N₂O emissions during the freezing and thawing periods from six fields in a livestock farm, southern Hokkaido, Japan

Abstract

In many countries, high nitrous oxide (N₂O) emissions have been observed during soil freezing and thawing periods. Quantification of those emissions is crucial to evaluate annual N₂O emissions. For this study, we measured N₂O and nitric oxide (NO) fluxes along with soil N₂O concentrations at a corn (Zea mays L.) field and five grasslands (Phalaris arundinacea L. and Phleum pratense L.) during a winter-spring period in southern Hokkaido, Japan. We also measured denitrification activities of the soils from those sites. During the observation period, the soils froze to a maximum depth of 370 mm under saturated conditions and the lowest soil temperature at a 50 mm depth was −4.5°C. After 6 March 2005, daily air temperature rose above 0°C, but the soil temperature remained approximately 0°C for about two weeks. These two weeks were defined as the “transition period,” while the periods before and after the transition period were defined as the “freezing” and the “thawing” periods, respectively. During the freezing and transition periods, N₂O concentration increased in the frozen soils relative to the unfrozen soils, and the highest values were observed in the frozen soils during the transition period. During the thawing period, the N₂O concentration in the soils decreased. N₂O emissions were much higher during the thawing period than during the freezing and transition periods, and remarkably higher N₂O emissions were observed at the corn site compared to those at the grassland sites. NO emissions were also observed during the thawing period but at much lower levels than N₂O emissions at all the sites. N₂O-N/NO-N ratio exceeded one at all the sites during the entire period, indicating N₂O production through denitrification. At the corn site, denitrification activity was much lower and N₂O/(N₂O+N₂) was much higher than at the grasslands. The result indicated that high N₂O emissions at the corn site were caused by complementary processes: (1) high accumulated N₂O through denitrification in the frozen soil during the freezing and transition periods, and (2) low N₂O reduction rate during the thawing period.

Keywords:

• corn field
• denitrification activity
• freezing and thawing periods
• grassland
• nitrous oxide.

INTRODUCTION

Nitrous oxide (N₂O) is a major greenhouse gas that is also responsible for destruction of stratospheric ozone (Crutzen 1970). Global N₂O emissions are estimated at 17.7 Tg N y⁻¹ and of those, an estimated 6.7 Tg N y⁻¹ are from anthropogenic sources (Denman et al. 2007). Agricultural soils are the main source of N₂O emissions, at an estimated 42% of total anthropogenic N₂O emissions (Denman et al. 2007).

In soils, N₂O is produced mainly by nitrification and denitrification. These biological processes are influenced by soil environmental factors, such as moisture condition, oxygen (O₂) level, soil temperature, nitrogen (N) availability, organic matter content, and pH (Mosier 1998; Sahrawat and Keeney 1986). The soil freeze-thaw cycle has also been recognized as having an influence on soil N₂O emissions, since Bremner et al. (1980) observed N₂O emissions during a winter period. Kaiser and Ruser (2000) summarized annual N₂O emission data and winter N₂O losses observed on arable soils in Germany from 1995 to 1999. The annual N₂O losses ranged from 0.53 to 16.8 kg N ha⁻¹ and N₂O losses during the winter ranged from 7% to 89%. We also summarized the reported data and the N₂O losses during the winter ranged from 0% to 93% (Table 1). These results indicate a high concentration of N₂O emissions during the winter period and uncertainty of N₂O emissions.

Table 1. Nitrous oxide (N₂O) emissions during the winter and cropping periods and N₂O losses during the winter period for the investigated experiments

Reference	Country	N2O emissions during winter period (kg N ha^−1)	Annual N2O emissions (kg N ha^−1 y^−1)	Emission during winter (%)
Grassland
Wagner-Riddle et al. 1997	Canada	−0.10–0.090	−0.070–0.12	−75–143
Kaiser et al. 1998	Germany	0.57–0.92	1.3–2.4	39–53
Goossens et al. 2001	Belgium	0.59–2.4	2.6–32	7–23
Regina et al. 2004	Finland	0.060–6.7	2.6–9.9	1–68
Syvasalo et al. 2004	Finland	0.54–2.2	1.5–3.9	33–59
Katayanagi et al. 2008 ^†	Japan	0.35–2.4	7.3–82	2–33
Cropland
Goossens et al. 2001	Belgium	0.17–0.59	0.78–1.5	11–76
Wagner-Riddle et al. 2007	Canada	0.42–2.9	0.89–3.3	24–88
Regina et al. 2004	Finland	3.3–19	6.2–24	53–79
Syvasalo et al. 2004	Finland	1.6–7.0	3.7–7.5	42–93
Flessa et al. 1995	Germany	1.1–4.6	9.4–17	11–46
Röver et al. 1998	Germany	1.1–2.3	1.8–3.5	67–78
Kaiser et al. 1998	Germany	1.7–2.0	3.3–6.2	30–61
Teepe et al. 2000	Germany	2.8	4.8	58
Ruser et al. 2001	Germany	0.73–3.3	1.3–6.9	40–58
Sehy et al. 2003	Germany	0.30–1.7	3.1–10	10–20
Koga et al. 2004	Japan	0.00–2.1	0.09–2.4	0–89
Bremner et al. 1980	United States	0.030–0.44	0.34–2.0	8–26
Katayanagi et al. 2008 ^†	Japan	0.010–0.77	1.1–12	1–24
^†The winter period emission and the emission rate during the winter are unpublished data.

