Download Nicholson et al., 1997 - University of Minnesota Duluth

Rift-wide correlation of 1.1 Ga Midcontinent rift
system basalts: implications for multiple mantle
sources during rift development
Suzanne W. Nicholson, Steven 6. Shirey, Klaus J. Schulz, and John C. Green
Abstract: Magmatism that accompanied the 1.1 Ga Midcontinent rift system (MRS) is attributed to the upwelling and
decompression melting of a mantle plume beneath North America. Five distinctive flood-basalt compositions are recognized
in the rift-related basalt succession along the south shore of western Lake Superior, based on stratigraphically correlated
major element, trace element, and Nd isotopic analyses. These distinctive compositions can be correlated with equivalent
basalt types in comparable stratigraphic positions in other MRS localities around western Lake Superior. Four of these
compositions are also recognized at Mamainse Point more than 200 km away in eastern Lake Superior. These regionally
correlative basalt compositions provide the basis for determining the sequential contribution of various mantle sources to
flood-basalt magmatism during rift development, extending a model originally developed for eastern Lake Superior. In
this refined model, the earliest basalts were derived from small degrees of partial melting at great depth of an enriched,
ocean-island-type plume mantle source ( E , , ( ,
value of about 0), followed by magmas representing melts from this
plume source and interaction with another mantle source, most likely continental lithospheric mantle ( E , , ( ,
< 0). The
relative contribution of this second mantle source diminished with time as larger degree partial melts of the plume became
value of about 0). Towards the end of magmatism,
the dominant source for the voluminous younger basalts (E,,(,
= 0 to +3).
mixtures of melts from the plume and a depleted asthenospheric mantle source became dominant ( E , , ( ,
,,,
,,,
,,,
,,,
RCsumC : Le magmatisme contemporain de la mise en place du systkme du rift Midcontinental, il y a 1,l Ga, est attribuk
au soulkvement et a une fusion par baisse de pression d'un panache mantellique localis6 sous I'AmCrique du Nord. Le
long de la rive sud dans la partie occidentale du lac SupCrieur, cinq compositions distinctes ont kt6 reconnues parmi les
coulees de basalte de la succession basaltique associke au rift, sur le fondement de correlations stratigraphiques Ctablies
pour les elements majeurs et en traces, et les analyses isotopiques du Nd. Ces compositions distinctes peuvent etre
mises en correlation avec les types de basalte equivalents qui occupent des niveaux stratigraphiques comparables dans
les autres localitks du systkme du rift Midcontinental voisines du secteur occidentale du lac Superieur. On a identifie
Cgalement quatre de ces compositions a Mamainse Point, c'est-a-dire a plus de 200 km de distance, dans la region
orientale du lac SupCrieur. Ces correlations regionales fondkes sur les compositions des basaltes fournissent les elements
de base pour determiner la contribution sequentielle des diffkrentes sources mantelliques au magmatisme des coulees de
basalte durant le dkveloppement du rift, perfectionnant le modkle developpk initialement pour la region orientale du lac
SupCrieur. Dans ce modkle raffine, les premiers basaltes formes derivaient de faibles degres de fusion partielle a grande
profondeur d'un panache mantellique enrichi de type oceanique-insulaire (E,,(,
proche de 0), suivis de magmas
composes des liquides de fusion partielle du panache et d'une interaction avec une autre source mantellique, plus
probablement issue du manteau lithospherique continental (E,,(,
< 0). La contribution relative de cette deuxikme
source mantellique diminuait avec le temps, au fur et a mesure qu'augmentait la proportion de fusion partielle du
panache qui finalement est devenue la source dominante alimentant les volumineuses coulees de basalte plus jeunes
proche de 0). Les derniers stades du magmatisme ont kt6 domines par les melanges de magmas derives du
= 0 a + 3).
panache et d'une source appauvrie du manteau asthCnosphCrique (E,,(,
[Traduit par la redaction]
,,,
,,,
,,,
,,,
Introduction
The
Ga Midcontinent rift system (MRS)
at least
rocks (Cannon 1992) which were
lo6 km3 of
Received April 19, 1996. Accepted November 28, 1996.
S.W. Nicholson' and K. J. Schulz. United States Geological
Survey, National Center, MS 954, Reston, VA 20192,
U.S.A.
S.B. Shirey. Department of Terrestrial Magnetism, Carnegie
Institution of Washington, 5241 Broad Branch Road NW,
Washington, DC 20015, U.S.A.
J.C. Green. Department of Geology, University of
Minnesota, Duluth, Duluth, MN 558 12, U.S.A.
I
Corres~ondingauthor (e-mail: s w n i c h-@ u s ~ s . ~ o v ~ .
w
Can. J . Earth Sci. 34: 504-520 (1997)
u
u
z
emplaced within about 2 2 Ma (Davis and Green 1997), and
provides an example of a structurally intact continental rift
that nearly formed an oceanic basin. The last decade has been
a period of intense data gathering about all aspects of the rift.
This effort coincided with the publication of rapidly changing
concepts regarding the general development of continental
rifts, flood-basalt provinces, and mantle plumes (Cox 1980;
R.S. White et a1. 1987; McKenzie and Bickle 1988; Richards
et al. 1989; R.S. White and McKenzie 1989, 1995; Campbell and Griffiths 1990; Arndt and Christensen 1992; R.S.
White 1992; Turner and Hawkesworth 1995). Models developed from these concepts suggest that progressive melting of
a mantle plume head can produce voluminous flood basalts
distributed over a region up to 1 x l o 6 km2 or more (e.g.,
R.S. White and McKenzie 1989: R.S. White 1997). The
I
-
-
1997 NRC Canada
Nicholson et al.
sequential compositions of erupted flood basalts derived
from melting at this scale provide a basis for evaluating the
nature and dynamics of mantle sources during rifting.
Most of the MRS is buried by water or Paleozoic strata,
but recent deep crustal seismic reflection, refraction, and
gravity profiles (Behrendt et al. 1988; Cannon et al. 1989;
Hinze et al. 1990; Hutchinson et al. 1990; McGinnis and
Mudrey 1991; Allen et al. 1995) show that it extends to as
much as 30 km depth and is segmented into a number of
subbasins. High-precision U - Pb zircon dating (Paces and
Miller 1993; Davis and Green 1997; Zartman et al. 1997)
has constrained the absolute time frame of magmatism and
allows for resolution of individual magmatic pulses and hiatuses (Miller et al. 1995~).Geochemical and isotopic analysis (Brannon 1984; Dosso 1984; Green 1986; Sutcliffe 1987;
Paces and Bell 1989; Nicholson and Shirey 1990; Klewin
and Berg 1991; Lightfoot et al. 1991; Shirey et al. 1994) tied
to detailed stratigraphy provide the basis for geochemical
correlations that allow the nature and interaction of magma
sources during development of the Midcontinent rift to be
explored.
Abundant geologic, geochronologic , geochemical, and
isotopic data for rift basalts on the south shore of western
Lake Superior are now available. The relationships observed
in this stratigraphic section can be compared with other MRS
exposed sections, in particular the well-documented section
at Mamainse Point. The correlation of geographically isolated sequences provides a more complete view of the
succession of magma types during rift development than can
be observed at any one locality. The regional correlation of
flood-basalt compositions extends the model for the succession of magmatic sources derived for eastern Lake Superior
(Klewin and Berg 1991; Shirey et al. 1994) and leads to a
general model for the interaction of mantle and crustal
sources such as enriched continental lithospheric mantle of
probable Archean age, enriched plume-type mantle, and
depleted convecting mantle.
of volcanic flows was reduced near the margins of the rift
basin and in the vicinity of these crustal blocks, At about
1060 Ma, a period of compression and thrust faulting
inverted the central graben (Cannon et al. 1 9 9 3 ~ ) .
Rift-related volcanic and sedimentary rocks, called the
Keweenawan Supergroup, are present at the surface only in
the Lake Superior region, primarily along the margins of the
rift (Fig. I). Southeast and southwest of Lake Superior, riftrelated rocks are covered by Paleozoic sedimentary rocks;
samples are available only from drill core. Most of the
exposed outcrop of MRS volcanic and sedimentary rocks
occur in western Lake Superior (Fig. I). Volcanism was
dominated by subaerial tholeiitic basalt flows with lesser
intermediate and rhyolitic rocks.
Historically, a regionally recognized shift in magnetic
polarity from reversed to normal has provided a ready correlation tool (e.g., Halls and Pesonen 1982). Recent highprecision U-Pb zircon ages have constrained the rift's
duration of magmatism from about 1108 to about 1086 Ma
and placed the regionally recognized magnetic polarity shift
at about 1100 Ma (Miller et al. 1995a; Davis and Green
1997). Recognition of other magnetic reversals in the section
and the acquisition of zircon ages have clarified the overall
time frame for rift development and magmatism, but also
have raised new questions. Tools such as stratigraphically
controlled geochemical and isotopic data have become
increasingly important in identifying regional correlations.
A reasonably complete stratigraphic section of MRS volcanic and sedimentary rocks about 17 km thick is preserved
in northern Michigan and Wisconsin on the southern margin
of the lake (Fig. 2). Although the outcrop is discontinuous,
detailed mapping and sampling provide the basis for regional
correlation of geochemical and isotopic data with welldocumented stratigraphy, constrained by magnetic polarity
and precise U - Pb geochronology .
Geological setting
Over a period of more than 10 years, several hundred MRS
samples have been collected around Lake Superior by United
States Geological Survey personnel and analyzed for major
and trace elements by United States Geological Survey
laboratories, using inductively coupled plasma - atomic
emission spectrometry (ICP - AES), rapid rock,' X-ray fluorescence spectrometry (XRF), and instrumental neutronactivation analysis (INAA) techniques as described in
Baedeker (1987). From this suite, nearly 70 volcanic
samples from the south shore of western Lake Superior and
35 samples from elsewhere around Lake Superior have been
analyzed for Sr, Nd, and Pb isotopic composition. Concentrations of Rb, Sr, Nd, and Sm were determined by isotope
dilution along with isotopic compositions of Sr, Nd, and Pb
at the Department of Terrestrial Magnetism, Carnegie Institution of Washington, following procedures similar to those
given in Carlson (1984) and Nicholson and Shirey (1990).
These largely unpublished results, in addition to published
and unpublished chemical data provided by J.C. Green based
on three decades of research on the North Shore Volcanic
Group, as well as the chemical analyses and (or) Nd isotopic
data reported by Brannon (1984), Dosso ( 1984), Sutcliffe
(1987), Paces (1988), Paces and Bell (1989), Nicholson and
The arcuate 1.1 Ga Midcontinent rift system extends more
than 2000 km from Kansas north beneath Lake Superior and
then southeast beneath the Michigan basin to the Grenville
Front (Fig. l), cutting across crustal terranes ranging in age
from 3.6 to 1.65 Ga (Van Schmus and Hinze 1985; Van
Schmus et al. 1987). Recent seismic reflection profiles
(Behrendt et al. 1988; Cannon et al. 1989; Allen et al. 1995)
show that the deepest portion of the rift subsided along large
normal growth faults to a depth of about 30 km, accommodating at least 20 km of rift-related volcanic rocks and
10 km of overlying sedimentary rocks. Broad shallow basins
filled with rift-related volcanic and sedimentary rocks flank
these deep central graben segments. In the western Lake
Superior region, individual central graben segments are
asymmetric, alternating across major accommodation or
,transform zones (Cannon et al. 1989; Allen et al. 1995),
whereas in eastern Lake Superior, the central graben is symmetric (Mariano and Hinze 1994). Two basement blocks
(White's ridge and Grand Marais ridge: WR and GMR in
Fig. 1) acted as topographic high points as the volcanic
basins subsided around them (Allen et al. 1995). Deposition
Sampling and analytical techniques
@ 1997 NRC Canada
506
Can. J. Earth Sci. Vol. 34, 1997
Fig. 1. (a) Map of Lake Superior region illustrating major exposures of volcanic and sedimentary rocks associated with the Midcontinent
rift (after Green et al. 1987) and the major structures as determined from seismic reflection studies (Allen et al. 1995). DF, Douglas
fault; GMR, Grand Marais ridge; IRF, Isle Royale fault; KF, Keweenaw fault; LOF, Lake Owen fault; MF, Michipicoten fault;
TF, Thiel fault; WR, White's ridge. (6) The stippled region represents the geophysical expression of the Midcontinent rift (after
Van Schmus et al. 1987).
