An assessment of scale issues related to the configuration of the ACRU model for design flood estimation

A continuous simulation modelling approach to design flood estimation has many advantages and overcomes many limitations of commonly used design event approaches. A major concern with continuous simulation using a hydrological model is the scale at which modelling should take place. According to researchers, the level of representation that will preserve the physical chain of hydrological processes, both in terms of scale of representation and level of description of the physical parameters for the modelling process, is a critical question which must be addressed. Objectives of this research were to determine the optimum levels of catchment discretization and soil and land cover information and to assess the optimum use of daily rainfall stations for the configuration of the Agricultural Catchments Research Unit (ACRU) agrohydrological model when used for design flood estimation. Results obtained for selected quaternary catchments in the Thukela catchment and Lions River catchment indicated that modelling at the level of hydrological response units (HRUs), using area-weighted soils information and more than one driver rainfall station where possible, produced the most realistic streamflow volume results when compared with observed streamflows. Design flood estimates from simulated peak flows did not compare well with observed data.