the Yangtze River delta region, and the Pearl River delta region (three key regions) were the emission hotspots in China due to intensive fossil fuel consumption because of high population density and high intensity of economic activities. Figure 4 shows the cumulative curve of the grid cells contribution to total emission. This clearly illustrates the high degree of emission clustering. Accumulated emissions of the top 10 grid cells accounted for 12% of total emissions, emissions of the top 100 grid cells accounted for 29% of total emissions, and emissions of the top 1000 grid cells accounted for 70% of total emissions. The emission map indicated that the supervision of 1% of total land could enable the management of about 70% of emissions in China. The central and local government should reallocate their monitoring and enforcement of emission reductions investment in national and provincial level and put more financial and 7088
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
Figure 3. China 10 km gridded CO2 emissions map for 2007.
Figure 4. Cumulative percentage of grid emissions accounting for total emissions. The grid cells were ranked first in terms of emission by descending order before the cumulative percent calculation.
mined the spatial pattern of the gridded CO2 emissions map. The industrial emissions are completely derived from the point source data and hence are comparatively accurate spatially. The spatial pattern of the energy combustion CO2 emissions from industrial enterprises can be found in Supporting Information
human resources on these top 1000 grid cells to make optimal results out of reduction effort. Energy combustion in industry accounted for 71% of the total CO2 emissions of China. The spatial pattern of CO2 emissions from industrial energy combustion largely deter7089
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
Figure 5. Comparison between 10 km gridded CO2 emissions (left) and EDGAR data (right).
Figure 6. Comparison between 10 km gridded CO2 emissions and EDGAR data in the Taklimakan desert: (a) MODIS image; (b) 10 km gridded CO2 emissions data; (c) EDGAR data. The color scheme of (a) and (b) are the same.
smoothly distributed than that of ours. The EDGAR data set, as it states, use point/line emission sources, population density grids, and land use data at various resolutions as spatial proxies to finish the gridded CO2 emissions data. Therefore, the gradient distribution of the EDGAR data is probably due to its using population density or other spatial proxies to allocate the majority of CO2 emissions. These operations might explain why the EDGAR map has large areas of barren land with relatively higher CO2 emissions than the same places of ours, which almost approach to zero in our map, and relatively lower values in the areas that correspond to our hotspots. A region of the Taklimakan desert in Xinjiang province is enlarged to show the significant difference between the EDGAR data and ours (Figure 6). The Taklimakan desert is the largest desert in China and is thus devoid of the widespread of anthropogenic CO2 emissions that the EDGAR data show. Our data agrees well with reality, with almost no emissions in most of this area. There are some high emission dots in (b) of Figure 6 that correspond to oil exploration (Tazhong operation site of the Tarim oilfield) in the center of this region and represent other human settlements in oases in the northernmost and southwest part of this region. However, these emission dots are indiscernible in the EDGAR data.
