This work is copyright. Apart from any use as permitted under the Copyright Act 1968, no part may be reproduced by any process without prior written permission from the Commonwealth

15 ppb 122 ppb

0.015 g/s emissions as volume source from doors 0.022 g/s emissions as volume source from roof ridge modelling includes building wake effects

2 ppb 37 ppb

4. Uncertainties in the fugitive emission estimate

Because the fugitive emission estimate (≈1000 kg p.a.) is determined as the difference between two large numbers (27,000 and 25,950), it is critical that the impact of uncertainty in these large numbers be taken into account in the calculation. The EML report does not include any uncertainty estimates. [Note that the EML report uses both kg

1. 
   and g/sec emission estimates; rounding errors in the conversion produce small differences when comparing the numerical values estimated in different units.] Uncertainties in the estimate of the magnitude of fugitive emissions arise from (in approximate order of importance):
   
   1. 
      uncertainty in the NPI estimate of total formaldehyde emissions;
      
   2. 
      uncertainty in the total stack emission rates derived from the stack testing on up to 16 separate stacks;
      
   3. 
      uncertainty due to differences in the production rate assumed for the NPI estimates and that prevailing at the time the stack testing was undertaken.
      

1. 
   Uncertainty in NPI estimates
   

given in the manual, it would appear reasonable to conclude that these descriptors correspond to an uncertainty of at least 20% in the NPI estimates (i.e. 27,000 ± 5,400 kg p.a.). In the estimate below, we consider the impact of uncertainties of 20%, 10% and 5% in the NPI estimate.

2. 
   Uncertainty in total stack emission rate
   

NPI estimate, we ignore this uncertainty in estimating the uncertainty in the fugitive emission rate.

3. 
   Uncertainty due to differences in the production rate
   

in the fugitive emission rate.

The EML report determined the fugitive emission rate to be 0.037 g/sec. This is equal to 1170 kg p.a. Taking into account possible uncertainties of 20%, 10% and 5% in the NPI total emission rate would give the following averages and range of values for the fugitive emissions:

• 
  1170 kg p.a. (range 0 – 6600 kg p.a.) for 20% uncertainty in NPI estimate
  
• 
  1170 kg p.a. (range 0 – 3900 kg p.a.) for 10% uncertainty in NPI estimate
  
• 
  1170 kg p.a. (range 0 – 2500 kg p.a.) for 5% uncertainty in NPI estimate.
  

These indicate that even a 5% uncertainty in the NPI estimate could more than double the fugitive emissions. Scaling the PEC results in Table 18A-1 produces the following range of maximum 24-hour average PECs:

• 
  0 kg p.a. fugitives 33 ppb (only stack impact)
  
• 
  2,500 kg p.a. fugitives 79 ppb
  

That is, the model results are very sensitive to uncertainties in the NPI estimate. Because the 24-hour average PECs are significant compared to NEPM Air Toxics formaldehyde investigation level (40 ppb, 24-hour average) and are the largest identified in the NICNAS report, further work is needed to reduce the uncertainty in the fugitive emission estimate.

5. Possible method for reducing uncertainty in fugitive estimate

It is very difficult to see how to reduce the uncertainty in the fugitive emission estimate computed as the difference between the NPI estimated total emissions and the measured stack total.

The alternative is to use some other method for estimating fugitive emissions, for example to measure/estimate volume flows (air exchange rates) and 24-hour average formaldehyde concentrations in the fugitive emissions from the roof ridge and doors.

6. Justification for 40:60 split of fugitives between door:roof ridge

No justification is provided for using a 40%:60% split of the fugitive emissions between the door and roof ridge. Is this important? This could be determined by looking at the sensitivity of the PEC estimates to changes in this split?

7. Ausplume Modelling

We consider that the set-up of the model, the inclusion of buildings, stacks and fugitive sources is appropriate.

The 24-hour average PECs were calculated on a radial grid at 5º increments and at spacings from 100 m to 1000 m at 100 m increments. This is considered to be suitable for determining the maximum 24-hour average PEC.

However, the annual average concentrations were only determined at a single discrete receptor, which was located 100 m east of the main plant building. This is indicated by

the data listed in Attachment 3 (Air Dispersion Model Outputs). Although this may be the location of the highest annual average PEC for the current fugitive/stack emission characteristics it is possible that this point of maximum concentration will change for other configurations. The annual average PECs should be calculated on the same grid as that used for the 24-hour averages.
The EML Ausplume modelling used a 1997/1998 meteorological file for Paisley in the west of Melbourne. This is the same as that used by CSIRO. Although meteorology from this site was appropriate for the CSIRO assessment of the impact of emissions from a wide range of industries as well as urban and roadway sources, it is recommended that for modelling the impact of the largest formaldehyde emitter (in the wood and paper manufacturing industries), it would be more appropriate to use a local meteorological file (either from measurements or generated by TAPM) for the site of the largest emitter.

8. Other comments

The end of Section 2 of the EML report lists a number of observations, which it is suggested indicate “that the modelling by CSIRO for NICNAS has generally over- predicted ground-level concentrations of formaldehyde at the boundary of wood product industries, particularly in the case of the largest emitters”. We list these observations and provide our response to each.

• 
  EML report: Typical Australian State EPA dispersed design criteria for formaldehyde as 3-minute average < 33 ppb or 1-hour average ≤ 18 ppb, which would be substantially lower if modelled at the longer averaging period of 24-hours and annual.
  

CSIRO response: The implication of including this statement is that all plants meet current EPA design criteria. However, even the modelling in the EML report indicates that emissions from the stacks alone (ignoring the contribution from fugitives) produce 24-hour average PECs up to 33 ppb at a distance of 100 m from the source. This is substantially larger than the 18 ppb 1-hour average listed as an EPA criterion. It is modelling and measurements that are relevant for determining PECs, not current standards.

• 
  EML report: Fugitive emissions concentrations are predominantly reflected by measurement of workplace exposure. The wood panels industry has data indicating that most workplaces are < 20 ppb (8-hour average). At the perimeter, 100 m from the sources of these fugitive concentrations, the ambient concentration on a 24-hour and annual average will be much lower.
  

from “most workplaces”. Furthermore, the relevance of workplace exposure measurements to the concentration in the fugitive emissions depends on the ventilation arrangements in the facility. Ventilation systems are generally designed to minimise workplace exposure, for example by drawing fresh air in around the workspace and

exhausting it through roof vents. Thus measurements of workplace concentration could be

far lower than concentrations in fugitive exhaust air.

• 
  EML report: Ambient measurements at the boundary of two wood panel facilities in Australia have reported <10 ppb 24-hour averages.”
  

measurements made under conditions when the maximum 24-hour averages could be