aerosols (Watson, 1979), Coal combustion from plants (Klein, et al., 1975), Secondary aerosol. 1 Central Western (CW), ..... 999 Kwai Chung Road, Kwai Chung.
Feasibility of Establishing Air Monitoring Supersites in Hong Kong Draft Report
Prepared by S.C. Leea, J.G. Watsonb, J.C. Chowb, K.F. Hoa, T. Wanga, A. Lauc, H., Guoa, S. Liud
a
Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong b Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, USA c Atmospheric Research Center, The Hong Kong University of Science and Technology d Academia Sinica, Taipei, Taiwan
Prepared for： Hong Kong Environmental Protection Department (HKEPD) 33/F., Revenue Tower, 5 Gloucester Road, Wan Chai, Hong Kong
September 18, 2008
S-1.
EXECUTIVE SUMMARY
Hong Kong (HK), the Pearl River Delta (PRD) Region, and much of China face important air quality problems, especially for PM2.5 and O3. These problems can only be understood and solved with a thorough understanding of the sources, transport, chemical transformations, ambient concentrations, and effects on visibility and human health. Although some of the science and several of the air quality management and emission reduction measures developed in the U.S. and Europe can provide starting points, Hong Kong and the rest of China have too many special situations to expect cost-effective results by primarily adopting these methods. A supersite is recognized by: 1) a design that answers specific questions and develops conceptual models; 2) continuous measurements with shorter durations, greater frequency, lower detection limits, and more observables than those used for air quality compliance monitoring; 3) frequent and detailed analysis for chemical components of particulate matter (PM) and volatile organic compounds (VOCs); 4) consistence and quality assured long-term monitoring (at least 1 year at the same location) to experience a wide range of atmospheric environments; 5) public dissemination and documentation of validated and quality-assured data; and 6) reports and journal articles that use supersite data to test hypotheses. Supersite monitoring is not separate from other necessary activities, such as emission inventory improvement, source apportionment, chemical transport modelling, air quality forecasting, and evaluation of control strategy effectiveness. Long-term supersite monitoring provides a perspective for such programs because it captures many different meteorological and emissions conditions and allows generalizations to be made from better studied pollution episodes. The objectives of the Hong Kong Supersites Program are: 1) create collaborative relationships, including with Project 863, to obtain, analyze, and model measurements that address the questions cited above; 2) evaluate, compare, and deploy modern measurement methods that go beyond compliance monitoring;. 3) obtain measurements with specified accuracy, precision, and validity that permit data analysis, source, and receptor; 4) analyze data, and apply source and receptor models using acquired data to inform emissions reduction decisions; 5) and support studies related to health, visibility, and material damage effects of air pollution in the PRD. This study is limited to this evaluating the feasibility and practicality of adding a Hong Kong Supersites program to HKEPD’s air quality management portfolio. Air quality planning requires a large amount of knowledge that does not presently exist. The Hong Kong Supersites program will provide quantitative and defensible information on: Source Contributions: The largest contributors will be identified and their contributions quantified by time of day, time of year, and location. Limiting precursors will be determined, and the point of inflection when another precursor becomes limiting will be identified. Smog and visibility: The key pollutants will be identified and their contributions quantified by time of day and location. It will help to provide a scientific basis for policy-making of control policy in Hong Kong. S-1
Control Measures and Strategies: Local, urban, and regional scale emissions that affect ambient concentrations will be determined. Local controls are easier to implement than regional controls, and justification will be provided for these decisions. Human Exposure: Human exposures may vary with location, both in the horizontal and the vertical. Supersites can be surrounded by inexpensive and portable satellite sites to determine their zones of representation and to identify hot spots. Effects on Health: Much of what is known about health effects is based on studies in the U.S. and Europe. The long-term continuous record from supersites allows for better relationships between health end-points and several air pollution indicators that can improve understanding of how well results from non-Asian areas can be interpreted within an Asian context. Improving Air Quality Models: Improved data are needed as inputs to and for evaluation of air quality source models. These improved models can be used to forecast air quality and to give advice to sensitive populations and to test different emissions reduction scenarios for their effects on pollution levels. Improving Compliance Networks: Current Hong Kong networks are modelled on the U.S. EPA’s Federal Reference Methods that are not easily adaptable for multi-purpose monitoring. EPA’s National Monitoring Strategy calls for future multi-purpose monitoring, and Hong Kong can use the supersites to become a leader in this new monitoring concept. Benefits similar to those accrued from the U.S. supersites program are anticipated for the Hong Kong and PRD supersites programs. These include: Emission reduction decisions based on scientific demonstration in a holistic view. Increased collaboration among Hong Kong, mainland, and U.S. scientists and academics that will increase the probability of success for future projects. Training of a future air quality workforce, through the participation of university students that will benefit industry and regulatory agencies. More than 530 scientific articles and technical reports related to Hong Kong/PRD were identified and catalogued, and their abstracts were reviewed in this study. Information related to emissions, meteorology, and ambient concentrations was summarized. This information is used to answer the questions to the extent possible based on findings from previous studies and to identify contradictions and knowledge gaps among these studies. The existing monitoring networks with emphasis on the existing Hong Kong compliance monitoring sites were analyzed. These are classified as representing middle-, neighbourhood-, urban-, and regional scales. Also the existing meteorological monitoring network and its ability to represent local phenomena (e.g., upslope and downslope flows, sea breezes), regional-scale transport conditions, and coupling of surface layers to layers aloft were analyzed.
S-2
Information on advanced air quality measurements was acquired from publications, technical reports, interviews with instrument users, and from vendors. Measurement principles, detection limits, accuracy, precision, standards, power and space requirements, and costs are tabulated. Moreover, the personnel needs, phase-in schedules, quality assurance activities were synthesized into a supersite plan. A preliminary estimate of sampling methods and locations is given in report. Measurements would be acquired at seven locations representing upwind/background sites, maximum emissions impact sites, maximum concentration sites, and extreme downwind monitoring sites. Potential Supersite locations are: Tsuen Wan, Yuen Long, Tung Chung, Mong Kok, Hok Tsui, HKUST campus, and Tai Mo Shan. Owing to the novel nature of many of the measurement systems, substantial effort is needed prior to routine operations to develop procedures and expertise in their use. Since different techniques are used to quantify the same observables at different locations, it is important to collocate these instruments for a period of time to evaluate their comparability. Three approaches have been identified for scheduling Hong Kong supersite installation and operation, each with its own advantages and disadvantages. All the specifications for monitoring equipment and operational approach options, etc has been addressed in this report. The overall goal of supersite program is to improve our understanding of source/receptor relationships and atmospheric processes leading to PM/O3 accumulation on urban and regional scales, and thus provide the scientific understanding for modelling and data analysis efforts to support the future development of air quality programs in HK. The supersite program can be a multi-year, multi-task program which need a long term financial commitment. The funding sources of HK supersite program can be solely from HK government, University, or even a government-Univeristy partnership. The Government-University partnership can be beneficial for both HKEPD and Universities in terms of resource allocations, equipment sharing, current monitoring site and operation at monitoring sites, staff and graduate student training, etc. The Government-University partnerships can also provide a win-win situation for both HKEPD and Universities in HK.and the concerted efforts will positively help solve the local and regional air quality problems in the long run.
S-3
执行概要 香港，珠三角和中国很多地区面临严重的空气质量问题，尤其是 PM2.5 颗粒物和臭 氧污染。为了科学地解决大气污染的问题，我们需要深入了解污染物的排放源，大气 迁移转化过程，环境浓度水平，及其对能见度和人类健康的影响。尽管美国和欧洲国 家提倡的科学方法和空气质量管理、减排措施可以为我们提供一些初步建议，但是由 于香港和中国其他地区具体情况的特殊性，盲目地引用这些策略将无法收到预期的效 益。 超级测站的特征： 1）能具体地回答空气污染问题，建立具香港特式的概念模型； 2）提供比常规空气监测站周期更短的、频率更高的、检测限更低的、监测内容更广的 持续测量； 3）针对颗粒物和挥发性有机污染物的化学成分作周期性的详细分析； 4）针对大范围大气环境的连续的有质量保证的监测（同一站点至少 1 年）； 5）经过验证、有质量保证的数据的分发并文档化； 6）利用超级测站的数据来检验假定的报告和学术文章。 超级测站并不独立于其他必要活动，例如排放清单的改进，污染源分摊，化学传输模 型，空气质量预报，和控制策略有效性评估等等。 长期的超级测站监测能捕获很多不同的气象和排放条件，通过研究污染事件进行 归纳总结，所以为前述的这些项目提供了一个远景。香港超级测站的目标包括： 1）建立合作关系，包括与 863 计划，以获取、分析及模拟那些有关上述空气污染问题 的测量结果； 2）评价、比较并利用那些超过常规监测水平的现代测量手段； 3）获得有具体准度、精度和有效性的数据，从而实现数据，源和受体分析； 4）分析数据，应用获得的数据支持源-受体模型，并提出减排意见； 5）支持珠三角区域内进行的空气污染造成的健康、能见度及材料损害影响的研究。 本研究仅限于评价在香港环保署大气质量管理现有基础上增设超级测站项目的可 行性和实用性。 大气质量的规划所需要的大量知识与信息至今仍不完善。香港超级测站项目可以 在以下几个方面提供量化的可靠的信息，包括： 源贡献：确定最大的污染源，量化它们每日，每年及不同地区地区的污染贡献情 况，确定限制性前驱物，识别其他前驱物成为限制因素的转变点。 煙霧和能見度: 对于造成烟雾与能见度损害之污染物作科学分析（日夜变化，季节 变化，地域变化）。并为香港环保署科学提供作为制定未来管制政策的依据。 控制措施及策略：确定局地、城市及区域尺度中影响环境污染物浓度的各种排放 。局地尺度控制相对区域尺度控制更容易实施，超级测站项目将提供支持这些决 策的依据。 人体暴露：地区不同，在水平和垂直范围内的人类暴露也不尽相同。超级测站可 以与其外的轻便的卫星测站一起决定它们代表的区域并确定污染物最高浓度之暴 露位置。 S-4
健康影响：目前大多数对人类健康影响的了解都基于美国和欧洲国家的研究。超 级测站长期连续的监测数据可以更好地揭示人类健康和空气污染指标之间的关系 ，并提高对于非亚洲地区的研究结果应用于亚洲地区是否合适的认识。 改进空气质量模型： 空气质量源模型需要更多改进的有效数据作为输入。这些改 进的模型可以用来预报空气质量，为敏感人群提供建议，还能检验不同排放控制 方案的有效性。 改进常规监测网络。目前香港的网络是仿照美国环保署的联邦参考方法，不能很 好的适应多用途监测。美国环保署的国家监测战略开始倡导未来多用途监测，香 港可以利用超级测在这个新的监测概念中成为领导者。 香港及珠三角超级测站项目预期可获得与美国超级测站相类似的益处，包括： 基于全面检视科学论证来决定污染减排决策。 香港，内地以及美国科学家和学者间日益增多的合作，增加了未来项目成功的可 能性。 通过大学研究生的参与，培养新一代空气质量工作者，有益于香港空气质量人才 之培养。 本研究回顾了超过 530 篇与香港/珠三角相关的科学文献及技术报告。我们总结了 有关排放，气象，环境浓度等方面的信息，并利用这些信息来回答以往研究基础上提 出的科学问题，同时确定现有研究之间的矛盾和知识缺陷。 我们分析了已存在的监测网络，重点分析了香港现有的常规监测站点。这些站点 被分类代表中等尺度，邻近城市，市区及区域尺度。我们还分析了现有气象监测网络 ，及其反演局地现象（如顺风，逆风，海风等）、区域传输情况、地表层到上层大气的 耦合情况的能力。 我们从以往的文献、技术报告及跟仪器商的交流中获取了先进空气质量测量的各 种信息，对测量原理、检测限、精度、准度、测量标准、功率及空间要求、成本等信 息汇总成表。有关于人员需求、日程安排、质量控制措施的信息也包含在超级测站计 划中。 报告还给出了采样方法和地点的初步设计。测量将在 7 个代表地点进行，代表逆风 （上风向）/背景站，最大排放影响站，最大浓度站，极端顺风（下风向）站。可能的 超级测站地点位于：荃湾，元朗，东涌，旺角，鹤咀，香港科大校园以及大帽山。 基于许多监测系统的复杂性，在日常操作以前，我们需要在制定操作程序和培训 专员方面做出很多努力。因为不同的技术将用来定量不同站点的同种污染物，把这些 仪器收集在一起评价它们的可比性是很重要的工作。我们确立了三种香港超级测站安 装及操作的方法，每种有它自身的优点和缺点。本报告也阐述了所有监测仪器的规格 和操作的具体方法。 超级测站计划之目标是帮的香港了解污染物源与受体之关系，城市及区域之大气 气溶胶/臭氧之积累过程，并为未来香港空气污染的发展提供了模型与数据的支持。超 级测站是一个多样性项目需要长期的财政支持，经费来源可以单从香港政府支持，大 学支持或政府/大学共同支助。由香港政府与大学共同支持，才能有利于双方在资源分 配，设备共享，现有监测站点的运作，并工作人员和研究生培训等的发展工作。由香 港环保署与大学共同支持超级测站计划可共创双赢，并同时改善香港及珠三角之远空 气污染问题。 S-5
TABLE OF CONTENTS Section
Title
Page No.
S-1 EXECUTIVE SUMMARY…………………………………………….........................S-1 1. INTRODUCTION .......................................................................................................... 1-1 1.1 Background.............................................................................................................. 1-1 1.2 Supersite Objectives ................................................................................................ 1-2 1.3 Study Design Objectives ......................................................................................... 1-3 1.4 Overview of Technical Approach ........................................................................... 1-3 2. SCIENTIFIC QUESTIONS AND ANSWERS.............................................................. 2-1 2.1 Air Quality Questions.............................................................................................. 2-1 2.1.1 What are the pollutants of greatest concern in Hong Kong and why are they important? .............................................................................................. 2-1 2.1.2 When and where pollutants attain excessive levels? ..................................... 2-4 2.1.3 What are the levels of ultrafine particles and how do they evolve from primary emissions and secondary aerosol formation? ................................. 2-10 2.1.4 What are the characteristics of different particle size fractions (PM mass, number, surface area as a function of temporal and spatial variability) that affect regional haze? What are the spatial and temporal characteristics of visibility (total extinction and its components)? How do chemical constituents contribute to the light extinction budget?............ 2-11 2.1.5 What synoptic and local meteorological conditions are associated with observed PM2.5 and O3 standard exceedences in HK? Which variables are most important for each situation?......................................................... 2-13 2.2 Source Questions ................................................................................................... 2-14 2.2.1 What are the major source types and their emittants that contribute to excessive O3 and PM2.5? How well are their emissions inventoried? ....... 2-14 2.2.2 How accurate and representative are emission factors, activity levels, source profiles, and temporal distributions for these emissions? ................ 2-16 2.2.3 How much do the major source types contribute to primary PM2.5 and directly emitted precursors of secondary O3 and PM2.5? ............................. 2-17 2.2.4 What are the major source areas affecting HK PM2.5 in different seasons?........................................................................................................ 2-23 2.2.5 How does pollution transport on local and regional scales interact with vertical mixing to affect the diurnal variations in pollutant levels?............. 2-24 2.3 Control Strategy Questions.................................................................................... 2-25 2.3.1 Wha are the limiting precursors (e.g., VOC, NOx, NH3, HNO3, SO2) for PM2.5 and O3 formation in Hong Kong? ...................................................... 2-25 2.3.2 How will reducing the O3 precursors affect PM2.5 levels? .......................... 2-25 2.4 Measurement Questions ........................................................................................ 2-26
i
2.4.1 What are the most practical methods that will assist in answering the important questions? .................................................................................... 2-26 2.4.2 How comparable are measurements from different instruments and operating procedures? .................................................................................. 2-26 2.4.3 How should the Hong Kong EPD network be enhanced to serve better answer these questions in the future, within existing resource constraints? .................................................................................................. 2-27 2.4.4 How can collaborative relationships be established with Guangdong regulatory and research entities, including the 863 Major Project, to extend supersite monitoring over a larger region?....................................... 2-27 3. MONITORING NETWORK.......................................................................................... 3-1 3.1 Network Design for Multiple Purposes................................................................... 3-1 3.2 Air Quality Monitoring Networks ........................................................................... 3-1 3.2.1 Hong Kong EPD Air Quality Monitoring Network....................................... 3-1 3.2.2 Utility Industry Monitoring Network for SO2 and NO2 in Hong Kong......... 3-8 3.2.3 Hong Kong International Airport (HKIA) Air Quality Monitoring Network........................................................................................................ 3-10 3.2.4 PRD Regional Air Quality Monitoring Network......................................... 3-11 3.3 Meteorological Networks ...................................................................................... 3-15 3.3.1 HKEPD Meteorological Network................................................................ 3-15 3.3.2 HKIA Meteorological Network ................................................................... 3-15 3.3.3 Hong Kong Observatory (HKO) Meteorological Network ......................... 3-17 3.3.4 PRD Meteorological Network ..................................................................... 3-22 3.4 Visibility Monitoring Network.............................................................................. 3-25 3.5 Vertical Aerosol Measurement (LIDAR and Sun Photometer) ............................ 3-25 3.6 VOC Grid Monitoring ........................................................................................... 3-25 3.7 Supersite Measurement Enhancements ................................................................. 3-26 4. MODERNMEASUREMENT TECHNOLOGIES ......................................................... 4-1 4.1 Particle and Visibility Measurements...................................................................... 4-1 4.1.1 Advances in Filter-Based Integrated Samplers.............................................. 4-1 4.1.2 Continuous Particle Analyzers....................................................................... 4-4 4.2 Continuous Gaseous Analyzers............................................................................... 4-4 4.2.1 Continuous Nitrogen Species, SO2, CO, and CO2 ......................................... 4-4 4.2.2 Non-methane Hydrocarbon and Volatile Organic Compounds (VOCs) Speciation....................................................................................................... 4-5 4.3 Proposed Advance Measurements for Hong Kong Supersites Program ................. 4-6 4.4 Proposed Additional Special Studies..................................................................... 4-11 5. SUPERSITES CONFIGURATION AND OPERATION .............................................. 5-1 5.1 Space and Power Requirements .............................................................................. 5-1 5.2 Standard Operating Procedures ............................................................................... 5-7 5.3 Staffing Requirements ........................................................................................... 5-10
ii
5.4 Data Management and Validation ......................................................................... 5-12 5.5 Installation and Operations Schedule .................................................................... 5-13 5.5.1 Install and operate less complex instruments before more complex instruments................................................................................................... 5-13 5.5.2 Begin operation of all measurements and locations simultaneously ........... 5-14 5.5.3 Install and operate the TW site prior to or after the YL site ........................ 5-16 5.6 Quality Control and Quality Assurance ................................................................ 5-16 5.7 Project Management.............................................................................................. 5-26 5.8 Manpower and Budget .......................................................................................... 5-27 6. SUMMARY.................................................................................................................... 6-1 7. REFERENCES AND BIBLIOGRAPHY....................................................................... 7-1 APPENDIX A: Pearl River Delta Air Quality Studies ....................................................... 7-1 APPENDIX B. U.S. AND TAIWAN SUPERSITE MEASUREMENTS..........................B-1 APPENDIX C. INSTRUMENT DESCRIPTIONS ............................................................C-1 APPENDIX D. SUPERSITE MEASUREMENT SPECIFICATIONS ............................. D-1
iii
LIST OF TABLES Table Table 2-1. Table 2-2.
Title Page No. Air quality standards and objectives for China and Hong Kong. ..................... 2-2 Air quality standards and guidelines for the European Union, the United States (USA), and the World Health Organization (WHO). ....................................... 2-3 Table 2-3. Summary of PM2.5 mass and chemical composition (2000-2005 special study data) in Hong Kong .......................................................................................... 2-7 Table 2-4. Mean particulate matter and carbon concentrations in Hong Kong. ................ 2-8 Table 2-5. Hong Kong annual emissions estimates for 2006 (EPD, 2008b) and for the Pearl River Delta Region for 2003 and 2010 (Guangzhou EPB and EPD) in 1000kg/yr. ........................................................................................................................ 2-15 Table 2-6. Summary of PM source apportionment studies in Hong Kong...................... 2-19 Table 3-1. Summary of the Compliance Air Quality Monitoring Stations and TMS monitoring station in Hong Kong. .................................................................... 3-4 Table 3-2. Summary of Gaseous and Particle Measurements at the 14 Compliance Monitoring Station in Hong Kong.................................................................... 3-5 Table 3-3. Measurement Specifications for the HKEPD Monitoring Stations.................. 3-6 Table 3-4. Sampling and Analysis Methods Used in Measuring Toxic Air Pollutants by HKEPD............................................................................................................. 3-7 Table 3-5. Summary of Industrial SO2 and NO2 monitoring network............................... 3-9 Table 3-6. Measurement Specifications for the Utility Industry SO2 and NO2 Monitoring Network ............................................................................................................ 3-9 Table 3-7. Measurement Specifications for the HKIA Monitoring Network .................. 3-11 Table 3-8. Summary of the Regional Air Quality Monitoring Network in Pearl River Delta Region............................................................................................................. 3-13 Table 3-9. Gaseous and RSP Measurement Methods for the PRD Region Air Quality Monitoring Stations. ....................................................................................... 3-14 Table 3-10. Summary of the Hong Kong Observatory meteorological stations in Hong Kong. ........................................................................................................................ 3-20 Table 3-11. Summary of the meteorological stations in Pearl River Delta Region.......... 3-24 Table 3-12. Proposed locations for the Hong Kong Supersites ......................................... 3-27 Table 4-1. Summary of the thermal/optical analysis protocols for IMPROVE_A, IMPROVE, Hong Kong Governmental Laboratory (HKGL) and Hong Kong University of Science and Technology (HKUST-3). ....................................... 4-2 Table 4-2. Summary of Proposed Instrument for Hong Kong Supersites Program .......... 4-7 Table 4-3. Proposed Special Study Instrument................................................................ 4-11 Table 5-1. Space requirements for additional Supersite instruments at existing monitoring sites. .................................................................................................................. 5-1 Table 5-2. Power requirements for supersite instruments at each site................................ 5-1 Table 5-3. Standard operating procedures (SOPs) for the Hong Kong Supersite Program.5-8 Table 5-4. Measurements added at each site and phase-in periods. ................................. 5-14 Table 5-5. Elements of a Quality Assurance Project Plan (QAPP) .................................. 5-17
iv
Table 5-6. Table 5-7.
Quality Assurance Activities for the Hong Kong Supersites Program. .......... 5-20 Manpower requirements by category and project period. ............................... 5-28
v
LIST OF FIGURES Figure Figure 1-1.
Figure 2-1. Figure 2-2. Figure 3-1. Figure 3-2. Figure 3-3. Figure 3-4. Figure 3-5. Figure 3-6. Figure 3-7. Figure 3-8. Figure 3-9. Figure 5-1. Figure 5-2. Figure 5-3. Figure 5-4. Figure 5-5. Figure 5-6. Figure 5-7. Figure 5-8.
Title
Page No.
Worldwide supersites as identified from published literature indicating that the six supersite objectives have been attained. (A more detailed tabulation of their measurements is needed to verify compliance with the criteria.)..............1-2 Monthly variations of PM10 in seven HKEPD monitoring stations from 2001-2007.........................................................................................................2-23 Variations of PM10 in four different seasons at YL and TW sites during 2001-2007.........................................................................................................2-24 Locations of Hong Kong EPD’s Compliance Air Quality Monitoring Stations (and TMS PM2.5 monitoring station)..................................................................3-3 SO2 and NO2 Monitoring Network Operated by Two Utility Industries in Hong Kong ...................................................................................................................3-8 Hong Kong International Airport Air Quality Monitoring Network ...............3-10 Locations of PRD Regional Air Quality Monitoring Stations .........................3-12 Locations of Surface and Upper-Air Meteorological Stations in Hong Kong .3-15 Locations of Surface and Upper Air Meteorological Stations in HKIA ..........3-17 Locations of Meteorological Stations in Pearl River Delta Region. ................3-23 Locations of proposed Hong Kong Supersite Program....................................3-26 Average PM10 and chemical composition for the several sites in Hong Kong (2001-2007)......................................................................................................3-27 Additional supersite instruments to be located at the TW site (previous AQO room). .................................................................................................................5-2 Additional supersite instruments to be located at the TW site (inside shelter) ..... 1 Additional supersite instruments to be installed at the YL site (inside shelter). 5-3 Proposed environmental shelter (1.2m (L) × 1.2m (W) × 1.5m (H)) to be placed at the YL site (outside AQO station). .....................................................5-3 Additional supersite instruments to be located at the inside of the MK site (inside AQO station). .........................................................................................5-4 Additional supersite instruments to be located: a) on the rooftop and b) in the outside environmental shelter at the MK site (outside AQO station). .................. 1 Proposed mini environmental shelter for the TMS site......................................5-6 Hong Kong Supersites program management structure...................................5-27
vi
LIST OF APPENDICES APPENDIX A:
Summary of Pearl River Delta Air Quality Studies.
Table A-1. Summary of the Literature Reviewed for the Hong Kong and the Pearl River Delta Region Economic Zones. ............................................................... A-1 Table A- 2. Summary of Aerosol Characterization Studies. ................................................. A-37 Table A- 3. Summary of Aerosol Characterization Studies in PRD region .......................... A-70 Table A- 4. Summary of Ozone Characterization Studies. ................................................... A-82 Table A- 5. Summary of VOCs Characterization Studies..................................................... A-100 Table A- 6. Summary of Emissions Studies.......................................................................... A-127 Table A- 7. Summary of PM Source Apportionment Studies Using Models ....................... A-139 Table A- 8. Summary of Meteorological Studies.................................................................. A-149
Appendix B: Summary of U.S. and Taiwan Supersites Measurements. Table B- 1. Summary of U.S. and Taiwan Supersites Measurements................................... B-1 Table B- 2. Supersite Objectives in U.S. and Taiwan ........................................................... B-8 Table B- 3. The Project Roles in US Supersites (Watson et al., 2001) ................................. B-12
Appendix C: Measurement Method Descriptions and Analytical Specifications. Table C- 1. Analytical specifications for continuous PM2.5 mass and mass surrogate instruments (From Chow et al., 2008a). ............................................................. C-1 Table C- 2. Analytical specifications for continuous sulfate and multi-ions. (From Chow et al., 2008a)............................................................................................. C-6 Table C- 3. Measurement and analytical specifications for continuous carbon （From Chow et al., 2008a） .......................................................................................... C-12 Table C- 4. Measurement and Analytical Specifications for Gaseous NO/NOx, NOy, HNO3, PAN/NO2, and NH3 ................................................................................ C-19 Table C- 5. Measurement and Analytical Specifications for Gaseous CO, CO2 and SO2..... C-24 Table C- 6. Measurement and Analytical Specifications for Gaseous NMHC and VOCs ............................................................................................................................ C-28 Table C- 7. Summary of volatile organic components acquired in Hong Kong. .................. C-31
Appendix D: Site Summaries and Proposed Measurements. Table D-1. Measurements and Analytical Specification for the Proposed Hong Kong Supersites Program ..............................................................................................D-1
vii
Table D- 2. Power Requirements and Instrument Configurations for the Proposed Hong Kong Supersites Program ............................................................................................D-14
viii
1.
INTRODUCTION
1.1
Background
Hong Kong (HK), the Pearl River Delta (PRD) Region, and much of China face important air quality problems, especially for PM2.5 and O3. These problems can only be understood and solved with a thorough understanding of the sources, transport, chemical transformations, ambient concentrations, and effects on visibility and human health. Although some of the science and several of the air quality management and emission reduction measures developed in the U.S. and Europe can provide starting points (Bachmann, 2007; Chow et al., 2007a), Hong Kong and the rest of China have too many special situations to expect cost-effective results by blindly adopting these methods. The recently completed U.S. supersites program (Allen and Turner, 2008; Chow et al., 2008a; Demerjian and Mohnen, 2008; Fine et al., 2008; Russell, 2008; Solomon and Sioutas, 2008; Turner and Allen, 2008; Watson et al., 2008a; Wexler and Johnston, 2008) has been widely recognized as an important contributor to U.S. air quality science and control strategy development. More than 350 reports and publications resulting from this program have helped air quality decision-makers justify important emission reduction strategies such as the Clean Air Interstate Rule (U.S. EPA, 2005) and State Implementation Plans (SIPs), to attain U.S. National Ambient Air Quality Standards (NAAQS). Supersites were established and operated in the U.S. beginning in 1999 to: 1) evaluate new measurement technologies (Chow et al., 2008a; Demerjian and Mohnen, 2008); 2) improve understanding of atmospheric processes (Allen and Turner, 2008; Simon et al., 2008); and 3) support studies that relate air quality to adverse human health (Chow et al., 2006b; Pope, III and Dockery, 2006) and visibility impairment (Chow et al., 2002a; Watson, 2002a). A supersite is recognized by: 1) a design that answers specific questions and develops conceptual models; 2) continuous measurements with shorter durations, greater frequency, lower detection limits, and more observables than those used for air quality compliance monitoring; 3) frequent and detailed analysis for chemical components of particulate matter (PM) and volatile organic compounds (VOCs); 4) consistence and quality assured long-term monitoring (at least 1 year at the same location) to experience a wide range of atmospheric environments; 5) public dissemination and documentation of validated and quality-assured data; and 6) reports and journal articles that use supersite data to test hypotheses. A supersites program is not limited to a single location; it is embedded in an existing or temporary network to evaluate and understand the meaning of results from the network. It may add measurements of increased complexity at existing sites, as well as field instruments of low complexity (e.g., inexpensive battery-operated monitors, passive samplers) to better understand the zones of influence of certain sources and the zones of representation for existing compliance sites (Chow et al., 2002b; Chow and Watson, 2008). Supersite monitoring is not separated from other necessary activities, such as emission inventory improvement, source apportionment, chemical transport modelling, air quality forecasting, and evaluation of control strategy effectiveness. Supersite measurements enhance the accuracy and validity of these activities by supplying input data and data to verify their results.
1-1
Supersite programs do not replace short-term intensive air quality studies such as Program of Regional Integrated Experiments of Pearl River Delta Region (PRIDE-PRD) (Zhang et al., 2008). Long-term supersite monitoring provides a perspective for such programs because it captures many different meteorological and emissions conditions and allows generalizations to be made from better studied pollution episodes. Other countries have established supersites, and several long-term observatories (e.g., Mace Head in Ireland, Mauna Loa in Hawaii, Cape Grimm in Australia) have recently been recognized as potential supersites (Figure 1-1). A HK supersite would provide a valuable addition to this expanding global network.
Figure 1-1. Worldwide supersites as identified from published literature indicating that the six supersite objectives have been attained. (A more detailed tabulation of their measurements is needed to verify compliance with the criteria.) 1.2
Supersite Objectives
The goal of the HK Supersites Program is to provide a sound technical basis for cost-effective region wide emission reduction strategies. Given the size, population, and industrial activity in the PRD, attaining this goal will require increasing cooperation with air quality scientists and regulators in neighbouring Guangdong Province and its major population centers. China Project 863 has potential to complement the HK Supersites Program if common objectives and methods can be agreed to and acted upon. Project 863 includes supersites as one component in a more comprehensive program to upgrade compliance monitoring sites, improve emission inventories, develop air quality models, and objectify enforcement measures. HKEPD has counterpart efforts in all of these categories
1-2
except for the supersite component. This study is limited to the evaluation of feasibility and practicality of adding a Hong Kong Supersite program to HKEPD’s air quality management portfolio. Specific objectives of the HK Supersite program are: Evaluate, compare, and deploy modern measurement methods that go beyond compliance monitoring. Obtain measurements with specified accuracy, precision, and validity that permit data analysis, source, and receptor modelling to answer scientific questions relevant to air quality management. Describe relationships among emissions, meteorology, and atmospheric chemistry that cause excessive concentrations and how these vary in space and time. Quantify source contributions by time of day, time of year, and location. Determine limiting precursors for secondary O3 and PM2.5 formed from directly emitted SO2, NOx, VOC, and NH3 and how these precursors vary in space and time. Evaluate control strategy effectiveness through long-term tracking of source markers, source contributions, and limiting precursor concentrations. Improve utility of compliance network measurements through introduction of newer measurement technologies and better use of existing monitors, after thorough testing at supersites. Support research studies related to health, visibility, and material damage effects of air pollution in Hong Kong. 1.3
Study Design Objectives
The goal of this study is to develop a plan for Hong Kong Supersites program that will advance air quality planning for the PRD region. Specific objectives are: Develop and attempt to answer questions about causes of excessive HK pollutant concentrations. Determine where and when measurements should be taken to answer those questions. Estimate the capital and operating costs to obtain those measurements. Outline the necessary elements for supersite data acquisition, documentation, validation, management, and dissemination. Specify how the acquired measurements can be used in source models, receptor models, and other data analyses to answer and refine the outstanding questions. 1.4
Overview of Technical Approach
Policy-relevant questions were formulated in consultation with the EPD staff, civic groups, as well as local and international air quality researchers. More than 530 scientific articles and technical reports related to PRD were identified and catalogued, and their abstracts were reviewed. Copies of relevant articles were obtained and catalogued on a project website (access controlled owing to copyright restrictions). Information related to 1-3
emissions, meteorology, and ambient concentrations was summarized in consistent tables that appear in Appendices A through D of this report. This information is used in Section 2 to answer the questions to the extent possible based on findings from previous studies and to identify contradictions and knowledge gaps among these studies. Also summarized in tabular form are results from U.S. supersites that addressed similar questions, and these are evaluated to specify approaches that might also be applied to HK supersite measurements. Section 3 analyzes the existing PRD monitoring network, with emphasis on the existing Hong Kong compliance monitoring sites. These are classified as representing middle-, neighbourhood-, urban-, and regional- scales (Chow et al., 2002b). Also analyzed are the existing meteorological monitoring network and its ability to represent local phenomena (e.g., upslope and downslope flows, sea breezes), regional-scale transport conditions, and coupling of surface layers to layers aloft. Information on advanced air quality measurements was acquired from publications, technical reports, interviews with instrument users, and from vendors. Measurement principles, detection limits, accuracy, precision, standards, power and space requirements, and costs are tabulated. Section 4 analyzes these data to select instruments that have the potential to operate for extended periods without high maintenance costs and that provide data needed to answer the Section 2 questions. Information from these tasks is synthesized into a supersite plan in Section 5, which analyzes the personnel needs, phase-in schedules, quality assurance activities, and costs for a two year monitoring program that could be extended to four years. Section 6 summarizes the results of this feasibility study within the context of the fourteen tasks specified in HKEPD tender AS 07-386. Extensive references and bibliogaphy are provided in Section 7.
1-4
2.
SCIENTIFIC QUESTIONS AND ANSWERS
Several questions need answers to make important air quality management decisions in Hong Kong. These questions are divided into four categories related to air quality, source contributions, control strategies, and measurement characterizations. Some questions are subservient to other larger questions. For each question listed below, information is drawn from previous studies (Appendices A, Tables A1 to A7) to explain what is believed to be known or unknown about the topic. Examples are given from supersite and other studies about how the question might be answered using a combination of advanced air quality and meteorological measurements and data analysis methods. Measurements and data analysis activities for the Hong Kong Supersites Program are then recommended. 2.1
Air Quality Questions
Answers to these questions will provide a better understanding of when and where high pollutants occur, how they might affect haze and public health exposures, and how they are related to meteorological regimes. Answers to these questions will increase decision-makers ability, to separate local from regional source contributions, to develop better forecasting models, and to estimate improvements that might accrue from emission reductions on Hong Kong sources. 2.1.1 What are the pollutants of greatest concern in Hong Kong and why are they important? Primary air quality standards are established to protect public health as determined by epidemiological and toxicological studies (Byrd and Joad, 2006; Chow et al., 2006b; Davidson et al., 2005; Delfino et al., 2005; Grahame and Schlesinger, 2005; Koranteng et al., 2007; Mauderly and Chow, 2008; McDonald et al., 2007; Pope, III and Dockery, 2006; Ris, 2007; Romieu et al., 2008; Schlesinger et al., 2006; Schlesinger, 2007; U.S.EPA, 2006b; U.S.EPA, 2006c; U.S.EPA, 2006d; U.S.EPA, 2008; Zhao et al., 2006). The elements of an air quality standard are: 1) an indicator (e.g., TSP, PM10, or PM2.5 mass for suspended particles); 2) one or more averaging times; 3) a maximum level; and 4) permitted exceedances. The indicator needs to be practically measured at a large number of locations. In the case of suspended particles, there may be PM components that are more toxic than is indicated by the mass, or there may be particle sizes that allow particles to penetrate to more sensitive parts of the body, but there are insufficient measurements to establish concentration/health relationships. Even though they are not yet regulated, high numbers of ultrafine particles have been shown to overwhelm the human body’s defense mechanisms, travelling across the lungs and into vital organs such as the heart and brain (Alessandrini et al., 2006; Biswas and Wu, 2005; Cascio et al., 2007; Chow et al., 2005d; Chow and Watson, 2006; Kreyling et al., 2006; Rundell et al., 2007). There is also growing concern about the effects of coarse particles (PM10-2.5) on health (Gerlofs-Nijland et al., 2007; Kan et al., 2007; Lipsett et al., 2006; Schwarze et al., 2007). In the U.S., PM2.5 mass is the main indicator of PM adverse health effects because: 1) it penetrates into the lung; 2) single measurement locations can represent large exposure areas; and 3) this size fraction contains components that are believed to be toxic (e.g., heavy metals, acids, PAHs and other
2-1
organic compounds). PM2.5 also includes the particle sizes and chemical components accounting for most of the light scattering and absorption that causes urban and regional hazes (Chow et al., 2002a; Deng et al., 2008; Wang, 2003; Watson, 2002a). Hazardous Air Pollutants (HAPs) include heavy metals and certain VOCs that have resulted in a large number of adverse health effects (England et al., 2001; Kyle et al., 2001; Leikauf, 2002; U.S.EPA, 2001; Yan et al., 2002). The European Union has established one-year average ambient air quality standards for benzene (0.5 µg/m3), arsenic (120 ng/m3), cadmium (5 ng/m3), nickel (20 ng/m3), and benzo(a)pyrene (1 ng/m3) (Chow et al., 2007a). The U.S. regulates HAPs emissions rather than ambient concentrations. China’s air quality standards and Hong Kong’s air quality objectives (AQOs) are compared in Table 2-1. China has three grades for its standards: 1) Grade I applies to nature reserves, scenic spots and other areas in need of special protection; 2) Grade II applies to residential areas; commercial, transportation and residential mixed areas, cultural areas and general industrial areas specified in urban planning as well as rural areas; and 3) Grade III applied to large industrial zones. Most of the PRD cities are to achieve the Grade II levels. As seen in Table 2-1, China’s limits are more restrictive than Hong Kong’s air quality objectives. Table 2-2 summarizes air quality standards from European Union, the U.S., and the World Health Organizations. In many cases, China’s air quality standards are more restrictive than those of more developed countries, especially with respect to allowed exceedances. Table 2-1.
