Odekon M. Encyclopedia of paleoclimatology and ancient environments

Подождите немного. Документ загружается.

to such behavior has become a high priority. Yet after more

than two decades of research, the underlying mechanisms

remain elusive.

Millennial clim ate cycles: history and description

Heinrich events

Heinrich (1988) was the first to recognize that glacial sediments

in a deep-sea core from the Dreizack Seamount (Figure M25)were

punctuated by six distinct layers demarking brief (1,000 years)

episodes of extreme ocean surface cooling, dramatic reductions

in abundances of planktic foraminifera, depletions in planktic

O, and exceptionally large increases in IRD (ice-rafted debris)

produced by melting of armadas of debris-laden icebergs dis-

charged into the subpolar North Atlantic. Later, these six episodes

were named Heinrich events by Broecker et al., (1992).

Building on Heinrich’s discoveries, Bond et al. (1992, 1993)

demonstrated that IRD maxima and the strong depletions in

O in the six Heinrich events (H1–H6) occurred suddenly

at the end of cooling ramps containing bundles of progressively

colder Dansgaard/Oeschger (D/O) cycles (Figure M26a–e).

The cooling ramps were followed abruptly by an unusually

large warming that raised sea surface temperatures by several

degrees Celsius to nearly Holocene values. Cores containing

these features extend from the Labrador Sea eastward across

the entire North Atlantic between latitudes of about 40



N and



N, the so called IRD belt (Figure M25). Petrologic and geo-

chemical tracers in the IRD point to sources of the massive

Heinrich iceberg discharges in eastern Canada, probably in

the Hudson Strait region.

Heinrich events are sometimes defined as corresponding to

the full range of the increases in IRD. Detailed analyses of

the anatomy of Heinrich layers, however, demonstrate that the

initial increase in IRD does not contain tracers linked to

Hudson Strait sources. To avoid confusion over the age of

Heinrich events the term should apply only to layers rich in

IRD derived mainly from eastern Canada. For an excellent

summary of Heinrich events see Hemming (2004) and this

volume (Heinrich events).

Bond cycles

Named by Broecker (1994), a Bond cycle, first identified in

Greenland ice cores and North Atlantic temperature records,

begins with the post-Heinrich event warming, followed by a gra-

dual cooling ramp that contains a bundling of progressively

colder D/O cycles, and then ends abruptly at the peak cooling

of the next Heinrich event (Figure M26a,b). The atmosphere

and ocean surface temperature drop over the course of a

Bond cycle is a few to several degrees Celsius. IRD concentra-

tions in North Atlantic sediments follow this pattern, with

short cycles of progressively larger amplitude, probably correla-

tive with D/O cycles, developing as temperatures decrease

(Figure M26c).

The shapes of Bond cycles were the basis for correlating

temperature records from Greenland ice cores and North

Atlantic deep-sea sediments over the entire last glacial cycle

(Figure M26a,b). The correlation documented for the first time

the relation between D/O cycles in the ice cores and Heinrich

events in the ocean.

D/ O cycles

The typical evolution of these abrupt events, also named by

Broecker (1994), begins with a rapid ocean surface and atmo-

spheric warming, completed usually within a few decades, fol-

lowed by gradual cooling over a few hundred years, and a

distinct cold phase afterwards for a few centuries to a millennium

(Figure M26a). Approximately 23 D/O oscillations occurred dur-

ing the last glacial cycle. The well-known Younger Dryas (YD)

event is the last of these and was followed abruptly by the onset

of the present interglacial or Holocene (Figure M26a). A recent

revision of ice core temperature calibration suggests that the YD

ended with a remarkable atmospheric warming of 10



C in dec-

ades or less (Grachev and Severinghaus, 2005).

Within the D/O stadials that terminate Bond cycles, the

large IRD peaks of Heinrich events are confined to the last few

hundred years, suggesting that the massive discharges of Heinrich

icebergs came after several hundred years of cold temperatures

(Bond et al., 1993). Cold phases of the other D/O cycles also

contain peaks in the concentrations of IRD, although they are

not as large as those during Heinrich events (Figure M26c). The

sources of IRD in D/O cycles are diverse, including Europe, east

Greenland, and Iceland (Hemming, 2004). Percentages of detrital

carbonate remain low during the cold phases of D/O cycles

(Figure M26d), suggesting that, in contrast to Heinrich events,

sources in eastern Canada were not important.

Persistent multi-cent ury to millennial cycles

High-resolution measurements of two IRD grain types,

hematite-stained grains derived from red bed deposits and fresh

volcanic glass derived from Iceland and or Juan Mayen have

revealed robust oscillations that do not follow the D/O cycle

pattern, but instead punctuate both cold and warm phases of

those cycles (Figure M26f). Surprisingly, these persist across

the glacial-interglacial transition and continue through the

entire Holocene (Bond et al., 2001). The pacing of the cycles

is highly irregular, varying from about 900 years to about

2,000 years. The petrologic changes are thought to reflect

repeated southward advections of cooler waters carrying ice

with debris rich in the two grain types from the polar seas north

of Iceland. These shorter cycles are sometimes incorrectly

called Bond cycles.

Figure M25 Locations of cores mentioned in text and location of the

Ruddimann IRD belt (dashed lines).

MILLENNIAL CLIMATE VARIABILITY 569

At least some of these persistent oscillations also occur in

the Greenland ice core records where subtle variations in

O and in estimates of temperature punctuate many of the sta-

dial and interstadial intervals of D/O cycles (North Greenland

Ice Core Project members, 2004; Landais et al., 2004). Promi-

nent depletions in d

O also are present in early Holocene ice at

about 11,300, 10,900, 10,300, 9,300, 8,200 years.

Global distribution of millennial variations

Heinrich events, Bond cycles and D/O cycles have now been

found in climate records far from the North Atlantic where they

were first identified. Voelker (2002) compiled evidence that

revealed their presence in both hemispheres extending from

high polar latitudes to the tropics. The best records, however,

Figure M26 (a) Records from GISP2 ice core, and (b–f) DSDP 609. Npl., N. pachyderma (s.); HSG, hematite-stained grains; H, Heinrich event. Dashed

lines mark boundaries between increases in detrital carbonate (Heinrich events) and precursory events as defined by HSG record. See text for

discussion (modified from Bond et al., 1999).

570 MILLENNIAL CLIMATE VARIABILITY

are from stalagmite deposits, which can be dated accurately

with uranium-thorium methods, and from ice cores whose

chronology is constrained in part by layer counting. A map of

the stalagmite and ice core sites (Figure M27) documents the

global footprint of these three types of millennial variations.