Figure S5. The pattern of energy related CO2 emissions matches to the pattern of CO2 emissions in Figure 3. The JingJin-Ji region, the Yangtze River delta region, and the Pearl River delta region accounted for large volumes of fossil fuel combustion due to a high density of energy intensive industrial plants. According to our spatial analysis, the grids with no less than 50 plants that emit more than 100,000 tons of CO2 dominated these three key regions mentioned above. The high density of energy intensive plants made the three key regions the hotspots of energy consumption as well as CO2 emissions. Comparison to Previous Gridded Data. This 10 km gridded CO2 emissions data can be compared to 0.1° gridded CO2 emissions data from the EDGAR database, which is a joint project of the European Commission Joint Research Centre and The Netherlands Environmental Assessment Agency. The EDGAR database provides the closest data publically available to our 10 km spatial resolution CO2 emissions data set of China. Extensive studies have been conducted based on the EDGAR database.8,10−12 Figure 5 shows the comparison between our 10 km gridded CO2 emissions data and the EDGAR data. The overall similarity of spatial pattern can be seen from the visual comparison. However, evident differences do exist. The EDGAR data are much more evenly and spatially 7090
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
can be calculated that the uncertainty of our total emission estimation is no more that 8% according to the error propagation equation in the IPCC Guidelines.32A more detailed discussion on the data quality and uncertainty analysis can be found in the Supporting Information. In addition, we have more confidence on the accuracy of the geodistribution of the gridded data because the geolocations of plants in 4 key energy-intensive sectors were checked systematically and manually. The spatial distribution of CO2 emissions from urban residential sector was developed at the county/district level with the support of population density grids. This processing method improved the spatial accuracy of the gridded data compared with the country-level or provincial level spatial distribution, since the average area of counties/ districts in eastern China is less than 1,500 km2 (15 grids). Eastern China is the place where the majority of CO2 is released. The spatial distribution of the transport sector and the agriculture/rural household sector had a lower spatial accuracy compared to other data. The transport tracks including road, railway, shipping line, and flight route spread widely across the country, and the track density may not be an appropriate proxy for transport energy activities and CO2 emissions. The agriculture/rural household sector had similar issues, with agriculture land covering a large proportion of almost every province. Nevertheless, CO2 emissions from these two sectors accounted for no more than 9% of China’s total CO2 emissions. The spatial distribution of CO2 emissions from these two sectors has a trivial effect on the national pattern of spatial distribution, not only because of its small share of total emissions but also because of their comparatively inherent homogeneous spatial pattern. Implications. The 10 km gridded CO2 emission map of China has wide-ranging implications for carbon cycle as well as carbon mitigation policies. It will be a useful representation of terrestrial anthropogenic carbon flux in China for climate change models and regional carbon budget research. It will facilitate the inversion of CO2 emissions using surface CO2 concentration data from monitoring instruments that were established by the Ministry of Environmental Protection in all 31 provincial capital cities of China in 2012. It can also be used to optimize the future site selection of the surface CO2 concentration monitoring. The high resolution CO2 emissions data for China are also significant in the context of the burden sharing of CO2 emissions reduction in China for the provinces and the joint control and mitigation of CO2 emissions by adjacent provinces or cities. It provides solid foundation for the comparison and benchmark setting between provinces. It also strongly supports the inventorying of CO2 emissions at the local level, especially for cities in China. Cities have great potential for CO2 emissions reduction and have been recognized as the major sources of reduction opportunities. China has started two batches of low-carbon pilot cities that include 36 cities, such as Beijing, Tianjin, Chongqing, and Shenzhen. Low quality energy statistics and a variety of data sources generate inconsistent and incomparable results in these cities. The gridded data build up a uniform data platform for cities’ CO2 emission estimation and comparison. In addition, CO2 emission gridded data could be helpful for differentiated policies in different regions based on carbon emission zoning. The carbon emission zoning should be performed by synthesizing the spatial distribution of CO2 emissions, population density, and affluence. The development