It has been assumed that N₂O would be produced in an unfrozen soil layer beneath a frozen layer and released from the soil surface during a thawing period once the frozen soil cover has disappeared in situ (Bremner et al. 1980; Burton and Beauchamp 1994; Kaiser et al. 1998). But based on a soil column experiment, Teepe et al. (2001) and Wagner-Riddle et al. (2007) hypothesized that N₂O was produced and trapped in the frozen soil during a freezing period and then released during a thawing period. Recently, Yanai et al. (2011) found an increase in N₂O concentration with a decrease in O₂ concentration in frozen soil during the freezing and thawing periods.

To investigate the N₂O production process during the freezing and thawing conditions, soil incubation experiments have been conducted by many researchers. Yanai et al. (2004b) evaluated the nitrification potential of frozen soils and reported that soil freeze-thaw cycles did not inhibit the nitrification potential of soils that experienced a smaller decrease in microbial biomass following the freeze-thaw cycles. The source of ammonium (), which is the base of nitrification, may be soil organic matter (McGarity 1962; Groffman and Tiedje 1989; Neilson et al. 2001) or destructed microbes or substrates from the destructed microbes (McGarity 1962; Groffman and Tiedje 1989; Schimel and Clein 1996; Yanai et al. 2004a). Christensen and Christensen (1991) evaluated soil carbon availability by denitrifiers, using an acetylene block method under freeze-thaw conditions. Müller et al. (2002) and Ludwig et al. (2004) reported that high N₂O emissions were caused by the enrichment of nitrate () through ¹⁵ addition experiments. Based upon an experiment with a ¹⁵N tracer, Wagner-Riddle et al. (2008) concluded that denitrification activity during a thawing period was the main process of N₂O production. Mørkved et al. (2006) also conducted a ¹⁵ addition experiment and reported that denitrification was the main N₂O source in freeze-thaw-affected soil and that soluble carbon could play a significant role in the freeze-thaw cycle. Öquist et al. (2004) observed higher N₂O emissions after freezing at −4°C in soils with a high moisture content compared to soils with a low moisture content.

Despite significant research on N₂O emissions from soil, there are only a few studies that have measured both N₂O concentration in soil and N₂O flux at a field and evaluated denitrification potentials as well. The objective of this study was to determine the cause of high N₂O emissions from soils during a thawing period. We measured N₂O fluxes and N₂O concentrations in soils at a livestock farm in southern Hokkaido, Japan. The study was conducted after a cropping period in 2004, because high N₂O emissions were observed from 2000 to 2003 (Table 1). Furthermore, an incubation experiment was conducted just after soil thawing to measure denitrification activities in soils.

Table 2. Soil properties at 0–100 mm depth at the study site in Shizunai Experimental Livestock Farm

Site	Land-use type^†	Total carbon^‡ (g kg^−1)	Total N^‡ (g kg^−1)	Bulk density^‡ (Mg m^−3)	D/D0 ^§
C2	C	60.8	4.5	0.60	0.0635
CG	G	41.8	3.5	0.78	0.00684
G2s	G	32.2	3.0	0.81	0.00113
G2c	G	42.4	3.4	0.68	0.00429
G2n	G	49.7	4.3	0.63	0.00132
G3	G	75.2	5.4	0.65	0.00143
^†Cornfield (Zea mays L.), C; Grassland (Phalaris arundinacea L. and/or Phleum pratense L.), G; Nitrogen, N.
^‡The values were reported by Katayanagi et al. (2008)
^§Relative gas diffusion coefficients (D/D0) of soil core samples were measured using the method proposed by Osozawa (1998). The data was measured using soils collected on 19 April 2005.

