Introduction

Nitrous oxide is involved in global warming and stratospheric ozone depletion (Forster et al. 2007). Agricultural soils are the major source of atmospheric N2O, accounting for 60% of the global N2O emission (IPCC 2007). The North China Plain, with an area of 35 million hectares and a population of 130 million, is one of the most important agricultural regions in China with the main soil type belonging to fluvo-aquic soil (Du et al. 2009). A double-cropping rotation consisting of winter wheat (Triticum aestivum L.) and summer maize (Zea mays L.) is the main cropping system adopted in this major grain-producing region (Cai et al. 2002a). Due to the rapid economic development and the less labor cost of applying mineral N fertilizer since 1980s, the application rate of urea has markedly increased (Gong et al. 2009). However, a low percentage of mineral fertilizer N such as urea, ammonium bicarbonate is taken up by crops (Zhu 1997; Meng et al. 2005). Previous field studies showed that the total gaseous loss (due to N2O, NH3, and N2) of N from urea and ammonium bicarbonate applied to maize and wheat was in a range of 4–54% and 22–30%, respectively (Zhang et al. 1992; Cai et al. 1998, 2002a; Fan and Zhu 1998). Nitrous oxide emission is an important pathway for gaseous N loss from wheat–maize rotation ecosystems with N fertilizer application in the North China Plain (Cai et al. 2002a, b; Meng et al. 2005).

Nitrous oxide in soil can be produced through nitrification (Inubushi et al. 1996), denitrification (Tiedje 1988; Firestone and Davidson 1989; Cavigelli and Robertson 2001), and dissimilatory nitrate reduction to ammonium (Silver et al. 2001). Some researchers have reported that nitrification (nitrifier denitrification) was the major source of N2O production under aerobic soil conditions (Wrage et al. 2001; Shaw et al. 2006; Ma et al. 2007), while others found denitrification to be the major source of N2O production under anaerobic soil conditions (Webster and Hopkins 1996; Zaman et al. 2008). In soil of the North China Plain, previous studies suggested that N2O mainly comes from nitrification process (Dong et al. 2001; Cai et al. 2002a; Meng et al. 2005), which is controlled primarily by soil moisture during the maize growing season and by soil temperature during the wheat growing season (Meng et al. 2005; Ding et al. 2007a). Nitrous oxide emission factor of chemical N fertilizer, estimated to be 0.70% or so in fulvo-aquic soil (Cai et al. 2002a; Ding et al. 2007a), is slightly higher than the national average value of 0.60% in China (Xing 1998). During the maize growing season, N2O emission factor (1.05–1.34%) is obviously higher than that of the wheat growing season (0.24–0.26%) despite the shorter growing period of maize than wheat (Ding et al. 2007a). Therefore, the reduction of N2O emission from soil during the maize growing season in the North China Plain should improve efficiency of N fertilizer and mitigate N2O emission from agricultural fields. One such approach is to apply urea fertilizer with urease inhibitor and nitrification inhibitor to delay urea hydrolysis and nitrification, respectively, which will in turn reduce N2O emission.

Nitrification inhibitor, dicyandiamide (DCD), by depressing the activities of ammonia monooxygenase enzyme in soil, can delay the oxidation of ammonia to hydroxylamine that is further oxidized into nitrite (NO −2 ) and then nitrate (NO −3 ) (Prasad and Power 1995), thus reducing N2O emissions directly by decreasing nitrification or indirectly by reducing the availability of NO −3 for denitrification and leaching. As a consequence, DCD can increase N use efficiency by increasing plant growth and N uptake (Asing et al. 2008; Li et al. 2009). Urease inhibitor, N-(n-butyl) thiophosphoric triamide (NBPT), can be quickly converted in soil to its more effective oxygen analog N-(n-butyl) phosphoric triamide (NBPTO), which then forms a tridentate ligand with the urease enzyme, slowing conversion of urea to NH +4 and then reducing N losses by ammonia volatilization (Manunza et al. 1999; Zaman et al. 2009). This inhibitor can also affect nitrification and subsequent denitrification rates (Zaman et al. 2008; Li et al. 2009). Therefore, the use of NBPT with urea-based fertilizer may be a potential management strategy to lower N2O emissions (Menéndez et al. 2009). A combined application of DCD and NBPT with urea can maintain N as NH +4 form in soil for a longer time with more opportunity of the fertilizer derived-N being taken up by plants, fixed by clays or immobilized to organic N with reduction of the gaseous N loss (Gioacchini et al. 2002; Xu et al. 2002; Blennerhassett et al. 2006; Zaman and Blennerhassett 2010).

Studies on N2O emissions as influenced by various urease and nitrification inhibitors have been carried out in different agro-ecosystems including rice paddy fields (Majumdar et al. 2000; Xu et al. 2002; Boeckx et al. 2005), wheat cultivations (Xu et al. 2000; Majumdar et al. 2002), vegetable farmings (Asing et al. 2008), grasslands (Merino et al. 2002; Menéndez et al. 2009) and drylands treated with urea (Majumdar et al. 2002), slurry (Merino et al. 2002), or animal urine patches (Di et al. 2007). To date, there is little information available as to the effect of nitrification and urease inhibitors on N2O emission from arable soil in the North China Plain. The objective of this study was to evaluate the influence of the application of urea with DCD, NBPT and NBPT + DCD on N2O emission during the maize growing season and to determine soil factors that may affect N2O emissions from soil in the North China Plain. Inhibitor rates were higher than those commonly used because their cost is cheap and nitrification rate was high in the studied soil.

Materials and methods

Study site and soil properties

In 2009, a field experiment was conducted in a maize field located at the Fengqiu State Key Agro-Ecological Experimental Station, Chinese Academy of Sciences, Henan Province, China (35°00′ N, 114°24′ E), a typical region of the North China Plain. The 30-year mean annual temperature was 13.9°C, with a range varying from −1.0°C in January to 27.2°C in July. The mean precipitation was 615 mm, two thirds of which fell between June and September. The soil, derived from alluvial sediments of the Yellow River and classified as aquic inceptisol, had a sandy loam texture (clay 11%, silt 72%, and sand 17%), bulk density 1.4 g cm−3, pH 8.6, total N 1.5 g N kg−1, and total organic C 12.0 g C kg−1. The cropping rotation system was wheat (T. aestivum L.) in winter and maize (Z. mays L.) in summer.

Experimental design and field management

The field experiment included four treatments: CK (the control, urea alone), NBPT (urea with NBPT), DCD (urea with DCD), and NBPT + DCD (urea with NBPT and DCD). The control without urea was not included since there was the limited area of the experimental field, and meanwhile, background N2O emission from the studied soil and N2O emission factor of N fertilizer applied during the maize growing season had been measured in the previous study (Ding et al. 2007a). Each treatment had three replicates. Each plot measured 5 × 3.5 m. All plots were arranged in a completely randomized block design. Urea was applied at 200 kg N ha−1, and NBPT (Hengshuo Chemical, Wuhan, China) and DCD (Sunnyfield Chemicals, Ningxia, China) were applied at a rate of 0.2% and 10% of the applied urea N (w/w), respectively. Inhibitor(s) and urea were thoroughly mixed and then incorporated into soil as the supplemental fertilizer using the drill-hole method on July 24, 2009 (day 46). Irrigation was carried out immediately after fertilization to reduce ammonia volatilization. No basal fertilizer was applied since no-tillage practice was adopted. Maize seeds were directly sown into each plot using the drill-hole method by hand on June 8, 2009 (day 0) after returning wheat straw to soil surface. The mature maize was harvested on September 29, 2009 (day 113).

Sampling and measurements

Gas samples were collected using the static closed-chamber method. Immediately after sowing, a PVC chamber base (30 × 30 × 10 cm) was inserted into the soil about 5 cm deep between maize rows in the center of each plot. The PVC chamber (30 × 30 × 15 cm) was placed on to the chamber base by inserting the flange of the chamber into the water trough at the upper end of the chamber base. The chamber was equipped with two ports: a small, silicon-sealed vent for sampling and a second port for measuring chamber temperature. Gas samples were taken twice a week in the morning between 09:00 and 12:00. During specific periods, such as after N fertilizer application and irrigation, gas samples were collected at an interval of 1 to 2 days. Each time, four samples of the chamber air were manually pulled into 50-ml syringes at 0, 10, 20, and 30 min after closure, and then injected into 20-ml pre-evacuated vials fitted with butyl rubber stoppers.

Nitrous oxide concentrations were analyzed on a gas chromatograph (Agilent 7890, Santa Clara, CA, USA) equipped with an electron capture detector. The interfering oxygen contained in the injected gas sample (1.0 ml) was separated by a pre-column (1 m) in combination with an analytical column (3 m). Both columns, packed with Porapak Q (80/100 mesh), were attached directly to the six-port valve to control the backflush. The temperatures of column oven, injector and detector were 40°C, 100°C, and 300°C, respectively. The flow rate of carrier gas (95% Argon +5% CH4) was 40 ml min−1. Nitrous oxide gas standards were supplied by the National Research Center for Certified Reference Materials, Beijing, China.

Soil temperatures at vertical depths of 5, 10, and 15 cm were measured with a digital thermometer (Model 2455, Yokogawa, Japan). Field soil moisture, expressed as water-filled pore space (WFPS), was measured directly using time domain reflectometry probes positioned vertically through the soil.

Three soil samples were taken from the 0–20 cm soil layer at different positions in each plot just after flux measurement using a 5-cm diameter stainless steel soil sampler and then all samples from each plot were thoroughly mixed to form a composite sample. After visible roots and litter materials were picked up, soil samples were passed through the 2 mm sieve and then extracted with 2 M KCl (soil/KCl suspension of 30:100 ratio) for 1 h on a rotary shaker. The extracted solutions were filtered and stored in a deep freezer (−18°C) until analysis. The NH +4 and NO −3 concentrations were measured using a colorimetric method and a segment flow analyzer of Skalar (SAN++, The Netherlands) (Ding et al. 2007b; Williams et al. 2009). Maize yields and straw weights in all treatments were determined at harvest by drying at 70°C for 72 h.

Statistical analysis

Nitrous oxide fluxes were calculated using the following equation:

$$ F = \rho \times (P/760) \times (V/A) \times (\Delta C/\Delta t) \times \left[ {{273}/({273} + T)} \right] $$

where F is the N2O flux (μg N2O–N m−2 h−1), ρ is the density of N2O at 0°C and 760 mm Hg (kg m−3), V is the chamber volume (m3), A is the area from which N2O was emitted into the chamber (m2), ∆C/∆t is the rate of N2O accumulation in the chamber (ppbv N2O–N h−1), T is the chamber air temperature in Celsius, and P is the air pressure of the experimental site (mm Hg). The altitude of the experimental site for this study is very close to that of sea surface, so P/760 ≈ 1 (Zheng et al. 2000). Total N2O emission was calculated by summing the products of the averaged two neighboring fluxes, multiplied by their interval time during the maize growing season (Ding et al. 2007a). Mean N2O flux during the maize growing period was calculated by total N2O emission divided by the experimental duration.

Statistical analyses were conducted using the SPSS (version 12.0) software package for Windows. Statistically significant differences were tested using analysis of variance (ANOVA) and least significant difference calculations at a 5% significant level. Flux data were log-transformed as needed to normalize the distributions prior to statistical analysis. Correlation and nonlinear regression analyses were used to test relationships between N2O fluxes and other factors.

Results

Maize yield and aboveground biomass

Maize grain yield was increased by 15.1%, 14.2%, and 8.4% in the NBPT, DCD, and NBPT + DCD treatments over urea alone, respectively, while the correspondent increasing rates for aboveground biomass was 18.4%, 19.2%, and 9.2%, respectively (Table 1). However, the difference among four treatments was not significant, indicating that both maize grain yield and total aboveground biomass were not significantly affected by the application of NBPT, DCD, and NBPT + DCD in the North China Plain.

Table 1 Effect of urea applied with NBPT, DCD, and combination of both NBPT and DCD on maize yield and aboveground biomass
Full size table

Soil temperature and moisture

Temporal variations of surface soil moisture (WFPS) and soil temperature at 5 cm depth are presented in Fig. 1. Soil moisture levels were highly variable during the maize growing season with WFPS values varying from 20.2% to 75.4%. Soil temperature at 5 cm depth remained above 15°C throughout the maize growing season and the highest soil temperature with the value of 30°C occurred at the end of June (day 19) when WFPS rapidly reduced to 20% (Fig. 1).

Fig. 1
figure 1

Temporal variations of soil WFPS (a) and temperature (b) at 5 cm depth during the maize growing season. Vertical bars denote the standard error of the mean (n = 3). The dash line arrow represents the time of irrigation

Full size image

Soil NH +4 and NO −3 concentrations

Temporal variations of NH +4 and NO −3 concentrations in soils during the maize growing season are shown in Fig. 2. Soil NO −3 contents ranged from 1.04 to 29.4 mg N kg−1 with a mean of 8.38 mg N kg−1, while NH +4 contents varied from 0.38 to 7.36 mg N kg−1 with an average of 3.15 mg N kg−1 during the experimental period. There was no significant difference in NH +4 and NO −3 concentrations among four treatments before urea application. However, soil NO −3 concentration dramatically increased after N fertilizer application. The levels of NO −3 contents in the NBPT + DCD treatment were relatively low from 1 week after N fertilizer application to maize harvest compared to those in other treatments but gradually increased in the NBPT treatment probably due to decomposition and then declining the effectiveness of NBPT. In contrast, there was no remarkable difference in soil NH +4 concentration among four treatments. Soil NH +4 concentrations were just slightly lower in the NBPT treatment and higher in the DCD and NBPT + DCD treatment than the values in the CK treatment possibly because the hill-hole application of N fertilizer may lower representative of soils sampled in each plot, which in turn masked the difference in soil NH +4 concentration among treatments as compared to the broadcast application (Zaman et al. 2008).

Fig. 2
figure 2

Temporal variations of NH +4 (a) and NO −3 (b) concentrations in soil during the maize growing season. Vertical bars denote the standard error of the mean (n = 3). The solid line arrow represents the time of N fertilizer application

Full size image

N2O emission

Seasonal variations of N2O flux during the maize growing season are presented in Fig. 3. Nitrous oxide fluxes from inhibitor(s)-incorporated treatments showed almost the same pattern of the control with urea alone. Two seasonal peaks of N2O fluxes were identified in each treatment during the maize growing season (Fig. 3). Soil moisture (WFPS) rose from 20% to 70% after irrigation on day 25 (Fig. 1). Accordingly, the first N2O flux peak with relatively smaller value of 42.5–66.5 μg N2O–N m−2 h−1 was observed. No basal fertilizer was applied before sowing in this study due to adoption of no-tillage practice, so there was no significant difference in N2O flux between treatments before fertilization on day 46. The second seasonal peak of N2O fluxes was detected after N fertilizer application, and varied from 110.4 to 497.0 μg N2O–N m−2 h−1 in four treatments. Comparing with the control with urea alone, the application of inhibitor(s) obviously reduced the second flux peaks by 33.2%, 41.4%, and 77.3% in the NBPT, DCD, and NBPT + DCD treatments, respectively. Nitrous oxide flux was negligible at the late stage of the maize growth in all treatments. Correlation analysis showed that natural logarithmic N2O flux was significantly correlated with soil WFPS in the DCD and NBPT + DCD treatments, but not with soil temperature and NH +4 or NO −3 concentrations in all treatments during the growing season (Table 2). When the four data points measured after N fertilizer application were excluded from the analysis, the natural logarithmic N2O fluxes were significantly related to WFPS in all treatments except the CK treatment. During the period from fertilization to harvest, N2O fluxes were significantly negatively correlated with soil NO −3 concentrations in the CK treatments, but not with soil NH +4 concentration, WFPS and temperature in all treatments.

Fig. 3
figure 3

Temporal variations of N2O flux from four treatments during the maize growing season. Vertical bars denote the standard error of the mean (n = 3). The dash and solid line arrows indicate the timing of irrigation and N fertilizer application, respectively

Full size image
Table 2 Relationships between ln [N2O flux] and soil WFPS or temperature (T) at 5 cm depth, soil NH +4 and NO −3 concentrations in the 0–20 cm soil layer during the whole growing season and during the period from fertilization to harvest
Full size table

Mean N2O fluxes in the CK, NBPT, DCD, and NBPT + DCD treatments were 28.4, 17.7, 17.3, and 15.1 μg N2O–N m−2 h−1 during the whole growing season, respectively, and 38.6, 21.8, 19.3, and 15.5 μg N2O–N m−2 h−1 during the period from fertilization to harvest, respectively (Table 3). Mean N2O flux in the CK treatment was significantly higher than the values in other treatments, indicating that the application of DCD, NBPT and NBPT + DCD played an important role in reducing N2O emission from the soil. The treatment with DCD or NBPT was significantly lower effective than that with NBPT + NBPT in inhibiting N2O evolution. One-way ANOVA analysis showed that there was no significant difference in N2O fluxes between NBPT and DCD treatments.

Table 3 Mean N2O flux and total N2O emission during the maize growing season and during the period from fertilization to harvest
Full size table

Cumulative N2O emissions from different treatments are also listed in Table 3. Total N2O emission from the CK, NBPT, DCD, and NBPT + DCD treatments was 0.77, 0.48, 0.47, and 0.41 kg N2O–N ha−1, respectively. It can be seen that total N2O emission in plot added with urea alone was significantly higher than those of inhibitor(s)-treated plots. Total N2O emission during the maize growing season was reduced by 37.7%, 39.0%, and 46.8% after the application of NBPT, DCD, and NBPT + DCD, respectively. The inhibition efficiency of combination of DCD and NBPT was significantly higher than that of the application of DCD or NBPT alone during the maize growing season.

Discussion

Effect of inhibitor(s) on N2O emission

Nitrous oxide emission is directly related to the amount of mineral N available in soil and the application of inhibitor(s) with urea can effectively regulate the NO −3 and NH +4 contents (Fig. 2; Li et al. 2009). After N fertilizer application on day 46, the NH +4 concentration slightly increased in the presence of DCD (Fig. 2) because DCD delayed the conversion of NH +4 to NO −3 (Vogeler et al. 2007; Asing et al. 2008; Zaman et al. 2008). Relatively low NH +4 content was found after NBPT application, indicating that urea hydrolysis was retarded by NBPT in studied soil. Therefore, urease inhibitor NBPT and nitrification inhibitor DCD, which retarded the hydrolysis of urea applied and slowing down the oxidation of NH +4 immediately following the hydrolysis of urea, respectively, jointly attenuated subsequent formation of N2O (Table 3).

In the tested soil of the North China Plain, most of N2O was emitted 2 weeks after fertilization during the maize growing season (Fig. 3; Meng et al. 2005; Ding et al. 2007a). The half-life of DCD was estimated to be 18–25 days at 20°C and the higher soil temperature measured during our experiment may have shortened the half-life of DCD (Di and Cameron 2004; Kelly et al. 2008). Therefore, the inhibitive effect of urease and nitrification inhibitors on N2O emission was higher in the initial than in the final periods (Li et al. 2009).

The combination of urease and nitrification inhibitors has already been shown to be a better approach to reduce N2O emission than urease or nitrification inhibitor application alone. Tiessen et al. (2006) found that the presence of NBPT and DCD suppressed the nitrification and increased the recovery of NH +4 -forming N fertilizer (fall-banded urea) in a 2-year field study. Urea applied with NBPT and DCD reduced N2O emissions by up to 37.0% over urea alone in a 3-month field experiment (Zaman et al. 2008). Zaman et al. (2009) reported that the combination of NBPT and DCD was more effective in reducing N2O emissions from urine in grazed pastures and this was attributed to the limited availability of NH +4 by NBPT followed by NH +4 retention by DCD. In the present study, we have found that the application of NBPT + DCD significantly reduced N2O emission by 46.8% over urea alone, which was higher than the 37.7% for NBPT and 39.0% for DCD (Table 3). Our study clearly showed that the combination of NBPT and DCD with urea seems promising in mitigating N2O emission from arable soils in the North China Plain (Table 3). It should be mentioned that the application rate of inhibitors in this study was rather high compared to the previous experiment (Zaman and Blennerhassett 2010).

Effect of soil factors on N2O emission

Nitrous oxide is mainly produced by microbial nitrification and denitrification in soil. Any factor such as fertilizer application, soil temperature, and soil moisture affecting N transformation can alter N2O emissions. The highest peak of N2O fluxes was found following fertilization since the amount of mineral N available in soil can directly affect N2O emission (Fig. 3; Meng et al. 2005; Dalal et al. 2009; Li et al. 2009). In the present study, we have found that ln [N2O flux] was positively, albeit not significantly, correlated with soil NH +4 concentration during the maize growing season or during the period from fertilization to harvest, but negatively correlated with soil NO −3 concentration (Table 2). This poor relationship was also observed by Barton et al (2008), and was attributed to the fact that no representative soil samples were collected during the experimental period especially following the greatest N2O fluxes as occurred in this study. Dobbie and Smith (2003) pointed out that denitrification in agricultural soils was N-limited only at NO −3 concentration <5 mg N kg−1 soil, which was smaller than the values measured in our study at most time. This suggests that denitrification activity was not determined by NO −3 concentration in tested soil. Previous studies have shown that nitrification was the major process producing N2O in soil of the North China Plain (Dong et al. 2001; Cai et al. 2002a, b; Ding et al. 2007a). Apparently, N2O production and emission was mainly controlled by NH +4 concentration rather than NO -3 concentration in tested soil.

Soil moisture affects both aerobic and anaerobic conditions that will favor the occurrence of nitrification or denitrification (Menéndez et al. 2009). During the maize growing season, there is a significant positive relationship between ln [N2O flux] and WFPS in the DCD and NBPT + DCD treatments (Table 2). When N2O fluxes measured during the 11-day peak emission period after fertilization from days 46 to 57 were excluded from analysis, the natural logarithmic N2O flux in all treatments except the CK treatment was significantly correlated with WFPS (P < 0.05). Barton et al. (2008) observed a similar correlation in a cropped soil in a semiarid climate, Australia. This confirms that soil moisture is a key factor affecting N2O emission during the maize growing season excluding the 11-day peak emissions after N fertilizer application, and soil N source plays an important role in N2O evolution immediately after fertilization (Meng et al. 2005; Abdalla et al. 2010).

Granli and Bøckman (1994) suggested that the primary source of N2O is nitrification at WFPS levels ranging from 30% to 70% and denitrification at WFPS ranging from 70% to 90% in cultivated soils. Ding et al. (2007a) concluded that nitrification in tested soil occurred mainly at a WFPS of 45–60%, whereas denitrification happened primarily when WFPS exceeded 75%. In the present study, soil moisture was no more than 75% WFPS (Fig. 1). This further verified that N2O in this study was mainly produced by nitrification and soil WFPS controlled N2O emission by affecting the nitrification rate during the maize growing season in the North China Plain (Ding et al. 2007a).

Soil temperature in this study was close to the optimum value for N2O production, which ranges from 25°C to 40°C (Fig. 1; Granli and Bøckman 1994). No significant correlation between N2O flux and soil temperature occurred probably because soil temperature did not show great variability during the maize growing season (Table 2). This indicates that soil temperature was not a key factor causing temporal variation of N2O fluxes during the maize growing season (Meng et al. 2005).

Conclusions

The application rates of NBPT and DCD were higher than those commonly used because their cost is cheap and nitrification rate was high in the studied soil. Nitrous oxide emission was considerably reduced by the combined application of NBPT and DCD with urea during the maize growing season. A significant relationship between ln [N2O flux] and soil moisture was found during the maize growing season excluding the 11-day peak fluxes after fertilizer application, indicating that N2O flux was mainly affected by soil moisture. In tested soil, N2O flux was mainly controlled by NH +4 rather than NO −3 concentrations. The application of urea with NBPT, DCD, and NBPT + DCD reduced N2O emissions by 37.7%, 39.0%, and 46.8%, respectively, and increased the maize grain yield by 15.1%, 14.2%, and 8.4%, respectively, over urea alone. The combination of NBPT and DCD was the most effective in reducing N2O emissions, controlling urea hydrolysis, and keeping urea N as NH +4 . Based on our findings, the combination of DCD and NBPT together with urea would be an optimum agricultural strategy for reducing N2O emission and improving maize yield in soils of the North China Plain.