Karen E. Dobbie and Keith A. Smith. Institute of Ecology and ...... McTaggart I.P., Clayton H., Parker J., Swan L. and Smith K.A.. Clayton H., McTaggart I.P., Parker ...
Nutrient Cycling in Agroecosystems 67: 37–46, 2003. 2003 Kluwer Academic Publishers. Printed in the Netherlands.
37
Impact of different forms of N fertilizer on N 2 O emissions from intensive grassland Karen E. Dobbie* and Keith A. Smith Institute of Ecology and Resource Management, University of Edinburgh, Darwin Building, Mayfield Road, Edinburgh EH9 3 JG, UK; * Author for correspondence (e-mail: karen.dobbie@ ed.ac.uk) Received 5 March 2002; accepted in revised form 14 August 2002
Key words: Fertilizer, Grassland, Inhibitor, Nitrogen, Nitrous oxide, Soil
Abstract Nitrous oxide (N 2 O) emissions were measured over two years from an intensively managed grassland site in the UK. Emissions from ammonium nitrate (AN) and urea (UR) were compared to those from urea modified by various inhibitors (a nitrification inhibitor, UR(N), a urease inhibitor, UR(U), and both inhibitors together, SU), as well as a controlled release urea (CR). N 2 O fluxes varied through time and between treatments. The differences between the treatments were not consistent throughout the year. After the spring and early summer fertilizer applications, fluxes from AN plots were greater than fluxes from UR plots, e.g. the cumulative fluxes for one month after N application in June 1999 were 5.2 6 1.1 kg N 2 O-N ha 21 from the AN plots, compared to 1.4 6 1.0 kg N 2 O-N ha 21 from the UR plots. However, after the late summer application, there was no difference between the two treatments, e.g. cumulative fluxes for the month following N application in August 2000 were 3.3 6 0.7 kg N 2 O-N ha 21 from the AN plots and 2.9 6 1.1 kg N 2 O-N ha 21 from the UR plots. After all N applications, fluxes from the UR(N) plots were much less than those from either the AN or the UR plots, e.g. 0.2 6 0.1 kg N 2 O-N ha 21 in June 1999 and 1.1 6 0.3 kg N 2 O-N ha 21 in August 2000. Combining the results of this experiment with earlier work showed that there was a greater N 2 O emission response to rainfall around the time of fertilizer application in the AN plots than in the UR plots. It was concluded that there is scope for reducing N 2 O emissions from N-fertilized grassland by applying UR instead of AN to wet soils in cool conditions, e.g. when grass growth begins in spring. Applying UR with a nitrification inhibitor could cut emissions further.
Introduction Nitrous oxide (N 2 O) is environmentally important. It is one of the greenhouse gases whose emissions need to be quantified by countries who have signed the Kyoto Protocol and it also contributes to ozone depletion in the stratosphere. The atmospheric concentration of N 2 O has increased by 17% since 1750 and continues to increase (Houghton et al. 2001). Kroeze et al. (1999) estimated that global N 2 O emissions have increased from 10.7 Tg N y 21 in 1850 to 17.7 Tg N y 21 in 1994 and suggested that this was mainly caused by expansion and intensification of agriculture. Direct emissions from agricultural soils are believed to account for 2.1 Tg N y 21 (Mosier et al. 1998). Improving the estimates of anthropogenic N 2 O
emissions, including those from agricultural soils, and seeking measures to reduce them, are priorities. N 2 O is produced in soils by nitrification (the oxidation of ammonium (NH 41 ) to nitrate (NO 32 )) and denitrification (the reduction of NO 2 3 to dinitrogen gas (N 2 )) (Firestone and Davidson 1989). N 2 O is an intermediate product of both processes. The rates of these processes and of N 2 O production depend on the amount of NH 41 and NO 32 in the soil, as well as other factors including soil water-filled pore space (WFPS) (Davidson 1991; Dobbie and Smith 2001) and soil temperature (Keeney et al. 1979; Dobbie and Smith 2001). The addition of nitrogen (N) fertilizers to agricultural soils increases the potential for N 2 O emissions (Granli and Bøckman 1994). Bouwman (1996) found
38 a linear relationship between fertilizer N applied and N 2 O emission for the relatively small number of reported experiments of at least one year in duration. This relationship, giving an emission factor of 1.25 6 1% of the N applied, was subsequently adopted by the Intergovernmental Panel on Climate Change (IPCC) (IPCC 1995) as the ‘default emission factor’, which could be used to calculate emissions on the basis of the amount of N fertilizer used in a country or a region. The emission factor (EF) was taken to be independent of crop type and the chemical form of N used. However, several studies have since shown that annual EFs can be greatly affected by short-term weather patterns and can show consistent differences ´ between crop types (McTaggart et al. 1997; Henault et al. 