Dissolved inorganic nutrients in the western Mediterranean Sea (2004–2017)

Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 © Au ho (s) 2020. This wo k is dis ibu ed unde he C ea i e Commons A ibu ion 4.0 License. Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea (2004–2017) Malek Belgacem1,2, Jacopo Chiggia o1, Mi eno Bo ghini1, B uno Pa oni2, Gab iella Ce a i3, F ancesco Ac i1, S e ano Cozzi4, Albe o Ribo i5, Ma a Ál a ez6, Si K. Lau se 7, and Ka in Sch oede 1 1ISMAR-CNR, A senale Tesa 104, Cas ello 2737/F, 30122 Venice, I aly 2Dipa imen o di Scienze Ambien ali In o ma ica e S a is ica, Uni e si à Ca’ Fosca i Venezia, Campus Scien i ico, Mes e, 30170 Venice, I aly 3Depa men o Sus ainabili y, S Te esa Ma ine En i onmen Resea ch Cen e, ENEA, 19032 Pozzuolo di Le ici (SP), I aly 4ISMAR-CNR, A ea Science Pa k, Baso izza, 34149 T ies e, I aly 5IAS-CNR, Loc. Sa Ma dini snc, To eg ande, 9170 O is ano, I aly 6Ins i u o Español de Oceanog a ía, IEO, A Co uña, Spain 7NORCE No wegian Resea ch Cen e, Bje knes Cen e o Clima e Resea ch, Be gen, No way Co espondence: Jacopo Chiggia o (jacopo.chiggia o@isma .cn .i ) Recei ed: 5 Augus 2019 – Discussion s a ed: 3 Sep embe 2019 Re ised: 16 June 2020 – Accep ed: 19 July 2020 – Published: 4 Sep embe 2020 Abs ac . Long- e m ime se ies a e a undamen al p e equisi e o unde s anding and de ec ing clima e shi s and ends. Unde s anding he complex in e play o changing ocean a iables and he biological implica ion o ma ine ecosys ems equi es ex ensi e da a collec ion o moni o ing, hypo hesis es ing, and alida ion o modelling p oduc s. In ma ginal seas, such as he Medi e anean Sea, he e a e s ill moni o ing gaps, bo h in ime and in space. To con ibu e o illing hese gaps, an ex ensi e da ase o dissol ed ino ganic nu ien obse a ions (ni a e, phospha e, and silica e) was collec ed be ween 2004 and 2017 in he wes e n Medi e anean Sea and subjec ed o igo ous quali y con ol echniques o p o ide o he scien i ic communi y a publicly a ailable, long- e m, quali y-con olled, in e nally consis en biogeochemical da a p oduc . The da a p oduc includes 870 s a ions o dissol ed ino ganic nu ien s, including empe a u e and salini y, sampled du ing 24 c uises. De ails o he quali y con ol (p ima y and seconda y quali y con ol) applied a e epo ed. The da a a e a ailable in PANGAEA (h ps://doi.o g/10.1594/PANGAEA.904172, Belgacem e al., 2019). Da a co e age and pa ame e measu ed A ailable a : h ps://doi.o g/10.1594/PANGAEA.904172 Co e age: 44◦N–35◦S, 6◦W–14◦E Loca ion name: wes e n Medi e anean Sea Da e/ ime s a : May 2004 Da e/ ime end: No embe 2017 1 In oduc ion Dissol ed ino ganic nu ien s play a c ucial ole in ma ine ecosys em unc ioning. They se e as egula o s o ocean biological p oduc i i y and a e ace elemen s o biogeo- chemical cycling as well as o na u al and an h opogenic sou ces and anspo p ocesses (Bé houx, 1989; Bé houx e al., 1992). They a e also non-conse a i e ace s, since hei dis ibu ion a ies acco ding o bo h biological (such as p ima y p oduc ion and espi a ion) and physical (such as con ec ion, ad ec ion, mixing, and di usion) p ocesses. Ve y schema ically, ino ganic nu ien s a e con inuously con- Published by Cope nicus Publica ions. 1986 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea sumed by phy oplank on (due o p ima y p oduc ion) a he sea su ace and egene a ed in he mesopelagic laye by bac- e ia and animals (due o espi a ion). Mo eo e , he sinking o o ganic ma e and i s decomposi ion inc ease he nu ien concen a ions in he in e media e- and deep-wa e masses o e ime. To iden i y he limi ing ac o s o biological p o- duc ion in he oceans, we need o unde s and he unde ly- ing chemical cons ain s and especially he mac o- and mi- c onu ien spa ial and empo al a ia ions. Dissol ed ino - ganic nu ien s may be used as ace s o wa e masses like salini y and empe a u e o assess mixing p ocesses and o unde s and he biogeochemical ci cums ances o hei o ma- ion egions. Unde s anding he complex in e play o chang- ing ocean a iables and he biological implica ion o ma ine ecosys ems is a di icul ask and equi es no only modelling bu also ex ensi e da a collec ion o moni o ing, hypo hesis es ing, and alida ion. Moni o ing gaps s ill emain in bo h ime and space, especially o ma ginal seas such as he A c- ic Ocean o he Medi e anean Sea. The Medi e anean Sea has been iden i ied as a egion sig- ni ican ly a ec ed by ongoing clima ic changes, like wa m- ing and a dec ease in p ecipi a ion (Gio gi, 2006). In addi- ion, i is a egion pa icula ly aluable o clima e change esea ch because i beha es like a minia u e ocean (Bé houx e al., 1999) wi h a well-de ined o e u ning ci cula ion cha - ac e ized by spa ial and empo al scales ha a e much sho e han o he global ocean, wi h a u no e o only se e al decades. Being an in e con inen al sea and subjec ed o mo e e es ial nu ien inpu s ( i e uno , subma ine g oundwa- e discha ge) and a mosphe ic deposi ion, he Medi e anean Sea has a ni a e o phospha e N :P a io ha is anomalously high compa ed o he “classical” wo ld oceans Red ield a- io, indica ing a gene al P-limi a ion egime, which becomes s onge along a wes - o-eas g adien . The Medi e anean Sea is he e o e a po en ial model o s udying global pa - e ns ha will be expe ienced in he nex decades wo ldwide, no only ega ding ocean ci cula ion bu also he ma ine bio a (Lejeusne e al., 2010). Se e al en i onmen al a iables can ac as s esso s o ma ine ecosys ems, by which clima ically d i en ecosys em dis u bances a e gene a ed (Boyd, 2011). These changes a ec , among o he hings, he dis ibu ion o biogeochemical elemen s (including ino ganic nu ien s) and he unc ioning o he biological pump and CO2 egula ion. Wi hin his con ex , he aim o his pape is o compile an ex ensi e da ase o dissol ed ino ganic nu ien obse - a ions (ni a e, phospha e, and silica e) collec ed be ween 2004 and 2017 in he wes e n Medi e anean Sea (WMED); o desc ibe he quali y con ol echniques; and o p o ide he scien i ic communi y wi h a publicly a ailable, long- e m, quali y con olled, and in e nally consis en biogeo- chemical da a p oduc , con ibu ing o p e iously published Medi e anean Sea da ase s like MEDAR/MEDATLAS ( ime pe iod 1908–1999; Fichau e al., 2003) and he Medi e - anean Sea – Eu ophica ion and Ocean Acidi ica ion ag- g ega ed da ase s 2018 ( ime pe iod 1911–2017) p o ided Figu e 1. Map o he wes e n Medi e anean Sea showing he bio- geochemical s a ions (in blue) and he i e e e ence c uise s a ions (in ed). by EMODne Chemis y (Gio ge i al., 2018) a ailable a h ps://www.seada ane .o g/P oduc s/Agg ega ed-da ase s (las access: Ma ch 2020). 2 Dissol ed ino ganic nu ien da a collec ion 2.1 The CNR dissol ed ino ganic nu ien da a in he WMED Long- e m ime se ies, such as he OceanSITES global ime se ies (h p://www.oceansi es.o g, las access: May 2019), a e a undamen al p e equisi e o unde s anding and de ec - ing clima e shi s and ends. Howe e , biogeochemical ime se ies a e s ill limi ed o he no he n pa o he wes e n Medi e anean Sea (MOOSE ne wo k; Coppola e al., 2019). Ye , ino ganic nu ien s in he Medi e anean Sea ha e e- cei ed mo e a en ion in ecen yea s, and a ious da ase s ha e been compiled o unde s and hei unique cha ac e is- ics such as he one buil by he PERSEUS p ojec conso - ium (Policy-o ien ed ma ine en i onmen al esea ch in he sou he n Eu opean seas”, EU FP7 p ojec , g an ag eemen (GA) no. 287600), which included 100 c uises collec ed du - ing he p ojec ’s li e ime, in addi ion o hose om o he p ojec s like SESAME (EU FP7 p ojec , GA no. GOCE- 036949) and da a p oduc s such as MEDAR/MEDATLAS. In addi ion o ha , he da a assembly sys em EMODne Chemis y, a leading in as uc u e, is suppo ed by he pan- Eu opean Di ec o a e-Gene al o Ma i ime A ai s and Fish- e ies (DG MARE; Ma ín Míguez e al., 2019; Tin o é e al., 2019). The da ase p esen ed he e consis s o 24 oceanog aphic c uises (Fig. 1, Table 1a and b) conduc ed in he WMED on boa d esea ch essels un by he I alian Na ional Resea ch Council (CNR) and he NATO Science and