No abstract

No abstract

Results of a field campaign and numerical simulations are used to show how physical mechanisms impose length scales and timescales that determine the dominant biogeochemical process. As an example, the dynamics of the Snake River inflows into Brownlee Reservoir is investigated to explain the onset and maintenance of an oxygen-depleted region (the oxygen block) in the surface layer of the upstream part of the reservoir. The oxygen block was located in a region of the reservoir in which the surface layer was warmer as a result of smaller wind stresses and reduced evaporation rates. Numerical simulations reproduced the hydrodynamic field observations resulting from inflow, outflow, wind stress, and atmospheric heat fluxes. When the wind stress opposed the inflow, the surface layer was arrested, forming a zone of convergence, stagnating the fluid and allowing the biological oxygen demand in the water to deplete the dissolved oxygen (DO) in the surface water; direct measurements showed that vertical mixing was small and contributed only marginally to the oxygen depletion. Net DO production in the water column was consistent with the observed variation with the buoyant inflow pattern, that is, a sink during overflows and overcast days and a source during interflows and intense sunlight. These observations provided further evidence that the water in this region was biologically isolated as confirmed by a scaling analysis. Modern numerical hydrodynamic simulations have reached a level of accuracy where they may be used to identify and quantify ecological niches.

No abstract

A desalination plant is proposed to be the major water supply to the Olympic Dam Expansion Mining project. Located in the Upper Spencer Gulf, South Australia, the site was chosen due to the existence of strong currents and their likely advantages in terms of mixing and dilution of discharged return water. A high-resolution hydrodynamic model (Estuary, Lake and Coastal Ocean Model, ELCOM) was constructed and, through a rigorous review process, was shown to reproduce the intricate details of the Spencer Gulf dynamics, including those characterising the discharge site. Notwithstanding this, it was found that deploying typically adopted 'direct insertion' techniques to simulate the brine discharge within the hydrodynamic model was problematic. Specifically, it was found that in this study the direct insertion technique delivered highly conservative brine dilution predictions in and around the proposed site, and that these were grid and time-step dependent. To improve the predictive capability, a strategy to link validated computational fluid dynamics (CFD) predictions to hydrodynamic simulations was devised. In this strategy, environmental conditions from ELCOM were used to produce boundary conditions for execution of a suite of CFD simulations. In turn, the CFD simulations provided the brine dilutions and flow rates to be applied in ELCOM. In order to conserve mass in a system-wide sense, artificial salt sinks were introduced to the ELCOM model such that salt quantities were conserved. As a result of this process, ELCOM predictions were naturally very similar to CFD predictions near the diffuser, whilst at the same time they produced an area of influence (further afield) comparable to direct insertion methods. It was concluded that the linkage of the models, in comparison to direct insertion methods, constituted a more realistic and defensible alternative to predict the far-field dispersion of outfall discharges, particularly with regards to the estimation of brine dilution in the immediate vicinity of an outfall location.