Biogeosciences 16 755ndash768 2019httpsdoiorg105194bg-16-755-2019copy Author(s) 2019 This work is distributed underthe Creative Commons Attribution 40 License

Multi-year effect of wetting on CH4 flux at taigandashtundra boundaryin northeastern Siberia deduced from stable isotope ratios of CH4

Ryo Shingubara1a Atsuko Sugimoto234 Jun Murase5 Go Iwahana26 Shunsuke Tei23 Maochang Liang1bShinya Takano1 Tomoki Morozumi1 and Trofim C Maximov78

1Graduate School of Environmental Science Hokkaido University Sapporo 060-0810 Japan2Arctic Research Center Hokkaido University Sapporo 001-0021 Japan3Faculty of Environmental Earth Science Hokkaido University Sapporo 060-0810 Japan4Global Station for Arctic Research Global Institution for Collaborative Researchand Education Hokkaido University Sapporo 060-0808 Japan5Graduate School of Bioagricultural Sciences Nagoya University Nagoya 464-8601 Japan6International Arctic Research Center University of Alaska Fairbanks Fairbanks 99775-7340 USA7Institute for Biological Problems of Cryolithozone Siberian Branch Russian Academy of Sciences Yakutsk 677890 Russia8Institute of Natural Sciences North-Eastern Federal University Yakutsk 677000 Russiaacurrently at Graduate School of Environmental Studies Nagoya University Nagoya 464-8601 Japanbcurrently at College of Horticulture and Gardening Yangtze University Jingzhou 434025 China

Correspondence Atsuko Sugimoto (atsukoseeshokudaiacjp)

Received 23 October 2018 ndash Discussion started 7 November 2018Revised 7 January 2019 ndash Accepted 24 January 2019 ndash Published 11 February 2019

Abstract The response of CH4 emission from natural wet-lands due to meteorological conditions is important becauseof its strong greenhouse effect To understand the relation-ship between CH4 flux and wetting we observed interan-nual variations in chamber CH4 flux as well as the concen-tration δ13C and δD of dissolved CH4 during the summerfrom 2009 to 2013 at the taigandashtundra boundary in the vicin-ity of Chokurdakh (7037prime N 14755prime E) located on the low-lands of the Indigirka River in northeastern Siberia We alsoconducted soil incubation experiments to interpret δ13C andδD of dissolved CH4 and to investigate variations in CH4production and oxidation processes Methane flux showedlarge interannual variations in wet areas of sphagnum mossesand sedges (36ndash140 mg CH4 mminus2 dayminus1 emitted) IncreasedCH4 emission was recorded in the summer of 2011 when awetting event with extreme precipitation occurred Althoughwater level decreased from 2011 to 2013 CH4 emission re-mained relatively high in 2012 and increased further in 2013Thaw depth became deeper from 2011 to 2013 which maypartly explain the increase in CH4 emission Moreover dis-solved CH4 concentration rose sharply by 1 order of mag-nitude from 2011 to 2012 and increased further from 2012

to 2013 Large variations in δ13C and δD of dissolved CH4were observed in 2011 and smaller variations were seenin 2012 and 2013 suggesting both enhancement of CH4 pro-duction and less significance of CH4 oxidation relative tothe larger pool of dissolved CH4 These multi-year effectsof wetting on CH4 dynamics may have been caused by con-tinued soil reduction across multiple years following the wet-ting Delayed activation of acetoclastic methanogenesis fol-lowing soil reduction could also have contributed to the en-hancement of CH4 production These processes suggest thatduration of water saturation in the active layer can be impor-tant for predicting CH4 emission following a wetting eventin the permafrost ecosystem

1 Introduction

Atmospheric CH4 has an important greenhouse effect(Myhre et al 2013) The largest source of atmospheric CH4is the emission from natural wetlands which is considered tobe the main driver of interannual variations in the global CH4emission depending on meteorological conditions such as

Published by Copernicus Publications on behalf of the European Geosciences Union

756 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

air temperature and precipitation (Ciais et al 2013) For in-stance Dlugokencky et al (2009) reported that high temper-atures in the Arctic and high precipitation in the tropics ledto high CH4 emissions from natural wetlands which causedthe observed large growth rates in atmospheric CH4 concen-tration during 2007 and 2008 Atmospheric CH4 has been in-creasing from 2007 through the present (Nisbet et al 2014)

Methane flux from wetland soil to the atmosphere (we de-fine a positive flux value as CH4 emission) is determinedby three processes CH4 production oxidation and trans-port (Lai 2009) Methane is produced by strictly anaer-obic Archaea (methanogens) mainly via hydrogenotrophicmethanogenesis (4H2+CO2rarr CH4+2H2O) or acetoclasticmethanogenesis (CH3COOHrarr CH4+CO2) as an end prod-uct of organic matter decomposition (Lai 2009) In the soilrsquosaerobic zone CH4 is oxidized to CO2 by methanotrophicbacteria utilizing O2 which reduces CH4 emission to the at-mosphere (Lai 2009) Underground CH4 is transported tothe atmosphere via bubble ebullition diffusion through soillayers and surface water and via aerenchyma of vascularplants (Lai 2009)

High water levels can lead to development of reducingconditions in soil which can promote CH4 production or de-press CH4 oxidation both leading to increases in CH4 flux(Lai 2009) This is reflected in the widely observed posi-tive relationship between water level and CH4 flux foundin a meta-analysis across the circum-Arctic permafrost zone(Olefeldt et al 2013) Meanwhile Desyatkin et al (2014)observed increases in CH4 flux during the second consecu-tive year of flooding at a thermokarst depression in borealeastern Siberia Treat et al (2007) reported observations ata temperate fen in the northeastern USA showing that highwater level coincided with high CH4 flux in interannual vari-ations However water level correlated negatively with CH4flux over shorter timescales namely as monthly means orindividual measurements These observational results implythat wetting is not directly related to CH4 flux in wetlands Tounderstand the relationship between wetting and CH4 flux itis necessary to assess the underlying processes

Stable isotopes of CH4 have been used to estimate produc-tion pathways of CH4 (Sugimoto and Wada 1993 1995 Mc-Calley et al 2014 Itoh et al 2015) determine the fractionof oxidized CH4 versus produced CH4 (Marik et al 2002Preuss et al 2013) and study mechanisms of CH4 transportby plants (Chanton 2005) When CH4 in soil is lost by oxi-dation or diffusion both δ13C and δD of the remaining CH4increase While the hydrogen isotope ratio increases morethan that of carbon during oxidation both ratios are consid-ered to change to the same extent during diffusion Thus it isuseful to analyze both carbon and hydrogen isotopes of CH4to distinguish the effects of both of these processes (Chanton2005)

The taigandashtundra boundary ecosystem (or transition zone)contains vegetation types of both taiga and tundra ecosys-tems Liang et al (2014) reported that the distribution of

vegetation types at the taigandashtundra boundary on the lowlandof the Indigirka River in northeastern Siberia is controlledby soil moisture which corresponds to microtopographyLarches the dominant tree species in the taiga forests of east-ern Siberia grow on microreliefs with higher ground leveland drier soil while wetland vegetation such as sphagnummosses and sedges typically seen in wet tundra (van Huis-steden et al 2005 van der Molen et al 2007) dominateslower and wetter microreliefs Thus it is reasonable to as-sume that the taigandashtundra boundary ecosystem has variousmicroreliefs in terms of interannual variation in soil wetnessconditions always wet microreliefs always dry microreliefsand microreliefs with large interannual wetness variationsHence this ecosystem is a suitable area to evaluate the pro-cesses controlling CH4 flux in relation to soil wetting andordrying on an interannual timescale

In this study to understand relationships between CH4flux and environmental factors we observed interannualvariations in chamber CH4 flux along with the concentra-tion δ13C and δD of dissolved CH4 during the summerfrom 2009 to 2013 at the taigandashtundra boundary located onIndigirka River lowlands in northeastern Siberia We alsoconducted soil incubation experiments to investigate howδ values of CH4 reflect CH4 production and oxidation pro-cesses in this ecosystem In 2011 a wetting event with asignificant amount of precipitation occurred We focused inparticular on the responses of CH4 flux and other underlyingprocesses to this unusual wetting event

2 Methods

21 Study sites

The taigandashtundra boundary on the lowlands of the IndigirkaRiver was selected as our study area Observations and sam-pling were conducted at three sites (V Verkhny Khatis-takha K Kodac B Boydom) in the vicinity of Chokurdakh(7037prime N 14755prime E) Republic of Sakha (Yakutia) Russia(Fig 1 and Table 1) The sites are located in the RussianArctic with an annual mean air temperature of minus139 Cand an annual mean precipitation of 208 mm for the pe-riod of 1950ndash2008 according to the Baseline Meteorologi-cal Data in Siberia database (Yabuki et al 2011) Sites V Kand B are beside the Indigirka River or its tributary and treedensity decreases from site V to site B

These study sites are underlain by continuous permafrost(Iwahana et al 2014) Normally snowmelt and the start ofactive layer thawing begin in the latter half of May throughthe first half of June and the growing season occurs fromthe end of June through the beginning of August Air tem-perature and surface soil temperature (10 cm in depth) peakin July whereas the maximum thaw depth occurs from thelatter half of August to the first half of September The freez-ing of the active layer starts in the latter half of September to

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 757

Table 1 Observation points of chamber CH4 flux Concentration and isotopic compositions of dissolved CH4 were also observed in thefollowing wet areas

Site Landscape Observation points and Dominant Volumetric Thawsurface conditions vegetation water depth

content ()b (cm)c

V (Verkhny tree mound_V Green moss 17plusmn 5 23plusmn 3Khatistakha) Larix gmelinii (n= 3) (n= 5)

7015prime N Larch forest sedge_V Carex spp 48plusmn 4 56plusmn 314728prime E and wetland (wet area) Eriophorum (n= 3) (n= 4)

angustifolium

tree mound_K Green moss 21plusmn 06 23plusmn 4K (Kodac)a Larix gmelinii (n= 4) (n= 9)

7034prime N Typical taigandash sphagnum_K Sphagnum 42plusmn 5 31plusmn 814816prime E tundra boundary (wet area) squarrosum (n= 6) (n= 15)

sedge_K Eriophorum 44plusmn 4 32plusmn 13(wet area) angustifolium (n= 6) (n= 28)

tree mound_B Green moss 6plusmn 2 20plusmn 4B (Boydom) Low-centered Larix gmelinii (n= 5) (n= 8)

7038prime N polygon sedge_B Eriophorum 46plusmn 2 36plusmn 914809prime E (wet area) angustifolium (n= 5) (n= 8)

a Site K was previously named Kryvaya (Iwahana et al 2014) or Kodak (Liang et al 2014) b Observed for the surface soil layer down to20 cm on 1 to 3 days in July 2011 at each observation point (see Table S2 for detailed observation dates) Standard deviations are shownc Observed from early July to early August during 2010ndash2013 (see Table S1 for the interannual variation and Table S2 for detailedobservation dates) Standard deviations are shown

Figure 1 Locations of study sites (a) Location of study region in northeastern Siberia (the Generic Mapping Tools 500) (b) Satellite imageof Indigirka River lowland around Chokurdakh village (Ckh 7037prime N 14755prime E) from Landsat 8 Observation sites (V K B) were selectedin this region beside the main stem and a tributary of the Indigirka River

October and the whole active layer freezes from Novemberto December

Observations of CH4 flux were conducted at seven pointswith three typical vegetation types as summarized in Ta-ble 1 These vegetation types are distributed in patches cor-

responding to microtopography and soil moisture (Liang etal 2014) Microrelief with a higher ground level is coveredby green moss larch trees and shrubs of willows or dwarfbirches Conversely lower microrelief is covered by wet-land vegetation of sphagnum moss or sedges In this study

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

758 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

the former vegetation type was termed ldquotree moundrdquo andthe latter type was termed ldquowet areardquo Observation pointsin tree mounds were selected at each of the sites V Kand B and termed ldquotree mound_Vrdquo ldquotree mound_Krdquo andldquotree mound_Brdquo (Table 1) For observation points of wet ar-eas a microrelief covered by sphagnum moss at site K wastermed ldquosphagnum_Krdquo and points covered by sedges includ-ing in particular cotton sedges (Eriophorum angustifolium)at sites V K and B were termed ldquosedge_Vrdquo ldquosedge_Krdquo andldquosedge_Brdquo respectively Measurements of volumetric watercontent in the surface soil layer (0ndash20 cm) by TDR (timedomain reflectometry TDR-341F Fujiwara Scientific Com-pany Japan) showed that tree mounds were drier than wetareas this will be described in Sect 31 (Table 1)

22 Field observations and samplings

Methane flux was observed using the chamber method ineach of the typical vegetation types described in Sect 21during the summer from 2009 to 2013 A transparent cylin-drical flux chamber (acrylic resin base area 47times 102 cm2height 25 cm) was installed on the ground The headspacegas of the chamber (ca 12 L) was circulated with a pump(ca 1 L minminus1) The chamber was closed for 15ndash30 min andheadspace gas was sampled two to three times after chamberclosure In most cases the chamber was closed for 30 minand headspace gas was collected at 0 15 and 30 min afterclosure Samples were kept in pre-evacuated glass vials withbutyl rubber septa To minimize soil disturbance we steppedon wooden boards at observation points In 2009 and 2010CH4 flux measurements were conducted in the latter half ofJuly and from 2011 to 2013 observations were conductedcontinuously from early July to the end of July or early Au-gust For all of these years the observation period includedthe warmest season when CH4 emission was expected to bethe most active (Table S2)

