当前位置:首页 >> 电力/水利 >>

Ecosystema nd physiologicalc ontrolso ver methane production in northern wetlands


RESEARCH, VOL. 99, NO. D1, PAGES 1563-1571, JANUARY 20, 1994

Ecosystem and physiological controlsover methane production
in northern
David W.



NaturalResource EcologyLaboratory, Colorado StateUniversity,Boulder,Colorado

Elisabeth A. Holland and David S. Schimel
NationalCenterfor Atmospheric Research, Boulder,Colorado

Peatchemistry appears to exertprimarycontrolovermethane production ratesin the Canadian Northern Wetlands Study(NOWES)area.We determined laboratory methane production rate potentials in anaerobic slurries of samples collected froma transcot of sites through the NOWESstudy area.We related methane production rates to indicators of resistance to microbial decay (peat C:N andlignin:N ratios) andexperimentally manipulated substrate availability for methanogenesis using ethanol (EtOH)andplantlitter.We alsodetermined responses of methane production to pH and temperature. Methane production potentials declined alongthe gradient of sites fromhighrates in thecoastal fens to low rates in theinterior bogs andweregenerally highest in surface layers.Strong relationships between CH4 production potentials andpeatchemistry suggested that methanogenesis waslimited by fermentation rates. Methane production at ambient pH responded strongly to substrate additions in the circumneutral fenswith narrowlignin:Nand C:N ratios(0CI?/0EtOH = 0.9-

2.3ragg4) and weakly in theacidic bogs withwideC:N andlignin:N ratios (0CI?/0EtOH = -0.04-0.02 mg g4). Observed Qx0 values ranged from 1.7 to 4.7 andgenerally increased with increasing substrate availability, suggesting that fermentation rateswere limiting.Titrationexperiments generally demonstrated
inhibition of methanogenesis by lowpH. Our results suggest thatthe low ratesof methane emission observed

in interior bogs during NOWES likely resulted from pH andsubstrate quality limitation of thefermentation step in methane production andthus reflect intrinsically low methane production potentials. Low methane emission
rates observed during NOWESwill likelybe observed in other northern wetland regions with similar vegetation

areas[Aselmann and Crutzen,1989; Fung et .I., 1991], from uncertainty aboutthe atmospheric destruction rate [Vaghjianiand Ravishankara , 1991] and from new informationsuggesting a larger soil sink [Steudter et aI., 1989; Mosier et aI., 1991). Three-dimensional global analysesof atmosphericmethane concentrations [Fung et al., 1991] and isotopic composition [Tyler, 1986, 1989, 1991; Quay et al., 1991] are the most powerfultechniques for analyzing current sources andsinks,and provide hintsas to controls overemissions. Theseanalyses and otherstudies have suggested an importantbut highly uncertain
methane source from northern wetlands. The Canadian Northern

but would complement global analyses by linking edaphic, biological,and climaticcontrols to flux rates and providea Uncertaintyin the global methanebudgetarisesfrom the mechanistic basis bothfor spatial analysis andfor projection of scarcity of source measurements, especially integrated overl?rge future changes. Process models, when coupledwith data
describingpalcowetlandextent, also provide a basis for
understanding pastchanges in methanefluxes.


Methane production is carried outby methanogenic bacteria, which areobligate anaerobes capable of using a limited number of relatively simple substrates supplied by fermentation processes.

Several reviews of themechanisms of methanogenesis recently
have beencompiled[Oreroland,1988;Conrad, 1989; Knowles, 1993]. Conrad [1989] enumerated the controls over

biogeochemistry now existfor any ecosystem type. Process the controlsover methaneproductionfrom a seriesof sites in models of microbial transformations and relevanttransport northern wetland ecosystems and present results from processes have notbeen widely used in methane biogeochemistry experimental manipulations of severalfactors. The siteschosen

WetlandsStudy (NOWES) and two American Atmospheric over methaneproductionrate. While the controlsover methane BoundaryLayer Experiments (ABLE 3A and 3B) constrain production at themicrobial levelarereasonably well known,the bounds onthatuncertainty. relativequantitative importance of the differentfactors is not well In addition, no predictive or diagnostic modelsof methane known, especially under fieldconditions. In thispaper we analyze

methanogenesis as(1)substrate availability, (2) temperature, (3) pH, and (4) competitive electron acceptors.He arguedthat population sizeof methanogens wasnot likely a majorcontrol

contrasted in plantspecies composition, productivity, hydrology,

1Also at Natural Resource Ecology Laboratory, Colorado State
University,Fort Collins.

andpH [Klingeret aI., this issue]. Controlsover methaneproductionfall into two classes that must be considered togetherin predictivemodelsof methane

Copyright 1994by the American Geophysical Union.
Papernumber 93JD00391.
0148-0227/94/93JD-00391 $05.00

production: (1) controlsover substrate availability (i.e., fermentation rateasaffected by temperature, litterinput,andlitter decomposition rate); and (2) physiological controls over methanogenic activity(pH and redoxpotential)from a given
supply rate of substrate.



The principal substrates formethanogenesis are CO2+H 2 and


acetate [Orernland, 1988; Conrad, 1989; KnowIes, 1993). Substrate abundance is constrained by theslowest link in a chain StudySite of eventsstartingwith inputsof detritalorganicmatter and including several decomposition andfermentation steps resulting Thefourstudy sites werelocated in theHudson Baylowlands in methanogenic substrates. While manystudies identifythe alonga 95-km transect southwest of Moosonee,Ontario,Canada
importanceof substratelimitation as a control over methane production [AseImann and Crutzen,1989; Conrad, 1989], few

