Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 CRITICAL REVIEW DISCUSSION
Measurement Methods to Determine Compliance with Ambient Air Quality Standards for Suspended Particles
INTRODUCTION
John G. Watson Critical Review Sub-Committee Chairman and Research Professor, Desert Research Institute PO Box 60220 Reno, NV 89506
At the Air & Waste Management Association’s annual meeting in San Antonio, Texas, Dr. Judith C. Chow provided a comprehensive review and evaluation of methods to measure particles suspended in the atmosphere. This review was timely in that the U.S. Environmental Protection Agency (EPA) intends to revise the current National Air Quality Standards (NAAQS) for suspended particles in 1997. In her presentation, Dr. Chow summarized from the written Critical Review! technical findings that supported seven recommendations for future compliance monitoring methods. These recommendations were as follows:
1. Compliance monitoring methods should be based on performance standards, not on design standards.
2. The upper particle size cut for a compliance monitor should be between 2 and 3 um.
3. Particle mass should be the primary property measured for compliance with the 1997 NAAQS,
4. Standard operating procedures need to be part of the performance standard and need to be specific with tespect to sample handling, storage, equilibration, and filter acceptance testing.
5. Compliance monitors should be capable of determining chemical components of suspended particles,
6. Trade-offs between increased sample density, frequency and duration, and decreased measurement precision need to be evaluated.
7, Research is needed on continuous measurement of major chemicals in suspended particles to support the next round of NAAQS review.
Dr. Chow was specifically charged by the Critical Review subcommittee with drawing conclusions in these areas, in full recognition that others might have alternate interpretations of the published literature. In her presentation, she provided examples drawn from nearly 1,000 literature citations to support her recommendations. Discussants were recruited to offer additional opinions on the content of the review and Dr. Chow’s conclusions. Additional reactions were solicited from those attending the session and from the ARWMA membership. This article summarizes the substance of these presentations.
666 Journal of the Air & Waste Management Association
Each of the invited discussants was asked to participate based on demonstrated experience and scientific contributions in the areas of particulate monitoring, health effects research, and regulatory planning.
Dr. George Thurston is an Associate Professor at the New York Institute of Environmental Medicine, New York University. He currently conducts research on air quality monitoring and the effects of pollution on human health, and has extensively published in this area. Dr. Thurston received his doctoral degree from the Harvard School of Public Health where he analyzed data acquired in the Harvard Six Cities Study.
Mt. Neil Frank is principal monitoring advisor with the Office of Air Quality Planning and Standards (OAQPS) at the EPA. His current responsibilities involve the development of measurement strategies to determine compliance with a new particle standard, Prior to this assignment, Mr. Frank was chief of the OAQPS monitoring section. Mr. Frank has been at EPA since its inception and carried out many of the statistical analyses that justified the form of the current particulate standard for PM, 9.
Dr. James Lodge is an air quality consultant and Editor of the Intersociety Committee’s Methods of Air Sampling and Analysis.2, This book, which is updated every five to ten years, codifies current knowledge of air pollution measurement methods, including those for suspended particles. Dr. Lodge’s career includes more than forty years of experience in aerosol measurement and atmospheric science. He was a principal editor of Atmospheric Environment since its inception, for 35 years, and is highly regarded as a consultant in atmospheric chemistry. He has authored or co-authored over 200 publications, many of which have dealt with aerosol measurement. Dr. Lodge received his B.S. in Chemistry from the University of Illinois and his Ph.D degree from the University of Rochester.
Dr. Russell Wiener is an aerosol physicist at EPA and former Chief of the aerosol physics and chemistry branch. The discussant presentation in San Antonio was given by Mr. Frank McElroy, an environmental engineer with the Office of Research and Development (ORD) at EPA, where he has been a principal in developing EPA’s reference and equivalent methods program over the past two decades. Dr. Wiener and Mr. McElroy is currently examining different measurement alternatives for determining compliance with a revised particulate standard.
Dr. Michael Kleinman is Professor of Community and Environmental Medicine at the University of California at Irvine, where he is involved in a wide array of epidemiological and toxicological studies. Dr. Kleinman received his
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
doctoral degree in environmental health from New York University where he developed receptor-oriented source apportionment methods for suspended particles.
The comments received from other contributors follow those of the invited discussants. Though the invited discussants have had an opportunity to review and revise their transcripts, the Critical Review sub-committee chairman has condensed and edited these comments to meet space and style requirements. Written and transcribed comments from other contributors have also been edited and condensed. Where the intent has been misinterpreted, discussants and commentators may make clarifications via letters to the editor.
References 1. J. C. Chow, “Critical review: measurement methods to determine compliance with ambient air quality standards for suspended particles,” J. Air & Waste Manage. Assoc. 45:320-382 (1995). 2. J.P. Lodge, Jr. (ed.), Methods of Air Sampling and Analysis, Third Edition, Lewis Publishers, Inc., Chelsea, MI (1989).
ISSANT PRESENTATIONS
Dr. George Thurston. Associate Professor New York Institute of Environmental Medicine Long Meadow Rd., Tuxedo, NY 10987
Dr. Chow has provided an excellent and thorough review of the issue of Particulate Matter (PM) measurement and its implications for the regulatory process. She's presented a comprehensive review of the historical background of PM measurement, including the rationale for its need, the known characteristics of particles, human exposure and inhalation factors, ambient particulate matter sampling methods and variables, and laboratory analysis methods. These accomplishments will make this Critical Review a widely referenced and relied upon treatise on particulate matter measurements in the years to come. In addition, Dr. Chow has provided seven recommendations regarding particulate matter monitoring in the future, and my comments address these recommendations.
The first recommendation, that compliance monitoring should be based on performance rather than design standards, makes excellent sense. New methods should not have to replicate the errors of past sampling methods, such as the particle volatilization and artifact problems known to plague existing filter-based methods.
The second recommendation, that a standard based on a of 2 to 3 um size fraction should be implemented, is a logical extension of what we now know about the characteristics of particles. It is justified by the greater strengths of fine particle associations with adverse human health effects exhibited in recent epidemiological studies of particulate matter and human morbidity and mortality. An example is the Harvard Six City Study that recently reported stronger PM associations
Volume 45. September 1995
1995 Critical Review Discussion
with mortality for fine particle mass measures and sulfates.! This confirms past work looking cross-sectionally at mortality throughout the U.S. in 1980, which also found stronger mortality associations with fine particles and sulfates.2 These results are also consistent with research that I’ve conducted on hospital admissions in Toronto, Ontario, in which we found non-significant associations with coarse particles, but statistically significant respiratory hospital admissions associations with fine particles, sulfates, and (especially) fine particle acidity.3 It is apparently the fine particle mass and its constituents that are driving the health effects relationships that we've found with PM), (particles less than 10 microns in aerodynamic diameter) and TSP (Total Suspended Particulate, particles less than ~30 microns in aerodynamic diameter) in the past. Thus, setting a fine particle standard is an appropriate step.
The third recommendation, that particle mass concentration should be the primary property measured, is not consistent with the real, but unmet, need for one or more composition-based particulate matter standards. The 1995 Critical Review cites the 1973 Critical Review by Wesley Hemeon,? noting that he said, “The vague term particulate matter with no designation of its chemical and physical nature violates every principle of toxicology.” Hemeon also observed: “The term particulate matter is entirely improper as a definition of an air pollutant in the primary standard, i.e., causing adverse effect on health, especially in acute episodes. Any agent nominated as a health hazard should certainly be identified according to its chemical composition.”
I agree with Hemeon, and feel, as I wrote over a decade ago, that: “The grouping of all particles together as total suspended particulate matter may be as misleading with respect to pollution exposures as would the grouping together of all gases as total gases.” We certainly recognize in the present standards that there is a health effects distinction between ozone and oxygen, and we ought to start recognizing there are also health effects differences among the particles that we’re weighing on a filter. The fact that particulate matter is not one pollutant, but is a class of pollutants, needs to be recognized in our health-based ambient standards.
While the 1987 PM,, standard eliminated the larger nonthoracic particles from regulation, all thoracic particles are still regulated to the same extent, irrespective of their toxicity. This means that, under current regulation of this pollutant class, some particle types are overregulated while others are underregulated. This must necessarily have led to inefficiencies in the control of particulate matter and, hence, wasted money and/or insufficiently protected public health. The EPA and other regulatory agencies should stop ignoring the composition and toxicity differences among particles from different sources and both: 1) set composition based particle matter standards; and 2) commence particle composition sampling on a regular basis throughout their ambient sampling networks.
Journal of the Air & Waste Management Association 667
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
The Critical Review states that the epidemiological studies have not found definitive relationships between specific chemical species and ill health. I disagree. In particular, the evidence summarized in the EPA’s 1995 draft Air Quality Criteria for Particulate Matter identifies adverse health effects by acid aerosols in both animal and human studies, providing a compelling argument for the immediate implementation of an acid aerosol ambient air quality standard, in addition to an appropriate particle mass standard. The epidemiological evidence of significant adverse health effects by acid aerosols (primarily as sulfuric acid and ammonium bisulfate) includes direct measurements of acid aerosols collected by Comminsé during the historical London fog mortality and morbidity episode of 1962. More recent studies show associations between acid aerosol exposures and increased symptoms by asthmatics,” decreased lung function in children,’ increased respiratory hospital admissions,3 and increased daily mortality.!° There is a pressing need to set an ambient air quality standard to protect public health from the deleterious effects of acidic particles.
The fourth recommendation, that specific standard operating procedures need to be a part of the particulate matter performance standard, is a reasonable and essential requirement.
The fifth recommendation, that compliance monitors should be capable of determining the chemical components of suspended particles, is on target as far as it goes. As noted previously, it is essential that we know the composition of particles if we are to know their health consequences and regulate their sources appropriately. Therefore, I would modify this recommendation to require that the chemical composition of collected particles be determined. This need not be done on every sample. Such characterization might be required only on every sixth day and for samples on which the particulate matter mass exceeds a designated concentration. I understand that some states already do this to some extent. Such chemical characterizations of PM samples should be conducted throughout the U.S. in order to provide a basis for more effective protection of public health.
The sixth recommendation, that compliance monitoring strategies need to consider sampling density, frequency and duration, is appropriate but does not go far enough. Recent epidemiological studies suggest that the greatest effects of particulate matter air pollution may be due to multiday episodes rather than to a single 24-hour peak. At this time, regulators in most cities are only required to monitor PM once every sixth day. Monitoring and regulating this pollutant class using measurements from one day out of six does not make sense from a public health perspective. There is a need for continuous or daily sampling at all particulate matter monitoring stations.
The seventh recommendation, that research is needed to develop continuous measurement methods for major chemicals in suspended particles, is exactly correct. If we
668 Journal of the Air & Waste Management Association
are to make the transition from a non-specific mass-based standard to more relevant standards based on chemical composition, automated continuous measurement methods will be absolutely essential.
