Info

e-sind)

AAAAAAAA,

Natural Archive core J recovery analysis and calibration

FIGURE 6.14 Schematic diagram illustrating the relationship between orbital forcing and the signal that is eventually preserved in the sedimentary record.

then to examine the coherency spectrum of the tuned record with the spectrum of orbital forcing at frequencies other than those used in the tuning procedure. In the tuned records obtained by Imbrie et al., (1984), coherency was not only very high in the tuning bands but also in the -100 ka eccentricity band, which was not used in tuning at all.

This tuning procedure was applied to a set of 8lsO records, which had previously been "stacked" (aligned with respect to important stratigraphic features, but without a timescale), to produce a standard reference chronostratigraphy for the late Quaternary (Fig. 6.15). The resulting "SPECMAP" (Spectral Mapping Project) record thus enables other 8180 records to be adjusted to fit the reference chronology, even if no dates are available for the new records (Prell et al., 1986; Martinson et al., 1987). Table 6.2 gives the estimated ages of major stage boundaries over the last 250 ka based on the orbital (or astronomical) tuning approach. The isotopic record can be viewed (in terms of its major low-frequency components) as being

NM ri ri Tf\ftí i(j ui toiri r; N r; CO f> N » ^ U>

Specmap Stack Mis

NM ri ri Tf\ftí i(j ui toiri r; N r; CO f> N » ^ U>

FIGURE 6.15 The SPECMAP composite chronology for a set of seven stacked (superimposed) 8I80 records from different ocean basins of the world.The records were stacked according to their common strati-graphic characteristics (like those used to derive Fig. 6.1 I) but without a reliable chronology. The chronology applied here is derived by orbital tuning that assumes that the primary forcing controlling the frequency characteristics of each sedimentary record is related to Milankovich orbital variations and that there has been a constant phase lag between the orbital forcing and system response over the last 300,000 yr. The tuning was first applied to subpolar Indian Ocean core RC11-120 and the chronology then transferred to the composite set of normalized records (i.e., each had a mean of 0 and a standard deviation of I). Major stadial episodes (isotopic stages 2,4,6, and 8) are shaded. Vertical lines are times of transition between stages or substages as defined by Martinson et al. (1987); substages 5a to 5e are centered on lines 5.1 to 5.5, respectively. The age error estimates on the stacked stage boundaries average -±5000 yr but this varies from one part of the record to another, as shown in Table 6.2 (Martinson et al., 1987).

6 MARINE SEDIMENTS AND CORALS TABLE 6.2 Estimated Ages of Oxygen Isotope Stage Boundaries and Terminations

Boundary0

Termination11

Estimated ages (x Ar Bd

103 years) Ce

Df

Error«

2.0

I

13

11

11

12.05

3.14

3.0

32

29

27

24.11

4.93

4.0

64

61

58

58.96

5.56

5.0

75

73

72

73.91

2.59

5.1

79.25

3.58

5.2

90.95

6.83

5.3

99.38

3.41

5.4

110.79

6.28

5.5

123.82

2.62

6.0

II

128

127

128

129.84

3.05

7.0

195

190

188

189.61

2.31

8.0

III

251

247

244

244.18

7.11

8-9

297

276

279

9-10

IV

347

336

334

10-11

367

352

347

11-12

V

440

453

421

12-13

472

480

475

13-14

502

500

505

14-15

542

551

517

15-16

VI

592

619

579

16-17

627

649

608

17-18

647

662

671

18-19

688

712

724

Brunhes/Matuyama boundary

700

728

Jaramillo (top)

908

Jaramillo (bottom)

983

Olduvai (top)

1640

Olduvai (bottom)

1820

" Isotope boundary designation 2.0 (Pisias et al., 1984) is alternatively referred to as 1-2.

b Terminations from Broecker and van Donk (1970). They defined terminations on the basis of their interpretation of the saw-toothed character of the oxygen isotope record.

c Estimates from Shackleton and Opdyke (1973) by linear interpolation in core V28-238, using a mean sedimentation rate of 1.7 cm per thousand years.

d Estimates from Hays et al. (1976), Kominz et al. (1979), and Pisias and Moore (1981), based on the assumption that variations in the tilt of the Earth's axis (obliquity) have resulted in variations of global ice volume, and that the phase shift between the Earth's tilt and the 41,000-yr component of the isotopic record has remained fixed with time. See Section 6.8 for discussion.

e Estimations from Morley and Hays (1981) based on adjustments to maintain a constant-phase relationship between variations in oxygen isotope ratios and changes in obliquity and precession.

f Derived from orbital tuning by Martinson et al. (1987); paleomagnetic transition ages from Shackleton et al. (1990). * Error estimates for the stacked record of Pisias et al. (1984) derived from the orbital tuning of Martinson et al. (1987) applied to those records; error averages ± 5000 yr over the last 300,000 yr, but is lower in some segments than in others.

made up of periods of gradually increasing 8lsO separated by shorter, relatively abrupt episodes when 8180 values decrease. The curve is thus "saw-toothed" in character, the slow increase in 8180 resulting from the gradual build-up of ice on the continents, followed by a period of rapid déglaciation when isotopically light water was returned to the oceans. Broecker and van Donk (1970) referred to the sharp decreases in 8lsO as Terminations, signifying the end of a glacial period, the most recent déglaciation being Termination I. Estimated ages of other Terminations are included in Table 6.2.

