All?gre Claude J. Isotope Geology

Подождите немного. Документ загружается.

caps. For benthic species, th e dominant factor is the isotopic £uctuation of the ocean as

Shackl eton (1967b) had surmised. In addition, Duple ssy’s group and Shackleton estab-

lished that an additional isotope fractionation occurred which was characteristic of each

spe cies of foraminifera studied. But those ‘‘ vital e¡ects’’ were calibrated and so isotopic

measurementson di¡erentspecies could be made consistentw ith each other.

Exercise

We have just seen that the d

O variation of foraminifera was mostly due to d

O variation

because of melting ice. Let us look more closely at the quantitative inﬂuence of melting polar

ice on d

O. Imagine an intense glacial period when the sea level falls by 120 m. What would

be the volume of polar ice and the d

O value of sea water?

Answer

The ocean surface area is 3.61  10

. The volume of the ocean is 1370  10

. The

volume of present-day polar ice is 29  10

. If the sea level is 120 m lower, 46  10

has been stored in the ice caps, corresponding to a mass fraction of the hydrosphere of 3.3%.

The polar ice caps were 1.6–2 times larger than today.

If we take the d

O value of ice as 50ø, then 50ø 0.033 ¼1.65ø. There is indeed a

difference in d

O of this order of magnitude between glacial and interglacial periods. Notice

that, as with radiogenic isotopes, these effects could not be detected if we did not have a very

precise method for measuring d

Glaciers

Another interesting application of this fractionation wasbegun on glaciers independently

by Sa m uel Eps tei n ofthe California Institute of Technology and (more systematically and

continuously) byWil li Dan sgaard of Copenhagen University.When a core ofpolar glacier

ice is taken, it has layers ofstrati¢ed ice which can be dated by patient stratigraphy and var-

ious radiochronological methods. Now, the study of these ice strata reveals variations in

OanddD (Dansgaard, 196 4; Epstein et al., 1965; Dansgaard and Tanber, 1969)

(Figures7.29 and7.3 0).

For a single region such variations are analogous and mean the sequence of one glacier

can be matched with the sequence of a neighboring glacier. An isotope stratigraphy of

glaciers c an be de¢ned. We can venture an interpretation of these facts in two ways.

Either we accept that the origin of precipitation has varied over recent geological time

and we then have a way of determining variations in the meteorological cycle of the past.

Or we consider that the fractionation factor has varied and therefore the temperature has

varied.

Research by Dansgaard and his team on the ice ¢rst of Greenland and then of

Antarctica showed that th e te mpe rat ure e¡e c t i s predo mina nt. By simultaneously

measuring isotope composition and temperature, he showed that the d

OanddD

correlation did indeed correspond to this e¡ect. Moreover, the qualitative rule is the

reverse of the carbonate rule: when the temperature rises, both d

OanddD increase,

because fractionation diminishes with temperature; but the d values are negative and so

move clos er tozero(Figure 7.31). A simple empiricalruleisthatwhenever thetemperature

rises by18C, d

O increasesby 0.7ø.

409 Paleoclimatology

We can investigate why d

£uctuation is very important for foramin ifera and why it is

the local temperature e¡ect that dominates with ice. Because isotope fractionation at very

low temperatures becomes very large and dominates isotope £u ctuation related to the

water cycle. But we shall see thatthis assertion mustbequali¢ed. Modern studiesof isotope

£uctuations of glaciers use a co mbination of both e¡ects, loc al temperatures and isotopic

changes in thewater cycle, as for foraminifera,butwith di¡erentrelativeweightings.

7.7.2 Systematic isotope paleoclimatology of the Quaternary

Wehavejustseentherearetwowaysofrecording pastte mperatures.

(1) Oneisbasedon 

Oanalysisoffoss ilshellsinsedimentaryseries (marineand continen-

tal cores).

(2) Theother uses 

Oanalysis ofaccu mulated layers of icein theice caps.

