ArticlePDF Available

Abstract

Considerable uncertainty remains into how increasing atmospheric CO2 and anthropogenic climate changes are affecting open‐ocean marine ecosystems from phytoplankton to top predators. Biological time series data are thus urgently needed for the world’s oceans. Here, we use the carbon stable isotope composition of tuna to provide a first insight into the existence of global trends in complex ecosystem dynamics and changes in the oceanic carbon cycle. From 2000 to 2015, considerable declines in δ13C values of 0.8 to 2.5‰ were observed across three tuna species sampled globally, with more substantial changes in the Pacific Ocean compared to the Atlantic and Indian Oceans. Tunas not only recorded the Suess effect, i.e. fossil fuel‐derived and isotopically‐light carbon being incorporated into marine ecosystems, but also profound changes at the base of marine food webs. We suggest a global shift in phytoplankton community structure, e.g. a reduction of 13C‐rich phytoplankton such as diatoms, and/or a change in phytoplankton physiology during this period, while this does not prevent other concomitant changes at higher levels in the food webs. Our study establishes tuna δ13C values as a candidate essential ocean variable to assess complex ecosystem responses to climate change at regional to global scales and over decadal timescales. Finally, this time‐series will be invaluable in calibrating and validating global earth system models to project changes in marine biota.
Glob Change Biol. 2019;00:1–13. wileyonlinelibrary.com/journal/gcb  
|
 1
© 2019 John Wiley & Sons Ltd
Received:20Decem ber2018 
|
  Revised:2 0September2019 
|
  Accepted:30September2019
DOI : 10.1111/gcb .14858
PRIMARY RESE ARCH ARTICLE
Trends in tuna carbon isotopes suggest global changes in
pelagic phytoplankton communities
Anne Lorrain1| Heidi Pethybridge2| Nicolas Cassar1,3 | Aurore Receveur4|
Valérie Allain4| Nathalie Bodin5,6 | Laurent Bopp7| C. Anela Choy8|
Leanne Duffy9| Brian Fry10| Nicolas Goñi11| Brittany S. Graham12|
Alistair J. Hobday2| John M. Logan13 | Frederic Ménard14 |
Christophe E. Menkes15 | Robert J. Olson9| Dan E. Pagendam16 | David Point17 |
Andrew T. Revill2| Christopher J. Somes18 | Jock W. Young2
1IRD,CNRS ,Ifremer,LEMAR,UnivBrest,Plouzané ,France
2CSIROOceansandAtmosphere,Hobart ,Tas.,Australia
3Divisio nofEarthandOce anSciences,Ni cholasS chooloft heEnvironment ,DukeUni versit y,Durham,NC,USA
4Pacifi cCommunity,Ocea nicFish eriesProgram me,Nou méa,Fra nce
5IRD,FishingPort,Vic toria,Mahe,RepublicofS eychelles
6SeychellesFish ingAuth orit y(SFA),Vic toria ,Mahe,Re publicofSeychelles
7Labor atoiredeMétéorologieD ynamique(LMD),I nstitutPierr e‐SimonL aplace(IPSL),EcoleNor maleSupérieu re/PSLRe s.Univ.,CNRS,
EcolePoly technique,S orbon neUniversité,P aris,France
8IntegrativeOceanographyDivision,ScrippsInstitutionofOceanography,Unive rsityofCalifornia,SanDiego,LaJoll a,CA ,USA
9Inter‐Ameri canTropic alTunaCommis sion,(I ATTC),LaJolla,CA,USA
10AustralianR iversInstitu te,GriffithUniver sity,Nath an,Qld ,Australia
11AZ TI,MarineResearch,Pasaia ,Gipuzkoa,Spai n
12Nationa lInstituteofWaterandAtmos phericResearch,Ltd.(NIWA),Wellington,NewZeala nd
13Massa chuset tsDivisionofMarineFisheries,NewBedford,MA,US A
14AixMarseill eUniver sity,Universit yofToulon,CNRS,IRD,MIO,UM110,Mars eille,Fr ance
15IRD,UMRENT ROPIE,N ouméacedex,NewCale donia
16CSIRO,ComputationalInformatics,Brisb ane,Qld,Aust ralia
17Obser vatoireMidi‐P yréné es,GE T,UMRCNRS5563/IRD234,Université́PaulSabat ierToulouse3,Toulouse,France
18GEOMARH elmholtzCentreforOceanResearchKiel,Kie l,Germany
Correspondence
AnneLorrain,IRD,CNRS ,Ifrem er,LEMAR,
UnivBre st,F‐29280Plouzan é,France .
Email:anne.lorrain@ird.fr
Funding information
LabexMER,Grant/AwardNumber:ANR‐10‐
LABX‐19;FrenchGover nment
Abstract
Considerable uncertainty remains over how increasing atmospheric CO2 and
anthropoge nic climate changes are af fecting ope n‐ocean marine ecos ystems from
phytoplanktontotoppredators.Biologicaltimeseriesdataarethusurgentlyneeded
forthe world's oceans.Here,weusethecarbonstableisotopecompositionoftuna
toprovide afirstinsight intothe existence ofglobal trends in complex ecosystem
dynamicsand changes in the oceanic carbon cycle. From 2000to 2015, consider‐
abledeclinesinδ13Cvaluesof0.8‰–2.5‰wereobservedacrossthreetunaspecies
sampled globally,with moresubstantial changesin the Pacific Ocean comparedto
theAtlanticandIndianOceans.TunarecordednotonlytheSuesseffect,thatis,fossil
fuel‐derivedandisotopicallylightcarbonbeingincorporatedintomarineecosystems,
2 
|
   LORRAIN et AL .
1 | INTRODUCTION
Ov e rthe p ast5 0 yea r s , 90%o fthe h e ata s soci a tedw i thg l o b alw a r m
ing,and30%ofthefossilfuelcarbonemissionshavebeenabsorbed
bytheoceans(LeQuereetal.,2018).Suchprocessesarepredicted
toseverelyimpactmarine biota (Poloczanska et al., 2016)through
enhancedoceanstratificationandacidification.Unfortunately,there
are large un certaint ies on how oceanic e cosystems have ch anged
or may change i n the future. Fo r example, th e current gene ration
of eart h system mod els simulates a wi de range of futu re changes
inglobaloceannetprimaryproductivity(NPP),withbothincreases
anddecreasesofupto20%by2100(Boppetal.,2013;Kwiatkowski
etal., 2017),highlightinglargediscrepancies inthetrends ofsimu
latedNPP.Onlyalimitednumberofempiricaldatasetsrecordtrends
inthephytoplanktoncommunitycompositionorphysiology(Gregg
&Rousseaux,2014;Gregg,Rousseaux,&Franz,2017;Rousseaux&
Gregg,2015).Themagnificationofrelativechangesin phytoplank
tondynamicsacrosstrophiclevelshasrarelybeeninvestigatedwith
fewavailableempiricalmethodscapableofquantifyingecosystem‐
level resp onses. Biolo gical time ser ies dataset s are imperati ve for
understandingpast responsesofthe world'soceans andfor quan
titati ng uncert ainty in fu ture climate p rojectio ns (Bonan & Do ney,
2018).