Table 3. Slurry and manure application after the cropping season and nitrogen (N) surplus in 2004 at the study site in Shizunai Experimental Livestock Farm. Slurry was applied on 19 October 2004 and manure was applied on 1 November 2004

	Slurry	Manure	Slurry + Manure	Nitrogen Surplus^†
Site	(kg N ha^−1)	(kg N ha^−1)	(kg N ha^−1)	(kg N ha^−1)
C2	9	123	132	16
CG	0	38	38	99
G2s	0	0	0	24
G2c	0	0	0	24
G2n	0	0	0	24
G3	0	43	43	−12
^†Katayanagi et al. (2008).

MATERIALS AND METHODS

Study area

The study was conducted at the 458-ha Shizunai Experimental Livestock Farm of the Field Science Center for the Northern Biosphere, Hokkaido University, Japan (42°25′54N, 142°28′54E). The site is an experimental station, but also a working production facility where young animals are reared to maturity (e.g., for beef production and preserving native species of Hokkaido horses) and crops (primarily corn (Zea mays L.) and grass (Phalaris arundinacea L. and/or Phleum pratense L.)) are grown to support the animals. The farm is located in the watershed of the Kepau River, which flows through the farm from an upstream forested area. The average annual mean temperature at the study site is 7.9°C, the minimum monthly temperature is −8.1°C in February, and the maximum monthly temperature is 23.6°C in August. The average annual precipitation is 1365 mm.

Measurements were taken at a corn field (C₂), a grassland that was recently converted from a corn field in 2003 (CG), and four grasslands (G_2s, G_2c, G_2n and G₃) which have been used as a meadow for more than three years. Three grasslands G_2s, G_2c, and G_2n were divided from G₂ and named G_2s for the southern site, G_2c for the center site, and G_2n for the northern site (see Supporting Information, Fig. S1). These three grasslands were adjacent to one another and were all managed in the same way. The soil types of the research sites (Katayanagi et al. 2008) are Vitric Andosols at CG, G_2s, G_2c, and G_2n and Histosols at C₂ and G₃ (FAO 1988) (Table 2). The dominant vegetation type is corn (Zea mays L.) at C₂, reed canary grass (Phalaris arundinacea L.) at G_2s, G_2c, G_2n and G₃, and timothy-grass (Phleum pratense L.) at CG. Grass was harvested twice annually and corn was harvested once a year.

After the cropping season in 2004, manure was applied to CG, G₃, and C₂ on 27 October 2004 and slurry was applied to C₂ on 18 October 2004 (Table 3). N surpluses at G₂, G₃, CG, and C₂ calculated by Katayanagi et al. (2008) are also provided in Table 3.

Temperature, precipitation, and snow depth

Daily precipitation measurements from the farm's Sasayama Weather Station (42°25′54N, 142°28′54E) (Fig. 1) were used for the analysis of gas data. Daily air temperature was also observed at the station but there were many missing data due to equipment problems. Snow depth was not observed at the station. Therefore, the data measured at the Shizunai Weather Station (42°20′36N, 142°21′42 E) were used instead. For calculation of gas fluxes, air temperature inside a static chamber was measured with a digital thermometer during the gas measurement. Soil temperature at a depth of 50 mm was measured continuously during the study period at hourly intervals using a small waterproof temperature logger (TR-52, T&D, Nagano, Japan) for one replicate per field. Depth of soil freezing was measured using an acrylic tube that was filled with a 0.03% methylene blue solution. A polyvinyl chloride (PVC) pipe was used to protect the acrylic tube. Petroleum jelly was spread thinly on the acrylic tube to prevent it from sticking to the PVC pipe.

Figure 1. Temporal variability of precipitation observed at the Sasayama Weather Station in the Shizunai Experimental Livestock Farm, and air temperature and snow depth observed at the Shizunai Experimental Station, from December 2004 to April 2005.

Soil sampling

Disturbed soil samples for measurements of soil chemical properties were collected from a depth of 0−50 mm. The sampling was conducted with three to six replicates per site before and after freezing (08 or 18 December 2004 and 19 April 2005, respectively). Only a single sample was collected on 31 March 2005 when the soil started to thaw because only surface soil from approximately 0 to 10 mm depth had thawed, and it was difficult to collect a large volume of soil samples at that time. The collected soil samples were preserved at 4°C and analyzed within a week of sampling.

Analysis of soil chemical properties

The disturbed soil samples from each site were extracted in deionized water (at a 1:5 ratio of soil to water) in bottles and filtered through a 0.2-μm membrane filter using suction. Nitrate-N (-N), -N and dissolved organic carbon (DOC) concentrations in the extracted solution were determined using ion-exchange chromatography (QIC analyzer, Dionex, Sunnyvale, CA, USA), colorimetry with indophenol-blue, and total organic carbon analysis (TOC-5000 A, SHIMADZU, Kyoto, Japan), respectively. Soil pH was measured with a glass electrode (F-22, Horiba, Kyoto, Japan) using infiltrated samples remaining in the bottle after the filtration.

