View all publications

Chapter 7: Land use, land-use change and forestry (LULUCF)

7.1 Sector overview

In 2008, net removals by the land use, land-use change and forestry (LULUCF) sector were 26,176.8 Gg carbon dioxide equivalent (CO2-e). This is made up of net removals of 26,219.1 Gg carbon dioxide and emissions of 38.2 CO2-e of methane (CH4), and 4.1 Gg CO2-e of nitrous oxide (N2O).

Net removals have decreased by 4,889.5 Gg CO2-e (15.7 per cent) from the 1990 level of 31,066.3 Gg CO2-e (Figure 7.1.1). This is largely due to the harvesting and replanting of plantation forests in the five years prior to 2008 as this lowered the average age and therefore the CO2 absorption capacity of planted forests in 2008. The decrease is also due to direct emissions from deforestation. Figure 7.1.2 shows the changes in emissions and removals by land-use category from 1990 to 2008. The increase in emissions in the grassland land-use category is primarily the result of the increased deforestation and conversion to grassland of plantation forests that occurred in the five years prior to 2008, as emissions from land-use change are reported in the ‘land converted to’ category.

Carbon dioxide emissions and removals in the LULUCF sector are primarily controlled by the uptake from plant growth, emissions from harvesting production forests, deforestation and the decomposition of organic material. Nitrous oxide can be emitted from the ecosystem as a by-product of nitrification and de‑nitrification, and the burning of organic matter. Other gases released during biomass burning include CH4, carbon monoxide (CO), other oxides of nitrogen (NOx) and non-methane volatile organic compounds (NMVOCs).

All emissions and removals from the LULUCF sector are excluded from national totals unless otherwise specified. This is consistent with the Climate Change Convention reporting guidelines.

New Zealand has adopted the six broad categories of land use as described in Good Practice Guidance for Land Use, Land-Use Change and Forestry (IPCC, 2003), hereafter referred to as GPG-LULUCF.

The land-use categories of forest land remaining forest land, conversion to forest land, grassland remaining grassland and conversion to grassland are key categories for New Zealand in 2008.

Figure 7.1.1 New Zealand’s annual emissions and removals from the LULUCF sector from 1990 to 2008


Figure 7.1.1 New Zealand’s annual emissions and removals from the LULUCF sector from 1990 to 2008

Year Net removals
Gg CO2-equivalent
1990 -31,066.3
1991 -31,534.6
1992 -30,544.9
1993 -30,652.6
1994 -29,891.8
1995 -28,768.1
1996 -28,708.8
1997 -29,871.6
1998 -31,793.1
1999 -32,450.1
2000 -31,281.2
2001 -30,206.9
2002 -27,591.0
2003 -28,423.1
2004 -27,899.6
2005 -23,792.7
2006 -19,714.1
2007 -16,820.7
2008 -26,176.8

Note: At this scale, emissions on cropland, wetland, settlements and other land are inseparable.

Figure 7.1.2 Change in New Zealand’s emissions and removals from the LULUCF sector from 1990 to 2008

Figure 7.1.2 Change in New Zealand’s emissions and removals from the LULUCF sector from 1990 to 2008

Land-use category Net emissions/removals (Gg CO2-equivalent )
  1990 2008 % change in emissions since 1990
Forest land -32,856.7 -29,757.9 9.4
Cropland 29.9 -23.7 -179.0
Grassland 1,742.4 3,557.0 104.1
Wetlands 0.0 0.8 NA
Settlements 6.6 20.0 204.4
Other land 11.5 26.9 133.0

7.1.1 Land use in New Zealand

Before human settlement, natural forests were New Zealand’s predominant land cover, estimated at 75 per cent of total land area. Today, natural forest covers around 30 per cent of the total land area of New Zealand (see Table 7.1.1.1). Nearly all lowland areas have been cleared of natural forest for agriculture, horticulture, plantation forestry and urban development.

Table 7.1.1.1 Land use in New Zealand in 2008
Land-use category Subcategory Net area in 2008 (ha) Proportion of total area (%)
Forest land Natural forest 8,118,004 30.2
  Pre-1990 planted forest 1,430,286 5.3
  Post-1989 forest 568,775 2.2
  Subtotal 10,117,064 37.7
Cropland Annual 334,159 1.2
  Perennial 88,541 0.3
  Subtotal 422,700 1.6
Grassland High producing 5,813,712 21.6
  Low producing 7,701,148 28.7
  With woody biomass 1,056,975 3.9
  Subtotal 14,571,835 54.2
Wetlands   644,135 2.4
Settlements   206,288 0.8
Other land   889,068 3.3
Total   26,851,090 100.0

Note: Areas as at 31 December 2008 and include deforestation of post-1989 forest since 1990.

Pasture establishment is thought to have slightly increased mineral soil carbon levels. However, losses of carbon due to erosion as a result of land-use change are also possible (Tate et al, 2003c). New Zealand soils are naturally acidic with low levels of nitrogen, phosphorus and sulphur. Consequently, soils used to grow crops and pasture need to be developed and maintained with nitrogen-fixing plants (such as clover), fertilisers and, often, lime to sustain high-yield plant growth.

New Zealand has a substantial estate of planted forests that are intensively managed for timber-supply purposes. In 2008, plantation forests covered approximately 2.0 million hectares, around 7.5 per cent of New Zealand’s total land area. This includes areas not managed for timber supply, for instance, areas planted for erosion control.

The description of the land-use subcategories mapped by New Zealand is provided in Table 7.1.2.6. The national threshold that New Zealand uses to define forest land for both Climate Change Convention and Kyoto Protocol reporting are: a minimum area of 1 hectare, a crown cover of 30 per cent and a minimum height of 5 metres (Ministry for the Environment, 2006).

There continues to be considerable land-use change in New Zealand, and while the harvesting of natural forests has been restricted under the 1993 amendments to the Forests Act 1949 and the Forests (West Coast Accord) Act 2000, there are very few other regulations in place to influence other land-use change. Recently, the following regulations and government initiatives have been put in place to either encourage forest establishment or discourage deforestation:

  • Climate Change Response Act 2002 (updated 8 December 2009)
  • Permanent Forest Sink Initiative (Ministry of Agriculture and Forestry, 2008b)
  • Afforestation Grant Scheme (Ministry of Agriculture and Forestry, 2009b).

7.1.2 Methodological issues for New Zealand

Recalculation of the 1990–2007 LULUCF inventory

For this submission, New Zealand has recalculated its emission and removal estimates for the LULUCF sector from 1990 to 2007 to incorporate improved New Zealand-specific methods and data. This follows the introduction in 2008 of a new data collection and modelling programme for the LULUCF sector, the Land Use and Carbon Analysis System (LUCAS) (see below and Annex A3.2 for further details).

With the change to using LUCAS information, there have been major recalculations to New Zealand’s LULUCF emission and removal estimates. These have resulted in a decrease of 12,927.8 Gg CO2-e in net emissions (an increase in removals) on the 1990 estimate, and an increase of 7,015.2 Gg CO2-e in net emissions (a decrease in removals) on the 2007 estimate (see Table 7.1.2.1).

Table 7.1.2.1 Recalculations to New Zealand’s total net removals
  Reported net removals Change in estimate
2009 submission (Gg CO2-e) 2010 submission
(Gg CO2-e)
(Gg CO2-e) (%)
1990 –18,138.5 –31,066.3 –12,927.8 –71.3
2007 –23,836.0 –16,820.7 +7,015.3 +29.4

The LUCAS system has enabled New Zealand to make significant improvements to LULUCF estimates, including new mapping of land use and land-use change since 1990, and the use of improved New Zealand-specific methods, activity data and emission factors. This has improved the accuracy, completeness and transparency of the estimates.

The main differences between this submission and previous estimates of New Zealand’s LULUCF emissions and removals are the result of:

  • new mapping of 1990 and 2007 land use and land-use change to improve the identification of ‘land-use remaining’ and ‘land-use change’ areas in the six land-use categories, and, in particular, of changes to and from forest land, resulting in the new mapping of 275.6 kilo hectares of previously unidentified forest land in 1990 (revising the total area of 1990 forest land from 9,368.9 kha to 9,644.6 kha in the current submission)
  • changes in the land-use subcategories New Zealand is reporting on, to improve the alignment between New Zealand’s forest land, grassland and wetlands categories, the IPCC land-use categories and the Kyoto Protocol forest definition. Previously, New Zealand had reported land-cover categories (Table 7.1.2.7)
  • improved measurement of deforestation up to 1 January 2008 has been based on land-use change mapping rather than relying as previously on a range of information sources. Deforestation is also now reported on in the ‘land converted to’ category, whereas previously it had been included in the forest land remaining forest land subcategory
  • methodological improvements to the calculation of emissions and removals by using the LUCAS ‘calculation engine’ to consistently estimate annual changes in carbon and non-carbon emissions, for the five carbon pools associated with annual land-use changes, using a master set of New Zealand-specific and IPCC default emission factors (see Tables 7.1.2.3, 7.1.2.4, 7.1.2.5 and 7.1.2.10)
  • improvements to the accuracy of emission factors, in particular, the age-based carbon yield table for post-1989 forests and the national average carbon levels for natural forests
  • the inclusion of soil carbon stock estimates using a Tier 2 method with New Zealand-specific georeferenced soil pedon data.

The impact of these recalculations on net CO2-e removals in each land-use category is provided in Table 7.1.2.2. The differences shown are a result of recalculations for all carbon pools used for Climate Change Convention and Kyoto Protocol reporting for the whole time series for the LULUCF sector. This table only includes recalculations from 1990 to 2007, to enable a comparison of the two approaches.

Table 7.1.2.2 Recalculations to New Zealand’s net emissions and removals for 1990 and 2007
Land-use category Net emissions and removals (Gg CO2-e) Change in 1990 estimate (%) Change in 2007 estimate (%)
2009 submission:
1990 estimate
2010 submission:
1990 estimate
2009 submission:
2007 estimate
2010 submission:
2007 estimate
Forest land –18,649.2 –32,856.7 –24,527.9 –30,651.5 +76.2 +25.0
Cropland –477.7 29.9 –510.3 17.5 –106.3 –103.4
Grassland 863.9 1,742.4 1,063.7 13,618.3 +101.7 +1180.3
Wetlands 0.7 0.0 0.7 4.2 –98.5 +482.7
Settlements 97.2 6.6 97.2 102.7 –93.2 +5.7
Other land 26.7 11.5 40.6 88.0 –56.8 +116.7
Total –18,138.5 –31,066.3 –23,836.0 –16,820.7 +71.3 –29.4

Detailed information on the recalculations is provided below in the relevant source-specific recalculations sections, and in chapter 10.

Methodological approaches to calculating emissions and removals

New Zealand uses a combination of Tier 1 and Tier 2 methodologies for estimating and reporting emissions and removals for the LULUCF sector. The Tier 1 approach, based on a simple land-use change matrix, has been used to estimate carbon for the four biomass pools for all land-use categories except for natural forest, pre-1990 planted forest and post-1989 forests as these all use a Tier 2 approach. A Tier 2 modelling approach has also been used to estimate carbon in the mineral soil component of the soil organic matter pool for all land-use categories, except for other land. The other land category uses a Tier 1 approach. Carbon in the organic soil component is not estimated independently from mineral soils, as this makes up just 0.9 per cent of New Zealand’s total land area.

New Zealand is estimating carbon stock change for each of the five Kyoto Protocol carbon pools and aggregating the results to the three Climate Change Convention reporting pools. Table 7.1.2.3 summarises the methods being used to estimate carbon by pool for each land use.

Table 7.1.2.3 Relationships between carbon pool, land-use category, LULUCF activity and model calculations used by New Zealand
Climate Change Convention reporting pool
Living biomass
Dead organic matter
Soils
Kyoto Protocol reporting pool Above-ground biomass Below-ground biomass Dead wood Litter Soil organic matter
Land-use category Natural forest Allometric equations % of above-ground biomass Allometric equations Lab analysis Soil carbon model
Natural forest [D] Look-up table based on natural forest national average tonnes C ha–1  
Pre-1990 planted forest A NEFD-based yield table and the C_Change model Soil carbon model
Pre-1990 planted forest [D] Look-up table based on Forest Carbon Predictor model (table is split by tree age, stocking and site index)
Post-1989 forest [AR] Forest Carbon Predictor model Per cent of above-ground biomass Forest Carbon Predictor model Soil carbon model
Post-1989 forest harvesting Forest Carbon Predictor model, with emissions parameter Per cent of above-ground biomass Forest Carbon Predictor model, with emissions parameter  
Post-1989 forest [D] Look-up table based on Forest Carbon Predictor model (table is split by tree age, stocking and site index)  
Cropland IPCC Tier 1 default parameters Not estimated Not estimated Not estimated Soil carbon model
Grassland (high and low producing) IPCC Tier 1 default parameters IPCC Tier 1 default parameters Not estimated Not estimated Soil carbon model
Grassland with woody biomass Allometric equations Per cent of above-ground biomass Allometric equations Allometric equations
Wetlands IPCC Tier 1 default parameters IPCC Tier 1 default parameters Not estimated Not estimated Soil carbon model
Settlements IPCC Tier 1 default parameters IPCC Tier 1 default parameters Not estimated Not estimated Soil carbon model
Other land IPCC Tier 1 default parameters IPCC Tier 1 default parameters Not estimated Not estimated IPCC Tier 1 default parameters

Notes: AR = afforestation/reforestation, D = deforestation and NEFD = the National Exotic Forest Description (Ministry of Agriculture and Forestry, 2009a). See the methodology sections on soils (section 7.1.2) and forests (section 7.2.2) for explanations of the soil carbon, C_Change and Forest Carbon Predictor models.

LUCAS Data Management System

New Zealand has established a data collection and modelling programme for the LULUCF sector called the Land Use and Carbon Analysis System (LUCAS) (see www.mfe.govt.nz/issues/climate/lucas/). This programme addresses the lack of information for some land-use categories, and includes the use of field plot measurements for natural and planted forests and airborne scanning LiDAR (Light Detection and Ranging) for planted forests (Stephens et al, 2007, 2008); use of allometric equations and models to estimate carbon stock and carbon-stock change in natural and planted forests respectively (Beets et al, 2009; Kimberley and Beets, 2008); wall-to-wall land-use mapping for 1990 and 2008 using satellite and aircraft remotely sensed imagery; a New Zealand-specific soil carbon model to estimate changes in soil organic matter with changes in land use; and development of databases and applications to store and manipulate all data associated with LULUCF activities.

Details of the natural forest allometric equations, and the planted forest growth models and carbon allocation models, are provided in the forest land section (section 7.2). This section provides details about the database, including where the data is stored, the calculations performed, the soil carbon model, liming and biomass burning.

The LUCAS Data Management System stores, manages and retrieves data for international greenhouse gas reporting for the LULUCF sector. The system comprises three primary applications: the Geospatial System, the Gateway and the Calculation and Reporting Application (Figure 7.1.2.1). These systems are used for managing the land-use spatial databases and the plot and reference data, and for combining the two sets of data to calculate the numbers required for Climate Change Convention and Kyoto Protocol reporting. Details on these databases and applications are provided in Annex 3.2.

The LUCAS Data Management System:

  • provides a transparent system for the storage and management of LULUCF activity data
  • provides a transparent means for the versioning and validation of land-use data, plot measurements, reference data and emissions factors
  • calculates carbon stocks, emissions and removals by land use and carbon pool for both Climate Change Convention and Kyoto Protocol reporting
  • calculates biomass burning and liming emissions by land use based on spatial and emission factors stored in the Gateway
  • produces the output required to populate the common reporting format tables for the LULUCF sector and reporting under Article 3.3 of the Kyoto Protocol.

Figure 7.1.2.1 New Zealand’s LUCAS Data Management System


Figure 7.1.2.1 New Zealand’s LUCAS Data Management System

Figure 7.1.2.1 describes the LUCAS Data Management System, which stores, manages and retrieves data for New Zealand’s international greenhouse gas reporting for the LULUCF sector.

The system comprises three primary applications: the Geospatial System, the Gateway and the Calculation and Reporting Application (CRA), supported by document management and server and hosting services.