Carbonstableisotopes(δ13Cvaluesor13C/12C)havebeenused
to reconstruct the oceanic carbon cycle using direct measure
mentsormarinearchives(e.g.,marinesediments,corals)frompa
leoclimatestothecurrentanthropogenicperturbation(Ehleringer,
Buchman n, & Flanagan , 2000 ; Freeman & Hayes, 1992; Keel ing,
2017;Wuetal.,2018).SincetheIndustrialRevolution,therisein
atmosphericCO2hasbeenaccompaniedbyadecreaseinthecar
bonisotope ratioof atmospheric CO2, known as the Suessef fect
(Keeling, 1979).Thisdecreaseisat tributedtotheatmosphericre
leaseofisotopically lightcarbonfromfossilfuelcombustion.Due
totheoceanic uptake ofthis 13C‐depletedCO2,the oceanicδ13 C
value of dissolved inorganic carbon (δ13CDIC) is decreasing (Quay,
Sonnerup,Munro, & Sweeney,2016;Quay et al., 2007).Changes
in δ13 CDICvaluesarerecordedinphytoplanktonδ13Cvaluesafter
accounting for an isotopic fractionation factor associated with
photosynthesis (defined as εp). Isotopic fractionation is depen
dent on seawater characteristics, phytoplankton composition,
and physio logy. The primar y factors that are b elieved to affec t
the isotopic values of phytoplankton are: (a) the concentration
and δ13Cvaluesofdissolved CO2([CO2]aq;Fr y,1996;Laws,Popp,
Bidigar e, Kennicutt, & M acko, 1995;P opp et al., 1998); (b) phy
toplankton community composition and cell morphology (Popp
etal.,1998);and(c)cellulargrowthrate(Bidigareetal.,1997;Fry,
1996).Secondary physiologicaltraits (e.g., decreases in bicarbon
ateu pta keo ri nca r bon co nce nt r ati ngm ech ani sma c tiv ity)ca nal so
impact isotopicvalues, but are difficult tomodel (Cassar,Laws,&
Popp,2006).
Carbonisotopic changes at the baseoffood webs are trans
ferredtohighertrophiclevelswit hvaluesincreasingslightly(ty p
ically 0.5‰–1‰)with each trophic transfer (Fr y,2006; Graham
etal.,2010).Metabolically activetissuesof consumers (e.g., fish
muscle) integrate the stable isotope values of thisbase through
their diet (C herel & Hobso n, 2007; Graham e t al., 2010). While
nitrogenisotope (δ15N)values are commonly usedtoinvestigate
changes in trophic levels, δ13C values provide information on
animal die ts and on spat ial variation s at the base of food w ebs
(Cherel&Hobson,2007;MacKenzie,Longmore,Preece,Lucas,&
Trueman,2014;Trueman,MacKenzie,&Palmer,2012).Historical
studies focusing on baseline changes haveex amined accretion
arybioarchivesthatsufferlittledegradationafterformationsuch
askeratin baleenplates, feathers or teethdentin of marinecon
sumers thatreflectthefood theyingest and therefore,theδ13C
valuesofphytoplank ton(Jaeger& Cherel,2011;Newsomeetal.,
2007;Schell,20 01).Thesestudiesdemonstratetheutilityofiso
tope meas urements to rec onstruct p ast and present o cean pri
mary productivity,andprovideevidenceof past climatechanges
atregional scales(Hobson,Sinclair,York,Thomason,& Merrick,
2004;Newsomeetal.,2007;Schell,20 01).Finally,metabolically
inert but inorganic accretionary str uctures (e.g., bivalve shells,
co ralsk ele to n s,s cl e ro spon ge s ,o rfi sho to lit hs)ca nal sor ef l ect the
δ13CDICvalueof th ee nv ir on me nt ac ro ssth ei rl ifet im e(Frailee ta l. ,
but also recorded profound changes at the base of marine food webs. Wesuggest
a global shif t in phytoplankton community structure, for example, a reduction in
13C‐richphytoplankton suchas diatoms,and/ora changeinphytoplankton physiol‐
ogyduringthisperiod,althoughthisdoesnotruleoutotherconcomitantchangesat
higherlevelsinthefoodwebs.Ourstudyestablishestunaδ13Cvaluesasacandidate
essentialoceanvariabletoassesscomplexecosystemresponsestoclimatechangeat
regionaltoglobalscalesandoverdecadaltimescales.Finally,thistimeserieswillbe
invaluableincalibratingandvalidatingglobalearthsystemmodelstoprojectchanges
inmarinebiota.
KEYWORDS
albacoretuna,AtlanticOcean,bigeyetuna,biogeochemicalcycles,carboncycle,IndianOcean,
PacificOcean,phy toplankton,Suesseffect,yellowfintuna
    
|
 3
LORRA IN et AL .
2016;Swart,2010)astheyusuallyprecipitateinequilibriumwith
seawater, although vital effects can complicate environmental
reconstruction (Lorrain, 2004;McConnaughey &Gillikin, 2008).
Similarly,δ13C values of metabolically active tissues may reflect
trendsinphysio‐chemicalprocesses (CO2 and δ13Caqvalues)and
biologicalprocesses(phytoplanktonδ13Cvalues).
Theaimofourstudy wastoassesstrendsinatimeseriesof
stableisotopevaluesofmetabolicallyactivetunatissues,andto
test if thi s could be used to d etect ecos ystem‐level re sponses
over deca dal time scales at r egional to global sc ales. For this
purpos e, we analyze d δ13C and δ15Nvalues ofmusclesamples
from thre e species of tuna (ye llowfin tun a, Thunnus albacares;
bigeye tuna,T. obe su s;and albacore tuna,T. alalunga) collected
throughout tropical, subtropical, and temperate oceans from
200 0 to 2015 (n=4,477;Figure1).Eachofthesespecieshas
different vertical and foraging distributions (from surface to
mesopelagic depths)(Olson et al., 2016). Therefore, our study
seekstoresolvebroadhorizontalandverticalspatialpatternsin
oce anicf oo dwebs .A stunaar ewide lydistributedandha rves te d
globall y (Majkowski, 20 07), they are goo d candidate s to study
how obser ved and suspec ted changes in phys ical and biologi
calprocessesat globalandoceanbasinscales maybereflected
inconsumerδ13C values. We developed a theoretical modelto
decompose the observed temporal changes in consumer δ13C
valuesintoputativecausal contributors. The model accounted
for(a)known temporaltrends infossilfuel–derivedcarbon (the
Suesseffect)andCO2availability;(b)possiblechangesinphyto
planktondynamicsincludingcommunitycompositionandgrowth
rates;and(c)potentialchanges inthetrophicfractionation fac
tor.Our study,which focusesoncarbonbutdraws on nitrogen
isotopes to a ssess potenti al changes in tun a trophic posi tions,
suggestslarge‐scaleshifts inphytoplankton communitiesfrom
2000to2015.
2 | MATERIALS AND METHODS
2.1 | Tuna carbon isotope data
We assemble d a global dat abase using pu blished an d unpublishe d
regional carbon isotope studies resulting in 4,477 records from
2000to 2015. Det ails on isotopic methods andpredatorsampling
are provid ed in Pethybridge e t al. (2018) who analy zed the same
globaldatasetbutforδ15Nvalues.Asforδ15Nandotherglobalcom
pilation studies (Birdetal., 2018), we assumedthat the agreement
between δ13C values generated ac ross different lab oratories was
<0.2‰–0.3‰.Tunasize(forkleng th,incm)wasmeasuredforeach
individual. Tuna were sam pled from three oce an basins (Atlantic,
Indian,andPacificOceans)withalbacoretunaoccupyingmoretem‐
perate waterscomparedto thetropicalyellowfinandbigeyetunas
(Figure1;Olsonetal.,2016).
