Consultancy Proposal: Air Quality in the Tuggeranong Valley, ACT

Preliminary Assessment of Wintertime Air Quality in the TuggeranongValley, ACT

Prepared by

Howard Bridgman

Conjoint Associate Professor

School of Environmental and Life Sciences

University of Newcastle

Fellow, Clean Air Society of Australia and New Zealand

For

Director, Health Protection Service

ACT Health

December 21, 2009
EXECUTIVE SUMMARY

This report provides a summary of particle matter air quality in the TuggeranongValley, ACT, its measurement, the main sources and emissions, and its relationship to meteorology and the measurements at Civic and Belconnen. In general, the ACT is well-serviced by its three air quality monitoring stations, and by the quality of the equipment used for measurement. There are higher concentrations of PM on winter evenings in the TuggeranongValley, mainly due to solid fuel smoke from residential emissions.

Below are the recommendations for improving the quality of the measured data, for communication to the public, and for future planning:

Recommendation 1: The data sets created by the instruments need more attention. A revisit of the methods used to handle the measurement instruments and the monitoring data, with the objective to improve data availability and data quality, especially for the GRIMM and LoVol data sets, would be of considerable benefit. Major differences between the HiVol, TEOM and GRIMM winter measurements should be resolved. Good comparisons between measurements from different instruments will provide considerable flexibility and backup possibilities over time.

Recommendation 2: Evaluating the relationship between PM10 and meteorology for the other winter months when hourly PM10 is available, to compare with July 2007, is very important. If the relationship remains strong, ACT Government can use these data to issue daily forecast alerts to the public associated with Don’t Light Tonight and similar management campaigns.

Recommendation 3: Efforts to reduce the use of solid fuel burners in TV in winter have been successful and should continue. Natural gas can create a similar social cozy winter night environment, and is a much cleaner fuel. Wood pellets, substituted for wood logs, burn more efficiently and create minimal smoke emissions.

Recommendation 4: The NSW DECC AQI calculation for PM10 and PM2.5 should be adopted for Canberra, and daily information issuedbased on measurement locations that meet NEPM standards. Forecasts of expected poorer AQI can be included with BOM forecasts of low wind speeds, low temperatures, and high humidity, especially in winter, and especially for the Tuggeranong Valley.

Recommendation 5: Upgrading the Health Protection Service web site to include details about where, what, how and over what period of time air quality monitoring has been and is occurring, would be of benefit to the community.

Recommendation 6: Providing air quality measurements from all stations on-line, in real-time mode, along with AQI information, and if possible, weather measurements, would provide current information useful for the community.

Recommendation 7: Consideration be given to measuring particulate chemistry at the Monash site for a period of at least one winter to determine the composition of the PM. Comparison measurements for one other season would be useful.

1.0 Introduction

Residential wood smoke under stable winter conditions creates air quality problems in many cities and towns in Australia (Todd 2007), and also on the South Island of New Zealand (Wilton et al. 2009). Residents use wood heaters to keep warm on clear chilly winter nights. Low temperatures, low wind speeds and higher atmospheric moisture content do not allow the smoke from the wood heater emissions to disperse (Senarantne et al. 2005). The problem can be enhanced under hill and valley conditions, when overnight cold air drainage from the hill tops combine with heat loss from the earth’s surface creating an inversion (increase in temperature with height) (Sturman and Tapper 2006). This condition further traps air pollutants near the surface and prevents dispersion.

The purpose of this report is to provide an analysis of air pollution in the Tuggeranong Valley (TV), ACT, where in winter, air pollution can be a problem. The TV suffers from occasional high particulate matter (PM)concentrations in winter, due to residential burning of solid fuel for heating. Where appropriate, comparisons with measurements at Civic and Belconnen are made. Three air pollution monitoring sites, located from north to south across Canberra and the ACT, provide a good representative indication of air quality in the area.

The Tuggeranong Valley (TV) is located south of the City of Canberra in an area of rolling hills and valleys. It is one of the two fastest growing areas in the ACT in terms of population. Air quality, therefore, is an important environmental concern, especially in terms of human health.

This report will focus on particulate pollution, especially PM10 (particles less than 10 μm in diameter), PM2.5 (particles less than 2.5 μm in diameter, fine particles) and also include some information about PM1 (ultrafine particles) concentrations. Monitoring and meteorological data have been provided by Health Protection Service (HPS), ACT Health. The National Pollution Inventory (NPI)is also used to compare sources and emissions of particles matter. Analysis is by summary tables, graphs and simple statistics, including correlations where appropriate.

