LAKE PIRACY ACROSS CONTINENTAL WATERSHED DIVIDE
IN SUMMIT COUNTY, OHIO
Eckstein Yoram
Department of Geology, Kent State University, Kent, Ohio, 44242
yeckstei@geology.kent.edu
Matyjasik Barbara
Utah Geological Survey, Salt Lake City, Utah, 84114
Matyjasik Marek
Department of Geosciences, Weber State University, Ogden, Utah 84408 mmatyjasik@weber.edu
ABSTRACT
Crystal Lake of Bath Township, Summit County, Ohio, located within the Cuyahoga
River drainage basin, only a few hundred meters west-north-west of the continental surface-
and ground water-divide between the St. Lawrence River and the Ohio-Mississippi River
drainage basins, experienced substantial drop in water table during the years 1991-94.
Hydrological mass balance for the lake indicated that at the end of September 1991 the water
level in the lake was about 20 inches (50.8 cm) lower than the projected lake level calculated
on the basis of the actual local seasonal hydro-meteorological conditions. Hence, it was
assumed that the decline of the lake level must be attributed to a decline in ground water
level. Drop of the water table in the lake, in fact coincided with a significant decline in the
ground water table recorded in a number of observation wells penetrating the shallow glacial
aquifer in the region.
Mapping of the ground water table disclosed an expanding cone of depression
centered at the wells of the Montrose Well Field of Summit County. During the ten years,
from 1985 through 1994, combined annual production rates increased consistently, from a
low of 65 million gallons in 1985, to a peak production of 246 million gallons in 1993. The
increase in production rates was accompanied by consistent trend of ground water level
decline recorded in all the observation wells of the area. Expansion of the cone of depression
to the west resulted in dislocation of the ground water divide between the Cuyahoga River
1
(discharging to Lake Erie) and Tuscarawas River (a tributary of the Ohio River), "pirating"
Crystal Lake from the Great Lakes drainage into the Ohio River Drainage Basin.
History of the drawdowns generated by the increase in ground water production from
the Montrose Well Field was simulated using Modflow numerical model, using average
aquifer hydraulic properties obtained from interpretation of pumping tests carried out in the
two production wells. The model was calibrated and verified using ground water levels
recorded through the years in the observation wells. Results of the modeling and the
dislocation of the water drainage divide are presented using Geographic Information System.
The model, along with the ground water level surveys, suggests that Crystal Lake converted
during the ten years (1985-1994) from an effluent to an influent surface water body. Also, the
model suggests that the excessive drawdowns in the two Montrose production wells resulted
in drop of the water level below the semi-confining boundary of the till covering the kame
gravel aquifer, changing the aquifer conditions from a semi-confined to unconfined.
Key Words: continental watershed divide, ground water diversion.
INTRODUCTION
The term "piracy" is commonly used in fluvial geomorphology, in reference to
natural process resulting in diversion of a stream from one drainage basin to another.
Thornbury (1969) states that "stream piracy may take place by abstraction, headward
erosion, lateral planation, and subterranean diversion" (p. 148). The case of subterranean
diversion is probably most notably represented by the portion of the Danube waters diverted
within the Black Forest through a system of bedrock fractures across the continental
drainage divide into the adjacent Rhein River Channel. In the following we present a case
where a lake located close to a continental drainage divide between the Great Lakes and the
Mississippi River basins was temporarily "pirated" across the divide by a man-made
subterranean diversion.
Aside for illustrating a rare hydrogeological phenomenon, the case has the potential to
contribute to the controversial issues associated with cross-drainage-divide export of fresh
water. The case presents an additional dimension, particularly to the recent conflict, and still
on-going discussion of the City of Akron plan to export fresh water from Lake Rockwell
Reservoir on the Cuyahoga River (draining into Lake Erie) across the continental drainage
divide to the western outskirts of the city located within the Ohio River Basin.
