CO C O N F E R E N C E E D I T I O N ST 39236 O F COST OF POLLUTION PO LLU IN CHINA ECONOMIC ESTIMATES OF PHYSICAL DAMAGES T IO N IN CHI N A The State Environmental Protection Administration 115 Xizhimen Nanxiaojie, Beijing 100035, P. R. China Tel: 86 (10) 6653.2331 Fax: 86 (10) 6653.2424 www.sepa.gov.cn Rural Development, Natural Resources and Environment Management Unit, East Asia and Pacific Region, The World Bank 1818 H Street, NW, Washington DC 29433, USA Tel: + 1 (202) 458.4073 Fax: + 1 (202) 477.2733 www.worldbank.org/eapenvironment THE The World Bank Office, Beijing 16th Floor, China World Tower 2 WORLD No. 1 Jianguomenwai Avenue Beijing 100004, P. R. China Tel: + 86 (10) 5861.7600. Fax: + 86 (10) 5861.7800. BANK www.worldbank.org.cn THE GOVERNMENT OF THE THE WORLD BANK PEOPLE'S REPUBLIC OF CHINA Environmental and Social Development Unit East Asia & Pacific Region Ph: 202-458-5660 Fax: 202-522-1666 e-mail: jnygard@worldbank.org COST OF POLLUTION IN CHINA ECONOMIC ESTIMATES OF PHYSICAL DAMAGES The World Bank State Environmental Protection Administration, P. R. China This publication is available online at www.worldbank.org/eapenvironment. Front cover photos: John D. Liu. From the film "A Green Call" prepared by the Environmental Edu- cation Media Project in Beijing in cooperation with the World Bank. Cover design: Circle Graphics, Jostein Nygard Rural Development, Natural Resources and Environment Management Unit East Asia and Pacific Region The World Bank Washington, D.C. February, 2007 This volume is a product of an expert team from China, international experts from various countries and the staff of the World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of the World Bank of the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions of all of this work without permission may be a violation of applicable law. The World Bank encourages dissem- ination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopying or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Table of Contents ACKNOWLEDGMENTS v ABBREVIATIONS AND ACRONYMS vii FOREWORD ix EXECUTIVE SUMMARY xi 1 Overview 1 2 Health Impacts of Ambient Air Pollution 19 3 Health Impacts of Water Pollution 33 4 Valuation of Environmental Health Risks 67 5 Non-Health Impacts of Water Pollution 79 6 Non-Health Impacts of Air Pollution 111 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N iii Acknowledgments This Report is the result of a collaborative worked on statistical health data. A team from research effort by a joint Chinese and interna- the Rural Water Supply Central Disease Control tional expert team being contracted by the (CDC) lead by Fan Fucheng and Tao Yong pro- World Bank. vided assistance on particularly drinking water In Beijing, the overall team was lead by Guo and partly health related data. Xiaomin, a senior advisor to the State Environ- In Shanghai, a team lead by Prof. Peng Xizhe mental Protection Administration (SEPA). His at the Fudan University, included Chen Yan, Tian team included Yu Fang from the China Acad- Wenhua and Cheng Yuan. In Chongqing, a team emy of Environmental Planning (CAEP), who lead by Ass. Prof. Chen Gangcai at the Chongqing has handled the overall technical coordination, Academy of Environmental Science included Zhou Guomei coordinated the Valuation of Wang Fei, Ran Tao, Zhou Zhien, Liu Lanyu, and Environmental Health Risk (VEHR) study Chen Derong in addition to Yang Xioalin, Xiang together with Zhang Kai, Zhou Jun and Wu Xinzhi and Qin Lei from Chongqing CDC and Yuping from the Policy Research Center for Tang Guil from Chongqing MoH. Environment & Economy. Pan Xiaochuan at International experts have included Haakon the Medical College of Peking University lead a Vennemo and Henrik Lindhjem (ECON), team on dose response function development, Kristin Aunan and Hans Martin Seip (CICERO), which included Wang Lihua, and Jiang Jinhua. Alan Krupnick, Sandy Hoffmann and Michael Monitoring data was provided by the China McWilliams (RFF), Bjorn Larsen and Ramon National Monitoring Centre by Zhuo Jianping, Ortiz (independent consultants). Ling Lixin, Fu Deqing and WuHuaimin. Zhao At the World Bank, the project was coordi- Yaoming has participated from the Ministry of nated by Jostein Nygard, task team leader Agriculture. (EASRE) under the overall supervision of A team from the Water Resources and Magda Lovei (EASOP). Substantive inputs were Hydropower Planning and Design Institute of provided by Maureen Cropper (DEC), Tamer the Ministry of Water Resources (MWR) lead Samah Rabie (ECSHD), while technical sup- by Li Yuanyuan, which also included Zhou Zhi- port was provided by Marija Kuzmanovic and wei, Cao Jianting and Zhangwei provided assis- Andrew Murray (EASEN/EASRE). tance on water scarcity subjects. Gao Jun and The current report has mainly be written by Xu Ling from the Ministry of Health (MoH) Maureen Cropper, Tamer Rabie, Haakon Ven- C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N v A C K N O W L E D G M E N T S nemo, Kristin Aunan, Hans Martin Seip, Yu EASRE), Anjali Acharya and Giovanni Ruta Fang, Guo Xiaoming and Jostein Nygard, while (ENV) and Charles E. Di Leva (LEGEN). the extensive Chinese expert team has mainly Coordination of the study within SEPA, has been writing the progress and background been made by their Foreign Economic Cooper- reports that this report builds upon. The RFF, ation Office (FECO) with Wang Xin and Xie Shanghai and Chongqing teams have been writ- Yongming. Personnel within SEPA's Planning ing the "Willingness to Pay for Reduced Mortal- and Finance, Pollution Control and Science and ity Risk Reduction in Shanghai and Chongqing" Technology departments in addition MoH per- study, which is also being published as a sepa- sonnel have reviewed the report extensively. rate World Bank discussion paper report. The report was edited by Robert Livernash, Mainly based upon work by Bjorn Larsen, a consultant. Circle Graphics designed and man- separate discussion paper report "China Health aged desktopping. Production was supervised by Effects of Indoor Air Pollution" is also being Jaime Alvarez. Photos provided by John Liu, the published. Environmental Education Media Project, from Peer reviewers included Chris Nielsen (Har- a World Bank-contracted film "A Green Call". vard University), Hao Jiming (Tsinghua Uni- Chinese translation was provided by the transla- versity), Kseniya Lvovsky (World Bank, tion desk at SEPAs Department of International SASES), Rita Klees (World Bank, ENV), and Cooperation. Anil Markandy (ECSSD). Additional reviews Finally, we would like to express our gratitude and comments were provided by David Dollar, to the Government of Norway and Finland, Bert Hofman and Andres Liebenthal (World which provided the main trust funds (TFESSD) Bank, Beijing), Maria Teresa Serra (EASES/ to carry out the study. The study was also sup- EAPVP) Julien Labonne and Jian Xie (EASES/ ported by the World Bank's own funding. vi C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N Abbreviations and Acronyms ACS American Cancer Society AHC Adjusted Human Capital BOD Biological Oxygen Demand BOH Bureau of Health (at local levels) CAEP Chinese Academy for Environmental Planning CAES Chongqing Academy of Environmental Sciences CDC Center for Disease Control and Prevention CECM Chinese Environmental Cost Model CEVD Cerebrovascular Disease CNHS China National Health Survey CO Carbon Monoxide COD Chemical Oxygen Demand COI Cost of Illness COPD Chronic Obstructive Pulmonary Disease CSMI Clear Water and Sewage Mixed Irrigation CV Contingent Valuation CVD Cardiovascular Disease DALY Disability-Adjusted Life Year DC Dichotomous Choice Method DSP Disease Surveillance Point ECM Environmental Cost Model EU European Union EV Emergency Visit GDP Gross Domestic Product GIOV Gross Industrial Output Value HEI Health Effects Institute HH Household ICD International Classification of Disease C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N vii A B B R E V I A T I O N S A N D A C R O N Y M S IWQI Integrated Water Quality Index MoA Ministry of Agriculture MoH Ministry of Health MWR Ministry of Water Resources NAPAP National Acid Precipitation Assessment Program NBS National Bureau of Statistics NOx Nitrogen Oxides O3 Ozone OPV Outpatient Visit OR Odds Ratio PC Payment Card Method PM Particulate Matter PM10 Particulate Matter of Less than 10 m in diameter PPP Purchasing Power Parity PSI Pure Sewage Irrigation QALY Quality Adjusted Life Year RD Respiratory Disease RFF Resources for the Future RMB Chinese Currency, Yuan RR Relative Risk SCE Standard Coal Equivalent SEPA State Environmental Protection Administration SO2 Sulphur Dioxide TSP Total Suspended Particulates TVEs Town and Village Enterprises UNEP United Nations Environmental Programme USEPA United States Environmental Protection Agency VEHR Valuation of Environmental Health Risk VSL Value of Statistical Life WHO World Health Organization WTP Willingness to Pay viii C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N Foreword to the Conference Edition This is a draft edition of the Cost of Pollution in of the economic impacts of air and water pollu- China: Economic Estimates of Physical Damages tion in China, to provide relevant policy infor- report, which will be presented at the interna- mation to decision makers and to enable the tional conference on Sustainable Development in Chinese government to make optimal resource Beijing, China on March 2, 2007. The purpose allocations for environmental protection. of this conference edition is to present the findings Prior to the publication of this report, com- of the studies undertaken in China over the prehensive comments have been received by past about 3 years as well as to obtain relevant both the Chinese Government, particularly the comments and feedback from the conference State Environmental Protection Administration participants that could be included in the final (SEPA) and independent Chinese and Non- edition of the report. Chinese reviewers. Some of the subjects that This report traces its origin to 1997, when the have been carefully developed during the course World Bank published the China 2020 ­ Clear of implementation, including certain physical impact estimations as well as economic cost cal- Water Blue Skies report. This work underscored culations at local levels have been left out of this the economic implications of environmental conference edition due to still some uncertain- degradation by estimating that the cost of air and ties about calculation methods and its applica- water pollution in China is between 3.5 and 8 tion. How to possibly make use of these percent of GDP. Following these findings, the materials will be continuously worked on during Chinese government requested the World Bank and after the conference. Moreover, the com- to collaborate with a number of Chinese and prehensive reference material that has been international research institutes to develop an developed by joint Chinese and International environmental cost model (ECM) using expert team (including progress reports and var- methodologies specific to the China context. ious background reports), is going to be attached This work includes an in-depth review of in a CD-ROM in the final edition. international ECM studies, and development Wish you good reading of this edition and and application of new methodologies (and soft- looking forward to receiving your comments. ware) for annual estimations of water and air pollution in China at both central and local lev- Report Authors els. The aim of this work is to increase awareness February 2007 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N ix Executive Summary In recent decades, China has achieved Rapid Economic Growth Has Had Positive Environmental rapid economic growth, industrializa- Impacts but Also Created New Environmental Challenges tion, and urbanization. Annual in- Considering China's strong economic growth over the last 20­25 years, there creases in GDP of 8 to 9 percent have is no doubt that it has had positive impacts on the environment. Along- lifted some 400 million people out of side economic growth, technology improvements over this period have cre- dire poverty. Between 1979 and 2005, ated much-improved resource utilization. Energy efficiency has improved China moved up from a rank of 108th drastically--almost three times better utilization of energy resources in to 72nd on the World Development 2000­02 compared to 1978. As a result of the changing industrial structure, Index. With further economic growth, the application of cleaner and more energy-efficient technologies, and pollu- most of the remaining 200 million tion control efforts, ambient concentrations of particulate matter (PM) and people living below one dollar per day sulfur dioxide (SO2) in cities have gradually decreased over the last 25 years. may soon escape from poverty. Al- Implementation of environmental pollution control policies--particularly though technological change, urban- command-and-control measures, but also economic and voluntarily ization, and China's high savings rate measures--have contributed substantially to leveling off or even reducing suggest that continued rapid growth pollution loads, particularly in certain targeted industrial sectors. is feasible, the resources that such At the same time, new environmental challenges have been created. Fol- growth demands and the environmen- lowing a period of stagnation in energy use during the late 1990s, total energy tal pressures it brings have raised consumption in China has increased 70 percent between 2000 and 2005, grave concerns about the long-term with coal consumption increasing by 75 percent, indicating an increasingly sustainability and hidden costs of energy-intensive economy over the last few years. Moreover, between 2000 growth. Many of these concerns are and 2005, air pollution emissions have remained constant or, in some associated with the impacts of air and instances, have increased. The assessment at the end of the tenth five-year water pollution. plan (2001­05) recently concluded that China's emissions of SO2 and soot were respectively 42 percent and 11 percent higher than the target set at the beginning of the plan. China is now the largest source of SO2 emissions in the world. Recent trends in energy consumption, particularly increased coal use, provide a possible explanation for the increase in SO2 emissions. Water pollution is also a cause for serious concern. In the period between 2001 and 2005, on average about 54 percent of the seven main rivers in China contained water deemed unsafe for human consumption. This repre- C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N xi E X E C U T I V E S U M M A R Y sents a nearly 12 percent increase since the early pollution, it is striking that the areas with the 1990s. The most polluted rivers occurred in the highest per capita exposure are almost all located northeast in areas of high population density. in northern China (Qinghai, Ningxia, Beijing, The trends in surface water quality from 2000 to Tianjin, Shaanxi, and Shanxi). The exception is 2005 suggest that quality is worsening in the Hunan, which is located in the South. In Fig- main river systems in the North, while improv- ure 1, the color of the provinces on the map ing slightly in the South. This may partly be the shows the percentage of the urban population result of rapid urbanization (the urban popula- exposed to air pollution, while the bars indicate tion increased by103 million countrywide from the absolute number of people exposed. 2000 to 2005), which caused COD loads from Similarly, the most severely polluted water urban residents to increase substantially and, basins--of the Liao, Hai, Huai, and Songhua hence, surpass the planned targets for 2005. rivers--are also located in northern China (see Rapid industrialization probably also plays a part. figure 2 for surface water quality). North China also has serious water scarcity problems. Some provinces--including Beijing, Shanxi, Ningxia, Northern China Bears a Double Tianjin, and Jiangsu--seem to face the double Burden from Air and Water Pollution burden of exposure to high levels of both air and While the most populous parts of China also water pollution. However, while air pollution have the highest number of people exposed to air levels may be directly associated with population F I G U R E 1 . Urban Population Exposed to PM10 levels, 2003 Heilongjiang Neimeng Jilin Xinjiang Liaoning Gansu Beijing Tianjin Ningxia Hebei Shanxi Shandong Qinghai Shaanxi Henan Jiangsu Anhui Shanghai Sichuan Hubei Zhejiang Chongqing Hunan Jiangxi Guizhou Fujian Guangdong Yunnan Guangxi Pollution Exposure Population Exposed to Pollution Hainan 0 - 10% 11 - 30% 200,000 31 - 45% 46 - 60% 61 - 70% 71 - 80% 81 - 90% 91 - 100% xii C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N E X E C U T I V E S U M M A R Y F I G U R E 2 . Water Quality Levels, 2004 exposure, the same does not necessarily apply to GDP. This assumes that premature deaths are surface water pollution. This is because popula- valued using the present value of per capita GDP tions generally have different drinking water over the remainder of the individual's lifetime. sources that may allow them to escape high levels If a premature death is valued using a value of of contamination. About 115 million people in a statistical life of 1 million yuan, reflecting rural China rely primarily on surface water as people's willingness to pay to avoid mortality their main source of drinking water. Surface risks, the damages associated with air pollution water as a drinking water source is more vulner- are 3.8 percent of GDP. These findings differ in able to possible pollution compared to other, two important ways from previous studies of the safer drinking sources. burden of outdoor air pollution in China. First, they are based on Chinese exposure-response functions, as well as on the international litera- Air and Water Pollution ture; and second, they are computed for indi- have Severe Health Impacts vidual cities and provinces. Previous estimates According to conservative estimates, the eco- by WHO (Cohen et al. 2004) were based on nomic burden of premature mortality and the assumption that increases in PM beyond morbidity associated with air pollution was 100 g/m3 of PM10 caused no additional health 157.3 billion yuan in 2003, or 1.16 percent of damage.( In the base case considered by WHO, C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N xiii E X E C U T I V E S U M M A R Y relative risk does not increase beyond 50 g/m3 piped water is significantly associated with excess of PM2.5, which is approximately equivalent to cases of diarrheal disease and deaths due to diar- 100 g/m3 of PM10.) This assumption implies rheal disease in children under 5 years of age. that the WHO estimates cannot be used to eval- Although there are many indications that surface uate the benefits of specific urban air pollution and drinking water pollution problems con- control policies. tribute to serious health impacts, the lack of Two-thirds of the rural population is without monitoring data on specific pollutants and data piped water, which contributes to diarrheal disease on household behavior regarding avoiding expo- and cancers of the digestive system. The cost of these sure to polluted drinking water make it difficult health impacts, if valued using a VSL of 1 million, to quantify all of the health effects of water pol- are 1.9 percent of rural GDP. Analysis of data lution. Specifically, the lack of exposure data from the 2003 National Health Survey indicates makes quantifying the relationship between that two-thirds of the rural population does not chemical and inorganic pollution and the inci- have access to piped water. The relationship dence of chronic diseases almost impossible. Pre- between access to piped water and the incidence liminary estimates suggest that about 11 percent of diarrheal disease in children under the age of of cases of cancer of the digestive system may be 5 confirms this finding: the lack of access to attributable to polluted drinking water. More F I G U R E 3 . Rural Households with No Access to Piped Water & Diarrhea Incidence Rural HH NTW by County 0 - 3458 3459 - 7800 7801 - 13574 13575 - 21886 21887 - 41341 Incidence of Diarrhea by Province 0 - 72,061 Counties with no shading were 72,062 - 208,769 categorized as 'Urban' or 208,770 - 393,469 'Urban Center with Rural 393,470 - 633,312 Surroundings', which account 633,313 - 893,222 xiv C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N E X E C U T I V E S U M M A R Y attention, however, needs to be given at the pol- proportionately on the less economically ad- icy level to reinforcing the surveillance capacity vanced parts of China, which have a higher share for chronic exposures and disease incidence. of poor populations. As shown in Figure 1, Ningxia, Xinjiang, Inner Mongolia, and other low-income provinces are more affected by air Health is Highly Valued pollution on a per capita basis than high-income by the People in China provinces such as Guangdong and other The mortality valuation surveys conducted in provinces in the southeast. Shanghai and Chongqing as part of this study From another perspective, analysis of the suggest that people in China value improve- 2003 National Health Survey showed that ments in health beyond productivity gains. The 75 percent of low-income households in rural value of a statistical life estimated in these China with children under 5 years of age have no surveys--the sum of people's willingness to pay access to piped water, compared to 47 percent in for mortality risk reductions that sum to one sta- the higher-income categories. This implies that tistical life--is approximately 1 million yuan. low-income households rely more on other This number supports results of other studies, drinking water sources. In fact, about 32 percent which suggest that the value of an avoided death of households within the lowest income quartile is greater than what is implied by the adjusted rely primarily on surface water as their primary human capital approach, which is approximately source of drinking water, compared to 11 per- 280,000 Yuan in urban areas. Evaluation of the cent in the highest income quintile. This health losses due to ambient air pollution using means that the rural poor are at a substantially willingness-to-pay measures raises the cost to higher risk from surface water pollution than 3.8 percent of GDP. the non-poor. It is remarkable that the willingness to pay is The fact that water quality in the North is about the same in locations as different as worse than in the South may explain the slightly Shanghai and Chongqing, which differ greatly higher diarrheal prevalence seen in lower income in per capita GDP with a ratio as high as 5:1. groups in northern China (2.1 percent) com- (However, sample per capita incomes showed a pared to southern China (1.9 percent). How- more modest ratio of 2:1.) Furthermore, these ever, when focusing on differences between new findings illustrate that the urban Chinese income groups in the North, the data clearly population has a willingness to pay to reduce show that the poor (lowest income quartile) have mortality risk comparable in PPP terms to the a much higher diarrheal prevalence (2.4 percent) levels seen in several developed countries with in households using surface water compared to much higher per capita incomes. This means the highest income groups, where no diarrhea that the Chinese people highly value their health cases have been recorded. status and their longevity. Pollution Exacerbates Water China's Poor Are Scarcity, Costing Disproportionately Affected by 147 Billion Yuan a Year Environmental Health Burdens Water scarcity is a chronic problem, especially in Although the objective of this study was not to the North. It is closely related to problems of compare the impacts of air and water pollution water pollution. Surface water pollution has put on the poor versus the non-poor, the findings pressure on the use of groundwater for agricul- suggest that environmental pollution falls dis- tural and industrial purposes. The depletion of C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N xv E X E C U T I V E S U M M A R Y F I G U R E 4 . Groundwater Depletion and Polluted Water Supply Ground Water Depletion & Polluted Water Supply, 2003 N W E S The sum of groundwater depletion and polluted water supply (in 100 million cubic meters) 0 - 10 10 - 20 20 - 30 30 - 50 >50 nonrechargeable groundwater in deep freshwater Air and Water Pollution aquifers imposes an environmental cost, since it Cause Significant Crop depletes a nonrenewable resource and increases and Material Damage future costs of pumping groundwater. It can also lead to seawater intrusion and land subsidence. This study makes clear that the impacts of air Estimates of the cost of groundwater deple- and water pollution on health are severe in both tion suggest that it is on the order of 50 billion absolute and in economic value terms. Although yuan per year, while estimates of the costs of we acknowledge that not all non-health-related using polluted water to industry are comparable impacts can be quantified, the impacts of pollu- in magnitude, bringing the overall cost of water tion on natural resources (agriculture, fish and scarcity associated with water pollution to forests) and manmade structures (e.g. buildings) 147 billion yuan, or about 1 percent of GDP. are estimated to account for substantially lower These new findings indicate that the effects of damages in economic terms. water pollution on water scarcity are much more Acid Rain costs 30 billion yuan in crop damage severe than previous studies have estimated. and 7 billion in material damage annually. It is xvi C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N E X E C U T I V E S U M M A R Y estimated that acid rain, caused mainly by The figures presented in the summary table at increased SO2 emissions due to increased fos- the end of this chapter suggest that outdoor air sil fuel use--causes over 30 billion yuan in dam- pollution poses a very serious problem in urban ages to crops, primarily vegetable crops (about areas. This is not surprising when one compares 80 percent of the losses). This amounts to the levels of ambient PM10 in Chinese cities with 1.8 percent of the value of agricultural output. other large cities across the world. With annual Damage to building materials in the South average PM10 concentrations of over 100Ìg/m3, imposed a cost of 7 billion yuan on the Chinese several selected cities in both northern and economy in 2003. In addition to the human southern China are among the most polluted health effects reported above, these damages pro- cities in the world (see figure 5). vide an additional impetus for controlling SO2. Although the health damages associated with Damages to forests could not be quantified due water pollution are smaller, in total, and as a per- to lack of monitoring data in remote areas and cent of rural GDP, they are still 0.3 percent of adequate dose-response functions. rural GDP if conservatively valued and 1.9 per- Six provinces account for 50 percent of acid rain cent of rural GDP when valued using a 1 million effects. The burden of damages from acid rain yuan VSL. Both figures ignore the morbidity is also unevenly distributed. Over half of the associated with cancer and therefore underesti- estimated damages to buildings occur in three mate the health costs associated with water pol- provinces: Guangdong (24 percent), Zhejiang lution. However, relative to other developing (16 percent), and Jiangsu (16 percent). Almost half countries, China's diarrheal prevalence in rural of the acid rain damage to crops occurs in three areas is quite low, actually lower than in coun- provinces: Hebei (21 percent), Hunan (12 per- tries where a larger percentage of the rural pop- cent), and Shandong (11 percent). However, the ulation has access to piped water supply (see impacts of acid rain extend across international figure 6). boundaries and also affect neighboring countries. Irrigation with polluted water costs 7 billion yuan per year. This study has quantified part of The Benefits of Sound Policy the damage caused by the use of polluted water Interventions May Exceed the Costs for irrigation in agriculture and a portion of the impact of water pollution on fisheries. The This study report shows that the total cost of air impact of irrigating with polluted water in desig- and water pollution in China in 2003 was 362 nated wastewater irrigation zones--considering billion yuan, or about 2.68 percent of GDP for only the impact on yields and produce quality, the same year. However, it should be noted that but not on human health--was estimated to this figure reflects the use of the adjusted human reach 7 billion yuan in 2003. capital approach, which is widely used in Chi- The cost to fisheries is estimated at 4 billion nese literature, to value health damages. If the yuan. The impact of acute water pollution inci- adjusted human capital approach is replaced by dents on commercial fisheries is estimated at the value of a statistical life (VSL) based on stud- approximately 4 billion yuan for 2003. The ies conducted in Shanghai and Chongqing, the impact of chronic water pollution on fisheries amount goes up to about 781 billion yuan, or could not be estimated for lack of exposure data about 5.78 percent of GDP. as well as adequate dose-response information. Setting priorities for cost-effective interventions. Air Pollution Poses a Large Health Risk in Interventions to improve the environment in Urban Areas and Water Pollution a Significant China are likely to yield positive net benefits. Health Risk in Rural Areas Indeed, one of the advantages of the environ- C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N xvii E X E C U T I V E S U M M A R Y F I G U R E 5 . Annual average PM10 concentrations observed in selected cities worldwide, 2004, 2005 Source: China Environmental Yearbook 2005 and WHO 2005. mental cost model developed in this project is sure could be calculated, using the techniques that it can be used to evaluate the benefits of developed in this study, and compared with the specific pollution-control policies and assist in costs. designing and selecting appropriate targeted Targeting high-risk areas. The findings from intervention policies. Once the impact on this project suggest that a focus on northern ambient air quality of a policy to reduce partic- China is essential, particularly the North China ulate emissions has been calculated, the tools Plain and areas located northeast and northwest used to calculate the health damages associated of the plain, where the study shows that there is with particulate emissions can be used to com- a double burden from both air and water pollu- pute the benefits of reducing them. To illus- tion. This problem is further magnified by the trate, researchers have examined the costs and presence of disparities between the poor and impacts on ambient air quality of measures to non-poor. On this basis, it seems relevant that control SO2 emissions and fine particles in stronger policy interventions should be de- Shijiazhuang, the capital of Hebei Province veloped to address air and water pollution (Guttikunda et al. 2003). The monetized value problems. In addition, these efforts should be of the health benefits associated with each mea- complemented with emphasis on improving xviii C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N E X E C U T I V E S U M M A R Y Figure 6. Diarrheal Prevalence and Access to Piped Water Supply 100 90 water 80 piped 70 no w/ 60 50 households 40 rural of 30 20 Percentage 10 0 4 5 00 a 2000 2003 2/2003 2003 200 03/ a 2003 2003 2003 20 2003 200 2000 2002 bique 200 20 Peru Ghana Keny o 2003-2004 China Bolivia Egypt Jordan Cambodi occ Philippines Colombia Mozam agascar Indonesia Mor Mad Piped water Diarrhea Prevalence last 2 weeks Source: ORC Macro, 2006. MEASURE DHS STATcompiler. http://www.measuredhs.com, July 3 2006. access to clean water, with a specific focus on the concrete knowledge of the environmental im- lowest income groups. pacts and costs. By providing new, quantitative Responding to people's concerns. This study information based on Chinese research under suggests that the Chinese value the avoidance of Chinese conditions, this study has aimed to health risks beyond productivity gains. This reduce this information gap. At the same time, implies that people's preference for a clean envi- it has pointed out that substantially more infor- ronment and reduced health risks associated mation is needed in order to understand the with pollution are stronger than past policies health and non-health consequences of pollu- appear to have acknowledged. Growing con- tion, particularly in the water sector. It is criti- cerns about the impacts of pollution are increas- cally important that existing water, health, and ingly expected to guide national policies as well environmental data be made publicly available as local actions. Public disclosure of envi- so the fullest use can be made of them. This ronmental information such as emissions by would facilitate conducting studies on the polluting enterprises, as well as ambient envi- impacts of water pollution on human and ani- ronmental quality data by local authorities, mal health. Furthermore, surveillance capacity at could be an important tool for responding to the local and national levels needs to be people's concerns and creating incentives for expanded to improve the collection of environ- improving local conditions. mental data, especially data on drinking water Addressing the information gap. Past policies quality. These efforts will further improve the and decisions have been made in the absence of analysis begun in this project. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N xix E X E C U T I V E S U M M A R Y Developing an environmental-health action presented in this report. The plan should include plan. At present, an environmental-health action a focus on the geographical areas identified in plan is being jointly drafted by the State Envi- northern China, where there is a double burden ronmental Protection Administration (SEPA) of both air and water pollution. Furthermore, and the Ministry of Health (MoH). This plan particular focus should be put on areas where should take into consideration the mortality and poor populations are adversely affected from morbidity impacts from water and air pollution lack of access to clean water and sanitation. xx C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 1 Overview AIR AND WATER To illustrate, China has not been able to meet 10 of its 13 critical 10th five- POLLUTION IN CHINA year-plan targets for air and water pollution control (see table 1.1). The most pressing off-target performance is the drastic increase in industrial-based SO2 In the last 25 years, China has achieved rapid economic growth, emissions, which has reversed the downward trend in SO2 levels, and industrialization, and urbanization, degraded air quality and the increase in domestic COD loads, which have with annual increases in GDP of caused water quality to deteriorate. 8 to 9 percent. During the same China is the world's second largest energy consumer after the United period, advances in technology States. Almost 68 percent of its energy comes from coal, much of which is and economic efficiency, coupled with pollution control policies, have T A B L E 1 . 1 Environmental Targets for the 10th Five Year Plan vs. positively affected air and water Environmental Performance (million tons) pollution loads. However, great challenges remain in further Actual 2005 Comparison improving China's environmental Actual Planned (completed by with Planned status. Indicators 2000 2005 6/17/06) 2005 (+/- %) Air Pollution SO2 emissions 19.9 17.9 25.5 42 Industry 16.1 14.5 21.7 50 Domestic 3.8 3.5 3.8 9 Soot Emissions 11.7 10.6 11.8 11 Industry 9.5 8.5 9.5 12 Domestic 2.1 2.1 2.3 10 Industrial Dust Emissions 10.9 8.98 9.1 1 Water Pollution COD discharge 14.5 13.0 14.1 8 Industry 7.0 6.7 5.5 -18 Domestic 7.4 6.5 8.6 32 Ammonia Nitrogen 1.8 1.65 1.5 -9 Industry 0.8 0.7 0.525 -25 Domestic 1.1 0.9 0.973 8 Source: Estimations based upon China Environmental Yearbook 2001 and 2006, the 10th Five Year Plan for Environmental Protection and status of the China environment report, 2005 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 1 O V E R V I E W burned in thermal power plants or in industrial medium-sized Chinese cities, beginning in boilers. This has led to continuously high levels 1980). (The averages in each year are arithmetic of SO2 and particulate air pollution. In addition, averages--unweighted by population--of avail- water pollution and water scarcity problems are able readings for "major cities." The set of cities also very severe, particularly in North China, varies from 53 to 97, depending on the year.) Sep- where the region faces some of the most severe arate averages are reported for northern and water quality and quantity challenges in the world southern cities. Suspended particulate levels are today. This section provides a brief overview of higher in northern cities, due in part to industrial these challenges. activity, but also to geographic and meteorologi- cal conditions that make these cities more vulner- able to particulate pollution than cities in the Air Pollution Trends south of China, holding emissions constant Although levels of SO2 and particulates have (Pandey et al. 2005). In both northern and south- declined since the 1980s, China's cities still rank ern cities, particulate concentrations show a among the most polluted in the world. Figure 1.1 downward trend from 1980 until the early 1990s shows trends in annual average total suspended and then remain relatively flat. Sulfur dioxide and particulates (TSP, SO2, and NOx in large and NOx concentrations also show a downward trend F I G U R E 1 . 1 Ambient Air Pollution Levels in China's Major Cities (annual averages) Compared to Chinese Class II Air Quality Standards Total Suspended Particulates (µg/m3) Nitrogen Oxides[1] (µg/m3) 250 3,000 2,500 200 2,000 150 3 3 gm 1,500 gm 100 1,000 50 500 0 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 1980 1983 1986 1989 1992 1995 1998 2001 2004 Sulfur Dioxide (µg/m3) 600 Average of Southern Cities 500 Average of Northern Cities 400 Annual Average Standard 3 24-hour Average Standar 300 g/m Average 200 Vertical bars indicate ranges of values for all cities; the highest horizontal mark shows the most polluted of the Chinese cities. 100 0 [1] In the Nitrogen Oxides chart, data for 2001 and 2004 are for NO2. 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Source: China Environmental Year Books 2004 & 2005 2 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W F I G U R E 1 . 2 TSP and SO2 Concentrations in China, 2002 Source: Abstracted from www.sepa.gov.cn/ in northern cities. Since 2003, however, NOx and urban population--reported annual average particularly SO2 concentrations have increased. PM10 levels in excess of 100 µg/m3, which is twice When measured in terms of the number of the U.S. annual average standard. Twenty-one cities violating Chinese air quality standards, air percent of cities reported annual average levels in quality has shown some improvement since excess of 150 µg/m3. Only 1 percent of the coun- 1999. Table 1.2 shows the number of cities vio- try's urban population lives in cities with annual lating at least one air quality standard (cities clas- average PM10 levels below 40 µg/m3. sified as Grade III or worse than Grade III) since Sulfur dioxide levels in cities measure up bet- 1999. The number of cities worse than Grade III ter in terms of international standards. In 2003, has declined steadily since 1999. Nevertheless, almost three-quarters of cities had sulfur dioxide in 2005 about 50 percent of China's cities still levels below the U.S. annual average standard did not meet air quality standards. (60 µg/m3), suggesting that particulate air pollu- Table 1.3 presents the distribution of moni- tion is likely to be a more important health con- tored cities by PM10 and SO2 levels in 2003 and cern in the future. 2004. In 2003, 53 percent of the 341 monitored A direct consequence of air pollution from SO2 cities--accounting for 58 percent of the country's and NOX is acid rain, which remains a serious T A B L E 1 . 2 Trends in Air Quality in China's Cities (%) Air Quality Standards 1999 2000 2001 2002 2003 2004 2005 Grade II (Up to the standard) 33 37 34 36 42 39 52 Grade III 26 30 33 34 31 41 38 Worse than grade III 41 33 33 28 27 20 10 Source: Status of China Environment reports 1999­2005 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 3 O V E R V I E W figure 3). However, recent data (see table 1.1) sug- T A B L E 1 . 3 Distribution of PM10 and SO2 Levels in 341 Cities, gest that sulfur dioxide emissions are increasing 2003 and 2004 due to the high demand for coal in a rapidly grow- ing economy. Emissions in 2005 were over 25 mil- % of Cities lion tons, 28 percent higher than in 2000, and 42 percent higher than the 2005 target. Distribution of PM10 Levels 2003 2004 Despite increased SO2 emissions over the last PM10 100 µg/m3 three years (up 32 percent from 2001 to 2005), 46 47 100 < PM10 150 µg/m3 33 39 it should be noted that the number of cities PM10 > 150 µg/m3 21 14 reaching acceptable SO2 concentration standards (i.e. reaching class II) has in fact increased in the Distribution of SO2 Levels SO2 control zone and remained about the same in the acid rain control zone (see table 1.4). This may SO2 60 µg/m3 74 74 60 < SO2 100 µg/m3 14 17 indicate that SO2 emission from high point SO2 > 100 µg/m3 12 9 sources have increased, while emissions from low point sources and area sources have decreased. Source: China Environmental Yearbooks 2004 and 2005. Water Pollution Trends and Quality problem in China. Figure 1.3 shows the distribu- Surface water quality in China is poor in the most tion of rainfall by pH level in China in 2001, densely populated parts of the country, in spite of 2003, and 2005. The problem remains serious in increases in urban wastewater treatment capacity. the south and southeastern portions of the coun- Water quality is monitored by the State Environ- try. As illustrated below, there are some indications mental Protection Administration (SEPA) in that the main areas affected are gradually moving about 500 river sections and by the Ministry of from southwest to southeast. Over half of China's Water Resources in more than 2,000 sections sulfur dioxide emissions come from electric utili- across the main rivers. It is classified into one of ties (Sinton, 2004). Total sulfur dioxide emissions five categories based on concentrations of the 30 declined in the late 1990s, largely due to stricter substances listed in Annex 2. Recent trends sug- standards on emissions of SO2 by coal-fired power gest that quality is worsening in the main river sys- plants and to the "Two Zones" control program tems in the North, while improving in the South designed to reduce acid rain by controlling SO2 (see figure 1.4). For all the five main river systems emissions in cities with high ambient SO2 levels in the North (Songhua, Liao, Hai, Huai, and (see the second map in figure 1.2 and the maps in Huang rivers), sections with class IV to VI ranked F I G U R E 1 . 3 Distribution of Acid Rain in China, 2001, 2003, and 2005 4 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W T A B L E 1 . 4 Distribution of SO2 Levels Among Cities in the Two Air Pollution Control Zones, 1998­2005 (in %) SO2 Concentrations 1998 2000 2002 2003 2004 2005 In the SO2 control zone: Reaching Class II standards: 33 48 41 39 41 45 (SO2 0.6 mg/m3) Reaching Class III standards: 30 25 31 25 30 34 (0.06 mg/m3 < SO2 0.10 mg/m3) Below Class III standards: 37 27 28 36 29 21 (SO2 > 0.10 mg/m3) In the acid rain control zone: Reaching Class II standards: 70 81 79 75 73 74 (SO2 0.6 mg/m3) Reaching Class III standards: 14 6 14 15 20 22 (0.06 mg/m3 < SO2 0.10 mg/m3) Below Class III standards: 16 13 7 10 7 4 (SO2 > 0.10 mg/m3) Source: Status of China Environment reports 2000­05 F I G U R E 1 . 4 Surface Water Quality, 2000 and 2004 songhuajiang liaoriver hairiver northwest huangriver huairiver southwest 2000 i-iii yangziriver southeast 2000 > iii 2004 i-iii hujiang 2004 > iii Source: China--Water Quality Management--Policy and Institutional Considerations (World Bank, 2006) C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 5 O V E R V I E W water--i.e., non-potable water sources, but that Pollution of sea water and lakes is also serious. may be used by industry (class IV) and agriculture Thirty percent of sites at which sea water quality (class V)--increased, while the better class I­III is monitored have quality poorer than Grade III. ranked water--i.e. suitable for drinking water, Seventy-five percent of the lakes in China swimming and household use, and which also can exhibit some degree of eutrophication. Among support aquatic life--increased in the South. the 27 major lakes and reservoirs monitored in The overall trend for the period 1990 to 2005 2004, none met the Grade I water quality stan- indicates that water quality has become substan- dard, only two (7.5 percent) met the Grade II tially better in the water-rich south, but has not water quality standard, and five (18.5 percent) improved and may even have worsened in the met the Grade III quality standard. Most sites water-scarce north (see figure 1.5). have lower quality levels: four (14.8 percent) are In 2004, about 25,000 km of Chinese rivers Grade IV quality, six (22.2 percent) are Grade failed to meet the water quality standards for V, and ten (37.0 percent) failed to meet the aquatic life and about 90 percent of the sections Grade V quality standard. The "Three Lakes" of rivers around urban areas were seriously pol- (Taihu, Chaohu, and Dianchi) were among the luted (MWR 2005). Many of the most pol- lakes failing to meet the Grade V water quality luted rivers have been void of fish for many standard; total nitrogen and phosphorus were years. Among the 412 sections of the seven the main pollution indicators contributing to major rivers monitored in 2004, 42 percent poor water quality (SEPA 2004). met the Grade I­III surface water quality stan- From a health perspective, it is drinking water dard (that is, water that is safe for human con- quality that matters more than surface water qual- sumption), 30 percent met Grade IV­V ity. Although the last major, nationwide survey of standards, and 28 percent failed to meet Grade drinking water quality in China occurred in the V. Figure 3.2 (chapter 3) shows for 2004 the 1980s, monitoring of drinking water and the location of monitoring stations that failed to sources of drinking water in 300 rural counties, meet Class I to III standards. The bulk of the together with data on disease incidence, suggest violations occurred in the north in areas of high that polluted drinking water continues to be a population density. problem in rural areas. Due to inadequate treat- F I G U R E 1 . 5 Average Water Quality in Southern and Northern Rivers, 1991­2005 100 100 90 South China 90 V ­ V* 80 80 North China 70 70 V ­ V* 60 South China 60 III ­ IV 50 50 40 40 North China 30 30 III ­ IV 20 20 South China 10 I ­ II 10 North China I ­ II 0 0 3 1 3 0 4 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 200 2004 2005 199 1992 199 1994 1995 1996 1997 1998 1999 200 2001 2002 2003 200 2005 South China I ­ II South China III ­ IV South China V ­ V* North China I ­ II North China III ­ IV North China V ­ V* Source: China Water Quality Management--Policy and Institutional Considerations (World Bank 2006). 