Silurian conglomerate and tectonics within the Fennoscandian continental margin, Sunnhordland, western Norway

ROALD B. FÆRSETH & RONALD J. STEEL

Færseth, R. B. & Steet, R. J.: Silurian conglomerate sedimentation and tectonics within the Fennoscan­ dian continental margin, Sunnhordland, western Norway. Norsk Geologisk Tidsskrift, Vol. 58, pp. 145- 159. Oslo 1978. ISSN 0029-I%X. A thick succession of Silurian conglomerates on Stord, show ing clast-supported textures but no internat stratification, has been interpreted as resedimented. A distinct upwards change in composition from mainly sedimentary clasts to mainly igneous clasts, with upwards increasing , suggests a shift in direction of drainage through time, with the lates! phase of disposal from an uplifted igneous complex along the basin's northwest margin. The succession of conglomerates overlying tholeiitic basalts, graptolitic shales, and carbonates is interpreted in terms of an intercontinental rift basin sequence, marginal to a continent-based magmatic are. The conglomerates record the rifting, early dispersal from the cratonic side of the basin, the later uplift of the magmatic are, and the gradual unroofing of its plutonic core.

R. B. Færseth, Geologisk institutt, avd. A, Universitetet i Bergen, 5014 Bergen, Norway. Present adress: Norsk Hydro A/S, P. O. Box 490, 1301 Sandvika, Norway. R. J. Steet, Geologisk institutt, avd. A, Universitetet i Bergen, 5014 Bergen, Norway.

The sedimentary history of the earl y Palaeozoic conglomeratic part of which is discussed here, geosynclinal belt in western Norway is still situated in a narrow zone in the centre (Figs. poorly documented. A recent synthesis of the l and 7). The Sunnhordland lgneous Massif con­ succession in the Trondheim area, together with sists of acidic (455 ± 5 Ma, Priem & Torske 1973) new data on the chemistry of the widespread and basic lavas, pyroclastics, gabbros, granites, Cambrian to Lower Ordovician basic volcanic and local peridotites. Studies of the volcanics rocks in the sequence, led & Roberts (1974) (Furnes & Færseth 1975) have suggested an in­ to propose that the early Palaeozoic geography tracontinental environment. Within the Dyvik­ of western Norway was dominated by a western vågen Group there are tholeiitic basalts belt (Cambrian/early Silurian) composed of an (post-lower Llandovery) with a trace element island are complex developed on oceanic crust geochemistry characteristic of ocean floor basalt with a back-are eugeosynclinal pile, and an east­ (Furnes & Færseth 1975) and these are interpret­ em belt (late Precambrian/Ordovician) of mio­ ed in terms of crustal thinning and separation geosynclinal character which developed on con­ within the Dyvikvågen basin, presently lying on tinental crust. the southeast margin of the Sunnhordland lg­ Based on lithological similarities a common neous Massif. The conglomerates of the Utslet­ origin has been assumed for the lower Palaeozo­ tefjell Formation dominate the Dyvikvågen ic successions in the Trondheim and Sunnhord­ Group on Stord (Fig. 1). The northwestern and land regions (e.g, Strand 1960, 1972). However, southeastern contacts of the conglomerate belt recent studies of the allochthonous lower Palae­ have previously been considered to be tectonic, ozoic succession in the region south of Bergen, possibly thrusts, and it has generally been as­ on Stord and BØmlo (Fig. 7), suggest that there sumed that the northwesterly contact was of the his tory of events was different from what has greater importance since it separated been suggested for the Trondheim area. from igneous rocks (e.g. Strand 1972). We have In the Stord/Bømlo region there are three main remapped these boundaries and suggest that the groups of rocks: the Sunnhordland lgneous Mas­ northwesterly one was a major Silurian normal sif in the west, a more metamorphosed and de­ which shows evidence of post-conglomer­ formed groupof dominant!y phyllites and volcan­ ate reactivation. As a corollary there has been no ics in the east, and the Dyvikvågen Group, the major horizontal movement along this tectonic

lO- Geologisk Tidsskr. 3/78 146 R. B. Færseth & R. J. Steet NORSK GEOLOGISKTIDSSKRIFT 3 (1978)

Leg end 1 OYVIKVÅGEN GROUP---""( �·��::;� ��:���:it:��}U���� l��6�LL

r:;�-�s�;\J NONSHAUG FORMATION

- VIKANES FORMATION

� LIMBUVIK FORMATION

F----·-1 FAULT

LEGE NO:

f{(;{;}:jDYVIKVÅGEN

EillillACID IGNEOUS R r---i BASIC/INTERMEDIATE � VOLCANICS � �GABBRO

le�I PEAT

N

______Q�----s�q�o 1o�,oo __ 2��om NORSK GEOLOGISKTIDSSKRIFT 3 (1978) Silurian conglomerate sedimentation and tectonics 147 line and the Dyvikvågen Group squarely belongs bedding is steep and the foliation is seen to be to the igneous belt (Fig. 7). sub-parallel with general NE-SW trends. Where the bedding is at a lower angle of dip, the long axes of the clasts, which Iie within the Sl folia­ The Dyvikvågen Group - tion, are often oriented at a marked angle to the stratigraphic context bedding. Late Dl slides parallel to the Sl folia­ tion are common within parts of the conglomer­ The Dyvikvågen Group consists of four forma­ ate and usually coincide with local zones of tions: strong deformation. Such movements

