Scholarly article on topic ' Assessment of winter fluxes of CO 2 and CH 4 in boreal forest soils of central Alaska estimated by the profile method and the chamber method: a diagnosis of methane emission and implications for the regional carbon budget '

Assessment of winter fluxes of CO 2 and CH 4 in boreal forest soils of central Alaska estimated by the profile method and the chamber method: a diagnosis of methane emission and implications for the regional carbon budget Academic research paper on "Earth and related environmental sciences"

0
0
Share paper
Academic journal
Tellus B
Keywords
{""}

Academic research paper on topic " Assessment of winter fluxes of CO 2 and CH 4 in boreal forest soils of central Alaska estimated by the profile method and the chamber method: a diagnosis of methane emission and implications for the regional carbon budget "

Tellus (2007), 59B, 223-233

Printed in Singapore. All rights reserved

© 2007 The Authors

Journal compilation © 2007 Blackwell Munksgaard

TELLUS

Assessment of winter fluxes of CO2 and CH4 in boreal forest soils of central Alaska estimated by the profile

method and the chamber method: a diagnosis of methane emission and implications for the regional

carbon budget

By YONGWON KIM1*, MASAHITO UEYAMA1, FUMIKO NAKAGAWA2, URUMU TSUNOGAI2, YOSHINOBU HARAZONO1 and NORIYUKI TANAKA1, 1 International Arctic Research Center, University of Alaska Fairbanks, Fairbanks, AK 99775-7335, USA; 2Division of Earth and Planetary Science, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Hokkaido, Japan

(Manuscript received 9 January 2006; in final form 18 August 2006) ABSTRACT

This research was carried out to estimate the winter fluxes of CO2 and CH4 using the concentration profile method and the chamber method in black spruce forest soils in central Alaska during the winter of 2004/5. The average winter fluxes of CO2 and CH4 by chamber and profile methods were 0.24 ± 0.06 (SE; standard error) and 0.21 ± 0.06 gCO2-C/m2/d, and 21.4 ± 5.6 and 21.4 ± 14 /u,gCH4-C/m2/hr. This suggests that the fluxes estimated by the two methods are not significantly different based on a one-way ANOVA with a 95% confidence level. The hypothesis on the processes of CH4 transport/production/emission in underlying snow-covered boreal forest soils is proven by the pressure differences between air and in soil at 30 cm depth. The winter CO2 emission corresponds to 23% of the annual CO2 emitted from Alaska black spruce forest soils, which resulted in the sum of mainly root respiration and microbial respiration during the winter based on the 513CO2 of —22.5%c. The average wintertime emissions of CO2 and CH4 were 49 ± 13 gCO2-C/m2/season and 0.11 ± 0.07 gCH4-C/m2/season, respectively. This implies that winter emissions of CO2 and CH4 are an important part of the annual carbon budget in seasonally snow-covered terrain of typical boreal forest soils.

1. Introduction

Winter CO2 flux is a large carbon source in snow-covered ecosystems, such as arctic tundra (Zimov et al., 1993; Oechel et al., 1997; Fahnesock et al., 1999), alpine and subalpine regions (Solomon and Cerling, 1987; Sommerfeld et al., 1993; Brooks et al., 1996; Mast et al., 1998; Wickland et al., 1999, 2001), boreal forests (Winston et al., 1997) and temperate forests (Hirano, 2005; Takagi et al., 2005). Winter flux measurements have rarely been observed due to the difficulty of making measurements in snow-covered ecosystems; furthermore, winter CO2 emissions through the snowpack were assumed to be negligible or were underestimated (Raich and Schlesinger, 1992; Raich and Potter, 1995). The magnitude of the flux through the snowpack to the

* Corresponding author. e-mail: kimyw@iarc.uaf.edu DOI: 10.1111/j.1600-0889.2006.00233.x

atmosphere is smaller than that estimated in the growing season. However, the contribution of winter flux corresponded to 1030% of the annual soil respiration in alpine and tundra regions, where the snow-covered period is more than 200 days (Oechel et al., 1997; Mast et al., 1998; Wickland et al., 1999, 2001). These findings suggest that the winter CO2 emission to the atmosphere might contribute a significant portion of the annual carbon budget.

Winter CH4 flux also acts as a large carbon source in temperate peatlands (Dise, 1992; Shannon and White, 1994), subalpine wetlands (Mast et al., 1998; Wickland et al., 1999, 2001), boreal peat bogs and peatlands (Alm et al., 1999; Panikov, 1999; Panikov and Dedysh, 2000) and boreal forests (Whalen and Reeburgh; 1988; 1992). These ecosystems are well known as CH4 sources under anaerobic environments. Conversely, the soils under aerobic environments have also become famous as CH4 sinks (e.g. CH4 oxidation) in various ecosystems (Born et al., 1990; Dorr and Miinnich, 1990; Deflas et al.,

1999). However, winter CH4 emission and production in the snow-covered boreal forest soils are still poorly understood. The winter emissions of CH4 as well as CO2 in boreal forest and tundra soils might be significant carbon sources on regional and global scales.

Within the anaerobic environments of Alaska bogs and tundra wetlands, Duddleston et al. (2002) and Strom et al. (2003) demonstrated that acetate was the primary organic terminal product of anaerobic decomposition, and acetate was ultimately produced by this reduction: (CH3COOH ^ CH4 + CO2; CO2 + 4H2 ^ CH4 + 2H2O). The produced CO2 and CH4 gases were released to the atmosphere by three main pathways: molecular diffusion, bubble ebullition and through emergent vascular plants during the growing season. During the winter, when boreal forest soils were frozen with time, the frozen soils sometimes changed to suboxic/anoxic environments. CO2 and CH4 produced in the frozen soil may be emitted to the atmosphere only by vascular plants; in this case tussock tundra plants at the tops of tussocks. Here, I provide a novel report of net methane emission through the vascular plants over tussock soils to the snowpack in a boreal forest of central Alaska. This study has been conducted in parallel with tower observations that have monitored the fluxes of CO2 and CH4 between the boreal forest and the atmosphere since November 2002 on the wooded campus of the University of Alaska Fairbanks.

