Geochemistry and petrogenesis of basaltic rocks from Coppermine River area, Northwest Territories

Geochemistry and petrogenesis of basaltic rocks from Coppermine River area, Northwest Territories

J. DOSTAL Department of Geology, Saint Mary's University, Halifax, N.S., Canada B3H 3C3

W. R. A. BARAGAR Geological Survey of Canada, 588 Booth St., Ottawa, Ont., Canada KIA OE4

AND

C. DUPUY Centre Gkologique et Ge'ophysique, USTL, P1. E . Bataillon, 34060 Montpellier, France

Received August 30, 1982 Revision accepted November 26, 1982

Proterozoic basaltic flows (>2000 m thick) and associated dykes and sills from the Coppermine River area, Northwest Territories have chemical compositions typical of continental tholeiites. The low Mg/Fe ratio and abundances of Ni and Cr indicate that the lavas were extensively fractionated prior to extrusion. The variations of incompatible elements such as K, Rb, REE, Y, Zr, Nb, and Th suggest that the rocks were affected by interaction with continental crust. The samples least affected by contamination have trace-element compositions very similar to those of P-type mid-ocean ridge basalts. It is suggested that continental tholeiites have been generated from the same source as P-type oceanic tholeiites, and geochemical features, such as the enrichment of some lithophile elements in many of these rocks, may be related to crustal contamination. The variations within the volcanic pile of the Coppennine River area are related to those of an exposed part of the Muskox layered intrusion. I

Les coulCes basaltiques d'Lge ProtCrozoique (Cpaisseur >2000 m) et les dykes et filons-couches de la rCgion de la rivikre Coppermine, Territoire du Nord-Ouest, posskdent des compositions chimiques typiques des tholCiites continentales. Le faible rapport Mg/Fe et les teneurs en Ni et Cr indiquent que les laves avaient CtC considCrablement fractionnkes avant leur extrusion. Les variations parmi les ClCments incompatibles c o m e le K, Rb, terres rares, Y, Zr, Nb et Th rCvklent que ces roches furent chimiquement contaminCes par une interaction avec la crofite continentale. Les Cchantillons les moins affect& par cette contamination possbdent des compositions en 6lCments traces trks semblables 9 celles des basaltes des chaines mCdio-ocCaniques , du type-P. I1 est suggCrC que les tholCiites continentales et les tholkiites ocCaniques du type-P possbdent une source commune, et que les caractkristiques gCochimiques c o m e l 'e~chissement en quelques ClCments lithophiles de plusieurs de ces roches peuvent dkpendre de la contamination par la crofite. Les variations h I'intCrieur de I'empilement volcanique dans la rkgion de la rivibre Coppermine sont relikes a celles de I'intrusion stratifiee Muskox.

Can. J . Eakth Sci., 20, 684-698 (1983) [Traduit par le journal]

Introduction Some continental tholeiites, including plateau

basalts, have relatively high and variable contents of incompatible elements. Although several processes such as crustal contamination, melting of enriched "plume" mantle source, and "wall-rock reactions have been invoked (e.g., Green and Ringwood 1967; Baragar 1977; Thompson et al. 1980), the causes of this enrichment and variation are not well understood and more data on continental tholeiites are needed to enhance the understanding of their geochemistry and origin. The purpose of this paper is to present geochemical data on Proterozoic plateau basalts and associated dykes and sills from Coppennine River, Mackenzie District, Northwest Territories and to discuss their origin.

Geology of the Coppermine River basalts Geological setting

The geological setting of the Coppermine River

basalts shown in Fig. 1 is based on mapping by Baragar and Donaldson (1973) and modifications by Kerans et a1. (1981) and Kerans (1983).

The Coppermine River Group is part of a gently north-dipping Helikian succession, dated at about 1200 Ma (Baragar 1972), that rests uncomformably upon the eroded surface of the Aphebian Wopmay orogen (Fraser et al. 1972). Underlying units are the Hornby Bay Group of predominantly quartzites at the base of the succession and the Dismal Lakes Group of predominantly dolomites 'immediately below the Coppermine River Group. These record a complex secular change in depositional environment from fluvial initially to shallow marine at the onset of Coppermine River magmatism (Kerans et al. 1981). The Copper- mine River Group and its intrusive equivalents represent a voluminous magmatic episode, the Mackenzie magmatic event (Fahrig and Jones 1969). The group comprises a succession of plateau basalts 2000-3500 m thick (Copper Creek Formation), which interfingers

DOSTAL ET AL.

FIG. 1. Geological map of the Coppermine River area after Baragar and Donaldson (1973) with modifications by C. Kerans (personal communication, 1982). Locations of sections 1 and 6 are shown.

upward into a mixed assemblage of red sandstones and sparse basaltic flows at least 1200 m thick (Husky Creek Formation). Except for the lowermost one or two flows it was emplaced subaerially. Intrusive counterparts include the Muskox intrusion, the Mackenzie dyke swarm, and minor sills intrusive into the Dismal Lakes Group (Dismal Lakes sills). Formerly, following work by Irvine (1970a), the Muskox intrusion was considered to predate the Dismal Lakes Group although it had earlier been thought to be comagmatic with the Coppermine River lavas. Recently, however, reinter- pretation of Irvine's evidence by Hoffman (1980) and new work by Kerans (1983) has re-established the

essential contemporaneity of Muskox and Coppermine River magmatism.

Overlying the Coppermine River Group with gentle discordance are the Rae Group of Hadrynian age and the associated Coronation sills.

Stratigraphic and petrographic relationships The Copper Creek Formation comprises about 150

flows most commonly ranging from 8 to 25 m thick and each typically formed of a massive base and amygdaloidal top. The flows follow one upon the other with no intervening sediments to the top of the formation where they interfinger over a narrow stratigraphic

686 CAN. J . EARTH SCI. VOL. 20, 1983

interval with red fluviatile sandstones of the Husky Creek Formation. Intermittent renewal of volcanism with long periods of quiescence is recorded by lava flows distributed sparsely within the sandstone sequence.

