Effect of crop residue C:N ratio on N₂O emissions from Gray Lowland soil in Mikasa, Hokkaido, Japan

Abstract

We studied the effect of crop residues with various C:N ratios on N₂O emissions from soil. We set up five experimental plots with four types of crop residues, onion leaf (OL), soybean stem and leaf (SSL), rice straw (RS) and wheat straw (WS), and no residue (NR) on Gray Lowland soil in Mikasa, Hokkaido, Japan. The C:N ratios of these crop residues were 11.6, 14.5, 62.3, and 110, respectively. Based on the results of a questionnaire survey of farmer practices, we determined appropriate application rates: 108, 168, 110, 141 and 0 g C m⁻² and 9.3, 11.6, 1.76, 1.28 and 0 g N m⁻², respectively. We measured N₂O, CO₂ and NO fluxes using a closed chamber method. At the same time, we measured soil temperature at a depth of 5 cm, water-filled pore space (WFPS), and the concentrations of soil NH⁺ ₄-N, NO⁻ ₃-N and water-soluble organic carbon (WSOC). Significant peaks of N₂O and CO₂ emissions came from OL and SSL just after application, but there were no emissions from RS, WS or NR. There was a significant relationship between N₂O and CO₂ emissions in each treatment except WS, and correlations between CO₂ flux and temperature in RS, soil NH⁺ ₄-N and N₂O flux in SSL and NR, soil NH⁺ ₄-N and CO₂ flux in SSL, and WSOC and CO₂ flux in WS. The ratio of N₂O-N/NO-N increased to approximately 100 in OL and SSL as N₂O emissions increased. Cumulative N₂O and CO₂ emissions increased as the C:N ratio decreased, but not significantly. The ratio of N₂O emission to applied N ranged from −0.43% to 0.86%, and was significantly correlated with C:N ratio (y = −0.59 ln [x] + 2.30, r ² = 0.99, P < 0.01). The ratio of CO₂ emissions to applied C ranged from −5.8% to 45% and was also correlated with C:N ratio, but not significantly (r ² = 0.78, P = 0.11).

Key words:

• C:N ratio
• emission fraction
• N₂O
• residue
• soil

INTRODUCTION

Nitrous oxide (N₂O) is one of the major greenhouse gases and has a 296-fold higher greenhouse effect than carbon dioxide (CO₂) over a 100-year time frame, and its concentration in the air is increasing to 0.8 p.p.b. y⁻¹ (Intergovernmental Panel on Climate Change 2001a). The estimated global N₂O emission is reported to be 16.2 Tg N y⁻¹, of which 25% is emitted from agricultural soils (Mosier et al. 1998). However, estimated values of N₂O emissions have a large uncertainty (Intergovernmental Panel on Climate Change 2001a; Mosier et al. 1998).

The application of crop residues to soil has been shown to increase N₂O emission (Aulakh et al. 1991; Cochran et al. 1997; Huang et al. 2004; Mori et al. 2005; Potthoff et al. 2005). In general, N₂O is produced by nitrification and denitrification in soil. Because , NH⁺ ₄, NO⁻ ₃ and organic C are used in these processes, applying crop residues to the soil and mineralizing it generally increases N₂O production. Application of crop residues to soil activates aerobic decomposition. Following that, the drawdown of oxygen activates denitrification (Potthoff et al. 2005). Vigil and Kissel (1991) reported that applying residues with a low C:N ratio encouraged mineralization, but applying residues with a higher C:N ratio advanced N immobilization. Akiyama and Tsuruta (2003) reported a negative logarithmic relationship between N₂O emission and the C:N ratio of applied organic matter at the same N application rate in an Andisol field, whereas Huang et al. (2004) reported that the ratio of N₂O emission to applied residue N increased

Table 1 Chemical properties of the soil horizons

Horizon	Depth (cm)	pH (H2O)	CEC cmolc kg^−1	Sand (%)	Silt (%)	Clay (%)	C (g kg^−1)	N (g kg^−1)	C:N
Ap1	0–10	5.8	25.5	12.1	51.2	36.7	32.1	2.80	11.5
Ap2	10–33	5.2	24.6	11.3	50.9	37.8	35.1	2.80	12.7
Bg	33–64	4.4	27.1	2.90	47.9	49.2	49.1	3.10	16.0
C	64–100	4.7	33.1	6.10	56.3	37.6	76.6	4.40	17.4
Cation exchange capacity (CEC).

