Developing revised emission factors for nitrous oxide emissions from agricultural pasture treated with nitrification inhibitors

Literature review on effectiveness of nitrification inhibitors

Dicyandiamide (DCD)
New Zealand peer-reviewed literature
International peer-reviewed literature
Application rate for nitrification inhibitors
Formulation and application of nitrification inhibitors
Recording the area where nitrification inhibitors are applied
Recording variables to estimate nitrogen fertilizer application and excretion rates
Comparison of emissions factors from other countries
Revised emission factors and emissions inventory: Method 1 - ‘Annualised’ revisions of EF_3(PRP), EF₁ and Frac_LEACH using nitrification inhibitors
Revised emission factors and emissions inventory: Method 2 - Disaggregating nitrogen application onto soils for nitrification inhibitor responses
Revised emission factors and emissions inventory: Method 3 - Disaggregating the EF_3(PRP) and EF₁ data for revision by nitrification inhibitor response

Nitrogen (N) is applied to agricultural soils in two forms; namely, excreta deposited as urine and dung by sheep and cattle during year-round grazing outdoors and fertilizer. This review is related to N losses from soil as nitrous oxide (N₂O) emissions. These occur directly to the atmosphere and indirectly because of nitrate leaching. For computation of New Zealand’s N₂O emissions inventory, urine and dung excreta are aggregated but these components are available separately on monthly bases. For the 2003 inventory, fertiliser application to agricultural soils was 337 Gg N (14 % of the total N input to these soils). Corresponding values for urine and dung excreta were 796 (32 % of the total) and 408 Gg N (16 %) for sheep, 217 (9 %) and 111 Gg N (5 %) for beef cattle and 444 (18 %) and 159 Gg N (6 %) for dairy cattle. The percentage of N applied to soils which is directly emitted as N₂O is called the direct N₂O emissions factor. For urine and urea fertiliser, based on 17 field trials conducted in New Zealand, the direct N₂O emissions factors were indistinquishable and the overall average was 1.0 % (Kelliher et al. 2005a, Kelliher and de Klein 2006). For dairy cattle and sheep dung, the corresponding averages from six field trials were 0.2 % and zero (Kelliher et al. 2005a). To our knowledge, there have been no measurements of N leaching following dung excretion onto soils. Because dung N content is only 3 – 4 %, we’d expect no N leaching associated with its application to soils. Returning to the 2003 inventory and keeping in mind the N₂O emissions factors and N leaching, urine and fertiliser applications to agricultural soils were 1,457 and 337 Gg N, respectively. For this reason, we will focus our review on urine excreta.

A large number of chemical compounds are marketed as nitrification inhibitors and these are used in agricultural systems including dicyandiamide (DCD), 3, 4-dimethylpirazol phosphate (DMPP), neem oil, sodium thiosulfate, sulphur, acetylene, thiourea, 2-amino-4-chloro-6-methylpyrimidine (AM), 4-chloro-3-methylpyrazole (CIMP) and nitrapyrin amongst others. Nitrification inhibitors, through the inhibition of nitrification of ammonium, reduce leaching losses of nitrate. By slowing nitrification, N₂O production rate is reduced as a by-product of nitrification. Reduction of nitrate also reduces potential for the denitrification of nitrate and the production of N₂O, an intermediary in the denitrification pathway.

Most nitrification inhibitors have not been assessed for their effectiveness in reducing N₂O emissions from grazed pasture systems. This rendered much of the international literature irrelevant to the project. Further, it is essential that application of the nitrification inhibitor to New Zealand soils is sustainable including no deleterious environmental consequences. Dicyandiamide (DCD) (chemically written as C₂H₄N₄) has been studied for more than 80 years (for example, McGuinn 1924) and it has been subject to many tests with no reported environmental side effects (Suter et al. 2006). Suter et al. (2006) determined that DCD and DMPP were the available nitrification inhibitors most suited for use in pastoral systems. In New Zealand, Suter et al. reported DMPP is only available as a coated ammonium nitrate fertiliser. They concluded that this form of DMPP delivery would greatly limit its efficacy with respect to urine excreta patches in soils beneath grazed pasture and also make the inhibitor’s use cost prohibitive. Research trials using DCD have recently been conducted in New Zealand’s pastoral agricultural system and, in agreement with Suter et al. (2006), we focus our review on DCD.

The literature relating to DCD is extensive. An examination of the CAB International bibliographic database provided >500 abstracts for articles related to ‘dycandiamide and soil’ in agricultural systems over the period 1910 through January 2007. However < 20 of these are relevant because they deal with either urine and/or pasture systems.

We have excluded DCD literature that does not apply to grazed pastoral systems and in particular its use in conjunction with urine or dung. We do not include slurry (stored animal excreta) as the practice of storing excreta and applying it as slurry onto soils is not done in New Zealand, and slurry is chemically distinct from freshly deposited manure. We also excluded literature covering the use of DCD in rice paddies and for arable and horticultural crop production. Finally, comparative trials of DCD with other nitrification inhibitor chemicals/products and urease inhibitors are not considered here.

Sparseness of the literature led us to make some exceptions to our selection criteria for two key factors relevant to the effectiveness of DCD application to soils. This more catholic approach was necessary to obtain enough data to independently analyse the crucial and controversial relation between DCD decomposition and soil temperature. This was deemed essential for the review and synthesis of seasonal DCD efficacy field trials that have been conducted across New Zealand.

Dicyandiamide (DCD)

DCD is a white or colourless crystal that inhibits the first stage of nitrification (Amberger 1989). For application to soils at 10 kg/ha, DCD is considered to have a low water solubility. This can be illustrated by a simple calculation using the water solubility of DCD, 23 grams/litre at 13 °C according to Amberger (1989). At this temperature, DCD application in aqueous solution to one hectare of soil includes 435 litres of water! In contrast, the water solubility of sucrose at 13 °C is around 190 kg/litre. DCD contains 65% nitrogen and in soils, it is decomposed via guanyle urea and guanidine to urea, a conventional nitrogen fertilizer (Amberger, 1989). DCD is practically non-toxic according to Amberger (1989) (LD50/LC50: Oral, mouse: LD50 = >10 g/kg). In comparison, for salt (sodium chloride), the LD50: Oral, mouse = 4 g/kg. After DCD has decomposed to urea, it ultimately hydrolyses, catalysed by the enzyme urease, to form ammonium bicarbonate and hydroxide ions. DCD has a bacteriostatic mode of action, so it does not kill soil bacteria but rather inhibits or reduces their activity. In soils, microbes use the DCD molecule as a nitrogen source, and it is biotically mineralised or decomposed by specific enzyme activity (Schwarzer and Haselwandter, 1991).

New Zealand peer-reviewed literature

Effects on nitrate leaching (Frac_LEACH)

In a lysimeter study that used a free-draining Lismore stoney silt loam soil the use of DCD decreased nitrate leaching by 76% for urine applied in the autumn and by 42% for urine applied in the spring (Di and Cameron, 2002). In this trial, urea was also applied in eight split dressings to the lysimeters (200 kg N/ha/y). DCD was applied in solution form at the rate of 7.5 kg/ha after each split application of urea or at the higher rate of 15 kg/ha with each urine application (Di and Cameron, 2002).

Concentrations of inorganic-N were monitored in Wakanui silt loam soil over 100 days in urine-amended plots with and without DCD at either 10 or 25 kg/ha (Cookson and Cornforth, 2002). Commencing in March (day 60 of the year), at a depth of 0.075 m, daily average soil temperature decreased from 18 ^oC to 6 ^oC. In the urine only plots soil ammonium decreased to the background level found in control plots within 28 days. When urine was applied with DCD it took 60 days for soil ammonium concentrations to reach the levels seen in the control plots. DCD application reduced nitrate leaching (below a depth of 0.2 m) by 46 % (10 kg DCD/ha) and 74 % (25 kg DCD/ha).

Lysimeters, containing Templeton fine sandy loam soil, were treated with cow urine (1000 kg N/ha) and urea N (200 kg N/ha/y split into 8 applications) and then treated with DCD (15 kg/ha) in autumn or treated with DCD in autumn and spring, in solution form (Di and Cameron, 2004b). These treatments were designed to determine the effectiveness of DCD in reducing nitrate leaching as affected by one or two applications. There was no difference in the quantity of nitrate leached for the two DCD treatments with a 76% reduction in nitrate leaching.

Lysimeters, containing Templeton fine sandy loam soil, were treated with urea N (200 kg N/ha/y split into 8 applications) and urine (1000 kg N/ha) applied in a single application in the autumn. DCD treatments applied included ‘no DCD’, and DCD applied as a fine particle suspension at either 5 or 10 kg/ha of DCD (Di and Cameron, 2005). Nitrate-N leaching losses were not significantly reduced when DCD was applied at 5 kg/ha but they were significantly reduced at 10 kg/ha with nitrate-N leaching losses reduced by 68%.

Smith et al. (2005) found that DCD (15 or 30 kg N/ha) applied with synthetic urine (580 kg N/ha) was effective in limiting nitrification in a poorly-drained Pukemutu soil for more than 100 days. This indicated the potential for reduced nitrate leaching when treatments were applied in late spring. At a depth of 0.1 m, daily average soil temperature increased from 10 ^oC to 18 ^oC by day 80 of the year (21 March).

A summary of experimental protocols for relevant published field studies performed in New Zealand, and their treatments, is presented in Table 1. (Note the field study of Smith et al. (2005) is presented and discussed separately below).

Table 1: Experimental overviews and treatments of lysimeter and field studies using DCD. Fertiliser N (urea) was applied to all treatments, at a rate of 200 kg N ha^-1 y^-1 split into 8 dressings, the exception to this was in (Di and Cameron, 2002), where treatment 1 received nil fertiliser, and treatments 7 and 8, where the urea was split evenly into 4 dressings. Urine-N was applied in all treatments unless stated, in the season shown, at rate of 1000 kg N ha^-1. DCD was applied as a solution in studies except in Di and Cameron (2005; 2006) and Di et al. (2006) where a fine particle suspension (FPS) was used.