Based on statistical analysis and in situ investigation, large emitters such as energy intensive plants are concentrated in central areas and are not continuously distributed. This is the main reason for the spatial discontinuity in our CO2 emissions map. This means that the spatial discernibility of our data is higher than the EDGAR data under almost the same spatial resolution. The smooth spatial distribution of CO2 emissions in the EDGAR map blurred the predominant significance of the hotspots. Spatial Characteristics. Spatial autocorrelation analysis was used to measure the spatial characteristics of CO2 emissions in China, and the Global Moran’s I and Local Moran’s I were used as the indicators for spatial autocorrelation. The Global and Local Moran’s I measure spatial autocorrelation based on both feature locations and feature values simultaneously. They evaluate whether the pattern expressed is clustered, dispersed, or random. The Z score (standard deviations) indicates statistical significance. A Moran’s I value near +1.0 indicates clustering, while a value near −1.0 indicates dispersion. The spatial autocorrelation analysis showed that the global Moran’s I is approximately 0.028 (Z value = 40.28, P < 0.01), which indicates that the spatial pattern of China’s CO2 emissions in 10 km resolution has positive spatial autocorrelation. The CO2 emissions were spatially clustered rather than randomly distributed. Based on the local Moran index calculation (significance test and the spatial map of the local Moran index are provided in the Supporting Information Figure S6), significant positive spatial autocorrelation of CO2 emissions exist in some regions of China. Clustering is most pronounced in the three key regions. This means that some hotspot areas have substantial influence on surrounding areas and that their high economic activity and energy consumption spur peripheral CO2 emissions. Almost all of the hotspot areas were located around key cities. The negative spatial autocorrelation with significant level around some emission hotspots further confirms our analysis regarding the spatial distribution of the energy intensive plants, which is clustered locally and discontinuously distributed at larger scales. Caveats. The FCPSC was jointly conducted and compiled by the Ministry of Environmental Protection, the National Bureau of Statistics, and the Ministry of Agriculture and was led by the vice-premier of China. Like the National Population Census and National Economic Census, it is a national census with strong law enforcement. This ensured to a large extent that the individual enterprise would submit their real information as the decree of the State Council of the People’s Republic of China (No. 508) required. The FCPSC has established strict and standard procedures in data collection, field investigation, data analysis, and data quality assurance/ quality control. As our activity data, the uncertainty of the FCPSC data is no more than 6%, according to the results of a two tiered data quality check.31 Another uncertainty is associated with the emission factors we adopted from the available literature.21 Emission factors for energy combustion were derived from the official emission factors21 for the national inventory of China 1994. However, the changes in the emission factors for energy combustion were not significant. The emission factors we used for the industrial process emissions were much more direct because we calculated the emissions directly from the chemical reaction process. The uncertainty of our activity data and emission factors is no more than 6% and 5% (suggested by ref 21), respectively; it 7091
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
emission map could be shared on request within the GHG inventory and carbon cycle research community.
of a carbon market should also start in high emission zones supplemented by low emission zones as offset regions and gradually moved to national market. The high resolution gridded data also supplement the formation of China’s urban development strategies, especially its spatial layout, in key regions and nationwide. The spatial pattern of China’s CO2 emissions is mainly influenced by three key regions of China. Therefore, policy priority should be paid to these regions. In fact, air pollution and carbon emissions are two serious environmental problems faced by China. The CO2 emissions and SO2 emissions largely come from combustion of fossil fuels and follow the simultaneous, colocated, and co-originated mechanism. Consequently, integrating addressing climate change into environmental management is the inevitable and optimal pathway for green development in China. The three key regions have already been identified as the focuses of joint prevention and control of air pollution in Air Pollution Prevention Plan in Key Regions (2011−2015) issued by the Ministry of Environmental Protection in 2012. Therefore, it is proposed that China could consider comanagement of CO2 and SO2 at these key regions and carefully consider the spatial adjustment of the distribution of carbon emission intensive industry. Based on the FCPSC data set, 10 km resolution China CO2 emission gridded data is built from the bottom up. The resultant emission map has proved to be more accurate and spatially discernible when compared with the EDGAR data. Spatial distribution of CO2 emissions in China is highly unbalanced and is mainly influenced by key cities and the three key regions of China. China’s CO2 emissions were spatially agglomerated by the hotspot cities. Provincial administrative boundaries have no significant effect on the spatial distribution of China’s CO2 emissions. The high spatial resolution CO2 emissions map is of great importance both for scientific research and policy formation. Future research should update the point emission sources to gradually build up a time series of emissions maps, which will provide greater support for spatial research and policy implementation.
