DataY.-W. Luo et al.: Database of diazotrophs in global ocean 49 known as “puﬀs” or “tufts” or as free trichomes (Carpenter, 1983). Heterocystous cyanobacteria (Nostocales,

Earth Syst Sci Data 4 47ndash73 2012wwwearth-syst-sci-datanet4472012doi105194essd-4-47-2012copy Author(s) 2012 CC Attribution 30 License

History of Geo- and Space

SciencesOpen

Acc

ess

Advances in Science amp ResearchOpen Access Proceedings

Ope

n A

cces

s Earth System

Science

Data Ope

n A

cces

s Earth System

Science

Data

Discu

ssions

Drinking Water Engineering and Science

Open Access

Drinking Water Engineering and Science

DiscussionsOpe

n Acc

ess

Social

Geography

Open

Acc

ess

Discu

ssions

Social

Geography

Open

Acc

ess

CMYK RGB

Database of diazotrophs in global ocean abundancebiomass and nitrogen fixation rates

Y-W Luo1 S C Doney1 L A Anderson2 M Benavides3 I Berman-Frank4 A Bode5 S Bonnet6K H Bostr om7 D Bottjer 8 D G Capone9 E J Carpenter10 Y L Chen11 M J Church8 J E Dore12L I Falc on13 A Fernandez14 R A Foster15 K Furuya16 F Gomez17 K Gundersen18 A M Hynes19

D M Karl 8 S Kitajima16 R J Langlois20 J LaRoche20 R M Letelier21 E Maranon14D J McGillicuddy Jr 2 P H Moisander22 C M Moore23 B Mouri no-Carballido14

M R Mulholland 24 J A Needoba25 K M Orcutt 18 A J Poulton26 E Rahav4 P Raimbault6A P Rees27 L Riemann28 T Shiozaki16 A Subramaniam29 T Tyrrell 23 K A Turk-Kubo 22

M Varela5 T A Villareal 30 E A Webb9 A E White21 J Wu31 and J P Zehr22

1Department of Marine Chemistry and Geochemistry Woods Hole Oceanographic Institution Woods HoleMassachusetts 02543 USA

2Department of Applied Ocean Science and Engineering Woods Hole Oceanographic Institution Woods HoleMassachusetts 02543 USA

3Instituto de Oceanografıa y Cambio Global Universidad de Las Palmas de Gran Canaria 35017 Las Palmasde Gran Canaria Spain

4The Mina and Everard Goodman Faculty of Life Sciences Bar-Ilan University Ramat Gan 52900 Israel5Instituto Espanol de Oceanografıa Centro Oceanografico de A Coruna 15080 A Coruna Spain

6IRD-INSU-CNRS Laboratoire drsquoOceanographie Physique et Biogeochimique UMR 6535 CentredrsquoOceanologie de Marseille Aix Marseille Universite France

7Department of Natural Sciences Linnaeus University 39182 Kalmar Sweden8School of Ocean and Earth Science and Technology University of Hawaii Honolulu Hawaii 96822 USA

9Department of Biological Sciences and Wrigley Institute for Environmental Studies University of SouthernCalifornia Los Angeles California 90089 USA

10Romberg Tiburon Center San Francisco State University Tiburon California 94920 USA11Department of Marine Biotechnology and Resources National Sun Yat-sen University Kaohsiung 80424

Taiwan12Department of Land Resources and Environmental Sciences Montana State University Bozeman MT 59717

USA13Laboratorio de Ecologıa Bacteriana Instituto de Ecologıa Universidad Nacional Autonoma de Mexico

Mexico14Departamento de Ecoloxıa e Bioloxıa Animal Universidade de Vigo 36310 Vigo Spain

15Max Planck Institute for Marine Microbiology Celsiusstrasse 1 28359 Bremen Germany16Department of Aquatic Bioscience University of Tokyo Yayoi Bunkyo Tokyo 113-8657 Japan

17Instituto Cavanilles de Biodiversidad y Biologıa Evolutiva Universidad de Valencia PO Box 22085 46071Valencia Spain

18Department of Marine Science University of Southern Mississippi Stennis Space Center Mississippi 39529USA

19MIT-WHOI Joint Program in OceanographyApplied Ocean Science and Engineering Woods HoleOceanographic Institution Woods Hole Massachusetts 02543 USA

20Leibniz Institute for Marine Sciences Duesternbrooker Weg 20 24105 Kiel Germany21College of Oceanic and Atmospheric Sciences Oregon State University Corvallis Oregon 97331 USA

22Ocean Sciences Department University of California Santa Cruz California 95064 USA

Published by Copernicus Publications

48 Y-W Luo et al Database of diazotrophs in global ocean

23University of Southampton National Oceanography Centre Southampton European Way Southampton SO14 3ZH UK24Department of Ocean Earth amp Atmospheric Sciences Old Dominion University Norfolk VA 23529 USA

25Division of Environmental and Biomolecular Systems Oregon Health amp Science University Portland Oregon 97239 USA26National Oceanography Centre Waterfront Campus European Way Southampton SO14 3ZH UK

27Plymouth Marine Laboratory Prospect Place The Hoe Plymouth PL1 3DH UK28Marine Biological Section University of Copenhagen 3000 Helsingoslashr Denmark

29Lamont Doherty Earth Observatory Palisades New York 10964 USA30Marine Science Institute University of Texas at Austin Port Aransas Texas 78382 USA

31Rosenstiel School of Marine and Atmospheric Science University of Miami Miami FL 33149 USAlowastcurrently at Department of Marine Sciences University of Georgia Athens GA 30602 USA

lowastlowastcurrently at Department of Biology University of Massachusetts Dartmouth North Dartmouth Massachusetts 02747 USA

Correspondence to Y-W Luo (yluowhoiedu)

Received 27 December 2011 ndash Published in Earth Syst Sci Data Discuss 13 February 2012Revised 25 July 2012 ndash Accepted 14 August 2012 ndash Published 31 August 2012

Abstract Marine N2 fixing microorganisms termed di-azotrophs are a key functional group in marine pelagicecosystems The biological fixation of dinitrogen (N2) tobioavailable nitrogen provides an important new source ofnitrogen for pelagic marine ecosystems and influences pri-mary productivity and organic matter export to the deepocean As one of a series of efforts to collect biomass andrates specific to different phytoplankton functional groupswe have constructed a database on diazotrophic organismsin the global pelagic upper ocean by compiling about 12 000direct field measurements of cyanobacterial diazotroph abun-dances (based on microscopic cell counts or qPCR assaystargeting thenifH genes) and N2 fixation rates Biomass con-version factors are estimated based on cell sizes to convertabundance data to diazotrophic biomass The database is lim-ited spatially lacking large regions of the ocean especiallyin the Indian Ocean The data are approximately log-normaldistributed and large variances exist in most sub-databaseswith non-zero values differing 5 to 8 orders of magnitudeReporting the geometric mean and the range of one geomet-ric standard error below and above the geometric mean thepelagic N2 fixation rate in the global ocean is estimated to be62 (52ndash73) Tg N yrminus1 and the pelagic diazotrophic biomass inthe global ocean is estimated to be 21 (14ndash31) Tg C fromcell counts and to 89 (43ndash150) Tg C fromnifH-based abun-dances Reporting the arithmetic mean and one standard errorinstead these three global estimates are 140plusmn92 Tg N yrminus118plusmn18 Tg C and 590plusmn70 Tg C respectively Uncertaintiesrelated to biomass conversion factors can change the estimateof geometric mean pelagic diazotrophic biomass in the globalocean by aboutplusmn70 It was recently established that themost commonly applied method used to measure N2 fixationhas underestimated the true rates As a result one can expectthat future rate measurements will shift the mean N2 fixationrate upward and may result in significantly higher estimatesfor the global N2 fixation The evolving database can never-theless be used to study spatial and temporal distributions

and variations of marine N2 fixation to validate geochemi-cal estimates and to parameterize and validate biogeochem-ical models keeping in mind that future rate measurementsmay rise in the future The database is stored in PANGAEA(doi101594PANGAEA774851)

1 Introduction

N2 fixation is the biological conversion of dinitrogen (N2) gasinto two molecules of ammonia by diazotrophic organismsOver geological time scales N2 fixation is important for reg-ulating fixed N concentrations in the ocean and thereby sus-taining ocean fertility (Tyrrell 1999) The rate of pelagic N2

fixation in the contemporary ocean has been estimated to be100ndash200 Tg nitrogen (N) yrminus1 which constitutes about halfof the total external source of bioavailable N to the ocean(Gruber and Sarmiento 1997 2002 Karl et al 2002 Gal-loway et al 2004 Deutsch et al 2007 Gruber 2008) It isgenerally accepted that cyanobacteria are the major N2-fixingmicroorganisms in the ocean (Karl et al 2002 Zehr 2011)However non-cyanobacterial prokaryotic plankton may alsoconduct N2 fixation in the ocean as revealed by the presenceand transcription ofnifH genes (encoding the iron proteincomponent of the nitrogenase enzyme) (Zehr et al 1998Riemann et al 2010 Farnelid et al 2011 Fernandez et al2011) albeit their relative contribution to global N2 fixationremains to be determined

There are three major types of marine planktonic dia-zotrophic cyanobacteria (Sohm et al 2011b Zehr 2011)(1) non-heterocystous filamentous cyanobacteria (2) hete-rocystous cyanobacteria and (3) unicellular cyanobacteriaAmong them most field and laboratory research has focusedon non-heterocystous filamentousTrichodesmium(Oscilla-toriales) (Dugdale et al 1961 Carpenter and Romans 1991Capone et al 1997)Trichodesmiumis an abundant dia-zotroph in the open ocean (Capone et al 2005 LaRocheand Breitbarth 2005) which can be found as large colonies

Earth Syst Sci Data 4 47ndash73 2012 wwwearth-syst-sci-datanet4472012

Y-W Luo et al Database of diazotrophs in global ocean 49

known as ldquopuffsrdquo or ldquotuftsrdquo or as free trichomes (Carpenter1983)

Heterocystous cyanobacteria (Nostocales of the generaRichelia and Calothrix) are frequently found in severaloceanic diatom genera includingRhizosoleniaand Hemi-auluswhere cyanobacteria live within the diatom frustule butoutside the cell wall or as epiphytes onChaetocerosdiatoms(Mague et al 1974 1977 Venrick 1974 Villareal 1990Foster and OrsquoMullan 2008)

The most recently characterized diazotrophic phylotypesare the unicellular cyanobacteria (UCYN) Zehr et al (2001)first successfully amplified a fragment of thenifH gene andnifH transcripts from thelt10microm size fraction of the wholewater samples that demonstrated the presence of unicellu-lar diazotrophs Subsequently Montoya et al (2004) mea-sured high rates of N2 fixation by UCYN in the PacificOcean Three distinct phylogenetic groups have been iden-tified from UCYN includingCrocosphaera watsonii(some-times referred to as Group B UCYN-B) uncultivated GroupA (UCYN-A) (Zehr et al 2001) and Group C (UCYN-C)(Langlois et al 2005 Foster et al 2007) which have onlyrecently been cultured (Taniuchi et al 2012)

Although marine cyanobacterial diazotrophs play a criti-cal role in the oceanic N cycle primary productivity and or-ganic matter export (eg Karl et al 1997 2002 Capone2000 Gruber 2008) there is still no database synthesizingthe many field measurements of diazotrophic abundances andN2 fixation rates in the global ocean Such a database is fun-damental to understand the spatial distribution and tempo-ral variability of diazotrophic biomass and activity More-over a more comprehensive set of direct measurements maybe useful in evaluating basin- and global-scale geochemi-cal estimates of diazotrophic N inputs which have gener-ally found the global N budget to be in deficit with total Nsources significantly lower than the N sinks (Gruber 2008)The database can also be expected to provide useful infor-mation with which to investigate the controlling mechanismsfor marine diazotrophic distribution and activities

In this paper we present a database compiling data on theabundance biomass and N2 fixation rate of diazotrophs inthe global ocean This effort is part of the Marine EcosystemModel Intercomparison Project (MAREMIP) in which fieldmeasurement-based databases are constructed for biomassand related process rates for phytoplankton functional types(PFTs) (Buitenhuis et al 2012) The databases named ldquoMA-Rine Ecosystem DATardquo (MAREDAT) include nine PFTs di-atomsPhaeocystis coccolithophores diazotrophs picophy-toplankton bacterioplankton mesozooplankton macrozoo-plankton and pteropods and also a database for dissolvedorganic carbon In addition to the database for diazotrophspresented in this paper other databases are presented in otherpapers of this special volume The MAREDAT databases willbe used for the future model intercomparison studies in theMAREMIP projects and are also available for public use

In Sect 2 the database is described including informa-tion about database construction data measurement meth-ods data quality control and conversion from diazotrophabundances to carbon (C) biomass In Sect 3 we presentand discuss synthesized results from the database including(1) the results of quality control (2) spatial and temporaldistribution of cell counts of diazotrophs N2 fixation ratesandnifH-based abundances (3) general characteristics of thedatasets including mean N2 fixation rates and diazotrophic Cbiomass as a function of geographical location and depth(4) estimates for global N2 fixation rates and diazotrophic Cbiomass and (5) limitations of the database and the globalestimates In Sect 4 we draw conclusions and provide rec-ommendations for appropriate use of this database

2 Data and methods

The database is available at PANGAEA(doi101594PANGAEA774851)

21 Database construction

211 Database summary

Data comprised of three types of direct measurementscell counts of diazotrophs N2 fixation rates andnifH-based abundances via quantitative polymerase chain reac-tion (qPCR) assays were compiled from the scientific lit-erature and personal communication with researchers work-ing in the field The database contains a total of 12 151 datapoints including three sub-databases (1) cell counts of di-azotrophs with 5326 data points (Table 1a) (2) N2 fixationrates with 3624 data points (Table 1b) and (3) diazotrophicabundances estimated fromnifH copy abundances (referredto asnifH-based abundances hereafter) with 3201 data points(Table 1c) Note that the counts forTrichodesmiumwerereported in number of colonies trichomes or cells yet wemaintain usage of the term ldquocell countsrdquo to distinguish count-based methods fromnifH-based abundances The diazotrophabundances based on cell counts andnifH genes were con-verted into C biomass using conversion factors (discussed be-low) In each sub-database the data were grouped into threetaxonomic typesTrichodesmium UCYN and heterocystouscyanobacteria (Table 1) A separate grouping is maintainedfor those N2 fixation rates measured from whole seawatersamples (Table 1b)

Each volumetric data point is identified by its samplingdate geographic location (latitude and longitude) and depthDepth-integrated values were calculated for those verticalprofiles with measurements available at three or more depthsBy doing this the profiles were linearly interpolated fromsurface to bottom sampling depth ie N2 fixation rates anddiazotrophic abundances are considered to be negligible be-low the bottom depth These calculated depth-integrated datapoints were not counted in the total points in Table 1 as

wwwearth-syst-sci-datanet4472012 Earth Syst Sci Data 4 47ndash73 2012


Table 1a Summary of data points for cell counts of diazotrophs including volumetric measurements ofTrichodesmium unicellularcyanobacteria and heterocystous cyanobacteria and their depth-integrals

Investigator Cell counts for diazotrophs Region References

Trichodesmium Unicellular Heterocystous Depth TotalIntegrala Pointsc

Bar-Zeev 26 26 Mediterranean Sea Bar-Zeev et al (2008)Benavides 29b 29 Subtropical N Atlantic Benavides et al (2011)Borstad 521 73 521 Tropical W Atlantic Borstad (1978)Brzezinski 7 1 7 Subtropical N Pacific Brzezinski et al (1998)Carpenter 377 44 377 Tropicalsubtropical

