Verbatim Responses o f expeRts the        Clarkson Airshed Expert Consultation MAY 28 2009 The Expert Consultation Background The expert consultation used the following process: In 2000, MOE began a four-part study designed to › The Inter-Agency Health and Environment Working Group invited five lead the way to improving air quality for residents distinguished academics to provide expert advice, and developed a tentative in the Clarkson area. The study approach list of questions that these experts deliberated on. combines science with community participation › The questions were sent separately and anonymously to the experts in each and support. area (air quality/human health experts, and air monitoring/ modeling experts). The experts were asked to provide a brief written response on each question, The Clarkson Airshed Study looked at how including any rewording of or additions to the questions asked. emissions from local industries, vehicles, › The responses from each group were summarized and the summary given to residences and sources outside the study area each participant for review. Each expert was then be given the opportunity to contribute to air quality in the Clarkson area. modify or expand on his/her responses, and provide final advice on each point. This included identifying major sources of target pollutants and looking at the data from ambient air monitoring and air quality modeling. List of Experts Participating Although the Clarkson Airshed Study provided Dr. Michael Jerrett, Associate Professor, UC Berkeley, Environmental Health some answers, several key questions remain: Sciences (note he has a PhD from U of Toronto, 1996 and BS from Trent University); › Are people in the Clarkson Airshed research interests include spatial analysis of disease-exposure associations experiencing adverse health effects because of using Geographic Information Science; geographic exposure modeling, land use the air quality conditions? characterization. › Who is responsible for the observed occasional Dr. Michael Brauer, Professor, School of Environmental Health and Department of high PM2.5 concentrations (and those of other Medicine/Atmospheric Science Programme, College for Interdisciplinary Studies, pollutants)? University of British Columbia. His research interests include environmental and › Was the computer simulation model appropriate occupational exposure assessment, air pollution health effects, and development of sampling and analysis methods for air pollutants. for the situation? Was it appropriately applied? › Are the principal air pollution sources really Dr. Stuart A. McKeen, Atmospheric VOC Research Group, Cooperative Institute for vehicular? Research in Environmental Sciences (CIRES)/University of Colorado and NOAA/AL, Boulder, Colorado; research interests in the modeling and monitoring of PM . 2.5 Dr. Gary Foley, Director, National Center for Environmental Research, Office of Research and Development, US EPA (Triangle Park, NC). Oversees all of EPA’s air Inter-Agency Health quality modeling activities, nationwide. Dr. Ann McMillan is currently on a special assignment with the Department of and Environment Fisheries and Oceans Canada working on aspects of Climate Change and the Arctic. Previously, she had been with Environment Canada for almost 20 years supervising air quality modeling, air quality assessments and science to policy Working Group work. Ann is the Canadian Cochair of the International Air Quality Advisory Board of the IJC. She was previously Ontario Chair of the Air Pollution Control Association The ministry established an Inter-Agency Health (now the AWMA). and Environment Working Group composed of senior members of Ministry of the Environment, Peel Region Health Dept. and Halton Region Health Dept. This Report The first task of this group was consulting with and getting solid, scientific answers from environmental This report contains the verbatim responses of the experts to the questions. experts on the adequacy of the Clarkson Airshed Although the layout and format is large, it was the only way we could provide the Study. reader with an easy to use presentation of their responses.                                                         Ogilvie, Ogilvie & Company | The independenT public FaciliTaTor First Round Questions        1How does air quality Michael Brauer Gary Foley in the “Clarkson” Airshed Although overall PM annual averages are low Nothing seems unusual but I do not have specific compare to: relative to other urban areas (in comparison to other examples that I can do a comparison with. Bert › Other urban areas with sites in Canada the concentrations are typical for Guindon at Natural Resources Canada in Ottawa has a significant industrial eastern Canada and higher than in the West; the been working with the International Air Quality Advisory sector? concentrations are on the low end of what is seen Board to do a landscape mapping of the cities on the › Other urban areas in or in the US. and generally much lower than European Great Lakes, including looking at the “urban form” nearby large volume traffic locations and far below those in Asia and factors that affect “vehicle miles travelled” and air corridors (including a busy S America) there is evidence of elevated PM pollution. Most of the cities are done. I do not know if international airport)? concentrations, especially clearly elevated 24-hr it will give insights into the Clarkson Airshed but it is › Other nearby municipalities concentrations – and exceedences of the CWS. worth asking him about it. (e.g., Mississauga, PM is not especially variable spatially so it is difficult Brampton and Oakville) to determine local influence without measurement of or similar