Radioisotopic Methods

spontaneous radioactive decay by the loss of nuclear particles (a or 3 particles) and, as a result, they may transmute into a new element.5 For example, 14C decays to 14N, and 40K decays to 40Ar and 40Ca. Furthermore, the decay rate is invariant so that a given quantity of the radioactive isotope will decay to its daughter product in a known interval of time; this is the basis of radioisotopic dating methods. Providing that the radioisotope "clock" is started close to the stratigraphically relevant date, measurement of the isotope concentration today will indicate the amount of time that has elapsed since the sample was emplaced. The amount of time it takes for a radioactive material to decay to half its original amount is termed its half-life. Table 3.1 lists the half-lives of some radioisotopes that have been used in the context of dating. In the case of radiocarbon (14C) the half-life is 5730 ± 30 yr. Thus a plant that died 5730 yr ago has only half its original 14C content remaining in it today.6 After a further 5730 yr from today it will have only half as much again, that is, 25% of its original 14C content, and so on (Fig. 3.2).

For a radioactive isotope to be directly useful for dating it must possess several attributes: (a) the isotope itself, or its daughter products, must occur in measurable quantities and be capable of being distinguished from other isotopes, or its rate of decay must be measurable; (b) its half-life must be of a length appropriate to the period being dated; (c) the initial concentration level of the isotope must be known; and (d) there must be some connection between the event being dated and the start

TABLE 3.1 Half-lives of radioisotopes used in dating — although the half-life of l4C is calculated to be 5730 ± 40 yr, by convention the "Libby half-life" of 5568 ± 30 yr is used6

14C

5.73 x 103 years

238U

4.51 X 109 years

235y

0.71 X 109 years

40K

1.31 X 109 years

5 An a particle is made up of two protons and two neutrons (i.e., a helium nucleus) and a |3 particle is an electron. Neutrons may decay to produce a (3 particle and a proton, thereby causing a transmutation of the element itself.

6 When Libby (1955) expounded the principles of radiocarbon dating he calculated a half-life for 14C of 5568 yr. This was the average of a number of estimates up to that time, and was adopted by all radiocarbon dating laboratories. By the early 1960s, further work had demonstrated that the original estimate was in error by 3%, and the half-life was closer to 5730 yr (Godwin, 1962). To avoid confusion, it was decided to continue using the "Libby half-life" (rounded to 5570 yr) and this practice has continued. For practical purposes it is not a significant problem as all dates are now reported in the journal Radiocarbon using the Libby half-life. However, when comparing "radiocarbon years" with calendar years (historical, archeological, and/or astronomical events) and dates obtained by other techniques, adjustments are necessary (see Section 3.2.1.5).

T,„ 2T,„ 3T,„ 4T„i 5T1(! TIME 5570 11140 16710 22280 27850 yr 5730 11460 17190 22920 28650 yr

FIGURE 3.2 The decay of radiocarbon with time. Both the Libby and revised half-lives (Tl/2) are shown on the abscissa.The ordinate scale is in disintegrations per minute per gram of carbon (Olsson, 1968).

T,„ 2T,„ 3T,„ 4T„i 5T1(! TIME 5570 11140 16710 22280 27850 yr 5730 11460 17190 22920 28650 yr

FIGURE 3.2 The decay of radiocarbon with time. Both the Libby and revised half-lives (Tl/2) are shown on the abscissa.The ordinate scale is in disintegrations per minute per gram of carbon (Olsson, 1968).

of the radioactive decay process (the "clock"). The relevance of these factors will be made clear in the ensuing sections.

In general terms, radioisotopic dating methods can be considered in three groups (see Fig. 3.1): those that measure (a) the quantity of a radioisotope as a fraction of a presumed initial level (e.g., 14C dating) or the reciprocal build-up of a stable daughter product (e.g., potassium-argon, and argon-argon dating); (b) the degree to which members of a chain of radioactive decay are restored to equilibrium following some initial external perturbation (uranium-series dating); and (c) the integrated effect of some local radioactive process on the sample materials, compared to the value of the local (environmental) flux (fission-track and luminescence dating). Each of these methods will be considered separately.

