Beachville, Oxford County Particulate Matter Sampling. Contact information:

Transcription

1 Beachville, Oxford County Particulate Matter Sampling Requestor: Request prepared by: Date: 26-Apr-16 Peter Heywood, Manager, Health Protection, Oxford County Public Health James Johnson MPH, Environmental Health Analyst Sunil Varughese, MSc, Environmental Health Analyst Ray Copes, MD, Chief, Environmental and Occupational Health Contact information: Key Findings Quarry activity and dust events tend to produce particles in size fractions larger than PM 2.5 such as PM 10 and Total Suspended Particulate (Total Dust, or TSP), the concentration of these larger particles varied across the sites sampled. In contrast concentrations of PM 2.5 did not show appreciable variation and likely reflect regional ambient concentrations. Analysis for increases in PM over short periods of time revealed evidence of these occurrences at some but not all sites. These likely originate with a neighbourhood source and are localized elevations. Resampling of a site that experienced dust events in the summer showed dust events occurring in one but not both fall reassessments.

2 Introduction In response to local residents concern over exposure to particulate matter (e.g., dust clouds), Oxford County Public Health called on the Environmental and Occupational Health (EOH) Team at Public Health Ontario (PHO) to assess Oxford County air quality. It was recognized from the outset of this assessment that peak events, rather than changes to average concentrations of particulate matter were of immediate concern to the local community. Through the summer and fall of 2015, Public Health Ontario assisted Oxford County in conducting air monitoring for 2-week periods at 8 properties located in the Beachville area. Monitoring was performed to measure concentrations of particulate matter (PM) of different sizes at homes of local residents who volunteered to participate. As part of these monitoring activities, PHO conducted the following: We collected and analyzed PM data and local meteorological data to better characterize day-today dust levels. We looked for evidence of short-term (5-60 minute) increases in PM levels ( dust events such as dust clouds). Particulate Matter PM is a complex mixture of particles and liquid droplets of varying sizes. The size of particles is directly linked to the ability of PM to cause health problems. Health agencies and regulatory bodies are especially concerned with particles that are 10 micrometers in diameter or smaller (PM 10 ). PM 2.5 is a component of PM 10, and is often referred to as the fine fraction. The coarse fraction includes particles that have a diameter between 2.5 and 10 micrometers. PM 10 is of importance from a health perspective because it includes the particles that are able to pass through the throat and nose and enter the lungs. Air quality objectives or standards have been established for both PM 2.5 and PM 10. PM 2.5 consists mainly of combustion particles from motor vehicles and the burning of coal, fuel oil and wood, but will also contain crustal materials from finely crushed road dust and soils. The coarse fraction of PM 10 consists mainly of crustal particles generated mechanically from agriculture, mining, construction, road traffic and other related sources, as well as particles of biological origin 1. A more detailed outline of these particles is shown in Figure 1. LOCAL VERSUS REGIONAL POLLUTION Larger particles tend to settle out of the air, and are not as likely to contribute to regional pollution 2. As a result, larger fractions of dust (PM and larger) are more likely to be due to local sources (sources within metres to a few kilometres) than PM 2.5, which can be generated both by local sources and regionally (a few kilometres to tens or hundreds of kilometres). In this report, local events are referring

3 to events that are likely due to local road, quarry or agricultural activity, due to the increased presence of particulate matter in fractions larger than 2.5 microns. Figure 1 - Description of different particulate matter fractions Methods Locations Sampled Short term monitoring was conducted between October 2015 and November 2015 at volunteer sites as part of an ongoing monitoring effort in the Oxford County area (Figure 2).

