Greenhouse gas fluxes in a drained peatland forest during spring frost-thaw event

Biogeosciences, 7, 1715–1727, 2010 www.biogeosciences.ne /7/1715/2010/ doi:10.5194/bg-7-1715-2010 © Au ho (s) 2010. CC A ibu ion 3.0 License. Biogeosciences G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en M. K. Pihla ie1, R. Kiese2, N. B ¨ uggemann2, K. Bu e bach-Bahl2, A.-J. Kieloaho1, T. Lau ila3, A. Lohila3, I. Mamma ella1, K. Minkkinen4, T. Pen il¨ a5, J. Sch¨ onbo n2,6, and T. Vesala1 1Depa men o Physics, Uni e si y o Helsinki, P. O. Box 48, Uni e si y o Helsinki, 00014, Helsinki, Finland 2Ins i u e o Me eo ology and Clima e Resea ch, A mosphe ic En i onmen al Resea ch (IMK-IFU), Ka ls uhe Ins i u e o Technology, Ga misch-Pa enki chen, Ge many 3Finnish Me eo ological Ins i u e, P. O. Box 503, 00101, Helsinki, Finland 4Depa men o Fo es Ecology, Uni e si y o Helsinki, P. O. Box 27, Uni e si y o Helsinki, 00014, Helsinki, Finland 5Finnish Fo es Resea ch Ins i u e, Van aa Uni , Finland 6Me eo ological Ins i u e, Albe -Ludwigs-Uni e si y F eibu g, F eibu g, Ge many Recei ed: 29 May 2009 – Published in Biogeosciences Discuss.: 23 June 2009 Re ised: 8 Ap il 2010 – Accep ed: 28 Ap il 2010 – Published: 25 May 2010 Abs ac . Fluxes o g eenhouse gases (GHG) ca bon diox- ide (CO2), me hane (CH4) and ni ous oxide (N2O) we e measu ed du ing a wo mon h campaign a a d ained pea - land o es in Finland by he eddy co a iance (EC) echnique (CO2and N2O), and au oma ic and manual chambe s (CO2, CH4and N2O). In addi ion, GHG concen a ions and soil pa- ame e s (mine al ni ogen, empe a u e, mois u e con en ) in he pea p o ile we e measu ed. The aim o he measu emen campaign was o quan i y he GHG luxes du ing eezing and hawing o he op-soil, a ime pe iod wi h po en ially high GHG luxes, and o compa e di e en lux measu e- men me hods. The o es was a ne CO2sink du ing he wo mon hs and he luxes o CO2domina ed he GHG ex- change. The pea soil was a small sink o a mosphe ic CH4 and a small sou ce o N2O. Bo h CH4oxida ion and N2O p oduc ion ook place in he op-soil whe eas CH4was p o- duced in he deepe laye s o he pea , which we e un ozen h oughou he measu emen pe iod. Du ing he os - haw e en s o he li e laye dis inc peaks in CO2and N2O emis- sions we e obse ed. The CO2peak ollowed igh ly he in- c ease in soil empe a u e, whe eas he N2O peak occu ed wi h a delay a e he hawing o he li e laye . CH4 luxes Co espondence o: M. K. Pihla ie ([email p o ec ed]) did no espond o he hawing o he pea soil. The CO2 and N2O emission peaks we e no cap u ed by he manual chambe s and hence we conclude ha high ime- esolu ion measu emen s wi h au oma ic chambe s o EC a e neces- sa y o quan i y luxes du ing peak emission pe iods. Sub- canopy EC measu emen s and chambe -based luxes o CO2 and N2O we e compa able, al hough he luxes o N2O mea- su ed by EC we e close o he de ec ion limi o he sys em. We conclude ha i luxes a e high enough, i.e. g ea e han 5–10µgNm−2h−1, he EC me hod is a good al e na i e o measu e N2O and CO2 luxes a ecosys em scale, he eby minimizing p oblems wi h chambe enclosu es and spa ial ep esen a i eness o he measu emen s. 1 In oduc ion D ainage o pea lands o o es y has been a common p ac- ice in Fennoscandia du ing he pas 100 yea s. In Fin- land, mo e han hal o he o iginal pea land a ea has been d ained o o es y o ag icul u al use since he 1920s (Paa - ilainen and P¨ ai ¨ anen, 1995; Joos en and Claa ke, 2002). D ainage lowe s he g oundwa e able and imp o es he ae a ion o he pea , which inc eases he g ow h o ees. The eby, d ainage also changes g eenhouse gas dynamics o he pea land, as a la ge pa o he decomposi ion o he pea Published by Cope nicus Publica ions on behal o he Eu opean Geosciences Union. 1716 M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en swi ches om anae obic o ae obic condi ions wi h a shi om me hane (CH4) o ca bon dioxide (CO2) as he end- p oduc o decomposi ion (Moo e and Dal a, 1993; Sil ola e al., 1996; Minkkinen e al., 2002; Roule e al., 1993, Ma - ikainen e al., 1995; Nyk¨ anen e al., 1998). S imula ed ae - obic decomposi ion o he pea eleases nu ien s, especially ni ogen, o he soil, which may lead o ele a ed emissions o ni ous oxide (N2O) (Ma ikainen e al., 1993; Sil ola e al., 1996; Laine e al., 1996). Howe e , he changes in N2O emissions a e d ainage seem o depend on he e ili y o he o iginal pea land, i.e. i s ni ogen con en o he C:N a io o he pea , and he le el o he wa e able a e he d ainage ( on A nold e al., 2005a, b). D ained pea lands which ha e been used o ag icul u e i s and hen plan ed wi h ees (a o es ed pea soils) a e s ong poin sou ces o N2O. These N2O emissions a e o he same o de o magni ude as he emissions om d ained pea lands which a e s ill used o ag icul u e (Maljanen e al., 2001; Regina e al., 2004; M¨ aki an