P inputs determine denitrifier abundance explaining dissolved nitrous oxide in reservoirs

Limnol. Oceanog . 9999, 2023, 1–16 © 2023 The Au ho s. Limnology and Oceanog aphy published by Wiley Pe iodicals LLC on behal o Associa ion o he Sciences o Limnology and Oceanog aphy. doi: 10.1002/lno.12381 P inpu s de e mine deni ifie abundance explaining dissol ed ni ous oxide in ese oi s Elizabe h Le on-Palme o , 1 * ,a Ra ael Mo ales-Baque o , 1 Isabel Reche 1,2 * 1 Ins i u o del Agua and Depa amen o de Ecología, Uni e sidad de G anada, G anada, Spain 2 Resea ch Uni Modeling Na u e (MNa ), Uni e sidad de G anada, G anada, Spain Abs ac Rese oi s a e impo an si es o ni ogen p ocessing, especially hose loca ed in ag icul u al and u ban wa e sheds. Ni ogen inpu s p omo e N 2 O p oduc ion and emission, bu he mic obial pa hways con olling N 2 O ha e been seldom s udied in ese oi wa e columns. We de e mined N 2 O concen a ion in he wa e column o 12 ese oi s du ing he summe s a ifica ion and win e mixing. We explo ed he po en ial mic obial sou ces and sinks o N 2 O by quan i ying key genes in ol ed in ammonia oxida ion (bac e ial and a chaeal amoA) and deni ifica ion (ni S and nosZ). Dissol ed N 2 O a ied up o h ee o de s o magni ude (4.7–2441.2 nmol L 1 ) ac oss sys ems, om unde sa u a ed o supe sa u a ed alues (37%–24,174%) depending on ese oi s and dep hs. N 2 O concen a ion depended on ni ogen and oxygen a ailabili ies, wi h he lowes and highes N 2 O alues a suboxic condi ions. Ammonia-oxidizing a chaea domina ed o e ammonia-oxidizing bac e ia bu we e no ela ed o he dissol ed N 2 O. In con as , he abundance o he ni S gene was significan ly ela ed o N 2 O concen a ion, and h ee o de s o magni ude highe han amoA abundance. Deni i ying bac e ia appea ed consis en ly in he wa e column o all ese oi s. The ni S and nosZ genesappea edinoxicandsuboxicwa e s,bu heywe e mo e abundan in suboxic wa e s. The ni a e concen a ion, and ni S and nosZ ela i e abundances explained he dissol ed N 2 O. Besides, ni S abundance was ela ed posi i ely wi h o al phospho us and cumula i e chlo ophyll a, a p oxy o esh o ganic ma e . The e o e, P inpu s, no jus N inpu s, p omo ed N 2 O p oduc ion by deni ifica ion in he wa e column o ese oi s. The an h opogenic p oduc ion o e ilize has doubled he inpu s o ni ogen (N) o he Ea h’s su ace, inc easing hei expo s o eshwa e s and boos ing he p oduc ion o ni ous oxide (N 2 O) (e.g., Sei zinge e al. 2000; Beaulieu e al. 2011). N 2 O is a po en g eenhouse gas wi h 298 imes he wa ming e ec o CO 2 in a 100-y ime ho izon (IPCC 2013) and he p ima y d i e o s a osphe ic ozone deple ion (Ra ishanka a e al. 2009). The s udies on N 2 O emissions in eshwa e s ha e mainly ocused on s eams and i e s (Beaulieu e al. 2015), wi h an es ima ed N 2 O emission o 0.68 Tg N 2 O-N y 1 (Beaulieu e al. 2011). Ne e heless, he longe wa e esidence ime o len ic sys ems such as lakes and ese oi s esul s in subs an ial N p ocessing (We zel 2001). The N 2 O emission om lakes and ese oi s was ecen ly es ima ed a 0.3 Tg N 2 O-N y 1 based on a sca ce da ase (DelSon o e al. 2018). Rese oi s a e enla ged wa e bodies behind a dam, and hei numbe has inc eased significan ly o e he pas 60 yea s, eaching o e 16.7 million globally (Lehne e al. 2011). In spi e o his, ese oi s a e pa icula ly unde - s udied, al hough hey p ocess a disp opo iona ely high ac- ion o N om he ca chmen in compa ison o lakes due o hei highe ca chmen a ea, and highe d ainage a io (i.e., he ca chmen a ea: lake o ese oi su ace a ea) (Ha ison e al. 2009). They may suppo excep ionally high N 2 Oemis- sions due o subs an ial N loadings om he ag icul u al and u ban a eas in hei wa e sheds, exceeding punc ually he cli- ma ic o cing p oduced by CH 4 emissions (e.g., Izn aja ese oi *Co espondence: eleonpalm[email p o ec ed];[email p o ec ed] This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion-NonComme cial License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed and is no used o comme cial pu poses. Addi ional Suppo ing In o ma ion may be ound in he online e sion o his a icle. a P esen add ess: No dcee, Depa men o Biology, Uni e si y o Sou he n Denma k, Odense, Denma k Au ho Con ibu ion S a emen : E.L.-P., R.M.-B., and I.R. con ibu ed o da a acquisi ion du ing he ese oi samplings. E.L.-P. analyzed he samples and p ocessed he da a. E.L.-P., R.M.-B., and I.R. analyzed he da a and discussed he esul s. E.L.-P. w o e he fi s d a manusc ip , which was complemen ed by significan con ibu ions o R.M.-B. and I.R. I.R. and R.M.B. designed he s udy and ob ained he unds. All au ho s ha e gi en app o al o he final e sion o he manusc ip . 