Evidence for hydraulic heterogeneity and anisotropy in the mostly carbonate Prairie du Chien Group,...

www.elsevier.com/locate/sedgeo

Sedimentary Geology 18

Evidence for hydraulic heterogeneity and anisotropy in the mostly

carbonate Prairie du Chien Group, southeastern Minnesota, USA

Robert G. Tipping a,*, Anthony C. Runkel a, E. Calvin Alexander Jr.b,

Scott C. Alexander b, Jeffery A. Green c

a Minnesota Geological Survey, 2642 University Avenue, Saint Paul, MN 55114, United Statesb University of Minnesota, Department of Geology and Geophysics, 310 Pillsbury Drive SE, Minneapolis, MN 55455, United States

c Minnesota Department of Natural Resources, Division of Waters, Rochester, 2300 Silver Creek Road, Rochester, MN 55906, United States

Abstract

In southeastern Minnesota, Paleozoic bedrock aquifers have typically been represented in groundwater flow simulations as

isotropic, porous media. To obtain a more accurate hydrogeologic characterization of the Ordovician Prairie du Chien Group, a new

approach was tested, combining detailed geologic observations, particularly of secondary porosity, with hydraulic data. Lithologic

observations of the depositional and erosional history of the carbonate-dominated bedrock unit constrained characterization of both

primary (matrix) and secondary porosity from outcrops and core. Hydrostratigraphic data include outcrop and core observations

along with core plug permeability tests. Hydrogeologic data include discrete interval aquifer tests, borehole geophysics, water

chemistry and isotope data, and dye trace studies. Results indicate that the Prairie du Chien Group can be subdivided into the

Shakopee aquifer at the top, consisting of interbedded dolostone, sandstone and shale, and the underlying Oneota confining unit

consisting of thickly bedded dolostone. The boundary between these two hydrogeologic units does not correspond to lithostrati-

graphic boundaries, as commonly presumed. Groundwater flow in the Shakopee aquifer is primarily through secondary porosity

features, most commonly solution-enlarged bedding planes and sub-horizontal and vertical fractures. Regional scale preferential

development of cavernous porosity and permeability along specific stratigraphic intervals that correspond to paleokarst were also

identified, along with a general depiction of the distribution of vertical and horizontal fractures. The combination of outcrop and

core investigations, along with borehole geophysics, discrete interval aquifer tests, water chemistry and isotope data and dye trace

studies show that the Prairie du Chien Group is best represented hydrogeologically as heterogeneous and anisotropic. Furthermore,

heterogeneity and anisotropy within the Prairie du Chien Group is mappable at a regional scale (N15,000 km2).

D 2005 Elsevier B.V. All rights reserved.

Keywords: Dolostone; Groundwater; Paleokarst; Prairie du Chien Group; Dual porosity; High-resolution flowmeter

1. Introduction

The Lower Ordovician Prairie du Chien Group,

along with the underlying Cambrian Jordan Sandstone,

is a primary source of groundwater in southeastern

0037-0738/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.sedgeo.2005.11.007

* Corresponding author. Fax: +1 612 627 4778.

E-mail address: [email protected] (R.G. Tipping).

Minnesota and southwestern Wisconsin (Fig. 1). This

part of Minnesota was largely unglaciated during the

most recent Pleistocene Late Wisconsin glacial ad-

vance. As a result, bedrock is at depths of less than

15 m across most of the area, with overlying unconsol-

idated material consisting primarily of loess and weath-

ered bedrock residuum. The combination of shallow-to-

bedrock conditions and permeable bedrock make these

4 (2006) 305–330

Fig. 1. Map showing extent of the Prairie du Chien Group, southeastern Minnesota. Areas where the Prairie du Chien Group is the uppermost

bedrock layer is shown, as well as areas where it is overlain by younger bedrock. Locations are shown for boreholes, dye traces, quarries and other

features discussed in the text.

R.G. Tipping et al. / Sedimentary Geology 184 (2006) 305–330306

aquifers highly productive, but also highly susceptible

to contamination from activities at the land surface.

Paleozoic aquifers in southeastern Minnesota have

typically been treated in groundwater flow models as

equivalent porous media. Although models based on

this assumption work well for estimates of aquifer yield,

they have limited usefulness for predicting precise flow

paths and groundwater velocities. The goal of this study

was to test the hypothesis that the Prairie du Chien

Group is best characterized as hydraulically heteroge-

neous and anisotropic. The hydrogeology of thick, car-

bonate-dominated units has proven quite difficult to

characterize. Because the Paleozoic strata in this portion

of the central mid-continent are relatively undeformed,

R.G. Tipping et al. / Sedimentary Geology 184 (2006) 305–330 307

well exposed along river systems, and mapped exten-

sively, they provide a unique opportunity to combine

borehole hydrostratigraphic and hydraulic data with

outcrop observations of porosity and permeability. A

complex distribution of fractures and solution features,

combined with widely variable hydraulic properties

found in this study, is inconsistent with classification

of the Prairie du Chien Group as a single hydrogeologic

unit. Of particular importance to this study was the

identification of high permeability intervals within the

unit, along with intervals that confine underlying re-

gional aquifers. A variety of methods at different scales

were used to identify aquifer heterogeneity at different

scales, from cm to km.

2. Data and methods

Recent improvements to the hydrogeologic charac-

terization of Paleozoic aquifers have been based on

correlating hydraulic conductivity to stratigraphic fea-

tures (e.g., Muldoon et al., 2001). Our investigation of

Prairie du Chien Group hydrogeology follows the

methods applied to other Paleozoic rocks in southeast-

ern Minnesota (Runkel et al., 2003; Runkel et al.,

2006—this volume). Specifically, our characterization

of hydrogeologic attributes of the Prairie du Chien

Group is based on hydraulic data interpreted within

the context of hydrostratigraphic attributes.

According to Seaber (1988), hydrostratigraphic attri-

butes are assigned based on rock properties alone, and,

with the exception of plug-scale permeability testing,

are not based on hydraulic measurements of any kind.

This study assembles matrix porosity and permeability

estimates from plug samples taken from core and out-

crop (MUGSP, 1980; Setterholm et al., 1991). Second-

ary porosity in the form of fractures and solution

cavities was described directly from observations of

core, outcrop and indirectly from an inventory of sink-

holes and spring locations and stratigraphic positions.

Lithofacies are broadly grouped into three matrix

hydrostratigraphic components: (1) a carbonate compo-

nent consisting of limestone and dolostone; (2) a coarse

clastic component consisting of fine- to coarse-grained

quartz sandstone; and (3) a fine clastic component

consisting of very fine quartz sandstone, siltstone and

shale. Borehole geophysical methods included an as-

sessment of fracture density using video logs and mea-

sured fracture/solution feature aperture using caliper

logs.

Hydraulic data assembled for this study included

borehole-scale characterization based on high resolu-

tion electromagnetic (EM) flowmeter measurements,

discrete interval packer testing, hydrochemical logging,

temperature and fluid resistivity measurements, and a

review of historic conventional aquifer tests. Most log-

ging was performed on existing test wells that were free

of pumps or other obstructions. Additional logging was

performed on three test holes drilled as part of the

project.

The EM flowmeter was used to measure vertical

flow within the open borehole; positive values mea-

sured with the flowmeter indicate upflow in the bore-

hole, and negative values indicate downflow. Unlike

impeller-type flowmeters which require several liters

per minute in order to produce measurable flow, the

EM flowmeter can measure flows as low as to 0.01 l/

min. After fractures and zones within the open borehole

of equal diameter have been identified with a caliper

log, the flowmeter is btrolledQ upwards at a constant

speed to identify shifts that may indicate water moving

into or out of the borehole. Measurements are then

made at stationary positions above and below fractures,

in areas of equal hole diameter. Flow either up or down

from one transmissive zone to another indicates con-

finement between the two zones. In this way, flowmeter

data can be used to identify hydraulically active zones

within the borehole. Flowmeter logging was done under

ambient and stressed (either pumping or injection) con-

ditions. Using methods outlined by Paillet et al. (2000),

flowmeter logs under these conditions can be used to

estimate transmissivity for individual high permeability

zones by providing flow measurements under two dif-

ferent head gradients. Flow measurements from an

individual fracture under different head conditions

eliminates head dependence in the calculation of trans-

missivity from flow. In this study, high permeability

zones within the Prairie du Chien Group were typically

single fractures, or clusters of fractures, or zones of

cavernous porosity formed by a combination of frac-

tures and dissolution. Transmissivity estimates from

flowmeter logs for these zones were converted to hy-

draulic conductivity by dividing transmissivity by zone

thickness. In areas where zone thickness could not be

determined, a maximum thickness of approximately 0.3

m was used.

In addition to flowmeter logging, discrete interval

packer tests and hydrochemical logging was conducted

at two of the test holes drilled as part of this project.

Packers were used to isolate 4 m sections of the bore-

hole in order to measure yields and head levels. Loca-

tions of tested intervals were chosen in portions of the

open hole where upper and lower packers could be set

in competent, relatively unfractured bedrock. As a re-

sult, zones of major cavities in the lower Shakopee


Formation could not be tested. Pumping rates were

adjusted until water levels within the interval stabi-

lized. With the pump off, head levels were checked in

the open borehole by alternately opening the upper

and lower packer and measuring ambient head level.