4 Site Dates Sampled 1 June 4-22, June 4-22, June 22 - July 9, July 9 - July 27, July 9 - July 27, July 27 - August 12, July 27 - August 12, (2) October 6-22, (2) October 6-22, (3) October 22 - November 6, October 22 - November 6, 2015 Figure 2 Sites sampled in Oxford County in the summer and fall of 2015 Instruments Used For this assessment, DustTrak 8533 Desktop monitors were used for assessment of particulate matter fractions of PM 2.5 and PM 10. Wind direction, speed, temperature and relative humidity were measured at the same time as PM 2.5 and PM 10 assessments with the Kestrel 4500 Weather Meter. DUSTTRAK INSTRUMENT TSI DustTrak monitors are used by Public Health Ontario for particulate matter assessments because of their ability to provide reliable data on particulate matter concentration with portability and relative ease in operation and maintenance compared to other measurement methods. An important advantage of this instrument is the continuous, direct reading capability which allows for the determination of very short term (e.g., 5 minute) changes in PM concentrations. They can also be set up fairly easily at any site where household power is available. These monitors are not approved by the U.S. Environmental Protection Agency under its Federal Reference Methods for PM 2.5 for verifying compliance with air quality standards. However, the DustTrak measurements provide useful information to apply to community risk management and assessing potential exposures to people in different areas of the community. Gravimetric monitors which are larger and retain a sample of the dust collected for later chemical analysis if required. They are more precise than the Dust Trak and are approved for verifying compliance

5 with air quality standards. However, they do not give minute by minute readings and it is difficult to set them up at multiple sites close to each other to check for differences in PM within small areas such as neighbourhoods. In Southern Ontario, previous studies have documented that readings from a DustTrak tend to be 2-3 times higher than particulate levels sampled with gravimetric methods (which are considered superior) 3. This number is referred to as a correction factor. A correction factor of 2.4 for PM 2.5 has been previously established for the Toronto Area by members of the Southern Ontario Centre for Atmospheric Aerosol Research (SOCAAR) research group at the University of Toronto 4. No correction factor is documented for the Oxford County Area, and no comparable measurement could be used to establish a correction factor due to equipment limitations. For this report, a conservative correction factor of 2 was used in reporting. This number is conservative because it is the lower end of the range of reported corrections necessary for measurements taken in the Southern Ontario airshed 3 (see Appendix 4). KESTREL 4500 WEATHER METER The Kestrel 4500 Weather Meters are used by Public Health Ontario for measuring weather information because of their accuracy and portability. Relevant parameters gathered include wind direction, wind speed, temperature and relative humidity. Sampling Strategy Sampling sites were volunteered by members of the community and chosen based on proximity to potential sources of community dust and to identify spatial variation that may be occurring. To assist this, when possible, measurements were taken simultaneously at two sites in the Beachville area. Wind direction, wind speed, temperature and relative humidity data were collected alongside the particulate matter data. The purpose of this additional data was to assist in identifying areas where particulate matter levels were influenced by wind direction and speed, where possible. Measurements were taken every five minutes for particulate matter data, wind direction and wind speed. For reporting purposes and for comparisons to CAAQS, particulate matter measurements were averaged to 24-hour segments based on the date of the sample. Measurements were also reported in the five-minute time segments. For the fall assessment, residents also reported information in a diary about local events that may contribute to particulate matter levels. These reported events were compared to graphs of daily particulate matter to check for associations and were provided to residents in separate reports. Neither the summer of fall analyses were designed to associate elevated particulate events with a given source.

6 Reference Guidelines Guidelines used for comparison in this assessment were outlined in documents produced by the World Health Organization (WHO) 5 and the Standards development branch of the Ministry of Environment and Climate Change 6. The Canada Ambient Air Quality Standard (CAAQS) for PM 2.5 (28 µg/m 3 ) 7 is slightly higher than the WHO guideline (25 µg/m 3 ). Standards for recommended peak values are not used by either organization; the MOECC document clearly states that no conversion can be made from this standard to shorter time periods 6. For the purpose of this assessment, results were compared to the CAAQS for PM 2.5 and the interim Ambient Air Quality Criteria (AAQC) guideline value for PM 10 using both 24-hour averages and peak-events. These levels were used as a benchmark for comparison in the absence of short term guidelines. For this report, it is important to note that peak events above the 24- hour guideline values and single days above the CAAQS for PM 2.5 would not necessarily constitute an exceedance of the CWS. According to the WHO, the health effects resulting from PM exposure are broad, but are predominantly in the respiratory and cardiovascular systems 5. For smaller particulate sizes (PM 2.5 and PM 10 ), these health effects are the rationale for the guideline values. In larger fractions, visibility is more important than health outcomes because these particles are too large to enter the respiratory system 5. It is also stated that it is unlikely that any standard or guideline will lead to complete protection in every individual against all possible adverse health effects,and that the standards setting process needs to aim to set the lowest concentrations possible in the context of local constraints, capabilities and public health priorities. 5