Pollutant CO SO2
Air quality standards and objectives for China and Hong Kong. China Grade Ⅰ/Ⅱ/Ⅲ Standards Grade I/II/III Avg Exceed 3 10/10/20 mg/m 1 hr 0/yr
H.K. Air Quality Objectives Conc Avg Exceed 3 30 mg/m 1 hr 3/yr
4/4/6 mg/m3
10 mg/m3
24 hr 3
150/500/700 µg/m 3
50/150/250 µg/m 20/60/100 µg/m NO2
24 hr
3
120/240/240 µg/m3 3
80/120/120 µg/m 40/80/80 µg/m PM10 PM2.5 TSP TSP Pb
160/160/200 µg/m 3
1/yr
1 hr
3/yr
3
24 hr
1/yr
800 µg/m
350 µg/m
1 yr
0/yr
3
80 µg/m
1 yr
0/yr
1 hr
0/yr
300 µg/m3
1 hr
3/yr
0/yr
3
150 µg/m
24 hr
1/yr
0/yr
3
1 yr
0/yr
1 hr
3/yr
1 yr 3
0/yr
8 hr
3
0/yr
24 hr
3
O3
1 hr
0/yr
1 hr
0/yr
80 µg/m
3
240 µg/m
3
50/150/250 µg/m
24 hr
0/yr
180 µg/m
24 hr
1/yr
40/100/150 µg/m3
1 yr
0/yr
55 µg/m3
1 yr
0/yr
None
None 3
120/300/500 µg/m
24 hr
0/yr
260 µg/m3
24 hr
1/yr
80/200/300 µg/m3
1 yr
0/yr
80 µg/m3
1 yr
0/yr
1.5 µg/m3
3 months
0/yr
1.5 µg/m3
3 months
0/yr
1 yr
0/yr
3
1.0 µg/m
2-2
CO, SO2, Pb and most of the HAPs are directly emitted, so reducing their emissions results in reductions in ambient concentrations. CO, NOx, and TSP Pb levels have been substantially reduced as a result of improved engine controls and fuels. HAPs levels have been reduced owing to reformulated fuels and surface coatings, improved fuel/chemical storage and transfer facilities, and precipitators and baghouses on industrial facilities. O3 and large fractions of PM2.5 and NO2 result from atmospheric transformation of directly emitted gases. While direct SO2 and NOx emissions may be sufficient to attain their ambient standards, these gases are also major causes of excessive PM2.5 and O3. Although they do not have air quality standards, ammonia (NH3) and a wide range of volatile organic compounds (VOCs) contribute to the formation of excessive O3, as well as secondary ammonium nitrate (NH4NO3), ammonium sulphate ((NH4)2SO4), and organic carbon (OC) in PM2.5. The processes that form secondary pollutants are not linear, and decreases in emissions may result in no effect, or even an increase, on the end-product until an inflection point has been passed. There may be cases in which the reduction of one precursor emissions (e.g., NOx) reduces PM2.5 nitrate, but increases O3 levels. The “limiting precursors” may differ by time and location (e.g. Watson et al., 1994). Table 2-2. Air quality standards and guidelines for the European Union, the United States (USA), and the World Health Organization (WHO). Pol-
European Union
lutant
Conc
CO mg/m3 SO2
10 350
3
µg/m
Avg 8 hr Max/day 1 hr
USA Exceed 0/yr 24/yr
125
24 hr
3/yr
NO2
200
1 hr
18/yr
µg/m3
40
1 yr
0/yr
O3 3
µg/m PM10
120 50
3
µg/m
PM2.5 3
µg/m
40
8 hr Max/day 24 hr 1 yr
25/3-yr 35/yr 0/yr
25
1 yr
0/yr
0.5
1yr
0/yr
WHO
Conc
Avg
Exceed
40
1 hr
1/yr
10
8 hr
1/yr
365
24 hr
1/yr
80
1 yr
0/yr
100
1 yr
0/yr
8 hr
th
4 3-yr
max/day
Avg
160 150
24 hr
35
24
2nd /3-yr
Conc
500 20/50/125
[a]
0/yr
24 hr
0/yr
200
1 hr
0/yr
1 yr
0/yr
100/160/240[b]
8 hr
0/yr
50/75/100/150[c]
24
3/yr
Avg
20/30/50/70
98/%
25
[c]
[c]
3 yr
0/3-yr
10/15/35/35
1.5
3 mo
0/quarter
None
µg/m3 [a] Guideline/Interim target-2/-1 [b] Guideline/Interim target-1 (daily maximum)/High level (Daily maximum)
2-3
10 Min
40
15
[c] Guide-line/Interim target-3/-2/-1
Exceed
None
TSP Pb
Avg
1 yr 24 hr
3/yr
1 yr
0/yr
In principle, the data from PAMS network could be extremely useful for the regulatory and scientific communalities. However, it appears that the full potential of the data has yet to be realized. Questions concerning the accuracy of the VOC and NOx concentrations obtained from the PAMS instrumentation have limited the ability of researchers to use the data to empirically assess the relationships between ambient VOC and NOx concentrations and O3 formation. In spite of those limitations, The PAMS data sets can be probably still be used to evaluate trends, but this type of evaluation generally requires a record of measurements of a decade or more, and the PAMS network is just now reaching that level of maturity. In summary, a larger number of air quality observables is needed other than the pollutants listed in Tables 2-1 and 2-2. Chemical compositions in various size fractions are needed to determine PM toxicity, origins, HAPs content, and effects on visibility. VOCs need to be quantified to evaluate their effects on secondary organic aerosol, HAPs, and O3 formation. SO2 and secondary SO4= levels need to be correlated with each other and with NH3 levels to understand the formation of (NH4)2SO4. NO, NO2, PAN, HNO3, NH3, and NOy levels need to be correlated to understand the limiting precursors for secondary O3 and secondary NH4NO3. Ultrafine particle numbers and size distributions are needed to better understand the health risk and to elucidate the formation mechanisms for secondary aerosol formation. Light scattering and absorption are needed along with size distribution and PM2.5 chemical composition to evaluate the chemical components that cause poor visibility and determine which emissions reductions will improve it most rapidly. Also lacking from the compliance measurements of Tables 2-1 and 2-2 are the meteorological observables that permit an understanding of how direct emissions are transported, mix together, and transform by the time they reach a receptor location. 2.1.2 When and where pollutants attain excessive levels? EPD (2008a) operates and reports data from an extensive network of air quality monitors (See Section 3) to determine human exposure and compliance with the AQOs in Table 2-1. During 2007, EPD (2008a) showed no exceedances of its one-hour CO and SO2 AQOs at any of the 11 general exposure and roadside stations. The hourly NO2 AQO was exceeded at all three roadside stations, and two concentrations > 300 µg/m3 were quantified at the Kwun Tong station. The hourly O3 AQO was exceeded at the Central/Western, Sha Tin, Tung Chung, Yuen Long, and Tap Mun stations, with the highest level (331 µg/m3) quantified at the Sha Tin station. AQOs for 24-hour averaging times were achieved for SO2 and TSP, but were exceeded for NO2 at the three roadside sites and for PM10 (RSP) at the Tung Chung, Yuen Long, and Causeway Bay sites. Annual average AQO were attained for all pollutants except for NO2 at the three roadside sites, for TSP at Kwai Chung, Kung Tong, Yuen Long, and Mong Kok, and for PM10 at Eastern, Sham Shui Po, Tsuen Wan, Yuen Long. Causeway Bay experienced the highest PM10 level at 85 µg/m3. Although EPD (2008a) demonstrates attainment of Hong Kong AQOs at many sites, nearly every site exceeds the Chinese Grade III levels for SO2, NO2, O3, and PM10. Many sites would exceed EU and U.S. air quality standards. Only CO and TSP Pb are within international consensus limits. Roadside monitors often show the highest levels.
2-4
GDEMC (2008) summarizes 2007 results from the 16 stations of Pearl River Delta Air Quality Monitoring Network that includes Hong Kong’s Tung Chung, Tsuen Wan, and Tap Mun stations (See Section 3). This network gives a broader geographical picture of where and when pollutants achieve high levels. Exceedances of the Chinese standards during 2007 were most common at the stations near Foshan, a highly industrialized area. The annual SO2 standard was exceeded at Guangzhou, Foshan, Zhaoqin, and Zhongshan, with the Hong Kong sites meeting the Chinese Grade II levels. Hourly and 24-hour NO2 Grade III standards were exceeded at most sites, including Hong Kong’s Tsuen Wan and Tung Chung sites. Foshan again showed the highest levels and the largest number exceedances. The annual average NO2 standard was not exceeded at any of the monitors. The hourly Grade II O3 standard was exceeded numerous times at all of the sites, with 243 hours above 200 µg/m3 at Zhongshan. Hourly O3 exceeded 200 µg/m3 10 times at Tsuen Wan, 59 times at Tap Mun, and 73 times at Tung Chung sites. Twenty-four hour PM10 exceeded the Grade II standard of 150 µg/m3 at all sites during 2007, with 128 exceedances at Foshan. The level was exceeded 7 times at Tsuen Wan, 3 times at Tap Mun, and 9 times at Tung Chung. An examination of seasonal differences shows that pollutants have minimal concentrations during the wet summer season, except when tropical cyclones create high pressure conditions that accumulate and recirculate air pollutants. These synoptic conditions often extend into October and November when PM10 and O3 reach their highest levels. PM10 generally remains high through May, with some evidence of PM10-2.5 incursions from Asian dust (Cao et al., 2003a). High O3 levels are again more frequent during the spring period, often associated with trough low and ridge high pressure systems. High NO2 levels near roadside sites are most intense in winter, but are also evident in fall and spring EPD, 2008a reports annual average levels of hexavalent chromium, lead, benzene, benzo[a]pyrene, 1,3,-buadiene, formaldehyde, perchloroethylene, and dioxins at the Tsuen Wan and Central Western sites. Benzene levels of 1.75 and 1.39 µg/m3 at Tsuen Wan and Central Western are higher than the 0.5 µg/m3 embodied in the EU standards. A more extensive analysis of Hong Kong HAPs can be found- by Lau et al., 2003. It is evident that SO2, NO2, O3, and PM10 levels can be excessive in Hong Kong and throughout the PRD. High values can occur throughout the year, although summertime has a lower frequency of occurrence than the other seasons. A larger number of exceedances are observed when Hong Kong levels are compared to China’s air quality standards. Exceedances are more common at other monitoring locations in the PRD. Moreover, PM2.5, PM10-2.5, ultrafine articles, and HAPs levels are not well characterized. Table 2-3 summaries previous PM2.5 speciation studies in Hong Kong. Organic mass (i.e., OC×1.4) and SO4= were the two largest components of PM2.5. These components accounted for 21-38% and 12-39% of PM2.5, respectively, while secondary (NH4)2SO4 and NH4NO3 are more homogeneously distributed in the region. PM2.5 elemental carbon (EC) exhibited large spatial variability, account for 3-8%; 9-16%; and 12-33% of PM2.5 at the regional; urban; and middle-scale sites, respectively. The impacts of vehicle exhaust on PM2.5 are apparent, especially at the road-side stations. PM2.5 crustal materials were in the range of 2-8%. PM2.5 sea salt ranged from 3-9%, with low NO3- in the range of 1-5%. These studies showed that secondary inorganic aerosol (i.e., (NH4)2SO4, NH4NO3) and carbonaceous aerosol are the main contribution to PM2.5 in 2-5
HK. Many studies only report PM masses and carbon concentrations, as summarized in Table 2-4. For studies conducted between 2000 and 2005, annual averages of PM2.5 ranged from 42 to 58 μg/m3 at the middle-scale roadside sites (i.e., MK, PU), 34-43 μg/m3 at the urban-commercial/residential sites (i.e., TW, KT), and 24-28 μg/m3 at the regional-scale HT site (Chow et al., 2002c; Ho et al., 2006a; So et al., 2007).
2-6
Table 2-3.
Summary of PM2.5 mass and chemical composition (2000-2005 special study data) in Hong Kong
Type of site
Site
Reference
(% of PM2.5 mass concentration) Regional HT Ho et al., 2006a HT Ho et al., 2006a HT Chow et al., 2002c HT So et al., 2007 Urban - scale TW Chow et site al., 2002c TW So et al., 2007 KT KT Middle-scale roadside site
MK MK PU PU
Ho et al., 2006a Ho et al., 2006a Chow et al., 2002c So et al., 2007 Ho et al., 2006a Ho et al., 2006a
Sampling Season/yr
Crustal materiala,b
OMc
EC
(NH4)2SO4d
NH4NO3e
SO4=
NO3-
NH4+
Sea saltf
Unidentified
Nov-Dec, 00 and Jan-Feb, 01 June-Aug, 01
5
21
3
43
4
31
3
4
9
23
6
25
4
22
0
16
0
3
5
41
2001/02
5
25
7
51
4
37
3
2004/05
4
21
8
44
4
32
3
2001/02
4
36
16
37
5
27
4
2004/05
3
27
15
47
5
34
4
10
Nov-Dec, 00 and Jan-Feb, 01 Jun-Aug, 01
8
28
9
37
5
27
4
5
6
14
42.6
6
31
15
22
1
16
1
3
5
23
42.6
2001/02
3
39
33
21
4
16
3
1
2004/05
2
31
26
33
6
24
5
8
Nov-Dec, 00 and Jan-Feb, 01 Jun-Aug, 01
4
29
12
35
6
25
5
6
3
15
58.3±2.7 (annual) 53.0±2.7 (annual) 42.2l
4
38
19
17
1
12
1
3
3
22
42.21
HT – Hok Tsui; TW – Tsuen Wan; KT – Kwun Tong; MK – Mong Kok; PU – The Hong Kong Polytechnic University a Crustal material=1.89xA1+2.14xSi+1.4xCa+1.43xFe for Chow et al., 2002c and So et al., 2007 b Crustal material=A1/7.26% for Ho et al., 2006 c OM (organic material)=OCx1.4 d (NH4)2SO4 = 1.38 [SO4=] e NH4NO3 = 1.29 [NO3-] f Sea salt=Na×2.54
2-7
8 11
8 2 2
2
Average PM2.5 (μg/m3) 28.2 28.2 (annual) 23.7±14.5 (annual) 28.4±2.4 (annual) 34.1±19.1 (annual) 39.0±2.0 (annual)
Table 2-4.
Reference
Mean particulate matter and carbon concentrations in Hong Kong. PM2.5 Mass
TC
OC
EC
OC/EC
TC/PM2.5
PM10 Mass
TC
OC
EC
OC/EC
TC/PM10
24-hr, Jun-Jul/02, IMP_TOR 24-hr, once every third day by Mini-vol, from 1/Aug/02 to 2/Aug/02, IMP_TOR
40.1±19.7
10.2±2.8
6.3±2.3
3.9±1.6
1.7
0.25
40.8±15.6
12.1±3.5
7.4±2.9
4.7±2.1
1.6
0.30
60.4±22.9
16.7±5.4
10.6±3.7
6.1±1.8
1.7
0.28
78.4±34.1
19.7±6.9
12.7±4.7
7.0±2.6
1.8
0.25
24-hr, Jun-Jul/02, IMP_TOR 24-hr, once every sixth day by Mini-vol, from 1/Aug/02 to 2/Aug/02, IMP_TOR
30.8±7.6
8.8±0.9
5.6±0.8
3.2±0.4
1.8
0.29
38.6±12.8
10.6±1.2
6.7±1.1
3.9±0.5
1.7
0.27
48.5±24.7
12.8±11.0
8.4±6.3
4.4±4.7
2.3
0.26
55.7±19.3
11.0±2.0
7.6±1.3
3.4±0.7
2.2
0.20
24-hr, Jun-Jul/02, IMP_TOR 24-hr, once every sixth day by Mini-vol, from 1/Aug/02 to 2/8/02, IMP_ 24-hr, by high-vol, 16/Aug/04-17/Dec/04, NOISH-TOT
15.8±2.4
4.1±0.3
3.4±0.3
0.7±0.1
4.7
0.26
31.9±4.9
5.2±0.6
4.1±0.6
1.1±0.1
3.8
0.16
41.3±20.0
8.2±3.4
6.3±2.6
1.9±0.9
3.3
0.20
82.9±17.9
12.1±2.0
9.1±1.0
3.0±1.0
3.2
0.15
31.9±26.0
7±3.9
5.6±3.8
1.4±0.9
3.8±0.8
0.22
Study Location*
Sampling period/Study period/Carbon analysis method
PU campus PU campus
Cao et al., 2004 Cao et et., 2003b
BU urban
Cao et al., 2004 Cao et al., 2003b
HT rural
Duan et al., 2007
HT rural
Hongkong Cao et al., 2004 Cao et al., 2003b
BU urban
HT rural
2-8
Table 2-4 (Cont’d) PM2.5 Mass
TC
OC
EC
OC/EC
TC/PM2.5
24-hr, by high-vol, 16/Aug/04-17/Dec/04, NOISH_TOT 24-hr, by high-vol, 2/Jan/05-8/Mar/05, NOISH_TOT
50.0±33.1
16.3±6.5
11.8±6.4
4.5±1.0
2.7±1.4
0.33
49.9±22.1
14.0±5.8
11.4±5.7
2.6±1.1
4.9±2.1
0.28
TC urban
24-hr, once every sixth day by RAAS, Oct, Dec/02
32.5±14.8
8.3±3.5
6.3±3.4
2.0±1.0
3.2
0.26
Hagler et al., 2006
CW urban
24-hr, once every sixth day by RAAS, Oct, Dec/02 & Mar, June/03, NOISH_TOT
34.3±14.1
8.5±3.3
6.6±3.1
1.9±1.0
3.5
0.25
Hagler et al., 2006
TM rural
24-hr, once every sixth day by RAAS, Oct, Dec/02 & Mar, June/03, NOISH_TOT
28.7±13.8
5.7±2.8
4.9±2.8
0.8±0.5
6.1
0.20
Duan et al., 2007
YL urban
24-hr, by high-vol, 16/Aug/04-17/Dec/04, NOISH_TOT
54.8±38.5
15.6±6.5
11.3±6.5
4.3±0.8
2.6±1.2
0.28
Reference
Study Location
Sampling period/Study period/Carbon analysis method
Duan et al., 2007
TW urban
Duan et al., 2007
TW urban
Hagler et al., 2006
*
PM10 Mass
TC
OC
EC
PU – The Hong Kong Polytechnic University; BU – Hong Kong Baptist University; HT – Hok Tsui; TW – Tsuen Wan; TC – Tung Chung; CW- Central/Western; TM – Tap Mun; YL – Yuen Long
2-9
OC/EC
TC/PM10
2.1.3 What are the levels of ultrafine particles and how do they evolve from primary emissions and secondary aerosol formation? As stated above, ultrafine particles are a new health concern. They also are indicators of nearby combustion emissions and secondary aerosol formation, and were extensively examined in the U.S. supersite program (Bein et al., 2005; Bein et al., 2006; Biswas and Wu, 2005; Cabada et al., 2004; Chakrabarti et al., 2004b; Cho et al., 2005; Dillner et al., 2005; Drewnick et al., 2005; Eiguren-Fernandez et al., 2003; Fine et al., 2004a; Fine et al., 2004b; Garnes and Allen, 2002; Gasparini et al., 2004; Gaydos et al., 2005; Geller et al., 2002; Geller et al., 2005; Gupta et al., 2004; Herner et al., 2005; Jimenez et al., 2003; Jung et al., 2006; Kane et al., 2002; Kim et al., 2002; Kuhn et al., 2005a; Kuhn et al., 2005b; Lake et al., 2003; Laurent and Allen, 2004; Lipsky and Robinson, 2005; McMurry and Woo, 2002; Miguel et al., 2004; Miguel et al., 2005; Misra et al., 2002; Nel, 2005; Park et al., 2005a; Phares et al., 2003; Qian et al., 2007; Rhoads et al., 2003; Sakurai et al., 2005; Sardar et al., 2004; Sardar et al., 2005a; Sardar et al., 2005b; Stanier et al., 2004a; Stanier et al., 2004b; Tolocka et al., 2004a; Tolocka et al., 2004b; Tolocka et al., 2005; Watson et al., 2006a; Watson et al., 2006b; Watson and Chow, 2002a; Weitkamp et al., 2005; Westerdahl et al., 2005; Yu et al., 2004c; Zhang et al., 2004b; Zhang et al., 2005a; Zhang et al., 2004c; Zhang et al., 2005b; Zhao et al., 2005; Zhou et al., 2004c; Zhou et al., 2005a; Zhu et al., 2002a; Zhu et al., 2002b; Zhu et al., 2004; Zhu et al., 2005). Gaydos et al. (2005) analyzed measurements from the Pittsburgh supersite using a NH3-H2SO4-H2O nuclei formation model and demonstrated reasonable agreement between modelled and measured growth rates. Using these results, they estimated that NH3 emission reductions would decrease the frequency of secondary ultrafine particle formation events during summer and winter, with a greater efficiency in summer. The response to changes in emissions of SO2 during the summer is counterintuitive. SO2 emission reductions, on the other hand, would result in SO4= reductions, but might increase the frequency of summertime secondary ultrafine particle formation events. During winter, however, SO2 reductions would decrease the number of ultrafine formation events. Watson et al. (2006a and 2006b) measured 3-407 nm size distributions continuously for two years at the Fresno supersite and correlated them with short-duration (5 min) PM2.5, SO4=, NO3-, black carbon (BC), particle-bound polycyclic aromatic hydrocarbons (PAHs), NOx, CO, O3, and meteorological data to describe ultrafine particle events and their primary and secondary sources. Four types of events were found: 1) 3- to 10-nm morning secondary formation; 2) 10- to 30-nm primary morning traffic; 3) 10- to 30-nm afternoon secondary photochemical; and 4) 50- to 84-nm evening primary home heating, including residential wood combustion. Intense examples of the first type were observed on 29 days, nearly always during the summer. The second type of event was observed on more than 73 days and occurred throughout the year. The third type was observed on 36 days, from spring through summer. The fourth type was found on 109 days, all of them during the winter. Although SO2 emissions in Central California are low, the small residual amounts in gasoline and diesel fuel are apparently sufficient to initiate nucleation events. These were measured in the morning, soon after the shallow surface inversion coupled with layers aloft where nucleation probably was initiated. PM2.5 concentrations were poorly correlated with nanoparticle number.
2-10
Some special studies of ultrafine particles have been conducted in Hong Kong, mostly near roadways (Chan et al., 2007b; Schnell et al., 2006; Wong et al., 2003; Yao et al., 2001; Yao et al., 2005; Yao et al., 2006a; Yao et al., 2006b; Yao et al., 2007a; Yao et al., 2007b). Outside of the tunnels on roadways with heavy traffic, ultrafine particle counts peaked at 10 nm, with a smaller peak at 50 nm. Inside tunnels, 10 nm ultrafine particle counts decreased after an abrupt increase with distance into the tunnel. The 50 nm particles numbers increased from entrance to exit. These large variations indicate the dynamic nature of ultrafine particles that makes it difficult to relate them to health effects. Ultrafine particle number counts are needed in several Hong Kong locations to better understand how human exposure varies spatially and is related to primary vs. secondary emissions sources. Detailed size distributions are also needed to better understand the secondary aerosol formation processes and primary source contributors. The effects of potential reductions in precursor gases on these particles can be evaluated to evaluate how they might affect these formation mechanisms. 2.1.4 What are the characteristics of different particle size fractions (PM mass, number, surface area as a function of temporal and spatial variability) that affect regional haze? What are the spatial and temporal characteristics of visibility (total extinction and its components)? How do chemical constituents contribute to the light extinction budget? As noted earlier, visibility is highly affected by scattering and absorption of fine particles, especially in the PM2.5 size fraction. The relationships between particle size, composition, and radiative properties was studied extensively in the U.S. supersites program and related studies (Adam et al., 2004; Cabada et al., 2004; Carrico et al., 2003; Chen et al., 2003; Chen et al., 2006; Chow, 1995; Chow, 2002a; Chow, 2002b; Chow et al., 2000; Chow et al., 2002a; Chow et al., 2004a; Chow et al., 2006d; Chow et al., 2008b; Eatough et al., 1997; Javitz et al., 1988; Kidwell and Ondov, 2001; Koloutsou-Vakakis et al., 2001; Lowenthal et al., 1995; Lowenthal et al., 2000; Lowenthal and Kumar, 2003; Lowenthal and Kumar, 2004; Lowenthal and Kumar, 2006; Malm et al., 1993; Mathai et al., 1990; Mauderly and Chow, 2008; Mazzera et al., 2001a; Mazzera et al., 2001b; Moosmuller et al., 1998a; Moosmuller et al., 1998b; Park et al., 2006a; Pitchford et al., 2007; Ryan et al., 2005; Schmid et al., 2001; Shah et al., 1984; Sioutas et al., 2000; Sloane et al., 1991; Watson, 2002a; Watson, 2002b; Watson et al., 1990; Watson et al., 1994; Watson et al., 2001b; Watson et al., 2002b; Watson et al., 2008b; Watson and Chow, 1994; Watson and Chow, 2002a; Watson and Chow, 2003; Watson and Chow, 2006). Dry light scattering, which is often measured with a nephelometer after removing liquid water associated with the particles, is highly correlated with PM2.5 mass (Chow et al., 2006c), but the relationship varies with season and location because the particle sizes and compositions are different. These relationships can be established by limited measurement of PM2.5 on filters collocated with the continuous nephelometers. The IMPROVE formula (Pitchford et al., 2007) is a weighted sum of chemical components measured from filters that intends to reproduce light extinction at one or more locations. This “chemical extinction” has the advantage that the contributions to haze can be broken down into component parts of 2-11
sulfates, nitrates, geological material, sea salt, elemental carbon and organic carbon. These contributions may vary with space and time, so it is important to have measurements of these major PM2.5 components at several locations. Several visibility studies have been conducted in Hong Kong (Chang and Koo, 1986; Cheng et al., 1997; Cheng et al., 2006a; Cheung et al., 2005; Choi and Chan, 2002; Deng et al., 2008; Lai and Sequeira, 2001; le Clue, 2004; Lee and Sequeira, 2001; Lee and Sequeira, 2002; Li et al., 2005; Liu et al., 2008a; Sequeira and Lai, 1998; Wang, 2003; Wu et al., 2005; Wu et al., 2007; Yu, 2002), mostly examining existing data to detect trends. Wang (2003) conducted an in-depth study of the meteorological and chemical factors causing visibility degradation in Hong Kong. The main results from this study are summarized below: Seven weather patterns associated with poor visibility in Hong Kong have been identified. Visibility is poorer in the western airport than in urban Victoria Harbor. The IMRPROVE formula (Watson, 2002a) has been applied to apportion the different components of air pollution to light extinction coefficient after a preliminary evaluation of formula’s applicability to Hong Kong’s situation. PM2.5 ammonium sulfate was the largest contributor to visibility degradation in Hong Kong, on average accounting for 51% and 33% of the total light extinction in the upwind rural and urban areas, respectively. Organic carbon and elemental carbon contributed to 17% (in rural area)-21% (in urban area) and 12% (in rural area)-26% (in urban area), respectively. Under the prevailing easterly and northeasterly flows, background PM2.5 concentrations (measured at Hok Tsui) accounted for about 79% of the total mass measured in urban Tsuen Wan. The regional sources contributed nearly 100% of the PM2.5 sulfates. Although the IMPROVE formula has been tested at a site in western Hong Kong, it’s applicability in other parts of PRD needs to be tested for a wide range of conditions. Continuous particle size measurements, similar to those needed to understand ultrafine particles, but extending the size distribution into the coarse particle size range, are needed at one or more locations within the haze to evaluate changes during the 24-hour filter monitoring period. Supetsite studies would be needed to derive chemically specific size distributions and to determine growth factors for Hong Kong aerosols with changes in composition and relative humidity. The Supersite network can be used to: determine chemical compositions of specific size fractions; monitor the total mass concentration of PM2.5 and; identify sources of visibility impairment in the region. The contribution of human activities to visibility impairment is accessed by monitoring the concentrations and aerosol compositions of PM in this region and deriving so called light extinction coefficients and visibility ranges from these measurements. The proposed Supersite
2-12
programs will definitely enhance the current visibility measurement in Hong Kong by providing more consistence, more frequently aerosol data for visibility evaluation in Hong Kong. To address the problems with the filter-based techniques, U.S.EPA has implemented a number of PM Supersite in USA where more detailed research is addressing the scientific uncertainties associated with fine PM. Supersite program can provide more time-resolved measurement for visibility measurement which will provide more daylight hours and night time hours. (visibility is most important during the daylight hours when the RH is typical lower than during the night. bext contributions from different chemical components will differ between daylight and nighttime hours. Much shorter average time Chemical composition as well as size-resolved mass concentration. Most of existing HKEPD field programs measure mass concentrations of major aerosol species sampled with filters. But with the real time instrument of HK supersite network, the b ext can be more accurate calculated with most of the real time analyzers. 2.1.5 What synoptic and local meteorological conditions are associated with observed PM2.5 and O3 standard exceedences in HK? Which variables are most important for each situation? Excessive O3 and PM2.5 levels are affected by different meteorological scales and conditions in the HK and PRDR. Large regional SO4= clouds have been observed in the eastern U.S. during summer that were studied at supersites. Prolonged stagnation events with shallow (20 m) surface layers underneath a ~500 m valleywide layer were found in California’s central valley (Watson and Chow, 2001b). These layers allowed accumulation of locally generated pollutions at night when the layers were decoupled, with mixing of these primary pollutants aloft soon after sunrise when the layers mixed. This mixing also brought NH4NO3 to the surface that had formed and circulated aloft throughout the valley (Watson and Chow, 2002a). Several attempts have been made to associate high pollutant levels with synoptic and local weather conditions in the HK and PRDR (Chan et al., 1998a; Chan et al., 2003a; Chan and Chan, 2000; Chan and Kwok, 2000; Cheng and Lam, 1999; Cheung et al., 2005; Chin, 1986; Chung et al., 1999; Deng et al., 2006; Ding et al., 2004; Fang et al., 1999; Feng et al., 2007; Huang et al., 2005; Jackson, 2008; Jeary A.P., 2008; Lam et al., 2005; Liu et al., 2001; Liu and Chan, 2002a; Lo et al., 2006; Qu et al., 2007; Shun and Chan, 2008; Wai and Tanner, 2005; Wang and Lu, 2006; Wang et al., 1998; Yan, 2007; Yeung et al., 1990; Zhang and Zhang, 1997; [Anon], 1920). An understanding of the local and regional circulation, including the complex interaction between the local land see breeze (LSB) circulation and the urban environment in the PRD, is critical. Although researchers suspect the LSB circulation system has a significant influence on air flow pattern and may affect the air quality in this region (Fung et al., 2005; Lam et al., 2006; Liu and Chan, 2002b), the physical mechanism of this phenomenon has never been fully understood or quantified. In calm weather days, 78% of the RSP in Hong Kong are contributed by regional sources outside Hong Kong, of which 2-13
surface emissions from Guangzhou are the largest contributor. On the other hand, in typical-wind condition with northeasterly surface wind, 71% of the RSP are contributed by regional transport. The climate of Hong Kong is controlled by the East Asian monsoon. During late summer and early fall (i.e., August and September) when O3 episodes are most common, significant weather systems affecting Hong Kong includes tropical cyclones / typhoons over the South China Sea or the Northwestern Pacific, regional high pressure systems (anticyclones) to the north over mainland China, and low pressure systems (troughs) to the south and east over the South China Sea. Studies have shown that some of these systems can have significant impact on O3 episodes in Hong Kong (Chan and Chan, 2000; Huang et al., 2005; Lee and Hills, 2003). However, a systematic and quantitative evaluation on how these weather systems impact on the occurrence of O3 episodes in Hong Kong is still lacking. In particular, most of the earlier air quality studies are based on case studies, but it is often hard to gauge the representative of individual case. As outlined in Section 3, Hong Kong has an extensive network of surface and upper-air meteorological stations, although the data need to be unified for a comprehensive analysis. It lacks the corresponding air quality measurements to exploit this data base. Data need to be obtained for shorter averaging times than one hour (2 to 5 minutes is preferable) to detect plume impacts from industrial sources and vehicle exhaust. More measurements are needed at elevated locations to quantify reservoirs of pollutants stored aloft that may mix down when surface and upper air layers couple. Concurrent SO2 and SO4= measurements are needed under clear air and within fogs to elucidate the dry air and aqueous transformation mechanisms for coal-fired power plant plumes from Hong Kong stations as well as more distant power generators in the PRD. 2.2
Source Questions
Answers to these questions will improve emissions estimates, determine source contributions to O3 and PM2.5 at different times and places, and estimate how emission reductions that reduce levels for one pollutant might increase levels of other pollutants. Improved answers to these questions will allow decision-makers to better estimate how far emissions reductions must go before improvements in secondary air pollutants are observed. It will also provide the technical justification for potentially expensive and intrusive emission reduction strategies that will affect citizen’s costs and lifestyles. 2.2.1 What are the major source types and their emittants that contribute to excessive O3 and PM2.5? How well are their emissions inventoried? Table 2-5 summarizes current and projected emission rates for Hong Kong and the larger PRD. PRD emission data are limited and do not have the source sector resolution available for Hong Kong emissions. Given the large PRD economic growth over the last decade, the 2003 emissions probably do not represent the current or future situation. Projections for 2010 show a substantial decrease Projections for Hong Kong emissions
2-14
Hong Kong SO2 emissions are dominated by coal-fired electrical generators that currently lack SO2 scrubbers. Hong Kong SO2 emissions are projected to be 90% removal efficiencies expected of modern SO2 controls (Davis, 2000). PRD governments have also committed to installing power plant SO2 scrubbers, but the reductions between 2003 and 2010 do not reflect the reduction efficiencies of modern pollution control technology. These reductions should result in important and detectable reductions in both SO2 and SO4= levels, as was found in response to major SO2 reductions in the eastern U.S. (Malm et al., 2002; Watson et al., 2002a). Hong Kong NOx emissions are expected to be ~15% of the total PRD emissions by 2010 and only slightly less than 2006 emissions. Hong Kong’s coal-fired power stations are the largest component of total NOx emissions, followed by vehicle emissions and shipping. On-road vehicle emissions are expected to decrease as EURO IV standard vehicles penetrate the existing fleet, thereby increasing the importance of power generation and navigation emissions in the future. Primary PM10 emissions are dominated by PRD emissions, with Hong Kong contribution only ~2% of the total. It is not clear how the PRD emission estimates is made or what is included. It is well known that U.S. fugitive dust emission factors overestimate PM10, especially when scaled up larger areas (Watson et al., 2000; Watson and Chow, 2000). Primary PM10 emissions from power generation exceed those from motor vehicle exhaust, implying that approximately equal contributions from these sources would be measured at urban- and regional-scale monitors. This is at odds with point source contributions quantified by receptor modelling (Chow and Watson, 2002) and short duration measurements at U.S. supersites (Bein et al., 2006; Bein et al., 2007; Eatough et al., 2007; Lee et al., 2006b; Ondov et al., 2006; Paatero and Hopke, 2002; Park et al., 2005b). Table 2-5. Hong Kong annual emissions estimates for 2006 (EPD, 2008b) and for the Pearl River Delta Region for 2003 and 2010 (Guangzhou EPB and EPD) in 1000kg/yr. Source Category
SO2
NOx
RSP (PM10)
VOC
CO
HK Electricity Generation
66,000
41,800
1,860
416
3,370
HK Road Transport
956
21,800
1,810
8,080
62,300
HK Navigation HK Civil Aviation HK Other Fuel Combustion1 Non-combustion2 Total Hong Kong (2006) Hong Kong (2003)3
3,920 294 2,660 NA 73,900 91,800
16,700 5,020 9,530 NA 94,800 110,900
499 21 925 750 5,860 7,100
304 261 998 31,200 41,200 44,300
2,290 2,020 4,260 NA 74,200 NA
837,000
758,800
532,700
517,300
NA
30,200
92,800
4,700
31,000
NA
461,600
596,500
212,200
209,200
NA
PRD Region (2003)3 3
Hong Kong (2010 Projection)
3
PRD Region (2010 projection)
2-15
1
Industrial, Commercial, Domestic and Off-road Transport.2006 SO2 emissions are based on projected energy data. 2 RSP: includes Quarrying and Construction Aggregate Processing, Type, Brake and Road Surface Wear, VOC: mainly consists of Consumer Products, Paint and Printing VOC. 3 Mid-term review report with updated emissions data from PRD and HK for 2003
According to Table 2-5, ~15% of PRD VOC emissions will come from Hong Kong in 2010. This represents a ~25% reduction since 2006. Most VOC emissions are estimated to derive from non-combustion sources such as solvent use, surface coatings, dry cleaners, and other activities. VOC emission factors are notoriously inaccurate, and these emission rates are uncertain. The compositions of VOCs from different sources are also needed to determine source contributions and their effects on O3 formation, secondary nitrate, and secondary organic aerosol. There are limited information for several important source categories, such as primary PM2.5 emissions from biomass burning and cooking, non-road engines, and cold starts and poorly tuned engines. The navigation emissions are highly suspect, as are few emission factors for shipping emissions and the types of fuels and engines are largely uncharacterized. Supersite measurements are needed that can provide some verification of the important sources and how they vary with meteorology. High-time resolution measurements coupled with CO2 measurements to indicate the presence and absence of combustion plumes are useful in this regard. High-time resolution SO2, SO4=, and PM2.5 mass measurements can detect plumes from ships and power stations and correlate them with wind directions to determine their impacts. Measuring PM2.5, RSP and VOC source marker compounds can permit receptor model source contribution estimates for PM2.5, RSP, and VOCs (Watson et al., 2001a; Watson et al., 2002a; Watson et al., 2008a). 2.2.2 How accurate and representative are emission factors, activity levels, source profiles, and temporal distributions for these emissions? Several studies of Hong Kong and PRD emissions rates and profiles have been conducted (Cai and Xie, 2007; Chan et al., 2007b; Chan and Ning, 2005; Cheng et al., 2006b; Chiu et al., 2006; Dobridge et al., 2001; EPD, 2008b; Fung, 1989; Guangdong EPB and EPD, 2007; He et al., 2004; HKPolyU, 2005; Ho et al., 2006a; Ho et al., 2006d; Ho, 1990; Huang et al., 2004; Huang et al., 2006b; Hung et al., 2002; Hung et al., 2007; Lee and Wang, 2004; Li et al., 2007; Li et al., 2008; Ma et al., 2005; Ning and Chan, 2007; Sheng et al., 2008; Tan et al., 2004; Tang and Wang, 2006; Tian et al., 2008; Tsai et al., 2005; Wang et al., 2004; Wang et al., 2005b; Wang et al., 2006; Wang et al., 2007; Yang et al., 2008; Zeng et al., 2007; Zhang et al., 2001; Zhi et al., 2008; Zou et al., 2000), but it is not clear how representative these are of current conditions or how they are incorporated into EPD’s emission inventories. The U.S. supersites program presented opportunities to evaluate emissions profiles and to estimate emission factors based on top-down inverse modelling (Braun et al., 2008; Chow et al., 2007b; Fitz et al., 2004; McGaughey et al., 2004; Murphy and Allen, 2005; Simon et al., 2008; Weitkamp et al., 2005; Zhang et al., 2005a; Zhu and Hinds, 2005). Murphy and Allen (2005) applied time series analyses to compare average annual flow rates for air pollutant emissions with those released during reported emission events. The results indicate that the 2-16
magnitude and frequency of VOC emissions are an important element in accurately reflecting O3 precursor emission patterns in the Houston-Galveston area. More than 50% of the reported episodic (event) emissions of VOCs were composed of ethene and about a third was associated with propene. Isomers of butene and 1,3-butadiene were also included in these events, which were of less than 24-hour duration. The amount of VOCs released ranged from 100 to more than 50,000 kg from chemical manufacturing. If the emission variability occurs at times and locations where atmospheric conditions were amenable to O3 formation, the chemical industry emission events might result in higher O3 levels. Peak O3 levels could increase by as much as 100 ppb for a large VOC emission event. A top-down approach that estimates contributions from combustion of carbon-containing fuels is possible when simultaneous CO2 measurements are available along with air quality measurements. CO2 is the main product of complete combustion, and ratios of increments in combustion-related pollutants (e.g., PM2.5, CO, SO2, NOx) to CO2 can be used to develop urban-scale fuel-based emission factors (Moosmuller et al., 2003). The Hong Kong supersites program could provide the CO2 at regional-, urban-, and source-dominated sites to provide this top-down estimate of combustion emissions that could be used to refine the emissions inventories. Continuous VOC measurements at urban-scale and source-oriented sites could develop emissions profiles and potentially identify non-inventoried sources as was done at the Houston supersite. 2.2.3 How much do the major source types contribute to primary PM2.5 and directly emitted precursors of secondary O3 and PM2.5? U.S. supersites conducted several source apportionment studies for PM2.5 and precursor gases, especially VOCs (Anderson et al., 2006; Bein et al., 2006; Braun et al., 2008; Buzcu et al., 2006; Chen et al., 2007; Chow et al., 2007b; Dutkiewicz et al., 2006; Eatough et al., 2007; Fine et al., 2004b; Fraser et al., 2003a; Fraser et al., 2003b; Gorin et al., 2006; Ito et al., 2004; Jaeckels et al., 2007; Ke et al., 2008; Kim et al., 2004; Lane et al., 2007; Lee et al., 2006b; Lee and Hopke, 2006; Lemire et al., 2002; Li et al., 2004; Liu et al., 2006; Marmur et al., 2007; Millet et al., 2005; Ogulei et al., 2005; Ogulei et al., 2006; Paatero and Hopke, 2002; Park et al., 2005b; Park et al., 2006b; Pekney et al., 2006a; Pekney et al., 2006b; Pekney et al., 2006c; Polidori et al., 2006; Robinson et al., 2006a; Robinson et al., 2006b; Robinson et al., 2006c; Subramanian et al., 2006; Subramanian et al., 2007; Vayenas et al., 2005; Watson et al., 2008a; Zhao et al., 2004b; Zheng et al., 2007; Zhou et al., 2004a; Zhou et al., 2004b; Zhou et al., 2004c; Zhou et al., 2005a; Zhou et al., 2005b). Table 2-6 summarizes results from 10 Hong Kong and PRD source apportionment studies that are detailed in Table A-6 of Appendices A. Secondary aerosol is one of the largest contributors, accounts for 20-38% of PM10. (Cheng, 2007; Yuan et al., 2006a) and 13% of PM10-2.5 (Cheng, 2007). PM10 (NH4)2SO4 are a factor of 2 to 4 higher than those of NH4NO3. Vehicle exhaust is also one of the major contributor, account for 7% TSP (Fung and Wong, 1995), 1-26 % of PM10 (Lee et al., 1999; Yuan et al., 2006a) and 11% of PM10-2.5 (Cheng, 2007). Only one study (Cheng, 2007) resolved vehicle exhaust into 2/3 of diesel-fuel vehicles and 1/3 of gasoline vehicles. Fuel combustion varied from 0.2-28%, whereas coal combustion only reported in two studies with 2-17
17% in TSP (Fung and Wong, 1995) and 13% in PM10 (Cheng, 2007). As expected, crustal maerial contribution varies by particle sizes, account for 25% TSP, 3-10% of PM10, and 37% of PM10-2.5. Marine aerosol accounts for 11 to 22% (Yuan et al., 2006) of PM10. Regional combustion accounts for 8 to 22% (Yuan et al., 2006) of PM10. These limited studies varied in sampling locations, monitoring methods, sampling periods, analytical methods, and source apportionment models, with only one roadside study, segregated into PM2.5 and PM10-2.5. Site specific and representative source profiles are often lacking to obtain reliable source contribution estimates. Hong Kong supersites need to measure more specific markers for primary PM2.5 and VOC emissions. Organic compounds typical of secondary organic aerosol must also be quantified to combine with the ultrafine particle nucleation event data to better understand the contributions to secondary aerosol.