3.2.1 Radiocarbon Dating

For studies of late Quaternary climatic fluctuations, 14C or radiocarbon dating has proved to be by far the most useful. Because of the ubiquitous distribution of 14C, the technique can be used throughout the world and has been used to date samples of peat, wood, bone, shell, paleosols, "old" sea water, marine and lacustrine sediments, and atmospheric C02 trapped in glacier ice. Furthermore, the useful timeframe for radiocarbon dating spans a period of major, global environmental change that would be virtually impossible to decipher in any detail without accurate dating control. Radiocarbon dating is also ideal for dating man's development from paleolithic time to the recent historical past and it has therefore proved invaluable in archeological studies. In addition, variations in the 14C content of the atmosphere are of interest in themselves because of the implications these have for solar and/or geomagnetic variations through time and hence for climatic fluctuations.

3.2.1.1 Principles of ,4C Dating

Radiocarbon (14C) is produced in the upper atmosphere by neutron bombardment of atmospheric nitrogen atoms:

The neutrons have a maximum concentration at around 15 km and are produced by cosmic radiation entering the upper atmosphere. Although cosmic rays are influenced by the Earth's magnetic field and tend to become concentrated near the geomagnetic poles (thus causing a similar distribution of neutrons and hence 14C), rapid diffusion of 14C atoms in the lower atmosphere obliterates any influence of this geographical variation in production. The 14C atoms are rapidly oxidized to 14C02, which diffuses downwards and mixes with the rest of atmospheric carbon dioxide and hence enters into all pathways of the biosphere (Fig. 3.3). As Libby (1955) stated, "Since plants live off the carbon dioxide, all plants will be radioactive; since the animals on earth live off the plants, all animals will be radioactive. Thus ... all living things will be rendered radioactive by the cosmic radiation."

During the course of geological time, an equilibrium has been achieved between the rate of new 14C production in the upper atmosphere and the rate of decay of 14C in the global carbon reservoir. This means that the 7.5 kg of new 14C estimated to be produced each year in the upper atmosphere is approximately equal to the weight of 14C lost throughout the world by the radioactive decay of 14C to nitrogen, with the release of a (3 particle (an electron):

The total weight of global 14C thus remains constant.7 This assumption of an essentially steady concentration of radiocarbon during the period useful for dating is fundamental to the method though, in detail, this assumption is invalid (see Section 3.2.1.5).

Plants and animals assimilate a certain amount of 14C into their tissues through photosynthesis and respiration; the 14C content of these tissues is in equilibrium with that of the atmosphere because there is a constant exchange of new 14C as old

7 Prior to atomic bomb explosions in the atmosphere the equilibrium quantity of 14C was estimated to be -62 metric tons. Since the 1950s, the amount of artificially produced 14C has increased by perhaps 3-4%, though most of this has, as yet, remained in the atmosphere; consequently 14C levels there have almost doubled (Aitken, 1974).

COSMIC RAYS I

COSMIC RAYS I

cells die and are replaced. However, as soon as an organism dies this exchange and replacement of 14C ceases. From that moment on the 14C content of the organism declines as the 14C decays to nitrogen, and the 14C content is henceforth purely a function of time; the radioactive "clock" has been activated. Because the 14C content declines at a negative exponential rate (see Fig. 3.2) by the time that ten half-lives have elapsed (57,300 yr) the sample contains less than 0.01% of the original 14C content of the organism when it was alive. To put this in a more practical way, in a 1 g sample of carbon with a 14C content equivalent to modern levels, decay of the radiocarbon atoms in the sample will produce about 15 particles per minute, a rate that is relatively easy to count. By contrast, 57,300 yr after an organism has died, 1 g of its carbon will produce only about 21 (3 particles per day (Aitken, 1974). It is this ever-decreasing quantity of 14C with increasing sample age that makes conventional radiocarbon dating so difficult; it simply becomes impossible to separate disintegrations of the sample from the extraneous background radiation (typically ~132 (3 particles per day in a modern counter) as the signal-to-noise ratio (S/N) becomes too small.