Outside of the North Atlantic the imprint of the persistent

millennial cycles has been found in late Holocene sediments

from the Cariaco Basin off Venezuela, in Holocene lake

records from Ecuador, in glacial-Holocene loess deposits from

the Mississippi Valley, China and Tibet, in tree ring records

from Finland, and in a stalagmite record from Yemen.

Mechanisms

Oscillatory ice sheet dynamics

Although no consensus has been reached on the cause of millen-

nial climate variability, one widely held view maintains that the

fundamental mechanism arises from oscillatory internal dynamics

of glaciers flowing into the northern North Atlantic. Recurring

surges or collapses of those glaciers are thought to have caused

repeated and rapid increases in the flux of icebergs (i.e., fresh-

water) into the ocean. Upon reaching the convecting regions

north and south of Iceland, the freshwater injections are presumed

to have lowered ocean surface salinities enough to force ocean

circulation across a threshold, induce a hysteresis behavior, and

abruptlyreduce or shutdownNorth Atlantic Deep Water (NADW)

formation (Ganopolski and Rahmstorf, 2001). Models of this

process have shown that a shutdown of NADW production trig-

gers a precipitous dropin ocean surface temperature that is broadly

consistent with paleotemperature data.

To explain the oscillatory dynamics, MacAyeal (1993) pro-

posed a free oscillation mechanism that has come to be known

as the binge-purge model. A large glacier or ice sheet frozen to

bedrock will slowly build up during the binge phase. The purge

phase (i.e., the Heinrich event) occurs when geothermal heat

melts the basal sediment and produces a lubricated discharge

pathway. The model produced massive collapses of the ice

every 7,000 years, agreeing reasonably well with the observed

timing of Heinrich events. In a more elaborate development of

the model, Greve and MacAyeal (1996) found that free oscilla-

tions from 1,000 to 10,000 years could occur within a large ice

sheet, thereby potentially providing an explanation for both

Heinrich events and the faster-paced D/O cycles.

Climate-centered mechanisms

The elegant glacier-centered hypothesis was called into ques-

tion, however, by the results of efforts to test two of its corol-

laries. If the hypothesis were correct, the IRD increases in

both Heinrich events and D/O cycles should occur at the

same time as, or slightly lead, the ocean-climate response,

and synchronicity of discharges from different glaciers would

be unlikely owing to the vagaries of internal glacier dynamics.

Instead, it was found in high-resolution North Atlantic records

that the onset of ocean surface coolings actually preceded, by at

least several hundred years, the IRD increases, and that the

IRD in D/O cycles discharged from different glaciers at the

same time (Bond et al., 1999; Hemming, 2004). Moreover,

each Heinrich event appeared to have been immediately pre-

ceded by an increase in IRD (precursory events) whose compo-

sition recorded the same simultaneous discharges of icebergs

from different sources that were found in D/O cycles (e.g.,

Figure M26c,f).

Those findings raised the specter of the chicken or egg

problem. The lead of ocean surface coolings and the synchro-

nous IRD discharges from different glaciers during D/O events

seemed best explained by the effects of a climate-ocean mec-

hanism that operated upon glaciers flowing into the North

Atlantic. The observation of D/O-like IRD discharges immedi-

ately preceding each Heinrich event implied that even Heinrich

events were triggered by the same climate mechanism.

Attempts to identify the climate triggering mechanism have

invoked changes in salinities in the Gulf Stream (Broecker

et al., 1990), stochastic resonance (Alley et al., 2001), collapse

of an ice shelf in the Labrador Sea (Hulbe et al., 2004), and

super El Niños driven by ocean surface salinity changes in

the tropical Pacific (Stott et al., 2002). None of these mechan-

isms, however, has been generally accepted.

The enigmatic 1,470-year cycle

Mayewski et al. (1997) were the first to recognize the presence of a

spectral peak in d

O records from Greenland ice cores centered

on about 1,470 years. The peak is narrow and statistically sig-

nificant, suggesting to them that the cycle is truly periodic. Bond

Figure M27 Locations of ice cores and stalagmite records with

evidence of Heinrich events, Bond cycles, and D / O cycles.

MILLENNIAL CLIMATE VARIABILITY 571

et al., (1997, 1999) found that the pacing of the two petrologic

tracers, hematite-stained grains and Icelandic glass, through the

last glacial cycle had a mean value of 1,470 500 years (1 sigma).

Whether the cycle is truly periodic is unclear due to error in the

radiocarbon-dated time series. Recently, Rahmstorf (2003), using

non-spectral methods, demonstrated that the spacing of D/O

cycles in the GISP2 ice core was exactly 1,470 years or even multi-

ples of 1,470 years, thereby clearly tying the D/O oscillations to

the 1,470-year cycle (Figure M28). From that finding, Rahmstorf

argued that D/O events are discrete events paced by the

1,470-year cycle rather than being discrete cycles themselves.

The cause of the 1,470-year cycle remains a mystery.

Rahmstorf (2003 ) argued that no internal processes in a com-

plex non-linear system such as the Earth’s climate system

could produce such a regular oscillation, a conclusion sup-

ported by results of modeling experiments. He suggested that

the cycle might have been produced by orbital variations within

the Milankovitch band or by solar forcing.

The Holocene dilemma

Once regarded as a long period of climate stability, it is now

clear from ice core, marine, and terrestrial records that the entire

12,000 years of the Holocene is punctuated by a series of robust

cycles of millennial duration. The Holocene dilemma arises from

evidence that, in many records, these cycles have about the

same 1,500-year pacing found in records from the last glacial.

In particular, after extending their analyses of the two petrologic

tracers, hematite-stained grains and Icelandic glass, to the present,

Bond et al. (1999, 2001) found that the Holocene cycle pacing

fell within the same range as documented for the glacial record,

1,470  500-years. In addition, the amplitude of the cycles is

about the same in both the glacial and Holocene intervals. One

interpretation of those findings is that the Holocene and glacial

1,470-year cyclicity was produced by the same mechanism.

Given the absence of large glaciers and ice sheets during most

of the Holocene, the mechanism could not have been linked to

oscillatory ice sheet dynamics. Moreover, the mechanism

underlying the cyclicity must have operated independently of the

glacial-interglacial climate state. Those arguments have been criti-

cized, however, because the Holocene cycles do not have the pre-

cise 1,470-year periodicity of the D/O oscillations (e.g., Schulz

and Paul, 2002).