suburban chemical composition. Still, there does appear to be “bedroom” communities clear evidence of local impacts (traffic and industrial) located close to major that are significant. Based on Clarkson Airshed transportation corridors Study (CAS), if background is assumed to be Ind and airports? East site (6.9 ug/m3), then the Ind Centre increment is 4.4 ug/m3 (64% increase) and QEW east (traffic) increment is 7.1 ug.m3 (103% increase). Given that these increments are for annual averages and are for PM which incorporates multiple components and impacts from different sources (regional, urban, local) this local scale variability is quite substantial. The windrose analysis indicates that Stn Res can be impacted by traffic sources (QEW), industrial emissions and residential combustion and some very high peak concentrations have been measured. Ozone levels can also be elevated, but these seem quite similar to other areas in the region – as expected given that ozone is regional in nature. Overall, NO concentrations are rather low, but this 2 is highly dependent upon the locations of monitoring sites. Based on the CAS, traffic hotspots (QEW) were identified. These are important from an exposure and health perspective but are not unique to this region or airshed. Some VOC levels are elevated, but in general similar are quite similar to industrial sectors of other urban areas. The CAS suggests industrial source impacts on VOC levels, but these are likely relatively confined to industrial and bordering areas, given typical fate and transport of VOCs. Note that these local sources of VOCs also contribute to ozone formation downwind of the immediate Clarkson Airshed. It should also be noted, that but overall VOC exposure typically dominated by indoor sources (furnishings, consumer products) and activities (smoking, hobbies, etc.) and not by outdoor air pollution. Although the monitoring program is very typical, note that only a relatively limited number of pollutants have actually been measured. These are best thought of as general indicators of air quality. If specific sources are of concern, monitoring of specific chemical source tracer compounds may reveal added information regarding the air quality in this airshed.    2 Michael Jerrett Stuart McKeen Ann McMillan The available information here was somewhat limited by a fairly I use TEOM PM measurements I note that the report “Air Quality in 2.5 sparse network of monitors, which generally lacked samples in from the U.S. EPA network in my Ontario – 2007” provides a comparison direct downwind impacted areas and lacked many important research/work, as well as NOx of air quality at 40 cities world-wide. It pollution constituents. Prevailing winds are from the south and measurements collected in a variety summarizes “Overall, the air quality of west, but most of the monitors operated for this study were not of settings, from clean remote to the three Ontario cities, Windsor, Toronto situated directly to the northeast of the industrial areas in the urban. The annual average PM and Ottawa, was generally better than 2.5 populated residential zones. This limits the ability of the study to concentrations for all 6 sites expect the other cities used in this analysis for assess the direct impacts of industry in the area. Stn QEW east, as well as the number the parameters compared”. This applies There was also no monitoring of PAH particle species, elemental of 30 µg/m3 24-hour exceedances, to Clarkson as a subset of Toronto. carbon, transition metals, or ultrafine particles (UFP – 0.1 of are rather typical for 119 monitors in Compared to cities such as Los micrometer diameter or less). All of these constituents could the northeast U.S. I am familiar with. Angeles and Mexico City, which both be present in elevated levels around the busy roadways and These monitors are primarily located industrial sources. As a result, my comments are limited to those in urban and suburban settings have settings that make them prone pollutants that were measured (which may not characterize for exposure monitoring. I am not to pollutionenhancing meteorological the full potential for health effects on the nearby populations) familiar with annual average NOx situations, Clarkson is pretty clean. That and more generally limited because of the sparseness of the levels. But comparing to summertime being said, the shoreline of Lake Ontario monitoring directly downwind of the industrial sources. can have “lake breeze” situations on field study measurements, these some days, most often in the summer, annual 24-hr averages, the maximum As a general note, the maps and graphics in the Part II report in which pollutants that have blown out were for the most part illegible in my copy. This limits some of levels, as well as the diurnal patterns over the lake are blown back in again my comments as well to the textual aspects of the Part II report – reported in CAS Part II, are all later in the day. These special cases some of the Part III graphics helped to fill these gaps, but I was consistent with NOx observations in resulted in Lakeview Generating Station not able to examine many of the important maps and charts in several U.S. urban areas (Boston, operating under a special system where detail. suburban New York/New Jersey, it shut down during a certain number of Nashville, Dallas, and Houston) Levels here in the CAS are comparable to other Canadian cities especially bad days to reduce pollution. having significant and diverse with near-source industry for particulate matter and NOx species. industrial emission contributions. Some of the sites had annual averages that would be above the Air quality in Clarkson has improved California Air Resources Board Annual