Measurement of denitrification enzyme activity

Using the disturbed soil samples collected on 31 March 2005, denitrification activity was measured by the acetylene block technique proposed by Tiedje (1994). A total of 15 g of fresh soil sample was thoroughly mixed with a 15-mL chloramphenicol solution (1 g L⁻¹) in a 200-mL serum bottle to prevent enzyme production (Tiedje 1994; Lowrance 1992; Murray and Knowles 1999). To determine the denitrification activity under field conditions, available carbon and nitrogen were not added to the soil samples. The serum bottle was evacuated and flushed four times with N₂ to ensure anaerobic conditions. For this study, the incubations were conducted using non-acetylene and acetylene treatments to determine the denitrification activities of N₂O production and N₂+N₂O production, respectively. N₂O production was estimated using N₂O emissions from the non-acetylene-treated samples and N₂+N₂O production was estimated using N₂O emissions from the acetylene-treated samples, because acetylene inhibits N₂O reduction to N₂. Acetylene gas was added to a final concentration of 10% (10 kPa) in the headspace for the acetylene treatment. Incubations were conducted in triplicate at 5°C, which was the soil temperature at the time of collection from the field. At 1 and 2 h after the incubation start time, 15 mL of headspace gas was sampled from each bottle into an evacuated glass vial using a 25-mL syringe. N₂O concentration of the gas samples, which were kept in evacuated glass vials, was analyzed by gas chromatography with an electron capture detector (GC 14B, Shimadzu, Kyoto, Japan) at 375°C. PorapackQ filled in a stainless steel column (2 m length and 3 mm inner diameter) was used as stationary compound and 5% methane in argon was used as the carrier gas. The minimum detectable concentration of N₂O was 0.007 ppmv, as determined from the deviation obtained by repeatedly analyzing the concentration of N₂O standard gas (atmospheric level).

Measurement of N₂O and NO fluxes

Measurements of N₂O and nitric oxide (NO) were conducted once a month in December 2004 and February 2005 and once a week in March and April 2005. Each flux measurement was replicated at four locations per site for G₃, CG, and C₂ and at six locations per site for G_2s, G_2c, and G_2n.

The measurements were conducted using a closed-chamber technique (Rolston 1986). A cylindrical stainless steel chamber (diameter 300 mm, height 350 mm) was used for measurements. Detailed information about this chamber design is reported in a previous study (Katayanagi et al. 2008). The deepest snow depth on the monitoring days was only 150 mm, and the chambers were placed on the snow during the freezing period. During the soil freezing conditions, gas samples were collected at 0 and either 40 or 60 min after closing the lid; during the other periods, samples were collected at 0 and 20 min after closing the lid. The chamber height from the snow surface to the top of the chamber was measured for four replicates to calculate gas flux. A 500-mL gas sample was taken using a 50-mL syringe and injected into a 500-mL Tedlar® bag. The bags were then brought to the laboratory and 20 mL of each gas sample was immediately transferred into a 10-mL evacuated glass vial, using a 25-mL syringe, for N₂O analysis.

N₂O concentrations of the samples kept in evacuated glass vials were analyzed using the gas chromatograph as described earlier. The gas samples kept in Tedlar® bags were analyzed by a chemoluminescence nitrogen oxide analyzer (MODEL-265 P, Kimoto Electric, Osaka, Japan) within 24 h of sampling to measure NO concentrations in the samples.

Measurement of soil N₂O concentration

The soil N₂O concentration was measured concurrently with the gas flux measurement. Stainless steel tubes (inner diameter 10 mm and outer diameter 12 mm; the lengths of the tubes were the target depth plus 100 mm; there was no side hole on the pipe) were installed in the soil at 100, 200, 300, 400, 500, and 600 mm depths with three replications each. A stainless steel stick (20 mm longer than the pipe) was inserted in the pipe during installation, in order to avoid clogging the pipe with soil and to provide air space (ca. 2 mL) at the bottom of the pipe. The stick was removed after installation. A three-way cock was connected to the aboveground part of an installed pipe using a vinyl tube (100 mm), a straight connector and a silicone tube (50 mm); 50 mL of air was drained from inside the pipe through the cock just after the installation. The pipe was left in place throughout the freezing and thawing periods, and the cock was closed and covered using a cap made of polyethylene terephthalate to avoid freezing the cock from snow and ice exposure. A 50-mL soil gas sample was collected from each pipe. The gas samples were collected slowly to prevent contamination by gas from non-target depths. However, the volumes of the pipes were small and the gas under the target depth would be collected together with the gas at the target depth. The samples that were collected from the same depth were placed into a Tedlar® bag together. A 300-mL air sample from the soil surface was collected near the pipes at the same time. N₂O was analyzed using gas chromatography, as described earlier. Because of those operations, we had only one replicate sample per depth at each site; therefore, standard variations could not be reported in this paper.

