Author(s): Y. Q. Zhang; N. R. Viney

Linked Author(s):

Keywords: Catchments; Runoff; Rainfall-runoff modelling; Regionalisation; Stratification; SIMHYD

Abstract: Predicting runoff in ungauged catchments is one of the most important tasks in surface hydrology. This study comprehensively evaluates runoff prediction results in ungauged catchments by using rainfall-runoff modelling and regionalisation in 685 unregulated catchments widely distributed across Australia. First, a rainfall-runoff model, SIMHYD, is calibrated against recorded monthly runoff in each catchment. Then, model parameter sets optimised at the spatially nearest donor catchments are used for a target ungauged catchment for regionalisation (or model cross-validation). Last, the regionalisation results are evaluated in all the 685 catchments and its subsets obtained by stratification using catchment attributes. The results show that using multiple donor catchments can improve runoff estimation accuracy, compared to the single donor catchment. The model stratification results show that the catchment attributes including catchment area, precipitation and aridity index are suitable for stratification of model simulation results, and the model performs much worse for dry and large catchments than for wet and small catchments. These results suggest that it is very useful to stratify the large dataset of catchments for comprehensively evaluating regionalisation results.

DOI:

Year: 2011