Graduation Year


Document Type




Degree Name

MS in Civil Engineering (M.S.C.E.)

Degree Granting Department

Civil and Environmental Engineering

Major Professor

Jeffrey J. Cunningham, Ph.D.

Committee Member

Matthew Pasek, Ph.D.

Committee Member

Mark Ross, Ph.D.


ASR, Groundwater Quality, Tracer Test, Dispersivity, Push-Pull Test


To use Aquifer Storage and Recovery (ASR) as a water treatment and storage method, we must understand the fate and behavior of potential contaminants in the system. Knowing more about the aquifer properties such as dispersivity will enable us to assess how the target chemicals behave in the system. Characterizing the aquifer by using a tracer is a good way to estimate dispersivity since the tracer is not adsorbed or degraded. This study aims to build a lab-scale physical model of a confined aquifer to simulate an ASR system, verify its hydraulic functionality, and characterize the hydraulics of the aquifer by estimating the apparent dispersivity during simulated ASR conditions.

In order to represent the real aquifer conditions as much as possible, I used a rectangular water tank, a mixture of 2 different types of sand for the permeable layers, impermeable kaolin clay as confining unit, and plastic tube attached to bubbler stones as injection/extraction wells. I installed eight of my wells at the edges as boundary wells to reduce the wall effect of the tank and to mimic a real aquifer of large areal extent.

After building my system, I conducted several injection and extraction tests using only groundwater to verify that the system acts like a real aquifer and there is no clogging inside of the tubes. As a result of these tests, I proved that the hydraulic conductivity of my confined aquifer is high enough to let water flow, and it is possible to inject or extract water in desired volumes. Fortunately, the boundary wells successfully allowed water to enter or leave the system.

Then, I performed two tracer tests by injecting groundwater solutions with bromide into the aquifer in order to quantify the recovery of bromide at the end of the tests and to estimate the dispersivity. In Test 1, I introduced 4 liters of groundwater with 77.6 mg/L bromide into the system and set the resting time for 4 days before proceeding to the extraction phase. In Test 2, 4 liters of groundwater with 100 mg/L bromide was injected and I waited for 15 days before extracting and sampling.

Ion chromatography was used to quantify the bromide levels in the samples, and an empirical equation was employed to estimate dispersion coefficients for both tests.

Results indicated that the recovery of bromide tracer mass for 2 different waiting periods is different than each other. I recovered 80% of bromide for Test 1 whose storage time was 4 days, but only 63% for Test 2 that had a 15 days resting period. It was an expected outcome since the longer the bromide resides in the system, the more it undergoes diffusion and dilution.

Modeling the extracted concentrations of bromide with a previously published equation suggested apparent longitudinal dispersivity of 0.05 m for Test 1 and 0.2 m for Test 2. Even though these values are not very far from each other, I was not expecting to find different estimations for apparent dispersivity. This conclusion indicates that the formula I used may not be the best prediction method for dispersivity because it does not account for the bromide storage time.