Navigation:  Demonstration Exercises and Benchmark Tests >

Exercise 13: Multi-Layer Aquifer

 

NOTE: The multi-layer aquifer analysis pumping test solution is available in the AquiferTest Pro edition.

This tutorial provides an example of a multi-layer aquifer analysis based on a data set from a numerical model generated by Visual MODFLOW Flex. For more general information on this solution, please refer to Multi-Layer-Aquifer-Analysis section.

This tutorial also assumes that you are familiar with the basics of navigating the AquiferTest interface, and as such, the steps have been abbreviated to focus just on the steps/inputs that are required for a Multi-Layered Aquifer analysis.

A theoretical pumping test was performed in a multi-layer leaky confined aquifer with the conceptual model shown as below.

The pumping well is screened across the lower aquifer, and was pumped for a constant rate of 10 L/s, for 30 days. The observations were taken in the pumped aquifer and middle aquifer, located at 20 m away from the pumping well. Visual MODFLOW was used to generate a time-drawdown data set at the observation points, which were loaded into AquiferTest in order to verify the Multi-Layer analysis.

A sample project has already been created with the well locations, discharge, and time-water level data set.  You will need to open and start with this project for this exercise.

If you have not already done so, double-click the AquiferTest icon to start AquiferTest

When you launch AquiferTest, the AquiferTest welcome page will load, with the option to create a new project or open an existing project.

Click the Open other project... button, and browse to the file:

"C:\Users\Public\Documents\AquiferTest Pro\Exercises\Multi-Layer-Aquifer.HYT"

 

The project will load with the Pumping Test tab selected.

Click on the Water levels tab; you will see 3 wells listed; there are no data defined for the pumping well.

If you select OW-20-pumped-aq, you will see the time-water level data from the observation point that is located in the pumped aquifer;

Click on OW-20-unpumped-aq and you will see the time-water level data from the observation point that is located in the unpumped (upper) aquifer. Note that a data filter is applied to both data sets, which will keep only 15 data points per logarithmic time scale.

Click on the Analysis tab; by default, you should see a Time-Drawdown plot displaying the data sets from both observation points.

Click on Analysis/Create a New Analysis from the main menu .

From the Analysis Method panel on the right side, choose Multilayer. You should then see a Multi-Layer Settings appear as shown below.

 
In this example, you will analyze a two aquifer system, so it is not necessary to change the first setting "Number of Aquifers".

Next you will define the layer types.  You will start with the first (topmost) layer, which is set to Aquiclude by default.

Left-click twice on the cell in the very top left; you should see the value change to "Aquitard bounded top s=0"

Next, left click three times on the cell in the bottom left (the last, bottommost layer); the cell value should become "Aquitard bounded bottom impervious"

If assigned correctly, your window should now appear as shown below.

 
Next you will assign each observation well to the appropriate aquifer. This is done in the "Wells" column (last column on the right)

Locate the second row in the table, which corresponds to the "Aquifer" layer (this is the upper aquifer, which is not pumped). In the wells column, left-click twice and a dropdown arrow should appear on the right side.

Click on this dropdown arrow, and choose OW-20-unpumped-aq from this list (as shown below). Then immediately click on the cell below this row (note: this is required in order to register the checkbox selection).
 

 

Next locate the fourth row in the table, which is the "Aquifer (pumped)". In the Wells column, left-click twice and a dropdown arrow should appear.

Click on the dropdown arrow and choose OW-20-pumped-aq from the list (as shown below), then immediately click on the cell below this row. (note: this is required in order to register the checkbox selection).
 

 

Next you will define some default starting parameters based on your knowledge of the aquifer and aquitard conditions.

Enter the following values for each layer type:

For Aquitards, you need to define start S and c (hydraulic resistance values)

For Aquifers, you need to define start T and S values

Define the values as per the table below.

 

Layer

T [m2/d]

S

c [s]

Aquitard
bounded top s=0

 

1E-4

1E7

Aquifer
 

1E2

1E-2

 

Aquitard
 

 

1E-7

5E5

Aquifer (pumped)
 

2E2

1E-4

 

Aquitard
bounded bottom impervious

 

1E-7

1E7

 

Once complete, your settings window should appear as shown below.
 