Orbital tuning has now been extended back over 2.5 Ma yr for selected cores (Fig. 6.16) (Shackleton et al, 1990; Hilgen, 1991; Imbrie et al, 1993b; Chen et al, 1995). A number of important features are apparent in such records. First, 8lsO values have rarely been lower than Holocene levels, indicating that there have been very few occasions when there has been less continental ice than there is at present; these episodes were principally in isotope stages 5e, 9, 11, 31, and 37 (Raymo, 1992). From 3.1 to 2.6 Ma B.R an increase in 8180 indicates a progressive cooling, probably associated with the growth of the Antarctic ice sheet, and of continental ice sheets in the northern hemisphere. Around 2.7 Ma, a dramatic increase in the production of ice-rafted debris is seen in marine sediments from the North Atlantic. This increase in frequency corresponds with an increase in 8lsO values to levels typical of recent stadials. Only since isotope stage 22 (-800 ka ago) have there been episodes of continental glaciation comparable in magnitude to the most recent ice age (stage 2). Finally, the record provides a remarkable perspective on the increased importance of variance in the 100 ka (eccentricity) frequency

FIGURE 6.16 Benthic oxygen isotope record from equatorial Atlantic core ODP-607 for the past 3.2 Ma. Dashed horizontal lines indicate the 8lsO values for the Holocene (upper line), stage 5c (middle line), and the stage 2/1 boundary (lowest line), as recorded at this site. Note the increased importance of the — 100 ka cycle within the last I M yr (Raymo, 1992).

FIGURE 6.16 Benthic oxygen isotope record from equatorial Atlantic core ODP-607 for the past 3.2 Ma. Dashed horizontal lines indicate the 8lsO values for the Holocene (upper line), stage 5c (middle line), and the stage 2/1 boundary (lowest line), as recorded at this site. Note the increased importance of the — 100 ka cycle within the last I M yr (Raymo, 1992).

band over the last 1 M years, when continental ice sheets were much larger (as registered by higher levels of 8lsO) compared to the preceding period. It seems unlikely that this is due simply to a change in the amplitude of eccentricity variations because over this interval there has actually been a shift away from variance in the 100 ka eccentricity frequency band, and an increase in variance at lower frequencies (-412 ka) (Imbrie et al., 1993a). The 100 ka period in 8lsO is therefore likely to result from feedbacks internal to the climate system, which amplified the orbitally driven radiative forcing. From ice core records, greenhouse gases are clearly implicated in this amplification process (see Section 5.4.3). Ruddiman et al., (1986) also point to the possibly critical effects of an increase in the elevation of major mountain ranges (Himalayas-Tibet and the north-south mountain ranges of western North America). As these ranges became higher, their effect on the general circulation would have been to establish a more meridional circulation regime, which may have favored faster ice sheet growth under certain orbitally driven radiation regimes.

An important by-product of orbitally tuning the 8lsO record is that it then provides age estimates on the timing of paleomagnetic polarity boundaries recorded within the sediments, which are independent of radiometric age determinations. Thus, if the chronology of the tuned ODP record by Shackleton et al. (1990) is correct, revisions in the paleomagnetic timescale are called for, such that the Brunhes-Matuyama boundary was -780,000 B.P., the Jaramillo lasted from 1.07 M to 0.99 Ma B.P., the Olduvai from 1.95 to 1.77 Ma B.P. and the Matuyama-Gauss boundary was at -2.6 Ma B.P. In fact, recent reassessments of radiometric data indicate that the orbitally derived dates are compatible with new high resolution dates on paleomagnetic reversals (Tauxe et al., 1992; Chen et al., 1995). The marine isotope chronology can also be used to date other stratigraphic events, such as the level at which particular species became extinct (its last appearance datum, or LAD) (Berggren et al., 1980). These biostratigraphic events may then be used as chrono-stratigraphic markers in their own right, independent of both radioisotopic and stable isotope analyses on the sedimentary record in question. For example, extinction of the radiolarian Stylatractus universus has been found by stable isotope stratigraphy to have occurred throughout the Pacific and Atlantic Oceans at 425,000 ± 5000 yr B.P. (Hays and Shackleton, 1976, Morley and Shackleton, 1978). Similarly, the coccolith Pseudoemiliania lacunosa became globally extinct in the middle of isotope stage 12, at -458,000 yr B.P. and the coccolith Emiliania huxleyi made its first appearance at -268,000 yr B.P., late in isotope stage 8 (Thierstein et al., 1977).