The Pacific Ocean had the most extensivesampling with 2,504
individualsandnogapfrom2000to2015,exceptforalbacorewhere
datawerenotavailablefor4years.IntheIndianandAtlanticOceans,
datawere more scattered(Table S1). Pelagictuna are mobile preda‐
torsandthestablecarbonisotopiccompositionoftheirmuscletissue
representsanintegrationoftheirforagingenvironmentoverapproxi‐
mately6monthsto1year(Houssardetal.,2017).Tunamuscletissue
δ13Cvalueswerecorrectedforlipidsinallsampleseitherwithchemi‐
calextractionorusingamassbalanceequationforelevatedlipidcon‐
tentsamples(C/N>3.5)withparametersderivedfromAtlanticbluefin
tuna(T. thynnus)muscle(Loganetal.,2008).
2.2 | Temporal trends in tuna δ13C values
Time series analyses based on multiple linear regression analysis,
performedusingthe R‐3.2.4 software (R Development CoreTeam,
2016) and the nlme package (Pinheiro et al., 2018), wereused to
examineandtestforsignificantlineartrendsintunacarbonisotope
values.Toensurethat tuna length(size)did nothaveany effecton
potentialtemporaltrends,aninteractionbetweensizeandyearwas
testedandwasnotfoundtobesignificantforthePacificorAtlantic
Ocean.Fortheglobaldataset,wetestedamodelwithtuna sizein
cluded,byspeciesandoceanbasins,andthenfittedamodelexplain
ingtheresidualsofthis firstmodelasafunctionofyear.Theslopes
weresimilar to those obtainedwithout theeffectofsizeincluded,
meaningthattheadditionofsizedoesnotchangetheobservedpat
ternsand that this factor has a small impact on temporal trendsin
δ13Cvalues.Wefinallytestedforthreevariables:year(quantitative),
ocean with three levels (Atlantic, Indian, and Pacific Oceans), and
tuna species with threelevels (albacore, bigeye, andyellowfin). All
combinationsweretestedandthefinalmodelwaschosenusingthe
Akaikeinformationcriterion.Weaddedanautocorrelationstructure:
FIGURE 1 Mapofglobalstudyarea
andlocationsof4,477samplesfor
threetunaspecies.Theblacksquare
delineatestheNewCaledonia–Fijiregion
usedforafocusedspatialandtemporal
analysis
40ºS
20ºS
0
20ºN
40ºN
60ºN
60ºE 100ºE 140ºE 180º 140ºW 100ºW 60°W 20ºW
Tuna species
Albacore
Bigeye
Yellowfin
4 
|
   LORRAIN et AL .
a one‐degree autoregressive integrated moving average (ARIMA;
Pinheiro et al., 2018)via the gls function fitted by groups of tuna
sampled at the same date and position.Autocorrelation structures
onresidualswerecheckedwith anautocorrelationfunction. Finally,
toaccountforpossiblespatialbiases(a)inyearsofsamplingaccord
ingtolocations(FigureS1) or(b)due tobaselineisotopicvariations
acrossspace(McMahon,LingHamady,&Thorrold,2013),tunaδ13C
trendswerealsodeterminedatasmallerspatialscalebyconsidering
oneregionwheresufficientdatawereavailableperyear,thatis,New
Caledonia and Fiji (see Figure1 for selected region area) andwith
similarisotopevaluesat thebaseofthe foodweb (Houssardet al.,
2017; Magozzi, Yool, Vander Zanden, Wunder, & Trueman, 2017).
Furthermore,all New Caledonia and Fiji samples were analyzed in
thesamelaboratoryandtimeperiod.
2.3 | Modeling the factors influencing tuna
δ13C values
Wedevelopedatheoretical model to explain thepotentialeffects
of various f actors and p rocesses know n to explain tren ds in tuna
δ13Cvalues.First,weconsideredtheisotopevalueofphytoplankton
(δ13Cp)that has beenshown to be driven by the magnitude of car‐
bonisotopicfractionationduringphotosynthesis(εp)andtheisotope
valueofCO2(δ13Caq,thatis,theSuesseffect),withεpdependenton
thecarbonisotopefractionationassociatedwithcarbonfixation(εf)
andthespecificgrowthrate(µ;Lawsetal.,1995).
with
wherebisaconstant(mM/day)reflectingthedegreeofdependence
offractionation on theCO2concentration,and isbelievedto vary
between species and as a function of growthconditions (Bidigare
etal.,1997;Cullen,Rosenthal,&Falkowski,2001).Whiletheparam
etervaluesarearbitrary,theyarewithintherangeofvaluesreported
intheliterature(Table1).Aninitialvalueof120wasusedforbwhich
isconsistentwiththerangeofvalues(52.6–137.9)fromPoppetal.
(1998)forEmiliana huxleyi(εp = 24.6–137.9 µ/CO2)themostcommon
species of coccolithophoreglobally (Beardall & Raven, 2013), and
Phaeodactylum tricornutum(εp = 25.5–52.6 µ/CO2),adiatommostly
usedinlaboratorystudiesbutnotrepresentativeoftheglobalocean.
Fortheintrinsicfractionationduringphotosynthesisbytheenzyme
Rubisco(εf),weused avalueof25‰,whichhas beenestimated to
rangebetween~22‰ and 30‰ depending on species (e.g., Popp
etal.,1998),withvaluesaslowas11‰fortheRubiscoofthecocco
lithophoreE. huxleyi(Boller,Thomas,Cavanaugh,&Scott,2011).The
valueof25‰forεfhasoftenbeenproposed(Bidigareetal.,1997)in
studiestryingtounderstandtemporaltrendsinmarinechronologies
(Schell,2001).
(1)
δ
13Cp=
1
(
1+𝜀p
1,000 )
(δ13Caq 𝜀p)
,
(2)
𝜀
p=𝜀f
b
𝜇
CO
2
,
Factors (x)
Starting
value or
equation
used
Sensitivity
(d
𝛅
13C
tuna
𝛅13C
tuna /
dx
x
)
Imposed
change
%
Tun a
Δδ13C,
‰,
16‐years
% Change explained
NC‐Fiji PO AO IO
δ13Caq Quayetal.
(2016)
−0.08 NA 0.30 14 12 22 38
CO2Cassaretal.
(200 6)
0.21 NA 0.06 3 2 4 8
Bidigare
(1997 )
NA 0.27 13 11 20 35
Growth
rate(µ)
0.3 −0.2 −5 0.20 10 814 26
−10 0.30 14 12 22 38
−15 0.48 23 19 35 62
Carbon
fixation
fractionation
factor(εf)
25 1.53 +1 0. 24 12 10 17 31
+2 0.48 23 19 35 61
+3 0.72 35 29 52 92
+5 1.20 58 48 87 154
bFactor 120 −0.2 −2 0.10 5 4 7 13
−4 0.20 10 814 26
−8 0.40 19 16 29 51
−10 0.50 24 20 36 64
Trophic
fractionation
factor(εfc)
4−0.23 −5 0.20 10 814 26
−10 0.40 19 16 29 51
−15 0.59 29 24 43 76
TABLE 1 Parameterandtimeseries
datausedtorunvariousscenariosof
imposedchangesinphytoplankton
dynamic s,andtheeffectsonthedifferent
spatialareasexamined(NewCaledonia–
Fiji[NC‐Fiji]region,PacificOcean[PO],
AtlanticOcean[AO],andIndianOcean
[IO]).Forexample,δ13Caqexplained
12%ofthetunaδ13CdecreaseΔδ13C, in
16yearsinthePO,whileanimposed5%
changeincarbonfixationfractionation
factor(εf)explained48%ofΔδ13C,this
factorhavingthelargestsensitivity
(1.53)comparedtoallfactorstested(see
Section2formoredetails)
    
|
 5
LORRA IN et AL .
