Trophic structure of neuston across tropical and subtropical oceanic provinces assessed with stable isotopes.

F on ie s in Ma ine Science | www. on ie sin.o g 1 Janua y 2021 | Volume 7 | A icle 606088 ORIGINAL RESEARCH published: 25 Janua y 2021 doi: 10.3389/ ma s.2020.606088 Edi ed by: Xa ie Pochon, Caw h on Ins i u e, NewZealand Re iewed by: Ch is ophe Somes, GEOMAR Helmhol z Cen e o Ocean Resea ch Kiel, Ge many Eleono a Puccinelli, Uni e si é de B e agne Occiden ale, F ance *Co espondence: Rui Albuque que [email p o ec ed] †ORCID: An onio Bode o cid0000-0002-9535-254 Special y sec ion: This a icle was submi ed o Ma ine Ecosys em Ecology, a sec ion o he jou nal F on ie s in Ma ine Science Recei ed: 14 Sep embe 2020 Accep ed: 15 Decembe 2020 Published: 25 Janua y 2021 Ci a ion: Albuque que R, Bode A, González-Go dillo JI, Dua e CM and Quei oga H (2021) T ophic S uc u e o Neus on Ac oss T opical and Sub opical Oceanic P o inces Assessed Wi h S able Iso opes. F on . Ma . Sci. 7:606088. doi: 10.3389/ ma s.2020.606088 T ophic S uc u e o Neus on Ac oss T opical and Sub opical Oceanic P o inces Assessed Wi h S able Iso opes RuiAlbuque que 1 *, An onioBode 2†, JuanIgnacioGonzález-Go dillo 3, Ca losM.Dua e 4 and Hen iqueQuei oga 1 1 Depa amen o de Biologia, CESAM – Cen o de Es udos do Ambien e e do Ma , Uni e sidade de A ei o, A ei o, Po ugal, 2 Ins i u o Español de Oceanog a ía, Cen o Oceanog á ico de A Co uña, A Co uña, Spain, 3 Ins i u o Uni e si a io de In es igación Ma ina (INMAR), Campus de Pue o Real, Uni e sidad de Cádiz, Cádiz, Spain, 4 Red Sea Resea ch Cen e (RSRC) and Compu a ional Bioscience Resea ch Cen e (CBRC), King Abdullah Uni e si y o Science and Technology, Thuwal, Saudi A abia The ma ine neus on, o ganisms li ing in he icini y o he ocean su ace, is one o he leas s udied zooplank on g oups. Neus on occupies a es ic ed ecological niche and is a ec ed by a wide ange o endogenous and exogenous p ocesses while also being a ood sou ce o zooplank on ish mig a ing om he deep laye s and seabi ds. In his s udy, he neus onic communi ies we e cha ac e ized along he Malaspina global expedi ion sampling opical and sub opical oceanic p o inces using s able ca bon and ni ogen iso opes o explo e hei ophic s uc u e and ela ionships wi h en i onmen al a iables. The di e ences in s able iso opes mi o ed he pa e ns in en i onmen al cha ac e is ics o each p o ince. High δ13C alues we e associa ed wi h a mosphe ic ca bon inpu s, while he p esence o dino lagella es, coccoli hopho ids, and upwelling in luence is ela ed o low δ13C alues. Simila ly, p o inces p esen ing high δ15N alues we e associa ed wi h deni i ica ion and ni a e di usi e luxes, whe eas he p esence o low δ15N is a ibu able o ni ogen supplied h ough N2 ixa ion by diazo ophs. Neus on showed a la ge o e lap among he iso opic niches o ou unc ional g oups, wi h chae ogna hs and de i i o es gene ally exhibi ing a smalle deg ee o o e lap compa ed o ca ni o es and omni o es/ he bi o es. These esul s suppo he hypo hesis o a common ophic s uc u e in he neus on communi y ac oss he ocean. Howe e , he size o he niche, small in coas al a eas and hose in luenced by upwelling and la ge in oligo ophic egions, and hei o e lap, low in mo e p oduc i e p o inces and high in oligo ophic p o inces, may beassocia ed wi h ood a ailabili y. Small ophic niches a e associa ed wi h a dominance o specialized o e -oppo unis ic eeding in p oduc i e en i onmen s. Keywo ds: s able iso opes, ophic g oups, neus on, biogeochemical p o inces, niche size Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 2 Janua y 2021 | Volume 7 | A icle 606088 INTRODUCTION The neus on, one o he less desc ibed and known aqua ic ecological g oups, is pa adoxically he closes o ou sampling pla o ms as i inhabi s he uppe cen ime e s o he ocean. The e m neus on was coined in 1972 (Hempel and Weike , 1972) o de ine he pelagic o ganisms ha occupy he icini y o he su ace laye , albei o en in a empo ally- es ic ed and a iable manne . Neus on occupies a delimi ed ecological niche and is gene ally g ouped in o h ee ecological ca ego ies: (a) euneus on: o ganisms wi h maximum abundance in he icini y o he su ace on which hey eside day and nigh ; (b) acul a i e neus on: o ganisms concen a ing a he su ace only du ing ce ain hou s o he day, usually du ing da kness; and (c) pseudoneus on: o ganisms wi h maximum concen a ions a deepe laye s bu eaching he su ace laye a leas du ing ce ain hou s (Ma shall and Bu cha d , 2005). The neus onic communi y s uc u e is condi ioned by sunligh and an a ay o endogenous (o ganic ma e , espi a o y, pho osyn he ic, decomposi ional p ocesses) and exogenous (a mosphe ic deposi ion, ino ganic ma e , winds, wa e ac ion, p ecipi a ion, UV adia ion, oceanic cu en s, su ace empe a u e) a iables and p ocesses a ec ing nu ien inpu s and ecycling (Ma shall and Bu