The Geospatial System comprises: image quality assurance and quality control (QA/QC), an image server, a land-use mapping (LUM) staging database, LUM QA/QC, a LUM production database, and geospatial analysis and reporting.

The LUCAS gateway comprises a data layer, a validation layer and an extraction layer.

The CRA comprises the natural forest, soils, planted forest and non-carbon datasets, which feed into the ‘joint calculations’ emissions analysis, which provide the inputs for LULUCF emissions and removals reporting.

A range of different datasets feed into all of these applications. Images and land-use maps feed into the Geospatial System; emission factor datasets feed into the LUCAS Gateway, including forest plot data, reference data, soils data, IPCC defaults, parameters and non-carbon data; and method and reference documents are stored in the document management system.

These systems are used for managing the land-use spatial databases and the plot and reference data, and for combining the two sets of data to calculate the numbers required for Climate Change Convention and Kyoto Protocol reporting. Details on these databases and applications are provided in Annex 3.2.

Note: LUM means land-use map, and joint calculations are described below.

‘Joint calculations’ refers to the process New Zealand uses to estimate national average carbon values by carbon pool for each land-use category and subcategory. The joint calculation process is performed within the Calculation and Reporting Application (CRA). For further details refer to Annex 3.2.

Calculation of national emission and removal estimates

To calculate emissions and removals for the New Zealand LULUCF sector, the following data are used:

  • annual land use and land-use change area data
  • biomass carbon stocks per hectare prior to land-use conversion, and annual growth in biomass carbon stocks per hectare following conversion (see Tables 7.1.2.4 and 7.1.2.5)
  • age-based carbon yield tables for pre-1990 planted forests and post-1989 forests
  • emission factors and country-level activity data on biomass burning and liming
  • IPCC default conversion factors.

The formula used to calculate emissions from biomass changes is:

(Loss of biomass present in previous crop) X Activity data (Area)) + (Annual growth in biomass carbon stocks X Activity data (Area))

The formula used to calculate emissions from soil changes is:

Soil carbon at steady state in the original land use - Soil carbon at steady state in the converted to land use/ 20 years (tranisition period) X Activity data (Area)

For example, the annual change in carbon stock from the conversion of 100 hectares of low-producing grassland to perennial cropland would be calculated as follows:

Biomass change = (–3.05 x 100) + (2.25 x 100) = –80 t C

Soil change = (((117.66 – 114.91) / 20) x 100) = 13.75 t C

Total carbon change = –66.25 t C

Total emissions = (carbon stock change / 1000 x –1) x (44/12)

Total emissions = (–66.25 / 1000 x –1) x (44/12) = 0.2429 Gg CO2-e

These calculations are performed to produce estimates of annual carbon stock and carbon stock changes since 1990 to inform the Climate Change Convention and Kyoto Protocol Article 3.3 reporting.

Emission factors

The emission factors required to estimate carbon stock changes using the Tier 1 and Tier 2 equations are provided in Tables 7.1.2.4 and 7.1.2.5. These are split into biomass carbon stocks by land use prior to conversion and annual growth in stocks after land-use change. The values used for the previous submission are also provided for comparison, to illustrate the recalculations that have been made to the biomass emission factors used by New Zealand since the previous submission.

Table 7.1.2.4 New Zealand’s biomass carbon stock emission factors in land use before conversion
Land-use category Land-use subcategory 2009 submission emission factors
(t C ha–1)
2010 submission emission factors
(t C ha–1)
Carbon pools Source/reference
Forest land Natural forest 182 173 Live tree biomass only (2007)
All biomass pools (2008)
Hall et al, 2001 (for the 2007 value)
Beets et al, 2009 (for the 2008 value)
Pre-1990 planted forest 222 Based on an age-based carbon yield table All biomass pools Wakelin, 2008
Post-1989 forest NE Based on an age-based carbon yield table All biomass pools Kimberley et al, 2009
Cropland Annual 0 5 Above- and below-ground biomass Table 3.3.8, GPG-LULUCF IPCC, 2003
Perennial 63 63 Above-ground biomass Table 3.3.2, GPG-LULUCF IPCC, 2003
Grassland High producing 1.35 6.75 Above-ground biomass (2007)
Above- and below-ground biomass (2008)
Table 3.4.9, GPG-LULUCF, IPCC, 2003
Low producing 0.8 3.05 Above-ground biomass (2007)
Above- and below-ground biomass (2008)
Table 3.4.9, GPG-LULUCF, IPCC, 2003
With woody biomass 29 29 All biomass pools Wakelin, 2004
Wetlands   NE NE NA Section 3.5.2.2 and Annex 3A, GPG-LULUCF, IPCC, 2003
Settlements   NE NE NA Section 3.6.2, GPG-LULUCF, IPCC, 2003
Other land   NE NE NA Section 3.7.2.1, GPG-LULUCF, IPCC, 2003

Notes: NE is not estimated and NA is not applicable. All biomass pools include above- and below-ground biomass, litter and dead organic matter pools. See below in section 7.1.2 and under Methodological issues in each category-specific section for further details.

Table 7.1.2.5 New Zealand’s emission factors for annual growth in biomass for land converted to another use
Land-use category Land-use subcategory 2009 submission emission factor
(t C ha–1)
2010 submission emission factor
(t C ha–1)
Years to reach steady state Carbon pools Source/ reference
Forest land Natural forest 0.24 NA NA All biomass pools (2007) Annex 3A, IPCC, 2003
Pre-1990 planted forest 8.9 Based on age-based carbon yield table 28 Above-ground biomass (2007)
All biomass pools (2008)
Annex 3A, IPCC, 2003
Post-1989 forest NE Based on age-based carbon yield table 28 All biomass pools (2008) Kimberley et al, 2009
Cropland Annual 5 5 1 Above- and below-ground biomass Table 3.3.8, IPCC, 2003
Perennial 2.1 2.25 28 Above-ground biomass Table 3.3.2, IPCC, 2003
Grassland High producing 6.75 6.75 1 Above- and below-ground biomass Table 3.4.9, IPCC, 2003
Low producing 3.05 3.05 1 Above- and below-ground biomass Table 3.4.9, IPCC, 2003
With woody biomass NE 1.04 28 All biomass pools Wakelin, 2004
Wetlands   NE NE NA NA Assume steady state, IPCC, 2003
Settlements   NE NE NA NA Assume steady state, IPCC, 2003
Other land   NE NE NA NA Assume steady state, IPCC, 2003

Note: NE is not estimated and NA is not applicable See below in section 7.1.2 and under Methodological issues in each category-specific section for further details.

Representation of land areas

In this submission, the total land area of New Zealand used for all estimates of activity data is 26,851.1 kha. This value includes all significant New Zealand land masses, and comprises the North Island, South Island, Stewart Island, Great Barrier Island, Little Barrier Island and the Chatham Islands. All other small, outlying islands are excluded from the calculation of New Zealand’s land-use areas as they are not subject to land-use change. The excluded area is 74,430 hectares, comprising less than 1 per cent of the total land area of New Zealand (66,637 hectares of which is accounted for by the Auckland Islands and Campbell Island).

New Zealand has used a mix of Approaches 2 and 3 to map land-use changes between 1 January 1990 and 31 December 2008 (IPCC, 2003, chapter 2.3.2.3). The areas of forest as at 1 January 1990 and 1 January 2008 are based on wall-to-wall mapping of satellite and aircraft remotely sensed imagery taken in, or close to the start of, 1990 and 2008. Land-use changes during 2008 are then interpolated from other sources. This is described in further detail under the Land-use change during 2008 section.

In this submission, the land-use subcategories mapped are different from those used in earlier submissions. The land-use subcategories used in this submission are defined in Table 7.1.2.6.

Table 7.1.2.6 New Zealand’s definitions for land-use subcategories as mapped
Land-use subcategory Definition
Natural forest Areas that on 1 January 1990 were:
  • Tall forest on Department of Conservation land, including self-sown exotic trees.
  • Short forest or shrubland (with potential to reach ≥5 metres at maturity in situ) on Department of Conservation land.
  • Roads/tracks less than minimum width on Department of Conservation land, within the above two categories.
  • Tall non-planted forest (≥30 per cent cover) on other (non-Department of Conservation) land.
  • Broadleaved hardwood shrubland (eg, mahoe (Melicytus ramiflorus), wineberry (Aristotelia serrata), Pseudopanax spp., Pittosporum spp.), manuka/kanuka (Leptospermum scoparium/Kunzea ericoides) shrubland or other woody shrubland (≥30 per cent cover, with potential to reach ≥5 metre at maturity in situ) on other (non-Department of Conservation) land under current land management.
Pre-1990 planted forest
  • Radiata pine (Pinus radiata), Douglas-fir (Pseudotsuga menziesii), eucalypts (Eucalyptus spp.), or other planted species (with potential to reach ≥5 metre height at maturity in situ). This includes riparian or erosion control plantings that meet the forest definition.
  • Harvested areas within pre-1990 planted forest (assumes these will be replanted, unless deforestation is later detected).
  • This includes roads/tracks/skids less than minimum area/width of 30 metres within pre-1990 planted forest areas.
Post-1989 forest
  • Includes forests that meet the forest definition and have either been planted or established on or after 1 January 1990 onto land that was non-forest land as at 31 December 1989. Generally, these forests are planted with exotic species, but they may arise from natural regeneration of indigenous tree species as a result of management change after 1 January 1990.
  • For exotic forest, may include radiata pine (Pinus radiata), Douglas-fir (Pseudotsuga menziesii), eucalypts (Eucalyptus spp.), or other planted species (with the potential to reach ≥5 metres height at maturity in situ).
  • Includes roads/tracks/skids less than a minimum area/width of 30 metres within post-1989 forest areas.
Cropland – annual
  • All annual crops.
  • All cultivated bare ground.
  • Linear shelterbelts associated with annual cropland.
Cropland – perennial
  • All orchards and vineyards.
  • Linear shelterbelts associated with perennial cropland.
Grassland – high producing
  • Grassland with exotic species (eg, Perennial Ryegrass (Lolium perenne L).
  • Excludes linear shelterbelts that are larger than the minimum area/width criteria. (These are mapped separately as grassland – with woody biomass.)
Grassland – low producing
  • Low fertility grasses on hill country.
  • Tussock grasslands (eg, Chionochloa and Festuca spp).
  • Montane herbfields at either an altitude higher than above-timberline vegetation, or where the herbfields are not mixed up with woody vegetation.
  • Excludes linear shelterbelts that are larger than the minimum area/width criteria. (These are mapped separately as grassland – with woody biomass).
  • Other areas of limited vegetation cover and significant bare soil.
Grassland – with woody biomass
  • Grassland with tall tree species (<30 per cent cover), such as golf courses in rural areas (and except where the Land Cover Database (LCDB) has classified these as settlements).
  • Grassland with riparian or erosion control plantings (<30 per cent cover).
  • Grassland with matagouri (Discaria toumatou) and sweet briar (Rosa rubiginosa), broadleaved hardwood shrubland (eg, mahoe (Melicytus ramiflorus), wineberry (Aristotelia serrata), Pseudopanax spp., Pittosporum spp.), manuka/kanuka (Leptospermum scoparium/Kunzea ericoides), manuka/kanuka (Leptospermum scoparium/Kunzea ericoides) shrubland and other woody shrubland (<5 metres and any per cent cover) where under current management it is expected that the forest criteria will not be met over a 30–40 year time period.
  • Above timberline shrubland vegetation and intermixed with montane herbfields (does not have the potential to >5 metres height in situ).
  • Linear shelterbelts that meet area/width criteria of 30 metres.
Wetland – open water
  • Lakes and rivers.
Wetland – vegetated
  • Herbaceous and/or non-forest woody vegetation that may be periodically flooded. Includes scattered patches of tall tree-like vegetation in the wetland environment where cover <30 per cent.
  • Estuarine/tidal areas including mangroves.
Settlements
  • Built-up areas and impervious surfaces.
  • Grassland within ‘settlements’ including recreational areas.
  • Urban parkland and open spaces.
Other land
  • Largely bare ground (if not cropland).
  • Montane rock/scree.
  • Any other remaining land.

The new 2008 land-use subcategories were chosen as they better conform to the dominant land-use types in New Zealand while still enabling reporting under the land-use categories specified in IPCC (2003). The alignment of these new land-use subcategories to those used in previous reports is shown in the Table 7.1.2.7 below.

Table 7.1.2.7 Change to land-use categories since the 2009 submission
Land-use categories in the 2009 submission In this submission
Forest land Forest land
Natural forest and (some) grassland – with woody biomass Natural forest
Plantation crop and plantation understory Pre-1990 planted forest
Post-1989 forest
Cropland Cropland
Cropland – annual Cropland – annual
Cropland – perennial Cropland – perennial
Grassland Grassland
Wetlands
Grassland – high producing Grassland – high producing
Grassland – low producing and (some) grassland with woody biomass and (some) wetland – vegetated Grassland – low producing
Grassland – with woody biomass
Wetlands Wetlands
Wetlands – managed Wetlands – unmanaged (‘vegetated’ and ‘open water’)
Wetlands – unmanaged
Settlements Settlements
Other land Other land

Note: Mapping between the subcategories used in the 2009 submission and in this submission is not 1:1 because some areas that were reported as separate subcategories in this submission were reported across more than one subcategory in the 2009 submission.

Wetlands have been split into ‘wetland – open water’ and ‘wetland – vegetated’ in the land-use mapping but they are reported on together in the common reporting format tables as ‘wetlands’.

Further refinements are planned to improve these estimates of land-use change, as stated at the end of this section under planned improvements. Land areas reported as ‘converted’ and ‘remaining’ within each land-use category are the best current estimates and will be improved should additional activity data become available.

Land-use mapping

Land-use mapping – 1990

The 1990 land-use map is derived from 30 metre spatial resolution Landsat 4 and Landsat 5 satellite imagery taken in, or close to, 1990. The first of the images used were taken in November 1988 and the last in February 1993. In addition to orthorectification and atmospheric correction, the satellite images were standardised for spectral reflectance using the Ecosat algorithms documented in Dymond et al (2001), Shepherd and Dymond (2003) and Dymond and Shepherd (2004). These standardised images were used for the automated mapping of woody biomass, and then used to map woody biomass classes into the land-use subcategories being used for reporting. These land-use subcategories at 1990 included natural forest, pre-1990 planted forest and grassland with woody biomass.

This classification process was validated and improved using 15 metre resolution Landsat 7 ETM+ imagery acquired in 2000–2001, and SPOT 2 and 3 data acquired in 1996–1997. The use of this higher-resolution imagery (coupled with the use of concurrent aerial photography) enabled more certain land-use mapping decisions to be made. A detailed description of this mapping process is provided in chapter 11, section 11.2.2.

To determine the spatial location of the other land-use categories and subcategories as at 1990 and 2008, information from two Land Cover Databases, LCDB1 (1996) and LCDB2 (2001) (Thompson et al, 2004), the New Zealand Land Resource Inventory (NZLRI) (Eyles, 1977) and hydrological data from Land Information New Zealand (a government agency) have been used (Shepherd and Newsome, 2009a, b).

The NZLRI database was used to better define the area of high- and low-producing grassland. Areas tagged as ‘improved pasture’ in the NZLRI vegetation records were classified as grassland – high producing in the land-use maps. All other areas were classified as grassland – low producing. Figure 7.1.2.2 illustrates this mapping process.

Figure 7.1.2.2 New Zealand’s land-use mapping process

Figure 7.1.2.2 New Zealand’s land-use mapping process

Figure 7.1.2.2 illustrates New Zealand’s land-use mapping process. The land-use map is derived from standardised satellite imagery, which is used to map areas of woody biomass, and then automatically classify these woody biomass areas into the land-use subcategories being used for reporting (natural forest, pre-1990 planted forest, post-1989 forest and grassland with woody biomass). The areas of other, non-woody land-use subcategories (annual cropland; perennial cropland; vegetated, non-forest wetlands; settlements; and other land) are then in-filled with information from two Land Cover Databases, LCDB1 (1996) and LCDB2 (2001), and the wetlands – open water areas are then in-filled with hydrological data from Land Information New Zealand (LINZ). The remaining unmapped areas are then in-filled with the two grassland subcategories (high-producing and low-producing) based on information from the New Zealand Land Resource Inventory (NZLRI).