1 in Le on-Palme o e al. 2020a). The e o e, a be e unde s and- ing o he con olling ac o s ha lead o N 2 Op oduc ionand consump ion in he wa e column o ese oi s is needed, espe- cially ega ding he inc easing end in ese oi cons uc ion a a global scale (Za fle al. 2015). Mic obial ni ifica ion and deni ifica ion p oduce N 2 O. Ni ifica ion consis s o he oxida ion o ammonia o ni i e (i.e., ammonia oxida ion) and o ni a e (i.e., ni i e oxida ion). Ammonia oxida ion is he a e-limi ing s ep, and i is pe o med by ammonia-oxidizing bac e ia (AOB) and ammonia-oxidizing a chaea (AOA) (Kowalchuk and S e- phen 2001; Könneke e al. 2005). The global significance o his p ocess and he ela i e con ibu ion o AOB and AOA ha e been deduced om he abundance o he bac e ial and he a chaeal amoA genes, which encode he subuni A o he key enzyme ammonia monooxygenase (Kowalchuk and S e- phen 2001; F ancis e al. 2005). N 2 O is a by-p oduc o AOB and AOA in oxic condi ions h ough di e en pa hways (Wa d 2013a; S ein 2019). While N 2 O p oduc ion in AOB occu s h ough enzyma ic ac i i y ia he wo obliga e in e - media es hyd oxylamine (NH 2 OH) and ni ic oxide (NO) (Ca an o and Lancas e 2017); he N 2 O o ma ion in AOA seems o be a hyb id p ocess, whe e NO eac s wi h NH 2 OH (Kozlowski e al. 2016), al hough specific de ails pe - sis con o e sial. A low oxygen condi ions, AOB also pe - o ms ni ifie deni ifica ion, which consis s o educing ni i e o NO and N 2 O, inc easing he yield o N 2 O p oduced ela i e o he ammonia oxidized (W age e al. 2001; Wa d 2013a). Fo bo h cases, N 2 O p oduc ion a es eached maxima a suboxic condi ions (< 10 μmol L 1 ) (Hink e al. 2017). AOA a e p esen in ma ine wa e s and sedimen s in la ge numbe s, con ibu ing significan ly o he N 2 O p o- duc ion (e.g., F ancis e al. 2005; Lösche e al. 2012). How- e e , he con ibu ion o ammonia-oxidizing mic oo ganisms o dissol ed N 2 O in ese oi s has no been s udied. Deni ifica ion consis s o he subsequen educ ion o ni a e o ni i e, NO, N 2 O, and dini ogen (N 2 ), gene ally coupled o he oxida ion o o ganic ma e . I can be a sou ce o a sink o N 2 O depending on he a e o N 2 O p oduc ion compa ed o he a e o N 2 O educ ion o N 2 . The genes ha code he ni i e educ ases (i.e., ni S, ni K) du ing deni ifica- ion a e widely used o in e he abundance and he con ibu- ion o he deni i ying bac e ia o he dissol ed N 2 O budge , while he gene ha code he ni ous oxide educ ase (i.e., nosZ) is used o add ess he capaci y o educing N 2 O o N 2 (Hallin e al. 2018). Besides, deni ifica ion is usually con- side ed as a acul a i e anae obic espi a ion because oxygen egula es he sequence o he deni ifica ion enzymes, espe- cially he ni ous oxide educ ase coded by nosZ, which is inhibi ed e en a e y low oxygen concen a ions (Bonin e al. 1989; Dalsgaa d e al. 2014). Consequen ly, he s udies on deni ifica ion in eshwa e s ha e ocused on anoxic wa e s, pa icula ly, on sedimen s (Piña-Ochoa and ´ Al a ez- Cobelas 2006). Howe e , deni ifie s a e also p esen in he wa e column o lakes (Junie e al. 2008; Kim e al. 2011; Paja es e al. 2017), and he influence o O 2 concen a ion on he deni i ying ac i i y appea s o di e om one bac e ium o ano he (Lloyd 1993). Lakes and ese oi s can be essen ial playe s in he N emo al a he landscape scale, ac ing as N 2 O sou ces in a eas subjec ed o high N inpu s (McC ackin and Else 2011)o N 2 O sinks as, o ins ance, in bo eal lakes (Soued e al. 2015). Howe e , only some s udies ha e ocused on he ole o ese - oi s as sinks o sou ces o N 2 O, specially ega ding hei wa e columns. He e, we quan ified he dissol ed N 2 O concen a ion du ing he summe s a ifica ion and he win e mixing in he wa e column o 12 empe a e Medi e anean ese oi s co e - ing a b oad spec um o ages, mo phome ies, chemical cha - ac e is ics, and wa e shed land uses. We examined he main d i e s o N 2 O concen a ion in he wa e column. We also s udied he ela ionship be ween he ni ifie s and deni ifie s abundance and he N 2 O concen a ion. Finally, we connec ed he landscape and he chemical and biological p ope ies o p o ide an ecosys emic pe spec i e o N 2 O cycling in ese oi s. Ma e ial and me hods S udy ese oi s, mo phome y, and wa e shed land uses We sampled he wa e column o 12 ese oi s loca ed in sou he n Spain be ween July 2016 and Augus 2017, once du - ing summe s a ifica ion and once du ing win e mixing. The loca ion o hese ese oi s is shown in Suppo ing In o ma- ion Fig. S1a,b. The s udy ese oi s we e buil be ween 1932 and 2003 o wa e supply and ag icul u e i iga ion, and hey di e in mo phome y, wa e chemis y, and ophic s a us. In Supplemen a y Table S1, we showed he geog aphical coo di- na es, age, and he mo phome ic desc ip ion o he s udy es- e oi s. They a e in wa e sheds wi h con as ing land uses, anging om o es ed landscape (e.g., San Clemen e ese oi in Suppo ing In o ma ion Fig. S1c) o ag icul u al and u ban landscape (e.g., Izn aja ese oi , in Suppo ing In o ma ion Fig. S1d). We p o ided mo e in o ma ion and de ailed maps in Le on-Palme o e al. (2019,2020a,b,c,d). Wa e column sampling Ve ical p ofiles and wa e column sampling we e pe - o med nea he dam, in he open wa e s o he ese oi , selec ing he same loca ion du ing he s a ifica ion and he mixing pe iod. Fi s , we pe o med he e ical p ofile using a Sea-Bi d 19plus CTD p ofile . Then, based on he empe a u e and oxygen p ofiles, we selec ed six o nine dep hs ep- esen ing he epilimnion, me alimnion (oxycline), and hypo- limnion. We ook he wa e samples a hese dep hs using a 5-l UWITEC sampling bo le o he chemical and biological analysis explained below. We selec ed 178 dep hs in o al, 96 dep hs du ing s a ifica ion and 82 dep hs du ing mixing. F om hese samples, he concen a ions o he dissol ed N 2 O Le on-Palme o e al. N 2 O in he wa e column o ese oi s 2 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License and chlo ophyll a(Chl a), and he abundance o p oka yo es we e analyzed. Fo he analysis o he majo nu ien