Iron and phosphorus co-limit þ nitrogenﬁxationinthe eastern tropical...

..............................................................

Iron and phosphorus co-limitnitrogen fixation in theeastern tropical North AtlanticMatthew M. Mills1, Celine Ridame1, Margaret Davey2,3, Julie La Roche1

& Richard J. Geider2

1Marine Biogeochemistry, IFM-GEOMAR Leibniz-Institut furMeereswissenschaften, Dusternbrooker Weg 20, D-24105 Kiel, Germany2Department of Biological Sciences, University of Essex, Colchester CO4 3SQ, UK3Marine Biological Association of the United Kingdom, Plymouth PL1 2PB, UK.............................................................................................................................................................................

The role of iron in enhancing phytoplankton productivity in highnutrient, low chlorophyll oceanic regions was demonstrated firstthrough iron-addition bioassay experiments1 and subsequentlyconfirmed by large-scale iron fertilization experiments2. Ironsupply has been hypothesized to limit nitrogen fixation andhence oceanic primary productivity on geological timescales3,providing an alternative to phosphorus as the ultimate limitingnutrient4. Oceanographic observations have been interpretedboth to confirm and refute this hypothesis5,6, but direct experi-mental evidence is lacking7. We conducted experiments to testthis hypothesis during the Meteor 55 cruise to the tropical NorthAtlantic. This region is rich in diazotrophs8 and stronglyimpacted by Saharan dust input9. Here we show that communityprimary productivity was nitrogen-limited, and that nitrogenfixation was co-limited by iron and phosphorus. Saharan dustaddition stimulated nitrogen fixation, presumably by supplyingboth iron and phosphorus10,11. Our results support the hypoth-esis that aeolian mineral dust deposition promotes nitrogenfixation in the eastern tropical North Atlantic.

There is long-standing uncertainty as to whether N or P is thenutrient that limits phytoplankton productivity in the sea12. Ontimescales of one or two days, nutrient-enrichment experimentsindicate that primary productivity is N-limited in the SargassoSea13. On geological timescales, however, N2 fixation can increasethe nitrate inventory of the ocean, thus increasing primary pro-duction. In turn, N2 fixation may be limited by either P (ref. 4) or Fe(ref. 3). These two essential nutrients are in sparse supply inoligotrophic oceans, and both P and Fe have been implicatedindividually as the nutrient that ultimately limits oceanic primaryproduction and carbon dioxide sequestration3,4. However, directexperimental assessment of the relative importance of P and Fe incontrolling N2 fixation in natural plankton populations is lacking7.

In the oligotrophic North Atlantic5,6,14 N2 fixation may fuel 50%of the export production15. Atmospheric transport of Saharan dustis a major source of dissolved Fe to the tropical North Atlantic9 andTrichodesmium spp. are most abundant in areas of high Saharandust deposition8,16. In this region, Fe may be supplied in excess ofthe Trichodesmium spp. growth requirements5,14 and this has beeninterpreted to support the P-limitation hypothesis5. Furthermore,in the subtropical and tropical Atlantic, the rate of nitrogen fixationhas been correlated with the phosphorus content (but not the ironcontent) of filamentous non-heterocystous Trichodesmium spp5.These oceanographic observations have been interpreted to supportP-limitation of diazotrophy. In fact, the low dissolved inorganicphosphate concentrations in surface waters of the Sargasso Sea havebeen used to argue that the phytoplankton community, as a whole,is P-limited in this region6.

Nutrient-addition bioassays, designed to investigate which nutri-ent (N, P or Fe) limits primary production and nitrogen fixation,were conducted at three stations in the tropical Eastern Atlantic (seeSupplementary Information) during October–November 2002. Afully replicated, factorial design of the nutrient additions using

trace-metal clean techniques was implemented to assess the indi-vidual and combined effects of N, P and Fe on the community as awhole and on the diazotrophs. Nitrogen was supplied as NH4

þ andNO3

– to allow for species-specific preferences, common among thepicophytoplankton17. Labelling with 15N2 and 14CO2 permitted theassessment of nitrogen fixation18 by diazotrophs and carbon fixa-tion19 by whole plankton community. Biomass was assessed fromchlorophyll a concentration20. Additions of Saharan soils were madeto assess the potential impact of dust deposition on the primary anddiazotrophic production.

At all locations, chlorophyll and nutrient concentrations reflectedoligotrophic conditions (see Supplementary Information). Ironconcentrations ranged between 1 and 3 nM (P. Croot, personalcommunication), and are typical of surface waters around theseareas16. Non-heterocystous filamentous cyanobacteria of the generaTrichodesmium and Katagnymene were the dominant diazotrophs,but based on nifH gene sequencing21, unicellular cyanobacterial andheterotrophic diazotrophs were also present.

In all experiments, phytoplankton production and biomass werelimited by the availability of fixed nitrogen as demonstrated by the2–3-fold stimulation of CO2 fixation and the 1.5–2.5-fold increaseof chlorophyll concentration in the treatments amended withN. In contrast, additions of P, Fe, or P þ Fe did not increase carbonfixation or chlorophyll concentration (Fig. 1). However, once N-limitation was artificially removed by adding nitrate andammonium, additions of P and Fe further increased productivityand chlorophyll. The largest effect was found when N, P and Fe wereadded simultaneously, which led to 5–10-fold increases in CO2

fixation rates (Fig. 1a–c) and corresponding increases in chlorophyll(Fig. 1d–f).

Our experiments demonstrate that N was the proximate limit-ing nutrient in this region, contrary to recent suggestions of P-limitation throughout the North Atlantic6,22. Nitrogen limitation ofprimary production accentuates the potential importance of diazo-trophy throughout this region. Nitrogen fixation was detected at allstations, but was lowest at the westernmost site (Fig. 1g). Additionof P and Fe together resulted in a 2–3-fold enhancement of the N2

fixation rate relative to the control at all three stations (Fig. 1g–i).Addition of either P alone or Fe alone did not stimulate N2 fixationat the two easternmost sites. At these locations, N2 fixation was co-limited by both P and Fe.

Saharan dust additions stimulated N2 fixation as much as twofold(Fig. 1g–i), implying that these treatments relieved the co-limitationof N2 fixation by P and Fe. We calculated that the supply of dissolvedP and Fe from the dust additions was sufficient to meet the demandsof the enhanced N2 fixation rates (see Methods and SupplementaryInformation). We cannot rule out, however, that the alleviation ofP-limitation of diazotrophy by the dust addition was due to theaddition of another micronutrient, such as Zn, which is a cofactor inmany alkaline phosphatases7.

The additions of ammonium and nitrate inhibited N2 fixation(Fig. 1h, i). Diazotrophy may have been suppressed throughphysiological feedback inhibition of the nitrogenase by dissolvedinorganic nitrogen8 in these treatments. Alternatively, other micro-organisms may have outcompeted the diazotrophs for the limited Pand Fe on alleviation of N-limitation. We cannot differentiatebetween these alternate explanations, but the repression of N2

fixation by dissolved fixed nitrogen suggests that N-limitation ofcommunity primary productivity is a pre-requisite for high rates ofdiazotrophy in surface waters.

The tropical Atlantic is subjected to some of the highest mineraldust deposition rates in the world9, and has high dissolved ironconcentrations in surface waters relative to other oceanic basins. Assuch, this is a region where the phytoplankton community as awhole23, and the diazotrophs in particular14, are least likely to beFe-limited. Given the reported importance of P-limitation in thesubtropical and tropical Atlantic5,6 and the high Fe concentrations

letters to nature

NATURE | VOL 429 | 20 MAY 2004 | www.nature.com/nature292 © 2004 Nature Publishing Group

measured in our study (see Supplementary Information), we weresurprised to find that Fe addition was required to stimulatediazotrophy. However, our results are consistent with the recentsuggestion that iron may control the abundance of diazotrophs inthe South China Sea, another region subject to high dust depo-sition24. As it is generally argued that Fe concentrations in our studyarea are in excess of diazotroph iron requirements14,25, our findingssuggest that total dissolved Fe concentration is a poor index ofbioavailability26, perhaps due to temporal variation in the chemicalspeciation of dissolved Fe (ref. 7). It is also possible that the level ofiron required to saturate diazotroph growth has been underesti-mated. The important role of Fe in our study region implies that thecontrol of N2 fixation by Fe should be even greater in other oceanicregions6,14 that receive less dust deposition.

Whether P and Fe co-limitation of N2 fixation is peculiar to theeastern tropical North Atlantic during the time that we sampled, oris common throughout the oligotrophic ocean, can only be ascer-tained by further research. It is usually assumed that P is suppliedprimarily from deep waters by mixing or diffusion, whereas Fe issupplied both from deep waters and by mineral dust depositionfrom the atmosphere. Temporal variability in these sources ofnutrients may cause transients in the ratio of P supply to Fe supply,and thus affect the relative importance of these two elements incontrolling N2 fixation. Previous suggestions of P-limitation ofdiazotrophy in the North Atlantic5,6 were based on observationsmade during spring when dust deposition is at its highest. Althoughdust deposition in the tropical Atlantic is high relative to otherlocations, the seasonal low deposition during our study in autumnmay have shifted the balance towards P and Fe co-limitation. (seeSupplementary Information).

A common source of P and Fe, through dust deposition, couldalso lead to co-limitation of nitrogen fixation. Our observation ofstimulation of nitrogen fixation by the addition of Saharan dust isconsistent with this suggestion. To be valid, this hypothesis requires

that aeolian supply of the bioavailable forms of P and Fe mustroughly match the diazotrophic demand for P and Fe. It iscommonly assumed that the dissolution of dust can supply the Ferequired to support nitrogen fixation, but would not be able tosupply the necessary P. This is because crustal material containsabout 30 times as much Fe as P (3.5% versus 0.11%), whereasdiazotrophs require 30–300 times less Fe than P. However, there areuncertainties on both the supply and demand sides of this com-parison. The availability of these nutrients on contact with sea waterdepends on the dust source, the dust load to the surface of the ocean,prevailing pH conditions during atmospheric processing as well asthe mode of deposition (wet or dry)11. Due to solubility differences,more phosphate then Fe can be released from dust (see Supplemen-tary Information). Furthermore, we expect the dissolved phosphateto be immediately bioavailable. The bioavailability of the Fe releasedfrom dust is less certain26. On the demand side, estimates of P and Ferequirements of Trichodesmium14,27 vary, whereas those of therecently discovered marine diazotrophs21 are not yet characterized.The possibility that simultaneous release of P and Fe from dust maystimulate diazotroph production in regions of high dust depositionremains an open question (see Supplementary Information).

