Odekon M. Encyclopedia of paleoclimatology and ancient environments

Подождите немного. Документ загружается.

proportion of

Sr is somewhat variable because it is a decay pro-

duct of

Rb. The decay constant l for

Rb is 1.42  10

–11

–1

resulting in a half-life T of 48.8 billion years (Ga).

The present-day quantity of isotope 87 (

) depends

on the initial stock (

) and the amount of radiogenic Sr

generated by decay of

Rb over time t

Rb e

l t

 1



ð1 Þ

Because of the design of mass spectrometers, the above

quantities are measured as a ratio to the stable isotope



 1



ð2 Þ

The young Earth, at the time of its formation  4.55 Ga ago,

inherited the (

Sr/

Sr)

ratio of 0.699 from its chondritic

meteorite building blocks.

From Equation 2, it is evident that the term (

Sr /

Sr)

for

coeval rocks that originate from the same source and inherit

the same (

Sr/

Sr)

ratio depends only on their Rb/Sr ratio.

The latter is about six times larger for the fractionated rocks

that form the continents than for the basalts of the oceanic crust

(0.15 vs. 0.027). As a consequence, the average (

Sr/

Sr)

for continents is about 0.720 while for the basalts of the

oceanic crust it is 0.703 (Figure S32). The temporal evolution

of these two geologic entities is the principal control of chemi-

cal and isotopic composition of seawater, reflecting the relative

inputs from rivers and submarine hydrothermal circulation,

respectively.

Strontium isoto pic evolution of seawater

The present day

Sr/

Sr ratio of seawater is 0.7092 and is

uniform worldwide. This is because the average time that an

atom of Sr spends in seawater, the so-called residence time, is

somewhere between 2 and 5 million years, while the mixing

rate of the oceans is only thousands of years. Thus, from

the point of view of Sr, the seawater at any given time is

well mixed. Even marginal marine basins, such as the almost

brackish Hudson Bay of northern Canada, have a Sr isotopic

composition identical to that of the open sea. This is because

the Sr concentrations in riverine waters are 2–3 orders of

magnitude less than that of seawater, at 8 ppm, and any

dilution by river water, while lowering the Sr concentration,

will not impact the

Sr/

Sr ratio because almost the entire

Sr load originates from seawater.

Note that the

Sr/

Sr ratio of modern seawater is only

about 1/3 of the way up the scale between that of oceanic

and continental crust (Figure S32), despite the fact that modern

river input of Sr is about 6.5 times larger than its hydrothermal

flux (2.73 vs. 0.42  10

–1

), and therefore dominates

the oceanic budget. This is because the average

Sr/

Sr of

river water, at 0.711, is much less radiogenic than that of

the average continental crust. The discrepancy is a reflection

of the fact that carbonate rocks almost always account for a

substantial proportion of the lithology in the watersheds of

major rivers. Carbonate rocks, being marine sediments, inherit

the Sr isotopic composition of seawater at the time of their for-

mation. The oceanic, and particularly the continental crust, was

less radiogenic in the early history of the Earth (Figure S33)

and this was reflected in the contemporaneous seawater and

in the sediments that precipitated from it. Subsequently, weath-

ering of ancient marine carbonates, their redeposition in the sea

and renewed uplift and erosion resulted in perpetual recycling

of this “ancient” Sr, thus “retarding” the

Sr/

Sr signal of

seawater relative to the development of the continental crust.

Unfortunately, we do not have samples of past seawater that

would enable us to measure directly its past chemical and iso-

topic composition. On the other hand, we do have sediments

that precipitated from seawater, either directly (salts, gypsum)

or with the help of biological intermediaries that precipitated

carbonate (phosphate, siliceous) minerals as shells or cry-

stallites. We can therefore utilize shells for about 540 Ma and

carbonate rocks (limestones, dolostones) for almost 4 Ga to

decipher the Sr isotopic composition of ancient seawater. The

matter is complicated by the fact that many of these shells

and rocks were post-depositionally (diagenetically) altered

and this often results in alteration of the

Sr/

Sr signal. Tech-

niques do exist for at least partial evaluation of the degree of

such alteration, but none is foolproof. Usually, the alteration

process shifts the

Sr/

Sr values upwards and it is common

practice to accept the lowest measured value for any suite of

samples as the best approximation of the

Sr/

Sr ratio of the

coeval ocean.

In the early stages of isotope geology it was believed that the

Sr/

Sr of seawater evolved linearly during geologic history

(Wickman, 1948) and it was hoped that such a linear trend could

be utilized for dating of (bio)chemical marine sediments. Due to

the low sensitivity of early mass spectrometers, this proposition

could only be tested some 20 years later, when it was realized

that we were dealing with a widely oscillating signal (Peterman

et al., 1970). This was subsequently documented by Veizer and

Compston (1974, 1976) and Burke et al. (1982). The latest edi-

tions of these curves for the entire 4 Ga of Earth history were

published by Shields and Veizer (2002)(Figure S33) and for

the Phanerozoic by Veizer et al. (1999)(Figure S34).

The most important features of the billion-year trend in

Sr/

Sr (Figure S33) are the departure of seawater from

mantle values somewhere around the Archean/Proterozoic

transition, 2.5 Ga ago; the relatively sharp rise until about

1.8 Ga ago; the flat trend to 900 Ma ago; and another steep

rise during the Neoproterozoic, followed by a trough-like struc-

ture during the Phanerozoic. The Precambrian portion of the

0.720

0.710

0.700

Continental

crust

Rivers

“Upper

mantle”

Erosion

Seawater

Sediments

Diagenesis

Subduction

Dissolved

load

rivers

Weathering

volcanism

hydrothermal alteration

Figure S32 Schematic presentation of Sr isotopic surficial cycle

(modified after Wadleigh, 1982).