Calculation of N₂O and NO fluxes

N₂O and NO fluxes were calculated by a linear regression using the following equation:

where F is the flux (mg N m⁻² h⁻¹), ρ is the gas density (ρN₂O-N = 1.26 × 10⁶ and ρNO−N = 0.63 × 10⁶ mg N m⁻³), V is the volume of the chamber (m³), A is the area of the chamber (m²), Δc/Δt is the ratio of change in the gas concentration inside the chamber (10⁻⁶ m³ m⁻³ h⁻¹), T is the air temperature inside the chamber (°C) and P is the air pressure (mm Hg). For this study, we defined positive fluxes as emissions and negative fluxes as uptake by soil or snow and calculated N₂O-N/NO-N when both N₂O and NO fluxes were positive during the observation period.

Statistical analysis

Statistical analyses were performed with Kyplot 5.0 (KyensLab Inc. Tama, Tokyo, Japan). The -N, -N, and DOC concentrations, measured before freezing and at the end of thawing, were compared using paired t-tests. Differences between N₂O and N₂+N₂O fluxes, which were the results of denitrification enzyme activity measurement, were also compared using paired t-tests and differences of N₂O and N₂+N₂O fluxes among each site were compared by Tukey-Kramer test. Correlation analyses between the -N, -N, and DOC concentrations and the manure plus slurry N application and surplus N were performed using the same software.

RESULTS

Temporal variability of air temperature, soil temperature, precipitation, snow depth and frozen soil layers

The daily air temperature decreased from December to February and increased from March to April (Fig. 1). Daily mean soil temperature at 50 mm depth dropped below 0°C from 20 December 2004 and remained below zero until around 06 March 2005 (Fig. 2). The lowest recorded daily soil temperatures were −4.5, −1.8, −2.3, −4.7, −3.2, and −3.3°C at C₂, CG, G_2s, G_2c, G_2n, and G₃, respectively. After 06 March 2005, the soil temperature did not increase and remained approximately 0°C for about two weeks, even though air temperature rose above 0°C (Fig. 2a). We defined these two weeks as the “transition” period while the periods before and after were defined as the “freezing” and “thawing” periods, respectively. On this definition, the freezing period started from 20 December, and the period prior to 20 December 2004 was defined as the “unfreezing” period.

Figure 2. Temporal variability of (a) soil temperature at 50 mm depth, (b) nitrous oxide (N₂O) flux, (c) nitric oxide (NO) flux, (d) N₂O-N/NO-N ratio, and (e) N₂O concentration in soil with soil freezing depths at one cornfield (C₂) and three grasslands (CG [converted from a cornfield in 2003], G₂ [including G_2s, G_2c and G_2n], and G₃) from December 2004 to April 2005. The lines in e) indicate the edge of the frozen soil layers and the soil between the lines froze.

The maximum snow depth during the whole period was 250 mm (Fig. 1) and soil freezing was observed from 24 January 2005. The depth of the frozen soil layer got deeper after soil freezing started (Fig. 2e); the maximum depths of the frozen layers were observed in March and were 370, 230, 220, 220, 340, and 350 mm at C₂, CG, G_2s, G_2c, G_2n, and G₃, respectively. The frozen soils were very hard during the freezing period due to the saturated soil conditions caused by high soil water content before freezing (water-filled pore space was 72, 92, 79, 81, 82 and 79% at C₂, CG, G_2s, G_2c, G_2n, and G₃, respectively) and also by water supply from rainfall and melted snow when air temperature increased above 0°C. Soil thawing was observed from 24 March 2005 at both the soil surface and the bottom of the frozen layers (Fig. 2e). Frozen soil layers were not observed after 19 April 2005 at all sites.

N₂O and NO fluxes

N₂O emissions during the thawing period were much higher than the emissions during other periods (Fig. 2b). During the thawing period, fluxes at all sites showed three patterns: (i) remarkably high N₂O emissions were observed at C₂ and CG; (ii) N₂O emissions increased through the transition and thawing periods at G_2s, G_2c, and G_3, and the fluxes at those sites were much lower than fluxes at C₂ and CG; and (iii) N₂O peaks were observed during the transition and thawing periods and high N₂O uptake (negative N₂O flux) was also observed at G_2n. N₂O emissions were also observed during the unfreezing and freezing periods at all sites and were higher at C₂ than at the other sites (Fig. 2b), but the emissions were lower during the freezing period than during the thawing period.