(Figure1). The regionhas beengradually exposed by the
recedingHudson Bay during isostaticrebound since the last

havesystematically analyzed its importance [Yavittand Lang, glacial maximum [Webber et aI., 1970],giving anagesequence 1990]. If most methane is produced largely fromrecent organic alongthe transect extending back approximately 4000 years matter, as14C results suggest [Jeris and McCarty, 1965; Wahlen [Klinger et aI., thisissue]. From youngest (closest to James Bay) et al., 1987], thenmethane production should ultimatelybe to oldestthe four sitesstudiedwere the coastalfen, interiorfen, closely coupled to terrestrial primary production and Carling Lake, and KinosheoLake. The older two sites were decomposition rates. NOWES provides a foundation for ombrotrophic Sphagnum bogs whiletheyounger twoweresedgemodeling thelinkage of methane production to ecosystem change dominated minerotrophic fens. (through changing plantproduction andchemistry) as well as to The vegetation and peat characteristics of the sites are

described in detail byKlinger etal. [this issue] sowepresent only a summary here(Table1). The relativecovers of sedges, nonSphagnum bryophytes, deciduous shrubs, andpteridophytes are [1989]reported thattheeffectof increasing temperature on the highest in theinterior fen,decreasing slightly in thecoastal fen rateof methane production is largely offsetby a correspondingandareverylowor absent in bothof thebogsites.Conversely, increase in Km (decreased substrate affinity)of methanogenicthe relativecoversof sphagnum, ericaceous shrubs, and lichens enzyme systems. Thustheobserved effectof temperature will aremuch higher in thebogsites (especially Kinosheo Lake)than likely be manifested only through increased availabilityof in the fen sims. The covers of forbs and coniferous shrubs and methanogenic substrate aslongasreadily decomposable reduced treeswere fairly constant across sites. Peatdepthincreased carbon compounds areavailable. These interactions areimportant approximately monotonically fromabout 0.7 m in thecoastal fen because theycould leadto unexpected results, e.g.,temperature site to about 2.9 m at Kinosheo Lake.
increases could lead to increased instantaneous rates of

climatechange. Temperature directlyaffectsmethanogenesis ratesas well as decomposition and fermentationrates, but Westermann et aI.

methanogenesis via enhanced fermentation, butthese higher rates Procedures could outstrip supply rates of readily fermentable compounds and Intact cores were taken during July1990using a 10.2-cm 2 box ultimately leadtoreduced methane production rates. corer from each of the four sites. Four cores were taken from the Appropriate redoxpotentials for methanogenesis require

theaerobic zone in which CH4 oxidation canoccur butalsothe bog sites. All sampleswere vacuumsealed in 0.13-mm-thick portion of the soil or peatprofilein whichmethanogenesis is
possible. Conversely, redoxpotentialalsomay limit substrate production because very low redoxpotentials tend to retard

Lake and Kinosheo Lake.In addition, wesampled thetop competing, more electronegative electron acceptors (NO3-, SO42-, Carling atboth Fe3+). Water table fluctuations influence not only the thickness of 1 m of peatfromthesideof a peatpit in 0.1-mincrements

saturated or nearly saturated soil conditions and the absence of

coastalfen site, three from the interior fen, and two each from

decomposition. Since organic matter chemistry varies withdepth in most profiles, water table depth will influence both thequantity
of organic matter formethanogens andthemaximum rateat wlfich that organicmatter may be processed. We considered this


interaction only to the extent that we surveyedmethane production potentials by depth andamended samples fromthe
uppertwo peatdepths.


In thefield,methane emission is influenced by production but is modulated by transmission to theatmosphere which occurs via diffusion, solution transport, bubbles or vascular transport in plants [Kelleret aI., 1986; Conrad, 1989;Chanton andDacey, 1991;Fechner andHernon& 1992]. Transport, oxidation and water table depthwere considered by other investigators in NOWES[KIinger et aI., thisissue; RouIet et al.., thisissue]. Detailed process studies of thesortdescribed in thispaper are an important complement to field experiments not onlybecause they aid in modeldevelopment but alsobecause they aid in
interpretingfield observations.Methane flux rates observedin




I 0
Kinosheo Site

I I0

I 20

I 34)

I 40

I 50 km

Carling Lake

Interior ?


Coastal Fen Site


NOWES weremuchlowerthanexpected [Rouletet al., this

issue] and ourprocess studies help explain why. Ourstudies help establish whether fluxeswerelow because of intrinsically low rates of methane production (determined frommaximal laboratory
methaneproductionrates). This will be of importancein extrapolating NOWESresults to other regions.


81 ?00' W

82 ? O0'W

Fig. 1. Location of study sites.Thetwoend-point sites; coastal fenand Kinosheo Lake were used fordetailed methane process studies.



TABLE1. Summary of Site Characteristics forSites Used in Regional Survey of Methane
Production Potentials


Distance From

Water Table



Basal Peat

Interior fen







80?37' W

Lonl?itude tkm)





80?52.8' W


Carling Lake

51?29.8' N
81?45.7' W






81 ?49.5' W

*From Klingeret al. [thisissue].

period following sparging varied between 0.25 polyethylene bags,thenkept at 4?C until analysis. The peat period.The assay
horizons were classified using the von Post scale of peat humification [Bglanger et al., 1988] (Table2). Moisturecontent was determined gravimetrically by drying to constant weight at 70?C. Total carbonandnitrogen concentrations weredetermined ondriedandground samples at 10 cmincrements on a CarloErba NA 1500 CN analyzer. Lignin and cellulose contents were determinedat 10-cm intervals using the Goering-Van Soest

and 2 hours, depending onrates of CH 4 production. At theendof
the assayperiodwe used35-mL polypropylene syringes fitted with nylon stopcocks to sample12 mL of the headspace. If two

samples wererequired, aswhenCO2 production rates weremore rapidthanCH4 rates,then12 mL of UHP N2 wasadded as a
makeup gasfollowingthefirst sampling. Gassamples wereanalyzed within3 hours followingsampling on a ShimadzuGC-8 gas chromatograph fitted with a flame

procedure [Goering andVanSoest, 1970; KirkandObst, 1988].
Methaneproduction potentials in slurtiedincubations were
assessed for all cores at 10-cm intervals at 20?C. Incubations were

ionization detector (FID) forCH4 and a Carle Instruments thermal conductivity detector (TCD)forCO 2. We used a 3.175-mm OD
Ni column packed with HayeSep D in an 60?Covenwith a carrier

performed in slurries of 20 g (fresh weight) subsamples of wellmixedpeat. The peatandenough deionized water(to 100 mL)

gas (He)flowrate of 35cm 3min -1. TheFID was kept at200?C

were added to a 250-mL Erlenmeyer flask under a nitrogen and the TCD at 60?C. Concentrations were corrected for trace of CH4 (generally 0.2 parts permillionby volume atmosphere. The flaskswere stoppered with siliconestoppers quantities contained in theUHPN2 used fortheheadspace. fittedwitha stopcock, septum, andin a subset of replicates, redox (ppmv)) Duringthe incubation period,production potentials generally (platinum) andpH electrodes. Theseallowed monitoring of redox potentials to ensure thatconditions sufficiently reduced for increasedfor the first several weeks, after which they declined methanogenesis prevailed in the flasks. Slurries initially were modestlyin most slurries. No initial rapid methaneproduction sparged withultrahigh pure(UHP) N2 for 2 minutes andthen rates were observed, suggesting that disturbance-enhanced of volatile fatty acids were not reflected in our continuously shaken on anorbital shaker in refrigerated incubators concentrations

throughout the incubation period. Samples weresubsequentlydata. We alsoassessed process controls overCH4 production in the allowed to preincubate for 2 weeks before measurement to allow exhaustion of initiallyhighvolatilefatty acidlevels,afterwhich top two von Posthorizonsof Coastalfen and KinosheoLake
incubations lasted 4-6 weeks.Duringtheincubation period,redox
changed lessthan0.2 units.

