Decision and Information Sciences Division

Argonne National Laboratory

Argonne, IL 60439, U.S.A.

Department of Environmental Engineering

Tsinghua University

Beijing 100084, P.R.C.

DRAFT FINAL REPORT

September 10, 1998

Report for the China-MAP Project

Sponsored by the National Aeronautics and Space Administration

through a Contract with the Georgia Institute of Technology

______________________________

where E represents the energy, in PJ, and r is the annual rate of growth. By manipulating this function algebraically, we find that:

E₁₉₉₅ = (E₁₉₉₀ * E₂₀₀₀)^_

These regional estimates were used to distribute the national total found by combining IEA trend data with RAINS-Asia data from 1990.

Data for biofuel consumption were projected from regional data for 1990 (Streets and Waldhoff, 1998), per capita consumption trend data from previous years (Lu, 1993; Sinton, 1996), and total Chinese population (UN, 1995). Per capita combustion of biofuels in 1995 was determined using the formula:

B₁₉₉₅ = B₁₉₉₀ (1 + r)⁵

where B is biofuel use, per capita, and r is the annual rate of growth, which in this case is negative. It was found that while the per capita rate of biofuel combustion has declined since 1990, the population has grown at a faster rate, causing absolute levels of biofuel use to rise.

Because CO is a product of incomplete combustion, emission factors are dependent on burning conditions and have the potential to vary widely. Many sources were consulted for sector- and fuel-specific emission factors, and those which seemed to most accurately represent emissions in developing nations such as China were chosen. Emission factors for the domestic sector were variable among (and within) source types (Table 1). It was decided that emission factors for domestic fossil-fuel combustion presented in AP-42 (EPA, 1995) represented Chinese emission factors fairly well. For domestic biofuel combustion, however, emission factors that had been determined experimentally (Smith, 1988) in conditions similar to actual combustion conditions in China, were determined to be the best available. These emission factors were then applied to the energy data to yield CO emissions.

Regional fossil-fuel combustion in industrial boilers was calculated for 1995 by combining IEA trend data with RAINS-Asia regional energy data as described above. Total coal and oil consumption were approximately 15,000 PJ and 2,800 PJ, respectively. Emission factors from AP-42 (EPA, 1995) were combined with information about energy values of average Chinese coal (Sinton, 1996) to produce China-specific emission factors. CO emissions from coal dominate this category because in China coal is more readily available than fuel oil and CO emission factors are nearly eight-times higher for coal than for oil.

In 1995 China was the second largest steel producer in the world. Regional production data were obtained from several sources (China Statistical Yearbook, 1996; UNEP, 1997). Data were divided into large point sources (LPS), defined as steel plants with production greater than 250 thousand tonnes per year, and area sources, representing the remaining steel produced by small facilities (<250 kt/yr). The precise location of each LPS steel plant was determined (see Table 2) and area sources were allocated according to regional population.

Iron and steel manufacturing consumes nearly as much carbon as iron on a weight basis, and much more in terms of moles. CO emissions result in part from incomplete combustion, but the largest source is from CO used in materials processing. In industrialized nations, most of this CO is recycled as fuel or as process gas or flared to complete the combustion process. In China, however, this is often not the case. Siddiqi et al. report that 8-9% of blast-furnace gas is released to the atmosphere, so CO emission factors from this source are considerably higher than in industrialized countries (EPA, 1995; UNEP, 1997; Siddiqi et al., 1994).

CO emission factors were calculated for several individual processes in iron and steel manufacturing. Pig iron production requires four main processing components: coke plant, pellet plant, sinter plant, and blast furnace (see Table 3). Emission factors for each step were calculated (EPA, 1995; UNEP, 1997) to determine a combined CO emission factor for pig iron production of 0.045 t-CO * t-pig iron^-1.

Steel is manufactured using two primary methods, the basic oxygen furnace or the electric arc furnace. The electric arc furnace is a newer, cleaner technology (see Table 3), but is not yet in wide use in China—only 6% of steel is produced with this method. The basic oxygen furnace emits more than four times as much CO per tonne of steel produced than the electric arc furnace (0.030 t-CO * t-steel^-1 compared to 0.007 t-CO * t-steel^-1). If all Chinese steel were produced using the electric arc furnace (94% of current production switched from BOF to EAF), CO emissions would be less by 2 million tonnes per year. The appropriate emission factors for Chinese processes were applied to point-source and regional pig iron and steel production estimates for 1995 (China Statistical Yearbook, 1996; UNEP, 1997) to yield CO emission estimates.

