Quaternary Research 57, 244?252 (2002)
doi:10.1006/qres.2001.2311, available online at http://www.idealibrary.com on
n
a
0
Com
used t
ing its
about
bathym
the Mc
dicate
in vol
times.
corres
with g
Ojibw
areal e
Sea. D
of wat
lets w
Laure
tigated
1600 a
the Ty
about
Key
Lak
forme
tinent
ern m
Agass
times
Manit
Dakot
et al.,
is bel
mate
Hoste
1 To
Earth a
Institut
e
l
0033-58
Copyrig
All righChanges in the Bathymetry and Volu
between 9200 and 77
David W. Leverington,1 Jason D. Ma
Department of Geological Sciences, University of Manitob
Received May 16, 2
puter reconstructions of the bathymetry of the lake were
o quantify variations in the size and form of Lake Agassiz dur-
final two phases (the Nipigon and Ojibway phases), between
9200 and 7700 14C yr B.P. (ca. 10,300?8400 cal yr B.P.). New
Lake
the d
and b
Thor
and metric models for four Nipigon Phase stages (corresponding to
Cauleyville, Hillsboro, Burnside, and The Pas strandlines) in-
that Lake Agassiz ranged between about 19,200 and 4600 km3
ume and 254,000 and 151,000 km2 in areal extent at those
A bathymetric model of the last (Ponton) stage of the lake,
ponding to the period in which Lake Agassiz was combined
lacial Lake Ojbway to the east, shows that Lake Agassiz?
ay was about 163,000 km3 in volume and 841,000 km2 in
xtent prior to the final release of lake waters into the Tyrrell
uring the Nipigon Phase, a number of catastrophic releases
er from Lake Agassiz occurred as more northerly (lower) out-
ere made available by the retreating southern margin of the
ntide Ice Sheet; we estimate that each of the four newly inves-
Nipigon Phase releases involved water volumes of between
nd 2300 km3. The final release of Lake Agassiz waters into
rrell Sea at about 7700 14C yr B.P. is estimated to have been
163,000 km3 in volume. C? 2002 University of Washington.
Words: Lake Agassiz; Lake Ojibway; bathymetry; volume.
INTRODUCTION
e Agassiz was the largest of the proglacial lakes that
d in North America during the last deglaciation, as con-
al drainage was impounded against the retreating south-
argin of the Laurentide Ice Sheet (LIS). The size of Lake
iz varied considerably during its 4000-yr history, at various
covering parts of the Canadian provinces of Saskatchewan,
oba, Ontario, and Que?bec, and the U.S. states of North
a, South Dakota, and Minnesota (e.g., Elson, 1967; Teller
1983; Smith and Fisher, 1993) (Fig. 1). Lake Agassiz
ieved to have played a role in determining regional cli-
during the last deglaciation (Teller, 1987; Hu et al., 1997;
tler et al., 2000). Releases of large volumes of water from
whom correspondence should be addressed. Present address: Center for
nd Planetary Studies, National Air and Space Museum, Smithsonian
ion, Washington, DC 20560-0315. E-mail: leveringtond@nasm.si.edu.
Deep
et al.,
et al.,
In t
maps
of the
that t
place
8400
extend
metric
determ
As
Amer
age w
about
began
Hudso
water
Bluem
served
segme
sidera
ing ic
closin
and T
of Lak
to the
into th
(Vinc
1994;
At
umes
lower
24494/02 $35.00
ht C? 2002 by the University of Washington.
ts of reproduction in any form reserved.me of Glacial Lake Agassiz
00 14C yr B.P.
n, and James T. Teller
, Winnipeg, Manitoba, Canada, R3T 2N2
01
Agassiz at various times in its history, associated with
glaciation of new and lower outlets, impacted the rivers
asins that received the lake?s overflow (e.g., Teller and
eifson, 1983; Lewis et al., 1994; Fisher and Smith, 1994)
ay have influenced ocean circulation and North Atlantic
Water production (e.g., Broecker et al., 1989; Licciardi
1999; Barber et al., 1999; Leverington et al., 2000; Clark
2001).
his paper, area and volume calculations and bathymetric
are given for six stages of glacial Lake Agassiz (four stages
Nipigon Phase and two stages of the Ojibway Phase)
ogether span the final 1500 yr of the lake, which took
between about 9200 and 7700 14C yr B.P. (ca 10,300?
cal yr B.P., using Stuiver and Reimer, 1993). This research
s the work of Leverington et al. (2000), in which bathy-
models for seven earlier stages of Lake Agassiz were
ined.
BACKGROUND
the LIS retreated during the last deglaciation of North
ica, meltwaters collected in proglacial lakes where drain-
as impeded by ice (e.g., Teller, 1987). Lake Agassiz formed
11,700 14C yr B.P. when the Red River Lobe of the LIS
its final retreat northward across the divide between the
n Bay and Mississippi River drainage basins, ponding
against the southern margin of the LIS (Elson, 1967;
le, 1974; Clayton, 1983). Based on investigations of pre-
offshore lake sediments, outlet channels, and strandline
nts, the size of Lake Agassiz is known to have varied con-
bly during its history, driven by isostatic rebound, chang-
e-sheet configurations, and the consequent opening and
g of drainage routes (e.g., Elson, 1967; Zoltai, 1967; Teller
horleifson, 1983; Smith and Fisher, 1993). The last stages
e Agassiz involved confluence with glacial Lake Ojibway
east (Elson, 1967) and the ultimate drainage of these waters
e Tyrrell Sea (Hudson Bay) by about 7700 14C yr B.P.
ent and Hardy, 1979; Dredge and Cowan, 1989; Veillette,
Barber et al., 1999).
numerous times in the history of Lake Agassiz, large vol-
of water were released when lake levels dropped after
outlets were deglaciated. We believe that most releases
involved
single ca
outlet deg
ing multi
catastrop
day, exte
river vall
(Teller an
The hi
phases: L
(Fenton e
Phase ex
the Herm
southern
et al., 19
was thro
Elson, 19
The M
from abo
a signific
following
Lakes th
and Thor
Moorhea
to the No
FIG. 1.
