Estimating Sediment and Nutrient Loads in Gippsland Lakes Catchments Using E2 Modelling Framework

The deterioration of water quality in Gippsland Lakes, indicated by severe algal blooms, and influenced by the flow of nutrients from the Lakes catchments, has received increasing attention. For priority setting on appropriate management actions, a clear understanding is needed of spatial and temporal dynamics of sediment and nutrient fluxes in the catchments of Gippsland Lakes. A number of studies have been performed over recent years to investigate the sources and amounts of nutrient and sediment load entering the Gippsland Lakes (e.g. Grayson et al., 2001; Grayson and Argent, 2002). Following the establishment of various targets for load reductions to the Gippsland Lakes, attention has now turned to assessment of amelioration and load reduction options. The work reported here is a part of a larger undertaking of DPI Victoria to quantify the impacts of selected 'Best Management Practices' on water quality in the Gippsland Lakes catchments, which uses Bayesian techniques to investigate the impact of grazing and fertiliser management on loads generated from irrigated and non-irrigated farmland. This work builds upon an investigation using the CSS modelling system (Grayson and Argent, 2002) on flow and loads into Gippsland Lakes, and aimed to provide the capability required to support the catchment-scale assessment of the impacts of farm-scale management actions. To achieve this, the model structure and code base were updated using a new modelling system and two new component models, created for the Gippsland Lakes application, were developed. Flow and constituent loads (Total Suspended Solids (TSS), Total Phosphorous (TP) and Total Nitrogen (TN)) were simulated for the catchments of Gippsland Lakes using the E2 software, from the Catchment Modelling Toolkit (www.toolkit.net.au). E2 is a flexible catchment modelling software that supports construction of a range of models of different complexity using an approach built around selection and/or creation of particular component models. In this case, the Gippsland Lakes E2 model runs on a monthly time-step using twenty years of data covering high, medium and low flow years. Stream discharge and water quality loads are simulated using historic climate data, land use, and management information. The sources of sediment and nutrient loads are identified. E2 provides the capability to have different component models for flow, constituent generation and filtering for each functional unit (FU). The FUs within each sub-catchment are identified according to land use, and monthly flow and constituent loads are calculated for each FU. The loads from FUs within each subcatchment are lumped to form sub-catchment loads, which are then transferred downstream through a node-link structure. Following Grayson and Argent (2002), model outputs are available for major catchments (Latrobe, Thomson/Macalister, Avon, Mitchell, Nicholson and Tambo), for Western catchments, for Eastern catchments, and for the ?whole of Gippsland Lakes?. As part of model testing, evaluation of results for specific sub-catchments as well as ?whole of catchment? were undertaken. These results for major catchments of the Gippsland Lakes were compared with those of the CSS modelling system. Compared to the CSS model, E2 was generally found to be at least as robust or better. Long-term average annual loads were well predicted but there were many inconsistencies for months. Overall, the modelling results provide confidence that the model does capture the basic temporal and spatial variability of the system well. Point source data were not included in the E2 modelling reported here, as point source loads have changed significantly over recent years. Apart from this difference, the E2 system used similar data, timesteps and complexity and so has a similar level of certainty to the results of Grayson and Argent (2002). Their caution therefore continues to hold, i.e. the model should (therefore) be used for making relative assessments of the impact of various management actions to reduce long-term average nutrient and suspended loads.