• Open Access

Carbon budget and methane and nitrous oxide emissions over the growing season in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan

Authors


J. Ryan Stewart, tel. +1 217 265 5461, fax +1 217 244 3469, e-mail: rstewart@illinois.edu

Abstract

Species in the Miscanthus genus have been proposed as biofuel crops that have potential to mitigate elevated atmospheric carbon dioxide (CO2) levels and nitrous oxide (N2O) and methane (CH4) emissions. Miscanthus sinensis is widespread throughout Japan and has been used for biomass production for centuries. We assessed the carbon (C) budget and N2O and CH4 emissions over the growing season for 2 years in a M. sinensis-dominated grassland that was naturally established around 1972 in Tomakomai, Hokkaido, Japan, which is near the northern limit for M. sinensis grassland establishment on Andisols. Average C budget was −0.31 Mg C ha−1, which indicates C was released from the grassland ecosystem to the atmosphere. Dominant components in the C budget appeared to be aboveground net primary production of plants (1.94–2.80 Mg C ha−1) and heterotrophic respiration (2.27–3.11 Mg C ha−1). The measurement of belowground net primary production (BNPP) of plants in the M. sinensis grassland was extremely variable, thus only an approximate value could be calculated. Mean C budget calculated with the approximated BNPP value was 1.47 and −0.23 Mg C ha−1 for 2008 and 2009, respectively. Given belowground biomass (9.46–9.86 Mg C ha−1) was 3.1–6.5 times higher than that of aboveground biomass may provide additional evidence suggesting this grassland represents a C sink. Average CH4 emissions across years of −1.34 kg C ha−1 would indicate this grassland acts as an atmospheric CH4 sink. Furthermore, average N2O emissions across years were 0.22 kg N ha−1. While the site may contribute N2O to the atmosphere, this value is lower compared with other grassland types. Global warming potential calculated with the approximated BNPP value was −5.40 and 0.95 Mg CO2 Eq ha−1 for 2008 and 2009, respectively, and indicates this grassland could contribute to mitigation of global warming.

Nomenclature:
C=

carbon

CO2=

carbon dioxide

CH4=

methane

N2O=

nitrous oxide

ANPP=

aboveground net primary production

BNPP=

belowground net primary production

GWP=

global warming potential

AET=

accumulative effective temperature above 10 °C

Introduction

Using bioenergy instead of fossil fuels could potentially mitigate human-induced global climate change. Although fossil fuels emit carbon (C), the primary source of C for bioenergy is atmospheric carbon dioxide (CO2). Also, due to the steadily rising global demand for energy and the dwindling global reserve of petroleum (Owen et al., 2010), production of bioenergy from plant biomass has increasingly been considered as a viable alternative source of fuel (Milliken et al., 2007; Heaton et al., 2008). Perennial grassland plant species, such as switchgrass (Panicum virgatum) or Miscanthus×giganteus, have been evaluated as nonfood bioenergy feedstock due to their potential for large biomass production and concomitant reduction of greenhouse gas (GHG) emissions (Davis et al., 2009). Using the model predictions from RothC, which is used to calculate projections of soil C turnover, and Miscanthus yields from MISCANFOR, a recently developed Miscanthus growth model (Hastings et al., 2009), Dondini et al. (2009) reported soil under M. ×giganteus production has potential to store, on average, 2–3 Mg C ha−1 yr−1 over a 20-year period.

The Miscanthus genus, which utilizes the energy-efficient C4 photosynthetic pathway (Naidu et al., 2003), is comprised of several species that are considered potential bioenergy crops because of their relative low-nutrient requirements (Lewandowski et al., 2003; Heaton et al., 2004), high water-use efficiencies (Clifton-Brown et al., 2002), and high productivity (Clifton-Brown et al., 2001; Jones & Walsh, 2001; Stewart et al., 2009). M. ×giganteus contributes to C accumulation (0.59–1.12 Mg C ha−1 yr−1) in soil (Hansen et al., 2004; Clifton-Brown et al., 2007), which indicates the potential of this species not only as a bioenergy crop, but also as a C sequester. Application of nitrogen (N) fertilizer often results in reduction of CH4 absorption and increase in N2O emission from soil (Schimel et al., 1993; Conrad, 1995; Bouwman et al., 2001). Given Miscanthus can be produced with relatively little N fertilization (Heaton et al., 2004), species within this genus represent a viable alternative to other grasses, not only because their production can contribute to reductions in GHG emissions and energy use by requiring less N fertilizer. In addition, as Miscanthus is perennial, it does not require annual energy-intensive management such as tillage and seedbed preparation. All these factors can contribute to a relatively low global warming potential (GWP) of Miscanthus as a bioenergy crop.

Over the past 20–30 years, considerable interest has been directed toward M. ×giganteus as a source for bioenergy due to some of the advantages previously mentioned. Interestingly, M. ×giganteus, which is a sterile allopolyploid hybrid, was first collected in Japan in 1935 and then cultivated throughout Europe as an ornamental plant (Scally et al., 2001). In Japan, natural and seminatural grasslands, which constitute 4% of the land area, are comprised of several graminoid and forb species, including Miscanthus sinensis, which is one of the parent species of M. ×giganteus (Himiyama et al., 1995). Nearly 24% of these grasslands are dominated by M. sinensis (National Parks Association of Japan, 1996). Extensive research in Japan over the past several decades focused on the use of M. sinensis as thatching material for roofs of traditional houses and buildings, organic fertilizer, and livestock feed (Stewart et al., 2009).

Past studies also focused on biomass production and vegetation characteristics of M. sinensis grasslands (e.g., Numata, 1975; Midorikawa, 1978). Within the past 10 years, research has been directed towards understanding the effects of these grasslands on climate change (cf. Yazaki et al., 2004; Toma et al., 2010a). However, there are few studies on C-cycling in M. sinensis grasslands (Yazaki et al., 2004). There is also no information available concerning the impact of cultivation of M. sinensis on CH4 and N2O emissions from the soil. In Europe and the United States, considerable effort has been devoted over the past 10–20 years to understanding C sequestration in M. ×giganteus fields (Hansen et al., 2004; Clifton-Brown et al., 2007). However, those studies were conducted in relatively short-term research plots (i.e., 16 years or less) and may not represent long-term conditions. The slower rates of turnover of the humic fractions of organic matter in soil suggest the greatest long-term benefits arise from sequestration in more recalcitrant soil C pools (Clifton-Brown et al., 2007). In contrast to relatively short duration of time cultivated fields of M. ×giganteus have been established, M. sinensis-dominated grasslands in Japan have been actively managed, mostly through burning and mowing, for decades to hundreds of years (Otaki, 1999).

Although M. sinensis exhibits high productivity (1.8–13 Mg ha−1 yr−1, Stewart et al., 2009), biomass production depends on environmental conditions such as temperature and general soil properties (Numata & Mitsudera, 1969). However, there is interest in producing bioenergy crops in lands that may be marginal for other more traditional agricultural crops. The ability of M. sinensis to thrive in limiting environments may represent an important niche for this species as different environments are considered for bioenergy production. However, it is critically important to increase our fundamental understanding of C budgets and GHG emissions under such limiting environments.

