View all publications

Chapter 6: Agriculture

6.1 Sector overview

The agriculture sector emissions totalled 37,203.24Gg CO2 equivalent and represented 49.4% of all greenhouse gas emissions in 2003. Emissions in this sector are now 15.6% higher than the 1990 level of 32,193.76Gg CO2 equivalent (Figure 6.1.1). The increase is attributable to a 9.6% increase in CH4 emissions from enteric fermentation and a 29.4% increase in N2O emissions from the agricultural soils category.

Figure 6.1.1 Agricultural sector emissions from 1990 to 2003

 

Year Gg CO2 equivalent
1990 32,193.76
1991 32,119.70
1992 32,140.22
1993 32,739.36
1994 33,316.76
1995 33,770.26
1996 33,993.67
1997 34,223.33
1998 34,341.70
1999 34,757.85
2000 35,509.21
2001 36,349.64
2002 36,762.40
2003 37,203.24

Emissions of CH4 from enteric fermentation dominate the sector producing 63.4% of CO2 equivalent emissions in the sector (Figure 6.1.2) and 31.3% of New Zealand's total emissions. N2O emissions from agricultural soils are the other major component of the sector comprising 34.9% of agricultural emissions.

Agriculture is the principal industry of New Zealand and agricultural products are the predominant component of exports. This is due to several factors: the favourable temperate climate, the abundance of agricultural land and the unique farming practices used in New Zealand. These practices include the extensive use of all year round grazing systems and a reliance on nitrogen fixation by legumes rather than nitrogen fertiliser.

Figure 6.1.2 Emissions from the agricultural sector in 2003 (all figures Gg CO2 equivalent)

Category Gg CO2 equivalent Percent of total
Enteric Fermentation 23592.21 63.4
Manure Management 610.51 1.6
Agricultural Soils 12968.99 34.9
Field Burning of Agricultural Residues 29.93 0.1
Prescribed Burning of Savannas 1.0 0.0

Since 1984, there have been changes in the balance of livestock species. There has been a trend for increased dairy production and deer numbers for meat and velvet production due to the prevailing good world prices. This has been counterbalanced by land coming out of sheep production and consequently decreased sheep numbers. Beef numbers have remained relatively static. There have also been productivity increases across all major animal species and classes. At the same time there has been an expansion of the land used for plantation forestry. The land area used for horticulture has not changed significantly since 1990 although the types of produce grown have changed with less grain but more vegetables, fruit and grapes for wine production.

New Zealand uses a June year for all animal statistics and reports a rolling three year average in the inventory. The June year reflects the natural biological cycle for animals in the southern hemisphere. To maintain consistency, a single livestock population characterisation is used as the framework for estimating CH4 emissions from enteric fermentation, CH4 and N2O emissions from manure management and N2O emissions from animal production. A complete time-series of the agriculture data is shown in Annex 8.4 and information on livestock population census and survey procedures are included in Annex 3.1.

6.2 Enteric fermentation (CRF 4A)

6.2.1 Description

In 2003, emissions from enteric fermentation comprised 23,592.21Gg CO2 equivalent. This represents 31.3% of New Zealand's total CO2 equivalent emissions. The category is dominated by emissions from cattle which represent 57.6% of emissions from enteric fermentation. The current level of emissions from enteric fermentation is 9.6% above the 1990 level, however there have been large changes within the category. The largest increase has been in emissions from dairy cattle which have increased 70.3% since 1990. This increase has been offset by decreases in emissions from sheep (-18.9%), goats (-86.1%) horses (-16.0%) and swine (-13.8%).

CH4 is produced as a by-product of digestion in ruminants e.g. cattle and some non-ruminant animals such as swine and horses. Ruminants are the largest source of CH4 as they are able to digest cellulose. The amount of CH4 released depends on the type, age and weight of the animal, the quality and quantity of feed, and the energy expenditure of the animal.

CH4 emissions from enteric fermentation have been identified as the largest key category for New Zealand in the level assessment (excluding LULUCF). In accordance with good practice (IPCC, 2000), the methodology for estimating CH4 emissions from enteric fermentation in domestic livestock was revised to a Tier 2 methodology for the 2001 and subsequent inventories.

