Atmospheric methane uptake by tropical montane forest soils and the contribution of organic layers

Transcription

1 Biogeochemistry (212) 111: DOI 1.17/s Atmospheric methne uptke y tropicl montne forest soils nd the contriution of orgnic lyers Ktrin Wolf Heiner Fless Edzo Veldkmp Received: 24 June 211 / Accepted: 8 Novemer 211 / Pulished online: 3 Novemer 211 Ó The Author(s) 211. This rticle is pulished with open ccess t Springerlink.com Astrct Microil oxidtion in eroic soils is the primry iotic sink for tmospheric methne (CH 4 ), powerful greenhouse gs. Although tropicl forest soils re estimted to glolly ccount for out 28% of nnul soil CH 4 consumption (6.2 Tg CH 4 yer -1 ), limited dt re ville on CH 4 exchnge from tropicl montne forests. We present the results of n extensive study on CH 4 exchnge from tropicl montne forest soils long n elevtion grdient (1,, 2,, 3, m) t different topogrphic positions (lower slope, mid-slope, ridge position) in southern Ecudor. All soils were net tmospheric CH 4 sinks, with decresing nnul uptke rtes from 5.9 kg CH 4 Ch -1 yer -1 t 1, m to.6 kg CH 4 C h -1 yer -1 t 3, m. Topogrphy hd no effect on soil tmospheric CH 4 uptke. We detected some unexpected fctors controlling net methne fluxes: positive correltions etween CH 4 uptke rtes, minerl nitrogen content of the minerl soil nd with CO 2 emissions indicted tht the lrgest CH 4 uptke corresponded with K. Wolf (&) E. Veldkmp Büsgen Institute-Soil Science of Tropicl nd Sutropicl Ecosystems, Georg-August University of Göttingen, Büsgenweg 2, 3777 Göttingen, Germny e-mil: kwolf@gwdg.de H. Fless Johnn Heinrich von Thünen-Institut, Federl Reserch Institute for Rurl Ares, Forestry nd Fisheries, Institute of Agriculturl Climte Reserch, Brunschweig, Germny fvorle conditions for microil ctivity. Furthermore, we found indictions tht CH 4 uptke ws N limited insted of inhiited y NH 4?. Finlly, we showed tht in contrst to temperte regions, sustntil high ffinity methne oxidtion occurred in the thick orgnic lyers which cn influence the CH 4 udget of these tropicl montne forest soils. Inclusion of elevtion s co-vrile will improve regionl estimtes of methne exchnge in these tropicl montne forests. Keywords Methne uptke Cron dioxide Trce gses Tropicl montne forests Orgnic lyers Elevtion grdient Topogrphic grdient Gs diffusion rrier Introduction Methne (CH 4 ) is trce gs with glol wrming potentil 21 times greter thn tht of cron dioxide (CO 2 ) nd prticiptes in chemicl rections producing tropospheric ozone (Forster et l. 27; Topp nd Pttey 1997). Oxidtion processes re the min sink of tmospheric CH 4, with the rection of hydroxylrdicls (OH) in the troposphere ccounting for pproximtely 9% of nnul CH 4 removl (Prther et l. 21). Although soils cn ct s sinks nd sources for CH 4, microil oxidtion (methnotrophy) in erted uplnd soils is the primry iotic sink for tmospheric CH 4 (Potter et l. 1996). Two forms of methnotrophic oxidtion re known in erted soils.

2 47 Biogeochemistry (212) 111: The first, low ffinity oxidtion, is only found ner methne sources (e.g. in pet soils nd rice pddies) nd is performed y methylotrophic cteri t CH 4 concentrtions higher thn 4 ll l -1 (Conrd 1996;Le Mer nd Roger 21). The second, high ffinity oxidtion, occurs t tmospheric CH 4 concentrtions nd is responsile for tmospheric CH 4 consumption y erted soils (Conrd 1996). The net CH 4 exchnge etween soils nd the tmosphere is controlled y soil physicl nd chemicl properties tht determine gs diffusivity, microil ctivity nd the lnce etween the processes of CH 4 production (methnogenesis) nd uptke (methnotrophy). Gs diffusivity is considered the primry regulting fctor for CH 4 consumption in soils, s the potentil consumption of tmospheric CH 4 y high ffinity methnotrophs generlly exceeds the rte of CH 4 to diffuse from the tmosphere into the soil (Striegl 1993). Soil texture nd soil moisture minly control diffusivity in soils y regulting permeility nd gs trnsport resistnce, s CH 4 diffuses 1 4 times fster in ir thn in wter (Dörr et l. 1993; Topp nd Pttey 1997). Orgnic lyers re considered to ffect CH 4 oxidtion primrily s gs diffusion rriers ecuse they disply little, if ny, CH 4 oxidtion cpcity (Butterch-Bhl et l. 22; Murer et l. 28; Sri et l. 1998; Steinkmp et l. 21). The removl of the orgnic lyer led e.g. to enhnced CH 4 uptke of temperte forest soils (Brumme nd Borken 1999; Sri et l. 1998). Besides sustrte vilility, microil ctivity in soils is regulted y temperture, where incresing tempertures led to higher microil ctivity, nd nutrient vilility (Conrd 1996). N vilility cn indirectly ffect CH 4 uptke, y competition of NH? 4 with CH 4 for methne oxidizing enzymes nd the locking of enzyme functions y products of NH? 4 oxidtion (Schnell nd King 1994). Negtive correltions etween CH 4 uptke rtes nd extrctle NH? 4 or nitrifiction rtes hve een reported in tropicl regions (Mosier nd Delgdo 1997; Veldkmp et l. 21). Furthermore, net CH 4 fluxes in soils re the result of CH 4 production nd CH 4 oxidtion which re regulted y sptil nd temporl vriility of neroic nd eroic microsites in the soil. Aneroic microsites in uplnd soils my occur during wet periods or due to consumption of soil oxygen ecuse of high microil ctivity (Verchot et l. 2). An enhnced occurrence of neroic microsites cn turn soils from net sinks for tmospheric CH 4 into net sources (Dvidson et l. 24; Keller nd Reiners 1994; Keller et l. 25). Most tropicl lowlnd rinforest soils re sinks for tmospheric CH 4 (Kiese et l. 23, 28; Werner et l. 26). Uptke of tmospheric CH 4 in eroic tropicl forest soils is estimted to ccount for out 28% of the glol nnul soil consumption (6.2 Tg yer -1 ) (Dutur nd Verchot 27). Montne forests cover out 9% of the tropics (FRA 2) nd the few studies tht hve een conducted on CH 4 fluxes show tht their soils re generlly CH 4 sinks (Delms et l. 1992; Ishizuk et l. 25; Kiese et l. 28; Puropuspito et l. 26). However, recent study conducted y our group in the sme region s the present study showed tht CH 4 produced in the cnopy y Bromelids my chnge the source-sink lnce of these ecosystems (Mrtinson et l. 21). The extreme vrition in elevtion grdients nd topogrphy re likely to ffect methne fluxes t the soil surfce (e.g. incresing precipittion nd decresing tempertures t higher elevtions, chnges of soil moisture nd the ccumultion of orgnic mtter in the orgnic lyers). We re not wre of study on the potentil contriution of orgnic lyers to oxidize methne in tropicl montne forest soils even though thick orgnic lyers frequently occur in these ecosystems (Edwrds nd Gru 1977). Here, we present the results of n extensive study on CH 4 exchnge from tropicl montne forests soils long n elevtion grdient from 1, to 3, m nd long topogrphic grdients (lower slope, mid-slope, nd ridge) in southern Ecudor. We tested the following hypotheses: (1) net uptke of tmospheric CH 4 decreses with elevtion due to n incresing soil moisture nd decresing temperture, (2) net uptke of tmospheric CH 4 decreses with incresing orgnic lyer thickness, (3) net uptke of tmospheric CH 4 is inhiited y high N vilility nd tht (4) orgnic lyers contriute not or only little to the uptke of tmospheric CH 4 in tropicl montne forests soils. We investigted soil CH 4 exchnge, possile iogeochemicl controls, nd the contriution of orgnic lyers to the tmospheric CH 4 uptke using comintion of field mesurements nd lortory incutions of minerl soil nd orgnic lyers.

