no climate paradox under the faint young sun - rosing et al

7/28/2019 No Climate Paradox Under the Faint Young Sun - Rosing Et Al

1/6

LETTERS

No climate paradox under the faint early SunMinik T. Rosing1,2,4, Dennis K. Bird1,4, Norman H. Sleep5 & Christian J. Bjerrum1,3

Environmental niches in which life first emerged and later evolvedon the Earth have undergone dramatic changes in response toevolving tectonic/geochemical cycles and to biologic interventions13,as well as increases in the Suns luminosity of about 25 to 30 per centover the Earths history4. It has been inferred that the greenhouseeffect of atmospheric CO2 and/or CH4 compensated for the lowersolar luminosity and dictated an Archaean climate in which liquidwater was stable in the hydrosphere58. Here we demonstrate,however, that the mineralogy of Archaean sediments, particularlythe ubiquitous presence of mixed-valence Fe(IIIII) oxides (mag-netite) in banded iron formations9 is inconsistent with such highconcentrations of greenhouse gases and the metabolic constraintsof extant methanogens.Promptedby this,and theabsence of geologicevidence for very high greenhouse-gas concentrations1013, we hypo-thesize that a lower albedo on the Earth, owing to considerably lesscontinental area and to the lack of biologically induced cloud con-densation nuclei14, made an important contribution to moderatingsurface temperature in the Archaean eon. Our model calculationssuggest that the lower albedo of the early Earth provided environ-mental conditions above the freezing point of water, thus alleviatingthe need for extreme greenhouse-gas concentrations to satisfy thefaint early Sun paradox.

The Earths surface environment over the approximately 4 billionyears (Gyr) recorded in geologic formations appears to have beenmaintained within a relatively narrow range in which liquid waterwas stable. This is surprising because the factors that determine sur-face temperature have evolved owing to temporal variations of theSuns irradiance, the Earths albedo and cloud cover, and concentra-tions of atmospheric greenhouse gases over geologic time. It is notreadily apparent to what extent this apparent thermostasis can beattributed to physico-chemical feedback mechanisms, metabolicinterventions from living organisms, or combinations of unrelatedsecular changes.

A 2530% lower solar luminosity from the early Sun inferred fromstandard stellar evolution trends4 should have caused a 2126 K lowerradiative equilibrium temperature for the early Earth, with a planetaryalbedo equal to the present value5,7. If we also assume that the atmo-spherehad a transmittance similarto thepresent, then theaverageearlyArchaean surface temperature would have been below the freezingpoint of water. Thecontrast betweengeologic evidencefor thepresenceofliquid wateron theEarths surface in the deep past15,16 withthe lowerirradiance from a faintearly Sunhas been cast as a paradox, promptinghypotheses58,17 of stronger greenhouse effects from high partial pres-sures of CO2, NH4, CH4, and/or C2H6. Secular changes in the contri-bution from clouds to the Earths planetary albedo have also beenproposed as important parameters controlling climate on the earlyEarth18,19.

Extreme concentrations of greenhouse gases would be required tosupport marine temperatures of 706 15 uC, as suggestedfromoxygenisotopes of cherts 3.53.2 Gyrold and the assumption that the oxygen

isotope composition of Archaean sea water was the same as atpresent20,21. Suchhigh temperatures and requiredgreenhouse gas con-centrations, as well as the consistency of the oxygen isotope composi-tion of sea water over geologic time, are not universally accepted 22,23,and would have required a partial pressure of atmospheric CO2 ofpCO2


2/6

and Fe(III) compounds would break down owing to reduction ofFe(III). This oxidizes the environment by consuming H2 throughreactions of the type:

Fe2O3 [inminerals]1H2 [gas]52FeO [inminerals andsolutions]1H2O (1)

Likewise, increasing CO2 wouldfavour formationof siderite (FeCO3)relative to Fe(II) silicates, and Fe(III)-containing minerals such asmagnetite (FeOFe2O3) would react to form siderite while consuminghydrogen:

FeOFe2O3 [inminerals]1 3CO2 [gas]1H2 [gas]5

3FeCO3 [inminerals]1H2O (2)

Ferric iron is present in the Earths mantle and in mantle-derivedmelts where equilibria involving ferrous iron silicates and Fe(III)-richspinels (nominally magnetite) define the intrinsic oxidation state ofthe Earths mantle and crust. Ferrous iron silicates can be oxidized bywater to form magnetite while releasing aqueous silica and hydrogen,which may escape to the oceanatmosphere system through hydro-thermal vents (the reverse of reaction (1)). Reactions involving fluidsthat have equilibrated with silica-rich rocks and/or CO 2-rich reser-voirs such as the oceanatmosphere system will, on the contrary,consume hydrogen as magnetite reacts to form Fe(II) carbonatethrough reaction (2) and to form Fe(II) silicates as illustrated byreaction (3), taking into account the fayalite component in olivine.