Average standard of 12 considerably since air quality studies ug/m3 for PM . Canada wide standards of 30 ug/m3 are for only were done in the 80’s to look at the 2.5 24 hour averaging periods, which were also exceeded on several impact of the Lakeview Generating occasions. The U.S. EPA standard of 15 ug/m3, and the California Station on SO levels in the area. At 2 standard as mentioned is 12 ug/m3 for annual averages may be that time no measurements of many of more important when considering health effects. Most available the other pollutants were taken and the evidence suggests a linear dose-response in the health response modelling focussed on the single high to PM exposure so the levels observed here may be expected 2.5 stack at the generating station. to be associated with cardio-respiratory and other health effects (see below for more details). › Other urban areas with a significant industrial sector? Elevated levels of NO, NO , NOx were observed near the 2 roadways and in some areas near the industry. At ambient › Other urban areas in or nearby large levels observed here, pollutant levels are fairly characteristic volume traffic corridors (including a of levels found in other cities in North American and display busy international airport)? the distance decay pattern of lower levels away from source › Other nearby municipalities (e.g., that might be expected. It is important to note that at ambient Mississauga, Brampton and Oakville) levels, these NOx pollutants are unlikely to be the most important or similar suburban “bedroom” toxic agent, but they do associate closely with EC, UFP, and communities located close to major other pollutant constituents that are probable toxic agents at transportation corridors and airports? ambient levels – therefore these NOx species probably represent markers of potentially more harmful pollution constituents that were not measured directly in this study. Recent evidence from Toronto indicates that the levels here of NO and the elevations 2 observed at some stations may be associated with premature mortality – probably from the other toxic agents that follow similar spatiotemporal patterns as NO . 2 › Other urban areas in or nearby large volume traffic corridors (including a busy international airport)? › Other nearby municipalities (e.g., Mississauga, Brampton and Oakville) or similar suburban “bedroom” communities located close to major transportation corridors and airports? These are difficult questions to answer in general because much would depend on the actual location of the monitors in relation to emission sources and prevailing meteorology patterns. The levels of PM and NOx species appear comparable, the levels of VOCs may be higher – although there is little empirical evidence available because any the monitoring data routinely available is measured away from the source. Recent evidence implicates jets as a major source of ultrafines and VOCs and since there are no direct monitoring data near the airport that were supplied here, it is difficult to assess. Peak patterns are in line with what would be observed in these other locations during the summer and diurnally during the peak rush hours.                                                      3 Ogilvie, Ogilvie & Company | The independenT public FaciliTaTor       First Round Questions 2Is the scientific Michael Brauer Gary Foley approach of Yes, in general the approach was appropriate but The scientific approach, as described in the the Clarkson Airshed study generally clearly more detailed monitoring and sample analysis documents, appears to be well thought out and well appropriate? Is the could provide more information. Specifically, source designed. The modeling and monitoring, based on modelling and monitoring apportionment (PM and VOCs) would have been the findings and conclusions, appear to be sufficiently sufficiently rigorous to helpful to identify specific source contributions. In rigorous. The apportionment of the emission sources characterize and apportion addition, more saturation-type monitoring (as in looks good. The results, however, tell me that an even emission sources by Windsor study) would be helpful to map out hotspots more rigorous study may be warranted. contributing sector or related to specific source categories (traffic, industrial land use? point sources). Overall, there has not been a rigorous apportionment of source impacts, and a more source- oriented MONITORING approach would be desirable. The source-oriented modelling that was conducted could, as a first step, be applied to the population (not just the monitoring sites) to provide more information on population-weighted source impacts. This may provide a different picture than the current analysis.    4 Michael Jerrett Stuart McKeen Ann McMillan As I mentioned above, there are problems The measurement approach The scientific approach is generally a state-of-the-art regulatory with the sparseness of the monitoring sites within CAS Part II seems entirely approach. As such it is sufficiently rigorous to characterize and and the pollutant species measured. appropriate, giving a balanced apportion emission sources, but the accuracy of this apportionment sampling of highly polluted (Stn will be subject to debate. For the modeling, there are concerns QEW east, Stn Industrial Centre, about over-reliance on self-reported Little description is provided regarding the emissions data base, for Stn Industrial Downwind) versus emissions from industrial operators example. I assume that the emissions compiled are average and more upwind or residential monitors. who may have under-reported or used not time dependent. Does the preprocessing component introduce The measurement techniques