pollution reduction program 4.2 particulate … coal dust reports...katestone environmental pty ltd...

Pollution Reduction Program 4.2 Particulate Emissions from Coal Trains Prepared for

Australian Rail Track Corporation Pty Ltd May 2013

Final Prepared by Katestone Environmental Pty Ltd ABN 92 097 270 276 Ground Floor, 16 Marie St PO Box 2217 Milton, Queensland, Australia 4064

www.katestone.com.au [email protected]

Ph +61 7 3369 3699 Fax +61 7 3369 1966

Document Quality Details

Job Number: 12048

Deliverable Number: D12048-15

Title: Pollution Reduction Program 4.2 Particulate Emissions from Coal Trains

Client: Australian Rail Track Corporation Pty Ltd

Document reference: D12048_15_ParticulateMonitoringProgram_Finalv1.1.docx

Prepared by: Natalie Shaw, Tania Haigh, Frank Quintarelli and Andrew Vernon

Reviewed by: Simon Welchman Revision Date Approved Signature

Rev 1.1 30/05/2013 SW

Disclaimer This document is intended only for its named addressee and may not be relied upon by any other person. Katestone Environmental Pty Ltd disclaims any and all liability for damages of whatsoever nature to any other party and accepts no responsibility for any damages of whatsoever nature, however caused arising from misapplication or misinterpretation by third parties of the contents of this document. This document has been prepared with all due care and attention by professional scientists and engineers according to accepted practices and techniques. This document is issued in confidence and is relevant only to the issues pertinent to the subject matter contained herein. Katestone Environmental accepts no responsibility for any misuse or application of the material set out in this document for any purpose other than the purpose for which it is provided. Where site inspections, testing or fieldwork have taken place, the report is based on the information made available by the client, their employees, agents or nominees during the visit, visual observations and any subsequent discussions with regulatory authorities. The validity and comprehensiveness of supplied information has not been independently verified except where expressly stated and, for the purposes of this report, it is assumed that the information provided to Katestone Environmental Pty. Ltd. is both complete and accurate. Copyright This document, electronic files or software are the copyright property of Katestone Environmental Pty. Ltd. and the information contained therein is solely for the use of the authorised recipient and may not be used, copied or reproduced in whole or part for any purpose without the prior written authority of Katestone Environmental Pty. Ltd. Katestone Environmental Pty. Ltd. makes no representation, undertakes no duty and accepts no responsibility to any third party who may use or rely upon this document, electronic files or software or the information contained therein. Copyright Katestone Environmental Pty. Ltd.

Katestone Environmental Pty Ltd D12048-15 Australian Rail Track Corporation Pty Ltd

May 2013 Page i

Contents Executive Summary ...................................................................................................................... vi

1. Introduction ....................................................................................................................... 1

2. Work program ................................................................................................................... 3

3. Monitoring Methodology ................................................................................................ 4

3.1 Objective and overview of methodology ........................................................ 4 3.2 Particulate monitoring equipment ..................................................................... 4 3.3 Meteorological monitoring equipment ............................................................. 5 3.4 Monitoring location .............................................................................................. 5 3.5 Train monitoring .................................................................................................... 6 3.6 Monitoring period ................................................................................................. 6

4. Data management .......................................................................................................... 7

4.1 Data capture ........................................................................................................ 7 4.2 Train data capture ............................................................................................... 9 4.3 Data validation ................................................................................................... 10 4.4 Algorithms for combining train, meteorological and pollution data .......... 10

4.4.1 Synchronisation of data......................................................................... 10 4.4.2 Determination of train type ................................................................... 10 4.4.3 Train pass-by time ................................................................................... 11

4.5 Particulate averaging period ........................................................................... 13

5. Assessment Methodology ............................................................................................. 14

5.1 Calculation of concentration ........................................................................... 14 5.2 Assessment of contribution by train type ........................................................ 14 5.3 Assessment of contribution by other variables ............................................... 15

6. Results ............................................................................................................................... 16

6.1 Train movements ................................................................................................ 16 6.2 Particulate matter concentration by train type ............................................ 16 6.3 Variations in concentration with train speed ................................................. 19 6.4 Variations in concentration with wind direction ............................................ 19 6.5 Variations in concentrations with wind speed ............................................... 20 6.6 Variations in concentrations minus background ........................................... 20 6.7 Influence of rainfall on particulate concentrations....................................... 23

7. Comparison with Pilot Study ......................................................................................... 25

8. Limitations ........................................................................................................................ 26

9. Conclusions ..................................................................................................................... 27

10. References ...................................................................................................................... 29

Appendix A Data validation ................................................................................................... 51


May 2013 Page ii

A1 Full scale deflection ........................................................................................... 51 A2 Corrupt data ....................................................................................................... 52 A3 Elevated concentrations ................................................................................... 53

Appendix B Statistical analysis of concentrations ............................................................... 57

B1 Variations in concentrations with train speed, wind direction and wind speed ................................................................................................................... 57

Tables Table 1 Data capture rate for TSP, 30 November 2012 up to and including

29 January 2013 ....................................................................................................... 7

Table 2 Data capture rate in days for TSP, PM10 and PM2.5, 30 November 2012 up to and including 29 January 2013 ......................................................................... 9

Table 3 Data capture rate by hour of the day (%) .......................................................... 9

Table 4 Data capture rate by day of the week (%) ......................................................... 9

Table 5 Total number of trains during the PRP 4.2 monitoring period .......................... 10

Table 6 Example of train data received from ARTC and train identification ............. 11

Table 7 Example of train data received from ARTC and train characteristics ........... 12

Table 8 Number of 6 second average particulate records associated with each train type for a single pass-by and multiple pass-bys used in the analysis of particulate concentrations ............................................................................. 16

Table 9 Particulate concentrations by train type ........................................................... 18

Table 10 Particulate concentrations by train type minus background (µg/m³) .......... 22

Table 11 Rainfall recorded at Maitland Visitors Centre ................................................... 23

Table 12 PM10 and PM2.5 24-hour average concentrations recorded during the pilot monitoring program and during the PRP 4.2 monitoring program (µg/m³) .................................................................................................................... 24

Table 13 Average PM10 concentrations recorded at the ARTC Metford site during the two study periods (µg/m³) ............................................................................. 25

Table A1 Data removed from the raw data set due to full scale deflection................ 51

Table A2 Data removed due to data corruption ............................................................. 52

Table B1 Particle concentrations for all trains by train speed ......................................... 58

Table B2 TSP concentration for all data (trains and no trains) by wind direction ........ 59

Table B3 PM10 concentration for all data (trains and no trains) by wind direction ...... 59

Table B4 PM2.5 concentration for all data (trains and no trains) by wind direction ..... 60


May 2013 Page iii

Table B5 Concentration by train type when wind direction is between 150 degrees to 300 degrees ....................................................................................... 61

Table B6 Particle concentration for all trains by wind speed .......................................... 62 Figures

Figure 1 Site plan showing Metford monitoring location ................................................ 30

Figure 2 Metford particulate and meteorological monitoring station ......................... 31

Figure 3 Comparison of average TSP concentrations of loaded and unloaded coal trains in the 8 minutes prior to and 8 minutes after passing the trackside monitor ................................................................................................... 32

Figure 4 Comparison of TSP concentrations by train type ............................................. 33

Figure 5 Comparison of PM10 concentrations by train type ........................................... 34

Figure 6 Comparison of PM2.5 concentrations by train type .......................................... 35

Figure 7 Comparison of TSP concentration upper and lower 95% confidence limits by train type ........................................................................................................... 36

Figure 8 Comparison of PM10 concentration upper and lower 95% confidence limits by train type .................................................................................................. 37

Figure 9 Comparison of PM2.5 concentrations upper and lower 95% confidence limits by train type .................................................................................................. 38

Figure 10 Comparison of TSP concentrations, 95th, 75th, 25th, and 5th percentiles by train type ................................................................................................................ 39

Figure 11 Comparison of PM10 concentrations, 95th, 75th, 25th, and 5th percentiles by train type ................................................................................................................ 40

Figure 12 Comparison of PM2.5 concentrations, 95th, 75th, 25th, and 5th percentiles by train type ................................................................................................................ 41

Figure 13 Concentrations of TSP, PM10 and PM2.5 by wind direction ................................ 42

Figure 14 Comparison of TSP concentrations upper and lower 95% confidence limits by train type minus background ............................................................... 43

Figure 15 Comparison of PM10 concentrations upper and lower 95% confidence limits by train type minus background ............................................................... 44

Figure 16 Comparison of PM2.5 concentrations upper and lower 95% confidence limits by train type minus background ............................................................... 45

Figure 17 Comparison of TSP “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type ...................................................................................... 46

Figure 18 Comparison of PM10 “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type ...................................................................................... 47

Figure 19 Comparison of PM2.5 “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type ...................................................................................... 48

Figure 20 Daily rainfall during the study period (mm) ....................................................... 49

Figure 21 Daily rainfall during the pilot monitoring program (mm) ................................. 50


May 2013 Page iv

Figure A1 Example of data validation: example of full scale deflection limit, data on dashed line was removed from analysis ...................................................... 54

Figure A2 Example of data validation: data between the dashed lines was excluded from the analysis .................................................................................. 55

Figure A3 Timeseries of TSP and PM10 during 10 January 2013 ......................................... 56

Figure B1 TSP concentrations by train type and train speed ........................................... 63

Figure B2 PM10 concentrations by train type and train speed ......................................... 64

Figure B3 PM2.5 concentrations by train type and train speed ........................................ 65

Figure B4 Comparison of TSP concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees .... 66

Figure B5 Comparison of PM10 concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees ................................................................................................................... 67

Figure B6 Comparison of PM2.5concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees ................................................................................................................... 68

Figure B7 Concentrations of TSP, PM10 and PM2.5 by train type for wind directions between 150 degrees and 300 degrees ........................................................... 69

Figure B8 TSP concentrations by train type and wind speed ........................................... 70

Figure B9 PM10 concentrations by train type and wind speed ........................................ 71

Figure B10 PM2.5 concentrations by train type and wind speed ....................................... 72


May 2013 Page v

Glossary

Term Definition µg/m3 micrograms per cubic metre

µm microns km kilometre

km/hr kilometre per hour m metre

m/s metres per second NE northeast NW northwest PM10 Particulate matter with a diameter less than 10 micrometres PM2.5 Particulate matter with a diameter less than 2.5 micrometers SE southeast SW southwest TSP Total suspended particulates

Abbreviations ARTC Australian Rail Track Corporation Ltd BoM Bureau of Meteorology EPA Environmental Protection Authority PRP Pollution Reduction Program NSW New South Wales OEH Office of Environment and Heritage


May 2013 Page vi

Executive Summary

Katestone Environmental Pty Ltd (Katestone) has been commissioned by Australian Rail Track Corporation Ltd (ARTC) to conduct monitoring of particulate matter on its behalf in accordance with the Pollution Reduction Program (PRP) 4.2 – Particulate Emissions from Coal Trains. The objective of PRP 4.2 was to determine whether:

- Trains operating on the Hunter Valley rail network are associated with elevated particulate matter concentrations; and

- Loaded coal trains operating on the Hunter Valley rail network have a stronger association with elevated particulate matter concentrations than unloaded coal trains or other trains on the network (and by inference contributing to ambient rail corridor particulate levels).