924 STRONTIUM ISOTOPES

curve is in all likelihood composed of higher order oscillations

similar to those of the Phanerozoic, but we do not yet have the

geochronological resolution to decipher it. For this reason, we

can discuss only the lower bound of the trend. The mantle-

like values of seawater during the early Archean indicate that

either the river input was negligible, or the hydrothermal flux

was much stronger than today, or both. The geological interpre-

tation is that the continents could have been much smaller

and seawater circulation via submarine hydrothermal systems

(in order to dissipate heat from the hotter mantle) more vigor-

ous. This was followed by the major interval of generation

of continental crust, between 2.7 and 1.8 Ga ago, the cumu-

lative size of continents reached near-modern extent towards

the end of this time interval. Simultaneously, the chemical

and isotopic composition of the oceans evolved from being

“mantle-buffered” to “river-buffered,” and this development

n = 5059

4,000 3,000 2,000 1,000 0

e (Ma)

0.710

0.708

0.706

0.704

0.702

0.700

Barites

Mantle evolution

Seawater evolution

Sr/

Figure S33 Strontium isotope record of carbonate rocks (seawater) during geologic history. Full symbols: chronologically well constrained; empty

symbols: poorly constrained samples (reproduced from Shields and Veizer, 2002).

Figure S34 Sr isotope record for Phanerozoic seawater based on 4,055 samples of calcite shells (reproduced from Veizer et al., 1999).

STRONTIUM ISOTOPES 925

may also have been the cause of other contemporary phenom-

ena, such as the transition from oxygen-poor to oxygen-rich

atmospheric composition. The steep rise in

Sr /

Sr in the

Neoproterozoic may be related to another pronounced phase

of continent formation, the Pan African orogeny.

Th e o rigin of t he troug h-like P han er ozoic struct ure

( Figure S34) with its superimposed oscillations is at this

stage enigmatic. Attempts to correlate the details of this struc-

ture with tectonic development of the Earth have not led to

any unequivocal conclusions. A great deal of attention was

paid particularly to the steep Tertiary rise of

Sr /

Sr values,

that was assumed to be a reflection of the collision between

India and Asia, with the resulting uplift of the Himalayas.

The corollary of such an interpretation was a potential role that

this uplift could have played in increased global erosion that, in

turn, would have resulted in atmospheric CO

drawdown and

the overall Tertiary cooling of the globe (Ruddiman, 1997).

However, the modeling simulations do not appear to support

the validity of such a scenario.

Isoto pe stratigrap hy

Despite the fact that the

Sr /

Sr trend for seawater is not

linear, the sections of this oscillating trend that show a particu-

larly steep slope can be utilized for dating and correlation

of marine (bio)chemical sediments, the approach known as iso-

tope stratigraphy (McArthur, 1994). In particular, the Cenozoic

portion of the trend ( Figure S34) enabled correlations at levels

comparable to biostratigraphic techniques.

Ján Veizer

Bibliogra phy

Burke , W. H., D enison, R.E., H etherington, E .A., K oepnick, R .B ., Ne ls on, H.F. ,

and O tto, J .B., 1982. Va ri ation of s eawater

Sr /

Sr throughout

Phane rozoic time. Geology, 10 , 516–519.

Faure, G., 1986. Principles of Isotope Geology. New York: Wiley, 589pp.

McArthur, J.M., 1994. Recent trends in strontium isotope stratigraphy.

Terra Nova, 6, 331–358.

Peterman, Z.E., Hedge, C.E., and Tourtelot, H.A., 1970. Isotopic composi-

tion of strontium in seawater throughout Phanerozoic time. Geochim.

Cosmochim. Acta, 34, 105– 120.

Ruddiman, W.F., 1997. Tectonic Uplift and Climate Change. New York:

Plenum Press, 535pp.

Shields, G., and Veizer, J., 2002. Precambrian marine carbonate isotope

database: Version 1.1. Geochem. Geophys. Geosyst., 3(6), 12 ( http://

g-cubed.org /gc2002 / 2001GC000266).

Veizer, J., and Compston, W., 1974.

Sr /

Sr composition of seawater

during the Phanerozoic. Geochim. Cosmochim. Acta, 38, 1461–1484.

Veizer, J., and Compston, W., 1976.

Sr /

Sr in Precambrian carbonates as

an index of crustal evolution. Geochim. Cosmochim. Acta, 40,

905– 914.

Veizer, J., Ala, D., Azmy, K., Bruckschen, P., Buhl, D., Bruhn, F.,

Carden, G.A.F., Diener, A., Ebneth, S., Goddéris, Y., Jasper, T.,

Korte, C., Pawellek, F., Podlaha, O.G., and Strauss, H., 1999.

Sr /

Sr,

C and d

O evolution of Phanerozoic seawater. Chem. Geol.,

161,59–88.

Wadleigh, M.A., 1982. Marine Geochemical Cycle of Strontium. M.Sc.

Thesis, University of Ottawa, ON, 187pp.

Wickman, P.E., 1948. Isotope ratios: A clue to the age of certain marine

sediments. J. Geol., 56,61–66.