For measurements of dissolved CH4 surface water andsoil pore water were sampled in wet areas from 2011 to 2013Surface water was directly taken up by a 50 mL plastic sy-ringe with a three-way cock attached to its tip whereas soilpore water was sampled by a 50 mL syringe (with a three-way cock attached) through a plastic tube inserted in the soilSoon after collecting water samples dissolved CH4 was ex-tracted inside the syringes with the headspace method af-ter adding 15ndash35 mL of the atmosphere prepared in a 10 Laluminum bag This atmosphere was collected beforehandat Chokurdakh village or our observation sites and filteredusing Molecular Sieves 5Aring (116 pellets FUJIFILM WakoPure Chemical Corporation Japan) The atmosphere was an-alyzed later for CH4 concentration and isotopic compositionsas a background sample (20ndash43 ppm for CH4 concentra-tion minus53 permil to minus45 permil for δ13C of CH4 and minus168 permil tominus78 permil for δD of CH4) The syringes were vigorously shakenfor 1 min and left standing for 5 min to ensure equilibration

Finally headspace gas in the syringes was preserved in 10ndash20 mL pre-evacuated glass vials with rubber septa

Concurrently with each flux measurement soil tempera-ture around the flux chamber was measured with a temper-ature sensor in an ORP electrode (PST-2739C DKK-TOACorporation Japan) with an ORP meter (RM-30P or RM-20P) After flux measurement samples were collected thawdepth was observed on the same day around each chamberby inserting a steel rod into the ground From 2011 on wa-ter level was also measured after flux measurements aroundeach chamber in wet areas using a scale The water level wasexpressed as height relative to the ground surface or the mosssurface Observation dates of these environmental factors areshown in Table S2

23 Soil incubation experiments and microbialcommunity analysis

Soil incubation experiments were conducted to estimate δ13Cand δD of produced CH4 and fractionation factors of CH4oxidation for carbon and hydrogen isotopes For CH4 pro-duction experiments surface soil was sampled in all thewet areas in Table 1 (sedge_V sphagnum_K sedge_K andsedge_B) during summer 2013 Samples were taken at 10 cmin depth at each sampling location To observe vertical vari-ations in δ values of produced CH4 within the thaw layer wealso collected samples from two additional depths (20 and30 cm) at sedge_K which is a location typical of the taigandashtundra boundary region These samples were from organiclayers except for the samples from 30 cm which were fromthe top of the mineral layer

Approximately 10 mL of soil was directly transferred intoeach plastic syringe (60 mL maximum capacity) along within situ water (approximately 50 mL) to prevent the samplefrom being oxidized by the atmosphere Syringes were pre-served in water to ensure no leakage and were immediatelypreincubated for 4ndash8 days and then incubated in triplicate for8 days Preincubation and incubation temperatures were setat 5 C We also incubated syringes at 10 C for samples from10 cm in depth at sedge_K to investigate temperature depen-dence of δ values of produced CH4 For each of these sevenincubation treatments (sphagnum_K sedge_V sedge_K andsphagnum_K 10 cm in depth 5 C sedge_K 20 and 30 cmin depth 5 C sedge_K 10 cm in depth 10 C) three repli-cate soil samples were prepared Water in each incubationsyringe was sampled twice at the start and the end of incuba-tion and dissolved CH4 was extracted using the headspacemethod described in Sect 22 As a consequence dissolvedCH4 samples were collected in triplicate for each of the ini-tial and final conditions of one incubation treatment

To interpret CH4 production in the incubation experiments(Sect 23) phylogenic composition of methanogens in thesurface soil was additionally analyzed in 2016 using 16SrRNA gene sequencing In July 2016 soil samples from10 cm in depth were collected in 10 mL plastic tubes in trip-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 759

licate in the same four wet areas as the anaerobic incuba-tion experiments and kept frozen until analysis DNA wasextracted from 3 g of the soil samples as described by Ikeda etal (2004) Extracted DNA was purified using the OneSteptrade

PCR Inhibitor Removal Kit (Zymo Research California) andquantified using the Quant-iT PicoGreen dsDNA assay kit(Invitrogen Carlsbad California) Amplicon sequencing wasconducted targeting the V3ndashV4 regions of 16S rRNA genes(Caporaso et al 2011) Sequences obtained were processedthrough the QIIME pipeline (Caporaso et al 2010) A rep-resentative sequence was picked from each operational tax-onomic unit (OTU) and the Greengenes reference database(version 138) was used to assign taxonomic information andcalculate the relative abundance of methanogenic archaeapresent

For CH4 oxidation surface organic layers (0ndash13 cm indepth) were sampled at sphagnum_K and sedge_K inJuly 2012 and then kept in a refrigerator until the experi-ment (6 days) These soil samples were cut into small piecesand mixed well with air A total of 10 g (about 40 mL) ofsoil sample was transferred into plastic syringes (maximum120 mL) in quadruplicate for each sampling location Ap-proximately 80 mL of air and 02ndash2 mL of 25 CH4 gaswere added to each syringe so the total volume in each sy-ringe was 120 mL and the headspace CH4 concentration was50times 102ndash48times 103 ppm Syringes were preserved in waterand incubated at 8 C for 8 days Headspace gas was sampledon day 0 day 4 and day 8 from each syringe into 20 mL pre-evacuated glass vials with rubber septa Consequently qua-druplicate gas samples were collected for each location andeach sampling day

24 Sample analysis and data processing

Methane concentrations in air samples were analyzed us-ing a gas chromatograph (HP6890 series G1530A HewlettPackard USA) equipped with a flame ionization detector anda CP-carboPLOT capillary column (Varian USA) Methaneflux was calculated from CH4 concentration in chamberheadspace by a linear regression of two to three concentra-tion values against the time elapsed since chamber closureThe detection limit of CH4 flux for each observation was cal-culated as 08ndash24 mg CH4 mminus2 dayminus1 based on whether thechange of chamber CH4 concentration during the observationwas significant relative to the precision of CH4 concentrationanalysis Regression r2 was calculated (formally) as ge 087when the flux value was larger than 2 mg CH4 mminus2 dayminus1Dissolved CH4 concentrations were obtained from calcula-tion of the headspace method for which equilibrations ofCH4 between gas and water phases are described by the Bun-sen absorption coefficient of CH4 (Yamamoto et al 1976)

Carbon and hydrogen isotope ratios of in situ dissolvedCH4 and CH4 samples from both incubation experimentswere analyzed on a GCndashGCndashCndashIRMS (modified after Sug-imoto 1996) ndash which is a continuous flow system con-

sisting of two gas chromatographs a combustion reactorand an isotope ratio mass spectrometer (MAT253 ThermoFisher Scientific USA) ndash and on a GCndashGCndashPndashIRMS (P py-rolysis in a HTC reactor of GC IsoLink Thermo FisherScientific) respectively Carbon and hydrogen isotope ra-tios obtained were represented relative to Vienna PeedeeBelemnite (VPDB) and Vienna Standard Mean Ocean Wa-ter (VSMOW) respectively Precisions of the analyses wereplusmn02 permil and plusmn2 permil for δ13C and δD respectively When cal-culating δ13C and δD of dissolved CH4 the effect of CH4 inbackground air was removed based on the mass balance Inthe aerobic incubation experiments the fractionation factorsof CH4 oxidation for carbon and hydrogen were calculatedusing the following Rayleigh distillation equation

lnRt

R0=

(1αoxminus 1

)ln

[CH4]t[CH4]0

(1)

where R0 and Rt represent isotope ratios under initial condi-tions and at time t respectively αox is the fractionation factorfor CH4 oxidation (defined so that αox gt 1) and [CH4]0 and[CH4]t are CH4 concentrations under initial conditions andat time t respectively

All statistical tests for detecting differences in CH4 fluxesor dissolved CH4 concentrations were conducted usingR software (version 333) MannndashWhitneyrsquos U test was ap-plied to compare magnitudes between 2 years of data andSteelndashDwassrsquos multiple comparison test was used to comparemagnitudes among 3 years or more of data

25 Meteorological data

Air temperature and precipitation observed at a weather sta-tion in Chokurdakh (WMO station 21946) were used to in-vestigate interannual variations in meteorological conditionsduring our observation period of CH4 flux (2009ndash2013) andduring the preceding 2 years (2007ndash2008) The distance be-tween the weather station and our farthest observation site(site V) is approximately 45 km (Fig 1) These data were ob-tained from GHCN-Daily a NOAA database (Menne et al2012a b)

3 Results

31 Environmental factors

Soil wetness conditions and thaw depth differed among veg-etation types (Table 1) Tree mounds had lower surface watercontent (21 ndash17 ) than wet areas (42 ndash48 ) Amongthe two types of wet areas the water level was lower in wetareas of sphagnum mosses than those of sedges (Fig 2)Wet areas of sedges experienced water levels higher thanthe ground surface (defined as 0 cm) reaching more than10 cm above the ground surface Corresponding with soil wa-ter content the thaw depth was shallower at dry tree mounds

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

756 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

air temperature and precipitation (Ciais et al 2013) For in-stance Dlugokencky et al (2009) reported that high temper-atures in the Arctic and high precipitation in the tropics ledto high CH4 emissions from natural wetlands which causedthe observed large growth rates in atmospheric CH4 concen-tration during 2007 and 2008 Atmospheric CH4 has been in-creasing from 2007 through the present (Nisbet et al 2014)

Methane flux from wetland soil to the atmosphere (we de-fine a positive flux value as CH4 emission) is determinedby three processes CH4 production oxidation and trans-port (Lai 2009) Methane is produced by strictly anaer-obic Archaea (methanogens) mainly via hydrogenotrophicmethanogenesis (4H2+CO2rarr CH4+2H2O) or acetoclasticmethanogenesis (CH3COOHrarr CH4+CO2) as an end prod-uct of organic matter decomposition (Lai 2009) In the soilrsquosaerobic zone CH4 is oxidized to CO2 by methanotrophicbacteria utilizing O2 which reduces CH4 emission to the at-mosphere (Lai 2009) Underground CH4 is transported tothe atmosphere via bubble ebullition diffusion through soillayers and surface water and via aerenchyma of vascularplants (Lai 2009)

High water levels can lead to development of reducingconditions in soil which can promote CH4 production or de-press CH4 oxidation both leading to increases in CH4 flux(Lai 2009) This is reflected in the widely observed posi-tive relationship between water level and CH4 flux foundin a meta-analysis across the circum-Arctic permafrost zone(Olefeldt et al 2013) Meanwhile Desyatkin et al (2014)observed increases in CH4 flux during the second consecu-tive year of flooding at a thermokarst depression in borealeastern Siberia Treat et al (2007) reported observations ata temperate fen in the northeastern USA showing that highwater level coincided with high CH4 flux in interannual vari-ations However water level correlated negatively with CH4flux over shorter timescales namely as monthly means orindividual measurements These observational results implythat wetting is not directly related to CH4 flux in wetlands Tounderstand the relationship between wetting and CH4 flux itis necessary to assess the underlying processes

Stable isotopes of CH4 have been used to estimate produc-tion pathways of CH4 (Sugimoto and Wada 1993 1995 Mc-Calley et al 2014 Itoh et al 2015) determine the fractionof oxidized CH4 versus produced CH4 (Marik et al 2002Preuss et al 2013) and study mechanisms of CH4 transportby plants (Chanton 2005) When CH4 in soil is lost by oxi-dation or diffusion both δ13C and δD of the remaining CH4increase While the hydrogen isotope ratio increases morethan that of carbon during oxidation both ratios are consid-ered to change to the same extent during diffusion Thus it isuseful to analyze both carbon and hydrogen isotopes of CH4to distinguish the effects of both of these processes (Chanton2005)

The taigandashtundra boundary ecosystem (or transition zone)contains vegetation types of both taiga and tundra ecosys-tems Liang et al (2014) reported that the distribution of

vegetation types at the taigandashtundra boundary on the lowlandof the Indigirka River in northeastern Siberia is controlledby soil moisture which corresponds to microtopographyLarches the dominant tree species in the taiga forests of east-ern Siberia grow on microreliefs with higher ground leveland drier soil while wetland vegetation such as sphagnummosses and sedges typically seen in wet tundra (van Huis-steden et al 2005 van der Molen et al 2007) dominateslower and wetter microreliefs Thus it is reasonable to as-sume that the taigandashtundra boundary ecosystem has variousmicroreliefs in terms of interannual variation in soil wetnessconditions always wet microreliefs always dry microreliefsand microreliefs with large interannual wetness variationsHence this ecosystem is a suitable area to evaluate the pro-cesses controlling CH4 flux in relation to soil wetting andordrying on an interannual timescale

In this study to understand relationships between CH4flux and environmental factors we observed interannualvariations in chamber CH4 flux along with the concentra-tion δ13C and δD of dissolved CH4 during the summerfrom 2009 to 2013 at the taigandashtundra boundary located onIndigirka River lowlands in northeastern Siberia We alsoconducted soil incubation experiments to investigate howδ values of CH4 reflect CH4 production and oxidation pro-cesses in this ecosystem In 2011 a wetting event with asignificant amount of precipitation occurred We focused inparticular on the responses of CH4 flux and other underlyingprocesses to this unusual wetting event

2 Methods

21 Study sites

The taigandashtundra boundary on the lowlands of the IndigirkaRiver was selected as our study area Observations and sam-pling were conducted at three sites (V Verkhny Khatis-takha K Kodac B Boydom) in the vicinity of Chokurdakh(7037prime N 14755prime E) Republic of Sakha (Yakutia) Russia(Fig 1 and Table 1) The sites are located in the RussianArctic with an annual mean air temperature of minus139 Cand an annual mean precipitation of 208 mm for the pe-riod of 1950ndash2008 according to the Baseline Meteorologi-cal Data in Siberia database (Yabuki et al 2011) Sites V Kand B are beside the Indigirka River or its tributary and treedensity decreases from site V to site B