a e al., 2007). D ained o es ed pea lands co e 25% o o es a ea in Finland mak- ing hese ecosys ems po en ially impo an sou ces o g een- house gases. Du ing he las wo decades he e has been de- ba e whe he he d ainage o pea lands o o es y u ns hem om ne sinks o ca bon in o ne sou ces, and whe he N2O makes up an impo an pa o he o al g eenhouse gas bal- ance. In ensi e measu emen s o GHG emissions om d ained pea land o es s a e sca ce. Also, compa isons o di e en measu emen echniques in hese ecosys ems a e almos non- exis en . Mos o hes udies ha ebeen conduc edwi h cham- be echniques using weekly o mon hly measu ing in e als. This measu emen equency may se e ely miss impo an emission e en s, so called “ho momen s”, ela ed especially o N2O emissions om soils (see e.g. Ma zne and Bo ken, 2008; Papen and Bu e bach-Bahl, 1999), such as os - haw pe iods which could be subs an ial in bo eal en i onmen s (see e.g. Koponen e al., 2004, 2006). As a esul , calcula- ions o seasonal o annual budge s o g eenhouse gases may be biased and po en ially unde es ima ed i he equency o measu emen s o spa ial co e age is no su icien o co e a ia ions. We epo esul s o g eenhouse gas emissions (CO2, CH4 and N2O) om a d ained pea land o es in Kale ansuo, sou he n Finland. The measu emen campaign las ed wo mon hs om he end o Ap il un il he end o June 2007, and was un unde he Ni oEu ope IP EU-p ojec . The main aim was o quan i y he o al GHG balance du ing a po en ially high peak season in he sp ing, when he pea is mel ing and os - haw d i en N2O luxes a e likely o occu . We mea- su ed ne CO2exchange abo e and below he o es canopy and N2O exchange below he o es canopy by he eddy co- a iance (EC) me hod and compa ed hese luxes o soil CO2, CH4, and N2O luxes measu ed simul aneously wi h au o- ma ed and manual chambe echniques. Ou aim was o es ima e he ne GHG exchange and he impo ance o di e en C and N lux componen s on he o- al GHG balance du ing he wo-mon h measu ing pe iod. We hypo hesise ha N2O is an impo an componen o he ecosys em g eenhouse gas exchange due o “ho momen ” emissions such as os - haw e en s. Ou second aim was o e alua e he sui abili y o sub-canopy EC-based N2O mea- su emen s as a sophis ica ed al e na i e o adi ionally used chambe me hods in his en i onmen . The quali y con ol and lux e o analysis o he EC N2O measu emen s a he si e a e p esen ed in his issue in Mamma ella e al. (2010). 2 Ma e ials and me hods 2.1 Si e desc ip ion The measu emen s we e conduc ed a a Kale ansuo d ained pea land o es classi ied as an omb o ophic dwa -sh ub pine bog. The si e is loca ed in sou he n Finland (60◦390N, 24◦220E), whe e he mean annual p ecipi a ion is 606mm and he mean annual empe a u e is 4.3◦C. The bog was d ained o o es y in 1971 by open, abou 1m deep di ches dug wi h app oxima ely 40m spacing be ween he pa allel di ches. In 1973 he si e was e ilised wi h phospho us and po assium, ollowing he guidelinep ac ises o d ained pea - lands. D ainage esul ed in a lowe ed wa e able down o app ox. 40cm om he pea su ace, and a changed compo- si ion o g ound ege a ion om ypical bog ege a ion o- wa ds mo e o a o es unde s o ey. Howe e , some ea u es such as he abundance o pea land dwa sh ubs and ai ly high co e age o Sphagnum species s ill dis inguish he si e om upland o es s. Cu en ly he heigh o he ee s and is 15–18m, a e age basal a ea is 18m2ha−1, and a e age s em densi ies a e 900, 750, and 40 s ems pe ha o he dominan Sco s pine (Pinus syl es is L.) ees and he smalle unde - s o ey downy bi ch (Be ula pubescens) and No way sp uce (Picea abies L.) ees, espec i ely. The o al LAI in he si e is app oxima ely 2m2m−2(Mamma ella e al., 2010). Fo es loo ege a ion consis ed mainly o hummock dwa sh ubs (Vaccinium i is-idaea, Vaccinium my illus, Empe um nig um, Vaccinium uliginosum, Ledum palus e and Be ula nana), sedges like E iopho um agina um and mosses (Pleu ozium sch ebe i,Dic anum polyse um, Sphag- num ussowii, Spagnum capilli olium and Sphagnum angus- i olium). The dep h o he well decomposed Sphagnum pea a he si e is app oxima ely 2.5m wi h pea a pH o 5.0 and C/N a- io o 41 in he li e laye and 45 in he op 10cm o he pea soil. 2.2 Flux measu emen s In ensi e GHG measu emen s we e ca ied ou om 25 Ap il o 27 June 2007 wi hin a homogenous and ep esen a i e ap- p ox. 1ha plo o he o es ( o al a ea o app ox. 60ha). The Biogeosciences, 7, 1715–1727, 2010 www.biogeosciences.ne /7/1715/2010/ M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en 1717 main measu emen s included mic ome eo ological eddy co- a iance (EC) measu emen s o CO2abo e and below he o es canopy and N2O luxes below he canopy, au oma ed as well as manual chambe -based measu emen s o CO2, CH4and N2O luxes. The loca ions o he di e en mea- su emen sys ems a e shown in Fig. 1. The abo e canopy