Cross- refere nces

Atmospheric Evolution, Earth

Mountain Uplift and Climate Change

Stable Isotope Analysis

SULFUR ISOTOPES

Sulfur isotope systematic s

The element sulfur has four stable isotopes

S and

with the following natural abundances: 95.02%, 0.75%, 4.21%,

and 0.02% (Hoefs, 1997). Most commonly, the two major iso-

topes (

S) are being measured and results expressed as

S, defined by the following equation

S %½¼



Sample



Standard

 1

1 ; 000 ð1 Þ

and reported as per mil difference with respect to the V-CDT-

standard (defining the Zero-Point of the sulfur isotope scale).

Due to analytical improvements, the minor sulfur isotopes

and

S can now be measured routinely with their delta values

defined analogous to equation 1.

Sulfur is present in the atmosphere, hydrosphere, biosphere

and lithosphere ( Figure S35). In addition to natural processes,

the present day sulfur cycle is strongly affected by anthropo-

genic sulfur inputs. The transfer of sulfur between the different

reservoirs is mediated through abiological and biological pro-

cesses. Such processes are generally associated with a change

in the redox state and an isotopic fractionation of variable

magnitude, depending on the mode of isotope exchange (i.e.,

equilibrium vs. kinetic isotope effect). From the geological per-

spective, research focuses on sulfate as the most oxidized and

sulfide as the most reduced form of sulfur. Respective pro-

cesses are the precipitation of evaporitic sulfate from seawater

and the bacterial cycling of sulfur in sediments.

The pure precipitation of sulfate minerals from evaporating

seawater is only associated with a small, almost negligible iso-

tope effect (Claypool et al., 1980). As a consequence, marine

evaporites and sulfate incorporated into marine chemical and

biological precipitates are regarded as proxy signals for the iso-

topic composition of seawater sulfate (e.g., Strauss, 1999;

Kampschulte and Strauss, 2003).

In contrast, variable but generally substantial isotopic fractio-

nation occurs during bacterial sulfate reduction (e.g., Canfield,

2001). At the expense of sedimentary organic matter, sulfate

is reduced to hydrogen sulfide, which subsequently reacts with

Figure S35 Schematic view of the sulfur cycle showing d

S values for

modern reservoirs.

926 SULFUR ISOTOPES

Fe to form sedimentary pyrite (FeS

). The resulting sulfide

is depleted in

S with respect to ambient sulfate due to the

preferential utilization of

S by sulfate-reducing bacteria.

The size of this isotope effect is dependent on the type of organ-

ism, the reaction rate, and the availability of sulfate and reac-

tive organic matter (Canfield, 2001). Experimental data from

pure cultures indicate an isotopic fractionation of 2–42% (e.g.,

Detmers et al., 2001), whereas natural populations display an iso-

tope effect of 19–43% (Habicht and Canfield, 2001). For mod-

ern marine sediments, however, the isotope difference between

porewater sulfate and sedimentary pyrite amounts to 31–56%

(Habicht and Canfield, 2001). This additional isotopic frac-

tionation accompanies the disproportionation of partly oxidized

intermediate sulfur compounds: elemental sulfur, thiosulfate,

and sulfite (e.g., Canfield and Thamdrup, 1994; Habicht et al.,

1998). Measurements in recent marine sediments indicate that

41–85% of the apparent isotopic fractionation between seawater

sulfate and sedimentary sulfide are the result of bacterial sulfate

reduction, with the remaining part being due to disproportiona-

tion reactions (Habicht and Canfield, 2001 ).

Progressive bacterial sulfate reduction under sulfate-limiting

conditions (e.g., porewaters in deep sediments) results in a con-

tinuous depletion of

S. As a consequence, the d

S becomes

increasingly more positive. This explains part of the sulfur iso-

tope record for reduced sedimentary sulfur.

The oxidative weathering of sulfur-bearing minerals and

subsequent riverine delivery as dissolved sulfate from the

continents to the ocean is not associated with an isotope effect.

The long-term evolution of the global sulfur cycle can be

evaluated via simple isotope mass balance calculations. Con-

sidering the two principal output functions as oxidized (sulfate)

and reduced (sulfide) sulfur (see Figure S35), this mass balance

is defined as:

Input

¼ f

Sulfide

 1  f

Sulfide

ðÞd

Sulfate

ð2Þ

with d

Input

representing the average isotopic composition of

crustal sulfur (assumed constant at + 2%, Holser et al., 1988)

and f

Sulfide

and (1–f

Sulfide

Sulfate

being the fraction and

isotopic composition of sulfide and sulfate sulfur. Fluctuations

in the sulfur isotopic composition of seawater sulfate can then

be interpreted as a consequence of temporal variations in the

fractional burial of reduced (sulfide) sulfur.

Sulfur isotopic evolution of seawater

The average sulfur isotopic composition of dissolved oceanic

sulfate lies at + 21% (e.g., Rees et al., 1978). Lateral and

vertical homogeneity of this value is a consequence of

a long residence time for sulfur in the ocean (3 million

years) in comparison to the short mixing time for the ocean

(1,000 years). Only marginal seas, such as the eastern Mediter-

ranean or the Baltic Sea, represent exceptions. Their isotopic

composition is more strongly affected by the riverine input of

dissolved sulfate (on average + 8 to + 10%). However, part

of this input stems from anthropogenic sources (e.g., Grinenko

and Krouse, 1992).

Despite its homogeneity at any given point in time (e.g.,

Nielsen, 1989), empirical data clearly show that the sulfur iso-

topic composition of seawater sulfate has fluctuated widely

throughout four billion years of Earth’s history (e.g., Strauss,

2002). However, the existing isotope record is highly variable

in its temporal resolution and becomes increasingly fragmen-

tary for the early part of Earth’s history (Figure S36).

