Global Biogeochemical Cycles

N2O emission and CH4 uptake in arable fields managed under conventional and reduced tillage cropping systems in northern Japan

Authors

  • Nobuhisa Koga,

    1. Crop Production Research Team, Department of Upland Agriculture Research, National Agricultural Research Center for Hokkaido Region (NARCH), Kasai, Hokkaido, Japan
    Search for more papers by this author
  • Haruo Tsuruta,

    1. Greenhouse Gas Emission Team, Department of Global Resources, National Institute for Agro-Environmental Sciences (NIAES), Tsukuba, Ibaraki, Japan
    2. Now at Center for Climate System Research, University of Tokyo, Tokyo, Japan.
    Search for more papers by this author
  • Takuji Sawamoto,

    1. Greenhouse Gas Emission Team, Department of Global Resources, National Institute for Agro-Environmental Sciences (NIAES), Tsukuba, Ibaraki, Japan
    2. Now at Faculty of Dairy Science, Rakuno Gakuen University, Ebetsu, Hokkaido, Japan.
    Search for more papers by this author
  • Seiichi Nishimura,

    1. Greenhouse Gas Emission Team, Department of Global Resources, National Institute for Agro-Environmental Sciences (NIAES), Tsukuba, Ibaraki, Japan
    Search for more papers by this author
  • Kazuyuki Yagi

    1. Greenhouse Gas Emission Team, Department of Global Resources, National Institute for Agro-Environmental Sciences (NIAES), Tsukuba, Ibaraki, Japan
    Search for more papers by this author

Abstract

[1] Nitrous oxide (N2O) emission and methane (CH4) uptake were measured in an experimental long-term tillage field (Andosol) in Hokkaido, northern Japan, to assess their contributions to net global warming, associated with arable crop production. From May 2001 to August 2002, the field was cultivated with winter wheat, adzuki bean, sugar beet, potato, and cabbage, where the total N applied was 110, 40, 150, 60, and 220 kg N ha−1 yr−1, respectively. Under conventional tillage (CT) cropping systems, basal N fertilization and plowing for residue incorporation had little effect on N2O fluxes, but vigorous N2O emission was observed when rotary harrowing was used for incorporating N-rich cabbage residues into soil in summer. Also, high N2O emissions occurred when there was heavy rainfall after a large amount of N fertilizer had been applied to sugar beet and also when there was thawing of frozen soil and snow in the winter wheat treatment. Despite the differing N2O flux patterns among the crops, the annual N2O emissions from each crop were positively correlated with the total N applied as fertilizer. Under CT systems, across all five crops, the mean N2O emission factor (the percent ratio of N2O-N emitted out of total N applied as fertilizer) was 0.36%. Under reduced tillage (RT) cropping systems, where crop residues were left on the ground over winter, large quantities of N2O were emitted from adzuki bean and sugar beet residues when the frozen soil and snow thawed. Therefore, total N2O emissions from adzuki bean and sugar beet cultivated under RT systems were much greater than under CT systems. The rates of CH4 uptake by arable soils were less sensitive to crop type, field management practices, and fertilizer application rates, but the rates were strongly influenced by long-term tillage management. For fallow, winter wheat, adzuki bean, and sugar beet treatments, the CH4 uptake rates in the CT soils (1.36 kg CH4 ha−1 yr−1), which had a 20-year history of intensive plowing, were lower than those in the RT soils (2.40 kg CH4 ha−1 yr−1). Thus RT production systems improved CH4 uptake by arable soils, although they adversely affected N2O emissions for adzuki bean and sugar beet production.

1. Introduction

[2] Global warming resulting from greenhouse gas (mainly CO2, N2O, and CH4) emissions is currently one of the major environmental issues facing managers of agricultural systems [Robertson et al., 2000; Organization for Economic Co-operation and Development (OECD), 2001; Flessa et al., 2002], and mitigation options are being studied increasingly [Smith et al., 1997; Robertson et al., 2000]. In arable-land farming systems, the production and uptake of N2O and CH4 within cultivated lands occurs in association with soil microbial activities [Smith et al., 1997; Hütsch, 2001]. The N2O-N emitted from agricultural soils as a result of nitrification and denitrification processes arises predominantly from N fertilizers applied to the field [Smith et al., 1997; Mosier et al., 1998; Akiyama and Tsuruta, 2003]. Therefore the quantity of fertilizer-N applied can be an important index for estimating N2O emissions from agricultural lands [Intergovernmental Panel on Climate Change (IPCC), 1997]. In contrast, atmospheric CH4 is taken up by dry field soils through the action of CH4-oxidizing bacteria [Hütsch, 2001]. The rates of CH4 uptake are affected by soil factors such as moisture content, temperature, and mineral N content [Dobbie and Smith, 1996; Kessavalou et al., 1998].