These study sites are underlain by continuous permafrost(Iwahana et al 2014) Normally snowmelt and the start ofactive layer thawing begin in the latter half of May throughthe first half of June and the growing season occurs fromthe end of June through the beginning of August Air tem-perature and surface soil temperature (10 cm in depth) peakin July whereas the maximum thaw depth occurs from thelatter half of August to the first half of September The freez-ing of the active layer starts in the latter half of September to

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 757

Table 1 Observation points of chamber CH4 flux Concentration and isotopic compositions of dissolved CH4 were also observed in thefollowing wet areas

Site Landscape Observation points and Dominant Volumetric Thawsurface conditions vegetation water depth

content ()b (cm)c

V (Verkhny tree mound_V Green moss 17plusmn 5 23plusmn 3Khatistakha) Larix gmelinii (n= 3) (n= 5)

7015prime N Larch forest sedge_V Carex spp 48plusmn 4 56plusmn 314728prime E and wetland (wet area) Eriophorum (n= 3) (n= 4)

angustifolium

tree mound_K Green moss 21plusmn 06 23plusmn 4K (Kodac)a Larix gmelinii (n= 4) (n= 9)

7034prime N Typical taigandash sphagnum_K Sphagnum 42plusmn 5 31plusmn 814816prime E tundra boundary (wet area) squarrosum (n= 6) (n= 15)

sedge_K Eriophorum 44plusmn 4 32plusmn 13(wet area) angustifolium (n= 6) (n= 28)

tree mound_B Green moss 6plusmn 2 20plusmn 4B (Boydom) Low-centered Larix gmelinii (n= 5) (n= 8)

7038prime N polygon sedge_B Eriophorum 46plusmn 2 36plusmn 914809prime E (wet area) angustifolium (n= 5) (n= 8)

a Site K was previously named Kryvaya (Iwahana et al 2014) or Kodak (Liang et al 2014) b Observed for the surface soil layer down to20 cm on 1 to 3 days in July 2011 at each observation point (see Table S2 for detailed observation dates) Standard deviations are shownc Observed from early July to early August during 2010ndash2013 (see Table S1 for the interannual variation and Table S2 for detailedobservation dates) Standard deviations are shown

Figure 1 Locations of study sites (a) Location of study region in northeastern Siberia (the Generic Mapping Tools 500) (b) Satellite imageof Indigirka River lowland around Chokurdakh village (Ckh 7037prime N 14755prime E) from Landsat 8 Observation sites (V K B) were selectedin this region beside the main stem and a tributary of the Indigirka River

October and the whole active layer freezes from Novemberto December

Observations of CH4 flux were conducted at seven pointswith three typical vegetation types as summarized in Ta-ble 1 These vegetation types are distributed in patches cor-

responding to microtopography and soil moisture (Liang etal 2014) Microrelief with a higher ground level is coveredby green moss larch trees and shrubs of willows or dwarfbirches Conversely lower microrelief is covered by wet-land vegetation of sphagnum moss or sedges In this study

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

758 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

the former vegetation type was termed ldquotree moundrdquo andthe latter type was termed ldquowet areardquo Observation pointsin tree mounds were selected at each of the sites V Kand B and termed ldquotree mound_Vrdquo ldquotree mound_Krdquo andldquotree mound_Brdquo (Table 1) For observation points of wet ar-eas a microrelief covered by sphagnum moss at site K wastermed ldquosphagnum_Krdquo and points covered by sedges includ-ing in particular cotton sedges (Eriophorum angustifolium)at sites V K and B were termed ldquosedge_Vrdquo ldquosedge_Krdquo andldquosedge_Brdquo respectively Measurements of volumetric watercontent in the surface soil layer (0ndash20 cm) by TDR (timedomain reflectometry TDR-341F Fujiwara Scientific Com-pany Japan) showed that tree mounds were drier than wetareas this will be described in Sect 31 (Table 1)

22 Field observations and samplings

Methane flux was observed using the chamber method ineach of the typical vegetation types described in Sect 21during the summer from 2009 to 2013 A transparent cylin-drical flux chamber (acrylic resin base area 47times 102 cm2height 25 cm) was installed on the ground The headspacegas of the chamber (ca 12 L) was circulated with a pump(ca 1 L minminus1) The chamber was closed for 15ndash30 min andheadspace gas was sampled two to three times after chamberclosure In most cases the chamber was closed for 30 minand headspace gas was collected at 0 15 and 30 min afterclosure Samples were kept in pre-evacuated glass vials withbutyl rubber septa To minimize soil disturbance we steppedon wooden boards at observation points In 2009 and 2010CH4 flux measurements were conducted in the latter half ofJuly and from 2011 to 2013 observations were conductedcontinuously from early July to the end of July or early Au-gust For all of these years the observation period includedthe warmest season when CH4 emission was expected to bethe most active (Table S2)

For measurements of dissolved CH4 surface water andsoil pore water were sampled in wet areas from 2011 to 2013Surface water was directly taken up by a 50 mL plastic sy-ringe with a three-way cock attached to its tip whereas soilpore water was sampled by a 50 mL syringe (with a three-way cock attached) through a plastic tube inserted in the soilSoon after collecting water samples dissolved CH4 was ex-tracted inside the syringes with the headspace method af-ter adding 15ndash35 mL of the atmosphere prepared in a 10 Laluminum bag This atmosphere was collected beforehandat Chokurdakh village or our observation sites and filteredusing Molecular Sieves 5Aring (116 pellets FUJIFILM WakoPure Chemical Corporation Japan) The atmosphere was an-alyzed later for CH4 concentration and isotopic compositionsas a background sample (20ndash43 ppm for CH4 concentra-tion minus53 permil to minus45 permil for δ13C of CH4 and minus168 permil tominus78 permil for δD of CH4) The syringes were vigorously shakenfor 1 min and left standing for 5 min to ensure equilibration

Finally headspace gas in the syringes was preserved in 10ndash20 mL pre-evacuated glass vials with rubber septa

Concurrently with each flux measurement soil tempera-ture around the flux chamber was measured with a temper-ature sensor in an ORP electrode (PST-2739C DKK-TOACorporation Japan) with an ORP meter (RM-30P or RM-20P) After flux measurement samples were collected thawdepth was observed on the same day around each chamberby inserting a steel rod into the ground From 2011 on wa-ter level was also measured after flux measurements aroundeach chamber in wet areas using a scale The water level wasexpressed as height relative to the ground surface or the mosssurface Observation dates of these environmental factors areshown in Table S2

23 Soil incubation experiments and microbialcommunity analysis

Soil incubation experiments were conducted to estimate δ13Cand δD of produced CH4 and fractionation factors of CH4oxidation for carbon and hydrogen isotopes For CH4 pro-duction experiments surface soil was sampled in all thewet areas in Table 1 (sedge_V sphagnum_K sedge_K andsedge_B) during summer 2013 Samples were taken at 10 cmin depth at each sampling location To observe vertical vari-ations in δ values of produced CH4 within the thaw layer wealso collected samples from two additional depths (20 and30 cm) at sedge_K which is a location typical of the taigandashtundra boundary region These samples were from organiclayers except for the samples from 30 cm which were fromthe top of the mineral layer

Approximately 10 mL of soil was directly transferred intoeach plastic syringe (60 mL maximum capacity) along within situ water (approximately 50 mL) to prevent the samplefrom being oxidized by the atmosphere Syringes were pre-served in water to ensure no leakage and were immediatelypreincubated for 4ndash8 days and then incubated in triplicate for8 days Preincubation and incubation temperatures were setat 5 C We also incubated syringes at 10 C for samples from10 cm in depth at sedge_K to investigate temperature depen-dence of δ values of produced CH4 For each of these sevenincubation treatments (sphagnum_K sedge_V sedge_K andsphagnum_K 10 cm in depth 5 C sedge_K 20 and 30 cmin depth 5 C sedge_K 10 cm in depth 10 C) three repli-cate soil samples were prepared Water in each incubationsyringe was sampled twice at the start and the end of incuba-tion and dissolved CH4 was extracted using the headspacemethod described in Sect 22 As a consequence dissolvedCH4 samples were collected in triplicate for each of the ini-tial and final conditions of one incubation treatment

To interpret CH4 production in the incubation experiments(Sect 23) phylogenic composition of methanogens in thesurface soil was additionally analyzed in 2016 using 16SrRNA gene sequencing In July 2016 soil samples from10 cm in depth were collected in 10 mL plastic tubes in trip-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 759

licate in the same four wet areas as the anaerobic incuba-tion experiments and kept frozen until analysis DNA wasextracted from 3 g of the soil samples as described by Ikeda etal (2004) Extracted DNA was purified using the OneSteptrade

PCR Inhibitor Removal Kit (Zymo Research California) andquantified using the Quant-iT PicoGreen dsDNA assay kit(Invitrogen Carlsbad California) Amplicon sequencing wasconducted targeting the V3ndashV4 regions of 16S rRNA genes(Caporaso et al 2011) Sequences obtained were processedthrough the QIIME pipeline (Caporaso et al 2010) A rep-resentative sequence was picked from each operational tax-onomic unit (OTU) and the Greengenes reference database(version 138) was used to assign taxonomic information andcalculate the relative abundance of methanogenic archaeapresent

For CH4 oxidation surface organic layers (0ndash13 cm indepth) were sampled at sphagnum_K and sedge_K inJuly 2012 and then kept in a refrigerator until the experi-ment (6 days) These soil samples were cut into small piecesand mixed well with air A total of 10 g (about 40 mL) ofsoil sample was transferred into plastic syringes (maximum120 mL) in quadruplicate for each sampling location Ap-proximately 80 mL of air and 02ndash2 mL of 25 CH4 gaswere added to each syringe so the total volume in each sy-ringe was 120 mL and the headspace CH4 concentration was50times 102ndash48times 103 ppm Syringes were preserved in waterand incubated at 8 C for 8 days Headspace gas was sampledon day 0 day 4 and day 8 from each syringe into 20 mL pre-evacuated glass vials with rubber septa Consequently qua-druplicate gas samples were collected for each location andeach sampling day

24 Sample analysis and data processing

Methane concentrations in air samples were analyzed us-ing a gas chromatograph (HP6890 series G1530A HewlettPackard USA) equipped with a flame ionization detector anda CP-carboPLOT capillary column (Varian USA) Methaneflux was calculated from CH4 concentration in chamberheadspace by a linear regression of two to three concentra-tion values against the time elapsed since chamber closureThe detection limit of CH4 flux for each observation was cal-culated as 08ndash24 mg CH4 mminus2 dayminus1 based on whether thechange of chamber CH4 concentration during the observationwas significant relative to the precision of CH4 concentrationanalysis Regression r2 was calculated (formally) as ge 087when the flux value was larger than 2 mg CH4 mminus2 dayminus1Dissolved CH4 concentrations were obtained from calcula-tion of the headspace method for which equilibrations ofCH4 between gas and water phases are described by the Bun-sen absorption coefficient of CH4 (Yamamoto et al 1976)

Carbon and hydrogen isotope ratios of in situ dissolvedCH4 and CH4 samples from both incubation experimentswere analyzed on a GCndashGCndashCndashIRMS (modified after Sug-imoto 1996) ndash which is a continuous flow system con-

sisting of two gas chromatographs a combustion reactorand an isotope ratio mass spectrometer (MAT253 ThermoFisher Scientific USA) ndash and on a GCndashGCndashPndashIRMS (P py-rolysis in a HTC reactor of GC IsoLink Thermo FisherScientific) respectively Carbon and hydrogen isotope ra-tios obtained were represented relative to Vienna PeedeeBelemnite (VPDB) and Vienna Standard Mean Ocean Wa-ter (VSMOW) respectively Precisions of the analyses wereplusmn02 permil and plusmn2 permil for δ13C and δD respectively When cal-culating δ13C and δD of dissolved CH4 the effect of CH4 inbackground air was removed based on the mass balance Inthe aerobic incubation experiments the fractionation factorsof CH4 oxidation for carbon and hydrogen were calculatedusing the following Rayleigh distillation equation

lnRt

R0=

(1αoxminus 1

)ln

[CH4]t[CH4]0

(1)

where R0 and Rt represent isotope ratios under initial condi-tions and at time t respectively αox is the fractionation factorfor CH4 oxidation (defined so that αox gt 1) and [CH4]0 and[CH4]t are CH4 concentrations under initial conditions andat time t respectively

All statistical tests for detecting differences in CH4 fluxesor dissolved CH4 concentrations were conducted usingR software (version 333) MannndashWhitneyrsquos U test was ap-plied to compare magnitudes between 2 years of data andSteelndashDwassrsquos multiple comparison test was used to comparemagnitudes among 3 years or more of data

25 Meteorological data

Air temperature and precipitation observed at a weather sta-tion in Chokurdakh (WMO station 21946) were used to in-vestigate interannual variations in meteorological conditionsduring our observation period of CH4 flux (2009ndash2013) andduring the preceding 2 years (2007ndash2008) The distance be-tween the weather station and our farthest observation site(site V) is approximately 45 km (Fig 1) These data were ob-tained from GHCN-Daily a NOAA database (Menne et al2012a b)

3 Results

31 Environmental factors

Soil wetness conditions and thaw depth differed among veg-etation types (Table 1) Tree mounds had lower surface watercontent (21 ndash17 ) than wet areas (42 ndash48 ) Amongthe two types of wet areas the water level was lower in wetareas of sphagnum mosses than those of sedges (Fig 2)Wet areas of sedges experienced water levels higher thanthe ground surface (defined as 0 cm) reaching more than10 cm above the ground surface Corresponding with soil wa-ter content the thaw depth was shallower at dry tree mounds

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 757

Table 1 Observation points of chamber CH4 flux Concentration and isotopic compositions of dissolved CH4 were also observed in thefollowing wet areas