In summary, I feel Dr. Chow has done an admirable job of reviewing present knowledge of PM measurement methods for compliance with ambient standards. In general, | also agree with her recommendations for PM compliance monitoring in the future, with the notable additions that I feel there is also an immediate need: 1) to require daily PM sampling, and; 2) to supplement the existing size-selective mass-based standard with composition-based PM standards, such as for acidic aerosols.
References
1. D.W. Dockery, C.A. Pope, III, X. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.G. Ferris, Jr., and RE. Speizer, “An association between air pollution and mortality in six U.S. cities,” N. Engl. J. Med. 329:17531759 (1993).
2. _H. Ozkaynak and G.D. Thurston, “Associations between 1980 U.S. mortality rates and alternative measures of airborne particle concentration,” Risk Anal. 7:449-461 (1987).
3. G.D. Thurston, K. Ito, C.G. Hayes, D.V. Bates, M. Lippmann, “Respiratory hospital admissions and summertime haze air pollution in metropolitan Toronto, Ontario,” Environ. Res. 65:271-290 (1994).
4. W.C.L, Hemeon, “A critical review of regulations for the control of particulate emissions,” JAPCA 23:376-387 (1973).
5. G.D. Thurston, “A source apportionment of particulate air pollu-
tion in metropolitan Boston,” Doctoral Dissertation, Harvard School
of Public Health, Boston, MA (1983).
B.T. Commins, Personal Communication. January 16, 1991 (1991).
B.D. Ostro, MJ. Lipsett, M.B. Wiener, and J.C. Selner, “Asthmatic
responses to airborne acid aerosols,” Am.J. Public Health 81:691-702
(1991).
8. L.M. Neas, D.W. Dockery, P. Koutrakis, DJ. Tollerud, and FE. Speizer, “The association of ambient air pollution with twice daily peak expiratory flow rate measurements in children,” Am. J. Epidemiol. 141:111-122 (1995).
9. MJ. Studnicka, T. Frischer, R. Meinert, A. Studnicka-Benke, K. Hajek, J.D. Spengler, and M.G. Neumann, “Acidic particles and lung function in children,” Am. J. Respir. Crit. Care Med. 151:423-430 (1995).
10. M. Lippmann and K. Ito, “Separating the effects of temperature and season on daily mortality from those of air pollution in London: 1965-1972,” Inhalat. Toxicol. 7:85-97 (1995).
NS
Mr. Neil Frank Principal Monitoring Advisor Office of Air Quality Planning and Standards U.S. Environmental Protection Agency Research Triangle Park, NC 27711
This commentary represents my views and those of other U. S. Environmental Protection Agency (EPA) staff. It does not necessarily represent official agency policy. My comments to the Critical Review start with a brief overview of differences in monitoring networks used for assessing ambient air quality. With this introduction I will respond to the recommendations in the context of new measurement methods for fine particles.
In accordance with the Clean Air Act, each state is responsible for a State Implementation Plan (SIP) to implement, maintain and enforce the NAAQS. The SIP must
Volume 45 September 1995.
Downloaded by [66.67.115.93] at 07:08 26 November 2015
include provisions to establish and operate an ambient air quality surveillance system. Monitoring regulations embodied in 40 CFR (Code of Federal Regulations) 58 specify that this system consist of a network of monitoring stations designated as State and Local Air Monitoring Stations (SLAMS) that measure ambient concentrations of those pollutants for which NAAQS have been established. The principal objectives for the SLAMS networks are to determine: 1) the highest concentrations expected to occur in the area covered by the network; 2) representative concentrations in areas of high population density; 3) the impact on ambient pollution levels from significant sources or source categories; and 4) general background concentration levels. A subset of the SLAMS are the National Air Monitoring Stations (NAMS) which are located in the largest urban areas. There are currently 1215 monitors identified as SLAMS for particulate matter, 293 of which are designated as NAMS.
The primary role of the SLAMS is to judge attainment and non-attainment with the NAAQS. SLAMS monitoring must utilize either the Federal Reference Method described in 40 CFR 50 or equivalent methods in accordance with 40 CFR 53. In addition, SLAMS monitors must conform to standardized monitoring schedules, instrument siting criteria, and quality assurance procedures. Because of the varied air surveillance objectives and the large number of SLAMS monitors used nationwide, the SLAMS networks must be costeffective and, therefore, utilize relatively simple monitoring methods that are easy to implement at minimum cost.
In addition to the goal of general air quality surveillance, SIP monitoring is also needed to support air quality models and to help develop pollution control strategies for attainment and maintenance of the NAAQS. For these and other SIP purposes, current monitoring regulations are less specific than those designated for monitoring compliance. To best serve the needs of individual geographic areas with specific PM problems, more specialized, non-reference methods may be operated on different sampling schedules; these do not need to be part of the routine compliance monitoring program. Similarly, research monitoring is used for various purposes including the support of health, ecological, and atmospheric programs and is not an objective of SLAMS monitoring. Research monitoring tends to be highly specialized, can involve emerging and expensive monitoring technologies, and is often difficult to incorporate into routine monitoring programs.
I disagree with the first recommendation for a performance rather than a design standard. Health effects have been observed in association with measurements of PM. Until the precise causal agents in suspended particles are identified, however, EPA uses a surrogate measure, referred to as an indicator, to represent the agents of concern. In the case of PM, this indicator is PM mass concentration in a specified size fraction. In order to treat the regulated community fairly and to provide a uniform level of health
Volume 45 September 1995
1995 Critical Review Discussion
protection across the nation, the indicator must be consistently defined in terms of stable, reproducible measurements. It is difficult to accomplish this using a performance standards approach.
The performance standards approach used for the current Federal Reference Method has allowed for approval of several PM,, monitoring instruments and also allowed for innovative technologies to emerge that provide useful PM measurement capabilities. However, this process has resulted in systematic relative biases of 5% to 15% or more among approved reference methods. These biases have been attributed to differences in sampler design as well as to differences in maintenance and operating procedures. Together with PM,) measurement precisions of ~+10%, the performance standard approach has yielded relatively large differences among 24-hour PM,, concentrations when samplers are collocated in the field. Recognition of this situation led EPA to adopt a 1988 policy stating that PM,) measurements produced by collocated reference instruments with different designs are not interchangeable and represent different monitoring sites for regulatory purposes.
Even more dramatic are large systematic differences between approved equivalent samplers and reference instruments for PM, 9. Relative differences greater than 40% between field measurements produced by collocated reference and equivalent samplers have been reported. Such large differences are only apparent under certain environmental and ambient aerosol sampling conditions, and these were not anticipated by the performance and approval criteria promulgated in 1987. Experience shows that there are unanticipated variables that can produce real effects in the field. The time constraints associated with revision of the current PM NAAQS add to the difficulty of ensuring that appropriate performance standards can be developed to achieve acceptably low variability among future regulatory measurements.
Current thinking among EPA staff for a revised PM NAAQS is a hybrid approach for a new fine particle (PMgne) measurement method. This would involve specification of a samplier by design for the Federal Reference Method, while allowing the designation of equivalent methods based on performance standards and field comparability. As is current practice, both reference and equivalent methods could be used to judge attainment and non-attainment with NAAQS. EPA staff believe that a design approach for the Federal Reference Method is the best way to minimize variability among PMane Measurements. The design approach should have appropriate quality assurance (QA) and standard operating procedures (SOPs) as recommended in the Critical Review. Furthermore, the inherent imprecision of the reference method design (as well as approved equivalent methods) must be addressed in equivalency assurance and certification procedures. With such a hybrid approach, the revised regulations should provide sufficient flexibility in usable
Journal of the Air & Waste Management Association 669
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
measurement methodology, permit measurement technology to advance, but at the same time achieve nationally consistent and reproducible PM indicator measurements.
EPA staff have established a variety of design standard goals for a potential new PMgne Measurement method, including the following characteristics and measurement objectives: 1) filter based gravimetric measurement with the capability for chemical analysis; 2) amenable to everyday (sequential) sampling and comparable to existing manual methods (PM, PM, ;); 3) readily available, inexpensive, and easy to use; 4) applicable to a 24-hour measuring period; 5) able to operate under a wide range of ambient conditions (temperatures and pressures); 6) possessing a sufficiently sharp particle size cut-point to minimize biases from particles larger than specified size fraction (especially for a new PMgine indicator); and 7) possess proven quality assurance and standard operating procedures.
A PMgine Sampler that achieves these objectives can be quickly incorporated into SLAMS networks, be operated on a daily basis, and be capable of providing at least preliminary information needed to characterize the emissions sources that affect PM concentrations. More importantly, such measurements can be comparable to existing PM data bases used in health studies that form the basis of the NAAQS. These measurements would ensure uniform health protection nationwide and provide a consistent target for developing control programs. Although alternative, equivalent methods will be encouraged under this process, the selected design essentially becomes the reference standard against which the other systems will be judged, thus ensuring consistency.
The Critical Review concludes that “...newly developed methods should not be constrained to showing equivalence with a less-perfect method embodied in design standard.” Systematic differences among approved monitoring methods will be very undesirable from a health point of view. Significant relative bias among reference and equivalent methods translates into a change in the definition of the indicator pollutant. This in turn changes the health protection afforded by the NAAQS and the impact of pollution control decisions faced by the States. I do not want a changing yardstick for judging attainment or nonattainment with the NAAQS, so desirable improvements in the measurement capability for a revised NAAQS would be accomplished within the context of measurements comparable to those produced by the design standard. Although the new process for approval of equivalent samplers must address relevant issues, such as particle volatilization, this would be done in the context of performance characteristics of the promulgated design method.
With respect to the second recommendation concerning the size fraction associated with PM,,,., the current draft of the EPA criteria document concludes that a cut point between 1 and 3 is appropriate. This document is currently undergoing public and scientific review. EPA’s Clean Air
670 Journal of the Air & Waste Management Association
Scientific Advisory Committee discussed this issue at a public meeting in August 1995, but the results of this discussion are not available at the time of this publication. EPA staff are currently working with a PM, ; size fraction, given the limited time available for the PM NAAQS review and because the existing PMane database and many health studies are based on PM,,, measurements. However, EPA is considering other size fractions and chemically specific indicators. Until scientific review is completed, EPA has not ruled out the possibility of other size fractions (e.g. PM, 0).
I agree with the third recommendation that mass is the most practical and cost-effective property of PM that can be measured in widely deployed network. A filter based gravimetric measurement is stated as one of the goals for a Federal Reference Method and EPA staff are currently working under this assumption.
I agree with the fourth recommendation that more specific standard operating procedures (SOPs) are needed to ensure the highest quality ambient air quality data. These SOPs must be part of an integrated QA program. Inlet maintenance, filter handling and storage, lab analysis, and instrument performance testing are all needed to reduce differences among PM measurements. For example, revised procedures might require filters to be loaded into filter holders in a clean laboratory and sealed prior to sampling and analysis. Some consideration might also be given to refrigeration or rapid collection of samples under certain circumstances in order to reduce particle volatilization. As part of an overall QA program, there is a need for effective filter labeling and tracking systems, especially in networks where large number of filters are involved. The inclusion of SOPs as part of the reference method also forces the development of high-quality laboratory facilities.
