A rainfall runoff model is used to derive runoff for a particular area from inputs of rainfall and potential evapotranspiration (or areal potential evapotranspiration). All rainfall runoff models in Source are conceptual models that represent catchment hydrological response to rainfall as a series of mathematical relationships. They provide runoff output from each functional unit as total discharge, which is split into quick flow (surface flow) and slow flow (baseflow) proportions. Refer to the Source Scientific Reference Guide for more detail.

Choosing the right model

Choosing and calibrating your rainfall runoff model is essential. The quality of your rainfall runoff calibration can interact with, and affect output from other models. For example, where constituent load is calculated as the product of flow and concentration, any errors in flow estimates will be propagated through to load estimates.

Consider the following when choosing a rainfall runoff model:

• What models do you and other people in your organisation have experience with;
• Have any rainfall runoff models worked well for your catchment in the past;
• What type of catchment are you modelling? (urban, forested, mixed land use);
• What data do you have available? You need to parameterise any models you choose; and
• Do you have any information on previous rainfall runoff model calibrations? If calibrated models already exist, you can re-use the models and parameter sets.

The following might also be useful:

• Run a range of different models and determine which ones work best; and
• Consider using different models or parameters, in different sub-catchments or functional units. What works best in one functional unit may not work well in another. For example, different models may be useful where sub-catchments/FUs have different soil types (different soil water holding capacity) or different proportions of urban areas (large differences in impervious runoff).

Configuring rainfall runoff models

The total discharge generated from rainfall runoff depends on which model is specified for the sub-catchment/FU combination. Configuring rainfall runoff models in Source involves 3 steps: assigning the model, adding input data to it, and parameterising the data (optional). These are described next.

Assign rainfall runoff models

In Source, a functional unit in a sub-catchment can be assigned a rainfall runoff model by choosing Edit > Rainfall-Runoff > Assign Models.... (Figure 133) shows the resulting window.

Choose the appropriate model using the drop-down menu in the Rainfall runoff column. There is an option to assign a rainfall runoff model, using either a table a map to select sub-catchments, then assigning a model to the FUs. These methods are similar to those outlined for assigning FU areas.

If runoff is not of interest, leave the entries in the Rainfall runoff column set to Nil runoff. Figure 133 shows some sub-catchment/FU combinations assigned a model, whereas others have been left with Nil runoff.

Assign inputs to rainfall runoff models

To assign inputs (PET and rainfall) to the selected rainfall runoff model, choose Edit > Rainfall-Runoff > Assign Inputs...(Figure 134).

The Grid-Based Input Assignment method allows you to load a time series of input data for each FU and sub-catchment combination in the model. You can assign inputs using a table or map. To assign input data using a table, do the following:

• ln the PET or Rainfall columns, select a sub-catchment or FU cell for which you have climate data;
• From the drop-down menu in the cell, choose Load;
• Navigate to the input file and click Open; and
• Choose the input file, which now appears on the drop-down menu for all cells in that column.

If required, refer to Using the Apply-to options to assign data to multiple sub-catchments and FUs.

The Climate Data Import tool allows you to upload the input data using the import tool. Refer to Using the Climate Data Import Tool.

Parameterise rainfall runoff models

Each model has a different set of input data requirements, which is explained later. You can parameterise rainfall runoff models if desired by choosing Edit > Rainfall-Runoff > Parameters....

A parameter set consists of a set of variables that you can assign values to. You can accept the defaults or change these values to create a new parameter set, which can be applied to each of the rainfall runoff models that you have previously selected. If the same model needs a different set of parameters because for example, some parameters are spatially dependent, simply create another parameter set with values appropriate for a different geographic region:

• Click a cell in the Model column. This specifies the model you wish to create a parameter set for.
• Choose New... from the Parameter set drop-down menu (Figure 135), This opens a window showing the default values for the parameter set for that model.
• To create a new parameter set, enter the desired values into the relevant parameters (Figure 136). Give the new parameter set a descriptive name.

Click OK and use the Table or Map tabs to assign the values from that parameter set to any or all of the functional units that have the relevant model assigned to them.