Intrusive members of the Coppermine River suite shown in Fig. 1, the diabase dykes, the Dismal Lakes dolerite sills, and the Muskox intrusion, appear to concentrate in a zone along the axis of the latter. This is also the zone that contains the greatest thickness of lavas and probably represents the source region of the surface flows. The Muskox intrusion is a major layered intrusion (Smith and Kapp 1963; Smith et al. 1967; Irvine and Smith 1967; Irvine and Baragar 1972; Irvine 1970b, 1980) postulated by Irvine (1970b) and Irvine and Baragar, (1972) to be a fractionated magma source for surface eruptions. Because of the time difference thought to separate emplacement of the Muskox intrusion and extrusion of the Coppermine River flows, Irvine (1970b) proposed that the surface counterparts had been removed by erosion. With contemporaneity now re-established by Kerans (1983), on the basis of stratigraphy, the logical surface counterpart of the Muskox intrusion is the volcanic succession of the Coppermine River Group. The Dismal Lakes sills comprise four or five closely spaced sills of short lateral extent within the zone of most intensive intrusive activity. They are cut by Mackenzie dykes and may well be the immediate precursors of the surface eruptions. The dykes themselves penetrate the lower contact of the Copper Creek Formation in great profusion but few emerge from its upper boundary into the Husky Creek Formation. Obviously they are feeders to the Coppermine River flows.

The basaltic flows are composed mainly of augite, plagioclase, and titanomagnetite. Potash feldspar is commonly a minor constituent in the interstices between the other minerals. However, some systematic petrographic variations upward through the Copper Creek Formation are evident. Altered olivine and what are assumed to be pseudomorphs of orthopyroxene are sparsely present in the lower 1000 m but not at higher levels. Augite is the prime phenocryst throughout the sequence and is joined by plagioclase principally above about 1000 m. The modal content of iron oxide increases upward in the succession and native copper appears as a constituent of the flows in the upper half of the formation only. Some of the lowermost flows of the sequence have an extraordinary aspect; granophyric patches and interstitial fillings are major constituents and opaques are predominantly platy ilmenite. In this respect they are greatly similar to the Dismal Lakes sills, strengthening the view that the latter are precursors to the flows. Flows interbedded with the red sediments of the Husky Creek Formation do not appear to be petrographically

continuous with those of the main part of the sequence. The reappearance in them of olivine pseudomorphs and a lower Fe-Ti oxide content than in uppermost Copper Creek flows would suggest a return to magmatic compositions more in keeping with the lower than the upper part of the Copper Creek Formation. The dykes are petrographically similar to the flows except for the presence of sulphides in some of them in place of native copper in the flows.

Metamorphism of both flows and dykes is variable but of low grade. Augite is almost invariably primary but plagioclase is generally replaced to a greater or lesser degree by saussurite or clays. In the lowermost flows and in the Dismal Lakes sills the rocks are extensively altered and few primary minerals survive.

Sampling and analytical notes The samples analyzed for this paper are from

collections systematically taken from all members of the Coppermine River suite in the course of mapping (Baragar and Donaldson 1973) and previous geochemical studies (Baragar 1969). The succession of flows were sampled along six section lines that were surveyed across the Coppermine River Group; the analytical results (rapid analyses; major and trace elements) from five of these sections have been published (Baragar 1969). For the present study samples were selected from two of the sections (sections 1 and 6, Fig. 1) and span the stratigraphic range exposed. In addition, representative samples of the dykes and Dismal Lakes sills are included. The dyke samples are of chilled margins.

The major-element compositions of the samples are given in Table 1 and their trace-element abundances are reported in Table 2. Major elements were determined by wet methods. Li, Rb, Sr, V, Cr, Co, Ni, Cu, and Zn were analyzed by atomic absorption, Ba, Zr, Nb, Y, La, and Ce by X-ray fluorescence, and rare-earth elements (REE), Th, U, Hf, and Sc, by instrumental neutron activation. The precision and accuracy of the atomic absorption and X-ray fluorescence methods were given by Dupuy et al. (1979) and those of neutron activation were reported by Dostal and Capedri (1979). The precision of trace-element data can be expected to be better than 10%.

Geochemistry Major- and some trace-element geochemistry was

previously discussed by Baragar (1969) on the basis of mainly rapid-method major-element analyses and spectrographic trace-element analyses. Those results are supplemented by the more precise analyses done for this study, which are presented in Tables 1 and 2.

Major elements The analyzed flows and dykes have basaltic

compositions with SiOz generally less than 50%; most

TA

BL

E 1. M

ajor

-ele

men

t com

posi

tion

of C

oppe

rmin

e R

iver

bas

alts

and

ass

ocia

ted

rock

s

Lav

as

Sect

ion

1 Se

ctio

n 6

Cop

per C

reek

H

usky

C

oppe

r C

reek

H

usky

D

ykes

Si

lls

Cre

ek

Cre

ek

24-6

6 3-

66

42-6

6 11

5-66

15

7-66

31

-69

3-69

13

-69

17-6

9 29

-69

829

818

741

823

718

716

Si0

2 5

1.5

94

8.8

0

47.2

7 51

.46

49.6

5 53

.74

49.8

4 48

.60

49.3

2 48

.60

49.7

4 47

.93

48.4

1 49

.24

55.3

5 53

.75

Ti0

2 1.

97

1.74

3.

74

2.71

1.

34

2.02

2.

23

3.44

2.

52

3.05

1.

94

3.74

3.

29

2.40

2.

40

2.72

M

2o3

12.0

0 13

.58

11.9

6 12

.24

13.4

4 12

.35

12.7

4 12

.00

12.2

3 12

.57

13.2

9 11

.96

12.1

6 12

.55

12.1

4 12

.38

Fe2

03*

11.8

6 13

.90

17.9

6 16

.28

13.1

9 12

.28

14.2

2 16

.37

16.3

0 15

.77

15.4

0 17

.89

16.9

3 16

.08

12.6

8 14

.05

MnO

0.