cha d , 2005; Rawlinson e al., 2005; Rezai e al., 2019). Fu he mo e, he neus on p o ides a ood sou ce o he zooplank on mig a ing om deepe laye s o he su ace (Hempel and Weike , 1972), as well as o seabi ds oaming o e he oceans (Cheng e  al., 2010). Fo hese easons, he neus onic communi y is belie ed o play a c i ical ole on he s uc u e and unc ion o ma ine ood webs. Ye , esea ch on neus on communi ies o da e ocused p edominan ly on geog aphically- limi ed egions o he ocean (Zai se , 1971; Hempel and Weike , 1972; Holdway and Maddock, 1983; Ebbe s and Wing, 1997; Rezai e  al., 2019) o coas al a eas (B odeu , 1989; Le Fe e and Bou ge , 1991; Padma a i and Goswami, 1996). Consequen ly, neus on complexi y is s ill poo ly unde s ood as s udies on he communi y s uc u e and he axonomical composi ion o o ganisms inhabi ing his ecological niche emain ew (Rezai e  al., 2019), and global scale analyses a e ye lacking. The neus onic animals o m a subse o he zooplank on communi y, which plays a pi o al ole in he unc ioning o ma ine ecosys ems. Zooplank on a e pa ially esponsible o he ac i e ene gy lux be ween supe icial and deep laye s o he ocean (Tu ne , 2002; Jónasdó i e  al., 2015; He nández-León e  al., 2020). Zooplank on species composi ion, biomass, and seconda y p oduc ion in luence a wide ange o ophic le els in ma ine communi ies, as hey cons i u e a link be ween p ima y p oduc ion and seconda y consume s (Li chman e  al., 2013; Benede i e  al., 2016; de Oli ei a Sod é and Bozelli, 2019). Copepods cons i u e he mos abundan zooplank on axon in e ms o biomass and di e si y wo ldwide (Kiø boe, 2011; Neumann-Lei ão e al., 2018); he e o e, changes in hei communi y composi ion can hus impac he biogeochemical cycles (Bianchi and Mislan, 2016) and migh be indica i e o clima e a iabili y impac s on ecosys em unc ioning (Hoo and Pe e son, 2006). His o ically, zooplank on assemblages esea ch has ocused mainly on axonomic s udies and hose ela ed o communi y s uc u e (Pome leau e  al., 2015). Howe e , ecen ly, esea ch has ee ed owa d an al e na i e ai -based app oach (Pome leau e  al., 2015; Benede i e  al., 2016; Campos e  al., 2017), p o iding a pe spec i e mo e ocused on g oups o species wi h analogous unc ional ai s. This allows indi iduals o be classi ied in o ypes cha ac e ized by he p esence/absence o ce ain alleles o a gene, in o size classes, ecological guilds, o unc ional g oups (FGs; Tuomis o, 2010). Func ional ai s a e pheno ypes a ec ing o ganism i ness, g ow h, su i al, and ep oduc i e abili y (Violle e  al., 2007; de Oli ei a Sod é and Bozelli, 2019). These a e egula ed by he exp ession o genes wi hin species, and he exp ession o ai s egula e, in u n, he species i ness unde con as ing bio ic and abio ic ci cums ances (Ba on e al., 2013). Mo eo e , a speci ic unc ional ai can also de elop om he in e ac ions be ween o he ai s and en i onmen al condi ions (Kiø boe, 2011), leading o a gi en ai g ouping being a o ed unde ce ain condi ions. Zooplank on ai s can beclassi ied in acco dance o ecological unc ions – eeding, g ow h, ep oduc ion, su i al, and o he cha ac e is ics such as mo phology, physiology, beha io , o li e his o y (Li chman e  al., 2013; Hun e  al., 2015; B un e al., 2016). Pa icula ly, eeding s a egies and ophic g oups a e ele an o asce ain eeding e iciency and associa ed p eda ion isk (B un e  al., 2017). Addi ionally, hey acili a e he unde s anding o ecosys em se ices associa ed wi h zooplank on, such as he dis ibu ion o ishe ies o biogeochemical cycling (P owe e al., 2019) while also allowing he posi ioning o zooplank on axa in he ood web (Benede i e  al., 2016, 2018). S able iso ope analysis (SIA) has been widely used o explo e he ood web s uc u e; o iden i y an o ganism’s ophic posi ion; o quan i y ca bon, ni ogen, and ene gy luxes; and o cha ac e ize ophic niches (F y, 2006; Bouillon e  al., 2011; Middelbu g, 2014). Gene ally, s able iso opes unde go a p edic able ophic en ichmen be ween p ey and consume (Minagawa and Wada, 1984) and e lec he o ganism’s die o e a conside able pe iod o ime (Vande Zanden e  al., 2015). Ra ios o ca bon iso opes (δ13C) gene ally ha e a low ophic en ichmen and a e commonly used o iden i y ca bon sou ces (DeNi o and Eps ein, 1978; Vande Zanden and Rasmussen, 2001; Pos , 2002a). Ni ogen iso ope a ios (δ15N) show p og essi e en ichmen be ween p ey and consume s (DeNi o and Eps ein, 1978; Minagawa and Wada, 1984; Pos , 2002b; McCu chan e  al., 2003) and we e hus employed o es ima e ophic posi ions (F y and She , 1984; Minagawa and Wada, 1984; Pe e son and F y, 1987; Owens, 1988; Pos , 2002b). Thus, by measu ing he a ios o δ13C and δ15N, i is possible o in e he ophic s uc u e o ma ine ood webs (F y, 2006). Recen ly, SIA has been inc easingly employed o he cha ac e iza ion o ophic niche (Layman e al., 2007a,b, 2012; Hun e  al., 2015). The