No significant difference in N₂O emission, fertilizer-induced N₂O emission factor and CH₄ absorption between anaerobically digested cattle slurry and chemical fertilizer applied timothy (Phleum pratense L.) sward in central Hokkaido, Japan

Abstract

Nitrous oxide (N₂O) and methane (CH₄) fluxes from a fertilized timothy (Phleum pratense L.) sward on the northern island of Japan were measured over 2 years using a randomized block design in the field. The objectives of the present study were to obtain annual N₂O and CH₄ emission rates and to elucidate the effect of the applied material (control [no nitrogen], anaerobically digested cattle slurry [ADCS] or chemical fertilizer [CF]) and the application season (autumn or spring) on the annual N₂O emission, fertilizer-induced N₂O emission factor (EF) and the annual CH₄ absorption. Ammonium sulfate was applied to the CF plots at the same application rate of NH₄-N to the ADCS plots. A three-way ANOVA was used to examine the significance of the factors (the applied material, the application season and the year). The ANOVA for the annual N₂O emission rates showed a significant effect with regard to the applied material (P = 0.042). The annual N₂O emission rate from the control plots (0.398 kg N₂O-N ha⁻¹ year⁻¹) was significantly lower than that from the ADCS plots (0.708 kg N₂O-N ha⁻¹ year⁻¹) and the CF plots (0.636 kg N₂O-N ha⁻¹ year⁻¹). There was no significant difference in the annual N₂O emission rate between the ADCS and CF plots. The ANOVA for the EFs showed insignificance of all factors (P > 0.05). The total mean ± standard error of the EFs (fertilizer-induced N₂O-N emission/total applied N) was 0.0024 ± 0.0007 (kg N₂O-N [kg N]⁻¹), which is similar to the reported EF (0.0032 ± 0.0013) for well-drained uplands in Japan. The CH₄ absorption rates differed significantly between years (P = 0.014). The CH₄ absorption rate in the first year (3.28 kg CH₄ ha⁻¹ year⁻¹) was higher than that in the second year (2.31 kg CH₄ ha⁻¹ year⁻¹), probably as a result of lower precipitation in the first year. In conclusion, under the same application rate of NH₄-N, differences in the applied materials (ADCS or CF) and the application season (autumn or spring) led to no significant differences in N₂O emission, fertilizer-induced N₂O EF and CH₄ absorption.

Key words:

• emission factor
• methane
• nitrous oxide
• slurry
• timothy

Introduction

More efficient use and appropriate management of livestock excreta is required in the context of nutrient utilization and demand for reducing environmental pollution (e.g. nitrogen leaching, emission of greenhouse and other pollutant gases). The management system of livestock excreta can be roughly classified into two groups, aerobic treatment (e.g. composting, purification in aeration system) and anaerobic treatment (e.g. digestion treatment).

Anaerobic digestion of livestock slurry by biogas plants provides a valuable energy source of methane (CH₄) and reduces the emission of CH₄ from stored slurry (Clemens et al. 2006). The product produced after the anaerobic digestion of cattle excreta is called “anaerobically digested cattle slurry (ADCS)”, and contains nutrient elements essential for plant and crop growth (Matsunaka et al. 2002). Moreover, the concentration of bad odor and pathogenic bacteria in ADCS is lower than that in undigested slurry (Pain et al. 1990; Saxena et al. 1989). In Japan, biogas plants using stock farming as feedstock have not been widely introduced because of the high cost of building such plants (Tsukahara et al. 2005). However, understanding the nutrient effect and environmental pollution followed by the application of ADCS to upland and grassland is important for promoting biogas plant systems.