The purposes of this study are to (1) compare the profile method and the chamber method for the winter fluxes of CO2 and CH4 in snow-covered boreal forest soils, such as the tussock and sphagnum moss regime of typical boreal forest soils, central Alaska, (2) clarify the emission of CO2 and CH4 in soil, snow-pack, and tower air samples by using stable carbon isotopes of CO2 and CH4, and (3) evaluate the contribution of winter carbon to the annual regional carbon budget in a typical boreal forest of interior Alaska.

2. Materials and methods

2.1. Sampling locations and methods

The study site was located at a tower station (64°52'N, 147°51'W) on the campus of the University of Alaska Fairbanks in central Alaska. The site represents typical boreal forest and the dominant vegetation there is black spruce, Picea mariana, which are 6 m high. The forest floor is a mixture of tussocks, vascular plants, shrubs, sphagnum and feather mosses, and lichen. Specifically, those plants include Betula glandulosa, Ledum palustre, Vaccinium vitis-idaea, Carex lugens, Sphagnum spp., Thuidium abietinum and Cladina stellaris on the permafrost regime. The densities of black spruce, sphagnum and feather mosses are 4500 trees/ha, 4.4 individual/cm2 and 0.8 individual/cm2, respectively. The vegetation characteristics of the site are shown in Table 1.

a) Tussock

60 cm 50 40 30 20 10 0

Permafrost

Fig. 1. Sampling scheme in (a) tussock and (b) sphagnum moss regimes of the boreal forest, central Alaska. Sampling depths in snowpack and soil were described in text.

The annual average air temperature was -3.3 °C, and the average temperature in July was 16.4 °C during the last 40 years (1964-2004). Annual average precipitation at the site is 270 mm and the average precipitation in August is 47 mm during the last 40 yr.

The average height of each tussock is 20 cm. The soil surrounding tussocks is so compacted clayey that vascular plants only inhabit the top of tussocks (Fig. 1a). Hence, we could not collect soil air samples in the tussock regime. Sphagnum moss photosynthesizes and has a large water-holding capacity in the surface soil. Individual living moss strands are 7-8 cm long, and the dead moss in the peat phase exists to depths of 50 cm below the living moss. The soils are characterized by an organic matter mat (50 cm in depth), a loess parent material (<1cm depth) and permafrost. The mineral soil consists of loess that was blown from the Alaska Range and fluvial plains during the Holocene. The mineral loess soil extends at least to the top of the permafrost. Permafrost exists beneath the covered mineral loess soil (Fig. 1b). Tussock and sphagnum moss in Alaska's boreal forest completely freezes ordinarily in late January (V. Romanovsky, personal communication, 2003). The contents of biomass, carbon and nitrogen in aboveground and below-ground vegetations in the black spruce forest of central Alaska are listed in Table 1.

The flux chamber was deployed within a few metres of the in-soil probes at sphagnum moss and the chamber diameter was 50 cm with a cross section area of 0.2 m2 and height of 10 cm

Table 1. The contents of biomass, carbon, and nitrogen in abovegound and belowground of mass regime in the black spruce forest, central Alaska

Component Blomass(kg/m2) Carbon content (kgC/m2) Nitrogen content (kgN/m2)

Mean SD Mean SD Mean SD

Black spruce (n = 3) 0.10 0.061 0.055 0.033 0.002 0.002

Moss (n = 5) 0.47 0.104 0.216 0.05 0.002 0.001

A0 layer (n = 4) 5.57 0.717 2.564 0.33 0.056 0.007

(plastic material). The chamber top has two fixed nozzles (1 mm ID and 3 mm OD; 300 cm long); one is used to collect air samples and the other is used to balance the pressure between the inside and outside chamber. The chamber was placed over the snowpack surface at the beginning of each measurement series as not to disrupt the surface snowpack at an interval of 2 m during a steady state condition. Four air samples were drawn from the headspace in the chamber at approximately 10 min intervals after the deployment. The 10 mL samples are collected using a 50 mL plastic syringe at a rate of 5 mL per minute. The flux was calculated from the gradient of the relationship between the concentration variations of a gas and sampling time. The correlation coefficient was above 0.99 and the curve was linear (Kim and Tanaka, 2003).

Stainless steel tubes (1 mm ID, 3 mm OD; 50 cm long; Swagelok, USA) used for concentration profiles of CO2 and CH4 were horizontally set on the ground between the tussocks (Fig. 1a) before the snow fell. The tubes were installed in sphagnum moss in March 2005. The stainless steel tube was connected with polyethylene tubing (1 mm ID and 3 mm OD; 150 cm long; Nalgene, USA), and the end of the tubing was attached to a three-way stopcock. The sampling depths were 0, 10, 20, 30, 40, 50 and 60 cm above the soil surface.

The tubes under the soil were vertically built at the sphagnum moss regime. The stainless steel tube, the polyethylene tubing and the three-way stopcock were connected using the same method executed at the tussock area (Figure 1b). The sampling depths were 5, 10, 20,30 and 40 cm below the surface. The sampling period was an interval of 10 d and consisted in part of snow pit-wall observations, where we examined the snow crystals, snow density and snow temperature. However, the sampling strongly depended on local weather conditions. The monitoring of soil temperature and moisture has been carried out since June

2004 for the sphagnum moss regime. The monitoring of pressure at 2 m above the soil surface and at soil 30 cm depth using loggers (HOBO Pressure, Onset Instrument Corp., USA) and of temperature (HOBO 4-Channel External, Onset Instrument Cop.) at 5, 10, 30, 40 and 50 cm depth below the soil surface and at 85 cm above the surface have been conducted since October

2005 at the sphagnum moss site.

A tower site was built at the typical black spruce forest on the discontinuous permafrost in November 2002; the tower is 10 m high and the sampling inlets are at the 2 and 8 m levels.

The height of the canopy is 6 m. The ambient air samples were collected to apre-vacuum vial (10 mL, GL Science, Japan) using the pump from two levels.

Ten millilitres of air sample from soil and snowpack were collected in the vials with a plastic syringe at a rate of 5 mL per minute at the tussock and sphagnum moss regimes. The vials were transferred to the laboratory for the analyses of CO2 and CH4 concentration. The air samples of soil, snowpack and tower were separately collected for the stable carbon isotope of CO2 and CH4 every 2 to 3 weeks.