14

0.21

0.

24

0.24

0.

20

0.28

0.

26

0.24

0.

26

0.25

0.

21

0.20

0.

27

0.21

0.

18

0.21

M

gO

8.96

6.

84

4.64

4.

83

8.37

5.

56

5.95

5.

30

5.23

5.

74

6.13

5.

30

5.30

6.

23

2.72

3.

40

CaO

5.

38

9.60

8.

60

5.78

8.

85

4.96

7.

25

8.55

7.

31

9.90

10

.00

8.22

8.

22

8.85

4.

58

4.45

N

a20

2.30

2.

18

2.16

2.

93

1.63

2.

40

3.54

2.

19

2.63

2.

12

1.95

1.

82

2.20

1.

85

1.55

2.

83

K20

1.

56

0.68

0.

87

1.68

0.

40

3.37

1.

54

1.05

1.

62

0.65

0.

45

1.37

0.

77

1.33

3.

44

3.57

$

pzos

0.

25

0.17

0.

49

0.39

0.

17

0.24

0.

23

0.42

0.

35

0.37

0.

20

0.51

0.

39

0.29

0.

43

0.34

$

H20

- 0.

39

0.20

0.

16

0.13

0.

63

0.12

0.

15

0.10

0.

08

0.26

0.

10

0.05

0.

04

0.02

0.

13

0.16

>

H

2O+

3.11

1.

66

1.16

1.

81

2.22

2.

18

1.68

1.

16

1.56

0.

57

0.86

0.

59

1.59

0.

74

3.58

1.

22

r Z

99

.51

99.5

6 99

.25

100.

48

100.

09

99.5

0 99

.63

99.4

2 99

.41

99.8

5 10

0.27

99

.58

99.5

7 99

.79

99.1

8 99

.08

[Mgl

62

.3

52.3

36

.6

40.0

58

.5

50.1

48

.3

42.0

41

.7

45.1

47

.4

40.0

41

.1

46.6

31

.8

35.0

Q 6.

2 2.

2 6.

6 6.

2 5.

1 7.

0 6.

8 3.

0 5.

6 5.

1 6.

8 7.

0 4.

2 18

.7

8.2

Or

9.7

4.2

5.3

10.2

2.

5 20

.7

9.4

6.4

9.9

3.9

2.7

8.3

4.7

8.0

21.5

21

.8

Ab

20.5

19

.1

18.9

25

.5

14.3

21

.1

30.9

19

.1

23.0

18

.3

16.8

15

.8

19.2

16

.0

13.9

24

.7

An

18.7

26

.1

21.1

15

.7

29.3

13

.5

14.8

20

.4

17.4

23

.3

26.7

20

.9

21.7

22

.5

16.9

10

.9

Di

6.3

18.1

16

.7

9.5

12.5

8.

7 17

.2

17.0

15

.1

20.2

18

.8

14.8

15

.0

17.1

3.

8 8.

3 H

Y

28.8

21

.5

15.0

20

.4

28.9

19

.2

13.8

15

.1

19.8

15

.1

20.7

17

.3

17.8

21

.1

13.3

13

.6

01

3.1

n 3.

9 3.

4 7.

4 5.

3 2.

7 4.

0 4.

4 6.

7 5.

0 5.

9 3.

8 7.

3 6.

5 4.

7 4.

8 5.

3 M

t 5.

3 4.

9 7.

9 6.

3 4.

3 5.

3 5.

6 7.

4 6.

0 6.

8 5.

1 7.

8 7.

2 5.

8 6.

0 6.

3 AP

0.6

0.4

1.1

0.9

0.4

0.5

0.5

1.0

0.8

0.8

0.5

1.2

0.9

0.7

1.0

0.8

NO

TE

S:

[Mg]

= 1

00M

g/(M

g +

Fez+

); Fe

z+ st

anda

rdize

d as

Fe3

+/Fe

Z+ =

0.1

5.

TA

BL

E 2. T

race

-ele

men

t abu

ndan

ces i

n C

oppe

nnin

e R

iver

bas

alts

and

ass

ocia

ted

rock

s

Lav

as

Sect

ion

1 Se

ctio

n 6

Cop

per

Cre

ek

Hus

ky

Cop

per

Cre

ek

Hus

ky

Dyk

es

Sills

C

reek

C

reek

24

-66

3-66

42

-66

115-

66

157-

66

31-6

9 3-

69

13-6

9 17

-69

29-6

9 82

9 81

8 74

1 82

3 71

8 71

6

WSTAL ET AL.

FIG. 2. AFM (NazO + K20 - FeO,, - MgO) diagram for the Coppennine River lavas (solid circles), dykes (crosses), and sills (empty circles).

are quartz normative. Major-element composition of these rocks is typical of continental tholeiites (Irvine and

I Baragar 197 1; Sukheswala and Poldervaart 1958; Annells 1974; Ghose 1976; Wright et al. 1973). Compared with typical oceanic tholeiites the analyzed rocks have higher contents of Ti02 (1.3-3.7%) and K20 (0.4-3.4%). The Mg numbers (100Mg/(Mg + Fe2+) with Fe2+ standardized as / ~ e ~ + = 0.15) vary between 62 and 37 with most rocks having values of less than 50, indicative of a differentiated character. On the AFM diagram (Fig. 2) most samples fall into the tholeiitic field, showing a typical iron-enrichment trend.

Stratigraphic variations in composition in the Coppermine River basalts are consistent with the changes noted in the petrography. As previously shown (Baragar 1969) Fe and Ti increase and Mg, Cr, Ni, Si, and K all decrease upward in the stratigraphic sequence of the Copper Creek Formation. Flows interlayered with the Husky Creek Formation are again more similar in composition to those low in the Copper Creek Formation than to those immediately below, in its upper part. This is illustrated in Fig. 3 where the Mg number, representing a measure of differentiation, and K20 are 1 plotted against the stratigraphic level for sections 1 and 6. (Note that the Mg number in Fig. 3 represents atomic

I Mg/(Mg + Fe,,,,).) Previously reported analyses (Baragar 1969) and unpublished analyses from section 6

are included. Plotted below the Copper Creek sequence are corresponding values for the Dismal Lakes sills shown in their approximate stratigraphic position. The points to be noted in Fig. 3 are the following.