Reference^a	Soil surface texture	Soil pH	Pasture age^b (y)	Irrigation (mm)	Treatments	Urine application season	Total N (kg)	DCD^c (kg ha^-1)	Total DCD kg)
(Di and Cameron, 2002)	silt-loam	5.9	4	100 mm flood every 3 weeks 1364 total 564 rain 800 irrig’n	1	nil	0	0	0
					2	nil	200	0	0
					3	Aut.	1200	0	0
					4	Aut.	1200	7.5^d/15.0^e	75
					5	Aut.	1200	0	0
					6	Aut.	1200	7.5^d/15.0^e	75
					7	Spr.	1200	0	0
					8	Spr.	1200	7.5^d/15.0^e	45

(Di and Cameron, 2003)	silt loam	5.9	5	50 mm spray every 2 weeks 850 total 490 rain 360 irrig’n	1	Aut.	1200	0	0
					2	Aut.	1200	15^f	15
					3	Aut.	1200	15^g	30
					4	Aut.	1200	15^h	15
					5	Spr.	1200	0	0
					6	Spr.	1200	15^f	15
					7	Spr.	1200	15ⁱ	75
					8	Spr.	1200	7.5^j	75
(Di and Cameron, 2004b)	silt loam	5.8	>10	50 mm spray every 2 weeks 1600 total	1	Aut.	1200	0	0
					2	Aut.	1200	15^f	15
					3	Aut.	1200	15^k	30
(Di and Cameron, 2005)	Sandy loam	5.8	>10	50 mm spray every 2 weeks 1200 total 500 rain 700 irrig’n	1	Aut.	1200	0	0
					2	Aut.	1200	5^k	15
					3	Aut.	1200	10^k	30

(Di and Cameron, 2006)	silt loam sandy loam	5.9	5	30 mm spray every week 1200 total 500 rain 700 irrig’n	1	Aut.	1200	0	0
					2	Aut.	1200	7.5^k	15
					3	Aut.	1200	10^k	20
					4	Aut.	1200	15^k	30
					5	Win.	1200	0^k	0
					6	Win.	1200	10^k	20
					7	Aut.	1200	0^k	0
					8	Aut.	1200	10^k	20
					9	Aut.	1200	10^l	20
(Di et al., 2006)	fine-sandy-loam	5.8	not stated	Supplementary rainfall	1	nil	0	0	0
					2	nil	0	10	10
					3	Win.	1000	0	0
					4	Win.	1000	10	10

	stony-silt-loam	5.9	not stated	Supplementary rainfall	1 2	Aut. Aut.	1000 1000	0 10	0 20^a

	Silt Loam	5.5	Not Stated	nil	1	nil	0	0	0
					2	nil	0	10	20^a
					3	Aut.	1000	0	0
					4	Aut.	1000	10	20^a

	pumice Sand	5.6	Not Stated	nil	1	nil	0	0	0
					2	nil	0	10	10
					3	Win.	700	0	0
					4	Win.	700	10	10

^aDCD applied in autumn and in winter
^bPastures were all perennial ryegrass (Lolium perenne) - white clover (Trifolium repens).
^cDCD = dicyandiamide,
^dkg per urea application,
^ekg per urine application,
^fafter urine application,
^g15 kg after urine application and 15 kg in late winter,
^hmixed with urine,
ⁱ15 kg mixed with urine application then 15 kg quarterly,
^j7.5 kg after urine application and 7.5 kg after each urea application.

Table 2 summarises nitrate leaching from studies where DCD was applied with urine.

The reduction in nitrate leaching ranged from 13 to 77%. If the lowest rate of DCD is excluded (i.e. the 5kg/ha rate used in treatment 2 by Di and Cameron (2005)), the range becomes 42 to 77%. All the leaching measurements were performed after autumn/winter applications of urine, the sole exception being Di and Cameron (2002), treatment 8 where the urine was applied in spring. A spring application severely reduces the likelihood of nitrate leaching and it could be considered that there is also greater opportunity for plant uptake of urine-N over the growing season. DCD should be most effective in reducing nitrate leaching when application coincides with the lowest soil temperatures (see below) and highest drainage rates, i.e. during from autumn through winter.

We thus eliminate the spring data (42% from Di and Cameron (2002), treatment 8), so the range becomes 68 to 77% with a mean of 74% reduction in nitrate leaching (Frac_LEACH) when DCD was applied (n=5, Std Dev. 4 %). The average result coincides with 75% reduction in nitrate leaching reported by Cookson and Cornforth (2002). The DCD application rates were 10 and 15 kg/ha with the lower rate achieving the minimum leaching reduction (68%). The application dose effect will be explored later.

The experiments reviewed here did not include cattle grazing following DCD application. The reduction in nitrate leaching thus did not account for recycling of ‘conserved’ nitrogen via animal excreta. It follows that plant ingestion by grazing cattle leads to N excretion onto soils. The soil thus receives further N application(s) but no accompanying DCD application(s). Hence, for grazing after DCD application, nitrate leaching may be greater than expected from the reviewed experiements. This is predicated on the ‘conserved’ nitrogen being taken up by pasture plants. We note that on high fertility soils, ryegrass begins growth at 5 °C according to (Kerr, 2000). Pasture growth can thus be limited by low temperature during May - September. Finally, we note the experiments reviewed included multiple urea fertiliser dressing following DCD application, although these dressing involved only 25 kg N/ha, while urine application was 700 – 1200 kg N/ha..

The experiments reviewed here are intended to represent the results of DCD application onto agricultural pasture. For dairy farms, in the South Island and possibly the Waikato and Taranaki regions in future, grazing may take place during May – September on ‘support’ land. This land may not be located within a farm’s boundary. There may also be a ‘cut and carry’ feeding regime during ‘May – September’ including feed produced outside the farm’s boundary. This may involve ‘stand off’ and feed pads. We do not know the importation rate of palm kernel (by-product of the palm oil industry in South East Asia, mainly in Malaysia) that may be used as cattle feed, but this may be another consideration.

Finally, we analyse the tabulated data to determine FracLEACH. As stated earlier, the New Zealand-specific value for FracLEACH is 0.07. Following urine application in autumn 2001 to the free-draining Lismore stoney silt loam, where no DCD was applied, 43% of the applied N leached from the soil as nitrate according to Di and Cameron (2002)(Table 2). Consequently, FracLEACH was 0.43. Following urine application in autumn 2002 and 2003 to the free-draining Templeton fine sandy loam, where no DCD was applied, 7 and 11 % of the applied N leached from the soil as nitrate according to Di and Cameron (2004b) and Di and Cameron (2005), respectively. Hence, FracLEACH averaged 0.09 in these studies.

Table 2: Summary of the mass of NO₃-N leached, the percentage reduction in NO₃-N leaching when DCD was used, the fraction of

N applied leached; N₂O-N emissions and their respective reductions in emissions when DCD was used in the various research trials following

cow urine application to lysimeters and field plots^.

Reference^b	Soil	Treatment	Total N Applied (kg)	DCD^c (kg ha^-1)	NO₃-N leached (kg)	Leaching reduction (%)	Fraction of N applied leached	N₂O-N (kg ha^-1)	Reduction in N₂O Loss (%)
(Di and Cameron, 2002)	Lismore	1	0	0	4.8	-	-	-	-
	Lismore	2	200	0	7.9	-	0.040	-	-
	Lismore	3	1200	0	516	-	0.430	-	-
	Lismore	4	1200	7.5^d/15^e	128	75	0.107	-	-
	Lismore	5	1200	0	488	-	0.407	-	-
	Lismore	6	1200	7.5^d/15^e	112	77	0.093	-	-
	Lismore	7	1200	0	397	-	0.331	46.0	-
	Lismore	8	1200	7.5^d/15^e	230	42	0.192	8.5	82

(Di and Cameron, 2003)	Lismore	1	1200	0	-	-	-	26.7	-
	Lismore	2	1200	15^f	-	-	-	7.0	74
	Lismore	3	1200	15^g	-	-	-	7.6	72
	Lismore	4	1200	15^h	-	-	-	4.5	83
	Lismore	5	1200	0	-	-	-	18.0	-
	Lismore	6	1200	15^f	-	-	-	4.5	75
	Lismore	7	1200	15ⁱ	-	-	-	4.8	73
	Lismore	8	1200	7.5^j	-	-	-	2.5	86
(Di and Cameron, 2004b)	Templeton	1	1200	0	85	-	0.071	-	-
	Templeton	2	1200	15^f	22	74	0.018	-	-
	Templeton	3	1200	15^k	20	76	0.017	-	-

(Di and Cameron, 2005)	Templeton	1	1200	0	134	-	0.112	-	-
	Templeton	2	1200	5^k	116	13	0.097	-	-
	Templeton	3	1200	10^k	43	68	0.036	-	-

(Di and Cameron, 2006)	Lismore	1	1200	0	-	-	-	23.1	-
	Lismore	2	1200	7.5^k	-	-	-	8.2	65
	Lismore	3	1200	10^k	-	-	-	6.9	70
	Lismore	4	1200	15^k	-	-	-	6.2	73
	Lismore	5	1200	0^k	-	-	-	31	-
	Lismore	6	1200	10^k	-	-	-	8.4	73
	Templeton	7	1200	0^k	-	-	-	37.4	-
	Templeton	8	1200	10^k	-	-	-	14.6	61
	Templeton	9	1200	10^l	-	-	-	16.3	56
(Di et al., 2006)	Templeton	1	0	0	-	-	-	1.0
	Templeton	2	0	10	-	-	-	0.8	20
	Templeton	3	1000	0	-	-	-	20.9
	Templeton	4	1000	10	-	-	-	5.7	73

	Lismore	1	1000	0	-	-	-	8.7
	Lismore	2	1000	10	-	-	-	2.9	67

	Horotiu	1	0	0	-	-	-	0.27
	Horotiu	2	0	10	-	-	-	0.12	56
	Horotiu	3	1000	0	-	-	-	6.2
	Horotiu	4	1000	10	-	-	-	2.4	61

	Taupo	1	0	0	-	-	-	0.18
	Taupo	2	0	10	-	-	-	0.15	17
	Taupo	3	700	0	-	-	-	1.01
	Taupo	4	700	10	-	-	-	0.31	69

^bPastures were all perennial ryegrass (Lolium perenne) - white clover (Trifolium repens).
^cDCD = dicyandiamide,
^dkg per urea application,
^ekg per urine application,
^f15 kg after urine application,
^g15 kg after urine application and 15 kg in late winter,
^h15 kg mixed with urine application,
ⁱ15 kg mixed with urine application then 15 kg quarterly,
^j7.5 kg after urine application and 7.5 kg after each urea application,
^kafter urine application and in mid-winter,
^lten days after urine application and mid-winter.

