Low-carbon energy policy and ambient air pollution in Shanghai, China: A health-based economic assessment

Low-carbon energy policy and ambient air pollution in Shanghai, China: A health-based economic assessment

Science of the Total Environment 373 (2007) 13 – 21 www.elsevier.com/locate/scitotenv Low-carbon energy policy and ambient air pollution in Shanghai,...

206KB Sizes 1 Downloads 37 Views

Science of the Total Environment 373 (2007) 13 – 21 www.elsevier.com/locate/scitotenv

Low-carbon energy policy and ambient air pollution in Shanghai, China: A health-based economic assessment Changhong Chen a , Bingheng Chen b , Bingyan Wang a , Cheng Huang c , Jing Zhao c , Yi Dai c , Haidong Kan b,⁎ a

b

Shanghai Academy of Environmental Sciences, Shanghai 200233, China Department of Environmental Health, School of Public Health, Fudan University, Shanghai 200032, China c East China University of Science and Technology, Shanghai 200237, China Received 28 May 2006; received in revised form 12 November 2006; accepted 18 November 2006 Available online 3 January 2007

Abstract Energy and related health issues are of growing concern worldwide today. To investigate the potential public health and economic impact of ambient air pollution under various low-carbon energy scenarios in Shanghai, we estimated the exposure level of Shanghai residents to air pollution under various planned scenarios, and assessed the public health impact using concentrationresponse functions derived from available epidemiologic studies. We then estimated the corresponding economic values of the health effects based on unit values for each health outcome. Our results show that ambient air pollution in relation to low-carbon energy scenarios could have a significant impact on the future health status of Shanghai residents, both in physical and monetary terms. Compared with the base case scenario, implementation of various low-carbon energy scenarios could prevent 2804–8249 and 9870–23,100 PM10-related avoidable deaths (mid-value) in 2010 and 2020, respectively. It could also decrease incidence of several relevant diseases. The corresponding economic benefits could reach 507.31–1492.33 and 2642.45–6192.11 million U.S. dollars (mid-value) in 2010 and 2020, respectively. These findings illustrate that a low-carbon energy policy will not only decrease the emission of greenhouse gases, but also play an active role in the reduction of air pollutant emissions, improvement of air quality, and promotion of public health. Our estimates can provide useful information to local decision-makers for further costbenefit analysis. © 2006 Elsevier B.V. All rights reserved. Keywords: Low-carbon development; Air pollution; Public health; Economic evaluation

1. Introduction Energy and related health issues are of growing concern worldwide today. Fossil fuels, the primary source of world energy, are the greatest source of ⁎ Corresponding author. Current address: Epidemiology Branch, National Institute of Environmental Health Sciences, P.O. Box 12233, Mail Drop A3-05, Research Triangle Park, NC 27709, US. E-mail address: [email protected] (H. Kan). 0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2006.11.030

ambient air pollution, producing particulate matter (PM), nitrogen oxides (NO2) and sulfur oxides (SO2). These pollutants have been related with increased mortality and morbidity from cardiorespiratory diseases (Brunekreef and Holgate, 2002). The burning of fossil fuels is also the major source of carbon dioxide (CO2), a primary contributor to global warming (Cifuentes et al., 2001). Coal has been China's long-term primary energy source. The relatively poor energy technology currently in use in China has caused high emissions of both local air

14

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

pollutants (LAP) and CO2 during energy consumption. With the high speed of economic growth, energy demand will increase concomitantly with economic development. Clearly, today's decisions regarding energy policies in China will have a significant effect on future CO2 emission, air pollution and the health status of local people. Shanghai, the largest city in China, is in the leading position of economic development in the country. Emissions of CO2 and LAP per unit of gross domestic product (GDP) in Shanghai are much higher than those in developed countries. Clearly, a low-carbon development strategy is beneficial for the city because it can reduce the emissions of CO2 and LAP, while simultaneously meeting the energy demand of economic growth. In the present study, we strive to evaluate the public health impact of ambient air pollution under various low-carbon energy scenarios in Shanghai, and to put monetary values on the estimated health effects. Our results can provide useful information to local decision-makers for further costbenefit analysis of low-carbon development policy. 2. Methods 2.1. Development of low-carbon energy scenarios In this study, a LEAPs (long-range energy alternatives planning system) model was used to study the effect of low-carbon energy scenarios on the emissions of CO2 and LAP. In contrast with other optimization tools, such as MARKet ALlocation (MARKAL), LEAPs is an accounting tool that balances production and consumption of energy in an energy system. As opposed to the MARKAL model, LEAPs has fewer degrees of freedom, and usually has only one solution to the proposed problems. Both LEAPs and MARKAL models have been successfully applied in energy scenario analysis in developing countries (ERC, 2006). Although LEAPs is basically an accounting framework, users can go beyond simple accounting and perform sophisticated simulations. Details of LEAPs are available from the Stockholm Environmental Institute — Boston (LEAPs, 2001). LEAPs uses the concepts of sector, sub-sector, end use and devices, and therefore analysis of demand and resource at each stage, from end use to primary energy level, may be carried out. Our study forecasts the future energy demand in Shanghai according to the driving factors of economic growth, city development scale, industry structure and vehicle population, and estimates LAP and CO2 emission according to energy consumption. Details of the application of the LEAPs model in Shanghai have been discussed previously (Wang et al., 2004).