During the study period, NO emissions at all sites showed a similar temporal variation: low NO emissions were observed at the beginning of the transition period and higher NO emissions were observed during the latter part of the transition and thawing periods (Fig. 2c). NO emissions were also observed at all sites during the unfreezing and freezing periods as well as during the thawing period (Fig. 2c). However, while NO emissions were observed during the whole observation period, the emission level was much lower than N₂O (Fig. 2b and c).

N₂O-N/NO-N values during the unfreezing and freezing periods were lower than during the thawing period (Fig. 2d). The values were distributed from 2.0 to 13.6 except at C₂ during the unfreezing period. The highest value at C₂ during the unfreezing was 69.8. During the thawing period, higher N₂O-N/NO-N was observed at C₂ than at the other sites and the highest value at C₂ was 151. The values at the other sites were distributed from 8.5 to 63.3 during the thawing period. The temporal variation of N₂O-N/NO-N was similar to that of N₂O at C₂, CG, G_2s, and G_2c (Fig. 2b, d). At G_2n and G₃, there was no apparent pattern due to few samples and negative N₂O emissions.

Soil N₂O concentrations

During the freezing and transition periods, N₂O concentrations increased in the frozen soils and were higher during those periods than during the unfreezing and thawing periods at all sites except G_2c, where gas samples in the frozen soil could not be collected (Fig. 2e). The highest N₂O concentrations were observed in the frozen soils during the transition period at all the sites except G_2c. N₂O concentrations also increased in the unfrozen soil layers below the frozen layers at C₂, CG, G_2s and G_2n, but, the concentrations in the unfrozen soils were lower than those in the frozen soils (Fig. 2e). At G_2c and G₃, the N₂O concentration in the unfrozen soil layers just below the frozen soil layers and at the edge of the frozen soil layers was lower than that of the atmosphere (i.e., approximately 0.3 ppmv).

After the start of soil thawing, N₂O concentrations in the thawed and unfrozen layers decreased at C₂, CG, G_2s and G_2n. At G_2c and G₃, the N₂O concentrations increased to and above the level in the atmosphere (Fig. 2e).

Soil -N, -N and DOC concentrations during soil freezing

Figure 3 shows the soil -N, -N and DOC concentrations during the unfrozen period measured on 08 or 18 December 2004, at the beginning of the thawing period measured on 31 March 2005, and at the end of thawing period measured on 19 April 2005.

Figure 3. Ammonium-nitrogen (-N), nitrate-nitrogen (-N), and dissolved organic carbon (DOC) concentrations in soil at a 0−50 mm depth at the research sites in the Shizunai Experimental Livestock Farm. Measurements were carried out on 8th (C₂, CG and G₃) or 18^th (G_2s, G_2c, and G_2n) December 2004 before soil freezing, on 31^st March 2005 just after the start of soil thawing, and on 19th April 2005 at the end of the thawing period. Paired t-test was used to test for statistically significant differences between the data observed on 8^th Dec 2004 and 19^th April 2005 for each site. Levels of significance were as follows: *<0.05, **<0.01, ***<0.001.

Soil -N values were remarkably high at the beginning of the thawing period compared with the values during the unfreezing period and at the end of the thawing period for all sites (Fig. 3); however, these data couldn’t be statistically compared with the other data. The value at the end of the thawing period (p = 0.004) was significantly higher than before freezing at C₂, while there was not a significant difference in the values at the other sites (Fig. 3). The -N concentration did not show any correlation with slurry + manure N and N surplus, as shown in table 3.

Soil -N values at the beginning of the thawing period were higher than those before the freezing period at C₂, CG, and G_2n; the values at G_2s, G_2c and G₃ did not show appreciable differences (Fig. 3) but these data couldn’t be statistically compared with the other data. The highest -N concentrations were observed at the end of the thawing period for all sites, and the values at C₂, CG, G_2n and G₃ were significantly higher than those before the freezing period (Fig. 3; the p values were 0.0005, 0.0384, 0.0499, and 0.0034, respectively). The -N concentration showed a positive correlation with the application rates of slurry + manure N (Table 3) for all sampling dates (r = 0.9031 and p = 0.0136 on 08 and 18 December 2004, r = 0.9272 and p = 0.0078 on 31 March 2005, and r = 0.9035 and p = 0.0135 on 19 April 2005, respectively).

The soil DOC concentrations measured at the beginning of the thawing period were higher than those measured during the unfreezing period and at the end of the thawing period for all sites (Fig. 3). There was no significant difference between the values for the freezing period and those for the end of the thawing period. The DOC concentrations did not show any correlation with slurry + manure N and N surplus.