s and unc ional genes, we selec ed h ee o ou ep esen a i e dep hs o he epilimnion, me alimnion (oxycline), and hypo- limnion, o bo om wa e s du ing he s a ifica ion pe iod (n=41), and h ee o ou equi alen dep hs du ing he mixing pe iod (n=36). Dissol ed N 2 O in he wa e column Samples o he dissol ed N 2 O analysis we e ca e ully col- lec ed in 125 o 250 mL ai igh Winkle bo les in duplica e (250 mL) o in iplica e (125 mL), p ese ed wi h a solu ion o HgCl 2 (final concen a ion 1 mmol L 1 ) o inhibi biologi- cal ac i i y, and sealed wi h Apiezon ® g ease o p e en gas exchange. We s o ed hese samples in he da k a con olled empe a u e (25C) un il analysis. We measu ed dissol ed N 2 O using headspace equilib a ion in a 50 mL ai igh glass sy inge (Agilen P/N 5190–1547; Sie a e al. 2017). We ob ained wo eplica es o each 125 mL Winkle bo le and h ee eplica es o each 250 mL Winkle bo le. We ook a quan i y o 25 g o wa e (0.01 g) using he ai igh sy inge and added a quan i y o 25 mL o a s anda d gas mix u e ha had a N 2 O concen a ion simila o a mosphe ic alues (0.3 ppm ) o comple e he olume o he sy inge. The sy inges we e shaken o 5 min (Vib oma ic, Selec a) o ensu e mixing, and we wai ed 5 min o each comple e equilib ium. Then, he gas in he sy inge was injec ed manually in o he gas ch o- ma og aph (GC; B uke ® GC-450). We daily calib a ed he de ec o s using h ee s anda d gas mix u es wi h N 2 O concen- a ions o 305, 474, 2000 ppb , made and ce ified by Ai Liquide (F ance). The p ecision o he gas mix u e o N 2 O used in he headspace equilib ium (0.3 ppm ) was 7.8% (n=108). We calcula ed he gas concen a ion in equilib ium using he Bunsen solubili ies o N 2 O (Weiss and P ice 1980). The p ecision o he measu emen o he dissol ed N 2 O con- cen a ion, ha included he analy ical p ocessing o he sam- ples and he equilib a ion s ep, was 3.1% (N 2 O) o ou o six eplica es o each sample. We calcula ed he sa u a ion alues (%) as he a io be ween he concen a ion o he dissol ed gas measu ed and he gas concen a ion expec ed in equilib- ium conside ing he empe a u e, salini y, and ba ome ic p essu e o each ese oi . Chl aand majo nu ien analysis in he wa e column Wa e samples o chemical and biological analysis we e main ained a 4C un il a i al a he labo a o y. We de e - mined Chl aconcen a ion by fil e ing he pa icula e ma e ial o 500 o 2000 mL o wa e h ough 0.7 μm po e-size Wha man GF/F glass-fibe fil e s. Then, we ex ac ed he pig- men s om he fil e s wi h 95% me hanol in he da k a 4C o 24 h (APHA 1992). We measu ed Chl aabso p ion a he wa eleng h o 665 nm using a Pe kin Elme UV-Lambda 40 spec opho ome e , and we co ec ed he solu ion sca e ing a 750 nm. To ob ain he cumula i e Chl ain he whole wa e column (mg Chl am 2 ), om he disc e e dep hs, we added he concen a ion o Chl a om each s a um using he apezoidal ule as desc ibed p e iously (Le on-Palme o e al. 2020b). We used he fil e ed wa e o de e mine he concen a ions o dissol ed nu ien s, and he unfil e ed wa e o o al nu i- en s. We acidified he samples o dissol ed o ganic ca bon (DOC), o al dissol ed ni ogen (TDN), and o al ni ogen (TN) concen a ion wi h phospho ic acid (final pH < 2). We measu ed DOC, TN, and TDN by high– empe a u e ca aly ic oxida ion using a Shimadzu o al o ganic ca bon analyze (Model TOC-V CSH) coupled o ni ogen analyze (TNM-1; ´ Al a ez-Salgado and Mille 1998). We measu ed he ni a e (NO 3 ) concen a ion using he ul a iole spec opho ome - ic me hod wi h a Pe kin Elme UV-Lambda 40 spec opho- ome e a he wa eleng h o 220 nm and including co ec ions o DOC abso bance a 275 nm (APHA 1992). We measu ed he ni i e (NO 2 ) concen a ion by induc i ely coupled plasma op ical emission spec ome y. In his wo k, we calcula ed he dissol ed ino ganic ni ogen as he addi ion o he NO 3 and NO 2 concen a ions. We measu ed o al phospho us (TP) concen a ion by iplica e using he molyb- denum blue me hod a e diges ion wi h a mix u e o po as- sium pe sul a e and bo ic acid a 120C o 30 min (APHA 1992). Abundance o p oka yo es The abundances o o al p oka yo es we e de e mined using flow cy ome y in unfil e ed wa e ollowing he p ocedu es p oposed by Gasol and del Gio gio (2000). We collec ed and fixed he samples wi h a mix u e o 1% pa a o maldehyde and 0.05% glu a aldehyde o 30 min in he da k a 4C. Then, we oze he samples in liquid ni ogen and s o ed hem a 80C un il analysis. We analyzed he samples in iplica e in a FACScalibu flow cy ome e equipped wi h he BD CellQues P o so wa e o da a analysis. Be o e analysis, we s ained he samples (500 μL) o 10 min in he da k wi h a DMSO dilu ed SYBR G een I (Molecula P obes) s ock (1 : 200) a 10 μmol L 1 final concen a ion. We used yellow–g een 0.92 μm la ex beads (Polysciences) as an in e nal s anda d o con ol he cy ome e pe o mance e e y day. Samples we e un a low speed o 2 min and de ec ed by hei signa u e in bi a ia e plo s SSD s. FL1 (g een fluo escence o he DNA s ained wi h SYBR G een I). Quan i a i e PCR (qPCR) assays We quan ified he abundance o unique unc ional genes in ol ed in N 2 O cycling using qPCR. We also calcula ed he a io be ween he abundances o he specific genes (copies mL 1 ) and o al p oka yo es (cell mL 1 ) o dis inguish he e ec o di e en chemical and biological d i e s on he o al abundance o p oka yo es s. he abundance o he specific g oups. We