The geochemistry of Lower Proterozoic siliciclasticturbidites from the Rombak Window: implicationsfor palaeogeography and tectonic settingsEDWARD W. SAWYER & ARE KORNELlUSSEN

Sawyer, E.W. & Korneliussen. A. 1989: The geoc hemistry of lower Proter ozoic siliciclastic tur­bidites from the Rombak Window : implications for palaeogeography and tecton ic settings . Nor.geol. unoers. Bull. 415. 23-38.

Lowe r Proterozoic metamorpho sed. suicictasuc turbidites at Rombaksbotn and Gautelis in thewestern part of the Rombak Window have a high sand/clay rat io. whereas those from Ruvsso t inthe east have a larger silt- and clay-size component. The source rocks to the turbidites exper ien­ced little chemical weathering , and the presence of unweathered feldspar in the clay-sized fract ionof the Gautelis and Rombaksbotn turbidites suggests either a contr ibution of material directly fromvolcanic eruptions or weathering in a cold climate. The presence of thin calc-alkali volcanic layersin the turbidites at Gautelis and Rombaksbotn favours the volcanic source .

The Gautelis and Rombaksbotn bulk sediments contain both a large component of materialderived from a high K,O calc-alka line intermediate to tetstc volca nic suite (SN volcan ites), and asmall component «5%) chrom ite-bearing ultramafics. Howe ver. the character ist ic smooth REEpatterns of the Gautelis turb idites (and their lower K,O and Rb contents) indicate an additionalcomponent of materiat derived from a tonalitic sourc e, similar to that exposed near Gautelisvatnetand believed to be the baseme nt to the supracrustals. The large component of SN votcanites isconsistent with formation of the Rombaksbotn and Gautelis turbid ites in an active marginal basinadjacent to a mature volcanic arc. In the case of the Gautelis turbidites the associated volcanicarc formed on continental crust, and an active continental marg in sett ing of Andean-type is proposed.

The geochemical signature of the Ruvssot turbidites indicates that they formed in a basin thatreceived material from maf ic to intermediate sources. but not from evolved teisic volcanic or cont i­nental source s. An intra-ocea nic selt ing near to primiti ve calc-a lkaline volcanism is therefore sugge­sted for the Ruvssot turbidite basin.

Compos itional fractionation due to sorti ng in the grad ed turb idite beds is considerable . TheRombaksbotn and Gautelis pelites are enriched in a variety of major and trace elements and theHREE relative to the greywackes . similar , thoug h less pronounced features are seen in the Ruvs­sot turbidites.

E.W. Sawyer, Sciences de te Terre, Universite du Quebec a Cnicoutimi, Cnicoutimi, Quebec G7H2B l . Canada.A. Kornettussen, Norges ge%giske unaerseketse, P.O.Box 3006-Lade. N-7002 Trondheim, Norway.

IntroductionThe compos ition of any part icular clast ic sedi­ment can, in general, be related to the cumula­tive effect of four principal factors:

1) Bulk composition of the provenance area;this is determined by the rock types presentand their relative abundances.

2) The degree of chemical weathering thatoccurred in the provenanc e area, includingweathering during sediment transport.

3) Sedimentary processes occurring duringsediment transport and deposition.

4) Post-depositional modification by elementredistribution during diagenesis, metamor­phism or deformation.

Thus, in order to use the compos itions ofancient clastic sediments to infer the characte­ristics of provenance, areas, such as theirtectonic setting (Bhatia 1983, Bhatia & Crook1986), or the rock types present (Jenner etal. 1981), the compos itional variations due tothe other three factors must be considered ,as pointed out by Bhatia (1983), Reimer (1985,1986), Sawyer (1986a) and Argast & Donnelly(1987).

Some degree of chemical weathering is tobe expected in all clastic sediments as weath­ering is the principal means of rock disaggrega ­tion. However, there are some except ions

24 Edward W.Sawyer & Are Korne liussen

such as island arc volcaniclastics where vol­canic fragmentation has a dominant role (Crookpers comm, 1988). Nesbitt et al. (1980) haveshown that the relative degree of chemicalweather ing between samples can be estimatedbecause of the chemical fractionation betweencertain elements (e.g. CaO and K:O) increaseswith the degree of weathering. Similarly, bysampling all the grain-size fractions in theircorrect abundances, and by using sedimenttypes that have not been extensively rework­ed, the modification to the bulk compos itionof disintegration products due to sedimentaryprocesses can be recognised (Sullwold 1961,Reimer 1985, Sawyer 1986a). The post-deposi­tional redistribution of elements is, however,more difficult to assess , but by sampling toexclude clearly altered rock s from near oredepos its, fractures and shear zones, and byconsidering the less mobile elements such asTiO:, AI:O" REE, Th, Zr, Y, Sc and Cr, theproblem may be lessened (Nesbitt 1979, Rei­mer 1986, Sawyer 1986a, Bhatia & Crook1986). Argast & Donelly (1 987) compared thecompos itions of North American Palaeozoicturb idites with unlithified turb idites from theBlack Sea, and concluded that little, if any,major element mobility occurred during tur­bidite diagenesis.

The purpose of this study is to use thegeochemistry of Lower Proterozo ic rnetatur­bidites from three localities to investigate theEarly Proterozoic geological evolution of theRombak Window in northern Norway (Fig. 1).However, before the contribution of sourcerock types to the sedimentary detritus is evalu­ated, the effects of other processes on sedi­ment composition, most notably chemicalweathering and sediment transport/deposition,are considered. The composition of the tur­bidites is then compared to the poss ible sour­ce rocks found in the Rombak Window, anddifferences between the three turb idite sequen­ces discussed in terms of their tectonic andgeographic settings during the Early Proterozo ­ic.

Geological Sett ingThe Rombak Window is a slightly elongatedculmination of Lower Proterozoic rocks thatare exposed within the Caledonian Nappesequence 20 km southeast of Narvik (Fig. 1).Much of the window consists of large plutons

NGU · BULl.415. 1989

ROMBAK WINDOW

Fig. 1. Geological map of the Rombak Window . in partafter Birketanc (1976) with modifications based on the pre­sent writers ' field mapping. Sample locations of the Rom­bak turbidites: (1) Rombaksbotn. (2) Gautetis, and (3) Ruvs­sot. Other locat ions mention ed in the text: C - Cainhajavrre .G • Gautelis. H • Stasjonsnolmen, M • Muoh taguob la. R ­Rombaksbotn. S - Sorcalen. Z - Sjangeli. The dashed linelabelled MTZ marks the western limit of the MuohtaguoblaTectonic Zone.

of coarse-gra ined granite and syenite that intru­ded older volcanosedimentary rocks at 1.78Ga(Rb-Sr method , Gunner 1981). Only relativelysmall north -south or iented remnants of thevolcanosedimentary sequences are preserved.The Lower Proterozoic rocks are locally overla­in by a thin autochthonou s sequence of LateProterozoic to Early Cambrian sediments cal­led the Dividalen Group (Birkeland 1976). TheLower Proterozo ic rocks together with theirDividalen Group cover were subsequently over­ridden by Caledonien nappes at about 400Ma.

At present, correlation of the Lower Protero­zoic supracru stal sequences within the Rom­bak Window and with the adjacent parts ofNorway and Sweden is restr icted by the lackof reliable age dates. The sedimentary andvolcanic rocks of the Ruvssot-Sjangeli area in

NGU- BULL.415, 1989 The geochemistry ofLower Proterozoic stuctctesttc turbidites 25

the east of the window, for which Romer(1987) has obtained and Rb-Sr date of 2.3Ga,have compositions that are markedly differentfrom those in the west (this paper; Kornelius­sen & Sawyer, this volume). The volcanic rocksin the western part of the Rombak Windowhave geochemical similarities with the 1.91-1.88Ga volcanic rocks of northern Sweden repor­ted by Frietsch & Perdahl (1987). Unfortunate­ly the contact relationship between the twogeochemically distinct supracrustal sequencesin the Rombak Window is obscured by amajor N-S tectonic zone between Muohtaguob­la and Ruvssot (MTZ, Fig. 1).

Volumetrically the two most important consti­tuents of the supracrustal sequences of theRombak window are siliciclastic turbidites andvolcanic rocks. Generally the turbidites and thevolcanic rocks occur in distinct domains sepa­rated by the later plutons, but at Rombaks­botn and Gautelisvatnet volcanic and tuffitelayers are present in the turbidites. In the areaswhere volcanic rocks predominate (e.g. Serda­len) volcaniclastics and conglomerate/debrisflows are interbedded with lavas.