The results presented here have important implications forunderstanding controls on marine N2 fixation and how it relatesto CO2 fixation in the North Atlantic. First, contrary to recentsuggestions6,22, our experiments demonstrate that the total primaryproductivity of the natural plankton community in the tropicalAtlantic is N-limited. Second, they demonstrate that N2 fixation isco-limited by Fe and P in a region where mineral dust deposition ishigh and iron should be in excess. This has not been reported before.Further studies are required to determine whether this co-limitationis widespread. Finally, our results suggest that dust, when suppliedat high levels locally, can relieve Fe and P co-limitation of diazo-trophy. The tropical North Atlantic is a region of high dustdeposition. It is also considered one of the most important areas

Figure 1 Effect of nutrient additions during bioassay experiments. Measurements were

taken at three sites in the tropical Atlantic during October–November 2002. a–c, Net CO2

fixation rates; d–f, chlorophyll a concentration; g–i, N2 fixation rates. Carbon fixation,

nitrogen fixation and chlorophyll were measured from separate triplicate bottles, such that

nine bottles were incubated for each nutrient treatment. Shown are means ^ standard

errors, n ¼ 3 for all variables except where n ¼ 2 (as indicated by an asterisk). Treatment

means were compared using a one-way ANOVA and a Fisher PLSD means comparison

test. Means that are not significantly different are labelled with the same letter

(a ¼ 0.05).

letters to nature

NATURE | VOL 429 | 20 MAY 2004 | www.nature.com/nature 293© 2004 Nature Publishing Group

globally for N2 fixation15. Dust deposition at this location is highlyepisodic, and has varied widely on geological timescales28. If dustdeposition can to some extent relieve both P and Fe limitation ofdiazotrophy, the postulated link between climate-driven changes indust deposition and N2 fixation may be even stronger then initiallysuggested3,29. A

MethodsTrace-metal clean techniques were strictly used throughout the preparation and executionof the experiments as this is crucial to the good survival of diazotrophs and forTrichodesmium spp. in particular27. Surface sea water was collected (1–3 m) after darkusing a trace-metal clean diaphragm pump. Sea water was pumped into 60-l carboys fromwhich it was siphoned into 1.18 l acid-washed polycarbonate bottles. Under a laminar flowhood, nutrients were added alone and in combination to final concentrations of 1.0 mMNH4

þ þ 1.0 mM NO32, 0.2 mM NaH2PO4, and 2.0 nM FeCl3. Saharan dust treatments were

also conducted with final concentrations of 0.5 mg l–1 (D1) and 2 mg l21 (D2). Theseconcentrations were chosen to simulate concentrations in the upper 1 m of the watercolumn after a strong Saharan aerosol deposition event11. The dust consisted of the finefractions of surface soils collected in the Hoggar region (Southern Algeria) with a grain-size distribution and chemical composition typical for Saharan aerosols collected far fromthe source10,11. The measured P and Fe content of the dust was 0.14 ^ 0.01% and4.97 ^ 0.49% (^ one standard error). The phosphate liberated from the dust treatmentswas approximately 2.7 and 10.8 nmol l21, and Fe released was 0.9 and 3.6 nmol l21, for the0.5 and 2.0 mg l21 dust treatments respectively (see Supplementary Information). Thebottles were then sealed gas tight and placed in an on-deck incubator with circulatingsurface sea water. Light was attenuated to 20% of incident surface values with blue filters(Lagoon Blue, Lee Filters #172). For each treatment, parallel incubations for each variable(carbon fixation, nitrogen fixation and biomass) were run in triplicate over 48 h with ratemeasurements made during the final 24 h and chlorophyll concentration determined at48 h. At each of the three study sites, over 100 bottles were incubated. For measuring netnitrogen fixation rates, 1.0 ml of 99% 15N2 was introduced to each bottle through a butylseptum using a gas-tight syringe. 15N2 uptake measurements may underestimate nitrogenfixation if significant release of dissolved nitrogen occurs30. However, this effect should beminimal in our experiments because, in a N-limited oligotrophic system, our 24-h ratemeasurements should allow released labile dissolved N to be reincorporated intoparticulate matter. For measuring primary productivity, 0.1 mCi 14C-bicarbonate wasadded to each bottle. All incubations were conducted from dawn-to-dawn and stopped bygentle filtration.

Received 17 September 2003; accepted 5 April 2004; doi:10.1038/nature02550.

1. Martin, J. H. & Fitzwater, S. E. Iron deficiency limits phytoplankton growth in the north-east Pacific

subarctic. Nature 331, 341–343 (1988).

2. Boyd, P. et al. A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron

fertilization. Nature 407, 695–702 (2000).

3. Falkowski, P. G. Evolution of the nitrogen cycle and its influence on the biological sequestration of

CO2 in the ocean. Nature 387, 272–275 (1997).

4. Tyrrell, T. The relative influences of nitrogen and phosphorus on oceanic primary production. Nature

400, 525–531 (1999).

5. Sanudo-Wilhelmy, S. A. et al. Phosphorus limitation of nitrogen fixation by Trichodesmium in the

central Atlantic Ocean. Nature 411, 66–69 (2001).

6. Wu, J., Sunda, W., Boyle, E. A. & Karl, D. M. Phosphate depletion in the Western North Atlantic

Ocean. Science 289, 759–762 (2000).

7. Morel, F. M. M. & Price, N. M. The biogeochemical cycles of trace metals in the oceans. Science 300,

944–947 (2003).

8. Capone, D. G., Zehr, J. P., Paerl, H. W., Bergman, B. & Carpenter, E. J. Trichodesmium, a globally

significant cyanobacterium. Science 276, 1221–1229 (1997).

9. Gao, Y., Kaufman, Y. J., Tanre, D., Kolber, D. & Falkowski, P. G. Seasonal distributions of aeolian iron

fluxes to the global ocean. Geophys. Res. Lett. 28, 29–32 (2001).

10. Guieu, C., Loye-Pilot, M. D., Ridame, C. & Thomas, C. Chemical characterization of the Saharan dust

end-member: Some biogeochemical implications for the western Mediterranean Sea. J. Geophys.

Res. Atmos. 107, 4258 (2002).

11. Ridame, C. & Guieu, C. Saharan input of phosphate to the oligotrophic water of the open western

Mediterranean Sea. Limnol. Oceanogr. 47, 856–869 (2002).

12. Redfield, A. C. The biological control of chemical factors in the environment. Am. Sci. 46, 205–221

(1958).

13. Graziano, L. M., Geider, R. J., Li, W. K. W. & Olaizola, M. Nitrogen limitation of North Atlantic

phytoplankton: Analysis of physiological condition in nutrient enrichment experiments. Aquat.

Microb. Ecol. 11, 53–64 (1996).

14. Berman-Frank, I., Cullen, J. T., Shaked, Y., Sherrell, R. M. & Falkowski, P. G. Iron availability, cellular

iron quotas, and nitrogen fixation in Trichodesmium. Limnol. Oceanogr. 46, 1249–1260 (2001).

15. Gruber, N. & Sarmiento, J. L. Global patterns of marine nitrogen fixation and denitrification. Glob.

Biogeochem. Cycles 11, 235–266 (1997).

16. Tyrrell, T. et al. Large-scale latitudinal distribution of Trichodesmium spp. in the Atlantic Ocean.

J. Plankton Res. 25, 405–416 (2003).

17. Palinska, K. A. et al. The signal transducer P-II and bicarbonate acquisition in Prochlorococcus marinus

PCC 9511, a marine cyanobacterium naturally deficient in nitrate and nitrite assimilation.

Microbiology 148, 2405–2412 (2002).

18. Montoya, J. P., Voss, M., Kahler, P. & Capone, D. G. A simple, high-precision, high-sensitivity tracer

assay for N2 fixation. Appl. Environ. Microbiol. 62, 986–993 (1996).

19. Lohrenz, S. et al. A comparison of in situ and simulated in situ methods for estimating primary

production. J. Plankton Res. 14, 201–221 (1992).

20. Welschmeyer, N. Fluorometric analysis of chlorophyll a in the presence of chlorophyll b and

pheopigments. Limnol. Oceanogr. 38, 1985–1992 (1994).

21. Zehr, J. P. et al. Unicellular cyanobacteria fix N2 in the subtropical North Pacific Ocean. Nature 412,

635–638 (2001).

22. Ammerman, J. W., Hood, R. R., Case, D. A. & Cotner, J. B. Phosphorus deficiency in the Atlantic: An

emerging paradigm in oceanography. EOS Trans. Am. Geophys. Union 84, 165–170 (2003).

23. Coale, K. H., Fitzwater, S. E., Gordon, R. M., Johnson, K. S. & Barber, R. Control of community

growth and export production by upwelled iron in the equatorial Pacific Ocean. Nature 379, 621–624

(1996).

24. Wu, J. F. et al. Dissolved inorganic phosphorus, dissolved iron, and Trichodesmium in the oligotrophic

South China Sea. Global Biogeochem. Cycles 17, 1008 (2003).

25. Kustka, A., Sanudo-Wilhelmy, S., Carpenter, E. J., Capone, D. G. & Raven, J. A. A revised estimate of

the iron use efficiency of nitrogen fixation, with special reference to the marine cyanobacterium

Trichodesmium spp. (Cyanophyta). J. Phycol. 39, 12–25 (2003).

26. Rue, E. & Bruland, K. Complexation of Fe(III) by natural ligands in the central north Pacific as

determined by a new competitive ligand equilibrium/absorptive cathodic voltammetry method. Mar.

Chem. 50, 117–138 (1995).

27. Kustka, A. et al. Iron requirements for dinitrogen- and ammonium-supported growth in cultures of

Trichodesmium (IMS 101): Comparison with nitrogen fixation rates and iron: carbon ratios of field

populations. Limnol. Oceanogr. 48, 1869–1884 (2003).

28. Mahowald, N. et al. Dust sources and deposition during the last glacial maximum and current climate:

A comparison of model results with paleodata from ice cores and marine sediments. J. Geophys. Res.

Atmos. 104, 15895–15916 (1999).

29. Karl, D. et al. Dinitrogen fixation in the world’s oceans. Biogeochemistry 57, 47–98 (2002).

30. Glibert, P. M. & Bronk, D. A. Release of dissolved organic nitrogen by marine diazotrophic

cyanobacteria, Trichodesmium spp. Appl. Environ. Microbiol. 60, 3996–4000 (1994).

Supplementary Information accompanies the paper on www.nature.com/nature.