Matsunaka et al. (2003) revealed that the nutrient effect of ADCS as N on the dry matter production of grass was similar to conventional slurry or chemical fertilizer, and that the effect depended on the net applied ammonium nitrogen (NH₄-N), which was the difference between the applied NH₄-N and the nitrogen loss through ammonia (NH₃) volatilization. A lysimeter experiment on timothy sward, conducted in central Hokkaido (northern island) Japan, showed that nitrate (NO₃-N) leaching associated with ADCS application in autumn tends to be higher than that in spring, owing to snow cover and snow melting, resulting in higher N-use efficiency of the ADCS in the spring application (Matsunaka et al. 2006). It is also well known that N application and soil environmental factors (e.g. temperature, moisture, pH, oxidation–reduction potential, availability of organic carbon, NH₄-N and NO₃-N) affect nitrous oxide (N₂O) and CH₄ fluxes from the soil (Le Mer and Roger 2001; Sahrawat and Keeney 1986). From these facts, it is suggested that N₂O and CH₄ fluxes may be affected by the applied material and the application season, even if the same level of NH₄-N is applied to the grassland soil. To date, these effects have not been examined.

In addition, only a few peer-reviewed papers investigating both N₂O emission and CH₄ absorption from Japanese grasslands over the year have been published (Mori et al. 2005, 2008). These studies were conducted in Nasu, Tochigi prefecture, central Japan, and reported annual N₂O emission and CH₄ absorption. However, these studies did not report the fertilizer-induced N₂O emission factor (EF). In Japan, most (92%) grassland for livestock production is located in Hokkaido (Statistics Bureau 2008). Investigations of annual N₂O emission rates and CH₄ absorption rates and EFs in Hokkaido grassland are important for the estimation of greenhouse gas emissions from Japanese grassland.

In Hokkaido, Japan, timothy swards for making silage and/or hay are commonly fertilized in spring (after snow melting) and in early summer (June, after first cutting), and harvested in early summer (June) and late summer (August–September). However, in farms producing ADCS, it is necessary to apply ADCS to the sward in autumn because the storage tank should be emptied before the snow season. The effects of the applied materials and the application season on N₂O emission and CH₄ absorption have not been demonstrated. In this paper, we report the results of a 2-year field experiment in central Hokkaido, Japan, using three materials (control, ADCS or chemical fertilizer [CF]) and two application seasons (autumn or spring). The objectives of the present study were to obtain annual N₂O and CH₄ emission rates and to examine the effect of the applied material and the application season on the annual N₂O and CH₄ emission and fertilizer-induced N₂O EF.

Materials and Methods

Site description

A field study was carried out in timothy (Phleum pratense L.) grassland established in 2002 at the Rakuno Gakuen University, central Hokkaido, Japan (43°04′N, 141°30′E, 61 m a.s.l.). The soil was gray upland soil (Aeric, Typic, Epiaquults). The Ap horizon was at a depth of 0–20 cm. The bulk density and texture were 0.81 Mg m⁻³ and loam (clay 14%, silt 43% and sand 43%), respectively. The pH (H₂O), total carbon content, total nitrogen content and cation exchange capacity were 6.3, 30 g kg⁻¹, 2.5 g kg⁻¹ and 39 cmol₍₊₎ kg⁻¹, respectively. The 5-year (2003–2007) averages of precipitation and air temperature were 928 mm year⁻¹ and 7.0°C, respectively (from the nearest weather station, 7 km north-east of the site [Japan Meteorological Agency 2008]).

Experimental design and field management

Experimental plots using a randomized block design were established in mid-October 2003, with three replicates (blocks) and five treatments (Fig. 1). Each plot was 9.0 m² in area (3.0 m × 3.0 m). The treatments were as follows: (1) CF (as (NH₄)₂SO₄) application in autumn (27 October 2003, 25 October 2004), (2) ADCS application in autumn (dates as above), (3) no nitrogen application (Control), (4) CF application in spring (22 April 2004, 28 April 2005), (5) ADCS application in spring (dates as above).

Figure 1 Outline of the experimental field, the measurements taken and the dataset.

