Environ. Sci. Technol. 1999, 33, 3056-3061
Secondary Benefits of Greenhouse Gas Control: Health Impacts in China§ XIAODONG WANG† AND K I R K R . S M I T H * ,†,‡ Environmental Health Sciences, University of California, Berkeley, California 94720-7360, and East-West Center, Honolulu, Hawaii 96848
In addition to long-term benefits of greenhouse gas (GHG) reductions in the form of avoided health and ecosystem damage, there are important near-term benefits resulting from the reduction in health-damaging pollutants (HDP) that can accompany GHG reductions. This paper estimates such health benefits, using the power and household energy sectors of China as a case study for the method. Four policy scenarios were compared: business as usual, leastcost energy efficiency, least-cost per unit global-warmingreduction fuel substitution, and least-cost per unit exposurereduction fuel substitution. The health benefits were estimated by converting PM10 emissions first to human exposures and then to avoided mortality and morbidity with existing exposure-response relationships. Our results demonstrate that the near-term health benefits from GHG reductions in China could be substantial but are highly dependent on the technologies and sectors chosen. Such near-term benefits provide the opportunity for a true “no-regrets” GHG reduction policy. The results of this study also have important implications for the current international negotiations to cut GHG emissions by demonstrating that GHG reduction strategies can have substantial local and national as well as global benefits.
Introduction One of the most controversial aspects of the climate change protocol signed at Kyoto in December 1997 is its provision for development of an emissions trading framework that would allow countries to invest in greenhouse gas (GHG) reduction projects in other countries and share part of the emissions-reduction credits (1). Discussions of such schemes in the form of “joint implementation” and “clean development mechanisms” were indeed prominent in the following meeting held in Buenos Aires (Nov 1998). Since any particular project will have impacts beyond just modification of GHG emissions, a critical part of this debate revolves around the nature and extent of “secondary benefits” that might be created by different GHG reduction strategies. The most telling argument for spending resources to reduce current GHG emissions is that the benefits in the § This article is a summary of the October 1998 WHO report “Nearterm Health Benefits of Greenhouse Gas Reductions: A Proposed Assessment Method and Application in Two Energy Sectors of China” (5). * Corresponding author phone: (510)643-0793; fax: (510)642-5815; e-mail:
[email protected]. † University of California. ‡ East-West Center.
3056
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999
FIGURE 1. Two pathways by which reduced greenhouse gas emissions can improve health. Over the long term, reduced emissions may lead to fewer hazards from global warming. In the near term, however, they lead to reduction of health-damaging pollutants. form of reduced impacts of climate change will be greater. Among the important benefits, which also include avoiding ecosystem effects that could have significant indirect impacts on humanity, are avoiding or reducing the impacts on human health (Figure 1). Studies have shown that reduction in current greenhouse gas emissions could lead to reduction in long-term health risks in the form, for example, of lessened spread of malarial mosquitoes, fewer extreme climate events, and lower impacts on food production compared to what might occur if today’s GHG emission trends continue (2). Each step of the causal chain from GHG emissions through global warming to health effects is not understood with certainty, however, leading many observers to view the overall connection with skepticism. Such skepticism still plays a role in holding back international agreements to significantly alter current patterns of GHG emissions. This is particularly so in developing countries, which must contend today with many urgent problems related to human health and welfare. There is another and more certain health benefit from GHG reductions, however, in the concomitant reduction in health-damaging air pollutants (HDP) that will occur (Figure 1). The same combustion processes that produce GHG emissions such as carbon dioxide and methane also generate local and regional HDP, such as particulates and sulfur oxides (3). Thus, reduction in GHG with benefits several decades from now can be expected to achieve HDP reduction as well, potentially bringing health improvements with a much smaller mean delay, days to years depending on the disease. These near-term benefits may allow for “no-regrets” GHG reduction scenarios, which achieve a significant degree of near-term benefits as well as GHG reduction so that immediate action can be justified even if it later develops 10.1021/es981360d CCC: $18.00
1999 American Chemical Society Published on Web 08/04/1999
that the human vulnerability to climate change wrought by GHG additions is less than now thought (4). To obtain an idea of the extent of the HDP reduction from GHG control, we estimated the near-term human health benefits of GHG reductions resulting from changes in energy efficiency and structure of energy use in China, both as a case study for the method and because China is such an important actor in global GHG scenarios (5). In addition, because of its extensive use of solid fuels, a significant fraction of the nation’s current burden of ill-health can be attributed to air pollution exposure (6). Much of the exposure occurs indoors from burning solid fuels, as coal in the industrial, power, and commercial sectors and as unprocessed household fuels (coal and biomass) for cooking and heating (7).