Field and petrographic relationsof the turbidite sequences

Rombaksbotn areaLower Proterozoic metasedimentary rocks arewell exposed along the south side of Rombaks­botn (Fig. 1). In the west they are overthrustby Caledonian nappes, but to the east theyare interbedded with, and grade into, a vol­canic sequence of intermediate to acidic com­position. Beds are typically 2 cm to 1 m thickand are right Way up (Le. they young to thewest) and dip steeply to the west or north­west. The metasediments sampled are meta­greywackes and metapelites that commonlyexhibit graded bedding and, in places, preser­ve structures chartacteristic of Bouma sequen­ces found in turbidites. Small scours at thebase of the coarsest beds are present in someoutcrops. Several thin, bluish-green, calc-slll­cate beds are also present in the metagrey­wacke-pelite sequence, but are volumetricallyminor «0.5%).

The original mineralogy and textures of thesediments have been severely modified bymetamorphism, and to a lesser extent by pene-

trative deformation. In the metagreywackes themaximum grain-size is about 4mm, although0.5mm is typical; generally the largest grainsconsist of quartz or plagioclase of equant orslightly elongate form. A schistosity is presentin all the samples examined and is of thedomainal type. Grain-size in the pelites is com­paratively uniform, and the schistosity penetra­tive and planar. A few of the Rombaksbotnsamples contain the greenschist-facies mine­rals chlorite, quartz and muscovite, but becau­se these phases have not been observed tobe in mutual contact they are not regardedas diagnostic of metamorphic grade. Plagio­clase compositions in the metagreywackes andmetapelites are in the range An" to An2, .

Metamorphism took place at lower amphibo­lite facies conditions since the interbeddedcale-silicate layers contain the assemblage:

calcite + quartz + diopside + clinozoisite +hornblende.

Chemical compositions of the metagrey­wackes and metapelites are presented in Tab­le 1 and details of the analytical methods usedis given in Appendix I. An estimate of thebulk chemical composition of the turbiditesequences is given in Table 2, and is basedon the field observation that they contain about80% metagreywacke and 20% metapelite.Samples ES68 and ES69 represent the twoextremes of composition resulting from thedepositional process.

Gautelis areaWell-bedded, graded metagreywackes andmetapelites are exposed at the northern endof Gautelisvatnet. It is thought that the Gaute­lis metasediments (Fig. 1) rest unconformablyon a tonalitic basement and its overlyingcover sequence of dolomitic carbonates, butshear zones in the area preclude a clear in­terpretation of the contact relationships. Theclastic rocks are thin- to thick-bedded, gradedunits that contain remnants of Bouma sequen­ce structures, hence they are interpreted asturbidites. Dips are generally steep and to thewest or northwest, but younging directions areto the east; thus the beds are inverted, atleast locally. The foliation in the metasedi­ments dips steeply to the west-northwest.Several thin conglomeratic layers are presentand contain predominantly tonalitic clasts.

The principal minerals in both the metagrey­wackes and metapelites - quartz, plagioclase

26 Edward W. Sawyer &Are Kometiussen NGU· BULL. 415. 1989

Table 1. Analyses of greywactu (g) and pelt tu (p) frOl1ll theRombak WtndtN

GAUTELl5 ROMllAX580TN RIlf<BAX580TN RUV550T

E557g E558g E560p ESOlp E557g E568g E569p E570p E5729 E573g E571* E5131p E51329 E51339 E513'9

5tOI 72.25 70.57 57.17 SC.70 65.15 69.'0 53.17 60.11 69.08 68.31 5•• 88 51.85 58.0' 58.'1 58.31

Z:~S30.7' 0.85 0.92 1.00 0.75 0.70 0.8' 0.75 0.62 0.88 0.'5 1.21 0.85 0.90 0.88

\l.se II.59 16.80 11.74 1•• 08 13.23 18.08 18.07 13.59 12.78 12.21 13.39 13.97 14.55 14.61

~~S03 6.38 5.80 9.15 le.I7 6.95 6.18 10.'0 7.97 5.73 7.05 4.83 12.81 11.09 9.63 9.740.06 0.06 0.09 C.09 0.06 0.05 0.08 0.07 0.05 0.07 0.58 0.13 0.09 0.08 0.09

MgO 2.36 2.42 4.22 '.65 3.27 3.01 5.33 '.20 2.78 3.08 2.43 8.61 6.75 5.65 4.61C.O 2.71 2.78 2.75 1.86 I ... 1.07 2.'0 1.08 2.31 1.85 20.66 6.00 1.58 1.99 2.85

~~SO2.30 2.50 3.IO 3.60 3.00 3.50 3.'0 3.60 3.10 2.90 bd 2.91 2.55 3.51 2.711.82 1.86 3.41 3.90 I.70 2." '.17 3.73 2.27 I.59 bd 1.80 5.27 4.35 4.79

rS~50.13 0.08 0.12 C.12 0.13 0.12 0.13 0.11 0.11 0.15 0.17 O.II 0.15 0.13 0.130.71 0.79 1.25 :.27 1.0' 0.88 1.5' 1.38 0.70 0.72 4.57 0.78 1.18 0.70 0.91

TOTAL 101.30 100.15 98.88 10:.20 99.51 100.64 99.5' 99.0' 100." 100." 100.78 99.81 99.52 100.0' 99.69

V 139 118 229 27; I" 131 198 158 112 155 67 25. 141 II8 132c- 262 187 130 25, Ias 301 212 210 215 381 138 503 332 279 228Se 12 12 23 ZJ 15 14 24 21 14 13 21 37 22 21 23In III 111 188 IU 103 82 147 123 86 91 68 132 62 47 .8NI .5 5' 113 1~: 75 72 12. 82 67 63 48 211 139 109 98C. 8. 51 bd . 8 \l 109 18 59 35 32 8. 87 76 12Co I' IJ 18 23 20 19 35 19 18 21 10 .6 39 30 32l. 3d.0 25.0 30 22.3 52 33.1 28.7 16.' 28.0 61 27 10.8 23.9 20 27.3C. 73.9 52.0 54 ss.s 91 77.3 62.' 36.7 59.3 109 '6 25.0 50.8 50 51.8Nd 1'.8 12.4 1;.0 28.0 23.9 10.0 22.8 12.0 19.0 I •• O$m 5.30 •••2 '.18 4.91 4.82 2.64 4.3' 3.20 3.60 4.11E. 1.00 0.92 !.OO 0.92 0.93 0.69 0.90 1.00 C.86 1.00Tb 0.87 0.80 C.57 0.57 0.53 0.51 0.40 0.70 0.54 0.51Yb 2.21 1.62 2.51 1.91 2.07 2.32 1.85 2.31 1.49 1.5/Lu 0.31 0.29 C... 0.29 0.3' 0.39 0.3' 0... 0.37 0.30Y 20 18 26 2' 2. 25 18 22 15 30 33 28 19 16 18lr 252 167 141 13: 113 201 106 134 147 297 127 103 1I1 11' lISHr 5.75 3.98 J.32 5.10 2.IO 3.14 3.37 2.30 2.86 3.13Nb 8 9 12 I' II 11 17 11 8 11 10 9 7 8Ta 0.74 0.70 :.01 0.8' 0.99 0.99 0.85 0.41 0.75 0.858. 778 621 956 95: 704 873 598 876 697 775 bd 785 650 429 726C. 2.5 2.9 '.0 5.3 13.7 1I.1 7.0 5.6 8.1 8.5Rb 75 77 133 12; 110 88 205 151 105 10I bd 66 190 161 1915r 231 246 283 zs: 149 173 227 155 IC7 168 I05 157 118 153 112Th 9.8 7.6 bd s.s bd 10.7 10.7 10.4 8.1 17 bd 2.6 9.8 5 10.3U 3.2 2.7 bd '.0 bd 2.8 3.11 2.7 2.5 bd bd I.' 1.9 3 2.'Pb 25 18 32 22 II 23 32 II 19 14 bd 13 9 10 12CIA 52.6 53.0 54.3 S!.5 57.3 56.1 55.8 57.5 53.8 53.9 43.0 5I.3 50.8 49.6Eu/Eu* 0.58 0.62 :.74 0.65 0.68. 0.76 0.77 0.88 0.76 0.81

Qtz ...3 I: •• 35.6 10.9Plag 33.7 4:.3 37.1 Cl.98h 21.1 'Z.9 17.0 46.4

legend: ... -not determtned: bd-below :-!~ect1on level.

(Anl l to AnlO) and biotite - are not diagnosticof metamorphic grade. However, a metamor­phosed mafic dyke in the turbidites containsthe amphibolite-facies assemblage:

biotite + epidote + i1menite + plagioclase(AnJO-,,) + diopside + hornblende.

Field estimates indicate that, like the Rom­baksbotn turbidites, about 80% of the clasticmaterial is metagreywacke and 20% metape­lite. Thus, although samples ES57 and ES61represent the extremes of composition observ­ed (Table 1), the bulk sediment compositionlies close to the metagreywackes with about68% SiO, (Table 2).