W AtlanticCarpenter and Price (1977)

CarpenterRomans 15 15 Subtropical N Atlantic Carpenter and Romans (1991)CaponeCarpenterSubramaniam

94b 94 Subtropical N Pacific Sohm et al (2011a)

CaponeCarpenterSubramaniam

402 176 101b+166 679 TropicalsubtropicalN Atlantic

Carpenter et al (2004)Capone et al (2005)

Capone 22 22 Arabian Sea Capone et al (1998)Chang 85 61b 146 East China Sea Chang et al (2000)Chen 47b 47 South China Sea Chen et al (2003)Chen 20 20 Upstream Kuroshio and

South China SeaChen et al (2008)

Chen 35b 35 Northwest Pacific (Kuroshio) Chen et al (2011)Dong 36b 36 South China Sea Dong et al (2008)Dugdale 15 15 Subtropical N Atlantic

(Stn S)Dugdale et al (1961)

FernandezMourino-CarballidoMaranon

201 201 Cross N-S Atlantic Fernandez et al (2010)Mourino-Carballido et al (2011)Fernandez et al (2012)

FuruyaKitajima 30 30 60 TropicalsubtropicalN Pacific

Kitajima et al (2009)

Goering 51 6 51 Tropical Atlantic Goering et al (1966)Gomez 439 33 439 Tropicalsubtropical N

PacificGomez et al (2005)

Gomez 83 12 83 Tropicalsubtropical S Pacific unpublished dataHegde 135 135 Indian Ocean Hegde et al (2008)Hulburt 19 19 Subtropical N Atlantic Hulburt (1962)Hulburt 8 8 Subtropical N Atlantic Hulburt (1966)Hulburt 26 26 Caribbean Sea Hulburt (1968)Kromkamp 44 2 44 W Indian Ocean

(African coast)Kromkamp et al (1997)

Lugomela 98 12 98 W Indian Ocean(African coast)

Lugomela et al (2002)

LetelierKarl 23 23 Subtropical N Pacific(ALOHA)

Letelier and Karl (1996)

MarumoAsaoka 26b 26 Cross N-S Pacific Marumo and Asaoka (1974)McCarthyCarpenter 65 13 65 Subtropical N Atlantic McCarthy and Carpenter (1979)Moore 6 6 Tropicalsubtropical

AtlanticMoore et al (2009)

OrcuttGundersen 35 2 37 Subtropical N Atlantic(BATS)

Orcutt et al (2001)

Poulton 341 341 682 Madagascar Basin Poulton et al (2009)Sato 9 9 Tropical Atlantic Sato et al (1966)Scranton 7 7 Subtropical N Atlantic Scranton (1984)StevenGlombitza 69 69 Tropical Atlantic Steven and Glombitza (1972)Tyrrell 315 16 315 Cross N-S Atlantic

(AMT1-10)Tyrrell et al (2003)

Villareal 242 242 62 484 Subtropical N Pacific Villareal et al (2011)Villareal 186 9b+25 195 Subtropical N Pacific Villareal et al (unpublished data)Villareal 20 30 50 Gulf of Mexico Villareal et al (unpublished data)Whilte 52 52 22 104 Gulf of California White et al (unpublished data)Wu 21 3 21 South China Sea Wu et al (2003)

Total 3274 0 1612 930 5326

a Computed from vertical profiles unless marked for those reported by data providers as depth-integralsb Data are reported by data providers as depth-integrated biomassc Do not include depth-integrated profiles unless they are reported by data providers as depth-integrated values (as marked byb)



Table 1b Summary of data points for N2 fixation rates including volumetric measurements ofTrichodesmium unicellular cyanobacteriaand heterocystous cyanobacteria and their depth-integrals

Investigator N2 Fixation Rates Region References

Tricho- Uni- Hetero- Whole Depth Total Methoddesmium cellular cystous Seawater Integrald Pointsf

Bar-Zeev 12 12 C2H2 reduction Mediterranean Sea Bar-Zeev et al (2008)Benavides 53 60 32e 145 C2H2 reduction and

15N2 assimilationSubtropical N Atlantic Benavides et al (2011)

Berman-FrankRahav 15 1 15 15N2 assimilation Red Sea Rahav et al (2012b)Berman-FrankRahav 45 6 45 15N2 assimilation Mediterranean Sea Rahav et al (2011)

Yogev et al (2011)Rahav et al (2012a)

Bonnet 27 20 10 47 15N2 assimilation Tropical Pacific Bonnet et al (2009)Bonnet 80 9 80 15N2 assimilation Mediterranean Sea Bonnet et al (2011)ChurchKarl 332 50 332 15N2 assimilation Subtropical N Pacific

(ALOHA)Church et al (2009)

Carpenter 157 39 157 C2H2 reduction TropicalsubtropicalW Atlantic

Carpenter and Price (1977)


180 40e+33 220 C2H2 reduction and15N2 assimilation

Subtropical N Pacific Sohm et al (2011a)


410 185 119e+118 714 C2H2 reduction TropicalsubtropicalN Atlantic

Capone et al (2005)

Church 7 7 15N2 assimilation Subtropical N Pacific Fong et al (2008)Dore 13 24 6 37 15N2 assimilation Subtropical N Pacific

(ALOHA)Dore et al (2002)

Falcon 24 6 24 15N2 assimilation Subtropical N Pacific(ALOHA) andSubtropical N Atlantic

Falcon et al (2004)


115 32 115 15N2 assimilation Cross N-S Atlantic Fernandez et al (2010)Mourino-Carballido et al(2011) Fernandez et al(2012)

Foster 1 1 15N2 assimilation Subtropical N Pacific Needoba et al (2007)FuruyaKitajima 24 29 53 C2H2 reduction Tropicalsubtropical N

PacificKitajima et al (2009)

FuruyaShiozaki 103 11 103 15N2 assimilation Tropicalsubtropical Pacific Shiozaki et al (2010)Gandhi 28 7 28 15N2 assimilation Indian Ocean Gandhi et al (2011)Goering 40 4 40 15N2 assimilation Tropical Atlantic Goering et al (1966)Gunderson 24 2 24 C2H2 reduction Subtropical N Pacific Gundersen et al (1976)Karl 12 12 12 9 36 15N2 assimilation Subtropical N Pacific

(ALOHA)Grabowski et al (2008)

Kromkamp 15 15 C2H2 reduction W Indian Ocean(African coast)

Kromkamp et al (1997)

Mague 8 1 8 C2H2 reduction Subtropical N Pacific Mague et al (1974)Mague 46 8 46 C2H2 reduction Subtropical N Pacific Mague et al (1977)McCarthyCarpenter 24 5 24 C2H2 reduction Subtropical N Atlantic McCarthy and Carpenter

(1979)Montoya 16 1 16 15N2 assimilation Subtropical N Pacific Montoya et al (2004)Montoya 5 5 15N2 assimilation Arafura Sea (near

Australia)Montoya et al (2004)

Moore 46 10 46 15N2 assimilation TropicalsubtropicalAtlantic

Moore et al (2009)

Mulholland 30e 30 15N2 assimilation Subtropical NW Atlantic Mulholland et al (2012)Mulholland 181 26 181 15N2 assimilation Subtropical NW Atlantic Mulholland et al

(unpublished data)Mulholland 185 36 185 15N2 assimilation Subtropical N Atlantic Mulholland et al

(unpublished data)OrcuttGundersen 31 31 15N2 assimilation Subtropical N Atlantic

(BATS)Orcutt et al (2001)

Raimbault 158 24 158 15N2 assimilation TropicalsubtropicalS Pacific

Raimbault and Garcia(2008)

Raimbault 197 33 197 15N2 assimilation Subtropical S Pacific Garcia et al (2007)Raimbault 71 12 71 15N2 assimilation Mediterranean Sea Sandroni et al (2007)Rees 5 5 15N2 assimilation Eastern Mediterranean Sea Rees et al (2006)Rees 21 7 21 15N2 assimilation Subtropical N Atlantic

(FeeP)unpublished data

Rees 70 70 15N2 assimilation Cross N-S Atlantic(AMT14amp15)

unpublished data

Rees 2 2 15N2 assimilation Western English Channel Rees et al (2009)Rees 57 5 57 15N2 assimilation Tropical E Atlantic Turk et al (2011)TurkZehr 11 2 11 15N2 assimilation Tropical Atlantic Goebel et al (2010)WhiteLetelier 11 12 4 23 15N2 assimilation Subtropical N Pacific White et al

(unpublished data)WhiteLetelier 10 43 53 15N2 assimilation Subtropical N Pacific Watkins-Brandt et al

(2011)White 134 19 134 15N2 assimilation Gulf of CA and Tropical

PacificWhite et al (2011)

Total 689 275 205 2234 757 3624

d Computed from vertical profiles unless marked for those reported by data providers as depth-integralse Data are reported by data providers as depth-integrated N2 fixation ratesf Do not include integrated profiles unless they are reported by data providers as depth-integrated values (as marked bye)



Table 1c Summary of data points fornifH-based abundances from qPCR assays including volumetric measurements ofTrichodesmiumunicellular cyanobacteria and heterocystous cyanobacteria and their depth-integrals

Investigator nifH-based abundances Region References

Trichodesmium Unicellular Hetero- Depth Totalcystous Integral Pointsg

Bonnet 12 23 12 47 TropicalsubtropicalS Pacific

Bonnet et al (2008)

BostromRiemann 58 58 Baltic Sea Bostrom et al (2007)Church 5 16 3 21 Subtropical

N Pacific(ALOHA)

Church et al (2005a)

Church 24 48 24 8 40 SubtropicalN Pacific(ALOHA)

Church et al (2005b)

Church 8 16 8 50h 82 SubtropicalN Pacific

Fong et al (2008)

FosterZehr 66 198 66 50 330 Tropical Atlantic Foster et al (2007)FosterNeedobaZehr 18 3 18 Subtropical

N PacificNeedoba et al (2007)

FosterZehr 6 6 Red Sea Foster et al (2009a)FosterSubramaniamZehr

19 40 18 10 95 Tropical E Atlantic Foster et al (2009b)

FosterZehr 17 42 17 14 76 Gulf of California White et al(unpublished data)

Liu 25 49 14 15 88 South China Sea Kong et al (2011)LangloisLaRoche 140 420 80 560 Tropicalsubtropical

N AtlanticLanglois et al (2008)

MazardScanlan 24 4 24 Arabian Sea Mazard et al (2004)MoisanderZehr 54 105 36 18 195 South China Sea Moisander et al

(2008)MoisanderZehr 178 367 110 81 655 Tropicalsubtropical

S PacificMoisander et al(2010)

Mulholland 22 51 37 110 SubtropicalNW Atlantic

Mulholland et al(2012)

OrcuttGundersen 18 18 Gulf of Mexico Orcutt et al(unpublished data)

ReesTurk 29 58 27 114 Tropical E Atlantic Turk et al (2011)TurkZehr 154 300 154 80 608 Tropical Atlantic Goebel et al (2010)

Total 759 1793 599 416 3201

g Do not include integrated profiles unless they are reported by data providers as depth-integrated values (as marked byh)h Data are reported by data providers as depth-integrated values

they are derived values An additional 709 data points of di-azotroph cell counts and N2 fixation rates that were origi-nally reported as depth-integrated values were also includedThese 709 depth-integrated data points are counted in the to-tal in Table 1 as they are independent from other data pointsEach depth-integrated data point is identified by its samplingdate geographic location and integral depth

The database also provides total diazotrophic C biomass(from cell counts andnifH-based abundances) for eachsample by summing values from the three defined dia-zotrophic typesTrichodesmium unicellular and heterocys-tous cyanobacteria Total N2 fixation rates are also providedwhen whole seawater N2 fixation rates are available they

are used as total N2 fixation rates Otherwise the total N2

fixation rates are calculated by summarizing values fromthe three defined diazotroph types In many samples mea-surements were not available for all the three defined dia-zotrophic types Also these three defined types do not repre-sent the full diazotrophic community Thus the derived totals(via summation) can be considered as the lower limits of di-azotrophic biomass and activity

Accessory data (including temperature salinity and con-centrations of nitrate phosphate iron and chlorophyll) arealso provided if available



212 Cell counts

Cell counts for diazotrophs (Table 1a) were largely per-formed by standard light microscopy whilst a number ofsamples were counted using epifluorescence microscopywith blue or green excitation (Orcutt et al 2001 Chen etal 2003 2008 2011 Carpenter et al 2004 Sohm et al2011a Villareal et al unpublished North Pacific data) Thecell counts are limited toTrichodesmiumand heterocystouscyanobacteria but no cell counts are available for UCYN(Table 1a) as UCYN-B can only be directly identified byepifluorescence microscopy and UCYN-A have not been mi-croscopically identified UCYN-C has only recently been mi-croscopically identified (Taniuchi et al 2012)

Most counts forTrichodesmiumwere reported in numberof colonies or trichomes per volume and in a few datasetsin cell densities In order to use a unified biomass conver-sion factor forTrichodesmium(discussed later) all theTri-chodesmiumcounts were converted to number of trichomesassuming commonly used conversion factors of 200 tri-chomes colonyminus1 and 100 cells trichomeminus1 (Carpenter 1983Letelier and Karl 1996 LaRoche and Breitbarth 2005 Be-navides et al 2011) An exception is for the dataset of Car-penter et al (2004) where conversion factors were measuredin selected vertical profiles in three cruises in the tropicalNorth Atlantic with averages of 137 (71ndash267) 224 (89ndash411)and 148 (56ndash384) trichomes per colony respectively In thiscase the measured conversion factors are used for this spe-cific dataset Notably the assumed conversion factor of 200trichomes colonyminus1 is consistent with values reported in Car-penter et al (2004)

The cell counts for heterocystous bacteria were groupedinto two major generaRichelia and Calothrix Counts fortheRicheliaandCalothrix are provided as heterocyst abun-dances There are several datasets (Brzezinski et al 1998Gomez et al 2005 Poulton et al 2009 Villareal et al2011 Villareal et al unpublished Gulf of Mexico data)in which abundances of host diatomHemiaulusand Rhi-zosoleniawere reported whileRicheliaheterocysts were notcounted As aHemiaulusdiatom typically contains 2Riche-lia filaments whereas aRhizosoleniadiatom can contain1ndash32Richeliafilaments (Sundstrom 1984 Villareal 19891990 Foster and OrsquoMullan 2008) the abundances of hetero-cysts for these datasets were derived from cell counts of theirhost diatoms by assuming that eachHemiaulusor Rhizosole-nia cell contains 2 or 5Richeliafilaments respectively Inone dataset (Gomez et al 2005) abundances ofChaetoceroswere counted but the associated heterocystous cyanobacteriawere found to beRichelia An average ratio of 05Richeliaheterocyst perChaetoceroscell was reported by the datasetand was used to calculate theRicheliaabundance The tri-chomes or filaments ofRichelia or Calothrix are typicallycomposed of 3ndash4 vegetative cells and 1 terminal heterocyst(Foster and Zehr 2006) Thus abundances of cells withinthese genera are estimated by multiplying the heterocyst

abundances by 5 ie assuming 5 cells per filament Note thatit may underestimate heterocystous cell abundances as theRicheliasymbionts ofRhizosoleniain some cases can con-tain more vegetative cells (near 10) (Villareal 1989 1992Janson et al 1999)