1998; Kaiser et al. 1998; Dobbie et al. 1999; Goossens et al. 2001; Ruser et al. 2001). Consequently, the IPCC (IPCC 2001) acknowledges that data on direct soil emissions of N 2 O obtained in particular countries may provide a better basis for calculation than the default EF. The majority of these studies also conclude that year-round sampling is required to obtain a robust annual EF estimate, and thus support the judgement reached earlier by Bouwman (1996) when selecting the flux data to be used for his EF calculation. With regard to the impact of the chemical form of N applied, Eichner (1990) reported a connection between the magnitude of emissions and the type of N fertilizer applied. Later work in the UK by Clayton et al. (1997) and McTaggart et al. (1997), and in France ´ by Henault et al. (1998), suggested that there are significant differences in emissions between fertilizer forms under some conditions. Also, work on controlled-release fertilizers in Japan by Tsuruta et al. (1992) and Minami (1994), and in several countries with fertilizers containing nitrification inhibitors (Smith et al. 1997), has shown that emissions can be reduced by modifying the form of N applied. This paper describes an experiment on intensive grassland in which the N 2 O emissions from the two conventional N fertilizers most commonly used in the UK, ammonium nitrate (AN) and urea (UR), were compared with those from urea modified by various additives (a nitrification inhibitor, UR(N), a urease inhibitor, UR(U) and both inhibitors together, SU), as well as those from a controlled release form of urea (CR). Materials and methods The field experiment was carried out at Glencorse
Mains Farm, located on the Bush Estate, 11 km SSW of Edinburgh. The soil was an imperfectly drained gleysol (FAO-UNESCO classification) of clay loam texture. The crop was perennial ryegrass (Lolium perenne L.) managed as for silage production since 1991. The forms of N fertilizer and inhibitors used are shown in Table 1. Six replicates of one (AN) and three replicates of the other six treatments were laid out in a randomised block design. (The AN treatment was replicated six times, as it was also part of another experiment.) Each plot measured 4 m 3 2.4 m. N was applied at a rate of 100 kg ha 21 around the beginning of June and August 1999, April, June and August 2000 and April 2001. This provided two full years of study from June 1999 to June 2001 (Table 2). The AN, UR, SU and CR were all in granular form and were applied directly to the respective plot surfaces by hand. The Didin Fluid (nitrification inhibitor) and Agrotain (urease inhibitor) (see Table 1) were supplied in solution. The appropriate amount of urea for each plot was therefore dissolved in 16 l of water and the required amount of inhibitor added to the solution: 7.7 ml of Didin Fluid (corresponding to the recommended application rate of 8 l ha 21 ) or 1 ml of Agrotain (corresponding approximately to the recommended application of 5.2 l tonne 21 of urea). After thorough mixing, the solutions were applied using a watering can. The grass was cut around eight weeks after each fertilizer application and the cuttings were removed. Gas fluxes were measured using the closed chamber technique (Clayton et al. 1994; Dobbie et al. 1999). Cylindrical polypropylene chambers, 40 cm high and 20 cm diameter, were inserted to a depth of 5 cm in the soil. One chamber was inserted in each plot. The chambers remained in the soil and open to the atmosphere between samplings, except when removed for grass cutting and fertilizer application. The chambers were covered with gas-tight aluminium lids, normally for 40 min. At the end of the closure period, one gas sample was taken from each chamber through a sampling port into a 30 ml evacuated aluminium sampling tube (Scott et al. 1999). Ambient gas samples were also taken. Flux calculations were based on the assumption that there was a linear increase in N 2 O concentration with time in the closed chamber. Earlier investigations by our group, and by others in similar conditions, had confirmed this to be the case (Dobbie et al. 1999). Sampling was carried out generally between 10.00 and 12.00 A.M., daily for a week after fertilizer application, then every other day for another
39 Table 1. Forms of N fertilizer used in the experiment. Treatment
N form
AN UR UR(N) UR(U) SU CR CO 1
Modification
Ammonium nitrate Urea Urea Urea Urea Urea Unfertilized control
Nitrification inhibitor: Didin Fluid [DCD (dicyandiamide)] 1 2 Urease inhibitor: Agrotain [NBPT (N-(n-butyl) thiophosphoric triamide)] Combined nitrification and urease inhibitors: SuperU [DCD & NBPT] 2 Coating for controlled release 3
OMEX Agriculture, King’s Lynn, Norfolk, UK; 2 Agrotain International LLC, Montgomery, IN, USA; 3 Haifa Chemicals Ltd, Haifa, Israel.