Ruvssot areaSiliciclastic rocks are part of a supracrustalsuccession that is exposed on the south sideof Ruvssot. Generally, the siliciclastic rocksare pervasively foliated, but locally gradedbedding and structures of the Bouma sequen-

ce are preserved, from which deposition byturbidity currents is inferred. Minor, thin, blue­green, calc-silicate bands are present in theturbidites, but are usually disrupted into trainsof boudins.

Two types of mineral assemblage occur inthe Ruvssot turbidites. In the coarsest-grainedand lightest-coloured metagreywackes theprincipal phases are:

i1menite + biotite + quartz + plagioclase(AnlO_40)·

In contrast some darker metagreywackes.and most of the metapelites, contain the morecomplex amphibolite-facies assemblage:

epidote + ilmenite + hornblende + quartz +biotite + plagioclase (An,._40)'

The bulk composition of the arnphlbole­bearing rocks is AI,O)- and K,O-depleted, butCaO-enriched relative to the amphibole-freeassemblages.

Field observations indicate that the lighter­coloured, least mica-rich parts of the turbidites

NGU-BULL.415,1969 Thegeochemistry ofLower Proterozoic siliciclastic turbidites 27

Table 2. Est1mates of bulk sediment composH1ons forGautells (1). Rombaksbotn (2) and Ruvssot (3) turbldHes.