Continuous temperature, pH, and chloride profiles of

portions of these two wells were measured using

hydrochemical logging sondes, providing additional

information on water movement within the open

borehole.

A combination of dye trace investigations con-

ducted as part of this study near a spring-fed fish

hatchery in Lanesboro, Minnesota, along with data

from previous dye traces provided useful information

at both borehole and regional scales on the stratigraph-

ic position of preferential flow within the Prairie du

Chien Group, and time-of-travel estimates. The Lanes-

boro trace was conducted in response to problems

encountered during the construction of a private well

near the hatchery. Shortly after encountering two 3 m

intervals of cavernous porosity and the loss of circu-

lation mud during drilling, the hatchery’s Main Spring

turned cloudy, indicating a connection between the

conduits encountered while drilling and the spring.

To quantify the connections to the spring and to assure

the efficacy of the grouting, the fluorescent water

tracing dye fluorescein was added with the gravel in

the lower conduit (at ~94 m below the land surface)

and eosin was added with the gravel in the upper

conduit (at ~70 m) and samples were taken at the

nearby spring.

Additional regional-scale hydraulic information as-

sembled as part of this study included hundreds of

specific capacity data from the state water well

database converted to hydraulic conductivity using

methods described in Bradbury and Rothschild

(1985). The results of borehole investigations, obser-

vations from core and outcrop, dye tracing and

conventional aquifer test and specific capacity data

were correlated regionally to provide a comprehen-

sive hydrogeologic characterization of the Prairie du

Chien Group.

3. Geology of the Prairie du Chien Group

The Prairie du Chien Group, ranging in thickness

from approximately 35 to 60 m, is divided into the

lower Oneota Dolomite and upper Shakopee Formation

(Fig. 2). An unconformity that separates the two for-

mations corresponds to the Sauk C1–Sauk C2 subse-

quence boundary (Smith et al., 1996). The bulk of the

Oneota Dolomite consists of thick beds of very fine

grained dolostone. Stromatolites and thrombolites are

common, and the lowermost part contains some silici-

clastic interbeds. The Shakopee Formation is a mixed

carbonate-siliciclastic unit. It contains thin to medium

beds of dolomudstone, oolitic dolostone, shale, sandy

dolostone, and sandstone. Much of the Shakopee For-

mation contains classic features of btropicalQ, peritidal,mixed siliciclastic/carbonate facies, including stromato-

lites, oolites, intraclastic beds, and desiccation cracks.

The Oneota Dolomite also contains features similar to

modern tropical carbonate settings, but differs from the

Shakopee Formation in the presence of thick, internal

structureless burrow-mottled, and thrombolitic beds

lacking siliciclastics that were deposited in a relatively

deep subtidal setting.

3.1. Karst features

Karst processes active early in the history of the

Prairie du Chien Group initiated the development of a

regionally extensive zone of high permeability both

above and below the sequence boundary between the

Shakopee Formation and Oneota Dolomite. First, sub-

areal exposure of lithified Oneota Dolomite sediments

prior to Shakopee Formation deposition exposed these

rocks to extensive weathering and dissolution over a

period of approximately 5 my (Smith et al., 1996).

Karst processes occurred at or near the water table

and saltwater interface (e.g., Choquette and James,

1988). At the end of Shakopee Formation deposition,

there was a still longer period of subareal exposure of

over 15 million years, corresponding to the Sauk-

Tippicanoe Sequence Boundary (Smith et al., 1996;

Sloss, 1963). Many of the karst features formed earlier

were likely re-activated and expanded.

The two episodes of major karst dissolution in the

Ordovician Period took place over a large area that

includes all of southeastern Minnesota and adjacent

areas, and under relatively stable geologic conditions.

The resulting paleokarst zone has a distinct character

and stratigraphic position within the Prairie du Chien

Group. A photograph from a quarry (Fig. 3) shows

weathering in the Oneota Dolomite, just below the

Shakopee Formation–Oneota Dolomite contact. Cavi-

ties in the upper Oneota paleokarst zone at this site are

1 to 2 m high, 1 to 4 m wide and occur around the entire

perimeter of the quarry. The cavities are partly or

entirely filled with a variety of sediment, including

fine-grained dolostone, carbonate and sandstone brec-

cia, loosely cemented sandstone, and unconsolidated

sediment ranging in grain size from clays to boulders.

A map of a cave in the Prairie du Chien Group that

Fig. 2. Bedrock stratigraphic column of the Prairie du Chien Group, southeastern Minnesota. The Prairie du Chien Group rock record, composed of

the Shakopee Formation and Oneota Dolomite, is early Ordovician in age, bounded above by the Middle Ordovician St. Peter Sandstone, and below

by the Upper Cambrian Jordan Sandstone. The Oneota Dolomite is a thickly bedded dolostone, in contrast to the Shakopee Formation, composed of

more thinly bedded sandy dolostone with subordinate beds of sandstone and shale. The Shakopee–Oneota contact is unconformable, characterized

above and below by paleokarst features such as collapse breccias and large cavities partially or completely filled with both lithified and

unconsolidated sediment.


occurs in the same stratigraphic position as the paleo-

karst horizon shown in Fig. 3 provides a plan view of

the high-permeability zone (Fig. 4). The maze-like,

anastomosing morphology of this cave is characteristic

of caves that form in exposed carbonate platforms.

Similar caves are found in the Yucatan Peninsula and

other exposed carbonate platforms where large cavities

have developed at or just below either the present day,

or earlier Pleistocene water tables (Mylroie and Carew,

2000; Frank et al., 1998). This water table–subwater

Fig. 3. Photograph of a quarry north of Plainview in south-central Wabasha County, Minnesota, showing the paleokarst horizon just below the

contact of the Shakopee Formation and Oneota Dolomite. Cavities range in size from 1 to 2 m in height and 1 to 4 m in diameter. The morphology

of these caves is distinctly oblong, with broad openings occurring horizontally in series but apparently disconnected. Note person in lower left of the

photograph for scale. (Quarry location shown in Fig. 1.)


table (phreatic) cave network is distinctly different from

vadose (above the water table) joint controlled karst

cave systems, where enlarged orthogonal joints are

typical. Similar paleokarst horizons in Lower Ordovi-

cian dolostones have been recognized elsewhere across

North America (Kerans, 1988; Palmer and Palmer,

1989; Wilson et al., 1992), demonstrating the regional

significance of this process in early Ordovician time,

and the ubiquitous position of paleokarst within the

Prairie du Chien Group.

4. Hydrogeology of the Prairie du Chien Group

4.1. Hydrostratigraphic attributes: matrix porosity and

permeability

The Oneota Dolomite and overlying Shakopee For-

mation consist largely of carbonate rock with a matrix

porosity of less than 10% (Runkel et al., 2003). Vertical

hydraulic conductivity based on plug tests is low to

very low, with values ranging from 8.24�10�8 to

8.24�10�5 m/day (MUGSP, 1980; Setterholm et al.,

1991) (values converted from md to m/day using con-

version of 1 md=8.24�10�4 m/day at 20 8C). Hori-zontal permeability can be as much as 10 times greater

than vertical permeability in laminated samples (Runkel

et al., 2003). Fine- to coarse-grained clastic interbeds

are present in the lowest part of the Oneota Dolomite,

the Coon Valley Member, and as a subordinate compo-

nent of the Shakopee Formation (Fig. 2). Individual

sandstone beds within the Shakopee Formation are

typically less than 1 m thick. In the lower part of the

Shakopee Formation, siliciclastic beds can reach a

thickness of 12 m locally, where they are referred to

as the New Richmond or Root Valley Sandstone. Al-

though not measured, the New Richmond Sandstone is

expected to have a high matrix hydraulic conductivity,

with values as high as 8.24�10�1 m/day, measured in

the Jordan Sandstone (Runkel et al., 2003). The perme-

ability of subordinate shale beds within the Shakopee

Formation has also not been measured, but based on

measurements made of similar rocks, values of vertical

hydraulic conductivity as low as 8.24�10�5 m/day are

expected (Runkel et al., 2003), with horizontal perme-

ability at least an order of magnitude greater than

vertical permeability.

Fig. 4. Plan view of Kruger’s Cave, Plainview area, Wabasha County, Minnesota, showing the maze-like, anastomosing nature of the high

permeability zone. Note that bedrock continues east of the area shown. Plan depicts known passages as of November 1983. From a drawing by

David Gerboth, Minnesota Speleological Survey. (Cave location shown in Fig. 1.)


4.2. Hydrostratigraphic attributes: secondary porosity

The distribution of fractures and solution features in

the Paleozoic bedrock of southeastern Minnesota has

been investigated and summarized by Runkel et al.

(2003). Fractures are divided into systematic and non-

systematic types. Systematic fractures are flat-sided

openings oriented perpendicular to bedding. They are

also referred to as bjointsQ, and are typically the most

prominent fractures in large outcrops—commonly evi-

dent in side view even at distances of hundreds of

meters as straight, vertical openings with apertures up

to several centimeters and a spacing of several meters.