ophic niche o a single species, communi y, o ecosys em is he agg ega e o he in e ac ions be ween i s cons i uen s and he ecosys em (El on, 1927), hence ep esen a i e o he cha ac e is ics o i s habi a and ophic posi ion (Leibold, 1995). The ophic niche can be in e ed om he iso opic niche (e.g., he δ13C–δ15N bi-plo space) o unco e ele an aspec s o ophic s uc u e. Fo ins ance, Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 3 Janua y 2021 | Volume 7 | A icle 606088 niche-based quan i a i e me ics can be used o es ecological heo y and ophic esponses o an h opogenic impac s (Layman e  al., 2007a; Schmid e  al., 2007), including he deg ee o o e lap be ween dis inc ophic niches (Jackson e  al., 2011). To ou knowledge, he use o iso opic niche me ics on he s udy o he neus onic zooplank on communi y has ne e been a emp ed be o e. He e, we b idge his gap by explo ing he a ia ions in he size o he ophic niche o neus onic zooplank on ac oss sub opical and opical oceanic p o inces. Ou main hypo hesis is ha he ophic s uc u e o he neus onic communi y is p ese ed in spi e o di e ences in nu ien sou ces and p oduc i i y along oceanic p o inces. Mo e speci ically, in his s udy we (1) cha ac e ize a ios o C and N o neus on ac oss sub opical and opical oceanic p o inces, (2) de e mine ophic s uc u e simila i ies ac oss oceanic p o inces, and (3) analyze he ela ionships be ween selec ed en i onmen al a iables and he ophic s uc u e among oceanic p o inces. Wedo so based on he samples collec ed along he Malaspina Ci cumna iga ion Expedi ion, which ci cumna iga ed he sub opical and opical ocean in 2010–2011 (Dua e, 2015). MATERIALS AND METHODS Sample Collec ion The neus on samples we e collec ed along he Malaspina 2010 Expedi ion, which ci cumna iga ed he globe and was ca ied ou be ween Decembe 2010 and July 2011 ac oss opical, sub opical, and empe a e egions o he A lan ic, Indian, and Paci ic Oceans be ween 35° N and 40° S (Dua e, 2015). Sampling s a ions we e dis ibu ed o cha ac e ize pelagic communi ies ac oss egions o he open ocean in he no he n and sou he n hemisphe e (Dua e, 2015). In o de o allow o he in e compa abili y o he obse a ions and a oid ad e se wea he du ing sampling, he c uise was scheduled o isi mos egions du ing hei sp ing-summe mon hs. The sampling loca ions we e assigned o Longhu s Biogeochemical p o inces (Longhu s , 2007). The e, ou biomes (Pola , Wes e lies, T ades and Coas al) and a o al o 56 p o inces we e iden i ied, based on he cha ac e iza ion o p ima y p oduc ion, mixed dep h laye , nu ien s a ailabili y, pho ic dep h, algal biomass, B un - Väisälä equency, and he Rossby adius o in e nal de o ma ion. Speci ically, he neus on samples epo ed he e included 10 oceanic Longhu s p o inces (Longhu s , 2007). Fou p o inces in he A lan ic [No heas A lan ic sub opical gy al (NASE; n=16); No h A lan ic opical gy al (NATR; n=14); Wes e n opical A lan ic (WTRA; n=15); Sou h A lan ic gy al (SATL; n=12)], ou in he Paci ic [Sou h Paci ic gy e (SPSG; n=18); Paci ic equa o ial di e gence (PEQD; n = 12); No h Paci ic opical gy e (NPTG; n=14); No h Paci ic equa o ial coun e cu en (PNEC; n=15)], and wo in he Indian Ocean [Indian Sou h sub opical gy e (ISSG; n = 18); Sou h sub opical con e gence (SSTC; n=16); Figu e1; Supplemen a y Table S1]. Be ween h ee and ou s a ions we e sampled wi hin each p o ince. Samples we e collec ed wice a day a 12 pm and 4 am, by owing a neus on sample wi h a mou h opening o 80 × 30 cm and a mesh size o 200 μm, a 2–3 kno s o 10–15 min, which sampled he i s 15 cm o he wa e a a dis ance o 5 m om he s a boa d o he essel (González-Go dillo e  al., 2012). The con en o each sample FIGURE1 | Loca ion o he s a ions sampled in his s udy, indica ed by g een ci cles. S a ions we e g ouped in dis inc biogeochemical p o inces acco ding o Longhu s (2007) wi h ac onyms depic ed. Shape ile adap ed om Flande s Ma ine Ins i u e (2009). Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 4 Janua y 2021 | Volume 7 | A icle 606088 was s o ed in 4% o maldehyde un il analysis. En i onmen al da a we e acqui ed using CTD cas s deployed a each s a ion om he su ace o 4,000 m o 100 m abo e he sea loo when his was shallowe han 4,000 m. Sample P ocessing, Taxonomical Iden i ica ion, and Func ional G ouping Samples we e s ained wi h Bengal Rose in o de o acili a e he iden i ica ion o he o ganisms o axonomic le el o a amily which, whene e possible, was de e mined unde a s e eomic oscope (Olympus SZX16) using app op ia e guides (Bol o skoy, 1999; Cas ellani and Edwa ds, 2017). The mos abundan and ep esen a i e axa selec ed o SIA we e copepods o he amilies Aca iidae, Calanidae, Co ycaeidae, Oncaeidae, and Pon ellidae, and Phylum Chae ogna ha. The la e we e included as a ep esen a i e o op plank onic p eda o s. Each o hese axa ha e speci ic ophic oles, acco ding o he li e a u e. Fo ins