3 Biogeochemistry (212) 111: Tle 1 Site chrcteristics nd properties of the orgnic lyer nd the upper minerl soil (Ah-horizon) t the topogrphic positions cross the elevtion grdient 1, m 2, m 3, m Lower slope Mid-slope Ridge Lower slope Mid-slope Ridge Lower slope Mid-slope Ridge Coordintes 4 6 S, W 3 58 S, 79 4 W 4 6 S, W Elevtion.s.l. (m) 99 1,2 1,8 2,1 2,8 3, Air temperture ( C) Precipittion (mm yer -1 ) 2,23 1,95 4,5 Orgnic lyer thickness (cm) 2.5 (.7) 2.6 (1.)A 6.5 (1.9)A 4. (.9) 12.2 (3.1)B 24. (2.9)B 6.6 (.7) 14.2 (1.6)B 22.2 (3.4)B Orgnic lyer ph (H 2 O) 4.9 (.5) 4.8 (.6) 3.7 (.1) 4.6 (.2) 4.8 (.5) 3.4 (.1) 4. (.3) 3.7 (.1) 3.6 (.1) C:N 18.5 (.7)A 18.6 (1.5)A 2.5 (.7)A 15.7 (.8)B 15.6 (.6)A 21.1 (.7)A 2.5 (.8)A 23.9 (1.4)B 26.7 (1.3)cB Totl P (mg g -1 ) 1.1 (.1) 1.2 (.1)A.6 (.2)A 1.2 (.1) 1.1 (.2)A.5 (.)AB.8 (.1)A.6 (.1)B.4 (.2)B Minerl soil (Ah) ph (H 2 O) 5. (.3)A 5.1 (.4)A 4.2 (.1) 4.6 (.2)AB 4.7 (.4)AB 3.7 (.1) 3.7 (.1)B 3.7 (.)B 3.7 (.1) C:N 16.3 (.6) 15. (.7) 21.3 (1.2)AB 13.2 (1.1) 14.1 (1.5) 17. (.8)A 14.2 (.2) 18. (1.1) 22.3 (1.)B Totl P (mg g -1 ).7 (.2).8 (.2)A.6 (.1).5 (.1). 7 (.1)AB.2 (.).3 (.).2 (.)B.1 (.) Snd (%) 42.4 (8.6) 46.9 (8.9) 33. (7.3) 26.7 (2.1) 25.8 (5.3) 23.9 (2.9) 21.8 (3.9) 32.1 (6.7) 36. (4.8) Silt (%) 3.6 (7.5)A 35.8 (8.2) 35.6 (5.8)A 55.3 (3.)B 53.1 (3.8) 6.9 (1.8)B 57.5 (.3)B 49.7 (2.5) 5. (2.)AB Cly (%) 26.9 (2.6) 17.3 (2.3) 31.5 (2.1)A 18.1 (2.1) 21.1 (2.1) 15.2 (2.1)B 2.7 (3.6) 18.2 (4.3) 14.1 (3.3)B Moser et l Mens (± SE, n = 6) within rows followed y different letter indicte significnt differences mong topogrphic positions t ech elevtion (smll letters) nd differences mong elevtions t ech topogrphic position (cpitl letters) (two-wy ANOVA with Tukey HSD test t P B.5)

4 472 Biogeochemistry (212) 111: Methods Study re The study ws crried out long n elevtion grdient from 1, to 3, m within nd close to Podocrpus Ntionl Prk. The prk is locted in the Cordiller del Consuelo, which forms prt of the estern chin of the Andes in southern Ecudor. Three study sites were selected (Tle 1): premontne rinforest, from 99 to 1,2 m.s.l. (herefter the 1, m site ), lower montne forest, from 1,8 to 2,1 m.s.l. (herefter the 2, m site ) nd upper montne forest, from 2,8 to 3, m.s.l. (herefter the 3 m site ). At ech study site (1,, 2,, 3, m) we estlished 18 plots t three lndscpe positions: lower slope, mid-slope nd ridge. Annul precipittion showed little sesonlity (Emck 27). For most of the study re, Pleozoic metmorphic schist nd sndstones with some qurtz veins form the prent mteril for soil development. However, t the 1, m site, soil prent mteril of hlf of the plots consists of deeply wethered grnitic rock of the Jurssic Zmor grnitoide formtion (Litherlnd et l. 1994). Orgnic lyer thickness incresed with elevtion nd from lower slope positions to ridge positions (Tle 1), ut displyed lrge sptil vrition, rnging from to 3 cm t the 1, m site, nd from.5 to 75 cm t the 2, nd 3, m sites. At the ridge positions of the 2, nd 3, m sites nd to some extend t the ridges of the 1, m site, thick orgnic lyers led to Humic nd Histic chrcteriztion of soil types. Soils t well-drined positions of the 1, nd 2, m sites clssified s Eutric or Dystric Cmisols. Soils with hydromorphic properties (Stgnic Cmisols or Plnosols) were found on the ridges of the 2, m site nd t ll topogrphic positions of the 3, m site (Wolf et l. (211)). Experimentl design At ech elevtion we estlished 18 plots, six t ech lndscpe position (lower slope, mid-slope nd ridge positions). All plots were situted in mture, closedcnopy forest tht ws representtive for the respective topogrphic position. In ech plot we instlled four chmer ses for gs exchnge mesurements in rndom design long two orthogonl trnsects, in totl 216 chmers (54 plots 9 4 chmers per plot). Chmer ses were instlled 6 months prior to the first mesurements, inserting them pproximtely.2 m into the soil nd remined in plce until the end of the smpling period. We used n inserting depth of only.2 m in order to minimize root disturnce nd ecuse the dense cnopy provided n efficient wind protection. Chmer lids were equipped with Luer-Lock smpling port nd vent for pressure equilirtion. Methne, soil respirtion, soil moisture content, soil temperture, nd soil minerl N were mesured i-monthly over period of 1 yer (My 28 My 29). Site chrcteriztion Soil chrcteristics of the plots were determined from 41 soil profiles ( m) t the eginning of the study (28). We dug soil pit t ech plot except for very similr plots which were locted t the sme site, t the sme slope position nd which hd the sme inclintion. For such similr plots we smpled only one common soil pit. All together 41 soil profiles were nlyzed to chrcterize soil properties. Bulk densities of the minerl soil nd the orgnic lyers were determined ccording to Blke nd Hrtge (1986). For chemicl nlysis, smples from the orgnic lyer (s pooled smple covering the whole depth of the orgnic lyer) nd from the Ah horizon were ovendried (45 C for 3 dys), sieved nd stored in plstic gs until ir trnsported to Germny. Totl cron (C tot ) nd totl nitrogen (N tot ) were determined using ground smples nd CNS Elementl Anlyzer (Elementr Vrio EL, Hnu, Germny). Totl P (P tot ) nd totl se ctions were mesured using Inductively Coupled Plsm Atomic Emission Spectrometer (ICP AES; Spectro Anlyticl Instruments, Kleve, Germny) fter pressure digestion in 65% HNO 3. Soil ph ws mesured in suspension of soil in H 2 O with 1:1 nd 1:1 rtio for sieved minerl soil nd orgnic lyer, respectively. Soil texture of the Ah horizon ws determined y sedimentry frctiontion of the cly frction (25 C, 21 h, 3 cm fll height) following the Attererg method fter wet sieving of the snd nd corse silt frction (63, 2, 63 nd 2 lm) nd destruction of orgnic mtter with 3% H 2 O 2 nd Fe oxides with 4% N Dithionite Citrte solution (Schlichting et l. 1995). Fine nd mid silt frctions (2 2 lm) were determined y drying nd weighing of the remining mteril in the Attererg Cylinders