a ge ed he amoA gene, which encodes he ca a- ly ic subuni o ammonia monooxygenase (Kowalchuk and Le on-Palme o e al. N 2 O in he wa e column o ese oi s 3 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License S ephen 2001), o s udy he AOB (bac-amoA), and he AOA (a ch-amoA). To s udy he deni ifie abundance, we a ge ed he ni S and he nosZ genes. The ni S gene encodes he ni i e educ ase ha ca alyzes he ans o ma ion o ni i e o NO, while he nosZ gene encodes he ni ous oxide educ ase, ha educes N 2 O oN 2 . We chose specific p ime s om p e ious s udies pe o med in eshwa e s. Pu e cul u es we e used as posi i e con ols in PCR, and o qPCR s anda d p epa a ion. P ime s, amplicon leng h, annealing empe a u e, and pu e cul u es a e shown in Supplemen a y Table S2. We p o ide u he de ails on he DNA ex ac ion and qPCR p ocedu e in he Suppo ing In o ma ion (Ex ended Ma e ials and me hods). S a is ical es s and so wa e We used linea eg ession analysis and gene alized addi i e models (GAMs) (Wood 2006). We examined he concu i y among p edic o s and fi ed he models o minimize he Akaike in o ma ion c i e ion and he gene alized c oss alida- ion c i e ion alues. We conduc ed all he s a is ical analysis in R so wa e (R Co e Team 2019). We also used R and Inkscape™(Inkscape P ojec 2017) o plo he esul s and c e- a e he figu es. We p o ide u he de ails on he s a is ical analysis pe o med in he Suppo ing In o ma ion (ex ended ma e ials and me hods). Resul s and discussion P ofile desc ip ion: N 2 O sinks s. sou ces We ound p ominen changes in he dissol ed N 2 O con- cen a ion among ese oi s, dep hs, and seasons (Figs. 1, 2, Suppo ing In o ma ion Figs. S2–S11 and summa ized in Sup- plemen a y Table S3). The N 2 O concen a ion and % o sa u a- ion anged up o h ee o de s o magni ude om 4.7 o 2441.2 nmol L 1 , and om 37% o 24,174%. To acili a e he esul p esen a ion, we g ouped he ese oi s as sinks i he sum o he N 2 Ofluxes du ing he summe s a ifica ion and win e mixing was ≤0; o sou ces i he sum was > 0 acco ding o ou p e ious s udy on N 2 Ofluxes in hese 12 ese oi s (Le on-Palme o e al. 2020a). The eigh ese oi s ha ac ed as N 2 O sinks (he ea e , sinks) we e: San Clemen e, La Bole a, F ancisco Abell an, J andula, Neg a ín, El Po illo, Rules, and Los Be mejales (Fig. 1; Suppo ing In o ma ion Figs. S2–S8). The ou ese oi s ha ac ed as sou ces o N 2 O (he ea e , sou ces) we e: Izn aja , Bézna , Cubillas, and Colome a (Fig. 2; Suppo ing In o ma ion Figs. S9–S11). N 2 O concen a ion in he wa e column o he sinks a ied om 4.7 o 46.1 nmol L 1 (median =12.3 nmol L 1 ), and om 12.6 o 2441.2 nmol L 1 in he sou ces (median =20.3 nmol L 1 ). While he sou ces we e always supe sa u a ed in N 2 O (109%–24,174%), he sinks showed supe sa u a ion and unde sa u a ion a di e en dep hs o he same p ofile (37%–366%). The median concen a ion and % o N 2 O in he sinks we e lowe han in he sou ces (Suppo ing In o ma ion Fig. S12). We de ec ed he minimum alue o N 2 O in he hypolimnion o San Clemen e (4.7 nmol L 1 , 37%, Fig. 1a), and he maximum alue in he hypolimnion o Izn aja (2441.2 nmol L 1 , 24,174%, Fig. 2a); bo h o hem a suboxic condi ions (i.e., DO < 10 μmol L 1 ). These wo ese oi s also showed he minimum and maximum N 2 Ofluxes in Le on- Palme o e al. (2020a). In addi ion, we also obse ed N 2 O unde sa u a ion a o he dep hs o he wa e column in oxic condi ions, as in La Bole a, Neg a ín, and El Po illo ese oi s (Suppo ing In o ma ion Figs. S2b, S5b, S6b). In San Clemen e, La Bole a, F ancisco Abell an, and J andula ese oi s (sinks), we ound a N 2 O peak abo e he oxycline (Fig. 1a; Suppo ing In o ma ion Figs. S2–4a), and a dec ease in N 2 O o unde sa u a ed alues below he oxycline, eaching 37% in San Clemen e ese oi , and 84% in La Bole a ese oi . In bo h sys ems, we de ec ed high abundances o he nosZ gene ha could explain he ne consump ion o N 2 O du ing he las s ep o deni ifica ion (i.e., he con e sion o N 2 O oN 2 ). In con as , N 2 O inc eased below he oxycline in ese oi s ac ing as sou ces, leading o massi e accumula ions o N 2 O (Fig. 2a; Suppo ing In o ma ion Figs. S9a–S11a). In Izn aja , Bézna , and Cubillas ese oi s we de ec ed high abundances o a chaeal amoA gene and ni S gene a hese si es, sugges ing ha he accumula ion o N 2 O may be p oduced by ammonia oxidize s o deni ifie s (Fig. 2a; Suppo ing In o ma ion Figs. S9a, S10a). Bo h p ocesses ha e he highes yields o N 2 O a low oxygen concen a ions. We show he dis ibu ion o N 2 O, dissol ed oxygen (DO), and o he chemical and biologi- cal a iables in Supplemen a y Tables S3–S5 and in Suppo ing In o ma ion Fig. S12. The wide ange in N 2 O concen a ion and % o sa u a ion ound in his s udy co e s alues epo ed in empe a e and sub opical ese oi s (Deeme e al. 2011; Liang e al. 2019), and i is b oade han he a iabili y ound in alpine and sub- opical ese oi s (Diem e al. 2012; Musenze e al. 2014). The maximum N 2 O concen a ion ound in Izn aja ese oi was highe han he maximum alues de ec ed in su ace and deep wa e s in o he s udies (Diem e al. 2012; Musenze e al. 2014; Beaulieu e al. 2015; Liang e al. 2019). P e ious s udies also de ec ed peaks o N 2 O a he oxic–anoxic in e ace in s a ified lakes and ese oi s (Beaulieu e al. 2015), and la ge accumula- ions o N 2 O below he oxycline o ese oi s (Deeme e al. 