2FeOFe2O3 [magnetite]13SiO2 [aqueous or quartz]12H2 [aqueous]53Fe2SiO4 [olivine]12H2O (3)

Themantle andits magmatic derivatives will thus be reducing relativeto silica- and/orCO2-poor fluids, but oxidizing relative to silica- and/or CO2-rich fluids. The reverse of reaction (3) represents hydro-thermal oxidation of ultramafic rocks producing magnetite andreleasing H2, which, together with CO2, provides important nichesfor methanogenic bacteria1,2 according to:

CO21 4H25CH41 2H2O (4)

Although the Gibbs energy of reaction (4) is dependent on temper-

ature and the relative partial pressures (concentrations) of CO 2, H2and CH4, extant methanogenic organisms that metabolize H2 in sub-surface and shallow marine environments are able to deplete H2 to,1025 bar (Fig. 1) where the Gibbs energy of reaction is about915 kJmol21 (refs 2729). This would be in the range of about23.5 to 4.5 at pCO2 compatible with magnetitesiderite stability.This is the order of magnitude for the lower limit of predicted atmo-spheric H2 on the early Earth determined by coupled atmosphericecosystem models involving metabolic processes of methanogens,acetogens and anoxygenic phototrophs30,31. Conversion of H2 toCH4 or organic matter by hydrogen-consuming methanogens andphototrophs is in effect an oxidation of the environment, and likeoxygen-producing phototrophs they build redox gradients in theirenvironments, with higher oxidation levels in the photic zone and

lower at deeper levels in the ocean. Such redox contrasts in the earlyArchaean ocean have been documented32 and perhaps the presence ofH2-metabolizing phototrophs might have controlled the atmosphericpH2 at values close to 10

25 bar (Fig. 1).We conclude from the phase relations in Fig. 1, and the prevalence

of magnetite in Archaean chemical sediments, that CO2 and H2mixing ratios in the early Earths atmosphere were not significantlydifferent from present-day values. IfpH2 is limited by the metabolicprocesses of methanogens (,1025 bar) then CO2 values of the orderof about 3 times PAL are consistent with magnetitesiderite phaserelations at about 25 uC (green shading in Fig. 1). Given the un-certainty in paragenetic relations, it appears that the geologic/thermodynamic constraints imposed by the mineralogy of soils13

and weathering products of fluvial sediments12 are in accord with

the mineralogy of shallow marine chemical sediments. Consequently,

Siderite

a

Laboratory culture

(ref. 28)Haematite

Magnetite

FaintearlySun

constraint(re.7)

Lowerlimit

(re.8)

Upper-limit2.2-Gyr-old

palaeosols(re.13)

Lower-limit3.2-Gyr-old

fuvialgravels(re.12)

Presentatmosphere

R

angeof2.2-Gyr-old

palaeosols(ref.13)

Lower-limitmethanogens

35C

Upper stability of H2O, ptotal = 1 bar

logpCO2

logpH2

5 4 3 2 1 0

0

2

4

6

Atmospheric environments

15C

25C

Fayalite

logmCO2(aq)

6 5 4 3 2 1

Fa

ya

lite

Mag

netite

Haematite

Siderite

3

5

7

9

60C

Subsurface hydrothermal 60 C

(ref. 29)

Marine sediments

(ref. 27)

0

logmH2(aq) 25 C

25C

Aquatic environmentsb

Lower-limitmethanogens

Figure 1 | Constraints on the partial pressures for CO2 and H2 in theatmosphere of the early Earth. Stability of minerals in the systemFeOFe2O3SiO2CO2H2O at 1 bar total pressure for the giventemperatures. a, b, Partial pressures (bar) (a) and aqueous concentrations(molalities) (b) of H2 and CO2. Solid squares indicate observed conditionsfor extant methanogens. The yellow band indicates the lower limit of H2 formethanogenesis with a slope constrained by the stoichiometry of reaction(4) (see text). Model estimates for pCO2 are indicated by arrows and verticalbars in a. The green shading denotes the stability range ofmagnetite1 siderite between 15 uC and 35 uC at a pH2 controlled bymethanogens.

NATURE | Vol 464 | 1 April 2010 LETTERS

745

Macmillan Publishers Limited. All rights reserved2010


3/6

the greenhouse effect of Archaean CO2 levels was too small to com-pensate for the faint early Sun.

Methane has been suggested as an important greenhouse gas5,6,8.Haqq-Misra et al.8 suggests an upper limit of CH4:CO2, 10.6 in theatmosphere if the positive greenhouse effect should not be counter-balanced by a negative albedo effect from organic haze, and predicteven lower mixing ratios of CH4:CO2. This requires CO2 concentra-tions in excess of 100 PAL to account for the faint early Sun, whichprecludes the stability of mixed-valence Fe(IIIII) oxides and requires

other factors that could influence the radiation balance of the Earthand counteract the faint early Sun (Fig. 1). Methane and CO2 con-centrations allowed by the mineralogic and metabolic constraints ofCO2 and H2 outlined in Fig. 1 are unable to provide sufficient green-house warming to offset the effect of the faint early Sun.