Two or more joint sets with a preferred orientation

intersect one another to create an orthogonal network

in plan view. Nonsystematic fractures are irregular,

curved, to straight fractures that most commonly inter-

sect bedding at a lower angle than systematic fractures,

appear more randomly distributed, are more variable in

their orientation and shape, and generally have aper-

tures of less than 1 cm. They are more densely distrib-

uted than systematic fractures and commonly are no

more than a few meters in length. Secondary porosity

features preferentially aligned parallel to bedding are

referred to as bedding plane bfracturesQ in this report.

Secondary porosity and permeability within the

Prairie du Chien Group is present in solution-enlarged

systematic and nonsystematic fractures and along bed-

ding planes. Unlike younger bedrock in southeastern

Minnesota, tracing individual beds in outcrop for more

than a few meters within the Prairie du Chien is often

not possible, due to the recrystallized (dolomitized)

nature of this bedrock unit and because original bed-

ding was lenticular and irregular in thickness. Al-

though bedding planes are often obscured, linear

arrangements of multiple cavities within the Prairie

du Chien are not, and they typically can be traced

over distances of several tens of meters along a line

parallel to master bedding. Cavities along the same

plane range from a few centimeters to more than a

meter in height (Fig. 3).

In addition to preferential flow along bedding

planes, the origin of these cavities may also be attrib-

utable to the former presence of rock layers that are

more prone to dissolution, such as evaporite minerals.

Macroscopic secondary porosity features within the

Oneota Dolomite are chiefly restricted to a stromato-

litic facies that is as much as 20 m thick in the

uppermost part of the formation and along individual

beds less than a meter thick in the middle and lower

part of the formation. These high porosity intervals

within the Oneota Dolomite are separated from one

another by thick (N5 m) intervals of very fine-grained

dolostone, with minimal secondary porosity (Runkel et

al., 2003). Macroscopic secondary porosity features

within the Shakopee Formation are more ubiquitous


and include oomoldic pores and decimeter scale cavi-

ties in addition to enhanced permeability along stro-

matolitic horizons.

The distribution of macroscopic secondary porosity

is consistently associated with specific intervals within

the Prairie du Chien Group even in conditions where

the Prairie du Chien Group is buried under as much as

70 m of overlying bedrock in Minnesota, and under as

much as 300 m of overlying bedrock in Iowa (Des

Moines Water Works, 1995). Caliper and core logs

for selected wells in southeastern Minnesota illustrate

the characteristic distribution of secondary porosity

within the Prairie du Chien Group (Fig. 5). Enhanced

development of secondary porosity within the Shako-

pee Formation and upper Oneota Dolomite, particularly

near the Shakopee Formation–Oneota Dolomite contact

is clearly distinguishable from the lower Oneota Dolo-

mite, where fractures and solution features are less

common. Caliper logs are particularly useful in distin-

guishing between the Shakopee Formation and the

Oneota Dolomite. Unlike other Paleozoic rocks in

southeastern Minnesota where the presence of feld-

spathic strata at specific intervals makes the natural

gamma log a useful tool for stratigraphic correlation,

the scarcity of feldspathic strata within the Prairie du

Chien Group make it difficult to use natural gamma

logs to distinguish between Shakopee Formation and

Oneota Dolomite (compare natural gamma to caliper

logs, Figs. 7–9) Correlation of caliper logs as shown in

Fig. 5 clearly illustrates the characteristic distribution of

secondary porosity within the Prairie du Chien Group

over a broad region.

In shallow-to-bedrock (b60 m below bedrock sur-

face) conditions, both vertical and bedding plane frac-

tures are more ubiquitous and are the dominant

secondary porosity feature often regardless of lithology.

Most of what we know about the distribution of hori-

zontal and vertical fractures is based largely on natural

outcrops, road cuts and quarries, in settings where the

Prairie du Chien occurs as uppermost bedrock. Typical

fractures in such a setting, representative of a large area

of southeastern Minnesota, are shown in Fig. 6. The

Shakopee Formation is the uppermost bedrock unit, and

contains a high density of vertical nonsystematic and

systematic fractures, and closely spaced bedding plane

fractures. Based on a limited amount of data from core,

video and caliper logs, and outcrop, the size and fre-

quency of fractures diminish with depth. In the thickly

bedded Oneota, exposed at the bottom of the quarry

wall, only solution enlarged vertical systematic joints

have open apertures. In other quarries, where the same

thick beds are overlain by several tens of meters of

bedrock, apertures of the systematic joints close down-

ward and commonly do not penetrate these thick beds.

4.3. Hydraulic attributes: borehole-scale investigations

Hydraulic attributes of the Prairie du Chien Group at

the borehole-scale were largely determined using data

collected with the EM flowmeter. These results of this

logging demonstrate that at borehole scale, flow is

dominated by a few discrete (b2 m) intervals of sec-

ondary porosity that are hydraulically active. Intervals

of relatively unfractured rock serve as confining units.

Results of a borehole investigation on a well near the

city of Rochester, Minnesota (Paillet et al., 2000) are

shown in Fig. 7. Data were collected under ambient and

pumping conditions of 7.5 l/min. Slight upflow was

measured under ambient conditions, with water entering

the borehole from discrete fractures near the bottom of

the Shakopee Formation (well penetrates nearly full

thickness of the Shakopee Formation) and at two addi-

tional fractures, before exiting in fractures just below the

well casing (hydraulically active zones k1–k4 in Fig. 7).

Measurements repeated under pumping conditions show

similar trends. Inverse modeling using flowmeter mea-

surements under ambient and pumped conditions was

used to estimate transmissivity for individual fractures

(Paillet et al., 2000). Estimated hydraulic conductivities

for individual fractures shown in Fig. 7 were calculated

by dividing transmissivity estimates by 0.3 m. Results

show a range of values from 3 to 134 m/day for indi-

vidual fractures within the Shakopee Formation. Inter-

estingly, the largest fracture in the borehole as measured

by the caliper log is not hydraulically active (Fig. 7).

Flowmeter measurements made while changing hy-

draulic conditions surrounding the borehole in addition

to changing conditions within the borehole itself also

reveal distinct differences in hydraulic character of the

Shakopee Formation and Oneota Dolomite. The results

from a borehole investigation in Faribault, Minnesota

(Paillet et al., 2000) are shown in Fig. 8. Here, mea-

surements again were made under both ambient and

pumping conditions of 7.5 l/min. Large ambient down-

flow of up to 30 l/min was measured, with contribu-

tions coming from fractures just below the casing in the

upper part of the Shakopee Formation (zones k7, k6,

k5, Fig. 8). Water exits the borehole at several fractures

in the lower Shakopee Formation and one in the upper

Oneota Dolomite as indicated by step-wise reductions

in measured downflow below these zones (zones k4,

k3, k2 and k1, Fig. 8). Additional measurements were

made at this well while pumping a nearby municipal

well at several thousand liters per minute. The munic-

Fig. 5. Representative caliper and core logs showing the distribution of secondary porosity features in the carbonate-dominated Prairie du Chien Group across southeastern Minnesota. The depth of

burial beneath the bedrock surface (in meters) is listed beside each log. The sites range from deep bedrock (N60 m below bedrock surface) to shallow bedrock (b60 m below bedrock surface)

conditions. The distribution of pores, largely dissolution cavities, is stratigraphically controlled: the Shakopee Formation and upper part of the Oneota Dolomite have a high density of large cavities.

In contrast, secondary porosity in the middle to lower parts of the Oneota Dolomite is much lower.

R.G.Tippinget

al./Sedimentary

Geology184(2006)305–330

313

Fig. 6. Quarry exposing carbonate rock of the Shakopee Formation and Oneota Dolomite near Mankato in Blue Earth County. Nonsystematic

fractures are abundant in the upper part of the bedrock exposed in the quarry. Only widely spaced, systematic fractures are evident in the lower part

of the quarry. The depth to which nonsystematic and systematic fractures extend continuously beneath the bedrock surface will vary from place to

place in southeastern Minnesota (Runkel et al., 2003). (Quarry location shown in Fig. 1.)


ipal well, located 15 m away from the measured well, is

open to the full thickness of the Prairie du Chien and

underlying Jordan Sandstone. Pumping in the munici-

pal well produced turbulent flow in the monitoring

well, with downflow measurements exceeding the mea-

surement capacity of the EM flowmeter. No flow was

measured below hydraulic zone k2, with water exiting

along bedding plane fractures in the lower Shakopee,

connected to the open hole in the nearby municipal well

(Paillet et al., 2000).

Flowmeter results from wells with open holes

extending from the Shakopee Formation through the

Oneota Dolomite and into the underlying Jordan Sand-

stone show head differences between the Shakopee

Formation and Jordan Sandstone, indicating that por-

tions of the Oneota Dolomite act as a confining unit.

