AIR MONITORING DURING THE HAZELWOOD MINE FIRE
AIR MONITORING LOCATIONS
Figure 4.16 demonstrates the location of air monitoring sites during the Hazelwood mine fire.
Figure 4.16 Latrobe Valley monitoring sites – air25
Figure 4.17 demonstrates the location of carbon monoxide monitoring sites during the Hazelwood mine fire.
Figure 4.17 Latrobe Valley monitoring sites – carbon monoxide26
AIR MONITORING EQUIPMENT
Air quality monitoring equipment varies from continuous monitors installed in air-conditioned buildings, to mobile vehicles, right through to hand-held monitors. The equipment varies according to the pollutant measured, the method of monitoring and the timing and accuracy of the measurement. Monitoring equipment also varies in how complex it is as a piece of technology. Some monitoring equipment can be used instantaneously, such as pre-calibrated hand-held monitors, whereas other equipment can take anywhere from a day to a number of weeks to set up and calibrate in order to capture data accurately.
The EPA used a suite of air monitoring equipment to test air quality during the Hazelwood mine fire. Equipment included three full reference/fixed monitoring stations at Traralgon, Hourigan Road, Morwell (East) (see Figure 4.18), and the Morwell Bowling Club (South) (see Figure 4.19).
Figure 4.18 Fixed Air Monitoring Station – Hourigan Rd, Morwell (East) 27
Figure 4.19 MoLab – Morwell Bowling Club (South)28
In addition to the three fixed monitoring stations, a range of mobile air monitoring equipment was deployed in Morwell (and surrounding areas) that produced indicative data. This included the DustTrak that was placed at the Morwell Bowling Club (South) to measure PM2.5 (pending the arrival of the MoLab), canisters to measure volatile organic compounds, air visibility monitors to measure concentrations of particles, Area RAE monitors to measure carbon monoxide, and a TravelBLANkET from Environment Tasmania to measure PM2.5 (see Figures 4.20–4.24).29
This is not an exhaustive list of the air monitoring devices used, but gives the reader some indication of the diversity and in some cases, complexity of these pieces of scientific equipment.
Figure 4.20 DustTrak – Morwell Bowling Club (South)30
Figure 4.21 VOC Canister – Maryvale Crescent, Morwell31
Figure 4.22 ADR 1500 – Air Visibility Monitor32
Figure 4.23 Area RAE (carbon monoxide) at Morwell Bowling Club (South)33
From 20 February 2014, the TravelBLANkET was used to log data. Dr Torre told the Board that BLANkET stands for ‘Baseline Air Network of EPA Tasmania.’34
Figure 4.24 TravelBLANkET35
AIR MONITORING RESULTS
Figure 4.25 Validated and indicative visibility reduction levels for the Latrobe Valley from
9 February 2014 – 31 March 2014 (one hour average)36
Figure 4.25 adapted from an EPA graph shows validated data (solid line), and indicative data plotted retrospectively (dotted line) for one hour averages for visibility reduction in the Latrobe Valley from 9 February 2014 to 31 March 2014.
The State Environment Protection Policy (Ambient Air Quality) (State Ambient Air Quality standard) for visibility reduction, measured as the minimal visibility distance, is 20 kilometres for one hour.37
This standard is indicated in Figure 4.25 by the red line. It is not clear from the EPA graph how the standard of 20 kilometres for one hour relates to the red line. It is clear that the one hour averages of both validated and indicative data are above the standard on a number of occasions. However in his evidence to the Board, Dr Torre stated (in relation to 16 February 2014):
On that Sunday when I did come down, it was around about 5 o’clock, the visibility was down to, oh, less than a kilometre. We’ve got a table that sort of guides people in terms of trying to understand those levels, and it was at levels where, if you look at the categories, it’s called ‘hazardous’ – that’s very high levels.38
The Board is not in possession of the table Dr Torre refers to. However, according to the Californian Wildfire Smoke Protocol that was discussed during the hearings, the table used for estimating visibility reduction states that visibility less than one mile (approximately 1.6 kilometres) is classified as ‘hazardous’.39
Figure 4.26 Validated and indicative carbon monoxide levels for Morwell and Traralgon from
9 February 2014 – 31 March 2014 (8 hour average)40
The above Figure adapted from an EPA graph shows validated data (solid line) and indicative data plotted retrospectively (dotted line). The State Ambient Air Quality standard for carbon monoxide is 9 ppm averaged over eight hours. This standard is indicated in Figure 4.26 by the red line.