2-18
Table 2-6.
Summary of PM source apportionment studies in Hong Kong
Locations/Reference Model
Crustal
Vehicle emission
Industry
Marine
Regional combustion Secondary aerosol
Others
Unidentified PM (µg m-3)
23
NA
material PM Source Contribution Estimates (% PM mass) Diesel-fueled
Gasoline-fueled
vehicle
vehicle
Coal-
Fuel
oil Smelters Fresh
combustion burning
sea salt
Coal
salt
fired/SOC SO4=
NO3-
NA
NA
waste/
Secondary
Secondary
Aged sea Biomass/
vehicle Five urban sites in Principal factor 25 New including (SH),
7
17
28
NA
NA
NA
NA
NA
(Construcion
Territories, analysis San Hui (Fung
TM,
work)
and
Kelvin Wong, 1995)
28.6 (TSP, n=40)
a
Tower (KT), Hung Shui Kiu (HSK), and Au Tau (AT) Eleven AQMN sites, PMF-CMB including a roadside (Lee site (MK) and nine 1999)
et
6
0.8
NA
0.2
1.2
al.,
b
urban site (CW, JB, TP, SSP, ST, TST, HKS, KC, KT, TW, MK)1
2-19
6.9
14.3
NA
NA
37.8
NA
1.4
31.4
15.2
(particulate
(PM10,
Br+Cu)
n=1516)
Table 2-6 (Cont’d) Locations/Reference Model
Crustal
Vehicle emission
Industry
Marine
Regional combustion Secondary aerosol
Others
Unidentified PM (µg m-3)
NA
2
material PM Source Contribution Estimates (% PM mass) Diesel-fueled vehicle
Gasoline-fueled vehicle
Coal-
Fuel
oil Smelters Fresh
combustion burning
sea salt
Aged sea Biomass/ salt
waste/
Coal
Secondary =
Secondary
fired/SOC SO4
NO3-
7
12
vehicle Ten
AQMN
sites, PMF-CMB
5
25
NA
1
NA
4
10
9
25
site (MK) and nine 2006a)
36.4 (PM10,
including a roadside (Yuen et al., c
n=1950)
urban sites (CW, KC, KT, SSP, ST, TC, TP, TW, and YL) Ten
AQMN
sites, UNMIX
7
26
NA
3
NA
4
18
20
22
NA
NA
site (MK) and nine 2006a)
36.4 (PM10,
including a roadside (Yuen et al., c
n=1950)
urban sites (CW, KC, KT, SSP, ST, TC, TP, TW, and YL) Ten
AQMN
sites, PMF-CMB
including a roadside (for
3
40
NA
1
NA
13
4
4
19
5
NA
4
36.4
n=1950)
site (MK) and nine (Yuen et al., urban sites (CW, KC, 2006a)
7
(PM10,
summer)
c
KT, SSP, ST, TC, TP, TW, and YL)
2-20
Table 2-6 (Cont’d) Locations/Reference Model
Crustal
Vehicle emission
Industry
Marine
Regional combustion Secondary aerosol
Others
Unidentified PM (µg m-3)
NA
2
material PM Source Contribution Estimates (% PM mass) Diesel-fueled
Gasoline-fueled
vehicle
vehicle
Coal-
Fuel
oil Smelters Fresh
combustion burning
sea salt
Coal
salt
fired/SOC SO4=
NO3-
9
17
waste/
Secondary
Secondary
Aged sea Biomass/
vehicle Ten
AQMN
5
sites, PMF-CMB
including a roadside (for
17
NA
0.4
NA
4
7
13
26
(PM10,
winter)
n=1950)
site (MK) and nine (Yuen et al., urban sites (CW, KC, 2006a)
36.4
c
KT, SSP, ST, TC, TP, TW, and YL) PU roadside site
(Cheng, 2007) PU roadside site
10
PMF-CMB
13
13
8
NA
0
NA
NA
NA
20
NA
10
55.5 (PM2.5, n=40)
37
PMF-CMB (Cheng, 2007)
26
d
11
NA
NA
NA
17
NA
NA
13
NA
22
d
14.4 (PM10-2.5, n=40)
PU roadside site
EV-CMB (for 1
72
10
NA
NA
1
NA
NA
14
air mass from
2
(Biomass -8
39.9
burning)
n=40)
9
78.3
(PM2.5,
south) (Cheng, 2007)d,e PU roadside site
EV-CMB (for 1
55
14
NA
NA
1
NA
NA
20
(Biomass -6
burning)
air mass from north) (Cheng, 2007)d,e
2-21
n=40)
(PM2.5,
Table 2-6 (Cont’d) Locations/Reference Model
Crustal
Vehicle emission
Industry
Marine
Regional combustion Secondary aerosol
Others
Unidentified PM (µg m-3)
material PM Source Contribution Estimates (% PM mass) Diesel-fueled
Gasoline-fueled
vehicle
vehicle
Coal-
Fuel
oil Smelters Fresh
combustion burning
sea salt
Aged sea Biomass/
Coal
salt
fired/SOC SO4=
waste/
Secondary
Secondary NO3-
vehicle a
Observable: Zn, Pb, Mg, Ca, Na, S, Mn, Se, As, V, Cd, Sr, Ba in TSP
b
Observable: SO4=, Cl-, Br-, NH4+, Na+, K+, Al, Mg, Ca, Mn, Fe, Zn, Ba, Cd, Cu, Ni, Pb, and V in PM10
c
Observable: OC, EC, SO4=, NO3-, Cl-, Br-, NH4+, Na+, K+, Al, Mg, Ca, Mn, Fe, Zn, Ba, Be, Cd, Cr, Cu, Ni, Pb, and V in PM10
d
Observable: OC, EC, SO4=, NO3-, Cl-, NH4+, K+, forty elements from Na to U in PM2.5
e
Source profile: Vehicle emission (HKEPD, 2004)), Paved soil dust (Ho et al., 2003b), Tire wear (Hildemann, et al, 1991), Brake lining (Ondov, et al., 1982), Cooking fumes (HKEPD, 2006), Biomass burning (Core et al., 1989), Residual oil combustion, marine
aerosols (Watson, 1979), Coal combustion from plants (Klein, et al., 1975), Secondary aerosol 1
Central Western (CW), Junk Bay (JB), and Tai Po (TP) in residential areas; Sham Shui Po (SSP), Shatin (ST) and Tsim Sha Tsui (TST) in mixed commercial and residential areas; Hong Kong South (HKS) in a mixed residential and industrial area; Kwai Chung (KC),
Kwun Tong (KT) and Tsuen Wan (TW) in industrial area; Mongkok (MK) in a roadside station.
2-22
2.2.4 What are the major source areas affecting HK PM2.5 in different seasons? From HKEPD air monitoring data (2001-2007), PM10 levels at different EPD stations are very similar. PM10 levels increased from 2001 to 2004 and then standed steady during 2004-2007 (Figure 2-1). The examination of seven given PM10 chemical composition data revealed that OC, nitrate, sulfate, and ammonium followed the same patten as PM10 mass concentrations. EC levels at MK roadside station are higher than other stations due to the diesel emssions from MK roadside station. In general, SO2 levels in EPD stations show a slightly reduction trend during 2001-2007. However, the sulphate did not show any reduciton trend, with a stable level during this periods (2001- 2007). Sum of carbon (OC/EC) and ions (SO4=, NO3- and NH4+) account for 60% of PM10 in HK. Annual average PM10 in HK during 2003-2005 were 20% higher than those before 2003. Seasonal PM10 levels are generally high in winter and low in summer. Previous studies (Cheng, 2007) found that very high percentage of OC in PM0.1 (85%) as compared to OC in PM2.5 (57%) and in PM10-2.5. Sulfate is dominated in the PM2.5 mode while nitrate, sea salt and crustal materals are dominated under PM10-2.5.
Figure 2-1. Monthly variations of PM10 in seven HKEPD monitoring stations from 2001-2007. Figure 2-2 shows the variations of PM10 in four different seasons at YL and TW sites during 2001-2007. The PM10 levels are higher in winter and lower in summer. We have also reviewed the PM10 mass and detailed chemical compositions at CW, KT, SSP, TC and MK. The seasonal pattens at each site are very similar with that of YL or TW. The major aerosol compositions at YL/TW sites are organic matter (OC × 1.4), crustal materials and sulfate. The findings are similar to other studies which carbonaceous aerosols accounted for approximately half of the PM10 mass and dominated the seasonal and spatial variations of PM10. A number of past PM2.5 studies (Cao et al., 2006; Ho et al., 2003a; Leung, 1999) 2-23
100 YL
TW
80 60 40 20
Unidentified Crustal material (Fe*20) Sea salt Ammonium Nitrate Sulfate EC OM (OC*1.4)
fa ll w in ter
in ter sp rin g su m m er
w
fa ll
0 sp rin su g m m er
3
PM10 mass concentration (µg/m )
concluded that OC, EC and sulfate were the most abundant PM2.5 species. OC has sources of local vehicular exhaust while it is also influenced by polluted air masses that transport from the Asian Continent. The wintertime OC is two times higher than the summertime OC in HK (Ho et al., 2003a; Yu et al., 2004a) because of the transported carbonaceous aerosol from northern China in winter. EC mainly originates from local vehicle exhaust and ship emissions. The seasonality of EC is much weaker and variable among the monitoring sites. EC seasonality was found found to be dependent on the distance from and relative location of nearby road (Louie et al., 2005) and the city’s container port (Yu et al., 2004a).
Figure 2-2. Variations of PM10 in four different seasons at YL and TW sites during 2001-2007. 2.2.5 How does pollution transport on local and regional scales interact with vertical mixing to affect the diurnal variations in pollutant levels? Aerosol vertical distribution is an important piece of information to improve aerosol retrieval from satellite remote sensing. Aerosol extinction coefficient profile and its integral form, aerosol optical depth (AOD), as well as atmospheric boundary layer (ABL) height and haze layer height can be derived using lidar measurements. In Hong Kong, about 64% of monthly mean aerosol optical depths were contributed by aerosols within the mixing layer (with a maximum [~76%] in November and a minimum [~55%] in September) revealing the existence of large abundance of aerosols above ABL due to regional transport (He et al., 2008). The significant differences in diurnal and monthly variations between the AOD and surface aerosol extinction coefficients, especially in the winter months, indicate the importance of understanding aerosol vertical distribution in order to better establish the relationship between these two parameters. The surface aerosol extinction coefficients estimated on the basis of AOD and ABL height from lidar backscatter profiles reveal somewhat better correlation (r2 = 0.65) with measured ones than those with AOD only. An 2-24
even higher correlation coefficient (r2 = 0.78) with a slope ~0.82 and intercept ~0.03 km-1 between the measured and estimated surface extinction coefficients obtained using AOD, ABL height, and haze layer depth suggests that two-layer aerosol model could be very useful to provide continuous and accurate estimate of aerosol extinction coefficient near the surface for the urban pollution research. In the future, it is expected that by combining meteorological information with lidar measurements, more detailed analysis on aerosol distribution and evolution processes could be obtained (He et al., 2008). 2.3
Control Strategy Questions
Since O3 and portions of PM2.5 are of secondary origin and are not linearly related to emissions, it is important to estimate how emission reductions that benefit one pollutant might affect other pollutants. The following questions address limiting precursors and explain how supersites measurements can estimate how emission controls will affect the secondary pollutant levels. 2.3.1 What are the limiting precursors (e.g., VOC, NOx, NH3, HNO3, SO2) for PM2.5 and O3 formation in Hong Kong? Tropospheric ozone is a photochemical reactive oxidant mainly formed by chemical reactions involved by VOCs and NOx in the presence of sunlight at ground level. Several studies have been conducted to identify the limiting precursor in the formation of ground-level O3 in Hong Kong. They used methods ranging from analysis of morning ratio of VOC to NOx and the O3 formation potential of VOCs (Guo et al., 2007; So and Wang, 2004) to elaborate observation-based (Zhang et al., 2004a; Zhang et al., 2007) and emission-based modeling (Huang et al., 2005; Lam et al., 2005; Wei et al., 2007). VOCs are found to be the limiting precursors in most of the cases, while increasing NOx would reduce the peak O3 concentrations. So and Wang (2004) also evaluated the maximum incremental reactivity (MIR) of the VOCs in urban and rural coastal areas of Hong Kong. Among VOCs, reactive aromatics, toluene, xylene, and ethylbezene are the dominant VOCs. However, naturally emitted isoprene could be important during hot summer when the emission from trees and other vegetations is the strongest. Zhang et al. (2007) applied an observation-based model (OBM) to atmospheric measurements of O3, CO, NMHCs, and NO conducted in urban Hong Kong and Tai O and suggested that in autumn the O3 formation was limited by VOCs, which were dominated by the contribution from reactive aromatics. Since previous data were mostly collected in or near urban areas, more studies are needed to examine the role of biogenic sources in the formation of O3 (and PM2.5). Also, analysis of cases in different seasons/meteorological conditions will be needed in order to derive a more robust conclusion on the key precursors that control the formation of O3 in Hong Kong. 2.3.2 How will reducing the O3 precursors affect PM2.5 levels? Limited studies have suggested that O3 formation in Hong Kong is largely VOC limited, however, this may varies by season and by locations. Thus reducing VOC emissions by HKEPD’s current control measures may have positive effect on decreasing peak O3 2-25
concentrations. Reactive aromatics – the dominant VOCs leading to the formation of O3 – may also contribute significantly to the OC fraction of PM2.5 after they are oxidized in the atmosphere. Thus controlling the reactive aromatics is also expected to benefit lowering PM2.5, although it is difficult to quantify the effect at present. NOx control would probably lead to an increase in peak O3 concentrations in urban areas, but may reduce O3 levels in downwind rural areas. Table 2-3 shows that nitrate comprise of only a small (100 Transport Phenomena HKUST (UST) Regional-Scale >100 Transport Phenomena Tai Mo Shan (TMS) Elevated Regional 100-1000 Elevated Transport Scale Characteristics (958 m above mean sea level) a See Tables 3-1 and Table 3-2 for detailed Site location and measurement b denotes the Anchor Sites, others are satellite sites.
Zone Representation
120
3
PM10 mass concentration (µg/m )
Site Name (Code)
Proposed locations for the Hong Kong Supersites
a
Unidentified Crustal material (Fe*20) Sea salt Ammonium Nitrate Sulfate EC OM (OC*1.4)
100 80 60 40 20 0 YL
TW
SSP
KT
TC
MK
Figure 3-9. Average PM10 and chemical composition for the several sites in Hong Kong (2001-2007)
3-27
In addition, according to U.S. EPA, the PAMS network should be fashioned to supply measurements which will assist in understanding and solving ozone nonattainment problems. U.S.EPA has determined that for the larger areas, the minimum network which will provide data sufficient to satisfy a number of important monitoring objectives should consist of four sites: Site 1 - Upwind and background characterization site. These sites are established to characterize upwind background and transported ozone and its precursor concentrations entering the area and will identify those areas which are subjected to overwhelming incoming transport of ozone. The 1 Sites are located in the predominant morning upwind direction from the local area of maximum precursor emissions and at a distance sufficient to obtain urban scale measurements. Typically, these sites will be located near the upwind edge of the photochemical grid model domain. Site 2 - Maximum ozone precursor emissions impact site (YL). These sites are established to monitor the magnitude and type of precursor emissions in the area where maximum precursor emissions representative of the MSA/CMSA are expected to impact and are suited for the monitoring of urban air toxic pollutants. The 2 Sites are located immediately downwind (using the same morning wind direction as for locating Site 1) of the area of maximum precursor emissions and are typically placed near the downwind boundary of the central business district (CBD) or primary area of precursor emissions mix to obtain neighborhood scale measurements. Additionally, a second Site 2 may be required depending on the size of the area, and should be placed in the second-most predominant morning wind direction. Site 3 - Maximum ozone concentration site (YL and TC). These sites are intended to monitor maximum ozone concentrations occurring downwind from the area of maximum precursor emissions. Locations for 3 Sites should be chosen so that urban scale measurements are obtained. Typically, these sites are located 10 to 30 miles from the fringe of the urban area. Site 4 - Extreme downwind monitoring site (TC). These sites are established to characterize the extreme downwind transported ozone and its precursor concentrations exiting the area and will identify those areas which are potentially contributing to overwhelming ozone transport into other areas. The 4 Sites are located in the predominant afternoon downwind direction from the local area of maximum precursor emissions at a distance sufficient to obtain urban scale measurements. Typically, these sites will be located near the downwind edge of the photochemical grid model domain. The Yuen Long (YL) and Tung Chung (TC) sites represent newly developed suburban residential/commercial neighborhood in the northwestern and western HK. The population density is much lower than those of well-established neighborhood like the TW site. The YL and TC sites started operation in 1995 and 1999, respectively. Both sites experienced elevated O3 and PM2.5. Enhanced photochemical measurements at these two sites provide opportunities to establish relationships between O3 and PM2.5.
3-28
The Mong Kok (MK) satellite site is selective to address human exposure in a street canyon. This site is situated two kilometers west of Kowloon Peninsula and southwest of the TW site, right next to a bus stop with heavy vehicle and pedestrian traffics, gas refueling station, restaurant cooking, in addition to the urban mixture. The Hok Tsui (HT) site and Hong Kong University of Science and Technology (UST) site are selected to represent regional transport phenomenon. The HT site is located in a semi-rural area and at an elevation of 60 m above the sea level and at the southeast end of Hong Kong Island. The UST site, located inside the HKUST campus at Clear Water Bay, is in the eastern coast of the Kowloon Peninsula. The sampling site will be located on the roof of an academic building (20 m above the ground and 65 m above sea level) ~500 m from the coast, and ~500 m from a lightly trafficked road and ~5 kilometers from the closed industrial area. Prevailing easterly and south easterly wind in winter allow the study of pollution transport at these two regional transport sites. The Tai Mo Shan (TMS) is a newly established (since 2005) elevated (958 m above ground) site. This site is located on a mountain top in a midst of forest with little or no anthropogenic sources (e.g., cooking, marine ship, traffic). Under most stagnant winter period, the TMS will be above the inversion layer. The enhanced measurement at this site provides information on vertical concentration gradients of air pollution and the extent of downwind mixing.
3-29
4.
MODERNMEASUREMENT TECHNOLOGIES
4.1
Particle and Visibility Measurements
As part of the U.S. EPA Supersites Program and related studies in the past decade, much progress has been made along the goals and objectives of the measurement method evaluation (Brock et al., 2004; Chow et al., 2004b; Chow and Solomon, 2006; Geller and Solomon, 2006; Middlebrook et al., 2004; Ondov et al., 2004; Pandis et al., 2005; Solomon et al., 2003a; Solomon and Allen, 2004; Stanier and Solomon, 2006; Wittig and Solomon, 2006). Building from the previous review (Chow, 1995; Fehsenfeld et al., 2004; McMurry, 2000; Solomon et al., 2003a; Wilson et al., 2002). Chow et al (2008a) documented and evaluated recent developed integrated and continuous PM measurements focusing on the past 7-10 years. In addition to measurement methods for compliance with NAAQS, Chow et al. (2008a) summarized measurements related to aerosol properties (e.g., number concentrations, particle density, light extinction/scattering/absorption efficiencies, refraction index, particle surface and size distributions, gas/particle equilibrium, and chemical/physical compositions). Data from these measurement technologies are used in source and receptor models to : 1) identify and quantify source contributions (Watson et al., 2008a); 2) better understand atmospheric process and secondary particle formation (Fine et al., 2008); 3) evaluate site zone of influences and regional transport of aerosol (Turner and Allen, 2008); and 4) test model components (e.g., the inorganic thermodynamic algorithms) and evaluate air quality modeling system (e.g., Community Multiscale Air Quality [CMAQ], Comprehensive Air Quality Model [CAMx] (Russell, 2008). 4.1.1
Advances in Filter-Based Integrated Samplers
As of 2008, the U.S. EPA designates 9 Federal Reference Method (FRM) and Federal Equivalence Method (FEM) with PM10 inlets (e.g., Louvered low-volume [16.7 L/min] PM10 inlet, high-volume Size-Selective inlet). There are 17 FRMs and FEMs with PM2.5 inlets (e.g., Well-Impactor Ninety-Six [WINS], Sharp Cut Cyclone [SCC], and Very Sharp Cut Cyclone [VSCC]) (Watson and Chow, 1993; Watson and Chow, 2001a). Field test was conduct to inter-compare the capacity and comparability among different PM2.5 inlets (Chow et al., 2008a). Very Sharp Cut Cyclone (VSCC [BGI, Inc], 2.5 µm inlet with Geometric Standard Deviation of 1.16) became approved FEM inlet. Laboratory testes showed that the VSCC cut points was not affected by high loadings. Many of the PM speciation samplers in U.S. have been equipped with VSCC. Since 1997, there are nine integrated sampling systems (including inlet, surface material, gaseous denuder, filter medium, filter holder, and pump) become commercially available. Several of them have been applied in U.S. urban Chemical Speciation Network (CSN) (Flanagan et al., 2006; Frank, 2006) and non-urban Interagency Monitoring of PROtected Visual Environments (IMPROVE) network (Eldred et al., 1998). Several instrument such as R&P Partisol 2025 Dichotomous Sequential Air Sampler have been applied by HKEPD at the three compliance sites (i.e., TW, YL, and TC sites). To have a better understanding of PM chemical composition as a function of the particle sizes for visibility assessment, the following instrument are needed: 1) Micro-Orifice 4-1
Uniform Deposit impactor (MOUDI, 0.056-18 µm in 10 stages as MOUDI, 0.010-0.056 µm in 3 stages as nano-MOUDI); 2) Low pressure impactor (0.03-10 µm in 13 stages); 3) Electronic Low Pressure Impactor (0.007-10 µm in 12 stages); and 4) Davis Rotating-Drum Uniform Size Cut Monitor (0.1-2.5 µm in 3 stages and 0.09->5 µm in 8 stages). Chow and Watson (2007) summarizes the ~20 studies reporting particle size distribution of PM chemical compositions throughout the world. In Hong Kong, several studies address size distribution for PM ionic and elemental species (Cheng et al., 2008; Zhuang et al., 1999a; Zhuang et al., 1999b), dicarboxylic acids (Yao et al., 2002), water-soluble organic carbon (Yu et al., 2004b), and on vehicle exhaust plumes (Yao et al., 2007a). In addition to the commonly applied chemical speciation methods for elements, anions (e.g., Cl-, NO3-, SO4=) and cations (Na+, Mg++, K+, Ca++, NH+), and organic and elemental carbon (OC, EC), efforts have been made in carbon speciation. Within HK and PRD area, several different carbon analysis protocols (e.g., HK Governmental Laboratory, HKUST-3, NOISH-like Thermal/Optical Transmission [TOT], IMPROVE and IMPROVE_A Thermal/Optical Reflectance [TOR] methods) have been applied. Table 4-1 showed that these protocosl varied by temperature and analysis time per temperature protocol. As of April, 2006, U.S. EPA has formally switched the Speciation Trend Network (STN)-TOT carbon protocol to IMPROVE_A protocol (U.S.EPA, 2006a). To maintain consistency, the HK Supersites Program will only apply carbon analysis following the IMPROVE_A protocol (Chow et al., 1993; Chow et al., 2001; Chow et al., 2004b; Chow et al., 2005b; Chow et al., 2007a). Table 4-1. Summary of the thermal/optical analysis protocols for IMPROVE_A, IMPROVE, Hong Kong Governmental Laboratory (HKGL) and Hong Kong University of Science and Technology (HKUST-3). Carrier
Carbon
IMP_A_TOR/TOTa
IMP_TOR/TOTa
HKGL_TOTa
HKUST-3_TOTa
Gas
Fraction
Temperature, Time
Temperature,
Temperature,
Temperature,
Time
Time
Time
30 °C, 90 s
30 °C, 90 s
30 °C, 90 s
He-1
OC1
140 °C, 150-580 sc
120 °C, 150-580 sc
350 °C, 70 s
250 °C, 150 s
He-2
OC2
280 °C, 150-580 s
250 °C, 150-580 s
550 °C, 70 s
550 °C, 150 s
He-3
OC3
480 °C, 150-580 s
450 °C, 150-580 s
850 °C, 110 s
650 °C, 150 s
He-4
OC4
580 °C, 150-580 s
550 °C, 150-580 s
Cool Ovens
850 °C, 110 s
-
-
O2/He-1b
EC1
580 °C, 150-580 s
550 °C, 150-580 s
550 °C, 10 s
650 °C, 150 s
O2/He-2
EC2
740 °C, 150-580 s
700 °C, 150-580 s
600 °C, 50 s
750 °C, 150 s
O2/He-3
EC3
800 °C, 150-580 s
He-Purge
He-5
30 °C, 90s
Cool Oven
800 °C, 150-580 s
700 °C, 40 s
850 °C, 150 s
O2/He-4
-
750 °C, 30 s
890 °C, 150 s
O2/He-5
-
800 °C, 30 s
-
850 °C, 70 s
4-2
a
IMP_TOR
(IMPROVE_TOR; Not listed on Table, same temperature and times as IMP_TOR/TOT) Thermal/optical reflectance analysis following the IMPROVE (Interagency Monitoring of Protected Visual Environments) protocol using DRI/OGC analyzers (Desert Research Institute, Reno, NV). IMPROVE_TOR does not advance from one temperature to the next until a well-defined carbon peak has evolved (Chow et al., 1993, 2001, 2004b). Filter reflectance is monitored throughout the analysis; pyrolyzed OC (OP) is defined as the carbon evolving between the introduction of oxygen (O2) in the helium (He) atmosphere and the return of reflectance to its initial value (the OC/EC split). OP is reported as a positive value if the OC/EC split occurs after the introduction of O2, and as a negative value if the OC/EC split occurs before O2 is introduced. In either case, OC equals OC1+OC2+OC3+OC4+OP and EC equals EC1+EC2+EC3-OP. Eight well-defined fractions of carbon, including four OC fractions (OC1, OC2, OC3, and OC4), three EC fractions (EC1, EC2, and EC3), and OP are reported as part of the IMP_TOR protocol.
IMP_TOR/TOT.................Same as the IMP_TOR protocol but using a DRI Model 2001 thermal/optical carbon analyzer (Atmoslytic, Calabasas, CA). The DRI Model 2001 performs charring correction through both reflectance and transmittance and reports as OPR and OPT, respectively. Subsequently, OC and EC calculated from OPR (OPT) are referred to as OCR and ECR (OCT and ECT), respectively. IMP_A_TOR/TOT
Same as the IMP_TOR/TOT protocol using a DRI Model 2001 thermal/optical carbon analyzer.
Temperatures are revised to more accurately reflect actual sample
temperatures (Chow et al., 2005b, 2007a). HKGL_TOT
Thermal/optical transmission analysis following the HKGL (Hong Kong Government Laboratory) protocol with a Sunset Aerosol Analysis Lab Instrument (Sunset Laboratory Inc., Tigard, OR). The HKGL transmittance protocol (HKGL_TOT) is similar to STN_TOT or NIOSH 5040 (NIOSH, 1999) but with different thermal protocols and combustion atmospheres (Chow et al., 2005; Watson et al., 2005).
HKUST-3_TOT ................Thermal/optical transmission analysis following the HKUST-3 (Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong, China) protocol using a Sunset Laboratory Carbon Aerosol Analysis Lab Instrument. The HKUST-3 transmittance protocol has a short and fixed residence time per temperature plateau and does not usually report distinguishable carbon fractions. b
2% O2 in He for IMPROVE and STN protocols, and 1% O2 in He for HKUST-3 protocol.
c
The residence time at each temperature in the IMPROVE protocol depends on when the flame ionization detector (FID) signal returns to the baseline to achieve well-defined carbon fractions.
Efforts have been made to explain the analytical methods for water-soluble OC (WSOC) and to establish comparability between traditional solvent extraction (SE) - and thermal desorption (TD) methods followed by Gas Chromatography/Mass Spectrometry (GC/MS) (Chow et al., 2007c; Hays et al., 2004). The TD-GC/MS provides organic speciation of ~130 non-polar compounds (e.g., n-alkanes, alkenes, hopanes, sterenes, and polycyclic aromatic hydrocarbons [PAHs]) using small portion (~ 0.5 cm2) of filter substrate from existing archived filter samples. The method was developed by Professor Yu of HKUST
4-3
(Ho and Yu, 2004) and has been applied in U.S. for epidemiological studies. Speciation of WSOC using ion-chromatography (IC)–based instrument can replace labour-intensive solvent extraction methods to obtain specific source markers. 4.1.2 Continuous Particle Analyzers Advances have been made on conventional Tapered-Element Oscillating Microbalance (TEOM), using diffusion Nafion dryer to remove particle-bound water (i.e., Filter Dynamic Measurement System [FDMS]). Table C-1 of Appendix C summarized the accuracy, precision, Minimum Detection Limit (MDL), interferences, comparability, and data completeness of the continuous PM2.5 mass and mass surrogate instruments. Reasonable good agreements on PM2.5 mass were achieved between FRM and FDMS-TEOM or Beta Attenuation Monitor (BAM) (Chow et al., 2008a). For ionic species, over 10 different single and multi-ions continuous analyzers have been developed. Table C-2 of Appendix C summarizes analytical specifications for sulfate and for multi-ions. These instruments suffer the same issue with respect to 15-30% underestimation at high concentration as compared to collocated filter sampling and well-established laboratory analysis. Not much development has been made for continuous OC and EC analysers. Table C-3 of Appendix C shows that R&P 5400 has been discontinued from the market. Whereas semi-continuous Sunset OC/EC is not compatible with those of filter-base IMPROVE_A protocols (Park et al., 2006a). For continuous light absorption (reported as black carbon [BC]), there are several instruments have been applied at various U.S. Supersites (e.g., Dual and Seven-wavelength Aethalometer, Particle-soot Absorption Photometer [PSAP], Multi-angle Absorption Photometer [MAAP], and Photoacoustic analyzer). Park et al. (2006a) showed that the manufacture supplied absorption coefficient (σabs) varied by different environment under different meteorological conditions. Site-specific σabs need to be developed in HK to convert the particle light absorption (M m-1) to BC concentrations (µg/m3). 4.2
Continuous Gaseous Analyzers
4.2.1 Continuous Nitrogen Species, SO2, CO, and CO2 Little progress has been made for the criteria pollutant gaseous monitors. Tables C-4 of Appendix C summarizes measurements and analytical specifications for gaseous NO/NOx, NOy, HNO3, PAN/NO2 and NH3. For NOy analyzer, the manganese dioxide converter is placed at the sampling inlet of the conventional NO/NOx analyzer to minimize loss in the sampling system. The converter converts reactive oxidant of nitrogen and other organic nitrogen compounds into NO before it can be chemically react with the sampling system. It alternates between the two channels: 1) total reactive nitrogen species or NOy (i.e., NOy = PAN+HNO3+NO+NO2), and 2) HNO3 denuded channel to acquire NOy minus HNO3. The difference between the two channels is HNO3. The NH3 analyzer is based on the combination of a high performance NH3 converter and proven chemiluminescence detection to measure NH3, oxides of nitrogen (NOx) and total nitrogen compounds (Nx). Nx is the sum of NO+NO2+ NH3 i.e. total oxides of nitrogen 4-4
including NH3. To measure Nx concentration, NO2 and NH3 are converted to NO in a quartz converter heated to 750°C. In a separate reaction, NOx (NO+NO2) is passed through a molybdenum converter heated to approximately 325°C. The resulting NH3 concentration is determined by subtracting the Nx result from the NOx. Table C-4 of Appendix C compares different SO2 measurements. The MDLs of UV fluorescence SO2 analyzer can be reduced from conventional analyzers at 1 ppb to high-sensitivity SO2 at 0.05 ppb level. The CO and CO2 analyzer measures low ranges of pollutants by comparing infrared energy absorbed by a sample to that absorbed by a reference gas according to the Beer-Lambert law. This is accomplished with a Gas Filter wheel which alternately allows a high energy light source to pass through a CO/CO2 filled chamber and a chamber with no CO/CO2 present. The analyzers typically using a microprocessor control with nondispersive infrared (NDIR) detector at a MDL of 20 to 100 ppb and 0.002 to 200 ppb for CO and CO2, respevtively (Table C-5). 4.2.2 Non-methane Hydrocarbon and Volatile Organic Compounds (VOCs) Speciation The analytical specifications for the real-time non-methane hydrocarbon (NMHC) and speciated volatile organic compounds (VOCs) are given in Table C-6 of Appendix C. The CH4/NMHC analyzer is used to provide a direct measurement of both methane and non-methane hydrocarbons without the use of scrubbers or catalysts from 0 to 2,000 ppm. The proprietary column design is unaffected by oxygen and provides complete recovery of low volatility compounds. Continuous VOCs analyzer is capable of measuring all target VOC compounds of the requirement of PAMS in one hour averaging periods and without any need for costly laboratory analysis. The system employs two gas chromatographs (GC) which allow analysis of light and heavy VOC compounds with excellent quality (precision ±3%; accuracy ±15%) for separation and high sensitivity via flame ionization detector (FID). Three approaches to VOC monitoring have emerged in Hong Kong during the last decade, but with different species being analyzed for each study as shown in Table C-7 of Appendix C: TAPs project (since 1997): collect 24-hr integrated samples into canisters by active VOC sampler (Xontech 910A/RM 910A), with subsequent analysis in Hong Kong Government Laboratory by Gas Chromatography/Mass Spectrometry (GC/MS). Total of 143 C3–C12 species (including alkanes, alkenes, aromatics and halocarbons) are determined. The limit of detection (LOD) and practical quantification limits were 0.02 and 0.10 ppbv, respectively. Spatial Studies: acquired air samples of 3 hrs or 24 hrs in duration with canisters, followed by subsequent analysis at laboratory by Prof. Donald Blake, University of California, Irvin, USA. These GCs equipped with five different column/detector combinations. Total of 50 C1-C10 species are determined (including CO, CO2, CH4, alkanes, alkenes, aromatics, halocarbons and alkyl nitrates). The LOD is 5 ppbv for CO, 0.003 ppbv for NMHCs, 0.00002 ppbv for alkyl nitrates, and 0.00001-0.01 ppbv for halogenated species.
4-5
Continuous VOC speciation at Tung Chung (TC) site: real-time sampling and on-line analysis of ambient air, using GC (Synspec Model GC955-600-800). Total of 30 C2–C10 VOCs including alkanes, alkenes, aromatics are determined by sample systems and two separate column systems with the LODs 0.15 ppbv for benzene. These sampling approaches are valuable and have their specific advantages and disadvantages. Integrated and grab sampling has low time resolution but lower detection limits. This approach is thus well suited for extensive monitoring networks covering large areas. Canister sampling suffers the losses of isoprene if sampled air is not analyzed immediately. Real-time monitoring provides a good time resolution but is costly, since GC system and a skilled operator are needed (WMO, 1995). However, sampling protocols and the measured species should be harmonized to allow for best comparability of data. In Table C-7 of Appendix C, the VOC species requested by U.S. EPA for Method TO-14A and PAMS (Photochemical Assessment Monitoring Stations) are shown. To maintain consistency across the network and target on VOC species related to photochemical reactions and air toxics, the proposed 87 VOC species for the Hong Kong Supersites Program, which includes both U.S.EPA list of 61 VOC species (PAMS) and 34 VOC species for TO-14 methods are selected (Table C-7). 4.3
Proposed Advance Measurements for Hong Kong Supersites Program
Table 4-1 summarizes the proposed Supersites measurements for the three anchor sites (i.e., Tsuen Wan (TW), Yuen Long (YL), Tung Chung (TC)) and four satellite sites (i.e., Mong Kok (MK), Hok Tsui (HT), The Hong Kong University of Science and Technology (UST), and Tai Mo Shan (TMS)). The existing HKEPD measurements are also included in Tables 4-2, along with the averaging time and sampling frequency. The measurement and analytical specifications (e.g., accuracy, precision, minimum detectable limits [MDLs], interference) for the proposed Hong Kong Supersites Program are documented in Table D-1 of Appendix D. There are 21 types of additional instrument are proposed, as shown in Table 4-1. The instruments were selected based on the evaluation summarized in Tables C-1 through C-6 of Appendices C.
4-6
Table 4-2. Summary of Proposed Instrument for Hong Kong Supersites Program Sampling Sites Averaging
TW1
YL2
TC3
MK4
Instrument
Time
Frequency
TSP
Hi-vol sampler (General Metals 2310)
24hrs
1/6days
+
+
+
+
RSP
Hi-vol sampler (General Metals 2310)
24hrs
1/6days
+
+
+
+
PM10
Hi-vol sampler
24hrs
1/6days
+
+
+
PM2.5
Hi-vol sampler
24hrs
1/6days
+*
+*
+*
PM2.5
Low-vol R&P2025
24hrs
1/6days
+
+
+
PM2.5
Mini-Volume sampler (AirMetrics or BGI Mini-Vol Air Sampler)
24hrs
1/6days
xx
PM2.5
Reference Ambient Air Samplers 400 (Special Study)
24hrs
1/6days
+
Continuous ambient air monitor with two FDMS (Thermo Scientific Model
Hourly
Daily
x
HT5
UST6
xx
xx
TMS7
Mass Concentration Integrated PM Mass
xx
Continuous PM Mass PM2.5 and PM10 mass
1405-DF) PM2.5 mass
Continuous Ambient Particulate Monitor (Thermo Scientific Model: 1400)
Hourly
Daily
+
+
+
+
+
PM10 mass
Continuous Ambient Particulate Monitor (Thermo Scientific Model: 1400)
Hourly
Daily
+
+
+
+
+
PM2.5 Surrogate
Continuous Dust Track TSI
Hourly
Daily
x
Sulfate
R&P 8400S, Flash volatilization of sulfate
30 mins
Daily
+
Sulfate
Continuous Sulfate Particulate Monitor with PM2.5 Inlet (Thermo Scientific Model:
90sec
Daily
Nitrate
R&P 8400N Flash volatilization of nitrate
30 mins
Daily
x
x
x
Ions +
+ x
5020 SPA)
4-7
+
+
+
x
Table 4-2.