Both th ese methods have progressively converged to allow very precise studies of cli matic

£uctuations in the Quaternary and more especially for the last million years. Nick

Shack leto n andWilli Dansgaard sharedthe CrafoordPrizeinrecognitionoftheircomple-

mentaryachievement. Each method has its limits, and itis onlygradually that wehavebeen

able to compareand usebothtypesofrecords in a complementary way to decipher climatic

variations thathavea¡ectedour planetoverthe last mi llionyears andwhich consistin alter-

nati ng glacial andwarmer inte rglacial periods.

O(‰)

Depth (m)

3 5

GlacialGlacial

7 9

C(‰)

3 5 7 9

-1

Inter-

glacial

Inter-

glacial

Cibicides sp.

Neogloboquadnina. pachyderma

Figure 7.29 Comparison of d

C and d

O ﬂuctuations in a pelagic (blue circles) and a benthic (white

circles) species of foraminifer from the Antarctic Ocean. The d

C variation (above) is represented to

show there is a shift for d

O (below) but not for d

410 Stable isotope geochemistry

2100

2000

1900

1800

1700

1600

1500

–32 –31 –30 –29 –28 –27 –26 –25 –24 –23

Depth (cm)

δO

‰

2100

2000

1900

1800

1700

1600

1500

–25 –24 –23 –22 –21 –20 –19 –18

Depth (cm)

δD%

(S?)

Figure 7.30 Variations in d

O and dD with depth in an Antarctic glacier. Summers (S) can be

distinguished from winters. (Caution! dD is in percent!) After Epstein and Sharp (1967).

–200

–30

–20

–40

–150

–250

–300

–350

–50 –40 –30

Temperature (°C)

δD (‰)

O (‰)

–20 –10

Antarctic

(Terre Adélie)

Greenland

Figure 7.31 Relationship between temperature, 

O, and D in snowfall at the poles. After Lorius and

Merlivat (1977) and Johnsen

.(1989).

411 Paleoclimatology

Core sampling (sequential rec ords)

Cores of marine sediments canbe taken from all latitudes and longitudes (in the ocean and

from continentsand lakes); however, two conditions restricttheir use. First, sedimentation

must have occurred abovethe‘‘carbonate compensation depth’’ for there tob e any measur-

able fossil tests left. And second, sedimentation must have been very rapid to provide a

recordwithgoodtime resolution. Sedimentary coreshaveno time limits other thanthe life-

span of the ocean £oor. Quaternary, Tertiary, and Secondary cores can be studied up to

120 Ma, which is the age of the oldest remnants of oceanic crust that have not been swal-

lowe d up bysubduction (ancientcores are c ompacted andtransformed intohard rocks and

sotime resolution is not asgood).

For ice caps,the¢rstproblem is,ofcourse, theirlimitedgeographicalandtemporal extent.

Geographically, records are primarily from the glaciers of Antarctica and Greenland.

Mountainglaciershavealso recordedclimatic eventsbutovermuchshortertime-spans.

Ice caps are limited in time. For a long time, the longest core was one from Vostok

in Antarctica covering 420 000 years. A new c ore of EPICA has been drilled and covers

700 000 years. Cores from the big mountain glaciers go back a mere 2000 years or so. For

b oth types of record ^ se diments and ice ^ precise, absolute dating is essential, but here

again many di⁄culties arise. Especially because as research advances and as studies

becomeever more re¢nedforeversmallertime-spans,theneedfor precision increases con-

stantly.There is scope for

C dating and radioactive disequilibrium methods on sedimen-

tary cores, buttheir precision leaves something tobe desired. Useful cross-checki ng canbe

done with paleomagnetism and well-calibrated paleontological methods. In turn, the oxy-

gen isotopes ofawell-dated core canbeused todatethelevelsofothercores. Thus,gradually,

a more orless reliable chronologyis established, which mustbe constantly improve d. Dating

is di⁄culton ice caps exceptfor the mostrecentperiodswhereannual layers canbe counted.