Lake Agass
Kaministik
Superior ba
glacial Lake Ojibway at about the Ponton stage of Lake Agassiz.rcross strandline (Thorleifson, 1996).BATHYMETRY AND VOLUME CHANGE, LAKE AGASSIZ 245
relatively rapid outlet deglaciation and occurred as
tastrophic events. However, releases involving slower
laciation (e.g., at more complex outlet systems involv-
ple channels) may have occurred as a series of smaller
hic events that occurred over a longer time frame. To-
nsive boulder and cobble lags are preserved in modern
eys that once acted as Lake Agassiz overflow channels
d Thorleifson, 1983).
story of Lake Agassiz can be divided into five main
ockhart, Moorhead, Emerson, Nipigon, and Ojibway
t al., 1983; Teller and Thorleifson, 1983). The Lockhart
tended from about 11,700 to 11,000 14C yr B.P., and
an beaches outline the extent of Lake Agassiz in the
end of the basin at this time (Upham, 1895; Fenton
83; Hobbs, 1983). Drainage during the Lockhart Phase
ugh the southern outlet (Minnesota River Valley; e.g.,
67; Fenton et al., 1983) (Fig. 1).
oorhead Phase was a low-water phase that extended
ut 11,000 to 10,100 14C yr B.P. and was initiated by
ant drop in lake level from the Herman strandlines
deglaciation of an eastern overflow route to the Great
rough the Kaministikwia route (Clayton, 1983; Teller
leifson, 1983; Thorleifson, 1996) (Fig. 1). During the
d Phase, isostatic rebound gradually raised lake level
The Emerson Phase extended from about 10,100 to 9400
14C yr B.P. and was initiated after drainage to the east was
blocked by ice readvances. Most overflow during the Emerson
Phase was through the northwestern outlet (Fig. 1), although
changing ice configuration, isostatic rebound, and outlet erosion
resulted in several short episodes of southward overflow (Teller,
2001). Recognized strandlines of the Emerson Phase include the
Norcross, Tintah, and Upper Campbell.
The next phase of Lake Agassiz, the Nipigon Phase, extended
from about 9400 to 8200 14C yr B.P. and was initiated by the
deglaciation of the Kaiashk outlet, which allowed overflow into
the Nipigon and Superior basins (Teller and Thorleifson, 1983).
This led to the final abandonment of the southern and north-
western outlets. During the Nipigon Phase, the waters of Lake
Agassiz gradually shifted northward with the retreat of the LIS,
and the elevation of the lake surface incrementally dropped each
time new (lower) outlets were opened and as outlet channels
were deepened by erosion (Elson, 1967). Commencing with the
Lower Campbell level, a series of about a dozen strandlines
developed during the Nipigon Phase; each formed as a trans-
gressive beach (Teller, 2001), and each was abandoned when
lower outlets were opened (Leverett, 1932; Johnston, 1946;
Elson, 1967; Fenton et al., 1983).
The Ojibway Phase began when Lake Agassiz combined withMap showing the total geographic coverage of Lake Agassiz over its 4000-yr histor
iz outlets mentioned in the text are labeled as follows: 1, northwestern outlet (Cle
wia route; 4, Kaiashk system; 5, Kopka system; 6, Pikitigushi system; and 7, Kinoje?
sins are identified (?Nip.? and ?Sup.?).y (modified after Teller et al., 1983, and Dredge and Cowan, 1989).
arwater spillway); 2, southern outlet (Minnesota River Valley); 3,
vis outlet to the Ottawa River Valley. Locations of the Nipigon and
246 N
Lake
the Ot
to nor
1979)
possib
believ
gin at
Hardy
et al.,
In
mode
and (
to the
Camp
here b
mode
the M
two st
level
Our in
ogniz
and w
throug
For
ing a
elevat
from
bound
et al.,
metry
water
of mo
isosta
volum
(Leve
Cal
tated
with t
gin of
in tim
Nipig
ice m
(2) an
shifte
ice m
by 1?
define
volum
and p
and a
ice m
of the
b
e
h
m
ti
i
a
d
a
s
u
u
b
l
e
6
o
s
e
,
re
o
e
e
a
ic
j
p
T
l
s
iLEVERINGTON, MANN, A
Agassiz?Ojibway drained through the Kinoje?vis outlet to
tawa River Valley until the LIS no longer provided a barrier
thward outflow into the Tyrrell Sea (Vincent and Hardy,
. Catastrophic northward drainage into the Tyrrell Sea,
ly initiated in the region north of modern James Bay, is
ed to have followed the collapse of the confining ice mar-
about 7700 14C yr B.P. (Barber et al., 1999; Craig, 1969;
, 1977; Vincent and Hardy, 1979; Veillette, 1994; Barber
1999).
