COUPLING A NEAR FIELD OUTFALL MODEL WITH A FAR FIELD HYDRODYNAMIC MODEL

The fate of outfall discharges in coastal waters involves hydrodynamic phenomena that act over wide ranges of temporal and spatial scales. Near field mixing occurs within a few minutes and a few tens of meters after which the plume is either trapped by density stratification or reaches the water surface. It then drifts with the current and is diffused by oceanic turbulence in the far field by processes that occur on time scales of hours to days and length scales of hundreds of meters to kilometers. It is not feasible to encompass all these scales in one model, so separate models are needed for the near and far fields that must be coupled. Near field mixing depends on the wastewater flow rate, current speed and direction, and density stratification. These can vary rapidly, resulting in wide variations of plume rise height and dilution; they can be simulated by the model NRFIELD. Far field transport depends on large-scale circulation patterns driven by winds, tides, and local freshwater sources; these processes can be simulated by the hydrodynamic model Delft3D. In this paper we describe the coupling of NRFIELD and Delft3D to produce a state-of-the-art system for predicting the impacts of wastewater discharges in coastal waters. Three steps are involved. First, Delft3D-FLOW is run to generate time series of currents and density stratifications near the diffuser. These time series are then used as inputs to NRFIELD, producing a time series of plume behavior, especially its rise height, thickness, and dilution. This time series in turn becomes an input to the transport modules of Delft3D. Mass sources of, for example bacteria, are apportioned into the grid cells that overlap the plume with initial concentrations and volumes computed from the near field dilution. The far field advective transport is then predicted by a model that uses the currents from the FLOW module, and either the gradient-diffusion type water quality model WAQ, or the particle-tracking model PART. For bacteria, the particle tracking module is best as it does not suffer from high numerical diffusion that can considerably overestimate impacts at specific locations and decay models of bacterial mortality can be readily incorporated. In this paper we describe the procedure for coupling NRFIELD and Delft3D and give an example application to San Francisco, California.