Scenarios related to this study included base-case (BC) and various low-carbon scenarios. The base-case scenario only considers the economic growth in Shanghai, assuming that GDP growth rate is be 9– 11% during 2000–2005, with a 0.5%–1% decline of GDP growth rate on the base of 9–11% for every five years afterward between 2005 and 2020. The lowcarbon scenarios consist of three sub-components: energy efficiency improvement (EE), expanding natural gas use for final sectors (GAS), and wind electricity generation (WIND). In the EE scenario, an annual average improvement of 2% is assumed in final energy use sectors from 2000 to 2020. In the GAS scenario, direct burning of hard coal and coal gas is switched to natural gas. The detailed settings are as follows: ▪ Primary industry: The share energy derived from hard coal falls by 5% each year. ▪ Other industry: In industry sectors apart from primary industry, the dependence on hard coal falls by 0.5% each year from 2001 to 2005 and by 13% each year after 2005. ▪ Commercial use: The need for hard coal falls by 5% each year, and the dependence on coal gas will drop to zero after 2005. ▪ Residential use: For the urban area of Shanghai, the use of hard coal falls by 5% each year from 2001 to 2005 and by 10% after 2005; while coal gas utilization decreases by 5% yearly. In the rural areas, the use of hard coal falls by 5% yearly from 2001 to 2005 and by 10% yearly after 2005. The WIND scenario sets the capacity of power generation according to “Eleventh Five-Year Plan of Electricity Supply in Shanghai”. The detailed settings are shown in Table 1. The detailed elements of various scenarios are summarized in Table 2. 2.2. Concentrations of ambient air pollutant Emissions of particulate matter less than 10 μm (PM10) and SO2 provided by the LEAPs model were summarized in Table 3. Based on the principle of transfer matrix, one type of air quality model \ the exposure level model was developed to link emission scenarios of the LEAPs model and air pollutant concentrations. The fundamental matrix was input through a longrange transport and deposition model (ATMOS model)

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21 Table 1 Electricity generation capacity in WIND scenario (unit: 10,000kW)

Power plant

Combination of heat and power (CHP) Total a

a

Year

Coal

Natural gas

Wind

Total

2006 2010 2030 2006 2010 2030 2006 2010 2030

699.9 805.9 1165.9 447.5 747.5 2347.5 1147.4 1553.4 3513.4

50 50 50 150 150 150 200 200 200

20 20 60 0 0 0 20 20 60

769.9 875.9 1275.9 597.5 897.5 2497.5 1367.4 1773.4 3773.4

15

Table 3 Emissions of PM10 and SO2 for selected scenarios in Shanghai (unit: 1000 ton/year) Scenarios

Pollutants

2000

2010

2020

BC

PM10 SO2 PM10 SO2 PM10 SO2 PM10 SO2

144 461 144 461 144 461 144 461

247 745 216 666 168 544 154 503

386 1111 293 871 169 569 154 530

EE GAS WIND

Combination of power plant and CHP.

for primary PM10 and SO2. The ATMOS model is a Lagrangian parcel model with three vertical layers (Calori and Carmichael, 1999). For the Shanghai project, the ATMOS model provided a 4 km × 4 km resolution for the concentration of primary PM10 and SO2. The total area of Greater Shanghai, 6341 km2, was divided into 487 grids. Two transfer matrices for use in exposure level prediction were produced: a region-to-grid matrix for the area sources, and a large point source-to-grid matrix for the elevated point sources. Based on matrix output of the ATMOS model, the Shanghai Exposure Level model was developed in Excel to link the emission prediction of the LEAPs model and provide exposure levels for the health impact analysis. The year 2000 was selected as the base period in this analysis. Air quality changes in 2010 and 2020 were estimated under the following scenarios: BC, EE, GAS and WIND. 2.3. Human exposure level to PM10

of Shanghai residents in each 4 km × 4 km grid cell was then made for the assessment based on the population data collected from the Shanghai Bureau of Statistics. Combining the PM10 level and population number in each cell, we estimated the population exposure level to outdoor air pollution under various scenarios in 2010 and 2020 in Shanghai. 2.4. Estimation on health effects To develop estimates of public health impact of air pollution, we relied on published peer-reviewed studies on air pollution and health, using concentrationresponse (C-R) coefficients derived from studies conducted in China or worldwide. Since most of the epidemiologic studies linking air pollution and health endpoints are based on a relative risk model in the form of a Poisson regression, the cases at a given concentration C, could be given by: E ¼ expðb  ðC−C0 ÞÞ  E0