I also agree with the fifth recommendation, that chemical measurements are needed and could be very helpful in creating the databases needed to support health studies. Chemical measurements do not necessarily need to be made with the same monitoring method that will be used for judging NAAQS attainment and non-attainment; such measurements do not need to be part of routine monitoring programs.
The importance of chemical analysis capability is acknowledged as one of the design goals of the Federal Reference Method, as previously enumerated. This is necessary for preliminary assessments of high concentrations produced with routine ambient air quality monitoring (e.g. identify culpable emission sources, determine the existence of special conditions, verify data validity, etc.). As noted in my introductory remarks, however, monitoring and analytical methods that are not designated as reference or equivalent can be used to properly characterize the chemical components of ambient aerosols and to perform more detailed and accurate assessments. A more flexible monitoring approach can allow the use of a variety of monitoring methods,
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
including those that utilize multiple filter media or denuders. Because these measurements may provide more accurate assessment of individual chemical species, they can exhibit different (e.g., higher) PMg,. mass concentrations than those produced by the approved reference method. Under these circumstances, such measurements would be representative of a different PMg,. indicator pollutant.
Consequently, technical assistance guidelines must be developed to describe acceptable methodologies for chemical characterization and to establish linkages between such non-equivalent methods and the reference method. Guidance for control strategy development and other regulatory purposes is especially needed when detailed chemical analyses are performed on different days than those on which the NAAQS exceedances occurred. The choice of method may depend on the intended data use (control strategy development, characterization of upwind and precursor concentrations, emissions tracking, and inventory reconciliation or exposure assessment) and may vary according to sampling location or geographic region.
Some possibilities for a revised PM NAAQS include a requirement for periodic chemical analyses using the approved reference or equivalent methods at selected locations (e.g., at NAMS) or in certain designated nonattainment areas (as is currently done for precursors of ozone with the Photochemical Assessment Monitoring Stations, PAMS). There is still time to discuss this issue, since such requirements could be promulgated subsequent to a revision of the current PM NAAQS.
Talso agree with the sixth recommendation, that improvements in network design are needed, but I also concur with the Critical Review observations that this is difficult to fully consider within the time frame provided for the current NAAQS review. Such analyses must also be somewhat conceptual because limited PMg,, data are currently available and much of these data are from special studies. A monitoring strategy consists of much more than a compliance monitoring method. The regulations for Ambient Air Quality Surveillance (40 CFR 58) specify schedules for the required frequency of monitoring, number of required monitors, allowable scales of measurement, sampler placement, and QA requirements. Guidelines have always been used and will again be produced to assist with the design of networks and to describe the role of alternative (i.e. non-reference) methodologies.
Several concepts are currently under discussion by EPA staff to permit more flexibility in new PM regulations and to optimize the use of scarce monitoring resources. Different sampling requirements (and correspondingly different monitoring data interpretation) might be instituted for a core network of NAMS/SLAMS monitors compared to other, special purpose monitors. Inexpensive, less precise, portable samplers might be used to characterize spatial patterns, but not to judge attainment. Surrogate samplers (like nephelometers), calibrated against reference samplers on a site by site or regional basis, might be used to design
Volume 45 September 1995
1995 Critical Review Discussion
networks, encourage the need for additional sampling or to initiate abatement and emission control programs.
I agree with the final recommendation that new technologies are needed to supply the information to set and comply with air quality standards in the 21st century. It is essential to do parallel sampling using current manual methods together with existing and emerging PM measurement technologies. Measurements are needed for different averaging times and alternative indicators of particulate matter (including dry aerosols) than are commonly used today. This is particularly important in areas where epidemiological studies are planned or are in progress to help understand relationships between PM and health and to acquire data sets for the next PM NAAQS review. However, this should not be part of a regulatory compliance monitoring program nor is it the responsibility of state and local monitoring agencies. It belongs in the research arena, but should be supported and promoted to the extent possible by the EPA.
Monitoring methods used to measure the PM indicator pollutant should be nationally consistent, sufficiently reproducible, and easy to implement. The most consistent measurements of particulate matter can be accomplished with a design standard for the Federal Reference Method, accompanied by comparable equivalent instruments to be judged on the basis of performance standards. The design standard would explicitly define the PMgine indicator, against which all compliance monitoring will be fairly and equitably judged for the effective duration of a revised NAAQS.
The principal function of compliance monitoring in NAMS/SLAMS networks is the determination of attainment or nonattainment of the NAAQS, and this must be performed with reference or equivalent instruments. Equivalent methods used for regulatory purposes must be comparable to the reference method, but alternative non-reference methods could and should be utilized for improved chemical, spatial and temporal characterization of the ambient aerosol. A new monitoring strategy must recognize the need for alternative, cost-effective measurement methodologies and resources must be allocated among various data requirements.
Dr. James P. Lodge, Jr. Consultant in Environmental Chemistry and Editor of Methods of Air Sampling and Analysis
801 Circle Dr. Boulder, CO 80302
Dr. Chow has done an enormous job in collecting the applicable literature. The weight of this commentary is a series of loud “amens” to points that she has made, and they should appear in large print, boldface, framed, and underlined. As they are printed now, it is all too easy to ignore them if they are inconvenient or embarrassing. My adverse comments may take up more space, but they are for the most part
Journal of the Air & Waste Management Association 671
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
quibbles that require explanation, and hence verbiage. I would sacrifice most of them if necessary to get the points of agreement accepted.
I applaud her strong stand for continuing the use of performance specifications for the reference method sampler. Rumors have suggested that EPA has been thinking seriously of abandoning this very rational step. While there is evidence that they have never been comfortable with it, it is an approach that I think the community is within its rights to demand.
I am also very pleased by her clear stand to stop the practice of reducing measured concentration to “reference conditions.” There is no justification for this.! Her suggestion of different standards for different altitudes is a reasonable one, provided it is not used as an excuse simply to tighten the standard without justification. As one who advocated a more permissive standard,? I know that many enthusiasts never heard my reasoning, simply because it did not “tighten the screws” enough for their taste.
Dr. Chow begins her review by quoting, without comment, the current NAAQS for PM, .3 I said at the time that the NAAQS was published, and I say again, that the use of arithmetic average is a statistical abomination. Concentrations of particulate matter are not distributed normally, but lognormally. Hence, the use of the arithmetic mean is wrong, actually concealing information. The sole reason for use of the arithmetic mean is that it yields bigger numbers than the geometric mean. As to the actual level of the standard, I have commented on this in the past,? and will not indulge in further polemic, beyond observing that it appears that the sole grounds for credibility in studies of health effects are that they indicate ever-lower thresholds of adverse effects.
No case is made in the review for measuring a size fraction lower than 2 um. The attitude seems to be that, thanks to an essentially arbitrary decision made many years ago, we have some information on PM, ,, so it is a foregone conclusion that this will become the basis of the new standard. Prior to setting the PM,, standard, a great deal of effort was spent measuring PM,,, and pressure was brought to make that the basis for a revised standard. The technology was simply proven not to work. PM,, was settled on as the minimum change that the community would accept, despite good reasons for going to PM,.‘ I had a lengthy discussion with the late K. T. Whitby’ over the optimum upper boundary for fine particles, which he held to be “...1.8 ym, or, if you want a purer sample of fine particles, 1 ym.” If fine particles are the cause of health effects, and if routine monitoring data are used in epidemiological studies, as Dr. Chow anticipates, we must advocate the use of PM, o to clarify the composition of the population of the fine particle modes.
Dr. Chow ably discusses the mechanisms that allow the “accumulation mode” to accumulate: constant input at the low end by coagulation; an upper limit set by lack of coagulation, and sedimentation at the high end. She does not
672 Journal of the Air & Waste Management Association
cover the lack of removal by washout and rainout over the nominal size range of the population.$ If the dispersions of the accumulation and coarse modes are similar, the minimum marks the point where the mass of geologic material from the coarse mode that contaminates the accumulation mode equals the mass of condensed material from the accumulation mode that is collected with the coarse material. Decreasing that contamination is almost certainly more important than maximizing the capture of the accumulation mode. Of course, from the viewpoint of the enforcers, it is always bad to decrease the collected mass, at any price. We can safely ignore the fact that no commercial PM, » sampler presently exists; if there is a clear indication that this will be specified, the highly competitive field of sampler development will provide several in ample time to acquire the samplers and get the data.
If it is necessary to refrigerate samples to protect their integrity, then a sampler that heats the intake airstream cannot be assumed to collect a valid sample for any purpose, whether for gravimetry or for subsequent analysis.
lam getting very tired of the overuse of the word “quantify.” My desk dictionary’ lists “quantitate,” defined to mean precisely what is connoted by “quantify” in this paper, but it never states that the two words are synonymous. When I studied chemistry, we learned not how to quantify iron, but how to determine iron (or whatever). It is a good phrase; it is time that the chemists started using it again.
There are some errors of fact, for which I offer some corrections below. My primary client at the moment is Wedding & Associates, Inc., and this gives me access to their products (and those of some competitors) to clarify some information in the Critical Review. While this may be taken to imply that I have a vested interest in promoting my client’s products, I believe that the consultant who is seriously prejudiced by his employment soon loses all credibility.
The Critical Review does not discuss oversampling by the Andersen high volume inlet® at high wind speeds, which was shown to fit exactly the oversampling curve predicted by Davies.? The EPA declined to make performance at this speed a part of acceptance testing because its wind tunnel could not achieve it. With vortex tubes lined with oil-saturated porous material, Wedding inlets don’t need “greasing,” and need re-oiling only every few years. Other clarifications include:
¢ InTable 3, the Wedding Intermediate Flow inlet does have a cleaning port on its top, and comes supplied with a “bottle brush” to use in cleaning it.
¢ In Table 4, the “Millipore Polyvinyl Chloride membrane” is characterized as a “thin membrane of cellulose nitrate.” This appears to be in error.
e In Table 6, the Constant Flow Controller for an Aerosol Sampler using a Filter (CFCASF, U.S. Patent #5,317,930) method should be added. The CFCASF feeds back the pressure drop across the filter to the
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
flow-controlling orifice. Flow is constant within a few percent over an extremely large range of filter pressure drops. This permits the use of a wide variety of filters, in rotation if desired, without recalibration. Since the CFCASF is an orifice rather than a critical throat, to use the Critical Review’s terminology, it cannot be used on the hivol sampler.
References
1. J.B. Wedding, M.A. Weigand, YJ. Kim, D.L. Swift, and J.P. Lodge, Jr., “A critical flow device for accurate PM,, sampling and correct indication of PM,, dosage to the thoracic region of the respiratory tract,” JAPCA 37:254-258 (1987).
2. ‘J.P. Lodge, “A rational basis for National Ambient Air Quality Standards for particulate matter,” Aerosol Sci. Technol. 5:459-468 (1986).
3. USE.P.A., “Revisions to the National Ambient Air Quality Standards for particulate matter,” 52 FR 24634-24669 (July 1, 1987).