The unique settings of both long-term management and limiting environmental conditions found in some M. sinensis grasslands in northern Japan represent a valuable opportunity to study C sequestration and GHG emission parameters. Although the soil and climatic conditions in Japan where these grasslands naturally occur likely differ from where they could be cultivated in the United States and Europe, it should be acknowledged that they provide unique and incomparable information in determining how Miscanthus influences soil dynamics, at least until more long-term data can be acquired in soils more suited for agriculture. Also, this information could ultimately be related to long-term production of Miscanthus for biomass or biofuel. Thus, our objective was to characterize the C budget and CH4 and N2O emissions in a naturally established M. sinensis grassland growing on nutrient-poor soils in a cool climatic region.

Materials and methods

Site description

A M. sinensis-dominated grassland located in central Hokkaido, Japan (42°40′N, 141°45′E, 2 m asl), in which M. sinensis vegetation has naturally grown for more than 30 years on sandy, siliceous, mesic typic endoaquands (USDA, 2010), was selected near the northern natural limit of M. sinensis-type grassland (139°47′N–148°45′N) (Numata, 1969) as the study site (0.1 ha). The study was conducted during 2008 and 2009. Selected physical and chemical properties of the soil are presented in Table 1. The top 16 cm layer (horizons Ah1 to C) was deposited in 1739 by a volcanic eruption of Mt. Tarumae (Soya, 1972). The 16–40 cm soil depth increment (horizons 2Ab to 2C2) was deposited in 1667, also by an eruption of Mt. Tarumae (Soya, 1972). Except for the top layer, which consisted of sandy loam soil, most of the profile was sand (Table 1). Bulk density of the Ah1 horizon (0–5 cm) was 0.2 g cm−3, which increased with depth from 0.68 to 1.05 g cm−3. The top 30 cm layer had 55.7 Mg C ha−1 and 3.9 Mg N ha−1. Bray(2) P in the top horizon was low (5.1 mg P kg−1) compared with deeper horizons (24.7–50.6 mg P kg−1). Exchangeable K in the top three horizons was 165–189 mg K kg−1. The site had a 30-year mean annual precipitation of 1190 mm and air temperature of 7.6 °C (Table 2). The 30-year mean temperature of the growing season, which is generally from May 1 to November 1, where daily mean temperatures are ≥10 °C, was 14.7 °C. Accumulated effective temperature (AET), which is the sum of mean daily temperatures over 10 °C, was calculated following the method of Stewart et al. (2009).

Table 1.   Physical and chemical characteristics of soil under a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan
HorizonDepth
(cm)
Bulk
density
(g cm−3)
Soil texture (%) pHCEC
(cmolc kg−1)
TC
(g C kg−1)
TN
(g N kg−1)
C/NP (Bray 2)
(mg P kg−1)
Exchangeable
K (mg K kg−1)
SandSiltClay
Ah10–50.2076.823.20SL5.518.794.767.3912.85.10165
Ah25–130.6886.014.00S5.659.4863.254.5913.842.5189
C13–160.8997.42.600S5.762.079.550.5318.024.7174
2Ab16–180.9093.65.361.07S5.616.4927.431.8015.233.434.5
2Bb18–220.9199.00.520.52S5.723.806.750.3420.150.627.0
2C122–290.9499.50.510S5.820.572.650.1124.836.144.9
2C2+291.0599.50.510S5.690.562.200.1120.831.829.0
Table 2.   Climate data of a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan
YearMean temperature
(°C)
Precipitation
(mm)
AET* (°C)AET from May 1 to
panicle formation (°C)
  • The growing seasons in 2008 and 2009 were May 6 to November 16 and April 9 to November 8, respectively.

  • *

    Accumulative effective temperature (AET) is accumulative daily mean temperature over 10°C.

  • †Panicle formation began on August 18, 2008 and August 5, 2009.

  • ‡30-year average was from 1980 to 2009.

Annual
 30-year average7.61190  
 20087.91075  
 20098.01263  
Growing season
 30-year average14.78702522 
 200814.681926121471
 200913.896725891275

Plant species in the study site included M. sinensis (25% canopy coverage), woody plant species [Myrica gale (13%), Spiraea salicifolia (3.5%), and Lonicera caerulea (1.0%)], fern species [Thelypteris palustris (33%) and Osmunda cinnamomea (3.2%)], and herbaceous species [Aster tripolium (14%), Sanguisorba tenuifolia (3.4%), Rubia jesoensis (1.3%), and Lycopus maackianus (0.9%)]. Canopy coverage of each species was calculated following the method of JSGS (2004) in August 2009. The site was originally a wetland, but was drained with canals beginning in 1972, and subsequently had no additional management. After draining, it is assumed M. sinensis naturally established at the study site.

Treatments

Four 1 m × 1 m areas within the study site were randomly assigned as vegetation (V) plots for measurement of CH4 and N2O emissions. Three bare-plot treatments (B1, B2, and B3), were randomly assigned to three respective 1 m × 1 m plots for measurement of CO2 emission. After a glyphosate [potassium N-(phosphonomethyl)glycinate] herbicide application, which was applied 50 cm beyond the perimeter of the plots, on June 9, 2008, aboveground dead vegetation in each herbicide-treated area was cut and removed. Root-resistant, water-permeable sheets (Toyobo BKS9812, Toyobo Specialties Trading Co. Ltd., Osaka, Japan) were installed at a 20 cm soil depth around each of the bare-plot treatments to prevent root growth into the plots. After this initial procedure, the B1 plot remained undisturbed. In the B2 plot, the top 20 cm of soil was removed, sieved with 2 mm mesh to remove all belowground plant material, and then returned to the excavated area. In the B3 plot, the same amount of disturbance as in B2 was produced by sieving, but all soil and plant materials were returned to the excavated area.

Measurement of aboveground and belowground biomass

Aboveground and belowground biomass of M. sinensis and other plant species were collected from eight randomly selected 1 m2 plots within the study site on May 18, September 1, and November 4, 2008. Aboveground biomass was collected on May 11, September 1, and October 28, 2009, whereas belowground biomass was collected only on May 11, 2009. After aboveground biomass of M. sinensis and other species were collected, living and dead plant tissues of M. sinensis were separated based on differences in color in September, whereas aboveground biomass of other plants included both living and dead plant material. Belowground biomass was calculated from materials collected from sieving (4 mm mesh) the top 20 cm of soil (Hayashi et al., 1981; Shimizu et al., 2009). Rhizomes of M. sinensis were separated from roots. Given that roots of M. sinensis could not be distinguished accurately from other roots, no separation was attempted. All plant samples were oven-dried at 70 °C for 48 h to constant weight and were then weighed, ground, and analyzed for C with an elemental analyzer (Vario EL III, Elemental, Hanau, Germany).