We nowinvestigatethe role of theEarths planetaryalbedo, which isa major factor determining the Earths surface temperature. Albedo isdetermined by the nature of the materials on the Earths surface, andtheabundance andtypeof clouds18. Thedifference in planetary albedoover land and oceans is minor today, partly because most land iscovered by vegetation and partly because the top-of-the-atmospherealbedo is dominated by the contribution from clouds. Because theconcentrations and type of cloud condensation nuclei (CCN) haveevolved over geologic time, clouds over the early Earth had different

characteristics. The droplet size in clouds is determined by the avail-ability of CCN and the H2O contents in the atmosphere. Few CCNresults in larger droplets, less scattering of incoming short-wave radi-ation and a shorter lifetime of clouds. In our present atmosphere themajority of CCN are produced by atmospheric oxidation of gasesreleased by plants and eukaryotic algae. With a non-oxygenic atmo-sphere and a biosphere devoid of plants and algae, CCN concentra-tions would probably have been considerablyloweron the early Earththan they are today. This biotic effect on cloud formation throughCCN can be observed over low-productivity gyres in the ocean today,where the transparency of the atmosphere for incoming short waveradiation is greater than in areas with high biologic productivity14.With a more transparent atmosphere the low albedo of the oceanwould be more strongly expressed in the planetary albedo, as illu-

strated in Fig. 2 (see Methods).We have modelled the sensitivity of the Precambrian radiationbalance to changes in the fraction of the Earths surface covered bycontinents as well as biological evolution over geologic time. TheEarths continents are formed by fractionation of the mantle througha suite of geological processes taking place over some span of time.There is little consensus on when the first continents emerged, or therate of growth since continental nucleation. However, mass-balanceconsiderations for radioactive tracers and their daughter isotopes inthemantle and crust have allowed constructionof continental growthmodels, most of which support a sigmoidal increase in the mass of thecontinents, beginning at about 4 Gyr ago with a rapid growth ratebetween 3.5 and 1.5 Gyr ago and levelling out from about 1 Gyr agoto the present, as illustrated in Fig. 2a. There is no simple relationshipbetween the mass of continental material extracted from the mantleandthe surface area of exposed land, because a number of geophysicalparameters, includingthe rheologyof theEarthsmantle,influence thearea/volume ratio for the continents. We have chosen to use thepresent-day area/volume relationship (Fig. 2a), which probably over-estimates the continental area, and in consequence, the albedo for theearly Earth. Because the timing and rate of growth of the Earthscontinents is a matter of debate, we have included a scenario in whichthe surface area occupied by continents is constant over geologic timeas one end-member in our model (see Methods).

Our model, which is based on that of Caballero and Steder33, takesinto account the effects of lower production of biogenic CCN, andlower average surface albedo (Fig. 2b) owing to less continental areato predict the temporal evolution of planetary albedo through geo-logic time (Fig. 2c) (see Methods). The results are summarized inFig. 2d, where CO2 and CH4 are both 900parts permillion by volume

(p.p.m.v.). This represents an intermediate pCO2 level within therange of environmental conditions represented by the green shadingin Fig. 1a. The model provides similar results as long as CO2 is sub-stituted for CH4 in such a way that an increase of 7 p.p.m.v. units ofCO2 is balanced by a decrease of 1 p.p.m.v. unit of CH4. At the high-CO2 extreme of the range in Fig. 1a, this corresponds to 800 p.p.m.v.for CH4 and 1,600 p.p.m.v. for CO2.

Our model (Fig. 2) shows that the albedo effects of the growth ofthe Earths continents, amplified by higher transparency for short-wave radiation of the early atmosphere, was sufficient to counter-balance the secular changes in irradiance from the Sun 4 throughgeologic history. We suggest that the evidence for the presence ofliquid water under the faint early Sun is not paradoxical 5. The earlyEarth was dominated by oceans and lacked plants and algae, whichresulted in a lower planetary albedo. Absorption of a higher fraction

0.3

0.2

0.1

0

0.15

0.1

0.05

0.22

0.26

planetary

surface

land

0.18

20

d

c

b

a

10

0

10

204 3.5 3 2.5 2 1.5 1

Tsurface

(C)

Time (Gyr before present)

Figure 2 | Climatic modelsfor theEarthsince 3.8Gyr ago. a, Thecontinentfraction (gland) of theEarths surface as a function of geologic time. Thesolidline is based on geochemical models (see Methods). The shaded areaincludes uncertainties in the estimate of continental emergence and growthas well as the possibility that the continent area has remained constant.b, Surface albedo (asurface) as a function of geological time (see Methods).The land albedo (aland) is assumed to be granitic desert of albedo 0.3gradually changing to the present vegetated land of albedo0.21 from 1.0Gyr

ago (solid line) (the contribution from ice albedo is assumed to be constantat thepresent level). Thesolidline is based on thegrowth model in a,andtheshadedareais the range covered bythe shadedareain a, assuming icealbedoproportional to gland for the lower bound. c, The Earths calculated averageplanetary albedo(aplanetary) as a function of geological time fora droplet sizeof30mm and900 p.p.m.v. CO2 and900 p.p.m.v. CH4. d, Surface temperature(Tsurface) versus geologic time assuming increasing solar irradiance, asurfacescaled to the extent of continents from b and different average cloud dropletsizes.The blue lines arecomputedfor 375p.p.m.v.CO2 and1.7 p.p.m.v. CH4and droplet sizes of 20 mm (dashed line) and 12 mm (present-day size and

water content; dotted line). The green lines are computed for droplet sizes of20mm (solid line) and 30 mm (dashed line) and 900p.p.m.v. CO2 and900 p.p.m.v. CH4. The grey shading is the asurface range in b. The black barrepresents the model from ref. 18 with dynamically adjusted clouds butpresent optical properties of clouds.