An interpretation guide for automated and visual interpretation was prepared and used to ensure a consistent basis for all mapping processes (Dougherty et al, 2009). Independent quality control was performed for all mapping. This involved an independent agency looking at randomly selected points across New Zealand and using the same data as the original operator to decide what land use the point fell within. The two operators were in agreement at least 95 percent of the time. This is described in more detail in GNS Science (2009).

Land-use mapping – 2008

The 2008 land-use map (land use as at 1 January 2008) is derived from 10 metre spatial resolution SPOT 5 satellite imagery and was processed into standardised reflectance images, using the same approach as for the 1990 imagery. The SPOT 5 imagery was taken over the summers of 2006–07 and 2007–08 (November to April), to establish a national set of cloud-free imagery. Where the SPOT 5 imagery pre-dates 1 January 2008, a combination of aerial photography, Moderate Resolution Imaging Spectroradiometer (MODIS) satellite imagery and field verification has been used to identify where deforestation has occurred to ensure that the 2008 land-use map is as accurate as possible. Further details are provided below under the Mapping of deforestation and harvesting section.

A list of the land-use categories and subcategories used to map land use and land-use change from 1990 to 2008 is provided in Table 7.1.2.6. Maps showing the land use in New Zealand as at 1 January 1990 and 1 January 2008 are shown in Figures 7.1.2.4 and 7.1.2.5.

Figure 7.1.2.3 Land-use map of New Zealand as at 1 January 1990

Figure 7.1.2.3 Land-use map of New Zealand as at 1 January 1990

Figure 7.1.2.3 provides the land-use map of New Zealand as at 1 January 1990, as derived from the LUCAS mapping programme. The areas of each land use are provided in the table below.

Land-use category Subcategory Area in 1990 (000 ha) % of total area
Forest Land Natural forest                                 8,152.550 30.4%
  Pre-1990 planted forest                                 1,480.345 5.5%
  Post-1989 forest                                              -   0.0%
Cropland annual                                    339.191 1.3%
  perennial                                      78.727 0.3%
Grassland high-producing                                 5,856.474 21.8%
  low-producing                                 8,016.862 29.9%
  with woody biomass                                 1,184.680 4.4%
Wetlands                                      644.060 2.4%
Settlements                                      203.439 0.8%
Other Land                                      894.785 3.3%
Total area of New Zealand                               26,851.090 100.0%

Note: The insert map is of the Chatham Islands, which lie approximately 660 km south-east of the Wairarapa coast, or 800 km due east of Banks Peninsula.

Figure 7.1.2.4 Land-use map of New Zealand as at 1 January 2008

Figure 7.1.2.4 Land-use map of New Zealand as at 1 January 2008

Figure 7.1.2.4 provides a land-use map of New Zealand as at 31 December 2008, as derived from the LUCAS mapping programme. The areas of each land use are provided in the table below.

Land-use category Subcategory Area in 2008 (000 ha) % of total area
Forest Land Natural forest                                 8,118.004 30.2%
  Pre-1990 planted forest                                 1,430.286 5.3%
  Post-1989 forest                                    580.523 2.2%
Cropland annual                                    334.159 1.2%
  perennial                                      88.541 0.3%
Grassland high-producing                                 5,813.712 21.7%
  low-producing                                 7,701.148 28.7%
  with woody biomass                                 1,056.975 3.9%
Wetlands Wetlands                                    644.135 2.4%
Settlements Settlements                                    206.288 0.8%
Other Land Other Land                                    889.068 3.3%
Total area of New Zealand                               26,851.090 100.0%

Note: The insert map is of the Chatham Islands, which lie approximately 660 km south-east of the Wairarapa coast, or 800 km due east of Banks Peninsula.

Mapping of deforestation and harvesting

New Zealand has used a combination of data sources to identify the location and timing of deforestation prior to 1 January 2008. Land-use data generated from classification of SPOT 5 satellite imagery acquired between November 2006 and April 2008 was used to identify the conversion of land from a forest land use to a non-forest land use. Evidential information to confirm land-use change was collected using higher-resolution aerial photography and field visits. This is illustrated in Figure 7.1.2.5.

Figure 7.1.2.5 New Zealand’s identification of deforestation

Figure 7.1.2.5 New Zealand’s identification of deforestation

Figure 7.1.2.5 illustrates New Zealand’s mapping process to identify deforestation. New Zealand has used a combination of data sources to identify the location and timing of deforestation prior to 1 January 2008, including satellite images, which are shown in the top two pictures. The top-left 1990 image is a 30 m resolution Landsat 4 satellite image acquired 25 December 1990. The arrow points to a pre-1990 planted forest polygon as at 1 January 1990. The top-right 2008 image is a 10 m SPOT 5 satellite image acquired 29 December 2007. The arrow points to a grassland – high producing polygon as at 1 January 2008, which was previously pre-1990 planted forest.

Land-use maps are generated from the classification of satellite imagery such as these, and used to identify the conversion of land from a forest land use to a non-forest land use (the middle two images, which show the mapping of land-use categories in 1990 and 2008).

Evidential information to confirm land-use change is then collected using higher-resolution aerial photography and field visits. This is illustrated by the bottom photograph, taken of a grassland area which shows evidence of forest harvesting, which was taken on 25 June 2009 during field verification of areas identified as deforested prior to 1 January 2008.

To map deforestation and harvesting between 2008 and 2012, a similar approach will be used. Two years after harvesting has been mapped there will be an inspection of areas mapped as harvested, based on satellite imagery. The first exercise will be done using imagery acquired in December 2009 and January 2010. This exercise will identify harvesting and deforestation that will have occurred during 2008 and 2009.

Areas of possible deforestation will be confirmed using aerial photography, airborne scanning LiDAR and digital aerial photography. Supporting information from regional councils, Ministry of Agriculture and Forestry district offices and forestry consultants will also be searched to see if deforestation or restocking can be confirmed. This process is shown in Figure 11.2.2.1 in chapter 11.

Where areas of harvesting are unable to be confirmed as either restocked or deforestation, the proportion of potential deforestation in 2010 and 2011 will be estimated based on the harvesting and deforestation data collected for the years 2008 and 2009. These estimates will be validated in 2013 when wall-to-wall mapping of land use as at 31 December 2012 occurs.

For this submission, the estimate of the total area harvested each year between 1990 and 2007 is based on the area harvested as reported in the National Exotic Forest Description, a survey conducted by the Ministry of Agriculture and Forestry (Ministry of Agriculture and Forestry, 2009a). Data for the year ending 31 December 2008 was not available so a combination of roundwood statistics (the volume of roundwood harvested, also produced by the Ministry of Agriculture and Forestry) and the ratio of roundwood volume to area harvested over the five-year period 2002–2007 was used to estimate the area harvested in 2008 from the volume of roundwood produced. The harvesting values for 2008 will be updated in next year’s submission when finalised data for 2008 becomes available.

The total area harvested was then split by forest type.

  • Natural forest: In 2008, 0.05 per cent of New Zealand’s total forest timber production was from the harvesting of natural forests (Ministry of Agriculture and Forestry, 2009c).
  • Post-1989 forest: There is no published information available for the area of post-1989 forest harvesting in New Zealand, but most post-1989 forest harvesting is of eucalypt species for the supply of pulp for export, or to local pulp and paper mills. Experts in the various regions where eucalypts are commercially grown were contacted and asked about the level of harvesting they believed was occurring. Where possible, these expert opinions were corroborated with publically available information from companies’ websites and various other reports.
  • Pre-1990 planted forest harvesting: This was estimated as the difference between total harvesting (based on statistics from the Ministry of Agriculture and Forestry, as outlined above) and the amount of post-1989 forest harvesting estimated.
Land-use change
Land-use change during 2008

The 2008 land-use map shows land use and land-use change up to 1 January 2008. Land-use change occurring during 2008 was not mapped due to the expense of a national annual mapping programme. To fill this information gap for all land-use subcategories except pre-1990 planted forest and post-1989 forest, the average annual land-use change from 1990 to the end of 2007 was used to extrapolate an estimation of change in 2008. As there have been changes in the drivers for land-use changes for pre-1990 planted forest and post-1989 forest (including changes to legislation to discourage deforestation and initiatives to encourage afforestation), using the average annual change was not appropriate for these land-use subcategories, and so activity data was taken from other data sources.

The data sources used to estimate land-use change for pre-1990 planted forests and post-1989 forests (new planting, ie, afforestation and reforestation; and deforestation) are listed below.

  • For pre-1990 planted forest and post-1989 forest deforestation: a combination of data from the 2008 Deforestation Survey (Manley, 2009) and unpublished work by Scion (NZ Forest Research Institute) was used. The work by Scion is referred to in Wakelin, 2008.
  • For post-1989 afforestation/reforestation: data from the National Exotic Forest Description as at 1 April 2008 (Ministry of Agriculture and Forestry, 2009a) was used. The data for 2008 is still provisional and will be updated for the 2011 submission.

Land-use change from 1990 to 2008

Table 7.1.2.8 is a land-use change matrix for the years 1990 to 2008 using the 1990 and 2008 land-use maps, and activity data on land change during 2008 from the Ministry of Agriculture and Forestry.

Some prominent land-use changes between 1990 and 2008 include:

  • forest establishment of 580,524 hectares (classified as post-1989 forest) that has mostly occurred on land that was previously grassland, primarily low-producing grassland. Approximately 11,500 hectares has subsequently been deforested.
  • deforestation of 96,375 hectares. This has occurred mainly since 2004. Between 1990 and 2004 there was very little deforestation in New Zealand, due to market conditions.

Table 7.1.2.9 shows a similar land-use change matrix, for 2007 to 2008.

Table 7.1.2.8 New Zealand’s land-use change matrix from 1990 to 2008 (1, 2, 3, 4, 5, 6)

Table 7.1.2.8 New Zealand’s land-use change matrix from 1990 to 2008 (1, 2, 3, 4, 5, 6)
                                         

1990



2008

Forest land Cropland Grassland Wetlands Settlements Other land Net area 31 Dec 2008 (kha)
Natural Pre-1990 Planted Post-1989 Annual Perennial High producing Low producing With woody biomass Wetlands Settlements Other land
Forest land Natural 8,118.0                     8,118.0
Pre-1990 planted   1,430.3                   1,430.3
Post-1989     NA 0.0 0.0 106.4 348.7 120.5 0.0 0.0 4.8 568.8
Cropland Annual       334.0   0.1 0.0       0.0 334.2
Perennial   0.2 0.0 5.2 78.6 4.2 0.2 0.1       88.5
Grassland High producing 7.0 45.3 10.5     5,735.7 0.0 14.9 0.1   0.1 5,813.7
Low producing 24.1 3.9 1.0     0.0 7,641.4 30.3 0.0   0.4 7,701.1
With woody biomass 3.1 0.0 0.0     8.2 25.9 1,018.5   0.0 1.2 1,057.0
Wetlands Wetlands   0.0 0.0           643.9 0.1 0.1 644.1
Settlements Settlements 0.1 0.4 0.1   0.1 1.7 0.2 0.2 0.0 203.3   206.3
Other land Other land 0.3 0.3 0.1     0.1 0.1 0.1     888.1 889.1
Area as at 1 Jan 1990 (kha) 8,152.6 1,480.3 0.0 339.2 78.7 5,856.5 8,016.9 1,184.7 644.1 203.4 894.8 26,851.1
Net change 1 Jan 1990–31 Dec 2008 –34.5 –50.1 568.8 –5.0 9.8 –42.8 –315.7 –127.7 0.1 2.8 –5.7 0.0
Net change 1990–2008 (%) –0.4 –3.4 NA –1.5 12.5 –0.7 –3.9 –10.8 0.0 1.4 –0.6 0.0

Table 7.1.2.9 New Zealand’s land-use change matrix from 2007 to 2008 (1, 2, 3, 4, 5, 6)

Table 7.1.2.9 New Zealand’s land-use change matrix from 2007 to 2008 (1, 2, 3, 4, 5, 6)
                                         

2007



2008

Forest land Cropland Grassland Wetlands Settlements

Other land

Net area 31 Dec 2008 (kha)

Natural Pre-1990 planted Post-1989 Annual Perennial High producing Low producing With woody biomass Wetlands Settlements

Other land

Forest land Natural 8,118.0                  

 

8,118.0

Pre-1990 planted   1,430.3                

 

1,430.3

Post-1989     579.5 0.0 0.0 0.2 0.6 0.2 0.0 0.0

0.0

580.5

Cropland Annual       334.2 0.0 0.0 0.0      

0.0

334.2

Perennial   0.0 0.0 0.3 88.0 0.2 0.0 0.0    

 

88.5

Grassland High producing 0.4 1.9 0.8     5,809.8 0.0 0.8 0.0  

0.0

5,813.7

Low producing 1.3 0.2 0.1     0.0 7,698.0 1.6 0.0  

0.0

7,701.1

With woody biomass 0.2 0.0 0.0     0.4 1.4 1,055.0   0.0

0.1

1,057.0

Wetlands Wetlands   0.0 0.0           644.1 0.0

0.0

644.1

Settlements Settlements 0.0 0.0 0.0   0.0 0.1 0.0 0.0 0.0 206.1

 

206.3

Other land Other land 0.0 0.0 0.0     0.0 0.0 0.0    

889.0

889.1

Net area as at 31 Dec 2007 (kha) 8,119.8 1,432.4 568.8 334.4 88.0 5,810.8 7,700.0 1,057.6 644.1 206.2

889.1

26,851.1

Net change 31 Dec 2007–31 Dec 2008 –1.8 –2.1 0.1 –0.3 0.5 2.9 1.1 –0.6 0.0 0.1

0.0

0.0

Net change 2007–2008 (%) 0.0 –0.1 –0.1 0.6 0.1 0.0 –0.1 0.0 0.1

0.0

0.0

Notes: (1) Units in 000’s hectares. (2) The minimum area shown for land-use change is 100 ha, however, areas are mapped to 1 ha resolution. (3) Zeros are displayed where land-use changes are of less than 100 ha, blank cells indicate no land-use change during the period. (4) Land-use values since 1990 do not sum to total New Zealand area due to double-counting of 11,749 hectares of deforestation of post-1989 forest. Columns and rows may not total due to rounding. (5) Shaded cells indicate land remaining in each category. (6) Land-use change values refer to change over the course of the year. Land-use area values are as at point in time indicated (31 December for 2007 and 2008; 1 January for 1990).

Methodological change

The total land area of New Zealand used in this submission is 26,851.1 kha. This differs from the area used in the 2009 submission that reported 26,821.6 kha (a difference of 0.1 per cent). The difference arises from the use of alternative coastal boundaries. In the 2009 submission, the New Zealand boundaries were derived from the Land Cover Databases (LCDBs) that are based on satellite imagery. Where LCDB imagery stopped short of the coastline, the area between the LCDB edge and the coastline was ignored (not counted), creating a shortfall in the total area mapped. In this submission, we have been able to correct this as the imagery collected for LUCAS covers the total area of New Zealand (the edges of the imagery extend beyond the coastline). This means that the total area of New Zealand in this submission is the same as that derived from the official coastal boundary provided by Land Information New Zealand, the official government agency responsible for cadastral mapping in New Zealand (www.linz.govt.nz). 

While New Zealand previously represented land areas using Approach 3, land areas were calculated using previously existing land cover databases – LCDB1 (c. 1996) and LCDB2 (c. 2001) (Thompson et al, 2004). The LCDBs were not specifically developed for use in Climate Change Convention reporting. The area of forest as at 1990 and 2008 presented in this submission is based on wall-to-wall mapping of satellite and aircraft remotely sensed imagery taken in, or close to, the start of 1990 and 2008. LCDB1 and 2 data, as well as information from the New Zealand Land Resource Inventory (Eyles, 1977) and hydrological data from Land Information New Zealand, have been used to map the non-forest categories. The approach used to map land use is described in Shepherd and Newsome (2009a, b).