Braun et al. (2005) have proposed an intriguing solution to

this problem. Building on evidence in Bond et al. (2001)thatthe

Holocene cycles may have been forced by variations in solar irra-

diance, Braun and his colleagues made the remarkable discovery

that their climate system model, CLIMBER-2, when forced by

the DeVries solar cycle (210 years) and the Gleissberg solar cycle

(88 years), can produce robust D/O events with a precise period

of 1,470 years under glacial boundary conditions, but not under

Holocene boundary conditions. Their surprising results have

far-reaching implications for the fundamental mechanisms that

underlie abrupt climate change, and they clearly deserve much

further investigation and rigorous testing.

Summary

Views on what causes millennial climate variability fall into

two camps. One is glacier-centered and assumes that free oscil-

lations in glaciers and ice sheets cause repeated discharges of

icebergs into the ocean. The iceberg freshwater injections force

recurring changes in the North Atlantic’s deep-water circula-

tion that in turn produce a series of robust abrupt climate

changes with a global or nearly global footprint. The other is

climate-centered and assumes that a climate mechanism of

unknown origin forces increases in iceberg discharge either

through direct effects on glacier mass balance or through

changes in the North Atlantic deep circulation. In this case

the glacier activity and its effects on NADW production oper-

ates as a non-linear positive feed back that amplifies the climate

signal and transmits a series of abrupt climate shifts well

beyond the North Atlantic region. Both views have been criti-

cized and neither can fully account for the 1,470-year pacing

of the D/O cycles and the presence of similar cycles in the

Holocene when the large glaciers and ice sheets were absent.

Gerard Bond

Bibliography

Alley, R.B., Anandakrishnan, S., and Jung, P., 2001. Stochastic resonance

in the North Atlantic. Paleoceanography, 16( 2), 190–198.

Bond, G.C., Heinrich, H., Broecker, W., Labeyrie, L., McManus, J.,

Andrews, J., Huon, S., Jantschik, R., Clasen, S., Simet, C.,

Tedesco, K., Klas, M., Bonani, G., and Ivy., S., 1992. Evidence for

massive discharges of icebergs into the North Atlantic ocean during

the last glacial. Nature, 360, 245–249.

Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J.,

and Bonani, G., 1993. Correlations between climate records from

North Atlantic sediments and Greenland ice. Nature, 365, 143–147.

Figure M28 (a) The 1,470-year timing of D / O events from Rahmsdorf (2003). Dashed lines mark out exact 1,470-year intervals. Markers with

solid dots denote onset of D /O cycles as defined by warming above a threshold defined by amplitude and range. (b) Deviation in years of onset of

D / O cycles from an exact 1,470-year pacing (modified from Rahmsdorf, 2003).

572 MILLENNIAL CLIMATE VARIABILITY

Bond, G.C., Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P.,

Priore, P., Cullen, H., Hajdas, I., and Bonani, G., 1997. A pervasive mil-

lennial-scale cycle in North Atlantic Holocene and Glacial climates.

Science, 278, 1257–1266.

Bond, G.C., Showers, W., Elliot, M., Evans, M., Lotti, R., Hajdas, I.,

Bonani, G., and Johnson, S., 1999. The North Atlantic’s1–2 Kyr cli-

mate rhythm: Relation to Heinrich events, Dansgaard/ Oeschger cycles

and the Little Ice Age. In Clark, P.U., Webb, R.S., and Keigwin, L.D.

(eds.), Mechanisms of global climate change at millennial time scales.

Washington, DC: AGU, Geophysical Monograph Series 112, pp.

35–58.

Bond, G.C., Kromer, B., Beer, J., Muscheler, R., Evans, M., Showers, W.,

Hoffmann, S., Lotti-Bond., R., Hajdas, I., and Bonani, G., 2001. Persis-

tent solar influence on North Atlantic climate during the Holocene.

Science, 294, 2130–2136.

Braun, H., Christl, M., Rahmstorf, S., Ganopolski, A., Mangini, A.,

Kubatzki, C., Roth, K., and Kromer, B., 2005. Possible solar origin

of the 1,470-year glacial climate cycle demonstrated in a coupled

model. Nature, 438, 208–211.

Broecker, W.S., 1994. Massive iceberg discharges as triggers for global

climate change. Nature, 372, 421–424.

Broecker, W.S., Bond, G., Klas, M., Bonani, G., and Wolfl, W., 1990.

A salt oscillator in the glacial North Atlantic? 1. The concept. Paleocea-

nography, 5, 469–477.

Broecker, W.S., Bond, G.C., Klas, M., Clark, E., and McManus, J.F., 1992.

Origin of the northern Atlantic’s Heinrich events. Clim. Dynam., 6,

265–273.

Ganopolski, A., and Rahmstorf, S., 2001. Rapid changes of glacial climate

simulated in a coupled climate model. Nature, 409, 153–158.

Grachev, A., and Severinghaus, J., 2005. A revised +10  4



C magnitude

of the abrupt change in Greenland temperature at the Younger Dryas

termination using published GISP2 gas isotope data and air thermal

diffusion constants. Quaternary Sci. Rev., 24, 513–519.

Greve, R., and MacAyeal, D.R., 1996. Dynamic/thermodynamic simula-

tions of Laurentide ice-sheet instability. Ann. Glaciol., 23, 328–335.

Heinrich, H., 1988. Origin and consequences of cyclic ice rafting in the

northeast Atlantic Ocean during the past 130,000 years. Quaternary

Res., 29, 142–152.

Hemming, S.R., 2004. Heirich events: Massive late Pleistocene detritus

layers of the North Atlantic and their global climate imprint. Rev.

Geophys., 42,1–43.

Hulbe, C.L., MacAyeal, D.R., Denton, G.H., Kleman, J., Lowell, T.V.,

2004. Catastrophic ice shelf breakup as the source of Heinrich

event icebergs. Paleoceanography, 19(1). doi: PA1004 10.1029/

2003PA000890.

Landais, A., Barnola, J.M., Masson-Delmotte, V., Jouzel, J., Chappellaz, J.,

Caillon, N., Huber, C., Leuenberger, M., and Johnsen, S.J., 2004. A

continuous record of temperature evolution over a sequence of

Dansgaard-Oeschger events during marine stage 4 (76–62 ky

BP).

Geophys. Res. Lett., 31,1–4, L22211. doi: 10.1029 /2004GL021193.