1.1 Objectives of the report

The following is a list of objectives to be met in this report.

1. To assess air quality and its trends in the ACT

This section provides information on changes in PM over time, and differences between measurement locations, using time series graphs and tables. Meteorology was used to assess conditions under which pollution episodes in the TV occur.

2. Compare TV sources and emissions of PM with that from other Australian jurisdictions

Comparison with Launceston, Tasmania, andBallarat, Victoria, examples of two locations with long-term winter air quality problems, provides a perspective for TV. These locations all have residential solid fuel smoke emissions problems.

3. Impacts of air quality protection measures and policy interventions (i.e Don’t Light Tonight)

The data are assessed to try to determine whether such policies have had any success.

4. Introduction of an Air Quality Index (AQI)

AQIs are used in many urban locations around Australia as a way to communicate the general level of air pollution to the public.

5. Potential health impacts in relation to air quality

Potential relevant health impacts (respiratory and cardiovascular) from hospital admissions are compared by spatial location, and comments are made about the methodologies needed for a detailed analysis of the relationship between health and air quality.

6. Advice on strategic directions for the future for the HPS Air Quality Monitoring

Some recommendations regarding future directions are included, especially for TV and Belconnen, the two fastest growing areas in the ACT.

2.0 Data availability, quality, and analysis methodology

Table 1. Data provided by Health Protection Service for this report.

Type of Data / Time Period* / Instrument / Data Available (% of Possible) / Comments and Location
PM1024 hours / April 1995 – March 2009 / High Volume Air Sampler (HiVol) / 85 / Measurements every six days
Monash
Civic
PM10 Hourly / Feb 27 2002 – July 1, 2009 / ThermoElectric Oscillating Microbalance (TEOM) / 91 / A small number of negative values
Monash
PM2.5 24 hours / Jan 1, 2004 – June 3, 2009 / Low Volume Air Sampler (LowVol) / 69 / Low data availability
Monash
PM1, PM2.5, PM10 Hourly / Dec 27, 2005 – Oct 29, 2009
Jan 19, 2006- Oct 30, 2009
Jan 20, 2006 –Oct 30, 2009 / GRIMM Size Distribution Sampler / 73 (Monash)
77 (Civic)
76 (Belconnen) / Variable data availability,
Monash
Civic
Belconnen
Meteorology / Jan 1, 2006 – Oct 30, 2009 / Bureau of Meteorology standards / 100 / Temperature
Wind
Relative Humidity
Monash

* Represents the day or hour before the stated time

Table 1 presents the data sets used in this analysis plus a summary of data availability. The measurement site in the TV was originally at Gowrie (1995-1999), but was moved to Monash in May 1999. Comparative measurements for daily PM10 are available from Civic, and for the GRIMM data from both Civic and Belconnen. Otherwise, all data was measured at Monash. Currently, only the Monash site meets National Environmental Protection Measure (NEPM) requirements. However, data from Civic and Belconnen can be used for spatial and temporal comparison.

The percentage of data available meets the NSW Department of Environment and Climate Change (NSW DECC) guideline of 85% for Daily PM10 (HiVol) and Hourly PM10 (TEOM). 24-hour PM10 HiVol measurements occur only every six days, which is less than 20% of possible days. While this meets NSW DECC monitoring requirements, it creates severe limitations to more detailed analysis, but can be used as a check against GRIMM and other measurements.

For the other data sets, the amount of missing data creates limitations in how the data could be used. Hourly measurements of PM1, PM2.5 and PM10 by the GRIMM instruments provide the most interesting comparison possibilities. A preliminary quality assurance of these data was completed by subtracting PM2.5 from PM10 measurements. If the hourly result was negative, then these data were considered to be suspect, and where possible, were eliminated from the data set.

An acceptable assessment of trends requires at least ten years of data from one location. In terms of the objectives of this report, HiVol PM10 was used to illustrate trends. Despite the every-six-day limitation, this was the longest data set in terms of time. All other data sets were less than five years old, and were not considered long enough to establish trends. For the more detailed assessments, the GRIMM data sets provided the base data, and the other data sets were used where appropriate for comparison. The author has familiarity with the quality of GRIMM instruments from a comparison of measurement instruments in the HunterValley, NSW (Bridgman et al. 2005)

For Objective 2, sources and emissions data from the National Pollution Inventory ( were used for Tuggeranong, Launceston, and Ballarat.