(Cont’d) Sampling Sites TW1
Averaging Instrument
Time
Frequency
PILS-IC (Applikon Model: ADI 2081)
Hourly
Daily
x
OC/EC
R&P 5400 ambient carbon particulate monitor
2 hrs
Daily
+
OC/EC
Sunset field analyzer
hourly
Daily
Grimm Technologies Wide Range Aerosol Spectrometer
5mins
Daily
Grimm Technologies CPC 5.403
5mins
Daily
Light absorption
Magee Scientific Aethalometer (7-Wavelength)
5mins
Daily
Light absorption
Magee Scientific Portable Aethalometer (Dual-Wavelength)
5mins
Daily
Water soluble ions
YL2
TC3
MK4
HT5
x
x
UST6
TMS7
OCEC + +
Aerosol Size / Particle number Size distribution (0.0045
x
to 20 μm ) Particle Count (0.0045 – 3μm) Light absorption x
x x
x
Light scattering Light Scattering
Ecotech Aurora-3000 3-Wavelength Nephelometer
Daily
x
Ozone Ozone
2B Technologies Model 205 Dual Beam Ozone Monitor
1min
Daily
Ozone
2B Technologies Model 306 Ozone Calibration Source
1min
Daily
Ozone
API 400 UV Absorption
1min
Daily
+
+
+
NO/NOx
Teledyne API Technologies API 200A
1min
Daily
+
+
+
NO2
2B Technologies Model 401 NO2 Converter
1min
Daily
x
x
x
NO
2B Technologies Model 400 NO Monitor
1min
Daily
x
x
x
x
x
x
x
NO/NOx/NO2/PAN/NOy
4-8
+
Table 4-2.
(Cont’d) Sampling Sites TW1
Averaging
YL2
TC3
Instrument
Time
Frequency
NO
2B Technologies Model 408 NO Calibration Source
1min
Daily
PAN/NO2
PAN Detector Drummond Technology LPA-4D
5mins
Daily
x
NOy
Thermo Scientific NOy Analyzer Model: 42i-Y
10sec
Daily
x
+
Ammonia EcoTech Model: EC9842r
1min
Daily
x
CO
Thermo 48C
1min
Daily
+
+
+
CO2
Thermo 410i-D
30sec
Daily
x
SO2
Thermo Scientific Model: 43i-TLE (High Sensitivity)
10sec
Daily
SO2
TECO 43A/ API 100E
1min
Daily
NMHC (CH4, Total HC)
Thermo Scientific 55C
1min
VOCs (C2-C12)
American Ecotech VOC AirmoZone (PAMS)
VOCs (C2-C10) VOCs (C1-C6)
MK4
HT5
UST6
TMS7
x
Ammonia NH3 CO/CO2 + x
x
SO2 x +
+
+
Daily
x
+
30mins
Daily
x
The Syntech Spectras Model: GC955 Series 600/800 POCP
Hourly
Daily
SRI Environmental and BTEX GC Systems (PID and FID/DELCD) TO-14
Hourly
Daily
Cimel CE-318 automatic sunphotometer AOD
Hourly
6hours
+
Gaseous Hydrocarbon x
+ x
Satellite Aerosol Optical Depth Light
Detection
And
MPL Micropulse Lidar
30sec
Ranging Meteorology
4-9
Daily
# Φ
x
Table 4-2.
(Cont’d) Sampling Sites Averaging
TW1
YL2
TC3
MK4
+
+
+
+
Instrument
Time
Frequency
MetOne Instrument
3sec
Daily
Meteorology (WS, WD,
Davis Instrument Wireless Vantage Pro2 Plus with 24-Hr Fan Aspirated Radiation
10sec
Daily
Temp, RH, Solar)
Shield Model: 6163
Visibility
Point Transmisometer (ASOS)
Hourly
Daily
-
Remtech Doppler Sodar Model: PA1 Antenna
10mins
Daily
x
Meteorology (WS, WD,
HT5
UST6
+
Temp, RH, Solar)
Sonic
Detection
and
Ranging Note: * denotes PM2.5 Hi-vol sampler to be provided by HKEPD in total of 3 units + denotes the instrument at the site location is currently running by HKEPD x denotes the instrument is proposed to be installed at the monitoring site # denotes NASA sunphotometer measurement Φdenotes HKUST Aerosol extinction to backscatter measurement, the second unit is located in Sha Tau Kok8 (STK) Site -denotes visibility measurement is operated by the Hong Kong University of Science and Technology
1
2
3
4
5
TW - Tsuen Wan (PM2.5) YL - Yuen Long (O3) TC - Tung Chung (O3) MK - Mong Kok (Hotspot) HT - Hok Tsui 6
7
8
(Transport) UST - Hong Kong University of Science and Technology (Transport) TMS - Tai Mo Shan (Elevated) STK - Sha Tau Kok (rural)
4-10
TMS7
x
4.4
Proposed Additional Special Studies
Measurements listed in Table 4-1 do not include enhanced measurements that could be added to the long-term monitoring as special studies during a short sampling periods (e.g., two weeks intensive study with short [2-4 hr] sampling durations for particle size distribution of mass and chemical composition (e.g., using MOUDI and/or Nano MOUDI) and for visibility assessment. In addition to measurement acquired at the TW site, particle number/size distribution (e.g., using Nano-Scanning Mobility Particle Sizer [SMPS], Standard SMPS, Aerodynamic Particle Sizer, and Wide Range Spectrometer) at the Yuen Long (YL) and HT or UST transport sites can be used to better understand the particle nucleation and formation in a large spatial scale. The expansion of the muti-ions analyzer from Tsuen Wan (TW) site to Yuen Long (YL) and Tung Chung (TC) allow a better understanding of the diurnal variation of inorganic secondary ions. Spatial variations of PM2.5 and chemical species (including water soluble organic compounds [WSOC] and organic speciations) can be obtained by densely located battery-powered mini-vol samplers (e.g., 20 sites) in a neighborhood for daily sampling of two weeks duration. Continuous NH3 and HNO3 measurement at the TW site can be reinforced with denuder difference sampling using integrated speciation samplers. To better understand cloud formation and its chemical composition, a CalTech Cloud Collector can be used to sample cloud water at the TMS site. Table 4-2 lists several of the advanced instruments that may be collaborated with professors from HKPU and HKUST for future special studies. Table 4-3. Proposed Special Study Instrument Observables
Instrument
Operated (Owner)
Partical Size/Number
TSI Nano SMPS, SMPS, HKPU (Lee) Wide Range Aerosol Spectrometer Mass and Chemical Mono-MOUDI,MOUDI HKPU (Lee) and HKUST Particle Size Distribution (Chan) Diurnal Variation of URG AIM and Metorhm HKPU (Wang) and Muti-Ions Ion Analyzer HKUST (Yu) Cloud Water CalTech Collector HKPU (Wang) NH3 an HNO3 Thermo Scientific HKUST (Yu) Reference Ambient Air Sampler (RAAS) or MetOne (SASS) Speciation Monitor * NOT included in the proposed budget
4-11
5.
SUPERSITES CONFIGURATION AND OPERATION
5.1
Space and Power Requirements
Tables 5-1 and 5-2 summarize the additional space and power requirements for each site, respectively, based on the enhanced measurements specified in Table 4-2. All of the proposed sites have sufficient site security and accessibility. Table 5-1. Space requirements for additional Supersite instruments at existing monitoring sites. Site
Outdoor Space Requirements
Indoor Space Requirements
200cm x 200cm (Shelter) +
TW
100cm x 100cm (FDMS-TEMO)
YL TC MK HT UT
65cm x 65cm (Sodar)
39U (1940mm) + 27U (1407mm)
-
-
-
19" rack (600mmx600mm) 27U (1407mm)
-
19" rack (600mmx600mm) 45U (2207mm)
19" rack (600mmx600mm) 39U (1940mm) 19" rack (600mmx600mm) x 2
TMS
30cm x 30cm (Integrated Meteorology Instrument)
19" rack (600mmx600mm) 27U (1407mm)
-
19" rack (600mmx600mm) 39U (1940mm)
See Table D-2 of Appendix D for detailed space requirements U is a unit of rack (1.75")
Table 5-2. Site TW YL TC MK HT UST TMS Total
Power requirements for supersite instruments at each site. Outdoor Power Requirements
Indoor Power Requirements
5kW
12kW
3kW
10kW
-
-
3kW
6kW
-
12kW
3kW
6kW
-
8kW
15kW
68kW
See Table D-2 of Appendix D for detailed power requirement. Includes indoor power for two air conditioning systems at each site.
For TW, the particle sizing instrument and multi-ions analyzer will be placed on two working tables (e.g., 0.5 m (D) × 1 m (L)) in the room which are currently located but not occupied at the present TW site (Figure 5-1). In addition, one standard instrument rack needs to be placed within the shelter to accommodate the aethalometer, nephelometer, CO2 and
5-1
NH3 analyzer (Figure 5-2). In addition, two AirMetic PM2.5 mini-vols will be installed on the roof top of the shelter.
PILS‐IC
Wide Range Aerosol Spectrometer
Figure 5-1. Additional supersite instruments to be located at the TW site (previous AQO room).
Figure 5-2. shelter)
Additional supersite instruments to be located at the TW site (inside
5-2
For YL, one standard instrument rack needs to be placed within the shelter to accommodate the speciated VOCs instruments (Figure 5-3). The instrument rack also needs to host calibration and data acquisition system. The space requirement also includes the gas cylinders to operate and calibrate the continuous analyzers.
Figure 5-3. shelter).
Additional supersite instruments to be installed at the YL site (inside
In addition, an environmental shelter is needed for the YL site. The configuration of instrument within the temperature-controlled environmental shelter is shown in Figure 5-4. Two Airmetrics mini-vols will be placed on the roof top of the shelter. Since the TC site is well equipped with PM and photochemical related gaseous measurements, data from the TC site will be used to compare and contrast measurements from the YL sites. Therefore no additional space or power is required.
Figure 5-4. Proposed environmental shelter (1.2m (L) × 1.2m (W) × 1.5m (H)) to be placed at the YL site (outside AQO station).
5-3
An instrument rack is required at the MK site to accommodate the DustTrak, Condensation Particle Counter (CPC), and Aethalometer (Figure 5-5) and an environmental shelter (1.2m (L) × 1.2m (W) × 1.5m (H)) is needed for NMHC, and VOCs (portable GC) which is placed outside the AQO station (Figure 5-6). Two Airmetrics mini-volumes will be placed on a tripod, next to the existing integrated samplers.
Figure 5-5. Additional supersite instruments to be located at the inside of the MK site (inside AQO station).
5-4
Figure 5-6. Additional supersite instruments to be located: a) on the rooftop and b) in the outside environmental shelter at the MK site (outside AQO station). The old instrument shelter (2×3 m) from the The Hong Kong Polytechnic University is planned to be upgraded and used at the HT site to accommodate two instrument racks to house DustTrak, Aethalometer, Thermo Scientific Model 5020 SO4=, CPC, high-sensitivity SO2, NO/NOx, and O3 analyzers. Since a HKO surface meteorological station is nearby, so
5-5
there is no plan to add any meteorological measurements at this site. Two AirMetic Mini-vols will be placed on the roof top of the shelter. A new instrument shelter is needed at the UST site. One instrument rack is needed to accommodate DustTrak, O3, and NO/NO2 measurements. One meteorological monitoring station (including WS, WD, temperature, and relative humidity) from Davis Instrument, along with two AirMetic Mini-vols are planned to install on the roof top of the shelter. Limited space is available for the TMS site. A compact, environmental-controlled instrument enclosure, similar to the ones constructed by Airpointer (Air Monitors Ltd., UK), is proposed. Figure 5-7 show that the enclosure could accommodate the Thermo Scientific Model 5020 sulfate, portable Aethalometer, O3, SO2, and CO2 monitors and data acquisition system. Since the site are frequently situated in a mist of fogs and clouds, the PM2.5 inlets may need to be equipped with smart heaters that can initiate the heating element as relative humidity exceed pre-set level (e.g. 65%). Dimensions and power requirement for individual instrument are summarized in Table D-2 of Appendix D. Since most of the in-situ continuous instrument are operated with low flow rates, the required power are typically 95% data recovery with precisions 2 μm in diameter). Size distributions of elements were examined by Cheng et al. (2008). It is reported that Mg, Al, Si, Ca, Ti and Na, Cl are mainly present in coarse mode; K, V, and Ni distributed in both fine and coarse modes; S, Pb, As, Se, Zn, Cu, Cd, and Ba have a single peak at around 1.0 µm.
Figure A-3. Typical size distributions: (a) SO4=, (b) NH4+, and (c) NO3-. (Zhuang et al., 1999a). Water-soluble organic compounds (WSOC) Chemical characterization of the water-soluble organic compounds (WSOC) was also performed (Louie et al., 2002; Yu et al., 2004b; Ho et al., 2006c). Yang et al. (2003) compared two A-9
methods for WSOC analysis: one based on a total organic carbon analyzer (TOC) and the other based on an aerosol carbon analyzer (ACA). Measurements of standard compounds and aerosol samples show that both are suitable for the determination of WSOC in aerosol samples and acid treatment to remove carbonate carbon was unnecessary for the determination of WSOC in PM2.5 aerosols. The WSOC accounts for a significant portion of the total organic carbon mass, ranging from 14% to 64% (Louie et al., 2002). The WSOC concentrations are comparable in their concentration levels at the roadside, urban, and rural sites. This is similar to the spatial distribution of secondary inorganic aerosols in Hong Kong (i.e., SO4=, NH4+ and NO3-). These observations seem to suggest that the WSOC fraction is mainly of secondary origin. The WSOC/OC ratio shows a clear spatial pattern, highest at rural sites and lowest at roadside sites (Louie et al., 2002; Ho et al., 2006c). The OC/EC ratio also exhibits the same spatial pattern (Ho et al., 2002a). Both observations indicate that the contribution of secondary organic aerosol is more prominent at the rural site. The size distribution of WSOC was investigated by Yu et al. (2004b). They found that the WSOC exhibited bimodal size distributions, centred at 0.7±0.1 and 4.0±0.3 µm, respectively. The fine WSOC accounted for 2/3 to 4/5 of total WSOC, comprising of 21%, 39%, and 40% of low, medium, and high MW polar compounds and coarse WSOC (> 2.1 µm) was largely made of the low MW polar compound group. The low MW group is likely to associate with sea-salts, and high MW group is secondary in origin, formed during cloud processing. Solvent extractable organic compounds (SEOC) Several studies (Zheng and Fang, 2000; Yao et al., 2002; Sin et al., 2005; Ho et al., 2006b) have investigated the solvent extractable organic compounds (SEOC) and related analysis methods (Sin et al., 2002). The results of a 12-month study of more than 100 SEOC in PM2.5 collected at three air monitoring stations located at roadside, urban, and rural sites in Hong Kong are reported in Sin et al. (2005)’s study. The total SEOC accounted for ~8-18% of OC. The two largest components of SEOC were UCM and FAs, accounting for 32-71% and 23-56% of the total SEOC, respectively. Distinct seasonal variations (summer/winter differences) were observed with higher concentrations of the total and each class of SEOC in the winter and lower concentrations in the summer. Spatial variations are also obvious, with the roadside samples having the highest concentrations of SEOC and the rural samples having the lowest concentrations in all seasons. Characteristic ratios of petroleum hydrocarbons, such as carbon preference index, unresolved to resolved components, and carbon number with maximum concentration, suggest that PM2.5 carbon in Hong Kong originates from both biogenic and anthropogenic sources. Vehicular exhaust is the dominant urban emitter of carbon at MK roadside site and TW urban site, while greater biogenic contributions in summer and greater continental outflow in winter at HT rural site. The long-term trend of PM2.5 mass and chemical composition The long-term trend of PM2.5 mass and chemical composition was investigated recently (So et al., 2007). Comparing PM2.5 data in 2004/2005 with those collected in 2000/2001, a reduction in PM2.5 mass concentration at the roadside (8.7%) but an increase at the urban (15%) and rural (20%) sites were observed. The reduction of PM2.5 at the roadside was attributed to the decrease of carbonaceous aerosols (organic carbon and elemental carbon) (>30%, Table A-1), indicating the effective control of motor vehicle emissions over the period. On the other hand, the apparent increase over the territory was found for NO3-, SO4=, and NH4+, indicating the deterioration of
A-10
regional air quality. Over 36% growth in sulfate concentration was found from 2000/2001 to 2004/2005. Table A-1 Twelve-month average concentrations (µg/m3) of OC, EC, TC, OC/EC ratios, SO4=, NO3-, and NH4+ for PM2.5 samples collected at the MK roadside site, TW urban site, and HT rural site during the two study periods (2000/2001 and 2004/2005) (So et al., 2007).
Wind effects and long-range transport effects on PM levels It is found that Hong Kong PM and some chemical component levels were influenced by wind effects: the high loadings are apparent in the days with the northerly and westerly winds from the continent, while the days with the southerly winds possess lower loadings (Cheng and Lam, 1998; Lee and Hills, 2003; Fung et al., 2005; Wai et al., 2005; Louie et al., 2005b). Since Hong Kong’s climate is controlled by the Asian monsoon, the wind patterns are distinctive in four seasons, with prevailing southerly wind in summer and northerly and north-easterly winds in spring, summer, and winter. Therefore, the seasonal variations of PM and some chemical component levels (i.e., SO4=, NO3-, NH4+, and OM) normally show high loading in winter autumn, and spring, and low loading in summer. A box model study reported, on average, ~40% of SO4= and NH4+ in Hong Kong was from continental air masses and non-local sources (Pathak et al., 2003). Extremely events, like Asian dust storm, tropical cyclones, and pollution episodes, also impact the air quality, leading to day-to day variations of PM and chemical species. Two studies (Fang et al., 1999; Cao et al., 2003a) compared the levels of aerosols in Hong Kong between Asian dust storm and normal days. The normal day samples exhibited local characteristics while dust storm
A-11
day samples suggested outside sources. The concentrations of crustal elements and anthropogenic species were 2-14 higher in dust storm than non dust storm period (Cao et al., 2003a). Moreover, the eolian input includes high plant wax content; a low ratio of C18:1/C18:0 fatty acids in the dust samples indicated ageing of aerosols due to long-range transport (Fang et al., 1999). Investigating the cool season pollution episodes showed high concentrations of As, Mn, V, non-sea salt SO4= and NH4+_N, indicating the impact of fossil fuel burning, except for mineral dust (Lee and Hills, 2003; Fung et al., 2005). Particle Numbers Short-term measurements of particle numbers were conducted in urban, tunnels, and rural areas (Yao et al., 2005; Yao et al., 2006b; Tsang et al., 2008). According to a previous study (Yao et al. 2005) with an Engine Exhaust Particle Sizer (5.6 – 560 nm), the aerosol representative of rural area particles consisted of a single mode at 50 nm, which was formed by a combination of condensation and coagulation growth of nucleation mode particles. In high concentration gasoline vehicle plumes in urban streets, a bi-modal size distribution with a dominant mode at 10 ± 1 nm and a minor mode at 50 ± 5 nm was frequently measured, and in high concentration heavy-duty diesel vehicle plumes, a bi-modal size distribution with a dominant mode at 30 ± 4 nm and a minor mode at 10 ± 1 nm was observed. Inside the Tate Cairn’s Tunnel, 10 nm particles gradually decreased after an abrupt increase, while 50nmparticles increased from entrance to exit. The hetero-coagulation process between 10 nm particles and particles >30 nm is proposed to play a key role in the transformation of particle modes in the tunnel. Tsang et al. (Tsang et al., 2008) measured the number concentrations of particles in range of 5 nm-2 µm using a water-based condensation particle counter (WCPC) at three locations in MK district on 30 July 2005. The results were listed in the Table A-2. Atmospheric dilution, increased distance from exhaust pipes, and time contribute to rapidly decreasing particle counts in the ultrafine and nano-particle range after emission. High particle counts occur when vehicles accelerate, after stopping. Table A-2. Particle number concentrations (5 nm-2 µm) at the three locations Location and distance to road (m) A: HKEPD station, 1 B: Pioneer Center, 5 C: Primary School, 6
WCPC (particle/cm3) 89,705±75,546 55,453±29,060 45,088±7,358
Reviews on EPD AQO and AQMN Data for PM To protect Hong Kong's air quality, since 1989, EPD established sets of Air Quality Objectives (AQOs) for seven pollutants. The Total Suspended Particulates (TSP) and Respirable Suspended Particulates (RSP/PM10, particles with aerodynamic diameter less than 10 μm) were included as two air standards. The annual AQOs are 80 μg m-³ for TSP and for 55 μg m-³ RSP; the 24-hr AQOs are 260 μg m-³ for TSP and 180 μg m-³ for RSP (Table A-3). The air quality standards have not since been revised. Recently, WHO has found that PM2.5 is more hazardous than larger ones in terms of mortality and cardiovascular and respiratory illness. The United States, the European
A-12
Union and Australia have all recently reviewed, or are in the process of reviewing, their PM standards (see summary in Table A-3). They have all chosen to develop PM2.5 standards because it was felt that existing PM10 standards did not adequately protect public health. Hong Kong currently has PM10 standards but does not have any PM2.5 guidelines or standards. Table A-3. Summary comparison of PM standards in different countries (all figures in µg/m3)
Since 1984, the Hong Kong Environmental Protection Department (HKEPD) established a Hong Kong Air Quality Monitoring Network (AQMN). Up to now, the AQMN comprises fourteen fixed monitoring stations, including eleven general stations and three roadside stations. The 24-hr average TSP and PM10 data was measured by High-vol samplers, once every sixth day. A large amount of data of chemical compositions of aerosols, mainly focus on TSP and PM10, has been obtained. Levels of PM10 at roadside monitoring stations in Hong Kong are still a major concern, which is mainly attributed to the high concentration of diesel vehicles in urban areas. However, levels are on a downward trend, with a 13% drop in average concentrations over the past five years (see Figure A-4). This is largely due to the various vehicle emission controls implemented in Hong Kong since 1999. In Figure A-4, the RSP concentrations recorded in the territory showed a primarily downward trend between 1995 and 2002, rebounded afterwards to a higher level in 2004, then dropped again afterwards. In Hong Kong, high level of RSP at roadside is a major air pollution concern, which is mainly attributed to the high concentration of vehicles especially diesel vehicles in urban areas. As a result of the implementation of various vehicle emission control measures in recent years, the annual average of RSP at roadside in 2006 reduced by 13% compared with 1999. A-13
Figure A-4. Long-term trend of annual average PM10 (HKEPD 2006) Qin et al. (1997) analyzed the chemical compositional data of PM10 measured regularly at 11 AQMN stations from 1990 to 1994 and the chemical compositions, seasonality, and spatial variations of PM10 were studied. He found that most of the chemical species show seasonal variations which reflect the weather conditions. Carbonaceous aerosols accounts for approximately half of the mass. SO4=, NH4+ and NO3- account for a quarter of PM10, mainly from long distance transport controlled by the East Asian monsoons. Some species, like Cu and vehicle-related species (C, Zn, Pb, Br- and Ni) had large spatial variations in Hong Kong territory. VOLATILE ORGANIC COMPOUNDS (VOCs) Review on VOC in Hong Kong Due to high density of population and rapid economic growth since 1980s, Hong Kong encounters urban air pollution problems like most mega-cities. Large amounts of VOCs are released from both local vehicles and regional industrial activities everyday. In 2002, there were over 524,000 vehicles registered in Hong Kong, from which 8360 tonnes of non-methane VOCs (NMVOCs) were emitted. In addition, activities from power generation, marine vessels and aircraft generated 1229 tonnes of NMVOCs. Although the Government has been working for years to control emissions from motor vehicles, improvements from those measures have tended to be offset by the increasing emission levels from the continuously growing vehicle numbers and kilometres driven.
A-14
NMVOCs play an important role in physic-chemical processes of the troposphere as they largely contribute to the formation of ozone and other photochemical oxidants. Several other effects of VOCs are recognized such as their contribution to stratospheric ozone depletion, tropospheric photochemical ozone formation, toxic and carcinogenic human health effects and enhancement of the global greenhouse effect. Therefore, to impose efficient emission restrictions on the most dominant local and regional NMVOC sources and to establish the relative influence of regional versus local sources on the ambient air quality, it is important to understand the chemical composition of NMVOCs in the PRD region, identify major source regions of air pollution, and quantify the relative contribution of each source sector to ambient NMVOC levels. Ambient concentration levels in the PRD region A comparison of hydrocarbon mixing ratios from various sampling sites in the PRD region is presented in Table A-4. In addition to the concentration levels, other sampling details (sampling technique, sampling period etc) are also compiled. It is clear that canister sampling is the main method for VOC collection, due to its applicability as the whole air sampling and trace level analysis, and duplicated analysis from canister if necessary. A number of studies have been conducted to understand the characteristics of VOCs in Hong Kong, one of the largest metropolitans in the PRD region. Sin et al. (2000) reported that, at the Tsuen Wan and Central/Western stations, the annual average concentration of the measurable VOCs was well within the range of 0.20–5.0 ppbv. However, individual measurements of toluene concentration occasionally exceeded 20 ppbv. Guo et al. (2004a) further reported that the benzene, toluene, ethylbenzene, and xylenes (BTEX) concentrations at the Tsuen Wan and Central/Western stations were similar. The annual mean concentration of BTEX was found to fall in the range of 0.57–13.27μg /m3. Toluene was the most abundant VOCs in all the samples, and the maximum daily value was up to 53μg /m3. At a remote coastal site (Hok Tsui), Wang et al (2003a) reported that the two dominant NMHCs (C2-C8) were ethane (mean: 2368 pptv) and ethyne (mean: 1402 pptv), followed by propane (814 pptv), toluene (540 pptv), benzene (492 pptv), ethene (498 pptv), and n-butane (326 pptv). The most abundant halocarbon was CH3Cl (mean: 821 pptv), while 2-BuONO2 and i-PrONO2 were the two dominant alkyl nitrates species with a mean mixing ratio of 20 pptv and 19 pptv, respectively. Guo et al. (2006) demonstrated that large variations in the measured NMVOCs were observed at Tai O. The total average NMVOC concentration was 25.5 ppbv, in which alkanes accounted for 40%, alkenes 10%, alkynes 11%, aromatics 35% and measured halocarbons (C2Cl4 and CH3Cl) 4%. The most abundantly measured 10 compounds were toluene, ethyne, ethane, propane, ethene, n-butane, CH3Cl, ethylbenzene, benzene and i-pentane. These 10 NMVOCs, accounted for 76% of the total NMHCs. In particular, toluene alone accounted for 22% of the average. Ho et al. (2004) reported the VOCs levels at PolyU campus (PU), Kwun Tong (KT), and Hok Tsui (HT). Toluene was the most abundant atmospheric aromatic hydrocarbon, followed by benzene in both PU and KT stations. The concentrations of toluene at PU ranged from 14.4 to 54.3μg /m3 in winter and 11.6 to 39.2μg/m3 in summer, respectively. KT has the highest toluene concentrations (means 64.3μg/m3), especially in summer. The KT site has high traffic volume and its BTEX concentrations are only slightly lower than those in PU. At HT, which is classified as background, toluene, benzene, methylchloride and ethylene chloride were found to be the top four species in winter and their concentrations were 3.2μg/m3, 2.3μg/m3, 2.1μg/m3and 1.6μg/m3, respectively. Other VOC concentrations were less than 1.0μg /m3 since there were no significant anthropogenic sources. A-15
Table A-4. Mixing ratios of selected hydrocarbons in the PRD region Location
Hok Tsui, HKa
Tai O, HKb
Sampling periods
3 Mar to 26 Apr , 2001
Aug 2001 to Dec 2002
Sampling methods
Canister
Canister
mean±SD,ppbv (range)
mean±SD,ppbv (range)
Central / Western, HKc
Tsuen Wan, HKc
Road sites, HKd
43 Chinese citiese
Jul 1997 to Jun 1998
Jul 1997 to Jun 1998
Jan and Feb, 1998
Jan / Feb, 2001
Canister
Canister
Multi-sorbent tube
Canister
mean±SD,
mean±SD,
ppbv(Cmax) Ethane
2.37±0.51 (1.2-3.39)
Propane
0.81±0.31 (0.16±1.43) 0.22±0.11 (0.39±0.49) 0.33±0.17 (0.05±0.75)
i-Butane n-Butane i-Pentane n-Pentane
0.08±0.05 (0.02±0.24)
n-Hexane n-Heptane n-Octane n-Nonane n-Decane Ethene
0.50±0.36
ppbv(Cmax)
2.12±0.99 (0.08-2.01) 2.05±2.16 (0.02-12.99) 0.80±0.93 (0.01-6.05) 1.64±2.13 (0.01-12.79) 0.80±1.44 (0.01-17.25) 0.45±0.65 (ND-5.59) 0.50±0.67 (ND-4.74) 0.33±0.50 (ND-4.36) 0.06±0.06 (ND-0.43) 0.07±0.08 (ND-0.59) 0.09±0.15 (ND-1.47) 1.67±1.68
Dongguanf
Guangzhoug
Sep, 2005
Sep, 2005
Apr, 2005
Canister
Canister
Canister
mean±SD,ppbv (range)
mean±SD,ppbv (range)
mean,ppbv
1.89±0.1 (0.89-3.47) 6.79±0.73 (1.3-19.8) 21.6±0.23 (0.47-6.14) 3.5±0.36 (0.76-9.77) 2.47±0.28 (0.54-8.57) 1.12±0.13 (0.22-3.83) 0.84±0.92 (0.14-2.25) 0.68±0.87 (0.1-2.14) 0.2±0.03 (0.03-0.71) 0.13±0.03 (0.02-1.07) 0.12-0.02 (0.02-0.92) 3.97±0.42
1.6±0.06 (0.64-2.32) 2.46±0.19 (0.48-8.13) 1.07±0.08 (0.19-3.36) 1.89±0.17 (0.37-6.5) 1.42±0.12 (0.32-4.66) 0.7±0.1 (0.12-4.95) 0.69±0.09 (0.07-2.98) 0.62±0.1 (0.03-3.23) 0.15±0.02 (0.02-0.95) 0.08±0.01 (0.02-0.28) 0.08±0.01 (0.02-0.26) 3.01±0.27
3.90±1.17
range, ppbv mean±SD,ug/ m3(Cmax) 3.7-17.0 1.5-20.8 0.4-4.6 0.6-14.5 0.3-18.8 0.2-7.7 15.6±15.4 (49.8) 3.2±5.1 (19.0)
0.1-3.2
2.5±5.1 (16.9)
0.04-1.3
2.4±5.8 (24.8)
0.04-0.7
1.5±5.5 (31.7)
0.03-0.4
0.06-3.4
2.1-34.8
A-16
Guangzhouf
11.29±5.69 4.48±2.31 6.31±3.21 3.81±1.93 1.76±0.87 1.24±0.59 1.12±0.65
8.6±4.28
Location
Hok Tsui, HKa
Tai O, HKb
Sampling periods
3 Mar to 26 Apr , 2001
Aug 2001 to Dec 2002
Sampling methods
Canister
Canister
mean±SD,ppbv (range)
mean±SD,ppbv (range)
Central / Western, HKc
Tsuen Wan, HKc
Road sites, HKd
43 Chinese citiese
Jul 1997 to Jun 1998
Jul 1997 to Jun 1998
Jan and Feb, 1998
Jan / Feb, 2001
Canister
Canister
Multi-sorbent tube
Canister
mean±SD,
mean±SD,
ppbv(Cmax) Propene Ethyne
(0.10±2.09) 0.06±0.05 (0.02±0.21) 1.40±0.57 (0.42±2.78)
Isoprene
ppbv(Cmax)
Benzene
0.87±0.92 (0.02-10.32) Toluene 5.67±7.13 (0.01-48.98) Ethylbenzene 0.87±1.23 (ND-8.12) m/p-Xylene 0.96±1.86 (ND-1.46) a (Wang et al., 2003a); b (Guo et al., 2006); e (Barletta et al., 2005); f (Barletta et al., 2008);
0.2-8.2 2.9-58.3 0.04-1.7
0.79±0.47 (2.3) 4.5±3.1 (20)
0.90±0.44 (2.2) 5.0±2.3 (16)
26.7±33.0 (128.6) 77.2±74.4 (320.0) 3.1±6.7 (36.5)
0.45±0.32 0.56±0.42 (0.93) (2.1) 0.91±0.67 0.97±0.562.3 12.1±19.4 (2.5) () (106.0) c d (Sin et al., 2000); (Chan et al., 2002b); g (Tang et al., 2007)
A-17
Guangzhoug
Sep, 2005
Sep, 2005
Apr, 2005
Canister
Canister
Canister
mean±SD,ppbv (range)
mean±SD,ppbv (range)
mean,ppbv
(0.78-11.1) 0.81±0.09 (0.12-2.21) 4.95±0.48 (1.71-13.4) 1.63±0.17 (0.18-4.46) 0.14±0.02 (0.01-0.62) 2.05±0.24 (0.65-6.8) 5.87±0.74 (0.71-19.6) 1.24±0.17 (0.14-4.6) 1.53±0.19 (0.25-4.9)
(0.65-10.7) 0.53±0.04 (0.11-1.31) 4.27±0.41 (1.02-16.7) 0.68±0.12 (0.11-4.74) 0.13±0.02 (0.02-0.52) 1.26±0.14 (0.27-6.45) 6.13±0.81 (0.53-25.3) 1.06±0.17 (0.06-7.45) 1.47±0.16 (0.11-6.95)
range, ppbv
0.6±2.2 (12.9) 0.49±0.24 (0.17-1.15) 0.54±0.48 (0.11-1.78) 0.06±0.05 (0.01-0.26)
Dongguanf
mean±SD,ug/ m3(Cmax)
(0.03-10.53) 0.22±0.29 (0.01-2.18) 2.77±1.99 (0.08-11.74) 0.43±0.73 (ND-5.35)
a-Pinene
Guangzhouf
0.7-10.4 0.4-11.2 0.1-2.7 0.4-15.3
2.36±1.34 9.8±3.95 0.27±0.14
2.75±1.19 10.02±4.69 1.91±1.04 3.03±1.69
In other areas of the PRD region, especially Guangzhou, many studies were also conducted in recent years. In general, the mixing ratios of most hydrocarbons in the PRD did fall within the ranges of hydrocarbons measured in the 43 Chinese cities (Barletta et al., 2005; Tang et al., 2007; Barletta et al., 2008; Liu et al., 2008b; Liu et al., 2008c). Chan et al. (2006b) performed a VOC study in industrial (T1), industrial-urban (T2), and industrial-suburban (T3) atmospheres of the PRD region. The sampling sites are throughout Dongguan, Foshan, Guangzhou, Jiangmen, and Zhongshan. Toluene was the most abundant NMHC quantified (T1: 13.5ppbv; T2:11.5ppbv; T3: 7.3ppbv). Ethane, ethene, ethyne, propane, n-butane, i-pentane, benzene, and m-xylene were the next most abundant VOCs. Tang et al. (2007) reported that both toluene and ethyne are among the three most abundant hydrocarbons in all three PRD sites. Propane is the most abundant hydrocarbon in Guangzhou, while ethene and ethane are the third abundant hydrocarbons in Panyu and Dinghu Mountain, respectively. For the roadside samples, Guangzhou had much higher level of propane and butanes, while Qingxi Township, Dongguan, had much higher level of aromatic hydrocarbons. Barletta et al. (2008) measured NMHCs concentrations in Guangzhou and Dongguan, two important urban centres of the PRD region. Propane was the most abundant species in Guangzhou, with an average mixing ratio of 6.8 ppbv (±0.7 ppbv), compared to 2.5±0.2 ppbv in Dongguan. Toluene was the most abundant hydrocarbon in Dongguan (6.1±0.8 ppbv, compared to 5.9±0.7 ppbv in Guangzhou). When compared the levels of VOCs with a few urban areas round world, NMHCs levels in Hong Kong were generally lower or comparable to those of other overseas cities. Guo et al. (2004a) reported that the average concentrations of most alkanes and alkenes at Tsuen Wan and Central/Western were generally close to those measured in Europe (Derwent and Davies, 2000), but much lower than those found in Asia and South America (Blake and Rowland, 1995; Grosjean et al., 1998; Morikawa et al., 1998; Barletta et al., 2002). For example, the mean propane concentrations in Mexico City, Santiago, Porto Alegre and Karachi were 10.4, 270.1, 99.4, 80.5 μg/m3, respectively. In this study, the propane concentration was only 4.2l μg/m3. The propane concentration in these cities is 20–79 times that observed in the study of Guo et al. (2004a). Tang et al. (2007) concluded that the mixing ratios of most hydrocarbons in Guangzhou did fall within the ranges of hydrocarbons measured in the 43 Chinese cities. Toluene (10.0 ppbv) and i-butane (4.5 ppbv) in Guangzhou were at the upper end of the ranges (0.4–11.2 ppbv and 0.4–4.6 ppbv, respectively) in 43 Chinese cities. Propane, toluene, ethyne, ethene and i-butane were observed to be higher (1.1–10 times) in Guangzhou than in most other Asian cities i.e. Ulsan, Korea (Na et al. 2001); Kathmandu, Nepal (Sharma et al. 2000); Hong Kong (So and Wang, 2004); Ahmedabad, India (Sahu et al. 2006); and Kaohsiung, Taiwan (Chang et al. 2005), except for Taipei, Taiwan (Wu et al. 2006) and Karachi, Pakistan (Barletta et al. 2002). Diurnal and Seasonal variations Most studies investigated the diurnal and seasonal variations of the VOCs in the PRD region. The total NMHC levels at the urban sites in Hong Kong generally varied according to the season, with high levels in winter and low levels in summer (Lee et al., 2002a; So and Wang, 2004; Guo, et al., 2004a; Guo et al. 2004b; Guo et al. 2007). The higher winter values A-18
were due to the contributions of emissions from the Pearl River Delta under the influence of Asian monsoon system, the weaker vertical mixing and the slower photochemical reaction. In contrast to the urban sites, So and Wang (2004) reported that the roadside site had the highest levels of total NMHCs and alkanes in summer. In particular, compounds such as butane, isobutane and isopentane, which have their major sources from vehicle exhausts, the evaporation of gasoline, and from leakages of liquefied petroleum gas (LPG), all showed the highest levels at roadside in summer and the lowest concentrations in winter. As the roadside site is situated by a road with heavy traffic and surrounded by buildings, the NMHC concentrations at this site are thus dominated by the strength of the emissions from vehicles and are less affected by the changes in regional-scale air flow patterns, as compared to other three sites. The high temperatures in summer could enhance the evaporation of fuel, contributing to a higher level of butane, isobutane, isopentane, and other NMHCs in the summer season, similar observations of elevated summertime concentrations of typical fuel components such as butanes and pentanes. Tang et al. (2007) performed a study on diurnal variation of NMHCs in the PRD. The measurements revealed that the diurnal patterns of the hydrocarbons differed between each of the sampling locations. High levels were observed in the morning and evening for most hydrocarbons and total NMHC in Guangzhou. The two peak pattern indicates that the major sources of these hydrocarbons were traffic emissions. Spatial variations of ambient VOCs in the PRD region The spatial patterns of NMHCs have been conducted in many studies in the PRD region. So and Wang (2004) performed one study to investigate the spatial variations of four sites in Hong Kong: Hok Tsui (rural); Central / Western (residential); Tsuen Wan (industrial), and Mong Kok (roadside). The results showed that the total NMHCs and the three major functional groups of NMHCs were highest at the roadside site and lowest at the rural site. In general, the average concentrations of hydrocarbons were, in descending order: roadside > industrial > residential > rural. The fact that the highest total NMHC level was found at the roadside area is attributed to the heavy vehicular emissions and to the presence of the surrounding buildings, which is thought to prevent the vehicular emissions from dispersing effectively. The industrial site (TW) felt the impact of emissions from surrounding industrial areas and local sources of traffic. The total NMHC level at this site was slightly lower than those collected from the roadsides, because the samples were collected on the roof-top, and where there was less influence from direct emissions. Compared to the roadside and industrial sites, the total NMHC level at the residential site (C / W) was even lower, because of fewer industrial activities and a lower volume of traffic around the site. As expected, the lowest level of total NMHC was observed at the rural HT site. Guo et al. (2007) reported VOCs at four sites ranging from urban to rural areas in Hong Kong. The four sampling sites are Tap Mun (TM, rural), Central / Western (C / W, urban), Tung Chung (TC, sub-urban), and Yuen Long (YL, sub-urban), respectively. As expected, the average levels of all VOCs at YL were the highest except isoprene, while the average levels of all VOCs at TM were the lowest except isoprene. The average isoprene A-19
level at TM was the highest due to the biogenic emissions. The average VOC levels at TC and C/W were similar except butanes, and in between the two extremes of TM and YL. The highest VOC average levels at YL are likely attributed to strong local urban activities and regional pollutants from the rapidly developing PRD industrial region, especially under the prevailing north-westerly wind during the seasonally high pollution period. C/W and TC are typical urban and suburban areas, respectively, and their ambient VOC levels represent the general VOC patterns in Hong Kong. The comparable VOC levels at TC and C/W suggest an insignificant VOC contribution to TC from the nearby aircraft activities. TM is a rural area without major anthropogenic sources and appears to be less strongly influenced by PRD pollution. As a result, TM generally has the lowest VOC levels compared to the other sites. The characteristics and major sources of VOCs in roadside microenvironments have been conducted in many studies. Chan et al. (2002b) performed a study on the characteristics and concentrations of VOCs in the roadside microenvironments of Hong Kong. VOCs samples were collected in four central businesses, commercial, residential and industrial districts of metropolitan Hong Kong. The VOC concentrations, especially toluene (74.9 μg m-3), benzene (25.9 μg m-3) and chlorinated VOCs in Hong Kong were high when compared with those in most developed cities. The average benzene/toluene ratio in Hong Kong was 0.5 suggesting that vehicular emission was the dominant VOC source in most areas of Hong Kong. There were strong deviations in benzene/toluene, benzene/ethylbenzene and benzene/(m+p-xylene) ratios in the commercial district, and highly chlorinated VOC in the industrial and commercial districts. These suggest that there were other benzene and VOC sources overlying on the high background VOC concentrations in these districts. The common usage of organic solvents in the building and construction industries, and in the small industries in the industrial and commercial districts were believed to be important sources of VOC in Hong Kong Lee et al. (2002a) also measured VOCs levels at five roadside sampling sites in Hong Kong. The concentrations of VOCs ranged from undetectable to 1396 μg/m3. Among all of the VOC species, toluene has the highest concentration. Benzene, toluene, ethylbenzene and xylenes (BTEX) were the major constituents (more than 60% in composition of total VOC detected), mainly contributed from mobile sources. Similar to other Asian cities, the VOC levels measured in urban areas in Hong Kong were affected both by automobile exhaust and industrial emissions. High toluene to benzene ratios (average T/B ratio > 5) was also found in Hong Kong as in other Asian cities. In general, VOC concentrations in the winter were higher than those measured in the summer (winter to summer ratio > 1). The VOCs levels in roadside environments in three cities (Guangzhou, Macao, and Nanhai) in the PRD region were reported by Wang et al. (2002). Guangzhou had the highest benzene levels in its roadside environment, followed by Macao and Nanhai. Ethylbenzene and xylenes showed excellent correlations (r2>0.94) in all three cities, indicating that the two have common vehicular sources. On the other hand, the correlation of benzene with ethylbenzene was much weaker, especially in Guangzhou and Zhuhai. This may suggest that benzene had important sources other than vehicular emissions in Guangzhou and Nanhai.