To achieve the objective of PRP 4.2, a continuous particulate monitoring station was installed to measure particulate levels in the rail corridor adjacent to tracks carrying various types of trains. The time that the train passed the monitoring station was recorded in addition to other details of each train so that any change in particulate levels associated with the train pass-by could be accurately quantified. Monitoring of particulate matter using an OSIRIS and meteorological conditions was conducted at Metford, NSW, from 30 November 2012 to 29 January 2013. Particulate concentration data was grouped by train type and a statistical analysis was used to determine the change in particulate concentration associated with each train type and the significance of the change relative to the no trains group and relative to each train type. The influence of meteorological conditions on measurements was also investigated. This report presents the monitoring methodology, analysis methodology and results of monitoring conducted in accordance with the work program. The findings of monitoring program are as follows:

• Passenger trains were not associated with a statistically significant difference in TSP, PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station

• Freight trains were not associated with a statistically significant difference in TSP,

PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station.

• Loaded coal trains were not associated with a statistically significant difference in

PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station. However, loaded coal trains were associated with a statistically significant difference in TSP concentrations when compared with concentrations when no train was passing the monitoring station.

• Unloaded coal trains were associated with a statistically significant difference in TSP, PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station.


May 2013 Page vii

• Average concentrations of TSP associated with loaded coal trains, unloaded coal trains and freight trains were higher by 3.2 µg/m3, 6.1 µg/m3 and 4.5 µg/m3, respectively compared to when no train passes the monitoring station.

• Average concentrations of PM10 associated with loaded coal trains, unloaded coal

trains and freight trains were higher by 2.3 µg/m3, 4.5 µg/m3 and 3.0 µg/m3, respectively compared to when no train passes the monitoring station.

• Average concentrations of PM2.5 associated with loaded coal trains, unloaded coal


• Loaded coal trains operating on the Hunter Valley rail network, when measured at

Metford, did not have a statistically stronger association with elevated TSP, PM10 and PM2.5 concentrations than other trains.

• There was a statistically significant difference in concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and passenger trains. However, there was no statistically significant difference in concentrations of TSP, PM10 and PM2.5 between the other train types.

• There was no increasing or decreasing trend in average concentrations with respect to train speed.

• When the wind direction was between 150 degrees and 300 degrees (blowing from

the rail tracks towards the monitoring station) there was a statistically significant difference in the average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains and unloaded coal trains compared with concentrations recorded when no trains were passing the monitoring station.

• There was a statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and passenger trains when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• There was no statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between freight trains and passenger trains when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• There was a statistically significant difference in average concentrations of TSP between loaded coal trains and passenger trains; however, there was no statistically significant difference in average concentrations of PM10 and PM2.5 between loaded coal trains and passenger trains when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• Average concentrations of TSP, PM10 and PM2.5 associated with unloaded coal trains were higher than concentrations when no train was passing the monitor by 10.1 µg/m3, 7.6 µg/m3 and 2.1 µg/m3, respectively, when wind direction was from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 23%, 24% and 21% for TSP, PM10 and PM2.5, respectively.


May 2013 Page viii

• Average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains were higher than concentrations when no train was passing the monitor by 6.0 µg/m3, 4.3 µg/m3 and 1.1 µg/m3, respectively, when wind direction was blowing from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 14%, 14% and 11% for TSP, PM10 and PM2.5, respectively.

• There was no increasing or decreasing trend in average concentrations with respect to ambient wind speed.

Limitations of the PRP 4.2 study are as follows:

• In general, the light scattering photometers that are available to measure continuous particulate concentrations cannot measure a true TSP size fraction.

• Recording of 6 second average data was a considerable challenge for the Osiris, which has limited on-board data storage that affected the Osiris’s functionality and data recording capacity during download.

• The Metford particulate monitor is located at unequal distances from the coal and main rail lines.

• Trains with the commodity label “Unknown” were omitted from the analysis. • Pass-by duration was determined by the speed and train length recorded by the

wayside monitor. This assumes constant train speed and does not account for acceleration or deceleration of the train. In cases of accelerating and decelerating trains, this may lead to a minor error in train pass-by time.


May 2013 Page 1

1. Introduction

Katestone Environmental Pty Ltd (Katestone) has been commissioned by Australian Rail Track Corporation Ltd (ARTC) to conduct monitoring of particulate matter on its behalf in accordance with the Pollution Reduction Program (PRP) 4.2 – Particulate Emissions from Coal Trains The key actions required by PRP 4.2 are as follows: Action 4.2A The licensee will conduct a monitoring program to determine whether:



Timeframe – For a continuous period of at least 30 days between 1 November 2012 and 29 January 2013. Action 4.2B A revised work program for the monitoring program must be submitted to the EPA. Monitoring must be consistent with that work program. The revised work program must be based on the previous EPA approved work plan submitted by ARTC in compliance with Action 4.1B above from PRP 4.1, and account for the learnings from and limitations identified with the pilot program. The licensee will also publish the work program on its website. Timeframe – The licensee must submit the work program to the EPA within four weeks of inclusion of the PRP on the licence. The licensee must publish the work program on its website within four weeks of the EPA providing written comment to the licensee on the draft work program. Action 4.2C On completion of the monitoring program required by 4.2A the licensee will submit to the EPA for comment a report on the results of the monitoring. Timeframe – On or prior to 15 March 2013. Action 4.2D The licensee will submit a final report on the monitoring program required by 4.2A to the EPA. The report must address any comments provided by the EPA and include a “plain English” Executive Summary at the start of the document. The licensee will also publish the final report in full on its website. Timeframe – The report will be submitted to the EPA and published on the licensee’s website within four weeks of the EPA providing written comment to the licensee on the draft report. In response to Action 4.2B, Katestone developed a work program to guide the monitoring program in response to PRP 4.2 (Katestone 2012). The work program was submitted to the


May 2013 Page 2

EPA for comment. A final version of the work program was made available on ARTC’s website. This report presents the monitoring methodology, analysis methodology and results of monitoring conducted in accordance with the work program. Monitoring of particulate matter and meteorological conditions was conducted at Metford, NSW, from 30 November 2012 to 29 January 2013.


May 2013 Page 3

2. Work program

In accordance with Action 4.2B of PRP 4.2, a work program was developed to guide monitoring required under PRP 4.2. The work program was required to be based on the work program prepared for PRP 4.1 and to account for the learnings from and limitations identified during the pilot monitoring program (Environ, 2012) conducted at Mayfield and Metford. The following learnings from the pilot program were accounted for in the work program for PRP 4.2:

• Results of monitoring conducted at Mayfield were unreliable due to: o The lack of a wayside monitor and; hence, the use of the 4TRAK monitoring

system to identify train pass-bys results in a likely misalignment of the times of particulate measurements and train pass-bys. In particular, the 4TRAK system suffers from the following inadequacies: It cannot provide train lengths It cannot provide train speeds It records pass-by time rounded to the nearest minute

o Slower train speeds of coal trains entering the port and queuing of trains at Mayfield results in few single pass-bys and; hence, a relatively small number of data points could be included in the analysis

• An averaging period for particulate concentration that is shorter than 30 seconds would improve the accuracy of measurements of all trains and particularly those that have short pass-by times. This shorter averaging period would only be of value where accurate train pass-by information was also available

• To achieve greater data resolution, data would need to be downloaded remotely on a frequent basis

The Work Program specified the monitoring and assessment methodology, the statistical techniques to be used, reporting requirements as well as limitations and assumptions for the PRP 4.2 program in light of the findings of the pilot monitoring program.


May 2013 Page 4

3. Monitoring Methodology

3.1 Objective and overview of methodology

The objective of the work program was to specify the work that would be completed to address the requirements of PRP 4.2. The work program was used to guide the conduct of the monitoring program and in particular:

• Monitoring methodology • Assessment methodology and statistical techniques • Reporting requirements • Limitations and assumptions • Publication

The work program was based on the previous EPA approved work plan submitted by ARTC in compliance with Action 4.1B from PRP 4.1. The work program accounted for the learnings from and limitations identified in the pilot program. To achieve the objective of PRP 4.2, a continuous particulate monitoring station was installed at one location to measure particulate levels in the rail corridor adjacent to tracks carrying various types of trains. The time that the train passed the monitoring station was recorded in addition to other details of each train so that any change in particulate levels associated with the train pass-by could be accurately quantified. Particulate concentration data was grouped by train type and a statistical analysis was used to determine the change in particulate concentration associated with each train type and the significance of the change relative to the no trains group and relative to each train type. The influence of meteorological conditions on measurements was also investigated. 3.2 Particulate monitoring equipment

Particulate monitoring was conducted using an Osiris, a light scattering laser photometer that is capable of measuring at a frequency of once every 6 seconds. The following mass fractions were measured and analysed:

• Total Suspended Particulates (TSP) • Particulates with aerodynamic diameter less than 10 micrometres (PM10) • Particulates with aerodynamic diameter less than 2.5 micrometres (PM2.5)

The Osiris monitor is not a compliance monitor as it does not satisfy the requirements of the USEPA Federal Reference or Equivalence Method. Consequently, the results obtained are not suitable for comparison against the standards specified in the National Environmental Protection (Ambient Air Quality) Measure 1998 (NEPM). However, the purpose of PRP 4.2 was to determine the relative contributions of passing trains to particulate matter concentrations. Given the relative pass-by times of trains on the Hunter Valley rail network, from seconds to around 2 minutes, the Osiris monitor was chosen because it has the capability to measure at the requisite time resolution. Whereas, the available USEPA Federal Reference or Equivalence Methods can achieve a time resolution of at best 5-10 minutes and up to 24 hours.


