Non-equilibrium Isotopic Fractionation between Seawater and Planktonic Foraminiferal Tests

N. J. Shackleton et al.

Editor’s Note

Harold Urey suggested in 1947 that, because the differential uptake of the two stable isotope of oxygen ( ¹⁶ O and ¹⁸ O) by marine microorganisms that form calcium carbonate exoskeletons depends on temperature, measuring this isotope ratio for these structures preserved in seafloor sediments could reveal the temperature of the seawater in which they formed, opening a window on past climate. Here Nicholas Shackleton at Cambridge and his coworkers cast doubt on whether this will work for foraminifera, because they don’t form their exoskeletons under chemical equilibrium. Several subsequent reports cast doubt on the use of oxygen isotope measurements to gauge palaeotemperature. But in the early 1980s it was shown that, with proper calibration, these did not preclude the technique after all. 中文

UREY ¹ has suggested that the temperature dependence of the isotopic fractionation factor between the oxygen in water and the oxygen in calcium carbonate could be used as a geological thermometer. For estimating Earth surface temperatures, biologically deposited calcium carbonate is more widely available than inorganic precipitates. So it was necessary to discover whether or not organisms deposit carbonate under equilibrium conditions. Epstein et al. ^2,3 investigated this using molluscs; although they found one case where a mollusc seemed to have deposited some carbonate in non-equilibrium conditions, they ascribed this to special circumstances. Apart from this one case, they inferred that the Mollusca deposit calcium carbonate in isotopic equilibrium with the surrounding water. 中文

Until recently it has been assumed that this is also true of the Foraminifera. Support for this assumption comes from the seemingly reasonable temperature values which Emiliani ^4,5 derived making this assumption. Indeed, whereas the first work on molluscs was performed with the intention of investigating the isotopic fractionation as a function of temperature, the first work on planktonic foraminifera ⁶ was an investigation of the depth (temperature) habitat of recent foraminifera, performed on the assumption that the isotopic fractionation factor was known. We now find that this was an unwarranted assumption, and that a substantial portion of the variation in isotopic composition between one species and another in foraminiferal death assemblages is due to different fractionation factors rather than to different life habitats. 中文

Oxygen Isotope Analyses

Duplessy et al. ⁷ have previously reported differential isotopic fractionation among benthonic foraminiferal species, and van Donk ⁸ obtained evidence suggesting that the planktonic species may also deposit calcite out of equilibrium with water. But although a fossil population such as that studied by Duplessy et al. enables a comparison to be made between benthonic species, it cannot be used for planktonic species because they do not derive from the same temperature habitat. For this reason, it is only possible to compare the departure from isotopic equilibrium among planktonic species if they are collected in plankton tows within the isothermal layer of the ocean. 中文

With the exception of a few samples analysed by van Donk all oxygen isotope analyses of planktonic foraminifera so far published have been performed using fossil or sub-fossil material. The reason is that it is in general difficult to collect large enough samples from plankton tows for conventional isotope analysis. In the present work we have been able to analyse samples as small as 0.04 mg carbonate, and so have been able to make use of a series of plankton tows made in the Indian Ocean by US Coastal and Geodetic Survey Ship Oceanographer in 1967. The stations from which we have used samples in this study are given below in Table 1. Some samples are shown in Fig. 1. 中文

中文

中文

Where possible, the samples for isotope analysis were divided and two independent analyses made. This provides a check on the reproducibility of the analyses, which we considered desirable because plankton have not previously been analysed at the Cambridge laboratory. Samples were roasted in vacuo for 30 min at 450℃ prior to analysis. Carbon dioxide for mass spectrometric analysis was released by the action of 100% orthophosphoric acid at 50℃. The oxygen isotopic composition of the gas was compared with that of an aliquot from a standard bulk gas sample, using the mass spectrometer described by Shackleton ⁹ , and the results calibrated by analysing standard carbonates under the same conditions. 中文

In this work Emiliani’s belemnite standard B1 has been used as a calibration standard, assuming its ¹⁸ O content to be +0.1‰ with respect to the PDB standard. Analyses listed in Table 2 are referred to the PDB standard on this basis. PDB is a standard based on a belemnite from the Carolina Peedee Formation ¹⁰ . 中文

Fourteen of these figures are the mean of two independent analyses. The standard error of a single analysis may be estimated from the difference between pairs as ±0.11‰.

