View all publications

Appendix A: Methodology report for agriculture projections as provided by the Ministry of Agriculture and Forestry

Projected emissions from the agriculture sector over the first commitment period (2008-2012)

Table A1: Summary of emission projection scenarios

  

1990

 

2010

 

Baseline

Lower scenario

Most likely

Upper scenario

 

*Mt CO2-e

Mt CO2-e

Mt CO2-e

Mt CO2-e

Total projected emissions

32.117

36.059

39.756

44.447

Projected emissions above 1990

 

3.942

7.639

12.331

Percent change from 1990

 

12.3

23.8

38.4

Methane total

22.159

23.694

25.971

28.798

Nitrous Oxide total

9.958

12.365

13.785

15.649

Figures based on 2004 national greenhouse gas inventory methodology

*Million tonnes carbon dioxide equivalents

a. Summary

Emissions from the agriculture sector are projected to range from 180.29 Mt CO2-e to 222.24 Mt CO2-e over the first commitment period (2008-2012) with a most likely value of 198.78 Mt CO2-e. Average annual emissions over the first commitment period are projected to range from 36.06 Mt CO2-e to 44.45 Mt CO2-e with a most likely value of 39.76 Mt CO2-e.

The most likely value for annual emissions at 2010 is 0.64 Mt CO2-e lower than projected in 2005. This small decrease is attributed to a decrease in the emission factor for nitrogen fertiliser. There has also been a significant increase in the range between the upper and lower projection scenarios. The difference is now 8.4 Mt CO2-e compared to 3.5 Mt CO2-e in the 2005 estimate. This is due to the use of a new methodology for calculating the uncertainty in projected animal numbers and the use of a new econometric model to project future nitrogen fertiliser use.

b. Introduction

These projections are based on the methodology used in the New Zealand national greenhouse gas inventory submitted to the United Nations Framework Convention on Climate Change (UNFCCC) annually and on econometric models developed by the Ministry of Agriculture and Forestry (MAF). The inventory methodology conforms to the Good Practice Guidance methodologies developed by the Intergovernmental Panel on Climate Change and adopted by the UNFCCC.

Projections are driven by future estimates of:

  • animal numbers by species: dairy cattle, beef cattle, sheep and deer in 2010 using the MAF Pastoral Supply Response Model (PSRM)
  • enteric methane emissions per animal based on changes in past emissions per animal between 1990 and 2004
  • nitrogen output per animal based on changes in past nitrogen output per animal between 1990 and 2004
  • nitrogen fertiliser use based on an econometric model projecting future use using projected animal numbers.

Two further scenarios of projected emissions in 2010 have also been produced. These represent emission estimates using the upper and lower bounds of methane and nitrous oxide emission factors and animal numbers.

c. Changes in methodology since last year's assessment

Changes in the methodology have been instigated to take into account new approaches and information obtained since the last update in May 2005.

Changes were implemented in four areas:

  • no post-model adjustments of land area change were incorporated into the projected animal numbers that were derived from the PSRM model
  • a new methodology was implemented for upper and lower scenario values for animal numbers
  • the projections of nitrogen fertiliser usage in 2010 was based on a new methodology developed by MAF
  • updating of the national greenhouse gas inventory methodology in keeping with UNFCCC requirements for maintaining "good practice".

The national greenhouse gas inventory implemented two changes in the agricultural sector inventory in 2006, reporting on the 2004 emission levels. These were:

  • decreasing the emission factor for nitrous oxide of nitrogen fertiliser applied to agricultural soils from 1.25 percent to 1 percent, and
  • implementing a new "Tier 2" system for accounting for methane emissions from waste produced by ruminant animals.

The above changes have resulted in an increase in the difference between the upper and lower scenarios.

d. Development of the most likely scenario

i. Projection of most likely animal population in 2010

The PSRM is used to project animal numbers. The PSRM is currently used for livestock projections that contribute to the Treasury's economic and fiscal updates twice yearly. It is a system of equations that capture the biological and market behaviours of the New Zealand pastoral sectors (dairy, beef, sheep, and deer). Changes in livestock inventory are derived from exogenous shocks to the model that includes real farm-gate prices and weather.

Baseline scenario

The MAF livestock forecasts provide farm-gate price forecasts out to 2014. These forecasts take into account the macro-economic assumptions (eg, economic growth, exchange rate, inflation) provided by Treasury's Half Yearly Economic and Fiscal Update in December 2005 and the international supply and demand factors which subsequently affect international commodity prices.

The weather variable in the PSRM uses the Daily Soil Moisture Deficit (DSMD) series supplied by National Institute of Water and Atmospheric Research. For the forecast period (2006-2020), DSMD uses the average value over the 1973-2005 period.

The price assumptions are exogenous in the PSRM. Information from experts in the forestry sector suggests that land conversion only marginally impacts the pastoral sectors. Therefore, any land use changes in the forecast period are assumed to be captured by the residuals (errors) of the forecasts.