A-20
Source apportionment and emission inventory Many studies on NMHCs distribution have been conducted in China, especially in South China. A study in 43 Chinese cities indicated that vehicular emissions were the major source of NMHCs in 10 cities, while coal and bio fuel combustions were the major sources in 15 other cities (Barletta et al., 2005). A study of NMHCs in the PRD industrial, industrial-urban, and industrial- suburban atmospheres showed that industrial emissions had great influence on the ambient levels of NMHCs (Chan et al., 2006b). Barletta et al. (2008) reported that vehicular emission appears to be the dominant source of NMHCs measured in Guangzhou. By contrast, selected species (including toluene) in many of the Dongguan samples were influenced by an additional source, most likely related to industrial activities. A specific B/T ratio ( 0.8)
organic and inorganic
sites in seven days of 1993-1994; the sampling time is
A poor correlation between V and B(b,k)F, suggesting that vehicular emission and
species at five AQMN
the same period of HKEPD TSP samples
not heavy oil combustion is the most important source of B(b,k)F
sites (Zheng and Fang,
Up to 80 organic species of n-alkanes (C14-C36), fatty
For NH4+ and SO4=, R close to 0.8 is found with Pb, K, As, and B(b,k)F
2000)
acids, alkanols and PAHs by GC-MS
NO3- behaves differently when compared to SO4= and NH4+, suggesting that NH4+
EPD data: Al, Ca, Mg, Pb, V, Mn, Fe, Ni, Zn, EC, Cd,
associates with secondary SO4= and not with NO3-
correlations
the
CW, HKS, KT, KC, MK of the AQMN
Ba, Cu, As, Be, Hg, and Cr by ICP and AAS EPD data: Na+, Cl-, NH4+, NO3-, SO4=, Br-, K+ by IC Correlation coefficients of certain spices versus other species Chemical mass closure were used to estimate the sources contributions
A-40
Table A–2. (Cont’d) Locations*
Objective
Characterize
the
HT rural site
Measurement+ and Data Analysis Methods
24-hr averaged TSP and PM10 samples were collected
Major Finding
The seasonal variation in NO3−, SO4=, NH4+-N mass concentrations shows a
chemical composition
simultaneously using two High-vol samplers every
summer minimum and a winter maximum
of major ions,
third day from 04/95 to 04/96 to obtain 120 sets of
The Br and I show low levels in summer and high levels in winter. The sources of
halogen elements and
samples
Br are mainly marine but the sources of I are from natural and anthropogenic
−
− 3
=
+ 4
+
+
++
++
dusts in TSP and PM10
Cl , NO , SO4 , NH -N, Na , K , Mg , Ca
at HT rural site (Cheng
Al, Br, I, Fe, As, Sb, Se, Sc, V, Zn and total Mg, Ca,
by IC
Four sources for TSP: marine (Pc1, 60%), fossil combustion, vehicle, municipal
et al., 2000)
Na, Cl by INAA
incineration and nonferrous metal production emission (Pc2, 16%), fossil fuel
Pb by ICP-MS
combustion including possible biomass burning (Pc3, 6%), fossil combustion and
Factor analysis (SPSS for window 6.0 release SPSS
ocean emission (Pc4, 4%)
Inc, 1993)
Four sources for PM10: fossil combustion and anthropogenic emission (Pc1, 58%),
sources, including possible biomass emission
sea salt (Pc2, 16%), fossil fuel combustion, vehicle and possible biomass emissions (Pc3, 7%), mineral dust (Pc4, 5%)
A-41
Table A–2. (Cont’d) Objective
Locations*
Measurement+ and Data Analysis Methods
A-42
Major Finding
Table A–2. (Cont’d) Locations*
Objective
Reviews
of
air
No sampling
Measurement+ and Data Analysis Methods
No sampling
Major Finding
In the HKEPD’s PM2.5 database, the annual average PM2.5 concentrations on the
pollution in mega cities
roadside and at urban and rural sites were 58, 34 and 24 µg/m3, respectively
in China
At the roadside and urban sites, carbonaceous aerosols accounted for 52-75% of
(Chan and Yao, 2008)
the PM2.5 mass, followed by (NH4)2SO4 (assuming complete neutralization of SO4= aerosols) accounting for 23-37% of the PM2.5 mass. At the background site, (NH4)2SO4 was the largest contributor to the PM2.5 mass, accounting for 51% of the PM2.5 mass, while the carbonaceous aerosols only accounted for 32% of the PM2.5 mass The estimated regional contribution to PM10 collected at HKEPD stations varied from about 40% to more than 60% in various studies due to the different approaches used
A-43
Table A–2. (Cont’d) Locations*
Objective c.
Measurement+ and Data Analysis Methods
Major Finding
Meteorological
effects on PM and its chemical species Analyze
the
Measurements of hourly SO2 and TSP for once every
The highest TSP in winter, followed by fall, spring, and summer
six days were collected by HKEPD.
The high TSP are apparent in the days with the northerly and westerly winds from
concentrations (Cheng
24-hr averaged SO2 and TSP at the three stations from
the continent, while the days with the southerly winds possess lower TSP
and Lam, 1998)
1986 to 1995 were most intensively analyzed in this
The 70% of the severe TSP days occur in the days with the northerly and westerly
study
winds
Meteorological data were obtained from Hong Kong
The industry in north and west region of HK is responsible for 22% of TSP
affect on TSP
winds
Three sites (CW, KT, SSP) of AQMN
mass
Observation. Pollution roses for SO2 and TSP Grouping data based on wind direction Characterize
UST rural site
24-hr
inorganic and organic
Lantau Island in the south-west region
simultaneously by two High-vol samplers at both sites
compared to normal days
aerosols in TSP during
of HK
on
For dust samples, enrichment factors of Ca, Ni, Co, and V are less than or close to 1,
dust
storm
the
events
(Fang et al., 1999)
averaged
TSP
samples
01/04/96-02/04/96
were
(normal
collected
day)
and
High levels of Al, Ca, Fe, S, and Cl and low levels of K and Zn during dust storm
09/05/96-10/05/96 (Dust storm), obtaining a total of
while Cu, Pb, Cd, and S are higher than 1
four samples
The normal day samples exhibited local characteristics while dust storm day
Aliphatics, PAHs, esters and alkanols by GC-MS
samples suggested outside sources, in terms of SEOC
Elements by XRF and ICP-MS
The aeolian input includes high plant wax content; a low ratio of C18:1/C18:0 FAs in
The wind field calculations were made using the Penn State/NCAR
(National
Center
for
Atmospheric
Research) Mesoscale Model (MM5) (Grell et al., 1994); The initial Þeld and boundary conditions were extracted
from
National
Meteorological
(NMC) synoptic model in 12-h intervals
A-44
Center
the dust samples indicated aging of aerosols due to long-range transport
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
Major Finding
Enrichment factors CPI, Cmax (carbon number maximum), U:R and correlation coefficients among the indices Investigate the effects
Seven urban sites of AQMN: YL, TP,
24-hr averaged TSP and PM10 samples were collected
Anthropogenic emissions together with mineral dust can be transported to HK from
of Asia dust storm on
TW, ST, SSP, KT, and TM
every sixth day by a High-vol sampler at each site
Mainland China during dust storm event
As, Ba, Ca, Cd, Cr, Cu, Fe, Mg, Ni, Pb, Zn by ICP-AES
Concentrations of crustal elements and anthropogenic species were 2-14 higher in
chemical species of
−
−
= 4
+ 4
+
+
PM10
and AAS; Cl , NO3 , SO , NH , Na , K by IC; TC by
dust storm than non dust storm period
(Cao et al., 2003a)
TGM
Mineral dust contributed 41% to PM10 during dust storm event, 2 times higher than
Hourly PM10 mass concentration by TEOM at each site
normal days
from 15/04/98 to 24/04/98 Enrichment factor Pie charts Back trajectory by HYSPLIT (Draxler and Hess, 1998) Investigate the cool season episodes
pollution (Lee
Hills, 2003)
and
CT urban site of AQMN
Hourly PM10 mass concentration was recorded by EPD
Seven pollution episodes were found with PM10 concentrations exceeding the HK
at CT from 1996 to 2002
AQO
Meteorological data was obtained from Hong Kong
Six of episodes are related to the continental northeast monsoon system, and one is
Observation
caused by cyclogenesis over southwestern China
Environmental factors related to pollution episodes: the
High concentrations of As, Mn, V, non-sea salt SO4= and NH4+_N indicate the
northeast monsoon, dust storm coastal sea breezes
impact of fossil fuel burning, while enhanced Al, Ca, Fe, and Mn show the effect of
Air mass back trajectories by HYSPLIT (Draxler and
mineral dust
Hess, 1998) Web-based satellite images by TOMS
A-45
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
Major Finding
Study a severe air
Four urban sites of AQMN: two western
Hourly PM10 mass concentrations was recorded by
The air pollution episode on 28/12/99-30/12/99 caused by a local forest fire and
pollution episode on
sites of TC and YL, two central and
EPD
long-rang transport
28/12/99-30/12/99
eastern sites of CW and KC
Speciation data was also from EPD
The pollutants were brought down to the surface by the local land-sea breeze
Meteorological data was obtained from Hong Kong
circulation
Observation
High loadings of K+
Multi-techniques including remote sensing techniques,
OC/EC ratio is 7 and 6 at TC and YL, respectively
(Fung et al., 2005)
chemical species characterization, a three-dimensional mesoscale meteorological model MM5 and back trajectory Effect
of
Mixing
Height
(MH)
and
long-rang transport on NO3−, SO4= , and NH4+,
UST coastal/rural site
UST: Seven 24-hr averaged PM2.5 samples in the
Under the influence marine air masses, the measured concentrations of SO4=, NH4+,
TST urban site with significant traffic
03/00-05/00 and 06/00-07/00; Another seven 6/12-hr
and NO3− correlate well with the MH
HMT urban site
PM2.5 in 10/00-11/00
~40% of SO4= and NH4+ in Hong Kong was from continental air masses and
TST: ten 24-hr averaged PM2.5 samples from 05/12/00
non-local sources
A-46
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
(Pathak et al., 2003)
Major Finding
to 16/12/00
The influence of the northly/easterly continental air masses on NO3− concentration
HMT: eleven 24-hr averaged PM2.5 samples in
and acidity was not obvious
04/01-05 /01
The strong acidity of PM2.5 in Hong Kong was higher than the acidity reported
Using a Harvard honeycomb denuder/filter-pack
elsewhere in the literature
system; In the honeycomb denuder system, the first denuder (1) was coated with 1% Na2CO3/1% glycerol in a 50% H2O/50% methanol mixture to absorb SO2, HNO3, and HCl gases. The second denuder (2) was coated with 4% citric acid and 2% glycerol in methanol solution to absorb NH3 Cl−, NO3−, SO4= , NH4+, Na+ by IC The estimated in situ acidity of PM2.5 ([H+]AIM2) predicted by the thermodynamic aerosol inorganic model (AIM2; Clegg et al., 1998) Air mass back trajectory by HYSPLIT (Draxler and Hess, 1998) Investigate the effect of
MK roadside site
meteorological –
characteristics PM2.5, characteristics
on seasonal
2m from Lai Chi Kok
Road and 2m above ground level TW urban site
and –
on the rooftop of a
The sampling and chemical analysis methods are the
PM2.5 mass concentrations were highest in winter, followed by fall, at the
same to Louie et al. (2005a)
urban-scale and regional-scale sites
Air parcel back trajectory by HYSPLIT (Draxler and
Winter high PM2.5 level resulted from the elevated OM (OC ×1.4) and NH4NO3
Hess, 1998) and residence time analyses
EC was spatially inhomogeneous, increasing with the nearby traffic density from
Bar charts for seasonal variations of species
the regional-scale to the middle-scale roadside sites, but showed limited seasonal
regional transport of
government building at ~15-18m above
Calculation of EF (defined as the ratio of the episode
variability
PM2.5 (Louie et al.,
ground level
mass fraction to the average mass fraction of a species)
Local mobile source emissions dominated ambient OC and EC concentrations, but
2005b)
HT rural site –
vegetative burning also contributes in winter on a hillside platform
Local and regional sources both contributed to the formation of a pollution episode
A-47
Table A–2. (Cont’d) Locations*
Objective Examine
the
Major Finding
One-hr BC mass concentrations were continuously
Hourly BC concentrations ranged from 63.0 ng/m3 to 17.3 mg/m3
influences of long-rang
recorded by a Model AE-42 Aethalometer at HT site
During the winter, high BC concentrations occurred frequently as a result of
transport
from 04/06/04 to 31/05/05
southward long-range transport of polluted air masses in the boundary layer over
concentrations of BC
Air parcel back trajectory by HYSPLIT (Draxler and
the Asian continent. Anthropogenic emissions in coastal areas of southeastern
in four seasons at HT
Hess, 1998)
China were the major potential sources for the observed pollutants.
rural site (Cheng et al.,
Potential source contribution to air pollution at HT site
During the summer, high BC concentrations were measured occasionally when the
2006c)
by PSCF model (Ashbaugh, 1993)
air masses came from the northwest. These anthropogenic pollutants were found to
on
HT rural site
Measurement+ and Data Analysis Methods
the
be regional in nature, originating from sources in the Pearl River Delta (PRD) region, which included emissions from residential and agricultural combustion, industry, power plants, motor vehicles, and ships. Study the influences of
231 24-hr averaged PM10 samples, once every sixth day
Continental air masses associated with the East Asian winter monsoon contain high
using High-vol samplers from 1998 to 2001
contents of mineral dust and anthropogenic species such as nss-SO4=
(LRT) and the effects
Cl−, NO3−, SO4=, NH4+, Na+, K+ by IC
Wet scavenging significantly depletes the concentrations of chemical species.
of
Ca, Mg, Al, Mn, Fe++, Fe3+ and by ICP_AES
Marine air masses associated with the summer monsoon or high-pressure systems
Comparison with other study
near Japan contain large amounts of sea-salt species
Surface pressure maps
Stagnant air is not the only factor to govern the elevated species levels in aerosols,
Five-day back trajectory at a height of 500m and
but the nature of air masses affecting Hong Kong before sampling should also be
1000m above ground by HYSPLIT4 (Draxler and
taken into consideration
long-range
transport
different systems
weather on
the
concentrations of ions in
PM10 (Wai
and
Tanner, 2005)
CW of AQMN, represent a regional site
Hess, 1998) for data grouping Scatter plots between ions in different group Chemical transport modeling results Relevant meteorological data at Waglan Island
d. Carbon: OC, EC, WSOC, SEOC
A-48
Table A–2. (Cont’d) Objective Investigate PM2.5 and
Locations* PU roadside site
PM10 mass and carbon – 1.5m above the ground content (Ho et al., 2002a)
–
1m away from the major road
KT urban site –
25m high rooftop
HT background site
Measurement+ and Data Analysis Methods
Major Finding
At PU and KT, 24-hr averaged PM2.5 and PM10 samples were collected simultaneously using two High-vol samplers every sixth day from 23/11/00 to 23/02/01 At HT, 24-hr averaged PM2.5 and PM10 were collected simultaneously using two High-vol samplers every twelfth day from 23/11/00 to 23/02/01 A total of 70 samples were obtained OC and EC by TMO Two-tailed t-tests show no significant difference between the mass concentrations collected by the High-Vol sampler and R&R 2000 sampler
The average concentrations at the three sites ranged from 73 to 84 µg/m3 for PM10
Correlation coefficients were given between PM2.5 and
and from 42 to 57 µg/m3 for PM2.5
PM10.
PM2.5 constitutes 53-78% of PM10
OC/EC ratios
OC and EC are major constituents of PM2.5 The correlation between PM2.5 and PM10 was high at KT and HT, suggesting similar source impact For PM2.5, the highest mean concentration of OC was at KT while the highest mean concentration of EC at PU OC/EC ratios for PM2.5 and PM10 were less than 2 at PU and KT sites while the ratio exceeded 3 at HT background site
A-49
Table A–2. (Cont’d) Locations*
Objective Study the seasonal and
PU campus
spatial distribution of – 6m above ground and 8m away from OC, EC, WSOC, and isotopes in PM10 and (Ho
PM2.5
et
al.,
the Hong Chong road
Measurement+ and Data Analysis Methods
Major Finding
24-hr averaged PM10 and PM2.5 samples were collected simultaneously every sixth day at PU and KT, but every 12th day at HT, by High-vol and R&P 2000
KT urban site of AQMN
samplers from 11/00 to 02/01 and 06/01 to 08/01,
HT rural site
obtaining a total of 100 samples OC/EC by IMP_TOR and TMO; WSOC by IC;
2006c)
δ13C(OC) and δ13C(EC) by Finnigan MAT252 The difference between the High-vol and the R&P 2000 sampler was less than 7% for PM10 and PM2.5 mass Ternary diagrams for PM2.5 WSOC, WIOC, and EC in TC to display the different sources of carbon between 80-90% of the PM10 WSOC was in the PM2.5 fraction; wintertime WSOC was
urban and rural sites WSOC/EC as a indicator of aged aerosol
~1.5-2 times higher than in summer with less spatial variation
Based on the scatter plots of δ C(OC) versus
WSOC accounts for 41-54% of TC at HT, 13-20% at PU and 21-24% at KT
δ C(EC), grouping data
WSOC/EC 50%) and exhibited
collected by two Partisol sampler every sixth day at
the largest site-to-site variations
each site for a year from 06/11/00 to 26/10/01
Ammoniated sulfate concentrations showed limited urban-rural contrasts,
OC and EC by IMP_TOR; Ions (Cl-, NO3-, SO4=) by IC;
suggesting regional distribution
NH
+ 4
+
+
by AC; Na and K by AAS; 40 elements from
The local influence of vehicle exhaust are significant at the urban sites Long-range transport of (NH4)2SO4 and Asian dust cause regional impacts at all
government building at ~15-18m above
Na to U by XRF
ground level
Correlations and scatter plots for measurement validity
three sites
Chemical mass closure for PM2.5
Marine influences accounted for 7% of the mass at the HT background site
HT background site –
on a hillside platform
Ternary diagrams for Conceptual PM2.5 source model
A-53
Table A–2. (Cont’d) Locations*
Objective Investigate
the
PU campus
Measurement+ and Data Analysis Methods
Major Finding
24-hr averaged PM10 and PM2.5 samples were collected
Mass closure improved when separate factors (1.4 and 1.9 respectively) were used
seasonal variations and – 6m above ground and 8m away from
simultaneously every sixth day at PU and KT sites, but
to convert water-soluble organic carbon (WSOC) and water-insoluble organic
mass closure of PM10 the Hong Chong road
every twelfth day at HT, by two High-vol samplers
carbon (WINSOC) into corresponding organic masses
and PM2.5 (Ho et al.,
KT urban site of AQMN
from 11/00 to 02/01 and 06/01 to 08/01
The urban sites showed high percentages of water-soluble ions (>50%) in winter
2006a)
HT rural site
OC/EC by IMP_TOR and TMO; Ions (Cl--, NO3-, SO4=,
and high percentages of carbonaceous species (>28%) in summer
+ 4
+
+
NH ) by IC; Na and K by AAS; other elements (Al,
At HT the major components of PM10 were sea salt (25%), sulfate (18%) and crustal
Ca, Fe, Ti, Mg, Zn, As, Cr, Cu, Sr, and V) by ICP-MS
matter (9%) while the major component for PM2.5 during winter is sulfate (31%)
Statistics for PM, OC, EC mass concentrations
and organic aerosols (21%)
Bar charts for species mass concentrations and percentages Mass closure for PM10 and PM2.5 24-hr averaged PM10 every sixth day by High-vol
The annual average SOC as estimated by the PMF method was 4 μg/m3 while the
seasonality of SOC and TC, TP, TW, and YL, on roof tops of four
sampler at each site from 1998 to 2001
summer average was 2 μg /m3 and the winter average was 7 μg /m3. In comparison,
its relationship with to six stories) and one roadside site (MK)
OC and EC by NIOSH 5040 (Sin et al., 2002);
Investigate
secondary
the Nine general sites (CW, KC, KT, SSP, ST,
sulfate of AQMN
(Yuan et al., 2006b)
−
−
−
water-soluble ions (Br , Cl , NO3 , SO
= 4
+ 4
+
, NH , K ,
+
the method that uses EC as a tracer for primary carbonaceous aerosol sources to derive SOC overestimated SOC by 70–212% for the summer samples and by
Na ) by IC; elements (Al, Ba, Be, Ca, Cd, Cr, Cu, Fe,
4–43% for the winter samples
Mg, Mn, Ni, Pb, V and Zn) by ICP-AES; and As, Hg
The overestimation of SOC by the EC tracer method is due to a mixture of primary
and Se by FIA-AA
carbon sources varying in time and space
Comparison of SOC between PMF and EC tracer
Based on PMF, the wintertime SOC is over 4 times summertime SOC, which is due
method
to the meteorological conditions unique to HK (high rainfall, MH, and WD in
Time series and scatter plots for the relationship between SOC and secondary SO
= 4
summer versus high temperature, sunlight, and WD in winter) The wintertime primary OC is 1.6 times summertime
Conceptual model for the the relative production ratio
SOC and secondary sulfate had synchronous seasonal variation and were correlated
of SOC and secondary sulfate
in individual seasons, suggesting common factors that control their formation The presence of SOC was enhanced more than that of sulfate in the winter is a
A-54
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
Major Finding combined result of the temperature-induced favorable partitioning of SOC into the aerosol phase and more abundant SOC precursors (total non-methane hydrocarbon (TNMHC))
Determine long-term
MK roadside site
24-hr (0000-0024 LST) averaged PM2.5 samples were
Statistical analysis showed that the mean difference in PM2.5 between 00/01 and
trend
TW urban site (on the rooftop of a
collected by two Partisol sampler every sixth day at
04/05 studies at any of the three sites was significant at 2σ level (two standard
15-18m government building)
each
errors)
and
variations
spatial of
PM2.5
site
during
11/6/00-10/26/01
and
03/11/04-29/10/05
Substantial increases in PM2.5 mass concentration were found in TW (from 34±3 to
composition (So et al.,
OC/EC by IMP_TOR
39±2 μg /m3) and HT (from 23±2 to 28±2 μg /m3). In contrast, a reduction was
2007)
Cl-, NO3-, SO4= by IC
observed in MK (from 58±2 to 53±3 μg /m3)
NH4+ by AC
The TC levels decreased at MK from 00/01 to 04/05 (29%-32%) due to the
Na+ and K+ by AAS
formulation and implementation of a series of vehicular emission control strategies
Forty elements (from Al to U) by XRF
The apparent increase of NO3-, SO4=, NH4+ over the territory indicated the
Bar charts for PM2.5 mass concentrations comparison
deterioration of regional air quality
between annual means
The ambient PM2.5 levels in HK (from 34±3 μg /m3 in 00/01 to 39±2 μg /m3 in
Mass closure at each site
04/05) were high as compared to the US NAAQS standards of 15 μg /m3
mass
and
chemical
HT rural/coastal site
Study the uncertainties
HK
TSP samples were collected in HK, GZ, and on the
Neither of the two assumptions is valid: (1) PEC (pyrolytically generated EC)
in charring correction
GZ in China
ACS during a cruise over periods of 4-72 h at flow
evolves before native EC evolves in the analysis or (2) PEC and native EC have the
in the analysis of OC
Nanjing (NJ) in China
rates of 0.61-1.13 m3/min using a High-vol sampler
same apparent light absorption coefficient (ó) at the monitoring light wavelength
and
South China Sea (SCS)
PM2.5 samples were collected in NJ and HK over 12-h
Minimizing charring improves the accuracy of the OC/EC split in thermal/optical
EC
by
Thermal/Optical
Cheju (CJ) in South Korea
3
periods at a flow rate of 1.13 m /min using a High-vol
method (Yang and Yu,
sampler
2002)
PM5 samples were collected in CJ at a flow rate of 0.50 m3/min
using
a
High-vol
particle
trap
impactor/denuder sampler designed by CalTech Charcoal aerosols were generated in UST lab by
A-55
methods
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
Major Finding
passing charcoal powder (-20+50 mesh) suspended in a methanol solution through the nebulizer followed by a drying column Carbon analysis by a Sunset thermal/optical aerosol carbon analyzer. The following table lists the two thermal methods used in this study, Ace-Asia and UST-3
Characterize
UST coastal/rural site (on the top of a
12-hr averaged PM2.5 samples were collected by a
WSOC accounts for 13-66% of charring, while hexane extractable organic
charring of OM in
the
building)
High-vol sampler from 07/02/01 to 13/02/01 at UST
compounds produce little charring (2%)
thermal analysis (Yu et
Nanjing (NJ) University campus
12-hr averaged PM2.5 samples were collected by
The extent of charring from WSOC, defined as the ratio between PEC to the total
High-vol sampler from 31/01/01 to 05/02/01 at NJ
WSOC, is found to increase with the WSOC loading in each analysis when the
University
loadings are below a 7 µg/m2. For the WSOC loading greater than 7 µg/m2, the
OC and WSOC by TOT (Sunset), see the forth column
percentage of charring remains unchanged, with an average value of 31% and a
al., 2002)
standard deviation of 3% High variability in the extent of charring among aerosol samples from different locations as well as among samples from a single location collected at different times Inorganic salts (e.g., NH4HSO4) enhances the charring of starch and cellulose but reduces the charring of levoglucosan Explore
the
carbon
MK roadside site
The sampling and chemical analysis methods are the
A-56
TC from the IMP_TOR, STN_TOT/TOR, and HKGL_TOT protocols show good
Table A–2. (Cont’d) Locations*
Objective
Measurement+ and Data Analysis Methods
Major Finding
measurement methods
TW urban site
same to Louie et al. (2005a)
agreement within ~10%
for Hong Kong PM
HT background site
Carbon analysis by two laboratories: HKGL with
Best agreement in EC (1.5 it is ammonium poor (AP); if ratio80 ppbv) were observed, with the
outflow of air pollution to Hong
mostly during 12:00-13:00 from 3 March to 26
highest 1-h O3 mixing ratio of 142 ppbv.
Kong (Wang et al., 2003a)
April 2001
The two dominant NMHCs (C2-C8) were ethane (mean: 2368 pptv) and ethyne
Samples analyzed by GC-MS system (University
(mean: 1402 pptv), followed by propane (814 pptv), toluene (540 pptv), benzene
of California, Irvine for analysis)
(492 pptv), ethene (498 pptv), and n-butane (326 pptv). *ethane (mean: 2.91 μg/m3) and ethyne (mean:1.61 μg/m3 ), followed by propane
Investigate
different
emission
sources on air quality in the PRD
The HK site (22.2°N,114.3°E) - The southeastern
tip of Hong
Correlation analysis
(1.46 μg/m3), toluene (2.03 μg/m3), benzene (1.57 μg/m3), ethene (0.57 μg/m3), and
Case studies
n-butane (0.77 μg/m3).
Samples collected in March 2001
The O3/NOz ratio of 25 was found to be a reasonable threshold to separate
Samples analyzed by GC/MS system (University
NOx-limited and VOC limited regimes concerning O3 chemistry.
A-111
Table A–5.
(Cont’d)
Objectives
Locations*
region (Wang et al., 2005b)
Kong Island The
Measurement and Data Analysis Method of California, Irvine for analysis) GZ
Major Finding The urban area is VOC-limited and the non-urban area is NOx limited.
station
(23.16°N,113.28°E)
The
NCAR/Penn
State
Fifth-Generation
Mesoscale Model (MM5) run in these two domains to produce the meteorological fields needed to drive STEM-2K1 The NCEP-NCAR reanalysis data
used to drive
MM5
Correlation plots between individual compounds Modeling simulation Characterize the composition of
Tai O, a rural/coastal area
1-min samples collected from 08/2001 to 11/2002
Chemical ratios such as C2H2/CO, propane/ethane, and toluene/benzene were much
the ‘background’ atmosphere over
- A hill 80 m above sea level
186 canister samples were taken, most samples
higher in winter (C2H2/CO:~7 pptv / ppbv) and lower in summer (C2H2/CO: ~1
the South
acquired between October and December
pptv / ppbv) , reflecting the impact of the outflow of urban/industrial emissions in
China Sea (SCS) and of pollution
Samples analyzed by a 6-column multiple GC/MS
winter and the dominance of aged maritime air masses in summer.
outflow from the industrialized
system (University of California, Irvine for
An analysis of selected ratios of NMHCs indicates that the southern China air mass
PRD region and southern China
analysis)
had undergone intense atmospheric processing. The enhanced ratios of ethyne/propane, benzene/ propane, and toluene/benzene,
(Wang et al., 2005a) Seasonal variation of VOCs
suggested a relatively large emission of ethane and benzene from mobile sources
Comparison with other observations
and of toluene from the use of toluene-rich gasoline and solvents.
The concentration ratio of [n-butane]/[ethane] versus
[propane]/[ethane]
for
analyzing
atmospheric process Investigate hydrocarbon reactivity (Zhang et al., 2007)
Samples collected by canisters from 1 October to
The reactivity of the VOCs is dominated by anthropogenic VOCs, with VOCs from
– a residential and commercial
CW
31 December, 2002.
natural or biogenic sources making a minor contribution during fall, when this field
district
Samples analyzed by GC-MSD (University of
study was conducted.
A-112
Table A–5.
(Cont’d) Locations*
Objectives
YL
Measurement and Data Analysis Method California, Irvine for analysis)
Of the anthropogenic VOCs, reactive aromatics dominate, of which xylenes and
Back trajectory for air mass
An Observation-Based Model (OBM) is used to calculate the sensitivity of the O3
Diurnal variations of VOCs
production to changes in the concentrations of the precursor compounds. Generally
Application of the Observation Based Model
the production of O3 throughout much of the Hong Kong area is limited by VOCs,
- a residential area TC - a residential area TO
Major Finding
toluene are the most important.
while high nitric oxide (NO) concentrations suppress O3 concentration.
- A rural/coastal area TM - remote/rural area Source Apportionment Studies Investigate source apportionment of
ambient
non-methane
hydrocarbons (Guo et al., 2004b)
CW
Samples collected for 24h by canister from
The results of APCS receptor model indicated that 39% and 48% of the total
- roof top station 17~18 m above
10/01/2001 to 30/12/2001.
NMHCs mass concentrations measured at CW and TW were originated from
the ground level
61 air samples collected in CW station
vehicle emissions, respectively. 32% and 36.4% of the total NMHCs were emitted
59 air samples collected in TW station
from the use of solvent and 11% and 19.4% were apportioned to the LPG or
- roof top station 17~18 m above
Samples analyzed by GC-MSD (The Government
natural gas leakage, respectively. 5.2% and 9% of the total NMHCs mass
the ground level
Laboratory of Hong Kong for analysis)
concentrations were attributed to other industrial, commercial and domestic
TW
sources, respectively. Statistics (principal component analysis)
Vehicle emissions and LPG or natural gas leakage were the main sources of
APCS receptor model for source apportionment
C3–C5 alkanes and C3–C5 alkenes while aromatics were predominantly released from paints. The PCA/APCS model was applicable for estimation of sources of
Identify
regional
and
Local
Tai O, a rural/coastal area
1-min samples collected by 2L canister from
The regional and local source contributions to ambient NMVOC levels at the site
- A hill 80m above sea level.
08/2001 to 11/2002
were significantly different.
Most samples acquired between October and
Air
compounds (NMVOCs) (Guo et
December
approximately 39% of the total NMVOC levels, followed by industrial emissions
al., 2006)
Few samples taken in April-July
(35%), gasoline evaporation (14%) and commercial/domestic liquefied petroleum
A 6-column mutiple GC-MS system used to
gas/natural gas use (12%).
contributions
to
ambient
non-methane
volatile
organic
A-113
masses originating from HK, vehicular emissions accounted for
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding
identify and quntify NMVOCs (the University of
Air masses originating from the PRD the industrial emissions accounted for 43%
California , Irvine for analysis)
of the total NMVOC burden, followed by vehicular emissions (32%) and biomass burning (25%).
The ratios of CO to NOy to trace the origin and transport pathway of air mass PCA/APCS
receptor
model
for
source
apportionment Overview
of
atmospheric
TM
24-hr Samples collected by canisters from 09/2002
Urban and sub-urban sites generally gave relatively higher VOC levels when
C/W
to 08/2003.
compared to the rural area.
apportionment of VOCs (Guo et
TC
A total of 248 ambient VOC samples collected
The principal component analysis (PCA) with absolute principal component
al., 2007)
YL
4-hr average PM2.5 & PM10 using two high-Vol
scores (APCS) technique indicated that vehicular emissions made a significant
processing
and
source
samplers at each site from 11/23/2000-2/23/2001
contribution to ambient non-methane VOCs (NMVOCs) levels in urban areas
Every 6 days for a year
(65±36%) and in sub-urban areas (50±28% and 53±41%). At the rural site, almost
Samples analyzed using a combination of gas
half of the measured total NMVOCs was due to combustion sources.
chromatography with flame ionization detection and mass spectrometric detection (the University of California , Irvine for analysis)
Scatter plots PCA/APCS receptor model for quantifying the impact of relevant sources. Emission Studies Investigate the impact of biogenic
The non-hydrostatic meteorological model MM5
The estimated biogenic emissions of isoprene, terpene, and other reactive VOCs
VOC emissions on a ozone episode
used in mesoscale atmospheric phenomena and
(ORVOCs) during this tropical cyclone-related episode are 8500, 3400, and 11
in the PRD region (Wei et al., 2007
tropical cyclone system simulation
300 ton / day, respectively. The ratio of isoprene to the total BVOCs was 36.4%.
A regional chemical transport model, CMAQ,
The simulations show that Guangdong province, particularly the Pearl River Delta
A-114
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding
used to simulate regional air quality in the PRD
(PRD) region, was the area most reactive to biogenic emissions in South China.
region and its surrounding areas
The net formation of O3 from 9:00 to 15:00 h was the highest near the surface and could reach 38 ppb, which include 4 ppb attributed to biogenic impact. The surface O3 budget was dominated by the vertical transport and dry deposition. The horizontal transport and gas-phase chemical production were relatively small in the surface layer.
Investigate
a
biogenic
compound
emission
VOC
A static enclosure measurement approach to
Plant emission data were combined to estimate annual BVOC emissions of 8.6×
inventory
measure isoprene emission rates of 13 tree species
109 gC for Hong Kong. Isoprene, monoterpenes, and other VOCs contributed
Isoprene emissions sampled by
(Tsui et al., 2008)
solid-phase
about 30%, 40%, and 30% of the estimated total annual emissions, respectively.
microextraction sampler (SPME) equipped with
10 tree species contribute about 76% of total annual VOC emissions despite
carobxen / polydimethylsiloxane (CAR/PDMS)
hundreds of plant species are found in Hong Kong country parks.
fiber for 10min, and transferred to a HP 5890 gas
The results can be applied for studying BVOC emissions in nearby southern
chromatograph with flame ionization detector
China and Asian regions that share similar climate and plant distributions.
(GC-FID). ESRI ARCMAP GIS program used Monitoring
VOCs
in
the
43 cities in China
A total of 2-min 158 Samples collected by
Most of the identified NMHCs correlated with ethyne or i-pentane suggesting that
atmosphere of 43 Chinese cities
canister in Jan / Feb 2001
their primary source is combustion or gasoline evaporation.
(Barletta et al., 2005)
Samples analyzed using the detectors being two
10 cities with a B/T in the range of 0.4–0.8, The B/T ratio of about 0.6
FIDs, two ECDs and a MSD (the University of
good correlation with i-pentane, suggest vehicular emissions (vehicular
California, Irvine for analysis)
combustion and gasoline evaporation) are likely the aim sources of the dentified
Correlation plots between individual compounds
15 cities with a B/T greater than 1, Combustion is still the main source of
Spatial distribution of VOCs
NMHCs, the main combustion source is likely coal and/or biofuel usage. It also
and the
hydrocarbons.
appears that natural gas leakage or some other methane/ ethane source impacts NMHC concentrations in those cities.
A-115
Table A–5.
(Cont’d) Locations*
Objectives Investigate
ambient
Guangzhou
and
VOCs
at
Dongguan
(Barletta et al., 2008)
Guangzhou
Measurement and Data Analysis Method districts
(Liwan,
Major Finding
2-min samples collected by 2L canisters in
Propane was the most abundant species in Guangzhou, with an average mixing
Baiyun, Haizhu, Tianhe, Yuexiu,
September 2005
ratio of 6.8 ppbv (±0.7 ppbv S.E.), compared to 2.5 ± 0.2 ppbv in Dongguan.
and Huanpu)
96 samples collected from Guangzhou and
Toluene was the most abundant hydrocarbon in Dongguan (6.1 ± 0.8 ppbv,
Dongguan districts(urban area,
Dongguan
compared to 5.9 ± 0.7 ppbv in Guangzhou)
Shilong, Houjie, Changping, and
Samples analyzed using the detectors being two
Zhongtang)
FIDs, two ECDs and a MSD (the University of
Toluene 10.98 ±1.44 μg/m3 compared to 10.62 ±1.25 μg/m3 in Dongguan.
California, Irvine for analysis)
Based on an analysis of the correlation between vehicular-emitted compounds and
* Propane (12.23 ±1.25 μg/m3) compared to 4.49 ±0.36 μg/m3 in Dongguan.
the measured NMHCs, together with the benzene-to-toluene (B/T) ratio, vehicular emission appears to be the dominant source of NMHCs measured in Guangzhou. Correlation plots between individual compound
By contrast, selected species (including toluene) in many of the Dongguan samples were influenced by an additional source, most likely related to industrial activities. Toluene accounted for one-third of ozone production in Dongguan samples that were strongly influenced by industrial emissions. In the remaining Dongguan samples and in Guangzhou, toluene contribution to ozone formation was 16–19%. Although propane was the most abundant NMHC measured in Guangzhou, it had a minor role in ozone production, with a contribution of about 1%.
Investigate aromatic VOCs in
Subway
Samples collected by stainless steel adsorbent
The mean VOC exposure levels in roadway transports were about several times
public transportation modes in
Taxis
tubes in both non-rush hours (14:00–16:00) and
higher than those in railway transport due to difference related to fuel type,
Guangzhou (Chan et al., 2003c)
Non-air-conditioned buses
evening
driving lane and vehicle height.
Air-conditioned buses
five-consecutive day period in May 2001
Vehicular emission is the major source of aromatic VOCs.
