Quaternary lithostratigraphy and correlation, introduction, Cainozoic of north-east Scotland

From Earthwise
Jump to navigation Jump to search
Merritt, J W, Auton, C A, Connell, E R, Hall, A M, and Peacock, J D. 2003. Cainozoic geology and landscape evolution of north-east Scotland. Memoir of the British Geological Survey, sheets 66E, 67, 76E, 77, 86E, 87W, 87E, 95, 96W, 96E and 97 (Scotland).

Contributors: J F Aitken, D F Ball, D Gould, J D Hansom, R Holmes, R M W Musson and M A Paul.


Key to the most common morpho-lithogenetic symbols used on recently published drift maps of the district. P915252.

Some of the more recently published drift editions of sheets covering the district accommodate lithostratigraphical units of glacigenic deposits in addition to the conventional morpho-lithogenetic categories (P915252). For example, several glacigenic formations have been mapped on Sheet 66E Banchory and Sheet 67 Stonehaven. On other sheets, such as 87W Ellon and 76E Inverurie, conventional morpho-lithogenetic categories of deposit have been assigned to one or more lithostratigraphical group on the basis of gross lithological characteristics. It is beyond the scope of this publication to define and describe every lithostratigraphical unit, but some discussion of stratigraphical correlation is included, especially of the tills, because it is important for deciphering the history of events that have taken place. A link to the generalised maps showing the distribution of drift groups and a selection of deposits, landforms and localities mentioned in the text can be found here:

Elgin district. P915371.
Sheet 96W Portsoy. P915372.
Sheet 96E Banff. P915373.
Sheet 97 Fraserburgh. P915374.
Sheet 86E Turriff. P915375.
Sheet 87W Ellon. P915376.
Sheet 87E Peterhead. P915377.
Sheet 76E Inverurie. P915378.
Sheet 77 Aberdeen. P915379.
Sheet 66E Banchory. P915380.
Sheet 67 Stonehaven. P915381.

Definition of terms[edit]

Lithostratigraphy involves the description, definition and naming of rock units (Whittaker et al., 1991). It is fundamental to stratigraphy because accuracy in biostratigraphy (based on fossils), geochronometry (based on measurements of age in years) or chronostratigraphy (apportioning units of strata to intervals of time) relies on the correct recognition of the relative spatial relationships of rock units, as does subsequent palaeogeographical reconstruction. Individual units are normally described and defined using their gross lithological characteristics and by their interrelationships with adjacent units, but this is not always possible for Quaternary glacigenic sequences, especially when they have been glacitectonised and/or occur as large glacial rafts. Such units are defined primarily on the basis of their unconformable bounding surfaces, rather than on lithology alone (e.g. Whitehills Glacigenic Formation, see below). Strictly speaking, the result is a hybrid of lithostratigraphy and allostratigraphy (Hedberg, 1976; Owen, 1987), the latter being based on the recognition of units bounded by discontinuities and their correlative conformities, such as in offshore seismic sequence stratigraphy.

Correlation of lithostratigraphical units in north-east Scotland. P915347.

In the Quaternary succession of north-east Scotland, discrete lithological units are named, as far as possible, after localities lying close to characteristic sections. They are then ranked in a formal hierarchy of Bed, Member, Formation and Group (P915347). The Formation is commonly defined as the basic mappable unit in the hierarchy. Type sections (stratotypes) are chosen to illustrate typical lithologies, characteristics and relationships to adjacent units, but in complex and commonly poorly exposed areas a single section or borehole sequence rarely suffices and several partial type sections and reference sections may be required. Where formations and groups embrace numerous lesser ranked units, or there is poor exposure, a type area is given instead of a type section. Many type sections and type areas are defined in the descriptions of important localities. All units established by BGS have been entered into the BGS Lexicon of named rock units and the index of computer codes (BGS, 2002).

Numerous units have been named in the literature; most are informal and as far as possible each one has been allocated here to one of the groups and assigned formation, member or bed status (P915347). Names generally have not been changed unless ambiguity exists, where they have been used already for other lithostratigraphical units, or cardinal rules have been broken (e.g. the use of ‘lower’ and ‘upper’ qualifiers).