Early Archean (3.5 billion years ago) barite occurrences from

Australia, South Africa and India constrain our knowledge

about the sulfur isotopic composition of seawater sulfate from

the early part of Earth’s history at a value around þ 4% (Strauss,

2002). Sulfate abundance in the global ocean is considered to

have been very low (<1 mM, Canfield et al., 2000). This inter-

pretation is supported by a lack of substantial sulfur isotopic

fractionation between sulfate and sedimentary sulfide (e.g.,

Strauss, 2002). Still, contrasting views of an Archean high-

sulfate ocean have been proposed (e.g., Ohmoto, 1997).

The sulfur isotopic composition of seawater sulfate

increases throughout the Precambrian. Due to the fragmentary

record, it is unclear whether this increase is linear or episodic.

A significant the sulfur isotopic composition of seawater sul-

fate is considered to occur between 2.4 and 2.2 billion years

ago. It is thought to coincide with a substantial increase in

the oxygen abundance in the atmosphere, termed the Great

Oxidation Event (Holland, 2002). This would have initiated

Figure S36 Sulfur isotopic evolution of seawater sulfate.

SULFUR ISOTOPES 927

the oxidative weathering on the continents, delivering dis-

solved sulfate to the ocean.

The resulting increase in oceanic sulfate abundance would

have triggered an increase in sulfur cycling via bacterial sulfate

reduction (Canfield and Raiswell, 1999). Supporting evidence

stems from the observation that large sulfur isotopic fractiona-

tions between sulfate and sedimentary sulfide commence at that

time (e.g., Strauss, 2002). These are comparable in size to those

recorded from younger times and modern marine sediments.

Towards the end of the Precambrian (600–500 million years

ago), the average sulfur isotopic composition of seawater

sulfate reaches its all-time maximum at + 35%. This high

S value continues into the Phanerozoic.

Substantial fluctuations in the sulfur isotopic composi-

tion of seawater sulfate characterize the Phanerozoic (the last

545 million years). The high d

S values at the beginning of this

era are followed by a long-term decrease to a minimum value

of + 11% in the Permian (some 250 million years ago) and a sub-

sequent rise to the present day value. A detailed isotope record

with good temporal resolution, based on structurally substituted

sulfate (SSS) in biogenic carbonates, reveals additional higher

order oscillations (Kampschulte and Strauss, 2003).

Following the simple isotope mass balance (Equation ( 2))

these temporal variations in d

S of seawater sulfate reflect var-

iations in the fractional burial of reduced sedimentary sulfur.

Potential causes for such variations include habitat size (near

shore shelf area) and availability of reactive organic matter.

Evolution of biological sulfur cycling

The observation of a large sulfur isotope effect associated with

bacterial sulfate reduction has prompted researchers to utilize

this as a proxy for biogenicity in their studies of sedimentary

pyrite. The resulting sulfur isotope record (Figure S37) reveals

a temporal evolution that parallels the evolution of seawater

sulfate in many respects.

Early Archean sedimentary pyrite generally displays a very

limited variation in d

S near 0% and only a small difference

to the sulfur isotopic composition of barite sulfur (e.g., Strauss,

2002). Lack of a large isotope effect suggests that bacterial sul-

fur cycling played only a minor role in sedimentary systems.

Starting in the late Archean/early Proterozoic (about 2.8–2.4

billion years ago) a substantial spread in d

S for sedimentary

sulfide indicates sizeable sulfur isotopic fractionations. These

are attributed to the activity of sulfate-reducing bacteria.

Throughout the Proterozoic (2.5–0.5 billion years ago), the sul-

fur isotopic composition of sedimentary pyrite remains highly

variable. Maximum isotopic fractionation between oceanic

sulfate and sedimentary pyrite ranges between 30 and 40%,

reflecting bacterial sulfate reduction. The Neoproterozoic

(1,000–545 million years ago) witnesses not only large spreads

in d

S but also strongly positive,

S enriched sulfur isotope

values for sedimentary sulfide. Considering minimum d

values for sedimentary pyrite and d

S values of + 35% for

Neoproterozoic seawater sulfate results in a maximum isotopic

fractionation of 60%. Such a large apparent isotope fractiona-

tion reflects the isotope effect associated with disproportiona-

tion reactions in addition to bacterial sulfate reduction.

The Phanerozoic (the last 545 million years) sulfur isotope

record for sedimentary sulfur displays comparable large varia-

tions in d

S and an apparent isotopic fractionation that reflects

sulfate reduction and disproportionation. The sulfur isotope

record for reduced sulfur parallels roughly the respective record

for seawater sulfate (Figure S36): high d

S values in the early

Phanerozoic, followed by a decrease towards the Permian and a

subsequent rise during Mesozoic and Cenozoic. The compar-

able temporal evolution of both sulfur isotope records at least

during the last 600 million years suggests that bacterial sulfate

reduction and disproportionation of intermediate sulfur com-

pounds have been the key processes in biological sulfur cycling

in sedimentary environments.

Multiple sulfur isotopic evidence for an atmospheric

impact on the early sulfur cycle

The measurement of multiple sulfur isotopes (

reveal that all physical, chemical and biological processes on

Earth result in mass-dependent fractionation (MDF) of these

isotopes. Hence, mass-dependent relations can be defined as:

TFL

¼ 0:515  d

S ð3Þ

−10

−20

−30

−40

4,000 3,500 3,000 2,500 2,000

Geolo

ic a

e [Ma]

S [‰, CDT]

1,500 1,000 500 0

Figure S37 Sulfur isotopic composition of reduced sedimentary sulfur.