5 Biogeochemistry (212) 111: fter the elutrition of the cly frction. We did not seprte fine nd mid silt frctions. Soil moisture content, soil minerl N, nd soil temperture were determined t the depth of to.5 m of the surfce (orgnic lyer or minerl soil), within one meter distnce of the chmer ses concomitntly with every gs smpling. We expressed soil moisture s grvimetric moisture content. Soil minerl N (mmoni (NH 4? ) nd nitrte (NO 3 - )) ws determined on composite soil smple, which consisted of four susmples per plot. In the field, prt of the composite smple ws dded to polyethylene ottle contining 15 ml of.5 molr potssium sulfte (K 2 SO 4 ) solution. The remining prt of ech smple ws stored in plstic g for determintion of grvimetric moisture content in the lortory fter drying for 24 h t 15 C. Extrction of soil minerl N ws finished within 12 h fter field smpling to void shifts in NH 4? nd NO 3 - concentrtions in the soil over time (Arnold et l. 28). Extrcts were filtered through filter ppers (4 lm nominl pore size) fter 1 h of gittion, nd frozen immeditely fter drop of chloroform ws dded to prevent further N turnover. Smples remined frozen during trnsporttion y ir to Germny, where nlysis ws conducted. Ammonium nd NO 3 - concentrtions of the extrcts were nlyzed using continuous flow injection colorimeter (Cenco/Sklr Instruments, Bred, Netherlnds) (Arnold et l. 29). Field mesurement of net CH 4 exchnge nd soil respirtion We mesured net exchnge rtes of CH 4 nd CO 2 imonthly nd once dy over time period of 1 yer (My 28 My 29). We used sttic closed chmers with se re of.44 m 2 nd closed chmer volume of 13 l. Gs smples were tken 2, 14, 26, nd 38 min following chmer closure using pre-evcuted glss continers (6 ml). Pressure ws checked with mnometer during smpling to ssure tht continers hd een evcuted properly nd were filled to tmospheric pressure. Gs smples were nlyzed within 48 h following smpling using gs chromtogrph (Shimdzu GC-14B, Duisurg, Germny) equipped with flme ioniztion detector (FID) nd n electron cpture detector (ECD) comined with n utosmpler (Loftfield et l. 1997). Gs concentrtions were clculted y compring integrted pek res of smples with four stndrd gses (1.5, 2., 1 nd 2 ll l -1 stndrd concentrtions of CH 4 ; nd 35, 7, 15, 5 ll l -1 stndrd concentrtions of CO 2 ; Deuste Steiniger GmH, Mühlhusen, Germny). Methne exchnge rtes were clculted from the liner chnge of CH 4 concentrtion in the chmer ir over time, while CO 2 emissions were clculted y fitting oth liner nd qudrtic regression models if CO 2 concentrtions incresed symptoticlly (Wgner et l. 1997). Positive flux rtes represent net emissions nd negtive vlues represent net uptke rtes y soils. Bsed on the Akike Informtion Criterion the qudrtic model ws found to e sttisticlly more dequte nd therefore pplied in out 5% of the cses. By pplying this dt screening nd model fitting, we minimized underestimtions of CO 2 emissions tht my hve occurred y using liner models uncriticlly with sttic chmer flux dt (Livingston et l. 26). Due to the low sesonlity of gs fluxes, nnul fluxes were pproximted from the yerly verge of the six i-monthly mesurements, ssuming constnt fluxes per dy. Potentil uptke of tmospheric CH 4 nd soil respirtion t different soil depths To determine the potentil uptke of tmospheric CH 4 nd soil respirtion t different soil depths nd to determine the influence of temperture on gs flux rtes, we conducted lortory incution experiment. Undistured smples from minerl soils nd orgnic lyers were tken from selected soil profiles using stinless steel cores (1 cm 3 ). We smpled 12 soil profiles of the minerl soil t the lower slope positions of the 1, m site, where no orgnic lyer ws present. We smpled six minerl soil profiles from soils on deeply wethered grnite nd six from young, shllow soils over schist. The minerl soil profiles consist of smples from four different soil depths: 5, 5 1, 1 15 nd 15 2 cm. Furthermore, we smpled 15 orgnic lyer profiles from the ridge positions of the three elevtions. The orgnic lyers profiles were smpled in 5-cm depth intervls from the surfce to the minerl soil nd include two depth of the underlying minerl soil ( 5 nd 5 1 cm). We smpled three orgnic lyer profiles over wethered grnite t 1, m, seven orgnic lyer profiles t 2, m nd five orgnic lyer profiles t 3, m. Additionlly,