2.2. Analyses of the concentrations and stable carbon isotopes ofCO2 and CH4

These gas-filled vials were transported to the laboratory for later CO2 and CH4 analyses. Concentrations of CO2 were measured with a gas chromatograph equipped with a thermal conductivity detector (TCD-GC, GC-14B, Shimadzu Ltd., Japan) with a column packed with Porapack Q (80/100 mesh). Calibration was done with a series of standard gases containing 338 ± 7, 491 ± 10 and 5000 ± 100 ppmv of CO2 in synthetic air (Scotts Gas Co., USA). Precision with the calibration range was usually better than 1%.

Concentrations of CH4 were measured with a gas chromatograph equipped with a flame ionization detector (FID-GC, GC-14B, Shimadzu Ltd., Japan) using a column packed with Molecular Sieve 13X (80/100 mesh). Calibration was done with a series of standard gases containing 0.79 ± 0.016,1.61 ± 0.020 and 2.43 ± 0.048 ppmv of CH4 in synthetic air (Scotts Gas Co.). Precision with the calibration range was usually better than 2%.

In order to analyze the contents of carbon and nitrogen from above- and below-ground vegetations, after each bulk sample was dried at 85 °C and the sample was homogeneously ground into powder using agate mortar and pestle, the contents of biomass, carbon and nitrogen in each sample were determined using a CN analyzer (SUMIGRAPH NC-1000, Japan). The analytical error was usually less than 0.1%.

The stable carbon isotope compositions of CO2 and CH4 were measured using a GC/C/IRMS [Gas Chromatography (Varian 3400)/Combustion/Isotope Ratio Mass Spectrometry (Finnigan MAT delta S)] system (Nakagawa et al., 2002; Ishimura et al., 2004). In this system, air samples for determination of 13CofCO2 were injected directly without any pre-treatment. Components

in the injected air samples were separated through a capillary column (Pora Plot Q, Chrompack). Column-separated CH4 was oxidized in a CuO furnace (800 °C) to CO2 and H2O. After removal of H2O from the mixture by a -100 °C cryogenic trap, the purified CO2 was introduced into a mass spectrometer for S13C measurements. To determine 13C in CO2, CuO was not heated; the CO2 was separated using a gas chromatograph and then introduced into the IRMS. The precision of the techniques was estimated to be ±0.10%c for 13CO2 and ±0.30%c for 13CH4.

Carbon isotope ratios in delta notation are calculated as:

sample

where R is the molar ratio of heavy to light isotope (13C/12C), and the international standard for air samples of soil, snowpack and tower is CO2 and CH4 from the Vienna Pee Dee Belemnite (VPDB) standard. The S values are conveniently presented in parts per thousand (%o).

CH4 flux measured by tower observation was estimated by the aerodynamics flux method of Ueyama et al. (2006), who modified the various parameters that were suitable for the Alaska boreal forest, while the vertical wind is weak and the atmospheric stability is near neutral during the winter season.

2.3. Flux measurements ofCO2 and CH4 by the chamber method

The flux of CO2 and CH4 from the snowpack surface can be calculated as the rate of concentration increases in the chamber: dC V d7 ' A '

where J is the flux of a gas, V is the volume of air in the chamber, m3, A is the area covered by the chamber, m2, C is the chamber concentration of a gas, ppm and dC/dt is the rate of concentration change in the chamber air for each gas, ppm/m.

The fluxes of CO2 and CH4 estimated by the chamber method were calculated with eq. (2). The average fluxes of CO2 and CH4 were 0.24 ± 0.05 gC/m2/d [coefficient of Variation (CV): 22%] and 0.023 ± 0.009 mgCH4-C/m2/h (CV: 41%) during the winter period (from DOY 284 to DOY 484), respectively. The fluxes of CO2 and CH4 were measured at an interval of 2 hr for October 8, as is shown in Figure 2. The fluxes of CO2 and CH4 determined by the chamber method were found to be emitted through the snowpack to the atmosphere, suggesting that the soil may play a significant role as a source of CO2 and CH4 in the boreal forest. The fluxes of CO2 and CH4 showed similar diurnal variations during a 24 hr flux measurement. The diurnal variation of CH4 flux similarly showed continuous CH4 flux measured by the tower (shown in Fig. 2b during 24 hr CH4 flux observation), indicating that the black spruce forest soil may be a CH4 source during the winter. CH4 oxidation was observed at a sphagnum moss carpet of black spruce forest soils and at deciduous aspen and birch forest soils (Whalen et al., 1991; Kim and Tanaka,

a) CO2 flux

Oct-08 21:00 Oct-09 05:00 Date

f: " 0

" -2 '3

3 00:00 Oct-08 12:00 Oct-09 00:00 Oct-09 12:00 Oct-10 00:00 Date

Fig. 2. Diurnal variations of (a) CO2 flux by the chamber method and (b) CH4 fluxes by the chamber method and tower observation in the early winter. These suggest that boreal forest soils play a significant role as a source during the winter.

2003) during the growing season, such as during annual CH4 flux monitoring by the tower observation. Ueyama et al. (2006) reported the diurnal variations of CH4 flux measured by tower in black spruce forest of central Alaska; a weak emission on snowpack, an obvious emission around spring thaw, and CH4 uptake in the growing season.