ance, copepod amilies we e assigned o FGs acco ding o Benede i e  al. (2016, 2018). Fou FGs we e conside ed and axa we e assigned as ollows: FG1=de i i o es (Oncaeidae); FG2 = he bi o es/omni o es (Aca iidae and Calanidae); FG3 = ca ni o es (Co ycaeidae and Pon ellidae); and FG4 = p eda o s (Chae ogna ha). All specimens selec ed o analysis we e adul s acco ding o hei mo phological aspec unde he mic oscope. S able Iso ope Analysis Analysis o he na u al abundance o ca bon and ni ogen iso opes we e pe o med on p e iously d ied (50°C, 48 h) neus on samples. E en when specimens we e classi ied a genus o species le el, we pooled indi iduals om he same sample s a ion, including day and nigh samples, a amily (Aca iidae, Calanidae, Co ycaeidae, Oncaeidae, and Pon ellidae) o phylum (Chae ogna ha) le el in o de o ob ain ca. 1 mg d y weigh o iso opic de e mina ion and o achie e a minimum o h ee da a poin s o each unc ional g oup and minimize he cons ain s o small sample size (Jackson e  al., 2011). D ied samples we e packed in o in capsules and measu ed in an elemen al analyze (Ca lo E ba CHNSO 1108) coupled o an iso ope- a io mass spec ome e (Finnigan Ma Del a Plus). Iso opic analyses we e pe o med by he Se icio de Análisis Ins umen al o he Uni e sidade da Co uña (Spain). Samples we e no acidi ied in o de o emo e ca bona es since he selec ed axa we e only sligh ly calci ied, and i has been demons a ed ha his p ocedu e could impac ni ogen measu emen s (Ma eo e  al., 2008). No co ec ions we e made o he possible e ec o o maldehyde on he s able iso ope composi ion as mos s udies speci ically made on ma ine zooplank on samples poin ou o no signi ican e ec s in δ15N and gene ally a dec ease o less han 2‰ in δ13C a e se e al yea s o s o age (e.g., Mullin e  al., 1984; Bicknell e al., 2011; de Lecea e al., 2011). Taking in o conside a ion ha all samples ha e been p ese ed simila ly and ha he ime be ween collec ion and SIA la gely exceeded 1 yea , we assumed ha di e ences in s able iso ope alues e lec genuine a ia ions in he p o inces and selec ed FGs. Values o na u al abundance o s able iso opes we e exp essed as δ13C and δ15N (‰) ela i e o Vienna Pee Dee Belemni e and a mosphe ic ni ogen, espec i ely (Coplen, 2011). Ce i ied iso ope s anda ds (USGS40 and L-alanine) we e analyzed along wi h in e nal ace anilide and sample s anda ds wi h s anda d de ia ion (SD) be ween ce i ied and measu ed alues <0.1‰. The p ecision [s anda d e o (SE)] o eplica e de e mina ions o s anda ds and samples was <0.05‰ o bo h iso opes (n=4). As he C:N mass a io o mos samples exceeded 3.5 ( hus sugges ing a signi ican and a iable lipid con en ), δ13C alues we e no malized using an empi ical linea eg ession wi h he sample C:N alue de e mined o aqua ic o ganisms (Pos e al., 2007). This p ocedu e aimed a emo ing he e ec o he low δ13C associa ed o lipids and was p e e ed due o low biomass cons ain s ha hinde ed he use o lipid emo al me hodologies such as dichlo ome hane and an accele a ed sol en ex ac ion sys em (Bodin e  al., 2009). En i onmen al Va iables En i onmen al a ibu es o each o he sampled s a ions we e cha ac e ized by a numbe o a iables collec ed in si u o sa elli e-de i ed. (Supplemen a y Table S2). These a iables we e al eady employed in a p e ious analysis o Malaspina 2010 c uise, whe e me hodological de ails can be ound (Mompeán e al., 2013, 2016b; Fe nández-Cas o e al., 2015). In b ie , he s a i ica ion o he wa e column was ep esen ed by he dep h o he mixing laye (MLD, m), he mean squa ed B un –Väisälä equency (N2, s−2), and he dep h o he chlo ophyll maximum (DCM, m), all es ima ed om e ical p o iles o a CTD equipped wi h a luo escence senso (Fe nández-Cas o e al., 2015). Simila ly, nu ien inpu s om deep laye s we e es ima ed by di usi i y due o u bulence (KT, m2 s−1), de e mined om e ical cas s o a mic os uc u e u bulence p o ile (Fe nández-Cas o e  al., 2015). Phy oplank on biomass was ep esen ed by su ace and pho ic- zone in eg a ed chlo ophyll-a (Chlas and Chlai, mg m−3 and mg m−2, espec i ely), de e mined om ace onic ex ac s o phy oplank on (Es ada e al., 2016). Mic oplank on (40–200μm) cha ac e is ics we e indica ed by an abundance o he ni ogen- ixe T ichodesmium (T icho, cells ml−1), ca bon biomass (C40–200, mg C m−3), and na u al abundance o ni ogen iso opes (δ15N40–200, ‰) measu ed in samples collec ed by e ical ows o a plank on ne be ween he su ace and 200 m dep h (Mompeán e  al., 2013, 2016b). Sa elli e de i ed a iables we e conside ed o ep esen condi ions p e ailing o e la ge spa ial and empo al scales han hose conside ed du ing he speci ic sampling o each s a ion and included annual a e ages o p ima y p oduc ion (PP) o 2010 (mg C m−2 d−1) and mean mon hly a mosphe ic dus deposi ion (MDU, g m−2mon h−1). The o me ep esen ed egional p oduc i i y and was de i ed om he da a p o ided by he Ocean P oduc