Info

cal. ka BC

FIGURE 3.10 The atmospheric AI4C record of the last 30,000 yr derived from both dendrochronologically based wood samples and corals dated by 230Th/234U, with a 400-yr correction for the reservoir age of tropical surface waters. The coral data are shown as circles with 2ct error bars (from Stuiver and Braziunas, 1993).

10,500

10,000

O 9,000

0.500

8,500

10.000

10.500

11,000

11,500

12,000

Calendar age (yr B.R)

FIGURE 3.1 I The very detailed radiocarbon chronology of tree rings from German oaks (squares) and pines (open circles) has been used to place the "floating" chronology of annually laminated (varved) sediments from the Cariaco Basin, north ofVenezuela (solid circles) into a firm chronological framework (r = 0.99). By finding the optimum fit between both records, through matching the "wiggles" in the tree-ring record with those of the varve record, the varve sequence (which extends 3000 years further back in time) has been fixed in time. As the varves can be counted, giving calendar year ages, and they also have been AMS dated (using forams in the sediments), they can be used to further extend the calibration of the radiocarbon timescale.Tim-ing of the Younger Dryas— Pre-boreal transition (recognized in the varves) is shown by the vertical shaded line (from Hughen etai, 1998).

8,500

10.000

10.500

11,000

11,500

10,500

10,000

O 9,000

0.500

8,000 9,000

12,000

Calendar age (yr B.R)

FIGURE 3.1 I The very detailed radiocarbon chronology of tree rings from German oaks (squares) and pines (open circles) has been used to place the "floating" chronology of annually laminated (varved) sediments from the Cariaco Basin, north ofVenezuela (solid circles) into a firm chronological framework (r = 0.99). By finding the optimum fit between both records, through matching the "wiggles" in the tree-ring record with those of the varve record, the varve sequence (which extends 3000 years further back in time) has been fixed in time. As the varves can be counted, giving calendar year ages, and they also have been AMS dated (using forams in the sediments), they can be used to further extend the calibration of the radiocarbon timescale.Tim-ing of the Younger Dryas— Pre-boreal transition (recognized in the varves) is shown by the vertical shaded line (from Hughen etai, 1998).

13.000

12.000-

9.000 8.000

9,000 10.000 11.000 12,000 13,000 14,000 15,000 Calendar age (yr B.P)

FIGURE 3.12 Calendar age vs radiocarbon age for Cariaco Basin varved sediments (solid circles) compared to uranium-series (TIMS) and radiocarbon ages of corals from the Atlantic and Pacific Oceans (open symbols) (coral data from Bard et al., 1993, 1996; Edwards et al., 1993).The thin diagonal line shows the expected relationship if no changes in atmospheric MC occurred.The vertical shaded lines show the time (and duration of each transition) at the start and end of the Younger Dryas episode, as recorded in the varves (from Hughen etai, 1998).

levels (Edwards et al., 1993). Indeed, models of past variations in the geomagnetic field reproduce remarkably well the observed 14C variations manifested in the Cariaco varve sequence. This is discussed further in the next section.

A very significant feature of the radiocarbon-calendar year calibration is the presence of prolonged 14C age plateaus centered at -11,700, -11,400 and -9,600 radiocarbon yr B.P. (and to a lesser extent at -8750 and -8250 14C yr B.P.) (Becker et al., 1991; Lotter, 1991; Kromer and Becker, 1993; Hughen et al., 1998). There is also an interval from 10.6 to 10 ka 14C B.P. when varve ages change by -1600 yr (Fig. 3.12). These 14C age plateaus document periods of time when atmospheric 14C concentrations temporarily increased, so that organisms acquired higher levels of 14C at those times. Consequently, they now appear to be the same age as organisms that are, in fact, several hundred years younger. In effect, this means that two samples with 14C ages of 10,000 and 9600 yr B.P., for example, could in reality differ by as much as 860 years or as little as 90 years. The changes in atmospheric 14C beginning at -10.6 ka (radiocarbon) B.P. (+40-70%o in <300 yr) correspond to the onset of the Younger Dryas cold episode (Goslar et al., 1995). A sudden reduction in North Atlantic Deep Water (NADW) formation at that time (or more probably a shift from deep water transport to intermediate water) would have changed the equilibrium atmospheric 14C concentration by reducing the ventilation rate of the deep ocean (Hughen et al., 1996b, 1998).

