The Effect of Stream Network Simplification and Gauge Error on River Model Performance

Abstract

Recent development of a Simplified Murray-Darling Basin Model along with an auto-calibration capacity has advanced the ability to rapidly conceptualise and parameterise river system models. However in any process of conceptualisation, some simplification may be warranted or desirable. In some cases the exact nature of stream connection in an anastomosing river network is difficult to determine. To more thoroughly assess the effect of simplification or error in stream network conceptualisation a range of river network conceptual models were assessed, ranging from as complex as the gauging network permits, to a simple single reach. Model performance is assessed by agreement with observed stream gauge data at various points for separate calibration and evaluation periods. The Lachlan catchment in central NSW was used as the test basin for this study. Results show that simplification can result in similar goodness of fit to less simple conceptual models. Care must be taken to avoid inclusion of gauges that increase model error in some downstream gauges. Errors in gauge data are propagated to downstream reaches, but the nature of the errors will determine how far the error persists as calibration moves further downstream. Flow based errors (eg. gauge rating error) and high frequency errors tend to become less detectable at more distant downstream gauges.

Key Words:

• Auto-calibration
• error propagation
• river
• river losses
• routing
• gauge error
• simplification

Additional information

Notes on contributors

J D Hughes+

Dr Justin Hughes is currently a Research Scientist at CSIRO working on water resource projects in river systems modelling, irrigation, and non-stationarity in runoff prediction. He has a background in soil management and catchment hydrology. He has hydrological research experience in a number of contexts ranging from, agricultural land-use and deep drainage, salinity, hydrological processes and runoff generation to bauxite mining and rehabilitation in southwest Western Australia. He was awarded a PhD from Charles Sturt University in 2009 for his thesis “Streamflow generation and salt mobilisation processes in the Central Murray-Darling Basin”.

D Dutta

Dr Dushmanta Dutta is a Senior Research Scientist in CSIRO’s Land and Water Division in Canberra. He holds a PhD (in Hydrology) from the University of Tokyo, Japan, and a MEng degree (in Water Resources Engineering) from the Asian Institute of Technology, Thailand. Dushmanta has 20 years of research experience in the fields of hydrology, river system modelling, hydrodynamic modelling, flood risk analysis, and water resources engineering and management. He is the author of more than 160 research publications, including 4 books, 40 referred journal articles and 90 conference papers.

S S H Kim

Shaun Kim is a Research Scientist and has been working in CSIRO Land and Water since 2010. Shaun holds a MAppSc (Coastal Management) and BSc (Marine Science) from University of Sydney. Since his studies, Shaun has worked in Research and Development in the field of hydrodynamic and biogeochemical modelling. While working at CSIRO, Shaun has helped develop the hydrological modelling tool, eWater Source, and has worked in developing river system models for the Murray-Darling Basin and North Queensland. Recently, Shaun has been involved in model development and calibration of the Australian Water Resources Assessment Modelling System in collaboration with the Bureau of Meteorology.

J Vaze

Dr Jai Vaze is a Principal Research Scientist in CSIRO Land and Water in Canberra, where he leads the catchment hydrology team within the surface water hydrology research program and the surface water assessments stream within the water for a healthy country fl agship. Jai has more than 20 years’ experience in the fi eld of water resources management, inland landscape modelling, distributed hydrologic modelling, predictions in ungauged basins, hydroclimate modelling and spatial modelling. He has published more than 100 scientific papers and reports. Prior to joining CSIRO in 2009, Jai managed the catchment and climate modelling group with the New South Wales Government. He has worked on various projects that contribute to key national initiatives, including the CSIRO Sustainable Yields projects, Australian Water Resource Assessment project, the eWater CRC’s River Systems project, and the South Eastern Australia Climate Initiative

G M Podger

Geoff Podger leads the Integrated Basin Modelling Stream within the Integrated Water Resources Management Theme of CSIRO’s Water for a Healthy Country Flagship. He has over 30 years’ experience in hydraulic and hydrologic modelling. He is leading teams of scientists researching integrated water resource management in south Asia. There are currently four international projects that Geoff overseas in south Asia. They include capacity building in Bangladesh, and a number of trans-boundary projects, including piloting eWater’s Source modelling system in three major river basins including the Koshi River Basin which runs through China, Nepal and India, the Brahmani Baitarni River Basin in eastern India, and the Indus River Basin in India and Pakistan. Geoff has been involved in the implementation and development of numerous hydraulic and hydrologic computer models and is the principal author of the Integrated Quantity and Quality Model. He has also been involved in the implementation of river models across New South Wales, and assisted in implementations throughout Queensland. He has also been involved in the implementation of the Mekong River Basin model from Chiang Saen, Thailand, to the Mekong Delta, Vietnam, and the Kunduz River Basin model in Afghanistan. Geoff has played a key role in the development of several catchment modelling toolkit products including rainfall runoff library, forest cover fl ow change tool and E2 (catchment model). As research leader for source development within the eWater CRC, he was involved in the development of the source river modelling platform. Geoff has developed a river modelling framework that connects over 70 surface and groundwater models that describe the water resources of the Murray-Darling Basin. These models were used in a landmark study to assess the water availability of the Murray-Darling Basin. Geoff led the River Modelling team for the Murray-Darling Basin Sustainable Yields project. The study considers risks to the shared resource that include climate change, plantation forestry development, groundwater development and farm dam development. This framework has been adopted by Murray-Darling Basin Authority and used to develop the Murray-Darling Basin Plan. As a Stream Leader he has overseen the implementation of Source river systems models for the entire Murray-Darling Basin, and the Flinders and Gilbert catchments in North Queensland. He has also been managing projects on uncertainty, risk assessment and management in river system models.