Methods based on radioactive isotopes such as

Be,

Cl,

Kr, and

Ar are used, but

theyare extremelydi⁄culttoimplementboth analytically (ice is averypure material!)andin

terms ofreliability. Switzerland’s Hans Oeschger (and his team) is associatedwith the devel-

opment of these intricate techniques for dating ice, which, despite their limitations, have

broughtaboutdecisiveadvancesin decipheringtheicerecord (Oeschger,1982).

These clari¢cations should make it understan dable that establishing ti me sequences of

recordsis a di⁄cultand lengthy jobthatis constantlybeing improved.A ll reasoning should

make allowancefor this.

Deciphering sedimentaryseries and the triumph of Milankovitch’s theory

Between 1920 and 1930, the Yugoslav mathematician and astronomer Milutin

Milankovitch developed a theory to account for the ice ages that had already been identi-

¢ed by Quaternary geologists (see Milankovitch, 1941). These periods were thought to

be colder.The polar ic e extended far to the south and mountain glacie rs were more exten-

sive too (Figure 7.32). Alpine glaciers stretched down as far as Lyon in France. These

glacial traces can be identi¢ed from striated rock blocks forming what are know n as

moraines.

They have been used by Lony Thompson of Ohio State University for careful study of recent tempera-

ture ﬂuctuations (see his 1991 review paper).

412 Stable isotope geochemistry

Mil ankovitch’s theory gave rise to vehement controversy (as vehement as that over

We g e n e r ’s theoryofcontinentaldrift

). And yetthisvision ofthe pioneers ofthe1920s was

largelyaccurate.This isnottheplacetosetoutthistheoryin full.Itcanbefoundintextbooks

on paleoclimatology, forexample, Bradley(1999). However, we shal l outlinethe mainprin-

ciples and theterms toclarify what wehaveto sayaboutit.

The Earth’s axisofrotation isn otperpendicular toitsplaneofrotation aroundthe Sun. It

deviates from it by 238 on average. But the axis of this deviation rotates around the vertical

over a period of 23 000 years. A further movement is superimposed on th ese, which is the

£uctuation of the angle of deviation between 21.88 and 24. 48.The period of this £uctuation

is 41 000 years.The ¢rst of these phenomena is termed preces sion,thesecondisobl iqu it y.

A third phenomenon is the variation in the ell ipiticity of the Earth’s orbit,withaperiod

of 95 000 years. These three phenomena arise from the in £uence on th e Earth of the

Sun, Jupiter, and the other planets and the tides. They not only combine but are

Today

18 000 years ago

Figure 7.32 Worldwide distribution of ice in glacial and interglacial times.

Wegener had been the ﬁrst, with his father-in-law Ko

ppen, to suggest an astronomical explanation for

the ice ages before becoming an ardent defender of Milankovitch.

413 Paleoclimatology

0.05

400

Total

Precession

Obliquity

Excentricity

200

413

100

600 800

0.02

0.0

24.5

23.3

22.0

-0.07

-0.02

0.04

2.7

0.0

2.7

Time (ka BP)

Fourier spectrum

Summer

solstice

Winter

solstice

Autumn

equinox

Winter

equinox

Sun

Perihelion

Precession

Obliquity

Aphelion

Age (ka)

III

Insolation in June

1230

950

80°N

60°N

40°N

20°N

0°N

20°N

40°N

60°N

80°N

1130

900

1140

930

880

740

870

520

360

280

100

1080

910

1000 200 300

Age (ka)

III

Insolation in December

100

80°N

60°N

40°N

20°N

0°N

20°N

40°N

60°N

80°N

1000 200 300

370

270

670

520

810

740

1080

880

1150

930

1150

910

1260

980

Figure 7.33 The principle of Milankovitch’s theory. (a) The three parameters that change: eccentricity of

the orbit, obliquity, and precession. (b) Variations in the three parameters calculated by astronomical

methods with their Fourier spectrum on the right. (c) Variation in the sunlight curves in June and

December with latitude, calculated by the theory.