Effects on direct N₂O emissions from urine applied to soils (EF_3(PRP))

Table 3 summarises direct N₂O emission factors obtained from field studies of four soils based on various treatments with DCD.

In the lysimeter study that used a free-draining Lismore stony silt loam soil, DCD application decreased direct N₂O emissions (hereafter, N₂O emissions) by 82% from urine patches over the course of two spring urine deposition events (Di and Cameron, 2002). Urea was also applied in eight split dressings to the lysimeters (200 kg N/ha/y). DCD was applied in solution form at the rate of 7.5 kg/ha after each split application of urea or at the higher rate of 15 kg/ha with each urine application.

A lysimeter study, that also used a free-draining Lismore stony silt loam, showed DCD reduced N₂O emissions by 76% following urine application in autumn and by 78% following a spring urine application (Di and Cameron, 2003). Repeated applications of DCD after urine application or mixing DCD with urine had effects similar to a single application immediately after urine deposition. DCD was applied in solution form at the rate of 15 kg/ha after each urine application.

Templeton and Lismore soils were again used in lysimeter studies where urea (eight split dressings; 200 kg N/ha/y) and cow urine (1000 kg N/ha) were applied (Di and Cameron, 2006). Urine treatments were applied in late autumn or winter. DCD was applied twice, initially in late autumn and again in winter to all treatments. Where urine was applied in late autumn the first DCD application was applied immediately after the urine except for one treatment where the DCD was applied 10 days after the urine. Where urine was applied in winter the first DCD application occurred in late autumn followed by the second application in winter. DCD was applied at three rates to the Lismore soil (7.5, 10 and 15 kg/ha) when urine was applied in late autumn and at 10 kg/ha to late autumn urine applications on the Templeton soils. The DCD was applied as a fine particle suspension and sprayed evenly over the soil surface prior to irrigation (10 mm). On the Lismore soils all three rates of DCD were effective in reducing N₂O emissions from the urine applied in late autumn by 65-73% while the 10kg/ha rate of DCD also reduced N₂O emissions from the Lismore soil in winter by 73%. In the Templeton soil DCD (10 kg/ha) reduced late autumn applied urine emissions by 61%. DCD applied 10 days after urine deposition was equally effective with a 56% reduction in N₂O emissions.

Di et al. (2006) examined the effectiveness of DCD in reducing N₂O emissions from urine patches on four different soil types located in the Waikato (Horotiu silt loam), Canterbury (Templeton fine sandy loam and a Lismore stony loam) and Taupo (Taupo pumice sand) regions. The authors noted that soil temperatures differed between soils over the experimental periods. Analysis of these data indicated that, on average, the Horotiu soil was 2 – 3 ^oC warmer but no differences were statistically significant (see further discussion of these data in a subsequent section). At all sites urine was applied (1000 kg N/ha) either in autumn or winter. DCD was applied as a fine particle suspension on the same day as the urine treatments with the exception of the Templeton soil where an 18 day delay was instigated to simulate the possible delay between the end of a grazing event and a DCD application. Reductions in the N₂O emissions for the Templeton, Lismore, Horotiu and Taupo soils were 73, 67, 61 and 69% respectively.

When DCD was applied, the reduction in N₂O emissions ranged from 56 to 86% (n = 17, mean 71 %, Std. Dev. 8 %). This data analysis includes the four soils called Lismore (n = 12), Templeton (n = 3), Horotiu (n = 1) and Taupo (n= 1). Excluding the value of 56 % obtained for the Templeton soil when DCD was applied 10 or 18 days after the urine, the range becomes 61 to 86% (n = 16, mean 72 %, Std. Dev. 7 %).

We now refine the data analysis by considering the soils separately. For the Templeton soil, N₂O reductions ranged from 56 to 73 % (n = 3, mean 63 %, Std. Dev. 9 %). The largest data set is available for the Lismore soil where N₂O reduction ranged from 65 to 83 % (n = 12, mean 74 %, Std. Dev. 6 %). For the Horotiu and Taupo soils, single trials for each soil yielded N₂O emission reductions of 61 and 69% respectively. Combining the mean and single values of the direct N₂O emissions reduction (EF_3(PRP)) when DCD was applied to the four soils, the range was 61 to 74 % (n = 4, mean 67 %, Std. Dev. 6 %).

As before, the experiments reviewed here did not include cattle grazing following DCD application. Smith et al. (2007) provided a draft manuscript reporting the effectiveness of DCn^TM to reduce soil nitrate accumulation and N₂O emissions within a grazed pasture system. In the first year, during late April 2004, urea or urea+DCD treatments were applied after grazing. Both treatments had a urea application rate of 50 kg/ha and the DCD application rate was 10 kg/ha (42 kg DCn^TM/ha). Plots were grazed (for 3 hours by 2.7 dairy cattle/ha as described by de Klein et al., 2006; confirmed by personal communication on 18 April 2007) prior to and one month after treatment applications. In year two the same treatments were applied three times on 4^th March, 18^th April, and 25^th August. The plots were grazed just before each of the three treatment application and one month after the third treatment application (for one day, less time spent being milked twice, by 2.7 dairy cattle/ha). The experimental design of Smith et al. (2007) is thus unique because grazing took place after DCD application and direct N₂O emissions were measured.

In year 1 N₂O emissions were generally insignificant and spatially variable. This may have reflected the water content exceeding field capacity in the poorly drained soil throughout 2004. In year 2 with an elevated number of replicates, differences in N₂O emissions were detected for 70% of the measurement sets and DCD application corresponded with significantly lower emissions. The soil remained very wet but drier than year 1, especially during the spring of year 2. In year one few differences in soil ammonium N were detected, with a brief period where soil nitrate-N was elevated in the urea only treatment. In year 2 soil ammonium-N varied little in March and April but was higher in the DCD treated plots in spring. Soil nitrate-N was reduced significantly with DCD use, but not after a second grazing following the August application in spring (29^th September). Reductions in N₂O emissions ranged between 75 and 91% over a 2 to 3 month period following application (n = 3, mean 82%, Std. Dev. 8). These results should also be considered for judgement about direct N₂O emissions reduction (EF_3(PRP)) following DCD application.

Table 3: Emission factors for N₂O-N calculated as the mass of N₂O-N divided by the mass of N applied, either gross-N (fertiliser + urine) or urine-N only

Reference	Treatment	DCD^c (kg ha^-1)	EF (gross N)	EF (urine-N)	Reduction in EF (%)
(Di and Cameron, 2002)	7	7.5^d/15^e	0.038	0.02	82
	8	0	0.007

(Di and Cameron, 2003)	1	0	0.022	-	-
	2	15^f	0.006	-	74
	3	15^g	0.006	-	72
	4	15^h	0.004	-	83
	5	0	0.015	-	-
	6	15^f	0.004	-	75
	7	15ⁱ	0.004	-	73
	8	7.5^j	0.002	-	86

(Di and Cameron, 2006)	1	0	0.019	0.023	-
	2	7.5^k	0.007	0.008	65
	3	10^k	0.006	0.007	70
	4	15^k	0.005	0.006	73
	5	0^k	0.026	0.027	-
	6	10^k	0.007	0.007	73
	7	0^k	0.031	0.036	-
	8	10^k	0.012	0.014	61
	9	10^l	0.014	0.016	56

	1	0	-	-	-
	2	10	-	-	-
	3	0	-	0.020	-
	4	10	-	0.005	75

	1	0	-	0.008	-
	2	10	-	0.003	63

	1	0	-	-	-
	2	10	-	-	-
	3	0	-	0.006	-
	4	10	-	0.002	67

	1	0	-	-	-
	2	10	-	-	-
	3	0	-	0.009	-
	4	10	-	0.003	67

^ekg per urine application,
^f15 kg after urine application,
^g15 kg after urine application and 15 kg in late winter,
^h15 kg mixed with urine application,
ⁱ15 kg mixed with urine application then 15 kg quarterly,
^j7.5 kg after urine application and 7.5 kg after each urea application,
^kafter urine application and in mid-winter,
^lten days after urine application and mid-winter.

Effects on herbage yields

In a lysimeter study that used a free-draining Lismore stoney silt loam soil the use of DCD increased herbage yields in all treatments relative to non-DCD treatments (Di and Cameron, 2002). Urea was also applied in eight split dressings to the lysimeters (200 kg N/ha/y). DCD was applied in solution form at the rate of 7.5 kg/ha after each split application of urea or at the higher rate of 15 kg/ha with each urine application (Di and Cameron, 2002). The application of DCD increased annual herbage yields in the autumn treatments by an average 49% and by an average 18% in the spring treatments.

For a poorly-drained Pukemutu soil, Smith et al. (2005) found no significant differences in pasture production for treatments that included artificial urine (580 kg N/ha) and DCD at 15 or 30 kg N/ha in a solution form. A urine-only treatment was not included, so it is not possible to compare the effect of urine +DCD against urine only in terms of DM yield. In addition this study was performed in late spring when pasture production rate should be at a maximum (the 156-day-long trial began 30 October 2003). Rainfall maintained soil water content near the field capacity for the trial’s first 30 days but thereafter, soil water content was only about one-quarter of the field capacity. At a depth of 0.1 m, minimum, maximum and mean soil temperatures were 7, 18 and 15 ^oC, respectively. The relatively dry and warm soil may have affected pasture herbage response to the additional N by DCD treatment. Also with respect to the earlier studies, these 156-day-long results may not be directly comparable to those obtained over an entire year.