METHODOLOGY
a previous study (Leverington et al., 2000), bathymetric
ls of seven stages of the Lockhart, Moorhead, Emerson,
earliest) Nipigon phases (including those corresponding
Herman, Norcross, Tintah, Upper Campbell, and Lower
bell strandlines) were generated. The research presented
uilds on the previous study by generating bathymetric
ls of four stages of the Nipigon phase (corresponding to
cCauleyville, Hillsboro, Burnside, and The Pas levels) and
ages of the Ojibway phase (corresponding to the Kinoje?vis
and possible Fidler level) (Teller and Thorleifson, 1983).
formal ?Kinoje?vis stage? roughly corresponds to the rec-
ed Ponton strandline (e.g., Teller and Thorleifson, 1983)
as used to define Lake Agassiz?Ojibway when it drained
h the Kinoje?vis outlet (Fig. 1).
each stage, lake bathymetry was calculated by subtract-
stage-specific rebound surface from a database of modern
ions. A rebound surface is defined by values interpolated
isobase data and describes the relative glacio-isostatic re-
that has occurred over a region since a given time (Mann
1999). The geometry of a rebound surface matches the geo-
of the corresponding (and now differentially rebounded)
plane. The subtraction of a rebound surface from a database
dern topography adjusts topography for the effects of
tic rebound, providing a basis for the estimation of lake
e and area, as well as the positions of paleo-shorelines
rington et al., 2000, in press).
culations of volumes and areas of Lake Agassiz necessi-
the definition of the lake?s northern (ice-contact) margins
he LIS. Because the configuration of the ice-contact mar-
Lake Agassiz is not precisely known for specific points
e, calculations of lake parameters were made for each
on stage using three different ice margins: (1) a ?favored?
argin (based on previously published margins; see below);
ice margin with the same form as the favored margin but
d by 1? of latitude (about 111 km) to the north; and (3) an
argin with the same form as the favored margin but shifted
of latitude to the south. The use of three ice margins to
each stage helps to emphasize the uncertainty with which
e and area calculations for Lake Agassiz are carried out
Data
Th
searc
Team
tion
(Has
conta
ing L
lakes
lake
datab
grid
for H
Rebo
Re
angu
from
curv
(199
lated
gins.
bathy
Pont
isoba
imat
1979
fine
(Kin
Ice M
Th
Phas
drain
The
the O
Dyke
extra
Bay.
Vo
Agas
8400
given
in whrovides what we feel are realistic ranges for lake volumes
reas. For comparison purposes, calculations using offset
argins were additionally made for the seven earlier stages
lake investigated previously (Leverington et al., 2000).
by 1?
Table
for th
fromD TELLER
DATA COLLECTION
ase of Modern Elevations
primary source of modern elevations used in this re-
was Version 1.0 of the GLOBE database (GLOBE Task
, 1999). The GLOBE database is a global digital eleva-
odel with a latitude?longitude grid spacing of 30 arc sec
ngs and Dunbar, 1999). The GLOBE database does not
n bathymetric data for Hudson Bay nor for lakes, includ-
ke Winnipeg. While supplementary bathymetric data for
was not considered necessary for this exercise (average
epths in the region are less than 15 m), the ETOPO5
se (National Geophysical Data Center, 1988; 5 arc min
pacing) was used as a source of modern bathymetric data
dson Bay.
nd Surfaces
ound surfaces were spatially interpolated using the ?tri-
ated irregular network? algorithm (see Mann et al., 1999)
point data taken from isobases and strandline rebound
s plotted by Teller and Thorleifson (1983) and Thorleifson
); where necessary, rebound data were linearly extrapo-
to the north when working with north-shifted ice mar-
The form of the rebound curve used to reconstruct lake
metry at the Kinoje?vis stage was based on that for the
n stage (Thorleifson, 1996), with the trends of associated
es modified in the east so that the 300-m isobase approx-
ly intersects the Kinoje?vis outlet (see Vincent and Hardy,
their fig. 3-I). The number of isobase points used to de-
bound surfaces ranged from 270 (McCauleyville) to 437
je?vis).
argins
ice margins of the four investigated stages of the Nipigon
were modified from those of Thorleifson (1996), with
ge allowed through appropriate eastern outlets (Fig. 1).
e margin used to help define the two investigated stages of
ibway Phase was based on the margins of Dredge (1983),
and Prest (1987), and Vincent and Hardy (1979), with
olation from these margins in the region west of James
he faces of all ice margins were treated as vertical.
AREA, VOLUME, AND BATHYMETRY
umes, areas, and maximum depths for 13 stages of Lake
iz in the period 11,700 to 7700 14C yr B.P. (ca. 13,600?
cal yr B.P.) are given in Table 1. Values in this table are
for ?favored? ice margins, as well as for ice configurations
ch these preferred ice margins were shifted north and south
of latitude. Bathymetric maps for the first seven stages in
1 are given in Leverington et al. (2000). Bathymetric maps
e last six stages in this table, corresponding to the period
9200 to 7700 14C yr B.P. (ca. 10,300?8400 cal yr B.P.),
Area
Lake phase
Lockhart
Moorhead
Emerson
Nipigon
Ojibway
a
?North
Laurentide
b Combi
c Possibl
are show
below is
the ?favo
The vo
(e.g., He
modern L
although
as large a
1984). Vo
Campbel
the total
Herdend
the begin
c
a
S
b
u
s
h
8
a
e
d
v
t
r
e
u
d
h
u
s
j
d
,0
s
)
p
c
iBATHYMETRY AND VOLUME CHANGE
TABLE 1
, Volume, and Maximum Depth Values for 13 Stages
of Lake Agassiz
14C yr Ice Area Volume Max.