Stable isotope stratigraphy has also enabled volcanic ash horizons to be accurately pin-pointed in time (e.g., the -53,000 year B.P. Z2 ash layer in the North Atlantic) so that they can also be used as independent chronostratigraphic markers (Kvamme et al., 1989). This is particularly important where the ash can also be found in terrestrial deposits, such as ice cores, loess or lake sediments, enabling direct correlation of land and marine records to be carried out (Gronvald et al., 1995).

6.3.4 Sea-level Changes and §l80

It was noted earlier that the bulk of the 8180 signal in benthic foraminifera is related to changes in the isotopic composition of the oceans under the influence of changing continental ice volume. As ice sheets grew on the continents, the 8lsO of the ocean increased and global sea level fell. Hence, there should be some sort of relationship between 8lsO in forams, continental ice volume, and sea-level change. However, the connections between these three phenomena are complex; the mean isotopic composition of ice sheets no doubt changed over time, depending on the ice-sheet location (latitudinally) and its mean elevation. If ice sheets remained in a steady state for extended periods, the mean 8lsO of ice lost at the margin (representing older ice formed at lower elevations) would probably have been higher than precipitation falling later on the high elevation accumulation zone of the ice sheet, leading to a systematic enrichment of the ocean in lsO without any change in ice volume. There is therefore a non-linear relationship between ice-sheet volume and oceanic 8lsO composition (Mix and Ruddiman, 1984). Furthermore, the 8180 of benthic forams is not only influenced by oceanic S180; if water temperature (or salinity) changed, this would affect 8lsO and such effects may have been more important at certain times than at others. Finally, estimates of paleo-sea level are generally based on coral reefs found on rising coastlines so that the record of former low sea-level stands are now found at locations well above present sea level. Understanding how sea level at various times in the past relates to global ice volume changes requires not only a tectonic model for local uplift but also an understanding of global sea-level distribution, relative to the mean geoid. Temporally synchronous sea-level terraces may not be at equivalent heights above present sea level due to geophysical constraints on the distribution of water on the planet (Peltier, 1994).

Notwithstanding these difficulties, the relationship between 8lsO and sea level has been considered by Chappell and Shackleton (1986) and Shackleton (1987). By comparing paleo-sea-level estimates from U-series dated coral terraces on the Huon Peninsula of New Guinea (Fig. 6.17) with the record of 8lsO in benthic foraminifera, it is apparent that the relationship between sea level change and 8lsO has not been constant over the last glacial-interglacial cycle. In fact, it seems likely that temperatures in the deep ocean must have been at least 1.5 °C cooler in glacial and interstadial times (-110-20 ka B.P.) than during the interglacials (isotope stages 1 and 5e). If it is further assumed that such a change has been typical of earlier glacial/interglacial cycles, it is then possible to estimate (from benthic 8lsO) the relative magnitudes of ice volumes in glacial and interglacial extremes back in time, for which no well-dated sea-level terraces exist. Such an analysis leads to the conclusion that continental glaciation during marine isotope stage 6 was slightly greater than in stage 2, but that in stages 12 and 16 it was even larger (Shackleton, 1987). However, there is little evidence that previous interglacials 7, 13, 15, 17, or 19 were significantly warmer (i.e., had much less ice remaining on the continents) than the present one. Stages 1, 5e, 9, and 11 were all similar (isotopically) though independent evidence suggests that stages 5e and 11 (at least) were somewhat warmer with sea level > 6 m above present levels.

Chappell Sea Level

FIGURE 6.17 Raised coral terraces, indicative of former sea levels, on the Huon Peninsula of Papua New Guinea.The coastline in this area has been continuously rising at the same time as sea level has been rising and falling due to ice volume changes on the land (eustatic sea-level changes). Upper Terrace Vila (-140 ka B.R) is slightly above Terrace Vllb (-125 ka B.P.). Terrace VI (-107 ka B.P.) is a small fringing reef above Terrace V (-85 ka B.P.).The dated sea-level record is shown in Fig. 6.13 (photograph courtesy of A.L. Bloom).

FIGURE 6.17 Raised coral terraces, indicative of former sea levels, on the Huon Peninsula of Papua New Guinea.The coastline in this area has been continuously rising at the same time as sea level has been rising and falling due to ice volume changes on the land (eustatic sea-level changes). Upper Terrace Vila (-140 ka B.R) is slightly above Terrace Vllb (-125 ka B.P.). Terrace VI (-107 ka B.P.) is a small fringing reef above Terrace V (-85 ka B.P.).The dated sea-level record is shown in Fig. 6.13 (photograph courtesy of A.L. Bloom).

0 0

Post a comment