Crystal Lake is a small, elongated reservoir, 2500 ft long, located in Bath Township,
Summit County, Ohio, within the Cuyahoga River drainage basin. The lake is, however, only
2
a several hundred meters WNW from the continental surface- and the concomitant ground
water-divide between the St. Lawrence River and the Ohio-Mississippi River drainage basins
(Fig. 1). The lake, during the years 1991-94, experienced substantial drop in water table. The
map of water table prior to the decline and location of monitoring wells are presented in
Figure 2. The map indicates the watershed divide line between the two drainage basins.
Crystal Lake is recharged from groundwater flowing in the NW direction, away from the
divide line. Hydrological mass balance for the lake (Carlson, 1992) indicated that at the end
of September 1991 the water level in the lake was over 0.5 m (about 20 inches) lower than the
projected lake level calculated on the basis of the actual local seasonal hydro-meteorological
conditions. Hence, it was assumed that the decline of the lake level must be attributed to a
decline in ground water level. Drop of the water table in the lake, in fact coincided with a
significant decline in the ground water table recorded in a number of observation wells
penetrating the shallow glacial aquifer in the region.
GEOLOGY AND HYDROGEOLOGY OF THE STUDY AREA
The shallow subsurface of the study area consists of two major sedimentary
formations: Quaternary poorly stratified heterogeneous, uncosolidated glacial and fluvio-
glacial deposits blanketing the Paleozoic sedimentary bedrock (Fig. 3). The young
unconsolidated sediments fill a wide buried valley, deeply carved in the surface of the
consolidated bedrock beneath the valley of Yellow Creek located just NW of the study area.
The bedrock floor of the buried valley is at the depth from about 15 to over 90 m (50 to 300
feet) below the ground surface. The valley has been filled up with sediments by a number of
glacial ingressions and recessions (in the form of glacial till) and by fluvio-glacial deposits
(Fig. 3). Glacial deposits in the area around Crystal Lake formed during the Wisconsinan
stage of glaciation, represented in the neighborhood of the lake by Mogadore Till, Hayesville
Till, and Hiram Till. The area NE of the lake is covered by alluvial deposits consisting mainly
of silt and silty sand, while the surface sediments south, west and north of the lake belong to
kames and kame terraces (White, G. W., 1984). Kames and kame terraces consist mainly of
well-sorted and stratified sand and gravel interlayered with tills of the ground moraine.
The approximately uppermost 30 m (100 feet) of the geologic profile near Crystal
Lake consist of clay, sand and some gravel. The top of the sandy aquifer is located at a depth
of 7.25 - 10.21 m (23.8–33.5 feet) below sediments characterized by a relatively low hydraulic
conductivity, acting as a semi-confining layer.
Bedrock formations underneath the glacial deposits in the study area consist
3
primarily of a thick sequence of shale and siltstones belonging to the Mississippian Cuyahoga
Formation, topped occasionally by erosional remnants of conglomeratic sandstone belonging
to the Sharon Member of the Pennsylvanian Pottsville Formation.
GROUND WATER WITHDRAWALS
Mapping of the ground water table disclosed an expanding cone of depression
centered around the Summit County production wells of the Montrose Well Field. During
the ten years, from 1985 through 1994, combined annual production rates increased
consistently, from a low of approximately 246,000 m3 (65 million gallons) in 1985, to a peak
production of over 931,000 m3 (246 million gallons) in 1993 (Fig. 4). The increase in
production rates was accompanied by consistent ground water level decline recorded in all
the observation wells of the area.
In October 1996 operation of the Montrose Well Field was taken over by the City of
Akron Public Utilities Bureau, and the production from the wells was discontinued shortly
thereafter. The wells were taken out of service in April and sealed in May 1997.
