The Layered Series


Introduction

The Layered series (LS) is the most voluminous of the primary Skaergaard subdivisions and dominates the exposed parts of the intrusion. It displays a multitude of magmatic structures related to the cumulus and postcumulus stages of the evolution. The magmatic layering, which is the most prominent feature of the LS, displays strong similarities to sedimentary layering and defines the stratigraphy of the series. A suite of included blocks (autoliths and xenoliths) fallen from the roof of the intrusion during crystallisation locally disturb and disrupt the layering with abundant impact structures. Locally the layering is disrupted by unconformities or displays evidence of slumping and redeposition. In places, the layering is transgressed by mafic pegmatites and pods and streaks of anorthosite that apparently postdate its deposition. These structures appear to have partially replaced the original cumulates.
 

Subdivision and nomenclature

The LS forms a stratigraphic cumulate succession of around 2500 metres developed on the magma chamber floor. It evolves systematically from high-temperature cumulates at it's base towards progressively lower temperature cumulates towards it's top. The uppermost stratigraphic level where the LS meets the Upper Border series is known as the Sandwich horizon, which is believed to represent the level where the last residual melt crystallised.

[more]

The LS is subdivided according to changes in the cumulus mineral assemblage, which represents the liquidus assemblages of the Skaergaard magma. As such, it reflects directly the evolution of the Skaergaard magma. The primary subdivision of the LS relates to the presence of olivine in the cumulates. The Lower zone (LZ) has olivine as a fractionating phase, the Middle zone (MZ) has no olivine but locally pigeonite, and the Upper Zone (UZ) is marked by the return of olivine.
 

Stratigraphic relations

Layering is developed throughout most parts of the LS from the deepest exposed levels up to the base of UZc. Several distinct styles can be recognised that obviously formed by distinctly different processes. Where some types are clearly distinguishable by differences in their style and occurrence, others appear to be variations of one of the major types. Commonly distinct types of layering are superimposed on the cumulates. Common for all styles of layering is their orientation parallel to the assumed crystallisation front. The development of layering appears largely to depend on the shape of the magma chamber and the cumulus mineral assemblage, and consequently there are stratigraphic variations in the different major types throughout the LS. Layering is developed on three distinctly different scales that also appear to have different origins.

Rhythmic layering consists of modally graded planar sheets of material (layers) of typically 5-50 cm thickness interbedded with variable amounts of modally uniform (isomodal) gabbro. The layers display strong density stratification and a weaker grain size stratification in their mineral constituents consistent with gravity sorting in suspension currents (Irvine, 1987; Irvine et al., 1998). The individual layers have iron-titanium oxides at their bases followed upwards by olivine, pyroxenes, and finally plagioclase towards their tops. Individual layers can be followed for a maximum of some 300 m along strike after which they taper out and end with slightly upturned edges in ridges of isomodal gabbro (fig. ). Both the upper and lower boundaries of the individual layers are sharp to within millimetres. The rhythmic layering displays abundant structures around included blocks that resemble structures found in clastic sedimentary rocks (described below).

Microrhytmic layering defines a modal stratification on the scale of the crystal size (centimetre scale) within the cumulates. The layering appears mostly as a banding of cumulus mineral crystals within isomodal or modally graded rhythmic layers. The layering is mostly faint, but prominent examples have been found in LZa (McBirney and Noyes, 1979) and LZb (fig. XX) on Uttental Plateau, in MZ below Pukugaqryggen, and in UZb below Basistoppen. The microrhythmic layering is somewhat similar to the inch-scale layering in the Stillwater complex (Hess, 1960). For the Stillwater complex, it has been suggested that this type of layering formed during postcumulus coarsening by a process similar to Ostwald ripening (Boudreau, 1987).