[3] The effects of field management practices on N2O and CH4 fluxes in agricultural lands have been studied intensively. Significant N2O emissions from arable soils were noted after field operations such as tillage [Kessavalou et al., 1998; Ball et al., 1999; Baggs et al., 2003], fertilizer application [Smith et al., 1997; Akiyama and Tsuruta, 2003], and residue incorporation [Baggs et al., 2000]. Soil disturbance by such agricultural practices can also change the CH4 uptake rates in arable soils [Hütsch, 1998]. For example, deep tillage significantly reduces CH4 uptake, and therefore no till or reduced tillage can be an effective means of reducing atmospheric CH4 [Cochran et al., 1997; Hütsch, 1998; Kessavalou et al., 1998; Ball et al., 1999]. In addition, climatic events considerably influence the fluxes of these gases. Changes in soil moisture content following rainfall [Kessavalou et al., 1998; Sawamoto and Hatano, 2000] and thawing of frozen soil [Christensen and Tiedje, 1990; Flessa et al., 1995] are crucial in determining the fluxes of N2O and CH4. Moreover, soil type is also an important factor affecting N2O emission rates: at Mikasa, in the middle of the Hokkaido region of Japan, the N2O emission factor was 2.8% for gray lowland soil planted with onions (Allium cepa. L.) [Kusa et al., 2002], but less than 0.5% for volcanic ash soils in Tsukuba, central Japan [Akiyama and Tsuruta, 2002; Cheng et al., 2002].

[4] Located in northern Japan, the Tokachi region of Hokkaido is the primary site of arable crop production in Japan. There, four major crops, winter wheat (Triticum aestivum L.), beans (adzuki bean: Vignaangularis (Willd.) Ohwi and Ohashi; kidney bean: Phaseolus vulgaris L.; soybean: Glycine max Merr.), sugar beet (Beta vulgaris L.) and potato (Solanum tuberosum L.) are cultivated in rotation [Yamagata, 2001]. Recently, farmers have expanded their cultivation of vegetables, including cabbage (Brassica oleracea L.), to increase their incomes. As farmers use considerable quantities of chemical fertilizers and practice a wide variety of soil-disturbing field operations (seedbed preparation, plowing, fertilization, sowing, harvesting, and residue management), N2O and CH4 fluxes may vary substantially.

[5] In winter, soils in the Tokachi region of Hokkaido freeze and snowfall events occur. Consequently, when frozen soils and snow thaw in early spring, soils become saturated. In regions where soils undergo a freeze-thaw cycle, significant N2O emissions have been shown to occur during spring thaws [Goodroad and Keeney, 1984; Christensen and Tiedje, 1990; Flessa et al., 1995; Wagner-Riddle and Thurtell, 1998]. Two possible explanations have been proposed for thaw-induced N2O emissions: (1) the short-term release of N2O derived from residual NO3 remaining from the previous autumn [Burton and Beauchamp, 1994; Wagner-Riddle and Thurtell, 1998] or (2) the rapid decomposition of organic matter occurring at the wet ground surface in the presence of C available for denitrification [Christensen and Tiedje, 1990; Christensen and Christensen, 1991]. It was of interest to assess whether arable lands of northern Japan presented an N2O emission burst at thawing. This study sought to quantify net annual N2O and CH4 fluxes in arable fields under conventional cropping systems of the Hokkaido region in northern Japan and to identify which management practices, including reduced tillage, had the greatest influence on these fluxes.