Lake stage B.P. margina (km2) (km3) depth (m)
Herman 10,900 North 214,000 24,700 321
Favored 134,000 10,900 231
South 66,000 3500 157
Early Moorhead 10,700 North 193,000 23,700 338
Favored 117,000 10,800 247
South 61,000 3700 173
Late Moorhead 10,300 North 271,000 36,500 363
Favored 185,000 19,700 258
South 120,000 10,100 227
Norcross 10,100 North 254,000 27,300 275
Favored 166,000 13,300 243
South 94,000 5400 180
Tintah 9900 North 276,000 30,100 274
Favored 184,000 15,700 233
South 114,000 7100 184
Upper Campbell 9400 North 382,000 39,500 281
Favored 263,000 22,700 233
South 156,000 10,700 190
Lower Campbell 9300 North 355,000 33,600 262
Favored 240,000 19,100 214
South 142,000 9000 173
McCauleyville 9200 North 329,000 29,700 250
Favored 219,000 16,400 199
South 128,000 7700 160
Hillsboro 8900 North 368,000 34,500 260
Favored 254,000 19,200 199
South 163,000 10,500 160
Burnside 8500 North 309,000 22,300 210
Favored 202,000 10,300 147
South 115,000 4900 131
The Pas 8200 North 238,000 12,100 203
Favored 151,000 4600 96
South 75,000 1800 63
Kinoje?visb 7700 Favored 841,000 163,000 773
Fidlerc <7700 Favored 408,000 49,900 421
? and ?South? refer to 1? latitude shifts from ?favored? positions of
Ice Sheet margins.
ned Lake Agassiz?Ojibway.
e final stage.
n in Figs. 2a?2f. Except where noted, the discussion
given with respect to lake parameters estimated using
red? ice margins.
lumes of most of the smaller stages shown in Table 1
rman and Burnside) are roughly comparable to that of
ake Superior (about 12,200 km3; Herdendorf, 1984),
the smallest investigated stage (The Pas) is only about
s modern Lake Michigan (about 4900 km3; Herdendorf,
with gla
consider
of the LI
and lowe
Ojibway
area of a
This vol
estimate
Veillette
200,000
mates. T
were abo
vided at
The b
ble final
Cochran
and Har
stage is b
a morain
below th
Fidler le
or any o
and Tho
last stag
have occ
as a few
the Kino
the Coc
(Agassiz
This wo
fore the
possible
Thorleif
stage are
present u
the Kino
deposits
stable Fi
terpreted
lake stag
calculate
and 408
Maxim
for inve
(Table 1
imum de
m, respe
spite the
Ojibway
cally. Thlumes of the lake when at the Upper Campbell, Lower
l, and Hillsboro strandlines are roughly comparable to
volume of the modern Great Lakes (about 22,700 km3;
orf, 1984). The size of Lake Agassiz increased greatly at
ning of the Ojibway Phase, when Lake Agassiz merged
773 m. T
Agassiz
see also L
For th
Burnside, LAKE AGASSIZ 247
ial Lake Ojibway to the east; this combined lake grew
bly during the Ojibway Phase, since northward shifts
southern margin during this phase did not expose new
r drainage routes. The Kinoje?vis stage of Lake Agassiz?
(Fig. 2e) dwarfed all other investigated stages, with an
out 841,000 km2 and a volume of about 163,000 km3.
me is less than, but roughly consistent with, previous
of the final size of Lake Agassiz?Ojibway produced by
(1994) (230,000 km3) and Barber et al. (1999) (about
km3), who used more coarse techniques for their esti-
e volumes of the Agassiz and Ojibway sides of this lake
ut 102,000 km3 and 61,000 km3, respectively, when di-
0?W longitude.
thymetric model of the Fidler stage (Fig. 2f), a possi-
lake stage, is given here only for regions west of the
II ice advance (this advance is described by Vincent
y, 1979, and Veillette, 1994). The level of the Fidler
ased on Klassen?s (1983) identification of a beach along
e in northern Manitoba, which today lies about 80 m
e Ponton beach. As reconstructed in this research, the
el is too low to have been associated with the Kinoje?vis
her known outlet in the region (see also Klassen, 1983,
leifson, 1996). If the Fidler beach does represent the
of Lake Agassiz, the final drainage of the lake would
rred in two steps, perhaps separated in time by as much
ecades. In this reconstruction, complete drainage from
je?vis stage must have been prevented by ice related to
rane II advance, causing some waters in the western
) side of the lake basin to be temporarily held back.
ld have allowed the Fidler strandline time to form be-
final drainage into the Tyrrell Sea. Alternatively, it is
that the trends of the isobases we used (cf. Teller and
on, 1983; Thorleifson, 1996) to reconstruct the Fidler
not valid in the eastern region and that, contrary to our
nderstanding, the Fidler strandline in fact did intersect
e?vis outlet. Finally, it is also possible that the localized
in northern Manitoba currently cited as evidence for a
ler standline (Klassen, 1983, p. 111) have been misin-
and that the Kinoje?vis stage was instead the final stable
e. The Fidler stage as reconstructed in this research is
d to have had a volume and area of about 49,900 km3
00 km2, respectively.
um lake depths ranged between about 200 and 260 m
tigated stages from the Herman to the Hillsboro
. The subsequent Burnside and The Pas stages had max-
ths that were markedly less than this range (147 and 96
tively), resulting in relatively small lake volumes de-
r considerable surface areas. After lakes Agassiz and
combined, maximum lake depths increased dramati-
e maximum depth of the Kinoje?vis stage was about
he deepest waters of all 13 investigated stages of Lake
were generally located along the LIS margin (Fig. 2;
everington et al., 2000).
e McCauleyville (Fig. 2a), Hillsboro (Fig. 2b), and
(Fig. 2c) stages, overflow from Lake Agassiz was
248
FIG
(f) sho
menti
exten
of Jam
2e?2f
eastw
chan
and
(Fig.