6 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W ment, drinking water standards are often violated pollution may still be a serious problem. Figure 1.6 even in piped water in townships and villages contrasts mortality rates from esophageal, stom- across China. Concerning non-piped water, mon- ach liver, and bladder cancers in different parts of itoring data from rural areas show extremely large China with world averages. Death rates due to violations of guidelines. The main problem is stomach, liver, and bladder cancers in rural China land-based contamination. Approximately two- are considerably higher than world averages and fifths of the rural population does not have piped also much higher than in large cities in China. drinking water, according to the 2005 China Health Yearbook. Analyses presented in Chapter Energy use, industrialization, and 3 of this report suggest a correlation between urbanization affect environmental levels of bacteria and total coliform in drinking performance water and absence of piped water, as well as a Trends in energy use offer a possible explanation clear relationship between lack of access to piped for the recent increase in SO2 emissions described water and prevalence of diarrhea in children. above. Following the economic slowdown in the When it comes to infectious diseases associated late 1990s, the economy grew by about 9 percent with drinking water pollution, however, the an- each year. Total energy consumption in China nual incidence rates have shown a marked down- increased by 70 percent between 2000 and 2005 ward trend in the last 20 years. (see figure 1.7). Coal consumption accounted for Although information is not readily available 75 percent of this increase, while the fraction on the percent of the population exposed to vari- of energy consumption met by hydropower ous levels of chemical and inorganic pollutants, decreased during the 2001­05 period. Moreover, mortality rates associated with cancers of the diges- following a marked decrease in the energy inten- tive system (stomach, liver, and bladder cancers) in sity of GDP between 1978 and 2001--measured rural areas in China suggest that drinking water in standard coal equivalents (SCE) used to F I G U R E 1 . 6 Mortality Rates for Diseases Associated with Water Pollution (per 100,000) in China in 2003 and World Averages in 2000 35 30 Major cities Medium/small cities 25 Rural World average 20 15 10 5 0 Oesophagus cancer Stomach cancer Liver cancer Bladder cancer Source: MoH 2004 and WHO 2006. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 7 O V E R V I E W F I G U R E 1 . 7 Total Energy Consumption in China, 1978­2005 of 2.5 tons 2 (10,000 1.5 SCE) 1 Production 0.5 0 Energy 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Year Source: Calculations based upon China Statistical Yearbooks, Various Years. produce 10,000 Yuan GDP--energy intensity tributed to increases in urban COD and ammo- increased in the 2002­05 period (see figure 1.8). nia nitrogen loads. Although the rate of urban Production of 10,000 Yuan GDP in 1978 water treatment is increasing (up to 45 percent in required energy equal to 8.43 tons SCE. This 2005), the absolute number of urban residents was reduced to 2.58 tons in 2001--a 3.2-fold not linked to water treatment systems has also reduction. However, energy intensity increased increased. Moreover, the share of the industries to 2.76 tons in 2005. that contribute most to water pollution loads-- China has also experienced an unprecedented pulp and paper, food production & processing, increase in the rate of urbanization. From 2000 textiles, and mining and tanning--have all to 2005, China's urban population increased by retained their respective Gross Industrial Output 103 million (see table 1.5). This has likely con- Value (GIOV) in the industrial process. This F I G U R E 1 . 8 Energy Use (SCE) to Produce 10,000 Yuan of GDP Energy Use (SCE) in China per 10,000 Yuan of GDP 9.000 of 8.000 Yuan 7.000 6.000 10,000 5.000 GDP 4.000 per 3.000 use 2.000 1.000 Energy 0.000 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 Year Source: Calculations based upon China Statistical Yearbooks, Various Years. 8 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W T A B L E 1 . 5 China's Urbanization and Industrialization Urban Population % Urban GIOV Values (Bio RMB in GIOV Values Year Total Population (million) Population constant 1990 prices) (indexed) 1978 963 172 18 255 100 1985 1,059 251 24 502 197 1990 1,143 302 26 686 269 1995 1,211 352 29 1723 675 2000 1,267 459 36 2753 1071 2004 1,300 543 42 4083 1600 2005 1,308 562 43 4594 1800 Source: Calculations based upon China Statistical Yearbook various years. implies that China has yet to realize a substantial Water depletion and consumption of unsafe reduction in industry-based water pollution due water are linked responses to water scarcity. In to changes in industrial structure favoring cleaner some areas of China, authorities do not supply downstream production. unsafe water, with the implication that ground- water depletion increases. For example, this happens in the lower reaches of the Yangtze. It WATER SCARCITY AND THE USE OF is estimated that 25 billion cubic meters of non- POLLUTED WATER FOR IRRIGATION rechargeable deep-aquifer groundwater were Generally speaking, China's water resources are mined in China in 2000, 90 per cent of which most abundant in the southern and western re- was used for agricultural purposes. gions of the country and scarce in the north. The In other areas, polluted water is used to the northeast plain areas account for one-third of maximum extent and water depletion is less than GDP, but only 7.7 percent of national water it would have been otherwise. Wastewater irri- resources, while the southwestern areas account gation zones are spreading in China and now for 21.3 percent of national water resources, but account for about 4 million hectares of agricul- only 8.7 percent of GDP. tural land. The produce is likely to contain heavy To cope with water scarcity, China has metals such as mercury, cadmium, lead, copper, developed strategies that have to some degree chromium, and arsenic. put pressure on the environment. There are three ways that water scarcity harms the envi- The Chinese Environmental ronment. First, water scarcity may lead to deple- Pollution Impact Model tion of groundwater. In some areas of China, the groundwater table has fallen 50 meters since This report represents the culmination of a 1960, and it continues to fall 3 to 5 meters joint effort between the Chinese government annually. Second, water scarcity may lead to and a team of Chinese and international experts excessive consumption of unsafe, polluted water. to assess the costs of environmental degradation Consumption of unsafe water in China runs to in China. The team (see figure 1.9) consisted of billions of cubic meters every year. As a third staff members from China's State Environ- consequence, water scarcity may lead to indus- mental Protection Administration (SEPA) and try, agriculture, and households being periodi- affiliates--the Chinese Academy for Environ- cally rationed. mental Planning, the Policy Research Center of C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 9 O V E R V I E W F I G U R E 1 . 9 Institutions Involved in the Project ECON / ECM VEHR RFF CICERO Beijing Chongqing Shanghai SEPA + Affiliates CAES MoH CDC Fudan University CDC MWR CAEP BoH Peking Univ. Sch. PH Environment and Economy, and the China impacts; (b) contribute to the development of National Environment Monitoring Center--as a National Environmental Accounting System; well as other government agencies such as the and (c) contribute to provincial comparisons of Ministry of Water Resources (MWR), Ministry environmental performance. of Health (MoH), and the Center for Disease To accomplish these aims, the project was Control and Prevention (CDC). The team also designed to fulfill a set of technical objectives: included staff from the World Bank, Resources for the Future (USA), CICERO (Norway), and 1. To formulate, based on Chinese as well as ECON (Norway). It was formed with the international studies, a Chinese Environmen- intention of both assessing current environ- tal Cost Model (CECM) that would calculate mental damages from air and water pollution the damages associated with air and water pol- and developing the tools that would enable lution, by pollutant, sector, and province. these damages to be calculated on a continuing 2. To undertake pilot studies on the valuation basis at both the national and provincial levels. of health risk (VEHR) that would estimate The project, supported by the World Bank, willingness to pay (WTP) for reductions in adopted a multi-sectoral approach to assessing premature mortality for use in the CECM. the magnitude of air and water pollution in 3. As an integrated part of the CECM, to develop China, with critical data and inputs from SEPA a software tool that would standardize and (and its affiliates) and affiliates under the MWR make operational the calculation of environ- and MoH including CDC). mental costs. As part of the multiyear effort to refine 4. To build capacity for environmental cost methodologies and estimate the costs of pollu- calculation in China through collaboration tion, an environmental cost model was devel- between China's national expert team and an oped to (a) help monitor annual environmental international expert team. 10 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W F I G U R E 1 . 1 0 Main Government Partners in the Project Local Environment Environmental Protection Bureau Monitoring Environmental Cost From Pollution Project MOH HQ CDC (MOH) 5. To identify gaps in knowledge--both gaps in of an ECM for China has been aided by three research and in the collection of environmen- factors: tal data--that must be filled if the ECM is to form a basis for decision making in China. · The advancement of methods for assessing environmental costs over the past 20 years. It should be emphasized that the outputs of the Methods to calculate the burden of disease project can be used for three purposes: (1) to cal- attributable to air and water pollution have culate the total damages associated with air and advanced significantly, as have methods of water pollution; (2) as an input to China's Green estimating the economic costs of environ- National Accounts; and (3) to calculate the ben- mental degradation. efits of programs to reduce air and water pollu- · The expansion of studies of pollution tion. Box 1.1 summarizes how similar analyses damages--for example, of the health effects have been used in other countries. of air pollution--by Chinese researchers. This report summarizes the results of the Previous studies of environmental damage in environmental cost model (ECM) and valuation China (World Bank 1997; Cohen et al. of environmental health risks (VEHR) studies 2004) have relied largely on transferring and also describes the methods, data, and litera- dose-response functions from the interna- ture that have been used to calculate environ- tional literature to China. A hallmark of the mental costs in this project. The development current project is its reliance on studies con- C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 11 O V E R V I E W B O X 1 . 1 Environmental Cost Models: International Experience The goal of this project--to quantify environmental degradation using a damage function approach--parallels efforts undertaken by international agencies and governments throughout the world. This box summarizes these efforts. Global burden of disease due to environmental factors. The World Health Organization (WHO) has calculated (by region) mortality and morbidity associated with both indoor and out- door air pollution using the same methods as this study. In the case of outdoor air pollution, WHO has estimated annual average PM10 concentrations for over 3,000 cities around the world and has used concentration-response functions from Pope et al. (2002) to translate these into premature deaths associated with air pollution. These are calculated by comparing current annual average PM10 levels in each city with a reference level of 15 µg/m3, the same reference level used in the CECM. To calculate the burden of disease associated with indoor air pollution (which is the focus of a separate study), odds ratios from the international literature were applied to the relevant populations exposed to biomass fuels. WHO converts cases of illness and premature mortality into disability-adjusted life-years-saved (DALYs) rather than monetizing cases of illness and premature death. Benefit-cost analyses of environmental regulations. The United States, United Kingdom, and other members of the European Union regularly conduct benefit-cost analyses of environmental regulations. The techniques used in this report to calculate the health impacts of reducing pollu- tion from current levels to background concentrations--the approach used in calculating the global burden of disease--can also be used to calculate the benefits of smaller reductions in air pollution that are likely to be delivered by various pollution control programs. In the United States (and the EU), the methods described in Chapter 5 of this report are used to monetize health benefits and compare them to costs. In the United States, benefit-cost analyses must be conducted for all "economically significant" regulations (those costing more than $100 million per year), and are routinely conducted for air quality regulations, following the same protocols used in Chapters 2 and 4 of this report. Benefit- cost analysis is typically used to judge the acceptability of a regulation (do benefits exceed costs?) and sometimes to rank regulatory options--for example, different maximum contaminant levels for arsenic in drinking water (USEPA 2000). ducted in China, studies that are more appro- (including buildings and monuments). Air pol- priate to the Chinese context. lution or pollution of rivers and lakes may also · The improvement in monitoring and environ- detract from recreation and aesthetic experiences. mental data collection in China. Improvements The CECM focuses on air and water pollution-- in monitoring of air and water pollution have both surface and drinking water pollution--but made it possible to quantify exposures to envi- does not include solid waste pollution or radiation ronmental pollution and estimate associated at this time. The main sectors for which damages damages. are estimated are health, agriculture, forests, fish- eries, materials, and water resources. Project Components In the case of air pollution, the model focuses on particulate matter (TSP or PM10), sulfur diox- Pollution costs are typically classified by pollu- ide (SO2), and acid rain. China is the world's tion medium and by the sector affected. Pollu- largest producer and consumer of coal, much of tion media include air, surface water, drinking which has high sulfur content. PM10 and SO2 water, land-based pollution (solid waste), as well from coal burning, with attendant acid rain, as noise and heat. Pollution damages are usually have caused severe pollution problems in China classified according to their effects--on human for decades. Particulate matter is the key air pol- health, agriculture, forests, fisheries, or materials lutant that has been studied in relation to human 12 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W health. Associations have been documented hepatitis A and dysentery. Another goal is to link between PM and premature mortality; incidence surface water pollution to impacts on fish popu- of chronic bronchitis, heart attack, and stroke; lations and to agriculture. The use of polluted respiratory and cardiovascular hospital admis- surface water for irrigation reduces both the sions; and restricted activity days. Acid rain, quantity of agricultural output that is suitable caused by SO2 reacting in the atmosphere with for human consumption and the quality of out- water, oxygen, and other substances, can reduce put. Pollution of surface water may also increase crop and timber yields and forest canopy and pressure on groundwater resources, contributing damage buildings and monuments, as can SO2 to the problem of water scarcity. in gaseous form. The goal of the CECM is to quantify and, In the case of water pollution, a variety of pol- where possible, to monetize the effects of air lutants are monitored in China, both in surface and water pollution listed in Table 1.6. using a and drinking water. These include biological damage function approach. This entails five pollutants such as coliform bacteria, which are steps: (1) identifying the nature of the pollution associated with fecal contamination, and chem- problem--for example, high annual average PM10 ical pollutants, including naturally occurring ele- concentrations in the ambient air or concentra- ments such as arsenic and fluoride, heavy metals tion of arsenic in drinking water; (2) identifying (such as mercury), ammonia, nitrates, and toxic the specific endpoints affected (cardiovascular petroleum compounds. From a health perspec- mortality in the case of PM10, or liver cancer in tive, it is drinking water quality that matters most. the case of arsenic) and estimating an exposure- Epidemiological studies have linked virtually all response function that links exposure to each of the drinking water pollutants in Appendix 2 to endpoint; (3) estimating population exposures either chronic or acute health effects. Eventually, (numbers of persons exposed to various PM10 the goal of the CECM is to link specific drink- concentrations or concentrations of arsenic in ing water pollutants to health endpoints such as drinking water); (4) calculating the physical cancers of the liver and digestive system; to other effects of exposure (deaths due to PM10 exposure chronic diseases, such as diabetes and cardio- or cases of liver cancer attributed to arsenic expo- vascular disease, which have been associated with sure); and (5) assigning a monetary value to the arsenic; as well as to acute illnesses, such as physical effects. T A B L E 1 . 6 Sectors and Pollutants Included in the CECM Environmental Sectors Health Agriculture Materials Forestry Water Resources Fishery Pollutants Air pollutants TSP (PM10) SO2 Acid rain Water pollutants In drinking water In surface water Source: the project team. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 13 O V E R V I E W F I G U R E 1 . 1 1 Flow Chart for Estimating the Economic Cost of Pollution Dose-response Polluted relationship Pollution condition Physical Monetary (concentration) impact impact area Exposed population and activity Source: the project team. Step 1: Identify the pollution factors, polluted areas than to estimate the effects of chronic expo- area, and related conditions. sure to arsenic in drinking water. Step 2: Determine affected endpoints and estab- In China, PM10/TSP and SO2 are regularly lish dose-response relationships for pol- monitored in 341 cities, some of which also lution damage. monitor nitrogen oxides (NOx). Dose-response Step 3: Estimate population (or other) expo- functions linking these pollutants to a variety of sures in polluted areas. health outcomes have been estimated by Chi- Step 4: Estimate physical impacts from pollu- nese and international researchers. As a result, tion using information from steps 2 estimating the health impacts of air pollution in and 3. urban areas is relatively straightforward, at least Step 5: Convert pollution impacts in physical for acute health effects. In the case of arsenic or terms to pollution costs in monetary other pollutants in drinking water, monitoring terms. data are more difficult to obtain, and the defini- tion of an exposure metric is more complicated The measurement of physical effects attributable than for air pollution. to pollution depends crucially on the existence Drinking water is monitored in a sample of of dose-response functions linking pollution counties by the Chinese Center for Disease Con- exposure to physical effects, and also on the abil- trol and Prevention in Beijing, but the samples ity to characterize exposures. This has been done are not sufficient to obtain an accurate estimate of more successfully in the case of human health the fraction of the population exposed to differ- and air pollution and material damage and air ent concentrations of pollutants in their drinking pollution than in other areas. For material dam- water throughout the country. Thus, although age, exposure-response functions are available for there are epidemiologic studies linking arsenic to most building materials. However, a compre- liver cancer, it is difficult to apply them, as indi- hensive exposure assessment is more difficult due cated in Figure 1.11, for lack of exposure data. to lack of data on the amount and surface area of The absence of dose-response functions materials in use. Concerning human health, the becomes more of a problem when examining the availability of dose-response functions and data effects of pollution in non-human populations. on exposure differ greatly among pollutants and For example, the literature linking fish popula- health endpoints. For example, it is much easier tions to surface water pollution (either to acid to estimate the health effects of PM10 in urban rain, or to eutrophication of lakes due to nitrogen 14 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W loadings) is sparse. So is the literature linking acid dren under 5 living in rural areas of China rain or SO2 to timber yields and to tree cover. with availability of piped water. This makes it difficult--in China, but also in Chapter 4. Valuing Environmental Health Effects. Western countries--to quantify the effects of air An important goal of the CECM/VEHR and water pollution on forests and fisheries. For project is to contribute to the literature on these reasons, it has not been possible to quantify health valuation in China. This chapter sum- all of the effects checked in Table 1.6. marizes the results of original studies conducted The remainder of this report summarizes the in Shanghai and Chongqing to estimate peo- current state of analysis of the effects of air and ple's willingness to pay to reduce risk of pre- water pollution in the CECM. It is divided into mature death. The chapter also discusses the 6 chapters, organized as follows: Adjusted Human Capital (AHC) approach-- the official approach used to value health Chapter 2. The Health Impacts of Ambient Air costs in China, and uses both approaches to Pollution. The CECM quantifies cases of value premature mortality associated with chronic bronchitis, premature mortality, and air pollution. Estimates of the value of air- respiratory and cardiovascular hospital admis- pollution-related morbidity are also pre- sions associated with PM10 in urban areas in sented, as well as the health impacts of water China. This is a bottom-up analysis, con- pollution. ducted at the city level, and aggregated to the Chapter 5. The Non-Health Impacts of Water provincial and national levels. A distinguish- Pollution. This chapter concentrates on the ing feature of the CECM is its use of Chinese impacts from water pollution, where pollu- concentration-response functions rather than tion of surface water bodies can reduce agri- relying solely on dose-response transfer from cultural yields and harvests of fish. It estimates the international literature. the damages associated with acute pollution Chapter 3. The Health Impacts of Water Pollu- incidents affecting fisheries and the damages tion. As noted above, it is not possible to mea- associated with the use of sewage-contaminated sure population exposures to the pollutants water for irrigation of crops. It also deals with listed in Table 1.6 from available data. This the related issue of water scarcity caused by chapter presents an overview of surface water pollution. pollution in China, as well as information on Chapter 6. The Non-Health Impacts of Air Pollu- the source of drinking water and the nature tion. This chapter focuses on the non-health of drinking water treatment. Information on impacts from air pollution, including SO2 and the levels of specific pollutants in drinking acid rain damage to buildings and other ma- water is presented for a sample of rural terials and their impacts on crop and timber counties, as well as for selected districts in yields. It values damages to buildings in South Chongqing. Information on the incidence of China and crop losses due to acid rain and SO2 diseases that have been associated with vari- pollution throughout the country using Chi- ous drinking water pollutants is presented, nese dose-response information. Effects on together with a disease matrix summarizing forests are not quantified due to lack of data on associations found in the Chinese and inter- the area planted, by species, and lack of appro- national literature. An attempt is made to priate dose-response functions. compute the impact of polluted drinking water on cancer incidence in rural areas. The Table 1.7 below highlights some important types chapter concludes with original research link- of environmental damages that were not quanti- ing incidence of diarrheal disease among chil- fied due to lack of sufficient data. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 15 O V E R V I E W T A B L E 1 . 7 Environmental Damages in the CECM Quantified Damages Non-quantified Damages Why Not Quantified Health effects of ambient PM10 Health effects of ambient ozone 1 Diarrheal disease associated with no Health effects associated with 1 piped water connection; cancers chemical and inorganic water associated with water pollution pollutants SO2 and acid rain damage to crops SO2 and acid rain damage to forests 1,2 SO2 and acid rain damage to buildings SO2 and acid rain damage to other 1 types of construction Acute effects of water pollution on fish Chronic effects of water pollution 1,2 on fish Agricultural damages from wastewater irrigation 1 = Effect not quantified due to insufficient information about exposure 2 = Effect not quantified due to insufficient information about dose-response Source: the project team. A N N E X 1 . Concentration Values of Pollutants in Ambient Air Concentration Values Name of Pollutant Time Class I Class 2 Class 3 Concentration Level Unit SO2 Yearly average 0.02 0.06 0.10 Daily average 0.05 0.15 0.25 Hourly average 0.15 0.50 0.70 TSP Yearly average 0.08 0.20 0.30 Mg/m3 Daily average 0.12 0.30 0.50 PM10 Yearly average 0.04 0.10 0.15 Daily average 0.05 0.15 0.25 NOx Yearly average 0.05 0.05 0.10 Daily average 0.10 0.10 0.15 Hourly average 0.15 0.15 0.30 NO2 Yearly average 0.04 0.04 0.08 Daily average 0.08 0.08 0.12 Hourly average 0.12 0.12 0.24 CO Daily average 4.00 4.00 6.00 Hourly average 10.00 10.00 20.00 O3 Hourly average 0.12 0.15 0.20 Pb Seasonal average 1.50 Yearly average 1.00 µg/m3 B(a)P Daily average 0.01 Daily average 7a F Hourly average 20a Monthly average 1.8b 3.0c µg/dm2 . d? 1.2b 2.0c a. Urban area b. Pasturing area, or Part Farming--Part Pasturing, or Silkworm-mulberry producing area c. Farming and Forestry Area 16 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N O V E R V I E W A N N E X 2 . List of Pollutants Monitored in Surface Water and Their Standards (mg/L) Categories No. Parameters I II III IV V Basic requirements All water bodies should not contain substances from non-natural causes as listed below: a. Any substance that can subside and form offensive sediments b. Floating matter, such as fragments, floating scum, oils, or any other materials that can offend sense organs c. Any substance that produces offensive color, odor, taste, or turbidity d. Any substance that can harm human beings, animals, and plants, or cause toxic or adverse physiological reactions e. Any substance that can easily cause the breeding of offensive aquatic organisms 1 Water temperature (C°) Temperature changes in the water environment induced by human activities should be within: Summer weekly average maximum temperature rise 1 Winter weekly average maximum temperature down 2 2 pH 6.5 8.5 (mg/L) 6 9 (mg/L) 3 Sulfate (as SO4 )* 2- below 250 250 250 250 250 4 Chloride (as CI- )* below 250 250 250 250 250 5 Soluble iron* below 0.3 0.3 0.5 0.5 1.0 6 Total manganese* below 0.1 0.1 0.1 0.5 1.0 7 Total copper* Below 0.01 1.0 (0.01 for 1.0 (0.01 for 1.0 1.0 fishery) fishery) 8 Total zinc* 0.05 1.0 (0.1 for 1.0 (0.1 for 2.0 2.0 fishery) fishery) 9 Nitrate (as N) Below 10 10 20 20 25 10 Nitrate (as N) 0.06 0.1 0.15 1.0 1.0 11 Non-ionic ammonia 0.02 0.02 0.02 0.02 0.02 12 Kjeldahl nitrogen 0.5 0.5 1.0 2.0 2.0 13 Total phosphorus (as P) 0.02 0.1 (0.025 for 0.1 (0.05 for 0.2 0.2 reservoirs reservoirs and lakes) and lakes) 14 Permanganate value 2.0 4.0 6.0 8.0 10.0 15 Dissolved oxygen 90% of 6.0 5.0 3.0 2.0 saturation value 16 Chemical oxygen Demand (CODCr) Below 15 Below 15 15 20 25 17 Biological oxygen Demand (BOD5) Below 3.0 3.0 4.0 6.0 10 18 Fluoride (as F-) Below 1.0 1.0 1.0 1.5 1.5 19 Selenium (IV) Below 0.01 0.01 0.01 0.02 0.02 20 Total arsenic 0.05 0.05 0.05 0.1 0.1 21 Total mercury** 0.00005 0.00005 0.00001 0.001 0.001 22 Total cadmium*** 0.001 0.005 0.005 0.005 0.01 23 Chromium (VI I) 0.01 0.05 0.05 0.05 0.1 24 Total lead** 0.01 0.05 0.05 0.05 0.1 25 Total cyanide 0.0005 0.05 (0.005 0.02 (0.005 0.2 0.2 for fishery) for fishery) 26 Volatile phenol** 0.002 0.002 0.005 0.01 0.1 27 Oils** (Petroleum ether extraction) 0.05 0.05 0.05 0.5 1.0 28 Anionic surfactant Below 0.2 0.2 0.2 0.3 0.3 29 Coli-index*** (Individuals/L) 10000 30 Benzo (a) pyrene (pg/L) 0.0025 0.0025 0.0025 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 17 O V E R V I E W A N N E X 3 . Pollutants Monitored in Drinking Water in China and Drinking Water Standards Drinking Water Pollutants Class I Class II Class III Chrome (degree) 15.0 20.0 30.0 Turbidity (degree) 3.0 10.0 20.0 Total dissolved solids (mg/L, CaCO3) 450.0 550.0 700.0 Iron (mg/L) 0.3 0.5 1.0 Manganese (mg/L) 0.1 0.3 0.5 COD (mg/L) 3.0 6.0 6.0 Chlorate (mg/L) 250.0 300.0 450.0 Sulfate (mg/L) 250.0 300.0 400.0 Fluoride (mg/L) 1.0 1.2 1.5 Arsenic (mg/L) 0.1 0.1 0.1 Nitrate (mg/L) 20.0 20.0 20.0 Total bacteria (/mL) 100.0 200.0 500.0 Total coliform (/L) 3.0 11.0 27.0 18 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 1 2 Health Impacts of Ambient Air Pollution This chapter reviews the health effects Energy consumption, especially coal consumption, is the main source of air associated with particulate matter, pollutants such as particles, SO2, NOx, and CO in most cities of China. As summarizes population exposure to the primary energy source, coal has accounted for about 65 to 70 percent PM10 in China and describes the (China Statistical Yearbook 2004) of total energy consumption in recent techniques used to estimate the years, which has caused many environmental and human health problems. health damages associated with PM10 Crude oil consumption has been increasing because of the rapid expansion exposure in 2003. Specifically, the of the motor vehicle fleet in many cities. In recent years, epidemiological CECM quantifies cases of chronic bronchitis, premature mortality, and studies conducted around the world have demonstrated that there are close respiratory and cardiovascular hospital associations between air pollution and health outcomes. PM10 and SO2 are admissions associated with PM10 in chosen in many studies as the indicative pollutants for evaluating the health urban areas in China. effects of ambient air pollution. Although the mechanisms are not fully understood, epidemiological evidence suggests that outdoor air pollution is This is a bottom-up analysis, a contributing cause of morbidity and mortality. Epidemiological studies conducted at the city level, and have found consistent and coherent associations between air pollution and aggregated to the provincial and various outcomes, including respiratory symptoms, reduced lung function, national levels. A distinguishing chronic bronchitis, and mortality. feature of the CECM is its use of In China, epidemiological studies have been conducted beginning in the Chinese concentration-response 1980s and 1990s in Beijing, Shenyang, Shanghai, and other cities. These functions rather than relying solely include two time-series analyses of the relationship between daily air pol- on dose-response transfer from the international literature. The premature lution and hospital outpatient visits/emergency room visits and daily cause- deaths and cases of illness quantified specific population mortality in urban areas of Beijing (Chang et al. 2003; using the techniques described in this Chang, Wang, and Pan 2003), a meta analysis of exposure-response func- chapter are valued in Chapter 4. tions between air pollutants and cause-specific mortality derived from Chi- nese studies, and a regression analysis of environmental monitoring data and population mortality data for over 30 cities of China. (See CD-ROM A.1). These study results suggest that urban air pollution in China causes significant public health impacts and economic damage to the exposed pop- ulations. They provide a good foundation for further evaluation of the health impacts of air pollution in China. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 19 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N HEALTH OUTCOMES FROM selected all-cause mortality, hospital admissions AIR POLLUTION for respiratory and cardiovascular disease, and incidence of chronic bronchitis as endpoints Epidemiological research has found consistent because of the availability of exposure-response and coherent associations between air pollution functions. Health endpoints can be classified in and various health endpoints, or health effects. broad disease groups or specified in detail accord- These include reduced lung function, respiratory ing to ICD codes. Different studies on exposure- symptoms, chronic bronchitis, cardiovascular and response relationships may address more or less cerebrovascular diseases, hospitalization, outpa- specific health endpoints. Typically, studies report tient visits, work and school absenteeism, and pre- steeper exposure-response coefficients when cause- mature death. Although the mechanisms are still specific health endpoints are addressed, as opposed not fully understood, research during the past 10 to studies focusing on broader groups of end- to 20 years suggests that outdoor air pollution points. For these endpoints, we therefore have to contributes to morbidity and mortality linked apply a relatively crude classification, which in- with respiratory, cardiovascular, and cerebrovas- creases the uncertainty of the results. cular illness and diseases. Some effects may arise In health cost estimation, it is also important from short-term exposure, while others are asso- to make sure that the endpoints in the exposure- ciated with long-term exposure. When we select response functions are consistent with the end- health endpoints to be accounted for in the envi- points for which statistical data are available. At ronmental cost model, the basic principles are as present, the health data from regular surveillance follows: is often insufficient, and the system for reporting prevalence of morbidity is not complete, espe- · First priorities should be given to the health cially for some chronic diseases. This limits the endpoints that are registered in Chinese cities choice of health endpoints. Since there is no on a regular basis and classified by ICD-9 code requirement from the Ministry of Health for (or by ICD-10, the latest revision of the classi- cause-specific registration for emergency visits fication system). This will ensure data availabil- (EVs) and outpatient visits (OPVs), OPVs and ity and enable comparisons between regions. EVs for respiratory and cardiovascular diseases These data include population mortality, cannot enter endpoint lists despite documented hospital admissions, and hospital outpatient/ studies on their dose-response coefficients. emergency visits. In line with the above principles, the health · There are well-documented studies of expo- endpoints evaluated in this project are described sure-response functions between concentra- as follows: tions of air pollutants (exposure) and the given · Mortality. all-cause mortality health endpoints (response). · Morbidity. respiratory and cardiovascular hospi- · The methodologies applied in the epidemio- tal admissions; incidence of chronic bronchitis logical studies forming the basis for exposure- response functions should be as similar as Two endpoints related to hospitalization are possible to studies in other countries to facil- selected, covering the bulk of hospital admissions itate comparison. attributable to air pollution. The two endpoints are hospital admissions due to cardiovascular dis- As noted above, the selection of health endpoints eases and hospital admissions due to respiratory is restricted by the availability of exposure- diseases. A broad range of diagnoses, specified by response studies. In this assessment we have their ICD-9 code, were included in the studies 20 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N from which the exposure-response functions are violating the Class-II standard (100µg/m3). derived, including cerebrovascular diseases, pul- Annual average NO2 concentrations of all monary heart diseases, ischaemic heart disease, monitored cities met the Class-II standard COPD, and pneumonia. The prevalence of (50µg/m3). This suggests that particulate chronic respiratory symptoms and diseases in matter has become the air pollutant of pri- a population is related to long-term, integrated mary concern in China. exposure. Prevalence rates are often higher in 2) Different air pollutants may have a synergetic adults than in children. We selected the end- effect on human health. For instance, the point chronic bronchitis as an endpoint pre- combined effect of SO2 and PM10 may be sumably representing an important share of the higher (or lower) than the sum of the two economic impact and human suffering associated components when they occur in isolation. with air pollution. Chronic bronchitis typically Moreover, a part of PM10 may be sulfate, constitutes the largest share of cases of chronic which is converted SO2. In spite of a large obstructive pulmonary diseases (COPD) and body of studies, the contribution of each of covers a range of sub-diagnoses, which are all these pollutants to health damage is difficult likely to entail substantial reduced well-being to disentangle. In our view, adding the health and restricted activity. cost from, respectively, PM10 and SO2 may lead to double counting. 3) The trial calculation results showed that the CAUSAL AGENTS AND THRESHOLD VALUES health cost estimated for SO2 (based on the dose-response coefficients in the December 2002 Progress Report of Chinese Environmental Causal Agents in Air-Pollution- Cost Model) represented only about one-tenth Related Disease of the total health cost due to air pollution. Although adverse effects on human health from particulate matter, SO2, O3, NOx, and CO are Because particulate matter under 10 µm is an documented, most studies have focused on the important vector for several toxic and hazardous relationship between SO2, particulate matter, air pollutants and because of the close relationship and respiratory and cardiovascular diseases. After between PM10 and health effects found in many thorough consideration, we decided to choose epidemiological studies, we exclude SO2 from the PM10 as the single air pollutant index for the fol- final estimation to avoid double counting. lowing reasons: Air Pollutant Thresholds 1) Ambient SO2 concentrations in most Chinese cities have greatly decreased compared with a According to WHO (2000), there is no level few years ago, and are in many cities now below which particulate matter may not result in lower than the WHO Air Quality Guideline health effects in the susceptible population, but (2000) of 50µg/m3. The air quality monitor- there is a lower limit to the level at which results ing results from Chinese cities in 2003 showed have been reported in epidemiological studies. that, among the 341 monitored cities, the We define 15µg/m3 as the lower threshold value annual average ambient SO2 concentration for PM10 effects, given that the lowest PM10 con- exceeded the Class-II standard (60µg/m3) in centration observed in the ACS cohort study by 26 percent of the cities. Fifty-five percent of the Pope (1995) is 15µg/m3. This lower threshold is cities had annual average PM10 (TSP) levels also applied by WHO (Cohen et al. 2004). C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 21 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N POPULATION EXPOSURE uate the effects of outdoor air pollution on the human health of the population in rural areas. China is the largest developing country in the Chapter 3 of this report estimates the health world. Of the more than 1.3 billion inhabitants, impacts of indoor air pollution in rural areas about 40 percent live in urban areas (China Sta- in China. tistical Yearbook 2004). With the growing econ- omy, many cities in China have to face the challenge of air pollution. The health impacts of EXPOSURE-RESPONSE air pollution, especially in cities, are gradually RELATIONSHIPS being acknowledged by researchers, government, and the public. It is difficult to estimate the num- Review of Epidemiological Evidence ber of people exposed to high levels of air pollu- The effects of air pollution on human health tion at a national level, because there are large include the chronic effects of long-term exposure variations across different geographic and meteo- rological areas, as well as across socioeconomic and the acute effects of short-term exposure. In groups. Generally, ambient air pollution is closely the past two decades, a large number of studies-- associated with industrialization and urbaniza- especially short-term, time-series studies--have tion. Hence, the urban population is likely to be reported exposure-response relationships between the primary group exposed to high levels of ambi- air pollution exposure and human health. Long- ent air pollution. Moreover, the state-controlled term cohort studies provide the best method to air pollution monitoring sites are distributed in evaluate the chronic effects of air pollution on urban areas only, so the air quality data represent human health, whereas time-series studies are air pollution levels in cities. "Exposed popula- appropriate for revealing the acute effects of short- tion" in this health damage valuation refers to term fluctuations in pollution levels. Exposure- urban residents, defined as the population of response coefficients from cohort studies of urban districts as given in the China City Statis- premature mortality are typically several times tical Yearbook (2004). higher than coefficients reported in time-series Table 2.1 shows the percentage of the urban studies. We assumed that the short-term effects population exposed to different classes of PM10 found in time-series studies are embedded in the levels in the 31 provinces of mainland China. Fig- long-term effects on mortality rates derived from ure 2.1 maps the percentage of urban population cohort studies. exposed to Class III and > Class III PM10 levels. A large number of time-series studies of mor- Over half of the urban population in China is tality have been published in the past 20 years, exposed to annual average PM10 levels greater than but only a few cohort studies have appeared. In or equal to the Class III standard (100 µg/m3). China, there are some time-series studies and sev- Over 11 percent are exposed to PM10 levels in eral cross-sectional mortality studies, conducted in excess of 150 µg/m3, which is three times the U.S. cities such as Beijing (Chang et al. 2003; Chang, annual average standard. The provinces with the Wang, and Pan 2003; Dong et al. 1995; Dong largest percentage of people exposed to PM10 lev- et al. 1996; Gao et al. 1993; Xu et al. 1995; Xu els greater than or equal to the Class III standard et al. 1994), Shanghai (Kan and Chen 2003; are generally in the north, while eastern and south- Kan and Chen 2004), Shenyang (Wang, Lin, ern provinces with high population densities-- and Pan 2003; Xu et al. 1996a; Xu et al. 2000; Shandong, Guangdong, and Jiangsu--have the Xu et al. 1996b), and Chongqing (Venners highest numbers of people exposed. et al. 2003). There are few air pollution monitoring sta- To derive exposure-response functions for air tions in rural areas in China, so we cannot eval- pollution and mortality applicable to the entire 22 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N T A B L E 2 . 1 PM10 Pollution Exposure of the Urban Population (population in 10,000's) I Class II Class III Class >III Class PM10< PM10: PM10: PM10 Total Provinces Item 40µg/m3 40­100µg/m3 100­150µg/m3 >150µg/m3 Population/% Beijing Population 0 0 1,079 0 1,079 % 0.00 0.00 100.00 0.00 100 Tianjin Population 0 0 759 0 759 % 0.00 0.00 100.00 0.00 100 Hebei Population 0 384 1,650 496 2,529 % 0.00 15.16 65.23 19.60 100 Shanxi Population 0 148 322 796 1,267 % 0.00 11.68 25.45 62.87 100 Neimeng Population 0 144 311 240 694 % 0.00 20.67 44.74 34.59 100 Liaoning Population 0 1,615 1,265 78 2,958 % 0.00 54.61 42.75 2.64 100 Jilin Population 0 807 473 407 1,687 % 0.00 47.80 28.06 24.14 100 Heilong Jiang Population 0 627 841 179 1,647 % 0.00 38.08 51.04 10.88 100 Shanghai Population 0 1,278 0 0 1,278 % 0.00 100.00 0.00 0.00 100 Jiangsu Population 0 639 3516 458 4,613 % 0.00 13.84 76.22 9.94 100 Zhejiang Population 0 1,532 1,782 0 3,314 % 0.00 46.22 53.78 0.00 100 Anhui Population 0 927 1062 0 1,990 % 0.00 46.61 53.39 0.00 100 Fujian Population 0 1,243 385 79 1,707 % 0.00 72.81 22.57 4.62 100 Jiangxi Population 0 448 800 159 1,407 % 0.00 31.85 56.85 11.30 100 Shandong Population 0 3,610 1,546 190 5,345 % 0.00 67.53 28.92 3.55 100 Henan Population 0 792 1,706 738 3,236 % 0.00 24.47 52.72 22.81 100 Hubei Population 0 1,520 2,351 0 3,871 % 0.00 39.26 60.74 0.00 100 Hunan Population 0 317 1,594 358 2,269 % 0.00 13.97 70.23 15.80 100 Guang Dong Population 0 5,005 293 0 5,298 % 0.00 94.47 5.53 0.00 100 Guangxi Population 204 473 689 254 1,619 % 12.57 29.20 42.55 15.68 100 Hainan Population 354 113 0 0 467 % 75.86 24.14 0.00 0.00 100 Chong Qing Population 0 0 1,488 0 1,488 % 0.00 0.00 100.00 0.00 100 Sichuan Population 0 489 1,337 1,276 3,103 % 0.00 15.77 43.09 41.14 100 Guizhou Population 0 357 582 0 939 % 0.00 38.00 62.00 0.00 100 (continued) C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 23 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N T A B L E 2 . 