92'
90'
88'
86'
Bayfield Group and Jacobsville Sandstone (undivided)
Oronto Group sedimentary rocks
Gabbroic-anorthositic intrusive rocks
..... . .....*._..
'
:.;.__.
Basalt flows, lesser andesites and rhyolites,
and interflow sediments
Sibley Group sedimentary rocks and Nipigon sills
Crystalline rocks (Early Proterozoic and Archean)
Shirey (1990), Klewin and Berg (1991), Lightfoot et al.
(1991), and Shirey et al. (1994), provide the basis for the
subsequent discussion.
South shore of western Lake Superior
A review of the detailed stratigraphy of the south shore of
western Lake Superior is important at this point, in light of
recent geochemical distinctions made within the earliest formations (Nicholson and Shirey 1992; Nicholson et al. 1994),
a newly refined age for the early flows (Davis and Green
1997) and new ages for some of the youngest units on the
south shore (Zartman et al. 1997), and the recent recognition
of a possible additional magnetic reversal in this section.
These newly acquired refinements provide constraints on the
nature and duration of rift magmatism, particularly the earliest magmas, and provide critical criteria with which to draw
correlations with other stratigraphic sections throughout
the rift.
Stratigraphy
In Wisconsin and Michigan, the basal unit of the Keweenawan Supergroup is the Bessemer Quartzite, a thin lacustrine
sandstone (Fig. 2). Overlying the Bessemer is the basal volcanic unit, the Siemens Creek Volcanics (SCV), which consists of two informally defined upper and lower members.
The lower member contains only basalt, including a newly
recognized picrite. The upper member is dominated by basalt
but includes minor basaltic andesite and andesite. No rocks
with Si02 > 57 wt. % have been found in this formation.
O 1997 NRC Canada
Nicholson et al.
507
Fig. 2. Stratigraphic column for Keweenawan volcanic rocks on the south shore of western Lake Superior. Also included are the
magnetic polarity and U-Pb age dates that provide constraints on regional correlation. The U-Pb age dates are from the following
sources: a, Davis and Green (1997); b, Zartman et al. (1997); c, Davis and Paces (1990).
Stratigraphic
Thickness
(km)
18
-
17
-
16
-
Upper Michigan and
North-Central Wisconsin
Magnetic
Polarity
-
n
3
2
C3
Northwest
Wisconsin
Jacobsville Sandstone
Orienta Sandstone
Freda Sandstone
Freda Sandstone
H
0
+a.
15
-
14
-
13
-
12
-
11
-
:
C
Nonesuch Shale
-
Copper Harbor
Conglomerate
N
10
-
9
-
8
-
7
-
6
-
Nonesuch Shale
Volcanics
n
3
2
C3
C
1094.0 Ma
C
a
-
1094.6 Mab
P
Q,
m
Portage Lake
Volcanics
Chengwatana
Volcanics
1096.2 MaC
5
4
-
3
-
2
-
1
-
0
-
The lower member is thin ( < 5 0 m thick), has limited
regional extent, and is distinctive because it contains augite
phenocrysts, a rarity among younger MRS basalts, including
those of the upper member of the SCV. The upper member
is more widespread than the lower member and considerably
thicker (up to 1.5 krn). Near Ironwood, Michigan, the augitephyric basalts of the lower member disappear to the east and
the basalts that directly overlie the Bessemer Quartzite are
upper Siemens Creek flows that do not contain augite phenocrysts. Pillows formed in the basal few basalt flows that
directly overlie the Bessemer Quartzite, regardless of
whether the flow composition is lower or upper Siemens
Creek. The SCV is characterized by reversed magnetic
polarity (Palmer and Halls 1986). No ages have been determined for the SCV, but by correlation with the earliest ages
from the compositionally similar units of the Osler Group
and the Nipigon sills north of Lake Superior, volcanism was
probably initiated about 1108 Ma (Davis and Green 1997).
The Kallander Creek Volcanics (KCV) overlies the SCV,
and also has been informally divided into upper and lower
O 1997 NRC Canada
Can. J. Earth Sci. Vol. 34, 1997
members. Both members range from basalt to intermediate
rocks and rhyolite. The lowermost 1.5 km of the KCV consists of flood basalt, with minor andesite and rhyolite. The
basalts typically have plagioclase phenocrysts, which, in
some flows, form large radiating clusters. A rhyolite flow
near the top of the lower KCV gives an age of 1 107.3 _+
1.6 Ma (Davis and Green 1997). The upper member (about
2 km thick) is dominated by andesite and is considered to
be the extrusive products of a localized magmatic system
(Cannon et al. 19936). On the east side of the Mellen Intrusive Complex, a thick and laterally extensive rhyolite at the
top of the upper member of the KCV yields a date of 1099.0
2.6 Ma (Zartman et al. 1997). The Mellen granite and
Mellen gabbro, which intrude the KCV, have ages of
1101.5 f 2.9 and 1102.1 f 3.5 Ma, respectively (Zartman
et al. 1997).
Volcanic rocks in the KCV are reversely polarized, with
two exceptions (Fig. 2), both of which have been recognized
during reconnaissance fieldwork in the last 2 years. There
appears to be a thin interval of flows with normal polarity at
the base of the upper member. The stratigraphic thickness of
this interval has not yet been delineated and exposure is
limited; however, most of the upper KCV elsewhere has
reversed magnetic polarity. The second exception occurs at
the top of the formation. Historically, the regionally recognized shift in magnetic polarity from reversed (lower
Keweenawan) to normal (upper Keweenawan) has been
placed at the boundary between the KCV and the overlying
Portage Lake Volcanics (PLV) on the south shore (e.g.,
Books 1968). However, a recent reconnaissance survey of
magnetic polarity identified the shift from reversed to normal
polarity as occurring a few flows below the 1099 Ma rhyolite
that forms the top of the KCV. This age corroborates ages
from the Duluth Complex that provide a lower limit of
1099 Ma on the age of the magnetic reversal (Paces and
Miller 1993) and from the North Shore Volcanic Group that
suggest the reversal is no younger than 1100.2 Ma (Davis
and Green 1997).
The PLV and all younger Keweenawan formations have
normal magnetic polarity. The PLV is exposed laterally for
about 200 km from the Mellen Intrusive Complex east to the
tip of the Keweenaw Peninsula. Near the Mellen Intrusive
Complex, the PLV disconformably overlies the KCV
(Hubbard 1975) and is concordantly overlain by the Copper
Harbor Conglomerate. Where the PLV is exposed along the
Keweenaw Peninsula and in northern Wisconsin, it consists
of 3-5 km of subaerial tholeiitic basalt flows with minor
rhyolites and few intermediate rocks, but is truncated by the
high-angle reverse Keweenaw fault (Fig. 1). Seismic reflection profiles across Lake Superior suggest that the thickness
of the PLV rocks may have been 6-7 km near the axis of
the rift (Cannon et al. 1989). Two stratigraphically separated
basalt flows from the PLV on the Keweenaw Peninsula yield
ages of 1094.0 f 1.5 and 1096.2 f 1.8 Ma (Davis and
Paces 1990).
The Porcupine Volcanics overlies the PLV in the vicinity
of the Porcupine Mountains in Michigan (Fig. l), and consists of basalt, abundant andesite, and rhyolite (Johnson and
White 1969; Hubbard 1975; Cannon and Nicholson 1992).
This unit has a lateral extent of about 35 km and maximum
thickness of at least 4 km. The overlying Copper Harbor
+
Conglomerate thins dramatically where the Porcupine Volcanic~is present. A rhyolite at the top of the Porcupine Volcanic~yields an age of 1093.8 f 1.4 Ma (Zartman et al.
1997).
The Chengwatana Volcanics, which lies in the Ashland
syncline to the west of the Mellen Intrusive Complex
(Fig. l), occupies the same stratigraphic position as the PLV,
which lies to the east of the Mellen Intrusive Complex
(Fig. 2). The Chengwatana Volcanics is at least 14 km thick
on the southeast limb of the syncline (Allen et al. 1995),
although the true thickness is unknown due to truncation by
the Lake Owen fault. A Chengwatana rhyolite near the Lake
Owen fault yields an age of 1094.6
2.1 Ma (Zartman
et al. 1997), an age comparable to the upper part of the PLV.
Near the base of the Copper Harbor Conglomerate, both
in the Porcupine Mountains area and off the north shore of
the Keweenaw Peninsula, a few to several dozen basaltic
andesite to andesite flows are interspersed within the conglomerate (Paces and Bornhorst 1985). North of the
Keweenaw Peninsula these flows are informally called the
Lake Shore traps (Lane 191 I), and one flow yields an age of
1087.2 f 1.6 Ma (Davis and Paces 1990).
+
Geochemistry
To provide a consistent basis on which to compare magma
compositions among formations, only those samples with
Si02 < 53 wt. % are considered in the following discussion.
Regional hydrothermal alteration that accompanied native
copper deposition clearly affected the mobile alkali elements
(e.g., Paces 1988; Nicholson 1990). However, less mobile
trace and major elements can be used to characterize basalt
compositions. A summary of average compositions for basalts
in each formation east of the Mellen Intrusive Complex on
the south shore appears in Table 1. Five distinctive floodbasalt compositions are recognized and designated basalt
types I -V, in order to ease the comparisons among detailed
stratigraphic successions at various localities.
Compared to basalts from other south shore formations,
the basalts of the lower member of the SCV (basalt type I)
are characterized by substantially lower A1203, the highest
MgO content, the largest MgO range (8 - 17 wt. %), and the
steepest rare earth element (REE) slopes (Table 1). They
have the highest Cr and Ni abundances and moderately high
incompatible trace element abundances (Fig. 3). One sample
from this unit collected just west of the Mellen Intrusive
Complex is picritic (MgO = 16.9 wt. %; Mg# = 0.72,
where Mg# = molar Mg/(Mg+ Fe)) and is the only MRS
picrite identified on the south shore. A single basalt sample
from this unit has been analyzed for Nd isotopic composition
and yields an EN^( 100) value of -0.6.
Basalts of the upper SCV (basalt type 11) are characterized
by higher Al2O3,lower Ti02 and MgO, and lower chondritenormalized Ce/Yb (CeN/YbN)than basalts in the lower
Siemens Creek, and still have low heavy REE (HREE) abundances (Table 1). Compatible and incompatible trace element
abundances are less overall than for the lower SCV. In addition, the upper SCV basalts show slight negative Nb and Ta
anomalies (Fig. 3), in contrast to the smooth incompatible
trace element pattern of the lower SCV basalts. Six samples
analyzed for Nd isotopic composition yield an average
E ~ d ( 1 1 0 0 )value of -3.6 (range = -1.5 to -6.9).