May 2013 Page 5

3.3 Meteorological monitoring equipment

Wind speed and wind direction was monitored using a weather station and datalogger. The anemometer was mounted at the same height as the sample inlet of the Osiris monitor. The meteorological monitoring equipment was capable of measuring the following parameters at a frequency of 1 second:

• Wind speed at a height of 4 metres • Wind direction at a height of 4 metres

3.4 Monitoring location

Selection of the monitoring location was based on the pilot program work plan (Environ, 2011) and findings of the pilot program (Environ, 2012). The pilot program work plan identified two preferred monitoring locations following a detailed site selection study. Monitoring was conducted at both locations, namely: Mayfield and Metford, during the pilot monitoring program. Given the constraints identified in conducting the pilot program and the findings, monitoring during the PRP 4.2 study program was not conducted at Mayfield. This decision was made based on the following findings during the Mayfield pilot monitoring program (Environ, 2012):

• Relatively few (44%) pass-bys were single trains. This meant that 66% of trains passing the Mayfield monitor during the pilot monitoring program were excluded from the analysis

• Train speeds were relatively slow indicating that the monitoring site was not representative of the operating conditions elsewhere in the network

• Significant difficulties in relating train pass-bys to monitoring data. This was because the system of recording train pass-bys (known as 4TRAK) could not record pass-bys to the nearest second. The pass-by time had to be assumed and resulted in poor accuracy in relating train type to particulate concentrations at the Mayfield monitoring location.

The Mayfield site was not replaced in this follow-up study because there is no existing alternative site in the Lower Hunter with a wayside monitor that would allow the learning's from the pilot program to be addressed. Additionally, there are significant cost and time implications to setup a wayside monitor for the purpose of this follow-up study. Further, the original study was a pilot and aimed to address issues that had not previously been investigated or considered. Given the flaws of the Mayfield location and the strengths of the Metford location (detailed below) only a single monitoring site, at Metford, was selected for the PRP 4.2 study. There were a number of substantial advantages of the Metford site that were evident from the pilot monitoring program (Environ, 2012):

• The site is located so it is passed by trains travelling on both the Hunter Valley and North South rail lines

• Relatively few multiple pass-bys occurred (8%) • Good distribution of train types • Greater number of coal train pass-bys and fewer multiple pass-bys • The Metford site uses train monitoring devices (wayside monitoring) which are able

to record train pass-by to the nearest second. This made pairing with the Osiris particulate concentrations more accurate.


May 2013 Page 6

The disadvantages of this site were noted in the work plan for the pilot program (Environ, 2011). The disadvantages are not likely to compromise the achievement of the objective of PRP 4.2. The monitoring site at Metford was located in the rail corridor, adjacent to the four rail tracks, at the same location as was used for the pilot study. The location is shown in Figure 1. A photograph of the site is shown in Figure 2. The particulate monitoring equipment and anemometer were mounted at approximately 4 metres above track height and a horizontal distance of approximately 4 metres from the nearest track and 6 metres from the edge of the rail corridor. The distance of the monitor to each track in the rail corridor is as follows:

• 4 metres - UP MAIN (UM) • 8.5 metres - DOWN MAIN (DM) • 13 metres - UP COAL (UC) • 17.5 metres - DOWN COAL (DC)

TSP and the finer particle size fractions will remain suspended for many tens of metres downwind of the emission source. Hence, the distance from each of the tracks to the Osiris monitor is unlikely to result in substantially different concentrations of TSP, PM10 and PM2.5. 3.5 Train monitoring

The Metford monitoring site was collocated with a fixed wayside monitoring station for logging train movements. Data from the wayside monitoring station was used to identify:

• Train consists/train types • Pass-by time • Train speed

3.6 Monitoring period

PRP 4.2 required monitoring to be conducted for a continuous period of at least 30 days. The monitoring equipment was installed on the 30 November 2012 and decommissioned on 29 January 2013, resulting in a monitoring period of 62 days. This was extended over the required 30 days due to poor data capture early in the study. The reasons for poor data capture are discussed in Section 4.


May 2013 Page 7

4. Data management

4.1 Data capture

The data capture rates for TSP for data collected during the monitoring period is presented in Table 1. Data loss occurred throughout the monitoring program due to the following:

• Storm activity in the area between 30 November and 2 December 2012 caused damage to the data logger and modems

• Loss of communications with the monitoring site requiring equipment reset at the site • Recording of 6 second average data was a considerable challenge for the Osiris,

which has limited on-board data storage that affected the Osiris’s functionality and data recording capacity during download

• Manual data download was required every 27 hours with 1 hour of data loss during download

There were 47 days when the data capture rate was over 50%. Of these days, 35 days had a data capture rate over 70%. Capture rates for PM10 and PM2.5 were very similar to TSP. Overall there was a 68% capture rate of TSP over the monitoring period. This equated to 42 whole days of data out of the 62 day period of monitoring (Table 2). The capture rate by hour of the day and day of the week is presented in Table 3 and Table 4, respectively. The tables show that data capture was relatively evenly spread across the hours of the day, with no single hour having a significantly lower data capture rate. A similar picture is seen in the capture rate with day of week, with no one day having particularly lower data capture. Hence, it has been concluded that the data loss during the project has not caused the outcomes to be biased in any particular way.

Table 1 Data capture rate for TSP, 30 November 2012 up to and including 29 January 2013

Date Capture rate Number of 6 second average records

30-Nov 64.3% 9255 1-Dec 46.6% 6713 2-Dec 0.0% 0 3-Dec 53.4% 7696 4-Dec 53.6% 7723 5-Dec 0.0% 1 6-Dec 44.2% 6372 7-Dec 68.9% 9925 8-Dec 13.7% 1968 9-Dec 88.9% 12797 10-Dec 73.1% 10520 11-Dec 58.3% 8401 12-Dec 56.7% 8164 13-Dec 47.3% 6810 14-Dec 53.3% 7670 15-Dec 0.0% 1 16-Dec 0.0% 0 17-Dec 64.0% 9212 18-Dec 51.4% 7406


May 2013 Page 8

Date Capture rate Number of 6 second average records

19-Dec 39.9% 5743 20-Dec 28.8% 4148 21-Dec 82.1% 11820 22-Dec 42.9% 6173 23-Dec 93.6% 13483 24-Dec 85.3% 12277 25-Dec 88.8% 12787 26-Dec 73.9% 10647 27-Dec 93.2% 13423 28-Dec 82.7% 11907 29-Dec 88.0% 12678 30-Dec 99.9% 14379 31-Dec 7.3% 1048 1-Jan 0.7% 98 2-Jan 58.6% 8441 3-Jan 83.4% 12011 4-Jan 94.5% 13611 5-Jan 99.9% 14379 6-Jan 94.8% 13650 7-Jan 94.1% 13556 8-Jan 94.8% 13647 9-Jan 52.5% 7567 10-Jan 94.7% 13643 11-Jan 14.4% 2074 12-Jan 96.5% 13894 13-Jan 95.4% 13731 14-Jan 95.0% 13685 15-Jan 94.5% 13605 16-Jan 94.8% 13649 17-Jan 63.90% 9202 18-Jan 95.10% 13695 19-Jan 95.50% 13756 20-Jan 94.30% 13578 21-Jan 95.10% 13694 22-Jan 94.80% 13650 23-Jan 92.90% 13384 24-Jan 95.10% 13693 25-Jan 88.50% 12745 26-Jan 99.90% 14380 27-Jan 94.80% 13648 28-Jan 94.30% 13586 29-Jan 92.80% 13370

Average / Total 67.7% 604769


May 2013 Page 9

Table 2 Data capture rate in days for TSP, PM10 and PM2.5, 30 November 2012 up to and including 29 January 2013

Parameter Value Equivalent days 42

Days above 50% capture 47 Days above 70% capture 35

Table 3 Data capture rate by hour of the day (%) Hour Capture Rate (%)

1 60.6 2 59.4 3 63.6 4 66.6 5 63.6 6 62.4 7 64.8 8 64.8 9 63

10 62.4 11 66 12 67.2 13 66.6 14 67.2 15 68.4 16 69 17 66 18 68.4 19 74.4 20 72.6 21 71.4 22 70.2 23 72.6 24 70.8

Table 4 Data capture rate by day of the week (%) Day Capture Rate (%)

Monday 73.8 Tuesday 70.2

Wednesday 57.0 Thursday 64.8

Friday 64.2 Saturday 64.8 Sunday 73.8

4.2 Train data capture

The train data captured at the Metford wayside monitor was provided in spreadsheet format. Data was analysed firstly by commodity type (either coal, freight, passenger or unknown) and then by track location (either UM, DM, UC or DC) to indicate the direction of train travel and to determine if coal trains were loaded or unloaded. For example, a coal train on the UC (UP COAL) track was classed as a loaded coal train as it was heading to the port. Conversely, a coal train on the DC track was classed as an unloaded coal train heading to the mines.


May 2013 Page 10

The total number of trains recorded by the wayside monitor for each train type is shown in Table 5.

Table 5 Total number of trains during the PRP 4.2 monitoring period Train type Number of trains recorded during PRP 4.2 study period

Loaded coal 2,550 Unloaded coal 2,969 Freight 515 Passenger 5,043 Unknowns 168 Total 11,245 4.3 Data validation

Data deemed to be invalid was removed and was not included in the analysis presented in this report. A summary of the data removed and the justification is provided in Appendix A. 4.4 Algorithms for combining train, meteorological and pollution data

Details of algorithms for combining the different datasets is described below. These algorithms resulted in a consolidated dataset for the entire monitoring period. 4.4.1 Synchronisation of data

Train data from the Metford wayside monitor was received from ARTC in Australian Daylight Savings Time (UTC + 11 hours). The internal clock of the data logger connected to the meteorological station was synchronised with the wayside monitoring equipment. The Osiris recorded data in Australian Eastern Standard Time (UTC + 10 hours). All train and meteorological data was adjusted to Australian Eastern Standard Time and then combined with the particulate concentration data. This adjustment was such that trains recorded as passing by at 11 am Australian Daylight Savings Time were associated with Osiris data recorded at 10 am Australian Eastern Standard Time. The monitoring results were subject to spot-checks to determine provide further confidence that train pass-bys were synchronized with measurements of particulate concentrations. 4.4.2 Determination of train type

Train types were determined from data collected by ARTC’s wayside monitor at Metford. A sample of the train data received is shown in Table 6. Katestone developed data processing algorithms to extract information from the wayside monitor data. An example of this is given below. The final column of Table 6 states the train direction and type extracted from the two examples. Inconsistencies were noted between the train directions indicated by "Trn_Site_Initials" and "Trn_Direction", and between train types indicated by "Trn_Site_Initials" and "Commodity". The algorithm used to identify train direction and type extracts either "U" or "D" from "Trn_Site_Initials" to determine the train direction and then refers to "Commodity" to determine whether it is a coal, freight, or other type of train.