Technology O - ganiza ion (STO) Cen e o Ma i ime Resea ch and Expe i- men a ion (CMRE). All c uises we e me ged in o a uni ied da ase wi h 870 nu ien s a ions and ∼9666 da a poin s o e a pe iod o 13 yea s (2004–2017). The o e all spa ial Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 1987 Table 1. (a) C uise summa y able and pa ame e s lis ed wi h numbe o s a ions and samples. C uises we e iden i ied wi h an ID numbe and expedi ion code (EXPOCODE o o ma AABBYYYYMMDD – AA: coun y code; BB: ship code; YYYY: yea ; MM: mon h; DD: day indica i e o c uise s a ing day). (b) Da a sou ces and links o he epo s (las access: June 2020). (a) C uise Common name EXPOCODE Resea ch essel Da e s a and end S a ions Samples Samples Samples Maximum bo om Chie scien is (s) ID no. (RV) ni a e phospha e silica e dep h (m) 1 TRENDS2004 and MEDGOOS8leg2 48UR20040526 U ania 26 May–14 Jun 2004 36 255 253 255 3499 Mi eno Bo ghini 2 MEDGOOS9 48UR20041006 U ania 6–25 Oc 2004 68 627 626 627 3610 Mi eno Bo ghini 3 MEDOCC05 and MFSTEP2 48UR20050412 U ania 12 Ap –16 May 2005 68 828 828 828 3598 Mi eno Bo ghini 4 MEDGOOS10 48UR20050529 U ania 29 May–10 Jun 2005 36 577 577 577 3505 Angelo Pe illi 5 MEDGOOS11 48UR20051116 U ania 16 No –3 Dec 2005 14 143 143 143 2810 Angelo Pe illi, Mi eno Bo ghini, Massimiliano Dibi e o 6 MEDOCC06 48UR20060608 U ania 8 Jun–3 Jul 2006 66 787 785 787 2881 Mi eno Bo ghini 11aSIRENA06 06A420060720 Alliance 20 Jul–6 Aug 2006 35 208 208 209 1854 Je Haun 8 MEDGOOS13 and MEDBIO06 48UR20060928 U ania 28 Sep–8 No 2006 37 519 520 520 2862 Albe o Ribo i 9 MEDOCC07 48UR20071005 U ania 5–29 Oc 2007 71 977 977 979 3497 Angelo Pe illi, Mi eno Bo ghini, Albe o Ribo i 10 SESAMEIT4 48UR20080318 U ania 18 Ma –7 Ap 2008 11 164 164 164 2882 Chia a San inelli 11 SESAMEIT5 48UR20080905 U ania 5–16 Sep 2008 12 74 74 74 536 S e ania Spa nocchia, Gian Pie o Gaspa ini, Mi eno Bo ghini 12 MEDCO08 48UR20081103 U ania 3–24 No 2008 24 342 350 348 2880 Albe o Ribo i 13 TYRRMOUNTS 48UR20090508 U ania 8 May–3 Jun 2009 41 430 441 440 2559 Gian Pie o Gaspa ini 14 BIOFUN010 48UR20100430 U ania 30 Ap –17 May 2010 26 405 405 405 3540 Elena Mannini, S e ano Aliani 15 VENUS1 48UR20100731 U ania 31 Jul–25 Aug 2010 32 431 432 428 3544 Gian Pie o Gaspa ini, Mi eno Bo ghini 16 BONSIC2010 48UR20101123 U ania 23 No –9 Dec 2010 18 144 143 143 3540 Albe o Ribo i 17 EUROFLEET11 48UR20110421 U ania 21 Ap –8 May 2011 28 277 275 277 3540 Gian Pie o Gaspa ini, Mi eno Bo ghini 18 BONIFACIO2011 48UR20111109 U ania 9–23 No 2011 13 180 180 181 3541 Albe o Ribo i, Gina La Spada, Mi eno Bo ghini 19 TOSCA2011 48MG20111210 Ma ia G azia 10–20 Dec 2011 21 310 310 309 2728 Mi eno Bo ghini 20 ICHNUSSA12 48UR20120111 U ania 11–27 Jan 2012 21 353 352 323 3551 Albe o Ribo i 21 EUROFLEET2012 48UR20121108 U ania 8–26 No 2012 53 429 434 434 2633 Mi eno Bo ghini 22 ICHNUSSA13 48UR20131015 U ania 15–29 Oc 2013 37 405 404 405 3540 Albe o Ribo i 23 OCEANCERTAIN15 48QL20150804 Mine a Uno 4–29 Aug 2015 71 531 531 531 3513 Jacopo Chiggia o 24 ICHNUSSA17 and INFRAOCE17 48QL20171023 Mine a Uno 23 Oc –28 No 2017 31 251 254 254 3536 Albe o Ribo i, S e ania Spa nocchia, Mi eno Bo ghini h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 1988 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Table 1. Con inued. (b) C uise Expedi ion o iginal name PI(s) o chie scien is (s) Speci ic linka(las access: June 2020) ID no. 1 TRENDS2004 and MEDGOOS8leg2 Mi eno Bo ghini h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=5821 h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=4935 2 MEDGOOS9 Mi eno Bo ghini Repo submission in p og ess h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=5823 h ps://doi.o g/10.17882/70340 3 MEDOCC05 and MFSTEP2 Mi eno Bo ghini h p:// ice ca.isma .cn .i /CRUISE_REPORTS/2005/URANIA_MEDOCC05.pd h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=4936 4 MEDGOOS10 Angelo Pe illi h p://www.sea o ecas .cn .i /i /obse a ion_i .h m h ps://doi.o g/10.17882/70340 5 MEDGOOS11 Angelo Pe illi, Mi eno Bo ghini, Mas- similiano Dibi e o h p:// ice ca.isma .cn .i /CRUISE_REPORTS/2005/URANIA_MEDGOOS11_05_REP.pd h ps://doi.o g/10.17882/70340 6 MEDOCC06 Mi eno Bo ghini h p://www.sea o ecas .cn .i / epo s/Medocc06CR.pd h ps://seada a.bsh.de/Cgi-cs / e ie e_sdn2/ iewRepo .pl?cs e =20106010 7 SIRENA06 Je Haun Repo submission in p og ess 8 MEDGOOS13 and MEDBIO06 Albe o Ribo i h p://www.sea o ecas .cn .i / epo s/Mebio06-Medg13_CR.pd h ps://doi.o g/10.17882/70340 9 MEDOCC07 Angelo Pe illi, Mi eno Bo ghini, Al- be o Ribo i h p://www.sea o ecas .cn .i / epo s/Medocc07-MedCo07_Rapp.pd h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=5146 10 SESAMEIT4 Chia a San inelli h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=5148 h ps://emodne -chemis y.ma is.nl/sea ch/de ails.php?s ep=0012004~0022017~0153~057l04001~058 din,n a,phos, slca~00445~0056~00617~00734~0541&coun =3592&page=1000&so =0&heade =no 11 SESAMEIT5 S e ania Spa nocchia, Gian Pie o Gas- pa ini, Mi eno Bo ghini h ps://is ama .ocean.o g.il/pe seus_da a/C uiseIn o.aspx?c iuseid=5147 12 MEDCO08 Albe o Ribo i h p://www.sea o ecas .cn .i / epo s/MedCO08_Rapp.pd 13 TYRRMOUNTS Gian Pie o Gaspa ini Repo submission in p og ess 14 BIOFUN010 Elena Mannini, S e ano Aliani h p://www.isma .cn .i /p oduc s/ epo s-campagne/2010-2019 15 VENUS1 Gian Pie o Gaspa ini, Mi eno Bo ghini Repo submission in p og ess 16 BONSIC2010 Albe o Ribo i h p://www.sea o ecas .cn .i / epo s/Boni acio2010Sic_Rapp.pd 17 EUROFLEET11 Gian Pie o Gaspa ini, Mi eno Bo ghini Repo submission in p og ess 18 BONIFACIO2011 Albe o Ribo i, Gina La Spada, Mi eno Bo ghini h p://www.sea o ecas .cn .i / epo s/Boni acio2011_Rapp.pd 19 TOSCA2011 Mi eno Bo ghini Repo submission in p og ess 20 ICHNUSSA12 Albe o Ribo i h p://www.sea o ecas .cn .i / epo s/Ichnussa2012_Rapp.pd 21 EUROFLEET2012 Mi eno Bo ghini Repo submission in p og ess 22 ICHNUSSA13 Albe o Ribo i h p://www.sea o ecas .cn .i / epo s/Ichnussa2013_Rapp.pd 23 OCEANCERTAIN15 Jacopo Chiggia o h ps://doi.o g/10.1594/PANGAEA.911046 24 ICHNUSSA17 and INFRAOCE17 Albe o Ribo i, S e ania Spa nocchia, Mi eno Bo ghini Repo submission in p og ess aThe speci ic links a e subjec u7 o upda es. Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 1989 dis ibu ion o he s a ions co e s he whole WMED, bu he ac ual dis ibu ion s ongly a ies depending on he speci ic c uise and mos o he da a a e collec ed along sec ions. A all s a ions, p essu e, salini y, and empe a u e we e measu ed wi h a CTD- ose e sys em consis ing o a CTD SBE 911plus and a Gene al Oceanics ose e wi h 24 Niskin bo les o 12L capaci y. Tempe a u e measu emen s we e pe o med wi h he SBE 3F he mome e wi h a esolu ion o 10−3◦C; con- duc i i y measu emen s we e pe o med wi h an SBE 4 sen- so wi h a esolu ion o 3×10−4Sm−1. The p obes we e cal- ib a ed be o e and a e each c uise. Du ing all CNR c uises, edundan senso s we e used o bo h empe a u e and salin- i y measu emen s. Seawa e samples o dissol ed ino ganic nu ien mea- su emen s we e collec ed du ing he CTD upcas a s anda d dep hs (wi h sligh modi ica ions acco ding o he dep h a which he deep chlo ophyll maximum was de ec ed). The s anda d dep hs a e usually 5, 25, 50, 75, 100, 200, 300, 400, 500, 750, 1000, 1250, 1500, 1750, 2000, 2250, 2500, 2750, and 3000 m. No il a ion was employed; nu ien samples we e immedia ely s o ed a −20 ◦C. No e ha sample s o - age and eezing du a ion a ied g ea ly om one c uise o ano he (Table 3 shows c uises whe e his exceeded 1 yea ). 