Calculation of aboveground (ANPP) and belowground net primary production (BNPP)

Throughout Japan, M. sinensis initiates growth in spring and senescence before the onset of winter (Stewart et al., 2009). Considering this seasonal phenology of M. sinensis, ANPP of M. sinensis was estimated by its biomass C at the end of the growing season. Since dead biomass of the previous growing season was not removed at the study site in Tomakomai, some equations were used to not include these data in the calculation of ANPP of the current growing season. ANPP of M. sinensis was calculated as follows:

image

where MBCnew is M. sinensis biomass C at the end of the growing season (November) and MBCold is M. sinensis biomass C in standing biomass from the previous growing season at the end of the current growing season. We assumed standing M. sinensis biomass from the previous season decreased at a constant rate during the current growing season (Koike et al., 1975). Hence, MBCold was estimated by the decrease of dead M. sinensis biomass C from May to September as follows:

image

where MBCMay and MBCSep are dead M. sinensis biomass C in May and September, respectively, Δd is days between May and September and Δd′ is days between May and November.

Peak biomass production of other plant species, which included both annual and perennial plants, occurred during the summer. Considering no dead standing biomass was present from the previous year, ANPP of other plant species for each growing season was calculated by the following equation:

image

where OPBCSep is biomass C of other plant species in September and OPBCMay is biomass C of other plant species in May.

BNPP was also calculated from the seasonal difference in biomass C of rhizome or roots of M. sinensis and other plant species. However, the time interval to determine BNPP was different for roots and rhizomes, given peak growth occurs at different times. Matumura (1998a) reported rhizome biomass decreased before panicle formation because initial M. sinensis growth and development occurred from resources stored in rhizomes. After panicle formation, M. sinensis generally replenishes depleted energy reserves (Matumura, 1998a). Since peak growth of rhizomes occurs after panicle formation and before dormancy, BNPP of Miscanthus rhizomes (BNPPRhizome) was determined for the period between September and November.

image

where RhiCSep and RhiCNov are biomass C of rhizomes in September and November, respectively. However, root development is rapid in the growing season to secure adequate access to water and nutrients for the rapidly growing shoot, but root growth diminishes once plants reach reproductive stages (Eissenstat & Yanai, 2002). Thus, BNPP of roots (BNPPRoot) was determined during the peak period of growth between May and September.

image

where RCMay and RhiCSep are biomass C of rhizome in May and September, respectively. Total BNPP was calculated as follows:

image

Measurement of CO2, CH4, and N2O fluxes

Fluxes of CH4 and N2O were measured in the V plots with a closed-chamber method described by Toma & Hatano (2007). Also, CO2 flux was measured in the B1, B2, and B3 plots. Stainless-steel bases were installed on April 21, 2008 in the V plots and in July 31, 2008 in the three bare plots. Living plant material was not included in the bases. Gas samples for calculating gas fluxes were collected four to five times per month from April to November in both years. Gas fluxes were measured from 10:00 to 14:00 hours. Gas samples (volume=250 mL) were collected into a tedlar bag (volume=500 mL) for CO2 determination, 0 and 6 min from the time the chambers were deployed (Nakano et al., 2004). CO2 concentrations were measured with a CO2 analyzer (ZFP-9, Fuji Electric Systems, Tokyo, Japan). Vacuumed 10 mL vials sealed with butyl rubber stoppers (SVF-10, Nichiden-Rika, Kobe, Japan) were used to collect 20 mL CH4 and N2O gas samples at 0, 15, and 30 min after chamber deployment. Methane (CH4) and N2O concentrations were determined with a gas chromatograph equipped with a flame-ionization detector (GC-8A, Shimadzu, Kyoto, Japan) and an electron capture detector (GC-14B, Shimadzu), respectively.

Fluxes of CO2, CH4, and N2O were calculated with the following equation:

image

where F is the flux (mg m−2 h−1); ñ is the gas density (1.977 × 106 mg m−3 for CO2, 0. 717 × 106 mg m−3 for CH4, and 1.978 × 106 mg m−3 for N2O); V is the volume of the chamber (m3); A is the cross-sectional area of the chamber (m2); Δct is the ratio of change in the gas concentration (c) inside the chamber per unit time (t) during the sampling period (m3 m−3 h−1); T is the air temperature ( °C), and α is a conversion factor for CO2 to C (=12/44), CH4 to C (=12/16), or N2O to N (=28/44).

Calculation of heterotrophic respiration, and cumulative CH4 and nitrous oxide (N2O) emissions

Under natural grassland conditions, CO2 emission from soil surface, which is considered soil respiration (Rs), is the combination of heterotrophic respiration (Rh) and root respiration (Hanson et al., 2000; Subke et al., 2006). In addition, disturbance of soil structure (e.g. when roots, rhizomes, etc. are removed from soil) often induces CO2 emission from soil (e.g. Reicosky et al., 1997; Curtin et al., 2000). The bare-plot treatments were established to accurately estimate Rh, which is often difficult to measure separately from Rs under field conditions (Hanson et al., 2000; Subke et al., 2006), in the V plots. Components of soil CO2 flux from each of the bare-plot treatments were determined as follows:

image
image
image

Thus, Rh was estimated the following equation:

image

CO2 flux from soil strongly depends on temperature (Boone et al., 1998). Thus, cumulative Rh will be underestimated if daily variation of soil temperature is not considered. Considering that soil temperature in the V plots may have been lower than in the bare-plot treatments due to the removal of aboveground biomass, Rh in the V plots was estimated using hourly soil temperature data measured at 5 cm depth in the V plots and a regression model between soil temperature at 5 cm depth and CO2 flux in the bare-plot treatments (Shimizu et al., 2009). The regression model, which is a correction function for CO2 flux (Shimizu et al., 2009), is the following:

image

where f(T) is the CO2 flux (mg C m−2 h−1) and T is soil temperature at 5 cm depth ( °C). Thus, the Rh at t °C in V plots inline imagewas estimated using the regression model f(T) for each bare plot {B1 [fB1(t)], B2 [fB2(t)] and B3 [fB3(t)]}.

image

where t is the soil temperature at 5 cm depth ( °C) in V plots. Hourly data of soil temperature was measured with a thermocouple (n=12) connected to a data logger (FreeSlot-68KD, M.C.S, Hokkaido, Japan) beginning on March 30, 2008 until the experiment was over. The regression model was created by using CO2 flux and soil temperature derived from manual measurements (n=5) at 5 cm depth with a digital thermometer (CT-413WR, CUSTOM, Tokyo, Japan) near the soil chamber at the time of gas measurement in the bare-plot treatments. Cumulative Rh, CH4, and N2O emissions were calculated by linear integration of flux measurements during the growing season of both years (Toma & Hatano, 2007; Toma et al., 2010b).

Calculating C budget and GWP

C budget was calculated as follows:

image

where ANPP and BNPP are the sums of both M. sinensis and other plant species.

Based on estimates that CH4 and N2O GWP are 23 and 296 times higher than CO2, respectively, GWP was calculated by the following equation in CO2 equivalents (IPCC, 2001)

image

Negative GWP values indicate mitigation of global warming.

Ancillary measurements

A composite of three core soil samples from each plot was collected from the 0–10 to 10–20 cm depth increments at each gas sampling measurement. Soil samples were analyzed for NH4+ and NO3 (Toma et al., 2010b). Soil samples were extracted with deionized water (1 : 5) and 2 m KCl (1 : 10) for determining NO3 and NH4+ concentrations, respectively, in the soil. Samples for soil water-filled pore space (WFPS) and volumetric water content calculations were also collected from each plot at the time of gas sampling.