Table 2 Quantities of nitrogen and carbon applied in crop residues in each treatment

Treatment (plot name)	N (%)	C (%)	C:N	N application (g m^−2)	C application (g m^−2)
Onion leaf	3.39	39.3	11.6	9.30	108
Soybean stem & leaf	2.98	43.3	14.5	11.6	168
Rice straw	0.63	39.2	62.3	1.76	110
Wheat straw	0.37	40.5	110	1.28	141
No residue	0	0	–	0	0

with increasing C:N ratio of residues. As views on the effect of crop residue quality on N₂O emissions vary, we investigated these effects under field conditions.

MATERIALS AND METHODS

Site description

We selected fallow land with Gray Lowland (GL) soil in Mikasa, Hokkaido, Japan (43°14.4′ N, 141°50′ E). Onion had been cultivated in the field for more than 10 years, and the field was fallow for 10 months before our experiments. Mean annual temperature and precipitation are 7.4°C and 1,212 mm, respectively, and the soil does not freeze in winter. At a depth of 0–10 cm, the soil pH (H₂O) was 5.8 and the cation exchange capacity (CEC) was 25.5 cmolc kg⁻¹. Soil C and N concentrations were 32.1 and 2.8 g kg⁻¹, respectively, and sand, silt and clay comprised 12.1%, 51.2% and 36.7%, respectively, at a depth of 0–10 cm (Table 1).

Experimental plots

We set up five experimental plots with four types of crop residue, onion leaf (OL), soybean stem and leaf (SSL), rice straw (RS) and wheat straw (WS), and no residue (NR). Each residue was dried completely at 70°C and cut into 5-cm pieces. The C:N ratios were 11.6, 14.5, 62.3 and 110, respectively. From the results of a questionnaire survey of farmer practices, we estimated the residue N application rate to each field. Then, the appropriate application rates, which did not exceed the maximum application rate, were determined. The residue C or N application rates were 108, 168, 110, 141 and 0 g C m⁻² and 9.3, 11.6, 1.76, 1.28 and 0 g N m⁻² (Table 2).

We set up the experimental plots on 1 September 2005. We dug out all soil to a depth of 5 cm from each plot (1 m × 1 m) and mixed it thoroughly. We divided the soil into five equal parts by weight, and mixed one residue into each part (leaving the NR unamended). We placed root-resistant, water-permeable sheets (Toyobo BKS9812) at the bottom of each hole to avoid mixing the treated soil and subsoil, and then laid each mixed soil onto the sheet. The saturated hydraulic conductivity of the sheet (4.8 × 10⁻⁶ m s⁻¹) was higher than that of the soil in the same field (Hu et al. 2002). Finally, we set up three bases in each plot for the gas sampling chambers (Fig. 1). Each base measured 20 cm in diameter and was made of stainless steel, and the upper part had a slight depression (1 cm deep, 1 cm wide). The lower part (3 cm deep) remained in the soil throughout the experimental period (Fig. 1).

Measuring N₂O, CO₂ and NO fluxes and estimating N₂O and CO₂ emissions

N₂O, CO₂ and NO fluxes from each plot were measured using a closed chamber method (Kusa et al. 2002). We sampled three times in the first week (2, 5 and 8 September), two times in the next week and a half (13 and 18 September), and then once per week until 30 October (six times). The gas sampling chamber was 20 cm wide and 25 cm high, and was made of white-painted stainless steel (Fig. 1). The cover, made of white acryl, was fitted with a sample collector, a pressure-regulating bag, and a Tedlar sample bag (0.5 L). Gas samples were withdrawn from the sample collector by syringe into the Tedlar bag. The chamber was set on the base's depression, which was filled with water to prevent aeration during measurement. We took gas samples before

Figure 1  Outline of the chamber and base used in the closed chamber method.

closing the chamber (0 min) for analysis of N₂O, CO₂ and NO. After the chamber was closed, we took one gas sample for CO₂ after 6 min, and another for N₂O and NO after 15 min (Kusa et al. 2002; Nakano et al. 2004). All samples were taken in triplicate (one per base).