1 PM10 Pollution Exposure of the Urban Population (population in 10,000's) (Continued) I Class II Class III Class >III Class PM10< PM10: PM10: PM10 Total Provinces Item 40µg/m3 40­100µg/m3 100­150µg/m3 >150µg/m3 Population/% Yunnan Population 64 789 76 0 929 % 6.91 84.95 8.14 0.00 100 Xizang Population 0 14 0 0 14 % 0.00 100.00 0.00 0.00 100 Shaanxi Population 0 147 721 340 1,207 % 0.00 12.19 59.69 28.13 100 Gansu Population 0 124 339 272 735 % 0.00 16.92 46.04 37.04 100 Qinghai Population 0 0 107 12 119 % 0.00 0.00 89.92 10.08 100 Ningxia Population 0 0 72 157 229 % 0.00 0.00 31.34 68.66 100 Xinjiang Population 0 181 278 226 684 % 0.00 26.37 40.66 32.96 100 Total Population 622 23,720 27,422 6,716 58,480 % 1.06 40.56 46.89 11.48 100 Source: authors calculations. Note: The PM10 pollution exposure is computed based on data from 660 cities. Because air pollution monitoring data in China are available for 341 cities, the air pollution levels of non-monitored county-level cities refer to the data of their upper-level prefecture cities. country, we undertook a systematic literature with an increase in total mortality, cardiopul- review and analyzed the available studies by monary mortality, and lung cancer mortality in means of meta-analysis and statistical trend analy- adults. These cohort studies include the Harvard sis, and made a final selection according to the six-city study (Dockery et al. 1993), the ACS criteria mentioned above. cohort study (Pope et al. 1995), and the ACS extended study (Pope et al. 2002). The main Cohort studies of long-term exposure background information and results are shown Cohort studies take advantage of spatial varia- in Tables 2.2 and 2.3. tion in air pollution concentrations to compare the incidence of disease and death in popula- Ecological studies of air pollution tions exposed over the long term to differing and human health levels of air pollution. By following large pop- There is no cohort study in China and only three ulations for many years, cohort studies estimate cross-sectional studies that reflect the effects of both numbers of deaths and, more impor- long-term air pollution exposure on mortality. tantly, mean reduction in life span attributable In China, Jing et al. (1999), Xu et al. (1996a, to air pollution. 1996b, 2000), and Wang et al. (2003) investi- Evidence from cohort studies of populations gated the chronic effects of air pollution on mor- in the United States indicates that long-term tality in Shenyang and Benxi. They estimated exposure to outdoor air pollution is associated relative risks by comparing mortality rates in the 24 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N F I G U R E 2 . 1 Urban Population Exposed to Class III and > Class III PM10 Levels, 2003 Source: Based upon Table 2.1 T A B L E 2 . 2 Background of Cohort Studies in the United States Authors Year Locations Pollutants Concentration Ranges Study Design Dockery et al. 1993 U.S. 6 cities PM10 18.246.5ug/m3 Cohort study Pope et al. 1995 U.S. 61 cities PM2.5 9.033.5ug/m3 Cohort study Pope et al. 2002 U.S. 61 cities PM2.5 Mean=17.7ug/m3 Cohort study Sources: Dockery et al. 1993; Pope et al. 1995; Pope et al. 2002. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 25 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N T A B L E 2 . 3 Main Results of Long-Term Cohort Studies in the U.S.A. Authors Health End Points Pollutants RR 95% C.I. Beta Std Error. Dockery et al. All Cause PM10 1.26 1.08,1.47 0.82 0.28 Lung Cancer 1.37 0.81,2.31 1.10 0.94 Cardiopulmonary 1.37 1.11,1.68 1.10 0.37 Pope et al. All Cause PM2.5 1.17 1.09,1.26 0.64 0.15 Lung Cancer 1.03 0.80,1.33 0.12 0.53 Cardiopulmonary 1.31 1.17,1.46 1.10 0.23 Pope et al. All Cause PM2.5 1.04 1.01,1.08 0.40 0.16 Lung Cancer 1.08 1.011,1.16 0.79 0.35 Cardiopulmonary 1.06 1.02,1.11 0.57 0.22 Source: Dockery et al. 1993; Pope et al. 1995; Pope et al. 2002. Note: In the study by Dockery et al., RR is the mortality-rate ratio for the most polluted of the cities as compared with the least polluted. In the studies by Pope et al., RR is the relative risk associated with a 10 µg/m3 change in particu- late pollution. Beta is the percentage increase in health effect per µg/m3 increment of air pollutant. worst-polluted and the least-polluted areas of ratio of 0.55 in 28 cities in China. We apply a each city. The background and main results are conversion ratio of 0.60 for PM2.5 to PM10 and a shown in Tables 2.4 and 2.5. ratio of 0.50 for PM10 to TSP. The results are shown in Tables 2.6 and Table 2.7. Transformation of TSP and PM2.5 to PM10 The particulate matter indices, including PM10, Time-series Studies of Short-term PM2.5, and TSP, differ in the above cohort stud- Exposure and Morbidity ies and ecological studies. In order to be applied in the ECM and compared with each other, we Time-series studies have been conducted to ana- convert to the uniform indicator index--PM10. lyze the relationship between daily rates of health Aunan and Pan (2004) suggest that the conver- events, such as hospital admissions or deaths, sion ratio of TSP to PM10 is 0.60. In Dockery's and daily concentrations of air pollutants and six-city study (Dockery et al., 1993), the ratio of other risk factors (e.g., weather). In time-series PM2.5 to PM10 is 0.60 to 0.64. Lvovsky et al. studies, individual factors--such as smoking, (2000) suggest that the ratio is 0.65. In the recent nutrition, behavior and genetic characteristics-- Chinese four- city study (Qian et al., 2001), the are unlikely to be confounders because they are ratio is 0.510.72. Wan (2005) found an average generally constant throughout the study period. T A B L E 2 . 4 Background of Ecological Studies in China Authors Year Locations Pollutants Concentration Ranges Study Design Jing et al. 1999 Benxi TSP 290620ug/m3 Cross-sectional ecological study Xu et al. 1996 Shenyang TSP 353560ug/m3 Cross-sectional ecological study Wang et al. 2003 Shenyang TSP 200540ug/m3 Cross-sectional ecological study Source: Jing et al. 1999; Xu et al. 1996; Wang et al. 2003. 26 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N T A B L E 2 . 5 Main Results of Cross-Sectional Ecological Studies in China Authors Health End Points Pollutants RR 95% C.I. Beta Std Error. Jing et al. All Cause TSP 1.08 1.02,1.14 0.077 0.028 COPD 1.24 1.04,1.44 0.22 0.083 CVD 1.24 1.08,1.41 0.22 0.068 CEVD 1.08 1.00,1.15 0.077 0.036 Xu et al. All Cause TSP 1.20 1.15,1.24 0.059 0.0063 COPD 1.22 0.065 CEVD & CVD 1.21 0.062 Coronary-heart-disease 1.11 0.034 Wang et al. CVD TSP 1.01 1.00,1.02 0.024 0.0087 Source: Jing et al 1999; Xu et al. 1996; Wang et al. 2003. Note: Beta is the percentage increase in health effects per 1µg/m3 increment of TSP. Various regression techniques are used to esti- impacts of SO2 on human health. In recent years, mate a coefficient that represents the relationship air pollution in urban areas in China has been between exposure to air pollution and human transformed from coal-smog air pollution into a health outcomes. The usual regression methods mixture of coal-smog and automobile exhaust. model the logarithm of the response variable, Emissions of SO2 have decreased gradually and such as daily deaths or hospital admissions, to esti- particulate matter has become the principal pol- mate the relative risk, or proportional change in lutant of concern in most cities in China. the outcome per increment of ambient pollution concentration. Table 2.8 presents the results of meta-analysis of time series morbidity studies T A B L E 2 . 6 Summary of the Results of Long-Term Exposure Studies (PM10) conducted in China (Aunan and Pan 2004). Authors Health End Points Beta Std Error Limitations of Previous Studies Dockery et al. 1993 All Cause 0.82 0.28 The exposure-response functions mentioned Lung Cancer 1.11 0.94 above are based on research conducted in China Cardiopulmonary 1.11 0.37 Pope et al. 1995 All Cause 0.38 0.09 and other countries during the past 10 years. A Lung Cancer 0.07 0.32 range of factors may affect the magnitude of the Cardiopulmonary 0.66 0.14 exposure-response coefficients. These factors Pope et al. 2002 All Cause 0.24 0.10 Lung Cancer 0.47 0.21 may have changed since the older studies were Cardiopulmonary 0.34 0.13 carried out, including the general health status Jing et al. 1999 All Cause 0.15 0.06 and living conditions of the population, and COPD 0.43 0.17 the level and composition of air pollution. The CVD 0.43 0.14 CEVD 0.15 0.07 main limitations of previous studies are the Xu et al. 1996 All Cause 0.12 0.01 following: COPD 0.13 CEVD 0.12 Change of air pollution level and type Coronary-heart-disease 0.07 Wang et al. 2003 CVD 0.041 0.02 Historically, air pollution in urban areas in China has come primarily from coal combustion. Up to Source: Dockery et al. 1993; Pope et al. 1995; Pope et al. 2002; Jing et al. 5 to 10 years ago, research was focused on the 1999; Xu et al. 1996; Wang et al. 2003. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 27 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N T A B L E 2 . 7 Exposure-Response Relationship for Long-Term Impact of PM10 on Mortality Rates Dockery Pope, 1995 Pope, 2002 Jing et al. Xu et al. Wang et al. 1 2 3 4 5 6 All Cause 0.82 0.38 0.24 0.15 0.12 Lung Cancer 1.11 0.072 0.47 Cardiopulmonary 1.11 0.66 0.34 COPD 0.43 0.13 CVD 0.43 0.041 CEVD 0.15 0.12 PM10 (µg/m3) 18.246.5 37.7 27.2 145310 178280 100270 Source: Dockery et al.1993; Pope et al.1995; Pope et al.2002; Jing et al.1999; Xu et al.1996; Wang et al.2003. Note: Beta is the percentage increase in health effect per 1µg/m3 increment of PM10. Limitation of study areas have not been undertaken in China. Most cross- The meta-analysis by Aunan and Pan (2004) sectional studies in China have been ecological was primarily based on the results from several studies, in which no detailed information at the large cities and not on middle and small cities. individual level is collected. This implies that the However, the characteristics of air pollution in studies in different locations may not be compa- middle and small cities are often different from rable due to site-specific characteristics. More those in large ones, and the age structure and importantly, the studies do not control for con- susceptibility to air pollution of the local popu- founding factors that may affect mortality (such lation may also vary with city size. So the extrap- as socioeconomic status), which may also be cor- olation of the exposure-response functions to the related with air pollution. For this reason, we rely other cities should be considered carefully. on the results of cohort studies from the U.S. (1995, 2002) in the manner described below. Limitations of methodology Large-sample epidemiological cohort studies, sim- Proposed Exposure-Response ilar to those carried out in the U.S. to study the Coefficients effects of long-term exposure to PM on mortality, Exposure-response coefficients for long-term exposure and mortality Table 2.8 Exposure-Response Relationships of PM10 Since impacts on all-cause mortality are reported and Morbidity Outcomes both in long-term cohort studies and ecological Health Endpoints Diseases Beta Standard Errors studies (the latter presumably representing the chronic effects of air pollution on mortality rates), Hospital admission RD 0.12 0.02 we select all-cause mortality as an endpoint in CVD 0.07 0.02 our assessment. New Cases Chronic Bronchitis 0.48 0.04 There are indications that the percentage change in the mortality rate per 1µg/m3 increment Source: Aunan and Pan (2004). Note: Beta is the increased percent of health effects per µg/m3 increment of PM10 changes with the concentration level. The of PM10. studies in the U.S. are all carried out in areas with 28 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N lower PM10 concentrations compared to Chinese in the logarithm of PM10 (Ostro 2004), imply- studies. The relative risk (RR) of dying at a PM10 ing that the relative risk function is given by concentration of C, compared to the threshold concentration of 15 µg/m3 for the studies reported RR = exp( + lnC ) exp( + ln15) in Table 2.6, is given by equation (2.1) = (C 15) (2.2) RR = exp(C) exp(15) = exp(C ) (2.1) Ostro adds 1 to both concentrations, to avoid where C is the difference between the current taking the logarithm of zero, so that equation (2.2) becomes: PM10 concentration and the threshold. This func- tion is plotted in blue in Figure 2.2 for = .0024 RR = ((C + 1) 16) (2.3) from the Pope et al. (2002) study. The Pope rel- ative risk function reaches 1.38 at a concentration When the data from Pope et al. (1995) are fit to of 150 µg/m3, implying (as explained below) that the log linear relative risk function, = 0.073 28 percent of deaths are premature deaths attrib- (s.e. = 0.028) (personal communication from utable to air pollution. This is clearly an implau- Rick Burnett, July 2006). This relative risk func- sible result. WHO (2004) dealt with this issue by tion (labeled Ostro RR) is plotted in Figure 2.3. assuming that the RR function becomes horizon- It coincides with the RR function based on (2.1) tal at approximately 100 µg/m3 of PM10, as shown with = .0012 at 150 µg/m3 and yields higher in pink on the graph. This assumption implies relative risks at lower PM10 levels. Figure 2.3 that there are no health benefits from reducing compares this RR function with the RR func- PM10 from 150 to 100 µg/m3! tion implied by equation (2.1) with = .0012. One alternative is to use as the RR function Equation (2.3) is used to compute the relative equation (2.1) with = .0012 from a meta- risks of PM10 concentrations in the CECM. analysis of cross-sectional Chinese studies (see figure 2.3.) These studies, however, were con- Exposure-response coefficients ducted in cities where average PM10 levels were for hospital admissions well above 150 µg/m3 and may not be applica- Few studies have been carried out in China ble to PM10 levels below 150 µg/m3. A compro- addressing hospitalization associated with air mise solution is to assume that exposure is linear pollution (HEI 2004; Aunan and Pan 2004). We F I G U R E 2 . 2 Comparing Relative Risk Functions 1.6 1.5 1.4 .0024 RR 1.3 .0024 TR 1.2 1.1 1 0 50 100 150 200 250 Source: Authors calculation. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 29 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N F I G U R E 2 . 3 Relative Risk Functions Based on U.S. and Chinese Studies COMPARING RELATIVE RISK FUNCTIONS 1.3 RISK 1.2 OSTRO RR .0012 RR 1.1 RELATIVE 1.0 0 50 100 150 200 250 PM10 Source: Authors calculation. apply the functions derived in Aunan and Pan Lanzhou, Wuhan, and Benxi). In the studies, (2004) to estimate the number of annual excess the definition of bronchitis was not precise in cases of hospital admissions for cardiovascular dis- terms of ICD-9 (or ICD-10) code, but was eases and respiratory diseases. The functions are described as "chronic" or "diagnosed by a physi- based on two time-series studies in Hong Kong cian." We assume that the endpoint approxi- and indicate a 0.07 percent (S.E. 0.02) increase in mates chronic bronchitis, and use the relative hospital admissions due to cardiovascular diseases risk function (2.1) with = .0048 for chronic per µg/m3 PM10 and a 0.12 percent (S.E. 0.02) bronchitis in adults. increase in hospital admissions due to respiratory diseases per µg/m3 PM10. The relative risks for CALCULATING HEALTH DAMAGES hospital admissions are given by (2.1) with the values of = .0007 and = .0012, respectively. With the identified health endpoints and exposure-response coefficients proposed earlier, Exposure-response coefficients the health effects from PM10 pollution consist of for chronic bronchitis three parts: (1) all-cause premature death; (2) hos- Aunan and Pan (2004) report an exposure- pital admissions for respiratory disease (RD) and response coefficient of 0.48 percent (S.E. 0.04) cardiovascular disease (CVD); and (3) new cases per µg/m3 PM10 for bronchitis in adults and of chronic bronchitis. 0.34 percent per µg/m3 PM10 (S.E. 0.03) for The number of cases of each health endpoint bronchitis in children. Altogether, eight cross- attributed to air pollution (E) is calculated as the sectional questionnaire surveys addressing a range size of the exposed population (Pe) times the dif- of persistent/chronic respiratory symptoms and ference between the current incidence rate ( fp) diseases were included in Aunan and Pan (2004). and the incidence rate in a clean environment All surveys were carried out in Chinese cities, and ( ft) [equation (2.4)]. The latter is calculated covered both urban and suburban areas. The from the fact that the current incidence rate coefficients for bronchitis are the result of a meta- equals the "clean" incidence rate times the rela- analysis of the sub-sample of odds ratios esti- tive risk, RR. Substituting (2.6) in (2.5) implies mated for this particular endpoint (given for that excess deaths are the product of current 30 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N deaths ( fpPe) times the fraction of deaths attrib- Premature mortality utable to air pollution--(RR-1)/RR. Formally, Current mortality rates, which vary by city size, E = ( f - f t Pe ) (2.4) are obtained from the China Health Statistical p Yearbook. f p = ft RR (2.5) Chronic bronchitis implying Calculating annual cases of chronic bronchitis E = ((RR - 1) RR) f pPe associated with air pollution requires an esti- (2.6) mate of the incidence of chronic bronchitis by city. Because only prevalence rates are available, we approximate the annual incidence of chronic Calculation of Baseline Incidence (fp) bronchitis by dividing the prevalence rate by the average duration of the illness (23 years). Hospital admissions This yields an incidence rate of approximately The Health Statistical Yearbook (Ministry of 0.00148. Health 2004) provides only all-cause hospital admissions by province, and not hospital admis- Excess Cases of Premature Mortality sions for specific diseases such as respiratory dis- and Morbidity Attributable ease. Another problem is that hospital admissions to Air Pollution by province include both rural and urban areas, whereas only the urban population is used to cal- By combining baseline cases of each health end- culate the health costs of air pollution. We esti- point with the selected relative risk functions, we mate hospital admissions for respiratory disease arrive at estimates of the number of excess cases in urban areas in two steps. First, we estimate the of premature mortality, hospital admissions, and number of hospital admissions due to respira- chronic bronchitis attributable to PM10. In addi- tory diseases by multiplying all-cause hospital tion to calculating the mean number of cases admissions by the ratio of respiratory diseases to attributable to outdoor air pollution, the 5th and all diseases by province. The percentage of respi- 95th percentiles of cases are also calculated. The ratory disease to all diseases is reported in the monetary value of these damages is presented in Analysis Report of the Third National Health Chapter 4. Services Survey (Ministry of Health Statistical Information Center 2003). This is based on an References assumption that the share of patients being admitted to the hospital for respiratory diseases Aunan, K. and X. C. Pan. 2004. "Exposure-response func- tions for health effects of ambient air pollution applica- resembles the share of people suffering from res- ble for China--a meta-analysis." Science of the Total piratory diseases among all people who are ill. Environment 329(1­3): 3­16. Second, we estimate the number of hospital Chang, G., et al. 2003. "Time-series analysis on the rela- admissions due to respiratory disease in the tionship between air pollution and daily mortality in Beijing." Journal of Hygiene Research 32(6): 567­567. urban population from the ratio of the urban Chang, G., L. Wang, and X. Pan. 2003. "Study on the asso- population to the total population. This is based ciation between ambient air pollutants and hospital out- on an assumption that the hospitalization rate patient visits or emergency room visits in Beijing, China." per case of disease is the same in urban and rural Chinese Journal of School Health 17(4): 295­97. areas, which is a crude approximation. Annex A China Statistical Yearbook 2004. Beijing: China Statistical Press. at the end of this chapter presents a detailed China City Statistical Yearbook 2004. Beijing: China Statis- description of the data sources. tical Press. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 31 H E A L T H I M P A C T S O F A M B I E N T A I R P O L L U T I O N Cohen, A. J., et al. 2004. "Urban air pollution." In M. Ezzati, Pope, C. A. et al. 1995. "Particulate air pollution as a pre- A. D. Rodgers, A. D. Lopez, and C. J. L. Murray, eds. dictor of mortality in a prospective study of U.S. adults." Comparative quantification of health risks: Global and American Journal of Respiratory and Critical Care Medi- regional burden of disease due to selected major risk factors, cine 151: 669­74. Vol 2. Geneva: World Health Organization. Pope, C. A. et al. 2002. "Lung cancer, cardiopulmonary Dockery, D. W., et al. 1993. "An association between air mortality, and long-term exposure to fine particulate air pollution and mortality in six U.S. cities." New England pollution." Journal of the American Medical Association Journal of Medicine 329(24): 1753­9. 287(9): 1132­41. Dong, J., et al. 1995. "Association between air pollution and Qian, Z., et al. 2001. "Long-term ambient air pollution lev- daily mortality in urban areas in Beijing in 1990­1991." els in four Chinese cities: inter-city and intra-city con- Journal of Hygiene Research 24(4): 212­14. centration gradients for epidemiological studies." Journal Dong, J., et al. 1996. "Association of air pollution with of Exposure Analysis and Environmental Epidemiology unscheduled outpatient visits in Beijing Longfu Hospital 11(5): 341­51. in 1991." Chinese Journal of Epidemiology 17(1): 13­16. Venners, S. A., et al. 2003. "Particulate matter, sulfur diox- Gao, J., et al. 1993. "Association between air pollution ide, and daily mortality in Chongqing, China." Envi- and mortality in Dongcheng and Xicheng districts in ronmental Health Perspective 111(4): 562­67. Beijing." Chinese Journal of Preventive Medicine 27(6): Wan, Y. 2005. "Integrated assessment of China's air pollu- 340­43. tion-induced health effects and their impacts on national HEI. 2004. Health Effects of Outdoor Air Pollution in Devel- economy." Ph.D. Thesis. Tokyo: Department of Social oping Countries of Asia: A Literature Review. Health Effects Engineering, Graduate School of Decision Science and Institute International Scientific Oversight Committee, Technology, Tokyo Institute of Technology. Special Report 15. Wang, H., G. Lin, and X. Pan. 2003. "Association between Jing, L., et al. 1999. "Association between air pollution and total suspended particles and cardiovascular disease mortality in Benxi City." Chinese Journal of Public Health mortality in Shenyang." Journal of Environment and 15(3): 211­12. Health 20(1): 13­15. Kan, H. and B. Chen. 2003. "Air pollution and daily mor- WHO (World Health Organization). 2000. Guidelines for tality in Shanghai: A time-series study." Archives of air quality. Geneva: World Health Organization. Environmental Health 58(6): 360­67. Xu, X., et al. 1995. "Association of air pollution with hospital Kan, H., J. Jia, and B. Chen. 2004. "The association of outpatient visits in Beijing." Archives of Environmental daily diabetes mortality and outdoor air pollution in Health 50(3): 214­220. Shanghai, China." Journal of Hygiene Research 67(3): Xu, X., et al. 1994. "Air pollution and daily mortality in res- 21­25. idential areas of Beijing, China." Archives of Environ- Lvovsky, K., et al. 2000. "Environmental Costs of Fossil mental Health 49(4): 216­22. Fuels--A Rapid Assessment Method with Application Xu, Z., et al. 1996a. "Effects of air pollution on mortality in to Six Cities." Washington, DC: World Bank. Shenyang City." Chinese Journal of Public Health 15(1): Ministry of Health, P.R. China. 2004. China Health Statis- 61­64. tical Yearbook 2004. Beijing: Chinese Academy of Med- Xu, Z., et al. 1996b. "Relationship between air pollution ical Sciences Press. and morbidity of chronic diseases in Shenyang City." Ministry of Health, Statistical Information Center, P.R. Chinese Journal of Public Health 15(2): 123­25. China. 2004. Analysis report of National Health Services Xu, Z., et al. 2000. "Air pollution and daily mortality in Survey 2003. Beijing: Chinese Academy of Medical Sci- Shenyang, China." Archives of Environmental Health ences Press. 55(2): 115­20. 32 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 1 3 Health Impacts of Water Pollution The poor quality of China's scarce China is facing a severe water shortage. In 45 percent of the national terri- water resources, which is increasingly tory, annual precipitation is less than 400mm (Zonggu Zhang and Dehong attributed to nonpoint sources such Shi). With a rapidly growing economy and burgeoning populations, the as agricultural runoff and municipal country's scarce water resources are seriously affected by pollution from wastewater, has a significant health the vast discharges of industrial and domestic wastewater, indiscriminate impact. The impact is particularly solid waste disposal, and runoff from an agricultural sector characterized by high in rural areas, where about excessive use of fertilizer and pesticides and large-scale livestock breeding. 300 million people lack access to piped water, as well as among Some 300­500 million people in rural areas do not have access to piped vulnerable groups, such as children water and are exposed to severe health risks related to polluted drinking under 5 years of age and women. water. Most urban residents have access to piped water that has been sub- This study attributes excess cases of ject to treatment. In smaller cities and townships, the drinking water qual- diarrhea and excess deaths due to ity guidelines are frequently violated in piped water and--to an even larger diarrhea among children under 5 in extent--in nonpiped types of water. rural areas (only in rural areas) to The complex and fragmented system for monitoring drinking water lack of safe water supply. The study resources (using different classification systems and sometimes showing also estimates the number of cancer contradictory patterns) complicates a comprehensive assessment of the health deaths in rural areas that are due to effects of polluted drinking water. In this chapter, we have attempted to the use of poor quality surface water quantify the health burden related to water pollution on excess diarrhea as a drinking water source. morbidity and mortality in children under 5 years of age as well as water- related cancer mortality in the general population. Although it seems clear that there are large health risks associated with water pollution in China, it could well be that the lion's share of the costs to society of polluted drinking water are avoidance costs--ranging from the cost of building treatment plants to the cost to households of buying bottled water and small-scale treatment devices. Water quality is monitored in more than 2,000 river sections across the main rivers in China (MWR 2005). About 25,000 km of Chinese rivers failed to meet the water quality standards for aquatic life and about 90 percent of the sections of rivers around urban areas were seriously polluted (MWR 2005). C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 33 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N Many of the most polluted rivers have been void and capacity of industrial wastewater treatment of fish for many years. Among the 412 sections facilities. However, discharges from the numer- of the seven major rivers monitored in 2004, 42 ous town and village enterprises (TVEs) and percent met the Grade I­III surface water qual- municipal sources are increasing rapidly and are ity standard, 30 percent met Grade IV­V, and causing extensive pollution of water bodies. 28 percent failed to meet Grade V (see figure The main pollutants are changing from heavy 3.1).1 Among these seven rivers, the Zhujiang metals and toxic organic chemicals, which are typ- River (79 percent) and Yangtze River (72 per- ically related to discharges of industrial wastewater, cent) had the largest share of sections meeting to pollutants from nonpoint sources. Runoff from Grade I­III. The Haihe River was the most pol- agriculture, including pesticides and fertilizers, is luted, with 57 percent of the monitored sections the single greatest contributor to nonpoint-source failing to meet Grade V (Figure 3.1). Major pol- pollutants. The consumption of chemical fertiliz- lutants contributing to poor water quality are ers nearly doubled in the period 1990­2004, and ammonia nitrogen, oxygen-demanding organic in the same period the use of nitrogenous fertiliz- substances (measured by BOD5), permanganate ers grew by 40 percent (China Statistical Yearbook value, and toxic petroleum compounds (SEPA 2005). Nonpoint sources are difficult to monitor 2004). Water pollution has penetrated beyond and in many cases more difficult to control than infecting the surface water found in lakes, rivers, point sources (Yu et al. 2003). and streams, and over half of the cities now have Treatment of domestic sewage has been lim- polluted groundwater (Siciliano 2005). ited until recently but is increasing steadily. In The amount of wastewater discharged from 1999, China had 266 modern wastewater treat- larger, regulated industries has leveled off since ment plants with treatment capacity account- the early 1990s due to an increase in the number ing for only 15 percent of the total 20.4 billion F I G U R E 3 . 1 Water Quality in Seven Major Rivers in China (percentage of river sections in different water quality classes) 100 % 80 % 60 % >V IV-V I-III 40 % 20 % 0 % Zhujiang river Yangtze river Yellow river Liaohe river Haihe river Songhuajiang Huaihe River river Source: SEPA, 2004 34 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N tons of domestic sewage. By the end of the tenth DRINKING WATER-- Five-Year Plan period (2001­2005), the gov- ACCESS AND QUALITY ernment put special emphasis on improving the There is a close relationship between water and situation in the three-river, three-lake drainage health. Water is an essential ingredient for main- area. (`The three rivers' are Huaihe, Liaohe and taining human life, but, when contaminated, it Haihe, and `the three lakes' are Taihu, Dianchi is also an important medium for the spread of and Chaohu. The total drainage area of the disease. To what extent polluted water resources three rivers and three lakes is 810,000 sq km, actually have an impact on people's health traversing 14 provinces with a total population depends on the society's capacity to treat sewage of 360 million.) Four-hundred-and-sixteen sew- and industrial discharges and to purify drinking age treatment plants in these areas have been water. It also depends upon whether people take completed or are under construction, with a action on an individual level to avoid consuming daily treatment capacity of 20.93 million tons polluted drinking water when treatment is absent (State Council 2006). By the end of 2004, the or not satisfactory. most people in China's urban rate of urban sewage treatment in China had areas have access to piped water--95 percent in reached 46 percent. 2003, according to the China National Health As evident from Figure 3.2, the major river Survey (CNHS).2 The corresponding figure in systems in the North are more heavily polluted rural areas is unclear. According to the survey, than those in the South, due to serious water 34 percent of the rural population has piped scarcity in northern China (MWR 2005). Pollu- water, while the reported average percentage on a tion levels are particularly high in the case of national level is about 50 percent (Table 3.1). ammonia nitrogen, dissolved oxygen, BOD These estimates are close to those based on data and permanganate, with 17­33 percent of from the nationwide China County Population monitoring sites not meeting class III drinking Census in 2000, which reported that 33 percent of water quality, mainly in the North (see figures rural households had access to piped water and 3.2A­2H). Moreover, population densities are 54 percent of the population in China on average had access to piped water. higher in the North, thus implying larger dis- The census data, which is based on a survey of charges of municipal wastewater into the rivers about 33 million households across rural and (see figure 3.3). urban China, also indicate that a large share of Among the 27 major lakes and reservoirs the population does not have access to adequate being monitored in 2004, none of them met the sanitary facilities, another factor of importance Grade I water quality standard; only two met the for the spread of waterborne infectious diseases Grade II water quality standard (7.5 percent); five (Table 3.2) (ACMR 2004). The health survey, met the Grade III quality standard (18.5 percent); supported by the census data, implies that about four met the Grade IV quality standard (14.8 per- 500 million people in rural areas do not have cent); six met the Grade V quality standard access to piped water. According to the Ministry (22.2 percent), and ten failed to meet the Grade V of Health, however, 61 percent of the rural pop- quality standard (37.0 percent). The "Three ulation had piped water in 2004 (Ministry of Lakes" (Taihu Lake, Chaohu Lake, and Dianchi Health 2005), which implies that about 300 mil- Lake) were among the lakes failing to meet the lion people in rural areas did not have access to Grade V water quality standard. The main pol- piped water. This is in accordance with figures lution indicators contributing to poor water from the Ministry of Water Resources.3 What- quality were total nitrogen and total phospho- ever the real figure is, it is clear that a substantial rus (SEPA 2004). number of people in rural areas still rely on well C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 35 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N Kilometers Miles 2,000 1,250 2004 Category Quality Quality Water 625 Water 1,000 N/A I II III IV V VI Overall Overall 500 312.5 3.2 0 0 FIGURE 36 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N B D 2004 2004 Levels Oxygen Levels Lead Pollutant Pollutant Dissolved Water Water A C ) continued( 2004 2004 2004 Levels Levels Quality Nitrogen Oxygen Water Pollutant Pollutant Ammonia Biological Water Water Overall 3.2 FIGURE C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 37 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F H 2004 2004 Value Levels Levels Oils Pollutant Pollutant Permanganate Water Water 2004 Value Permanganate E G ) continued( 2004 2004 2004 Levels Levels Phenol Quality Mercury Pollutant Pollutant Volatile Water Water Water Overall 3.2 FIGURE 38 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F I G U R E 3 . 3 Population Densities in China (number of people per square kilometer) Source: China County Population Census (ACMR 2004). water or water from rivers, lakes, and ponds. wise, dependence on other drinking water types Figure 3.4 shows the share of households with may not necessarily imply a health risk if the water access to piped water for each county in China source is protected from contamination. The cov- (census data). erage and technologies of treatment facilities for Having access to piped water, however, does piped water differ significantly across regions in not guarantee access to clean drinking water. Like- China. The most comprehensive treatment entails B O X 3 . 1 Pollution Accidents In addition to the continuous discharges of pollution into river systems, accidents may lead to tem- porary high levels of pollutants. In the aftermath of the accident in the Songhua River in northeast- ern China in November 2005, where a chemical plant released about 100 tons of the highly toxic chemical benzene into the river, the government carried out an inspection of 127 major chemical and petrochemical plants. The inspection found that many plants were located too close to major bodies of water and that 20 of the inspected plants had serious environmental safety problems. These plants included oil refineries and ethylene and methanol factories along the Yangtze River, the Yellow River, and the Daya Bay near Hong Kong. In the period November 2005 to April 2006, 76 more water pollution accidents were reported by the Chinese government (Associated Press 2006). C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 39 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N T A B L E 3 . 1 Proportion of Drinking T A B L E 3 . 2 Availability of Tap Water Water Types Among and Sanitary Facilities Households (urban in China (% of total and rural) households surveyed) Water Source Percent of Households Water type Piped water 45.7 Piped water 49.7 Nonpiped water 54.3 Hand pump 25.8 Bathing facility (hot water) Well 6.5 Centralized support of hot water 0.9 Rain collection 2.6 Water heater in home 15.4 Other facility for bath 9.7 Other (surface) 15.4 No facility for bath 74.0 Total 100.0 Lavatories WC in home 18.0 Sharing WC with neighbors 0.7 Source: 3rd National Health Service Survey 2003. Other type of lavatory in home 49.3 Sharing other type of lavatory with neighbors 3.9 No lavatory 28.0 Source: China County Population Census (ACMR, 2004). F I G U R E 3 . 4 Households with Access to Piped Water (% of total in county) Source: China County Population Census (ACMR 2004). 40 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N both sedimentation and disinfection, while much hepatitis may in the long run lead to cirrhosis and of the piped water is subject only to partial treat- liver cancer. High levels of chemical pollutants ment (either sedimentation or disinfection). The can cause acute poisoning, whereas--perhaps simplest form of treatment is chlorination. more importantly--long-term exposure to lower As seen in Figure 3.5, the level of coliform levels may lead to chronic health effects such bacteria is lower in groundwater than in most as cancers and may enhance the risk of adverse other drinking water sources. The level of fluo- pregnancy outcomes such as spontaneous abor- ride and arsenic, however, is highest in ground- tions and birth defects. water. With respect to the mean values for all The disease matrix presented in Figure 3.6 on 300 counties for the different drinking water the next page summarizes the different health types, drinking water quality guidelines (Class I) outcomes that may result from drinking polluted are violated only for total bacteria and total col- water. The matrix is based on information collated iform bacteria (see table 4). from WHO's Guidelines for Drinking Water Qual- The violations are large for some water types. ity (WHO 1996) for a number of biological and Among the other indicators, however, there is chemical pollutants, as well as a comprehensive large variability in the measurements for many literature search, which was conducted in order water types, implying that guidelines must be vio- to investigate associations between biological/ lated in many samples (see figure 3.5). It should heavy metal pollutants in drinking water and be noted that Class I of the drinking water qual- possible health outcomes. Most of the retrieved ity guidelines (Table 3.3) represents the na- studies gave a measure of effect or provided the tional standard and applies to urban areas. This means to calculate it (see appendix 1 for details of means that piped water produced by treatment the studies). The objective of this exercise was not plants in urban areas should not violate the Class to retrieve all possible publications on the associ- I standard. In rural areas, Class III applies. ation between drinking water pollutants and As evident from Figure 3.5, nitrate is leaking health outcomes (i.e. this was not a systematic into groundwater, since the level in groundwater review), but rather to collect enough evidence and spring water is not lower than in surface water bodies. T A B L E 3 . 3 Drinking Water Quality Standards for China CAUSAL AGENTS AND IMPACT PATHWAYS Class I Class II Class III Water pollutants can be categorized into two main Chrome (degree) 15.0 20.0 30.0 types--biological pollutants (including microor- Turbidity (degree) 3.0 10.0 20.0 ganisms causing infectious hepatitis A or E, dysen- Total dissolved solids (mg/L, CaCO3) 450.0 550.0 700.0 tery, typhoid fever, cholera and diarrhea),4 and Iron (mg/L) 0.3 0.5 1.0 chemical pollutants (including inorganic sub- Manganese (mg/L) 0.1 0.3 0.5 stances such as nitrates, phosphates, mercury, COD (mg/L) 3.0 6.0 6.0 Chlorate (mg/L) 250.0 300.0 450.0 arsenic, chrome, fluorine and lead, and organic Sulfate (mg/L) 250.0 300.0 400.0 compounds such as phenols, benzene and other Fluoride (mg/L) 1.0 1.2 1.5 aromatic compounds, and oil). While infectious Arsenic (mg/L) 0.1 0.1 0.1 Nitrate (mg/L) 20.0 20.0 20.0 diseases typically occur as an acute effect, expo- Total bacteria (/mL) 100.0 200.0 500.0 sure to biological pollutants may also have long- Total coliform (/L) 3.0 11.0 27.0 term health implications. For instance, chronic intestinal infections may develop and infectious Source: SEPA C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 41 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N 2004 water water spring spring Areas, data. groundwater Rural groundwater the in ni Reservoir Reservoir errors season Water season be rainy- rainy- Lake may Lake bacteria Drinking Fluoride water there Total pool of pool or water spring or and water dike dike new Sources River groundwater River relatively 0 is 000 000 000 000 000 000 000 000 000 000 8 6 4 2 20 18 16 14 12 10 Reservoir 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0.00 Different Lm/UFC L/gm in season water rainy- Lake Nitrate drinking arsenic) water water pool spring or spring rural and water of dike (April­October) nitrate, groundwater groundwater River Season monitoring as 5.00 0.00 Reservoir 30.00 25.00 20.00 15.00 10.00 Reservoir season L/gm Rainy bacteria, season rainy- caution the rainy- in total Lake Lake with Coliform Arsenic Total pool pool or or Quality coliform, water water dike dike interpreted ater be W (total River River should 3.5 0 800 600 400 200 000 800 600 400 200 Beijing. 1 1 1 1 1 0.035 0.030 0.025 0.020 0.015 0.010 0.005 0.000 data Lm/airetcab fo oN L/gm CDC, These FIGURE Source: Note: 42 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N to show possible associations. Studies relating to At concentrations as low as 0.9­1.2 mg/liter, mortality as an outcome have not been incorpo- fluoride could cause dental fluorosis. However, rated in the presented matrix. with higher concentrations, it may cause skeletal From the matrix, it is clear that there are pos- fluorosis and with even more elevated levels it itive associations between exposure to chemical may result in crippling skeletal fluorosis. There pollutants, namely nitrate/nitrite and arsenic, is inconclusive evidence on fluoride carcino- and at least nine different malignancies. How- genicity in humans. ever, studies on arsenic seemed to be more estab- Lead has been shown to cause renal disease lished than those on nitrate/nitrites. Arsenic is and may in some cases be associated with chronic responsible for inducing several malignant tumors nephropathy, particularly with prolonged expo- affecting epithelial tissues including skin, liver, sure. Furthermore, strong associations have been lung, bladder and kidney. It is also associated documented between blood lead levels in the with cardiovascular, respiratory and neurologi- range of 7­34 microgram/dl and hypertension. cal disorders. From another perspective, there is very little evi- Much research has been conducted on the dence, if any, that lead is a human carcinogen. effect of nitrate in drinking water on human The international literature clearly docu- health; but, with considerable controversy over ments associations between fecal coliforms/total some of the outcomes investigated, particularly bacteria and diarrhea. Moreover, associations gastric cancers. Nevertheless, several epidemio- between these biological pollutants and other logical studies have demonstrated increased risk digestive outcomes such as cholera, dysentery, for bladder, ovarian and colorectal cancers asso- gastroenteritis, giardia, salmonella, typhoid, and ciated with ingestion of nitrates in drinking shigella have also been established water. Furthermore, there is strong evidence that nitrates are also associated with an increased risk HEALTH AND CHEMICAL of non-Hodgkin's lymphoma. WATER POLLUTANTS Many epidemiological studies have looked into the impact of prolonged ingestion of fluoride in Chemical pollution of water resources may be drinking water and have concluded that it pri- due to natural conditions. In China, chronic marily affects skeletal tissues in different degrees endemic arsenism is among the most serious depending on the concentration levels detected. endemic diseases related to drinking water (Xia T A B L E 3 . 4 Exceeding Drinking Water Quality Standards for Total Bacteria and Total Coliform Bacteria in Drinking Water Types in China (ratio between the mean value of samples from 300 rural counties and the guideline value, Class I) Piped Piped Water Piped Water Quality Water (partially Water Nonpiped Nonpiped Nonpiped Indicator (treated) treated) (untreated) (by machine) (manual) (hand pump) Total bacteria 6.6 7.2 5.8 8.3 8.0 4.9 Total coliform 11.4 23.7 18.7 20.4 103.1 12.4 Source: National CDC, Beijing. Note: These data should be interpreted with caution as monitoring of rural drinking water is relatively new and there may be errors in the data. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 43 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F I G U R E 3 . 6 Matrix for Biological/Chemical Drinking Water Pollutants and Health Outcomes Biological Pollutants Chemical Pollutants Health Outcome Fecal Total Nitrate/ Fluoride Lead Arsenic coliform bacteria Nitrite Malignancies Bladder Cancer · · Colorectral Cancer · Gastric Cancer · Liver Cancer · Lung Cancer · Renal Cancer · Skin Cancer/ Pre-malignant Lesions · Ovarian Cancer · Non-Hodgkin Lymphoma · Cardiovascular Peripheral Vascular Disease · Hypertension · · · Respiratory Bronchiectasis · Bone/Skeletal Deformities Bone Deformity · Dental Fluorosis · Skeletal Fluorosis · Neurological Central Nervous System Defects · Mental Retardation · Peripheral Neuropathy · Digestive Cholera · Diarrheal Diseases · · Dysentery · · Hepatitis · · Typhoid Fever · · Hepatomegaly (enlarged liver) · Pregnancy-Related Adverse birth outcomes · · Spontaneous Abortion · Other Renal Dysfunction · · Diabetes Mellitus · Source: WHO 1996 and authors' calculations. 