Pikit
Kino
Rive
and
have
Dred. 2. Bathymetric models for six stages of Lake Agassiz during the Nipigon and Ojibway phases, at (a) 9200, (b) 8900, (c) 8500, (d) 8200, (e) 7700, and
rtly after 7700 14C yr B.P. Maximum depth is 773 m (Lake Ojibway portion of the Kinoje?vis stage). Grid cells are 2? by 2?. The locations of two eastern outlets
oned in the text are labeled as follows: 1, the Kaiashk system; 2, the Kopka system. These two outlet positions are based on their approximate westernmost
ts along the continental divide. The outline of modern-day Hudson Bay and James Bay is given in 2e and 2f. In 2e and 2f, the ice-margin segment southwest
es Bay speculatively treats the ice in this region as a western component of the Cochrane II surge. The scales for 2a?2d are different from the scales for
.
ard into the Nipigon basin through progressively lower
nels of the Kopka system (#5, Fig. 1; #2, Fig. 2), (cf. Teller
Thorleifson, 1983). For the The Pas stage of Lake Agassiz
2d), drainage was into the Nipigon basin through the
igushi system (east of the area shown in Fig. 2d). For the
je?vis stage (Fig. 2e), overflow was south into the Ottawa
dam along the eastern side of Lake Agassiz (e.g., Thorleifson,
1996).
The lake parameters presented in Table 1 highlight the wide
range in lake size that is possible for individual lake stages,
depending on the selected position and configuration of the
northern ice margin of the lake. Typically, a 1? northward shift ofLEVERINGTON, MANN, ANr Valley through the Kinoje?vis outlet (e.g., Vincent
Hardy, 1979). For the Fidler stage (Fig. 2f), drainage may
been beneath the stagnant LIS into the Tyrrell Sea (e.g.,
ge, 1983; Klassen, 1983), or around the remaining ice
ice fro
areas t
config
60% oD TELLERm the favored ice margins results in an expansion of lake
o about 150% of the areas estimated using the favored ice
urations; a 1? southward shift results in a decrease to about
f the areas estimated using the favored ice configurations.
49BATHYMETRY AND VOLUME CHANGE, LAKE AGASSIZ 2FIG. 2? Continuned
250 N
The s
200%
confi
RE
Us
were
of the
(Tabl
new o
down
cause
(sout
slowl
Lake p
Lock
Emer
Nipig
Ojibw
a
?No
Lauren
b He
c Vo
level.
a
d
a
i
e
d
e
t
t
a
o
n
l
a
s
d
r
m
a
s
s
r
d
b
h
e
e
e
a
n
r
p
d
s
te
,
l
g
in
le
s
j
eLEVERINGTON, MANN, A
ame shifts typically result in lake volumes that are about
and 45% of the volumes estimated using the favored ice
gurations, respectively.
LEASES OF WATER FROM LAKE AGASSIZ DURING
THE NIPIGON AND OJIBWAY PHASES
ing the bathymetric databases discussed above, estimates
made of the volumes of water released at the terminations
four Nipigon stages and two Ojibway stages investigated
e 2). During retreat of the LIS during the Nipigon Phase,
utlets were periodically opened, resulting in a rapid draw
of the lake?s surface and a reduction in lake volume. Be-
subsequent isostatic rebound led to a rise in lake level
h of the isobase passing through the active outlet), waters
y deepened again and Lake Agassiz expanded until the LIS
TABLE 2
Volumes of Water Released at Terminations of Selected
Stages of Lake Agassiz
14C yr Ice Volume Released
hase Lake stage B.P. margina (km3)
hart Hermanb 10,900 North 17000
Favored 9500
South 3400
son Norcross 10,100 North 9300 to 11600
Favored 7500 to 9300
South 3200 to 3800
Tintah 9900 North 9700
Favored 5900
South 3100
Upper Campbellb 9400 North 3700 to 10500
Favored 2500 to 7000
South 1500 to 2500
on Lower Campbell 9300 North 5400
Favored 3700
South 3100
McCauleyville 9200 North 3200
Favored 2100
South 1200
Hillsboro 8900 North 2500
Favored 1600
South 1100
Burnside 8500 North 3600
Favored 2300
South 1300
The Pas 8200 North 2600
Favored 1600
South 800
ay Kinoje?vis 7700 Favored 163,000 (or 113,100c)
Fidler <7700 Favored (49,900c)
rth? and ?South? refer to 1 ? latitude shifts from ?favored? positions of
retre
imum
stran
fell e
any d
of th
base
at th
outle
temp
to L
the f
lowi
Thor
McC
side
At
lease
occu
volu
it is
for a
relea
occu
woul
side,
whic
vanc
of th
of th
W
natio
given
given
histo
Cam
cusse
relea
tima
30 m
and S
Ca
Nipi
these
term
9300
(Tab
relea
Kino
the rtide Ice Sheet margins.
rman and Upper Campbell after Leverington et al. (2000).
