Supervising Scientist Report 136
Tong GD & Williamson B
Supervising Scientist, 1998
ISBN 0 642 24339 5
- SSR136 - Modelling of the hydrodynamics and chemistry of Macquarie Harbour, western Tasmania - main report (PDF - 389 KB)
- Appendix Part 1 - Figs 2.1 to 3.46a (PDF - 3,920 KB)
- Appendix Part 2 - Figs 4.6a to 5.39 (PDF - 3,210 KB)
This report documents the work carried out under Task 14 (Modelling of the hydrodynamics and chemistry of Macquarie Harbour, Western Tasmania) as part of the joint Federal/Tasmanian 'Mount Lyell Remediation Research and Demonstration Program' (MLRRDP). The report describes all aspects of the work-strategy carried out to achieve the study objective of developing a hydrodynamic and chemical model capable of estimating the effect on water quality in Macquarie Harbour as a result of changes in pollutant levels, particularly copper, entering the harbour from the King River.
Macquarie Harbour is a large roughly rectangular harbour (30 km x 8 km x up to 50 m deep) on Tasmania's rugged World Heritage West Coast. It is connected to the Southern Ocean through a relatively narrow and shallow entrance channel. Being in the midst of the 'roaring forties' and with two major west coast rivers (King River and Gordon River) discharging to it, the harbour exhibits complex three-dimensional hydrodynamics under tidal, wind and freshwater forcing.
Additional to this natural system has been the discharge of copper mine tailings into the Queen River and down the King River to Macquarie Harbour at the rate of 1 million tonnes per year (copper load 2000 kg/day) for the last century.
Herein lies the backdrop to this study which sets the study objective of a working computational model. The objective has been achieved and the major tasks constituting the modelling approach are briefly reviewed here.
- Review of the available physical and chemical data.
- Configuration, verification and use of a hydrodynamic model, C3, capable of simulating the 3D hydrodynamics of Macquarie Harbour.
- Conceptualisation of the chemical speciation model and its implementation through the equilibrium model MINTEQA2.
- Embedding of the chemical model within the hydrodynamic modelling.
- Use of the integrated physical/chemical model to further understand harbour processes making full use of the available DELM data set.
- Use of the model in predicting the effects of treated loadings.
1 - Review of data
The study commenced with a review of all available data to gain an understanding of the harbour in terms of its geometry, hydrodynamics and chemistry. Tidal, wind and freshwater forcing were also reviewed and much of the available data were plotted and presented in a compendium for easy access.
The chemistry data collected over DELM's 4½ month intensive monitoring period (early December 1994 to mid April 1995) was particularly valuable as a set of test data and this period was chosen as the study's test period.
2 - Hydrodynamic modelling
The proposed model C3 from Seaconsult Marine Research Ltd, Vancouver, was configured to Macquarie Harbour with a 400 m x 400 m horizontal grid resolution and a vertical resolution (from water surface) of 6x2 m + 1x3 m + 7x5 m for the deepest part of the harbour.
Taking advantage of the mode-split nature of the model and using a time-step of 2 minutes (adopted throughout the study) preliminary runs were undertaken to establish configuration of the model to the harbour. The hydrodynamic configuration process addressed freshwater plume propagation, harbour water level response, entrance channel behaviour and development of the typical steep gradient of the observed halocline structure. A detailed account of this configuration process was given in the required Interim Report MH-1-9/95.
Once configured, the model was used to assist in understanding the complexities of the hydrodynamics of the harbour. Pre-empting the modelling of the chemistry, the concentration of copper was first used as a scalar tracer (no chemical reactions taking place) to investigate the hydrodynamic response of the model (in terms of likely copper plume migration) for two prevailing wind conditions under a common tidal forcing. All hydrodynamic tests are fully described in section 3.
3 - Chemical modelling
Following a review of the chemistry of Macquarie Harbour, the US EPA model MINTEQA2 was used to simulate the two processes considered to be the most important in the copper speciation process within the harbour-adsorption onto hydrous metal oxides and complexation with dissolved organic matter (DOC). Details of the application of MINTEQA2 are given in section 4.
The adsorption process was shown to be dominant and the settling out of the river-borne colloidal iron oxides was modelled by the introduction of a settling velocity to the particulate iron oxides formed by the flocculation of the colloids as they move into the higher ionic strength of the mixing zone. For the King River sediments discharging to Macquarie Harbour, this was taken from the literature to be 10-4 m/s for fine flocculating sediments.