May 2013 Page 11

Table 6 Example of train data received from ARTC and train identification Parameter

Train identification Trn_Site_Initials Trn_Time Trn_Direction Commodity

DC 22/01/2013 7:38:49 Up General

Freight Freight train travelling

in 'Down' direction

DC 22/01/2013 8:05:01 Up Coal Coal train travelling in

'Down' direction 4.4.3 Train pass-by time

Pass-by duration was determined by the speed and train length recorded by the wayside monitor. This assumes constant train speed and does not account for acceleration or deceleration of the train. In cases of accelerating and decelerating trains, this may lead to a minor error in train pass-by time.


May 2013 Page 12

Table 7 Example of train data received from ARTC and train characteristics

Trn_Site_Initials Trn_Time

Trn_ Directio

n Trn_

Commodity Trn_

Locomotives Trn_

Carriages Length Carr Count

Carr_ Speed

Carr_Mass _Sum

Carr_Mass _Avg

UC 24/01/2013 0:06:03 Up Coal 2 28 444.686 30 50.62 3,072.27 102.41

UC 24/01/2013 0:28:52 Up Coal 3 96 1539.817 99 58.63 14,246.03 143.90

UC 24/01/2013 0:58:32 Up Coal 2 82 1408.086 84 57.27 12,460.99 148.35

UC 24/01/2013 1:35:58 Up Coal 3 90 1515.45 92 61.21 13,319.66 144.78

UC 24/01/2013 1:55:24 Up Coal 3 96 1540.907 99 52.68 14,075.66 142.18

UC 24/01/2013 2:40:06 Up Coal 2 82 1406.585 84 47.85 11,960.34 142.38

UC 24/01/2013 2:51:50 Up Coal 1 63 1024.865 48 39.94 5,916.64 123.26


May 2013 Page 13

4.5 Particulate averaging period

To calculate a representative concentration of TSP, PM10 and PM2.5 that may have been associated with a train pass-by, it was important to consider the potential entrainment in the air of particulates after a train passed the monitor. There is no standard methodology to account for the entrainment of particles in this situation. The following methodology was adopted for the PRP 4.2 study and represents a conservative approach based on the available data:

• From the time that the train passed the Metford monitor, each 6-second average particulate concentration was summed and averaged over the duration of the train pass-by time

• For all non passenger trains (i.e. loaded coal trains, unloaded coal trains and freight trains) the number of 6-second average readings was increased threefold to account for the entrainment of particulates in the atmosphere after the train had passed

• The number of 6-second average readings for passenger trains was not increased threefold because of the short duration of a pass-by (2 seconds) relative to the Osiris averaging period of 6 seconds

• The concentration of particulates for no trains was determined from the average of all the 6-second average readings when there was no train passing the Osiris. The data associated with entrainment was excluded from the no trains average concentrations to avoid double counting.

The threefold increase to the number of 6 second averages associated with all non-passenger trains was based on analysis of the average TSP concentration during the 8 minutes prior to and 8 minutes after a train passed the trackside monitoring station, as shown in Figure 3. The analysis shows that the time taken for the average TSP concentration to return to the same level recorded when the train pass-by started was approximately three times the average train pass-by time, indicating the entrainment of particulates. Whilst the period of time when entrainment occurred would in reality vary for each train, a threefold increase in the number of 6 second averages captured the train’s influence on particulate concentrations. The methodology described above varies slightly from the methodology in the pilot study which increased the pass-by time based on the ambient wind speed. The pilot study increased a train pass-by threefold if the wind speed was less than 2 metres per second. Based on the findings of this study with respect to wind speed (Section 6.5), which showed no increasing or decreasing trend in average concentrations with respect to ambient wind speed, and the fact that wind speeds of less than 2 metres per second were recorded for 78% of the study period, it was deemed conservative to apply the threefold increase to all non-passenger train pass-bys.


May 2013 Page 14

5. Assessment Methodology

5.1 Calculation of concentration

Concentrations of TSP, PM10 and PM2.5 for each train type were calculated by averaging the number of records of TSP, PM10 and PM2.5 recorded whilst the train passed the monitoring station (including the increased pass-by time for all non-passenger trains described in Section 4.5). For example:

• A loaded coal train may have ten 6 second averages. The total averaging time of particulate concentrations was made to be three times the pass-by time (thirty 6 second averages) to determine the particulate concentration associated with that specific train

• A passenger train passing within two seconds. Due to the short pass by duration of the passenger train the closest 6 second average to the clock pass-by time of the train was assigned to the train. It may also be possible for a passenger train to pass-by over two 6 second averages. In this instance, both 6 second averages were assigned to the train.

• No trains. All the6 second averages when no train pass-bys were recorded were assigned as no trains. The threefold increase in 6 second averages associated with all non-passenger trains were excluded from the no trains data to avoid double counting.

• A multiple pass-by. If two or more trains passed the monitoring station at the same time (including the time of the threefold increase in pass-by for all non-passenger trains) the concentrations were averaged for the duration of the pass-by and assigned as a multiple pass-by event.

If concentration monitoring data were not present in the dataset for the entire duration of a train pass-by, the available concentration data that were present were averaged and assigned to that train. 5.2 Assessment of contribution by train type

Particulate concentration data (as TSP, PM10 and PM2.5) were grouped according to the following categories related to train types:

• Loaded coal • Unload coal • Freight • Passenger • Multiple pass-by • No trains • Unknowns


May 2013 Page 15

Grouped particulate concentration data was subjected to statistical analysis with the exception of unknown trains, which were removed. The following statistics were reported for each grouped dataset:

• Average • Median • Standard deviation • Upper confidence level on average (95%) • Lower confidence level on average (95%) • 5th percentile • 95th percentile • Maximum concentration

The grouped datasets have been used to characterize the dust concentrations that are likely to be associated with each train population. Overlap of upper and lower confidence levels has been used to determine whether there is statistical significance in the differences between the average concentrations of the grouped samples that have been measured to be representative of each population. The grouped data was also analysed using graphical techniques such as:

• Bar charts • Box and whisker plots

5.3 Assessment of contribution by other variables

The data were analysed to identify whether relationships exist between particulate concentration and the following parameters:

• Train speed • Wind speed • Wind direction

The results of these analyses are presented in Appendix B.


May 2013 Page 16

6. Results

6.1 Train movements

For the period 30 November 2012 up to including 29 January 2013 TSP, PM10 and PM2.5 data were collected for each train type, including when no trains passed. Table 8 provides information on the number of trains and data points that have been used in this analysis.

Table 8 Number of 6 second average particulate records associated with each train type for a single pass-by and multiple pass-bys used in the analysis of particulate concentrations

Train type Number of trains with valid data

Number of 6 second average records TSP PM10 PM2.5

Loaded coal 916 41,102 41,102 41,102 Unloaded coal 1,109 48,524 48,524 48,524 Freight trains 135 6,052 6,052 6,052 Passenger 2,624 2,641 2,641 2,641 Multiple pass-bys 1,377a 75,715 75,715 75,715 Unknowns 93 496 496 496 No trains - 555,464 555,460 555,464 Table note: a For multiple pass-bys 1,377 is the number of instances when there are two or more trains passing the detector simultaneously.

6.2 Particulate matter concentration by train type

Table 9 presents statistics for TSP, PM10 and PM2.5 based on the following train types: unloaded coal, loaded coal, freight, passenger and multiple pass-bys. Also included are the statistics for times when no trains were passing. Differences in the average, median, 95th percentile and maximum concentrations by train type are shown for TSP, PM10 and PM2.5 in Figure 4 to Figure 6, respectively. Also included in the figures are particulate concentrations for times when no trains were passing the monitoring station. Figure 7 to Figure 9 present the upper and lower 95% confidence limits by train type for TSP, PM10 and PM2.5, respectively. Also included in the figures are particulate concentrations for times when no trains were passing the monitoring station. Figure 10 to Figure 12 present box and whisker plots of the 95th, 75th, 25th and 5th percentile concentrations by train type for TSP, PM10 and PM2.5, respectively. The box represents the 75th and 25th percentile concentrations and the upper and lower whiskers represent the 95th and 5th percentile concentrations. The median is represented by a dash within the box. The results show:

• The range in the median concentrations between all train types and concentrations measured when no trains were passing the monitoring station was 8.3 µg/m3 for TSP, 4.8 µg/m3 for PM10 and 1.6 µg/m3 for PM2.5.. This corresponds to a difference of 26%, 20% and 20% in median concentrations of TSP, PM10 and PM2.5 compared to concentrations when no trains were passing the monitoring station.


May 2013 Page 17

• The average, median and 95th percentile concentrations of TSP and PM10 were highest for multiple pass-bys

• The average and median concentration of PM2.5 was highest for multiple pass-bys • The 95th percentile concentration of PM2.5 was highest for freight trains • The average and 95th percentile concentrations of TSP for freight trains, unloaded

coal trains and multiple pass-bys were higher than loaded coal trains • The average and 95th percentile concentrations of PM10 for freight trains, unloaded

coal trains and multiple pass-bys were higher than loaded coal trains • The average and 95th percentile concentrations of PM2.5 for freight trains, unloaded

coal trains and multiple pass-bys were higher than loaded coal trains • There was no statistically significant difference in the average concentrations of TSP,

PM10 and PM2.5 associated with freight and passenger trains compared with concentrations recorded when no trains were passing the monitoring station

• There was no statistically significant difference in the average concentrations of PM10 and PM2.5 associated with loaded coal trains compared with concentrations recorded when no trains were passing the monitoring station

• There was a statistically significant difference in the average concentration of TSP associated with loaded coal trains and unloaded coal trains compared with concentrations recorded when no trains were passing the monitoring station

• There was a statistically significant difference in the average concentrations of PM10 and PM2.5 associated with unloaded coal trains compared with concentrations recorded when no trains were passing the monitoring station

• Average concentrations of TSP associated with loaded coal trains, unloaded coal trains and freight trains were higher by 3.2 µg/m3, 6.1 µg/m3 and 4.5 µg/m3, respectively. This corresponds to an increase in average concentrations when no train passes the monitoring station of 9%, 16% and 12% for loaded coal trains, unloaded coal trains and freight trains, respectively.

• Average concentrations of PM10 associated with loaded coal trains, unloaded coal trains and freight trains were higher by 2.3 µg/m3, 4.5 µg/m3 and 3 µg/m3, respectively. This corresponds to an increase in average concentrations when no train passes the monitoring station of 8%, 16% and 11% for loaded coal trains, unloaded coal trains and freight trains, respectively.