2.2 Analy ical me hods o ino ganic nu ien s Fo all c uises, nu ien de e mina ion (ni a e, o hosilica e, and o hophospha e) was ca ied ou ollowing s anda d col- o ime ic me hods o seawa e analysis, de ined by G assho e al. (1999) and Hansen and Ko ole (1999). Fo ino - ganic phospha e, he me hod is based on he eac ion o he ions wi h an acidi ied molybda e eagen o yield a phospho- molybda e he e opoly acid, which is hen educed o a blue- colou ed compound (abso bance measu ed a 880 nm). Ino - ganic ni a e is educed (wi h cadmium g anules) o ni i e ha eac s wi h an a oma ic amine, leading o he inal o - ma ion o he azo dye (measu ed a 550nm). Then, he ni i e ha is sepa a ely de e mined mus be sub ac ed om he o- al amoun measu ed o ge he ni a e concen a ion only. The de e mina ion o dissol ed silicon is based on he o - ma ion o a yellow silicomolybdic acid educed wi h asco - bic acid o a blue-colou ed complex (measu ed a 820nm). Nu ien analysis was pe o med in h ee labo a o ies. F om 2004 o 2013, all c uises nu ien s we e analysed by ENEA, while o hose o 2015 (c uise no. 23) and 2017 (c uise no. 24), nu ien concen a ions we e analysed by ISMAR-CNR (ISMAR is he Ins i u e o Ma ine Sciences). Re e ing o Table S1 in he Supplemen , ou di e en mod- els o au oanalyse we e used. Measu emen s om he au o- analyse we e epo ed in mic omoles pe li e. Ino ganic nu- ien concen a ions we e con e ed o he s anda d uni o mic omoles pe kilog am, using sample salini y om CTD and a mean labo a o y analy ical empe a u e o 20 ◦C. Da a om nu ien analysis we e hen me ged wi h ancilla y CTD bo le da a. Table 2. C uise summa y able o he e e ence c uise collec ion used in he seconda y quali y con ol, collec ed om 2001 o 2016. Common name EXPOCODE Da e s a and end S a ions Ni a e Phospha e Silica e Sou ce Nu ien PI(s) Chie scien is (s) sample sample sample M51/2 06MT20011018 18 Oc –11 No 2001 6 79 79 82 GLODAP 2 Be nd Schneide Wol gang Roe he TRANSMED_LEGII 48UR20070528 28 May–12 Jun 2007 4 78 77 78 CARIMED (no ye a ailable) S e ano Cozzi, Vale ia Ibello Mau izo Azza o M84/3 06MT20110405 5–28 Ap 2011 20 339 343 – GLODAP 2 Giuseppe Ci i a ese Tos e Tanhua HOTMIX 29AH20140426 26 Ap –31 May 2014 18 144 140 144 CARIMED (no ye a ailable) Xosé An on Ál a ez-Salgado Ja ie A is egui TAlP o-2016 29AJ20160818 18–28 Aug 2016 42 293 293 293 MedSHIP p og amme Lau en Coppola Loïc Jullion, Ka in Sch oede h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 1990 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Figu e 2. O e iew o he e e ence c uise spa ial co e age and e ical dis ibu ions o he ino ganic nu ien s. (a) Geog aphical dis ibu ion map; (b) e ical p o iles o ni a e (µmolkg−1); (c) e ical p o iles o phospha e (µmolkg−1); (d) e ical p o iles o silica e (µmolkg−1). 2.3 Re e ence ino ganic nu ien da a In addi ion o he da a collec ed du ing he abo e-men ioned c uises and in o de o pe o m he seconda y quali y con ol (desc ibed below), we iden i ied i e e e ence c uises (Ta- ble 2), based on hei spa ial and empo al dis ibu ion and he eliabili y o he measu emen s (see Fig. 2 and Table S3 and Fig. S1 in he Supplemen ). C uises 06MT20110405 and 06MT20011018 a e he only wo Medi e anean c uises in- cluded in he publicly a ailable Global Ocean Da a Anal- ysis P ojec e sion 2 (GLODAP 2; Olsen e al., 2016). These c uises, conduc ed on boa d he R/V Me eo , p o- ide a eliable e e ence because nu ien analysis s ic ly ollowed he ecommenda ion o he Wo ld Ocean Ci cula- ion Expe imen (WOCE) and he GO-SHIP p o ocols (Hy- des e al., 2010; Tanhua e al., 2013). C uises 29AH20140426 and 48UR20070528 a e o be included in he CARIMED da a p oduc (pe sonal communica ion by Ma a Ál a ez, in p epa a ion bu no ye a ailable) and ha e unde gone igo ous quali y con ol ollowing GLODAP ou ines. Fi- nally, 29AJ20160818 was ca ied ou in he amewo k o he MedSHIP p og amme (Sch oede e al., 2015), and i s da a a e a ailable a h ps://doi.o g/10.1594/PANGAEA.902293 (Tanhua, 2019). 3 Quali y assu ance and quali y con ol me hods Combining ino ganic nu ien da a om di e en sou ces, collec ed by di e en ope a o s, s o ed o di e en amoun s o ime, and analysed by mul iple labo a o ies is no a s aigh o wa d ask. This is widely ecognized in he biogeo- chemical oceanog aphic communi y. Since he 1990s, se - e al s udies and p og ammes (e.g. Wo ld Ocean Da abase, Wo ld Ocean A las, WOCE) ha e been de o ed o acili a ing he exchange o oceanog aphic da a and de eloping quali y con ol p ocedu es o compile da abases by he es ima ion o sys ema ic e o s (Gou e ski and Jancke, 2000) o inc ease in e -compa abili y, gene a e consis en da ase s, and accu- a ely obse e long- e m change. An example o a i s quali y con ol p ocedu e is he use o e e ence ma e ials ha a e a ailable o salini y (IAPSO – In e na ional Associa ion o Physical Sciences o he Ocean, salini y s anda d by OSIL – Ocean Scien- i ic In e na ional Limi ed) and empe a u e (SPRT – s an- da d pla inum esis ance he mome e ). As o he ino ganic ca bon, o al alkalini y (Dickson e al., 2003), and ino ganic nu ien s (Aoyama e al., 2016), ce i ied e e ence ma e i- als (CRMs) ha e been ecen ly made applicable o oceano- g aphic c uises. Howe e , since CRMs a e no always a ail- able o used o biogeochemical oceanog aphic da a, Lau- se and Tanhua (2015) de eloped a seconda y quali y con- ol ool o iden i y biases in deep da a. The me hod sugges s Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 1991 adjus men s ha educe c uise- o-c uise biases, inc ease ac- cu acy, and allow o he in e -compa ison be ween da a om a ious sou ces. This app oach, based on a c osso e and in- e sion me hod (Gou e ski and Jancke, 2000; Johnson e al., 2001), was used o gene a e he CARbon dioxide IN he A - lan ic Ocean (CARINA; see Hoppema e al., 2009), GLO- DAP 2.2019 (Olsen e al., 2019), and PACIFICA (Suzuki e al., 2013) da a p oduc s. 3.1 P ima y quali y con ol Each indi idual c uise was i s subjec ed o a p ima y qual- i y con ol ( i s QC) ha included a check o appa en and ex eme ou lie s in CTD salini y, ni a e, phospha e, and sili- ca e. Each pa ame e included a quali y con ol lag, ollow- ing s anda d WOCE lags (Table 3). The su ace, in e medi- a e, and deep laye we e e alua ed sepa a ely because nu i- en obse a ions e ol e di e en ly in each laye . The coe - icien o a ia ion (CV, de ined as s anda d de ia ion o e mean) was compu ed o each dep h laye . Coe icien s o a ia ion in he su ace laye (0–250db) we e high (ni a e CV =1.16, phospha e CV =1.005, silica e CV =0.75) due o ai –sea in e ac ion (Muniz e al., 2001) occu ing in his laye ende ing i di icul o lag. These in luences a e o e- duced impo ance in he in e media e laye (250–1000db; ni- a e CV =0.23, phospha e CV =0.31, silica e CV =0.24) and he deep laye (>1000 db; ni a e CV =0.15, phospha e CV =0.22, silica e CV =0.14), dec easing he o al a i- ance. Flags in he uppe and in e media e laye we e hus se based on ou lie s wi hin p essu e anges de ined acco ding o s anda d p essu es (0–10, 10–30, 30–60, 60–80, 80–160, 160–260, 260–360, 360–460, 460–560, 560–1000 db). Below 1000 db, lagging included an inspec ion o ni a e o phospha e (N :P) and ni a e o silica e (N :Si) a ios. The median and median absolu e de ia ion (MAD) we e com- pu ed by classes o p essu e: we conside ed ou lie s o be any a ypical obse a ion and any alue ha depa s om he median by mo e han h ee MADs in he di e en p essu e anges o each c uise. An o e iew o he nu ien dis ibu ion is p o ided wi h sca e plo s, showing also he lagged measu emen s (Fig. 3). Each measu emen was lagged as 2 (“Accep - able/measu ed”) o lagged as 3 (“Ques ionable”); 4.1% o ni a e da a, 3.37 % o phospha e da a, 3.16 % o silica e da a, and 0.07 % o CTD salini y da a we e conside ed ou lie s and lagged as 3. As highligh ed by Tanhua e al. (2010), he p i- ma y QC can be subjec i e depending on he expe ise o he pe son lagging he da a, hus lagging could b ing in some unce ain ies. In o de o ha e a i s assessmen o he p ecision o each c uise se o measu emen s, he s anda d de ia ion o obse - a ions deepe han 1000 db was calcula ed along wi h a - e ages and s anda d de ia ions o each c uise and by sub- egion o ha e an o e iew o nu ien con en a iabili y in he deep laye and o he obse a ions’ spa ial sp ead o indi idual c uises (Table 4). Following he subdi ision o Manca e al. (2004), he WMED has been di ided in o sub- egions (Fig. S2, Table S2) acco ding o he gene al ci cu- la ion pa e ns (de ails in Manca e al., 2004). Table 4 dis- plays he compa ison o s anda d de ia ion o deep mea- su emen s o each c uise and wi hin sub egions. The o e - all