213 N2 fixation rates

N2 fixation rates were measured directly by15N-labeled N2

gas (15N2) assimilation (Montoya et al 1996) or indirectlyby the acetylene (C2H2) reduction assay (Capone 1993) (Ta-ble 1b) The15N2 assimilation method tracks the conversionof 15N2 to particulate N The15N2 tracer is added into theambient pool of N2 the 15N14N ratio is measured in theparticulate N after incubation and compared to the naturalabundance of N isotopes in unlabeled particulate materialThe C2H2 method estimates N2 fixation rate indirectly bymeasuring the reduction of C2H2 (a competitive inhibitor ofN2) to ethylene (C2H4) which is then converted to a N2 fixa-tion rate assuming 3 or 4 moles C2H2 reduced per 1 mole N2fixed depending on the extent of nitrogenase-linked hydro-gen production (Postgate 1998) Generally the direct15N2

assimilation is a precise and sensitive method hence it hasbeen used to generate the majority of rates (Table 1b) Di-rect comparison of these two methods showed the direct15N2

assimilation method generally yields lower rates than thoseestimated from C2H2 reduction assay (see summary in Mul-holland 2007) Discrepancies between these measures couldbe because the15N2 assimilation method measures the netrate of conversion of reduced N to cellular N or net N2 fix-ation while the C2H2 reduction method measures gross N2

fixation which includes the reduced N both stored in cellsand excreted as ammonium or dissolved organic N during in-cubation (Mulholland et al 2004 Mulholland 2007) Morerecently it has been suggested that the direct15N2 assim-ilation method significantly underestimates the N2 fixationrates because the15N2 bubbles injected in seawater do not at-tain equilibrium with surrounding water (Mohr et al 2010)which will be discussed later We include the N2 fixationrates acquired by either the15N2 assimilation method or theC2H2 reduction assays in the database considering that theydistribute within a similar range of magnitude (Fig 1) Butthis analysis does not consider different sampling sites Fur-ther investigations using pair-wise comparison of the meth-ods are needed to evaluate effects of these two methods onN2 fixation rate measurement Users have to be careful whenusing the database to study N2 fixation rates aggregated fromtwo different methods

The collected N2 fixation rates were mostly measured forwhole seawater samples (Table 1b) Some samples were fil-tered and N2 fixation rates were measured for organismsin the lt10microm size-fraction which we have assigned tounicellular diazotrophs (Table 1b) Note that UCYN-B canform colonies and may not be included in this size frac-tion It is also possible that some diatoms with associated



75

Figure 1

Figure 1 Histogram of volumetric whole seawater N2 fixationrates (non-zero values) from15N2 assimilation assays and C2H2 re-duction assays Data values are on logarithmic scale

heterocystous cyanobacteria may be included inlt10micromfractions N2 fixation rates were also measured in somedatasets specifically forTrichodesmiumand heterocystouscyanobacteria Most N2 fixation data were reported as dailyrates except for 11 datasets that were reported as hourlyratesTrichodesmiumfixes N2 exclusively during the lightperiod while the diel patterns of N2 fixation are unclear forother diazotrophs (Carpenter and Capone 2008 and refer-ences therein) Thus those hourly N2 fixation rates were con-verted to daily rates by multiplying by 12 h which how-ever could be conservative if diazotrophs other thanTri-chodesmiumcould fix N2 during night (eg Montoya et al2004 Zehr et al 2007)

214 NifH-based abundances

NifH abundances were estimated by qPCR targeting thenifHgene (Church et al 2005a b Foster et al 2007) Cellu-lar DNA was extracted and the gene sequences were tar-geted for different diazotrophic groups MostnifH-basedabundances were estimated for the three major diazotrophtypes Trichodesmium UCYN groups and heterocystousgroups (Table 1c) Gene copies for UCYN were identi-fied as UCYN-A UCYN-B andor UCYN-C groupsNifH-based abundances were also estimated for different groupsof heterocystous cyanobacteria based on threenifH gene se-quences (het-1 het-2 and het-3) which have been identifiedin symbioses with diatomsRichelia-Rhizosolenia Richelia-HemiaulusandCalothrix-Chaetoceros respectively (Churchet al 2005b Foster and Zehr 2006) Note that there is onedataset (Bostrom et al 2007) that reports the abundance ofheterocystous genusNodularia The sensitivity of the qPCR

thermocycler is usually in the range of 6ndash12nifH copies perreaction When accounting for sample filtration volume fi-nal elution volume and the volume of the DNA extract usedin the qPCR reaction the detection limit of the assay is sig-nificantly higher Due to the variety of factors that determinethe assay detection limit it varies between laboratories andeven datasets and is typically on the order of 25ndash100nifHcopies lminus1 Thus for those data points reported as ldquodetectedbut not quantifiablerdquo 10nifH copies lminus1 could be a conserva-tive estimate

To estimate diazotrophic abundancesnifH gene copies areconverted to number of diazotrophic cells on 1 1 basis as itis for those diazotrophic genomes that have been sequenced(egTrichodesmiumand UCYN-B) that there is onenifHgene copy per genome (Zehr et al 2008) and assuming onegenome copy per cellNifH genes are present in both the veg-etative cells and the heterocysts of heterocystous cyanobac-teria (Foster et al 2009b) Thus this estimate accounts forabundances of the total cells not just the heterocysts of theheterocystous cyanobacteria Limitations are associated withthis extrapolation Evidence indicates that this extrapolationoverestimates the diazotrophic abundances and can only betreated as an upper limit of the cell density because of thepresence of multiplenifH gene copies per cell in some dia-zotrophs such asClostridium pasteurianum(Langlois et al2008) Little information is available on the variability ofgenome copies per cell for allnifH phylotypes (Langlois etal 2008) It is possible that when there is excess phospho-rus or if diazotrophic cells are carbonenergy limited ratherthan nutrient limited they might accumulate more than onegenome copy per cell However this extrapolation can alsounderestimate diazotrophic abundances as it is known thatDNA and RNA extractions are not 100 efficient and mayvary among species (Foster et al 2009b) The extraction ef-ficiencies and even genome copies per cell are currently un-der investigation (J Zehr personal communication 2011)

In addition non-replicating deoxyribonucleic acid (DNA)can comprise up to 90 of the total DNA in the oligotrophicregions (Winn and Karl 1986) and DNA sampled in natu-ral environments may represent non-living (detrital or non-replicating) particulate matter (Holm-Hansen et al 1968Holm-Hansen 1969 Winn and Karl 1986 Bailiff and Karl1991 Arin et al 1999) There is also evidence suggestingthat the non-living DNA is less important than originallythought (Dortch et al 1983)

22 Log-normal distribution and quality control

Diazotrophic abundances and N2 fixation rates in the oceancan range from 0 when diazotrophs are below detection ortruly absent at that location and time to very high valuesduring diazotrophic bloom phases Cell abundances and N2

fixation rates hence vary by several orders of magnitudeare often not normally distributed and are positively skewed(long tail of high values) However the datasets (excluding



Table 2 Estimated default and the upper and lower bounds for the biomass conversion factors and their impacts on global biomass estimates(based on geometric mean)

Conversion Factor (pg C cellminus1) Global Biomass Estimatea (Tg C)default lower upper default lower upper

Cell-count-based biomass

Trichodesmiumb 300 100 500Richelia 10 2 80Calothrix 10 5 20

21 08 35

nifH-based biomass

Trichodesmium 300 100 500UCYN-A 02 01 03UCYN-B 20 4 50UCYN-C 10 5 24UCYN all groupsRichelia 10 2 80Calothrix 10 5 20

89 28 170

a The lowhigh biomass is calculated when all their subtypes use lowhigh conversion factorsb Assuming 100 cells trichomeminus1

zero-value data points) are approximately log-normally dis-tributed (Figs 2 and S1) as is typical of many biological andecological properties that are induced by biological mecha-nisms (Koch 1966 Campbell 1995) Mathematically cal-culating the (arithmetic) mean and standard error of the log-transformed data and then back-transforming them exponen-tially results in geometric mean macrxg and geometric standarderror SEg which should be used in estimating the mean andthe error of the mean for a log-normal distribution in formatof xgtimes SEg (xg multiplied and divided by SEg) instead ofusingxaplusmnSEa (xa arithmetic mean SEa arithmetic standarderror) (Limpert et al 2001 Doney et al 2009) Thus in thisstudy the error range for a geometric mean macrxg is representedas between macrxgSEg and xgtimesSEg

We control for data quality by using Chauvenetrsquos crite-rion to flag suspicious outliers (Glover et al 2011) whichgenerally applies to normally distributed datasets and rejectsdata whose probability of deviation from the mean is lessthan 1(2n) wheren is the number of data points Howeverconsidering (1) the datasets are log-normally distributed and(2) valid diazotrophic abundances and N2 fixation rates canbe infinite low or zero Chauvenetrsquos criterion in our practiceis applied to the log-transformed non-zero data to flag theoutliers only on the high side ThosenifH-based abundancesthat were reported as ldquodetected but not quantifiablerdquo arenot included in the application of Chauvenetrsquos criterion Al-though not used in the application of Chauvenetrsquos criteriondata with zero values are kept in the database as they repre-sent valuable ecological information Note that the criterionis processed separately for the volumetric data points of eachtaxonomic type The criterion is also applied to the depth-

integrated total N2 fixation rates and the depth-integrated to-tal biomass estimated from cell counts andnifH-based abun-dances as these depth-integrated values are used later for theglobal estimates First the mean macrxlog and the standard devi-ationσlog of the log-transformed data are calculated whichare used to calculate the critical valuexlowastlog with a probabil-ity of a half of 1(2n) that values would exceed the meanxlog by this amount assuming normal distribution (in log-transformed space) One-half of 1(2n) is used because Chau-venetrsquos criterion is two-tailed test and we only reject data atone tail on the high side Thus all data with log-transformedvalues higher than macrxlog+ xlowastlog are flagged

All the data points not flagged by the Chauvenetrsquos criterionare accepted However not all the suspicious outliers flaggedby the Chauvenetrsquos criterion have to be rejected as ourdatasets are not strictly log-normal distributed and the dis-tribution estimated from the existing samples may not wellrepresent the true distribution especially at the sites with un-usual environmental conditions Thus for each flagged out-lier we evaluate whether its extraordinarily high value isreasonable or spurious based on the specific environmentalconditions andor the discussion with the original data con-tributor

23 Biomass conversion

The cell counts andnifH-based abundances are convertedto C biomass using conversion factors (Table 2) As dis-cussed above for C biomass estimates all the counts ofTri-chodesmiumin colonies and cells have been converted tonumber of trichomes by assuming 200 trichomes colonyminus1



(a) (b)

(c) (d)

(e)

Figure 2 Histogram of data points on logarithmic scale (blue bars) and the critical values for quality control using Chauvenetrsquos criterion(dashed red lines) Values higher than the critical values are rejected(a) Trichodesmiumcell counts(b) TrichodesmiumN2 fixation rates(c) UCYN N2 fixation rates(d) whole seawater N2 fixation rates and(e) UCYN-B nifH-based abundance See Supplement Fig S1 forfigures for other types



Table 3 Biomass conversion factors estimated forTrichodesmiumcells based on size measurements of cultured Woods HoleTrichodesmiumspecies (Hynes et al 2012) Carbon contents are calculated using the Verity et al (1992) model

Biovolume (microm3) Carbon content (pg C cellminus1)

Species meanplusmn standard range meanplusmn standard rangedeviation deviation

T thiebautii 680plusmn 380 280ndash1200 120plusmn 57 55ndash190T tenue 610 110T pelagicum 1100plusmn 130 1000ndash1200 190plusmn 18 170ndash200T hildebrandtii 1600plusmn 140 1500ndash1700 250plusmn 9 240ndash250T erythraeum 340plusmn 200 190ndash920 65plusmn 32 42ndash160T contortum 1300plusmn 51 1300ndash1400 210plusmn 7 210ndash220

and 100 cells trichomeminus1 which were widely used as an es-timate forTrichodesmium(LaRoche and Breitbarth 2005)To determine the biomass conversion factor for trichomesof Trichodesmium we utilize size measurements of culturedcells of differentTrichodesmiumspecies (Hynes et al 2012)and estimate carbon content using the model of Verity etal (1992) (Tables 3 and S1) The format of the Verity etal (1992) model is

C= 0433middotV0863 (1)

where C is the cell carbon content in pg andV is thecell volume inmicrom3 This model was used because it wasbased on many data points across a wide range of cell sizesand species type including cyanobacteriaT erythraeumisthe smallestTrichodesmiumspecies with estimated C con-tent of 65 pg C cellminus1 (Table 3) or 65 ng C trichomeminus1 (us-ing 100 cells trichomeminus1) which is consistent with the back-calculated 42plusmn1 pg C cellminus1 from measurements of the iron-to-carbon ratio and iron content per cell (Tuit et al 2004Goebel et al 2008) Thus the Verity et al (1992) modelused here appears to be suitable forTrichodesmium How-everT erythraeumis relatively small compared to otherTri-chodesmiumspecies The estimated carbon content for otherTrichodesmiumspecies is 110ndash250 pg C cellminus1 (Table 3) or11ndash25 ng C trichomeminus1 (using 100 cells trichomeminus1) Carboncontent forTrichodesmiumwas derived from elemental anal-ysis coupled with direct trichome counts at 40 stations in thetropical Atlantic in 1994 and 1996 (see the dataset associatedwith Carpenter et al 2004 in our database) yielding a con-version factor of 10plusmn12 ng C trichomeminus1 Some other stud-ies observed higher C contents forTrichodesmiumcoloniessuch as 97microg C colonyminus1 in the Pacific (Mague et al 1977)and 109 and 116microg C colonyminus1 for Trichodesmiumpuffsand tufts in Atlantic (McCarthy and Carpenter 1979) whichresults in a conversion factorsim50 ng C trichomeminus1 by assum-ing 200 trichomes colonyminus1 Another direct analysis ofTri-chodesmiumC content at the BATS station in the SargassoSea in 1995ndash1997 shows 83plusmn35microg C colonyminus1 (n= 23)for Trichodesmiumpuffs 99plusmn25microg C colonyminus1 (n= 14) forTrichodesmiumtufts and 45plusmn24microg C colonyminus1 (n= 6) for

T erythraeum(Orcutt and Gundersen unpublished data)which by assuming 200 trichomes colonyminus1 leads to a con-version factor of 41 49 and 22 ng C trichomeminus1 respectivelyThe synthesis of laboratory and field data by LaRoche andBreitbarth (2005) provides values of 50ndash58 ng C trichomeminus1To accommodate all these estimates we use default con-version factors of 300 pg C cellminus1 (30 ng C trichomeminus1 using100 cells trichomeminus1) for Trichodesmium with its range esti-mated to 100ndash500 pg C cellminus1 (Table 2)

The biomass conversion factor for UCYN-A is difficultto calculate because there is no isolate in culture UCYN-A cells are spherical in shape and the estimate ofsim1micromin diameter by Goebel et al (2008) is the only measure-ment for UCYN-A cell size which was determined usingfluorescence-activated cell sorting (FACS) coupled with realtime-qPCR This gives a cell size estimate of 05microm3 and adefault conversion factor of 02 pg C cellminus1 for UCYN-A us-ing the Verity et al (1992) model with its range estimatedby varying the default conversion factor byplusmn50 ie 01ndash03 pg C cellminus1 (Table 2)