Utslettefjell Formation - 1000 m polymict con­ make accurate measurements of the succession glomerates. thickness impossible. However, detailed sedi­ Nonshaug Formation - max. 220 m tholeiitic mentological measurements across the conglom­ volcanics, including pillow Javas (Furnes & erate body suggest no major fault repetitions and Færseth 1975),. the total thickness of the sequence probably Vikanes Formation - max. 110 m graptolitic exceeds 1000 m. shales, early Llandovery age (Skevington pers. comm. in Færseth & Ryan 1975). Limbuvik Formation - minimum 25 m shelly The conglomerates of the 1929). limestones, Ashgillian (Kiær Utslettefjell Formation The conglomerates dominate the Group and are well exposed for 20 km both to the northeast Two general aspects of thick conglomeratic suc­ of Dyvikvågen and on Bømlo, the island south of cessions often provide dues to the geographic Stord (Fig. 7). At Dyvikvågen (Fig. l) the con­ and tectonic his tory of the area containing them. tact between the graptolitic shales of earl y Llan­ In the conglomeratic portion of clastic strata, dovery age and the overlying conglomerate is distinctive compositional features of their source gradational and in places the appear are usually accentuated (Walton 1963) and, al­ to interfinger. though associated may come from The conglomerates have undergone two pha­ more distant sources, the clasts in conglomeratic ses of Caledonian folding; the first phase was sequences therefore provide an unusually clear accompanied by a lower greenschist facies me­ record of the sites and ex tent of parent drainage tamorphism. areas (e.g. Bluck 1969). On the other hand, the

A Rb-Sr age ( Åab = 1.39 X lO- Il yr- l) of very coarseness of such cl�stic sequences is like­ 409 ± 15 Ma has been reported for a secondary ly to be an indication of contemporary, rapid cleavage developed in greenschist facies meta­ basin subsidence while the variation of coarse­ morphosed sediments on Hardangervidda (And­ ness may provide additional information about resen et al. 1974). This age was interpreted to the type or intensity of tectonism (e.g. Steel & represent a late phase of the Caledonian orogeny Wilson 1975). In addition, details of the internal affecting both the Hardangervidda-Ryfylke organisation of conglomerates, such as the tex­ Nappe System and the autochthonous sedi­ ture and structure of individual beds, may pro­ ments. As the rocks of the Dyvikvågen Group vide information about the mechanisms and apparently have been subjected to this phase of processes involved in their transportation and deformation and , we assume that deposition (e.g. Bluck 1967, Davis & Walker the conglomerates were deposited between early 1974). Considering the relative ease with which Llandoverian and late Silurian time. most of the se properties can be measured (assu­ In the rocks of the Dyvikvågen Group major ming bedding can be identified) it is surprising noncylindrical folds and a regional foliation (Sl) how few detailed studies have been made in were developed during the first deformation conglomerate sequences. phase (Dl) affecting these rocks. During this The present study of the conglomerates of the phase of deformation, the fragments having an Utslettefjell Formation has dealt largely with oblate shape were rotated into a position parallel general aspects. Tectonic deformation, espedal­ to the axial planes of the folds. Normally the ly in the lower part of the succession (quartzite

Fig. I. Map of the Dyvikvågen Group, showing its position between the igneous complex on NW Stord and the sedimentary/vol­ canic sequence of SE Stord. loset, simplified geological map of Stord after Kvale (1937) and Skordal (1948). 148 R. B. Færseth & R. J. Steel NORSK GEOLOGISKTIDSSKRIFf 3 (1978)

/:!7 Granitic

fSTI2l L=U Quartzit•

Sondstont

Basic

Acid volcanic • Othors

Mean max, clost Sondst.7. Clast composition size (cm.} NORSK GEOLOGISKTIDSSKRIIT 3 (1978) Silurian conglomerate sedimentation and tectonics 149 conglomerates), made detailed examination ex­ are rare in the con­ tremely time-consuming. Five profiles, involving glomerates but cross-stratification or parallel detailed logging of pebble composition, sedimen­ lamination is found in some of the sandstones tary structures, maximum particle size, and- in interbedded with the finer conglomerates. Mas­ places - bed thickness, were made over various sive beds are most common in the parts of the body (Fig. l). The main aim was to coarser parts of the succession. The quartzite identify any significant vertical variations conglomerate, apart from its basal 100 m, de­ through the conglomerates and to interpret these creases in coarseness upwards (Figs. 2 and 4). in terms of the prevailing tectonic situation. In The conglomerate matrix and the sandstone addition, at several localities, details of the tex­ beds are mainly lithic and feldspathic grey­ ture of individual conglomerate beds were inves­ wacke. Conglomerate clasts, which are usually tigated. well-rounded, include quartzite, sandstone, The Utslettefjell Formation consists of two limestone, granite, granodiorite, quartz diorite, compositionally distinct sequences, now inform­ gabbro, basic volcanics, and chert. ally named the quartzite and granitic conglom­ A detailed analysis of the conglomerate facies erates respectively (Fig. 2). The latter usually across the entire body has not been attempted, overlies the former along a well-defined contact but detailed observation at a number of localities (Fig. l), although the two parts of the succession indicates that there is considerable variation in interfinger at the northeastern end of the area the nature of conglomerate beds or units present. (Figs. l and 3), indicating that no stratigraphic Bedding is usually marked by variation in the break separates the two conglomerates. At the grain size of adjacent units, either by sandstone southwestern end of the body the conglomerate interbedded with conglomerate or, more often, succession appears to rest conformably upon by interbedding of fine and coarse-grained con­ Lower Silurian graptolite schists of the Vikanes glomerates. The latter feature, interbedding of Formation (Fig. 1). In the same area, however, conglomerates with differing mean and max­ about O.S km north of Limbuvik, an unconform­ imum parti eie size, is a very stri king aspect of able junction has been interpreted in terms of a the succession at most localities. This basic al­ large erosional-channel infilling (Færseth & Ry­ ternation together with vertical gradients of grain an l97S). size either within a bed or within a group of beds, variation in the amount of matrix in the grain population, and the presence or absence of !ami­ nation in the associated sandstones, produces a Quartzite conglomerate considerable range of sedimentary units in the The quartzite conglomerate sequence appears to conglomerate body. Those illustrated in Figs. S have a maximum thickness of approximately 600 and 6 have been distinguished in the coastal suc­ m. Detailed logging at certain localities (Fig. l) cession (A-A', Fig. l) at the SW end of Stord. shows that bed thickness rarely exceeds 4 m, at Individual beds with abrupt or erosional upper which point mean maximum clast size reaches and lower boundaries con sist either of conglom­ 30-40 cm. In general there is a decrease in bed erate or sandstone. The former are usually mas­ thickness with decreasing clast size (Fig. 4). The sive, reverse- or normal-graded (in order of sandstone percentage (as discrete sandstone abundance). The latter may occur in single sets beds) in the sequence varies from 5 to 15% and (Fig. 6A, 6D) or in cosets (Fig. 5A, 5C) and may this also appears to be inversely related to con­ be parallel or cross-laminated (Fig. SD, 6D) or glomerate coarseness (Fig. 2). As regards the normal-, reverse- or non-graded (Fig. 5C, 5D). composition of , it is noticeable that Beds commonly occur together in groups, where coarser beds typically contain more granitic the beds often have indistinct or non-erosional clasts. boundaries and where there is a trend (discontinuous) of grain size. Such groups of beds probably have a close depositional or