For purposes of standard comparison, the Federal Government National Environmental Protection Standard (NEPM) for PM10, 50 μg m-3over one day, was used. For PM2.5, the current advisory standard of 25 μg m-3over one day was used(

3.0 Assessment of Air Quality and its Trends

3.1Hourly measurement comparisons

The GRIMM data provides the base data for hourly measurement comparisons, with the main focus on winter PM concentrations. Due to problems with missing or incomplete data (see Table 1), and also the size of the data set, two winter and two summer seasons were chosen for analysis. These are Summer 2006-2007, Winter 2007, Summer2007-2008, and Winter 2008. Summer analyses are included to provide some comparisons with winter concentrations. The seasons are defined by standard calendar months.

For detailed evaluation within these seasons, analysis was by months: February 2007 and 2008 (summer), and July 2007 and 2008 (winter). These months had a high level of data availability and data quality, with the exception of Belconnen for Winter 2008.

Examples of PM concentrations for winter evening episodes are also presented.

Table 2 presents summary averages for the GRIMM data sets used in this analysis. There are three important results from Table 2. First, winter values for all PM pollutants are much higher at Monash compared to the other two stations, especially in winter 2008. This suggests an extra local source of pollution exists in the TV compared to the the other two locations.

Second, summer concentrations are similar at all sites, although the averages differ between seasons. This suggests that the Canberra air quality in summer is subject to similar sources, no matter what the location.

Third, the percentages of PM2.5 in PM10 and PM1 in PM2.5 average 90% for wintertime at Monash, compared 56-82% at the other two stations and in summer at all stations. These high percentages suggest that the main source of particles is from burning processes (Scott and Sturman 2006), creating fine particles, rather than non-burning sources such as wind-blown dust.

Table 2. PM10, PM2.5 and PM1 seasonal averages (ug m-3) for the Grimm data sets used in this report, including the percentage of PM2.5 compared to PM10, and PM1 compared to PM2.5.

Season:
Data Type / Sum 2006-07 / Winter 2007 / Sum 2007-08 / Winter 2008
Monash PM10 / 15.9 / 17.8 / 9.8 / 25.4
PM2.5 / 11.3 / 16.0 / 6.7 / 23.2
PM1 / 8.7 / 14.4 / 4.7 / 21.1
PM10-PM2.5 / 4.6 / 1.9 / 3.1 / 2.2
PM2.5-PM1 / 2.5 / 1.5 / 1.2 / 2.1
%PM2.5 in PM10 / 71 / 89 / 71 / 91
%PM1 in PM2.5 / 77 / 90 / 69 / 91
Civic PM10 / 13.4 / 7.8 / 8.5 / 6.2
PM2.5 / 8.7 / 4.8 / 4.8 / 4.1
PM1 / 6.7 / 3.5 / 2.9 / 3.1
PM10-PM2.5 / 4.7 / 3.0 / 3.7 / 2.0
PM2.5-PM1 / 2.0 / 1.3 / 1.9 / 1.0
%PM2.5 in PM10 / 65 / 61 / 56 / 66
%PM1 in PM2.5 / 77 / 74 / 61 / 75
Belcon PM10 / 18.4 / 8.7 / 5.8 / NA
PM2.5 / 11.0 / 7.0 / 4.2 / NA
PM1 / 8.2 / 5.7 / 2.9 / NA
PM10-PM2.5 / 7.4 / 1.7 / 1.6 / NA
PM2.5-PM1 / 2.8 / 1.2 / 1.3 / NA
%PM2.5 in PM10 / 60 / 80 / 72 / NA
%PM1 in PM2.5 / 75 / 82 / 70 / NA

3.1.1 Detailed summer comparisons

Figure 1. Scatter plot of hourly PM2.5 for Monash compared to Civic and Belconnen, February 2008. The line of fit and the very high R2 value shows a highly significant relationship, meaning very similar concentrations for all stations.

______

Figure 1 represents the correlations between hourly PM2.5 measured at all three locations in February 2008. Scatter plots for PM1 and PM10 are very similar, for both summer seasons. The line of fit and the highly significant correlation (R2 = 0.785) emphasises on a detailed scale that there are only small differences in PM concentrations between the three locations in summer.