MacAyeal, D.R., 1993. Binge /purge oscillations of the Laurentide ice

sheet as a cause of the North Atlantic’s Heinrich events. Paleoceano-

graphy, 8(6), 775–784.

Mayewski, P.A., Meeker, L.D., Twickler, M.S., Whitlow, S., Yang, Q.,

Lyons, W.B., and Prentice, M., 1997. Major features and forcing of

high-latitude Northern Hemisphere atmospheric circulation using

a 110,000 year-long glaciochemical series. J. Geophys. Res., 102,

26345–26366.

North Greenland Ice Core Project Members, 2004. High-resolution

record of northern hemisphere climate extending into the last intergla-

cial period. Nature, 431, 147–157.

Rahmstorf, S., 2003. Timing of abrupt climate change: A precise clock.

Geophys. Res. Lett., 30(10), 1510. doi: 10,1029/2003GL.017115.

Schulz, M., and Paul, A., 2002. Holocene climate variability on centennial-

to-millennial time scales 1. Climate records from the North-Atlantic

realm. In Wefer, G., Berger, W.H., Behre, K.-E. and Jansen, E. (eds.),

Climate Development and History of the North Atlantic Realm. Berlin:

Springer, pp. 41–54.

Stott, L., Poulsen, C., Lund, S., and Thunell, R., 2002. Super ENSO

and global climate oscillations at millennial time scales. Science, 297,

222–226.

Voelker, A.H.L., 2002. Global distribution of centennial-scale records

for marine isotope stage (MIS) 3: A database. Quaternary Sci. Rev.,

21(10), 1185–1212.

Cross-references

Binge-Purge Cycles of Ice Sheet Dynamics

Dansgaard-Oeschger Cycles

Heinrich Events

Ice-Rafted Debris (IRD)

Quaternary Climate Transitions and Cycles

Sun-Climate Connections

Younger Dryas

MINERAL INDICATORS OF PAST CLIMATES

Introduction

A mineral is a naturally occurring crystalline solid created by

geological or biogenic processes (e.g., calcite/aragonite in mol-

lusk shells or coral). Minerals that form at or near the Earth’ssur-

face are products of chemical weathering, evaporation, authigenic

crystallization, and bio-mineralization. They reflect ambient con-

ditions at the Earth-atmosphere interface. Therefore they can fur-

nish important clues about former climates. Minerals are utilized

as paleoclimate indicators or proxies in several different ways:

1. To infer past climates and changes over time.

2. To deduce changes in atmospheric composition.

3. As mineralogical “markers” in provenance studies.

4. As “hosts” for climate-sensitive stable isotopes and trace

elements.

The most direct mineral indicators or proxies are those that

are generated under relatively narrow climatic ranges or within

restricted environmental settings. Examples include chemical

precipitates such as evaporites, low temperature minerals such

as ikaite and hydrohalite, residual soils (iron and aluminum

oxyhydroxides, kaolinite), and authigenic minerals (forming

in situ, e.g., glauconite, bertherine, phillipsite, barite). Evapor-

ite minerals (e.g., gypsum, anhydrite, halite, sylvite) are gener-

ally indicative of arid to hyperarid climates. They form in

restricted marine basins, narrow rift valleys, coastal lagoons,

sabkhas (salt flats), and internally-drained continental interiors,

where rates of evaporation exceed rates of water supplied by

rivers or rainfall (McLane, 1995; Parrish, 1998; Evaporites).

The distribution of modern evaporites closely corresponds

to the distribution of deserts. In addition, diagnostic of aridity

are chemically-resistant minerals like quartz that are concen-

trated into eolian sand deposits (and ultimately sandstones) by

wind transport and deposition. Sedimentary rocks of biogenic

origin (e.g., chalks, cherts, phosphorites) provide information

on paleotemperatures, productivity, and ocean circulation.

Clay minerals derived from chemical weathering of rocks are

relevant climate indicators, since they denote prevailing condi-

tions at the time of formation. Kaolinite, for example, although

widespread in many soils, is characteristic of tropical weather-

ing, whereas chlorite is usually more abundant in colder

regimes, where physical weathering predominates. Changes in

composition of marine or lacustrine clays often parallel climate

fluctuations (e.g., Mallinson et al., 2003; Robert and Kennett,

1997; Yuretich et al., 1999, see below).

Minerals that are sensitive to oxidation may provide infor-

mation about past atmospheres. Pyrite and uraninite oxidize

and alter rapidly under current atmospheric conditions. Their

preservation in detrital grains from 3 billion year old sediments

implies much lower atmospheric oxygen than at present (Fleet,

MINERAL INDICATORS OF PAST CLIMATES 573

1998; Kirk et al., 2003). The abundance of banded iron forma-

tions (BIFs) older than 2 billion years also suggests reduced

oxygen levels at that time. The accumulation of large BIF

deposits required transport of dissolved ferrous iron over long

distances, followed by precipitation in locally, more oxyge-

nated environments (see Banded iron formations and the early

atmosphere; Atmospheric evolution, Earth; McLane, 1995).

Minerals act as markers of provenance in paleoclimate studies.

For example, changing pyroxene to amphibole ratios in late

Quaternary Nile delta sediments point not only to variations in

the relative contributions from the Blue and White Nile tributaries,

but also to oscillations in rainfall in the two source areas that

are linked to larger-scale climatic changes (Foucault and Stanley,

1989). An important clay mineral tracer is the latitudinally-

sensitive kaolinite/chlorite ratio, one of several parameters used

to establish an eastern Asian desert source for ice-age dust in

the GISP2 (Greenland) ice cores (e.g., Biscaye et al., 1997).

Stable isotopes and trace elements housed in minerals

furnish quantitative paleoclimate data. Oxygen isotope ratios

in foraminiferal calcite are related to seawater temperatures

and global ice volume (see Stable isotope analysis). Oxygen

and carbon isotope ratio fluctuations in speleothem calcite

reflect variability in atmospheric circulation, precipitation pat-

terns, and vegetation cover (Genty et al., 2003; Speleothems).

Oxygen isotope ratios in marine barite have been attributed to

altered redox conditions, changes in continental shelf area,

and sea level during the Plio-Pleistocene (Turchyn and Schrag,

2004). Magnesium/calcium and strontium/calcium ratios in

foraminiferal calcite are also important paleo-ocean temperature

indicators.

Mineral paleoclimate indicators are briefly described below

and summarized in Table M1.