Both Dr Torre and independent expert Ms Claire Richardson, Managing Director and Principal Consultant, Air Noise Environment Pty Ltd, advised the Board that the levels of carbon monoxide recorded in Morwell on 21 February 2014, 22 February 2014 and 26 February 2014, exceeded the State Ambient Air Quality standard. They also advised that carbon monoxide levels were likely to have been exceeded on 15 February 2014 and 16 February 2014 (when only indicative data was available).41
The maximum value of carbon monoxide recorded (eight hour rolling average) in Morwell during the mine fire was 14 ppm. However, there are estimated readings of up to 34 ppm (eight hour rolling average) on the evening of 15 February 2014.42 There is no data on carbon monoxide levels available from 9 February to 11 February 2014.43
Particulate matter (PM2.5)
Figure 4.27 Validated and indicative PM2.5 levels for Morwell and Traralgon from
9 February 2014 – 31 March 2014 (daily averages)44
Figure 4.27 adapted from an EPA graph shows validated data (solid line) and indicative data plotted retrospectively (dotted line) for daily averages of PM2.5 in Morwell from 9 February to 31 March 2014. The Victorian standard (advisory) for PM2.5 is 25 µg/m3 measured over one day. This standard is indicated in Figure 4.27 by the red line.
The graph shows that there were three peaks (above the red line) of increased levels of PM2.5 between 15 February 2014 and 16 February 2014, and around 21 February 2014 and 26 February 2014.
In their joint expert report to the Board, Dr Torre and Ms Richardson advised that from 14 February 2014 until 31 March 2014, in the area south of Commercial Road, Morwell, there were:
- 21 days where the levels of PM2.5 exceeded the advisory reporting standard (greater than 25 µg/m3)
- seven days that would be classified as hazardous in the PM2.5 Health Protection Protocol (equal or greater than 157 µg/m3)
- four days where the levels (indicative and validated) were in the extreme category in the PM2.5 Health Protection Protocol (greater than 250 µg/m3).45
Due to the need for scientific calibration of data, indicative data was not available to the EPA and the Department of Health until after the fire was controlled (which was on 10 March 2014).46
The highest validated recording of PM2.5, as shown in Figure 4.27 was over 400 µg/m3 (on about 21 February 2014). This is approximately 16 times the daily National Ambient Air Quality standard of 25 µg/m3.
The highest indicative recording of PM2.5, as shown in Figure 4.27, was over 700 µg/m3 between 15 and 16 February 2014. This is approximately 28 times the daily National Ambient Air Quality standard of 25 µg/m3.
Figure 4.28 Indicative PM2.5 levels from the DustTrak at the Morwell Bowling Club (South) from 13 February 2014 – 20 February 2014 (hourly averages)47
Figure 4.28 adapted from an EPA graph shows indicative data for hourly averages of PM2.5 from the DustTrak monitor that was located at the Morwell Bowling Club (South) from 13 February 2014, until the MoLab was made ready and transported there as a fixed monitoring station. The maximum recorded PM2.5 level was just below 1,200 µg/m3 on 17 February 2014. There is no hourly standard for PM2.5.
Figures 4.29 and 4.30 below show the instantaneous PM2.5 levels measured by the TravelBLANkET while driving around Morwell on 22 February 2014. Height and colour represent concentrations of PM2.5 measured by the TravelBLANkET. Red markers indicate higher concentrations and blue markers indicate lower concentration levels. These images clearly show a greater level and distribution of fine particles closest to the mine site, in the southern part of Morwell. These images were not made publicly available by the EPA during the course of the mine fire.
Figure 4.29 Instantaneous PM2.5 levels measured by the TravelBLANkET on 22 February 2014 (aerial view)48
Greater than 250 µg/m3
Between 100 – 250 µg/m3
Between 0 – 100 µg/m3
Figure 4.30 Instantaneous PM2.5 levels measured by the TravelBLANkET on 22 February 2014 (looking east)49
Greater than 250 µg/m3
Between 100 – 250 µg/m3
Between 0 – 100 µg/m3 50
Particulate Matter (PM10)
Figure 4.31 Validated and indicative results for PM10 levels in Morwell and Traralgon from
9 February 2014 – 31 March 2014 (daily averages)51
Figure 4.31 adapted from an EPA graph shows validated data (solid line) and indicative data plotted retrospectively (dotted line) for daily averages of PM10 in Morwell and Traralgon from 9 February 2014 to 31 March 2014. The Victorian standard for PM10 is 50 µg/m3 averaged over one day. This standard is indicated on Figure 4.31 by the red line.