3.2.1.6 Causes of Temporal Radiocarbon Variations

Table 3.2 lists some of the possible causes of 14C fluctuations, and these are discussed in some detail by Damon et al. (1978). In general terms they can be considered in two groups: factors internal to the earth-ocean-atmosphere system (II and III in Table 3.2) and extraterrestrial factors (group I in Table 3.2). Although it is probable that all these different factors have played some part in influencing 14C concentrations through time, it would appear that most of the variance in the record, as it is currently known, can be accounted for by changes in the intensity of the Earth's magnetic field (dipole moment) (Mazaud et al., 1991; Trie et al., 1992) and by changes in solar activity (Stuiver and Quay, 1980; Stuiver, 1994). The former factor is primarily related to low-frequency (long-term) 14C fluctuations, and the latter factor to higher-frequency (de Vries-type) fluctuations in 14C. Evidence for changes in magnetic field intensity has come mainly from archeological sites through studies of thermoremanent magnetism in the minerals of baked clay (Bucha, 1970; Aitken, 1974). Although there are uncertainties in this chronology, it appears that there is a strong inverse correlation between magnetic field variations and 14C concentration, such that as the magnetic field strength decreases (thereby allowing more galactic cosmic rays to penetrate the upper atmosphere) 14C concentration increases (Stuiver et al., 1991; Sternberg, 1995).

On a shorter timescale, variations in solar activity also influence 14C concentrations. This was well illustrated by Stuiver and Quay (1980), who used recorded sunspot data to demonstrate a convincing relationship between periods of low solar activity and high 14C concentrations. Annual records of A14C (and 10Be, an other cosmogenic isotope) also show very strong periodicities related to the sunspot cycle of solar activity (Stuiver, 1994; Beer et al., 1994). Solar magnetic activity is reduced during periods of low sunspot number and this allows an increase in the intensity of galactic cosmic rays incident on the Earth's outer atmosphere, thereby increasing the neutron flux and 14C (and 10Be) production (Fig. 3.13). Thus, during the Maunder, Sporer, and Wolf periods of minimum solar activity (A.D. 1654-1714, 1416-1534, and -1280-1350, respectively) 14C concentrations were at their maximum levels for the past thousand years (Eddy, 1976; Stuiver and Quay, 1980). By subtracting the low frequency geomagnetic signal from the overall A14C record, a residual component is revealed that primarily reflects solar (heliomagnetic) effects (Fig. 3.14). This shows a number of Maunder and Sporer-type A14C maxima throughout the Holocene, each lasting 100-200 yr (Stuiver et al., 1991).

Whatever the underlying reasons for the high frequency variations in A14C, they are extremely important for the interpretation of 14C dates in terms of calendar years. The 14C variations or "wiggles" may not permit the assignment of a unique

FIGURE 3.13 Global l4C production rates Q derived from measured neutron fluxes (1937-70) in relation to sunspot numbers S (plotted inversely). During periods of higher solar activity cosmic-ray bombardment of the upper atmosphere is reduced, causing HC production to decrease.The broken line represents the long-term change in l4C production during solar minima (Stuiver and Quay, 1980).

YEAR

FIGURE 3.13 Global l4C production rates Q derived from measured neutron fluxes (1937-70) in relation to sunspot numbers S (plotted inversely). During periods of higher solar activity cosmic-ray bombardment of the upper atmosphere is reduced, causing HC production to decrease.The broken line represents the long-term change in l4C production during solar minima (Stuiver and Quay, 1980).