There has been one New Zealand field trial quantifying the effects of repeated use of DCD on pasture production and quality (Moir et al. 2007). For four years (2002 – 2006 at the Lincoln University Dairy Farm), DCD was applied to a Wakanui silt loam soil beneath pasture at 10 kg ha^-1 in early May in addition to dairy cattle urine (1000 kg N ha^-1) with DCD applied again in early August. A herd of dairy cattle (3.3 cows/ha in year one increasing annually up to 4.3 cows/ha in year four) grazed the plots at approximately 21 day intervals during September - May (9 months) each year with no grazing during June – August. Comparisons were made with control plots that received no DCD. All plots had an area of 100 m² and a (‘nova flow’) drainage pipe installed at a depth of 0.6 m. Each year, DCD applications consistently corresponded with increased pasture herbage dry matter yield that ranged from 19 to 36 % (n = 8, mean 24 %, Std. Dev. 6 %) on an annual basis including urine patches and inter-urine areas. On average, calculated on whole paddock and annual bases, application of DCD corresponded with a 21 % increase of dry matter production. Pasture nitrogen, metabolisable energy and fibre contents were not affected by the DCD applications.

Table 4 summarises the herbage yield data with respect to DCD treatment.

Table 4: Increases in DM yields under DCD applications and the average total annual yields.

Reference	Treatment	DCD^c (kg ha^-1)	Increase in DM (%)	%N	Average DM yield (tonne ha^-1 y^-1)
(Di and Cameron, 2002)	1	0	-	3.5^c	11^c
	2	0	-	-	-
	3	0	-	-	-
	4	7.5^d/15^e	49^a	4.1^d	15^d
	5	0	-	-	-
	6	7.5^d/15^e	-	^-	-
	7	0	-	-	-
	8	7.5^d/15^e	18^b	^-	-

(Di and Cameron, 2004d)	1	0	-	3.3	15.9
	2	15^f	14	3.5	18.2
	3	15^k	33	3.1	21.1

(Smith et al., 2005)	1	0^l	-	1.9	3.0^m
	8	10^l	-	3.6	6.7^m
	9	20^l	-	3.6	7.2^m

(Di and Cameron, 2006)	1	0	-	2.9	15.3
	2	5^k	0	2.9	15.3
	3	10^k	33	3.1	20.3
(Moir et al., 2007) Year 1	1^a	0		3.5	10.2
	2^b	10	21	3.5	12.3
	3^c	0		3.6	11.6
	4^d	10	30	3.4	15.1
(Moir et al., 2007) Year 2	1	0		3.6	9.0
	2	10	17	3.6	10.5
	3	0		3.8	10.0
	4	10	23	3.8	12.3
(Moir et al., 2007) Year 3	1	0		3.5	10.3
	2	10	25	3.7	12.9
	3	0		3.9	13.3
	4	10	36	4.0	18.1
(Moir et al., 2007) Year 4	1	0		3.6	11.8
	2	10	19	3.8	14.1
	3	0		4.1	14.9
	4	10	23	4.3	18.4

^aaverage of autumn urine treatments + DCD
^baverage of spring urine treatments + DCD
^cwithout DCD
^dwith DCD
^ekg per urine application,
^f15 kg after urine application,
^kafter urine application and in mid-winter,
^l artificial urine applied at nil (treatment 1) or 580 kg N/ha with DCD in solution form),
^mDM yield is for tonnes/ha over 156 days.

Effects on ammonia volatilization

Using a free-draining Templeton fine sandy loam soil, lysimeters were treated with cow urine (1000 kg N/ha) and urea N (200 kg N/ha/y split into 8 applications) and then treated with DCD (15 kg/ha) in autumn or treated with DCD in autumn and spring, in solution form (Di and Cameron, 2004d). Ammonia volatilization losses were not increased by the DCD treatment with ammonia-N volatilization losses equal to 1.7 - 3.5% of the urine-N applied. The authors suggested that since urine is a liquid it would rapidly permeate into the soil after deposition.

It is understood that ammonia volatilization is affected by temperature, wind speed and soil pH, buffer capacity and the concentrations of ammonia or ammonium in soil solution. The factors affecting ammonia volatilization have been reviewed elsewhere (Haynes and Sherlock, 1986; Jarvis and Pain, 1990). For urine applied at 500 kg N/ha during summer, autumn and winter trials at Lincoln with Templeton silt loam soil, and including no DCD applications, ammonia volatilisation ranged from 10 to 37 % of the applied N (n = 9, Ave. 20 %, Std. Dev 8 %)(Sherlock, 1984).

Given that the use of DCD slows down the rate of nitrification in the urine patch and thus prolongs the presence of ammonium and the period of elevated soil pH, it might be expected that such conditions would favour volatilization. However, the extra ammonium available could also undergo other transformations and be fixed, immobilized or be taken up by plants, rather than be volatilized. In fact a prolonged elevation of soil pH was observed when urine was treated with DCD but this did not translate into higher ammonia losses (Di and Cameron, 2004d). The field study of Wakanui silt loam by Cookson and Cornforth (2002) included measurement of soil pH following the application of DCD at 10 or 25 kg/ha with urine (450 kg/ha). The controls had a pH of 5.3. After DCD application, the soil’s pH increased by less than 1 pH unit. These measurements were made on sieved samples, taken from the surface to a depth of 0.1 m, in the laboratory.

Effects on other soil cations

Using a free-draining Templeton fine sandy loam soil, lysimeters were treated with cow urine (1000 kg N/ha) and urea N (200 kg N/ha/y split into 8 applications) and then treated with DCD (15 kg/ha) in autumn or treated with DCD in autumn and spring, in solution form (Di and Cameron, 2004d). In the DCD treated urine patches, there were 38 – 56 % and 21 – 42 % reduction in the leaching of calcium and magnesium, respectively, but no change in potassium leaching. Decreased cation leaching was postulated to be attributable to decreased nitrate leaching and a reduced leaching requirement for counter ions.

Lismore stoney silt loam lysimeters were treated with urea (200 kg N/ha/y) split into 8 applications throughout the year and urine (1000 kg N/ha) applied in the autumn. DCD was applied in solution form 7.5 kg/ha after each urea application and immediately after urine application (15 kg N/ha). Leaching losses of calcium, potassium and magnesium were reduced by 50, 65 an 53% respectively when DCD was applied to urine treated lysimeters DCD (Di and Cameron, 2004c).

Templeton fine sandy loam lysimeters were treated with urea N (200 kg N/ha/y split into 8 applications) and urine (1000 kg N/ha) applied in a single application in the autumn. DCD treatments applied included ‘no DCD’, and DCD applied as a fine particle suspension at either 5 or 10 kg/ha of DCD (Di and Cameron, 2005). Calcium and magnesium leaching was reduced by 51 and 31 %, respectively, when DCD was applied at 10 kg/ha (Di and Cameron, 2005).

Effects on the soil microbial community

In an incubation study DCD (7.5 or 15 kg/ha) was applied in conjunction with urea (25 kg N/ha) and urine (1000 kg N/ha) to a Lismore silt loam at field capacity field, at either 8 or 20 ^oC. Soil microbial biomass carbon and nitrogen contents were not affected by DCD treatment (Di and Cameron, 2004e).

Effects on nitrate accumulation in pastoral soils and herbage

Smith et al.(2005) found no difference between the solid and liquid forms of DCD applied with urine onto pasture in terms of slowing nitrate formation in 0.1 m surface layer of soil (Table 5). A ‘urine- nil DCD’ treatment would have been beneficial and enabled the effect of the DCD treatment relative to urine only to be determined.

Smith et al. (2005) also found that DCD application reduced nitrate accumulation in herbage to safe levels for ingestion by grazing stock, as well as a trend to increasing herbage magnesium and calcium concentrations.

Table 5 from Smith et al. (2005) showing the treatments used in the spring formulation study

Treatment	DCD rate (kg/ha)	N application rate (kg/ha)
Treatment	DCD rate (kg/ha)	DCD	Urea	Urine
Control	Nil		nil	nil
Urea	Nil		50	nil
Urea+urine	Nil		50	580
Super U®^a	1	0.7	50	nil
Super U® + urine	1	0.7	50	580
Coated N (25%)^b	30	20	42	nil
Coated N (25%) + urine	30	20	42	580
Liquid DCD + urine^c	15	10	50	580
Liquid DCD + urine	30	20	50	580
Zeolite (25% DCD w/w) + urine^d	15	10	50	580
Zeolite (25% DCD w/w) + urine	30	20	50	580

International peer-reviewed literature

Effects of DCD on direct N₂O emissions from urine applied to soils (EF_3(PRP))

Williamson and Jarvis (1997) applied cow urine to grassland, in autumn, to a silty clay loam at the extremely low rate of 60 kg N/ha along with 7.9 kg DCD/ha and N₂O emissions were reduced by 74%. Mean soil temperature was 10 ^oC.

De Klein and Van Logtestijn (1994) measured N₂O emission rates following the application of artificial urine (400 kg urine-N/ha) to a perennial rye-grass sward on sandy soil in the Netherlands. Urine was also applied with DCD to distinguish between N₂O emission from denitrification and nitrification. When DCD was added to the urine, N₂O emissions over the following 14 days were reduced by 50 to 89 % compared to the urine only treatment.

Akai et al. (2001) conducted an experiment on a pasture sown with Italian ryegrass (Lolium multiflorum), tall fescue (Festuca arundinacea) and red clover (Trifolium pratense). The treatments included cow urine and cow urine + DCD. The nitrogen application rate was 50 kg/ha except for the urine + DCD, which had a nitrogen application rate of 55 kg/ha. Nitrous oxide generation was reduced by 66 and 40% in 1998 and 1999, respectively, in the cow urine + DCD treatment, compared to that in the cow urine treatment. Soil temperature was not reported in the abstract.

