Introduction
The Groundwater Interaction module in Source is a collection of modules and protocols/guidelines that simulate the exchange of groundwater and salt between rivers and the underlying groundwater systems.It determines the exchange flux of water between a river and the underlying aquifer for each link of Source at each time-step. The estimated flux accounts for interactions between groundwater and surface water along the entire length of the link. The direction of flux can either be from the river to aquifer or vice versa, that is, the river either loses water to the groundwater system or it gains water from the groundwater system.
| Info | ||
|---|---|---|
| ||
| Note: The Groundwater interaction module was formerly known as the Groundwater Surface water interaction tool (GSWIT). |
The various components of the exchange flux may be calculated within Source or they can be imported as a time series from field monitoring. The exchange fluxes may include the following components:
- Continuous natural exchange between the river and the aquifer, which is driven by the head difference;
- Flux out due to evapotranspiration;
- Flux out due to groundwater pumping;
- Flux in due to irrigation and diffuse recharge; and/or
- Exchange fluxes during (within bank and overbank) flood events.
The process for selecting the appropriate methodology to estimate GW-SW exchange flux is mainly dictated by the type of connection between a river and the underlying aquifer, and is influenced by data availability. A decision-making flow chart (shown in Figure 1) is a guide in selecting the appropriate methodology for any situation.
Figure 1. GW-SW decision-making flowchart
Configuring groundwater in Source
Exchange fluxes are assigned to a Source link once storage routing has been enabled on the link (refer to Types of links routing).
To configure groundwater, open the link's feature editor. Figure 2 shows the contextual menu for the Groundwater item. In this example, no groundwater model has been selected.
Figure 2. Storage link routing, Groundwater
Types of groundwater models
There are four types of groundwater models which can be configured in Source; these are described next.
| Info | ||
|---|---|---|
| ||
| Note: When modelling groundwater, a positive flux represents a discharge out of the river while a negative flux represents a recharge into the river. |
Flux input
This represents the daily exchange flux between the reach and its aquifer, which can be specified as a single value, a time series (using Data Sources), or via an expression using the Function Editor (Figure 3). Note that flux values can be both positive and negative, so the reach alternates between gaining and losing.
Figure 3. Groundwater, Flux Input
Heads input
One method for estimating the groundwater flux is based on knowledge of the head difference between the river stage and the water table in the underlying aquifer. The Heads Time Series input model (Figure 4) uses the head difference between the current calculated groundwater head and the interpolated (from rating curve) river stage height to calculate groundwater flux. It requires:
- A conductance value (either set or calculated prior to run time, or calculated dynamically at run time); and
- A time series of groundwater heads covering the period of simulation (which can be specified as a single value, via a file using Data Sources, or via an expression using the Function Editor).
Figure 4. Groundwater, Heads Time Series Input
Unsaturated connection
The Unsaturated Connection model (Figure 5) uses the head difference between the stream and the aquifer, together with a conductance value to calculate the groundwater flux. This model is used when there is a permanent unsaturated connection between stream and the adjacent aquifer, that is, the stream must always be a losing one.
There are three main steps to configuring the model:
- Enter a conductance value (similar to Heads Time Series Input) ; and
- Set the groundwater level as a long term average groundwater level, or via a simple forecast model. For an unsaturated connection model, the groundwater level must always be less than the river water level.
| Info | ||
|---|---|---|
| ||
| Note: When using multiple rating curves, it is important that each curve has a unique start date to ensure that the model matches the current time-step with the relevant rating curve. The model gives a warning if a rating curve has the same start date. |
Under unsaturated conditions, the groundwater head value can not be larger than the current stream height. Groundwater head is always below the stream bed level under unsaturated conditions.
Figure 5. Groundwater, unsaturated
Saturated Connection
In a Saturated Connection model (shown in Figure 6), you may select any of the eight distinct groundwater and floodplain processes to run simultaneously. At each time-step the model evaluates the impacts of each selected process and uses assumed linearity and superposition to aggregate the impacts and calculate a flux. It is the most complex of all the groundwater models.