In the present assessment, PM10 was selected as an indicator of air pollution to estimate the relevant health effects, since there has been strong epidemiologic evidence to support its association with adverse health effects among all air pollutants. All people living in Greater Shanghai were considered as the exposed population in this analysis. An estimate of the number

ð1Þ

In Eq. (1), C and C0 are the PM10 concentration under one specific scenario and baseline scenario, respectively, and E and E0 are the corresponding health effect cases under the concentration of C and C0. The health effect (benefit/damage) under the scenario with respect to baseline scenario is the difference between E

Table 2 Elements of the scenarios Economic Energy Natural gas Wind growth efficiency availability electricity improvement generation Base-case (BC) Low carbon scenarios EFF GAS Wind



√ √ √

√ √ √

√ √



Fig. 1. Model to derive number of cases under different scenarios.

16

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

Table 4 Percent of population exposure to PM10 level under different scenarios in 2010 and 2020 (%) PM10 level (μg/m3)

2010 BC

EE

GAS

WIND

BC

EE

GAS

WIND

0–10 10–20 20–30 30–40 40–50 50–60 60–70 70–80 80–90 90–100 100–110 110–120 120–130 130–140 Total

0.85 8.40 11.23 7.85 5.20 6.09 23.36 36.41 0.62 – – – – – 100

1.62 10.60 12.96 6.58 6.82 24.08 37.33 0.02 – – – – – – 100

3.83 18.50 9.98 19.28 46.58 1.81 0.02 – – – – – – – 100

5.06 19.69 9.60 33.17 32.46 0.02 – – – – – – – – 100

0.03 2.32 5.27 6.50 7.87 5.17 3.40 2.96 4.79 4.31 11.48 30.25 15.63 0.02 100

0.20 5.32 9.42 9.79 4.94 3.85 5.74 14.84 36.72 9.16 0.02 – – – 100

5.09 19.52 10.17 41.60 21.99 1.61 0.02 – – – – – – – 100

6.15 20.07 11.23 58.39 4.07 0.09 – – – – – – – – 100

2020

and E0. The value could be obtained if the following data components are available: exposure-response functions (β), population exposure levels (C and C0), and baseline rate (E0) (Fig. 1). Exposure-response functions (β) link air quality changes and heath outcomes. The preference for this analysis was to select C-R functions from Chinese studies whenever they were available. Only when the selected endpoints could not be found in Chinese literature, the results of international peer-reviewed studies were used. If there were several studies describing the C-R function for the same health

endpoint, we used the pooled estimate to get the mean and 95% confidence interval (CI) of the coefficient. This meta-analysis method was based on the variance weighted average across the results of studies with available quantitative effect estimates (coefficients or relative risks). Studies with lower standard errors had more weight in the resulting joint estimate. Although PM10 was selected as the indicator of air pollution in this analysis, some studies depended on other measures of particulate matter [e.g. total suspended particles (TSP), or particulate matter less than 2.5 μm (PM2.5)] for exposure assessment. Therefore, if

Table 5 Exposure-response coefficients and baseline rate (per person) used in the analysis Health outcome (age group)

Mean (95% CI)

Reference

Frequency

Reference

Long-term mortality (adult ≥30) Chronic bronchitis (all ages) Respiratory hospital admission (all ages) Cardiovascular hospital admission (all ages) Outpatient visits-internal medicine (all ages) Outpatient visitspediatrics (all ages) Acute bronchitis (all ages) Asthma attack (children b15 years) Asthma attack (adults ≥15 years)

0.00430 (0.00260, 0.00610)

Dockery et al. (1993), Pope et al. (1995)

0.01077

0.00450 (0.00127, 0.00773)

Ma and Hong (1992), Jin et al. (2000)

0.01390

0.00130 (0.00010, 0.00250)

Zmirou et al. (1998), Wordley et al. (1997)

0.01240

0.00130 (0.00070, 0.00190)

Wordley et al. (1997), Prescott et al. (1998)

0.00850

0.00034 (0.00019, 0.00049)

Xu et al. (1995)

3.26000

0.00039 (0.00014, 0.00064)

Xu et al. (1995)

0.30000

0.00550 (0.00189, 0.00911)