6.2.2 Methodological issues

New Zealand's methodology uses a detailed livestock population characterisation and livestock productivity data to calculate feed intake for the four largest categories in the New Zealand ruminant population (dairy cattle, beef cattle, sheep and deer). The amount of CH4 emitted is calculated using CH4 emissions per unit of feed intake (Figure 6.2.1). The calculation process is explained in the full description of the Tier 2 approach in Annex 3.1 and Clark et al.(2003).

There has been a gradual increase in the IEFs for dairy cattle and beef cattle from 1990 to 2003. This is to be expected because the methodology is able to use animal performance data that reflects the increased levels of productivity achieved by New Zealand farmers since 1990. Increases in animal weight and animal performance (milk yield) require increased feed intake by the animal to meet energy demands. Increased feed intake produces increased CH4 emissions per animal. The increases in productivity are shown in the agricultural worksheets in Annex 8.4 and in the detailed description in Annex 3.1.

Figure 6.2.1 Schematic of New Zealand's enteric methane calculation methodology

6.2.3 Uncertainties and time-series consistency

Animal numbers

Many of the calculations in this sector require livestock numbers. Both census and survey methods are used with surveys occurring in the intervening years between each census. Detailed information from Statistics New Zealand on the census and survey methods is included in Annex 3.1.2.

Methane emissions from enteric fermentation

In the 2001 inventory, the CH4 emissions data from domestic livestock in 1990 and 2001 were subjected to Monte Carlo analysis using the software package @RISK to determine the uncertainty of the annual estimate (Clark et al., 2003; Table 6.2.1). For the 2003 inventory, the uncertainty in the annual estimate was calculated using the 95% confidence interval determined from the Monte Carlo simulation as a percentage of the mean value i.e. in 2001, the uncertainty in annual emissions was ± 53%.

The overall inventory uncertainty analysis shown in Annex 7 (Good Practice Table 6.1) demonstrates that the uncertainty in annual emissions from enteric fermentation is 12.0% of New Zealand's total emissions and removals for 2003, and is the largest single component affecting the national total. However, in the trend from 1990 to 2003 the uncertainty from enteric fermentation is only 1.9% of the trend in emissions and removals. The uncertainty between years is assumed to be correlated, therefore the uncertainty is mostly in the emission factors and the uncertainty in the trend is much lower than uncertainty for an annual estimate.

Table 6.2.1 Uncertainty in the annual estimate of enteric CH4 emissions in 1990, 2001 and 2003 and the 95% confidence interval (± 1.96 standard deviations from the mean) estimated using Monte Carlo simulation

Year Enteric CH4 emissions (Gg/annum) 95% CI Min 95% CI Max

1990

1,015.5

478.1

1,552.9

2001

1,099.4

517.6

1,681.2

2003

1,123.4

530.0

1,718.8

Note: The CH4 emissions used in the Monte Carlo analysis exclude those from swine and horses.

Uncertainty in the annual estimate is dominated by variance in the measurements used to determine the 'CH4 per unit of intake' factor. For the measurements made of this factor, the standard deviation divided by the mean is equal to 0.26. This uncertainty is thought to be mostly natural variation from one animal to the next. Uncertainties in the estimation of energy requirements, herbage quality and population data are thought to be much smaller (0.005 - 0.05), so these variables play a much smaller role.

6.2.4 Source-specific QA/QC and verification

CH4 emission rates measured for 20 selected dairy cows scaled up to a herd have been corroborated using micrometeorological techniques. Laubach and Kelliher (in press) used the integrated horizontal flux (IHF) technique and the flux gradient technique to measure CH4 flux above a dairy herd. Both techniques are comparable, within estimated errors, to scaled-up animal emissions. The emissions from the cows measured by IHF and averaged over three campaigns are 329 (±153) g CH4/day/cow compared to 365 (± 61) g CH4/day/cow for the scaled-up measurements reported by Waghorn et al. (2002) and Waghorn et al. (2003).