2011; Beaulieu e al. 2015). In a su ey o 20 ese oi s, Beaulieu e al. (2015) also ound ha some ese oi s we e supe sa u a ed in N 2 O, while o he s p esen ed unde sa u a ion alues. Biogeochemical con ol on N 2 O concen a ion N 2 O concen a ion was a posi i e powe unc ion o he ni ogen con en in di e en chemical o ms in he wa e col- umn (Supplemen a y Table S6). In pa icula , TN o ni a e concen a ion explained 43% o he a iance in he N 2 O con- cen a ion (Suppo ing In o ma ion Fig. S13). Besides, N 2 O was also a nega i e powe unc ion o he oxygen a ailabili y Le on-Palme o e al. N 2 O in he wa e column o ese oi s 4 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License (Suppo ing In o ma ion Fig. S14). Combined in a GAM, ni a e and DO concen a ions explained 64% o he a iance in he N 2 O concen a ion, and up o 81% o he a iance when he in e ac ion be ween bo h was included in he model (Fig. 3; Supplemen a y Table S7). The pa ial esponse o N 2 O wi h espec o ni a e concen a ion shows a posi i e unc- ion (Fig. 3b), bu nega i e wi h espec o DO concen a ion (Fig. 3c). Mo eo e , he model in Fig. 3a shows he in e ac ion Dep h (m) (a) San Clemen e (s a i ica ion pe iod) nosZ (x103 copies mL-1) Dep h (m) (b) San Clemen e (mixing pe iod) 15 25 15 25 (μmol L-1) bac- amoA a ch- amoA ni S nosZ Chl aTP Chl aTP TN TDN NO3- NO2- TN TDN NO3- NO2- N2O (nmol L-1)N2O (%) N2O sa u a ion (%) N2O sa u a ion (%) Tempe a u e (ºC) Chl a (μg L-1) Dissol ed O2 (μmol L-1) Dissol ed N2O (nmol L-1) TN (μmol-N L-1) TP (μmol-P L-1) TP (μmol-P L-1) N2O (nmol L-1)N2O (%) Tempe a u e (ºC) Dissol ed O2 (μmol L-1) Dissol ed N2O (nmol L-1) TN (μmol-N L-1) Chl a (μg L-1) nosZ (x103 copies mL-1) bac- amoA a ch- amoA ni S nosZ amoA (x103 copies mL-1) ni S (x103 copies mL-1) amoA (x103 copies mL-1) ni S (x103 copies mL-1) Fig. 1. Ve ical p ofiles o physicochemical and biological a iables in San Clemen e ese oi . N 2 O concen a ion (nmol L 1 , mean s anda d e o , ci - cles), N 2 O sa u a ion (%, mean s anda d e o , diamonds), and a mosphe ic equilib ium concen a ion (discon inuous line); wa e empe a u e (C); DO concen a ion (μmol L 1 ); Chl aconcen a ion (μgL 1 ) and TP concen a ion (μmol-P L 1 ); TN concen a ion (μmol-N L 1 ); abundance o he amoA genes (bac e ial amoA and a chaeal amoA,10 3 copies mL 1 , mean s anda d de ia ion); and abundance o he ni S and nosZ genes (10 3 copies mL 1 , mean s anda d de ia ion) du ing he s a ifica ion pe iod (a) and he mixing pe iod (b). No e ha he gene abundance axes a e in loga i hmic and di e en scale. The g ay a ea ep esen s he suboxic zone (DO < 10 μmol L 1 ). Le on-Palme o e al. N 2 O in he wa e column o ese oi s 5 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License be ween ni a e and DO concen a ions. N 2 Owasconsumed a low oxygen condi ions when he ni ogen concen a ion was lowe (i.e., sinks), bu N 2 O p oduc ion inc eased a low oxygen condi ions when he ni ogen concen a ion was highe (i.e., sou ces). The eigh ese oi s ha expe ienced suboxic condi ions (DO < 10 μmol L 1 ) in hei hypolim- nions du ing he summe s a ifica ion showed his end: ou o hem we e classified as sinks (i.e., Fig. 1a; Dep h (m) (a) Iznája (s a i ica ion pe iod) Dep h (m) (b) Iznája (mixing pe iod) amoA (x103 copies mL-1) ni S (x103 copies mL-1) amoA (x103 copies mL-1) ni S (x103 copies mL-1) N2O (nmol L-1)N2O (%) (μmol L-1) N2O sa u a ion (%) N2O sa u a ion (%) TN TDN NO3- NO2- TN TDN NO3- NO2- Tempe a u e (ºC) Chl a (μg L-1) Dissol ed O2 (μmol L-1) Dissol ed N2O (nmol L-1) TN (μmol-N L-1) N2O (nmol L-1)N2O (%) Tempe a u e (ºC) Chl a (μg L-1) Dissol ed O2 (μmol L-1) Dissol ed N2O (nmol L-1) TN (μmol-N L-1) TP (μmol-P L-1) nosZ (x103 copies mL-1) bac- amoA a ch- amoA ni S nosZ TP (μmol-P L-1) nosZ (x103 copies mL-1) bac- amoA a ch- amoA ni S nosZ Chl aTP Chl aTP Fig. 2. Ve ical p ofiles o physicochemical and biological a iables in Izn aja ese oi . N 2 O concen a ion (nmol L 1 , mean s anda d e o , ci cles), N 2 O sa u a ion (%, mean s anda d e o , diamonds), and a mosphe ic equilib ium concen a ion (discon inuous line); wa e empe a u e (C); DO concen a ion (μmol L 1 ); Chl aconcen a ion (μgL 1 ) and TP concen a ion (μmol-P L 1 ); TN concen a ion (μmol-N L 1 ); abundance o he amoA genes (bac e ial amoA and a chaeal amoA,10 3 copies mL 1 , mean s anda d de ia ion); and abundance o he ni S and nosZ genes (10 3 copies mL 1 , mean s anda d de ia ion) du ing he s a ifica ion pe iod (a) and he mixing pe iod (b). No e ha he N 2 O, and he gene abundance axes a e in loga i hmic and di e en scale. The g ay a ea ep esen s he suboxic zone (DO < 10 μmol L 1 ). Le on-Palme o e al. N 2 O in he wa e column o ese oi s 6 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License Suppo ing In o ma ion Figs. S2–4a) and ou as sou ces (Fig. 2a; Suppo ing In o ma ion Figs. S9–S11a). Ni ogen and oxygen a ailabili ies con ol dissol ed N 2 Oin he wa e column because hey a ec he mic obial p ocesses ha de e mine he p oduc ion and consump ion o N 2 O. These mic obial p ocesses will be u he discussed in he nex sec ion. P e ious s udies demons a ed ha N 2 Ois closely linked o N a ailabili y in eshwa e s (e.g., Sei zinge e al. 2000; Beaulieu e al. 2011,2015; Zhou e al. 2021). Beaulieu e al. (2015) ound a posi i e ela ionship be ween mean epilimnion N 2 O concen a ion and bo