UCYN-B (Crocosphaera) cells are also spherical in shapewith a reported diameter range of 3ndash5microm in laboratory iso-lates (Goebel et al 2008) and 3ndash8microm in natural samples andcultures (Webb et al 2009 Moisander et al 2010) ThusUCYN-B cells have a range in volume from 14ndash270microm3 andcellular carbon content from 4ndash50 pg C cellminus1 using the Ver-ity et al (1992) model (Table 2) By assuming a diameterof 5microm and thus a volume of 65microm3 we also estimate thedefault conversion factor of 20 pg C cellminus1 for UCYN-B

The only successful isolation and laboratory culture ofa UCYN-C strain designated TW3 (Taniuchi et al 2012)show that the cells are 25ndash30microm in width and 40ndash60microm inlength which gives cellular volume of 20ndash42microm3 and cellu-lar carbon content of 5ndash11 pg C cellminus1 for UCYN-C by usingthe Verity et al (1992) model However this range is an es-timate from only one UCYN-C strain For example thenifHgene of UCYN-C is most similar to thenifH gene of the ben-thic Cyanothece(Zehr 2011) and the cell dimensions of twoCyanothecestrains BH63 and BH68 have been reported as4ndash5microm width by 7ndash8microm length (Reddy et al 1993) which



Table 4 Biomass conversion factors estimated forRicheliaandCalothrixtrichomes based on biovolume measurements by Foster et al (2011)and assuming trichome composition of one heterocyst and three five or ten vegetative cells Carbon contents are calculated using the Verityet al (1992) model Numbers are meanplusmn standard deviation when applicable

Species Heterocyst Vegetative cell Vegetative cells Biomass conversionBiovolume C content Biovolume C content per trichome (pg C cellminus1)

(mum3) (pg C cellminus1) (microm3) (pg C cellminus1) mean range

Richelia (n= 26) 3 13plusmn 18 2ndash86110plusmn 180 22plusmn 31 42plusmn 76 9plusmn 14 5 12plusmn 16 2ndash78

10 11plusmn 15 2ndash71

Calothrix (n= 5) 3 12plusmn 5 7ndash1845plusmn 35 11plusmn 8 47plusmn 19 12plusmn 4 5 12plusmn 5 7ndash17


equals 90ndash160microm3 and leads to a conversion factor of 15ndash24 pg C cellminus1 by using the Verity et al (1992) model Bymerging these two estimates the default conversion factorof 10 pg C cellminus1 with a range of 5ndash24 pg C cellminus1 is used forUCYN-C (Table 2)

The trichomes ofRicheliaandCalothrix are comprised ofthree to ten vegetative cells and one terminal heterocyst (Jan-son et al 1999 Foster and Zehr 2006) The sizes of vege-tative cells and heterocysts have been measured for multipleRicheliaandCalothrix samples (Foster et al 2011) We usethese values to estimate the C contents of vegetative cellsand heterocysts by using the Verity et al (1992) model (Ta-bles 4 and S2) As the sizes and the C contents are differ-ent in vegetative cells and heterocysts average C content perRicheliaor Calothrixcell is calculated by assuming each tri-chome is comprised of one heterocyst and three five or tenvegetative cells (Table 4) The number of vegetative cells pertrichome does not greatly impact the estimate of average Ccontent per cell (Table 4) Based on these estimates a defaultbiomass conversion factor of 10 pg C cellminus1 with a range of2ndash80 pg C cellminus1 is used forRichelia and of 10 pg C cellminus1

with a range of 5ndash20 pg C cellminus1 for Calothrix (Table 2)

3 Results and discussion

31 Results of quality control

Most data types are well approximated by a log-normal dis-tribution (Figs 2 and S1) except forCalothrix cell countsandCalothrix nifH-based abundances which only have lim-ited non-zero data points (Fig S1b and h) By applying Chau-venetrsquos criterion there are only 9 data points flagged in-cluding 1 volumetricTrichodesmiumcell count (Fig 2a)1 volumetricTrichodesmiumN2 fixation (Fig 2b) 2 volu-metric N2 fixation rates by UCYN (Fig 2c) 2 volumetricwhole seawater N2 fixation rate (Fig 2d) and 3 volumetricUCYN-B nifH genes (Fig 2e) Thus from a statistical pointof view most data are acceptable The flagged 1 data pointof Trichodesmiumcell count of 44times1010 trichomes mminus3

and 1 data point ofTrichodesmiumN2 fixation rate of31 391micromol N mminus3 dminus1 were sampled simultaneously in thewestern Indian Ocean near the Kenyan coast (Fig 3a and b)(Kromkamp et al 1997) The authors reported that a mas-sive bloom which seemed to be associated with a front wasencountered and large streaks withTrichodesmiumcolonieswere floating at the surface As both the high abundance andN2 fixation rate were observed these two data points werevery likely real The flagged data points of UCYNrsquos N2 fix-ation rate of 360 and 960micromol N mminus3 dminus1 (Montoya et al2004) were fromsim10 Ssim130ndash135 E the western tropicalPacific in the Arafura Sea (Fig 3b) High N2 fixation ratesare commonly found in this region such as another flaggedwhole seawater N2 fixation rate of 610micromol N mminus3 dminus1 (Bon-net et al 2009) found in the nearby sea region at 6 S147 E near Papua New Guinea (Fig 3b) Recent mea-surements in the southwestern Pacific near New Caledoniaalso show extremely high N2 fixation rates (Bonnet unpub-lished data) The possible reason for these high N2 fixationrates could be the iron supply from volcanoes or via up-welling Another flagged whole seawater N2 fixation rateof 13 500micromol N mminus3 dminus1 (Gandhi et al 2011) was sam-pled from surface water atsim17 N 73 E near the Indiancoast (Fig 3b) In the same vertical profile the N2 fixa-tion rates sampled in 2 m and deeper were all less than32micromol N mminus3 dminus1 and all the other surface N2 fixation ratessampled in the surrounding area on this cruise were below540micromol N mminus3 dminus1 We evaluate that this high rate is ex-traordinary and should be removed from the database Theflagged three data points of UCYNnifH-based abundancesof sim2ndash9times1011 copies mminus3 (Orcutt et al unpublished data)are from the Mississippi Sound (Fig 3c) although most othernifH gene data were sampled from the open ocean Missis-sippi Sound is a very shallow partially land-locked coastalenvironment and is very different from open ocean watersAlso the three flagged data points of the extreme peaks innifH-based abundance were sampled in June and July 2009during extremely high water temperature events (sim30C) to-wards the end of the summer This is also a time of the year



(a)

(b)

(c)

Figure 3 Spatial distributions of collected diazotrophic data in number of data points (binned on to 1 times1 grid) including cell counts(panela) N2 fixation rates (panelb) and nifH-based abundances (panelc) Blue diamonds mark the location for the data rejected byChauvenetrsquos criterion

when the lowest seasonal dissolved inorganic N P-ratios(approximately 05) occurred although the ratio increases to-wards 3ndash6 by the end of the late summer period (likely dueto N2 fixation) The extreme peaks innifH-based abundancealso coincided with peaks in phytoplankton chlorophylla (5ndash6microg lminus1) Thus we believe most flagged data points exceptthe high surface N2 fixation rate found near the Indian coast

(Gandhi et al 2011) are due to the specific environment andnot necessarily related to the data quality Hence we haveretained these values in the database These points howeverare not included in our later analyses because their extremelyhigh values would influence the mean values It is also im-portant to note that we have not excluded data based on an



assessment of the protocols of sampling handling preserva-tion or measurement

32 Data distribution

Figure 3 shows the spatial distribution of the three sub-databases The Atlantic Ocean has the best data coverage inall three sub-databases especially in the North Atlantic Inthe Pacific Ocean the coverage is limited especially in theSouth Pacific There is almost no data coverage for the In-dian Ocean except four datasets in the Arabian Sea (Caponeet al 1998 Mazard et al 2004) near the Kenyan coast(Kromkamp et al 1997) and in the Madagascar Basin (Poul-ton et al 2009) There are also some data points in innerseas N2 fixation rates were measured almost every month atthe BATS station in 1995ndash1997 and at Station ALOHA from2005ndash2010 These are the only two sites with long-term sus-tained time series of N2 fixation measurements

Most data were collected in the tropical and subtropi-cal regions with latitudinal coverage of 50 Sndash50 N forcell counts of diazotrophs 40 Sndash60 N for N2 fixation ratesand 30 Sndash60 N for nifH-based abundances (Fig 4) Mostdata were collected in the 1990s and 2000s with some cellcounts and N2 fixation rates collected in 1960s and 1970sand very limited data points of cell counts collected in 1980s(Fig 5a c and e) The monthly distribution of the data tendsto be random for cell counts (Fig 5b) and N2 fixation rates(Fig 5d) while most ofnifH gene data were collected inMarch April and July (Fig 5f)


After being binned onto 3 times3 grids and calculating the ge-ometric means of the data in each bin depth-integrated N2

fixation rates are found to be highest in the western trop-ical Atlantic near the Caribbean Sea and in the subtropi-cal North Pacific near the Hawaiian Islands on an order of100ndash1000micromol N mminus2 dminus1 (Fig 6a) In most other regionsdepth-integrated N2 fixation rates are on an order of 1ndash100micromol N mminus2 dminus1 (Fig 6a)

The volumetric N2 fixation rates are also analyzed on3 times3 grids in the five vertical layers of 0ndash5 m 5ndash25 m25ndash625 m 625ndash1375 m and 1375ndash250 m The geometricmean values of each grid box are illustrated in Fig 6bndashfNote that the depth-integrated and volumetric values are notfrom exactly the same data sources They overlap for somedata sources but also have locations that are not in com-mon as some of the data were originally reported as depth-integrated values and the other depth-integrated values werecalculated herein N2 fixation rates generally decrease withdepth (Fig 6bndashf) When compared horizontally in each layerthe database reveals that there are several regions with highN2 fixation rates the subtropical North Pacific in all the lay-ers surface waters of the western Pacific and the tropical At-lantic (especially in the west) in 0ndash25 m (Fig 6bndashf)

(a)

(b)

(c)

Figure 4 Latitudinal distribution of the data for(a) cell counts fordiazotrophs(b) N2 fixation rates and(c) nifH-based abundances

34 Diazotrophic abundance and biomass

Cell counts andnifH-based abundances are used to estimatediazotrophic C biomass using the default biomass conver-sion factors (Table 2) except for a few datasets where the



Figure 5 Temporal distribution of the data points in year andmonth for (a)ndash(b) cell counts for diazotrophs(c)ndash(d) N2 fixationrates and(e)ndash(f) nifH-based abundances

contributors measured the conversion factors The cell countdata demonstrate thatTrichodesmiumis the dominant dia-zotroph except that the abundance ofCalothrix can be aslarge asTrichodesmium(Table 5) But the number of depth-integrated cell count samples forCalothrix are very lim-ited (Table 5) and are limited to one cruise in the subtropi-cal North Pacific (Villareal et al unpublished) As the av-erageTrichodesmiumcell size and thus the biomass con-

version factor is much larger than heterocystous cyanobac-teria (Table 2)Trichodesmiumconstitute more than 97 of diazotrophic biomass based on cell count data (Table 5)Cell count data do not include UCYN groups The geometricmean ofnifH-based abundances shows that UCYN-A is themost abundant among all the diazotrophic groups includingTrichodesmium and the abundances of UCYN-B UCYN-C and Richelia are also comparable toTrichodesmiuminthe volumetric data (Table 5) HoweverTrichodesmiumstilldominate the diazotrophic biomass (Table 5) because oftheir higher biomass conversion factor than other groupsComparison of the geometric means shows thatnifH-basedabundances are mostly one order of magnitude higher thancell-count-based abundances forTrichodesmiumand het-erocystous cyanobacteria except that the volumetricTri-chodesmiumabundances are comparable in both cell-count-based andnifH-based geometric means (Table 5) In theNorth Atlantic Ocean where both cell counts andnifH-basedabundances were frequently measured (Fig 3a and c) his-tograms ofTrichodesmiumabundances derived from boththe cell count data (assuming 100 cells trichomeminus1) andnifH-based data are in agreement (Fig 7) This evidence providessome support for the reliability ofnifH-based abundance forTrichodesmium

For comparison the arithmetic mean and standard errorare also calculated for each diazotrophic group (Table 5)Note that zero-value data points have to be excluded for cal-culating geometric means while these data points can beused to calculate arithmetic means The arithmetic meansare mostly one to several orders of magnitude larger thanthe geometric means especially in thenifH-based datasets(Table 5) The estimated biomass of UCYN groups increasesgreatly and becomes comparable to that ofTrichodesmiumifthe arithmetic means are used (Table 5) which however iscertainly due to the high values within the approximate log-normal distributions dominating the calculation of the arith-metic mean

Total diazotrophic C biomass estimated from cell counts isdisplayed in spatial maps for a given depth-integral and forthe five vertical layers of 0ndash5 m 5ndash25 m 25ndash625 m 625ndash1375 m and 1375ndash250 m by showing the geometric meansof each 3times3 grid bin (Fig 8andashf) The depth-integrated cell-count-based diazotrophic biomass is higher in the westernthan in the eastern Atlantic (Fig 8a) The depth-integratedcell-count-based diazotrophic biomass is also high in thesubtropical North Pacific near Hawairsquoi while it is low inother regions of the subtropical North Pacific because mostlyonly heterocystous cyanobacteria were counted in thesesampling sites (Fig 8a) The cell-count-based diazotrophicbiomass shows maxima in the surface and decreases withdepth (Fig 8bndashf) The surface cell-count-based diazotrophicbiomass is high in the tropical Atlantic (Fig 8b and c) whichis consistent with the high N2 fixation rates found in the sameregion In the southern Atlantic the cell-count-based dia-zotrophic biomass is low in all layers (Fig 8bndashf) The high



(a) (b)

(c) (d)

(e) (f)

Figure 6 (a) Geometric mean depth-integrated N2 fixation rates binned on 3times3 grid Also geometric mean N2 fixation rates on 3times3 gridfor 5 vertical layers of(b) 0ndash5 m(c) 5ndash25 m(d) 25ndash625 m(e) 625ndash1375 m and(f) 1375ndash250 m The color bar is in logarithmic scaleZero values are marked with black triangles

cell-count-based diazotrophic C biomass in the Arabian Seais from one dataset (Capone et al 1998) reporting an exten-sive Trichodesmiumbloom which may not represent meanlevel of diazotrophic biomass in the Arabian Sea

The total diazotrophic biomass estimated fromnifH-basedabundances is presented for a given depth-integral and forthe five vertical layers of 0ndash5 m 5ndash25 m 25ndash625 m 625ndash1375 m and 1375ndash250 m by showing the geometric meansof each 3 times3 grid bin (Fig 9andashf) Both the depth-integralsand results for vertical layers show highnifH-based Cbiomass in the tropical Atlantic (Fig 9andashf) which is consis-tent with the high N2 fixation rates found in the same regionHigh nifH-based biomass was also found in the southwesternPacific (Fig 9a c and d) where the diazotrophic cell countdata are not reported ThenifH-based biomass is generallyhigh in 0ndash625 m (Fig 9bndashd) and decreases below 625 m(Fig 9e and f)

35 First-order estimates for global N2 fixation rate anddiazotrophic biomass

Many analyses can be conducted withsim12 000 data points inthe database as per the objectives of the users Here we showa simple example using the database to conduct first-orderestimates of the global N2 fixation rate and diazotrophicbiomass We divide the global ocean into 6 regions the Northand the South Atlantic Ocean the North and the South Pa-cific Ocean the Indian Ocean and the Mediterranean Sea(Tables 6 7 and 8) Two methods geometric mean and arith-metic mean were used for the estimation As the data arenot evenly distributed in space and intensive samplings weremade in some regions (Fig 3) the depth-integrated valuesare first binned to 3 times3 grid to partially avoid this biasGeometric and arithmetic means are calculated for each binwhich are then used to calculate geometric and arithmeticmeans for each region Areal sum of the total N2 fixation