6.2.5 Source-specific recalculations

For the 2003 inventory, there are revisions due to a modification of the Tier 2 model to use industry data for milk yield rather than MAF estimates, using differential growth rates for stock up to weaning. This includes an allowance of N in milk powder fed to calves and correcting small errors in population models and data inputs. The provisional livestock population data for 2003 was updated to actual numbers and the corresponding three year average populations for 2001, 2002 and 2003 updated.

Sheep numbers used for the calculations between 1993 and 1996 were corrected. In these years a category 'other' appeared on the Statistics New Zealand data. Expert opinion is that these animals were probably new born lambs that were on the farm when the census is completed. In the 2003 inventory, these animals were treated as lambs and added to the lamb total for the purposes of estimating CH4 and N2O emissions.

6.2.6 Source-specific planned improvements

A national inter-institutional ruminant CH4 expert group was formed to identify the key strategic directions for research into the CH4 inventory and mitigation to maximise the benefit of the existing programmes, and to develop a collaborative approach to improve the certainty of CH4 emissions. This is funded through the MAF. A private sector funded Pastoral Greenhouse Gas Research Consortium has been established to carry out research primarily into mitigation technologies and management practices but also on-farm inventory considerations. The implementation of the Tier 2 approach for CH4 emissions from enteric fermentation and manure management is a consequence of the research conducted by the expert group.

6.3 Manure management (CRF 4B)

6.3.1 Description

CH4 and N2O are produced during the anaerobic decomposition and storage of manure. Emissions from the manure management category comprised 2.3% of emissions from the agriculture sector.

Livestock manure is composed principally of organic material. When the manure decomposes in the absence of oxygen, methanogenic bacteria produce CH4. The emissions of CH4 are related to the amount of manure produced and the amount that decomposes anaerobically. CH4 from manure management has been identified as a key category for New Zealand in the 2003 level and trend assessments (Tables 1.5.2 and 1.5.3).

This category also includes emissions of N2O related to manure handling before the manure is added to the soil. The amount of N2O released depends on the system of waste management and the duration of storage. With New Zealand's extensive use of all year round grazing systems, this category is relatively small at 62.0Gg CO2 equivalent of N2O in 2003. In comparison, agricultural soil emissions of N2O totalled 12,969.0Gg CO2 equivalent.

6.3.2 Methodological issues

Methane

Estimates of CH4 emissions from manure management for cattle, sheep and goats are derived from Joblin and Waghorn (1994). Joblin and Waghorn used stock numbers, feed intake and digestibility data from Ulyatt (1992) to estimate total faecal output from cattle, sheep, goats and deer at approximately 16 million tonnes dry weight in 1990. The CH4 estimates are based on the maximum potential emissions of CH4 from animal waste. Therefore, actual emissions are likely to be substantially lower than the values reported (Joblin and Waghorn, 1994). The same emission factors are used for each year of the inventory (Table 6.3.1).

Table 6.3.1 Derivation of New Zealand emission factors for CH4 emissions from manure

Animal class Faecal dry matter (1000 t) Estimated maximum CH4 potential (1000 t) Emissions factor (kg/animal/year)

Dairy cattle

2683.6

3.1

0.889

Non-dairy cattle

3647.5

4.2

0.909

Sheep

9009.1

10.3

0.178

Goats

115.4

0.1

0.119

Deer

313.5

0.4

0.369

Total

15769.1

   

Although these emission factors are much lower than the IPCC default values, which are 32 and 6 kg CH4/head/year for dairy and beef respectively, the IPCC defaults are not applicable to New Zealand conditions. For example, in New Zealand, dairy cows and beef animals are managed similarly on the pasture yet the IPCC emissions factor varies by a factor of 5. New Zealand specific emission factors are not available for CH4 emissions from manure management for swine, horses and poultry. Emissions estimates for these species use IPCC default emission factors (refer to the agricultural worksheets in Annex 8.4).