• Average concentrations of PM2.5 associated with loaded coal trains, unloaded coal trains and freight trains were higher by 0.6 µg/m3, 1.2 µg/m3 and 0.7 µg/m3, respectively. This corresponds to an increase in average concentrations when no train passes the monitoring station of 6%, 13% and 8% for loaded coal trains, unloaded coal trains and freight trains, respectively.

• There was a statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and passenger trains.

• There was no statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and loaded coal trains and freight trains.

• There was no statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between loaded coal trains, freight trains and passenger trains.


May 2013 Page 18

Table 9 Particulate concentrations by train type

Parameter Units TSP PM10 PM2.5

Loaded coal

Unloaded coal Freight Passenger Multiple

pass-bys No trains Loaded coal


pass-bys No trains Loaded coal


pass-bys No trains

Number of Trains # 916 1109 135 2624 1,377 - 916 1108 135 2624 2,843 - 916 1108 135 2624 2,843 -

Mean

µg/m³

40.6 43.5 41.9 36.8 44.0 37.4 30.2 32.4 30.9 27.4 32.4 27.9 10.0 10.6 10.10 9.1 10.7 9.4 Median 36.2 39.3 36.8 30.9 39.8 31.5 27.1 28.4 26.7 24.0 29.2 24.4 8.9 9.0 8.6 8.1 9.5 7.9

Maximum 369.9 1173.4 103.7 1124.9 517.9 5538.4 188.2 770.9 78.6 1121.2 360.9 5538.3 59.1 249.3 48.8 175.3 93.8 640.5 Standard Deviation 47.3 60.9 47.3 50.8 52.2 39.5 34.8 44.5 35.0 39.3 38.1 30.4 11.9 14.4 12.2 11.5 12.7 10.9

Upper 95% Confidence

Level 43.7 47.0 49.9 38.8 46.8 37.5 32.4 35.0 36.8 28.9 34.4 28.0 10.7 11.4 12.2 9.6 11.3 9.4

Lower 95% Confidence

Level 37.6 39.9 33.9 34.9 41.3 37.3 27.9 29.8 25.0 25.9 30.4 27.9 9.2 9.7 8.0 8.7 10.0 9.4

95th percentile 79.1 81.8 81.8 81.60 89.6 84.0 60.5 62.5 63.8 57.4 65.7 58.9 22.1 22.8 25.3 20.3 23.2 21.4

5th percentile 12.1 12.0 13.8 5.8 12.9 6.3 9.0 8.6 9.4 5.2 9.1 5.6 2.6 2.5 2.3 2.0 2.8 1.9

Time of Maximum

date and time

10/01/13 9:19

10/01/13 9:37

8/01/13 21:40

17/12/12 8:37

10/01/13 8:56

3/01/13 12:56

10/01/13 9:19

10/01/13 9:37

11/12/12 4:44

17/12/12 8:37

10/01/13 8:56

3/01/13 12:56

3/01/13 13:01

15/01/13 13:40

27/12/12 12:58

16/01/13 0:27

5/01/13 4:38

3/01/13 12:56

Wind Speed m/s 2.2 1.7 2.9 1.0 2.8 3.8 2.17 1.71 ND 0.96 2.81 3.8 2.52 1.57 ND 0.59 0.34 3.8 Wind

Direction ° 77 71 295 240 65 112 77 71 ND 240 65 112 93 95 ND 224 231 112

Table note: ND indicates that wind speed and wind direction data was not collected at the time of the maximum


May 2013 Page 19

6.3 Variations in concentration with train speed

Concentrations of TSP, PM10 and PM2.5 have been grouped into the following train speed categories:

• Less than 30 km/hr • 30 to 60 km/hr • 60 to 90 km/hr • Above 90 km/hr

Statistical analysis of this data is presented in Appendix B for all trains. These data were then further categorised by train type. Box and whisker plots of concentrations of TSP, PM10 and PM2.5 by train type and train speed are presented in Appendix B. The results show:

• That there was no increasing or decreasing trend in average concentrations with respect to train speed.

6.4 Variations in concentration with wind direction

Initial analysis of the data considered concentrations of TSP, PM10 and PM2.5 from the eight compass points of N, NE, E, SE, S, SW, W and NW in order to determine whether there were other sources in the local area which may be contributing to the measured concentrations other than trains. Figure 13 presents a box and whisker plot for TSP, PM10 and PM2.5 for all data (trains and no trains) for the eight wind directions. The figure illustrates that average concentrations of TSP, PM10 and PM2.5 were higher when winds were from the S, SW, W and NW directions. The railway tracks are located between 4 metres and 17.5 metres to the south of the monitoring station and is aligned along the axis 150 degrees and 300 degrees. Further analysis of concentrations based on when wind was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station) was undertaken based on train type. Statistical analysis of this data is presented in Appendix B for each train type. The results show:

• There was no statistically significant difference in the average concentrations of TSP, PM10 and PM2.5 associated with passenger trains and freight trains compared with concentrations recorded when no trains were passing the monitoring station when wind direction was from the rail tracks towards the monitoring station

• There was a statistically significant difference in the average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains and unloaded coal trains compared with concentrations recorded when no trains were passing the monitoring station when wind direction was from the rail tracks towards the monitoring station.

• There was no statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between train types when wind direction was from the rail tracks towards the monitoring station, with the exception of unloaded coal trains for TSP,


May 2013 Page 20

PM10 and PM2.5 which showed a statistical difference to passenger trains and loaded coal trains for TSP which showed a statistical difference to passenger trains.

• Average concentrations of TSP, PM10 and PM2.5 associated with unloaded coal trains

were higher than concentrations when no train was passing the monitor by 10.1 µg/m3, 7.6 µg/m3 and 2.1 µg/m3, respectively, when wind direction was from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 23%, 24% and 21% for TSP, PM10 and PM2.5, respectively.

• Average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains

were higher than concentrations when no train was passing the monitor by 6.0 µg/m3, 4.3 µg/m3 and 1.1 µg/m3, respectively, when wind direction was blowing from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 14%, 14% and 11% for TSP, PM10 and PM2.5, respectively.

• It should be noted that other sources of particulates are located outside the rail

corridor surrounding the Osiris monitor, such as an industrial estate and a cemetery. Emissions of particulates from other sources in the area could have contributed to the particulate concentrations recorded by the Osiris monitor.

6.5 Variations in concentrations with wind speed

Concentrations of TSP, PM10 and PM2.5 have been grouped into the following wind speed categories:

• 0 to 0.5 m/s • 0.5 to 1 m/s • 1 to 2 m/s • 2 to 5 m/s • 5 to 10 m/s

Statistical analysis of this data is presented in Appendix B for all trains. The results show:

• The results show that there was no increasing or decreasing trend in average concentrations with respect to ambient speed.

6.6 Variations in concentrations minus background

The analysis so far has not excluded the existing ambient levels of particulates in the area at the time of train pass-by. In order to differentiate between concentrations linked directly to a train, additional analysis was undertaken where the concentrations of TSP, PM10 and PM2.5 when no train was passing the monitor (representing the “background”) was subtracted from the concentrations of TSP, PM10 and PM2.5 recorded during the pass-by of a train. In order to account for the natural variability in existing ambient levels background concentrations were determined by averaging the 6-second data for the two minutes prior to the arrival of a train. This value was then subtracted from the average concentration calculated over the length of the train pass-by (including the threefold increase for all non-passenger trains) to determine a “train only” concentration. This was done for all instances when there was a train (single pass-by only). If during the two minutes prior to a train arrival,


May 2013 Page 21

another train was present, this data was omitted from the analysis. Therefore, in this section train only concentrations was compared to the background (the data two minutes prior to the arrival of a train). Table 10 presents statistics for concentrations of TSP, PM10 and PM2.5 due to “train only” for the following train types:

• Unloaded coal • Loaded coal • Freight • Passenger

Figure 14 to Figure 16 present the upper and lower 95% confidence limits by train type for TSP, PM10 and PM2.5, respectively. Figure 17 to Figure 19 present box and whisker plots of the 95th, 75th, 25th and 5th percentile concentrations by train type for TSP, PM10 and PM2.5, respectively. The box represents the 75th and 25th percentile concentrations and the upper and lower whiskers represent the 95th and 5th percentile concentrations. The median is represented by a dash within the box. Also included is the concentrations for the background. The results show:

• For TSP there is a statistically significant difference in concentration between unloaded coal trains and passenger trains. For all other train types there is no statistically significant difference in TSP concentrations

• For PM10 and PM2.5 there is no statistically significant difference in concentrations between train types

• The largest increase in TSP, PM10 and PM2.5 concentrations was due to unloaded coal trains, with an average concentration of 3.3 µg/m3, 2.5 µg/m³ and 0.6 µg/m³, respectively

• The average concentration changed by 1.3 µg/m3, 0.6 µg/m3 and -0.2 µg/m3 for TSP, PM10 and PM2.5 in the presence of a loaded coal train

• Passenger trains had negligible impact on average concentrations


May 2013 Page 22

Table 10 Particulate concentrations by train type minus background (µg/m³)

Parameter Units TSP PM10 PM2.5

Loaded coal Unloaded coal Freight Passenger Loaded coal Unloaded coal Freight Passenger Loaded coal Unloaded coal Freight Passenger

Number of Trains* # 906 1071 133 2610 906 1071 133 2610 906 1071 133 2610 Mean

µg/m³

1.3 3.3 0.9 0.2 0.6 2.5 0.6 0.1 -0.2 0.6 0.3 -0.1 Median 1.9 2.0 0.7 -2.4 1.1 1.2 0.9 -0.7 0.2 0.2 0.1 -0.1

Maximum 182.4 763.0 75.1 1030.1 116.7 554.3 49.5 1057.8 34.0 242.9 40.0 159.2 Standard Deviation 28.5 30.1 18.5 30.5 26.0 23.4 13.5 24.8 9.7 9.6 6.7 5.8

Upper 95% Confidence Level 3.2 5.1 4.0 1.4 2.3 3.9 2.9 1.1 0.5 1.2 1.4 0.1

Lower 95% Confidence Level -0.6 1.5 -2.3 -1.0 -1.1 1.1 -1.7 -0.8 -0.8 0.0 -0.9 -0.3

95th percentile 20.9 23.3 29.9 30.5 14.6 16.1 20.0 13.1 4.4 5.2 5.9 1.8 5th percentile -18.3 -20.0 -25.1 -22.9 -12.0 -12.4 -18.8 -12.6 -2.8 -4.2 -4.3 -2.2

Table note: * The number of trains in each category will vary from the number of trains presented in Table 9 due to the presence of a train in the 2-minutes prior to a train arrival


May 2013 Page 23

6.7 Influence of rainfall on particulate concentrations

Rainfall information recorded at the Bureau of Meteorology (BoM) Maitland Visitors Centre weather station has been investigated to determine the potential impact of rainfall on particulate concentrations. Daily rainfall measured at the Maitland Visitors Centre during the PRP 4.2 period (30 November 2012 to 29 January 2013) and during the pilot monitoring program (14 February to 20 March 2012) is shown in Table 11, Figure 20 and Figure 21, respectively.