Fig. 2. A composite profile through the UtslettefjeU Formation time-relationship and are here called units. As (C-C' /B-B' are located on Fig. l) showing mean maximum regards the relationship of sandstone beds to particle size, sandstone percentage, and clast composition (of such conglomerate units, it is extremely com­ the 'maximum particle size' fraction) trends. Note the major mon to tind laminated sandstones abruptly cap­ clast size and clast composition change at the base of the granite conglomerate. ping .units (Fig. SA, SD, 6D), white massive 150 R. B. Færseth & R. J. Steet NORSK GEOLOGISKTIDSSKRIFf 3 (1978)

SW NE LOC. D LOC. E

20 40 60

Meon max. clost· size ( cm l

B%i��Z;a Gronite conglomerote

l ;::j Quortzite

Sandstone m 20 40 1 M.C.S.

Fig. 3. Profiles through the north-eastern end of the conglomerate body (D-D', E-E' in Fig. l) showing probable interfingering.

sandstones usually occur at the base of re­ Small-pebble populations at the top of these verse-graded units (Fig. 5C, 50). units are often hetter sorted and less well­ Coarsening-upwards conglomerate beds or cemented than similar size fractions in coars­ units are common. Where there are only two ening-upwards units. The or beds or grain populations in the unit, the up­ conglomerates of fining-upwards units are often wards coarsening is often restricted to a thin poorly sorted frameworks (Fig. 6E) or are mat­ bottom portion (Fig. 6A, bottom unit). In other rix-rich (Fig. 6C, top). Sometimes there is a thin, cases it occurs on a larger scale, and here the basal coarsening-upwards portion to a fi­ coarsest part of the unit may be either mat­ ning-upwards unit (Fig. 6C, middle unit). rix-rich (Fig. 6B, bottom unit) or a tight frame­ The examples of conglomerate beds and units work (Fig. 5B). Where there are three beds or described above (Figs. 5 and 6) may not be rep­ populations in the coarsening-upwards unit, resentative for the whole conglomerate body. each part may show little overlap in mean or Most of them have been selected from only 100 range of grain size (Fig. 6A, top unit), or the m of a well-exposed coastal profile (Fig. 4) which middle population may appearto be a mixture of is relatively fine-grained. In the inland segments the upper and lower (Fig. 6B, top unit). of the conglomerate body, which are both coars­ The fining-upwards conglomerate units are er grained and more thickly bedded, we expect usually more obviously composite than those that a correspondingly greater thickness of the which coarsen upwards because their two or succession is represented by massive beds. three component beds are often bounded by However, the log of mean maximum clast size in more obvious erosion surfaces (Fig. 6D). the central area emphasises that a dominant NORSK GEOLOGISKTIDSSKRIFT 3 (1978) Silurian conglomerate sedimentation and tectonics 151

Q) .... ca ... Q) E o a c: o o

Q) .... c: ca ...

., ....

ca ...

., E o

a c: o 100 o

....

N ...... ca :l o 10 20 30 o M. P. S. cm.

·• f.

Fig. 4. Sequence through part of the UtslettefjeU Formation on the SW coast of Stord (A-A' in Fig. 1). Examples of textures discussed and figured (Figs. 5, 6) are located in this profile. Note the overall fining-upwards trend in the quartzite conglomerate succession. 152 R. B. Færseth & R . .f. Steel NORSK GEOLOOISKTIDSSKRIFT 3 (1978)

@ © @ � d='0-12 . //·/ ////// )/• � is�·.·��:;5i:),Z,�'IfS_{ '· ' ' -<��j9· ��· · � . "/.

·· .. �: .. ·:-' · : .. ...· ·- .· .":.·.··-:.' ·. ·

d=7-10

Fig. 5. Examples of some of the different types of sandstone beds in the Utslettefjell Formation. Arrows indicate direction of tining within beds or units. The mode of the 'maximum particle size' fr action in conglomerate beds is shown (D). Erosive or sharp boundaries are shown by solid lines. Conglomerate textures are large!y diagrammatic. Some of the examples are located in Fig. 4.

attribute of the sequence here also is the rapid range. Locally, granite of maximum vertical alternation of coarse and fine-grain po­ diameter 1-2m occur. Plutonic clasts, particular­ pulations (Fig. 2). ly of granite and gabbro, now dominate the suc­ cession, while both quartzite and chert clasts are relatively rare (Fig. 2). The body has a thickness of approximately 400 Granitic conglomerate m, but thins out both southwestwards and north­ Conglomerate beds in the granitic conglomerate eastwards (Fig. l). The Iower half of the succes­ body resemble those of the quartzite conglom­ sion appears to coarsen upwards white the top erate in having an abundance of pebble or cobble half fines upwards (Fig. 2). frameworks and a Jack of cross-stratification. Deformation in the granitic conglomerate se­ Again, the characteristic feature of 'organisa­ quence is evident from the steeply dipping strata tion' is a grain-size differentiation (Fig. 6E). In and from several NE-SW trending slides probab­ contrast, discrete sandstone beds are extremely ly of similar age to the late Dl slides in the rare in this succession (Fig. 2) and the matrix quartzite conglomerate. Pebble deformation and which infills the pebble or cobble frameworks cleavage are conspicuously absent. consists, in most instances, of small pebble or grade rather than coarse or medium-­ grained sandstone. Transporting mechanisms In general, granitic conglomerate beds are thicker than in the quartzite conglomerates and The conglomerates described above have an ap­ the mean clast size is increased to the 40-70 cm parently anomalous combination of attributes. NORSK GEOLOGISKTIDSSKRIFf 3 (1978) Silurian conglomerate sedimentation and tectonics 153