Nitrous oxide

For the N2O calculation, six alternative regimes for treating animal manure, known as animal waste management systems (AWMS) are identified in the 1996 IPCC Guidelines. New Zealand farming uses four AWMS: (1) anaerobic lagoons, (2) pasture, range and paddock, (3) solid storage and dry-lot, and (4) other systems (poultry without bedding and swine deep litter). With the exception of dairy cattle, animals were allocated to the different AWMS according to the information provided in the IPCC 1996 guidelines for the Oceania region as these were applicable to New Zealand farming practices. For dairy cattle, New Zealand specific data from Ledgard and Brier (2004) were used.

The pasture, range and paddock AWMS is the predominant regime for animal waste in New Zealand. All sheep, goats, deer and non-dairy cattle excreta are allocated to the pasture, range and paddock AWMS. For dairy cattle, 95% of excreta are allocated to pasture, range and paddock and 5% is allocated to anaerobic lagoons. Emissions from the pasture, range and paddock AWMS are reported in the 'Agricultural soils' category.

The calculation for nitrogen in each AWMS is shown in the agricultural worksheets in Annex 8. A time-series of Nex values used for calculating animal production N2O emissions is also shown in the worksheets. The Nex values show an increase over time caused by the increases in animal production. The nutrient input/output model OVERSEER® is used to determine the annual quantities of nitrogen deposited in excreta by grazing animals. The OVERSEER® model uses the same animal populations and feed intake from the Tier 2 model used to determine methane emissions (Clark et al., 2003), and an assessment of feed nitrogen content.

6.3.3 Uncertainties and time-series consistency

Emission factors from manure and manure management systems, the livestock population, nitrogen excretion rates and the usage of the various manure management systems are the main factors causing uncertainty in N2O emissions from manure management (IPCC, 2000). New Zealand uses the IPCC default values for EF3 (direct emissions from waste) for all AWMS except for AWMSPRP (manure deposited on pasture, range and paddock) which is a country specific factor (0.01 kg N2O-N/kg N). The IPCC default values have uncertainties of

-50% to +100% (IPCC, 2000).

The overall inventory uncertainty analysis shown in Annex 7 (Good Practice Table 6.1) demonstrates that the effect of uncertainty in annual emissions from manure management is relatively minor compared to the effect of CH4 emissions from enteric fermentation and N2O from agricultural soils.

6.3.4 Source-specific QA/QC and verification

CH4 emissions from manure management were identified as a key category for New Zealand in the 2002 inventory. In preparation of the 2003 inventory, the data for this category underwent a Tier 1 QC checklist (refer Annex 6 for examples).

6.3.5 Source-specific recalculations

There were substantial recalculations for this sector in the 2003 inventory. The major change was a redistribution of waste between AWMS for dairy cows. In previous inventories, 11% of dairy excreta were in anaerobic lagoons and 89% on pasture, range and paddock. A MAF commissioned study by Ledgard and Brier (2004) showed that values of 5% in lagoons and 95% on pasture are a more accurate representation of New Zealand dairy farming. The change in allocation of excreta was peer-reviewed by an independent scientist before being incorporated in the New Zealand inventory.

The CH4 emissions factor for goats was changed from 0.12kg CH4/head/yr (the IPCC value for a cool climate) to 0.18kg CH4/head/yr (the IPCC value for temperate climate). All other livestock using IPCC values are allocated to a temperate climate. The change was made to make the allocation of goats consistent with other livestock. Revisions to the Tier 2 model to calculate dry matter intake also affected the Nex and consequently the N2O emissions from manure management (refer to section 6.2.5).

6.3.6 Source-specific planned improvements, if applicable

New Zealand has developed a Tier 2 methodology to improve the estimate of CH4 emissions from manure management. However, time constraints meant this method was not used for the 2003 inventory. Estimates using the Tier 2 method will be included in the next inventory submission in 2006.

6.4 Rice cultivation (CRF 4C)

6.4.1 Description

There is no rice cultivation in New Zealand. The 'NO' notation is used in the CRF.