scientifically invalid methods to estimate reasonable diurnal variations, etc? Do the emissions reflect appear sound and consistent with emissions. The model cross-validation limitations in the need to report to NPRI? the best available technology. The was done at a limited number of sites, results and conclusions based on The approach does not explore many issues that could provide so again it is difficult to tell whether the the measurements are probably insight into what’s going on, but perhaps that is beyond the scope/ model is predicting accurately. The robust, and appropriate for the budget of the current study. findings indicate the model conforms with evaluations of the sector-specific an accepted standard of performance Usually, air quality concentrations of pollutants tend to be “episodic” emissions presented in that report. +/- 2 times the observed value. Although in nature, that is, very high values occur only a few times in a year. The modeling work in CAS Part III is this is a reasonable standard for model That is why the 98th percentile is used as a standard. While there not sufficiently rigorous to be useful calibration over larger areas, it is more is controversy for some pollutants as to which exposures cause the for source apportionment, and in problematic for smaller areas and this most significant health effects, long term exposures to low levels, some cases is inconsistent with the model may lack the precision over small or shorter term exposures very high levels, only long term average analysis done in CAS Part II. For areas to accurately predict pollution concentrations are really studied here. PM in particular, transboundary 2.5 concentrations. import is assigned a single value The reason for the episodic nature of air pollution is often The finding that vehicles contribute (3.8 µg/m3) based on the difference meteorological in origin. Because of this is it is often revealing to pull the majority of the PM mass is open between annually average PM2.5 out the meteorological conditions for the “worst” pollution events 2.5 to question. Most gradient studies by from the CalPuff modeling and the over the period and work them up in more detail. From previous roadways indicate only modest elevations observations. However, the times of work that I did at Lakeview Generating Station, I know that there of PM near the sources, usually less 30.0 µg/m3 PM2.5 exceedances at are lake breeze cases along the northern shore of Lake Ontario in 2.5 several Canadian monitors in and which the wind starts off being off shore, and then swings onshore than 20%. It was not clear if the CalPuff around Toronto during the summer during the day resulting in pollution (which is not necessarily the model was predicting the large vehicular of 2004 occur during the multi-day result of long range transport) sweeping back over the area off source contribution because of secondary “smog” episodes described in CAS of the lake. These days are typically warm and rather stagnant formation of the vehicular emissions or Part II. These are periods when with a pronounced temperature inversion over the shoreline area. because of primary emissions. If it is U.S. levels, and regional-scale Lakeview did have a monitoring program that looked for these types predicting primary, the large contribution contribution to southwest Ontario, are of days and the station shut down to avoid polluting the surrounding from vehicles is out of sync with findings much higher than 4 µg/m3. The model area. If that program was still running in 2004, perhaps the data from Britain and the US indicating that source apportionment depends could be obtained from Ontario Power Generation as another vehicles contribute only between 2.5-10% critically on the accuracy of the source of data for this work. Obviously a defined source sector of the PM mass. 2.5 emission inventory, which has huge approach does not work well for these situations. Another important consideration, not uncertainties for some sectors. The Air quality models typically model “normal” air pollution days rather covered in the modeling report, deals source contribution numbers in CAS well, but fail miserably at modelling the few outlier days described with auto and truck trip generation. The Part III are dubious at best. above. While some evaluation of CALPUFF model performance is major industries in the area likely serve as given, a bit deeper discussion of this and examination of some of major origins and destinations for many the individual cases in which high values were measured but not vehicles (commuters driving to and from predicted would add considerably to the credibility of the work. their work place) and for trucks servicing the industries. Therefore, the assertion This study uses American regulatory models as tools. Their that the industrial sources contribute less performance as shown with Q Q plots is not that good and to the overall pollution levels in the area performance references are both old and American. While there is an overstatement unless the models are good reasons to use these models, it would be useful https:// specifically account for the trip generation zcs1.mail.uoguelph.ca/zimbra/public/frameOpenerHelper. caused by the industrial facilities. jsp?id=3&async=true 23/03/09 4:45 PM to confirm that the models indeed are the “best” in some sense for this work. I do not see a list of modelling experts that have been consulted on this matter or, indeed, any exploration of the other tools that are available for doing work of this nature. As a Canadian scientist I think we should start by looking at applicable Canadian science and take the step of ensuring that the US science is “better” before using it. Veronique Bouchet of Environment Canada at CMC in Dorval leads an applied air quality modelling group who use the latest techniques for studies such as these (but maybe not for the very small scales considered here). It would be useful to have opinions from them as to whether the Canadian air quality models, based on the operational weather forecasting models run by CMC, would provide a useful comparison. In Section Six, Figure 6.1(a) and (b) show isopleths stopping at the US/Canada boundary. It would be useful to see the Canadian continuation of these isopleths.                                                      