Soil porosity and bulk density was determined from samples collected on August 8, 2008 (porosity=0.821 cm3 cm−3 for 0–10 cm depth and 0.761 cm3 cm−3 for 10–20 cm depth, bulk density=0.274 g cm−3 for 0–10 cm depth and 0.515 for 10–20 cm depth). Air temperature and precipitation data were collected from a weather station in the Tomakomai area.

Statistical analysis

All statistical analyses were performed using ‘r’ (version 2.10.1, R Development Core Team, 2005). Differences in belowground total C biomass across the three measurement times (May 18, September 1, and November 4) in 2008 and across the 2 years (2008 and 2009) in May were analyzed by anova. Regression analysis between soil temperature and CO2 flux was performed by the exponential least squares method. Differences in the regression curves between 2008 and 2009 were also analyzed. The one-sided 95% confidence interval was calculated as follows:

image

The two-sided 95% confidence interval of Rh was estimated from the regression equation of soil temperature and CO2 flux. The two-sided 95% confidence interval of ANPP, CH4, and N2O emissions were determined as follows:

image

where df is the degree of freedom (df=3 for CH4 and N2O emissions, df=7 for ANPP), and t (df, 0.05) is the t-value at 5% significant level with a two-sided alternative.

Results

Although mean annual temperatures in 2008 and 2009 were 0.3 and 0.4 °C higher than the 30-year average, respectively, mean temperature for the growing season was near average in 2008 and 0.9 °C lower than average in 2009 (Table 2). Precipitation in 2008 was 115 mm lower than the 30-year average, whereas in 2009, precipitation was close to average (Table 2). Cumulative precipitation in July and August in 2008 and 2009 represented 127% (458 mm) and 130% (463 mm) of the 30-year mean precipitation in July and August, respectively. Accumulative effective temperature was 90 °C and 67 °C higher than the 30-year average in 2008 and 2009, respectively (Table 2). Panicle formation of M. sinensis started on August 18, 2008 and August 5, 2009. Accumulative effective temperature from May 1, which was when growth began in both years, to panicle formation was 1471 and 1275 °C in 2008 and 2009, respectively (Table 2).

Variation in soil temperature at 5 cm depth was similar to that of air temperature (Figs 1 and 2a). Water-filled pore space was generally below 40% for both surface (0–10 cm depth) and subsurface (10–20 cm depth) soil layers across the growing season in 2008. However, in 2009, only the surface layer was generally below 40% WFPS (Fig. 2c and d). The subsurface layer in 2009 was normally around 40% WFPS with higher levels between July and August due to large precipitation events (Fig. 1). Soil NH4+ concentrations were fairly low and constant across 2008 at the surface and subsurface layers of the soil (Fig. 2e). Similarly, NO3 concentrations in the subsurface were low and constant in 2008, whereas the surface layer had low concentrations through June with a modest increase thereafter (Fig. 2g). In 2009, NH4+ levels were higher than in 2008 especially for the surface soil layer (Fig. 2f). Unlike 2008, surface layer NO3 concentrations in 2009 were generally higher before June (Fig. 2h).

Figure 1.

 Seasonal change of air temperature (black line) and precipitation (gray bar) in 2008 (a) and 2009 (b) in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan.

Figure 2.

 Seasonal change of soil temperature at 5 cm (a, b), water-filled pore space (WFPS) (c, d), ammonium (NH4+) concentration (e, f), and nitrate (NO3) concentration (g, h) in the vegetation plots in 2008 and 2009 in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan. Filled circles and black lines represent the values of 0–10 cm depth of soil. Open circles and dotted lines represent the values of 10–20 cm depth increment of soil. Error bars represent SD values.

Amounts of aboveground biomass-C of plant materials in both years and belowground biomass-C in 2008 are provided in Tables 3 and 4, respectively. Although belowground biomass was 3.1–6.5 times higher than aboveground biomass, there were no significant differences in biomass-C of rhizome or roots of M. sinensis and other plant species across sampling times (rhizome: P=0.48, roots: P=0.59). Rhizome-C did not significantly differ between September and November (P=0.72) nor for root-C between May and September (P=0.57). In addition, belowground biomass-C did not significantly differ between May 2008 and May 2009 (rhizome, P=0.64; roots, P=0.13).

Table 3.   Seasonal change of mean biomass carbon (C) of living and dead aboveground biomasses of Miscanthus sinensis and other plants in a grassland in Tomakomai, Hokkaido, Japan
YearDateAboveground biomass C (Mg C ha−1)
M. sinensisOther plants
LivingDead
  1. Values between parentheses represent standard deviations (n=8).

2008May 181.00 (0.61)0.45 (0.44)
September 10.54 (0.19)0.45 (0.26)2.12 (0.31)
November 41.28 (0.83)1.31 (0.30)
2009May 110.83 (0.30)0.52 (0.32)
September 10.55 (0.41)0.29 (0.26)1.54 (0.14)
October 280.94 (0.62)0.85 (0.41)
Table 4.   Seasonal change of mean biomass carbon of belowground biomass of rhizomes of Miscanthus sinensis and roots of M. sinensis and other plant species in a grassland in Tomakomai, Hokkaido Japan
YearDay-MonthBelowground
Rhizome of
M. sinensis
Root of
M. sinensis and
other plant species
Total
  1. Values between parentheses are standard deviations (n=8).

200818-May1.57 (0.62)7.89 (1.66)9.46 (1.66)
1-Sep1.18 (0.40)8.68 (1.71)9.86 (1.76)
4-Nov1.34 (0.58)8.23 (1.50)9.57 (1.50)
200911-May1.32 (0.75)7.49 (4.03)8.81 (4.04)

There was significant correlation between soil temperature and CO2 flux in the bare-plot treatments (Fig. 3a–c). In the B1 plot, exponential regression models significantly differed between years (Fig. 3a). Estimated values of Rh increased during the summer (Fig. 4a and b). Average Rh in 2009 was 1.37 times higher than in 2008 (Table 5). Decomposition of dead belowground plant material and CO2 flux induced by soil disturbance in 2008 and 2009 varied from 0.29 to 0.55 and −0.56 to −0.90 Mg C ha−1 growing season−1, respectively.

Figure 3.

 Relationships between soil temperature at 5 cm depth and CO2 flux in B1 (a), B2 (b), and B3 (c) plots in 2008 and 2009 (b) in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan. Filled and open circles represent values in 2008 and 2009, respectively. Solid and dotted lines represent exponential regression models between CO2 flux and soil temperature in 2008 and 2009, respectively. The model is F=a × exp(b ×T), where F is the CO2 flux (mg C m−2 h−1) and T is soil temperature at 5 cm depth (°C). *Significant difference in regression models among years (**P<0.01).

Figure 4.

 Seasonal change of heterotrophic respiration in the vegetation plots in 2008 (a) and 2009 (b) in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan.

