This practice note is one of a set developed to provide consistency and transparency of river system models being used within the Murray–Darling Basin. The notes cover modelling practices, such as naming conventions for folder structures, to model methods, such as for flow routing and residual inflow estimation, and have been developed through a collaboration between the MDBA and Basin States.
Produced in collaboration with:
This practice note, 'Routing Calibration', describes the general principles and a high-level method that should be adopted when undertaking the calibration of routing parameters during reach calibration.
Background
Upstream and downstream observed flows define a river reach. Various methods are available for matching the attenuation of flows between two gauges and the most appropriate method depends on the purpose of the model. In Source storage routing can be implemented using either a lookup table or a power function. For more background see Link storage routing - SRG.
Helpful information about deciding on which gauging sites should be used during routing calibration can be found here: Practice note: Selecting flow data.
Purpose
Routing parameters are used to simulate the timing and shape of the hydrograph at the downstream flow gauge. The purpose of this practice note is to outline the general principles that should be applied when calibrating routing parameters for a river reach.
Scope
This practice note covers:
- Tests of attenuation of flow and choice of parameterisation (lookup table or power function).
- Location of routing links within the river reach.
- Identification of an appropriate calibration period or calibration events.
- Relationship between routing and estimation of other reach calibration parameters.
- Criteria for acceptable flow routing parameters.
Associated Practice Notes and references
General Principles
- Routing parameters should be calibrated to achieve the best match of the timing and attenuation of flows at the downstream flow gauge.
- The physical characteristics of the reach, data availability, and model purpose should provide some guidance as to the most appropriate routing model and parameters.
- Reach lengths should be determined from catchment maps or from reported values.
- Tests for attenuation of flow should be used to determine if routing is needed (see below).
- Consideration should be given to the location of the routing links with respect to other nodes in the reach. In many cases, routing parameters will only be included on a single link and the location of this link may impact on water availability at other model nodes. It may also affect how inflows derived using a reach water balance calculation can be distributed within the reach. The sensitivity of the model results to location of the routing link should be considered.
- Appropriate events should be identified for calibrating routing parameters or determining the lookup table. These events should cover the entire flow range and where possible represent periods of minimal diversions and other phenomena that will influence the shape of the downstream hydrograph.
- Events should be identified that allow the modeller to determine the travel time between the upstream and downstream flow gauges.
- Where there will be individual routing reaches associated with different upstream inflows, then events should be identified which will best enable the different routing reaches to be parameterised.
- When there is a need to account for detention and loss of water, routing calibration may need to be iterative with the loss model.
- Generally, routing parameters should remain the same for the period of record, however, there may be cases in calibration where different routing is used in different periods, such as before and after a dam is built.
- Criteria for acceptable flow routing parameters should be established prior to commencement of calibration.
- Reach subdivision should be guided by stability criteria.
Suggested tests and approaches for setting parameters
Identification of an appropriate calibration period (or calibration events)
The modeller will be required to define a calibration period. They should try to:
- Identify events without major extractions or other phenomena that influence the shape of the downstream hydrograph.
- Consider a range of events of different sizes, but focus more closely on events in the flow range of most interest.
- Understand potential difficulties associated with matching the flows during a particular event at the downstream gauge. (e.g. flows greater than highest rating, major extractions between upstream and downstream gauge).
Tests of attenuation of flow
The modeller should undertake certain tests to decide if storage routing is required. They should:
- Undertake visual inspections of the rising and falling limbs of hydrographs. When the rising limbs of the upstream and downstream gauges are parallel then attenuation is not significant, and storage routing may not be required.
- Plot a time series graph of all the reach inflows (upstream gauge, tributaries, and, residual catchments) and the downstream gauge to be calibrated to determine an initial estimate for the lag in each link. When the shape of the hydrographs at the upstream and downstream gauges are similar, then lag routing may be appropriate.
Tests of linear or non-linear attenuation
If storage routing is required, a decision is required as to whether linear or non-linear routing should be used. The modeller should:
- Compare the percentage of downstream peak to upstream peak for a range of events (assuming losses and gains between the gauges have been accounted for). In a linear system the percentage should be similar.
- Calibrate assuming linear and assess the results to see if non-linear attenuation is required.
- Compare reach storage and outflow to see if the relationship is linear.
Examples of dummy hydrographs for visual inspection.
| No Routing Required | Lag | Attenuation |
 |  |  |
|---|
Selection of Muskingum routing (power function) or variable parameter Muskingum routing (lookup table)
More information on the routing is available in Source, see Link storage routing - SRG and Storage routing sections in the eWater Source documentation.
When selecting what type of storage routing should be used, the modeller should consider:
- Quality of flow data available.
- The time available for calibration. The modeller should minimise the time spent to achieve an acceptable outcome when calibrating routing links.
The presence of major overbank flow. Variable parameter Muskingum routing may be required when there are clearly different travel times for different flow rates.
Criteria for acceptable flow routing parameters
When assessing routing the modeller should consider the following:
- How well the rising limbs of the hydrographs match across the flow range at the downstream gauge.
- Match the shape and timing of the peak of the hydrographs at the downstream gauge.
- Match the shape of the recessions at the downstream gauge.
Steps for determining piecewise relationship for routing when using variable parameter Muskingum routing in Source
In many models within the MDB, it is likely that variable parameter Muskingum routing (lookup table) will be adopted. The method below outlines a method for determining the required parameters.
- The maximum index flow should be above the maximum recorded flow value.
- Index flows should be set based on the upstream rating curve. A maximum of 10 points should be selected. Points should be selected based on points of change in the rating curve.
- The first segment on the piecewise relationship should be flat, and the zero flow travel time should equal that of the first index flow.
- Observed data from the upstream gauge should be used to set the final two points of the piecewise relationship. In this high flow range the peaks can be identified and the travel time easily determined. The upper section of the piecewise relationship should also be flat (i.e. the two largest index flows should have the same value).
- Travel times for the remaining points can be set by either:
- Optimisation, or,
- Analysis of the travel times between the upstream and downstream gauge for peaks at the different index flows.
- Where routing needs to be split between reaches (e.g. due to intermediate flow gauges, location of tributary inflows or large diversions) travel time can be proportioned based on length of reaches.
- Inflow bias should initially be set to 1.
- Number of divisions should be set to 2 times the travel time to ensure stability (Link storage routing).
References
Close, A.F. (1996) A new daily model of flow and solute transport in the River Murray. Proc. 23rd Hydrology and Water Resource Symposium. Hobart, 21-24 May: 173-178. Institution of Engineers, Australia
Laurenson, E.M. (1959) Storage analysis and flood routing in long rivers. Journal of Geophysical Research, 64(12): 2423-2431, doi:10.1029/JZ064i012p02423.