Ambient flow measurements from 3 wells with this

construction are shown in Fig. 9. In Well A, upflow of

several liters per minute was measured, with water

entering the borehole from the Jordan Sandstone, mov-

ing past the lower Oneota Dolomite, with most of the

upflow exiting the borehole near the Shakopee Forma-

tion–Oneota Dolomite contact. In addition, downflow

was measured in the same borehole, with water enter-

ing in fractures just below the casing, and exiting

though a series of fractures in the upper Shakopee

Formation. In Well B, located approximately 1 km

from Well A, upflow was again measured from the

Jordan Sandstone, past the Oneota Dolomite and into

the lower Shakopee Formation, this time at rates of up

to 12 l/min. In Well C, strong ambient downflow of up

to 15 l/min was measured in the upper Shakopee

Formation, with water entering in two fractures at the

top of the open hole, and exiting through a series of

fractures in the upper Oneota Dolomite.

In addition to borehole geophysical measurements,

discrete interval packer tests and water chemistry mea-

surements were made at several test holes as part of this

Fig. 7. Results of a borehole geophysical investigation of the carbonate-dominated strata of the Shakopee Formation in the city of Rochester (Fig. 1)

(Paillet et al., 2000). Flowmeter data from the open hole were collected under ambient conditions and while pumping 7.5 l/min out of the borehole.

Note that vertical borehole flow increases at discrete horizons, typically less than 30 cm thick, with estimated hydraulic conductivity as high as 134

m/day (see text for discussion of hydraulic conductivity estimates). Minnesota unique number 485610.


study in order to investigate the hydraulic attributes of

the Prairie du Chien at the borehole scale. Results from

a test hole drilled in Northfield, Minnesota illustrates

the nature of groundwater flow within the Prairie du

Chien Group under shallow bedrock conditions (Fig.

10). At this site, the uppermost bedrock layer is the

Shakopee Formation. Depth to bedrock is 5.4 m. The

borehole is open to the full thickness of the Shakopee

Formation and Oneota Dolomite, and 14 m of the

underlying Jordan Sandstone. Video and caliper logs

show characteristic distribution of horizontal to sub-

horizontal fractures and solution cavities, with relative-

ly fewer major cavities in the Oneota Dolomite com-

pared to the Shakopee Formation. Largest voids in the

borehole occur in the lower Shakopee Formation. Pack-

er test pumping rates ranged from zones with yields

below the limit of the pump in the lower Oneota

Dolomite to 3.7 l/min in the upper Shakopee Forma-

tion. With the pump off, head levels were checked by

alternately opening the top and then bottom packer, and

measuring head levels under each condition. An up-

ward gradient measured through the Oneota Dolomite,

with the largest gradient of 2.4 m was measured in its

lowermost section (Fig. 10). A downward gradient

(maximum 1.5 m) was measured in the upper portion

of the borehole. Flowmeter measurements recorded

upflow out of the Jordan and past the Oneota at a rate

of 40 l/min. Upflow exited the borehole near the top of

the large zone of major cavities in the lower Shakopee

Formation, as indicated by a reduction in measured

upward (positive) flow above this zone, and a shift in

the trolling log (Fig. 10).

Fig. 8. Results of a borehole geophysical investigation of the carbonate dominated strata of the Shakopee Formation and uppermost Oneota

Dolomite strata at the city of Faribault (Fig. 1) (Paillet et al., 2000). Flowmeter data from the open hole were collected under ambient conditions and

while pumping 7.5 l/min out of the borehole. Note that vertical borehole flow increases through discrete horizons, typically less than 30 cm thick,

with hydraulic conductivity as high as 312 m/day. Ambient inflow to the well occurs in the upper half of the Shakopee Formation and outflow

occurs in the lower Shakopee Formation and uppermost Oneota Dolomite, demonstrating hydraulic separation of conduits within the Shakopee

Formation. Minnesota unique number 625327.


Additional temperature, pH, conductivity and chlo-

ride logging at the Northfield site was conducted in

April 2001 during spring snowmelt. Increasing temper-

ature and decreasing chloride concentrations and a shift

in pH with depth in the upper portion of the borehole

record (Fig. 10) the gradual influx of snowmelt water

(high chloride from road salt) through fractures in the

upper Shakopee Formation, mixing with warmer, more

dilute upflow from the underlying Jordan Sandstone.

4.4. Hydraulic attributes: regional-scale investigations

Results from dye traces conducted as part of this

study, along with results from previous investigations

provide hydraulic attribute information on the Prairie

du Chien Group at greater-than-borehole scale. At the

Lanesboro trace, the distance from the well where the

dyes were input to the spring at the hatchery is approx-

imately 680 m. This trace takes place through conduits

Fig. 9. Results of a borehole geophysical investigations of three boreholes open from the Shakopee Formation to the Jordan Sandstone. (A) Upflow

measured out of the Jordan, past the Oneota and exiting in the Shakopee formation. Minnesota unique well number 676444. (B) Upflow measured

out of the Jordan, past the Oneota and exiting in the Shakopee formation. Abrupt jumps in station measurement values indicate water entering or

exiting the borehole through fractures. Minnesota unique well number 676445. (C) Strong downflow measured from the upper to lower Shakopee

Formation. Caliper log shows characteristic smooth borehole wall through the Oneota, contrasted with an increase in log deviations in the borehole

wall in the Shakopee Formation, where there is a greater frequency of horizontal to sub-horizontal bedding plane fractures. Minnesota unique well

number 508116.


Fig. 10. Results of a borehole geophysical investigation of the carbonate-dominated strata of the Prairie du Chien Group and underlying coarse clastic strata of the Jordan Sandstone at Carleton

College in Northfield, Minnesota.

R.G.Tippinget

al./Sedimentary

Geology184(2006)305–330

318


in the lower Oneota Dolomite, in a stratigraphic posi-

tion considered to be part of the Oneota confining unit

under deep bedrock conditions. The presence of con-

duits within the Oneota Dolomite at this site is consid-

ered to be the result of its extensive fracturing and

karstification in bluff-side, shallow bedrock conditions.

Borehole dye input locations are shown in Fig. 11.

Flow is upward from the dye input locations to the

spring. The leading edge of the fluorescein dye from the

lower conduit reached the main spring in about 1.5

h and the peak of the breakthrough curve reached the

spring in 2.4 h. The leading edge of the eosin dye from

the upper conduit reached the spring in 1.1 h and the

breakthrough curve peaked at 1.6 h. These values doc-

ument straight-line flow velocities of about 600 m/h in

the Prairie du Chien conduits. Assuming a curvature

Fig. 11. Construction diagram of a well used for the dye trace near the State

684602. After circulation was lost, drilling operations were halted until a

completed without causing further damage to the conduit system. Ultimately

the bottom of the lower conduit. Pea gravel was then added to the outside of t

the top of the lower conduit to the bottom of the upper conduit; pea grave

remainder of the well was grouted to the surface. Conduits are located in the

with enhanced weathering and karstification of the Oneota near bedrock val

factor of 1.5, the actual water flow velocity in the

conduits may be about a 1 km/h.

Similar velocities measured from dye traces con-

ducted over longer distances were found at the Oronoco

land fill (3 km), near Rochester, Minnesota (Fig. 12A)

(Donahue and Associates, Inc., 1991), and in Fillmore

County near the Lanesboro trace (4 km) (Fig. 12B)

(Wheeler, 1993). The location of the Lanesboro trace is

shown on Fig. 12B. Horizontal flow speeds of ground-

water at the Oronoco site were as rapid as 244 m/day

(Donahue and Associates, Inc., 1991). The Fillmore

County study by Wheeler (1993) measured flow speeds

as rapid as 10 km/day. At both sites, flow occurred

roughly along the Shakopee Formation–Oneota Dolo-

mite contact considered to be within the zone of high

permeability associated with paleokarst near the contact.

Fish Hatchery, Lanesboro, Minnesota. Minnesota unique well number

suitable resolution could be found that allowed for the well to be

, the well was completed by grouting from the bottom of the casing to

he casing until it filled the lower conduit. Grouting was continued from

l and cement grout were used to fill the upper conduit and then the

lower Oneota and near the Prairie du Chien–Jordan contact, consistent

leys, and the location of springs in the region. See text for discussion.

Fig. 12. Hydrogeologic character of the Prairie du Chien Group and Jordan Sandstone in shallow to deep bedrock conditions at a landfill near Oronoco (Olmsted County), and in eastern Fillmore

County. See Fig. 1 for location. (A) At Oronoco, a water-table aquifer lies in the Shakopee Formation and uppermost Oneota Dolomite, which have abundant fractures and dissolution features. The

Jordan Sandstone at this site is an intergranular aquifer with a regional groundwater system. The two aquifers are hydraulically separated by Oneota Dolomite with relatively few secondary porosity

features. This depiction is based on borehole videos, natural gamma and caliper logs, cuttings, dye tracing, water chemistry, and potentiometric levels. Modified from Donahue and Associates, Inc.

(1991). (B) Conduit flow in the Prairie du Chien Group in eastern Fillmore County based on dye trace investigations. Note the similarity of hydrogeologic conditions to those at Oronoco. Modified

from Alexander and Lively (1995).