LETTERS NATURE | Vol 464 | 1 April 2010

746



4/6

of theincomingsolarradiation in combinationwith only slightly higherconcentrations of greenhouse gases, probably CO2 and methane, pro-vided the Earth with a clement climate.

METHODS SUMMARY

Mineral assemblages in chemical sediments preserved in the geologic recordreflectthe fugacitiesof dissolved gases in thePrecambrian shallowoceanthroughreactions (1) to (3). Equilibrium fugacities of H 2 and CO2 were calculated usingthe SUPCRT computer program, which calculates equilibrium constants forgiven temperatures and pressures using an internally consistent thermodynamicdatabase for minerals, aqueous species and gases. In this study we used theSLOP98 database (see Methods). The effects of the changes in solar luminosity,in radiative transparency of the atmosphere by reduced CCN from plants andalgae and the changes in surface albedo caused by continental growth and emer-gence of land vegetation were calculated by a radiativeconvective equilibriummodel based on the CliMT climate modelling toolkit33 using the NCAR CCM3radiationcode. Details, referencesand links tothe computerprogram sourcesaregiven in the Methods section.

Full Methods and any associated references are available in the online version ofthe paper at www.nature.com/nature.

Received 13 July 2009; accepted 9 February 2010.

1. Canfield, D. E., Rosing, M. T. & Bjerrum, C. Early anaerobic metabolisms. Phil.

Trans. R. Soc. B 361, 18191834 (2006).

2. Sleep,N. H.& Bird, D. K. Nichesof thepre-photosynthetic biosphereand geologic

preservation of Earths earliest ecology. Geobiology5, 101117 (2007).

3. Lowe,D. R.& Tice,M. M.Geologic evidencefor Archean atmosphericand climatic

evolution: fluctuating CO2, CH4, and O2, with an overriding tectonic control.

Geology 32, 493496 (2004).

4. Gough, D. O. Solar interior structure and luminosity variations. Sol. Phys. 74,

2134 (1981).

5. Sagan, C. & Mullen, G. Earth and Marsevolution of atmospheres and surface

temperatures. Science 177, 5256 (1972).

6. Kiehl, J. T. & Dickinson, R. E. A study of the radiative effects of enhanced

atmospheric CO2 and CH4 on early Earth surface temperatures. J. Geophys. Res.

92, 29912998 (1987).

7. Kasting, J. F. Earths early atmosphere. Science 259, 920926 (1993).

8. Haqq-Misra, J. D., Domagal-Goldman, S. D., Kasting, P. J. & Kasting, J. F. A.

Revised, hazy methane greenhouse for the Archean Earth. Astrobiology 8,

11271137 (2008).

9. Klein, C. & Beukes, N. in Proterozoic Crustal Evolution Vol. 10 Developments in

Precambrian Geology (ed. Condie, K. C.) 383418 (Elsevier, 1992).

10. Rye, R., Kuo, P. H. & Holland, H. D. Atmospheric carbon-dioxide concentrationsbefore 2.2-billion years ago. Nature 378, 603605 (1995).

11. Sleep, N. H. & Zahnle, K. Carbon dioxide cycling and implications for climate on

ancient Earth. J. Geophys. Res. Planets 106, 13731399 (2001).

12. Hessler, A. M., Lowe, D. R., Jones, R. L. & Bird, D. K. A lower limit for atmospheric

carbon dioxide levels 3.2 billion years ago. Nature 428, 736738 (2004).

13. Sheldon, N. D. Precambrian paleosols and atmospheric CO2 levels. Precambr. Res.

147, 148155 (2006).

14. Kump, L. R. & Pollard, D. Amplification of cretaceous warmth by biological cloud

feedbacks. Science 320, 195 (2008).

15. Rosing, M.T., Rose, N. M.,Bridgwater,D. & Thomsen, H. S. Earliest part of Earths

stratigraphic record: a reappraisal of the .3.7 Ga Isua (Greenland) supracrustal

sequence. Geology 24, 4346 (1996).

16. Peck, W.H., Valley, J.W., Wilde,S. A.& Graham, C. M.Oxygen isotope ratiosand

rare earth elements in 3.3 to 4.4 Ga zircons: ion microprobe evidence for high

delta O-18 continental crust and oceans in the Early Archean. Geochim.

Cosmochim. Acta 65, 42154229 (2001).

17. Catling, D. C., Zahnle, K. J. & McKay, C. P. Biogenic methane, hydrogen escape,

and the irreversible oxidation of early Earth. Science 293, 839843 (2001).

