Parameterisation of a simple semi-distributed model for assessing the impact of land-use on hydrologic response

In this study, a simple semi-distributed model is combined with a landscape/climate regionalisation strategy with the aim of developing a predictive capacity regarding the range of effects caused by varying types and intensities of land-use at the regional-catchment scale. Using a case-study application within a 1260 km2 catchment in eastern Australia, it is demonstrated that even if distinct landscape types are randomly distributed within a (sub)catchment, it is possible to correctly parameterise their distinct hydrological responses via streamflow optimisation, provided the variability in hydrologic regime associated with variable climatic inputs is explicitly represented. Using a rigorous Monte Carlo style investigation of parameter uncertainty, it is demonstrated that the strength of the identified response ‘signatures’ is significantly improved when the information content of multiple ‘internal’ streamflow gauges is utilised. Along with the resultant improvement in streamflow predictability, the optimal landscape parameterisation is shown to be associated with response characteristics that are consistent with expected variations within the study area. An empirical connection between two regionalised model parameters and physical soil-depth data is also demonstrated. Such qualitative evidence provides improved confidence in extrapolating model predictions to investigate land-use change scenarios, and highlights the level of insight that can be gained into large-scale catchment behaviour when simple models are combined with ‘internal’ landscape and hydrologic response information.