O 1997 NRC Canada
O
icl
z
\D
-
a
1
-1
15.6
633k313
299 + 209
50 + 6
233f71
18f 3
24.54 f4.47
365 f 123
57.94f 9.55
111.68f 18.94
48.64k 10.01
9.17k1.66
2.37f 0.42
0.87 f 0.17
1.72 k0.46
0.22kO.04
5.65k 1.32
3.55 k0.45
4.98f 1.08
0.78 k0.33
0.59
48.62f 1.28
2.41 k0.26
9.92k1.40
2.08f 0.1 1
11.77f0.61
0.22f 0.01
11.06k3.06
10.30f 1.29
2.06f 0.86
1.28 f 0.29
0.28f 0.06
100.00
-3.60
6
-6.9 to - 1.5
5 .o
201 k 8 3
107k29
17k6
137f 20
21f4
26.14f 2.06
261 k 180
20.57f 4.67
40.16f 8.09
20.24f3.08
5.01 k0.65
1.59k0.21
0.71 k0.05
1.88k0.18
0.26k0.03
3.25f 0.38
1.09k0.33
2.34f 0.65
0.5 1f 0.47
0.56
51.20+ 1.35
1.70f 0.22
15.13k0.99
1.71kO.11
9.68k0.60
0.18k0.01
6.77f 1.06
9.82f 0.93
2.74k0.50
0.84f 0.50
0.23 f0.04
100.00
Upper ( n = 18;
basalt type 11)
-0.90
1
9.6
32 f 33
49+22
50k28
290 + 89
31 k 5
18.25k4.80
471 f 315
53.11 f25.69
106.46 f 47.65
51.87f 16.60
11.16k2.67
3.21 f 0.73
1.26k0.24
2.52k0.52
0.33 k0.06
7.08f 2.15
3.56k2.01
5.86k2.74
1.54k0.71
0.39
50.92k 1.25
3.62 k0.63
14.28+ 1.01
2.05f 0.20
11.64f 1.11
0.18 k0.03
4.08f0.99
7.11 f0.90
4.18k0.96
1.40f 0.73
0.54f 0.10
100.00
Lower ( n = 10;
basalt type 111)
4
-1.8to-1.2
- 1.50
3.1
84 f 36
97 f 4 3
20 k 8
266 k 93
43f 15
33.81 f 3.67
477 f283
27.26f9.80
57.84f 21.08
31.90f 10.82
8.87 f 2.90
2.32 k0.56
1.36k0.42
4.26f 1.25
0.60f 0.16
6.42f2.23
1.32f 0.59
3.55f 1.68
0.93 k0.48
0.46
49.37 k 1.04
2.99f 0.93
14.06f 1.26
2.20f 0.25
12.48f 1.44
0.24f0.02
5.99f 1.05
7.91 f 1.41
2.77 f0.48
1.55f 0.67
0.45 f 0.17
100.00
Upper
( n = 7)
Kallander Creek Volcanics
0.4
8
-0.6to+l.O
-2.2
2.7
186f51
172k 4 0
9f 2
111f36
24k6
28.68k2.18
262.00f219
11.96f2.31
25.74f4.59
15.21k2.66
4.09 k0.57
1.33k0.17
0.70kO. 1 1
2.19k0.31
0.32 f0.05
2.66f0.59
0.61 k0.09
1.17f 0.39
0.22 f 0 . 17
0.58
48.56f 1.10
1.70k0.22
16.81k0.66
1.75f 0.12
10.44k0.68
0.19k0.05
7.90k 1.17
8.40f2.15
2.68 k0.74
1.37f1.21
0.19f 0.03
100.00
Low TiO,
( n = 31;
basalt type IV)
0.6
4
Oto+1.5
2.6
71 k 5 8
75 3 3 6
16f3
200 k 35
44f 1 1
31.60f 3.77
379f 163
20.94f4.40
44.78f8.95
25.31 f 4.00
6.88 k 1.08
2.00k0.25
1.16k0.20
3.88f0.81
0.58 f 0.10
5.00 k0.84
1.01 f0.23
2.56k0.87
0.56kO. 19
0.47
49.37f1.51
2.68f 0.31
15.21 f 0.37
1.84k0.13
12.13 f0.88
0.24 k0.06
6.05 k0.81
7.32f 1.67
3.01 k0.88
1.81f 1.15
0.36f0.10
100.00
High TiOz
( n = 8)
Portage Lake Volcanics
+
8
-11.4 to -8.3
- 10.0
4.7
90f 77
63 f 32
15f 8
237 f 127
43f19
29.73f3.19
1141f777
43.16f 24.89
86.19f 47.01
40.58f 21.87
9.04f 4.58
2.52 f 1.26
1.29f 0.60
4.19f 1.67
0.60f 0.24
5.70k2.86
1.02 k0.57
5.38f3.00
0.90f 0.64
0.49
50.41f1.40
2.14k0.72
15.12f 1.25
2.02f0.22
11.43 f 1.26
0.20f 0.05
6.05 f 1.36
7.00f 1.97
2.90f 0.61
2.11f 1.26
0.63 f0.48
100.00
Porcupine
Volcanics
( n = 27)
2.2
2
1.3 to 3.4
2.2
159k78
194k86
7k3
97 k 4 5
21 k 9
28.15 f 5.10
107 f 4 0
8.97 k4.88
18.87f 10.07
12.08k6.25
3.32f 1.57
1.14 f0.38
0.58f 0.29
1.90f 0.85
0.27 f0.12
2.30f 1.23
0.48 kO.22
0.92 f0.72
0.11 fO.OO
0.63
48.80f 0.88
1.33f0.50
17.16f 1.45
1.56k0.28
8.81 f 1.58
0.16f 0.03
8.61 f 1.82
10.26f 1.58
2.53 f0.27
0.66f 0.65
0.14k0.06
100.00
Late secondphase basalts
( n = 6;
basalt type V)
Note: Due to limitations of space, individual analyses cannot be reported here and instead are represented by averages. which are used to define the five distinctive flood-basalt compositions.
FeO) = 0.85. and trace elements in
Individual analyses are available on request from the first author. Major elements are reported in weight percent on an anhydrous basis. assuming FeO/(FelO,
parts per million as analyzed. Uncertainties on the averages are la. Mg#, MgIFe + Mg (mole percent); Ce,/Yb,, chondrite-normalized CeIYb; n. number of samples.
CeN/YbN
Nd isotopic composition
EN^( I loo)
n
Range
Outlier
Cr
Ni
Nb
Zr
Y
Sc
Ba
La
Ce
Nd
Sm
Eu
Tb
Yb
Lu
Hf
Ta
Th
U
Mg#
Sum
P2°,
FeO
MnO
MgO
CaO
Na20
K2O
Fe20,
A1203
SiO,
Ti02
Lower ( n = 5;
basalt type I)
Siemens Creek Volcanics
Table 1. Average basalt compositions for formations on the south shore of western Lake Superior.
CD
o
2
a,
3
0
cn
o_
5
zo
Can. J. Earth Sci. Vol. 34, 1997
Fig. 3. Trace element compositions for basalts erupted on the
south shore of western Lake Superior east of the Mellen Intrusive
Complex, based on average compositions reported in Table 1.
Normalizing factors for C1 chondrites from Sun and McDonough
(1989), except P (Thompson et al. 1983). (a) Trace element
compositions for basalts erupted during the first magmatic phase.
(b) Trace element compositions for basalts erupted during the
second magmatic phase. Stippled field delineates the field for
basalts illustrated in (a).
I
x
Late second-phasebasalts
Porcupine VoIcanics
II
The lower KCV contains the most evolved basalts in the
south shore section, as evidenced by these basalts having the
highest overall Ti02 content, the lowest MgO, Cr, and Ni,
and the lowest average Mg#. This compositional group
(basalt type 111) has a chondrite-normalized incompatible
trace element pattern that is very similar to that of the lower
SCV, but elemental abundances are higher (Fig. 3). Only
one sample of the lower KCV has been analyzed for Nd iso1
value of -0.9.
topic composition, yielding an E ~ d ( lm)
Basalts of ,the upper KCV have a major element composition similar to that of basalts from the lower KCV, with high
Ti02 and low MgO, but are less fractionated than the lower
KCV and have higher Mg#, Cr, Ni, and Sc (Table 1).
However, there is a marked contrast between the two units
in incompatible trace element composition (Fig. 3), with the
upper KCV basalts showing negative Nb and Ta anomalies,
which are not present in the lower KCV. In addition, the
REE slopes are distinctly different, and the upper KCV
basalts have higher HREE than the lower KCV basalts
(Fig. 3). Four samples of the upper KCV yield a narrow
range around an average E ~ d ( lm)
1
value of - 1.5 (Table 1).
The exposed PLV basalts can be divided into two groups
based on major and incompatible trace element composition,
informally described as high-Ti02 ( > 2.5 wt. % Ti02) and
low-Ti02(< about 2.0 wt.% Ti02) basalts (Nicholson
1990). The low-Ti02 PLV basalts are high in alumina
(Table l ) , and some samples have Mg#'s as high as 0.66.
With the exception of one group of younger basalts, the
low-Ti02 PLV basalts have the lowest abundances of
incompatible trace elements of any basalt group (Table 1).
The high-Ti02 PLV basalts are somewhat more differentiated (lower Mg#) and have higher incompatible trace element abundances. Slopes of REE patterns are similar for
both compositional groups (CeN/YbN= 2.6 and 2.7), and
both compositional groups have similar average E ~ d ( 1lm)
values (+0.4 vs. +0.6; Table 1). The laterally extensive
low-Ti02 basalts make up at least 90% of the PLV, whereas
the far less abundant high-Ti02 basalts are geographically
localized. Therefore, for the purposes of this discussion, the
low-Ti02PLV basalt composition is taken to be representative of basalt type IV. The average upper KCV basalt composition is nearly identical to that of the high-Ti02 PLV
basalts (Fig. 3), differing only in EN^( lm) value (Table 1).
Basalts of the Porcupine Volcanics are relatively enriched
in Th and Ba compared with other basalt types (Table 1) and
have relatively high incompatible trace element abundances
and distinct negative Nb and Ta anomalies (Fig. 3). In contrast to other MRS basalts, these basalts have strongly negative ENd(l lm) values ( - 10; Table l ) , suggesting that these
rocks represent crustally contaminated compositions.
Preliminary evaluation of chemical analyses of basalts
from the Chengwatana Volcanics on the southeast limb of the
Ashland syncline suggests that the distinctions between lowand high-Ti02 basalts determined for the PLV can be applied
to this unit as well. The bulk of the Chengwatana Volcanics
is high-alumina basalt and, for the purposes of this discussion, the PLV and Chengwatana Volcanics will be considered correlative. However, there is a basalt composition in
the upper part of ,the Chengwatana that has not been recognized in the PLV. This composition, informally called the
"late second-phase" basalt (Table l ) , refers to a group of
basalts about 0.5 krn thick near the top of the formation. This
composition (basalt type V) has only been recognized in a
few dikes elsewhere on the south shore, but flows of similar
composition appear at the top of the North Shore Volcanic
Group (NSVG) in Minnesota. Although chemical and isotopic data are limited for this unit on the south shore, it does
represent a distinctive composition 'characterized by high
Mg#, low incompatible trace element abundances, and positive E ~ d ( l1m ) values (+2.2; Table 1).
