Quaternary oxygen isotope chronology Unlike the continental sedimentary record, ocean-floor sediments can often provide an undisturbed record of environmental change for the entire Quaternary. One of the main techniques used for correlation and climatic reconstruction during the past 40 years has been oxygen isotope analysis of calcareous foraminifera within deep-sea cores.
Foraminifera are marine animals which range in size from less than 0.40 mm to 10 cm. Many species secrete a shell of calcium carbonate. When these animals die, their shells preserve a record of the nature of the water, including its oxygen content, during their lifetime. Two isotopes of oxygen are important in this form of analysis:
16O (oxygen-16) and
18O (oxygen-18). The lighter
16O is, on average, about 500 times more common than the heavier
18O. There is a preferential uptake of
16O into the atmosphere during evaporation. During glacial periods the ocean tends to become enriched in the heavier
18O because much of the evaporate containing the lighter
16O is locked within the ice sheets. During interglacial periods the ice caps melt and return the isotopically light water to the oceans, where it mixes rapidly.
The difference in the oxygen isotopic composition of the ocean between glacial and interglacial periods is about 1.0‰ (1 part per thousand). This is shown quite clearly by the isotopic composition of the skeletons of the foraminifera in sediments on the sea floor. This isotopic composition depends, in fact, on two main factors: the isotopic composition of the water during life, and the temperature. When the temperature is low there is a greater fractionation of
18O relative to
16O, and the foraminifera shell contains a higher proportion of the heavier
18O than it would if the temperature were higher.
Benthic species of foraminifera are subject to a much smaller range of temperature during their life than planktonic species which float freely in the upper 50 m of ocean, because the temperature of oceanic deep water rarely fluctuates by much. For this reason the changes in the
18O/
16O ratio of benthic foraminifera are thought mostly to represent ocean isotopic changes which are caused by the changing volumes of continental ice sheets.
The amount of
18O/
16O is measured during a mass spectrometer on hand-picked foraminifera, usually from the same species. The result is then related to a standard (for deep-sea cores this is usually a reference carbonate) and the results are expressed as relative deviations in parts per millilitre relative to the standard. For example, a result of –4‰ would mean that the sample was 4 parts per millilitre deficient in
18O relative to the standard.
There are some problems with stable oxygen analysis of foraminifera shells from deep-sea sediments:(1) Some species of foraminifera migrate through the water column during their life cycle, which means that juveniles and adults of the same species can have different
18O/
16O ratios.(2) Physiological differences exist between different types of foraminifera which result in large differences in the
18O/
16O ratio between species.(3) Disturbances of the top 20 cm of sediment by burrowing animals can blur the isotopic signal in the record.(4) Below a depth of between 3 and 5 km, the
carbonate compensation depth (CCD), the carbonate in foraminifera shells begins to dissolve.(5) Near to continents and on steep slopes the sediments can be disturbed by terrestrial sedimentation, slumping, or downslope currents.These factors can be minimized by taking cores with high sedimentation rates from ocean basins above the CCD and by using foraminifera from the same life stage of a single species. Often, however, benthic species may be present in very small numbers relative to planktonic species. It may then be necessary to use mixed assemblages or, alternatively, planktonic species whose ecology and life style are well documented.
Oxygen isotope analyses have now been performed on cores from all the main oceans and seas of the world. The most surprising result from these analyses is the degree of similarity in the
18O/
16O trace between oceans, reflecting the dominance of the signal caused by the changing isotopic composition of the oceans in response to changes in ice volume on the continents. Because of the essential synchronism of the record, once a reference curve has been calibrated for age it can then be used for global correlation.
The usual independent dating control is to use radiocarbon dating in the part of the core that is younger than about 35 000 BP. In addition, the first major magnetic reversal in the core is then identified and assigned an age of about 730 000 bp, assuming that this represents the change in polarity between the Brunhes and Matuyama epochs. Levels in between are then interpolated. Further checks can be employed using other data, such as the Barbados high sea-level stands, which have been dated to about 125 000, 130 000, and 82 000 bp using uranium-series dating of raised coral reefs. These high stands should correspond to peaks in the
18O/
16O diagram.