There is a dearth of international literature on the interaction between DCD, bovine urine, nitrate leaching (Frac_LEACH) and N₂O emissions (EF_3(PRP)). The three studies above included very low N application rates, compared to that associated with cattle urine excretion during grazing in New Zealand. Nevertheless, as under New Zealand conditions, DCD was effective in reducing N₂O emissions in these international pastoral soils trials.

DCD degradation and microbiology

It has been reported that DCD is degraded in soil via guanylurea, guanidine and urea to yield carbon dioxide and ammonium (Rathsack, 1955; Rodgers et al., 1985; Vilsmeier, 1980). The first step in the degradation cascade was assumed to be catalyzed by the interaction of metal oxides, e.g. Fe(OH)₃, rather than being due to microbial and hence enzymatic mineralization (Amberger and Vilsmeier, 1979; Hauser and Haselwandter, 1990). However, some soil bacteria species can utilise and thus degrade DCD (Schwarzer and Haselwandter, 1991). One of the most efficient isolates is strain EK1 of Mycobacterium sp. (Teaumroong et al., 1997). DNA probes have been developed to detect EK1 and related or identical species of soil bacteria (Teaumroong et al., 1997). Bacterial cultures have also been isolated from composts (Mycobacterium sp., Pseudomonas sp.) and shown to degrade DCD (200mg N/ml) within 3 days, with two metabolic pathways observed, one being consistent with guanyl urea metabolism (Hallinger et al., 1990). DCD mineralization in culture (strain EK1) was also shown to be enhanced under anaerobic conditions with the rate of mineralization decreasing with increasing concentration of DCD, following normal degradation kinetics in batch cultures as will be discussed further below (Hauser and Haselwandter, 1990).

DCD-N is only slowly mineralized in acid soils ca. pH 4.0 (Rodgers et al., 1985).

The interim products of DCD decomposition (guanylurea and guanidine) have been reported to have little if any effect on nitrification compared with DCD (McCarty and Bremner, 1989).

In laboratory experiments with nitrifying cultures of Nitrosomonas sp. nitrification was completely inhibited , but numbers of ammonium oxidizing bacteria were not significantly affected by a 48 h treatment with DCD (Rodgers and Ashworth, 1982).

DCD has been shown to not inhibit growth and respiration of N-fixing bacteria (Rhizobium leguminosarum and Azotobacter chroococcum) in cell suspensions with 400 μg/ml of DCD (Zacherl and Amberger, 1990). While DCD applied at high rates had the potential to affect the N₂ fixation process in nodules of alfalfa it was not likely to be of practical significance if DCD was used at rates normally required to inhibit nitrification (Rice and Olsen, 1988).

Laboratory experiments have demonstrated that repeat applications of DCD to soil have little or no effect on the rate of DCD decomposition or the ability of DCD to inhibit nitrification (Rodgers, 1986). Repeated field application of DCD resulted in no differences in the sensitivity of ammonium oxidizing bacteria to DCD (Rodgers, 1986).

In sterile soil at 30^oC the applied DCD concentration remained constant after 36 days (Rajbanshi et al., 1992) when soil was reinoculated it disappeared within 7 days. Addition of Fe₂O₃ powder to the sterilized soil had no effect on DCD degradation. Suggestion of an inducible metabolic degradation occurred since pretreated soils degraded DCD faster.

A relationship was established between the nitrification capability of microscopic soil fungi and the appearance of phytotoxic properties of micromycetes. The phytotoxic activity of pure cultures of micromycetes was found to change under the influence of nitrification inhibitors. The influence of nitrification inhibitors on growth, accumulation of biomass, and formation of nitrates and nitrites by microscopic fungi appeared at concentrations 1-4 orders of magnitude higher than with regard to autotrophic nitrifying bacteria. The tested preparations were arranged in the following order in terms of effectiveness of action of fungi: nitrapyrin, carbamoylmethylpyrazole, dicyandiamide, and 4-amino-1,2,4 triazole. (Kurakov and Popov, 1996).

DCD degradation and temperature

The fate of DCD applied to soils depends strongly on temperature. This dependence is complex, in turn varying with temperature. A progressive series of analyses will be done to synthesise the available data and further develop DCD application criteria for New Zealand conditions.

Hauser and Haselwandter (1990) measured the concentration of DCD, applied as five doses, over time since application to a nutrient solution maintained at 25 °C and containing the soil bacteria strain EK1. To quantify the reported relationship, we fitted their tabulated data to a first-order exponential decomposition model. Term ‘c’ is the DCD concentration and ‘c₀’ is the initial concentration of DCD at time zero. For the representative data portrayed in Figure 1, the model predicted that DCD concentration had declined to half of its initial value (called the half life, t_½, t_½ = Ln(2)/k where Ln denotes natural logarithm (Ln(2) = 0.69) and k is a decomposition constant determined by regression analysis) in 10 days. For the five doses, additional regression analysis showed that as c₀ increased from 1.5 to 3 g/L, t_½ increased by a factor of 4.3.

The relation between DCD concentration and time since application to a nutrient solution maintained at 25 decrees C and containing the soil bacteria strain EK1

Figure 1 The relation between DCD concentration and time since application to a nutrient solution maintained at 25 °C and containing the soil bacteria strain EK1. These data came from Table 1 of Hauser and Haselwandter (1990). The curve is a first-order exponential decomposition model, c(t) = c₀ e^-kt where c(t) is DCD concentration as a function of time, t, c₀ is DCD concentration when t is zero (0) and k is a decomposition constant. Regression analysis yielded c(t) = 2.7 e^-0.073t with a coefficient of determination (R²) = 0.97.

The relation between the half life (t½) of DCD, and the initial DCD concentration applied to a nutrient solution (c0) maintained at 25 degrees C and containing the soil bacteria strain EK1

Figure 2 The relation between the half life (t_½) of DCD, and the initial DCD concentration applied to a nutrient solution (c₀) maintained at 25 °C and containing the soil bacteria strain EK1. Using a model explained in the caption of Figure 1, t_½ values were calculated from the data in Table 1 of Hauser and Haselwandter (1990). Regression analysis yielded the exponential curve, t_½ (c₀) = 0.62 e^0.976c₀, that explained 95 % of the variance.

The rate of a biologically mediated process can also be described using Michaelis-Menton kinetics. For DCD, under constant environmental conditions, kinetics can define a relation between decomposition rate (D) and substrate concentration (S). When S is low, D increases with S as a hyperbolic curve, indicating a first-order reaction. For DCD, as shown, this means its decomposition rate in soil can depend on the applied quantity. However, when S is high, D is relatively constant, indicating a zero-order reaction. When saturated with substrate, D is at a maximum (D_max). A parameter, K_m, is the substrate concentration corresponding with D equal to half of D_max. The relation, exhibited by the data of Schwarzer and Haselwandter ((1991); see their Figure 1) based on pure culture studies with the soil bacteria strain EK1, may be written as:

D = [D_max + S] / [K_m + S].

To determine D_max and K_m, data are plotted according to the Lineweaver-Burke transformation with 1/S on the abscissa (X axis) and 1/D on the ordinate. The relation is thus transformed to a line with slope and intercept equal to K_m/D_max and 1/D_max, respectively.

Schwarzer and Haselwandter (1991) reported that DCD decomposition occurred only in a temperature range between 10 and 33 °C. They did not specify the incubation period(s) nor report half lives. At 25 °C, they obtained an optimum (fastest decomposition) rate, meaning the rate increased with increasing temperature between 10 and 25 °C but declined with increasing temperature between 25 and 33 °C. In contrast, they noted the inorganic DCD decomposition rate increased with increasing temperature up to at least 90 °C. Earlier, Amberger (1989) tabulated data (but did not quantitatively analyse it) showing that DCD decomposition occurred in an unspecified soil over a temperature range between 0 and 12 °C. Zero- and first-order decomposition models fitted these data well and yielded similar half lives (Figure 3). For 0 and 12 °C, t_½ was 147 and 42 days, respectively.

3 Relations between the mass of DCD remaining in soil, maintained at 6 degrees C, and time since its application (the rate was not given) based on data (shown here as symbols) tabulated by Amberger (1989)

Figure 3 Relations between the mass of DCD remaining in soil, maintained at 6 °C, and time since its application (the rate was not given) based on data (shown here as symbols) tabulated by Amberger (1989). The data were fitted to a first-order decomposition model by regression that accounted for 94 % of the variance. This is portrayed as the dashed curve, M(t) = M₀ e^-kt where M(t) is mass of DCD in the soil as a function of time, t, M₀ is mass of DCD applied to the soil when t is zero (0)(M₀ was normalised to unity on the ordinate) and k is a decomposition constant (0.0071 per day). The first-order model predicted that inhibitor mass had declined to half its application value in 98 days, t_½, = Ln(2)/k where Ln denotes natural logarithm. A zero order decomposition model was also fitted to the data by regression (M(t) = M₀ – kt, so t_½ = 1/[2k]) that accounted for 91 % of the variance and yielded t_½ = 109 days.

For a Lismore silt loam soil sampled beneath grazed pasture near Lincoln and incubated for 135 days at 8 and 20 °C, DCD decomposition was measured on six occasions (Di and Cameron, 2004a). These data were fitted to the first-order first-order exponential decomposition model. The soil was fertile with organic carbon and total nitrogen contents of 36.5 and 3.5 grams per kg of soil, respectively, and a pH of 5.9. The DCD application rates were 5.77 and 11.54 mg per kg of soil, equivalent to 7.5 and 15 kg DCD per hectare over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. The samples were maintained near the field capacity water content of 0.3 kg of water per kg of soil. For 8 °C, with 7.5 and 15 kg DCD per hectare, t_½ was 111 and 116 respectively, while for 20 °C, t_½ was 26 and 18 days (Figure 4). Consequently, doubling the DCD application rate did not correspond with a significant increase of t_½ as portrayed in Figure 2 based on the data of Hauser and Haselwandter (1990). For the data of Di and Cameron (2004a) at 8 and 20 °C, regression accounted for 70 and 94 % of the variance, respectively. Confidence limits were not stated for t_½ but an approximation of k ± 10 %, based on the data of Irigoyen et al. (2003), see Table 3), suggested a t_½ confidence interval of 100 – 122 days at 8 °C.