Since setting up the scheme outlined here, a revised correlation of Quaternary deposits in Scotland has been published by the Geological Society of London as part of a wider review of the British Isles (Sutherland, in Bowen, 1999). No groups are recognised in that publication, the ‘member’ becomes the basic mappable unit and formations are really ‘type sequences’. For example, the ‘Kirkhill Formation’ embraces all units at Kirkhill irrespective of lithology, provenance, genesis and age. Sutherland’s scheme is not adopted here for several reasons. The main objection is that most of his formations are not mappable units. Some include all drift deposits occurring in an area (e.g. Kirkhill) whereas others include only minor elements of the local drift sequence (e.g. Castleton and Camp Fauld). The omission of lithological qualifiers makes it difficult to comprehend what the units are and how they interrelate. There are also numerous spelling mistakes (e.g. Elton instead of Ellon Member) and inconsistencies (e.g. similar locality names appear at both formation and member level). Another problem is that several important papers have been published since the Geological Society Report was in preparation and recent mapping has helped clarify spatial relationships of many deposits away from type localities. For example, new information at Boyne Bay and Gardenstown indicates that the ‘Castleton Formation’ is almost certainly a collection of glacial rafts and not in situ, and the organic deposits at Camp Fauld and Crossbrae are now placed in a different stratigraphical context. In general, Sutherland’s nomenclature has not been adopted here unless new names are required for formerly inappropriately named, or unnamed units.

The groups[edit]

Five lithostratigraphical groups of predominantly glacigenic (glacial, glaciofluvial, glaciolacustrine) deposits have been established in north-east Scotland. They reflect lithology and provenance rather than age and may encompass units formed during several glaciations. Although some nonglacigenic materials such as palaeosols and periglacial deposits are included, these units are closely associated with the glacigenic sequence and represent important events in the geological record of Quaternary events in the district. A ‘top down’ approach has been adopted in order to allow sufficient room in the classification for complex successions and to accommodate more information in the future. It also allows traditional morpho-lithogenetic units to be embraced on sheets where formal lithostratigraphical division has not been attempted.

Generalised flow-lines of ice during the Main Late Devensian glaciation. P915251.

Traditionally, it has been the practice in north-east Scotland to relate the glacigenic deposits to one of three distinct bodies of ice that are thought to have existed in the region (P915251). Three ‘series’ have been informally recognised (Sutherland, 1984a; Hall and Connell, 1991). Ice moving north-eastwards along the east coast led to the deposition of the Red Series (Jamieson, 1906; Synge, 1956), which include a variety of materials of a typically vivid reddish brown colour. Ice moving eastwards along the Moray Firth impinged onto the coastal lowlands and was responsible for laying down a suite of dark grey deposits assigned to the Bluegrey Series (Synge, 1956). In order to complete the picture, the typically sandy, yellowish brown diamictons laid down in the interior by ice flowing from the Grampian Highlands were assigned to an Inland Series (Hall, 1984a).

The tripartite division is retained in the present scheme, but the series are renamed as groups (‘series’ is a chronostratigraphical term). The Blue-grey series becomes the Banffshire Coast Drift Group and the Inland series becomes the East Grampian Drift Group. The Red series warrants division in order to separate deposits that were laid down by ice that moved onshore from the North Sea basin (Logie-Buchan Drift Group) from those that were derived entirely from ice flowing within Strathmore (Mearns Drift Group). Additionally, the deposits laid down by ice emanating from the Central Highlands and entering the area via the Spey valley are assigned to the Central Grampian Drift Group. Clast composition is the most important attribute in deciding from which ice stream a deposit was derived. Matrix colour remains a useful parameter for general classification. However, both methods require care in their application because distinctive clasts may have been reworked from older deposits and incorporation of local rock or sediment can locally change the typical colours of matrices.

The distribution of the five drift groups is shown approximately in P915251 and on the 11 maps mentioned above. The relative positions of the ice streams changed with time during the last glaciation, and the ice streams of earlier glaciations probably did not cover exactly the same ground as in the Late Devensian. There is consequently local, but stratigraphically significant interdigitation of deposits belonging to different groups.