2. Materials and Methods

2.1. Description of the Study Site

[6] A field experiment to measure fluxes of N2O and CH4 in arable lands of the Tokachi region was conducted at the National Agricultural Research Center for Hokkaido Region (NARCH), Japan (143°03′E, 42°53′N), from May 2001 to August 2002. The soil was Typic Hapludands [Natural Resources Conservation Services (NRCS), 1998], which is a well-drained volcanic ash soil. The field had been continuously managed under two tillage systems, conventional tillage (CT) and reduced tillage (RT) since 1980. Each year, the CT plots were harrowed twice to a depth of roughly 10 cm for seedbed preparation in early spring and plowed once to a depth of roughly 25 cm after harvesting to incorporate crop residues. In RT plots, the soil was harrowed once, and plowing was omitted. The resulting soil properties under both tillage treatments are listed in Table 1. As a result of continuous tillage treatments over a 20-year period, the RT soils had a higher C content than did the CT soils.

Table 1. Properties of CT and RT Soils at NARCH
 CTRT
0–5 cm Depth5–10 cm Depth10–20 cm Depth0–5 cm Depth5–10 cm Depth10–20 cm Depth
T-C, g kg−129.830.230.236.134.731.8
T-N, g kg−12.42.52.53.12.82.5
C/N ratio12.412.112.111.612.412.7
pH (H2O)5.75.75.85.55.55.6
Bulk density, g cm−30.700.890.970.770.770.82

2.2. Crop Cultivation and Field Management

[7] In CT plots, adzuki bean (cv. Erimoshozu), sugar beet (cv. Megumi), potato (cv. Kita-akari), and cabbage (cv. Early Ball) were sown or transplanted on 15 May 2001, and winter wheat (cv. Hokushin) was sown on 21 September 2001 with basal fertilization. After harvesting, the residues of adzuki bean, sugar beet, and potato were incorporated into the soil by moldboard plowing, and residues of winter wheat and cabbage were incorporated to a depth of roughly 10 cm by rotary harrowing. In RT plots where winter wheat, adzuki bean, and sugar beet were cultivated, the residues of adzuki bean and sugar beet were spread uniformly after harvesting, and remained on the soil surface over the winter. In this region, wheat straw is normally removed from the field for livestock; therefore only stubble was left in the field. The plant density and fertilizer application rates used are presented in Table 2. Commercial compound fertilizers containing N, P, K, and Mg were applied as basal fertilizer at a depth of 5 to 7 cm below plant rows as is conventional for this region. Extra fertilizer in the form of granular ammonium sulfate or a compound fertilizer containing N and K was broadcast twice each during the growth of winter wheat and cabbage, respectively. In fallow treatments, there were no fertilizer applications and no crops cultivated.

Table 2. Planting Density and Fertilizer Application Rates (kg ha−1) Used in This Study
CropPlanting Density,a Plant m−2Row Spacing, cmBasal FertilizationExtra FertilizationN in Total
NH4-NNO3-NP2O5K2ONH4-NK2O
  • a

    In g seeds m−2 for winter wheat.

Fallow--0000000
Winter wheat12.030600150100500110
Adzuki bean16.760400200800040
Sugar beet6.760816925016000150
Potato4.47545152001200060
Cabbage5.2609317210135110135220

2.3. Gas Flux Measurements

[8] Flux measurements began on 9 May 2001 (30 August 2001 for winter wheat) just before sowing or transplanting, and continued for 1 year, except for the period of 19 December 2001 to 13 March 2002, as no major N2O fluxes during soil freezing had been reported in Ontario, Canada, in the 3-year field experiment by Wagner-Riddle and Thurtell [1998]. Gas sampling was conducted weekly by a closed chamber technique, using rectangular chambers with an area of 4500 cm2 and a height of 43.5 cm, with two replicate chambers per plot. On each sampling day, three gas samples were collected from each chamber with a disposable syringe at 0, 10, and 20 min, and collected samples were transferred into evacuated glass vial bottles with butyl rubber stoppers. Air temperature inside the chambers was also recorded at every sampling occasion. One row of each crop (two rows for winter wheat) was included in the chamber. Tall chambers, 83 cm in height, were used for winter wheat, sugar beet, and potato in the middle of their growing seasons.

[9] The concentration of N2O was analyzed by a gas chromatograph equipped with an electron capture detector (GC-8A or GC-14A, Shimadzu Corp., Kyoto, Japan). The concentration of CH4 was determined by a gas chromatograph equipped with a flame ionization detector (GC-9A, Shimadzu Corp.).