Table 5.   Heterotrophic respiration, decomposition of dead belowground plant material, and CO2 flux induced by soil disturbance during the growing season in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan
 Heterotrophic
respiration
Decomposition of
dead belowground
plant material
CO2 flux
induced by soil
disturbance
Year(Mg C ha−1 growing season−1)
  1. Values between parentheses represent one-sided 95% confidence interval.

20082.27 (0.31)0.55 (0.86)−0.56 (2.03)
20093.11 (1.25)0.29 (0.36)−0.90 (0.97)

CH4 flux was zero or negative except during August (19.4 μg C m−2 h−1) and September (25.2 μg C m−2 h−1) in 2008 (Fig. 5a and b). Average CH4 emission across years calculated from the values in Table 6 was −1.34 kg C ha−1 growing season−1. During the growing season, N2O flux tended to peak in August (13.8 and 14.5 μg N m−2 h−1 in 2008 and 2009, respectively) (Fig. 5c and d). In 2009, high N2O fluxes were also observed in April (47.5 and 52.3 μg N m−2 h−1) (Fig. 5d). Average N2O emission across years calculated from the values in Table 6 was 0.22 kg N ha−1 growing season−1. There was significant, but weak positive correlation between CH4 flux and WFPS in the top 10 cm of soil throughout the study period (n=179, R2=0.03, P<0.05).

Figure 5.

 Seasonal change of methane (CH4) (a, b) and nitrous oxide (N2O) (c, d) fluxes in vegetation plots in 2008 and 2009 in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan. Error bars represent SD values.

Table 6.   Average methane (CH4) and nitrous oxide (N2O) emissions during the growing season in a Miscanthus sinensis grassland in Tomakomai, Hokkaido, Japan
YearCH4 emission (kg C ha−1
growing season−1)
N2O emission (kg N ha−1
growing season−1)
  1. Values between parentheses represent one-sided 95% confidence interval.

2008−1.07 (0.59)0.07 (0.06)
2009−1.62 (0.25)0.36 (0.18)
Average−1.340.22

Averaged across years, total values of ANPP of M. sinensis and other plant species calculated from values in Table 7 were 1.04 and 1.34 Mg C ha−1 growing season−1, respectively. Differences in biomass-C of rhizomes between the periods May–September and September–November in 2008 calculated from values in Table 4 were −0.39 and 0.15 Mg C ha−1, respectively. However, biomass-C values of roots of M. sinensis and other plant species for the same time periods were 0.79 and −0.45 Mg C ha−1, respectively. Given the distinct peak growth periods of these different belowground organs, BNPP values of rhizome, root, and the sum of both were calculated to be 0.15, 0.79, and 0.94 Mg C ha−1 growing season−1, respectively. Unfortunately, due to extreme variability in BNPP measurements, these BNPP values were considered only an approximation due to lack of statistical differences in belowground biomass of rhizomes of M. sinensis and roots of M. sinensis and other plant species among the three sampling times. Thus, the C budget presented in Table 7 was a partial calculation that did not include BNPP. Although the C budget in 2008 was positive (0.53 Mg C ha−1 growing season−1), in 2009, it was negative (−1.17 Mg C ha−1 growing season−1) (Table 7). The average of the C budget across years was −0.31 Mg C ha−1 growing season−1. Net GWP values in 2008, 2009, and the average across years were −1.95, 4.40, and 1.22 Mg CO2 Eq ha−1 growing season−1, respectively (Table 8). Including the approximation of BNPP in the calculation of the C budget for 2008 would have resulted in a C budget of 1.47 Mg C ha−1 growing season−1. Given the high variability of belowground biomass measurements, and that biomass-C in May 2009 was not different than in May 2008, it is possible to assume the estimated BNPP value in 2008 would be similar for 2009. Under this assumption, including BNPP in the C budget calculation for 2009 would have resulted in a C budget of −0.23 Mg C ha−1 growing season−1. The average of the C budget across years would be 0.62 Mg C ha−1 growing season−1. Similarly, if the estimated BNPP value were included in the calculations of GWP in 2008 and 2009, GWP would have been −5.40 and 0.95 Mg CO2 Eq ha−1 growing season−1, respectively. The average of GWP across years would have been −2.22 Mg CO2 Eq ha−1 growing season−1.

Table 7.   Aboveground net primary production (ANPP) of Miscanthus sinensis and other plant species, belowground net primary production (BNPP), heterotrophic respiration (Rh), methane (CH4) emission, and carbon budget during growing season in a M. sinensis grassland in Tomakomai, Hokkaido, Japan
YearANPP (Mg C ha−1
growing season−1)
BNPP (Mg C ha−1
growing season−1)
Rh (Mg C ha−1
growing season−1)
CH4 emission (kg C ha−1
growing season−1)
Carbon budget (Mg C ha−1
growing season−1)
M. sinensisOther plants speciesRhizomes of M. sinensisRoots of M. sinensis and other plants
  • Negative values indicate C was emitted from the ecosystem. Values between parentheses represent one-sided 95% confidence interval.

  • *

    Values were not included into the C budget

20081.14 (1.17)1.66 (0.64)0.15 (0.06)*0.79 (0.24)*2.27 (0.31)−1.07 (0.59)0.53 (2.69)
20090.93 (0.74)1.01 (0.44)3.11 (1.25)−1.62 (0.25)−1.17 (1.41)
Average1.041.342.69−1.34−0.31
Table 8.   Above- and belowground net primary production (ANPP and BNPP) of Miscanthus sinensis and other plant species; heterotrophic respiration (Rh); methane (CH4) and nitrous oxide (N2O) emissions; and global warming potential (GWP) in a M. sinensis grassland in Tomakomai, Hokkaido, Japan
YearANPP (Mg CO2 Eq ha−1
growing season−1)
BNPP (Mg CO2 Eq ha−1
growing season−1)
Rh (Mg CO2 Eq ha−1
growing season−1)
CH4 emission
(Mg CO2 Eq ha−1
growing season−1)
N2O emission
(Mg CO2 Eq ha−1
growing season−1)
Net GWP
(Mg CO2 Eq ha−1
growing season−1)
M. sinensisOther plants speciesRhizomes of M. sinensisRoots of M. sinensis and other plants species
  • Negative values indicate the mitigation of global warming. Values between parentheses represent one-sided 95% confidence interval.

  • *

    Values were not included in the calculation net GWP.