18. Rossow, W. B., Henderson-sellers, A. & Weinreich, S. K. Cloud feedbacka

stabilizing effect for the early Earth. Science 217, 12451247 (1982).

19. Charlson,R.J., Lovelock, J.E., Andreae, M.O. & Warren, S.G. Oceanicphytoplankton,

atmospheric sulfur, cloud albedo and climate. Nature 326, 655661 (1987).

20. Knauth, L. P. & Lowe, D. R. High Archean climatic temperature inferred from

oxygen isotope geochemistry of cherts in the 3.5Ga Swaziland Supergroup, South

Africa. Geol. Soc. Am. Bull. 115, 566580 (2003).

21. Tice, M. M. & Lowe, D. R. Photosynthetic microbial mats in the 3,416-Myr-old

ocean. Nature 431, 549552 (2004).

22. Sleep, N. H. & Hessler, A. M. Weathering of quartz as an Archean climatic

indicator. Earth Planet. Sci. Lett. 241, 594602 (2006).

23. Jaffres, J. B. D., Shields, G. A. & Wallmann, K. The oxygen isotope evolution of

seawater: a critical review of a long-standing controversy and an improved

geological water cycle model for the past 3.4 billion years. Earth Sci. Rev. 83,

83122 (2007).

24. Kasting, J. F. et al. Paleoclimates, ocean depth, and the oxygen isotopic

composition of seawater. Earth Planet. Sci. Lett. 252, 8293 (2006).

25. Rye, R. & Holland, H. D. Paleosols and the evolution of atmospheric oxygen: a

critical review. Am. J. Sci. 298, 621672 (1998).

26. Ohmoto, H. Watanabe, Y. & Kumazawa, K. Evidence from massive siderite beds

for a CO2-rich atmosphere before 1.8 billion years ago. Nature 429, 395399

(2004).

27. Hoehler, T. M, et al. Comparative ecology of H2 cycling in sedimentary and

phototrophic ecosystems. Antonie Van Leeuwenhoek Int. J. Gen. Molec. Microbiol.

81, 575585 (2002).

28. Kral, T. A., Brink, K. M., Miller, S. L. & McKay, C. P. Hydrogen consumptions bymethanogens on the early Earth. Origins Life Evol. B 28, 311319 (1998).

29. Chapelle, F. H. et al. A hydrogen-based subsurface microbial community

dominated by methanogens. Nature 415, 312315 (2002).

30. Kasting, J. F., Pavlov, A. A. & Siefert, J. L. A coupled ecosystem-climate model for

predicting the methane concentration in the archean atmosphere. Origins Life

Evol. B 31, 271285 (2001).

31. Charecha, P., Kasting, J. & Seifert, J. A coupled atmosphere-ecosystem model of

the early Archean Earth. Geobiology 3, 5376 (2005).

32. Rosing, M. T. & Frei, R. U-rich Archaean sea-floor sediments from Greenland

indications of . 3700 Ma oxygenic photosynthesis. Earth Planet. Sci. Lett. 217,

237244 (2004).

33. Caballero, R. & Steder, M. CliMT: An object-oriented Climate Modelling and

diagnostics Toolkit. http://maths.ucd.ie/,rca/climt/ (2008).

Acknowledgements This study was carried out at the Nordic Center for EarthEvolution funded by the Danish National Research Foundation and the researchwas supported by an Allan C. Cox Professorship to M.T.R. and endowment funds

from the Department of Geological and Environmental Sciences from StanfordUniversity to D.K.B., and NSF grants to N.H.S. We are grateful for comments fromD. Lowe, D. Canfield, F. Selsis, P. Ditlevsen and the Ice and climate group of the

Niels Bohr Institute. We thank R. Caballero and J. Bendtsen for comments on theradiation balance model. We thank J. Kasting, A. Lenardic and N. Sheldon forconstructive reviews.

Author Contributions All authors have contributed to developing the ideas

presented. M.T.R. and D.K.B. carried out the thermodynamic modelling, and C.J.B.carried out the radiative climate modelling, based on surface albedo change(M.T.R.) and low CCN (C.J.B.).

Author Information Reprints and permissions information is available atwww.nature.com/reprints. The authors declare no competing financial interests.Correspondence and requests for materials should be addressed to M.T.R.([email protected]).

NATURE | Vol 464 | 1 April 2010 LETTERS

747

http://www.nature.com/naturehttp://maths.ucd.ie/~rca/climthttp://maths.ucd.ie/~rca/climthttp://maths.ucd.ie/~rca/climthttp://www.nature.com/reprintsmailto:[email protected]:[email protected]://www.nature.com/reprintshttp://maths.ucd.ie/~rca/climthttp://www.nature.com/nature


5/6

METHODSThermodynamic modelling of phase equilibria. The partial pressures of CO2and H2 have been calculated assuming they were in equilibrium with minerals

and aqueous solutions through reactions (1) to (3), which are relevant for

describing the Earths early surface environments. The equilibrium constants

KP,T (where Pis pressure and Tis temperature) for the limiting reactions (1) to(3)havebeen calculatedat specified pressures andtemperatures. We assume that

the total atmospheric pressure was 0.1 MPa and that the temperature was either

15 uC,25 uC or 35 uC as shown in Fig. 1a. The calculations were carried outusingthe Supcrt computer programme34, using the internally consistent thermo-dynamic database SLOP98 (http://geopig.asu.edu/supcrt92_data/slop98.dat).