In order to estimate the amount of crop residue that is burned as a means of disposal each year in China, several pieces of information were combined. First, the 1995 regional production of rice, wheat, and corn were determined based on provincial data (USDA, 1998). These regional estimates were then multiplied by crop-specific, residue-to-production ratios (Lu, 1993). Although it is believed that rice stalks are the major component of residue burned, it was impossible to determine the exact proportions of specific crop residues that were burned, so a flat rate of 23% (Crutzen and Andreae, 1990) was applied to all grain residues. There is a wide range of CO emission factors between residue types. Both the burning conditions and the types of residue affect combustion conditions. AP-42 (EPA, 1995) has CO emission factors for many specific residues, but because of uncertainty in the proportions of crop residues burned in China, an emission factor for open burning of "unspecified" crop residues was used here.

Transportation is a major source of CO emissions in China, primarily from the use of gasoline and diesel vehicles for highway passenger and freight transportation. Trains, airplanes, and ships are much smaller sources. The number of vehicles in China is growing rapidly, and by 1995 it had reached 10.4 million (China Statistical Yearbook, 1996). These represented about 1.8 million cars, 4.6 million light-duty vehicles, and 4.1 million heavy-duty vehicles. In addition, there were about 10 million motorcycles (Tsinghua University, 1997).

Although the number of highway vehicles in China is small relative to western countries, most vehicles on the road are of domestic construction, inefficient and high polluting, having CO emission levels typically 5-10 times higher than the international level. Laboratory testing (Tsinghua University, 1997) has determined that the average CO emission factor for automobiles is 44 g/km and for motorcycles, 165 g/km. One reason for the high emission factors is the low average speed of vehicles on Chinese roads (23 km/hr).

The Tsinghua University report for the World Bank estimates that the total CO emissions from highway vehicles in 1995 were 23 million tonnes, apportioned as follows: automobiles 4 million tonnes, light-duty vehicles 7 million tonnes, heavy-duty vehicles 10 million tonnes, and motorcycles 2 million tonnes. These values have been adopted as controlling estimates at the national level.

This national highway-vehicle estimate was apportioned into three categories—cars, trucks, and motorcycles—using information from the Tsinghua University study and national statistics on vehicle ownership. No statistical information is available on regional emissions, so surrogate measures were developed to disaggregate the national totals. Provincial statistics on car and truck registrations (China Statistical Yearbook, 1996) were used to allocate national emission estimates to RAINS-Asia regions. Population estimates were used to separate urban and non-urban vehicle usage. The registration metric proved more robust than other tested metrics, such as total passenger miles traveled and freight tonnage shipped, and was better able to reproduce emissions for Beijing and Guangzhou discussed in the Tsinghua University paper. Unfortunately, the most useful parameter, vehicle miles traveled, is unavailable. Emissions from motorcycles were allocated to regions according to a method that took into account preferential use in (a) urban areas, (b) more affluent areas, and (c) the southern parts of China that have a conducive climate.

The resulting regional estimates somewhat underestimate emissions in urban areas. Our allocation method yields an estimate of 1.0 million tonnes of CO for Beijing—compared with 1.4 million tonnes in the Tsinghua University study—and 300,000 tonnes for Guangzhou—as compared with 400,000 tonnes. These differences can probably be attributed to a combination of lower driving speeds and higher vehicle usage in the urban areas. However, pending the availability of regionally detailed driving surveys, this method yields the best regional distributions available.

Emissions from steam locomotives in 1995 were estimated from the quantity of coal burned in the transportation sector according to the RAINS-Asia Model (interpolated between 1990 and 2000), the emission factor derived by the EDGAR team (Olivier et al., 1996), and regional distributions according to turnover volume of freight traffic on national railways (China Statistical Yearbook, 1996). Emissions from locomotives are about two orders of magnitude less than from motor vehicles. CO emissions from airplanes, ships, and other vehicles were assumed to be negligible for China.