44 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N and Liu 2004). High levels of arsenic (As) in Mortality rates in China for cancers associated drinking water are attributable to the geological- with water pollution are shown in Figure 3.7, geochemistry environment. In China, high lev- along with the world average rates. For stomach, els of As in groundwater are mainly found in the liver, and bladder cancer, the rates are highest in (a) plain of the Great Bend of the Yellow River rural areas. For liver and stomach cancer in par- and the Hu-Bao Plain in the Inner Mongolia ticular, the mortality rates in China are well above Autonomous Region; (b) the Datong basin of the world average. Liver cancer is the most preva- Shanxi Province; (c) the floodplain of the north- lent type of cancer in rural China.6 ern side of the Tian Mountain of Xinjiang Uygur Autonomous Region; and (d) the southwest HEALTH AND BIOLOGICAL coastal plain of Taiwan (Lin et al. 2002). Epi- WATER POLLUTANTS demiological studies have shown that high levels of As in drinking water are associated with skin Drinking contaminated water is typically only cancers and other cancer, hypertension, and one of several ways of contracting infectious dis- peripheral vascular diseases. Dose-response rela- eases. Pathogens may also be spread by food and tionships are reported for some health endpoints. by flies, and because pathogens are spread by It is estimated that 2.3 million people are exposed direct contact, hygiene is of primary importance. to high levels of As (>0.05mg/L)5 through drink- Figure 3.8 below portrays the famous F-diagram, ing water (Xia and Liu and references therein, which captures the potential exposure pathways MWR 2005). About 7,500 patients were diag- for fecal-oral transmission. nosed with arsenism in the areas surveyed by the The F-diagram clearly shows that breaking the Ministry of Health in 2003 (MoH 2004). fecal-oral transmission route, and thus reducing Geological conditions may also lead to high infections, is not entirely dependent on the avail- levels of fluorine (F) in drinking water. Accord- ability of clean water but also depends on other ing to the Ministry of Water Resources (2005), factors, including safe disposal of feces (safe 63 million people drink water with high con- sanitation), hygiene behaviors--especially hand- centrations of fluorine. Endemic dental fluoro- washing with soap after defecation--and safe sis related to drinking water affected 21 million food handling and storage. Generally, disease people in 2003, whereas 1.3 million suffered from incidence is high in areas, where basic sanitary drinking-water-related skeletal fluorosis (MoH facilities are lacking.7 Water scarcity may also 2004). Water-pollution-related fluorosis is more enhance the spread of infectious diseases. Conve- prevalent in the northeast and central part of the nient access to sufficient water quantity encour- country, but cases are reported in nearly all ages better hygiene and, therefore, limits the provinces. spread of disease.8 In the environmental cost model, we do not As opposed to the chronic diseases arising from include the natural water contaminants but long-term exposures to carcinogenic pollutants, focus on anthropogenic pollutants in drinking the incidence rates for infectious diseases may vary water and their potential health risks. Although substantially from one year to the next and from the health effects of natural and anthropogenic one season to another (see Box 3.2 for Chongqing pollutants may overlap, and it may be difficult study). Outbreaks of the disease may cause very to disentangle the individual contribution of high incidence rates in an area during a limited the two types--for instance, for cancers--it is period of time. The fatality rates for these diseases believed that anthropogenic pollution of drink- are, however, relatively low. The case-fatality rates ing water is the most important in today's for dysentery, typhoid/paratyphoid, and cholera China. in China in 2003 were on average 0.05 percent, C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 45 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F I G U R E 3 . 7 Mortality Rate for Diseases Associated with Water Pollution (1/100,000) in China, 2003 (world averages in 2000) 35 30 Major cities Medium/small cities 25 Rural World average 20 15 10 5 0 Oesophagus cancer Stomach cancer Liver cancer Bladder cancer Source: MoH, 2004, and WHO, 2006, GLOBOCAN, 2000 0.06 percent, and 0.41 percent, respectively. As shown in Box 1, the incidence rate of water- pollution-related infectious diseases is highest in children. The death toll is also highest in children, F I G U R E 3 . 8 F-Diagram for Fecal-oral Transmission particularly for diarrheal diseases. In China, the mortality due to diarrhea in children under five in rural areas is nearly twice the rate in urban Fluids areas (1.35 vs 0.75 deaths per 100,000 children) (MoH 2004). Figure 3.9 shows the distribution of cases of Hepatitis A, dysentery, and typhoid/paratyphoid Fingers fever across provinces of China in 2003. Gener- ally, higher rates prevail in western parts of the New Feces Food country. Dysentery is the most frequent of the Host water-related infectious diseases. In spite of outbreaks every year, the occurrence Flies of dysentery has fallen dramatically in China in the last decades and now seems to be stabilizing (see figure 3.10A). The worst outbreak occurred in 1975, when an incidence rate of 1,000 per Fields 100,000 was reported. As evident from Table 3.5 and Figure 3.10 outbreaks of typhoid/paraty- phoid fever are rarer than dysentery, with an inci- 46 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N B O X 3 . 2 Drinking Water and Waterborne Infectious Diseases in Rural Areas in Chongqing As part of this study, specific work in Chongqing found that the rural population in this province faces many challenges with respect to drinking water supply. Not only do rural areas have limited access to piped water as compared to urban areas, but most of the piped water undergoes little treatment and has significant levels of contamination. This makes the rural population more sus- ceptible to some waterborne infectious diseases. Drinking Water in Chanqing People in Chongqing Province get their main drinking water supply from a variety of sources, including centralized piped water as well as wells, ponds, rivers, and ditches. Generally, the decentralized (nonpiped) water sources undergo no treatment and are less safe for drinking water purposes. Monitoring of drinking water across rural areas in Chongqing shows that the level of the coliform group bacteria (an indicator of fecal contamination) in nonpiped drinking water is about ten times the level in piped water, and there are more frequent incidents of extreme bacteria levels in the rural nonpiped water. In Chongqing, only around 30 percent of the population has access to piped water (China Cen- sus 2000), and, as elsewhere in China, most of this population resides in the largest urban centers. However, only a fraction of the piped water supply undergoes comprehensive treatment before it reaches the end users. Among the 10 counties and urban districts for which detailed information of water supply is available, the share of the population that has access to comprehensively treated piped water is less than a third in all counties/districts, and on average only 15 percent of the population has access to comprehensively treated piped water. Share of population with centralized and decentralized water system according to a) treatment (for centralized) and means of distribution (for decentralized), and b) source (surface water or underground water) Decentralized Pump 6% Piped groundwater Decentralized by 2% Machine 9% Decentralized Piped No Treatment Piped surface Decentralized groundwater 14% 35% manual 42% 48% Piped Disinfection 2% Piped Sedimentation & Disinfection 6% Piped Comprehensive Decentralized treatment surface 15% 21% (a) (b) Source: Authors calculations. The degree of treatment of piped water also varies between urban and rural areas. While drink- ing water treatment plants in cities and to some extent in smaller townships provide comprehensive treatment of the water through sedimentation and disinfection, a large share of the piped water in townships and villages undergoes only limited treatment (either sedimentation or disinfection or chlorination). The overall effectiveness of treatment is, therefore, very limited. According to 2001­04 monitoring data from 100 township treatment plants in 14 countries upstream of the Three Gorges area, the levels of a number of contaminants--such as arsenic, fluoride, and nitrate--were not sig- nificantly affected by treatment. Treatment did reduce the total bacteria content, but the resulting water still had on average 83 percent more coliform bacteria than permitted by the national stan- dard for drinking water quality (Class I). The mercury in the treated water was on average 38 percent above the standard, and the levels of heavy metals like arsenic (As) and cadmium (Cd) were on aver- age approximately 3 percent higher than the standard. (continued) C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 47 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N B O X 3 . 2 Drinking Water and Waterborne Infectious Diseases in Rural Areas in Chongqing (Continued) The urban/rural discrepancy is also evident in the water sources, which have strong implications on drinking water quality. While the piped water in the 10 counties and urban districts mainly comes from surface water, the decentralized sources are mainly underground water (wells and springs) (see figure above). Most of the underground water is from shallow wells, however, which are easily contaminated by wastewater and runoff from industry, agriculture and households. In figure 1b, the combination of source and treatment most likely associated with the highest risk of infectious diseases is decentralized surface water, on which 21 percent of the population is depen- dent. In addition, untreated piped surface water probably entails a correspondingly high risk. As nearly 40 percent of the piped water is sent untreated to the end-users, and most of the piped water is from surface water bodies, a substantial share of the total population in these areas-- around 13 percent--has untreated surface water in their tap. The measurements showed that the water quality in Chongqing varies substantially from year to year and between seasons. The study found that the median values of main contaminants, as Coliform bacteria, total bacteria, As, and Hg, did not fluctuate very much from year to year in the period 2001­2004. The mean values did, however, fluctuate considerably, indicating that during some of the years, there were incidents of very high pollution levels for a shorter period of time. For most of the water pollution indicators, the noncompliance rate is higher in the flooding sea- son compared to the dry season. Waterborne infectious diseases in Chongqing Hepatitis A, dysentery,9 and typhoid/paratyphoid fever are three main types of infectious diseases associated with polluted drinking water. Fatality rates for all three diseases in Chongqing are low, indicating that few people die from these diseases, but the annual incidence rates vary. Whereas the incidence rate of Hepatitis A is somewhat higher in Chongqing compared to the average inci- dence rates in China (12.3 vs. 7.4 cases per 100,000 in 2003), the incidence rates for the two other diseases are lower (for dysentery 27.9 versus 34.5 cases per 100,000, and for typhoid/paratyphoid 0.9 versus 4.2 cases per 100,000) (MoH 2004). The incidence rates are, however, high relative to those found in European countries and the United States, which indicates that there is still a lot of potential for improvement. The study found that outbreaks of infectious diseases vary considerably from year to year and are generally more frequent in the flooding season as compared to the dry season. As shown in the figure below, there may be large differences between counties when it comes to outbreaks. In Chongqing, outbreaks of the three different diseases occurred independently during the period 2001­04, with no spatial correlation between them. Children in Chongqing are much more likely than adults to contract infections diseases, particu- larly dysentery and typhoid fever. The incidence rate for children under 5 years of age is 10 per- cent higher than the rest of the population. For hepatitis A, incidence rates are also markedly higher in children and adolescents; however, a considerable share of the cases also occur in the older age groups (see figure below). The data available from Chongqing, though limited, show a significant correlation between the level of total bacteria in drinking water10 and incidence rates for dysentery. While the inci- dence rate of typhoid was associated with total bacteria to some extent (for females), hepatitis A did not show a clear association with the total level of bacteria or coliform group bacteria as mon- itored in the rural drinking water. (continued) 48 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N B O X 3 . 2 Drinking Water and Waterborne Infectious Diseases in Rural Areas in Chongqing (Continued) Incidence of (a) Dysentery, (b) Hepatitis A, and (c) Typhoid in 19 Counties in Chongqing 250 45 2001 2002 40 2001 2002 200 2003 2004 35 2003 2004 30 150 25 20 1/100,000100 1/100,000 15 10 50 5 0 0 0- 1- 5- 10 15 20 25 30 35 40 - - - - - - -45-50 55 60 65 70 - - - - -75- 80 85 - ?? ?? 0- 1- 5- 10 15 - - 20- 25-30 -35- 40-45-50-55 -60-65-70-75 80 - -85 ?? ?? 30 2001 2002 25 2003 2004 20 15 1/100,000 10 5 0 0- 1- 5- 10- 15 20 - - 25- 30 35 - -40- 45-50- 55 - 60- 65 70 - - 75- 80- 85 ?? ?? Incidence rates of (a) Dysentery, (b) Hepatitis A, and (c) Typhoid Among Female Age Groups in Chongqing, 2001­04 dence rate of around 4 cases per 100,000 in 2003. CHINESE STUDIES OF HEALTH This rate is also much lower than the world aver- EFFECTS OF DRINKING age. The annual incidence of typhoid worldwide WATER POLLUTION at present is estimated to be about 283 cases per 100,000.11 The incidence rate of cholera has been The basis for reliably estimating the full public low in China in recent years--0.02 cases per health implications of drinking water pollution 100,000 were reported in 2003. on a population level in China is limited for sev- Domestic sewage and agricultural runoff may eral reasons. First, there are relatively few studies lead to eutrophication of water bodies. In addition in China and other developing countries address- to the health risk associated with the eutrophio- ing the exposure-response relationship between cation agents themselves, including nitrates and drinking water pollution and health effects. Water phosphates, eutrophication supports the growth pollution epidemiology and its application is of cyanobacteria that can produce toxins such as severely hampered by the fact that a range of other microcystins. These are potent liver cancer pro- factors contribute to disease. Contaminated water moters and are directly hepatotoxic to humans. is typically one of several ways of contracting Microcystins in drinking water cannot be com- infectious disease and is closely linked to sanitation pletely removed by common disinfection and and hygiene, as discussed above. Similarly, in the heating (MWR 2004 and references therein: case of chemical pollutants, contaminated water is Wang et al. 1995; Ling 1999). also one of several ways of getting ill. Enhanced C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 49 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F I G U R E 3 . 9 Incidence Rates of Hepatitis A, Dysentery, and Typhoid/Paratyphoid Fever in China in 2003 (1/100,000) Source: MoH 2004. rates of disease may as well be related to occu- amination with hepatitis A virus and other pational exposure, smoking, food, and other microorganisms. As the environmental resistance, life-style factors. and thus lifetime, of the virus is higher than fecal Secondly, even when exposure-response func- bacteria, it may not be a quantitative association tions are available, the exposure assessment is between the content of virus and bacteria down- complicated by the fact that--in contrast to air stream of the discharge area (Fernández-Molina pollution--there is a larger scope for people et al. 2004). Clearly, this implies that an expo- being protected from exposure. Finally, the assess- sure assessment based on water quality at a lim- ment of attributable risk from water pollution is ited number of monitoring sites will be rather complicated by the fact that in many cases the uncertain. causal agent of disease is not monitored directly, In China, most studies addressing health effects and indicators of exposure are needed. For from water pollution have looked at drinking instance, even though hepatitis A is one of the water pollution and cancers. Su De-long (1980) most prevalent waterborne infectious diseases in explored causal factors of liver cancer in Qidong developing countries, China included, hepatitis county of Jiangsu Province, and found that the A virus is rarely monitored directly. Instead, fecal morbidity of liver cancer was closely related to bacteria are used as an indicator of possible cont- drinking water contamination. Xu Houquan et al. 50 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N F I G U R E 3 . 1 0 Incidence Rates for Dysentery (A), Typhoid/Paratyphoid Fever (B), and Cholera (C) in China, 1985­2003 350 300 250 200 (1/100,000) rate 150 100 Incidence 50 0 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 16 14 12 10 (1/100,000) 8 rate 6 Incidence 4 2 0 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 3.5 3 2.5 2 (1/100,000) rate 1.5 1 Incidence 0.5 0 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 Source: MoH 2004. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 51 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N mortality rates for liver and esophagus cancer T A B L E 3 . 5 National Average Incidence Rates of Hepatitis A, among residents relying on groundwater that was Dysentery, Typhoid/ contaminated by sewage was significantly higher Paratyphoid Fever, and compared to people in the control area. Pan and Cholera in China in 2003 (1/100,000) Jiang (2004) investigated the correlation between various water quality indices in drinking water Incidence Rate and the mortality rates of a range of cancer types Disease (1/100,000) in the Yangtze and Huai He river basins in the period 1992­2000. They found a significant Hepatitis A 7.37 positive correlation between the level of COD Dysentery (viral and amebic) 34.52 Typhoid and paratyphoid 4.17 (chemical oxygen demand), fluorine, and chlo- Cholera 0.02 ride, and male stomach cancer. A number of studies have also examined the Source: MoH 2004. effects of water pollution on infectious diseases in China. Pan and Jiang (2004) investigated the cor- (1995) carried out a case-control study of risk relation between coliform group bacteria and the factors of liver cancer around the Nansi Lake, integrated water quality index (IWQI), which Shandong Province. They showed that drinking includes a wide range of water quality indicators, lake water, getting in touch with lake water, in drinking water and the incidence rates of infec- drinking alcohol, and eating fish were all risk fac- tious diseases in the Yangtze and Huai He river tors for liver cancer. The estimated odds ratios basins in the period 1992­2000. They found were 6.55, 3.24, 1.86, and 2.55, respectively. Xu significant correlations between the level of col- Houquan et al. (1994) carried out a retrospective iform group bacteria in drinking water and the cohort study on the relationship between water incidence rates of diarrhea, and between the IWQI pollution and tumors and showed that the mor- and incidence rates of typhoid/paratyphoid and tality rates of stomach, esophagus, and liver can- diarrhea for both men and women. No correlation cer for people drinking lake water were higher was found for either index regarding the inci- than those drinking well water. The relative risks dence of hepatitis A. They also show that there (RR) were 1.56, 1.50, and 1.63. The nationwide was a strong correlation between the level of col- study on organic pollution of drinking water and iform group bacteria in surface water and drink- liver cancer by Wang Qian et al. (1992) showed ing water in rural areas in the two river basins. that mortality due to liver cancer for men and Because monitoring sites were changed during women was positively correlated with the chem- the period, the number of counties for which ical oxygen demand (COD) in drinking water. both disease data and drinking quality data were In a 16-year retrospective cohort study in an area available was somewhat limited in the study. with enhanced stomach cancer incidence rates, In the present study, we found no statistically Wang Zhiqiang et al. (1997) found that mortal- significant relationship between the level of total ity due to stomach cancer in people drinking coliform bacteria in rural drinking water and inci- river water was significantly higher than in peo- dence rates for infectious diseases. However, due ple drinking well water. Monitoring data showed to lack of data, it was not possible to control for the high levels of ammonia, nitrite, chloride, COD, range of possibly important confounding factors and heavy metals like lead and mercury, suggest- in the analysis. The data included in the analysis ing that drinking polluted water is one of the were incidence rate data for 2004 for infec- causal factors of stomach cancer. In Baoding city tious diseases from the National Infectious Re- in Hubei Province, Hu (1994) reported that the porting System (dysentery, hepatitis A, typhoid/ 52 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N paratyphoid, diarrhea) and data for rural water Exposure assessment quality from the National Rural Water Quality The survey categorized drinking water sources into Monitoring System provided by the China Cen- five types, as shown in Table 3.6 below. Although ter for Disease Control and Prevention (CDC). all sources except for surface water (drawn directly from rivers, lakes, pools, canals, ditches, and VALUATION MODELS IN THE ECM house drains) are considered to be relatively safe in China, this analysis took a more conservative Estimating Excess Diarrheal Disease approach in considering piped water as the only Morbidity in Children safe drinking water source. An earlier survey esti- mated that about 50 percent of the population in Data from the Third National Health Service Class IV rural areas still drinks water not meet- Survey--prepared by the Health Statistics and ing the national sanitary standards.12 This could Information Center in the Ministry of Health mean that at least 30 percent of this population (MoH)--were used to derive exposure-response (considering that surface water accounts for nearly functions for diarrhea in children under 5 years 20 percent) are using unsafe water sources that of age in rural China. The survey was carried out includes hand pumps, wells (i.e. underground in 95 counties in 31 provinces and incorporated water), and rain collection. Therefore, classifying questions pertaining to household characteristics, piped water as the only safe source is the most individual factors, disease prevalence, and costs conservative approach. for treatment. Since no indicators directly related to water quality were included in the survey, we Sample population description decided to use a conservative approach in the After data cleaning and reduction, we conducted analysis by considering access to piped water as a descriptive analysis to highlight the socioeco- the safest drinking water source. We recognized nomic and demographic profile of the 7,103 that piped water does not inherently equal safe, sampled rural households with children less than clean water, and that there are sometimes viola- 5 years of age (see table 3.7). tions of drinking water quality standards, partic- From the table, it is clear that a small number ularly in rural China. But given the absence of of the surveyed households (2.9 percent) were data on water quality in the survey, we used piped officially listed as poor. However, the majority of water as a proxy indicator for safe water. The role these households (84 percent) had an income of of sanitation was considered in the analysis and less than 3,000 RMB, while half of them were was controlled for using multivariate modeling. spending more than 500 RMB on health-related The role of hygiene, however, was not assessed costs--indicating that a very high share of a due to the lack of any indicators related to it, limited income was used for health care. especially handwashing. The present analysis and Nearly 88 percent of mothers in these house- results are based on a rural household-level rather holds had some education (most had completed than an individual dataset to mitigate any disease- either primary or secondary education). Nearly clustering effects that may exist. one-fifth of the surveyed population did not have access to safe sanitation and hence relied on defe- Outcome of interest cation in the open. Household diarrhea prevalence was estimated by calculating the proportion of households with one Two-week household diarrhea prevalence or more diarrhea cases in children under 5 years The two-week prevalence for household diarrhea relative to the total number of households with in rural China was computed and accounted for children less than 5 years of age in rural China. 2.2 percent of the total number of households. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 53 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N · Maternal education. An ordinal categorical T A B L E 3 . 6 Water and Sanitation in Rural China variable comparing risk of diarrhea for the dif- ferent educational levels relative to the baseline Source Percentage group of no education. Water (n=7,036) Piped Water 33.6 Table 3.8 shows the results of the binomial Hand Pump 33.3 Well 8.2 regression quoting risk ratios (and 95 percent Rain Collection 3.3 confidence intervals) and their respective P-values Other (Surface) 21.7 for diarrhea. As the table indicates, the regression analy- Source: Third National Health Service Survey, 2003 sis found that the risk for diarrhea in house- holds with piped water (safe water proxy) is Exposure-Response Functions 0.66 times less than households with no access (i.e., risk is 1.52 if comparing no access to piped Multivariate modeling water versus access), which is significant at the A binomial regression model was used to estimate 5 percent level the risk of diarrhea in households with no access to piped water versus those with access after controlling for confounders. Estimating excess annual number The final model included piped water as the of diarrheal episodes in rural China exposure variable (comparing those with piped water to those without) and included the follow- Given that: ing covariates (confounders): ­ The crude two-week prevalence for household · Sanitation. A binary variable comparing risk of diarrhea is 2.2 percent; diarrhea in households with a sanitation facility ­ The risk for diarrhea in households with no relative to those with none. piped water is 1.52 more than households with · Income. A continuous variable showing percent piped water supply; and of diarrhea risk change for every unit increase ­ The proportion of households with no piped in income. water is 0.664. T A B L E 3 . 7 Socioeconomics and Demographics Listed Poor Expenditure Households Rural Class Income on Health (n=7,042) (n=7,042)(%) (n=7,042) (%) (n=7,042)(%) 2.9% High Economic 18.7 <1000 RMB 26.4 <100 RMB 5.2 Medium-High 28.1 1000­1699 RMB 34.2 100­299 RMB 27.4 Medium-Low 35.4 1700­2999 RMB 23.1 300­499 RMB 18.1 Low Economic 17.8 3000 RMB 16.4 500 RMB 49.3 Maternal None 12.4 Ethnicity Han 78.1 Education Primary 32.8 (n=7,028) Other 21.9 (n=6,188) Secondary 47.7 Sanitation Safe 81.5 Higher 7.1 (n=7,038) Unsafe 18.5 Source: Third National Health Service Survey, 2003. 54 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N And if: T A B L E 3 . 8 Binomial Regression for Risk Ratio AFp = [Pe (RR - 1)] [1 + Pe (RR - 1)] for Diarrhea Where: Risk Ratio 95% CI P-value AFp = population attributable fraction Pe = Prevalence of risk exposure in the pop- Piped Water 0.66 0.44­0.99 0.046 Sanitation 0.64 0.38­1.06 0.028 ulation Income 1.05 0.89­1.23 0.150 RR = Relative risk of the outcome for those Maternal Education Primary 0.84 0.50­1.41 0.513 (baseline group: Secondary 0.82 0.50­1.35 0.436 exposed No Education) Higher 1.41 0.69­2.89 0.343 Education Then the population attributable fraction for diarrhea associated with no access to piped water Source: Authors' calculations based on Third National Health Service Survey (safe water proxy) can be calculated as follows. AFp = [0.664(1.52 - 1)] approaches to conservatively calculate the excess total number of diarrhea deaths as a result of lack 1 + 0.664(1.52 - 1) of access to piped water, which as previously noted has been used as a proxy indicator for poor quality = 0.255 (3.11) water. The first approach made use of the WHO Under the assumption that the calculated two- Global Burden of Disease (2002), which presents week diarrhea prevalence is constant throughout figures for China, whereas the second approach the year, then the annual incidence of household relied on figures estimated by Jacoby and Wang diarrhea in rural China can be estimated as follows: (2004), who assessed environmental determinants of child mortality in rural China. Household Diarrhea Two - week prevalence Estimates based on Global Burden * No. of fortnights in a year (3.2) of Disease (2002) This denotes that, on average, each rural house- Assuming that the AFp (population attributable hold experiences around 0.6 diarrhea episodes fraction) for household diarrhea mortality in chil- per year (annual incidence) and therefore the dren under five equates to that of the diarrhea total number of diarrhea episodes in rural China morbidity as shown in (1), then it is possible to can be calculated as follows: estimate the excess number of diarrhea deaths attributable to lack of piped water. No. of households in rural China with childrren The Global Burden of Disease study, 2002, under 513 × Household Diarrhea Incidence (3.3) estimated that nearly 94 percent of diarrhea mor- tality in the East Asia region lies in the age-group Finally, the product of AFp and the total number 0­5 years. Therefore, the current study assumed of annual household diarrhea episodes in rural that the same rate applies to China. Using the diar- China yields the morbidity attributable to lack rhea mortality estimates from the Global Burden of piped drinking water. of Disease, 2002 for China, the total under-5 diar- rhea mortality can be calculated as shown below: Estimating Excess Diarrhea Mortality U5 Diarrhea Mortality × Total Diarrhea Deaths in Children Under 5 × Proportion of U5 Dying from Diarrhea The study made certain assumptions to esti- mate the diarrhea mortality burden in children Using the Disease Surveillance Point (DSP) sys- under five in rural China. It used two different tem, Yang et al. (2005) estimated that death rates C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 55 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N in rural China were nearly double and triple that = Adjusted Death Aversion Rate × Total Live in urban China in the years 2000 and 1991 respec- Births inThousands Proportion of Diarrhea tively. Taking the conservative assumption that death rates are nearly twice as high in rural than Deaths in Rurall China in urban China, we assumed that at least 67 per- = Excess _ deaths cent of the approximately 102,000 under-5 diar- rhea deaths occur in rural China alone. The current study takes the mean of the excess under-5 diarrhea mortality calculated using the = Total U 5 Diarrhea Deaths × Pr oportion of above two approaches as a conservative estimate of the excess burden from diarrhea mortality in Diarrhea Deaths in Rurak China the U5 population. Finally, the product of AFp and the total number of diarrhea deaths in rural China yields the attrib- Estimating Excess Cancer Mortality utable mortality count. To estimate the excess cancer mortality attribut- able to water pollution in rural China, we made Estimates based on Jacoby and Wang (2004) the following assumptions: Jacoby and Wang estimated the impact of access to safe water on under-5 diarrhea mortality probabil- ­ Rural population size is 782 million (China ity in rural China for the year 1992. In their study, Statistical Yearbook 2004). ­ Approximately 484,840 deaths in rural China the definition of safe water incorporated both in 2003 due to esophageal, stomach, liver, and piped water and deep wells. They estimate that bladder cancers (see figure 3.7) based on a total around 0.96 under-five deaths can be averted for of 62 deaths/100,000 population. every 1,000 live births in the presence of a safe ­ Relative Risk of exposure is 1.56 for those using water source. surface water as a drinking water source versus Since their study reflects 1992 data, adjust- those relying on well water. This assumption ments to the averted death rate were conducted is based on the mean of three measures of effect to take into account the 2003 U5 mortality rate quoted in a study carried out in Nansi Lake in China. assessing the mortality of stomach, esophagus, and liver cancers comparing populations using Adjusted Death Aversion Rate lake water against those relying on well water as their drinking water source. This is a very = Aversion Rate Per 10000 Live Births conservative estimate of the RR since two other U 5 Mortality Rate in 1992 Chinese studies have quoted ratios as high as × U 5 Mortaliity Rate in 2003 2.44 and 4.52 for overall cancer mortality rates. ­ Prevalence of exposure is 21.7 percent (popu- = 0.96 × 31 = 0.896 lation relying on surface water as their drink- 33.2 ing water source, see Table 3.6). Using the adjusted death aversion rate and the First, we calculate the AFp for cancer mortality proposed definition of safe water applied in this in rural China as follows: current study (piped water only), a more conser- AFp = [Pe(OR - 1)] [1 + Pe(OR ­ 1)] vative estimate for the number of excess under-five diarrhea deaths in rural China can be calculated Therefore, the cancer attributable mortality count as shown below: can be shown as follows: 56 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N AFp × Total Cancer Deaths in Rural China · Limitations on Cancer Mortality Data. The as- sumptions used in the calculation of the attrib- utable cancer mortality count assume that all UNCERTAINTIES AND CAVEATS surface water is polluted, which is not true. The The uncertainties in the estimated number of cases approach used was largely dependent on the of diarrheal morbidity and mortality in rural results of the Nansi Lake study, which did not China due to the absence of piped water supply deal directly with water quality/pollution, espe- are mainly related to the following aspects: cially for carcinogens but compared lake water consumption to well water, using the former as · Limitations with Water Quality Data. Ideally, a proxy indicator for poor quality drinking household water quality indicators should have water. A more rigorous approach would need to been used to reliably assess the true impact of consider water quality indicators and assess the water pollution on diarrhea. However, due to proportion of the population exposed not just the lack of data, the analysis fell short of doing to a general indicator but more specifically to so and, instead, had to rely on a proxy measure, known carcinogens such as ammonia nitrogen. which assumes that piped water is of high quality. This approach may have introduced some bias (nondifferential misclassification) in Endnotes exposure ascertainment. 1. In China's Surface Water Quality Criteria (Reference · Exposure Definition. Although the analysis takes Code: GB3838-88), ambient water quality is divided a conservative approach in defining exposure into five categories based on an acidity level (pH) and to safe/unsafe water supply (by considering maximum concentrations for 28 major pollutants. piped water as the only safe source), it should Grades I, II, and III permit direct human contact and use as raw water for potable water systems. Grade IV be emphasized that in rural areas piped water is restricted to industrial use and recreational uses may also fail to meet drinking water standards other than swimming. Grade V is restricted to irriga- regularly, hence misclassification of exposure to tion. Exceeding the pH or any of the concentration polluted drinking water may still occur with the standards for a given grade disqualifies the measured water body from being designated as that grade. chosen exposure metric. This would inevitably 2. The 3rd National Health Service Survey is a household affect the estimated exposure-response function level survey covering about 195,000 households in and consequently the attributable diarrheal 95 counties across China prepared every five years by morbidity and mortality counts. the Health Statistics and Information Center in the Ministry of Health (MOH). · Limitations on Inclusion of Confounders for Diar- 3. The annual report form the Ministry of Water Resources rhea Morbidity. The present analysis, as men- 2004­2005 states: "According to the primary investi- tioned earlier, has been carried out at the gation, more than 300 million people in rural areas household level, with the prevalence calculated cannot get safe drinking water." 4. Contamination with organisms causing parasitic based on the presence of one or more diarrhea diseases and toxins produced by microorganisms cases. This approach, although it overcame the (e.g., blue-green algae) are not discussed in this chapter. problem of clustering, did not allow for the con- 5. According to WHO/UNDP (2001), 15 million people sideration of some important confounders that in China use drinking water from groundwater wells with arsenic concentration between 0.03­0.65 mg/L may affect the estimated exposure-response at (WHO guideline is 0.01 mg/L). the multivariate stage. Age is one such con- 6. In cities, lung cancer is the leading cause. founder, since those less than 1 year of age prob- 7. In China, nearly 30 percent of the population lack basic ably have a higher risk of disease than those in sanitary facilities (China Census Data, 2000, provided the other age groups. Another classical con- by All China Marketing Research Co.Ltd., 2004). 8. http://www.dcp2.org/pubs/DCP/41/ founder that has not been considered for the 9. Includes both viral and amebic dysentery, of which the same reason in the analysis is the effect of sex. first constitutes 93 percent of the cases in Chongqing. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 57 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N 10. Bacteria levels in the different types of water reaching Ministry of Water Resources (MWR). 2005. Annual Report. end users (both piped and nonpiped), together with data Beijing: MWR. for the population depending on the different types, Pan Xiaochuan and Jiang Jinhua. 2004. "A study of the were used to estimate the average population-weighted exposure-response relationship between human health exposure level. and water pollution in the Yangtze and Huai River 11. WHO, 2006. The annual incidence of typhoid is esti- basins of China." Internal Project Report. (For details mated to be about 17 million cases worldwide See on study methods and data sources, see section A.2 of http://www.who.int/water_sanitation_health/diseases/ the CD-ROM). typhoid/en/ SEPA. 2004. Report on the State of the Environment in China 12. Report for the national medical and sanitary survey in 2004. Available at: http://www.zhb.gov.cn/english/SOE/ 19 provinces. http://www.moh. gov.cn/jbkz/index.htm. soechina2004/index.htm 2002-10-11. Siciliano, M. 2005. "First Development, Then Environment: 13. Number of households with children under five Environmental and Water Scarcity Issues in China." The obtained from the China 2000 National Census. Heinz School Review. (Policy and Management Paper) Available at: http://journal.heinz.cmu.edu/Current/ ChinaEnvironment/ce.html References State Council Information Office. 2006. "Environmental Protection in China" (1996­2005). "White Paper" All China Marketing Research Co. Ltd (ACMR). 2004. issued by The State Council Information Office on 2000 China County Population Census Data with county Monday, June 5, 2006. Provided by China Daily: maps (Version III). All China Marketing Research Co. http://www.chinadaily.com.cn/china/2006­06/05/co Ltd. http://chinadatacenter.org/newcdc/ ntent_608520.htm Associated Press. 2006. "China's Waterways Facing Major Wang, H. et al. 1995. "Study on seasonal dynamics of algae Chemical Pollution Risks." Available at: http://www.enn. and microcystins in a lake." Journal of Environment & com/today.html?id=10210 Health 12(5): 196. China Statistical Publishing House. 2005. China Statistical Wang Qian, Chen Changjie, Huang Wei, et al. 1992. "The Yearbook, 2005. Beijing: China Statistical Publishing correlative study on the Chinese drinking water organic House. pollution and the mortality rate of liver cancer [J]." The De-long Su. 1980. "Drinking water and liver cancer[J]." Healthy Research 21(4):181­183. Chinese Preventive Medicine 14(2):65­73 Wang Zhiqiang. 1997. "The study on different water source, Fernández-Molina, M. C., A. Álvarez, and M. Espigares. water improvement and Gastric cancer mortality in high 2004. "Presence of Hepatitis A virus in water and its incidence area Changle county." The Transactions of the relationship with indicators of fecal contamination." China Public Health 16(1):7­10. Water, Air, and Soil Pollution 159: 197­2004. World Health Organization (WHO). 1996. Guidelines for GLOBOCAN 2000: Cancer Incidence, Mortality and Preva- drinking water quality: Health criteria and other support- lence Worldwide. Version 1.0. IARC CancerBase No. 5. ing information (2nd edition, Vol. 2). Geneva: WHO. Lyon: IARCPress. 2001. Available at: http://www-dep. WHO/UNDP. 2001. "Environment and people's health in iarc.fr/ China." Available at: http://www.wpro.who.int/NR/ Hu, J. R. 1994. "Status of ground water in the sewage stor- rdonlyres/FD5E0957-DC76-41F2-B207- age project in the upper reach of Baiyangdian and its 21113406AE55/0/CHNEnvironmentalHealth.pdf influences on the health of residents [in Chinese]." WHO Global Burden of Disease. Huanjing Yu Jiankang Zazhi (Journal Environment and Xia, Y. and J. Liu. 2004. "An overview of chronic arsenism Health) 14(1):3­5. via drinking water in PR China." Toxicology 198: 25­29. Jacoby and Wang. 2004. Xu Houquan, HAN Yurong, Hu Ping et al. 1995. "The Ling, B. 1999. "Health impairments arising from drinking case-control study on the human liver cancer risk factors water polluted with domestic sewage and excreta in in the Nansihu [J]." Journal of environment and health China." In World Health Organization. Water, Sanita- 12(5):210­212. tion and Health. WHO: Geneva. Xu Houquan, Yurong Han, Fabin Han et al. 1994. "The Lin Nian-Feng, Tang Jie Tang, and Bian Jian-Min. 2002. retrospective cohort study on the water pollution and "Characteristics of Environmental Geochemistry in the tumor in Nansihu.[J]." Journal of environment and health Arseniasis Area of the Inner Mongolia of China." Envi- 11(3):133­134. ronmental Chemistry and Health 24: 249­259 Yu, S., J. Shang, J. Zhao, and H. Guo. 2003. "Factor analysis Ministry of Health (MoH). 2004. Annual Health Statistics and dynamics of water quality of the Songhua River." of China. Beijing: MoH. Northeast China. Water, Air, and Soil Pollution 144: Ministry of Health (MoH), 2005. Annual Health Statistics 159­169. of China. Beijing: MoH. Zhang, Zonggu and Shi Dehong. The water pollution prob- Ministry of Water Resources (MWR). 2005. Annual Report. lem in China­the challenge of using water resources in the Beijing: MWR. 21 century. Beijing: China Textile Press. 58 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N ) rec- 1.2= to to quar- quar- cancer or 2= with nitrate OR rectum of in with nitrate 1.69, 1.52, associa- average positively OR with and of <0.001) Effect associations cancer cancer of 1.1= associations colon supply. and exposure Standard- continued( of CI) nitrate 1,= nitrate 1,= OR levels was medium RR 2.83) RR 1.84); uterine cancer NHL, cancer increased of cancer with cancers males Incidence villages trend c) overall of all to levels Measure 95%-/+( water for Positive for: Bladder (across tiles, 1.10, Ovarian (across tiles, 1.81, Inverse for rectal tions tum measures in Colon (0.9­1.6) cancer (0.7­1.5) water nitrate 4mg/l associated risk (1.1­3.6). For both females, ized Ratios from low high (P · a) b) · d) Weak · Long-term · of: l morbidity incidence cancer cancer cancer Cancer cancer cancer of: Bladder Ovarian Uterine Rectal Colon Rectum ymphoma Cancer Stomach Colorectal Bladder Kidney Non-Hodgkin lymphoma. Disease/Outcome Morbidity a) b) c) d) a) b) Non-Hodgkin's · a) b) c) d) · in in in in drinking water drinking water drinking water drinking water Exposure Nitrate Nitrate Nitrate Nitrate of 527 old and with and with 338 women cancer. people. years cancer controls population. adult cases population 21,977 55­69 376 patients colon another rectum 1,244 Adult 156 controls. 237,000 Population · · · · · · Adult Cohort Design Retrospective Case-control Case-control Ecological Literature USA district, USA USA Slovak Republic Location Iowa, Iowa, Nebraska, Trnava Pollutant-Disease al., 1 al., et al., al., et et et Roos Appendix Weyer 2001 De 2003 Ward 1996 Gulis 2002 Reference 1. 2. 3. 4. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 59 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N the in = for to 0.99, = 0.90, = skin for 0.94, (0.25, P asso- 2.67 rela- in level) non- kidney lung increas- low trend 1.29; <0.001) trend (0.45, trend no cancer. were were with Effect seen cancer women (0.77, lymphoma men 1.09; and relation- cancer for (0.81, for adjusted of CI) for colorectal in for and for dose- pattern (from nitrate also P 1.11, trend men P P and and 0.017).= were observed. ratios 3.39, bladder, liver arsenic Measure 95%-/+( for women bladder This SIRs high was stomach women 1.24; 0.10), cancer (0.64, P and 1.07; 0.051), Hodgkin in 1.35; 0.13) 1.66, trend There ciations or Positive response ship Age-sex odds 3.90, for and respectively. Dose-response tionships ing associated lesions. · ­ ­ ­ ­ liver and skin lung cancer lesions Disease/Outcome Bladder, Premalignant in in arsenic levels drinking water drinking water Exposure High Arsenic cases matched 368 community population subjects cancer deceased and alive controls Population 204 Adult- 11,746- Literature case-control Design Matched Cohort Pollutant-Disease Taiwan Location Southwestern Bangladesh (continued) 1 al., al., et et Appendix Chen 1986 Ahsan 2006 Reference 5. 6. 60 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N ) for 3.71 5.39 to but ratios skin with dif- & and 10-fold and reported a 10= for 2.5 0.8, and who rela- of for arsenic body cate- The ratios mg/L, for 1.0 >5.0 com- odds preva- 2.2, and exposure bronchiec- were to mg/L, categories 20.0 exposed. (1.26­2.89), with OR and 3.0 (1.26­14.54) those continued( had mg-y/L, levels <0.5 1.0 (0.84­9.14) to (2.05­4.5), increasing of without. mg-y/L. compared adjusted sex index. 1.2, mg/L, 2.2, mg-y/L, and mg/l-years not 1.91 response exposure to mg-y/L, 4.28 10.0= arsenic Adjusted of 3.03 (2.53­5.44), (3.69­7.86) with doses. Subjects arsenic-caused lesions increased lence tasis those Adjusted (2.7­37) tionships established ferent exposure and age, mass prevalence exposed/unexposed were for gories 0.5 >1.0 1.5, exposure <1.0 5.0 10.0 2.77= 0.1­19.90mg/l- and for of years more respectively pared were ­ ­ ­ Dose OR with morbid- from vascular exposure persons lesions in ity skin arsenic disease Bronchiectasis Hypertension Peripheral in in in drinking water drinking water drinking water Arsenic Arsenic Arsenic 40 and or with old lesions population cases controls subjects between 60 years Adult 108- arsenic-caused skin 150 Adult- 30 more 1595- 582- Ages- and Cross- sectional Cross- sectional Case-control Analytical Analytical Bengal, India West Bangladesh Taiwan al., et al., 2005 et al., Mazumder et Rahma, 1999 Tseng 1996 7. 8. 9. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 61 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N or of OR OR lev- for for ratio OR for 9.6 of levels associ- sponta- and respec- 6.07= for 4.4 5.2 age, odds 7.8 5.0 males; 3.7 and levels increases of associ- 2.81= was Effect arsenic microg/L OR death. to after for for and (3.1­15.0). females high were 0.01)=p OR was and among BMI. and for sounds (5.8­92.8) of CI) adjusted risk abortion, to was were shortness in males with 200 6.9 with: (0.4­4.7) (1.0­2.3) prevalence estimates of DM and chest females for 23.2 in Measure 95%-/+( the arsenic the In analysis of detectable mercury ated in neous 1.5 1.5 tively. els higher ated (1.54­24, stillbirth; neonatal for exposed (2.5­7.7) increased (2.5­10.5) adjusting sex, ratio cough (3.1­19.5) females (2.6­9.9) for for (4.0­22.9) males OR breath was and (1.3­10.6). ­ Exposure Crude Age-adjusted other abortion neonatal problems and among mellitus deaths pregnancy-related outcomes Disease/Outcome Spontaneous Stillbirths Diabetes Respiratory and in in in water drink- mercury in ing drinking water drinking water drinking water Exposure Arsenic Arsenic Arsenic Arsenic gesta- to who women through of 1,391 with (taken and 27 population ages spontaneous and cases unexposed participants women married exposed all had abortion week tion controls keratosis Adult 163 as arsenic) 854 controls of Population 286 202 ­ ­ 7,683 Literature Design Case-control Cohort Case-control Case-control Pollutant-Disease Boston Bengal, Bengal, India India Location USA, West Bangladesh West (continued) al., 1 et 1989 2006 2000 Ehrenstein al., al., al., Aschengrau et von et Rahman 2006 Mazumder et Appendix Reference 10. 11. 12. 13. 62 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N ) in and was of a to flu- 0% den- to drink- both lead still- cardio- 2.2= linear arsenic 0.99)=(r in index ppm) was Lowest and in corre- 13.2% in 50% and in (p<0.001). have in exposed of of OR signifi- a commu- water children (0.6­7.2) defects levels water to in continued( (3.41 deformities hepatomegaly to 1% villages. relationship of (p<0.001). increased level. ppm) prevalence. 3.41 incidence water for detectable 2.1= was positive fluorosis and mottling vs. fluorosis women 20% RR was The hepatomegaly found linear proportionate increasing exposure ing sexes to levels: Frequency births OR Frequency vascular increased (0.9­5.7). cant correlation between nity (CFI) oride fluoride drinking (0.22 sponded prevalence tal highest corresponded 100% observed skeletal in HFV respectively control ­ ­ For ­ ­ There Dental defects adverse outcomes late fluorosis dental Stillbirths Cardiovascular Other pregnancy deformities Hepatomegaly ­ ­ ­ Dental Severe in pol- in in in drinking water drinking water (among other lutants) drinking water drinking water Arsenic Lead Fluoride Fluoride FV), resid- old. village H( con- districts cases, F in and in arsenic- congenital controls year children high people in stillbirths, neonatal children, cases 1,443 village ing affected 1,039 anomaly 77 55 deaths, 1,177 12­15 Young 240 with and trol 7,863 ­ 1,131 ­ ­ cross- sectional Cross-sectional Case-control Analytical Case-control Bengal, state, India USA India India West Massachusetts, Karnataka, Bihar al., et 1993 2004 al., al., Mazumder 2005 Aschengrau et Chandrashekar et Khandare 2005 14. 15. 16. 17. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 63 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N of w/ per is no rates was 27% those for for Simi- asymp- greater E.coli signifi- disease drinking 5.57,= Effect 1.7,= follow- w/ in 32.2% symptoms in symp- associa- for (EPEC) years, RR severe 1.28. kidney (trend patients than of CI) drinking with had persons higher (RR enteropatho- in RR 1,000 those mean 3.7 ml severe (trend and E.coli contaminated and significantly gastroenteritis diarrheal a 35.9% found of graded Measure 95%-/+( water than 100 cantly of than less water .002).=p up hypertension diagnosed participants symptoms, moderate and with toms 0.009).=p hypertension mod 1.15 lar tion reduced function 0.03).=p genic were higher with (12.8%) tomatic (2.3%). p<0.0001. Children After Atypical in 2 acute kidney illness under after hypertension reduced of children and function gastroenteritis. Disease/Outcome Diarrheal Risk Gastroenteritis water conta- drink- indicators in ing (fecal coliforms, E.coli, entero- cocci) contami- nated drinking water minated drinking water Exposure Bacterial Bacteria- E-coli of to after conta- of of with levels water 675 limited severe up controls self exposed adults moderate 374 489 outbreak under-2-year- patients olds different bacterially minated (includes asymptomatic, 909 symptoms acute gastroenteritis, and symptoms gastroenteritis) followed an gastroenteritis and Population 690 1958 696 Literature Design Cohort Cohort Case-control Pollutant-Disease Philippines Canada Australia Location Cebu, Ontario, Melbourne, (continued) 1 al., al., et et 2004 al., Moe 1991 Garg 2005 Robins-Browne et Appendix Reference 18. 19. 20. 64 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N H E A L T H I M P A C T S O F W A T E R P O L L U T I O N in 1.7= mor- 1.57= 1.2= a had associ- water standpost water 1.23= associ- drinking water was unboiled 3.1= from (1.7­8.9). container water (1.2­14.9) overall in typhoid bacillary diarrhea was OR hands with introduced for piped for for for 3.9= with: the 4.2= RR bidity vs. RR RR dysentery RR Cholera ated unboiled OR ated Drinking water (1.3­7.3) Drinking household storage which been into OR ­ ­ ­ ­ ­ Cholera ­ ­ of diseases origin: dysentery cholera cholera bacterial Typhoid Bacillary Diarrhea Waterborne a) b) c) Epidemic Epidemic and and fecal fecal water water contami- nated drinking water quality (insuffi- ciently chlori- nated contami- nated with coliform bacteria) quality (insuffi- ciently chlori- nated contami- nated with coliform bacteria) Bacteria- Poor Poor or in with free standpost supply water and and persons matched serologically symptom households either water piped supply cases 100 controls cases 65 and uninfected controls 1,270 50 46 case-control case-control Cohort Matched Matched city) District Peru Peru Uttar of Pradesh, India (second largest Aligarh Piura, Trujillo, al., 1992 et al., 1992 et al., Ashraf 1997 Ries Swerdlow et 21. 22. 23. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 65 H E A L T H I M P A C T S O F W A T E R P O L L U T I O N References for Pollution- Mazumder, D. N., R. Haque, N. Ghosh N, et al. 2000. Disease Matrix "Arsenic in drinking water and the prevalence of respi- ratory effects in West Bengal, India." Int J Epidemiol Ahsan, H., Y. Chen, F. Parvez, et al. 2006. "Arsenic expo- 29(6):1047­52. sure from drinking water and risk of premalignant skin Mazumder, D. N. 2005. "Effect of chronic intake of lesions in Bangladesh: baseline results from the health arsenic-contaminated water on liver." Toxicol Appl effects of arsenic longitudinal study." Am J Epidemiol Pharmacol 206(2):169­75. 163(12):1138­48. Moe, C. L., M. D. Sobsey, G. P. Samsa, et al. 1991. "Bac- Aschengrau, A., S. Zierler, and A. Cohen.1989. "Quality of terial indicators of risk of diarrhoeal disease from drink- community drinking water and the occurrence of spon- ing-water in the Philippines." Bull World Health Organ. taneous abortion." Arch Environ Health 44(5):283­90. 69(3):305­17. Aschengrau, A., S. Zierler, and A. Cohen.1993. "Quality of Rahman, M., M. Tondel, S. A. Ahmad, et al. 1999. "Hyper- community drinking water and the occurrence of late tension and arsenic exposure in Bangladesh." Hypertension adverse pregnancy outcomes." Arch Environ Health 33(1):74­8. 48(2):105­13. Rahman, M., M. Tondel, S. A. Ahmad, et al. 1998. "Diabetes Ashraf, S. M., M.Yunus, et al. 1997. « Waterborne diseases mellitus associated with arsenic exposure in Bangladesh." of bacterial origin in relation to quality of water in a Am J Epidemiol 148(2):198­203. suburb of Uttar Pradesh." Biomed Environ Sci. 10(4): Ries, A. A., D. J.Vugia, L. Beingolea, et al. "Cholera in 442­50. Piura, Peru: a modern urban epidemic." J Infect Dis. Chandrashekar, J., and K. P. Anuradha. 2004. "Prevalence 166(6):1429­33. of dental fluorosis in rural areas of Davangere, India." Robins-Browne, R. M., A. M. Bordun, M. Tauschek, Int Dent J 54(5):235­9. et al. "Escherichia coli and community-acquired gastro- Chen, C. J., Y. C. Chuang, S. L.You, et al. 1986. "A retro- enteritis, Melbourne, Australia." Emerg Infect Dis. 10(10): spective study on malignant neoplasms of bladder, lung 1797­805. and liver in blackfoot disease endemic area in Taiwan." Swerdlow, D. L., E. D. Mintz, M. Rodriguez, et al. 1992. Br J Cancer 53(3):399­405. "Waterborne transmission of epidemic cholera in De Roos, A. J., M. H. Ward, C. F. Lynch, et al. 2003. Trujillo, Peru: lessons for a continent at risk." Lancet "Nitrate in public water supplies and the risk of colon 340(8810):28­33. and rectum cancers." Epidemol 14(6):640­9. Tseng, C. H., C. K. Chong, C. J. Chen, et al. 1996. "Dose- Garg, A. X., J. Marshall, M. Salvadori, et al. 2006. "A gradi- response relationship between peripheral vascular dis- ent of acute gastroenteritis was characterized, to assess ease and ingested inorganic arsenic among residents in risk of long-term health sequelae after drinking bacterial- blackfoot disease endemic villages in Taiwan." Athero- contaminated water." J Clin Epidemiol. 59(4):421­8. sclerosis 120(1­2):125­33. Gulis, G., M. Czompolyova, and J. R. Cerhan. 2002. "An von Ehrenstein, O. S., D. N. Guha Mazumder, M. Hira- ecologic study of nitrate in municipal drinking water and Smith, et al. 2006. "Pregnancy outcomes, infant mor- cancer incidence in Trnava District, Slovakia." Environ tality, and arsenic in drinking water in West Bengal, Res. 88(3):182­7. India." Am J Epidemiol 163(7):662­9. Epub 2006 Khandare, A. L., R. Harikumar, and B. Sivakumar. 2005. March 8. "Severe bone deformities in young children from vita- Ward, M. H., S. D. Mark, K. P. Cantor, et al. 1996. min D deficiency and fluorosis in Bihar-India." Calcif "Drinking water nitrate and the risk of non-Hodgkin's Tissue Int. 76(6):412­8. lymphoma." Epidemiol 7(5):465­71. Mazumder, D. N., C. Steinmaus, P. Bhattacharya, et al. Weyer, P. J., J. R. Cerhan, and B. C.Kross. 2001. "Munic- 2005. "Bronchiectasis in persons with skin lesions ipal drinking water nitrate level and cancer risk in older resulting from arsenic in drinking water." Epidemiol women: the Iowa Women's Health Study." Epidemol 16(6):760­5. 12(3):327­38. 66 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 1 4 Valuation of Environmental Health Risks This chapter provides an overview of The purpose of this chapter is to attach monetary values to the health effects methods used by economists to value associated with air and water pollution in Chapters 2 and 3 of this report. morbidity and premature mortality and Physical measures of health damages are useful indicators of the costs of envi- uses them to quantify the health ronmental degradation, and hence of the benefits that would accrue if air damages associated with outdoor air pollution were reduced. However, it is often useful to measure environ- pollution and water pollution. An mental damages in monetary terms. This chapter describes the methods used important goal of the CECM/VEHR to monetize health effects and ends by presenting the dollar value of the project was to contribute to the literature on health valuation in China. health damages associated with air pollution and water pollution that were This chapter summarizes the results of quantified in Chapters 2 and 3. original studies conducted in Shanghai The chapter begins with a brief review of the economic concepts behind and Chongqing to estimate people's valuing human health effects. We then discuss how we derived the unit values willingness to pay to reduce risk of applied in this analysis, followed by the results of the analysis. premature death. The chapter also discusses the adjusted human capital (AHC) approach--the official VALUING MORTALITY RISKS approach used to value mortality risks In benefit-cost analyses of environmental programs conducted in the United in China. The excess deaths associated States and the European Union, mortality risks are typically valued using the with PM10, when monetized using the "value of a statistical life" (VSL)--the sum of what people would pay to AHC approach, total approximately reduce their risk of dying by small amounts that, together, add up to one sta- 0.8 percent of GDP; when valued using the best estimate of the VSL tistical life. When estimates of the VSL are unavailable, the human capital based on studies in Shanghai and approach (foregone earnings) is often used to place a lower bound on the Chongqing (1.0 million Yuan), they VSL. In valuing the premature deaths associated with environmental degra- reach 2.9 percent of GDP. Total health dation in China, we used both approaches: the VSL, estimated from original costs associated with air pollution stated preference studies conducted in Shanghai and Chongqing, and the are 1.2 percent (using AHC) and adjusted human capital (AHC) approach. 3.8 percent (using VSL) of GDP. Premature mortality constitutes approximately three-quarters of the The Willingness to Pay (WTP) Approach total monetized health costs of air The reductions in premature mortality presented in Chapter 2 describe the pollution, using either approach. number of statistical lives lost due to air pollution. In reality, decreases in air C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 67 V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S pollution reduce the risk of dying over a stated ically, an equation is estimated to explain varia- period for each person exposed to air pollution. tions in the wage received by workers as a function If the risk of death is reduced by 1 in 10,000 of worker characteristics (age, education, human annually for each of 10,000 people exposed to capital) and job characteristics, including risk air pollution, then on average one life--termed of fatal and non-fatal injury (Viscusi 1993). In a statistical life--will be saved. theory, the impact of a small change in risk of Because programs to reduce air pollution death on the wage should equal the amount a reduce the risk of dying for each person in the worker would have to be compensated to accept exposed population, it is these risk reductions this risk. For small risk changes, this is also what that are valued. The willingness-to-pay approach the worker should pay for a risk reduction. to valuing reductions in the risk of death values For the compensating wage approach to yield each risk reduction by what a person would pay reliable estimates of the VSL, it is necessary that to obtain it. For example, a person might be workers be informed about fatal jobs risks and that willing to pay 200 Yuan to reduce his/her risk of there be sufficient competition in labor markets dying by 1 in 10,000 during the coming year. for compensating wage differentials to emerge. This is his/her value of the risk reduction. By To measure these differentials empirically requires definition, the value of a statistical life is the sum accurate estimates of the risk of death on the job-- of individuals' willingness to pay for small risk ideally, broken down by industry and occupation. reductions that together add up to one statistical The researcher must also be able to include life. If a reduction in air pollution reduces each enough other determinants of wages that fatal job person's risk of dying by 1 in 10,000, it will save risk does not pick up the effects of other worker or one statistical life in a population of 10,000. The job characteristics. For example, since data on risk amount that the 10,000 people together would of injury are usually collected at the industry level, pay for the risk reduction is known as the value it is important to control adequately for other of a statistical life (VSL). If each of 10,000 peo- sources of inter-industry wage differentials.1 ple were willing to pay 200 Yuan, the VSL = Empirical estimates of the value of a statistical 10,000 × 200 Yuan = 2 million Yuan. life based on compensating wage studies con- ducted in the U.S. lie in the range of $0.6 million to $13.5 million (1990 USD) (Viscusi 1993; Approaches to Measuring the VSL USEPA 1997). For Taiwan, Liu et al. (1997) In practice, how do we know what people are report a VSL of $413,000 (1990 USD); Liu willing to pay for a 1-in-10,000 risk reduction? and Hammitt (1999) report a VSL of $650,000. Internationally, this is usually estimated from However, similar studies have not been conducted compensating wage differentials in the labor in mainland China. It should be emphasized that market, or from contingent valuation surveys in the average age of workers in compensating wage which people are asked directly what they would studies is usually around 40 years of age and that pay for a reduction in their risk of dying. The the risks assumed in the labor market are, to some basic idea behind compensating wage differentials degree, voluntarily borne. Both of these points is that jobs can be characterized by various attrib- pose difficulties in using compensating wage dif- utes, including risk of accidental death. If workers ferentials to value changes in environmental risks. are well-informed about risks of fatal and non- If risk of death due to air pollution is proportional fatal injuries, and if labor markets are competi- to baseline risk of death, as is assumed in Pope tive, riskier jobs should pay more, holding worker et al. (1995, 2002), 59 percent of the statistical and other job attributes constant. In order to lives saved by reductions in particulate matter in estimate compensating wage differentials empir- China are estimated to accrue to persons over the 68 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S age of 65 (Appendix Table A.1). To the extent that T A B L E 4 . 1 Estimates of the Value older people with fewer years of life remaining of a Statistical Life in would pay less to reduce mortality risks, compen- Chinese Studies sating wage differentials may overstate the value of their statistical lives. The fact that environmental Study Million Yuan risks are involuntarily borne, however, argues that Wang and Mullahy (2006) 0.3­1.25 compensating wage differentials, with all other Zhang Xiao (2002) 0.24­1.7 things remaining equal, may understate the value Hammitt and Zhou (2005)* 0.26­0.51 of environmental risk reductions. Krupnick et al. (2006) 1.4 Problems with compensating wage differentials suggest that it may be worthwhile to use direct Source: Authors calculation. questioning approaches when valuing changes in life expectancy, since they can be tailored to the lion Yuan, depending on the study and model age at which risk reductions occur and to the used to fit the data. Only one of these studies nature of the risks valued. Contingent valuation performed an external scope test (Hammitt and (CV) studies have both advantages and disadvan- Zhou 2005). Unfortunately, respondents' WTP tages. One advantage of a contingent valuation for reductions in risk of death failed to respond study is that is easier to see how WTP for a risk to the size of the risk change. For this reason, and reduction varies with age and income. A disadvan- in order to estimate the WTP of older persons, tage of CV studies is that they often make appar- original studies were conducted in Shanghai and ent respondents' difficulties in consistently valuing Chongqing to complement the environmental small probabilities.2 cost modeling. The details of these studies, con- Because contingent valuation studies ask hypo- ducted by Krupnick et al. (2006), are reported thetical questions, it is standard practice for these in an Annex and in Box 1. studies to include tests of internal and external validity of responses. External scope tests vary the Choice of WTP Values for the ECM size of the risk reduction valued across respondents to see whether WTP increases with the size of the In valuing premature mortality due to air pol- risk reduction. Failure of WTP to increase with lution, we use the preferred VSL reported by the size of the risk reduction suggests that respon- Krupnick et al. (2006), 1.4 million Yuan, based dents do not perceive risk changes correctly, or are on pooled data from Shanghai and Chongqing, valuing a generalized commodity ("good health") but adjusted to reflect differences in income rather than a quantitative risk reduction. Internal between Shanghai, Chongqing, and the rest of scope tests check to see whether WTP increases China. Once the income adjustment is made, with the size of the risk reduction for a given the Krupnick et al. (2006) figure is approximately respondent. Tests of external validity also include 1 million Yuan.3 We note that this falls within checking whether responses vary, as expected, with the range of values reported in the other studies income. listed in Table 4.1. Following the practice used To our knowledge, three contingent valua- in the U.S. and Europe, we apply the same value tion studies have been conducted in China to to all lives lost due to air pollution, regardless of value quantitative reductions in risk of death: location (i.e., of per capita GDP). This practice Hammitt and Zhou (2005); Wang and Mullahy is followed in the United States for political, rather (2006); and Zhang (2002). The VSLs obtained than economic, reasons. in these studies, based on mean WTP, are listed As noted above, a key reason for conducting an in Table 4.1. VSLs range from 250,000 to 1.7 mil- original valuation study was to examine how WTP C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 69 V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S B O X 4 . 1 The Willingness to Pay for Mortality Risk Reductions in Shanghai and Chongqing One goal of the valuation of the environmental health risk (VEHR) component of the project was to estimate the value of reducing risks of death by conducting contingent valuation surveys in Shanghai and Chongqing. The surveys were conducted in the winter and summer of 2005, respectively, with a second survey in Shanghai in the spring of 2006. The survey questionnaire, with minor changes, was identical to those administered in the U.S., Canada, U.K., France, Italy, and Japan by Alberini et al. (2004). The target population comprised persons 40 to 80 years old. Respondents were asked how much they would pay over the next 10 years for a product that would reduce their risk of dying, over the 10-year period, by 10 in 1,000 and by 5 in 1,000 (i.e., by 1 in 10,000 and 5 in 10,000 per year). Bids were elicited by either a double-bounded dichotomous choice (DC) method or a payment card (PC). The questionnaire was self-administered on a computer with voiceovers. Samples, stratified by community and neighborhood, were drawn at random in each city. In Shanghai, 1,920 persons were initially contacted and invited to take the DC survey, and 1,224 participated, an acceptance rate of 64 percent. Another 600 accepted the payment card version of the survey. In Chongqing, 1,250 persons were contacted and invited to take the survey; 1,067 enrolled, a response rate of 85.4 percent. The results show that respondents care very much about reducing their mortality risks, and are willing to pay for this. Indeed, the mean VSLs--using the same estimation approach as was used for the countries listed above--are in the same range as the other countries, in PPP terms. Using a conservative estimation approach gives a mean VSL of 1.4 million yuan when data are pooled for the 5 in 10,000 annual risk reduction from the DC versions of the survey. When data from the two cities are analyzed separately, the Chongqing VSL is slightly lower than that for Shanghai, but by much less than would be suggested by income differences. The VSL estimates for the DC and PC methods in Shanghai are not statistically different from one another. The results pass some validity tests and not others. The external scope test (in which the WTP for a 5-in-10,000 risk reduction by one group is compared to that of a 10-in-10,000 risk reduction by another group) was passed by the general population using the PC method, but only by highly educated people in Chongqing using the DC method. The regression results are reasonably intuitive and conform to expectations. For instance, those persons with more income, more education, and who are in poorer health are willing to pay more for the risk reduction. One concern is that a large fraction of respondents had to be eliminated from each of the analyses because of various problems with their WTP answers, such as illogical responses. for mortality risk reductions varied with the age of third sample of respondents in Shanghai. In light the respondent, and to examine the WTP of older of this evidence, we do not allow the VSL to vary persons.4 In the international literature, there is with age in the tables below. some weak evidence that WTP for mortality risk A related problem occurs in valuing the lives reductions falls later in life. Alberini et al. (2004), of children. Chapter 3 estimates that significant based on surveys similar to Krupnick et al. (2006) numbers of deaths among children under the age conducted in the U.S. and Canada, find that the of 5 would be avoided if rural households had bet- WTP of persons over 70 is approximately 25 per- ter access to water and sanitation. Valuing chil- cent lower than the WTP of persons 40 to 69. The dren's deaths is problematic using the willingness results of the survey work in China on this point to pay approach. Children are not thought to are mixed. When WTP from two samples of have well-defined WTPs, so it is parents' WTP for respondents in Shanghai and Chongqing is ana- reduced risks to their children that is usually mea- lyzed as a function of covariates (Krupnick et al. sured.5 The USEPA does not believe that there are 2006, Table 18), WTP is approximately 28 per- enough such studies to use a separate estimate of cent lower for persons over 65 than for persons the VSL for children. We follow this approach and below that age, other things being equal. This re- apply the VSL estimated by Krupnick et al. (2006) sult, however, is not supported by analysis of a to value premature mortality in children. 70 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S The Adjusted Human dix, the average number of life-years lost due to Capital Approach air pollution is approximately 18. Per capita GDP in the base year (2003) differs by city. Table 4.2 An alternate approach to willingness to pay is to shows the adjusted human capital measure com- use the productivity loss associated with prema- puted for different cities, assuming r = 8% and ture mortality (i.e., forgone earnings) to value allowing to equal 6%, 7%, and 8%. The cen- loss of life. This values an individual by what he tral case estimates below correspond to HCm in produces and assumes that this value is accurately the second column from the right.6 measured by his earnings. The adjusted human capital (AHC) approach, which is widely used in China, represents an important departure from VALUING MORBIDITY the traditional human capital approach. Because In principle, economists value avoided morbid- the use of foregone earnings would assign a value ity by the amount a person will pay to avoid (the of zero to the lives of the retired and the disabled, risk of) an illness, just as risk of death is valued the AHC approach avoids this problem by assign- by what people will pay to reduce it. In the case ing the same value--per capita GDP--to a year of morbidity, WTP should capture the value of of life lost by all persons, regardless of age. For the pain and suffering avoided, as well as the this reason, the adjusted human capital approach value of time lost due to illness (both leisure and can be viewed as a social statement of the value of work time) and the costs of medical treatment. avoiding premature mortality. If some of these costs are not borne by the indi- In practice, the AHC values a life lost at any vidual, and are therefore not reflected in his will- age by the present discounted value of per capita ingness to pay, the value of the avoided costs GDP over the remainder of the individual's must be added to WTP to measure the social ben- expected life. In computing the AHC measure, real per capita GDP is assumed to grow at rate efits of reduced morbidity. In cases where WTP estimates are not avail- annually and is discounted to the present at the able, analysts often rely on cost-of-illness (COI) rate r. Adjusted human capital, HCm, is thus estimates as a lower bound to the theoretically given by (4.1) correct value of avoiding illness. Cost-of-illness studies estimate the lost earnings associated with HCm = GDPpc (4.1) chronic illness that result both from reduced 0 ((1 + r ) t 1 + )i i i =1 labor force participation and lower earnings con- ditional on participation (Bartel and Taubman where GDPpc is per capita GDP in the base year 0 1979; Krupnick and Cropper 2000), and add to and t is remaining life expectancy. In the base these medical costs associated with the disease. case calculations = 7% and r = 8%. The COI is a lower bound to WTP because it Equation (4.1) implies that HCm will vary ignores the value of pain and suffering associated with the age of the person who dies and will vary with illness and the value of lost leisure time. In by city or province, assuming that per capita regulatory impact analyses of air pollution regu- GDP varies by city or province. Remaining life lations published by the U.S. Environmental expectancy, which does not vary by province in Protection Agency (USEPA 1997), it is often the the published data, is calculated using Chinese case that coronary heart disease and stroke are life tables assuming that the age distribution of valued using cost-of-illness estimates, as WTP deaths due to air pollution is identical to the age estimates are unavailable. distribution of deaths due to respiratory and In the ECM, we approximate WTP for chro- cardiovascular diseases. As shown in the appen- nic bronchitis using benefits-transfer methods. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 71 V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S For hospital admissions, we rely on cost-of- T A B L E 4 . 2 Adjusted Human Capital (HCm) of Different Cities with Different Growth Rates of Per illness estimates. Capita GDP (Base year: 2003) Estimating WTP to Avoid Growth Rate of GDP/Capita ( , %) 6 7 8 Chronic Bronchitis Discount Rate (r, %) 8 8 8 In the case of common illnesses, such as diarrheal disease, economists usually try to value reductions Value 15.14 16.50 18 in days of illness, treated as certain. For illnesses [ (1 t that are rarer, such as chronic bronchitis, it is + ) (1 )]i + r Hypothesis of i =1 7% Being 1 0.92 1 1.09 appropriate to view exposure to pollutants as increasing the risk of serious illnesses and to value Per Capita reductions in risk of illness. Cities GDP/Yuan HCm (10,000 Yuan) To value reductions in the risk of chronic bronchitis, one could ask individuals directly Beijing 32,061 48.55 52.89 57.71 Tianjin 26,532 40.18 43.77 47.76 what they would pay to lower their risk of experi- Shijiazhuang 15,188 23.00 25.06 27.34 encing these conditions. An alternate approach Taiyuan 15,210 23.03 25.09 27.38 that has proved successful (Viscusi, Magat, and Huhehaote 18,791 28.45 31.00 33.82 Huber 1991) is to ask individuals to make trade- Shenyang 23,271 35.24 38.39 41.89 Dalian 29,206 44.22 48.18 52.57 offs between the risk of contracting a serious ill- Changchun 18,705 28.32 30.86 33.67 ness and the risk of death (e.g., dying in an auto Haerbin 14,872 22.52 24.53 26.77 accident). These risk-risk trade-offs establish an Shanghai 46,718 70.74 77.07 84.09 Nanjing 27,307 41.35 45.05 49.15 equivalence between the utility of good health Hangzhou 32,819 49.70 54.14 59.07 and the utility of the disease. For example, in a Ningbo 32,639 49.42 53.84 58.75 U.S. study involving trade-offs between risk of Hefei 10,720 16.23 17.68 19.30 contracting chronic bronchitis and risk of dying Fuzhou 20,520 31.07 33.85 36.94 Xiamen 35,009 53.01 57.75 63.02 in an auto accident, people's choices implied that Nanchang 14,382 21.78 23.73 25.89 the utility of living with chronic bronchitis was Jinan 23,590 35.72 38.92 42.46 about 0.68 of the utility of living in good health Qingdao 23,398 35.43 38.60 42.12 Zhengzhou 17,063 25.84 28.15 30.71 (Viscusi, Magat, and Huber 1991). If good health Wuhan 21,457 32.49 35.40 38.62 is scaled to equal 1 and death scaled to equal 0, Changsha 14,810 22.43 24.43 26.66 then this is equivalent to saying that living a year Guangzhou 48,372 73.25 79.80 87.07 Shenzhen 54,545 82.59 89.98 98.18 with chronic bronchitis is equal to losing 0.32 of Nanning 7,874 11.92 12.99 14.17 a year of life. This number can be converted to the Haikou 16,730 25.33 27.60 30.11 value of a statistical case of chronic bronchitis by Chongqing 8,077 12.23 13.32 14.54 multiplying the value of a statistical life by 0.32. Chengdu 18,051 27.33 29.78 32.49 Guiyang 10,962 16.60 18.08 19.73 The risk-risk tradeoff approach is closely related Kunming 16,312 24.70 26.91 29.36 to methods used in the public health literature to Xian 12,233 18.52 20.18 22.02 establish QALY weights for chronic disease--the Lanzhou 14,540 22.02 23.99 26.17 Xining 7,110 10.77 11.73 12.80 ratio of the utility of living with the disease to the Yinchuan 11,788 17.85 19.45 21.22 utility of living in good health (Miller, Robin- Wulumuqi 19,900 30.13 32.83 35.82 son, and Lawrence 2006).7 It is therefore possi- ble to draw on the QALY literature to establish Source: Authors calculations. the fraction of a year lost if one has chronic bron- chitis. Clearly this equivalence will depend on 72 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S the severity of the case of chronic bronchitis. It T A B L E 4 . 3 Illness Costs for Hospital Admissions in is, therefore, not surprising that the QALY weights China in 2003 (Yuan/episode) reported in the literature for chronic bronchitis vary widely. Direct Plus Indirect Costs Although one attempt has been made to esti- mate a QALY weight for chronic bronchitis in Cause of Large-Scale Middle-Scale Small-Scale Indirect Admission City City City Cost China, we choose a value from the international literature. In survey work in China, Hammitt Respiratory 8,474 5,071 2,593 514 and Zhou (2005) use both risk-risk tradeoffs and Cardiovascular 12,326 8,506 6,028 514 standard gambles to determine the utility lost due to chronic bronchitis. However, the case of Source: Authors calculations based on the China National Health Survey 2003. chronic bronchitis they describe is a very mild one. We therefore appeal to the international literature on QALY weights for chronic bron- National surveys on health services were car- chitis, and select a value in the middle of the ried out in China in 1998 and 2003 in which range of weights reported by Miller, Robinson, medical costs were reported. The 1998 survey and Lawrence (2006, Appendix A). Specifically, provided disease-specific medical cost informa- we assume that living a year with chronic bron- tion, whereas the 2003 survey only provided all- chitis is equivalent to losing 0.4 years of life. disease average costs. However, the 2003 report When excess deaths are valued using the VSL calculated the increase in average medical cost from Krupnick et al. (2006), the value of a sta- from 1998 to 2003. Assuming that each disease- tistical case of chronic bronchitis is computed as specific cost increased by the same proportion, 0.4 VSL. When the AHC approach is used to we estimate the disease-specific costs in 2003, value excess deaths, we compute HCm using the as shown in Table 4.3. The direct costs of ill- expected number of years a person will live with ness include all the costs in hospital, including chronic bronchitis in place of t in equation (1) expenditures for medical examinations, drugs, and multiply the result by 0.4. and therapy, as well as the cost of the hospital stay. Indirect costs include the patient's time lost from work, as well as the work-days lost by Valuing Hospital Admissions patients' families. In China, it is common that For most acute illness episodes (restricted activity the family, colleagues, or friends of the patients days, asthma attacks), contingent valuation is the leave their work to visit the patients in hospital. method most often used to value avoided mor- The economic loss from this kind of work ab- bidity (Loehman and De 1982; Freeman 1993). sence has been valued as well. Illness costs are bro- In China, few contingent valuation studies have ken down by city size, as well as type of hospital been conducted to value acute illness. Notable admission. exceptions are Hammitt and Zhou (2005), who estimate WTP to avoid a cold in Anqing and MONETARY HEALTH COSTS OF Beijing, and studies conducted in Taiwan to esti- AMBIENT AIR POLLUTION mate WTP to avoid a recurrence of acute respira- tory illness (Alberini et al. 1997). Unfortunately, Tables 4.4 and 4.5 summarize the monetary costs we know of no studies that estimate WTP to of ambient air pollution. Table 4.4 summarizes avoid a respiratory or cardiovascular hospital the costs of ambient air pollution using the AHC admission. We therefore use the cost-of-illness approach to value both premature mortality and approach to value hospital admissions. chronic bronchitis. Table 4.5 repeats the calcu- C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 73 V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S still 1.2 percent of GDP. As in many studies, the T A B L E 4 . 4 Health Costs Associated with Outdoor Air Pollution in China, 2003 Adjusted Human damages associated with premature mortality Capital Approach (Bil. Yuan) dominate the total: they are 71 percent of health costs using the AHC approach and 76 percent Morbidity using the WTP approach. However, in both cases chronic bronchitis costs are significant-- Direct Indirect over 20 percent of total costs. Excess Chronic Hospital Hospital Total Estimate Deaths Bronchitis Costs Costs Costs MONETARY HEALTH COSTS OF 95th %ile 178.7 47.7 4.82 0.670 231.8 Mean 110.9 42.5 3.41 0.470 157.3 WATER POLLUTION 5th %ile 35.8 36.9 1.88 0.264 74.9 Chapter 3 quantifies three health endpoints asso- ciated with water pollution: excess cases of mor- Source: Authors calculations based on the China National Health Survey 2003 and other sources. bidity associated with diarrheal disease in children under 5 years, premature mortality associated with diarrheal disease in this age group, and premature lations using the VSL to monetize premature mortality due to cancers of the digestive system. mortality and chronic bronchitis. The mean esti- Here we monetize the premature mortality associ- mates and 5th and 95th percentiles refer to the ated with water pollution and the morbidity asso- uncertainty bounds for the number of cases of ciated with diarrheal disease. Cancer morbidity is mortality and morbidity. not monetized due to the difficulty in calculating Several points are worth noting. The mean the cost of treating an episode and the percent of total health cost associated with ambient air pol- episodes treated. For this reason, the estimates lution in urban areas of China in 2003 is 157 bil- below must be regarded as lower bounds to the lion Yuan if the adjusted human capital approach total health costs associated with water pollution to valuation is used, and 520 billion if WTP in China. estimates from the VEHR study are used. Use of To monetize premature mortality using the WTP increases total costs by a factor of 3.3, AHC approach requires an estimate of rural per bringing health costs to 3.8 percent of 2003 capita GDP. There are no official data on urban GDP. Using the AHC approach, health costs are and rural GDP in China. However, based on our calculations GDP is approximately 5,384Yuan.10 We assume, following the Green National Ac- T A B L E 4 . 5 Health Costs Associated with Outdoor Air Pollution in China, 2003 Willingness to Pay counting Report, that premature mortality due Approach (Bil. Yuan) to cancers of the digestive system results in a loss of 21 years of life. This implies, for the central Morbidity case of = .07 and r = .08, that HCm = 102,242 Yuan for a statistical cancer death. Table A.1 Direct Indirect implies a loss of approximately 78 years of life for Excess Chronic Hospital Hospital Total Estimate Deaths Bronchitis Costs Costs Costs a child who dies of diarrheal disease before age 5. Using the same per capita rural GDP figure 95th %ile 641.1 136.7 4.82 0.670 783.3 implies that HCm = 297,251 yuan for a statistical Mean 394.0 122.1 3.41 0.470 519.9 death due to diarrheal disease. Using the VSL 5th %ile 135.6 106.2 1.88 0.263 243.9 approach, both deaths are valued at 1.0 million Yuan. These assumptions lead to the results Source: Authors calculations based on the China National Health Survey 2003 and other sources. reported in Table 4.6. 74 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S risk changes as economists expect them to. T A B L E 4 . 6 Health Costs Associated 3. This adjustment is made using the ratio of average dis- with Water Pollution in posable income in China to average disposable income China, 2003 (Bil. Yuan) in Shanghai and Chongqing. The income elasticity of 0.48 from Krupnick et al. (2006) is used to make the AHC VSL adjustment. Morbidity Mortality Mortality 4. For this reason, over one-third of sample respondents Disease Cost Cost Cost were chosen to be 60 years of age and older. 5. For a summary of this literature see the USEPA's Hand- Diarrhea 0.22 4.16 14.0 book on Valuing Children's Health at http://yosemite. Cancer N/A 5.31 52.0 epa.gov/ee/epa/eed.nsf/webpages/HandbookChildrens Total 9.47 66.0 HealthValuation.html 6. It should be noted that the adjusted human capital Source: Authors Calculations values in Table 4.2 pertain to cities, whereas the results below are reported for provinces and municipalities. 7. One such approach is the standard gamble approach, Morbidity due to diarrheal disease in children used by Hammitt and Zhou (2005). This approach asks a person, were he to contract chronic bronchitis, what under the age of 5 is valued at a cost of two days risk of death he would accept to undergo an operation of caregivers' time. This was calculated as 29.5 that would cure the disease with probability 1-. Yuan per case, the pro-rated value of per capita 8. Total Health Costs = Cases of Premature Mortality rural GDP.11 Cost per Case + Cases of Chronic Bronchitis Cost per Although an underestimate of the total health Case + Direct Cost of Hospital Admissions + Indirect Cost of Hospital Admissions. costs of water pollution, the costs in Table 4.6 are 9. This assumes, strictly speaking, that the slope of the about an order of magnitude smaller than the Ostro relative risk function in Table 2.3 is approxi- health costs associated with outdoor air pollu- mately linear over the relevant range of ambient con- tion. This is true even when outcomes are valued centrations. 2 using the same VSL for persons in rural and 10. That is, 29.5 = (5384/365) urban areas. Compared to the health cost of air pollution, the health cost of water pollution are References relatively low. This does not, of course, mean that individual projects to improve rural drink- Alberini, A., M. L. Cropper, T. Fu, A. Krupnick, J.-T. Liu, D. Shaw and W. Harrington. 1997. "Valuing health ing water quality will necessarily yield smaller net effects of air pollution in developing countries: The case benefits than specific projects to improve urban of Taiwan." Journal of Environmental Economics and air quality. It should also be noted that improve- Management 34:107­126. ments in surface water quality, which are one way Alberini, A., M. Cropper, A. Krupnick, and N. Simon. 2004. "Does the value of statistical life vary with age and of reducing the costs of drinking water treatment, health status? Evidence from the U.S. and Canada." will yield non-health as well as health benefits. Journal of Environmental Economics and Management 48: 769­792. Bartel, A., and P. Taubman. 1979. "Health and labor Endnotes market success: the role of various diseases." Review of Economics and Statistics 61:1­8. 1. Estimates of compensating wage differentials are often Black, Dan et al. 2003. How Robust Are Hedonic Wage quite sensitive to the exact specification of the wage Estimates of the Price of Risk? Report to the USEPA equation. Black et al. (2003), in a reanalysis of data [R82943001]. Washington, DC: Government Printing from U.S. compensating wage studies requested by the Office. USEPA, conclude that the results are too unstable to be Freeman, A. M. III. 1993. The measurement of environmen- used for policy. tal and resource values: theory and methods. Washington, 2. For example, WTP for a reduction in risk of death sel- D.C.: Resources for the Future. dom increases in proportion to the size of the risk Hammitt, J. K., and Y. Zhou. 2005. "The economic value of change, which suggests that respondents do not perceive air-pollution-related health risks in china: a contingent C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 75 V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S valuation study." Environmental and Resource Economics study of U.S. adults." American Journal of Critical Care 33:399­423. Medicine 151:669­74. Krupnick, A., and M. Cropper. 2000. "The social costs of Pope, C. A. et al. 2002. "Lung cancer, cardiopulmonary chronic heart and lung disease." In Valuing Environ- mortality, and long-term exposure to fine particulate air mental Benefits, Selected Essays of Maureen Cropper. pollution." Journal of the American Medical Association Cheltenham, UK: Edward Elgar. 287:1132­41. Krupnick, A., S. Hoffmann, B. Larsen, X. Peng, R. Tao, U.S. Environmental Protection Agency. 1997. The benefits and C. Yan. 2006. The willingness to pay for mortality and costs of the clean air act, 1970 to 1990. Report to the risk reductions in Shanghai and Chongqing, China. U.S. Congress. Washington, DC: Government Printing Washington, D.C.: Resources for the Future. Office. Loehman, E., and V. De. 1982. "Application of stochastic Viscusi, W. K. 1993. "The value of risks to life and health." choice modeling to policy analysis of public goods: a Journal of Economic Literature 31:1912­1946. case study of air quality improvements." Review of Viscusi, W. K., W. Magat, and J. Huber. 1991. "Pricing Economics and Statistics 64:474­480. environmental health risks: A survey assessment of risk- Li, Y., M. Bai, W. Zhang, K. Yang, and X. Wang. 2002. risk and risk-dollar tradeoffs for chronic bronchitis." "Analysis on the influence factors of residents' willing- Journal of Environmental Economics and Management ness to pay for improving air quality in Beijing." China 21:32­51. Population, Resources and Environment 12: 123­126 Wang, H., and J. Mullahy. 2006. "Willingness to pay for Liu, J.-T., J. K. Hammitt, and J-L. Liu. 1997. "Estimated reducing fatal risk by improving air quality: a contingent hedonic wage function and value of life in a developing valuation study in Chongqing, China." Science of the country." Economics Letters 57:353­358. Total Environment 367:50­57. Miller, W., L. A. Robinson, and R. S. Lawrence, eds. 2006. Zhang, X. 2002. Valuing mortality risk reductions using the Valuing health for regulatory cost-effectiveness analysis. contingent valuation methods: evidence from a survey of Washington, D.C.: The National Academies Press. Beijing Residents in 1999. Beijing: Centre for Environ- Pope, C. A., M. Thun, M. Namboodiri, D. Dockery, J. ment and Development, Chinese Academy of Social Evans, F. Speizer, and C. Heath. 