Ice

Ice (H

O), the solid form of water at or below 0



C and under

normal atmospheric pressure, occurs in massive accumulations

in polar ice sheets, and also in mountain glaciers, ice shelves,

and sea ice (see Cryosphere). The lower density of ice

(0.92 g cm

–3

) relative to water results from a relatively open

crystal lattice, due to H-bonding of H

O molecules. Ice,

although solid, deforms readily under the force of gravity,

which sets glaciers into motion.

Because the temperature of the water-ice phase transition

lies within the diurnal or seasonal range of many regions, ice

is sensitive to climate change on many time scales. The extent

of the cryosphere has varied markedly during Quaternary

glacial-interglacial cycles. The growth and decay of polar ice

sheets can be reconstructed from the distribution of geomor-

phological features and marine oxygen isotope ratios that

record paleosea-level variations (see Glacial geomorphology,

Stable isotope analysis). Ice cores from Greenland, Antarctica,

and mountain glaciers contain important archives of past tem-

peratures and atmospheric trace gases (see Ice cores, Antarctica

and Greenland; Ice cores, mountain glaciers).

Iron oxides and oxyhydroxides

Hematite

Hematite (a-Fe

) is a common constituent of sedimentary rocks

formed under oxidizing conditions. Hematite occurs in Phanero-

zoic ironstones, iron-rich rocks that also contain goethite, berthier-

ine, and siderite. Ironstones may be deposited in lakes, bogs or

coal swamps, or on shallow marine shelves or under semi-arid

or seasonally-alternating wet/dry climates (Parrish, 1998). Hema-

tite coats quartz and clay grains, and cements detrital particles

in continental red beds (iron-stained sandstones, siltstones, and

shales). These rocks occur in a variety of geologic settings, includ-

ing desert sandstones, evaporitic tidal flats, lagoons, and savanna

soils. Ferric iron, initially deposited as goethite, gradually trans-

forms into hematite, intensifying the reddish color. The chemical

reaction is enhanced by increasing temperature, alkalinity, and

dehydration. Although once believed to be an indicator of

arid climates, red beds are now thought to develop also in warm

Table M1 Mineral indicators

Mineral

Paleoclimatic and/or paleo-environmental

significance

Oxides, hydroxyoxides

Ice Cryosphere volume, glacial-interglacial cycles

Hematite Oxic conditions; tropical/ semi-tropical/ temperate

weathering

Goethite Oxic conditions; tropical /semi-tropical/ temperate

weathering

Lepidocrocite Oxic conditions; tropical/semi-tropical /temperate

weathering

Maghemite Magnetostratigraphic dating

Magnetite Magnetostratigraphic dating

Gibbsite Component of bauxite; humid, tropical chemical

weathering

Boehmite Component of bauxite; humid, tropical chemical

weathering

Birnessite Component of desert varnish; aridity

Sulfides

Pyrite Reducing (anoxic) conditions; low atmospheric O

(as detrital grains)

Halides

Halite Evaporite; aridity

Sylvite Evaporite; aridity

Hydrohalite Evaporite; low temperature, hypersalinity

Antarcticite (rare) Evaporite; low temperature, hypersalinity

Carbonates

Calcite Paleotemperatures (oxygen isotope, trace element

ratios),”calcite seas – greenhouse climates”

Aragonite Like calcite;”Aragonite seas – Icehouse climates”

Dolomite Arid, hot-saline lakes, sabkhas

Siderite Reducing (suboxic-anoxic) conditions

Ikaite (glendonite) Low temperature, near-freezing

Sulfates, phosphates

Gypsum Evaporite; aridity

Anhydrite Evaporite; aridity

Mirabilite Evaporite; aridity, hypersalinity, low temperature

Barite Ocean paleoproductivity

Francolite Ocean paleoproductivity

Silica minerals

Quartz Eolian sand, sandstones – aridity; eolian dust

(marine sediments, ice cores) – wind strengths,

direction.

Opal, chert Ocean paleoproductivity

Clay minerals

Kaolinite Tropical/semi-tropical chemical weathering

Chlorite High latitude physical weathering

Sepiolite Warm, semi-arid to arid climates

Palygorskite Warm, semi-arid to arid climates

Glauconite Weakly reducing (suboxic) conditions, low

sedimentation rates

Berthierine Reducing environments

Zeolites

Analcime Desert soils, alkaline lakes–aridity

Chabazite Alkaline lakes; aridity

Phillipsite Alkaline lakes; aridity

Clinoptilolite Alkaline lakes; aridity

574 MINERAL INDICATORS OF PAST CLIMATES

climates with seasonal precipitation (Parrish, 1998; see also Red

beds,thisvolume).

Hematite also concentrates within nodular horizons in

ferricrete – an iron-rich lateritic duricrust (also termed “cuirasse”)

that is a product of intense chemical weathering in semi-

tropical to tropical climates, with seasonally varying rainfall

(T ¼ 25–30



C, P ¼ 1,500 mm yr

–1

,RH¼ 65% (Tardy,

1992). Most silicates and even resistant minerals such as quartz

are leached out, enriching the soil in residual iron oxides

(hematite, goethite), and minor kaolinite.

Hematite-stained quartz and feldspar grains derived from

sedimentary red beds have been used as lithologic tracers for

late Quaternary to Holocene North Atlantic ice-rafted debris

(IRD) (Bond et al., 1997). The hematite-stained grains were

probably derived from eastern Greenland and Spitsbergen, in

contrast to volcanic glass – another lithic tracer – from Iceland

or Jan Mayen. Peaks in the record of hematite-coated grains

provide a reliable and consistent proxy for estimating the

pacing of IRD events.

Hematite, magnetite (Fe

) and siderite (FeCO

) alternate

with bands of chert (SiO

) in early Proterozoic banded iron for-

mations (BIFs). The accumulation of large BIF deposits

implies the transport of soluble iron (as Fe

), likely derived

from submarine volcanic sources, over long distances. Upwel-

ling subsequently brought dissolved iron onto locally more

oxygenated, shallow marine shelves, where oxidation of iron

and precipitation occurred (McLane, 1995). Since most BIFs

are older than 2 billion years, the availability of ferrous iron

implies lower atmospheric oxygen levels than at present.

Maghemite

Maghemite (g-Fe

) is a magnetic polymorph of hematite

(with the same chemical composition, but a different crystal

structure). It is a widespread, although minor component of

paleosols and loess. Along with magnetite, it carries the mag-

netic signal employed in magnetostratigraphic dating.

Goethite

Goethite (a-FeOOH) is an oxyhydroxide, widely distributed in

sedimentary rocks, including ironstones, red beds, and also found

in lateritic soils. Upon burial, it dehydrates and inverts to hematite.