Acknowledgements We would like to acknowledge the captain and crew of the F/S Meteor and the

scientists aboard the Meteor 55 SOLAS cruise, especially D. Wallace, K. Lochte, H. Bange, P. Croot,

M. Voss, F. Malien, R. Langlois and P. Fritsche. We thank D. Wallace for insightful comments on

this manuscript. Additionally, we acknowledge W. Balzer for loaning the clean container used

during M55. This work was supported by the Deutsche Forschungsgemeinschaft’s Meteor

Schwerpunktprogramm, a Natural Environment Research Council grant to R.J.G., and a Marie

Curie Post Doctoral fellowship to C.R.

Authors’ contributions The first three authors made equal contributions to the success of the

experiments. This manuscript is the product of an equal collaboration between the groups of J.L.

at IfM-Geomar and R.J.G. at University of Essex.

Competing interests statement The authors declare that they have no competing financial

interests.

Correspondence and requests for materials should be addressed to R.J.G. ([email protected]).

..............................................................

Evidence for ecology’srole in speciationJeffrey S. McKinnon1, Seiichi Mori2, Benjamin K. Blackman3*,Lior David3*, David M. Kingsley3, Leia Jamieson1, Jennifer Chou1

& Dolph Schluter4

1Biological Sciences, University of Wisconsin-Whitewater, Whitewater,Wisconsin 53190, USA2Biological Laboratory, Gifu-Keizai University, Ogaki, Gifu prefecture 503-8550,Japan3HHMI and Stanford University School of Medicine, Beckman Center B300,279 Campus Drive, Stanford, California 94305-5329, USA4Department of Zoology and Centre for Biodiversity, University of BritishColumbia, Vancouver, BC V6T 1Z4, Canada

* Present addresses: Department of Biology, Indiana University, Jordan Hall Rm 325, 1001 E. 3rd Street,

Bloomington, Indiana 47405, USA (B.K.B,); Department of Biochemistry, Stanford University School of

Medicine, Stanford, California 94305, USA (L.D.)

.............................................................................................................................................................................

A principal challenge in testing the role of natural selection inspeciation is to connect the build-up of reproductive isolationbetween populations to divergence of ecologically importanttraits1,2. Demonstrations of ‘parallel speciation’, or assortativemating by selective environment, link ecology and isolation3–5,but the phenotypic traits mediating isolation have not beenconfirmed. Here we show that the parallel build-up of mating

letters to nature

NATURE | VOL 429 | 20 MAY 2004 | www.nature.com/nature294 © 2004 Nature Publishing Group

cells fully induced with 0.5% galactose) of Gal3p and Gal80p, respectively (SupplementaryFigs S2 and S4).

Determination of galactose consumption rateTo determine the galactose consumption rate, aliquots from cultures were filtered and thegalactose concentration of the cell-free medium was analysed as follows. b-Galactosedehydrogenase was used to oxidize galactose in the presence of 2.5 mM NADþ dissolved ina buffer containing 50 mM imidazole and 5 mM MgCl2 pH 7.0 (ref. 30). Conversion ofNADþ into NADH was followed spectrophotometrically at 340 nm.

Received 2 February; accepted 9 March 2005; doi:10.1038/nature03524.

1. Freeman, M. Feedback control of intercellular signalling in development. Nature 408, 313–319

(2000).

2. Davidson, E. H. et al. A genomic regulatory network for development. Science 295, 1669–1678 (2002).

3. Xie, H., Ye, M., Feng, R. & Graf, T. Stepwise reprogramming of B cells into macrophages. Cell 117,

663–676 (2004).

4. Xiong, W. & Ferrell, J. E. Jr. A positive-feedback-based bistable ‘memory module’ that governs a cell

fate decision. Nature 426, 460–465 (2003).

5. Shykind, B. M. et al. Gene switching and the stability of odorant receptor gene choice. Cell 117,

801–815 (2004).

6. Markevich, N. I., Hoek, J. B. & Kholodenko, B. N. Signalling switches and bistability arising from

multisite phosphorylation in protein kinase cascades. J. Cell Biol. 164, 353–359 (2004).

7. Gardner, T. S., Cantor, C. R. & Collins, J. J. Construction of a genetic toggle switch in Escherichia coli.

Nature 403, 339–342 (2000).

8. Becskei, A., Seraphin, B. & Serrano, L. Positive feedback in eukaryotic gene networks: cell

differentiation by graded to binary response conversion. EMBO J. 20, 2528–2535 (2001).

9. Atkinson, M. R., Savageau, M. A., Myers, J. T. & Ninfa, A. J. Development of genetic circuitry

exhibiting toggle switch or oscillatory behavior in Escherichia coli. Cell 113, 597–607 (2003).

10. Isaacs, F. J., Hasty, J., Cantor, C. R. & Collins, J. J. Prediction and measurement of an autoregulatory

genetic module. Proc. Natl Acad. Sci. USA 100, 7714–7719 (2003).

11. Angeli, D., Ferrell, J. E. Jr. & Sontag, E. D. Detection of multistability, bifurcations, and hysteresis in a

large class of biological positive-feedback systems. Proc. Natl Acad. Sci. USA 101, 1822–1827 (2004).

12. Lai, K., Robertson, M. J. & Schaffer, D. V. The sonic hedgehog signaling system as a bistable genetic

switch. Biophys. J. 86, 2748–2757 (2004).

13. Biggar, S. R. & Crabtree, G. R. Cell signaling can direct either binary or graded transcriptional

responses. EMBO J. 20, 3167–3176 (2001).

14. Ideker, T. et al. Integrated genomic and proteomic analyses of a systematically perturbed metabolic

network. Science 292, 929–934 (2001).

15. Suzuki-Fujimoto, T. et al. Analysis of the galactose signal transduction pathway in Saccharomyces

cerevisiae: interaction between Gal3p and Gal80p. Mol. Cell. Biol. 16, 2504–2508 (1996).

16. Timson, D. J., Ross, H. C. & Reece, R. J. Gal3p and Gal1p interact with the transcriptional repressor

Gal80p to form a complex of 1:1 stoichiometry. Biochem. J. 363, 515–520 (2002).

17. Peng, G. & Hopper, J. E. Gene activation by interaction of an inhibitor with a cytoplasmic signaling

protein. Proc. Natl Acad. Sci. USA 99, 8548–8553 (2002).

18. Mizutani, A. & Tanaka, M. Regions of GAL4 critical for binding to a promoter in vivo revealed by a

visual DNA-binding analysis. EMBO J. 22, 2178–2187 (2003).

19. Johnston, M., Flick, J. S. & Pexton, T. Multiple mechanisms provide rapid and stringent glucose

repression of GAL gene expression in Saccharomyces cerevisiae. Mol. Cell. Biol. 14, 3834–3841 (1994).

20. Wieczorke, R. et al. Concurrent knock-out of at least 20 transporter genes is required to block uptake

of hexoses in Saccharomyces cerevisiae. FEBS Lett. 464, 123–128 (1999).

21. Ozbudak, E. M., Thattai, M., Lim, H. N., Shraiman, B. I. & van Oudenaarden, A. Multistability in the

lactose utilization network of Escherichia coli. Nature 427, 737–740 (2004).

22. Bhat, P. J. & Hopper, J. E. The mechanism of inducer formation in gal3 mutants of the yeast galactose

system is independent of normal galactose metabolism and mitochondrial respiratory function.

Genetics 128, 233–239 (1991).

23. Rohde, J. R., Trinh, J. & Sadowski, I. Multiple signals regulate GAL transciption in yeast. Mol. Cell. Biol.

20, 3880–3886 (2000).

24. Blake, W. J., Kærn, M., Cantor, C. R. & Collins, J. J. Noise in eukaryotic gene expression. Nature 422,

633–637 (2003).

25. Raser, J. M. & O’Shea, E. K. Control of stochasticity in eukaryotic gene expression. Science 304,

1811–1814 (2004).

26. Hasty, J., Pradines, J., Dolnik, M. & Collins, J. J. Noise-based switches and amplifiers for gene

expression. Proc. Natl Acad. Sci. USA 97, 2075–2080 (2000).

27. Bialek, W. Stability and noise in biochemical switches. Adv. Neural Info. Process. 13, 103–109 (2001).

28. Kepler, T. B. & Elston, T. C. Stochasticity in transcriptional regulation: origins, consequences, and

mathematical representations. Biophys. J. 81, 3116–3136 (2001).

29. Kramers, H. A. Brownian motion in a field of force and the diffusion model of chemical reactions.

Physica 7, 284–304 (1940).

30. Sillje, H. H. et al. Effects of different carbon fluxes on G1 phase duration, cyclin expression, and

reserve carbohydrate metabolism in Saccharomyces cerevisiae. J. Bacteriol. 179, 6560–6565 (1997).

Supplementary Information accompanies the paper on www.nature.com/nature.

Acknowledgements We thank M. Thattai for stimulating discussions and helpful suggestions on

measuring the escape rates; J. Pedraza, M. Kardar, O. Ozcan and S. Cabi for useful discussions;

B. Kaufmann for help with image analysis; and B. Pando for comments on the manuscript. M.A.

acknowledges support by an MIT Presidential Fellowship endowed by Praecis Pharmaceuticals,

Inc. A.B. is a Long Term Fellow of the Human Frontiers Science Program. This work was

supported by a NSF-CAREER and a NIH grant.

Competing interests statement The authors declare that they have no competing financial

interests.

Correspondence and requests for materials should be addressed to A.v.O. ([email protected]).

..............................................................

corrigendum

Iron and phosphorus co-limitnitrogen fixation in the easterntropical North Atlantic

Matthew M. Mills, Celine Ridame, Margaret Davey, Julie La Roche& Richard J. Geider

Nature 429, 292–294 (2004)..............................................................................................................................................................................