R.G.Tippinget

al./Sedimentary

Geology184(2006)305–330

320


Additional information on regional hydraulic attri-

butes of the Prairie du Chien Group comes from hy-

draulic conductivity estimates based on the state water

well database specific capacity data. Hydraulic conduc-

tivity was calculated from specific capacity for 807

wells, using the methods of Bradbury and Rothschild

(1985). Depths and calculated hydraulic conductivities

from Prairie du Chien wells in shallow and deeply

buried bedrock conditions are shown in Fig. 13. Wells

are finished almost exclusively within the Shakopee

Formation of the Prairie du Chien Group; under condi-

tions in which the top of the Prairie du Chien is the

uppermost bedrock layer and conditions where it is

overlain by younger bedrock, the majority of wells

penetrate less than 10 m into the Prairie du Chien

Group (Fig. 13A, C). Wells finished in the Shakopee

under conditions where the Prairie du Chien is upper-

Fig. 13. (A) Box plot of depth (m) in wells from the top of the Prairie du Ch

Prairie du Chien is the uppermost bedrock unit. (B) Box plot of hydraulic c

wells in conditions where Prairie du Chien is the uppermost bedrock unit. (C

the bottom of the casing (start of well open hole) in conditions where Prairie

conductivity values in m/day calculated from specific capacity data for wells i

most bedrock have higher yields than Shakopee wells

where the Prairie du Chien is not first bedrock, support-

ing the hypothesis that the Shakopee has greater frac-

ture density under shallow bedrock conditions (Fig.

13B, D). Wells are completed over a wider range of

depths in the Prairie du Chien Group under conditions

where the Prairie du Chien is not first bedrock (Fig.

13C, A), presumably because under deeper bedrock

conditions, the Shakopee Formation alone does not

provide sufficient yields.

5. Hydrogeologic synthesis

The results of observations from core and outcrop,

borehole investigations, dye tracing and conventional

aquifer test and specific capacity data provide an im-

proved hydrogeologic characterization of the Prairie du

ien to the bottom of the casing (start of open hole) in conditions where

onductivity values in m/day calculated from specific capacity data for

) Box plot of depth (m) in wells from the top of the Prairie du Chien to

du Chien is not the uppermost bedrock unit. (D) Box plot of hydraulic

n conditions where Prairie du Chien is not the uppermost bedrock unit.


Chien Group at both borehole and regional scales.

Unlike previous studies where the entire Prairie du

Chien Group and underlying Jordan Sandstone have

traditionally been treated as a single aquifer in Minne-

sota (e.g., Kanivetsky, 1978; Delin and Woodward,

1984), this study separates the Prairie du Chien

Group into the Shakopee aquifer consisting of the

Shakopee Formation and upper one-third of the Oneota

Dolomite, and the Oneota confining unit, consisting of

the lower two-thirds of the Oneota Dolomite.

Certain hydrogeologic characteristics are common to

both the Shakopee Formation and the Oneota Dolomite.

As seen from plug tests, intergranular porosity and

permeability are negligible in the recrystallized, dolos-

tone beds of both formations. The largest volumes of

groundwater flow are therefore almost entirely restrict-

ed to discrete intervals of parallel to bedding fractures,

joints, and secondary porosity features ranging in width

from centimeters to meters. The distribution of these

features changes from shallow (less than 60 m below

the bedrock surface) to deep (greater than 60 m below

the bedrock surface) settings. In shallow settings (Figs.

3 and 6), there is a greater density of nonsystematic

fractures, solution-enlarged conduits parallel and per-

pendicular to bedding than under deep settings. Cav-

ernous porosity is present in both shallow and deep

settings, and is characterized by anastomosing channels

as seen in cave maps (Fig. 4) and long-dye trace

breakthrough curves. Hydraulically active secondary

porosity intervals are separated by thicker intervals of

carbonate rock with few secondary pores, some of

which provide confinement at least at the borehole

scale.

A compilation of data tied to stratigraphic attributes

unique to the Shakopee Formation and Oneota Dolo-

mite reveals regional and therefore mappable and pre-

dictable hydrogeologic characteristics of these rocks.

The Shakopee Formation, composed of thin to medium

beds of sandy dolostone with minor sandstone and

shale, has a higher frequency of solution-enlarged bed-

ding plane fractures and secondary porosity features

than the thick, internally structureless beds of the

Oneota Dolomite. Secondary porosity and permeability

in the form of solution-enhanced bedding plane frac-

tures and cavernous porosity is present in the paleokarst

horizon both above and below the Shakopee Forma-

tion–Oneota Dolomite contact. This zone of high per-

meability is present in both shallow and deep bedrock

conditions, and can be traced regionally (Fig. 5). Tests

conducted on several deep wells in Iowa attribute the

primary production interval to a fracture zone near the

Shakopee Formation–Oneota Dolomite contact. Here,

the Prairie du Chien Group is deeply buried, with the

high production interval located between 700 and 716

m below the land surface, and over 300 m below the

bedrock surface (Des Moines Water Works, 1995).

Hydraulic conductivities calculated from water well

specific capacity data, along with the distribution of

open-hole intervals also reflect the regional distribution

of porosity and permeability within the Prairie du Chien

Group. Under shallow bedrock conditions, wells are

mostly finished within the uppermost Shakopee Forma-

tion (Fig. 13A) with hydraulic conductivities higher

than those found in Shakopee Formation wells in

deep bedrock conditions (Fig. 13C, D). Higher hydrau-

lic conductivites in shallow to bedrock Shakopee For-

mation wells are presumably the result of more

integrated fracture networks than under deep bedrock

conditions. Under deep bedrock conditions, casing

lengths are more variable (Fig. 13C), presumably the

result of decreased abundance of fractures perpendicu-

lar to bedding; drillers extend wells further into the

Prairie du Chien Group in order to intercept hydrauli-

cally active bedding plane fractures and/or cavernous

porosity. Under both shallow to bedrock and deep

bedrock conditions, wells are almost exclusively fin-

ished within the Shakopee Formation. No wells were

found that were open only to the Oneota Dolomite,

indicative of its regionally extensive low porosity and

permeability.

5.1. The Shakopee aquifer

Hydraulic data presented in this study, interpreted

in the context of fracture and secondary porosity fea-

ture distribution within the Prairie du Chien Group,

supports the subdivision of these rocks into the Sha-

kopee aquifer consisting of the Shakopee Formation

and upper one-third of the Oneota Dolomite. This

aquifer is broadly characterized by groundwater flow

through discrete bedding plane fractures, separated by

thicker intervals of carbonate rock, some of which

provide confinement within the aquifer, at least at

the borehole scale. Inflow and outflow in the Shako-

pee Formation near the city of Rochester was found to

be limited to a few fractures over 40 m of open

borehole (Paillet et al., 2000; Fig. 7). Calculated hy-

draulic conductivity for these fractures ranged over

two orders of magnitude (Paillet et al., 2000). Step-

wise jumps in the flowmeter station measurements

within the Shakopee Formation, indicative of sharp

breaks in flow in and out of the borehole, were also

found in other wells tested as part of this study (Fig.

9), and are characteristic of flowmeter profiles in open


holes where flow occurs within a few discrete hydrau-

lically active fractures.

Typically, water enters (inflow) or exits (outflow) the

borehole in the upper part of the Prairie du Chien Group

(mostly Shakopee Formation), with the lower most

occurrence of inflow/outflow either just above or

below the Shakopee Formation–Oneota Dolomite con-

tact (Fig. 9). This lower most hydraulically active zone

corresponds to the paleokarst zone within the Prairie du

Chien Group discussed earlier. It marks the lower

boundary of the Shakopee aquifer, and is extends into

the upper one-third of Oneota Dolomite. This zone

often has the highest porosity and permeability within

the Prairie du Chien Group (Figs. 9 and 10).

Additional evidence for enhanced porosity and per-

meability near the Shakopee Formation–Oneota Dolo-

mite contact comes from dye trace studies presented

earlier. Rapid groundwater flow speeds measured with-

in this horizon at the Oronoco site and Fillmore County

Study attest to the high degree of hydraulic heteroge-

neity within the Prairie du Chien Group (Fig. 12).

These and other dye traces in shallow bedrock condi-

tions of southeastern Minnesota yielded relatively wide

breakthrough curves that have long tails and a fine-

scale structure. These breakthrough curves are different

from the narrow breakthrough curves and short tails

seen in the conduits of karst systems in Upper Ordovi-

cian and Devonian age rocks in Minnesota, and are

consistent with movement through complex anastomos-

ing, turbulent flow systems. The anastomosing charac-

ter of voids creates multiple flow paths and reservoirs

where dye can become trapped until mobilized by

subsequent recharge events.

5.2. The Oneota confining unit

In Minnesota, differences in static levels between the

upper part of the Prairie du Chien Group and Jordan

aquifer have been noted in at least seven counties (Hall

et al., 1911; Kanivetsky and Palen, 1982; Kanivetsky,

1984, 1988; Kanivetsky and Cleland, 1990, 1992;

Donahue and Associates, Inc., 1991; RMT, Inc., 1992;

Barr Engineering, 1996; Zhang and Kanivetsky, 1996).