2.4. Climatic and Soil Parameters

[10] The climatic data on precipitation and snow height were collected near the study site at NARCH. The soil temperature in the experimental field at a 10 cm depth was recorded, and the depth of the frozen soil layer was monitored by the method of Löfvenius et al. [2003].

[11] The soil mineral N content was determined from soils sampled with an auger to a depth of 10 cm from between individual crop plants in the planted rows. The soil was passed through a 2-mm sieve and extracted with 10% (w/v) KCl. The NO3 concentration in the extracts was determined by the copper-cadmium reduction method using a Flow Injection Analyzer (K-1000, Hitachi, Tokyo), and the NH4+ concentration by the indophenol blue method [Japanese Industrial Standards Committee, 1998]. Further soil samples were taken in duplicate from between rows in each plot, using a 10-cm-tall, 200-mL core sampler, to measure water-filled pore space (WFPS) as an index of soil moisture [Linn and Doran, 1984]. Soil samples for mineral N and WFPS were collected every other week.

2.5. Data Analysis

[12] The effect of the tillage and crop treatments (fallow, winter wheat, adzuki bean, and sugar beet) on annual N2O emission and CH4 uptake was tested with two-way ANOVA. The relationship between crop type (differing in N application rates) and annual N2O emission was determined by regression analysis.

3. Results

3.1. Seasonal Variation in Climatic and Soil Parameters

[13] May and June 2001 were dry, but large rainfall events occurred from July to October 2001 (Figure 1), which is typical of this region. Snow accumulated in the field from 30 November 2001 to 29 March 2002. The maximum snow height was 85 cm. Soil freezing occurred between 28 November 2001 and 2 April 2002, and the maximum thickness of the soil freezing layer was 21 cm. The monthly averaged soil temperature at a depth of 10 cm ranged from −0.2° to 22.9°C over the study period.

Figure 1.

(a) Monthly precipitation (bars) and mean soil temperature at a depth of 10 cm (dots) during the study period. (b) Thickness of the frozen soil layer and snow height during the winter.

[14] In cultivated plots, the soil NH4-N content rose rapidly after basal fertilization (Figure 2). In sugar beet, potato, and cabbage soils, the NO3-N content rose immediately after the basal fertilization, as 46, 25, and 15%, respectively, of their total N content was contributed by NO3-N (Table 2). In adzuki bean plots, where basal fertilizer application consisted solely of NH4-N, the NO3-N content increased gradually, as a result of nitrification. Soil NH4-N virtually disappeared by the middle of August, while soil NO3-N decreased until August, presumably as a result of plant uptake and rain-driven leaching.

Figure 2.

Seasonal changes in (left) NH4-N and (right) NO3-N contents in soil. Note the different x axis for winter wheat. Arrows indicate different management practices: basal fertilization (B), extra fertilization (E), and residue incorporation (under CT systems) by plowing (P) and rotary harrowing (R). Five soil samples were collected from planted rows for each treatment, mixed well, and used for determining mineral N content.

[15] The WFPS of fallow CT plots ranged between 34 and 57%, while the mean WFPS values of CT and RT fallow soils over the study period were 43 and 52%, respectively (Figure 3). Similar trends were observed in planted soils.

Figure 3.

Seasonal changes in WFPS (0–10 cm) of fallow soils at CT and RT plots. Two core soil samples were collected for each treatment, and each dot in the figure represents the mean of two WFPS values.

3.2. Time Series of N2O Fluxes

[16] Under the cropping systems studied, there were notable responses of N2O fluxes to various agricultural practices and climatic events, depending on the crop species (Figure 4). Under winter wheat production, N2O fluxes after basal and extra N fertilizations were less than 4 g N ha−1 d−1, but large increases in N2O flux occurred under CT (50 g N ha−1 d−1) and RT (19 g N ha−1 d−1) systems during the thawing of frozen soil and snow in March. In the case of sugar beet, N2O emissions occurred predominantly in the rainy season (August to October), and small peaks in N2O fluxes (<2 g N ha−1 d−1) occurred after residue incorporation by plowing. In cabbage production, large increases in N2O production occurred after the application of extra fertilizer (2–7 g N ha−1 d−1) and also after residue incorporation by rotary harrowing in August (58 g N ha−1 d−1). A total of 599 g N2O-N ha−1 was emitted from 69.9 kg N ha−1 of incorporated cabbage residues (Table 3). Conventionally tilled soils planted with adzuki bean or potato exhibited no large increases in N2O flux over the study period. Under RT, where crop residues overwintered on the ground surface, thaw-induced N2O emissions occurred in adzuki bean (9 g N ha−1 d−1) and sugar beet (280 g N ha−1 d−1) plots. Total N2O emissions from RT adzuki bean and sugar beet plots were significantly higher than in comparable CT plots. For fallow, winter wheat, adzuki bean, and sugar beet, however, the results of ANOVA showed that there was no significant difference in annual N2O emission between CT and RT treatments (Table 4).