4 - The hydrodynamic and chemical model interface
MINTEQA2 is an equilibrium model and the proposed modelling strategy was to embed MINTEQA2 within the hydrodynamic model for use, in a pseudo-transient manner, by continually updating it with a changing chemical field. The validity of this strategy depends on the equilibrium speeds of the MINTEQA2 chemistry and the hydrodynamic transport of the four quantities involved in the adsorption and complexation processes (concentrations of Cu, Fe, DOC and salinity for pH).
The approach to embedding MINTEQA2 within C3 was via the pre-computation of a set of 'look-up' tables to be interrogated for % adsorbed and % complexed total Cu and defined over the expected range of the 4-parameter space for Macquarie Harbour. Interrogation of the tables was efficiently achieved by two bi-linear interpolations, one in the Cu-pH plane and the other in the Fe-DOC plane. Toxic copper (free copper ions) was also uniquely available from the MINTEQA2 speciation.
5 - Model proving tests
The model was run over the 4½ month intensive monitoring period with the appropriate hydrodynamic forcing and the measured King River chemical loadings. Results were compared with the measured harbour data and presented in a range of graphical formats (section 5).
Two sets of plots are of particular importance since they compare model results with measured results.
The time-series plots over the full period at 6 DELM monitoring stations in the northern harbour for top, mid and bottom of the water column (figs 5.1 to 5.6). These plots show good agreement with the weekly copper measurements and indicate both the dominance of the adsorption/fall-out mechanism and its validity in the northern harbour.
The profile plots at 11 DELM monitoring stations in the southern harbour (figs 5.28 to 5.38) which show the conservative copper model as applicable on and south of the Sophia Pt- Liberty Pt transect.
6 - Model predictions
The model has been used to predict the effect of selected reduced copper loadings and commensurate reduction in loadings of the hydrous metal oxide adsorbing surfaces (Fe, Al and Mn) provided as two-point data by project management.
For the representative electrowinning solvent extraction process with a pH of 5.5, these tests were for reduced loadings designated Stock 2 (80% treated) and Stock 3 (99% treated). Time-series of loadings were derived by assuming flow dependency and linearly interpolating between the two given data points corresponding to Burbury Power Station being either off or on. (For Stock 2, these were scenarios 8 and 11 respectively whilst for Stock 3, they were scenarios 9 and 12 respectively as designated by project management.)
Results have been consistently compared with the untreated case running over the adopted 4½ month test period of DELM's monitoring of the King River loadings and weekly profiling measurements of the harbour's responses. The comparisons show the expected marked decrease in water column copper concentrations as a time-series over the test period.
In the longer term and perhaps more interestingly, since adsorption with fall-out of colloidal sediments is dominant, is the comparative cumulative deposition plot at the end of the report which shows the decrease of copper to sediments for the treated cases to be consistent with the percentage reductions in loadings. This result reinforces the finding that most of the copper deposits out with the settling of the sediments that form the King River delta.
7 - Concluding remarks
The modelling procedure has proved better than expectations held at the commencement of the study. This positive assessment is largely based on the time-series plots (figs 5.1 to 5.6) and the profile plots (figs 5.28 to 5.38) that constitute a comprehensive comparison of chemical processes modelled, untreated/treated scenarios and the DELM field measurements taken in Macquarie Harbour over the 4½ month test period.
No doubt the study could benefit from refinements such as further calibration of the hydrodynamics from moored current meter data and refined grid modelling of the northern harbour, particularly around the King River delta, to better resolve the deposition process. An addition to the model expected to yield improved results would be the inclusion of the uptake of copper from the sediments throughout the harbour.
Nonetheless the study has yielded some interesting results as follows.
- King River plumes disperse quickly with the majority of the copper adsorbing to hydrous metal oxides which undergo colloidal sedimentation and fall-out close to the mouth forming the delta region. (Residual water column copper being well-modelled by the adsorption, deposition and complexation process.)
- Toxic conditions (indicatively 3x10-9 M) were reached throughout the southern harbour over the 4½ month test period and were well-modelled by the conservative case.
- Typical proposed treatment scenarios for acid mine drainage will achieve the expected beneficial results (commensurate with the level of treatment) on the water quality in Macquarie Harbour with much less copper depositing and transporting within the water column.
- The proposed treatment scenarios will achieve reduced concentration levels in the region around Yellow Bluff which are generally below the indicative level for potentially 'toxic' conditions (3x10-9 M) for typical tidal conditions and moderate to strong south-westerlies and north-easterlies. This result is of interest in regard to the region as a potential fish-farming site.