In this Letter to Nature, the chlorophyll a data presented in Fig. 1d–fand in Table S1 of the Supplementary Information are an order ofmagnitude too low owing to a calculation mistake. This error doesnot alter the conclusions of our paper. A

letters to nature

NATURE | VOL 435 | 12 MAY 2005 | www.nature.com/nature232© 2005 Nature Publishing Group

© 2005 Nature Publishing Group

Anthropogenic ocean acidification overthe twenty-first century and its impact oncalcifying organismsJames C. Orr1, Victoria J. Fabry2, Olivier Aumont3, Laurent Bopp1, Scott C. Doney4, Richard A. Feely5,Anand Gnanadesikan6, Nicolas Gruber7, Akio Ishida8, Fortunat Joos9, Robert M. Key10, Keith Lindsay11,Ernst Maier-Reimer12, Richard Matear13, Patrick Monfray1†, Anne Mouchet14, Raymond G. Najjar15,Gian-Kasper Plattner7,9, Keith B. Rodgers1,16†, Christopher L. Sabine5, Jorge L. Sarmiento10, Reiner Schlitzer17,Richard D. Slater10, Ian J. Totterdell18†, Marie-France Weirig17, Yasuhiro Yamanaka8 & Andrew Yool18

Today’s surface ocean is saturated with respect to calcium carbonate, but increasing atmospheric carbon dioxideconcentrations are reducing ocean pH and carbonate ion concentrations, and thus the level of calcium carbonatesaturation. Experimental evidence suggests that if these trends continue, key marine organisms—such as corals andsome plankton—will have difficulty maintaining their external calcium carbonate skeletons. Here we use 13 models of theocean–carbon cycle to assess calcium carbonate saturation under the IS92a ‘business-as-usual’ scenario for futureemissions of anthropogenic carbon dioxide. In our projections, Southern Ocean surface waters will begin to becomeundersaturated with respect to aragonite, a metastable form of calcium carbonate, by the year 2050. By 2100, thisundersaturation could extend throughout the entire Southern Ocean and into the subarctic Pacific Ocean. When livepteropods were exposed to our predicted level of undersaturation during a two-day shipboard experiment, theiraragonite shells showed notable dissolution. Our findings indicate that conditions detrimental to high-latitudeecosystems could develop within decades, not centuries as suggested previously.

Ocean uptake of CO2 will help moderate future climate change, butthe associated chemistry, namely hydrolysis of CO2 in seawater,increases the hydrogen ion concentration [Hþ]. Surface ocean pHis already 0.1 unit lower than preindustrial values. By the end of thecentury, it will become another 0.3–0.4 units lower1,2 under the IS92ascenario, which translates to a 100–150% increase in [Hþ]. Simul-taneously, aqueous CO2 concentrations [CO2(aq)] will increase andcarbonate ion concentrations ½CO22

3 � will decrease, making it moredifficult for marine calcifying organisms to form biogenic calciumcarbonate (CaCO3). Substantial experimental evidence indicates thatcalcification rates will decrease in low-latitude corals3–5, which formreefs out of aragonite, and in phytoplankton that form their tests(shells) out of calcite6,7, the stable form of CaCO3. Calcification rateswill decline along with ½CO22

3 � owing to its reaction with increasingconcentrations of anthropogenic CO2 according to the followingreaction:

CO2 þCO223 þH2O! 2HCO2

3 ð1Þ

These rates decline even when surface waters remain supersaturatedwith respect to CaCO3, a condition that previous studies havepredicted will persist for hundreds of years4,8,9.

Recent predictions of future changes in surface ocean pH andcarbonate chemistry have primarily focused on global averageconditions1,2,10 or on low latitude regions4, where reef-building coralsare abundant. Here we focus on future surface and subsurfacechanges in high latitude regions where planktonic shelled pteropodsare prominent components of the upper-ocean biota in the SouthernOcean, Arctic Ocean and subarctic Pacific Ocean11–15. Recently, it hasbeen suggested that the cold surface waters in such regions will beginto become undersaturated with respect to aragonite only whenatmospheric CO2 reaches 1,200 p.p.m.v., more than four times thepreindustrial level (4 £ CO2) of 280 p.p.m.v. (ref. 9). In contrast, ourresults suggest that some polar and subpolar surface waters willbecome undersaturated at ,2 £ CO2, probably within the next 50years.

ARTICLES

1Laboratoire des Sciences du Climat et de l’Environnement, UMR CEA-CNRS, CEA Saclay, F-91191 Gif-sur-Yvette, France. 2Department of Biological Sciences, California StateUniversity San Marcos, San Marcos, California 92096-0001, USA. 3Laboratoire d’Oceanographie et du Climat: Experimentations et Approches Numeriques (LOCEAN), CentreIRD de Bretagne, F-29280 Plouzane, France. 4Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543-1543, USA. 5National Oceanic and AtmosphericAdministration (NOAA)/Pacific Marine Environmental Laboratory, Seattle, Washington 98115-6349, USA. 6NOAA/Geophysical Fluid Dynamics Laboratory, Princeton, NewJersey 08542, USA. 7Institute of Geophysics and Planetary Physics, UCLA, Los Angeles, California 90095-4996, USA. 8Frontier Research Center for Global Change, Yokohama236-0001, Japan. 9Climate and Environmental Physics, Physics Institute, University of Bern, CH-3012 Bern, Switzerland. 10Atmospheric and Oceanic Sciences (AOS) Program,Princeton University, Princeton, New Jersey 08544-0710, USA. 11National Center for Atmospheric Research, Boulder, Colorado 80307-3000, USA. 12Max Planck Institut furMeteorologie, D-20146 Hamburg, Germany. 13CSIRO Marine Research and Antarctic Climate and Ecosystems CRC, Hobart, Tasmania 7001, Australia. 14Astrophysics andGeophysics Institute, University of Liege, B-4000 Liege, Belgium. 15Department of Meteorology, Pennsylvania State University, University Park, Pennsylvania 16802-5013, USA.16LOCEAN, Universite Pierre et Marie Curie, F-75252 Paris, France. 17Alfred Wegener Institute for Polar and Marine Research, D-27515 Bremerhaven, Germany. 18NationalOceanography Centre Southampton, Southampton SO14 3ZH, UK. †Present addresses: Laboratoire d’Etudes en Geophysique et Oceanographie Spatiales, UMR 5566 CNES-CNRS-IRD-UPS, F-31401 Toulouse, France (P.M.); AOS Program, Princeton University, Princeton, New Jersey 08544-0710, USA (K.B.R.); The Met Office, Hadley Centre, FitzRoyRoad, Exeter EX1 3PB, UK (I.J.T.).

Vol 437|29 September 2005|doi:10.1038/nature04095

681


Changes in carbonate

We have computed modern-day ocean carbonate chemistry fromobserved alkalinity and dissolved inorganic carbon (DIC), relying ondata collected during the CO2 Survey of the World Ocean CirculationExperiment (WOCE) and the Joint Global Ocean Flux Study(JGOFS). These observations are centred around the year 1994,and have recently been provided as a global-scale, gridded dataproduct GLODAP (ref. 16; see Supplementary Information).Modern-day surface ½CO22

3 � varies meridionally by more than afactor of two, from average concentrations in the Southern Ocean of105 mmol kg21 to average concentrations in tropical waters of240 mmol kg21 (Fig. 1). Low ½CO22

3 � in the Southern Ocean is dueto (1) low surface temperatures and CO2-system thermodynamics,and (2) large amounts of upwelled deep water, which contain high[CO2(aq)] from organic matter remineralization. These two effectsreinforce one another, yielding a high positive correlation of present-day ½CO22

3 � with temperature (for example, R2 ¼ 0.92 for annual

mean surface maps). Changes in ½CO223 � and [CO2(aq)] are also

inextricably linked to changes in other carbonate chemistry variables(Supplementary Fig. S1).

We also estimated preindustrial ½CO223 � from the same data, after

subtracting data-based estimates of anthropogenic DIC (ref. 17)from the modern DIC observations and assuming that preindustrialand modern alkalinity fields were identical (see SupplementaryInformation). Relative to preindustrial conditions, invasion ofanthropogenic CO2 has already reduced modern surface ½CO22

3 � bymore than 10%, that is, a reduction of 29 mmol kg21 in the tropicsand 18 mmol kg21 in the Southern Ocean. Nearly identical resultswere found when, instead of the data-based anthropogenic CO2

estimates, we used simulated anthropogenic CO2, namely themedian from 13 models that participated in the second phase ofthe Ocean Carbon-Cycle Model Intercomparison Project, orOCMIP-2 (Fig. 1c).

To quantify future changes in carbonate chemistry, we usedsimulated DIC from ocean models that were forced by twoatmospheric CO2 scenarios: the Intergovernmental Panel on ClimateChange (IPCC) IS92a ‘continually increasing’ scenario (788 p.p.m.v.in the year 2100) and the IPCC S650 ‘stabilization’ scenario(563 p.p.m.v. in the year 2100) (Fig. 1). Simulated perturbations inDIC relative to 1994 (the GLODAP reference year) were added to themodern DIC data; again, alkalinity was assumed to be constant. Toprovide a measure of uncertainty, we report model results asthe OCMIP median ^ 2j. The median generally outperformed

Figure 1 | Increasing atmospheric CO2 and decreasing surface ocean pH and[CO3

22]. a, Atmospheric CO2 used to force 13 OCMIP models over theindustrial period (‘Historical’) and for two future scenarios: IS92a (‘I’ in band c) and S650 (‘S’ in b and c). b, c, Increases in atmospheric CO2 lead toreductions in surface ocean pH (b) and surface ocean ½CO22

3 � (c). Results aregiven as global zonal averages for the 1994 data and the preindustrial(‘Preind.’) ocean. The latter were obtained by subtracting data-basedanthropogenic DIC (ref. 17) (solid line in grey-shaded area), as well as bysubtracting model-based anthropogenic DIC (OCMIP median, dotted linein grey-shaded area; OCMIP range, grey shading). Future results for the year2100 come from the 1994 data plus the simulated DIC perturbations for thetwo scenarios; results are also shown for the year 2300 with S650 (thickdashed line). The small effect of future climate change simulated by the IPSLclimate–carbon model is added as a perturbation to IS92a in the year 2100(thick dotted line); two other climate–carbon models, PIUB-Bern andCommonwealth Scientific and Industrial Research Organisation (CSIRO),show similar results (Fig. 3a). The thin dashed lines indicating the ½CO22

3 �for sea water in equilibrium with aragonite and calcite are nearly flat,revealing weak temperature sensitivity.

Figure 2 | The aragonite saturation state in the year 2100 as indicated byD[CO3

22]A. The D½CO223 �A is the in situ ½CO22

3 � minus that for aragonite-equilibrated sea water at the same salinity, temperature and pressure. Shownare the OCMIP-2 median concentrations in the year 2100 under scenarioIS92a: a, surface map; b, Atlantic; and c, Pacific zonal averages. Thick linesindicate the aragonite saturation horizon in 1765 (Preind.; white dashedline), 1994 (white solid line) and 2100 (black solid line for S650; black dashedline for IS92a). Positive D½CO22

3 �A indicates supersaturation; negativeD½CO22

3 �A indicates undersaturation.