i i y websi e (h p://www.science. o egons a e.edu/ocean.p oduc i i y/index.php) in a g id o 0.17°×0.17° including each s a ion posi ion. Dus deposi ion, a p oxy o a mosphe ic inpu s o key nu ien s as Fe o P, was es ima ed om Aqua-MODIS Ae osol Op ical Dep h a 550 nm and Ae osol Small Mode F ac ion da a p o ided by he Gio anni online da a sys em (NASA Godda d Ea h Sciences) Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 5 Janua y 2021 | Volume 7 | A icle 606088 om a g id o 1°×1° nea each s a ion (Mompeán e al., 2016b). In dep h in o ma ion on sampling and analy ical me hodology employed h oughou he Malaspina 2010 c uise can be ound in Mo eno-Os os (2012). S a is ical Analysis Di e ences in δ13C and δ15N a ios o Oncaeidae among oceanic p o inces we e es ed wi h ANOVA and Bon e oni pos hoc es on S a is ica 12 (S a So , Inc., Tulsa, OK, USA). Analysis o he di e ences in s able iso ope alues among oceanic p o inces was assessed using a pe mu a ional mul i a ia e analysis o a iance (PERMANOVA) and pai wise es s on PRIMER 6.0 (Cla ke and Go ley, 2006) and he add-on package PERMANOVA+ (Ande son e  al., 2008; Supplemen a y Table S3). Fo subsequen analysis, a ep esen a i e p ima y consume was used o no malize s able iso ope alues acco ding o Cla k and F i z (1997) and S asko e  al. (2018) o accoun o spa ial he e ogenei y in ca bon and ni ogen a ios. To his end, consume δ13C alues we e no malized ela i e o a pelagic baseline (Δ13Cpel) as: ∆13 3 1000 1000 110Cpel w c =+ +−          × d d, we e δC is he consume δ13C alue and δW is he p o ince-speci ic mean δ13C alue o he copepod amily Oncaeidae, conside ed as he e e ence baseline. Consume δ15N alues we e no malized by sub ac ing he p o ince-speci ic mean δ15N alue o he copepod amily Oncaeidae om he consume δ15N, ep esen ed he ea e as δ15Nn. The use o Oncaeidae as baseline is no in ended o e lec he base o he ood web bu a he o se a homogeneous s a ing poin along he δ13C and δ15N con inuum. All s a is ical analyses we e om his poin on pe o med on Oncaeidae- no malized alues. Analysis o he iso opic niche size we e made in he iso opic space de ined by Δ13Cpel–δ15Nn o each p o ince/ unc ional g oup combina ion wi h he package SIBER (Jackson e  al., 2011). Bayesian es ima es o he s anda d ellipse a ea (SEAb) we e calcula ed o each combina ion o cha ac e ize he ull a iabili y in o aging habi s and used esou ces (Layman e  al., 2007a, 2012). These es ima es accoun ed o unce ain y due o he numbe o samples. A minimum o h ee samples was equi ed o hese es ima ions. SEAb s a is ics we e compu ed om 104 simula ions pe p o ince/ unc ional g oup a angemen , al hough only he maximum likelihood ellipses we e used o g aphical ep esen a ions. Pai wise compa isons o SEAb (Supplemen a y Table S4) we e pe o med by calcula ing he p opo ion o ellipse size ha di e ed be ween wo gi en combina ions being in e p e ed as a di ec p oxy o he p obabili y ha one combina ion is di e en om he o he (Jackson e  al., 2011). Because o he limi a ions imposed by he a ailable sample biomass o iso opic analyses, da a o some FGs could no be ob ained o all p o inces. P incipal componen analysis (PCA) was used o in es iga e he ela ionships be ween en i onmen al a iables and he niche size (as a p oxy o ophic s uc u e) o he FGs among oceanic p o inces using PRIMER 6.0 (Cla ke and Go ley, 2006). Fi s , no malized en i onmen al a iables measu ed on each sampling si e we e employed o compu e he eigen alues, eigen ec o s, and p incipal componen (PC) coo dina es. Then, Pea son’s co ela ion coe icien be ween SEAb a e age and 95% c edible in e als (CI) ange o each unc ional g oup/p o ince combina ion and he en i onmen al PC coo dina es we e compu ed using he co () unc ion on R 3.6.3 (R Co e Team, 2020). Finally, he co ela ion coe icien s be ween he niche me ics and he PCA axes we e plo ed on he a iable space o he PCA i s wo p incipal componen coo dina es axis. RESULTS δ13C and δ15N Ra ios The mean δ13C o Oncaeidae a ied signi ican ly among p o inces (Figu e2A). Bon e oni pos hoc es e ealed h ee dis inc g oups (a, b, and c) o dec easing mean alues (−21.2, −21.8, and −22.3‰, espec i ely). The highes alues we e ound in NATR and SPSG p o inces, while he lowes alues occu ed in NASE. Simila ly, signi ican di e ences in mean alues o δ15N be ween p o inces we e also ound (Figu e2B). All δ15N alues we e posi i e, wi h maximum alues in SSTC p o ince, ollowed by hose in PNEC and NPTG, and minimum alues (<2‰) in NATR. Bon e oni pos hoc es s also e ealed h ee g oups (d, e, and ) o dec easing mean δ15N (10.1, 5.4, and 3.5 ‰, espec i ely). A B FIGURE2 | Mean and SE alues o δ13C (A) and δ15N (B) o Oncaeidae by Longhu s P o inces. G oups o signi ican ly di e en means (ANOVA and Bon e oni pos hoc es , p<0.05) a e indica ed by di e en le e s (a, b, c, d, e, and ). Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 6 Janua y 2021 | Volume 7 | A icle 606088 Iso opic Niche Size o Func ional G oups The analysis o he iso opic niches e ealed di e ences be ween p o inces and FGs in all oceans. The maximum likelihood ellipses we e dis inc om each o he in all p o inces and FGs bu showed la ge o e laps be ween hem in some p o inces/ unc ional g oup a angemen s. A lan ic P o inces The e was a clea sepa a ion o he ellipses o op p eda o s (FG4) and de i i o es (FG1) in he p o inces whe e hese g oups we e analyzed (Figu e 3). In con as , ellipses o he bi o es/omni o es (FG2) and ca ni o es (FG3) showed la ge o e laps. The a iabili y o SEAb (Bayesian s anda d ellipse a ea) was highes o FG3 and lowes o FG2 (Table 1). Pai wise compa isons o he iso opic niche a eas using SEAb e ealed ha only FG3 a ied signi ican ly om small alues in NASE, in e media e in NATR and high alues in SATL, while no signi ican di e ences esul ed when compa ing o he FG among A lan ic p o inces (Supplemen a y Table S4). Indian P o inces Dis inc maximum likelihood ellipses we e p esen in bo h p o ince/FGs a ays; none heless, signi ican o e laps a e disce nible h oughou he isospace (Figu e4; Table 1). SEAb was consis en ly la ge o all FGs in he ISSG p o ince and he la ges ophic a iabili y was ound o FG2. Mo eo e , SEAb we e he smalles o bo h FG1 and FG3 in SSTC. The e we e no signi ican di e ences in SEAb o equi alen FG among Indian p o inces (Supplemen a y Table S4). Paci ic P o inces The isospace a ea o he Paci ic p o inces ollows he end desc ibed o A lan ic and Indian p o inces, wi h pe cep ible di e ences in he dis inc p o inces/FGs a angemen s, bu he la ge o e lap was ound in many cases, especially in PNEC and SPSG (Figu e 5). The la ges alues o SEAb we e ound o FG4 in SPSG and he minimum alues o FG2 in PEQD (Table1). NPTG and PEQD p esen ed na owe SEAb compa ed o SPSG and PNEC. Pai wise compa isons o SEAb indica ed ha he alues o FG1 we e signi ican ly smalle o PNEC compa ed o hose o SPSG and ha hose o FG2in NPTG and PEQD we e smalle han hose in PNEC and SPSG, while hose o FG3 in PNEC and SPSG we e la ge han hei equi alen a eas in NPTG and PEQD. Finally, SEAb o FG4in PNEC and SPSG we e la ge han hei compa able g oup in NPTG (Supplemen a y Table S4). En i onmen al Va iables and T ophic S uc u e The PCA on he en i onmen al a iables e ealed dis inc i e condi ions among p o inces, as s a ions a e clea ly clus e ed wi hin each p o ince (Figu e 6). The i s wo p incipal componen s o he PCA accoun ed o 66.4% o he o al a iance (Table 2). Su ace chlo ophyll had he la ges posi i e FIGURE3 | Isospace o Δ13Cpel and δ15Nn including indi idual sample no malized alues o neus onic zooplank on o biogeochemical p o inces in he A lan ic Ocean: NASE, NATR, WTRA, and SATL. The maximum likelihood es ima es o he s anda d ellipse a eas ( hick lines) o each unc ional g oup (FG) a e shown. FG 1=de i i o es; FG 2=he bi o es/omni o es; FG 3=ca ni o es; FG 4=p eda o s. Dec eases in Δ13Cpel e lec inc eases in aw δ13C alues due o no maliza ion. Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 7 Janua y 2021 | Volume 7 | A icle 606088 loadings wi h he i s componen (PC1), ollowed by he mic oplank onic ca bon biomass and u bulence di usi i y (Figu e7). Con e sely, PC1 was nega i ely co ela ed wi h he dep h o chlo ophyll maximum. The second componen (PC2) was posi i ely co ela ed wi h p ima y p oduc ion, δ15N40–200, and he mixed laye dep h, while showing nega i e co ela ions wi h a mosphe ic dus deposi ion, T ichodesmium abundance, and pho ic-zone in eg a ed Chl-a. The emaining en i onmen al a iables ha e a mino ele ance in explaining he epo ed di e ences (Figu e 7). The size o he iso opic niche (indica ed by median SEAb alues and 95% CI) o FG4 was posi i ely co ela ed wi h PC1, while hose o FG3 showed nega i e co ela ions wi h his componen . In u n, SEAb o FG1 and FG2 we e mainly nega i ely co ela ed wi h PC2 (Figu e 7). DISCUSSION In he cu en s udy, we p o ide he i s compa ison o he ophic s uc u e o he neus onic communi y in opical and sub opical p o inces o he A lan ic, Indian, and Paci ic Oceans (Longhu s , 2007; Dua e, 2015). The analysis o s able ca bon and ni ogen iso opes o neus on FGs allowed o he cha ac e iza ion o ca bon and ni ogen a ios, he quan i ica ion o he size o he iso opic niche as a p oxy o ophic s uc u e, and i s ela ionships wi h en i onmen al a iables ac oss dis inc oceanic p o inces. No wi hs anding he limi a ions imposed by he analy ical equi emen and, hus, he small numbe o samples employed, as well as he samples p ese a ion me hod, he esul s o his s udy allow us o es ablish an app oxima ion o he ophic in e ac ions in he neus on and hei ole in pelagic ood webs a a global scale. δ13C and δ15N Ra ios The di e ences ound in ca bon and ni ogen appea o e lec a ia ions in he en i onmen al egimes o he dis inc p o inces. P o ided