week, then twice per week and finally weekly. Gas fluxes were measured on 120 occasions over the twoyear period. N 2 O concentrations were determined using a Unicam 610 Series gas chromatograph fitted with an electron capture detector and an automated sample injector system (Arah et al. 1994). Daily flux measurements for each treatment were calculated using the arithmetic mean of the three (or six) chambers and are quoted along with standard errors of the mean. Cumulative fluxes were calculated by plotting daily fluxes through time, interpolating linearly between them, and integrating the area under the curve. Cumulative standard errors were also determined in this way. Emission factors (EF) were calculated as the amount of N 2 O-N emitted as a percentage of the fertilizer N applied. The EFs from the AN and UR treatments were compared using a t-test as described in Zar (1999). Linear regression equations were compared by comparing the slopes of the regression lines, again as described in Zar (1999). Soil samples were taken using a 3 cm diameter gouge auger to a depth of 10 cm. One core was taken from each plot and bulked into treatment samples. Gravimetric soil water content was determined for each sampling occasion by the method described in
Klute (1986). Soil mineral nitrogen (NH 1 4 -N and NO 2 -N) content was measured by colorimetric analy3 sis of 1 M KCl extracts of field-moist soil, using a Bran and Luebbe AutoAnalyzer for NH 41 -N and a Perstorp automated flow analyzer (a flow-injection system) for NO 2 3 -N. These analyses were carried out on soil samples taken at least once a week during the growing season and less frequently thereafter. Soil temperature was measured at 10 cm depth using a digital thermometer. Soil bulk density was measured using the core method described in Klute (1986), allowing soil WFPS to be calculated from the soil water content values.
Results The N 2 O fluxes varied widely through time and between treatments (Figure 1). They peaked soon after N fertilizer application, then generally tailed off, remaining low until the next N application. Fluxes from the AN plots reached higher values than those from any other treatment immediately after fertilizer application, with peaks up to 560 6 57 g N 2 O-N ha 21 d 21 . At the other extreme, fluxes from the unfertilized control were consistently low throughout the experiment, averaging 0.88 6 0.10 g N 2 O-N ha 21 d 21 .
Table 2. Values of soil WFPS, temperature and NO 2 3 -N concentration at the time of each fertilizer application, and the average values for the month following N application. Fertilization date
16.06.99 10.08.99 06.04.00 30.05.00 15.08.00 26.03.01
Soil WFPS (%)
2 21 Mean soil NO 2 3 concentration (mg NO 3 -N kg dry soil) in the month following fertilisation
Soil temperature (8C)
At fertilizer application
Following month (mean)
At fertilizer application
Following month (mean)
AN
UR
UR(N)
UR(U)
SU
CR
CO
95 60 85 67 73 90
79 64 82 75 77 91
15.1 14.0 4.8 10.5 15.4 2.6
14.8 14.2 6.3 12.9 14.7 5.7
9.9 24.4 18.6 40.8 21.2 22.7
1.8 9.6 1.2 10.2 16.7 2.7
1.5 2.7 0.5 1.8 2.3 2.4
10.1 13.1 6.2 8.1 8.8 2.2
2.0 3.4 0.8 8.6 15.1 2.4
0.6 1.8 0.3 1.9 0.5 2.1
0.6 1.4 0.1 1.6 0.2 2.2
40
Figure 1. N 2 O emissions from the different N fertilizer treatments for 2000. (a) AN, ammonium nitrate, (b) UR, urea, (c) UR(N), urea with nitrification inhibitor, (d) UR(U), urea with urease inhibitor, (e) SU, urea with both nitrification and urease inhibitors, and (f) CR, controlled release urea. Arrows in (a) and (d) show dates of N application and are applicable to all other treatments.
