JOurn.31 ofr in.nerr.3du.3a- Re--search
..oluin-e 5, issue 1 - Octoier 21:11:13
The Optimization of Aquifer Storage and Recovery Wetlse
Meeting the public's every increasing demand for potable water is becoming a major problem throughout the
United States. The water sources available are already under stress; however, use of Aquifer Storage Recovery
(ASR) wells may partially resolve the problem in as they can be used for the cyclic storage and recovery of
water. The function of these wells is to store the water in the subsurface when it is readily available from the
rainy season in order to have a sufficient supply of water when there is a shortage during the dry season. Right
now this technology is only being used for the storage of treated drinking water. It could be expanded to store
water that is not potable but suitable for other purposes such as irrigation. This water could be stored in
This project focuses on determining the relationship between the volume of water withdrawn from the ASR well
and the recovered water quality as expressed in chloride concentrations. Modeling this relationship will enable
system operators to predict the volume of recoverable water from an ASR well and in turn lead to more
efficient operation of ASR systems. The amount of usable water that can be withdrawn varies between wells with
the duration of storage. From analyzing data from current ASR wells a relationship between water quality and
the volume of water recovered was discovered. For each well we will now be able to know exactly when to
stop recovering water.
In the United States the public water supply demand is growing. This is a problem because the groundwater supply
is already under stress. In Southwest Florida they are considering placing caps on groundwater production in order
to limit the environmental impacts that the increased demand has caused. Examples of these adverse impacts
are saltwater intrusion and lower lake levels. In order to meet this higher demand new techniques are
being researched. One technique that is being tested is the use of Aquifer Storage Recovery technology.
ASR wells are used for the cyclic storage and recovery of water. Throughout the year, there are rainy seasons
and dry seasons. The function of these wells is to store the water in the subsurface when it is readily available
from the rainy season in order to have a sufficient supply of water when there is a shortage during the dry
season. Right now this technology is only being used for the storage of treated drinking water. It could be
expanded to store water that is not quite up to drinking standards but would be suitable for other purposes such
as irrigation. This water could be stored in brackish aquifers.
Theoretically the volume injected into these wells should be equal to the volume that can be recovered. In reality
this is not the case. As the volume of water drawn out increases, the quality of that water decreases. This
depletion of water quality is due to an increased level of chloride concentration. The quality of water extracted is
not the same as the quality of water injected because the water extracted is always a blend of the water injected
and ambient groundwater. The quality of water injected is higher than the quality of the ambient groundwater
so there comes a specific time during the recovery process when the fraction of groundwater is as high as it can
be tolerated before the water becomes impotable. The goal of this research is to determine that time.
In order to understand what happens to the water after it is injected into the well a simple model
was developed. As the water is injected it is assumed that it disperses into the shape of a cylinder
with radius C centered around the ASR well. The cylinder has a length L which is the aquifer
thickness. The radius of this cylinder can be calculated by:
Q1 Y 2
where Q1 is the injection flow rate (L3/T), to is the length of time of the injection period, and 0 is
the effective porosity of the aquifer.
This cylinder of water does not idly sit with its center at the well. It is influenced by the natural flow
of water in the aquifer. After a known storage time, ts, the center of the cylinder moves a distance B
from the well in the direction of that natural gradient. B can be calculated by:
where q is the specific discharge of the aquifer under natural gradient conditions. q is a function of
the hydraulic conductivity of the aquifer K [L/T], the hydraulic head 0, and the direction of
the groundwater flow, s [L]. This is calculated by:
Following the storage period is a recovery period. During this time, water is extracted through the
ASR well. A cylinder of length L also models the extracted water. This cylinder has a radius A which
can be calculated by:
A _Q__ - t, Y
where QR [L3/T] is the flow rate of the water being extracted, and tr is the time duration of the
Since the center of the cylinder of injected water moved down gradient during the storage time, part
of the volume of this cylinder contains the injected water and part of the volume contains the
ambient groundwater. These parts are proportional to the circumference of the cylinder of
extracted water. If the two cylinders, injected and extracted, are superimposed together, the fraction
of the extracted water that consists of ambient groundwater can be calculated by the fraction of
the circumference of the extraction cylinder that is outside the boundary of the injection cylinder.
The center angle containing this outside fraction of the circumference, WI, can be calculated by:
[1' A +C -B^ (B +C -A4
' -2- arecos --- + arccos ----
Now the fraction of ambient groundwater extracted over this recovery period can be determined.