5102 68.32 65.23 56.19T10~ 0.75 0.76 0.97Al2 3 13.21 14.38 14.13

~~~03 6.81 7.28 10.820.07 0.06 0.10

M90 2.80 3.50 6.42CaD 2.76 1.82 3.11Na~o 2.60 3.18 2.95K2 2.20 2.83 4.06

r~~50.10 0.13 0.130.85 0.97 0.B9

TOTAL 100.47 100.14 99.77

V 154 148 164Cr 230 272 336se 15 16 26Zn 122 102 72Ni 65 80 139Cu 44 60Co 17 23 37La 28.9 39.3 20.5Ce 60.3 77.3 46.2Nd 22.7 18.3Srn 4.72 3.64Eu 0.97 0.95Tb 0.81 0.59Vb 2.05 1.84Lu 0.33 0.37V 20 22 20Zr 195 185 114Hf 4.56 3.83 2.76Nb 10 12 6Ta 0.78 0.88 0.69Ba 754 729 650Cs 3.0 7.7 7.4Rb 87 122 1525r 249 193 135Th 8.9 11.7 6.9U 3.2 2.8 2.2Pb 27 24 11CIA 53.1 55.3 48.7

No estimates of the Nd, Srn, Eu, Tb, Vb and Lu contents inthe Rombaksbotn bulk sed1ment are made because samples E567 andE573 (for wh1ch these REE were not determIned) have much h19herLa and Ce contents than the samples for which the REE weredeterm1ned.

predominate over the finer-grained and moremicaceous parts. Unfortunately the scatteredoutcrops and degree of deformation (thechange in relative proportion between mica­rich and mica-poor units during deformationis discussed by Sawyer & Robin 1986) preclu­des an accurate estimate of the relative abun­dance of metapelitic and metagreywacke mate­rial, although the coarse material predomina­tes. Samples ES133and ES131 (Table 1) repre­sent the extremes of grain-size, colour andmica content observed in the field. The bulkchemical composition of the sediment is esti­mated by a simple average (Table 2), but issimilar to the dominant rock type, ES 132.

Interpretation of the geochemistryof the Rombak turbiditesIn a simplistic model the provenance areabulk composition is the starting compositionfrom which the composition of clastic sedi-

ments diverge through the action of variousprocesses. Firstly, the provenance area bulkcomposition is modified by the process ofchemical weathering, which breaks the sourcerocks down to a detritus consisting of a mixtu­re of clay weathering products and unweather­ed grains (quartz, feldspar, etc.). Secondly,during sediment transport and deposition afurther modification takes place because thecoarse- and fine-grained material have diffe­rent hydraulic properties. Thirdly, there is thepreferential preservation of grains that aremechanically resistant to abrasion; quartz andzircon, for example.

The effect of chemical weatheringBecause the constituent elements of a rockundergoing chemical weathering are not equal­ly accommodated in the clay weathering pro­ducts, a chemical fractionation occurs duringweathering. Those elements not accommo­dated in the weathering products are lost tothe weathering solutions (Nesbitt et al. 1980).Thus, it follows that the more intense the chemi­cal weathering, the greater is the differencebetween provenance area bulk compositionand the composition of the detritus derivedfrom it. Since feldspars are the dominant labilemineral type in the earth's crust, Nesbitt &Young (1982) have argued that the main ef­fect of chemical weathering is the breakdownof feldspar to clay. Therefore, the ratio of theleast easily removed element in feldspar (Le.AI) to the more easily mobilized elements (Le.Ca. Na and K) provides a measure of chemi­cal weathering. Nesbitt & Young (1982) pro­posed a chemical index of alteration (CIA)based on the molecular proportions of themajor oxides:

CIA = AI20) / (AI20) + CaO· + Na20 + K20) /100 where CaO· represents CaO in the sili­cate phases only.

The CIA values for the Rombak clastic sedi­ments (Table 1) range between 43 and 58,and are comparable to the CIA values ob­tained from unweathered feldspar and igenousrocks (feldspar 50; basalts 40; granites 45-55).It is concluded that the detritus from whichthe Rombak clastic sediments were derivedhad undergone little chemical weathering. LowCIA values could be due to rapid erosion anddeposition preventing extensive chemical weat­hering, or they could also result from erosionin a cold (Nesbitt & Young 1982) and/or arid

28 Edward W.Sawyer &Are Korneliussen NGU-BULL.415.1989

3025

/

TO HICROCLlNE

/

20

. *- D~

~- 0::. 0- sorting::. "'1 ~o lines

~.... /

o 86;./

~.."/

/ /sorting line

BULK SEDIHENTS*ROHBAKSBOTN [ ~

* GAUTELlS t~'* RUVSSOT [~

/

TOHICROCLINE

®

QUARTZ, CALCITE

@

6

2

8

oN

~ 4

120Clu 8+0 ~

N cCl4

....z ~

120 QUARTZ

0 5

climate. Table 2 shows that the CIA value forthe Rombaksbotn bulk sediment (55.3) is slight­ly higher than that for the Gautelis detritus(53.1); this may indicate a greater degree ofchemical weathering for the Rombaksbotndetritus, as also indicated by a higher iIIitecontent (Fig. 2A).

The Ruvssot samples are enriched in K,O(Fig. 2A) and Rb relative to their AI,O, con­tents when compared to the other turbidites;this could be due to either an additional K-richdetrital phase, or to the introduction of K,Oand Rb. A K,O-rich detrital mineral would im­ply an acidic source area, which is not suppor­ted by the SiO" Hf, Zr, and MgO contents ofthe turbidites. The metasomatic introductionof K,O and Rb during regional metamorphismseems unlikely since nearby marbles and ba­sic rocks are not affected. An alternative, andperhaps more likely explanation for the K,Oand Rb enrichment is the introduction of aphillipsite cement, or the replacement of vol­canic glass fragments by phillipsite, duringearly diagenesis on the sea floor. Thus, thelower CIA values of the Ruvssot sampleswhen compared to either the Gautelis or theRombaksbotn turbidites (Tables 1 & 2) maybe due to diagenetic effects, or a more maficsource, rather than to a lower degree of chemi­cal weathering.

The effect of sortingSorting selects material by grain-size, but forany chemical fractionation to occur there mustalso be a sorting by phase. If detritus containsa complete range of grain-size for each phasepresent, then it is possible that no chemicaldifferences develop when grain-size sortingoccurs. However, if certain elements are en­riched in a particular size fraction (e.g., AI,O"Fe,OH MgO, K,O, Se, HREE in clays, or SiO"CaO and Na,O in coarse quartz and plagio­clase), then a chemical fractionation will occurduring grain-size sorting.

For the Rombak turbidites the effects ofgrain-size sorting are most clearly seen in theGautelis and Rombaksbotn samples. Compa­ring the coarse-grained base of graded unitswith the fine-grained top, it can be seen (com­pare ES57 with ES61 and ES68 with ES69,Table 1) that the metapelites are enriched inTiO" AI,O" Fe,O" MgO, K,O, P,O" Rb, Nb,V, Se, Co, Ni, Zn and the HREE (Fig. 3) relati­ve to the metagreywackes. With the exception

Fig. 2. Feldspar and clay components in the Early Proterozo­ie clastic turbidites of the Rombak Window. (A) AI,O, ver­sus K,O; the difference in K,O contents between the pelitesand greywackes from Rombaksbotn and Gautelis can berelated to iIIite separation along the sorting trend (dashedline). The Ruvssot samples may have contained detritalK-feldspar. (8) AI,O, versus (CaO+Na,O); for the Rombaks­botn and Gautelis samples the sorting trends (dashed li­nes) show that quartz, plagioclase and illite are affectedby sorting. Note that quartz tends to occur in the coarser­grained size fraction, but that plagioclase and illite occurin the fine-grained fraction. Symbols: circles - Rombaks­botn, squares - Gautelis, triangles - Ruvssot, open syrn­bols indicate fine-grained rocks (pelites) and closed sym­bols indicate coarse-grained samples (greywackes).

of P,O, and Nb, metapelites are enriched inthe elements found in, or adsorbed on, claysderived from chemical weathering. Metagrey­wackes are enriched in SiO" LREE (Fig. 3),Zr, Hf, and Cr relative to pelites. With theexception of the LREE, these elements probab­ly reflect the preferential preservation of me­chanically resistant quartz and zircon anddense chromite grains in the metagreywackeportion of graded beds.

NGU - BULL. 415. 1989 The geochemistry ofLower Proterozoic siliciclastic turbidites 29

Fig. 3. Chondrite-normalized REE contents of the Rombakclastic turbidites. Open circles indicate greywackes. closedcircles indicate pelites. Chondrite-normalizing factors fromTaylor & Gorton (1977), and Gd values interpolated.

volcanic ash, has been pointed out to us byone of the reviewers (Crook). Such an ex­planation is in accord with our inference ofexplosive volcanic activity close to the sedi­mentary basin.

The more pelitic parts of the Ruvssot tur­bidites are depleted in Si02, LREE, Zr, Ta, andTh, but enriched in Ti02, Fe203, MgO, V, Cr,Sc, Zn, Ni, Co and HREE (Fig. 3) relative tothe coarser grained portions. In contrast tothe Rombaksbotn and Gautelis turbidites bothNi and Cr are enriched in the Ruvssot meta­pelites, and it is inferred that, in this case,both elements are contained in the clays. Thehigh CaO, but low AI203, K20 and Rb contentsof sample ES131 may indicate the presenceof a fine-grained carbonate component andless clay in the original sediment (Fig. 2B).

GAUTELlS

10

10

YbLuNd Srn Eu TbLaCe

zWW0::

100

100

100

Figure 2A shows that the metapelites andmetagreywackes from Rombaksbotn and Gau­telis lie on a trend away from the illite apex.Thus, the difference in K20 contents betweenthe coarse- and fine-grained portions of gra­ded beds from Rombaksbotn and Gautelis isdue to the concentration of iIIite into the pe­litic part during sorting of the bulk sediment.However, the elements associated with plagio­clase show an interesting relationship. BothCaO and Sr have similar abundances in thecoarse-grained and fine-grained fractions ofthe Rombaksbotn and Gautelis turbidites. Inthe Gautelis clastics Na20 is actually enrichedin the metapelites, but has comparable levelsin both the Rombaksbotn metapelites and themetagreywackes. Furthermore, the total feld­spar content of the metapelitic portions of thegraded beds from Gautelis and Rombaksbotnis greater than in the metagreywacke portions(Table 1). This is contrary to the normal relati­onship (Blatt 1985) where pelitic rocks havelower feldspar contents than greywackes (e.g.the Archaean Quetico metaturbidites, Sawyer1986a fig. 2).

Because neither CaO, Na20 nor Sr are ac­commodated in the clays resulting from chemi­cal weathering they are removed in the weat­hering solutions (Nesbitt et al. 1980). Thus,rocks with a large proportion of weatheredmaterial (Le. clay-rich pelites) should be Na20,

CaO and Sr depleted relative to rocks contain­ing a large proportion of unweathered materi­al (Le. greywackes). The metagreywacke-bulksediment-metapelite sorting lines in Fig. 2Bshow that all three components are affectedby sorting; quartz is enriched in the metagrey­wackes, but illite and plagioclase are enrichedin the metapelites. Quartz must, therefore,occur mainly in the coarse grain-size fraction,but a large proportion of the plagioclase ispresent in the same size fraction as illite (Le.clay-sized). Therefore, it is concluded that theRombaksbotn and Gautelis bulk sediment con­tained a large proportion of clay-sized, un­weathered feldspar in addition to clay pro­duced by the Chemical weathering of feld­spar. Fine-grained unweathered feldspar impli­es a mechanical disaggregation process, twopossibilities are explosive volcanic eruptionsnear to the sedimentary basin, or that erosionin the provenance area took place in a cold cli­mate. An alternative explanation that the highmodal plagioclase contents in the metapelitesmay result from zeolitic alteration of silt-sized

30 Edward W. Sawyer &Are Korneliussen NGU· BULL. 415. 1989

Table J. Representatlve analyses of possible source rocks to tile Rombakturbtef tes

legend: - • not determined: bd • below detection Ltmtt ,The major oxides were determined by ):RF using fused glass ct scs , Y, Zr, Nb, ee,aa, Sr. Pb, Y, Cu, Zn and Ni were det:ermlned by XRF using pressed powderpellets. Se, Co, c-, cs , Hf, ra, Th, U, la, ce , Nd, Srn, Eu, re, Vb and lu wereoetermtned by Instrumental neutron ac:thatlon ane tys ts ,

1

CD

10

\ALKALI­°BASALT

TRACHYTE

o

.11

RHYOLlT[ vv

0·1 1Nb/Y

0·01 0·1Zr/Ti02

DAClTE

RHYODAClTE •• *0.) *

'" .00o

ANDES/TE

RHYOLlTE

SUB-ALKALINE 0 0

BASALT

RHYODAC/TE ei".,DAC/TE .Ift... v

•----- ..,.."'- .ANDES/TE .~

- - - ---".""ANDES/TE ,: •BASALT "

SUB- ·44ALKALINE rBASALT'/ ALKALI BASALT

80 T---------.---..,.,

400·001

Gautelis tetsic volcanites (G-type). Fine-grainedfelsic rocks that are thought to be of volcanicorigin occur within the Gautetis tonalitic com­plex. It is possible that the G-type volcanitesare intruded by the tonalite, but field relation­ships are not clear. These rocks have low K,Obut high Na,O contents, and 50 resemble theGautelis tonalitic complex, but contrast withthe generally potassic composition of theSN-type volcanic rocks (Table 3). The G-typevolcanites have LREE-enriched, calc-alkali­type REE patterns (Fig. 6).

1,..----...---..,.-.......---....,

0·01

Fig. 4. Winchester & Floyd (1977) discrimination diagramsfor volcanic rocks found in various parts of the RombakWindow. Symbols: • and * = Serdalen, 0 = Stasionshot­men. _ = Cainhavarre, t; = Rombaksbotn, 0 = Muohtaguob·la. 'V = Gautelis.

70

50

No(/)

60

No!::0·1L..

N

0·0010·01

G-type T·type

1::101.5 K140.5

75.70 67.290.22 0.54

13.01 14.911.27 4.470.02 0.050.29 1.541.36 3.606.17 4.501.03 2.120.03 0.130.53 0.83

99.62 99.98

13 56

"' s• B7 352 73 11

11 6'3 3179 6326 21'.0 3.'0.78 0.980.58 0.582.25 1.440.32 0.18

17 1618' 180

5.05 4.5717 12

1.62 1.13886 967

D.' I.'15 65

290 J7711 85 2

12 12

2'

M,"'"'"'

IS 213bd 2778 21

50 36

"' 80

"' 11

"' 36BD 1413 42556 1612.1 3.'0.97 0.97I.' 0.275.1 I.'0.B5 0.26

61 13t6l '3

15

667 Si

151 M'8 10724 381J bd18 bd

RS-type

~S19.3 Rl.3 Rl2.3

72.16 48.57 45.550.30 1.30 0.25

13.60 9.88 7.493.38 13.04 10.630.04 0.09 0.160.29 7.75 20.561.09 6.86 8.853.9 4.1 0.54.93 0.17 0.030.04 0.12 0.020.35 6.86 4.28

)00.28 98.74 98.32

16B2200

269B

1000B

890.621.00.'0.190.131.00.198

25

The carbonate was subsequently consumedduring the metamorphic reactions that produ­ced amphibole.

Tonalitic Complex (T-type). Clasts derived fromtonalitic rocks in the conglomerates within theGautelis turbidites indicate that sialic crust wasexposed in the area during the Early Protero­zoic. The Gautelis tonalitic complex consistsof a suite of low-K,O, but high Na,O plutonicrocks containing from 60 to 72% SiO, (Table3), that underlies the turbidite sequence. Thus,the tonalite could have contributed detritus tothe turbidites.

Rock types in the provenance areaTwo possible sources can be considered forthe detritus from which clastic turbidite form­ed: 1) Erosion of continental crust. 2) Erosionof coeval and predominantly volcanic rocks inthe general vicinity. Within these two con­straints, four possible source rock types canbe recognised in the Rombak Window. Chemi­cal analyses of a representative group of pos­sible source rocks for the turbidites are presen­ted in Table 3.

SN-type vetcenf tes

E54) E$48 KID4.'

51°2 58.35 65.33 68.41

~i~531.03 0.6' 0.56

17.16 14.60 13.'3

