Investigation of Fine Particulate Matter Characteristics and Sources in Edmonton, Alberta

Transcription

1 Investigation of Fine Particulate Matter Characteristics and Sources in Edmonton, Alberta Executive Summary Warren B. Kindzierski, Ph.D., P.Eng. Md. Aynul Bari, Dr.-Ing. 19 November 2015

2 Executive Summary This study investigated characteristics of air quality and various source contributions to ambient fine particulate matter (PM 2.5 ) in the Edmonton Capital Region. The study used historical data measured at National Air Pollution Surveillance (NAPS) air monitoring stations in Edmonton, including a chemical speciation monitoring station (Edmonton McIntyre station). Three objectives were sought: 1. Investigate the characteristics and trends of individual air pollutants, including fine particulate matter (PM 2.5 ), in order to understand how bad or good air quality is in Edmonton. 2. Investigate the sources of PM 2.5 at an Environment Canada National Air Pollution Surveillance (NAPS) chemical speciation monitoring site in Edmonton, including identifying the contribution from coal combustion sources. 3. Investigate origins and causes of PM 2.5 concentration differences in Edmonton during 2010 relative to other years. Objective 1 Investigate characteristics and trends of individual air pollutants in Edmonton Trends in Environmental Canada s National Pollutant Release Inventory reported industrial emissions in the Edmonton Capital Region for PM 2.5, oxides of nitrogen (NO X ), sulfur dioxide (SO 2 ), ammonia (NH 3 ), volatile organic compounds (VOCs) and major trace elements were investigated over the last decade ( ): Statistically significant decreasing trends were observed for the industrial combustion pollutants NO X (p 0.01) and SO 2 (p 0.05) over the last decade. A statistically significant decrease (p 0.05) was observed for arsenic (As). Statistically significant downward trends were observed for other industrial elements e.g., vanadium (V), manganese (Mn), chromium (Cr), copper (Cu) and cobalt (Co). Surrogate data (~20,600 additional motor vehicle registrations annually) suggest an increasingly important role of transportation sector emissions in Edmonton Capital Region over the past decade. This is opposite to reported industrial emissions trends described above. Statistically significant trends (p 0.05) for hourly average percentile concentrations (50 th, 65 th, 80 th, 90 th, 95 th and 98 th percentiles (%iles)) of air pollutants measured using continuous monitors were observed at Edmonton central and east stations (17-year period of record: ) and at Edmonton south and McIntyre stations (9-year period: ) (Table ES1): Decreasing trends were observed for hourly average percentile concentrations of nitrogen dioxide (NO 2 ), SO 2, total hydrocarbon (THC), and carbon monoxide (CO). Air quality in Edmonton has improved for these air pollutants over the past 17 years. No change was observed for PM 2.5 at any of the monitoring stations. During 2009 equipment at many of the original continuous PM 2.5 monitoring stations in Alberta was upgraded to improve capture of some components of fine particulate matter (i.e., semi-volatile material) which were lost under the previous method. Thus despite absolute PM 2.5 levels at air monitoring stations being bumped higher in 2010 and subsequent years relative to previous years as a result of equipment upgrades in 2009, no statistically significant trends were observed at any of the stations. For ground-level ozone (O 3 ), a small increasing trend was found at lower hourly percentile concentrations (50 th to 80 th %iles) at Edmonton central, a small decreasing trend was detected at higher percentile concentrations (at 80 th to 98 th %iles) at Edmonton east and no trend was observed at Edmonton south at all percentile concentrations. These characteristics indicate evidence of spatial variation of O 3 precursor concentrations e.g., NO X, VOCs across Edmonton rather than any type of consistent trend in O 3 concentrations in Edmonton over the past 17 years. i