These differences in potentiometric head levels have

typically been attributed to impermeable beds of limited

extent in the Prairie du Chien Group (for example

Kanivetsky and Palen, 1982; Kanivetsky, 1984, 1988;

Kanivetsky and Cleland, 1990, 1992), but no strati-

graphic evidence is provided to support the inter-

pretation that such beds are genuinely only locally

distributed. The results of this study support a different

interpretation: that the differences in head levels noted

in these many different areas reflect the ability of an

interval of rock that is of regional extent across south-

eastern Minnesota, the Oneota Dolomite, to provide

confinement.

Flowmeter results from wells with open holes

extending from the Shakopee Formation through the

Oneota Dolomite and into the underlying Jordan Sand-

stone show head differences between the Shakopee

Formation and Jordan Sandstone, indicating that por-

tions of the Oneota Dolomite act as a confining unit at

least at the borehole scale (Fig. 9A, B). Results in wells

in Fig. 9A and B show inflow into the borehole from

the Jordan Sandstone, strong upflow past the massive

dolostone beds of the lower Oneota Dolomite and

outflow into fractures in the Shakopee Formation.

Water does not enter or leave the open hole within

the lower portion of the Oneota Dolomite. Thickly

bedded, unfractured carbonate bedrock of the lower

Oneota Dolomite confines the underlying Jordan aqui-

fer, resulting in strong ambient upflow into the overly-

ing Shakopee Formation.

Low porosity and permeability in the lower Oneota

Dolomite at the borehole scale is also shown by the

results of packer tests at the Northfield site (Fig. 10).

Isolated 4-m sections of the open hole within the Oneota

Dolomite typically had yields less than the minimum

pumping rate of 1 l/min. Packers also showed that head

differences between the underlying Jordan aquifer and

the overlying Shakopee Formation were maintained

throughout the lower Oneota Dolomite, with a maxi-

mum head difference of 2.4 m in just above the Oneota

Dolomite–Jordan Sandstone contact. Maximum head

difference located over a relatively thin section of the

Oneota Dolomite may be due to a lithologic change in

the bedrock. The Coon Valley Member of the Oneota

Dolomite contains feldspathic sands and shales that may

inhibit the propagation of vertical fractures from the

Oneota Dolomite to the underlying Jordan Sandstone.

These results are similar to those studied by Eaton

(2002), who found sharp head differences over thin

beds within the Ordovician Maquoketa Formation, sug-

gesting that head differences may be due to the termi-

nation of vertical fractures within these beds.

Interpreting borehole results at a larger scale sup-

ports characterization of the Prairie du Chien Group

into a Shakopee aquifer and Oneota confining unit. A

cross section of the Prairie du Chien Group and the

Jordan Sandstone, extending from the Northfield test

hole (Fig. 10) to the Cannon River 425 m away, is

shown in Fig. 14. Snowmelt recharge and upflow from

the Jordan Sandstone exiting the test hole is inter-

preted to be flowing through the lower Shakopee

Fig. 14. Schematic cross section that includes the test hole at Carleton College from the test hole location to the Cannon River, approximately 425 m to the west of the site. Caliper, video,

temperature, borehole flowmeter, and chemistry logs from Fig. 10 have been included at the same vertical scale as the cross section. Water moving upward from the Jordan Sandstone and downward

from fractures in the upper Shakopee Formation exits the borehole near the elevation of the nearby Cannon River.

R.G.Tippinget

al./Sedimentary

Geology184(2006)305–330

324

Fig. 15. Potentiometric surfaces of the St. Peter aquifer (dashed line) and Jordan aquifer (solid line) in northeastern Iowa, demonstrating that part(s)

of the intervening Prairie du Chien Group strata serve as a confining unit that creates hydraulic separation of the two aquifers. Modified from Horick

(1989).


Formation towards the Cannon River. Additional flow-

ing Jordan wells in the area support the conclusion

that Oneota Dolomite hydraulically separates ground-

water in the Shakopee Formation from groundwater in

the Jordan Sandstone. Aquifer tests in similar hydro-

geologic settings (Prairie du Chien uppermost bedrock

Fig. 16. Summary of hydraulic conductivity estimates along with additional information on travel times for the Prairie du Chien Group under deep and shallow bedrock conditions (Appendix A).

R.G.Tippinget

al./Sedimentary

Geology184(2006)305–330

326


near a large river) have also measured confinement in

the lower Oneota (SEH, 2001; Barr Engineering,

1996). Additional aquifer tests in other settings show

a vertical bulk conductivity in the lower Oneota Do-

lomite of about 10�3 to 10�4 m/day (Fig. 15, Appen-

dix A, Fig. 16).

Further evidence for regional confining characteris-

tics of the Oneota come from water chemistry and

isotope analyses, where chemical and groundwater

age differences have been identified in nest wells of

Shakopee–Jordan pairs (Donahue and Associates, Inc.,

1991; Setterholm et al., 1991; Wall and Regan, 1994)

along with head level differences between wells in

similar configurations (Kanivetsky, 1988; Donahue

and Associates, Inc., 1991; Barr Engineering, 1996).

A study in northern Iowa by Horick (1989) provided

evidence that the lower part of the Prairie du Chien

Group provides effective confinement at a regional

scale (Fig. 15). Horick showed that the potentiometric

surface of water in the St. Peter aquifer is 15 to 60 m

higher than that of the Jordan Sandstone across much of

northeastern Iowa along its border with Minnesota. In

the absence of any known beds in the lower St. Peter

aquifer that could create such confinement, Horick

concluded that some part of the Prairie du Chien

Group is a bconfining intervalQ.As with all other confining units, the Oneota Do-

lomite provides confinement where it is not breached

by interconnected networks of secondary porosity fea-

tures, a situation that most commonly occurs locally in

shallow bedrock or bluff-side conditions. At the Lanes-

boro trace, the well used for dye input up gradient

from the spring encountered two conduits at different

levels within the Oneota Dolomite, both shown by

traces to be connected to the spring (Fig. 11). These

results indicate that the density of conduits in this part

of the flow system is very high. It also further illus-

trates the complexity of this karst conduit system as no

dye was found in Triple Spring, the other water source

for the hatchery. In 1992, Rhodamine WT was poured

into a stream sink 3400 m due south of the hatchery

and was detected in both springs. In the case of the

Lanesboro trace, groundwater flow occurs within the

Oneota Dolomite in bluff-side conditions, where frac-

tures and secondary porosity features extend to greater

depths.

The Oneota confining unit can be traced across all of

southeastern Minnesota area using outcrops, cores,

borehole logs, and cuttings. Although discrete beds

with a relatively great density of secondary porosity

features in the Oneota confining unit may be of high

horizontal hydraulic conductivity, there is strong evi-

dence that bulk vertical conductivity is low enough to

provide confinement, analogous in that respect to the

hydrogeologic attributes of other well established con-

fining units in Minnesota. Observations of secondary

porosity features in outcrop and in boreholes showing

diminished abundance, vertical extent and aperture of

fractures with depth in the Oneota is consistent with

confining characteristics in this unit measured with the

flowmeter and head measurements and aquifer tests,

and is consistent with hydraulic conductivity estimates

and open hole interval data obtained from the state

water well database.

6. Implications for groundwater management

One of the most important results of aquifer charac-

terization studies is that they help to identify paradigms

that have become intransigent in the interpretation of

hydrogeologic data. Hydrogeologic investigations of

the Prairie du Chien Group and Jordan Sandstone

have typically calculated transmissivity estimates

using the combined thicknesses of these units and a

lumped hydraulic conductivity value (for example,

Ruhl, 1999), or treated head differences between for-

mations as local perturbations, even though regional

data do not support these interpretations (Fig. 14),

and that head differences show considerable variation

within formations (Fig. 10).

These misinterpretations become most apparent,

and are perhaps most significant in groundwater mod-

eling studies used to calculate travel times to wells,

and the extent of wellhead protection areas. Dye traces

within the Prairie du Chien Group have measured flow

speeds up to 10 km/day (Wheeler, 1993). Clearly,

traditional methods of modeling carbonate aquifers

as equivalent porous media, with uniform hydraulic

conductivity over tens of meters of aquifer thickness,

misrepresents actual groundwater flow within aquifers

that have measured flow speeds this high. Ground-

flow dominated by a few discrete fractures or prefer-

ential flow over a long open hole interval has been

documented in carbonate rock (e.g., Gianniny et al.,

1996; Michalski and Britton, 1997; Morin et al., 1997;

Runkel et al., 2003) and in siliciclastics both at local

(Swanson, 2001) and regional scales (Runkel et al.,

2006—this volume). In addition, travel time estimates

based on uniform hydraulic conductivity over long

open holes has resulted in velocities that are orders

of magnitude too slow (Bradbury et al., 2000; Rayne

et al., 2001). Recognition of fracture flow within the

Shakopee Formation, along with a zone of high per-

meability near the Shakopee Formation–Oneota Dolo-


mite contact is essential for developing conceptual

models of groundwater flow in the Prairie du Chien

Group, and should be included in well head protection

and contaminant transport studies.