This model contains eight processes that can be enabled or disabled using the contextual menus shown in Figure 6. More detail is provided about them in the next section.
| Info | ||
|---|---|---|
| ||
| Note: Removing a process may result in the loss of any information entered for that process. |
The parameters for the saturated connection model fall into one of two categories:
- Shared ones (labelled Common Controls in Figure 6). These parameters are only set once per link and are shared by all of the individual processes in the model. Depending on the process that is activated, some parameters become inactive. For the example shown in Figure 6, the Floodplain Width parameter is only active parameter because it is relevant to the enabled (Evapotranspiration) process; and
- Unshared parameters only pertain to a single process.
Figure 6. Groundwater, saturated
Bank Storage
There are two ways to generate a bank storage:
- Calculation method - this calculates the bank storage flux at run time from shared aquifer parameters and stream flow. At each time-step, the model takes the current stage height for the stream, the previous stage height for the stream, the floodplain width, aquifer transmissivity and aquifer diffusivity, and calculates a bank storage flux; or
- Load a known time series of bank storage fluxes. At each time-step, the model will then retrieve a bank storage flux value and add it to the total groundwater flux generated for other saturated connection model processes under flooding conditions.
Figure 7. Groundwater, saturated (bank storage)
Diffuse recharge
Diffuse recharge (Figure 8) is applied across the whole of the floodplain simultaneously. It infiltrates through the surface soil and recharges the aquifer, and if sufficient head difference is achieved, causes the aquifer to discharge water to the stream. The process model requires three shared aquifer parameters (Transmissivity, Specific yield and Floodplain width) and four process-specific unshared parameters:
- Name - a unique text string representing the pump name or number;
- Start time - the day on which the diffuse recharge commenced;
- End time - the last day of recharge; and
- Recharge rate.
Figure 8. Groundwater, saturated (diffuse recharge)
Evapotranspiration
The Evapotranspiration process model (Figure 9) calculates the amount of water which is lost from the aquifer by evapotranspiration (ie. plant transpiration and/or bare soil evaporation). The model requires Floodplain Width as a shared aquifer parameter and three process specific unshared parameters:
- Extinction Depth - the depth to water table at which ET no longer happens;
- Average Depth of Adjacent Groundwater Table - the average depth of the water table; and
- Gradient - the hydraulic gradient of the aquifer as a percentage.
Groundwater evapotranspiration is assumed to decrease with increasing depth to water table and the extinction depth is the depth at which groundwater evapotranspiration attains a value of zero.
Figure 9. Link (Groundwater, saturated, ET)
Low flow
A river may continuously lose water to, or gain water from the nearby aquifer; neutral cases are also a possibility (no head gradient with zero exchange). Consider a whole river-aquifer system that is at a steady-state (ie. recharge into the entire aquifer is equal to discharge from the aquifer to the river system), thus this exchange flux (gain or loss) would remain constant with time. It is given by:
Δ hxC
M
Where, Dh is the head difference, C is the hydraulic conductance of the river-aquifer interconnection (m2/day) and M is the thickness of river bed sediments (m).
Figure 10 shows the parameters that must be configured for low flow.
Figure 10. Groundwater, saturated (Low flow)
Overbank
Overbank (Figure 11) occurs when the stream stage height exceeds the height of the stream bank. Water flows over the stream bank and out onto the adjacent floodplain filling up any depressions and channels and creating inundated floodplain areas. The water in the inundated floodplains either evaporates or infiltrates through the soil and finds its way back to the stream. At each time-step the model determines:
- Whether an overbank event has occurred;
- How much water is in any of the inundated floodplain areas;
- How much of this water is lost to evaporation; and
- How much water is returned to the stream (as a groundwater flux).
Figure 11. Groundwater, saturated (Overbank)
For each inundated floodplain area that you wish to model, an additional six parameters are required :(outlined in Table 1).