Radiocarbon dates from Late-glacial sites in the district
Site Grid reference Laboratory number Age (year BP) Dated material and setting Reference
Rothes cutting NJ 277 498 Beta 8653 11 110 ± 70 peat under remobilised till Appendix 1
Garral Hill, Keith NJ 444 551 Q-104 10 808 ± 230 peat under remobilised till Godwin and Willis (1959)
Garral Hill, Keith NJ 444 551 Q-103 11 098 ± 235 peat under remobilised till Godwin and Willis (1959)
Garral Hill, Keith NJ 444 551 Q-102 11 308 ± 245 peat under remobilised till Godwin and Willis (1959)
Garral Hill, Keith NJ 444 551 Q-101 11 888 ± 225 peat under remobilised till Godwin and Willis (1959)
Garral Hill, Keith NJ 444 551 Q-100 11 358 ± 300 peat under remobilised till Godwin and Willis (1959)
Woodhead, Fyvie NJ 788 384 SRR-1723 10 780 ± 50 peat under remobilised till Connell and Hall (1987)
Howe of Byth NJ 822 571 SRR-4830 11320 peat beneath gravel Hall et al. (1995)
Moss-side, Tarves NJ 833 318 I-6969 12 200 ± 170 peat under remobilised till Clapperton and Sugden (1977)
Loch of Park NO 772 988 HEL-416 10 280 ± 220 kettlehole infill Vasari and Vasari (1968)
Loch of Park HEL-417 11 900 ± 260 kettlehole infill Vasari and Vasari (1968)
Mill of Dyce NJ 8713 1496 SRR-762 11 550 ± 80 kettlehole infill Harkness and Wilson (1979)
Mill of Dyce NJ 8713 1496 SRR-763 11 640 ± 70 kettlehole infill Harkness and Wilson (1979)
Glenbervie NO 767 801 GX-14723 12 460 ± 130 peat under remobilised till Appendix 1
Glenbervie NO 767 801 SRR-3687a. (humic) 12 305 ± 50 peat under remobilised till Appendix 1
Glenbervie NO 767 801 SRR-368Th (humin) 12 340 ± 50 peat under remobilised till Appendix 1
Brinzieshill Farm NO 7936 7918 SRR-387 12 390 ± 100 peat under remobilised till Auton et al. (2000)
Rothens NJ 688 171 SRR-3803 10 680 ± 100 kettlehole infill Appendix 1
Rothens NJ 688 171 SRR-3804 11 640 ± 160 kettlehole infill Appendix 1
Rothens NJ 688 171 SRR-3805 11 760 ± 140 kettlehole infill Appendix 1

Organic deposits, buried soils and periglacial sediments and structures[edit]

The lower relief areas of north-east Scotland are notable for the occurrence of a significant number and variety of Middle and Late Pleistocene nonglacial, terrestrial sediments. These serve as important stratigraphical marker horizons and provide evidence for ice-free episodes within the Pleistocene succession. Additionally, they provide important, and in some cases unique, evidence from which to reconstruct the climate of these periods.

Organic deposits[edit]

Peat, organic mud and organic-rich sand occur at numerous sites buried by later sediments. Most commonly found are Late-glacial organic deposits (Gunson, 1975), chiefly lake or pond muds and thin peat beds, organic muds and sands (see table above Garral Hill, Byth, Woodhead, Mill of Dyce, Loch of Park and Glenbervie). The pollen record from these sites reveals that in the Windermere Interstadial, opening at around 13 000 BP, an initial phase of open habitat vegetation was succeeded by closed heath and scrub vegetation, with local growth of tree birch and pine. After around 11 500 BP, the temperature began to decline and tundra vegetation had become established by the start of the Loch Lomond Stadial at 11 000 BP.

No deposits have yet been found in north-east Scotland representing the last nonglacial period preceding the Main Late Devensian Glaciation. This ‘Sourlie Interstadial’ (Jardine et al., 1988), dated to around 30 000 BP, has been recognised from sites in western and central Scotland, where good evidence of tundra flora and fauna is preserved. The buried soil at Teindland and the peat at Crossbrae (P915347) were formerly thought to date, at least in part, from this period, but the original relatively young radiocarbon dates from the sites probably result from contamination.

Evidence of older organic sediments has been found, but these deposits are beyond the range of radiocarbon dating. As only a few uranium series and luminescence dates have been obtained, the ages of these deposits are poorly known. However, detailed pollen investigations, and an increasing number of studies of Coleoptera allow tentative biostratigraphic correlation. Evidence of Early Devensian interstadial conditions is found at Camp Fauld, Crossbrae and Burn of Benholm (P915347). The organic materials at these sites appear to be equivalent in age to Oxygen Isotope Stages (OIS) 5c and 5a, but correlation with one or other interstadial remains tentative. Birch and pine woodland was probably extensive in the region during the warmest parts of these interstadials. Episodes with dominant dwarf birch and willow scrub mark cooler episodes and the terminations of the interstadials are presaged by a decline into tundra conditions.

