Absolute Dating of Deep-Sea Cores by the Pa231/Th230Method

Abstract

Oxygen isotopic analysis of Globigerina-ooze cores from the Atlantic and adjacent seas showed that surface ocean temperatures underwent numerous, apparently periodical, variations during the past few hundred thousand years. C14 dating showed that the last temperature minimum of the deep-sea cores was synchronous with the last major glaciation, the Main Würm. Previous attempts to date deep-sea cores were based on the decay of uranium-unsupported Th230 (ionium). This method requires, among other conditions, that the supply of uranium-supported Th230 in sea water and the rate of non-carbonate sedimentation remained essentially constant over the time interval to be dated. Attempts to correct for possible variations in the non-carbonate sedimentation rate have been made by using such ratios as Th230/Th232 or Th230/Fe2O3. The validity of these corrections is questionable because Th230 produced in sea water by the decay of U238 and U234 has a geochemical history different from that of Th232 and Fe2O3. The requirements mentioned above need not be met if the ratio $$Pa^{231}/Th^{230}$$ is used. Since Pa231 and Th230 are daughters of the same element, uranium, and since they decay at different rates, their ratio is a function of time alone. While information from deep-sea cores, bearing directly on Pleistocene history, has been obtained almost exclusively by isotopic and micropaleontological analysis of the foraminiferal component of Globigerina-ooze cores, dating by the decay of uranium-unsupported Th230 or by the ratio $$Pa^{231}/Th^{230}$$ is based on the clay component where these nuclides are concentrated. Therefore, dating, by these two methods, of the stratigraphic record given by the foraminiferal component requires synchronism between the two components. Such synchronism may be exceptional, for the clay component may frequently or even generally contain some or much reworked material, even when the foraminiferal record is undisturbed. In such cases, the ages obtained may be generally greater than the ages of the events to be dated. $$Pa^{231}/Th^{230}$$ dating of two deep-sea cores from the Caribbean, about 600 km. apart, has given a set of dates which are internally consistent; identical, within the limits of error, in stratigraphically equivalent levels of the two cores; and coincident with the C14 chronology. This set of dates is believed to provide a reliable, absolute time scale, extending from the present to about 175,000 years ago. $$Pa^{231}/Th^{230}$$ and C14 measurements on deep-sea cores, C14 measurements on continental material, paleotemperature analysis of deep-sea cores, and correlation of the temperature record of the deep-sea cores with continental events provide the following ages for Pleistocene stages: postglacial, 0-10,000 years; Late and Main Wurm, 10,000-30,000 years; Main Würm-Early Würm interval, 30,000-50,000 years; Early Würm, 50,000-65,000 years; Riss/Würm interglacial, 65,000-100,000 years; Riss, 100,000-130,000 years; and Mindel/Riss interglacial, 130,000-175,000 years. These ages are very close to or identical with the ages given by Emiliani (1955a, 1958). Correlation between temperature variations of the deep-sea cores and continental stages preceding the last interglacial, however, is only tentative. The apparent identity of the C14 and $$Pa^{231}/Th^{230}$$ chronologies over the entire range of the C14 method indicates that the cosmic-ray flux did not change by more than a factor of 2 during the past 60,000 years. $$Pa^{231}/Th^{230}$$ dating of a deep-sea core from the North Atlantic gave ages which are consistently about 30,000 years greater than the $$Pa^{231}/Th^{230}$$ ages obtained from the two Caribbean cores and the C14 chronology. This is believed to result from contamination by reworked clay, an effect which may actually exist in most deep-sea cores. Rates of sedimentation of the carbonate fraction larger than 62 μ, the carbonate fraction smaller than 62 μ, and the non-carbonate fraction, calculated for the intervals between selected dated levels, appear not to have changed markedly when averaged over time intervals of some tens of thousands of years. The rates of sedimentation during the last 11,000 years, however, were lower than during previous time intervals. A generalized temperature curve, calibrated in terms of the C14-$$Pa^{231}/Th^{230}$$ chronology, is presented. This curve is very close to the curve previously constructed by Emiliani (1955a, 1958).