Thermal desorption (TD) and GC-MSD employed
The VOC levels from evening peak-hour commutes were only slightly higher
for the analysis of the targeted VOCs
than those in afternoon non-peak hour commutes.
78 samples collected by 2L stainless steel canister
Toluene was the most abundant NMHC quantified. Ethane, ethene, ethyne,
Characterize
NMHCs
in
the
Dongguan
rush
hour
(17:00-19:30)
A-116
in
a
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding
between 0900 and 1700 from 11 August to 19
propane, n-butane, i-pentane, benzene, and m-xylene were the next most abundant
Guangzhou
September 2000
VOCs.
Jiangmen
Samples analyzed by GC/MS system
Zhongshan
University of California, Irvine for analysis)
atmosphere of the PRD region
Fosan
(Chan et al., 2006b)
(the
Correlation analysis Ratios
Direct emissions from industrial activities were found to greatly impact the air quality in nearby neighborhoods. Good correlations between isoprene and several short-lived combustion products were found in industrial areas, suggesting that in addition to biogenic sources, anthropogenic emissions may contribute to urban isoprene levels. Investigate
Guangzhou (urban site)
2-3 min samples collected by 2-L canisters in
Elevated regional mixing ratios of many species, including CFCs, HFCs, HCFCs,
sources of halocarbons in urban,
mixing
ratios
and
Panyu (rural site)
2001 and 2004
Halons and other halocarbons when compared with the regional mixing ratios of
semi-urban, and rural sites (Chan
Dinghu Mountain (rural site)
Samples
performed
with
GC-MSD
A-117
(the
the western Pacific and East Asian coast, and the background levels of the
Table A–5.
(Cont’d) Locations*
Objectives et al., 2006a)
Road sites and vehicular tunnels
Measurement and Data Analysis Method University of California, Irvine for analysis)
Major Finding Northern Hemisphere.
in PRD region and Hong Kong Correlation plots between individual compounds Comparison with other cities
There are significant sources of CFC-12, CFC-113, CH2Br2 and CHBr3 in the tunnels and roadside sites, and the excessive CH2Br2 and CHBr3 are related to vehicular exhaust emissions. High levels of methyl halide were simultaneously observed in the oceanic air masses that originated from the coastal areas of Guangdong province and had passed over the Pearl River estuary and Pearl River. Characterize
2-3 h samples collected by cartridge during
The total concentrations of carbonyls were highest in the residential area
compounds and their sources in
ambient
carbonyl
- a bus station near the Guangzhou
Tianhe district
June-September 2003
(67.49±11.90 μg /m3), followed by the downtown area (66.06±9.20 μg /m3),
Guangzhou (Feng et al., 2005)
East Railway Station
Samples were collected between 11:00 a.m.
the industrial area (62.24±11.43 μg /m3), the Botanical Garden (52.66±8.22
and13:00 p.m. every day on 4–7 consecutive days
μg /m3) and the semi-rural area (49.31±15.58 μg /m3).
in each place.
Acetone was the most abundant carbonyl in the ambient air, accounting for
19 carbonyl compound identified by HPLC
24–40% of the total carbonyls. Acetone, formaldehyde and acetaldehyde were
system (HP1100)
the dominated carbonyls and accounted for 63–72% of the total carbonyls. The
Panyu District - near a restaurant Baiyun District - a small downtown location Huangpu District - a industrial area
long lifetime of acetone resulted in higher concentrations in the atmosphere. Figures for concentration levels
A-118
Table A–5.
(Cont’d) Locations*
Objectives
Tianhe District - in a garden
Measurement and Data Analysis Method
Major Finding
Comparison concentration levels with other studies
Panyu District - a semi-rural area Liwan District -a residential area
Good correlations between acetaldehyde, propionaldehyde, 2-butanone, butyraldehyde and benzaldehyde indicated that they came from the same source, e.g. vehicular exhaust. Formaldehyde had bad correlations with other carbonyls, indicating that photochemical reactions might be another important source at noon in summer other than primary emissions from vehicular exhaust. Cooking exhaust from all kinds of restaurants might be the major source of high-molecular weight aldehydes in urban air. Investigate
VOCs
in
the
GZ
Samples collected by
atmosphere of the PRD region (Liu
XK
from 04 October to 03 November 2004
et al., 2008b)
CH
Samples
HZ
pre-concentrator and HP 6890 equipped with two
FS
columns and two detectors (FID &MS)
ZS
134 VOC species identified and quantified
analyzed
stainless steel canister
by
using
DG Correlation analysis Ratios
A-119
a
The total VOC levels varied from 10 ppbv to over 200 ppbv. GZ had a very high level of propane, whereas Xinken, the suburban site, had high mixing ratios of
cryogenic
aromatics.
Table A–5. Objectives
(Cont’d) Locations*
Measurement and Data Analysis Method
Major Finding
Alkanes constituted the largest percentage (>40%) in mixing ratios of the quantified VOCs at six sites; the exception was one major industrial site that was dominated by aromatics (about 52%).
Of the anthropogenic VOCs, alkenes played a predominant role in VOC reactivity at GZ, whereas the contributions of reactive aromatics were more important at XK. VOC correlations suggest that the ambient VOCs at GZ came directly from local sources (i.e., automobiles); those at XK were influenced by both local emissions and transportation of air mass from upwind areas.
A-120
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Investigate source apportionment
GZ
Samples collected by
of VOCs in the PRD region (Liu et al., 2008c)
Major Finding
stainless steel canister in
Vehicle exhaust was the largest source of VOCs, contributing to >50% of ambient
XK
the fall of 2004
VOCs at the three urban sites (Guangzhou, Foshan, and Zhongshan).
CH
Samples analyzed by using HP 6890 combined
LPG leakage played an important role, representing 8–16% of emissions at most
HZ
with GC-MS/FID, 58 VOC species measured
sites in the PRD.
FS ZS
Solvent usage was the biggest emitter of VOCs at Dongguan, an industrial site, Diurnal variations of VOCs
contributing 33% of ambient VOCs. Similarly, at Xinken, a non-urban site, the
Chemical mass balance (CMB) receptor model
evaporation of solvents and coatings was the largest emission source, accounting for 31% of emissions, probably because it was downwind of Dongguan. Local biomass combustion was a noticeable source of VOCs at Xinken; although its contribution was estimated at 14.3%, biomass combustion was the third largest VOC source at this site.
Investigate the role of VOCs and
Xinken
Samples collected by 6L canisters from May to
Local emissions and long-distance transportation of pollutants play important
NOx
GZ (Baiyun hill, Huadu)
September 2000
roles in the regional distribution of VOCs.
ground-level ozone (Shao et al.,
There samples per day collected at 6:00-9:00,
Ambient O3 production is significant in urban areas and also downwind of cities.
2008)
11:00-14:00, and 15:00-18:00, at 1-week intervals
The relative incremental reactivities (RIRs), determined by an observation-based
Samples analyzed by GC-MS system
model, showed that ground-level ozone formation in the Guangzhou urban area is
in
the
production
of
generally limited by the concentrations of VOCs, but there are also measurable Correlation plots between individual compounds
impacts of NOx.
Modeling simulation (OBM) Characterize
NMHCs
in
the
GZ
Samples collected by
stainless steel canister in
Vehicular and industrial emissions were the main sources of NMHCs. The effect
atmosphere of the urban, suburban,
-the urban center surrounded by
the April of 2005
of vehicular emission on the ambient air varied among the three PRD sites. The
and rural sites (Tang et al., 2007)
residential buildings and business
Five canister samples were collected at 8:00,
impact of industrial emissions was widespread and they contributed greatly to the
offices
11:00, 14:00, 17:00 and 20:00 local time each day
high levels of aromatic hydrocarbons, especially toluene.
Samples
Leakage from vehicles fueled by LPG contributed mainly to the high levels of
PY
analyzed
by
using
automated
-the rooftop of an environmental
GC–MS/FID system with two columns (a PLOT
propane and n-butane at the urban GZ site.
monitoring chamber (about 4m
and a DB-1)
Diurnal variations of NMHCs showed that the contribution from vehicular
A-121
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding emissions varied with traffic conditions and were more influenced by fresh
above ground). DM -the hilltop in the Dinghushan
Diurnal variation of VOCs
emissions at the urban site and by aged air at the suburban and rural sites.
Comparison VOCs level with other cities
Ethene, toluene and m/p-xylene were the main contributors to the OFP at the three PRD sites.
Investigate the changing urban and rural emissions on NMHCs in the PRD region (Tang et al., 2008)
DM - the hill top Guangzhou (Liwan district) - the rooftop of a nin-story building Guangzhou (Dongshan district) - the rooftop of a 21-story building
Most 2-min Samples collected by canisters during
NMHC profiles showed significant differences in the exhaust samples between
the spring of 2001 and 2005
gasoline- and LPG-fueled taxis. Propane (47%) was the dominant hydrocarbon in
Samples
performed
with
GC-MSD
(the
University of California, Irvine and
the
that of gasoline-fueled taxis.
Research Center for Environmental Changes
Leakage of LPG is a major source of propane and n-butane in GZ and on DM.
(RCEC), Academia Sinica, Taiwan for analysis)
The increasing use of LPG-fueled buses and taxis, especially in GZ, has
Roadside and tunnel sites in Guangzhou
the exhaust of the LPG-fueled taxis, while ethene (35%) was the dominant one in
significantly affected ambient levels of propane. Comparison VOCs level with other cities Wind roses for analyzing climate condition Correlation plots of individual compound
Characterize
Guangzhou
Most 30min Samples collected by sorbent tubes in
The urban roadside BTEX levels in Guangzhou and Macau were much higher
hydrocarbons in the PRD region
roadside
aromatic
Nanhai
July and November 1996.
than those in Nanhai and reported levels in other North American and European
(Wang et al., 2002)
Macao
Most
Samples taken at the curbsides at
14:00-14:30 each day
least 20m away from bus stops,
Samples
junctions and traffic lights and
(Guangzhou
collected 1.2m above the ground
analysis)
samples
collected
performed institute
at
9:00-9:30
and
cities. Mean concentrations of benzene, toluene, ethylbenzene and xylenes in the
TD-GC-MSD
95.6 μg/m3 (Macau); 20.0, 39.1, 3.0 and 14.2 μg/m3 (Nanhai), respectively. The
three cities were 51.5, 77.3, 17.8 and 81.6 μg/m3 (Guangzhou); 34.9, 85.9, 24.1, with of
geochemistry
for
difference of these BTEX levels in urban roadside microenvironments might be explained by their source strengths and dilution conditions. In Guangzhou and Macau, the high levels of BTEX, particularly benzene, which
level
were probably related to the street canyon effect and traffic density. Macau emissions were predominantly traffic related, and there were sources other than vehicle emission in Guangzhou and Nanhai. Characterize the ozone precursors
Guangzhou(urban site)
Hourly samples collected from 16 October to 5
A-122
In Guangzhou, peak concentrations were usually seen after dusk at the traffic rush
Table A–5.
(Cont’d)
Objectives
Locations*
by observing NMHCs in the PRD
- the roof of a 17-story building Xinken (non-urban site)
(Wang et al., 2008)
Measurement and Data Analysis Method
Major Finding
November 2004.
hour, whereas the maximum in Xinken occurred between late night and early
Samples analyzed by GC-MSD and GC-FID
morning (a: Guangzhou, b: Xinken)
- the rooftop of a 2-story building Principle component analysis Scatter plots of individual compound
High correlation between NOx, CO, and volatile organic compounds (VOCs) in Guangzhou suggests that motor vehicle exhaust is the dominant source of NMHCs in Guangzhou. It is also found that propane, iso-butane, and n-butane accounted for nearly 40% of the total NMHC concentration in Guangzhou, possibly due to the wide use of liquefied petroleum gas (LPG) as motor vehicle fuel in Guangzhou. The air in Xinken is substantially more aged than that in Guangzhou, consistent with the relatively smaller percentage of alkenes found in Xinken than in Guangzhou. 15 min VOC samples collected by 2-L canisters at
Formaldehyde (3.70±4.38 μg /m3) and acetaldehyde (3.33±3.65 μg /m3)
carbonyls and VOCs in forest park
-located in the northern part of the
3-h intervals from 9 to 12 July, 2004
were the two most abundant carbonyls, while the most abundant VOCs were
in south china (Yu et al., 2008)
Huadu District
3-h carbonyl samples collected by cartridges from
isoprene (14.90 μg/m3), followed by o-xylene
9 to 14 July, 2004
The highest concentrations were observed from 03:00 to 06:00 for 1-butene,
Characterize
atmospheric
Prince Hill Forest Park
VOC
samples
chromatography/mass
analyzed spectrometry
by
gas
from 06:00 to 12:00 for isoprene, and from 12:00 to 15:00 for a-pinene. The
(GC/MS,
highest levels for aromatic hydrocarbons occurred during midnight and the
Agilent 6890N/5973N)
lowest in late afternoon.
Carbonyl compounds identified by HPLC system (HP1100)
A-123
(2.21 μg/m3).
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding
20 carbonyls and 8 VOCs identified and quantified
The comparison of concentration levels with other studies Diurnal variations of VOCs and carbonyls
Most C3-C10 carbonyls had higher concentrations from 09:00 to 15:00, and their levels were lower during night-time and often reached the lowest in early morning. Emissions from vegetation and photo-oxidation of gas-phase hydrocarbons were the main sources for some carbonyls and VOCs in this region. Other compounds, such as formaldehyde, acetaldehyde and BTEX, showed anthropogenic sources. Characterize
Streets selected in four districts of
30-mins Samples collected by sorbent tubes
Pedestrian exposure to some toxic VOCs (for example, benzene) was relatively
hazardous VOCs while walking
Guangzhou,
between 07:00 and 19:00 on 1 and 2 February
high. Monocyclic aromatic hydrocarbons were found to be the most abundant
along streets in Guangzhou (Zhao
Yuexiou, Dong Shan and Tianhe
2002
VOCs, and 71% of the samples had benzene levels higher than 30 μg /m3
Samples analyzed by a thermal desorption system
The good correlations between BTEX, PM10 and CO in the streets indicated that
coupled to HP 5972 GC/MSD (The Guangzhou
automotive emission might be their major source.
et al., 2004a)
the
exposure
to
namely
Liwan,
A-124
Table A–5.
(Cont’d) Locations*
Objectives
Measurement and Data Analysis Method
Major Finding
institute of geochemistry for analysis) Overview
regional
integrated
Guangzhou (urban site)
30-60 min samples collected by canisters from 4
Total oxidant was of regional scale covering at least the entire PRD. Its
experiments on air quality over
- a height of about 50 m above
October to 5 November 2004
relationship with O3 precursors was highly non-linear. Photochemical production
PRD (Zhang et al., 2008)
street level
The samples analyzed by GC-FID and GC-MS for
of O3 was sensitive to VOCs at both the Guangzhou urban site and Xinken
55 species of NMHCs (Peking University and
downwind rural site.
Research Center for Environmental Changes,
Hong Kong has been found also in the VOC-limited regime, it is highly likely that
Taiwan for analysis)
the production of O3 in the entire PRD region is VOC-limited.
Xinken (non-urban site)
Modeling simulation (observation based model)
Notes 1
VOC concentrations in tables are expressed in units of μg /m3, if paper is exhibited in units of ppbV, the conversion will be done by assuming ambient conditions of 1 atm and 25 ℃
*
Sampling sites
CB………………. Causeway Bay, an important commercial area with heavy traffic density and many high-rise commercial buildings. CW………………. Central and Western, a residential and commercial district on Hong Kong Island. KC………………. Kwai Chung, urban industrial area with industrial and residential activities, high traffic flow of heavy trucks, good vehicles, buses and commercial van. KT………………. Kwun Tong, Kwun Tong (KT), close to metal and printing industries, residential buildings and roads with light and heaey good vehicles. HT………………. Hok Tsui, located on a hillside platform facing the South China Sea at the southeast end of Hong Kong Island, represents a rural background/regional/transport environment. MK……………….Mongkok, urban area with active commercial and residential activities, high pedestrian and traffic flow of buses, minibuses, vans and taxies. PU………………. Polytechnic University, close to roads with heavy diesel bus, diesel taxi, and passenger car traffic; represents a roadside microenvironment in a downtown area. SSP……………….Sham Shui Po, old commercial and residential area with Mixed commercial and residential activities, medium traffic flow of buses, minbuses, vans and taxies. Tai O…………….. A rural/coastal area in southwest HK , 32km away from the HK urban center to the east and approximately the same distance from Macau/ Zhuhai to the west. TC……………….Tung Chung, a residential area in a new town, but adjacent to the highway and to railway lines that connect the airport to the other islands of Hong Kong. TM……………….Tap Mun, located on a remote island just off the northeast coast of Hong Kong. The surrounding area is rural and sparsely populated.
A-125
TST……………….Tsim Sha Tsui, central business area with mainly commercial activities with relatively medium traffic flow of taxies and private cars. TW………………. Tsuen Wan, a fairly newly developed town located in a mix of residential, commercial and light industrial area in the New Territories. YL……………….Yuen Long, located in a new town in the northwestern part of New Territory Island.
CK……………….Xinken, lied in a less populated coastal area; it is a rural site located~50 km to the southeast of the city center. CH……………….Conghua, a rural site which is located upwind of the PRD region. DM……………….Dinghu mountain, a rural site which is about 85km to the west of GZ and 18km to the northeast of Zhaoqing (a less industrialized city compared with other cities in PRD with about 3.7 millions inhabitants). DG……………….Dongguan, a famous industrial manufactory center of China with about 6.6 million inhabitants) to the west, 30km from Zhongshan (an important industrial city in PRD with 2.4 millions inhabitants) to the south. FS……………….Foshan, a urban site. GZ………………. Guangzhou, situated at the coast of the South China Sea (21~23° N) and experiencing a typical sub-tropical climate. HZ……………….Huizhou, s suburban site which is located upwind of the PRD region. PY……………….Panyu, a suburban site which is is about 20 km from the urban center of GZ to the north, 25km from ZS……………….Zhongshan, a urban site.
A-126
Table A- 6. Summary of Emissions Studies. Objective
Monitoring Station
Measurement and Data Analysis Methods
Major Finding
Hong Kong Biogenic
VOC
Isoprene emission rates of 13 Hong Kong tree species
Biogenic VOC (isoprene)
Isoprene emission rates were measured by static
emission
13 local tree species were measured for their
enclosure measurement approach
Emission
Emission rate
inventory for HK
isoprene emission potential
A branch of one seedling of each species was enclosed in
rate
(µg
(Tsui et al., 2008)
The enclosure system was housed in a growth
a glass chamber at 30°C and 1000 µmol/m2/s light
Genus_species
(µgC/g/h)
isoprene/g/h)
chamber (Conviron, ND, USA)
intensity for 1 hr
Acacia confusa
UD
UD
Tested tree species:
2 samples for each species (total 26 samples)
Bauhinia
Isoprene emissions were sampled by solid-phase
candida
20±2.4
22.7±2.7
microextraction sampler (SPME)
Castanopsis fissa
UD
UD
Cinnamomum burmannii
2.5±0.1
2.8±0.1
Cinnamomum camphora
UD
UD
Cyclobalanopsis edithiae
UD
UD
Eucalyptus citriodora
6.4±1.1
7.3±1.2
Eucalyptus robusta
10±1.1
11.3±1.2
values were applied. For monoterpene emission rates of
Eucalyptus torelliana
4.2±0.6
4.8±0.7
the 13 species, and isoprene and monoterpene emission
Ficus hispida
3.3±0.7
3.7±0.8
rates of all other species, the values from literature were
Ficus microcarpa
UD
UD
used.
Mallotus paniculatus
UD
UD
Emission rate: µg isoprene measured / leaf dry
Macaranga tanarius
UD
UD
Isoprene was analyzed by GC/FID and the detection was confirmed by GC/MS
variegata
var.
Standard emission factors of isoprene, total monoterpene (TMT) and other volatile organic compounds (OVOCs) were assigned to 148 Hong Kong tree species. For tested 13 species, measured isoprene emission rate
weight/sampling hour
UD = undetected
The investigated period was 01/08/2004 to 31/07/2005.
Annual Biogenic VOC emissions: 8.6 x 109 tonneC/yr and contribute 21% of
Emission rate of an area
the total annual VOC load in HK Isoprene: 30%, Monoterpenes: 40% and other VOCs: 30% (of the estimated
= isoprene emission rate of plant
A-127
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods
Major Finding total annual emissions)
D = foliar density [g (leaf day weight) / m2 (ground)] = coefficient of light intensity influence on emissions = coefficient of temperature influence on emissions
10 tree species contribute about 76% of total annual VOC emissions. They are: Gordonia axillaris
Baeckea frutescens
Tree distribution was estimated for country park areas
Schima superba
Cinnamomum parthenoxylon
based on field survey data
Lophostemon confertus
Syzygium jambos
Plant emission data obtained from measurements and the
Cratoxylum cochinchinense
Schefflera heptaphylla
literature, tree distribution estimation data, land use
Litsea rotundifolia
Acacia confusa
information, and meteorological data were combined to estimate annual Biogenic VOC emissions
A-128
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods
Major Finding
Characteristics of
Sampling locations: two famous temples in HK
Four sets of samples for each temple, 2 sets (duplicates)
The pollutant levels of the two temples during peak period were significantly
emissions of air
Sampling height: 1.6m
for peak periods and 2 sets for non-peak periods [i.e.
higher than those during non-peak period. For NOx level, the concentration
pollutants
PM (PM10 and PM2.5), VOCs (38 species),
totally 8 sets of samples]
during the peak period was 2 to 3 times of the non-peak one.
carbonyls, CO, NOx, CH4, NMHC, inorganic
Sampling time: 09:00 to 17:00 (8 hours)
Average PM2.5/PM10 ratios were approximately 82%
ions / temples
On site temperature and relative humidity: Q-trak
The total mass of inorganic ions, organic carbon, and elemental carbon
CO, NOx, CH4 and NMHC were collected by air bags
accounted for about 71% in PM2.5 and 72% in PM10.
from
burning of incense
in
Measured inorganic ions:
temples in HK (Wang
et
al.,
-
+
+
−
-
+ 4
2-
Na , K , Cl , NO3 , NH and SO4
Measured carbonyls:
2007)
Carbonyls were collected by DNPH-silica cartridges
The average OC/EC rations ranged from 2.6 to 17 in PM10 and from 4.2 to 18
VOCs were collected by canisters
in PM2.5.
PM were collected by mini-volumes
For measured VOC, benzene methylchloride, and toluene were the most
CO: analyzed by CO analyzer
abundant VOC species.
NOx: analyzed by Chemiluminescence NO-NO2-NOx
Formaldehyde was the most abundant carbonyl compounds, followed by
Analyzer
acetaldehyde
CH4 and NMHC: Direct Methane, Non-Methane
The highest average CO level (10078±1718 µg/m3) was obtained at one
Hydrocarbons Analyzer
temple during peak period exceeded IAQO 8-h Good Class criteria (10000
Carbonyls were measured by HPLC
µg/m3).
VOCs were measured by GC/MSD (according to USEPA method TO-14) OC/EC were quantified by Thermal/Optical Carbon Analyzer Ions were measured by IC
Emission
factor
Source: Vehicles
35 pairs of samples: 9 in summer and 26 in winter of
Carbonyls emission factors: 21.7 to 68.9 mg/veh/km with an average of
of carbonyls in a
Location: Shing Mun Tunnel
2003
35.8±11.9 mg/veh/km
HK tunnel
The inlet sampling station located 686m inside
Sampled with DNPH cartridge
Carbonyl emissions from Diesel-fueled vehicle: 75.1 mg/veh/km
(Ho et al., 2007)
the entrance of the tunnel and the outlet
Sample collected at 800-900mL/min for 1or 2h
Carbonyl emissions from Non-Diesel-fueled vehicle: 10 mg/veh/km
A-129
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods
Major Finding
sampling station was located 350m upwind of
Analyzed with HPLC
Higher emission factors of carbonyls during summer
the exit. (One pair sample)
Traffic count analysis: traffic compositions and volume
The five carbonyls with the largest diesel-fueled vehicle emission factor
Sampling height: 1.5m
were determined by manual counts at the entrance of the
were, in decreasing order, formaldehyde (38.3 mg/veh/km), acetaldehyde
15 Carbonyls:
tunnel at 15-min intervals during sampling periods.
(11.4
Traffic speed surveys wer e conducted using car chasing
mg/veh/km) and benzaldehyde (2.0 mg/veh/km). These five carbonyl
method by a Darwin microwave speed sensor and the
compounds together accounted for 87% of the sum of all diesel-fueled
tachometer sensor equipped vehicle.
vehicle carbonyl emission factors.
The emission factors from tunnel were calculated
The five most abundant carbonyls in non-diesel-fueled vehicle emission at
according to the method of Pierson
the tunnel were, in decreasing order, formaldehyde (3.5 mg/veh/km), acetone
mg/veh/km),
acetone
(5.3
mg/veh/km),
crotonaldehyde
(5.2
(1.8 mg veh/km), methyl ethyl ketone (1.6 mg/veh/km), m,p-tolualdehyde (1.0 mg/veh/km) and acetaldehyde (mg/veh/km). They accounted for 85% of the sum of all non-diesel-fueled vehicle carbonyl emission factors. EFveh = average vehicular emission factor Cout = mass concentrations at exit Cin = mass concentration at entrance A = area of tunel cross-section in m2 U = wind speed in m/s t = sampling duration N = total traffic number L = distance between 2 monitoring stations in km
Emission
of
carcinogenic
Concentrations of chemicals in samples (µg/m3)
Kitchen exhaust air samples were collected
Sampling number: 18
during the peak hours in lunch (12:00-14:00) or
Sampling: high volume air sampling and sampling train
components from
dinner (19:00-21:00) periods.
Sampling train was equipped with a cyclone to remove
commercial
Commercial Kitchens:
particles with PM greater than 2.5µm.
kitchens in HK
- 6 Chinese restaurants
Flow rate: 30L/min
A-130
AHs(total)
Chinese 0.41 1.34 4.47 10.65
Western 5.24 13.27 1.90 0.30
Exotic 4.37 17.32 5.60 15.75
1.81
2.67
5.27
Table A–6. (Cont’d) Objective (To et al., 2007)
Monitoring Station - 6 Western restaurants: 2 local steak houses, 2 fried chicken chain stores, 1 fast-food
Measurement and Data Analysis Methods
Major Finding 3.14
0.96
4.33
PAHs(total)
0.021 0.004 0.30 0.31 0.041 0.39
0.43 0.26 0.22 0.054 0.40 0.079
0.26 0.12 0.65 0.27 0.41 0.27
FAs(total)
** 57.36 46.48 ** 22.51
20.83 28.77 ** 30.49 72.88
** 901.38 ** ** **
AAs(total)
* * * * * 0.184
0.86 4.20 * 0.082 0.27 0.024
* 6.09 5.51 * * 4.47
AHs, PAHs and FAs were analyzed by GC/MS Restaurant number obtained from Government
restaurant and 1 café. - 6 exotic food-servicing areas: 3 Chinese barbecue centres, 1 Korean and 1 Thai and 1 Indian restaurant Target pollutants: AHs, PAHs, FAs, AAs FAs were transesterificated to fatty acid methyl esters (FAMEs) Tested FAMEs - lauric acid methyl ester (C12:0) - myristicacid methyl ester (C14:0) - pentadecanoic acid methyl ester (C15:0) - palmitic acid methyl ester (C16:0)
*AA samples were not collected at these sites. **FA samples were not collected at these sites. N.A. - not applicable.
- palmitoleic acid methyl ester (C16:1n9c) - stearic acid methyl ester (C18:0) - elaidic acid methyl ester (C18:1n9t) - oleic acid methyl ester (C18:1n9c)
Oil fumes from commercial kitchens contain carcinogens such as PAHs and
- linoleic acid methyl ester (C18:2n6c)
AAs and relatively less AHs.
- arachidic acid methyl ester (C20:0)
Concentration of FAs was about 1 to 2 order of magnitude larger than that of
- cis-11-eicosenoic acid methyl ester (C20:1)
AHs and AAs.
- benehic acid methyl ester (C22:0)
Western restaurants: Particulate phase AH constituted 52% of total weight of
- erucic acid methyl ester (C22:1n9)
AHs Exotic food servicing: Particulate phase AH constituted 27% of total weight of Ahs
A-131
Table A–6. (Cont’d) Objective 2006 Hong Kong Air
Pollutants
Monitoring Station
Measurement and Data Analysis Methods
Major Finding
Target pollutants: SO2, NOx, RSP, VOC and
The emission inventory was collected from EPD
Air Pollutant emission inventory in 2006
CO
webpage which did not provide the methodology for the
Tonne/yr
Emission
Sources: Electricity generation, road transport,
emission estimation.
(%)
Inventory
navigation,
Method:
(HKEPD, 2006)
combustion, non-combustion
civil
aviation,
other
fuel
based
on
USEPA
AP-42
EMEP/CORINAIR Emission Inventory Guidebook.
and
Pollutant Source Categories
SO2
NOx
RSP
VOC
CO
Electricity
66,000
41,800
1,860
416
3,370
Generation
(89%)
(44%)
(32%)
(1%)
(5%)
Road
956
21,800
1,810
8,080
62,300
Transport
(1%)
(23%)
(31%)
(20%)
(84%)
3,920
16,700
499
304
2,290
(5%)
(18%)
(9%)
(1%)
(3%)
Civil
294
5,020
21
261
2,020
Aviation
(0.4%)
(5%)
(0.4%)
(1%)
(3%)
2,660
9,530
925
998
4,260
(4%)
(10%)
(16%)
(2%)
(6%)
750
31,200
NA
NA (13%)
(76%)
Navigation
Other Fuel Combustion
1
NonCombustion 2
NA
73,900
94,800
5,860
41,200
74,200
(100%)
(100%)
(100%)
(100%)
(100%)
Total 1. Industrial, Commercial, Domestic and Off-road Transport. 2006 SO2 emissions are based on projected energy data. 2. RSP: includes Quarrying and Construction Aggregate Processing, Type, Brake and Road Surface Wear. Paint and Printing VOC.
A-132
VOC: mainly consists of Consumer Products,
Table A–6. (Cont’d) Objective
Monitoring Station
Carbonyl
15 commercial kitchens were sampled
Measurement and Data Analysis Methods
Major Finding
Sampling during lunch or dinner peak hours
Estimated annual total carbonyls emission rate: 31.9 (tonne/yr)
Sampling duration: 15-60 min
Hong Kong-style fast-food shops contributed the highest total carbonyl
commercial
PFBHA sampling tubes: 20 mL/min; PFPH sampling
emissions per year.
cooking in HK
tubes: 100 mL/min
Annual emission rate of total carbonyls
(Ho et al., 2006d)
Sampled by PFBHA and PFPH sampling tubes
Type of Restaurant
Analyzed by GC/MS
General Chinese Restaurants
9.52
Restaurant-based emissions in moles per day were
HK-Style Fast Food Shop
10.826
estimated using the following formula:
Chinese Barbeque
1.192
Western Fast-Food Shops
4.610
Non-Chinese Restaurants
5.713
Total
31.861
emissions
from
f = exhaust volumetric flow rate C = carbonyl molar concentration t = operation period of the restaurants The total carbonyls emission from commercial cooking Kitchen exhaust air samples were collected
in HK was estimated by surrogate with number of
Targets pollutants: 13 carbonyls
restaurants
A-133
Emission (tonne/yr)
Table A–6. (Cont’d) Objective
Monitoring Station
PM2.5
and
PM2.5: 16 pairs of samples: 4 runs in summer (08/03) and
PM2.5 emission factor (Filter method): 131±37 mg/veh/km (n=16)
12 runs in winter (01/04-02/04)
PM2.5 emission factor (Dust trak): 150±45 mg/veh/km (n=16)
The inlet sampling station located 686m inside
Sampling time: 2h for summer and 1h for winter
NOx emission factor: 878±308 mg/veh/km (n=7)
Mun
the entrance of the tunnel and the outlet
PM2.5
CO emission factor: 1845±434 mg/veh/km (n=4)
sampling station was located 350m upwind of
- Dust trak
Annual vehicle emission flux of PM2.5: 1466.0±414.1 tonne/yr, about 81% of
the exit. (One pair sample)
- Filter-based PM sampling
the 1810 tonne/yr reported by EPD Emission Inventory (2006) for RSP.
Sampling height: 1.5m
NOx
Target pollutants: PM2.5, NOx, CO
- Chemiluminescence NO-NO2-NOx analyzer
Tunnel in HK et
Source: Vehicles Location: Shing Mun Tunnel
emissions
(Cheng
Major Finding
in
gaseous
Shing
Measurement and Data Analysis Methods
al.,
2006b)
CO ambient analyzer The emission factors from tunnel were calculated according to the method of Pierson
EFveh = average vehicular emission factor Cout = mass concentrations at exit Cin = mass concentration at entrance A = area of tunel cross-section in m2 U = wind speed in m/s t = sampling duration N = total traffic number L = distance between 2 monitoring stations in km
Marine Source for
Target pollutant: SO2
The study developed an improved pollution rose, called
Two dominant modes of SO2 variation were identified. One is associated
SO2 levels in HK
Sources: Marine ships
CPWN, and combines it with both PCA and SVD
with weak southerlies and southwesterlies, and is localized around the
(Lau et al., 2005)
Using the data measured in the existing
techniques to analyze the pollutant data recorded in14 air
Kowloon peninsula. A second mode of variation is associated with moderate
A-134
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods
monitoring stations
Major Finding
quality monitoring stations and wind data from two
to strong northwesterlies, and affects most of the territory of the HKSAR.
automatic weather stations
The first mode is associated with the nearby combustion of residual fuel oil
PCA and RPCA were used to identify the dominant
from marine sources in the vicinity of the large container terminal at Kwai
modes of variations in atmospheric and environmental
Chung. The second mode associated with a mixture of sources but mainly
datasets
related to the regional transport from the northwest.
The
proportion of
the total
SO2
concentrations
Residual oil combustion from marine vessels around Kwai Chung terminal
contributed by each of the source regions was estimated
container port and local power plants are responsible, respectively, for 36%
by SVD
and 7% of the total SO2 concentrations measured at local general air quality
To help identify the pollution sources contributing to
monitoring stations.
SO2 levels measured at the stations, simple receptor
With regard to specific locations in the heart of the urban area, local SO2
modeling was carried out suing chemical speciation data
sources, in particular marine vessels burning high sulphur residual fuel oil
from 24-hour high volume PM10 samples collected by
appear to be the major source.
every six days at the urban sites.
The analysis did not identify road vehicle or power plants emissions as important sources of contribution to the local SO2 levels in Hong Kong.
PM
&
VOC
Emission Profiles for
Vehicular
Monitoring Station 1 Ambient site -
Sources in HK (HKPolyU, 2005)
Tsuen
Wan
EPD
Monitoring Station 3 roadsides -
Mongkok EPD Roadside Monitoring Station
-
HKPU
Roadside
Monitoring Station -
Lok Ma Chau Roadsie
Air measurement was carried out at the sampling stations
Emission factors
from 05/2003 to 09/2003 and from 11/2003 to 02/2004
- VOCs: 130.5±44.3 mg/veh/km
VOCs
- Total measured gaseous PAHs: 1.175±0.366 mg/veh/km
Ambient Volatile Organic Canister Sampler (AVOCs)
- Total particulate PAHs: 0.106±0.098.3 mg/veh/km
was used to collected air samples into canisters. The
- Total measured carbonyls: 39.07±21.29 mg/veh/km
samples were analyzed by GC/MSD/FID/ECD with
- PM2.5: 114.80±62.67 mg/veh/km
USEPA Method TO-14
- CO: 1809.68±572.37 mg/veh/km
PAHs
- NOx: 989.78±498.28 mg/veh/km
PAHs were sampled with an Andersen GPSIX PUF
- SO2: 160.39±82.63 mg/veh/km
sampler followed by solvent extraction and HPLC/UV
A-135
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods detector analysis
Station 6 vehicular sources
Carbonyls
-
HKPUCar Park
Carbonyls were collected with DNPH-silica cartridge.
-
Wan Chai LPG Refilling
The cartridges were analyzed by extraction and HPLC according to the USEPA TO-11A method
Station -
Shau Kei Wan Minibus
Portable particle samples were used in ground-based
Station -
PM2.5 Mass and Chemical Species
Cheung Sha Wan Whole Food Market
source sampling of motor vehicles exhausts in tunnel. The collected samples were analyzed by gravimetric,
-
The Peak
x-ray fluorescence, anion chromatography, automated
-
Tuen Mun Bus Depot
colorimetry,
atomic
absorption
spectrophotometry,
thermal/optical reflectance combustion, and GC/MS for
2 highway tunnels -
Shing Mun Tunnel
the chemical speciation. Chemical species to be
-
Tseung Kwan O Tunnel
identified were as follow:
Sources: vehicles
Ammonia
Organic Carbon
Phosphorus
Elemental
Target pollutants: PM2.5, VOCs (115 species), PAHs (17 species), carbonyls (15 species),
Chloride
Carbon
Sulfur
NH3, CO, NOx, SO2
Nitrate
Total Carbon
Chlorine
Sulphate
Sodium
Potassium
Ammonium
Magnesium
Calcium
Soluble Sodium
Aluminum
Titanium
Potassium
Silicon
Vanadium
Chromium
Manganese
Iron
Cobalt
Nickel
Copper
Bromine
Yttrium
Tin
Soluble
A-136
Major Finding
Table A–6. (Cont’d) Objective
Monitoring Station
Measurement and Data Analysis Methods Barium
Lanthanum
Lead
Zinc
Antimony
Major Finding
SO2: Pulsed Florescence SO2 Analyzer CO: Gas Filter Correlation CO Ambient Analyzer CO2: Q-Trak monitor NOx: Chemiluminescence NO-NO2-NOx Analyzer Simple Mass Balanced (SMB) Model was used in determining emissions factors of vehicles inside tunnel
Emission strengths
for
Target pollutants: NOy, CO
Sample period: 18/10/94, 08/11/94 and 13/11/94
Emission rate for urban plume NOy: 5.4 x 1025 molecules/s
Estimated from an aircraft study of HK air
CO measurement: Thermo Electron analyzer (Model 48)
Emission rate for urban plume CO: 1.8 x 1026 molecules/s (2.64 x 105
quality
NOy
pollutants
as
The National Center for Atmospheric Research
conversion/detection technique that all higher oxides of
CO/NOy ratio for HK urban plume was 3.3 and lower than other cities which
estimated
from
(NCAR) Beechcraft King Air aircraft served as
nitrogen were first reduced to NO with CO.
may be due to the predominance of diesel vehicles having low CO and high
aircraft study of
research platform
Emission rates of NOy and CO were determined by
NO emissions.