Correlations in western and central Lake
Superior
It is only within the last decade that a critical body of
stratigraphically well-controlled geochemical and isotopic
data for exposed MRS magmatic rocks has been amassed.
These data coupled with more than two dozen high-precision
U-Pb zircon ages (Fig. 4) and the recognition of multiple
changes in direction of magnetic polarity now allow correlations to be drawn on a regional scale (Fig. 5). An important
result of the high-precision zircon dating effort is the recognition that rift magmatism was episodic over a period of
about 22 Ma (Fig. 4) (Miller et al. 1995a; Davis and Green
1997).
First magmatic phase: about 1108 - 1105 Ma
The oldest ages determined for MRS magmatic rocks are
1108.2 f 0.9 Ma for the basaltic and peridotitic Logan sills
(Davis and SutcYiffe 1985; Davis and Green 1997) and
O 1997 NRC Canada
+
1108 1 Ma for the alkaline Coldwell Complex in Ontario
(Heaman and Machado 1992). At the base of the Osler
Group is a rhyolite that has been dated at 1107.5:; Ma
(Davis and Sutcliffe 1985). Numerous ages for rhyolites and
plutonic rocks in the NSVG cluster between 1107 and
1108 Ma (Fig. 4); a rhyolite in the KCV yields an age of
1107.3 Ma, and a rhyolite near the top of the central suite
in the Osler Group has been dated at 1105.3 Ma (Davis and
Green 1997).
The three basalt compositional groups erupted during the
first phase of magmatism on the south shore of Lake Superior
are also found in other localities around western Lake
Superior (Fig. 5). The augite-phyric flows of the lower SCV
are products of the first preserved erupted MRS magma
(basalt type I), a distinctive magma composition exposed
elsewhere only in western Lake Superior. Flows of this earliest phase are found at the base of the Keweenawan Supergroup on the south shore (lower SCV), as the basal units in
the Ely's Peak (Kilburg 1972) and Grand Portage areas
(Green 1972, 1977, 1995) of the NSVG, and as the basal unit
(lower suite) of the Osler Group in Ontario (Lightfoot et al.
1991) (Fig. 5). This unit is generally less than 100 m thick
and overlies a quartz sandstone. The basal flows are commonly pillowed, indicating eruption into what was probably
a shallow lake. On Black Bay Peninsula, the lower suite is
about 750 m thick and overlies a thin conglomerate
(McIlwaine and Wallace 1976; Lightfoot et al. 1991). The
basal augite-phyric unit everywhere has reversed magnetic
polarity (Green and Books 1972; Halls 1974; Halls and
Pesonen 1982; Palmer and Halls 1986). The chemical composition of this unit is consistent among localities (low
and HREE, smooth, broadly concave-down, chondritenormalized trace element pattern; Fig. 6), as is the Nd isotopic composition, which yields an average EN^( 100, of -0.1
(n = 10 samples) with a narrow range (-0.7 to +0.7;
Fig. 7).
The second composition (basalt type 11), exemplified by
the upper Siemens Creek basalts, also can be correlated with
basalts of similar composition in the NSVG, the Osler
Group, and possibly at Ely's Peak. Everywhere these basalts
have reversed magnetic polarity. On the northeast limb of the
NSVG, basalts with this composition are found in the upper
part of the Grand Portage lavas and in the overlying Hovland
lavas (Fig. 5). This compositional group forms the basal
lavas west of Grand Portage, where the augite-phyric flows
pinch out, much as the upper SCV forms the basal flows on
the south shore, where the lower SCV is absent. South shore
and north shore representatives of this compositional group
show similar chondrite-normalized trace element patterns,
including slightly negative Nb and Ta anomalies, which distinguish these rocks from the lower Siemens Creek and lower
Kallander Creek compositions (Fig. 6). This compositional
type differs from most other MRS basalt units by its broad
range of EN^( 100) values ( - 1.4 to -6.9; Fig. 7) and negative average E N * ( , 100)of - 3.8 (n = 14 samples).
On Black Bay Peninsula, the central suite overlies the
basal augite-phyric basalts. Although much of the central
suite has been shown to be affected by variable amounts of
crustal contamination (Lightfoot et al. 1991), the most primitive representatives of the central suite are broadly comparable to the upper Siemens Creek composition (Fig. 6), as
Fig.
age dates for MRS
- 4. Compilation o f published U - P b
volcanic and plutonic rocks using data from Davis and Green
(1997), Paces and Miller (1993), Zartman et al. (1997), Van
Schmus et al. (1990), Heaman and Machado (1992), Davis and
Sutcliffe (1985), Davis and Paces (1990), and Palmer and Davis
(1987).
5(8)
Volcanic Rocks
n=18
(b)
Plutonic Rocks
n=12
would be expected from its stratigraphic position. Nd isotopic data have not yet been published for the central suite.
On the south shore, the third basalt composition (basalt
type 111) is represented by the lower member of the KCV,
which overlies the SCV. The lower KCV has higher
and Ti02 and very low MgO, suggesting substantial fractionation, but is otherwise similar geochemically to the lower
Siemens Creek (Figs. 3, 6). No augite phenocrysts are
present, but large tabular plagioclase phenocrysts are common (Green 1982). This composition is also present in the
Grand Portage and Hovland lavas on the northern end of the
north shore (Fig. 5). An age of 1107.3 Ma from the top of
the lower KCV is indistinguishable from the 1107.7 Ma age
determined for the top of the Hovland lavas. This composition is also present in the Ely's Peak area, but may be intercalated with flows of the lower Siemens Creek composition
(Kilburg 1972). Only a few representative chemical analyses
are available for the Osler Group (Lightfoot et al. 1991), but
this composition does not appear to be present (or possibly
is overprinted by crustal contamination). Everywhere these
basalts have reversed magnetic polarity. At Ely's Peak and
on the south shore, this composition has an c N d ( \
of about
-0.6 (n = 3 samples; Fig. 7).
63 1997 NRC Canada
Can. J. Earth Sci. Vol. 34, 1997
51 2
Fig. 5. Schematic correlation of MRS volcanic rocks in western and eastern Lake Superior based on comparison of stratigraphic
position of distinctive flood-basalt compositions, magnetic polarity, and, except for Mamainse Point, absolute age where possible. To
adequately represent the numerous magmatic events that occurred between 1110 and 1105 Ma as well as between 1100 and 1095 Ma,
these two sections of the age scale have been expanded. Dotted lines running across the diagram represent lines of equal age. Dashed
lines in (u) separate lower and upper members of Kallander Creek and Siemens Creek volcanics. Relative formation thicknesses are
schematic: approximate preserved thicknesses of volcanic sections are reported at the bottom of diagram. Roman numerals I-V refer to
the five distinctive laterally extensive flood-basalt compositions identified on the south shore of western Lake Superior (see text for
discussion). Where equivalent basalt compositions occur in other stratigraphic successions, the appropriate Roman numeral is noted.
Additional basalt compositions not identified as one of the five distinctive ones may also be present at any given stratigraphic horizon at
a specific locality. Chemical data used for correlation are from the following sources: (a) Nicholson and Shirey (1990), Paces (1988),
and unpublished data from S.W. Nicholson; (b) Brannon (1984), and unpublished data from J.C. Green; (c) Green (1995), Vervoort
and Green (1997), and unpublished data from J.C. Green; (d) Huber (1973) for Isle Royale, Lightfoot et al. (1991) for the Osler
Group on Black Bay Peninsula, and unpublished data from S.W. Nicholson; ( e ) Annells (1974) for Michipicoten Island, Klewin and
Berg (1991) for Mamainse Point, and Shirey et al. (1994). U-Pb age dates from Davis and Green (1997), Zartman et al. (1997),
Davis and Paces (1990), Davis and Sutcliffe (1985), and Palmer and Davis (1987). Abbreviations in "Dated rock" column broadly
refer to material from which zircon or baddeleyite was recovered for dating: P, plutonic rock; B, basalt; R, rhyolite. The shaded
regions represent (i) intervals in which plutonic rocks obscure contacts between volcanic rocks (northeast Minnesota), or (ii) contacts
that are covered (Isle Royale, Black Bay, Lake Nipigon, and eastern Lake Superior). Por. Volc., Porcupine Volcanics.
....-...-......
Kallander Creek
APPROX. THICKNESS
OF VOLC. SECTION:
13 km
Evidence for a volcanic hiatus: 1105- 1100 Ma
A compilation of high-precision U -Pb zircon ages for MRS
magmatic rocks (Fig. 4) suggests that after an early period
of volcanic eruptions and emplacement of intrusions, there
was a period of diminished magmatic activity from at least
6 km
Isle Royab: 3.5 km
Osler Group: 3 km
Mich. Island:
Mamainse Pt:
4 km
5 km
1105 Ma until about 1100 Ma in western Lake Superior
(Miller et al. 1995a; Davis and Green 1997). This apparent
hiatus is best documented within the NSVG where on the
northeast limb, within about 100 m of stratigraphy, there is
an apparent 7 Ma gap between the top of the Hovland basalts
O 1997 NRC Canada
Nicholson et al
Fig. 6. Comparison of trace element patterns for each of the
five distinctive mantle-derived basalt compositions identified in
both western and eastern MRS. Solid symbols represent average
compositions for the south shore, north shore, and Osler Group
in western Lake Superior. Open symbols identify specific groups
recognized on Mamainse Point. Data source for Mamainse Point
in eastern Lake Superior is Shirey et al. (1994). Data sources
for western Lake Superior are as follows: (a) Brannon (1984),
Nicholson (1990), and J.C. Green and S.W. Nicholson,
unpublished data; Normalizing factors as in Fig. 3; (b) Brannon
(1984), Nicholson (1990), and S.W. Nicholson, unpublished
data; (c) J.C. Green and S.W. Nicholson, unpublished data;
(d) Lightfoot et al. (1991), and J.C. Green and S.W. Nicholson,
unpublished data; and (e) Kilburg (1972), Lightfoot et al.
(1991), and J.C. Green and S.W. Nicholson, unpublished data.
1000
Basalt type V
1
1
1
1
1
1
5
(a)
Late second-phase basalts and
equivalent rocks
oo
Fig. 7. Compilation of initial E,, data for MRS rocks with SiOz <
55 wt. % for western and eastern Lake Superior. Data sources
include S.W. Nicholson (unpublished data), Brannon (1984),
Paces and Bell (1989), Nicholson and Shirey (1990, 1992), and
Shirey et al. (1994). The fields labeled CC, CLM, PM, and DM
represent the estimated Nd compositions various source
reservoirs would have had at 1100 Ma. CC, field for Archean
crust; CLM, field for continental lithospheric mantle (Shirey
1997); PM, plume; DM, depleted asthenospheric mantle source
(like that of modern mid-ocean-ridge basalt) (Nicholson and
Shirey 1990; Shirey et al. 1994).
CC CLM
PM
DM
Group 7
Mamainse Point
1
1
1
1
1
1
1
1000 !