6.5 Agricultural soils (CRF 4D)

6.5.1 Description

The agricultural soils category is the source of most N2O emissions in New Zealand comprising 12,969.0Gg CO2 equivalent in 2003. Emissions are 29.4% over the level in 1990. The category comprises three sub-categories:

  • direct N2O emissions from animal production (the pasture, range and paddock AWMS)
  • indirect N2O from nitrogen lost from the field as NOx or NH3
  • direct N2O emissions from agricultural soils as a result of adding nitrogen in the form of synthetic fertilisers, animal waste, biological fixation, inputs from crop residues and sewage sludge.

All three of these sub-categories have been identified as key sources for New Zealand (Tables 1.5.2 and 1.5.3). Direct soil emissions from animal production is the fourth largest key category comprising 7,455.5Gg CO2 equivalent, indirect N2O from nitrogen used in agriculture comprised 3,329.4Gg CO2 equivalent and direct N2O emissions from agricultural soils comprised 2,185.5Gg CO2 equivalent.

CO2 emissions from limed soils are included in the LULUCF sector.

6.5.2 Methodological issues

N2O emissions are determined using the IPCC 1996 approach where emission factors dictate the fraction of nitrogen deposited on the soils that is emitted into the atmosphere as N2O. The two main inputs in New Zealand are from nitrogen fertiliser and the excreta deposited during animal grazing.

The worksheets for the agricultural sector (Annex 8.4) document the emission factors and other parameters used in New Zealand's calculations. Two New Zealand specific factors/parameters are used: EF3PRP and FracLEACH. These factors were extensively reviewed for the 2001 submission, and a new value for FracLEACH was used from the 2001 inventory onwards and back-calculated to 1990.

Animal production (N2O)

Direct soil emissions from animal production refers to the N2O produced from the pasture, range and paddock AWMS. This AWMS is the predominant regime for animal waste in New Zealand as 95% of dairy cattle and 100% of sheep, goats, deer and non-dairy cattle are allocated to it. The emissions calculation is based on the livestock population multiplied by nitrogen excretion (Nex) values and the percentage of the population on the pasture, range and paddock AWMS. The Nex and allocation to AWMS are discussed in section 6.3.2 - nitrous oxide. The Nex values have been calculated using the model OVERSEER® based on the same animal intake values used for calculating CH4 emissions for the different animal classes and species. This ensures that the same base values are used for both CH4 and N2O emission calculations.

New Zealand uses a country-specific emissions factor for EF3PRP of 0.01 (Carran et al., 1995; Muller et al., 1995; de Klein et al., 2003; and Kelliher et al., 2003). Considerable research effort has gone into establishing a better quantification of EF3PRP. Field studies have been performed as part of a collaborative research effort called NzOnet. The parameter EF3PRP has been measured by NzOnet researchers in the Waikato (Hamilton), Canterbury (Lincoln) and Otago (Invermay) regions for pastoral soils of different drainage class (de Klein et al., 2003). These regional data are comparable because the same measurement methods were used at the three locations. The percentage of applied nitrogen (as urine or dung) emitted as N2O, and environmental variables, were measured in three separate trials that began in autumn 2000, summer 2002 and spring 2002. Measurements were carried out for up to 250 days or until urine treated pasture measurements dropped back to background emission levels.

Kelliher et al. (2003) assessed all available EF3PRP data and their distribution with respect to pastoral soil drainage class to determine an appropriate national, annual mean value. The complete EF3PRP data set of NzOnet was synthesised using the national assessment of pastoral soils drainage classes. This study recognises that (1) environmental (climate) data are not used to estimate N2O emissions using the IPCC 1996 methodology, (2) the N2O emissions rate can be strongly governed by soil water content, (3) soil water content depends on drainage that can moderate the effects of rainfall and drought, and (4) as a surrogate for soil water content, drainage classes of pastoral soils can be assessed nationally using a geographic information system. In New Zealand, earlier analysis showed the distribution of drainage classes for pasture land is highly skewed with 74% well-drained, 17% imperfectly drained and 9% poorly drained (Sherlock et al., 2001).