2.4. Flux measurements of CO2 and CH4 by the profile method

The fluxes of CO2 and CH4 through snowpack to the atmosphere under the assumption that the surface soil was the top of a tussock were obtained by applying the following equation under a steady state condition described by Fick's first law:

j=d-----t • e,

where J is flux of a gas, D is the diffusivity of a gas corrected only for the in-situ temperature within the snowpack, cm2/s, dC/dt is the vertical concentration gradient of a gas observed within the snowpack, ppmv/cm, t is tortuosity and e is the snow porosity. Porosity was calculated from the density of ice (aice = 0.91) and the water contents of the snowpack over the gradient interval. Tortuosity is difficult to measure and, in soils, usually is described as a function of porosity with values ranging from e1/3 to e2/3 (Striegl, 1993). In this study, the tortuosity of the snowpack was estimated from the theoretical relation t = e1/3 (Millington, 1959), which yielded values ranging from 0.74

to 0.92. These values are similar to the range of 0.68 to 0.90 through the whole observation period for a boreal forest snow-pack in central Alaska. Sommerfeld et al. (1993) and Mast et al. (1998) reported similar data (0.74-0.90) in a subalpine and alpine snowpack, Wyoming. The effective diffusivities of CO2 and CH4 through the whole snowpack calculated in this study were 0.042 to 0.074 cm2/s and 0.062 to 0.110 cm2/s under an assumption of 1 atm at site during the winter, respectively. Mast et al. (1998) calculated the effective diffusivities of both gases to be 0.061 to 0.144 cm2/s for CO2 and 0.086 to 0.203 cm2/s for CH4 at subalpine soils of Colorado, during 3-year study. These data are much higher than data in our study. This suggests the different physical factors at the observed site, such as snow transformation with time, snow depth, temperature gradient through the snowpack and so on.

During the winter of 2004/5, which featured a thicker snowpack, the average fluxes of CO2 and CH4 estimated by the profile method were 0.22 ± 0.05 gCO2-C/m2/d (CV: 23%) and 0.018 ± 0.005 mgCH4-C/m2/hr (CV: 28%), respectively.

3. Results and discussion

3.1. Temporal variation ofCO2 and CH4 concentration in snowpack and soil

Temporal variations of CO2 and CH4 concentrations at the University of Alaska Fairbanks tower site in central Alaska

within the snowpack at the tussock site (Figs. 3a and 3b) and at soil of sphagnum moss site (Figs. 3c and 3d) during the winter of 2004/5 (using Ocean Data View software of Schlizer, 2004) are shown in Figure 3. The data for above the surface (depth = 0) described the concentrations of CO2 and CH4 from the bottom soil of intertussocks. Because intertussock zones feature very hard and compact soil, the soil air probes could not be mounted in the soil. The vertically empty column in each figure indicates no observation due to extreme cold (>-20 °C), and the horizontally empty row in Figures 3c and 3d denotes no data because of frozen ground layers 10 and 30 cm in depth. However, the data at 5, 20 and 40 cm depths were shown in Figures 3c and 3d, suggesting the unfrozen layer may be due to a peat phase of thick dead Sphagnum moss, which had a higher pore space than soil.

The concentration gradients of CO2 and CH4 within the snowpack at the tussock site are shown in Figure 4, indicating that the depth that shows a maximum concentration is 20 cm at the top of the tussock and emissions of these gases occur through the vascular plant at the tussock tops during winter. The gradients and the depth of CO2 and CH4 have different patterns with time. While the maximum CO2 emits through the vascular plants to the snowpack during the early winter (DOY 322 and 356), the maximum CH4 emits during the early and middle of winter (DOY 312, 322 and 386). This suggests that organic matter (e.g. acetate) in the rhizosphere at the tussock is decomposed into CO2 and CH4 during the early winter and then CO2 is reduced to CH4 during the middle winter (Duddleston et al., 2002; Strom

Fig. 3. Temporal variations of concentrations of (a) CO2 and (b) CH4 in snowpack above the tussock surface and of (c) CO2 and (d) CH4 in soil below the sphagnum moss mat. The vertically empty column indicates no observation due to the severe weather condition, and the horizontally empty row denotes no data owing to the frozen soil, which has no pore space. The contour interval is 100 ppmv within the snowpack for CO2 profile, and 0.5 ppmv within the snowpack CH4 profile during the winter season.

m w HI <■•!

№r№Jllli*¥ ' UtV

OOVtta-jMn^Wfi. NDb

350 400 450 500 550 600 650 700 750 CO concentration (ppm)

3.0 3.5 4.0 4.5 5.0 CH concentration (ppm)

et al., 2003). Or this suggests that the pressured CO2 and CH4 that were produced at a sphagnum site may be transported to a tussock and emitted to the snowpack through tussock vascular plants.

The average tussock height is approximately 20 cm at the site, and the vegetation at the top of the tussocks is vascular plants. Precipitation levels were as follows: snow began to accumulate in early October 2004 and the maximum depth was 78 ± 5 cm (n = 10) in the middle of February 2005. The higher CO2 concentration (>2000 ppm) showed at 40 cm depth of soil in the early winter (DOY 284 to DOY 302). In sphagnum moss, the CO2 concentration varied from 420 ppm at 5 cm (DOY 386) to 1200 ppm at 40 cm (DOY 322) throughout the whole period. In the snowpack over tussocks, CO2 concentration ranged from 370 to 740 ppm at 20 cm (DOY 322). In tower air, the CO2 concentration varied from 471 ppm (DOY 357) to 554 ppm (DOY 388), which is much higher than that (380 ppm) of background air at the NOAA/CMD site in Barrow. Oechel et al. (1997) demonstrated the contribution of winter CO2 emission to the atmosphere measured in tussocks and wet sedges of arctic tundra soil.

In the snowpack, a higher CH4 concentration (6.4 ppm) was recorded at 20 cm depth on the top of tussocks (DOY 386). In the moss regime, the concentration ranged from 1.8 ppm at 40 cm (DOY 322) to 4.6 ppm at 5 cm (DOY 434). In tower observations, the CH4 concentration varied from 2.7 ppm (DOY 436) to 3.6 ppm (DOY 357), which is higher than that (1.9 ppm) of background air at NOAA/CMD in Barrow. Because most of the winter CH4 flux measurements were conducted in anaerobic environments such as bogs, fens, peatlands, and wetlands, the winter CH4 emission is still poorly understood in typical boreal forest soils of central Alaska. Sommerfeld et al. (1993) indicated that CH4 concentration profiles in alpine and subalpine snowpack tended to decrease with depth in Wyoming during a 2 yr winter season, demonstrating that alpine and subalpine aerated soils act as a sink of atmospheric CH4. Whalen et al. (1991), Whalen and Reebugh (1996) and Kim and Tanaka (2003) reported CH4 oxidation in typical boreal forest soils, central Alaska, during the growing season. Therefore, it is necessary to make clear the origins of CO2 and CH4 in boreal forest soils using the stable isotopic discrimination of 13CO2 and 13CH4 from the air

Fig. 4. Temporal variations of (a) CO2 and (b) CH4 concentration profiles within the snowpack at tussock site during the winter. During the middle of winter, the maximum concentrations of CO2 and CH4 are shown at the top (20 cm) of tussock, suggesting that CO2 and CH4 emit through the vascular plants to the snowpack and terminally to the atmosphere.