(1) The Mg number tends to decline upward in the sequence. I

(2) K20 declines rapidly upward in the lower part of the sequence then remains constant or increases slightly

I in its upper part.

(3) The Dismal Lakes sills are possibly precursors to the lowermost flows.

(4) Lava flows in the Husky Creek Formation are at about the same stage of differentiation (Mg number) as are the lowermost flows of the Copper Creek Formation but do not show the same potash enrichment.

Variations in composition of major elements with stratigraphic level in the Copper Creek Formation are generally consistent with fractionation in a high-level magma reservoir. However, the rapid decline in potash content is not, and must be indicative of other influences.

Trace elements The behavior of trace elements in the Coppermine

River suite described below is based on the analyses given in Table 2.

Li varies widely but shows overall positive

690 CAN. J. EARTH SCI. VOL. 20, 1983

Stratigraphic level

metree

H U S K Y CREEK

FORMATION

OPPERCREEK

FORMATION

DISMAL LAKES

FIG. 3. Variation of Mg/(Mg + Fe,J ratio and K20 (analyses recalculated to 100%, water-free) with stratigraphic level in sections 1 and 6. Plotted points are analyses done for this study (diamonds) and for previous studies (Baragar 1969, and unpublished, 1972) (circles). Patterns are the same as for Fig. 1.

correlation with H20. A similar correlation of Li with water has been observed in some ophiolitic complexes and ocean-floor basalts (Dostal et al. 1977) and can be attributed to alteration and (or) low-grade metamorphism. The abundances of Rb and Ba are also highly variable with the K-rich samples (such as 31-69, 3-69, 718, and 7 16) having the highest content. Compared with continental tholeiitic basalts in general (Erlank and Hofmeyer 1968), the high-K rocks are richer in Rb, whereas the other samples are within the reported range

(Nathan and Fruchter 1974; Erlank and Hofmeyer 1968). There is no distinct correlation between the K/Rb ratio and differentiation as represented by Mg number. Except in high-K samples, Sr shows a negative correlation with Ca, indicating that fractionation was dominated by clinopyroxene. This is consistent with the petrography, which shows clinopyroxene as phenocrysts throughout the sequence and plagioclase mainly in the upper part. The relatively immobile incompatible elements (Zr, Hf, Nb, Th, Y, La, and Ce) also show

DOSTAL ET AL.

FIG. 4. REE abundances normalized to chondrites (Frey et al. 1968) in lavas and sills.

large variations in the lavas and dykes, with an overall increase with increasing degree of differentiation. Their concentrations are well within the range reported for continental tholeiitic basalts (Nathan and Fruchter 1974). The ratios involving trace elements of differing degrees of incompatibility such as La/Zr and Th/La are varied. Th/La ratios in the analyzed samples are high (B0.1) and comparable to continental tholeiites described from Australia (Frey et al. 1978).

Overall, the highest contents of the incompatible elements Rb and Ba, like that of K (Fig. 3), are in the lowermost flows of both sections and the Dismal Lakes sills located beneath them. This is undoubtedly attributable to their high content of granophyre as noted earlier. Samples from flows interlayered with red beds at the top of both sequences (157-66 and 29-69) are among the lowest in incompatible element abundances. In detail, the variation is complex, and samples enriched or depleted in lithophile elements occur at various levels of both sequences. To a certain degree the variations seem to be independent of the degree of differentiation and

1 there is no obvious compositional difference between the two sections.

1 REE abundances in several representative samples normalized to chondrites (Frey et al. 1968) are shown in

Fig. 4. REE patterns of the Coppermine River dykes and lavas can be subdivided into two groups according to their La/Yb ratios. The first type has relatively low light-REE (LREE) contents with a low La/Yb ratio (2.9-4.2) and only slightly fractionated heavy-REE (HREE) (samples 3-66, 157-66, and 829). The REE pattern of these rocks resembles the P-type mid-ocean ridge basalts (MORB) (Sun et al. 1979) and LREE-enriched tholeiites from Iceland (Schilling et al. 1978). The second type has higher LREE abundances, higher La/Yb ratios (5.8- 10.3), and fractionated HREE. In general, the REE patterns and contents of these rocks are intermediate between typical oceanic island alkali basalts (Shibata et al. 1979) and some Hawaiian (Leeman et al. 1980) as well as typical continental tholeiites (standard rock BCR-1, Herrmann 1970).

The content of transition elements is variable and is related to the degree of differentiation. Like Ti, V increases with differentiation. The Dismal Lakes sills with low Mg numbers are, however, low in V, which could reflect their high content of ilmenite rather than magnetite. The Ti/V ratio, which increases with differentiation, generally varies between 30 'and 60, the values intermediate between those of MORB and alkali basalts (Langmuir et al. 1977). With the exception of

692 CAN. I. EARTH SCI. VOL. 20, 1983

sample 24-66, Ni and Cr in the analyzed rocks are below 100 and 500 ppm, respectively, reflecting the fractionated character of the samples. These two elements exhibit positive correlation with the Mg numbers. The significantly steeper decrease of Cr with differentiation as compared with Ni indicates dominant pyroxene fractionation in agreement with Ca-Sr intercorrelation and consistent with the presence of clinopyroxene phenocrysts throughout the sequence. Crystallization of olivine does not appear to have had a great influence on the variations in chemical composition of lavas at the time they were emplaced, as is also suggested by the rather constant Co abundances in most samples. Petrographically this is borne out by the rarity of altered olivine in the lavas. The Zn content ranges between 100 and 200ppm and increases with differentiation. Compared with typical tholeiitic basalts (Andriambolo- lona and Dupuy 1978) the abundances of Cu are on average very high. Cu varies between 37 and 700ppm and exhibits an unusual regular increase with differentiation.