Both ADCS and CF were manually applied onto the surface of the grassland. The ADCS was applied at a rate of 6.0 kg m⁻² (60 Mg ha⁻¹), which is the recommended rate in Hokkaido (Matsunaka et al. 1998). The concentration of ammonium nitrogen (NH₄-N) and the total N were measured at each application. Ammonium sulfate was applied at the same application rate of NH₄-N in ADCS. The applied NH₄-N ranged from 71 to 102 kg N ha⁻¹ (Table 1). For the ADCS treatments, four types of N application rate were calculated by considering NH₃ volatilization immediately after application (mostly within 48 h; Matsunaka et al. 2008) (Table 1). Phosphorus (superphosphate) and K (potassium sulfate) were applied to all plots at a rate of 60 kg P₂O₅ ha⁻¹ year⁻¹ and 180 kg K₂O ha⁻¹ year⁻¹, respectively; two-thirds was applied in spring (at the end of April) and the remainder was applied after the first harvest. The timothy was harvested twice per year, in June (24 June 2004, 16 June 2005) and in August (26 August 2004, 23 August 2005).

Although there were five treatments in the field, six treatment datasets were obtained by considering two periods (autumn start and spring start) for the Control (Fig. 1). As shown in Fig. 1, Y2004 and Y2005 are the datasets for the first and second years, respectively. For the autumn application, Y2004 and Y2005 were obtained from 25 October 2003 to 16 October 2004 and from 17 October 2004 to 27 October 2005, respectively. For the spring application, Y2004 and Y2005 were obtained from 21 April 2004 to 24 April 2005 and from 25 April 2005 to 28 April 2006, respectively.

Field measurements and laboratory analyses

Measurement of the N₂O and CH₄ fluxes and soil sampling for the determination of the ammonium (NH₄-N) and nitrate (NO₃-N) contents in the soil were conducted from 25 October 2003 to 28 April 2006, including over the snow season. The sampling frequency was higher (at least twice per week) just after the application of CF and ADCS than in the growing season (usually once per week, at least once every 2 weeks). The total number of field samplings was 75. On most sampling dates, the flux measurement was conducted in the morning and soil sampling was conducted in the afternoon.

Table 1 Application of anaerobically digested cattle slurry (ADCS) and chemical fertilizer (CF) (NH4)2SO4

Year and application season	ADCS								CF N application rate NH4 (kg N ha^−1)
	Application rate (Mg ha^−1)	N concentration^†			N application rate^‡
		NH4-N (g kg^−1)	Total-N (g kg^−1)	Org-N (g kg^−1)	NH4–NH3 (kg N ha^−1)	NH4 (kg N ha^−1)	TN–NH3 (kg N ha^−1)	TN (kg N ha^−1)
Y2004
Autumn in 2003	60	1.18	2.94	1.76	48	71	154	176	71
Spring in 2004	60	1.67	3.27	1.60	68	100	164	196	100
Y2005
Autumn in 2004	60	1.20	2.77	1.57	49	72	143	166	72
Spring in 2005	60	1.70	3.41	1.71	69	102	172	205	102
^†Nitrate (NO3-N) was not detected. ^‡NH3 volatilization factor was assumed to be 0.32 (Matsunaka et al. 2008). NH4–NH3 is the amount of applied NH4-N minus NH3 volatilization; NH4 is the amount of applied NH4-N; TN–NH3 is the amount of applied total N minus NH3 volatilization; TN is the amount of applied total N.

The closed chamber technique was used to determine the N₂O and CH₄ fluxes from the soil surface to the atmosphere. In each plot, a round shape metal collar (0.285 m in inside diameter or 0.0638 m² in area) equipped with a small ditch on the top side was inserted into the soil surface in mid-October 2003. These collars were not removed until the end of April 2006. At each sampling time, a cylindrical polyvinyl chloride chamber (49.0 cm in height) was inserted into the ditch with some water to make it airtight, and the chamber was in place for 0.50 h. The total volume inside the chamber was 0.0392 m³. The gas samples inside the chamber (∼25 cm³) were collected into evacuated glass bottles (15 cm³) with butyl rubber stoppers using a portable syringe at 0.00 and 0.50 hours after chamber placement. A gas chromatograph equipped with an electron capture detector (GC-14B; Shimadzu, Kyoto, Japan) and a gas chromatograph equipped with a flame ionization detector (GC-14B; Shimadzu) were used to determine the N₂O and CH₄ concentrations, respectively. The flux was calculated using the following equation:

where F is the gas flux (μg N₂O-N m⁻² h⁻¹ or μg CH₄ m⁻² h⁻¹), ρ is the gas density under STP (273.15 K and 101325 Pa) (1.26 × 10³ g m⁻³ for N₂O-N, 0.717 × 10³ g m⁻³ for CH₄), V and A are the volume (0.0392 m³) and area (0.0638 m²) of the chamber, respectively, C _0.50 and C _0.00 are the N₂O or CH₄ concentrations (μm³ m⁻³) at 0.50 and 0.00 h after chamber placement, respectively, t is the time of chamber placement (0.50 h), and T is the average temperature (K) measured inside the chamber. It is important to note that CF and ADCS were manually applied to both the inside and outside of the chamber collars at the rates presented above. In the snow seasons, a similar collar was temporarily set on the snow surface in each plot and the same procedure was conducted.

The topsoil at a depth of 0–15 cm was collected on each sampling date to determine the soil NH₄-N and NO₃-N contents. Field-moist soil samples were extracted in 100 g L⁻¹ KCl solution (soil : solution = 1:5) and the NH₄-N and NO₃-N concentrations in the extracted solution were determined using a flow injection analyzer (FIA star 5012 analyzer; FOSS, Hillerød, Denmark). Soil NH₄-N and NO₃-N contents (mg N kg⁻¹) were calculated using these concentrations and the soil moisture content measured by drying soil at 105°C for >48 h. The water-filled pore space (WFPS) was calculated using soil moisture content, bulk density (0.81 Mg m⁻³) and soil particle density (assumed to be 2.65 Mg m⁻³).

At each harvest, timothy grass (1.0 m² area) in each plot was cut at a height of 5 cm from the soil surface, and collected and weighed. One hundred grams of the grass was taken and dried at 70°C for >48 h. After weighing, a portion of the grass was milled and the N content was determined by wet ashing and the steam distillation technique. Finally, the N uptake (kg N ha⁻¹) by timothy grass was calculated.

Daily precipitation (mm) and the daily mean air temperature (°C) of the experimental sward were obtained from the weather station at Rakuno Gakuen University.

Fertilizer-induced N ₂ O emission factor and statistical analyses

The fertilizer-induced N₂O-N EF was calculated in each block and year. For the ADCS treatments, four types of N application rate were calculated by considering NH₃ volatilization (Table 1). According to these four values, four types of EFs can be calculated for the ADCS plots as follows:

where EF is the fertilizer-induced N₂O EF (kg N₂O-N [kg N]⁻¹) and N₂O-N is the annual fertilizer-induced N₂O emission rate (kg N₂O-N ha⁻¹ year⁻¹), which was obtained by subtracting the emission from the control plot in each block and year; is the amount of applied NH−⁴N minus NH₃ volatilization (kg N ha⁻¹ year⁻¹); N_TN is the amount of applied total N (kg N ha⁻¹ year⁻¹). In this paper, these four types of EF were calculated. For the CF plots, was calculated, but all four EF types are assumed to be equal, , because there is no NH₃ volatilization from the CF plots (Matsunaka et al. 2008).

A three-way ANOVA was used to examine the significance of the factors; the material (control, ADCS, CF), the season (autumn, spring) and the year (Y2004, Y2005). The PC software SigmaStat (SPSS 1997) was used for the ANOVA.

Results

Seasonal changes in N ₂ O and CH ₄ fluxes

The daily precipitation and mean air temperature from October 2003 to April 2006 are shown in Figs 2,3. Maximum daily precipitation (94.5 mm) was recorded on 7 September 2005. The highest (25.8°C) and lowest (−10.9°C) daily mean air temperatures were recorded on 9 August 2004 and 23 January 2006, respectively. The amount of precipitation in the second year (2005, 807 mm) was greater than that in the first year (2004, 564 mm), and the air temperature showed a similar pattern (Fig. 4). The mean air temperature of the ADCS application dates ranged from 5 to 11°C and precipitation within 12 h after the application of ADCS was not more than 1 mm. These results show that using only one value (0.32) for NH₃ volatilization factor (Table 1) is possible.

Ammonium and nitrate contents in the topsoil fluctuated (Fig. 2). Both contents sharply increased and gradually decreased within 1 month after the application of ADCS or CF, although the magnitudes of the variation differed. These increases and decreases result from the addition of NH₄ with the applied materials, nitrification of the added NH₄ and uptake by timothy. Little or no change was observed after harvesting of the grass.