Methods To estimate near-term health benefits, it is necessary to link each specific technological option taken in a particular GHG reduction scenario with the accompanying HDP emissions reductions. The health impact of these emissions, however, is closely dependent on the sector of the economy in which the emissions occur. This is because the degree of human exposure created by a unit of HDP emissions depends on where they are released in relation to where people spend time. That is, their exposure or dose effectiveness can vary dramatically (8). Thus, a ton of emissions averted in the household sector close to where people live much of the time will generally cause a much greater reduction in human exposure than a ton of outdoor emissions averted in the industrial sector. The methodological framework followed in this analysis is shown in Figure 2. The first steps involve choosing realistic scenarios for GHG reductions. The GHG reduction targets chosen for analysis here are 10% below business-as-usual (BAU) by 2010 and 15% below by 2020, which were proposed by the U.S.A. and Australia for developing countries before the Kyoto Protocol but do not represent the official position of China or any other developing country. Unlike the commitments asked of developed countries at Kyoto, this hypothetical commitment for China would be only to reduce the growth of emissions, not to reduce them below current levels. For meeting these targets, we examined the two most likely policy approaches, energy efficiency and fuel substitution, in two sectors, household and power. Four scenarios are examined overall: BAU, least-cost energy efficiency, leastcost per unit global-warming-reduction fuel substitution, and least-cost per unit human-air-pollution-dose-reduction fuel substitution. The BAU scenario accounts for improvements in efficiency and emissions that are expected to occur regardless of GHG control strategies. The energy efficiency scenario maintains the same fuel mix as the BAU scenario but accelerates the improvement in supply side energy efficiency to achieve the GHG reduction target. The fuel substitution scenario is intended to explore the pathway of fuel switching to reach the same GHG reduction target. We examined two fuel switching pathways: least-cost per unit global warming potential (GWP) reduction scenario and leastcost per unit dose reduction scenario. “Least-cost” refers to a pathway in which the cheapest options are taken first until exhausted, followed by the next cheapest, etc. These comparisons allow us to examine the relative health benefits achieved by different technological and policy approaches to meeting the same GHG reduction targets. The next step is to conduct a set of technology assessments comparing HDP emissions (particulates and SO2), GHG emissions (CO2 and CH4), and economic costs of each energy technology option on a per-unit-delivered-energy basis. The technologies chosen in both sectors represent realistic nearterm alternatives for China (9). The environmental com-
FIGURE 2. Analysis framework for this assessment. The policy approach combines with the technological characteristics of various energy options to produce alternative energy scenarios for each of the four principal sectors. Here, however, we only examine the household and power sectors. These scenarios, in turn, produce a certain mix of greenhouse emissions and health-damaging emissions, which, after adjusting by, respectively, their global warming potentials and exposure effectiveness, produce estimates of total global warming and human dose averted. Each scenario is taken to the point at which the chosen GHG reduction target is met. At that point, the dose averted is then converted to health benefit for comparison. parisons in this research address the entire fuel cycle, that is, fuel mining, processing, transporting, conversion, transmission, distribution, and end-use. Only the operational stage in each step is considered, however, and not the impacts from construction, dismantling, or long-term waste management. We base our analysis on PM10 (particles less than 10 µm in diameter) and assume that 50% of SO2 emissions are converted to sulfate compounds in the fine particles fraction. The economic cost figures primarily come from Wang (9), and the detailed methodology can be found in the full report (5). Then, the GHG emissions are converted to overall global warming potential (GWP) by appropriate weightings according to the action of each GHG in the atmosphere. Used here is the 20-year GWP of methane, which is 23. This means that 1 g of carbon as methane has the global warming effects of 23 g of carbon as carbon dioxide over a 20-year period (10). In parallel, the HDP emissions are converted to human inhaled doses. This is because human health impacts are not directly related to emissions but rather to the exposures, which indicate how many grams of pollutants actually reach the places where people are breathing over a certain period of time. For ease of comparison, calculated exposure concentrations related to a source are here converted to nominal (administered) population doses by multiplying the following factors: exposure concentration, mean breathing rate, number of affected people, and the duration of exposure. For the same weight of emissions, the nominal population VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3057