Discussion Several processes are discussed below to evaluate

whether or not they could account for the observed compositional variations in the Coppermine River rocks. They include low-grade metamorphism and (or) alteration, fractional crystallization, partial melting of upper mantle, mixing of magmas, and crustal contamination.

Low-grade metamorphism or alteration The Coppermine River lavas are affected by alteration

that could have changed the original chemical composition. Baragar (1969) noted changes in composition from flow interiors to the highly altered tops, including a decrease of A1 and Sr and an increase in COz. The observed variation of Li and its correlation with H20+ can probably also be attributed to secondary processes. However, the secondary compositional changes are probably rather limited. The lavas were collected from the interiors of the flows and are among the freshest samples studied by Baragar (1969). Their compositions are also closely comparable to those of the associated relatively fresh dykes and indicate that diagenetic alteration did not affect the distribution of most elements. Likewise, the good correlation between immobile and some mobile elements suggests that the secondary processes did not alter significantly the composition of the studied rocks. This is also indicated by an overall positive correlation of the ratios of elements of variable degrees of mobility such as La/Yb or Zr/Y with K/Ba or Rb/Sr.

the samples indicate that the rocks were extensively fractionated prior to extrusion. Fractionation dominated by pyroxene separation continued throughout the eruptive history of the Copper Creek Formation. However, the substantial compositional variations among the studied rocks are difficult to reconcile with simple fractional crystallization as a sole process relating the samples to each other and to a single parental magma. This mechanism cannot explain the differences between the two types of REE patterns as fractional crystallization cannot effectively fractionate LREE from HREE without having a strong effect on major and transition elements (cf. Frey et al. 1974). There are other features that are inconsistent with such a process. They include substantial variation in P and Ti and large differences in element ratios such as La/Yb or Zr/Y in the rocks of similar Mg numbers.

Partial melting Two types of REE patterns observed in the

Coppermine River rocks differ in slope. The patterns of the least differentiated rocks (with lower slopes) cross at the HREE end so that rocks with the highest LREE content have the lowest Lu abundances. Such a diversity of slopes can result from variable degrees of partial melting of the same source provided that the proportions of garnet relative to olivine and pyroxenes differed in the source and (or) melt.

In order to generate such parental magmas with a degree of melting 110% (Green and Ringwood 1967) the upper mantle source would have to have a chondrite-normalized pattern enriched in LREE. HOW- ever, such a mechanism cannot account for the differences in ratios of trace elements of the same degree of incompatibility (e. g., Nb /La or Th/La), which vary among the analyzed rocks by factors of up to two.

Dynamic melting The dynamic melting model was originally postulated

by Langmuir et al. (1977) to account for the significant variations of lithophile elements in a volcanic pile that exhibited relatively constant abundances of transition elements. This mechanism was applied by Wood et al. (1979), Strong and Dostal (1980), and Dupuy et al. (1981) to explain the element variations in ophiolites and oceanic and continental tholeiites. The dynamic partial melting of a rising upper mantle diapir can produce melts with REE patterns comparable to those observed in the studied lavas and dykes. However, the large observed variations of several trace-element ratios such as Th/La or Nb/La cannot be produced by this process since it cannot markedly alter ratios of elements

Fractional crystallization kith the same degree of incompatibility (Sun et al. The low Mg numbers and abundances of Ni and Cr in 1979).

DOSTAL ET AL.

FIG. 5. Chondrite-normalized abundances of incompatible elements in lavas 24-66 (solid circles), 157-66 (empty circles), and 3-66 (stars). The mix (solid squares) of samples 24-66 and 157-66 in the proportions 17:83 to approximate the content of sample 3-66 is shown for comparison. Elements are arranged according to degree of incompatibility of Sun et al. (1979). Normalizing values are also from Sun et al. (1979).

Mixing The process of mixing of magmas can explain the

observed variations including the differences in some element ratios (e.g., Th/La, K/Ba) that are difficult to fractionate by differentiation and melting processes. A binary mixing model was tested by using the simple mixing equation of Allbgre and Minster (1978). For the calculation of the element concentration in the intermediate members, the lavas with the highest Mg numbers and large differences in the abundances of lithophile elements were taken as two mixing components. Sample 24-66 is from near the base of the Copper Creek flows within the region of high potash content, whereas 157-66 is from the Husky Creek Formation where the flows appear to belong to a new cycle of primitive magma. The results of such a mixing are graphically shown in Fig. 5, normalized to chondrites. It shows that a composition with element ratios similar to that of sample 3-66 may be produced by mixing a high-K sample (24-66) with one relatively depleted in lithophile elements (157-66) in the proportions of 17:83. Whether or not the other rock compositions can be produced by such a process may be tested by comparing the variations of element ratios that minimize the effects of fractional crystallization. In Fig. 6, where La/Y ratios are plotted against La/Zr, the

analyzed samples plot along a straight line, suggesting that the composition of most of the samples can be explained by simple mixing. In fact, it seems that a mixing process can explain the abundances of trace elements and their ratios.

Origin of magmas I

A mixing model implies the presence of at least two magmas with different compositions. Two relatively I

primitive rocks that were used as representatives of the two mixing magmas have similar Mg numbers but significantly different contents of lithophile elements. These differences cannot be produced by simple models of fractional crystallization or partial melting from a I

common parent. The abundances and relative distribution of incom-

patible elements, except for slightly enriched K in sample 157-66 (Fig. 7), are closely comparable to the P-type MORB (Sun etal. 1979; Wood etal. 1979). Such a distribution of incompatible elements in the melt can be produced by partial melting of the upper mantle. I However, the generation of magma with a composition similar to sample 24-66 by partial melting would require a significantly different upper mantle source. Such a parent would have had to be highly enriched in most lithophile elements but depleted in Nb or this element

CAN. J. EARTH SCI. VOL. 20, 1983

1 l l l l l l l l l , l l l l l l ,

5 10 15 20

I O O x La/Zr

FIG. 6. La/Y vs. La/Zr diagram for the Coppermine River basaltic rocks of section 1 . Lavas (solid circles), dykes (crosses), sills (empty circles), and average of Precambrian quartz-feldspathic rocks from Baffin Island (Shaw et al. 1976) (solid square). The linear trend is consistent with the mixing of two end members.

would have had to be retained in residual phases. However, the presence of several different sources seems unlikely considering the close spatial and temporal association of the rocks. Thus, assuming that magmas were indeed derived from a single upper mantle source, the observed compositional differences require that the magma was affected by a secondary process such as contamination. In fact, the enrichment of K and some other incompatible elements in sample 24-66 may be due to crustal contamination.