Table A2: Projected most likely animal numbers in 2010

 

1990

baseline

(000)

2010 most likely

scenario

(000)

Dairy cattle

3,391

5,526

Beef cattle

4,597

4,076

Sheep

57,861

39,736

Deer

1,036

1,448

ii. Projections of enteric methane emissions

Projections of methane emissions per animal in 2010 are derived from linear trends of the methane emissions per animal using data in the national greenhouse gas inventory from per animal emissions from 1990 to 2004 extrapolated to 2010.

These per-animal emissions have been derived from the model developed by Clark et al (2003) that is used to estimate methane emissions in the national inventory (Figure A1).

Figure A1: Model for deriving methane emissions (Clark et al, 2003)

Thumbnail of image. See figure at its full size (including text description).

The model determines animal feed intakes in monthly time steps for age classes of each animal species based on the mean national animal performance derived from national statistics relevant to each species. For example, dairy cattle inputs include: animal liveweight, total milk production and milk fat and protein percentages. For each animal species, an empirical relationship has been derived for the amount of enteric methane produced per unit of feed intake. These relationships have been developed in New Zealand for: deer, beef and dairy cattle, and sheep guided by the SF6 technique that enables assessment of methane emissions in the field. The estimated annual methane emissions per animal take into account changes in animal performance over time.

The estimated methane emission factors for each of the species in 1990, 2010 and the correlations of the trend over time from 1990 to 2004 used to derive the projected value are presented below.

Table A3: Projected most likely methane emission per animal in 2010

  1990 (kg/methane/head/annum) 2010 Projected (kg/methane/head/annum)

Correlation (r) 1990 to 2004

Dairy cattle

70.8

83.4

0.963

Beef cattle

51.0

58.9

0.929

Sheep

8.9

11.6

0.996

Deer

21.0

23.4

0.864

iii. Methane from ruminant animal waste

Methane emissions also arise from faecal material deposited on pasture and, in the case of lactating dairy cows, from animal waste management systems (effluent ponds).

Table A4: Projected most likely methane emissions from animal waste in 2010

  1990 (kg/methane/head/annum) 2010 Projected (kg/methane/head/annum)

Correlation 1990 to 2004

Dairy cattle

2.93

3.47

0.977

Beef cattle

0.65

0.75

0.925

Sheep

0.09

0.12

0.997

Deer

0.20

0.22

0.865

iv. Projections of nitrous oxide emissions

Nitrous oxide emissions are derived from animal nitrogen output. This is a function of animal feed intake and nitrogen content of the pasture eaten minus any nitrogen stored in animal product. Models developed by Clark et al (2003) for methane emissions also estimate nitrogen output per animal. Projections of nitrogen output per animal in 2010 are derived from linear trends of nitrogen outputs per animal using data in the national inventory for the period 1990 to 2004. Nitrous oxide emissions are then calculated using the methodology used for the national inventory.

Table A5: Projections of most likely nitrogen output per animal in 2010

 

1990 (kg/N/head/annum)

2010 Projected (kg/N/head/annum)

Correlation 1990 - 2004

Dairy cattle

106.2

122.1

0.955

Beef cattle

65.2

76.1

0.927

Sheep

12.2

15.9

0.995

Deer

27.4

30.6

0.871

v. Projections of nitrogen fertiliser use

A new method of projecting nitrogen fertiliser use was developed for this round of projections. In order to select the best-practice model for forecasting nitrogen fertiliser use, the consistency between the forecast for nitrogen fertiliser and livestock number forecasts was considered. In the empirical nitrogen model, the nitrogen use series was inserted into the PSRM model and livestock forecasts used to derive the nitrogen use forecast.

Empirical model

After investigating the impact of many different variables on nitrogen usage, the main drivers of nitrogen usage were found to be the number of dairy animals. Information from industry supports this conclusion. In 2005 average nitrogen fertilizer application on dairy farms was 140kg/ha whereas on sheep and beef farms, an average of 12kg/ha was used. Also, from the 2004 Agricultural Production Survey, 66 percent of urea was applied to farms defined as solely dairy.

The projected most likely value for nitrogen usage in 2010 was 403,709 tonnes.

vi. Other animal species and greenhouse gas sources

No projections were derived for the emissions of minor animal species present in the national inventory, ie, horses, goats, pigs, and poultry. This was also the case for nitrous oxide emissions from crop stubble burning, savannah burning and nitrogen fixing crops. These emission sources make up less that 4 percent of the agricultural sector emissions. There was no basis to assume that any of these emission sources would be significantly different from the present levels. The impact of even large changes in any of these small emission sources on total national emissions would be small and so 2004 inventory emission levels were used for 2010 projections.

e. Development of lower and upper scenarios

Two further scenarios were developed: an upper and lower scenario. The upper scenario combined the upper 95 percent confidence interval values for animal numbers, methane emission per head, nitrogen output per head and nitrogen fertiliser use. The lower scenario combined the lower 95 percent confidence interval values for animal numbers, methane emission per head, nitrogen output per head and nitrogen fertiliser use. These two scenarios give an estimate of the upper and lower bounds of future projected emissions.

i. Animal numbers

To derive livestock forecasts for different scenarios, exogenous price uncertainty is introduced into the PRSM model. Price uncertainty is introduced into the simulation through specifying the possible movements in commodity prices for the forecast period. Variation of prices (or standard deviation of the 95 percent confidence interval) during the last ten-year period for each price series is used. This gave estimations for the upper and lower bounds of the stochastic forecasts that could be considered as upper and lower scenarios due to the movement in prices.