ARTICLES NATURE|Vol 437|29 September 2005

682


individual models in OCMIP model–data comparison (Supplemen-tary Fig. S2). By the year 2100, as atmospheric CO2 reaches788 p.p.m.v. under the IS92a scenario, average tropical surface½CO22

3 � declines to 149 ^ 14 mmol kg21. This is a 45% reductionrelative to preindustrial levels, in agreement with previous predic-tions4,8. In the Southern Ocean (all waters south of 608 S), surfaceconcentrations dip to 55 ^ 5 mmol kg21, which is 18% below thethreshold where aragonite becomes undersaturated (66 mmol kg21).

These changes extend well below the sea surface. Throughout theSouthern Ocean, the entire water column becomes undersaturatedwith respect to aragonite. During the twenty-first century, under theIS92a scenario, the Southern Ocean’s aragonite saturation horizon(the limit between undersaturation and supersaturation) shoals fromits present average depth of 730 m (Supplementary Fig. S3) all theway to the surface (Fig. 2). Simultaneously, in a portion of thesubarctic Pacific, the aragonite saturation horizon shoals fromdepths of about 120 m to the surface. In the North Atlantic, surfacewaters remain saturated with respect to aragonite, but the aragonitesaturation horizon shoals dramatically; for example, north of 508N itshoals from 2,600 m to 115 m. The greater erosion in the NorthAtlantic is due to deeper penetration and higher concentrations ofanthropogenic CO2, a tendency that is already evident in present-daydata-based estimates17,18 and in models19,20 (Supplementary Figs S4and S5). Less pronounced changes were found for the calcitesaturation horizon. For example, in the year 2100 the average calcitesaturation horizon in the Southern Ocean stays below 2,200 m.Nonetheless, in 2100 surface waters of the Weddell Sea becomeslightly undersaturated with respect to calcite.

In the more conservative S650 scenario, the atmosphere reaches2 £ CO2 in the year 2100, 50 years later than with the IS92a scenario.In 2100, Southern Ocean surface waters generally remain slightlysupersaturated with respect to aragonite. However, the models also

simulate that the Southern Ocean’s average aragonite saturationhorizon will have shoaled from 730 m to 60 m, and that the entirewater column in the Weddell Sea will have become undersaturated(Fig. 2). In the north, all surface waters remain saturated underthe S650 scenario. North of 508N, the annual average aragonitesaturation horizon shoals from 140 m to 70 m in the Pacific, whereasit shoals by 2,000 m to 610 m in the North Atlantic. Therefore, undereither scenario the OCMIP models simulated large changes in surfaceand subsurface ½CO22

3 �: Yet these models account for only the directgeochemical effect of increasing atmospheric CO2 because they wereall forced with prescribed modern-day climate conditions.

In addition to this direct geochemical effect, ocean ½CO223 � is also

altered by climate variability and climate change. To quantify theadded effect of future climate change, we analysed results from threeatmosphere–ocean climate models that each included an oceancarbon-cycle component (see Supplementary Information). Thesethree models agree that twenty-first century climate change will causea general increase in surface ocean ½CO22

3 � (Fig. 3), mainly becausemost surface waters will be warmer. However, the models also agreethat the magnitude of this increase in ½CO22

3 � is small, typicallycounteracting less than 10% of the decrease due to the geochemicaleffect. High-latitude surface waters show the smallest increases in½CO22

3 �; and even small reductions in some cases. Therefore, ouranalysis suggests that physical climate change alone will not sub-stantially alter high-latitude surface ½CO22

3 � during the twenty-firstcentury.

Climate also varies seasonally and interannually, whereas ourprevious focus has been on annual changes. To illustrate how climatevariability affects surface ½CO22

3 �; we used results from an oceancarbon-cycle model forced with the daily National Centers forEnvironmental Prediction (NCEP) reanalysis fields21 over 1948–2003 (see Supplementary Information). These fields are observa-tionally based and vary on seasonal and interannual timescales.Simulated interannual variability in surface ocean ½CO22

3 � isnegligible when compared with the magnitude of the anthropogenicdecline (Fig. 3b). Seasonal variability is also negligible except in thehigh latitudes, where surface ½CO22

3 � varies by about ^15 mmol kg21

Figure 3 | Climate-induced changes in surface [CO322]. a, The twenty-first

century shift in zonal mean surface ocean ½CO223 � due to climate change

alone, from three atmosphere–ocean climate models—CSIRO-Hobart(short dashed line), IPSL-Paris (long dashed line) and PIUB-Bern (solidline)—that each include an ocean carbon-cycle component (seeSupplementary Information). b, The regional-scale seasonal andinterannual variability is simulated by an ocean carbon-cycle model forcedwith reanalysed climate forcing.

Figure 4 | Key surface carbonate chemistry variables as a function ofpCO2

. Shown are both ½CO223 � (solid lines) and [CO2(aq)] (dashed lines) for

average surface waters in the tropical ocean (thick lines), the SouthernOcean (thickest lines) and the global ocean (thin lines). Solid and dashedlines are calculated from the thermodynamic equilibrium approach. Forcomparison, open symbols are for ½CO22

3 � from our non-equilibrium,model-data approach versus seawater pCO2

(open circles) and atmosphericpCO2

(open squares); symbol thickness corresponds with line thickness,which indicates the regions for area-weighted averages. The nearly flat, thindotted lines indicate the ½CO22

3 � for seawater in equilibrium with aragonite(‘Arag. sat.’) and calcite (‘Calc. sat.’).

NATURE|Vol 437|29 September 2005 ARTICLES

683


when averaged over large regions. This is smaller than the twenty-first-century’s transient change (for example, ,50 mmol kg21 in theSouthern Ocean). However, high-latitude surface waters do becomesubstantially less saturated during winter, because of cooling (result-ing in higher [CO2(aq)]) and greater upwelling of DIC-enricheddeep water, in agreement with previous observations in the NorthPacific22. Thus, high-latitude undersaturation will be first reachedduring winter.

Our predicted changes may be compared to those found in earlierstudies, which focused on surface waters in the tropics8 and in thesubarctic Pacific22,23. These studies assumed thermodynamic equi-librium between CO2 in the atmosphere and the surface waters attheir in situ alkalinity, temperature and salinity. If, in the equilibriumapproach, the pCO2

is taken only to represent seawater pCO2; then the

results agree with our non-equilibrium approach when the sets ofcarbonate chemistry constants are identical (Fig. 4). However,assuming equilibrium with the atmosphere leads to the predictionthat future undersaturation will occur too soon (at lower atmos-pheric CO2 levels), mainly because the anthropogenic transient in theocean actually lags that in the atmosphere. For example, withthe equilibrium approach, we predict that average surface waters inthe Southern Ocean become undersaturated when atmospheric CO2

is 550 p.p.m.v. (in the year 2050 under IS92a), whereas our non-equilibrium approach, which uses models and data, indicates thatundersaturation will occur at 635 p.p.m.v. (in the year 2070). Despitethese differences, both approaches indicate that the Southern Oceansurface waters will probably become undersaturated with respect toaragonite during this century. Conversely, both of these approachesdisagree with a recent assessment9 that used a variant of the standardthermodynamic equilibrium approach, where an incorrect inputtemperature was used inadvertently.

Uncertainties

The three coupled climate–carbon models show little effect of climatechange on surface ½CO22

3 � (compare Fig. 3a to Fig. 1) partly becauseair–sea CO2 exchange mostly compensates for the changes in surfaceDIC caused by changes in marine productivity and circulation. Insubsurface waters where such compensation is lacking, these modelscould under- or over-predict how much ½CO22

3 � will change as a

result of changes in overlying marine productivity. However, themodels project a consistent trend, which only worsens the decline insubsurface ½CO22

3 �; that is, all coupled climate models predictincreased evaporation in the tropics and increased precipitation inthe high latitudes24. This leads to greater upper ocean stratification inthe high latitudes, which in turn decreases nutrients (but not to zero)and increases light availability (owing to more shallow mixed layers).Thus, at 2 £ CO2 there is a 10% local increase in surface-to-deepexport of particulate organic carbon (POC) in the Southern Oceanusing the Institut Pierre Simon Laplace (IPSL)-Paris model25. Sub-sequent remineralization of this exported POC within the thermo-cline would increase DIC, which would only exacerbate the decreasein high-latitude subsurface ½CO22

3 �: For the twenty-first century,these uncertainties appear small next to the anthropogenic DICinvasion (see Supplementary Information).

The largest uncertainty by far, and the only means to limit thefuture decline in ocean ½CO22

3 �; is the atmospheric CO2 trajectory. Tobetter characterize uncertainty due to CO2 emissions, we comparedthe six illustrative IPCC Special Reports on Emission Scenarios(SRES) in the reduced complexity, Physics Institute University ofBern (PIUB)-Bern model. Under the moderate SRES B2 scenario,average Southern Ocean surface waters in that model become under-saturated with respect to aragonite when atmospheric CO2 reaches600 p.p.m.v. in the year 2100 (Fig. 5). For the three higher-emissionSRES scenarios (A1FI, A2 and A1B), these waters become under-saturated sooner (between the years 2058 and 2073); for the twolower-emission scenarios (A1T and B1), these waters remain slightlysupersaturated in 2100. Thus, if atmospheric CO2 rises above600 p.p.m.v., most Southern Ocean surface waters will becomeundersaturated with respect to aragonite. Yet, even below this level,the Southern Ocean’s aragonite saturation horizon will shoal sub-stantially (Fig. 2). For a given atmospheric CO2 scenario, predictedchanges in surface ocean ½CO22

3 � are much more certain than therelated changes in climate. The latter depend not only on the modelresponse to CO2 forcing, but also on poorly constrained physicalprocesses, such as those associated with clouds.

Ocean CO2 uptake

With higher levels of anthropogenic CO2 and lower surface ½CO223 �;

the change in surface ocean DIC per unit change in atmospheric CO2

(mmol kg21 per p.p.m.v.) will be about 60% lower in the year 2100(under IS92a) than it is today. Simultaneously, the CO22

3 =CO2ðaqÞratio will decrease from 4:1 to 1:1 in the Southern Ocean (Fig. 4).These decreases are due to the well-understood anthropogenicreduction in buffer capacity26, already accounted for in oceancarbon-cycle models.