7 Table 1-24-Hour Guidelines for Particulate Matter 24-hour Suspended Particulate (<44 µm diameter) 24-hour PM 10 Guideline (AAQC 5 and WHO 5 ) Guideline value 120 µg/m 3 50 µg/m 3 (0.05 mg/m 3 ) 24-hour PM 2.5 Guideline 25 µg/m 3 (WHO) 5 28 µg/m 3 (CAAQS) 5 Rationale for guideline value Visibility Health Health Analysis All analysis was conducted in R version and Microsoft Excel. Wind rose plots were created with the openair package in R 8. Final analysis included three parts: Comparisons of sampled data to applicable 24 hour standards. Analysis of peak events, with percentiles in different fractions of particulate matter. This method has been reported previously as a method of monitoring events from local environmental activity 9. Wind direction and wind speed information captured alongside the pollution information was interpreted using pollution roses.

8 Concentration (µg/m³) Results Guideline Comparisons Averages captured from each 2-week sampling period are presented in figure 3. To illustrate the spread of concentrations, the maximum sampled 24-hour reading at each site is included alongside the 2-week averages for PM 10 and PM 2.5. The highest 24-hour reading was at Site 3, with maximum value of µg/m³ over 24 hours. This trend in PM2.5 was similar to a period of elevated readings captured at an air quality monitoring station operating in London, Ontario (19.25 µg/m³), which suggests that the elevation was at least partly due to a regional elevation rather than local PM sources 10. A figure of boxplots showing of the range of concentrations at each site is included in Appendix 1. The levels of PM 10 at all sites closely followed PM 2.5, and none of the days approached or exceeded the 50 µg/m³ guideline value for PM (2) 7(2) 7(3) 8 PM10 PM2.5 Maximum 24- hour PM2.5 PM2.5 Canada Ambient Air Quality Standard Figure 3 - Mean PM 10 and PM 2.5 concentrations for summer and fall 2015 sites Concentrations presented are the average reading for each site over the periods sampled. Also presented are the maximum 24-hour readings captured at each site. This figure includes both the measurements taken in the summer (Sites 1-7) and the fall (Sites 3(2), 7(2),7(3) and 8).The concentrations are very similar for both PM 10 and PM 2.5, indicating PM in this area was mainly made up of fine PM (PM 2.5 ).

9 Analysis of Peak events To further identify peak events that may be a concern to local residents, a peak event analysis was conducted on the two-week samples taken at each site, using percentiles. A percentile of the data represents a cut-off, where a certain percentage of the readings fall. To illustrate this concept, in a list of numbers from one to a hundred, the 95 th percentile would be 95. Potentially confusingly, this means that the 95 th percentile is actually a cut-off for the highest five percent of data. Because it represents a cut-off for higher values, the percentile is a useful tool for investigating peak dust events. This method has been identified in a previous study that investigated resident complaints about industry dust 9. Summaries of the percentile assessments are presented in Figure 4, and numerical summaries are in Appendix 2. In the context of our measurements, a percentile can be interpreted as a period of time. For example, in a 2-week assessment, a reading of 47 µg/m³ of 99.9 th percentile means that 99.9 per cent of the time, the levels were below 47 µg/m³, and 0.1 percent of the time (meaning for about 20 minutes), the levels were above 47 µg/m³. Other examples are provided in Table 2. An increase in concentrations towards the 99.9 percentile indicates the presence of infrequent, local shortterm dust events, especially when occurring in larger fractions of PM. Table 2 - Approximate amount of time represented by a percentile for a two-week interval Percentile Amount of time levels sampled were at or above the level indicated (approximate) 95 % 16 hours and 38 minutes 99% 3 hours and 22 minutes 99.9 % 20 minutes Figure 4 presents the percentiles of dust by fraction for each site sampled. On the x-axis of these graphs are increasing percentiles. These can be interpreted as PM matter levels of increasing rarity, with the 70 th percentile is a level that occurred 30 percent of time, and the 99.9 th percentile occurring only 0.1 percent of the time. Sites 3,4,6 and 7 all demonstrated evidence of peak events occurring in the summer assessment In the fall assessment, site 3 demonstrated evidence of peak events and site 7 also showed evidence of peak events in the first (but not the second) round of sampling. Because of the short time interval represented by the 99.9 th percentile, the increases in the summer at sites 3 and 6 were likely due to single dust events.