~~~O3 7.22 6.63 4.940.10 0.14 0.07

.,0 3.29 1.95 0.76CaD 6.00 4.19 1.46

~;503.7 3.8 '.1

2.596 2.56 4.89

~5~50.25 0.14 0.100.91 0.64 0.75

TOTAL 100.60 100.67 100.07

V 117 83 30Cr 136 60 13Se 17 15 8Z, ID' Z'I 62N1 22 11 sC" 6 98 6'Co 22 I' 8la 35.5 44.6 61C. 76.9 94.4 132N' 29.7 38.9 53Srn 6.61 7.66 8.8

'" l.JO 0.93 1.10Tb 0.77 1.10 1.20Vb 2.82 3.47 4.19tu 0.390 0.568 0.71V 28 '0 51Zr 189 207 337Hf 4.55 5.27Nb 10 14 17Ta 0.91 1.09Ba 650 435 882C. 6.3 12.0

." US tss na5r 562 202 116Th '.8 15.0 19U '.6 6.8 bdPO 3' " 23

NGU - BULL. 415, 1989 The geochemistry ofLower Proterozoic siliciclastic turbidites 31

100

o HAFIC• ULTRAHAFlC

RS-TYPEEXTRUSIVES 10

SN-TYPE VOLCAN/CSoAClO• INTERHEOIATE

La Ce Nd SmEu Tb""'--r"---'T""-"T"""1r--""---"""T'"'1~1

YbLu

10

Fig. 6. Chcndrite-norrnalized REE patterns for a representa­tive group (Table 3) of possible source rocks to the Rom­bak turbidites. Note the LREE-enriched calc-alkali-like pat­terns for the SN- and G-type volcanic rocks, and the relati­vely smooth patterns from the tonalitic complex, Normali­zing factors from Taylor & Gorton (1977), Gd values interpo­lated.

represents a significant proportion of the sup­racrustal sequence that includes the Rombaks­botn turbidites.

Low-potassium, mafic and ultramafic extru­sives (RS-type)In the eastern part of the Rombak Window,notably in the Ruvssot-Sjangeli area (Fig. 1),

100

100 10

z GAUTEL/S AREAW o TONAL/TIC COHPLEXW • (i-TYPE VOLCANICSCl::

100 10

8070

A

o~

; .. . \••••. .a.".

•• 't...

50 60Si0 2

40

Fig. 5. SiO, versus (Nap + K,Oj diagram for the Rombakvolcanic rocks showing the tendency of mafic and interme­diate types to plot in the alkaline field, but felsic rocks toplot in the subalkaline field. Symbols as for Fig. 4.

Felsic to mafic volcanic suite (SN-type). In thewestern part of the Rombak Window (serca­ten, Stasjonsholmen, Cainhavarri, Muohtagu­obla and Rombaksbotn, Fig. 1) volcanic rocksthat range from 49% to 76% Si0 2 (Table 3)are an important constituent of the supracrus­tal sequences. On the Winchester & Floyd(1977) variation diagrams (Fig. 4) these vol­canites range in composition from subalkalineand alkali basalt through andesite and trachy­andesite to dacite, rhyodacite and rhyolite. thematte and intermediate members of the suiteplot in the alkaline field, whereas the felsicmembersare distinctlysubalkalinewhenconsid­ered on a Si01versus (Na10 + K10) diagram(Fig. 5). However, the andesites and dacitesare K20-enriched and have LREE-enrichedREEpatterns (Fig. 6) similar to those shown by thecalc-alkaline magma suite. Thus, the SN-typevolcanic rocks are interpreted as K-rich, calc­alkali andesites and dacites similar to thosedescribed by Gill (1984).

Significant Si01 mobility either into, or outof, the SN-type volcanites seems unlikely be­cause most of the samples plot in the samefields in Figs. 4A & B. Notable exceptions arethe intermediate volcanites from the Muohtagu­obla area, which occur in a region of extensi­ve greenschist-facies retrograde metamor­phism (Sawyer 1986b). The Muohtaguoblasamples appear to have had Si0 2 added andare, therefore, excluded from further consider­ation in this paper. The SN-type volcanic suite

14-r----------.....,12

'0 10N

~ 8~ 6

N

~ 42o..L....:.'T----:..-,-----.---r---r-......I

32 Edward W.Sawyer &Are Korneliussen NGU·BUll.415,1989

o

40>-

60

20

300

80

200.0a:::

100

o2·01·50·5o1-50·5

•• * ROI1BAKSBOTN •.~ •0

6.... * CiAUTELlS ".. • •

'0 • "* RUVSSOT • •••~.* •• • •.* • • •• • • ~

•••• ••0 • • . ,.- ..,. • •

# .- * I •• •• ••.. ... *.• • 6.: . • •

•• 0 6 00 0.- ••• 06

lllP e® @lIIlO 0 0

", SOURCE ROCKS•• • SN-TYPE0 RS-TYPE

•• • 6 o-TYPE••.. • 0 T-TYPE• • •• • r...... ,• • •.. ,. • • • '\ •• ••

·0 ~... 6 •• , •.a: •• ~ ••i' •• .1- • -o ;k • • •~. •• If:' :..

O~ 0 •• o •~* .. o • 0

0 00

r10Cl © rm @oo

50

70

40

80

120

NoU) 60

80o

.....J

Fig. 7. TiO, versus SiO" Rb, La and Y variation diagrams comparing the Rombaksbotn, Gautelis and Ruvssot bulk sedi­ment compositions (Table 2) with the four possible source rock types,

mafic to ultramafic rocks are an importantconstituent of the supracrustal sequence. Thebasaltic rocks show pillow structures in a fewplaces. The basalts have low K20 contentsand REE patterns that are LREE-enriched;thus, they could be of calc-alkaline affinity.The ultramafic rocks, on the other hand, haveREE patterns that are different (Fig. 6) andresemble those of ocean ridge material. Onthe basis of their noble element patterns (Bar­nes et al. 1988), major and trace element con­tents (Korneliussen & Sawyer this volume), theultramafic rocks probably represent deformedand metamorphosed komatiites, or rocks ofboninitic affinity (Korneliussen & Sawyer thisvolume). Romer (1988) also considers the ultra­mafic rocks to be of boninitic affinity on thebasis of their I> Ndvalues.

Bulk sediment composition and sourcerock contributionsThis section examines which, if any, of thepossible source rocks could have contributedmaterial to the detritus from which each of theRombak turbidite sequences were derived. Theapproach adopted here is to compare thebulk sediment composition with the compositi­ons of the potential source rocks. Use of thebulk sediment compositions eliminates theeffects of sorting and maturity. The compositi­ons of the source rocks in Figs. 7 and 8 arefrom Korneliussen & Sawyer (this volume),who discuss the geochemistry and petrogene­sis of the Rombak igneous rocks.

In general, the contribution of a particularsource rock to sedimentary detritus can onlybe recognised if that source rock has a distinc-

NGU • BULL. 415, 1989 The geochemistry of Lower Ptoterozoic siliciclastic turbidites 33

Fig. 8. Rb versus Ni diagram for discriminating possiblesource rock types to the Rombak clastic turbidites, sho­wing that the bulk sediment compositions are Ni-enrichedrelative to the SN· and G-type volcanites and the T-typesource. Note breaks on the vertical axis. Symbols as forFig. 7.

would give rise to much higher MgO (>4%),CaO (>4.5%) and Sr (>400ppm) contents in thesediments than is observed. Although it canbe argued that both CaO and Sr would be lostto the weathering solutions, this is not thecase for MgO (Nesbitt et al. 1980, Sawyer1986a). The Rombaksbotn and Gautelis bulksediment compositions could be more sucess­fully modelled if the bulk sediment containeda small component (2 to 5%) of a very Ni­and Cr-enriched (1400ppm and 2600ppm, re­spectively) source, such as the RS-type ultra­mafics. Unfortunately, no distinction can bemade between a small component of an RS­type ultramafic source for the Ruvssot bulksediment, and a large component of the Ni­and Cr-enriched SN-type volcanic source, onthe basis of either Rb, K20, Sr, CaO, MgOor Ti02 contents.

The REE and other high ionic potential ele­ments (e.g. Th, Zr, Y, Ta and Nb) have beenuseful for determining metasediment sources(Jenner et al. 1981, Taylor & McLennan 1981a,Sawyer 1986a), or tectonic settings (Bhatia &Taylor 1981, Bhatia 1981, Bhatia & Crook1986) of clastic sediments because they arethe least mobile elements during weathering,

100 0

11I

300:~ .