The research and analysis to date indicates that if excreta is separated into urine and dung, EF3 for urine and dung could be set to 0.007 and 0.003 respectively. However, it is recognised that the dung EF3 data are limited. Research by NzOnet is continuing into the possibility of separate emission factors for dung and urine. Combining urine and dung EF3 values, the dairy cattle total excreta EF3 is 0.006. Conservatively rounding the total excreta EF3 of 0.006 provides a country-specific value of 0.01 for EF3PRP. The IPCC default value of EF3PRP is 0.002.

Indirect N2O from nitrogen used in agriculture

The N2O emitted indirectly from nitrogen lost from agricultural soils through leaching and run-off is shown in the agricultural worksheets in Annex 8. This nitrogen enters water systems and eventually the sea with quantities of N2O being emitted along the way. The amount of nitrogen that leaches is taken as a fraction that is deposited or spread on land (FracLEACH).

Research studies in New Zealand together with a literature review have shown lower rates of nitrogen leaching than is suggested in the IPCC guidelines. In inventories reported prior to 2003, a New Zealand parameter for FracLEACH of 0.15 was used. However, IPCC-based estimates for different farm systems were found on average to be 50% higher than those estimated using the OVERSEER® nutrient budgeting model (the model provides average estimates of the fate of nitrogen (N) for a range of pastoral, arable and horticultural systems. In pastoral systems, N leaching is determined by the amount of N in fertilizer, farm dairy effluent and that excreted in urine and dung by grazing animals. The latter is calculated from the difference between N intake by grazing animals and N output in animal products, based on user inputs of stocking rate or production and an internal database with information on the N content of pasture and animal products). The estimates were closer for farms using high rates of N fertilizer, indicating that the IPCC-based estimates for N leaching associated with animal excreta were too high. When the IPCC methodology was applied to field sites where N leaching was measured (four large scale, multi-year animal grazing trials), it resulted in values that were two-fold higher than measured values. This indicated that a value of 0.07 for FracLEACH more closely followed actual field emissions (Thomas et al., 2002) and this value was adopted and used for all years as it reflected New Zealand's national circumstances.

Direct N2O emissions from agricultural soils

Direct emissions from agricultural soils are calculated in the five tables of worksheet 4.5.

The emissions arise from synthetic fertilizer use (SN), spreading animal waste as fertilizer (AW), nitrogen fixing in soils by crops (BW) and decomposition of crop residues left on fields (CR). All of the nitrogen inputs are collected together and an emissions factor applied to calculate total direct emissions from non-organic soils.

Nitrogen fertiliser use is determined by the New Zealand Fertiliser Manufacturers' Research Association (FertResearch) from sales records for 1990 to 2004. A rolling three year average is used to calculate inventory data. There has been a five fold increase in nitrogen fertiliser use over the 12 years, from 51,786 tonnes in 1990 to 298,380 tonnes in 2003. The calculation of N2O that is emitted indirectly through synthetic fertilizer and animal waste being spread on agricultural soils is shown in the agricultural worksheets in Annex 8. Some of the nitrogen contained in these compounds is emitted into the atmosphere as ammonia (NH3) and NOx through volatilisation, which returns to the ground during rainfall and is then re-emitted as N2O. This is shown as an indirect emission of N2O.

The FAW calculation for animal waste includes all manure that is spread on agricultural soils irrespective of which AWMS it was initially stored in. This includes all waste in New Zealand except for emissions from the pasture range and paddock AWMS. New Zealand uses the IPCC default value of EF1 of 0.0125 kg N2O-N/kg N.

Direct N2O emissions from organic soils are calculated by multiplying the area of cultivated organic soils by an emissions factor. Recent analysis identified 202,181 hectares of organic soils of which it is estimated that 5% (i.e. 10,109 ha) are cultivated on an annual basis (Kelliher et al., 2003). New Zealand uses the IPCC default emissions factor (EF2 equal to 8 kg N2O-N/kg N) for all years of the time-series.