5 Ogilvie, Ogilvie & Company | The independenT public FaciliTaTor       First Round Questions 3Is the current Michael Brauer Gary Foley regional-scale With the exception of ozone, the existing network is This is hard to respond to. I am not familiar with the ambient air monitoring system (Air not sufficient to represent local conditions, except the Canadian AQI. I know it is very different from the US Quality Index – AQI) temporal variability. The regional monitoring system AQI. An AQI is an indicator. If it gives frequent high sufficient to represent local is really not intended to characterize local conditions values, then I would want to re-evaluate whether there air quality conditions or or impacts. It is best applied to identify temporal is a need to enhance the current monitoring system. to identify local air quality changes (both short and long-term) in air quality and impacts? If there are data seems to meet this need (although it was not clear gaps, what are the gaps from my reading if there was historical VOC data). To and how can they be filled? adequately characterize local air quality conditions and the spatial patterns in air quality locally is best done with special studies, such as the CAS, but with more detailed saturation sampling and/or land use regression modelling. Overall, the monitoring network is relatively sparse (in terms of numbers of monitoring sites – in BC there are more than 80 monitoring sites, although some are only seasonal or limited to a single pollutant; Metro Vancouver operates 24 stations in one metropolitan areas alone) given concerns, sources and population in the region, but the supplementation with specific studies (CAS) can help fill this gap. The routine monitoring sites then are best used to indicate trends and indicate some evidence of decreases in SO in 2 Brampton, otherwise no real indication of decreases during 2001-6. Ozone is mainly regional and is well-characterized with the existing monitoring. From a health perspective, the AQI (and AQHI) overall provide very little information on long-term exposures to air pollution and on spatial differences in air pollution concentrations – nor are they designed to provide this type of information.    6 Michael Jerrett Stuart McKeen Ann McMillan The first question is unclear – is meant The air monitoring system (monitors This study is done based on the AQI methodology that has to suggest that the current network of and measurements) is a system, been used for years by the Ministry. While it does provide government monitors is adequate, or not an index. This question is poorly useful information, it is being replaced in many parts of that the current AQI is adequate or some worded, making it difficult to respond. Canada with a new Air Quality Health Index developed combination of both is unclear. I have By comparison with other datasets, federally which uses the data in a somewhat different way answered both questions. and aircraft data in particular, the to provide a forecast that is more closely linked to health argument can easily be made that the effects. The current monitoring network is probably monitoring network in the eastern U.S. adequate to assess daily and seasonal The Air Quality Health Index (AQHI) is a new Index and southeast Canada is sufficiently variations in pollutants for time series designed to provide a real-time estimate of the risk of dense and reliable to identify regional studies of acute health effects. In terms air pollution, and a forecast for the next day. It has been air quality impacts. From simple scaling of assessing spatial variation, the major developed by examining air pollution statistics and arguments, an hourly average monitor component of chronic health exposures, mortality rates for a twenty year period across Canada reading will be representative of ~20 km the network is inadequate. Even the and estimating what effect a combination of pollutants in horizontal scale at mid-day. Monitoring temporary network with the monitors set (particulate matter, ozone, nitrogen dioxide) has on the next to a freeway or within a dense urban up in the CAS is probably incapable of health of Canadians. core is an entirely different matter and measuring the actual spatial variation outside my area of expertise. While mortality is the health indicator used, hospital visits in many pollutants, which may occur on were also tested and found to provide similar results. scale of 30-300 m. For 24-hr average PM the dependence 2.5 Mortality is used however because statistics on it are highly of exceedance events on larger scale The AQI is probably inadequate as a standardized across the country and these are available meteorological conditions tends to wash measure of potential health effects since in a database that can be accessed on an ongoing basis. out local differences between monitors in it represents an amalgam of various Other adverse effects, while considered to be similarly different (urban vs. suburban) locations. pollutants that may act independently or in affected, are not accessible in an appropriate manner. This is not the case for max 8-hr average synergy to produce different health effects. The Index is based on the air pollution-health link, and ozone or for 24-hr average NO . 