FIGURE 3. (a) Assumptions used in determining exposure concentrations from power plant emissions. The fraction of time spent outdoors is multiplied by the exposure concentration calculated for the outdoor location and added to the fraction of time spent indoors multiplied by the indoor exposure concentration. (3b) For household emissions, the fraction of time spent outdoors is multiplied by the exposure concentration calculated for the outdoor location and added to the fraction of time spent indoors multiplied by the indoor exposure concentration. dose (or exposure effectiveness-the fraction that actually concentrations in Chinese households are derived from reaches people’s breathing zones) from different sources can “Indoor Air Pollution Database for China” (7). To take into vary by several orders of magnitude (8). account that improvements may have occurred in indoor Therefore, when comparing technologies of different levels since the studies making up the database were done, types, it is necessary to convert HDP emissions to exposures we took values at the low end for measured concentrations. or doses to obtain an accurate estimate of human health To calculate the particulate dose effectiveness, however, we effects. Here we multiply particulate nominal dose effectiveneed to have the particulate emissions data to link with the ness (PDE), which is defined as grams of particulates inhaled concentrations. Unfortunately, the database did not provide by humans per ton of emissions, by HDP emissions to obtain fuel use data corresponding to the indoor concentration a dose measure that is more strongly related to health effects owing to a lack of documentation in the original measurement than emissions alone. A different PDE is calculated for each reports. For a first approximation, we used the average fuel major type of technology depending on its location relative use per household per year for cooking and heating and the to the surrounding population. particulate emission factor to estimate the emissions. Outdoor For air pollution from coal-fired power plants, we used concentrations from local households are assumed to be a Gaussian plume model with Chinese meteorological data 10% of indoor level. This could be improved if better to estimate the changes in particulates concentration resultinformation becomes available about ground-level “neighing from marginal changes in emissions (similar to the borhood” pollution from local residential/commercial sources. approach in ref 11). We developed an exposure-distance curve According to our estimates, the health benefits of a 1 ton so that air pollution exposures out to a 50 km boundary from reduction in particulate emissions from household stoves each power plant could be included (5). A population density are at least 40 times larger than those from coal-fired power of 1250 people/km2 was assumed (11). Figure 3a shows the plants, as summarized in Figure 4. The method is detailed calculation of indoor and outdoor exposures from power in the WHO report (5). Since there is much variation plant emissions. throughout China in the factors that make up this comIndoor and outdoor exposures from household stoves are parison, these calculations should be considered only also determined, as shown in Figure 3b. Indoor particulates indicative. 3058
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999
FIGURE 4. Comparison of dose effectiveness and health effects of emissions from household stoves and power plants. Per ton of emissions, stoves seem to produce about 40 times more dose and health effects than do power plants.
FIGURE 5. The global warming potential (GWP: kg C as CO2) and particulate dose (PD: mg) per unit energy (GJ) delivered for various household cooking options. The figure is designed such that a potential change from one option to another will be equivalent in energy service terms (i.e., cooking) and thus directly indicate the difference in GWP and PD that would result by such a shift. For illustration, the arrows indicate how shifts from coal stoves to other options change both types of emissions for the same energy service output. In a similar fashion, the impact of shifts from any one technology to any other can be determined. Figure 5 shows GWP and HDP dose of different household cooking options to illustrate the type of information generated. Note that there is a substantial difference in the dose/ GWP ratios of different mixes of fuels/technologies. The GWP and dose created by any particular mixture of these technologies can be determined from such a figure by simply multiplying by the total energy production of each technology. The dose and GWP implications of technological shifts can easily be derived from the figure. Such a shift is denoted
by an arrow from one technology to the other. The arrow in Figure 5, for example, shows that a shift from traditional coal stoves to electric stoves (powered by coal-fired power plants) results in a substantial decrease in particulate dose but only limited reduction in GWP. A shift from traditional coal stoves to natural gas stoves, on the other hand, would result in substantial reductions in both types of pollution. A shift to traditional nonrenewable biomass stoves, in contrast, would result in an increase in both types of pollution. VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3059
TABLE 1. Annual Avoided Premature Deaths by 2020a scenario efficiency fuel substitution least-cost GWP fuel substitution least-cost dose
sector
low
central
high
power household power
1 500 62 000 1 700
4 400 150 000 4 900
13 000 460 000 15 000
household power
47 000 1 800
120 000 5 200
360 000 16 000
household
70 000
180 000
530 000
a
Estimated mortality/population in China in 2020: 14 million/1470 million (15).