s anda d de ia ion be ween c uises in he deep laye a - ied be ween 0.51 and 1.41 µmolkg−1 o ni a e, be ween 0.1 and 1.64 µmolkg−1 o silica e, and be ween 0.025 and 0.078 µmolkg−1 o phospha e. Regional s anda d de ia- ion o ni a e measu emen s below 1000 db a ied be ween 0.08 µmolkg−1in he Gul o Lion (DF2) wi h c uise no. 9 and 1.6µmolkg−1in he Balea ic Sea (DS2) obse a ions o c uise no. 14. The lowes phospha e egional s anda d de ia- ion was 0.01 µmolkg−1 ound in he obse a ions o c uise no. 9 in he Gul o Lion (DF2), c uise no. 10 in he Balea ic Sea (DS2) and Alge ian Wes (DS3), c uise no. 14 and c uise no. 15 in he Ty henian Sou h (DT3), and c uise no. 18 in Alge o-P o ençal (DF1) and he Sa dinia Channel (DI1), while he highes s anda d de ia ion was 0.1µmolkg−1in he obse a ions o c uise no. 12 in he Alge ian Wes (DS3). As o silica e, he lowes s anda d de ia ion was 0.02µmolkg−1 obse ed in c uise no. 9 measu emen s o he Gul o Lion sub egion (DF2) and he highes deep s anda d de ia ion was obse ed in c uise no. 6 in all i s sub egions oge he wi h c uise no. 5 measu emen s in he Ty henian No h (DT1) wi h a 1.83 µmolkg−1s anda d de ia ion. C uises no. 3, no. 6, and no. 9 had he la ges spa ial ex en- sion (see igh side o Fig. 9) wi h a high numbe o samples o e mo e han se en sub egions (Table 4); he geog aphi- cal a iabili y in he dis ibu ion o dissol ed ino ganic nu- ien s esul s hus in he la ges s anda d de ia ions. Con- e sely, c uises wi h smalle spa ial co e ages ha e lowe s anda d de ia ions. The e o e, a ela i ely small spa ial co - e age and high s anda d de ia ion is conside ed as indica i e o da a wi h low p ecision (Olsen e al., 2016). This applies o c uises no. 1, no. 5, and no. 16. Despi e he small spa ial co - e age, samples o ni a e and phospha e o c uise no. 5 ha e an o e all s anda d de ia ion o 1.35 and 0.07 µmolkg−1, espec i ely, a high s anda d de ia ion was also poin ed ou in he egional s anda d de ia ion o deep measu e- men s in he Ty henian No h (DT1) and Sou h (DT3). C uise no. 1, wi h ew s a ions in he Ty henian No h (DT1) and Sou h (DT3) sub egions and 21 samples below 1000 db, has an o e all s anda d de ia ion o 1.25 µmolkg−1 o ni a e, 0.06µmolkg−1 o phospha e, and 1.64 µmolkg−1 o silica e. The egional s anda d de ia ion was ela i ely high o ni a e (0.51–1.32 µmolkg−1), phospha e (0.02– 0.065 µmolkg−1), and silica e (0.53–1.83 µmolkg−1). A compa ison wi h he de ia ions om e.g. c uise no. 2, ca - ied ou in he same yea and e.g. c uise no. 17 (wi h a sim- ila c uise ack) con i ms he lowe p ecision o he da a o c uise no. 1. Simila conside a ions apply o he quali y o ni a e samples (0.87–1.02µmolkg−1) and silica e (0.87– 0.9 µmolkg−1) om c uise no. 16, co e ing a small a ea in h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 1992 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Table 3. WOCE lags used in he o iginal da a p oduc and in he adjus ed p oduc . WOCE lag alue In e p e a ion in o iginal da ase In e p e a ion in adjus ed p oduc 2 Accep able/measu ed Adjus ed and accep able 3 Ques ionable/no used Adjus ed and ecommended ques ionable 9 No measu ed/no da a Figu e 3. Sca e plo s o (a) phospha e s. ni a e (µmolkg−1) and (b) silica e s. ni a e (µmolkg−1). Da a ha ha e been lagged as ques ionable ( lag 3) a e in ed; he colou ba indica es he p essu e (dba ). The black lines ep esen he bes linea i be ween he wo pa ame e s, and he co esponding equa ions and 2 alues a e shown in each plo . A e age esul ing N :P a io is 20.87; a e age esul ing N:Si a io is 1.05 (whole dep h). he Ty henian No h (DT1) and Sou h (DT3), compa ed o c uise no. 17, ca ied ou in he same egions ( igh side o Fig. 9 and Table 4). Deep silica e measu emen s o c uise no. 6 ha e wice he o e all s anda d de ia ion o silica e da a o c uise no. 8 om he same yea . Adding o ha , in he se en sub egions, he egional s anda d de ia ion o deep silica e obse a ions was he highes , be ween 1.04 and 2 µmolkg−1, which was el- a i ely high compa ed o he su ounding c uises ha ha e obse a ions in he same sub egions. This is again sugges- i e o limi ed p ecision. On he o he hand, ying o ex- plain he sou ce o ela i ely high s anda d de ia ions in spe- ci ic c uises is no always s aigh o wa d, as hey could s em om a a ie y o sou ces, sampling, and conse a ion and analysis me hods. The bo om wa e in he WMED exhibi s a high nu ien con en below 1000 db (Table 4), due o he longe esidence ime. Di iding he WMED in o sub egions has e ec i ely emo ed he na u al spa ial change in nu i- en s, making he in e p e a ion o he s anda d de ia ion a ma e o he p ecision o he measu emen s only. In Table 4, deep a e ages by sub egions show ha o e all nu ien concen a ion luc ua ed a a ound 7.4± 0.9 µmolkg−1 o ni a e, 0.3±0.06 µmolkg−1 o phos- pha e, and 7.7±0.8 µmolkg−1 o silica e; simila indings we e epo ed by Manca e al. (2004). Compa ing c uise a - e ages in each egion enabled he iden i ica ion o “suspec ” c uises. C uise no. 24 has he lowes deep a e age in ni- a e in he Alge o-P o ençal (DF1) and Ty henian No h (DT1) sub egions and in he Sa dinia Channel (DI1). Sili- ca e o c uises no. 24 and no. 16 was e y low compa ed o he o e all egional a e age in he Ligu o-P o ençal (DF3) and Ty henian Sou h (DT3) sub egions. The deep a e age o phospha e did no show any ou lie c uises in all sub e- gions. Di e en easons could explain he low p ecision in he samples; eezing is one. Al hough i is a alid p ese - a ion me hod (Do e a al., 1996), he e o is highe when samples a e no analysed immedia ely (Segu a-Nogue a e al., 2011), so he s o age ime could be in luen ial. 3.2 Seconda y quali y con ol – he c osso e analysis The me hod used o pe o m he seconda y QC on he WMED dissol ed ino ganic nu ien da ase makes use o he quali y-con olled e e ence da a and he c osso e analysis oolbox de eloped by Tanhua (2010) and Lau se and Tan- hua (2015). The compu a ional app oach is based on com- pa ing he c uise da ase o a high-quali y e e ence da ase o quan i y biases, desc ibed in de ail in Tanhua e al. (2010). He e, we summa ize he echnique wi h emphasis on in- o ganic nu ien s. The i s s ep consis ed o selec ing e - e ence da a, as desc ibed in Sec . 2.3. The second s ep is he c osso e analysis ha was ca ied ou using a MAT- Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 1993 Table 4. A e age and s anda d de ia ions o ni a e, phospha e, and silica e measu emen s by c uise and o each egion wi h numbe o samples deepe han 1000 db included in he second QC. A e age s o age ime is he minimum s o age ime de ined as ime di e ence be ween he c uise ending day and he i s day o he labo a o y analysis. C uise EXPOCODE/ Regional SD Regional a g SD Regional SD No. samples A g s o age ID egion a g ni a e ni a e phospha e phospha e a g silica e silica e (in days) (µmolkg−1) (µmolkg−1) (µmolkg−1) (µmolkg−1) (µmolkg−1) (µmolkg−1) 1 48UR20040526/ 1.25 0.062 1.64 21 131 DT1 – Ty henian No h 6.07 1.32 0.26 0.065 6.92 1.83 16 DT3 – Ty henian Sou h 7.03 0.51 0.31 0.02 7.66 0.53 5 2 48UR20041006/ 0.59 0.029 0.81 21 251 DT1 – Ty henian No h 7.68 0.53 0.41 0.031 8.74 0.75 15 DT3 – Ty henian Sou h 8.17 0.60 0.41 0.025 9.31 0.87 6 3 48UR20050412/ 1.15 0.050 1.41 233 135 DF2 – Gul o Lion 7.89 0.98 0.40 0.044 8.17 1.065 24 DF3 – Ligu o-P o ençal 7.45 1.08 0.41 0.05 7.72 1.10 66 DS2 – Balea ic Sea 7.44 1.14 0.40 0.039 7.68 1.47 21 DF1 – Alge o-P o ençal 7.87 1.16 0.41 0.043 8.88 1.96 42 DS3 – Alge ian Wes 7.7 0.816 0.39 0.048 8.14 0.941 23 DT1 – Ty henian No h 6.57 1.065 0.36 0.047 7.41 1.15 21 DT3 – Ty henian Sou h 6.52 1.12 0.36 0.05 7.56 1.42 22 DI1 – Sa dinia Channel 7.22 1.065 0.40 0.04 8.08 1.11 14 4 48UR20050529/ 1.13 0.057 1.08 205 314 DS1 – Albo an Sea 6.4 1.15 0.38 0.041 6.26 1.02 32 DS3 – Alge ian Wes 7.6 1.13 0.41 0.06 7.33 0.99 73 DS4 – Alge ian Eas 7.48 1.13 0.41 0.06 7.50 1.23 47 DT1 – Ty henian No h 7.24 0.44 0.42 0.03 7.91 0.56 16 DT3 – Ty henian Sou h 7.70 0.38 0.41 0.03 7.55 0.36 14 DI1 – Sa dinia Channel 7.58 1.08 0.43 0.049 7.42 0.82 23 5 48UR20051116/ 1.35 0.078 0.98 16 738 DT1 – Ty