Appendix 6. Odour Buffer Assessment [As Published] · 132007.2399

                                                                                                             [21]
                    Odour Buffer Assessment
                    StarCo Developments


                                     Figure 8.1: Field Observation Locations

                              Table 8.2: Odour Intensity Observations – Details
Observation 2                MGA 1          Time       Duration of            Site Description
                                                      Observation
     1 – Dec 21         508458. 7078519.          16:00              Observation aborted – odour not detected
     2 – Jan 11         508521. 7078477.          10:20         15 mins                   Northwest of STP
     3 – Jan 11          508507. 7078502          11:15              Observation aborted – odour not detected
     4 – Jan 11         508459. 7078462.          11:24         15 mins                   Northwest of STP
     5 – Jan 11         508458. 7078519.          11:50         15 mins                   Northwest of STP
1.    Australian Geodetic Datum (1966)
2.    Observations 2-5 corresponded to odour sampling by Airlabs


Analysis of the odour intensity data detailed in Appendix B allows estimation of mean odour
concentration, which can be compared to odour modelling predictions. The observations also allow
estimation of the maximum odour concentration during each observation period, and the peak-to-
mean ratio. The results are contained in Table 8.3.

          Table 8.3: Estimated odour concentrations from observed odour intensity observations 1
                                     Mean odour                            Maximum odour
                                                                                                    Peak:Mean
                                  concentration (ou)                      concentration (ou)
Observation
                         Max.            Min.              Median              Average               Average
                       Estimate        Estimate           Estimate             Estimate              Estimate
          2               0.91            0.75              0.84                   2.9                   3.5
          4               0.63            0.47              0.54                   1.8                   3.3
          5               0.48            0.32              0.39                   1.2                   3.1
     1.       Average of paired observations. The relationship between odour and intensity as determined by
              Airlabs is expressed as: I = 2.07 Ln(Conc) + 1


A comparison between the predicted model results and field odour observations can be made in
order to validate the results of odour dispersion modelling. This requires comparing the observed
(estimated) and modelled odour concentrations at the observation location under the specific
weather conditions at the time of the observations.

Any comparison of modelled and observed concentrations must take into account the uncertainty
inherent in both sets of data. The uncertainty in the estimated odour concentrations based on
intensity observations can be expressed simply by the range of estimated mean concentrations.

Models predict so-called ensemble mean concentrations. For a given set of meteorological
conditions, averaged over 1 hour, the resulting ground level concentration predicted by the model
is the average that would occur under those conditions (assuming the model is accurate). In any
specific hour, however, turbulence in the atmosphere will cause variations in actual concentrations,
so that even a “perfect” model will not capture these random variations. Hence, when the model
predicts a certain result, the actual concentrations in the field under the same conditions may differ
significantly from the model result, purely by this chance or random factor. Only by gathering data
from many similar situations will the model result and the average results from the real world be
certain to coincide (provided that all of the data are accurate). Whether an individual observation
(e.g., of 30-minute averaged odour concentrations) matches the model prediction for that single
event is strongly influenced by random factors. Any individual real-world result may differ from the
model by at least a factor of two due to turbulence-based variability (Stein and Wyngaard, 2001).
Hence, a factor of two around the model prediction is a fair indication of the range of values one
might expect when comparing the model result with an actual measurement. Furthermore, the
results of the model prediction are dependent on the observer’s location relative to the direct




1898a report final.doc                                                                    PAE Job Number 1898a