Modelling Electrical Conductivity Variation in the Intermittent, High-Salinity Scotts Creek Catchment, South-Eastern Australia Using a Travel Time Distribution Approach

Focusing on water travel time as an important characteristic of water transport through a catchment provides an important underpinning for modelling of solute transport and water quality. This concentrates effort on a key catchment feature instead of trying to model all complexities. Past assessments of the travel time approach have focussed on relatively small perennial catchments and have not considered intermittent catchments. To address this gap, the Scotts Creek catchment in SW Victoria, Australia, was modelled. In this analysis, a 20-year, high-frequency electrical conductivity record was used to assess the performance of a travel time-based EC model. The EC data presents a complex pattern of response. Behaviour includes positive EC-discharge correlations early in the runoff season after flow commences and is still at very low discharge levels, while for the majority of the runoff season more typical dilution effects dominate. A time-variant StorAge Selection (SAS) function was applied within a modified version of the Tran-SAS. SAS function parameters were determined by selecting periods of time with consistent EC-discharge variations, reflecting probable changes in EC transport processes. EC concentration was modelled with an age-based concentration model. By varying the relationship between SAS function parameters and discharge under different flow conditions and including a high salinity source for part of the year, the simulation of complicated EC dynamics in Scotts Creek's catchment was improved, particularly for very low flow conditions. The effect of different patterns of interannual variability—either continuously dry or alternating dry and wet years—was explored, demonstrating that the hydroclimatological conditions in previous years affect EC response. The behaviour of simulated travel time distributions and residence time distributions for ten time points across the typical annual pattern of variation is also explored. The final outcomes present good consistency with the existing hydrogeological conceptual model of this catchment. An important learning is that EC appeared to behave non-conservatively in some parts of the year, and it is important to account for this when fitting the model.