2008−4.18 (4.29)−6.09 (2.35)−0.55 (0.22)*−2.90 (0.88)*8.32 (1.12)−0.03 (0.02)0.03 (0.03)−1.95 (5.02)
2009−3.41 (2.71)−3.71 (1.61)11.4 (4.60)−0.05 (0.01)0.17 (0.08)4.40 (5.58)
Average−3.80−4.909.86−0.040.101.22

Discussion

Environmental factors for C budget

The average C budget during this study was −0.31 Mg C ha−1 growing season−1, which indicates C was released from the grassland ecosystem to the atmosphere. However, including the approximate value of BNPP, the average C budget of the M. sinensis grassland in Tomakomai would have been 0.62 Mg C ha−1 growing season−1. Dominant components in the C budget appeared to be ANPP, Rh, and BNPP (Table 7). Total ANPP of M. sinensis and other plant species in our study site was low (1.94–2.80 Mg C ha−1 growing season−1) compared with other studies. Yazaki et al. (2004) reported ANPP ranging from 11.4 to 12.1 Mg C ha−1 yr−1 for a M. sinensis grassland on an Andisol soil in Nagano, Japan (36°31′N, 138°21E, 1315 m asl). They suggested the high ANPP of the M. sinensis grassland, relative to that of 10 western North American grasslands reported by Sims & Singh (1978), might have been due to high precipitation. Jones & Walsh (2001) reported dry weight yields of M. ×giganteus increased with increasing precipitation and cumulative temperature during the growing season (May 1–October 31). Dry weight yields of M. ×giganteus (19 Mg ha−1) maximized at approximately 400 mm of precipitation with a cumulative temperature of 2740 °C during the growing season based on data collected from several fields site throughout Europe (Clifton-Brown et al., 2001). Moreover, Sala et al. (1988) and Paruelo et al. (2010) reported a significant positive linear relationship between annual precipitation and ANPP in grassland ecosystems in the US Midwest and south-central South America. Although Yazaki et al. (2004) considered high precipitation (1288–1396 mm) to contribute to high ANPP in the M. sinensis grassland in Nagano, ANPP at the Tomakomai study site was lower despite having comparable precipitation (1075–1263 mm). In addition, mean annual temperature at our study site ranged from 7.9 to 8.0 °C, which is more conducive for growth than the mean annual temperatures of 6.1–6.5 °C in the study by Yazaki et al. (2004). It appears the low ANPP at our study site was not the result of these climatic conditions.

Numata & Mitsudera (1969) reported height of M. sinensis, which can be used to estimate aboveground biomass (Yoshida, 1976), increased with increasing thickness of surface–soil A horizon in Miyagi, Japan. Thicker A horizons are often the result of soil development that can increase not only fertility levels but other soil properties that are important for plant development. Thickness of the A horizon in the study field of Yazaki et al. (2004) was 62 cm (Suzuki et al., 1999), which was deeper than that of our study site. The M. sinensis grassland at Tomakomai was established on a young volcanic ash soil that started to develop 269 years ago. In addition to the shallow depth of the A horizon at the study site, the entire rooting depth was limited by a water table 40 cm below the soil surface. Unfortunately, the study by Yazaki et al. (2004) provides no information about soil-fertility parameters. In an adjacent study to the Tomakomai grassland study site, application of P fertilizer elicited a significant yield response (data not published). Although Miscanthus is considered to have low-nutrient requirements (Lewandowski et al., 2003; Heaton et al., 2004; Stewart et al., 2009), it is important to acknowledge insufficient availability of nutrients could negatively impact M. sinensis productivity. Insufficient fertility, possibly along with other soil or environmental constraints, may have limited ANPP at the M. sinensis grassland in Tomakomai.

ANPP of M. sinensis and other plant species was lower in 2009 than in 2008 (Table 7). This was probably due to lower temperatures during the growing season in 2009 than in 2008 (Table 2). It is well known that net photosynthetic rates of C3 and C4 plant species increase with increasing temperature when temperatures are between 10 and 20 °C (Berry & Björkman, 1980; Hikosaka, 2004). In addition, AET during the 2009 growing season, including the period of panicle formation, was 196 °C lower than in 2008 (Table 2). Matumura (1997) reported biomass production of M. sinensis increased with increasing AET from spring until panicle formation.

Heterotrophic respiration was 2.27 and 3.11 Mg C ha−1 growing season−1 in 2008 and 2009, respectively (Table 7). Owing to snow cover and low temperatures (approximately 0 °C) during the winter season at the Tomakomai study site (Fig. 3), CO2 flux was likely very low. Moreover, nearly 0 mg C m−2 h−1 was reported in a field of reed canarygrass (Phalaris arundinacea) under similar growing conditions in Hokkaido as the M. sinensis grassland in Tomakomai during the winter season (Shimizu et al., 2009). Therefore, calculated Rh at the M. sinensis grassland likely approximates that of the entire year. In addition to the 20 cm deep root barriers installed in the soil around the bare plots, herbicide was applied 50 cm beyond the perimeter of the plots. Hence, the difference in Rh between 2 years might not be due to the intrusion of living plant roots into the bare plots. Although calculating Rh by removing roots has been widely adopted (e.g., Subke et al., 2006), the possibility exists that CO2 flux measured with our modified root-exclusion method may have been influenced by soil disturbance. However, estimates of Rh in our study are nonetheless valid since the effects of both of soil disturbance and dead belowground plant materials on Rh were likely minimal compared with the more commonly used root-exclusion method. However, it should be acknowledged that root-exclusion methods need to be continually improved and refined.

There is generally an optimal soil moisture range where soil CO2 flux is maximized (Gulledge & Schimel, 1998; Harper et al., 2005; Yan et al., 2010). Harper et al. (2005) reported soil CO2 flux increased with increasing soil temperature, but reached a maximum at approximately 30% of volumetric soil water content in a silty clay loam soil. Toma et al. (2010b) reported Q10, which is a relative increase of CO2 flux for every 10 °C change in soil temperature, increased from 1.11 to 2.38 with increasing annual precipitation that increased from 1015 to 1576 mm. At the Tomakomai study site, precipitation during the growing season in 2009 was 20% higher (148 mm) than in 2008 (Table 2). In addition, exponential regression models of the B1 plot in 2009 was significantly higher than in 2008 (Fig. 3a), which may indicate higher moisture conditions in 2009, relative to 2008, were responsible for the increase in Rh during 2009.

It should be considered that CO2 emissions from the study site might have included C that originated from the oxidation of organic matter accumulated when the site was a wetland. Increase in CO2 emissions from drained wetlands is a well-known phenomenon (Nykänen et al., 1995; Wright & Reddy, 2001). However, CO2 emission in our study was lower compared with other studies. Shimizu et al. (2009) reported Rh ranged from 3.6 to 3.9 Mg C ha−1 yr−1 in a field of reed canarygrass in Hokkaido, Japan (42°26′N, 142°29′E) on Andisols with the same parent material as the M. sinensis grassland in Tomakomai. Frank et al. (2004) reported Rh ranged from 4.52 to 5.09 Mg C ha−1 yr−1 in a switchgrass field on Mollisols in South Dakota, USA (46°46′N, 100°55′W). In those sites, mean annual temperatures were higher than at the Tomakomai site (Table 9). Although soil CO2 flux is influenced by temperature (Boone et al., 1998), these studies were all in cooler northern latitudes where temperatures may not have such a strong influence given that the increasing rate of CO2 flux has been shown to be low at low temperatures compared with that at higher temperatures (Boone et al., 1998). More importantly, however, soil organic-C levels are considered to be the primary source of CO2 produced in soil. At the Tomakomai site, soil C in the 0–30 cm depth increment was 55.7 compared with 76.6 Mg C ha−1 in a reed canarygrass field reported by Shimizu et al. (2009). Soil CO2 flux levels reported by Shimizu et al. (2009), which were higher than those measured in the M. sinensis grassland in Tomakomai, were likely due to the higher soil-C levels at their study site.