henian No h 5.68 1.26 0.19 0.08 6.30 0.92 10 DT3 – Ty henian Sou h 6.71 1.51 0.20 0.06 6.86 1.065 5 DI1 – Sa dinia Channel 6.29 0 0.26 0 7.53 0 1 6 48UR20060608/ 1.16 0.054 1.47 221 27 DF2 – Gul o Lion 7.69 1.02 0.42 0.04 7.089 1.04 27 DF3 – Ligu o-P o ençal 8.08 0.78 0.43 0.04 7.41 1.21 35 DS2 – Balea ic Sea 8.06 0.9 0.43 0.03 7.07 1.18 30 DF1 – Alge o-P o ençal 7.97 1.16 0.44 0.05 7.34 1.32 61 DS3 – Alge ian Wes 8.39 0.9 0.42 0.03 8.5 2 28 DT3 – Ty henian Sou h 6.39 1.28 0.36 0.06 6.86 1.7 26 DI1 – Sa dinia Channel 8.04 0.85 0.43 0.04 7.77 1.25 14 7a06A420060720 – – – – 1367 8 48UR20060928/ 0.71 0.036 0.76 179 606 DS2 – Balea ic Sea 7.97 0.17 0.33 0.017 7.84 0.27 4 DF1 – Alge o-P o ençal 8.17 0.22 0.33 0.026 8.11 0.3 22 DS1 – Albo an Sea 8.2 0.14 0.35 0.02 8.59 0.35 47 DS3 – Alge ian Wes 7.93 0.89 0.33 0.03 8.09 0.91 70 DS4 – Alge ian Eas 7.98 0.68 0.34 0.04 8.01 0.7 28 DT3 – Ty henian Sou h 6.2 1.51 0.28 0.04 6.71 1.45 3 DI1 – Sa dinia Channel 7.66 0.6 0.28 0.02 8.00 0.49 5 9 48UR20071005/ 0.89 0.040 0.86 302 751 DF2 – Gul o Lion 8.41 0.08 0.31 0.01 7.43 0.02 4 DF3 – Ligu o-P o ençal 8.17 1.08 0.31 0.03 7.64 1.08 81 DS2 – Balea ic Sea 8.17 0.43 0.31 0.02 7.58 0.39 29 DF1 – Alge o-P o ençal 8.33 0.6 0.32 0.03 7.79 0.69 82 DS4 – Alge ian Eas 8.41 0.2 0.33 0.018 7.90 0.26 19 DT1 – Ty henian No h 7.83 0.41 0.28 0.03 8.26 0.55 26 DT3 – Ty henian Sou h 7.49 1.22 0.28 0.05 7.71 1.26 38 DI1 – Sa dinia Channel 7.92 1.05 0.33 0.02 8.26 0.41 23 h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 2000 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Figu e 8. Da ase compa ison be o e (black) and a e (blue) adjus men , showing e ical p o iles o (a) ni a e (µmolkg−1), (b) phospha e (µmolkg−1), and (c) silica e (µmolkg−1). Sca e plo s o he adjus ed da a om all dep hs a e i s and second quali y con ol o (d) phos- pha e s. ni a e (µmolkg−1) and (e) silica e s. ni a e (µmolkg−1). The black lines ep esen he bes linea i be ween he wo pa ame e s, and he co esponding equa ions and 2 alues a e shown in each plo . A e age esul ing N :P a io is 22.09; a e age esul ing N :Si a io is 0.94 (whole dep h). C uise no. 1 (48UR20040526). The adjus ed alues a e s ill lowe han he e e ence (Figs. 5–7, S3–S5) and a e s ill ou side he 2% accu acy ange. This c uise had s a ions in he S ai o Sicily, Ty henian No h and Sou h, and Ligu ian Eas sub egions (Fig. 9, igh side), and only ou s a ions we e deepe han 1000db ( hose wi hin he Ty henian Sea). The low p ecision o his c uise had al eady been e idenced du ing he i s QC (Sec . 3.1). We ecommend lagging his c uise as ques ionable ( lag 3). C uise no. 5 (48UR20051116). This c uise ook place be- ween he S ai o Sicily and he Ty henian No h and Sou h (Fig. 9, igh side). Ni a e, phospha e, and silica e da a we e lowe han hose om o he c uises (no. 3 and no. 4) un he same yea (Figs. 5–7, S3–S5) and a e s ill biased a e adjus - men s. Conside ing he limi ed p ecision and he low numbe o c osso e s, i is ecommended o lag he c uise as ques- ionable ( lag 3). C uise no. 6 (48UR20060608). This c uise had an o se wi h i e c uises, gi ing e idence ha adjus men s o 1.05 o ni a e, 0.86 o phospha e, and 1.26 o silica e a e needed. The silica e bias was educed a e adjus men bu emained la ge wi h espec o he accu acy limi (Figs. 7, S5). This c uise has a wide geog aphic co e age, wi h s a ions along nine sec ions (Fig. 9, igh side). Conside ing also he high s anda d de ia ion (Table 4), which is pa ially a ibu ed o he spa ial co e age o he c uise, he e is s ill unce ain y abou he quali y o he samples. I is ecommended o lag silica e da a o c uise no. 6 as ques ionable ( lag 3). C uise no. 12 (48UR20081103). Phospha e da a ha e low accu acy wi h espec o he e e ence c uises (Figs. 6, S4). This c uise has s a ions along a longi udinal sec ion om he S ai o Sicily o he Albo an Sea, which migh explain he la ge s anda d de ia ion o deep phospha e samples (Ta- ble 4). C uise no. 12 was gi en a co ec ion o 1.08 o ni- a e, 1.12 o silica e, and 1.38 o phospha e. The mean o se om i e c osso e s compu ed wi hin he Ty henian Sou h, Sa dinia Channel, Alge ian Eas , Alge ian Wes , and Albo an Sea sub egions sugges s ha his c uise has lowe nu ien alues han he e e ence c uise. A e adjus men , c uise no. 12 is wi hin he accep able ange o ni a e and silica e bu no o phospha e as highligh ed in Sec . 3.2. In addi ion, conside ing he ela i ely high numbe o s a- ions >1000 db and a plausible end in phospha e, i is ec- ommended o lag he phospha e da a as good/accep able ( lag 2). C uise no. 15 (48UR20100731). This c uise has 149 s a- ions along a simila ack o c uise no. 12 bu shows la ge o se s o phospha e and silica e (Figs. 6–7, S4– Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 2001 Figu e 9. h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 2002 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Figu e 9. Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 2003 Figu e 9. h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 2004 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Figu e 9. Ve ical p o iles o he ino ganic nu ien s in he da ase a e adjus men s and spa ial co e age o each c uise ( e e ence o c uise ID is abo e each map). The whole WMED adjus ed p oduc is shown in black, while he da a o each indi idual c uise a e shown in blue ( lag 2) and g een ( lag 3). Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 2005 S5) compa ed o c uise no. 12. Conside ing ha deep sili- ca e da a we e no o low quali y (small s anda d de ia ion; see Table 4) and ha deep phospha e da a all wi hin he phospha e end discussed abo e, hese da a a e lagged as good/accep able ( lag 2). C uise no. 16 (48UR20101123). The c uise shows la ge o se s o phospha e and silica e (Figs. 6–7, S4–S5), simila o c uise no. 15. Conside ing ha he o e all c uise s anda d de ia ion o silica e samples below 1000 db was ela i ely high (1.02 o e 14 samples; see Table 4); ha i has only one c osso e be ween he Ty henian No h and Sou h sub- egions (Table 6); and ha when compa ing deep egional a e ages, his c uise had he lowes a e age silica e alue, i is ecommended o lag silica e da a o c uise no. 16 as ques- ionable ( lag 3). As o phospha e, he c uise is pa o he phospha e end and is he e o e lagged as good/accep able ( lag 2). C uise no. 24 (48QL20171023). This c uise has he la ges o se o ni a e e en a e adjus men . I is e y likely due o a di e ence be ween labo a o ies (calib a ion s anda ds) conce ning ni a e, which needs o be lagged as ques ionable ( lag 3) in he inal p oduc . The e a e se e al sou ces o bias in he obse a ion. One o he main easons o an upwa d o downwa d bias would be he di e ence in he nu ien ’s chemical analy ical me hod and he lack o use o CRMs in all c uises as also no ed in CARINA (Tanhua e al., 2009) o in he mos ecen global compa abili y s udy by Aoyama (2020). C uises discussed in his sec ion we e no emo ed om he inal p oduc bu we e e ained along wi h hei ecom- mended quali y lag (Table 3), de ailed abo e and in he Sup- plemen – Pa 2, A2. We ha e ca ied ou he e alua ion o hei o e all quali y bu lea e up o he use s how o app o- p ia ely use hese da a. 4.5 P oduc assessmen – compa ison wi h MEDATLAS A e age wa e mass biogeochemical p ope ies ha e been compu ed om he adjus ed p oduc (Table 7) and compa ed o he MEDAR/MEDATLAS annual clima ological p o iles, downloaded om he I alian Na ional Oceanog aphic Da a Cen e (NODC) websi e (h p://doga.ogs. ies e.i /meda /, las access: Feb ua y 2020) gi en by Manca e al. (2004), in o de o e alua e and assess he new p oduc . Since nu ien p ope ies exhibi di e ences wi h dep hs, we compa ed a - e age nu ien concen a ions o he h ee main wa e masses in 12 sub egions o he WMED (Table 7, Fig. S2). The esul s o Table 7 compa e wa e mass biogeochemi- cal p ope ies wi h he e e ence clima ology. The new p od- uc ag ees well wi h he MEDATLAS clima ology. Howe e , he e a e some dis inc ions. The su ace laye (0–150db) is cha ac e ized by a low nu ien con en . The su ace ni a e a ies be ween 0.69 and 2.75 