Lepidocrocite (g-FeOOH), a polymorph of goethite, is also com-

mon is soils. Ferrihydrite (Fe

·2FeOOH·2.6·H

O) is a poorly

crystallized iron oxyhydroxide found in recent soils, sediments,

and iron concretions. Since it is metastable, it slowly dehydrates

and recrystallizes to goethite or hematite (Evans, 1992).

Magnetite

Magnetite (Fe

), a di- and tri-valent iron oxide, forms under

somewhat more reducing conditions than hematite, in the

absence of carbonate or sulfide (McLane, 1995). Magnetite,

along with maghemite is the dominant source of magnetic sig-

nal in soils or sediments (Maher, 1998). Magnetic susceptibility

(the ratio of induced magnetization to the inducing magnetic

field) is a measure of the ability of a substance to be magne-

tized. It is directly proportional to the quantity of strongly mag-

netic minerals (i.e., magnetite and maghemite) present in loess,

paleosols, and marine sediments.

Magnetite formation and soil magnetization increase in well-

drained soils overlying Fe-rich source rocks that are exposed to

seasonally-varying wet and dry cycles. On the other hand, exces-

sively arid, wet, or acid soils show little magnetic enhancement.

Furthermore, magnetically-enhanced soils show some correlation

between maximum soil susceptibility and annual precipitation

(Maher, 1998). Thus, paleo-susceptibility values in carefully-

selected soil or loess samples, as in the Loess Plateau, China, pro-

vide a useful proxy of annual rainfall, and potentially of changes in

monsoonal patterns and atmospheric circulation.

The Earth’s magnetic field has reversed episodically over

time, the last major reversal (Matuyama-Brunhes) occurring

780,000 yr

BP. The timing of changes in direction of magnetiza-

tion of rocks over time has therefore been used as a dating tool

in paleoclimatology (see Dating, magnetostratigraphy). The

time-varying magnetic properties of eolian loess deposits in a

stratigraphic sequence also record a high-resolution history of

climate change, including precipitation patterns. Paleoclimate

studies show that paleosols formed under relatively warm,

humid, interglacial conditions are characterized by relatively

high magnetic susceptibilities that correlate with high illite/

chlorite ratios in the southern Loess Plateau, China, signifying

more intense chemical weathering (Zhao et al., 2005) and with

more negative oxygen isotope ratios in deep sea sediments

(Reynolds and King, 1995). The reverse is found during cold,

dry glacial periods.

Aluminum oxides and oxyhydroxides

Gibbsite (Al(OH)

) and boehmite (AlOOH) are major constitu-

ents of bauxite. Bauxite (Figure M29) is a residual, lateritic soil

that is formed by intense chemical weathering of rocks under

wet, tropical climates with annual mean precipitation above

1,700 mm yr

–1

, whereas ferricretes (see above) develop under

hot, but seasonally-variable precipitation ranging between

1,300–1,700 mm yr

–1

(Tardy and Roquin, 1992). However,

others have suggested that both bauxites and laterites were gen-

erated under similar tropical climates, with one or more dry

Figure M29 Bauxite with pisolitic texture, Arkansas (6.4 cm  6.4 cm).

Bauxite is composed of a mixture of aluminum oxides and

oxyhydroxides, resulting from intense chemical weathering under wet,

tropical climates (specimen, V. Gornitz; photograph, John Betts).

MINERAL INDICATORS OF PAST CLIMATES 575

months (Parrish, 1998). Under these climate regimes, most sili-

cates and even resistant minerals such as quartz are destroyed,

leaving insoluble aluminum minerals, such as gibbsite and

boehmite, with lesser quantities of diaspore (AlOOH), kaolinite,

hematite, and goethite.

The geographic distribution of bauxites has changed over

time, reflecting not only changing climates, but also the north-

erly motion of the African and American plates since the open-

ing of the Atlantic Ocean in the Jurassic period. Quaternary to

Recent bauxites in Africa and South America are nearly equa-

torial, whereas pre-Quaternary bauxites on both continents tend

to lie to the northeast (and also south) of their present positions

(Tardy and Roquin, 1992).

Manganese oxides and oxyhydroxides

Manganese minerals occur in diverse environments, including

terrigenous soils, rock varnish, and oceanic nodules. Among

the more common phases in soils and sediments are pyrolu-

site (MnO

), manganite (MnO·OH), vernadite (d-MnO

)or

((Mn,Fe,Ca,Na)(O,OH)

·nH

O), birnessite ((Na,Ca)

0.5

(Mn

)

·1.5H

O), hollandite (Ba

(Mn

,Mn

)

), and

todorokite (Mn

,Ca,Mg)Mn

·H

O. Manganese (and iron)

oxides are concentrated in lateritic soils, ferricretes, and duri-

crusts (Tardy and Roquin, 1992). Iron-manganese nodules

are also widespread in clay-rich soils of temperate to warm

climates (Stiles et al., 2001). Initially depositing as hydrous

phases, such as birnessite, these eventually dehydrate and

invert to anhydrous phases such as hollandite during diagen-

esis. Total Fe content of Fe-Mn nodules in both modern

and paleo-Vertisols has been used as a proxy for mean annual

precipitation, showing a positive correlation (whereas total

Mn varies inversely) with this climate parameter (Stiles

et al., 2001).

Exposed rock surfaces in desert environments are frequently

coated by thin laminations of manganese and iron oxides.

Desert (or rock) varnish consists of around 70% clays and

30% Fe and Mn oxides, mainly hematite, birnessite, and lesser

todorokite (McKeown and Post, 2001). Rock varnish is pro-

duced by oxidizing bacteria that extract iron, manganese, and

other metals from airborne dust, and precipitation. The manga-

nese content of varnish increases with mean annual precipita-

tion, in contrast to that of soil Fe-Mn nodules. Accurate

radiometric dating and chemical analyses of varnish lamina-

tions show promise as a paleoclimate proxy (Broecker and

Liu, 2001; see also Desert varnish as a climate proxy).

Oceanic manganese nodules are rounded concretions that

develop in areas of low sediment accumulation rates on the

deep seafloor. The nodules are a complex intergrowth of poorly

crystalline hydrous manganese and iron phases, including bir-

nessite, todorokite, and vernadite (Li and Schoonmaker,

2003). Most nodules have accreted on nuclei of sharks’ teeth,

fish bone, shell or rock fragments.