3 Table ES1. Summary of linear trends for hourly average percentile concentrations of air pollutants at Edmonton air monitoring stations. Concentration Edmonton Central ( ) Edmonton East ( ) Edmonton South ( ) Edmonton McIntyre ( ) (percentile) Trend Change/year R 2 Trend Change/year R 2 Trend Change/year R 2 Trend Change/year R 2 NO 2 50 th 0.71 ppb ppb ppb th 0.77 ppb ppb ppb th 0.75 ppb ppb ppb th 0.70 ppb ppb ppb th 0.68 ppb ppb 0.75 n/a n/a 98 th 0.72 ppb ppb 0.62 n/a n/a SO 2 * 50 th 0.11 ppb 0.61 n/a n/a 65 th 0.17 ppb 0.85 n/a n/a 80 th 0.15 ppb 0.65 n/a n/a 90 th 0.15 ppb ppb th 0.14 ppb ppb th 0.17 ppb 0.31 n/a n/a PM th n/a n/a n/a n/a n/a n/a n/a n/a 65 th n/a n/a n/a n/a n/a n/a n/a n/a 80 th n/a n/a n/a n/a n/a n/a n/a n/a 90 th n/a n/a n/a n/a n/a n/a n/a n/a 95 th n/a n/a n/a n/a n/a n/a n/a n/a 98 th n/a n/a n/a n/a n/a n/a n/a n/a O 3 50 th 0.38 ppb 0.75 n/a n/a n/a n/a 65 th 0.33 ppb 0.68 n/a n/a n/a n/a 80 th 0.26 ppb ppb 0.47 n/a n/a 90 th n/a n/a 0.34 ppb 0.55 n/a n/a 95 th n/a n/a 0.44 ppb 0.57 n/a n/a 98 th n/a n/a 0.55 ppb 0.54 n/a n/a THC 50 th n/a n/a <0.1 ppm 0.35 n/a n/a 65 th <0.1 ppm 0.35 <0.1 ppm 0.52 <0.1 ppm th <0.1 ppm 0.28 <0.1 ppm 0.53 n/a n/a 90 th <0.1 ppm 0.40 <0.1 ppm 0.56 <0.1 ppm th <0.1 ppm 0.35 <0.1 ppm 0.38 n/a n/a 98 th <0.1 ppm 0.25 <0.1 ppm 0.35 n/a n/a CO 50 th <0.1 ppm 0.93 <0.1 ppm 0.69 n/a n/a 65 th <0.1 ppm 0.93 <0.1 ppm 0.68 n/a n/a 80 th <0.1 ppm 0.92 <0.1 ppm 0.84 n/a n/a 90 th <0.1 ppm 0.93 <0.1 ppm 0.66 <0.1 ppm th 0.10 ppm 0.91 <0.1 ppm 0.78 <0.1 ppm th <0.1 ppm 0.92 <0.1 ppm 0.92 <0.1 ppm 0.92 *not measured at Edmonton central; Direction of trend (p = 0.05): no change; increasing; decreasing; n/a: not applicable ii

4 Eight-year trends in chemical species present in 24 h integrated PM 2.5 samples collected at Edmonton McIntyre station were investigated (Table ES2): Statistically significant decreasing trends for 24 h concentrations of OC (organic carbon), EC (elemental carbon), oxalate, barium (Ba) and lead (Pb) (p 0.05) and cadmium (Cd) (p 0.10) were observed. A statistically significant increase (p 0.05) was observed for NaCl (a component of road-salt). No statistically significant changes were observed for all other chemical species examined. Concentrations of K + and Zn exhibited strong and significant seasonal variability with higher concentrations in winter than in summer. Seasonal patterns with high winter levels of these tracer elements likely reflect wood smoke origins more than other potential sources in the Edmonton Capital Region. Table ES2. Temporal trends (8-year) in ambient concentrations of PM 2.5 chemical species at Edmonton McIntyre air monitoring station. Chemical species Unit 2007 gmean 2014 gmean Slope unit/year % Change per year Significance level p-value OC μg/m % 0.05 EC μg/m % SO 4 μg/m n.s NO 3 μg/m n.s NH 4 μg/m n.s NaCl μg/m % 0.05 K + μg/m n.s Oxalate μg/m % 0.05 Al ng/m n.s As ng/m n.s Ba ng/m % 0.01 Cd ng/m % Co ng/m n.s Cr ng/m n.s Cu ng/m n.s Fe ng/m n.s Mn ng/m n.s Mo ng/m n.s Ni ng/m n.s Pb ng/m % 0.05 Sb ng/m n.s Se ng/m n.s Sn ng/m n.s Sr ng/m n.s Ti ng/m n.s V ng/m n.s Zn ng/m n.s gmean = geometric mean Objective 2 Investigate fine particulate matter sources in Edmonton The U.S. Environmental Protection Agency Positive Matrix Factorization (PMF) model version 5.0 was run under two scenarios (a base and a constrained run) using data from the Edmonton McIntyre station for the time period of The plausibility and interpretability of solutions with six to eleven factors (sources) were examined and a 9-factor solution best represented the makeup of ambient PM 2.5 sources at Edmonton McIntyre station (Table ES3 and Figure ES1). iii