414 Stable isotope geochemistry

superimposed, leading to complexphenomena.Thus, at present, the Earth is closestto the

Sun on 21 December, but the Earth’s axis is aligned away from the Sun, and so, in all, the

northern hemisphere receives little sunlight. It is winter there, but other conjunctions

also occur. Thus we can calculate the sunshine received during the year at various lati-

tudes. Cel estial mechanics mean such calculations can be made precisely (see Figure 7.33

for a simpli¢ed summary). As Milankovitch understood, if little sunshine reaches the

Earth at high latitudes in summer the winter ice will remain, the white surface will re£ect

solarradiation, andth e coolinge¡ectwillbeampli¢ed.This isagoodstar ting pointforcli-

matic cooli ng.

What should be remembered is that when we break down the complex signal of sunlight

received by the Earth using Fourier analysis methods (that is, when we identify the si ne-

wave frequencies that are super imposed to make up the signal) we ¢nd peaks at 21 000,

41000,and 95 000years.Whenwe conduct a similarFourierdecompositionfor d

Ovalues

recorded by foraminifera in sedimentary series, we ¢nd the same three frequencies

(Figure7.34 ).

The d

O variations re£ect those of the Earth’s temperatures. This ¢n ding con¢rm s

Mil ankovitch’s theory (at least as a ¢rst approximation) and so fully bears out the early

studies of Emiliani. In complete agreement with the theory, th e sedimentary series also

showed that climatic variations were very marked at the poles (several tens of degrees),

very low in the intertropical zone, and intermedi ate in the temperate zones (of the order

of 5 8C) .

Figures7.35 and 7.36 giveafairlycompletesummaryofthe essentialisotopicobservation s

made fromsedimentarycores.

The sedimentarycorerecord(Figure7.35)also shows in detail how thetemperaturevar-

iations evolved. Cooling i s slow, followed by sudden warming. Finer £uctuations are

superimposed on these trends but their frequencies match those of the Milankovitch

cycles.

Con¢rmation of Milankovitch cycles byAntarctic isotope records

Itwassome considerabletimebeforeMilankovitchcycleswere con¢rmedintheicerecords,

for two reasons.There were no ice cores long enough and so covering a long enough time-

span andthe dating methodsweretooimprecise.

It was only after the famous Vostok core from Antarctica was studied by the Franco-

Russ ianteam thatevi dence of Milankovitch cycles was foundin the ice record.Butthe core

yielded much morethanthat: it allowed climaticvariations tobe correlated withvariations

ofother parameters:



dustcontent: itwas realizedthat during ice ages therewas much more dust andth erefore

morewindthanduring interglacialperiods.



greenhouse gas (C O

and CH

) content in air bubbles trapped inthe ice ^ when the tem-

peratureincreases there is an increase in CO

(inthe absence ofhumanactivity!).

This last question on the debate about the in£uence of human activity on the greenhouse

e¡ect and so on climate is a fundamental one.Which in creased ¢rst, temperature or CO

levels? It is a di⁄cult problem to solve because temperature is measured by dDinice

415 Paleoclimatology

100 30 15 10 7.5 6

Amplitude of climatic cycle

highlow

Cycle duration (ka)

Figure 7.34 Fourier spectrum of paleotemperatures using oxygen isotopes.

