How does the glaciologist and geologist know, when, for example, the Scandinavian ice sheet of the Last Glaciation has melted away? How does he know, when the Pleistocene Epoch there has ended, and when the Holocene Epoch has begun? – Usually – among other things -, by radiocarbon dating the oldest sediments at the bottom of lakes and ponds, left behind on the deglaciated area by the melting ice. How reliable are these radiocarbon dates? And how reliable is the time-scale of the late Pleistocene and early Holocene, based on these radiocarbon dates? How trustworthy are they? What have scientists found out about this now?
D. G. Sutherland, University of Edinburgh, states under the heading, “Problems of Radiocarbon Dating Deposits from Newly Deglaciated Terrain: Examples from the Scottish Lateglacial”, in Studies in the Late Glacial of North-west Europe, Oxford, 1980, pp. 139-149:
“Discussions of the accuracy of the basal radiocarbon dates often include statements to the effect that the dates may have been influenced by the incorporation of ancient carbon derived from the neighbouring bedrock, either by subaquatic photosoynthesis (the well-known ‘hard-water error’ first documented by Deevey et al., 1954), or by the inclusion of carbon such as graphite in unaltered form. This section examines the possibility that newly deglaciated terrain is likely to give rise to hard-water error in basal radiocarbon dates.” (1980:142).
“The possibility of hard-water error, it is sometimes argued, is diminished by the lack of calcareous rocks or erratics within the relevant catchment area. This is an important consideration, even in the Grampian Highlands of Scotland, for limestones are widely distributed in the Dalradian rocks. It must not be forgotten, however, that few rocks exist that do not contain carbon. The Handbook of Geochemistry (1966; pp. 6-E-1, 6-E-2, 6-M-1) gives lists of abundance of carbon in rocks. Of particular interest for the Scottish Highlands are the carbon contents of granitic and high grade metamorphic rocks, the average contents of which are very similar, being in the case of granitic rocks an elemental carbon content of between 110 and 270 ppm C and a carbonate content of between 200 and 1,100 ppm CO2. The overall total carbon mean figure is 600 ppm (0.06%).
“The likelihood of such apparently small quantities of carbon influencing radiocarbon dates depends upon (1) release of the carbon from the rocks and (2) incorporation into biogenic material. It is argued below that a glacier sliding over bedrock provides a mechanism to release the carbon and hence that newly deglaciated terrain is characterized by soils and lakes in which ancient carbon is in relative abundance and available for synthesis. Evidence from both recently and formerly deglaciated terrain is cited to demonstrate that such ground is chemically and hence biologically distinct.” - Sutherland, G. Dd. (1980:142).
“A temperate glacier, sliding across its bed, moves past obstacles in the bed partly by deformation of the ice and partly by pressure melting on the stoss side of the obstacles with accompanying regelation on the lee side (Patterson, 1969 p. 119). This latter process results in the solution of bedrock material on the abraded stoss side where meltwater is produced followed by incorporation of the solutes into the regelation ice forming on the lee side. An increasing number of studies have been made of the chemical composition of regelation ice (Souchez et al., 1973; Souchez and Lorrain,1975, 1978; Hallet et al., 1978) and such ice is found to be enriched in Na, Ca, K and Mg. An extreme example occurs in areas of calcareous bedrock, the stoss side meltwater becoming supersaturated with CaCO3 and on pressure release on the lee side precipitating the excess carbonate (Ford et al., 1970; Hallet, 1976) to form small flutings of calcium carbonate which are dissolved upon deglaciation over a 10-100 year timescale, allowing the ancient carbon to be incorporated into soils, by plants, and into lake sediments.”
Comment: Regelation: The freezing again of water derived from ice melting under pressure, when the pressure is relieve.
“The immediate post-glacial environment is strongly influenced by the chemical characteristics of the bare mineral soil exposed on glacier retreat. In front of glaciers presently retreating, pH values of over 8 are frequently recorded from unvegetated soils (e.g. Goldthwaite et al., 1966; Moiroud and Gonnet, 1977), with alkalinity on bare soils not necessarily being dependent upon nearby calcareous bedrock as is specifically pointed out by Braun-Blanquet (1965). In formerly deglaciated terrain chemical analyses of basal sediments reveal relatively high values of Ca, Mg and Na at the base (Pennington et al., 1972). Pollen analyses of basal sediments indicate that vegetation typically today of calcareous or base-rich conditions was typical of pioneer plant communities (Gray and Lowe, 1977). The analysis of the lipid fraction of sedimentary organic matter (Cranwell, in Pennington, 1977b) shows that basal sediments were deposited in lakes of high trophic status. Cranwell supposed that these nutrient-rich conditions resulted from leaching of glaciated rock surfaces, in addition to which, it is suggested here, a considerable contribution must initially have resulted from the glacial deposits.” (1980:142, 143).