1995. "Particulate air Sciences. pollution as a predictor of mortality in a prospective 76 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N V A L U A T I O N O F E N V I R O N M E N T A L H E A L T H R I S K S A P P E N D I X A . 1 Average Life Years Lost due to Air Pollution RD CVD CEVD Remaining Age Life Lost Life Years Lost Life Years Lost Life Years Groups Expectancy Deaths × Deaths Deaths × Deaths Deaths × Deaths 0­ 78.79 1680.41 132393.79 266.69 21011.27 89.25 7031.41 1­4 78.51 518.07 40675.18 130.56 10250.80 21.93 1722.13 5­9 74.71 237.68 17756.99 68.13 5089.58 9.08 678.61 10­14 69.83 138.12 9644.70 138.12 9644.70 30.47 2127.51 15­19 64.92 195.99 12723.77 229.12 14874.26 99.38 6451.49 20­24 60.01 253.55 15215.41 548.45 32911.59 212.21 12734.64 25­29 55.15 480.56 26502.25 964.25 53176.60 436.88 24092.96 30­34 50.31 987.78 49696.39 2019.46 101601.51 1153.98 58058.01 35­39 45.50 1274.53 57991.06 3279.82 149232.08 2359.17 107342.37 40­44 40.73 1797.80 73231.00 4369.61 177990.75 4169.36 169833.60 45­49 36.07 3519.17 126930.67 7359.60 265448.36 8821.92 318191.93 50­54 31.49 4903.51 154393.74 8674.55 273130.10 10787.44 339657.12 55­59 27.06 6598.43 178537.66 9180.92 248413.60 12547.36 339501.39 60­64 22.86 14205.99 324757.07 16643.19 380473.08 21818.92 498793.14 65­69 19.01 25778.30 489933.95 29385.85 558497.67 36225.79 688495.36 70­74 15.64 41228.53 644924.21 39312.80 614957.14 47827.45 748148.97 75­79 12.96 46403.88 601328.17 40830.47 529104.65 48111.39 623455.11 80­85 11.07 41399.67 458171.04 34687.21 383884.08 36150.15 400074.47 85­ 10.72 36785.43 394314.94 31578.30 338497.99 25073.53 268771.30 Total 228387.42 3809121.98 229667.09 4168189.83 255945.66 4615161.51 Average lost statistical years 16.68 18.15 18.03 Note: Deaths are the product of the population in the survey report of the national 5th population census and the disease-specific death rates in the Health Statistical Yearbook. RD = Respiratory disease; CVD = Cardiovascular disease; CEVD = Cerebrovascular disease C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 77 5 NON-HEALTH IMPACT OF WATER POLLUTION 5.1 Water Scarcity and Pollution Water scarcity is most prevalent in Water scarcity is predominantly an issue in northern China. While most of northern China. High pollution in this China's water resources are in the south, the greatest need for these resources region exacerbates water scarcity. is in the northern and eastern part of the country, where most of the people Polluted water is held back from live. The four northern river basins contain less than 20 percent of national supply and becomes a source of water resources, but account for two-thirds of the farmland and 45 percent water scarcity. However, some water of GDP. By contrast, the southwestern areas contain slightly more water is allowed in the supply despite being resources (21.3 percent), but account for only 8.3 percent of GDP.1 too polluted; in such a case, pollution becomes a consequence of water The concentration of people and economic activity leads to water scarcity. scarcity. Groundwater depletion is a Water scarcity has a number of different definitions. The United Nations partly overlapping consequence of Environment Programme (UNEP) defines it as a state in which "the amount water scarcity that also creates a of water withdrawn from lakes, rivers or groundwater is so great that water major environmental problem in supplies are no longer adequate to satisfy all human or ecosystem require- China. We found that between 2000 ments, bringing about increased competition among potential demands." and 2003 polluted water supply (http://freshwater.unep.net.) The statement invites an interpretation in constituted about 47 billion cubic which there is a deficit of water: water withdrawn is larger than supply, which meters of water while polluted water is no longer adequate. held back from supply constituted An economist, on the other hand, would routinely define water scarcity as a about 25 billion cubic meters. situation in which demand for water, or water withdrawn, exceeds supply at a Groundwater depletion constituted price of zero. This means that the available water is not sufficient for everybody about 24 billion cubic meters. The economic cost of the pollution- to meet their needs at no financial cost. The economist's definition is echoed in related sources of water scarcity is part of the UNEP definition regarding competition among potential demands. estimated to be 147 billion RMB In the absence of sufficient quantities, there must be competition between yuan, with a 95 percent confidence demands and an associated opportunity cost as reflected by the price of water. interval relative to uncertainty in In this chapter, we define water scarcity as a state in which available water valuation of 95 and 199 billion RMB. resources per capita fall (far) below sustainable levels. Under such circum- The cost of groundwater depletion stances, there is a competition among potential demands and not all human comes at a further 92 billion RMB. and ecosystem requirements are met, as suggested by the UNEP definition. There is also a real risk that water supplies are no longer adequate to meet demand, and there is insufficient capacity to satisfy everyone's needs at a price of zero, as suggested by the economist's definition. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 81 W A T E R S C A R C I T Y A N D P O L L U T I O N A combination of historical and contemporary polluted water is withheld at the expense of more trends can explain water scarcity in China's north groundwater depletion. Furthermore, ground- and east. For various historical reasons, people water itself is often polluted by both natural have settled and prospered in the northern and and anthropogenic sources. An investigation of eastern parts of the country, despite the low water drinking water in 118 cities carried out by the resources, so contemporary China has inherited Ministry of Water Resources (MWR) found that relatively high population density in these regions. groundwater was polluted to varying degrees in Recent population growth and high economic 97 percent of the cities. Figure 4.6 shows that the growth have further increased the demand for two main groundwater pollutants are arsenic and water, while pollution of water basins as well as fluoride. technical deficiencies in the water supply facilties, We sought to estimate the cost of ground- such as leaky pipes and canal, have reduced avail- water pollution as a way of estimating pollution able water resources. as a source of water scarcity. We also estimated In this chapter, we are focusing on pollution as total groundwater depletion, because it is a seri- a cause of water scarcity. We attempt to estimate ous environmental problem and the method for the economic costs that arise from the inability to separating out pollution related to groundwater make productive use of polluted water. In addi- depletion has some uncertainties. The cost of tion, we seek to identify environmentally un- total groundwater depletion is not added to the sustainable responses to water scarcity. In China, total environmental cost estimate, but is avail- about 10 percent of the current water supply is able as stand-alone information. too polluted to be usable. Pollution, therefore, increases the volume of water that is held back WATER RESOURCES IN CHINA from use, but some of this water is supplied despite pollution, which also has a high economic To understand the nature of water scarcity in cost. Industry and agriculture are the largest China, it is useful to begin with a survey of avail- consumers of polluted water, despite fairly lax able, rechargeable water resources. A common standards for what passes as acceptable water. We indicator is the "water crowding" index of pop- also attempted to estimate the environmental cost ulation per million cubic meters per year. Here of consumption of such highly polluted water. we use the inverse measure of cubic meters per In addition to pollution in the sense of not meet- person. Levels of 1,000­1,700 cubic meters per ing the standards, there is also an issue of water person indicate water stress, and less than 1,000 that passes as acceptable but is fairly dirty. We cubic meters per person indicates extreme water do not attempt to estimate the environmental scarcity (Vörösmarty et al. 2006; Falkenmark cost of that water. 1997). Figure 5.1, with data from NBS (2004), Water scarcity in China also leads to depletion shows that in six provinces in China, per capita of groundwater resources. Depletion of ground- water resources fall below 500 cubic meters.2 In a water resources, particularly deep aquifers, is further five provinces, water resources falls below another environmentally unsustainable response 1,000 cubic meters, meaning that one-third of and adds to China's environmental costs. In some China's provinces qualify for extreme and more areas of China, the groundwater table has fallen than extreme water scarcity.3 Available resources more than 50 meters since 1960, and it contin- depend on precipitation. Data for water resources ues to fall two meters annually. are from 2003, which was an average year in terms Groundwater depletion is, to some extent, of water resources nationwide.4 linked to the pollution problem. In several The figure shows that per capita water re- provinces, like in the lower reaches of the Yangtze, sources are lowest in the Huang-Huai-Hai river 82 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N F I G U R E 5 . 1 Per Capita Water Resources in China in 2003 N W E S per capita water resources (cu.m/person) <500 500 - 1000 1000 - 1700 1000 0 1000 2000 Kilometers >1700 basins, especially the lower reaches. The lower POLLUTION AS A SOURCE reaches of the Huang River, also called the Yellow OF WATER SCARCITY River, used to dry up until the construction of the Xiaolangdi Dam (Berkoff 2003). The middle As discussed in chapter 3, the most polluted reaches of the Huang River have also experienced water basins in China are located in the north- dry spells (Zhu 2006). The middle and lower ern and eastern parts of the country in the same reaches of some tributaries of the Hai River tend regions that have low water resources per capita. to be dry all year round. Since water is polluted, less is available for con- On the other hand, several provinces in south- sumption in households, industries and agricul- western China have abundant water resources, ture, which further exacerbates the serious water including Yunnan, Qinghai, and Tibet. The scarcity situation. Yangtze River (Changjiang), Pearl River, and The decrease in water consumption below the rivers in the east and south together have levels needed by households, industries and agri- 80 percent of the water resources of China (see culture, called repressed demand, is one possible Table 5.1) (MWR 2005a). The average per capita impact of low water resource availability. Another water resources for all of China was 2,131 cubic impact may be increased groundwater depletion, meters in 2003. which may happen if water authorities consider The scarcity of water resources in the Huang- it a priority to maintain water supply. Increased Huai-Hai is particularly pressing in dry years. reliance on groundwater is also the decentralized The Hai and Huai flows fall to 70 percent of response of farmers, who may dig wells when average in one year in four and to 50 percent one they are not allowed surface water for irrigation. year in twenty (Berkoff 2003). Dry years tend to Anecdotal evidence and the expert opinion of come in succession, accentuating the problem. MWR suggest that groundwater depletion is in C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 83 W A T E R S C A R C I T Y A N D P O L L U T I O N T A B L E 5 . 1 The Quantity of Water Resources in China, Average 1956­2000 Shallow Rechargeable Non-Overlapping Water Resource Precipitation Surface Water Groundwater Quantity Total Region (I class) (billion m3) (billion m3) (billion m3) (billion m3) (billion m3) Songhuajiang River 471.9 129.6 47.8 19.6 149.2 Liaohe River 171.3 40.8 20.3 9.0 49.8 Haihe River 171.2 21.6 23.5 15.4 37.0 Yellow River 355.5 59.4 37.8 11.2 70.7 Huaihe River 276.7 67.7 39.7 23.9 91.6 Changjiang River 1937.0 985.7 249.2 10.2 996.0 Southeast rivers 437.2 265.4 66.5 2.7 268.1 Pearls River 897.2 472.3 116.3 1.4 473.7 Southwest rivers 918.6 577.5 144.0 0.0 577.5 Northwest rivers 542.1 117.4 77.0 10.2 127.6 Total 6178.7 2737.4 822.1 103.6 2841.2 Source: Ministry of Water Resources. fact a common alternative to polluted surface lower reaches of the Yangtze, mentioned in the water. introduction as an area where groundwater deple- There are no readily available statistics on pol- tion substitutes for polluted water. luted water that is held back from water supply in China. To estimate the amount of water held IMPACT OF WATER SCARCITY: back, we rely on the assumption that ground- WATER POLLUTION IN SUPPLY water depletion and repressed demand are the two responses to holding back water. See Box 5.1 Water scarcity has led China to make use of for details about the method. excessive amounts of polluted water in its water Table 5.2 presents the estimate of non- supply. Polluted water is supplied to households, supplied polluted water by province. Hebei industry and, in particular, agriculture. Water Province and Shandong Province are estimated for households and industry is in most cases to have the largest volume of non-supplied pol- treated before consumption. Impacts of house- luted water. Ningxia Autonomous Region and hold consumption of polluted water are dis- Shanghai both have zero non-supplied polluted cussed in chapter 3. Impacts of wastewater water. The reasons are rather different, however. irrigation are discussed in this chapter. Impacts In Ningxia, the situation is so tight that all avail- on industries include lower product quality and able water resources, including all available pol- production stoppages. For instance, a report by luted resources, are used in supply. In Shanghai, Chang, Seip, and Vennemo (2001) from a the supply of water is greater than demand for Chongqing silk production plant found that the water, and there is no recorded depletion of raw silk became yellow when polluted water was groundwater. Shanghai's situation should prob- used, and its quality fell from 5A to 3A or 2A. ably be viewed in the context of neighboring Another silk production plant and a fertilizer Jiangsu Province, which has the third highest plant were forced to stop production. non-supplied polluted water volume in the This chapter provides a comprehensive pic- country. Jiangsu Province is the home of the ture of the extent of polluted water in supply. 84 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N B O X 5 . 1 Constructing an Estimate of Polluted Water Held Back from Supply We distinguish between water-scarce water basins on the one hand, and water-abundant water basins on the other. Water-scarce areas not only have little water resources per capita--the indicator we emphasized above--but also have high consumption rates relative to their water resources. In the literature, a consumption rate of 20­40 percent is considered medium to high (World Meteoro- logical Organisation 1997; Vörösmarty et al. 2006). We define a water-scarce basin in China as one in which the consumption rate is above 40 percent. We made separate calculations for each of the 73 water basins and provinces distinguished by MWR. In water-scarce basins, we assume that the quantity of polluted water that is held back equals either the quantity of depleted groundwater or a measure of repressed demand, whichever is largest, times the share of polluted water in the water resource. This share is defined as the share of water of quality IV and worse. That is, all water that is unsuitable for bodily contact is consid- ered polluted. To measure the share, we take the weighted sum of river sections in which mea- sured water is polluted divided by the total weighted sum. The weights are the lengths of the river sections. In effect then, the shares give polluted water in the rivers and basins measured by length. In symbols: (1) PW = max (G,RD ) In this equation, PW represents polluted water that is withheld, G is groundwater depletion, RD is repressed demand, and is the share of polluted water in the resource. The idea here is that the more pollution there is in the resource, the greater is the share of groundwater depletion or repressed demand that can be attributed to non-supplied polluted water. If the share is zero, then none of the (presumably low) amount of water depletion or repressed demand should be attrib- uted to pollution. If the share is one, then all water depletion or repressed demand should be attributed to pollution. Repressed demand is calculated as the difference between notional demand, which is a plan- ning indicator of MWR, and sustainable supply; that is, current supply excluding groundwater depletion and supply of polluted water. For example, in Shanxi Province repressed demand is estimated to 0.76 billion cubic meters. That is larger than the volume of groundwater depletion, estimated at 0.54 billion cubic meters. In Shanxi Province, the share of polluted water is 71 percent. Polluted non-supplied water is estimated to be 71 percent of 0.76 billion cubic meters, or 0.54 billion cubic meters. By coincidence, that number corresponds to the volume of groundwater depletion. That is not the case in all provinces. There is one exception to equation (1). It is possible to estimate how much water is withheld in a water basin in total. If that volume is lower than the estimate coming out of equation (1), there is obviously a problem with the estimate from equation (1). The volume of polluted water that is held back cannot be higher than the volume of all water that is held back. To eliminate this possi- bility, we assume that when the total volume is lower than the largest of groundwater depletion and repressed demand, then polluted withheld water equals total available water times the pol- luted share. In symbols (2) PW = TW , TW max (G,RD ) It is a complicated matter to estimate TW, available total water. First, we calculate the total amount of polluted water in the resource, using estimates of the total surface water resource (the basis for the 40 percent or more consumption rates) and the polluted share. From this volume, we subtract the amount of polluted water that is supplied. In some basins, there is also some residual clean water not supplied, which is added to the total resource. The net result of these operations is the estimate of available total water TW in a water basin. For example, in Tianjin the available total water is estimated to be 0.212 billion cubic meters. That is slightly lower than either repressed demand or groundwater depletion, which both amount to about 0.215 billion cubic meters. The estimate of 0.212 in available total water equals (continued) C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 85 W A T E R S C A R C I T Y A N D P O L L U T I O N B O X 5 . 1 Constructing an Estimate of Polluted Water Held Back from Supply (Continued) the total surface water resource of 1.1 billion cubic meters times a polluted share of 86 percent. From this number is subtracted the amount of polluted water that is supplied, which is as high as 0.744 billion cubic meters (see Table 5.). There is no surplus clean water in Tianjin. With 0.212 as the estimate of total available water in Tianjin, we use equation (2) and estimate polluted withheld water to be 86 percent of 0.212 billion cubic meters, which is 0.18 billion cubic meters. Equations (1) and (2) apply to river basins in which consumption is 40 percent or more of the rechargeable resource. When consumption is lower than 40 percent, there is no resource-oriented reason for groundwater depletion. The expert judgment of MWR and SEPA is that in river basins of less than 40 percent consumption, the reason for groundwater depletion is that available resources are too polluted for use. There is normally no repressed demand in these river basins. Accordingly, we assume that in these basins non-supplied polluted water equals water depletion. In symbols (3) PW = G, consumption 40% Fifteen southern provinces--including Jiangsu, Sichuan, and Guangdong--are in a situation where consumption is lower than 40 percent. The procedure to estimate non-supplied polluted water relies on a number of untested assumptions. Yet SEPA and MWR consider that it gives a rough indication of non-supplied pol- luted water in China. By polluted water in supply, we refer to water Heilongjiang and Ningxia). Note that the cor- that exceeds the water quality standard relevant relation is specified between a per capita measure for the purpose. Polluted water for households and an aggregate measure. The amount of pollu- refers to water worse than class III supplied tion in Ningxia is particularly remarkable, since (after treatment) to households; polluted water it has a very small population. Jiangsu is situated for industrial purposes refers to water worse in the lower reaches of the Yangtze River. On its than class IV supplied (after treatment) to border is the final destination of the Huai River, industry; and polluted water for agricultural which never reaches the sea. Ningxia is in the purposes refers to water worse than class V sup- upper reaches of the Huang (Yellow) River. plied to agriculture. While the correlation is striking, there are Figure 5.2 shows the volume of water supply also differences with the map of water resources (in millions of cubic meters) that does not meet per capita. Supplies of polluted water in the supply standards for each province. The map Hebei-Beijing-Tianjin area are large, but relative was produced using MWR survey data from to other provinces the situation is better than it 2000 to 2003. The multiyear coverage allows us is in terms of water resources. Heilongjiang to account for annual variations due to rainfall Province, by contrast, has a serious problem and other factors. with the supply of polluted water compared to The map shows significant correlation with its water resources. Altogether, close to 50 bil- the map of water resources. For instance, Ningxia lion cubic meters of water that did not meet the and Shanghai are the provinces in the country pollution standard were supplied annually dur- with the lowest per capita water resources in ing the 2000­03 period. This figure is close to 2003. Ningxia, Jiangsu, and Heilongjiang are the 10 percent of average national water consump- provinces with the largest supply of polluted tion in the period, which was 566 billion cubic water (9.5 billion cubic meters in Jiangsu, 4.0 in meters. 86 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N ing sediments as fertilizer encourages excessive use T A B L E 5 . 2 Non-Supplied Polluted Water by Province of irrigation water (Yang, Zhang, and Zehnder 2003). Pollution of irrigation water is related to Polluted, about 62 billion cubic meters of sewage in China Non-Supplied Water in 2000 (MWR, 2000), of which only 24 per- Province (million m3) cent was treated up to standard. Beijing 138.1 Tianjin 184.2 IMPACT OF WATER SCARCITY: Hebei 3,618.7 Shanxi 540.6 DEPLETION OF GROUNDWATER Inner Mongolia 2,498.1 Liaoning 613.7 Water scarcity has forced China to rely increas- Jilin 317.4 ingly on groundwater, which has led to depletion Heilongjiang 970.8 of groundwater reservoirs. It is useful to distin- Shanghai 0.0 Jiangsu 2,165.8 guish between rechargeable groundwater in the Zhejiang 496.4 shallow freshwater aquifer (phreatic water), and Anhui 1,647.7 non-rechargeable groundwater in the deep fresh- Fujian 349.4 water aquifer (confined water). Shallow ground- Jiangxi 319.3 Shandong 2,716.9 water recharges from, and/or discharges to, Henan 1,991.4 precipitation and surface water flows; for exam- Hubei 367.1 ple, compare the columns shallow rechargeable Hunan 353.5 Guangdong 1,083.9 groundwater and non-overlapping quantity in Guangxi 572.5 Table 5.1. Depletion of shallow groundwater Hainan 135.9 occurs when consumption exceeds sustainable Chongqing 311.1 Sichuan 697.1 levels. Deep groundwater recharges/discharges Guizhou 361.6 extremely slowly. Replenishment rates can be in Yunnan 1,058.1 the order of thousands of years. Depleting deep Tibet 175.6 groundwater is similar to mining a nonrenewable Shaanxi 520.5 Gansu 400.2 resource. Qinghai 32.0 Depletion of groundwater may have serious Ningxia 0.0 consequences for the environment. One is salin- Xinjiang 124.9 Total 24,762.4 ity intrusion, as declining groundwater resources are substituted by brackish water that often lies Source: Authors Calculation. between the shallow and deep groundwater Note: Polluted water is water of class IV or worse. tables. Salinity intrusion is also caused by sea- water intruding from the outside. Land sub- sidence following compaction of the geological A breakdown of polluted water supply by formation containing groundwater (so-called consumption sector shows agriculture receiving aquitard) is another unfortunate consequence of two-thirds of the water, and industry receiving groundwater depletion. In China, salinity intru- 20 percent. The ratios differ by province. In the sion is a chronic problem, such as in the Hai River two high consumption provinces of Jiangsu and Basin (Zhu 2006). In some locations, intrusion Ningxia autonomous region, agricultural con- of brackish water has been monitored at a rate sumption constitutes 62 percent and 98 percent of 0.5­2 meters per year for the past 20 years (respectively) of all consumption. In Ningxia, it (Foster et al. 2004). In turn, salinity intrusion has been reported that a local tradition of apply- poses problems for the waterworks and for human C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 87 W A T E R S C A R C I T Y A N D P O L L U T I O N F I G U R E 5 . 2 Polluted Water in Supply in China N W E S water supply not meeting standard 0 - 500 500 - 1500 1500 - 3000 3000 - 4000 >4000 2000 0 2000 Kilometers Source: Ministry of Water Resources and agricultural use. A recent episode of this kind Besides the environmental aspects, ground- occurred in the Pearl River delta in the south. water depletion carries an economic cost. As the MWR (2005) reports that in 2003­04, follow- groundwater table falls, the cost of pumping it ing a period of 30 percent lower precipitation becomes high, especially for agricultural purposes. than normal, the Pearl River delta suffered a case Figure 5.3 shows groundwater depletion by of severe salinity intrusion. The salinity intru- province in the period 2000­03 and refers to the sion restricted operation of the waterworks of survey of MWR mentioned above. The figure Pearl River and Macao for 170 continuous days. refers to depletion of shallow and deep ground- More than 5 million people, as well as industries water in total. and agriculture, were affected to varying degrees. The figure shows that depletion of ground- Both for natural and man-made reasons, the water extends from the Huang-Huai-Hai plain quality of groundwater often is poor. An investi- to almost every province in the north, including gation of drinking water in 118 cities carried out heavy depletion in Inner Mongolia, and sub- by MWR found that groundwater was polluted stantial depletion in Xinjiang. It reinforces the to varying degrees in 97 percent of the cities. In message that the north and east have the most 64 percent of cities, groundwater was seriously serious problems of water scarcity and are the polluted. Data from MWR indicates that in main source of water depletion. Particular prob- 30 percent of the area supplied by groundwater, lems are evident in Hebei Province and sur- people should not use the water for drinking rounding provinces, most of which belong to the purposes.5 This area contains a disproportion- Huang-Huai-Haiplain. ate number of cities. In a further 30 percent of In Hebei, 6 billion cubic meters of ground- the area, groundwater needs to undergo water water were depleted annually in 2000­03. Zhu treatment.6 (2006) comments that part of the aquifer in 88 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N T A B L E 5 . 3 Supply of Surface Water that Does Not Meet Pollution Standards (106m3) Urban Rural Large Other Domestic Domestic Industrial Irrigation Agricultural Surface Province Use Use Use Use Use Total Beijing 0 0 0 100 0 100 Tianjin 0 0 13 731 0 744 Hebei 0 0 14 910 1159 2083 Shanxi 30 69 105 342 612 1158 Inner Mongolia 13 0 67 468 194 742 Liaoning 41 0 381 2049 233 2703 Jilin 309 0 55 190 1084 1639 Hailongjiang 289 0 1190 354 2211 4045 Shanghai 1184 74 1018 0 646 2922 Jiangsu 707 609 2281 280 5629 9508 Zhejiang 379 215 1253 805 1011 3663 Anhui 133 254 134 0 1919 2440 Fujian 1 12 0 0 0 13 Jiagxi 4 0 141 0 67 211 Shandong 52 2 14 15 503 586 Henan 226 5 408 639 1396 2674 Hubei 101 5 334 10 54 504 Hunan 305 270 648 0 627 1850 Guangdong 1138 50 1276 0 100 2563 Guangxi 19 53 124 0 88 284 Hainan 0 0 0 0 0 0 Chongqing 69 69 103 0 36 277 Sichuang 247 126 150 0 73 596 Guizhou 1 0 50 0 1 52 Yunnan 0 0 0 0 0 0 Tibet 0 0 0 0 0 0 Shaanxi 4 3 0 290 0 297 Gansu 252 46 167 0 531 997 Qinghai 12 1 14 0 189 216 Ningxia 0 2 135 2162 1655 3953 Xinjiang 0 0 0 32 0 32 Total 5516 1866 10073 9348 20018 46821 Source: Authors calculation. Hebei and Beijing are nearly dried up, and in implies a deep-aquifer depletion of more than other parts the groundwater table is sinking 50 meters. 3­5 meters annually. In most of the Huang-Huai- Table 5.4 indicates amounts of annual ground- Hai plain, the groundwater table has dropped. water depletion in 2000­03 between provinces, The drop is 2­3 meters in some areas and as much as well as its distribution between urban and rural as 10­30 meters in others. Foster et al. (2004) households, industry, and agriculture. Ground- state that in rural areas of the Huang-Huai-Hai water depletion totaled 24 billion cubic meters. plain "an average value for deep aquifer ground- As can be seen from the table, a common use water-level decline of more than 3 m/year dur- of groundwater is for irrigation of agriculture. In ing the period 1970­80 has now reduced to fact, 74 percent of all groundwater depletion is for 2 m/year." Compounded over 35 years, this agricultural purposes. In many areas irrigated by C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 89 W A T E R S C A R C I T Y A N D P O L L U T I O N F I G U R E 5 . 3 Groundwater Depletion by Province (million cubic meters) N groundwater depletion (million cubic meters) 0 - 300 300 - 600 600 - 1500 1500 - 3000 >3000 Source: Ministry of Water Resources groundwater, each pump serves only a small num- meters of water that does not meet quality stan- ber of farmers (Yang, Zhang, and Zehnder 2003). dards is nevertheless supplied to households, It is a sign both of the strain on water resources industry, and agriculture. A further 24 billion and of the spread of groundwater depletion that cubic meters of water beyond rechargeable quan- in 1997 alone, 221,000 wells were drilled on the tities is extracted from wells and creates ground- Huang-Huai-Hai plain, while 100,000 wells were water depletion. Although there are some deserted. In Beijing and Tianjin, the numbers of overlaps, close to 100 billion cubic meters of newly drilled wells were outstripped by those water in China is affected by pollution and deserted (Yang and Zehnder 2001). other environmental stress. By comparison, the total supply of water in China is approximately 550 billion cubic meters (NBS 2006). THE ENVIRONMENTAL COST Most experts agree that water scarcity substan- OF WATER SCARCITY tially restricts economic development. According We have found that approximately 25 billion to Zhu (2006), water scarcity becomes "an obsta- cubic meters of polluted water in China is held cle for the enhancement of people's living stan- back from water consumption, contributing to dard as well as construction and development of problems of repressed demand and ground- big water-consuming industrial enterprises." water depletion. As much as 47 billion cubic Furthermore, water scarcity "restricts agricul- 90 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N tural development in North China and threat- T A B L E 5 . 4 Depletion of Groundwater ens food safety." Another impact of water scarcity is to increase Depletion of Groundwater (106m3) the frequency and force of droughts. Droughts lead to economic loss and human strain. In Quantity 2004, for example, 218 million mu of cropland were damaged by drought, causing 19.9 million Urban Rural Province Domestic Domestic Industrial Irrigation Total tons of damage to grain production and an eco- nomic loss of 24.7 billion yuan. As a result of Beijing 74 8 64 114 261 this drought, 23.4 million people were left tem- Tianjin 45 8 49 113 215 porarily without drinking water supply (MWR, Hebei 316 331 909 4553 6109 2004). According to MWR, more severe droughts Shanxi 37 35 108 363 543 Inner Mongolia 75 87 159 2406 2728 had preceded the one in 2004. Liaoning 127 51 280 823 1280 Taking arguments like these forward, several Jilin 37 28 121 419 605 authors have tried to systematically assess the Heilongjiang 67 39 305 1078 1489 Shanghai 0 0 0 0 0 value of water in China. Box 5.2 describes some Jiangsu 64 44 364 703 1175 of these efforts. Zhejiang 11 6 32 66 116 Our conclusion from going through the evi- Anhui 21 47 144 318 530 Fujian 0 0 0 1 2 dence is that generally 1­5 RMB yuan per cubic Jiangxi 0 0 0 0 0 meter seems a fair value for water in China. To Shandong 138 153 480 1894 2665 be more concrete, one should consider what Henan 139 242 434 1655 2469 would have happened to the water if the envi- Hubei 7 4 31 58 101 Hunan 2 3 9 32 46 ronmental problem was not there. In terms of Guangdong 47 24 121 268 459 polluted water held back from supply, the value Guangxi 5 10 18 111 144 of marginal water seems a reasonable indicator. Hainan 7 8 5 102 122 Chongqing 0 0 0 0 0 To value this water, we use the paper of He and Sichuang 1 1 2 5 9 Chen (2005), which besides producing estimates Guizhou 0 0 1 2 3 that are a priori reasonable compared with other Yunnan 3 4 6 49 63 Tibet 0 0 0 0 0 sources, is a recent and comprehensive attempt. Shanxi 62 57 166 686 971 Given the rapid economic growth in China, Gansu 17 19 86 486 608 value added per unit of water increases consid- Qinghai 0 0 0 3 4 erably year by year. Recall, for instance, the Ningxia 2 1 8 146 158 Xinjiang 15 14 26 1309 1363 increase over time in agricultural output per unit Sum 1320 1224 3930 17763 24236 of water. It is, therefore, particularly important to use a recent estimate.7 Source: Authors Calculation. To value polluted water that is included in supply, we could in principle also use the esti- mates of He and Chen (2005). However, pol- sumption, and the cost of treatment is part of luted water used for irrigation purposes in the environmental cost. Other costs related to agriculture, particularly in so-called wastewater polluted supply for households, including health irrigation zones, is discussed later in this chap- costs, are discussed in chapter 4. Costs to indus- ter. To avoid overlap, we focus on polluted try in addition to treatment, such as halts in pro- supply for households and for industry, which duction, are not included. amount together to around 17 billion cubic Groundwater depletion can ideally be valued meters. This water needs treatment before con- by its environmental effects. As noted above, C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 91 W A T E R S C A R C I T Y A N D P O L L U T I O N B O X 5 . 2 Efforts to Value Water in China Relevant for Water Scarcity Efforts to estimate the value of water relevant for water scarcity usually follow one of two approaches. One approach is to estimate the value of marginal water. Estimating the value of marginal water means finding the economic value added if a little extra water were available in the economy--or the value deducted if a little less water were available. Using this approach and a so-called computable general equilibrium model, He and Chen (2005) find that a cubic meter of water adds between 2.1 and 5.2 RMB yuan of value. The value differs between water basins. The highest values are obtained in the Huang-Huai-Hai basins, which is reasonable, since these are the most water scarce basins. By comparison, an earlier paper by Liu and Chen (2003) uses a linear pro- gramming model and 1999 data to find that water for industrial purposes adds between 0.12 and 9.07 RMB yuan per cubic meter between provinces. The highest value is obtained for Ningxia Province. Several authors use the marginal value method in an informal way. Foster et al. (2004) argue that agricultural irrigation is the marginal use of water on the Huang-Huai-Hai plain. It is the sector that accommodates additional water (rainfall) and the sector that suffers during droughts. It is also the sector that would notice the shortfall if groundwater depletion was disallowed. The main crop affected by water scarcity is winter wheat. Liu and He (1996) report that on the Huang-Huai-Hai plain, 1.2 kg of wheat is grown per cubic meter of water (cited in Yang and Zehnder 2001). Jia and Liu (2000) estimate that in Shaanxi Province the figure is 1.3 kg, and increasing. (Their estimate for 1981 is 0.6 kg). Other authors, including Foster et al., use a somewhat lower number. With a national wholesale price of wheat of 1.15 RMB yuan/kg (see chapter 6 on wastewater irrigation), the implied value of water is currently approximately one RMB yuan per cubic meter, but as mentioned earlier, this estimate depends on irrigation of winter wheat being the marginal use of water. The marginal value of water can also be indicated by its price. It is likely that water is purchased to the extent that the value of water to the consumer is at least as high as the price; that is, the marginal value of water equals the price. With repressed demand, however, the marginal value is probably higher than the price. Increasing water prices is, therefore, included in MWR's strategy for a "water-saving society." Currently the average price of water in China's 36 major large and medium-sized cities is 2.1 RMB yuan per cubic meter (MWR 2006). This price refers to urban and domestic use. Prices have recently increased 10 percent annually, and will have to increase even more to make a serious impact on water consumption. Still, 2.1 RMB yuan is an estimate of the value of water from the side of the price. Of course, in some circumstances water has a much higher price--up to 2 RMB yuan per liter for bottled drinking water. However, in such cases the circumstance and packaging is part of the product. Another main approach to valuation of water is estimating the cost of current mitigation mea- sures. While in some cases estimating environmental cost via the cost of mitigation contains an ele- ment of circular reasoning, it is in other cases a useful measure of the political willingness to pay, or valuation of water. Other things being equal, it is the approved mitigation measures with the highest cost that come the closest to expressing a political valuation for water. A main element in China's strategy to end groundwater depletion is the South-North Water Diversion Project, which will trans- fer water from the Yangtze River to the Huang-Huai-Hai basin. The project will move up to 45 billion cubic meters annually to the basin. That number equals half of current water scarcity as estimated by this chapter. Water demand in China is likely to be significantly higher upon completion of the proj- ect (around 2050) than it is now. Still the project indicates a Chinese willingness to reduce and perhaps end pollution-related problems of water scarcity. The investment cost of the South-North Water Diversion Project is tentatively set at RMB yuan 486 billion, but after two years of investment it faced a 20 percent cost overrun (China Daily 2004). Ignoring cost overruns, Berkoff (2003) finds that the implied annual value of water is 0.7­0.9 RMB yuan per cubic meter. Allowing 20 percent higher investment cost increases these numbers to 0.9­1.1; adding in 0.06­0.38 RMB yuan annual operation and maintenance cost per cubic meter increases them to 1.2­1.3 RMB yuan per cubic meter. These estimates assume a 12 percent rate of return on the investment, which despite high economic growth in China and associated high return to capital, still could be on the high side. A high rate of return in the investment implies a high value of the future benefit stream, which is a high value of the diverted water. (continued) 92 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N B O X 5 . 2 Efforts to Value Water in China Relevant for Water Scarcity (Continued) Other mitigation options that have been discussed in the literature are related to agriculture in particular. They include agricultural crop changes (Yang and Zehnder 2001), efficient water irriga- tion techniques (Foster et al. 2004) and desalination (Zhou and Tol 2003), and import of agricul- tural produce. Import of agricultural produce is sometimes referred to as import of "virtual water," since it is an indirect way of transporting water to an area (Allen 1993). When polluted water is supplied, it requires treatment. A survey by CAEP (2006) of about 1,000 enterprises in ten provinces has estimated that the treatment cost for domestic purposes is about 2.6 and that for industry is 4.6 RMB yuan per cubic meter. The estimates are preliminary and will need more in-depth analysis. groundwater depletion may lead to salination any cost of groundwater depletion, it goes with- and to compaction of land. The main cost, how- out saying that depletion ends up higher than ever, may relate to groundwater's existence value a social planner would have desired. In terms and the speed of resource exhaustion; that is, of institutional distortions, MWR has recently water that has been accumulating underground emphasised the need to develop a water resources over thousands of years is being spent by only a management system on the basis of the theory of few generations. This is a cost that is relevant water rights and water markets (MWR 2005), and of concern not only for the Chinese, but the and the World Bank has long advocated water international community as well. management reform in China. There are thus rea- The environmental cost of groundwater deple- sons to assume that present groundwater deple- tion can be estimated directly, if one focuses on tion is not the outcome of rational choice. the salinity issue, but ignoring the importance of Using the values that we have described, existence value may lead to significant under- Table 5.5 sets out the environmental cost of estimation of the real cost. Therefore, we consider water scarcity by province. In the final estimate the marginal economic value of groundwater of environmental cost, we do not include the on the assumption that the environmental cost cost of groundwater depletion. One reason of groundwater is as least as high as its eco- is that groundwater depletion is not wholly a nomic value. pollution-related item. Another reason is that To say that the cost of groundwater depletion there is overlap between groundwater depletion is higher than its economic value is an untested and polluted water held back from supply. assumption that relies on water depletion being a Including both items in the estimate of environ- rational choice of society. Had it been a rational mental cost would imply double-counting. Note choice, the implicit value of untapped ground- also that the cost of wastewater irrigation is dis- water would have been lower than the economic cussed in section 5.2, while the health cost of value. While untested, there is general agree- water pollution is discussed in chapter 3. ment among experts that the present depletion of The environmental cost of water scarcity groundwater is the consequence of decentral- related to pollution amounts to 147 billion ized decisions without appropriate incentives for RMB. Among the provinces, Hebei and Jiangsu conservation. For instance, individual farmers in provinces have the largest environmental cost. groundwater irrigated areas usually do not pay for Hebei is dominated by the Hai River basin. the water itself. Farmers only pay for power and Jiangsu is split almost equally between the Huai equipment (Yang, Zhang, and Zehnder 2003). and Yangtze. The cost of groundwater depletion When farmers are not informed by the market of amounts to 92 billion RMB. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 93 W A T E R S C A R C I T Y A N D P O L L U T I O N T A B L E 5 . 5 The Environmental Cost of Water Scarcity Cost of Polluted Cost of Polluted Environmental Cost of Water Held Back Water in Cost--Sum Groundwater Province from Supply (1) Supply (2) of (1) and (2) Depletion (3) Beijing 721 0 721 1,360 Tianjin 961 60 1 022 1,122 Hebei 18,790 63 18,853 31,744 Shanxi 2,547 565 3,112 2,595 Inner Mongolia 9,431 344 9,775 10,735 Liaoning 1,881 1,872 3,753 3,925 Jilin 971 1,072 2,044 1,853 Heilongjiang 2,971 6,277 9,248 4,556 Shanghai 0 7,850 7,850 0 Jiangsu 5,745 12,437 18,182 3,217 Zhejiang 1,081 6,806 7,886 270 Anhui 4,648 971 5,619 1,438 Fujian 785 4 789 4 Jiangxi 921 662 1,583 0 Shandong 9,477 202 9,679 9,060 Henan 7,384 2,489 9,874 8,372 Hubei 1,061 0 1,061 290 Hunan 1,020 3,807 4,827 133 Guangdong 2,863 8,921 11,784 1,213 Guangxi 1,515 625 2,140 381 Hainan 359 0 359 321 Chongqing 660 659 1,319 0 Sichuan 1,478 1,349 2,826 19 Guizhou 1,022 236 1,257 9 Yunnan 2,623 0 2,623 156 Tibet 556 0 556 0 Shanxi 2,196 11 2,207 3,988 Gansu 1,237 1,440 2,677 1,861 Qinghai 126 98 223 12 Ningxia 0 625 625 722 Xinjiang 400 0 400 4,360 Sum 85,429 61,258 146,687 92,356 Note: For polluted water held back from supply the marginal value approach is used, He and Chen (2005). For polluted water in supply the mitigation cost approach is used, that is treatment cost, CAEP (2006). For groundwater depletion He and Chen (2005) is again used. Groundwater depletion is only available by province. Water basin values of He and Chen are aggregated using supply per water basin in a province as weights. UNCERTAINTIES AND SENSITIVITY limited knowledge of the sampling method, we subjectively estimate the uncertainty to about There are uncertainties both in the quantity and ±20 percent. The quantity of polluted water held value aspect of the environmental cost of water back from supply is estimated by the authors and scarcity. Quantities for polluted water in supply the uncertainty of the calculation is estimated to and groundwater depletion come from the MWR about ±40% percent. survey data from 2000­03, so there is significant The values used also are uncertain. The average uncertainty associated with these figures. With price used to value the cost of polluted water held 94 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N W A T E R S C A R C I T Y A N D P O L L U T I O N back from supply is 3.45 yuan RMB per cubic ation of ±1 on the estimate of 3.45 (3.81) seems meter (marginal production value). The average reasonable, and (1.45, 5.45) would then give a price used to value polluted water in supply is 95 percent confidence interval. 