Table A6: Lower and upper scenarios of animal number projections in 2010

 

Lower 2010

(000)

Most Likely 2010

(000)

Upper 2010

(000)

Dairy cattle

5,281

5,526

5,813

Beef cattle

3,502

4,076

4,761

Sheep

37,349

39,736

43,106

Deer

1,095

1,448

1,871

ii. Methane emissions

Upper and lower estimates of methane emissions per animal were obtained from the 95 percent confidence interval for the linear regression of emissions from 1990 to 2004. This gave an upper and lower bound for projected methane emissions per head in 2010.

Table A7: Lower and upper scenarios of per-animal methane projections to 2010

 

Lower 2010

(kgCH4/head/annum)

Most Likely 2010

(kgCH4/head/annum)

Upper 2010

(kgCH4/head/annum)

Dairy cattle

82.0

83.4

84.8

Beef cattle

57.7

58.9

60.0

Sheep

11.5

11.6

11.7

Deer

22.7

23.4

24.2

iii. Nitrogen output

Upper and lower estimates of nitrogen output per animal were obtained from the 95 percent confidence interval for the linear regression of emissions from 1990 to 2004. This provided an upper and lower bound for projected methane emission per head in 2010.

Table A8: Lower and upper scenarios of per-animal nitrogen output projections to 2010

 

Lower 2010

(kgN/head/annum)

Most Likely 2010

(kgN/head/annum)

Upper 2010

(kgN/head/annum)

Dairy cattle

120.1

122.1

124.0

Beef cattle

74.6

76.1

77.6

Sheep

15.8

15.9

16.1

Deer

29.5

30.6

31.5

iv. Nitrogen Fertiliser

Upper and Lower scenarios for future nitrogen fertiliser use were obtained from the upper and lower projections of dairy animal numbers derived within the PSRM model.

Table A9: Lower and upper bounds of projected nitrogen fertiliser use in 2010

Lower 2010

Tonnes N per annum

Most likely 2010

Tonnes N per annum

Upper 2010

Tonnes N per annum

299,659

403,709

541,707

f. Overall assumptions and limitations of projections

The projections need to be assessed within the uncertainties of biological systems and economic circumstances of the agricultural industry, which is largely driven by overseas markets. For example, an assumption implicit in the projections is that the rate of increase in productivity per animal over the next 6 years will not be dissimilar to the rate of increase in animal performance over the past 14 years, and therefore a linear extrapolation of methane emissions per animal is appropriate. It is possible that the rate of increase in animal performance may decline over time and to test this other non-linear relationships were looked at; however, no improvement in relationship was gained.

Mitigation technologies that reduce emissions at an individual animal level may emerge over the next six years. These include products such as Monensin, a bloat control agent that has been show to reduce methane emissions, or the widespread adoption of the nitrification inhibitor, DCD that has also been shown to reduce nitrous oxide emissions. None of these mitigation technologies has been factored into the projections as it is believed that they may be counter balanced by improvement in animal productivity growth. Industry strategy plans, particularly the dairy industry, are seeking productivity that is, however, higher than that achieved over the last 14 years ie, since 1990. There is also considerable uncertainty around the level of adoption of potential mitigation technologies.

In terms of nitrogen fertiliser usage, future changes such as limitations on nitrogen use in some catchments, ie, Lake Taupo and Lake Rotorua, the continuing conversion of pastoral land to forestry, the Dairying Clean Streams Accord, Regional Council initiatives, industry policies on good fertiliser practice, are likely to limit the steep upward trend in fertiliser nitrogen use apparent in recent years.

On the other hand, a pest found in New Zealand in 1996 has been reducing the nitrogen fixation in clover, New Zealand's main source of nitrogen for pasture. The response of affected farmers has been to increase the use of nitrogen fertiliser and increase feed supplements. An economic impact assessment (NZIER, 2005) determined that under a medium-impact scenario, the clover root weevil increased greenhouse gas emissions by 0.92 Mt CO2-e during the first commitment period. With the revised emission factor for nitrogen fertiliser, clover root weevil is expected to increase total emissions from nitrogen fertiliser by 0.74 Mt CO2-e.

g. References

Clark, H.; Brookes, I. and Walcroft, A. 2003, Enteric methane emissions for New Zealand Ruminants 1990-2001 calculated using an IPCC Tier Two approach. Report prepared for the Ministry of Agriculture and Forestry by AgResearch Ltd.

Ministry of Agriculture and Forestry, 2006, Briefing on Methodology for Forecasting Livestock Numbers and Nitrogen Fertiliser Use. 8pp.

NZIER, 2005, Clover Root Weevil: Economic Impact Assessment. For Biosecurity New Zealand, Ministry of Agriculture and Forestry.