Figure 4.

Seasonal changes in N2O fluxes in fallow and cultivated plots. Note the different x axis for winter wheat. Arrows indicate different management practices: basal fertilization (B), extra fertilization (E), harvesting (H), and residue incorporation (under CT systems) by plowing (P) and rotary harrowing (R).

Table 3. Characteristics of Crop Residues Returned to the Field
CropResidueDry Weight, t ha−1C Content, %N Content, %N Returned, kg N ha−1C/N Ratio
CT
Winter wheatstubbles4.5631.20.4319.672.6
Adzuki beanleaves, stems, and pods1.9340.52.0740.019.6
Sugar beetleaves4.4940.61.9487.120.9
Potatoleaves and stems0.7333.51.279.326.4
Cabbagemain roots and leaves2.4832.52.8269.911.5
 
RT
Winter wheatstubbles4.3331.30.4921.263.9
Adzuki beanleaves, stems, and pods2.3741.51.8644.122.3
Sugar beetleaves6.5740.71.97129.420.7
Table 4. Annual N2O Emissions Under Tillage and Crop Treatments (n = 2)a
CropN2O Emission, kg N ha−1 yr−1
CTRTRT/CT
  • a

    For fallow, winter wheat, adzuki bean, and sugar beet, the results of ANOVA showed that there was no significant difference in annual N2O emission between CT and RT treatments. N.D.: no data.

Fallow−0.020.13-
Winter wheat0.560.530.96
Adzuki bean0.190.311.62
Sugar beet0.342.366.93
Average (fallow, winter wheat, adzuki bean, and sugar beet)0.270.833.13
Potato0.09N.D.-
Cabbage0.82N.D.-

3.3. Time Series of CH4 Fluxes

[17] CH4 fluxes were negative in most cases, but some positive fluxes were measured between July and September (Figure 5) when precipitation was high (Figure 1). CH4 fluxes were less sensitive to field management practices such as fertilization, harvesting, and residue incorporation than were N2O fluxes. For fallow, winter wheat, adzuki bean, and sugar beet, two-way ANOVA showed that there was a significant difference in annual CH4 uptake between tillage treatments (P < 0.05), but none between crop treatments (Table 5).

Figure 5.

Seasonal changes in CH4 fluxes in fallow and cultivated plots. Note the different x axis for winter wheat. Arrows indicate different management practices: basal fertilization (B), extra fertilization (E), harvesting (H), and residue incorporation (under CT systems) by plowing (P) and rotary harrowing (R).

Table 5. Annual CH4 Uptake Under Tillage and Crop Treatments (n = 2)a
CropCH4 Uptake, kg CH4 ha−1 yr−1
CTRTRT/CT
  • a

    For fallow, winter wheat, adzuki bean, and sugar beet, the results of ANOVA showed that there was a significant difference in annual CH4 uptake between CT and RT treatments (P < 0.05). N.D.: no data.

Fallow1.302.371.82
Winter wheat0.852.492.94
Adzuki bean1.812.211.22
Sugar beet1.492.521.70
Average (fallow, winter wheat, adzuki bean, and sugar beet)1.362.401.76
Potato1.44N.D.-
Cabbage1.61N.D.-

4. Discussion

4.1. Relationship Between Field Management Practices and N2O Fluxes

[18] Nitrification, the oxidation of NH4-N to NO3-N, is one of the important processes leading to N2O emissions from agricultural soils [Bouwman, 1996]. Significant increases in N2O emissions after basal N fertilization have been reported [Yamulki et al., 1995; Akiyama and Tsuruta, 2002; Cheng et al., 2002]. In our experiment, the rapid decrease in soil NH4-N content indicated that nitrification was active from May to July (Figure 2), but no major N2O fluxes were recorded after basal fertilization for the four crops planted in May (Figure 4). Therefore, nitrification after basal fertilization was not an important source of N2O in this case. It has been shown that NO emissions are greater than N2O emissions in Andosol soils under low WFPS conditions [Hou et al., 2000; McTaggart et al., 2002]. As WFPS values measured in May and June were low due to very limited precipitation (Figure 3), more NO might have been emitted than N2O. However, since McTaggart et al. [2002] found no such increase in NO production in gray lowland soils after N fertilization, it appears that soil type likely plays an important role in the balance of N2O and NO emissions after N fertilization.