®

�������d�=!1-�2� � . •. -�.. - ,:·;:::: ·:· .'.·�. ���:-\ :�_,: ;:_, • ::::.':: · l :>'i<.<:-=..-<:H"> . 75 ��d@-=1-2 cm d=1-2 · : . . .

d=1Z-14

®

Fig. 6. Examples of some of the different types of conglomerate beds in the Utslettefjell Formation. Textures are largely diagrammatic. Some of the examples are located in Fig. 4. Symbols as in Fig. 5.

The conspicuous lack of stratification within ment to be important and have interpreted in conglomerate beds suggests a lack of turbulence terms of fully turbulent currents (Walker 1970, and bed-load traction· during sedimentation (at Rocheleau & Lajoie 1974). least in the sense in which equally coarse-grained Of considerable interest is the recent sugges­ fluvial conglomerates become abundantly strati­ tion that a continuum may exist between certain fied (e.g. Bluck 1967, Steell974), while the abun­ types of mass flow and turbidity currents (Mid­ dance of clast-supported beds and the prominent dleton 1970, Middleton & Hampton 1973). In a alternation of coarse and fine-grained conglom­ recent discussion of some Cambro-Silurian con­ erates in the sequences suggest the importance glomerates (with many characteristics similar to of a sorting process during sedimentation: On those of the Utslettefjell Formation), Davis & the one hand a mass flow or other process invol­ Walker (1974: 1211) have proposed a 'turbulent ving deposition from suspension or dispersion is suspension (that is a flow transitional between suggested; on the other hand, a process invol­ grain flow and turbidity current)' to account for ving fluid turbulence or some other mechanism their origin. According to Davis & Walker capable of sorting sediment. (1974), a combination of fluid turbulence and The same problem has evidently faced other dispersive pressure enabled gravet to be marine conglomerate workers, and they have supported above the bed during transportation. often judged one or the other of the above sets of Clasts were oriented parallel to flow with long attributes to be more prominent in the rocks, or axes dipping upstream and clast sorting took to be more informative about sedimentation place. mechanisms. Some have emphasised mass flow We suggest that similar mechanisms and proc­ properties and have invoked (Hen­ esses may account for the apparently contradic­ dry 1973) or grain flow (Aalto & Dott 1970); tory features of the conglomerates of the Utslet­ others have considered grain-by-grain move- tefjell Formation. Because tectonic deformation 154 R. B. Færseth & R. J. Steet NORSK GEOLOOISK TIDSSKRIFf 3 (1978) has often caused some pebble flattening and ro­ that resedimented cobble and cobble/boulder tation we cannot apply one of their arguments, conglomerates do not always exist in their own namely that dast a-axis alignment parallel to right (cf. Davis & Walker 1974: 1206) but may flow precluded dast rolling on the bed. However, have a dose relationship to finer grained assem­ the features of the conglomerates described blages. We have not analysed enough beds to above, particularly the lack of stratification, demonstrate this confidently as a general feature suggests that bed-load traction was relatively of the Stord conglomerates, but we have identi­ unimportant during their transportation. fied particular examples of it. Reverse graded beds, and possibly also some Two important characteristics of the conglom­ of the reverse-graded units (Figs. 5 and 6), em­ erate beds throughout most of both sequences phasise that the grave! dispersions had a high are the presence of pebble or cobble frameworks sediment concentration (Fisher 1971). We sug­ and the conspicuous Jack of sedimentary struc­ gest that the massive conglomerates may have tures indicative of traction currents. As regards been deposited under similar conditions. On the the latter point it should be noted that gra veis are other hand it is likely that sandstone beds which apparently not always able to 'take up' many of cap some of the units and the conglomeratic the structures which appear in (B. J. Bluck upper parts of some others, were deposited by pers. comm. 1976) although, on the other hand, normal traction currents because of the parallel as discussed below, the re are many documented or inclined lamination in the sandstones and the examples of coarse conglomerates with abun­ fact that the upper parts of conglomerate beds dant structures. Sedimentary structures are con­ are often lensoid, are finer grained, better sorted fined to some of the sandstones which accompa­ and have less matrix content than the underlying ny conglomerate beds or to the uppermost part main part of the conglomerate units. of the conglomerate beds themselves. In dealing with their resedimented conglomer­ As suggested above the conglomerates of ate sequences, Davis & Walker (1974) have sug­ the Utslettefjell Formation are probably gested two depositional models, an inverse­ resedimented in a marine depositional environ­ to-normally graded one and a graded-stratified ment. A his tory of reworking is sugge sted by the one; they have also appealed for data from rese­ good rounding of the clasts, particularly those in dimented conglomerates in other areas for the the quartzite conglomerates, and by the fact that testing of their descriptive models. From our these rounded dasts now form the framework of limited measurements and observations on the a deposit which is poorly sorted. That their final Stord resedimented conglomerates we offer the site of deposition was marine is strongly following comments: suggested by their association with pillow lavas Unlike Davis & Walker (1974) we have obser­ and particularly their gradual transition and pro­ ved reverse grading in coarse pebble conglome­ bable interfingering contacts with the underlying rates and in granule sandstones (contrast Figs. 5 marine Llandovery shales. According to the and 6 with fig. 2 of Davis & Walker 1974). These simple scheme suggested by Walker (1975), the occurrences are both within beds as well as with­ clast-supported nature of the Utslettefjell con­ in what we acknowledge to be units. Even in glomerates suggests that they are not de bris flow the latter case we suggest that the components of deposits, while their rare cross-stratification and units have been deposited with a dose time rela­ the fairly common