Iron sulfides

Pyrite

Since iron can exist in two different oxidation states, iron

minerals are potentially sensitive indicators of soil or sediment

redox conditions, which are linked to biogeochemical cycles,

and thus indirectly to climate. Pyrite (FeS

), in particular,

has been used an indicator of soil/sediment redox state.

Decomposition of organic matter in sediments creates a locally

anaerobic environment, in which bacterial reduction of sulfates

to hydrogen sulfide (H

S) takes place. The H

S reacts in turn

with dissolved iron to precipitate pyrite or marcasite. Marcasite

(FeS

) is less stable in acidic solutions than pyrite, and is there-

fore less abundant in the geologic record. Deposition of pyrite

is favored by euxinic lacustrine or marine environments

(restricted water circulation), presence of organic-rich sedi-

ments, locally elevated CH

levels, and (on land), poorly-

drained or water-logged soils. The degree of pyritization

(DOP)

has frequently been taken as a marker for soil/sediment

redox state. DOP values <0.46 represent aerobic conditions,

with 0.46–0.75 intermediate, and DOP > 0.75 euxinic condi-

tions (Raiswell et al., 1988). However, DOP may also depend

on iron availability. Furthermore, in certain shallow water

environments, such as tidal salt marshes, DOP values may

reflect pore water redox state and sediment chemistry rather

than bottom water oxygen levels (Roychoudhury et al., 2003).

Pyrite is rapidly oxidized in today’s atmosphere and thus

generally does not occur as a detrital mineral. Therefore,

well-rounded “detrital” grains of pyrite, also uraninite (UO

and siderite (FeCO

), from the Witwatersrand basin in South

Africa have been widely regarded as evidence for much lower

than present atmospheric oxygen levels during the Archean eon

(>2,500 million y

BP). The significance of these grains as

paleo-atmosphere indicators has been controversial, since a

hydrothermal origin was proposed for “placer” gold nuggets

associated with these minerals in conglomerates. However,

several recent studies strengthen the case for the placer theory.

Precise

187

Re/

187

Os radiometric dating shows that the rounded

gold nuggets (and associated pyrite and uraninite grains)

are 3 billion years old as compared to significantly younger

ages of 2.76–2.89 billion years for the enclosing host conglom-

erates, and that therefore the heavy minerals must be detrital

(Kirk et al., 2003). Textural and isotopic evidence also favors

the placer theory. Rounded, As-bearing pyrite grains with trun-

cated compositional growth bands clearly demonstrate that the

grains were mechanically abraded (Fleet, 1998). The rounded

pyrite exhibits a broader range in dS

values, suggestive of a

placer origin, in contrast to crystalline pyrite overgrowths,

which probably were produced by later hydrothermal alteration

or metamorphism (England et al., 2002).

Halides

Halite

Halite (NaCl) crystallizes from a supersaturated brine at the air-

water interface as thin square plates, which grow faster at their

edges. These expand into pyramidal crystals with hollowed

centers (hoppers), and eventually coalesce into floating rafts

and sink. The hoppered crystals continue to grow on the basin

floor, ultimately creating a distinctive layered, chevron-like pat-

tern in sedimentary salt deposits (Figure M30). In addition to

precipitating directly from brine, halite can also grow within

soft mud or silt into large displacive cube crystals, often with

hollow faces (Figure M31; Gornitz and Schreiber, 1981).

Displacive halite is well-known from the geologic record. The

presence of such crystals in sedimentary rocks suggests an

origin within shallow marine sediments or from the intertidal

to supratidal zone, in fairly hot, arid climates (Gornitz and

Schreiber, 1981).

DOP = %Fe

pyrite

/(%Fe

pyrite

+ reactive Fe), where Fe

pyrite

is the fraction of

iron associated with pyrite, while reactive Fe is the fraction of iron that

reacts with dissolved sulfur.

576 MINERAL INDICATORS OF PAST CLIMATES

Halite crystals often trap droplets of the brine from which

they precipitate. Such fluid inclusions, if primary (i.e., trapped

at the time of growth), can supply useful information about

the temperature and composition of the brine. For example,

primary fluid inclusions in “chevron”-type halite from shallow

ephemeral saline lakes or salt pans from the Permian Nippewalla

Group, western Kansas yield homogenization temperatures in

the range 21–50



C. (Benison and Goldstein, 1999). Since these

shallow saline pools were exposed to the surface, these values

closely reflect atmospheric temperatures and support other

paleogeographic evidence pointing to a hot, arid climate of

the mid-North American continent during the mid-late Per-

mian. Primary inclusions in marine chevron halite have also

been used to track major oscillations in Phanerozoic paleosea

water chemistry (Lowenstein et al., 2001). Mg

/Ca

ratios of

Phanerozoic seawater derived from the fluid inclusions show

systematic variations over time, with peak Mg

/Ca

ratios

occurring during the late Precambrian, Permian, and Neogene.

These periods of high Mg

/Ca

values were associated with

periods of enhanced aragonite (CaCO

) and MgSO

deposition,

as well as slower seafloor spreading rates.

Other halides

Sylvite (KCl) is an evaporitic salt, often associated with halite,

which forms under similar conditions of high aridity. Hydroha-

lite (NaCl·2H

O) is a form of salt that precipitates from highly

saturated brines at low temperatures (Roberts et al., 1997).

Because of its solubility, it is rarely preserved in the geologic

record, but its former presence may be inferred from pseudo-

morphs of halite or other minerals. Antarcticite (CaCl

·6H

O),

is a rare soluble chloride, discovered in saline lakes from the

dry valleys of Antarctica (Torii and Ossaka, 1965). It precipi-

tates from saturated brines at below-freezing temperatures

(see also Evaporites).

Carbonates

Calcite, aragonite, and dolomite

Calcite and aragonite are the common polymorphs of CaCO

(Vaterite is a much rarer form of CaCO

). Aragonite is less

stable than calcite, but is stabilized by magnesium in seawater

(McLane, 1995). High Mg-calcite, a variety of calcite, may

contain up to 10–18 mol-% MgCO

. In dolomite (CaMg

(CO

)

), the molar proportion of magnesium to calcium is

1:1. Calcite, aragonite, and dolomite occur over a broad range

of geologic settings, including karst terrains, arid soils, lacus-

trine, coastal, and shallow to deep-water marine environments.

These minerals can form as inorganic chemical precipitates and

through biogenic processes.

Figure M30 Chevron texture in bedded halite from the mid-Permian

Nippewalla Group, western Kansas formed by precipitation of rafts

of hoppered halite in supersaturated brines (photograph, Kathy

Benison, Central Michigan University; reproduced with permission).