peats.Temperature effects wereanalyzed on slurries at 10?C and
or 8.22 mmol ethanol (39.4 mg or 197.3 mg C equivalent,

of substrate limitation wasassessed by adding1.64 potentials remained very low (-300 mV) andpH generally 20?C. Degree

Flaskheadspaces weresampled weeklyduring theincubation period. Before sampling, each flask was sparged withUHPHe by vigorously bubbling the slurryfor 2 minutes. This procedure
reduces but doesnot eliminatedegassing problemsduring the

assay period [Kiene andCapone, 1985],minimizes problems of pressure variation among flasks due to CO2 and CH4
accumulation,establishessimilar bicarbonateequilibria among

sampling periods, andlessens concentration gradients driving
diffusive lossesacrossthe silicone stoppersduring the assay

respectively) as an acetateandhydrogen source [Bryantet al. 1967;Large, 1983;Svensson, 1984]. Although ethanol cannot be useddirectlyby methanogens, it is easilyconverted to hydrogen andacetate by otherorganisms [Bryantet al., 1967;Svensson, 1984]. For our experiments, ethanol also presented two advantages over acetate:first, protonated aceticacid is toxic to microbial populations, implyingthatonly very smallamounts of acetate couldbe added at low pH; second, ethanol wouldprovide a prolonged supply of substram duringthe incubation. Because

TABLE 2. Depths and Von Post Indices (VPI) forPeats Used inAmended Incubations,
Coastal Fen Kinosheo Lake

Peat Horizon Surface Index 2-3

cm 0-10.5

Index 1

cm 12






Peats were sampled byVonPost Index, and thetoptwouniform VPI horizons were used in experimental studies. Fulldescriptions ofpeat profiles are inthework ofKlinger
etal. [this issue]. Chemical characteristics of peat profiles areshown in Table3.



acetic acid is a weak acid, moreover, a significant fraction of addedacetatesaltswould protonateand raisepH in the low pH slurties. Substrate limitationto fermenting populations was also assessed by adding 0.5 g of driedandground plantlitter samples collectedon site. Effectsof pH on production potentialswere assessed by loweringthepH of coastal fen slurries to -?5.5 or -?4

[Klingeret al., this issue]. The trend also agreedwith in situ

measurements of CH4 emissions, except for anomalously low
emissionsfrom the interior fen [Klinger et al., this issue] compared to the production potentials (Figure 2). This

discrepancy probably resulted fromCH4 oxidation dueto thelow
watertableanddeepaerobic zone(Table 1).

withH3PO 4 andraising thepH of Kinosheo Lakeslurtics to -?5.5
or -7 with KOH. We alsotestedtheresponse of Kinosheo Lake

slurties to direct H2 enhancements. Additions of 10 mL of H2 brought headspace partialpressures to 0.05 atm.,but no CH4

In thecoastal andinterior fens,methane production potentials were highestin the surface peatsand declinedwith depthto a plateau below0.2 m (p<0.0.05, Figure2). The depthpatternwas

less clear in thebogsites; lowCH4 production potentials occurred

peats fromthebogsites, CarlingLake, and production response occurred (12.1 versus19.2 and 13.4 versus at all depths.Surface 8.3 gg/g for surface and subsurfacelayers, respectively). Kinosheo Lake had slightly higher productionpotentials, thandid intermediate depths. Similardepth patterns of Because of the lack of response we dropped H? treatments however,
thereafter in favor of ethanol amendments.

CH 4 production have also been observed in spruce-peatlands in

Statistical significance of experimental results weredetermined westcentralAlberta[Hogg et al., 1992]. usinganalysis of variance. Rates of CO2 production exhibited thesame spatial pattern as methane,supportingthe argumentthat methaneproductionis linkedto carbon substrate available for decomposition (seebelow)

(Figure 2b).Decreases in surface peat CO 2production potentials
from coastalto inland sites were less marked than for methane,

Landscape Patterns in CH4 andCO2 Production
Methane productionpotentialswere highestin Coastal fen slurtiesand dropped rapidly with distance inland to the bogsat CarlingLake and Kinosheo Lake (p<0.01, Figure2). This trend corresponded to patterns of soil solution pH (highest at thecoast), NPP (highestat .thecoast),and plant communitycomposition (sedgesand bryophytesin the coastal and interior fen sites, Sphagnum moss,and ericaceous shrubs in the inland bog sites)

though significant.Subsurface peats had similarandlow rates
along the entire transect,as with methane. More methane was

produced perunitCO 2 produced in thecoastal than in theinland
sites (p<0.005),perhaps because of morefavorable physiological conditions for methanogen activityandpopulation growthin the
neutralcoastal peats.
Process Controls

I?] 0-10 cm


? 20-40 cm

Organic matter quantityand quality. Methane production potentialsalong the transectand with depth corresponded to changes in peatchemistry andits influence on decomposition and fermentation rates. Ratesof methanogenesis depend on ratesof substrate supply from fermentation,and rates of fermentation dependon the amountand decomposability of organic matter
[Conrad 1989; Sass et al., 1990, 1991; Svensson, 1976, 1980;