2 Health Canada has invested considerable is combined with air pollution on an hourly basis giving effort in developing alternative indicators the public a means by which they can judge current and that better represent health effects. These near-term air health risk, and suggestions on how they can health-based indicators are a step in the reduce their health risk should they choose to do so. While right direction beyond the AQI. much of the Index is oriented towards those considered at higher risk than average (those with cardio-respiratory disease and children especially), it is also designed for use by the general public with health messages targetted at various groups within the population. (Barry Jessiman, Health Canada, personal communication). Attached is further information on the AQHI.                                                      7 Ogilvie, Ogilvie & Company | The independenT public FaciliTaTor       First Round Questions 4What do air levels of Michael Brauer Gary Foley common pollutants Ozone episodes and high annual average ozone have The levels that exceed the standards from the Clarkson in the south Mississauga/south Oakville both been associated with increased mortality. The study indicate that there may be a risk, especially to area mean for population occurrence of ozone episodes and their contribution sensitive subpopulations. It is important to note that health risk? to the annual average would be expected to have for US NAAQS (ambient air quality standards), there such impacts in this airshed. Quantitative health is generally a safety factor built in such that most of impact assessment could be conducted using CMA/ the population is not likely to be at risk if there is a OMA ICAP model, perhaps also including more recent slight exceedence of the standard. This is true for all evidence on respiratory mortality associated with criteria air pollutants except for ozone. For ozone, the ozone. background level is generally too high to allow much of a safety factor. In order to better understand potential High PM during much of summer smog season population health risk, exposure modeling of the (perhaps as many as 60 days) is of concern and is general population and the sensitive subpopulations likely clearly associated with health impacts. One needs to be done using a model like MENTOR which would expect increased daily mortality as well as is available from EOHSI at Rutgers University in New impacts on acute cardiovascular disease events Jersey (contact Panos Georgeopolis). (hospital admissions, myocardial infarction incidence). As with ozone, the impacts could be quantitatively estimated with ICAP or similar model. In addition, a growing body of literature indicates traffic proximity impacts in specific areas within an airshed. These studies suggest increased development of childhood asthma, allergic diseases, exacerbation of existing asthma, increased childhood respiratory and middle ear infections, a relatively small impact on reproductive outcomes (low birthweight and pre- term births) and some suggestions of cardiovascular disease impacts (increased progression of atherosclerosis and increased cardiac mortality). 5What, from a Michael Brauer Gary Foley public health Based on Clarkson Airshed Study (CAS) – PM I believe that the PM is the pollutant of greatest perspective, are 2.5 2.5 the air contaminants concentrations are elevated and sometimes exceed concern. However, I also believe that the there may of greatest concern? the CWS. Although this occurs mostly during smog be synergies between PM2.5 and various organic Are there particular events and is in part due to regional transport – the compounds and that this is a very difficult area to contaminants on which heavy regional impacts suggests that extra attention focus on. the Ministry of the should be paid to local sources as the airshed is Environment and other already under stress. The industrial impact (estimated stakeholders should at 25%) seems quite high compared to other urban focus? areas. Traffic-related pollutants in general – can be characterized by NOx, but perhaps some assessment of diesel PM is warranted. Ozone is mainly regional and is already well- characterized with the existing monitoring.    8 Michael Jerrett Stuart McKeen Ann McMillan As in many urban areas around the world, the current levels of ambient Health risks are no higher, and are This question is beyond my area air pollutants in this area might be contributing to excess mortality, probably somewhat lower (due to of expertise, but I will mention exacerbation of asthma symptoms, declines in lung function growth frequency and occurrence of the that I chaired the CCME Working in children, and possibly to incident asthma, adverse birth outcomes, relatively clean westerly and north- Group on Air Quality Objectives and adverse cardiovascular outcomes including atherosclerosis, blood westerly wind sector) than living in and Guidelines (WGAQOG) when pressure changes, and altered autonomic function with decreased heart any other North American urban they did the studies to develop rate variability and consequent increased risk of heart attack. The current area. the “reference levels” for PM and 10 state of the evidence on air pollution health effects probably qualifies “as PM and recall that they were 2.5 suggestive but insufficient to establish causal association” between air chosen as levels below which pollution and these or other potential health outcomes. (This is a term no health effects were found used by the U.S. Surgeon general to indicate likely causality between a in