EC CO2 lux measu emen sys em (ECabo e) included a METEK USA-1 ul a sonic anemome- e (METEK GmbH, Elmsho n, Ge many) moun ed on he op o a 21.5m elescopic mas and a LI-7000 CO2/H2O analyze (Li-Co , Inc., Lincoln, NE, USA) moun ed a 6m heigh in he owe . Ai was d awn om he p oximi y o he sonic o he LI-7000 CO2/H2O analyze using a Be -A- Line IV ubing (The moplas ic p ocesses, S i ling, NJ) wi h an inne diame e o 3.1mm. The s o age lux o CO2was calcula ed om he concen a ion da a measu ed a heigh s o 21.5m and 6m, he la e being measu ed wi h a LI-820 CO2 analyze (Li-Co Inc., Lincoln, NE, USA). The s o age lux was added o he measu ed ne ecosys em exchange (NEE), he ea e NEE e e ing o he sum o u bulen and s o age luxes. The mas was loca ed in he cen e o he measu e- men si e (Fig. 1). The sub-canopy EC measu emen s (ECsub) we e con- duc ed a 4m heigh . The sub-canopy mas was loca ed ap- p oxima ely 100m sou hwes o he all mas , and app oxi- ma ely hal way be ween he all mas and he au oma ic soil chambe s (see Fig. 1). The CO2 luxes we e measu ed wi h a Li-7500 Open-Pa h In a ed CO2/H2O Gas Analyze (Li- Co , Inc., Lincoln, NE, USA) and a CSAT3 Sonic Anemome- e (Campbell Scien i ic Inc., Logan, UT, USA). EC mea- su emen s o N2O luxes we e conduc ed a he same mas using he same CSAT3 anemome e and a unable diode lase spec ome e (TGA-100A, Campbell Scien i ic Inc., Logan, UT, USA). Fo es loo (soil and g ound ege a ion) luxes o CO2, N2O and CH4we e measu ed wi h he enclosu e me hod using au oma ic ( anspa en ) and manual (opaque) cham- be s. The au oma ic chambe sys em consis ed o a al e-d i en sampling sys em (cus om-made by IMK-IFU) o nine soil chambe s wi h dimensions o 50×50×15cm (leng h ×wid h ×heigh ). The au oma ic chambe s we e loca ed app ox. 170m sou hwes o he all EC mas , and ap- p ox.100msou hwes om hesub-canopyEC mas (Fig. 1). The chambe s we e connec ed o a gas ch oma og aph (SRI Ins umen s, To ance, CA, USA) equipped wi h an elec on cap u e de ec o (ECD) o N2O and a lame ioniza ion de- ec o (FID) o CH4, and a GMD20D in a ed CO2analyze (Vaisala, Van aa, Finland). The nine chambe s we e spli in o 3 se s o 3 chambe s. One measu emen cycle included clo- su es o 3 chambe s and a simul aneous calib a ion wi h a e e ence gas. Each chambe was closed o 48min, and he mean sampling in e als we e 6, 18, 30, and 42min a e he closu e. The measu emen sys em is desc ibed in mo e de ail in Kiese and Bu e bach-Bahl (2002) and We ne e al. (2007). The ege a ion inside he au oma ic chambe s was Fig. 1. Map o he measu emen si e showing he loca ions o abo e canopy eddy co a iance (EC) mas (ECabo e), sub-canopy EC mas (ECsub), manual chambe s (MC, squa e) and au oma ic chambe s (AC, ci cle). Do ed line nex o one o he manual cham- be g oups show he place o soil gas concen a ion pi s, and g ey line a ound he sub-canopy EC mas show he oo p in a ea om which 85% (a 30m) o he sub-canopy N2O luxes o igina e (see Mamma ella e al., 2010). simila han in he pea land gene ally, howe e , all dwa sh ubs we e no p esen . De ailed ege a ion su ey was no conduc ed o he au oma ic chambe s. Manual chambe measu emen s we e conduc ed once a week du ing Ap il o June 2007, and o nigh ly du ing July o Sep embe 2007. In o al 16 ci cula me al colla s we e loca ed in g oups o ou app ox. 30–60m om he all EC mas in he ou main di ec ions, and 10–150m no h-eas om he sub-canopy EC mas (Fig. 1). The colla s we e in- s alled in 2004 a soil dep h o 3–5cm, on op o he oo laye . Du ing chambe measu emen s, a 30cm high ci - cula me al chambe was placed on he colla . Volume o he chambe was app ox. 27L. Ai inside he chambe was mixed wi h a an, and he empe a u e inside he chambe was moni o ed wi h a he mome e in o de o co ec he luxes. Gas samples (100ml) we e collec ed wi h a sy inge a 2, 15, 25 and 35min in e als and ans e ed immedi- a ely in o 12-ml glass ials (Labco Exe aine ®, Labco Lim- i ed, Buckinghamshi e, UK). Nine y ml o he gas sample was used o lush he ai in he ial wi h wo needles. The es 10ml o he gas sample was used o o e -p essu ize he ial a e emo ing he lushing needle. Gas samples we e analyzed wi hin one week o N2O and CH4by a www.biogeosciences.ne /7/1715/2010/ Biogeosciences, 7, 1715–1727, 2010 1718 M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en gas ch oma og aph (Agilen 6890 GC, Agilen Technologies Finland, Espoo, Finland) equipped wi h an ECD o N2O and an FID o CH4. 