3.2.1.2 Measurement Procedures, Materials, and Problems

Until the early 1980s, nearly all radiocarbon-dating laboratories used so-called "conventional" methods — either proportional gas counters or liquid scintillation techniques. In the former method, carbon is converted into a gas (methane, carbon dioxide, or acetylene), that is then put into a "proportional counter" capable of detecting p particles (variations in output voltage pulses being proportional to the rate of p-particle emission). In liquid scintillation procedures, the carbon is converted into benzene or some other organic liquid and placed in an instrument that detects scintillations (flashes of light) produced by the interaction of p particles and a phosphor added to the organic liquid. In both methods, stringent measures are necessary to shield the sample counters from extraneous radioactivity in the instrument components, laboratory materials, and surrounding environment, including cosmic rays penetrating the Earth's atmosphere from outer space. Indeed, the difficulty of separating the sample (3-particle signal from environmental "noise" was one of the major obstacles to the development of 14C dating, particularly of older samples, which have very low levels of 14C anyway (Libby, 1970). Lead shielding, electronic anticoincidence counters (to alert the counter to particles entering the counting chamber from outside) and construction of laboratories beneath the ground are common strategies to help keep background radiation levels as low as possible.

One of the problems of dating very old samples by conventional methods is the large sample size needed to obtain enough radiocarbon for its 3 activity to be counted. Technical difficulties place an upper limit on the volume of gas or liquid that can be accurately analyzed; hence, for very old samples, some means of concentrating the 14C is needed to reduce the volume. One solution (no longer in common usage) is to concentrate a gas containing the 14C (e.g., 14C02) by thermal diffusion, "enriching" the sample and reducing the required volume. Effectively, the gas containing the heavier isotope is encouraged to collect in the lower chamber of a thermal diffusion column; in this way, the radioactive component is concentrated, reducing the total volume of gas necessary for accurate counting. However, the procedure is very time consuming; a 6-fold enrichment may take up to 5 weeks! Nevertheless, this procedure has enabled samples as old as 75,000 yr (13 half-lives) to be dated (Stuiver et al., 1978; Grootes et al., 1980). The main limitation is that the initial sample must be large enough to yield 100 g of carbon for analysis, and there must be very low background 14C activity to minimize any statistical uncertainty in the calculated age. Furthermore, if an infinitely old sample has even 0.1% contamination by modern carbon, it will yield a finite date of ~55,500 yr B.P., illustrating the inherent dangers of interpreting extremely old 14C dates. In practice, 14C dates of >45,000 yr B.P. should be viewed with caution.

Radiocarbon dating underwent a technological revolution in the late 1970s and early 1980s when a method for dating very small organic samples was developed, using an accelerator coupled to a mass spectrometer (AMS dating) (Muller, 1977; Nelson et al., 1977; Litherland and Beukens, 1995). Instead of measuring the quantity of 14C in a sample indirectly, by counting p-particle emissions, the concentrations of individual ions (12C, 13C, 14C) are measured. Ions are accelerated in a tandem electrostatic accelerator to extremely high velocities; they then pass through a magnetic field that separates the different ions, enabling them to be distinguished (Stuiver, 1978a; Elmore and Phillips, 1987). Sample sizes used in this technique are much smaller than in conventional 14C dating (only 1 mg of carbon is required) so that dates on small samples of foraminifera, pollen grains isolated from their surrounding matrix, or even individual seeds can be dated (Brown et ai, 1992; Reg-nail, 1992). There are now many accelerators designed specifically for 14C dating applications and several labs can produce results within days of receiving a sample. This rapid turn-around time allows field workers to quickly reassess sampling strategies, thus making the most of time in the field. On the other hand, it may not be too long before a field-portable radiocarbon analyzer is available, at least for first-order estimates of sample age (Robertson and Griin, 1994).