Table 5 Abundances and estimated biomass (using default conversion factors in Table 2) of each diazotrophic group from cell counts andnifH-based data shown by geometric and arithmetic mean Note that data points with zero-values are excluded when calculating geometricmeans Error ranges for geometric means shown in parentheses are estimated by dividing and multiplying the geometric means by onegeometric standard error Standard errors for arithmetic means are shown in parentheses Depth-integrated and volumetric data points areanalyzed separately

Depth-integrated Data n nof zero-value data Abundance (109 cells mminus2) Biomass (mg C mminus2)Geometric Arithmetic Geometric Arithmetic

cell-count-based data

Trichodesmiumlowast 620 59 016 (014ndash018) 12 (012) 43 (38ndash49) 310 (27)Richelia 288 41 0010 (0008ndash0013) 33 (088) 0099 (0078ndash013) 33 (88)Calothrix 17 9 0084 (0050ndash014) 0096 (0050) 084 (050ndash14) 096 (050)

nifH-based data

Trichodesmium 103 10 14 (11ndash19) 88 (15) 400 (310ndash530) 2600 (440)UCYN-A 104 24 15 (088ndash24) 3400 (2200) 029 (018ndash048) 680 (450)UCYN-B 103 20 011 (0071ndash016) 31 (093) 21 (14ndash32) 63 (19)UCYN-C 31 18 0055 (0020ndash016) 032 (016) 055 (020ndash16) 33 (16)Richelia 63 2 024 (015ndash038) 34 (12) 24 (15ndash39) 340 (120)Calothrix 11 0 021 (0093ndash046) 058 (015) 21 (093ndash46) 58 (15)

Volumetric Data n nof zero-value data Abundance (106 cells mminus3) Biomass (microg C mminus3)Geometric Arithmetic Geometric Arithmetic



nifH-based data

Trichodesmium 759 163 28 (25ndash32) 87 (12) 850 (740ndash970) 26 000 (3700)UCYN-A 771 334 55 (45ndash68) 23 000 (9400) 11 (089ndash14) 4600 (1900)UCYN-B 792 322 11 (095ndash13) 280 (75) 22 (19ndash27) 5600 (1500)UCYN-C 226 141 12 (090ndash17) 48 (11) 12 (90ndash17) 48 (11)Richelia 535 160 16 (13ndash19) 240 (63) 16 (13ndash19) 2400 (630)Calothrix 64 52 033 (017ndash062) 024 (010) 33 (17ndash62) 24 (10)

lowast Assuming 100 cells trichomeminus1

rate for each region is calculated by multiplying geometricor arithmetic means with ocean area of that region Note thatthe volumetric data points are not used in this simple exam-ple although they can also provide valuable information forthe global estimates Global N2 fixation rate and diazotrophicbiomass are then estimated by summing the estimates fromall the 6 regions As 99 of the data in the database were col-lected within a latitudinal span ofsim40 Sndash55 N we assumethe oceanic N2 fixation is negligible outside of this latitudi-nal range Given known temperature constraints on diazotro-phy this seems like a reasonable assumption however onecould further refine global estimates if significant N2 fixationis found outside this latitudinal range in the future

In the estimates for N2 fixation rate geometric mean ofthe depth-integrated rates for the North Atlantic Ocean sur-prisingly is lower than that for the North and the SouthPacific (Table 6) although the North Atlantic has attracted

most studies of diazotrophs By using the geometric meansthe total N2 fixation rate in the North Atlantic is estimatedto 17 Tg N yrminus1 one order lower than the estimates for theNorth and the South Pacific Ocean (35 and 24 Tg N yrminus1respectively) (Table 6) The arithmetic mean of N2 fixa-tion rate in the North Atlantic however is one order higherthan the geometric mean which makes the total arithmeticrate of 32 Tg N yrminus1 more comparable to the North Pacific(56 Tg N yrminus1) and the South Pacific (46 Tg N yrminus1) (Table 6)The data indicate that although high N2 fixation rates wereobtained in the North Atlantic especially in the western trop-ical North Atlantic (Fig 6a) low N2 fixation rates weremore frequently found in this region More measurementsare needed to confirm the high estimates of N2 fixation ratefor the Pacific Ocean as the measurements in the North Pa-cific were mostly in the subtropical gyre near the HawairsquoiIslands (Fig 6a) and the sampling in the South Pacific has



been limited to data mostly from two cruises (Garcia et al2007 Raimbault and Garcia 2008) The N2 fixation rateswere also low in the South Atlantic (Table 6) There are inad-equate depth-integrated N2 fixation rates collected in the In-dian Ocean to estimate the means of this basin By summingthe geometric mean rates from all the regions the global N2

fixation rate (not including the Indian Ocean) is estimated tobe 62 (error range 53ndash73) Tg N yrminus1 (Table 6) which is atthe low end of the current geochemical estimates of 100ndash200 Tg N yrminus1 for marine pelagic N2 fixation (Gruber andSarmiento 1997 2002 Karl et al 2002 Galloway et al2004 Deutsch et al 2007 Gruber 2008) Note that as de-scribed in Sect 22 the error range for a geometric meanin this study is represented as between the geometric meanmultiplieddivided by the corresponding geometric standarderror The estimate of global N2 fixation rate using arith-metic mean is higher at 140 (standard error 92) Tg N yrminus1

(Table 6) which is within the range of the current geochem-ical estimates

In a manner similar to our global estimates of the N2 fix-ation rate we have also calculated global mean diazotrophicbiomass from cell-count-based biomass data (Table 7) Asin many sampling sites in the Pacific Ocean only hetero-cystous cyanobacteria are counted (Table 1a) and thus verylow total diazotrophic biomass was reported for these sites(Fig 8a) We give special consideration for the Pacific by es-timating mean biomass forTrichodesmiumand heterocystouscyanobacteria separately before summing the values fromthese two taxonomies (Table 7) The average diazotrophicbiomass in the North Atlantic is similar to or higher thanthat in the North Pacific as shown by both geometric andarithmetic means (Table 7) The global diazotrophic biomass(not including Mediterranean Sea) is estimated from the cell-count-based data as 21 (error range 14ndash31) Tg C using ge-ometric mean or 18 (standard error 18) Tg C using arith-metic mean (Table 7)

The global diazotrophic biomass is also estimated fromdepth-integratednifH-based data First the depth-integrateddata points ofnifH-based biomass are limited with no or in-adequate data points in the South Atlantic the Indian Oceanand the Mediterranean Sea (Table 8) and thus the estimatecould be highly biased The geometric meannifH-based di-azotrophic biomass is extremely high in the South Pacificcompared to other regions (Table 8) The global estimate(the Pacific plus the North Atlantic) of diazotrophic biomassfrom nifH-based data is 89 (error range 52ndash150) Tg C us-ing geometric mean or 590 (standard error 70) Tg C (usingarithmetic mean) which are dominated by the estimate of theSouth Pacific (Table 8)

By comparing the global estimate of N2 fixation rates withthe global estimates of cell-count-based ornifH-based di-azotrophic biomass (using geometric means) (Tables 6 7and 8) and assuming molar C N ratio of 106 16 for dia-zotrophic cells the turnover time of diazotrophic cellular N

(a)

(b)

Figure 7 Histogram of depth-integrated (panela) and volumet-ric (panelb) non-zero data points of cell-count-based (blue) andnifH-based (red)Trichodesmiumabundances in the North Atlantic(10 Sndash50 N) Each Trichodesmiumtrichome is assumed to becomprised of 100 cells Data values are on logarithmic scale Thoseldquodetected but not quantifiablerdquonifH-based abundances which areassigned 10times103 cells mminus3 (see text for details) are not included

due to N2 fixation is estimated to be 2 days and 92 days re-spectively

36 Limitations

There are a number of limitations with the database and theglobal estimates First and foremost these data points are notuniformly distributed in the world ocean Historically sci-entists have also sought out regions with high diazotrophicabundances and hence there is a higher possibility of artifi-cially elevated N2 fixation activity and diazotrophic biomassrelative to true regional means As the current database doesnot cover some regions such as the coastal upwelling zones



(a) (b)

(c) (d)

(e) (f)

Figure 8 (a) Geometric mean depth-integrated cell-count-based diazotrophic carbon biomass binned on 3 times3 grid Also geometric meancell-count-based diazotrophic carbon biomass on 3 times3 grid for 5 vertical layers of(b) 0ndash5 m(c) 5ndash25 m(d) 25ndash625 m(e)625ndash1375 mand(f) 1375ndash250 m The color bar is in logarithmic scale Zero values are marked with black triangles

our estimate could be changed substantially if values differwithin these biogeographic domains

As mentioned above it was recently established that themost commonly applied method used to measure N2 fixa-tion the 15N2 assimilation method (Table 1b) has under-estimated the true rates because the15N2 bubbles injectedin seawater do not attain equilibrium with surrounding wa-ter (Mohr et al 2010) Two independent method compar-isons conducted in the Pacific and Atlantic Oceans demon-strated that the measured rates are significantly higher whenthe 15N2 tracer is added in a dissolved form to the incuba-tions resulting in an instantaneous15N isotopic equilibriumin the seawater (Groszligkopf et al 2012 Wilson et al 2012)The variability in the underestimation was linked to the com-position of the diazotrophic community (Mohr et al 2010Groszligkopf et al 2012) indicating that it will be difficult tocorrect the historical rate measurements made with the15N2

method However these measurements can be treated as min-

imum estimates that may increase in the future These mea-surements are still valuable when comparing N2 fixation ratesin different regions as they are approximately log-normallydistributed and range insim6 orders of magnitude (Fig 1)

It is also a notable issue that the difference between thegeometric and the arithmetic mean N2 fixation rate is largerin the North Atlantic than those in the other basins (Table 6)A strict log-normal distribution is symmetric in the log spaceand its geometric mean is located at the peak of the distri-bution ie at the highest probability However the depth-integrated total N2 fixation rates in our database after binnedon 3 times3 grid are not strictly log-normally distributed ineach basin (Fig 10) The North Atlantic has highest spatialcoverage and the N2 fixation rates in this basin differ in 6 or-ders of magnitude from 10minus3ndash103 micromol N mminus2 dminus1 (Fig 10)Its distribution is not symmetric with more values observedon the left (lower) side Thus the geometric mean N2 fixationrate for the North Atlantic is one order of magnitude lower



(a) (b)

(c) (d)

(e) (f)

Figure 9 (a) Geometric mean depth-integratednifH-based diazotrophic carbon biomass binned on 3 times3 grid Also geometric meannifH-based diazotrophic carbon biomass on 3 times3 grid for 5 vertical layers of(b) 0ndash5 m (c) 5ndash25 m(d) 25ndash625 m(e) 625ndash1375 m and(f) 1375ndash250 m The color bar is in logarithmic scale Zero values are marked with black triangles

Table 6 Estimates of N2 fixation rate for the global oceans based on geometric and arithmetic means Note that data points with zero-values are excluded when calculating geometric means Depth-integrated N2 fixation rates are first binned to 3 times3 grid and geometricand arithmetic means are calculated for each bin which are then used to calculate geometric and arithmetic means for each region Arealsum is calculated by multiplying geometric or arithmetic means with ocean area Error ranges for geometric means shown in parenthesesare estimated by dividing and multiplying the geometric means by one geometric standard error Standard errors for arithmetic means areshown in parentheses Error range for global geometric mean is estimated by simply summing lower and upper bounds of the error ranges ofeach region Errors for global arithmetic means are propagated from (independent) errors in each region ND no data NQ not qualified tocalculate areal sum because of too few samples

Region Latitudinal Ocean area n of n of zero- N2 fixation rate (micromol N mminus2 dminus1) Areal Sum (Tg N yrminus1)range (times1012 m2) bins value bins Geometric Arithmetic Geometric Arithmetic

North Atlantic Ocean 0ndash55 N 36 125 0 92 (70ndash12) 170 (40) 17 (13ndash22) 32 (74)South Atlantic Ocean 40ndash0 S 27 15 0 79 (63ndash98) 13 (44) 11 (09ndash14) 18 (06)North Pacific Ocean 0ndash55 N 89 45 3 78 (67ndash90) 120 (22) 35 (30ndash41) 56 (98)South Pacific Ocean 40ndash0 S 72 26 0 64 (54ndash76) 130 (46) 24 (20ndash28) 46 (17)Indian Ocean 40 Sndash25 N 62 4 0 120 (29ndash490) 590 (320) NQ NQMediterranean Sea 30ndash45 N 25 10 0 18 (12ndash28) 45 (21) 02 (01ndash04) 06 (03)Global 62 (52ndash73) 137 (92)



Table 7 Estimates of diazotrophic biomass for the global oceans using cell-count-based data Same method as Table 6 is used except thatthe estimates forTrichodesmiumand Heterocystous cyanobacteria (CYN) are calculated separately in the North and South Pacific OceanND no data NQ not qualified to calculate areal sum because of too few samples

Region Latitudinal Ocean area n of n of zero- Cell-count-based diazotrophic Areal Sum (Tg C)range (times1012 m2) bins value bins biomass (mg C mminus2)

Geometric Arithmetic Geometric Arithmetic

North Atlantic Ocean 0ndash55 N 36 103 4 24 (18ndash32) 310 (66) 087 (067ndash12) 11 (24)South Atlantic Ocean 40ndash0 S 27 8 4 11 (67ndash18) 75 (39) 030 (018ndash050) 021 (011)

North Pacific Ocean 0ndash55 N 89Trichodesmium 52 15 10 (65ndash16) 71 (19)Heterocystous CYN 63 11 002 (002ndash003) 042 (012)Total 10 (66ndash16) 71 (16) 091 (058ndash14) 63 (15)

South Pacific Ocean 40ndash0 S 72Trichodesmium 4 1 011 (009ndash013) 009 (003)Heterocystous CYN 16 13 006 (004ndash011) 002 (001)Total 017 (013ndash025) 010 (001) 0013 (0009ndash0018) 0007 (0001)

Indian Ocean 40 Sndash25 N 62 2 0 190 (160ndash240) 350 (110) NQ NQMediterranean Sea 30ndash45 N 25 ND ND ND ND ND ND

Global 21 (14ndash31) 18 (18)

Table 8 Estimates of diazotrophic biomass for the global oceans usingnifH-based data Same method as Table 6 is used ND no data NQnot qualified to calculate areal sum because of too few samples

Region Latitudinal Ocean area n nof zero- nifH-based diazotrophic biomass Areal Sum (Tg C)range (times1012 m2) of bins value bins (mg C mminus2)