3.93 yuan RMB per cubic meter (treatment On the side of treatment cost, CAEP is cur- cost). The average price used to value ground- rently reanalyzing the data. Based on preliminary water depletion is 3.81 (marginal production analysis ±0.5 is a reasonable standard deviation, value, but different composition between and thus (2.93, 5.93) is a 95 percent confidence provinces than polluted water held back from interval. The uncertainty is larger on the side of supply). While it is reassuring that the prices industry treatment cost. from different sources are similar, there are sev- Treating the quantities as givens, we obtain a eral sources of uncertainty. On the side of pro- 95 percent confidence interval on the environ- duction value, the price depends on which sector mental cost of (95, 199) billion RMB yuan. and activity is marginal in the Chinese economy. In addition to the quantified uncertainty, there If the marginal activity is agriculture, in particu- are several omissions that contribute to making lar winter wheat, the value of water may go as low our estimate of environmental cost imprecise. For as one RMB yuan per cubic meter. If the mar- instance, the cost of treating water that is within ginal value is industry, the value may go above quality limits but is polluted is omitted from our six RMB yuan per cubic meter. A standard devi- estimate. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 95 5.2 Crop Loss Due to Wastewater Irrigation A combination of water scarcity, The output and quality of agricultural crop production is in many areas in growing demand for agricultural China seriously affected by water pollution from wastewater irrigation. As products, and readily available noted in other chapters of this report, there is a serious shortage of water wastewater supply has contributed resources in China, especially in the north. To mitigate the problem, it is to continuous increase in wastewater quite common to use wastewater or sewage for irrigation. irrigation in China for a number of Although water pollution damage to farm crops is recognized as a com- decades. Wastewater-irrigated areas mon problem in China, little research is carried out to document the prob- increased by a factor of 1.6 between 1982 and 1995. This report estimated lem (ECON 2000). Two comprehensive investigations of effects on crops in that in 2003, wastewater irrigation areas irrigated with water from sewage pipes, industrial plants, and other waste- areas totaled about 4.05 million water sources have been conducted. The wastewater/sewage irrigation zone hectares. It is estimated that the refers to farmland with an area over 20 hectares irrigated with water that either economic cost of wastewater does not meet the government standards for water quality in farmland irri- irrigation on four major crops (wheat, gation or for any other reason may lead to the death of aquatic species such corn, rice, and vegetables) in China is as fish or shrimp (GB5084-92)[2]. The first survey (MoA 1984) was con- about 7 billion RMB annually. This ducted over 20 years ago and the results are probably not a very good reflec- cost estimate was arrived at by tion of the present situation. A second survey in 1998 included two parts: accounting for the impact of polluted (1) a general survey of national wastewater irrigation areas; and (2) some data water on crop quantity and quality, on irrigation water quality, pollution conditions of the farmland, and crop including fitness for consumption as quality in representative wastewater irrigated areas (MoA 2001). Although well as the impact on the crops' nutrition quality. Economic losses due the second survey is also not very recent, in the absence of more up-to-date to ecosystem degradation and data, we decided to use it in the present evaluation. damage to human health were not In the second survey, water samples were collected for short time spans and included in this analysis, which not regularly throughout the entire growth periods. Growth periods may last means that the total economic cost several months, during which water quality may change irregularly, hence the associated with wastewater irrigation representativeness of the measurements is rather uncertain. Furthermore, since is most likely larger. the sampling locations usually were not routinely monitored sections, it is hard to derive quantitative relationships between pollutant concentrations and their effects. Due to our inability to relate damage to specific pollutant levels, we base the calculation mainly on the area that is being irrigated with wastewater, applying the results from the second survey and some other Chinese studies to 96 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N estimate the reduction in quantity and quality is expressed as a percentage of crop reduction and associated with wastewater irrigation for given percentage of production having reduced quality. crops (Gu 1984; Yang 1984; Chen 2001; Wang 2002; Gao 1997, Sun 2001; Fu 1999). Calculation Model CAUSAL AGENTS, IMPACT PATHWAY, PSI and CSMI areas in each province are AND CALCULATION MODEL obtained from the Second National Survey Report of Wastewater-irrigated Area (MoA 2001). As Causal Agents mentioned, economic losses from crop damage caused by water pollution stem from reductions The comprehensive pollution index or weighted in both quantity (reduced yield) and quality comprehensive pollution index (Pc) is usually (excess pollutants and substandard nutritional used to indicate water quality. However, due to value). Since crops with reduced nutritional limited data, dose-response functions for effects quality may or may not have excess pollutants, on crops applying these indices cannot be con- we calculated the loss due to reduced nutri- structed at present. tional quality according to equation. 5.1, i.e. The second survey of wastewater-irrigated as the mean value of the above two possibilities areas distinguished between two kinds of sewage as presented in equations.5.3 and. 5.4. The water irrigation: (1) clear water and sewage mixed effect of introducing equation 5.4 is to avoid a irrigation (CSMI), and (2) pure sewage irrigation double counting of economic loss otherwise (PSI). Generally speaking, the water used in PSI likely to occur when the percentage quantity of was of poorer quality and thus more dangerous a crop that is contaminated (contains pollutant to crops; water qualities of CSMI, although vary- levels above health guidelines) is high. The three ing considerably, were less dangerous than PSI to equations are described below. crops. In this project, we calculated damage to PSI and CSMI areas in each province are farm crops caused by both types of irrigation. obtained from the Second National Survey Report of Wastewater-irrigated Area. Losses of farm crops Impact Pathway caused by water pollution consist of three parts, which can be calculated with the following for- Use of polluted water for irrigation affects agri- mulas. Since crops with excess of pollutants may cultural production both by reducing the quan- or may not have reduced nutritional quality, we tity and the quality of output. The reduction in propose the third loss expressed in equation 5.5 quality is related to two factors: (1) an excess of as the mean value of the above two possibilities pollutants in crops, originating from heavy met- as presented in equation 5.3 and equation 5.4. als or other toxic substances in wastewater, mak- ing the crop unsuitable for human consumption; (1) Economic loss due to yield reduction and (2) substandard nutritional quality, with less protein, amino acids, Vitamin C, and other 4 (5.1) nutrients. For example, rice of poor quality pro- L1 = SiQi Pi 100 1t i =1 duced more brown and damaged grains. Wheat of poor quality produced less flour and gluten. (2) Economic loss due to excess pollutants in Vegetables of poor quality have an unpleasant crops taste and contain more nitrate and nitrite (Gao 1997; Sun 2001; Zhang 1999; Bai 1988). The 4 L2 = (1 4 damage to farm crops caused by water pollution - 1 ) 2 2 Si Qi Pi 10 (5.2) i i i i =1 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 97 C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N (3) Economic loss due to nutritional quality DOSE-RESPONSE COEFFICIENTS decline AND OTHER PARAMETERS 4 L3 = Areas in China Irrigated u (1 - 1 ) 3 3 Si i i i i =1 with Wastewater Q P 10 4 (5.3) i i A report that recorded the area irrigated with wastewater from 1949 to 1982 states that the total 4 L3 = wastewater-irrigated area was 1398.7 kha in 1982 l (1 - 1 ) (1 - 2 ) 3 3 i i i i i =1 (Dong 1985). A second report from 1996 to S Q P 10 6 1999 indicates that the total area was 3639.3 kha (5.4) i i i (with 1995 as the base year) MoA 2001). During L3 = (L3 + L3 ) the 13­14 years between the two timeperiods of 2 (5.5) u l the studies, the total area increased by 2219.4 kha, that is, by a factor of 1.6. The sewage-irrigated These three equations are put together below areas in each province from the second survey to determine the total economic loss to crops are listed in Table 5.6, and the yearly increase in from polluted water irrigation: the sewage-irrigated area is listed in Table 5.7. (Dong 1985). (4) Total economic loss from crops irrigated Table 5.7 shows that the average annual with polluted water increase in wastewater irrigation area tended to accelerate during the period from 1949 to 1982. L3 = L1 + L2 + L3 (5.6) The average increase in the area of wastewater irrigation was only 2.81kha between 1949 and where L is the economic cost of reduced agricul- 1963. A sharp increase occurred in the period tural yield and reduced crop quality caused by pol- from 1979 to 1982, reaching 355.13kha. After luted irrigation water, in tens of thousands of 1982, the rate of change slowed down. RMB; L1 is the economic loss due to yield reduc- The increase of wastewater irrigation area is tion; L2 is the economic loss due to excess pollu- related to three factors:(1) increases in planted tants in crops; L3 is the economic loss due to area, (2) limited water resources, and (3) available nutritional quality reduction; L3 is the economic u wastewater in China. Given the relative con- loss only due to nutritional quality reduction; straints or driving forces of these three factors, we L3 is the economic loss when the crop both has l assume the change in area irrigated with waste- excess levels of pollutants and decreased nutri- water in China is consistent with the Pearl growth tional levels; Pi is the market price of crop i, in function (Pearl and Reed 1920). The general RMB/kg; Si is the wastewater irrigated area of crop expression for the Pearl growth function is: i (ha), Qi is the yield per unit area of crop i in clean region, kg/ha; 1 is the fractional quantity reduc- i tion of crop i from water pollution; 2 is the frac- Y = YC 1 + EXP(a + bx) [ ] (5.7) i tional quantity of crop i that is contaminated (contains pollutant levels above health guidelines); where Y is the area irrigated with wastewater; YC 3 is the fractional quantity of crop i with nutri- i is the maximum value of the wastewater irriga- tional quality decline; 2 is the coefficient of value i tion area; and x is the year. loss of crop i due to contamination8; and 3 is the i When we performed the regression analysis coefficient of value loss of crop i due to poor qual- using the Pearl function, we first assumed a value ity, determined by the degree of quality decline. for YC, and then performed a regression using 98 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N available data. In practice, we used data from T A B L E 5 . 6 Wastewater-Irrigated Areas by Province (kha) between 1949 and 1990 as the regression sam- ple points and performed the regression using Region CSMI PSI Total different YC's. Results provided by Eview statis- tical software are given in Table 5.8. From these Beijing 0 13.60 13.6 values, we obtain the relationship, given by Tianjin 119.15 114.88 234.0 equation 5.2.9, between YC and Y1995 by regres- Hebei 96.68 18.50 115.2 Shanxi (Tai-yuan) 79.34 6.64 86.0 sion. Using the observed value for Y1995, the esti- Inner Mongolia 104.67 24.00 128.7 mated YC is obtained. Liaoning 477.1 57.90 535.0 Jilin 0.72 0.00 0.7 Heilongjiang 75.48 13.13 88.6 YC = -2959 + 1.9497 Y1995 Shanghai 14.4 0.00 14.4 Jiangsu 71.12 6.30 77.4 R2 = 0.9998 (5.8) Zhejiang 14.73 0.00 14.7 Anhui 638.19 0.27 638.5 According to the second wastewater irrigation Jiangxi 15.05 4.96 20.0 Fujian 0.55 0.05 0.6 area investigation, the Y1995 value is 3,639.3kha. Shandong 262.68 85.83 348.5 Inserted into formula 5.2.8, the appropriate Henan 670.36 45.48 715.8 value of YC is 4,136. Then the final regression Hubei 30.3 11.70 42.0 function to predict Y is presented as formula 5.9. Hunan 211.28 58.80 270.1 Guangdong 5.91 1.90 7.8 Guangxi 2.25 0.56 2.8 Hainan 0 0 0 Y = 4136 [1 + EXP(8.70963 - 0.23253X )] Sichuan 55.1 2.85 58.0 R2 = 0.96307 (5.9) Chongqing 3.04 0.07 3.1 Guizhou 5.28 0.00 5.3 Yunnan 10.01 6.30 16.3 Using the above equation, the wastewater- Shanxi (Xi-an) 119.74 33.89 153.6 irrigated area in 2003 was predicted to be Gansu 28.77 0.36 29.1 4,050 kha. Qinghai 6.66 0.00 6.7 Ningxia 4.16 2.00 6.2 Xinjiang 4.97 1.63 6.6 Total 3,127.69 511.60 3639.3 Total Sewage-irrigated Cropland (S) and Per Unit Area Yield (Q) Source: Second National Survey Report of Wastewater- irrigated Area. Because of lack of more specific data on areas planted with various crops in every wastewater irrigated region, the total area of sewage-irrigated cropland in each province was calculated based T A B L E 5 . 7 Sewage-Irrigated Area in China, 1949­1995 /kha Year 1949 1963 1972 1976 1979 1982 1990 1995 Total area 0.7 40.0 93.3 180.0 333.3 1398.7 3333 3639.3 Annual average increase 2.81 5.92 21.68 51.10 355.13 241.79 61.26 Source: Agricultural Environmental Protection Institute, Ministry of Agriculture. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 99 C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N that CSCI would not lead to a reduction in crop T A B L E 5 . 8 Regression Results with Different Yc Values yield, but, on the contrary, would increase the yields to some extent, about 450­750 kg per Yc A B Y1995 hectare. This is mainly due to the presence of nutrient elements such as N, P, K, Cu, and Zn, 3,500 8.869709 -0.25527 3312.67 which are essential to crop growth, in waste- 4,000 8.711862 -0.23506 3563.93 4,500 8.725913 -0.22765 3831.45 water. On average, the content of N in waste- 5,000 8.771965 -0.22333 4088.74 water is 15.2 mg/L; P2O5 is 2.8mg/L; and K2O 5,500 8.827719 -0.22041 4332.20 is 2.4 mg/L. Other studies, however, indicate that PSI will lead to yield reduction. According to field exper- on information from the China Agriculture Year- iments that were made in wheat lands irrigated book (China Agricultural Yearbook 2004). It was with wastewater from the Liangshui River, the assumed that the ratios of areas planted with Tonghui River and the Wanquan River by Bai wheat, rice, corn, and vegetables in wastewater- Ying et al. (1988), there were 11 cases of yield irrigated areas were the same as for the planted reduction among 15 cases, and the reduction per- areas as a whole within each province, province- centage was about 8.0­17.1 percent. Similar municipality, or autonomous region. So for each experiments were conducted in the Gaobeidian province we get: region of the Tonghui River and the Yizhuang region of the Lianghe River, where there were Si = ci St (5.10) 20 years of sewage-irrigation history. Yields of both wheat and rice grown in relatively unpol- luted soils in the sewage-irrigated area decreased where Si is the area planted with crop i in a waste- 10 percent compared to clean water-irrigated water-irrigated region, in kha; ci is the fraction of areas; yields of wheat and rice grown in polluted area planted with crop i to the total planted soil irrigated with sewage decreased even more, area in one province, province-municipality, or 40.6 percent and 39.0 percent respectively, com- autonomous region; and St is total wastewater- pared to clean irrigation areas. irrigated area in the province, kha. Chang et al. (2001) also indicates that sewage- The yield per unit planted area (Q) is the aver- irrigation can cause production reduction. Their age yield of different provinces or municipalities, study proposes expressing reductions as a function calculated from information in the China Agri- of the comprehensive water pollution index (P), culture Yearbook 2004. when P >1.0 sewage irrigation caused yield reduc- tions of 10 percent for wheat and 30 percent Identification of 1i , 2i , and 3i for rice and vegetables. The term 1 represents the percent by which the i Sewage irrigation can affect the growth of roots quantity of crop i has been reduced as a result of and seedlings in rice crops and tillers in wheat environmental pollution. crops. The height, leaf area, and dry matter can be Our determination of crop loss caused by reduced. Because of reduced leaf surface area, the wastewater irrigation was mainly based on data photosynthesis of wheat is reduced. All these fac- from field experiments, and the final estimates tors directly affect crop production. In conclu- are conservatively adjusted results of these data. sion, the negative effects on the yield of wheat and The documents of the first National Agricul- rice mainly occur as a reduction in the number of tural Environmental Quality Investigation in ears per unit area, number of seeds per ear, and Wastewater Irrigation Areas (MoA 1984) showed seed weight. The clean water in CSCI can allevi- 100 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N ate the damage, however, and generally does not T A B L E 5 . 9 Ratio of Crops Exceeding Standards from lead to large yield reductions (Bai 1988). 2nd Survey of Sewage Irrigation Based on the above considerations, we suggest using the a1 values given in Table 5.2.5 until more Percentage i information becomes available. Yield of Yield (2) Identification of 2 --percentage quantity Failing Failing i Type of Total to Meet to Meet reduction of crop i from exceedance of pollutant Sewage Reported Pollution Pollution criteria Crops Irrigation Yield /t Standards /t Standards As mentioned, the term 2 represents the i Wheat CSMI 305,466 52,218 17 percent of crop i that contains levels of pollutants PSI 291,540 84,603 29 above health guidelines, which is probably the Corn CSMI 789,766 128,608 16 most severe effect of wastewater irrigation. The PSI 242,912 43,608 18 Rice CSMI 187,480 79,442 42 accumulation of harmful pollutants in farm PSI 182,267 92,321 51 crop products renders large amounts of prod- Vegetable CSMI 188,723 22,838 12 ucts unsuitable for human consumption or PSI 160,580 42,755 27 even useless. The results of the second survey of wastewater irrigation show that the main pol- Source: Authors calculation. lutants in wastewater-irrigated areas were heavy metals, such as Hg, Cd, Pb, Cu, Cr, and As. The term a3 refers to the reduction in useful i The primary pollutants that exceed allowable yield of crop i from quality decline due to reduced thresholds in wheat are Hg, Cd, Pb, and Cu; in nutrient content rice, Hg, Cd, and Pb; in corn, Cd and Pb; and The results of both field experiments and sur- in vegetables, Hg, Cd, and As. The extent to veys show that wastewater irrigation results in which pollutants in the four crops exceed allow- more brown and damaged rice grains, some even able thresholds is shown in Table 5.9. The field with disagreeable tastes. Wastewater irrigation experiments showed that contents of NO3 and - causes low gluten in wheat and low flour pro- NO2 in vegetables from sewage-irrigated areas - duction. The results of field experiments showed were considerably higher than those prevailing that the contents of protein and amino acids in in clean water-irrigated regions (Bai 1984). wheat produced from wastewater-irrigated areas The data in Table 5.9 show that the pollution were lower than in clean-water irrigated areas. levels in 17 percent and 29 percent of the wheat Suspended substances in wastewater appar- crops in the CSPI regions and PSI regions ently affect soil porosity, lowering activity and respectively exceeded allowable thresholds; for respiration of wheat roots and leading to a lower rice, the values were 42 percent and 51 percent; protein content. Rice belongs to the helophytes; for corn, 16 percent and 18 percent; and for veg- the roots get oxygen not only from air in soil but etables, 12 percent and 27 percent. It is clear also from the atmosphere through leaves and from the table that the damage to crops is seri- stems, so protein content is not affected. ous for both CSCI and PSI. For all crops, those The effects of wastewater irrigation on pro- irrigated with PSI exceeded allowable levels of duction and quality of farm crops are summa- pollution by a far greater rate than those irri- rized in Table 5.10. gated with CSCI. In this project, we estimated only economic losses caused by yield reduction, Identification of and excess pollutants, and poor quality, and did not 2i 3i include losses due to ecological environmental The term 2 represents the price-loss coeffi- i degradation and damage to human health. cient for crops that exceed allowable pollution C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 101 C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N T A B L E 5 . 1 0 Effects of Wastewater Irrigation on Production and Quality of Farm Crops (%) Percent by Percent by Percent by Which a Which a Which a Crop is Crop is Crop is Reduced by Type of Reduced by Reduced Decline in Sewage Environmental by Excess Nutrient Major Crops Irrigation Pollution (a1 )i Pollutants (a2 ) i Content (a3 ) i Pollutants Wheat CSMI 0 17 0 Cd, Hg, Pb, PSI 10 29 10 Cu, As, Zn, F Corn CSMI 0 16 0 Cd, Hg, Cu, PSI 10 18 10 Pb, As Rice CSMI 0 42 0 Cd, Hg, Pb, Cu, PSI 20 51 5 Cr, As, F Vegetable CSMI 0 12 5 Cd, Hg, Pb, Cr, PSI 15 27 15 Cu, F, NO3 - Source: Authors calculation thresholds. If the thresholds are exceeded, the and crops that fail to meet standards of allowable crops become inedible. Wheat, rice, and corn that pollution levels, as well as deterioration of agro- fail to meet quality standards can be put into ecological environments. The effects on agro- industrial use, with a value half that of products ecological environments include soil pollution that do meet the standards, which means 2 will i of farmland, and destruction of soil structure be 0.5. However, the vegetables become waste and groups of soil microorganisms. If the agro- and 2 will be 1. ecological environment is degraded, recovery is i The terms 3 represents the price-loss coef- very difficult to achieve. This project does not i ficient for crops with reduced quality due to include agro-ecological environmental deteriora- nutrient reduction. Market prices indicate that tion; it only calculates economic losses of wheat, the price of crops of this kind is moderately corn, rice, and vegetables caused by wastewater lower than that of high quality products. We irrigation. The calculation results are shown in set 3 of wheat, corn, and rice to 0.8, and that table 5.11. Table 5.12 presents the economic i of vegetables to 0.7. mid-loss of every province caused by wastewater irrigation in 2003. Prices of agricultural products Tables 5.11 and 5.12 show that the direct According to the data on the national agricul- economic losses for crops in 2003 was about tural information website issued at the beginning 6.7 billion RMB. The losses of four crops are: of 2004, the average prices (wholesale price) of the wheat, 0.4 billion RMB; corn, 0.5 billion RMB; above four kinds of crops in 2003 were as follows: and rice, 1 billion RMB. Loss of vegetables dom- rice 1.20RMB/kg, corn 1.15RMB/kg, wheat inates with about 73.5 percent of the total. Eco- 1.14RMB/kg and vegetables 1.42RMB/kg. nomic loss caused by failure to meet pollution standards is 5.2 billion RMB, about 78.5 per- cent of the total agricultural economic loss. RESULTS By adding the loss caused by quality decline, Effects of wastewater irrigation on agriculture 85.1 percent of the total loss is obtained. The include production reduction, poor quality crops, high and low total crop losses are respectively 102 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N T A B L E 5 . 1 1 Economic Losses of Wheat, Corn, Rice, and Vegetables in 2003 (10,000 RMB) Loss Wheat Corn Rice Vegetables Total Output reduction 4,463 7,188 8,211 79,630 99,491 Excessive pollution levels 31,833 37,247 86,190 368,696 523,967 Poor nutrient quality 687 1,177 245 42,220 44,329 Total 37,099 45,729 94,729 494,861 672,419 672,419 36,983 45,613 94,646 490,546 667,787 667,787 36,867 45,496 94,562 486,231 663,155 663,155 T A B L E 5 . 1 2 Economic Mid-Loss Caused by Wastewater Irrigation by Province in 2003 Economic Loss /104 Yuan Agricultural Percentage to Output/ Agricultural Regions Wheat Corn Rice Vegetable Total 108 Yuan Output Beijing 252 442 37 14,925 15,657 88.8 1.76 Tianjin 3,496 4,982 1,339 121,851 131,669 88.2 14.93 Hebei 1,846 1,686 186 21,222 24,940 958.3 0.26 Shanxi (Tai-yuan) 629 1182 13 6,203 8,028 249.5 0.32 Inner Mongolia 271 2,441 356 4,551 7,620 336 0.23 Liaoning 109 14,678 17,816 76,397 109,000 497.3 2.19 , Ji-in 0 25 13 25 63 438.3 0.00 Heilongjiang 98 1,084 2,476 3,038 6,697 502.9 0.13 Shanghai 27 11 755 3,196 3,989 98.2 0.41 Jiangsu 764 313 4,673 10,129 15,879 981.2 0.16 Zhejiang 11 12 854 1,942 2,819 529.4 0.05 Anhui 5,051 2,464 22,642 25,791 55,947 617.9 0.91 Jiangxi 0 0 27 60 88 466.8 0.00 Fujian 2 3 2,088 1,353 3,446 383.7 0.09 Shandong 7,159 5,585 1,050 87,234 101,027 1,599.3 0.63 Henan 14,059 7,168 5,937 64,456 91,621 1,137.7 0.81 Hubei 126 139 3,133 6,064 9,462 733.4 0.13 Hunan 84 597 25,530 24,981 51,193 671.7 0.76 Guangdong 0 11 692 1,310 2,013 851.7 0.02 Guangxi 0 10 186 247 443 500.8 0.01 Hainan 0 0 0 0 0 152.7 0.00 Sichuan 8 18 119 176 321 270.1 0.01 Chongqing 302 342 2,521 3,721 6,887 804.7 0.09 Guizhou 10 39 98 169 316 275.5 0.01 Yunnan 67 198 652 934 1,852 433.9 0.04 Xizang 25.3 0.00 Shanxi (Xi-an) 2,133 1,823 1,299 8,120 13,375 334.4 0.40 Gansu 247 179 10 1,479 1,915 275.8 0.07 Qinghai 52 0 0 238 291 29.7 0.10 Ningxia 79 98 101 332 610 54.1 0.11 Xinjiang 100 81 41 400 622 482.8 0.01 Total 36,983 45,613 94,646 490,546 667,787 14,870.1 0.45 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 103 C R O P L O S S D U E T O W A S T E W A T E R I R R I G A T I O N 6.7 and 6.6 billion RMB; the gap between them based on a scientific methodology and a strict is about 1.39 percent to the medium estimation. research process. In current research literature The above data fully reflect that large amounts concerning the relationship between water of polluted crops have been produced from waste- pollution and the quantity and quality of agri- water irrigation areas every year. These inferior cultural products, methods are not identical, products constitute a great threat to food safety which causes some difficulties in the selection and human health if they enter markets and the of dose-response functions. potential negative effect on human health may Second, dose-response functions for greatly exceed the direct economic loss. Thus, wastewater irrigation are derived for a par- the phenomenon merits high attention by the ticular case, for example, a particular region, related sectors. and a given mix of pollutants, therefore the China is a country lacking freshwater resources results are not necessarily generalizable. and water for agricultural irrigation is in extreme Water pollution may damage agricultural shortage. Due to the successive droughts for many crops in several ways and systematic research years in northern China, north of the Yangtze is lacking. The effect estimates in this guide- River, wastewater has become an important line are mainly based on data and analysis of resource for agricultural irrigation and the area the second national wastewater irrigation using wastewater irrigation has continued to investigation. As there is no data on extent of expand in recent years. In China, about 480 hun- damage of agricultural crops and correspond- dred million tons of wastewater are discharged ing irrigation water quality, dose-response every year, as discussed elsewhere in this report functions have not been established between (SEPA 2004). For better utilization of water and the composite water quality index and quan- fertilizers in wastewater, it is important to seek tity and quality of agricultural crops. Instead better treatment and strengthened control of effects are estimated simply for two types of wastewater to meet the standards of water qual- irrigation: pure wastewater irrigation (PSI) ity for farm irrigation. One should also strive to and mixed clear and waste water irrigation gradually decrease wastewater-irrigated areas. (CSMI). Clearly, the uncertainties in the esti- mates are large. (2) Determination of value loss coefficients UNCERTAINTIES The -values given above were estimated The uncertainties in estimating wastewater irri- from market investigations. However, there gation effects and the economic cost of agricul- are clearly large variations in degree of damage tural production reduction due to pollution are and demand for the products of reduced mainly related to the following issues: quality so the values are highly uncertain. Uncertainty in the values is among the (1) Selection of dose-response functions. First, the main sources of uncertainty in the final appropriate dose-response function must be result. 104 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 5.3 Fishery Loss Due to the episodic nature and easily Water pollution can damage both marine fisheries and inland waters fisheries. measurable effects of acute pollution, The marine fishery zone includes coastal sea areas and offshore: the inland its impact on fisheries is much better waters fishery includes cultivation in rivers, lakes, estuaries, and reservoirs. understood than the effects of chronic These fisheries can suffer both acute and chronic toxic effects of pollution. In pollution. Analyses conducted by the marine waters, acute toxicity mainly refers to the extensive death of aquatic Ministry of Agriculture and SEPA animals and plants caused by red tide (see box 5.3). In inland waters, acute estimated that fishery losses due to toxicity is generally triggered by excessive discharge of high concentration pol- acute pollution accidents in 2003 amounted to over 4.3 billion RMB, lutants in inland waters. Chronic toxicity of fisheries in both cases results including 713 million RMB in direct from the long-term accumulation of pollutants and mutagenic substances in economic losses and more than the water bodies. This pollution-related damage to the fisheries results in a 3.6 million RMB in indirect losses loss of production, which can be described in terms of direct and indirect eco- (MoA and SEPA 2003). While not nomic loss. Direct economic loss means that pollution sources contaminate insignificant, these figures may the fishing zone, killing or damaging valuable and/or rare and endangered greatly underestimate the total aquatic wildlife such as fish, shrimp, crab, shellfish, and algae. Indirect eco- economic cost of fishery loss due nomic loss refers to the possible loss of fishery production caused by reduced to pollution. First, chronic pollution reutilization of natural fishery resources, reduced capacity to reproduce, and costs are likely to be much higher decreased breeding grounds. The direct and indirect losses have been esti- than the acute. Secondly, the mated by the fishery supervision and management agency (MoA 1996). methodology employed in these Acute pollution accidents may cause a high death rate, which tends to studies--such as the application of a rule that stipulates that the attract public attention. When large numbers of aquatic animals and plants indirect costs cannot be higher than die over a short time, the ensuing economic loss can be more easily measured three times the direct--may further and calculated. However, the fish morbidity rate resulting from chronic dam- underestimate the true cost of age caused by prolonged exposure to polluted water may be more serious, pollution. with the extent of damage depending on the degree of pollution. Because chronic damage takes place over a long period of time and is not easily observed, it tends to be ignored. Furthermore, the lack of systematic research on the exposure­response relationship for aquatic animals and plants in the polluted water body makes it difficult to evaluate the loss due to chronic dam- age. With respect to both the quality and quantity of damage, chronic water pollution may result in greater losses than those caused by acute toxicity. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 105 F I S H E R Y L O S S B O X 5 . 3 Background Knowledge Red Tide and Its Damage Mechanisms Red tide occurs when environmental conditions in the ocean change and cause a bloom of phyto- plankton algae, which in turn changes the color of the water. Not all red tide is red. The actual color of waters experiencing red tide depends on the species of algae that blooms. Red tide can have either natural or anthropogenic origins. Natural causes include shifts in climatic factors, sea temperature, salinity, and seawater exchange. Human-induced red tide generally stems from marine aquaculture and pollution. Red tide causes severe damage to the marine ecosystem and even endangers human health through five damage mechanisms: (1) Mucus excretion. The algae excrete mucus, which adheres to the gills of marine animals, impeding respiration and causing them to die by suffocation. (2) Chemical excretion. The algae excrete chemical substances (such as ammonia, hydrogen sulfide) that harm the water body and poison other organisms. (3) Toxin production. The algae produce toxins that directly poison cultivated animals and plants and/or are transferred through the food chain to damage human health by poisoning. (4) Oxygen depletion. The algae use the oxygen in the water body or cause water to contain a great deal of hydrogen sulfide and methane so that cultivated organisms may die from oxygen depletion or poisoning. (5) Absorption of solar rays. The algae absorb sunshine and shade the sea surface, causing aquatic plants to die from insufficient sunshine, which could further reduce fish populations. Chronic pollution not only directly causes exces- research on the exposure­response relationship sive levels of pollutants in aquatic animals and between pollution severity of the water body and plants, but may also cause changes in the bio- fish growth and reproduction, chronic damage logical community affecting the ecological bal- still cannot be estimated. Therefore, fishery loss ance of the whole water body. Furthermore, it in the following only refers to fishery damage may bring about so-called secondary pollution caused by acute pollution accidents. loss problems when humans eat the aquatic In addition, it should be noted that overfish- products with pollutants exceeding standards. ing and poor fisheries management practices, as To quantitatively evaluate pollution-induced well as irrational water resource development and fishery loss, the Chinese Ministry of Agriculture utilization, can also affect the sustainability of issued the Regulations on Calculation Method of fish resources and change ecological environments Fishery Loss Caused by Pollution Accidents in Water and migration routes for fishes. In some cases, Area in 1996. The calculation method takes into aquatic animals and plants may be pushed toward account the type of water (marine or inland), extinction. Such losses are caused by poor man- hydrographic conditions, the size of the polluted agement and are not evaluated in this project. area, and the type of damaged resources. The reg- ulations provide a basis for scientific and rational ESTIMATED FISHERY LOSS DUE TO calculation of fishery loss caused by pollution acci- ACUTE POLLUTION EPISODES dents, as well as guidelines for how to handle such accidents. The China Fishery Ecological Environ- Estimation Method mental Condition Bulletin, issued jointly by the Ministry of Agriculture and the State Environ- Direct economic loss includes loss of aquatic mental Protection Administration each year, products, loss of additional pollution protection publishes the fishery loss caused by pollution facilities, loss of fishing gear, pollution removal accidents as calculated according to the regula- cost, and actual cost of evidence collection and tions. But as described above, due to the lack of identification by monitoring departments. The 106 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N F I S H E R Y L O S S aquatic product loss is quantified according to the still seriously polluted by nitrogen, phosphorus, market retail price provided by the local business petroleum, and some heavy metals. In 2003, administration department. Loss of aquatic prod- fishery pollution accidents occurred, resulting in ucts includes the quantity of fish killed by expo- fishery losses totaling over 4.3 billion RMB-- sure to toxic pollutants, the quantity that has including 713 million RMB in direct economic apparent toxic symptoms but still can survive, losses and more than 3.6 billion in indirect losses and the quantity that has been rendered inedible related to measurable damage to natural fishery due to pollution. The loss includes both finished resources, of which 896 million RMB was from products and semi-finished products, as well as inland waters and 2.7 billion RMB from natural loss of offspring. The quantified loss is expressed ocean fishery resources. in terms of loss of finished products, with con- Due to the difficulties in estimating the fish- version factors for offspring and semi-finished ery loss from the chronic effects from pollution, products determined by the Fishery Supervision the stated loss by the Bulletin may be only a small and Management Agency according to different fraction of the total loss, since chronic effects are species and actual local situation. For fish culti- believed to be more serious than acute effects. vation in cages or in rice fields, loss is quantified Another big estimation bias is from the method as follows: to evaluate the indirect loss. The regulation that the economic loss for natural fishery resources Loss of aquatic product = shall not be less than three times the direct eco- local market price × loss quuantity nomic loss is a rather subjective judgment. The cost of death of a parent strain and original Endnotes seed of cultivation species shall be set 50 to 1. Where no reference is given, data in this chapter is from 500 percent higher than the general commodity personal communications with staff at the Ministry of price, depending on its degree of importance. Water Resources, China. The specific price is determined by the Fishery 2. The six are Beijing, Tianjin, Hebei, Shanxi, Shanghai Supervision and Management Agency. and Ningxia. 3. Lioaning, Jiangsu, Shandong, Henan and Gansu. The costs of pollution protection facilities, 4. Data from 2005 indicate that even Shandong is below 500, fishing equipment, pollution removal, and evi- while Gansu is above 1000. (NBS, 2006). 2005 was also dence collection and identification by the Fish- an average year in terms of nationwide water resources. ery Supervision and Management Agency are 5. Worse than class IV in the groundwater standard (GB/T14848-93). calculated according to actual expenditures. 6. Worse than class III in the same standard. Economic loss for natural fishery resources is 7. The range 2.1­5.2 produced by He and Chen-- determined in accordance with the local resource depending on river basin--may also be compared with situation by the Fishery Supervision and Manage- values used by the World Bank (2001)--depending on final consumption. The World Bank (2001) values range ment Agency, but shall not be less than three times from 0.8 in agriculture to 6 in urban industry. the direct economic loss obtained from the esti- 8. Because not all substandard crops are discarded but used mated reduction in amount of aquatic products. for other purposes with lower quality criteria such as fod- der and industrial raw materials, this factor is introduced. Estimation Result References According to the China Fishery Ecological Envi- ronmental Condition Bulletin, the overall ecolog- Allan, J. A. 1993. "Fortunately there are substitutes for water: Otherwise our hydro-political futures would be impos- ical situation for fisheries in China was good in sible." Priorities for Water Resources Allocation and Man- 2003, but at the local level, some waters were agement. London: ODA. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 107 F I S H E R Y L O S S Bai, Ying. 1984. "Effects of sewage irrigation on farm crop China." Philosophical Transactions of the Royal Society quality." Agricultural environmental protection 1: 11­13. 361: 835­842. Bai, Ying et al. 1988. Pollution of irrigation water and its Mai, Yongbin et al. 1984. "Investigation of agro-environ- effects. Beijing: Beijing Agriculture University Press. mental quality in major wastewater irrigated areas in Berkoff, J. 2005. "China: The South-North water transfer China." Agricultural environmental protection 5: 1­4. project­is it justified?" Water Policy 5: 1­29. Ministry of Agriculture (MoA). 1984. National Agricultural Cai, Siyi et al. 1994. "Wastewater treatment and agricultural Environmental Quality Investigation in Wastewater utilization." Agricultural environmental protection 13(1): Irrigation Areas. Beijing: Agricultural Environmental Chen, Zhujun et al. 2001. "Application of wastewater irri- Protection Institute, Ministry of Agriculture. gation in agriculture in Israel." Agro-Environmental Pro- Ministry of Agriculture (MoA). 1996. "Regulations on Cal- tection 20(6): 462­464. culation Method of Fishery Loss Caused by Pollution China Agriculture Yearbook Compilation Committee. Accidents in Water Area, 1996." 2004. China Agriculture Yearbook 2004. Beijing: China Ministry of Agriculture (MoA). 2001. Second National Sur- Agriculture Press. vey Report of Wastewater-irrigated Area Beijing: Agricul- Chinese Academy of Environmental Planning (CAEP). tural Environmental Protection Institute, Ministry of 2006. Surveying data of enterprises the National survey Agriculture. of Green Accounting and Pollution cost valuation in ten Ministry of Agriculture and State Environmental Protection candidate provinces. Beijing: Chinese Academy of Envi- Administration (MoA and SEPA). 2003. China Fishery ronmental Planning. Ecological Environmental Condition Bulletin. Chang, Y., H. M. Seip, and H. Vennemo. 2001. "The envi- Ministry of Water Resources (MWR) 2004. 2004 Statistical ronmental cost of water pollution in Chongqing, China." Bulletin on the National Water Activities. Beijing: Ministry Environment and Development Economics 6: 313­333. of Water Resources. Available at: http://www.mwr.gov. China Daily. 2004. "Beijing may tap Yangtze water by the end cn/english1/pdf/2004Bulletin.pdf of the decade." China Daily, November 17. Available at: Ministry of Water Resources (MWR). 2005a. The Survey www.mwr.gov.cn/english/20041117/42972.asp. and Assessment of Water Resources and Utilization in China. Dong, Keyu et al. 1985. Effects of wastewater irrigation on Beijing: General Institute of Water Resource Planning, agro-environment and countermeasures Beijing: Beijing Ministry of Water Resources. Agricultural Environment Protection Institute. Ministry of Water Resources (MWR). 2005b. 2004­2005 ECON. "An environmental cost model, 2000." Oslo (under Annual Report. Beijing: Ministry of Water Resources. contract with the World Bank) Available at: http://www.mwr.gov.cn/english1/2006 Falkenmark, M. 1997. "Society's interaction with the water 0404/20060404085729BEKKEQ.doc cycle: A conceptual framework for a more holistic Ministry of Water Resources (MWR). 2006. "Water prices approach. Hydrological Sciences Journal 42: 451­66. likely to be raised again. " Beijing, April 25. Available at: Foster, S., H. Garduño, R. Evans, D. Olson, Y. Tian, www.mwr.gov.cn/english/20060425/70858.asp. W. Zhang, and Z. Han. 2004. "Quaternary aquifer of National Bureau of Statistics (NBS). 2004. China Statistical the North China Plain­assessing and achieving ground- Yearbook 2004. Beijing: National Bureau of Statistics. water resource sustainability." Hydrogeology Journal National Bureau of Statistics (NBS). 2006. China Statistical 12: 81­93. Yearbook 2006. Beijing: National Bureau of Statistics. Fu, Yuhua et al. 1999. "Investigation on vegetable pollution Pearl, R., and L. J. Reed. 1920. "On the rate of growth of in suburbs of Shenyang city." Agricultural environmental the population of the United States since 1870 and its protection 18(1): 36­37 mathematical representation." Proceedings of the National Gao, Xiali et al. 1997. "Study on effects of wastewater Academy of Sciences 6: 275­288. irrigation on physiology and cells of vegetable." Chinese Qu, Ziqing. 1989. "Construction of wastewater-land treat- Environmental Science 17(5): 443­445. ment system based on present situation of wastewater Gu, Fangqiao, et al. 1984. "Discussion on development irrigation in China." Agricultural environmental protec- approach of wastewater irrigation." Agro-Environmental tion 8 (1). Protection 1: 1­3 State Environmental Protection Administration (SEPA). He, Jing and Chen Xikang. 2005. "Calculation of Chinese 2004. China environment statistics annals 2003. Beijing: Shadow price of Water Resources Based on Dynamic State Environmental Protection Administration. Computable Equilibrium Models." Systems Engineering- Sun, Hua et al. 2001. "Appraisal on heavy metal pollution theory & Practice 25(5):49­54. of paddy field irrigated by wastewater in Guixi city, Liu, C. and X. He. 1996. Projection of Water for the 21st Jiangxi province." Agricultural environmental protection Century in China. Beijing: Science Press. 20(6): 405­407. Ma, J., A. Y. Hoekstra, H. Wang, A. K. Chapagain, and D. Vörösmarty, C. J. et al. 2006. "Fresh water." In Hassan, R., Wang. 2006. "Virtual versus real water transfers within R. Sholes, and N. Ash, eds. Ecosystems and human well- 108 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N F I S H E R Y L O S S being. Findings of the Condition and Trends Working Yang, H., and A. Zehnder. 2001. "China's regional water Group of the Millennium Ecosystem Assessment, scarcity and implications for grain supply and trade." vol 1. Washington, DC: Island Press. Available at: Environment and Planning A 33: 79­95. http://www.millenniumassessment.org/en/Products. Yang, Dingguo. 1984. "Wastewater irrigation and agro- Global.Condition.aspx environmental pollution." Agro-Environmental Protection Wang, Qingren et al. 2002. "Study on situation and reason 5: 2­4 of soil pollution by heavy metals in several mining areas Yang, H., X. Zhang, and A. Zehnder. 2003. "Water and sewage irrigated areas." Journal of Environmental scarcity, pricing mechanism and institutional reform in Science. 22(3): 354­358. northern China irrigated agriculture." Agricultural World Bank. 2001. China: Agenda for water sector strategy Water Management 61: 143­161. for North China. Report prepared for the World Bank, Zhang, Naiming et al. 1999. "Ecological effects of waste- Washington, D.C. (Sinclair Knight Merz & Egis Con- water irrigation and its profit and loss analysis." Agri- sulting, Australia; General Institute of Water Resources cultural environmental protection 18 (4). & Hydropower Research, Beijing; Institute of Hydrology Zhou, Y., and R. Tol. 2003. "Implications of desalination and Water Resources, Nanjing; and Chinese Research to water resources in China­an economic perspective." Academy for Environmental Sciences, Beijing.) World Working paper FNU-22. Hamburg, Germany: Ham- Bank Report No. 22040-CHA. Four volumes. Wash- burg University. ington, DC: World Bank. Zhu, R. 2006. "China's South-North water transfer project World Meteorological Organization. 