180-100 cm

Svensson and Rosswall, 1980; Taylor et al., 1991]. Decomposition of peatis slowunderunperturbed conditions, and substrate for heterotrophic processes is usuallylimiting[Svensson, 1980]. Severalinvestigators have foundthatmethane production or emissionwas highest at sites where NPP, plant density, or


biomass washighest [Klinger et al., thisissue; Morrissey and
Livingston,1992;Svensson, 1976, 1980; Whitingand Chanton, 1992]. The role of NPP is clearlythrough the annualsupplyof relativelylabilefractions of plantdebris, though vascular transport playsa role at somesitesaswell. The decomposability of the litter produced influences ratesof decomposition, fermentation, andmethanogensis. The chemical availability of carbonfor decomposition is usuallyquantified by two indicesof decomposability: C:N and lignin:N ratios of the materialto be decomposed [Parnas, 1975, Melillo et al., 1982;



0 50

Bosatta and Berendse 1984; ,?gren and Bosatta, 1987' Berendse
et al., 1987]. The indiceshavebeenusedfrequently to describe the substrate qualityof litter. We usethemhere to describe the decomposability of peat,whichis partlydecomposed plantlitter. We also measuredcellulosecontentof peats, as cellulose is a relativelyavailable C fraction, routinelymeasured aspart of the

lignindetermination. Threeresults fromthisstudy emphasize the of substrate supplyin controlling spatialpatterns of Coastal Fen Interior Fen Carling Lake Kinosheo Lake importance methanogenesis: (1) CH4 production potentials uniformly Site
Fig.2. (a) Methane and(b)CO 2 production potentials under unamended
conditions for all sitesalong the regionaltransectand as a function of depth. Errorbarsare standard errorsof themean(n-4 for CF, 2 for IF, 3 for CL, 3 for KL).

increased withsubstrate addition, including both EtOHandlitter amendments (Figure 3); (2) CH4 production potentials increased
with increasing substrate quality(decreasing lignin:N ratiosand

increasing cellulose) (Table3); (3) CH4 production potentials







::::L 10



pH 7 5.5 4

7 5.5 4

7 5.5 4

0.0 mg C

39.4 mgC


Fig. 3. The response of methane production to added EtOH andpH at two depths (see Table2 for horizon depths anddescriptions) areshown for (a) neutralcoastal fen peat(n=3), (b) acidKinosheo Lakepeat(n=3), and (c) KinosheoLake degradational feature ("black hole") peat (low sample availability prohibited replication)."NA" indicates thata giventreatment combination was not cardedout, due to insufficientremainingsample.
Error bars are standard errors of the mean.

increased with increasing NPP ofthe peat collection site (Figure 1
and Klinger et al. [this issue]. Both the C:N and the lignin:N ratios of surface peat increase and, presumably, substrate

decomposability declines, with distance inland (Table 3).
Strongly decreasingN contentsinland raised the C:N ratio at KinosheoLake to nearly triple that of the coastalfen. Lignin:N

ratiosonly doubled alongthe sametransect because of partially
offsetting decreasesin lignin content inland. The C:N and lignin:N ratios suggest that carbon is more available for fermentation and subsequent methanogenesis at the coastalsites

than at the inlandsites. Higher CH4 and CO2 production potentials forthecoastal compared totheinland sites support this. The highestCH4 and CO2 production potentials always
occurred in surface peats(p<0.05) anddeclinedwith depth. The

lowbutnonzero CH4 production potentials at lowerdepths in all
sitesindicatethat theselayers can serveas important,long term

contributors to CH4 stored in pore water andepisodically released
duringwatertablefluctuations [Rouletet al., thisissue]. The surface peat layer receivesall of the aboveground litter inputsas well asmostof the inputs of freshorganicmatterfrom root turnover and most closely reflects the chemistry of the existingvegetation.Changes in peat chemisu3, with depthresult from historicmicrobial transformations and the chemistryof the vegetationfrom which the peat formed. As indicatedby the



pattern of CH4 and CO2 production rates, decomposability of
organic matter also decreaseswith depth. In the bog sites, production potentialsdecreased with depth, paralleling strong increasesin C:N and lignin:N ratios with depth that may have servedto restrictdecomposition rates. Neither the C:N nor the lignin:N ratio reflectedthe depthtrend in the fen sites,however. A similarlackof patternin differentsitesled Hogg et al. [ 1992] to suggest thatotherfactors,suchas higherP, K, and carbohydrate contents in recently dead organic matter, controlled decomposability withinthepeatcolumn. The positive correlation


Depth 1 Depth 2



between CH4 production potentials andcellulose content in the
fen sitesseemto support this idea. This overalllandscape patternsuggests that two different sets of controls over decompositionand methane production may 'T' dominateat the two site types. The recalcitrantchemistryof organicmatter, low pH, and likely low nutrientlevels severely restrictmicrobial activitiesin the bog sites,resultingin the very




low overallCH4 and CO2 production potentials. The more
favorable peat chemistry, high pH, and likely higher nutrient levels at the fen sites remove this physiological restriction on microbial activity, allowing the rapid decomposition and methanogenesis ratesto be limited only by the supplyof substrate. 0 Fermentation controlsthe rate of substrate supply,which in 10oc 20oc 10oc 20oc turn limits methanogenic rates. Our experimental ethanol 0.0 m? c 39.4 mg c additionsindicated that substrateavailability limited methane production ratesundercertainconditions.In coastalfen surface Fig. 4. Response of methaneproduction at two depths(seeTable 2 for peat slurries, methaneproduction ratesincreased (Figure 3a) and horizondepths anddescriptions) to (a) temperature in the coastal fen peats and (b) temperature andsubstrate (asEtOH) in Kinosheo Lakepeats. carbon dioxide productionrates decreasedin the presenceof ethanol, although neither responsewas significant (p>0.1). However,the proportion of respired carbonevolvingas methane