h NO 2 and NO 3 concen a ions in ese oi s. In addi ion, hey also ound a ne consump ion o N 2 O in he hypolimnions wi h low ni ogen con en , bu a ne p oduc ion in he hypolim- nions wi h highe ni ogen con en . Zhou e al. (2021) de ec ed a posi i e ela ionship be ween he N a ailabili y (TN and ni a e) and N 2 Ofluxes in shallow lake sedimen s. In o he aqua ic en i onmen s, as he Chesapeake Bay, he p o- duc ion o N 2 O was also con olled by ni ogen and oxygen a ailabili ies (Ji e al. 2018). Mic obial g oups a ec ing N 2 O balance Dis ibu ion and d i e s o he mic obial g oups in he wa e column The abundances o he bac e ial and a chaeal amoA (i.e., AOB and AOA), and he ni S and nosZ (i.e., deni i ying bac e ia) genes a e shown in Figs. 1, 2, Suppo ing In o ma- ion Figs. S2–S11 and summa ized in Supplemen a y Table S5. The a chaeal amoA gene appea ed in all he s udy ese oi s, bu no in all he dep hs (i.e., 56 ou o 77 samples), while he bac e ial amoA gene appea ed only in ou samples. The median abundance o he a chaeal amoA gene was 301 copies mL 1 , anging om 0 o 1.9 10 4 copies mL 1 . The maxi- mum abundances we e in Los Be mejales, Neg a ín, and La Bole a ese oi s du ing he win e mixing (Suppo ing In o - ma ion Figs. S8b, S5b, S2b). A chaeal amoA abundance did no show a significan ela ionship o he DO (n=77, p=0.147, Suppo ing In o ma ion Fig. S15a), o he ni i e concen a- ion (n=77, p=0.484, Suppo ing In o ma ion Fig. S16). Tha may sugges ha ammonium oxida ion is coupled o ni i e oxida ion. P e ious s udies in empe a e lakes also ound ha AOA domina ed o e AOB (Small e al. 2013; Palacin-Liza be e al. 2019), while Paja es e al. (2017) ound simila abundances o bo h g oups in a opical oligo ophic lake, wi h an abundance o he a chaeal amoA gene simila o ou s udy. Recen in es iga ions poin ed ou ha AOA may domina e o e AOB in oligo ophic en i onmen s wi h a low ammonia supply (i.e., ocean; Ma ens-Habbena e al. 2009). The ni S gene was ubiqui ous, appea ing in conside able abundances in he wa e column o all he ese oi s du ing bo h pe iods. We de ec ed he maximum abundances o he ni S gene in Izn aja , Bézna , and Cubillas ese oi s du ing he s a ifica ion pe iod oge he o high concen a ions o N 2 O (Fig. 2a; Suppo ing In o ma ion Figs. S9a, S10a). The median abundance o he ni S gene was 3.2 10 5 copies mL 1 , ang- ing om 0 o 1.1 10 8 copies mL 1 . We did no de ec he ni S gene only in 8 samples ou o 77, and we did no include hese samples in he s a is ical es s since hey we e s a is ical 2.5 2.0 1.5 1.0 0.5 0.0 0.25.20.15.1 Log10(Ni a e , μmol-N L-1) Log10(DO +1, μmol L-1) 0.0 -0.5 -1.0 1.0 0.5 (a) (b) 0.0 -0.5 -1.0 1.0 0.5 0.0 (c) Log10(Ni a e , μmol-N L-1) Log10(DO +1, μmol L-1) s(log10(Ni a e)) s(log10(DO +1)) Fig. 3. GAM fi ed o N 2 O concen a ion (log 10 N 2 O, nmol L 1 ), as unc ion o ni a e (log 10 NO 3  ,μmol-N L 1 ), and DO concen a ion (log 10 DO, μmol L 1 ). (a) Con ou plo showing he ela ionship be ween ni a e (x-axis) and DO (y-axis) wi h N 2 O concen a ion (z-axis, con ou lines). The model includes he in e ac ion be ween ni a e and DO concen a ions. Pa ial esponse plo s showing he pa ial e ec s o (b) ni a e and (c)DOonN 2 O con- cen a ion. Rugs on x-axis a e he obse ed da a poin s. The lines a e he smoo hing unc ions, and he shaded a eas indica e he 95% confidence in e - als. Mo e de ails in Supplemen a y Table S6. Le on-Palme o e al. N 2 O in he wa e column o ese oi s 7 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License ou lie s (G=0.25, p< 0.001). Besides, he ni S abundance was a nega i e powe unc ion o he DO concen a ion (n=69, p< 0.001; Suppo ing In o ma ion Fig. S15b). The concen a ion o ni a e, which is he main subs a e o deni- ifica ion, was no significan ly ela ed o he abundance o ni S gene, o al p oka yo es, o he ni S : p oka yo es a io (Supplemen a y Table S8; Suppo ing In o ma ion Fig. S17a–c). Howe e , he ni S gene, and he ni S : p oka yo es a io we e nega i ely ela ed o he ni i e concen a ion when he con- cen a ion was o e de ec ion le el (Suppo ing In o ma ion Fig. S17e, ). We also ound ha TP concen a ion and cumula- i e Chl a, which is a p oxy o he o al phy oplank onic bio- mass expo ed om he wa e column, we e posi i e and significan ly ela ed o he abundance o he ni S gene, and, specifically, he ni S : p oka yo es a io, bu hey we e no ela ed o he o al abundance o p oka yo es (Fig. 4). The e- o e, he e ec s o TP and cumula i e Chl aa e specific on he deni i ying bac e ia (i.e., ni S gene). Toge he , DO and TP concen a ions explained he abundance o deni ifie s in he wa e column o hese ese oi s (n=69, adj R 2 =0.68, Suppo ing In o ma ion Fig. S18; Supplemen a y Table S9), showing he highes ni S abundances a low DO and high TP concen a ions (Suppo ing In o ma ion Fig. S18a). On he o he hand, we de ec ed he nosZ gene in 37 ou o 77 samples analyzed in oxic and suboxic condi ions (Suppo ing In o ma- ion Fig. S15c). The abundance o he nosZ gene anged om 0 o 3.1 10 4 copies mL 1 (median =0 copies mL 1 ). The maximum abundances o he nosZ gene we e in he hypolim- nia o San Clemen e and La Bole a ese oi s a suboxic condi- ions, whe e we also de ec ed he lowes N 2 O sa u a ion alues (i.e., 37%, and 86%; Fig. 1a; Suppo ing In o ma ion Fig. S2a). The abundance o ni S gene is a bioma ke o he den- i i ying communi