grading suggest resedimenta­ tionship to each other. tion rather than fluvial or shoreline processes. We have observed pebble conglomerates grad­ ing up into cobble conglomerates (Fig. 6B. Cont­ rast with Davis & Walker 1974: 1204). Provenance and direction of Up to 40% of the sandstone beds in the Stord sediment dispersal succession are non-laminated. These usually occur at the base of reverse-graded conglomer- . Due to the deformation of the conglomerate ate units, and immediately overlie the promi­ body and the present steep dip of the beds, direct nently laminated sandstone which cap units (eg. palaeocurrent measurements are of little value. Fig. 5D). The laminated sandstones are usually Systematic pebble composition measurements less well cemented than the apparently massive were therefore of prime importance in attempt­ sandstones. ing a palaeogeographic reconstruction. In a gen­ In the light of our observations it is possible eral sense such measurements indicate the natu- NORSK GEOLOGISK TIDSSKRIFT 3 (1978) Silurian conglomerate sedimentation and tectonics 155 re of source areas and, indirectly, pro bable di­ supply across the northwestern boundary was rection of sediment dispersal (bearing in mind dominant. The quartzite conglomerates, domi­ that structural deformation may cause a relative nant at an earlier period, are likely to have been displacement of source and depositional areas). derived in the opposite direction, across the When measured systematically, compositional southeastern boundary of the bas in. trends may suggest sudden or gradual, verti­ cal/lateral uncovering of the source area. The value of the compositional parameter is in­ creased if evaluated together with grain size, Some tectonic implications particularly with the size of the coarsest fraction of each bed. In the case where the composition Any thick sequence of coarse clastic sediments profile shows distinct influxes of different pebble implies the prior existence of a deep sink within types, consideration of grain size may allow the which the sediment accumulated or relative sub­ distinction between a situation in which different sidence of the sub-stratum during the sedimenta­ rock types were being tapped from the same tion. In either instance, vertical movements of source area and one where two different source the earth 's crust are implied. In the case of the areas, at significantly differing distances from thick Utslettefjell conglomerate succession, the the locus of sedimentation, supplied their own periodic influxes of coarse plutonic debris from clasts. the northwest margin of the basin, in the upper The suggested provenance of the main com­ half of the sequence (Fig. 2), indicate contem­ ponents of the conglomerate body is shown in poraneous subsidence. In this respect the verti­ Fig. 7. Granitic and basic plutonic clasts are cal variation of conglomerate coarseness in easily matched to the igneous complex presently reflecting varying proximity or relief of the lying along the northwest margin of the source area is likely to be a useful tectonic conglomerates. Two types of basic volcanic rock index, particularly if that variation is present that differ particularly in trace element content throughout the length of the body. (Furnes & Færseth 1975) present! y occur respec­ Throughout much of the quartzite conglomer­ tively to the northwest and to the southeast of ate succession there is a trend of fining-upwards the conglomerate belt. Chemical analyses of (Figs. 2 and 4), despite the likelihood of some of greenstone pebbles have revealed that clasts of the coarsest modes being due to minor influxes both types occur in the conglomerates, sugges­ of granite conglomerate from the north west side ting sediment contribution across both margins. of the basin (Figs. 2 and 7). This is suggestive of Distinctive chert clasts occur preferentially in gradually diminishing n!lief in the (quartzite) the lower part of the succession and are compa­ . source area or source-wards retreat of the slopes rable to cherts in the area to the southeast. Se­ across which the grave! was supplied. It is not dimentary clasts such as quartzite and sandstone certain that either of these changes need have are particularly common low in the succession. been directly tectonically controlled. They may The latter are probably at !east partly rather passively reflect the Jack of any renewed intraformational, and the former, although more uplift of the drainage areas to the southeast. problematic, are most likely to have been der­ The sudden influx of much coarser sediment at ived from the cratonic or south and southeastern the base of the granite conglomerate succession side of the basin. probably indicates sudden uplift and the creation There is a distinctive compositional change in of a new and major drainage area across the the conglomerate sequence through time. At the northwest margin of the basin. Debris shed base of the granite conglomerate, plutonic clasts southeastwards off the igneous complex then become dominant (Fig. 2) and indicate the sud­ dominated the basin in the area studied (Fig. 7). den new importance of the igneous complex as a There is no obvious time trend of coarseness in drainage area (Fig. 7). Despite this sharp compo­ this conglomerate sequence common to each of sitional change, occasional influxes of coarse­ the measured sections. However, these con­ grained plutonic material are evident at lower glomerates, compared to the quartzite con­ stratigraphic levels (Fig. 2) and influxes of glomerates, are generally very coarse grained quartzitic clasts at higher levels (Fig. 2). This (Fig. 2) and are probably proximal represen­ implies that sediment was commonly shed from tatives of the facies, so that small changes in a more than one direction, but that at later times local variable such as direction of sediment dis- 156 R. B. Færseth & R. :1. Steet NORSK GEOLOGISK TIDSSKRIFT 3 (1978)

Fig. 7. A possible. simplified map of the palaeogeology and palaeoge­ ography during the deposition of the lower and upper parts of the conglomerate of the Utslettefjell Formation succession in Sunnhord­ land. The size of the arrows reflects the relative importance of the sedi­ ment contribution across the basin margins.

N Branile !Sunnhordland Gabbroolcanics Massiflgneous

l� l? ....