There are several indications in support of this process. The sample has element ratios such as Rb/Sr, Ba/La, and Th/La typical of crustal rocks and has a chondrite-normalized pattern for most lithophile elements similar to that of the continental crust even in its depletion of Nb (Fig. 7). Likewise, on a graph such as Fig. 6 where two element ratios are plotted against each other, the compositions of many crustal rocks lie on an extension of the trend of the studied rocks. Stratigraphically, the samples with high contents of K and related elements in each section (24-66 and 3 1-69) are from the lowermost flows. Furthermore, two samples from the Dismal Lakes sills immediately beneath the flow sequence are even more enriched in these elements than samples from the basal zones. The petrography of the lowermost lavas and of the Dismal

Lakes sills is not only consistent with an origin by contamination but is highly suggestive of it. As noted previously, both contain extraordinary amounts of granophyre, which, in its relationship to the textural framework of the rock, has the aspect of a foreign substance.

To relate these results to the geology observed in the field it is necessary to take into account the possible role of the Muskox intrusion in the evolution of the lava sequence.

Irvine (1970b) has shown that the layered sequence of the Muskox intrusion is divisible on the basis of order of crystallization into essentially two stratigraphic succes- sions. The lower succession is marked by the crystallization order olivine, clinopyroxene, plagioclase, and orthopyroxene (intercumulus) and the upper succession by the order olivine, orthopyroxene, clinopyroxene, and plagioclase. Layers resulting from these crystallization orders form sequences, which, from bottom to top, are in the order of: (1) dunite, olivine clinopyroxenite, and gabbro, and (2) peridotite, orthopyroxenite, websterite, and gabbro, respectively. Full or partial sequences are repeated a number of times throughout each succession, a feature attributed by Irvine to periodic recharge of the magma chamber from a common primitive source. With each recharge, the

DOSTAL ET AL.

FIG. 7. Chondrite-normalized abundances of incompatible elements in lavas 24-66 (solid circles) and 157-66 (empty circles) and an average of the Precambrian Canadian Shield (Shaw et al. 1976) (solid squares). Normalizing values are from Sun et al. (1979).

' magma displaced is assumed to have been expelled to the surface as lava flows. The upper margin of the

, intrusion is a thin, irregular, granophyric layer largely ' or entirely derived by partial melting of the country

rock. The change in the order of crystallization during

deposition of the layered series has been shown by Irvine (1970b) to be achievable by sialic contamination of the original magma. Double diffusive convection within the magma chamber as recently postulated by k i n e (1980) would greatly increase the effectiveness of distribution of the contaminant from the roof to deep within the chamber.

The Coppermine River flows, therefore, as the surface counterparts of the Muskox intrusion can be expected to show the effects of both fractionation and

contamination. With this in mind the following sequence of events can be envisaged. (1) Emplacement of the proto-Muskox intrusion with

accompanying partial melting of the host and collection of the resulting sialic melt at the roof of the intrusion.

(2) Enlargement of the magma chamber with addition of new magma and partial discharge of contaminated roof magma to higher levels; the Dismal Lakes sills and first-formed flows at surface.

(3) Introduction of a succession of rapidly repeating refills and discharges such that the effects of contamination are minimized in the corresponding lavas at surface.

(4) Reduction in the rate of refill and discharge with some accompanying increase in contamination evident in flows near the top of the Copper Creek Formation.


(The effects of fractional crystallization are complicated by the repeated refills of the magma chamber but are progressive overall and manifested in the changing composition of the growing lava pile.)

(5) Cessation of eruptions through the Muskox intrusion and its eventual solidification.

(6) Intermittent renewal of magmatism from more primitive sources and interlayering of the resulting lavas with Husky Creek sediments.

Conclusion The chemical compositions of the lavas and dykes of

the Coppermine River complex indicate that the rocks were extensively fractionated and also were affected by interaction with continental crust. Many of the observed variations may be explained by the mixing of magmas with strongly contrasting degrees of sialic contamination or by variable degrees of contamination of a single magma. The mixing of upwelling basaltic liquid with granitic melt extracted from the crustis probably not the dominant mechanism of contamination although a certain amount of granitic partial melt was probably mixed with the magmas. The large variations of K and related elements in basalts would require incorporation of significant amounts of granitic material while the variation of Si02 is relatively small. It seems more likely that contamination of basaltic magma took place largely by diffusion of selective elements. Figure 7 indicates that highly mobile elements such as K and Rb were more affected than relatively immobile elements such as Zr. This is in agreement with the data of Fratta and Shaw (1974) and Dostal and Fratta (1977) who documented a similar process on a local scale in a Precambrian dyke of continental tholeiites. This mechanism has also been invoked in studies of other continental lavas (e.g., Thompson 1975; Thompson et al. 1980; McDougall 1976; Leeman et al. 1976; Moorbath and Welke 1969; Moorbath and Thompson 1980).

The geochemical characteristics of the studied lavas and dykes are typical of continental tholeiites. They exhibit a relatively wide compositional range in which samples with the lowest abundances of incompatible elements are similar to oceanic tholeiites. For example, lava 157-66, which was affected only by rather limited crustal contamination, has a chondrite-normalized pattern of incompatible elements (Fig. 7) very similar to that of P-type mid-ocean ridge basalts (Sun et al. 1979; Wood et al. 1979), even in its relative depletion of Th and Rb. This suggests that continental tholeiites could well have been generated from the same source as the P-type MORB, and their different geochemical features such as the enrichment of some lithophile elements in many of these rocks may be related to crustal contamination as recently suggested by Allbgre et al. (1981). Consequently, it is possible that the continental

upper mantle contains abundances of incompatible elements and their element ratios are comparable to the upper mantle source for the P-type MORB.