[19] Top dressing of N fertilizers has also been reported to increase soil N2O emissions [Bouwman, 1996]. In our study, N2O was emitted from cabbage and winter wheat treatments when NH4-N fertilizer was broadcast on the surface as extra fertilizer (Figure 4). The fact that N2O emissions from cabbage soils were much greater than those from winter wheat can be attributed to the greater rainfall following fertilizer application to the cabbage plot than occurred after fertilizer application to the winter wheat, in addition to the higher rate of application for cabbage. Furthermore, in the case of cabbage production, some old leaves fell to the ground as the cabbage grew. Vigorous N2O release from rotting vegetable leaves was noted by Komada and Takeuchi [2003], and the contribution of old cabbage leaves to the N2O production should be investigated.

[20] During the rainy season of July to October 2001, significant N2O emissions were recorded from the sugar beet plot to which a large quantity of basal N fertilizer (150 kg N ha−1) had been applied (Figure 4). Similarly, considerable N2O emissions were also observed during the rainy season (July to September) in an onion field in Mikasa, central Hokkaido, which had received large quantities of N (about 300 kg N ha−1 yr−1) [Sawamoto and Hatano, 2000; Kusa et al., 2002]. In our study, such elevated N2O emissions were not measured for adzuki bean and potato plots, to which less N fertilizer had been applied. Therefore, given the large quantities of N fertilizers applied to crops in the Hokkaido region, heavy rainfall events can have a significant role in stimulating significant N2O emissions from these soils. Assuming that nitrification would already be complete by the rainy period (Figure 2), and any remaining NO3-N would have been leached to deeper soil layers by the greater rainfall, denitrification would be the most likely process responsible for the increased N2O fluxes during the rainy season.

[21] In August, a sharp spike in N2O flux was detected under cabbage production after its residues were incorporated into a shallow soil layer (roughly 10 cm depth; Figure 4). Assuming all N2O-N to have been produced through the decomposition of the residues, N2O emissions measured after residue incorporation (9 August to 30 August 2001), 599 g N2O-N ha−1, represented 73% of total annual N2O emission, and 0.86% of total N in the residues incorporated (Table 3). By contrast, N2O emissions were much smaller following the late-October incorporation by plowing of adzuki bean and sugar beet residues into a deeper (roughly 25 cm) soil layer: 52 and 50 g N2O-N ha−1 (Figure 4), respectively, representing 0.13 and 0.06% of total residue N incorporated (Table 3). It is likely that the C/N ratio of the residues, soil temperature after residue incorporation, and the depth to which the residues were incorporated were key factors influencing the differences in incorporated residue-derived N2O emissions (Figure 1 and Table 3). Consistent with our results, greater N2O emissions were reported when lettuce residues were incorporated into soil by rotary tillage (5 cm depth) than by plowing (35 cm depth) [Baggs et al., 2000]. One possibility is that the incorporation of moist vegetable residues into a shallow soil layer by rotary harrowing creates anaerobic microsites. During the rapid microbial decomposition of residues, the N2O produced near the ground surface can readily diffuse out of the soil to the atmosphere, whereas N2O emitted after deeper incorporation has more opportunity for reduction to N2 before being emitted from the soil [Baggs et al., 2000].