4. J.B. Wedding, M.A. Weigand, and J.P. Lodge, Jr., “A thoracic particle inlet (D,, = 6 um, D, = 10 wm) for the high-volume sampler,” Atmos. Environ, 17:1203-1204 (1983).
5. K.T. Whitby, personal communication, 1980.
6. S.M. Greenfield, “Rain scavenging of radioactive particulate matter from the atmosphere,” J. Meteor., 14, 115-125 (1957).
7. KC. Mish, ed., Webster’s Ninth New Collegiate Dictionary, MerriamWebster, Inc., Springfield, MA (1984).
8. J.B. Wedding, M.A. Weigand, and YJ. Kim, “Evaluation of the Sierra-Andersen 10 ym inlet for the high-volume sampler,” Atmos. Environ., 19:539-542 (1985).
9. C.N. Davies, “The sampling of aerosol,” Staub - Reinhalt, Luft (English Ed.), 28 (6]:1-9 (1968).
Dr. Russell W. Wiener Aerosol Physicist Mr. Frank F McElroy Environmental Engineer National Exposure Research Laboratory U.S, Environmental Protection Agency Research Triangle Park, NC 27711
Although the research information described in these comments has been funded by the United States Environmental Protection Agency, the comments have not been subjected to Agency review and therefore do not necessarily reflect the views of the Agency and no official endorsement should be inferred. Mention of trade names or commercial products does not constitute endorsement or recommendation for use.
We find the 1995 Critical Review to be comprehensive and well written. It is our contention, however, that when pertinent issues are considered, a new Federal Reference Method for fine particulate matter (PMgn-) Should employ a sampler that is specified primarily by design, not by performance. This position is in disagreement with the first recommendation of the Review.
The Critical Review intends to identify issues related to measurement methods for determining compliance with standards. These issues should include political and economic realities that present substantial limitations. Using the PM,y standard as an example, several performance specifications
Volume 45 September 1995
1995 Critical Review Discussion
were set forth. For example, measurement precision had a performance requirement of +5 g/m: for concentrations less than 80 pg/m3 and +7% for concentrations over 80 g/m}. Equivalent method samplers had similar requirements along with a requirement for a regression slope of 1+0.1 and a correlation of 0.97 or higher when compared with a collocated reference method. In practice, these specifications have not been achieved, as the Review acknowledges.
The Critical Review concludes that: “Compliance monitoring data are used for more purposes than determining whether or not an ambient concentration is higher than or lower than a given number. They are used to apportion excessive particulate concentrations to their sources, to monitor the effectiveness of control strategies, and to determine relationships between air pollution and health. All these uses need to be considered when measurement methods are selected.” The fraction of all samples utilized for purposes beyond the determination of compliance is extremely small. Better results are obtained with sampling devices designed for a specific purpose than with devices designed for a large number of possible purposes.
The Review also notes that: “The performance standard for PM,) compliance monitors has allowed and encouraged the development of new measurement methods to meet specific needs....” We agree, but there are mechanisms to encourage sampler development other than a performance standard, and such incentives should not be part of compliance monitoring. The sole criteria for a compliance monitoring should be based on the efficacy of the field data for that purpose. If we wish to support epidemiological studies, the monitoring network should be tailored to the objectives of those studies.
The Review finds that “Collocated measurement studies often show differences between sampling systems of similar as well as different design. Inlet maintenance, filter handling and storage, laboratory analysis, and operating procedures are just as important variables as the sampler design in explaining these differences, Inlet characteristics and particle volatilization properties are the most important variables that cause mass concentration to differ.” We strongly agree. It is extremely important to develop user-friendly instruments and to provide adequate quality assurance, quality control, and rigorous standard operating procedures for all aspects of sampling and analytical work.
We disagree with the first recommendation for a performance standard for compliance monitors. No one has yet proposed a significant definition for performance characteristics (other than comparison to a reference method). Further, funding is not in place to properly evaluate instruments against some as-yet undefined performance standard. Though the Review advocates definition of relevant properties of the measured particles and reasonable tolerances for their measurement, this definition is lacking. One
Journal of the Air & Waste Management Association 673
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
cannot adequately define sampling performance parameters and instrumentation when one hasn’t defined what one is sampling for. The Review states that “Newly-developed methods that show better performance should not be constrained to showing equivalence with a lessperfect method embodied in a design standard.” However, if sampler performance has not been defined, then “better” performance is also indeterminate. The Review suggests that “A revised performance standard should be more specific with respect to conditions that minimize particle volatilization between sampling and analysis, in addition to the inlet, filter, and flow rate specifications in the current performance standard.” We agree, though this moves in the direction of a design standard, not a performance standard; set-
Wedding High Volume PM,, Sampler
50 100 150
‘Secondary
ting operational parameters is not performance.
We do not believe that the fifth recommendation for chemical characterization capability is essential. Compliance monitors should be the best monitors available to measure the quantity specified by the standard; other uses are clearly secondary. With respect to the seventh recommendation for research on continuous monitoring methods for major chemical components, we definitely agree and encourage the aerosol community to keep emphasizing this need.
Table 1 summarizes comparison statistics from collocated PM, o Mass concentrations used in compliance monitoring networks that reported to the EPA AIRS data base for the years 1991 through 1993. The first three rows in Table 1 correspond to collocated reference samplers of the same design while the final five rows show similar results from reference samplers of different design.
As shown in Table 1, collocated precisions for monitors of the same make and model range from about 10% to 14%, which doesn’t quite meet the +7% specified by the performance
Figure 1. Comparison of PM,, measurements acquired with two collocated Wedding high volume samplers. Data are from compliance monitoring sites reporting to the EPA AIRS network for 1991 through 1993.
standard, but is considered generally acceptable for compliance monitoring. The 95% probability limits range from about -29% to +26%. For monitors of different make and model, the agreement between collocated measurements is poorer, with precisions ranging from about 15% to 26%, with the 95% probability limits from -67% to about +82%.
Figures 1 and 2 show some of these measurements as scatterplots. Figure 1 compares two like collocated samplers, in this case Wedding high volume samplers, showing a regression slope within 10% of 1,0. Figure 2 compares dissimilar samplers, in this case Wedding and Sierra-Andersen high volume samplers, with obviously poor correlation. Figure 3 illustrates a case where there is very good agreement between dissimilar samplers. The key difference for this case is that the measurements were taken in a very carefully controlled research situation, not a compliance monitoring
Table 1. Comparison statistics for collocated samplers in compliance monitoring networks.
Lower Upper Bias 95% 95% (Avg. % Precision Prob. Prob. Method 1 Method 2 Number Difference) (CV, %) (%) (%) Same Sampler Wedding Wedding 1440 -0.037 11.2 -22.6 21.3 SA1200 SA1200 2757 -1.30 14.2 -29.1 26.5 SA321B SA321B 2330 0.327 9.68 -18.7 19.3 Different Sampler SA1200 Wedding 223 35.0 23.7 -11.4 81.4 SA1200 SA321B 10 44.9 19.1 75 82.3 SA321B Wedding 246 -1.63 15.6 -32.2 28.9 SA231B SA1200 28 -16.4 26.3 -67.8 35.1 SA321B SA246B 153 29.6 23.9 -18.2 75.5
674 Journal of the Air & Waste Management Association
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
We have examined several samplers that
SA 1200
might achieve these goals. A commerciallyavailable sampler that might form a good basis for the reference method is the Rupprecht and Patashnick Company’s Partisol® sampler; this sampler would have to be modified and adapted as necessary to meet the specific FRM needs. Other samplers are also under consideration.
EPA should provide for designation of alternative samplers or measurement techniques as equivalent methods, based on a perfor-
Wedding
mance comparison to a reference method sampler. Such equivalent methods would also allow for sequential samplers to meet the need for daily PM, ; samples without daily chang-
sites reporting to the EPA AIRS network for 1991 through 1993.
Figure 2. Comparison of PM,, measurements acquired with an Andersen SA1200 and a Wedding collocated high volume samplers. Data are from compliance monitoring
ing of the filter. Equivalent method requirements or specifications have not yet been determined. We would expect quite modest
S/A PM,, vs. Wedding
Birmingham, Alabama: December 1990
requirements for a sequential sampler that was otherwise very similar to the reference method sampler, while the full equivalent method requirements would apply to other samplers or alternative instruments. Equivalence require- + ments should include wind tunnel tests for
| inlet sampling dynamics, static laboratory tests for flow audits, reliability, and filter equilibration, and field comparisons of mass concentration, flow rates, and possibly chemical
concentrations. A Squared = 0.98 Slope 16 Intercept 94 Dr. Michael T. Kleinman
0 20 40 60 80 100 120 140 Wedding PM,, (g/m)
Community and Environmental Medicine University of California at Irvine Irvine, CA 92717
Figure 3. Comparison of PM,, measurements acquired with an Andersen SA1200 and a Wedding collocated high volume sampler under highly controlled research condi-
tions. Though precision is good, a systematic bias still exists.
situation, emphasizing the need for effective quality assurance and quality control.
Mr. Frank has enumerated the goals and characteristics for a Federal Reference Method. these include appropriate quality assurance requirements for both reference and equivalent methods. Weighing all the considerations, we believe the reference method should be a mass based, gravimettic standard and that it should specify a single reference method sampler that is defined primarily by design. The sampler design should be published in the Code of Federal Regulations and freely available to any manufacturer. There must be an on-going quality assurance requirement for precision and accuracy to maintain certification of the samplers.
Volume 45 September 1995
Figure 2 in the Critical Review clearly shows the particle sizes that deposit in the lung, with peaks in the 0.02 to 0.04 um (ultrafine) and 2 to 4 um regions. To determine the amount of particles that is actually likely to deposit in the lung, however, one must take the convolution of the ambient particle size distribution in Figure 1 of the Review with the lung deposition curve in Figure 2. This allows one to determine what fraction of the currently-measured PM,, that deposits in the lung would be measured by other size fractions. From Figures 1 and 2, a PM, ; fraction would include ~70% of the PM ,o depositing in the lung and a PM, size fraction would include ~50% of the deposited PM. Most of this difference is due to the larger 2 to 4 um peak in Figure 2. I am not advocating either a PM, , or a PM, 9 size fraction because, in reviewing the health base literature, it is very difficult to find overwhelming evidence that fine particles are entirely
Journal of the Air & Waste Management Association 675
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
involved in all of the health effects. Nevertheless, there is quite a bit of action in terms of lung deposition in the 1 to 3 ym size range, even though this is a minimum in most ambient size distributions. Figure 2 of the Review also shows deposition in the nose and upper respiratory tract that includes particles as large and larger than 10 um. We should certainly maintain the PM,, standard in some form, as well as the long-term monitoring that characterizes human exposure.
From my perspective as an atmospheric chemist, it would be beneficial to design a sampler that separates the fine and the coarse mode particles. The fine mode particles tend to be more acidic, or partially neutralized salts of acids and metal ions. The coarse fraction tends to consist of metal oxides which are basic. When the fine and coarse fractions are mixed, the basic particles neutralize the acidic particles and the actual characteristics of the suspended particles are obscured.