3.2.1.3 Accuracy of Radiocarbon Dates

It is tempting to accept a radiocarbon date as the gospel truth, particularly if it confirms a preconceived notion of what the sample age should be! Radiocarbon dates are, however, statements of probability (as are all radiometric measurements). Radioactive disintegration varies randomly about a mean value; it is not possible to predict when a particular 14C atom will decay, but for a sample containing 10101012 atoms of 14C a certain number of disintegrations will occur, on average, in a certain length of time. This statistical uncertainty in the sample radioactivity (together with similar uncertainty in the radioactive decay of calibration samples and "noise" due to background radiation) is inherent in all 14C dates. A single "absolute" (i.e., numerical) age can therefore never be assigned to a sample. Rather, dates are reported as the midpoint of a Poisson probability distribution; together with its standard deviation, the date thus defines a known level of probability. A date of 5000 ± 100 yr B.P., for example, indicates a 68% probability that the true (radiocarbon) age is between 4900 and 5100 yr B.P., a 95% probability that it lies between 4800 and 5200 yr B.P., and a 99% probability of it being between 4700 and 5300 yr B.P. In conventional dating, the use of large samples, extended periods of counting, and the reduction of laboratory background noise will all improve the precision of the age determination. However, even rigorous 14C analysis cannot account for all sources of error and these must be evaluated before putting a great deal of confidence in the date obtained. A sample age may be precisely determined (analytically) but it may not be an accurate reflection of the true age if the sample is contaminated, or if appropriate corrections are not made, as discussed in the following sections.

3.2.1.4 Sources of Error in l4C Dating

(a) Problems of Sample Selection and Contamination

It is self-evident that a contaminated sample will give an erroneous date, but it is frequently very difficult to ascertain the extent to which a sample has been contaminated. Some forms of contamination are relatively straightforward: modern rootlets, for example, may penetrate deep into a peat section and without careful inspection of a sample and removal of such material, gross errors may occur. More abstruse problems arise when dating materials that contain carbonates (e.g., shell, coral, bone). These materials are particularly susceptible to contamination by modern carbon because they readily participate in chemical reactions with rainwater and/or groundwater. Most molluscs, for example, are primarily composed of calcium carbonate in the metastable crystal form, aragonite. This aragonite may dissolve and be redeposited in the stable crystal form of calcite. During the process of solution and recrystallization, exchange of modern carbon takes place and the sample is thereby contaminated (Grant-Taylor, 1972). This problem also exists in corals, which are all aragonite. Commonly, x-ray diffraction is used to identify different carbonate mineral species, and materials with a high degree of recrystallization are discarded. However, Chappell and Polach (1972) have noted that recrystallization can occur in two different modes: one open system and therefore susceptible to modern 14C contamination; and one closed system, which is internal and involves no contamination. The former process tends to be concentrated around the sample margins, as one might expect, and so a common strategy is to dissolve away the surface 10-20% of the sample with hydrochloric acid and to date the remaining material. For shells thought to be very old, in which recrystallization may have affected a deep layer, the remaining inner fraction should ideally be dated in two fractions (an "outer inner" and an "inner inner" fraction) to test for consistency of results. However, even repeated leaching with hydrochloric acid may not produce a reliable result in cases where recrystallization has permeated the entire sample. It is worth noting in this connection that "infinite"-aged shells (those well beyond the range of 14C dating techniques) contaminated by only 1% of modern carbon will have an apparent age of 37,000 yr (Olsson, 1974). Thus, even very small amounts of modern carbon can lead to gross errors, and many investigators consider that dates of >25,000 yr on shells should be thought of as essentially "infinite" in age. While such a conservative approach is often laudable, it does pose the danger that correct dates in the 20,000-30,000 year range may be overlooked. A possible remedy to this problem of dating old shells is to isolate a protein, conchiolin, present in very small quantities in shells (1-2%) and to date this, rather than the carbonate (Berger et al., 1964). Carbon in conchiolin does not undergo exchange with carbon in the surrounding environment and hence is far less likely to be contaminated. Unfortunately, to isolate enough conchiolin, very large samples (>2 kg) are needed, and these are often not available.