Blue shading indicates cells where models within FUs have had parameter values applied. Figure 137 shows a parameter set applied to all FUs of type "Urban".

Note that once you apply a parameter set to the models in a FU or sub-catchment, the models’ parameter values are fixed. Any subsequent changes to the parameter set values do not change the values applied to the models. If you have to change the parameters in a parameter set, you must re-apply the changed parameter set with the Using the Apply-to options.

Types of rainfall runoff models

There are 10 rainfall runoff models available in Source. The next section provides an overview of each of them.

AWBM

The Australian Water Balance Model (AWBM) generates daily total streamflow, quickflow and slowflow. It contains five stores - three surface stores to simulate partial areas of runoff, a baseflow store and a surface runoff routing store. The model also requires a baseflow index and 2 recession constants. Overall, the model has seven parameters and is one of the easier models to calibrate in Source.

This model requires rainfall and evapotranspiration (ET) data as input, the formats of which are provided in Table 103. Note that input data format for rainfall, ET and potential ET (PET) are the same This data must be continuous and overlapping.

GR4J

IHACRES Classic

Nil runoff

This is the default model in Source, which generates no flow (no quick or slow flows).

Observed runoff

The Observed Runoff model allows you to include an observed time series for rainfall runoff. It uses a digital filter to separate the quick (surface flow) and slow flow (baseflow) from the observed input data This model allows the inclusion of an observed time series for rainfall runoff and uses a digital filter to separate the quick and slow flow. This is done using two user defined parameters (k and C), which Source uses to identify the base flow component of flow and when subtracted from the total flow also gives the quick flow proportion.

This model requires daily flow data for input.

Observed surface runoff

The Observed Surface Runoff model is used to input an observed runoff sequence of quick flow in place of runoff generated by a model (ie, there is no slow flow component). It generates daily total stream flow and quick flow, and sets slow flow to 0. This model can be used in situations where there is only surface flow affecting the sub-catchment such as surface flow from a sewage treatment plant, or where only part of a catchment is being modelled. The observed runoff is used to represent inflow from areas that are not being explicitly modelled.

Note that if a base flow component is required, the Observed Runoff runoff model should be used.

For input data, a time series of observed surface flow over the FU per time step is required. Additionally, the time-series data must contain continuous daily values.

Sacramento

This is a catchment water balance model that was originally developed for the Sacramento River in the USA. It contains five soil water stores and performs flow routing via a unit hydrograph and a channel water store. In Source, the model has 16 parameters, but it is possible to have a larger number of parameters using higher increments in the unit hydrograph.

The Sacramento model has previously been embedded in the IQQM river system modelling framework and has been widely used for water resources planning in Australia, particularly in Queensland and New South Wales. If you are modelling a catchment in either of these states, its likely that the state government authorities have previously established a Sacramento model and you can obtain its parameters.

Note that due to a large number of parameters, there are lots of flows between internal stores within the Sacramento model, making it more difficult to calibrate.

The model is a continuous rainfall runoff model used to generate daily stream flow from rainfall and evaporation records using daily data. Input requirements include daily rainfall and PET data (Table 103).

SIMHYD

This model contains a rainfall interception and soil water store and estimates daily stream flow from daily rainfall and areal PET data. An additional 7 input parameters are required, including base flow coefficient, impervious threshold and recharge coefficient. SIMHYD has been widely used in Australia and was applied for generating runoff for the Murray Darling Basin Sustainable Yields study in 2006-2008.

Just as with the Sacramento and AWBM models, inputs required are daily rainfall and PET data (Table 103).

SIMHYD with routing

SMARG

SURM

SURM is a simplified version of the SIMHYD rainfall runoff model and caters for separate runoff generation processes in the impervious and pervious areas of the catchment. The model contains stores for interception loss, soil moisture and groundwater stores and an additional 8 parameters. SURM is the default rainfall runoff model used in music.

Inputs required are daily rainfall and PET data(Table 103).

Where: date is the observation date (dd/mm/yyyy)
rain is the observed rainfall or evapotranspiration or potential evapotranspiration in millimetres per day for the associated date.