Cumulative fluxes over the second field season were higher than over the first season in all treatments except UR(U) (Figure 2). The differences between treatments were not consistent throughout the year (Figure 3). After the April and June fertilizer applications, fluxes from the AN plots were greater than those from the UR plots; however, after the August
Figure 2. Annual N 2 O emissions from the different N fertilizer treatments. AN, ammonium nitrate, UR, urea, UR(N), urea with nitrification inhibitor, UR(U), urea with urease inhibitor, SU, urea with both nitrification and urease inhibitors, CR, controlled release urea, and CO, unfertilized control.
fertilizer application, there was no difference between the two treatments. After all three applications, fluxes from the UR(N) plots were much smaller than those from either the AN or the UR plots. Fluxes from UR(U) and SU were not substantially different from those from the untreated urea at any time (SU data not shown). The CR plots showed very small fluxes, together with poorer grass growth in the spring and early summer, but better grass growth in the autumn, than in the plots receiving conventional N fertilizers. In the autumn of 2000, fluxes of up to 100 g N 2 O-N ha 21 d 21 were observed from the CR plots, much greater than those from the other treatments at this time (Figure 1). The soil WFPS and temperature at the times of N addition, and the average values for the month after N was applied, are shown in Table 2. The average values for soil NO 2 3 -N concentration for the same period are also shown in Table 2. In the AN plots, soil NO 2 3 -N concentration was highest immediately after N application (up to 125 mg kg 21 dry soil). It then decreased over about a month to background levels. This is illustrated in Figure 4. Soil NO 2 3 -N concentrations in the month after N application in the AN plots were never , 5 mg kg 21 dry soil. In contrast, in June 1999, April 2000 and April
41
Figure 3. Cumulative N 2 O emissions for four weeks after N application for AN, UR, UR(N) and UR(U) treatments.
2001, soil NO 2 3 -N concentrations in the UR plots were always , 5 mg kg 21 dry soil. However, in August 1999 and June and August 2000, NO 2 3 -N values, although initially low, increased over two weeks and averaged $ 10 mg kg 21 dry soil for the month after N application (Figure 4). NO 32 -N concentrations in the UR(U) plots followed a similar pattern to UR plots and averaged . 5 mg kg 21 dry soil for the month after N application in all periods except April 2001, when they were , 5 mg
kg 21 dry soil. Plots fertilized with SU had soil NO 2 3 N concentrations , 5 mg kg 21 dry soil except in June and August 2000, when they showed a similar pattern to the UR plots. In the UR(N), CR and CO plots, NO 32 -N values rarely exceeded 5 mg kg 21 (Figure 4). Soil NH 41 -N concentrations were generally higher than soil NO 32 -N concentrations in all treatments, except in the AN plots where they were similar (data not shown). Concentrations were a maximum directly after N application and decreased through time. In 4
Figure 4. Soil NO 2 3 -N concentration following N application in August 2000 for AN, UR, UR(N), UR(U) and CO treatments.
42 out of the 6 periods studied, in the UR based plots NH 1 4 -N concentrations were in the order UR(N) . UR(U) . UR.