[22] In regions where soils undergo a freeze-thaw cycle, significant N2O emissions have frequently been reported at thawing [Goodroad and Keeney, 1984; Christensen and Tiedje, 1990; Flessa et al., 1995; Wagner-Riddle and Thurtell, 1998]. We observed two cases of thaw-induced N2O emissions. Under winter wheat production, regardless of tillage practice, elevated N2O emissions occurred during the March thaw, despite low soil temperatures (Figure 4). Thaw-induced N2O emissions (15 March to 8 April 2002) from CT and RT winter wheat plots were 448 and 216 g N2O-N ha−1 (Figure 4), respectively, accounting for 0.75 and 0.36% of basal fertilizer N applied. A 15N-labeling experiment [Nira and Nishimune, 1998] previously conducted in this region had shown that early spring uptake of basal N fertilizer by winter wheat plants was only 6–10 kg N ha−1 out of a total application of 60 kg N ha−1. Therefore, large amounts of residual mineral N might remain in cultivated soils planted with winter wheat. Wagner-Riddle and Thurtell [1998] showed that N2O fluxes during freeze/thaw cycles in Ontario, Canada, were positively correlated with soil NO3 content measured in the previous autumn. Thus the presence of mineral N under the frozen soil layer during winter might be one of the factors influencing the sharp and vigorous N2O fluxes at thawing.

[23] Thaw-induced N2O emissions were also observed in RT adzuki bean and sugar beet plots, but these were not found in the CT plots (Figure 4). Under RT-managed adzuki bean and sugar beet, N-rich crop residues overwintered on the soil surface (Table 3). Thaw-induced N2O emissions (15 March to 8 April 2002) from adzuki bean and sugar beet residues were 167 and 2095 g N2O-N ha−1 (Figure 4), respectively, representing 0.38 and 1.62% of total residue N returned. Unlike with winter wheat, adzuki bean and sugar beet residues remaining on the soil surface under RT management decomposed under the snow, so when the surface soil thawed and became water-saturated in the spring, a pool of residue C was available for denitrification. Additionally, according to Christensen and Christensen [1991], bioavailable C for denitrification, derived from microorganisms killed by freezing, has a crucial role in N2O production at thawing. Thus, under RT systems, crop residues containing large quantities of C and N could have supplied a readily available source of C that stimulated denitrification and induced high N2O emission rates during the thaw.

4.2. Annual N2O Emissions From Each Crop Production

[24] Applied N fertilizers are the main source of N2O emissions from agricultural soils [IPCC, 1997; Bouwman, 1996; Smith et al., 1997]. Under the CT cropping systems typical of the Tokachi region, total N2O emissions were positively correlated with the total N applied as fertilizer (Figure 6). The mean N2O emission factor for CT cropping systems was 0.36%. This factor is similar to those reported in Andosol fields in Japan [Akiyama and Tsuruta, 2002; Cheng et al., 2002], but significantly lower than the 1.25% default value used by the IPCC to estimate direct N2O emissions from agricultural soils [IPCC, 1997]. Generally, fine-textured soils have higher N2O fluxes than coarser-textured soils [Davidson, 1991]. McTaggart et al. [2002] showed that N2O production was much lower in coarse-textured Andosol soils (bulk density; ∼0.7 g cm−3) than in fine-textured gray lowland soils (bulk density; 1.1 g cm−3) by an incubation experiment. Actually, field measurements showed that in the gray lowland soil (bulk density of 1.1 g cm−3) planted with onions at Mikasa, in the central portion of Hokkaido, the emission factor was 2.8% [Kusa et al., 2002], whereas in our study on an Andosol the emission factor was only 0.36% (Figure 6). This illustrates the importance of soil physical properties such as soil texture and bulk density on N2O emissions. In general, such physical properties have a significant impact on soil moisture conditions and air-filled porosity, which in turn affect the activities of N2O-producing bacteria in the soil. These findings indicate that differences in physical properties of soils are very important in determining N2O emission factors.

Figure 6.

Relationship between total N applied as fertilizer and annual N2O emissions under CT cropping systems of Hokkaido (n = 2). Vertical bars indicate the deviation of two replicates.

[25] Owing largely to increased N2O fluxes during the decomposition of N-rich crop residues at thawing (Figure 4), annual N2O emissions under RT cropping systems for adzuki bean and sugar beet were 1.62 and 6.93 times larger, respectively, than those under CT cropping systems (Table 4). However, with winter wheat, given the large N2O fluxes which occurred under both tillage systems (Figure 4), annual N2O emissions under CT and RT systems were similar. Therefore the use of RT systems for adzuki bean and sugar beet cultivation adversely affects N2O emission rates in areas of soil freezing and snowfall, such as the Tokachi region of Hokkaido. However, it did reduce CO2 emissions by reducing fuel consumption for tillage practices [Koga et al., 2003].

