Date of Award


Degree Name

Master of Science


Geological and Environmental Sciences

First Advisor

Dr. Duane R. Hampton

Second Advisor

Dr. Mohamed Sultan

Third Advisor

Dr. Alan Kehew


Slug tests, unconfined aquifers, hydraulic conductivity (K) analysis, water level displacements, four-way analysis of variance (ANOVA) examination

Access Setting

Masters Thesis-Open Access


This research presents a hydraulic conductivity (K) analysis of unconfined aquifers using slug tests. Slug tests are used to determine in situ aquifer hydraulic conductivity more quickly and economically than by a pump test. This study examines how to best conduct a slug test using a physical slug. Different common slug test analysis methods are compared, including Bouwer and Rice (1976), Hvorslev (1951), Dagan (1978) and Kansas Geological Survey (KGS, 1994). Questions that motivated this study include: Which methods are better for performing and analyzing slug tests? Does the size of the physical slug affect the results? Do large initial water level displacements produce better results than smaller displacements?

Slug tests were performed at two sites: 1- Asylum Lake in Kalamazoo, Michigan, in a well 0.33 ft in diameter and 97 ft deep. 2- Another unconfined aquifer in Portage, Michigan, in wells 0.167 ft in diameter and relatively shallow depth of 16 ft. Both sites were slug tested using two different sizes of physical slug rods. The smaller slug was 5 ft in length with 0.12 ft diameter. The larger slug was 6.91 ft long and 0.125 ft in diameter.

The four-way analysis of variance (ANOVA) examination of 39 slug tests on wells in these two aquifers confirmed that log K depends highly significantly upon slug in/out, and as expected, the aquifer tested. What was not expected was the average log K for tests in 5-cm diameter wells using slug in was highly significantly larger than average log K for tests using slug out; these means also differed significantly by slug size. Importantly, mean log K’s produced by the Hvorslev (1951) method are highly significantly larger than mean log K’s produced by the other three methods above. The size of the initial water-level deflection, Yo, does not affect the results for log K. The use of a larger or smaller slug does not significantly affect the log K values. Finally, assuming that a large K value better represents the true value, closer to a pump test K value, Hvorslev (1951) gives the largest K values.

Included in

Hydrology Commons