Table 11 Rainfall recorded at Maitland Visitors Centre

Parameter PRP 4.2

(30 November 2012 to 29 January 2013)

Pilot Monitoring Program (12 February to 20 March 2012)

Monitoring period length (days) 62 35 Total rainfall (mm) 186 142 Average daily rainfall (mm) 3 4 Number of dry days 38 21 Percentage of dry days 62% 60% The total amount of rainfall during the PRP 4.2 study period was 186 mm with an average daily rainfall of 3 mm. However, 133 mm of rain, or 71% of the total rainfall amount, was recorded over three days of the study period. A total of 108 mm was measured on the last two days of the study period and therefore could not have contributed to lower dust levels in the study period outside of these two days. During the pilot monitoring program there was a total of 142 mm of rainfall with an average daily rainfall of 4mm. There were 9 days, or 26 % of the time, when daily rainfall was between 5 mm and 35mm. Hence, the majority of the PRP 4.2 study period was substantially drier than the pilot monitoring program and, with the exception of the last two days of the monitoring program, rainfall during the program was well below average (78 mm vs average for December and January of 112.7 mm). For five months prior to the PRP 4.2 study period (July to November 2012), below average rainfall was recorded at the Maitland Visitors Centre. The total for the five months prior to the PRP 4.2 study was 133 mm compared with the long-term average for these months combined of 273.4 mm. For 1.5 months prior to the pilot monitoring program (1 January to 11 February 2012), above average rainfall was recorded at the Maitland Visitors Centre, with 211 mm of rainfall recorded compared with the average for January and February of 167.4 mm. The rainfall over the two monitoring periods indicates that the PRP 4.2 study period was drier than the pilot monitoring program and that the months prior to the PRP 4.2 study period were drier than the months prior to the pilot monitoring program. Overall the PRP 4.2 study period would be more representative of dry periods, conducive to higher potential for the generation of particulate matter. Further evidence of the impact of rainfall is investigated through comparison between particulate concentrations measured during the two monitoring periods. The 24-hour average PM10 and PM2.5 data recorded at the Metford site during the pilot program and the PRP 4.2 study has been compared to one another and to the equivalent data recorded concurrently at OEH sites in the Lower Hunter Valley. The three OEH sites are Beresfield, Wallsend and Newcastle.


May 2013 Page 24

Table 12 provides the average, maximum and 70th percentile 24-hour average concentrations recorded at each site (only the average concentration was presented in the pilot monitoring program). It should be noted that the NSW OEH sites use monitors that are compliant with NEPM monitoring requirements whereas the Osiris monitor used at Metford is non-compliant. The values presented below at the ARTC Metford site should not be compared to air quality objectives to assess compliance.

Table 12 PM10 and PM2.5 24-hour average concentrations recorded during the pilot monitoring program and during the PRP 4.2 monitoring program (µg/m³)

Pollutant Statistics ARTC Metford site

NSW OEH Data Beresfield Wallsend Newcastle

Pilot monitoring program - 14 February to 20 March 2012 PM10 Average 18.7 16.4 13.6 15.5

Maximum - 31.8 21.5 26.2 70th percentile - 18.1 15.5 18.6

PM2.5 Average 5.5 4.4 3.6 - Maximum - 6 6.1 -

70th percentile - 5.0 4.1 - PRP 4.2 monitoring program - 30 November 2012 to 29 January 2013 PM10 Average 29.1 22.2 18.9 28.4

Maximum 51.6 53.4 41.1 50.7 70th percentile 35.1 25.9 21.4 31.9

PM2.5 Average 9.8 7.9 7.3 - Maximum 20.8 13.9 17.6 -

70th percentile 11.9 9.2 8.5 - The data shows the following:

• Higher average concentrations of PM10 and PM2.5 were recorded at the ARTC Metford site during the PRP 4.2 study period than during the pilot monitoring program

• Higher concentrations of PM10 and PM2.5 were recorded at the NSW OEH sites during the PRP 4.2 study period than during the pilot monitoring program

• Comparison of the measured particulate concentrations with rainfall during the two monitoring periods suggests that rainfall events during and before the pilot program resulted in lower particulate concentrations at Metford and across the region.


May 2013 Page 25

7. Comparison with Pilot Study

The measured PM10 and PM2.5 concentrations detailed in Table 12 show that the PRP 4.2 monitoring program recorded higher concentrations than the pilot monitoring program. This is shown in more detail in Table 13, which compares the average PM10 concentrations for each train type during each monitoring period. The table clearly shows that the PRP 4.2 study period recorded higher PM10 concentrations for each train type.

Table 13 Average PM10 concentrations recorded at the ARTC Metford site during the two study periods (µg/m³)

Average PM10 concentration (µg/m³) Loaded coal

Unloaded coal Freight Passenger

Multiple pass-bys

PRP 4.2 (30 November 2012 to 29 January 2013)

30.2 32.4 30.9 27.4 32.4

Pilot Monitoring Program (12 February to 20 March 2012)

20.7 20 20.6 18.6 22.1

There are a number of differences between the pilot monitoring program and the PRP 4.2 monitoring study that may affect the results, including:

• Longer sampling period of 62 days compared with 35 days • Change in Osiris monitor averaging period from 30 seconds to 6 seconds • Rainfall prior to and during the study periods • Increase in percentage of multiply pass-bys from 8% (pilot study) to 30%

The Osiris monitor averaging period was increased to 6 seconds for the PRP 4.2 monitoring program. This fivefold increase in data capture significantly increases the resolution of concentrations related to passing trains and is more likely to record short-term spikes in concentrations. This is evident in larger maximum concentrations that were recorded, especially for the no trains category. The rainfall that occurred prior to and during the study periods is likely to be the main contributor to the difference in recorded concentrations. As discussed in Section 6.7, the amount of rainfall that occurred prior to and during the majority of the pilot program was greater than the PRP 4.2 study period. This is supported by the particulate concentrations at the ARTC Metford site and the NSW OEH monitoring stations. As discussed in Section 5.1 the methodology used to account for entrainment of dust behind a train varied from the methodology used in the pilot study which increased the pass-by time based on the ambient wind speed. The pilot study increased a train pass-by threefold if the wind speed was less than 2 metres per second. In this study a threefold increase to the number of 6 second averages associated with all non-passenger trains was included to capture the train’s influence on particulate concentrations. This resulted in an increase in train length and an increase in occurrences of multiple pass-bys.


May 2013 Page 26

8. Limitations

Limitations of the program include:

• In general, the light scattering photometers that are available to measure continuous particulate concentrations cannot measure a true TSP size fraction

• Recording of 6 second average data was a considerable challenge for the Osiris, which has limited on-board data storage that affected the Osiris’s functionality and data recording capacity during download

• The Metford particulate monitor was located at unequal distances from the coal and main rail lines

• Trains with the commodity label “Unknown” were omitted from the analysis • Pass-by duration was determined by the speed and train length recorded by the

wayside monitor and extended by threefold to account for particle entrainment. This assumes constant train speed and does not account for acceleration or deceleration of the train. In cases of accelerating and decelerating trains, this may lead to a minor error in train pass-by time


May 2013 Page 27

9. Conclusions

Monitoring of particulates has been conducted at Metford on the Hunter Valley rail network in response to PRP 4.2– Particulate Emissions from Coal Trains. The objective of PRP 4.2 was to determine whether:



The findings of monitoring program are as follows:

• Passenger trains were not associated with a statistically significant difference in TSP, PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station

• Freight trains were not associated with a statistically significant difference in TSP,

PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station.

• Loaded coal trains were not associated with a statistically significant difference in

PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station. However, loaded coal trains were associated with a statistically significant difference in TSP concentrations when compared with concentrations when no train was passing the monitoring station.

• Unloaded coal trains were associated with a statistically significant difference in TSP, PM10 and PM2.5 concentrations when compared with the concentrations recorded when no train was passing the monitoring station.

• Average concentrations of TSP associated with loaded coal trains, unloaded coal


• Average concentrations of PM10 associated with loaded coal trains, unloaded coal


• Average concentrations of PM2.5 associated with loaded coal trains, unloaded coal


• Loaded coal trains operating on the Hunter Valley rail network, when measured at

Metford, did not have a statistically stronger association with elevated TSP, PM10 and PM2.5 concentrations than other trains.

• There was a statistically significant difference in concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and passenger trains. However, there was no


May 2013 Page 28

statistically significant difference in concentrations of TSP, PM10 and PM2.5 between the other train types.

• There was no increasing or decreasing trend in average concentrations with respect to train speed.

• When the wind direction was between 150 degrees and 300 degrees (blowing from

the rail tracks towards the monitoring station) there was a statistically significant difference in the average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains and unloaded coal trains compared with concentrations recorded when no trains were passing the monitoring station.

• There was a statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between unloaded coal trains and passenger trains when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• There was no statistically significant difference in average concentrations of TSP, PM10 and PM2.5 between freight trains and passenger when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• There was a statistically significant difference in average concentrations of TSP between loaded coal trains and passenger trains; however, there was no statistically significant difference in average concentrations of PM10 and PM2.5 between loaded coal trains and passenger trains when the wind direction was between 150 degrees and 300 degrees (blowing from the rail tracks towards the monitoring station).

• Average concentrations of TSP, PM10 and PM2.5 associated with unloaded coal trains were higher than concentrations when no train was passing the monitor by 10.1 µg/m3, 7.6 µg/m3 and 2.1 µg/m3, respectively, when wind direction was from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 23%, 24% and 21% for TSP, PM10 and PM2.5, respectively.

• Average concentrations of TSP, PM10 and PM2.5 associated with loaded coal trains

were higher than concentrations when no train was passing the monitor by 6.0 µg/m3, 4.3 µg/m3 and 1.1 µg/m3, respectively, when wind direction was blowing from the rail tracks towards the monitoring station. This corresponds to an increase in average concentrations when no train passes the monitoring station of 14%, 14% and 11% for TSP, PM10 and PM2.5, respectively.

• There was no increasing or decreasing trend in average concentrations with respect to ambient wind speed.