Figure M31 (a) Displacive halite, hopper (hollow) cube with mud inclusions, Dead Sea, Israel (6.3 cm  6.3 cm). (b) Displacive halite, Dead

Sea, Israel (2.5 cm  3.2 cm). Mud inclusions have been incorporated into the halite along crystallographic directions during growth,

forming a Maltese cross pattern (specimens, V. Gornitz; photographs, John Betts).

MINERAL INDICATORS OF PAST CLIMATES 577

Calcite and, to a lesser extent, aragonite are the most common

minerals in cave speleothems (i.e., stalactites and stalagmites).

The concentric, layered rings in speleothems record annual cli-

matic or environmental variations, much like tree rings. Thus,

trace elements or oxygen and carbon isotope ratios in spe-

leothems have been used as paleoclimate proxies (e.g., Genty

et al., 2003; Speleothems). In arid to semi-arid climates, calcite

and dolomite concentrate within hardpan soil layers, forming

calcrete (or “caliche”), a type of duricrust (Milnes, 1992; Duri-

crusts). Calcite and aragonite constitute the bulk of carbonate-

rich eolianites, which accumulate as wind-blown dunes adjacent

to mid-latitude beaches (see Eolianite). Eolianites develop along

dry, low relief coasts, with high wave and wind energy. Mg-

calcite and aragonite are characteristic cements of beachrock,

a clastic rock that is lithified within the intertidal zone of tropi-

cal to subtropical coasts (McLane, 1995; see Beachrock).

Beachrock has been used as a paleo-sealevel indicator.

Calcite is the dominant bio-mineral of deep-sea calcareous

oozes or chalk, which consists mainly of skeletal coccolitho-

phores and foraminifera. Aragonite is derived from pelagic

pteropods. Modern shallow-water carbonates, largely confined

to subtropical and tropical climates, are dominated by aragonite

and high Mg-calcite (Morse, 2003). Aragonite is produced by

scleractinian coral, calcareous green algae, and some mollusks,

whereas high Mg-calcite is created by benthic foraminifera and

sea urchins. Oolitic carbonates (composed of aragonite or high

Mg-calcite) are generally non-biogenic (Morse, 2003).

A number of trace elements, such as magnesium, strontium,

or barium are incorporated into marine carbonates. Mg/Ca and

Sr/Ca ratios in foraminiferal calcite serve as paleotemperature

proxies, and Cd/Ca and Ba/Ca ratios as proxies of ocean

paleoproductivity and circulation (Kastner, 1999; Parrish,

1998). Oxygen and carbon isotopes in foraminifer calcite yield

information on ocean paleotemperatures, polar ice volume, and

the carbon budget (Parrish, 1998). Coccoliths have been used

for biostratigraphic dating of marine sediments.

The abundance of biogenic carbonates has varied consider-

ably during the Phanerozoic eon. Periods of calcite dominance

(“calcite seas”) from the Ordovician through Devonian periods

and Jurassic-Cretaceous have alternated with periods of arago-

nite dominance (“aragonite seas”) during the Carboniferous-

Triassic and more recently, the Neogene (Montañez, 2002).

Calcitic corals and stromatoporoids dominated the reefs of

calcite seas, in contrast to the aragonitic sponges, corals, algae,

and Mg-calcite red algae that occupied the reefs of aragonite

seas. Modern reefs are populated by aragonitic scleractinian

corals and Mg-calcite coralline algae. The calcite-dominant

periods are closely associated with “greenhouse” climates,

whereas the aragonite-rich periods are linked to “icehouse”

conditions. These temporal variations in marine bio-mineralogy

have been attributed to changes in seawater Mg/Ca ratios,

shifts in long-term rates of sea-floor spreading, eustatic fluctua-

tions, and climate cycles. Periods when the seawater Mg/Ca

mole ratio exceeds two favor the production of high Mg-calcite

and aragonite, and vice-versa (Lowenstein et al., 2001).

Although dolomite occurs massive in sedimentary rocks, it

forms at present in restricted environments such as highly sal-

ine lakes or sabkhas. These are coastal salt flats, or shallow,

tidal pools in hot, arid regions (e.g., the Persian Gulf, also

around the Caribbean, Southern California) that are periodi-

cally flooded by seawater and concentrated by evaporation.

Gypsum (sometimes anhydrite) precipitates first, leaving the

residual brine enriched in magnesium, which then reacts with

calcium carbonate sediments, resulting in the deposition of

dolomite (McLane, 1995).

Ikaite (glendonite)

Ikaite (CaCO

·6H

O) is a rare mineral that is stable only at

very low temperatures and several kilobars pressure. It occurs

in nature at temperatures up to 7



C in alkaline, phosphatic

marine and continental waters (De Lurio and Frakes, 1999).

Although ikaite has generally been replaced by calcite (“glen-

donite”), its original crystal habit of individual bipyramids or

radiating clusters has been preserved in the sediments

(Figure M32; Swainson and Hammond, 2001). Because of

the rather restricted conditions under which ikaite/glendonite

forms, it has been used as a proxy for near-freezing water tem-

peratures (De Lurio and Frakes, 1999).

Siderite

Siderite (FeCO

) occurs in minor amounts in reducing, low sul-

fide, lacustrine, deltaic environments (McLane, 1995), and on

marine continental margins, where it forms during late diagenesis

in suboxic to anoxic pore fluid waters near a source of detrital Fe

(Kastner, 1999). Sphaerosiderite is a distinctive millimeter-sized,

spherulitic variety of siderite found in silty mudstones from redu-

cing soils in continental wetlands (Ludvigson et al., 1998).

The d

O ratios of sphaerosiderite and co-existing calcite have

been used to estimate paleotemperature and ground-water d

values (Ludvigson et al., 1998). Careful selection of unaltered

samples is necessary to avoid errors in the paleothermometry

introduced by diagenetic alteration (Ludvigson et al., 2000).

Siderite figures prominently in the debate surrounding the

composition of the Earth’s early atmosphere. The existence of

ancient oceans presumes that the reduced luminosity of the

Figure M32 Glendonite (calcite pseudomorph after ikaite), Olenitsa

River, Kola Peninsula, Russia, (2.5 cm) (specimen, V. Gornitz;

photograph, John Betts). Ikaite, or glendonite, is an indicator of

near-freezing water temperatures.

578 MINERAL INDICATORS OF PAST CLIMATES