Table 1. Saturated connection, Overbank parameters
| Parameter | Description |
|---|---|
| Bed thickness |
| The surface soil depth |
| of the area |
| . |
| Hydraulic conductivity |
| The hydraulic conductivity of the surface soil in the area |
| . |
| Orthogonal distance |
| The orthogonal distance |
| from the area centroid to the stream |
| . |
| Overbank length |
| Threshold |
| A (unique) height (in mAHD) that river stage height must exceed to increase the area’s water volume |
| . |
| YDistance |
| The distance from the area centroid to a line perpendicular to the stream through the upstream node (not currently used by the model but included for completeness). |
Additionally, you must specify the river's cross section using a table of level, volume, area relationships for the area such that, in each successive row of the table, the level is greater than that in the previous row and all level values are greater than the threshold. Refer to Cross Section Editor.
Irrigation recharge
Irrigation recharge (Figure 12) works in a similar way as an unconfined aquifer except that, when recharge occurs, the cone of depression is inverted to form a groundwater mound, recharging the aquifer. If a sufficient groundwater head is developed the recharge water returns to the stream via aquifer discharge.
Since it is modelled in the same manner as pumping from an unconfined aquifer, they share similar parameters (with the exception that Pumping Rate is called Recharge Rate) and the controls for entering these parameters are almost identical. As with pumping, you can add irrigation recharge locations either manually or from a file. In reality, irrigation recharge is a distributed source, but the Source software treats it as a point source.
Figure 12. Groundwater, saturated (irrigation recharge)
Pumping
Groundwater pumping is one of the most important processes that impacts the exchange flux between groundwater and surface water. Pumping-induced river depletion is defined as the reduction of river flow due to induced infiltration of stream water into the aquifer or the capture of aquifer discharge to the river. As the cone of depression progresses towards a nearby river, groundwater discharge to the river gradually reduces and when the cone of depression reached the river, groundwater discharge to the river ceases. River water eventually start infiltrating the aquifer contributing to groundwater pumping and the rate of groundwater level drop decreases. Surface water may even start to infiltrate into the aquifer thus marking the start of river depletion.
After a long period of pumping, the cone of depression takes its final shape (ie. a steady-state is reached), and a portion, or in some cases, all of the pumping will be balanced by a reduction in, or reversal of flow, from the aquifer to the river. The proportion of pumping met by river depletion in the steady-state case will depend on various factors, including the proximity of the bore to the river compared to the distance between the bore and other stresses to the groundwater system (eg recharge, ET).
Source allows for two types of groundwater pumping:
- Pumping from an unconfined aquifer (Figure 13) with or without a no-flow boundary - This caters for pumps which tap directly into an aquifer (with or without a no-flow boundary) which is hydraulically connected to the stream. The pump is at a given distance from the stream, it switches on at some time, pumps at a constant rate for a period (the pumping period may occur prior to the simulation start time and the effects are still modelled) and then switches off; and
- Pumping from a semi-confined aquifer (Figure 1014) - used to model pumps which tap into a semi-confined aquifer which is hydraulically connected to the stream. Again, the pump switches on at some time, pumps at a constant rate and switches off.
You could enter the parameters for each process, or import a .csv file containing this information.
Parameter definitions (Figures 9 13 and 11):
- Name - a unique text string representing the pump name or number;
- Start time - the day on which the pumping commenced;
- End time - last day of the pumping;
- Rate - Rate of pumping;
- Orthogonal distance - the orthogonal distance (m) from the area centroid to the stream;
- Y distance - the distance from the area centroid to a line perpendicular to the stream through the upstream node (not currently used by the model but included for completeness);
- River to no-flow boundary distance - Distance away from the river where effects will no longer be observed at the river.
Figure 13. Groundwater, saturated (Unconfined pumping)
Parameters definitions (Figure 1014):
- Name - a unique text string representing the pump name or number;
- K* - non-dimensional variables used in the Hunt (2003) equation;
- λ - streambed resistance coefficient;
- ε - non-dimensional variables used in the Hunt (2003) equation;
- a - the orthogonal distance between the river and the pump;
- Start time - the day on which pumping commenced;
- End time - the last day of pumping;
- Pumping Rate - Rate of pumping;
- Orthogonal distance - the orthogonal distance (m) from the area centroid to the stream;
- K' - aquitard permeability;
- b - nominal stream width;
- σ - aquitard porosity;
- B' - aquitard saturated thickness;
- B'' - aquitard thickness beneath stream.