200Rb

. .,: I:. .. .. ._. ....,

100

. ** ..-

o

200

1600Io

140010

105010

250

tive composition. Although Fig. 7A (Ti0 2 versusSi02) clearly shows that acidic source rockswere the largest contributors to both the Rom­baksbotn and the Gautelis bulk sediments, itcannot distinguish between the three possibleacidic sources. However, the high Ti02 but lowSi02 contents of the Ruvssot turbidites indi­cates that neither the G-type, the T-type northe felsic SN-type source rocks are a signifi­cant component in these sediments.

Mineralogically, the principal difference bet­ween the acidic source rock types is the type Z

of feldspar they contain: Fig. 7B (Ti0 2 versusRb) differentiates between the K-feldspar-richSN-type and the plagioclase-rich G-type vol­canites and the tonalitic complex. The Gaute-lis bulk sediments are Rb- (Fig. 7) and K20­

depleted relative to the trend of the SN-typevolcanites and are displaced towards the fieldof low-Rb acidic source rocks. This impliesthat either the G-type volcanites or the tonali-tic complex provided a significant acidic com­ponent to the Gautelis turbidites. The Rom­baksbotn bulk sediment is also Rb- and K20­

depleted relative to the SN-volcanic trend;however, the acidic component of the tonaliticcomplex and/or the G-type volcanites is muchsmaller, and may even be absent. In con­trast, the Ruvssot bulk sediment plots withinthe SN-type volcanic trend, but because theseturbidites may have experienced K20 and Rbaddition during diagenesis, the Ti02 versusRb plot cannot be used to distinguish betweensource components.

The Ruvssot-Sjangeli mafic and ultramaficrocks also have tow K20 and Rb contents, andplot close to, but separately from, the G-typevolcanites and the tonalitic complex in Fig. 7B,nevertheless, a greater dispersion between thesource rock types with low Rb contents isdesirable. The high Ni and Cr content of theRS-type source rocks (particularly the ultrama­fic rocks) is distinctive, and in Fig. 8 the RS­type rocks form a separate field. All threebulk sediment compositions are enriched inNi (Fig. 8) and Cr relative to the tonalitic comp­lex, the G-type and the SN-type volcanites.Hence, the presence of a component derivedfrom a Ni- and Cr-enriched source is inferredfor all three bulk sediments. However, the Ni­and Cr-enriched members of the SN-type vol­canites cannot be the source for the Ni andCr present in the Gautelis and Rombaksbotnbulk sediments. This is because the largevolumes required of such a source rock type

34 Edward W. Sawyer & Are Korneliussen

diagenesis and metamorphism (Nance & Tay­lor 1976, McLennan et al. 1980).

Figure 7C (TiO, versus La) shows conside­rable range of La contents within the possiblesource rock types, especially for the SN-type.Inspection of Fig. 7C shows that most of themafic and intermediate members of the SN­type sample set have La contents between25 and 50ppm, although some samples havein excess of 100ppm. Because the frequencydistribution of La contents in the intermediateto mafic members of the SN-type volcanitesis not Gaussian but log-normal, a representati­ve value is more accurately given by the esti­mated geometric mean (43ppm) of the dataset, and not by the arithmetic mean (51ppm).Similarly for the acidic members (SiO, >66%)the estimated geometric mean is 67ppm La.The REE patterns for the SN-type volcanites(Fig. 6) are, therefore, of samples closest tothese estimated geometric means. The variati­on in Y contents (Fig. 70) suggests that theHREE are also log-normally distributed.

The comparatively low total REE contentsof the Gautelis bulk sediment indicates that alarge component of the bulk sediment wasderived from a source rock having low REEcontents, such as the G-type volcanites or thetonalitic complex (incidently, confirming theTiO, versus Rb data). Alternatively, a largecomponent of REE-deficient material, such asquartz, is diluting the REE levels in all thesediments. It can be argued that the shapeof the Gautelis bulk sediment REE pattern(smooth from La to Lu) implies that the tonali­tic complex (smooth REE pattern) and not theG-type volcanites (LREE-enriched REE pat­tern) was the principal source of acidic materi­al. The Gautelis bulk sediment compositioncan be derived from a mixture of tonaliticcomplex, SN-type volcanites (intermediate tomafic composition and without extremely frac­tionated types) and a small component ofRS-type ultramafic rock.

The Rombaksbotn sediments all have REEpatterns that are flat from Tb to Lu, and hen­ce resemble those of the SN-type volcanicrocks (compare Figs. 3 and 6). The REE pat­tern is, therefore, consistent with a large pro­portion of SN-type material in these sedi­ments, and confirms the conclusions basedon the distribution of other trace elements(e.g. Fig. 7). The Ni and Cr data indicate thepresence of a RS-type ultramafic source com­ponent. There is no conclusive evidence for

NGU-BULL.415.1989

either tonalitic complex or G-type volcanitederived material in the Rombaksbotn bulksediment.

The Ruvssot sediments have lower and flat­ter REE patterns than either the Rombaksbotnor the Gautelis bulk sediments, and so musthave had a different provenance area. The lowLa and total REE contents of the Ruvssotturbidites excludes the possibility that the Cr­and Ni-enriched SN-type volcanites form asignificant component in the sediments be­cause such volcanic rocks contain >43ppm La.However, on the basis of their REE contentsthe RS-type extrusives are a suitable sourcefor the Cr, Ni and MgO in the Ruvssot tur­bidites. The REE patterns of the Ruvssot tur­bidites are sufficiently low as to exclude theSN-type volcanites as a source componentaltogether. This is because if the SN-typevolcanites were a source component, thenconsiderable dilution by phases lacking REE(e.g. quartz and carbonate) would be requiredto produce the observed low REE contents,and the major elements contents of these tur­bidites does not support this. It is concludedthat a major component of the sediment in theRuvssot turbidites came from a source thatis not presently recognised. However, it ispossible to infer that this principal source forthe Ruvssot turbidites was volcanic and hadan intermediate, LREE-enriched bulk composi­tion with 20-30ppm La: we suggest a calc­alkaline arc volcanic source that was moreprimitive (less evolved) than the SN-type.

Tectonic setting and basindevelopmentThe field and compositional characteristics ofthe turbidites in the Rombak Window provideconstraints on the palaeogeography and typeof tectonic environment in which they formed.However, the lack of precise age determinati­ons from the volcanic sequences in the win­dow precludes any meaningful interpretationof the sequence of basin development. Theage of basin formation is poorly constrainedby the 2.3 ± 0.08Ga Rb-Sr age for the RS­type extrusives (Romer 1987) and the 1.78 ±0.085 Ga Rb-Sr age for the granites and syeni­tes that intrude the supracrustal sequences(Gunner 1981). Outside the Rombak Windowtwo periods of Lower Proterozoic volcanismare known; the first, in northern Norway betwe­en 2.5 and 2.1 Ga, and the second, in north-

NGU· BULL. 415, 1989 Thegeochemistry ofLower Proterozoic siticictestic turbidites 35

central Sweden, between 1.91 and 1.88 Ga(Skiold & Cliff 1984, Pharaoh & Pearce 1984,Claesson 1985; Ski61d 1986, 1987). By analogywith volcanic rocks of similar compositiondescribed from Sweden (Frietsch & Perdahl1987, Widenfalk et al. 1987) the SN-type vol­canites of the Rombak Window may belongto the younger 1.91-1.88 Ga event.

The Rombaksbotn and Gautelis turbiditesare comparatively thinly bedded with a largesand/clay ratio; thus they resemble turbiditesfrom active margin basins rather than Atlantic­type passive margins. However, other factors,e.g. denudation rates, can affect sand/clayratios (Stow et al. 1985, Normark et al. 1985).The Rombaksbotn and Gautelis turbidites areinterbedded with andesitic and dacitic volcanicrocks and have compositional characteristicswhich suggest that part of their fine-grainedsize fraction was either volcanic glass, orunweathered feldspar. Hence, an origin nearto paroxymal volcanism is inferred, and conse­quently these metasediments are unlikely torepresent trailing edge-type passive marginbasins.

Comparison of the turbidite compositionwith possible source rocks found in the Rom­bak Window indicate that both the Rombaks­botn and the Gautelis sediments contain alarge proportion of SN-type volcanic materialand a small «3%) Ni- and Cr-enriched maficto ultramafic component. If the high Cr con­tents of the greywackes are indeed due todetrital chromite, then an ultramafic source ismore likely. Korneliussen & Sawyer (this vo­lume) argue that, on the basis of negative Ta,Nb and Ti anomalies shown by the SN-typevolcanites when compared to primordial man­tle, these volcanites are subduction-relatedmagmatic arc rocks. Furthermore, by analogywith modern arc rocks described by Gill (1984),the calc-alkaline REE patterns and high K20and Rb contents of the SN-type volcanitessuggest that they represent a late, or mature,stage of arc development.