[26] Of all the soil parameters analyzed, only WFPS showed a significant positive correlation with N2O fluxes (Figure 7). According to Davidson [1991], nitrification and denitrification are the predominant process for N2O production in soils at lower and higher WFPS, respectively. In our study, most of the N2O fluxes were measured from soils ranging from 30 to 60% WFPS. Although some very low fluxes were found even at the higher WFPS, the highest N2O fluxes occurred at WFPS higher than 60% during thawing, presumably through denitrification. At lower WFPS, no such high fluxes of N2O were observed, except for a high N2O flux after cabbage residue incorporation. This indicates that soil moisture content is usually the most crucial factor in controlling N2O fluxes through the year, even though various field management practices such as fertilization, tillage, and residue incorporation are conducted. The dependence of N2O emissions on WFPS has also been reported in a number of field studies [Kessavalou et al., 1998; Smith et al., 1998; Dobbie et al., 1999].

Figure 7.

Relationship between WFPS (0–10 cm) and fluxes of N2O and CH4 in CT (n = 118) and RT (n = 82) plot.

4.3. CH4 Uptake Under CT and RT Cropping Systems

[27] The CH4 uptake rates under CT and RT systems averaged across fallow, adzuki bean, sugar beet, and winter wheat treatments were 1.36 and 2.40 kg CH4 ha−1 yr−1, respectively (Table 5). This significant difference in annual CH4 uptake between tillage treatments (P < 0.05), showed that long-term tillage management operations over the last 20 years had a significant impact on the magnitude of CH4 uptake by the Japanese Andosol soil. Similar results showing larger CH4 uptake under no till and reduced tillage systems than under plow-based tillage systems have been noted in a number of cases [Cochran et al., 1997; Hütsch, 1998; Kessavalou et al., 1998; Ball et al., 1999]. Dobbie and Smith [1996] showed that there was no immediate recovery in CH4 oxidation rates when cultivation and fertilization were abandoned. Similarly, Hütsch [1998] pointed out that the activity of CH4-oxidizing bacteria responsible for CH4 uptake was susceptible to soil disturbance by intensive tillage like plowing, and was very slow to recover. Therefore tillage practices can significantly reduce CH4 uptake rates in arable soils. In our results, no significant differences in annual CH4 uptake were noted between fallow and fertilized plots (Table 5), signifying that fertilizer application rates and crop vegetation had little effect on the CH4 uptake.

[28] There was a positive correlation between WFPS (0–10 cm) and CH4 fluxes under both tillage systems (Figure 7). RT soils showed higher WFPS than CT soils (Figure 3), but total CH4 uptake was greater in RT soils than in CT soils, due principally to a higher potential for CH4 uptake in RT soils.

5. Conclusion

[29] • Under CT cropping systems in the Hokkaido region of northern Japan, N2O fluxes from arable soils were largely influenced by (1) field management practices, for example, cabbage residue incorporation by harrowing and N fertilizer application to sugar beet, a crop which receives more N fertilizer than others, and (2) climatic events: (a) heavy rainfall during the growing season of sugar beet, and (b) winter wheat at thawing when basal N fertilizer was applied in the autumn.

[30] • For the five crops cultivated under CT systems, annual N2O emissions from an Andosol field were positively correlated with the total N applied as fertilizer, and the mean N2O emission factor was 0.36%.

[31] • As a result of the decomposition of N- and C-rich residues remaining on the soil under the RT system, greater N2O production occurred at thawing (and overall) under RT systems than under CT systems for adzuki bean and sugar beet crops.

[32] • The CH4 uptake rates in arable soils were strongly influenced by long-term tillage management. Under CT soils, which had been tilled by plowing for 20 years, the annual CH4 uptake was reduced by 43%, compared to RT soils.

[33] • Of soil parameters analyzed, WFPS was positively correlated with both N2O and CH4 fluxes.

Acknowledgments

[34] We would like to express our special thanks to M. Takasugi, S. Soma, and M. Kikuchi in NARCH for supporting gas and soil sampling and to S. Banzawa, A. Yoshizawa, and N. Matsuoka in NIAES for gas analysis. We are also grateful to I. P. McTaggart in Department of Global Agricultural Sciences, University of Tokyo, for helpful comments on this paper.

Ancillary