ha di e en phy oplank on axa a o he use o dis inc ca bon sou ces o p ima y p oduc ion, i is plausible ha i could lead o a ia ions in δ13C ha can be passed upon he ood web. Fo ins ance, hea ie δ13C alues ha e TABLE1 | Iso opic niche a ea (‰2) es ima es o each unc ional g oup and p o ince. Ocean P o ince FG SEAb 95% CI A lan ic NASE 1 0.446 0.130 o 2.128 2 0.339 0.088 o 1.144 3 0.377 0.115 o 0.999 4 0.212 0.040 o 0.929 NATR 2 0.481 0.173 o 1.403 3 1.259 0.526 o 3.921 WTRA 1 0.866 0.198 o 4.276 2 0.832 0.211 o 3.497 3 0.597 0.250 o 1.540 4 0.524 0.140 o 2.592 SATL 2 0.596 0.194 o 2.094 3 4.508 1.695 o 13.427 Indian ISSG 2 1.411 0.598 o 3.888 3 1.001 0.500 o 2.510 4 0.468 0.134 o 2.273 SSTC 1 0.177 0.047 o 0.821 2 0.826 0.348 o 1.973 3 0.054 0.015 o 0.238 4 0.221 0.061 o 1.047 Paci ic NPTG 2 0.338 0.119 o 1.253 3 0.554 0.167 o 1.541 4 0.250 0.063 o 1.284 PEQD 2 0.088 0.024 o 0.393 3 0.543 0.142 o 1.709 4 0.998 0.220 o 4.597 PNEC 1 0.397 0.107 o 2.010 2 1.258 0.367 o 4.325 3 2.913 1.027 o 8.495 4 2.232 0.487 o 10.505 SPSG 1 2.162 0.444 o 10.884 2 1.897 0.802 o 4.899 3 3.177 0.938 o 7.988 4 3.361 0.598 o 14.937 Es ima es a e p esen ed as he mode o he Bayesian s anda d ellipse a ea (SEAb). Lowe and uppe 95% CI indica e he unce ain y in he SEAb es ima es. FIGURE4 | Isospace o Δ13Cpel and δ15Nn including indi idual sample no malized alues o neus onic zooplank on o biogeochemical p o inces in he Indian Ocean: ISSG and SSTC. The maximum likelihood es ima es o he s anda d ellipse a eas ( hick lines) o each FG is shown. FG 1=de i i o es; FG 2=he bi o es/ omni o es; FG 3=ca ni o es; FG 4=p eda o s. Dec eases in Δ13Cpel e lec inc eases in aw δ13C alues due o no maliza ion. Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 8 Janua y 2021 | Volume 7 | A icle 606088 been associa ed wi h highe phy oplank on ac iona ion o ino ganic ca bon, and subsequen sinking o o ganic ma e ela i ely deple ed in 13C om he su ace wa e s led o high alues in he uppe ocean a sub opical and opical wa e s (G ube e al., 1999; Schmi ne e al., 2013). Mo eo e , up ake o a mosphe ic CO2 ia ai -sea exchange and consequen ac iona ion in he con e sion p ocess o HCO−3 can esul in high δ13C-dissol ed ino ganic ca bon (Zhang e  al., 1995). Such p ocesses could explain he en ichmen in 13C (i.e., highe δ13C) ound o he iso opic baseline (Oncaeidae) in he SPSG and NATR. Addi ionally, he en ichmen in δ13C measu ed in SPSG could also be ela ed o he p esence o dia oms, as epo ed in he egion du ing he same c uise (Es ada e  al., 2016). Al hough hey we e no he dominan phy oplank on axa (Es ada e  al., 2016), hey may ha e disp opo iona ely con ibu ed o he me azoan ood web, while he dominan picocyanobac e ia would con ibu e o he mic obial ood web. While cyanobac e ia up ake o ino ganic ca bon occu s ia di ec HCO−3 anspo , dia oms a e also capable o anspo ing CO2 de i ed om he ca alyzed dehyd a ion o HCO−3 (To ell and Mo el, 2002). This implies a la ge iso opic ac iona ion (i.e., less nega i e alues o δ13C) in dia oms when compa ed wi h cyanobac e ia o o he mic oalgae (F y and Wain igh , 1991). Ano he hypo hesis o he δ13C en ichmen measu ed in NATR p o ince is ela ed o he abundance o Sa gassum spp. on he wes e nmos s a ions (Gou êa e  al., 2020). FIGURE5 | Isospace o Δ13Cpel and δ15Nn including indi idual sample no malized alues o neus onic zooplank on o biogeochemical p o inces in he Paci ic Ocean: NPTG, PEQD, PNEC, and SPSG. The maximum likelihood es ima es o he s anda d ellipse a eas ( hick lines) o each FG a e shown. FG 1=de i i o es; FG 2=he bi o es/omni o es; FG 3=ca ni o es; FG 4=p eda o s. Dec eases in Δ13Cpel e lec inc eases in aw δ13C alues due o no maliza ion. FIGURE6 | P ojec ion o samples on he space o he i s (PC1) and second (PC2) componen s o he p incipal componen analysis (PCA) o en i onmen al a iables (DCM: dep h o he chlo ophyll maximum (m), δ15N40–200: na u al abundance o mic oplank on ni ogen iso opes (‰), PP: annual mean p ima y p oduc ion (mg C m−2 d−1), MLD: mixed laye dep h (m), C40–200: mic oplank on (40–200μm) ca bon biomass (mg C m−3), KT: di usi i y due o u bulence (m2 s−1), Chlas: su ace chlo ophyll-a (mg m−3), N2: mean squa ed B un -Väisälä equency (s−2), Chlai: pho ic-zone in eg a ed chlo ophyll-a (mg m−2), T icho: abundance o T ichodesmium (cells ml−1), MDU: mean mon hy a mosphe ic dus deposi ion (g m−2 mon h−1)) de e mined o neus on sampling s a ions. P o ince ac onyms as in Supplemen a y Table S1. Albuque que e al. T ophic S uc u e o Neus on F on ie s in Ma ine Science | www. on ie sin.o g 9 Janua y 2021 | Volume 7 | A icle 606088 This mac oalgae has a δ13C anging om −16‰ o −18‰, and i is concei able ha i could ha e in luenced he δ13C baseline in NATR (Cabanillas-Te an e  al., 2019). Howe e , di e ences in δ13C could be he esul o many ac o s associa ed wi h changes in phy oplank on CO2 ixa ion, which