? l

l NORSK GEDLOGISK TIDSSKRIFT 3 (1978) Silurian conglomerate sedimentation and tectonics 157 persal may significantly influence sedimentation Silurian time many areas show evidence of the time trend at these localities. accumulation of shallow marine clastic and car­ The regional tectonic context of sedimenta­ bonate sequences (Table 1). According to the tion, an uplifting magmatic are of continental model of Gale & Roberts (1974), based on data margin type within which there is a developing mainly from the Trondheim area, this lower Pa­ rift basin, is discussed below. At this stage it is laeozoic succession accumulated in a back-are worth commenting on some of the possible eau­ basin behind an island-are/trench system which ses of pronounced vertical crust movements in existed some way off the present west and cen­ such a situation. Vertical motions of crust and tral Norwegian coast. lithosphere are commonly related to changes in Within the area studied, the present authors crustal thickness, in thermal regime and in the suggest additional complexity in Ordovician time conditions for isostatic balance (Dickinson when there appears to have been a local thinning I974b). In the present case the conglomerates of the continental crust with later extrusion of immediately overlie basalts of ocean-floor type Silurian tholeiitic basalts of ocean floor type and (Fig. l) but are banked against and derived from accumulation of conglomerates in a deep, fault­ continental source areas (Fig. 7). In such a situa­ controlled basin. We interpret the present geo­ tion where thin oceanic crust had been created graphical distribution of rock groups in the area adjacent to thick continental crust a sediment (Fig. 7) as follows: trap could have been predicted, simply on the The Ordovician igneous activity represented basis of crustal elevation differences. The magni­ by the Sunnhordland lgneous Massif shows fea­ tude of the sediment trap was probably further tures of continental rift volcanism producing a exaggerated by increased uplift of the continen­ thick pile of rhyolites (ignimbrites as well as lava tal crust on the northwest side of the basin, as flows) and basalts. On northern Stord the suggested by the major influx of the granite con­ volcanic succession, interpreted as subaerial glomerates in the upper part of the succession. (Kvale 1937), exceeds 3500 m in thickness. Ig­ This may have been an istostatic response to the nimbrites have also been identified on Bømlo thickening of the crust by intrusion of the igne­ (Songstad 1971). On Stord, recent analysis of the ous bodies, an effect apparently more common basic volcanics indicates alkaline/calc-alkaline for continental margin magmatic arcs than for composition (Furnes & Færseth 1975). The initi­ intraoceanic arcs underlain by thin crust (Dick­ al Sr17/Sr&' ratio of 0.7071 obtained by Priem & inson 1974a). Torske (1973) is within the range of continental rhyolites and the igneous activity probably cor­ responds to stages of thinning of the continental crust. On Bømlo there are local pillow Javas and The tectonic context on both Bømlo and Stord there are local but The lower Palaeozoic succession of western thick accumulations of volcano-clastic sedi­ Norway (Table l) displays some characteristics ments, suggestive of local basin development of accumulation in geosynclinal basins of within the igneous complex. These supracrustal back-are or Japan Sea-type (Gale & Roberts rocks were subsequently intruded by gabbros, 1974). Extrusion of submarine basic volcanics, granitic rocks, basaltic dykes, and local perido­ initially of ocean-floor type, was widespread in tites. The entire massif has been affected by late Cambrian to earliest Ordovician times. Dur­ greenschist facies metamorphism, and, on Stord, ing the remainder of the Ordovician, flyschtype Kvale (1937) showed that the original features of sediments accumulated, although island-are type the igneous rocks are best preserved in the basic and acid volcanics are locally important. In southeast and that the degree of deformation most regions during this time parorogenic increases towards the northwest. movements are suggested by the presence of The Dyvikvågen Group occurs in a narrow local polymict conglomerate sequences, al­ belt on the southeast side of the igneous complex though in the Bergen area a major phase of de­ (Fig. 7). It is interpreted in terms of relatively formation is implied prior to the deposition of the late stages in the evolution of a rift basin (Ashgil­ Moberg conglomerate by the presence of peb­ lian-post-Lower Llandovery). Above a se­ bles having a pre-conglomerate metamorphic quence of thin carbonates there are graptolitic fabric and by the angular unconformity at its shales and pillow lavas which are compositional­ base (Kvale 1960, Sturt & Thon 1976). By early ly comparable to ocean-floor tholeiites (Furnø� 158 R. B. Færseth & R. J. Steel NORSK GEOLOGISK TIDSSKRIFT 3 (1978)

Table l. Correlation between generalised Lower Palaeozoic successions in western Norway.

Age Trondheim area Sunnfjord area Bergen area Sunnhordland area P. " -"' ... "' Graptolite shales - Shallow marine Shallow marine Resedimented - P. · u· bl) ::l " "' 5 ... o Shallow marine �g clastics clastics " conglomerates o ... ""bl) .3� :tO clastics �� Graptolite shales Submarine basic �P. Shallow marine ·- ::l volcanics > o » ... carbonates 00 Graptolite shales Shallow marine carbonates " ·;; o p.. ::l Mostly flysch Mostly flysch Basic and acidic l: o . ... "" igneous rocks ::>O (basic and acidic " volcanics are � (Sunnhordland " "' p.. "' locally important) ::l igneous massif) l: o :g " > ·;; ;jtJ o o p.. Most!y flysch Conglomerates Conglomerates "E ::l l: o "" (Stubseid) (Moberg) o " ...i Cl "' stratigraphic break major strati- l: p.. " ::l Mostly flysch graphic break l: o pre-conglomerate ...i Cl deformation of "' the substrate " " p.. "p.. " ::l " ... o Submarine basic ::l Submarine basic Submarine basic !Sl ... "'> ...o til O volcanic tectonic ti) O volcanics volcanics contact "'" "'" Late Precambrian- .s p.. Mostly pelites Mostly pelites ] B "' e Ordovician, "' o "' "" ... u Gula Group :tO