Relationships between the mass of DCD remaining in a Lismore silt loam soil, maintained near the field capacity water content (approximately 0.3 kg water per kg soil), and time since its application for two temperatures after Di and Cameron (2004a)

Figure 4 Relationships between the mass of DCD remaining in a Lismore silt loam soil, maintained near the field capacity water content (approximately 0.3 kg water per kg soil), and time since its application for two temperatures after Di and Cameron (2004a). These data were fitted to a first-order exponential model, M(t) = M₀ e^-kt where M(t) is mass of DCD in the soil as a function of time, t, M₀ is mass of DCD applied to the soil when t is zero (0)(M₀ was normalised to unity on the ordinate) and k is a decomposition constant determined by regression analyses (0.00625 per day and 0.0266 per day for 8 and 20 °C and portrayed as solid and dashed curves, respectively, when M₀ was 5.77 mg DCD per kg soil, equivalent to 7.5 kg DCD per hectare over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3). For 8 and 20 °C, the model predicted that inhibitor mass had declined to half its application value (called the half life, t_½, t_½ = Ln(2)/k where Ln denotes natural logarithm) in 111 and 26 days, respectively.

For silt loam soil sampled beneath pasture near Giessen, Germany, DCD decomposition was measured over 90 days at 10, 20 and 30 °C (Rajbanshi et al., 1992). These data were fitted to a zero-order decomposition model recognising three time points; namely at DCD application, at the end of a lag between application and the beginning of active mineralisation and at the end of active mineralisation. The model may be written as an equation, M_E = M_L - kt where M_E is mass of DCD in the soil at the end of active mineralisation, M_L is mass of DCD in the soil at the end of the lag, k is again a decomposition constant obtained by regression analysis and t is time since application. The equation may be used to determine t_½ from k when M_E is zero and M_L normalised to unity, so re-arrangement yields t (t_½) equal to 1/k. The application rates were equivalent to 10, 25 and 50 mg DCD per kg soil including the reported values of DCD-N mass, where N denotes nitrogen, divided by 0.65 for conversion to DCD mass. For comparison with Di and Cameron (2004a), these DCD application rates corresponded to 13, 32 and 65 kg DCD per hectare over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. This soil had organic carbon and total nitrogen contents of 29 and 2.5 g/kg of soil, respectively, and a pH of 6.1. During incubation, the samples were maintained at 0.4 kg of water per kg of soil. For 10, 20 and 30°C, with 13 and 32 kg DCD per hectare, corresponding values of t_½ were 52, 16 and 13 days and 70, 18 and 12 days. For 20 and 30°C with 65 kg DCD per hectare, t_½ was 22 and 15 days.

The value of t_½ reported by Rajbanshi et al. (1992) for 10 °C with 13 kg DCD per hectare (ie 52 days) was about half that reported by Di and Cameron (2004a) for 8 °C with 15 kg DCD per hectare (ie 116 days). The Giessen soil contained significantly less nitrogen than the Lismore soil. If the Giessen soil’s microbial community was relatively N-limited, it follows that the N-rich DCD would be mineralised faster following its application. Rajbanshi et al. also found that pretreatment of the soil with DCD reduced t_½ to 7 and 13 days at 10 °C with 13 and 32 kg DCD per hectare, respectively, and there was no lag between application and mineralisation. In contrast, Rogers (1986) found no pretreatment effect when applying 10 mg DCD per kg of soil for 4 years.

For a Decatur silt loam soil sampled in Alabama, USA, DCD decomposition was measured for 75 days at three temperatures by Bronson et al. (1989). The cropping soil had an organic carbon content of only 8.0 g/kg of soil and a pH of 6.2. The soil’s nitrogen content was not reported but the very low organic carbon content should be indicative. During incubation, the samples were maintained at 0.2 kg of water per kg of soil. These data were fitted to a zero-order model. The application rate, M₀, was 2 mg DCD per 20 g soil, equivalent to 130 kg DCD/ha over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. The decomposition constant, k, was 0.0382, 0.0775 and 0.1276 mg DCD per 20 g soil per day for 8, 15 and 22 °C, respectively, and the corresponding values of t_½ (t_½ = M₀/2k, equivalent to 1/k given the units used for M₀ and k) were 26, 13 and 8 days. These short half lives were exceptional and may reflect a very N-limited microbial community.

For a sandy loam soil sampled in Spain, following DCD application, ammonium ammonium was considered a proxy for DCD and its decomposition was measured for 105 days at three temperatures (Irigoyen et al., 2003). These data were fitted to the zero-order model according to Bronson et al. (1989). The application rate of DCD was 4 mg/kg soil, equivalent to 5.2 kg/ha over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. The cropping soil had organic carbon and Kjeldahl nitrogen contents of 7.3 and 1.3 g/kg of soil, respectively, and a pH of 7.5. During incubation, samples were maintained at 75% of field capacity or 0.1 kg of water per kg of soil. For 10, 20 and 30 °C, t_½ was > 105, 18 and 7 days, respectively. At 10 °C, the t_½ of ammonium was broadly similar to that expected for DCD in Lismore silt loam (Di and Cameron, 2004a).

For a soil sampled in France, following DCD application, ammonium decomposition was measured for 364 days at three temperatures (Guiraud and Marol, 1992). These data were fitted to a first-order model. The DCD application rate was 15 mg/kg soil, equivalent to 19.5 kg/ha over a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. An equal measure of ammonium thiosulphate was applied with the DCD in order to increase its effectiveness according to the authors. The cropping soil contained 21% clay, had organic carbon and total nitrogen contents of 11 and 1.3 grams per kg of soil, respectively, and a pH of 7.9. During incubation, samples were maintained at 67% of the field capacity water content. For 10, 15 and 20°C, the respective regressions accounted for 72, 98 and 94% of the variance and t_½ was 231, 77 and 14 days. Thus, when ammonium thiosulphate was applied with DCD, ammonium half lives were longer than generally found for DCD when DCD was applied to soils alone (e.g. Di and Cameron, 2004a).

To synthesise the data reviewed here, an inverse exponential relation was fitted to a plot of DCD half life and soil temperature (Figure 5). The sixteen measurements included a temperature range of 0 – 30 °C but with only five < 10 °C (Amberger, 1989; Di and Cameron, 2004a; Irigoyen et al., 2003). To quantify the decrease in half life per unit increase of temperature, the function portrayed in Figure 5 was differentiated with respect to temperature (Figure 6). This showed that the half life change is inversely proportional to the temperature but also that the change depends strongly on the temperature. At weather stations throughout New Zealand, beneath grass that was mowed regularly, soil temperature was measured daily (at 0900 h) at depths of 0.1 and 0.3 m (New Zealand Meteorological Service, 1983). Measurements at the 0.l m depth were considered most representative of a pasture plant root zone. During autumn, winter and early spring, soil temperature ranged from 4 – 13 °C according to the climatological data in Tables 6 and 7. For the efficacy of DCD, this is a significantly wide temperature range. For example, during May, average soil temperature varied from 7 °C at Invercargill up to 13 °C at Dargaville (Table 7). The relation fitted to the data in Figure 5 suggested that the corresponding values of DCD half life nearly halved from 81 to 44 days. Our analysis was based on data obtained under controlled temperatures. To our knowledge, there have been no reported measurements of DCD decomposition in the field. Moreover, the field efficacy of DCD application in reducing direct nitrous oxide emissions has not been statistically analysed with respect to soil temperature.

Finally, five of six DCD application field trials conducted in NZ had an average soil temperature of 8 °C and recorded 75 ± 11 % (average ± 95 % confidence limit) reduction in direct N₂O emissions, while corresponding values for the sixth trial were 11 °C and 61 % (Table 8, Di and Cameron 2006, Smith et al. 2007). The differences were not statistically significant. During late spring and summer, soil temperature in New Zealand ranges from 9 – 21 °C (Table 9).

Relation between the half life (t½, days) of DCD mixed into soil samples, incubated under controlled conditions, and the corresponding temperature (T, deg C)

Figure 5 Relation between the half life (t_½, days) of DCD mixed into soil samples, incubated under controlled conditions, and the corresponding temperature (T, °C). Over the period denoted t_½, the mass of DCD had declined to half its application value. These data are described in the text but exclude those of Bronson et al. (1989), Guiraud and Marol (1992) and the 65 kg DCD per hectare treatment data of Rajbanshi et al. (1992)) The New Zealand data of Di and Cameron (2004a) are portrayed by open symbols. The range of application rates was 5 – 32 kg DCD ha^-1 based on a soil depth of 0.1 m and bulk density of 1.3 Mg m^-3. Regression, portrayed as the curve, yielded the function t_½ (T) = 163 e^-0.1T that accounted for 91% of the variance.

Relation between the decrease of DCD half life (t½, days) and soil temperature (T, deg C) based on differentiating the functional relation portrayed in Figure 5

Figure 6 Relation between the decrease of DCD half life (t_½, days) and soil temperature (T, °C) based on differentiating the functional relation portrayed in Figure 5. The derivative, dt_½/dT = -16 e^-0.1T, is portrayed here as the curve.

Table 6 Five former weather stations of the New Zealand Meteorological Service located along a North to South transect and the period (years) of soil temperature measurement (New Zealand Meteorological Service, 1983).