928 SULFUR ISOTOPES

TFL

¼ 1:9  d

S ð4Þ

with data falling on a respective Terrestrial Fractionation Line

(TFL) in a three-isotope-plot. Multiple sulfur isotope analyses

of sedimentary sulfide and sulfate of Archean and early Paleo-

proterozoic age (>2350 million years) yielded clear deviations

from these mass-dependent relationships (e.g., Farquhar et al.,

2000). Such mass-independent fractionation (MIF) in a mea-

sured sample can be quantified as:

S ¼ d

measured

 d

TFL

ð5Þ

S ¼ d

measured

 d

TFL

ð6Þ

Based on experimental and modelling data, this mass-indepen-

dent isotope fractionation is believed to result from photoche-

mical reactions in the atmosphere involving gaseous sulfur

species, in particular sulfur dioxide. Furthermore, respective

reactions require the absence of an effective UV shield (like

ozone). Consequently, the presence of MIF-S has been taken

as evidence for the absence of free oxygen in the Archean

and early Paleoproterozoic atmosphere (e.g., Farquhar et al.,

2000; 2007). Modelling evidence suggests that the maximum

concentration of atmospheric oxygen had to be <10

5

of the

present atmospheric level (PAL) in order to yield MIF-S. Sedi-

mentary sulfur younger than 2350 million years displays only

mass-dependent sulfur isotopic fractionation, indicating that

the atmospheric oxygen level had increased to >10

2

PAL

(e.g., Pavlov and Kasting, 2002).

Harald Strauss and Ján Veizer

Bibliography

Canfield, D.E., 2001. Biogeochemistry of sulfur isotopes. In Valley, J.W.,

and Cole, D.R. (eds.), Stable Isotope Geochemistry. Rev. Mineral. Geo-

chem., 43, 607–636.

Canfield, D.E., and Raiswell, R., 1999. The evolution of the sulfur cycle.

Am. J. Sci., 299, 697–723.

Canfield, D.E., and Thamdrup, B., 1994. The production of

S-depleted

sulphide during bacterial disproportionation of elemental sulphur.

Science, 266, 1973–1975.

Canfield, D.E., Habicht, K.S., and Thamdrup, B., 2000. The Archean sulfur

cycle and the early history of atmospheric oxygen. Science, 288,

658–661.

Clark, I., and Fritz, P., 1997. Environmental Isotopes in Hydrogeology.

Boca Raton, FL: CRC Press, 328pp.

Claypool, G.E., Holser, W.T., Kaplan, I.R., Sakai, H., and Zak, I., 1980.

The age curve of sulfur and oxygen isotopes in marine sulfate and their

mutual interpretation. Chem. Geol., 28, 190–260.

Detmers, J., Brüchert, V., Habicht, K. S., and Kuever, J., 2001. Diversity of

sulphur isotope fractionations by sulphate-reducing prokaryotes. Appl.

Environ. Microbiol., 67, 888–894.

Farquhar, J., Bao, H. M., and Thiemens, M., 2000. Atmospheric influence

of Earth's earliest sulfur cycle. Science, 289, 756–758.

Farquhar, J., Peters, M., Johnston, D.T., Strauss, H., Masterson, A.,

Wiechert, U., and Kaufman, A.J., 2007. Isotopic evidence for

Mesoarchaean anoxia and changing atmospheric sulphur chemistry.

Nature, 449, 706–709.

Grinenko, V., and Krouse, H.R., 1992. Isotope data on the nature of river-

ine sulfates. Mitt. Geol.-Paläont. Inst. Univ. Hamburg, 72,9–18.

Habicht, K.S., and Canfield, D.E., 2001. Isotope fractionation by sulphate-

reducting natural populations and the isotopic composition of sulphides

in marine sediments. Geology, 29, 555–558.

Habicht, K.S., Canfield, D.E., and Rethmeier, J., 1998. Sulphur isotope

fractionation during bacterial sulphate reduction and disproportionation

of thiosulphate and sulfite. Geochim. Cosmochim. Acta, 62,

2585–2595.

Hoefs, J., 1997. Stable Isotope Geochemistry. Berlin: Springer, 201pp.

Holland, H.D., 2002. Volcanic gases, black smokers, and the great oxida-

tion event. Geochim. Cosmochim. Acta, 66, 3811–3826.

Holser, W.T., Schidlowski, M., Mackenzie, F.T., and Maynard, J.B., 1988.

Geochemical cycles of carbon and sulfur. In Gregor, C.B., Garrels, R.M.,

Mackenzie, F.T., and Maynard, J.B. (eds.), Chemical Cycles in the Evolu-

tion of the Earth. New York: Wiley, pp. 105–173.

Kampschulte, A., and Strauss, H., 2003. The sulphur isotopic evolution of

Phanerozoic seawater based on the analysis of structurally substituted

sulphate in carbonates. Chem. Geol.

, 204(3–4), 255–286.

Nielsen, H., 1989. Local and global aspects of the sulphur isotope age

curve of oceanic sulphate. In Brimblecombe, P., and Lein, A.Y. (eds.),

Evolution of the Global Biogeochemical Sulphur Cycle. New York:

Wiley, pp. 57–64.

Ohmoto, H., 1997. When did the Earth’s atmosphere become oxic? The

Geochemical News, 93-Fall,12–28.