The EPA prioritised monitoring carbon monoxide and PM2.5 during the mine fire, as these are of most concern to human health.52 This explains why validated data for PM10 was not monitored at the Morwell Bowling Club (South) until around 27 February 2014.
The highest indicative recording of PM10 in Figure 4.31 is just below 1,600 µg/m3 (on around 15–16 February 2014). This is approximately 30 times the National Ambient Air Quality standard.
Figure 4.32 Validated sulphur dioxide levels for Morwell and Traralgon from
9 February 2014 – 31 March 2014 (1 hour average)53
Figure 4.32 adapted from an EPA graph shows that sulphur dioxide levels in Morwell and Traralgon did not exceed the State Ambient Air Quality standard over the course of the Hazelwood mine fire.54
Figure 4.33 Validated nitrogen dioxide levels for Morwell and Traralgon from
9 February 2014 – 31 March 2014 (1 hour average)55
Figure 4.33 adapted from an EPA graph shows that nitrogen dioxide levels in Morwell and Traralgon did not exceed the State Ambient Air Quality standard over the course of the Hazelwood mine fire.56
Volatile organic compounds
In his statement to the Board, Mr Merritt noted that all but one of 14 of the volatile organic compounds measured were many times lower than the State Ambient Air Quality standard. Only benzene exceeded the assessment criterion of nine parts per billion, at the following locations:
- Maryvale Crescent Preschool – 9.2 parts per billion on one occasion (date not specified). There were no children at the Preschool during the Hazelwood mine fire.
- Morwell Bowling Club (South) – 14 parts per billion on 26 February 2014 and 9.7 parts per billion on 27 February 2014.
The EPA informed the Department of Health and the Chief Health Officer of these results by email. Monitoring of benzene is included as part of the EPA’s monitoring in Morwell. This monitoring will continue for at least 12 months to March 2015.57
Figure 4.34 Validated ozone levels for Morwell and Traralgon from 9 February 2014 – 31 March 2014 (1 hour average) 58
Figure 4.34 adapted from an EPA graph shows that ozone levels in Morwell and Traralgon did not exceed the State Ambient Air Quality standard over the course of the Hazelwood mine fire.59
AIR QUALITY MONITORING IN AN EMERGENCY
Ms Richardson set out the three stages of an effective monitoring response in an emergency:
Stage 1: initial emergency response: subjective response and monitoring using hand-held emergency type instrumentation intended to screen for extremes in pollution exposure.
Stage 2: rapid deployment of relatively low cost portable monitors that offer a known degree of accuracy, are able to be installed quickly and easily, and can be installed at a number of locations to provide reasonable spatial coverage.
Stage 3: implementation of a high quality monitoring station to provide accurate data that is compliant with the relevant national standards.60
Ms Richardson advised the Board that this general approach was followed during the Hazelwood mine fire.61 However, she noted that where an incident is expected to cause significant air emissions that extend for longer than a day, additional monitors should be considered and the initial response should involve installation of temporary portable monitoring stations as soon as possible. Importantly, although these monitors would not be full ‘reference’ monitors that comply with the National Ambient Air Quality standard, the instruments would be expected to provide greater certainty in terms of data accuracy than the initial emergency response monitoring that occurred in this instance.62
CHALLENGES WITH AIR QUALITY MONITORING DURING THE HAZELWOOD MINE FIRE
In their joint report, Ms Richardson and Dr Torre told the Board that:
The Morwell South [monitoring] station was [in] the optimum position and allowed determination of the worst case community exposure. In an ideal world, the station would have been operational earlier in the incident.63
The monitoring conducted by the EPA from 13 February 2014 onwards at Hourigan Road, Morwell (East) also provided valuable data.