Site Landscape Observation points and Dominant Volumetric Thawsurface conditions vegetation water depth

content ()b (cm)c

V (Verkhny tree mound_V Green moss 17plusmn 5 23plusmn 3Khatistakha) Larix gmelinii (n= 3) (n= 5)

7015prime N Larch forest sedge_V Carex spp 48plusmn 4 56plusmn 314728prime E and wetland (wet area) Eriophorum (n= 3) (n= 4)

angustifolium

tree mound_K Green moss 21plusmn 06 23plusmn 4K (Kodac)a Larix gmelinii (n= 4) (n= 9)

7034prime N Typical taigandash sphagnum_K Sphagnum 42plusmn 5 31plusmn 814816prime E tundra boundary (wet area) squarrosum (n= 6) (n= 15)

sedge_K Eriophorum 44plusmn 4 32plusmn 13(wet area) angustifolium (n= 6) (n= 28)

tree mound_B Green moss 6plusmn 2 20plusmn 4B (Boydom) Low-centered Larix gmelinii (n= 5) (n= 8)

7038prime N polygon sedge_B Eriophorum 46plusmn 2 36plusmn 914809prime E (wet area) angustifolium (n= 5) (n= 8)

a Site K was previously named Kryvaya (Iwahana et al 2014) or Kodak (Liang et al 2014) b Observed for the surface soil layer down to20 cm on 1 to 3 days in July 2011 at each observation point (see Table S2 for detailed observation dates) Standard deviations are shownc Observed from early July to early August during 2010ndash2013 (see Table S1 for the interannual variation and Table S2 for detailedobservation dates) Standard deviations are shown

Figure 1 Locations of study sites (a) Location of study region in northeastern Siberia (the Generic Mapping Tools 500) (b) Satellite imageof Indigirka River lowland around Chokurdakh village (Ckh 7037prime N 14755prime E) from Landsat 8 Observation sites (V K B) were selectedin this region beside the main stem and a tributary of the Indigirka River

October and the whole active layer freezes from Novemberto December

Observations of CH4 flux were conducted at seven pointswith three typical vegetation types as summarized in Ta-ble 1 These vegetation types are distributed in patches cor-

responding to microtopography and soil moisture (Liang etal 2014) Microrelief with a higher ground level is coveredby green moss larch trees and shrubs of willows or dwarfbirches Conversely lower microrelief is covered by wet-land vegetation of sphagnum moss or sedges In this study

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

758 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

the former vegetation type was termed ldquotree moundrdquo andthe latter type was termed ldquowet areardquo Observation pointsin tree mounds were selected at each of the sites V Kand B and termed ldquotree mound_Vrdquo ldquotree mound_Krdquo andldquotree mound_Brdquo (Table 1) For observation points of wet ar-eas a microrelief covered by sphagnum moss at site K wastermed ldquosphagnum_Krdquo and points covered by sedges includ-ing in particular cotton sedges (Eriophorum angustifolium)at sites V K and B were termed ldquosedge_Vrdquo ldquosedge_Krdquo andldquosedge_Brdquo respectively Measurements of volumetric watercontent in the surface soil layer (0ndash20 cm) by TDR (timedomain reflectometry TDR-341F Fujiwara Scientific Com-pany Japan) showed that tree mounds were drier than wetareas this will be described in Sect 31 (Table 1)

22 Field observations and samplings

Methane flux was observed using the chamber method ineach of the typical vegetation types described in Sect 21during the summer from 2009 to 2013 A transparent cylin-drical flux chamber (acrylic resin base area 47times 102 cm2height 25 cm) was installed on the ground The headspacegas of the chamber (ca 12 L) was circulated with a pump(ca 1 L minminus1) The chamber was closed for 15ndash30 min andheadspace gas was sampled two to three times after chamberclosure In most cases the chamber was closed for 30 minand headspace gas was collected at 0 15 and 30 min afterclosure Samples were kept in pre-evacuated glass vials withbutyl rubber septa To minimize soil disturbance we steppedon wooden boards at observation points In 2009 and 2010CH4 flux measurements were conducted in the latter half ofJuly and from 2011 to 2013 observations were conductedcontinuously from early July to the end of July or early Au-gust For all of these years the observation period includedthe warmest season when CH4 emission was expected to bethe most active (Table S2)

For measurements of dissolved CH4 surface water andsoil pore water were sampled in wet areas from 2011 to 2013Surface water was directly taken up by a 50 mL plastic sy-ringe with a three-way cock attached to its tip whereas soilpore water was sampled by a 50 mL syringe (with a three-way cock attached) through a plastic tube inserted in the soilSoon after collecting water samples dissolved CH4 was ex-tracted inside the syringes with the headspace method af-ter adding 15ndash35 mL of the atmosphere prepared in a 10 Laluminum bag This atmosphere was collected beforehandat Chokurdakh village or our observation sites and filteredusing Molecular Sieves 5Aring (116 pellets FUJIFILM WakoPure Chemical Corporation Japan) The atmosphere was an-alyzed later for CH4 concentration and isotopic compositionsas a background sample (20ndash43 ppm for CH4 concentra-tion minus53 permil to minus45 permil for δ13C of CH4 and minus168 permil tominus78 permil for δD of CH4) The syringes were vigorously shakenfor 1 min and left standing for 5 min to ensure equilibration

Finally headspace gas in the syringes was preserved in 10ndash20 mL pre-evacuated glass vials with rubber septa

Concurrently with each flux measurement soil tempera-ture around the flux chamber was measured with a temper-ature sensor in an ORP electrode (PST-2739C DKK-TOACorporation Japan) with an ORP meter (RM-30P or RM-20P) After flux measurement samples were collected thawdepth was observed on the same day around each chamberby inserting a steel rod into the ground From 2011 on wa-ter level was also measured after flux measurements aroundeach chamber in wet areas using a scale The water level wasexpressed as height relative to the ground surface or the mosssurface Observation dates of these environmental factors areshown in Table S2

23 Soil incubation experiments and microbialcommunity analysis

Soil incubation experiments were conducted to estimate δ13Cand δD of produced CH4 and fractionation factors of CH4oxidation for carbon and hydrogen isotopes For CH4 pro-duction experiments surface soil was sampled in all thewet areas in Table 1 (sedge_V sphagnum_K sedge_K andsedge_B) during summer 2013 Samples were taken at 10 cmin depth at each sampling location To observe vertical vari-ations in δ values of produced CH4 within the thaw layer wealso collected samples from two additional depths (20 and30 cm) at sedge_K which is a location typical of the taigandashtundra boundary region These samples were from organiclayers except for the samples from 30 cm which were fromthe top of the mineral layer

Approximately 10 mL of soil was directly transferred intoeach plastic syringe (60 mL maximum capacity) along within situ water (approximately 50 mL) to prevent the samplefrom being oxidized by the atmosphere Syringes were pre-served in water to ensure no leakage and were immediatelypreincubated for 4ndash8 days and then incubated in triplicate for8 days Preincubation and incubation temperatures were setat 5 C We also incubated syringes at 10 C for samples from10 cm in depth at sedge_K to investigate temperature depen-dence of δ values of produced CH4 For each of these sevenincubation treatments (sphagnum_K sedge_V sedge_K andsphagnum_K 10 cm in depth 5 C sedge_K 20 and 30 cmin depth 5 C sedge_K 10 cm in depth 10 C) three repli-cate soil samples were prepared Water in each incubationsyringe was sampled twice at the start and the end of incuba-tion and dissolved CH4 was extracted using the headspacemethod described in Sect 22 As a consequence dissolvedCH4 samples were collected in triplicate for each of the ini-tial and final conditions of one incubation treatment

To interpret CH4 production in the incubation experiments(Sect 23) phylogenic composition of methanogens in thesurface soil was additionally analyzed in 2016 using 16SrRNA gene sequencing In July 2016 soil samples from10 cm in depth were collected in 10 mL plastic tubes in trip-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 759

licate in the same four wet areas as the anaerobic incuba-tion experiments and kept frozen until analysis DNA wasextracted from 3 g of the soil samples as described by Ikeda etal (2004) Extracted DNA was purified using the OneSteptrade

PCR Inhibitor Removal Kit (Zymo Research California) andquantified using the Quant-iT PicoGreen dsDNA assay kit(Invitrogen Carlsbad California) Amplicon sequencing wasconducted targeting the V3ndashV4 regions of 16S rRNA genes(Caporaso et al 2011) Sequences obtained were processedthrough the QIIME pipeline (Caporaso et al 2010) A rep-resentative sequence was picked from each operational tax-onomic unit (OTU) and the Greengenes reference database(version 138) was used to assign taxonomic information andcalculate the relative abundance of methanogenic archaeapresent

For CH4 oxidation surface organic layers (0ndash13 cm indepth) were sampled at sphagnum_K and sedge_K inJuly 2012 and then kept in a refrigerator until the experi-ment (6 days) These soil samples were cut into small piecesand mixed well with air A total of 10 g (about 40 mL) ofsoil sample was transferred into plastic syringes (maximum120 mL) in quadruplicate for each sampling location Ap-proximately 80 mL of air and 02ndash2 mL of 25 CH4 gaswere added to each syringe so the total volume in each sy-ringe was 120 mL and the headspace CH4 concentration was50times 102ndash48times 103 ppm Syringes were preserved in waterand incubated at 8 C for 8 days Headspace gas was sampledon day 0 day 4 and day 8 from each syringe into 20 mL pre-evacuated glass vials with rubber septa Consequently qua-druplicate gas samples were collected for each location andeach sampling day

24 Sample analysis and data processing

Methane concentrations in air samples were analyzed us-ing a gas chromatograph (HP6890 series G1530A HewlettPackard USA) equipped with a flame ionization detector anda CP-carboPLOT capillary column (Varian USA) Methaneflux was calculated from CH4 concentration in chamberheadspace by a linear regression of two to three concentra-tion values against the time elapsed since chamber closureThe detection limit of CH4 flux for each observation was cal-culated as 08ndash24 mg CH4 mminus2 dayminus1 based on whether thechange of chamber CH4 concentration during the observationwas significant relative to the precision of CH4 concentrationanalysis Regression r2 was calculated (formally) as ge 087when the flux value was larger than 2 mg CH4 mminus2 dayminus1Dissolved CH4 concentrations were obtained from calcula-tion of the headspace method for which equilibrations ofCH4 between gas and water phases are described by the Bun-sen absorption coefficient of CH4 (Yamamoto et al 1976)

Carbon and hydrogen isotope ratios of in situ dissolvedCH4 and CH4 samples from both incubation experimentswere analyzed on a GCndashGCndashCndashIRMS (modified after Sug-imoto 1996) ndash which is a continuous flow system con-

sisting of two gas chromatographs a combustion reactorand an isotope ratio mass spectrometer (MAT253 ThermoFisher Scientific USA) ndash and on a GCndashGCndashPndashIRMS (P py-rolysis in a HTC reactor of GC IsoLink Thermo FisherScientific) respectively Carbon and hydrogen isotope ra-tios obtained were represented relative to Vienna PeedeeBelemnite (VPDB) and Vienna Standard Mean Ocean Wa-ter (VSMOW) respectively Precisions of the analyses wereplusmn02 permil and plusmn2 permil for δ13C and δD respectively When cal-culating δ13C and δD of dissolved CH4 the effect of CH4 inbackground air was removed based on the mass balance Inthe aerobic incubation experiments the fractionation factorsof CH4 oxidation for carbon and hydrogen were calculatedusing the following Rayleigh distillation equation

lnRt

R0=

(1αoxminus 1

)ln

[CH4]t[CH4]0

(1)

where R0 and Rt represent isotope ratios under initial condi-tions and at time t respectively αox is the fractionation factorfor CH4 oxidation (defined so that αox gt 1) and [CH4]0 and[CH4]t are CH4 concentrations under initial conditions andat time t respectively

All statistical tests for detecting differences in CH4 fluxesor dissolved CH4 concentrations were conducted usingR software (version 333) MannndashWhitneyrsquos U test was ap-plied to compare magnitudes between 2 years of data andSteelndashDwassrsquos multiple comparison test was used to comparemagnitudes among 3 years or more of data

25 Meteorological data

Air temperature and precipitation observed at a weather sta-tion in Chokurdakh (WMO station 21946) were used to in-vestigate interannual variations in meteorological conditionsduring our observation period of CH4 flux (2009ndash2013) andduring the preceding 2 years (2007ndash2008) The distance be-tween the weather station and our farthest observation site(site V) is approximately 45 km (Fig 1) These data were ob-tained from GHCN-Daily a NOAA database (Menne et al2012a b)

3 Results

31 Environmental factors

Soil wetness conditions and thaw depth differed among veg-etation types (Table 1) Tree mounds had lower surface watercontent (21 ndash17 ) than wet areas (42 ndash48 ) Amongthe two types of wet areas the water level was lower in wetareas of sphagnum mosses than those of sedges (Fig 2)Wet areas of sedges experienced water levels higher thanthe ground surface (defined as 0 cm) reaching more than10 cm above the ground surface Corresponding with soil wa-ter content the thaw depth was shallower at dry tree mounds

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

758 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

the former vegetation type was termed ldquotree moundrdquo andthe latter type was termed ldquowet areardquo Observation pointsin tree mounds were selected at each of the sites V Kand B and termed ldquotree mound_Vrdquo ldquotree mound_Krdquo andldquotree mound_Brdquo (Table 1) For observation points of wet ar-eas a microrelief covered by sphagnum moss at site K wastermed ldquosphagnum_Krdquo and points covered by sedges includ-ing in particular cotton sedges (Eriophorum angustifolium)at sites V K and B were termed ldquosedge_Vrdquo ldquosedge_Krdquo andldquosedge_Brdquo respectively Measurements of volumetric watercontent in the surface soil layer (0ndash20 cm) by TDR (timedomain reflectometry TDR-341F Fujiwara Scientific Com-pany Japan) showed that tree mounds were drier than wetareas this will be described in Sect 31 (Table 1)