Discussion The results show very clearly that there is scope for reducing N 2 O emissions from N-fertilized grassland in moist environments, by selecting or modifying the form of N applied. The essential factor affected by this action was the soil NO 2 3 -N concentration, as the largest fluxes always occurred when this was high. The general impact of the addition of NO 2 3 -N to the soil on N 2 O fluxes was exemplified by the emissions from the AN plots following each N application in 2000. On each occasion soil NO 2 3 -N concentration was $ 10 mg kg 21 dry soil for the month after N application, and maximum fluxes exceeded 400 g N 2 O-N ha 21 d 21 . These high fluxes were associated with WFPS values which averaged $ 75% and soil temperatures which averaged ca. 6, 13 and 15 8C in April, June and August, respectively (Table 2). In contrast, in the CO plots, although soil WFPS and temperature were the same as in the AN plots, soil 21 NO 2 dry soil and fluxes 3 -N was always , 5 mg kg 21 21 were , 3 g N 2 O-N ha d . Figure 3 shows that substantial fluxes occurred from the AN plots in the month after every fertilizer application except that in August 1999. This particular month was drier than normal. Soil WFPS at the time of fertilizer application was 60% and exceeded 70% on only two sampling occasions in the following month. These data show that for N 2 O fluxes to exceed 100 g N 2 O-N ha 21 d 21 , it was necessary for soil WFPS to be . 65%, soil temperature to be . 4.5 8C, and soil 21 NO 2 dry soil. 3 -N concentration to be . 5 mg kg These observations provide support for the idea that there are threshold values for these variables, and if any one of them is below the threshold, large fluxes will not occur (Ryden and Lund 1980; Smith et al. 1998; Dobbie et al. 1999). The size of the fluxes from the AN plots, in conjunction with high WFPS values and soil NO 2 3 -N concentrations, suggests strongly that denitrification was the predominant source of N 2 O. In the UR plots no NO 32 -N was added to the soil, unlike the AN treatment, and overall this had a mitigating effect on N 2 O emissions. However, substantial fluxes did occur after some UR applications
(Figure 3). After hydrolysis of urea, nitrification produces NO 2 3 in the soil, emitting some N 2 O in the process, and this NO 32 can be subsequently denitrified if conditions are favourable, causing larger N 2 O fluxes. This mechanism could explain the pattern seen in the UR plots in June and August 2000. On these occasions, NO 2 3 -N concentrations were low immediately after N application, but increased over a couple of weeks to average concentrations of $ 10 mg kg 21 dry soil, accompanied by substantial N 2 O fluxes (Figure 1). However, in April 2000, when the average soil temperature was only 6 8C (compared with the 13 8C and 15 8C in June and August, respectively), soil NO 2 3 -N values and fluxes both remained low. The use of the nitrification inhibitor DCD with urea (UR(N)) proved to be an effective way of reducing emissions throughout the study. This confirms the earlier observations of McTaggart et al. (1997) and Velthof et al. (1997). Soil analytical data showed that the DCD in the UR(N) plots effectively depressed the nitrification reaction even in the warmer periods. The resulting low soil NO 2 concentrations would be 3 expected to limit denitrification, providing an explanation for the low N 2 O fluxes from this treatment (Figures 1 and 3; Table 2). In the UR(U) plots, the urease inhibitor NBPT proved ineffective as a means of reducing emissions. The rationale for its use was that slower hydrolysis of 1 urea to NH 1 4 , together with uptake of NH 4 by grass, would result in a generally reduced concentration of NH 1 4 in the soil. Thus less could potentially undergo nitrification and denitrification. However, in the 21 event, soil NO 2 dry soil for 3 -N averaged . 5 mg kg the month after N application in all periods except April 2001 and large N 2 O fluxes occurred at these times – with the exception of August 1999 when soil WFPS was low. The reasons are unknown. The SU treatment involved the use of both DCD and NBPT with urea, but gave rise to twice as much N 2 O than when DCD only was added. The plots emitted sizeable amounts of N 2 O in June and August 2000 (Figure 1), corresponding to soil NO 2 3 -N concentrations of . 5 mg kg 21 dry soil (Table 2). Again, the reasons are unknown. The controlled release urea fertilizer (CR) was designed for use at higher soil temperatures than we find in Scotland. Under Scottish conditions (mean soil temperature of 12.5 8C during the growing season), N release appeared to have been so slow that grass growth was adversely affected. The resulting continuation of N release in the autumn, when growth
43 was reduced by light and temperature limitations, would explain the relatively high N 2 O fluxes seen in the CR plots in the autumn of 2000 (Figure 1). However, there was no evidence of high NO 2 3 -N concentration in the soil. Combining the results of this experiment with data from Clayton et al. (1997) and Dobbie et al. (1999), and previously unpublished data of our own, gives us emission data from AN- and UR- treated grassland for seven field seasons at the Glencorse Mains site and for three years of experiments at other Scottish grassland sites. The results are compared in Figure 5a. Arable crops: winter wheat, potatoes and oilseed rape, for which emissions from AN were reported by Dobbie et al. (1999), also had UR treatments and the results are compared in Figure 5b. In the grassland soils, annual fluxes from UR plots were lower than those from AN plots in 6 out of the 13 sites / seasons studied, but were greater in 3 (Figure
Figure 5. Comparison between annual fluxes from AN- and URtreated plots in (a) grassland and (b) arable land.