Vogt 1945 Skjerlie 1969 Roberts et al. 1970 Skjerlie 1974 Kolderup & Færseth & Ryan 1975 Kolderup 1940 Gale & Roberts 1974 Indrevær & Steel l975 Kvale 1960 Present account

& Færseth 1975). This may be interpreted in the psammites, cherts, and local thin conglomerate context of crustal thinning and incipient conti­ layers, together with spilitic, submarine volcan­ nental separation, with the development of a rift ics, and quartz keratophyres. The succession is basin. From a consideration purely of the isostat­ of uncertain age, although on Bømlo (Fig. 7) an ic balance of the crust, such a location would be Ordovician age has been ass u med for a sequence a site of potentially thick sedimentation. The of similar rock types (Songstad 1971). These very thick overlying conglomerate sequence re­ rocks also exhibit a slightly higher grade of me­ cords the development of actively elevating high­ tamorphism and a more complex structural his­ lands, first on the southeast side (the cratonic tory than those belonging to the Dyvikvågen side) of the basin and later of the magmatic are Group, thus indicating that they are older than itself, with gradual unroofing of gabbroic and the late Ordovician/early Silurian rocks. granitic plutonic bodies. Because of the present The association of rock types is not diagnostic outcrop distribution of this sequence and, in par­ of a particular environment, but the occurrence ticular, the absence of its southeastern margin, it of local conglomerates and abundant calcareous is impossible to comment either on the width of material may suggest periods of shallow marine the trough during this period of uplift and subsi­ conditions. dence, or on the likelihood of the trough marking The contact between this southeastern sedi­ the position of transition from craton to magmat­ mentary and volcanic sequence and the rock ic are. groups to the northwest (Fig. 7) is of a tectonic The rocks occurring to the southeast of the origin. The contact to the southeast of the Ut­ Dyvikvågen Group (Fig. l) are a dominantly sed­ slettefjell conglomerates with greenstones of imentary sequence of phyllites, calcareous supposed Arenig age (Kiær 1929, Skordal 1948, NORSK GEOLOGISK TIDSSKRIFT 3 (1978) Silurian conglomerate sedimentation and tectonics 159

Strand 1972) is dipping 60°-70° NW where both record rifting and uplift of the magmatic are and rock types are strongly sheared in a zone about the grad ual unroofing of its plutonic c ore. 15 m in breadth. This contact could either repre­ The present outcrop of the Dyvikvågen sent a normal fault or a thrust which has been Group, probably only a fragment of the original subjected to later reactivation. basin, has, together with the igneous complex, As emphasized above, the present northwest­ possibly been later thrust eastwards onto a ern boundary of the conglomerate body, against sedimentary/volcanic succession of assumed the northern igneous complex (Fig. 7), has at no Ordovician age. time been a major thrust zone (contrast with Strand 1960, 1972). Any major post-early Siluri­ Aeknowledgements. - We thank Drs. K. Bjørlykke, B. J. Bluck, G. H. Gale, and D. Roberts for critically reading the an thrusting in this area is more likely to have manuscript and suggesting improvements. been located along the southeastern boundary of July 1976 the conglomerate.

References

Summary Alto, K. B. & Dott, R. H. 1970: Late Mesozoic conglomeratic flysch in southwestern Oregon, and the problem of transport The upper Ordovician/Silurian sedimentary suc­ of coarse grave! in deep water. Geo/. Assoe. Canada Spee. cession on Stord, western Norway, consists of P aper no. 7, 53--65. Andresen, A., Heier, K. S., Jorde, K., & Naterstad, J. 1974: shallow marine carbonates, graptolitic shales, Preliminary Rb/Sr geochronological study of the Hardang­ tholeiitic basalts, and a thick sequence of marine ervidda - Ryfylke nappe system in the RØldal area, S. Nor­ conglomerates. way. Nor. Geo/. Tidsskr. 54, 35-47. A study of the conglomerates shows that they Bluck, B. J. 1967: Deposition of some Upper Old Red ­ stone conglomerates in the Clyde area: a study in the signifi­ have an apparently anomalous combination of cance of bedding. Seottish J. Geo/ogy 3, 139-167. attributes. The lack of stratification and the Bluck, B. J. 1969: and otber Palaeozoic common presence of reverse-grading in beds conglomerates. Am. Assoe. Petr. Geo/. Mem. no. 12, suggests that they originated by a mass flow 711-723. Davis, l. C., & Walker, R. G. 1974: Transport and deposition process, while the abundance of clast frame­ of resedimented conglomerates: the Cap Enrage Formation, works and relatively good sorting suggest turbu­ Cambro-Ordovician, Gaspe, Quebec. J. Sed. Petro/ogy, 44, lent currents. 1200-1216. Examination of the pebble types suggests that Dickinson, W. R. 1974a: Sedimentation within and beside an­ cient and modern magmatic arcs. In Dott, R. H. Jr. & they were derived from source areas on either Shaver, R. H. (eds.), Modern aild ancient geosynclinal sed­ side of the basin. The composition time-trend imentation. Soc. Eeonomie Pa/eont. Mineral. Spee. Publ. indicates that initially mainly sedimentary mate­ 19, 230-239. rial was derived across the southeast margin but Dickinson, W. R. 1974b: Plate tectonics and sedimentation./n Dickinson, W. R. (ed.), Tectonics and sedimentation. Soe. that later, erosion and dispersal from an igneous Eeonomie Pa/eont. Mineral. Spee. Publ. 22, 1-27. complex dominated. Fischer, R. V. 1971: Features of coarse-grained, high-concen­ The overall time-trend of grain size through tration fluids and their deposits. J. Sed. Petrology 41, the succession, coarsening-upwards of the con­ 916-927. Furnes H. & Færseth, R. B. 1975: Interpretation of prelimina­ tribution from the igneous source, and a fining­ ry trace element data from lower Palaeozoic greenstone upwards in that from the sedimentary source sequences on Stord, west Norway. Nor. Geo/. Tidsskr. 55, suggest that the basin was being filled under the 157-169. influence of progressive uplift of the igneous Færseth, R. B & Ryan, P. D. 1975: The geology of the Dyvik­ vågen Group, Stord, Western Norway and its bearing on the area through time. litho-stratigraphic correlation of polymict conglomerates. The sedimentary sequence is interpreted in Nor. Geo/. Unders. 319, 39-45. terms of deposition in a rift basin within the Gale, G. H & Roberts, D. 1974: Trace element geochemistry of Fennoscandian shield; the igneous complex, Norwegian Lower Palaeozoic basic volcanics and its tecton­ ic implications. Earth Planet Sei. Letters, 22, 380-390. with early volcanic episodes (subaerial acidic Hendry, H. E. 1973: Sedimentation of deep water conglomer­ and basic) and later basic and acid intrusions, is ates in Lower Ordovician rocks of Quebec-composite bed­ seen as a continental margin are positioned large­ ding produced by progressive liquefaction of sediment. J. ly to the west of the basin. The ocean-floor tho­ Sed. Petrology 43, 125-136. Indrevær, G. & Steel, R. J. 1975: Some aspects of the sedimen­ leiitic basalts in the basin suggest local thinning tary and structural history of the Ordovician and Devonian and separation of the crust either behind or with­ rocks of the westernmost Solund islands. Nor. Geo/. Under. in the are area. The overlying conglomerates 317, 23-32. 160 R. B. Færseth & R. J. Steet NORSK GEOLOGISK TIDSSKRIFT 3 (1978)

