Overview of East Asian Air Pollutant Emission Grid Database (EAGrid2000)

Transcription

1 Overview of East Asian Air Pollutant Emission Grid Database (EAGrid2000) Akiyoshi Kannari(Freelance, former Institute of Behavioral Sciences) Yutaka Tonooka(Saitama University) Tsuyoshi Baba(Institute of Behavioral Sciences) Kentaro Murano(National Institute for Environmental Studies) This database was prepared as an outcome of developing the emissions inventory in the FY2002 through FY2004 research project Global Environment Research Fund, C-1 International Co- operative Survey to Clarify the Trans-boundary Air Pollution Across the Northern Hemisphere, (2) Research on the Elaboration and Verification of the Next-Generation Source Receptor Matrix. The following chapters provide an overview of the research and a description of the database. 1. Research Purpose EAGrid1995 is a database containing estimates of emissions per grid cell of air pollutants such as sulfur dioxide (SO 2 ), nitrogen oxides (NOx), non-methane volatile organic compounds (NMVOCs), and ammonia (NH 3 ) for China, Taiwan, Japan, the Republic of Korea (ROK), the Democratic People s Republic of Korea (DPRK), and Mongolia. This database was developed as part of a Global Environment Research Fund research project extending to FY2001, whose purpose was to enter emissions data into a long-range atmospheric transport model incorporating an atmospheric chemical reaction process meant to elucidate the mechanisms by which sulfates and nitrates are formed, transported, and removed. Subsequently, it became necessary to revise the basic database in light of changes such as economic development in China, and a need arose also to develop a database covering other air pollutants. In consideration of this situation, this research developed the 2000 emission inventory. In developing this inventory, we added the following substances to 1

2 those in EAGrid1995: carbon monoxide (CO), which is important as a tracer in the long-range transport process; particulates (here PM10), which are becoming increasingly important from the perspective of their impact on human health; and mercury (Hg), which is a particularly important heavy metal in terms of global pollution, and has especially high emissions in China. 2

3 2. China Region Source Inventory (1) Energy Statistics and Basic Activity Our research used the edition of China Energy Statistics, published in 2004, for the basic data on energy consumption. Statistics were taken from the province-specific energy supply and demand matrixes (production and consumption sectors energy types), but supplementary tables were also utilized for 3 biomass fuels (fuelwood, agricultural waste, and methane) as well as for conventional thermal power and hydropower. Additionally, the breakdown by industry type in the industrial sector was estimated from China s nationwide consumption by fuel type and from production in monetary values or quantity for each industry type and province. In particular, for the nonmetal ore products industry, data were recategorized for cement, glass, and other ceramics, while in the transport sector estimates were performed in sub-categories for each transportation mode to prepare a detailed province-specific energy matrix table that was used as the basic activity data for emission estimates. Supplementary estimates were performed for the Ningxia Autonomous Region, Tibet, and Hong Kong. The province-specific tables are likely to more accurately reflect actual consumption than the national tables. To obtain data on SO 2 emissions from industrial production processes, we used values reported in the China Environmental Yearbook as basic data. Evaporation sources from NMVOCs such as oil products, petrochemicals, and organic solvents were estimated from the production and handling volumes given as basic activity data in sources such as China Industrial and Economic Statistics Yearbook. (2) Estimation Method and Emission Factors a. SO 2 The basis for determining SO 2 is the percentage sulfur content in fuel, but because new information was obtained for coal sulfur content, this was used as a reference in reviewing data by province (as consumption areas) and setting new values.[1] The percentage remaining in ash was determined according to boiler size. 3