Uncertainties and time-series consistency

Due to constraints in time and resources, New Zealand has not completed a full accuracy assessment to determine uncertainty in the mapping data. However, the approach to mapping land-use change between 1990 and 2008 is based on a peer-reviewed and published work by Dymond et al (2008). With this approach, it was estimated that an accuracy of within ±7.0 per cent of actual afforestation can be achieved in mapping change in planted forests in New Zealand. One of the planned improvements for the activity data is to perform an accuracy assessment and determine the uncertainty for the woody biomass categories mapped under LUCAS. The levels of uncertainty for non-woody classes (±6.0 per cent) and for natural forest (±4.0 per cent) are similar to what was reported in previous submissions because the same data sources have been used.

The accuracy of mapping land-use changes between 1990 and 2008 has not been determined and will be included in the next submission.

Source-specific QA/QC and verification

Quality-control and quality-assurance procedures have been adopted for all data collection and data analyses, consistent with GPG-LULUCF and New Zealand’s inventory quality-control and quality-assurance plan. Data quality and data assurance plans are established for each type of data used to determine carbon stock and stock changes, as well as for the mapping of the areal extent and spatial location of land-use changes.

The 1990 and 2007 land-use mapping data have been independently checked to determine the level of consistency in satellite image classification to the requirements set out in the Guide to Mapping Woody Land Use Classes Using Satellite Imagery (Dougherty et al, 2009). Through this process, approximately 28,000 randomly selected points in the 1990 and 2008 woody classes were evaluated by independent assessors. From this exercise, 91 per cent of the time, independent assessors agreed with the original classification. Where there was disagreement, the points were recorded in a register and this has been used as the basis for preparing the improvement plan described in this report. The process does not determine errors of omission/commission that would provide an accuracy assessment and definitive level of uncertainty. (An error of commission is where a particular class has been mapped incorrectly, eg, as a result of similarities in spectral signatures; an error of omission error is where mapping has failed to detect a particular land use, eg, a planted forest block visible in imagery.)

The approach used to implement quality-assurance processes is documented in the LUCAS Data Quality Framework (PricewaterhouseCoopers, 2008).

Source-specific planned improvements

The quality-control and quality-assurance process followed during mapping exposed a number of limitations in the mapping method. Future improvements to both the 1990 and 2008 maps will focus on these areas.

  • The land-use mapping approach for both 1990 and 2008 mapping involved visual assessment and classification of all polygons greater than 5 hectares in woody classes. Polygons between 1 and 5 hectares, while classified, are likely to have a lower confidence limit and are tagged for further analysis during 2010. However, the area of New Zealand covered by polygons in the 1 to 5 hectare category only represents between 2–4 per cent of the area associated with the polygons larger than 5 hectares in size (Shepherd and Newsome, 2009a, b).
  • The mapping of 1990 land use presented challenges, particularly in identifying newly established exotic forests using Landsat satellite imagery. Where trees are planted within 3 years of the image acquisition date, they (and their surrounding vegetation) are unlikely to show a distinguishable spectral signature on 30 metre resolution imagery. For LUCAS mapping, this situation is compounded by the lack of ancillary data to support land-use classification decisions at 1990. Land-use mapping will be updated and improved as more detailed land-use information becomes available from the New Zealand Emissions Trading Scheme. The Ministry of Agriculture and Forestry is administering the forestry component of the Emissions Trading Scheme and applicants to the scheme will be providing new land-management and land-use information as at 1990.
  • Cropland areas were mapped using historical boundaries from LCDB2. It is expected that the mapping of this land-use category will be improved by either visually interpreting the satellite imagery at both 1990 and 2008, or by analysis of a New Zealand farm enterprise database (AgriBase) that has nationwide annual cropping spatial statistics from about 1999.

New Zealand will create a 2012 land-use map using high-resolution satellite data as the key source of information at the end of the first commitment period. This mapping will be used to make comparisons with the 2008 land-use map (prepared using similar high-resolution imagery) to improve the spatial determination of harvesting, deforestation and land-use changes between 1 January 2008 and 31 December 2012.

Soils

In this submission, New Zealand uses a Tier 2 method to estimate soil carbon stock, with the use of New Zealand-specific land-use and soil pedon data (Scott et al, 2002). This is an improvement from the method used in the 2009 submission, as the previous method used a New Zealand-specific reference soil carbon stock value. This is explained further under Methodological change below.

The resulting peer-reviewed Soil Carbon Monitoring System (Soil CMS) is used to quantify 1990 baseline soil carbon stocks for the organic fraction of the mineral soils and to estimate subsequent changes in soil carbon stocks associated with land-use change (Tate et al, 2003a, b; Tate et al, 2004).

The Soil CMS does not estimate carbon stock or carbon changes for organic soils, as it calculates the total carbon from the carbon concentration of soil within a fixed depth, rather than the soil’s total organic carbon mass (Tate et al, 2005). In order to accurately determine emissions from organic soils undergoing land-use change, changes in total organic carbon mass would need to be estimated, as any emissions would be caused by the loss of soil volume, rather than by changes to the soil’s carbon concentration.

As a result of this constraint in the methodology, and as limited New Zealand data currently exists on the impact of land-use change on the total organic carbon mass of organic soils, New Zealand has estimated emissions from organic soils by aggregating the activity data for the two soil types together, and using the emission factors for mineral soils as the default emission factors for organic soils. New Zealand has accordingly recorded the notation key IE (‘included elsewhere’) for organic soils in the common reporting format tables.

While this methodology may slightly underestimate New Zealand’s soil carbon emissions, it should be recognised that organic soils occupy a relatively small proportion of New Zealand’s total land area (0.9 per cent) and that between January 1990 and January 2008, only 2,560 hectares of land with organic soils underwent land-use change, representing just 0.3 per cent of the total area of land-use change in New Zealand.

Model

Based on the well-established premise that the concentration of soil carbon is largely controlled by soil type, climate and land use (Tate et al, 1999), the Soil CMS pre-stratified the country by these three factors, namely: soil class, climate and land use. This resulted in 39 combinations of these three factors, called cells, and describes 93 per cent of the New Zealand landscape (Tate et al, 2003a, b). An ‘erosion index’ (slope x rainfall) factor was later added (Tate et al, 2005). Geo-referenced soil carbon data (0–0.3 metre depth increment) are used to quantify average soil carbon (t C ha–1) for each combination using a General Linear Model.

A key assumption of the model is that the soil carbon values in the national soil pedon database represent equilibrium soil carbon values for each soil or land use combination, with a variety of tests indicating that this is a reasonable assumption (Tate et al, 2002, 2005). It is further assumed that change in land use is the key determinant of change in soil carbon (Tate et al, 2005). Estimation of change in soil carbon with change in land use is calculated based on the differences in equilibrium soil carbon values (t C ha–1) between the initial and final land use (Tate et al, 2002) over a 20-year period (IPCC default). The change in carbon is then multiplied by the area of change mapped (Table 7.1.2.8)

Data

The LUCAS soil dataset consists primarily of historical data extracted from the National Soils Database (Landcare Research Limited), that gives national coverage of undisturbed representative or modal soil pedons collected for soil survey purposes, and from the Forest Nutrition dataset (Forest Research Institute Limited). It has been supplemented with recent data collected specifically for the Soil CMS to fill gaps in the geographical coverage and to increase the number of data points for land uses of particular interest to help reduce uncertainty for these land uses (Baisden et al, 2006b).

The consolidated LUCAS dataset has recently been reclassified to use the current land-use categories. This reclassification was based on data from the original plot sheets. Potential bias in cropland soil carbon stock estimates from using data from undisturbed pedons has been removed by invalidating any records that were not specifically collected within the cultivated area (Fraser et al, 2009). The current dataset consists of 1,235 records, distributed across land uses as shown in Figure 7.1.2.6 and Figure 7.1.2.7.

Figure 7.1.2.6 New Zealand soil plot distribution (plots used for 2009 model estimation of soil carbon stocks)

Figure 7.1.2.6 New Zealand soil plot distribution (plots used for 2009 model estimation of soil carbon stocks)

Figure 7.1.2.6 illustrates the distribution of the soil plots used for the 2009 model estimation of soil carbon stocks in New Zealand. The current dataset consists of 1,235 records, distributed across land uses as shown in the table below.

Land-use category Number of soil plots
Natural forest 192
Post-1989 forest 9
Pre-1990 planted forest 64
Cropland - annual 3
Cropland - perennial 11
Grassland - high producing 526
Grassland - low producing 273
Grassland - with woody biomass 152
Wetlands 5
Total 1235

Figure 7.1.2.7 Number of New Zealand soil plot sites by land-use and dataset

Figure 7.1.2.7 Number of New Zealand soil plot sites by land-use and dataset

Land-use category Provenance of dataset      
  National Soil Database Forest Nutrition Database MfE Gap Filling Total
Natural forest 144 2 46 192
Post-1989 forest 0 0 9 9
Pre-1990 planted forest 11 47 6 64
Cropland - annual 3 0 0 3
Cropland - perennial 11 0 0 11
Grassland - high producing 522 2 2 526
Grassland - low producing 258 3 12 273
Grassland - with woody biomass 134 0 18 152
Wetlands 5 0 0 5
Total 1088 54 93 1235

Note: MfE is the Ministry for the Environment.

Testing and validation

Testing of the Soil CMS was completed to evaluate its ability to predict soil carbon stocks at regional and local scales. The results from the Soil CMS have been compared against independent, stratified soil sampling for 24,000 hectares of South Island low-producing grassland (Scott et al, 2002) and for an area of the South Island (about 6,000 hectares) containing a range of land-cover and soil-climate categories (Tate et al, 2003a, b). A regional-scale validation exercise has also been performed using the largest climate/soil/land-use combination cell (Moist Temperate Volcanic Grassland), with independent random sampling of 12 profiles taken on a fixed grid over a large area (2,000 km2). Mean values derived from the random sampling were well within the 95 per cent confidence limits of the database values (Wilde et al, 2004; Tate et al, 2005). Overall, tests have indicated that the Soil CMS estimates soil carbon stocks reasonably well at a range of scales (Tate et al, 2005).

The system has also been validated for its ability to predict soil carbon changes between land uses at steady state for New Zealand’s main land-use change, grassland converted to planted forest. This was done by comparing the Soil CMS results with estimates based on a ‘paired-site’ approach (Tate et al, 2003b; Baisden et al, 2006a). The paired-site approach compares two nearby sites that have reasonably uniform morphological properties, which originally were under a single land use, and where one site has since changed to a different land use, with sufficient time having elapsed for it to have reached steady state values for soil carbon (Baisden et al, 2006a, b). Therefore the influence that differing soil types, climatic conditions and erosion regimes may have on soil carbon are removed and any resulting changes in soil carbon can be attributed to the change in land use. Results indicate that, once a weighting for forest species type has been applied to the paired-site dataset (to remove potential bias as Pinus radiata was under-represented in the analysis), the predictions of mean soil carbon from the Soil CMS model and paired sites are in agreement within 95 per cent confidence intervals (Baisden et al, 2006a, b).

Model outputs

Table 7.1.2.10 gives the national soil carbon stock estimates for all land-use categories and the associated standard errors. All estimates are produced by the Soil CMS, except for the other land estimate, as it is an IPCC default value.

Table 7.1.2.10 New Zealand’s soil carbon stock (0–0.3 m) for land-use categories (after McNeill et al, 2009)
Land use Soil carbon stock
(t C ha–1)
Natural forest 111.85 ± 5.24
Planted forest (pre-1990, post-1989) 104.31 ± 6.44
Annual cropland 118.27 ± 22.47
Perennial cropland 114.91 ± 13.22
High-producing grassland 114.93 ± 3.56
Low-producing grassland 117.66 ± 12.56
Grassland with woody biomass 111.57 ± 4.29
Wetlands 104.62 ± 19.92
Settlements 117.66 ± 12.56
Other land 88
Methodological change

In previous submissions, New Zealand has used a country-specific reference soil carbon stock value, adjusted by IPCC land-use, management and input factor values (IPCC, 2003). However, land-use factors were double-counted in previous submissions for some land-use categories, and no land-use factors were available for key land-use categories (natural and planted forest). For this submission, New Zealand has used the Soil CMS. This represents a significant change in method from the previous submission.

The Soil CMS has been developed and refined over time to remove bias and increase accuracy. A recent refinement to the Soil CMS has been to remove the effect of bias from spatial clustering of soil samples. As the dataset used by the model consists primarily of historical data collected for specific purposes, it is not a random sample of soils in New Zealand, with some soil/climate/land-use combinations over-represented and some under-represented. As soil samples are correlated to some extent according to the distance between them, the use of the Soil CMS model with the dataset was resulting in estimates in soil carbon stocks that were biased (McNeill et al, 2009). A correction factor for spatial correlation between data points was incorporated into the model in 2009 to address this issue. This resulted in a decrease in the difference in stock estimates between low producing grassland and post-1989 forest (the major land-use change) from –18.43 t C ha–1 (Baisden et al, 2006b) to –13.35 t C ha–1 (McNeill et al, 2009).

Uncertainties and time-series consistency

Use of a General Linear Model allows estimates to be made of uncertainties associated with estimates of soil carbon changes. The standard errors for each land-use soil carbon estimate are given in Table 7.1.2.10. There is relatively little soil data for those land uses that comprise only minor land areas within New Zealand (croplands and wetlands), and the uncertainties associated with these estimates are correspondingly high. Uncertainties also arise from lack of soil carbon data for some soil/climate/land-use combinations (Scott et al, 2002), and from variations in site selection, sample collection and laboratory analysis with data from different sources and time periods (Wilde, 2003; Baisden et al, 2006a).

Other uncertainties in the Soil CMS include: the assumption that soil carbon is at steady state for all land uses, lack of soil carbon data and soil carbon changes estimates below 0.3 metres, potential carbon losses from mass-movement erosion, and a possible interaction between land use and the soil-climate classification (Tate et al, 2004, 2005).

Source-specific QA/QC and verification

Quality-control and quality-assurance procedures have been adopted for all data collection and data analyses, to be consistent with GPG-LULUCF and New Zealand’s inventory quality-control and quality-assurance plan.

  • Details of the quality-management system for data collection, laboratory analyses and database management of the National Soils Database, are given in Wilde (2003).
  • Recent data collection, analyses and management methods are subject to the soils quality-control and quality-assurance plan.
  • The consolidated soils dataset used within the Soil CMS has been subject to further quality-assurance procedures (Fraser et al, 2009).
  • The Soil CMS model has been subject to various forms of testing and validation (eg, Scott et al, 2002; Tate et al, 2005; Baisden et al, 2006a; McNeill et al, 2009), and has been published in peer-reviewed international journals (Scott et al, 2002; Tate et al, 2003a, b; Tate et al, 2005).
Source-specific planned improvements

Recent reviews of the Soil CMS identify a range of potential areas for improvement of the system (Baisden et al, 2006b; Kirschbaum et al, 2009). Those areas identified for future improvement include the following, and will be prioritised before any further improvements are agreed and funded.

  • Improvement in the model to better reflect the landform types where land-use change is occurring (eg, on eroding hill-country landscapes).
  • Improvement in the data by collecting more data for under-represented land-use and soil-climate cells, and to correct the data by removing sampling and analysis anomalies.
  • Further validation of the Soil CMS model by checking the results it gives against independent field soil/climate/land-use cell sampling and against targeted paired sites and time-series sampling to investigate soil carbon changes following specific land-use change.

Liming

In New Zealand, agricultural lime is mainly applied to acidic grassland and cropland soils to maintain or increase the productive capability of soils and pastures.

Information on agricultural lime (limestone) application is collected by the national statistics agency, Statistics New Zealand, as part of its annual Agriculture Production Survey. Previously, this survey has asked for the total weight (in tonnes) of lime applied but, for the first time in the 2008 survey, estimates of limestone use and dolomite use were reported separately. This showed that 1.2 per cent of total agricultural lime was dolomite. As this data is only available for one year and dolomite is such a small percentage of total lime use, lime is not separated from dolomite in this report or in the common reporting format tables.