Pavlov, A.A., Kasting, J.F., 2002. Mass-independent fractionation of sulfur

isotopes in Archean sediments: Strong evidence for an anoxic Archean

atmosphere. Astrobiology, 2,27–41.

Rees, C.E., Jenkins, W.J., and Monster, J., 1978. The sulphur isotopic com-

position of ocean water sulphate. Geochim. Cosmochim. Acta, 42,

377–381.

Strauss, H., 1999. Geological evolution from isotope proxy signals –

sulphur. Chem. Geol., 161,89–101.

Strauss, H., 2002. The isotopic composition of Precambrian sulphides –

seawater chemistry and biological evolution. Spec. Publ. Int. Assoc.

Sediment., 33,67–105.

Cross-references

Atmospheric Evolution, Earth

Carbon Isotope Variations over Geologic Time

Isotope Fractionation

Stable Isotope Analysis

SUN-CLIMATE CONNECTIONS

Introduction

All energy distributed in the climate system originates from

the Sun. Earth’s surface temperature is maintained by a balance

between incoming and outgoing radiation. Irradiance is the most

important parameter related to solar variability, as the relationship

between irradiance and climate in radiative balance is simple and

direct. The Earth currently receives an average of 1,367 W m

2

integrated over all wavelengths. Climate variability directly

associated with the radiative imbalance caused by changes in

solar radiation is “solar forced.” In contrast, a “solar influ-

enced” or “solar triggered” climate change depends mostly on

atmospheric or oceanic feedback mechanisms for the effect.

The Sun also generates strong magnetic fields. Through

their interaction with Earth’s magnetic field and charged parti-

cles such as cosmic rays, magnetic fields are central to both

observational and proxy records establishing solar variability.

A current challenge in Sun-climate research is the interpreta-

tion of long-term proxy records related to the Sun’s magnetic

activity. It will only be possible to confirm a Sun-climate

connection for past climates when the innermost workings of

the Sun are understood well enough to quantify the relation

between magnetic proxies and irradiance.

On millennial and shorter time scales, there is evidence

that small changes in irradiance affect climate. It has proven

difficult to identify the physical mechanisms by which small

observed and reconstructed irradiance changes (a few tenths

of a percent) could produce observed and reconstructed climate

changes, up to 2



C within 50 years (Bond et al., 2001). Over

SUN-CLIMATE CONNECTIONS 929

the past ten years, researchers have discovered that the atmos-

phere and ocean can undergo abrupt re-organizations when per-

turbed only slightly. This can cause large regional climate

change while global average changes are small. Positive feed-

back systems have been proposed involving the upper and

lower atmosphere, oceans, and polar ice. These can amplify

the effects of small irradiance changes, and are likely critical

for Sun-climate connections unless past solar variability was

much larger than present estimates.

Solar variability

Solar cycles and sunspots

Perhaps the best-known indirect index of solar variability is the

11-year sunspot cycle. Over this cycle, the number of dark

spots visible on the Sun’s surface ranges from 0 to over 200.

Sunspots are associated with strong local magnetic activity

inhibiting surface convection, leading to cooling and darkening

relative to the rest of the visible solar surface. Bright regions

called faculae surround sunspots, causing a net increase in total

solar irradiance (TSI) of about 0.1% during sunspot peaks rela-

tive to sunspot minima. Sunspot numbers from 1610 to present

are shown in Figure S38, with the relationship to TSI measured

by satellites since 1978 shown in Figure S39.

Though measured irradiance and sunspot number are highly

correlated, direct comparison is only possible for the past three

solar cycles. The relation between long-term trends in sunspot

numbers (Figure S38) and irradiance is unclear. Sunspot num-

bers return nearly to 0 during cycle minima, while during the

Maunder Minimum, from 1645–1715, hardly any sunspots

were observed in 50 years. If irradiance varied over this period,

it must be reconstructed from records of solar activity other

than sunspot numbers.

Solar magnetic activity and the geomagnetic field

Solar magnetic activity is evident on Earth due to interaction

with the geomagnetic field. The aa index (Figure S40)isa

measure of the Earth’s magnetic field constructed from surface

magnetometers.

A portion of the solar magnetic field is carried through

the solar system and beyond by the solar wind. This is the

interplanetary magnetic field (IMF), which influences Earth’s

magnetic field. Increasing IMF strength over the past 100 years

Figure S38 Sunspot numbers and global radiocarbon production rate show coherent variability. Carbon-14 production from nuclear weapons

testing limits comparison with sunspots to before 1950. Note the lack of sunspots during the Maunder Minimum, and the increasing amplitudes

of the solar cycles (data from National Geophysical Data Center, USA).

Figure S39 Total solar irradiance and sunspots are closely associated for the past three solar cycles (data from National Geophysical Data

Center, USA).

930 SUN-CLIMATE CONNECTIONS

is responsible for the long-term upward trend of the aa index

(Lockwood et al., 1998). Though there is an 11-year cycle in aa

in addition to the long-term trend, it is not directly convertible

to irradiance because the IMF has a different origin from

magnetic activity associated with sunspots (Lean et al., 2002).

Solar variability and cosmogenic isotopes

The Earth is continually being bombarded by galactic cosmic

radiation (GCR): high-energy charged particles originating

mostly from outside the solar system. The charge on cosmic

ray particles subjects them to deflection by solar and terrestrial

magnetic fields. The partial shielding of Earth by the coupled

solar and geomagnetic fields is modulated by variability in

those fields. This is clear from cosmic ray neutron monitor meas-

urements taken since the 1950s. Figure S40 shows the neutron

monitor counting rate (GCR flux) from Climax, Colorado, USA,

which varies inversely with sunspot number.