■
■
(1) Gurney, K. R.; Mendoza, D. L.; Zhou, Y.; Fischer, M. L.; Miller, C. C.; Geethakumar, S.; de la Rue du Can, S. High resolution fossil fuel combustion CO2 emission fluxes for the United States. Environ. Sci. Technol. 2009, 43 (14), 5535−5541. (2) Zhou, Y.; Gurney, K. A new methodology for quantifying on-site residential and commercial fossil fuel CO2 emissions at the building spatial scale and hourly time scale. Carbon Manage. 2010, 1 (1), 45− 56. (3) Gurney, K. R. Vulcan Science Methods Documentation, Version 2.0; Arizona State University/Purdue University: Lafayette, IN, 2011. (4) Oda, T.; Maksyutov, S. A very high-resolution global fossil fuel CO2 emission inventory derived using a point source database and satellite observations of nighttime lights, 1980−2007. Atmos. Chem. Phys. Discuss. 2010, 10, 16307−16344. (5) Guan, D.; Liu, Z.; Geng, Y.; Lindner, S.; Hubacek, K. The gigatonne gap in China’s carbon dioxide inventories. Nat. Clim. Change 2012, 2, 672−675. (6) IEA, CO2 Emissions from Fuel Combustion 2010; 2010. (7) Andres, R. J.; Marland, G.; Fung, I.; Matthews, E. A 1°×1° distribution of carbon dioxide emissions from fossil fuel consumption and cement manufacture, 1950−1990. Global Biogeochem. Cycles 1996, 10 (3), 419−429. (8) Marland, G.; Brenkert, A.; Olivier, J. CO2 from fossil fuel burning: a comparison of ORNL and EDGAR estimates of national emissions. Environ. Sci. Policy 1999, 2 (3), 265−273. (9) Doll, C. N. H.; Muller, J. P.; Elvidge, C. D. Night-time imagery as a tool for global mapping of socioeconomic parameters and greenhouse gas emissions. AMBIO 2000, 29 (3), 157−162. (10) Ohara, T.; Akimoto, H.; Kurokawa, J.; Horii, N.; Yamaji, K.; Yan, X.; Hayasaka, T. An Asian emission inventory of anthropogenic emission sources for the period 1980−2020. Atmos. Chem. Phys. Discuss. 2007, 7 (3), 6843−6902. (11) Parrish, D. D.; Kuster, W. C.; Shao, M.; Yokouchi, Y.; Kondo, Y.; Goldan, P. D.; de Gouw, J. A.; Koike, M.; Shirai, T. Comparison of air pollutant emissions among mega-cities. Atmos. Environ. 2009, 43 (40), 6435−6441. (12) Alonso, M. F.; Longo, K. M.; Freitas, S. R.; Mello da Fonseca, R.; Marécal, V.; Pirre, M.; Klenner, L. G. An urban emissions inventory for South America and its application in numerical modeling of atmospheric chemical composition at local and regional scales. Atmos. Environ. 2010, 44 (39), 5072−5083. (13) Olivier, J. G.; Van Aardenne, J. A.; Dentener, F. J.; Pagliari, V.; Ganzeveld, L. N.; Peters, J. A. Recent trends in global greenhouse gas emissions: regional trends 1970−2000 and spatial distributionof key sources in 2000. Environ. Sci. 2005, 2 (2−3), 81−99. (14) Rayner, P. J.; Raupach, M. R.; Paget, M.; Peylin, P.; Koffi, E. A new global gridded data set of CO2 emissions from fossil fuel combustion: Methodology and evaluation. J. Geophys. Res. 2010, 115 (D19306), 1−11. (15) Raupach, M.; Rayner, P.; Paget, M. Regional variations in spatial structure of nightlights, population density and fossil-fuel CO2 emissions. Energy Policy 2010, 38 (9), 4756−4764. (16) European Commission, Joint Research Centre (JRC), Netherlands Environmental Assessment Agency (PBL), Emission Database for Global Atmospheric Research (EDGAR). 2012. (17) Carbon Dioxide Information Analysis Center (CDIAC), Gridded Annual Estimates datasets, 2013. http://cdiac.esd.ornl.gov/ trends/emis/meth_reg.html (accessed Mar 2, 2014). (18) Zhao, Y.; Wang, S.; Duan, L.; Lei, Y.; Cao, P.; Hao, J. Primary air pollutant emissions of coal-fired power plants in China: Current status and future prediction. Atmos. Environ. 2008, 42 (36), 8442− 8452. (19) Zhao, Y.; Nielsen, C. P.; McElroy, M. B.; Zhang, L.; Zhang, J. CO emissions in China: Uncertainties and implications of improved
ASSOCIATED CONTENT
S Supporting Information *
A more detailed elaboration of the data sets and uncertainty analysis. This material is available free of charge via the Internet at http://pubs.acs.org.