While precise estimates ofεf or b are not available, this param‐
etrization provides a quantitativedemonstration of how even small
changesinphytoplanktoncommunitycompositionorphysiologymay
influence tuna muscle δ13C values, and hence emergent signals of
change(Table1).Growthrateµwas setat0.3 day−1asitis the me‐
dianvalueatStationALOHA(Lawset al., 2013)inthe central North
Pacific(Hawaii OceanTime‐series) and was alsousedbyseveral au‐
thors(Cullenetal.,2001).Rangesforµfrom0.1to1day−1havebeen
reported in the literature(Boyd et al., 2013; Laws et al., 2013). We
proposedadecreaseingrowthrateofupto15%overthe2000–2015
studyperiod(from0.30to0.26),whichisonthehighendofobserved
modernchanges(Greggetal.,2017)andpredicteddecreasesforthe
future(Kwiatkowskietal.,2017).Theδ13Caq valuesweretakenfrom
StationALOHA(Quayetal.,2016).
Changesintunaδ13Cvaluescaninturnbedescribedbythefol
lowingequation:
whereεfcistheoverallfractionationassociatedwithtrophictrans
fersandisconsideredtobelow(~0.5‰to1.8‰pertrophiclevel),
thereforeweused4‰fortunathatareconsideredtobeatatrophic
positionof~4(Olsonetal.,2016).Asacomparison,Birdetal.(2018)
found an average difference of4.6‰ between phytoplank ton and
sharksonaglobalscale.
CombiningEquation(2)withEquation(3)leadsto:
We used two dif ferent parametrizations developed in the litera
ture on the effect of the concentrationof CO2aqontheisotopeval
ues of phytoplankton (Bidigare et al., 1997; Cassar et al., 2006).
Indeed, while the first parametrization (Bidigare et al., 1997; Laws
et al., 1995) provides quantitative intuition for the dependence of
tuna δ13C value s on phytoplank ton physiology, it do es not account
for additional factors influencing the phytoplankton isotope val‐
ues, including changes in the carbon source (bicarbonate vs. CO2),
deactivation of carbon‐concentrating mechanisms in response
to increased CO2 availability, and changes in the growth condi
tions (nutrient vs. light limitation). For completeness, we also show
the predictions based on the parametrization presented in Cassar
et al. (20 06). The sensiti vity of tuna δ13C va lues to each fa ctor was
assessedbycalculatingtheratioofthepercentagechangeoftunaδ13C
values to p ercentage chan ge of each facto r. T he slope (‘m’) of eac h
curveisrelatedtotheratioofthepercentagechangeoftunaδ13C val
ues to perce ntage change of e ach factor a ccording to the fol lowing
equation:
Thissensitivityanalysisexamines the influence of one parameter at
a time on tuna isotopic composition with assumed initialvalues for
eachparameter basedonliteraturevalues.Toreinforcethis analysis,
weconductedaBayesianapproachthattakesintoaccounttheuncer‐
taintyofall parameterssimultaneouslytoexplaintuna isotopiccom‐
position with ranges and uncertainties taken from literature values
(AnalysisS1).Toreconstructtheobservedtrendsintunaδ13Cvalues,
anumberofscenarioswereruntosimulatepercentagechangesinthe
phytoplanktonparametersµ, b, and εf or εfc(Table1).Thesescenarios
wereusedtoresolvecompetinghypothesesfortheobservedpatterns.
3 | RESULTS
3.1 | Trends in δ13C values
Overtheentirerecord(Figure1),individualtunaδ13Cvaluesranged
from−19.9‰to−12.9‰.Mean annualδ13 Cvalues decreased by
0.8‰ to 2.5‰ within species and ocean basins from 20 00 to
2015(Table2;Figure 2). These negative trendsweresignificant,
withsimilarobservedslopesforeachtunaspeciesbyoceanbasin
(p<.0001;Table2).Thelargestdecreasewasobservedforthe
Pacific O cean and the l owest in the In dian Ocean, wi th the de
creaseintheAtlanticOceanbeingintermediate.Forreference,we
showedaglobaldecreasingtrendacknowledgingthatmostofour
observationswerefromthePacific(56%)(Table2;Figure2;1.8‰
decreasefrom2000to2015),whileobservationsfromtheIndian
andAtlanticOceanseachcomprised22%ofthedata(Table2).
In the regi on of New Caledo nia and Fiji, whe re we have the
most complete record to derive a temporal trend(Figure S1),the
same decreasing patternintunaδ13Cvalues was observedforall
threespecies(2.1‰decreasebetween200 0and2015;FigureS2)
asoverthe broader Pacific region (2.5‰ decrease).No temporal
changesinδ15Nvalu es wereo bser ve dfort hethreetunas pe ci esin
theNewCaledonia andFijiregions (FigureS3), suggestingnosig
nificanttunatrophicpositionchangesovertherecord.Someweak
trendsinδ15Nvalueswere foundfor some speciesandoceanba
si ns , whi chcou l da r is efr o mt h ei nte r act i onwit hot h ercon fou ndi n g
factor s such as tuna s ize and locat ion (Figure S 4). Twolikel y ex
planationsfortheobservedtunaδ13Ctrendsarediscussedbelow.
3.2 | Accounting for the observed Suess effect and
CO2 availability
Reported declines in δ13Caq values at Station ALOHA during the
2000–2015period(−0.3‰) explained 14%of thedecrease in tuna
Δδ13CvaluesobservedinourNewCaledonia–Fijiregion(Figure3a).
Assumingsimilar Suess effects inthe other ocean basins,12%,
22%, and 3 8% of the decreas e in tuna δ13C value s in the Pacific ,
Atlantic,andIndianOceans,respectively,canbeexplainedbyδ13Caq
(Table1).
An incre ase in the conce ntration of CO2aq obs erved at St ation
ALOHA also leadsto a decrease in tuna δ13C values by increasing
carbonisotopicfractionationduringphotosynthesisεp(Equation2).
(3)
δ
13Ctuna =
1
(
1+𝜀fc
1,000
)
(δ13Cp𝜀fc )
,
(4)
δ
13Ctuna =1
1+𝜀fc
1,000
1
1+𝜀fb𝜇
CO2
1,000
δ13Caq 𝜀fb𝜇
CO2
𝜀fc
.
(5)
dδ13Ctuna
δ13Ctuna
dx
x
=mx
δ13Ctuna
6 
|
   LORRAIN et AL .