May 2013 Page 29

10. References

Environ, 2012, Pollution Reduction Program (PRP) 4 – Particulate Emissions from Coal Trains, Prepared for Australian Rail Track Corporation, Prepared By Environ Australia Pty Ltd, Project Number: AS130301A, September 2012

Katestone, 2012, Work Program PRP 4.2 Particulate Emissions from Coal Trains, Prepared for Australian Rail Track Corporation, October 2012


May 2013 Page 30

Figure 1 Site plan showing Metford monitoring location

Location: Metford, NSW

Data source: GIS

Type: Site plan

Prepared by: Kyle Wright

Date: October 2012


May 2013 Page 31

Figure 2 Metford particulate and meteorological monitoring station

Location: Metford, NSW

Data source: Environ, 2012

Type: Photograph

Prepared by: Environ

Date: September 2012


May 2013 Page 32

Figure 3 Comparison of average TSP concentrations of loaded and unloaded coal trains in the 8 minutes prior to and 8 minutes after passing the trackside monitor

Location: Metford Averaging period: 6 second Data source: Osiris Units: µg/m³

Type: Line graph Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 33

Figure 4 Comparison of TSP concentrations by train type


Type: Bar chart Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 34

Figure 5 Comparison of PM10 concentrations by train type




May 2013 Page 35

Figure 6 Comparison of PM2.5 concentrations by train type




May 2013 Page 36

Figure 7 Comparison of TSP concentration upper and lower 95% confidence limits by train type


Type: Chart Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 37

Figure 8 Comparison of PM10 concentration upper and lower 95% confidence limits by train type




May 2013 Page 38

Figure 9 Comparison of PM2.5 concentrations upper and lower 95% confidence limits by train type




May 2013 Page 39

Figure 10 Comparison of TSP concentrations, 95th, 75th, 25th, and 5th percentiles by train type


Type: Box and whisker plot Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 40

Figure 11 Comparison of PM10 concentrations, 95th, 75th, 25th, and 5th percentiles by train type




May 2013 Page 41

Figure 12 Comparison of PM2.5 concentrations, 95th, 75th, 25th, and 5th percentiles by train type




May 2013 Page 42

Figure 13 Concentrations of TSP, PM10 and PM2.5 by wind direction

Location: Metford

Averaging period: 6 second

Data source: Osiris

Units: Degrees

Type: Box and whisker plot

Data period: 30/11/12 to 29/01/13

Prepared by: A. Vernon

Date: May 2013


May 2013 Page 43

Figure 14 Comparison of TSP concentrations upper and lower 95% confidence limits by train type minus background




May 2013 Page 44

Figure 15 Comparison of PM10 concentrations upper and lower 95% confidence limits by train type minus background




May 2013 Page 45

Figure 16 Comparison of PM2.5 concentrations upper and lower 95% confidence limits by train type minus background




May 2013 Page 46

Figure 17 Comparison of TSP “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type




May 2013 Page 47

Figure 18 Comparison of PM10 “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type


Type: Box and whisker Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 48

Figure 19 Comparison of PM2.5 “train only” concentrations, 95th, 75th, 25th, and 5th percentiles by train type


Type: Box and whisker Data period: 30/11/12 to 29/01/13 Prepared by: A. Vernon Date: May 2013


May 2013 Page 49

Figure 20 Daily rainfall during the study period (mm)

Location: Maitland Visitors Centre

Averaging period: Daily Data source: Bureau of Meteorology

Units: mm



May 2013 Page 50

Figure 21 Daily rainfall during the pilot monitoring program (mm)

Location: Maitland Visitors Centre

Averaging period: Daily Data source: Bureau of Meteorology

Units: mm



May 2013 Page 51

Appendix A Data validation

A1 Full scale deflection

From time to time the Osiris monitor reached its full scale deflection limit. This means that when particulate concentrations near the monitor reach a certain concentration, coincidence deflections occur within the Osiris, i.e. multiple particles are not differentiated and are measured as one. This was indicated by the following concentration measurements recorded by the Osiris:

• TSP is 6527.9 µg/m3 • PM10 is 6527.9 µg/m3 • PM2.5 is 652.79 µg/m3

Concentrations of this magnitude recorded by the Osiris indicate that particulate levels at the monitor were equal or greater to this concentration. An example of the data when the full scale deflection limit is shown in Figure A1. On the occasions when the measurements reached the full scale deflection limit, there was not a gradual increase in concentrations prior to the reading. These readings appeared to be isolated events (with the exception of 10 January 2013) and therefore were removed from the data set. Of these 29 records, only one train was passing at the time of the record. Table A1 presents the times when data due to the full scale deflection limit being reached was removed from the raw data set.

Table A1 Data removed from the raw data set due to full scale deflection Date and time record

removed 6 second TSP

µg/m3 6 second PM10

µg/m3 6 second PM2.5

µg/m3 3-Dec-12 12:08:16 6527.9 6527.9 652.79 3-Dec-12 12:08:22 6527.9 6527.9 652.79 3-Dec-12 12:08:28 6527.9 6527.9 652.79 12-Dec-12 14:15:27 6527.9 6527.9 652.79 12-Dec-12 14:15:33 6527.9 6527.9 652.79 12-Dec-12 14:15:39 6527.9 6527.9 652.79 12-Dec-12 14:15:45 6527.9 6527.9 652.79 12-Dec-12 14:15:51 6527.9 6527.9 652.79 12-Dec-12 14:15:57 6527.9 6527.9 652.79 12-Dec-12 14:16:03 6527.9 6527.9 652.79 12-Dec-12 14:16:09 6527.9 6527.9 652.79 12-Dec-12 14:16:15 6527.9 6527.9 652.79 12-Dec-12 14:16:21 6527.9 6527.9 652.79 12-Dec-12 14:16:27 6527.9 6527.9 652.79 30-Dec-12 15:49:12 6527.9 6527.9 652.79 30-Dec-12 15:49:18 6527.9 6527.9 652.79 30-Dec-12 15:49:24 6527.9 6527.9 652.79 30-Dec-12 15:49:30 6527.9 6527.9 652.79 2-Jan-13 13:00:14 6527.9 6527.9 652.79 3-Jan-13 12:55:31 6527.9 6527.9 652.79 3-Jan-13 12:55:37 6527.9 6527.9 652.79 3-Jan-13 12:55:43 6527.9 6527.9 652.79 3-Jan-13 12:55:49 6527.9 6527.9 652.79 3-Jan-13 12:55:55 6527.9 6527.9 652.79 3-Jan-13 12:56:01 6527.9 6527.9 652.79


May 2013 Page 52

Date and time record removed

6 second TSP µg/m3

6 second PM10 µg/m3

6 second PM2.5 µg/m3

3-Jan-13 12:56:07 6527.9 6527.9 652.79 10-Jan-13 9:34:14 6527.9 6527.9 259.48 10-Jan-13 9:35:08 6527.9 6527.9 411.47 16-Jan-13 8:40:21 6527.9 6527.9 652.79

A2 Corrupt data

On five days, very high concentrations were reached at the end of a data capture period (the time when the internal buffer of the Osiris reached capacity and the data was downloaded). An example of this behaviour is shown in Figure A2. While a ramp up in concentration values is observed on these occasions, advice from the Osiris supplier indicates these readings may be due to an error in the write procedure within the Osiris and therefore may not be accurate concentration data. On these occasions, data points noticeably higher than the typical values within the data capture period were removed. A total of 21 data points were removed for each pollutant (Table A2)

Table A2 Data removed due to data corruption Date and time record removed 6 second TSP

(µg/m3) 6 second PM10

µg/m3 6 second PM2.5

µg/m3 4-Dec-12 12:52:28 3032.7 3032.7 303.27 4-Dec-12 12:53:04 6451.4 1452.3 54.43 4-Dec-12 12:53:34 798.7 796.8 79.68 4-Dec-12 12:53:46 2415.9 2415.9 241.59 7-Dec-12 10:37:37 3032.7 3032.7 303.27 7-Dec-12 10:37:43 0 0 0 7-Dec-12 10:38:13 6451.4 1598.7 54.43 7-Dec-12 10:38:43 798.7 796.8 79.68 7-Dec-12 10:38:55 2415.9 2415.9 241.59 8-Dec-12 2:38:37 17.6 8-Dec-12 5:55:11 3032.7 3032.7 303.27 8-Dec-12 5:55:47 6451.4 1730.3 54.43 8-Dec-12 5:56:17 798.7 796.8 79.68 8-Dec-12 5:56:29 2415.9 2415.9 241.59 14-Dec-12 8:40:52 488 437 12.82 14-Dec-12 8:40:58 3032.7 3032.7 303.27 14-Dec-12 8:41:34 6451.4 2386.2 54.43 14-Dec-12 8:42:04 798.7 796.8 79.68 14-Dec-12 8:42:16 2415.9 2415.9 241.59 2-Jan-13 14:04:44 6451.4 4692.8 54.43 2-Jan-13 14:05:14 798.7 796.8 79.68


May 2013 Page 53

A3 Elevated concentrations

Elevated concentrations of TSP and PM10 were measured on 10 January 2013. Review of the data indicated that there was not one isolated elevated reading, but a one hour period between 8:42 am and 9:42 am when the 6 second data measured for both TSP and PM10 was elevated (Figure A3). The wind was from the northeast to east and winds were light to moderate. It is unlikely, that these elevated concentrations were due to phenomenon such as bushfires or dust storms due to the event being one hour in duration. The NSW EPA’s monitoring network did not indicate that air quality in the region was poor during this time either. A subcontractor attended the site at this time and noted substantial particulate was being produced by lawn mowing at a neighbouring business. It is likely, that this is the source of the elevated readings. These readings were kept in the data set.


May 2013 Page 54

Figure A1 Example of data validation: example of full scale deflection limit, data on dashed line was removed from analysis


Type: Scatter plot Data period: 30/11/12 to 29/1/13 Prepared by: N.Shaw Date: February 2013


May 2013 Page 55

Figure A2 Example of data validation: data between the dashed lines was excluded from the analysis


Type: Scatter plot Data period: 02/01/13 13:55 to 14:15 Prepared by: T. Haigh Date: February 2013


May 2013 Page 56

Figure A3 Timeseries of TSP and PM10 during 10 January 2013


Type: Timeseries Data period: 10/01/13 8am to 10am

Prepared by: N. Shaw Date: January 2013


May 2013 Page 57

Appendix B Statistical analysis of concentrations

B1 Variations in concentrations with train speed, wind direction and wind speed

Table B1 presents statistics for TSP, PM10 and PM2.5 based on train speed for all trains. Figure B1 to Figure B3 presents box and whisker plots of concentration based on train speeds for each train type for TSP, PM10 and PM2.5. Table B2 to Table B4 presents statistics for TSP, PM10 and PM2.5 based on wind direction for all data (train and no trains. Table B5 presents statistics for TSP, PM10 and PM2.5 when wind direction is from the rail tracks to the monitor based on train type. Also included are concentrations when no trains were passing the monitor. Figure B4 to Figure B6 presents the upper and lower 95% confidence limits when wind direction is from the rail tracks to the monitor for each train type for TSP, PM10 and PM2.5. Figure B7 presents box and whisker plots of concentration when wind direction is from the rail tracks to the monitor for each train type for TSP, PM10 and PM2.5 Table B6 presents statistics for TSP, PM10 and PM2.5 based on wind speed for all trains. Figure B8 to Figure B10 presents box and whisker plots of concentration based on wind speed for each train type for TSP, PM10 and PM2.5. The box represents the 75th and 25th percentile concentrations and the upper and lower whiskers represent the 95th and 5th percentile concentrations. The median is represented by a dash within the box.