Acknowledgements We thank Dr. M. Lambert for his critical comments.

ALLBGRE, C. J., and MINSTER, J. F. 1978. Quantitative models of trace element behaviour in magmatic processes. Earth and Planetary Science Letters, 38, pp. 1-25.

ALLBGRE, C. J., DUPRE, B., LAMBRET, B., and RICHARD, P. 1981. The subcontinental versus suboceanic debate, I. Lead-neodymium-strontium isotopes in primary alkali basalts from a shield area: the Ahaggar volcanic suite. Earth and Planetary Science Letters, 52, pp. 85-92.

ANDRIAMBOLOLONA, R., and DUPUY, C. 1978. Repartition et comportement des tlkments de transition dans les rochcs volcaniques. 1: Cuivre et zince. Butletin du Bureau de Recherches Gtologiques et Min ihs , 2, pp. 121-1 38.

ANNELLS, R. N. 1974. Keeweenawan volcanic rocks of Michipicoten Island, Lake Superior, Ontario. GeoIogical Survey of Canada, Bulletin 21 8, 141 p.

BARAGAR, W. R. A. 1969. The geochemistry of Coppermine River basalts. Geological S w e y of Canada. Paper 69-44, 43 p.

1972. Coppermine River basalts: geological setting and interpretation. In Rubidium-strontium isochron age studies, Report 1. Edited by R. K . Wanless and W. D. Loveridge. Geological Survey of Canada, Paper 72-23, pp. 21-24.

1977. Volcanism of the stable crust. In Volcanic regimes in Canada. Edited by W. R. A. Baragar, L. C. Coleman, and J. M. Hall. Geological Association of Canada, Special Paper No. 16, pp. 377-405.

BARAGAR, W. R. A., and DONALDSON, J. A. 1973. Coppermine and Dismal Lakes map-areas. Geological Survey of Canada, Paper 71-39, 20 p.

DOSTAL, J., and CAPEDRI, S. 1979. Rare-earth elements in high-grade metamorphic rocks from the western Alps. Lithos, 12, pp. 41-49.

DOSTAL, J., and FRATTA, M. 1977. Trace element geochemistry of a Precambrian diabase dike from western Ontario. Canadian Journal of Earth Sciences, 14, pp. 2941-2944.

DOSTAL, J., DUPUY, C., and CAPEDRI, S. 1977. K, U, and Li abundances in ultramafic rocks of the Tethyan ophiolites. Tschermaks Mineralogische und Petrographische Mittei- lungen, 24, pp. 161-168.

DUPUY, C., DOSTAL, J., and COULON, C. 1979. Geochemistry and origin of andesitic rocks from N.W. Sardinia. Journal of Volcanology and Geothermal Research, 6, pp. 375-389.

DUPUY, C., DOSTAL, J., and LEBLANC, M. 1981. Geochemistry of an ophiolitic complex from New Caledonia. Contributions to Mineralogy and Petrology, 76, pp. 77-83.

ERLANK, A. J., and HOFMEYER, P. K. 1968. K/Rb ratios in Mesozoic tholeiites from Antarctica, Brazil and India. Earth and Planetary Science Letters, 4, pp. 33-38.

FAHRIG, W. F., and JONES, D. L. 1969. Paleomagnetic

DOSTAL ET AL. 697

evidence for the extent of Mackenzie igneous events. Canadian Journal of Earth Sciences, 6, pp. 679-688.

FRASER, J. A., HOFFMAN, P. F., and IRVINE, T. N. 1972. The Bear Province. In Variations in tectonic styles in Canada. Edited by R. A. Price and R. J. W. Douglas. Geological Association of Canada, Special Paper 11, pp. 454-503.

FRATTA, M., and SHAW, D. H. 1974. 'Residence' contamination of K, Rb, Li, and TI in diabase dikes. Canadian Journal of Earth Sciences, 11, pp. 422-429.

FREY, F. A., HASKIN, M. A., POETZ, J. A., and HASKIN, L. A. 1968. Rare earth abundances in some basic rocks. Journal of Geophysical Research, 73, pp. 6085-6098.

FREY, F. A., BRYAN, W., and THOMPSON, G. 1974. Atlantic ocean floor: geochemistry and petrology of basalts from Legs 2 and 3 of the Deep-Sea Drilling Project. Journal of Geophysical Research, 79, pp. 5507-5527.

FREY, F. A., GREEN, D. H., and ROY, S. D. 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilites from southeastern Australia utilizing

, geochemical and experimental petrological data. Journal of Petrology, 19, pp. 463-513.

GHOSE, N. C. 1976. Composition and origin of Deccan basalts. Lithos, 9, pp. 65-73.

GREEN, D. H., and RINGWOOD, A. E. 1967. The genesis of basaltic magmas. Contributions to Mineralogy and Petrology, 15, pp. 103-190.

HERRMANN, A. G. 1970. Yttrium and lanthanides. In Handbook of geochemistry. Vol. 11-2. Edited by K. H. Wedepohl. Springer-Verlag, New York, NY, pp. 57-71.

HOFFMAN, P. F. 1980. On the relative age of the Muskox intrusion and the Coppermine River basalts, District of Mackenzie. In Current research, part A. Geological Survey of Canada, Paper 80-l A, pp. 223-225.

IRVINE, T. N. 1970a. Geological age and structural relations of the Muskox intrusion. In Report of activities, part A. Geological Survey of Canada, Paper 70-l A, pp. 145- 153.

1970b. Crystallization sequences in the Muskox intrusion and other layered intrusions. 1. Olivine-pyroxene-plagioclase relations. Geological Society of South Africa, Special Publication 1, pp. 441-476.