6 474 Biogeochemistry (212) 111: we took distured smples of ech profile depth to determine grvimetric moisture content fter drying for 24 h t 15 C. The undistured smples were trnsported to the lortory of the Technicl University in Loj nd were incuted the next dy. We incuted soil cores t field moisture nd two different tempertures (15 nd 25 C) under mient ir in closed glss flsks (.4 l) with septum smpling port. Soil cores were pre-incuted for 24 h t the respective incution temperture. If necessry, soil moisture ws re-djusted to the initil moisture content prior to incution, y weighing soil cores nd dding distilled wter. After 9 min of incution, gs smple (1 ml) ws tken from the hedspce of the flsk through septum with gs tight syringe (SEG Interntionl, Ringwood, Austrli), nd the concentrtion of CH 4 nd CO 2 ws mesured with the gs chromtogrph descried ove. Methne fluxes nd CO 2 emissions were clculted s the difference of the gs concentrtion mesured fter the incution of 9 min (T 1 ) nd the initil concentrtion in fresh ir efore the incution flsk ws closed (T concentrtion, fresh mient ir). The gs fluxes were relted to the soil core surfce re (24 cm 2 ) nd expressed s flux per squre meter. As CH 4 uptke ctivity is strongly controlled y sustrte vilility, this procedure proly resulted in n underestimtion of the CH4 uptke, however we were minly interested in the reltive differences in uptke ctivity etween soil cores t different depths, for which this procedure ws sufficient. After incution, soil mteril ws dried t 45 C, ground nd nlyzed for ph, C tot,n tot,p tot nd se ctions s descried for the soil smples ove. CH 4 concentrtion in soil ir of orgnic lyers To determine in situ soil ir CH 4 concentrtion t different orgnic lyer depths, we smpled soil ir in 1-cm steps from the minerl soil to the surfce t the sme smpling points t 2, nd 3, m selected for the incution experiment. We mesured three profiles per smpling point using steel tue (1.8, 6 8 cm, Alltech, Unterhching) nd pre-evcuted glss continers (2 ml), fter evcuting the tuing with syringe. Gs concentrtions were mesured using the gs chromtogrph descried ove. Mesurements were crried out during reltively dry period (Octoer Novemer 21) nd during wet period (Ferury Mrch 211). Sttisticl nlyses Sttisticl nlysis of dt from the field mesurements ws crried out on the plot men of trce gs fluxes (verge of four chmers). Liner mixed effects models (LME) were pplied to test the effect of: (1) Elevtion nd topogrphy on time series dt (i.e. i-monthly mesurements of trce gs fluxes, soil moisture nd soil minerl N), with elevtion nd topogrphic position s fixed effects, nd replicte plots nested in time s rndom effect. (2) Continuous vriles (soil moisture nd soil minerl N) on temporl vrince of the gs fluxes, with the continuous vrile s fixed effect nd replicte plots nested in time s rndom fctor. (3) Continuous vriles on the sptil vriility of nnul gs fluxes, including topogrphic positions nested into the elevtions s sptil rndom fctor. (4) Temperture nd soil texture on trce gs fluxes of the incuted smples, with profile depth s rndom fctor. (5) Wet nd dry seson, soil depth nd elevtion on CH 4 soil ir concentrtions, including the hierrchicl smpling structure s sptil rndom fctor. (6) We used LME for time series dt ecuse they ccount for temporl correltion mong oservtions on the sme experimentl unit (Piepho et l. 24) nd we used LME to test sptil correltions to ccount for the hierrchicl study design y including different error vrinces of the different sptil scles (Crwley 29). LME were tested for utocorreltion effects y including first order utoregressive structure nd for heteroscedsticity of residul vrince y including vrince functions (Bliese nd Ployhrt 22). Model residuls were tested for norml distriution nd dt ws trnsformed if necessry. For testing the elevtion nd topogrphy effects on dt tht were mesured once (i.e. soil chrcteristics) we used two-wy nlysis of

7 Biogeochemistry (212) 111: vrince (ANOVA). The reltionships mong soil chrcteristics nd nnul trce gs fluxes were ssessed y liner regression on the mens of six plots for ech topogrphic position (n = 9; 3 topogrphic positions of three elevtions). The effects of soil depth on trce gs fluxes were tested using one-wy ANOVA. For multiple comprison fter LME nd ANOVA Tukey HSD test ws pplied. Correltions of trce gs fluxes from the incuted smples with continuous vriles (i.e. cron concentrtions) were exmined using generlized lest squre models (GLS) on the elevtion men of ech profile depth. GLS models were tested for utocorreltion effects of profile depth y including sphericl correltion structure. Mens with ±1 stndrd error re given in the text. Effects were ccepted s sttisticlly significnt if P B.5. Anlyses were conducted using R version (R Development Core Tem 29) nd STAT- ISTICA Service Pck 8 (SttSoft Inc., Tusl, USA). Results Soil net CH 4 fluxes long elevtion nd topogrphic grdients Atmospheric CH 4 uptke of the soil decresed with incresing elevtion (P \.1), wheres topogrphy did not ffect tmospheric CH 4 uptke. Soil net CH 4 fluxes did not disply cler sesonlity (Fig. 1) nd soils of ll elevtions nd topogrphic positions were net sinks for tmospheric CH 4 with nnul men uptke rtes rnging from.6 ± 1.2 kg CH 4 C h -1 yer -1 t 3, m to 5.9 ± 1. kg CH 4 C h -1 yer -1 t 1, m (Tle 2). Positive CH 4 fluxes (emissions) occurred in 5.3% of ll mesurements nd incresed with elevtion from 2.1% t 1, m nd 5.2% t 2, m to 8.9% t 3, m. Vriility of fluxes ws highest t the lower slope position t 3, m where 22% of CH 4 fluxes mesured were positive. We found no effect of different soil prent mteril (grnite vs. schist) on net tmospheric CH 4 uptke rtes (P =.4) (Tle 3). To determine the effect of elevtion nd Fig. 1 Annul course of soil net CH 4 exchnge rtes nd grvimetric soil moisture (mens ± SE, n = 6) of the three topogrphic positions (lck downwrd tringles lower slope, lck circles midslope, lck upwrd tringles ridge) cross the elevtion grdient (1,, 2,, 3, m). Negtive vlues represent CH 4 uptke rtes CH 4 flux ( g C m -2 h -1 ) m 2 m 3 m -12 My 8 July 8 Sept 8 Nov 8 Jn 9 Mr 9 My 8 July 8 Sept 8 Nov 8 Jn 9 Mr 9 Dte 1 m 2 m 3 m Grvimetric moisture content (g g )