2.3 Soil measu emen s Concen a ions o N2O and CH4in he pea p o ile we e mea- su ed a wo pi s loca ed app oxima ely 40m sou hwes o he abo e-canopy EC mas . The concen a ions we e mea- su ed in he pea a 5cm, 22 and 45cm below he li e laye . Gas collec o cups we e 100ml in olume and made o s ain- lesss eel. Thecupswe eins alledho izon allyapp ox.20cm apa om each o he , upside down wi h an open end a he bo om in he soil and connec ed o he a mosphe e ia a 1/800 s ainless s eel ube. Gas samples we e collec ed weekly du ing Ap il o June om he dep hs 5 and 25cm and o - nigh ly du ing July o Sep embe om all dep hs (5, 22 and 45cm). A he ime o gas sampling, he 5–10ml gas ol- ume inside he ubing was disca ded a e which a 100ml gas sample was aken and ans e ed in o 12-ml glass ials as de- sc ibed abo e. When a gas collec o was below he g ound- wa e able, a wa e sample o 50ml was aken wi h he sy- inge. Then he gas dissol ed in he wa e was equilib a ed wi h 50ml o ambien ai by shaking he sy inge igo ously o 10min. A e shaking, 20ml o he gas sample was in- jec ed in o a p e-e acua ed 12-ml glass ial. Soil empe a u e and olume ic wa e con en s we e mea- su ed adjacen o he au oma ic chambe s in he li e laye and a 5 and 10cm dep hs o he pea (T ime®TDR IMKO and P -100, IMKO GmbH, E lingen, Ge many). In addi- ion, soil empe a u es in he li e laye , and a 5 and 30cm dep hs o he pea we e measu ed close o he all EC mas by FMI (Finnish Me eo ogical Ins i u e). The a ia ion o he g ound wa e le el nea he main EC mas was moni o ed by a PDCR 1830 le el p essu e senso (D uck Inc., New Fai - ield, CT, USA). Soil ammonium (NH4-N), ni a e (NO3-N) and o al dis- sol ed ni ogen con en s we e analysed om samples col- lec ed weekly du ing Ap il o June 2007, and mon hly du - ing July o Sep embe 2007. Soil samples om he li e laye and pea (0–10cm) we e collec ed in 5 eplica es: ou om close incini y o he manual chambe s (4 g oups) and one om close incini y o he au oma ic chambe s. F esh soil samples we e s o ed a +4◦C and ex ac ed wi h 1MKCl he nex day a e he sampling. The ex ac s we e ozen a −18◦C un il analysis by a low injec ion analyze (FIA 5012, Teca o ) a he Finnish Fo es Resea ch Ins i u e. To al ca - bon and ni ogen con en s we e analyzed om d ied (40◦C) soil samples using a a io MAX CN elemen al analyse . 2.4 Da a analysis Flux a es o manual and au oma ed chambe measu emen s we e calcula ed wi h he ollowing equa ion Fc=dC d h, (1) whe e Fcis he lux o he a ge gas (gm−2s−1), Cis he gas concen a ion in he chambe ai (gm−3) a s anda d p essu e (101325Pa) and empe a u e measu ed in he headspace, is closu e ime (s) and h he heigh o he chambe (m). The de- elopmen o he gas concen a ion inside he chambe s was linea o he majo i y o he measu emen s. Fo he man- ual chambe da a we compa ed luxes calcula ed based on quad a ic i and linea eg ession. The use o a quad a ic i esul ed in up o 30% highe luxes o CH4and 20% smalle luxes o N2O as compa ed o he linea eg ession. Due o only ou da a poin s and luxes close o ze o, we conside ed ha he linea eg ession me hod was mo e eliable o his da a and hence we calcula ed all he luxes by a linea e- g ession analysis (n=4). We il e ed ou bad quali y da a by emo ing da a wi h R2- alue 0.7 o less. EC luxes we e calcula ed as 30min a e age co a iances be ween he scala s (CO2and N2O) concen a ion and he e ical wind eloci y acco ding o he commonly accep ed p ocedu es (Aubine e al., 2000). The abo e canopy EC da a acquisi ion was done wi h a modi ied e sion o a p og am by McMillen (1986). Coo dina e o a ion and da a de ending by an au o eg essi e unning-mean il e wi h a 200-s ime cons an we e pe o med acco ding o McMillen (1988). The lag be ween he ime se ies esul ing om he anspo h ough he inle ube was aken in o accoun in he on-line calcula ion. An ai densi y co ec ion ela ed o he sensible hea lux is no necessa y, bu he co esponding co ec ion ela ed o he la en hea lux was made (Webb e al., 1980). Co ec ions o he sys ema ic high- equency lux loss ow- ing o he impe ec p ope ies and se up o he senso s we e ca ied ou o -line using ans e unc ions wi h empi ically- de e mined ime cons an s. The da a p ocessing p ocedu es ha e been p esen ed in mo e de ail by Lohila e al. (2007) and Au ela e al. (2009). The sub-canopy luxes we e calcula ed using so wa e de- eloped by he Mic ome eo ology g oup a he Uni e si y o Helsinki, Depa men o Physics. The so wa e is ou inely used o pos -p ocessing EC da a measu ed in se e al pe - manen si es and ield campaigns. I con ains all he upda e me hods and co ec ions acco ding o he Eu o lux me hod- ology (Aubine e al., 2000; Lee e al., 2004). Fo he p esen s udy, he so wa e was sligh ly modi ied in o de o handle wi h he lase da a, as epo ed by Mamma ella e al. (2010). All signals we e de ended o emo ing he a e age alues and ends. A simple linea de ending p ocedu e was used o calcula ing he CO2 lux. The N2O signal measu ed by he TDL gas analyze was cha ac e ized by s onge ends, caused mainly by ins umen al d i , which can gi e an ex a Biogeosciences, 7, 1715–1727, 2010 www.biogeosciences.ne /7/1715/2010/ M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en 1719 con ibu ion o he es ima