(i.e.,CH4/CO2) increased strongly andsignificantly (p<0.005) apparent Q10values were2 and-2.1 for surface andsubsurface
with added ethanol. Little or no response occurred from subsurface samples(Figure 3a). The same qualitative pattern occurred(althoughat much lower methaneproductionrates) in
most Kinosheo Lake slurties: an 80% increase in methane

production from the surface peats, with no responsefrom the subsurface samples (Figure 3b). Methaneproductionincreased eightfold from ethanol-treated KinosheoLake "black hole" slurties (p<0.05; Figure 3c). This responsiveness to substrate amendments suggests that in the field, methanogenic processes might outstripfermentation rates. In the black holes, fermentation and methanogensismay occur at different depths. The dark color, low albedo, and lack of vegetation cover all contribute to unusually warm surface temperatures of thesefeatures[Klinger et al., this issue],which accelerate decomposition. The closeproximityof the water table to the surface in these depressionsminimizes the distance substrate must descendto reach methanogenic populationsand

peats. The apparentdecline in methaneproductionpotentials from KinosheoLake subsurface peatswasnot significant (p>0.1) but, if real, may haveresultedfromrapiddepletion of an initially small pool of available carbon when microbial activity was enhanced by the warmer temperature.An additionalexperiment
examined interactions between substrate amendment and

temperature in Kinosheo Lake peats. Apparent Q10 values
increased slightly from 2 to 3 in surfacepeat slurties,and from -2.1 to 1.4 in subsurface peat slurties--when 39.4 mg of carbon(as

ethanol) was added (Figure4b), although the CH4 production
response to both the temperature changeand the ethanolwere not significant (p>0.1). Similar patterns have been observed in pure cultures of Methanosarcinabarkeri, in which Westermann et al. [ 1989] found thattemperature had little net effecton the conversion of acetate

tomethane.An offsetting increase in Kmapparently cancels out

nearly all of the increased methanogenic enzymeactivity with alsominimizes the oxidizingzonethrough whichCH4 must increasing temperature. Following this argument, apparent ascendto reach the atmosphere.Downward transportof labile temperature responsesof methane production are instead organicmatter can explain how thesepeatscould serve as such responses to increased substrate availabilityfrom temperaturestrong CH4 sources in thefield,despite low unamended methane enhanced rates of fermentation.

production potentials in lab incubations. Temperature. Many studies have demonstrated that temperature canbe an important controlovermethane production, yet correlationsbetween methane flux and temperature are typicallyweak [Williamsand Crawford, 1985; Crill et al., 1988;
Conrad, 1989; Moore et al., 1990; Moore and Knowles, 1990;

Responses topH. The response of CH4 production potentials
topH changes is important to understanding spatial variations in

CH4 production because wetlands tend tobecome more acidic as
their species composition and hydrology change through vegetation succession over time. Site differences in successional statusare a major source of landscape variability in wetlands [Klingeret al., 1990;Klingeret al., thisissue].Methanogenesis tendsto be mostrapid in the circumneutral pH range [Conrad, 1989;ConradandSchiitz,1988], although acidtolerant[Williams and Crawford, 1985; Patel et al., 1990] and moderately acidophilic [Maestrojudn and Boone, 1991] strains of methanogenicbacteria have been isolated. Conrad [1989]

WhalenandReeburgh, 1990;Klingeret al., thisissue].We found that temperaturehad a greater effect on methane production potentials in siteswheresubstrate qualitywashigher(Figure4).

Valuesof Q?0 (the ratio of methane production at T/methane
production at T-10) for thecoastal fen were4.7 and 1.7 for surface andsubsurface peats,respectively (Figure4a). At Kinosheolake,



patterns of response to suggested that p H limitationmay also be important in the Kinosheo"blackholes"showedconsistent

fermentation steps supplying substrate for methanogenesis. Our pH manipulation experiments generally suggested that neutrophilic methanogen populations dominatein northern wetlandpeats.Methaneproduction generallyincreased with increasing p H, evenin acidic peats.ThepH of thepeats ranged

substrate andp H, but quantitative responses varied with initial peatchemistry. The positiveresponse of methane production with respect to

substrate amendment addition atambient pH (3CH4/3EtOH; Table

4) suggests thatsubstrate availability playsa keyrole in regulating production.Siteswith moredecomposable peat due to from 6.5-7 in the fen sites down to almost 4 in the bog sites methane lignin:Nratios(coastal fen) couldmorereadilyrespond (Tables 1 and 3), correlating wellwithCH4 production potentials.narrower Lowering thepH of coastal fensurface peatslurries from7 to to substrateadditions, perhaps because of differences in sizeor maximalgrowth rate. circa 5.5cut CH 4production potentials almost sevenfold (p<0.05) population Methaneproduction with respectto pH at ambientsubstrate from1376 to208ggg-1 week-1, and further lowering pH to4 cut

wastypically strong andpositive (Table 4). The CH 4production potentials to5gg g-1 week-1 (Figure 3a).Similar(3CH4/3PH) at theacidic Kinosheo Lakesitewasmuch smaller patterns also occurred withsubsurface peatslurries. Raising the 3CH4/3PH atthecircumneutral coastal fensite.Similarly, 3CH4/3T was pH of Kinosheo Lakesurface peatslurries from4 to nearly 7 than
fen thanin the inlandsites(Table4). Thus increased (p<0.05) CH 4production potentials 20-fold from 0.5to largerin thecoastal

11 gg g-1week-1, whilea similar pH change in subsurface peat the ability of the substrate-limited site to respond to an
for methanogenesis waslimitedby the low ra?e of (Figure 3b). Although Kinosheo Lakeslurries produced more temperature) supply. methane in response to both increased pH andethanol additions substrate Litter chemistryandp H play importantroles in controlling individually, thecombined treatment typically produced less CH 4
than the ethanol treatment alone (Figure 3b). The weaker spatial and temporal variations in methaneproductionrates.

slurries increased CH 4potentials from 1.3 to3.2ggg-1 week-1improvement in

the physiological environment (pH


Spatially, peat chemistry reflectsthe chemistry of the peatland landscape gradient inCO 2 compared toCH 4 production suggests

with changes in lignin:Nratio reflecting the anadditive effect ofpH above thesubstram availability constraint. plantcommunity, shift in vegetation from the high N graminoids of the coastal fen The exception to the overallneutrophilic methanogenesis

to thelow N sphagnum species of the interiorbogs. patternoccurred with peatssampled from dark-colored community acidicandnutrient deficient conditions along degradational features ("black holes") at Kinosheo Lake. These With increasingly

black holes werestrong localmethane sources in situ [Klinger et al., thisissue; Roulet et al., thisissue], butuntreated slurries of

these peats produced littlemore CH 4 than other Kinosheo Lake peats.Methane production fromblack holeslurries notonly
failed to increasewith increased p H but actually decreased

slightly (p<0.05, Figure 3c). Similarly, the strong ethanol response observed in these slurries wasalmost completely offset (lignin:N)andpH havecoordinated andinteractive when pH wasincreased (interaction significant atp<0.05).This chemistry production, linkingmethane production to patternsuggests the presence of moderately acidophilic effectson methane succession through changes in plant chemistry. methanogens but alsomayhaveresulted fromtoxiceffects of
ethanol or theneutralizing titrant(KOH).