HK air quality
Average aircraft ground speed of 300 km/h
plume integration techniques.
primary
measurement:
accomplished
(Lind and Kok, 1999) Chemical Species: AAs
Aromatic amines
AHs
Aliphatic hydrocarbons
BVOC
Biogenic Volatile Organic Compound
CO
Carbon monoxide
FA
Fatty acids
NMHC
Non-methane hydrocarbons
NOx
Nitrogen oxides
A-137
through
a
tonne/yr)
PAHs
Polycyclic aromatic hydrocarbons
PM
Particulate matter
RSP
Respirable suspended particulates
SO2
Sulphur dioxide
VOC
Volatile Organic Compound
Other Abbreviation: CPWM:
Circular pollution wind map
DNPH:
2,4-dinitrophenylhydrazine
IAQO:
Indoor Air Quality Objectives
PFBHA:
O-(2,3,4,5,6-pentafluorobenzyl)-hydroxylamine
PFPH:
pentafluorophenylhydrazine
PCA:
Principal component analysis
RPCA:
Rotated principal component analysis
SVD:
Singular value decomposition
A-138
Table A- 7. Summary of PM Source Apportionment Studies Using Models Study Objective
Location*,
Period,
and Observables+
Source
Apportionment Major Findings ^
Measurements
Methods
Source apportionment of
Five urban sites in western New
Na, As, Se, Cd, and V by Principal factor analysis
TSP at five sites by
Territories: San Hui (SH), TM,
AAS and Zn, Pb, Fe, Mg,
Elements acquired at the
principal factor analysis
Kelvin Tower (KT), Hung Shui
Ca, S, Mn, Sr, and Ba by
five sites
(Fung and Wong, 1995）
Kiu (HSK), and Au Tau (AT)
ICP-AES
Seven-day averaged TSP were collected from 1986 to 1987 to obtain 40 data sets
A-139
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^
Source apportionment of
Eleven sites of AQMN:CW, JB,
25 species: elements by PMF (Paatero and Tapper,
PM10 at 11 sites of
TP, SSP, ST, TST, HKS, KC,
AAS and ICP-MS (Cd, Ni, 1994)
AQMN by PMF (Lee et
KT, TW, MK
As, V, Ba, Mn, Pb, Zn, Cu,
19 species (Al, Ca, Mg,
al., 1999)
24-hr average PM10 samples
Mg, Al, Fe, Ca) and Ions by
Pb, Na+, V, Cl-, NH4+,
-
-
were collected every sixth day
IC (Br , Cl ,
from 1992 to 1994 to obtain a
NH4+)
SO4=,
+
+
K , Na ,
SO4=, Br-, Mn, Fe, Ni, Zn, Cd, K+, Ba, Cu, and
total of 1516 samples
As) were included after eliminating species with large amount of missing data or below detection limit Marker species: Secondary aerosol-NH4+, SO4= Cl-
depleted
marine
+
aerosols-Na , Mg, SO4= marine
aerosols-Cl-, Unit:
+
%
Na , Mg
The PMF model can identify the sources of
Crustal/soil dust-Al, Ca,
particulate pollutants in Hong Kong, based on
Mg, Mn, Fe
the chemical species concentrations data
Nonferrous
metal +
smelter-Pb, K , As Vehicular emission-Pb, Ni, Zn, Cd
A-140
µg/m3
obtained from the HKEPD monitoring network comprising 11 air monitoring stations
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^ Particulate Br-BrParticulate Cu-Cu Fuel oil burning-V, Ni Unexplained residual
Source apportionment of
A roadside site (MK) and nine
OC and EC by NIOSH 5040 PMF (Paatero and Tapper,
PM10 at 10 sites of
urban sites (CW, KC, KT, SSP,
(Sin
AQMN by PMF (Yuan et al., 2006a)
ST, TC, TP, TW, and YL) of
et
−
water-soluble ions (Br , Cl , =
+
1994)
2002), −
−
AQMN
al.,
+
+
Data from all the stations
NO3 , SO4 , NH4 , K , Na )
were combined together
were
by IC, Sixteen elements (Al,
to
collected every sixth day from
Ba, Be, Ca, Cd, Cr, Cu, EC,
dataset
07/98 to 12/02
Fe, Mg, Mn, Ni, OC, Pb, V
samples in total
Number of samples (CW, 249;
and Zn) by ICP-AES, and
18
KC, 149; KT, 217; SSP, 274;
As, Hg and Se by FIA-AA
analyzed
24-hr
PM10
samples
construct
a
with
large 2199
species
were
ST, 146; TC, 225; TP, 150; TW,
The residual EV: blow
274; and YL, 266)
0.25 Marker species: Vehicle
exhaust-EC, Unit:
OM Residual oil-Ni, V -
µg/m (%)
Fresh sea salt-Cl
Overall, secondary sulfate and local vehicle
Aged sea salt-Na+, Mg,
emissions gave the largest contribution to PM10
SO4
A-141
3
=
in HK (25% each), followed by secondary
Crustal soil-Al, Ca, Fe,
nitrate (12%)
Mn, Mg
Contributions from other source types were
Secondary
below 10%. Regional transport was responsible
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^ sulfate-NH4+, SO4= Secondary
for about 60% of the ambient PM10 level on
Nitrate-NO3-
Biomass/Waste/Region
annual average basis, and could get up to 70% in wintertime
+
al Vehicle-K , Pb, Cd Coal-fired Combustion/SOC-As, Cd, Pb, OM Source apportionment of
PU roadside site
OC
PM2.5 at PU roadside site
24-hr averaged PM2.5 samples
IMPROVE,
by PMF (Cheng, 2007)
and −
EC
by
TOR PMF (Paatero and Tapper, Unidentified, 5.49 , 10%
water-soluble 1994) −
−
=
-
were collected every seventh
ions (Br , Cl , NO3 , SO4 ,
27 species including Cl ,
days from 10/04 to 09/05 to
NH4+, K+, Na+) by IC, Forty
NO3-, SO42-, NH4+, K+,
obtain a total of 80 samples
elements (from Na to U) by
OC, EC, Na, Mg, Al, Si,
XRF
K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, As, Br, Rb, Sn,
Gasoline-fueled vehicle, 7.16 , 13% Tire dust, 1.79 , 3%
Secondary aerosol, 10.82 , 20%
Sb, Ba, Pb Unidentified, 3.35 , 22%
Diesel-fueled
-3
Residual oil combustion, 4.54 , 8%
Coal combustion, 7.44 , 13%
Paved soil dust, 3.81 , 7%
Vehicle , 1.59 , 11%
-3
(14.37 ug m )
Paved soil dust, 2.39 , 17%
vehicle-EC, OC
-3
(in ug m , %)
Gasoline-fueled
Industry, 2.81 , 20%
vehicle-OC, EC
Field burning+second ary aerosol, 1.80 , 13%
Secondary NO3-,
NH4+ Coal combustion-multi-elem
A-142
a PM2.5 (50.02 ug m )
b PMcoarse
Maker species:
aerosol-SO4=,
Diesel-fueled vehicle, 14.44 , 26%
Marine aerosol, 2.43 , 17%
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^ ents Residual
oil
combustion-Ni, V Paved road dust-Mg, Al, Si, Ca, and Ti Tire dust-Zn Source apportionment of
PU roadside site
OC
and
PM10-2.5 at PU roadside
24-hr averaged PM2.5 samples
IMPROVE, −
EC
by
TOR PMF (Paatero and Tapper, Unidentified, 5.49 , 10%
water-soluble 1994) −
NO3−,
=
-
site by PMF (Cheng,
were collected every seventh
ions (Br , Cl ,
SO4 ,
27 species including Cl ,
2007)
days from 10/04 to 09/05 to
NH4+, K+, Na+) by IC, Forty
NO3-, SO42-, NH4+, K+,
obtain a total of 80 samples
elements (from Na to U) by
OC, EC, Na, Mg, Al, Si,
XRF
K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, As, Br, Rb, Sn,
Gasoline-fueled vehicle, 7.16 , 13% Tire dust, 1.79 , 3%
Secondary aerosol, 10.82 , 20%
Sb, Ba, Pb Unidentified, 3.35 , 22%
emission-EC,
-3
Residual oil combustion, 4.54 , 8%
Coal combustion, 7.44 , 13%
Paved soil dust, 3.81 , 7%
Vehicle , 1.59 , 11%
-3
(14.37 ug m )
Paved soil dust, 2.39 , 17%
OC
-3
(in ug m , %)
Secondary aerosol and field buring-SO4=, NO3-, +
NH4 , K Paved road dust-Mg, Al, Si, Ca, and Ti Industry-multi-elements Residual combustion-Ni, V
A-143
a PM2.5 (50.02 ug m )
b PMcoarse
Maker species: Vehicle
Diesel-fueled vehicle, 14.44 , 26%
oil
Industry, 2.81 , 20% Field burning+second ary aerosol, 1.80 , 13%
Marine aerosol, 2.43 , 17%
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^ marine aerosol-Cl-, Na and
(Watson, 1979;
MK roadside site, TW urban
PAHs
organic compounds in
site, HT rural site
GC/MS; AHs and FAs by Watson, et al., 2004)
PM2.5 by CMB (Zheng et
24-hr (0000-2400 LST) PM2.5
GC/FID; SEOC is defined
Source profiles of Diesel
al., 2006)
samplers were collected every
as the sum of identified
engine exhaust, Gasoline
sixth day using R&P 2025 at
AHs,
engine
Exhaust,
each site from 11/00 to 10/01
ROHs; OC and EC by
Vegetative
detritus,
Number of Samples: MK (58),
NOISH_TOT; WSOC by
Biomass
burning,
TW (57); HT (58)
IC
Cigarette, Smoke, Road
PAHs,
ROHs
by CMB
Source apportionment of
FAs,
and
dust,
Meat
cooking,
Natural gas combustion Diesel engine exhaust dominated fine organic were obtained from the carbon (57 ± 13% at urban sites and 25 ± 2% at previous studies in North the rural site), followed by meat cooking and America (Hildemann et biomass burning (14%) al., 1991; Rogge et al., The primary sources identified by this technique 1993a; 1993b; Schauer, explained 49%, 79%, and 94% of the measured 1998; Schauer et al., fine organic carbon mass concentration at the 1999a, 1999b, 2002b; rural, the urban, and the roadside sites, Zheng
et
al.,
2002; respectively. The unexplained fine OC is likely
McDonald, et al., 2003). due to secondary organic aerosol formation The coal source profile was obtained from China (Zheng, et al., 2005)
A-144
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements Source apportionment of PM2.5 at PU roadside site by CMB (Cheng, 2007)
PU roadside site 24-hr averaged PM2.5 samples were collected every seventh
Apportionment Major Findings
Methods^ OC
and
EC
TOR CMB
−
−
=
ions (Br , Cl , NO3 , SO4 , +
+
(Watson,
1979;
water-soluble Watson, et al., 2004)
IMPROVE, −
by
+
Source profile: Vehicle
days from 10/04 to 09/05 to
NH4 , K , Na ) by IC, Forty
emission
obtain a total of 80 samples
elements (from Na to U) by
2004)), Paved soil dust
XRF
(Ho et al., 2003b), Tire
(HKEPD,
wear (Hildemann, et al, 1991),
Brake
lining
(Ondov, et al., 1982), Cooking
fumes
(HKEPD,
2006),
Case I Marine sector
Source Vehicle emission Paved road dust Brake lining Tire wear Ammonium sulfate Ammonium bisulfate Ammonium nitrate Coal combustion Field burning Residual oil combustion
μg m-3 28.9 ± 7.7 0.0 ± 0.0 0.3 ± 0.4 0.0 ± 0.0 0.2 ± 0.8 4.3 ± 1.9 0.9 ± 0.4 3.9 ± 1.9 0.8 ± 0.7
Case II Continent sector % of mass 72 0 1 0 1 11 2 10 2
μg m-3 42.9 ± 13.2 0.7 ± 2.5 0.2 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 11.0 ± 2.2 4.5 ± 0.8 10.9 ± 2.7 7.2 ± 1.6
0.0 ± 0.1
0
Marine aerosol
0.4 ± 0.4
1
Calculated mass
39.9 ± 6.2
0.1 ± 0.1
Local sourcesa
29.3
73
43.9
Regional sourcesb
10.2
26
33.6
Unidentified
-3.0
-8
-4.4
R Chi square Percent mass
0.9 1.2 108.1
0.8 ± 1.0 78.3 ± 10.0
1.0 1.5 104.2
a
Local sources include vehicle exhaust, paved road dust, brake lining, tire dust, and residual oil comb
b
Regional sources include secondary aerosol, field burning, and coal combustion.
Biomass burning (Core The concentrations of both local and regional et al., 1989), Residual oil sources increase when air masses are from the combustion,
marine continent compared with when air masses are
aerosols (Watson, 1979), from marine Coal combustion from The contributions of regional sources to the plants (Klein, et al., PM2.5 at PU roadside site are less than 30% when 1975), aerosol
Secondary air masses are from marine direction, while they increase to nearly 50% when air masses are from continent direction
Source apportionment of PM2.5 at TW, MK, and HT by PCA and APCS (Guo, et al., 2008)
MK roadside site –
2m from Lai Chi Kok Road and 2m
above ground level TW urban site –
on the rooftop of a government
A-145
Table A–7. (Cont’d) Study Objective
Location*,
Period,
and Observables+
Source
Measurements
Apportionment Major Findings
Methods^
building at ~15-18m above ground level HT rural site –
on a hillside platform
*
Location
AQMN……………Hong Kong Air Quality Monitoring Station, including 14 air monitoring stations (11 urban sites: TM, YL, TP, ST, TW, KC, SS, KT, Eastern, TC, CW, and 3 roadside sites: MK, CB, and CC) CB………………..Causeway Bay is located on 1 Yee Woo Street, Causeway Bay. Site type: busy commercial/residential area surrounded by many tall buildings site CC………………..Central is located in junction of Des Voeux Road Central and Chater Road, Central. Site type: Busy commercial/financial area surrounded by many tall buildings site CW ………………Central Western is located on 1 High street, Si Ying Pun. Site type: mixed residential/commercial site Eastern……………Eastern is located on 20 Wai Hang Street, Sai Wan Ho. Site type: residential site HKS………………Hong Kong South HMT ….................Ho Man Tin, an urban site with moderate traffic HT………………..Hok Tsui is located on a hillside platform facing the South China Sea at the southeast end of Hong Kong Island, represents a rural background/regional/transport environment KC………………..Kwai Chung, located at 999 Kwai Chung Road, Kwai Chung. Site type: mixed residential/commercial/industrial site KT………………..Kwun Tong is located on 6 Tung Yan street, Kwun Tong. Site type: mixed residential/commercial/industrial site MK……………….Mong Kok is located in junction of Nathan Road and Lai Chi Kok Road. Site type: busy by many tall buildings commercial/residential area surrounded site PU………………..Polytechnic University, close to roads with heavy diesel bus, diesel taxi, and passenger car traffic; represents a roadside microenvironment in a downtown area SS………………. Sham Shui Po is located on 37A Yen Chow Street, Sham Shui Po. Site type: mixed residential/commercial site ST………………..Sha Tin is located on 11-17 Man Lai Road, Tai Wai, Sha Tin. Site type: residential site TC………………..Tung Chung is located on 6 Fu Tung Street, Tung Chung. Site type: residential site TM……………….Tap Mun. Site type: rural site. TP…………………Tai Po is located on 1 Ting Kok Road, Tai Po. Site type: residential site TST ………………Tsim Sha Tsui East, an urban site with moderate traffic TW……………….Tsuan Wan is located on 60 Tai Ho Road, Tsuen Wan. Site type: mixed residential/commercial/industrial site UST………………Hong Kong University of Science and Technology at Clear Water Bay, which is a sparsely populated area on the eastern coast of the New Territories
A-146
YL………………..Yuen Long is located on 269 Castle Peak Road, Yuen Long. Site type: residential site
+
Observable
FIA-AA………..Flow Injection Analysis Atomic Absorption GC-MS ................. Gas chromatograph/mass spectrometry for organic compounds IC ......................... Ion chromatography for anions
ICP-AES……….Inductively Coupled Plasma-Atomic Emission Spectroscopy IMP_TOR………...Thermal/optical reflectance method following the IMPROVE (Interagency Monitoring of Protected Visual Environments). It produces OC1, OC2, OC3, and OC4 at 120°C, 250°C, 450°C, and 550°C, in a 100% helium (He) atmosphere, and EC1, EC2, and EC 3 at 550°C, 700°C, 800°C in a 2% oxygen (O2) / 98% He atmosphere. Filter reflectance is monitored throughout the analysis; Pyrolized OC (POC) is defined as the carbon evolving between the introduction of oxygen and the return of reflectance to its initial value (the OC/EC split). IMPROVE OC is defined as OC1+OC2+OC3+OC4+POC; EC is defined as EC1+EC2+EC3−POC; (see Chow 1993 for more descriptions) NOISH_TOT……..NIOSH Method 5040 (Birch, 1998) (NIOSH, 1999), For OC, the oven temperature was raised stepwise from ambient temperature to 350 °C (70 s), 550 °C (70 s) and to 850 °C (110s) in a 100% helium atmosphere, and The oven was then cooled to 550 °C and raised gradually from 550 to 850 °C, with ramps at 600, 700, 750, 800 and 850C for periods of 50, 40, 30, 30, 70 s, respectively in a 5% v/v oxygen in helium to get EC. Correction for pyrolysis was made by continuously monitoring the filter transmittance of a laser beam (678 nm) throughout the analysis. STN_TOT………..Thermal/optical transmission analysis follows the STN (Speciation Trends Network, Peterson and Richards, 2002)) protocol. It produces OC and EC in 2% oxygen (O2) in helium (He). Filter transmittance is monitored to split OC and EC in STN_TOT; The STN protocol has a short and fixed residence time per temperature plateau and cannot report distinguishable carbon fractions
Chemical Species: AHs……………….Aliphatic hydrocarbons (C16-C38) Ca++ ........................Water-soluble calcium Cl- ..........................Chloride EC .............. ……..Elemental carbon FAs………………..Fatty acids Na+ .........................Water-soluble sodium NH4+ ........... ……...Ammonium NO3-........................Nitrate OC..........................Organic carbon OM............. ………Organic mass (= OC x a factor) PAHs……………...Polycyclic Aromatic Hydrocarbons PM………………..Particulate matters
A-147
ROHs……………..Alkanols SEOC……..............Solvent Extractable Organic Compounds SO4=............ ……....Sulfate SOC……………….Secondary organic aerosols TC .............. ………Total carbon THC ………………The total weight of extractable compounds
^
Source Apportionment Method
CMDL-NOAA….Climate Monitoring and Diagnostics Laboratory-National Oceanic and Atmospheric Administration EV_CMB………..Explained Variations_Chemical Mass Balance (Watson, 1990; 2004) EV……………….Explained Variations PCA………………Principal Component Analysis APCS……………..Absolute Principal Component Scores PMF_CMB……………..Positive Matrix Factorization Analysis (Paatero P., 1998; 2000); It is a variant of factor analysis with the contribution of each factor constrained to be non-negative (Paatero and Tapper, 1994). It solves the factor analysis problem by a weighted least square fit of the sum of individual scaled residuals. The use of PMF does not require a priori information about the component concentration profiles or their thermal properties.
A-148
Table A- 8. Summary of Meteorological Studies. Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings and
Meteorology: MM5 with
Lagragian
modeling analysis of
the PATH land use data
model:
a severe air pollution
for the two inner domains
simulation:
episode in western
(res. 1.5km and 4.5km,
Hong Kong (Fung et
covering HK and the PRD
PM10
al., 2005)
respectively) and USGS
western HK
Observational
dispersion
In the daytime, easterly to northeasterly
In the daytime, pollutants emitted in the
for
winds prevailed over eastern HK, while
PRD to the NE of HK could be transported
28-30
northwesterly wind was blowing over the
to the west part of HK. However, these
December 1999 of a
west of HK from the PRD. This
pollutants were prevented from traversing
northwesterly wind was part of the
to the eastern part of HK. The land-sea
Case
episode
over
land-sea breeze circulation encompassing
breeze could draw air northwesterly from
data for the two outer
western HK. The background weak
the PRD into HK, trapping the pollutants
domains (res. 13.5km and
easterly wind and the land-sea breeze
within its circulation or convergence zone.
40.5km,
covering
circulation over the Pearl River Estuary led
At
Guangdong province and
to convergence ozone in the western part of
smoothly across HK above the planetary
southern
HK over the north Lantau.
boundary layer height to its southwest
respectively). FDDA was
At night, no convergence zone was
under
employed.
discovered, boundary layer stabilized and
northeasterly wind affecting entire HK,
shrank to about 500m, and stronger
and no trapping of pollutants occurred.
northerly to northeasterly winds dominated
Composite visible and infrared satellite
much of HK.
images were examined to identify red spot
China
night,
the
pollutants
prevailing
were
advected
northerly
to
of wild fires in Guangdong province and a layer of smoke and haze along the southern coast of Guangdong near HK. The red spots are hot thermal anomalies derived from the infrared channels of the advanced very high resolution radiomater onboard NOAA polar orbiting satellites.
A-149
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings PATH
and process analysis
which includes,
chemistry:
of
Meteorology: MM5, with
Eulerian
in
the PATH land use data
Hong Kong (Huang et al., 2005)
ozone
typhoon-related episodes
model
system,
Dispersion
Numerical simulation
and
by
On 13 September 1999, before 1100 LST,
northerly/northwesterly wind over HK
ozone levels were rather uniform and low
under the influence of typhoon York.
over the PRD. About 1300 LST, one plume
CB-IV mechanism
Sea-breeze circulation developed along the
with ozone level about 83ppb appeared
for the two inner domains
Case
simulation:
coastline of HK in the afternoon, with one
downwind of Guangzhou.
(res. 1.5km and 4.5km,
11-13 September 1999
northwesterly established over Lantau
At 1400 LST a large range of evaluated
covering HK and the PRD
of
Island, one southerly over West Lamma
ozone levels about 90ppb appeared over
respectively) and USGS
ozone episode in HK
Channel, and the other over eastern part of
the
data for the two outer
Land use: compiled by
HK Island. Several convergence zones also
transported southeastward, reaching the
domains (res. 13.5km and
HK
developed
western part of the New Territories at
40.5km,
Department
covering
SAQM model
for a
with
cyclone-related
Planning with
The
flow
was
by
the
background winds.
dominated
sea
breeze
and
Pearl
River
Estuary
and
were
about 1500 LST. Another ozone-rich
Guangdong province and
resolution of 30 m over
plume traveled northeasterly to affect HK
southern
China
the PRD to supersede
at the same time.
respectively). FDDA was
the default land use
Process analysis: For domain 4 covering
employed.
information from the
HK, 30% of the total ozone production in
US Geological Survey.
the lower atmospheric layer (40-350m) was produced by photochemical reaction while 70% was transported by horizontal advection. The enhancement of ozone concentrations in HK was contributed mainly by the ozone-rich airmass traveling through the north boundary. Similar conclusion could be discovered in another ozone episode on 1 and 2 November 2003 which was associated with the passage of
A-150
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings severe tropical storm Melor. Emission: EMS-95. In HK domain, data was collected from HKEPD. In the PRD and southern China, data was extracted from the Global Emissions Inventory Activity (GEIA) database and Emission Database
for
Global
Atmospheric
Research (EDGAR) with additional data of major power stations in the southern China. US
With the rapid urbanization process, the
Urban land-use dataset and a proper urban
Survey
temperature in the lowest 2km atmosphere
land-use parameterization are critical for
cover
over urban areas increased by about 1.5
the mesoscale model to capture the major
use
characteristics database
degree. A stronger sea breeze circulation
features of the observed UHI effect and
treatment, a Lagrangian
use:
compiled in 1993 with 6
was developed in association with the
land–sea-breeze circulations in the PRD.
particle tranjectory model
major
classidications:
stronger urban heat island. The sea breeze
Stronger UHI in the PRD increases the
(MPTM)
urban,
Meteorology:
a mesoscale model
coupled with the NOAH
Geological
and the effects on the
land surface model with a
global
local circulation in
bulk
the Pearl River Delta region (Lo et al., 2007)
urban
MM5
Land
Urban modification in
land
land
cropland,
circulation front could penetrate further
differential temperature gradient between
grassland, forest, water
inland to overcome the prevailing easterly
urbanized areas and nearby ocean surface
body, and others – 0.5%
flow in the western HK.
and hence enhances the mesoscale SBC.
of the PRD is urban and
The SBC front consequently penetrates
45.6% is cropland
farther inland to overcome the prevailing
HK
Planning
easterly flow in the western part of inland
Department with spatial
Hong Kong. Additional sensitivity studies
resolution of 30m –
indicate
13% of the PRD is
development
A-151
that and
further
industrial
urbanization
will
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings urban and 22.7% is
strengthen the daytime SBC as well as
cropland
increase the air temperature in the lowest 2 km of the atmosphere.
A-152
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings
Simulation and observation of surface wind field valid at 1400 LST 31 Oct 2003 given by (a) URBAN-LSM, (b) SLABLANDU, and (c) SLAB experiments. Investigation
of
Meteorology:
MM5
Land use: compiled by
If there was no background wind, the sea
Conceptual model on pollutant dispersion:
Planning
breeze circulation over the Pearl River
In the afternoon of sunny days with weak
cross-city
coupled with the NOAH
HK
transport and trapping
land surface model with a
Department
with
Estuary strengthened in early afternoon
background wind, pollutants tend to rise
of air pollutants by
bulk
use
resolution of 30 m over
and extended up to 20km inland, pushing
over the urban areas, thereby drawing in air
coastal
urban
treatment, a Lagrangian
the PRD to supersede
over western half of HK. The circulation
from the surrounding oceanic areas,
breeze
particle tranjectory model
the default land use
could not extend to cover eastern half of
resulting in a convergent inflow at the
(MPTM)
information from the
HK even there was no background wind.
lower levels into the urban areas during the
US Geological Survey.
At night, the sea breeze circulation
day. The air will rise up to 1 to 2 km,
disappeared but the urban heat island
spread out horizontally, and eventually
convergence still remains near Guangzhou
sink back in the nearby coastal areas,
and Dongguan.
forming a closed vertical circulation.
enhanced
land-sea
and
circulations (Lo et al., 2006)
urban
land
A-153
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings The Pearl River Estuary linked the pollutants trapped in the land sea breeze circulation at all the coastal cities in the PRD well mixed. The mixed pollutants could be brought to the western HK through
the
return
branch
of
the
circulation. With the same weak wind frequency, the probability of pollutant trapping in the current PRD is much higher than two decades ago due to the rapid urbanization process. An
extremely
low
N/A
N/A
N/A
The
satellite
(MODIS)
data remote
visibility event over
sensing technique provides a global view
the
of aerosol distributions spatially with the
Guangzhou
region: A case study
resolution down to 1km.
(Wu et al., 2005)
Pollutant transport could be revealed by multiple satellite images. It was applied to examine the aerosol optical depth under a severe PM episode. The Aerosol optical depth in Guangzhou was about 1.2 on 2 November 2003, about twice of the seasonal average in the region. The
A-154
pollutant-rich
plumes
were
Table A – 8. (Cont’d) Objective
Model application and
Dispersion/Land use
Flow characteristics
Major Findings
settings transported from the central PRD to HK on 2 November 2003 by the northwesterly wind induced by the peripheral circulation of typhoon Melor. On the other hand, the normalized aerosol backscattering data from Micro-pulse Lidar could provide aerosol distribution at vertical direction. Boundary layer variation could be derived from Lidar observation. On 2 November 2003, the boundary layer height in HK was monitored only around 500m, which was caused by the subsidence airflow in the periphery of typhoon Melor. *
Locations
PRD………………. Pearl River Delta SBC………………. Sea-Breeze Circulation UHI………………. Urban Heat Island Model MM5……………... Mesoscale Model MPTM…………….Lagrangian Particle Trajectory Model
A-155
Feasibility of Establishing Air Monitoring Supersites in Hong Kong
APPENDIX B:
U.S. AND Taiwan Supersite Measurements
Prepared by S.C. Leea, J.G. Watsonb, J.C. Chowb, K.F. Hoa, T. Wanga, A. Lauc, S. Liud
a
Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong b Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, USA c Atmospheric Research Center, The Hong Kong University of Science and Technology d Academia Sinica, Taipei, Taiwan
Prepared for： Hong Kong Environmental Protection Department (HKEPD) 33/F., Revenue Tower, 5 Gloucester Road, Wan Chai, Hong Kong
September 10, 2008
APPENDIX B.
U.S. AND TAIWAN SUPERSITE MEASUREMENTS
This appendix reviews the supersites objectives and measurement for the 7 U.S. and 2 Taiwan supersites. Table B-1 intercompare the instrument access the 9 supersites. Table B- 1. Summary of U.S. and Taiwan Supersites Measurements.
Baltimore
Observable
Method/Instrument
Avg Time
Freq
Mass Conentration Aerosol Mass PM2.5 mass
Harvard Impactor
24-hr variable
Daily or 1-in-6 day variable
24-hr daily 24-hr
3rd day daily 3rd, 6th day
PM2.5 mass
Pittsburgh
PM2.5 mass PM2.5 mass PM2.5 mass PM2.5 mass
URG Filter Pack Sampler
24-hr
6th day
PM2.5 mass PM2.5 mass
Honeycomb Denuders, HDS HEADS (Harvard University)
24-hr 24-hr
6th day daily
PM2.5 mass
Dichotomous sampler
24-hr
6th day
PM2.5 mass
CMU sampler
24-hr
daily
x
PM2.5 mass
High-Volume sampler
24-hr
daily, 6th day
x
PM10 mass PM10 mass
Harvard Impactor Particle Concentrator (R&P Model 2400) R&P Partisol Sampler Honeycomb Denuders, HDS Dichotomous sampler
24-hr variable
daily variable
daily 24-hr 24-hr
daily 6th day 6th day
ChemTox
New York
St Louie
Houston
Los Angel
South Taiwan
North Taiwan
x
Particle Concentrator (R&P Model 2400) R&P Speciation 2300 sampler R&P Partisol Sampler Andersen RAAS FRM sampler
PM10 mass PM10 mass PM10 mass
ChemTox
Fresno
x x x x
x
x x x x x x
x
x x x x x
B-1
x
Table B–1. (Cont’d) Observable
Method/Instrument
Avg Time
Freq
PM10 mass
Organic Denuder Sampler, URG Inc
24-hr
6th day
PM10 mass
High-Volume sampler
24-hr
6th day
Baltimore
Fresno
Pittsburgh
New York
St Louie
Houston
Los Angel
South Taiwan
North Taiwan
x
x
x
x
x
x
x
x
x
x
Mass Conentration x x
rd
x
x
PM10 mass PM1 mass
STN sampler (Metone SASS) BGI PM1 sampler
24-hr 24-hr
3 day, daily daily
PMx mass PMx mass
MOUDI, MSP LPI
8-hr, 24-hr 24-hr
3rd day, daily daily
PMx mass PMx mass
High-Vol. Particle Size Classifier DRUM impactor
24-hr 3-hr
2nd, 6th day multidays
Hi-vol sampler
24-hr
12th day
x
R&P TEOM 1400
30
daily
x
x
x
30
daily
x
x
x
TSP Continuous Mass PM10 mass
x x x
x x x
x
x x
Particle
PM2.5 mass
R&P TEOM 1400
PM2.5 mass PM2.5 mass
R&P TEOM FDMS and T/RH modified MetOne Beta Attenuation Monitor (BAM)
10 min, min, 1-hr 10-min, min, 1-hr 30-min 1-hr
PM2.5 mass Ions Sulfate
Andersen CAMMS
1-hr
daily
R&P 8400S, Flash volatilization of sulfate
daily
x
Sulfate
Prototype Harvard
10 min, 30 min, 1-hr 10 min, 30 min
daily
x
Sulfate Sulfate Nitrate
Aerosol Dynamics Inc. (prototype) HSPH continuous monitor R&P 8400N Flash volatilization of nitrate
10 min, 30 min 1-hr 10 min, 30 min, 1-hr
daily daily daily
Nitrate
Prototype Harvard
10 min
daily
10 min
daily
Nitrate
x x
Aerosol Dynamic Inc. (prototype)
x
daily daily
x
x x x x
x
x
x
x
x
x x
x
x
x
x
x
x x x
x
x x
x x
B-2
x
x
x
Table B–1. (Cont’d) Baltimore
Fresno
Pittsburgh
New York
St Louie
Houston
Los Angel
x
x
x
x
Observable
Method/Instrument
Avg Time
Freq
Nitrate
HSPH continuous monitor
10 min, 1-hr
daily
Water soluble ions OCEC OC/EC
Khylstov steam sampler with IC analysis
1-2-hr
Continuous
R&P 5400 ambient carbon particulate monitor Aerosol Dynamic Inc. Sunset field analyzer
30 min
daily
x
x
x
30 min 1, 2, 4-hr
daily daily
x
x
x x
Classical Scattering Aerosol Spectrometer (PMS model CSAS-100-HV)
5, 10 and 15-min, 1 and 24-hr 5 and 15-min, 1 and 24-hr 5 and 15-min, 1 and 24-hr 5 and 15-min, 1 and 24-hr
daily
x
daily
x
x
x
x
x
x
x
x
x
x
x
x
x
x
x
5 and 15-min, 1 and 24-hr 10 min 5-min, 10 min 5 min 5 min 10 min 5-min
daily
5 and 15-min, 1 and 24-hr 5s 5 min 30 min
daily
5 and 30-min, 1 and 24-hr 2-hr
daily
OC/EC OC/EC Aerosol Size / Particle number Size distribution
Size distribution
TSI APS
Size distribution
TSI/Grimm SMPS
Size distribution
TSI ultrafine SMPS
Size distribution
DMA
Size distribution Size distribution Size distribution Size distribution Size distribution Particle number
EPLI Grimm Technologies OPC 1.108 CLIMET OPC (CI-500) PMS LasAir 1002/1003 PMS PCASP-X TSI Ultrafine Condensation Particle Counter
Particle number
TSI Condensation Particle Counter
Particle number Size distribution Surface area Light absorption Light absorption
TSI Aerosol Electrometer TSI Electrical Aerosol Detector Epiphaniometer
South Taiwan
North Taiwan
x x x
x
x
x
x
x
x
(TSI 3926)
daily
x
daily
(TSI 3936L10S)
x
daily daily daily daily daily daily
x x
x x x
x x x x
x
x
x
x
x
x
x
(TSI 3025A)
Light absorption Light scattering Light Scattering
Magee Scientific Aethalometer Coefficient of haze Nephelometer /TSI 3563
5-s, 1 or 5 min
x
daily daily daily
x
x x x x
x
x
x
x
x
(AE31)
daily
x daily
x
B-3
x
x
Table B–1. (Cont’d) Observable
Method/Instrument
Avg Time
Freq
Light scattering
OPTEC NGN3 / NGN2
5 or 10 min
daily
Light scattering Light scattering
Greentek or DUSTRAK photometer Radiance M903 heated nephelometer
5-min 5-min
daily daily
Light sacttering
Ecotech M9003 (Ecotech, USA),
1-min
daily
Others PM observables Metals Cloud condensation
SEAS with GFAA analysis DH Associates CCN
30 min 4-hr
-
Particle-bound PAH
EcoChem PAS 2000
5 min
daily
Individual particle
Time-of flight spectrometer
5 min
Baltimore
x
Fresno
Pittsburgh
x
x
New York
St Louie
Houston
Los Angel
x
x
x
x
x
x
x
x
x
(UC Riverside)
Individual particle
RSMS-III
10 min
daily
Particle morphology
SEM
24-hr
daily
Size composition
AMS, Aerodyne
resolved
x
x
x
PM2.5 biological material
Epi-fluorescent microscopy with assays
24-hr
daily
x
Toxic (metals, chromium VI, aldehydes) Ozone Ozone
Xontec920
24-hr
6th day
x
Ecotech 9810
1-min
daily
Ozone Ozone
Ozone Monitor Dasibi Model
daily daily
Ozone Ozone NO/NOX/NO2 Nitrogen Oxide (NO)
TECO API 400 UV Absorption
5 min 5, 15 min, 1, 4, 12 and 24 hr 5 min 5, 10 min
NO/NOx
TECO Model 42s
Monitor Labs Model 8840
15 min, 1, 4, 12, 24 hr 5 min
x
x x
x
x x
x x
daily daily
x
x x
x
daily
x
daily
x
B-4
North Taiwan
x
x x
x
South Taiwan
Table B–1. (Cont’d) Observable
Method/Instrument
Avg Time
Freq
Baltimore
NO/NOx
NO/NOx monitor
5 min
daily
x
Nitrogen Dioxide (NO2)
Photolytic titration modified TECO 42s
daily
daily
NO/NOx NO/NOx NO NOy NOy NOy NOy NOy Acidic gas and NH3 HNO3 HNO3 HNO3 Inorganic gases (HNO3 and NH3) NH3
TEI 42 Chemiluminescence API 200A TILDAS
5 min 10 min 5 min
daily daily daily
Moly. Modified TECO Model 42s Thermo 42 CY TEI 42C NOy monitor
5 min 1 min 5 min 5 min
daily daily daily daily
TEI 42 Chemiluminescence HNO3 scrubbed derivatization Denuder HNO3 CMU Sampler/ IC analysis
5 min 5 min 24-hr 4-6 hr
daily daily daily daily
ESA NH3 analyzer
1, 5 min
daily
x
NH3
Thermo 17C
1, 5 min
daily
x
NH3
5-min
daily
x
NH3 NH3 CO CO
API 200D high sensitivity with NOx scrubbers and oxidizers TEI 17C Denuder HNO3
5-min 24-hr
daily daily
x
Horiba APMA 360 (Horiba, Japan)
1, 3 min
daily
CO CO
Dasibi 3008 API 300
5 min 10 min
daily daily
CO CO CO
CO Monitor TECO Model 48s Thermo 48C
SO2 SO2 SO2
TECO Model 43b SO2 Monitor
daily 15 min, 1, 4, 12, 24 hr 5 min
Fresno
Pittsburgh
New York
St Louie
Houston
Los Angel
South Taiwan
x x x x
x x x x
x x x x x
x
x
x
x x
x x
daily daily daily
x
x x
x x
daily daily
x x
B-5
x x
North Taiwan
Table B–1. (Cont’d) Observable
Method/Instrument
Avg Time
Freq
SO2
Dasibi 1008 RS
5 min
daily
SO2 Other gases NMHC All hydrocarbons VOCs (C2-C12) VOCs (C2-C10) C10-C20 Carbonyls
API 100A
10 min
daily
TEI 55C UCB GC-FID/MSD Xontec canister sampler/ canister PE Auto GC TENAX sampler Xontec cartridge sampler/ DNPH sampling techniques AeroLaser 14001A UC Riverside Luminol Scrubbed derivatization HPLC Open path TILDAS LIF CSU monitor
5 min 1-hr 5-8 hrs, 24 hrs 40 min 5-8 hr 24-hr
daily
5 min 5 min 5 min 5 min 10 min 1-hr
daily daily daily daily daily
H2CO NO2/PAN HONO NO/CO or H2CO/HONO OH/HO2 Total soluble peroxides, hydrogen and organic peroxides Meteorology Relative Humidity Solar radiation (Total/UV) Temperature
Baltimore
Fresno
Pittsburgh
New York
St Louie
Houston
Los Angel
South Taiwan
North Taiwan
x x x x
x x
daily
x x
x
x x
x x x
x
x
x x x x x
daily daily
x x
x x
x x
x
x x
x x
x
x
x
x
x
x
x
x
x
x
x
x
Wind speed/direction
x
x
x
x
x
x
x
x
x
Sonic anemometer
x
Pressure,
x
Fog and composition Precipitation
x x
x
x
x
cloud x
B-6
x
x
The following sections briefly summarize the overall goal for the supersites and their objectives, QA/QC guidelines, project role and knowledge gaps for the supersite project.
U.S. Superistes In 1990s, USEPA has required more extensive monitoring and studies to support development of Sate Implementation Plans (SIPs) and the setting of National Ambient Air Quality Standards (NAAQS). A set of special studies is extending beyond the national regulatory networks for particulate matter to elucidate source-receptor relationships and atmospheric processes. Supersites should combine a mix of intensive or advanced measurements at a central location combined with a set of satellite sites in the same region. The network stations operated at community representative, roadside, and background locations are benefit to study the PM problems from urban to regional scale. The Supersites represent an important component to foster greater integration across several science research programs both in local and regional scale. In the United States, supersites are being established in seven urban areas: 1) Fresno, CA; 2) Los Angeles, CA; 3) Houston, TX; 4) St. Louis, MO; 5) Pittsburgh, PA; 6) Baltimore, MD; and 7) New York, NY. The goals of U.S. Supersite goals are to: Serve as a research platform for testing new instrument Better understand the atmospheric process Establish relationship between PM and health effects
Taiwan Supersites (Lin et al. 2008) Air pollution in Taiwan is characterized by Taiwan Air Quality Monitoring Network (TAQMN) since 1993 (Chang and Lee, 2006). TAQMN consists of more than 70 air stations throughout Taiwan (Taiwan EPA, 2007). Each air station measures hourly concentrations of criteria pollutants (i.e., SO2, CO, NO2, O3, PM10) and meteorological variables (i.e., wind speed, wind direction, temperature, relative humidity). Daily pollutant standards index (PSI) at each air station is calculated by Taiwan EPA based on the daily concentrations of the criteria pollutants at each station. Currently, there are two supersites in Taiwan as shown in Figure 4 and 5, the Northern Particulate Matter Supersite (Northern Supersite) and Southern Particulate Matter Supersite (Southern Supersite) with the same objectives. The Northern Supersite started monitoring in March, 2002 (Lee, 2002; Lee et al., 2006a; Chang et al., 2007), while the Southern Supersite at Kaohsiung City began in April, 2005 (Wu et al., 2005b; Lin et al., 2006a). The Southern Supersite consists of one core- and three satellite-stations. The core station, located at Fooyin University, is approximately at the center of the industrial activities in the Kao-Ping air basin. In Taiwan, supersites are being established in Northern and Southern Supersite. Four supersite stations have been operated in Southern Supersite successfully since 2005: B-7
Fooyin University (22.603° N, 120.388° E) (FY); 2) Chiautou (22.759° N, 120.298°E) (TO); 3) Chenjen (22.606° N, 120.301° E) (TZ); 4) Choujau (22.513° N, 120.529° E) (CZ); 5) Sinjhuang (25.01° N, 121.25° E) (SC) Establishment of each supersite hypotheses for air quality was specifically formulated related to its objectives. Measurement methods were explored and developed to provide the variables needed to test these hypotheses. The supersite network should cover a wide range scale of areas from ambient to regional. Supersite objectives for each station network were coordinated with human health and toxicological studies to assist in determining air pollution sources and to better understand the relationships between regional air quality and related disease symptom over the course of a day. Evaluating these relationships required frequent sampling over short durations for a multi-year monitoring period to satisfy monitoring objectives for US and Taiwan Supersites defined in Table B-2. Table B- 2. Supersite Objectives in U.S. and Taiwan Location Baltimore
Pittsburgh
Objectives To provide an extended, ultra high-quality multivariate data set, with unprecedented temporal resolution, designed to take maximum advantage of advanced new factor analysis and state-of-the-art multivariate statistical techniques; To provide important information on the potential for health effects of particles from specific sources and generic types of sources, To provide large quantities of well characterized urban PM for retrospective chemical, physical, biologic analyses and toxicological testing, To provide sorely needed data on the sources and nature of organic aerosol presently unavailable for the region To provide support to existing exposure and epidemiologic studies to achieve enhanced evaluation of health outcome-pollutant and -source relationships, Characterization of the PM in the Pittsburgh region. These characteristics include the PM size, surface, and volume distribution; chemical composition as a function of size and on a single particle basis; temporal and spatial variability Development and evaluation of current and next generation atmospheric aerosol monitoring techniques (single particle measurements, continuous measurements, ultrafine aerosol measurements, improved organic component characterization, and others). Quantification of the impact of the various sources (transportation, power plants, natural, etc.) to the PM concentrations in the area.