The Agriculture Production Survey has gaps in the time series. No survey was carried out in 1991, or between 1997 and 2001. Linear interpolation has been used to represent the data for these years. Since 2002, there has been a drop in the amount of lime applied. It is unclear why this occurred but quantities applied do vary from year to year depending on a number of factors, including farming profitability.

Analysis of the results of the Agriculture Production Survey indicate that, each year, around 94 per cent of agricultural lime used in New Zealand is applied to grassland, with the remaining 6 per cent applied to cropland. Emissions associated with liming are estimated using a Tier 1 method (GPG-LULUCF equation 3.4.11, IPCC, 2003), and the IPCC default emission factor for carbon conversion of 0.12.

Biomass burning

Biomass burning is not a significant source of emissions for New Zealand due to the nature of New Zealand’s climate and vegetation.

New Zealand reports on emissions from wildfire in forest land and grassland, and controlled burning associated with land-use change from grassland to forest land based on national data on the area burnt. Emissions from controlled burning in land converted to grassland are not reported in the inventory because of insufficient data on the area of land converted to grassland that is burnt. All emissions from the burning of crop stubble and controlled burning of savanna are reported in the agriculture sector (chapter 6).

Tier 2 methodologies are employed to estimate emissions from biomass burning in New Zealand. Country-specific biomass densities are applied (Wakelin et al, 2009) with IPCC defaults used for most emission factors (IPCC, 2003 sections 3.4.2.1.1.2 and 3A.1.12). Activity data (area of land-use change) for the grassland with woody biomass converted to forest category is based on annual land-use changes as estimated in section 7.1.2 – Representation of land areas. For the land remaining land categories, activity data is sourced from the National Rural Fire Authority database, which has data from 1992 onwards.

The average area burnt between 1992 and 2008 from this database is used as the estimate of area burnt for 1990 to 1991 as the estimates for this period are inaccurate because of incomplete coverage in data collection. The March year data is then converted to calendar years for use in the inventory (Wakelin et al, 2009).

There has not been a significant change in wildfire activity since 1990 (Wakelin et al, 2009). Natural disturbance (lightning) induced wildfires account for only 0.1 per cent of burning in grassland and forest land in New Zealand (Wakelin, 2006; Doherty et al, 2008). Emissions from these events are not reported because the subsequent regrowth is not captured in the inventory. In this situation, GPG-LULUCF (3.2.1.4.2) states that “if methods are applied that do not capture removals by regrowth after natural disturbances, then it is not necessary to report the CO2 emissions associated with natural disturbance events”. In pre-1990 planted forest and post-1989 forest, the stock change calculations account for emissions from wildfire if the affected stand is harvested or the area is left to grow on at a reduced stocking. This means emissions may be underestimated where a mature stand is damaged during a wildfire event without a subsequent reduction in its net stocked area. Given the few incidences of wildfire in New Zealand’s planted forest lands, these are not regarded as a significant source of error (Wakelin, 2008).

Controlled burning can be used to clear the slash residues (residual forest material) left behind after the forest land has been harvested, as part of site preparation for the next land use. In New Zealand, it is assumed that 25 per cent of grassland converted to forest land is cleared by controlled burning. Different biomass-density values for wildfire and controlled burning on grassland with woody biomass are used in the inventory. The differences are due to vegetation that is converted to forest, which is generally of a lesser stature when compared with other shrubland (Wakelin, 2008). The inventory does not report on-site preparation burning activities on forest land remaining forest land, because activity data is not available and the practice is not thought to be significant. Controlled burning of grassland with woody biomass for the establishment or re-establishment of pasture has also not been included. Conversions of planted forest land to grassland (pasture) have increased in the past four years; current research seeks to quantify emissions from this activity for reporting in future submissions (Wakelin, 2008).

Uncertainties and time-series consistency

Uncertainties arise from relatively coarse activity data for wildfires and a paucity of data for most controlled burning activities in New Zealand. Both liming and biomass burning statistics have gaps in the time series where data collection did not occur or survey methodologies changed. Assumptions are made for some biomass densities and burning factors where insufficient data exists.

Source-specific QA/QC and verification

Quality-control and quality-assurance measures are applied to the biomass burning and liming section of the inventory. The biomass burning dataset is scientifically verified whenever new data is supplied. In 2006 and 2009, the biomass burning parameters (biomass densities, burning and emissions factors), assumptions and dataset were scientifically reviewed and updated. Data validation rules and plausibility tests were then applied to the dataset (Wakelin et al, 2009).

Source-specific planned improvements

Emissions from controlled burning of planted forest harvesting residues, including those associated with planted forest land converted to grassland (pasture), are not reported in the inventory. Current research is seeking to quantify emissions from this activity for reporting in future submissions (Wakelin, 2008).

The LUCAS plot network is currently being analysed to develop a better estimate of biomass density for the grassland with woody biomass category.

7.2 Forest land (CRF 5A)

7.2.1 Description

In New Zealand’s Initial Report under the Kyoto Protocol (Ministry for the Environment, 2006), national forest definition parameters were specified as required by UNFCCC Decision 16/CMP.1. The New Zealand parameters are a minimum area of 1 hectare, a height of 5 metres and a minimum crown cover of 30 per cent. Where the height and canopy cover parameters are not met at the time of mapping, the land has been classified as forest land where the land-management practice/s and local site conditions (including climate) are such that the forest parameters will be met. New Zealand uses a minimum forest width of 30 metres from canopy-edge to canopy-edge. This removes linear shelterbelts from the forest land-use category. The width and height of linear shelterbelts can vary as they are trimmed and topped from time to time. Further, they form part of non-forest land uses, namely cropland and grassland (as shelter to crops and/or animals).

New Zealand has adopted the definition of managed forest land as provided in GPG-LULUCF: “Forest management is the process of planning and implementing practices for stewardship and use of the forest aimed at fulfilling relevant ecological, economic and social functions of the forest”. Accordingly, all of New Zealand’s forests, both those planted for timber production and natural forests managed for conservation values, are considered managed forests.

For inventory reporting three subcategories are used to cover all of New Zealand’s forests: natural forest (predominantly native forest pre-dating 1990), pre-1990 planted forest and post-1989 forest.

Forest land is the most significant contributor to carbon stock changes in the LULUCF sector. Forests cover 37.7 per cent (around 10 million hectares) of New Zealand. In 2008, forest land contributed 29,757.9 Gg CO2-e of net removals. This value includes removals from the growth of pre-1990 planted forests and post-1989 forests, emissions from the conversion of land to planted forest and emissions from harvesting and deforestation. Net removals from forest land have decreased by 3,098.9 Gg CO2-e (9.4 per cent) over the 1990 level of 32,856.7 Gg CO2-e. In 2008, forest land remaining forest land and conversion to forest land were key categories (trend and level assessment).

Table 7.2.1.1 New Zealand’s land-use change within the forest land category in 1990 and 2008, and associated CO2-e emissions
Forest land land-use category Net area in 1990 (ha) Net area in 2008 (ha) Change from 1990 (%) Net emissions/ removals (Gg CO2-e) Change from 1990 (%)
1990 2008
Forest land remaining forest land Natural forest 8,150,732 8,118,004 –0.4 NE NE NA
Pre-1990 planted forest 1,480,346 1,430,286 –3.4 –33,027.5 –12,430.5 –62.4
  Subtotal 9,631,078 9,548,290 –0.9 –33,027.5 –12,430.5 –62.4
Land converted to forest land Post-1989 forest 13,473 568,775 +4,121.6 166.2 –17,327.7 –10,525.8
Total   9,644,551 10,117,064 +4.9 –32,856.7 –29,757.9 –9.4

Notes: 1990 and 2008 areas are as at 31 December. Net area values include deforestation of post-1989 forest since 1990. Net removals/emission estimates are for the whole year indicated. Natural forest remaining natural forest is assumed at the national level to be at steady state, with no emissions estimated (NE).

Natural forest

Natural forest is the term used to distinguish New Zealand’s native and unplanted (self-sown or naturally regenerated) forests that existed prior to 1990 from pre-1990 planted and post-1989 forests. The category includes both mature forest and areas of regenerating vegetation that have the potential to return to forest under the management regime that existed in 1990. Natural forest ecosystems comprise a range of indigenous and some naturalised exotic species. In New Zealand, two principal types of natural forest exist: beech forests (mainly Nothofagus species) and podocarp/broadleaf forests. In addition, a wide range of seral plant communities fit into the natural forest category if they have the potential to succeed to forest in situ. Currently, New Zealand has an estimated 8.1 million hectares of natural forest (including these successional communities).

In 2008, 0.05 per cent of New Zealand’s total forest timber production was from harvesting of natural forests, as New Zealand’s wood needs are now almost exclusively met from planted production forests (Ministry of Agriculture and Forestry, 2009c). No timber is legally harvested from New Zealand’s publicly owned natural forests (an area approximately 5.5 million hectares in size). Most other harvesting of natural forests is required by law to be undertaken on a sustainable basis. The only natural forest harvesting that is not required by law to be on a sustainable basis is the harvesting of forests on land returned to Māori under the South Island Landless Natives Act 1906. These forests are currently exempt from provisions that apply to all other privately owned natural forests that require a sustainable forest management plan or permit before any harvesting. Approximately 57,500 hectares are covered by the South Island Landless Natives Act 1906.

Harvesting under the sustainable forest plans and permits is restricted to the removal of growth and sometimes takes place on a selective logging basis. This means the area from where trees are extracted still meets the forest definition chosen by New Zealand. Therefore, over the long term, the carbon stored in these forests is in steady state.

In 2008, the carbon stored in natural forest was 2,313,225.1 Gg C. Carbon stock has decreased by 9,843.9 Gg C (0.4 per cent) from the 1990 level of 2,323,069.0 Gg C. This is equivalent to emissions of 36,094.5 Gg CO2-e from natural forest since 1990. This loss is due to natural forest land being converted to other land uses. Carbon stock change in natural forest remaining natural forest is not estimated, as the carbon stored in natural forests is assumed to be in steady state. The emissions associated with the conversion of natural forest to other land uses are reported in the land-use category the land was converted to.

Table 7.2.1.2 New Zealand’s carbon stock change by carbon pool within the natural forest subcategory in 1990 and 2008
Carbon pool Carbon stock in 1990 (Gg C) Carbon stock in 2008 (Gg C) Change since 1990 (%)
Living biomass 1,133,204.4 1,128,402.5 –0.4
Dead organic matter 278,001.9 276,823.9 –0.4
Soil 911,862.7 907,998.7 –0.4
Total 2,323,069.0 2,313,225.1 –0.4

Note: 1990 carbon stock values are as at 31 December 1989 and 2008 values are as at 31 December 2008.

Pre-1990 planted forest

New Zealand has a substantial estate of planted forests created specifically for timber supply purposes. In 2008, pre-1990 planted forests covered an estimated 1.43 million hectares of New Zealand (5.3 per cent of the total land area). New Zealand’s planted forests are intensively managed and there is well-established data on the estate’s extent and characteristics. Having a renewable timber resource has allowed New Zealand to protect and sustainably manage its natural forests. Pinus radiata is the dominant species, making up about 90 per cent of the planted forest area. These forests are usually composed of stands of trees of a single age class and all forests have relatively standard silviculture regimes applied.

In 2008, the carbon stored in pre-1990 planted forest was 383,841.9 Gg C. Carbon stock has increased by 104,896.7 Gg C (37.6 per cent) from the 1990 level of 278,945.1 Gg C. This is equivalent to removals of 384.6 Gg CO2-e.

Table 7.2.1.3 New Zealand’s carbon stock change by carbon pool within the pre‑1990 planted forest subcategory in 1990 and 2008
Carbon pool Carbon stock in 1990 (Gg C) Carbon stock in 2008 (Gg C) Change from 1990 (%)
Living biomass 106,996.4 195,603.4 +82.8
Dead organic matter 17,533.8 39,045.3 +122.7
Soil 154,414.9 149,193.1 –3.4
Total 278,945.1 383,841.9 +37.6

Note: 1990 carbon stock values are as at 31 December 1989 and 2008 values are as at 31 December 2008.

Post-1989 forest

Between the 1 January 1990 and 31 December 2008, the net area of forest established as a result of reforestation activities was 568,775 hectares (taking account of deforestation). Based on the plots measured, 95 per cent of this forest subcategory comprises planted tree species (Paul et al 2009), with the remaining area comprising regenerating native tree species. Pinus radiata comprise 89 per cent of the planted tree species in this forest subcategory, with Douglas fir (Pseudotsuga menziesii) and Eucalyptus species being the two species making up most of the remainder (Paul et al 2009).

The new forest planting rate (land reforested) between 1990 and 2008 was, on average, 31,000 hectares per year (Figure 7.2.1.1). New planting rates were high from 1992 to 1998 (averaging 59,000 hectares per year). Since 1998, the rate of new planting rapidly declined and is now at very low levels. In 2008, it was estimated that 1,000 hectares of new forest was established (Ministry of Agriculture and Forestry, 2009a).

Figure 7.2.1.1 Annual areas of reforestation in New Zealand from 1990 to 2008

Figure 7.2.1.1 Annual areas of reforestation in New Zealand from 1990 to 2008

Figure 7.2.1.1 illustrates the annual areas of reforestation (new forest planting) in New Zealand from 1990 to 2008. The data used to produce the graph are provided in the table below.

Year Annual areas of reforestation in New Zealand from 1990 to 2008 (kha)
1990 13.473
1991 13.131
1992 42.805
1993 52.526
1994 83.747
1995 62.911
1996 71.439
1997 54.377
1998 43.579
1999 33.951
2000 28.458
2001 25.536
2002 18.693
2003 16.823
2004 8.895
2005 5.030
2006 2.180
2007 1.969
2008 1.000

Note: Annual planting estimates are derived from annual surveys of forest nurseries, as published in the National Exotic Forest Description (Ministry of Agriculture and Forestry, 2009a) and have been scaled downwards using a ratio derived from the LUCAS mapping of post-1989 forest area.

The area of new planting is expected to increase with the introduction of the Emissions Trading Scheme, Permanent Forest Sinks Initiative and Afforestation Grant Scheme that have been introduced since 2007 by the New Zealand Government to encourage new planting and regeneration of natural species (Ministry of Agriculture and Forestry, 2009b).

In 2008, the carbon stored in post-1989 forest was 116,849.3 Gg C (see Table 7.2.1.4). Carbon stock as at the start of 1990 was nil by definition. The growth in carbon stock since 1990 is equivalent to removals of 428,447.4 Gg CO2-e.

Table 7.2.1.4 New Zealand’s carbon stock change by carbon pool within the post‑1989 forest subcategory
Carbon pool Carbon stock in 1990 (Gg C) Carbon stock in 2008 (Gg C) Change since 1990 (%)
Living biomass 0.0 40,947.8 NA
Dead organic matter 0.0 14,189.3 NA
Soil 0.0 61,712.2 NA
Total 0.0 116,849.3 NA

Note: 1990 carbon stock values are as at 31 December 1989 and 2008 values are as at 31 December 2008. Post-1989 forest is land that has been afforested or reforested since 31 December 1989, and hence had approximately nil carbon stock on 1 January 1990.

The trend in removals is shown in Figure 7.2.1.2. This graph shows that the post-1989 forests do not become a net sink until 1995. This is due to the emissions from loss of biomass carbon stocks associated with the previous land use and the change (loss) of soil carbon with a land-use change to forestry (see Table 7.1.2.10), outweighing removals by forest growth.

Figure 7.2.1.2 New Zealand’s net removals by post-1989 forests from 1990 to 2008

Figure 7.2.1.2 New Zealand’s net removals by post-1989 forests from 1990 to 2008

Year Net removals
(Gg CO2)
1990 -166.2
1991 -166.0
1992 -480.0
1993 -387.2
1994 -292.6
1995 966.2
1996 2,552.5
1997 5,104.9
1998 7,865.9
1999 10,688.7
2000 13,063.9
2001 14,827.3
2002 15,950.0
2003 16,541.3
2004 16,887.2
2005 17,033.0
2006 17,074.1
2007 17,200.3
2008 17,327.7
Deforestation

In 2008, 4,818 hectares of forest land were converted to other land uses, primarily grassland. Table 7.2.1.5 below shows the areas of forest land subject to deforestation for conversion to other land uses in 2008, and since 1990.