The CO emissions estimated for highway vehicles in 1995 in this work, 23 million tonnes, is considerably higher than the EDGAR estimate of 9.9 million tonnes for 1990 (Olivier et al., 1996) and an undocumented estimate of 8.4 million tonnes for 1993 recently presented by The World Bank (1997). In fact, a previous estimate by Tsinghua University (He et al., 1996) was 9.7 million tonnes for 1995, but this has been improved by the use of laboratory-tested emission factors and the inclusion of motorcycles. The new study used here reveals that CO emissions from highway vehicles are more than twice as high as previously suspected. Control measures that include rapid replacement of older, inefficient, domestically manufactured vehicles and the provision of new highways to increase average speeds were recommended by Tsinghua University.

Using a similar method to that for domestic and industrial combustion, 1995 regional data for electric power generation were calculated by applying national IEA fuel and sector-specific trends to the regional data determined by applying the formula for fixed annual growth to 1990 and 2000 data from the RAINS-Asia model. Emission factors from AP-42 (EPA, 1995) were available for a variety of technologies, so emission factors for technologies that most accurately represent current practices in Chinese power generation were used. These emission factors were multiplied by the regional energy consumption to yield CO emissions from the electric power generation sector.

It is estimated that approximately 112.7 mt (million metric tonnes) of CO were released in China in 1995. This figure is 35% greater than 1995 CO emissions in the United States of 83.5 mt (EPA, 1996). Sectoral distribution is also significantly different, with the majority of U.S. emissions coming from the transportation sector, while in China the transportation sector is responsible for only 20% of total CO in 1995 (Table 4). In 1995, the Chinese domestic sector emitted 73.3 mt of CO, or 65% of the total CO emissions in China (Table 4). This sector contributed more than three times as much CO as the next highest emitting sector. There are a variety of reasons that domestic emissions are so high. Biofuels are a large portion of the energy consumed in this sector. When these fuels are burned in inefficient stoves that do not burn at high enough temperatures to fully combust the fuel, CO emissions can be quite high. CO emissions from biofuels amounted to approximately 62.6 mt in 1995. This is over 85% of the CO from the domestic sector, and nearly 54% of the total CO emitted in China. Approximately 9% of total CO emitted comes from the use of fossil fuels (particularly coal) in the domestic sector. This is due, again, to incomplete combustion in typical Chinese stoves. Another factor that affects the emission factor for domestic coal consumption is the prevalence of unwashed, poor quality coal in this sector, especially in rural areas.

The industrial sector, including iron and steel manufacture and combustion in industrial boilers, contributed 9% of total CO emissions in 1995. This makes it the third highest CO emitting sector in China. Approximately one-half of the emissions from industrial processes (iron and steel making) came from the 14 largest steel manufacturing plants (Table 2). Coal-burning boilers are responsible for 98% of the industrial combustion emissions.

Just over 5% of the CO emitted in China came from the open burning of crop residues in fields. This practice is used mainly to prevent the spread of crop diseases and for ease in disposal of residues. The most effective way to reduce CO emissions in the agricultural sector is to encourage farmers to plow under crop residues instead of burning them. Plowing under crop residues also contributes to greater organic content in the soil and decreases erosion (Lu, 1993). This practice has been growing in recent years, but with nearly 25% of crop residues still being burned in fields, there is clearly room for improvement.

The transportation sector was the second largest CO emitter, contributing 20.4% of total CO emissions in 1995. Trucks contributed the bulk of the emissions with 17 mt. Personal transportation, cars and motorcycles, contributed about one-third of that amount, with 4 and 2 mt, respectively. Rail transport had very low emissions, only 0.2% of total transportation emissions. Rail transportation tends to have lower CO emissions due to the high quality of coal used for this sub-sector and the efficiency of combustion. With personal cars likely to become more commonplace, emissions from passenger cars are expected to rise rapidly over the next decade and beyond. With 20.4% of total CO emissions coming from the transportation sector, and knowing that its percent contribution is likely to grow in the coming years, it would be wise to institute controls such as catalytic converters to prevent rapid growth in CO emissions from this sector.

Electric power generation contributed an almost insignificant amount of CO emissions in 1995. Emitting only 180 thousand tonnes of CO, it was the smallest contributing sector with 0.16% of the total emissions. Typically, power-sector boilers are large and operate at high temperatures, leading to relatively complete combustion.