y, while he nosZ gene is a p oxy o he bac e ia ha educes N 2 O oN 2 , because his gene code he ni ous oxide educ ase (Hallin e al. 2018). Ou esul s indi- ca e ha deni i ying bac e ia (i.e., ni S gene) a e ubiqui ous and e y abundan in he wa e column o he s udy ese - oi s. Besides, ni S abundance was 3 o de s o magni ude highe (median =3.2 10 5 copies mL 1 ) han he abundance o he amoA gene (median =301 copies mL 1 ), and he nosZ gene (median =0 copies mL 1 ). ni S gene codes he ni i e educ ase ha educe ni i e du ing deni ifica ion, and ha may explain i s ela ionship wi h he ni i e concen a ion. In his s udy we a ge ed only ni S- ype deni ifie s, howe e ni K- ype deni ifie s may be also p esen in hese sys ems. Fig. 4. E ec o he concen a ions o (a–c)TP(μmol-P L 1 ) and (d– ) cumula i e Chl a(mg m 2 ) on he abundance o (a, d) p oka yo es (cell mL 1 ), (b, e) ni S gene (copies mL 1 ), and he (c, ) ni S : p oka yo es a io (ni S copies: p oka yo ic cells). The b own do s s and o he ese oi s ac ing as sinks, while he g een do s s and o he ese oi s ac ing as sou ces o N 2 O. No e he log scales. Le on-Palme o e al. N 2 O in he wa e column o ese oi s 8 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License Few s udies ha e analyzed he occu ence o dis ibu ion o he ni S,ni K,o nosZ genes in he wa e column o lakes o ese oi s (Junie e al. 2008; Kim e al. 2011; Paja es e al. 2017; Mao e al. 2017). Junie e al. (2008) and Kim e al. (2011) de ec ed he ni S gene in he wa e column o lakes, wi h he highe di e si y in he epilimnion, and di e en den- i i ying communi ies in he wa e column and he sedimen s. The maximum abundance de ec ed in ou s udy was up o 3 o de s o magni ude highe han he maximum de ec ed by Paja es e al. (2017) in a opical lake, who also desc ibed a simila abundance o he ni K gene, bu he ni S abundance in ou s udy was simila o he abundance de ec ed by Mao e al. (2017). Mao e al. (2017) also de ec ed he nosZ gene in he wa e column o lakes, wi h an abundance se e al o de s o magni ude highe han he abundance de ec ed in ou wo k. In he s udy ese oi s, he occu ence o he gene nosZ was mo e limi ed han he occu ence o he ni S gene, and we only de ec ed nosZ in abou hal he samples analyzed. This finding may eflec a eal limi a ion in he dis ibu ion o he nosZ gene in he s udy ese oi s. Howe e , his could also be due o a me hodological limi a ion, as we may no ha e de ec ed all nosZ gene a ian s by using p ime s ha may only cap u e ypical nosZ a ian s (Clade I). Ni ous oxide educ- ases educing N 2 O oN 2 a e no always a filia ed wi h den- i i ying mic oo ganisms. Recen findings showed ha some o ganisms possess a Clade II (a ypical) nosZ gene, and his a ypical nosZ gene can e en domina e ma ine mic obiomes (Be agnolli e al. 2020). ni S and nosZ genes we e p esen in bo h oxic and suboxic condi ions, bu hey showed a p e e - ence o low oxygen en i onmen s. ni S abundance was nega- i ely ela ed o he oxygen a ailabili y, and he nosZ abundance showed he highes abundances a low oxygen condi ions. The e o e, DO concen a ion may a ec , bu no inhibi , deni ifica ion. In his ega d, p e ious s udies s a ed ha deni ifica ion is a acul a i e anae obic p ocess, and i s las s ep o N 2 O educ ion is s ongly sensi i e o e en aces o oxygen (Dalsgaa d e al. 2014). Howe e , o he s udies de ec ed he occu ence o deni i ying bac e ia (i.e., ni S ni K and nosZ genes) in oxic and anoxic condi ions in he lake wa e column (Junie e al. 2008; Kim e al. 2011; Mao e al. 2017; Paja es e al. 2017). Ae obic deni i ying bac e ia occu in di e se en i onmen s, and he e ec o DO may depend on he specific mic oo ganism (Lloyd 1993). In addi ion o DO, TP and cumula i e Chl aconcen a ions in he wa e column had a specific posi i e e ec on he deni- ifie abundance (i.e., ni S and ni S : p oka yo es a io). Paja es e al. (2017) also epo ed ha phospho us concen a ion was posi i ely ela ed o he abundance o deni ifie s in a opical lake. Phospho us is a a e-limi ing nu ien in na u al sys ems (Guigna d e al. 2017), and may a ec deni i ying bac e ia eplica ion o pe o mance. In ac , he s udy o A a e al. (2015) sugges ed ha P a ailabili y a ec s deni ifica ion gene exp ession. This s udy was based on he gene egula o y and me abolic ne wo k o he deni ifica ion pa hway in Pseudomonas ae uginosa PAO1 unde di e en en i onmen al condi ions. They also confi med expe imen ally ha phospha e concen a ion inc eased N 2 O p oduc ion in P. ae uginosa cul u es. Al hough his s udy was pe o med only wi h one bac e ium, o he deni i ying bac e ia may ha e sim- ila egula ion sys ems, and ha may explain he ela ionship be ween ni S gene abundance and TP concen a ion ha we ound in he s udy ese oi s. Much a en ion has been paid o he e ec o ni ogen on he N 2 O p oduc ion o deni ifica- ion, bu P esul s being a key nu ien , and i may ha e a mo e ele an ole in deni ifica ion ha expec ed. On he o he hand, in his s udy he abundance o he ni S gene did no depend on DOC concen a ion (Supplemen a y Table S8), bu i depended on he cumula i e Chl aconcen a ion. This ela ionship may sugges ha deni ifi- ca ion is s imula ed by esh au och honous o ganic ma e , ha can se e as labile ca bon sou ce, as hey a e he e o o- phic o ganisms. This idea was also sugges ed by p e ious s ud- ies in eshwa e s (McMillan e al. 2010; Chen