22 Field observations and samplings

Methane flux was observed using the chamber method ineach of the typical vegetation types described in Sect 21during the summer from 2009 to 2013 A transparent cylin-drical flux chamber (acrylic resin base area 47times 102 cm2height 25 cm) was installed on the ground The headspacegas of the chamber (ca 12 L) was circulated with a pump(ca 1 L minminus1) The chamber was closed for 15ndash30 min andheadspace gas was sampled two to three times after chamberclosure In most cases the chamber was closed for 30 minand headspace gas was collected at 0 15 and 30 min afterclosure Samples were kept in pre-evacuated glass vials withbutyl rubber septa To minimize soil disturbance we steppedon wooden boards at observation points In 2009 and 2010CH4 flux measurements were conducted in the latter half ofJuly and from 2011 to 2013 observations were conductedcontinuously from early July to the end of July or early Au-gust For all of these years the observation period includedthe warmest season when CH4 emission was expected to bethe most active (Table S2)

For measurements of dissolved CH4 surface water andsoil pore water were sampled in wet areas from 2011 to 2013Surface water was directly taken up by a 50 mL plastic sy-ringe with a three-way cock attached to its tip whereas soilpore water was sampled by a 50 mL syringe (with a three-way cock attached) through a plastic tube inserted in the soilSoon after collecting water samples dissolved CH4 was ex-tracted inside the syringes with the headspace method af-ter adding 15ndash35 mL of the atmosphere prepared in a 10 Laluminum bag This atmosphere was collected beforehandat Chokurdakh village or our observation sites and filteredusing Molecular Sieves 5Aring (116 pellets FUJIFILM WakoPure Chemical Corporation Japan) The atmosphere was an-alyzed later for CH4 concentration and isotopic compositionsas a background sample (20ndash43 ppm for CH4 concentra-tion minus53 permil to minus45 permil for δ13C of CH4 and minus168 permil tominus78 permil for δD of CH4) The syringes were vigorously shakenfor 1 min and left standing for 5 min to ensure equilibration

Finally headspace gas in the syringes was preserved in 10ndash20 mL pre-evacuated glass vials with rubber septa

Concurrently with each flux measurement soil tempera-ture around the flux chamber was measured with a temper-ature sensor in an ORP electrode (PST-2739C DKK-TOACorporation Japan) with an ORP meter (RM-30P or RM-20P) After flux measurement samples were collected thawdepth was observed on the same day around each chamberby inserting a steel rod into the ground From 2011 on wa-ter level was also measured after flux measurements aroundeach chamber in wet areas using a scale The water level wasexpressed as height relative to the ground surface or the mosssurface Observation dates of these environmental factors areshown in Table S2

23 Soil incubation experiments and microbialcommunity analysis

Soil incubation experiments were conducted to estimate δ13Cand δD of produced CH4 and fractionation factors of CH4oxidation for carbon and hydrogen isotopes For CH4 pro-duction experiments surface soil was sampled in all thewet areas in Table 1 (sedge_V sphagnum_K sedge_K andsedge_B) during summer 2013 Samples were taken at 10 cmin depth at each sampling location To observe vertical vari-ations in δ values of produced CH4 within the thaw layer wealso collected samples from two additional depths (20 and30 cm) at sedge_K which is a location typical of the taigandashtundra boundary region These samples were from organiclayers except for the samples from 30 cm which were fromthe top of the mineral layer

Approximately 10 mL of soil was directly transferred intoeach plastic syringe (60 mL maximum capacity) along within situ water (approximately 50 mL) to prevent the samplefrom being oxidized by the atmosphere Syringes were pre-served in water to ensure no leakage and were immediatelypreincubated for 4ndash8 days and then incubated in triplicate for8 days Preincubation and incubation temperatures were setat 5 C We also incubated syringes at 10 C for samples from10 cm in depth at sedge_K to investigate temperature depen-dence of δ values of produced CH4 For each of these sevenincubation treatments (sphagnum_K sedge_V sedge_K andsphagnum_K 10 cm in depth 5 C sedge_K 20 and 30 cmin depth 5 C sedge_K 10 cm in depth 10 C) three repli-cate soil samples were prepared Water in each incubationsyringe was sampled twice at the start and the end of incuba-tion and dissolved CH4 was extracted using the headspacemethod described in Sect 22 As a consequence dissolvedCH4 samples were collected in triplicate for each of the ini-tial and final conditions of one incubation treatment

To interpret CH4 production in the incubation experiments(Sect 23) phylogenic composition of methanogens in thesurface soil was additionally analyzed in 2016 using 16SrRNA gene sequencing In July 2016 soil samples from10 cm in depth were collected in 10 mL plastic tubes in trip-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 759

licate in the same four wet areas as the anaerobic incuba-tion experiments and kept frozen until analysis DNA wasextracted from 3 g of the soil samples as described by Ikeda etal (2004) Extracted DNA was purified using the OneSteptrade

PCR Inhibitor Removal Kit (Zymo Research California) andquantified using the Quant-iT PicoGreen dsDNA assay kit(Invitrogen Carlsbad California) Amplicon sequencing wasconducted targeting the V3ndashV4 regions of 16S rRNA genes(Caporaso et al 2011) Sequences obtained were processedthrough the QIIME pipeline (Caporaso et al 2010) A rep-resentative sequence was picked from each operational tax-onomic unit (OTU) and the Greengenes reference database(version 138) was used to assign taxonomic information andcalculate the relative abundance of methanogenic archaeapresent

For CH4 oxidation surface organic layers (0ndash13 cm indepth) were sampled at sphagnum_K and sedge_K inJuly 2012 and then kept in a refrigerator until the experi-ment (6 days) These soil samples were cut into small piecesand mixed well with air A total of 10 g (about 40 mL) ofsoil sample was transferred into plastic syringes (maximum120 mL) in quadruplicate for each sampling location Ap-proximately 80 mL of air and 02ndash2 mL of 25 CH4 gaswere added to each syringe so the total volume in each sy-ringe was 120 mL and the headspace CH4 concentration was50times 102ndash48times 103 ppm Syringes were preserved in waterand incubated at 8 C for 8 days Headspace gas was sampledon day 0 day 4 and day 8 from each syringe into 20 mL pre-evacuated glass vials with rubber septa Consequently qua-druplicate gas samples were collected for each location andeach sampling day

24 Sample analysis and data processing

Methane concentrations in air samples were analyzed us-ing a gas chromatograph (HP6890 series G1530A HewlettPackard USA) equipped with a flame ionization detector anda CP-carboPLOT capillary column (Varian USA) Methaneflux was calculated from CH4 concentration in chamberheadspace by a linear regression of two to three concentra-tion values against the time elapsed since chamber closureThe detection limit of CH4 flux for each observation was cal-culated as 08ndash24 mg CH4 mminus2 dayminus1 based on whether thechange of chamber CH4 concentration during the observationwas significant relative to the precision of CH4 concentrationanalysis Regression r2 was calculated (formally) as ge 087when the flux value was larger than 2 mg CH4 mminus2 dayminus1Dissolved CH4 concentrations were obtained from calcula-tion of the headspace method for which equilibrations ofCH4 between gas and water phases are described by the Bun-sen absorption coefficient of CH4 (Yamamoto et al 1976)

Carbon and hydrogen isotope ratios of in situ dissolvedCH4 and CH4 samples from both incubation experimentswere analyzed on a GCndashGCndashCndashIRMS (modified after Sug-imoto 1996) ndash which is a continuous flow system con-

sisting of two gas chromatographs a combustion reactorand an isotope ratio mass spectrometer (MAT253 ThermoFisher Scientific USA) ndash and on a GCndashGCndashPndashIRMS (P py-rolysis in a HTC reactor of GC IsoLink Thermo FisherScientific) respectively Carbon and hydrogen isotope ra-tios obtained were represented relative to Vienna PeedeeBelemnite (VPDB) and Vienna Standard Mean Ocean Wa-ter (VSMOW) respectively Precisions of the analyses wereplusmn02 permil and plusmn2 permil for δ13C and δD respectively When cal-culating δ13C and δD of dissolved CH4 the effect of CH4 inbackground air was removed based on the mass balance Inthe aerobic incubation experiments the fractionation factorsof CH4 oxidation for carbon and hydrogen were calculatedusing the following Rayleigh distillation equation

lnRt

R0=

(1αoxminus 1

)ln

[CH4]t[CH4]0

(1)

where R0 and Rt represent isotope ratios under initial condi-tions and at time t respectively αox is the fractionation factorfor CH4 oxidation (defined so that αox gt 1) and [CH4]0 and[CH4]t are CH4 concentrations under initial conditions andat time t respectively

All statistical tests for detecting differences in CH4 fluxesor dissolved CH4 concentrations were conducted usingR software (version 333) MannndashWhitneyrsquos U test was ap-plied to compare magnitudes between 2 years of data andSteelndashDwassrsquos multiple comparison test was used to comparemagnitudes among 3 years or more of data

25 Meteorological data

Air temperature and precipitation observed at a weather sta-tion in Chokurdakh (WMO station 21946) were used to in-vestigate interannual variations in meteorological conditionsduring our observation period of CH4 flux (2009ndash2013) andduring the preceding 2 years (2007ndash2008) The distance be-tween the weather station and our farthest observation site(site V) is approximately 45 km (Fig 1) These data were ob-tained from GHCN-Daily a NOAA database (Menne et al2012a b)

3 Results

31 Environmental factors

Soil wetness conditions and thaw depth differed among veg-etation types (Table 1) Tree mounds had lower surface watercontent (21 ndash17 ) than wet areas (42 ndash48 ) Amongthe two types of wet areas the water level was lower in wetareas of sphagnum mosses than those of sedges (Fig 2)Wet areas of sedges experienced water levels higher thanthe ground surface (defined as 0 cm) reaching more than10 cm above the ground surface Corresponding with soil wa-ter content the thaw depth was shallower at dry tree mounds

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 759

licate in the same four wet areas as the anaerobic incuba-tion experiments and kept frozen until analysis DNA wasextracted from 3 g of the soil samples as described by Ikeda etal (2004) Extracted DNA was purified using the OneSteptrade

PCR Inhibitor Removal Kit (Zymo Research California) andquantified using the Quant-iT PicoGreen dsDNA assay kit(Invitrogen Carlsbad California) Amplicon sequencing wasconducted targeting the V3ndashV4 regions of 16S rRNA genes(Caporaso et al 2011) Sequences obtained were processedthrough the QIIME pipeline (Caporaso et al 2010) A rep-resentative sequence was picked from each operational tax-onomic unit (OTU) and the Greengenes reference database(version 138) was used to assign taxonomic information andcalculate the relative abundance of methanogenic archaeapresent

For CH4 oxidation surface organic layers (0ndash13 cm indepth) were sampled at sphagnum_K and sedge_K inJuly 2012 and then kept in a refrigerator until the experi-ment (6 days) These soil samples were cut into small piecesand mixed well with air A total of 10 g (about 40 mL) ofsoil sample was transferred into plastic syringes (maximum120 mL) in quadruplicate for each sampling location Ap-proximately 80 mL of air and 02ndash2 mL of 25 CH4 gaswere added to each syringe so the total volume in each sy-ringe was 120 mL and the headspace CH4 concentration was50times 102ndash48times 103 ppm Syringes were preserved in waterand incubated at 8 C for 8 days Headspace gas was sampledon day 0 day 4 and day 8 from each syringe into 20 mL pre-evacuated glass vials with rubber septa Consequently qua-druplicate gas samples were collected for each location andeach sampling day

24 Sample analysis and data processing

Methane concentrations in air samples were analyzed us-ing a gas chromatograph (HP6890 series G1530A HewlettPackard USA) equipped with a flame ionization detector anda CP-carboPLOT capillary column (Varian USA) Methaneflux was calculated from CH4 concentration in chamberheadspace by a linear regression of two to three concentra-tion values against the time elapsed since chamber closureThe detection limit of CH4 flux for each observation was cal-culated as 08ndash24 mg CH4 mminus2 dayminus1 based on whether thechange of chamber CH4 concentration during the observationwas significant relative to the precision of CH4 concentrationanalysis Regression r2 was calculated (formally) as ge 087when the flux value was larger than 2 mg CH4 mminus2 dayminus1Dissolved CH4 concentrations were obtained from calcula-tion of the headspace method for which equilibrations ofCH4 between gas and water phases are described by the Bun-sen absorption coefficient of CH4 (Yamamoto et al 1976)

Carbon and hydrogen isotope ratios of in situ dissolvedCH4 and CH4 samples from both incubation experimentswere analyzed on a GCndashGCndashCndashIRMS (modified after Sug-imoto 1996) ndash which is a continuous flow system con-

sisting of two gas chromatographs a combustion reactorand an isotope ratio mass spectrometer (MAT253 ThermoFisher Scientific USA) ndash and on a GCndashGCndashPndashIRMS (P py-rolysis in a HTC reactor of GC IsoLink Thermo FisherScientific) respectively Carbon and hydrogen isotope ra-tios obtained were represented relative to Vienna PeedeeBelemnite (VPDB) and Vienna Standard Mean Ocean Wa-ter (VSMOW) respectively Precisions of the analyses wereplusmn02 permil and plusmn2 permil for δ13C and δD respectively When cal-culating δ13C and δD of dissolved CH4 the effect of CH4 inbackground air was removed based on the mass balance Inthe aerobic incubation experiments the fractionation factorsof CH4 oxidation for carbon and hydrogen were calculatedusing the following Rayleigh distillation equation

lnRt

R0=

(1αoxminus 1

)ln

[CH4]t[CH4]0

(1)

where R0 and Rt represent isotope ratios under initial condi-tions and at time t respectively αox is the fractionation factorfor CH4 oxidation (defined so that αox gt 1) and [CH4]0 and[CH4]t are CH4 concentrations under initial conditions andat time t respectively

