Another step on the difficult journey to regionalise a probability-based loss model

© CURRAN-CONFERENCE. All rights reserved. There is an industry need for guidance on loss model parameters to inform their application in ungauged catchments for design flood loss estimation. However, to date there has been only a limited number of studies which seek to understand how loss model parameter values are related to catchment characteristics in order to provide such guidance. Additionally, due to data restrictions when modelling losses across entire catchments, commonly used loss models (such as initial loss – continuing loss (IL/CL)) are typically quite simplistic and have only a limited physical basis. For this study, the relationship between parameter values of the loss model SWMOD and outputs from the Australian Water Resources Assessment – Landscape (AWRA-L) was investigated using 35 rural Australian catchments. SWMOD is a probability distributed model that represents the variability in soil moisture holding capacity within a catchment as a cumulative probability distribution. This structure gives SWMOD parameters a greater phyical basis than the parameters of IL/CL and allows modelled losses to reduce over the course of an event, which is consistent with the mechanism of saturation excess overland flow. Estimating the SWMOD parameter representing catchment soil moisture holding capacity directly from the range of the AWRA-L soil moisture time series was found to reproduce calibrated model performance in 25 of the 26 study catchments where calibrated values of this parameter were less than 300mm (which represents the upper range of soil moisture variability in AWRA-L). However, no consistent relationship between AWRA-L outputs and model performance or the SWMOD initial moisture parameter could be identified. The practicalities of using SWMOD for design flood loss estimation in ungauged catchments remain limited due to i) an inability to determine catchments where model performance will be acceptable a priori and, ii) an incomplete understanding of how the range and distribution of SWMOD parameters related to catchment attributes. Nevertheless, the strong relationship between AWRA-L outputs and the SWMOD soil moisture holding capacity parameter in the majority of the catchments considered indicates that this remains a promising avenue of future research.