3 Basalt type IV
(b)
: Portage Lake and equivalent rocks
1
5
a
C,
cn
a
1
1
1
1
1
I
~ a m a i n s ePoint
Lower Kallander Creek
and equivalent rocks
1
1
1
1
1
1
1
1
1
1000]
(d)
Basalt type 11
m
Upper Siemens Creek
and equivalent rocks
Creek Volcanics
10
Group 1
~Bmainsepoint
J
lo00
1
I
I
I
I
u
I
I
I
I
I
I
I
I
I
I
l
l
1
I
I
I
I
I
~
(e)
Basalt type I
I
I
I
I
Th NbTaLa Ce P SmZr Hf Ti Tb Y Yb
(1 107 Ma) and the base of the overlying Chicago Bay flows
(1 100 Ma) (Davis and Green 1997). Intrusion of the Duluth
Complex obscures any volcanic stratigraphy that might have
0
-12
-10
-8
-6
-4
-2
0
2
4
6
Initial F N ~
been erupted in this time interval on the southwest limb.
On the south shore, there is no obvious break in stratigraphy within the 3.5 km thick KCV that might mark a hiatus
in volcanism, although outcrop becomes increasingly sparse
upsection. The top of the 1.5 km thick lower KCV is about
1107 Ma and, given that the Mellen Intrusive Complex was
emplaced into the lower and upper KCV about 1102 - 1101
Ma (Zartman et al. 1997), the earliest upper KCV lavas had
to have erupted by 1102 Ma. Thus, any volcanic hiatus on
the south shore could be as much as 5 Ma (1 107 - 1102 Ma).
In the Osler Group, the Agate Point rhyolite at the top of
GI 1997 NRC Canada
Can. J. Earth Sci. Vol. 34. 1997
the reversely polarized central suite has been dated at
1105 Ma (Davis and Green 1997). Most of the upper suite
is also magnetically reversed, and a 10 m thick conglomerate
separates those flows from a few overlying flows of normal
magnetic polarity (Halls 1974). The reversely polarized
flows make up a stratigraphic thickness of nearly 3000 m. If
the magnetic reversal in the Osler Group took place at about
1100 Ma based on ages in the NSVG and KCV, then only
about 600 m accumulated during the period between the
youngest dated eruption of the central suite (1 105 Ma) and
the magnetic reversal (1 100 Ma).
Thus, there is an apparent diminishing of flood-basalt
eruptions in several localities around western Lake Superior
from about 1105 to about 1100 Ma. Based on available ages,
the duration is apparently about 5 -7 Ma in any given locality, but localized magmatic systems (individual central volcanoes and emplacement of plutons) may have been active
during this time (e.g., the upper member of the KCV and
related Mellen Intrusive Complex, and perhaps parts of the
central and upper suites of the Osler Group).
Second magmatic phase: 1100 - 1094 Ma
Beginning at about 1100 Ma, after the period of relative magmatic quiescence, flood-basalt volcanism and emplacement
of large intrusions resumed dramatically around the rift
(Fig. 4). This phase of MRS volcanism is the best exposed
phase in western Lake Superior and is represented by the
normally polarized PLV on the south shore and Isle Royale
(Huber 1973; Paces 1988; Nicholson 1990), the normally
polarized sections of the southwest and northeast limbs of the
NSVG (Brannon 1984; Dosso 1984; Green 1972, 1982,
1983), and the few flows of normally polarized volcanic
rocks at the top of the Osler Group (Lightfoot et al. 1991).
Much of the Duluth Complex was intruded about 1099 Ma,
followed shortly by the Beaver Bay Complex (Paces and
Miller 1993).
On the Keweenaw Peninsula 5-6 km of PLV is preserved, all normally polarized. Two basalt flows stratigraphically separated by about 2.5 km yield ages of 1094 and
1096 Ma (Davis and Paces 1990). The low-Ti02 Portage
Lake flood basalts (basalt type IV: high A1203, low REE
abundances; Fig. 6), which make up about 90% of the formation, are remarkably consistent in their chemical and
Nd isotopic composition over a very large outcrop area
(Keweenaw Peninsula to Isle Royale) as well as a substantial
stratigraphic thickness (Huber 1973; Paces 1988; Nicholson
1990). This compositional type has E ~ d ( 100)
1
values of + 1.0
to -0.6 (Table 1) and average EN^(, 100) of +0.4 (n = 8 samples with one outlier at -2.2). High-Ti02 PLV basalts
(about 10% of the formation) are more fractionated than the
low-Ti02 PLV basalts but also have a narrow range of
EN^( 100) values of + 1.5 to 0.0 (Table I), with an average
E ~ d ( 100)
1
of +0.6 (n = 4 samples). Chemical data are sparse
for the few normally polarized flows at the top of the Osler
Group, but those reported by Lightfoot et al. (1991) are similar to the composition of low-Ti02 PLV basalt. This correlation is corroborated by the single Nd isotopic analysis
available, which yields an E ~ d ( 100)
1
= -0.2.
On the southwest limb of the NSVG, nearly 10 km of
stratigraphy is preserved, of which about 9 km is normally
polarized. More than 6 km of these normally polarized flows
was erupted between about 1098 and 1096 Ma (Davis and
Green 1997), suggesting that the bulk of the PLV may be
slightly younger than the bulk of the NSVG on this limb.
Brannon (1984) analyzed 160 successive flows (about 4 km
thick in total), providing excellent compositional coverage
for nearly half of the stratigraphic thickness of the southwest
limb. Both low- and high-Ti02 basalts are well represented
in this section, whereas intermediate and felsic flows are few
and confined to the lower part. Brannon (1984) reported
E~d(1
100) values for basalts and basaltic andesites that range
from + 1.9 to - 1.1 (n = 12 samples, with one outlier at
-5.9), comparable to the values determined for the PLV
(Paces and Bell 1989; Nicholson and Shirey 1990).
The northeast limb of the NSVG differs substantially from
the southwest limb in two fundamental ways. First, the entire
stratigraphic section is condensed into about 6.5 km. Of the
4 km of normally polarized rocks, the lowermost several
hundred metres was erupted between about 1100 and
1098 Ma, and the remaining 3 km is broadly time equivalent
to the bulk of the southwest limb. Second, rhyolites make up
at least 25 % of the section (Green 1995; Vervoort and Green
1997), and intermediate rocks are abundant. Green (1995)
attributes much of the chemical variation observed in this
region to the development of crustal magma chambers in
which mantle magmas ponded, fractionated, and variably
assimilated Archean country rock. For three mafic samples
in the Keweenawan reference suite (Basaltic Volcanism
Study Project 1981) collected on this limb, Dosso (1984)
reported E~d(1100) values ranging from -0.9 to -9.5,
demonstrating a much broader range than is typical for this
unit elsewhere. Vervoort and Green (1997) show that the
rhyolites in this section become progressively more contaminated by older material with time. They suggest that large
crustal magma chambers, such as may be represented by
intrusions in the Duluth Complex, produced prolonged
crustal heating and melting, substantially affecting mantle
melts emplaced into the crust.
A late flood-basalt composition occurs near the top of the
volcanic sequence and has been recognized in the Schroeder
and Lutsen basalts at the top of the NSVG, in some dikes that
cut older MRS units on the south shore, near the top of the
Chengwatana on the south shore, and possibly as a few flows
near the top of the PLV on the Keweenaw Peninsula (Paces
1988). This composition (basalt type V) is more primitive
than other MRS basalt compositions, with the highest average A1203content and the lowest incompatible trace element
abundances (Table 1; Fig. 6). When taken as a group, these
samples are displaced towards more positive E ~ d ( 100)
1
values
( + 1.3 and + 3.4; n = 2 south shore samples) compared with
the Portage Lake equivalent rocks (Fig. 7). Miller et al.
(19956) have recently suggested that the Schroeder-Lutsen
basalts at the top of the NSVG may represent the northwest
edge of the PLV, erupted beyond the central rift graben. If
so, their composition as indicated here implies that they were
erupted near the end of the accumulation of the PLV, not at
its beginning. Ages for the late basalts at the top of the NSVG
and in the Chengwatana would provide much needed constraints on the minimum age of flood-basalt volcanism and
would facilitate correlations between the north and south
shores of western Lake Superior.
As the second phase of flood-basalt volcanism waned
@ 1997 NRC Canada
Nicholson et al.
around 1094 Ma, upper crustal magma chambers developed
locally near or at the top of the Portage Lake or equivalent
rocks, most notably in the Porcupine Mountains area (about
1094 Ma) of the south shore (W.S. White 1968; Green 1977;
Nicholson et a1. 1991). These central volcanoes produced
abundant intermediate and rhyolitic rocks. Even the most
mafic volcanic rocks associated with the Porcupine Volcanic~have E ~ d ( 1 values around - 10.0 (n = 8 samples),
suggesting substantial interaction with crustal sources.
Volcanism during thermal subsidence: about 1087 Ma
Flood-basalt volcanism waned abruptly after about 1094 Ma
(Fig. 4). Two outcrop areas in western Lake Superior record
the period of sedimentation that followed volcanism; one is
the south shore, where late-stage compression has exposed a
portion of the central graben, and the other is the stratigraphically equivalent section on Isle Royale, also an
upthrust section of the central graben. Very minor volcanism
occurred after flood-basalt volcanism ceased. Sedimentation
was well established on the south shore when the few flows
of the Lake Shore traps were erupted near the base of the
Copper Harbor Conglomerate. Davis and Paces (1990)
report an age of about 1087 Ma for one of these flows. This
is close to the age of a rhyolitic body (about 1086 Ma)
preserved on Michipicoten Island in eastern Lake Superior
(Palmer and navis 1987). In addition, there is a small andesite (the Bear Lake body) that intruded the Freda Sandstone
on the Keweenaw Peninsula, although this intrusion has not
been successfully dated.
Regional comparison between western
and eastern Lake Superior
Stratigraphic and geochemical correlation
We have shown that the stratigraphic succession of five distinct flood-basalt compositions recognized on the south shore
can be generally correlated with well-documented stratigraphic successions elsewhere around western Lake Superior.
However, a comparison of the south shore sequence and the
succession of basalt compositions reported for Mamainse
Point (Shirey et al. 1994) and Michipicoten Island (Annells
1974) in eastern Lake Superior points up several important
differences.
The first significant difference is that the basal augitephyric unit (lower Siemens Creek composition, basalt type I)
has not been recognized on Mamainse Point (Berg and
Klewin 1988; Klewin and Berg 1990, 1991; Massey 1980,
1983), Alona Bay and Cape Gargantua (Massey 1980), or
Michipicoten Island (Annells 1974) in eastern Lake Superior.
Instead, at Mamainse Point, the basal unit (group 1) has
some chemical characteristics that are similar to the upper
SCV composition (basalt type 11; Figs. 6, 7), although
picrites are found in this group at Mamainse Point (Berg and
Klewin 1988), but not in the equivalent unit in western Lake
Superior. Incompatible trace element abundances of group 1
at Mamainse Point are among the lowest of MRS basalt compositions, and distinct negative Ta and Nb anomalies appear
in group 1 (Klewin and Berg 1991) as they do in the similar
upper SCV basalts (Fig. 6). Isotopically, the upper SCV and
group 1 are also comparable: an average E ~ d ( 1 value of
about -4 determined for group 1 on Mamainse Point (Shirey
et al. 1994) is the same as the average for upper SCV equivalent rocks in western Lake Superior (Table I).
The second significant difference is that group 2 rocks on
Mamainse Point have no direct correlative unit on the south
shore of western Lake Superior. The incompatible trace element pattern of group 2 rocks is similar to the pattern for the
basalt type I1 (Fig. 6), except that the Nb and Ta anomalies
are smaller (Klewin and Berg 1991; Shirey et al. 1994). Yet,
isotopically, group 2 rocks have E ~ d ( 1 values (-0.6 to
-2.0) that are close to those of the lower Siemens Creek and
the lower Kallander Creek units (Fig. 7).