Jin et al. (2000)

0.39000

Shanghai Municipal Bureau of Public Health (2002) China Ministry of Health (1998) Shanghai Municipal Bureau of Public Health (2002) Shanghai Municipal Bureau of Public Health (2002) Shanghai Municipal Bureau of Public Health (2002) Shanghai Municipal Bureau of Public Health (2002) Wang et al. (1994)

0.00440 (0.00270, 0.00620)

Roemer et al. (1993), Segala et al. (1998), Gielen et al. (1997) Dusseldorp et al. (1995), Hiltermann et al. (1998), Neukirch et al. (1998)

0.06930

Ling et al. (1996)

0.05610

Ling et al. (1996)

0.00390 (0.00190, 0.00590)

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

17

Table 6 Health benefits in different scenarios with respect to BC Scenario in Shanghai in 2010 (mean and 95% CI)

Premature death Chronic bronchitis Respiratory hospital admission Cardiovascular hospital admission Outpatient visits (internal medicine) Outpatient visits (pediatrics) Acute bronchitis Asthma attack

EE

GAS

WIND

2804 (1766, 3811) 5828 (2048, 9069) 1570 (124, 2934) 796 (507, 1080) 111,300 (61,860, 158,400) 11,540 (4201, 18,970) 186,100 (0, 334,700) 3652 (3162, 4134)

7452 (4754, 9994) 15,450 (5558, 23,500) 4269 (341, 7906) 2169 (1385, 2935) 304,600 (169,500, 433,100) 31,590 (11,520, 51,820) 493,700 (0, 858,500) 9585 (8242, 10,890)

8249 (5274, 11040) 17,100 (6175, 25,900) 4745 (380, 8774) 2412 (1541, 3262) 339,000 (188,700, 481,900) 35,150 (12,820, 57,650) 546,400 (0, 944,500) 10,590 (9093, 12,040)

necessary, the following conversions were applied for different particulate matter indicators: PM10 ¼ TSP  0:65 and PM10 ¼ PM2:5 =0:6 (Teng et al., 1999) The final results of this analysis were given as the comparison of health effects under one specific scenario with respect to BC scenario in 2010 and 2020, respectively. 2.5. Economic valuation of the health effects Generally there are two ways to estimate the economic cost of health effects due to air pollution: one is the direct cost approach, the other is the general equilibrium approach. Given the difficulty and uncertainty associated with projecting future economic and technological changes in the general equilibrium approach, we used the direct cost approach. Although the general equilibrium approach could provide many insights that the direct cost approach cannot, it also introduces a significant level of additional uncertainty. Our choice of approach for economic assessment is also in line with other similar assessment (US EPA, 1999). The direct cost approach in our analysis was mainly based on the concept of willingness to pay (WTP), while cost of illness (COI) was also employed as an alternative choice for some morbidity endpoints that could not be valued based on existing WTP literature.

The effect of air pollution on mortality was assessed by using the value of a statistical life (VOSL). The literature on the VOSL, or on willingness to pay to avoid a statistical premature death, however, is mainly from the United States. Due to unavoidable constrains such as limited time and budget, our analysis relied on a contingent valuation study (CVM), which was conducted in Chongqing, China (Wang et al., 2001), for the estimate of Shanghai VOSL. In the Chongqing study, Wang et al. reported an average WTP for saving a statistical life being US $34,750. The marginal effect of income on WTP value was also reported as: with annual income increase of $145.80, the marginal increase for saving a statistical life was $14,550. Therefore, taking the annual income differences between Chongqing and Shanghai residents into account, which were $495.70 and $1234.50 respectively in 2000, we did a conversion based on Chongqing's coefficient between marginal WTP and income, and got an estimate for the VOSL in Shanghai. For different endpoints of morbidity, since no WTP studies on these endpoints are available in China, we used an alternative approach to infer the value from those used by the U.S. EPA after conversion (US EPA, 1999). The ratio for conversion was based on the per capita income of US and Shanghai residents, and the income elasticity was assumed to be 1. The COI was also calculated for the endpoint of hospital admissions and outpatient visits, using actual data from Shanghai. No existing studies based on WTP are available on the endpoints in Shanghai.