„'Ml

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 CH4 Flux by profile (mgCH 4-C/m2/h)

Fig. 5. Comparisons of (a) CO2 flux and (b) CH4 flux by the chamber method and the profile method, indicating that correlation coefficients (R) are 0.83 for CO2 and 0.83 for CH4, respectively.

samples of soil, snowpack and tower, as will be described in a chapter 3-3.

3.2. Winter fluxes of CO2 and CH4

Comparison of the winter fluxes of CO2 and CH4 measured by the chamber and the profile methods are shown in Figure 5. The fluxes estimated by two methods are not significantly different based on a one-way ANOVA with a 95% confidence level.

However, the flux ratios of profile to chamber methods were 1.28 ± 0.25 (CV 20%) for CO2 (ratio range: 0.45-3.00) and 0.84 ± 0.22 (CV 21%) for CH4 (range: -0.36-2.60) throughout the winter period, which means the CO2 flux estimated by the profile method is somewhat higher than that estimated by chamber method, and CH4 flux estimated by the chamber method is higher than that estimated by profile method. This indicates that the winter emissions of CO2 and CH4 estimated by chamber method explained 68% and 68% of the emissions of CO2 and CH4 evaluated by the profile method at boreal forest soils during winter. Mast et al. (1998) reported that the chamber fluxes were consistently lower than the gradient fluxes by nearly a factor of 3, resulting from the lower error in estimates of gradient fluxes, an inadequate seal between the chamber and the snow surface, and strong gust winds at the snow surface. However, because other studies were conducted at the Rocky Mountain sites, where season snowpacks typically reach several metres in depth, the snow surface did not reach a steady state condition due to strong winds, and thus deep snow depth and the estimates in chamber fluxes at subalpine sites might result in underestima-tions. In boreal forest soils, Winston et al. (1997) suggested that the winter CO2 flux was a higher chamber flux than the gradient flux due to the presence of melt channels and tree wells.

The CH4 fluxes estimated by the chamber and the profile methods and the tower observations were emitted through the snow-pack to the atmosphere during the winter, as shown in Figure 6. Based on the monitoring CH4 flux at the tower, these soils were a source of CH4 during the winter of 2003/4, whereas the flux was a sink during the growing season in these typical black spruce forest soils. Moreover, in winter, CH4, estimated by both methods, was also emitted to the atmosphere, suggesting that the soil-originated CH4 is a source of atmospheric CH4 during the winter in boreal forest soils of central Alaska. Conclusively, the boreal forest soils should be a significant carbon source

§ 0.00 JE x

¡X -0.05

-0.10 -0.15

0 100 200 300 400 500

DOY (1: January 1, 2004)

Fig. 6. Temporal variations of CH4 fluxes estimated by the chamber, profile and tower observations. The discontinuation of tower CH4 observation occurred due to problems with the CH4 analyzer since January 2005.

reservoir in the form of CH4 as well as CO2 during the winter. We suggest apossible transport mechanism of the soil-originated CH4 through the snowpack to the atmosphere. In early winter, the moss surface temperatures at 5 cm and 100 cm dropped below freezing and rose above freezing on 25 August 2004 (DOY 237); the temperature ranged from -4.8 °C at 04:00 to 25.6 °C at 14:00 ALT (Alaska local time), as shown in Figure 7a. Also, the sphagnum moss carpet began to freeze on 29 September 2004 (DOY 272). The moss column was completely frozen to 30 cm below the surface on 19 October 19 2004 (DOY 290), as shown in Figure 7b. The soil began to freeze on 12 October 2004 (DOY 285) at 10 cm below the surface and on 17 October 2004 (DOY 290) to 30 cm below the surface, indicating that the freezing rate between 10 cm and 30 cm was 4 cm/d (Fig. 7c). Therefore, the cold ambient temperature freezes the moss carpet fromthe surface downward, while permafrost freezes the dead moss layer (peat) upward with time, indicating a possible hypothesis of soil-originated CH4 transport and emission to the atmosphere: (1) the pressure in the carpet is steadily enhanced by the freezing from both directions with time, (2) the frozen moss surface seals off the air-filled pore space, (3) the atmospheric O2 does not reach the soil any more, (4) the soil CO2 begins to reduce to CH4 under suboxic/anoxic environments in the sphagnum moss, (5) the pressured CH4 (higher CH4 concentration) moves to the tussock and finally (6) the higher concentration CH4 releases to the vascular plants at a top of a tussock (20 cm) and then the atmosphere through the snowpack, as shown in Figure 3b. The comprehensive CH4 transport at sphagnum moss and tussock regimes in the black spruce forest soils is schematically shown in Figure 8. This implies the possibility of soil-originated CO2 and CH4 emissions through the vascular plants to the snow-pack and to the atmosphere during the winter. This finding of CH4 emission through the tussock to the snowpack in boreal forest soils of central Alaska reveals another contribution to the carbon budget at a regional scale. However, this phenomenon of winter CH4 emission is not continuous, but intermittent. The temporal variations of pressures at air and in soil 30 cm depth during a month of 2005, and of soil temperatures at soil 5, 10 and 30 cm depth below the surface, are shown in Fig. 9. The difference in pressure, the pressure at soil 30 cm subtracted from air pressure, during <30 d tends to decrease with time. This demonstrates that the pressure at soil 30 cm depth is enhanced. Also, the soil temperature at 5 cm below the surface is affected by colder air temperature and the time lag between soil 5 and 10 cm depths was approximately 3 hr. This suggests that the surface soil is downwardly affected by the extremely cold air temperature. Furthermore, the changes in soil temperatures below 30 cm (not described data) are almost similar, and the soil temperature at 30 cm was upwardly decreased to the end of January, when the whole soil column was frozen as described by Romanovsky (personal communication, Geophysical Institute, University of Alaska Fairbanks). Therefore, although these data are additional experiments, these data enable us to prove the

hypothesis on the process of CH4 transport/emission in boreal forest soils of central Alaska during the winter.