On the other hand, reduced export of CaCO3 from the highlatitudes would increase surface ½CO22

3 �; thereby increasing oceanCO2 uptake and decreasing atmospheric CO2. Owing to this effect,ocean CO2 uptake could increase by 6–13 petagrams (Pg) C over thetwenty-first century, based on one recent model study27 that incor-porated an empirical, CO2-dependant relationship for calcification7.Rates of calcification could decline even further, to zero, if watersactually became undersaturated with respect to both aragonite andcalcite. We estimate that the total shutdown of high-latitude arago-nite production would lead to, at most, a 0.25 Pg C yr21 increase inocean CO2 uptake, assuming that 1 Pg C yr21 of CaCO3 is exportedglobally28, that up to half of that is aragonite9,29, and that perhapshalf of all aragonite is exported from the high latitudes. The actualincrease in ocean CO2 uptake could be much lower becausethe aragonite fraction of the CaCO3 may be only 0.1 based onlow-latitude sediment traps30, and the latitudinal distribution ofaragonite export is uncertain. Thus, increased CO2 uptake fromreduced export of aragonite will provide little compensation fordecreases in ocean CO2 uptake due to reductions in buffer capacity.Of greater concern are potential biological impacts due to futureundersaturation.

Figure 5 | Average surface [CO322] in the Southern Ocean under various

scenarios. Time series of average surface ½CO223 � in the Southern Ocean for

the PIUB-Bern reduced complexity model (see Fig. 3 and SupplementaryInformation) under the six illustrative IPCC SRES scenarios. The results forthe SRES scenarios A1T and A2 are similar to those for the non-SRESscenarios S650 and IS92a, respectively.


684


Biological impacts

The changes in seawater chemistry that we project to occur duringthis century could have severe consequences for calcifying organisms,particularly shelled pteropods: the major planktonic producers ofaragonite. Pteropod population densities are high in polar andsubpolar waters. Yet only five species typically occur in such coldwater regions and, of these, only one or two species are common atthe highest latitudes31. High-latitude pteropods have one or twogenerations per year12,15,32, form integral components of foodwebs, and are typically found in the upper 300 m where they mayreach densities of hundreds to thousands of individuals per m3

(refs 11, 13–15). In the Ross Sea, for example, the prominentsubpolar–polar pteropod Limacina helicina sometimes replaceskrill as the dominant zooplankton, and is considered an overallindicator of ecosystem health33. In the strongly seasonal high lati-tudes, sedimentation pulses of pteropods frequently occur just aftersummer15,34. In the Ross Sea, pteropods account for the majority ofthe annual export flux of both carbonate and organic carbon34,35.South of the Antarctic Polar Front, pteropods also dominate theexport flux of CaCO3 (ref. 36).

Pteropods may be unable to maintain shells in waters that areundersaturated with respect to aragonite. Data from sediment trapsindicate that empty pteropod shells exhibit pitting and partialdissolution as soon as they fall below the aragonite saturationhorizon22,36,37. In vitro measurements confirm such rapid pteropodshell dissolution rates38. New experimental evidence suggests thateven the shells of live pteropods dissolve rapidly once surface watersbecome undersaturated with respect to aragonite9. Here we show thatwhen the live subarctic pteropod Clio pyramidata is subjected to alevel of undersaturation similar to what we predict for SouthernOcean surface waters in the year 2100 under IS92a, a markeddissolution occurs at the growing edge of the shell aperture within48 h (Fig. 6). Etch pits formed on the shell surface at the aperturalmargin (which is typically ,7-mm-thick) as the ,1-mm exterior(prismatic layer) peeled back (Fig. 6c), exposing the underlyingaragonitic rods to dissolution. Fourteen individuals were tested. Allof them showed similar dissolution along their growing edge, eventhough they all remained alive. If C. pyramidata cannot grow itsprotective shell, we would not expect it to survive in waters thatbecome undersaturated with respect to aragonite.

If the response of other high-latitude pteropod species to aragoniteundersaturation is similar to that of C. pyramidata, we hypothesizethat these pteropods will not be able to adapt quickly enough to livein the undersaturated conditions that will occur over much of thehigh-latitude surface ocean during the twenty-first century. Their

distributional ranges would then be reduced both within the watercolumn, disrupting vertical migration patterns, and latitudinally,imposing a shift towards lower-latitude surface waters that remainsupersaturated with respect to aragonite. At present, we do not knowif pteropod species endemic to polar regions could disappearaltogether, or if they can make the transition to live in warmer,carbonate-rich waters at lower latitudes under a different ecosystem.If pteropods are excluded from polar and subpolar regions, theirpredators will be affected immediately. For instance, gymnosomesare zooplankton that feed exclusively on shelled pteropods33,39.Pteropods also contribute to the diet of diverse carnivorous zoo-plankton, myctophid and nototheniid fishes40–42, North Pacificsalmon43,44, mackerel, herring, cod and baleen whales45.

Surface dwelling calcitic plankton, such as foraminifera andcoccolithophorids, may fare better in the short term. However, thebeginnings of high-latitude calcite undersaturation will only lag thatfor aragonite by 50–100 years. The diverse benthic calcareousorganisms in high-latitude regions may also be threatened, includingcold-water corals which provide essential fish habitat46. Cold-watercorals seem much less abundant in the North Pacific than in theNorth Atlantic46, where the aragonite saturation horizon is muchdeeper (Fig. 2). Moreover, some important taxa in Arctic andAntarctic benthic communities secrete magnesian calcite, whichcan be more soluble than aragonite. These include gorgonians46,coralline red algae and echinoderms (sea urchins)47. At 2 £ CO2,juvenile echinoderms stopped growing and produced more brittleand fragile exoskeletons in a subtropical six-month manipulativeexperiment48. However, the responses of high-latitude calcifiers toreduced ½CO22

3 � have generally not been studied. Yet experimentalevidence from many lower-latitude, shallow-dwelling calcifiersreveals a reduced ability to calcify with a decreasing carbonatesaturation state9. Given that at 2 £ CO2, calcification rates in someshallow-dwelling calcareous organisms may decline by up to 50%(ref. 9), some calcifiers could have difficulty surviving long enougheven to experience undersaturation. Certainly, they have not experi-enced undersaturation for at least the last 400,000 years49, andprobably much longer50.

Changes in high-latitude seawater chemistry that will occur by theend of the century could well alter the structure and biodiversity ofpolar ecosystems, with impacts on multiple trophic levels. Assessingthese impacts is impeded by the scarcity of relevant data.

Received 15 June; accepted 29 July 2005.

1. Haugan, P. M. & Drange, H. Effects of CO2 on the ocean environment. EnergyConvers. Mgmt 37, 1019–-1022 (1996).

Figure 6 | Shell dissolution in a live pteropod. a–d, Shell from a livepteropod, Clio pyramidata, collected from the subarctic Pacific and kept inwater undersaturated with respect to aragonite for 48 h. The whole shell (a)has superimposed white rectangles that indicate three magnified areas: theshell surface (b), which reveals etch pits from dissolution and resulting

exposure of aragonitic rods; the prismatic layer (c), which has begun to peelback, increasing the surface area over which dissolution occurs; and theaperture region (d), which reveals advanced shell dissolution whencompared to a typical C. pyramidata shell not exposed to undersaturatedconditions (e).

NATURE|Vol 437|29 September 2005 ARTICLES

685


2. Brewer, P. G. Ocean chemistry of the fossil fuel CO2 signal: the haline signal of“business as usual”. Geophys. Res. Lett. 24, 1367–-1369 (1997).

3. Gattuso, J.-P., Frankignoulle, M., Bourge, I., Romaine, S. & Buddemeier, R. W.Effect of calcium carbonate saturation of seawater on coral calcification. Glob.Planet. Change 18, 37–-46 (1998).

4. Kleypas, J. A. et al. Geochemical consequences of increased atmosphericcarbon dioxide on coral reefs. Science 284, 118–-120 (1999).

5. Langdon, C. et al. Effect of elevated CO2 on the community metabolism of anexperimental coral reef. Glob. Biogeochem. Cycles 17, 1011, doi:10.1029/2002GB001941 (2003).

6. Riebesell, U. et al. Reduced calcification of marine plankton in response toincreased atmospheric CO2. Nature 407, 364–-367 (2000).

7. Zondervan, I., Zeebe, R., Rost, B. & Riebesell, U. Decreasing marine biogeniccalcification: A negative feedback on rising atmospheric pCO2

. Glob.Biogeochem. Cycles 15, 507–-516 (2001).

8. Broecker, W. S. & Peng, T.-H. Fate of fossil fuel carbon dioxide and the globalcarbon budget. Science 206, 409–-418 (1979).

9. Feely, R. A. et al. The impact of anthropogenic CO2 on the CaCO3 system inthe oceans. Science 305, 362–-366 (2004).

10. Caldeira, K. & Wickett, M. E. Anthropogenic carbon and ocean pH. Nature 425,365 (2003).

11. Urban-Rich, J., Dagg, M. & Peterson, J. Copepod grazing on phytoplankton inthe Pacific sector of the Antarctic Polar Front. Deep-Sea Res. II 48, 4223–-4246(2001).

12. Kobayashi, H. A. Growth cycle and related vertical distribution of thethecosomatous pteropod Spiratella “Limacina” helicina in the central ArcticOcean. Mar. Biol. 26, 295–-301 (1974).

13. Pakhomov, E. A., Verheye, H. M., Atkinson, A., Laubscher, R. K. & Taunton-Clark, J. Structure and grazing impact of the mesozooplankton communityduring late summer 1994 near South Georgia, Antarctica. Polar Biol. 18,180–-192 (1997).

14. Fabry, V. J. Aragonite production by pteropod molluscs in the subarctic Pacific.Deep-Sea Res. I 36, 1735–-1751 (1989).

15. Bathmann, U., Noji, T. T. & von Bodungen, B. Sedimentation of pteropods in theNorwegian Sea in autumn. Deep-Sea Res. 38, 1341–-1360 (1991).

16. Key, R. M. et al. A global ocean carbon climatology: Results from Global DataAnalysis Project (GLODAP). Glob. Biogeochem. Cycles 18, 4031, doi:10.1029/2004GB002247 (2004).

17. Sabine, C. L. et al. The ocean sink for anthropogenic CO2. Science 305,367–-370 (2004).

18. Gruber, N. Anthropogenic CO2 in the Atlantic Ocean. Glob. Biogeochem. Cycles12, 165–-191 (1998).

19. Sarmiento, J. L., Orr, J. C. & Siegenthaler, U. A perturbation simulation of CO2

uptake in an ocean general circulation model. J. Geophys. Res. 97, 3621–-3645(1992).

20. Orr, J. C. et al. Estimates of anthropogenic carbon uptake from fourthree-dimensional global ocean models. Glob. Biogeochem. Cycles 15, 43–-60(2001).