Table 7 Health benefits in different scenarios with respect to BC Scenario in Shanghai in 2020 (mean and 95% CI)

Premature death Chronic bronchitis Respiratory hospital admission Cardiovascular hospital admission Outpatient visits (internal medicine) Outpatient visits (pediatrics) Acute bronchitis Asthma attack

EE

GAS

WIND

9870 (5911, 14,170) 20,700 (6636, 35,400) 5057 (387, 9782) 2537 (1600, 3477) 348,600 (193,000, 498,300) 36,210 (130,90, 59,950) 659,700 (0, 1,363,000) 13,580 (12,040, 15,130)

22,210 (13720, 30,860) 46,330 (15,730, 74,980) 12,030 (941, 22,770) 6075 (3856, 8273) 844,200 (468,700, 1,204,000) 87,620 (31,810, 144,400) 1,478,000 (0, 2,819,000) 29,590 (25,870, 33,270)

23,100 (14,300, 32,010) 48,160 (16,410, 77,610) 12,560 (985, 23,730) 6347 (4030, 8639) 882,700 (490,200, 1,258,000) 91,620 (33,270, 150,900) 1,537,000 (0, 2,913,000) 30,700 (26,800, 34,540)

18

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

Table 8 Summary of unit value for various endpoints in 2000 Endpoint

Mean (95% CI)

Premature death

108,500 (101,900, 115,100) Chronic bronchitis 6050 (807, 20,130) Respiratory hospital admission 710 ⁎ Cardiovascular hospital admission 1043⁎ Outpatient visits (internal medicine) 14⁎ Outpatient visits (pediatrics) 14⁎ Acute bronchitis 7.2 (2.6, 11.9) Asthma attack 5.3 (2.3, 8.3)

Approach

source of energy consumption was assessed. Those from other sources, such as natural sources, construction sites etc., were not included.

WTP

3.2. Estimation of health effect WTP COI COI COI COI WTP WTP

⁎ The available data in Shanghai did not provide the distribution of the values.

The economic value of a change in the incidence of a given adverse health outcome was calculated as the change in incidence (e.g. the number of avoidable deaths) multiplied by the unit monetary value (the value of a single case avoided). To deal with the inherent uncertainty in the health and economic impact assessment, the uncertainties in the effect estimates were quantified and the results were given as a range (mean and 95% CI). Since both health outcomes and unit values are distributions rather than constants, we performed the Monte Carlo simulation to calculate the economic values on the Analytic® environment (Lumina Corp, 2006). 3. Results 3.1. Exposure assessment to PM10 of general population in Shanghai Table 4 summarizes the percent of the population exposed to different levels of PM10 under various scenarios in 2010 and 2020, respectively. It should be emphasized that the PM10 levels in Table 4 are much lower than the actual concentrations in Shanghai, because in the present study, only the PM10 from the

Table 5 summarizes the PM10 exposure-response coefficients (mean and 95% CI) and the baseline rates of selected health outcomes in the analysis. The excess cases in each scenario, with respect to BC scenario, are computed based on the change of population exposure levels to PM10 under each scenario, exposure-response functions, and baseline rates for the health outcomes. Tables 6 and 7 show the health benefits in different scenarios with respect to the BC scenario in Shanghai in 2010 and 2020, respectively. It is clear that energy scenarios could have a significant impact on the health status for Shanghai residents in the future. Compared with the BC scenario, implementation of various energy scenarios in Shanghai could prevent 2804–8249 and 9870–23,100 PM10-related avoidable deaths (mid-value) in 2010 and 2020, respectively. 3.3. Economic valuation of the estimated health effects Table 8 summarized the unit values (mean and 95% CI) for various endpoints in 2000 in Shanghai, and the specific approach used in deriving them. Combining the health benefits and unit values described above, we compute the economic benefits under different scenarios with respect to the BC scenario. Tables 9 and 10 show the results in 2010 and 2020, respectively. 4. Discussion Our findings illustrate that low-carbon strategies could play an active role in the reduction of air pollutant emissions, improvement of air quality, and public

Table 9 Economic benefits in different scenarios with respect to BC Scenario in 2010 (millions of 2000 US $) (mean and 95% CI)

Premature death Chronic bronchitis Respiratory hospital admission Cardiovascular hospital admission Outpatient visits (internal medicine) Outpatient visits (pediatrics) Acute bronchitis Asthma attack Total a a

EE

GAS

WIND

450.40 (408.90–495.60) 49.45 (44.89–54.41) 1.65 (1.50–1.82) 1.23 (1.12–1.35) 2.31 (2.09–2.54) 0.24 (0.22–0.26) 2.00 (1.82–2.21) 0.03 (0.03–0.03) 507.31

1197.00 (1087.00–1317.00) 133.10 (120.90–146.50) 4.49 (4.07–4.94) 3.35 (3.04–3.69) 6.31 (5.73–6.95) 0.65 (0.59–0.72) 5.32 (4.83–5.86) 0.08 (0.07–0.08) 1350.30

1325.00 (1203.00–1458.00) 144.90 (131.50–159.40) 4.99 (4.53–5.49) 3.72 (3.38–4.10) 7.03 (6.38–7.73) 0.73 (0.66–0.80) 5.88 (5.34–6.47) 0.08 (0.08–0.09) 1492.33

Summing 5th and 95th percentile values yield a misleading estimate of the 5th and 95th percentile estimate of total health benefits. As a result, we only present the total mean.