Table 7.2.1.5 New Zealand’s forest land subject to deforestation
Forest land subcategory Area of forest in 1990 (hectares) Deforestation since 1990 Deforestation in 2008
Area (hectares) Proportion of 1990 area (%) Area (hectares) Proportion of 1990 area (%)
Natural forest 8,152,550 34,546 0.42 1,818 0.02
Pre-1990 planted forest 1,480,345 50,059 3.38 2,114 0.14
Post-1989 forest 0 11,749 NA 886 NA
Total 9,632,895 96,355 1.00 4,818 0.05

Notes: 2008 areas as at 31 December 2008, 1990 areas as at 1 January 1990, and therefore differ from 1990 area values in the common reporting format tables, which are at 31 December 1990. Post-1989 forest comprises land that has been afforested or reforested since 31 December 1989, and hence its area was nil by definition on 1 January 1990.

The conversion of forest land to grassland was most likely due to the relative profitability of some forms of pastoral farming (particularly dairy farming) compared with forestry. Figures 11.1.2 and 11.1.3 in chapter 11 show the area and net emissions respectively associated with deforestation between 1990 and 2008. These show the significant increase in deforestation of planted forests since 2003, with a significant decrease in 2008.

During the period 2008–2012, it is expected that the level of planted forest deforestation will be less than previous years (Manley, 2009). This is because policy measures are expected to be a disincentive to deforest.

The estimate of natural forest deforestation in 2008, which is based on previous trends, is also likely to be an overestimate. This is because land-use change during the 2008 calendar year was estimated by linear interpolation from the average land-use change mapped between 1 January 1990 and 1 January 2008. As there was no quantitative information on the annual rate of natural forest deforestation between 1990 and 2007, the same annual rate of change was assumed for the entire period (1,818 hectares per year), and extrapolated out to the end of 2008.

However, a number of factors suggest that the rate of natural forest deforestation is unlikely to have been constant over the 19-year period, but instead mostly occurred prior to 2002. The area available for harvesting (and potentially deforestation) was higher before amendments were made to the Forests Act 1949 in 1993. Further restrictions to the logging of natural forests were also introduced in 2002, resulting in the cessation of logging of publicly owned forests on the West Coast of New Zealand in 2002. Both of these developments are likely to have reduced natural forest deforestation since 2002.

The extrapolated estimate of natural forest deforestation will be updated in future submissions as new information becomes available, and will be replaced with an actual, mapped value in the 2013 submission at the latest, following production of the 2012 land-use map.

New Zealand assumes instant emissions of all biomass carbon at the time of deforestation, and soil carbon changes are modelled over a 20-year time period (refer to the previous section on soils). This approach is adopted because:

  • the majority of deforestation is from land converted to high-producing grassland, resulting in the rapid removal of all biomass, as the land is prepared for intensive dairy farming (see Figure 7.1.2.5)
  • it is not practical to estimate the volume of residues left on site after the deforestation activity, given the rapid conversion from one land use to another. Further estimating any residue biomass carbon pools and decay rates is difficult and costly
  • there is insufficient data prior to 2008 to estimate deforestation biomass residue coming into the first commitment period. If a different approach was adopted for deforestation before and after 2008 this might not meet GPG-LULUCF.

These deforestation emissions are reported in the relevant ‘land converted to’ category, as are all emissions from land-use change. See section 11.1 of chapter 11 for further information on deforestation.

7.2.2 Methodological issues

Forest land remaining forest land

Only natural forest and pre-1990 planted forest are described in this section because land in the post-1989 forest subcategory is included in the land converted to forest land category. Land transfers to the ‘land remaining’ category after having been in that land use for 28 years. New Zealand has chosen 28 years as the time period taken for land to reach a state of equilibrium (or maturity), as this is the average age that planted forests are harvested (Ministry of Agriculture and Forestry, 2008a). Lands categorised as post-1989 forest were 18 years old (at a maximum) in 2008.

New Zealand has established a sampling framework for forest inventory purposes based on a grid system established across the country. The grid has a randomly selected origin and provides an unbiased framework for establishing plots for field and/or LiDAR measurements. The grid is an 8 kilometre grid with divisions on a 4 kilometre grid being used for measurement of post-1989 forest areas.

Natural forest

A national monitoring programme to enable unbiased estimates of carbon stock and change for New Zealand’s natural forests was developed between 1998 and 2001 (see Goulding et al, 2001). There were 1,256 permanent plots installed systematically on the 8-kilometre grid across New Zealand’s natural forests and first measured between 2002 and 2007.

The plots were sampled using a method designed specifically for the purpose of calculating carbon stocks (Payton and Moss, 2001; Payton et al, 2004). As the plot network is remeasured, the data collected will be suitable for determining if New Zealand’s natural forests are carbon neutral, or whether they are a source or a sink of carbon. Where possible, the network incorporated plots that had been established previously and remeasured during the establishment phase of the national network to enable an initial assessment to be made of forest changes over time. Figure 7.2.2.1 shows the distribution of the carbon monitoring plots throughout New Zealand.

Figure 7.2.2.1 Location of New Zealand’s natural forest carbon monitoring plots

Figure 7.2.2.1 Location of New Zealand’s natural forest carbon monitoring plots

Figure 7.2.2.1 illustrates the locations of New Zealand’s natural forest carbon monitoring plots. There were 1,256 permanent plots installed systematically on the 8-kilometre grid across New Zealand’s natural forest. The map shows the plots distributed across the North Island, South Island and Stewart Island, where natural forest is occurring.

Remeasurement of the national plot network has begun. The remeasurement programme will run from 2009–2013. Once field work has been completed and the data has been quality assured and analysed, national carbon estimates will be updated in time for the 2012 submission (to be submitted in 2014).

At each plot, data is collected to calculate the volumes of trees, shrubs and dead organic matter present. These measurements are then used to estimate the carbon stocks for the following biomass pools:

  • living biomass (comprising above-ground biomass and below-ground biomass)
  • dead organic matter (comprising dead wood and litter).

Table 7.2.2.1 summarises the method used to calculate the carbon stock in each biomass pool from the information collected at each plot.

Table 7.2.2.1 Summary of methods used to calculate New Zealand’s biomass carbon stock from plot data
Pool Method Source
Living biomass Above-ground biomass Allometric equations (Beets et al, 2008b) Plot measurements; method (Beets et al, 2009)
Below-ground biomass Assumed to be 20 per cent of total biomass Coomes et al, 2002
Dead organic matter Dead wood Allometric equations (Beets et al, 2008a; Beets et al, 2008b) Plot measurements; method (Beets et al, 2009)
Litter Laboratory analysis of samples collected at plots.(Garrett et al, 2009) Plot measurements; method (Beets et al, 2009)
Living biomass

Living biomass is separated into two pools.

  • Above-ground biomass. The carbon content of individual trees and shrubs are calculated using species-specific allometric relationships between diameter, height and wood density (for trees), a non-specific conversion factor with diameter and height (for tree ferns), or volume and biomass (for shrubs). Shrub volumes are converted to carbon stocks using species and/or site-specific conversion factors, determined from the destructive harvesting of reference samples.
  • Below-ground biomass. Below-ground biomass is derived from above-ground biomass and is assumed to be 25 per cent of above-ground biomass (or 20 per cent of total biomass). This value is based on a few studies that report root to total biomass ratios of 9 to 33 per cent (discussed in Coomes et al, 2002). Coomes et al (2002) acknowledge more work is needed but use the average of the cited studies to justify allocating 20 per cent of total biomass to below-ground biomass.
Dead organic matter

Dead organic matter is separated into two pools.

  • Dead wood. The carbon content of dead standing trees is determined in the same way as live trees, but excludes branch and foliage biomass calculations. The carbon content of the fallen wood and stumps is derived from the volume of the piece of wood, species, if able to be identified, and what stage of decay it is at. Dead wood comprises woody debris with a diameter > 10 cm.
  • Litter. The carbon content of the fine debris is calculated by laboratory analysis of sampled material. Litter comprises fine woody debris (FWD) (dead wood from 2.5 to 10 cm diameter) and the litter (all material <2.5 cm diameter) and the fermented humic (FH) horizons. Samples were taken at approximately one-third of the natural forest plots.

Carbon stocks in New Zealand’s natural forests (excluding the soils pool) of 173 (±6) t C ha1 were estimated from the first full round of measurements (Beets et al, 2009) and those data are used for this report. The subset of plots for historic data that exist were separately analysed to estimate the change. Thirteen per cent of the natural forest LUCAS plots were used in the analysis, which found that natural forests in New Zealand were a carbon sink between 1990 and 2004 (Beets et al, 2009). Until the entire plot network has been remeasured, New Zealand will continue to report natural forests remaining natural forests as carbon neutral and therefore no removals or emissions are estimated in this submission.

Soil carbon

Soil carbon stocks in natural forest land remaining natural forest land are estimated using a Tier 2 method that uses New Zealand-specific land use and soil pedon data, as described in section 7.1.2. The soil carbon stock at equilibrium state is estimated to be 111.85 t C ha1 with a standard error of 5.24 (Table 7.1.2.10).

Natural forest carbon

Total carbon stocks in natural forest are determined by combining the biomass and soil carbon pools (of 173 and 111.85 respectively) to give a mean of 284.85 t C ha1 (±6.0 per cent). The mean value is then multiplied by the area of natural forest land remaining natural forest land to give a national total.

Pre-1990 planted forests

Living biomass and dead organic matter

New Zealand uses a Tier 2 method to estimate biomass carbon for pre-1990 planted forest. This involves:

  • data from the annual National Exotic Forest Description surveys
  • stem wood volume yield tables, compiled periodically for combinations of species, silvicultural regime and location
  • the C_Change model (Beets et al, 1999), which is used to derive forest biomass and carbon by pool from stem volume yield tables.

This method is essentially the same as used in the 2009 submission. The process is illustrated in Figure 7.2.2.2. The differences between the method used in the 2009 submission and this submission are as follows.

  • The age-based carbon yield table (derived from the National Exotic Forest Description and attributed to biomass pool using C_Change) is used in the LUCAS Calculation and Reporting Application, and not within the FOLPI modelling system (a forest estate modelling package for forest management planning).
  • In the 2009 submission, harvesting emissions were estimated as 85 per cent of the stem biomass, compared with 70 per cent of live biomass in this submission (the results of this change are minor). Future research is planned to verify this estimate.
  • In the 2009 submission, first and second rotation forests were treated separately but reported together. In this submission, all pre-1990 planted forests are treated the same (regardless of rotation number) but post-1989 forests are separated.

Inputs to the C_Change model include the National Exotic Forest Description stem volume yield tables, wood density classes for regions and species, and silvicultural regime details. The C_Change model is used to:

  • derive stem wood biomass increment from volume increment and wood density
  • apply an increment expansion factor to convert this to total biomass fixed
  • partition the total biomass to live biomass pools
  • calculate transfers from live to dead pools from mortality functions and regime details (ie, pruning/thinning)
  • apply decay functions to estimate dry-matter loss from dead pools.

Figure 7.2.2.2 New Zealand’s pre-1990 planted forest inventory modelling process

Figure 7.2.2.2 New Zealand’s pre-1990 planted forest inventory modelling process

Figure 7.2.2.2 illustrates New Zealand’s pre-1990 planted forest inventory modelling process. This process comprises the C_Change model, which is depicted as feeding into the national average yield table (by carbon pool), which in turn feeds into the LUCAS data storage and collection Gateway, which then feeds into the LUCAS Calculation and Reporting Application (CRA) for the ‘joint calculations’ of emissions and removals estimates.

The C_Change model, which is used to derive forest biomass and carbon by pool from stem volume yield tables, is depicted as incorporating data on: mortality thinning and pruning; annual biomass increment allocation; dead pools and decay. There are three main inputs to the C_Change model: stem volume yield tables by crop type; wood density class; and standardised national regimes.

National area figures, derived from the land-use matrix, are depicted as the input to the Gateway.

Wood density is an important variable in the estimation of carbon levels. It is related to the individual effects of temperature, stand age, nitrogen fertility and management factors (Beets et al, 2007a). Temperature and stand age have the greatest influence on wood density, followed by site fertility and stocking. The influence of the individual effects on wood density is provided in Table 7.2.2.2. The combined result of these individual effects can be large. For example, the 15-year growth sheath of a stand of standard genetics Pinus radiata, at a low stocking (200 stems ha–1) on a fertile (C/N=12), cool (8°C) site has a predicted wood density of 339 kg m-3, while a stand of the same age and genetics at a high stocking (500 stems ha–1) on a moderately fertile (C/N=25), warm (16°C) site has a predicted wood density of 467 kg m–3.

Table 7.2.2.2 Influence of individual site and management factors on predicted wood density for New Zealand
Factor affecting wood density Range in predicted density
(kg m–3) (% difference)
Temperature: 8°C versus 16°C 359–439 22
Age: 10 year old versus 30 year old 380–446 17
C/N ratio: 12 versus 25 384–418 9
Stocking: 200 versus 500 stems ha–1 395–411 4

The output from C_Change is a dry-matter yield table with estimates of dry matter per hectare by age class for each component. These are aggregated into the biomass carbon pools and converted to carbon using a carbon fraction of 0.5 (default value given in GPG-LULUCF, IPCC, 2003). These are then combined with the activity data (area of forest by age class) to estimate carbon stock by pool.

There are losses in this category of forest land remaining forest land associated with forest harvesting. The proportion of total stem volume removed at harvest varies, but is generally around 85 per cent and is equivalent to about 70 per cent of above-ground live biomass at typical rotation ages (Wakelin, 2008). This is treated as an instantaneous carbon loss. These losses are estimated as 70 per cent of above-ground live biomass, with the remaining 30 per cent of above-ground biomass being transferred to the dead organic matter pool and decayed in a linear manner over 20 years; however, work is currently under way to confirm these estimates.

Soil carbon

Soil carbon stock in pre-1990 planted forest land remaining pre-1990 planted forest land is estimated using a Tier 2 method as described in section 7.1.2 – Soils. The soil carbon stock in pre-1990 planted forests at equilibrium state is estimated to be 104.31 t C ha–1, with a standard error of ±6.44 (Table 7.1.2.10). Soil carbon change with harvesting is not explicitly estimated, as the long-term soil carbon stock for this land use includes any emissions associated with harvesting.

Non-CO2 emissions – Forest land remaining forest land

Non-CO2 emissions from drainage of soils and wetlands

New Zealand has not prepared estimates for this category as allowed for in the IPCC good practice guidance for LULUCF chapter 1.7.

Biomass burning

There are no emissions reported for controlled burning in forest land remaining forest land in New Zealand as this practice is not common and there is no activity data on this (Wakelin et al, 2009). The inventory reports only emissions resulting from wildfire for this category, and reports the notation key NE (‘not estimated’) for emissions from controlled burning in the common reporting format tables. New Zealand estimates emissions from wildfire using:

  • the IPCC default temperate forest fuel consumption rate of 45 per cent of total biomass (GPG-LULUCF Table 3A.1.12, IPCC, 2003)
  • wildfire activity data for April 1991 to March 2009. This data is collected and managed by the New Zealand Fire Service and the National Rural Fire Authority. The average over the period is then applied back to earlier years where no data is available. Activity data for wildfire is generally poor quality, but it is believed that there have not been major changes in wildfire occurrence since 1990 (N Challands, New Zealand Fire Service, pers comm; Wakelin et al, 2009).

Carbon dioxide emissions from wildfire in planted forest are captured by the stock change calculation at the time of harvest, as there is no reduction in carbon stock for areas burnt prior to harvesting or deforestation. Therefore, carbon dioxide emissions may be underestimated or overestimated using this approach. However, the total area of wildfires in planted forest is small and this is not regarded as a significant source of error.

Land converted to forest land

All land converted to forest land either by planting or as a result of human-induced changes in land-management practice (eg, removing grazing stock and allowing revegetation of tree species) since 1990 is included in the subcategory post-1989 forests.