e al. 2012). Deni ifica ion a es in he ocean depend on he o ganic ma - e expo ed om he pho ic zone de i ed om p ima y p o- duc ion, and a ec ed by he quali y and quan i y o o ganic ma e (Kal elage e al. 2013; Wa d 2013b; Babbin e al. 2014). In addi ion, he high algal biomass o eu ophic ese oi s can esul in hypoxia e en s ha also s imula e deni ifica ion. In he second place, his ela ionship may also sugges ha deni- ifica ion and N 2 O p oduc ion in he wa e column is enhanced by sinking pa icles de i ed om he phy oplank- on communi y (Zhou e al. 2019b). Mic obial con ol on N 2 O concen a ion We ound ha he N 2 O concen a ion (nmol L 1 ) depended on he ni S abundance (copies mL 1 ) ollowing a powe unc ion (N 2 O=0.27 ni S 0.33 ,n=69, adj R 2 =0.28, p< 0.001; Fig. 5b). N 2 O also depended on he ni S : p oka y- o es a io (n=69, adj R 2 =0.22, p< 0.001; Fig. 5e). Howe e , N 2 O was no significan ly ela ed o he abundances o amoA o nosZ sepa a ely (Fig. 5a,c). When we modeled he e ec o he h ee genes on he N 2 O using a GAM, we ound ha he abundance o nosZ also a ec ed significan ly he N 2 O concen- a ion, in combina ion wi h he ni S gene (Supplemen a y Table S10). Toge he , hey explained up o 34% o he a i- ance, and up o 40% when we used he gene: p oka yo es a ios (i.e., ela i e abundance) ins ead o he absolu e gene abundances. We also s udied he ela ionship be ween he amoA :ni S and he ni S :nosZ a ios wi h he N 2 O, bu he esul s we e no significan (Suppo ing In o ma ion Fig. S19). Finally, we modeled he N 2 O concen a ion in he wa e col- umn as unc ion o he concen a ion o he main subs a e o deni ifica ion (i.e., NO 3 ), and he ela i e abundance o deni ifie s, which a e in ol ed in he p oduc ion (i.e., ni S : p oka yo es) and he consump ion o N 2 O(i.e., nosZ : p oka yo es a ios) (Fig. 6). N 2 O concen a ion was a posi i e exponen ial unc ion o he ni a e and he Le on-Palme o e al. N 2 O in he wa e column o ese oi s 9 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License Weiss, R. F., and B. A. P ice. 1980. Ni ous oxide solubili y in wa e and seawa e . Ma . Chem. 8: 347–359. doi:10.1016/ 0304-4203(80)90024-9 We zel, R. G. 2001. Limnology: Lake and i e ecosys ems. Eos T ans Am Geophys Union 21:1–9. Wood, S. N. 2006. Gene alized addi i e models: An in oduc- ion wi h R. Chapman and Hall/CRC. W age, N., G. L. Vel ho , M. L. an Beusichem, and O. Oenema. 2001. Role o ni ifie deni ifica ion in he p oduc ion o ni ous oxide. Soil Biol. Biochem. 33: 1723–1732. doi:10.1016/S0038-0717(01)00096-7 Za fl, C., A. E. Lumsdon, J. Be lekamp, L. Tydecks, and K. Tockne . 2015. A global boom in hyd opowe dam con- s uc ion. Aqua . Sci. 77: 161–170. doi:10.1007/s00027- 014-0377-0 Zhou, S., Y. Zhang, T. Huang, Y. Liu, K. Fang, and C. Zhang. 2019a. Mic obial ae obic deni ifica ion domina es ni ogen losses om ese oi ecosys em in he sp ing o Zhoucun ese oi . Sci. To al En i on. 651: 998–1010. doi:10.1016/j. sci o en .2018.09.160 Zhou, Y., X. Xu, R. Han, L. Li, Y. Feng, S. Yee ken, K. Song, and Q. Wang. 2019b. Suspended pa icles po en ially enhance ni ous oxide (N 2 O) emissions in he oxic es ua- ine wa e s o eu ophic lakes: Field and expe imen al e i- dence. En i on. Pollu . 252: 1225–1234. doi:10.1016/j. en pol.2019.06.076 Zhou, Y., and o he s. 2021. Nonlinea pa e n and algal dual- impac in N 2 O emission wi h inc easing ophic le els in shallow lakes. Wa e Res. 203: 117489. doi:10.1016/j. wa es.2021.117489 Acknowledgmen s We especially hank Eulogio Co al and Alba Con e as-Ruiz o helping in he field, and labo a o y wo k, espec i ely. We also hank Ana Sie a, Jesús Fo ja and Teodo a O ega o helping wi h gas ch oma og aphy analy- sis a he Uni e si y o C adiz. We hank Amal Jayakuma om Wa d Lab a P ince on Uni e si y o he aining in qPCR. We hank he Con ede aci on Hid og  afica del Guadalqui i and he Agencia Andaluza del Medio Ambien e y Agua (AMAYA) o acili a ing he ese oi sampling. This esea ch was suppo ed by he Minis e io de Economía y Compe i i idad (HERA p ojec , g an no. CGL2014-52362-R), and he Minis e io de Ciencia, Inno aci on y Uni e sidades (CRONOS p ojec , RTI2018-098849-B-I00) o Spain o IR and RM-B. Elizabe h Le on-Palme o was suppo ed by a PhD el- lowship om he Minis e io de Educaci on, Cul u a y Depo e o Spain (g an nos. FPU014/02917), and a pos doc o al con ac om CRONOS p ojec , and la e om Danma ks F ie Fo sknings ond (DFF, 1026-00428B) a SDU. This manusc ip was imp o ed h ough eedback om an anony- mous e iewe , and D . Van Me e . Uni e sidad de G anada/CBUA unded he open access o his a icle. Conflic o In e es The au ho s decla e no compe ing financial in e es . Da a A ailabili y S a emen Da a suppo ing he findings o his s udy a e a ailable wi hin a icle, and in he Supplemen a y In o ma ion, and aw da a a e a ailable on eques om he au ho s. Submi ed 06 May 2022 Re ised 01 Feb ua y 2023 Accep ed 18 May 2023 Associa e edi o : Kimbe ly Van Me e Le on-Palme o e al. N 2 O in he wa e column o ese oi s 16 19395590, 0, Downloaded om h ps://aslopubs.onlinelib a y.wiley.com/doi/10.1002/lno.12381 by Uni e sidad De G anada, Wiley Online Lib a y on [30/06/2023]. See he Te ms and Condi ions (h ps://onlinelib a y.wiley.com/ e ms-and-condi ions) on Wiley Online Lib a y o ules o use; OA a icles a e go e ned by he applicable C ea i e Commons License