中文

Not all the species in each plankton tow sample have the same isotopic composition. We have performed an analysis of variance to isolate between species, between station and residual variation. 中文

Between species variation proves to be highly significant ( P <0.001). The Students T test was used to test variation between species (Table 3). 中文

中文

The residual variation is 0.1‰, which may be compared with the standard deviation of ±0.11‰ derived from the pairs of analyses. The apparent difference between G. ruber and G . sacculifera is only 0.11‰, too small to be estimated reliably in the present experiment. 中文

Implication for Depth-Habitat Studies

Emiliani ⁶ , Lidz et al. ¹¹ and Hecht and Savin ¹² have used ¹⁸ O/ ¹⁶ O ratio determinations in foraminiferal tests as a method of investigating their life habitats. In the course of a study of a core in the Indian Ocean, Oba ¹³ did the same for one sample. It is clear from our study that this is not a valid approach; indeed, as regards the four species discussed here the between species differences found in plankton samples from 50 m horizontal tows are indistinguishable from the variations found by Oba in a sediment sample. This means that in all probability none of the variation measured by Oba can be ascribed to different depth habitat among the four species; they all derive from populations living in the isothermal layer. 中文

In the Atlantic province, larger differences between the species have been measured ^6,11,12 ; it seems likely that in that case there is a contribution which may be safely ascribed to difference in depth habitat. At the same time we need to know much more about the differential fractionation effect before we can begin to use isotopic determinations in order to deduce life habitat. 中文

Hecht and Savin ¹² have gone further, and attempt to use isotopic determinations to discover whether morphological variation within a population of foraminifera is due to varying ecological stress; they deduce, for example, that specimens of Globigerinoides ruber with a diminutive final chamber are those which have lived in deeper (colder) water, by comparing their oxygen isotopic composition with that of normal individuals. Although this seems to be an ingenious test, we cannot exclude the possibility that whatever factor influences the shape of the final chamber also influences the departure from isotopic equilibrium in the test carbonate. This is by no means impossible, particularly because the suggestion made by Parker ¹⁴ , that no-equilibrium isotopic composition could be causally related to the presence of symbiotic zooxanthellae, has never been excluded. 中文

Implications for Palaeotemperature Studies

Because isotope analysis was not envisaged when the samples were collected, we do not have information on the temperature and isotopic composition of the water in which the foraminifera were living. This means that we cannot estimate exactly the isotopic composition which would have been measured had the test carbonate been deposited in isotopic equilibrium; however, an approximate estimate may be made. Craig and Gordon ¹⁵ show that in the equatorial region the observed variation in surface isotopic composition with salinity is small, about 0.11‰ for 1‰ change in salinity. Using Defant’s ¹⁶ plate V, we may assume that at stations 6, 7 and 15 the salinity was in the region of 34.5‰, and that the isotopic composition of the water was near zero on the PDB scale (+0.2‰ on the SMOW scale ¹⁵ ). At stations 3, 4 and 5 the salinity is likely to have been a little higher and the isotopic composition about +0.1‰ on the PDB scale. As regards temperature, the mean August surface temperatures (Defant ¹⁶ , Plate 3B) are about 27℃ for stations 3, 4, 5, 6 and 7, and about 25℃ for station 15. Using these estimates and the relation between temperature and isotopic composition given by Craig ¹⁷ yields the values in Table 4. Comparison with the measured isotopic composition for each species from Table 2 gives the extent of departure from isotopic equilibrium for each species. The estimates range from –0.50‰ for G. ruber to +0.06‰ for P. obliquiloculata . 中文

中文

In column 4 of Table 4 the isotopic composition of carbonates deposited at the temperature given in column 2, and in water having the isotopic composition of column 3, is estimated on the basis of the equilibrium relationship of Craig ¹⁷ . The remaining columns are the differences between the measurements from Table 2 and the values in column 4, and thus represent deviations from isotopic equilibrium for the species concerned. 中文

For this last species, the deviation is insignificant; it may well be that this species does in fact deposit its test in isotopic equilibrium with the surrounding water. If one could extract confidently from the sediment only those tests of P. obliquiloculata which lived in the isothermal layer of the ocean, this might well be an ideal species to use for palaeotemperature determination; it is possible that these could be recognized by the lack of the smooth external cortex which is a characteristic feature in tests from deeper water. 中文