Depend ing on the par ametrizat ion used to acco unt for this ef fect
(Bidigareetal., 1997;C assar etal., 2006), 2%–35%ofthe trendin
tunaδ13Cvalues can be explained bychanges in CO2availabilityin
the New Caledonia–Fijiregion (Figure 3b). Usingthe largerdegree
ofchangeinresponsetoCO2availability(Bidigareetal.,1997),per
centages of chang e similar to thos e from the Sues s effect c an be
explainedacrossthedifferentoceanbasins.Theadditiveimpactsof
CO2availabilit yandtheSuesseffectinexplainingtunaΔδ13Cvalues
are23%, 41%,and73%inthePacific,Atlantic,and IndianOceans,
respectively. IftheCO2availabilityef fectreportedbyCassar et al.
(2006)isused,inadditiontotheSuesseffect,thenonly14%,26%,
and 46% of the Δδ13C decrease inthe Pacific, Atlantic,and Indian
Oceans,respectively,canbeexplained.
3.3 | Hypothesized changes in phytoplankton
dynamics
According to theoretical models (see Section 2), the then unex
plainedtemporalchangesintunaδ13Cvalues(~27%–86%)mustbe
relatedto (a)a decreasein phytoplanktoncellulargrowthrates (µ)
orphysiology(e.g.,carbon‐concentrating mechanism activity);and/
or (b) potential changes in phytoplankton communities (through
changesinspecies‐dependentparametersb and εf)orinthetrophic
fractionation factor εfc. Based both on the sensitivityanalysis and
theBayesian inference,variations in the growthrate(µ) and in the
trophicfractionationfactor(εfc)haveasmalleffect ontunaisotope
values(Table1;AnalysisS1).Asanexample,an imposedsubstantial
15%decreaseinthe growthrate µover16yearsresultedinonlya
0.48‰decreaseintuna δ13Cvalues(Figure 3c;around20% ofthe
totaldecrease in theNewCaledonia–Fiji regionandin the broader
PacificOcean).This effectislargerin theAtlantic(35%)andIndian
(62%)Oceans.However, morereasonable declines of 5% and 10%
ofµover16yearsresultedinsmallerdecreasesof0.2‰and0.3‰
intunaδ13Cvaluesfrom2000to2015,respectively(Table1),which
explained 8%–38% of this overall signal in various ocean basins.
Variationsinthetrophicfractionationfactorεfcwereofsimilarorder
(Table1)withlargedecreasesofthisparameterneededtoexplainthe
tunaδ13Cpattern.
FIGURE 2 Timeseriesoftunamuscle
tissueδ13Cvalues(‰)withobservations
dividedbyoceanbasin.Theshadedarea
alongthelinearfitcorrespondstoa95%
confidenceinterval
Albacore Bigeye Yellowfin All oceans and species
2002 2006 2010 2014 2002 2006 2010 2014 2002 2006 2010 2014 2002 2006 2010 2014
−20
−18
−16
−14
Year
δ
13
C
Ocean
Atlantic
Indian
Pacific
Ocean basin/
region
Tun a
species
Intercept
(in 2000) Slope
Slope
stand
ard error r2
Temporal
change
(2000–
2015,
in ‰) n
Atlantic Albacore −17. 4 −0.092 0.0085 69.1 −1 . 3 8 608
Bigeye −16 . 6 126
Yel lo w fin −1 6 .6 256
Indian Albacore −16 . 8 −0.052 0.0077 69. 1 −0.78 24 8
Bigeye −16 . 1 237
Yel lo w fin −1 6 .1 498
Pacific Albacore −15 . 2 −0.166 0.0048 69. 1 −2.4 9 878
Bigeye −14 . 5 645
Yel lo w fin −14. 5 981
New
Cale donia–
Fiji
Albacore −15 . 3 −0 .138 0.0095 42.1 −2.0 7 364
Bigeye −14 . 8 120
Yel lo w fin −14. 9 331
Global All −15 . 4 0.120 0.0057 22.1 −1 . 8 0 4,477
TABLE 2 Regressionanalysisoutput
includingtheslopeandinterceptforeach
tunaspeciesandoceanbasin.Onlyone
valueisshownwhensimilarforseveral
species
    
|
 7
LORRA IN et AL .
Thecarbonfixationfractionationfactor(εf)andbvaluescanvary
widely among phytoplankton species (Popp et al., 1998). The sen‐
sitivit y analysis and the Bayesian model (that takes into account a
largerangeofvaluesfortheseparameters)showedthatthecarbon
fixationfractionationfactorεfhadthelar ges teffectont het unaiso
topevalues compared toallother factors(Table1;Analysis S1).As
anexample,wearbitrarilysetthechangesto5%forεfand10%forb
(Figure3d)toreflecttheirdifferentialimpactontunaΔδ13Cvalues.
Thissmall5% increasein εfresultedinalargedecreaseintunacar
bonisotopevaluesof1.2‰(i.e.,~50%ofthetunaΔδ13CintheNew
Caledonia–FijiregionandthePacific Ocean)(Figure3d; Table1).In
comparison,a 10%declinein‘b’values only causeda0.5‰decline
in tuna δ13C val ues, which ex plained 20%–36% in the P acific and
AtlanticOceans,and64%intheIndianOcean.
AfteraccountingfortheSuessandCO2availabilityeffects,sev
eral permutations for the parameters reflecting productivity (µ)
and spec ies composit ion (with chan ges in εf and b comb ined) may
accountfor the remaining Δδ13C changes. The combination of the
Suess ef fect, th e effect of inc reasing [CO2]aq on εp, and a chang e
of5%–10%inspecies‐specificparameters(10%forband5%forεf)
with no cha nge in produc tivity or t he trophic fr actionat ion factor
εfc,produced a 2.1‰decrease in tuna δ13C values,consistentwith
theobservedchangeinthePacificOcean(Figure3d).InthePacific
Ocean, w here we have the most ro bust dataset , changes in phy‐
toplank ton paramete rs seem to have occur red, unless we a ssume
that growth rates have changed by >70%. If we assume that no
changesingrowthratesandεfchaveoccurredandusetheBidigare
etal. (1997)parametrization, thenmore than 60% of Δδ13C has to
beexplainedbya changeinspecies compositioninthePacific and
AtlanticOceans,againstonly27%intheIndianOcean(Table1).The
use of the Cassaret al. (2006) parametrizationimplieseven larger
changes in s pecies comp osition. Aver aging all tu na species an d all
ocean basins,and both parametrizations used to calculate carbon
fractionationfromphytoplankton,theglobaltrendintunaδ13C val
ues(Δδ13 C)can,forexample,beexplainedby(a)theobservedSuess
effectand increasesinCO2aq(upto26%);(b)a5%decreaseinpro
ductivity (11%),a 10% decrease in thetrophicfractionation factor
(17%);and(c)imposedchangesof5%inspecies‐specificparameters
FIGURE 3 Predicted(colorline)versusobserved(blackline)changesintunamuscleδ13 Cvalues(Δδ13C,‰)intheNewCaledonia–Fiji
regionofthePacificOceanasafunctionofvariousprocesses.(a)TheSuesseffect.(b)IncreaseinCO2aqinseawaterundertwoscenarios
basedondifferentparametrizationsintheliterature(Bidigareetal.1997,Cassaretal.,2006;seeSection2fordet ails).(c)Adecreasein
phytoplanktoncellulargrowthrateof15%.(d)Achangeof5%forthecarbonfixationfractionationfactorεfand10%fortheconstantbused
tocalculatecarbonisotopefractionationduringphotosynthesis(seeSection2fordetails),andalsoallfactorsconsideredtogether,except
growthrate(blueline=Suesseffect+CO2aqfromBidigareetal.,1997+b + εf)
(c) Growth rate effect (d) Phytoplankton species (or physiological) effects
(a)
Suess effect
(b)
CO2 availability effect
2000 2005 2010 2015 2000 2005 2010 2015
−2.0
−1.5
−1.0
−0.5
0.0
−2.0
−1.5
−1.0
−0.5
0.0
Year
Observed tuna change
Observed tuna change Observed tuna change
Observed tuna change
All factors but growth rate
b
Growth rate
δ
13Caq effect
from Bidigare et al. (1997)
ε
f
from Cassar et al. (2006)
∆ δ C
13 tuna (per mil)
8 
|
   LORRAIN et AL .
indicatingashiftinspeciescomposition(46%;Figure4).Whilethisis
onepote ntials ce nario,inpa r ti nfor me da ndco ns tra in ed byobse rva
tionsintheliterature(Greggetal.,2017;Rousseaux&Gregg,2015),
thereisamultitudeofpermutationsthatmayfittheobservedtrend.