Weather station	Latitude, Longitude, elevation (masl)	Period
Dargaville	35° 57’ S, 173° 50’ E, 20	1951 - 1980
Rukuhia	37° 50’ S, 175° 18’ E, 66	1946 - 1980
Palmerston North DSIR	40° 23’ S, 175° 37’ E, 34	1939 - 1980
Lincoln	43° 39’ S, 172° 28’ E, 11	1943 - 1980
Invercargill Airport	46° 25’ S, 168° 20’ E, 0	1951 - 1980

Table 7 Monthly average soil temperature during autumn, winter and early spring at five locations along a North to South transect described in Table 1. Measurements were made daily at 0900 hours and the thermometer was located at a depth of 0.1 m beneath mown grass. The averages were computed from 30 – 41 years of data. During these months, rainfall generally exceeds evaporation (Scotter and Kelliher 2004), so soils become wet. If the soil’s water storage capacity is exceeded, there will be drainage.

Month Location	May	Jun	Jul	Aug	Sep
Dargaville	12.8	10.9	9.6	10.4	12.2
Rukuhia	11.1	8.7	7.6	8.5	10.7
Palmerston North	10.1	7.7	6.7	7.6	9.9
Lincoln	7.4	4.5	3.9	5.0	7.5
Invercargill	6.7	4.6	3.5	4.3	6.5

Table 8 Data from Di and Cameron 2006 and Smith et al. 2007

Location	Soil	Soil temperature (°C), depth of 0.1 m			% direct N₂O emissions reduction	Date of DCD treatment application	Period (days) when N₂O emissions from urine-amended soils greater than urine-amended soils treated with DCD
		Max.	Min.	Mean (± Std Dev)
Lincoln	Templeton	12.2	3.4	7.9 ± 2.2	73	23 Jun. 2005	82
Lincoln	Lismore	11.6	3.4	7.4 ± 2.1	67	29 Apr 2005	88
Hamilton	Horotiu	15.7	5.7	10.7 ± 2.8	61	15 May 2005	33
Taupo	Taupo	13.8	2.9	8.5 ± 2.4	69	3 Aug 2005	56
Invercargill	Pukemutu	16.2	3.6	7.2 ± 1.8	75	27 Apr 2004	55
Invercargill	Pukemutu	12.4	5.8	9.5 ± 1.5	91	25 Aug 2005	62

Table 9 Monthly average soil temperatures during late spring and summer. Measurements were made daily at 0900 hours and the thermometer was located at a depth of 0.1 m beneath mown grass. The averages were computed from 30 – 41 years of data.

Month Location	Oct	Nov	Dec	Jan	Feb	Mar	Apr
Dargaville	14.9	17.4	19.5	20.8	20.4	18.6	15.6
Rukuhia	13.4	15.8	18.1	19.5	19.5	17.6	14.3
Palmerston North	12.5	15.1	17.3	18.5	18.1	16.3	13.2
Lincoln	10.8	13.8	16.4	17.4	16.7	14.3	10.9
Invercargill	9.0	11.1	13.3	14.1	13.6	12.1	9.5

Inventory revisions for pasture treated with nitrification inhibitors

We re-iterate that the direct N₂O emission factor for N fertiliser, EF₁, is the fraction of N applied to soils that is emitted directly into the atmosphere (the New Zealand specific value is 0.01 kg N₂O-N/kg fertiliser N). Likewise, the direct N₂O emission factor for excreta deposited by farmed animals during grazing, EF3_PR&P, is the fraction of N excreted onto soils that is emitted directly into the atmosphere (the New Zealand specific value is also 0.01 kg N₂O-N/kg excreta N). Indirect N₂O emissions are mainly associated with N leaching. This depends on the total N applied to soils as fertiliser and excreta. The fraction of N applied to soils that is leached is called FracLEACH and the New Zealand specific value is 0.07 (Thomas et al., 2005)).

N₂O emissions and NO₃^- leaching depend on the quantity of N applied to soils. DCD should last the longest and be most effective when soil temperature is lowest. Field trials demonstrated that DCD application significantly reduced direct N₂O emissions and NO₃^- leaching from urine applied to pasture when the soil temperature averaged < 12 °C. During May - September, the soil temperature ranges from 4 – 13 °C according to climatological data (Table 7).

New Zealand’s agricultural soils N₂O emissions inventory could readily be computed on a monthly basis. The nitrogen excretion rates are determined on a monthly basis by animal type and we believe credible estimates could be made for the corresponding nitrogen fertiliser rates.

Application rate for nitrification inhibitors

Based on the peer-reviewed literature, we conclude that a DCD application rate of 10 kg/ha most effectively reduced direct N₂O emissions and nitrate leaching (Di and Cameron, 2005). This rate was predominantly based upon two applications per year in the autumn and late winter. The effective period following DCD application is discussed below as well as the effect of repeated applications.

Formulation and application of nitrification inhibitors

Smith et al. (2005) compared various granular formulations of DCD and DCD in solution. They found no formulation effect on pasture herbage dry matter response in synthetic urine patches. However, there was no urine-only treatment for comparison. As discussed earlier, the results may have been affected by relatively warm and dry soil for most of the trial. The response to N applied may have been reduced due to a relatively high soil organic matter mineralization rate. Nevertheless, all the DCD formulations containing 15 and 30 kg DCD/ha were effective in limiting the nitrification of ammonium-N to nitrate-N for more than 100 days. These results suggested granular- and solution-based formulations of DCD may be effectively applied to soils.

Di and Cameron (2005) compared three methods of DCD application; namely, (1) DCD dissolved in water and irrigated onto the soil surface, (2) DCD dissolved in cattle urine and irrigated onto the soil surface and (3) DCD ground into a fine powder then made into a suspension (FPS) and sprayed evenly across the soil surface. The soil treated by FPS was irrigated after application, per recommended practice to simulate rainfall. The three methods maximised DCD coverage of the soil and they were found to be equally effective in reducing nitrate leaching.

The optimum timing for DCD application is concurrently or as close as possible to the deposition of urine-N. In practice on the farm this will not always be possible. However, a 10 day lag between urine and DCD applications to Templeton silt loam soil at Lincoln did not significantly affect DCD performance because N₂O emissions were reduced by 56%, compared with 61% for concurrent applications of urine and DCD (Di and Cameron, 2006). For the same soil, an 18 day lag between urine and DCD applications corresponded with N₂O emissions reduced by 73% (Di et al., 2006).

In principle, maximum contact with soil bacteria should make DCD most effective. For perspective, we begin with some linear dimensions. In the rhizosphere, soil located in the vicinity of plant roots, the average distance between bacteria cells is 10 μm according to Watts et al. (2006). Following rainfall or irrigation, pores in soils that are drained by the suction of gravity (1 metre of suction) have diameters larger than 30 μm. One day after drainage, the soil obtains a water content called its field capacity. We can use this linear perspective to examine the application of DCD to soils. As an example, we consider a granular formulation. From a sample of 260 zeolite grains containing the DCD product called DCn^TM, the average weight was 2.3 mg. If the application rate was 50 kg/ha, on average, there would be 2174 grains/m² on the soil surface. For this hypothetical situation, assuming the sample measurements are representative, the average distance between grains would be 21 mm. This distance is 2,100 times greater than 10 μm, the average distance between bacteria cells in the rhizosphere. However, a distribution of distances between grains is more useful for analysis. A distribution statistic is the coefficient of variation (CV), a ratio of standard deviation and average values, commonly reported as a percentage. For a centrfugal fertilizer spreader, the CV in application to a rectangular paddock was 43 % according to Lawrence and Yule (2007). If the average distance between grains, containing DCD, was 21 mm and CV 43 %, the standard deviation was 9 mm. With 95 % confidence, this statistic indicates the minimum and maximum distance between grains was 3 (21 – [2 X 9]) and 40 (21 + [2 X 9]) mm, respectively. Because the applied DCD will horizontally disperse into non-treated areas of soil between grains by the process called diffusion, we can do further analysis. We begin by assuming DCD in the grain dissolves completely to create a point on the soil surface where DCD concentration is equal to the application rate. We then solve the so-called continuity equation, including Fick’s law of diffusion, to obtain a useful one-dimensional expression relating distance travelled by a diffusing solute and time (Nobel 1983, see page 16). The relation shows that the distance depends on the square root of time divided by a term called the diffusion coefficient (equal to 2.2 X 10^-9 m² s^-1 for DCD dissolved in water, based on the molecular weight, 84 grams/mole). The dependence is also determined by the fraction of solute that has diffused away from the point of application. For DCD solute diffusing horizontally in the soil between grains on the surface, located as close as 3 mm apart, the distance of interest is 1.5 mm or the halfway point between grains closest to one another. The relation shows, for example, that it would take 1.4 hours for 95 % of the DCD solute to diffuse from a grain to a point 1.5 mm away. Although for simplicity the solution and relation considers diffusion in one dimension (that is, horizontally along a line), the process take place throughout the soil. For grains located as far apart as 40 mm apart, the distance of interest is 20 mm and it would take 10 days for 95 % of the DCD solute to diffuse 20 mm. In summary, these calculations support two seminal results from studies of DCD formulation and application; namely, (i) granular-, solution- and fine powder suspension-based formulations of DCD have been effectively applied to soils and (ii) a 10 day lag between urine and DCD applications did not significantly affect DCD performance.

Recording the area where nitrification inhibitors are applied

The use of DCD includes a requirement for accurate and verifiable records of the treated pasture/land/soils area. Long-term record storage and availability for independent review are also required. A GPS system associated with application seems ideal. This system should be future proof and suitable for audit and accreditation of farm scale carbon credits. The system may have wider application to monitoring N use and losses from farms.

Recording variables to estimate nitrogen fertilizer application and excretion rates

Linked to the GPS record of land area covered by DCD application, there must be a grazing stock record. Hence the number and type of animals (dairy cattle, beef cattle and sheep) ‘treated’ with DCD could be calculated from the record and subsequently used in the IPCC inventory for New Zealand (Clough et al., 2007). From the number and type of animals ‘treated’ with DCD, compared to the national population, one could proportionally determine the required N excretion rate. Alternatively, at the farm scale, animal weight and production rate could be used for determination of pasture herbage N intake and output to product. Nitrogen fertilizer rate on the land ‘treated’ with DCD also needs to be recorded.