8 476 Biogeochemistry (212) 111: Tle 2 Men nnul soil CH 4 nd CO 2 fluxes (mens ± SE) nd prllel mesured prmeters of the top soil ( 5 cm) t the topogrphic positions cross the elevtion grdient 1, m 2, m 3, m Lower slope Mid-slope Ridge Lower slope Mid-slope Ridge Lower slope Mid-slope Ridge CH 4 (kg C h -1 yer -1 ) -5.5 (.7)A -5.4 (.9)A -5.9 (1.)A -2.3 (.3)B -4.3 (.9)A -2.7 (.3)B -.6 (1.2)B -1.6 (.4)B -1. (.1)C CO 2 (Mg C h -1 yer -1 ) 1.3 (.8)A 1.3 (.1)A 9.8 (.9)A 8. 8 (.4)A 7.6 (.6)B 6.7 (.7)B 6.4 (.4)B 5. 7 (.7)B 3.7 (.5)C Soil prmeters in 5 cm Grvimetric moisture.1 (.2).6 (.1) 1.7 (.3) NA NA NA NA NA NA minerl soil (g g -1 ) Grvimetric NA 1.7 (NA) 1.8 (.3)A 1.8 (.3)B 2.1 (.3) 4.3 (.1)B 2.1 (.2) 2.9 (.3) 4.5 (.3)B moisture orgnic lyer (g g -1 )? NH 4 (kg N h -1 ), 2.2 (.4)A 1.8 (.5) 1.5 (.5).8 (.2)B.9 (.3).9 (.2).6 (.1)B.7 (.1).8 (.1) (kg N h -1 ),.4 (.1)A.4 (.1)A.6 (.3)A.4 (.1)A.3(.1)A.4 (.2)B.B.2 (.1)B.3 (.1)B NO 3 - Mens (± SE, n = 6) within rows followed y different letter indicte significnt differences mong topogrphic positions t ech elevtion (smll letters) nd differences mong elevtions t ech topogrphic position (cpitl letters) (Liner mixed effects models with Tukey HSD test t P B.5) Prmeters from orgnic lyers or minerl soil slope position on soil moisture in the top.5 m of the soil we seprted results from the orgnic lyer nd the minerl soil. Grvimetric moisture content of minerl soils without orgnic lyer, which were only found t the 1, m site, ws lowest t the mid-slope position nd highest t the ridges (P =.3). Grvimetric soil moisture in the top.5 m of the orgnic lyers incresed with elevtion (P \.1) nd from the lower slope positions to the ridges t 2, nd 3, m (P \.1). We did not detect reltionship etween i-monthly mesured net CH 4 flux rtes nd soil moisture cross elevtions nd topogrphic positions, neither in minerl soil nor in the orgnic lyer. In ddition, we found no significnt correltion etween the nnul CH 4 uptke nd the men soil moisture content (dt not shown). Atmospheric CH 4 uptke of the soil incresed with incresing soil temperture long the elevtion grdient (P \.1, R 2 =.89), NO - 3 content (P \.1, R 2 =.72) nd NH? 4 content (P \.2, R 2 =.72) of the top.5 m (minerl soil or orgnic lyer) nd with incresing P (tot) (P \.1, R 2 =.66) nd ph vlue of the upper minerl soil (P =.2, R 2 =.51). Smller scle soil temperture chnges within the elevtion levels, orgnic lyer thickness, P tot, C:N rtios nd the ph vlue of the orgnic lyer did not ffect tmospheric CH 4 uptke. Annul CO 2 emissions nd nnul CH 4 uptke y soils with nd without orgnic lyers were correlted positively cross ll elevtions (P \.1, R 2 =.66, Fig. 5). When considering only gs fluxes from the soils covered y n orgnic lyer, this correltion ws even stronger (P =.2, R 2 =.74), while the correltion in minerl soils without orgnic lyer ws not significnt (P =.13, R 2 =.13). Potentil uptke of tmospheric CH 4 t different soil depths All incuted, undistured smples from the minerl soil profiles (profiles of the 1, m site, which hd no orgnic lyer) were net sinks for tmospheric methne (Fig. 2). Methne uptke rtes rnged from 3.5 to 6.5 lg CH 4 C m -2 h -1 with the highest uptke rte of tmospheric CH 4 t soil depth of 5 1 cm (P =.5). The incuted minerl soil profiles from the corse textured grnite sustrte displyed higher potentil uptke of tmospheric CH 4 thn the profiles from the fine textured schist sustrte (P \.1) (Fig. 2). Incuted orgnic lyer smples generlly cted s sinks for

9 Biogeochemistry (212) 111: Tle 3 Soil chrcteristics nd nnul flux rtes of CH 4 nd CO 2 (mens nd SE) of soils from two sustrtes (schist, grnite) of the topogrphic positions t 1, m Lower slope Mid-slope Ridge Schist Grnite Schist Grnite Schist Grnite Snd (%),c (2.36) 6.66 (5.25) (1.43) (3.92) (1.72) (.87) Silt (%),c 47.5 (3.84) 14.2 (.4) (1.4) (1.73) 48.5 (.99) 23.9 (3.47) Cly (%),c (1.48) (5.29) (2.19) 16.7 (4.43) (.73) (2.6) ph (H 2 O),c 5.4 (.2) 4.5 (.4) 5.9 (.3) 4.3 (.3) 3.9 (.1) 4.5 (.1) C:N,c 16.3 (.6) 16.4 (1.2) 14.5 (1.5) 15.6 (.2) 19. (.9) 23.7 (.8) Totl P (mg g -1 ),c 1.1 (.1).3 (.2) 1.4 (.2).3 (.1).8 (.4).4 (.1) NH? 4 (kg N h -1 ),d 2.8 (.6) 1.5 (.2) 1. (.4) 2.7 (.3) 2.7 (.2).4 (.1) NO - 3 (kg N h -1 ),d.4 (.2).3 (.3).2 (.4).5 (.2) 1.1 (.3).1 (.1) CH 4 (kg C h -1 yer -1 ) -5. (.7) -6. (1.4) -5.1 (1.3) -5.6 (1.5) -8. (.5) -3.8 (.8) CO 2 (Mg C h -1 yer -1 ) 1.1 (3) 1.4 (1.6) 8.7 (.3) 11.8 (2.7) 11.4 (.5) 8.25 (1.2) Mens (± SE, n = 3) within rows followed y different letter indicte significnt differences mong the sustrtes (smll letters) (one-wy ANOVA t P B.5) Mens (± SE, n = 3) within rows followed y different letter indicte significnt differences mong the sustrtes (smll letters) (Liner mixed effects models for time series dt t P B.5) c Prmeters were tken from the Ah horizon d Prmeters were tken from the top 5 cm of the soil (including minerl soil or orgnic lyer) Profile depth (cm) CH 4 -flux( gcm -2 h -1 ) Fig. 2 Men CH 4 uptke rtes of undistured field moist soil smples t different depths of the minerl soil profiles, which hd no orgnic lyer in schist (open symols) nd grnite (lck symols) mteril t 1, m, incuted t 25 C in the lortory. Mens (±SE, n = 6) within one depth followed y different letter indicte significnt differences mong the sustrtes (Kruskl Wllis-test t P B.5) tmospheric CH 4, except for the smples from the top 5 cm, which contined minly roots nd fresh plnt residues nd emitted CH 4 in 65% of the cses. All smples of the uppermost 5 cm of the orgnic lyer t the 1, m nd the 3, m sites emitted methne, while t the 2, m site only two of the seven uppermost profile smples emitted methne (Fig. 3). The incuted orgnic mteril from the 1, m site showed potentil CH 4 uptke rnging from -.2 to lg CH 4 C m -2 h -1. The potentil CH 4 exchnge in the orgnic mteril from 2, m rnged from?3.1 (emission!) to lg CH 4 C m -2 h -1 nd from?12.9 to lg CH 4 C m -2 h -1 in the orgnic mteril from 3, m. Potentil CH 4 uptke of tmospheric CH 4 in the minerl soil underlying the orgnic lyers ws higher t 1, m thn t 2, or 3 m (P \.1), with uptke rtes rnging from 7.5 to 7. lg CH 4 C m -2 h -1 t 1, m, from.8 to 24.4 lg CH 4 C m -2 h -1 t 2, m nd from -3.4 to 33.4 lg CH 4 C m -2 h -1 t 3, m. We found cler strtifiction of the potentil tmospheric CH 4 uptke rtes within the smpled orgnic lyer profiles. The CH 4 uptke rtes in the orgnic lyer profiles from ll elevtions incresed with incresing profile depth (P \., Fig. 3). At the 1, m site, potentil CH 4 uptke of the underlying minerl soil ws higher thn the orgnic mteril (P B.5, Fig. 3). At the 2, m nd the 3, m sites potentil tmospheric CH 4 uptke ws lrgesttthedeepestorgnic lyer overlying the minerl soil (Fig. 3, c). Orgnic lyer thickness did not ffect the potentil tmospheric CH 4 uptke of the underlying minerl soil.