The total 1995 CO emissions estimated in this study compare favorably with the EDGAR (Olivier et al., 1996) estimate for 1990. However, at the sector level, there are some obvious discrepancies. Our estimate of field burning of crop residues was 6.0 mt, less than one-quarter of the EDGAR estimate of 25.9 mt (Table 4). This difference can perhaps be partially explained by comparing total biofuel emissions from the domestic and field combustion sectors. The percent contribution of biofuels from both sectors is similar in the two studies; domestic and field combustion in our study contribute 60.9%, with the EDGAR study estimates these categories at 62.7% of total emissions. Thus, different ways of classifying biofuel combustion may be partly responsible for this discrepancy. A similar problem is found within the domestic sector, with the EDGAR estimate of fossil-fuel emissions being significantly higher, and biofuel emissions lower, than found for this study. Again, there appears to be a variance in distribution, but the percent contribution for the entire sector is comparable.

CO emissions are concentrated primarily in the eastern third of China. The three regions with the highest total emissions were HEHE (Henan, Hebei, and Anhui), NEPL (Heilongjiang, Jilin, and Liaoning), and SICH (Sichuan) with 16.7 mt, 11.4 mt, and 10.0 mt, respectively. As expected (since these are also the three most populated regions in China), these regions also had the greatest emission levels from the domestic sector, emitting 10.3 mt, 6.4 mt, and 7.8 mt, respectively. Industrial CO emissions in HEHE and NEPL, at 1.9 mt and 1.8 mt, contributed over 35% of the total industrial emissions in China. CO emissions from transportation were also greatest in HEHE and NEPL (3.3 mt and 2.4 mt), contributing more than the third, fourth, and fifth highest emitting regions combined.

The sectoral distribution of CO emissions in Beijing (Beij) and Shanghai (Shan), the two largest cities in China with similar populations, differed significantly in two main sectors, domestic and industry. Although the population of Shanghai is slightly greater than Beijing, domestic emissions in Beijing were nearly one-third greater than in Shanghai. This difference is probably primarily due to the greater heating needs in the northern climates. However, industrial emissions were much greater in Shanghai (0.91 mt compared to 0.62 mt in Beijing) due to the larger number of iron and steel manufacturing plants located there. Emissions from passenger vehicles were also 60% greater in Beijing (1.01 mt) than in Shanghai (0.63 mt).

Emissions were generally calculated based on 2020 energy projections from the RAINS-Asia model. For most sectors these energy projections were combined with emission factors for typical performance of facilities in the United States in 1990 (EPA, 1996). This method assumes a gradually slowing economic growth rate and that Chinese technology will continue to trail the United States by approximately 20-25 years. Table 5 contains the projected regional and national data for all sectors. The sector-specific methodology for 2020 is as follows.

Carbon monoxide emissions from fossil fuel combustion were calculated as described above, with the exception of oil emissions. Because CO emission factors from household oil combustion in China are currently quite low, the emission factors were assumed to remain constant through 2020.

Biofuels were also an exception to the above rule, since there are currently no RAINS-Asia projections for biofuel use in China in 2020. Several assumptions were used to estimate consumption of biofuels and the resulting CO emissions. The regional distribution and initial per capita consumption rates are based on a 1990 inventory (Streets and Waldhoff, 1998). The rate of per capita consumption was calculated from data for previous years (Sinton et al., 1996 and Lu, 1993). All rates showed biofuel use to be declining on a per capita basis. Because it is believed that biofuels will play less of a role in the future energy needs of China, the greatest level of decline (-0.01 kgce/cap/ year) was applied to each of the three main types of biofuel: fuelwood, crop residues, and animal waste. Per capita totals were then multiplied by 1,450 million, the Asian Development Bank’s projection for Chinese population in 2020 (Siddiqi et al., 1994). Total energy values were multiplied by appropriate emission factors (assumed to remain at 1995 levels) to yield 58 mt-CO from domestic biofuel combustion.

Carbon monoxide emissions from the combustion of coal and oil in the industrial sector were calculated as described above, using RAINS-Asia projections of energy use in this sector for 2020 and emission factors for the best current technology in the United States.