Daily mean N₂O fluxes ranged from −7 to +113 μg N₂O-N m⁻² h⁻¹ (Fig. 3). Compared with control plots, higher N₂O fluxes were observed in the ADCS and CF plots, particularly after the autumn application. These higher N₂O fluxes after the application of ADCS or CF coincided with fluctuations in the NH₄ and NO₃ contents in the topsoil. Little or no change was observed after harvesting or heavy rain.

Daily mean CH₄ fluxes ranged from −164 to +43 μg CH₄ m⁻² h⁻¹ (Fig. 3). Many CH₄ flux values showed negative values, meaning that the methane was absorbed from the atmosphere into the topsoil. There was no noticeable change in CH₄ flux after the application of ADCS or CF, harvesting and heavy rain. It appears that CH₄ fluxes in the first year were lower (higher absorption) than those in the second year.

Figure 2 (A) Seasonal changes in daily precipitation (bars) and mean air temperature (line), (B) NH4 and (C) NO3 contents in the topsoil (0–15 cm). Small letters and dotted lines denote management events; that is, a, autumn application of anaerobically digested cattle slurry (ADCS) or chemical fertilizer (CF); b, spring application of ADCS or CF; c, first harvest in June; d, second harvest in August. Ctrl, plots with no application; CF Autumn, plots with CF application in autumn; ADCS Autumn, plots with ADCS application in autumn; CF Spring, plots with CF application in spring; ADCS Spring, plots with ADCS application in spring. Circles and error bars show the mean and standard deviation (n = 3), respectively.

Nitrogen uptake by timothy grass

The annual dry matter production and N uptake by timothy ranged from 3.75 to 13.2 Mg ha⁻¹ year⁻¹ and from 59 to 204 kg N ha⁻¹ year⁻¹, respectively (Fig. 5). Dry matter production and N uptake until first harvest accounted for more than half of the annual uptake (data not shown). The ANOVA showed a significant interaction (season × material, P < 0.01) for both dry matter production and N uptake. This interaction appears to be caused by low growth in CF with autumn application (Fig. 5), which probably resulted from N loss through nitrate leaching followed by rainfall and snow melting before the growing season (Matsunaka et al. 2006). However, the reason for the relatively high dry matter production and N uptake in ADCS with autumn application is unclear. In addition to the significant interaction, the main effects were also significant (material, season, year, P < 0.01) for both dry matter production and N uptake, that is, the application of ADCS and CF increased the dry matter and N uptake of the grass. There was no significant difference in mean N uptake between the ADCS and CF plots (Fig. 5). The higher dry matter production and N uptake in the first year (2004) compared with the second year (2005) is most likely to have been caused by the higher air temperature in spring. The mean air temperature in May 2004 (12.0°C) was higher than that in May 2005 (9.2°C) (Fig. 4).

Figure 3 (A) Seasonal changes in daily precipitation (bars) and mean air temperature (line), (B) N2O and (C) CH4 fluxes from the soil surface to the atmosphere. Refer to Fig. 2 for a description of the abbreviations.

Figure 4 Monthly meteorological data (rain and mean air temperature) during the growing seasons.

Annual N ₂ O emission and CH ₄ absorption rate

Annual N₂O emission rates ranged from 0.349 to 0.926 kg N₂O-N ha⁻¹ year⁻¹ (Fig. 6). The ANOVA showed a significant effect of material (P = 0.042). The application of ADCS and CF increased the annual N₂O emission; the mean annual N₂O emission rates in the ADCS (0.708 kg N₂O-N ha⁻¹ year⁻¹) and CF plots (0.636 kg N₂O-N ha⁻¹ year⁻¹) were significantly higher than those in the control plots (0.398 kg N₂O-N ha⁻¹ year⁻¹) (Fig. 6). There was no significant difference in mean annual N₂O emission between the ADCS and CF plots. Autumn application plots had relatively higher (P = 0.134) N₂O emission rates than spring application plots.

