The pesticide diuron from urban and agricultural runoff is detected at concentrations above current water quality guidelines in the Great Barrier Reef (GBR) marine environment. We quantified the load of the pesticide diuron entering GBR waters using the GBR-Dynamic SedNet catchment model. After comparison of simulated distributions with observations at 11 monitoring sites we determined a half-life of diuron in GBR marine waters of 40 days.
The 1 km resolution eReefs marine model (GBR1: version GBR1_H2p0_Cq3pe1_Dhnd) was nested inside the 4 km eReefs marine model (GBR4: version GBR4_H2p0) to achieve stable and reliable boundary forcing. Diuron only enters the marine model from the catchments. Diuron at the ocean boundaries is set to 0 and if diuron leaves the 1km model boundary it does not return. The simulation period is from January 2016 to July 2018.
The GBR1 hydrodynamic model was forced by wind, atmospheric pressure gradients, surface heat fluxes and evaporation / precipitation fluxes, tides, and waves and calculates temperature, salinity, elevation and offshore currents, integrating on a 1.2 second barotropic time step. The hydrodynamic model calculates the mean volume flux through model cell faces over a 1-hour period (Herzfeld et al., 2016) that are used in a transport model to disperse diuron. The transport scheme was a flux-form semi-Lagrangian advection scheme that applies the fluxes in steps determined by the Lipschitz stability criterion, with a maximum of 1 hour time step. Each model cell contains their own value of diuron based on the hydrodynamics, and so diuron fluctuates hourly in each cell.
The eReefs marine model represents freshwater input from 16 rivers located in the GBR. For rivers with freshwater input, GBR1 uses a river boundary condition where freshwater was discharged in a brackish surface plume. This boundary condition provides realistic plume dynamics and avoids excessive offshore transport caused by overestimating the density difference between freshwater plumes and seawater (Herzfeld, 2015, Baird et al., 2017).
For the remaining minor rivers, diuron enters GBR coastal waters as loads as a surface flux (Figure 1B). Rivers containing high diuron loads that enter as surfaces fluxes were the Murray, Mossman, Proserpine and Burrum and Plane (very high risk). Some river catchments in GBR-SedNet do not contain diuron loads, (Normanby, Fitzroy, Calliope, Boyne, Styx, Shoalwater and Waterpark) as diuron was not one of the pesticides used or generally found within these catchment areas during frequent monitoring of all pesticides.
Individual river plume dispersal was quantified using conservative tracers. A river tracer was specified as a unit concentration in each river flow (e.g., 1 kg m-3). Therefore, a model grid cell with 0.20 concentration of river water will be composed of 20 % river water, and 80 % water that comes either from another river/s or the ocean. Each river has a unique tracer that was individually advected and diffused using a conservation flux-form scheme based on hourly-averaged 3D velocity fields. Similarly, the diuron tracer has a time-varying river concentration in μg L-1 specified by the SedNet catchment model and was advected and diffused in a similar manner to the river tracers.