All statistical tests for detecting differences in CH4 fluxesor dissolved CH4 concentrations were conducted usingR software (version 333) MannndashWhitneyrsquos U test was ap-plied to compare magnitudes between 2 years of data andSteelndashDwassrsquos multiple comparison test was used to comparemagnitudes among 3 years or more of data

25 Meteorological data

Air temperature and precipitation observed at a weather sta-tion in Chokurdakh (WMO station 21946) were used to in-vestigate interannual variations in meteorological conditionsduring our observation period of CH4 flux (2009ndash2013) andduring the preceding 2 years (2007ndash2008) The distance be-tween the weather station and our farthest observation site(site V) is approximately 45 km (Fig 1) These data were ob-tained from GHCN-Daily a NOAA database (Menne et al2012a b)

3 Results

31 Environmental factors

Soil wetness conditions and thaw depth differed among veg-etation types (Table 1) Tree mounds had lower surface watercontent (21 ndash17 ) than wet areas (42 ndash48 ) Amongthe two types of wet areas the water level was lower in wetareas of sphagnum mosses than those of sedges (Fig 2)Wet areas of sedges experienced water levels higher thanthe ground surface (defined as 0 cm) reaching more than10 cm above the ground surface Corresponding with soil wa-ter content the thaw depth was shallower at dry tree mounds

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

760 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 2 Interannual variations in precipitation (bars) and air tem-perature (solid lines) observed at a weather station at Chokurdakhfor the cold season with snow cover (C total from October in theprevious year to May in the current year) and the warm season(JJAS) and water level (dotted lines) measured in wet areas ofsedges (triangle) and sphagnum mosses (square) Water level wasvery low (ltminus12 cm) in the wet area of sphagnum in 2013 and couldnot be measured Error bars represent standard deviations Methaneflux was observed during the main summers (early July to early Au-gust) from 2009 to 2013

(20ndash23 cm) and deeper in wet areas (31ndash56 cm) In wet ar-eas thaw depth became deeper from 2011 (22plusmn 4 cm) to2012 (25plusmn8 cm) and 2013 (35plusmn7 cm) in observations madeduring mid-July (Table S1 in the Supplement) The overallaverage thaw depth observed on days when flux measure-ments were taken was 31plusmn12 cm (n= 77 9ndash58 cm between3 July and 9 August)

Figure 2 shows persistently low annual precipitation (162ndash173 mm) from 2007 to 2009 In 2010 July air tempera-ture was characteristically high (155 C) accompanying lowmonthly precipitation (8 mm) These show dry conditionsduring our flux observations in 2009 and 2010 Parmentieret al (2011) reported that water level was lower in sum-mer 2009 than the previous two summers at a tundra re-search station (Kytalyk) in the vicinity approximately 30 kmto the northwest of Chokurdakh In contrast precipitationin July 2011 was extremely high (94 mm) with a relativelymild temperature (130 C) which caused an unusual wet-ting High precipitation continued in August (74 mm) andSeptember (67 mm) of the same year Corresponding withthis heavy rainfall water levels were also high in 2011 andsubsequent observations show a clear decrease from 2011to 2013 in wet areas of sedges (p lt 0005) Water levels alsodeclined in wet area of sphagnum mosses with values ofminus9minus10 cm and less than minus12 cm in 2011ndash2013

32 CH4 flux and dissolved CH4 concentration

Obtained CH4 flux shows clear spatial and interannual vari-ations (Fig 3) Individual flux measurements (n= 143 in to-tal) are summarized as mean values for the main summerseasons From 2011 to 2013 continuous flux observations

Figure 3 Interannual variations in averaged CH4 flux in treemounds (denoted as ldquoTrrdquo) and wet areas (ldquoWtrdquo) for the main sum-mer seasons from 2009 to 2013 Replication numbers (ldquonrdquo) areshown for each averaged flux value and standard deviations are rep-resented by error bars Different letters show statistical interannualdifferences in the flux values for wet areas See Table S2 for fluxvalues at respective observation points

(Table S2) were conducted in concert with dissolved CH4analysis and the interannual variation during this period willbe discussed in detail

With regards to the spatial variation in CH4 flux treemounds had consistently small values around the detectionlimit for all measurements (minus49 to 19 mg CH4 mminus2 dayminus1)while wet areas showed CH4 emissions From 2009 to 2013the CH4 flux in wet areas showed large interannual vari-ations ranging from 36 to 140 mg CH4 mminus2 dayminus1 Theflux increased in 2011 when the wetting event occurredand then remained relatively large in 2012 (comparedto 2009 and 2010) Moreover the flux increased againfrom 20112012 to 2013 (p lt 005) No statistically sig-nificant correlation was found when CH4 flux was plottedagainst soil temperature (10 cm in depth) thaw depth or wa-ter level using all the data from wet areas (Fig S1 in theSupplement)

In addition to CH4 flux dissolved CH4 concentration in-creased after the wetting event in 2011 (Fig 4) From 2011to 2012 CH4 concentration in soil pore water at 10 cm indepth (Fig 4b) exhibited a sharp increase of 1 order of mag-nitude (p lt 0005) It remained high from 2012 to 2013 andthe concentrations in surface water and that at 20 cm in depth(Fig 4a and c) also increased significantly over the same pe-riod (p lt 005) No significant difference in concentrationwas observed at 30 cm in depth between 2012 and 2013 Interms of vertical variation the concentration in surface waterwas lower than that in soil pore water (10 20 and 30 cm indepth)

33 δ13C and δD of in situ dissolved CH4

Variability in both δ13C and δD of dissolved CH4 wassmaller in deeper layers showing different patterns betweenδ13C and δD and across years (Fig 5) The δ13C of dis-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 761

Figure 4 Interannual variation in dissolved CH4 concentrationin (a) surface water and soil pore water at (b) 10 cm (c) 20 cmand (d) 30 cm in depth in wet areas from 2011 (wetting event)to 2013 Different letters in each panel denote statistical differencesamong years in averaged concentration across the four wet areas(p lt 005) Error bars represent standard deviations See Table S3for numerical values of dissolved CH4 concentrations

solved CH4 had similarly large ranges (minus68 permil to minus40 permil)in surface water and at 10 cm and 20 cm in depth com-pared to a small range (minus53 permil to minus46 permil) at 30 cm indepth The δD of dissolved CH4 was variable only in sur-face water (minus415 permil to minus308 permil) and at 10 cm in depth(minus417 permil to minus341 permil) whereas it had a constant value ofaroundminus408 permil at 20 and 30 cm in depth Additionally δ13Cvalues approached a relatively high value (approximatelyminus50 permil) at depth while δD values converged to almost theirlowest value In terms of interannual variations in δ13C andδD of dissolved CH4 from 2011 to 2013 both δ13C and δDvalues in surface soil pore water (10 cm in depth) were scat-tered more widely in 2011 showing standard deviations (SD)of 66 permil and 24 permil respectively whereas their ranges weresmaller in 2012 and 2013 (SD 33 permil and 17 permil at maximarespectively)

As shown in Fig 6 convergence of δ13C and δD of dis-solved CH4 is associated with dissolved CH4 concentrationsThe δ13C and δD values of dissolved CH4 including surfacewater and 10 cm in depth converged at high CH4 concentra-tions to the values seen in deeper soil layers (δ13C=minus50plusmn5 permil and δD=minus408plusmn 5 permil at gt 200 micromol CH4 Lminus1)

Figure 5 In situ δ13C versus δD of dissolved CH4 in (a) surfacewater and soil pore water at (b) 10 cm in depth (c) 20 cm in depthand (d) 30 cm in depth from the wetting event in 2011 to 2013Individual delta values are shown here See Table S3 for averageddelta values for each observation point and each year

34 Soil incubation experiments and microbialcommunity analysis

In the anaerobic incubation experiment the CH4 produc-tion rate was different among sampling locations (Fig 7)the rate was higher for sedge_K and sedge_B (066plusmn015 and 043plusmn 009 micromol dayminus1 gminus1 (dry weight) respec-tively) than sedge_V and sphagnum_K (024plusmn 002 and008plusmn 001 micromol dayminus1 gminus1 (dry weight) respectively) Atsedge_K the sampling location tested in detail produc-tion was more rapid for shallower soil layers amongthe 10 20 and 30 cm in depth (066plusmn 015 033plusmn 0060003plusmn 0004 micromol dayminus1 gminus1 (dry weight) respectivelyp lt 001 in Welchrsquos ANOVA test) while no differencein the rate was found between incubation temperatures(066plusmn015 micromol dayminus1 gminus1 (dry weight) at 5 C and 074plusmn014 micromol dayminus1 gminus1 (dry weight) at 10 C p gt 05 in t test)When the CH4 production rate was high the δ13C and δDvalues of produced CH4 were less variable irrespective ofsampling location sampling depth or incubation tempera-ture The δ13C value of produced CH4 at a high produc-tion rate (gt 026 micromol dayminus1 gminus1) (dry weight) was minus55plusmn4 permil (n= 12) Similarly δD under rapid CH4 productionwas minus410plusmn 9 permil (n= 12) These δ13C and δD values ofCH4 obtained under rapid production were mostly compa-rable with the δ values of in situ dissolved CH4 that con-verged in deep soil layers (δ13C=minus50plusmn 2 permil at 30 cm indepth and δD=minus408plusmn 5 permil at 20ndash30 cm in depth Fig 5c

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

762 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Figure 6 In situ (a) δ13C and (b) δD versus concentration of dis-solved CH4 at four depths (surface water 10 20 and 30 cm) in wetareas from 2011 to 2013

and d) although δ13C values in the incubation experimentwere slightly lower than those in situ

In the microbial community analysis using 16S rRNAgene sequencing (Fig 8) soil with high rates of CH4 pro-duction shown in the incubation experiment (sedge_K andsedge_B as in Fig 7) had more abundant total methanogenswithin the detected archaea than that with slow CH4 pro-duction rates (sphagnum_K and sedge_V) Acetoclasticmethanogens in the order Methanosarcinales were higher inproportion among methanogens in sedge_K and sedge_Bwhere δ13C values of produced CH4 in the incubation werehigher In contrast Methanosarcinales were fewer in propor-tion in sphagnum_K where δ13C of the produced CH4 waslower

In the CH4 oxidation experiment CH4 concentration inheadspace declined continuously in every sample (Fig S2)As CH4 oxidation proceeded both δD and δ13C of the re-maining CH4 increased with a linear relationship betweenthem (Figs 9 and S2) Observed slope 1(δD)1(δ13C)was 11 indicating a much larger fractionation of hydro-gen than carbon regardless of vegetation types in wet ar-eas (sphagnum or sedge) The hydrogen isotope fractiona-tion factors of CH4 oxidation calculated from the data shownin Fig 9 were 125 and 116 for wet areas of sphagnum

Figure 7 (a) δ13C and (b) δD of produced CH4 versus CH4 pro-duction rate in the anaerobic soil incubation experiment Produc-tion rates are shown in micromoles of produced CH4 per day andper weight of dry soil in grams Soil samples were collected at fourobservation points (sphagnum_K sedge_V sedge_K and sedge_B)at three depths (10 20 and 30 cm) and incubated at two tempera-tures (5 and 10 C) These samples contain organic layers except forthose collected at 30 cm Error bars represent standard deviations

and sedge respectively while carbon isotope fractionationswere 1021 and 1015 respectively

4 Discussion

41 CH4 flux at tree mounds and wet areas at thetaigandashtundra boundary on the Indigirka Riverlowlands

Methane flux observed in our study was clearly larger in wetareas than at dry tree mounds (Table 1 Fig 3) Such differ-ences in CH4 flux between wetland vegetation and dry areaswith trees or shrubs is generally observed (van Huissteden etal 2005 van der Molen et al 2007 Flessa et al 2008) andis consistent with the fact that CH4 production requires re-ducing conditions in soil (Conrad 2007) Our CH4 flux inwet areas (36ndash140 mg CH4 mminus2 dayminus1) was comparable tothat reported for wet tundras (32ndash101 mg CH4 mminus2 dayminus1) or

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 763

Figure 8 Phylogenetic compositions of methanogenic archaea byorder in wet areas Soil samples (organic layers) were taken in tripli-cate from 10 cm in depth in each wet area in July 2016 See Table S5for detailed results

permafrost fens (42ndash147 mg CH4 mminus2 dayminus1) in a databaseacross permafrost zones compiled by Olefeldt et al (2013)In forests many studies have observed CH4 absorption in-stead of emission (King 1997 Dutaur and Verchot 2007Flessa et al 2008 Morishita et al 2014) However our ob-servations at tree mounds rarely found CH4 absorption oremission In addition CH4 was not consumed even underO2- and CH4-rich conditions in incubation experiments oftree mound soil from site K (Murase et al 2014) indicatingthat a lack of methanotrophic bacterial activity limited CH4absorption at this vegetation type Our results show that CH4emission from wet areas is expected to make a greater contri-bution to ecosystem-scale CH4 exchange at the taigandashtundraboundary on the Indigirka River lowlands

42 Methane flux production and oxidation responsesto the wetting event

In 2009 and 2010 the CH4 emission in wet areas was low(Fig 3) even at relatively high soil temperatures in 2010(Fig S1) under dry conditions that were not directly ob-served in this study The wetting event in 2011 initiated thehigh CH4 emission that continued through 2013 despite de-creasing water levels (Fig 2) Moreover a further increase influx was observed in 2013 accompanying a buildup of dis-solved CH4 (2011ndash2013) as shown in Fig 4