lumes associated with hypothetical two-stage release from Kinoje?vis
from
113,1
level o
TableD TELLER
ted and uncovered a lower outlet (Teller, 2001). The max-
extent of each transgression is marked by a Lake Agassiz
line (Teller, 2001). However, the level to which the lake
ch time a new lower outlet was opened was not marked by
scernable sediment or morphology. Thus, our calculation
drawdown of a specific Nipigon lake level was necessarily
on the difference between the elevation of the strandline
start of the draw down and the lowest elevation in the
system through which that draw down occurred. An at-
was made to select channel floor elevations that related
ke Agassiz outflow and not to subsequent incision. For
ur Nipigon stages investigated in this research, the fol-
g elevation drops were determined mainly from Teller and
eifson (1983, their fig. 2) and topographic maps: (1)
uleyville stage, 10 m; (2) Hillsboro stage, 7m; (3) Burn-
tage, 12 m; and (4) The Pas stage, 12 m.
the end of the Ojibway Phase, Lake Agassiz?Ojibway re-
its waters into the Tyrrell Sea. It is possible that this release
red in a single rapid and catastrophic event, with a total
e of about 163,000 km3 (Table 2). As discussed above,
lso possible that western (Agassiz) waters were retained
hort time behind stagnant ice of the LIS, after the initial
e of water from the level of the Kinoje?vis outlet. If this
red, the release at the termination of the Kinoje?vis stage
have involved the entire volume of the eastern (Ojibway)
ut only part (upper 103 m) of the western (Agassiz) side,
was held back by remaining ice of the Cochrane II ad-
(Fig. 2e). In this scenario, the subsequent release at the end
Fidler stage would have involved the remaining volume
western part of the basin, depicted in Fig. 2f.
ter volumes estimated to have been released at the termi-
s of the four Nipigon stages and two Ojibway stages are
in Table 2. For comparison purposes, calculations are also
for the terminations of selected stages from the earlier
y of Lake Agassiz (Herman, Norcross, Tintah, and Upper
bell). The Herman and Upper Campbell releases are dis-
in Leverington et al. (2000). The Norcross and Tintah
es through the northwestern outlet (#1, Fig. 1) were es-
d by assuming lake-level drawdowns of 40 to 52 m, and
respectively, at the terminations of these stages (see Fisher
mith, 1994, and Teller, 2001).
culated releases at the terminations of the investigated
on stages range between 1600 and 2300 km3 (Table 2);
volumes are less than those calculated for the pre-Ojibway
ations of the Herman (9500 km3), Norcross (7500 to
km3), and Upper Campbell stages (2500 to 7000 km3)
2; see also Leverington et al., 2000). The northward
e from Lake Agassiz?Ojibway at the termination of the
e?vis stage is estimated to have been about 163,000 km3. If
lease took place in two stages, the initial northward release
the level of the Kinoje?vis stage would have been about
00 km3 (Table 2), and the subsequent release from the
f the Fidler stage would have been about 49,900 km3 (see
s 1 and 2). The final release(s) of Lake Agassiz?Ojibway
,were into
Ocean.
Bathym
Nipigon
ital datab
bathyme
ify drain
magnitud
Nipigon
with a vo
smallest
stage, wi
The volu
was south
and 841,
Calcul
Ocean th
Nipigon
lets beca
northwar
the Tyrre
7700 14C
Waters fr
Hudson B
The in
rerouting
lation an
et al., 19
Barber et
Rind et a
addition
either ov
baseline
culation a
mates of
vide a q
Agassiz m
which it
from Lak
by Teller
The auth
comments
Manitoba G
Engineerin
Barber, D.
Kerwin,
n
h
fl
.
l
.
i
n
7
A
l
A
.
,
.
e
a
.
o
a
t
s
h
,
r
,
C
l
SBATHYMETRY AND VOLUME CHANGE
the Tyrrell Sea, and ultimately into the North Atlantic
SUMMARY AND CONCLUSIONS
etric models were generated for six stages of the
and Ojibway phases of Lake Agassiz by adjusting a dig-
ase of modern elevations for isostatic rebound. These
tric models were used to determine lake extents, to ver-
age routings, and to calculate lake volumes and the
es of abrupt drainage events. The largest of the four
Phase lake stages investigated was the Hillsboro stage,
lume of 19,200 km3 and an area of 254,000 km2. The
investigated stage of the Nipigon Phase was the The Pas
th a volume of 4600 km3 and an area of 151,000 km2.
me and area of Lake Agassiz? Ojibway when drainage
through the Kinoje?vis outlet were about 163,000 km3
000 km2, respectively.
ated releases from Lake Agassiz into the North Atlantic
rough the Nipigon and Superior basins during the
Phase (ca. 9400 ? 8000 14C yr B.P.), when lower out-
me ice free, range between 1600 and 3700 km3. The
d release of water from Lake Agassiz? Ojibway into
ll Sea at the termination of the Kinoje?vis stage at about
yr B.P. is estimated to have been about 163,300 km3.
om the final drainage of this lake were routed through
ay and Hudson Strait into the North Atlantic Ocean.
fluence that Lake Agassiz outbursts, and associated
s of baseline overflow, may have had on ocean circu-
d climate continues to be investigated (e.g., Broecker
89; Fanning and Weaver, 1997; Licciardi et al., 1999;
al., 1999; Leverington et al., 2000; Clark et al., 2001;
l., 2001; Teller et al., in press). Many suggest that the
of large volumes of water into the North Atlantic Ocean,
er short periods or as sustained (but smaller) changes in
runoff, may have substantially altered thermohaline cir-
nd production of North Atlantic Deep Water. The esti-
Lake Agassiz volumes and releases presented here pro-
uantitative basis for evaluating the impact that Lake
ay have had on North America and on the oceans into
flowed. The potential impacts of freshwater outbursts
e Agassiz on ocean circulation and climate are explored
et al. (in press).