North Atlantic Ocean 0ndash55 N 36 41 6 290 (150ndash560) 3800 (1200) 11 (55ndash20) 140 (44)South Atlantic Ocean 40ndash0 S 27 ND ND ND ND ND NDNorth Pacific Ocean 0ndash55 N 89 12 0 170 (78ndash380) 1700 (630) 15 (69ndash34) 150 (56)South Pacific Ocean 40ndash0 S 72 22 0 880 (560ndash1400) 4200 (1500) 63 (40ndash100) 310 (110)Indian Ocean 40 Sndash25 N 62 4 0 33 (16ndash69) 81 (63) NQ NQMediterranean Sea 30ndash45 N 25 ND ND ND ND ND NDGlobal 89 (52ndash150) 590 (70)

than the peak of the distribution The arithmetic mean in thisbasin is more than one order of magnitude higher than thegeometric mean In the North and South Pacific the N2 fix-ation rates only differ in 2 orders of magnitude without anydata lower than 10micromol N mminus2 dminus1 (Fig 10) Thus the differ-ence between the geometric and arithmetic means is smallerin the Pacific than in the North Atlantic Although the geo-metric mean N2 fixation rate in the North Atlantic is aboutone order of magnitude lower than those in the North andSouth Pacific the peaks of the distributions are actually muchcloser in these basins (Fig 10 also see Table 6) Comparedto the North Atlantic the Pacific is not intensively sampledfor N2 fixation Thus we need more samples with wider spa-tial coverage from the Pacific to better evaluate the global N2

fixation rateBased on depth-integrated data the global mean estimate

of diazotrophic biomass from cell counts is 1ndash2 orders ofmagnitude lower than that fromnifH-based abundances (Ta-bles 7 and 8) which are mostly contributed by the dom-

inant taxonTrichodesmium(Table 5) But considering thelog-normal distributions difference of one or two orders ofmagnitude is relatively small compared to the span of thedistribution especially when the number of samples is lim-ited (see Fig 7 as an example) More samples are needed inorder to confidently validate the usage ofnifH data for esti-mating diazotrophic abundances and the assumption of 1 1conversion fromnifH to cell densities

The above global estimates are based on default values ofthe biomass conversion factors In order to show the effectsof the variation of the biomass conversion factors we haveapplied the upper and lower bounds of these conversion fac-tors for all diazotrophic subtypes in order to revise global di-azotrophic biomass estimates following the same proceduredescribed above The estimated global geometric mean di-azotrophic biomass can vary by aboutplusmn70 from defaultestimates ranging in 08ndash35 Tg C 9 (default 21 Tg C) basedon cell count data or in 28ndash170 Tg C (default 89 Tg C) basedonnifH-based data (Table 2)



Figure 10 Histograms of depth-integrated total N2 fixation rates(after binned on 3 times3 grid) in North Atlantic (blue) South At-lantic (green) North Pacific (red) and South Pacific (black) Foreach basin filled diamond and empty circle mark the location ofthe geometric mean and the arithmetic mean respectively

4 Conclusions and recommendations for use

The first global database of oceanic diazotrophic measure-ments has been constructed with sub-databases for N2 fixa-tion rates cell counts for diazotrophs andnifH-based abun-dances This database provides useful information on thespatial patterns of N2 fixation and diazotrophic biomass inthe world ocean Depth-integrated values were used to con-duct a first-order estimate of the global N2 fixation rate anddiazotrophic biomass Spatial coverage is the main limita-tion of the database as there are still vast oceanic areaswhere diazotrophic activity and biomass have never beenmeasured For example measurements in the Indian Oceanthe South Atlantic and the subtropical southern Pacific area high priority in this regard A finer understanding of therange and applicability of fixed biomass conversion factorsis another concern Although careful analyses have been con-ducted to derive a reliable global biomass our estimates mayvary by plusmn70 depending on the conversion factor Moredirect elemental analyses are required to further narrow therange of biomass conversion factors Relatively higher abun-dances tend to be found fromnifH-based data than fromthe cell count data The estimated global mean diazotrophicbiomass fromnifH-based abundances is 1ndash2 orders of mag-nitude higher than that derived from cell counts Howeverin the well-sampled region of the North Atlantic theTri-chodesmiumabundances estimated from these two types ofdata match each other better Further investigations and moredata are needed to identify what causes the discrepancies be-tween these two estimates However it is clear that micro-scopic cell counts cannot at present account for the UCYN-A which is very abundant in the ocean

With the recent findings that most of the previous N2 fix-ation rates have been underestimated when the15N2 tracerwas added to the incubations as gas bubbles it can be ex-pected that higher N2 fixation rates will be reported in the

near future using the new method by adding15N2 in a dis-solved form Each data record needs to be well documentedfor the applied15N2 method so that future users can identifywhich data records are likely underestimated

The database is stored permanently at PANGAEA and canbe easily accessed by users It will be routinely updated withnew measurements The database can be used to study thelevel of oceanic N2 fixation activity and its temporal and spa-tial variations on local regional and global scales as well asfor constraining the relative contribution of new nitrogen in-puts from N2 fixation and other sources The diazotrophicbiomass data along with data from other functional groupscan be used to study phytoplankton community structureThe database can also be used to validate geochemical es-timates of N2 fixation and to parameterize and validate bio-geochemical models keeping in mind that N2 fixation ratemeasurements have been underestimating the true N2 fixa-tion rates

Supplementary material related to this article isavailable online athttp wwwearth-syst-sci-datanet4472012essd-4-47-2012-supplementpdf

Acknowledgements Y-WL SCD DB MJC DMKPHM KAT AEW and JPZ are funded by C-MORE project(NSF EF-0424599) LAA and DJM gratefully acknowledgesupport of NSF OCE-0925284 MB acknowledges fundingfrom projects CAIBEX (CTM2007-66408-CO2-O2) and CYFOD(CTM2008-00915-E) and FPI fellowship (BES-2008-006985) byMICINN (Spain) SB is funded by IRD and data are contributionsof the BOUM experiment and ISOFeRIX Project of the Frenchnational LEFE-CYBER program the European IP SESAME andthe international IMBER project MJC RML and DMKacknowledge NSF OCE OCE-08-50827 for supporting N2 fixationmeasurements at Station ALOHA RJL and JL are funded by theGerman BMBF project Sopran DMK PHM KAT and JPZare funded by the Gordon and Betty Moore Foundation CMM isfunded by NERC UK BM-C was supported by the Ramon y Ca-jal program from the Spanish Ministry of Science and TechnologyAPR received funding through NERC OCEANS2025 AEWacknowledges funding from NSF OCE-0726422 (to F Prahl OSU)We thank Nicolas Gruber and another anonymous referee for theirconstructive reviews

Edited by S Pesant

References

Arin L Berdalet E Marrase C Estrada M Guixa-BoixereuN and Dolan J Particulate DNA and protein relative to mi-croorganism biomass and detritus in the Catalano-Balearic Sea(NW Mediterranean) during stratification J Plankton Res 211299ndash1316 1999

Bailiff M D and Karl D M Dissolved and Particulate DNA Dy-namics during a Spring Bloom in the Antarctic Peninsula Region



1986ndash87 Deep-Sea Res 38 1077ndash1095doi1010160198-0149(91)90097-Y 1991

Bar-Zeev E Yogev T Man-Aharonovich D Kress N Herut BBeja O and Berman-Frank I Seasonal dynamics of the en-dosymbiotic nitrogen-fixing cyanobacteriumRichelia intracel-lularis in the eastern Mediterranean Sea Isme J 2 911ndash923doi101038ismej200856 2008

Benavides M Agawin N S R Arıstegui J Ferriol P and StalL J Nitrogen fixation byTrichodesmiumand small diazotrophsin the subtropical northeast Atlantic Aquat Microb Ecol 6543ndash53doi103354ame01534 2011

Bonnet S Guieu C Bruyant F Prasil O Van WambekeF Raimbault P Moutin T Grob C Gorbunov M YZehr J P Masquelier S M Garczarek L and ClaustreH Nutrient limitation of primary productivity in the South-east Pacific (BIOSOPE cruise) Biogeosciences 5 215ndash225doi105194bg-5-215-2008 2008

Bonnet S Biegala I C Dutrieux P Slemons L O and CaponeD G Nitrogen fixation in the western equatorial Pacific Ratesdiazotrophic cyanobacterial size class distribution and biogeo-chemical significance Global Biogeochem Cy 23 GB3012doi1010292008GB003439 2009

Bonnet S Grosso O and Moutin T Planktonic dinitrogen fixa-tion along a longitudinal gradient across the Mediterranean Seaduring the stratified period (BOUM cruise) Biogeosciences 82257ndash2267doi105194bg-8-2257-2011 2011

Borstad G A Some aspects of the occurrence and biology ofTri-chodesmium(cyanobacteria) in the western tropical Atlantic nearBarbados West Indies Doctor of Philosophy Marine ScienceCentre McGill University Montreal Quebec Canada 234 pp1978

Bostrom K H Riemann L Zweifel U L and Hagstrom AringNodularia spnifH gene transcripts in the Baltic Sea proper JPlankton Res 29 391ndash399doi101093planktfbm019 2007

Brzezinski M A Villareal T A and Lipschultz F Silica pro-duction and the contribution of diatoms to new and primary pro-duction in the central North Pacific Mar Ecol-Prog Ser 16789ndash104doi103354meps167089 1998

Buitenhuis E T Bednarsek N Doney S C Dunne J PLeblanc K Luo Y-W Moriarty R OrsquoBrien T D SchiebelR Peloquin J A and Vogt M MAREDAT Towards a worldocean atlas of marine ecosystem data Earth Syst Sci Data Dis-cuss in preparation 2012

Campbell J W The lognormal distribution as a model for bio-optical variability in the sea J Geophys Res-Oceans 10013237ndash13254doi10102995JC00458 1995

Capone D G Determination of nitrogenase activity in aquaticsamples using the acetylene reduction procedure in Handbookof Methods in Aquatic Microbial Ecology edited by Kemp PF Cole J J Sherr B F and Sherr E B Lewis PublishersBoca Raton Fla 621ndash631 1993

Capone D G The marine nitrogen cycle in Microbial ecologyof the ocean edited by Kirchman D Wiley-Liss New York455ndash493 2000

Capone D G Zehr J P Paerl H W Bergman B and Carpen-ter E JTrichodesmium a globally significant cyanobacteriumScience 276 1221-1229doi101126science276531612211997

Capone D G Subramaniam A Montoya J P Voss M Hum-borg C Johansen A M Siefert R L and Carpenter EJ An extensive bloom of the N2-fixing cyanobacteriumTri-chodesmium erythraeumin the central Arabian Sea Mar Ecol-Prog Ser 172 281ndash292doi103354meps172281 1998

Capone D G Burns J A Montoya J P Subramaniam AMahaffey C Gunderson T Michaels A F and Carpen-ter E J Nitrogen fixation byTrichodesmiumspp An im-portant source of new nitrogen to the tropical and subtropicalNorth Atlantic Ocean Global Biogeochem Cy 19 GB2024doi1010292004GB002331 2005

Carpenter E J Nitrogen fixation by marineOscillatoria (Tri-chodesmium) in the worldrsquos oceans in Nitrogen in the MarineEnvironment edited by Carpenter E J and Capone D G Aca-demic Press New York 65ndash103 1983

Carpenter E J and Capone D G Nitrogen Fixation in the MarineEnvironment in Nitrogen in the Marine Environment edited byCapone D G Bronk D A Mulholland M R and CarpenterE J Academic Press London 1197ndash1218 2008

Carpenter E J and Price C C Nitrogen fixation distributionand production ofOscillatoria(Trichodesmium) Spp in the west-ern Sargasso and Caribbean Seas Limnol Oceanogr 22 60ndash721977

Carpenter E J and Romans K Major role of the cyanobacteriumTrichodesmiumin nutrient cycling in the North Atlantic OceanScience 254 1356ndash1358doi101126science254503613561991

Carpenter E J Subramaniam A and Capone D G Biomassand primary productivity of the cyanobacteriumTrichodesmiumspp in the tropical N Atlantic Ocean Deep-Sea Res Pt I 51173ndash203doi101016jdsr200310006 2004

Chang J Chiang K P and Gong G C Seasonal variation andcross-shelf distribution of the nitrogen-fixing cyanobacteriumTrichodesmium in southern East China Sea Cont Shelf Res20 479ndash492doi101016S0278-4343(99)00082-5 2000

Chen Y L L Chen H Y and Lin Y H Distribution and down-ward flux ofTrichodesmiumin the South China Sea as influencedby the transport from the Kuroshio Current Mar Ecol-ProgSer 259 47ndash57doi103354meps259047 2003

Chen Y L L Chen H Y Tuo S H and Ohki K Sea-sonal dynamics of new production fromTrichodesmiumN2

fixation and nitrate uptake in the upstream Kuroshio andSouth China Sea basin Limnol Oceanogr 53 1705ndash1721doi104319lo20085351705 2008

Chen Y L L Tuo S H and Chen H Y Co-occurrence andtransfer of fixed nitrogen fromTrichodesmiumspp to diatoms inthe low-latitude Kuroshio Current in the NW Pacific Mar Ecol-Prog Ser 421 25ndash38doi103354Meps08908 2011

Church M J Jenkins B D Karl D M and Zehr J P Verticaldistributions of nitrogen-fixing phylotypes at Stn ALOHA in theoligotrophic North Pacific Ocean Aquat Microb Ecol 38 3ndash14doi103354ame038003 2005a

Church M J Short C M Jenkins B D Karl D M and ZehrJ P Temporal patterns of nitrogenase gene (nifH) expression inthe oligotrophic North Pacific Ocean Appl Envir Microbiol71 5362ndash5370doi101128Aem7195362-53702005 2005b

Church M J Mahaffey C Letelier R M Lukas RZehr J P and Karl D M Physical forcing of nitrogenfixation and diazotroph community structure in the North



Pacific subtropical gyre Global Biogeochem Cy 23 GB2020doi1010292008gb003418 2009

Deutsch C Sarmiento J L Sigman D M Gruber N andDunne J P Spatial coupling of nitrogen inputs and losses in theocean Nature 445 163ndash167doi101038nature05392 2007

Doney S C Lima I Moore J K Lindsay K Behrenfeld M JWestberry T K Mahowald N Glover D M and TakahashiT Skill metrics for confronting global upper ocean ecosystem-biogeochemistry models against field and remote sensing dataJ Marine Syst 76 95ndash112doi101016jjmarsys2008050152009

Dong J D Zhang Y Y Wang Y S Zhang S and Wang HK Spatial and seasonal variations of Cyanobacteria and theirnitrogen fixation rates in Sanya Bay South China Sea Sci Mar72 239ndash251 2008

Dore J E Brum J R Tupas L M and Karl D M Seasonaland interannual variability in sources of nitrogen supporting ex-port in the oligotrophic subtropical North Pacific Ocean Lim-nol Oceanogr 47 1595ndash1607doi104319lo200247615952002

Dortch Q Roberts T L Clayton J R and Ahmed S IRNADNA ratios and DNA concentrations as indicators ofgrowth rate and biomass in planktonic marine organisms MarEcol-Prog Ser 13 61ndash71 1983

Dugdale R C Goering J J and Ryther J H Nitrogen fixation inthe Sargasso Sea Deep-Sea Res 7 298ndash300doi1010160146-6313(61)90051-X 1961

Falcon L I Carpenter E J Cipriano F Bergman Band Capone D G N2 fixation by unicellular bacterio-plankton from the Atlantic and Pacific oceans Phylogenyand in situ rates Appl Environ Microbiol 70 765ndash770doi101128Aem702765-7702004 2004

Farnelid H Andersson A F Bertilsson S Abu Al-Soud WHansen L H Sorensen S Steward G F Hagstrom A andRiemann L Nitrogenase Gene Amplicons from Global MarineSurface Waters Are Dominated by Genes of Non-CyanobacteriaPlos One 6 e19223doi101371journalpone0019223 2011

Fernandez A Mourino-Carballido B Bode A Varela M andMaranon E Latitudinal distribution ofTrichodesmiumspp andN2 fixation in the Atlantic Ocean Biogeosciences 7 3167ndash3176doi105194bg-7-3167-2010 2010