Macrorhythmic layering consists of stratigraphic units of typically 0.5 5 m thickness. The layers have broadly isomodal compositions and are distinguished by different proportions of cumulus minerals. The layers are prominently developed in two successions, the first in the upper half of the LZb, the other in the upper half of the MZ (the so called zebra banding of Wager and Deer, 1939). Individual macrorhythmic layers can be traced for more than a kilometre along strike in the cliff face of Pukugaqryggen, their upper and lower contacts are gradational over decimetres, and the layering relations are easiest to observe from a distance. The individual macrorhythmic layers display no evidence of gravity sorting, but include frequent gravity stratified rhythmic layers. However, the relations of macrorhythmic layers around large blocks are similar to those of the rhythmic layers (depressed below, lapping up against the sides, and get streamlined over the top of large blocks) strongly suggesting that they too were produced during the primary cumulus stage of crystallisation (Irvine et al., 1998). The layers are sometimes compared with the megacyclic units of the Eastern Layered Series on the Isle of Rum, but appear to have closer resemblance to the kakortokite layering in the Ilimaussaq intrusion. No detailed studies have been carried out to clarify the origin of the macrorhythmic layering, but their lateral extent suggest that they formed in response to changes in the precipitation of cumulus crystals. Their origin must be related to fluctuations in the cotectic proportions of cumulus minerals unless a vary large scale process of crystal sorting was active. The layers have been attributed to rhythmic crystal nucleation (Maaløe, 1978), differences in the convective pattern of the magma (Naslund et al., 1991), and repeated gravitational collapse of crystal loaded suspensions from below the magma chamber roof (Brandeis and Jaupart, personal communication cited by Irvine, 1987).

The Triple Group is a prominent succession of very large scale layering occupying the uppermost 100 m of the MZ. The group terminates the succession of macrorhythmic layering in the upper half of the MZ, but has a much larger scale and more extreme modal variations. Furthermore, the individual layers can be recognised across the entire exposure of the LS, and diamond drilling has confirmed that they extend below the surface across the entire southern part of the intrusion. The succession has recently attracted attention because it hosts a potentially economic occurrence of gold and palladium (the Platinova Reefs, described below). The Triple Group members appear to consist of pairs of lower leucocratic and upper melanocratic units, each unit being some 5 15 m thick. The members are separated by variable amounts of mesocratic gabbro. A close examination revealed that the Triple Group members all host modally graded rhythmic layers, and that the interbedded mesogabbro displays macrorhythmic layering on a 5 m scale (Andersen, 1996).
 

Included blocks

The LS includes numerous blocks and fragments of material that appear to be of exotic origin. Abundant blocks of anortosite, gabbroic troctolite, and anorthositic gabbro are spread from the LZa to UZb and appear to be cognate xenoliths fallen from the UBS during crystallisation. The stoping appears to have been very intense in periods, where the blocks make up locally half of the exposure. On the whole, it has been estimated that some 15% of the entire MZ is made up of exotic material (McBirney, 1989). A few hornfelsed metabasaltic inclusions are exposed along the south coast of Kraemer Ø demonstrating that at that particular stage, the stoping had locally scaled away the entire roof section over that area.
The autolithic inclusions are distributed in large swarms with typically several tens or hundreds of blocks of various sizes and shapes. The swarms are stratigraphically controlled and occur at distinct stratigraphic levels in the cumulates. The lowermost swarm is exposed at the boundary between the LZa and LZb on Uttental Plateau, extensive swarms occur in the middle of LZb, in LZc, and throughout the lower half of the MZ. In the MZ, the stoping activity ceases, but solitary blocks are found scattered up until the base of the UZb. Apparently the blocks were dislodged from the UBS at distinct times when the roof of the intrusion became unstable. This could for example have occurred as a consequence of earthquake activity, but it is likely that the incorporation of mafic material would add enough weight to periodically destabilise the roof.
The blocks display a variety of sizes, shapes, and internal structures. The smallest fragments are only a few centimetres across and sometimes appear just as clusters of a few mineral grains barely distinguishable from crystals in their matrix material. Small blocks are commonly distributed in conglomerate beds (fragmental layers) that locally dominate the magmatic stratigraphy. The largest blocks are up to ½ kilometre across and display internal structures such as modal or textural layering, pegmatites, and even in some places smaller included blocks. In shape, the blocks range from angular boulders with straight, clear-cut boundaries to their host cumulates to rounded blocks that appear to have plastically deformed during impact. Larger blocks are mostly elongate slab-like rafts whereas smaller blocks may me more equilateral. Internal structures are abundant within the blocks. Larger blocks display well developed modal or textural layering; some have smaller blocks included; some have dikes that do not extend into the surrounding cumulates; some display hydrothermally altered fractures; and pegmatites and replacement anorthosites are widely developed.
 