The elevation of Crystal Lake is about 304.8 m (1000 feet) at the crest of the dam (NW
corner of the lake). Flood level is at an elevation of 306.32 m (1005 feet). Groundwater level
observations conducted in the network of the observation wells (Fig. 2) indicated that the
elevation of groundwater in the area around the lake varied in 1981 from 305.1 m (1001 feet)
to 305.7 m (1003 feet). Ground water divide between the Cuyahoga River Basin and the Ohio
River Basin is shown to run in a winding line from SW to NE, a short distance to the SE of
the highway I-77 and Medina Road intersection, leaving Crystal Lake within the Cuyahoga
River basin. Groundwater flows from the divide in the WNW direction, across Crystal Lake
at the levels higher than 304.8 m (1000 feet) at the crest of the dam, thus recharging the lake.
Between 1981 and 1994 the elevations of water in observation wells declined significantly in
the area around the lake. Figure 5 presents summary of the decline of water table in all
observation wells between 1981 and 1994. The highest decline has been documented in wells
located just east of highway I-77, around location of Montrose Well Field supplying water for
the City of Akron. Figure 6 demonstrates an example of detailed observations from well 7
located 548.6 m (1800 feet) SW of the Montrose Field Well #1.
Changes in groundwater table elevation correlate perfectly with the increase of
production rates from the water wells of the Montrose Well Field (Figs. 5, 6, 7). Water from
the lake has been declining since at least 1990 when production from the Montrose Well Field
exceeded the rate of about 552,600 m3 (146 million gallons) per year. No ground water level
4
data in the observation wells is available for the period beyond 1995. However, the lake level
has visibly recovered following termination of ground water production from Montrose Well
Field at the end of 1996.
MODELING OF GROUNDWATER FLOW
GW Modeler, an ArcView 3.2 extension (University of Wyoming, 1999) was used to
assist pre-processing and post-processing for the U.S. G.S. Modflow groundwater model to
simulate groundwater flow in the study area. Simulated maps of elevation of water table
were prepared within an area 5100 m x 4200 m, discretized into square elements 75 m x 75 m.
Simulations were conducted for three different stress conditions. The map presented on
Figure 7 shows simulated potentiometric surface representing conditions of unstressed
groundwater flow corresponding to water elevations measured in 1981. The map in Figure 8
represents simulated conditions in 1990 when the Montrose field produced 553,095 m3 (146
million gallons per year or 278 gpm). The map in Figure 9 represents simulated conditions in
1993, when the production rate peaked at 931,110 m3/year (246 million gallons per year). In
the simulations we assigned 63.6% of the discharge rate to the production well W1, and
36.4% of the total discharge to the production well #2, which is consistent with the
production potential of the two Montrose wells. The aquifer hydraulic parameters,
coefficient of transmissibility of 794 m2/day (63,970 gpd/ft) and storage coefficient of 10-5
were calculated based on the constant-rate aquifer pumping-tests in the two Montrose wells.
TEMPORARY DISLOCATION OF WATERSHED DIVIDE
Westward expansion of the cone of depression around the Montrose Well Field,
located in the Tuscarawas River drainage basin, resulted in dislocation of the ground water
drainage divide with the neighboring Cuyahoga River drainage basin. Crystal Lake under
normal, unstressed conditions is recharged from groundwater flowing NW, away from the
divide between the two watersheds. Under stressed conditions associated with the excessive
since 1991 withdrawal of ground water in the Montrose Well Field, the flow direction is
reversed, and water from Crystal Lake flows to the SE becoming part of the flow system in
the Tuscarawas River basin. Based on the measured elevations of ground water table in all
the observation wells and results of the modeling, the ground water divide was moved to the
NW by about 1.6 km (1 mile). As the Cuyahoga River discharges into Lake Erie, and the
Tuscarawas River is a tributary of the Ohio River, the dislocated divide is the major
continental divide between the St. Lawrence River Drainage Basin and the Ohio-Mississippi
5
River Drainage Basin.