The average wintertime emissions of CO2 and CH4 were 49 ± 13 gCO2-C/m2/(winter season) and 0.11 ± 0.07 gCH4-C/m2/(winter season) for the winter period of 212 d. This sug-40

100 cm

250 300 350 400 450 500 DOY (365: January 1, 2005)

(f) -10

b) 5 cm and 30 cm

1 5 cm

250 300 350 400 450 DOY (365: January 1, 2005)

300 350 400 450 DOY (365: January 1, 2005)

Fig. 7. Temporal variations of (a) 5 and 100 cm above sphagnum moss mat, (b) 5 and 30 cm below the moss layer for temperature and (c) 10 and 30 cm below the moss for soil moisture in near tower site from June 2004 to May 2005. Snow melting and freezing time can be analyzed with soil moisture, relating to the melting and freezing of soil surface condition. It suggests that the soil surface environment provides a clue to the production and emission of CH4 in boreal forest soils.

Freezing

^ * * * * *****

******************* ********* *********

Sphagnum moss

Tussock

Fig. 8. Schematic drawing of the soil-originated CH4 flow in tussock and sphagnum moss regime. The moss layer began to freeze the end of September, and the layer was completely frozen to 30 cm depth below the surface since 19 October 2005 based on the Figure 6b. The curves in the moss layer denote the freezing depth with time, and the depth is affected downward by the cold atmospheric temperature and upward by the permafrost. It suggests that the pore space in dead moss is getting narrower and then the soil air is steadily pressured with time. The stars demote the frozen soil with time. See the text for more detail.

gests that the winter emissions of CO2 and CH4 are an important part of the annual carbon budget in seasonally snow-covered terrain of the boreal forest. The winter CO2 flux corresponds to 23% of the annual CO2 emitted from Alaska black spruce forest soils based on the CO2 emission of 167 ± 89 gCO2-C/m2 during the growing season. Vogel et al. (2005) reported the winter respiration of 30-54 gCO2-C/m2/season during the winter, defined as snow-cover season, in mature black spruce forest soils of Bonanza Creek, central Alaska, these results are similar to values in this study. In alpine and subalpine soils of Wyoming, the average winter CO2 fluxes accounted for 8-23% of the gross annual CO2 emissions (Mast et al., 1998). In a boreal peatland in Finland, the winter fluxes corresponded to 20 to 23% of the total CO2 fluxes (Alm et al., 1999). During the winter season in arctic tundra ecosystems of the North Slope of Alaska, the seasonal emissions from moist tussock tundra and coastal wet sedge was 70 and 20 gCO2-C/m2, which is the main share of the total annual net emission of carbon (Oechel et al., 1997). The contribution of winter CO2 fluxes measured in this study are similar to previously reported values in alpine, boreal forest and tundra ecosystems.

There are no reports for winter CH4 emission in snow-covered boreal forest soils. Whalen and Reeburgh (1988) suspected that downward freezing toward an impermeable permafrost horizon played a role in concentrating the dissolved subsurface methane pool and that a void in the frozen moss matrix permitted an episodic methane release observed along a pond margin with standing water in Alaska. They reported that winter CH4 flux from the moss sites corresponded to 40% of the mean annual CH4 flux. However, their sites are not oxidized soils, but nearly anaerobic environments due to the existence of lake water. Considering the tussock and moss distribution areas in the Northern

970 L. 285

ffi o ro ¡n co ,

ID . i=

290 295 300 305 310 DOY (1: January 1, 2005)

b) Difference in pressure

-3.0 - y = 5 43 -0.022x

(R2 = 0.03) -4.0 i............

285 290 295 300 305 310 315 DOY (1: January 1, 2005)

350 400 450 500 DOY (365: January 1 2006)

Fig. 9. Temporal variations of (a) pressure at 2 m high above the soil surface, (b) daily average pressure difference that subtracts pressure at soil 30 cm depth from that at 2 m high during October 2005 and (c) soil temperatures at 5 (grey), 10 and 30 cm depth below the surface during 2005/6 in sphagnum moss regime. These data are additionally conducted to prove the hypothesis on the transport process of methane in boreal forest soils of central Alaska during the winter.

Hemisphere (6.5 x 1012 m2; Whalen and Reeburgh, 1998), the winter CH4 emission measured in this study should not be overlooked in estimations of the regional/global carbon budget.

3.3. Discriminations of CO2 and CH4

We used a simple mixing model developed by Keeling (1958, 1961) to calculate the isotope ratio of CO2 respired by a forest ecosystem:

[CO2]0 [CO2] f

(80 — SR) + 8 R,

where [CO2] is concentration of CO2 and 8 is the stable carbon isotope ratio of CO2, and the subscripts f and o represent the atmosphere within a forest boundary layer and the atmosphere above (outside) the forest boundary layer, respectively. It can be seen from eq. (4) that a plot of l/[CO2]f versus 8 f gives a straight line relationship with a slope, [CO2]o(8o—8R), and an intercept, 8R. We used this model to estimate 8R, the isotope ratio of CO2 respired by soils. Estimates were obtained from the y-intercept of a geometric mean linear regression (Flanagan et al., 1999) between 8 and 1/[CO2] values measured on air samples collected at different heights above the forest floor during the night when photosynthesis was not active. In this study, the same mixing model (eq. 4) was used to calculate the isotope ratio of respired CO2 from air samples collected during the change in CO2 in soil, snowpack and from the tower. It is possible, therefore, to use the "Keeling plot" approach to calculate separately the isotope ratio of respired CO2 at different spatial scales during the winter.