However, in her joint report with Dr Torre, Ms Richardson made the following statement:
Installation of a ‘reference’ monitoring station of the quality installed at Morwell East was not necessary to inform the emergency response. However, it was fortunate that the station could be re-commissioned quickly, and the monitoring provided useful additional measurements to assist with the overall community monitoring. I would be concerned if the re-commissioning of Morwell East took precedence over implementing the monitoring station at Morwell South and the broader monitoring using hand held instruments. I understand from discussions with Dr Torre that this is not the case. Dr Torre has explained that the implementation of Morwell South and the portable monitoring efforts were completed in parallel with the re-commissioning of Morwell East.64
Readiness of air monitoring response
Currently, the EPA lacks a comprehensive emergency response capability. As Dr Torre has noted:
It would be very rare for me, or other EPA Science Officers or Laboratory Emergency Response Officers to attend an emergency event, with air monitoring equipment. EPA has very limited air monitoring equipment for measuring air emissions from emergency incidents.65
Dr Torre also told the Board that when called out to an emergency event such as an oil spill, the EPA relies on the hand-held monitoring equipment it knows emergency service personnel carry on them as part of their occupational health and safety practices, to help with an initial assessment of the situation.66 In his evidence to the Board, Dr Torre explained that when he was first called to the Hazelwood mine fire he utilised the Country Fire Authority’s (CFA) equipment to measure the carbon monoxide levels around the perimeter of the mine.67 That is, the EPA used the CFA’s carbon monoxide detection equipment to measure levels in the community. Dr Torre stated that ‘it is not uncommon to work with the Fire Brigade to try and understand impact assessment’.68
In his statement to the Board, Mr Merritt commented that the EPA may need to consider updating its mobile monitoring and modelling equipment based on a history of previous incidents and predictions of future fire events.69
As Ms Richardson noted:
In this situation there needs to be an [sic] whole of government rapid response approach allow [sic] air quality to be monitored in a short period of time. This needs to include maintaining appropriate instrumentation resources that are suitable for rapid deployment, and trained personnel to operate the instrumentation during emergency events.70
There were some delays in establishing and maintaining monitoring at various locations in Morwell and surrounding areas due to the availability and serviceability of monitoring equipment. The EPA also encountered a number of technical challenges in relation to air monitoring. This is not unusual in itself as the equipment used is highly sensitive. However there was both an urgency and high degree of pressure involved in fixing issues as they arose, due to the nature of the event.
The first set of technical challenges occurred with the decision in the first week to set up a reference monitoring station close to the fire in the southern part of Morwell. Individual testing equipment had to be sourced from within the current network, and then checked, calibrated and brought up to standard in tight timeframes. The MoLab needed mechanical work. The MoLab site needed power and data communications needed configuring. None of these components presented a difficulty individually, but difficulty was experienced in trying to perform all of these functions very quickly. A process that might normally take four weeks was completed in four days.71
The MoLab that had become the fixed monitoring site at Morwell Bowling Club (South) developed a leak in the roof on two occasions. This was attended to in the first instance with tape, and in the second with a tarpaulin. There was no loss of data.72
There were early difficulties with the transfer of carbon monoxide data from the CFA to the EPA. Technologies at the two organisations were incompatible–such a transfer had never been attempted previously. The issue was addressed through a labour-intensive manual transfer, until an automated process could be finalised. By the second week of the response, an EPA officer was specifically tasked with doing this transfer each day. The process was streamlined by the second week of March 2014. Mr Merritt told the Board that in the future appropriate arrangements will be made with the CFA.73
- On 14 February 2014, the newly installed modem at Hourigan Road, Morwell (East) failed. This did not result in any loss of data (it was still being recorded inside the station), but it did result in a delay of 32 hours in getting the data streaming directly onto the EPA website.74
- On 16 February 2014, there was a software problem on the main data collection system that required a system restart. Again there was no loss of data, but the issue caused a 10 hour gap in the website display.75
- On 19 February 2014, the wind direction sensor at the Morwell Bowling Club (South) was twisted by strong winds and there was a loss of local wind direction data for 42 hours until it could be replaced.76
- On 22 March 2014, the modem in the particle monitor at Moe failed, and this took 60 hours to replace due to its specialised nature. Again data was not lost, however this data set was not available for the website.77
- On 27 March 2014, the air monitoring equipment at Morwell Bowling Club (South) had to be cleaned (due to the heavy smoke impacts since deployment). This resulted in a data loss of five hours.78
Access to laboratories to analyse data
The EPA has one contract with one National Association of Testing Authorities (NATA) accredited laboratory and a backlog was created with the number of samples the EPA was sending for analysis.79
Dr Torre told the Board that there was an issue with sufficient access to laboratories during the mine fire and that he considered that more laboratories were required.80
Data accuracy and analysis
During the mine fire, the EPA requested both interim and final results from its laboratories. ‘Interim’ and ‘final’ results refer to the life cycle of a data set that is produced from a fixed or permanent monitoring station.