Fig. 1a is computed using the gas standard states for H2 and CO2, where the

aqueous species standard state is adopted for these species in diagram Fig. 1b34.

Climate modelling. In an effort to model the sensitivity of the Earths surface

temperatures under a faint early Sun and with modest CO2 and CH4 mixingratios we have employed a simple standard radiativeconvective adjusted radi-

ative equilibrium model33. The model takes into account the effect of differentcloud characteristics during the Archaean owing to lower emissions of CCN

from the biosphere and lower surface albedo linked to the gradual growth ofcontinents through geologic time. We constrain the atmospheric CO2 mixing

ratios to be within the limits dictated by the ubiquitous presence of magnetite inArchaean banded iron formations, and allow an atmospheric CH 4: CO2# 1,

which is arbitrary but within the bounds suggested by refs 8 and 35, and issupported by the observed depletion in H2 in the Archaean ocean supposedly

caused by phototrophic methanogens32.

On the present Earththe hemispherical planetaryalbedo values are verysimilardespite the fact that the northern hemisphere has a larger proportion of landsurface area compared to the southern hemisphere. By integrating present-day

top-of-the-atmosphere albedo along a pure oceanic mid-Pacific longitudinalprofile we obtain values of,0.3 for an Archaean Earth with no continents.

This value is not much different from the present global mean top-of-the-atmosphere albedo. However, the probable low productivity of the Archean2,36

would have resulted in a droplet size much bigger than today14. As a result theatmosphere would be much more transparent to short-wave radiation, for the

same water content. The resulting radiation balance would be more sensitive tosurface albedo changes, further increasing surface temperatures. As pointed out

byref.14, theeffectof lowproductivityon droplet size andthus short-wave albedocanbe seen over thesubtropical gyres, in particular in thesouthPacific,wherethe

all-sky top-of-the-atmosphere albedo is about 0.2 locally down to 0.15, as low asthe clear-sky albedo. We evaluate the combined effect of changes in mean surfacealbedo and droplet size with a convective adjusted radiative equilibrium model.

Single-columnradiativeconvective model. Weuse a simplestandard model of

convective adjusted radiative equilibrium based on the CliMT modelling toolkitdeveloped by ref. 33. The model uses the NCAR CCM3 radiation code. A global

mean lapse rate of 6.5 K km21 and present-day CO2 and CH4 mixing ratios of375 and 1.7 p.p.m., respectively, is used37. From General Circulation Model

(GCM) calculations38 we find the mean distribution of relative humidity (79,60, and 50, respectively, for 1000900 mbar, 900700 mbar and 700100mbar

model intervals). Satellite data constrain the global radiation budget, withincident short-wave flux of,340Wm22 and mean outgoing long-wave radi-

ation of 235240W m22 (that is, planetary albedo of,0.31) (CERES EBAFEdition1A39,40). The global mean surface albedo is conservatively set to 0.135,

approximating a value of 0.144 extracted from model reanalysis data for 2003(CERES SRBAVG Edition 240). Following others41 the effective layer cloud water

content is adjusted to approximate the radiative effect on short- and long-waveradiative fluxes, respectively, about24 7 W m22 and,2 9 W m22 (CERES EBAF

Edition1A39

). For the clouds we assume a present-day droplet size of 12 mm(ref. 14). Low model clouds are set to fractional cloud cover of 0.5 and cloud

water contentof 11g m22. Highmodel clouds are setto fractional cloud coverof0.2 and cloud water content of 3g m22 . This results in cloud short- and long-wave radiative effects of245.8Wm22 and 27.8Wm22, respectively. These

values approximate observations (CERES EBAF Edition1A39,40). With these

parameters the calculated mean model surface temperature is 14.3uC and theplanetary albedo is 0.33.

Kump and Pollard14 estimate that droplet size would possibly increase to17mm during the low primary production of the Cretaceous period with fewer

CCN. We usethe relative humidity distribution from theGCM results38 based on4 times the PALpCO2 (relative humidity of 80,65 and51 at 1,000900mbar,900

700mbarand 700100 mbar intervals,respectively).Thena droplet size of 17mmand reduced cloud water content has the same radiative warming effect as in the

GCM experiments38, which included precipitation enhancement and resultingcloud cover reduction (not modelled explicitly here). Increasing droplet size

from 12mm to 17mm and decreasing cloud water content to 3.0 g m22 decreases

the planetary albedo to 0.23, as shown in Fig. 2c, and increases the model globalmean temperature by 10 uC, conservatively similar to the 12 uC of ref. 14.