According to the Keeling plot for the relationship between CO2 concentrations and 813C of CO2, despite different air samples as shown in Figure 10a, the intercept was —22.5%o, which corresponds to the value composed by soil respiration (i.e. the sum of root respiration and microbial respiration in soil) in a black spruce forest in central Alaska. This value is similar to respired 813CO2 (—23.56 ± 1.50%e) found in a moss regime in a black spruce forest during the BOREAS project (Flanagan et al., 1999). Also, when the ambient CO2 level was 385 ppm, the isotopic value was — 8%o, suggesting that the CO2 source is isotopically of the same origin and is formed by the decomposition of soil organic matter, such as acetate. Vogel et al. (2005) demonstrated that proportional contribution of roots in mature black spruce forests varied from 81% to 85% of total soil respiration using the trench method during the winter. Because their sites are similar to our sites, the 813CO2 measured in this study may be mainly originated by root respiration rather than by het-erotrophic respiration.

On the other hand, the relationship between CH4 concentrations and 813C of CH4 was sparse, as is shown in Figure 10b. The two groups are the heavier group (—43%o) and relative lighter group (—51%o) based on the Keeling plot on the relationship between CH4 concentrations and 813C of CH4. However, the CH4 concentrations from tower air samples were not constant,

1/C02*1000 (1/ppm)

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 1/CH4 (1/ppm)

Fig. 10. Comparison of the relationships between (a) I/CO2 concentration and the carbon stable isotope ratio of CO2, (b) I/CH4 concentration and the carbon stable isotope ratio of CH4 in air samples collected soil, snowpack and tower in the boreal forest of central Alaska during the winter season of 2004/5.

ranging from 2.7 ppm (513C: —43.0%c) at the 8 m level (DOY 436) to 3.6 ppm (513C: -49.3%0) at the 8 m level (DOY 357); these are much higher than the ambient concentration of 1.8 ppm (513C: -47.2 ± 0.2 %0) at NOAA/CMD in Barrow during the winter. This suggests that the origin of CH4 could not be identified exactly due to relatively lower tower height and because the tower air samples might be contaminated by another air mass that has higher CH4 concentration in this study area. The 1 d average wind speed and wind direction was monitored at the 8 m level of the tower. Generally, the wind direction during the winter was northwestern (NW). This may imply the influence (513C: —35%c to — 40%e) of the burning of fossil fuel. However, because the carbon isotopes are more enriched than the values from the tower air samples, these data could not definitely explain the origin of CH4 emission. Moreover, Sugimoto and Wada (1993) demonstrated that <513C varied widely in methane from anaerobic environments, mainly because of changes in methanogenic pathways (acetate fermentation: 513C = —30%c to —40%c and CO2 reduction: 513C < — 70%e), comparing our two groups (—43.0%o and 51%e) in soil, snowpack and tower air samples. Therefore,

in order to demonstrate these results, we additionally need to (1) collect background level ambient air samples higher than 10 m by using a high volume balloon for the constant values of CH4 concentrations and 513C of CH4 and (2) monitor the O2 concentration in the soil that is a significant key to understanding changes in soil environments with time during the winter.

4. Conclusions

The winter fluxes of CO2 and CH4 were measured in tussock and sphagnum moss sites of the boreal forest in central Alaska during the winter of 2004/5 using the chamber and the profile methods. While most of the winter CH4 emissions were measured in anoxic environments such as fens, bogs, peatlands and wetlands, the CH4 emission through the snowpack to the atmosphere is here first reported in boreal forest soils. Although the origin of CH4 production was not discerned using the stable carbon isotope of methane, our findings schematically suggest how soil-originated CH4 was transported to the atmosphere based on the CH4 profiles in soil and snowpack, the tower observation, pressure in soil and in air during the winter of 2005/6, and soil temperature and moisture data. Because tussock areas are widely distributed in the Northern Hemisphere, the winter CH4 emission observed in this study is of considerable significance in estimating the seasonally regional/global carbon budget along with the winter CO2 emission that accounts for >20% of the annual soil respiration during the cold season.

5. Acknowledgments

We would like to thank Dr G. Inoue at the National Institute of Environmental Science, Japan, Drs S. Aoki at the Tohoku University, T. Maximov at the Institute for Biological Problem of Cryolithozone and W. Oechel of San Diego State University for valuable discussions during the CO2 Conference. I greatly appreciate Prof. T. Nakazawa at the Tohoku University and an anonymous reviewer for beneficial comments and encouragements. A part of this study was financially supported by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC).

References

Alm, J., Saarnio, S., Nykanenm, H., Silvola, J. and Martikainen, P. J. 1999. Winter CO2, CH4 and N2O fluxes on some natural and drained boreal peatlands. Biogeochemistry 44, 163-186. Born, M., Dörr, H. and Levin, I. 1990. Methane consumption in aerated

soils of the temperate zone. Tellus 42B, 2-8. Brooks, P. D., Williams, M. W. and Schmid, S. K. 1996. Microbial activity under alpine snowpacks, Niwot Ridge, Colorado. Biogeochemistry 32, 93-113.

Deilas, C., Fenandez, M. C., Carretero, J. and Liger, E. 1999. Methane and carbon dioxide fluxes in soils evaluated by 222Rn flux and soil air concentration profiles. Atmos. Environ. 33, 4495-4502.

Dörr, H. and Miinnich, K. O. 1990.222 Rn flux and soil air concentration profiles in West-Germany. Soil222Rn as tracer for gas transport in the unsaturated soil zone. Tellus 42B, 20-28.

Dise, N. B. 1992. Winter fluxes of methane from Minnesota peatlands. Biogeochemistry 17, 71-83.

Duddleston, K. N., Kinney, M. A., Kiene, R. P. and Hines, M. E. 2002. Anaerobic microbial biogeochemistry in a northern bog: acetate as a dominant metallic end product. Global Biogeochemi. Cycles 16,1063, doi:10.1029/2001GB001402.

Fahnestock, J. D., Jones, M. D. and Welker, J. M. 1999. Wintertime CO2 efflux from arctic soils: implications for annual carbon budgets. Global Biogeochemi. Cycles 13, 775-780.

Flanagan, L. B., Kubien, D. S. and Ehleringer, J. R. 1999. Spatial and temporal variation in the carbon and oxygen stable isotope ratio of respired CO2 in boreal forest ecosystem. Tellus 51B, 367384.

Hirano, T. 2005. Seasonal and diurnal variations in topsoil and subsoil respiration under snowpack in a temperate deciduous forest. Global Biogeochemi. Cycles 19, GB2011, doi:10.1029/2004GB002259.