Denitrification activities

N₂O production was significantly lower at C₂ and CG, and higher at G₃ (p < 0.05); N₂O+N₂ production was significantly lower at C₂ and CG, and higher at G_2c and G₃ compared with the other sites (Table 4). N₂O+N₂ production was significantly higher than N₂O production at G_2c and G₃ and lower than N₂O production at CG (p < 0.05) (Tables 4). N₂O + N₂ production was also higher than N₂O production at G_2s and G_2n, but it was not significant because of the high standard deviation (Table 4). The ratio of N₂O/(N₂O+N₂) was lower at G_2s, G_2c, G_2n and G₃ than at C₂ and CG (Table 4). A regression analysis of N₂O and N₂O + N₂ productions and the amount of manure + slurry N and surplus N was also conducted, but the results did not show any relationship.

Table 4. Nitrous oxide (N₂O) and N₂O + dinitrogen (N₂) fluxes and N₂O/(N₂O+N₂) ratio measured by denitrification enzyme activity in soils collected from the study site at Shizunai Experimental Livestock Farm just after the soil thawing on 31 March 2005

	N2O flux^† (µg N kg^−1 h^−1)		N2O+N2 flux^† (µg N kg^−1 h^−1)		N2O/(N2O+N2)	P values^‡ N2O × N2O+N^‡
Site	n	Mean ± S.D.	n	Mean ± S.D.
C2	3	0.478 ± 0.122^a	3	0.519 ± 0.107^a	0.922	0.688^ns
CG	3	0.466 ± 0.106^a	3	0.229 ± 0.0485^a	2.03	0.0244*
G2s	3	1.60 ± 0.899^a,b	3	14.0 ± 11.4^a,b	0.114	0.132^ns
G2c	2	3.84 ± 1.53^a,b	2	73.6 ± 11.9^c	0.0522	0.0144*
G2n	3	5.45 ± 4.31^a,b	3	19.3 ± 32.3^a,b	0.282	0.503^ns
G3	3	6.63 ± 0.720^b	3	51.3 ± 15.0^b,c	0.129	0.00681**
^†The values are means with standard deviation (S.D.). Different letters indicate significant differences among the means for each site by Tukey-Kramer test at p < 0.05. There were no significant differences among N2O/(N2O+N2) at all sites.
^‡The difference between the means of N2O and N2O+N2 was tested by t-test for each site. Levels of significance were as follows: *<0.05; **<0.01; ***< 0.001; ns, not significant.

DISCUSSION

N₂O production and accumulation in frozen soil during the freezing and transition periods

During the freezing period, production and accumulation of N₂O in the frozen soils resulted from inhibition of gas diffusion by the frozen soil lids and from N₂O reduction to N₂. Higher N₂O concentrations in the soils, relative to the atmosphere (Fig. 2), indicated N₂O accumulation in the soils and disturbance of gas exchange by the frozen saturated soil, which functioned as a “lid” for all the sites. In the soil layers, the highest N₂O concentration was observed in frozen rather than unfrozen soils (Fig. 2); this indicated that higher N₂O production and/or lower N₂O reduction occurred in the frozen soils compared to unfrozen soils and that the produced or remaining N₂O in the frozen soils accumulated there. The results show that N₂O production occurs in the frozen soil in situ as well as in a laboratory (Christensen and Christensen 1991; Müller et al. 2002; Yanai et al. 2004 a and b; Öquist et al. 2004; Öquist et al. 2007), and support the hypothesis of Teepe et al. (2001) and Wagner-Riddle et al. (2007) that N₂O is produced and trapped in frozen soils during a freezing period.

Denitrification was identified as an important N₂O production process in the frozen soils, as demonstrated by the higher-than-one ratio of N₂O-N/NO-N during the freezing period (Fig. 2). Lipschultz et al. (1981) reported that the N₂O-N/NO-N ratio is considered an indicator of the contribution of nitrification and denitrification; when N₂O-N/NO-N is less than one, the main process of N₂O production is considered to be nitrification. Alternatively, when N₂O-N/NO-N is larger than 100, the main process of N₂O production is considered to be denitrification. The lid mentioned before must disturb O₂ diffusion from the atmosphere into the soils, and O₂ concentration in the soils must be low during the freezing period. Under such conditions, nitrification must be restrained by deficiency of O₂ as an electron acceptor, and N₂O production by denitrification would be accelerated. Yanai et al. (2011) observed an increase in N₂O concentration corresponding with a decrease in O₂ concentration in frozen soil during freezing and thawing periods, and Wrage et al. (2001) reported that low O₂ concentration inhibited N₂O reduction to N₂.