Estimated surface albedo through time. It is useful to look at how the presentalbedo can be approximated in simple terms, before calculating the conjectured

first-order variation of surface albedo through the Earths history. The present-

day global mean top-of-the-atmosphere albedo (aplanetary5 0.32) retrievedfrom satellite data can simply be approximated by the area-fraction-weightedplanetary albedo over land (aplanetaryland) and ocean (aplanetaryocean), that is,

glandaplanetaryland1 goceanaplanetaryocean, where gland5 0.3 is the presentarea frac-tion of the atmosphere underlain by land, and the ocean fraction is

gocean5 12gland. Mean top-of-the-atmosphere albedo of longitudinaltransects over land gives aplanetaryland< 0.35 (AfricaEurope 24u E) and overocean aplanetaryocean< 0.3 (Pacif ic Ocean, 145u W) (CERES EBAF

Edition1A39). This gives aplanetary< 0.315. Similarly the global mean surfacealbedo (asurface) can be approximated as the sum of the land fraction albedo

(gland3 0.23) and the ocean fraction albedo(gocean3 0.1), equivalent to thepresent-day global mean surface albedo of 0.14 as calculated from local short-

wave fluxes under all-sky conditions (model reanalysis data for 2003, CERESSRBAVG Edition 239). The mean surface albedo can also be calculated from thealbedo and fractional albedo contributions of different surfaces. The sea surface

with light wind waves has a global mean albedo of about asurface ocean< 0.06,

vegetation-covered continents have an albedo of about asurface land< 0.21, snow-and ice-covered regions have an albedo of asurfaceice< 0.5 (CERES SRBAVGEdition 237,39,42). Snow and ice cover 7% of the oceans (goceanice) and 11% ofland surface(glandice)

37.The resulting mean surface albedo of the Earth is again

found to be about asurface< 0.135:

asurface5 asurface ocean[gocean(12 goceanice)]1asurface land[gland(12 gland ice)]1

asurface ice(goceangoceanice1 glandglandice) (5)

We use equation (5) to calculate the surface albedo through time with appro-priate area fractions and surface albedos as follows: gland is from Fig. 2a,

gocean5 12gland, and with goceanice, glandice, aice and aocean as today. On the

early Earth the land surface albedo (aland) wouldhave been different from today.The first rocks and sediments would have been dominated by black basalts and

dark basaltic dust. Black basalt probably has an albedo similar to asphalt pave-ment (0.07,raging from 0.050.1 (ref. 37). Moist darksoil typically hasan albedo

of 0.1 (0.050.15). Once granitic rocks and quartz-rich sediment became pre-valent the albedo would have increased to 0.3, a value typical of dry desert soil

and wet sand (0.200.4). Not until the appearance of lichens and vegetationaround 1,000 to 450 million years ago would the land surface albedo decrease

to present values. To simplify matters, we conservatively assumed that the landsurface albedowas 0.3in thePrecambrian period.Snow-and ice-covered regionswere present on about 9% of the Earths surface, as today. In one endmember of

our uncertainty calculations we assume that gice on land and ocean are propor-tional to the land fraction. From equation (5) we then calculate that a conjec-

tured first-order variation of the surface albedo through the Earths historywould increase from ,0.08, gradually rising to 0.15 in conjunction with con-tinental crustal growth (Fig. 2b). The surface albedo would then decrease to

,0.13 once a terrestrial biosphere was established. The influence of the surfacealbedo on theplanetaryalbedois strongly enhanced in an atmospherewith fewer

CCNand thus a highershort-wave transparency.In ourmodelwe havetested theeffect of increased mean droplet sizes from the present-day value of 12 mm up to

17mm and 30mm. The effects of increasing the droplet size on the radiativeequilibrium temperature are shown in Fig. 2d.

Continental growth. The albedo model above is based on the assumption thatthe Earths continents have grown over geologic time. The continents on the

Earth are the products of fractionation of the silicate fraction of the planet,nominally theEarthsmantle.The mantle is dominated by MgFe silicate mineralswith minor CaAl silicate, oxide and sulphide minerals. Most other elements are

dissolved as trace components in these major phases. Mantle rocks are dense

(,3.4gcm23) and dark in colour. The continental crust is dominated by graniticrockswhichare composedof quartz andfeldspar (Ca, Na, K, Alsilicate) with onlyminor MgFe-bearing minerals suchas micaand amphibole. This rendersgranites

light in colour and low in density (,2.5gcm23). The granitic rocks of the con-tinents cannot be formed by direct partial melting of mantle rocks, because their

highsilica content is incompatible withthe minerals of the mantle. Partial meltingof the mantle produces magmas of the basalt family of rocks, which are dark in

colour and intermediate in density (,3 g c m23). Basalt is the overwhelminglydominant component of the oceanic crust, but forms only a minor fraction ofthe continents. Partial melting of basalt that has previously reacted with water to

formhydrous mineralscan leadto the formation of granites.The differentiation of

doi:10.1038/nature08955

http://geopig.asu.edu/supcrt92_data/slop98.dathttp://www.nature.com/doifinder/10.1038/nature08955http://www.nature.com/naturehttp://www.nature.com/naturehttp://www.nature.com/doifinder/10.1038/nature08955http://geopig.asu.edu/supcrt92_data/slop98.dat