"F

3000 B.C

FIGURE 3.14 Residual AHC from a ~400 yr spline applied to the record shown in Fig. 3.9.The spline (a type of low-frequency filter) is used to extract low-frequency changes thought to be the result of changes in the Earth's magnetic field.The residual represents changes due to solar activity. Prolonged episodes of reduced solar activity (and enhanced l4C production) similar to the most recent Maunder and Sporer Minima are denoted by M or S (from Stuiver et ai, 1991).

calendar age to a sample. This is illustrated in Fig. 3.15; a sample radiocarbon dated at 220 ± 50 yr B.P. cannot be assigned a single age range, within the probability margin of one standard deviation. Because of 14C fluctuations the actual calendar age of the sample could be from 150 to 210 yr B.P., from 280 to 320 yr B.P., or even from 410 to 420 yr B.P. (Porter, 1981a). In fact, only samples from a few decades around 300 years B.P. are likely to yield a unique radiocarbon date. In all other cases during the last 450 yr, multiple calendar dates, or a much broader spectrum of calendar ages, are derived from a single radiocarbon date (Stuiver, 1978b). This raises significant problems for studies attempting to resolve short-term environmental changes (such as glacier fluctuations; Porter, 1981a) and has important implications for the interpretation of radiocarbon dates at certain times in the Holocene (McCormac and Baillie, 1993). To help in taking such variations into account when calibrating 14C ages in terms of calendar years, Stuiver and Reimer (1993) have prepared a computer program that provides all possible calendar year ages for a particular 14C date and associated margins of error (1 or 2 a). They stress that the margin of error associated with a dated sample should not only be the analytical error term, but should also include an "error multiplier" factor that varies from 1 to 2, reflecting the reproducibility of results in each laboratory (Scott et al., 1990). This factor is readily accommodated in the program to compute the appropriate calendar year ages. Similar software has also been developed by Ramsey (1995).

0 100 200 300 400

FIGURE 3.15 Relationship between conventional ^C and calendar years. Shaded curve is twice the counting error in the measurements. For a radiocarbon date of 220 ± 50 yr B.R the actual calendar year represented by the date and its counting error could be anywhere within the range of 150-210,280-320, and 410-420 calendar years before 1950 (B.R) (from Porter, l98la;Stuiver, 1978b).

0 100 200 300 400

Calendar years before A.D. 1950

FIGURE 3.15 Relationship between conventional ^C and calendar years. Shaded curve is twice the counting error in the measurements. For a radiocarbon date of 220 ± 50 yr B.R the actual calendar year represented by the date and its counting error could be anywhere within the range of 150-210,280-320, and 410-420 calendar years before 1950 (B.R) (from Porter, l98la;Stuiver, 1978b).

One consequence of the nonlinear calendar year -14C age relationship is that a series of equally probable calendar dates may produce a histogram of strongly clustered 14C dates (Fig. 3.16). Similarly, at certain times in the past, a histogram of 14C dates showing two pronounced "events" may, in fact, correspond to a normal distribution around a single event (Bartlein et al., 1995). This is not the case for the entire period calibrated so far, because there are times when there is less (local) variability around the 14C-calendar year relationship, but certain key intervals (such as the Younger Dryas-Preboral transition) require very careful interpretation of radiocarbon dates to avoid misinterpretation of the true sequence of events. Similar caution is appropriate in dealing with estimates of rates of change during the late Glacial and early Holocene (Lotter et al., 1992). Indeed, as the calibration of radiocarbon dates is extended back into the period before 11,400 yr ago, other periods of near-constant radiocarbon age may yet become apparent.

3.2.1.7 Radiocarbon Variations and Climate

A number of authors have observed that periods of low solar activity, such as the Maunder minimum, correspond to cooler periods in the past (Eddy, 1977; Lean et ah, 1995). As variations in radiocarbon production seem to be related to

0 0

Post a comment