Table 9.   Tabulated values of carbon (C) accumulation under Miscanthus sinensis, Miscanthus x giganteus, and switchgrass (Panicum virgatum) fields
VegetationGrassland
type
LocationCoordinatesAir
temperature
(°C)
Precipitation
(mm)
Elevation
(m)
Soil typeManagementMethodC accumulation
(Mg C ha−1 yr−1)
Authors
  1. Natura, =grassland was naturally established; Semi-natural, grassland was naturally established, but mowing, fertilization, grazing or burning frequently occurred; Intensive, grassland was established artificially; E, ecological method; δ13C, isotopic δ13C analysis; SCC, soil C change.

Miscanthus sinensisNaturalHokkaido, Japan42°40′N, 141°45′E7.611902AndisolsNaturalE−1.16–0.53This study
Semi-naturalNagano, Japan36°31′N, 138°21′E6.511261315AndisolsMowingE−1.0–0.5Yazaki et al. (2004)
Miscanthus×giganteusIntensiveNorth Jutland, Denmark56°50′N, 9°26′E7.470630InceptisolsMowingδ13C0.78–1.12Hansen et al. (2004)
IntensiveCo. Tipperary, Ureland52°39′N, 7°50′W9.9100490InceptisolsMowing, fertilization (chemical)δ13C0.59Clifton-Brown et al. (2007)
SwitchgrassIntensiveQuebec, Canada42°25′N, 75°56′W6.51062370N/AMowing, fertilization (organic)SCC3.5Zan et al. (2001)
(Panicum virgatum)IntensiveSouth Dakota, USA44°10′N, 96°41′W6.3602470MolisolsMowing, fertilization (chemical)SCC2.4Lee et al. (2007)
Intensive      Mowing, fertilization (organic)SCC4Lee et al. (2007)
Intensive      MowingSCC0Lee et al. (2007)
IntensiveSouth Dakota, USA46°46′N, 100°55′W4404588MolisolsMowing, fertilization (chemical)E5.61–9.99Frank et al. (2004)

The amount of variability across sampling times made it difficult to estimate BNPP. Although it is a basic and important ecological variable, estimating BNPP has also been identified as a significant challenge in many other studies (Midorikawa, 1978; Lauenroth, 2000; Yazaki et al., 2004). Midorikawa (1978) and Yazaki et al. (2004) emphasized the difficulty of estimating BNPP in M. sinensis grasslands in Japan due to the large spatial variation in belowground biomass. The large amount of variability that can exist for this measurement was illustrated in a study on productivity levels of steppes and prairies in Uruguay, Argentina, and Brazil (Paruelo et al., 2010). BNPP ranged from 2.64 to 5.68 Mg C ha−1 growing season−1. Cahill et al. (2009) reported that the ratio of BNPP to ANPP of C3 (e.g. Bromus inermis Leyss., Phleum pratense L.) and C4 (e.g. Andropogen gerardii, switchgrass, Sorghastrum nutans) species varied from 16% to 342% and 21% to 388%, respectively, in Wisconsin, USA (43°4′N, 89°49′W). BNPP of rhizomes of M. sinensis and roots of the entire plant community in the M. sinensis grassland in Tomakomai were approximated based on the growth pattern of rhizome and roots (Table 7). The strategy of M. sinensis to use stored reserves in the rhizome for regrowth, reported by Matumura (1998a), has been observed for other species, including alfalfa (Medicago sativa L.), which uses stored reserves in its roots, and M. ×giganteus (Suzuki & Stuefer, 1999; Berg et al., 2009). Although our measurements provide only a crude estimate of BNPP, this value indicates the M. sinensis grassland ecosystem in Tomakomai might act as a sink of atmospheric C. Numata (1976) reported ANPP and BNPP of a M. sinensis grassland in Miyagi, Japan were 5.0 and 1.75 Mg ha−1 yr−1, respectively. Based on other studies of the same grassland, Numata (1976) also reported BNPP of M. sinensis was 25–30% of ANPP. Considering ANPP of M. sinensis in our study was 0.93–1.14 Mg C ha−1 growing season−1, if values reported by Numata (1976) are used, BNPP of M. sinensis could possibly range from 0.23 to 0.34 Mg C ha−1 yr−1. This value range would be higher than our estimated values and would indicate that our estimate is very conservative. Given belowground biomass of M. sinensis and other plant species in Tomakomai was large compared with aboveground biomass may provide additional evidence suggesting this grassland represents a C sink.

Comparison of C budget in M. sinensis grassland with other vegetations

Without taking into account BNPP, the average C budget was −0.31 Mg C ha−1 growing season−1. Similar values were obtained by Yazaki et al. (2004). They reported the C budget of a M. sinensis grassland established in an Andisol soil in Nagano, Japan, varied from −1.0 to −0.5 Mg C ha−1 yr−1. These values were also calculated without accounting for BNPP since it could not be measured due to large variability. When no mowing was done, the C budget varied from 3.57 to 4.03 Mg C ha−1 yr−1. Values of C budget at the Tomakomai site, and that reported by Yazaki et al. (2004), were low compared with those reported for M. × giganteus under production management conditions (Hansen et al., 2004; Clifton-Brown et al., 2007,Table 9). Hansen et al. (2004) reported accumulated soil-C, which was detected by changes in δ13C in soil under 9- and 16-year-old fields of M. ×giganteus grown on Inceptisols was 0.78 and 1.12 Mg C ha−1, respectively. In addition, Clifton-Brown et al. (2007) reported 0.59 Mg C ha−1 yr−1 of soil-C was derived from M. ×giganteus production over a 15-year period on Inceptisols in Ireland (52°39′N, 07°50′W), which was based solely on BNPP given aboveground biomass had not been included in the calculation. Average estimated yields of M. ×giganteus have been reported to be as high as 22.4 Mg ha−1 (Heaton et al., 2004), whereas aboveground biomass yield of M. sinensis ranges from 1.8 to 13 Mg ha−1 in grasslands in Japan (Stewart et al., 2009). Hence, C accumulation in M. sinensis grassland under natural conditions might be lower than that in cultivated M. ×giganteus fields. For cultivated switchgrass, C accumulation rate ranged from 0 to 10.0 Mg C ha−1 yr−1 (Zan et al., 2001; Frank et al., 2004; Lee et al., 2007,Table 9). Except for the reported values of C budget in a switchgrass field where organic fertilizer was applied (Lee et al., 2007), C accumulation rate in switchgrass fields might be larger than for M. sinensis grasslands (Table 9).