that successional gradient,species are favoredwhich are acid tolerant (e.g., mosses) and which have high nutrient use efficiency,thoughlow intrinsicgrowthrates [Kilnget, 1990; Chapin et al., 1987;VanCleve andVierek,1981;VanCleve et al., 1991]. This resultsin a positivefeedback,further reducing nutrient availability throughimmobilizationinto the slowly decomposing peat. As noted above, these changesin peat

Substrate limitation of methanogenesis in northernwetland

ecosystems mayalso contribute to seasonal variations in methane emissions. It is likely thatmuchmethane production is produced CONCLUSIONS fromrecently produced carbon (Figure5) [Klingeret al., this of freshliner will vary seasonally with We hypothesized thatrates of methane production should be issue]. The availability and with decomposition rates. constrained by substrate availability andregulated by temperature, timing of NPP and senescence variations in methane production andemission rates redoxpotential, andpH. Results from sitesother than the Thusseasonal

TABLE4. TheResponse of Methane Emission toManipulation of Substrate, pH, or Temperature asa Function of InitialLignin orpH WhileHolding Olher
Factors Constant

Coastalfen Coastalfen Kinosheo Lake KinosheoLake Coastalfen Coastalfen KinosheoLake KinosheoLake coastal fen Coastalfen Kinosheo Lake Kinosheo Lake

Depth Lignh?:N 0CH4/0EtOH Standard Error
1 2 1 2 17.8 19.7 31.4 23 47,985 25,399 32.6 54.9 27,173 8855 54 8.9

l 2 1 2 1 2 1 2

Initial pH
6.56 5.8 4.14 3.97 17.8 19.7 31.4 23

941 29.2 3.85 0.7 2.2 0.89 0.05 0.11

Standard Error
960 7.3 2.31 0.36 1.34 0.44 0.06 0.24

Depth Lignin:N ?CH4/OT ? Standard Error

Units are ,ug CH 4 (gpeat) -1per gEtOH-C, ?C, or?Hunit as appropriate.





Substrate and pH

IOepth 1

controlling methane emission rates. The importance of other factorsnot examinedhere, suchas competitiveelectronacceptors

(e.g.,8042), nutrient availability, water tablefluctuations,

o o







Peat Chemistry

-1 CH4


Utter InputRate

-r- 2
0 0.5

methanetransport, and methaneoxidationrates,will alsoneed to be included in futureprocess level methane emission studies. Finally, our resultsmay provide a partial explanationfor the low rates of methaneemissionsfrom interior bogs observed duringthe NOWES study[Klingeret al., this issue; Rouletet al., this issue]. Methane productionrates in thesesites are severely restricted at the fermentation stepby poor substrate quality and at themethanogenic stepby very low pH. This suggests thatsimilar
low emission rams will also be observed in other northern wetland

Ugnin?N - 18-20
C ?N - 20-25

regions with similarvegetation floristicsandchemistry.
Acknowledgments. We acknowledge the assistance of our colleagues Arvin Mosier, Nigel Roulet, John Pastor,Indy Burke, Lee Klinger, J'm Greenberg, and Pat Zimmermanfor assistance with the desi2n.execution, and interpretation of this study. By Brown providedassistance aufing field sample collection, andRod Hansen,ScottFeeley,Cory Cleveland, JasonNeff, and Becky Riggle provided expert laboratorywork. The comments of Mike Keller, Patrick Crill, and two anony]nous reviewers helpedimprovethis manuscript.The seniorauthorwas supported by a distinguished postdoctoral fellowship in globalchange awarded by theUS Department of Energyand administered throughOak Ridge Associated Universities. This studywas conceivedwhen the junior authorwas a
National Research Council senior fellow at NASA Ames Research Center

Temperature Substrateand pH






??Ot, ?PH 7 "Black. !_r ?I
40 apH4' Hole"
. I..
0.0 39.4 197.3

6o apH55


0 0 !m,l?l I.

_ !i:_il_l




I ?' _.1 CH 4

Peat Chemistry
Ugnin?N = 23-57
C ?N = 57-101


Substrate andpH



and the support of the Ecosystem Scienceand Technology Branchis acknowledged with gratitude. This researchwas supported by NASA (NAGW 1828 and NAGW 2662), the National Center for Atmospheric Research, andby the Canadian Northern Wetlands Study,whichprovided superb logistical support. TheNational Center for Atmospheric Research is sponsored by the NationalScienceFoundation.This is Scientific
Contribution number 92-19 of the Canadian Institute for Research in

mg C

Fig. 5. Processes controlling methane production for thelow N, low pH Kinosheo Lake,andthehighN, neutral coastal fen site. The pathway of methane production is shownwith experimental estimates of control functions adjacent.Solid linesindicatethe flow of carbon through the methane pathway. Dashed linesindicate thepointof influence of a control
function. Asterisk indicates treatments not included because of

Atmospheric Chemistry(CIRAC). The work described herein was undertaken as part of the CIRAC NorthernWetlandsStudywhichhas been generouslysupported by the Natural Sciencesand Engineering Council of Canada. Althoughthe research described in this article has beencarried outunder theauspices of CIRAC, it reflects onlytheviewsof the authors anddoesnot necessarily reflectthe official viewsof CIRAC.


insufficient sample.

J?gren, G. I., and E. Bosatta, Theoretical analysis of the long-term
dynamics of carbon andnitrogen in soils,Ecology, 68, 1181-1189,

microclimaticindicators[Crill et al., 1988; RoMe! e! al., 1992]. approach, Can.J. Bot.,65, 1116-1120, 1987. Our studyhas established quantitativerelationships between Bosatta, E., and F. Berendse, Energy or nutrient regulations of

will reflectnot only microclimatic controls overmethanogens and wetlands andrice paddies, theirnet primaryproductivity, seasonality fermenters(temperatureand water table fluctuations)but also and possible methane emissions, J. Atmos. Chem., 8,307-358,1989. variationsin substrate supply. These variations,in additionto Billanger, A., D. Polvin, R. Cloutier, M. Caron, andG. Th?hault,Peat-A Resource of the Future, Centrede Recherche Premier,Riviere-duchanges in porewatermethanecontentand methaneoxidation Loup,Quebec, 1988. rates, lead to the common failure of simple correlations of Berendse, F., B. Berg,andE. Bosatta, The effectof hgninandnitrogen on methaneemissionrateswith temperature, water table, and other thedecomposition of litterin nutrient-poor ecosystems: A theoretical decomposition: Implications for the minerahzation-immobilization response toperturbations, Soil Biol.Biochem., 16,63-67,1984. Bryant, M.P., E. A. Wolin, M. J. Wolin, and R. S. Wolfe,
Methanobacillus omelianskii, a symbiotic association of two species of
bacteria, Arch. Mikrobiol. , 59, 20-31, 1967.