Once the change in inhaled dose is determined, the resulting change in health risk (avoided mortality and morbidity) can be calculated using exposure-response information from epidemiological studies and other information. Unfortunately, most of such work to date has been done in developed countries. The few available Chinese studies, however, have generally found a somewhat lower risk per unit exposure. Because of the uncertainties of extrapolating the results of such studies from developed-country conditions to China, we thus utilize a range of possible risks per unit pollution resulting in high, central, and low estimates of morbidity and mortality. Our lower end of the exposure-response relationship range of mortality is derived from Chinese studies (11), and the higher end of the range is extrapolated from the chronic exposure-response relationship in the cross-sectional studies in the U.S. (12). A number of daily time-series studies in the U.S. have consistent results, which are intermediate (13). Thus, we use them as our central estimate. There are many uncertainties and complications of using these different kinds of studies to estimate actual public health risk, however (14). After choosing the cost-effective GHG mitigation options in the energy sector and examining their near-term health benefits, our report compared the marginal control costs to shift from conventional coal use to low carbon fuels with the marginal economic benefits of the improved health associated with such a change. Then, we estimated the marginal net economic costs of GHG reduction, which are the differences between these two. We used the method developed for this purpose by the World Bank (11), which applied the willingness to pay $3 million per statistical life saved in the U.S. to China by adjusting the income differences, and then the purchase power parity ratio was used to adjust for the differences in the levels of purchasing abilities between wages in the two countries.
Policy Implications We find that GHG reductions resulting from changes in Chinese energy use can be accompanied by substantial nearterm human health benefits. The degree of health benefit varies greatly with the choice of energy technologies and sectors, however. Shifting from conventional coal-fired power plants to natural gas, for example, has much greater relative reduction in HDP than in GWP, while shifting from coal power to hydropower results in the same percentage reduction in both HDP and GHG emissions. This variation in health benefits is even larger between sectors. Table 1 shows the estimated annual avoided premature deaths from the three alternative scenarios relative to the BAU case by 2020 (estimates of morbidity reduction are also provided in the full report). Depending on the scenario, there could be as much as a 4% reduction in the projected mortality 3060
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 18, 1999
of 14 million in China by 2020 (15). Contributions from HDP reductions in other sectors (transport, industry, and agriculture) not examined in our study would provide additional mortality and morbidity reductions. The wide range in the results show clearly the urgent need to pin down exposureresponse relationships under Chinese conditions and to better characterize the relationship between current emissions and exposures, both indoors and outdoors. We find that the economic benefits of improved health alone in the electric power sector seem not to be currently large enough to offset the incremental economic costs associated with introducing GHG reduction strategies. As incomes and willingness to pay rise, however, this situation could change. On the other hand, the economic benefits of improved health are currently substantially larger in the household sector than the costs of GHG control, indicating that efforts in this sector can be true no-regrets options. Premature mortality comparisons are difficult to interpret across populations and thus should be elaborated to show the actual lost life years, i.e., considering age distribution and life expectancy. There are also other important air pollutants and associated disease endpoints, for example lung cancer and arsenic poisoning from coal smoke. Complete analysis of the near-term health impacts of GHG control in the energy sector would also need to include such issues as changes related to occupational health, water pollution, and the risks of large accidents. The health impacts from these causes are likely to be much smaller in China at present, however, given the large impact of air pollution. A full accounting of secondary benefits of GHG control would need to include such factors as energy security, job creation, technology transfer, and ecosystem disruption. The near-term health benefits from GHG reductions in China could be substantial and are highly dependent on the technologies and sectors chosen. In China, much of the health benefit would occur by improving the efficiency of or switching away from traditional use of coal and biomass. Other countries with high dependence on solid fuels in the household and power sectors, India for example, could be expected to have a similar scale of benefits. Such near-term secondary benefits of GHG control provide the opportunity for a true no-regrets GHG reduction policy in which substantial advantages accrue even if the impact of humaninduced climate change itself turns out to be less than many people now fear. These results also have important implications for emissions trading in the form, for example, of joint implementation and clean development mechanisms. The near-term health improvements are local, i.e., they accrue nearly entirely to the nation in which GHG-control projects are undertaken. This is unlike the benefits of GHG reductions themselves, which accrue globally. Such large local benefits may provide a significant extra incentive for China and other developing countries to enter into arrangements by which local GHG controls are financed externally and the emissions credits are shared. Indeed, this study shows that a GHG reduction strategy can actually be consistent with such typical national development objectives as reducing local air pollution, increasing energy efficiency, and improving social equity by providing energy services to remote areas through renewable energy. Sharing GHG reduction credits in return for local benefits would obviously be even more attractive to developing countries if it involved access to financial flows not otherwise available to them. To achieve these benefits, however, considerations of health and other secondary benefits should be included from the start in designing GHG control strategies. Thus, healthbased analysis deserves a prominent place at the negotiating table in the current international debates about greenhouse gas control regimes.