Neve rtheless,changesint hec ar bonfixat ionfractionationfactor(εf)
had the largest effects in the simulations using bothmodeling ap‐
proaches,and better accountedforthe observed tunatrends than
changesinproductivity,thetrophicfractionationfactor(εfc),oreven
theknownSuesseffect.
4 | DISCUSSION
Ouranalysisrevealedthatchangesinthebiologicalcomponentofthe
marinecarboncyclecanbetracedinthetissuesofmarinetoppreda‐
tors.Weobservedsubstantialandwidespreaddeclinesintunamuscle
δ13Cvalues(by0.8‰–2.5‰)inthreetunaspeciesacrossthreeocean
basins.Suchatrendovera16‐yearperiodhasneverbeenrecordedin
metabolictissuesofamarinepredator.Theuseoftwoseparatemod‐
elingapproaches(sensitivityanalysisandBayesianinference)revealed
thattheparameterlinked to phytoplankton carbon fractionation (εf)
hadthelargestinfluenceontheobservedtemporaltrendintunamus‐
cle δ13Cvalues.Our calculationsthensuggest thatupto 60% ofthe
decrease in tuna δ13C values seems tobe due to a change in phy‐
toplankton parameters in the PacificOcean, compared to only 27%
inthe western Indian Ocean.Whileourmost robust dataset isfrom
the Pacific Ocean, the same decreasingpattern in tuna δ13C values
in all ocean basins (Pacific,Atlantic, and western Indian) suggests a
widespreadshiftinmarineplanktoncommunitiesorachangeintheir
physiology,butdoesnotexcludeotherfactorsthatmayactinsynergy
(e.g.,achangeinproductivityorinthetrophicfractionationfactor).
Previouslyreportedtemporal changesinδ13C valuesaregen
erallyattributedtotheSuesseffectorchangesinmarineproduc
tivity in various organisms and ecosystems (Fraile et al., 2016;
Newsome et al., 2007; Schell, 2001). For example, Schell (2001)
foundasignificantlong‐termdeclineinδ13Cvaluesininertbaleen
plates (~2.7‰) over a 30 year p eriod (bet ween 1965 and 1997)
attributingthisdeclinetoa~3 0%–40%declineinprimar yprodu c
tivit y in the Bering S ea. Cullen et al . (2001) propos ed that part
ofthe decrease observed by Schell (2001)was due to theSuess
effectanddue to the influence of changesinCO2 concentration
onph y to p la n kto np hys iol ogy (a ssh ow nin Fig ure3b and de s cri bed
herein). In co ntrast, t he Suess ef fect is relat ively small over o ur
timeperiod(0.3‰;Figure3a)andonlyexplains~12%–20%ofthe
observeddecreaseintunamuscleδ13Cvalues(intheAtlanticand
PacificOceans).Similarly,increasingCO2concentrationsonlyex
plain asmall percentage (~2%–18%; Table1)of theobservedde
creaseintunamuscleδ13Cvalues,usingthe Bidigareetal.(1997)
orCassaretal.(20 06)par ametrizationsforth eireffec tonthecar
bonfixationfactorfractionation.
AfteraccountingfortheobservedSuesseffectandchangesin
[CO2]aq availabili ty, our model w as used to expl ore how change s
inphytoplankton growth rates and species composition can fur
therreconstructourobserveddeclinesintunamuscleδ13Cvalues.
Thesensitivity,Bayesian,andscenarioanalysisdemonstratedthat
relativelysmallchangesinphytoplanktoncommunitycomposition
canleadtolargedeclinesintunaδ13Cval ues(Tabl e1;AnalysisS1),
while lar ger changes in pr oductivi ty or the trop hic fract ionation
factor w ould be need ed. In our st udy, a 15%d ecrease in p hyto
plankton productivity cannot explain the decline we observed
intuna muscle δ13C values, evenwhen combined with the Suess
effectand thecumulativeeffect of increasing [CO2]aq on εp. T he
changeinproductivitythatwetestedinthisstudyisatthehigher
endofpreviouslyreporteddeclines(typicallyrangingbetween0%
and 1.4% year−1attheregionaltooceanbasinscale;Behrenfeld,
2006; Gregg & Rousseaux, 2014; Joo, 2015).Other studieshave
sh ow nnore cen tt ren dsi ng lob alp rim ar ypro duc t ivi ty(Gre ggeta l.,
2017;Rousse au x&G re gg ,2017)andso mereg ionalincre aseshave
FIGURE 4 Synthesisofthepotential
effectsofvariousfactorsonthetuna
δ13Ctemporaltrend(Δδ13C).Different
combinationsarepossible(seetext
formoredetails)TunaillustrationLes
Hata©SPC
    
|
 9
LORRA IN et AL .
been rep orted historic ally (e.g., Pacific Ocean; Karl, Bidig are, &
Letelier,2001).Eveninmodelstudies,projectionsofglobalmarine
NPParehighlyuncertainwit hrelati vechangesbe tween−20%and
+20%overthe21stcentury(Kwiatkowskietal.,2017).
The rate ofdecrease observed inour tuna δ13Cvaluesrequires
concomitantchanges in phytoplankton species‐specific parameters
(εf and b).Young,B ruggeman,R ickaby,Erez,an dConte( 2013)alr ea dy
repor ted evidenc e of change in the bi ological c arbon isoto pic frac
tionation byphy toplankton(εp) with significant increases bet ween
1960and2010,inparticularinthesubtropics,wherethischangewas
thehighestcomparedtootherregions.Thechangeinbiologicalfrac
tionationestimatedthrough theirmodel (i.e.,amaximumof0.4‰in
16years)istwotofi vetimeslowerth anourob ser vationsintuna,de
pendingontheoceanbasin.Theirstudyisbasedonacompilationof
particulateorganiccarbon(POC)datafromseveraltransectsmostly
from theAtlantic Ocean with few data in our Pacific region, which
could explainthe differences between their model and our obser
vations. H owever,t hey showed a time s eries of δ13C POC (δ13
POC)
values in the North Atlanticoff Bermudawith a 2‰decreasefrom
1980to2007(i.e., ~1.2‰decreasein16years).Thisresultissimilar
tot he1.4‰decreaseintunaval ue sweobser vedinth eAtlanticfr om
2000to2015.