10 Concentration (µg/m³) Figure 4-70 th, 95 th, 98 th and 99.0 th th percentiles for each site, in the Total, PM10 and PM2.5 fractions A steep increase from the 99 th to the 99.9 th percentile indicates the presence of sporadic dust events. This was observed at Sites 3,4,6 and 7 in the summer, and in sites 3(2) and 7(3) in the fall.

11 Wind Direction and wind roses Wind direction readings and wind speed was paired with total dust with PM 2.5 trends removed in an effort to visualize movement of larger dust particles across the sites. Results are presented in appendix 3. Readings at site 7 demonstrated both the presence of dust events and the results were downwind of the quarry. This finding was consistently upon repeated sampling. Local obstructions such as trees, sheds and houses likely had an influence on the wind data that were obtained. Because of these local obstructions, limited interpretation is possible for the wind information. Discussion This assessment was not a compliance test. Instead, guidelines were used as a benchmark for comparison to see examples of how residents of Beachville experience particulate matter. Larger PM (PM larger than PM 2.5 ) is more likely to come from local sources and tends to deposit closer to where it is produced. For this reason, Total Suspended Particulate and PM 10 are useful information to capture near a quarry operation. Initial descriptive assessments of PM 2.5 data suggested that for PM 2.5, trends captured over 2-week periods in Beachville paired well between sites and with PM 2.5 data captured by the MOE in London for Air Quality Hazard Index readings 10. This finding could be further investigated when the MOE releases validated data for the 2015 year. At all sites sampled, particulate matter It was recognized in this assessment that a primary concern of the community was visible dust events rather than compliance with applicable standards. The advantage of using a DustTrak over a compliancebased gravimetric test was the ability of the Dusttrak to characterize concentrations over a short period of time, and by the size of the PM, allowing for the characterization of peak events. Of the sites sampled simultaneously in the summer, sites 1 and 2 showed similar trends, site 4 was dustier than site 5 and site 7 was dustier than site 6. It has been noted in a previous analysis that levels of PM can vary significantly by season and by month 9. Resampling of site 7 demonstrated this finding. As seen in Figure 3, site 7 showed the presence of dust events in two of the the three times it was sampled, but peak events were absent for the first resampling period in the fall (site 7(2)). This finding highlights the usefulness of ongoing monitoring as an investigation tool for PM. Because these assessments relied on volunteered sites from the community, and because of concerns about instrument security, the information captured in this assessment is primarily intended to provide residents with a snapshot of PM levels in an outdoor area where they would be expected to spend a reasonable amount of recreational time (in a backyard). As a further refinement, future assessments could use prevailing wind data and seek locally data collected from the MOECC as a reference for comparison.

