﻿ Pumping Test: Theory and Analysis Methods > Pumping Test Analysis Methods - Flexible Assumptions > Multi-Layer-Aquifer-Analysis

# Multi-Layer-Aquifer Analysis

The Multi-Layer solution can be applied to a layered aquifer system where aquifers are separated by aquitards, and bounded above and below by user-defined boundary conditions (see below). This method is useful when you have multiple aquifers, with pumping from just one of the aquifers, and observation wells are screened in various aquifers.

A conceptual illustration of one potential multi-layer aquifer system is shown below:

The solution implemented in AquiferTest uses the technique as described in Hemker and Maas (1987). For further information on this method, and prior to applying this solution to your analysis, you are encouraged to read the papers (see References section below).

In a multi-layer aquifer configuration, this solution can be used to estimate:

T and S of the pumped aquifer and unpumped aquifer(s)

S: Storage coefficient of the aquitard(s)

c (hydraulic resistance) of the aquitard. From this parameter, Kv can be calculated using also the thickness of the aquitard.(See page 24 in Kruseman and de Ridder [1991] for details)

The solution may also be applied for a variety of other aquifer and well conditions as described in the Hemker and Maas (1987) paper; however, for the purpose of using within AquiferTest, the focus is on the stacked/layered aquifer conditions.

An important requirement for this method is to define your conceptual model. This is done through the settings for the MultiLayer solution.

The settings to configure the multi-aquifer layer type and order can be accessed by clicking on the "Conceptual..." button as shown below.

The settings for the Multi-Layer configuration is shown below.

Define the number of aquifers you wish to analyze at the top of the window (must be at least one aquifer)

Define the conceptual model by specifying the appropriate layer type, where the topmost layer in this table corresponds to the upper most layers in your conceptual model. The Multi-Layer solution has requirements relating to the layer order and types:

The topmost layer must be one of the following:

Aquiclude: Impermeable materials, no flow conditions

Aquitard w/o storage bounded top s=0: Aquitard with no storage, and bounded above by a reservoir (constant head) boundary condition which yields no drawdown

Aquitard bounded top s=0: Aquitard with storage, and bounded above by a reservoir (constant head) boundary condition which yields no drawdown

Aquitard bounded top impervious: Aquitard which is bounded above by a no-flow boundary condition (impervious materials)

The intermediate layers may be one of the following:

Aquifer: an unpumped aquifer

Aquifer (pumped): this is the aquifer that is pumped

Aquitard without storage

Aquitard with storage

Note that all aquifers must be separated by aquitards.

The bottommost layer must be one of the following:

Aquiclude: Impermeable materials, no flow conditions

Aquitard w/o storage bounded bottom s=0: Aquitard with no storage, and bounded below by a reservoir (constant head) boundary condition which yields no drawdown

Aquitard bounded bottom s=0: Aquitard with storage, and bounded below by a reservoir (constant head) boundary condition which yields no drawdown

Aquitard bounded bottom impervious: Aquitard which is bounded below by a no-flow boundary condition (impervious materials)

After you configure each layer, it is suggested that you provide some reasonable default parameter values for each aquifer and aquitard.

Once you apply these settings and close the window, you will see some default type curves. One type curve will appear for each aquifer. You may then use the existing tools within AquiferTest for adjusting the fit to the data set.

Note that due to the number of parameters required for this solution (which increases with each additional layer), the automatic fit may not succeed in all cases.  You are advised to do a manual curve fit (using the parameter controls)  and also lock some parameters once you are confident with the estimated values.

A fundamental difference in the Multi-Layer solution lies in the way the aquifer parameters are estimated. Solutions like Theis, Neuman, etc. provide estimates for the parameters at each well; whereas the Multi-Layer analysis takes into account the best fit of all wells (specified in each aquifer) in order to determine a single set of parameters for each aquifer and aquitard; thus there is no need to average values from per well afterward as you would have to do with other solutions.  For this reason, on the Result Panel you will see estimated parameter values for each aquifer or aquitard; the results are presented in a top-down fashion, where the parameters correspond to the order in which the layers are defined in the Settings window.

## Example

An example of Multi-Layer Aquifer analysis is below.

An example project is available at:
C:\Users\Public\Documents\AquiferTest Pro\Examples\Multi-Layer-Aquifer-System.HYT

## Assumptions and Limitations

Important assumptions for this method include:

Within this system, water flows horizontally in the aquifers, which are separated by an aquitard

All layers are of infinite horizontal extent within the area of influence of the pumping test

Aquifer layers have homogeneous and isotropic transmissivity and storativity values

Aquitard layers have homogeneous vertical resistance and storativity values, and yield vertical flow only

Only saturated groundwater flow is considered

The top and base of the system have either no-drawdown or no-flow boundaries

Well screens fully penetrate the aquifer layer

Observation wells are screened in only one aquifer layer

Only drawdown (or build-up) as a result of pumping is considered

Darcy’s law is valid, except for turbulent flow near the well screen

Water from storage is discharged instantaneously with decline of head

Unsaturated zone flow does not influence drawdown

The change in water levels does not affect the transmissivity or storativity of any of the aquifers or aquitards

Seepage faces at the water-table well can be safely ignored

Horizontal-deformation process effects are negligible

The discharge rate in any screened layer of any pumping well is not affected by the drawdown caused by any other pumping well.

b < 1.0 (i.e. the radial distance from the well to the piezometers should be small

The implementation of this solution within AquiferTest has the additional limitations:

Single Pumping Well

Pumping well is fully screened over only one aquifer layer

Within AquiferTest, the following assumptions can be relaxed using superposition, in order to accommodate:

barrier/recharge boundary, and

variable discharge rate