Fernandez C Farıas L and Ulloa O Nitrogen Fixa-tion in Denitrified Marine Waters Plos One 6 e20539doi101371journalpone0020539 2011

Fernandez A Grana R Mourino-Carballido B Bode AVarela M Domınguez J F Escanez J de Armas D andMaranon E Community N2 fixation andTrichodesmiumsppabundance along longitudinal gradients in the Eastern Subtropi-cal North Atlantic in preparation 2012

Fong A A Karl D M Lukas R Letelier R M Zehr J Pand Church M J Nitrogen fixation in an anticyclonic eddyin the oligotrophic North Pacific Ocean ISME J 2 663ndash676doi101038ismej200822 2008

Foster R A and OrsquoMullan G D Nitrogen-fixing and nitrifyingSymbioses in the marine environment in Nitrogen in the Ma-rine Environment edited by Capone D G Bronk D A Mul-holland M R and Carpenter E J Academic Press London1197ndash1218 2008

Foster R A and Zehr J P Characterization of diatom-cyanobacteria symbioses on the basis ofnifH hetR and16S rRNA sequences Environ Microb 8 1913ndash1925doi101111j1462-2920200601068x 2006

Foster R A Subramaniam A Mahaffey C Carpenter E JCapone D G and Zehr J P Influence of the Amazon Riverplume on distributions of free-living and symbiotic cyanobac-teria in the western tropical north Atlantic Ocean LimnolOceanogr 52 517ndash532doi104319lo20075220517 2007

Foster R A Paytan A and Zehr J P Seasonality of N2 fixationandnifH gene diversity in the Gulf of Aqaba (Red Sea) LimnolOceanogr 54 219ndash233doi104319lo20095410219 2009a

Foster R A Subramaniam A and Zehr J P Distributionand activity of diazotrophs in the Eastern Equatorial At-lantic Environ Microbiol 11 741ndash750doi101111j1462-2920200801796x 2009b

Foster R A Kuypers M M M Vagner T Paerl R WMusat N and Zehr J P Nitrogen fixation and transfer in openocean diatom-cyanobacterial symbioses ISME J 5 1484ndash1493doi101038ismej201126 2011

Galloway J N Dentener F J Capone D G Boyer E WHowarth R W Seitzinger S P Asner G P Cleveland CC Green P A Holland E A Karl D M Michaels A FPorter J H Townsend A R and Vorosmarty C J Nitrogencycles past present and future Biogeochemistry 70 153ndash226doi101007s10533-004-0370-0 2004

Gandhi N Singh A Prakash S Ramesh R Raman MSheshshayee M S and Shetye S First direct measure-ments of N2 fixation during aTrichodesmiumbloom in theeastern Arabian Sea Global Biogeochem Cy 25 Gb4014doi1010292010gb003970 2011

Garcia N Raimbault P and Sandroni V Seasonal nitro-gen fixation and primary production in the Southwest Pa-cific nanoplankton diazotrophy and transfer of nitrogen topicoplankton organisms Mar Ecol-Prog Ser 343 25ndash33doi103354Meps06882 2007

Glover D M Jenkins W J and Doney S C Modeling Methodsfor Marine Science Cambridge University Press CambridgeUK 2011

Goebel N L Edwards C A Carter B J Achilles K Mand Zehr J P Growth and carbon content of three different-sized diazotrophic cyanobacteria observed in the subtropicalNorth Pacific J Phycol 44 1212ndash1220doi101111j1529-8817200800581x 2008

Goebel N L Turk K A Achilles K M Paerl R Hewson IMorrison A E Montoya J P Edwards C A and Zehr JP Abundance and distribution of major groups of diazotrophiccyanobacteria and their potential contribution to N2 fixation inthe tropical Atlantic Ocean Environ Microbiol 12 3272ndash3289doi101111j1462-2920201002303x 2010

Goering J J Dugdale R C and Menzel D W Estimates ofinsitu rates of nitrogen uptake byTrichodesmiumsp in the tropicalAtlantic Ocean Limnol Oceanogr 11 614ndash620 1966

Gomez F Furuya K and Takeda S Distribution of thecyanobacterium Richelia intracellularis as an epiphyte of the di-atom Chaetoceros compressus in the western Pacific Ocean JPlankton Res 27 323ndash330doi101093planktfbi007 2005

Grabowski M N W Church M J and Karl D M Nitrogenfixation rates and controls at Stn ALOHA Aquat Microb Ecol



52 175ndash183doi103354Ame01209 2008Groszligkopf T Mohr W Baustian T Schunck H Gill D

Kuypers M M M Lavik G Schmitz R A RWallace DW and LaRoche J Doubling of marine dinitrogen-fixationrates based on direct measurements Nature 488 361ndash364doi101038nature11338 2012

Gruber N The Marine Nitrogen Cycle Overview and Challengesin Nitrogen in the marine environment 2nd Edn edited byCapone D G Bronk D A Mulholland M R and Carpen-ter E J Elsevier Amsterdam 1ndash50 2008

Gruber N and Sarmiento J L Global patterns of marine nitrogenfixation and denitrification Global Biogeochem Cy 11 235ndash266 1997

Gruber N and Sarmiento J L Large-scale biogeochemicalphys-ical interactions in elemental cycles in The Sea Biological-Physical Interactions in the Sea edited by Robinson A R Mc-Carthy J J and Rothschild B J John Wiley amp Sons New YorkCity 2002

Gundersen K R Corbin J S Hanson C L Hanson M L Han-son R B Russell D J Stollar A and Yamadas O Structureand biological dynamics of the oligotrophic ocean photic zoneoff the Hawaiian Islands Pac Sci 30 45ndash68 1976

Hegde S Anil A C Patil J S Mitbavkar S Krishnamurthy Vand Gopalakrishna V V Influence of environmental settings onthe prevalence ofTrichodesmiumspp in the Bay of Bengal MarEcol-Prog Ser 356 93ndash101doi103354Meps07259 2008

Holm-Hansen O Determination of microbial biomass in oceanprofiles Limnol Oceanogr 14 740ndash747 1969

Holm-Hansen O Sutcliffe Jr W H and Sharp J Measurementof Deoxyribonucleic Acid in the Ocean and its Ecological Sig-nificance Limnol Oceanogr 13 507ndash514 1968

Hulburt E M Phytoplankton in thesSouthewestern Sagasso Seaand North Equatorial current Limnol Oceanogr 7 307ndash3151962

Hulburt E M The distribution of phytoplankton and its rela-tionship to hydrography between southern New England andVenezuela J Mar Res 24 67ndash81 1966

Hulburt E M Phytoplankton observations in the westernCaribbean Sea B Mar Sci 18 388ndash399 1968

Hynes A M Webb E A Doney S C and Waterbury J BComparison of culturedTrichodesmium(Cyanophyceae) withspecies characterized from the field J Phycol 48 196ndash210doi101111j1529-8817201101096x 2012

Janson S Wouters J Bergman B and Carpenter E JHost specificity in theRichelia-diatom symbiosis revealed byhetR gene sequence analysis Environ Microbiol 1 431ndash438doi101038ismej201126 1999

Karl D Michaels A Bergman B Capone D Carpenter ELetelier R Lipschitz F Paerl H Sigman D and StalL Dinitrogen fixation in the worldrsquos oceans Biogeochemistry5758 47ndash98doi101023A1015798105851 2002

Karl D M Letelier R Tupas L Dore J Christian J andHebel D The role of nitrogen fixation in biogeochemical cy-cling in the subtropical North Pacific Ocean Nature 388 533ndash538 1997

Kitajima S Furuya K Hashihama F Takeda S and Kanda JLatitudinal distribution of diazotrophs and their nitrogen fixationin the tropical and subtropical western North Pacific LimnolOceanogr 54 537ndash547doi104319lo20095420537 2009

Koch A L Logarithm in Biology 1 Mechanisms GeneratingLog-Normal Distribution Exactly J Theor Biol 12 276ndash290doi1010160022-5193(66)90119-6 1966

Kong L Jing H Kataoka T Sun J and Liu H Phylogeneticdiversity and spatio-temporal distribution of nitrogenase genes(nifH) in the northern South China Sea Aquat Microb Ecol65 15ndash27doi103354ame01531 2011

Kromkamp J De Bie M Goosen N Peene J Van Ri-jswijk P Sinke J and Duineveld G C A Primary pro-duction by phytoplankton along the Kenyan coast during theSE monsoon and November intermonsoon 1992 and the occur-rence ofTrichodesmium Deep-Sea Res Pt II 44 1195ndash1212doi101016S0967-0645(97)00015-5 1997

Langlois R J LaRoche J and Raab P A Diazotrophicdiversity and distribution in the tropical and subtropicalAtlantic ocean Appl Environ Microbiol 71 7910ndash7919doi101128Aem71127910-79192005 2005

Langlois R J Hummer D and LaRoche J Abundances anddistributions of the dominantnifH phylotypes in the North-ern Atlantic Ocean Appl Environ Microbiol 74 1922ndash1931doi101128Aem01720-07 2008

LaRoche J and Breitbarth E Importance of the diazotrophs asa source of new nitrogen in the ocean J Sea Res 53 67ndash91doi101016jseares200405005 2005

Letelier R M and Karl D M Role ofTrichodesmiumspp in theproductivity of the subtropical North Pacific Ocean Mar Ecol-Prog Ser 133 263ndash273doi103354meps133263 1996

Limpert E Stahel W A and Abbt M Log-normal distributionsacross the sciences Keys and clues Bioscience 51 341ndash352doi1016410006-3568(2001)051[0341LNDATS]20CO22001

Lugomela C Lyimo T J Bryceson I Semesi A K andBergman BTrichodesmiumin coastal waters of Tanzania di-versity seasonality nitrogen and carbon fixation Hydrobiologia477 1ndash13doi101023A1021017125376 2002

Mague T H Weare N M and Holm-Hansen O Nitrogen fixa-tion in the North Pacific Ocean Mar Biol 24 109ndash119 1974

Mague T H Mague F C and Holm-Hansen O Physiologyand chemical composition of nitrogen-fixing phytoplankton inthe central North Pacific Ocean Mar Biol 41 213ndash227 1977

Marumo R and Asaoka O Distribution of pelagic blue-green al-gae in the North Pacific Ocean Journal of the OceanographicalSociety of Japan 30 77ndash85doi101007bf02112896 1974

Mazard S L Fuller N J Orcutt K M Bridle O and ScanlanD J PCR analysis of the distribution of unicellular cyanobacte-rial diazotrophs in the Arabian Sea Appl Envir Microbiol 707355ndash7364doi101128Aem70127355-73642004 2004

McCarthy J J and Carpenter E JOscillatoria (Trichodesmium)thiebautii (cyanophyta) in the central North Atlantic OceanJ Phycol 15 75ndash82doi101111j1529-88171979tb02965x1979

Mohr W Grosskopf T Wallace D W R and LaRoche JMethodological underestimation of oceanic nitrogen fixationrates Plos One 5 e12583doi101371journalpone00125832010

Moisander P H Beinart R A Voss M and Zehr J P Di-versity and abundance of diazotrophic microorganisms in theSouth China Sea during intermonsoon ISME J 2 954ndash967doi101038ismej200851 2008



Moisander P H Beinart R A Hewson I White A E JohnsonK S Carlson C A Montoya J P and Zehr J P Unicellularcyanobacterial distributions broaden the oceanic N2 fixation do-main Science 327 1512ndash1514doi101126science11854682010

Montoya J P Voss M Kahler P and Capone D G A sim-ple high-precision high-sensitivity tracer assay for N2 fixationAppl Environ Microbiol 62 986ndash993 1996

Montoya J P Houll C M Zehr J P Hansen A Villareal TA and Capone D G High rates of N2 fixation by unicellu-lar diazotrophs in the oligotrophic Pacific Ocean Nature 4301027ndash1032doi101038nature02824 2004

Moore C M Mills M M Achterberg E P Geider R JLaRoche J Lucas M I McDonagh E L Pan X Poulton AJ Rijkenberg M J A Suggett D J Ussher S J and Wood-ward E M S Large-scale distribution of Atlantic nitrogen fix-ation controlled by iron availability Nat Geosci 2 867ndash871doi101038ngeo667 2009

Mourino-Carballido B Grana R Fernandez A Bode AVarela M Domınguez J F Escanez J Armas D D andMaranon E Importance of N2 fixation vs nitrate eddy diffu-sion along a latitudinal transect in the Atlantic Ocean LimnolOceanogr 56 999ndash1007doi104319lo20115630999 2011

Mulholland M R The fate of nitrogen fixed by diazotrophs inthe ocean Biogeosciences 4 37ndash51doi105194bg-4-37-20072007

Mulholland M R Bronk D A and Capone D G Dinitrogenfixation and release of ammonium and dissolved organic nitro-gen byTrichodesmiumIMS101 Aquat Microb Ecol 37 85ndash94doi103354ame037085 2004

Mulholland M R Bernhardt P W Blanco-Garcia J L Man-nino A Hyde K Mondragon E Turk K Moisander P Hand Zehr J P Rates of dinitrogen fixation and the abundanceof diazotrophs in North American coastal waters between CapeHatteras and Georges Bank Limnol Oceanogr 57 1067ndash1083doi104319lo20125741067 2012

Needoba J A Foster R A Sakamoto C Zehr J P and JohnsonK S Nitrogen fixation by unicellular diazotrophic cyanobac-teria in the temperate oligotrophic North Pacific Ocean Lim-nol Oceanogr 52 1317ndash1327doi104319lo200752413172007

Orcutt K M Lipschultz F Gundersen K Arimoto R MichaelsA F Knap A H and Gallon J R A seasonal study of the sig-nificance of N2 fixation byTrichodesmiumspp at the BermudaAtlantic Time-series Study (BATS) site Deep-Sea Res Pt II 481583ndash1608doi101016S0967-0645(00)00157-0 2001

Postgate J R The Fundamentals of Nitrogen Fixation CambridgeUniversity Press New York 1998

Poulton A J Stinchcombe M C and Quartly G D Highnumbers ofTrichodesmiumand diazotrophic diatoms in thesouthwest Indian Ocean Geophys Res Lett 36 L15610doi1010292009GL039719 2009

Rahav E Herut B Stambler N Bar-Zeev E Gertman I Mul-holland R M and Berman-Frank I Summertime dinitrogenfixation in a gyre and eddy of the eastern Mediterranean SeaGeophys Res Lett in review 2011

Rahav E Herut B Levi A Mulholland M and Berman-FrankI Controls on surface dinitrogen fixation across the Mediter-ranean during springtime in preparation 2012a

Rahav E Herut B Levi A Mulholland M and Berman-FrankI Dinitrogen fixation at the stratified and mixed periods of theGulf of Aqaba (Red Sea) in preparation 2012b

Raimbault P and Garcia N Evidence for efficient regenerated pro-duction and dinitrogen fixation in nitrogen-deficient waters of theSouth Pacific Ocean impact on new and export production esti-mates Biogeosciences 5 323ndash338doi105194bg-5-323-20082008

Reddy K J Haskell J B Sherman D M and Sherman LA Unicellular Aerobic Nitrogen-Fixing Cyanobacteria of theGenus Cyanothece J Bacteriol 175 1284ndash1292 1993

Rees A P Law C S and Woodward E M S High rates of ni-trogen fixation during an in-situ phosphate release experiment inthe Eastern Mediterranean Sea Geophys Res Lett 33 L10607doi1010292006GL025791 2006