Cosmic rays interact with Earth materials to produce cos-

mogenic isotopes, both in the atmosphere and in geologic

materials. Two atmospherically produced cosmogenic isotopes,

C and

Be, are among the most useful proxy records for solar

modulation of cosmic rays. Carbon-14 is produced from cos-

mic ray neutron reactions with nitrogen. It is rapidly oxidized

, and mixed throughout the atmosphere within about

one year. Living trees incorporate

C in their tissues during

photosynthesis. Measurements of

C in German oak tree rings

provide a continuous record of atmospheric

C concentration

back to almost 12,000 years. One complication in using atmo-

spheric

C as a proxy for solar variability is its sensitivity to

cycling of carbon in the ocean (Stuiver, 1994). In some ocean

basins, surface water is transported to the deep ocean via ther-

mohaline circulation (THC), where it may reside for more than

1,000 years before returning to the surface. Rapid changes in

THC can affect atmospheric

concentration independently

of its production. By increasing THC, more

-equilibrated

water is sequestered in the ocean, drawing down atmospheric

C. Atmosphere-ocean carbon cycling effects in tree ring

records may be evaluated by comparison with records of other,

non-cycling cosmogenic isotopes, such as

Be.

Beryllium-10 is produced from oxygen and nitrogen in the

atmosphere, rapidly sorbed to suspended aerosols, and flushed

out in precipitation. The highest quality

Be records are found

in polar ice cores. While ocean circulation changes cannot

directly affect

Be flux, the ocean can influence polar climate,

changing ice-core accumulation rates. Beryllium-10 concentra-

tion in the ice is directly convertible to

Be flux if ice-core

accumulation rates are known. Holocene ice-core accumula-

tion rates are believed to have been constant, allowing

variations to reflect changes in production. Figure S41 shows

C and

Be production rates for the past 11,000 years.

Ca II emission of sun-like stars

Conversion of the cosmogenic production rates to irradiance

remains a major challenge. One approach is to construct an

estimate of irradiance for some point in the record, which is

then used to scale back from modern irradiance. The most

widely cited estimates of irradiance beyond the observed

record come from observations of Sun-like stars at wavelengths

associated with brightness variations, such as singly ionized

calcium (Ca II) (Radick, 2003). Non-cycling stars are observed

to be dimmer than those with 8–15 year periods in Ca II

emission. Such observations suggest that the Sun was around

0.25% dimmer during the Maunder Minimum, when sunspot

cycles ceased (Lean et al., 1995, 1998; Figure S42).

A critical assumption remains unverified, that non-cycling

Sun-like stars are in a period of Maunder-type inactivity.

None of the non-cycling Sun-like stars observed over the past

two decades have resumed cycling. Given this uncertainty, true

irradiance change during the Maunder Minimum could be

either larger or equal to that during an 11-year cycle.

Correlations between solar variabili ty and climate

In a landmark paper, Eddy (1976) presented evidence suggest-

ing solar influence on climate for periods longer than the

11-year cycle. He noted that the Maunder Minimum corre-

sponded to the coldest years of the Little Ice Age, a period

of European history known for its climatic extremes. Eddy

described the striking correspondence between minima in sun-

spot observations and maxima in the tree-ring record of cos-

mogenic

C(Figure S38), suggesting that cold years during

the Maunder Minimum were linked in some way to reduced

solar activity.

In the modern observational record, there have been many

attempts to correlate the last three solar cycles with climate-

related variables. While some correlations appear convincing,

the observational records are typically too short for statistical

Figure S40 The geomagnetic aa index, neutron monitor count rate, and sunspot numbers demonstrate the influence of solar

magnetic variability on the geomagnetic field. Cosmic ray flux is lower during solar-geomagnetic maxima (data from National Geophysical

Data Center, USA).

SUN-CLIMATE CONNECTIONS 931

significance (Lean and Rind, 1999). The reviews of Lockwood

(2002), Solanki and Fligge (2002), and Haigh (2003) provide

excellent summaries of recent Sun-climate correlations using

observational records. One example is a study of irradiance

correlated with stratospheric temperature by Labitzke and

Matthes (2003; Figure S43).

Some correlations imply a physical mechanism other than

direct radiative forcing. For example, Marsh and Svensmark

(2000) correlated cosmic ray intensity with North Atlantic

regional cloud cover. Cloud cover is a potentially powerful

climate feedback, critically important in defining climate sensi-

tivity to external forcing in models. The authors suggest a

direct link between atmospheric ionization by cosmic rays

and cloud nucleation processes. Sun and Bradley (2002) argue

that these correlations do not hold using updated data sets.

Furthermore, large increases in cosmogenic isotope production

such as the Laschamp event around 40 ka appear unassociated

with climate. Carslaw (2002) has reviewed the mechanisms

of cloud response to cosmic rays in detail.

Paleoclimate and solar variability

Ocean sedimentcores with highaccumulation rates (>20 cm ka

1

)

can provide paleoclimate records with up to multi-centennial

resolution, currently limited by the accuracy of accelerator mass

spectrometer radiocarbon dating. Bond et al. (2001) presented

Holocene records of North Atlantic ice rafted detritus (IRD),

together with

Be flux and

C production rate (Figure S41).