“Of particular interest to radiocarbon dating is Cranwell’s observation that much of the (low) carbon content in Lateglacial or early Flandrian sediments is derived from aquatic plants and microbial sources. As these sources of organic carbon derive their carbon from the water in which they live, they are liable to act as agents for the assimilation of ancient carbon in the water. In postglacial sediments it is observed that the trophic status of the lake declines with the establishment of soil and vegetation cover, and the contribution of aquatic plants and plankton the organic component of the sediments decreases.
“Basal radiocarbon dates are typically derived from sediments containing 1-4% carbon (cf. Pennington et al., 1972). Olsson (1972, Fig. 6) gives a graph relating the amount of contaminant, the age difference between the sample age and the age of the contaminant, and the error in the measured age. From this graph it can be seen that c. 6% contamination with ancient carbon will produce a 500 year error in an assay and c. 12.5% contamination with similar material will give an error of 1,000 years. Thus for a dated sample containing only 1% organic carbon to be 500 years in error, it need contain only c. 0.06% carbon originally derived from bedrock. This sample suggests that with samples in which there are low carbon contents, even if only the average carbon concentration of graphitic rocks is made available for incorporation into the sediments being dated, a serious error could result.” - Sutherland, G. D. (1980:143).
Meltwater
How much ancient carbon does the meltwater in recently deglaciated areas contain? And how will this meltwater affect radiocarbon dating?
D. G. Sutherland: “The ice that melts from a kettle-hole or dead-ice hole and which provides much or all of the initial water in the kettle-hole is the basal ice of the ice mass that covered the area.
“Upon melting the CO2 in the meltwater would rapidly equilibriate with the atmospheric CO2 and would seem to be present no greater hazard to radiocarbon assay. Melt-out, however, as discussed below, takes place over a period of time and thus the meltwater constantly adds CO2 (which may be up to several thousand years old in the case of the ice sheet) to developing kettles. This mechanism therefore functions rather like ‘old’ ground water seeping into a lake and producing a hard-water effect.
“Carbonaceous material incorporated in glacial deposits or contained within the ice itself will be released on deglaciation and washed into lakes or kettle holes. Graphite derived from bedrock and organic carbon originating in previous interstadials or interglacials are the substances in this category most likely to influence basal radiocarbon dates, and though their contaminating effects have been realized elsewhere (Donner and Jungner, 1974) they have not been discussed in connection with radiocarbon dates from the Scottish Highlands.
“Graphite is widely distributed in the metamorphic rocks of the Scottish Highlands (Strahan et al., 1917) and the outcrop of the graphite shits and slates of Dalradian age is shown in Fig. 1 (= in his article) as an example. Graphite presents a problem to radiocarbon dating for it is a contaminant of ‘infinite’ age and cannot be removed by the standard pre-treatment of samples in radiocarbon laboratories. On samples of low carbon content relatively small amounts of graphite can produce a major dating error. For example, Ostrem (1965) indicated an error of over 3,000 years for carbonaceous material from ice in a Neoglacial ice-core moraine, the sample on later laboratory analysis having been shown to consist of only 0.07% weight of graphite. Ostrem was able to identify the presence of graphite by microscopic examination and produced a rough quantitative estimate of the amount of graphite present...
“Organic material derived from earlier interstadials or interglacials is a potential contaminant which may be particularly troublesome in the case of glaciers of the Loch Lomond Stadial. This can be said from the glaciers where short-lived and much of the deposits would be reworked ice sheet drift, which had previously been covered by soil and vegetation. This contamination will be difficult to identify, for example by pollen analyses, for the vegetation that immediately post-dated the Loch Lomond Stadial was rather similar to that which characterised the Lateglacial Interstadial.” – Sutherland, D. G. (1980:144).
Sediment Mixing
How are sediments mixed, when older carbon is mixed with younger carbon? How will this affect radiocarbon dating?
D. G. Sutherland: “Sediment mixing results in older carbon being moved up the sediment column and younger carbon being moved down. In a uniform column of sediment a sample slice from the middle may be presumed to have equal amounts of young and old carbon mixed, hence effectively cancelling each other out in the radiocarbon assay.” (1980:146).
Conclusion
What may we conclude now from these findings about radiocarbon dating sediments in recently deglaciated areas? How reliable are they?