May 2013 Page 58

Table B1 Particle concentrations for all trains by train speed

Parameter TSP PM10 PM2.5

5 to 30 km/hr

30 to 60 km/hr

60 to 90 km/hr

>90 km/hr

5 to 30 km/hr

30 to 60 km/hr

60 to 90 km/hr

>90 km/hr

5 to 30 km/hr

30 to 60 km/hr

60 to 90 km/hr

>90 km/hr

Number of Trains 4 1062 2398 1320 4 1062 2397 1320 4 1062 2396 1319

Mean 48.1 41.4 39.9 36.2 37.8 30.7 29.8 26.6 9.7 10.0 9.8 9.1 Median 55.1 37.2 34.4 30.5 41.5 27.6 26.1 23.8 10.1 8.9 8.5 8.3

Maximum 64.8 369.9 1173.4 390.9 54.2 188.2 1121.2 354.4 13.3 59.1 249.3 175.3 Standard Deviation ND 24.4 41.9 27.5 ND 17.5 33.2 18.6 ND 6.5 8.2 7.6


Level ND 42.9 41.6 37.6 ND 31.8 31.2 27.6 ND 10.4 10.1 9.5


Level ND 40.0 38.3 34.7 ND 29.7 28.5 25.6 ND 9.6 9.5 8.7

95th percentile ND 80.0 81.0 81.6 ND 60.9 61.1 54.8 ND 22.1 21.9 19.7

5th percentile ND 11.8 7.5 5.9 ND 8.8 6.2 5.3 ND 2.5 2.2 2.1

Table note: ND – Sample size is too small to calculate values


May 2013 Page 59

Table B2 TSP concentration for all data (trains and no trains) by wind direction

Parameter TSP

N NE E SE S SW W NW Number of samples 12891 46541 68063 93311 63793 45919 90647 43011

Mean 35.2 30.5 30.0 37.0 43.2 45.2 47.5 43.0 Median 27.0 22.7 23.8 31.0 37.6 40.4 41.7 36.3

Maximum 3183.7 4541.8 5538.4 2255.1 4543.9 1154.6 3605.9 1152.1 Standard Deviation 49.2 72.9 58.4 31.3 39.0 32.3 40.5 35.1

Upper 95% Confidence Level 36.1 31.2 30.4 37.2 43.5 45.5 47.8 43.4 Lower 95% Confidence Level 34.4 29.9 29.6 36.7 42.9 44.9 47.3 42.7

95th percentile 84.2 68.7 68.2 81.7 92.4 97.2 104.3 97.8 5th percentile 5.1 5.9 5.0 8.2 10.0 8.3 7.5 7.7

Table B3 PM10 concentration for all data (trains and no trains) by wind direction

Parameter PM10

N NE E SE S SW W NW

Number of Samples 12891 46541 68063 93310 63792 45919 90647 43011 Mean 26.1 23.0 23.1 28.6 33.0 33.9 33.9 30.1

Median 20.5 18.2 19.6 25.0 29.5 31.1 30.2 25.5 Maximum 2659.3 4453.1 5538.3 1766.5 4542.4 1122.6 3578.7 1145.6

Standard Deviation 34.7 55.5 46.0 21.7 32.2 22.9 30.6 24.0 Upper 95% Confidence Level 26.7 23.5 23.4 28.7 33.3 34.1 34.1 30.3 Lower 95% Confidence Level 25.5 22.5 22.7 28.4 32.8 33.7 33.7 29.9



May 2013 Page 60

Table B4 PM2.5 concentration for all data (trains and no trains) by wind direction

Parameter PM2.5

N NE E SE S SW W NW

Number of Samples 12891 46541 68063 93310 63792 45919 90647 43011 Mean 8.6 8.1 8.6 10.5 11.3 11.0 9.9 9.0

Median 6.5 6.7 7.5 8.9 9.4 8.9 7.8 6.6 Maximum 378.7 627.7 640.5 564.7 617.4 433.9 622.5 550.5

Standard Deviation 8.8 11.5 11.1 10.2 10.3 11.3 11.7 11.9 Upper 95% Confidence Level 8.8 8.2 8.7 10.6 11.4 11.1 10.0 9.1 Lower 95% Confidence Level 8.5 8.0 8.5 10.5 11.2 10.9 9.8 8.9



May 2013 Page 61

Table B5 Concentration by train type when wind direction is between 150 degrees to 300 degrees

TSP PM10 PM2.5

Loaded coal

Unloaded

coal Freight Passe

nger No

trains Loaded

coal Unload

ed coal

Freight Passenger

No trains

Loaded

coal

Unloaded coal

Freight

Passenge

r No

trains

Number of Trains 351 442 52 964 172923* 351 441 52 964 172923* 351 441 52 964 172923*

Mean 49.0 53.1 49.3 43.3 43.0 36.0 39.3 35.5 32.1 31.7 11.2 12.2 10.4 10.0 10.1 Median 47.4 50.9 49.8 38.0 38.0 34.0 37.5 38.4 28.5 28.6 9.6 10.6 10.2 8.6 8.2

Maximum 208.5 349.5 103.7 1124.9 4543.9 131.2 339.0 71.0 1121.2 4542.4 56.1 59.1 27.5 175.3 622.5

Standard Deviation 24.0 25.9 22.8 45.6 36.7 17.7 21.9 16.7 41.0 29.4 6.9 7.3 6.1 8.6 11.6


Level 51.5 55.5 55.5 46.2 43.2 37.9 41.4 40.1 34.7 31.8 11.9 12.9 12.0 10.5 10.1


Level 46.5 50.7 43.1 40.4 42.8 34.2 37.3 31.0 29.5 31.6 10.5 11.5 8.7 9.4 10.0

95th percentile 84.4 87.8 94.7 91.8 92.0 64.4 66.4 69.0 62.6 63.1 22.6 23.7 25.7 22.3 21.9

5th percentile 15.2 16.7 17.2 7.5 8.2 10.6 12.4 10.9 6.7 7.1 3.1 3.4 2.3 2.2 2.2 Table note: *Number of 6 second records


May 2013 Page 62

Table B6 Particle concentration for all trains by wind speed

Parameter TSP PM10 PM2.5

0 to 0.5 m/s

0.5 to 1m/s

1 to 2 m/s

2 to 5m/s

5 to 10m/s

0 to 0.5 m/s

0.5 to 1m/s

1 to 2 m/s

2 to 5m/s

5 to 10m/s

0 to 0.5 m/s

0.5 to 1m/s

1 to 2 m/s

2 to 5m/s

5 to 10m/s

Number of Trains 449 945 1115 1082 81 449 945 1114 1082 81 449 945 1114 1082 81

Mean 41.1 44.6 40.4 34.9 41.4 30.1 33.7 30.3 25.8 26.2 9.3 10.5 10.3 9.1 6.9 Median 35.5 38.6 34.7 29.5 36.4 26.3 28.5 26.6 22.9 24.0 7.9 8.9 9.2 7.9 5.4

Maximum 409.6 1124.9 1173.4 369.9 154.7 200.9 1121.2 770.9 339.0 78.7 37.0 175.3 249.3 61.3 24.3 Standard Deviation 29.3 44.1 42.7 28.9 25.4 18.8 40.9 28.8 20.1 13.3 5.9 8.7 9.3 6.3 5.0


Level 43.8 47.5 42.9 36.6 46.9 31.9 36.3 32.0 27.0 29.1 9.9 11.1 10.8 9.5 8.0


Level 38.3 41.8 37.9 33.2 35.8 28.4 31.1 28.7 24.6 23.3 8.8 10.0 9.7 8.7 5.8

95th percentile 81.9 86.3 81.7 77.5 84.0 60.2 64.9 60.9 53.4 47.7 21.4 23.1 21.7 21.1 20.0

5th percentile 10.7 10.0 8.7 6.9 8.0 7.8 7.9 7.3 5.8 7.5 2.3 2.6 2.6 2.5 2.6


May 2013 Page 63

Figure B1 TSP concentrations by train type and train speed

Location: Metford Averaging period: 6 second Data source: Osiris Units: km/hr

Type: Box and whisker plot Data period: 30/11/12 to 29/01/13

Prepared by: A. Vernon Date: May 2013


May 2013 Page 64

Figure B2 PM10 concentrations by train type and train speed



Prepared by: N. Shaw Date: May 2013


May 2013 Page 65

Figure B3 PM2.5 concentrations by train type and train speed





May 2013 Page 66

Figure B4 Comparison of TSP concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees

Location: Metford Averaging period: 6 second Data source: Osiris Units: Degrees

Type: Chart Data period: 30/11/12 to 29/01/13



May 2013 Page 67

Figure B5 Comparison of PM10 concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees





May 2013 Page 68

Figure B6 Comparison of PM2.5concentration upper and lower 95% confidence limits by train type for wind direction between 150 degrees and 300 degrees





May 2013 Page 69

Figure B7 Concentrations of TSP, PM10 and PM2.5 by train type for wind

directions between 150 degrees and 300 degrees

Location: Metford

Averaging period: 6 second

Data source: Osiris

Units: Degrees

Type: Box and whisker plot

Data period: 30/11/12 to 29/01/13

Prepared by: A. Vernon

Date: May 2013


May 2013 Page 70

Figure B8 TSP concentrations by train type and wind speed

Location: Metford Averaging period: 6 second Data source: Osiris Units: metres per second




May 2013 Page 71

Figure B9 PM10 concentrations by train type and wind speed





May 2013 Page 72

Figure B10 PM2.5 concentrations by train type and wind speed




pollution reduction program 4.2 particulate … coal dust reports...katestone environmental pty ltd...

Documents