ed lux in he case ha he luc ua- ions o he concen a ion a e co ela ed wi h he luc ua ions o he e ical wind eloci y. In o de o emo e he ins u- men al d i e ec and o educe he andom lux a iabili y, a unning mean il e (McMillen, 1988) was pe o med p io o calcula ion o he N2O lux. A mo e de ailed desc ip ion o he da a p ocessing o N2O EC signal is gi en in Mam- ma ella e al. (2010). Alag- imeo 2.3s was ob ained o he abo e-canopy CO2 signal, maximizing he c oss-co a iance unc ion be ween he CO2concen a ion and he e ical wind eloci y. The same p ocedu e was applied o he sub-canopy N2O signal, bu because he N2O emissions we e e y close o de ec ion limi o he sys em, i was no possible o clea ly de e mine expe imen ally he N2O lag ime. Then using a p ocedu e simila o Pihla ie e al. (2005), we used a ix lag ime o 1s. The same alue was ob ained by using he sample low and olumes o he inle ubing and he sample cell, o es i- ma ing he heo e ical N2O lag ime. The CO2 lux was co - ec ed o densi y luc ua ions e ec (WPL co ec ion; Webb e al., 1980), while such co ec ion was unnecessa y o N2O luxes, because o he p esence o high low sample d ye in he sys em (PD1000 Na ion®d ye , Campbell Scien i ic, Inc., Logan, UT, USA). Tempe a u e luc ua ions do no need o be co ec ed because hey can be assumed o be damped in he sampling ube (Rannik e al., 1997). No Bu ba co ec ion was used o he eddy co a iance da a om open pa h CO2 analyze e en hough he co ec ion may sligh ly inc ease he lux le els (Bu ba e al., 2008). The EC luxes we e co - ec ed o he high equency lux unde es ima ion acco ding o Mamma ella e al. (2010). Fo ypical mean wind eloci y in he sub-canopy laye , he lux loss was abou 5% and less han 10% o CO2and N2O, espec i ely. S a is ical es s (pai ed - es ) o he lux and soil mea- su emen da a was done wi h SPSS s a is ical p og am (SPSS Inc., Chicago, IL, USA). 3 Resul s 3.1 En i onmen al condi ions A he s a o he measu emen campaign pa o he pea was s ill ozen. The ai empe a u es a ied om below 0◦C in he end o Ap il o a maximum o 27◦C in he begin- ning o June (Fig. 2). P io o he s a o he measu emen campaign he soil had mel ed and ozen se e al imes. The i s p onounced eeze- haw cycle was eco ded in he end o Ma ch, one mon h p io o he measu emen campaign (da a no shown). Howe e , as indica ed by empe a u e measu e- men s o ai and li e laye , he pea su ace laye was s ill eezing and hawing du ing he measu ing campaign in he end o Ap il (Fig. 2). Du ing he in ensi e measu emen pe- iod (25 Ap il–27 June) he soil empe a u e inc eased om a ound 0◦C up o app ox. 16◦C in he uppe pa (5cm dep h) o he pea soil. Rain all du ing Ap il–June was low wi h low in ensi ies excep o wo e en s in mid Ap il and in he end o May, esul ing in sho inc eases in he soil wa e con en (Max. 22 ol%) and wa e able (see Fig. 2). De- spi e hese sho inc eases, he wa e able and soil mois u e dec eased (−25cm o −40cm; 16 o <10 ol%) du ing he in ensi e measu emen pe iod. 3.2 Concen a ion o soil ammonium, ni a e and o al dissol ed ni ogen Soil ni a e (NO− 3-N) concen a ions we e close o ze o h oughou he whole measu ing pe iod, whe eas soil am- monium (NH+ 4-N) and o al ni ogen ( o -N) concen a ions we e ele a ed a he beginning o he measu emen pe iod wi h a maximum du ing he os - haw e en in May, and de- c eased owa ds he end o he measu ing campaign (Fig. 6). The concen a ions o NO− 3-N, NH+ 4-N and o -N we e al- ways highe in he li e laye han in he pea a 0–10cm dep h (da a no shown). To al dissol ed ni ogen concen a- ions in he soil a ied be ween 50–230mgNkg−1d y soil, and we e app oxima ely one o de o magni ude highe han he concen a ions o NH+ 4-N in he soil. 3.3 CO2 luxes EC measu emen s abo e he o es canopy e ealed ha he si e was on a e age a ne sink o CO2du ing he mea- su ing campaign, om la e Ap il o la e June 2007 (see Fig. 3). The daily ne ecosys em exchange (NEE) o CO2 inc eased om app oxima ely −0.014mgCm−2s−1du ing Ap il o maximum o −0.064mgCm−2s−1in he middle o June. The d ained pea land o es was a weak sou ce o ca - bon (0.02mgCm−2s−1) on ew ainy days du ing he mea- su emen pe iod. O e all, he CO2exchange ollowed he changes in ai and soil empe a u es being highe (up ake) in wa m and lowe (up o emission) in cold days (see Figs. 2 and 3). In con as o he ne CO2up ake o he whole o es ecosys em, soil and g ound ege a ion oge he u ned ou o be a sou ce o CO2 o he a mosphe e. Bo h, CO2 luxes be- low he o es canopy measu ed by he EC and by au oma ic chambe s on he soil su ace showed an inc easing emission end om Ap il o June (Fig. 3). Fo es loo CO2 luxes (au oma ic chambe s) and sub-canopy luxes (sub-canopy EC) inc eased om a minimum o 0.001mgCm−2s−1in he end o Ap il o a maximum o 0.013mgCm−2s−1and 0.03mgCm−2s−1, espec i ely, in he end o May when also soil and ai empe a u es