Post-1989 forests

Survey data

As the majority of post-1989 forests are privately owned, field access to the forests has not been guaranteed and a double-sampling approach involving airborne scanning LiDAR (with digital aerial photography) and ground-based measurements has been used (Stephens et al, 2007; Stephens et al, 2008; Beets et al, 2010). This approach has allowed corrections to be made for unexpected loss of access to some field plots while simultaneously improving precision.

The double-sampling approach being used by New Zealand follows that described by Parker and Evans (2004) and Corona and Fattorini (2008), where LiDAR and conventional plot field measurements are used for the forest inventory. Double sampling (or two-phase sampling) involves field and LiDAR data of permanent sample plots (PSPs). All plots are sampled by LiDAR and a sample of the plots are measured in the field. A multiple linear regression between derived LiDAR metrics and plot carbon (Stephens et al, 2007; Kimberley et al, 2009) is then established. Standard double-sampling regression estimator procedures are then used to obtain an estimate of the average carbon stock per hectare and carbon in each biomass pool for the post-1989 forest estate with known precision (Kimberley et al, 2009).

The steps used in the inventory of post-1989 forests and to determine the average level of carbon per hectare for the PSPs are briefly described and are shown in Figure 7.2.2.3. The key steps involved in determining carbon per pool for post-1989 forests include the following.

  • Identification of the plots on the 4-kilometre grid that fell in post-1989 forest and seeking approval for field teams to access land (see Figure 7.2.2.4).
  • Training field and audit teams in the use of the post-1989 forest data collection manual (Payton et al, 2008) and in the use of the PSP data storing and checking software used in hand-held instruments. At each 4-kilometre grid point where a field plot was measured, four circular plots were established. The centre plot was 0.06 hectares in area and the other three were 0.04 hectares in area. While plots larger than 0.04 hectares were shown not to decrease the variance between plots (Moore and Goulding 2005), the 0.06 hectare plot was chosen for the central plot because this was deemed optimal for use with LiDAR.
  • Making and storing tree, dead wood, litter and soil fertility measurements on plots. Soil fertility measurements are made to assist in predicting wood density (Beets et al, 2007a).
  • Acquiring airborne scanning LiDAR and digital aerial photography measurements made of plots on a 4-kilometre grid (Stephens et al, 2007; Stephens et al, 2008). LiDAR data was acquired for at least 3 points m2 and the photography had a spatial resolution of 20 centimetres. The LiDAR and photography swath width was 170 metres.
  • Estimating total carbon per plot and per biomass pool using the ‘Forest Carbon Predictor’ (FCP) model (Kimberley and Beets, 2008). More information on this modelling system is provided in the Modelling and LiDAR double-sampling section.
  • Deriving LiDAR metrics per plot (vegetation height percentile, crown volume and canopy skewness – Stephens et al, 2007) and then determining multiple linear regression between metrics and carbon estimates (Kimberley et al, 2009).
  • Determining the average carbon content (t C ha1) for all plots on the 4-kilometre grid, using a LiDAR-based double-sampling regression estimator. This value provides the carbon stock, as at 1 January 2008, for post-1989 forests (Kimberley et al, 2009).

Figure 7.2.2.3 New Zealand’s approach used to inventory post-1989 forests and estimate the average carbon stock per pool for the plots within the forest

Figure 7.2.2.3 New Zealand’s approach used to inventory post-1989 forests and estimate the average carbon stock per pool for the plots within the forest

Figure 7.2.2.3 illustrates New Zealand’s approach used to inventory post-1989 forests and estimate the average carbon stock per pool for the plots within the forest, as described in section 7.2.2 of the Inventory report under Post-1989 forests – Survey data.

Figure 7.2.2.4 Location of New Zealand’s pre-1990 planted forest and post-1989 forest plots

Figure 7.2.2.4 Location of New Zealand’s pre-1990 planted forest and post-1989 forest plots

Figure 7.2.2.4 shows a map of New Zealand showing the location of New Zealand’s pre-1990 planted forest and post-1989 forest plots.

Quality assurance and quality control

Quality-assurance and quality-control activities were conducted throughout the post-1989 forest data capture and processing steps. These quality-assurance and quality-control activities are indicated in Figure 7.2.2.3, and were associated with the following: acquisition of raw LiDAR data and LiDAR processing; checking eligibility of plots; audits of field plot measurements; data processing and modelling; and regression analysis and double-sampling procedures (Brack, 2009). These activities are described more fully in section 7.2.4.

Modelling and LiDAR double sampling

The plot data collected was modelled using a forest carbon modelling system called ‘Forest Carbon Predictor’, version 2.2 (FCPv2.2) (Kimberley and Beets, 2008). This integrates the 300 Index Growth model (Kimberley et al, 2005), a wood density model (Beets et al, 2007a), and the C_Change model (Beets et al, 1999), to enable predictions of carbon stocks and changes in New Zealand’s planted forests. The 300 Index Growth model for radiata pine (Pinus radiata) generates a productivity index for a plot from the stand age, mean top height, basal area, stocking, and stand silvicultural history that it uses to predict gross and net stem volume under bark over a full rotation. The stem volume increment predictions are combined with estimates of the density of stem wood annual growth sheaths, which C_Change uses to estimate carbon stocks in four pools: above-ground biomass, below-ground biomass, dead wood and litter, taking into account natural and management-induced mortality and decay rates.

The use of the forest carbon modelling system enabled the production of plot-based estimates of carbon per plot at a specific date, and an area-weighted and age-based carbon yield table. Paul and Kimberley (2009) have demonstrated that using the FCPv2.2 for all planted forest tree species produces an average t C ha–1 value little different to using more specific carbon models/allometric equations for the non-radiata species (mainly Douglas fir (Pseudotsuga menziesii) and eucalypts (Eucalyptus spp.)). They established that there was a marginal decrease (0.77 t C ha–1) in the average amount of carbon removals per plot using the model for all planted forest species.

Good relationships were found between carbon pools estimated using ground-based tree measurements and carbon modelling using FCPv2.2 with airborne scanning LiDAR metrics for the post-1989 forests. The best fitting LiDAR metric for predicting total carbon was a height metric (the 30 per cent height percentile), but significant variation was also explained by a canopy cover metric (namely, per cent cover). A regression model explaining 74 per cent of the variation in total carbon was developed using these two LiDAR metrics. Beets et al (In press (b)) established strong relationships between LiDAR data and ground-based measurements of leaf area index, biomass carbon stocks, and annual carbon sequestration in radiata pine plots selected across a range of micro-sites differing in mean height and basal area. In this study, involving 36 plots independent of either post-1989 or pre-1990 planted forest plots, LiDAR metrics explained between 80–97 per cent of the variation in cumulative leaf area index, depending on the canopy depth examined, and the LiDAR data also explained 86 per cent of the variation in above-ground biomass carbon.

Regression models using the same model form were also fitted for each of the four biomass pools, providing good predictions for above-ground biomass carbon (R2=81 per cent and below-ground biomass carbon (R2=80 per cent), but less successful predictions for litter carbon (R2=38 per cent) and dead wood carbon (R2=21 per cent) (Kimberley et al, 2009). The R2 for a regression between the best LiDAR metric, 95th height percentile, and mean top height calculated from ground measurement was 96 per cent, with a root mean square error of 1.09 metres. Given this relationship, it has been assumed that the LiDAR and ground data have been well co-located.

These regression models were used to obtain estimates, as at 1 January 2008, of the national level of carbon stock in the post-1989 forests using double-sampling procedures, and to develop a national age-based and area-weighted carbon yield table for the resource. Carbon estimates from 246 ground plots were supplemented with LiDAR data from 46 additional plots. The regression estimators (using the LiDAR data) improved precision by 6 per cent compared with the ground-based estimates. The carbon stock estimate from using LiDAR and double sampling is 88.21 ± 2.76 t C ha–1 (at the 95 per cent confidence interval) and the comparable value from just the field measured plot data is 88.46 ± 2.94 t C ha–1 (Kimberley et al, 2009). This carbon stock estimate, while high, is consistent with the international comparisons provided in Table 3A.1.4 (GPG-LULUCF, IPCC, 2003), and reflects the composition of the forest of 95 per cent actively managed production forestry. The average age of post-1989 forest trees as at 1 January 2008 is 12 years. This submission’s yield table has been compared with the yield table used in the last submission (see section 7.2.5 for further details).

Living biomass

The living biomass pool is separated into two pools.

  1. Above-ground biomass. The carbon content of plantation crop trees and shrubs under crop trees is estimated using the FCPv2.2 model. For the other shrubs and non-crop tree species, the carbon content is estimated using species-specific allometric equations that enable carbon to be determined from diameter, height, wood density (for trees) and canopy cover measures (Paul et al, 2009). When non-forest land is converted to forest land, all living biomass that was present at the time of forest establishment is instantly emitted as part of the forest land preparation. Between 1990 and 2008, approximately 30 per cent of the non-forest land converted to post-1989 forest has been from grassland with woody biomass, and this land-use subcategory provides the largest source of emissions associated with land-use change to forestry.
  2. Below-ground biomass. This is derived from the above-ground biomass estimates. For plantation crop trees, the above- to below-ground biomass ratio is 0.2 (Beets et al, 2007b). The ratio for non-crop trees and shrubs is 0.25 (Coomes et al, 2002).
Dead organic matter

The dead organic matter carbon pools are separated into two pools.

  1. Dead wood. The carbon content of the dead wood pool is estimated using the FCPv2.2 model. Immediately following harvesting, 30 per cent of the above-ground biomass pool is transferred to the dead wood pool, with the other 70 per cent being instantaneously emitted. All material in this pool is decayed over a 20‑year period.
  2. Litter. The carbon content of the litter pool in post-1989 forests is estimated using the FCPv2.2 model.
Soil carbon

Soil carbon stocks in land converted to post-1989 forest are estimated using a Tier 2 method as described in section 7.1.2. There is not a specific post-1989 forest value as the number of data points in this land type is small and, as this forest is only 18 years old (at a maximum), it is not considered to have reached steady state. As 95 per cent of post-1989 forests are planted forest, the planted forest value has been used instead.

In the absence of country- and land-use specific data on the time rate of change, the IPCC default method of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original land use and planted forest land for any given period. For example, the soil carbon change associated with a land-use change from low-producing grassland (soil carbon stock 117.66 t C ha–1) to planted forest (soil carbon stock 104.31 t C ha–1), would be a loss of 13.35 t C ha–1 over the 20-year period.

Non-carbon dioxide emissions

Direct N2O emissions from nitrogen fertilisation of forest land and other

Nitrous oxide emissions from nitrogen fertilisation are covered in the agriculture sector under the agricultural soils category.

Biomass burning

It is estimated that 25 per cent of the grassland converted to forest land is cleared using controlled burning. A country-specific fuel consumption rate of 70 per cent of above-ground biomass (Wakelin et al, 2009) is used to estimate emissions from controlled burning. The remainder (30 per cent of above-ground biomass) and all biomass on unburned sites are assumed to decay over 20 years (IPCC default value, GPG-LULUCF Table 3.4.9, IPCC, 2003).

Emissions of carbon dioxide from controlled burns for afforestation are reported as a stock change in the grassland category. All non-CO2 emissions from wildfires in land converted to forest land are reported under the forest land remaining forest land category, as the annual area involved is relatively small and the activity data does not distinguish between the two forest land categories. Carbon dioxide emissions resulting from wildfire events are not reported, as the methods applied do not capture subsequent regrowth (GPG-LULUCF, section 3.2.1.4.2, IPCC, 2003).

7.2.3 Uncertainties and time-series consistency

Removals from forest land are 3.0 per cent of New Zealand’s total emissions and removals uncertainty in 2008 (Annex 7). Forest land introduces 2.6 per cent uncertainty into the trend in the national total from 1990 to 2008. This is the second largest impact on the trend following CO2 emissions from the energy sector.

Natural Forest

The uncertainty in mapping natural forest is ±4 per cent.

The natural forest plot network provides carbon stock estimates that are within 95 percent confidence intervals of 3.67 percent of the mean (±8.0 t/ha) in forests and 15.0 per cent of the mean (±8.6 t/ha) in shrublands (Beets et al, 2009).

No uncertainty estimates are currently available for emissions from harvesting of natural forests.

Table 7.2.3.1 Uncertainty in New Zealand’s 2008 estimates from natural forest
Variable Uncertainty at a 95% confidence interval (%)
Activity data uncertainty  
    Uncertainty in land area ±5.7
Emission factor uncertainty  
    Uncertainty in biomass carbon stocks ±3.7
    Uncertainty in soil carbon stocks ±4.7
    Uncertainty in liming emissions NO
    Combined emission factor uncertainty ±3.7
Implied uncertainty in N2O emissions ±6.0
Combined uncertainty ± 6.8

Notes: Lime application to natural forest does not occur (NO) in New Zealand. Nitrous oxide emissions are calculated as a proportion of carbon stock changes with the same uncertainty as for CO2, and therefore do not add to the combined uncertainty value. The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

Pre-1990 planted forest

Adopting a Tier 2 modelling approach has allowed the large body of plantation forestry knowledge in New Zealand to be applied to the Greenhouse Gas Inventory. For example, the wood density and yield table information as outlined in section 7.2.2.

Attempts have been made to quantify the uncertainties in the carbon dioxide removal estimates for planted forests but it is difficult to quantify the overall error due to the assumptions implicit in the models. Combining the uncertainties indicates that the proportional error in the carbon sequestration estimates is likely to be at least ±16.0 per cent. The total national planted area from the National Exotic Forest Description is considered to be accurate to within ±5.0 per cent (Ministry of Agriculture and Forestry, 2009a) and the National Exotic Forest Description yield tables are assumed to be accurate to within ±5.0 per cent.

A sensitivity analysis was conducted using the above accuracy ranges for total planted area and commercial yield, and a proportional uncertainty error of ±16.0 per cent. The C_Change model runs indicate that the precision of the carbon stock estimates could be of the order of ±25.0 per cent.

Table 7.2.3.2 Uncertainty in New Zealand’s 2008 estimates from pre-1990 planted forest
Variable Uncertainty at a 95% confidence interval (%)
Activity data uncertainty  
  Uncertainty in land area ± 9.9
Emission factor uncertainty  
  Uncertainty in biomass accumulation rates ±16.9 based on:
       C_Change model: wood density     ± 3.0
       C_Change model: carbon allocation     ± 15.0
       C_Change model: carbon content     ± 5.0
       NEFD yield table     ± 5.0
  Uncertainty in soil carbon stocks ±6.2
  Uncertainty in liming emissions NO
  Combined emission factor uncertainty ± 16.9
Implied uncertainty in N2O emissions ±17.9
Total combined uncertainty ± 19.5

Notes: NEFD is the National Exotic Forest Description (Ministry of Agriculture and Forestry, 2009a). Lime application to pre-1990 planted forest does not occur (NO) in New Zealand. Nitrous oxide emissions are calculated as a proportion of carbon stock changes with the same uncertainty as for CO2, and therefore do not add to the combined uncertainty value. The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

Post-1989 forest

Biomass

The models within the FCPv2.2 forest carbon modelling system that have been used for post-1989 forests have been individually validated.

  • The 300 Index Growth model has been extensively tested, although most of this work is so far unpublished (Mark Kimberley, Scion, pers comm). An unpublished industry report on validation of this model (Kimberley and Dean, 2005) is summarised in Kimberley et al (2005). The 300 Index Growth model is being used in New Zealand forest industry applications (MacLaren and Knowles, 2005; Palmer et al (In press)).
  • Validation of the C_Change model has been studied by Beets et al (1999). This work showed that above-ground stand carbon was highly correlated with that predicted by C_Change (r2=0.97, n=25, p<0.01). In this study, the understory and forest floor carbon were excluded. Uncertainties within the C_Change model have also been described by Hollinger et al (1993). These include ±3.0 per cent for wood density, ±15.0 per cent for carbon allocation and ±5.0 per cent for carbon content.