Recognition of a high permeability zone near the

Shakopee Formation–Oneota Dolomite contact also has

implications for land use management decisions that

effect water quality. This zone within the Prairie du

Chien Group has been associated sinkholes when close

to the land surface (Tipping et al., 2001) and is present

near the land surface in 3 documented cases of waste-

water treatment lagoon collapses in Minnesota (Alex-

ander and Book, 1984; Jannik et al., 1992; Alexander et

al., 1993).

The high permeability zone within the Prairie du

Chien Group is associated with a paleokarst horizon

and thus closely tied to the hydrologic history of the

rock. Porosity and permeability of Prairie du Chien

Group has evolved under a wide range of hydrologic

and hydrochemical conditions from diagenesis to

present day weathering by meteoric water. While

this hydrologic bimprintQ is recognized in the oil

industry as a means to identify zones of high produc-

tivity, it has received relatively less attention in

groundwater studies. An understanding of the geolog-

ic and hydrologic history of bedrock strata can greatly

improve its hydrogeologic characterization, which

improves management of groundwater resources in

the aquifer.

Acknowledgments

Funding for much of this research was provided by

the Legislative Commission on Minnesota Resources

Environmental Trust Fund, and from the Southeast

Minnesota Water Resources Board, administered by

Bea Hoffmann at Winona State University. The authors

would also like to thank Fred Paillet for his introduc-

tion to EM flowmeter operation and methods of data

analysis.

Appendix A. Summary of hydraulic conductivity

estimates along with additional information on

travel times for the Prairie du Chien Group under

deep and shallow bedrock conditions (Fig. 16)

Data are presented as a generalized stratigraphic

column of the Prairie du Chien Group in southeastern

Minnesota, showing matrix hydrostratigraphic compo-

nents, typical development of secondary porosity, and

hydraulic data compiled for this report. Figure is not to

scale. (Modified from Runkel et al., 2003.) Hydraulic

data for these figures (numbers in parentheses) are from

the following:

(1) Average value of conductivity calculated based

on specific capacity tests in the County Well

Index database.

(2) Fillmore County dye trace study by Wheeler

(1993).

(3) Standard aquifer test of Spring Grove municipal

well #4 (unique number 433257), Houston Coun-

ty (Eder and Associates, 1997).

(4) Discrete interval tests and dye trace studies at

Oronoco Landfill, Olmsted County, by Donahue

and Associates, Inc. (1991) and RMT, Inc.

(1992).

(5) Borehole flowmeter logging and pumping at

wells in: (A) Faribault (unique number 625327),

Rice County. (B) Rochester (unique number

485610), Olmsted County.

(6) Standard aquifer test of a well at Chatfield Fish

and Game Club in Fillmore County (unique num-

ber 227394).

(7) Twenty-six standard aquifer tests conducted in

southeastern Minnesota. Tests of 12 boreholes

located in the seven-county Twin Cities Metro-

politan area are reported by Runkel et al. (1999).

Tests of 14 boreholes outside of the metropolitan

area are from unpublished data compiled by the

U.S. Geological Survey and include the follow-

ing: Rochester Municipal wells 23 (unique num-

ber 220660), 27 (unique number 224212), 28

(unique number 180567), 29 (unique number

161425), 30 (unique number 239761), 31 (unique

number 434041), 32 (unique number 506819),

and 34 (unique number 463536), Rochester pub-

lic schools wells for Ridgeway (unique number

235583), Burr Oak (unique number 220615), and

Golden Hill (unique number 220679), all in

Olmsted County, and Rice County wells at Car-

leton College (unique number 171005), Dundas

(unique number 132294), and St. Olaf College

(no number).

(8) Standard aquifer tests at the New Brighton and

Arden Hills Twin Cities Army Ammunition Plant

site, Ramsey County, by Camp et al. (1991).

(9) Standard aquifer test by the Minnesota Depart-

ment of Natural Resources and Minnesota De-

partment of Health at Plainview in Wabasha

County, Minnesota.

(10) Seventy-two hour aquifer test of Woodbury Well

15 (unique number 676415), Washington County,

by Bonestroo and Associates (2003).


(11) Re-analysis of 1984 Burnsville aquifer test by

SEH (2001).

References

Alexander Jr., E.C., Book, P.R., 1984. Altura, Minnesota lagoon

collapses. In: Beck, B.F. (Ed.), Sinkholes: Their Geology, Engi-

neering, and Environmental Impact. A.A. Balkema, Rotterdam,

pp. 311–318.

Alexander Jr., E.C., Lively, R.S., 1995. Karst—aquifers, caves and

sinkholes. In: Lively, R.S., Balaban, N.H. (Eds.), Text Supplement

to the Geologic Atlas of Fillmore County, Minnesota Geological

Survey County Atlas C-8, Part C, pp. 10–18.

Alexander Jr., E.C., Broberg, J.S., Kehren, A.R., Graziani, M.M.,

Turri, W.L., 1993. Bellechester Minnesota lagoon collapses. In:

Beck, B.F. (Ed.), Applied Karst Geology: Proceedings of the

Fourth Multidisciplinary Conference on Sinkholes and the Engi-

neering and Environmental Impacts of Karst, Panama City, Flor-

ida, 25–27 January. A.A. Balkema, Rotterdam, pp. 63–72.

Barr Engineering, 1996. Dakota County groundwater model, sum-

mary report. 90 pp.

Bonestroo and Associates, 2003. Woodbury Well 15: 72 hour aquifer

test. Report to the Woodbury East Wellfield Technical Advisory

Committee, March 14, 2003, on file at Minnesota Geological

Survey.

Bradbury, K.R., Rothschild, E.R., 1985. A computerized technique

for estimating the hydraulic conductivity of aquifers from specific

capacity data. Ground Water 23 (2), 240–245.

Bradbury, K.R., Rayne, T.W., Muldoon, M.A., 2000. Field verifica-

tion of capture zones for municipal wells at Sturgeon Bay, Wis-

consin. Wisconsin Geological and Natural History Survey,

Informal Final Report to the Wisconsin Department of Natural

Resources, 29 pp.

Camp, Dresser, McKee, 1991. Phase 1A: multi-point source ground-

water remedial investigation, New Brighton/Arden Hills, Minne-

sota. Minnesota Pollution Control Agency Report 90-00007,

vol. 1. 453 pp.

Choquette, P.W., James, N.P., 1988. Introduction. In: James, N.P.,

Choquette, P.W. (Eds.), Paleokarst. Springer-Verlag, New York,

pp. 1–21.

Delin, G.N., Woodward, D.G., 1984. Hydrogeologic setting and the

potentiometric surfaces of regional aquifers in the Hollandale

Embayment, southeastern Minnesota, 1970–1980. U.S. Geologi-

cal Survey Water-Supply Paper 2219 (56 pp.).

Des Moines Water Works, 1995. Aquifer storage and recovery dem-

onstration project. Phase II. Demonstration test results. Informal

report submitted by CH2MHILL to the Des Moines Water Works

Department, unpaginated.

Donahue and Associates, Inc., 1991. Olmsted County dye-trace

investigations [prepared for Olmsted County]—Final report, vol-

ume 1. On file at the Minnesota Pollution Control Agency. 155 pp.

Eaton, T.T., 2002. Fracture heterogeneity and hydrogeology of the

Maquoketa aquitard, southeastern Wisconsin. PhD dissertation,

University of Wisconsin, Madison. 211pp.

Eder and Associates, 1997. 1996 annual monitoring report for Spring

Grove, Minnesota: response action plan. Project 536-01, on file at

the Minnesota Pollution Control Agency.

Frank, E.F., Mylroie, J.E., Troester, J., Alexander, E.C., Carew, J.L.,

1998. Karst development and speleogenesis, Isla de Mona, Puerto

Rico. Journal of Caves and Karst Studies 60 (2), 73–83.

Gianniny, G.L., Muldoon, M.A., Simo, J.A., Bradbury, K.R., 1996.

Correlation of high-permeability zones with stratigraphic features

in the Silurian dolomite, Sturgeon Bay, Wisconsin. Wisconsin

Geological and Natural History Survey Open-File Report 1996-

7. 102 pp., 1 pl.

Hall, C.W., Meinzer, O.E., Fuller, M.L., 1911. Geology and under-

ground waters of southern Minnesota. U.S. Geological Survey

Water-Supply Paper 256 (406 pp.).

Horick, P.J., 1989. Water resources of northeast Iowa. Iowa Depart-

ment of Natural Resources, Geological Survey Bureau Water

Atlas vol. 8, 133 pp.

Jannik, N.O., Alexander Jr., E.C., Landherr, L.J., 1992. The sinkhole

collapse of the Lewiston, Minnesota waste water treatment facility

lagoon. In: Quinlan, J., Stanley, A. (Eds.), Proceedings of the 3rd

National Ground Water Association Conference on Hydrogeol-

ogy, Ecology, Monitoring, and Management of Ground Water in

Karst Terranes. Water Well Journal Publishing Company, Dublin,

OH, pp. 715–724.