Interim results were requested as soon as they were available so that the EPA had as much information as possible as soon as possible. It is the experience of the EPA that it is very rare that final results differ from the interim results.81
Interim results are not to be confused with indicative results. Indicative data is data taken from hand-held or portable air quality monitoring equipment, rather than data logged from a permanent monitoring station that automatically streams data back to the EPA. Indicative data may not be 100 per cent accurate, but it provides the best estimate of air quality in the absence of data validated through the EPA’s processes.
There is merit in using indicative data in emergency response situations as it can be accessed promptly to help decision makers.
As Mr Merritt told the Board:
In circumstances where there is a lack of available data, the expectations and needs for accuracy are appropriately limited. During the early stages of the Hazelwood Mine Fire when air quality readings exceeded limits in the measurement standards, absolute accuracy of the readings was not the highest priority. Indicative data was sufficient.82
PM10 data collected from the permanent monitoring station in Traralgon during the mine fire was able to indicate what the air quality was like in Morwell. This is because Traralgon is east from Morwell (and the mine fire) and PM10 data readings correlate with PM2.5. In other words, if PM10 levels are raised in Traralgon, PM2.5 levels will be raised in Morwell. The PM10 data from the permanent monitoring station in Traralgon indicated that the peak air quality impacts were likely to have hit Morwell just after midday on 9 February 2014, with a slightly higher peak occurring around midday on 10 February 2014.83
Ms Richardson told the Board that a key feature of effective air quality monitoring is immediate access to monitoring data to inform the emergency response.84
In their joint report, Ms Richardson and Dr Torre also noted that having the capability to download carbon monoxide data remotely via a modem would allow for rapid access to that data.85
ONGOING AIR QUALITY MONITORING
The EPA is continuing to conduct air quality monitoring in Morwell and surrounding areas in addition to the permanent monitoring station in Traralgon (which now also captures PM2.5) and will do so for at least the 12 months following the Hazelwood mine fire. This monitoring will include:
- collecting data for PM2.5 and visibility as well as gaseous particles (sulphur dioxide, carbon monoxide, nitrogen oxides and ozone) at the Morwell Bowling Club (South)
- a subset of the above compounds will also be collected at Hourigan Road, Morwell (East)collecting data and interpreting results for both respirable silica and polycyclic aromatic hydrocarbons at Morwell (South)
- particlulate matter sampling at Moe and Churchill
- passive samplers at three locations across Morwell to collect data about volatile organic compounds.86
The reach and extent of smoke over an area is influenced by a number of factors, including the success of fire suppression, wind direction and speed, and temperature.
The amount of smoke emitted into the atmosphere and the impact of this smoke on Morwell and the Latrobe Valley varied over the 45 days that the Hazelwood mine fire burned. At its worst, visibility was less than one kilometre and a blanket of acrid choking smog infiltrated people’s homes, businesses and public buildings.87 On other days the air quality was considered good or fair, and visibility was reasonable.
Ms Brooke Burke, Morwell Business Owner, told the Board that ‘each day [of the mine fire] was very different, so some days it was clear at the studio, some days it was very bad at home. Some days at home it was clear and some days it wasn’t as good at the studio.’88
Dr Torre explained to the Board that during the mine fire south-westerly winds had the greatest impact on the distribution of smoke in Morwell.89
Figure 4.35 Validated PM2.5 levels and corresponding wind direction for Morwell from
21 February 2014 – 15 March 2014 (24 hour average)90
Figure 4.35 demonstrates the levels of PM2.5 recorded at the Morwell Bowling Club (South) air monitoring station between 22 February 2014 and 15 March 2014. The red line indicates the State Advisory standard for PM2.5.
The Figure shows that when the wind blew in a south-westerly direction the recorded levels of PM2.5 increased. This is particularly evident on 21–23 February 2014 and 26–28 February 2014 when the levels of PM2.5 were significantly higher than on other days.