ACKNOWLEDGMENTS
ors thank Peter Clark and an anonymous reviewer for their helpful
on this paper. This research was supported by a University of
raduate Fellowship to D. W. Leverington, and a Natural Sciences and
g Research Council (Canada) Research Grant to J. T. Teller.
and Gag
catastrop
Bluemle, J.
Geologic
Broecker, W
and Wol
during th
Clark, P. U
and Telle
the last g
Clayton, L
In ? Glac
Geologic
Craig, B. G
region. I
pp. 63 ? 7
Dredge, L.
its relatio
(J. T. Te
Canada S
Dredge, L.
tario shie
Ed.), pp
America
Dyke, A. S
the Laur
Elson, J. A.
(W. J. M
Fanning, A
fluences
Paleocea
Fenton, M.
stratigrap
? Glacial
logical A
Fisher, T. G
maximum
Science R
GLOBE T
(GLOBE
Adminis
Hardy, L. (1
que?be?coi
Quaterna
Hastings, D
Elevation
to Geop
spheric A
Herdendorf
Characte
Program
Hobbs, H.
and early
(J. T. Te
Canada,
Hostetler, S
(2000). SREFERENCES
C., Dyke, A., Hillaire-Marcel, C., Jennings, A. E., Andrews, J. T.,
M. W., Bilodeau, G., McNeely, R., Southon, J., Morehead, M. D.,
ice sheet
Hu, F. S., W
glacial La
tion. GeoLAKE AGASSIZ 251
on, J.-M. (1999). Forcing of the cold event of 8200 years ago by
ic drainage of Laurentide lakes. Nature 400, 344? 348.
P. (1974). Early history of Lake Agassiz in southwest North Dakota.
al Society of America Bulletin 85, 811? 814.
. S., Kennett, J., Flower, B., Teller, J. T., Trumbore, S., Bonani, G.,
i, W. (1989). Routing of meltwater from the Laurentide Ice Sheet
e Younger Dryas cold episode. Nature 341, 318? 321.
, Marshall, S. J., Clarke, G. K. C., Hostetler, S. W., Licciardi, J. M.,
r, J. T. (2001). Freshwater forcing of abrupt climate change during
aciation. Science 293, 283? 287.
(1983). Chronology of Lake Agassiz drainage to Lake Superior.
al Lake Agassiz? (J. T. Teller and L. Clayton, Eds.), pp. 291 ? 307,
al Association of Canada, Special Paper 26.
. (1969). Late-glacial and postglacial history of the Hudson Bay
? Earth Science Symposium on Hudson Bay ? (P. J. Hood, Ed.),
, Geological Survey of Canada, Paper 68 ? 53.
. (1983). Character and development of northern Lake Agassiz and
n to Keewatin and Hudsonian ice regimes. In ? Glacial Lake Agassiz?
ler and L. Clayton, Eds.), pp. 117 ? 131. Geological Association of
pecial Paper 26.
., and Cowan, W. R. (1989). Lithostratigraphic record on the On-
ld. In ? Quaternary Geology of Canada and Greenland ? (R. J. Fulton,
214 ? 235. Geological Survey of Canada, The Geology of North
Vol. K-1.
, and Prest, V. K. (1987). Late Wisconsinan and Holocene history of
ntide Ice Sheet. Ge?ographie physique et Quaternaire 41, 237? 263.
(1967). Geology of glacial lake Agassiz. In ? Life, Land, and Water ?
yer-Oakes, Ed.), pp. 36 ? 95. Univ. of Manitoba Press, Winnipeg.
F., and Weaver, A. J. (1997). Temporal-geographical meltwater in-
n the North Atlantic conveyor: implications for the Younger Dryas.
nography 12, 307? 320.
M., Moran, S. R., Teller, J. T., and Clayton, L. (1983). Quaternary
hy and history in the southern part of the Lake Agassiz basin. In
Lake Agassiz? (J. T. Teller and L. Clayton, Eds.), pp. 49 ? 74. Geo-
ssociation of Canada, Special Paper 26.
., and Smith, D. G. (1994). Glacial Lake Agassiz: Its northwest
extent and outlet in Saskatchewan (Emerson Phase). Quaternary
eviews 13, 845? 858.
sk Team. (1999). ? The Global One-Kilometer Base Elevation
) Elevation Model: Version 1.0.? National Oceanic and Atmospheric
ration, National Geophysical Data Center, Boulder, CO.
977). La de?glaciation et les e?pisodes lacustre et marin sur le versant
des basses terres de la baie de James. Ge?ographie Physique et
ire 31, 261? 273.
. A., and Dunbar, P. K. (1999). ? Global Land One-Kilometer Base
(GLOBE) Digital Elevation Model, Documentation: Vol. 1.0.? Key
ysical Records Documentation 34, National Oceanic and Atmo-
dministration, Boulder, CO.
C. E. (1984). ? Inventory of the Morphometric and Limnologic
istics of the Large Lakes of the World.? Ohio State Univ. Sea Grant
Technical Bulletin OHSU-TB-17.
. (1983). Drainage relationship of glacial Lake Aitkin and Upham
Lake Agassiz in northeastern Minnesota. In ? Glacial Lake Agassiz?
ler and L. Clayton, Eds.), pp. 245 ? 259. Geological Association of
pecial Paper 26.
. W., Bartlein, P. J., Clark, P. U., Small, E. E., and Solomon, A. M.
imulated interactions of proglacial Lake Agassiz with the Laurentide
11,000 years ago. Nature 405, 334? 337.
right, H. E., Jr., Ito, E., and Lease, K. (1997). Climatic effects of
ke Agassiz in the midwestern United States during the last deglacia-
logy 25, 207? 210.