5a). Average emission factors for AN-treated grassland were 2.75 6 0.56% and 2.12 6 0.44% for UR treated, and a t-test showed that these were not significantly different overall (p . 0.1). In the arable soils, again a t-test showed that there was no significant difference between the emissions from the two N forms, confirming the observations of McTaggart et al. (1997) from one field season on a soil under spring barley. The data for the Glencorse Mains site shows a clear contrast between emissions from AN and those from UR. In Figure 6, the emissions over the 4 weeks from the date of fertilizer application are plotted against the rainfall over a 4 week period beginning one week earlier, to take account of the impact of wet conditions already prevailing when N is applied. By curtailing the rainfall integration period in April 2000 by 2 days, to exclude 100 mm of rain falling in these 2 days (when the low concentration of soil NO 2 3 -N remaining nearly three weeks after application prevented any substantial flux response), both the AN and UR data points for this period fitted the general trend (Figure 6). This left one UR point for the unusually warm and wet August 1998 as an outlier. Excluding this point from the regression analysis, r 2 values of 0.624 and 0.721 were obtained for the AN and UR relationships with rainfall, respectively, with slopes that differed significantly (p , 0.01). Although the annual EFs for the AN- and URtreated grassland were not significantly different, the differences between the treatments were not consistent throughout the year. If seasonal fluxes are considered from the seven field seasons at the Glencorse Mains site, differences become clearer. In every year, after the spring N application, fluxes from the AN plots were higher than from the UR plots. After the late summer N application there was no substantial difference between the two N forms used. However, after the early summer application, in 5 out of 8 years, AN fluxes were higher than UR, while in the other 3, UR fluxes were the higher. This appears to be a response to climate. In spring in SE Scotland, soil temperatures are generally low, averaging around 6 8C in April. It is well established that at such low temperatures nitrification rates are low (e.g. Freney et al. 1979), but not too low to prevent denitrification, which is also promoted by the high soil WFPS values (e.g. Malhi et al. 1990). A NO 2 3 -containing fertilizer in such cool, wet conditions is therefore likely to give much higher N 2 O fluxes. In late summer, conditions tend to be warmer, allowing more nitrification to
44
Figure 6. Relationship between the cumulative flux in the four weeks after N fertilizer application and the amount of rainfall over the period from one week before to three weeks after N application. One outlier point for UR is not included.
occur, and therefore high fluxes are found from the urea plots as well. However, in early summer, climate is more variable. In the years where high emissions from AN were found, the rainfall in every case was higher than the average for the time of year, whereas in the years when the UR plots gave higher emissions, rainfall was lower than the average for the time of year. This suggests that, in Scottish conditions, if UR rather than AN is applied in spring, and then again in June if the soil is wet, fluxes should be lower than if AN is used throughout. Clayton et al. (1997) carried out an experiment for one year using UR on the first application followed by AN for the second and third applications. They showed that the total annual emission was not significantly different from the hypothetical emission calculated by adding the emissions accumulated by each N application separately. In other words, there was no ‘memory effect’ altering emissions from one application depending on the nature of the predecessor one. It was therefore possible to calculate that at this site, use of UR in the April application and AN in June and August would have reduced N 2 O emissions by 24% below those arising from the use of AN throughout. Use of UR in both April and also in June in the years when the June rainfall was higher than average for the time of year would have reduced emissions by 46% below those from AN use throughout. These findings are in keeping with those of Velthof et al. (1997), who found that a poorly drained sandy
soil in the Netherlands gave higher fluxes from NO 2 3 based fertilizers than from ammonium sulphate or UR in both April / May and June / July. They concluded 2 that the use of NH 1 4 -based instead of NO 3 -based fertilizers may greatly reduce N 2 O emissions from grassland in wet conditions. They also suggested that temperature could affect the relative uptake of NH 1 4 N and NO 32 -N by the sward, thereby affecting N 2 O emissions. On the basis of both our work and the Dutch work, there appears to be a real prospect for reducing N 2 O emissions from N-fertilized grassland, by substituting the use of UR for AN, at least on wet soils in spring when nitrification is slowed by lower temperatures. Use of N in the NH 1 4 form (or as UR) to mitigate N 2 O emissions is unlikely to cause agronomic problems in the British Isles. Watson and Adams (1986) suggested that under high rainfall conditions, early spring nitrogen should be applied in 2 the NH 1 4 form rather than the NO 3 form, to obtain more efficient N utilization and increased dry matter yield. Later, Watson et al. (1990) reviewed a number of comparisons of UR with AN and CAN (calcium ammonium nitrate), in the UK and Eire, and concluded that in general UR was as good as (C)AN for grass production in cool wet spring conditions, but less effective in the summer.