21. Kalnay, E. et al. The NCEP/NCAR 40-year reanalysis project. Bull. Am.Meteorol. Soc. 77, 437–-471 (1996).

22. Feely, R. A. et al. Winter–-summer variations of calcite and aragonite saturationin the northeast Pacific. Mar. Chem. 25, 227–-241 (1988).

23. Feely, R. A., Byrne, R. H., Betzer, P. R., Gendron, J. F. & Acker, J. G. Factorsinfluencing the degree of saturation of the surface and intermediate waters ofthe North Pacific Ocean with respect to aragonite. J. Geophys. Res. 89,10631–-10640 (1984).

24. Sarmiento, J. L. et al. Response of ocean ecosystems to climate warming. Glob.Biogeochem. Cycles 18, 3003, doi:10.1029/2003GB002134 (2004).

25. Bopp, L. et al. Potential impact of climate change of marine export production.Glob. Biogeochem. Cycles 15, 81–-99 (2001).

26. Sarmiento, J. L., Le Quere, C. & Pacala, S. Limiting future atmospheric carbondioxide. Glob. Biogeochem. Cycles 9, 121–-137 (1995).

27. Heinze, C. Simulating oceanic CaCO3 export production in the greenhouse.Geophys. Res. Lett. 31, L16308, doi:10.1029/2004GL020613 (2004).

28. Iglesias-Rodriguez, M. D. et al. Representing key phytoplankton functionalgroups in ocean carbon cycle models: Coccolithophorids. Glob. Biogeochem.Cycles 16, 1100, doi:10.1029/2001GB001454 (2002).

29. Berner, R. A. in The Fate of Fossil Fuel CO2 in the Oceans (eds Andersen, N. R. &Malahoff, A.) 243–-260 (Plenum, New York, 1977).

30. Fabry, V. J. Shell growth rates of pteropod and heteropod molluscs andaragonite production in the open ocean: Implications for the marine carbonatesystem. J. Mar. Res. 48, 209–-222 (1990).

31. Be, A. W. H. & Gilmer, R. W. Oceanic Micropaleontology Vol. 1 (ed. Ramsey, A.)733–-808 (Academic, London, 1977).

32. Dadon, J. R. & de Cidre, L. L. The reproductive cycle of the Thecosomatouspteropod Limacina retroversa in the western South Atlantic. Mar. Biol. 114,439–-442 (1992).

33. Seibel, B. A. & Dierssen, H. M. Cascading trophic impacts of reduced biomassin the Ross Sea, Antarctica: Just the tip of the iceberg? Biol. Bull. 205, 93–-97(2003).

34. Accornero, A., Manno, C., Esposito, F. & Gambi, M. C. The vertical flux ofparticulate matter in the polyna of Terra Nova Bay. Part II. Biologicalcomponents. Antarct. Sci. 15, 175–-188 (2003).

35. Collier, R., Dymond, J., Susumu Honjo, S. M., Francois, R. & Dunbar, R. Thevertical flux of biogenic and lithogenic material in the Ross Sea: mooredsediment trap observations 1996–-1998. Deep-Sea Res. II 47, 3491–-3520 (2000).

36. Honjo, S., Francois, R., Manganini, S., Dymond, J. & Collier, R. Particle fluxes tothe interior of the Southern Ocean in the western Pacific sector along 1708W.Deep-Sea Res. II 47, 3521–-3548 (2000).

37. Betzer, P. R., Byrne, R., Acker, J., Lewis, C. S. & Jolley, R. R. The oceaniccarbonate system: a reassessment of biogenic controls. Science 226,1074–-1077 (1984).

38. Byrne, R. H., Acker, J. G., Betzer, P. R., Feely, R. A. & Cates, M. H. Water columndissolution of aragonite in the Pacific Ocean. Nature 312, 321–-326 (1984).

39. Lalli, C. M. Structure and function of the buccal apparatus of Clione limacina(Phipps) with a review of feeding in gymnosomatous pteropods. J. Exp. Mar.Biol. Ecol. 4, 101–-118 (1970).

40. Foster, B. A. & Montgomery, J. C. Planktivory in benthic nototheniid fish inMcMurdo Sound, Antarctica. Environ. Biol. Fish. 36, 313–-318 (1993).

41. Pakhomov, E., Perissinotto, A. & McQuaid, C. D. Prey composition and dailyrations of myctophid fishes in the Southern Ocean. Mar. Ecol. Prog. Ser. 134,1–-14 (1996).

42. La Mesa, M., Vacchi, M. & Sertorio, T. Z. Feeding plasticity of Trematomusnewnesi (Pisces, Nototheniidae) in Terra Nova Bay, Ross Sea, in relation toenvironmental conditions. Polar Biol. 23, 38–-45 (2000).

43. Willette, T. M. et al. Ecological processes influencing mortality of juvenile pinksalmon (Oncorhynchus gorbuscha) in Prince William Sound, Alaska. Fish.Oceanogr. 10, 14–-41 (2001).

44. Boldt, J. L. & Haldorson, L. J. Seasonal and geographical variation in juvenilepink salmon diets in the Northern Gulf of Alaska and Prince William Sound.Trans. Am. Fisheries Soc. 132, 1035–-1052 (2003).

45. Lalli, C. M. & Gilmer, R. Pelagic Snails (Stanford Univ. Press, Stanford, 1989).46. Freiwald, A., Fossa, J. H., Grehan, A., Koslow, T. & Roberts, J. M. Cold-water

Coral Reefs: Out of Sight—No Longer Out of Mind (No. 22 in Biodiversity Series,UNEP-WCMC, Cambridge, UK, 2004).

47. Dayton, P. K. in Polar Oceanography, Part B: Chemistry, Biology and Geology (ed.Smith, W. O.) 631–-685 (Academic, San Diego, 1990).

48. Shirayama, Y. & Thornton, H. Effect of increased atmospheric CO2 on shallow-water marine benthos. J. Geophys. Res. 110, C09S09, doi:10.1029/2004JC002561 (2005).

49. Petit, J. R. et al. Climate and atmospheric history of the past 420,000 yearsfrom the Vostok ice core, Antarctica. Nature 399, 429–-436 (1999).

50. Pearson, P. N. & Palmer, M. R. Middle Eocene seawater pH and atmosphericcarbon dioxide concentrations. Science 284, 1824–-1826 (1999).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We thank M. Gehlen for discussions, and J.-M. Epitalon,P. Brockmann and the Ferret developers for help with analysis. All but theclimate simulations were made as part of the OCMIP project, which waslaunched in 1995 by the Global Analysis, Integration and Modelling (GAIM)Task Force of the International Geosphere–Biosphere Programme (IGBP) withfunding from NASA (National Aeronautics and Space Administration). OCMIP-2was supported by the European Union Global Ocean Storage of AnthropogenicCarbon (EU GOSAC) project and the United States JGOFS Synthesis andModeling Project funded through NASA. The interannual simulation wassupported by the EU Northern Ocean Carbon Exchange Study (NOCES) project,which is part of OCMIP-3.

Author Information Reprints and permissions information is available atnpg.nature.com/reprintsandpermissions. The authors declare no competingfinancial interests. Correspondence and requests for materials should beaddressed to J.C.O. ([email protected]).


686

447

It has long been recognized that the phys-ical and chemical interactions associatedwith placing heat-generating nuclear

waste in a geo-hydrologic environment arecomplex and difficult to predict (1) and thatproof of safety in an absolute sense isbeyond reach. The fundamental problem ofthe Yucca Mountain nuclear waste reposi-tory project, for which the life-cycle costwas last put at over $57 billion (2), is itslack of a robust proof of safety for a periodof hazard of a half-million years or longer.Proof of safety calls for a design of relativesimplicity based on well-understood physi-cal and chemical phenomena, careful test-ing and measurement, and adequate theoryfor extrapolations. The “proof ” ultimatelylies in performance assessments showingsuch low radiation doses as to indicate thatthroughout the period of hazard essentiallyall radioactive elements are contained nearthe point of waste emplacement. In ourview, the present repository design cannotmeet these tests.

To understand this, it is important toappreciate the Environmental ProtectionAgency’s evolving radiation protectionstandards for the Yucca Mountain project.In June 2001, the EPA issued standards set-ting the maximum allowed dose at 15 mil-lirems (mrem) per year but limiting compli-ance assessment to 10,000 years. Thesestandards were invalidated by the U.S.Circuit Court of Appeals for the District ofColumbia in its ruling of 9 July 2004. Thecourt found that the EPA, by limiting com-pliance assessment to 10,000 years, hadfailed its statutory obligation to heed rec-ommendations of the National Academy ofSciences. An academy panel (3) had foundthat “peak risks might occur tens to hun-dreds of thousands of years or even furtherinto the future” and that performance

assessments would be feasible.In response to this court ruling, the EPA

issued for public comment this past 9 Augusta proposed new standard that would establisha two-tiered radiation-protection regime,with a 15-mrem/year maximum dose for thefirst 10,000 years and a 350-mrem/year dosefor up to 1 million years thereafter. If this pro-posed standard is adopted, licensing maybecome easier, but credibility of project per-formance assessments will continue to faceserious challenge.

The Total System Performance Assess-ment (TSPA) issued by the project inDecember 2001 (4) was a key input to thedocuments supporting the formal selectionof the Yucca Mountain site and the joint res-olution by Congress upholding that selec-tion after the state of Nevada exercised itsright to veto it. The TSPA did not, however,promise virtually total containment ofradioactive elements out to hundreds ofthousands of years. The waste containersare to be protected by two major engineeredfeatures: first, a corrosion-resistant outerlayer of the nickel-based alloy-22 and sec-ond, a massive and continuous “drip shield”of titanium to be installed over the contain-ers just before closure of the repository,

about a hundred years after emplacement ofwaste in the repository has begun (see fig-ure below). The performance assessmentshowed the containers and the drip shieldbeginning to fail within the f irst severaltens of thousands of years (5). How, then,could safety be assured over times vastlylonger than that?

A plume of groundwater contaminatedby radioactive elements would formbeneath the repository and would migratesouthward down the hydraulic gradient,passing, over time, through the shallowaquifers of Armagosa Valley, where there iscurrently irrigated farming. Radiationdoses above the usual regulatory limit of 15mrem/year would begin to occur after about200,000 years (6).

Even a new TSPA could not get around thefact that the preferred design does not matchcharacteristics crucial to establishing proof ofsafety. It is not a simple design: The titaniumdrip shield itself adds substantial complexityto repository design and construction.Consider the 100-year delay (7) in installationof the drip shield. Will decision-makers fourgenerations hence, with priorities unlike ourown, choose to install a feature which in year-2000 dollars would add several billion dollarsto project costs?