 

Click OK to close the settings window and the values you defined will be applied and you should see some default type curves which correspond to the values you defined above.  

Set the plot to dimensionless units and your plot should be similar to the following:


 

The upper type curve corresponds to estimated time-drawdown for the upper (unpumped aquifer), whereas the lower type curve corresponds to the lower (pumped) aquifer

You may now adjust the fit to the type curves in order to more closely match the data set; if you choose to do an automatic fit, you may need to adjust the solver tolerance; this is due to the number of parameters that must be adjusted and the complexity of the Multilayer solution.

Locate the "Fit" button above the analysis graph, click on the dropdown arrow to expand the button, and select Fit Settings as shown below.

 

A Fit settings window will appear.

For the Maximum Number of Iterations, set this to 5000.

 

 

At this point, you may apply the automatic fit (by clicking on the Fit button), or further guide/control the solution by defining parameter ranges (lower and upper bounds)

For each parameter, you can use the scroll bars to set a reasonable lower and upper threshold which will be utilized during the automatic fitting routine. Some approximate recommended values are below (the parameters are displayed in the order of the layer they correspond to as defined in the Settings window above; the column "Layer" below provides some assistance for correlation)
 

Parameter

Layer

Lower Limit

Upper Limit

Hydr. resistance

Upper Aquitard

1E3

1E9

Storage coefficient

Upper Aquitard

1E-6

1E-2

Storage coefficient

Aquifer (unpumped)

1E-5

1E-2

Transmissivity

Aquifer (unpumped)

1E-2

 

1E3

Hydr. resistance

Middle Aquitard

1E4

1E8

Storage coefficient

Middle Aquitard

1E-7

1E-4

Storage coeff.

Aquifer (pumped)

1E-6

1E-3

Transmissivity

Aquifer (pumped)

1E-2

1E4

Hydr. resistance

Lower Aquitard

1E2

1E12

Storage coefficient

Lower Aquitard

1E-6

1E-1

 

Once complete, your window may look similar to what is shown below (note, the exact upper/lower bound values are not important, so long as you are in the same order of magnitude).

 

 

Then click Fit to register the changes. The program should then apply the automatic fit, and your Analysis graph should appear as shown below.

 

 

You are recommended to use your best judgment and intuition when reviewing the set of parameters that are estimated by the program.

 

Using this solution, you may find that two or more different conceptual models (Layer configurations) will fit the data equally.

An alternate approach would be to use the parameter controls in order to adjust the parameters one-by-one; you can then lock a parameter value, and apply an automatic fit, then repeat this in an iterative fashion.

 
The parameter controls window is shown below; in the screen image shown here, the window has been re-sized such that there are two "columns" of parameters, so that these can more easily be correlated to the layers in the Multi-Layer configuration.

 

 

Through adjusting the various parameters, you can see what impact this has on drawdown in the unpumped or pumped aquifer(s), and also see how this impacts early or late time-drawdown stages.

 

This concludes the Multi-Layer Aquifer Analysis Exercise. The next exercise introduces the Binkhorst and Robbins solution for slug tests in wells with partially submerged screens. You have the choice of exiting AquiferTest or continuing on to the next exercise.

 

References

Hemker C.J. and C. Maas, 1987.  Unsteady Flow to Wells in Layered and Fissured Aquifer Systems, Journal of Hydrology, vol. 90, pp. 231-249.

Hemker, C.J., 1999. Transient well flow in layered aquifer systems: the uniform well-face drawdown solution. Journal of Hydrology, 225: 19-44.

Carlson F., Randall, J., 2012.   MLU: A Windows Application for the Analysis of Aquifer Tests and the Design of Well Fields in Layered Systems. Groundwater Software Spotlight.

 

 


Page url:https://www.waterloohydrogeologic.com/help/aquifertest/index.html?example_13_multi-layer_aquifer.htm