eached hei maximum. In June a dec ease in empe a u e was ollowed by a dec ease in CO2 luxes, howe e , his was mo e p onounced in he sub-canopy EC luxes. In he end o June o es loo and sub-canopy luxes le eled a ound 0.01mgCm−2s−1, how- e e s ill ollowing changes in he ai and soil empe a u es (Figs. 2 and 3). Mean o es loo (0.008mgCm−2s−1) www.biogeosciences.ne /7/1715/2010/ Biogeosciences, 7, 1715–1727, 2010 1720 M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en Fig. 2. (a) Ai empe a u e, (b) soil empe a u es in li e laye (hummock and hollow) and in pea , and (c) soil mois u e ( ol/ ol), g ound wa e able dep h (WT) and p ecipi a ion a he d ained pea land pine o es du ing Ap il–Sep embe 2007 (in ensi e measu emen s 25 Ap il– 27 June). and sub-canopy CO2exchange (0.009mgCm−2s−1) o e he measu ing pe iod we e almos iden ical and a pai ed - es analysis did no e eal any s a is ical di e ences (Ta- ble 1). Fo es loo CO2and sub-canopy exchange co ela ed posi i ely wi h ai and soil empe a u es. The soil empe - a u e a 5cm dep h explained mos o he a iabili y in o - es loo CO2 lux a es ( =0.96, p<0.01). The co ela ion was less p onounced o sub-canopy EC based luxes due o a mo e sca e ed empo al emission pa e n also e lec ed in highe alues o CV% (Table 1, Fig. 3). Fu he mo e, we ound a nega i e co ela ion o o es loo CO2 luxes wi h soil mois u e (−0.60, p<0.01) and wa e able dep h (−0.76, p<0.01). These co ela ions we e no signi ican o he EC- based sub-canopy measu emen s. The measu emen campaign can be di ided in o wo dis- inc pe iods: a cold and a wa m pe iod. Du ing he cold pe iod (30 Ap il–10 May) he ne o es loo CO2 luxes, he sum o soil espi a ion and CO2pho osyn hesis o g ound ege a ion, and he CO2ne ecosys em exchange (NEE) abo e he o es canopywe esmall(Fig.4). Du ing hewa m pe iod (5 June–15 June) bo h he ne CO2emissions o he o es loo (Fig. 4c) and he ne CO2up ake o he o es canopy (Fig. 4d) inc eased. Du ing bo h cold and wa m pe- iods, he sub-canopy CO2 luxes ollowed a small bu clea diu nal end when he ne CO2emission dec eased du ing day- ime and inc eased du ing nigh - ime (Fig. 4a and c). The compa ison o he mean and median GHG exchange measu ed by abo e canopy EC and sub-canopy EC and by au oma ic o es loo chambe s du ing he en i e wo-mon hs measu emen pe iod is shown in Table 1, and he cumula i e luxes a e shown in Table 2. Du ing he pe iod o 25 Ap il– 21 June he cumula i e CO2 luxes measu ed by sub-canopy EC (42.5gCm−2) and o es loo chambe s (37.7gCm−2) did no s a is ically di e om each o he , and accoun ed o Biogeosciences, 7, 1715–1727, 2010 www.biogeosciences.ne /7/1715/2010/ M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en 1721 Table 1. Mean and median luxes o CO2, CH4and N2O and coe icien o a ia ion (CV%1) measu ed by eddy co a iance, and au oma ic and manual chambe s in Kale ansuo pea land o es du ing 25 Ap il–27 June 2007. ECaand ECss and o eddy co a iance abo e and below he canopy, espec i ely, and AC and MC s and o au oma ic and manual chambe s, espec i ely. mgCO2-Cm−2s−1µgCH4-Cm−2h−1µgN2O-Nm−2h−1 CO2ECaCO2ECsCO2AC CH4AC CH4MC2N2O ECsN2O AC N2O MC2 Mean3−0.031a0.009b0.008b−37.1a−18.5b3.2a4.5b6.8c Median −0.026 0.008 0.008 −35.6 −15.2 2.5 3.9 6.8 CV% 180 75.7 45.3 40.2 144 123 62.3 42.8 1Coe icien o Va ia ion was calcula ed as CV% = s de o he lux/mean lux×100. 2Measu emen pe iod 25 Ap il–18 June 2007. 3Di e en supe sc ip s indica e signi ican di e ences be ween lux a es o one componen measu ed wi h di e en me hods. Fig. 3. (a) Daily mean CO2exchange measu ed wi h eddy co a i- ance abo e he o es canopy (EC abo e) and inside he canopy (EC sub) and au oma ic chambe s (AC) a he d ained pea land pine o - es . E o ba s s and o s anda d de ia ions. 42 and 37% o he o al NEE (−102gCm−2), espec i ely (Table 2). 3.4 CH4 luxes Kale ansuo pea land o es was a small sink o CH4du ing he measu emen campaign (Fig. 5a). The CH4up ake mea- su ed wi h he au oma ic chambe s inc eased om a ound −30µgCm−2h−1 o a app oxima ely o −60µgCm−2h−1 in June. The CH4 luxes measu ed wi h manual chambe s we e cons an ly by a leas a ac o o wo smalle han he CH4 luxes measu ed wi h he au oma ic chambe s (Fig. 5). The luxes o CH4we e no a ec ed by hawing o he soil bu ollowed mo e closely he g oundwa e able and soil mois u e con en in he pea . CH4up ake co ela ed posi- i ely wi h soil wa e con en ( =0.38, p<0.01) and wa e able dep h ( =0.44, p<0.01), and nega i ely wi h soil em- pe a u es a 5cm and a 30cm dep h ( =−0.50, p<0.01; =−0.62, p<0.01), espec i ely, and CO2 luxes measu ed by he au oma ic chambe s ( =−0.50, p<0.01). Fig. 4. Daily ime cou se o CO2 luxes a he d ained pea - land pine o es measu ed wi h au oma ic chambe s (AC) and sub- canopy eddy co a iance (EC sub) (a), (c) and abo e canopy eddy co a iance (EC abo e) (b), (d) du