Aselmann,I., and P. J. Crutzen,Global distributionof naturalfreshwater

ecosystem andphysiological factors andmethane production rates in controlledlaboratoryconditions. Theseresultsform the basis for our overall model of methane production (Figure 5). Temperature, organic matterquality,andsupplyrate act to control fermentationand hence substratesupply rates to methanogens. Substrate supply rateandpH, in turn,control methanogenic rates. Thus we view methaneproductionas an ecosystem process coupledto plant speciescomposition, primary production,and overall heterotrophic metabolism. Spatial patternsin methane production also reflect plant physiological responses to successional andresource gradients that,in addition to climateand hydrology,influencethe decomposition of peat and litter. The observed interacting effects of substrate supply rate and temperature on methaneproduction rateswill requirevalidation throughfield studiesin order to confirm their importancein

Chanton, J.P., andJ. W. H. Dacey,Effects of vegetation onmethane flux, reservoirs, andcarbon isotopic composition, in TraceGasEmissions

by Plants, edited by E. A. Holland andH. A. Mooney,pp. 65-92,
Academic, San Diego, Calif., 1991.

Chapin, F. S. III, A. J. Bloom,C. B. Field,andR. H. Waring,Plant responses to multiple environmental factors, Bioscience, 37, 49-57,

Conrad, R., Controlof methane production in terrestrial ecosystems, in Exchange of TraceGases Between Terrestrial Ecosystems and the

Atmosphere, edited byM. O. Andreae andD. S. Schimel, pp.39-58,
John Wiley, New York, Chichester, 1989. Conrad,R., and H. Schfitz, Methodsof studying methanogenic bacteria andmethanogenic activities in aquatic environments, in Methods in



AquaticBacteriology,editedby B. Austin,pp. 301-343, JohnWiley,
New York, Chichester, 1988.

Crill, P.M., K. B. Bartlett, R. C. Hatriss,E. Gorham, E. S. Verry, D. I.
Sebacher, L. Madzar, and W. Sanner, Methane flux from Minnesota

peatlands, GlobalBiogeochem. Cycles, 2,371-384, 1988. Fechner, E. J., andH. F. Hemond, Methane transport andoxidation in the unsaturated zoneof a Sphagnum peatland, GlobalBiogeochem. Cycles,
6, 33-44, 1992. Fung,I., J. John,J. Lerner,E. Matthews,M. Prather,L. P. Steele,and P. J. Fraser, Three-dimensionalmodel synthesisof the global methane

Sass,R. L., F. M. Fischer,and P. A. Harcombe,Mitigation of methane emission from rice fields:Possible adverse effectsof incorporated rice straw, GlobalBiogeochem. Cycles, 5, 275-287, 1991. Steudler,P. A., R. D. Bowden, J. M. Melillo, and J. D. Aber, Influence of nitrogen fertilization on methaneuptake in temperateforest soils, Nature, 341,314-316, 1989. Svensson,B. H., Methane productionin tundra peat, in Microbial

Production andUtilization of Gases (H2, CH 4, CO), edited by H. G.
Schlegel, G. Gottschalk, andN. Pfennig, pp. 135-139E, Goltze KG, G6ttingen,1976.

Svensson, B. H., Carbon dioxide and methane fluxes from the cycle, J. Geophys. Res., 96 (D7), 13,033-13,065, 1991. Goering,H. K. and P. J. Van Soest,Foragefiber analysis,Agri. Handb. ombrotrophic parts of a subarctic mire,in Ecology of a Subarctic Mire, editedby M. Sonesson, Ecol.Bull. (Stockholm), 30,235-250, 1980. 379, Agri. Res. Serv., U.S. Dep. of Agri., Gov. Print. Off., Washington, D.C., 1970. Svensson, B.H., Different temperatureoptima for methaneformation Hogg, E. H., V. J. Lieffers,andR. W. Wein, Potential carbon losses from when enrichments from acid peat are supplemented with acetateor peat profiles: Effectsof temperature, droughtcycles,and fire, Ecol. hydrogen, Appl.Env.Microbiol.,48,389-394. Appl.,2,298-306, 1992. Svensson, B. H., andT. Rosswall,Energyflow through a subarctic mire at Jeris,J. S., and P. L. McCatry, The biochemistry of methane fermentation Strodalen,in Ecology of a SubarcticMire, edited by M. Sonesson, using 14C tracers, J.Water Pollut. Control Fed., 37,178, 1965. Ecol. Bull. (Stockholm), 30, 283-301, 1980. B. H., and T. Rosswall,In situ methane production from acid Keller, M., W. A. Kaplan, and S.C. Wofsy, Emissions ofN20, CH 4 and Svensson,

CO 2 from tropical forest soils, J. Geophys.Res.,91(Dll),
11,791-11,802, 1986.

peat in plant communities with different moistureregimesin a
subarctic mire, Oikos,43, 341-350, 1984.

Kiene,R. P., andD. G. Capone, Degassing of porewatermethane during sediment incubations, Appl.Environ.Microbiol. , 49, 143-147, 1985. Kirk, T. K., and J. R. Obst, Lignin determination, in Methods in Enzymology,editedby W. A. Wood and S. T. Kellogg, vol. 161, pp.
87-101, 1988.