Organic muds and sands associated with podzolic buried soils at Teindland and Kirkhill contain pollen of temperate woodland. At Teindland, the pollen record shows an early phase of pine, alder and hazel woodland with grassland clearings, followed by vegetation indicative of increasingly colder conditions. This appears to mark cooling at the close of the last interglacial, OIS 5e. At Kirkhill, an older interglacial (possibly OIS 7) is represented by thermophilous pollen of pine and alder and very rare grains of birch, lime and elm. This pollen is contained within a unit of organic-rich sand (Swineden Sand Bed), which is believed to have been redeposited from soil horizons in the vicinity. The pollen spectra are dominated by open grassland communities, which were probably present during the redeposition (information from J J Lowe, Royal Holloway University, London, 1990). This unit of sand is conformably overlain by sandy and gravelly gelifluction deposits (Corse Gelifluctate Bed) that formed in a periglacial environment (P915347). As at Teindland, the organic sand contains evidence of climatic deterioration.

Interglacial and interstadial buried soils[edit]

Southern face of Kirkhill Quarry in 1979. P528224.
Partially glacitectonised, podzolic Teindland Palaeosol Bed at Teindland quarry. P104118.

Buried soils and associated weathering profiles occur at several sites and record significant intervals of temperate conditions (P528224; P104118). Truncated podzolic soil profiles of interglacial status, developed in sand and gravel parent materials, occur in association with organic muds and sands at both Teindland (OIS 5e) and Kirkhill (possibly OIS 7) (P915347). A second, younger, truncated soil profile at Kirkhill (Fernieslack Palaeosol Bed) is considered to represent a gleyed ‘brown earth’ of interglacial status (OIS 5e) developed in diamicton parent material (Connell and Romans, 1984). Weathering profiles, with disintegration of clasts and mobilisation and movement of iron and manganese, are developed on the surface of older tills, gravels and sands of the East Grampian Drift Group at Teindland, Boyne Limestone Quarry, Howe of Byth, Moreseat, Tillybrex (Ellon), and Kirkhill (P915347). Around this last site, weathered till and gravel can be traced over an extensive area and represents both OIS 5e and earlier warm intervals.

Periglacial sediments and structures[edit]

A range of periglacial features have been recognised in units at several positions within the Pleistocene sequence. These include mass movement deposits, frost shattered bedrock, fluvial sands and gravels and structures indicative of the former presence of ground ice, notably cryoturbation structures and ice-wedge casts. It is clear that periglacial processes have had a significant impact on the landscape of north-east Scotland during most of the Quaternary, particularly across central Buchan. Intense periglacial activity last occurred during the Loch Lomond Stadial when marked slope instability, with destruction of soils and remobilisation of diamictons by gelifluction, led to slope-foot accumulations of significant thickness, locally burying organic sediments of Windermere Interstadial age.

Evidence of older periods of periglacial activity includes a bed of sandy gelifluctate with ice-wedge casts that lies beneath rafts and diamicton of the Whitehills Glacigenic Formation, and on slightly weathered gravel at Oldmill. These features probably represent one of the cold periods of the Middle or Late Devensian. Organic sediments of probable OIS 5e age at Teindland and either 5c or 5a age at Camp Fauld and Crossbrae record a decline to tundra conditions and the increasing slope instability is manifest as increased accumulation of sand, gravel and diamicton.

Undoubtedly the clearest record of recurring events is from Kirkhill, where quarrying activities over a period of 10 years allowed detailed resolution of multiple periglacial phases (P915347. Extensive mass movement occurred towards the end of the Main Late Devensian glaciation and possibly also in the Loch Lomond Stadial (Manse Gelifluctate Bed). Cryoturbation, solifluction and ice wedge growth is also recorded in the Corsend Gelifluctate Bed lower in the sequence. The latter bed contains elements that probably formed both at the close of the last interglacial (OIS 5e) and prior to the complex phase of glaciation that deposited the Corse Diamicton and overlying Hythie Till (P915347). ‘Silt-droplet fabrics’ have been revealed by micromorphological examination and probably record the development of an immature soil, or soils, close to a periglacial landsurface (J C C Romans, personal commnication in Connell, 1984c). These events may also be represented in an equivalent bed at Oldmill.

Another sandy gelifluctate bed in the Kirkhill/Leys sequence (Corse Gelifluctate Bed) is associated with an ice-wedge cast network. It has been dated by luminescence to 142 ± 19 ka yr BP (Duller et al., 1995) and if correct, marks an important periglacial period within OIS 6. The oldest periglacial sediment recognised at Kirkhill is the angular felsite rubble of the Kirkton Gelifluctate Bed, which must date to OIS 8 or an older cold period.


Full reference list