Similar problems are encountered in dating bone; because of exchange reactions with modern carbon, dates on total inorganic carbon or on apatite carbonate are unreliable (Olsson et al., 1974). As in the case of shells, a more reliable approach is to isolate and date carbon in the protein collagen. However, this slowly disappears under the influence of an enzyme, collagenase, so that in very old samples the amount of collagen available is extremely small and extra-large samples are needed to extract it. Furthermore, collagen is extremely difficult to extract without contamination and different extraction methods may give rise to different dates.

Another form of error concerns the "apparent age" or "hard-water effect" (Shotton, 1972). This problem arises when the materials to be dated, such as freshwater molluscs or aquatic plants, take up carbon from water containing bicarbonate derived from old, inert sources. This is a particularly difficult problem in areas where limestone and other calcareous rocks occur. In such regions, surface and ground water may have much lower 14C/12C ratios than that of the atmosphere due to solution of the essentially 14C-free bedrock. Because plants and animals existing in these environments will assimilate carbon in equilibrium with their surrounding milieu rather than the atmosphere, they will appear older than they are in reality, sometimes by as much as several thousand years. This problem was well illustrated by Shotton (1972), who studied a late-glacial stratigraphie section in North Jutland, Denmark. Dates on contemporaneous twigs and a fine-grained vegetable residue, thought to be primarily algal, fell neatly into two groups, with the algal material consistently 1700 yr older than the terrestrial material. The difference was considered to be the result of a hard-water effect, the aquatic plants assimilating carbon in equilibrium with water containing bicarbonate from old, inert sources.

Other studies have demonstrated further complexities in that the degree of "old carbon" contamination may change over time. For example, Karrow and Anderson (1975) suggested that some lake sediments in southwestern New Brunswick, Canada, studied by Mott (1975) were contaminated with old carbon shortly after déglaciation. Initial sedimentation was mainly marl derived from carbonate-rich till and carbonate bedrock, but as the area became vegetated and soil development took place the sediments became more organic and less contaminated by "old carbon." Dates on the deepest lake sediments are thus anomalously old and would appear to give a date of déglaciation inconsistent with the regional stratigraphy. This points to the more general observation that the geochemical balance of lakes may have changed through time and that the modern water chemistry may not reflect former conditions. This is particularly likely in formerly glaciated areas where the local environment immediately following déglaciation would have been quite different from that of today. One should thus interpret basal dates on lake sediments or peat bogs with caution. Equally, dates on aquatic flora and fauna, in closed lake basins which have undergone great size changes, must also be viewed in the light of possible changes in the aqueous geochemistry of the site.

It is important to recognize that not all types of contamination are equally significant; contamination by modern carbon is far more important than that by old carbon because of its much higher activity (Olsson, 1974). Figures 3.4 and 3.5 show the errors associated with different percentage levels of contamination by modern and old material, respectively. It will be seen that a 5000-yr-old sample, 20% contaminated with 16,000-yr-old carbon, would give a date in error by only -1300 years. By contrast, a 15,000-yr-old sample contaminated with only 3% of modern carbon would result in a dating error of about the same magnitude. Very careful sample selection is therefore needed; dating errors are most commonly the result of inadequate sampling.

(b) Variations in l4C Content of the Oceanic Reservoir

In the preceding section, it was discussed that some freshwater aquatic plants or molluscs may be contaminated by water containing low levels of 14C. In the case of marine organisms this problem is much more universal. First, when carbon dioxide is absorbed into the oceans a fractionation takes place that leads to an enrichment of 15%o (equivalent to -120 yr) in the 14C activity of oceanic bicarbonate

40000

30000

et 20000

10000

0 10000 20000 30000 40000 50000 °o

YEARS

AGE OF SAMPLE

FIGURE 3.4 Apparent ages as a function of sample age if there are different levels of contamination by modern material in the sample.Thus a 20,000 yr-old sample contaminated with 10% of modern material would appear to be ~ 15,000 yr old (from Olsson and Eriksson, 1972).