ing a cold pe iod in 30 Ap il– 10 May 2007 (a), (b) and a wa m pe iod in 5–15 June 2007 (c), (d). Do s ep esen median alues o each hou (AC, EC sub) o hal hou (EC abo e) o e he 10-day pe iod. E o ba s ep esen s anda d de ia ions. 3.5 N2O luxes Kale ansuo d ained pea land o es was a small sou ce o N2O du ing he measu emen pe iod om Ap il o June 2007. Mean emission a es a ied be ween 3.2µgNm−2h−1measu ed by he sub-canopy EC ech- nique, 4.5µgNm−2h−1by he au oma ic chambe s, and 6.8µgNm−2h−1by he manual chambe echniques (Fig. 6, Table 1). Independen o he measu ing echnique N2O emis- sions ha dly exceeded 10µgNm−2h−1excep o a sho pe- iod a he beginning o he measu ing campaign when ele- a ed N2O emissions could be de ec ed a leas wi h he em- po ally highly esol ed EC and au oma ic chambe measu e- men s (see Fig. 6). The ele a ed N2O emissions coinside www.biogeosciences.ne /7/1715/2010/ Biogeosciences, 7, 1715–1727, 2010 1722 M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en Table 2. Cumula i e g eenhouse gas CO2, CH4and N2O luxes a he Kale ansuo d ained pea land o es measu ed by eddy co- a iance and au oma ic chambe s du ing he in ensi e measu emen pe iod 25 Ap il–26 June 2007. Componen cumula i e lux, cumula i e lux, GWR100 CO2eq . Cm−2gGHGm−2 CO2ECa(NEE) −102 −373 −373 CO2ECs42.5 156 156 CO2AC 37.7 138 138 CH4AC −0.046 −0.062 −1.30 N2O AC 0.006 0.009 2.77 1Measu emen pe iod 25 Ap il–21 June 2007. 2GWP100 e e s o Global Wa ming Po en ial wi h a 100-yea ime ho izon. Fig. 5. (a) Daily mean luxes o CH4measu ed wi h au oma ic (AC, n=9) and manual (MC, n=16) chambe s, (b) soil concen a ions o CH4a h ee dep hs and in he ambien ai measu ed a he d ained pea land. Fi s column o he igu es ep esen he pe iod o in en- si e measu emen s, he second shows he da a ou side he measu e- men campaign. E o ba s s and o s anda d e o s o he mean. wi h he coldes pe iod (ai emp <0◦C) wi hin he mea- su ing pe iod and a apid inc ease in ai empe a u es up o 15◦C (Fig. 6). A signi ican up ake o a mosphe ic N2O was ne e de ec ed. In gene al, N2O luxes measu ed wi h he EC echnique we e mo e a iable han chambe based N2O luxes which is indica ed by a much highe CV% o 123 as compa ed o alues o CV% o 62.3 and 42.8 by he au o- ma ic and manual chambe s, espec i eley (Table 1). N2O emissions measu ed by he au oma ic chambe s co ela ed nega i ely wi h ai empe a u e ( =−0.50, p<0.01) and soil empe a u es in he li e laye , a 5cm and a 30cm dep hs ( =−0.48, p<0.01; =−0.47, p<0.01; =−0.46, p<0.01), espec i ely, soil mois u e con en ( =−0.46, p<0.01), and Fig. 6. (a) Mean soil N2O luxes measu ed wi h eddy co a iance, au oma ic and manual chambe s, (b) soil concen a ions o N2O a h ee dep hs and in he ambien ai , and (c) mine al ni ogen and o- alni ogen concen a ions in heli e laye o he soil du ing Ap il– Sep embe 2007 a he d ained pea land pine o es . Fi s column o he igu es ep esen he pe iod o in ensi e measu emen s, he sec- ond shows he da a ou side he measu emen campaign. E o ba s s and o s anda d e o s o he mean. CO2 luxes ( =−0.48, p<0.01). Posi i e co ela ions we e ound wi h wa e able dep h ( =0.40, p<0.01) and CH4up- ake ( =0.30, p<0.05). 3.6 CH4and N2O concen a ion in pea p o ile Du ing he in ensi e measu ing campaign om Ap il o June 2007 CH4and N2O concen a ions in he pea p o ile we e close o ambien ai concen a ions o ∼1.8ppm and ∼0.35ppm , espec i ely (Figs. 5b and 6b). In gene al, du - ing he in ensi e measu emen campaign he CH4concen- a ions dec eased (i.e. consump ion) and N2O concen a ion sligh ly inc eased (i.e. p oduc ion) wi h pea dep h in he op- soil. F om July o Sep embe he concen a ions o CH4in deepe pea laye s (22 and 45cm dep h) inc eased ma kedly. The highes concen a ion o 1400ppm was measu ed a 45cm dep h in Sep embe . A he same ime he CH4con- cen a ions in he li e laye we e close o he ambien ai concen a ions and he ne luxes measu ed by manual cham- be s showed ha he soil was s ill a sink o CH4(Fig. 5a and b). Biogeosciences, 7, 1715–1727, 2010 www.biogeosciences.ne /7/1715/2010/ M. K. Pihla ie e al.: G eenhouse gas luxes in a d ained pea land o es du ing sp ing os - haw e en 1723 Ni ous oxide concen a ions a 22cm dep h we e mos o he ime highe han he concen a ion jus below he li - e laye a 5cm (Fig. 6b). Concen a ions a 45cm dep h measu ed du ing July o Sep embe a ied be ween 0.210– 0.240ppm and we e much lowe han a 5 o 22cm dep hs and well below he a mosphe ic concen a ion. 4 Discussion 4.1 CO2 luxes Eddy co a iance (EC) measu emen s abo e he o es canopy e ealed ha he Kale ansuo d ained pea land pine o es was a ne sink o CO2du ing he measu ing pe iod om he end o Ap il o he end o June. The measu emen s below he o es canopy by sub-canopy EC and au oma ic chambe s showed ha he o es loo was a ne sou ce o CO2, how- e e , only a small pa