6.5.3 Uncertainties and time-series consistency

Uncertainties in N2O emissions from agricultural soils are assessed for the 1990, 2001 and 2002 inventory using a Monte Carlo simulation of 5000 scenarios with the @RISK software (Kelliher et al., 2003) (Table 6.5.1). The emissions distributions are strongly skewed reflecting that of pastoral soil drainage whereby 74% of soils are classified as well-drained, while only 9% are classified as poorly drained. For the 2003 inventory, the uncertainty in the annual estimate was calculated using the 95% confidence interval determined from the Monte Carlo simulation as a percentage of the mean value i.e. in 2002, the uncertainty in annual emissions was +74% and -42%.

Table 6.5.1 Uncertainties in N2O emissions from agricultural soils for 1990, 2002 and 2003 estimated using Monte Carlo simulation (1990, 2002) and the 95% CI (2003)

Year N2O emissions from agricultural soils (Gg/annum) 95% CI Min 95% CI Max

1990

31.9

17.2

58.2

2002

40.6

23.4

70.4

2003

41.8

24.2

72.7

The overall inventory uncertainty analysis shown in Annex 7 (Good Practice Table 6.1) demonstrates that the uncertainty in annual emissions from agricultural soils is a major contributor to uncertainty in the total estimate and trend from 1990. Uncertainty in N2O emissions contributes 9.1% of the uncertainty in New Zealand's total emissions and removals in 2003 and 1.23% to the trend in emissions and removals from 1990-2003. The uncertainty between years is assumed to be correlated, therefore the uncertainty is mostly in the emission factors and the uncertainty in the trend is much lower than uncertainty for an annual estimate.

The Monte Carlo numerical assessment was also used to determine the effects of variability in the nine most influential parameters on uncertainty of the calculated N2O emissions in 1990 and 2001. These parameters are shown in Table 6.5.2 together with their percentage contributions to the uncertainty. There was no recalculation of the influence of parameters for the 2003 inventory. The Monte Carlo analysis confirmed that uncertainty in parameter EF3PRP has the most influence on total uncertainty accounting for 91% of the uncertainty in total N2O emissions in 1990. This broad uncertainty reflects natural variance in EF3 determined largely by the vagaries of the weather and soil type.

Table 6.5.2 Percentage contribution of the nine most influential parameters on the uncertainty of total N2O emissions inventories for 1990 and 2001

Parameter 1990 % contribution to uncertainty 2001 % contribution to uncertainty

EF3PRP

90.8

88.0

EF4

2.9

3.3

Sheep N Excretion

2.5

1.8

EF5

2.2

2.8

Dairy N excretion

0.5

0.7

FracGASM

0.5

0.5

EF1

0.3

2.4

Beef N excretion

0.2

0.3

FracLEACH

0.1

0.2

6.5.4 Source-specific QA/QC and verification

The nitrogen fertiliser data obtained from FertResearch are corroborated by the MAF using urea production figures and industrial applications (including resin manufacture for timber processing) data.

6.5.5 Source-specific recalculations

Emissions from the agricultural soils category were recalculated for all years because of changes in the allocation of dairy cattle excreta between lagoons and pasture (refer 6.3.5), changes in the Nex values from modifications to the model for calculating dry matter intake (refer 6.2.5) and including an allowance for N stored in wool fleece, a correction of sheep numbers for 1993 through 1996 (refer 6.2.5), a recalculation of fertiliser sales figures for 2001 and 2002 from revised information, and correction of an error in inputs from crop residue. Previous inventories showed 50% of the residue from pulses and soyabeans being ploughed in but agricultural practice in New Zealand is to plough all residues into the soil.

N2O emissions from horse excreta were not included in New Zealand's previous inventories. Emissions for horses were included in 2003 and back-calculated for all years.

6.5.6 Source-specific planned improvements

The work of NzOnet will continue in order to better quantify N2O emission factors for New Zealand's pastoral agriculture.

6.6 Prescribed burning of savanna (CRF 4E)

6.6.1 Description

Prescribed burning of savanna is not a key category for New Zealand. The New Zealand inventory includes burning of tussock (Chionochloa) grassland in the South Island for pasture renewal and weed control as savanna burning. The amount of burning has been steadily decreasing since 1959 as a result of changes in lease tenure and a reduction in grazing pressure. In 2003, total emissions accounted for 1.00Gg of CO2 equivalent - a 70.0% reduction from the 3.33Gg CO2 equivalent estimated in 1990.

