Attachment 6 - Odour Impact Assessment - Katestone · 132005.638.06

    •    Geoscience Australia 9 second DEM terrain data

    •    TAPM default land cover data edited to be consistent with aerial imagery

    •    Default options selected for advanced meteorological inputs

    •    Data from the BoM AWS located at Tewantin assimilated over one vertical level with a radius of influence
         of 5km and a quality factor of 0.9.


A1.3 Comparison of TAPM output with observational data

A1.3.1 Overview

The model validation in this section compares observational data with data derived from running TAPM after
assimilating data from the BoM station at Tewantin. TAPM was originally run without data assimilation and the
predicted meteorological data was compared to observational data at the Tewantin station. The initial model
validation, prior to the assimilation of the BoM Tewantin data, showed that the model generally performed well at
simulating the meteorology across the region; however, the model was underpredicting calm winds (< 1 m/s) from
the south-southwest through to north-northwest as well as stronger winds from the south and northeast directions.
TAPM was also over-predicting the frequency of southeast direction.

The second iteration was run using the assimilation of the observational data at the BoM station at Tewantin in
order to improve the accurately of the model’s predictions to align more closely with the observational
meteorological data.


A1.3.2 Validation of TAPM modelling

The following section presents the results of the validation of the final TAPM run that was used for the dispersion
modelling assessment. The model validation compares the BoM Tewantin observational meteorological data with
data extracted from TAPM at the location of the Tewantin monitoring station.

Figure A4 shows probability density functions that graphically compare statistical distributions of meteorological
parameters between the TAPM output and BoM observational data. Table A1 presents statistical comparisons of
TAPM output (wind speed and temperature) to meteorological data recorded at the BoM monitoring station. The
TAPM output was extracted from the closest inner grid point to the location of the weather station.

The following statistical measures of model accuracy are presented in the tables:

    •    The mean bias, which is the mean model prediction minus the mean observed value. Values of the mean
         bias close to zero show good prediction accuracy.

    •    The root mean square error (RMSE), which is the standard deviation of the differences between predicted
         values and observed values. The RMSE is non-negative and values of the RMSE close to zero show
         good prediction accuracy. The RMSE is given by


                        1 N
         RMSE =            ( Pi − Oi ) 2
                        N i =1

         where N is the number of observations, Pi are the hourly model predictions and Oi are the hourly
         observations

    •    The index of agreement (IOA), which takes a value between 0 and 1, with 1 indicating perfect agreement
         between predictions and observations. The IOA is calculated following a method described in Willmott
         (1982), using the equation




 Katestone Environmental Australia Pty Ltd                                                    17 September 2026
 D25072-8 Noosa Council - Noosa Composting Odour Assessment - Final                                    Page 28