Nitrate, which is a substrate of denitrification, was supplied from slurry and manure, because application rates of slurry + manure N (Table 3) showed a positive correlation with -N on all the sampling dates. -N concentrations were higher in the surface soils just after thawing than before freezing (Fig. 3); this indicates that , which is a substrate of nitrification, was supplied through mineralization during the freezing period and high -N was caused by the nitrification activity. The source of the was not determined in this study. However, it would likely be the applied slurry and manure, soil organic matter (McGarity 1962; Groffman and Tiedje 1989; Neilson et al. 2001), destructed microbes or substrates from destructed microbes (McGarity 1962; Groffman and Tiedje 1989; Schimel and Clein 1996; Yanai et al. 2004 a). Yanai et al. (2004 b) evaluated the nitrification potential of frozen soils and reported that a decrease in microbial biomass after freeze-thaw cycles was less substantial and that the nitrification potential was not inhibited by the soil freeze-thaw cycles. Neilson et al. (2001) reported that soil freezing did not disturb nitrification. The nitrification process would also produce N₂O, but Mørkved et al. (2006) observed a low contribution of nitrification: N₂O produced through nitrification was −0.7% to 4.35% of N₂O emissions during the thawing period.

N₂O production and emission during the thawing period

High N₂O emissions during the thawing period were caused by diffusion of accumulated N₂O after disappearance of the frozen soil lids, and a low N₂O reduction rate during the thawing period. At C₂ and CG, higher N₂O concentrations in the surface soils were observed during the freezing and transition periods, and high N₂O emissions were observed during the thawing period after disappearance of the frozen soil lid (Fig. 2). However, the denitrification activity was much lower and N₂O/(N₂O+N₂) was much higher at those sites compared with other sites (Table 4). From these results, it is evident that the accumulated N₂O was not reduced to N₂ by denitrification activity, and high N₂O emissions were caused by the high N₂O concentration in soils and the high gas diffusion coefficients (Table 2) at the sites. Compared with the N₂O emissions at C₂ and CG, N₂O emissions at G_2s and G₃ were much lower, even though N₂O concentrations in the surface soils were high (Fig. 2). The low emissions must be attributed to high denitrification activities, which reduced N₂O to N₂ and resulted in low N₂O/(N₂O+N₂) (Table 4). The low gas diffusion coefficient at the sites (Table 2) would also accelerate N₂O reduction to N₂ before the emission. The N₂O concentration in the frozen soil at G_2c was not clear, but the low N₂O emissions during the thawing period would be attributed to the same reason as G_2s and G₃; the N₂O profile (Fig. 2), the low gas diffusion coefficient (Table 2), the high denitrification activity and the low N₂O/(N₂O+N₂) (Table 4) at G₂ were similar to the values at G₃. At G_2n, low N₂O emissions as well as N₂O uptake were observed during the thawing period (Fig. 2b). This would be caused by a relatively low N₂O concentration in the frozen soil (Fig. 2) and high denitrification activity (Table 4). In a soil column experiment, Teepe et al. (2001) and Wagner-Riddle et al. (2007) showed that N₂O trapped in frozen soil was released during a thawing period, which is consistent with our findings.

N₂O produced through nitrification and denitrification during the thawing period would also contribute to the N₂O emissions during the period, but the contribution must be low at the sites. The N₂O-N/NO-N values above one resulted from an increase in N₂O emissions during the thawing period (Fig. 2) and indicated a low contribution of nitrification. The contribution of denitrification to N₂O emissions was considered from the results, but high denitrification activity decreased N₂O to N₂ in the sites. Therefore, the contribution of N₂O production through denitrification during the thawing period would be low. Nitrification and the diffusion of accumulated N₂O was the main source of N₂O emissions during the period.

CONCLUSION

N₂O produced through denitrification activity during the freezing and transition periods accumulated in the frozen soils; this resulted from inhibition of N₂O diffusion by frozen soil lids and from N₂O reduction to N₂. N₂O emissions during the thawing period were caused by diffusion of the accumulated N₂O in the surface soil. During the thawing period, the contribution of N₂O produced by nitrification and denitrification to the emissions was low at the study sites.

ACKNOWLEDGMENTS

We thank Dr. H. Hata and the staff of the Shizunai Experimental Livestock Farm, Field Science Center, Hokkaido University, for their assistance. We thank Dr. M. Shimizu of the Soil Science Laboratory and Miss Y. Usui of the New Energy and Industrial Technology Development Organization for their assistance with our sampling. We also thank Dr. Y. Yanai of the National Agriculture and Food Research Organization (NARO), National Institute of Vegetable and Tea Science (NIVTS), for his helpful comments.

This study was partially supported by a special research grant provided by the project entitled “Establishment of good practices to mitigate greenhouse gas emissions from Japanese grasslands” funded by the Japan Racing Horse Association.

SUPPLEMENTARY MATERIAL

The supplementary material for this article is available online from: http://dx.doi.org/10.1080/00380768.2012.670810.