252 LEVERINGTON, MANN, AND TELLER
Johnston, W. A. (1946). ? Glacial Lake Agassiz with Special Reference to
the Mode of Deformation of the Beaches.? Geological Survey of Canada,
Bulletin 7.
Klassen, R. W. (1983). Lake Agassiz and the late glacial history of Northern
Manitoba. In ? Glacial Lake Agassiz? (J. T. Teller and L. Clayton, Eds.),
pp. 97 ? 115. Geological Association of Canada, Special Paper 26.
Leverett, F. (1932). ? Quaternary geology of Minnesota and parts of adjacent
states.? U.S. Geological Survey, Professional Paper 161.
Leverington, D. W., Mann, J. D., and Teller, J. T. (2000). Changes in
the bathymetry and volume of glacial Lake Agassiz between 11,000 and
9300 14C yr B.P. Quaternary Research 54, 174 ? 181.
Leverington, D. W., Teller, J. T., Mann, J. D. (in press). A GIS method for
reconstruction of late Quaternary landscapes from isobase data and modern
topography. Computers and Geosciences.
Lewis, C. F. M., Moore, T. C., Jr., Rea, D. K., Dettman, D. L., Smith, A. J., and
Mayer, L. A. (1994). Lakes of the Huron basin: Their record of runoff from
the Laurentide Ice Sheet. Quaternary Science Reviews 13, 891 ? 922.
Licciardi, J. M., Teller, J. T., and Clark, P. U. (1999). Freshwater routing by the
Laurentide Ice Sheet during the last deglaciation. In ? Mechanisms of Global
Climate Change at Millennial Time Scales ? (P. U. Clark, R. S. Webb, and L.
D. Keigwin, Eds.), pp. 171 ? 202. Am. Geophys. Union, Monograph, v. 112.
Mann, J. D., Leverington, D. W., Rayburn, J., and Teller, J. T. (1999). The volume
and
71?
Natio
me
Na
Rind,
(20
mo
Res
Smith
out
Stuiver, M., and Reimer, P. J. (1993). Extended 14C data base and revised CALIB
3.0 14C age calibration program. Radiocarbon 35, 215? 230.
Teller, J. T. (1987). Proglacial lakes and the southern margin of the Laurentide
Ice Sheet. In ? North America and adjacent oceans during the last deglacia-
tion? (W. F. Ruddiman and H. E. Wright, Eds.), pp. 39 ? 69. Decade of North
American Geology, Vol. K-3, Geological Survey of America.
Teller, J. T. (2001). Formation of large beaches in an area of rapid differen-
tial isostatic rebound: The three-outlet control of Lake Agassiz. Quaternary
Science Reviews 20, 1649? 1659.
Teller, J. T., and Thorleifson, L. H. (1983). The Lake Agassiz? Lake Superior
connection. In ? Glacial Lake Agassiz? (J. T. Teller and L. Clayton, Eds.),
pp. 261 ? 290. Geological Association of Canada, Special Paper 26.
Teller, J. T., Thorleifson, L. H., Dredge, L. A., Hobbs, H. C., and Schreiner,
B. T. (1983). Maximum extent and major features of Lake Agassiz. In ? Glacial
Lake Agassiz? (J. T. Teller and L. Clayton, Eds.), pp. 43 ? 45. Geological
Association of Canada, Special Paper 26.
Teller, J. T., Leverington, D. W., and Mann, J. D. (in press). Freshwater outbursts
to the oceans from glacial Lake Agassiz and climate change during the last
deglaciation. Quaternary Science Reviews.
Thorleifson, L. H. (1996). Review of Lake Agassiz history. In ? Sedimentology,
Geomorphology, and History of the Central Lake Agassiz Basin? (J. T. Teller,
L. H. Thorleifson, G. Matile, and W. C. Brisbin, Eds.), pp. 55 ? 84. Geological
Asso
epaleobathymetry of glacial Lake Agassiz. Journal of Paleolimnology 22,
80.
nal Geophysical Data Center. (1988). ETOPO 5 database: Data announce-
nt 88-MGG-02. In ? Digital Relief of the Surface of the Earth.? NOAA,
tional Geophysical Data Center, Boulder, CO.
D., deMenocal, P., Russell, G., Sheth, S., Collins, D., Schmidt, G., Teller, J.
01). Effects of glacial meltwater in the GISS coupled atmostphere-ocean
del: Part I. North Atlantic Deep Water response. Journal of Geophysical
earch 106, 27335 ? 27354.
, D. G., and Fisher, T. G. (1993). Glacial Lake Agassiz: The northwestern
let and paleoflood. Geology 21, 9? 12.
Mee
Upham
grap
Veillet
and
Vincen
Ojib
Zoltai,
Wat
Winciation of Canada Field Trip Guidebook for GAC/MAC Joint Annual
ting.
, W. (1895). ? The Glacial Lake Agassiz.? U.S. Geological Survey, Mono-
h 25.
te, J. J. (1994). Evolution and paleohydrology and glacial lakes Barlow
Ojibway. Quaternary Science Reviews 13, 945? 971.
t, J.-S., and Hardy, L. (1979). ? The Evolution of Glacial Lakes Barlow and
way, Quebec and Ontario.? Geological Survey of Canada, Bulletin 316.
S. C. (1967). Eastern outlets of Lake Agassiz. In ? Life, Land, and
r? (W. J. Mayer-Oakes, Ed.), pp. 107 ? 120. Univ. of Manitoba Press,
nipeg.