µmolkg−1wi h a maximum ound in he Ligu ian Eas (DF4) and a minimum in he Albo an Sea (DS1) sub egions; simila alues we e eco ded in he clima ology (0.61–3.00 µmolkg−1). The di e ences in ni a e a e ages in he su ace laye a e obse ed in he Gul o Lion (DF2) whe e he new p oduc is highe han he cli- ma ology and sligh ly lowe in he Ligu o-P o ençal (DF3). As o he su ace con en in phospha e, i a ied be ween 0.04 and 0.16 µmolkg−1wi h a maximum ound in he Lig- u ian Eas (DF1) and a minimum in he Albo an Sea (DS1), like he MEDATLAS clima ology, whe e phospha e a e - ages luc ua e be ween 0.05 and 0.19 µmolkg−1. The new p oduc is sligh ly lowe compa ed o he clima ology. As o he a e age su ace in silica e, i a ies be ween 1.36 and 2.91 µmolkg−1wi h a minimum ound in he Ligu ian Eas (DF4) and a maximum in he Gul o Lion (DF2), while in he clima ology, i a ied be ween 1.27 and 2.31 µmolkg−1 ( he minimum in he Ligu ian Eas – DF4 – and he maxi- mum in he Albo an Sea – DS1). The new p oduc is sligh ly highe in silica e. O e all, he di e ences in he su ace laye a e obse ed in he Gul o Lion (DF2), he Ligu o-P o ençal (DF3), and he Ligu ian Eas (DF4) egions which could be due o he in- ense a iabili y o he e ical mixing occu ing in he no h- e n WMED compa ed o in he o he sub egions. In he in e media e laye , a e ages we e compu ed om he dep h o he salini y maximum (Smax)±100 m om a egional a e age p o ile, indica i e o he Le an ine In e - media e Wa e (LIW) co e. The ni a e a e age a ied be- ween 4.94 and 9.32 µmolkg−1whe e he minimum con en was eco ded in he S ai o Sicily (DI3) and he maximum in he Alge ian Wes , (DS3) while in he MEDATLAS cli- ma ology, ni a e was be ween 5.14 and 8.60µmolkg−1. On a e age, he lowes con en in ni a e was in he Ty henian No h (DT1) and Sou h (DT3), Sa dinia Channel (DI1), and S ai o Sicily (DI3), while LIW o he Gul o Lion (DF2), Ligu o-P o ençal (DF3), Ligu ian Eas (DF4), Balea ic Sea (DS2), Alge o-P o ençal (DF1), Albo an Sea (DS1), and Al- ge ian Wes (DS3) and Eas (DS4) sub egions was ela i ely ich in ni a e compa ed o he MEDATLAS p oduc , hough he new p oduc was sligh ly highe mainly in he Gul o Lion (DF2), Ligu ian Eas (DF4), and Balea ic Sea (DS2). As o phospha e, LIW a e ages showed simila beha iou o ni a e: he lowes phospha e con en (0.21–0.27 µmolkg−1) was obse ed in he eas e n sub egions o he WMED (DI3, DI1, DT3, and DT1), while he maximum concen a ions (0.4–0.37 µmolkg−1) we e epo ed in he wes e n sub e- gions o he WMED (DS1, DS3 and DS4, DS2, and DF2). The la ge di e ences be ween he wo p oduc s we e in he Ligu ian Eas (DF4) and he Albo an Sea (DS1), sub egions wi h low numbe s o obse a ions. Conce ning silica e, he lowes a e age concen a ion (5.25 µmolkg−1) was obse ed in he LIW co e o he S ai o Sicily (DI3) and he maximum concen a ions (8.66– 8.77 µmolkg−1) we e in he Albo an Sea (DS1) and Gul o Lion (DF2); simila alues we e eco ded in he MEDATLAS clima ology (4.86–7.95 µmolkg−1). The e a e some disc ep- ancies whe e he new p oduc was highe pa icula ly in he h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 2006 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea Table 7. Wa e mass p ope ies and egional a e age concen a ions o ino ganic nu ien s: compa ison be ween he new adjus ed p oduc and he MEDAR/MEDATLAS clima ology (wi h s anda d de ia ions and numbe o obse a ions in b acke s). Region and wa e mass Ni a e (µmolkg−1) Phospha e (µmolkg−1) Silica e (µmolkg−1) A g new p oduc A g Meda A g new p oduc A g Meda A g new p oduc A g Meda DF2 – Gul o Lion Su ace wa e (0–150db) 2.68 ±2.53 (68)a1.7±1.1 0.15 ±0.06 (68) 0.13 ±0.04 2.91 ±1.33 (68) 1.72 ±0.64 LIW co e (Smax dep h ange: 300–500 db) 8.49 ±0.18 (17) 6.13 ±0.32 0.38 ±0.02 (17) 0.34 ±0.01 8.67 ±0.69 (17) 6.12 ±0.61 Deep wa e (>1500 db) 8.03 ±0.43 (33) 7.64 ±0.31 0.37 ±0.01 (33) 0.37 ±0.015 8.7±0.67 (33) 7.95 ±0.06 DF3 – Ligu o-P o ençal Su ace wa e (0–150db) 2.31 ±2.4 (205) 3.0±2.6 0.12 ±0.07 (205) 0.19 ±0.05 2.45 ±1.05 (205) 2.16 ±1.05 LIW co e (Smax dep h ange: 300–500 db) 8.05 ±0.18 (76) 7.74 ±0.13 0.36 ±0.01 (76) 0.35 ±0.01 7.49 ±0.55 (76) 6.26 ±0.60 Deep wa e (>1500 db) 8.18 ±0.25 (142) 7.79 ±0.04 0.37 ±0.02 (142) 1.03 ±1.29 8.98 ±0.39 (142) 7.60 ±0.21 DF4 – Ligu ian Eas Su ace wa e (0–150db) 0.7±0.69 (228) 0.61 ±1.03 0.05 ±0.02 (228) 0.18 ±0.02 1.37 ±0.45 (228) 1.27 ±1.86 LIW co e (Smax dep h ange: 300–500 db) 6.8±0.4 (23) 5.54 ±0 0.3±0.02 (21) 0.36 ±0.06 5.86 ±0.9 (24) 4.86 ±0 Deep wa e (>1500 db) – – – – – – DS2 – Balea ic Sea Su ace wa e (0–150db) 1.32 ±1.46 (196) 1.19 ±1.5 0.08 ±0.04 (196) 0.11 ±0.04 1.61 ±0.64 (196) 1.54 ±0.78 LIW co e (Smax dep h ange: 300–500 db) 8.32 ±0.32 (58) 6.92 ±0.12 0.37 ±0.02 (60) 0.39 ±0.003 7.31 ±0.9 (60) 7.55 ±0.62 Deep wa e (>1500 db) 8.2±0.35 (88) – 0.37 ±0.01 (88) – 8.71 ±0.51 (88) 8.45 ±0.8 DF1 – Alge o-P o ençal Su ace wa e (0–150db) 0.87 ±0.85 (372) 1.08 ±1.7 0.05 ±0.02 (372) 0.07 ±0.05 1.42 ±0.3 (372) 1.28 ±0.73 LIW co e (Smax dep h ange: 300–500 db) 8.07 ±0.34 (126) 7.51 ±0.18 0.36 ±0.02 (126) 0.34 ±0.008 6.84 ±0.95 (126) 5.96 ±0.77 Deep wa e (>1500 db) 8.36 ±0.27 (300) 7.87 ±0.13 0.38 ±0.02 (300) 0.38 ±0.001 9.01 ±0.33 (300) 8.18 ±0.10 DS1 – Albo an Sea Su ace wa e (0–150db) 2.75 ±2.87 (299) 2.51 ±2.23 0.17 ±0.11 (299) 0.16 ±0.07 2.07 ±1.38 (299) 2.31 ±1.14 LIW co e (Smax dep h ange: 400–600 db) 8.89 ±0.4 (77) 8.14 ±0.11 0.42 ±0.02 (77) 0.37 ±0.008 8.77 ±1.66 (76) 7.95 ±0.34 Deep wa e (>1500 db) 7.72 ±0.81 (65) – 0.36 ±0.04 (65) – 8.98 ±0.63 (65) 8.16 ±0 DS3 – Alge ian Wes Su ace wa e (0–150db) 1.8±1.88 (254) 1.82 ±2.01 0.11 ±0.05 (354) 0.11 ±0.06 1.71 ±0.68 (354) 2.10 ±0.91 LIW co e (Smax dep h ange: 400–600 db) 9.33 ±0.08 (70) 8.28 ±0.15 0.41 ±0 (73) 0.38 ±0.012 8.1±0.53 (72) 6.68 ±0.80 Deep wa e (>1500 db) 8.37 ±0.27 (246) 8.047 ±0.013 0.37 ±0.02 (246) 0.36 ±0.006 9.22 ±0.35 (246) 8.87 ±0.23 DS4 – Alge ian Eas Su ace wa e (0–150db) 0.94 ±0.77 (170) 0.75 ±1.26 0.07 ±0.02 (170) 0.05 ±0.03 1.53 ±0.12 (170) 1.35 ±0.52 LIW co e (Smax dep h ange: 400–600 db) 8.5±0.25 (43) 8.60 ±0.06 0.38 ±0.03 (43) 0.38 ±0.008 7.27 ±0.67 (42) 7.092 ±0.55 Deep wa e (>1500 db) 7.94 ±0.24 (132) 8.06 ±0.06 0.36 ±0.02 (132) 0.38 ±0.006 8.73 ±0.38 (132) 9.04 ±0.24 DT1 – Ty henian No h Su ace wa e (0–150db) 1.03 ±1.14 (231) 0.88 ±1.2 0.06 ±0.02 (231) 0.09 ±0.03 1.64 ±0.52 (231) 2.19 ±0.59 LIW co e (Smax dep h ange: 400–600 db) 5.95 ±0.49 (43) 5.86 ±0.36 0.27 ±0.03 (44) 0.308 ±0.02 7.06 ±0.08 (44) 6.76 ±0.59 Deep wa e (>1500 db) 7.75 ±0.37 (194) 7.12 ±0.47 0.36 ±0.03 (194) 0.40 ±0.02 9.19 ±0.47 (194) 7.51 ±0.49 DT3 – Ty henian Sou h Su ace wa e (0–150db) 1.21 ±1.38 (711) 1.23 ±1.80 0.06 ±0.03 (711) 0.061 ±0.04 1.58 ±0.61 (711) 1.55 ±1.05 LIW co e (Smax dep h ange: 300–500 db) 6.2±0.28 (225) 6.42 ±0.01 0.26 ±0.02 (225) 0.254 ±0.005 6.28 ±0.65 (224) 6.68 ±0.44 Deep wa e (>1500 db) 7.88 ±0.4 (227) 7.12 ±0.26 0.37 ±0.02 (227) 0.31 ±0.007 9.04 ±0.52 (227) 8.02 ±0.07 DI1 – Sa dinia Channel Su ace wa e (0–150db) 1.22 ±1.39 (271) 1.42 ±1.95 0.07 ±0.03 (271) 0.064 ±0.03 1.57 ±0.68 (271) 1.39 ±1.01 LIW co e (Smax dep h ange: 300–500 db) 6.52 ±0.17 (89) 6.45 ±0.22 0.27 ±0.02 (89) 0.250 ±0.01 6.36 ±0.67 (89) 6.27 ±0.70 Deep wa e (>1500 db) 7.91 ±0.62 (107) – 0.37 ±0.03 (107) 0.32 ±0 8.64 ±0.91 (107) – DI3 – S ai o Sicily Su ace wa e (0–150db) 0.87 ±0.68 (583) 0.77 ±0.81 0.06 ±0.02 (583) 0.063 ±0.02 1.53 ±0.29 (583) 1.44 ±0.58 LIW co e (Smax dep h ange: 200–400 db) 4.95 ±0.47 (80) 5.14 ±0.14 0.21 ±0.02 (78) 0.194 ±0.004 5.26 ±0.79 (81) 6.744 ±0.41 Deep wa e (>1500 db) – – – – – – aA e age (A g) ±s anda d de ia ion o ino ganic nu ien s ( he numbe o obse a ions wi hin dep h ange) o h ee laye s om he adjus ed o new p oduc and MEDATLAS e ical clima ological p o iles (called he e Meda ). Regions a e de ined acco ding o Manca e al. (2004; Table S2, Fig. S2). Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 2007 Figu e 10. RMSE egional a e ages o wa e mass p ope ies com- pu ed be ween he new adjus ed p oduc and MEDAR/MEDATLAS clima ology o ni a e (a), phospha e (b), and silica e (c). Gul o Lion (DF2), Ligu o-P o ençal (DF3), and Alge ian Wes (DS3) sub egions. This di e ence is explained by he limi ed numbe o obse a ions wi hin he dep h ange in he new p oduc compa ed o he obse a ions used in he clima- ology in hese sub egions. Re e ing o Manca e al. (2004), he LIW co e salini y alues a e ela i ely mo e p onounced in he S ai o Sicily (DI3), Sa dinia Channel (DI1), and Ty henian Sou h (DT3) and No h (DT1) sub egions, whe e nu ien s we e lowe han in he wes e n sub egions (DS3, DS4, DS1, DF1, DS2, DF4, DF3, DF2). The a e ages o nu ien wi hin he LIW co e ie in well wi h he MEDATLAS clima ology a e ages (Table 7), excep in sub egions wi h impo an e ical mix- ing. We ha e also e i ied a e age biochemical p ope ies in he deep laye (below 1500 db). The new p oduc is sligh ly highe in ni a e a e ages (7.74–8.37µmolkg−1) han he MEDATLAS clima ology (7.12–8.06µmolkg−1; Table 7). The la ges di e ence was ound in he Ty henian Sou h (DT3) and No h (DT1) sub egions. This di e ence could be due o he ac ha we a e compa ing wo di e en ime pe iods (2004–2017 and 1908–2001). As o he deep laye phospha e, a e age concen a ions a ied be ween 0.35 and 0.37 µmolkg−1and we e wi hin he clima ology lim- i s (0.31–0.40 µmolkg−1). In all sub egions, he e we e no la ge di e ences. O e all, phospha e was in acco dance wi h he MEDATLAS clima ology. Simila ly o ni a e, deep a e age silica e in he new p oduc (8.64–9.21µmolkg−1) was highe han he clima ology (7.51 o 9.04µmolkg−1). The la ges di e ence in a e age silica e was obse ed in he Ty henian No h (DT1) and Sou h (DT3) and Ligu o- P o ençal (DF3) sub egions. We hen used he oo mean squa e e o (RMSE) as a s a- is ical index o quan i y he di e ence be ween a e aged e- gional p o iles om he new p oduc and MEDATLAS p od- uc . The clima ology annual p o iles we e in e pola ed o he egional a e age p o iles o he new p oduc , and he a e - age RMSE o each laye and sub egion was calcula ed. Fig- u e 10 shows he egional e olu ion o RMSE in he main wa- e masses o he h ee nu ien s. Fo ni a e (Fig. 10a), he RMSE in he su ace laye a ied be ween 0.12 µmolkg−1(in he Ty henian No h – DT1) and 1.36µmolkg−1(in he Gul o Lion – DF2); in he in e media e laye , he RMSE was be- ween 0.07µmolkg−1(in he Sa dinia Channel – DI1) and 2.35 µmolkg−1(in he Gul o Lion – DF2) and was lowe in he deep laye , be ween 0.11 µmolkg−1(in he Alge ian Eas – DS4) and 0.79µmolkg−1( he Gul o Lion – DF2). The RMSE dec eases in he Alge ian Eas (DS4), Ty he- nian No h (DT1), Ty henian Sou h (DT3), Sa dinia Chan- nel (DI1), and S ai o Sicily (DI3). This illus a es he low di e ence be ween he wo p oduc s. Fo phospha e (Fig. 10b), he RMSE anges be ween 0.0022 µmolkg−1(in he Ty henian Sou h – DT3) and 0.12 µmolkg−1(in he Ligu ian Eas – DF4) in he su - ace laye and is be ween 0.003µmolkg−1(in he Ligu o- P o ençal sub egion – DF3) and 0.048µmolkg−1(in he Albo an Sea – DS1) a in e media e dep hs, while in he deep laye RMSE a ied be ween 0.0087 (in he Gul o Lion – DF2) and 0.057 µmolkg−1(in he Ty henian No h – DT1). Rega ding silica e RMSE (Fig. 10c) in he su ace laye , i a ied be ween 0.13 µmolkg−1(in he Alge o-P o ençal sub egion – DF1) and 3.5µmolkg−1(in he Ligu ian Eas sub egion – DF4); a lowe RMSE o be ween 0.10 µmolkg−1 (in he Sa dinia Channel – DI1) and 2.54 µmolkg−1(in he Gul o Lion – DF2) was epo ed in he in e media e laye ; he esul s in he deep laye we e be ween 0.33µmolkg−1(in he Alge ian Eas – DS4) and 1.43 µmolkg−1(in he Ligu o- P o ençal sub egion – DF3). The bes ag eemen be ween he wo p oduc s was ob- se ed in he in e media e and deep laye . The lowes RMSE was con ined o he deep laye in mos o he sub egions, while he highes di e ence was ound in he su ace laye since i is subjec ed o in ense e ical mixing mainly in he no he n WMED. Compa ing a e ages in sub egions showed simila di e ences in nu ien s be ween he wo p oduc s pa - icula ly in he Gul o Lion (DF2), he Ligu o-P o ençal (DF3), Ligu ian Eas (DF4), and Alge ian Eas (DS4), due o he ela i ely high a iabili y in nu ien concen a ions in hese sub egions. These di e ences a e no signi ican as he e is a disc epancy in he numbe o obse a ions used in he wo p oduc s. O e all, ino ganic nu ien s o he new p oduc ag ee e y well wi h he MEDAR/MEDATLAS cli- ma ology. The main ea u es o he spa ial dis ibu ion in he ino ganic nu ien s we e in acco dance wi h he indings o Manca e al. (2004), whe e he ela i ely high con en in nu ien s was ound mo e in he in e media e laye o he Alge ian sub egions (DF1, DS3, DS4) han in o he sub e- h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 2008 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea gions (Table 7). Besides, he highes concen a ions in deep- laye silica e we e epo ed in he Alge ian sub egions in he wo p oduc s (9.21µmolkg−1– DS3 – in he new p oduc ; 9.04 µmolkg−1– DS4 – in he clima ology), which is indica- i e o he poo egional en ila ion and o he longe esi- dence ime o deep wa e especially in hese sub egions. 5 Da a a ailabili y The inal p oduc is a ailable as a *.cs me ged ile om PANGAEA and can be accessed a h ps://doi.o g/10.1594/PANGAEA.904172 (Belgacem e al., 2019). Ancilla y in o ma ion is in he supplemen a y ma e ials wi h he lis o a iables included in he o iginal and inal p oduc . Table 1a and b summa izes all c uises included in he da ase . The da ase includes equen ly measu ed s a- ions and key ansec s o he WMED wi h in si u physi- cal and chemical oceanog aphic obse a ions. As men ioned, wo iles a e accessible; bo h include oceanog aphic a iables obse ed a he s anda d dep hs (see Supplemen Pa 2). O iginal da ase – CNR_DIN_WMED_20042017_ o iginal.cs . This is he o iginal da ase wi h a lag a iable o each o he ollowing pa ame e s: CTD salini y, ni a e, phospha e, and silica e om he p ima y quali y con ol (de ailed in Sec . 3.1). Adjus ed da ase – CNR_DIN_WMED_20042017_ adjus ed.cs . This is he p oduc a e p ima y quali y con ol and a e applying he adjus men ac o s om he seconda y quali y con ol. Recommenda ions o Sec . 4.4 a e included, as well as quali y lags. 6 Final ema ks An in e nally consis en da ase o dissol ed ino ganic nu i- en s has been gene a ed o he WMED (2004–2017). The accu acy en elope o ni a e and silica e was se o 2%, a p ede ined limi used in GLODAP and CARINA da a p od- uc s. Rega ding phospha e da a, hese we e almos en i ely ou side his limi , because o phospha e’s na u al a ia ions and he o e all e y low concen a ions in he WMED, a highly P-limi ed basin. Using a c osso e analysis (second QC oolbox) o compa e c uises wi h espec o eliable e - e ence da a imp o ed he accu acy o he measu emen s by minimizing he bias in he indi idual c uises. The new p od- uc was b oadly consis en wi h he ea lie clima ology o MEDAR/MEDATLAS. The publica ion o a quali y-con olled ex ensi e (spa- ially and empo ally) da abase o ino ganic nu ien s in he WMED was imely and ills a gap in in o ma ion ha p e- en ed baseline assessmen s on spa ial and empo al a iabil- i y o biogeochemical ace s in he Medi e anean. In com- bina ion wi h olde da abases in he same egion (e.g. bo le da a a ailable in he MEDAR/MEDATLAS da abase), his new da a p oduc will hus cons i u e a pilla on which he Medi e anean ma ine scien i ic communi y will be able o build wi h o iginal esea ch opics on biogeochemical luxes and cycles and hei ela ion o hyd ological changes ha oc- cu ed in he pe iod co e ed by he da ase . The da ase is also ele an o he modelling communi y as i can be used as an independen da a p oduc o assess eanalysis p oduc s o i can be assimila ed in new eanalysis p oduc s. Supplemen . The supplemen ela ed o his a icle is a ailable online a : h ps://doi.o g/10.5194/essd-12-1985-2020-supplemen . Au ho con ibu ions. MaB, MA, SKL, JC, and KS subs an ially con ibu ed o w i ing he manusc ip . SC, GC, and FA an he chemical analysis and con ibu ed o he manusc ip . MiB coo di- na ed he echnical aspec s o mos o he c uises. SC, GC, FA, AR, and BP con ibu ed o speci ic pa s o he manusc ip . Compe ing in e es s. The au ho s decla e ha hey ha e no con- lic o in e es . Acknowledgemen s. The da a ha e been collec ed in he ame- wo k o se e al na ional and Eu opean p ojec s, e.g. KM3NeT, EU GA no. 011937; SESAME, EU GA no. GOCE-036949; PERSEUS, EU GA no. 287600; OCEAN-CERTAIN, EU GA no. 603773; COMMON SENSE, EU GA no. 228344; EUROFLEETS, EU GA no. 228344; EUROFLEETS2, EU GA no. 312762; JERICO, EU GA no. 262584; and he I alian PRIN 2007 p og amme “Ty he- nian Seamoun s ecosys ems” and he I alian RITMARE lagship p ojec , bo h unded by he I alian Minis y o Educa ion, Uni e si y and Resea ch. We hank Sa ah Ju e s öm om he Swedish En i- onmen al Resea ch Ins i u e o in aluable help in quali y con ol discussions. We would like o exp ess ou app ecia ion o he IN- OCEN labo a o y eam a IEO o hei help and collabo a ion du - ing Malek Belgacem’s s ay he e. The au ho s a e deeply indeb ed o all in es iga o s and analys s who con ibu ed o da a collec ion a sea du ing so many yea s as well as o he PIs o he c uises (S e- ano Aliani, Ma io As aldi, Mau izo Azza o, Massimiliano Dib- i e o, Gian Pie o Gaspa ini, Annalisa G i a, Je Haun, Loïc Jul- lion, Gina La Spada, Elena Mannini, Angelo Pe illi, Chia a San- inelli, S e ania Spa nocchia), he cap ains, and he c ews o allow- ing he collec ion o his eno mous da ase ; wi hou hem, his wo k would no ha e been possible. Re iew s a emen . This pape was edi ed by Bi gi Heim and e- iewed by Tos e Tanhua and Ma ina Lipize . Re e ences Aoyama, M.: Global ce i ied- e e ence-ma e ial- o e e ence- ma e ial-scaled nu ien g idded da ase GND13, Ea h Sys . Sci. Ea h Sys . Sci. Da a, 12, 1985–2011, 2020 h ps://doi.o g/10.5194/essd-12-1985-2020 M. Belgacem e al.: Dissol ed ino ganic nu ien s in he wes e n Medi e anean Sea 2009 Da a, 12, 487–499, h ps://doi.o g/10.5194/essd-12-487-2020, 2020. Aoyama, M., Woodwa d, E., Malcolm, S., Bakke , K., Becke , S., Bjö kman, K., Daniel, A., Maha ey, C., Mu a a, A., Naik, H., Tanhua, T., Rho, T., Roman, R., and Sloyan, B.: Compa abili y o oceanic nu ien da a, Pos e Clus e Communi y Whi epape , CLIVAR Open Science Con e ence on “Cha ing he cou se o clima e and ocean esea ch”, 18–25 Sep embe 2016, Qingdao (China), 12 pp., 2016. Becke , S., Aoyama, M., Woodwa d, E. M. S., Bakke , K., Co - e ly. S., Maha ey, C., and Tanhua, T.: GO-SHIP Repea Hyd og- aphy Nu ien Manual: The p ecise and accu a e de e mina ion o dissol ed ino ganic nu ien s in seawa e , using Con inuous Flow Analysis me hods, in: The GO-SHIP Repea Hyd og a- phy Manual: A Collec ion o Expe Repo s and Guidelines, 56, h ps://doi.o g/10.25607/OBP-555, 2019. Belgacem, M., Chiggia o, J., Bo ghini, M., Pa oni, B., Ce a i, G., Ac i, F; Cozzi, S., Ribo i, A., Ál a ez, M., Lau se , S. K., and Sch oede , K.: Quali y con olled da ase o dis- sol ed ino ganic nu ien s in he wes e n Medi e anean Sea (2004–2017) om R/V oceanog aphic c uises, PANGAEA, h ps://doi.o g/10.1594/PANGAEA.904172, 2019. Bé houx, J. P.: Oxygen consump ion, new p oduc ion, e ical ad- ec ion and en i onmen al e olu ion in he Medi e anean Sea, Deep-Sea Res. P . I, 36, 769–781, h ps://doi.o g/10.1016/0198- 0149(89)90150-7, 1989. Bé houx, J. P., Mo in, P., Madec, C., and Gen ili, B.: Phospho- us and ni ogen beha iou in he Medi e anean Sea, Deep- Sea Res. P . I, 39, 1641–1654, h ps://doi.o g/10.1016/0198- 0149(92)90053-V, 1992. Bé houx, J. P., Mo in, P., Chaume y, C., Connan, O., Gen ili, B., and Ruiz-Pino, D.: Nu ien s in he Medi e anean Sea, mass balance and s a is ical analysis o concen a ions wi h e- spec o en i onmen al change, Ma ine Chem., 63, 155–169, h ps://doi.o g/10.1016/S0304-4203(98)00059-0, 1998. Bé houx, J. P., Gen ili, B., Mo in, P., Nicolas, E., Pie e, C., and Ruiz-Pino, D.: The Medi e anean Sea: a minia u e ocean o clima ic and en i onmen al s udies and a key o he clima ic uncioning o he No h A lan ic, P og. Oceanog ., 44, 131–146, 1999. Bé houx, J. P., Mo in, P., and Ruiz-Pino, D. P.: Tempo al ends in nu ien a ios: Chemical e idence o Medi e anean ecosys- em changes d i en by human ac i i y, Deep-Sea Res. P . II, 49, 2007–2016, h ps://doi.o g/10.1016/S0967-0645(02)00024- 3, 2002. Boyd, P. W.: Beyond ocean acidi ica ion, Na . Geosci., 4, 273–274, h ps://doi.o g/10.1038/ngeo1150, 2011. Coppola, L., Raimbaul , P., Mo ie , L., and Tes o , P.: Moni o - ing he en i onmen in he no hwes e n Medi e anean Sea, Eos, 100, 100–951, h ps://doi.o g/10.1029/2019EO125951, 2019. Dickson, A. G., A ghan, J. D., and Ande son, G. C.: Re e - ence ma e ials o oceanic CO2analysis: A me hod o he ce i ica ion o o al alkalini y, Ma ine Chem., 80, 185–197, h ps://doi.o g/10.1016/S0304-4203(02)00133-0, 2003. Do e, J. E., Houlihan, T., Hebel, D. V., Tien, G., Tupas, L., and Ka l, D. M.: F eezing as a me hod o sample p ese a ion o he analysis o dissol ed ino ganic nu ien s in seawa e , Ma ine Chem., 53, 173–185, 1996. Fichau , M., Ga cia, M. J., Gio ge i, A., Iona, A., Kuzne so , A., Rixen, M., and G oup, M.: MEDAR/MEDATLAS 2002: A Medi e anean and Black Sea da abase o ope a ional oceanog aphy, Else ie Oceanog aphy Se ies, 69, 645–648, h ps://doi.o g/10.1016/S0422-9894(03)80107-1, 2003. Gio ge i, A., Pa escano, E., Ba h, A., Buga, L., Ga i, J., Gio gi, G., Iona A., Lipize , M., Holdswo h, N., La sen, M.M., Schaap, D., Vinci, M., and Wenze , M.: EMODne Chemis y Spa ial Da a In as uc u e o ma ine obse a ions and ela ed in o ma- ion, Ocean Coas . Manage., 166, 9–17, 2018. Gio gi, F.: Clima e change ho -spo s, Geophys. Res. Le ., 33, 1–4, h ps://doi.o g/10.1029/2006GL025734, 2006. Gou e ski, V. V. and Jancke, K.: Sys ema ic e o s as he cause o an appa en deep wa e p ope y a iabili y: Global analysis o he WOCE and his o ical hyd og aphic da a, P og. Oceanog ., 48, 337–402, h ps://doi.o g/10.1016/S0079-6611(00)00049-5, 2000. G assho , K., K emling, K., and Eh ha d , M.: Me hods o seawa- e analysis (3 d edn.), Weinheim P ess, WILEY-VCH, 203–273, 1999. Hansen, H. P. and Ko ole , F.: De e mina ion o nu ien s, Me hods o Seawa e Analysis, 10, 159–228, 1999. Hoppema, M., Velo, A., an Heu en, S., Tanhua, T., Key, R. M., Lin, X., Bakke , D. C. E., Pe ez, F. F., Ríos, A. F., Lo Monaco, C., Sabine, C. L., Ál a ez, M., and Belle by, R. G. J.: Consis- ency o c uise da a o he CARINA da abase in he A lan ic sec o o he Sou he n Ocean, Ea h Sys . Sci. Da a, 1, 63–75, h ps://doi.o g/10.5194/essd-1-63-2009, 2009. Hydes, D. J., Aoyama, M., Amino , A., Bakke , K., Becke , S., Co - e ly, S., Daniel, A., Dickson, A. G., G osso, O., Ke ouel, R., an Ooijen, J., Sa o, K., Tanhua, T., Woodwa d, E. M. S., and Zhang, J. Z.: De e mina ion o Dissol ed Nu ien s (N, P, SI) in Sea- wa e Wi h High P ecision and In e -Compa abili y Using Gas- Segmen ed Con inuous Flow Analyse s, in: The GO-SHIP Re- pea Hyd og aphy Manual: A Collec ion o Expe Repo s and Guidelines. Ve sion 1, edi ed by: Hood, E. M., Sabine, C. L., and Sloyan, B. M., IOCCP Repo Numbe 14, ICPO Publica- ion Se ies Numbe 134, 87 pp., h ps://doi.o g/10.25607/OBP- 555, 2010. Johnson, G. C., Robbins, P. E., and Hu o d, G. E.: Sys ema ic adjus men s o hyd og aphic sec- ions o in e nal consis ency, J. A mos. Ocean. Tech., 18, 1234–1244, h ps://doi.o g/10.1175/1520- 0426(2001)018<1234:SAOHSF>2.0.CO;2, 2001. Key, R. M., Kozy , A., Sabine, C. L., Lee, K., Wanninkho , R., Bullis e , J. L., Feely, R. A., Mille o, F. J., Mo dy, C., and Peng, T. H.: A global ocean ca bon clima ology: Resul s om Global Da a Analysis P ojec (GLODAP), Global Biogeochem. Cycles, 18, 1–23, h ps://doi.o g/10.1029/2004GB002247, 2004. Lau se , S. K. and Tanhua, T.: A oolbox o sec- onda y quali y con ol on ocean chemis y and hyd o- g aphic da a, Limnol. Oceanog . Me hods, 13, 601–608, h ps://doi.o g/10.1002/lom3.10050, 2015. Lazza i, P., Solido o, C., Salon, S., and Bolzon, G.: Spa ial a i- abili y o phospha e and ni a e in he Medi e anean Sea: A modeling app oach, Deep-Sea Res. P . I, 108, 39–52, h ps://doi.o g/10.1016/j.ds .2015.12.006, 2016. Lejeusne, C., Che aldonné, P., Pe gen -Ma ini, C., Boudou esque, C. F., and Pé ez, T.: Clima e change e ec s on a minia- h ps://doi.o g/10.5194/essd-12-1985-2020 Ea h Sys . Sci. Da a, 12, 1985–2011, 2020