Basalt compositions in groups 3 -5 at Mamainse Point
were postulated to reflect a progressive increase in the
amount of crustal contamination (Klewin and Berg 1991;
Shirey et al. 1994; Shirey 1997). Given that the crustal component overprints the mantle signature, it is difficult to evaluate the original mantle magma type. Because the effects of
crustal contamination are strongly dependent on the age and
composition of the local basement rocks, direct compositional correlation with units on the south shore is not possible. However, the stratigraphic position and magnetic
polarity of groups 3 -5 as well as the association with the
Great Conglomerate suggest that these flows could be time
equivalent to the KCV.
Groups 6 and 7 rocks at Mamainse Point are stratigraphically and chemically correlative with the PLV (and equivalent
north shore rocks, basalt type 1V) and the late second-phase
basalts (basalt type V), respectively (Figs. 6, 7). Group 6
rocks are most similar to high-Ti02 PLV basalts, including
slight Nb and Ta anomalies (Fig. 6) and E ~ d ( l l * ) values that
center around 0 (Fig. 7). Group 7 rocks, like the youngest
NSVG basalts, are more primitive compositions and have
trace element abundances lower than in most other units. The
slight Nb and Ta anomalies present in group 6 rocks are
diminished or absent in the group 7 (Klewin and Berg 1991)
and later NSVG basalts. Groups 6 and 7 basalts have
E ~ d ( 1 1 values
~ )
that range from 0 to + 3 (Fig. 7).
Group 8 rocks of Mamainse Point have no stratigraphic
equivalent on the south shore of western Lake Superior.
However, despite their stratigraphic differences, the chemis1
to
try (Fig. 6) and isotopic characteristics ( ~ ~ ~ =( -0.4
- 1.7) of group 8 rocks are equivalent to those of the lower
Kallander Creek unit (basalt type 111), suggesting that their
origins are similar.
Correlation based on absolute age
The correlation between western and eastern Lake Superior
outlined above is based on stratigraphic, geochemical, and
isotopic similarities coupled with shifts in magnetic polarity.
Regional correlations based on magnetic polarity have been
complicated by the presence of two additional reversals
(R- N -R-N) at Mamainse Point that are generally not
recognized elsewhere around Lake Superior (Halls and
Pesonen 1982) . However, if the newly recognized normal
magnetic interval on the south shore is correlated with the
lower normal interval at Mamainse Point, then group 5
basalts at Mamainse Point could be time equivalent to the
normally polarized I-lows at the bottom of the upper KCV.
This correlation supports the inference that the Great
Conglomerate at Mamainse Point was deposited during the
period of diminished volcanic activity that preceded the
@ 1997 NRC Canada
Can. J. Earth Sci. Vol. 34, 1997
51 6
Table 2. Proposed correlation of MRS magma types between western and eastern Lake Superior and their inferred source reservoirs
(based on stratigraphic position and geochemical and Nd isotopic characteristics of basalts with Si02 < 53 wt. %).
Western Lake Superior, south shoreu
EN^( I 100)
Stratigraphic position
Eastern Lake Superiorh
Sources
Second-phase volcanism
No equivalent rocks
- 12 to -9
Porcupine Volcanics
Late second-phase basalts
0 to +4
Portage Lake Volcanics
- 1 to +2
PM
PM
PM
First-pase volcanism
-2 to - 1
- 1 to 0
-7 to - 1
- 1 to + I
PM f CC"
PM
PM
CLM
PM
Upper Kallander Creek
Lower Kallander Creek
Upper Siemens Creek
Lower Siemens Creek
+ CC"
+ DM
+
EN^( I 100)
Stratigraphic position
Sources
-2 to 0
No equivalent rocks
+ 2 to +4
0 to + 2
Group 8
PM
Group 7
Group 6
PM
PM
-9 to - 3
No equivalent rocks
-5 to 0
No equivalent rocks
Groups 3 - 5
PM
Groups 1, 2
PM
+ DM
+ DM
+ CC"
+ CLM
Notes: source abbreviations: PM,
c ~ ~ ~ n t i n e nlithospheric
tal
mantle; DM, depleted mantle; CC, continental crust.
"Data include published analyses for Portage Lake Volcanics (Nicholson and Shirey 1990) and the authors' unpublished data for the Siemens Creek,
Kallander Creek, and Porcupine volcanics and a few late basalts. Also included in the initial E. ranges are several unpublished Nd analyses from the
North Shore Volcanic Group for rocks supplied by J.C. Green which are equivalent to the Siemens Creek, lower Kallander Creek, and late basalts.
' ~ a t afor the eastern Lake Superior region from Mamainse Point, Ontario (Shirey et al. 1994).
' Local assimilation.
second phase of volcanism in western Lake Superior, and
that the uppermost magnetic reversal at Mamainse Point is
correlative with the regionally observed shift from reversed
to normal magnetic polarity (Klewin and Berg 1990; Shirey
et al. 1994).
This correlation appears to satisfy much of the available
stratigraphic, geochemical, and isotopic data. However, the
only U - Pb zircon age available for Mamainse Point is a date
of about 1096 Ma for a volcanic rock near the top of the
lowermost reversed section (Davis et al. 1995). This age
appears to be too young relative to the rock's inferred stratigraphic position based on the geochemical correlation outlined above. At the present time it is unclear whether this age
was determined on a flow, a tuff, or a dike. Our interpretation is consistent with this being a younger intrusive rock. If
it proves be the age of an extrusive rock, reinterpretation of
the regional correlation will be required.
Discussion
The massive volcanism of the Midcontinent rift has been
attributed to the emplacement of a mantle plume beneath
North America (e.g., R.S. White and McKenzie 1989, 1995;
Hutchinson et al. 1990; Nicholson and Shirey 1990; R.S.
White 1997), a model supported by a wide array of geophysical, geochemical, and geochronologic data. The comparison
of well-documented but geographically distinct stratigraphic
sections of MRS volcanic rocks reveals that five distinct
flood-basalt compositions are common to most if not all of
the sections, and that these compositions appear more or less
in the same stratigraphic order in each locality (Fig. 5).
Factors including plume heterogeneity and the presence or
absence of continental lithospheric mantle and its compositional variability over a broad region may contribute to
observable variations in magma compositions. Thus, it is
remarkable that a sequential progression of basalt compositions can be recognized in as widely dispersed a distribution
of MRS rocks as is seen throughout the Lake Superior basin.
At each locality there are also suites of volcanic rocks
whose chemical and isotopic characteristics reflect abundant
crustal assimilation, obscuring to various degrees the original mantle source characteristics. These crust-dominated
suites include the Porcupine Volcanics on the south shore;
much of the northeast limb of the NSVG; groups 3-5 at
Mamainse Point; and the central suite and lower part of the
upper suite of the Osler Group. The stratigraphic position of
these suites differs among localities and most likely reflects
the influence of local structural features, changes in the
tectonic setting and extension rate, and the effects of the
localized emplacement of intrusions. Because the ages and
compositions of crustal material differ across the region,
there is considerable compositional variation among these
crustally contaminated suites; thus these suites are less useful
for regional correlation than the distinctive flood-basalt compositions. Focusing on the regional stratigraphic eruption
sequence of distinctive flood basalt compositions, rather than
on the more localized volcanic products of mid- to uppercrustal staging chambers, allows construction of a model for
the nature and interaction of mantle sources through time as
the rift developed.
Recent workers (Klewin and Berg 1991; Klewin and
Shirey 1992; Shirey et al. 1994; Shirey 1997) have combined
detailed stratigraphic sampling with careful chemical and
isotopic modelling of MRS volcanic rocks, and proposed a
model for the progression of mantle sources during rifting
based on evidence observed at Mamainse Point. This model
begins with early interaction of plume melts and the continental lithospheric mantle, followed by melts dominated
largely by the plume component, and finally the mixing of
depleted mantle and plume components near the end of magmatism. (Table 2). The compiled data from western Lake
Superior now allow this model to be tested more broadly.
What is unaccounted for in this model developed for
Mamainse Point is the source of the initial magma type found
O 1997 NRC Canada
Nicholson et al
as basal units around western Lake Superior (the augitephyric basalts and picrite). Based on inversion of rare earth
element data from the lower suite in the Osler Group, R.S.
White and McKenzie (1995) calculated that this initial
magma type represents small-degree partial melts that last
equilibrated with a garnet-bearing residue. These earliest
melts have chemical characteristics that are analogous to
1
of about
modern ocean-island basalts (Fig. 8), and E ~ d ( 100)
0. They most likely represent melts directly from the plume
without any appreciable interaction with other sources.
Combining the results from studies in both western and
eastern Lake Superior leads to a unified model for the nature
and interaction of magma sources through time as the rift
developed. As the ascending plume head approached the base
of the lithosphere and began to melt, high-MgO melts were
produced at depths of more than about 75 km (R.S. White
and McKenzie 1995). A few of these earliest melts were
erupted through extensional fractures in the lithosphere
without much interaction with any other source (basalt
type I). Melting began at the bottom of the lithosphere and
magmas, which were large-degree partial melts from the
plume, interacted with another source, most likely the continental lithospheric mantle (Klewin and Berg 1991; Shirey
et al. 1994; Shirey 1997). These magmas (basalt type 11)
were erupted in large volumes (upper Siemens Creek composition, some early NSVG rocks, and Mamainse Point
group 1). Melting was still deep, as evidenced by the low
HREE abundances and the high CeN/YbNof these units
(Table 1; Fig. 6), and REE inversions of groups 1 and 2
compositions at Mamainse Point (R.S. White and McKenzie
1995; R.S. White 1997). Some small-degree partial melts of
the plume (basalt type 111) underwent fractionation (low
MgO, Mg#, Cr, Ni, and Sc, and high Ti02, P2O5, and
CeN/YbN), retaining their plume isotopic characteristics
(Figs. 7, 8). These highly evolved basaltic magmas were
erupted not only early in the rifting cycle (i.e., lower
Kallander Creek and equivalent rocks) but also as smallvolume eruptions near the end of volcanism (group 8,
Mamainse Point).
The first stage of extension and flood-basalt eruption
waned, leading to a period of diminished volcanic activity
lasting several million years. During this time, the crust was
being heated by magmas from the mantle staging in lowerto mid-crustal chambers. These chambers led to localized
volcanic centers with abundant intermediate and felsic rocks
(central suite and parts of the upper suite in the Osler Group;
groups 3-5 at Mamainse Point; upper Kallander Creek on
the south shore; some NSVG rocks) and some intrusions
(e.g ., Mellen Intrusive Complex). At about 1100 Ma the
initiation of the second phase of magmatism was marked by
renewed extension resulting in growth faults and grabens that
quickly filled with an enormous volume of basalt (Portage
Lake and equivalent rocks: basalt type IV). These magmas
were dominated by the plume component and last equilibrated at shallower levels than the earliest magmas. Assimilation of older (Archean) basement was confined to magmas
emplaced along the margins of the rift (e.g., NSVG). As the
plume finally dissipated, depleted asthenospheric mantle
mixed with the plume component and produced compositions
(basalt type V) that progressively became more depleted in
incompatible trace elements with time ( ~ ~ ~ (>~ 0).
~ 0 0 )
Fig. 8. Comparison of MRS magma compositions with modern
sources of oceanic island basalts (Sun and McDonough 1989) for
basalts erupted in the first (a) magmatic and second (b) phases.