These interannual variations from 2011 to 2013 could becaused by the development of reducing soil conditions overmultiple years after the wetting event Reducing soil con-ditions may have developed to some extent as a result ofthe extreme precipitation in the summer of 2011 (Fig 2)The surface soil layer particularly under high water levels

Figure 9 Enrichment of DH (CH4) and 13C12C (CH4) throughCH4 oxidation during the aerobic incubation experiment of surfaceorganic layers from wet areas of sphagnum mosses and sedges atsite K Individual delta values of the headspace CH4 from each in-cubated syringe and each day are shown Initial isotopic composi-tions of the headspace CH4 were minus66 permil to minus65 permil for δ13C andminus167 permil to minus162 permil for δD

could eliminate O2 from soil pore spaces due to water satu-ration These reducing conditions may have been preservedby freezing of the soil throughout the following winter Ad-ditionally a surface soil layer saturated with ice could haveprevented snowmelt water (rich in O2) from infiltrating thesoil during the spring thaw season of 2012 (Woo 2012)These processes would have led to the continuation of re-ducing conditions in the soil which were created in sum-mer 2011 into 2012 Through further decomposition of soilorganic matter with the consumption of O2 reducing soilconditions may have been exacerbated in the water-saturatedsoil layer to a greater extent in summer 2012 Water in thesaturated soil layer may be retained without exchange for arelatively long time at our study sites because lateral runoff isassumed to be small in the flat floodplain that comprises theIndigirka River lowlands (Nassif and Wilson 1975) In addi-tion deep percolation loss is prevented by the impermeablepermafrost layer (Woo 2012) From summer 2012 to sum-mer 2013 reducing conditions in the soil may have been sim-ilarly prolonged especially in the deep soil layer despite thedecrease in water level from summer 2011 to summer 2013This continuous soil reduction from 2011 to 2013 could havepromoted CH4 production andor decreased CH4 oxidationwhich may explain the increase in dissolved CH4 concentra-tion and CH4 flux in wet areas following the wetting eventand continuing until 2013 (Figs 3 and 4)

In 2011 δ13C and δD of dissolved CH4 (10 cm in depth)were scattered broadly across a wide range whereas in 2012and 2013 the ranges were narrower and they clustered arounda high δ13C value (minus50 permil) and a low δD value (minus408 permilFig 5b) Considering that δD increased much more rapidly

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

764 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

than δ13C in our oxidation experiment (Fig 9) δD can beconsidered a sensitive indicator of CH4 oxidation In con-trast δ13C is not a good indicator because its fractionationfactor of CH4 oxidation (1015ndash1021) was similar to thatof CH4 diffusion (1019 Chanton 2005) thus the effects ofCH4 oxidation and diffusion cannot be discerned by δ13CAdditionally δD of dissolved CH4 (Fig 5) was clearly lowerin deeper layers (20 and 30 cm in depth) than in shallowerlayers (surface water and 10 cm in depth) which indicates δDshowed CH4 oxidation in situ as well because shallower lay-ers are provided with O2 from the atmosphere and precipita-tion The effect of CH4 exchange between surface dissolvedCH4 and atmospheric CH4 can be excluded because all thedissolved CH4 observed in this study was highly oversatu-rated (gt 03 micromol Lminus1 Fig 4) compared to the equilibriumconcentration of atmospheric CH4 (4ndash5 nmol Lminus1 assuming1ndash10 C water temperature and 2 ppm atmospheric CH4 con-centration Yamamoto et al 1976) Thus δD values at 10 cmin 2011 were scattered broadly compared with those in 2012and 2013 that clustered around a low value suggesting thatCH4 oxidation was significant in the surface soil layer dur-ing the year of the wetting event (2011) In 2012 and 2013CH4 oxidation became insignificant relative to the largerpool of dissolved CH4 (Fig 4) In the CH4 production incu-bation experiment δ13C and δD of produced CH4 were lessvariable at higher production rates (δ13C=minus55plusmn 4 permil andδD=minus410plusmn9 permil as in Fig 7) Analogously those of in situdissolved CH4 converged at a high CH4 concentration aroundsimilar values (δ13C=minus50plusmn 5 permil and δD=minus408plusmn 5 permil inFig 6) This suggests that δ values of produced CH4 becamealmost constant under rapid CH4 production in situ and thatthe convergence of δ values of dissolved CH4 observed insitu reflect rapid CH4 production Hence the narrow rangesof δ13C and δD values of dissolved CH4 at 10 cm in depth ob-served in 2012 and 2013 (δ13C aroundminus50 permil and δD aroundminus408 permil Fig 5b) suggest enhanced CH4 production relativeto the wetting year (2011)

Multi-year effects of wetting on CH4 flux through soil re-duction processes have been previously proposed by Kuma-gai and Konno (1998) and Desyatkin et al (2014) as onepossible factor for explaining the increase in CH4 flux af-ter wetting Kumagai and Konno (1998) reported a CH4 fluxincrease at a temperate rice field in Japan 1 year after the ricefield was irrigated and restored from farmland that had beendrained for 8 years Desyatkin et al (2014) observed flux in-creases at a thermokarst depression in boreal eastern Siberiaduring the second consecutive year of flooding followinglarge volumes of precipitation Conversely studies at naturalwetlands in the northeastern USA (Smemo and Yavitt 2006Treat et al 2007 Olson et al 2013) and southern Canada(Moore et al 2011) reported that interannual variations inCH4 flux correspond with those of water level andor pre-cipitation in the current year In our study area multi-yearsoil reduction may be important because soil temperature isgenerally lower than 11 C (10 to 30 cm in depth Fig S1a

and Iwahana et al 2014) due to a shallow active layer un-derlain by permafrost Therefore decomposition of organicmatter can be slow (Treat et al 2015) which would slowlydecrease soil redox potential allowing it to remain relativelyhigh in the first year of wetting

In the following 2 years (2012ndash2013) we observed redoxpotential values lower thanminus100 mV in wet areas (Table S6)which are well below the upper limit for CH4 production insoil (Conrad 2007 Street et al 2016) Methane productionat a potential higher than minus100 mV can also occur becausesoil is heterogeneous and can have more reducing micrositesthan the rest of the bulk soil where redox potential can bemeasured (Teh et al 2005 Teh and Silver 2006)

In addition to the multi-year soil reduction it appears thatthe wetting event led to the thaw depth increase in wet ar-eas from 2011 to 2013 (Table S1) Although thaw depth in-creased summer air temperatures decreased from 2011 (77and 130 C as June and July mean temperatures respec-tively) to 2012 (74 92 C) and 2013 (66 105 C) as shownin Fig 2 The wetting event may have led to the CH4 flux in-crease from 2011 to 2013 (Fig 3) partly through the thawdepth increase by thickening the soil layer where CH4 pro-duction occurs (Nakano et al 2000 van Huissteden et al2005) However the clear increase in dissolved CH4 con-centration (Fig 4) and the enhanced CH4 production andless significant CH4 oxidation reflected in our isotopic data(Fig 5b) cannot be explained by the thaw depth increase Ad-ditionally in the incubation experiment of CH4 production(Fig 7) the CH4 production rate under anaerobic conditionswas slower in the deeper layer especially at 30 cm in depth(mineral soil) compared to 10 cm and 20 cm in depth (or-ganic soil) in sedge_K Treat et al (2015) also reported froma pan-Arctic synthesis of anaerobic incubations that differ-ences in soil types (organicmineral) and in substrate qualityalong depth are important controls on CH4 production rateOur results from the incubations suggests that the deep layercomprised of mineral soil where CH4 production becomesactive when thaw depth increases is not the main region forCH4 production

This study did not evaluate vegetation cover quantitativelyand the wetting event might have also led to some vegeta-tion change (such as increase in sedges) although no drasticchanges were found visually in the observed wet areas In-crease in cover by sedges might have raised CH4 flux partlyby providing labile organic substrate for CH4 production orconduits for the CH4 transport from the soil to the atmo-sphere (Chanton 2005 Lai 2009 Stroumlm et al 2015)

43 Process behind CH4 production response

When CH4 production is initiated after the onset of anoxia inrice paddy soil it first occurs via hydrogenotrophic methano-genesis and then by both hydrogenotrophic and acetoclas-tic methanogenesis which increases CH4 production rate(Conrad 2007) Afterwards the ratio of acetoclastic to hy-

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 765

drogenotrophic methanogenesis can stabilize (Roy et al1997) Considering that this ratio is an important control onisotopic compositions of produced CH4 stabilization of pro-duction pathways might explain the convergence in δ valuesof dissolved CH4 at our study sites under high CH4 concen-tration (Fig 6) and the reduced variability in δ values of pro-duced CH4 in our experiment under rapid production condi-tions (Fig 7) As acetoclastic methanogenesis leads to higherδ13C in produced CH4 than hydrogenotrophic methanogen-esis (Sugimoto and Wada 1993) acetoclastic methanogen-esis may have been activated when dissolved CH4 concen-tration or CH4 production rate were high during our studyThis interpretation is supported by the microbial commu-nity analysis (Fig 8) in which acetoclastic methanogens ofMethanosarcinales were more abundant in wet areas witha higher δ13C of produced CH4 in the incubation Thereforethe high δ13C values with smaller variation observed at 10 cmin depth in 2012 and 2013 (Fig 5b) suggest a greater con-tribution from acetoclastic methanogenesis compared to thewetting year (2011) Similar to findings from rice paddy soil(Conrad 2007) acetoclastic methanogenesis may have ex-perienced delayed activation after anoxic conditions beganin 2011 which could also have promoted CH4 productionin 2012 and 2013

5 Concluding remarks

At the taigandashtundra boundary on the Indigirka River low-lands we observed an increase in CH4 flux in wet areas fol-lowing the wetting event in 2011 and a further increase influx in 2013 Our results show interannual variations in δ13Cand δD of dissolved CH4 and when compared with our in-cubation experiments these variations suggest both enhance-ment of CH4 production and less significance of CH4 oxida-tion in 2012 and 2013 compared to 2011 This enhancementof production could be partly caused by activation of aceto-clastic methanogenesis following the development of reduc-ing soil conditions after the wetting event Analyses of iso-topic compositions of CH4 both in situ and in incubation ex-periments can be combined to investigate the effects of CH4production and oxidation on these isotopic compositions andto clarify the relationship between CH4 flux and wetting Inthe future measuring the δ13C of dissolved CO2 would beuseful to further validate activation of acetoclastic methano-genesis (Sugimoto and Wada 1993 McCalley et al 2014Itoh et al 2015) Outside of these processes the wettingevent might have affected CH4 flux partly via the thaw depthincrease or some amount of vegetation change It would beuseful to analyze δ13C and δD values of emitted CH4 in or-der to assess changes in CH4 transport (such as by increase insedge cover) and to investigate the relationship between dis-solved CH4 concentration and CH4 flux in detail (Chanton2005)

In recent years strong storm activity and wetting eventsin terrestrial ecosystems have been observed in northern re-gions (Iijima et al 2016) A wetting event at the taigandashtundraboundary can switch microreliefs with large interannual vari-ations in soil wetness conditions to significant CH4 sourceswe observed clear increases in CH4 flux in wet areas after thewetting event In order to predict CH4 flux following a wet-ting event in a permafrost ecosystem our results show themulti-year process of soil reduction affected by the durationof water saturation in the active layer

Data availability Our data of CH4 flux concentration and iso-tope ratios of dissolved CH4 soil incubation experiments phy-logenic compositions of methanogenic archaea soil temperaturethaw depth and water level have been deposited in the ArcticData archive System (ADS) Japan (httpsadsnipracjpdatasetA20190211-001 Shingubara et al 2019) Meteorological datashown in Fig 2 are available from GHCN-Daily a NOAA database(httpsdoiorg107289V5D21VHZ Menne et al 2012b)

Supplement The supplement related to this article is availableonline at httpsdoiorg105194bg-16-755-2019-supplement

Author contributions RS and AS designed the experiments andRS carried them out GI ST LM ST TM and TCM helped withsampling in situ data collection and preparing resources for thefieldwork JM contributed to the laboratory analysis RS preparedthe paper with contributions from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements This research was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) numbers 21403011and 16H05618 a Grant-in-Aid from the Global COE ProgramldquoEstablishment of Center for Integrated Field Environmental Sci-encerdquo (IFES-GCOE) funded by the Ministry of Education CultureSports Science and Technology ndash Japan (MEXT) JST (StrategicInternational Collaborative Research Program SICORP) EUcooperative research project ldquoDynamics of permafrost and methaneemission in Arctic terrestrial ecosystem in Eastern Siberiardquo theGreen Network of Excellence (GRENE) program funded byMEXT and the COPERA (C budget of Ecosystems Cities andVillages on Permafrost in Eastern Russian Arctic) project fundedby the Belmont Forum through JST We sincerely thank Alexan-der Kononov Roman Petrov Egor Starostin Alexandra Alexeevaand other members of the Institute for Biological Problems ofCryolithozone SB RAS and Tatiana Stryukova Sergey Ianyginand other staff at the Allikhovsky Ulus Inspectorate of NatureProtection for supporting our fieldwork in the vicinity of Chokur-dakh We also wish to acknowledge help from Yumi HoshinoSatori Nunohashi Kanako Tanaka Kayoko Saito Hanae Kudo andRuslan Shakhmatov in our research group at Hokkaido University

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

766 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Edited by Nobuhito OhteReviewed by two anonymous referees