One early standard record came from the upper 16 m of core V28–238 which was obtained from a depth of 3120 m on the Solomon Plateau of the western Pacific. Nicholas Shackleton of Cambridge University and Neil Opdyke of the Lamont-Doherty Geological Observatory published a much-cited paper in 1973 in the journal
Quaternary Research in which they identified 22 stages covering the Brunhes and Matuyama magnetic epochs. For the upper 230 cm, covering the past 130 000 years, they sampled the core at 5 cm intervals and hand picked samples of the foraminifera for oxygen isotope analysis. They used eight specimens of the planktonic foram
Globigerionoides sacculifera for each determination. Most horizons were analysed in triplicate, resulting in a precision of ±0.07‰. For the remainder of the core they sampled at 10 cm intervals and analysed in duplicate.
Shackleton and Opdyke interpreted the record in terms of continental ice-volume changes and assigned ages to each stage boundary. The core then became a template for correlation and was subsequently widely used by geologists interpreting the terrestrial record.
In 1976 a further seminal paper appeared in the journal
Science. The authors were James Hays of Columbia University and the Lamont-Doherty Geological Observatory, John Imbrie of Brown University, and Nicholas Shackleton of Cambridge. They analysed three measures of global climate from two cores, RC11–120 and E49–18, which were located centrally between Africa, Australia, and Antarctica. The indices used were the oxygen isotope record of planktonic foraminifera, an estimate of summer sea-surface temperature derived from analysis of the remains of radiolarians (a group of marine planktonic animals with silica shells), and the percentage of a particular radiolarian,
Cycladophora davisisana. Hays and his co-authors subjected the data to spectral analysis, which can show if there is any cyclicity in the record, and found that the variation of climate recorded in these cores could be explained by the interaction of three different cycles with periods of 23 000, 42 000, and 100 000 years. These periods are almost identical to the Milankovich cycles of precession of the equinoxes, axial tilt, and eccentricity of the orbit. It was concluded that changes in the Earth's orbit are the fundamental cause of the succession of Quaternary ice ages.
In 1984 a standard reference curve for the past 800 000 years was published (Fig. 1c). Instead of relying on just one curve, the SPECMAP (spectral mapping) curve was constructed by averaging (stacking) oxygen isotope profiles derived from planktonic foraminifera from several cores from low latitudes. John Imbrie and co-workers then used the Milankovich orbital cycles to adjust or ‘tune’ each curve in order to produce a type section against which other curves could be compared.
With the large number of high-quality hydraulic piston cores retrieved by the Deep Sea Drilling Program, and latterly the Ocean Drilling Program (ODP), spanning the whole Quaternary and beyond, the method of producing a reference curve tuned with respect to the orbital cycles has now been extended back some 5 million years into the Pliocene. Spectral analysis at intervals of approximately 0.5 Ma intervals from ODP core 659 confirms that for much of the Pliocene and Quaternary the axial tilt cycle of 41 000 years seems to have been the dominant factor in forcing climatic change. At about 1 million years ago, however, the 100 000-year eccentricity and 23 000-year precessional cycles become the major climatic forcing mechanisms.
The SPECMAP model has recently been challenged by a well-dated
18O record from a terrestrial source, Devils Hole, Nevada. Oxygen isotope measurements have been taken from a core (DH-11) of vein calcite (travertine) 36 cm long and fitted to a timescale based on 21 high-precision replicated uranium-series dates. This record reflects groundwater recharge rather than ice-volume fluctuations and is claimed by Isaac Winograd and his colleagues in the United States Geological Survey to be inconsistent with the Milankovich theory. There is in fact quite good agreement between the two records over the last 500 000 years. There are, however, two intervals of significant difference, around 140 000 and 450 000 bp (Fig. 2b). The discrepancy around 140 000 bp seems to be the most problematic since the uncertainties in the uranium-series dates are only about 2000 years at this point.
One view is that because the main discrepancy is restricted to one time interval it is unlikely that there is a continuous bias affecting one or other of the records. It could arise because the Devils Hole chronology reflects a regional rather than global response to climate change. The arguments about this discrepancy continue, and the outcome will be of paramount importance to our understanding of the mechanisms that force climate change.
B. A. Haggart
Bibliography
Dawson, A. G. (1992) Ice Age Earth: Late Quaternary geology and climate. Routledge, London