IRD consists of mineral grains once frozen in drifting glacier

or sea ice. Its occurrence in ocean sediment is evidence of

enhanced ice production or survivability, i.e., a cold period.

Though climate change could possibly influence ice core

Be and ocean circulation may influence atmospheric

C, it is

unlikely the agreement between nuclides with different produc-

tion and transport chemistries could be produced independent

of solar variability. The complex patterns of the co-varying

IRD and cosmogenic cycles make a strong argument for solar

influence on North Atlantic climate.

The Asian monsoon is a climate response to intense summer

heating of the Tibetan Plateau, leading to rising of air and low

pressure. This low pressure reverses the direction of prevailing

winds during summer to bring Indian Ocean moisture onshore.

A proxy record for upwelling of deep ocean water from a

sediment core off the coast of Oman indicates that during the

centennial to millennial coolings in the North Atlantic, winds

associated with the summer monsoon were weaker (Gupta et al.,

2003; Figure S44).

Figure S42 Two components of solar irradiance variability, background and the 11-year cycle, are summed to provide an estimate of total solar

irradiance (TSI) since 1610 (data from Lean et al., 1998).

Figure S41 Percentage of hematite-stained quartz grains increases during Holocene cosmogenic isotope maxima, indicating North Atlantic

cooling during solar minima. The most recent cold phase corresponds to the Little Ice Age in Europe (data from Bond et al., 2001).

932 SUN-CLIMATE CONNECTIONS

Carbon isotope indicators of humidity from lake sediments

(Hong et al., 2003) show drier conditions during North Atlantic

coolings. Furthermore, Neff et al. (2001) found that rainfall

intensity was weaker during solar minima based on a record

of oxygen isotopes in Oman cave calcite. These three compo-

nents of monsoon climate: winds, humidity, and rainfall, all

point to a weaker monsoon during periods of reduced solar

magnetic activity. Whether it is a direct response, as would

be expected only for a large irradiance change, or a response

to the solar-influenced climate of the North Atlantic is not yet

known.

Amplification of solar influence

The direct radiative temperature response to solar variability is

extremely small, around 0.06



C for a typical solar cycle irradi-

ance change of 0.1%. A measurable climate response therefore

requires amplifying mechanisms. One of the simplest ampli-

fiers yet proposed may be the climate system responding with

stochastic resonance (Benzi et al., 1982). According to this

model, a small signal (solar forcing) by itself would not pro-

duce climate change, but once added to a background level of

climate system “noise,” the signal crosses a response threshold

and is detectable. One candidate for climate noise may be the

El Niño, which has a recurrence interval (2–7 years) well sui-

ted to amplifying 11-year solar variability. Modeling has shown

that El Niño noise may amplify the solar cycle into decadal

variability of climate-related atmospheric anomalies, such as

the Pacific North American pattern (Ruzmaikin, 1999).

Another possible amplifier is stratospheric ozone. Irradiance

changes in ultraviolet (UV) wavelengths are up to ten times lar-

ger over the course of the solar cycle relative to total irradiance

change (Lean, 2000). The resulting stratospheric heating due to

UV absorption can change ozone concentrations, either ampli-

fying or damping the initial radiative forcing depending on

the altitude of ozone change. Aerosols from volcanic eruptions

may have affected observational data for the last two solar

cycles, so the actual ozone response to the solar cycle is still

unclear (Haigh, 2003).

Shindell et al. (2001) simulated climate of the Maunder

Minimum assuming a TSI reduction of 0.25%. Ozone concen-

trations were allowed to vary in response to the larger UV

reduction of spectrally resolved irradiance. In the model, the

irradiance decrease resulted in Maunder Minimum cooling via

two interacting mechanisms. First, stratospheric ozone changed

such that the initial negative radiative forcing was amplified.

Second, there was a dynamical response resulting in a surface

atmospheric pressure pattern resembling the negative phase of

the Arctic Oscillation/North Atlantic Oscillation (AO/NAO).

The modeled negative AO/NAO was associated with a cooling

of nearly 1–2



C during winter in Northern Europe, consistent

with estimates of the coldest Little Ice Age temperatures. Glob-

ally, modeled cooling was only around 0.35



C, emphasizing

the role of this amplification mechanism in producing regional

climate responses to solar variability.

IRD data of Bond et al. (2001) indicate that for the Little Ice

Age, the most recent of the Holocene millennial cycles, the

North Atlantic did not cool in the typical dipole pattern of the

AO/NAO. Instead, cooling appeared basin-wide. Basin-wide

cooling during the Holocene could be achieved via a reduction

in thermohaline circulation. Thermohaline circulation transports

heat to the North Atlantic region by surface ocean flow from

the tropics. It is sensitive to transport of fresher (lighter) water

into regions of density-driven convection. If solar influence

results in an atmospheric response capable of affecting North

Atlantic THC, the regional climate effect could be substantially

amplified. A persistent negative phase of the AO/NAO may

Figure S43 Stratospheric (70 hPa) temperatures at a point 30



N / 136



W are correlated with solar radio emission, a proxy for irradiance.

The consequences for surface climate depend on uncertain mechanisms of stratosphere-troposphere interaction (adapted from Labitzke

and Matthes, 2003).

Figure S44 Weak monsoon wind strength results in less nutrient-rich

water upwelling off the coast of Oman. This reduces abundance of

G. bulliodes, a hard-shelled plankton whose skeletons may be found in

sediment cores. Periods of weak monsoon winds (shaded) correspond

to North Atlantic coolings (adapted from Gupta et al., 2003).

SUN-CLIMATE CONNECTIONS 933