Structures associated with included blocks

The layering is commonly greatly affected by the presence of included blocks and displays abundant structures relating to their occurrence. It has been claimed (McBirney, 1989) that the layering is generally best developed in successions with included blocks, and indeed the many structures associated with the blocks are central for the investigation of the physical conditions during crystallisation.
Layers below blocks are generally depressed The abundance of small blocks in some layers
Structures associated with block impact
Block interactions with layering
 Impact structures
 Slumping
 

Replacement anorthosite

The LS locally displays transgressive bodies of anorthosite and leucocratic gabbro (commonly known as replacement anorthosites) that appear to have partially replaced the primary cumulates (cf., McBirney and Sonnenthal, 1990; Irvine et al., 1998). These replacement anorthosites are particularly abundant in the LZa and LZb on Uttental Plateau, where they form small local patches in the cumulates or up to 20 meter sized transgressive or diapiric bodies. Locally, the anorthosites appear to spread laterally with dendritic growth patterns along primary layering planes, and in places they include residual in-situ structures from their protolith (such as layering). The replacement anorthosites are commonly associated with mafic bodies that appear to be most commonly situated at their bases.
The replacement anorthosite bodies can be explained in the context of melting in a crystalline material fluxed by water (Yoder, 1970). Fluxing with hydrous fluids in the system diopside-anorthite-SiO2 results in a shift in the eutectic composition towards anorthite. Partial melting produces liquids rich in plagioclase, and the association between mafic bases and the anorthosites material can possibly be considered as zones of dissolution and reprecipitation of plagioclase (Irvine et al., 1998).
 

Mafic pegmatite

Mafic pegmatite appears to be related to the replacement anorthosites but have a much wider distribution in the LS. They still, however, only form a quantitatively minor rock unit in the intrusion amounting to less than a percent of the exposure. The pegmatites form podiform bodies; stringers and veins associated with the margins of included blocks; and semiconformable sheets following layer boundaries. There seem to be a morphological evolution of the pegmatites with stratigraphy in the LS, from dominantly podiform bodies in the lower parts to more conformable sheets towards the top. Crystal sizes are mostly around 10 cm, although in rare places they can exceed ½ meter. In composition the pegmatites are broadly gabbroic and generally display lower temperature mineral compositions than their host cumulates. The pegmatites are commonly strongly fractionated with plagioclase-rich margins, hydrous gabbroic zones, and granophyric cores; and commonly they are roofed by gabbro showing variable degrees of anorthositic replacement (Larsen and Brooks, 1994).
Podiform and vein-type pegmatites are from a few centimetres to some 10 m across and can have sharp or diffuse boundaries towards their host rocks. Some display well defined semi-circular bodies in places having smaller satellite bodies, others appear as diffuse sprays of coarse grained material. Semiconformable pegmatites are typically 10-20 cm thick and several tens to hundreds of meters long. They have generally sharp boundaries to their host rocks. Some appear to have replaced the felsic parts of modally graded layers, but others display evidence of having intruded with brittle failure along the layering planes. Commonly these pegmatites jump between adjacent layers having the occasional included blocks of their host material.
The pegmatites display equilibrium temperatures from 1006 to 766 ºC consistent with a formation at the late magmatic postcumulus stage (Larsen and Brooks, 1994), and fluid inclusions indicate the involvment of methane-bearing saline hydrous fluids (Larsen et al., 1992).
 

© 2003 skaergaard.org