4 Our assumption on emissions from cement kilns was that product desulfurization is 85% in large kilns and dry rotary kilns, while in medium-sized, small, and shaft kilns the rate is 75%. We assumed that flue gas desulfurization (FGD) equipment is installed in large coal-fired boilers, and calculated emissions by subtracting the removal amount reported in the China Environmental Yearbook [2] from the amount generated as estimated from the assumed sulfur content. Estimating the post-desulfurization emissions by province and industry type necessitated preparing the desulfurization amount from combustion sources as province- and industry-specific matrices. Based on the combustion-related SO 2 removal amount in the Environmental Yearbook, a matrix was used to estimate tables for each industry and province by applying the Fratar method (a method which uses successive approximations to estimate values in a 2-dimensional table from peripheral values). Results were used as the flue gas desulfurization amount. The ratios of raw coal consumption by province and industry were given as the initial values of the Fratar estimation, and emissions were calculated under the assumption that coal-fired boilers (using raw coal as fuel) were equipped with FGD equipment. Emission and removal amounts from production processes for each industry and province are contained in the Environmental Yearbook. Because the statistical values for basic activity which are needed to independently estimate SO 2 emissions from production processes are unavailable, we decided to directly incorporate into estimates the emissions generated according to the Environmental Yearbook. Just as with the FGD amount, we employed matrix estimation and the Fratar method to find province- and industry-specific emissions. Ultimate atmospheric emissions were determined using these potentially generated amounts, desulfurization levels, and production-process emissions. b. NOx Emission factors were fully reviewed as part of this work. Data on large boilers were obtained by classifying measurements taken at Chinese thermal power plants according to boiler size and taking weighted averages using the size mix of power plant boilers. For stationary sources we referred to the EU s new emission factors.[3] For motor vehicles we added regional differences to NOx emission factors for each 4

5 province in consideration of regional differences in the phase-in of low-emission vehicles. Although NOx emissions from biomass fuels were not included in the 1995 data, they have been included in these estimates. c. NMVOCs Emission factors were reviewed as per those for NOx. In particular,new EU emission factors [3] were used for small stationary emission sources. We set the emission factors for combustion-derived NMVOCs under the assumption that China would have higher emission factors than the examples for small furnaces cited by the EU. As a result, biomass fuels were given emission factors that are considerably larger than the previous values. The EU s largest value for biomass fuel combustion NMVOC is 1700 g/gj (36 g/kg). Traditional furnaces in China s rural areas are made of bricks, plaster, and other materials with high heat capacity, so burning biomass fuels with low calorific value is not likely to raise the furnace wall temperature significantly. As such, one would expect that the interior furnace temperature is low, and that sometimes on rainy days or immediately afterwards furnaces are given fuel that is not sufficiently dry, which would be a factor leading to a decline in furnace temperature due to latent heat of evaporation. Due to this, Chinese furnaces are expected to have lower temperatures than fireplaces and small residential boilers and stoves in the EU, making them more liable to emit NMVOC emissions. For that reason we converted the largest value of the EU emission factors given, 1700 g/gj, to a calorific value of 4000 kcal/kg and an emission factor of g/kg for both fuelwood and agricultural waste used in rural Chinese homes. It has been confirmed by the Energy Research Institute in Beijing that these statistical values were given in dry weight terms and do not require moisture correction. Therefore moisture correction was not performed when estimating emissions. Just as with NOx, for motor vehicles we took into account regional differences in the phase-in of low-emission vehicles and added regional differences to emission factors for each province. d. PM10 The state of dust collection at large stationary sources was considered in the light of actual measurements made at Chinese power plant boilers, Japan s average value, and other information, while for 5

6 small stationary sources it was assumed from the EU s latest emission factors [3,4,5] and other information that China s emissions factors are higher than those of the EU. In setting emission factors for mobile sources, we took into consideration information such as the number of regulation-compliant vehicles in service while also referring to EU emission factors. e. Hg China consumes a considerable quantity of coal with a high Hg content, and this consumption is anticipated to increase. We therefore estimated Hg emissions from coal combustion. Based on the recent report by Wang [6] on the Hg content of Chinese coal by region and the atmospheric emission rates from combustion, we estimated the average Hg content by province (0.03 to 0.34 mg/kg), and assumed the atmospheric emission rate to be 64% to 78% depending on the type of combustion facility. (3) Estimation Results Table 1 brings together the fuel combustion-derived atmospheric emissions accounting for most sources with the exception of NH 3. SO 2 emissions occur mainly from energy conversion (thermal power plants) and coal combustion by the manufacturing industries. Although motor vehicles emit less than 1% of SO 2, they account for 7% of NOx emissions, which is comparatively large, but this is still low compared with the 39% NOx contribution of motor vehicles in Japan. Results show that rural biomass combustion produces high NMVOC emissions, another big difference in comparison to Japan. Hg emissions from coal combustion were estimated to be 185 Mg/y, which is similar with the total European 1995 stationary combustion of Hg emissions according to Pacyna and Pacyna.[7] 6