CH4 and N2O emissions

During the study period, negative CH4 fluxes frequently occurred (Fig. 5a and b), which indicated atmospheric CH4 was absorbed into the soil. Positive CH4 fluxes occurred infrequently in August and September of 2008, which was likely due to relatively high soil moisture that may have caused hypoxic conditions. Low soil redox potential generally results in high CH4 flux due to increased anaerobic respiration (Schimel et al., 1993; Conrad, 1995). This was further confirmed by the significant positive correlation between CH4 flux and WFPS (0–10 cm) in the M. sinensis grassland in Tomakomai. Mori et al. (2008) also reported there was a significant positive correlation between CH4 flux and soil moisture content in mixed orchardgrass (Dactylis glomerata) grasslands established on Andisols in Tochigi, Japan (36°55′N, 139°55′W). At the Tomakomai study site, CH4 flux was approximately −50 μg C m−2 h−1. Mosier et al. (1991) reported CH4 flux ranged from −4.5 to −46 μg C m−2 h−1 in a shortgrass steppe in Colorado, USA (40°48′N, 104°45′W). Boeckx & Van Cleemput (2001) reported CH4 fluxes in European grasslands varied from −5.6 to −35 μg C m−2 h−1. Moreover, they suggested the average CH4 oxidation capacity in grassland soils was −1.88 kg C ha−1 yr−1 (−2.5 kg CH4 ha−1 yr−1) which was similar to our results in 2008 (−1.07 kg C ha−1 growing season−1) and 2009 (−1.62 kg C ha−1 growing season−1) (Table 6). Because CH4 emission was nearly 0 μg C m−2 h−1 during the winter in a field of reed canarygrass on Andisols in Hokkaido, Japan with the same parent material as the M. sinensis grassland in Tomakomai (Hatano et al., 2008), CH4 emission measured in Tomakomai during the growing season could be regarded as an annual CH4 emission. As such, these data indicate the M. sinensis grassland in Tomakomai might act as a sink of atmospheric CH4.

N2O flux varied from −7.7 to 52 μg N m−2 h−1 (Fig. 5), which is similar in magnitude to what has been reported in other studies (Mosier et al., 1991; Jørgensen et al., 1997; Mori et al., 2005). Jørgensen et al. (1997) reported that N2O flux from an unfertilized M. ×giganteus field on Inceptisols ranged from −1 to 35 μg N m−2 h−1 in Hornum, Denmark. Mori et al. (2005) reported N2O flux ranged from 1 to 122 μg N m−2 h−1 in an unfertilized, mixed orchardgrass grassland on Andisols in Tochigi, Japan. Mosier et al. (1991) reported N2O flux in an unfertilized shortgrass steppe in Colorado ranged from 0 to 45 μg N m−2 h−1. High N2O flux was observed in April 2009 at the M. sinensis grassland in Tomakomai, which is during the period of heavy snowmelt in central Hokkaido. Tiedje (1994) and Flessa et al. (1995) reported high N2O emissions from soil in the winter and early spring could be a general phenomenon in temperate and boreal climates where soils are subjected to periodic freezing and thawing. In August 2008 and 2009, N2O flux slightly peaked to 13.8 and 14.5 μg N m−2 h−1, respectively. N2O is usually produced by nitrification and/or denitrification processes in soil (Tiedje, 1994; Bouwman et al., 2001). In August, anaerobic conditions might have occurred in soil pore space due to high precipitation. Anaerobic conditions in August also induced denitrification. Average annual N2O emission at the Tomakomai study site was 0.22 kg N ha−1 growing season−1. Because low N2O emission during the winter, which ranged from 0.3 to 4.7 μg N m−2 h−1, was reported in a field of reed canarygrass under similar growing conditions in Hokkaido as the M. sinensis grassland in Tomakomai (Jin et al., 2010), N2O emission during the growing season at the Tomakomai study site likely represented CH4 emission for the entire year. Jørgensen et al. (1997) reported average N2O emission from an unfertilized M. ×giganteus field on Inceptisols from April to August was 0.14 kg N ha−1. Akiyama et al. (2006) estimated annual N2O emission in unfertilized upland fields in Japan was 0.36 kg N ha−1 yr−1 for well-drained soils and 1.40 kg N ha−1 yr−1 for poorly drained soils. Mori et al. (2005) reported that annual N2O emissions from orchardgrass, white clover, and mixed orchardgrass-white clover grasslands were 0.39, 1.59, 0.67 kg N ha−1 yr−1, respectively. Stehfest & Bouwman (2006) reported that estimated annual N2O emissions in grasslands and steppe was 0.48 kg N ha−1 yr−1, whereas annual N2O emission in temperate forest was 0.22 kg N ha−1 yr−1. It appears N2O emission in the M. sinensis grassland was lower than in unfertilized upland fields or grasslands and similar to that of temperate forests.

GWP in M. sinensis grassland

Average net GWP in this study was 1.22 Mg CO2 Eq ha−1 growing season−1, indicating that M. sinensis grassland at our study site accelerated global warming. However, including the approximated BNPP value results in average net GWP of −2.22 Mg CO2 Eq ha−1 growing season−1. Mosier et al. (2005) reported net GWP ranged from −0.80 to −0.48 Mg CO2 Eq ha−1 yr−1 in a restored prairie, whereas it varied from −0.63 to 0.31 Mg CO2 Eq ha−1 yr−1 in an upland fields under corn (Zea mays), soybean (Glycine max), and wheat (Triticum aestivum) rotation in Colorado, USA (40°22′N, 103°7′W). Mosier et al. (2006) also reported net GWP in an agricultural field under corn-soybean rotation ranged from −1.3 from 2.5 Mg CO2 Eq ha−1 yr−1 in Colorado, USA (40°39′N, 104°59′W). Mu et al. (2006) reported net GWP ranged from 0.75 to 1.8 Mg C ha−1 yr−1 in agricultural fields under different crop rotations in Hokkaido, Japan (43°14′N, 141°50′W). Relative to that reported above, net GWP in this study was relatively low when BNPP was considered in its calculation. This would indicate the M. sinensis grassland in Tomakomai may be helping to mitigate global warming. Similar to the C budget of the M. sinensis grassland in Tomakomai, NPP of M. sinensis and other species and Rh were the primary contributors to net GWP (Table 8).

Conclusions

Evaluation of the C budget and GHG emissions of the M. sinensis grassland in Tomakomai indicates the site may act as a C sink and help mitigate global warming. The most important factors for C budget and GWP in this grassland were ANPP, Rh, and BNPP, whereas CH4 and N2O emissions were not as important. Belowground biomass of M. sinensis and other plant species was large compared with aboveground biomass. The results of our study illustrate belowground biomass is not a trivial component for the characterization of C budgets and GHG emissions. The inherent variability in the measurement of belowground biomass makes it extremely difficult to accurately determine BNPP, which highlights the need to improve measurement techniques to better characterize the contribution of the belowground portion to the total C budget and GWP. Until variability in the measurement of the various components of the C budget and GWP parameters can be reduced or better measured, it will continue to be difficult to accurately estimate ecological benefits associated with these grasslands.

Acknowledgements

We are grateful to Tomatoh Corporation and its employees for their gracious assistance and for the use of land and resources to conduct our study. We would also like to express appreciation to Natsumi Iizuka and other staff members at the Experimental Farm of the Field Science Center for Northern Biosphere at Hokkaido University for assisting with field research. We would also like to thank Shotaro Kuwabara at Gifu University for providing taxonomic identification of plant species at the study site. This study was funded by the Energy Biosciences Institute.

Ancillary