■
REFERENCES
AUTHOR INFORMATION
Corresponding Authors
*Phone: 8610-84910681. Fax: 8610-84918581. E-mail:
[email protected]. *Phone: 8610-84947736-662. Fax: 8610-84947786. E-mail:
[email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was funded by the Project Study on Key Issues of China City Carbon Emission Inventory (No. 41101500). We sincerely thank the invaluable comments from the three anonymous reviewers. The China country boundary data is from National Administration of Surveying, Mapping and Geoinformation of China (http://www.sbsm.gov.cn). Our 7092
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
Environmental Science & Technology
Article
energy efficiency and emission control. Atmos. Environ. 2012, 49 (3), 103−113. (20) Zhao, Y.; Nielsen, C. P.; McElroy, M. B. China’s CO2 emissions estimated from the bottom up: Recent trends, spatial distributions, and quantification of uncertainties. Atmos. Environ. 2012, 59 (3), 214−223. (21) (NCCCC/NDRC) National Coordination Committee on Climate Change/National Development and Reform Commission, Energy Research Institute. China Greenhouse gas inventory Study; China Environmental Science Press: Beijing. 2007. (22) National Development and Reform Commission. The People’s Republic of China Second National Communication on Climate Change. 2012. (23) Cai, B.; Yang, W.; Cao, D.; Liu, L.; Zhou, Y.; Zhang, Z. Estimates of China’s national and regional transport sector CO2 emissions in 2007. Energy Policy 2012, 41, 474−483. (24) Department of Energy Statistics of National Bureau of Statistics. China Energy Statistical Yearbook 2008; China Statistics Press: Beijing, 2008. (25) Jiang, D.; Yang, X.; Wang, N.; Liu, H. Study on spatial distribution of population based on remote sensing and GIS. Adv. Earth Sci. 2002, 17, 734−738. (26) Wang, X.; Li, X.; Ma, M. Advance and Case Analysis in Population Spatial Distribution Based on Remote Sensing and GIS. Remote Sens. Technol. Appl. 2004, 19, 320−327. (27) Liu, H.; Jiang, D.; Yang, X.; Luo, C. Spatialization approach to 1 km grid GDP supported by remote sensing. Geo-information Sci. 2005, 7, 120−123. (28) Li, S.; Zhuang, D. A review on RS- and GIS-based population estimation methods. Prog. Geogr. 2006, 25, 109−121. (29) IEA, CO2 Emissions from Fuel Combustion 2013; 2013. (30) Ke, J.; McNeil, M.; Price, L.; Khanna, N. Z.; Zhou, N. Estimation of CO2 emissions from China’s cement production: Methodologies and uncertainties. Energy Policy 2013, 57, 172−181. (31) Editorial committee on First China Pollution Source Census. First China Pollution Source Census technical report; China Environmental Science Press: Beijing, 2011. (32) IPCC. 2006 IPCC Guidelines for National Greenhouse Gas Inventories; IGES: Japan, 2006.
7093
dx.doi.org/10.1021/es405369r | Environ. Sci. Technol. 2014, 48, 7085−7093