7 Table 1 Combustion-derived air pollutant emissions in China by sector (2000) ( 単位 :Gg/y, Hgは Mg/y) SO 2 NOx NMVOC PM 10 CO Hg Total 26,959 11,849 19,935 16, , Energy transformation 11,185 3, ,920 3, Final consumption 15,775 7,921 19,243 13, , Agriculture,forestry, fishery and farm , Industries 11,712 3,705 2,720 6,683 8, (1)Mining 1, ,298 7 (2)Manufacture 10,356 2,974 2,245 5,772 7, (3).Electricity, Gas, Water supply Construction ransportation,Storage,Postal,Telecommunictions 558 1,997 1, ,527 2 Transportation 327 1,650 1, ,034 0 (1)Rail Passenger (2)Rail Freight (3)Road Passenger (4)Road Freight ,117 0 (5)Domestic vessels (6)Domestic aviation Post, Telecommunication , Trade, Sales , Residential consumption 2, ,612 5,769 86, Urban ,705 6 Rural 1, ,067 5,372 80, Other consumption , ) Production-process emissions are included in SO 2 emissions 2)Stationary evaporative NMVOC emissions are follows; Fuel supply 302, Petro-chemical product 32, Solvent 1,318, Total 1,653Gg/y. Total amount including combustion emissions is 21,588Gg/y. 7

8 3. Refining the Japan Region Emission Inventory (1) Estimation Overview In order to clarify local domestic contributions to air pollution in simulations of trans-boundary air pollution arriving in Japan from other East Asian countries, it is desirable to refine the simulations by nesting a sub-grid covering Japan. In this inventory we improved and refined the estimation method used in EAGrid1995 in the following ways: Spatial temporal resolution improvement: The smallest spatial resolution for estimation is a 1-km grid, and units of time are by month and by time of day. More emission sources included: Open burning of agricultural waste, and machinery used in construction, industry, and agriculture were added. More air pollutants added: PM 10, CO, and Hg were added. Improved estimation method: The effects of air temperature, humidity, and catalyst deterioration were considered and cold-start emissions (additional emissions when engine is cold) were added to the estimation of motor vehicle emissions. (2) Estimation Methods Specific to Major Categories The following are the major differences with the previous EAGrid1995 estimation methodology. a. Motor Vehicles Motor vehicle emissions estimation in Japan is currently based on tailpipe emissions after engine warm-up, but more refined estimates are becoming possible thanks to research in the recent Japan Clean Air Program (JCAP). In these estimations we incorporated JCAP achievements by making a number of changes including: (1) modifying emission factors in view of the increased emissions due to catalyst deterioration in existing (gasoline) vehicles, (2) estimating cold-start emissions by taking into account the way vehicles are actually used, such as the distribution of engine start-up frequency and soak time (the length of time an 8

9 engine is stopped), (3) estimating emissions of evaporating NMVOCs such as diurnal breathing loss (DBL) due to gas tank respiration, and (4) incorporating new methods such as those for calculating estimates specific to seasons and regions, taking into account emission changes due to absolute temperature change (NOx) and gasoline-powered vehicle emission changes due to air temperature change (NOx, CO, NMVOC). To heighten the spatial resolution, we employed a method which uses detailed GIS data on main roads to distribute link-specific emissions on a 1-km grid. In terms of total emissions, NMVOCs, for example, increased substantially from 0.39 Tg/y at the time of the 1995 estimates to 0.47 Tg/y, but that increase is due to refinement of the estimation methodology. Additionally, due to differences in emission factors based on season and region, large seasonal and regional changes in emissions were estimated. b. Construction, Industrial, and Agricultural Machinery Basic data are not necessarily complete with regards to the coverage of emissions from construction, industrial, and agricultural machinery (off-road vehicles). However EAGrid2000 has estimated fleet size, utilization rates, average output, and other characteristics in consideration of application, machine type, size, and regulatory compliance, and from this then calculated emissions while applying the Ministry of Environment s emission factors. NOx is the pollutant which off-road vehicle emissions contribute most to total emissions, with the rate estimated to be about 8%. Of that, 56% is from construction machinery. The percentages of other pollutants for which off-road vehicles are responsible are all under 4%. c. Open Burning of Agricultural Waste Open burning of agricultural waste was not included in the previous work because of large uncertainties. However, because particulates were added this time, open burning was also added owing to its contribution to emissions, which is thought to be significant. We estimated the post-harvest amounts of residue and open burning for each crop, and then multiplied these by the appropriate emission factors. Although the Ministry of Agriculture, Forestry and Fisheries has investigated the amount of rice straw burned, there is not enough information on other crops, so the IPCC guidelines and other sources were used to set the amounts burned and emission factors. 9