MORB, mid-ocean-ridge basalt.
ean Island Basalt
E-Woe MORB
f
N-type MORB
0 Lower Siemens Creek Volcanics
+ Upper Siemens Creek Voicanics
Lower Kallander Creek Volcanics
Ocean Island Basalt
2a
m
N-type MORB
Second-phasebasalts
0 Upper Kallander Creek Volcanics
A Portage Lake Volcanics
0.1
The main conclusion of this paper is that throughout the
central portion of the MRS there is a similarity in the
sequence of basalts derived from mantle source types. Early
enriched plume melts interact with continental lithospheric
mantle; they are followed in turn by voluminous melts
derived chiefly from the plume and finally by mixing of
plume-derived melts with depleted upper mantle. That such
a parallel progression of mantle sources with magmatic
development of the rift can be extended across the central
600 km of the Midcontinent rift is remarkable considering
potential heterogeneities that occur in plumes, the geographical extent of the correlation, and the potential differences in
the age and composition of the continental lithosphere cut by
the MRS. Furthermore, a similar progression is found in
other flood-basalt provinces including the North Atlantic
(e.g., Gariepy et al. 1983; Dickin 1988; Gill et al. 1988) and
the Deccan (e.g., Mahoney 1988; Peng et al. 1994). Thus,
this style of interaction between mantle plume, continental
lithospheric mantle, and depleted convecting mantle appears
to have operated for at least 1.1 Ga in continental settings
@ 1997 NRC Canada
Can. J. Earth Sci. Vol. 34, 1997
ranging from failed rifts to successful rifts that lead to complete ocean basins.
Acknowledgments
We thank L.G. Woodruff for help with graphics preparation.
Discussions with J.H. Berg, W . F . Cannon, K.W. Klewin,
J.D. Miller, and L.G. Woodruff were instrumental in our
thinking about rift evolution. R. A. Ayuso, R. W . Baragar,
J.H. Berg, and W.F. Cannon provided thorough and insightful reviews.
References
Allen, D.J., Braile, L.W., Hinze, W.J., and Mariano, J. 1995. The
Midcontinent rift system, U.S.A. : a major Proterozoic continental rift. In Continental rifts: evolution, structure, tectonics.
Edited by K.H. Olsen. Elsevier, New York, pp. 375 -407.
Annells, R. N . 1974. Keweenawan volcanic rocks of Michipicoten
Island, Lake Superior, Ontario; an eruptive centre of Proterozoic age. Geological Survey of Canada, Bulletin 2 18.
Arndt, N.T., and Christensen, U. 1992. The role of lithospheric
mantle in continental flood basalt volcanism: thermal and
geochemical constraints. Journal of Geophysical Research, 97:
10967 - 10981.
Baedeker, P.A. (editor). 1987. Methods for geochemical analysis.
United States Geological Survey, Bulletin 1770.
Basaltic Volcanism Study Project. 1981. Pre-Tertiary continental
flood basalts. In Basaltic volcanism on the terrestrial planets.
Pergamon Press, Inc., New York, pp. 60-77.
Behrendt, J.C., Green, A.G., Cannon, W.F., et al. 1988. Crustal
structure of the Midcontinent rift system: results from
GLIMPCE deep seismic reflection profiles. Geology, 16:
81 -85.
Berg, J. H., and Klewin, K. W. 1988. High-MgO lavas from the
Keweenawan midcontinent rift near Mamainse Point, Ontario.
Geology, 16: 1003- 1006.
Books, K.G. 1968. Magnetization of the lowermost Keweenawan
lava flows in the Lake Superior area. United States Geological
Survey, Professional Paper 600D, pp. D245 - D254.
Brannon, J.C. 1984. Geochemistry of successive lava flows of the
Keweenawan North Shore Volcanic Group. Ph.D. thesis,
Washington University, St. Louis, Mo.
Campbell, I.H., and Griffiths, R.W. 1990. Implications of mantle
plume structure for the evolution of flood basalts. Earth and
Planetary Science Letters, 99: 79 - 93.
Cannon, W.F. 1992. The Midcontinent rift in the Lake Superior
region with emphasis on its geodynamic evolution. Tectonophysics, 213: 41 -48.
Cannon, W.F., and Nicholson, S.W. 1992. Revisions of stratigraphic nomenclature within the Keweenawan Supergroup of
northern Michigan. United States Geological Survey, Bulletin
1970-A.
Cannon, W.F., Green, A.G., Hutchinson, D.R., et al. 1989. The
North American Midcontinent rift beneath Lake Superior from
GLIMPCE seismic reflection profiling. Tectonics, 8: 305 - 332.
Cannon, W.F., Peterman, Z.E., and Sims, P.K. 1993a. Crustal
scale thrusting and origin of the Montreal River monocline - a
35-km-thick cross section of the Midcontinent rift in northern
Michigan and Wisconsin. Tectonics, 12: 728 - 744.
Cannon, W. F., Nicholson, S.W., Zartman, R.E., Peterman, Z. E.,
and Davis, D.W. 1993b. The Kallander Creek volcanics - a
remnant of a Keweenawan central volcano centered near
Mellen, Wisconsin. Institute on Lake Superior Geology, Annual
Meeting, Proceedings and Abstracts, 39: 20 -2 1.
Carlson, R.W. 1984. Isotopic constraints on Columbia River flood
basalt genesis and the nature of the subcontinental mantle.
Geochimica et Cosmochimica Acta, 48: 2357 -2372.
Cox, K.G. 1980. A model for flood basalt volcanism. Journal of
Petrology, 21 : 629 - 650.
Davis, D.W., and Green, J.C. 1997. Geochronology of the North
American Midcontinent rift in western Lake Superior and implications for its geodynamic evolution. Canadian Journal of Earth
Sciences, 34: 476-488.
Davis, D.W., and Paces, J.B. 1990. Time resolution of geologic
events on the Keweenawan Peninsula and implications for
development of the Midcontinent Rift system. Earth and Planetary Science Letters, 97: 54 -64.
Davis, D.W., and Sutcliffe, R.H. 1985. U -Pb ages from the Nipigon plate and northern Lake Superior. Geological Society of
America Bulletin, 96: 1572- 1579.
Davis, D.W., Green, J.C., and Manson, M. 1995. Geochronology
of the 1.1 North American Mid-continent rift. Institute on Lake
Superior Geology, Annual Meeting, Proceedings and Abstracts,
41: 9- 10.
Dickin, A.P. 1988. The North Atlantic Tertiary Province. In Continental flood basalts. Edited by J. D. Macdougall. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 1 11- 149.
Dosso, L. 1984. The nature of the Precambrian subcontinental
mantle: isotopic study (Sr, Pb, Nd) of the Keweenawan volcanism of the North Shore of Lake Superior. Ph.D. thesis, University of Minnesota, Minneapolis, Minn.
GariCpy, C., Ludden, J., and Brooks, C. 1983. Isotopic and trace
element constraints on the genesis of the Faeroe lava pile. Earth
and Planetary Science Letters, 63: 257 - 272.
Gill, R.C.O., Nielsen, T.F.D., Brooks, C.K., and Ingram, G.A.
1988. Tertiary volcanism in the Kangerdlugssuaq region, E
Greenland: trace-element geochemistry of the Lower Basalts
and tholeiitic dyke swarms. In Early Tertiary volcanism and the
opening of the NE Atlantic. Edited by A.C. Morton and L.M.
Parson. The Geological Society, Blackwell Scientific Publications, Oxford, U.K., pp. 161- 179.
Green, J.C. 1972. North Shore Volcanic Group. In Geology of
Minnesota: a centennial volume. Edited by P.K. Sims and
G.B. Morey. Minnesota Geological Survey, St. Paul, Minn.,
pp. 294-332.
Green, J.C. 1977. Keweenawan plateau volcanism in the Lake
Superior region. In Volcanic regimes in Canada. Edited by
W. R. A. Baragar, L.C. Coleman, and K. M. Hall. Geological
Association of Canada, Special Paper 16, pp. 407 -422.
Green, J.C. 1982. Geology of Keweenawan extrusive rocks. In
Geology and tectonics of the Lake Superior basin. Edited by R.J.
Wold and W.J. Hinze. Geological Society of America, Memoir
156, pp. 47-55.
Green, J.C. 1983. Geologic and geochemical evidence for the
nature and development of the Middle Proterozoic (Keweenawan) midcontinent rift of North America. Tectonophysics, 94:
413 -437.
Green, J.C. 1986. Lithogeochemistry of Keweenawan igneous
rocks. Minnesota Department of Natural Resources, Division of
Minerals Project 241-4, Open-file Report.
Green J.C. 1995. Chemostratigraphy at the fringe of the Midcontinent rift system: the northeast limb of the North Shore Volcanic
Group, Minnesota [abstract]. Proceedings, Field Conference
and Symposium, International Geological Correlation Programme project 336, Duluth, Minn., pp. 55-56.
Green, J.C., and Books, K.G. 1972. Paleomagnetic evidence for
the extent of the lower Keweenawan lavas in Minnesota. Institute on Lake Superior Geology, Annual Meeting, Proceedings
and Abstracts, 18, paper 8.
Green, J.C., Bornhorst, T.J., Chandler, V.W., Mudrey, M.G.,
Jr., Myers, P.E., Pesonen, L.J., and Wilband, J.T. 1987.
Keweenawan dykes of the Lake Superior region: evidence for
@ 1997 NRC Canada
Nicholson et al.
evolution of the Middle Proterozoic Midcontinent rift of North
America. In Mafic dyke swarms. Edited by H.C. Halls and
W. F. Fahrig. Geological Association of Canada, Special
Paper 34, pp. 289 - 302.
Halls, H.C. 1974. A paleomagnetic reversal in the Osler volcanic
group, northern Lake Superior. Canadian Journal of Earth
Sciences, 11: 1200- 1207.
Halls, H.C., and Pesonen, L.J. 1982. Paleomagnetism of Keweenawan rocks. In Geology and tectonics of the Lake Superior basin.
Edited by R.J. Wold and W.J. Hinze. Geological Society of
America, Memoir 156, pp. 173 -201.
Heaman, L.M., and Machado, N. 1992. Timing and origin of the
Midcontinent rift alkaline magmatism, North America: evidence
from the Coldwell Complex. Contributions to Mineralogy and
Petrology, 110: 289 - 303.
Hinze, W.J., Braile, L.W., and Chandler, V.W. 1990. A geophysical profile of the southern margin of the Midcontinent rift system
in western Lake Superior. Tectonics, 9: 303 - 3 10.
Hubbard, H .A. 1975. Lower Keweenawan volcanic rocks of Michigan and Wisconsin. Journal of Research of the United States
Geological Survey, 3: 529 -54 1.
Huber, N.K. 1973. The Portage Lake Volcanics (Middle Keweenawan) on Isle Royale, Michigan. United States Geological
Survey, Professional Paper 754-C, pp. C 1 - C32.
Hutchinson, D.R., White, R.W., Cannon, W. F., and Schulz, K.J.
1990. Keweenaw hot spot: geophysical evidence for a 1.1 Ga
mantle plume beneath the Midcontinent Rift System. Journal of
Geophysical Research, 95: 10 869 - 10 884.
Johnson, R.F., and White, W .S. 1969. Preliminary report on the
bedrock geology and copper deposits of the Matchwood quadrangle, Ontonagon County, Michigan. United States Geological
Survey, Open-file Report 1272.
Kilburg, J.A. 1972. Petrology, structure, and correlation of the
Upper Precambrian Ely's Peak basalts. M.S. thesis, University
of Minnesota, Duluth, Minn.