a thorough literature review, putative exposure and health effects.) even though these pollutants have no threshold for effects. Thus, personally I would not be concerned about health effects unless the levels were often substantially above these levels. Michael Jerrett Stuart McKeen Ann McMillan The current evidence base is substantial for respirable fine particulate Not my area of expertise. But I’m not a health expert, however matter, with suggestion that elevated exposures following a linear dose- the Canadian government has it has always seemed to me that response function that affects all the health outcomes mentioned above. established the new AQI taking into categorizing particulate matter Evidence of health effects from ozone is also suggestive but insufficient account ozone, PM and NO from on a size basis only, not on a 2.5 2 to establish a causal relationship with a wide rage of respiratory and the best available epidemiology chemical composition basis could cardiovascular outcomes. Beyond these pollutants, which have received and health impact studies. It should be glossing over some public the bulk of attention in the health effects literature, emerging evidence be relatively straightforward to health issues. I recognize that from toxicology and to a lesser extent epidemiology suggests that apply the Canadian AQI to the there is ongoing research in this elemental carbon, PAHs and transition metals all have the potential of measurements collected within area and it is probably too soon to inducing oxidative stress in the lung and potentially systemic pulmonary CAS Part II to determine both the consider it in a regulatory setting. inflammation. Over the long-term such inflammation may contribute to the degree, and sensitivity to each of There is growing concern with formation of many cardiorespiratory conditions. these 3 species. It should also be nanomaterials, and it would easy to do the same analysis for Diesel particulate is recognized as a probable carcinogen by be good to know if any of the more restrictive pollutant thresholds many national and international health or environmental protection industries in the area are using to accommodate uncertainties organizations. There should be more emphasis on assessing elemental them. within the AQI/health-impact carbon or other markers of diesel particulate. standardization. CAS Part II also shows certain known carcinogenic UFPs are capable of trans-locating into the circulation, and while the VOC may be a concern, which health effects of UFPs are only partly understood, there is substantial falls outside the AQI assessment biological plausibility and some toxicological evidence suggesting the for criteria pollutants. The analysis ultrafine fraction may be particularly important. Traffic is a major source of suggests this could be limited to ultrafines and possibly some of the industries in the area also contribute isolated receptor regions, but that to ultrafine concentrations. Future studies should be assessing the remains an open question. ultrafine fractions. In summary, there is a need to assess the constituents of the particles (noted above), the ultrafine particle numbers, and elemental carbon. Little is known about the potential effects of pollution mixtures, so this could also be a topic of future research.                                                      9 Ogilvie, Ogilvie & Company | The independenT public FaciliTaTor       First Round Questions 6If additional Michael Brauer Gary Foley monitoring is Focus on source impacts and understanding of spatial There are many ways to analyze monitoring data to warranted for the Clarkson Airshed, variability. This can be achieved by saturation sampling extract more information including specific source what should the and land use regression modeling, and source apportionment of PM10 samples using electron objective(s) and goal(s) apportionment of collected PM and VOC samples. microscopy and volatile organic compounds from of that monitoring be (i.e., individual monitoring stations by using source VOC I see little value in a specific (local) study to link air improved understanding signatures. Sometimes these bring insights about the quality to public health, except perhaps applying the of local air quality location, time and the point source that causes the CMA/OMA ICAP model specifically to this region (if this processes? Emissions high monitoring values. It can also bring insight into has not already been done in OMA reports). contributions by sector? the existence of previously unknown sources of fugitive Linkage of air quality to emissions of VOCs. public health? Improved There are many uncertainties in data and model understanding of multiple parameters that can affect the results of modeling source cumulative effects?) studies. Similarly, many of the air quality processes in the model are parameterized which can add uncertainties. Only the uncertainties in fugitive emissions appear to be mentioned in the report. There also is no discussion on how the air quality model was evaluated to determine that it was suitably configured and had the input data and parameters of sufficient quality to ensure upfront that the objectives of the study could be met. 7Please add any Michael Brauer Gary Foley other comments or As a general comment, the typical emissions information that you feel may be useful to us in inventory-based approach to assessing sources this review. might neglect high exposures due to close proximity of sources to population. Although there are no standardized approaches, a more exposure-focused approach (assessment of intake fraction, for example) might provide more useful insight if a major concern is population exposure.    10