Figure 1 also shows that there are only a few scattered hours when PM2.5 concentrations exceed the daily standard of 15 μg m-3. The result for PM10 is similar (compared to 50 μg m-3). No daily average would exceed either standard. The conclusion is that summer PM concentrations in Canberra are low and currently not of major concern. On individual days, higher concentrations most likely occur due to bushfires and dust storms.

3.1.2 Detailed winter comparisons

The winter PM concentrations are a different story, and more complex. In this section the analysis is divided into PM10, PM2.5 and PM1, with comparative results from other data sets included for PM10 and PM2.5 (see Table 1) where appropriate.

a. PM10: As a data quality check the hourly data for July 2007 and 2008 for the GRIMM was compared to that for the TEOM. Figure 2 shows the results representing each season. In both cases, the GRIMM data measurements are at least twice that of the TEOM when concentrations exceed 25 μg m-3, and GRIMM measurements above 50 μg m-3 exceed the TEOM by at least 4 times, especially in 2008. TEOM concentrations measured in the HunterValley (Bridgman et al. 2005) were also lower than the GRIMM results, although not to this degree. In that study, problems with the TEOM suggested that the instrument may not be as reliable as the GRIMM. As well, the GRIMM is more sensitive to changes in concentrations.

Correlations between PM10 data for Monash and that for Civic and Belconnen are presented in Figure 3 for July 2007 and July 2008. Higher concentrations (greater than 50 μg m-3) dominate at Monash. Almost no concentrations at this level occur at Civic. In July 2007 at Belconnen, there are several hours when PM10 reaches 50 μg m-3but the number is still considerably lower than at Monash.

Figure 4 is a time series plot of Monash hourly PM10 concentrations greater than 50 ug m-3 for both July periods. There is a wide variety in the number of hours in each episode. There are a higher number of episodes, and the concentrations of PM10 are higher, in 2008.

While the use of 50 μg m-3 is a convenient method to highlight hours where PM10 episode are occurring, it is important to emphasise that this PM10 standard is applicable to daily averages. The standard is exceeded on no July days in 2007, and on only three days in July 2008 (July 6 – 66.9; July 16 – 57.9; July 17 – 78.8 μg m-3). This result suggests that, despite several hours of higher PM10 during many winter evenings, exceedences of the daily PM10 standard are rare.

Figure 2. Monash PM10 measurements, GRIMM vs TEOM for July 2007 and July 2008. While in both cases the data are highly correlated, the GRIMM measurements are at least 2 times higher than those by the TEOM when concentrations exceed 25 ug m-3.

Figure 3. Monash hourly GRIMM PM10 compared to PM10 from Civic and Belconnen (missing for 2008) for July 2007 and 2008. The frequency of concentrations greater than 50 μg m-3 is much higher at Monash than the other two locations.

Figure 4. Time series of hourly PM10 for July 2007 and July 2008 at Monash, during hours when concentrations are greater than 50 μg m-3. Each dot represents one hour. The number of episode periods, and the strength of some episodes, is greater during July 2008.

b. PM2.5: As shown in Table 1, winter concentrations of PM2.5 make up approximately 90% of PM10. Since PM2.5 has been recognised as a potential respiratory and heart health problem (Pope and Dockery 2006), consideration of levels in the TV winter atmosphere is important.

Figure 5 shows scatter plots of PM2.5 measured at Monash compared to Civic and Belconnen for July 2007 and July 2008. There are many hours where the daily advisory standard of 25 μg m-3 is exceeded at Monash, but very few at Belconnen, and none at Civic.

LoVol daily data for PM2.5 was not available in July 2007. In 2008, these data showed only 3 days when the PM2.5 average exceeded the advisory standard, June 18, July 6, and July 27. Compared to the three PM10 exceedence dates (see above), the only match is July 6. Overall, the LoVol results are considerably lower that daily average PM2.5 from the GRIMM measurements, a similar problem to the comparison of TEOM and GRIMM PM10 data.

Figure 6 demonstrates that the temporal patterns for PM2.5 in July 2007 and July 2008 match those for the PM10 data (Figure 4) with minor exceptions. There are a considerable number of hours when the GRIMM PM2.5 measurements exceed the PM2.5 advisory standard. This further indicates the source of fine particles is burning.

c. PM1: PM1 is the particle pollution concern of the future. These very small particles can easily penetrate deep into the lung, causing respiratory problems for sensitive people (Pope and Dockey 2006). The main source from human activities is burning processes Air pollution experts are beginning to study PM1 seriously. There are no standards established yet, as the overall health impacts have not been fully established.