Aug 10, 1994 - CDepartmenf of Earlh Sciences, Memorial &fnivcrsity 0fiV~:~ c ,s;or!ndland, St. John's, i%'jld. ... asthenospheric source, while heterogeneity in the e Ndr La/Sm, ...... 5: OiB=ocean island basalts [data from Sun and McDonough ...
a,a 9
eb9
Tentrc GPoscientijqucde QuPhec-,CGC. 2700 rue Einstein, C.P. 7500, Ste-Fey, t&P. Cl K4C7, Canada bUniversitPdu Q&bee, INRS-Gkoressources I 2700 rue Einstein, C.P. 7500, Ste-Fey, QuP.GI V 4C7, Canada CDepartmenfof Earlh Sciences, Memorial &fnivcrsity0fiV~:~ c ,s;or!ndland,St. John’s, i%‘jld. A IB 3X5, Canada
Received 10 November 1993; revisionaaccepted 10 August 1994
There are two types of pre-Taconian mafic volcanics in t umber Zone of the northern Ap relatively smal!, l,cattered volumes of transitionai to alkaline basalts and hypabyssal intrusives; an though more voluminous, subalkaline tholeiitic volcanics, They formed during a Late Proterozoi zoic period of continental lithospheric extension which led to the break-up of Laurentia and to the opening of the Iapetus ocean, and were subsequently deformed and metamorphosed during the Ordovician Taconian and Late Silurian to iddle Devonian Acadian Orogenies. Subalkaline tholeiitic basalts have major-element and EE abundances similar to that of modern However, they are further characterized by variable [La/ IN (0.7-1.7) and +,d VdWS (+2.8 t0 -!Ma), andTi/Zr, Ta/La, Th/Ta an /Ce ratios forming more-or-less developed Ti, Ta and P troughs on mantlenormalized diagrams. Primitive s lkaline tholeiitic basaltic melts were most likely derived from a depleted asthenospheric source, while heterogeneity in the eNdrLa/Sm, Ti/Zr, Ta/La, Th/Ta and P/Cc values is attributed to variable degrees of assimilation and fractional crystallization of both lower- and upper-crustal material during ascent and eruption. nsitional to alkaline basalts are enriched in moderately to highly incompatible trace elements (k Ta, Th), but unlike tholeiites, they do not show evidence of crustal assimilation. Transitional b x..~~‘YNd vahies ( +3.3 to +4.9) than alkaline basalts ( 12< have less fractionated REE (3 < [La/Yb] N< 3) and lo-z;---=La/Yb]N 75% of such Grenvillian-type material.
5. I. Petrogenesis of tholeiiticbasalts
Any coherent model to account for the genesis of the Shickshock tholeiites has to explain the
tholeiites on primitive mantle-
ites is consistent w rzolite stability tie 1989). Nonetheless, the differences b
with decreasing Mg-numbers or transition-element contents. Other processes such as variations in the degree and/or mechanisms of partial melting (i.e. fractional vs. batch melting) are readily dismissed by the fractionation of trace elements (e.g., Th/Yb vs. Ta/Yb in the variable eNd values, which ca duced by simple variations in the degree of melta single mantle source (Fig. 5 j . in ding in the subcontinental lithosphere, either enriched by silicate alkaline magmas (i.e. lamproite, kimberlite, OIB), or carbonate meta-
G. Camirk et al. / Chemical
nM
Geology I I9 (I 995) 5.5- 77
* TAK basalts ? ? high Th/Ta enr. thnl. o low Th/Ta enr. thol, o deDleted tholeiites
Nb*!Th Fig. 6. A. This diagram of Th/Yb vs. Ta/Yb is used to show the characteristic enrichment of Th relative to Ta produced by crustal assimilation or subduction processes. In this figure, transitional alkaline basalts (T.4K) do not show any evidence of crustal assimilation, whereas the relative Th enrichment observed in the group of tholeiites may be explained by assimilation of an upper-crustal component. Unlike basalts from the oceanic Dunnage Zone [ROI= Bay of Islands ophiolite, data from Jenner et al. ( 1991) and Elthon ( 1991); TM=Thetford Mines ophiolite, data from Oshin and Crocket ( 1986) 1, the Shickshock tholeiites plot away from a calculated mixing line between a depleted mantle (D&f) and a strongly refractory subduction mantle (SM). ( I) =calculated mixing line between DM and SM; percents of mixing are also indicated. (2) =calculated AFC line between Appalachian sedimentary rocks (SED) and a primitive tholeiite (sample 1036, backfractionated); increments of AFC are also indicated. (3) =calculated mixing line between a depleted N-MORB source and an enriched OIB source, (4) =ca!culated fractional melting of a spine1 lherzolite. Sources of data and parameters of calculation are given in the Appendix, except for DM which is the average composition of estimates made by Wood (1979), Fitton and Dunlop ( 1985), Le Roex (1986), and Sun ( 1982). SE&sedimentary rocks from the Humber Zone, unpublished data from G. Camire except for samples 2017 and 4045 presented in Table 1. B. Zr/Y vs. Nb*/Th (Nb*= 17.4xTa and estimates Nb). (I) =calculated mixing line between DM and SM. (2) =calculated AFC line between SED and a primitive tholeiite (sample 1036, backfractionated). i 3) zcalculated AFC line between an estimate of the lower crust and a primitive tholeiite. (4) zcalculated mixing line bet4 .een N-MORB and OIB. (5) zfractional melting of a spine1 lherzolite; triangles to the left-hand side of the curve are % of ~?~~~iing. (6) =fractional melting of a garnet Iherzolite; marks on the right-hand side of the curve are % of melting. Note that curves (5) and (6) are superposed. Sources of data and parameters of calculation are given in the Appendix. See text for discussion.