12 References 1. Laden F, Neas LM, Dockery DW, Schwartz J. Association of fine particulate matter from different sources with daily mortality in six U.S. cities. Environ Health Perspect. 2000;108(10): National Center for Environmental Assessment-RTP Division. Integrated Science Assessment for Particulate MatterU.S. Environmental Protection Agency; Available from: 3. Wallace LA, Wheeler AJ, Kearney J, Van Ryswyk K, You H, Kulka RH, et al. Validation of continuous particle monitors for personal, indoor, and outdoor exposures. Journal of Exposure Science and Environmental Epidemiology. 2010;21(1): Evans G, Sabaliauskas K. consultation and follow up conversation on Correction Factors and Humidity considerations for DustTrak instruments World Health Organization. WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide and sulfur dioxide: Summary of Risk Assessment. Geneva, Switzerland: World Health Organization; Available from: 6. Standards Development Branch of the Ministry of Environment. Ontario's Ambient Air Criteria. PIBS # 6570e01 ed; Available from: 7. Canadian Council of Ministers of the Environment. GUIDANCE DOCUMENT ON ACHIEVEMENT DETERMINATION CANADIAN AMBIENT AIR QUALITY STANDARDS FOR FINE PARTICULATE MATTER AND OZONE; Available from: 8. Carslaw D, Ropkins K. Openair an R package for air quality data analysis. Environmental Modelling & Software.; Orkin A, Leece P, Piggott T, Burt P, Copes R. Peak event analysis: a novel empirical method for the evaluation of elevated particulate events. Environ Health. 2013;12:92,069X London: Hourly Fine Particulate Matter Readings [Internet].: QUEEN'S PRINTER FOR ONTARIO; 2016; cited April 12, 2016]. Available from: &start_month=7&start_year=2015&table=1&station_id=15026&submitter=update

13

14 Appendices Appendix 1: Distribution of 24-hour means at all sites Figure 3-24-hour comparisons at each site Shown here are all of the 24-hour averages taken, by site, presented as boxplots. The box contains half of the observations centered around the most frequently observed level (the median). The vertical lines running from the tops and bottoms of the boxes span the highest 25 percent of values and the lowest 25 percent of values, respectively. The dots above or below the boxes and lines are outliers, meaning they were very uncommon measurements. Highest readings were seen at site 3, with PM 10 following PM 2.5 closely at all sites sampled.

15 Appendix 2: Tables for Percentiles, by PM fraction Table 3 PM 2.5 Means, 95 th and 99 th percentiles, and Max (µg/m 3 ), by site Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 3(2) Site 7(2) Site 7(3) Site 8 Mean % % % Max Table 3 PM 10 Means, 95 th and 99 th percentiles, and Max (µg/m 3 ), by site Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 3(2) Site 7(2) Site 7(3) Site 8 Mean % % % Max Table 4 Total Dust Means, 95 th and 99 th percentiles, and Max (µg/m 3 ), by site Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 3(2) Site 7(2) Site 7(3) Site 8 Mean % % % Max * Table 5 Total Dust with PM 2.5 removed, Means, 95 th and 99 th percentiles, and Max (µg/m 3 ), by site Site 1 Site 2 Site 3 Site 4 Site 5 Site 6 Site 7 Site 3(2) Site 7(2) Site 7(3) Site 8 Mean % % % Max

16 Appendix 3: Wind roses for each site Figure 4 - Wind Direction and Total dust (no PM 2.5 ) summaries for each site (pollution data in µg/m 3 ) 8 These pollution roses are summaries of the total dust with PM 2.5 fraction removed paired with wind direction for the 2 weeks. The summer sites are labelled site2_1-site2_7, and the fall sites are labelled site 3_3-site3_8. The goal of this was to look at whether specific wind directions were associated with an increased proportion of peak events. In determining wind direction, the wind roses should be treated as arrows that point in the direction of the wind. For example, at site 3_7, wind was coming from west of south approximately 25 % of the time. While wind was blowing from the south, the concentrations of total PM ranged from µg/m 3, with the majority of samples in the 0-6 µg/m 3 range.

17 Appendix 4: Corrected Values for PM 2.5 and PM 10 Following a review of relevant literature 3, a correction factor of 2 was applied to the PM 2.5 readings, and to the PM 2.5 fraction of the PM 10 readings. Though a correction factor did not exist for PM 10 data, the values needed to be altered since PM 2.5 constitutes a portion of the PM 10 fraction. Keeping this in mind, the following formulas were used to alter the PM 2.5 and PM 10 data: CORRECTED VALUES FOR PM 2.5 ( ) ( ) CORRECTED VALUES FOR PM 10 First, the coarse fraction was established with the uncorrected readings: ( ) ( ) Then the coarse fraction was added to the corrected PM 2.5 readings: ( ) ( )