The IPCC Guidelines (1996) state that in agricultural burning, the CO2 released is not considered to be a net emission as the biomass burned is generally replaced by regrowth over the subsequent year. Therefore the long term net emissions of CO2 are considered to be zero. However the by-products of incomplete combustion, CH4, CO, N2O and NOx, are net transfers from the biosphere to the atmosphere.

6.6.2 Methodological issues

New Zealand has adopted a modified version of the IPCC methodology (IPCC, 1996). The same five equations are used to calculate emissions. However, instead of using total grassland and a fraction burnt, New Zealand uses statistics of the total amount of tussock grassland that has been granted a consent (a legal right) under the Resource Management Act (1991) for burning. Only those areas with a consent are legally allowed to be burned. Expert opinion obtained from land managers is that approximately 20% of the area allowed to be burnt is actually burnt in a given year.

Current practice in New Zealand is to burn in damp spring conditions which reduces the amount of biomass consumed in the fire. The composition and burning ratios used in calculations are from New Zealand specific research (Payton and Pearce, 2001) and the IPCC reference manual (1996).

6.6.3 Uncertainties and time-series consistency

The same sources of data and emission factors are used for all years. This gives confidence in comparing emissions through the time-series from 1990 and 2003. The major sources of uncertainty are the percentage of consented area actually burnt in that season, that biomass data from two study sites are extrapolated for all areas of tussock, and that many of the other parameters (i.e. the carbon content of the live and dead components, the fraction of the live and dead material that oxidise and the N:C ratio for the tussocks) are the IPCC default values. Uncertainty in the New Zealand biomass data has been quantified at ±6% (Payton and Pearce, 2001). However, many IPCC parameters vary by ±50% and some parameters lack uncertainty estimates.

6.6.4 Source-specific QA/QC and verification

There was no source specific QA/QC for this category.

6.6.5 Source-specific recalculations

There were no recalculations for the 2003 inventory.

6.7 Field burning of agricultural residues (CRF 4F)

6.7.1 Description

Burning of agricultural residues produced 30.53Gg CO2 equivalent in 2003. Emissions are currently 21.0% over the level in 1990. Burning of agricultural residues is not identified as a key source for New Zealand.

New Zealand reports emissions from burning barley, wheat and oats residue in this category. Maize residue is not burnt in New Zealand. New Zealand uses three-year averages of crop production in combination with the IPCC default emission ratios and residue statistics. Oats are included under the same emission factors as barley.

Burning of crop residues is not considered to be a net source of CO2 because the CO2 released into the atmosphere is reabsorbed during the next growing season. However, the burning is a source of emissions of CH4, CO, N2O and NOx (IPCC, 1996).

6.7.2 Methodological issues

The emissions from burning of agricultural residues are estimated in accordance with the IPCC guidelines (IPCC, 1996). The calculation uses crop production statistics, the ratio of residue to crop product, the dry matter content of the residue, the fraction of residue actually burned, the fraction of carbon oxidised and the carbon fraction of the residue. These figures are multiplied to calculate the carbon released. The emissions of CH4, CO, N2O and NOx are calculated using the carbon released and an emissions ratio. N2O and NOx emissions calculations also use the nitrogen to carbon ratio.

Good practice suggests that an estimate of 10% of residue burnt may be appropriate for developed countries but also notes that the IPCC defaults "are very speculative and should be used with caution. The actual percentage burned varies substantially by country and crop type. This is an area where locally developed, country-specific data are highly desirable" (IPCC, 2001). New Zealand estimates that common agricultural practice is that 50% of stubble is burnt. This figure is developed from expert opinion of MAF officials working with the arable production sector.

6.7.3 Uncertainties and time-series consistency

No numerical estimates for uncertainty are available for these emissions. The fraction of agricultural residue burned in the field is considered to make the largest contribution to uncertainty in the estimated emissions.

6.7.4 Source-specific QA/QC and verification

There was no source specific QA/QC for this category.

6.7.5 Source-specific recalculations

There are no recalculations for this category.