In the laboratory, the gas fluxes were calculated using a linear-recurrence method (Kusa et al. 2002). CO₂ was analyzed using an infrared CO₂ analyzer (Model ZFP5YA3I, Fuji Electric, Tokyo, Japan) within 12 h. The detection limit of CO₂ flux was ±3.7 mg C m⁻² h⁻¹. N₂O was analyzed using electron-capture detection gas chromatography (Model GC-14B, Shimadzu, Kyoto, Japan) within 1 month. The detection limit of N₂O flux was ±7.7 µg N m⁻² h⁻¹. NO was analyzed using a chemiluminescence nitrogen oxide analyzer (Model 265P, Kimoto Electric, Osaka, Japan) within 12 h. The detection limit of NO flux was ±0.99 µg N m⁻² h⁻¹.

N₂O and CO₂ emissions were calculated using the following equation: , where F _i is the N₂O or CO₂ flux at the ith sampling (N₂O: µg N m⁻² h⁻¹; CO₂: mg C m⁻² h⁻¹) and d _i denotes the day of the ith sampling.

Calculation of the emission fractions of N₂O (EF_N2O) and CO₂ (EF_CO2 )

We assumed that in each treatment the N₂O and CO₂ emissions originated from the N or C in the crop residues and soil organic matter. Therefore, we calculated the emission fractions of N₂O (EF_N2O) and CO₂ (EF_CO2 ) as:

Other measurements and soil sampling

Soil temperature at a depth of 5 cm was measured during each gas flux measurement. Soil samples were taken from 0–5 cm at three points in each plot. After measurement of water content, each sample was extracted with distilled water for measurement of the NO⁻ ₃-N and water-soluble organic carbon (WSOC) concentrations and with 2 mol L⁻¹ KCl for measurement of the NH⁺ ₄-N concentration. NO⁻ ₃-N and WSOC concentrations were analyzed using ion chromatography (QIC analyzer, Dionex Japan, Osaka, Japan) and a Total organic carbon (TOC) analyzer (Model TOC-5000A, Shimadzu, Kyoto, Japan), respectively. The NH⁺ ₄-N concentration was analyzed using colorimetry with indophenol-blue (Uvmini-1240, Shimadzu, Kyoto, Japan). Three replicate undisturbed soil samples (0–5 cm) were also taken with a steel corer (100 mL) on 1 September and 30 October and analyzed for dry bulk density. The dry bulk density increased by 30% during that interval owing to soil consolidation after setting the treatments. We assumed that the change was linear, and calculated water-filled pore space (WFPS) from the interpolated dry bulk density and water content. Daily precipitation data were obtained from the local Iwamizawa weather station (43°12.6′ N, 141°47.3′ E).

RESULTS

N₂O and CO₂ fluxes

Figure 2a,b shows N₂O and CO₂ fluxes over the study period. N₂O fluxes in OL (455 µg N m⁻² h⁻¹) and SSL (567 µg N m⁻² h⁻¹) reached their maximum on the third day (4 September) and then gradually decreased. However, in RS, WS and NR, N₂O fluxes were close to zero throughout the study period. CO₂ fluxes in OL (506 mg C m⁻² h⁻¹) and SSL (413 mg C m⁻² h⁻¹) also reached a maximum on the third day and then gradually decreased.

There were positive relationships between N₂O and CO₂ fluxes in all plots except for WS (Fig. 3a,b). There were strong correlations in OL (n = 33, r ² = 0.81, P < 0.001) and SSL (n = 33, r ² = 0.89, P < 0.001).