The Gautelis turbidites contain, in addition,geochemical and field evidence (presence ofclasts) of a component derived from tonalite.The tonalite, together with its associated, butvolumetrically minor, G-type volcanites, is ex­posed near Gautelisvatnet and could eitherbe part of the Archaean basement present inthe Kiruna·Lofoten area, perhaps equivalentto the Soppero gneisses (Pharaoh & Pearce1984, Ohlander et al. 1987), or alternatively,

it could be equivalent to the 1.89 Ga Kristine­berg and Jom subvolcanic complexes of Swe­den (Ohlander et al. 1987).

The Rombaksbotn and Gautelis turbiditesformed in active margin basins adjacent to amature volcanic arc; in the case of Gautelis,the arc may rest on a tonalitic continental crustin an Andean-type continental margin setting.The detritus from which the turbidites formedcame directly from volcanic eruptions and alsofrom erosion of the arc complex.

The Ruvssot turbidites were derived from asource that was quartz-poor, LREE-enrichedand generally of mafic to intermediate composi­tion. This more mafic source is also the likelyreason for the lower CIA values. The composi­tion of the Ruvssot turbidites indicates thatthey formed in a basin close to ultrarnaflc/mafic (RS-type) to intermediate rocks of calc­alkaline type, but distant from any source ofacidic, or very fractionated, LREE-enrichedmaterial. This suggests either an intra-oceanicenvironment adjacent to primitive island arcvolcanism, or perhaps a trench slope environ­ment.

Bhatia & Crook (1986) have used the compo­sitional characteristics of Palaeozoic grey­wackes from Australian turbidites to developa series of ratio and triangular plots for thediscrimination of the tectonic setting of sedi­mentary basins. Their diagrams for greywack­es (grain-size 0.06-2.0 mm) can be testedwith the Lower Proterozoic Rombak turbiditessince their tectonic settings have been tentati­vely established above.

Although the individual greywacke samplesshow some scatter on the Bhatia & Crook(1986) plots (Fig. 9A to D), they, and the bulksediment compositions, plot within the fieldcompatible with the field observations andgeochemical interpretation given above. Theonly exception is seen in the Sc/Cr versusLalY diagram (Fig. 9E) where Rombak sedi­ments plot out of the compositional fields alto­gether, except for some samples that lie inthe passive margin field. This suggests thatsediments derived from er- and La-enrichedsources cannot be represented on such dia­grams. Such a restriction excludes sedimentsderived from mature arcs: and probably alsoArchaean and Lower Proterozoic volcanicrocks, since evidence indicates a change inaverage composition through geological time(Condie 1981, Taylor & McLennan 1981 b, Gill1984), in particular higher MgO, FeO, Cr and

36 Edward W. Sawyer & Are Korneliussen NGU - BULL. 415.1989

an Andean-type setting is suggested. TheRuvssot turbidites are more primitive, and donot contain continental or fractionated acidicvolcanic material. Thus, they may have formedin an intra-oceanic setting by the erosion ofa primitive calc-alkaline volcanic arc far rerno­ed from a source of continental material.

©

.'.'

o~-,:-\

'D1IC I" " ,; 0 L>" '\

'i'l./ B "0,' I

, '0 ... _,.- __ =./_

", A

O-+-~~~~~

o Se/er 1·0

2 :' AQH

i>­<,e

...J

2r/l0

2r/l0

to Th

to Th©

Co

Se

Fig. 9. Rombak turbidites plotted on the trace element disc­riminant diagrams for greywackes (grain-size O.06-2.0mm)of Bhatia & Crook (1986). The fields outlined dashed are: A- ucean islands. B - continental island arcs, C - active conti­nental margins, D - passive margins. On diagram (E), Sc/Crversus La/Y, the Early Proterozoic Rombak and the Archae­an Quetico (AQM, Sawyer 1986) turbidites are Cr- andLa-enriched relative to the Palaeozoic turbidites upon whichBnatia & Crook (1986) base their discrimination diagrams,and hence plot out of the cornpositional fields of the vari­ous sedimentary basins. See text for discussion of thepetite samples plotted on these greywacke diagrams. Sym­bols as for Fig. 2.

®

AcknowledgementsThe first author would like to thank the Norges geologiskeundersekelse for a research contract (1984185) and NTNFfor a post-doctoral research fellowship (1985/86) held atNGU in Trondheim. We would also Ijke to express ourthanks to our colleagues Sarah-Jane Barnes and PierreCousineau (UQAC) for their constructive comments andcriticism at various stages of this work. Finally, we wouldlike to thank Drs. K. Crook, who pointed out the possibilityfor diagenetic zeolite in these sediments, M. Bhatia andA. Siedlecka for their constructive reviews of the manuscript.

Ni contents compared with Phanerozoic arcrocks. Contamination of the Rombak sampleswith Cr and La during sample preparation canbe excluded as the cause of the discrepancyin Fig. 9E (see Appendix).

The diagrams of Bhatia & Crook (1986) areintended for greywackes, and where the origi­nal silt- and clay-sized sediments can be recog­nised, as for example the tops of graded beds,the distinction between greywackes and pelitesmay be reltively simple. However, in the caseof sediments metamorphosed to the amphibo­lite facies the original grain-size may be consi­derably coarsened by recrystallisation (e.g.Chipera & Perkins 1988, p. 41), and consequ­ently the distinction between pelite and grey­wacke may be more difficult. The inadvertentplotting of pelites on these diagrams resultsin an increased scatter of samples on eachof the plots, as can be seen when the Rom­bak pelites are plotted in Fig. 9. For example,in Fig. 9D the Rombaksbotn and the Gautelisgreywacke bottoms of graded units plot in thefields of continental island arcs and activecontinental margins, but the pelitic tops plotin the oceanic island field. Therefore, in recrys­tallized metasediments care must be taken toexclude rocks that were originally mud-rich;these can then be treated separately (e.g.Bhatia 1985).

The separation of pelite from greywacke dueto the sorting of fine-grained material fromcoarse-grained is apparent on all the diagramsof Fig 9, and illustrates a potential weaknessof a grain-size based chemical classificationscheme for sediments. This is the possibilitythat, even in the sand-sized fraction, the clay­rich distal facies of a turbidite fan sequencecould be classified differently from the clay­poor proximal part, thus spuriouis correlationscould result.

ConclusionsThe tectonic setting in which the clastic tur­bidites of the Rombak Window formed canbe inferred from their compositions by determi­ning possible source rock types. However, thecompositional variation due to sorting mustbe eliminated by using bulk sediment composi­tions. The turbidites from Rombaksbotn andGautelis formed in an active marginal basinsetting adjacent to a mature volcanic arc. Thevolcanic arc that supplied the Gautelis tur­bidite basin formed on a tonalitic crust; and

NGU-BULL.415,1989 Thegeochemistry of Lower Proterozoic siliciclastic turbidites 37

AppendixAnalytical methodOne to two kg of rock were collected and reduced to2-5cm chips on the outcrop. Samples were then furtherreduced to 6 mm chips in a jaw crusher and pulverised to400 mesh in a an agate mill at the Norges geologiskeundersekelse (NGU)in Trondheim. Major oxides were deter­mined on fused glass discs and the trace elements Nb, Zr,Y, La, Ce, Th, U, s-. Rb, Ba, Zn, Co, Cr, Cu, Ni, V, Sc andPb on pressed powder pellets using standard XRF me­thods at NGU. La, Ce, Nd, Sm, Eu, Tb, Yb, Lu, Ta, Cs andHf were determined by instrumental neutron activation analy­sis (INAA) on selected samples using Becquerel Laboratori­es of Toronto; for these samples the values for er, eo,Sc, Th, U given in Table 1 are also by INAA. A sampleof Archaean greywacke from the Quetico metasedimentarybelt was also chrushed and pulverised at NGU and analy­sed by XRF at NGU and by INAA at Becquerel, the resultsagree to within 10% with those obtained earlier by the firstauthor at the University of Toronto. Previously pulverisedand analysed powders (UTB-l, UTA-l and UTR-l) wereused as internal standards in all the analytical runs andalso agreed to within 10% of previous estimates made atthe University of Toronto. Thus, systematic contamination,or errors in the analysis, of Sc,Cr, La and Y are not likely.

ReferencesArgast, s., & Donelly, T.W. 1987: The chemical discriminati­

on of clastic sedimentary components. J. Sed. Petrol.57, 813-823.

Barnes, S-J., Boyd, R., Korneliussen, A., Nilsson, L.-P.,Often, M., Pederson, R.B., & Robins, B. 1988: The useof mantle normalization and metal ratios in discrimina­ting between the effects of partial melting, crystal frac­tionation and sulphide segregation on the platiniumgroup elements, gold, nickel and copper: Examples fromNorway. In Prichard, H.M., Potts P.J., Bowles, J.FW.,& Cribb, S.J. (eds.) Proceedings of the symposiumGeo-platinium 87, Open University UK, April 1987. Else­vier Applied Science, Barking, 113-143.

Bhatia, M.R. 1983: Plate tectonics and geochemical compo­sition of sandstones. J. Geol. 91,611-627.