6/6

the silicate Earth to form the continents is thusa multi-stage process that relies onthe plate tectonic cycling of components between the Earths reservoirs. The con-tinents, as we understand them today, could not have formed as a protocrust, butmusthavesegregated duringa prolongedperiodof time toallowfor allthe stepsinthe multi-stage fractionation process to take place. The time required to build thepresent continental mass is a matter of considerable controversy. Continentalgrowth models vary between almost-instantaneous formation of the present con-tinental mass, in which recycling of crust to the mantle would balance the forma-tion of new continental crust4345, and other models based on the trace-elementfractionationhistory of the Earthsmantle,whichsuggestthat thecontinental crust

grewover geologictime, withvarying rates of growth and possiblyepisodically4649

.In a geologic sense, emergent land is not necessarily continental, and all

continental areas are not necessarily emerged, as exemplified by the continentalshelf, which is continental in composition but submerged under the sea, versusHawaii and Iceland, which are emergent but are composed of oceanic crust. Therelationship between the mass fraction of the silicate Earth that is compartmen-talized into the continents and the surface fraction of the Earth covered byemergent continents is also a matter of poor understanding and ongoing con-troversy50,51. From theperspectiveof theEarths albedo,the importantparameteris the surface fraction of emergent continental rocks, and in our models we havemade theassumption thatemergentcontinentalsurface area scales with themassof the continents. Note that in Fig. 2a, we have adopted various scenarios for thegrowthof continental mass as a function of geologic time, such that thefull rangebetween the proposed endmember models is tested in our model for the earlyEarths climate.

34. Johnson,J. W.,Oelkers, E. H. & Helgeson,H. C. Supcrt92a software package forcalculating the standard molal thermodynamic properties of minerals, gases,

aqueous species, and reactions from 1-bar to 5000-bar and 0-degrees-C to

1000-degrees-C. Comput. Geosci. 18, 899947 (1992).

35. Pavlov, A. A., Kasting, J. F., Brown, L. L., Rages, K. A. & Freedman, R. Greenhouse

warming by CH4 in the atmosphere of early Earth. J. Geophys. Res. 105,

1198111990 (2000).

36. Bjerrum, C.J. & Canfield,D. E.Oceanproductivity beforeabout 1.9Gyr agolimited

by phosphorus adsorption onto iron oxides. Nature 417, 159162 (2002).

37. Hartmann, D. L. Global Physical Climatology (Academic Press, 1994).

38. Kump, L. R., Kasting, J. F. & Barley, M. E. Rise of atmospheric oxygen and the

upside-down Archean mantle. Geochem. Geophys. Geosyst. 2, 2000GC000114

(2001).

39. Wielicki, B. A. et al. Clouds and the Earths radiant energy system (CERES): an

Earth observing system experiment. Bull. Am. Meteorol. Soc. 77, 853868 (1996).

40. Kiehl, J. T. & Trenberth, K. E. Earths annual global mean energy budget. Bull. Am.

Meteorol. Soc. 78, 197208 (1997).

41. Caldeira, K. & Kasting, J. F. The life-span of the biosphere revisited. Nature 360,

721

723 (1992).42. Peixoto, J. P. & Ooort, A. H. Physics of Climate (American Institute of Physics,

1992).

43. Armstrong, R. L. The persistent myth of crustal growth. Aust. J. Earth Sci. 38,

613630 (1991).

44. Wilde, S. A., Valley, J. W., Peck, W. H. & Graham, C. M. Evidence from detrital

zirconsfor theexistenceof continentalcrustand oceans onthe Earth4.4 Gyrago.

Nature 409, 175178 (2001).

45. Valley, J. W. et al. A cool early Earth. Geology 30, 351254 (2002).

46. Mclennan, S. M. & Taylor, S. R. Continental freeboard, sedimentation rates and

growth of continental crust. Nature 306, 169172 (1983).

47. Collerson, K. D. & Kamber, B. S. Evolution of the continents and the atmosphere

inferred from Th-U-Nb systematics of the depleted mantle. Science 283,

15191522 (1999).

48. Nagler, T. F. & Kramers, J. D. Nd isotopic evolution of theupper mantleduring the

Precambrian: models, data and the uncertainty of both. Precambr. Res. 91,

233252 (1998).

49. Condie, K. C. Episodic continental growth models: afterthoughts and extensions.Tectonophysics 322, 153162 (2000).

50. Flament, N., Coltice, N. & Rey, P. F. A case for late-Archaean continental

emergencefrom thermal evolution models and hypsometry. Earth Planet. Sci. Lett.

275, 326336 (2008).

51. Rey, P. F. & Coltice, N. Neoarchean lithospheric strengthening andthe couplingof

Earths geochemical reservoirs. Geology 36, 635638 (2008).

doi:10.1038/nature08955

Macmillan Publishers Limited All rights reserved2010
http://www.nature.com/doifinder/10.1038/nature08955http://www.nature.com/naturehttp://www.nature.com/naturehttp://www.nature.com/doifinder/10.1038/nature08955

no climate paradox under the faint young sun - rosing et al

Documents