10 478 Biogeochemistry (212) 111: Profile depth (cm) () () (c) CH 4 - flux ( g C m -2 h -1) Fig. 3 Men CH 4 fluxes (±SE) of undistured, field moist soil smples from different depths of the orgnic lyer (open symols) nd the minerl soil (lck symols) incuted t 25 C in the lortory. The smples were tken from the 1, m site (n = 3), the 2, m site (n = 7) nd c the 3, m site (n = 5) Cron dioxide emissions from the incuted minerl soil nd orgnic lyer profiles were lrgest in the top soil smples, rnging from 31.1 to 68.2 mg CO 2 C m -2 h -1 in the minerl soil nd from 28.9 to 14.9 mg CO 2 C m -2 h -1 in the orgnic lyer. Soil respirtion decresed with soil depth of ll profiles (dt not shown). Temperture mnipultion ffected soil respirtion from smples of the orgnic lyer from ll elevtions (P \.1), wheres it did not ffect soil respirtion of the minerl soil smples (P =.2). A temperture increse from 15 to 25 C resulted in douling of the CO 2 emissions rtes from smples of the orgnic lyer s shown y the douling of the slope of the regression function (Fig. 4). Temperture mnipultion did neither ffect CH 4 uptke rtes of smples from the orgnic lyer profiles during the incution (Fig. 4), nor did it ffect the CH 4 uptke rtes of the minerl soil profiles (dt not shown). Throughout the elevtions, potentil CH 4 uptke of the incuted minerl soil profiles nd orgnic lyer profiles incresed with decresing CO 2 emissions (P \.1, R 2 =.45, Fig. 5). Throughout the elevtion grdient, potentil tmospheric CH 4 uptke of the incuted orgnic mteril did not correlte with C tot content, ction concentrtion, grvimetric moisture content, or ph-vlues. However, potentil CH 4 uptke of orgnic mteril from the 1, m site correlted positively with P tot content of the orgnic mteril (P =.3, R 2 =.9). In contrst, potentil CH 4 uptke of orgnic mteril of the 2, m site incresed with decresing C tot content (P =.4; R 2 =.62), clcium (C 2?, P =.1; R 2 =.78) nd mgnesium (Mg 2?, P =.1; R 2 =.78) concentrtions nd phvlues (P \.1; R 2 =.91) of the orgnic mteril. CH 4 concentrtions in the soil ir of orgnic lyers In situ methne concentrtion in the soil ir decresed with soil depth (P \.1 for oth elevtions), oth during wet nd dry periods, with lowest CH 4 concentrtions ove the minerl soil (Fig. 6). Methne concentrtions were higher during the dry period mesurements compred to the wet period mesurements (P \.1 for oth elevtions)(fig. 6). Elevtion did not ffect CH 4 concentrtion in the soil ir of the orgnic lyers neither during the wet nor during the dry period. CH 4 concentrtion in soil ir ws lwys lower thn the tmospheric concentrtion nd there ws no evidence of in situ net production of CH 4 within the soil profile. Discussion Net exchnge of CH 4 under field conditions Annul CH 4 uptke rtes under field conditions t the 2, nd 3, m sites were comprle to those

11 Biogeochemistry (212) 111: CO 2 - flux (mg C m -2 h -1 ) () CO 2 -fluxt25 C(mgCm -2 h -1 ) Fig. 4 Reltion etween men (±SE) trce gs fluxes of undistured, field moist smples of the men orgnic lyer profile from ech elevtion (including the underlying minerl soil) incuted t 25 C (open symols) nd 15 C (lck -2-1 CH 4 -flux( gcm h ) () CH 4 -fluxt25 C( gcm -2 h-1) symols) for CO 2 fluxes (y = 5.5?.5 9 x, n = 24, P \.1, R 2 =.9) nd CH 4 fluxes (y =-1.4?.9 9 x, n = 24, P \.1, R 2 =.85) Exchnge of CH -C ( g m -2 h -1 ) () Emissions of CO 2 - C (mg m-2 h-1) Exchnge of CH 4 - C(mgm -2 h-1 ) () Log of CO 2 - C emissions (mg m-2 h -1) Fig. 5 Reltion etween soil CO 2 emission rtes nd soil CH 4 uptke rtes (mens ± SE) for flux rtes mesured in the field t different topogrphic positions (lck downwrd tringles lower slope, lck circles mid-slope, lck upwrd tringles ridge) cross the elevtion grdient (lck: 1, m; full grey: 2, m; open: 3, m) (y =-21.38?.8 9 x, n = 9, P =.5, R 2 =.66), nd for undistured, field moist soil smples t different depths of the orgnic mteril (open symols) nd the minerl soil (lck symols), incuted t 25 C in the lortory (y =-84.49? x, n = 32, P \.1, R 2 =.45). Minerl soil with overlying orgnic lyer lck circles, minerl soil without overlying orgnic lyer lck tringles reported for tropicl montne forest soils t comprle elevtion in Indonesi (1.5 nd 3.3 kg CH 4 C h -1 yer -1 ) (Ishizuk et l. 25; Puropuspito et l. 26). Annul CH 4 fluxes t the 1, m site were lrger thn previously reported for other premontne forests in Afric (Delms et l. 1992), Austrli (Kiese et l. 28), Indonesi (Puropuspito et l. 26) nd Chin (Werner et l. 26) nd were lso lrger thn those reported from lowlnd forests in Cost Ric (Keller nd Reiners 1994; Reiners et l. 1998) nd Brzil (Keller et l. 25). The kg CH 4 Ch -1 yer -1 CH 4 uptke rtes t our 1, m site were lso on the high end of the methne consumption rnges for temperte ecosystems (Fiedler et l. 25; Smith et l. 2). Dutur nd Verchot (Dutur nd Verchot 27), gve n estimte of 4.3 kg CH 4 Ch -1 yer -1 for the temperte forest iome. Both the incresed incidence of net CH 4 emission rtes with incresing elevtion nd generl decrese in CH 4 uptke rtes resulted in decresed nnul net CH 4 uptke with elevtion. This trend ws mirrored y the mximum CH 4 uptke rtes mesured for the incuted soil smples nd y the less pronounced decline of soil ir CH 4 concentrtions in the orgnic lyers t the drier 3, m site. The lck of influence of slope position on CH 4 uptke tht we detected supports results from forest sites in Cost Ric where no differences in CH 4 uptke etween topogrphic positions ws detected (Reiners et l. 1998). Furthermore, we found no evidence tht sesonl wterlogging turned net methne-oxidizing tropicl forest soils into net CH 4 sources s reported for sesonl lowlnd rin forests in Brzil (Dvidson et l. 24; Verchot et l. 2). Sesonl vrition in soil wter