Suppor tforourhypothesisofashiftinmarineplanktoncom
munitiesalreadyexists(McMahon,McCarthy,Sherwood,Larsen,
&G u i l d e r s on , 2 0 15 ; P o l ov i n a , Ho w e l l ,& A b e c a ss i s , 2 0 0 8 ;P o l o vi n a 
& Woodworth, 2012; Rousseaux & Gregg, 2015). Diatoms are
predicted to decrease in abundance in response to increased
seawaterstratificationwithareporteddeclineof1.22%year−1 in
the Nor th Pacific ( Rousseaux & G regg, 2015). Su ch a reducti on
intheabundanceofdiatoms, a 13C‐rich carbon sourceinmarine
food webs (Fry & Wainright, 1991), is expected to decrease the
δ13Cvaluesofconsumers,and this diatom contributionhas been
emphasize d in a recent mode l of phytoplan kton δ13C variations
in the glob al ocean (Magoz zi et al., 2017). Tuerena et al . (2019)
alsorecentlyfoundthatcellsizewastheprimarydeterminantof
δ13
POC in the Sout h Atlantic subtropical conver gence zone and
predicted that carbon isotopic fractionation will increase in the
future, l eading to lower δ13
POCthatmaypropagatethroughthe
food web. A decrease in the abundance of coccolithophores,
another 13C‐rich carb on source, migh t also explain so me of the
tuna δ13C trend. However,Rivero‐Calle, Gnanadesikan, Castillo,
Balch, and Guikema (2015) found an increase in the occurrence
ofcoccolithophoresintheNor thAtlantic,butdataare notavail
ableataglobalscale.Oceanbasindifferencesfoundinourstudy
in the temp oral slopes in tu na δ13C values bet ween the Indian
Ocean (0. 8‰) and the Atlanti c and Pacific O ceans (from 1.4‰
to2.5‰) could be due to a combination of severalfactors. The
magnitu de of the Suess e ffect may v ary region ally (Quay et al. ,
2016). Changesinphytoplanktoncommunitiesorphysiologyare
also depe ndent on regio nal‐scale p rocesses (Gr egg et al., 2017;
Siegel et al ., 2013). The use of spat ially resolve d models of the
ocean 13Ccyclewouldhelptounderstandtheregionaldifferences
(Tagliabue&Bopp,2008).
5 | CAVEATS AND LIMITATIONS
Phytoplanktonhavemanystrategiestotakeupcarbonasafunction
of growth co nditions t hat could aff ect frac tionatio n. For exampl e,
adecreaseinbicarbonate uptake or carbon‐concentratingmecha
nismactivityingeneralwouldbepredictedtoincreasetheapparent
fractionation.Ourpredictionsshouldthereforebeinterpretedwith
cautionasisotopicfractionationis not a single function of µ/CO2,
even within a single phytoplankton species (Cassar & Laws, 2007;
Cassaretal.,2006).Furthermore,otherpermutationsofεf or b may
fitthe observeddecrease. However,thisparametrizationtogether
with the Bayesian inference demonstrates how small changes in
phytoplanktoncommunitycompositionorphysiologymayinfluence
tunamuscleδ13Cvalues.
We also note that r egional variat ions cannot b e captured by
the time series of δ13CaqatStationALOHA(Hawaii,Pacific).
Long‐term declines of δ13Caqvaluesduetothecombinationof
the Sues s effect , vertical m ixing, and p rimary pro duction (re sid
ual carbon pool af ter POC production) have been documented
at other monitoring stations, with varying effects according to
regionandlatitude,inparticular in the southern regions (King &
Howard,2004).However,bothinstrumentalandproxyrecordsof
δ13Caqindicateaconsistentaveragedecreaseperyearof0.027‰
atfivePacificstationsfromHawaiitoAmericanSamoasince1980
(corresponding to an approximately0.4‰ decline in our 16 year
peri od ;Wu,2018) .Fur therm ore,Gru beretal.(1999) co mpare dt he
δ13Caqtrendsinseveraloceanicregi onsandfoundthatt hehighest
decreas e of 0.025‰ was in the s ubtropica l gyres (Be rmuda and
Hawaii) an d the lowest in t he equatoria l upwelling reg ion of the
Pacific (0.015‰),with the Indian Oceandisplayingadecreaseof
0.020‰ p er year. Therefore, t he predic ted ranges in al l oceanic
reg ionso f0.2‰–0 .4‰de cr easeoverou rs tu dyperiodof16y ea rs
are too smal l to explain the 1. 8‰ average decline in t una δ13C
values.
Other factors, related to food web ordietary processes, could
alsoinfluencethetunaδ13Ctrend.Sizedifferencesinsampledtuna
through timecould introduce abias, but no consistent relationship
betwee n tuna size and δ13C and δ15N val ues or any size changes
with time among tuna species were obser ved (Figures S5and S6).
Whiledecadalshiftsinthedietofyellowfintunahavebeenrecorded
in the eas tern Pacif ic Ocean from t he 1990s to the 20 00s (Ols on
et al., 2014), the sim ilar δ13C slopes obse rved for the thre e tuna
species in our study seem inconsistent with changes in foraging
location or diet. A shift in the tuna foragingrange or timing could
also be argued to explain the observed decrease in tuna δ13C val
uesastherearelargespatialandintraannualvariabilitiesintheδ13C
values of phy toplankton (Magozzi et al., 2017). Similarly,if all tuna
foraged deeperonmoremesopelagicpreythathadlowerδ13C val
ues than su rface prey (δ13CDIC is known to decreas e with depth;
Quayetal., 2003),tunaδ13Cvalueswoulddecrease.However,the
slopeofthisdecreasewouldvaryamongspeciesgiventhatyellowfin
tunamainlyinhabitsurfacewaters while bigeye tuna mostlyforage
inmesopelagicwaters(Olsonetal.,2016).Adecreaseinthetrophic
10 
|
   LORRAIN et AL .
fractionationfactorεfcwouldreducetunaδ13Cvaluesthroughtime
(Table 1; Figure 4). Var ious processes in cluding a change in foo d
chain leng th, food web s tructu re, quality of f ood, or tuna met ab‐
olism (Barnes, Sweeting, Jennings, Barry, & Polunin, 20 07) could
alter εfc. A change in the overall trophic fractionation factor could
thereforeoccuratmultiplelevelsofthefoodweb,drivenornot by
changesatthebaseofthefoodweb.Toourknowledge,thereareno
dataavailableintheliteraturetoexplorethisfurther.However,while
wecannot rule out thepossibility that changes in food web struc‐
ture are negated by changes in source 15N (e.g., denitrification vs.
N2fixation;Deutschetal.,2014;Somesetal.,2010),wedidnotsee
temporalchangesintunaδ15Nvaluesatglobaloroceanbasinscales,
suggestinglittlechangeinfoodchainlengthorstructure.
Finally,ourdatasethassomelimitationsinherenttothesampling
andweacknowledgethatourmostrobustanalysisisforthePacific
Ocean tha t covers a large are a with many ind ividuals by yea r and
species( TableS1).MoredataoverbroadreachesoftheAtlanticand
IndianOceans areneededtoprovide robustestimatesofbiological
changesintheseoceans.