I agree that samples should be chemically characterized on a regular basis. Certainly acidity, sulfate, and nitrate concentrations should be measured, as should some of the transition metal ions that have been implicated in health effects. These data would be extremely valuable, both for identifying the sources that we could potentially regulate, as well as providing a basis for identifying the causal agents that are involved in health effects.
Certain of the ultrafine particles apparently possess the characteristics of semiconductors. When deposited in a thin liquid layer, the ultrafine particles can generate free radicals. This has recently been hypothesized as an additional toxicity mechanism that merits further investigation.
Particles suspended in air are complex, and if we oversimplify the particle measurement problem by setting rigid design standards, we may miss important things. We need to encourage and allow the scientific community to be creative in developing and applying novel measurement methods and relating the results of those methods to health effects.
FLOOR DISCUSSION
Dr. Peter K. Mueller Air Quality Project Manager Electric Power Research Institute 3412 Hillview Ave. Palo Alto, CA 94304
The 1995 Critical Review is an encyclopedic tour de force on the technology of characterizing particles suspended in the ambient atmosphere relevant to situations influenced strongly by emissions from human activities. The review provides an extraordinary array of lookup tables ranging from the attributes of particles to the findings from comparative performance evaluations of sampling devices. This review is focused primarily on monitoring to determine the extent of compliance with air quality standards
676 Journal of the Air & Waste Management Association
because the standard is stated in terms of particles separated from the ambient aerosol. All of the current reference and equivalent methods depend on separating particles from the ambient aerosol. Therefore, most of Dr. Chow’s text is devoted to the methodological alternatives associated with such an approach.
When particulate matter standards were first set, concentrations in major urban and industrial centers were so high that the catch of particles separated from the air was an ample parameter, as was the concentration of such material not to be exceeded. Now that the application of this parameter has generally been successful in reducing directly emitted particles and the emissions of precursors leading to the formation of secondary particles, the classical parameters have become less relevant. Therefore, now armed with the knowledge so elegantly presented with reference to compliance sampling, I wish that a critical review could be prepared on the science of atmospheric aerosols. Such a review would specify what needs to be known about them so that emissions stemming from human activities can be managed to maintain an earthly atmosphere that is societally and environmentally acceptable.
Particles in the atmosphere play many beneficial as well as detrimental roles. From here on out, we must think about these before we set about reconsidering the current, somewhat anachronistic standards. Instead of a standard, the resultant science should drive the observational technology of the future.
The roles particles play in the atmosphere are generally well understood. They influence radiative transfer. In the horizontal that means visibility or atmospheric clarity, and in the vertical it means radiation balance, i.e. climate and weather, albeit very much less so than latitude and elevation. But what is it that influences radiative transfer? It is the particles in their suspended state.
When particles are separated from their aerosol by the so-called reference methods, many of them are adulterated. They are also adulterated by suction into the sensing chambers of continuous instruments, as well as by temperature change the aerosol typically undergoes in such instruments. Such adulteration not only confuses understanding radiative transfer influences; it also misleads the understanding of how respired particles react within airways prior to deposition on target tissues.
Atmospheric aerosol science has developed in recent years to the point of realization that a substantial fraction of the particulate material is continually in transition with the gas phase. With an understanding of the underlying reactions and thermodynamics, it has been demonstrated that it is now practical to derive the nature of the suspended particles at the location and time of sampling.
I maintain that the well-developed culture of particulate matter sampling and analysis, so thoroughly characterized by Dr. Chow, must be revamped. We must create and accept
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
a methodology for characterizing particles suspended in the atmosphere by coupling observation with derivation based on proven and testable scientific principles. I look forward to the time when air quality management practices will apply this concept. When that time comes, the current inconsistencies in setting, testing, and meeting standards and identifying sources will be resolved.
Mr. Alfred C. Schmidt MS, PE Schmidt Instrument Co. PO Box 111 San Carlos, CA 94070
Pump pulsations may also be responsible for causing mass concentration measurements to differ in collocated studies with different kinds of air samplers, and even with the same kinds of air samplers. Piston pumps and diaphragm pumps produce sinusoidal inlet velocity patterns that are bound to affect the performance of size-selective inlets. These pulsations may be observed by holding one’s hand over the pump exhaust, by installing an undamped vacuum gauge pressure sensor on the downstream side of the sampling filter, or by installing an electronic pressure sensor on the downstream side of the sampling filter. Until more is known about the effect of high frequency velocity variations on the cut point, I recommend that the amplitude of these variations be limited to the acceptable variation in the average velocity.
With respect to the fifth recommendation on the need for chemical characterization, I believe that compliance monitors should also be capable of determining radioactivity in the collected particles. For alpha and beta emitters this will require the use of surface collection (membrane) filters, that have more flow resistance than depth filters, and consequently require better flow control systems to keep the incoming flow constant.
On the sixth recommendation for an objective evaluation of trade-offs in network design, I believe that higher flow rates should be considered. I have found that a flow rate of 40 L/min works well with most 47 mm and 50 mm diameter membrane filters, and also works well with available 4-vane rotary vacuum pumps. Raising low-volume sampler flow rates from 16.7 to 40 L/min will result in larger, more easily weighed samples, and will help to offset the loss of weight from the coarser particles that will be excluded by new size-selective inlets.
Under “Filter Media and Filter Holders” in the main review, something should be said about the importance of leak testing filter holders before putting them into service. Vacuum leaks are not visible like water leaks, but they can have a large effect on reported particle concentrations. One kind of leak test uses an impervious membrane in place of the filter, and measures whether any flow gets through. Another kind of leak test measures the pressure drop across
Volume 45 September 1995
1995 Critical Review Discussion
the sampling filter and determines whether it is normal for the flow rate. This requires that the air mover be equipped with a vacuum gauge for indicating the filter load (which many commercial air movers regrettably do not have). Under “Flow Control” in Table 6 of the Review, it should be noted that critical orifices and critical venturi tubes are very sensitive to upstream filter loads. This is due to the pressure drop across the filter changing the density of the air in the orifice throat, without affecting the velocity of sound, which remains constant. Thus there is a drop in flow of 1% for each 4.1 in of water pressure drop ahead of a critical orifice or critical venturi, which corresponds to a 3.3% drop in flow for each 1 in Hg pressure drop. Also, Table 6 omitted the “constant vacuum” flow control method,! which is one of the oldest and still one of the best flow control methods in existence. There have been many improvements since 1955.
Reference
1. A.C, Schmidt and L.L. Wiltshire “A constant flow suction unit for aerosol sampling work,” AIHA Quarterly June, 1955).
Dr. Frederick W. Lipfert Environmental Consultant 23 Carll Ct. Northport, NY 11768
Dr. Chow is to be commended for the thoroughness and breadth of her Critical Review of PM measurement methods for particulate matter (PM).! Her bibliography is one of the most complete I have ever seen and constitutes a major accomplishment in its own right. I would like to offer comments on two relatively narrow aspects of the topic: 1) the degree to which aspects of human health effects interact with measurement methods; and 2) the need for recommendations on filter media and on specific methods to account for positive and negative filter artifacts.
I was surprised to see this review of measurement methods venture into the domain of human health effects, which has its own set of problems and issues, some of which are associated with measurements of air quality. This linkage raises a number of problems, including the speculation that was offered as to probable forms of revised NAAQS for PM (long-term standard for PM, , short-term standard for PM,.3). Such speculation was premature at the time that the Review was prepared, since the 1995 draft Criteria Document had not yet appeared and is still in a “do not quote or cite” status as these comments are being prepared. My reading of the recent epidemiological literature provides no support for Dr. Chow’s recommendation for an upper particle size of 2 to 3 um. Further, there seems to be some confusion between the critical need for more and better epidemiological research data and the potential need to regulate specific size fractions of the aerosol.
Journal of the Air & Waste Management Association 677
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
Much of the recent epidemiological literature on PM has involved analyses of daily time-series. Only a few of these have involved comprehensive analyses of multiple pollutants and competing effects. Only one study? specifically tested the hypothesis that particle size is a factor in premature mortality on a daily basis, and it found no significant difference between PMjo, PM, ;, and the difference between the two size fractions. Further, the elements associated with the coarse fraction seemed to be more strongly associated with daily mortality than those associated with the fine fraction (SO,?, H*). Lipfert and Wyzga? analyzed 25 short-term mortality studies and concluded that the mean effects on mortality were essentially the same for TSP, PM,o, the optical PM measures (COH, smoke, etc.), and NO,. The mean effects on mortality for PM,5, SO,2, SO,, and CO were slightly less, and that for O, was lower still. The implication here is that epidemiology cannot yet distinguish between the acute effects of particles and those of gases, let alone between different types of particles. Of course, in a heterogeneous population it is possible that all types of air pollution share some portions of the blame. Similar ambiguities have been noted in studies of hospital admissions, together with weather variables.*5
Long-term effects have been estimated by comparing the health statuses of different places in relation to their environmental quality. A major step forward in this regard was accomplished recently®’ by analyzing the survival of identified individuals rather than the community-wide mortality rates previously used by many studies.8.° However, all of the long-term studies have been forced to use communitywide air quality averages rather than individual exposures, so that they all suffer from the same kinds of uncertainties with regard to actual exposures.!° Lipfert and Wyzga3 also compared this group of studies, and found no compelling support for any particular PM index. Although one& of the studies had data on several different pollutants, the data were for only six different locations and the most polluted location had the highest values for several different pollutants, which precludes robust conclusions about the “responsible” pollutant. The other study’ analyzed a limited set of PM, data and a more complete set of SO, data; larger particles were not evaluated. The SO,? data were obtained with glass-fiber filters, so that their reported mean value was about double the true value.!! Lipfert? found TSP (multi-station averages for each city) to be a better predictor of mortality due to all diseases and due to respiratory diseases than PM,;, PM, 5, or SO,2. None of the long-term studies attempted to evaluate multiple pollutant models, and thus the “best” air pollution index of health effects is also poorly defined on a long-term basis.
The view that PM. Should be a better index of health effects than, say, PM, 9, is widely held in the environmental community and deserves further comment. Fine particles are known to penetrate deeper into the lung, may be coated
678 Journal of the Air & Waste Management Association
with toxic compounds, and also appear to penetrate more readily into indoor environments. This is the conventional wisdom and the apparent basis for Dr. Chow’s second recommendation. However, the conventional wisdom might be expected to propose a long-term standard for fine rather than coarse particles, since their expected mode of action in the deep lung is more characteristic of chronic than of immediate responses. Similarly, one might be able to argue that a broader based index such as PM,) might be more appropriate for acute responses, since they may involve loss of homeostasis rather than deep lung pathology per se.5 However, PM, and PM, are often highly correlated and it remains to be seen if the additional metric is really needed from a regulatory perspective. Moreover, given the more uniform spatial distribution of fine particles in the ambient, they should actually perform better than coarse particles as measures of community-wide air quality,!2 even if the actual effects on health were the same. However, this hypothesis and thus the conventional wisdom, is not supported by the epidemiology.