B-8
Table B–2. (Cont’d) Location Houston
Objectives To collect physicochemical data on fine PM that can be used to characterize spatial and temporal variability in fine PM source contributions and composition, in Southeastern Texas To characterize spatial and temporal variability in fine PM source contributions and composition, throughout the southeastern United States, and To examine the physical and chemical process that govern PM formation and transformation in Southeastern Texas To develop a combined database on PM, gas phase air pollutants and meteorological variables, suitable for testing models of the formation and fate of fine PM; this objective will be achieved by coordinating with a large, integrated ozone and PM field study planned by the Southern Oxidants Study To examine exposures to fine PM from specific source categories in Southeastern Texas; this objective will be achieved by coordinating with an exposure study currently underway in Houston, funded by the Mickey Leland National Urban Air Toxics Research Center, and To relate the physicochemical data on fine particulate matter to mammalian cell responses; this objective will be achieved by coordinating with an EPA funded project currently underway at the University of Texas Houston Health Science Center. Los Angeles To make use of ambient particle concentrators for exposure (South California assessment and toxicological research Particulate To emphasize chemical and physical characterization of PM, Centre) especially size distribution, in the study of the relationship between PM and associated morbidity and mortality To emphasize mechanistic toxicological research including in vitro, animal and human approaches for better understanding of the underlying bases for increased morbidity and mortality associated with exposure to PM; To conduct toxicological research which reduces the uncertainties in the identification of causative PM constituents; To investigate the role of allergens and bioaerosols in the morbidity and mortality associated with PM exposure; to better characterize the relationship between mobile sources and airborne PM in the Southern California air basin; To investigate the relation between outdoor, indoor and personal exposure to PM. New York To provide the opportunity to track the impact of emission controls and their effectiveness on air quality To verify that implemented PM2.5 primary and secondary B-9
Table B–2. (Cont’d) Location
Fresno
Objectives precursor (including ozone precursor) emission controls are performing according to specifications and verify that PM2.5 and ozone air quality has responded to the emission changes achieved as expected. To provide a unique and unparalleled opportunity to enhance our understanding of ozone/PM2.5-precursor relationships and track progress in current precursor emission control programs and assess their effectiveness in achieving expected air quality responses Testing and evaluation of non-routine monitoring methods, with the intent to establish their comparability with existing methods and determine their applicability to air quality planning, exposure assessment, and health impact determination; To increase the knowledgebase of aerosol characteristics, behavior, and sources so regulatory agencies can develop standards and strategies that protect public health; and To acquire measurements that can be used to evaluate relationships between aerosol properties, co-factors, and observed health end-points.
St Louis
Southern Supersite (Taiwan)
Northern Supersite (Taiwan)
To generate a data set suitable for a variety of analyses to advance our understanding of ambient particulate matter sources, burdens and effects. To address and integrate objectives of the atmospheric, health and exposure research communities. To disseminate its findings about measurement technologies to key stakeholders such as regional planning organizations and state and local agencies. • To characterize the physical and chemical properties of ambient PM • To provide high time-resolution measurements for the study of PM pollution, exposure and health effects. • To provide the opportunity to track the impact of emission controls and their effectiveness on air quality • To enhance our understanding of ozone and its precursor relationships • To characterize the physical and chemical properties of ambient PM • To make use of the particle data for human health exposure assessment. • To provide the opportunity to track the impact of emission controls and their effectiveness on air quality
B-10
Quality Assurance (QA) In United States, the supersite quality assurance was established on guidelines developed by USEPA’s Quality Assurance Division (QAD) within the Office of Research and Development (ORD), and was required to meet USEPA’s quality assurance requirements stated in Executive Order 5360.1 (April, 1984; updated in 1998). The US supersite’s QA was processed by NARSTO’s Quality System Management Plan (QSMP) that outlines a three-tiered QA approach for environmental data collection, data consumption, data storage, as well as data dissemination and application efforts: An overarching community level QSMP that establishes a framework and associated mechanisms. The NARSTO QSMP is the framework for designing the supersite QA program. A Program Quality Management Plan (PQMP) at the Supersites program level. The PQMP articulates basic program planning; implementation and organizational approaches; broad objectives; and data acquisition, evaluation and management. This planning document constitutes part of the overall PQMP, which will be fully developed through consultation with project Principal Investigators (PIs) and the NARSTO/DOE Oak Ridge Quality Systems Science Center. Quality Integrated Work Plans (QIWP) at individual project levels. Each PI will be responsible for developing a QIWP which minimally addresses Project Planning and Organization, Management Assessment, Implementation, Data Acquisition, Data Management, Routine Controls and Procedures, and Technical Assessment and Response. The NARSTO Quality Planning Handbook 4 provides templates for developing project specific QIWPs.
Project Roles The roles of Principal Investigators, QA manager, field and lab supervisors, database managers, and data analysts are described in Table B-3:
Knowledge Gaps During the last decades in the U.S., U.S.EPA acknowledged significant gaps in its knowledge regarding exposure to toxics and the potential benefits of further reductions. There is no clear emission limits was universally accepted in public due to the uncertainties of chemical speciation and emission estimates. In addition, it is difficult to be absolutely certain that spatial and temporal scales will satisfy the basis exactly. Supersites dataset is invaluable to further investigate that emissions of the compound in fact present a risk to public health or welfare, and also help in refining emission control strategy. Other knowledge gaps are shown as the followings:
B-11
• •
Health Impacts Health outcomes predominated by carcinogenic pollutants Causal physical/chemical species Exposure/dose/response relationship and thresholds Susceptibility factors Mechanisms initiating biological responses Relationship between air toxic and health
Technical Feasibility - Carbonaceous aerosol still represents the largest uncertainties in the ambient particulate measurements. Improvements in the in=situ continuous thermal evolution carbon analysis methods are needed in order to obtain measurements that are comparable to filter-based carbon measurements. - The organic sampling artefact remains not definitive. The quartz filter adsorption artefact exceeds the particle volatilization artefact by ~50%. More research is needed to better evaluate positive and negative carbon artifacts.
Table B- 3. The Project Roles in US Supersites Project Role
Responsibilities
Principal Investigators
To oversee all project tasks and has responsibility for the successful completion of Supersite measurements and interactions with other investigators that will use the measurements
Co-Investigators
To assist in project planning and facilitates field sampling, chemical analysis, data retrieval/reformatting/processing, and data analysis/modelling tasks
Data Manager
To assemble the project database. The responsibilities include: 1) database design [structure of the database; tables used to hold data; and conventions such as names, units, flags, time conventions, etc.], 2) data processing [convert data collected from various sources in various formats and conventions to meet Fresno Supersite database conventions], 3) data traceability [design data processing procedures and documentation to provide traceability from the database back to the original data], 4) level 0 statistical checks [perform minimum and maximum checks, jump checks, and flatness checks], and 5) database documentation [assemble internal and external documentation describing database structures and data processing procedures].
B-12
Table B– 3. (Cont’d) Project Role
Responsibilities
Data Analyst
To assist principal investigators and Co-investigators in assembling the data and applying the validation tests. Comparisons will be made between in-situ continuous measurements and integrated filter measurements to establish equivalence and comparability.
Quality Assurance Manager
To specifies primary, calibration, performance test, and audit standards and the frequency of their application. The job responsibilities include: 1) defines data validity flags that qualify the information based on internal and external consistency tests. 2)
uses data from performance audits,
performance tests, and validation checks to define the accuracy, precision, and validity of each data point. These measurement attributes are added to the project database. 3) conducts on-site and laboratory system audits for each measurement.
B-13
Feasibility of Establishing Air Monitoring Supersites in Hong Kong
APPENDIX C:
Instrument Descriptions
Prepared by S.C. Leea, J.G. Watsonb, J.C. Chowb, K.F. Hoa, T. Wanga, A. Lauc, S. Liud
a
Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong b Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, USA c Atmospheric Research Center, The Hong Kong University of Science and Technology d Academia Sinica, Taipei, Taiwan
Prepared for： Hong Kong Environmental Protection Department (HKEPD) 33/F., Revenue Tower, 5 Gloucester Road, Wan Chai, Hong Kong
September 10, 2008
APPENDIX C.
INSTRUMENT DESCRIPTIONS
Table C- 1. Analytical specifications for continuous PM2.5 mass and mass surrogate instruments (From Chow et al., 2008a). Instrument and Measurement Principle Inertia instruments Tapered Element Oscillating Microbalance (TEOM) Air is drawn through a size-selective inlet onto the filter mounted on an oscillating hollow tube. The oscillation frequency changes with mass loading on the filter, which is used to calculate mass concentration by calibrating measured frequency with standards. Filter Dynamics Measurement System-Tapered Element Oscillating Microbalance (FDMS-TEOM) A TEOM equipped with a diffusion Nafion dryer to remove particle-bound water. The flow is alternated between a base flow and a reference flow every 6 min. Particles collected during the base flow are allowed to volatilize during the reference flow. If a negative mass is measured during the reference flow, due to loss of volatiles from the filter, this mass is added to the mass measurement made during the prior particle-laden samples to obtain total PM2.5 concentration.
Averaging Time
Analytical Accuracya
Precisionb
Minimum Detection Limit (MDL)
Interferences
Comparability
Data Completeness
10 min 24 hrs
± 0.75%c
± 5 µg/m3 for 10-min averagec,d
0.01 µg, which is 0.06 µg/m3 for 1-hr averagec
Loses semi-volatile species at both 30 °C and 50 °C. SES-TEOM, while less sensitive to relative humidity, does not completely eliminate loss of semi-volatile species
Underestimated FRM mass by 20 to 35%
99%
0.01 µg, which is 0.06 µg/m3 for 1-hr averagec
n/a
9 to 30% higher than FRM mass Within 10% of mass by D-TEOM, PC-BOSS, RAMS and BAM
± 1.5 µg/m3 for 1-hr averagec,d
1 hr - 24 hrs
± 0.75%c
GreenTek A light source from bscat may 0.80). Nephelometers illuminates the DustTrak: ± 1 vary depending Comparability sample air and 3 DustTrak: µg/m for 24-hr on particle size, depends on >80% the scattered Greater of 0.1% averagej for shape and conversion factor 3 c,h light is detected or 1 µg/m DustTrak composition. used. at an angle (usually 90°) 95 to 98% for relative to the Light scattering GRIMM optical source. The by DustTrak particle counter signal is related proportional to to the dp6 for dp 30% at conc. 0.80), but
Droplet Techology)
= 0.4 Mm-1 for
interferes with that by
more than 40%
10-min
particles.
lower
average) at 532
by either removing
aethalometer,
nm
NO2 from sample line
MAAP and filter
using denuders or by
IMPROVE_TOR
doing
EC.
Suggests
need
for
Light absorption by particles in air results in a heating of the surrounding air. The expansion
3
of the heated air produces an acoustic (sound wave) signal which
is
detected
by
Accounted
a
periodic
background
a
(particle-free
microphone to determine babs,
subtraction.
C-16
air)
than
different σabs.
a
n/a
Table C–3. (Cont’d) Instrument and Measurement
Averaging
Analytical
Precisionb
Minimum
Principle
Time
Accuracy
Detection
a
Limit (MDL)
Interferences
Comparability
Data Completeness
which is confined to BC using σabs=5 m2/g for the 1047 nm instrument and σabs=10 m2/g for the 532 nm instrument. Photo-ionization instruments Photoionization
monitor
polycyclic
for
5 min
n/a
n/a
~3 ng/m3 j
n/a
n/a
>91%
aromatic
hydrocarbons (PAS-PAH)
manufacturer
The air stream is exposed to UV
specified
radiation, which ionizes the particle-bound PAH molecules. The
charged
particles
are
collected on a filter element and the
piezoelectric
proportional
current to
is the
particle-bound PAH. a
Accuracy is the ability of analytical methods to quantify the observable of a standard reference material correctly; does not refer to measurement accuracy, since no standards are available.
b
No specific details on how the precision was estimated; appears to be based on replicate analysis, may not represent overall collocated measurement precision
c
Collocated precision estimates based on variation in average ratios of replicate analysis using laboratory instrument and regression slopes (Slopes for OC = 1.01, EC = 0.82, TC = 0.94; R2 = 0.97 - 0.99) of collocated field measurements.
d
Estimated using collocated AE-21 and AE-31 BC measurements at Fresno, CA.
e
While the default manufacturer recommended conversion factor (or mass absorption efficiency, σabs) is 16.6 m2/g at 880 nm, Lim et al. assumed a value of 12.6 m2/g.
C-17
f
Assuming a σabs of 10 m2/g.
g
Collocated precision estimate based on the variability of the average ratio (0.99 ± 0.12).
h
Assuming a σabs of 6.5 m2/g.
i
Assuming a σabs of 10 m2/g at 532 nm and 5 m2g at 1047 nm.
j
Specified by manufacturer as “lower threshold”; needs to be calibrated with site-specific PAH. Typically used as a relative measure in terms of electrical output in femtoamps.
n/a: Not available.
C-18
Table C- 4. Measurement and Analytical Specifications for Gaseous NO/NOx, NOy, HNO3, PAN/NO2, and NH3 Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Time
Output Range
NO/NOx/NO2/PAN/NOy NO/NOx
Teledyne
API
Using the proven chemiluminescence detection
60 seconds
± 1-2%
±0.5%
0.5 l/min ±10%
0.4 ppb
0-50 ppb to 0-20,000
Serial Port 1: RS-232
ppb full scale,
(DB-9M) Serial Port 2:
Technologies
principle, coupled with microprocessor technology to
(API 200E)
provide the sensitivity, stability and ease of use
standard
needed for ambient or dilution CEM monitoring
optional
requirements.
(DB-9F), Ethernet
Environnement-
The analyzer is based on the chemiluminescence
60 seconds to
SA
reaction which occurs between NO and O3. The Nox
9999min
Chemiluminesc
concentration is measured by first passing the sample
ent Oxides of
througn a molybdenum NO2>NO high efficiency
Nitrogen
long life converter.
± 1-2%
n/a
0.6 l/min
0.35ppb
0- 50 ppb to 0-10,000 ppb
Analyzer (Model AC31M)
C-19
RS-232
RS232/RS422
or
RS-485
Table C–4. (Cont’d) Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Output
Time
Range
Thermo
Using chemiluminescence technology to measute
Scientific
the amount of NO in the air from sub-ppb levels up
NO-NO2-NOX
to 100ppm. It is a single Chamber, single
second
10 Status Relays, and
Analyzer
photomultiplier tube design that cycles between the
averaging
Power Fail Indication
(Model 42i)
NO and NOX modes.
time)
(standard).
60 seconds
±1%
full
±0.4 ppb
0.6 l/min.
scale
0.40
ppb
(60
0-50
ppb
to
0-100,000 ppb
Selectable
Voltage,
RS232/RS485, TCP/IP,
0-20
or
4-20
mA
Isolated Current Outout (optional)
NO/NOx
EcoTech
utilises microprocessor control and
NO-NO2-NOX
chemiluminescence to measure NO, NO2 and NOx
60 seconds
±1%
±1%
0.640 l/min
0.4 ppb
0-
50
ppb
0-20000 ppb
and
Multidrop RS232 port shared
between
Analyzer (Model
analysers for data, status
EC9841A)
and control. DB50
with
discrete
status, user controls and analogue output. NO2
2B Technologies
allow measurements of NO, NOx (NOx = NO +
5 min, 15 min, 1
NO2
NO2) and NO2 (as the difference between NOx and
hr
Monitor
(Model 401)
±2%
±2%
1 l/min
2 ppb
0-2000 ppb
RS-232, Flash Card Option
NO) in the range 2-2,000 ppb. The NO2 Converter also contains an internal scrubber for scheduled zeroing of the NO Monitor. The NO2 Converter makes use of a molybdenum ("moly") converter heated to 325 °C to reduce NO2 to NO.
NO
2B Technologies
Titration of NO with Ozone with Detection of
10
NO
Ozone Depletion by UV Absorption at 254 nm.
(Data averaging
Monitor
seconds
C-20
±2%
±2%
1 l/min
2 ppb
0-2000 ppb
RS-232, Flash Card Option
Table C–4. (Cont’d) Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Time (Model 400)
Output Range
options: 10 s, 1 min, 5 min, 1 hr)
NO2/PAN
Drummond
Additional nitric oxide is added in order to stabilize
Technology
the conversion of PAN to NO2 downstream of the
PAN
GC column. The analysis time for PAN and PAN
Detector
5 min
NA
±5%
0.2 l/min
PAN:
0 to 4000 ppb
0.01 ppb
(Model
analogs such as PPN12 is 5 minutes. The column
NO2:
LPA-4D)
effluent is heated to 115°C in the presence of 0.5
0.05 ppb
Analog output RS-232
ppm of NO in order to convert the PAN to NO2 with a known efficiency. The NO2 is detected by luminol chemiluminescence. A small constant amount of NO2 is added upstream of the reaction vessel in order to keep the detector in its linear range. Inside the reaction vessel, the NO2 in air reacts with a solution containing luminol, resulting in photons (~450nm) which are viewed by a photomultiplier tube.
NOy
Thermo Scientific
Using chemiluminescence technology to measures
120 seconds
± 1%
±1%
1 l/min
< 0.05 ppb
0-5 ppb to 0-1000
Selectable
Voltage,
ppb
RS232/RS485, TCP/IP,
NOy
the amount of nitrogen oxides in the air. It is a single
(120
Analyzer (Model
chamber, single photomultiplier tube design that
second
10 Status Relays, and
42i-Y)
measures NOy which includes most oxides of
averaging
Power Fail Indication
nitrogen with the exception of NH3 and N2O.
time)
(standard). 0-20
or
4-20
mA
Isolated Current Outout (optional)
C-21
Table C–4. (Cont’d) Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Time Teledyne
API
A remote catalytic converter, mounted at the
Output Range
120 seconds
± 1-2%
±0.5%
1.8 l/min ± 0.2
Technologies
sampling point, converts reactive oxide of nitrogen
l/min (NO &
Chemiluminesce
and other organic nitrogen compounds into NO
NOy)
nce
0.05 ppb
0-50 ppb to 0-2,000
Serial Port 1: RS-232
ppb full scale
(DB-9M) Serial Port 2: standard
RS-232
NO-NOy
before they can chemically react with the sampling
optional
Analyzer (Model
system. The converter is housed in a stainless steel
(DB-9F), Ethernet
200EU)
NEMA 4X (IP65) housing with a flexible metal
or
RS-485
umbilical, connecting the converter to the Flow Module, up to 50 feet (15 meters) away. The Flow Module provides bypass flow for consistent sampling as well as a unique purge for the calibration gas tubing (included). NOy
EcoTech
In conjunction with a NOy-1000 converter for the
NOy
conversion of NOy to NO, chemiluminescence
Analyzer
(Model EC9843)
120 seconds
±1%
±0.5%
0.640 l/min
0.05 ppb
0-50 ppb to 0-2000
Multidrop RS232 port
ppb
shared
between
analysers for data, status
technology is used to measure NO and NOy
and control. DB50
with
discrete
status, user controls and analogue output. Ammonia NH3
2 ppb
0-50 ppm / 0-20000
Multidrop RS232 port
reading
with
ppb
shared
chemiluminescence detection to measure NH3, NOx
with
Kalman
analysers for data, status
(Model:
and Nx. Nx is the sum of NO+NO2+ NH3 i.e. total
Kalman
filter
and control.
EC9842)
oxides of nitrogen including ammonia. To measure
filter
active
DB50
Nx concentration, NO2 and NH3 are converted to
active
EcoTech
Chemiluminescence NH3 Analyser combines a high
Ammonia
performance ammonia converter
Analyzer
120 seconds
and proven
NO in a quartz converter heated to 750°C. In a
NA
±1%
of
0.355 l/min
between
with
status, user controls and analogue output.
C-22
discrete
Table C–4. (Cont’d) Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Output
Time
Range
separate reaction, NOx (NO+NO2) is passed through a molybdenum converter heated to approximately
325°C.The
resulting
NH3
concentration is determined by subtracting the Nx result from the NOx. Teledyne
API
Technologies
Using the chemiluminescence principle and an
60 seconds
external ammonia converter and sampling system.
±2%
of
±0.5%
reading
reading
±1%
±0.4 ppb
of
1 l/min
1 ppb
0- 50 ppb to 0-2,000
RS-232
(DB-9)
ppb
connector
0-50 ppb to 0-20000
Selectable
ppb
RS232/RS485, TCP/IP,
Chemiluminesce nt NH3 Analyzer (Model 201E) Thermo
indicates the most sensitive changes in ammonia
Scientific
concentration
Ammonia
optimization and control.
Often plants will use
seconds
10 status relays and
Monitor (Model
conductivity or pH to control ammonia, however
average
power fail indicationn
17i)
these methods are negatively influenced by other
time)
(standard). 0-20 or 4-20
for
condensate
and
120 seconds
0.6 l/min
feedwater
1
ppb
(120
ionic species from the boiler treatment chemicals in
mA
the sample. EnvironnementSA
voltage,
isolated
output (optional) 180 seconds
Ammonia
±1%
n/a
0.7 l/min
1 ppb
0-100 ppb to 0-1000 ppb
Analyzer (Model AC32M-CNH3)
C-23
RS232/RS422
current
Table C- 5. Measurement and Analytical Specifications for Gaseous CO, CO2 and SO2 Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Output
Time
Range
CO/CO2 CO
Thermo
Based on the principle that CO absorbs infrared
Scientific
CO
60seconds
±1%
± 0.1 ppm
0.5 to 2 l/min.
0.04 ppm
radiation at a wavelength of 4.6 microns. Because
0-1
ppm
to
0-100000 ppm
Selectable voltages and RS-232 (standard) 4-20
Analyzer
infrared absorption is a nonlinear measurement
mA
(Model 48i)
technique, it is necessary for the instrument
isolated current RS-485
electronics to transform the basic analyzer signal into
(optional)
a linear output. Teledyne
API
It measures low ranges of carbon monoxide by
60 seconds
±2%
±0.5%
of
0.8 l/min ±10%
40 ppb
reading
0-1 ppm to 0-1,000
Serial Port 1: RS-232
ppm
(DB-9M);Serial Port 2:
Technologies
comparing infrared energy absorbed by a sample to
Gas
Filter
that absorbed by a reference gas according to the
standard
Correlation CO
Beer-Lambert law. This is accomplished with a Gas.
optional;RS-485
Analyzer
Filter wheel which alternately allows a high energy
(DB-9F), Ethernet
(Model 300E)
light source to pass through a CO filled chamber and
RS-232
or
a chamber with no CO present. The light path then travels through the sample cell, which has a folded path of 14 meters. Ecotech
Utilising NDIR Gas Filter Correlation photometry
CO Analyser
and microprocessor control to measure CO
60 seconds
NA
0.1 ppm or 1
(Model
%
reading
EC9830A)
of
1 l/min
50
ppb
with
0-100 ppb to 0-200
Multidrop RS232 port
ppm
shared
Kalman or
analysers for data, status
300sec
and control.
filter
Ethernet connection to a
active
TCP/IP network via an RJ45 connector
C-24
between
Table C–5. (Cont’d) Observable
Instrument
Measurement Principles
Averaging
Accuracy
Precision
Flow Rate
MDL
Measurement Range
Output
CO2
Thermo
Using new optical filter technology to monitor and
±1%
±1%
1 l/min
200 ppb
0-200
Selectable
Scientific CO2
report on CO2 stack gas levels utilizing either of the
Analyzer
accepted industry standard sampling methods,
10 Status Relays, and
(Model 410i)
Straight Extractive Sampling or Dilution Sampling.
Power Fail Indication
The gas detection scheme uses optically fixed
(standard).
bandpass interference filters and quantum detection
0-20
to analyze the gas stream, leading to a superior level
Isolated Current Outout
of performance and reliability. Add to that, an
(optional)
Time 90 seconds
ppm
to
0-10000 ppm
Voltage,
RS232/RS485, TCP/IP,
or
4-20
mA
expanded ambient temperature operating range and these instruments provide excellent performance over a wider range of concentrations.
Teledyne
API
Technologies CO2
Analyser
(Model 360E)
The Model 360E measures carbon dioxide (CO2) by
60 seconds
±1%
±0.5%
0.8 l/min ±10%
200 ppb
comparing infrared energy absorbed by a sample to
0-2 ppm to 0-2000
Serial Port 1: RS-232
ppm
(DB-9M) Serial Port 2:
that absorbed by a reference according to the
standard
Beer-Lambert law. This is accomplished by using a
optional
RS-232
Gas Filter Wheel which alternately allows a high
(DB-9F), Ethernet
or
RS-485
energy infrared light source to pass through a CO2-filled chamber and a chamber with no CO2 present. Ecotech
By using microprocessor control with NDIR gas
60 seconds
EC9820
filter correlation photometry detection to measure
±1%
CO2 Analyser
CO2 with minimal interference from CO and H2O.
reading
analogue
whichever
provides status
is
outputs, control inputs
C-25
n/a
±10ppm or of
greater
1 l/min
2 ppm
0-3000ppm
DB50
with
discrete
status, user controls and output
Table C–5. (Cont’d) Observable
Instrument
Measurement Principles
Averaging
Accuracy
Precision
Flow Rate
MDL
Measurement Range
Output
Time with
and 20mA current loop
Kalman
output
filter SO2 SO2
Using the proven chemiluminescence detection
Thermo Scientific
SO2
80 seconds
±1%
principle, coupled with microprocessor technology to
±1%
of
0.5 l/min
reading
0.05ppb
0-10 ppb to 0-1000
Selectable
(300
ppb
RS232/RS485, TCP/IP,
Voltage,
Analyzer
provide the sensitivity, stability and ease of use
second
10 Status Relays, and
(Model
needed for ambient or dilution CEM monitoring
averaging
Power Fail Indication
43i-TLE)
requirements.
time)
(standard). 0-20
or
4-20
mA
Isolated Current Output (optional) Trace
Incorporates UV fluorescence spectrometry with
Analyzer
microprocessor control for accurate and reliable
Ecotech SO2
(Model
60 seconds
NA
±2%
of
0.5-0.75 l/min
reading
0.2
ppb
with
Auto-ranging
for
0-2000 ppb
output 0-20 mA, 2-20
Kalman
measurement of sub-ambient levels of SO2
Menu selectable current
mA, 4-20 mA
filter
EC9850A)
active Teledyne
API
UV Fluorescence
120 seconds
±1%
±1%
SO2
l/min
0.05ppb
±10%
Technologies Trace
0.65
0-5 ppb to 0-20,000
,Serial Port 2: standard
ppb
RS-232 optional,RS-485
Level Analyzer
(DB-9F), Ethernet
(Model 100EU) Environnement-
UV fluorescent principle consists in detecting the
SA
characteristic fluorescence radiation emitted by SO2
UV Fluorescene
molecules. In the presence of a specific wavelength
Aanlyzer
of UV light (214nm) the SO2 molecules reach a
SO2
60 seconds
±1%
n/a
0.3 l/min
1ppb
0-0.1 0-10ppm
C-26
ppm
to
2RS232 OR RS422
or
Table C–5. (Cont’d) Observable
Instrument
Measurement Principles
Averaging
Accuracy
Time Model AF22M
temporary
C-27
Precision
Flow Rate
MDL
Measurement Range
Output
Table C- 6. Measurement and Analytical Specifications for Gaseous NMHC and VOCs Averaging Observable
Instrument
Measurement Principles
Measurement Accuracy
Precision
Flow Rate
MDL
Output
Time
Range
NMHC (CH4, Total HC) gas
chromatography
system
allows
70 seconds
±2%
±2%
0.5 l/min
20
ppb
0-20 ppm
Methane,
NMHC,
NMHC (CH4,
Thermo
Back-flushed
Total HC)
Scientific 55C
measurements of NMHCs at sub-ppm concentrations,
(methane)
even in the presence of much higher concentrations of
25
methane. To start an analysis cycle, a known volume of
(NMHC as
User-selectable
air is collected in the sample loop as shown in step
propane)
conc.
ppb3
Total Hydrocarbon, and
FID
signal,
A-Backflush. The eight port valve, which is located in a
ranges 0 - 10V, 5V,
150-2000°C detector oven, is then rotated to the position
IV
shown in step B-Inject. This injects the sample into a
(standard)
flowing stream of carrier gas.
4
or
-
20
0.1
mA
(optional), RS - 231 (optional)
The sample is carried to the separation column located in a separate oven kept at 650°C. As the sample is carried through the column, various hydrocarbons move at different velocities, based on their chemical and physical properties. Due to its low molecular weight and high volatility, methane is carried back to the detector oven and measured by the FED. The valve is then returned to the original position shown in step A-Backflush. This action reverse the direction of gas flow through the column, and "backflushes" the non-methane hydrocarbons to the FID for measurement. While NMHCs are being measured, the next sample is simultaneously collected in the sample loop.
C-28
Table C–6. (Cont’d) Observable
Instrument
Measurement Principles
Averaging
NMHC (CH4,
Environnement
The analyzers uses the principle of flame ionization
Total HC)
S.A. HC51M
detection (FID) to measure the concentrtaion of
Accuracy
Precision
Flow Rate
MDL
±1%
±2%
1.4 l/min
0.05
Measurement Range
Output
0-10
RS 232/ RS 422
Time 60
seconds
to 9999 min
ppm
(NMHC as
hydrocarbons in air. The analyzer's electrometer measures
ppm
to
0-1000 ppm
propane)
the current generated by the ionization of the carbon atoms in the flame fueled by a hydrogen/air mixture. To distinguish
between
"total"
and
"non-methane"
hydrocarbons, an optional selective coverter oven is used to oxidize all the non-methane hydrocarbons. VOCs VOCs
American
(C2-C12)
Ecotech
Airmo VOC monitor use a thermoelectric cooling system VOC
AirmoZone (species
for
PAMS)
1 hr
15%
3%
Airmo
15 ppt for
or
Uninterrupted sampling with pre-concentration on 1
0.02 l/min
ne
communication
absorbent tube. Gas chromatograph with 0.2mm ID
b)
metallic column and programmable temperature gradient
VOC
oven and pressure/flow control of the carrier gas by
C6-C12: 0.06
The
(C2-C10)
Spectras Model:
preconcentration
GC955
Airmo
protocol
l/min
The system one is a GC with a built-in cooled system.
Hydrocarbons
1 hr
NA
NA
are
0.0015 0.002 l/min
to
0.15 ppb
0 to 300 ppb
NA
0-
NA
for benzene
Series
preconcentrated on Carbosieves SIII a 5°C, desorbed
0.17
600/800 POCP
thermally and separated on a combination of two
for butene
(species
columns, a capillary film column and a capillary PLOT
PAMS)
MODBUS / JBUS
1,3-butadie
VOCs
for
0-1/10/100/1000ppb
VOC C2-C6:
piezo-valve. Syntech
a.)
to control the temperature of the sample trap.
ppb
column. In this way the low boiling hydrocarbons can be separated. Analysis is done by a photo ionisation detector and a flame ionisation detector.
VOCs (BTEX)
SRI
The 60-meter capillary column is the newest unbreakable,
30 min
C-29
NA
NA
NA
0.5 ppb
MGS1
Table C–6. (Cont’d) Observable
Instrument
Measurement Principles
Averaging
Accuracy
Precision
Flow Rate
MDL
NA
NA
0.07 l/min
0.3
Measurement Range
Output
Time Environmental
fused silica-lined, stainless steel technology, which gives
and BTEX GC
good separation of the TO-14 analytes with short run
Systems
times. The PeakSimple data system controls and
(PID
and
sequences the entire analysis, collecting the data from the
FID/DELCD)
three detectors, loading and desorbing the traps, then
TO-14
calculating and printing the results. Since it is small enough to take on-site for real-time measurements, you can perform the analysis right at the source, avoiding the need for expensive, labor-intensive canister sampling. The vacuum pump interface allows the data system to turn the external vacuum pump ON/OFF under PeakSimple control. The vacuum pump is used to draw ambient air through the traps for a precise amount of time, thus enabling the system to sample unattended. The built-in air compressor eliminates the hassle of transporting bulky air cylinders by providing an endless supply of combustion air for the FID/DELCD combination detector. To eliminate cylinders altogether, try the H2-50XR hydrogen generator for both carrier and FID combustion gases. The sensitive and non-destructive PID detector can run on air carrier, too.
VOCs (BTEX)
Environnement
It is based on the gas chromatographic separation of the
15 min, 30
S.A.
VOCs
compounds of interest combined with a detection
min
Analyzer Model
achieved by a photo ionization detector(PID)or a flame
VOC71M
ionization detector(FID)
μg/m3
for benzene
C-30
0-100
μg/m3
0-1000 μg/m3
and
RS 232/ RS 422
Table C- 7. Summary of volatile organic components acquired in Hong Kong.
Volatile Organic Compounds
HKGLa
Special Studyb
Proposed Synspec AirmoZone PAMSc TO-14d VOC species
1,1,1-Trichloroethane 1,1,2,2-Tetrachloroethane 1,1,2-Trichloroethane 1,1-Dichloroethane 1,1-Dichloroethene 1,2,3,5-Tetramethylbenzene 1,2,3-Trimethylbenzene 1,2,4,5-Tetramethylbenzene 1,2,4-Trichlorobenzene 1,2,4-Trimethylbenzene 1,2,4-Trimethylcyclohexane 1,2-Dibromoethane 1,2-Dichloroethane 1,2-Dichloropropane 1,2-Diethylbenzene 1,3,5-Trimethylbenzene 1,3-Butadiene 1,3-Diethylbenzene 1,4-Dichlorobutane 1,4-Diethylbenzene 1-Butene/Iso-Butylene 1-Butyne 1-Decene 1-Heptene 1-Hexene 1-Methylcyclohexene 1-Methylcyclopentene 1-Nonene 1-Octene 1-Pentene 1-Propyne 2,2,3-Trimethylbutane 2,2,4-Trimethylpentane 2,2,5-Trimethylhexane 2,2-Dimethylbutane 2,2-Dimethylhexane
C-31
Table C–7. (Cont’d) Volatile Organic Compounds
HKGLa
Special Studyb
2,2-Dimethylpropane 2,3,4-Trimethylpentane 2,3-Dimethylbutane 2,3-Dimethylpentane 2,4-Dimethylhexane 2,4-Dimethylpentane 2,5-Dimethylhexane 2-Butyl nitrate 2-Ethyl-1-Butene 2-Ethyltoluene 2-Methyl-1-Butene 2-Methyl-2-Butene 2-Methylbutane 2-Methylheptane 2-Methylhexane 2-Methylpentane 3,6-Dimethyloctane 3-Chlolopropene 3-Ethyltoluene 3-Methyl-1-Pentene 3-Methylheptane 3-Methylhexane 3-Methylpentane 4-Ethyltoluene 4-Methyl-1-Pentene 4-Methylheptane Benzene Benzyl chloride Bromodichloromethane Bromoethane Bromoform Bromomethane Bromotrichloromethane Butane Carbon monoxide Carbon tetrachloride
C-32
Proposed Synspec AirmoZone PAMSc TO-14d VOC species
Table C–7. (Cont’d) Volatile Organic Compounds
HKGLa
Special Studyb
Chlorobenzene Chloroethane Chloroethene Chloroform Chloromethane cis-1,2-Dichloroethene cis-1,2-Dimethylcyclohexane cis-1,3-Dichloropropene cis-2-Butene cis-2-Heptene cis-2-Hexene cis-2-Pentene cis-3-Heptene cis-3-Methyl-2-Pentene cis-4-Methyl-2-Pentene Cyclohexane Cyclohexene Cyclopentane Cyclopentene Decane Dibromochloromethane Dibromomethane Dichloromethane Dodecane Ethane Ethene Ethyl nitrate Ethylbenzene Ethyne Freon 11 Freon 113 Freon 114 Freon 12 Freon 22 Heptane Hexachlorobutadiene Hexane
C-33
Proposed Synspec AirmoZone PAMSc TO-14d VOC species
Table C–7. (Cont’d) Volatile Organic Compounds
HKGLa
Special Studyb
Hexylbenzene Indan Iso-Pentane Iso-Propyl nitrate Iso-Butane Iso-Butylbenzene Isoprene Iso-Propylbenzene m,p-Xylene m-/p-Chlorotoluene m-Dichlorobenzene M-Diethylbenzene Methane Methyl nitrate Methylcyclohexane Methylcyclopentane Methylene chloride Naphthalene n-Butylbenzene Nonane n-Propyl nitrate n-Propylbenzene o-Chlorotoluene Octane o-Dichlorobenzene o-Xylene Pamshc p-Cymene p-Dichlorobenzene P-Diethylbenzene Pentane Propane Propene Propylene Sec-Butylbenzene Styrene Tert-Butylbenzene
C-34
Proposed Synspec AirmoZone PAMSc TO-14d VOC species
Table C–7. (Cont’d) Volatile Organic Compounds
HKGLa
Special Studyb
Proposed Synspec AirmoZone PAMSc TO-14d VOC species
Tetrachloroethene Tetrachloroethene Tetrachloroethylene Toluene trans-1,2-Dichloroethene trans-1,2-Dimethylcyclohexane trans-1,3-Dichloropropene trans-2-Butene trans-2-Heptene trans-2-Hexene trans-2-Pentene trans-3-Heptene trans-3-Methyl-2-Pentene trans-4-Methyl-2-Pentene Trichloroethene Trichloroethylene Trichloromethane Undecane Total number of species 143 50 29 61 61 34 87 Routine Analysis Performed by Hong Kong Governmental Laboratories. b Special Study Analysis Based On the Analysis Performed by Professor Don Blake at University of California and Irvine. c VOC Acquired at U.S.EPA’s Photochemical Assessment and Monitoring Network. d U.S.EPA Air Toxic List. a
C-35
Feasibility of Establishing Air Monitoring Supersites in Hong Kong APPENDIX D: Supersite Measurement Specifications
Prepared by S.C. Leea, J.G. Watsonb, J.C. Chowb, K.F. Hoa, T. Wanga, A. Lauc, S. Liud
a
Department of Civil and Structural Engineering, Hong Kong Polytechnic University, Hung Hom, Hong Kong b Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512, USA c Atmospheric Research Center, The Hong Kong University of Science and Technology d Academia Sinica, Taipei, Taiwan
Prepared for： Hong Kong Environmental Protection Department (HKEPD) 33/F., Revenue Tower, 5 Gloucester Road, Wan Chai, Hong Kong
September 10, 2008
APPENDIX D.
SUPERSITE MEASUREMENT SPECIFICATIONS
Table D- 1. Measurements and Analytical Specification for the Proposed Hong Kong Supersites Program
Observable
Instrument
Measurement Principles
Averaging Time
Accuracy
Precision
Flow Rate
MDL
Measurement Range
Mass Concentration Integrated PM Mass TSP
Hi-vol (General 2310)
sampler Metals
When applied to a high volume air sampler, this flow control principal incorporates a smooth-wall venturi orifice that gradually opens to a recovery section. Vacuum is provided by a motor downstream of the venturi. Over 95% of the energy lost in differential pressures across the restricting orifice is recovered in this design.
NA