1980. Magmatic infiltration metasomatism, double- diffusive fractional crystallization and adcumulus growth in the Muskox intrusion and other layered intrusions. In Physics of magmatic processes. Edited by R. B. Hargraves. Princeton University Press, Princeton, NJ.

IRVINE, T. N., and BARAGAR, W. R. A. 1971. A guide to the chemical classification of the common volcanic rocks. Canadian Journal of Earth Sciences, 8, pp. 5 13-548.

1972. Muskox intrusion and Coppermine River lavas, Northwest Temtories, Canada. 24th International Geological Congress, Montreal, P.Q., Field Excursion A29 Guidebook.

WINE, T. N., and SMITH, C. H. 1967. The ultramafic rocks of the Muskox intrusion. In Ultramafic and related rocks. Edited by P. J. Wyllie. John Wiley and Sons Inc., New ' York,NY.

I KERANS, C. 1983. Timing of emplacement of the Muskox ' intrusion: constraints from Coppermine homocline cover I strata. Canadian Journal of Earth Sciences, 20, this issue.

KERANS, C., ROSS, G. M., DONALDSON, J. A., and

GELDSETZER, H. J. 1981. Tectonism and depositional history of the Helikian Hornby Bay and Dismal Lakes Groups, District of Mackenzie. In Proterozoic basins of Canada. Edited by F. H. A. Campbell. Geological Survey of Canada, Paper 81-10, pp. 157-182.

LANGMUIR, C. H., BENDER, J. F., BENCE, A. E., HANSON, G. N., and TAYLOR, S. R. 1977. Petrogenesis of basalts from the Famous area: Mid-Atlantic ridge. Earth and Planetary Science Letters, 36, pp. 133-156.

LEEMAN, W. P., VITALIANO, C. J., and PRINZ, M. 1976. Evolved lavas from the Snake River plain: craters of the Moon National Monument, Idaho. Contributions to Mineralogy and Petrology, 56, pp. 35-60.

LEEMAN, W. P., BUDAHAN, J. R., GERLACH, D. C., SMITH, D. R., and POWELL, B. N. 1980. Origin of Hawaiian tholeiites: trace element constraints. American Journal of Science, 280A, pp. 794-819.

MCDOUGALL, I. 1976. Geochemistry and origin of basalt from the Columbia River Group, Oregon and Washington. Geological Society of America Bulletin, 87, pp. 777-792.

MOORBATH, S., and THOMPSON, R. N. 1980. Strontium isotope geochemistry and petrogenesis of the early Tertiary lava pile of the Isle of Skye, Scotland and other basic rocks of the British Tertiary igneous province. Journal of Petrology, 21, pp. 295-321.

MOORBATH, S . , and WELKE, H. 1969. Lead isotope studies on igneous rocks from the Isle of Skye, northwest Scotland. Earth and Planetary Science Letters, 5, pp. 217-230.

NATHAN, S., and FRUCHTER, J. S. 1974. Geochemical and paleomagnetic stratigraphy of the Picture Gorge and Yakima basalts (Columbia River Group) in central Oregon. Geological Society of America Bulletin, 85, pp. 63-76.

SCHILLING, J. G., SIGURDSSON, H., and KINGSLEY, R. H. 1978. Skagi and western neovolcanic zones of Iceland: 2. Geochemical variations. Journal of Geophysical Research, 83, pp. 3983-4002.

SHAW, D. H., DOSTAL, J., and KEAYS, R. R. 1976. Additional estimates of continental surface Precambrian shield composition in Canada. Geochimica et Cosmochimica Acta, 40, pp. 73-83.

SHIBATA, T., THOMPSON, G., and FREY, F. A. 1979. Tholeiitic and alkali basalts from the Mid-Atlantic ridge at 43"N. Contributions to Mineralogy and Petrology, 70, pp. 127-142.

SMITH, C. H., and KAPP, H. E. 1963. The Muskox intrusion, a recently discovered layered intrusion in the Coppermine River area, Northwest Temtories, Canada. Mineralogical Society of America, Special Paper 1, pp. 30-35.

SMITH, C. H., IRVINE, T. N., and FINDLEY, D. C. 1967. Muskox intrusion. Geological Survey of Canada, Map 1213A.

STRONG, D. F., and DOSTAL, J. 1980. Dynamic melting of Proterozoic upper mantle: evidence from rare earth elements in oceanic crust of eastern Newfoundland. Contributions to Mineralogy and Petrology, 72, pp. 165-173.

SUKHESWALA, R. N., and POLDERVAART, A. 1958. Deccan basalts of the Bombay area, India. Geological Society of America Bulletin, 69, pp. 1475- 1494.

SUN, S. S., NESBITT, R. W., and SHARASKIN, A. YA. 1979.


Geochemical characteristics of mid-ocean ridge basalts. Earth and Planetary Science Letters, 44, pp. 119-138.

THOMPSON, R. N. 1975. Primary basalts and magma genesis. 11. Snake River plain, Idaho, U.S.A. Contributions to Mineralogy and Petrology, 52, pp. 213-232.

THOMPSON, R. N., GIBSON, I. L., MARRINER, G. F., MATTEY, D. P., and MORRISON, M. A. 1980. Trace-element evidence of multistage mantle fusion and polybaric fractional crystallization in the Palaeocene lavas of Skye, NW Scotland. Journal of Petrology, 21, pp. 265-293.

WOOD, D. A., TARNEY, J., VARET, J., SAUNDERS, A. D., BOUGAULT, H., JORON, J. L., TREUIL, M., and CANN, J. R. 1979. Geochemistry of basalts drilled in the North Atlantic by IPOD Leg 49: implications for mantle hetero- geneity. Earth and -planetary Science Letters, 42, pp. 77-97.

WRIGHT, T. L., GROLIER, M. J., and SWANSON, D. A. 1973. Chemical variation related to the stratigraphy of the Columbia River basalt. Geological Society of America Bulletin, 84, pp. 371-386.

Geochemistry and petrogenesis of basaltic rocks from Coppermine River area, Northwest Territories

Documents