3.0

5.0

4.0

3.0

4.0

5.0

0.0

0.2 0.4 0.6 0.8

200

2 6 8 10 12

18 20

400 600 800

3.0

4.0

5.0

O(‰) δ

O(‰)

Isotopic stages

-1

-2

200 0 400 600 800

Age (ka)

4 6 8

10 12 14 16 18 20 22

Site 677

Site 552

Site 607

Figure 7.35 Records of d

O in foraminifera and the synthetic reference curve. (a) Record of d

O for

benthic foraminifera at three sites: site 552 – 568 N, 238 W in the North Atlantic; site 607 – 418 N, 338 W

in the mid Atlantic; site 667 – 18 N, 848 W in the equatorial Paciﬁc. Correlation between the three cores

is excellent. (b) Synthetic reference curve produced by tuning, which consists in deﬁning the timescale

so that the Fourier decomposition frequencies of the d

O values of the cores match the astronomical

frequencies from Milankovitch’s theory. The period is then subdivided into isotopic stages. The odd

stages are warm periods and the even stages (shaded) glacial periods. (Notice that the interglacials

correspond to increased d

O values and ﬂuctuations are just a few per mill.) After various compilations

from Bradley (1999).

416 Stable isotope geochemistry

whereas CO

is measured from its inclusion in ice. Now, gaseous inclusions are formed

by the co mpacting of ice and continue to equilibrate with the atmosphere, that is, the air

samplesareyounger thanthe icethatentrapsthem (Figure 7.37).

We need, then, to be able to measure the temperature of inclusions directly and com-

pare it with the temperat ure measured fr om the D value of the ic e. A method has been

developedbySeveringhausetal.(2003)formeasuringthetemperatureof£uidinclusions

using

Ar/

Ar and

N isotope fractionation. After stringent calibration, the teams

at the Institut Simon-Laplace at Versailles University and at the Scripps Institution of

Oceanography (La Jolla, California) managed to show that the increase in CO

lags

behind the in crease in temperature by 800 years (Figure 7.38) and not the other way

round as asserted by the traditional greenhouse-e¡ect model (Severinghaus et al., 199 9;

Caillonetal.,2003).Now,weknowthatCO

solubility in sea water declines as tempera-

ture rises and that the characteristic time for renewal of the ocean water is 1000 years.

The ¢rst phase of temperature i ncrease followed by the increase in CO

,withalagof

800 years, can be readily understood, then, if we invoke the lag because of the thermal

inertia of the ocean.There may also be some feedbackof the CO

e¡ect on temperature.

Insolation (J)

(ppbv)

O atm (‰)

Temperature

(°C)

Age (years BC)

280

100 000 200 000 300 000 400 000

240

700

600

500

400

100

200

–2

–4

–6

–0.5

–0.0

0.5

1.0

Dust

Ice

volume

δD (‰)

Na (ppb)

O atm (‰)

100 000 200 000 300 000 400 000

–420

–440

–460

–480

0.0

0.5

1.0

1.5

1.0

0.5

110 ka 390 ka

5.4

11.24

5.1

5.3

5.5

7.1

7.3

7.5

8.5

9.1

9.3

11.3

11.1

100

1.0

0.5

0.0

–0.5

Figure 7.36 Various parameters recorded in the Vostok ice core. After Jouzel

.(1987) and Petit

(1999).

417 Paleoclimatology

Exercise

Argon and nitrogen isotope fractionations are caused by gravitational fractionation in the ice

over the poles. Using what has been shown for liquid–vapor isotope fractionation, ﬁnd the

formula explaining this new isotope thermometer.

Answer

If we write the fractionation:

Age (BC) for δ

Age in years (BC) for CO

235 000 240 000 245 000 250 000

200

220

240

260

2.20

2.15

2.10

800 years

2.05

2.00

Ar (‰)

(ppm)

Figure 7.38 Records of 

Ar and CO

from the Vostok core after shifting the CO

curve 800 years

backwards. After Caillon

.(2003).

Depth (m)

50-120

0.83

0.80

0.55

Density (g cm

–3

)

Surface

Free contact

with atmosphere

Isolated air

Diffuse contact

with atmosphere

Compaction

Δt = interval of time

Snow

Ice

Closure of

gas inclusion

Figure 7.37 As ice is compacted, the air trapped is younger than the snow.

418 Stable isotope geochemistry