Bhatia, M.R. 1985: Composition and classification of Pateo­zoic f1ysch mudrocks of eastern Australia: Implicationsin provenance and tectonic setting interpretation. Sed.Geol. 41, 249-268.

Bhatia, M.R., & Crook, K.AW. 1986: Trace element charac­teristics of greywackes and tectonic setting discriminati­on of sedimentary basins. Contrib. Mineral. Petrol. 92,181-193.

Bhatia, M.R., & Taylor, S.R. 1981:Trace-element geochemis­try and sedimentary provinces: a study from the Tas­manGeosyncline,Australia. Chem. Geol.33, 115-125.

Birkeland, T. 1976: Skjomen berggrunnsgeologisk kart N­O-M 1:100000. Nor. geol. unders.

Blatt, H. 1985: Provenance studies and rnudrocks. J. Sed.Petrol. 55, 69-75.

Chipera, S.J., & Perkins, D., 1988: Evaluation of biotite­garnet geothermometers: application 10 Ihe EnglishRiver subprovince, Ontario. Contrib. Mineral. Petrol. 98,40-48.

Claesson, L.A. 1985: The geochemistry of early Proterozo­ic metavolcanic rocks hosting massive sulphide depo­sits in the Skellefte district, northern Sweden. J. geol.Soc. London 142, 899-909.

Condie, K.C. 1981: Archean Greenstone Belts. Elsevier,Amsterdam, 434 pp.

Deer, W.A., Howie, R.A., & Zussman, J. 1967: An Introducti­on to the Rock-Forming Minerals. Longman, Green &Co. Ltd. London, 528 pp.

Frietsch, R., & Perdahl, J.-A. 1987: On the nature of thelower Proterozoic volcanic rocks in southern Norrbot­ten, northern Sweden. Abstr. IGCP 217 tuna, p. 35.

Gill, J.B. 1984: Orogenic Andesites and Plate Tectonics.Springer-Verlag, Berlin 392 pp.

Gunner,J.D. 1981:A reconnaissance Rb-Sr study of precam­brian rocks from the Sjangeli-Rombak window and thepattern of initial "Sr/ 16Sr ratios from northern Scandina­via. Nor. Geol. Tidsskr. 61,281-290.

Jenner, G.A., Fryer, B.J, & Mclennan,S.M. 1981:Geochemis­try of the Archaean Yellowknife Supergroup. Geochim.Cosmochim. Acta 45,1111-1129.

Korneliussen, A., Tollefsrud, J.I., Flood, B., & Sawyer, E.W.1986: Precambrian volcano-sedimentary sequences andrelated ore deposits, with special reference to the Gaute­Iisfjell carbonate-hosted gold deposit, Rombaken base­ment window. Nor. geol. unders. Report 86.193,46 pp.

Korneliussen, A., & Sawyer, E. W. 1989: The geochemistryof Lower Proterozoic basic to acidic extrusive and intru­sive rocks in the Rombak basement window. Nor. ge­01. unders. Bull, 415, 7-21 (this volume).

McLennan, S.M., Nance, W.B" & Taylor, S.R. 1980: Rareearth element-thorium correlations in sedimentary rocks,and the composition of the continental crust. Geochim.Cosmochim. Acta 44, 1833-1834.

Nance, W.B., & Taylor, S.R. 1976: Rare earth patterns andcrustal evolution I: Australian post-Archaean sedimenta­ry rocks. Geochim. Cosmochim. Acta 450,1539-1551.

Nesbitt, W.H. 1979: Mobility and fractionation of rare earthelements during weathering of a granodiorite. Nature279, 206-210.

Nesbitt, HW., Markovics, G., & Price, R.C. 1980: Chemicalprocesses affecting alkalis and alkaline earths duringcontinental weathering. Geochim. Cosmochim. Acta 44,1659-1666.

Nesbitt, HW., & Young, G.M. 1982: Early Proterozoic clima­tes and plate motions inferred from major element che­mistry of lutltes. Nature 299, 715-717.

Normark, W.R., Barnes, N,E., & Bouma, A.H. 1985: Com­ments and new directions for deep-sea fan research.In Bouma, A.H., Normark, W.R., & Barnes, N.E. (eds.)Submarine Fans and Related Turbidite Systems. Sprin­ger Verlag, New York, 341-343.

Ohtander, B., Ski6ld, T., Hamilton, P.J., & Claesson, L.-A.1987: The western border of the Archaean provinceof the Baltic shield: evidence from northern Sweden.Contrib. Mineral. Petrol. 956, 437-450.

Pharaoh, T.C., & Pearce, J.A. 1984: Geochemical evidencefor the geotectonic setting of early Proterozoic metavol­canic sequences in Lapland, Precamb. Res. 25,282-308.

Reimer, T.O. 1985: The suitability of shales as an indicatorfor the composition of the Archaean continental crust.Neues Jahrb. Mineral. Abh. 152, 211-223.

Reimer, T.O. 1986: Alumina-rich rocks from the early Pre­cambrian of the Kaapvaal craton as indicators of paleo­sols and as products of other decompositional reacti­ons. Precamb. Res. 32, 155-179.

38 Edward W. Sawyer & Are Korneliussen

Romer, R.L. 1987: The geology, geochemistry and meta­morphism of the Sjangeli area, a tectonic basementwindow in the Caledonides of northern Sweden. Rese­arch Report. 1987:16. LuleA University of Technology.124pp.

Romer. R.L. 1988: Disturbances of Pb. Sr, and Nd isotopesystems related to medium and low grade metamorp­hism illustrated with data from the polymetamorphicEarly Proterozoic (Rombak)-Sjangeli basement windowof the Caledonides of Northern Sweden. Doctoral the­sis, Lulea UniversityofTechnology, 1988:650. 124pp.

Sawyer. E.W. 1986a: The influence of source rock type,chemical weathering and sorting on the geochemistryof clastic sediments from the Quetico MetasedimentaryBelt, Superior Province, Canada. Chem. Geol. 55.77-95.

Sawyer, E.W. 1986b: Metamorphic assemblages and condi­tions in the Rombak basement window, Nordland. Nor.geol. unders. Report 86.168, 25 pp.

Sawyer, E.w. & Robin, P.-Y.F.1986:The subsolidus segrega­tion of layer-parallel quartz-feldspar veins in green­schist to amphibolite facies metasediments. J. Met.Geol. 4, 237-260.

SkiCild, T. 1986: On the age of the Kiruna greenstones,northern Sweden. Precamb. Res. 32. 35-44.

SkiCild, T. 1987: Aspects of the Proterozoic geochronologyof northern Sweden. Precamb. Res. 35. 161-168.

SkiCild, T., & Cliff, RA 1984: Sm-Nd and U-Pb dating ofearly Proterozoic matic-felsic volcanism in nothernmostSweden. Precamb. Res. 26. 1-13.

NGU - BULL. 415. 1989

Skonseng, E. 1985: Berggrunnsgeologisk kartlegging iGautelisomrfldet, Skjomen, Nordland. Feltrapport. Nor.geol. unders. Report 85.214. 16 pp.

Stow, DAV., Howall, D.G., & Nelson. C.H. 1985: Sedimenta­ry, tectonic and sea-level controls. In Bouma, A.H.,Normark, W.R., & Barnes, N.E. (eds.) Submarine Fansand Related Turbidite Systems. Springer-Verlag, NewYork. 15-22.

Sullwold, H.H. 1961: Turbidites in oil exploration. In Peter­son, J.A., & Osmond, J.C. (eds.), Geometry of Sandsto­ne Bodies. Am. Assoc. Petrol. Geol., 63-81.

Taylor, S.R. & Gorton, M.P. 1977: Geochemical applicationof spark source mass Spectrography-Ill. Element sensi­tivity, precision and accuracy. Geochim. Cosmochim.Acta 41. 1375-1380.

Taylor, S.R., & McLennan. S.M. 1981a: The rare earth ele­ment evidence in precambrian sedimentary rocks: impli­cations for crustaI evolution. In Kroner, A. (ed.) Precam­brian Plate Tectonics. Elsevier, Amsterdam, 527-548.

Taylor, S.R., & McLennan. S.M. 1981b: Evidence from rare­earth elements for the chemical composition of theArchaean crust. In Glover, J.E., & Groves, 0.1. (eds.)Archaean Geology: Second International Symposium.Perth. 1980.Spec. Publ. Geol. Soc. Aust. 7,255-262.

Widenfark, L., Claesson, L.-A., & SkiCild, T. 1987: Rocksassociated with uplift of a Proterozoic continental mar­gin, N. Sweden. Abstr. IGCP 217, uma, 92-93.

Winchester, J.A., & Floyd, P.A. 1977: Geochemical discrimi­nation ot different magma series and their differentiati­on products using immobile elements. Chem. Geol. 20.325-343.