The statistical problem of partitioning the effects of two or more collinear pollutants is not always straightforward.!3 We are dealing with extremely weak effects on health, and the relationships between pollutants (and weather variables) are often much stronger than those between pollution and health. Some authors have elected to use relative statistical significance as a figure of merit in assigning relative potency to collinear pollutants, a choice that is not always appropriate. The precision of measurement and the relative degree to which fixed-site, outdoor, measurements can serve as a surrogate for actual exposure can have a large effect on statistical significance, especially in multi-pollutant models. Numerical simulations! have shown that the pollutant with the lower measurement error always dominates in a twopollutant model, even when the underlying health relationships are identical.
Thus, particle size has statistical as well as health implications in epidemiological studies. The properties of particles which allow them to penetrate deeper into the lung and into buildings also give them a statistical advantage over coarser particles when used as a community-wide air quality index. Such indices have been used in all of the recent epidemiological studies, but the fact that PM,, and SO,2have actually performed no better (within statistical limitations) and often worse2.!5 than other air quality measures should thus be interpreted as demonstrating an underlying relationship with health that may actually be weaker than those of the other measures.
An additional method of evaluating the potency of different particle size fractions may be attempted by subtracting independent measurements (PM; - PM, ;, for example). Here again, statistical variability rears its ugly head: when two variables of similar magnitudes are subtracted, the variance of the difference is inevitably larger than either of the
Volume 45 September 1995.
Downloaded by [66.67.115.93] at 07:08 26 November 2015
originals. Thus, we would expect such a differential to display an inferior statistical relationship with health. (However, Lipfert? found TSP - SO,2 (mass) to be a better predictor of COPD mortality than TSP alone.) Thus, there appears to be no justification for a new NAAQS for fine particles based on the current epidemiological literature.
Dr. Chow’s final recommendation for “research . . . on continuous measurement of major chemicals in suspended particles . . .” can be supported by all researchers in this field, without reservation. I would add “by particle size” in lieu of her second recommendation on adoption of a 2-3 um size fraction for compliance monitoring. The lack of an adequate database for epidemiological purposes is a serious limitation to developing an improved understanding of the underlying relationships. PM is usually measured only every six days, and at least some of the agencies that determine SO,” are still using glass-fiber filters. Air quality monitors are often sited for individual source compliance purposes rather than to determine population exposures. Some of the high quality data sets that have figured prominently in the literature are privately held and have been unavailable for replication or further exploratory studies. It would thus appear that a more defensible recommendation would be to improve the publicly accessible epidemiological data base first and to consider regulation later.
Although the Critical Review listed a number of references on the determination of filtration artifacts, mainly on the glass-fiber filters used in high volume samplers, it did not attempt to quantify these effects or discuss their ramifications when such data are used in epidemiological analyses. Both of these topics are still relevant because some agencies continue to measure and report suspended sulfate concentrations on this basis. In addition, the review did not specifically identify the loss of nitrates from Teflon (PM,) filters and the apparent lack of a satisfactory routine method for measuring the nitrate component of urban aerosols. Specific recommendations on filter media per se are needed. Obviously, the lack of an adequate method for the measurement of nitrate particles precludes any epidemiological determination as to their effects.
Acknowledgments
The research supporting these comments was supported by the Electric Power Research Institute under RP 3253. However, the comments represent the opinions of the author and should not be attributed to any institution or agency.
References
1. J.C. Chow, “Methods to determine compliance with ambient air quality standards for suspended particles,” J. Air & Waste Manage. Assoc. 45:320-382 (1995).
2. D.W. Dockery, J. Schwartz, and J.D. Spengler, “Air pollution and daily mortality: associations with particulates and acid aerosols,” Envir, Res. 59:362-373 (1992).
3. BW. Lipfert and R.E. Wyzga, “Air pollution and mortality: issues and uncertainties,” Paper 95-MP19.03, presented at the annual meeting of the Air & Waste Management Association, San Antonio, TX (June 1995).
Voiume 45 September 1995
1995 Critical Review Discussion
al
EW. Lipfert, “A critical review of studies of the association between demands for hospital services and air pollution,” Environmental Health Perspectives Suppl. 101 (Suppl 2) pp. 229-268 (1993).
EW. Lipfert, Air Pollution and Community Health, Van Nostrand
Reinhold, New York (1994).
6. D.W. Dockery, C.A. Pope, Ill, X. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, B.G. Ferris, Jr, and FE. Speizer, “An association between air pollution and mortality in six U.S. cities,” N. Engl. J. Med. 329:17531759 (1993).
7. C.A. Pope Ill, MJ. Thun, M.M. Namboordiri, D.W. Dockery, and J.S. Evans, “Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults,” Am. J. Resp. Crit. Care Med. 151:669674 (1995).
8. _H. Ozkaynak and G.D. Thurston, “Associations between 1980 U.S. mortality rates and alternative measures of airborne particle concentration,” Risk Anal. 7:449-461 (1987).
9. EW. Lipfert, “Community air pollution and mortality: analysis of 1980 data from U.S. metropolitan areas I. particulate air pollution,” BNL 48446-R, prepared for U.S. Department of Energy, Brookhaven National Laboratory (August, 1993).
10. EW. Lipfert, “Estimating air pollution-mortality risks from crosssectional studies: prospective vs. ecologic study designs,” presented at the AWMA Symposium on Particulate matter: Health and Regulatory Issues, Pittsburgh, PA (April 1995).
11. EW. Lipfert, “Filter artifacts associated with particulate measurements: recent evidence and effects on statistical relationships,” Atmos. Environ. 28:3233-49 (1994).
12. EW. Lipfert and S.C. Morris, “Air pollution and mortality,” Science 255:606 (1991) and Science 256:722 (1992) (letters).
13. P.L. Kinney and H. Ozkaynak, “Associations of daily mortality and air pollution in Los Angeles County,” Env.Res. 54:99-120 (1991).
14. RW. Lipfert and R.E. Wyzga, “Uncertainties in identifying ‘responsible’ pollutants in observational epidemiology studies,” Inhalation Toxicology 7:67 1-89 (1995).
15. H. Ozkaynak, J. Xue, P. Severance, R. Burnett, and M. Raizenne,
“Associations between daily mortality, ozone, and particulate air
pollution in Toronto, Canada,” presented at the Colloquium on Par-
ticulate Air Pollution and Human Mortality and Morbidity, Irvine,
CA, January 24-5, 1994).
oad
Dr. Richard J. Thompson Executive Secretary Intersociety Committee on Methods of Air Sampling and Analysis PO Box 101 Harwood, TX 78632
Dr. Chow's perspective is that of a researcher at the cutting edge of developmental work in particle collection and classification. From this vista, Dr. Chow has delivered a work of awesome dimensions. The Review addresses sampling and analysis methods for characterizing atmospheric aerosols of varying diameters. The Review does not consider two important issues: 1) what is the objective of particle standards? and 2) how will particle standards achieve these objectives?
The first sentence in the introduction notes, “National Ambient Air Quality Standards were authorized to protect human health in the United States.” Actually, the Clean Air Act says “human health and welfare.” Consequently, there are primary (human health) and secondary (welfare) standards. The secondary standards have been designed to pro-
tect livestock and vegetation, and to lessen decreased visibility.
We do not know what, if any, specific size fractions of particulate matter in outdoor air produce definite, specific adverse health effects in humans. We do have evidence that
Journal of the Air & Waste Management Association 679
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
certain moieties, at certain sites, can cause undesirable effects to animals and vegetation.
My background includes direction of the group in EPA (and in predecessor agencies) that obtained the data upon which the NAAQS are based. Although we conducted special studies, and devised sampling and analytical methodologies germane to data delivery and quality, we were not funded nor “chartered” to do research. The monitoring data was not designed for, nor intended to, bear on the topic of adverse health effects to humans from exposure to ambient (outdoor) air and/or its constituents. Unfortunately, these data were often the only data of national scope existent. As Dr. Chow notes, these data were used to extract health implications beyond their merit for this purpose. Given this background, I offer my own recommendations for monitoring to determine compliance with particulate air quality standards.
My first recommendation is that the primary standard be based on the most reliable procedures available, and that size fractions should not be exactly specified. We do not know with assurance what size fraction, if any, of outdoor airborne particulate matter can cause adverse effects to human health or damage to plants and animals. We should at least acquire information with proven methods that have minimum uncertainties.
My second recommendation is that some measured values that exceed standards need to be discounted when evaluating compliance. Particulate concentrations attain high concentrations when wind speeds are very fast, owing to windblown dust. Accommodations for uncontrollable wind erosion conditions should be made when using measured concentrations for determining compliance. I was a professor at Texas Tech in Lubbock, TX, prior to joining the National Center for Air Pollution Control, and I assure you that Lubbock will always be a non-attainment area if a fixed gravimetric standard is strictly applied.
My third recommendation is that a national secondary standard be imposed on the “fine” fraction. Whatever method that affords a size fraction below 2 1m and yields accurate and precise data should be specified in the reference method. The objective of this secondary standard is to minimize the further decrease of visibility.
My fourth and final recommendation is that epidemiological studies be conducted in the ten counties and/or parishes with the most prevalent cases of disease that might be caused by excessive particulate loadings. Chemical analyses of sized fractions of particulate matter, and monitoring of airborne organics, may permit identification of causative agents. Samplers that acquire sized particle fractions of a gram a day have been used for health effects studies. This sample size permits biological testing and extensive chemical characterization of these particles. Using this approach, EPA researchers have found mutagens in diesel exhaust.
680 Journal of the Air & Waste Management Association
Dr. Purnendu K. “Sandy” Dasgupta Paul Whitfield Horn Professor Department of Chemistry and Biochemistry Texas Tech University Lubbock, TX 79409
This brings to your attention some work that we have cartied out! towards the automated sensitive determination of soluble gases. In particular, Simon and Dasgupta?4 show how soluble particles can be determined in near real time.
It is clear that the basis for changing from PM,, to PM, 5 lies in inhalation toxicology. It is also clear that based on technological or economic grounds, we are not likely to have a regulatory measure that will provide us with a complete chemical composition of the aerosol. This being the case, perhaps we should think about chemicals that are rapidly leached from the aerosol, including net acidity and materials that dissolve in a specified period of time that mimics pulmonary fluids. I have recently proposed building an aerosol protein analyzer that would help to distinguish between chemical and biological air pollution effects in an indoor environment.
References
1. Z. Vecera and P.K. Dazgupta, “Measurement of atmospheric nitric and nitrous acids with a wet effluent diffusion denuder and lowpressure ion chromatography-postcolumn reaction detection,” Anal. Chem, 63:2210-2216 (1991).
2. PK. Simon and P.K. Dasgupta, “Wet effluent denuder coupled liquid/ion chromatography systems: annular and parallel plate denuders,” Anal. Chem. 65:1134-1139 (1993).
3. P.K. Simon and P.K. Dasgupta, “Continuous automated measurement of the soluble fraction of atmospheric particulate matter,” Anal. Chem. 67:71-78 (1995).