This fraction at time tr will be called F:
For the above equations to be valid, C-B
be zero. In this case A would only encompass injected water, thus no ambient groundwater would
be extracted. A must be less than or equal to C+B because otherwise the recovery cylinder would
not intersect the injection cylinder. The model is not designed for this case.
To test this model, ASR well data was collected from five wells in the city of Cocoa in the years of
1994-2001. In order to evaluate the quality of the water, the parameter chosen to be measured was
the concentration of chloride. When the chloride concentration becomes too high, the water becomes
no longer potable. The data used was the extracted volume of water each day, the total injected
volume for the period, and the chloride concentration of the extracted water each day. The model
was applied to this data.
Next, the water quality was evaluated. Since the chloride concentration C of the recovered water
is monitored, and given that the chloride concentration of the injected water Ci and the
chloride concentration of the ambient groundwater Cb are both known, the actual fraction of
ambient groundwater recovered, Fdata, can be determined by:
To test the model, both F and Fdata were plotted versus the cumulative volume of water recovered.
Values of the ambient groundwater chloride concentration, effective porosity, and the
gradient displacement were calibrated to minimize the sum of the square difference between Fdata and F.
RESULTS AND DISCUSSION
As was expected the model closely follows the trend of the actual well data. Figure 1 shows the
results for well 1 in 1999 and Figure 2 shows the results for well 4 in 2001. The goal of the project was
to develop and test a simplistic model of the ASR system to be used to predict when extraction
should stop due to the quality of the water being withdrawn becoming too low for potable use. The
model can be used to predict this point; however, due to its simplicity the model has made
many assumptions that do not exactly follow actual conditions. It is a good starting point for
future research in this area.
Figure 1. Well 1 1999
0.4 - data
0.2 * model
-0.2 200000 400000 600000 80C000
0 0 0 0
Cumulative Volume (ft^3)
Figure 1. Well 1 1999.
Figure 2. Well 4 2001
0.4 -- + data
0.2 r 0*4 * model
-0.2 2E+06 4E+06 6E+06 8E+06 1E 07
Cumulative Volume (ft^3)
Figure 2. Well 4 2001.
In the future, improvements to the model can be made. The first involves the variable
hydraulic conductivity. Hydraulic conductivity has to do with the ease that the water moves through
the ground. It is not a constant variable throughout the natural ground as is suggested by the model.
Due to the simplicity of this model, an average conductivity was applied throughout the well.
Hydraulic conductivity is actually spatially variable. Moving down the well hydraulic conductivity
changes randomly meaning that some places water can move easily through the ground pore spaces
and other places it is more difficult for the water to move, unrelated to depth in the well. A
random hydraulic conductivity should be applied to different layers as depth is increased in the well
since the hydraulic conductivity will be higher in some places and lower in other places. This would
better model the true conditions in the well.
Another improvement that could be made to the model is associated with the injection phase. This
model assumes all the water is injected at one time and then it translates together as a cylinder
along the natural gradient of the aquifer. A better model of the actual conditions would be to inject
and translate the water in small increments. In the well the water does not wait around for the rest of
it to be injected so if smaller volumes of water injected over smaller time periods and each
then translated, the model should fit the data more closely.
I would like to thank Dr. Kirk Hatfield for all of his patience, help, and guidance.
Munis, Albert, P.E., Ziegler, William B., P.E. "Aquifer Storage and Recovery in Southeast
Florida." American Society of Civil Engineers Second International Symposium on Artificial Recharge
of Ground Water. Orlando, Florida. 17-22 July 1994, 311-317.
Pyne, R. David G. "Seasonal Storage of Reclaimed Water and Surface Water in Brackish Aquifers
using Aquifer Storage Recovery (ASR) Wells." American Society of Civil Engineers Second
International Symposium on Artificial Recharge of Ground Water. Orlando, Florida. 17-22 July 1994,
Vergara, Emilio D., Waller, Phillip L. "Aquifer Storage and Recovery Allows for Cost Effective
Development of a Regional Public Water Supply in a Water Short Area." American Society of
Civil Engineers Second International Symposium on Artificial Recharge of Ground Water. Orlando,
Florida. 17-22 July 1994, 299-310.
Ziegler, Gary J., P.E. " Artificial Recharge in New Jersey's Atlantic Coastal Plane." American Society
of Civil Engineers Second International Symposium on Artificial Recharge of Ground Water.
Orlando, Florida. 17-22 July 1994, 318-323.
Back to the Journal of Undergraduate Research
College of Liberal Arts and Sciences I University Scholars Program I University of
I UNIVERSITY of
'Pe Fourdon li:' :ort 're (flkr Natio
@ University of Florida, Gainesville, FL 32611; (352) 846-2032.