6 | CONCLUDING REMARKS
Weshowedthatδ13Cvaluesofmetabolicallyactivetissuesofmobile
marinepredatorslikelyreflectrecentchangesatthebaseofmarine
food webs.Wedetectedasubstantial worldwide decreaseintuna
δ13Cva luesov erth e2 000–2015p eriodwhichcanberelatedto va ri
ousprocessesknowntoinfluenceoceancarboncyclingintheglobal
oceans. Ouranalysissuggests thatphytoplanktonspecies(e.g.,dia‐
toms)th atun de rgoa la rgerfract ionatio nofcar bo nd uringphotos yn
thesis(andthushavehigherδ13Cvalues)havebeendecreasingover
recent dec ades or that the se phytoplank ton communities al tered
their physiologies. Whilewe cannot rule out a widespreaddecline
inphy toplankton productivity,we showedthatevena large (>15%)
decline wouldhaveasmall impact on tuna δ13C values and cannot
fullyexplaintheobservedglobaltrend.Whilerecognizingthatacon
comitantshift at higher levelsoffood webs (change in the trophic
fractionation factor or in tuna diet or physiology) could occur and
thatmoretunacarbonisotopedataareneededfromtheAtlanticand
IndianOceans,thepresentstudyexpandsourunderstandingofthe
mainf actor st ha taff ectth eisotopicva luesof to ppre da to rs an dp ro
videsaframeworktointerpretandmodelcarboncyclingatregional
to global sc ales. New ob servatio nal or modele d data that prov ide
estimatesofperiodicchanges in marine plankton communitieswill
enab leourmo deltoprovidees tim atesoftheothercontributingfa c
tors. Fi nally, the framewo rk presented h ere, through t he study of
tuna car bon and nitr ogen isotope s values, co uld suppor t develop‐
mentofauseful essentialocean variable(EOV)forimplementation
within aglobal oceanobservingsystem to documentcomplex eco‐
systemchangesatregionaltoglobalscalesandoverrelativelyshort
timescales(decadestocenturies).Theuseofpredatorisotopesasan
EOVwouldcomplementregionaleffortstoacquireinsitumeasure
mentsofplanktonabundanceanddiversity(Miloslavichetal.,2018).
ACKNOWLEDGEMENTS
We thank the observer programs, the many observers and re‐
searche rs who collec ted the sample s in each of the ocean b asins
and the Pa cific Marine S pecimen Tis sue Bank. Thi s work contrib‐
utes to CLIOTOP WG 3‐Task team 01. N .C. was suppo rted by the
“Laboratoire d'Excellence” LabexMER (ANR‐10‐LABX‐19) and co
funded byagrant fromtheFrenchGovernmentundertheprogram
“Investissement sd'Avenir.”Wethank P.Quaywhokindlyprovided
datafromStationALOHA,E.A.Lawsforhelpfuldiscussionsonthe
manuscript,andthethreereviewersfortheirthoroughreviews.
CONFLICT OF INTEREST
Theauthorsdeclarenocompetinginterests.
AUTHOR CONTRIBUTIONS
A.L .,A.R., N.C.,andH.P.analyzed thedataand interpreted the re
sults with thehelpof B.F.A.R .andD.E.P.performedthestatistical
analysis.A .L.,N.C., and H.P.wrotethemanuscriptwiththehelpof
A.J.H.andA.R.Allauthorscontributedtoandprovidedfeedbackon
variousdraftsofthepaper.
ORCID
Anne Lorrain https://orcid.org/0000‐0002‐1289‐2072
Heidi Pethybridge https://orcid.org/0000‐00027291‐5766
Nicolas Cassar https://orcid.org/0000‐0003‐0100‐3783
Aurore Receveur https://orcid.org/0000‐0003‐0675‐4172
Valérie Allain https://orcid.org/0000‐0002‐9874‐3077
Nathalie Bodin https://orcid.org/0000‐0001‐8464‐0213
Laurent Bopp https://orcid.org/0000‐0003‐4732‐4953
C. Anela Choy https://orcid.org/0000‐0002‐0305‐1159
Alistair J. Hobday https://orcid.org/0000‐0002‐3194‐8326
John M. Logan https://orcid.org/0000‐0002‐0590‐7678
Frederic Ménard https://orcid.org/0000‐0003‐1162‐660X
Christophe E. Menkes https://orcid.org/0000‐00021457‐9696
Dan E. Pagendam https://orcid.org/0000‐0002‐8347‐4767
David Point https://orcid.org/0000‐0002‐5218‐7781
Andrew T. Revill https://orcid.org/0000‐0003‐2486‐5976
Christopher J. Somes https://orcid.org/0000‐0003‐2635‐7617
REFERENCES
Barnes,C., Sweeting, C. J., Jennings,S., Barr y,J. T., &Polunin ,N.V.C.
(2007).Effectoftemperatureandrationsizeoncar bonandnitrogen
stableisotopetrophicfractionation.Functional Ecology, 21, 356–362.
https ://doi.org/10 .1111/j.1365‐24 35. 20 0 6.012 24. x
Beardall, J., & Raven, J. A. (2013). Calcification and ocean acidifica‐
tion: New i nsights fr om the cocco lithoph ore Emiliania huxleyi. New
Phytologist, 199,1–3.https://doi.org/10.1111/nph.12297
    
|
 11
LORRA IN et AL .
Behrenfeld,M.J.,O'Malley,R.T.,Siegel,D.A.,McClain,C.R.,Sarmiento,
J.L., Feldman, G . C., … B oss, E.S. (2006). Climate‐driventrendsin
contemporaryoceanproductivity.Nature, 444,752–755.http s://doi.
org /10.1038/nat ure 05317
Bidigare,R .R ., Fluegge,A., Freeman,K .H.,Hanson,K .L .,Hayes, J.M.,
Hollander,D.,…Millero,F.J. (1997).Consistentfractionationof13C
innatureandinthelaboratory:Growth‐rateeffectsinsomehapto
phyte algae. Global Biogeochemical Cycles, 11,279–292.https://doi.
org /10.1029/96gb0 393 9
Bird, C . S., Veríssimo, A ., Magozzi, S ., Abrantes , K. G., Ag uilar, A., Al‐
Reasi, H. , … Bouchoucha, M . (2018). A global per spective on th e
trophicgeographyofsharks.Nature Ecolog y & Evolution, 2, 299–305.
htt ps://doi.o rg /10.103 8/s41559‐ 017‐0 432‐z
Boller,A .J., Thomas,P.J.,C avanaugh,C .M.,&Scott, K.M.(2011).Low
stable carbon isotope fractionation by coccolithophore RubisCO.
Geochimica et Cosmochimica Acta, 75, 7200–7207. https://doi.
org/10.1016/j.gca.2011.08.031
Bonan, G.B.,&Doney,S.C.(2018).Climate,ecosys tems, andplanetary
futures:ThechallengetopredictlifeinEarthsystemmodels.Science,
359,eaam8328.https://doi.org/10.1126/science.aam8328
Bopp,L.,Resplandy,L.,Orr,J.C.,Doney,S.C.,Dunne, J.P.,Gehlen, M.,
…Tjiputra,J.(2013).Multiplestressors of ocean ecosystemsinthe
21st centu ry: Projections with C MIP5 models. Biogeosciences, 10,
6225–6245.ht tps://doi.org/10.5194/bg‐10 ‐6225‐2013
Boyd,P.W.,Rynearson,T.A.,Armstrong,E.A.,Fu,F.,Hayashi,K.,Hu,Z.,
…Thomas,M.K .(2013).Marine phytoplanktontemperature versus
growthresponsesfrompolartotropicalwaters–