Kanivetsky, R., 1978. Hydrogeologic map of Minnesota, bedrock

hydrogeology. Minnesota Geological Survey State Map S-2, 2

pls., scale 1:500,000.

Kanivetsky, R., 1984. Bedrock hydrogeology, pl. 4. In: Balaban,

N.H., Olsen, B.M. (Eds.), Geologic Atlas of Winona County,

Minnesota. Minnesota Geological Survey County Atlas C-2,

8 pls., scale 1:200,000.

Kanivetsky, R., 1988. Bedrock hydrogeology, pl. 5. In: Balaban,

N.H., (Ed.), Geologic Atlas of Olmsted County, Minnesota.

Minnesota Geological Survey County Atlas C-3, 9 pls., scale

1:100,000.

Kanivetsky, R., Cleland, J.M., 1990. Hydrogeology, pl. 5. In: Swan-

son, L., Meyer, G.N. (Eds.), Geologic Atlas of Washington Coun-

ty, Minnesota. Minnesota Geological Survey County Atlas C-5, 7

pls., scales 1:150,000 and 1:250,000.

Kanivetsky, R., Cleland, J.M., 1992. Bedrock hydrogeology, pl. 7. In:

Meyer, G.N., Swanson, L. (Eds.), Geologic Atlas of Ramsey

County, Minnesota. Minnesota Geological Survey County Atlas

C-7, 10 pls., scales 1:48,000 and 1:150,000.

Kanivetsky, R., Palen, B., 1982. Hydrogeology of Scott County, pl. 6.

In: Balaban, N.H., McSwiggen, P.L. (Eds.), Geologic Atlas of

Scott County, Minnesota. Minnesota Geological Survey County

Atlas C-1, 6 pls.

Kerans, C., 1988. Karst-controlled reservoir heterogeneity in Ellen-

burger Group carbonates of west Texas. American Association of

Petroleum Geologists Bulletin 72 (10), 1160–1183.

Michalski, A., Britton, R., 1997. The role of bedding fractures in the

hydrogeology of sedimentary bedrock—evidence from the New-

ark Basin, New Jersey. Ground Water 35 (2), 318–327.

Morin, R.H., Carleton, G.B., Poirier, S., 1997. Fractured-aquifer

hydrogeology from geophysical logs: the Passaic formation,

New Jersey. Ground Water 35 (2), 328–338.

MUGSP: Minnegasco Underground Gas Storage Project, 1980. Un-

published reports by the Minnesota Gas Company on the under-

ground gas storage project. On file at the Minnesota Geological

Survey.

Muldoon, M.A., Simo, J.A., Bradbury, K.R., 2001. Correlation of

hydraulic conductivity with stratigraphy in a fractured-dolomite

aquifer, northeastern Wisconsin, USA. Hydrogeology Journal 9,

570–583.

Mylroie, J.E., Carew, J.L., 2000. Speleogenesis in coastal and oceanic

settings. In: Klimchouk, A.B., Ford, D.C., Palmer, A.N., Drey-

brodt, W. (Eds.), Speleogenesis: Evolution of Karst Aquifers.

National Speleological Society, pp. 226–233.


Paillet, F.L., Lundy, J., Tipping, R.G., Runkel, A.C., Reeves, L.,

Green, J.A., 2000. Hydrogeologic characterization of six sites in

southeastern Minnesota using borehole flowmeters and other

geophysical logs. U.S. Geological Survey Water-Resources Inves-

tigation Report 00-4142, 33 pp.

Palmer, M.V., Palmer, A.N., 1989. Paleokarst of the United States.

In: Bosak, P., Ford, D.C., Glazek, J., Horacek, I. (Eds.), Paleo-

karst: A systematic and Regional Review. Elsevier, Amsterdam,

pp. 337–366.

Rayne, T.W., Bradbury, K.R., Muldoon, M.A., 2001. Delineation of

capture zones for municipal wells in fractured dolomite, Sturgeon

Bay, Wisconsin, USA. Hydrogeology Journal 9, 432–450.

RMT, Inc., 1992. Remedial investigation report: Olmsted County

sanitary landfill, Oronoco Township, Minnesota. On File at the

Minnesota Pollution Control Agency. 185 pp.

Ruhl, J.F., 1999. Hydraulic properties of the Prairie du Chien-Jordan

Aquifer, Shakopee Mdewakanton Sioux Community, southeastern

Minnesota. U.S. Geological Survey Water-Resources Investiga-

tions Report 99-4183, 10 pp.

Runkel, A.C., Kanivetsky, R., Tipping, R.G., 1999. Hydraulic and

hydrostratigraphic characteristics of the Prairie du Chien–Jordan

aquifer system in the Twin Cities metropolitan area, Minnesota:

Informal report submitted to the Minnesota Department of Health,

on file at Minnesota Geological Survey.

Runkel, A.C., Tipping, R.G., Alexander Jr., E.C., Green, J., Mossler,

J., Alexander, S.C., 2003. Hydrogeology of the Paleozoic bedrock

in southeastern Minnesota. Report of Investigations-Minnesota

Geological Survey 61 (105 pp., 2 pls).

Runkel, A.C., Tipping, R.G., Alexander Jr., E.C., Alexander, S.C.,

2006—this volume. Hydrostratigraphic characterization of inter-

granular and secondary porosity in part of the Cambrian sandstone

aquifer system of the cratonic interior of North America: improv-

ing predictability of hydrogeologic properties. Sedimentary Geol-

ogy 184, 281–304. doi:10.1016/j.sedgeo.2005.11.006.

Seaber, P.R., 1988. Hydrostratigraphic units. In: Back, W., Rosen-

shein, J.S., Seaber, P.R. (Eds.), Hydrogeology. Boulder, Colorado,

Geological Society of America, The Geology of North America

vol. O-2, pp. 9–14.

SEH, 2001. Re-analysis of 1984 Burnsville pumping test by Hickok

and Associates, SEH No. ABURNS0006.00 SP, SEH, Chippewa

Falls, Wisconsin, Technical memorandum by Erik Anderson to

Burnsville/Kraemer Well Groundwater Study, on file at Minnesota

Geological Survey.

Setterholm, D.R., Runkel, A.C., Cleland, J.M., Tipping, R.G.,

Mossler, J.M., Kanivetsky, R., Hobbs, H.C., 1991. Geologic

factors affecting the sensitivity of the Prairie du Chien-Jordan

aquifer. Minnesota Geological Survey Open-File Report 91-5.

29 pp., 12 pls.

Sloss, L.L., 1963. Sequences in the cratonic interior of North Amer-

ica. Geological Society of America Bulletin 74 (2), 93–113.

Smith, G.L., Byers, C.W., Dott Jr., R.H., 1996. Sequence stratigra-

phy of the Prairie du Chien Group, Lower Ordovician, Midcon-

tinent, USA. In: Witske, B.J., Ludvigson, G.A., Day, J. (Eds.),

Paleozoic Sequence Stratigraphy: Views from the North American

Craton, Geological Society of America Special Paper vol. 306,

pp. 23–33.

Swanson, S.K., 2001, Hydrogeologic controls on springflow near

Madison, Wisconsin. PhD dissertation, University of Wisconsin,

Madison. 436 pp.

Tipping, R.G., Green, J.A., Alexander, E.C. Jr., 2001. Karst features,

pl. 5. In: Runkel, A.C., project manager, Geologic Atlas of

Wabasha County, Minnesota. Minnesota Geological Survey

County Atlas C-14, pt. A, 7 pls., scale 1:100,000.

Wall, D.B., Regan, C.P., 1994. Water quality and sensitivity of the

Prairie du Chien-Jordan aquifer in western Winona County, Min-

nesota. Unpublished report by the Minnesota Pollution Control

Agency, Division of Water Quality, St. Paul, Minn., 65 pp.

Wheeler, B.J., 1993. Groundwater tracing in the Duschee Creek karst

basin in southeast Minnesota. Plan B Thesis paper submitted to

the Geology and Geophysics Department, University of Minne-

sota. 72 pp.

Wilson, J.L., Medlock, P.L., Fritz, R.D., Canter, K.L., Geesaman,

R.G., 1992. A review of Cambro-Ordovician breccias in North

America. In: Candelaria, M.P., Reed, C.L. (Eds.), Paleokarst,

Karst Related Diagenesis, and Reservoir Development: Examples

from Ordovician–Devonian Age Strata of West Texas and the

Mid-Continent, SEPM Publication vol. 92-33, pp. 19–29.

Zhang, H., Kanivetsky, R., 1996. Bedrock hydrogeology. pl. 6. In:

Falteisek, J., project supervisor, Geologic Atlas of Fillmore Coun-

ty, Minnesota. Minnesota Department of Natural Resources, Di-

vision of Waters County Atlas C-8, pt. B, 4 pls., scale 1:100,000.

doi:10.1016/j.sedgeo.2005.11.006

Evidence for hydraulic heterogeneity and anisotropy in the mostly carbonate Prairie du Chien Group,...

Documents

Transcript of Evidence for hydraulic heterogeneity and anisotropy in the mostly carbonate Prairie du Chien Group,...