relative to that of the atmosphere. However, ocean surface waters are not in isotopic equilibrium with the atmosphere because oceanic circulation brings 14C-depleted water to the surface to mix with "modern" water. Consequently, the 14C age of the surface water (the apparent age, or reservoir age) varies geographically (Fig. 3.6) (Bard, 1988). In the lower latitudes of all oceans, the mean reservoir age of surface waters is -400 yr, which means that 400 yr must be added to a 14C date on marine organic material from the mixed layer, in order to compare it with terrestrial material. At higher latitudes, this correction can be much larger due to up-welling of older water and the effect of sea ice, which limits the ocean-atmosphere exchange of C02. The modern North Atlantic is different from other high latitude regions because of advection of warmer waters (relatively enriched in 14C) from lower latitudes, and strong convection (deepwater formation), which limits any up-welling of 14C-depleted water.

The extent to which such 14C gradients have been constant over time is of great significance for dating older events in the marine environment and comparing them with terrestrial records. If North Atlantic deepwater formation ceased during the Last Glacial (see Section 6.10) then it is likely that the surface water reservoir age would

40000

30000

et 20000

10000

0 10000 20000 30000 40000 50000 °o

YEARS

AGE OF SAMPLE

FIGURE 3.4 Apparent ages as a function of sample age if there are different levels of contamination by modern material in the sample.Thus a 20,000 yr-old sample contaminated with 10% of modern material would appear to be ~ 15,000 yr old (from Olsson and Eriksson, 1972).

ERROR 1000 (years)

ERROR 1000 (years)

5 000 10 000 15 000 20 000 25 000

AGE DIFFERENCE (years)

FIGURE 3.5 The error in a radiocarbon date if a certain fraction of the sample (indicated by each curve) is contaminated by older material (having a lower l4C activity). Errors expressed as age differences (abscissa) between the sample and contaminant. For example, a 5000-yr old sample with 20% contamination by 16,000 yr-old material (i.e., material 11,000 yr older than the true age— see star on figure) will yield a date in error (too old) by 1300 yr (from Olsson, 1974).

5 000 10 000 15 000 20 000 25 000

AGE DIFFERENCE (years)

FIGURE 3.5 The error in a radiocarbon date if a certain fraction of the sample (indicated by each curve) is contaminated by older material (having a lower l4C activity). Errors expressed as age differences (abscissa) between the sample and contaminant. For example, a 5000-yr old sample with 20% contamination by 16,000 yr-old material (i.e., material 11,000 yr older than the true age— see star on figure) will yield a date in error (too old) by 1300 yr (from Olsson, 1974).

be similar to other high latitude areas, as depicted by the dashed line in Fig. 3.6. Hence, late Glacial planktonic samples from this area might require an age adjustment of as much as 1000 yr, and if deepwater switched on and off rapidly, age corrections would be equally volatile, possibly even leading to age inversions in otherwise strati-graphically undisturbed sediments. The matter was addressed by Bard et al. (1994) and Austin et al. (1995), who compared marine and terrestrial 14C-dated samples associated with the Vedde volcanic ash. This ash layer originated from an explosive Icelandic eruption and was widely distributed across the North Atlantic and adjacent land areas around 10,300 14C yr B.P., according to AMS dates on terrestrial samples from the time of ash deposition. However, all marine samples associated with the ash layer were dated -11,000 14C yr B.P., indicating that the reservoir effect then was -700 yr, vs -400 yr today, probably as a result of reduced North Atlantic Deep Water (NADW) formation, and/or increased sea-ice cover at that time.