12 48 Biogeochemistry (212) 111: Orgnic lyer depth (cm) () () dry wet CH4 concentrtion of soil ir (pp) condition in our study re ws rther limited ecuse of the lrgely non-sesonl climte. CH 4 fluxes nd their controls Fig. 6 Men soil ir CH 4 concentrtions (±SE) t different depths of the orgnic lyer during wet period (lck symols) nd dry period (open symols), from the 2, m site nd the 3, m site In contrst to the generl ide tht soil moisture content is n importnt driver for CH 4 fluxes (Butterch-Bhl et l. 24; Curry 27; Dvidson et l. 24; Keller nd Reiners 1994), neither temporl nor sptil vriility of net CH 4 exchnge could e explined y it. The min reson for this lck of correltion etween soil moisture content nd net exchnge of tmospheric CH 4 is proly the sence of pronounced dry seson nd the resulting smll vrition in soil moisture content. Furthermore, it is unlikely tht moisture contents in the highly porous, corse mteril of the upper orgnic lyer re high enough to restrict CH 4 diffusion into the soil. Our incution of minerl soil from different geologicl sustrtes corroortes findings tht corser soil texture leds to lrger potentil tmospheric CH 4 uptke s result of higher gs diffusivity (Dörr et l. 1993; Sri et l. 1997). However, this texture effect did not explin the sptil vriility of tmospheric CH 4 uptke oserved in our field mesurements. We found no evidence tht lndscpe differentition using the geologicl sustrte or soil texture would improve estimtes of soil CH 4 uptke in this region. This finding contrsts with studies in temperte forest ecosystems where soil texture ws identified s min control of soil uptke of tmospheric CH 4 (Sri et l. 1997). We found no indiction tht mmonium inhiition of methnotrophic cteri ws control mechnism for CH 4 uptke. On the contrry, we detected positive correltions etween soil minerl N content nd nnul CH 4 uptke which suggests tht methnotrophic ctivity my hve een N limited. This finding supports the findings of other studies crried out in severl ecosystems showing lower oxidtion ctivity due to nitrogen limittion (Aronson nd Helliker 21; Bodelier nd Lnroek 24). Furthermore, positive correltions with P tot content, ph-vlue of the Ah horizon nd field CO 2 emissions suggest tht higher CH 4 uptke occurred t those loctions where conditions were generlly more fvorle for microil ctivity. Strtifiction of potentil uptke of tmospheric CH 4 in minerl soil nd orgnic lyers The detected strtifiction of tmospheric CH 4 uptke in the incuted minerl soil profiles, with highest potentil CH 4 uptke t soil depth of 5 1 cm is in greement with findings of studies from temperte nd surctic forest soils (Schnell nd King 1994). The susurfce loction of mximum CH 4 oxidtion is proly due to edphic fctors in the surfce lyer, tht re suoptiml or inhiitory for the growth of CH 4 oxidizing cteri, e.g. occsionl droughts or the occsionl occurrence of neroic microsites. The negtive reltionship etween CO 2 production nd tmospheric CH 4 consumption of our incuted soil smples, my e the result of CH 4 inhiition y NH 4? relesed due to rpid N-minerliztion of orgnic mtter during trnsport nd pre-incution (Arnold et l. 28). We cn, however, not exclude tht this is n rtifct of our incution, s under field conditions n ccumultion of NH 4? would normlly not hppen due to plnt uptke. In ddition to inhiition y

13 Biogeochemistry (212) 111: ccumulted NH 4?, strtifiction of potentil CH 4 uptke my lso e due to the sensitivity of microorgnisms to temperture nd soil moisture chnges, which re more pronounced in the upper orgnic lyer thn in the deeper orgnic lyer or in the Ah horizon (Butterch-Bhl nd Ppen 22). Finlly, we lso found some evidence tht neroic microsites were creted, due to high microil ctivity in the uppermost orgnic lyers leding to low methnotrophic ctivity. The results of the lortory incutions of the orgnic lyer smples t the 1, m site pper to endorse the findings from temperte forest soils, tht orgnic lyers consume little, if ny, tmospheric CH 4 nd tht most of the CH 4 uptke cpcity is locted in the upper minerl soil (Conrd 1996). However, the results from our incutions of the 2, nd 3, m orgnic lyers reveled sustntil cpcity to oxidize tmospheric CH 4, with highest methnotrophic ctivity locted in the lowest orgnic lyers right ove the minerl soil, which contined minly strongly decomposed nd humified orgnic mtter. As the CH 4 concentrtions mesured in the soil ir were elow tmospheric, we conclude tht this methnotrophic ctivity ws of the high ffinity type. Conclusions Although we found s expected decrese in CH 4 uptke rtes with incresing elevtion, the widely? ccepted control mechnisms gs diffusivity, NH 4 inhiition nd the cretion of neroic microsites could not or only prtly explin vritions in tmospheric CH 4 uptke of these tropicl montne forest soils. In contrst we found some unexpected fctors controlling net methne fluxes. Net methne uptke occurred minly under conditions which re generlly fvorle for microil ctivity, nd we found indictions tht it ws N limited insted of inhiited y NH? 4. Finlly, we showed tht in the thick orgnic lyers sustntil high ffinity methne oxidtion occurred which cn influence the CH 4 udget of tropicl montne forest soils. Inclusion of elevtion s co-vrile will gretly improve regionl estimtes of CH 4 exchnge in tropicl montne forests. Acknowledgments We thnk Ptricio Sls, Angel Mcs, Fin Cuenc nd Mrco Silv for excellent field nd lortory ssistnce; Anke Müller nd Amnd Mtson for their support collecting field dt; the lortory stff of the Buesgen Institute, University of Goettingen, for their ssistnce in lortory nlysis; The Ministerio del Amiente for reserch permits, the Nture nd Culture Interntionl (NCI) in Loj for providing the study re nd the reserch sttion; The Universidd Técnic Prticulr de Loj for coopertion; The Deutsche Forschungsgemeinschft for funding this project s suprojects A2.4 nd A2.2 (Ve219/8-1 A2.2 nd Le762/1-1) of the reserch unit Biodiversity nd sustinle mngement of meg diverse mountin ecosystem in southern Ecudor (FOR 816). Open Access This rticle is distriuted under the terms of the Cretive Commons Attriution Noncommercil License which permits ny noncommercil use, distriution, nd reproduction in ny medium, provided the originl uthor(s) nd source re credited. References Arnold J, Corre MD, Veldkmp E (28) Cold storge nd lortory incution of intct soil cores do not reflect insitu nitrogen cycling rtes of tropicl forest soils. Soil Biol Biochem 4: Arnold J, Corre MD, Veldkmp E (29) Soil N cycling in oldgrowth forests cross n Andosol toposequence in Ecudor. For Ecol Mng 257: Aronson EL, Helliker BR (21) Methne flux in non-wetlnd soils in response to nitrogen ddition: met-nlysis. Ecology 91: Blke GR, Hrtge KH (1986) Bulk density. In: Klute A (ed) Methods of soil nlysis, prt 1. Physicl nd minerlogicl methods. Agronomy Monogrph, Soil Science Society of Americ, Mdison, pp Bliese PD, Ployhrt RE (22) Growth modeling using rndom coefficient models: model uilding, testing, nd illustrtions. Orgn Res Methods 5: Bodelier PLE, Lnroek HJ (24) Nitrogen s regultory fctor of methne oxidtion in soils nd sediments. FEMS Microiol Ecol 47: Brumme R, Borken W (1999) Site vrition in methne oxidtion s ffected y tmospheric deposition nd type of temperte forest ecosystem. Glo Biogeochem Cycles 13: Butterch-Bhl K, Ppen H (22) Four yers continuous record of CH4-exchnge etween the tmosphere nd untreted nd limed soil of N-sturted spruce nd eech forest ecosystem in Germny. Plnt Soil 24:77 9 Butterch-Bhl K, Rothe A, Ppen H (22) Effect of tree distnce on N2O nd CH4-fluxes from soils in temperte forest ecosystems. Plnt Soil 24:91 13 Butterch-Bhl K, Kock M, Willild G, Hewett B, Buhgir S, Ppen H, Kiese R (24) Temporl vritions of fluxes of NO, NO2, N2O, CO2, nd CH4 in tropicl rin forest ecosystem. Glo Biogeochem Cycles 18(3). Article No. GB312. doi:1.129/24gb2243 Conrd R (1996) Soil microorgnisms s controllers of tmospheric trce gses (H-2, CO, CH4, OCS, N2O, nd NO). Microiol Rev 6:69 64