1997. Comprehensive and its impacts on its economic and social development." assessment of the freshwater resources of the world. Geneva: Beijing: Ministry of Water Resources. Available at: http:// World Meteorological Organization. www.mwr.gov.cn/english1/20060110/20060110104100 Xia, Guang. 1998. An Economic Estimation for Environ- XDENTE.pdf mental Pollution Losses in China. Beijing: China Envi- Specification of agricultural irrigation criterion. Beijing: China ronmental Science Press. Agricultural Science and Technology Press. 1997. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 109 6 NON-HEALTH IMPACTS OF AIR POLLUTION 6.1 The Impact of Acid Rain and SO2 on Crop Loss Dose-response functions estimated Vegetation damages may be caused by direct exposure to gaseous or particu- from studies in the 1980s and 1990s late air pollutants or indirectly through soil acidification. Direct damage from for different crops and regions within SO2 emissions is very likely in some regions. Other possible causes of damage China show that crop losses due to include high concentrations of ozone and other photo-oxidants and, in some SO2 and acid rain accounted for areas, hydrogen fluoride in the air. In many countries, ozone is the main dam- about 30 billion RMB in 2003. About aging agent; this is an increasing factor also in China (Aunan et al. 2000; 80 percent of this economic loss was Wang and Mauzerall 2004). However, due to scarcity of monitoring stations, associated with lower vegetable especially outside cities, ozone damage is not considered in this report. yields, and nearly half of the total Indirect effects due to soil acidification may be from elevated levels of toxic cost was incurred by only three aluminum in soil water, increased leaching of plant nutrients (particularly provinces--Hebei, Hunan, and Shandong. This study did not magnesium) from soils, or reduced availability of phosphorus. Acidic mist or estimate the cost of acid rain on acidic cloud water can reduce tolerance of certain species to cold. In most pris- forests due to the lack of reliable tine forests, increased deposition of nitrogen will increase growth rates, but exposure-response functions and in if nitrate deposition becomes too high it may result in damage due to soil order to avoid attributing a cost, acidification, lack of other nutrients, or increased sensitivity to other stress which if based on timber loss alone factors. In many cases, vegetation damage is likely a combined effect of would significantly underestimate the anthropogenic and natural stressors (e.g., drought, frost, and pests). true cost. In this report, only effects of SO2 and acid rain are considered. There are large uncertainties in estimates of vegetation effects. Here we will propose a procedure for estimating effects on crops mainly based on Chinese studies. We warn against quantification of effects on forests at the present state of knowledge, but in Appendix X (to be included) we give tentative equations based on Chinese studies. CAUSAL AGENTS AND IMPACT PATHWAY There are many studies on effects of SO2 and acid rain on crops. However, the results are sometimes in conflict, and establishing dose-response func- tions--especially for acid rain--is difficult. A recent study showing clear effects of acid rain has been carried out in India (Singh and Agrawal 2004). However, in this report we will use results of Chinese studies. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 113 T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S Acid rain and SO2 damage to crops can be regime mimicked the conditions for the growth divided into acute injury and chronic injury. periods in 1955­85 for Nanning City in Guangxi Acute injury implies that the leaves get clear signs Province--thatis,usingaveragerainfalleachmonth of injury within a short time due to contact with and average number of days with rainfall greater acid rain or SO2. This kind of injury generally than 10mm (moderate rain). appears when the pollution levels are very high. The experimental outputs of crops cultivated Long-term exposure to lower pollution levels may in pots exposed to SO2 and acid rain are shown cause chronic injury, for example due to changes in Table 6.1. in chlorophyll or pigment. This destroys the nor- Based on the above data, we obtained the mal activity of the cells, causing cell death and/or dose-response relationships between SO2/acid symptoms such as early loss of leaves. rain pollution and crop yields (see table 6.2). In addition, though pollutants may affect soil Limits in SO2 concentration and pH identify conditions, e.g. cause soil acidification, it is not the type of pollution in the area. When [SO2] clear how much such indirect effects may reduce 0.04mg/m3 and pH 5.0, the crops are under yields. So, the possible indirect loss through soil combined acid rain/SO2 pollution stress. For changes caused by acid rain is not included in loss estimation, we suggest the set of functions this ECM-edition. in the right column of Table 6.2. When [SO2] 0.04mg/m3 and pH > 5.0, only SO2 has an effect Dose-Response Relationships on crops, and the functions in the left column of There were some studies in China on dose- Table 6.2 can be used. When [SO2] < 0.04mg/m3 response relationships between SO2/acid rain pol- and pH 5.0, the crop lossesare only from acid lution and crop yields in the 1980s and 1990s. In rain and can be estimated with the functions in one important experiment, crops cultivated in the middle column of Table 6.2. pots were exposed to various levels of SO2 and acid rain pollution. The acid rain used in the experi- Valuation Model ment simulated as far as possible natural acid rain n in South China in concentrations of SO4 , NO3 , 2- - Cac = a PSiQ0 100 (6.1) i i i Ca2 , Mg2 , K+, Na+, and Cl-. The weight ratio + + i =1 SO4 /NO3 was 9:1. The ratio is likely to be lower 2- - where: Cac--Economic cost of crop yield reduc- now (see Annex A for discussion). The watering tion caused by air pollution, 10,000 RMB; T A B L E 6 . 1 Yields of Crops Cultivated in Pots Exposed to SO2 and Acid Rain Crops SO2 Only Acid Rain Only SO2 and Acid Rain Rice Y = 26.01 - 2.85X1 -- Y = 26.61 - 4.82 X1 + 0.049 X2 Wheat Y = 23.52 - 6.33 X1 Y = 17.20 + 1.17 X2 Y = 17.84 - 7.14 X1 + 1.04 X2 Barley Y = 34.11 - 12.22 X1 Y = 27.29 + 1.55 X2 Y = 25.84-15.51 X1 + 1.53 X2 Cotton Y = 30.60 - 7.70 X1 Y = 24.07 + 1.26 X2 Y = 22.15-8.84 X1 + 1.62 X2 Soybeans Y = 40.82 - 11.75 X1 Y = 34.68 + 1.12 X2 Y = 30.57-13.24 X1 + 1.95 X2 Rape Y = 31.12 - 15.81 X1 Y = 16.85 + 2.71 X2 Y = 19.4-13.02 X1 + 1.83 X2 Carrots Y = 105.58 - 56.97 X1 Y = 54.96 + 9.67 X2 Y = 71.03-41.82 X1 + 5.22 X2 Tomatoes Y = 92.70 - 34.67 X1 Y = 72.82 + 3.78 X2 Y = 72.95-31.96 X1 + 2.60 X2 Kidney beans Y = 43.69 - 30.14 X1 Y = 9.00 + 6.39 X2 Y = 22.90-30.11 X1 + 3.01 X2 Source: Authors Calculations. Note: Y--Crop yield; X1--SO2 concentration in mg/m3; X2--pH value. 114 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S T A B L E 6 . 2 Dose-Response Relationship Between SO2/Acid Rain Pollution and Crop Yield Applied in ECM Percentage Yield Reduction (%) Pollution by SO2 Pollution by Acid Rain Combined Pollution of SO2 and Crops (mg/m3) (pH value) Acid Rain (mg/m3), pH Value Rice 0.1096 X1 0.0292 + 0.1793 X1 - 0.00182 X2 Wheat 0.2691 X1 0.2759 - 0.0493 X2 0.2461 + 0.3017 X1 - 0.043949 X2 Barley 0.3583 X1 0.2413 - 0.0431 X2 0.249 + 0.4508 X1 - 0.044466 X2 Cotton 0.2516 X1 0.2267 - 0.0405 X2 0.2906 + 0.2831 X1 - 0.051886 X2 Soybeans 0.2878 X1 0.1532 - 0.0273 X2 0.2632 + 0.3191 X1 - 0.047 X2 Cole 0.508 X1 0.4739 - 0.0846 X2 0.3457 + 0.4392 X1 - 0.061724 X2 Carrots 0.5396 X1 0.4963 - 0.0886 X2 0.2916 + 0.4171 X1 - 0.052064 X2 Tomatoes 0.374 X1 0.2252 - 0.0402 X2 0.1664 + 0.3652 X1 - 0.029711 X2 Kidney beans 0.6899 X1 0.799 - 0.1427 X2 0.424 + 0.7574 X1 - 0.075712 X2 Vegetables 0.5345 X1 0.481 - 0.0905 X2 0.294 + 0.5132 X1 - 0.0525 X2 Source: Authors Calculations. Note: X1--concentration of SO2, X2--pH value; The coefficients in the dose-response relationships for vegetables are average values for carrots, tomatoes, and kidney beans. Pi --Price of crop i, RMB/kg; ai --Reduction rate of crop i due to pollution Si --Planted area of crop i, 104 ha; (exposure-response relation), %; Q0 --Output per unit area of crop i in clean i n--Number of crop types, n = 6. region, kg/ha; Parameter Sources T A B L E 6 . 3 Parameters Used in Valuation Model for Crop Reduction by Acid Rain/SO2 Pollution and Their Sources Parameters Unit Data Sources Geographical Resolution Si: Planted area of crop i 104Mu Agricultural Statistics Yearbook City Qi: Production of unit kg/Mu Agricultural Statistics Yearbook Province area of crop i in clean region Pi: Price of crop i RMB/kg Agricultural StatisticsYearbook Nation i: Reduction rate of % Dose-response functions, Table 6.2 City crop i due to pollution, a = f (X1, X2) X1: Concentration of SO2 mg/m3 Environmental monitoring data City in planted areas X2: PH of rain in planted Environmental monitoring data City areas Source: Authors Calculations. Note: 1Mu = 1/15 ha C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 115 T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S Valuation Result for carrots, tomatoes, and kidney beans. If yield data can be obtained, economic losses of carrots, According to the statistics yearbooks in 2004 tomatoes, and kidney beans can be calculated of all provinces, there are only 16 provinces, separately. autonomous regions, and provincial cities for In the evaluation model, crop loss is calcu- which data of different agricultural products are lated as a percentage reduction in yield per unit given so that losses can be calculated at the city area attributable to the pollution. We apply level. In other areas, data for different agricul- the present yield per unit area to calculate a tural products are only available for the whole "hypothetical yield" under clean conditions in province. This implies that we can only calculate the given area, thus enabling an estimation pollution losses in these areas on the province of the crop loss. However, several factors influ- level. Annual average SO2 concentrations and ence the yield, such as land fertility and climate. annual pH values were calculated from values for These factors may enhance or reduce the effects cities. Using province-wide average concentra- of pollution, and the estimate of the hypotheti- tions of SO2 and pH may disguise crop damage cal yield under clean conditions thus becomes in pollution-intensive parts of a province and very uncertain. lead to underestimation of crop damage. We cor- Other main uncertainties in the results rect for a possible underestimation by increasing obtained by the described model originate from the pH limit in province-wide data from 5.0 to the following aspects: 5.6. It is generally agreed that above 5.6 no yield reduction occurs, but between 5.0 and 5.6 the 1) The dose-response relationships in the pro- functions of Table 6.2 usually show some dam- posed model have been obtained from pot age. If the suggested method yields negative eco- experiments, not from field studies. Results nomic losses, these are set to zero. The results are from other studies show quite large variations given in Table 6.4. and different results for different cultivars. The results show that the economic losses in The pot experiments simulated climatic con- agriculture caused by SO2 and acid rain pollu- ditions of South China, and there are espe- tion of China in 2003 were about 30 billion cially large uncertainties implied when the Yuan. About 80 percent of the total loss is due relationships are applied to estimate crop loss to impacts on vegetables. By region, 21 per- in the northern provinces. cent of the total loss is from crop loss in Hebei, 2) While most crops are grown in rural areas, 12 percent in Hunan, and 11 percent in monitoring data are generally only available Shandong. for urban areas. Using data for SO2 concen- trations and pH from urban areas will likely lead to overestimated crop losses. The error is UNCERTAINTIES likely to be most severe for SO2, which varies Although there are dose-response relationships more than pH with the distance from the for carrots, tomatoes, and kidney beans, only the cities. Limited monitoring data are available aggregate total output of vegetables is given for rural areas in China. However, there are in the agricultural statistical yearbooks. In the reasons to believe that the level of SO2 is sub- model, we therefore derive a dose-response func- stantially enhanced in large rural areas due to tion for vegetables from the arithmetic means of extensive use of coal in town and village the coefficients in the dose-response functions enterprises and as a main household fuel for 116 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S T A B L E 6 . 4 Crop Losses Caused by SO2 and Acid Rain Pollution of China, 2003 (10,000 Yuan) Economic Losses of Crops Total Economic Regions Rice Wheat Rape Seed Cotton Soybeans Vegetables Losses Beijing 8 341 0 78 162 24,060 24,649 Tianjin 55 807 0 2,633 391 34,413 38,299 Hebei 777 47,120 0 26,428 7,573 554,168 636,067 Shanxi/t 28 13,179 270 5,841 6,999 141,786 168,102 Neimeng 266 1,078 1,367 89 3,576 23,837 30,214 Liaoning 1,061 84 4,737 42 2,708 58,876 67,508 Jilin 233 0 0 0 1,485 8,302 10,021 Heilongjiang 1,173 9 0 0 3,943 10,016 15,142 Shanghai 465 97 306 18 80 15,030 15,997 Jiangsu 0 1,700 1,621 904 503 0 4,727 Zhejiang 0 1,226 8,218 1,346 3,594 178,582 192,966 Anhui 0 56 1,749 1,467 126 4,712 8,110 Fujian 0 34 73 0 476 0 583 Jiangxi 4,715 117 5,506 1,029 1,424 66,710 79,501 Shandong 385 17,024 117 11,853 2,201 309,677 341,257 Henan 335 29,423 1,919 5,282 2,601 155,366 194,926 Hubei 10,784 6,045 11,969 6,622 2,195 99,042 136,658 Hunan 79,245 982 16,609 13,955 9,230 241,061 361,082 Guangdong 0 24 55 0 490 568 Guangxi 38,553 33 362 18 2,031 69,310 110,307 Hainan 0 0 0 0 0 0 0 Chongqing 24,316 5,832 5,914 0 6,155 111,780 153,997 Sichuan 29,068 15,945 19,004 413 9,454 160,822 234,706 Guizhou 4,530 1,697 5,769 16 2,495 40,351 54,858 Yunnan 0 0 0 0 0 0 0 Xizang 0 0 0 0 0 0 0 Shanxi/x 155 14,732 1,719 4,483 1,312 41,149 63,550 Gansu 24 4,267 1,403 1,678 1,704 33,621 42,697 Qinghai 0 0 0 0 0 0 0 Ningxia 307 1,446 13 0 458 8,271 10,496 Xinjiang 38 649 13 516 61 2,852 4,129 TOTAL 196,521 163,945 88,714 84,713 73,429 2,393,794 3,001,117 Source: Authors Calculations. many people. In northern China, the fact FOREST DAMAGE that we use annual values instead of averages Studies in Europe for the growing season is likely to lead to and the United States overestimated crop losses because the pollu- tion level usually is higher during winter. Intensive research on possible effects of acidic 3) When lacking yield data on municipal levels, deposition (and its precursors) on forests have the use of provincial averages combined with been carried out over the last two to three decades environmental monitoring data on city levels both in Europe (UN/EC 2004) and the United introduces uncertainties. States (NAPAP 1998; Driscoll 2001). Menz and C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 117 T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S Seip (2004) give a short overview. Nonetheless, ducting quantitative research on the impacts in quantitative relationships between primary pollu- 11 provinces (Feng Zongwei et al. 1999; Chen tants and forest damage have been difficult to Chuying et al. 1993; Cao Hongfa et al. 1993). obtain. In Europe, assessment and monitoring of These studies revealed that acid rain and SO2 effects of air pollution on forests have been carried had obvious impacts on coniferous forest out in a joint UN-EC program since the late (mainly masson pine and Chinese fir forest). For 1980s (UN/EC 2004). Except for some areas in nine provinces, the volume loss rate attributable Eastern Europe, where direct effects of SO2 prob- to the impact of acid rain and SO2 on masson ably have played an important role in causing for- pine and Chinese fir was estimated by means of est defoliation, there are no clear long-term trends a multi-factor analysis of altitude, slope location, that can be related to acidic deposition. Fortu- slope, slope orientation, soil thickness, thickness nately, the dramatic forest dieback feared by some of black soil, SO2 (average and daily value), and scientists in the 1980s never materialized. Recent acid rain (average pH) (see table 6.5). These vol- improvements in tree vitality in some areas in ume loss rates were obtained in specific Eastern Europe--for example, in Poland--have provinces at certain times and under certain con- been partly ascribed to decreased pollution. To ditions of acid rain and SO2 pollution, and are date, investigations of possible effects of acidic not generally applicable. Using these volume loss deposition on forests in the northeastern United rates, the forest loss in one province in 2003 was States and in Canada have focused on red spruce estimated (see Annex). and sugar maple. There is evidence that acidic deposition has caused dieback of red spruce by decreasing their tolerance to cold (Driscoll et al. Uncertainties 2001). Damage to sugar maples may in some In spite of extensive studies, reliable relationships localized areas be caused, at least partially, by loss between forest growth and SO2 concentrations or of base cations (Ca2 , Mg2 ) from the soil. + + precipitation pH have not been established in In spite of considerable defoliation in some Europe or the United States. In the cost-benefit areas, European forests grow well. The European analysis for the Protocol to Abate Acidification, report (UN-EC, 2004) states: "Forest growth has Eutrophication, and Ground-Level Ozone in Europe increased across Europe. This means that today (Holland et al. 1999), effects on forests (timber in general both healthy and defoliated trees show production) were only estimated for ozone; the larger increments. The absolute growth level of effects of other pollutants were considered too the defoliated trees is, however, lower. Under cer- uncertain. tain stand and site conditions, nitrogen deposi- Although exposure-response functions have tion can contribute to this growth change, but been suggested in China based on Chinese also increasing temperature and carbon dioxide studies (see Appendix X), they are only tenta- concentration can have stimulating effects. It tive. The pH relationships are probably the has to be clarified whether this increased forest least reliable. The relationships are given sepa- growth leads to improved forest condition and rately for various provinces. This may in some functioning in the long term." cases reflect differences in environmental con- ditions, such as soils, but it is far from a satis- factory solution. Furthermore, the studies were Studies in China carried out more than a decade ago. At least in China started its research on the impact of acid some regions, there has been an increase in the rain on forests during the 1980s and 1990s, con- nitrate/sulfate ratio in precipitation. Tu et al. 118 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S T A B L E 6 . 5 Annual Average Timber Stocks Applied in Tree-by-Tree Investigation and Calculated Reduction Rates Reduction Rates % Annual Average Timber Stocks (baseline)(m3/ha) Masson Pine Chinese Fir Areas Masson Pine Chinese Fir SO2 Acid Rain SO2 Acid Rain JiangSu 5.75 6.76 8.45 5.20 5.73 4.77 ZheJiang 6.30 10.70 9.88 10.32 8.70 10.70 FuJian 8.49 8.33 3.04 4.86 1.74 5.36 JiangXi 6.30 6.35 4.40 2.60 6.19 5.91 AnHui 4.60 5.60 4.54 10.16 6.44 9.46 HuNan 4.80 5.40 4.01 6.29 6.56 11.74 HuBei 4.40 4.50 9.36 3.54 5.85 8.45 SiChuan* 4.74 3.11 16.68 30.20 GuiZhou* 5.77 5.07 9.38 14.20 Source: Feng Zongwei et al., 1999; *Chen Chuying et al., 1993. Note: *indicates data from 1984­86; others are data from 1992­93. (2005) found that the ratio between the nitrate Economic Evaluation and sulfate contributions to acidification in of Forest Damages precipitation in the Nanjing area increased In addition to possible loss of timber caused by from 0.1 in 1992 to 0.3 in 2003. Tang et al. SO2 and acid rain (see Appendix), forest damage (2004) report significant nitrate concentra- entails a number of other effects such as loss of tions. Since nitrogen is an important nutrient, non-timber forest products, carbon sequestration, nitrate deposition may increase growth. watershed protection, and recreation. The value of In conclusion, the present basis for estimat- these products is likely to be high. In a case study, ing forest damage caused by air pollution in Zhang (2001) estimated the total value to be China is not satisfactory. One reason is the lack 10 times the timber loss, but with very large uncer- of monitoring data in more remote areas (see tainty. Mahapatra and Tewari (2005) found the section on effects on crops). However, more net present value of non-timber forest products to studies to obtain more reliable exposure- be four to five times greater than potential timber response functions are needed. Acidity (pH) in revenue for two studied sites in India. The recent precipitation and SO2 concentrations in air are Millennium Assessment Report (MA 2005) com- not sufficient to determine possible forest dam- pared results from several countries. In most coun- age, even if good annual averages of these pa- tries, the marketed value of ecosystems associated rameters have been obtained at the actual with timber and fuelwood production was less forested sites (as opposed to the present situa- than one-third of the total economic value, includ- tion, when values are essentially from urban ing non-market values such as carbon sequestra- areas). If the equations in Appendix X are used, tion, watershed protection, and recreation. special consideration must be given to local It is likely that the ratio between the value of the conditions and the large uncertainties must be forest regarded as timber and the value of an envi- emphasized. ronmental good varies with the degree of damage. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 119 T H E I M P A C T O F A C I D R A I N A N D S O 2 O N C R O P L O S S A small reduction in tree growth rates, say less than loss of timber, most losses are very difficult to 10 percent, may have little or no effect on some estimate. As a first step, we suggest that loss in non-market values, e.g. soil erosion and recreation. CO2 capture can be calculated and used in com- Although the total value of losses related to bination with the best current value per ton CO2 forest damage is likely to be several times that of for monetization. 120 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 6.2 Material Damage Air pollution causes significant material damage in Air pollution causes material damage by corroding and deteriorat- southern China, where dry sulfur dioxide deposition ing materials. Atmospheric corrosion and deterioration of materi- corrodes or deteriorates a variety of materials, als is a cumulative, irreversible process that also takes place in the mainly building structures. This chapter, reporting absence of pollutants. The reactivity to various air pollutants varies on findings in 14 municipalities and provinces in greatly between different materials and pollutants. Together with southern China, estimated the economic cost of the level of air pollution, particularly SO2 and O3, and the pH in this damage to be about 6.7 billion RMB in 2003, precipitation, the deterioration processes also largely depend on with Guangdong, Zhejiang and Jiangsu bearing the meteorological conditions, especially the "time of wetness" (time highest economic burden and accounting for more fraction with relative humidity >80 percent and temperature >0°C) than 50 percent of the total damage incurred. (Kucera and Fritz 1993). Two processes are involved in deteriora- While there are a number of uncertainties associated with this cost estimation, if combined tion of materials. One is corrosion of metals, which are electro- with the significant economic burden from crop loss chemical processes depending on the presence of humidity. The due to acid rain, it is clear that the air pollution other is chemical reactions that alter the properties of materials. issue demands urgent attention. Materials with basic properties, such as calcium-rich rocks and con- crete, may be sensitive to acidic components. Photochemical oxi- dants such as ozone (O3) are also capable of damaging certain materials. Economically important materials that are susceptible to ozone damage include elastomers (natural rubber and certain syn- thetic polymers), textile fibers, and dyes. Culturally important materials, such as a number of artists' pigments and dyes, may also be damaged (U.S. EPA 1996). CAUSAL AGENTS AND IMPACT PATHWAY Causal Agents Previous studies show a relationship between a range of air pollu- tants and deterioration rates for different materials. The dose- response relationships presented below are generally based on variables such as the concentrations of SO2 and O3, the concentra- tion of H+ in the rain, and moisture. As there are few monitoring data for O3 in China, we decided to select functions that include only SO2 and pH in precipitation as variables. C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 121 M A T E R I A L D A M A G E Valuation Scope ing the seventh "five-year plan period," buildings and bicycles accounted for 55 percent and 40 per- The moisture in the air greatly affects the degree cent, respectively, of the total material loss in two to which acid deposition corrodes materials. In provinces of Guangdong and Guangxi in 1985. the north of China, where the climate is always The valuation model in the design phase of the quite dry and the days with relative wetness present work included both building materi- greater than 80 percent are infrequent, the dam- als and bicycles. During the trial computation aging effect of dry acid deposition on materials period, we found that the proportion of bicycle is probably very low. Referring to the final report loss to the total material loss was reduced to only from the joint Shanxi-Norwegian project Mas- 1.5 percent of the total for Guangdong and ter Plan Against Air Pollution in Shanxi Province, 3.4 percent for Sichuan. The reason for this the material losses represented only 0.19 percent of the total pollution cost (Shanxi Environmen- remarkable change may be that both building tal Information Centre, Norwegian Institute for areas and material prices have grown rapidly, while Air Pollution 1994). Thus, we limit the scope of the amount and price of bicycles increased little in the valuation to southern China, specifically the recent years. In addition, although vehicles and air provinces or municipalities of Shanghai, Jiangsu, conditioners are under long-term outdoor expo- Zhejiang, Fujian, Ahhui, Jiangxi, Guangdong, sure, they are usually replaced not because their Guangxi, Hunan, Hubei, Sichuan, Chongqing, lifetimes have been reduced as a result of air pol- Guizhou, and Yunnan. lution, but due to other reasons. With respect to ancient architectural structures, there is currently no standardized valuation approach. Therefore, Valuated Material Types building materials are the exclusive valuation A wide range of materials is exposed to polluted object in the following. ambient air, including the materials used in build- ings, bicycles, cables, ancient architectural struc- EXPOSED MATERIALS tures, railways, and bridges. Moreover, the types and amount of materials being used are increas- In a valuation at a national level, there are two ing. The numbers of vehicles and air conditioners main methods for estimating exposed stocks: have increased rapidly in recent years in China. (1) using the indicator of material stocks per All materials exposed in the acid environment will capita, or (2) looking at material stocks per unit to some extent be eroded. According to a Chinese construction area. While the valuation result from study (see Annex A.3) conducted in the 1980s on the latter indicator may be more reliable, the for- the effect of acid deposition on materials, the ele- mer one is more feasible because of deficient ments that are most damaged are building sur- updating and comprehensiveness of the statistical faces and bicycles. Another study (Henriksen data for urban construction areas in China. We et al. 1999: Kai et al. 1999) in Guangzhou also apply two different datasets for building material found that the loss of only three materials-- stocks per capita based on the material stocks sur- outdoor galvanized steel, painted and galvanized veys made in Jinan (see Annex A.3), Shanxi, and guardrails--represented nearly 80 percent of the the previous study in Guangzhou (Henriksen et total material loss. In conclusion, we only include al. 1999), as shown in Table 6.6, to calculate the in the valuation the materials with large amounts, material stocks of a) southeastern and b) southern extensive distribution area, and the support of cities in the southern acid rain region. exposure-response functions. The total area of exposed building materials With regard to the research finding of the is calculated from the exposed material stocks national key scientific and technical project dur- per capita in m2/person from Table 6.6 times the 122 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N M A T E R I A L D A M A G E population in all cities in the provinces that are T A B L E 6 . 6 Building Material Stocks Per Capita for included. Eastern Cities and Other Southern Cities (m2/ per capita) DOSE-RESPONSE FUNCTIONS AND MATERIAL COSTS Southeastern Other Southern Materials Cities Cities Several studies on the relationship between materials damage and air pollution in Western Cement 7.25 18.34 Brick 18.51 13.15 countries and China have provided a relatively Aluminum 10.03 3.2 robust basis for exposure-response functions for Painted wood 1.24 0.56 a wide range of materials, such as the ECE-ICP Marble/granite 9.14 0.47 program (Kucera et al. 1995; Zhang et al. 1993; Ceramics/Mosaic 40.97 7.76 Terrazzo /Cement 22.51 15.17 Wang et al. 1990). Functions for deterioration Painted plaster 18.08 20.59 rates of specific materials have been derived Tile 2.36 3.28 experimentally, in laboratory or under field con- Galvanized steel 0.29 -- Painted steel 6.69 0.28 ditions. From these functions we derived func- Painted steel as guardrail 13.82 13.82 tions for the relationship between air pollution Galvanized steel as guardrail 9.21 9.21 exposure and the service lifetime. By estimating maintenance and replacement cost related to Source: Authors Calculations. change in service time, the economic damage can be estimated (Kucera and Fitz 1993; Kucera et al. 1993). For those materials where a quantitative revised by field experiments--are presented in assessment of deterioration rate is difficult to Table 6.7. obtain, inspections of physical damage in the field have been used to directly estimate the rela- Exposure-Response Functions tionship between air pollution exposure and need from Europe for maintenance and replacement, i.e. service life- time (Kucera and Fitz 1993). In the following, ECON (2000) presented exposure-response func- we mainly draw upon two reports: one from tions derived from European studies, including China, and a review of dose-response functions the study by Kucera et al. (1995). These func- from Europe (ECON, 2000). tions were applied to estimate material loss in Guangzhou city by Tian et al. (1999). In order to Exposure-Response Functions facilitate the valuation of physical damage, Tian from China et al. adjusted the functions to better represent the Chinese situation and transferred them into ser- During the seventh five-year plan period, one of vice life-year reduction functions. Table 6.10 ren- the key scientific and technological projects was ders the functions applied by Tian et al. research on acid rain. A working group was orga- nized to conduct both indoor and field-exposure Exposure-Response Functions experiments on corrosion of different materials.1 Applied in the ECM The field exposure experiments were conducted in Liuzhou (a heavily polluted acid rain area), Since the material corrosion functions provided Nanning (a lightly polluted acid rain area), and by the Chinese project are based on the practical Guangzhou. tests and field experiments in southern China and The recommended dose-response relations-- consistent with the Chinese situation, we decided based mainly on indoor tests, but somewhat to apply the functions suggested by the Chinese C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 123 M A T E R I A L D A M A G E parameters used in the final valuation model are T A B L E 6 . 7 Exposure-Response Functions of Materials from summarized in Tables 6.9 and Table 6.10. the Study of the Chinese Acid Rain Project of the ECONOMIC COST ESTIMATION Seventh FYP Table 6.11 presents the exposed building mate- Materials Exposure-Response Functions rials of all provinces based on the building mate- rial stocks per capita in Table and population of Paint Y = 5.61 + 2.84[SO2] + 0.74 × 10 [H+]4 all cities in each province. Marble Y = 14.53 + 23.81[SO2] + 3.80 × 10 [H+] 4 Using equation 1 and the building material Galvanized Y = 0.43 + 4.47[SO2] + 0.95 stocks of each province (Table 6.11) and the steel × 10 [H+] 4 monitoring data for air pollution, we get the eco- Q235 steel Y = 39.28 + 81.41[SO2] + 21.2 × 10 [H+] nomic cost of each province in Table 6.12. 4 Aluminum Y = 0.14 + 0.98[SO2] + 0.04 The economic cost of material damage from × 10 [H+] 4 acid rain of all provinces in the southern acid rain region reached about 6.7 billion yuan in Source: Wang Wenxing, Zhang Wanhua et al., 1990 2003. Of all provinces in the southern acid rain Note: Y is the corrosion rate of material in a polluted region, Guangdong had the highest material cost area, µm/ year; [SO2] is the ambient concentration of SO2, mg/m3; and [H+] is the H+ concentration of rain, mol/l. of about 1.6 billion yuan, followed by Zhejiang and Jiangsu. Their material costs are both about 1.1 billion yuan. study in the valuation model. For materials like concrete and bricks that Chinese studies did not UNCERTAINTIES take into account, we apply the functions pro- vided by ECON (2000) and Tian et al. (1999). 1) There are great uncertainties in materials All exposure-response functions and associated inventory. One reason is that the types of T A B L E 6 . 8 Exposure-Response Functions of Materials Based on European Studies Materials Exposure-Response Functions Concrete If SO2 < 15 µg/m3, L = 50 years, else 40 years Bricks If SO2 < 15 µg/m3, L = 70 years, else 65 years Bricks with plaster L = 1000/(0.124 SO2 + 15.5 + 0.013 Rain H+) Painted wood L = 1000/(1.03 SO2 + 87.5 + 0.26 Rain H+) Marble L = 10000/(103.52 + 0.302 SO2 + 0.00487 Rain H+) Ceramics and mosaic If SO2 < 15 µg/m3, L = 70 years, else 65 years Concrete with stone grain If SO2 < 15 µg/m3, L = 50 years, else 40 years Paint for outer wall L = 1000/(0.28 SO2 + 18.8 + 0.07 Rain H+) Tiles If SO2 < 15 µg/m3, L = 45 years, else 40 years Galvanized steel as guardrail L = 30/(0.51 + 0.0015 TOW SO2 O3 + 0.0028 Rain H+) Painted steel as guardrail L = 1000/(1.37 SO2 + 103 + 0.35 Rain H+) Zinc ML = T0.92 (1.2[SO2]0.34exp(0.011RH + 0.062T - 0.9) + 0.21Rain[H+], T10°C ML = T0.92 (1.2[SO2]0.34exp(0.011RH - 0.028T - 0.9) + 0.21Rain[H+], T>10°C Sources: ECON, 2000; Tian et al., 1999. Note: L is the life expectancy in years; Rain is the annual rainfall in mm; H+ is the H+ concentration of rain in mol/l; TOW is the fraction of time when relative humidity exceeds 80 percent and temperature is greater than 0°C; and [SO2] is the ambient concentration of SO2 in µg/m3. 124 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N T A B L E 6 . 9 Exposure-Response Functions for Material Loss Valuation Materials Y (µm/year) or L(year) Literature Cement If SO2 < 15 µg/m3, L = 50 years, else 40 years 11, 13 Brick If SO2 < 15 µg/m3, L = 70 years, else 65 years 11, 13 Aluminum Y = 0.14 + 0.98[SO2] + 0.04 × 10 4[H+] 8 Painted wood Y = 5.61 + 2.84[SO2] + 0.74 × 10 4[H+] 8 Marble/granite Y = 14.53 + 23.81[SO2] + 3.80 × 10 4[H+] 8 Ceramics/Mosaic If SO2 < 15 µg/m3, L = 70 years, else 65 years 11, 13 Terrazzo/Cement If SO2 < 15 µg/m3, L = 50 years, else 40 years 11, 13 Painted plaster Y = 5.61 + 2.84[SO2] + 0.74 × 10 4[H+] 8 Tile If SO2 < 15 µg/m3, L = 45 years, else 40 years 11, 13 Galvanized steel Y = 0.43 + 4.47[SO2] + 0.95 × 10 4[H+] 8 Painted steel Y = 5.61 + 2.84[SO2] + 0.74 × 10 4[H+] 8 Painted steel as guardrail Y = 5.61 + 2.84[SO2] + 0.74 × 10 4[H+] 8 Galvanized steel as guardrail Y = 0.43 + 4.47[SO2] + 0.95 × 10 4[H+] 8 Source: Authors Calculation. Note: Y is the corrosion rate of material in a polluted area, µm/ year; L is the life expectancy in years; [SO2] is the ambient concentration of SO2, mg/m3; and [H+] is the H+ concentration of rain, mol/l. T A B L E 6 . 1 0 Parameters in the Valuation Model of Material Loss Materials CDL (1) Y0µm/Year (2) L0 Year (3) Y µm/Year (4) L Year (5) P Yuan/m2 Cement 50 40 22 Brick 70 65 65 Aluminum 10.0 0.141 (1)/(2) Table 3-3-4 (1)/(4) 200 Painted wood 13 5.63 (1)/(2) Table 3-3-4 (1)/(4) 20 Marble/granite 160 14.63 (1)/(2) Table 3-3-4 (1)/(4) 200 Ceramics/Mosaic 70 65 48 Terrazzo/Cement 50 40 26 Painted plaster 13 5.63 (1)/(2) Table 3-3-4 (1)/(4) 15 Tile 45 40 8 Galvanized steel 7.3 0.45 (1)/(2) Table 3-3-4 (1)/(4) 16 Painted steel 13 5.63 (1)/(2) Table 3-3-4 (1)/(4) 16 Painted steel as 13 5.63 (1)/(2) Table 3-3-4 (1)/(4) 16 guardrail Galvanized steel as 7.3 0.45 (1)/(2) Table 3-3-4 (1)/(4) 16 guardrail Source: Authors Calculations. Note: CDL is the critical damage limit of material, µm; Y0 is the corrosion rate of material in clean area, µm/year; Y is the corrosion rate of material in polluted area, µm/ year; L0 is the life expectancy of material in clean area, year; L is the life expectancy of material i in polluted area, year; P is the unit price of a single maintenance or replacement oper- ation, yuan/m2. B O X 6 . 1 Estimating the Cost of Corrosion and Deterioration of Building Materials The economic cost of corrosion and deterioration of building materials (in yuan/year) is calculated as: C = (1 L - 1 L0) × P × S (1) where L0 is the life expectancy of the material in clean areas (year); L is the life expectancy of the material in polluted area (year); P is the unit price of a single maintenance or replacement operation (yuan/m2), and S is the stock at risk (m2). C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 125 0 362 304 181 8,499 2,068 5,015 9,801 2,120 8,932 5,952 Yunnan 11,853 13,304 0 279 234 140 9,146 6,558 1,596 3,870 7,562 1,636 6,892 4,593 Guizhou 10,265 0 497 417 249 2,842 6,892 2,913 8,180 16,288 11,679 13,468 18,282 12,274 Chongqing 0 748 627 374 4,272 4,379 Sichuan 24,484 17,556 10,360 20,246 27,481 18,450 12,296 ) 2 m 0 591 496 295 3,374 8,183 3,459 9,712 Hunan 19,339 13,867 15,991 21,707 14,573 (10,000 0 617 Hubei 7,051 1,234 1,036 7,228 40,414 28,977 17,100 33,417 45,361 30,454 20,295 Region 0 Rain 392 329 196 9,193 2,237 5,425 2,293 9,662 6,439 Jiangxi 12,822 10,602 14,391 Acid 0 677 568 338 Anhui 3,867 9,378 3,964 22,163 15,891 18,326 24,876 16,701 11,130 Southern 0 the 510 428 255 2,914 7,067 2,987 8,388 in Guangxi 16,703 11,976 13,812 18,748 12,587 Provinces 5,140 9,761 1,211 30,036 76,709 41,570 37,881 93,278 74,933 27,741 57,287 38,191 All 169,816 Guangdong of 238 Stocks Fujian 5,909 8,178 1,011 7,452 1,920 5,457 7,513 15,091 33,407 18,350 14,741 11,270 508 Material 2,156 4,094 Zhejiang 12,599 32,176 17,436 15,889 71,230 39,126 31,431 11,636 24,029 16,019 Building 374 9,264 1,585 3,010 8,556 23,658 12,821 11,683 52,374 28,768 23,110 17,668 11,779 Shanghai Exposed 777 3,295 6,258 Jiangsu 19,256 49,178 26,650 24,286 59,801 48,040 17,785 36,727 24,484 108,870 Calculations. 6.11 as as Authors TABLE wood granite Mosaic Cement plaster steel steel steel guardrail steel guardrail Materials Cement Brick Aluminium Painted Marble/ Ceramics/ Terrazzo/ Painted Tile Galvanized Painted Painted Galvanized Source: M A T E R I A L D A M A G E 3) Since there are no monitoring data for some T A B L E 6 . 1 2 Material Loss of All Provinces in the Southern small cities, the cost valuation of these cities is Acid Rain Region based on monitoring data from neighboring (10,000 yuan) cities. Generally, the air quality of small cities is better than that in big cities, thus the total Provinces Material Losses effect may be overestimated. Jiangsu 104,882 Shanghai 54,841 Endnote Zhejiang 105,104 Fujian 22,037 1. In the indoor simulation tests, the variable factors were the Guangdong 158,266 acidity of precipitation and the concentration of SO2. Guangxi 15,246 Other factors were kept constant at the following values: Anhui 11,477 temperature 25°C, relative humidity 80 percent, velocity Jiangxi 18,246 of wind 0.6 m/s, concentration of O3 20 ppb, exposure Hubei 46,749 time 500 hours. The total exposure period was separated Hunan 38,509 into 42 cycles of 12 hours: rain for 0.5 hours, light for Sichuan 36,240 4 hours, moisture in the form of dew for 3.5 hours, and Chongqing 35,985 Guizhou 16,497 light again for 4 hours. Thus total exposure to light was Yunnan 10,327 336 hours, to rain 21 hours, and to dew 143 hours. Total 674,407 References Source: Authors Calculations. Aunan, K., T. K. Berntsen, and H. M Seip. 2000. "Surface buildings and materials used vary considerably ozone in China and its possible impact on agricultural crop yields." Ambio 29: 294­301. depending on the economic level and the area. China Statistical Yearbook 2004. Beijing. We have applied data from surveys in Jinan, China Agricultural Statistics Yearbook 2004. Beijing. Taiyuan, and Guangzhou, and the differences China's Rural Statistical Yearbook, 2004. Beijing. in material stocks per capita among the three (Statistical yearbooks for every province, autonomous region, and municipality.) cities are substantial. The average material Driscoll, C. T. et al. 2001. "Acidic deposition in the north- stocks per capita based on these three cities eastern United States: sources and inputs, ecosystem obviously cannot represent the national level. effects, and management strategies." BioScience 51: This implies that the urban material stocks of 180­198. ECON. 2000. An environmental cost model. ECON report different scales and economic levels are still the no. 16/2000. Oslo: ECON Centre for Economic critical issues for the valuation of material loss. Analysis. More data on the kinds and stocks of exposed Henriksen, J. F. et al. 1999. Mapping air pollution effects on material in Chinese cities is needed. materials including stock at risk in Guangzhou, China. Kjeller, Norway: Norwegian Institute for Air Research. 2) The dose-response coefficients applied in the Holland, M. R., D. Forster, and K. King. 1999. Cost- estimation are mainly from Chinese studies of Benefit Analysis for the Protocol to Abate Acidification, the 1980s. At that time, the main contributor Eutrophication and Ground-Level Ozone in Europe. to acid rain in China was sulfuric acid, while Abingdon, UK: AEA Technology. Available at http:// www.unece.org/env/tfeaas/4meeting/aea2_final.pdf now it is beginning to change due to increas- Feng Z. W., H. F. Cao, et al. 1999. Impact of acid deposition ing emissions of NOx in recent years. The con- on ecological environment and its recovery. Beijing: China tent of nitrate in the rain will increase and may Environmental Science Press. imply different effects on materials. The appli- Kai, T., H. Liangwan, G. Krigsvoll, and J. F. Henriksen. cation of the dose-response functions derived 1999. "Building materials pollution cost in Guangzhou." NORAD Project CHN 013. Technical Report B3. 20 years ago for the current situation implies Kjeller, Norway: Norwegian Institute for Air Research uncertainties. (NILU). C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N 127 M A T E R I A L D A M A G E Kucera, V. and S. Fitz. 1993. "Direct and indirect air pol- Pollution in Shanxi Province, 2004. Kjeller, Norway: lution effects on materials including cultural monu- Norwegian Institute for Air Pollution. ments." Water, Soil and Air Pollution 86: 153­165. Singh, B. and M. Agrawal. 2004. "Impact of simulated acid Kucera, V., J. F. Henriksen, D. Knotkova, and C. Sjöström. rain on growth and yield of two cultivars of wheat." 1993. "Model for calculations of corrosion costs caused Water, Air, and Soil Pollution 152: 71­80. by air pollution and its application in three cities." In Tang D. et al. 2004. "Integrated Monitoring Program on J. M. Costa and A. D. Mercer, eds. Progress in the Acidification of Chinese Terrestrial Systems­IMPACTS." Understanding and Prevention of Corrosion. London: Annual report­Results 2003. Institute of Materials. Tian, K., L. He, G. Krigsvoll, and J. F Henriksen. 1999. Kucera, V., J. Tidblad, J. F. Henriksen, A. Bartonova, and "Building materials Pollution Cost in Guangzhou." Air A. A. Mikhailov. 1995. "Statistical analysis of 4-year Quality Management and Planning System for materials exposure and acceptable deterioration and pol- Guangzhou. NORAD Project CHN 013. lution levels." UN/ECE International cooperative pro- Tu, J., et al. 2005. "Trends in chemical composition of pre- gram on effects on materials, including historic and cipitation in Nanjing, China, during 1992­2003." Atmospheric Research 73: 283­298. cultural monuments. Report No. 18. Stockholm: UN/EC. 2004. The Condition of Forests in Europe. Executive Swedish Corrosion Institute. Report 2004, Federal Research Centre for Forestry and MA. 2005. Millennium Ecosystem Assessment Synthesis Forest Products (BFH). Geneva: United Nations Eco- Report. Pre-publication final draft. March 23, 2005. nomic Commission for Europe. Available at: http:// Available at http://www.millenniumassessment.org/en/ www.icp-forests.org/pdf/ER2004.pdf Products.aspx. U.S. EPA. 1996. "Air quality criteria for ozone and related Mahapatra, A. K. and D. D. Tewari. 2005. "Importance of photochemical oxidants." Vol. II of III. Report EPA/ non-timber products in economic valuation of dry decid- 600/P-93/004bF. Research Triangular Park, NC: U.S. uous forest of India." Forest Policy and Economics 7: EPA. 455­467. Wang, W. P. and D. L. Mauzerall. 2004. "Characterizing Menz, F. C. and H. M. Seip. 2004. "Acid rain in Europe and distributions of surface ozone and its impact on grain pro- the United States: an update." Environmental Science & duction in China, Japan and South Korea: 1990 and Policy 7: 253­265. 2020." Atmospheric Environment 38: 4383­4402 NAPAP. 1998. National Acid Precipitation Program Biennial Wang Wenxing, Zhang Wanhua et al. 1990. "Study on Report to Congress. An Integrated Assessment. Silver Spring, material damage of acid rain in southern China." (Sev- Maryland. enth FYP project, 58-05-02-05). Rabl, A., J. V. Spadaro, and B. van der Zwaan. 2005. "Uncer- Zhang, Wanhua et al. 1993. "Research on the protection tainty of Air Pollution Cost Estimates: To What Extent techniques of air pollution." Does It Matter?" Environmental Science & Technology 39: Zhang X. D. 2001. "Value measurement on forest resources 399­408. in Changjiang basin." Paper presented at the Inter- Shanxi Environmental Information Center/Norwegian national Conference on Forest Environment Value Eval- Institute for Air Pollution. 2004. Master Plan Against Air uation, Beijing China, October 09­12, 2001. 128 C H I N A ­ E N V I R O N M E N T A L C O S T O F P O L L U T I O N