Figure 5 (A) Annual dry matter production and (B) N uptake by timothy grass. Bars and error bars show the mean and standard deviation (n = 3), respectively. The same superscript letters show no significant difference using a Fisher least significant difference test (P > 0.05). ACDS, anaerobically digested cattle slurry; CF, chemical fertilizer.

Figure 6 (A) Annual N2O emission rate and (B) CH4 absorption rate. Bars and error bars show the mean and standard deviation (n = 3), respectively. The same superscript letters show no significant difference using a Fisher least significant difference test (P > 0.05). ACDS, anaerobically digested cattle slurry; CF, chemical fertilizer.

Annual CH₄ absorption rates ranged from 1.65 to 4.59 kg CH₄ ha⁻¹ year⁻¹ (Fig. 6). The ANOVA showed a significant effect of year (P = 0.014). Mean CH₄ absorption rate in the first year (3.28 kg CH₄ ha⁻¹ year⁻¹) was significantly higher than that in the second year (2.31 kg CH₄ ha⁻¹ year⁻¹). The application of ADCS and CF did not affect the annual CH₄ emission rate.

Fertilizer-induced N ₂ O emission factor

When calculating the fertilizer-induced N₂O EF, the fertilizer-induced N₂O emission rate should be obtained by subtracting the emission from the control plot. A significant difference in the annual N₂O emission rate between the applied materials was found, indicating that it is possible to calculate the fertilizer-induced N₂O EF. The mean EF ranged from 0.0009 to 0.0089 (kg N₂O-N [kg N]⁻¹), and large variations were found (Fig. 7). It appears that the EFs in spring application plots were lower than those in autumn application plots. However, there were no significant factors (ANOVA; P > 0.05), although the P-values for season were relatively low (0.067–0.160). Table 2 shows the total mean (± standard error) of the four types of EFs. The total mean values ranged from 0.0024 to 0.0043.

Figure 7 Fertilizer-induced N2O-N emission factor (EF). Bars and error bars show the mean and standard deviation (n = 3), respectively. ACDS, anaerobically digested cattle slurry; CF, chemical fertilizer; TN, total nitrogen.

Table 2 Summary of the fertilized-induced N2O emission factors

Type of emission factor		Mean ± standard error
CF	ADCS
		0.0043 ± 0.0009 (n = 24)
		0.0034 ± 0.0008 (n = 24)
		0.0025 ± 0.0007 (n = 24)
	EFTN	0.0024 ± 0.0007 (n = 24)
Units are kg N2O-N (kg N)^−1. ACDS, anaerobically digested cattle slurry; CF, chemical fertilizer.

Discussion

N ₂ O and CH ₄ flux

In previous reports, the N₂O flux increased following rainfall after manure application in a grassland located in Tochigi (Mori et al. 2008). In addition, an increase in N₂O flux after heavy rain and/or harvest has been observed in crop fields located in Hokkaido (Katayanagi et al. 2008; Kusa et al. 2002, 2006; Mu et al. 2008). However, little or no change in N₂O flux was observed after harvesting or heavy rain in the present study. In another grassland with no fertilizer application located in Tochigi (Mori et al. 2005) and in an arable cropping system located in Hokkaido (Koga et al. 2004) the correspondence relationship between heavy rain and an increase in N₂O flux was not clear as it was in our study.

Prediction of the N mineralization rate from cattle manure was proposed (e.g. Shiga et al. 1985), but the rate from organic matter in ADCS applied to grassland surface is unknown. However, according to the results of Saigusa and Watanobe (2006), it is likely that decomposition of organic matter in the ADCS and subsequent mineral N supply to soil compensate the N loss through NH₃ volatilization. However, in the present study, a distinctive increase in mineral N (NH₄, NO₃) and N₂O flux in the ADCS plots was not observed in the growing seasons (July–September) (Figs 2,3).

Higher N₂O fluxes were observed mainly after the application of ADCS or CF, which coincided with fluctuations in the NH₄ and NO₃ contents in the topsoil. Therefore, the increase in the annual N₂O emission rates in the ADCS and CF plots compared with the control plots could be attributed to N₂O production via nitrification of added NH₄-N in the present study.