4. PK. Simon and PK. Dasgupta, “Continuous automated measurement of gaseous nitrous and nitric acids and particulate nitrite and nitrate,” Environ. Sci. Technol. 29:1534-1541 (1995).
Dr. David Marrack, M.D. Ft. Bend Medical Clinic P.O. Box 27907 Houston, TX 77277
In general, urban ambient air particulates are in perpetual flux. Particulate measurements are not made on a “steady state system.” This is an inherent problem with the measurements. The “particulate matrix” or “mass composition” may not be homogenous throughout a particle. Because of the microsponge structure of PM,, and especially PM; o, “snowflake-structures”!4 act as molecular sieves with huge23 and indeterminate surface areas per unit mass (the effective surface of activated carbon is abgut 1,000 m2/g). Extractable Organic Matter (EOM)5¢ and inorganic compounds are absorbed onto these particles. Diverse compounds occur in the EOM. Mutagens, polycyclic aromatic hydrocarbons (PAH) and nitrous compounds’4 have been identified but represent a small fraction of the EOM mass. Fifty to seventy-five
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
percent of the toxic species are in the PM, ;®!9! fraction. Seventy percent of these particles will reach the alveoli, to be taken into mobile cells and distributed about the body carrying some of their “piggy-backing” absorbed chemicals.
To understand the chemical heterogenicity of PM, both the particle matrix and the components they retain must be measured. Such real data can then be used in meaningful epidemiological studies. Lack of quantitative chemistry, not just speciation, confounds epidemiology. The molecular biology of the events that result in the observed health effects are unknown. This information is badly needed in order to form rational decisions to protect the environment and public health.
Some crude PM chemical speciation is needed to make better approximations of the relationship of PM, PM, , and acidic PM, ; to observed health effects. The weight of health studies is that acidic PM, , is associated with hospitalization excess and death. There are sound reasons to measure acidic PM, ; (not just fine particle sulfate) on many, if not all, airborne particulate samples. The PM, ; emissions from all oxidative combustors (incinerators, power utilities, and engines) should also be analyzed for EOM and mutagen content so that appropriate comparative evaluations of their probable health impacts can be made.
In response to Dr. Lipfert’s comments, it is not that the large particles of TSP have no health effects, but that their effects are much less than those of the equivalent mass of PM, ;. In response to Dr, Kleinman’s observations, it should be noted that though the fraction of TSP mass that is PM, ; mass may be low, the number of particles in the fine particle size fractions is orders of magnitude higher than the number of particles in the coarse size fraction. There may be very good biological reasons to regulate airborne particulates both by coarse particle mass and by find particle number, as well as by acidic particles, EOM mass, and mutagen content.?5 Pilot programs in several areas should be conducted using these as a basis for a later thorough analysis.
References
1. D. Michaud, M. Baril, G. Pettault, “Characterization of airborne dust from cast iron foundries by physico-chemical methods and multivariate statistical analyses,” J. Air Waste Manage. Assoc. 43:729 (1993).
2. R.G. Barton, W.D Clark, W.R Seeker, “Fate of metals in waste combustion systems,” Combust. Sci. and Tech. 74:327 (1990).
3. KS, Kumar, P.L Feldman, “Control of particulate trace metals in coalfired power plants,” Paper 93 MP-S/04, presented at the annual meeting of the Air & Waste Management Association, Denver, CO (June 1993).
4. G.A. Brinckman, “Control of particulate matter, acid gas, heavy metals emissions from medical waste incinerators,” Incineration Conf. Proc. Knoxville, TN, p. 369 (May, 1991).
5. C, Venkataramen, S.K. Friedlander, “Source resolution of fine particulate polycyclic aromatic hydrocarbons using a receptor model modified for reactivity,” J. Air & Waste Manage. Assoc. 44:103 (1994).
6. _T.R. Keyser, D.ES. Natusch, C.A. Evans, Jr, R.W. Linton, “Characterizing the surfaces of environmental particulates,” Environ. Sci. Technol. 12:68 (1978).
7. K. Yasuda, M. Takahashi, “The emissions of hazardous substances caused by medical waste incineration,” Paper 94 WA85.05 presented at the annual meeting of the Air & Waste Management Association, Cincinnati, OH, (June, 1994).
8. R. Williams, J. Meares, L. Brooks, R. Watts, P. Lemieux, “Priority pollutant PAH analysis of incinerator emission particles using HPLC and optimized fluorescence detection,” Intern. J. Environ. Anal. Chem, 54:299 (1993).
Volume 45 September 1995
1995 Critical Review Discussion
9. Zu-Wei Gu, Bao-Zheng, B.Nath, Wen-Zong Whong, W.E. Wallace, ‘Tong-Mon Ong, “Micronucleus induction and phagocytosis in mammalian cells treated with diesel emission particles,” Mutation Resch. 279,55:279 (1992).
10. D.A. Lane, D.N. Johnson, “Gas and particle phase measurements of toxic chemicals in the Great Lake Basin,” Paper 93 RP137.05 presented at the annual meeting of the Air & Waste Management Association, Denver, CO (June 1993).
11. R. Watts, B. Fitzgerald, G. Heil, H. Garabedian, R. Williams, S. Warren, L. Fradkin, J. Lewtas, “Use of bioassay methods to evaluate mutagenicity of ambient air collected near a municipal waste combustor,” JAPCA 39:1436 (1989).
12, D.M. DeMarini, J.E. Gallager, V.S. Houk, J.E. Simmons, “Toxicological evaluation of complex industrial wastes: implications for exposure assessment,” Toxicology Letters 49:199 (1989).
13. VS. Houk, D.M. Demarini, “Use of microscreen phage-induction assay to assess the genotoxicity of 14 hazardous waste industries,” Environ. and Molecular Mutagenesis, 11:13 (1988).
14. R.R. Watts, P.M. Lemieux, R.A. Grote, R.W. Lowans, L.R. Williams, S.H. Warren, D.M. Demarini, D.A. Bell, J. Lewtas, “Development of source testing, analytical and mutagenicity bioassay procedures for evaluating emissions from municipal and hospital waste combustors,” Environ. Health Perspectives, 98:227 (1992).
15. D.M. DeMarini, V.S. Houk, Hazardous waste detection, control treatment. R. Abbou, Ed. Elsevier Science Publishers, B.V. Amsterdam (1988).
16. D.M. DeMarini, R.W. Williams, L.R. Brooks, M.S. Taylor, “Use of cyanopropyl-bonded HPLC column for bioassay-directed fractionation of organic extracts from incinerator emissions,” Intern. J. Environ. Anal. Chem. 48:187 (1992).
17. MJ. Keane, $.G. Xing, S.G., J.C. Harrison, T. Ong, W.E.Wallace, “Genotoxicity of diesel exhaust particles dispersed in simulated pulmonary surfactants,” Mutation Research, 260:33 (1991).
18. _L.T. Crippitt, L.D. Claxton, L.D., T.E. Kleindienst, D.R. Smith, P.B. Shepson, “Transformation of Boise sources: the production and distribution of mutagenic compounds in wood smoke and auto exhaust,” Proc. EPA/AWMA International Symp. on Measurement of Toxic and Related Air Pollutants, VIP-10. Air & Waste Manage. Assoc., Pittsburgh: p. 85 (1988).
19. J. Lewtas, R.B. Zweidinger, L. Crippitt, “Mutagenicity, tumorgenicity and estimation of cancer risk from ambient aerosol source emissions from wood smoke and motor vehicles,” Paper 91-131-6 presented at the annual meeting of the Air & Waste Management Association, Vancouver, BC (June, 1991).
20. J. Lewtas, “Carcinogenic risks of polycyclic organic matter (POM) from selected emission sources - development of comparative potency method,” U.S. EOA HERL, Raleigh, North Carolina, Report 803 (1991).
21. P.B. Shepson, T.E. Kleindienst, E.R. Edney, “The production of mutagenic compounds as a result of urban photochemistry,” U.S. EPA Report no. 600/S3-87/020 (1987).
22. D.N. Wittoroff, A.W. Gertler, J.C. Chow, R.W. Barnard, H.A. Jongedyk, “The impact of diesel particulate emission on ambient particulate loadings,” Paper 94 WP91.01 presented at the annual meeting of the Air and Waste Management Association, Cincinnati, OH (June, 1994).
23. _E.W. Wallace, MJ. Keane, V.A. Hill, J. Xa, T. Ong, “Mutagenicity of diesel exhaust particles and oil shale particles dispersed in lecithin surfactant,” J. Toxicol. and Environ. Health, 21:163 (1987).
24. IARC, “Monographs on the evaluation of carcinogenic risks to humans, diesel and gasoline engine exhausts and some nitroarenes,” Vol. 46, IARC, Lyon, pp. 58-79 (1989).
25. E. Marrack, “All PM,, are not biologically equal,” in press (1995).
Dr. Robert A. Michaels President RAM TRAC Corporation 3100 Rosendale Rd. Schenectady, NY 12309
The Critical Review implies that particle-associated health impacts may depend on ambient conditions. Standardizing laboratory conditions for temperature and relative humidity is designed to help purify the mineral core
Journal of the Air & Waste Management Association 681
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
of particles, most notably by driving off moisture. Thus, the view that non-standard conditions might best serve as an indicator of health impacts confuses ambient particle collection conditions with sample handling conditions used to prevent contamination in the field and to achieve further purification in the laboratory.
Some would say that particles, being heterogeneous, are best represented by all of their components, including substances adsorbed to them under ambient conditions. This suggests that particle mass should be measured holistically, including water. This holistic approach overburdens the mass parameter as an indicator of particle impact. It prefers sample impurity over purity, adding undue variance to the relatively invariant mass of the particles’ mineral core. An alternative is either to measure or estimate selected adsorbed substances separately. The holistic approach may pose public health risks which can be compensated for only via an unnecessarily stringent particle regulation, at unnecessarily high cost to the regulated community.
The full toxicological potency of particles can only be gauged by analyzing their adsorbed substances. A particle measurement system that weighs particles holistically without also quantifying adsorbed metals, volatiles, semivolatiles, and acids will fail to indicate the degree of public health risk which may be posed. The EPA must recognize that particle mass alone, measured with or without adsorbed substances, is inadequate to distinguish the various toxic substances on particles from toxicologically inert water, let alone quantify them.
Health risks posed by many particle types, such as fly ash, include cardiopulmonary effects exerted by their mineral core and adsorbed acids; and lung cancer caused by metals, volatiles, and semivolatiles. The diversity of acute and chronic health risks suggests a two-pronged regulatory approach. First, an annual average would place an upper limit on particle mass, primarily to protect against lung cancer. Second, shorter-term averaging times would protect against acute toxic effects. The current PM,, regulation includes a mass limit for a 24-hour averaging time, but not for a one-hour averaging time. However, the new 24-hour value could be less stringent if combined with a one-hour value to protect against health effects, such as asthma, arising even short-term exposures.