CONCLUSIONS
The significant drop of the water table in Crystal Lake during the early 90's coincided
with the documented drop of the ground water table in a number of observation wells. Field
mapping of the ground water table suggested that the cause for the drop in the lake level was
associated with the consistently increasing since 1985 production rates from the Montrose
Well Field two water wells. Ground water simulations, using USGS Modflow coupled with
ESRI ArcView GIS software, indicate that expansion of the cone of depression around the
Montrose Well Field production wells resulted in dislocation of the continental ground water
divide. As a result, Crystal Lake was "pirated" from the Cuyahoga River - Lake Erie - St.
Lawrence River Basin to the Tuscarawas - Ohio River - Mississippi River Basin.
Furthermore, the simulations indicate that Crystal Lake, prior to the 1985-1994 excessive
production from the Montrose Well Field was an effluent lake, recharged by ground water
flow from the ground water divide to the NW. During the years 1985-1994 Crystal Lake
turned into an influent lake, loosing water to the ground water flow in the SE direction.
REFERENCES
Carlson, R. E., 1992. A Management Plan for Crystal Lake; unpublished consulting report.
McDonald, H.R.and A.W. Harbaugh, 1984, A Modular Three-Dimensional Finite-
Difference Ground-Water Flow Model; U.S. Geological Survey].
Thornbury, W.D., 1969. Principles of Geomorphology; Second Edition; John Wiley &
Sons, Inc.
White, G. W., 1984. Glacial Geology of Summit County, Ohio. Documents Of State of Ohio,
Department Of Natural Resources, Report No. 123.
University of Wyoming, 1999. GW Modeler for ArcView 3.1, ESRI, ArcView, GIS.
6
LIST OF FIGURES
FIGURE 1: Location of the study area on the continental divide between the St. Lawrence
River-Great Lakes and the Ohio-Mississippi drainage basins.
FIGURE 2: Water table map for 1981 prior to the decline of water in Crystal Lake, Summit
County, Ohio.
FIGURE 3: Geologic map of the study area.
FIGURE 4: Montrose Well Field annual production rates (million of gallons).
FIGURE 5: Decline of water table in the Crystal Lake area in observation wells between1981
and 1994.
FIGURE 6: Observed water table in monitoring well #7 located 548.6 m (1800 feet) SW of the
Montrose field main pumping well.
FIGURE 7: Simulated water table map corresponding to 1981, prior to the decline of water
table (water table contour interval 1 foot) .
FIGURE 8: Simulated water table map corresponding to 1990 with the discharge rate of 146
million gallons per year (water table contour interval 1 foot).
FIGURE 9: Simulated water table map corresponding to 1993 with the discharge rate of 246
million gallons per year (water table contour interval 1 foot).
7
Cleveland
St. Lawrence-Great Lakes
Drainage Basin
CUYAHOGA
WATERSHED
STUDY
AREA TUSCARAWAS
WATERSHED
Ohio-Mississippi Rivers
Drainage Basin
Miles
Kilometers
FIGURE 1: Location of the study area on the continental divide between the St. Lawrence
River-Great Lakes and the Ohio-Mississippi drainage basins.
8
2
D
A
O
R EXPLANATION
994
ek 1 observation well
1 re 3
5 C W1 pumping well
99
6 direction of
99
E
C groundwater flow
4 rysta
K
H
A 7
A
L 10
99 l
M
E
T
L
A
T La watershed divide
O
8 S
Y
ke W2 prior to pumping
99
W R
N w C
o W1
l
l 9
e
Y 99
8 9
6 7
Stoney0Hill
00
1
5
D
R Miles
R
01
O 0 Sho
A
D 1 C O P L E Y cal 0 12.5 25
K
og
C
2 O Run
0 1 R
0 H1
1 0 T 00
0 O 4
3 R 4107’30’’
o
25 50
8138’04’’
o Km
FIGURE 2: Water table map for 1981 prior to the decline of water in Crystal Lake, Summit
County, Ohio.
9
D
A
O
R Whag EXPLANATION
Wk
Wk ek Alluvium.
e
r
al al
C Silt and silty
d
H
E
K
A
Cr
and kame
Wk Kames and sand.