It is assumed that China will continue to be a major producer of iron and steel into the next century, but that processes will become increasingly more efficient and less polluting. In terms of total steel production, it is assumed that China's output will rise from 66 million tonnes in 1990 and 95 million tonnes in 1995 (International Iron and Steel Institute, 1998) to 140 million tonnes in 2020 (Worrell, 1995). This growth rate is slightly lower than what would be predicted on the basis of the RAINS-Asia overall industrial growth rate for China.

We further assume that by 2020 China's steel industry will have transformed itself to look something like the U.S. steel industry of 1990. This is consistent with some of the scenarios of Worrell (1995) and presumes that it takes more than 25 years to fully implement industrial change in China, considering the age of current equipment and rates of depreciation (Worrell, 1995). By making this assumption, we implicitly encompass a range of process types and emission rates. The U.S. produced 89.7 million tonnes of steel in 1990 (International Iron and Steel Institute, 1998), and total CO emissions from the ferrous metals processing industry in 1990 were 2.16 million short tons (U.S. EPA, 1997). This yields an average, industry-wide emission rate of 24.11 thousand short tons per million tonnes of steel produced. Note that this U.S. value had dropped to 20.27 by 1995, due to process improvements.

Application of the 1990 average emission rate in the U.S. to the 2020 estimated China production level, yields an estimated total CO emissions of 3.38 mt in 2020. To allocate these emissions to regional level, it was assumed that 80% of the emissions would be released from LPSs and 20% from area (smaller) sources; this is consistent with the expected closure of small, inefficient (and polluting) facilities and concentration of production at larger, efficient facilities. (In 1995, the split was 47% LPS and 53% area.) It is further assumed that the regional distribution of CO emissions from iron and steel production in 2020 will be the same as in 1995 for LPSs and area sources, separately. This implies that production will be generally located at the same places in the future as it was in 1995. There is no information on which to base any other assumption about future spatial allocations of production and emissions.

Because combustion of crop residues in the field is used as a means of disposing of this waste, and not as an energy source, it has the potential to vary significantly over a period of years. Efforts are currently being made to encourage farmers to plow under these wastes to increase the organic content and fertility of the soil.

Several factors needed to be examined in order to estimate emissions from this sector. The first factor is the projected Chinese grain production in 2020. The World Bank series, "China 2020" predicts that about 636 mt of grains (rice, wheat, and "coarse" grains) will be harvested in 2020. This is expected to produce a total of 737 mt of grain residues. This figure was used to calculate total crop residues in 2020. It was further assumed that efforts to educate farmers on the benefits of plowing-under versus burning crop residues would have a significant impact on farmers’ behavior, reducing the percent of residue burned in the fields in 2020 to one-half the rate in 1995 (from 23% to 11.5%). The CO emission factor for in this uncontrolled type of burning was assumed to remain equal to the 1995 rate. When this emission factor was applied to the total estimated crop residues subjected to burning, the result was 4.9 mt-CO (Table 5). Because of a lack of provincial projections, regional distribution was assumed to be the same as in 1995.

The estimation of CO emissions from the Chinese transportation system in 2020 requires a characterization of both the rates of growth of different modes and vehicles and the transformation of vehicle types to a more western picture of energy efficiency, fuels, and atmospheric emissions. For the purposes of this study, we have chosen to assume that vehicle performance in China in 2020 will approximate vehicle performance in the U.S. in 1990. This implies considerable improvement in Chinese vehicle performance.

Table 6 summarizes the vehicle data. Future growth in the number of gasoline-fueled passenger cars (LDGV) in China is expected to be extremely rapid. By 2020, we estimate 37.2 million such vehicles, with an additional 7.5 million diesel cars (LDDV). This estimate is obtained by extrapolation from data in the report by Tsinghua University (1997), and is consistent with the estimate by Siddiqi et al. (1994) of 44 million private cars by 2020. Though this number is large, it is still less than one-third the number of cars in the U.S. in 1990 (about 130 million) and one-twentieth on a per capita basis. Table 6 also shows the expected numbers of other types of highway vehicles in 2020. A large increase in the number of motorcycles, especially in southern Chinese cities, is anticipated.