Kiær, J. 1929: Den fossilførende ordovisiske-silurske lagrekke Songstad, P. 1971: Geologiske undersøkelser av den ordovi­ på Stord. Bergens Mus. Årbok 11, 8--75. ciske lagrekken mellom LØkling og Vikafjord, Bømlo, Kolderup, C. F. & Kolderup, N. H. 1940: Geology of the Sunnhordland. Unpublished cand. real. thesis, Universitetet Bergen Are System. Bergens Mus. Skr. 20, 1-137. i Bergen. Kvale, A. 1937: Et kaledonsk intrusiv- og effusivfelt på Stord. Steel, R. J. 1974: floodplain and piedmont Bergens Mus. Årbok, Naturvitensk. rekke no. l, 1-138. sedimentation in the Hebridean Province, Scotland.J. Sed. Kvale, A. 1960: The nappe area of the Caledonides in western Petro/ogy 44, 336--3 57. Norway. Exeursion guide. Nor. Geo/. Unders. 212e, 1-43. Steel, R. J. & Wilson, A. C. 1975: Sedimentation and tecton­ Lippard, S. J. 1976: Preliminary investigations of some Ordo­ ism (?Permo- Triassic) on the margin of the North Minch vician volcanics from Stord, West Norway. Nor. Geo/. Un­ Basin, Lewis.J. geo/. Soe. Lond., 131, 183-202. ders. 327, 41--66. Strand, T. 1960: Cambro-Silurian deposits outside the Oslo Middleton, G. V. 1970: Experimental studies related to prob­ region. In O. Holtedahl (ed.), Geology of Norway, Nor. lems of flysch sedimentation. Geo/. Assoe. Canada Spee. Geo/. Unders. 208, 170--284. Paper no. 7, 253-272. Strand, T. 1972: The Norwegian Caledonides: VI, areas with Middleton, G. V & Hampton, M. A. 1973: Sediment gravity Cambro-Silurian deposits along the west coast of south flows: mechanics of flow and deposition. In Middleton, G. Norway, 67-71./n Strand, T. & Kulling, O. (e d. ), Scandina­ V. , & Bouma, A. H. (eds. ), Turbidites and de ep water vian Caledonides. lnterscience Publishers. sedimentation. Paeifie Coast Seetion, Soe. Eeon. Paleont. Sturt, B. A. & Thon, A. 1976: A comment on 'The age of Mineral, Los Angeles. 1-38. orogenic deformation in the Swedish Caledonides'. Discus­ Priem, H. N. A., & Torske, T. 1973: Rb-Sr isochron age of sion. Am. J. Sei. 276, 385-390. Caledonian acid volcanics from Stord, Western Norway. Vogt, T. 1945: The geology of the Hølonda-Horg district, a Nor. Geo/. Unders. 300, 83 ...g5. type area in the Trondheim region. Nor. Geo!. Tidsskr. 25, Roberts.D., Springer, J. & Wolff, F. C. 1970: Evolution of the 449-528. Caledonides in the northern Trondheim region. Central Walker, R. G. 1970: Review of the geometry and facies organi­ Norway: a review. Geo/. Mag. 107, 13 3-145. zation of turbidites and turbidite-bearing basins. Geo/. As­ Rocheleau, M. & Lajoie, J. 1974: Sedimentary structures in soe. Canada Spee. Paper no. 7, 219-251. resedimented conglomerate of the Cambrian flysch, L'Islet, Walker, R. G. 1975: Conglome rate: sedimentary structures Quebec, Appalachians. J. Sed. Petro/ogy 44, 82�36. and facies models. In Depositional Environments as lnter­ Skjerlie, F. J. 1969: The pre-Devonian rocks in the Ask­ preted from Primary Sedimentary Structures and Stratifica­ voll-Gaular area and adjacent districts, Western Norway. tion Sequence, S.E.M.P. Short Course Notes No. 2. Nor. Geo/. Unders. 258, 325-359. Walton, E. K. 1963: Se dimentation and structure in the Skjerlie, F. J. 1974: The Lower Palaeozoic sequence of the Southern Uplands. In: The British Caledonides. Edinburgh, Stafjord district, Sunnfjord.Nor. Geo/. Unders. 302, 1-32. O liver and Boyd. 280 pp. Skordal, A. J. 1948: Vulkanitter og sedimenter på sør-østre de l av Stord. Univer. Bergen Årbok, Naturvitens. rekke no. 2, 1-58.