As mentioned earlier, for some dairy farms, grazing may take place during much of the ‘May – September period on ‘support’ land that may not be located within a farm’s boundary. There may also be a ‘cut and carry’ feeding regime during ‘May – September’ including feed produced outside the farm’s boundary. This may involve ‘stand off’ and feed pads.

Comparison of emissions factors from other countries

To our knowledge, no other country has revised its emission factors to account for the effect(s) of nitrification inhibitor application onto soils. However, in the absence of nitrification inhibitors, we can compare New Zealand’s specifc values of EF₁ and EF_3(PRP) (both equal to 0.01) with those of the IPCC and recently completed international peer-reviewed literature data syntheses. For Frac_LEACH, the reader is directed to Thomas et al. (2005).

Revised emission factors and emissions inventory: Method 1 - ‘Annualised’ revisions of EF_3(PRP), EF₁ and Frac_LEACH using nitrification inhibitors

The revised emission factors in this section will be ‘annualised’ figures, compatible with the current method of emissions inventory calculation that was described earlier. In turn, the revised emission factors will be used to revise the N₂O emissions inventory calculation. This revision of the emissions inventory will be called Method 1. Our recommendations are predicated on application criteria that have proved successful for DCD in research trials. Firstly, the nitrification inhibitor should thoroughly cover the soil. Each application rate should be at least 10 kg/ha. There should be two applications each year following grazing in autumn (May) and again in late winter (August). Given soil temperatures throughout New Zealand during May – September, we believe that effectiveness can be expected for these 5 months of the year. Hence, DCD is not used for seven months of the year (October – April), so the NZ specific values for emission factors do not require revision. We do not think there has been sufficient research to make recommendations for a range of application methods and rates. For the same reason, insufficient research prevents us from making recommendations based on the anticipated effects of soil drainage on nitrification inhibitor use efficacy (Kelliher et al. 2005a, 2005b). While the efficacy of nitrification inhibitor use has been evaluated in soils subjected to autumn and late winter applications, except for the pasture production data of Moir et al. (2007), the results from research trials in New Zealand that were available to us were limited to single years of study. Further research is needed to quantify the effect(s) of repeated nitrification inhibitor applications over several years.

When DCD was applied to four New Zealand pastoral soils, EF_3(PRP) was reduced by 61 to 74 % (n = 4, mean 67 %, Std. Dev. 6 %). These statistics define the uncertainty of nitrification inhibitor application to soils with respect to EF_3(PRP). The mean value determines our recommendation. Consequently, when DCD is applied as recommended, we revise EF_3(PRP) as follows

EF3_PRP 'plus DCD' = (7/12 x NZ specific EF3_PRP)+(5/12 x NZ specific EF3_PRP x (1-0.67)) = 0.007

For EF₁, revision should be the same as for EF_3(PRP). Urine is the primary N excreta constituent in the N₂O emissions inventory and there is no scientific evidence to suggest that urea from fertiliser behaves differently than urea from urine (Kelliher et al. 2005a, Kelliher and de Klein 2006). We also recommend estimating the uncertainty of EF₁ using the EF_3(PRP) uncertainty statistics, presuming N fertiliser applications will be done in autumn and late winter. We do not think there has been sufficient research to make any further recommendations for fertiliser. Consequently, when DCD is applied as recommended, we revise EF₁ as follows

EF₁ 'plus DCD' = (7/12 x NZ specific EF₁) + (5/12 x NZ specific EF₁ x (1-0.67)) = 0.007

When DCD was applied to New Zealand pastoral soils, Frac_LEACH was reduced by 68 to 77 % (n = 5, mean = 74% Std Dev. 4 %). These statistics define the uncertainty of nitrification inhibitor application to soils with respect to Frac_LEACH. The mean value determines our recommendation. Consequently, when DCD is applied as recommended, we revise Frac_LEACH as follows:

Frac_LEACH 'plus DCD' = (7/12 x NZ specific Frac_LEACH) + (5/12 x NZ specific Frac_LEACH x (1-0.074)) = 0.05

Revised emission factors and emissions inventory: Method 2 - Disaggregating nitrogen application onto soils for nitrification inhibitor responses

This method uses the same ‘annualised’ emission factors developed in the previous section. However, N application onto soils will be disaggregated for this emissions inventory method into two periods; namely, May – September when DCD is used and effective and October – April when DCD is not used. This method recognises that N excretion as urine and dung is computed monthly for dairy cattle, beef cattle, sheep and deer. This facilitates sums of figures for the seven months of October – April when DCD would not be applied according to our recommendations. Figures can also be summed for the five months of May – September when DCD would be applied and effective. We recognise that the monthly figures can be summed for other periods as well.

The application of N fertiliser onto soils also has to be disaggregated. Our figures come from expert judgement (Hilton Furness, FertResearch, Personal Communication, 8 March 2007). We begin by disaggregating N fertiliser application by animal type. Currently, 70 % of all N fertiliser that is sold annually is applied to soils associated with dairy cattle. Further, and currently, 10 % of all N fertiliser that is sold is applied to soils associated with beef cattle, sheep and deer. We approximate the partitioning as 5 % applied to soils associated with beef cattle and 5 % to sheep. Consequently, no N fertiliser is applied to soils associated with deer. Finally, in this report, we do not include 20 % of the N fertiliser sold annually because it is applied to soils associated with arable and horticultural crops. To re-iterate, this report includes only 80 % of the N fertiliser sold annually because this is the quantity thought to have been applied to soils associated with grazing animals. Finally, the current percentages of N fertiliser applied to soils associated with dairy cattle (70 %), beef cattle (5 %) and sheep (5 %) are used in our calculations for the years 1990, 2004 and 2010. To monitor these percentages, we recommend that FertResearch is the best available source of information.

The application of N fertiliser onto soils also has to be disaggregated for the calculation of nitrification inhibitor response. We re-iterate that there is no scientific evidence to suggest that urea from fertiliser behaves differently than urea from urine. For this report, N fertiliser was applied in two equal dressings (each half of the annual quantity) in early May and early August along with the DCD applications.

Revised emission factors and emissions inventory: Method 3 - Disaggregating the EF_3(PRP) and EF₁ data for revision by nitrification inhibitor response

This method uses the same (two period) disaggregation of N application onto soils as Method 2; namely, May – September when DCD is used and effective and October – April when DCD is not used. In addition, Method 3 uses this (two period) disaggregation for EF_3(PRP), EF₁ and Frac_LEACH.

For dairy cattle urine, data from the 17 NzOnet trials given in Table A.1 were disaggregated into the two seasonal periods; namely, spring + summer, representing the seven months of October – April, and autumn + winter, representing the five months of May – September. For spring + summer, the 9 trials yielded a geometric average of 0.006. For autumn + winter, the 8 trials yielded a geometric average of 0.014. For Method 3, these two seasonal values should be used when nitrification inhibitors are NOT applied to soils. In the absence of nitrification inhibitors, these two seasonal values should also be used for EF_3(PRP) of beef cattle urine excreted onto soils and EF₁ of nitrogen fertiliser applied to soils (fraction of applied nitrogen emitted to the atmosphere as nitrous oxide).

For dairy cattle dung, the 6 trials yielded a geometric average of 0.002 (Table A.2). There were not enough data for seasonal disaggregation. Consequently, a constant EF_3(PRP) value of 0.002 should be used throughout the year when nitrification inhibitors are NOT applied to soils. In the absence of nitrification inhibitors, this constant value should also be used for EF_3(PRP) of beef cattle dung excreted onto soils.

For sheep urine, the 4 trial values yielded a geometric average of 0.002 (Table A.3). The corresponding value for sheep dung was 0.0001 based on 2 trials. These values should be used throughout the year for the EF_3(PRP) of sheep urine and dung when nitrification inhibitors are NOT applied to soils.

We have recommended that DCD be applied to soils twice each year in autumn and late winter. We also recommended that DCD would be effective during the autumn + winter period, representing the five months of May – September. Accordingly, DCD application corresponds with no change of EF_3(PRP) or EF₁ during the seven months of October – April. When DCD was applied to New Zealand pastoral soils, EF_3(PRP) of cattle urine was reduced by 61 to 74 % (n = 4, mean 67 %, Std. Dev. 6 %). These statistics define the uncertainty of nitrification inhibitor application to soils with respect to EF_3(PRP). The mean value determines our recommendation. Consequently, when DCD is applied as recommended, we revise the cattle urine EF_3(PRP) during the five months of May – September as follows

Cattle urine EF_3(PRP) 'plus DCD' -= [0.014 X (1 – 0.67)] = 0.0046

EF₁ 'plus DCD' -= [0.014 X (1 – 0.67)] = 0.0046

In the absence of trials, we apply the same logic to the revision of sheep urine EF_3(PRP) in response to DCD application. Consequently, we write

Sheep urine EF_3(PRP) 'plus DCD' -= [0.002 X (1 – 0.67)] = 0.00066

Again, in the absence of trials, we cannot be sure if the same logic should be used for the revision of cattle and sheep dung EF_3(PRP) in response to DCD application. Consequently, it can only be an approximation when we write

Cattle dung EF_3(PRP) 'plus DCD' -= [0.002 X (1 – 0.67)] = 0.00066

Sheep dung EF_3(PRP) 'plus DCD' -= [0.0001 X (1 – 0.67)] = 0.000033

When DCD was applied to New Zealand pastoral soils, Frac_LEACH was reduced by 68 to 77 % (n = 5, mean = 74% Std Dev. 4 %). These statistics define the uncertainty of nitrification inhibitor application to soils with respect to Frac_LEACH. The mean value determines our recommendation. Consequently, when DCD is applied as recommended, we revise the Frac_LEACH during the five months of May – September as follows

Frac_LEACH 'plus DCD' -= [0.07 X (1 – 0.74)] = 0.0182

Contact for Enquiries

Sustainable Land Management and Climate Change
MAF
Pastoral House
25 The Terrace
PO Box 2526, Wellington
Tel: 0800 CLIMATE (254 628)
Contact this person

Developing revised emission factors for nitrous oxide emissions from agricultural pasture treated with nitrification inhibitors