10 d. Ships As before, the model estimates the in-port emissions of ships that have entered Japanese ports, but due to the heavy impacts on nearby cities of emissions from ships on oceangoing routes in semi-enclosed areas such as Tokyo Bay, Ise Bay, Osaka Bay, and the Seto Inland Sea, these emissions have also been estimated using origin-to-destination tables of inter-port freight and geographical information on routes. e. Hg Estimation Studies of coal-combustion facilities and waste incineration facilities by Yokoyama et al. [8] and Takaoka et al. [9] which investigated Hg emissions in Japan were used here to estimate average Hg content and atmospheric emission rates. Even when both these sources are combined emissions are only 9.5 Mg/y, a level far lower than Chinese emissions. (3) Estimation Results Table 2 shows the results of emission estimates by source. Basic data for emissions from large combustion sources were emissions of NOx, SO 2, and particulates according to a 1999 emissions survey and fuel consumption data. For pollutants other than NOx, SO 2, and particulates we applied average emission factors specific to furnace type and fuel type, and used annual corrections to arrive at FY2000 emissions. In the table, sources are categorized into power plants, waste incinerators, and other large combustion sources (mainly manufacturing facilities), which are further divided into 3 stack heights. The figures are in annual emissions. Some data which cannot be paired with emission grid cells are aggregated into a category called unknown cell facilities. Small combustion facilities signifies commercial facilities and residential combustion equipment. Other NMVOC evaporation sources corresponds to emissions from the use of organic solvents other than those in painting and printing, and from industrial processes. Among ammonia sources, those in the agriculture sector are emissions from livestock waste and the application of chemical fertilizers, while Other NH 3 sources signify emissions from fertilizer manufacturing processes and natural soil emissions. 10

11 The bottom row contains EAGrid1995 emissions. Although NOx and NMVOCs increased this time, caution is advised when comparing the EAGrid1995 figures with EAGrid2000 figures because the larger values are due to the enlargement of estimation sources and to differences caused by improvements in the estimation method. The change in SO 2 is due mainly to lower sulfur content in diesel fuel, while the lower figure for NH 3 emissions is due to the sector s lower emissions in agriculture. The largest air pollutant sources by pollutant type are motor vehicles (NOx, CO, PM10, PM 2.5 ), other large sources mainly in the industry sector (SO 2, CO 2 ), evaporation sources (NMVOCs), and agriculture (NH 3 ). Figure 1 summarizes change in emissions over time, which is the most distinctive feature of the estimation results. These graphs plot nationwide totals of hourly emissions as diurnal change by month, and different change characteristics are estimated depending on the pollutant. Reflected in these changes are the cold-start influence in motor vehicle emissions, the influence of autumn open burning, seasonal influences on NH 3 emissions, and others. Figure 2 shows an example of the emissions distribution over a 1-km grid. It is possible to use such detailed emissions data for a desired area in the Japan region. 11

12 Table 2 Emissions estimation results for the Japan region Source CO CO 2 NH 3 NMVOC NOx PM 10 PM 2.5 SO 2 Large combustion facilities Power plant: H<25m 778 9, , ,910 Power plant: 25<=H<100m 2,766 11, ,965 2,599 1,958 18,606 Power plant: H>=100m 9, ,700 2,543 3, ,737 5,886 4, ,511 Waste incineration: H<25m 25,392 5, ,265 2,409 1,286 12,675 Waste incineration: 25<=H<100m 157,001 27, ,426 39,187 7,112 4,211 16,587 Waste incineration: H>=100m 13,507 5, , ,191 Other large combustion: H<25m 263, , , ,371 16,905 11, ,408 Other large combustion: 25<=H<100m 292, , , ,157 18,882 13, ,011 Other large combustion: H>=100m 310, , , ,394 12,113 8, ,051 Unknown cell facilities 14,397 17, ,424 2,026 1,415 24,534 Small combustion facilities Small combustion 60, ,221 4,112 83,465 7,577 6,149 64,485 Small incineration 18,128 6,118 1,249 2,810 1, Open burning of crop residues 90,333 2,366 2,135 11,051 4,481 11,948 10, Moving sources Motor vehicle:exhaust 3,931, ,513 13, , ,682 64,224 54,591 25,811 Motor vehicle:evaporation 93,975 Motor vehicle:tyre and brake wear 25,000 8,269 Vessel 30,823 15,997 13, ,830 18,838 17, ,142 Aviation 16,737 3,625 5,076 20, Offroad vehicles 166,796 16,851 22, ,638 7,213 6,037 3,576 NMVOCevaporative sources Painting 782,673 Printing 183,000 Evaporation of fuels 235,118 Other evaporation 251,153 Ammonia sources Agriculture 285,661 Human and pets 101,278 Other NH 3 sources 37,037 Total 5,404,660 1,226, ,014 2,043,560 2,408, , , ,784 EAGrid ,372 1,947,107 2,101,550 1,023,202 H: stack height (Mgy -1 ) 12