Gravel
A
M
L
yst t
L
E al
Whag Hayesville Till.
A
T T
O S La
W
N w
Y
R
C
ke Silty till, very thin.
o
l
l
Whag
e
Y Wmoh Mogadore Till. till.
Coarse sandy
Stoney Hill
Wk
D
R Whag R
Sho
O
A
D C O P L E Y cal 0 12.5 Miles
og
K
C
O Ru
R n
Wmoh H
T
O
R al 4107’30
o 25 Kilometers
o
FIGURE 3: Geologic map of the study area.
10
300
250
Annual Production (MGal/Yr)
200
150
100
50
0
1985 1986 1988 1990 1991 1992 1993 1994 1995 1996
Years
FIGURE 4: Montrose Well Field annual production rates (million of gallons).
11
FIGURE 5: Decline of water table in the Crystal Lake area in observation wells between
1981 and 1994.
12
1010
1005
Elevation of Water Table (feet)
1000
995
y = -0.0058x + 1007.6
2
R = 0.8071
990
985
980
1981 1984 1987 1990 1993
975
0 1000 2000 3000 4000 5000 6000
Days (From January 1981)
FIGURE 6: Observed water table in monitoring well #7 located 548.6 m (1800 feet) SW of the
Montrose field main pumping well.
13
2
D
A
O
R
LEGEND
ek 1 observation well
1 e
r 3
C W1 pumping well
3 direction of
9 5
9 9 98 C
E
9 groundwater flow
4 rystal
K
H
A 9 7 A
L 6 10 5 3
M 9
9 0 0 0 watershed divide
E L
A La 0
1 0 0
T 9 T
ke W2 1 1
99 prior to pumping
O S
Y
W R
N w C
o 01 W1 watershed divide
l
l 10 8 9 after pumping
e
Y 03
10 6 7
Stoney Hill
05
10
07 5
10
D
R
R
O Sho Miles
A cal
D C O P L 008
E og 0 12.5 25
1 K
C
O
R
Run
H
T
O
R
0 25 50
o
o
4107’ Kilometers
8138’
FIGURE 7: Simulated water table map corresponding to 1981, prior to the decline of water
table (water table contour interval 1 foot) .
14
2
LEGEND
EXPLANATION
RO AD
1 observation well
997
ek
3 W1 pumping w ell
Cre
1
direction of
9969 5 groundwater flow
993
9 3
994
Cr y 99 10
L A KE
watershed div ide
99 5
HA ET WN
s
4 ta l L prior to pumping
M O
9 96
L
ak
C RYST A
e W2 watershed div ide
7
Yello w
99
after pumping
W1
8 9
8 6
99 7
Stoney Hill
99 9
5
Miles
1000
RD
0 12.5 25
ROAD
Sho
C O P L E Y cal
og
1
ROTHROCK
100 Run 0 25 50
02
10
o
o
4107’30’’
8138’04’’
Kilometers
FIGURE 8: Simulated water table map corresponding to 1990 with the discharge rate of 146
million gallons per year (water table contour interval 1 foot).
15
2
LEGEND
EXPLANATION
ROAD
1 observation well
993
ek
3
1 W1 pumping well
Cre
2
99 direction of
1 groundwater flow
Cr 99 0 9
LAKE
y st 99 98 10
HAMETOWN
al watershed divide
993
4 La prior to pumping
CRYSTAL
ke W2
watershed divide
Yellow
W1 after pumping
8 9
6 7
Stoney Hill
994
5
RD
ROAD
95 Sho
C O P L 9
E Y cal
o g
6
ROTHROCK
99
Ru n
9 97
998
1000
9 99
o
41 07’30’’
o
81 38’04’’
FIGURE 9: Simulated water table map corresponding to 1993 with the discharge rate of 246
million gallons per year (water table contour interval 1 foot).
16