Acknowledgments We appreciate research support by the World Health Organization, Fogarty International Center, and U.S. Environmental Protection Agency, but no official approval by any of these organizations of our study’s results should be inferred.
Literature Cited (1) Masood, E. Nature 1997, 390, 649. (2) Climate Change and Human Health; McMichael, A. J., Haines, E., Slooff, R., Kovats, S., Eds.; An Assessment Prepared by a Task Group on behalf of the World Health Organization, the World Meteorological Organization, and the United Nations Environment Programme. World Health Organization: Geneva, 1996. (3) Lancet 1997, 350, 1341. (4) Repetto, R.; Austin, D. The Costs of Climate Protection: A Guide for the Perplexed; World Resources Institute: Washington, DC, 1997. (5) Wang, X.; Smith, K. R. Near-term Health Benefits of Greenhouse Gas Reductions: A Proposed Assessment Method and Application in Two Energy Sectors of China; WHO/EHG/98.12; World Health Organization: Geneva, October, 1998. (6) Florig, H. K. Environ. Sci. Technol. 1997, 31, 276A. (7) Sinton, J.; Smith, K. R.; Hu, H.; Liu, J. Indoor Air Pollution Database for China; WHO/EHG/95.8; World Health Organization: Geneva, 1996. (8) Smith, K. R. Annu. Rev. Energy Environ. 1993, 18, 529.
(9) Wang, X. Comparison of Constraints on Coal and Biomass Fuel Development in China’s Energy Future. Ph.D. Dissertation, Energy and Resources Group University of California: Berkeley, 1997. (10) Intergovernmental Panel on Climate Change (IPCC). Radiative Forcing of Climate Change, and An Evaluation of the IPCC IS92 Emission Scenarios; Houghton, J. T., Meira Filha, L. G., Bruce, J., Lee, H., Callander, B. A., Haites, E., Harris, N., Maskell, K., Eds.; Cambridge University Press: Cambridge, U.K. 1995. (11) The World Bank. Valuing the Health Effects of Air Pollution, No. 8 in the series, China: Issues and Options in Greenhouse Gas Control; World Bank: Washington, DC, 1994. (12) Pope III, CA.; Dockery, D. Epidemiology of Chronic Health Effects: Cross-Sectional Studies. In Particles in Our Air: Concentrations and Health Effects; Wilson, R., Spengler, J., Eds.; Harvard University Press: Cambridge, MA, 1996. (13) Dockery, D.; Pope III, C. A. Epidemiology of Acute Health Effects: Summary of Time-Series Studies. In Particles in Our Air: Concentrations and Health Effects; Wilson, R., Spengler, J., Eds.; Harvard University Press: Cambridge, MA, 1996. (14) McMichael, A. J.; Anderson, H. R.; Brunekreef, B.; Cohen, A. Intl. J. Epidem. 1998, 27, 450. (15) Global Burden of Disease; Murray, C., Lopez, A., Eds.; Harvard University Press: Cambridge, MA, 1996.
Received for review December 31, 1998. Revised manuscript received May 7, 1999. Accepted May 12, 1999. ES981360D
VOL. 33, NO. 18, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3061