Ishimura, T., Tsunogai, U. and Gamo, T. 2004. Stable carbon and oxygen isotopic determination of sub-microgram quantities of CaCO3 to analyze individual foraminiferal shells. Rapid Commun. Mass Spectrom. 18, 2883-2888.

Keeling, C. D. 1958. The concentration and isotopic abundances f atmospheric carbon dioxide in rural areas. Geochim. Cosmochim. Acta 13, 322-334.

Keeling, C. D. 1961. The concentration and isotopic abundances f atmospheric carbon dioxide in rural and marine air. Geochim. Cosmochim. Acta 24, 277-298.

Kim, Y. and Tanaka, N. 2003. Effect of forest fire on the fluxes of CO2, CH4, and N2O in boreal foest soils, interior Alaska. J. Geophys. Res. 108, 8154, doi:10.1029/2001JD000663.

Mast, M. A., Wickland, K. P., Striegl, R. T. and Clow, D. W. 1998. Winter fluxes of CO2 and CH4 from subalpine soils in Rocky Mountain National Park, Colorado. Global Biogeochem. Cycles 12, 607-620.

Millington, R. J. 1959. Gas diffusion in porous media. Science 130, 100-102.

Nakagawa, F., Toshida, N., Nojiri, Y. and Makarov, V. N. 2002. Production of methane from alasses in eastern Siberia: implications from its 14C and stable isotopic compositions. Global Biogeochem. Cycles 16, doi:10.1029/2000GB001384.

Oechel, W. C., Vourlitis, G. and Hastings, J. 1997. Cold season CO2 emission from arctic soils. Global Biogeochem. Cycles 11, 163-172.

Panikov, N. S. 1999. Fluxes of CH4 and CO2 in high latitude wetlands: measuring, modeling and predicting response to climate change. Polar Res. 18, 237-244.

Panikov, N. S. and Dedysh, S. N. 2000. Cold season CH4 and CO2 emission from boreal peat bogs (West Siberia): winter fluxes and thaw activation dynamics. Global biogeochem. Cycles 14, 10711080.

Raich, J. W. and Potter, C. S. 1995. Global patterns of carbon dioxide emissions from soils. Global Biogeochem. Cycles 9, 23-36.

Raich, J. W. and Schlesinger, W. H. 1992. The global carbon dioxide flux in soil respiration and its relationship to vegetation and climate. Tellus 44B, 81-99.

Schlizer, R. 2004. Ocean Data View, available at http://www.awi-bremerhaven.de/GEO/ODV.

Shannon, R. D. and White, J. R. 1994. A three-year study of controls on methane emissions from two Michigan peatlands. Biogeochemistry 27, 35-60.

Solomon, D. K. and Cerling, T. E. 1987. The annual carbon dioxide in a montane soil: observations, modeling, and implications for weathering. Water Resour. Res. 23, 2257-2265.

Sommerfeld, R. A., Mosier, A. R. and Musselman, R. C. 1993. CO2, CH4 and N2O flux through a Wyoming snowpack and implications for global budgets. Nature 361, 140-142.

Striegl, R. G. 1993. Diffusional limits to the consumption of atmospheric methane by soils. Chemosphere 26, 715-720.

Strom, L., Ekberg, A., Mastepanov, M. and Christensen, T. R. 2003. The effect of vascular plants on carbon turnover and methane emissions from a tundra wetland. Global Change Biology 9, 1185-1192.

Sugimoto, A. and Wada, E. 1995. Hydrogen isotopic composition of bacterial methane-CO2 /H2 reduction and acetate fermentation. Geochim-ica et Cosmochimica Acta 59, 1329-1337.

Takagi, K., Nomura, M., Ashiya, D., Takahashi, H., Sasa, K. and coauthors. 2005. Dynamic carbon dioxide exchange through snow-pack by wind-driven mass transfer in a conifer-broadleaf mixed forest in northernmost Japan. Global Biogeochem. Cycles 19, GB2012, doi:10.1029/2004GB002272.

Ueyama, M., Harazono, Y., Okada, R., Ohtaki, E. and Miyata, A. 2006. Micrometeorological measurements of methane flux at boreal forest in central Alaska. Mem. Natl Inst. Polar Res., Spec. Issue, 59, 156-167.

Vogel, J. G., Valentine, D. W. and Ruess, R. W. 2005. Soil and root respiration in mature Alaska black spruce forests that vary in soil organic matter decomposition rates. Can. J. For. Res. 35, 161-174.

Whalen, S. C. and Reeburgh, W. S. 1988. A methane flux time series for tundra environments. Global Biogeochemi. Cycles 4, 399-409.

Whalen, S. C., Reeburgh, W. S. and Kizer, K. S. 1991. Methane comsuption and emission by taiga. Global Biogeochemi. Cycles 5, 261-273.

Whalen, S. C. and Reeburgh, W. S. 1992. Interannual variations in tundra methane emission: a 4-year time series at fixed sites. Global Biogeo-chemi. Cycles 6, 139-159.

Whalen, S. C. and Reeburgh, W. S. 1996. Moisture and temperature sensitivity of CH4 oxidation in boreal forest. Soil Biol. Biochem. 28, 1271-1281.

Wickland, K. P., Striegl, R. G., Schmidt, S. K. and Mast, M. A. 1999. Methane flux in subalpine wetland and unsaturated soils in the southern Rocky Mountains. Global Biogeochemi. Cycles 13, 101-113.

Wickland, K. P., Striegl, R. G., Mast, M. A., Clow, D. W. 2001. Carbon gas exchange at a southern Rocky Mountain wetland, 1996-1998. Global Biogeochemi. Cycles 15, 321-335.

Winston, G. C., Sundquist, E. T., Stephens, B. B. and Trumbore, S. E. 1997. Winter CO2 fluxes in a boreal forest. J. Geophys. Res. 102, 28795-28804.

Zimov, S. A., Zimova, G. M., Daviodov, S. P., Daviodova, A. I., Voropaev, Y. V. and co-authors. 1993. Winter biotic activity and production of CO2 in Siberian soils: a factor in the greenhouse effect. J. Geophys. Res. 98, 5017-5023.