5 Table ES3. Predicted sources and their contributions to PM 2.5 at Edmonton McIntyre station for Possible sources Key chemical species Base run Constrained run μg/m 3 % μg/m 3 % Factor 1 SOA OC, EC, arabitol, oxalate Factor 2 Secondary sulfate SO 2 + 4, NH Factor 3 Secondary nitrate NO + 3, NH Factor 4 Soil Ca +2, Mg +2, Al, Fe, Sr, Ti Factor 5 Traffic Ba, As, Cu, Sb, Co, EC, OC Factor 6 Biomass burning Levoglucosan, mannosan, K +, Cd, OC Factor 7 Road-salt Na +, Cl Factor 8 Refinery V, Mo Factor 9 Mixed industrial Cr, Cu, Mn, Fe, Mo, Co, Ni, Sn, Ti, Zn a b Winter Figure ES1. Average contributions of constrained PMF-derived sources at Edmonton McIntyre station for (a. seasonal, b. winter). iv

6 The major PM 2.5 sources identified in Edmonton were made up of secondary particulates (i.e., those that form in the atmosphere from other gaseous pollutants). These included secondary organic aerosol (SOA), secondary sulfate and secondary nitrate and together they contributed to two-thirds of the PM 2.5 mass concentrations on average (5.5 µg/m 3 ). Other PM 2.5 sources identified in Edmonton were made up of primary particulates (i.e., those that are directly released into the atmosphere). For these particles, soil, traffic and biomass burning emissions contributed to one-quarter of PM 2.5 mass concentrations on average (2.0 µg/m 3 ). Minor primary particle sources (road-salt, refinery and mixed industrial emissions) contributed to less than one-tenth of PM 2.5 mass concentrations on average (0.6 µg/m 3 ). Coal combustion emissions are associated with secondary particles. This is discussed further below: Secondary Organic Aerosol (SOA) The potential for SOA formation in coal combustion plumes is considered to be small or unimportant based on in-plume versus out-of-plume measurement studies published elsewhere. Consequently coal combustion emissions are not considered an important source of SOA identified in this study. Backward trajectory analysis was performed using the National Oceanic and Atmospheric Administration (NOAA) Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model. The trajectory analysis supports these published studies as it identified other plausible local and long range SOA sources. These include local sources such as vehicle exhaust and industrial activities, and distant sources such as the Yellowhead transportation corridor west of Edmonton, biogenic (rural) emissions, biomass burning (wildfire smoke), and residential wood fireplace burning. Secondary sulfate Secondary sulfate was interpreted to be related to background regional sulfate that is found in high abundance due to oil and gas extraction and production activities throughout Alberta. Based on backward trajectory analysis, only a small contribution to secondary sulfate was observed from the region immediately west of Edmonton where coal combustion sources are located. The backward trajectory analysis indicated that air parcels traveling over the region immediately west of Edmonton would be, on average, associated with lower concentrations of PM 2.5 for secondary sulfate at Edmonton McIntyre station compared to air parcels traveling over numerous other locations. Possible presence of local industrial sources and backward trajectory (long-range) analysis support that coal combustion sources west of Edmonton do not dominate the contribution to PM 2.5 for secondary sulfate at Edmonton McIntyre station. While the analysis undertaken here is insufficient to accurately quantify the contribution to secondary sulfate from coal combustion sources, their contribution is projected to be in the range of less than one-tenth to less than one-fifth of the secondary sulfate mass. This is consistent with a small contribution of tracer elements typically associated with coal combustion such as Se, As, Cd, Pb, and Sn observed with this factor. Secondary nitrate Secondary nitrate showed strong seasonality with the highest concentrations in winter. Correlations of secondary nitrate with NO 2, CO, THC and some VOCs such as benzene, toluene, ethylbenzene, xylene and other aromatic hydrocarbons (e.g., ethyltoluene isomers, trimethylbenzene isomers) as well as with alkanes suggest a strong influence of local sources such as vehicle exhaust and industrial activities. Backward trajectory analysis indicated that the region immediately west of Edmonton where coal combustion sources are located is not the only trajectory path associated with elevated levels of PM 2.5 for secondary nitrate at Edmonton McIntyre station. Other important regional precursor sources of secondary nitrate influencing Edmonton McIntyre station are located in Alberta south of Edmonton, northwestern British Columbia and southern Saskatchewan. Plausible explanations for these regional sources include oil and gas extraction and production activities (NO X emissions) and v