References

Caporaso J G Kuczynski J Stombaugh J Bittinger K Bush-man F D Costello E K Fierer N Pentildea A G GoodrichJ K Gordon J I Huttley G A Kelley S T KnightsD Koenig J E Ley R E Lozupone C A McDonaldD Muegge B D Pirrung M Reeder J Sevinsky J RTurnbaugh P J Walters W A Widmann J Yatsunenko TZaneveld J and Knight R QIIME allows analysis of high-throughput community sequencing data Nat Meth 7 335ndash336httpsdoiorg101038nmethf303 2010

Caporaso J G Lauber C L Walters W A Berg-Lyons DLozupone C A Turnbaugh P J Fierer N and Knight RGlobal patterns of 16S rRNA diversity at a depth of millions ofsequences per sample P Natl Acad Sci USA 108 4516ndash4522httpsdoiorg101073pnas1000080107 2011

Chanton J P The effect of gas transport on the isotope sig-nature of methane in wetlands Org Geochem 36 753ndash768httpsdoiorg101016jorggeochem200410007 2005

Ciais P Sabine C Bala G Bopp L Brovkin V Canadell JChhabra A DeFries R Galloway J Heimann M Jones CLe Queacutereacute C Myneni R B Piao S and Thornton P Carbonand other biogeochemical cycles in Climate change 2013 Thephysical science basis Contribution of Working Group I to theFifth Assessment Report of the Intergovernmental Panel on Cli-mate Change Cambridge University Press Cambridge UK andNew York NY USA 465ndash570 2014

Conrad R Microbial ecology of methanogens and methanotrophsin Advances in Agronomy vol 96 Academic Press BurlingtonSan Diego USA 1ndash63 2007

Desyatkin A R Takakai F and Hatano R Flood ef-fect on CH4 emission from the alas in Central Yaku-tia East Siberia Soil Sci Plant Nutr 60 242ndash253httpsdoiorg101080003807682014883486 2014

Dlugokencky E J Bruhwiler L White J W C EmmonsL K Novelli P C Montzka S A Masarie K ALang P M Crotwell A M Miller J B and Gatti LV Observational constraints on recent increases in the at-mospheric CH4 burden Geophys Res Lett 36 L18803httpsdoiorg1010292009GL039780 2009

Dutaur L and Verchot L V A global inventory of thesoil CH4 sink Global Biogeochem Cy 21 GB4013httpsdoiorg1010292006GB002734 2007

Flessa H Rodionov A Guggenberger G Fuchs H Magdon PShibistova O Zrazhevskaya G Mikheyeva N Kasansky OA and Blodau C Landscape controls of CH4 fluxes in a catch-ment of the forest tundra ecotone in northern Siberia GlobalChange Biol 14 2040ndash2056 httpsdoiorg101111j1365-2486200801633x 2008

Iijima Y Nakamura T Park H Tachibana Y and FedorovA N Enhancement of Arctic storm activity in relation to per-mafrost degradation in eastern Siberia Int J Climatol 364265ndash4275 httpsdoiorg101002joc4629 2016

Ikeda S Watanabe K N Minamisawa K and Ytow N Eval-uation of Soil DNA from Arable Land in Japan Using a Modi-fied Direct-extraction Method Microbes Environ 19 301ndash309httpsdoiorg101264jsme219301 2004

Itoh M Shimamura T Ohte N and Takemon Y Dif-ferences in hydrophyte life forms induce spatial hetero-geneity of CH4 production and its carbon isotopic signa-ture in a temperate bog peatland δ13CH4 variations in hy-drophyte species J Geophys Res-Biogeo 120 1177ndash1195httpsdoiorg1010022014JG002881 2015

Iwahana G Takano S Petrov R E Tei S Shingubara RMaximov T C Fedorov A N Desyatkin A R Nikolaev AN Desyatkin R V and Sugimoto A Geocryological char-acteristics of the upper permafrost in a tundra-forest transitionof the Indigirka River Valley Russia Polar Science 8 96ndash113httpsdoiorg101016jpolar201401005 2014

King G Responses of atmospheric methane consumption by soilsto global climate change Global Change Biol 3 351ndash362httpsdoiorg101046j1365-2486199700090x 1997

Kumagai K and Konno Y Methane emissions from rice paddyfields after upland farming Japan J Soil Sci Plant Nutr 69333ndash339 httpsdoiorg1020710dojo694_333 1998

Lai D Y F Methane dynamics in northern peatlands A re-view Pedosphere 19 409ndash421 httpsdoiorg101016S1002-0160(09)00003-4 2009

Liang M Sugimoto A Tei S Bragin I V Takano SMorozumi T Shingubara R Maximov T C KiyashkoS I Velivetskaya T A and Ignatiev A V Importanceof soil moisture and N availability to larch growth anddistribution in the Arctic taiga-tundra boundary ecosys-tem northeastern Siberia Polar Science 8 327ndash341httpsdoiorg101016jpolar201407008 2014

Marik T Fischer H Conen F and Smith K Seasonalvariations in stable carbon and hydrogen isotope ratios inmethane from rice fields Seasonal variations in CH4 iso-tope ratios from rice Global Biogeochem Cy 16 1094httpsdoiorg1010292001GB001428 2002

McCalley C K Woodcroft B J Hodgkins S B Wehr R AKim E-H Mondav R Crill P M Chanton J P Rich V ITyson G W and Saleska S R Methane dynamics regulated bymicrobial community response to permafrost thaw Nature 514478ndash481 httpsdoiorg101038nature13798 2014

Menne M J Durre I Vose R S Gleason B E and Hous-ton T G An overview of the Global Historical ClimatologyNetwork-Daily database J Atmos Ocean Tech 29 897ndash910httpsdoiorg101175JTECH-D-11-001031 2012a

Menne M J Durre I Korzeniewski B McNeal S ThomasK Yin X Anthony S Ray R Vose R S Glea-son B E and Houston T G Global Historical Cli-matology Network ndash Daily (GHCN-Daily) Version 322NOAA National Data Climatic Data Center Asheville USAhttpsdoiorg107289V5D21VHZ 2012b

Moore T R De Young A Bubier J L Humphreys E RLafleur P M and Roulet N T A multi-year record of methaneflux at the Mer Bleue Bog southern Canada Ecosystems 14646ndash657 httpsdoiorg101007s10021-011-9435-9 2011

Morishita T Matsuura Y Kajimoto T Osawa A ZyryanovaO A and Prokushkin A S CH4 and N2O dynamics of aLarix gmelinii forest in a continuous permafrost region of central

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019

R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia 767

Siberia during the growing season Polar Science 8 156ndash165httpsdoiorg101016jpolar201401004 2014

Murase J Sugimoto A Shingubara R and Maximov T CMethane oxidation potential of Arctic wetland soil of a taiga-tundra ecotone in Northeastern Siberia in Japan GeoscienceUnion Meeting 28 Aprilndash2 May 2014 Yokohama JapanACG36-P16 2014

Myhre G Shindell D Breacuteon F-M Collins W FuglestvedtJ Huang J Koch D Lamarque J-F Lee D MendozaB Nakajima T Robock A Stephens G Takemura T andZhang H Anthropogenic and natural radiative forcing in Cli-mate change 2013 The physical science basis Contribution ofWorking Group I to the Fifth Assessment Report of the Inter-governmental Panel on Climate Change Cambridge UniversityPress Cambridge UK and New York NY USA 658ndash740 2013

Nakano T Kuniyoshi S and Fukuda M Temporal variationin methane emission from tundra wetlands in a permafrostarea northeastern Siberia Atmos Environ 34 1205ndash1213httpsdoiorg101016S1352-2310(99)00373-8 2000

Nassif S H and Wilson E M The Influence of slope and rainintensity on runoff and infiltration Hydrolog Sci Bull 20 539ndash553 httpsdoiorg10108002626667509491586 1975

Nisbet E G Dlugokencky E J and Bousquet PMethane on the Rise ndash Again Science 343 493ndash495httpsdoiorg101126science1247828 2014

Olefeldt D Turetsky M R Crill P M and McGuire AD Environmental and physical controls on northern terres-trial methane emissions across permafrost zones Global ChangeBiol 19 589ndash603 httpsdoiorg101111gcb12071 2013

Olson D M Griffis T J Noormets A Kolka R and Chen JInterannual seasonal and retrospective analysis of the methaneand carbon dioxide budgets of a temperate peatland J GeophysRes-Biogeo 118 226ndash238 httpsdoiorg101002jgrg200312013

Parmentier F J W van Huissteden J van der Molen M KSchaepman-Strub G Karsanaev S A Maximov T C andDolman A J Spatial and temporal dynamics in eddy co-variance observations of methane fluxes at a tundra site innortheastern Siberia J Geophys Res-Biogeo 116 G01099httpsdoiorg1010292010JG001637 2011

Preuss I Knoblauch C Gebert J and Pfeiffer E-M Improvedquantification of microbial CH4 oxidation efficiency in arcticwetland soils using carbon isotope fractionation Biogeosciences10 2539ndash2552 httpsdoiorg105194bg-10-2539-2013 2013

Roy R Kluumlber H D and Conrad R Early initiation ofmethane production in anoxic rice soil despite the pres-ence of oxidants FEMS Microbiol Ecol 24 311ndash320httpsdoiorg101111j1574-69411997tb00448x 1997

Shingubara R Sugimoto A Murase J Iwahana G TeiS Liang M Takano S Morozumi T and Maximov TC Chamber CH4 flux observed at taigandashtundra boundarynear Chokurdakh in northeastern Siberia during the summerfrom 2009 to 2013 as well as isotopic compositions of dis-solved CH4 and soil incubation experiments 100 ArcticData archive System (ADS) Japan httpsadsnipracjpdatasetA20190211-001 last access 11 February 2019

Smemo K A and Yavitt J B A multi-year perspective onmethane cycling in a shallow peat fen in central New York

State USA Wetlands 26 20ndash29 httpsdoiorg1016720277-5212(2006)26[20AMPOMC]20CO2 2006

Street L E Dean J F Billett M F Baxter R Dinsmore K JLessels J S Subke J-A Tetzlaff D and Wookey P A Re-dox dynamics in the active layer of an Arctic headwater catch-ment examining the potential for transfer of dissolved methanefrom soils to stream water J Geophys Res-Biogeo 121 2776ndash2792 httpsdoiorg1010022016JG003387 2016

Stroumlm L Falk J M Skov K Jackowicz-Korczynski M Mas-tepanov M Christensen T R Lund M and Schmidt N MControls of spatial and temporal variability in CH4 flux in ahigh arctic fen over three years Biogeochemistry 125 21ndash35httpsdoiorg101007s10533-015-0109-0 2015

Sugimoto A GCGCCIRMS system for carbo isotope measure-ment of low level methane concentration Geochem J 30 195ndash200 httpsdoiorg102343geochemj30195 1996

Sugimoto A and Wada E Carbon isotopic composition of bacte-rial methane in a soil incubation experiment Contributions of ac-etate and CO2H2 Geochim Cosmochim Ac 57 4015ndash4027httpsdoiorg1010160016-7037(93)90350-6 1993

Sugimoto A and Wada E Hydrogen isotopic composi-tion of bacterial methane CO2H2 reduction and acetatefermentation Geochim Cosmochim Ac 59 1329ndash1337httpsdoiorg1010160016-7037(95)00047-4 1995

Teh Y A and Silver W L Effects of soil structure de-struction on methane production and carbon partition-ing between methanogenic pathways in tropical rainforest soils J Geophys Res-Biogeo 111 G01003httpsdoiorg1010292005JG000020 2006

Teh Y A Silver W L and Conrad M E Oxygen ef-fects on methane production and oxidation in humid trop-ical forest soils Global Change Biol 11 1283ndash1297httpsdoiorg101111j1365-2486200500983x 2005

Treat C C Bubier J L Varner R K and Crill P M Timescaledependence of environmental and plant-mediated controls onCH4 flux in a temperate fen J Geophys Res 112 G01014httpsdoiorg1010292006JG000210 2007

Treat C C Natali S M Ernakovich J Iversen C M Lu-pascu M McGuire A D Norby R J Chowdhury TR Richter A Šantruckovaacute H Schaumldel C Schuur EA G Sloan V L Turetsky M R and Waldrop M PA pan-Arctic synthesis of CH4 and CO2 production fromanoxic soil incubations Global Change Biol 21 2787ndash2803httpsdoiorg101111gcb12875 2015

van der Molen M K van Huissteden J Parmentier F J WPetrescu A M R Dolman A J Maximov T C KononovA V Karsanaev S V and Suzdalov D A The growing sea-son greenhouse gas balance of a continental tundra site in theIndigirka lowlands NE Siberia Biogeosciences 4 985ndash1003httpsdoiorg105194bg-4-985-2007 2007

van Huissteden J Maximov T C and Dolman A JHigh methane flux from an arctic floodplain (Indigirkalowlands eastern Siberia) Methane Flux Arctic Flood-plain Siberia J Geophys Res-Biogeo 110 G02002httpsdoiorg1010292005JG000010 2005

Woo M-K Permafrost hydrology Springer Berlin HeidelbergGermany httpsdoiorg101007978-3-642-23462-0 2012

wwwbiogeosciencesnet167552019 Biogeosciences 16 755ndash768 2019

768 R Shingubara et al Multi-year effect of wetting on CH4 flux at taigandashtundra boundary in NE Siberia

Yabuki H Park H Suzuki R and Ohata T Baseline Meteoro-logical Data in Siberia (BMDS) Version 50 500 Arctic Dataarchive System (ADS) Japan available at httpsadsnipracjpdatasetA20131107-002 (last access 31 October 2017) 2011

Yamamoto S Alcauskas J B and Crozler T E Solubility ofmethane in distilled water and seawater J Chem Eng Data 2178ndash80 httpsdoiorg101021je60068a029 1976

Biogeosciences 16 755ndash768 2019 wwwbiogeosciencesnet167552019