D. G. Sutherland: “In the above consideration of the factors that may affect a basal radiocarbon date an attempt has been made to show that newly deglaciated terrain is distinctive in its chemical and hence biological character. The importance of this to radiocarbon dating is twofold: firstly that glaciation makes available ‘old’ carbon from a variety of sources – rocks, ancient carbon dioxide, reworked organic matter; and secondly that the lack of soil development and the sparse vegetation on newly deglaciated terrain result in lake sediments being deposited that are low in carbon content, with a high proportion of this carbon being derived from aquatic sources.
“The low carbon content of the samples dated has resulted in the sample being vulnerable to relatively small amounts of contamination. Certain views, (e.g. Shoton, 1967; Bowen, 1978) have tended to diminish this influence of ancient carbon as a contaminant since the amount of ancient carbon necessary to produce a given error is considerably greater than the amount of modern contaminant that will produce the same magnitude of error (e.g. for a sample 12,000 years old, 10% ancient carbon produce an error of +850 years, whilst 10% of modern carbon produces an error of –2,400 years). With samples, however, of low total carbon content the relatively high levels of contamination necessary for ancient carbon to affect the dates can be achieved, especially in an environment favourable for its synthesis.
“Additionally, laboratory pre-treatment of samples is complicated by frequent lack of knowledge of the origins of the carbon in the sample. Graphite, for example, cannot consistently be removed from a sample even should it be identified.” - Sutherland, D. G. (1980:148).
Lake Manitoba
Old carbon in young sediments at Lake Manitoba, in southeastern Canada: How does it affect radiocarbon dating?
E.M.V. Numbudiri, James T. Teller, and W. M. Last, Dept. of Earth Sciences, University of Manitoba, Winnepeg, report: “The presence of pre-Quaternary microfossils and anomalously old radiocarbon dates from fine-grained organic material in Lake Manitoba sediment suggests that old noncarbonate carbon is contaminating these fine-grained deposits.
“These pre-Quaternary microfossils form as much as 80% of the total palynomorph assemblage in the early postglacial spruce-dominated zone. Their percentage gradually decreases upward in the sediment column until, at a depth of 5 m, no pre-Quaternary microfossils are present.” (1980:123, 124).
“Many of these old microfossils occur in the Upper Cretaceous sediments of the western interior of North America. The bedrock surface to the west and south of Lake Manitoba is composed dominantly of Cretaceous and Jurassic shales with interbedded lignites, siltstones, and carbonates. The bedrock surface is covered by a variable thickness of glacial sediment rich in Cretaceous shale components.
“The oldest known lacustrine sediment in the Lake Manitoba basin was deposited during the Lake Agassiz phase of the lake (about 12,500 to 9,000 B.P.). During most of this time, the lake boundaries were much expanded from those of today, and waves lapped up against the Cretaceous shale of the Manitoba escarpment, 60 km to the west and south.
“The large proportion of pre-Quaternary microfossils in the lower part of the sediment column was introduced into the basin at this time, as waves attacked the western shoreline and meltwater-swollen rivers flooded across the Cretaceous bedrock of the western interior. As the continental ice sheet retreated farther northward, the areal extent of the lake diminished, and the western shoreline no longer supplied significant quantities of sediment to the Lake Manitoba basin.” (1980:124, 125).
“Although contamination of Quaternary lake sediments by pre-Quaternary microfossils has been recognized by others, no one has acknowledged that such contamination is symptomatic of another problem, namely, contamination by other pre-Quaternary organic matter. It seems likely that, in pre-Quaternary microfossils are being eroded from rocks and glacial overburden of the watershed, other old organic matter also will be derived from the shales and, particularly, from the lignites [coals].
“The amount of organic matter in many of the shales of western Manitoba is large, and low-grade coal is common in the Tertiary rocks to the southwest of Lake Manitoba (Bannatyne, 1978). Although the actual percentage of the noncarbonate carbon (NCC) contamination is very difficult to establish when only finely disseminated components are present, we think that the percentage of old microfossils can be used as a guide.
“Most importantly, the presence of old NCC (noncarbonate carbon) in lucustrine muds will result in radiocarbon dates that are too old. In Lake Manitoba, radiocarbon-date ‘errors’ range from more than 25,000 year near the base of the postglacial lacustrine sediment to only about 2,000 year midway in the section. The fact that these dates do not accurately represent the age of the sediment is known from well-established geological and phytogeographical events in the region.
“For example, the decline of Picea (spruce) in the region occurred between 10,000 and 11,000 B.P. (Richie, 1976) and is found between the 10- and 11-m depth in the core... In addition, sediment below this depth contains abundant ice-rafted clasts attributable to the early Lake Agassize phase of the lake that ended about 11,000 yr ago. Therefore, the lab date of more than 29,000 B.P. at a depth of 10.2 m is in substantial conflict with other, independent evidence.” – Nambudiri, E.M.V. et al. (1980:125).