So far, we have focused primarily on surface water, but the same issues arise in deepwater changes. North Atlantic Deep Water and the Antarctic Bottom Water may remain out of contact with the atmosphere for centuries because of the overlying warmer water in mid and low latitudes. During this time, the 14C content of the deep water decreases so that deepwater samples commonly give 14C ages more than 1000 yr older than surface water (Fig. 3.7). Indeed, the gradual decline in 14C activity of oceanic water has been used to assess the former "ventilation rate" of the ocean, by comparing the age of 14C-dated planktonic and benthic foraminifera

LATITUDE

FIGURE 3.6 Apparent age of the surface waters of the major oceans, averaged by latitude, based on measurements of corals (open triangles), molluscs (closed circles), and sea-water SC02 samples (open squares); closed squares are reconstructions of pre-bomb AMC levels (Broecker et al., 1985a).The solid lines are least squares polynomials fitted to the data; the dashed lines represent a hypothetical profile for the Last Glacial Maximum (LGM) assuming there was no North Atlantic Deep Water (NADW) forming at that time (from Bard, 1988).

LATITUDE

FIGURE 3.6 Apparent age of the surface waters of the major oceans, averaged by latitude, based on measurements of corals (open triangles), molluscs (closed circles), and sea-water SC02 samples (open squares); closed squares are reconstructions of pre-bomb AMC levels (Broecker et al., 1985a).The solid lines are least squares polynomials fitted to the data; the dashed lines represent a hypothetical profile for the Last Glacial Maximum (LGM) assuming there was no North Atlantic Deep Water (NADW) forming at that time (from Bard, 1988).

FIGURE 3.7 Present-day radiocarbon age differences between surface waters and waters at 3 km depth (from Broecker et o/., 1988a).

(surface and deep-dwelling organisms, respectively) in marine sediments from the same site. Then, by comparing modern (core-top) samples with those from earlier periods, a record of changes in ventilation rate and ocean circulation can be established (see Section 6.9). In this way, Duplessy et al. (1989) found that the ventilation rate of the Pacific Ocean was less than today during the Last Glacial (resulting in a 1500-2500 yr age difference, vs -1300 yr today) but during the deglaciation (15,000-10,000 14C yr B.P.) it was greater (200-1000 yr age difference) (Shackleton et al., 1988). In the Atlantic Ocean, deepwater had a mean age of ~675 yr in the Last Glacial versus -350 yr today (Peng and Broecker, 1995).

Because there are still uncertainties about the timing and extent of past changes in reservoir effects in both surface and deep waters, many investigators only make corrections for the observed modern oceanic reservoir effect, arguing that so far there is an insufficient basis of knowledge to do anything else. Nevertheless, this can lead to difficulties in trying to determine the sequence of terrestrial and marine events, particularly at times of rapid environmental change, such as occurred at the end of the Last Glacial and around the time of the Younger Dryas episode (Austin et al., 1995). The problem is further compounded by the "plateau" found in 14C ages around 10,000 14C yr B.P., which tends to make events that were in fact di-achronous appear to be synchronous (see Section 3.2.1.5).

(c) Fractionation Effects

Basic to the principle of 14C dating is the assumption that plants assimilate radiocarbon and other carbon isotopes in the same proportion as they exist in the atmosphere (i.e., the 14C/12C ratio of plant tissue is the same as that in the atmosphere). However, during photosynthesis, when C02 is converted to carbohydrates in plant cells, an isotopic fractionation occurs such that 12C is more readily "fixed" than 14C, resulting in a lower 14C content in plants than that of the atmosphere (Olsson, 1974). This 14C "depletion" may be as much as 5% below atmospheric levels but this is not consistent among all organisms. The magnitude of the fractionation effect varies from one plant species to another by a factor of two to three and depends on the particular biochemical pathways evolved by the plant for photosynthesis (Lerman, 1972). This is discussed in further detail in Appendix A. Fortunately, some assessment of the 14C fractionation effect can be made relatively easily by measuring the 13C content of a sample. The 14C fractionation is very close to twice that of 13C (Craig, 1953), a stable isotope which occurs in far greater quantities than 14C and can hence be routinely measured by a mass spectrometer. The 13C content is generally expressed as a departure from a Cretaceous limestone standard (Peedee belemnite, see Appendix A):

+1 0

Post a comment