In general, it is considered that the application of organic matter to the soil increases N₂O emission because the easily degradable carbon enhances de-nitrification. For example, a grassland experiment under the same N application rate in Tochigi Japan showed that annual N₂O emission rates were significantly higher from manure plots than from chemical fertilizer plots (Mori et al. 2008). In contrast, there was no significant difference in mean annual N₂O emission between the ADCS and CF plots (Fig. 6) in the present study. In other words, it appears that organic matter (ADCS) application did not increase the N₂O emission under the same NH₄-N application rate. However, we should consider N loss just after ADCS application through NH₃ volatilization. Therefore, comparing in the CF plots with in the ADCS plots (Fig. 7) is suitable for checking the effect of organic matter. No significant effect of the material (P = 0.195) was found because of high variation; however, in the ADCS plots tended to be higher than in the CF plots (Fig. 7). These results suggest that the application of organic matter onto grassland surfaces increases N₂O emission to some extent. The effect of organic mater application onto grassland surfaces on N₂O emission should be analyzed and predicted quantitatively in the future.

Negative CH₄ fluxes were observed, which were considered to result from CH₄ consumption (oxidation) by methanotrophs in the soil. In general, it is well known that both methanotrophs and methanogens exist in soils and that this bacterial activity and the CH₄ flux are related to soil water content and redox condition (Le Mer and Roger 2001). A positive correlation between the CH₄ flux and soil water content (Mori et al. 2005, 2008) and an increase in CH₄ flux after heavy rain (Mori et al. 2008) have been observed in a grassland in Nasu, Tochigi prefecture, central Japan. In the present study, a noticeable increase in CH₄ flux after heavy rain was not recorded. However, the annual CH₄ absorption rate in the first year was significantly higher than that in the second year, which probably resulted from a smaller amount of precipitation and thus a more aerobic soil, and this was favorable for methanotroph activity in the first year. In fact, as shown in Fig. 8, soil WFPS and CH₄ flux tended to be higher in the second year (2005) than in the first year (2004) (P = 0.08 for WFPS, P < 0.01 for CH₄ flux using a t-test), and a significant positive relationship between WFPS and CH₄ flux was found.

Many papers have reported that CH₄ oxidation potential and CH₄ flux in upland soils are reduced and increased, respectively, by ammonium N fertilizer application (Hu et al. 2002; Le Mer and Roger 2001; Mosier et al. 1991) and by the NH₄-N content in the soil (Mori et al. 2005, 2008). However, in the present study, ammonium N fertilizer as ADCS or CF did not reduce the annual CH₄ absorption rate compared with the control plots. A possible explanation is that the period with high soil NH₄-N content was very short (approximately 1 month after the application).

Figure 8 Relationship between soil water-filled pore space (WFPS) and CH4 flux in the growing season (April–November) in 2004 and 2005. In the box plots, the boundary of the box closest to zero indicates the 25th percentile, the line within the box marks the median and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers above and below the box indicate the 90th and 10th percentiles, respectively.

Annual rate of N ₂ O emission and CH ₄ absorption

Annual N₂O emission rates ranged from 0.349 to 0.926 kg N₂O-N ha⁻¹ year⁻¹ in the present study. Compared with the N₂O emission rates in Japanese grasslands and pastures reported in peer-reviewed papers, the rates in the present study are similar to those in a grassland located in central Hokkaido (0.15–1.54 kg N₂O-N ha⁻¹ 6-months⁻¹; Mu et al. 2008) and in a grassland with no fertilizer application located in Tochigi (0.39–1.59 kg N₂O-N ha⁻¹ year⁻¹; Mori et al. 2005); and lower than those in a grassland with manure and chemical fertilizer application located in Tochigi (4.7–11.0 kg N₂O-N ha⁻¹ year⁻¹; Mori et al. 2008) and in grasslands and pastures located in southern Hokkaido (1.1–42.8 kg N₂O-N ha⁻¹ year⁻¹; Katayanagi et al. 2008).