7 animal feeding operations located south of Edmonton thru to the Alberta-Montana border and elsewhere (ammonia (NH 3 ) emissions). Levels of both secondary nitrate and sulfate particles tend to be simultaneously enhanced within plumes from coal combustion emissions relative to background. Again, while the analysis undertaken here is insufficient to accurately quantify the contribution to secondary nitrate from coal combustion sources west of Edmonton, their contribution is projected to be in the range of less than one-tenth to less than one-fifth of the secondary nitrate mass. Objective 3 Investigate origins and causes of PM 2.5 concentration differences in Edmonton during 2010 relative to other years A newspaper article circulated by the National Post newspaper in April 2015 reported that Edmonton had higher levels of a harmful air pollutant compared to Toronto, a city with five times the population and more industry. The article also stated that particulate matter (i.e., PM 2.5 ) exceeded legal limits of 30 µg/m 3 at two city monitoring stations on several winter days in 2010 through A limitation of PM 2.5 monitoring data presented for Edmonton in the National Post article is that it did not acknowledge changes in the operation of continuous PM 2.5 monitoring equipment during Equipment at many of the original continuous PM 2.5 monitoring stations in Alberta was upgraded to better capture some components of fine particulate matter (i.e., semi-volatile material) which were lost under the previous equipment operation methods. As a result of this change, PM 2.5 levels observed at air monitoring stations are higher in 2010 and subsequent years compared to previous years. For example, data comparing the newer and older 24 h PM 2.5 measurement method for 2010 showed good agreement during summer (slope = 1.06, R 2 = 0.99). While during winter the newer method measured 1.5 times (50%) higher concentrations than the older method. Analysis showed that a combination of changes in operation of continuous PM 2.5 monitoring equipment made during 2009 and a single major wildfire smoke event in 2010 (August 19 22) explain exceedances of 3-year averages of annual 98 th percentile 24 h average concentrations at the Edmonton central station. Results of the PMF model were examined further to identify how the contribution of identified sources (factors) for PM 2.5 at Edmonton McIntyre station differed in 2010 compared to other years in the analysis ( ) refer to Figure ES2. Several observations are made based on this examination: In order of relative importance, the secondary organic aerosol, secondary nitrate and biomass burning factors showed increased contributions during 2010 compared to The secondary sulfate factor showed no year-to-year variation over the 5-year period indicating that secondary sulfate precursor (i.e., SO 2 ) emissions influencing Edmonton McIntyre station were unchanged over the period. Other factors identified in the PMF analysis showed unimportant differences during 2010 compared to Multiple factors (i.e., secondary organic aerosol, secondary nitrate and biomass burning) other than secondary sulfate precursor emission sources showed increased contributions during 2010 compared to at Edmonton McIntyre station. Levels of both secondary nitrate and sulfate particles tend to be simultaneously enhanced within plumes from coal combustion emissions relative to background. Thus the observation of secondary sulfate displaying no year-to-year variation over the 5 year period provides vi

8 evidence that coal combustion emission sources would have played an unimportant role in explaining the year 2010 having a greater frequency of high PM 2.5 concentration events. On the other hand, increased contributions from secondary organic aerosol, secondary nitrate and biomass burning emission sources best explains the year 2010 having a greater frequency of high PM 2.5 concentration events. Increased factor contribution during 2010 relative to other years Figure ES2. Yearly average contributions of constrained PMF-derived sources at Edmonton McIntyre station for vii