The project’s most complicating featureis its preferred “hot repository” option,whereby waste containers would be spacedclosely enough for the heat from radioac-tive decay to bring the water in the nearbyrock above the boiling point. The waterwould become water vapor and migrateoutward, away from the waste emplace-

N U C L E A R W A S T E

Proof of Safety

at Yucca Mountain Luther J. Carter and Thomas H. Pigford

L. J. Carter is an independent journalist in Washington,DC, and the author of Nuclear Imperatives and PublicTrust: Dealing with Radioactive Waste (Resources forthe Future, Washington, DC , 1987); e-mail:[email protected]. T. H. Pigford is professor emeri-tus, Department of Nuclear Engineering, University ofCalifornia, Berkeley, CA 94720, USA; e-mail: [email protected]

POLICY FORUMC

RED

IT:M

ON

ITO

R S

CIE

NC

E

Tuff TuffWater-bearing fractures

Drippingwater

(seepage)

Drippingwater

(seepage)

Drip shield

Wastepackage

Wastepackage Gravel

(dry)

SandDiverted

water

Spentfuel

Spentfuel

Pallet

Proposed storage of nuclear waste at Yucca Mountain repository. (Left) Current DOE disposaldesign (4), including nuclear waste package and overlying titanium drip shield within an open tunnel.(Right) Capillary barrier design (8), in which the difference in permeability and capillary propertiesbetween two backfill materials assures diversion of groundwater flow around the waste packages (nodrip shield needed).

www.sciencemag.org SCIENCE VOL 310 21 OCTOBER 2005Published by AAAS

GUEST

CORRECTED 18 NOVEMBER 2005; SEE LAST PAGE

448

ment tunnels. But complicated flows of air,water vapor, and liquid water would be cre-ated, and chlorides and other corrosivechemicals would be mobilized by waterreacting with the hot rock. This design fea-ture is puzzling, because it could do noth-ing to prevent corrosion of waste contain-ers beyond the first 10,000 years when therepository will have cooled from thedeclining levels of radioactivity.

Is the design based on well-understoodphysical and chemical phenomena? Doesit allow reliable testing and measurement,and is there adequate theory for extrapolat-ing far beyond real-time data? To thesequestions, too, the answer is no. The proj-ect design team has worked hard on safety,but the drip shield, the container’s alloy-22outer shell, and the preferred hot reposi-tory design appear to have been chosen inan ad hoc manner.

What would be a better repositorydesign? First, the design must be carefullyadapted to the specific characteristics ofthe site. At Yucca Mountain the repositorywould be built within a mountain ridge ofvolcanic tuff that is high above the watertable in “unsaturated” rock where thewater present, although there is plenty of it(100 liters per cubic meter), does not fillall the pores in the rock (8). In this unsatu-rated or vadose zone, water moves quicklythrough open fractures but its movementwithin pores in the rock is kept extremelyslow by the capillary tension between airand water in a tightly confined space.

In this setting, the repository would berelatively dry and accessible by gentleramps and tunnels from the flanks of themountain ridge. But an accompanying dis-advantage is that air moves freely throughthe mountain via interconnected open anddry fractures permeating the ridge. Thus,waste containers may corrode wheneverthey become wet or damp.

A design strategy (9, 10) that wouldappear to have an excellent chance of satis-fying a robust proof of safety is based on aman-made capillary barrier (see f igure,page 447). The waste containers would becovered first by a layer of coarse gravel,then by a layer of fine sand or finely groundtuff. This barrier would work by virtue ofthe strong capillary forces present in thesand layer and by their absence in the gravellayer. Water dripping from the tunnel ceil-ing onto the sand layer would be seized bycapillary forces and caused to move veryslowly away through the sand above thegravel. The gravel layer is the key to com-puting radiation safety over the long term.

All the waste containers beneath thegravel will corrode over time from the watervapor and oxygen present. Eventually,radioactive elements dissolved in water will

emerge from the failed containers, diffusealong the gravel particle surfaces, and, aswe infer from a performance assessment of10 years ago, remain trapped there for hun-dreds of thousands of years. This assess-ment predicted that the radiation dose tofuture people from a repository using a cap-illary barrier would be lower at all times bya factor of one million than the dose from arepository similar to the one envisioned bythe Yucca Mountain project today (11).

By comparison with the project’s presentreference design, the capillary barrier systemwould be far simpler. It would also be farcheaper, from a use of locally obtainablematerials, without need for either the alloy-22 outer shell for the containers or a costlydrip shield. The average cost for each of the14,700 waste packages, with drip shield,would be about $900,000 (7). Given theintense radioactivity present, installing a dripshield or a capillary barrier would almostcertainly entail remote handling, but the cap-illary barrier would be easier and cheaper.The capillary barrier concept remainsuntried for disposal of spent fuel or high-level waste, although there has been interna-tional experience with capillary barriers fordisposal of low-level radioactive waste.There should also be tests of barrier integrityunder earthquake forces, but ground shakingwould be greatly attenuated deep in a geo-logic repository (12).

Another design concept worthy of seri-ous exploration at Yucca Mountain is that ofusing depleted uranium in waste containersas a sacrificial material to protect the spentfuel. But f inding a predictably enduringcorrosion-resistant material could be criti-cal, because proof of safety might turn onshowing experimentally that failure of thecontainer will not be by general corrosionbut by pitting or pinholes. Oxidation of thedepleted uranium should in that case occurslowly enough for the spent fuel to be pro-tected from degradation for hundreds ofthousands of years. The U.S. Department ofEnergy (DOE) has enormous stocks ofdepleted uranium from past uraniumenrichment, and this material could be usedin casks for storage and transport of spentfuel (13) but rigorous testing for its use incasks for final disposal has not been done.

If the Yucca Mountain project should berejected or abandoned, continued storage ofspent fuel in surface facilities is the defaultoption. The most likely new place for suchstorage is in Utah where, on 9 September2005, the U.S. Nuclear RegulatoryCommission (NRC) denied the state ofUtah’s appeals to stop a nuclear industry ini-tiative to store fuel on the Skull ValleyGoshute Indian reservation (14). The avail-ability of this new storage center, whereolder fuel from many of the 65 widely scat-

tered reactor stations could go, should makefor a better Yucca Mountain repository projectby dampening legal and political pressures thatmight otherwise lead to undue haste in designwork and the supporting experimentation.

Conditions similar to those at YuccaMountain are found in rock types around theworld wherever there is an arid climate, sub-stantial topographic relief, and a deep-lyingwater table. Success at Yucca Mountain couldpowerfully suggest that hosting disposal ofspent fuel and high-level waste can be a safeand profitable use for terrain previouslydeemed a wasteland, and that siting reposi-tories in the future will be much less difficultthan it has been in the past. The restraints on aglobal resurgence of nuclear power in responseto growing energy demand and concerns aboutgreenhouse warming are, to be sure, not lim-ited to waste disposal. Nonetheless, a proof ofsafety for such disposal that is seen interna-tionally as highly robust might relieve a matterof vexing and persistent concern.

References and Notes

1. J. D. Bredehoeft et al., “Geologic disposal of high-levelradioactive waste—Earth Science perspectives,” USGSCircular (no. 779), 3 (1978).

2. “Monthly summary of program financial and budgetinformation, as of November 30, 2004” (Office ofCivil ian Radioactive Waste Management, U.S.Department of Energy, Las Vegas, NV, 2004), pp. 1–4;(www.ocrwm.doe.gov/pm/budget/index.shtml).

3. R. W. Fri et al., Technical Bases for Yucca MountainStandards (National Academy Press, Washington, DC,1995), pp. 2, 6, 9.

4. Total System Performance Assessment—Analyses forDisposal of Commercial and DOE Waste Inventories atYucca Mountain—Input to Final EnvironmentalImpact Statement and Site Suitability Evaluation (Rev00, ICN 02, Bechtel SAIC Co., Las Vegas, NV, for DOE,Las Vegas, NV, December 2001).

5. Yucca Mountain Science and Engineering Report(DOE,Washington, DC, February 2002), pp. 11–16.

6. TSPA (4), Fig. 6-.7. Analysis of the Total System Life Cycle Cost of the

Civilian Radioactive Waste Management Program(DOE,Washington, DC, May 2001), pp. 2-2, 3-6.

8. J. C. S. Long, R. C. Ewing, Annu. Rev. Earth Planet. Sci.32, 363 (2004).

9. M. J. Apted, in Proceedings of the Fifth AnnualConference on High-Level Radioactive WasteManagement, Las Vegas, NV, 22 to 26 May 1994, p. 485.

10. W. Zhou, J. Conca, R. Arthur, M. Apted, Analysis andconfirmation of the robust performance for the flow-diversion barrier system within the Yucca Mountainsite” (Tech. Rep. 107189, prepared by QuantiSci, for theElectric Power Research Institute, Palo Alto, CA, 1996).

11. TRW Environmental Safety Systems, “Total systemperformance assessment—1995: An evaluation of thepotential Yucca Mountain repository” (B00000000-0717-2200-00136, Rev. 01, TRW, Las Vegas, NV,November 1995), pp. 9–84.

12. H. R. Pratt, in Proceedings of the Workshop on SeismicPerformance of Underground Facilities (SavannahRiver Laboratory,Aiken, SC, 1981), pp. 74 and 370.

13. C . W. Forsberg , L. R. Dole, in Proceedings of theAdvances in Nuclear Cycle Management III [CD-ROM],Hilton Head, SC, 3 to 8 October 2003 (AmericanNuclear Society, La Grange, IL, 2003), session 13–03,pp. 1–14.

14. S. Martin, Private Fuels Storage, LLC , personal communication and news release (www.privatefuel-storage.com/whatsnew/newsreleases/nr9-09-05.html).

10.1126/science.1112786

21 OCTOBER 2005 VOL 310 SCIENCE www.sciencemag.org

P O L I C Y F O R U M

Published by AAAS

1www.sciencemag.org SCIENCE Erratum post date 18 NOVEMBER 2005

post date 18 November 2005

ERRATUM

C O R R E C T I O N S A N D C L A R I F I C A T I O N S

Policy Forum: “Proof of safety at Yucca Mountain” by L. J. Carter and T. H.Pigford (21 Oct. 2005, p. 447). The figure parts should have been reversed. Theoriginal design is shown on the right and the capillary barrier design is shownon the left.

Iron and phosphorus co-limit þ nitrogenﬁxationinthe eastern tropical...

Documents

Transcript of Iron and phosphorus co-limit þ nitrogenﬁxationinthe eastern tropical...