Soil temperature, WFPS, soil NH⁺ ₄-N and NO⁻ ₃-N concentrations, WSOC and NO flux

Figure 2c–h shows the variation in soil temperature, WFPS, soil NH⁺ ₄-N, NO⁻ ₃-N, WSOC and NO flux, and Fig. 4a–j shows the relationship of N₂O and CO₂ fluxes with soil temperature, WFPS and soil NH⁺ ₄-N, NO⁻ ₃-N

Figure 2  Variations in (a) N2O flux, (b) CO2 flux, (c) soil temperature, (d) water-filled pore space (WFPS), (e) soil -N, (f) -N, (g) water-soluble organic carbon (WSOC) and (h) NO flux from plots treated with onion leaf (OL), soybean stem and leaf (SSL), rice straw (RS), wheat straw (WS) or no residue (NR). SD, standard deviation.

and WSOC. Soil temperature gradually decreased from 24.8°C to 10.3°C during the study period and was not correlated with N₂O fluxes in any plot. Only in the RS plot was there a positive correlation between soil temperature and CO₂ flux (n = 11, r ² = 0.44, P < 0.05). WFPS gradually increased from 25% (on 2 September) to 60% (on 30 October) and did not correlate with N₂O or CO₂ fluxes in any plot. Soil NH⁺ ₄-N concentration was at a maximum in OL (23.4 mg N m⁻²) and SSL (40.5 mg N m⁻²) on the third day. In RS, WS and NR, soil NH⁺ ₄-N was lower than in OL and SSL throughout the study period, peaking at 15.4, 13.6 and 15.5 mg N m⁻², respectively, from 4 to 14 October. There were significant positive correlations between soil NH⁺ ₄-N and N₂O flux in SSL (n = 11, r ² = 0.54, P < 0.01) and NR (n = 11,

Figure 3  Relationships between N2O and CO2 fluxes from plots treated with (a) onion leaf (OL) and soybean stem and leaf (SSL) and (b) rice straw (RS), wheat straw (WS) or no residue (NR).

r ² = 0.46, P < 0.05), and with CO₂ flux in SSL (n = 11, r ² = 0.46, P < 0.05). The maximum soil NO⁻ ₃ concentrations were 53.3 mg N m⁻² on 15 September in OL and 80 mg N m⁻² on 6 September in SSL. Soil NO⁻ ₃-N peaked after soil NH⁺ ₄-N in OL and SSL. In RS, WS and NR, soil NO⁻ ₃-N remained consistently low. Moreover, soil NO⁻ ₃-N concentration in RS and WS was lower than in NR. There were no relationships between soil NO⁻ ₃-N concentration and N₂O or CO₂ flux. Soil WSOC concentrations gradually decreased throughout the study period in all plots. There was a significant positive correlation between soil WSOC and CO₂ flux in WS (n = 11, r ² = 0.37, P < 0.05). The peaks of NO flux were 14.3 µg N m⁻² h⁻¹ in OL (day 14, 15 September), 15.0 µg N m⁻² h⁻¹ in SSL, and 4.48 µg N m⁻² h⁻¹ in RS (both day 5, 6 September). In contrast, in WS and NR, the NO fluxes peaked on the first day (1.57 and 2.85 µg N m⁻² h⁻¹, respectively), and remained low throughout the study period. When the N₂O flux reached its peak on September 4 in OL and SSL, the ratios of N₂O to NO flux (N₂O-N/NO-N) became 99.4 and 134, respectively (Figs 2,5).

Figure 4  Relationships between N2O (a–e) and CO2 (f–j) fluxes and soil temperature at 5 cm (a,f), water-filled pore space (WFPS) (b,g), soil -N (c,h), -N (d,i) and water-soluble organic carbon (WSOC) (e,j) in plots treated with onion leaf (OL), soybean stem and leaf (SSL), rice straw (RS), wheat straw (WS) or no residue (NR).

Table 3 Measured N2O and CO2 emissions, and estimated EF_N2O and EF_CO2 in five treatments

Treatment (plot name)	N2O emission (mg N m^−2)		CO2 emission (g C m^−2)		EFN2O (%)	EFCO2 (%)
	Mean	SD	Mean	SD
Onion leaf	84.5 a	28.8	98.0 a	10.6	0.86	45.4
Soybean straw & leaf	89.1 a	12.6	93.6 a	3.59	0.73	26.6
Rice straw	0.91 b	1.47	42.7 b	5.38	−0.18	−5.77
Wheat straw	−1.44 b	1.02	56.2 b	10.5	−0.43	5.08
No residue	4.11 b	2.17	49.0 b	3.87
Different letters indicate significant differences (P < 0.01). SD, standard deviation.

Figure 5  Relationships between N2O flux and N2O-N/NO-N in plots treated with onion leaf (OL), soybean stem and leaf (SSL), rice straw (RS), wheat straw (WS) and no residue (NR).

N₂O and CO₂ emissions and EF_N2O and EF_CO2

Table 3 shows the N₂O and CO₂ emissions and the estimated EF_N2O and EF_CO2 in each plot. N₂O and CO₂ emissions from each plot ranged from −1.44 to 89.1 mg N m⁻² per period and from 42.7 to 98.0 g C m⁻² per period. Both emissions in OL and SSL were larger than those in the other plots. There was a significant positive correlation between N₂O emission and crop residue N (y = 8.75x − 6.41, n = 5, r ² = 0.96, P < 0.01). Moreover, N₂O and CO₂ emissions increased as the C:N ratio decreased, but not significantly (N₂O: y = −0.99x + 92.2, n = 4, r ² = 0.83, P = 0.09; CO₂: y = −22.8 ln [x] + 152, n = 4, r ² = 0.84, P = 0.09). There were no relationships between CO₂ emission and C application rate (data not shown).

Emission fractions were −0.43% to 0.86% for N₂O and −5.77% to 39.9% for CO₂. EF_N2O and EF_CO2 were higher in OL and SSL than in the other plots. There was a negative logarithmic relationship between EF_N2O and

Figure 6  Relationships between (a) and (b) and crop residue C:N ratio.

residue C:N ratio (Fig. 6a, EF_N2O: y = −0.59 ln [x] + 2.30, n = 4, r ² = 0.99, P < 0.01), and between EF_CO2 and residue C:N ratio (Fig. 6b EF_CO2 : y = −18.3 ln [x] + 81.7, n = 4, r ² = 0.78, P = 0.11).

DISCUSSION

N₂O flux

N₂O fluxes increased greatly soon after the application of onion leaf and soybean residues, which had low C:N ratios. Similarly, CO₂ fluxes and soil NH⁺ ₄-N concentrations increased after the application of those residues, then soil NO⁻ ₃-N concentrations increased. The lack of change in N₂O and CO₂ fluxes in the NR suggests that onion leaf and soybean residues decomposed rapidly and nitrification took place. Aulakh et al. (1991) and Huang et al. (2004) reported that N₂O and CO₂ fluxes increased more after the application of residues with low C:N ratios compared with high C:N ratios. Vigil and Kissel (1991) reported that crop residue C:N ratio had a significant negative relationship to the rate of mineralization of residues. Eiland et al. (2001) reported that residues with a high C:N ratio posed a nutrient limitation on decomposers during the early stage of decomposition. We assume that residues with a lower C:N ratio were more easily mineralized and the probability of nitrification and denitrification was high, so N₂O was easily produced. Bouwman (1990) summarized the reports written by Anderson and Levine (1986) and Lipschultz et al. (1981), and reported that N₂O-N/NO-N was < 1 during nitrification and > 100 during denitrification. Because N₂O-N/NO-N increased with increasing N₂O flux (Fig. 5), we suggest that N₂O was mainly produced by denitrification in the OL and SSL (Bouwman 1990; Kusa et al. 2002, 2006). The production of N₂O during denitrification could have been enhanced by the supply of N and C from the decomposing onion leaf and soybean residues, which had low C:N ratios.

N₂O emission and EF_N2O

We found positive relationships between N₂O emission and crop residue N and between N₂O and CO₂ fluxes, but a negative logarithmic relationship between CO₂ emission and residue C:N ratio. Therefore, N₂O emission could have been affected by the decomposition of residues. N₂O emissions have been shown to increase with increasing fertilizer N application rate (Groenigen et al. 2004; Smith et al. 1998). In the present study, onion leaf and soybean residues, with more N than the other residues, could have been easily decomposed on account of their low C:N ratios. For that reason, more N was supplied to the soil and N₂O producers.

EF_N2O increased as the C:N ratio of residues decreased. Millar and Baggs (2005) applied six crop residues with the same amount of N and found significant negative correlations between N₂O emission and the residue's soluble C:N ratio. Akiyama and Tsuruta (2003) also reported a negative logarithmic relationship between N₂O emission and the C:N ratio of applied organic matter in an Andisol field. This relationship between EF_N2O and residue C:N ratio might arise because a lower C:N ratio makes the residue more easily decomposed (Eiland et al. 2001; Vigil & Kissel 1991). However, Huang et al. (2004) reported that the EF_N2O of residues was positively correlated with the residue C:N ratio, and suggested that most of the N₂O was produced by denitrification. The soil C concentration Huang et al. (2004) used was 21 g kg⁻¹, and sand and clay contents were 2.1% and 51%, respectively (they did not show WFPS). Therefore, the soil of their study was more clayey and had a lower C concentration than the soil used in our study. Groffman and Tiedje (1989) reported that clay loam soil compared with loam or sandy soils showed a high denitrification rate. Applying a C source, which is easily available, may increase the reduction of N₂O to N₂ (Weier et al. 1993). Thus, we believe that in their study, N and C supplied from residues with a low C:N ratio could have allowed active reduction of N₂O to N₂ and, thus, decreased the emission of N₂O from residues with a low C:N ratio. In our study, most of the N₂O was produced by denitrification in OL and SSL. Therefore, applying residues with a low C:N ratio to the soil may enhance the process of denitrification. However, the type of N (N₂O or N₂) emitted from the soil might be affected by soil moisture content or soil type (Cheng et al. 2004; Scholefield et al. 1997; Weier et al. 1993).

In our study, the EF_N2O values in RS and WS were negative. Moreover, EF_N2O became negative at a residue C:N ratio greater than 49.7. Soil NH⁺ ₄-N and NO⁻ ₃-N concentrations in RS and WS were lower than those in NR. Therefore, immobilization of NH⁺ ₄-N and NO⁻ ₃-N might have occurred in RS and WS. The application of residues with a high C:N ratio to the soil often immobilizes mineral N in soil (Yonebayashi 1997). Vigil and Kissel (1991) reported that the border of residue C:N ratio between mineralization and immobilization was approximately 40. However, our C:N ratios of rice and wheat straw were above 40 (62.3 and 110, respectively). In the regression model in Fig. 6a, EF_N2O became negative at C:N ratio > 49.7, indicating that the application of residues with high C:N ratio to the soil causes the immobilization of soil mineral N and makes EF_N2O negative.

Further study of the effects of soil texture and C and N concentrations in soil on the relationship between EF_N2O and residue C:N ratios is needed. However, it might be possible to estimate the EF_N2O of certain residues by using the relationship between EF_N2O and the residue C:N ratio.

Conclusion

N₂O emissions from crop residues applied to soil were highest from onion leaf and soybean stem and leaf. N₂O emission was higher in plots treated with residues with a low C:N ratio, possibly because residues with a low C:N ratio are more easily decomposed. There was a negative logarithmic relationship between EF_N2O and residue C:N ratio (y = −0.59 ln [x] + 2.30, n = 4, r ² = 0.99, P < 0.01), suggesting the possibility of estimating of residues from the residue C:N ratio.

ACKNOWLEDGMENTS

This work was financially supported by the Global Environmental Research Program of the Ministry of the Environment of Japan (No. S-2). The authors thank Akane Kagemoto and Yuu Hirose for help with collecting data, and Takeshi Morimoto for permission to use their onion fields.