Tyler, S.C., The globalmethane budget, in Microbial Production and 1990. Consumption of Greenhouse Gases: Methane, NitrogenOxides, and Halomethanes, editedby J. E. Rogers andW. B. Whitman,pp. 7-38, Knowles, R. Methane: processes of productionand consumption, in AmericanSociety for Microbiology, Washington, D.C., 1991. Agricultural Ecosystem Effectson Radiatively ImportantTraceGases G. L., andA. R. Ravishankara, New measurement of the rate andGlobalClimateChange, editedby L. A. Harper,A. R. Mosier,J. Vaghjiani, coefficient for the reaction of OH with methane, Nature, 350,406-409, M. Duxbury,andD. E. Rolston,Spec.Pub. 55, Am. Soc.of Agron.,in 1991. press, 1993. in relationto nutrient Large, P. J., Methylotrophy and Methanogenesis, Am. Soc. for Van Cleve, K., andL. A. Viereck,Forestsuccession Microbiol., Washington, D.C., 1983. cycling in theboreal forestof Alaska, in ForestSuccession: Concepts and Application, editedby D. West, H. Shugartand D. Botkin, pp. Maestrojuan, G. M., andD. R. Boone,Characterization of methanosarcina barkeri strains MST and 227, Methanosarcina mazei S-6T, and 185-211,Springer-Verlag, New York, 1981. Methanosarcina vacuolata Z-761 T, lnt. J. Syst.Bacteriol.,41,267-274, Van Cleve, K., F. S. Chapin III, C. T. Dyrness,and L. A. Viereck, 1991. Elementcyclingin taigaforests: State-factor control, Bioscience, 41,
Melillo, J. M., J. D. Aber, andJ. F. Muratore,Nitrogenandlignin controls 78-88, 1991. of hardwood leaf litter decompositiondynamics, Ecology, 63, Wahlen, M., N. Tanaka, R. Henry, T. Yoshinari, R. G. Fairbanks, A.
621-626, 1982.

Klinger, L. F., P. R. Zimmerman, J.P. Greenberg, L. E. Heidt, and A. B. Guenther,Carbontrace gasfluxes along a successional gradtentin the Hudson Bay lowland,J. Geophys. Res., thisissue. Klinger,L. F., S. A. Elias,V. M. Behan-Pelletier, andN. E. Williams, The bog climaxhypothesis: Fossilarthropod and stratigraphic evidence in peatsections from southeast Alaska, USA, HolarcticEcol.,13, 72-80,

Taylor, J. A., G. P. Brasseur, P. R. Zimmerman, and R. J. Cicerone,A study of the sources andsinks of methane andmethylchloroform using a global three-dimensional lagrangiantropospheric tracer transport model,J. Geophys. Res.,96, 3013-3044,1991. Tyler, S.C., Stablecarbonisotoperatiosin atmospheric methaneand some of its sources, J. Geophys. Res., 91, 13,232-13,238, 1986.

Tyler, S.C., 13C/12C ratios in atmospheric methane and some of its
sources., in StableIsotopes in Ecological Research, editedby P. W. Rundel, J. R. Ehleringer, andA. K. Nagy,Springer-Verlag, New York,
Ecol. Studies,68,395-409, 1989.

Shemesh, and W. S. Broecker, 13C,D and14C in methane
(abstract)EOS, Trans. AGU, 68, 1220, 1987.

Moore, T. R., and R. Knowles, Methane emissions from fen, bog, and

swamp peatlands in Quebec, Biogeochemistry, 11, 45-61, 1990.

Moore, T., N. Roulet, and R. Knowles,Spatialand temporalvariationsof Canada,Can. J. Earth Sci., 7, 317-325, 1970. methane flux from subarctic/northern borealfens,Global Biogeochem. Cycles, 4, 29-46, 1990. Westermann, P., B. K. Ahring,andR. A. Mah, Temperature compensation in Methanosarcina barkeri by modulation of hydrogen and acetate Morrissey, L.A., andA.P. Livingston, Methane emissions from Alaska affinity, Appl.Environ.Microbiol., 55, 1262-1266, 1989. Arctictundra:An assessment of localspatial variability, J. Geophys. Whalen, S.C., and W. S. Reeburgh, A methane flux transect along the Res., 97, 16,661-16,670, 1992. trans-Alaska pipelinehaul road,Tellus,42(B), 237-249, 1990. Mosier, A., D. Schimel, D. Valentine, K. Bronson, and W. Parton,
Methane and nitrous oxide fluxes in native, fertilized and cultivated

Webber, P. J., J. W. Richardson, andJ. T. Andrews, Postglacialuplift and substrate age at Cape Henrietta Maria, southeastern Hudson Bay,

Whiting, G. J., andJ.P. Chanton, Plant-dependent CH 4 emission in a
subarctic Canadian fen, GlobalBiogeochem. Cycles,6, 225-231, 1992. Williams,R. T., andR. L. Crawford,Methanogenic bacteria, includingan acid tolerantstrain,from peatlands, Appl. Environ. Microbiol., 39,
194-204, 1985.

grasslands, Nature, 350,330-332, 1991. Oremland,R. S., Biogeochemistry of mcthanogenic bacteria, in Biology of AnaerobicMicroorganisms, editedby A. J. B. Zehnder, pp. 641705, JohnWiley, New York, 1988. Parnas, H., Model for decomposition of organic material by microorganisms, SoilBiol. Biochem., 7, 161-169,1975. Patel,G. B., G. D. Sprott,andJ. E. Fein,Isolationandcharacterization of methanobactcrium espanolaesp. nov., a mesophilic, moderately acidophilic mcthanogen, lnt. J. Syst. Bacteriol., 40, 12-18,1990.

Yavitt, J. B., andG. E. Lang,Methaneproduction in contrasting wetland sites:Response to organic-chemical components of peatandto surfate
reduction,Geomicrobiol.J., 8, 27-46, 1990.

Quay,P. D., et al., Carbon isotopic composition of atxnosphehc CH4:
Fossil and biomassburning sourcestrengths,Global Biogeochem. Cycles,5, 25-47, 1991.
Roulet, N. T., R. Ash, and T. R. Moore, Low boreal wetlands as a source



and D.S.





of atmospheric methane, J. Geophys. Res.,97, 3739-3749,1992.
Roulet, N. T., J. Ritter, A. Jano,C. A. Kelly, L. F. Klinger, T. R. Moore, R. Protz, and W. R. Rouse, The role of the Hudson Bay lowland as a source of atmospheric methane, J. Geophys. Res., thisissue. Sass,R. L., F. M. Fischer,and P. A. Hatcombe,Methaneproduction and

Atmospheric Research, P.O. Box 3000, Boulder,CO 80307. D. Valentine,NaturalResource EcologyLaboratory, Colorado StateUniversity,Fort Collins, CO 80523.

emission ina Texas rice field, Global Biogeochem. Cycles, 4, 47-68, (Received May29, 1992;revised January 28, 1993;
1990. accepted February10, 1993.)