The average emission rates of these classes of vehicles are assumed to be those of typical similar classes of vehicles in the U.S. in 1990. U.S. CO emission rates (in grams/mile) were obtained for vehicles with average lifetimes (U.S. EPA, 1995). Where possible, these emission rates are compared with the emission rates measured from typical Chinese vehicles today. The average numbers of miles driven per year in the U.S. in 1990 for each class of vehicle were also obtained (U.S. Department of Transportation, 1998). Combining these estimates yields national CO emission estimates by vehicle type.

Table 5 shows that CO emissions from passenger cars are expected to increase from 4.0 mt in 1995 to 9.4 mt in 2020: the improved performance of the average car is counteracted by the vastly larger numbers of cars on the road. Similarly, CO emissions from motorcycles is projected to increase from 2 mt in 1995 to 4.5 mt in 2020, again due to the growth in volume. However, truck emissions are projected to decline rapidly, due to replacement of gasoline-powered trucks (HDGV) by diesel-powered trucks (HDDV), primarily on the basis of improved energy efficiency. The overall effect is that total national CO emissions from highway vehicles are projected to decline from 23.0 mt in 1995 to 17.6 mt in 2020. This future estimate is subject to a wide range of uncertainty that depends on the rate of growth of transportation energy demand, the growth of public transport, the vehicle performance levels of domestically produced vehicles, and the rate of turnover of existing vehicle stock. National emissions in 2020 were allocated to regions using the same distributions as in 1995 for each of the three major vehicle types.

Emissions from rail transportation are expected to decline to negligible levels by 2020, as existing coal-powered steam locomotives are replaced by diesel and electric locomotives, both of which have negligible CO emissions (at the vehicle level).

Carbon monoxide emissions from power generation were calculated as described above, using RAINS-Asia estimates for coal consumption in this sector in 2020 and EPA estimates for CO emission factors of the best current technology in the United States.

Total emissions of CO in China are expected to decrease from 112.7 mt in 1995 to 99.3 mt in 2020. This is largely due to the improvements in combustion efficiency China is expected to make over this period. If these gains are not made, CO emissions will rise considerably due to the expected increase use of fossil fuels.

Carbon monoxide emissions from all sectors are expected to decrease over this 25 year period. The domestic sector is still expected to dominate emissions with 70.8 mt-CO in 2020 (71% of total emissions). Although fossil-fuel emissions are expected to increase slightly, these will likely be off-set by the decrease in CO emissions from biofuel combustion in homes.

The same is true for the industrial sector. Total sector emissions are expected to decline, despite slight increases in fossil-fuel combustion related emissions. In this sector in particular, it is the expected improvements in combustion efficiency that will regulate the CO emissions. If these improvements are lagging, CO emissions are likely to jump significantly as the use of fossil fuels increases.

As more farmers are educated about the benefits of plowing-under versus burning of crop residues and better means of disease control are developed and introduced, combustion of crop residues in the fields as a means of disposal is expected to decline.

A large expected increase in the number of vehicles in China will be counteracted by improvements in vehicular fuel combustion efficiency. A rapidly expanding diesel fleet will also help to hold down CO emissions. It seems inevitable that emissions from passenger cars will increase, but a decline in truck emissions will compensate.

Due to the increased use of new, large, and efficient power plants over small, local generators, CO emissions in China from the power generation sector are expected to decline slightly by 2020.

It is known that there are at least two other significant sources of CO emissions in China. Spontaneous combustion in coal mines (Mao, 1995) and savannah burning and deforestation for the purpose of land clearing (Olivier et al., 1996) are well-known sources of CO emissions. By one estimate, 100 million tonnes of raw coal is burned in uncontrolled coal-mine fires each year (Mao, 1995), primarily in Xinjiang Province. This could produce upwards of 6 mt-CO annually. An estimate of CO produced from savannah burning and deforestation is of a similar magnitude (6.4 mt-CO in 1990, Olivier et al., 1996). Unfortunately, provincial breakdowns of these data are not readily available, and interannual variability is so large as to render an estimate for any particular year unsatisfactory. These two categories were, therefore, not included in the final estimates at this stage. However, we do believe that the national total of CO emissions would be roughly 10-15 mt greater if these sources could be better characterized and added to the inventory.

To allocate emissions from their regional distributions to a gridded distribution, the following procedures are recommended:

(1) LPS emissions to be allocated according to latitude and longitude coordinates contained in Table 3.