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

19

Table 10 Economic benefits in different scenarios with respect to BC Scenario in 2020 (millions of 2000 US $) (mean and 95% CI)

Premature death Chronic bronchitis Respiratory hospital admission Cardiovascular hospital admission Outpatient visits (internal medicine) Outpatient visits (pediatrics) Acute bronchitis Asthma attack Total a

EE

GAS

WIND

2346.00 (1934.00–2841.00) 260.30 (214.50–315.20) 7.87 (6.49–9.53) 5.80 (4.78–7.02) 10.69 (8.82–12.95) 1.11 (0.92–1.35) 10.53 (8.68–12.76) 0.16 (0.13–0.19) 2642.45

5280.00 (4352.00–6394.00) 580.50 (478.50–702.90) 18.72 (15.43–22.66) 13.88 (11.44–16.81) 25.90 (21.35–31.36) 2.69 (2.22–3.26) 23.28 (19.19–28.19) 0.34 (0.28–0.42) 5945.31

5490.00 (4525.00–6648.00) 613.50 (505.70–742.90) 19.54 (16.11–23.66) 14.51 (11.96–17.56) 27.08 (22.32–32.79) 2.81 (2.32–3.40) 24.31 (20.04–29.44) 0.36 (0.29–0.43) 6192.11

a

Summing 5th and 95th percentile values yield a misleading estimate of the 5th and 95th percentile estimate of total health benefits. As a result, we only present the total mean.

health. To our knowledge, this is the first study in China to integrate the analysis of low-carbon development, air pollution and public health. Our estimates may provide supportive evidence to the implementation of lowcarbon strategies in the city in that the resultant public health improvement is substantial in both physical and monetary terms. Actually some of the scenarios we proposed have been under consideration by the local decision-makers and will be implemented in Shanghai in the near future. Quantification of the impact of air pollution on public health has increasingly become a critical component in policy decision. Analyzing the total burden of ambient air pollution on public health in a community remains challenging, given the gaps in scientific knowledge about the health effects of air pollution, and the wide range of uncertainties characterizing many parts of the process. To assess the effects of air pollution, a complex mixture of pollutants, epidemiological studies use several indicators of exposure, (eg, NO2, CO, PM10, TSP, and SO2). These pollutants, however, are correlated. Hence, epidemiological studies cannot strictly allocate observed effects to single pollutants. But a pollutant-by-pollutant assessment would grossly overestimate the impact. Therefore, we selected only one pollutant to derive the public health impact. In this context PM10 is a useful indicator of several sources of outdoor air pollution. Our current estimation of public health impact associated with air pollution under various energy scenarios is conservative for three reasons. First, in the present analysis we only selected PM10 as an indicator of outdoor air pollution, which would probably overlook the adverse health effects due to exposure to other air pollutants, thus underestimating the health effects attributable to total air pollution. Although PM10 may be considered a good indictor of air pollution, there

is clear evidence that other pollutants, such as ozone, nitrogen oxides, and sulfur dioxide etc., may have independent health effects. In addition, we could not include estimates of synergistic effects between air pollutants and cofactors such as pollen and other allergens. Second, as we stated above, the ATMOS model we used could only simulate primary PM10 and SO2, thus leading to underestimation of the health effects attributable to secondary PM10, such as sulfate and nitrate. Previous study has shown that ammonium sulfate and nitrate accounted for substantial ratios of fine particles in Shanghai (Ye et al., 2003). Third, in the selection of relevant health endpoints, we only focused on those outcomes that could be quantitatively estimated and then translated into monetary values for further assessment. Some endpoints (e.g. sub-clinical symptoms, decrease in pulmonary function) were not included in this analysis, although there is evidence for an association between them and air pollution exposure. We did not estimate effects from cancer linked to exposure to ambient air pollution, although a recent cohort study in U.S. has suggested their association (Pope et al., 2002). Fourth, we focused our analysis on ambient air pollution and we did not consider the health impact due to exposure to indoor air pollution. Actually a substantial part of indoor air pollutants might be from outdoor sources in Shanghai. Thus excluding indoor air pollution from our analysis might also underestimate the total health impact under various energy scenarios. Some of the exposure-response functions we employed in this analysis were not available in Chinese studies. So we had to rely on international studies, conducted mostly in the U.S. and Western Europe. This raises the question of transferring the results from a developed country to a developing one. For example, compared to the studies in the U.S. and Western Europe, the Chinese studies generally reported lower coefficients for the exposure-response relationships