Conclusions These studies indicate that there is scope for reducing
45 N 2 O emissions from intensively managed grassland by applying UR instead of AN, particularly in cool conditions with wet soils. Applying UR with a nitrification inhibitor could cut emissions further. There is a strong case for a full-scale agronomic trial with UR and / or other NH 1 4 -based N fertilizers containing DCD, to investigate whether the N 2 O mitigation is achieved without any associated disadvantages in terms of crop yield and quality. The use of a urease inhibitor, a combination of urease and nitrification inhibitors, and a controlled release form of N had limited impacts on N 2 O emissions and therefore appear to be unsuitable as mitigation options in our conditions.
Acknowledgements This work was funded by the U.K. Department of Environment, Food and Rural Affairs, contracts CC0214 and CC0223. Technical assistance from Ed Robertson, Andy Gray, George Ritchie and John Parker is acknowledged.
References Arah J.R.M., Crichton I.J., Smith K.A., Clayton H. and Skiba U. 1994. Automated gas chromatographic analysis system for micrometeorological measurements of trace gas fluxes. J. Geophys. Res. 99: 16593–16598. Bouwman A.F. 1996. Direct emissions of nitrous oxide from agricultural soils. Nutr. Cycl. Agroecosyst. 46: 53–70. Clayton H., Arah J.R.M. and Smith K.A. 1994. Measurement of nitrous oxide emissions from fertilized grassland using closed chambers. J. Geophys. Res. 99: 16599–16607. Clayton H., McTaggart I.P., Parker J., Swan L. and Smith K.A. 1997. Nitrous oxide emissions from fertilised grassland: A 2year study of the effects of N fertiliser form and environmental conditions. Biol. Fertil. Soils 25: 252–260. Davidson E.A. 1991. Fluxes of nitrous oxide and nitric oxide from terrestrial ecosystems. In: Rogers J.E. and Whitman W.B. (eds), Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, and Halo-methanes. American Society for Microbiology, Washington, DC, pp. 219–235. Dobbie K.E., McTaggart I.P. and Smith K.A. 1999. Nitrous oxide emissions from intensive agricultural systems: variations between crops and seasons, key driving variables, and mean emission factors. J. Geophys. Res. 104: 26891–26899. Dobbie K.E. and Smith K.A. 2001. The effects of temperature, water filled pore space and land use on N 2 O emissions from an imperfectly drained gleysol. Eur. J. Soil Sci. 52: 667–673. Eichner M.J. 1990. Nitrous oxide emissions from fertilized soils: summary of available data. J. Environ. Qual. 19: 272–280. Firestone M.K. and Davidson E.A. 1989. Microbiological basis of NO and N 2 O production and consumption in soil. In: Andreae M.O. and Schimel D.S. (eds), Exchange of Trace Gases between
Terrestrial Ecosystems and the Atmosphere. John Wiley & Sons, Chichester, UK, pp. 7–21. Freney J.R., Denmead O.T. and Simpson J.R. 1979. Nitrous oxide emissions from soil at low moisture contents. Soil Biol. Biochem. 11: 167–173. Goossens A., De Visscher A., Boeckx P. and Van Cleemput O. 2001. Two-year field study on the emission of N 2 O from coarse and middle-textured Belgian soils with different land use. Nutr. Cycl. Agroecosyst. 60: 23–34. Granli T. and Bøckman O.C. 1994. Nitrous oxide from agriculture. Nor. J. Agri. Sci. Suppl. 12: 1–128. ´ Henault C., Devis X., Lucas J.L. and Germon J.C. 1998. Influence on different agricultural practices (type of crop, form of Nfertilizer) on soil nitrous oxide emissions. Biol. Fertil. Soils 27: 299–306. Houghton J.T., Ding Y., Griggs D.J., Noguer M., van der Linden P.J., Dai X. et al. (eds) 2001. Climate Change 2001: The Scientific Basis. Cambridge University Press, Cambridge, UK. IPCC (Intergovernmental Panel on Climate Change) 1995. IPCC Guidelines for National Greenhouse Gas Inventories. OECD/ OCDE, Paris. IPCC (Intergovernmental Panel on Climate Change) 2001. Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventories. Institute for Global Environmental Strategies, Kanagawa, Japan. ¨ Kaiser E.-A., Kohrs K., Kucke M., Schnug E., Heinemeyer O. and Munch J.C. 1998. Nitrous oxide release from arable soil: importance of N fertilization, crops and temporal variation. Soil Biol. Biochem. 30: 1553–1563. Keeney D.R., Fillery I.R. and Marx G.P. 1979. Effect of temperature on the gaseous nitrogen products of denitrification in a silt loam soil. Soil Sci. Soc. Am. J. 43: 1124–1128. Klute A. (ed.) 1986. Methods of Soil Analysis Part I, Physical and Mineralogical Methods. 2nd edn. American Society of Agronomy, Madison, WI. Kroeze C., Mosier A. and Bouwman L. 1999. Closing the global N 2 O budget: A retrospective analysis, 1500–1994. Global Biogeochem. Cycles 13: 1–8. Malhi S.S., McGill W.B. and Nyborg M. 1990. Nitrate losses in soils: effects of temperature, moisture and substrate concentration. Soil Biol. Biochem. 22: 733–737. McTaggart I.P., Clayton H., Parker J., Swan L. and Smith K.A. 1997. Nitrous oxide emissions from grassland and spring barley, following N fertiliser application with and without nitrification inhibitors. Biol. Fertil. Soils 25: 261–268. Minami K. 1994. Effect of nitrification inhibitors and slow-release fertilizers on emissions of nitrous oxide from fertilized soils. In: Minami K., Mosier A. and Sass R. (eds), CH 4 and N 2 O: Global Emissions and Controls from Rice Fields and Other Agricultural and Industrial Sources. NIAES, Yokendo, Tokyo, pp. 187–196. Mosier A.R., Kroeze C., Nevison C., Oenema O., Seitzinger S. and Van Cleemput O. 1998. Closing the global atmospheric N 2 O budget: nitrous oxide emissions through the agricultural nitrogen cycle. Nutr. Cycl. Agroecosyst. 52: 225–248. Ruser R., Flessa H., Schilling R., Beese F. and Munch J.C. 2001. Effect of crop-specific field management and N fertilization on N 2 O emissions from a fine-loamy soil. Nutr. Cycl. Agroecosyst. 59: 177–191. Ryden J.C. and Lund L.J. 1980. Nitrous oxide evolution from irrigated land. J. Environ. Qual. 9: 387–393. Scott A., Crichton I. and Ball B.C. 1999. Long-term monitoring of soil gas fluxes with closed chambers using automated and manual systems. J. Environ. Qual. 28: 1637–1643.
46 Smith K.A., McTaggart I.P. and Tsuruta H. 1997. Emissions of N 2 O and NO associated with nitrogen fertilization in intensive agriculture, and the potential for mitigation. Soil Use Manag. 13: 296–304. Smith K.A., McTaggart I.P., Dobbie K.E. and Conen F. 1998. Emissions of N 2 O from Scottish agricultural soils as a function of fertiliser N. Nutr. Cycl. Agroecosyst. 52: 123–130. Tsuruta H., Tanizaki T., Yagi K., Kanda K. and Minami K. 1992. N 2 O and NO emissions from fertilized soils. Proceedings of the 5th International Workshop on Nitrous Oxide Emissions, Tsukuba., pp. 281–285. Velthof G.L., Oenema O., Postma R. and Van Beutsichem M.L.
1997. Effects of type and amount of applied nitrogen fertiliser on nitrous oxide fluxes from intensively managed grassland. Nutr. Cycl. Agroecosyst. 46: 257–267. Watson C.J. and Adams S.N. 1986. Effect of simulated wet spring conditions on the relative efficiency of three forms of nitrogen fertiliser on grassland. J. Agri. Sci. Camb. 107: 219–222. Watson C.J., Stevens R.J., Garrett M.K. and McMurray C.H. 1990. Efficiency and future potential of urea for temperate grassland. Fert. Res. 26: 341–357. Zar J.H. 1999. Biostatistical Analysis. 4th edn. Prentice-Hall Inc., Upper Saddle River, NJ.