13 Figure 1 Diurnal variations by month for the emissions from the whole Japanese region Nagoya Osaka Figure 2 Example of emissions spatial distribution on a fine grid system 13

14 4. Emissions of Nearby Countries EAGrid2000 takes the following simplified approach to the air pollutant emissions, excluding NH 3 and CO, of Taiwan, ROK, DPRK, and Mongolia. a. Taiwan Publicly released data were used because it is reasonable to assume that the emission estimates of the Taiwanese government are reliable.[10] b. ROK Dr. Park Il Soo of the Korean National Institute of Environmental Research supplied us with the ROK s FY2000 emission estimates. For PM10 and CO we used the values estimated in the joint LTP (Long-range Transboundary Pollutants) research project of Japan, China, and ROK.[11] c. DPRK and Mongolia These countries were given simplified treatment due to their low emissions. Data for SO 2, NOx, and NMVOCs were given an annual correction with ratios of energy consumption to that in 1995 (DPRK 0.81, Mongolia 0.95). NH 3 emissions were estimated by using FAO statistics on livestock and fertilizer consumption and the same methodology as in EAGrid1995. For Hg we estimated emissions from coal combustion assuming the standard emission factor. 14

15 5. EAGrid2000 Figures 3 and 4 show the grid distribution of annual emission flux according to the 2000 emission database (EAGrid2000), which was obtained with the estimation procedures described above. Table 3 presents a comparison of the emissions of various countries according to estimates for 2000 by EAGrid2000 and by Streets et al.[12] The most significant characteristic of these results is that China s SO 2 and NMVOC emissions are higher than in previous estimates. Evaluation of the results should be practiced by comparing atmospheric process simulations with actual measurements. Table 3 Comparison of EAGrid2000 estimates and previous estimates China Taiwan Japan ROK DPRK Mongolia China, Taiwan, Japan and ROK (Mg/y for Hg, Gg/y for others) SO 2 NOx NMVOC NH 3 PM 10 CO Hg 26,959 11,849 21,588 12,167 16, , ,385 11,347 17,432 13,570 10, , , , ,371 2, , ,198 1, , , ,322 1, , , , ,765 15,739 25,153 12,880 17, , ,391 15,388 21,023 14,246 11, ,506 - Uppers:This work, lowers:streets et al.(2003); from Iowa Univ. for PM 10 15

16 Figure 3 Distribution of annual emissions of various pollutants according to EAGrid2000 (1) 16

17 Figure 4 Distribution of annual emissions of various pollutants according to EAGrid2000 (2) 17

18 EAGrid2000 File Notes (1) The database provides annual emissions by country and grid cell. Because some cells straddle national borders, they must be totaled depending on the purpose. (2) The database of anthropogenic sources comprises files for each type of pollutant, and its fields are country name, substance name, source sector, geographical coordinates at grid cell center, and annual emissions (kg). (3) Emissions of all substances for China and Japan, and emissions of NH 3 for all countries are given for each source sector, but there is also an All Sources record. Data for other countries have only the All Sources record. (4) When preparing China s grid emissions there were a number of cities for which the grid cells could not be identified. Therefore totals do not match the national totals in Table 3 (see Table 4). (5) Assuming little change in time-dependent biomass amounts since EAGrid1995, we used the EAGrid1995 data for the monthly and day/night emissions of biogenic NMVOCs. (6) The high-resolution database for the Japan region was created separately for EAGrid2000-Japan. Table 4 gives anthropogenic emissions according to the database. However, there are a number of caveats which must be observed regarding source sector definitions. (1) Field burning refers to open burning of agricultural waste. For Chinese agricultural waste we have calculated only residential emissions for energy applications. (2) Emissions of off-road vehicles in China are included in industrial sectors. (3) Soil NH 3 emissions in Japan are included in data even though this is not an anthropogenic source. 18