20

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21

between air pollution and adverse health effects (Kan et al., 2005). This is probably due to different levels of air pollution, local population sensitivity, age distribution and especially different air pollutant components. For instance the composition of the motor vehicle fleets in Western Europe and the U.S., where most of the epidemiological studies were performed, differs substantially from that in China. This, together with other differences such as the widespread use of coal in China, implies that the air pollution mixture differs substantially between China and the areas where most epidemiologic studies were conducted. Therefore, conceptually, when exposure-response functions from developed countries are applied to other regions, for example — Shanghai, they should be revised, taking into account local conditions, such as the physical (diameter, etc.) and chemical (components) characteristics of particles, social-economical status of local populations, etc. However, no reference data are available for such a revision. Until exposure-response functions derived locally become available, this will probably be the weakest part of this analysis. As no valuation studies on the health endpoints associated with air pollution in Shanghai have been performed, we had to estimate values from previous studies of similar changes. This procedure is often termed as benefit transfer or value transfer in economics. Characteristics of the concerned population, e.g. age distribution, income, health status, culture, may have contextual effects on the valuation results. If we directly transferred the U.S. VOSL into the Shanghai case after considering the income difference, a value of US $780,000 would be yielded, which is much higher than that estimated according to the Chongqing study. The value would be even higher if we used purchasing power parity (PPP) as the income definition here. It is obvious that the Chongqing result is better fitted to the Shanghai estimation than the US result in terms of economic and social situation. Therefore, the present analysis tries to employ the Chinese studies wherever they are available, and attempts to provide a conservative range of reasonable estimates. A major uncertainty that complicates the application of WTP estimates from the study site to the target site in the benefit transfer arises from difference between income levels. One of the fundamental issues in valuing the reduction of risk is that WTP rises with the income. The key question is the determination of income elasticity of the relevant WTP. The literature on the income elasticity of WTP for reducing the risk of damage to health is, however, extremely sparse. Different studies estimated the income elasticity from

0.26 (Loehman and De, 1996) to 1.1 (Viscusi and Evans, 1992). It is important to note the acute sensitivity of the social costs of ill health to the value of this parameter. Using an elasticity of 0.4 and 1.1 makes a difference of nearly 20 times in the final results. In view of the limited data source, it is considered prudent to maintain a degree of conservatism in this valuation exercise. Since there is no suggestive information that could be relied on for elasticity estimate, we have chosen to assume a higher income elasticity of 1 for morbidity costs estimates, so that attention is focused on difference in income. 5. Conclusions Energy and health is one of the biggest challenges for sustainable development in Shanghai during the high economic growth. Despite of the limitations described above, our analysis still emphasis the need to consider air pollution-related health effects as an important impact of low-carbon development strategy in Shanghai. Of course, selection of optimal low-carbon scenarios for Shanghai requires further cost-benefit analysis based on both our estimates and other analyses on the implementation cost of those scenarios. Our analyses also suggest that in a century moving toward sustainable development and health, close collaboration between public health and energy policy will enhance success in preventing avoidable health hazards. Acknowledgement The current project was founded by the U.S. Energy Foundation through grant G-0212-06632. The authors also appreciate the kind assistance of the NIH Fellows Editorial Board in reviewing the manuscript. References Brunekreef B, Holgate ST. Air pollution and health. Lancet 2002;360 (9341):1233–42. Calori G, Carmichael GR. An urban trajectory model for sulfur in Asian megacities: model concepts and preliminary application. Atmos Environ 1999;33:3109–17. China Ministry of Health. Report of the Second National Health Service Survey. Beijing; 1998 [in Chinese]. Cifuentes L, Borja-Aburto VH, Gouveia N, Thurston G, Davis DL. Climate change. Hidden health benefits of greenhouse gas mitigation. Science 2001;293(5533):1257–9. Dockery DW, Pope III CA, Xu X, Spengler JD, Ware JH, Fay ME, et al. An association between air pollution and mortality in six U.S. cities. N Engl J Med 1993;329:1753–9. Dusseldorp A, Kruize H, Brunekreef B, Hofschreuder P, de Meer G, van Oudvorst AB. Association of PM10 and airborne iron with