Klewin, K.W., and Berg, J.H. 1990. Geochemistry of the
Mamainse Point volcanics, Ontario, and implications for the
Keweenawan paleomagnetic record. Canadian Journal of Earth
Sciences, 27: 1194 - 1199.
Klewin, K.W., and Berg, J.H. 1991. Petrology of the Keweenawan
Mamainse Point lavas, Ontario: petrogenesis and continental rift
evolution. Journal of Geophysical Research, 96: 457-474.
Klewin, K.W., and Shirey , S.B. 1992. The igneous petrology and
magmatic evolution of the Midcontinent rift system. Tectonophysics, 213: 33 -40.
Lane, A.C. 1911. The Keweenawan series of Michigan. Michigan
Geological Survey, Annual Report of the Board of Geological
and Biological Survey (1909), Publication 6, Geological
Series 4.
Lightfoot, P.C., Sutcliffe, R.H., and Doherty, W. 1991. Crustal
contamination identified in Keweenawan Osler Group tholeiites,
Ontario: a trace element perspective. Journal of Geology, 99:
739 - 760.
Mahoney, J.J. 1988. Deccan Traps. In Continental flood basalts.
Edited by J. D. Macdougall . Kluwer Academic Publishers,
Dordrecht, The Netherlands, pp. 15 1 - 194.
Mariano, J., and Hinze, W.J. 1994. Structural interpretation of the
Midcontinent Rift in eastern Lake Superior from seismic reflection and potential-field studies. Canadian Journal of Earth
Sciences, 31: 619-628.
Massey, N.W .D. 1980. The geochemistry of some Keweenawan
metabasites from Mamainse Point, Ontario. Ph.D. thesis,
McMaster University, Hamilton, Ont.
Massey, N. W.D. 1983. Magma genesis in a late Proterozoic protooceanic rift: REE and other trace-element data from the
Keweenawan Mamainse Point Formation, Ontario, Canada.
Precambrian Research, 21 : 8 1- 100.
McGinnis, L. D., and Mudrey , M.G. 1991. Seismic reflection
profiling and tectonic evolution of the Midcontinent rift in Lake
Superior. Wisconsin Geological and Natural History Survey,
Miscellaneous Paper 9 1-2.
McIlwaine, W.H., and Wallace, H. 1976. Geology of the Black
Bay Peninsula area, district of Thunder Bay. Ministry of Natural
Resources, Ontario Division of Mines, Geoscience Report 133.
McKenzie, D., and Bickle, M.J. 1988. The volume and composition of melt generated by extension of the lithosphere. Journal
of Petrology, 29: 625 -679.
Miller, J.D., Jr., Nicholson, S.W., and Cannon, W.F. 1995a. The
midcontinent rift in the Lake Superior region. In Field trip
guidebook for the geology and ore deposits of the Midcontinent
rift in the Lake Superior Region. Edited by J. D. Miller. Minnesota Geological Survey, Guidebook Series, No. 20, pp. 1-22.
Miller, J.D., Jr., Chandler, V.W., Green, J.C. and Witthun, K.
1995b. The Finland tectono-magmatic discontinuity - a growth
fault marking the western margin of the Portage Lake Volcanic
basin of the Midcontinent Rift System. In Basement Tectonics
10, Proceedings of the 10th International Conference on Basement Tectonics, Duluth, Minn., August 1992. Edited by R.W.
Ojakangas, A.B. Dickas, and J.C. Green. Kluwer Academic
Publishers, Dordrecht, The Netherlands, pp. 35 -40.
Nicholson, S.W. 1990. Portage Lake rhyolites of the Midcontinent
rift system, Keweenaw Peninsula, Michigan: geology, petrogenesis, and implications for rift magmatism and mineralization.
Ph.D. thesis, University of Minnesota, Minneapolis, Minn.
Nicholson, S.W., and Shirey, S.B. 1990. Evidence for a Precambrian mantle plume: a Sr, Nd, and Pb isotopic study of the Midcontinent Rift System in the Lake Superior region. Journal of
Geophysical Research, 95: 10 851 - 10 868.
Nicholson, S.W., and Shirey , S.B. 1992. Nd and Pb isotopic evolution of basalts of the 1.1 Ga Midcontinent rift: evidence for a
region-wide model for plume -lithosphere -asthenosphere
interaction. In Workshop on the Application of Isotope Systems
to Geological Problems, Reston, Va., 14- 16 Sept. 1992.
United States Geological Survey, Open-file Report 92-525.
Nicholson, S.W., Schulz, K.J., Cannon, W.F., and Woodruff,
L.G. 1991. The Porcupine Mountains area, Michigan - a
Keweenawan central volcano? Institute on Lake Superior Geology, Annual Meeting, Proceedings and Abstracts, 37: 79 - 8 1.
Nicholson, S.W., Woodruff, L.G., and Cannon, W.F. 1994.
Geochemistry of the Kallander Creek Volcanics, 1.1 Ga Midcontinent rift system, Wisconsin and Michigan. Institute on
Lake Superior Geology, Annual Meeting, Proceedings and
Abstracts, 40: 50 - 5 1.
Paces, J .B. 1988. Magmatic processes, evolution and mantle source
characteristics contributing to the petrogenesis of Midcontinent
rift basalts: Portage Lake basalts, Keweenaw Peninsula, Michigan. Ph.D. thesis, Michigan Technological University, Houghton, Mich.
Paces, J.B., and Bell, K. 1989. Non-depleted sub-continental
mantle beneath the Superior Province of the Canadian Shield:
Nd -Sr isotopic and trace element evidence from Midcontinent
Rift basalts. Geochimica et Cosmochimica Acta, 53: 2023 2035.
Paces, J.B., and Bornhorst, T.J. 1985. Geology and geochemistry
of lava flows within the Copper Harbor Conglomerate,
Keweenaw Peninsula, Michigan. Institute on Lake Superior
Geology, Annual Meeting, Proceedings and Abstracts, 31:
71 -72.
Paces, J.B., and Miller, J.D. 1993. Precise U -Pb ages of Duluth
Complex and related mafic intrusions, northeastern Minnesota:
geochronological insights to physical, petrogenetic, paleomagnetic, and tectonomagmatic processes associated with the 1.1 Ga
Midcontinent rift system. Journal of Geophysical Research, 98:
13997 - 14013.
@ 1997 NRC Canada
Can. J. Earth Sci. Vol. 34, 1997
Palmer, H.C., and Davis, D.W. 1987. Paleomagnetism and U -Pb
geochronology of Michipicoten Island: precise calibration of the
Keweenawan polar wander track. Precambrian Research, 37:
157- 171.
Palmer, H.C., and Halls, H.C. 1986. Paleomagnetism of the
Powder Mill Group, Michigan and Wisconsin: a reassessment of
the Logan loop. Journal of Geophysical Research, 91: 11 571 11 580.
Peng, Z.X., Mahoney, J.J., Hooper, P., Harris, C., and Beane, J.
1994. A role for lower continental crust in flood basalt genesis?
Isotopic and incompatible element study of the lower six formations of the western Deccan traps. Geochimica et Cosmochimica
Acta, 58: 267-288.
Richards, M.A., Duncan, R.A., and Courtillot, V.E. 1989. Flood
basalts and hot-spot tracks: plume heads and tails. Science
(Washington, D.C.), 246: 103 - 107.
Shirey , S.B. 1997. Re - 0 s isotopic compositions of Midcontinent
rift system picrites: implications for plume - lithosphere interaction and enriched mantle sources. Canadian Journal of Earth
Sciences, 34: 489 - 503.
Shirey, S.B., Klewin, K.W., Berg, J.H., and Carlson, R.W. 1994.
Temporal changes in the sources of flood basalts: isotopic and
trace element evidence from the 1100 Ma old Keweenawan
Mamainse Point Formation, Ontario, Canada. Geochimica et
Cosmochimica Acta, 58: 4475-4490.
Sun, S.-S., and McDonough,W .F. 1989. Chemical and isotopic
systematics of oceanic basalts: implications for mantle composition and processes. In Magmatism in the ocean basins. Edited by
A. D. Saunders and M .J. Norry . Geological Society Special
Publication (London), no. 42, pp. 3 13 - 345.
Sutcliffe, R.H. 1987. Petrology of Middle Proterozoic diabases and
picrites from Lake Nipigon, Canada. Contributions to Mineralogy and Petrology, 96: 20 1- 2 1 1.
Thompson, R.N., Morrison, M.A., Dickin, A.P., and Hendry,
G.L. 1983. Continental flood basalts . . . Arachnids rule OK?
In Continental basalts and mantle xenoliths. Edited by
C.J. Hawkesworth and M.J. Norry. Shiva, Nantwich, U.K.
pp. 158- 185.
Turner, S., and Hawkesworth, C. 1995. The nature of the subcontinental mantle: constraints from the major-element composition of continental flood basalts. Chemical Geology, 120:
295 -3 14.
Van Schmus, W.R., and Hinze, W.J. 1985. The midcontinent rift
system. Annual Review of Earth and Planetary Sciences, 13:
345 -384.
Van Schmus, W. R., Bickford, M. E., and Zietz, I. 1987. Early
and middle Proterozoic provinces in the central United States.
In Proterozoic lithospheric evolution. Edited by A. Kroner.
American Geophysical Union, Geodynamics Series, 1,
pp. 43-68.
Van Schmus, W.R., Martin, M.W., Sprowl, D.R., Geisman, J.,
and Berendsen, P. 1990. Age, Nd and Pb isotopic composition
and magnetic polarity for subsurface samples of the 1100 Ma
midcontinent rift. Geological Society of America, Abstracts
with Programs, 22: A174.
Vervoort, J.D., and Green, J.C. 1997. Origin of evolved magmas
in the Midcontinent rift system, northeast Minnesota: Nd-isotope
evidence for melting of Archean crust. Canadian Journal of
Earth Sciences, 34: 52 1 -535.
White, R. S. 1992. Magmatism during and after continental breakup. In Magmatism and the causes of continental break-up.
Edited by B.C. Klewin, T. Alabaster, and R.J. Pankhurst.
Geological Society Special Publication (London), No. 68,
pp. 1- 16.
White, R.S. 1997. Mantle temperature and lithospheric thinning
beneath the Midcontinent rift system: evidence from magmatism
and subsidence. Canadian Journal of Earth Sciences, 34: 464475.
White, R.S., and McKenzie, D. 1989. Magmatism at rift zones: the
generation of volcanic continental margins and flood basalts.
Journal of Geophysical Research, 94: 7685 - 7729.
White, R.S., and McKenzie, D., 1995. Mantle plumes and flood
basalts. Journal of Geophysical Research, 100: 17 543 - 17 586.
White, R.S., Spence, G.D., Fowler, S.R., McKenzie, D.P., Westbrook, G.K., and Bowen, A.N. 1987. Magmatism at rifted
continental margins. Nature (London), 330: 439 - 444.
White, W.S. 1968. The native-copper deposits of northern Michigan. In Ore deposits of the United States, 1933 - 1967 (GratonSales volume). Vol. 1. Edited by J .D. Ridge. American Institute
of Mining, Metallurgy, and Petroleum Engineering, New York,
pp. 303 - 325.
Zartman, R.E., Nicholson, S. W., Cannon, W .F., and Morey , G.B.
1997. U - Th - Pb zircon ages of some Keweenawan Supergroup
rocks from the south shore of Lake Superior. Canadian Journal
of Earth Sciences, 34: 549 -56 1.
@ 1997 NRC Canada