AWRA-R version 5.0 Technical Report

[摘要] This report describes the AWRA-R v5.0 model and its application and benchmarking. The core objective of the AWRA-R model is to produce various fluxes and stores associated with river systems in regulated and unregulated systems to support the production of the national water accounts (NWA) by the BoM. The research and development of AWRA-R model started in FY2011-12 (Lerat et al., 2012) as part of WIRADA and since then four different versions of AWRA-R [versions 3.0 (Lerat et al., 2013), 4.0, 4.5 and 5.0] have been developed, tested and implemented so far. The AWRA-R model produces a large number of fluxes and stores for surface water accounts. This quantitative information provides a detailed understanding of major components of water balances at a reach level for water resources accounting, analyses and reporting purposes.The technical details of different components of AWRA-R v5.0 model and their governing equations are presented in section 2. The improvements/changes in AWRA-R over the period of the development since 2011 are also summarised in this section. A large amount of spatial and temporal data are required for AWRA-R modelling. The list of all input temporal variables, spatial and other parameters are listed in section 3 with the reference to appropriate governing equations presented in section 2. The model has eight calibration parameters, which are listed in section 3. The two calibration approaches (reach-by-reach and system calibration) used for model parameterisation are also described in this section. The system calibration approach has been newly developed and implemented in AWRA-R v5.0. The AWRA-R v5.0 model has been calibrated and validated in the MDB using both the reach-by-reach and system calibration approaches. Section 4 presents the collation of input data for the application of the model in the MDB. The AWRA-R v5.0 model in the MDB region included a total of 485 gauges and 33 large and medium size storages. The results of the model calibration and validation against the observed daily streamflow data are presented in section 5. A comprehensive benchmarking of AWRA-R v5.0 results was undertaken using available data from ground based and remotely sensed observed as well as the results of the earlier version of the model. The results of benchmarking are also presented in this section. A set of benchmarking scripts have been prepared to benchmarking AWRA-R results in both reach and regional scales for comprehensive understanding of the model performance and water balance. The results of the reach-by-reach calibration and validation of the model in the MDB show highly satisfactory performance of the model with median daily NSE of 0.60 and median annual bias of less than 1% for the period of calibration (1970-1991) and median daily NSE of 0.69 and median annual bias of 16% for validation period (1992-2014) for the MDB. Within the MDB region, median daily NSE varied between 0.54-0.83 in 18 different sub-regions under the calibration mode and the variation of median daily NSE was between 0.39-0.83 in the 18 sub-regions under the validation mode. The model performed best in the Murray catchment and worst in Loddon Avoca, where the quality of the observed stream data was poor with many missing data points. The mass balance error of the AWRA-R model was negligible.

[发布日期] 2015-10-07 [发布机构] CSIRO

[学科分类] 地球科学(综合)