The total mean of the annual CH₄ absorption rate was 2.80 kg CH₄ ha⁻¹ year⁻¹ (ranged from 1.65 to 4.59). Compared with the CH₄ absorption rates in Japanese uplands (grassland, arable field, forest) reported in peer-reviewed papers, the rates in the present study were higher than those in an onion field in Hokkaido (−0.78 to 0.36 kg CH₄ ha⁻¹ 6-month⁻¹; Hu et al. 2002), several cropping systems in Hokkaido (0–1.1 kg CH₄ ha⁻¹ 6-months⁻¹; Mu et al. 2006) and a grassland fertilized with manure and ammonium sulfate in Tochigi (0.16–0.84 kg CH₄ ha⁻¹ year⁻¹; Mori et al. 2008); similar to those in an arable cropping system in Hokkaido (1.4–2.4 kg CH₄ ha⁻¹ year⁻¹; Koga et al. 2004), non-fertilized grassland in Tochigi (1.8–2.4 kg CH₄ ha⁻¹ year⁻¹; Mori et al. 2005) and in 10 forests in Hokkaido (1.4–6.6 kg CH₄ ha⁻¹ 6-months⁻¹; Morishita et al. 2004); and lower than those in 25 forests in Japan (2.7–24.8 kg CH₄ ha⁻¹ year⁻¹; Ishizuka et al. 2009).

These annual N₂O emission and CH₄ absorption rates, including the rates from our study, show high variation, suggesting that N₂O and CH₄ fluxes from grassland are regulated by many factors, such as plant species, soil type and properties, climate condition, management practices, and any interactions between these factors. Collaborations between field measurements and model analyses (e.g. Saggar et al. 2009) will be required to estimate the amount of N₂O emission and CH₄ absorption in Japanese grasslands in the future.

Fertilizer-induced N ₂ O emission factor

High variation was observed in the calculated fertilizer-induced N₂O EFs, which could be a reason for the lack of significance of all factors in the analysis. However, the EFs in the autumn application plots tended to be higher than those in the spring application plots, which is considered to result from relatively higher N₂O emission and the lower N application in the autumn application plots compared with the spring application plots (Table 1). Relatively higher N₂O emission in the autumn application plots may be produced by higher N₂O fluxes just after the autumn application. The rate of grass growth (N uptake) following the spring application is usually higher than that following the autumn application (Matsunaka 1987), which would suppress soil NH₄-N or NO₃-N available for N₂O production and flux.

Which type of EF is the most suitable for evaluation of N₂O emission from the ADCS plots? In the present study, the NH₃ volatilization factor was assumed to be 0.32, but the rate can change with different field conditions, such as temperature and rainfall (Matsunaka et al. 2008). In addition, mineralization of organic matter in ADCS occurs, which can be considered in medium-term to long-term evaluations, although the mineralization rate is unknown at present. Taking these facts into account, EF_TN will be the most explicit, universal and convenient measure. The total mean ± standard error of EF_TN was 0.0024 ± 0.0007 (kg N₂O-N [kg N]⁻¹), which is similar to the reported EF (0.0032 ± 0.0013) for well-drained uplands in Japan (Akiyama et al. 2006). Akiyama et al. (2006) also reported a high EF (0.014) for poor-drained uplands in Japan. Toma et al. (2007) investigated variation in the EFs derived from chemical N fertilizer (EF_F) and organic matter (EF_O) in two soils located in Hokkaido. They suggested that EF_F and EF_O were related to the mean annual air temperature and mean annual relative humidity, respectively. Along with field measurements and model analyses, considering these factors for EF will contribute to increasing the precision of the estimation of N₂O emission from grasslands.

Conclusions

Under the same level of application of NH₄-N, differences in the applied materials (ADCS or CF) and application season (autumn or spring) did not lead to any significant difference in N₂O emission, fertilizer-induced N₂O EF or CH₄ absorption.

Acknowledgments

This work was partially supported by the New Energy and Industrial Technology Development Organization as part of a project entitled “Development of Technology to Assess and Verify Life Cycle CO₂Emissions; Development of demonstrate the CO₂emission output on lifecycle of products and others; Application of LCA methodology to local policies: Case studies on biogas plants in Betsukai-Chou, Hokkaido” (FY2003-2005).