is known o a y in ela ion o empe a u e, concen a ion o aqueous CO2, phy oplank on composi ion, and he a ailabili y o dissol ed ino ganic ca bon and nu ien s (Wong and Sacke , 1978; Descolas-G os and Fon ugne, 1990; F ancois e  al., 1993; Bu kha d e  al., 1999; Popp e  al., 1999). Bo h con inen al o ganic ma e and a mosphe ic CO2 a e cha ac e ized by ha ing lowe alues o δ13C han hei equi alen s in oceanic (Pe y e  al., 1999) o upwelled wa e s (G ube e  al., 1999). Ou esul s ag ee wi h he epo o highe δ13C alues in zooplank on om oligo ophic egions o he sub opical No h A lan ic du ing he same c uise (Mompeán e al., 2013). In con as , mo e nega i e δ13C alues o he Oncaeidae baseline, such as hose measu ed in NPTG and NASE, can beexplained by he dominance o dino lagella es and coccoli hopho es in hese p o inces (Es ada e  al., 2016) and by he in luence o upwelling in NASE (G ube e al., 1999; Mompeán e al., 2013). Ra ios o ni ogen also a ied be ween p o inces, acco ding o he di e ences ound in δ15N alues o Oncaeidae. Low δ15N alues could bedue o diazo ophy since he a mosphe ic ni ogen p esen s δ15N = 0‰, and i can be aced h ough he ood web (McClelland e al., 2003; Mompeán e al., 2016a; Bode and He nández-León, 2018). This p ocess is media ed by specialized p oka yo es – diazo ophs – which h i e in usually ni a e-poo , wa m and s a i ied wa e s such as he ones o he sub opical and opical gy es (Falkowski, 1997; Zeh e  al., 2003; Capone e  al., 2005; Luo e  al., 2012) and in oduce in o he ocean bioa ailable ni ogen deple ed in δ15N (Somes e al., 2010). Addi ionally, i is also known ha iso opic ac iona ion occu s du ing phy oplank on up ake, and his may cause low δ15N alues (<5‰) in plank on when dissol ed ni ogen concen a ions a e high, as epo ed a he ini ial phases o blooms (Wase e al., 2000). In u n, high concen a ions o 15N-en iched ni a e o ma ine o igin a e simila ly e lec ed in he δ15N alues o zooplank on (Owens, 1988; Mon oya e al., 2002), as ound in he SSTC samples. Ou esul s s ongly sugges ha he low δ15N alues measu ed o Oncaeidae e lec he ole o N2 ixa ion by diazo ophs in supplying N in some p o inces (mainly in NATR bu also in NASE, WTRA, SATL, SPSG, and PEQD), while he high δ15N alues measu ed in o he p o inces indica e inpu s o ni ogen de i ed om deni i ica ion p ocesses (e.g., PNEC, NPTG, and SSTC). These assump ions a e suppo ed by he consis ency wi h es ima ions o he abundance o he N- ixe T ichodesmium (Es ada e al., 2016; Mompeán e  al., 2016b) and N ixa ion a es and NO3− di usi e luxes ob ained o he Malaspina c uise (Fe nández- Cas o e  al., 2015), as well as ecen es ima es o ma ine ni ogen ixa ion and deni i ica ion a he scale o he global ocean (Knapp e  al., 2016; Bonne e  al., 2017; G ube , 2019; Wang e  al., 2019). Mo eo e , he δ15N alues epo ed he e o Oncaeidae a e consis en wi h hose o mic oplank on collec ed simul aneously du ing he Malaspina c uise (δ15N40–200) in he uppe 200 m o he wa e column, hus suppo ing ou assump ion o he ole o Oncaeidae as p ima y consume s and c i ical nodes in he ood web. Fu he mo e, he gene al cohe ence be ween he dis ibu ion pa e n o bo h Oncaeidae and mic oplank on δ15N ac oss p o inces (e.g., Mompeán e  al., 2016b; Supplemen a y Figu e S1) sugges s simila u no e imes o he s able iso opes (and hence g ow h a es) in epipelagic mic oplank on and, in ou case, mesoneus on (>200 μm). Spa ial Di e ences in Iso opic Niche The small laye o su ace wa e he neus onic communi y inhabi s would imply a sca ce a ailabili y o ood sou ces, pa icula ly in mo e oligo ophic egions, such as mos o p o inces in his s udy. This sugges s a s ong compe i ion o TABLE2 | Eigen alues, pe cen o o al a iance (%Va iance), and pe cen o cumula i e a ia ion (Cum. %Va ia ion) explained o he i s i e p incipal componen s (PC) o he p incipal componen analysis (PCA) on en i onmen al a iables. Componen Eigen alue %Va iance Cum. %Va ia ion PC1 4.49 40.8 40.8 PC2 2.82 25.6 66.4 PC3 1.13 10.3 76.7 PC4 0.86 7.8 84.5 PC5 0.72 6.5 91.1 FIGURE7 | Loadings ( ec o s) o he en i onmen al a iables on he space o he i s (PC1) and second (PC2) componen s o he PCA o en i onmen al a iables de e mined o neus on sampling s a ions and p ojec ions (do s) o median alues o niche size (SEAb) and 95% c edible in e als (CI) o he di e en FGs (indica ed by numbe s 1–4). DCM: dep h o he chlo ophyll maximum (m), δ15N40–200: na u al abundance o mic oplank on ni ogen iso opes (‰), PP: annual mean p ima y p oduc ion (mg C m−2 d−1), MLD: mixed laye dep h (m), C40–200: mic oplank on (40–200μm) ca bon biomass (mg C m−3), KT: di usi i y due o u bulence (m2 s−1), Chlas: su ace chlo ophyll-a (mgm−3), N2: mean squa ed B un -Väisälä equency (s−2), Chlai: pho ic-zone in eg a ed chlo ophyll-a (mg m−2), T icho: abundance o T ichodesmium (cells ml−1), MDU: mean mon hly a mosphe ic dus deposi ion (gm−2 mon h−1).