C. Chen et al. / Science of the Total Environment 373 (2007) 13–21 respiratory health of adults living near a steel factory. Am J Respir Crit Care Med 1995;152:1032–9. ERC (The Energy Research Centre at the University of Cape Town). The application of energy models in developing countries; 2006. Available from : http://www.erc.uct.ac.za/publications/paper%20%20CommendLEAPvMARKALv2.pdf, cited on Oct. 12, 2006. Gielen M, van der Zee S, van Wijnen J, van Steen CJ, Brunekreef B. Acute effect of summer air pollution on respiratory health of asthmatic children. Am J Respir Crit Care Med 1997;155:2105–8. Hiltermann TJ, Stolk J, van der Zee SC, Brunekreef B, de Bruijne CR, Fischer PH, et al. Asthma severity and susceptibility to air pollution. Eur Respir J 1998;11(3):686–93. Jin LB, Qin Y, Xu Z. Relationship between air pollution and acute and chronic respiratory disease in Benxi. Chin J Environ Health 2000;17(5):268–70 [in Chinese]. Kan H, Chen B, Chen C, Wang B, Fu Q. Establishment of exposureresponse functions of air particulate matter and adverse health outcomes in China and worldwide. Biomed Environ Sci 2005;18 (3):159–63. LEAP. Long-range energy alternatives planning system, User guide for LEAP version 2000. Boston, USA: Stockholm Environment Institute (SEI-B), Boston Center, Tellus Institute; 2001. Ling S, Peng J, Wang Z. A sample survey and multiple factor analysis on asthma in urban districts of Shanghai. Chin J Tuberc Respir Disease 1996;19(1):11–3 [in Chinese]. Loehman E, De VH. Application of stochastic choice modeling to policy analysis of public goods: a case study of air quality improvement. Rev Econ Stat 1996;64(3):474–80. Lumina Corp. Analytica: a software tool for uncertainty analysis and model communication; 2006. Available from http://www. lumina.com/software/ch10.9.PDF. Cited on Oct 21, 2006. Ma H, Hong C. Effects of particulate air pollution on respiratory disease. Chin J Public Health 1992;11(4):229–32 [in Chinese]. Neukirch F, Ségala C, Le Moullec Y, Korobaeff M, Aubier M. Shortterm effects of low-level winter pollution on respiratory health of asthmatic adults. Arch Environ Health 1998;53:320–8. Pope CA, Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, et al. Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am J Respir Crit Care Med 1995;151:669–74. Pope CA, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, et al. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 2002;287:1132–41.

21

Prescott G, Cohen GR, Elton RA, Fowkes FG, Agius RM. Urban air pollution and cardiopulmonary ill health: a 14.5 year time series study. Occup Environ Health 1998;55:697–704. Roemer W, Hoek G, Brunekreef B. Effects of ambient winter air pollution on respiratory health of children with chronic respiratory symptoms. Am Rev Respir Dis 1993;147:118–24. Segala C, Fauroux B, Just J, Pascual L, Grimfeld A, Neukirch F. Shortterm effect of winter air pollution on respiratory health of asthmatic children in Paris. Eur Respir J 1998;11:677–85. Shanghai Municipal Bureau of Public Health. Shanghai health yearbook 2001. Shanghai: Shanghai Science and Technology Literature Publishing House; 2002 [in Chinese]. Teng E, Hu W, Wu G, Wei F. The composing characteristics of elements in coarse and fine particle in air of the four cities in China. China Environ Sci 1999;19:238–42 [in Chinese]. US EPA. The benefits and costs of the Clean Air Act 1990 to 2010, Appendix H 21-26. Washington D.C.: US EPA Office of Air and Radiation; 1999. Viscusi WK, Evans WN. Utility functions that depend on health status: estimates and economic implication. Am Econ Rev 1992;80 (3):353–74. Wang J, Wang Q, Bi Z. A survey on acute respiratory disease. Chin J Epidemio 1994;15(3):141–4 [in Chinese]. Wang H, Mullahy J, Chen D, Wang L, Peng R. Willingness to pay for reducing the risk of death by improving air quality: a contingent valuation study in Chongqing, China. Presented at the third international health economic association conference in York, UK; 2001. Wang B, Chen C, Huang C, Zhao J, Dai Y. Local air pollutant and CO2 emission scenarios under low carbon development: Shanghai case study. Energy Res Info 2004;20(3):137–45 [in Chinese]. Wordley J, Walters S, Ayres J. Short term variations in hospital admissions and mortality and particulate air pollution. Occup Environ Health 1997;54:108–16. Xu X, Dockery D, Christiani D, Li B, Huang H. Association of air pollution with hospital outpatient visits in Beijing. Arch Environ Health 1995;50(3):214–20. Ye B, Ji X, Yang H, Yao X, Chan CK, Cadle SH, et al. Concentration and chemical composition of PM2.5 in Shanghai for a 1-year period. Atmos Environ 2003;37:499–510. Zmirou D, Schwartz J, Saez M, Zanobetti A, Wojtyniak B, Touloumi G, et al. Time-series analysis of air pollution and cause-specific mortality. Epidemiology 1998;9:495–503.