o he ne CO2up ake o he whole o es ecosys em. Du ing ew ainy days in he campaign (in o al 5 days) he Kale ansuo pea land o es u ned om a ne sink o ca bon o a ne sou ce. This inding is in line wi h he s udy by Lohila e al. (2007) whe e hey ound ha an a o es ed bo eal pea land u ned om a ne sink o a sou ce o ca bon du ing ainy days in he summe . To al NEE a he Kale ansuo d ained pea land o es om sp ing o ea ly summe (25 Ap il–21 June, −102gCm−2) is compa able o NEE alues epo ed om bo eal o es s g owing on mine al o pea soils (Suni e al., 2003; Lohila e al., 2007). In his s udy he diu nal a ia ion in he CO2exchange o he soil and o es loo ege a ion was e y small measu ed by he sub-canopy EC and non-exis en measu ed by he au- oma ic soil chambe s. Simila ly small diu nal a ia ion in he o es loo CO2exchange o a bo eal o es ecosys em has been measu ed ea lie by Launiainen e al. (2005) and Kulmala e al. (2008). Howe e , much s onge diu nal a i- a ion in he CO2exchange o soil and o es loo ege a ion has been measu ed in a empe a e o es ecosys em on min- e al soil (Subke and Tenhunen, 2004). In ou s udy he lack o diu nal a ia ion in he CO2exchange o he o es loo may esul om (1) a small pho osyn he ic ac i i y o he o - es loo ege a ion as compa ed o he soil and o es loo espi a ion, o (2) he possibili y o high pho osyn he ic ac- i i y du ing day- ime and a simul aneous inc ease in he soil espi a ion due o empe a u e dependency, which hen com- pensa es o he pho osyn hesis. The ne o es loo CO2 luxes measu ed by sub-canopy EC du ing Ap il–June pe iod compa e well wi h sub-canopy EC measu emen s ca ied ou in a bo eal pine o es (Launiainen e al., 2005), and chambe based measu emen s in o he d ained pea land o es s (Ma - ikainen e al., 1995; Alm e al., 1999). Co ela ion o o es loo CO2 luxes was highes wi h soil empe a u es in 5cm dep h. This shows ha a he he op-soil, ge ing esh li e inpu om ege a ion, is he majo sou ce o CO2as compa ed o he pea body i sel , hus, s imula ed decomposi ion o he pea due o ae a ion by d ainage has al eady diminished. In con as o N2O emissions no inc eases in CO2emis- sions ollowing hawing o he li e laye could be de ec ed. The in e mi en inc ease o CO2emissions in he end o Ap il can be ela ed o a signi ican inc ease in soil and ai empe a u es, howe e , in a pe iod when empe a u es we e ne e below 0◦C. As he measu emen s s a ed a e he i s eeze- haw cycles, i is unclea whe he such eeze- haw induced CO2peaks occu ed a he si e al hough he absence o less p onounced e ec o os - haw cycles on in si u CO2 emissions in o es ecosys ems is also epo ed in he e iew o Ma zne and Bo ken (2008). 4.2 CH4 luxes Au oma ic and manual chambe based measu emen s e- ealed ha he pea land o es was a sink o a mosphe ic CH4du ing he whole measu ing pe iod om end o Ap il o end o June 2007. This means ha he d ainage was deep enough o change he ae a ion s a us and, hus, he condi ions a ou able o me hanogenes o hose a ou able o me han- o ophs. The high in luence o he wa e able dep h on he CH4exchange o pea lands has been obse ed in o he s ud- ies (Ma ikainen e al., 1993, 1995) and is u he e lec ed by he signi ican posi i e co ela ion o CH4up ake a es wi h changes in wa e able dep h du ing he obse a ion pe iod. Maximum up ake a es o >60µgCH4-Cm−2h−1we e sig- ni ican ly highe han obse ed by Ma ikainen e al. (1995) o a d ainded en wi h compa able wa e able dep hs. In a la ge s udy combining da a om d ained and und ained pea land o es s in Finland Minkkinen e al. (2007) ound ha in gene al, und ained si es unc ioned as CH4sou ces whe eas d ained si es unc ioned ei he as CH4sinks o s ill as small sou ces o CH4. In hei s udy he mean CH4up- ake a es a ied om 1 up o 90µgCH4-Cm−2h−1. Fo he Kale ansuo si e Minkkinen e al. (2007) epo ed an an- nual CH4up ake o 0.2gCm−2. A simple linea ex apo- la ion om he cumula i e lux o a ull yea esul ed in an up ake o 0.09gCm−2y −1 o he au oma ic chambe s and 0.06gCm−2y −1 o he manual chambe s. This indica es ha his d ained pea land o es is a signi ican , bu sli gh ly smalle CH4sink as compa ed o bo eal o es s in gene al (−0.15gCm−2y −1) (Du au and Ve cho , 2007). We ound ha CH4was p oduced h oughou sp ing and summe a 22and25cm dep h in he pea p o ile. A hesame ime he ne lux o CH4was nega i e, showing CH4up- ake. This implies ha he Kale ansuo si e was well d ained and he oxic op-laye o he pea was su icien no only o oxidize he CH4p oduced in deepe laye s, bu also o oxi- dize addi ional a mosphe ic CH4. This obse a ion is in-line wi h obse a ions a o he si es, whe e also CH4concen a- ions well abo e a mosphe ic concen a ions we e de ec ed in deepe soil laye s, while soil was s ill ucn ioning as a ne sink o a mosphe ic CH4(Bu e bach-Bahl and Papen, www.biogeosciences.ne /7/1715/2010/ Biogeosciences, 7, 1715–1727, 2010