14 482 Biogeochemistry (212) 111: Crwley MJ (29) The R ook. Wiley & Sons, Chichester Curry CL (27) Modeling the soil consumption of tmospheric methne t the glol scle. Glo Biogeochem Cycles 21:1 15 Dvidson EA, Ishid FY, Nepstd DC (24) Effects of n experimentl drought on soil emissions of cron dioxide, methne, nitrous oxide, nd nitric oxide in moist tropicl forest. Glo Chnge Biol 1: Delms RA, Servnt J, Tthy JP, Cros B, Lt M (1992) Sources nd sinks of methne nd cron-dioxide exchnges in mountin forest in Equtoril Afric. J Geophys Res 97: Dörr H, Ktruff L, Levin I (1993) Soil texture prmeteriztion of the methne uptke in erted soils. Chemosphere 26: Dutur L, Verchot LV (27) A glol inventory of the soil CH4 sink. Glo Biogeochem Cycles 21:1 9 Edwrds PJ, Gru PJ (1977) Studies of minerl cycling in montne rin-forest in New-Guine: 1. J Ecol 65: Emck P (27) A climtology of south Ecudor with specil focus on the mjor Anden ridge s Atlntic Pcific climte divide. Nturwissenschftliche Fkultäten. Friedrich- Alexnder-Universität Erlngen-Nürnerg, Erlngen-Nürnerg, p 275 Fiedler S, Höll BS, Jungkunst HF (25) Methne udget of Blck Forest spruce ecosystem considering soil pttern. Biogeochemistry 76:1 2 Forster P, Rmswmy V, Artxo P, Berntsen T, Betts R, Fhey DW, Hywood J, Len J, Lowe DC, Myhre G, Ngng J, Prinn R, Rg G, Schulz M, Dorlnd RV (27) Chnges in tmospheric constituents nd in rditive forcing. In: Solomon S, Qin D, Mnning M, Chen Z, Mrquis M, Averyt KB, Tignor M, Miller HL (eds) Climte chnge 27: the physicl science sis. Contriution of working group I to the fourth ssessment report of the Intergovernmentl Pnel on climte chnge. Cmridge University Press, Cmridge FRA (2) Glol forest recources ssessment 2 min report. Food nd Agriculture Orgniztion of the United Ntions (FAO), Rome Ishizuk S, Iswndi A, Nkjim Y, Yonemur S, Sudo S, Tsurut H, Murdiyrso D (25) The vrition of greenhouse gs emissions from soils of vrious lnd-use/cover types in Jmi province, Indonesi. Nutr Cycl Agroecosyst 71:17 32 Keller M, Reiners WA (1994) Soil tmosphere exchnge of nitrous-oxide, nitric-oxide, nd methne under secondry succession of psture to forest in the Atlntic lowlnds of Cost-Ric. Glo Biogeochem Cycles 8: Keller M, Vrner R, Dis JD, Silv H, Crill P, Oliveir RC Jr, Asner AG (25) Soil tmosphere exchnge of nitrous oxide, nitric oxide, methne, nd cron dioxide in logged nd undistured forest in the Tpjos Ntionl Forest, Brzil. Erth Interct 9:1 28 Kiese R, Hewett B, Grhm A, Butterch-Bhl K (23) Sesonl vriility of N2O emissions nd CH4 uptke y tropicl rinforest soils of Queenslnd, Austrli. Glo Biogeochem Cycles 17: Kiese R, Wochele S, Butterch-Bhl K (28) Site specific nd regionl estimtes of methne uptke y tropicl rinforest soils in north estern Austrli. Plnt Soil 39: Le Mer J, Roger P (21) Production, oxidtion, emission nd consumption of methne y soils: review. Eur J Soil Biol 37:25 5 Litherlnd M, Aspen JA, Jemielit RA (1994) The metmorphic elts of Ecudor-Overses geology nd minerl resources. British Geologicl Survey, Keyworth, Nottinghm Livingston GP, Hutchinson GL, Sprtlin K (26) Trce gs emission in chmers: non-stedy-stte diffusion model. Soil Sci Soc Am J 7: Loftfield N, Fless H, Augustin J, Beese F (1997) Automted gs chromtogrphic system for rpid nlysis of the tmospheric trce gses methne, cron dioxide, nd nitrous oxide. J Environ Qul 26: Mrtinson GO, Werner FA, Scherer C, Conrd R, Corre MD, Fless H, Wolf K, Klose M, Grdstein SR, Veldkmp E (21) Methne emissions from tnk romelids in neotropicl forests. Nt Geosci 3: doi:71.138/ ngeo198 Murer D, Kol S, Humier L, Borken W (28) Inhiition of tmospheric methne oxidtion y monoterpenes in Norwy spruce nd Europen eech soils. Soil Biol Biochem 4: Mosier AR, Delgdo JA (1997) Methne nd nitrous oxide fluxes in grsslnds in western Puerto Rico. Chemosphere 35: Piepho HP, Büchse A, Richter C (24) A mixed modelling pproch for rndomized experiments with repeted mesures. J Agron Crop Sci 19: Potter CS, Dvidson EA, Verchot LV (1996) Estimtion of glol iogeochemicl controls nd sesonlity in soil methne consumption. Chemosphere 32: Prther M, Ehhlt D, Dentener F, Derwent R, Dlugokencky E, Hollnd E, Isksen I, Ktim J, Kirchhoff V, Mtson P, Midgley P, Wng M (21) Atmospheric chemistry nd greenhouse gses. In: Houghton JT, Ding Y, Griggs DJ, Noguer M, vn der Linden PJ, Di X, Mskell K, Johnson CA (eds) Climte chnge 21: the scientific sis. Contriution of working group I to the third ssessment report of the intergovernmentl pnel on climte chnge. Cmridge University Press, Cmridge, p 239 Puropuspito J, Veldkmp E, Brumme R, Murdiyrso D (26) Trce gs fluxes nd nitrogen cycling long n elevtion sequence of tropicl montne forests in Centrl Sulwesi, Indonesi. Glo Biogeochem Cycles 2:GB31 R Development Core Tem (29) R: lnguge nd environment for sttisticl computing. R Foundtion for Sttisticl Computing, Vienn Reiners WA, Keller M, Gerow KG (1998) Estimting riny seson nitrous oxide nd methne fluxes cross forest nd psture lndscpes in Cost Ric. Wter Air Soil Pollut 15: Sri A, Mrtikinen PJ, Ferm A, Ruusknen J, De Boer W, Troelstr SR, Lnroek HJ (1997) Methne oxidtion in soil profiles of Dutch nd Finnish coniferous forests with different soil texture nd tmospheric nitrogen deposition. Soil Biol Biochem 29: Sri A, Heisknen J, Mrtikinen PJ (1998) Effect of the orgnic horizon on methne oxidtion nd uptke in soil of orel Scots pine forest. FEMS Microiol Ecol 26: