Nutrient Cycling in Agroecosystems 63: 187–195, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands.

187

Nitrogen losses from fertilizers applied to maize, wheat and rice in the North China Plain ∗

G. X. Cai1 , D. L. Chen2 , H. Ding3 , A. Pacholski4 , X. H. Fan1 & Z. L. Zhu1 1 Institute

of Soil Science, Chinese Academy of Sciences, P.O.Box 821, Nanjing, People’s Republic of China; Resource Management, Forestry and Amenity Horticulture, The University of Melbourne, Vic 3010, Australia; 3 Institute of soil and fertilizer, Fujian Academy of Agricultural Sciences, Fuzhou 350013, People’s Republic of China; 4 Institute of Geoecology, Braunschweig Technical University, Langer Kamp 19c, 38106 Braunschweig, Germany (∗ Corresponding author; e-mail: [email protected]) 2 School

Key words: ammonia volatilization, denitrification loss, maize, N fertilizer, 15 N, rice, wheat

Abstract Ammonia volatilization, denitrification loss and total nitrogen (N) loss (unaccounted-for N) have been investigated from N fertilizer applied to a calcareous sandy loam fluvo-aquic soil at Fengqiu in the North China Plain. Ammonia volatilization was measured by the micrometeorological mass balance method, denitrification by the acetylene inhibition – soil core incubation technique, and total N loss by 15 N-balance technique. Ammonia loss was an important pathway of N loss from N fertilizer applied to rice (30–39% of the applied N) and maize (11–48%), but less so for wheat (1–20%). The amounts of unaccounted-for fertilizer N were in the order of rice > maize > wheat. Deep placement greatly reduced ammonia volatilization and total N loss. Temperature, wind speed, and solar radiation (particular for rice), and source of N fertilizer also affect extent and pattern of ammonia loss. Denitrification (its major gas products are N2 and N2 O) usually was not a significant pathway of N loss from N fertilizer applied to maize and wheat. The amount of N2 O emission (N2 O is an intermediate product from both nitrification and denitrification) was comparable to denitrification loss for maize and wheat, and it was not significant in the economy of fertilizer N in agronomical terms, but it is of great concern for the environment.

Introduction The consumption of nitrogen (N) fertilizers in China accounts for more than one quarter of total N consumption in the world (FAO, 1998). However, the efficiency of fertilizer N use is generally low and this results not only in financial losses for farmers but also a detrimental impact on the environment. Nitrogen can be lost mainly through leaching, runoff, denitrification and ammonia volatilization. It is essential to understand the relative importance of each process of N loss in order to develop effective strategies of N fertilizer management. Such information is also valuable for the development of comprehensive N models to address issues of sustainable agriculture and environmental quality. Studies in China (Zhu, 1997) showed that gaseous N losses, via NH3 volatilization and de-

nitrification, are mainly responsible for the fertilizer N losses. Rice, wheat and maize are the most important food crops in China, accounting for 27%, 26% and 22% of the total areas of food crops, and 39%, 21% and 26% of the total food production in China, respectively (China Agricultural Yearbook, 1999). The North China Plain is one of the major areas for cereal production in China. The main soil type is a fluvo-aquic soil (Aquic cambisol) with wheat as the main crop in winter and maize or rice in summer. Recently, the rate of application of N fertilizers has increased rapidly, up to around 200 kg N ha−1 per crop. Previous field investigations (Zhang et al., 1992; Cai et al., 1998; Fan and Zhu, 1998) showed that the total unaccountedfor N from N fertilizers applied to maize and wheat was in the range of 4–54% and 22–30%, respectively.

188 The loss through ammonia volatilization was in the range of 12–30% of applied N for maize and 1–9% for wheat, while the corresponding apparent denitrification loss was 15–18% and 13–29%, respectively. However, the denitrification loss measured by the direct, high abundance 15 N labeled technique (Siegel et al., 1982), in a previous investigation was exceptionally low, less than 0.4% (Cai et al., 1998; Fan and Zhu, 1998). The objective of this study was to quantify gaseous N losses through ammonia volatilization and denitrification by measuring them independently in irrigated maize and wheat in the North China Plain. Denitrification was measured using a direct method (intact soil cores with C2 H2 inhibition) (Parkin et al., 1984; Ryden et al., 1987). In addition, due to increasing concerns about the emission of N2 O from agro-ecosystems, the emission fluxes of N2 O were also measured. This paper summarizes the results obtained during the field experiments from 1998 to 2000, and from the previous studies in rice (Zhu et al., 1989), maize (Zhang et al., 1992) and wheat (Cai et al., 1998), and reviews the pathways of N losses.

Materials and methods Experimental sites and design Ammonia volatilization, denitrification loss and total N loss were investigated after application of N fertilizers to maize and wheat at Fengqiu. The field experiments were carried out on farmers’ fields adjacent to the Fengqiu Agroecological Experimental Station, Chinese Academy of Sciences, Fengqiu County, Henan Province, China, in 1998, 1999 and 2000. The soil is classified as fluvo-aquic soil, a calcareous sandy loam with a pH in the top soil (0–15 cm) of 8.4– 8.8, organic matter content of 8.1–11.6 g kg−1 , total N content of 0.52–0.66 g kg−1, available (Olsen) P of 3.0–5.8 mg kg−1 , available K of 62–75 mg kg−1 , and a cation exchange capacity of 7.3–8.1 cmolc kg−1 . Maize was sown in early June (without ploughing) and harvested in mid September. Wheat was sown in early October and harvested in early June. Three maize experiments were conducted to compare the losses of urea-N (1) applied by the different application methods, surface broadcast (SB) and deep point placement (DP) or broadcasting followed by irrigation (BI), (2) applied at different growing stages, seedling and panicle initiation stages, and (3) applied in different years (different climatic conditions) (Table

1). Two wheat experiments were carried out to compare the losses of urea-N applied to wheat by two application methods, surface broadcast (SB) and deep placement (DP) or broadcasting followed by irrigation (BI); and two application times- basal and top dressing (Table 1). The third wheat experiment was carried out to measure denitrification loss (Table 1). There were three treatments in each experiment, control (without N fertilizer), and two application methods, except for the last wheat experiment. Measurement of ammonia volatilization Ammonia volatilization was measured by the ammonia sampler method, based on a micrometeorological mass balance technique (Denmead et al., 1977; Denmead, 1983; Leuning et al., 1985). Two circular plots 20–30 m in diameter were set up for two treatments. The distance between the centres of the two plots was at least 60 m. As the background, the surrounding area of about 1.2 ha did not receive any N fertilizer during the measurement. Background values were determined upwind of the treated circles. The ammonia samplers were mounted at five heights, 0.4, 0.8, 1.2, 1.6 and 2.0 m above the soil surface, at the centre of two fertilized circular plots and in the upwind background area. The mass of ammonia collected in the sampler by the oxalic acid coating was measured by an ammonia electrode [Orion Research Inc, model 9512, 1983]. The measurement was started immediately after urea application and stopped when the ammonia flux was negligible. The period of measurement was 10 – 15 days with 6–48 h sampling intervals. The net mean horizontal flux density, uc (product of mean wind speed, u, and net mean concentration of ammonia, c, during the sampling interval, t, was calculated from: uc = M/(At), where M is the mass of ammonia collected (the difference between the treatment and the background), and A is the effective cross sectional area of the sampler. The net vertical flux density of ammonia (ammonia volatilization from the fertilizer), F, is given by:
z F = 1/X uc dz, 0

where X is the fetch, the distance traveled by the wind over the fertilized area (i.e. the radius of the circle, 10– 15 m), and z is the height of the air layer affected by ammonia emission.

189 Table 1. Ammonia volatilization, Denitrification loss and N2 O emission from applied urea (% of applied N) Crop

Application time

Maize

3 weeks after sowing (28/6/1998) 6 weeks after sowing (19/7/1998) 5 weeks after sowing (12/7/1999) At sowing (11/10/1998)

Maize

Maize

Wheat

Application Rate (kg N ha−1 )

Application method

Measuring ammonia loss

Measuring denitrification loss

Ammonia loss

Denitrification loss

N2 O emission

75

BIa

0.8

1.3

SBb DPc

44 11

1.1 1.2

0.9 1.0

150

SB DP

48 12

11 1.4

1.7

26 2.3

2.0

1.9

120

SB DP

30/6–7/7 (27.8 ◦ C, 2.1 mm) 20/7–14/9 (26.1 ◦ C, 376 mm) 14/7–16/9 (24.8 ◦ C, 106 mm)

18

200

28/6–8/7 (27.7 ◦ C, 2.1 mm)f 19/7–30/7 (27.7 ◦ C, 120 mm) 12/7–24/7 (22.3 ◦ C, 4.6 mm) 11/10–23/1 0 (17 ◦ C, 3.3 mm) 9/3–24/3 (6 ◦ C, 18.3 mm)

0.9

0.3

0.8

0.6

SB Wheat

Wheat

5 months after sowing (9/3/1999) At sowing and 5 months after sowing [11/10/1999 and 5/3/2000]

100

BI

220

SB DP+BId Urea + OFe

2.0 0.6 15 12/101999– 26/5/2000 (8.2 ◦ C, 101 mm)

a Broadcast followed by irrigation. b Surface broadcast. c Deep placement. d Urea of 120 kg N ha−1 was applied by DP at sowing and 100 kg N ha−1 was applied by BI five months after sowing. e Compost at a rate of 3.5 t ha−1 was mixed with plowed soil layer as basal dressing in addition to the application of urea. f Mean air temperature and amount of rainfall during the measurement period.

Measurement of denitrification and N2 O emission Denitrification loss was measured using the intact soil cores-acetylene inhibition technique (Parkin et al.,1984; Ryden et al., 1987; Chen et al., 1996). The N2 O-N flux in the presence and absence of C2 H2 was calculated as denitrification loss and N2 O emission, respectively. The incubator was a 150 mm diameter and 150 mm deep gas-tight PVC vessel. The lid was fitted with a rubber gasket and a rubber septum for gas sampling. Another rubber septum was incorporated on the side near the bottom of the PVC vessel for flushing the incubation gas. After all fresh soil core samples were taken, the sealed incubation vessel was flushed with the incubation gas (10% C2 H2 and 90% air) for 3 min (2 l/min). The incubation vessels were kept for 24 h in a hole of the same depth adjacent to the study area, and covered with a thin layer of soil to maintain the same temperature. A 24-h incubation avoided the

complicating effect of temperature fluctuation during day and night. After 24 h of incubation, a 20 ml gas sample was taken and transferred to a pre-vacuumed 18 ml gas-tight glass bottle for analysis of N2 O by GC. There were normally two replicates for each sampling. Small soil cores (39 mm diameter and 150 mm deep) were normally taken for the incubation. In the case of deep point placement, however, it was difficult to take representative small cores, so in the maze experiment in 1998 large soil cores (140 mm diameter × 150 mm deep) were taken at the points of fertilizer placement. It was assumed that the effect of N fertilizer was confined to a sphere 140 mm diameter, and that a measurement at the point of placement represented the total denitrification loss from the fertilizer. In the maize experiment in 1999, three sets of small soil cores according to the distance to the point of fertilizer placement were taken, incubated, and analyzed separ-

190 ately. Denitrification loss was calculated based on the three measurements and their area proportion. The measurement of denitrification was started 1–2 days after urea application and lasted 9 days in the first maize experiment and around 60 days (from the fertilizer application to harvesting) in the second and third maize experiment. In the wheat experiment, the measurement of denitrification was monitored in the whole cropping season. Flux densities of N2 O during the intervals between 2 sampling events were estimated as the averages of the two adjacent flux densities. Total N loss measurements Total N loss was measured in the micro-plots where labeled fertilizer was applied. There were two treatments with 4 replicates. The rate and methods of application were the same as the corresponding treatments in the large circular areas. Micro-plots (steel frames, 60 cm × 50 cm × 60 cm high for maize, and plastic frames, 20 cm in diameter with 40 cm high for wheat). ) were pressed into soil to a depth of 55 for maize and 35 cm for wheat and left five cm above the soil surface. Plant and soil samples were taken at harvest. Different plant parts as well as soil samples at different depths were taken, mixed thoroughly, then sub-samples were analyzed for total N and N isotopic ratio. All the soil from 0 to 20 cm depth was excavated, mixed thoroughly, and sub-samples were taken, and five soil cores of 4 cm diameter were taken from the 20–50 cm and 50–80 cm depths, mixed and subsampled for analysis in the maize experiment in 1998. In order to get more representative samples and the information of the fertilizer N distribution in soil profile the soil from 0 to 60 cm was divided into six sections of 10 cm deep each and all soil in each section was excavated, and five soil cores were taken from the 60–90 cm depth in the maize experiment in 1999.

15 N-

Results and discussion Ammonia loss from fertilizer N applied with different application method Table 1 shows that ammonia loss from urea applied to maize and wheat varied from 0.6 to 96 kg N ha−1 , and accounted for 0.6 – 48% of the applied N. Such great difference resulted from the different methods of fertilizer application, the different crops and the different weather conditions. Ammonia volatilization was much higher from the maize than from the wheat, and much

greater from the surface broadcast (SB) than from the deep placement (DP) or from the broadcast N fertilizer immediately followed by irrigation (BI). It is well known that the main factors governing ammonia volatilization are ammoniacal N concentration, pH and temperature of the surface soil (upland field) or in floodwater (flooded field), and wind speed (Freney et al., 1983). Fertilizer application methods mainly affect the ammoniacal N concentration in surface soil or in floodwater. In the maize experiments, the ammonia losses were significantly lower, 11–18 (mean±standard error: 13.3±1.6)% of applied urea, by using the DP or BI (the traditional local practice in the region) compared to 25–48 (37.4±5.2)% of the urea N applied by SB application (Table 1). A previous study in the region (Zhang et al., 1992) showed similar results. The study conducted in a maize field in Argentina also showed that ammonia volatilization was significantly reduced, from 11.5 to 6.2% of applied N, when urea was incorporated into soil compared to the surface application (Palma et al., 1998). For wheat crop, the common practice is to apply N fertilizer twice, as basal and top dressing. For the application prior to sowing the fertilizers were broadcast on the soil surface followed by ploughing, resulting in incorporation of most of the fertilizer at a depth of 10–15 cm, thus, basal dressing, in fact, is a deep placement. Top-dressing is to broadcast N fertilizer immediately followed by irrigation. The latter technique is very effective in moving urea down to the sub-surface soil and results in lower NH4 + concentrations at the soil surface (Li et al., 1984). The ammonia losses from the wheat field by DP and BI were in the range of 0.6–2.3 (1.5±0.9)% of the applied N compared to 15– 20 (17.5±2.5)% by SB. Katyal et al. (1987) reported that grain yield of wheat was increased and loss of N was greatly reduced by applying urea just prior to irrigation compared to following irrigation on alkaline coarse textured soils of India. In Australia, when urea was applied to sunflower fields through furrow or flood fertigation (dissolving urea in the irrigation water), the losses of ammonia were low (Freney et al., 1985; Smith et al., 1989). From the number of field experiments conducted in the past 15 years in the region, we conclude that the methods of N fertilizer application significantly affected the extent of ammonia volatilization. The point deep placement to maize, basal dressing for wheat (broadcast followed by ploughing), and the broadcasting of urea to maize or wheat immediately followed by irrigation were effective in reducing ammonia loss.

191

Figure 1. Ammonia loss from N fertilizer applied to different crops by farmers’ methods at Fengqiu. Mean±standard error of ammonia loss (% of applied N) for rice:34.6±4.5, maize:13.2±1.6, wheat:3.1±1.9. Source of published data: Zhu et al. (1989), Zhang et al. (1992), Cai et al. (1998).

Figure 2. Ammonia loss from the different fertilizers applied by farmers’ methods in China. Mean±standard error of ammonia loss (% of applied N) from urea and NH4 CO3 applied to rice were 19.5±10.7 and 28.7±10.5, respectively, applied to wheat were 1±0.1 and 6.9±1.9, respectively. Source of published data: Cai et al. (1986), Zhu et al. (1989), Cai et al. (1995), Cai et al. (1998).

Ammonia loss from fertilizer N applied to different crops at Fengqiu Figure 1 summarizes ammonia volatilization from N fertilizers applied to rice (Zhu et al., 1989), maize (Table 1 and Zhang et al., 1992) and wheat (Table 1 and Cai et al., 1998) at Fengqiu. Results from N fertilizers (urea or ammonium bicarbonate), using local traditional application methods, clearly indicated that ammonia volatilization in the region varied greatly among the crops (Figure 1). The losses were significantly higher from the flooded rice, 30– 39 (34.6±4.5)% of applied N, compared to the maize, 11–18 (13.2±1.6)%, and wheat, 1–9 (3.1±1.9)%. Such difference was mainly caused by the different application methods adopted for different crops and the prevailing temperature and solar radiation (for rice) immediately after the N application. Comparing to the wheat and maize the traditional application method, surface broadcasting and incorporation, used by the farmers in flooded rice fields was not efficient because most of the applied N remains in the floodwater (Zhu et al., 1989). The N in the floodwater is strongly subject to ammonia volatilization. Apart from wind speed, temperature is a very important factor influencing ammonia volatilization. The high ammonia losses of fertilizer N in rice and maize are consistent with the 10–20 ◦ C higher average temperature at application than when the fertilizer was applied to wheat. However, in flooded soils the solar radiation is probably the most important factor influen-

cing NH3 volatilization when N fertilizer is applied at transplanting. The solar radiation was very high when applying N fertilizers to the flooded rice at Fengqiu, which significantly promoted algal growth and elevated floodwater pH up to 10.5 (Zhu et al., 1989). For instance, a larger amount of N, 40% of the applied urea N, was lost as ammonia after the urea was applied to an acid paddy soil under the strong sunshine conditions in the Yingtan region of Jiangxi Province (Cai et al., 1992). In contrast, much less ammonia loss, only 9% of the applied urea N, was observed in a similar soil under cloudy conditions in the Danyang region of Jiangsu Province (Cai et al., 1986). Wind speed is an important factor influencing ammonia volatilization. Fillery et al. (1984) found that ammonia loss between two sites in the Philippines varied greatly due to the difference in wind speeds. However, it seems that wind speed is not a limiting factor for ammonia volatilization in the rice-growing season in China. The key factor controlling ammonia loss seems to be floodwater pH, which is greatly affected by solar radiation. Timing of fertilizer application also affects ammonia loss. Humphreys et al. (1988) found that, ammonia loss from urea applied at tillering was 21% compared with only 3% when applied at panicle initiation. The differences resulted from differences in floodwater pH and wind speed due to the large crop canopy at panicle initiation.

192 et al., 1981; Fillery and De Datta, 1986). In contrast, ammonia volatilization from urea took longer to reach the peak, with much lower but sustained rates. These differences resulted primarily from the difference in ammoniacal N concentrations in floodwater between the treatments. The ammonium fertilizers provide ammonium ions instantly available for volatilization, while when urea was applied the ammoniacal N concentration in the floodwater depended on the rate of urea hydrolysis (Cai et al., 1986; Zhu et al., 1989). Denitrification loss

Figure 3. Fertilizer N remaining in soil profile after surface application followed by a simulated single irrigation of 20 mm water (laboratory soil column experiment).
Urea, NH4 CO3 . Unpublished data.

Ammonia loss from different N fertilizers Ammonia volatilization losses were generally higher from ammonium bicarbonate than from urea (Figure 2). In the two separate field experiments with wheat, 9% N of ammonium bicarbonate was lost as ammonia volatilization compared to only 1% N from urea at Zhangbei in Hebei province (Cai et al., 1995), and 5% versus 1% at Fengqiu in Henan province (Cai et al., 1998). Higher ammonia losses from ammonium bicarbonate than urea were also observed in the flooded rice field at Fengqiu in Henan province, with corresponding loss of 39% and 30% of the applied N, respectively (Zhu et al., 1989) and 18% versus 9% at Danyang in Jiangsu province (Cai et al., 1986). In the two wheat experiments, N fertilizer was applied by broadcasting immediately followed by irrigation. Such technique is very effective for moving down urea N but not ammonium bicarbonate to subsurface soil (Figure 3). For the flooded rice field, the higher ammonia loss was mainly due to the instant availability of ammonium and high pH associated with the ammonium bicarbonate treatment. Furthermore, the pattern of ammonia volatilization from the two different N sources in flooded rice was different (Cai et al., 1986; Zhu et al., 1989). For the ammonium bicarbonate, the maximum rates of ammonia volatilization occurred almost immediately after the application and then sharply declined. The similar pattern of ammonia loss was observed for ammonium sulfate applied to flooded rice (Freney

The denitrification losses from maize and wheat over the measuring period ranged from 0.8 to 2% of applied N (Table 1) except for the SB treatment in the second maize experiment where much higher denitrification was measured, 11% of N applied. One of the major reasons was that the soil moisture was higher in the second maize experiment due to more rainfall (Table 1). By using a highly enriched 15 N technique in the same region, Fan and Zhu (1998) concluded that denitrification was not an important pathway of N loss. However, denitrification might be an important pathway of N loss if N fertilizer was applied by surface broadcasting and if rainfall was also high as in the SB treatment in the second maize experiment in 1998. Deep placement was apparently the most efficient way to apply N fertilizer in the region because of the lower NH3 and denitrification loss (Table 1). Mosier et al. (1986) and Mahmood et al. (1998) also reported low denitrification losses from the deep placement in maize crops with urea application. Apparent denitrification losses from maize and wheat estimated by the difference between the total N loss and ammonia volatilization were usually much higher than the values by direct measurement (Table 2). The apparent denitrification is subject to aggregation of the inherent errors in measuring total loss and ammonia loss. It should point out that the acetylene inhibition-soil core technique might have underestimated the denitrification because the original soil structure, aeration, anaerobic micro-sites may have been disturbed. The diffusion of acetylene may be reduced by compaction caused during the sampling. But more importantly, some peak fluxes of denitrification might be missed following the rainfalls because the field was too wet to take soil cores. A greater discrepancy between the two methods was found in a rice experiment in China. Only 1.8% of the applied N as (N2 +N2 O)-15N was measured by the

193 Table 2. Denitrification loss, ammonia volatilization and total unaccounted-for N from applied urea (% of applied N) Crop

Year

Treatment

Total unaccounted -for N

Ammonia loss

Denitrification loss

Apparent denitrification lossh

Ricea

1986

Maize

1998

Maize

1999

Wheat Wheat Wheat

1998 1999 1999–2000

Urea ABb DPc SBd DP SB DP BIe DP + BIf

63 72 10 (28, 62)g 67 (24, 9) 27 (38, 35) 42 (31, 27) 18 (50,32) 18 (62,20) –

30 39 11 48 12 26 2.3 0.6 –

– – 1.2 11 1.4 2.0 – – 0.9

33 33 0 19 15 16 16 17

a Zhu et al. (1989). b Ammonium bicarbonate. c,d,e,f Same as in Table 1. g Figures in the parenthesis are 15 N recovered in the crops and soil, respectively. h Calculated by the difference between the total unaccounted-for N and ammonia volatilization.

direct method from urea applied to a flooded rice field at Fuyang, while the corresponding apparent denitrification loss was up to 41% (Cai et al., 1991). Large discrepancies between apparent and directly measured denitrification have also been reported from applied urea in Philippines, Thailand and Indonesia (Busesh and De Datta, 1990; De Datta et al., 1991). Factors for the discrepancies were reviewed by Buresh and De Datta (1990) including entrapment of denitrification products in flooded soil (Samson et al., 1990; Cai et al., 1991; Chen et al., 1998), effect of rice plants on (N2 +N2 O)-15 N flux (Mosier et al., 1990) and errors inherent in the methodology for determining 15 N balance. N2 O emissions N2 O emission (in the absence of C2 H2 ) was in the range of 0.3 – 1.9% of N applied to maize and wheat (Table 1) with an average of 1.1%. This was higher than the average N2 O emission from fertilizer N applied to upland crops in China of 0.6%, as estimated by Xing (1998). But it is similar to the world average estimated by IPCC (1996) of 1.25 ±1.0%. Although N2 O emission is not significant in the economy of fertilizer N in agricultural terms, it is an active greenhouse gas and also it damages the ozone layer. Agricultural practices such as application of N fertilizer, and biological N fixation usually enhance N2 O emission (Granli and Bockman, 1994; Denmead et al., 1998).

It is well known that the major gas products by denitrification are N2 and N2 O, but N2 O can also be derived from nitrification. In the absence of C2 H2 , nitrification and reduction of N2 O to N2 can take place, and the N2 O flux is derived from both nitrification and denitrification. In the presence of C2 H2 , nitrification and the reduction of N2 O to N2 are inhibited, and the N2 O flux can be considered as the measure of total denitrification loss. The similar quantities of N2 O emissions and denitrification in these experiments (Table 1) suggested that the nitrification of the applied N significantly contributed to N2 O production. However, the present study was unable to distinguish between the relative contribution of nitrification and denitrification to N2 O production, and this warrants further study. Total nitrogen losses (unaccounted-for N) Total unaccounted-for N, ammonia volatilization and denitrification losses at Fengqiu are summarised in Table 2. The apparent denitrification loss, estimated by the difference between the total unaccountedfor N and ammonia loss, was also included. Total unaccounted-for N measured by 15 N balance technique was in the range of 10 – 72% of the applied N (Table 2). Consistent with the ammonia loss, total unaccounted-for N was higher from the flooded rice than from upland crops, higher from the maize than the wheat, higher when the fertilizer applied by the surface broadcast (SB) than by deep placement (DP). The losses by ammonia volatilization from N fertilizer

194 applied to maize were much higher than the corresponding denitrification losses simultaneously measured by the direct method (Table 1). In the case of wheat ammonia volatilization and denitrification was comparable, both were small. Results suggest that ammonia volatilization is the most important pathway of N loss from N fertilizers applied to flooded rice and maize at Fengqiu, but it is much less important for wheat. Denitrification seems to be a minor process of N loss from N fertilizer applied to maize, wheat or rice by the farmers’ methods. Substantial studies consistently reveal negligible denitrification from submerged rice soils during period of rice growth due to the limited rate of nitrification (Buresh and De Datta, 1990; Buresh et al., 1991).

Conclusions The N losses through ammonia volatilization varied greatly between crops, application methods, types of N fertilizer and weather conditions. A substantial amount of N was lost though ammonia volatilization from flooded rice and maize. However, the NH3 loss from wheat was comparatively small. Denitrification loss usually was not a significant pathway of N loss from N fertilizer applied to maize, wheat or submerged rice during period of rice growth. Deep placement of urea greatly reduced both ammonia volatilization and denitrification losses.

Acknowledgements We thank Mr Jiang Qiao of the Institute of Soil Science, Chinese Academy of Sciences and Dr Marco Roelcke of the Braunschweig Technical University for their assistance. This study was funded by the National Natural Science Foundation of China (39790100), DFG (Ri269/42-1/2) and GTZ (VN 810.12.840) in Germany, and the Australian Centre for International Agricultural Research (Project LWR1/96/164).

References Buresh RJ & De Datta SK (1990) Denitrification losses from puddled rice soils in the tropics. Biol Fertil Soils 9: 1–13 Buresh RJ, De Datta SK, Samson MI Phongpan S, Snitwongse P, Fagi AM & Tejasarwana R (1991) Ninitrogen and nitrous oxide flux from urea basally applied to puddled rice soils. Soil Sci Soc Am J 55: 268–273

Cai GX, Cao YC, Yang NC, Lu YH, Zhuan LJ, Wang XZ & Zhu ZL (1991) Direct estimation of nitrogen gases emitted from flooded soils during denitrification of applied nitrogen. Pedosphere 1: 241–251 Cai GX, Fan XH, Yang Z & Zhu ZL (1998) Gaseous loss of nitrogen from fertilizers applied to wheat on a calcareous soil in North China Plain. Pedosphere 8: 45–52 Cai GX, Peng GH, Wang XZ, Zhu JW & Zhu ZL (1992) Ammonia volatilization from urea applied to acid paddy soil in southern China and its control. Pedosphere 2: 345–354 Cai GX, Zhu ZL, Li RG, Bian XJ & Li XH (1995) Nitrogen loss from fertilizers applied to spring wheat growing on a meadow chestnut soil in the north of China. Academic Periodical Abstracts of China 1( Supplement):94 Cai GX, Zhu ZL, Trevitt ACF, Freney JR & Simpson JR (1986) Nitrogen loss from bicarbonate and urea fertilizers applied to flooded rice. Fert Res 10: 203–215 Chen D, Chalk PM & Freney JR (1998) Nitrogen transformations in a flooded soil in the presence and absence of rice plants: 2. Denitrification. Nutrient Cycling in Agroecosystems 51: 269–279 Chen D, White RE, Chalk PM, Heng LK, Fisher R & Helyar KR (1996) Measurement of gaseous N losses from grazed pastures. Proceedings Australian and New Zealand National Soils Conference, ‘Raising the Profile’ Volume 2. Melbourne. pp 41–2 China Agricultural Yearbook, 1999 De Datta SK, Buresh RJ, Samson MI, Obcemea WN & Real JG (1991) Direct measurement of ammonia and denitrification fluxes from urea applied to rice. Soil Sci Soc Am J 55: 543–548 Denmead OT (1983) Micrometeorological methods for measuring gaseous losses of nitrogen in the field. In: Freney JR & Simpson JR (eds) Gaseous Loss of Nitrogen from Plant–Soil Systems, pp 133–157. Martinus Nijhoff/Dr W. Junk Publishers: The Hague Denmead OT, Leuning R, Griffith DWT, Harper LA & Freney JR (1998) Verifying current estimates of non-CO2 greenhouse gas emissions from animals, landfills and pastures with direct measurements. CSIRO Land and Water, Canberra. Final Report to: National Greenhouse Gas Inventory Committee Denmead OT, Simpson JR & Freney JR (1977) A direct field measurement of ammonia emission after injection of anhydrous ammonia. Soil Sci Soc Am J 41: 1000–1004 Fan XH & Zhu ZL (1998) Investigation on the fate of nitrogen fertilizer and its nitrification-denitrification loss in summer maize on calcareous fluvo-aquic soils. In: Fu JP & Xu FA (eds). Agricultural Researches on the Huang Huaihai Plain, pp 48–51 Meteorological publisher: Beijing FAO (1998) Fertilizer. FAO yearbook Vol. 48 Fillery IRP & De Datta SK (1986) Ammonia volatilization from nitrogen surces applied to rice fields: 1. Methodology, ammonia fluxes, and nitrogen-15 loss. Soil Sci Soc Am J 50: 80–86 Fillery IRP, Simpson JR & De Datta SK (1984) Influence of field environment and fertilizer management on ammonia loss from flooded soil. Soil Sci Soc Am J 48: 914–920 Freney JR, Denmead OT, Watanabe I & Craswell ET (1981) Ammonia and nitrous oxide losses following application of ammonium sulfate to flooded rice. Aust J Agric Res 32: 37–45 Freney JR, Simpson JR & Denmead OT (1983) Volatilization of ammonia. In: Freney JR & Simpson JR (eds) Gaseous Loss of Nitrogen from Plant–Soil Systems, pp 1–32. Martinus Nijhoff/Dr W. Junk Publishers: The Hague Freney JR, Simpson JR, Denmead OT, Muirhead WA & Leuning R (1985) Transformations and transfers of nitrogen after irrigating a cracking clay soil with a urea solution. Aust J Agric Res 36: 685–694

195 Granli T & Bockman OC (1994) Nitrous oxide from agriculture. Norwegian Journal of Agricultural Sciences, Supplement 12 Humphreys E, Freney JR, Muirhead WA, Denmead OT, Simpson JR, Leuning R, Trevitt ACF, Obcemea WN, Wetselaar R & Cai GX (1988) Loss of ammonia after appliction of urea at different times to dry-seeded, irrigated rice. Fert Res 16: 47–67 IPCC (1996) Nitrous oxide from agricultural soils and manure management. In: Intergovernmental Panel on Climate Change for National Greenhouse Gas Inventories, Chapter 4 pp 1–125. OECD, Paris Katyal JC, Singh Bijay, Vlek PLG & Buresh RJ (1987) Efficient nitrogen use as affected by urea application and irrigatin sequence. Soil Sci Soc Am J 51: 366–370 Leuning R, Denmead OT, Freney JR & Simpson JR (1985). A sampler for measuring atmospheric ammonia flux. Atmos Environ 19: 1117–1124 Li ZL, Li AR & Cao ZH (1984) Effect of application method on efficiency of nitrogen fertilizers applied to spring wheat growing on a calcareous soil. Soils (in Chinese) 6: 134–137 Mahmood T, Tahir GR, Malik KA & Shamsi SRA (1998) Denitrification losses from an irrigated sandy-clay loam under a wheatmaize cropping system receiving different fertilizer treatments. Biol Fertil Soils 26: 35–42 Mosier AR, Guenzi WD & Schweizer EE (1986) Soil losses of dinitrogen and nitrous oxide from irrigated crops in northeastern Colorado. Soil Sci Soc Am J 50: 344–348 Mosier AR, Mohanty SK, Bhadrachalam A & Chakravorti SP (1990) Evolution of dinitrogen and nitrous oxide from the soil to the atmosphere through rice plants. Biol Fertil Soils 9: 61–67 Parkin TB, Kaspar AJ, Sexstone AJ & Tiedje JM (1984) A gas-flow soil core method to measure field denitrification rates. Soil Biol Biochem 16: 323–330

Palma RM, Saubidet MI, Rimolo M & Utsumi J (1998) Nitrogen losses by volatilization in a corn crop with two tillage systems in the Argentine Pampa. Commun Soil Sci Plant Anal 29: 2865– 2879 Ryden JC, Skinner JH & Nixon DJ (1987) Soil core incubation system for the field measurement of denitrification using acetylene-inhibition. Soil Biol Biochem 19: 753–757 Samson MI, Buresh RJ & De Datta SK (1990) Evolution and soil entrapment of nitrogen gases formed by denitrification in flooded soils. Soil Sci Plant Nutr 36: 299–307 Siegel RS, Hauck RD & Kurtz LT (1982) Determination of 30 N2 and application to measurement of N2 evolution during denitrification. Soil Sci Soc Am J 46: 68–74 Smith CJ, Freney JR, Chalk PM, Galbally IE, McKenney DJ & Cai GX (1989) Fate of urea nitrogen applied in solution in furrows to sunflowers growing on a red-brown earth:transformations, losses and plant uptake. Aust J Agric Res 39: 793–806 Xing GX (1998) N2 O emission from cropland in China. Nutrient Cycling in Agroecosystems 52: 249–254 Zhang SL, Cai GX, Wang XZ, Xu YH, Zhu ZL & Freney JR (1992) Losses of urea-nitrogen applied to maize grown on a calcareous Fluvo-Aquic soil in north China plain. Pedosphere 2: 171–178 Zhu ZL (1997) Fate and management of fertilizer nitrogen in agroecosystems. In: Zhu ZL, Wen QX & Freney JR (eds) Nitrogen in Soils of China, pp 239–279. Kluwer Academic Publishers, Dordrecht, The Netherlands Zhu ZL, Cai GX, Simpson JR, Zhang SL, Jackson AV, Chen DL & Freney JR (1989) Process of nitrogen loss from fertilizers applied to flooded rice fields on a calcareous soil in north-central China. Fert. Res 18: 101–115