Finally, a desired target level of public health protection can probably be achieved by less stringent mass limits at all three averaging times if analytical requirements are narrowly focused on the relatively invariant particle core measured at some standard temperature and humidity. The potential for adsorption of toxic substances may nonetheless be inferred based on knowledge of ambient conditions and particle surface areas corresponding to measured mass in the selected aerodynamic size range. Analyses to quantify adsorbed substances may also be made, but separately to preserve the accuracy of mass measurements. This is a win-win scenario, in which regulated entities can be subject to less stringent
682 Journal of the Air & Waste Management Association
particle mass limits through increased measurement accuracy associated with technologies now routinely used, while the public can still enjoy enhanced health protection.
Mr. Tom Moore PM Programs Coordinator WESTAR Council 1001 SW Fifth Ave. #1100 Portland, OR 97201
The Critical Review observes that the EPA’s “Reference Equivalent Method” PM,, sampler certification program has produced ten separate PM, sampling methods. These sampling methods build upon, share or use various sampler collection technologies and principles demonstrated to be “equivalent” through performance testing. The EPA certification program has allowed multiple Reference or Equivalent Methods to emerge by validating the equivalent performance of all the sampler models based on sideby-side test data from a few sites over short time periods. The nature of this testing format limits the range of ambient conditions under which sampler equivalency is decided.
In contrast to these certification tests, real-world NAAQS compliance monitoring applications using these approved samplers are long-term in nature, under more variable conditions. Changes in atmospheric conditions and emission sources produce unknown and unevaluated site-specific effects on flow rate, particle size fraction, and particle/gas equilbria that affect compliance monitoring data. PM,, sampler comparisons with Reference and Equivalent methods have shown great disagreement, especially at high ambient concentrations increase.
The EPA certification test format does not control the multivariate influences on data comparisons between the ten Reference or Equivalent Methods and may impair a clear analysis of cause and effect on data quality. However, the importance of inlet performance and maintenance in measuring PM, discussed in the Review needs special emphasis from an attainment determination perspective. Small variations in PM, inlet cutpoints are critical to compliance monitoring since the coarse particle mode at many sites peaks near ten mm. The importance of assuring proper maintenance of the various PM,, sampler inlets at hundreds of NAAQS compliance monitoring sites has not been emphasized in the Reference and Equivalent Method certification program or in EPA operational guidance documents. The effects of this variation cannot be evaluated quantitatively.
The undocumented effects of flow rate variability, inlet maintenance, and the resulting cutpoint variations have lead to data uncertainty and potential inequities in PM, standard nonattainment designations. Exceedances of the PM, standards resulting in a nonattainment designation due to
Volume 45 September 1995
Downloaded by [66.67.115.93] at 07:08 26 November 2015
positive sampler biases may cause unwarranted air pollution control costs, while an improper finding of attainment created from negative sampler biases may lead to inadequate public health protection.
The Critical Review points out that inlet cutpoint performance shifts may prove less critical, in general, for a one to three mm size fraction. However, long-term and widespread sampling in this size range has not generally been conducted in areas with high particulate loadings. Also, many of the state and local air quality agencies responsible for operating NAAQS compliance samplers do not have field experience with sampler operation, filter laboratory processes and interpretation complexities associated with this size range.
For these reasons, I believe that a single Reference Method with design-based specifications combined with a periodic sampler effectiveness re-certification requirement for each particulate matter NAAQS cutpoint is preferable to the performance-based approach advocated by the Critical Review. This design-based approach would more consistently and effectively balance three concerns: 1) public health data collection, analyses and exposure assessment; 2) regulatory and scientific data comparability and veracity; and 3) policy concerns of equity and effectiveness in regulatory costs.
Certification of Equivalent Method samplers using performance-based specifications could still be done on a national or regional basis, with samplers tested and certified in comparison to the single Reference Method. Certification tests for Equivalent Methods should also be combined with a “sunset clause” and a re-certification requirement to assure replacement of outmoded technologies, along with more rigorous and representative test criteria. Equivalent Method models could then be used for specific applications. These could include defining spatial and temporal variation of mass concentrations, source apportionment, pollution control evaluation, chemical characterization, and permitting.
An EPA requirement for a single design-based Reference Method sampler for each size fraction specified by the NAAQS accompanied by improved field, laboratory and data reporting protocols would better accomplish the primary mission of ambient particulate monitoring, which is to determine compliance with the NAAQS for the particulate matter indicator(s) to. assure that public health is protected.
Mr. Stanley Penkala Air Science Consultants, Inc. 347 Priestley Rd. Bridgeville, PA 15017
When we talk about a 24-hour sample, we’re talking about combining together on a filter materials that have been
Volume 45 September 1995.
1995 Critical Review Discussion
generated over a period of time. We need to recognize that this doesn’t necessarily represent what was in the air during sampling. For instance, in the Pittsburgh area we have a different aerosol collected overnight during the inversion hours than we do during the bulk of the day. Nevertheless, an “average” composition shows up the 24-hour average sample when it is analyzed. If we intend to use 24-hour samples for analysis, we need to keep in mind that this homogenization is taking place.
Dr. Ivar Tombach ENSR Consulting and Engineering 1220 Avenida Acaso Camarillo, CA 93012
Mr. McElroy’s figures comparing measurements from collocated samplers show linear regression slopes and intercepts that show positive or negative biases of one sampler with respect to another. I would argue that, using ordinary linear regressions, different conclusions about bias might be drawn if the X and Y variables were interchanged. The problem is that the ordinary linear regression approach is not the appropriate tool for comparisons in which neither variable is dependent on the other. A regression approach that minimizes the variance orthogonal to the regression line rather than the vertical distance to the regression line is needed to calculate unbiased and consistent estimates of slope and intercept. With this type of regression, the slopes will approach unity and be independent of the selection for X and Y variables. This will probably not change your overall conclusion that different sampler designs yield different values, however.
Ms. Lisa Vestman
Graduate Student
Rutgers University New Brunswick, NJ 08903
From what I have read it seems that the organic fraction of suspended particles has a lot of biological activity. I’m concerned about not characterizing particles by chemical composition that is associated with this biological activity, mutagenicity as well as immune system response.
Response from Dr. Kleinman
I fully agree that in addition to the inorganic components, we need to look at organic components. We're looking at this problem in our laboratory by exposing geriatric rats to different PM,, levels that have an organic component. We're finding, at least in our initial studies, that these older rats have a much stronger immune system response to PM,, than the younger, healthier rats that we used previously.
Journal of the Air & Waste Management Association 683
Downloaded by [66.67.115.93] at 07:08 26 November 2015
1995 Critical Review Discussion
Dr. Thomas Hauser Civil & Environmental Engineering Dept. University of Cincinnati Cincinnati, OH 45221
We debated these performance vs. design issues in the 1970s, and this led us to establish EPA's equivalency program. The equivalency program was established because the performance specifications claimed by more than thirty instrument manufacturers of sulfur dioxide monitors were not realistic. We wound up with four state agencies with monitors that simply didn’t work. We adopted a designated method against which new methods could be compared to guarantee the accuracy of the manufacturers’ specifications. This program has served us well, so let's not reinvent the wheel.
As a question to Mr. McElroy, is EPA considering the use of exposure monitors rather than fixed site monitors for determining compliance with a PM standard? The Critical Review pointed out some of the shortcomings of extrapolating from fixed site monitoring to the human exposure that actually dictates health effects. Isn’t it time to be moving in the direction of exposure monitoring?
Response from Mr. McElroy This is out of my area, but as I recall from the draft criteria document, the relationship between exposure measurements and health was not as clear as the relationships with particle concentrations from fixed sites. Part of the reason for this was attributed to poorer precision of the exposure monitors used in earlier studies.
SUMMARY
Dr. Judith C. Chow Desert Research Institute P.O. Box 60220 Reno, NV 89506
The Critical Review discussion has highlighted three issues: 1) information that was lacking from the review; 2) the design versus performance designation; and 3) the appropriate metric, in terms of particle size, chemical composition, and averaging time, for protecting public health.
First, I appreciate the additions from Mr. Schmidt, Dr. Dasgupta, and Dr. Lodge. The information in the Critical Review is continually being updated, and there will be opportunities to incorporate these and other additions into future publications. I would appreciate input from others who did not have the opportunity to respond in this summary.
Second, I believe that the design standard and method equivalency approach described by Mr. Frank and Mr. McElroy has potential to meet many of the same goals as a performance standard. The selected reference method must, however, adhere to strict performance standard that are at least as stringent as those for PM,9. Mr. Moore is correct in
684 Journal of the Air & Waste Management Association
suggesting that these performance standards need to be strictly adhered to, and that the reference method may need to be revised if it is found to fail these standards during subsequent use in a variety of environments. As pointed out in the Review, we need to guard against forcing new technologies to incorporate the mistakes of older ones.
I am also unclear about exactly what is included in the design standard. The R&P Partisol sampler was mentioned as a model, and this is certainly a well-designed instrument as are several others that are available. Does a design standard mean that every integrated circuit in its flow controller must be reproduced, or that any flow controller that has been shown to maintain a flow rate within a specified range under a variety of conditions will do? The latter performance specification makes more sense to me.
If we are to use a single sampler to determine non-attainment with the standard, with the option for subsequent sampling and analysis to determine control strategies, then we must not require that control strategies be developed around those values that determined non-compliance. I have witnessed too often a large effort put into the analyses of a few hivol filters that achieved high PM,, loadings, but that were never intended for the desired analyses. The argument is given that since the events characterized by those filters caused nonattainment, control strategies must be specific to those events. If a monitor is designed solely for compliance, then it should not be used for other purposes unless those purposes have been designed into it.
There was much discussion concerning the relationship between human health and particle measurements. Drs. Lodge and Lipfert advocate an upper size range lower than the 2 to 3 um range in my recommendations. Dr. Lipfert interprets the literature as saying that relationships between health and fine particles are not evident, while Dr. Thurston interprets the literature as showing that they are. My recommendation for particle size is based on two things: 1) the peak in the lung deposition curve shown in Figure 2 of the review; and 2) the availability of proven monitoring technology and inlets for a 2 to 3 ym size fraction. Given our past experience, I disagree with Dr. Lodge’s belief that PM, y inlets will be manufactured and tested sufficiently to meet the requirements of a 1997 standard.
Dr. Mueller has most elegantly summarized the real problem of particle measurement, as elaborated by Dr. Marrack, Dr. Michaels, Dr. Thompson, and Dr. Dasgupta. We need methods for the future that can better represent suspended particles in their natural states. I tried to summarize promising technologies in the Review, and we surely need to support those researchers who are developing those technologies. An interim goal, as noted by the discussants, is to quantify along with particulate matter samples those environmental variables (e.g., precursor gas concentrations, temperature, relative humidity) that affect the state of particles in the atmosphere and to apply appropriate chemical and physical models to infer their actual concentration, size, and composition.
Volume 45 September 1995
Publication citation
Watson, John G…. RA Michaels…, et al. Measurement methods to determine compliance with ambient air quality standards for suspended particles, 1995 Critical Review Discussion. Journal of the Air & Waste Management Association, 45:9, 666-84, doi: 10.1080/10473289.1995.10467395, September 1995;