Use of a PEST calibrated Source Catchments model and regional water quality data to improve the predictive capacity of constituent load generation in the Torrens River catchment

Abstract

Catchment models are useful decision support tools which simulate impacts of changes in a range of inputs on water quantity and quality issues. These issues are important to catchment managers for prioritising remediation on sub-catchments or landuses which contribute pollutants to receiving waters. This paper describes use of the eWater "Source Catchments" model to simulate water flows, sediment and nutrient loads in the Torrens River catchments of the Mount Lofty Ranges Watershed in South Australia. Rainfall and landuse are used as primary factors leading to run-off and pollutant transport. The Torrens River is a fairly complex system and includes regulated and unregulated catchments, reservoir in-flows, transfers between reservoirs and inter-basin transfer from the Murray River. The hydrology of the Source Catchments model was calibrated with the model-independent parameter estimation program (PEST) to best simulate this complexity. A range of data analysis approaches were also used to create surface and base flow constituent generation values for suspended solids, total nitrogen and total phosphorus for 12 landuse types. These values were generated from primary landuse as well as composite sampler load data from 16 catchments in the Adelaide and Mt Lofty Ranges region. Analysis of this local data; which spans more than 20 years at some locations; has generated locally relevant constituent data for major landuses in the Mt Lofty Ranges Watershed. This task was made easier through the use of eWater's "Water Quality Analyser" software tool. Better characterisation of both hydrology and constituent generation has improved the predictive capacity of the Source Catchments model from previous versions, which were manually calibrated and used non-local constituent data. The predictive capacity total nitrogen and total phosphorus has significantly improved although there was still considerable uncertainty around the suspended solids loads. The improved model will be used to target management actions and simulate impacts of strategic policy initiatives to reduce the pollutant loads to drinking water reservoirs in the Mount Lofty Ranges.