(2) All domestic, iron and steel area sources, industrial fuel combustion, transportation, and power sector emissions to be allocated according to population distribution within a particular region. This is an approximation for all but domestic emissions; however, no better methods exist at present.

(3) Emissions from combustion of crop residues to be allocated according to the gridded distribution of crop production levels.

1995

2020

The authors are grateful to John C. Molburg of Argonne National Laboratory for assistance with the development of emission estimates for iron and steel manufacturing and The World Bank for permission to use Figure 1. This work was funded by the National Aeronautics and Space Administration as part of the China-MAP program.

Arndt, R. L., Carmichael, G. R., Streets, D. G. and Bhatti, N. (1997) Atmospheric Environment 31(10) 1553-1572.

China Statistical Yearbook, China Statistical Publishing House, 1996.

Crutzen, P. J., Heidt, L. E., Krasnec, J. P., Pollock, W. H. and Seiler, W. (1979) Nature 282 253- 256.

Crutzen, P.R. and Andreae. M. O. (1990) Science 250 1669-1678.

Food and Agriculture Organization, (1995) FAOSTAT Agricultural Data, http://www.fao.org.

He, K., Hao, J., Fu, L., Lee, M. and Liu, Y. (1996) Motor Vehicle Related Air Pollution in China, preliminary report, Tsinghua University, Beijing, China.

International Energy Agency (1997) "Energy Statistics and Balances," IEA and OECD, Paris, France.

Lu, Y. (1993) Fueling One Billion: An Insider’s Story of Chinese Energy Policy Development, Washington Institute Press, Washington, D.C.

Mao, Y. (1995) Spontaneous Combustion in Chinese Coal Mines and Its Environmental Impact, in "Local and Regional Energy-Related Environmental Issues" World Energy Council, United Kingdom.

OECD Working Group (1991) Estimation of Greenhouse Gas Emissions and Sinks, Final Report from OECD Experts Meeting, February 1991.

Olivier, J. G. J., Bouwman, A. F., van der Maas, C. W. M., Berdowski, J. J. M., Veldt, C., Bloos, J. P. J., Visschedijk, A. J. H., Zandveld, P. Y. J. and Heverlag, J. L. (1996) EDGAR v.2.0, RIVM Report No. 771060 002 / TNO-MEP Report No. R96/119, Bilthoven, The Netherlands.

Siddiqi, T. A., Streets, D. G., Wu, Z. and He, J. (1994) National Response Strategy for Global Climate Change: People’s Republic of China, Asian Development Bank, Report No. TA 1690-PRC.

Sinton, J. E., ed. (1996) China Energy Databook, Report No. LBL-32822 Rev. 4, Berkeley, CA.

Smith, K.R. (1988) Environment 30, 16-35.

Smith, K.R., Health, Energy, and Greenhouse-Gas Impacts of Biomass Combustion, presented at BioResources ‘94, Bangalore, India, October, 1994.

Streets, D. G. and Waldhoff, S. T. (1998) Energy—The International Journal, in press.

Tsinghua University and others (1997) "China’s Strategies for Controlling Motor Vehicle Emissions," Report to the World Bank under China Environmental Technical Assistance Project B-9-3.

United Nations Environment Programme (1997) Steel Industry and the Environment: Technical and Management Issues, UNEP Technical Report No. 38.

U.S.D.A. Economics and Statistics System (1998)

U.S. Department of Transportation (1998) http://www.bts.gov/btsprod/nts/apxav.html.

U.S. Environmental Protection Agency (1995) AP-42, http://www.epa.gov.

U.S. Environmental Protection Agency Working Group (1996) National Air Pollutant Emission Trends,1900-1995, Office of Air Quality Planning and Standards, Research Triangle Park, NC.

The World Bank (1997) China 2020: Clear Water, Blue Skies, Washington D.C.

The World Bank (1997) China 2020: At China’s Table , Washington D.C.

Zhang, J. (personal communication, preliminary information) and Zhang, J., Smith, K. R., Kishore, V. V. N., Ma, Y., Rasmussen, R., Uma, R., Khalil, M. A. K., Kusam, J. and Thorneloe, S. T. Paper presented at the Air & Waste Management Association Conference on "Emission Inventory: Planning for the Future," Research Triangle Park, N.C., October 28-30, 1997.