Potential of using WATCH forcing data to model a low land river basin of the upper Murray-Darling basin in Australia

Abstract:

Scattered station based climate data is often not sufficient to describe the dynamics of the catchment processes and efficiently manage the water resources. Therefore, a lot of focus has been to identify alternative distributed data sources, such as; remotely sensed data or global re-analysis data. Hence, this study uses the Water and Global Change (WATCH) forcing data, based on 40 years ECMWF Re-Analysis (ERA-40), to model a semi-arid low land flood plain river basin in a data sparse region. The semi-distributed Soil Water Assessment Tool (SWAT) was used to model the river basin (Warrego, 52140.6 square km) located in the upper Murray-Darling basin in Eastern Australia. Multi station model calibration was achieved using the Sequential Uncertainty Fitting -2 (SUFI-2) algorithm with the Nash Sutcliffe Efficiency (NSE) as the goal function against monthly observed flow data. Modelling of a low land river system is highly challenging, due to topographic heterogeneity, nonlinear climatic behavior and sparse observed flow data with extended periods of zero flows. Preliminary simulation results indicate a NSE of 0.26 to 0.86 for the calibration period and 0.04 to 0.47 for the validation period. Furthermore, the volume fraction explained by the model ranged from 0.69 to 2.71 in the validation period. While the unsatisfactory results may be attributed to the SWAT modelling framework, which struggles with modelling flow in flat flood plains, the study does reveal the potential to use remotely sensed data in low land river basins with little or no climate data.