Rees A P Gilbert J A and Kelly-Gerreyn B A Nitrogen fix-ation in the western English Channel (NE Atlantic Ocean) MarEcol-Prog Ser 374 7ndash12doi103354meps07771 2009

Riemann L Farnelid H and Steward G F Nitrogenase genesin non-cyanobacterial plankton prevalence diversity and reg-ulation in marine waters Aquat Microb Ecol 61 225ndash237doi103354ame01431 2010

Sandroni V Raimbault P Migon C Garcia N andGouze E Dry atmospheric deposition and diazotrophy assources of new nitrogen to northwestern Mediterranean olig-otrophic surface waters Deep-Sea Res Pt I 54 1859ndash1870doi101016jdsr200708004 2007

Sato S Maryse N R and Eskinazi E On the mechanism of redtide of Trichodesmiumin Recife northwest Brazil with someconsiderations of the relation to the human disease ldquoTamadarefeverrdquo Trab Inst Oceanogr Univ Recife 5 7ndash49 1966

Scranton M I Hydrogen cycling in the waters near Bermuda therole of the nitrogen fixerOsciflutoriu thiebuutii Deep-Sea Res31 133ndash143 1984

Shiozaki T Furuya K Kodama T Kitajima S Takeda S Take-mura T and Kanda J New estimation of N2 fixation in thewestern and central Pacific Ocean and its marginal seas GlobalBiogeochem Cy 24 GB1015doi1010292009GB0036202010

Sohm J A Subramaniam A Gunderson T E Carpen-ter E J and Capone D G Nitrogen fixation byTri-chodesmiumspp and unicellular diazotrophs in the North Pa-cific Subtropical Gyre J Geophys Res-Biogeo 116 G03002doi1010292010JG001513 2011a

Sohm J A Webb E A and Capone D G Emerging patternsof marine nitrogen fixation Nat Rev Microbiol 9 499ndash508doi101038nrmicro2594 2011b

Steven D M and Glombitza R Oscillatory variation of a phyto-plankton population in a tropical ocean Nature 237 105ndash107doi101038237105a0 1972

Sundstrom B G Observations onRhizosolenia cleveiOsten-feld (Bacillariophyceae) andRichelia intracellularis Schmidt(Cyanophyceae) Bot Mar 27 345ndash355 1984

Taniuchi Y Chen Y-l L Chen H-Y Tsai M-L and OhkiK Isolation and characterization of the unicellular diazotrophiccyanobacterium Group C TW3 from the tropical western PacificOcean Environ Microbiol 14 641ndash654doi101111j1462-2920201102606x 2012



Tuit C Waterbury J and Ravizzaz G Diel variation of molyb-denum and iron in marine diazotrophic cyanobacteria LimnolOceanogr 49 978ndash990doi104319lo20044940978 2004

Turk K A Rees A P Zehr J P Pereira N Swift P Shel-ley R Lohan M Woodward E M S and Gilbert J Ni-trogen fixation and nitrogenase (nifH) expression in tropicalwaters of the eastern North Atlantic ISME J 5 1201ndash1212doi101038ismej2010205 2011

Tyrrell T The relative influences of nitrogen and phospho-rus on oceanic primary production Nature 400 525ndash531doi10103822941 1999

Tyrrell T Maranon E Poulton A J Bowie A R Harbour DS and Woodward E M S Large-scale latitudinal distributionof Trichodesmiumspp in the Atlantic Ocean J Plankton Res25 405ndash416doi101093plankt254405 2003

Venrick E L Distribution and Significance ofRichelia Intracellu-laris Schmidt in North Pacific Central Gyre Limnol Oceanogr19 437ndash445 1974

Verity P G Robertson C Y Tronzo C R Andrews M G Nel-son J R and Sieracki M E Relationships between cell vol-ume and the carbon and nitrogen content of marine photosyn-thetic nanoplankton Limnol Oceanogr 37 1434ndash1446 1992

Villareal T A Division Cycles in the Nitrogen-FixingRhizosole-nia (Bacillariophyceae)-Richelia(Nostocaceae) Symbiosis BritPhycol J 24 357ndash365doi101080000716189006503711989

Villareal T A Laboratory Culture and Preliminary Char-acterization of the Nitrogen-FixingRhizosolenia-RicheliaSymbiosis Mar Ecol 11 117ndash132doi101111j1439-04851990tb00233x 1990

Villareal T A Marine nitrogen-fixing diatom-cyanobacteria sym-bioses in Marine Pelagic CyanobacteriaTrichodesmiumandother diazotrophs edited by Carpenter E J Capone D G andRueter J G NATO ASI Series 362 Kluwer Academic Publish-ers Boston 163ndash175 1992

Villareal T A Adornato L Wilson C and Schoenbaechler CA Summer blooms of diatom-diazotroph assemblages and sur-face chlorophyll in the North Pacific gyre A disconnect J Geo-phys Res-Oceans 116 C03001doi1010292010jc0062682011

Watkins-Brandt K S Letelier R M Spitz Y H Church MJ Bottjer D and White A E Addition of inorganic or or-ganic phosphorus enhances nitrogen and carbon fixation in theoligotrophic North Pacific Mar Ecol-Prog Ser 432 17ndash29doi103354meps09147 2011

Webb E A Ehrenreich I M Brown S L Valois F W and Wa-terbury J B Phenotypic and genotypic characterization of mul-tiple strains of the diazotrophic cyanobacterium Crocosphaerawatsonii isolated from the open ocean Environ Microbiol 11338ndash348doi101111j1462-2920200801771x 2009

White A E Foster R A Benitez-Nelson C R Masque P Ver-deny E Popp B N Arthur K E and Prahl F G Nitrogenfixation in the Gulf of California and the Eastern Tropical NorthPacific Prog Oceanogr submitted 2012

Wilson S T Bottjer D Church M J and Karl D M Compara-tive assessment of nitrogen fixation methodologies conducted inthe oligotrophic North Pacific Ocean Appl Environ Microbiol78 6516ndash6523doi101128AEM01146-12 2012

Winn C D and Karl D M Diel Nucleic-Acid Synthesis and Par-ticulate DNA Concentrations ndash Conflicts with Division Rate Esti-mates by DNA Accumulation Limnol Oceanogr 31 637ndash6451986

Wu J F Chung S W Wen L S Liu K K Chen YL L Chen H Y and Karl D M Dissolved inorganicphosphorus dissolved iron andTrichodesmiumin the olig-otrophic South China Sea Global Biogeochem Cy 17 1008doi1010292002gb001924 2003

Yogev T Rahav E Bar-Zeev E Man-Aharonovich D Stam-bler N Kress N Beja O Mulholland M R Herut B andBerman-Frank I Is dinitrogen fixation significant in the Lev-antine Basin East Mediterranean Sea Environ Microbiol 13854ndash871doi101111j1462-2920201002402x 2011

Zehr J P Nitrogen fixation by marine cyanobacteria Trends Mi-crobiol 19 162ndash173doi101016jtim201012004 2011

Zehr J P Mellon M T and Zani S New nitrogen-fixing mi-croorganisms detected in oligotrophic oceans by amplification ofnitrogenase (nifH) genes Appl Environ Microbiol 64 3444ndash3450 1998

Zehr J P Waterbury J B Turner P J Montoya J P OmoregieE Steward G F Hansen A and Karl D M Unicellularcyanobacteria fix N2 in the subtropical North Pacific Ocean Na-ture 412 635ndash638doi10103835088063 2001

Zehr J P Montoya J P Jenkins B D Hewson I MondragonE Short C M Church M J Hansen A and Karl D MExperiments linking nitrogenase gene expression to nitrogen fix-ation in the North Pacific subtropical gyre Limnol Oceanogr52 169ndash183doi104319lo20075210169 2007

Zehr J P Bench S R Carter B J Hewson I Ni-azi F Shi T Tripp H J and Affourtit J P Glob-ally Distributed Uncultivated Oceanic N2-Fixing Cyanobacte-ria Lack Oxygenic Photosystem II Science 322 1110ndash1112doi101126science1165340 2008



23University of Southampton National Oceanography Centre Southampton European Way Southampton SO14 3ZH UK24Department of Ocean Earth amp Atmospheric Sciences Old Dominion University Norfolk VA 23529 USA

25Division of Environmental and Biomolecular Systems Oregon Health amp Science University Portland Oregon 97239 USA26National Oceanography Centre Waterfront Campus European Way Southampton SO14 3ZH UK

27Plymouth Marine Laboratory Prospect Place The Hoe Plymouth PL1 3DH UK28Marine Biological Section University of Copenhagen 3000 Helsingoslashr Denmark

29Lamont Doherty Earth Observatory Palisades New York 10964 USA30Marine Science Institute University of Texas at Austin Port Aransas Texas 78382 USA

31Rosenstiel School of Marine and Atmospheric Science University of Miami Miami FL 33149 USAlowastcurrently at Department of Marine Sciences University of Georgia Athens GA 30602 USA

lowastlowastcurrently at Department of Biology University of Massachusetts Dartmouth North Dartmouth Massachusetts 02747 USA

Correspondence to Y-W Luo (yluowhoiedu)

Received 27 December 2011 ndash Published in Earth Syst Sci Data Discuss 13 February 2012Revised 25 July 2012 ndash Accepted 14 August 2012 ndash Published 31 August 2012

Abstract Marine N2 fixing microorganisms termed di-azotrophs are a key functional group in marine pelagicecosystems The biological fixation of dinitrogen (N2) tobioavailable nitrogen provides an important new source ofnitrogen for pelagic marine ecosystems and influences pri-mary productivity and organic matter export to the deepocean As one of a series of efforts to collect biomass andrates specific to different phytoplankton functional groupswe have constructed a database on diazotrophic organismsin the global pelagic upper ocean by compiling about 12 000direct field measurements of cyanobacterial diazotroph abun-dances (based on microscopic cell counts or qPCR assaystargeting thenifH genes) and N2 fixation rates Biomass con-version factors are estimated based on cell sizes to convertabundance data to diazotrophic biomass The database is lim-ited spatially lacking large regions of the ocean especiallyin the Indian Ocean The data are approximately log-normaldistributed and large variances exist in most sub-databaseswith non-zero values differing 5 to 8 orders of magnitudeReporting the geometric mean and the range of one geomet-ric standard error below and above the geometric mean thepelagic N2 fixation rate in the global ocean is estimated to be62 (52ndash73) Tg N yrminus1 and the pelagic diazotrophic biomass inthe global ocean is estimated to be 21 (14ndash31) Tg C fromcell counts and to 89 (43ndash150) Tg C fromnifH-based abun-dances Reporting the arithmetic mean and one standard errorinstead these three global estimates are 140plusmn92 Tg N yrminus118plusmn18 Tg C and 590plusmn70 Tg C respectively Uncertaintiesrelated to biomass conversion factors can change the estimateof geometric mean pelagic diazotrophic biomass in the globalocean by aboutplusmn70 It was recently established that themost commonly applied method used to measure N2 fixationhas underestimated the true rates As a result one can expectthat future rate measurements will shift the mean N2 fixationrate upward and may result in significantly higher estimatesfor the global N2 fixation The evolving database can never-theless be used to study spatial and temporal distributions

and variations of marine N2 fixation to validate geochemi-cal estimates and to parameterize and validate biogeochem-ical models keeping in mind that future rate measurementsmay rise in the future The database is stored in PANGAEA(doi101594PANGAEA774851)

1 Introduction

N2 fixation is the biological conversion of dinitrogen (N2) gasinto two molecules of ammonia by diazotrophic organismsOver geological time scales N2 fixation is important for reg-ulating fixed N concentrations in the ocean and thereby sus-taining ocean fertility (Tyrrell 1999) The rate of pelagic N2

fixation in the contemporary ocean has been estimated to be100ndash200 Tg nitrogen (N) yrminus1 which constitutes about halfof the total external source of bioavailable N to the ocean(Gruber and Sarmiento 1997 2002 Karl et al 2002 Gal-loway et al 2004 Deutsch et al 2007 Gruber 2008) It isgenerally accepted that cyanobacteria are the major N2-fixingmicroorganisms in the ocean (Karl et al 2002 Zehr 2011)However non-cyanobacterial prokaryotic plankton may alsoconduct N2 fixation in the ocean as revealed by the presenceand transcription ofnifH genes (encoding the iron proteincomponent of the nitrogenase enzyme) (Zehr et al 1998Riemann et al 2010 Farnelid et al 2011 Fernandez et al2011) albeit their relative contribution to global N2 fixationremains to be determined

There are three major types of marine planktonic dia-zotrophic cyanobacteria (Sohm et al 2011b Zehr 2011)(1) non-heterocystous filamentous cyanobacteria (2) hete-rocystous cyanobacteria and (3) unicellular cyanobacteriaAmong them most field and laboratory research has focusedon non-heterocystous filamentousTrichodesmium(Oscilla-toriales) (Dugdale et al 1961 Carpenter and Romans 1991Capone et al 1997)Trichodesmiumis an abundant dia-zotroph in the open ocean (Capone et al 2005 LaRocheand Breitbarth 2005) which can be found as large colonies









2 Data and methods





































Orcutt et al (2001)




Total 3274 0 1612 930 5326




















Capone et al (2005)




Falcon et al (2004)













Moore et al (2009)










unpublished data





Total 689 275 205 2234 757 3624








Bonnet et al (2008)


N Pacific(ALOHA)





Fong et al (2008)















Total 759 1793 599 416 3201









212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References


































































































































































2 Data and methods





































Orcutt et al (2001)




Total 3274 0 1612 930 5326




















Capone et al (2005)




Falcon et al (2004)













Moore et al (2009)










unpublished data





Total 689 275 205 2234 757 3624








Bonnet et al (2008)


N Pacific(ALOHA)





Fong et al (2008)















Total 759 1793 599 416 3201









212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References

























































































































































































Orcutt et al (2001)




Total 3274 0 1612 930 5326




















Capone et al (2005)




Falcon et al (2004)













Moore et al (2009)










unpublished data





Total 689 275 205 2234 757 3624








Bonnet et al (2008)


N Pacific(ALOHA)





Fong et al (2008)















Total 759 1793 599 416 3201









212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References












































































































































































Capone et al (2005)




Falcon et al (2004)













Moore et al (2009)










unpublished data





Total 689 275 205 2234 757 3624








Bonnet et al (2008)


N Pacific(ALOHA)





Fong et al (2008)















Total 759 1793 599 416 3201









212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References
































































































































































Bonnet et al (2008)


N Pacific(ALOHA)





Fong et al (2008)















Total 759 1793 599 416 3201









212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































212 Cell counts














75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































75

Figure 1

















21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References
































































































































































21 08 35

nifH-based biomass


89 28 170










(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































(a) (b)

(c) (d)

(e)









C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References

































































































































































C= 0433middotV0863 (1)
























(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References












































































































































































(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































(a)

(b)

(c)














(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




































































































































































(a)

(b)

(c)













(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References



































































































































































(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































(a) (b)

(c) (d)

(e) (f)












nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References
































































































































































nifH-based data





nifH-based data














(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References


































































































































































(a)

(b)



36 Limitations




(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































(a) (b)

(c) (d)

(e) (f)









(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




























































































































































(a) (b)

(c) (d)

(e) (f)














Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References



































































































































































Global 21 (14ndash31) 18 (18)




















Edited by S Pesant

References




































































































































































Edited by S Pesant

References



















































































































































































































































































































































































































































































































































































































Documents

DataY.-W. Luo et al.: Database of diazotrophs in global ocean 49 known as “puﬀs” or “tufts” or as free trichomes (Carpenter, 1983). Heterocystous cyanobacteria (Nostocales,