19 Table 4 Emissions data from the EAGrid2000 grid database Pollutant(unit) Source Sector China DPRK Japan Mongolia ROK Taiwan Total NOx(Gg/y) All Sources 11, , , ,970 Energy Industry 3, ,079 Field Burning 4 4 Manufacture & Construction 3, ,462 Offroad Vehicle Other Stationary Sources 1, ,199 Residential Transport 1,983 1,298 3,280 SO 2 (Gg/y) All Sources 26, ,834 Energy Industry 11, ,258 Field Burning 1 1 Manufacture & Construction 11, ,258 Offroad Vehicle 4 4 Other Stationary Sources 1, ,169 Residential 2, ,257 Transport NMVOC(Gg/y) All Sources 21, , ,323 Chemical Product Energy Industry Field Burning Fugitive Fuel Manufacture & Construction 2, ,869 Offroad Vehicle Other Stationary Sources 1, ,277 Residential 13, ,615 Solvent Use 1,318 1,182 2,500 Transport 1, ,965 NH 3 (Gg/y) All Sources 11, ,527 Energy Industry 3 3 Fertilizer 5, ,646 Field Burning 2 2 Human & Pet Livestock 4, ,519 Manufacture & Construction 9 9 Other Stationary Sources 0 0 Soil Stationary Combustion Transport CO(Gg/y) All Sources 110,871 3,095 5, , ,703 Energy Industry 3, ,036 Field Burning Manufacture & Construction 9,574 1,062 10,636 Offroad Vehicle Other Stationary Sources 5, ,284 Residential 86, ,622 Transport 6,461 3,975 10,436 PM 10 (Gg/y) All Sources 16, ,675 Energy Industry 2, ,905 Field Burning Manufacture & Construction 6, ,801 Offroad Vehicle 7 7 Other Stationary Sources Residential 5, ,772 Transport Hg(Mg/y) All Sources Energy Industry Manufacture & Construction Other Stationary Sources Residential Transport

20 Acknowledgement This research was supported by the Japan Environment Ministry's Global Environment Research Fund (Subject C-1, International Co- operative Survey to Clarify the Trans-boundary Air Pollution Across the Northern Hemisphere, 2002~2004). References 1) China Clean Coal Technology, 2) China Environment Yearbook: China Environmental Science Press 3) EEA(2003) EMEP/CORINAIR Emission Inventory Guidebook, EMEPCORINAIR4/, and EEA(2004) Emission Inventory Guidebook, Small Combustion Installations, draft 4) Streets D.G., Shalini G., Waldhoff S.T., Wang Q,, Bond T.C., Bo Y. (2001) Black Carbon emission in China, Atms.Environ, 35, ) Bond, T. C., Streets, D. G., et al (2004)A technology- based global inventory of black and organic carbon emissions from combustion, J. Geophys. Res., 109, D14203, /2003 JD003697, ) Wang, Q. et al.(2000), Estimation of Mercury Emission from Coal Combustion in China, Environ. Sci. Technol., 34, ) Pacyna, E. G. and J. M. Pacyna(2002), Global emission of mercury from anthropogenic sources in 1995, Water, Air, and Soil Pollution 137, ) Yokoyama, T. et al.(2000), Mercury emissions from a coal-fired power plant in Japan, The Science of the Total Environment 259, ) Takaoka, M. et al.(2002), Control of mercury emissions from a municipal solid waste incinerator in Japan, J. Air & Waste Manade. Assoc., 52, ) Environmental Protection Administration Executive Yuan, R.O.C., Republic of China89, 25 years results Air Quality Control 20

21 11) The Secretariat of LTP Project(2004), The 7th Expert Meeting for the Long-range Transboundary Air Pollutants in Northeast Asia, October 28-30, ) Streets D.G., T.C. Bond, G.R. Carmichael, S.D. Fernandes, Q. Fu, D. He, Z. Klimont, S.M. Nelson, N.Y. Tsai, M.Q. Wang, J.-H. Woo, and K.F. Yarber(2003), An inventory of gaseous and primary aerosol emissions in Asia in the year 2000, J. Geophys. Res., 108(D22),