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 c apacity regarding the range of effects caused by varying types and intensit ies of land-use at the regional-catchment scale. Using a case-study applica tion within a 1260 km(2) catchment in eastern Australia, it is demonstrated that even if distinct landscape types are randomly distributed within a (s ub)catchment. it is possible to correctly parameterise their distinct hydro logical responses via streamflow optimisation, provided the variability in hydrologic regime associated with variable climatic inputs is explicitly re presented. Using a rigorous Monte Carlo style investigation of parameter un certainty, it is demonstrated that the strength of the identified response 'signatures' is significantly improved when the information content of mult iple 'internal' streamflow gauges is utilised. Along with the resultant imp rovement in streamflow predictability, the optimal landscape parameterisati on is shown to be associated with response characteristics that are consist ent 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 i n extrapolating model predictions to investigate landuse change scenarios, and highlights the level of insight that can be gained into large-scale cat chment behaviour when simple models are combined with 'internal' landscape and hydrologic response information. (C) 2001 Elsevier Science B.V. All rig hts reserved.