somatism, or else hydrated by subduction-related processes, is a possible but unlikely explanation for the observed variations in traceelement abundances. For example, if the subcontinental lithosphere source had been enriched by silicate alkaline magmas, then one would expect the Shickshock tholeiites to show variable enrichment in LREE but also depletion in HREE (e.g., Dautria et al., 1992), whereas all Shickshock tholeiites have flat HREE patterns. To produce basalts with N-MORB-type REE patterns and’major-element chemistry like those of the LREE-dep!eted Shickshock tholciites, considerable lithospheric extension and a rather high degree of partial melting are required. Although it is possible to End continental tholeiites
with a degree of LREE enrichment similar to that of some Shickshock tholeiites, most Shickshock tholeiites are so depleted in LREE that no subcontinental lithospheric mantle metasomatized by alkaline magmas can account for their chemical characteristics (e.g., Columbia River, Washington, U.S.A. - Hooper and Hawkesworth, 1993; Karoo Province, South Africa - Sweeney et al., 1994). In the same manner, one might expect large variations of the Zr/Hf ratio in the Shickshock tholeiites, if their mantle source had undergone carbonate metasomatism (e.g., Dupuy et al., 1992 ). Yet, the average Zr/Mf ratio of Shickshock tholeiites is 37.05 2.0 and essentially identical to that of MORB and OIB (46.6 2 2.9; Jochum et al., ! 985 ). Alternatively,
@TAK bosoits = high TWTa e,?r.thol. OlOW Th/To enr. thol, adepleted tholeiites
5.1-l. Asttenospheik geneities
mantle source hetero-
in the field of modern N-
cording to Bennett et al. ( 1993) 1. The eNdvalue for sample 403 7 ( + 4.9 ) is lower than one would expect, but there is no indication in the trace-element pattern of any enriched component, and it iS possible that the low eNdvahe refleCtS a more complex history for the source of this sample. The LREE-enriched tholeiites have some of the characteristics of T- and E-MORB: fractionated LREE patterns and variable &, values. In oceanic intraplate environments, subalkaline basalt sequences displaying various degrees of LREE enrichment and variable eNd Values are common, and are generally considered to be the prodrscts ofa heterogeneo-tis mantle soiirce or of mixing between two (or more) distinct endmember sources (i.e. depleted mantle and enriched OiR components; e.g., Le Roex et al.,
f/3-Th-Ta t~nary diagram of Woe+ et a?. ( 1979) ing a calculated AFC line between Appalachian sedimentary rocks (SED) and a primitive th.oieiite (sample JO.%, backfractionated). Transitional alkaline basalts of the Humber Zone plot within the field of intraplate basalrs near the average composition of E-MORB. On the other hand, tholeiites of the Shickshock Group plot at the limit between the fields of MORB and intraplate basaits away from both groups of TAK basalts and Dunnage Zone basalts. It is noteworthy that whether or not they are low-Ti basalts, subalkaline mafic volcanics from the Dunnage Zone (BOI=Bay of Islands ophiolite; ?W=Thetford Mines ophiolite) plot within the field of volcanic arc basalts and are relatively depleted in Ta when compared with the Shickshock tholeiites. Fields are: rl = MORB; B= within-plate basalts and E-MQRB; C= within-plate basalts; D-volcanic arc basans. N-MORB, .E-MORB and OIB values ace those of the Appendix.
1989 ). However, in contrast to what is observed in oceanic intraplate environments, the L enrichment of the Shickshock tholeiites is various degrees of Ti, P a d, Ta/La is much lo the LREE-enriched Shickshock tholeiites ~0.054, Table 1) than (0.07, Appendix ). The Ached Shickshock t that of the primit cDonough, 1989). Furthermore, Th/ Ta is not positively correlated with TaiLa, so that the possible contribution of an enriched asthen-
ospheric mantle component discarded.
(e.g., plume)
is
5.1.2, Crustal contamination processes Given the range of +,d values in the Shickshock tholeiites (from + 2.8 to + 7.0 ), the complex trace-element variations (e.g., variable Th/ Ta, LREE) and the potential tectonic setting of this volcanism, crustal assimilation processes may offer a likely explanation for the heterogeneity observed in them. In order to test this hypothesis, we have modelled assimilation of different crustal components by primitive Shickshock tholeiites, using incompatible trace elements and ENdvalues. On Fig. 5A, the 6Nd value of the Shickshock basaltic magmas before assimilation is approximated by that of a 540Ma-old N-MORB praduced by partial melting of depleted mantle [ cNd = + 7.8, in the less depleted part of the depleted mantle array of Bennett et al. ( 1993 ) 1. As an estimate of the crustal components potentially involved in assimilation, we have used Late Grenvillian monzonites (Emslie and Hegner, 1993 ), our own analyses of sedimentary rocks, and St. Lawrence and Manicouagan felsic gneisses of the Grenville Province (Dickin and Higgins, 1992 ). To a first order, the sedimentary rocks, which may have been partly derived from the above-mentioned Grenvillian monzonites and felsic gneisses, appear as the most likely contaminant when EN‘,values, Ti/Zr and La/Sm ratios are considered (see Fig. 5B and C). Assimilation and fractional crystallization (AFC) of crusta! components by basaltic melts was also modelled using Nielsen’s ( 199 1) CHAOS 5 @ software (see Appendix for parameters). Results of these calculations are represented in Figs, 6-8. For the modelling, an estimated primary magma (Mg#=0.7 1) was obtained by backfractionation of olivine, clinopyroxene and plagioclase from sample 1036 ( cwd= + 7.0 and [La/ SmlNzO.8 1). This was done because the Shickshock tholeiites have low Mg-numbers ( - 0.52 ) and are not directly representative of primary mantle melts. To produce a liquid with the traceelement composition of the estimated primary magma,