At the other end of the scale, G. ruber yields a value 0.5‰ lighter than it would if isotopic equilibrium prevailed. This is equivalent to an error of about 2.5℃. 中文

Before these values are used to correct estimates of palaeotemperature, it is important to establish whether the “vital effect” ¹ remains constant for each species. We have only measured the effect for foraminifera living at 50 m and in a restricted area of an ocean; it is of great interest to discover whether our results have general application. 中文

It is in any case not possible to use the data to correct measurements made in other laboratories because at present results from different laboratories do not seem to be related to each other. For example, specimens of Globigerinoides sacculifera from the top of core P6304-8 gave –1.29‰ when analysed by Emiliani ⁶ ; samples from the same level in the same core gave –1.69‰ when analysed by Lidz et al. ¹¹ and the same species from nearby core V12-122 yielded –2.2‰ (ref. 18). When corrected for the isotopic composition of the Caribbean water (+0.92‰) these figures yield temperatures from near 27℃ to near 32℃. At present these differences seem more serious than the fundamental problem posed by the present work. 中文

We conclude that whether the isotopic composition of a foraminiferal test is analysed with a view to determining its present-day life habitat, or with a view to elucidating the mysteries of climatic change in the past, it is not possible to translate the value obtained into an equivalent temperature on the basis of thermodynamic principles alone. In some manner yet to be determined the organism deposits carbonate of isotopic composition differing slightly from the thermodynamically predicted value. 中文

This work was supported by an NERC grant. We thank M. A. Hall for operation of the mass spectrometer, and the Oceanographic Sorting Center, Smithsonian Institution, for plankton aliquots. 中文

( 242 , 177-179, 1973)

N. J. Shackleton*, J. D. H. Wiseman† and H. A. Buckley†

* Sub-Department of Quaternary Research, University of Cambridge;

^† British Museum (Natural History), London

Received December 4, 1972.

---

References: G1LZeVOvUstvQKqxpCdEOVpiVGB1i0l/1DQW00L6PA36QsPb8BCTD5rKWYL5fUzq

1. Urey, J. C., J. Chem. Soc. , 562 (1947).
2. Epstein, S., Buchsbaum, R., Lowenstam, H. A., and Urey, H. C., Geol. Soc. Amer. Bull. , 62 , 417 (1951).
3. Epstein, S., Buchsbaum, R., Lowenstam, H. A., and Urey, H. C., Geol. Soc. Amer. Bull. , 64 , 1315 (1953).
4. Emiliani, C., J. Geol. , 63 , 538 (1955).
5. Emiliani, C., J. Geol. , 74 , 109 (1966).
6. Emiliani, C., Amer. J. Sci. , 252 , 149 (1954).
7. Duplessy, J. C., Lalou, C., and Vinot, A. C., Science , 168 , 250 (1970).
8. Van Donk, J., thesis, Columbia Univ. (1970).
9. Shackleton, N. J., J. Sci. Instrum. , 42 , 689 (1965).
10. Urey, H. C., Lowenstam, H. A., Epstein, S., and McKinney, C. R., Bull. Geol. Soc. Amer. , 62 , 399 (1951).
11. Lidz, B., Kehm, A., and Miller, H., Nature , 217 , 245 (1968).
12. Hecht, A. D., and Savin, S. M., Science , 170 , 69 (1970).
13. Oba, T., Sci. Rep. Tohoku Univ., S econd Ser. ( Geology ), 41 , 129 (1969).
14. Parker, F. L., Rep. Swedish Deep-Sea Exped. , 8 , 2, 219 (1958).
15. Craig, H., and Gordon, L. I., in Stable Isotopes in Oceanographic Studies and Palaeotemperatures (edit. by Tongiorgi, E.) (Pisa, 1965).
16. Defant, A., Physical Oceanography (Pergamon, Oxford, 1961).
17. Craig, H., in Stable Isotopes in Oceanographic Studies and Palaeotemperatures (edit. by Tongiorgi, E.) (Pisa, 1965).
18. Broecker, W. S., and Van Donk, J., Rev. Geophys. Space Phys. , 8 , 169 (1970).