Beets et al (In press (a)) tests the empirical accuracy and precision of carbon stock and change estimates and predictions derived using the Forest Carbon Predictor modelling system, using independent biomass data acquired from a range of sites in New Zealand. Carbon stocks from biomass measurements were compared with carbon stock estimates and predictions obtained using FCPv2.2. The overall model error was calculated by subtracting the carbon predictions from the biomass estimates, and therefore the error includes both model error and biomass estimation error. The error averaged –1.2 per cent for stem volume, 6.3 per cent for above-ground biomass carbon and 10.3 per cent for total carbon (excluding roots and mineral soil C) based on plot measurements obtained in the same year the biomass study took place. The prediction error, based on plot data acquired five years before or five years after the biomass study, was 5.0 per cent (not significant) greater than the estimation error obtained using the reference year.

Another potential source of error and bias can occur if some grid intersections located within the mapped forest are not sampled, either by LiDAR or field measurements. New Zealand has operational measures in place to ensure that this source of error is addressed, such as continually validating the plot network to ensure that relevant plots are included in analyses.

Soils

Ninety-nine per cent of land converted to post-1989 forest land is from grassland. There has been paired-site validation of the ability of the Soil CMS to predict soil carbon changes between grassland and planted forest land. There was reasonable agreement between modelled estimates and observed data for the 0–0.1 metre soil depth increment, but significant differences for the 0.1–0.3 metre increment. This is due partly to a lack of observed data for the 0.1–0.3 metre increment, as well as a greater emphasis in the observed data on species other than Pinus radiata (the dominant planted forest species in New Zealand). Results indicate that, once a weighting for forest species type had been applied (to remove potential bias in the paired-site dataset because Pinus radiata was under-represented), the Soil CMS model and paired-site predictions of mean soil carbon are in agreement within 95 per cent confidence intervals (Baisden et al, 2006a, b).

In addition, the Soil CMS has recently been modified to remove the effect of bias from spatial clustering of soil pedon data points. As the dataset used by the model consists primarily of historical data collected for specific purposes, it is not a random sample of the soils in New Zealand, with some soil/climate/land-use combinations over-represented and some under-represented. As soil samples are correlated to some extent according to the distance between them, the incorporation of a correction factor for spatial correlation into the Soil CMS model has resulted in a decrease in difference between stock estimates for grassland and planted forest, and improved agreement between the modelled estimates and the paired site observed data.

Table 7.2.3.3 Uncertainty in New Zealand’s 2008 estimates from post-1989 forest
Variable Uncertainty at a 95% confidence interval (%)
Activity data uncertainty  
  Uncertainty in land area ± 6.5
Emission factor uncertainty  
  Uncertainty in biomass accumulation rates ±11.9 based on:
      Modelling     ± 10.3
      Sampling     ± 5.9
      Forecasting     ± 1.0
  Uncertainty in soil carbon stocks ±6.2
  Uncertainty in liming emissions NO
  Combined emission factor uncertainty ± 10.1
Implied uncertainty in N2O emissions ±13.4
Total combined uncertainty ± 12.0

Notes: Lime application to post-1990 forest does not occur (NO) in New Zealand. Nitrous oxide emissions from lime application on deforested land is calculated as a proportion of carbon stock change, with the same uncertainty as for CO2, and therefore it does not add to the combined uncertainty value. The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.2.4 Category-specific QA/QC and verification

Carbon dioxide removals from both ‘forest land remaining forest land’ and ‘land converted to forest land’ are key categories (for both level and trend assessments). In the preparation of this inventory, the data for these emissions underwent Tier 1 quality checks.

For the pre-1990 planted forests, one of the primary input data sets used is the National Exotic Forest Description. The National Exotic Forest Description is New Zealand’s official source of statistics on planted production forests and, as such, is subject to formalised data-checking procedures. Each National Exotic Forest Description report is reviewed by a technical National Exotic Forest Description committee before publication. Broad comparisons of forest areas reported in the National Exotic Forest Description reports are made with independent sources of information such as the Land Cover Database (LCDB) estimates and the annual results of Statistics New Zealand’s Agricultural Production Survey. National Exotic Forest Description yield tables have been subject to review (eg, Jaakko Poyry Consulting, 2003; Manley, 2004) and have recently been revised.

For post-1989 forests, quality-assurance and control procedures were specified for both the field and LiDAR data (see Figure 7.2.2.3). The field measurements of the permanent sample plots were formally audited through the remeasurement of 10 per cent of the sites by a team independent of the field inventory contractor. The sites audited were randomly selected throughout the measurement period and revisited shortly after the original measurement. Audit results for a site were provided to the measurement contractor as soon as possible so any issues found could be addressed with the field team. The data pre-processing (in the Scion PSP system) and modelling (using FCPv2.2) were also independently checked (Woollons, 2009).

Quality-assurance and control procedures of the LiDAR data involved checking the raw data as it was acquired following the method outlined in Stephens et al (2008). The key characteristics considered included sensor calibration, positional accuracy, density of first return, data decimation, consistent classification of the ground returns within the point cloud and accurate data administration. The LiDAR sensor calibration was flown four times with 600 height difference samples taken, the point positioning tested on six occasions and a summary of first returns provided for eight delivery dates. Sites that failed to meet the required pulse density were reflown. FUSION LiDAR visual and analysis software (McGaughey et al, 2004) and ERDAS IMAGINE software were used for quality assurance of the delivered LiDAR data sets continuously throughout the operation, with results and feedback provided to the contractor within 10 days of data delivery. When the data were subsequently analysed on a plot-by-plot basis, the results of this analysis were also audited with the more than 30 LiDAR variables produced and checked by an independent agency for 10 per cent of the plots.

7.2.5 Category-specific recalculations

In this submission, New Zealand has recalculated its emissions and removal estimates for the whole LULUCF sector from 1990, including for the forest land category. These recalculations have involved improved country-specific methods, activity data and emission factors. The impact of the recalculations on net CO2-e emissions estimates for the forest land category is provided in Table 7.2.5.1. The differences shown are a result of recalculations for all carbon pools used in Climate Change Convention and Kyoto Protocol reporting for the whole time series for the LULUCF sector.

Table 7.2.5.1 Recalculations on New Zealand’s estimates for the forest land category in 1990 and 2007
  Net removals and areas Change from the 2009 submission
2009 submission 2010 submission   (%)
1990 net removals –18,649.2 Gg CO2-e –32,856.7 Gg CO2-e –14,207.5 Gg CO2-e +76.2

2007 net removals

–24,527.9 Gg CO2-e –30,651.5 Gg CO2-e –6,123.6 Gg CO2-e +25.0
1990 land areas 9,368,950 ha 9,644,551 ha +275,601 ha +2.9

2007 land areas

9,993,180 ha 10,131,746 ha +138,566 ha +1.4

For forest land, the reasons for the recalculation differences are explained below.

Land area

The estimate of land area in each of the forest land subcategories was updated as a result of the land-use mapping programme. A comparison of the estimates used in this submission compared with the previous submission is given in Table 7.2.5.1. There have been changes in all three forest subcategories because the definitions of the classes have changed since the previous submission (see Table 7.1.2.7). The key points to note are as follows.

  • Previous estimates of total planted forest land were derived from the Ministry of Agriculture and Forestry’s National Exotic Forest Description. This data is compiled from a voluntary postal survey of commercial forest owners. National Exotic Forest Description data is reported primarily as net stocking while the area of natural forest (gross area not net area) was derived for the LCDB created from the classification of satellite imagery.
  • The LUCAS mapping now estimates the total area of forest land to be 3 per cent higher than previously reported. This difference can be attributed to the difference between data sources for planted forest. The LUCAS mapping area estimates of planted forest represent gross areas (using the Kyoto Protocol forest definition) while the National Exotic Forest Description estimates represent net stocked area. In New Zealand, the difference between net and gross area is generally between 10 and 15 per cent. Additionally, the new estimate for planted forest land area also covers all planted forest areas, including riparian and erosion-control plantings that meet the Kyoto Protocol forest definition, not just areas of plantation crop as in the previous submission.

Age-based and area-adjusted carbon yield tables

The LUCAS programme has introduced a new approach to estimate carbon stock change in post-1989 forests, as described in section 7.2.3. This approach is based on a forest inventory of permanent sample plots to establish the carbon stock as at 1 January 2008, and a national area-weighted and age-based carbon yield table for this forest subcategory. This new approach has only been implemented for the post-1989 forests, but will be implemented for the pre-1990 planted forests during 2010, and the results will be in the 2012 submission.

From 2010, New Zealand will estimate carbon in the two subcategories of planted forests (pre-1990 and post-1989) in the same manner. A comparison between the two national carbon yield tables, namely, the previous National Exotic Forest Description-based table as used for the 2009 submission and the current LUCAS double-sampling, inventory-based table used for this submission, has been undertaken. This comparison was completed by Scion, an independent research provider (Steve Wakelin, Scion, pers comm). The result of this comparison is provided in Figure 7.2.5.1. The approach to develop the two yield tables was as follows.

  1. The National Exotic Forest Description carbon yield table was established from regional volume tables based on data voluntarily supplied by large forest owners. Each table was converted to a carbon yield table using the C_Change model (also part of FCPv2.2), and subsequently weighted by area to create a national average carbon yield table. This is used for pre-1990 planted forests.
  2. The LUCAS carbon yield table was established by using the forest carbon modelling system, FCPv2.2 (explained above) to estimate biomass carbon (t C ha1) for each of the inventory plots from planting to maturity. This was updated to provide an area-weighted table by LiDAR regression estimators using double-sampling procedures described in section 7.2 (Kimberley et al, 2009).

Figure 7.2.5.1 shows the differences between the current and previous yield tables. The National Exotic Forest Description yield table does not reflect the specific characteristics of the post-1989 forests that have been planted into ex-grassland, especially those with relatively high soil fertility. The LUCAS yield table has been constructed using national data collected specifically for carbon reporting.

Figure 7.2.5.1 New Zealand’s previous and current post-1989 forest age-based and area-weighted forest carbon yield tables

Figure 7.2.5.1 New Zealand’s previous and current post-1989 forest age-based and area-weighted forest carbon yield tables

New Zealand’s previous and current post-1989 forest age-based and area-weighted forest carbon yield tables (t C ha-1)
Age (years) LUCAS yield table NEFD  yield table
1 6.2 0.3
2 6.7 1.4
3 8.3 4.6
4 13.4 11.0
5 21.5 19.9
6 32.3 30.5
7 44.4 41.3
8 56.1 51.3
9 66.6 60.2
10 75.8 68.2
11 84.5 76.2
12 93.2 84.5
13 101.0 93.0
14 109.1 101.9
15 117.7 111.1
16 126.7 120.5
17 136.0 130.2
18 145.7 139.9
19 155.5 149.7
20 165.3 159.3
21 175.1 168.8
22 184.7 178.0
23 194.1 187.0
24 203.2 195.7
25 212.0 204.1
26 220.6 212.3
27 229.0 220.4
28 237.2 228.3
29 245.3 236.1
30 253.3 236.1
Calculation method

Both the previous and current approaches use an age-class distribution for planted forests. The previous approach used three age-class distributions: first rotation planted forests; second rotation planted forests; and first rotation post-1989 planted forests. These were modelled separately but reported with pre-1990 planted forest under plantation (crop).

The current approach uses two age-class distributions, one for the pre-1990 planted forests and one for the post-1989 forests. The previous approach used the latest National Exotic Forest Description statistics as at 1 April 2007 (Ministry of Agriculture and Forestry, 2008a) to establish the age-class distributions and back-cast these to 1990. For pre-1990 planted forests, the current approach used the 1991 National Exotic Forest Description statistics to establish the 1990 age-class distribution. For the post-1989 forests, the age data from the national plot network was used to establish the age-class distribution as at 1 January 2008. The data was then used with National Exotic Forest Description annual new planting information to establish the 1990–2008 planting profile for these recently established forests (see Figure 7.2.1.1).

The previous National Exotic Forest Description-based approach for post-1989 forests estimates more carbon stock over the period 1994 to 2002 when compared with the current LUCAS approach. However, the current LUCAS approach estimates more carbon stock over the period 2002 to 2013.

Soil carbon stock factors

These factors have been improved (see section 7.1 – Soils). The previous reference soil carbon stock value was 83 t C ha–1. This previous reference value had been calculated incorrectly, where the effect of management on soil carbon was applied twice. The Soil CMS stock value includes the effect of management. The current reference value is 117.66 ± 12.56 t C ha–1 (Table 7.1.2.10), and the stock values for natural forest and planted forest are 111.85 ± 5.24 t C ha–1 and 104.31 ± 6.44 t C ha–1 respectively. The previous approach used the reference soil carbon stock for both natural and planted forests.

Natural forest carbon stock estimates

Previous estimates of national carbon stock values for natural forest that were used to populate previous submissions were derived from historic data. These data were not collected specifically for the purpose of estimating carbon stocks at a national scale. Essential data had not necessarily been collected, and important vegetation types and some regions with substantial forest cover were significantly under-represented. The establishment and first measurement of the national grid-based network of permanent plots has provided an unbiased estimate of carbon stored in the natural forests. The estimate used in this report is based on data collected from a complete set of plots, includes the full set of data required for carbon calculations and the result is more accurate with reduced uncertainty. The methods used are fully documented, repeatable and will be comparable once the full network of plots has been remeasured.

Deforestation area

The area estimates of deforestation have been updated from the previous submission. These are reported in the ‘land converted to’ category, primarily high-producing grassland.

Estimates of pre-1990 planted and natural forest deforestation between 1990 and 2007 are based on new LUCAS land-use mapping, whilst deforestation of post-1989 forest has been estimated from a variety sources, such as the carbon inventory of New Zealand’s planted forests for the 2007 Greenhouse Gas Inventory (Wakelin, 2008) and the Deforestation Intentions Survey (Manley, 2009).

Additional deforestation data sourced from the National Exotic Forest Description (Ministry of Agriculture and Forestry, 2009a) has been used as the basis for estimating when the actual deforestation event occurred between 1990 and 2007.

Deforestation within 2008 of pre-1990 and post-1989 planted forest is sourced from the Deforestation Intentions Survey (Manley, 2009) and unpublished work by Scion (New Zealand Forest Research Institute).

The estimate of natural forest deforestation in 2008 was estimated by linear interpolation from the average land-use change mapped between 1 January 1990 and 1 January 2008. As there was no quantitative information on the annual rate of natural forest deforestation between 1990 and 2007, the same annual rate of change was assumed for the entire period, and extrapolated out to the end of 2008. The extrapolated estimate of natural forest deforestation will be updated in future submissions as new information becomes available, and will be replaced with an actual, mapped value in the 2013 submission at the latest, following production of the 2012 land-use map. For more detail on natural forest deforestation, see section 7.2.1.

Mapping methods and data sources are described in further detail in section 7.1.2.

7.2.6 Category-specific planned improvements

A natural forest remeasurement is under way. After this remeasurement is complete, New Zealand will be able to better illustrate whether its natural forests are a source, sink or carbon neutral.

The inventory of the pre-1990 planted forests is still in the planning stage but is scheduled for 2010. Given the experience of the carbon inventory of post-1989 forests (see Land converted to forest land – section 7.2.2), a similar double-sampling approach will be employed using LiDAR in combination with ground-based permanent sample plots. The pre-1990 approach will be consistent with the inventory approach used for post-1989 forests. It is likely that at least 200 ground sites (see Figure 7.2.2.4) will be sampled on the 8-kilometre grid where this intersects pre-1990 planted forests to ensure that a robust, unbiased regression model may be estimated. This sample intensity will be sufficient to estimate the stock of carbon to a PLE 9 of less than 10 per cent, excluding any error due to measurement or calculations of individual tree and carbon pool content. Following the completion of field measurements for the pre-1990 planted forests, the methods to estimate carbon of the two subcategories (pre-1990 planted and post-1989 forests) will likely be standardised.

To ensure that all of the 4-kilometre grid intersections located within mapped post-1989 forest are sampled, the mapping of these forests will be iteratively improved to ensure that carbon estimates are unbiased. Sampling at the grid intersections would be by LiDAR only, or with both LiDAR and field measurements. The post-1989 forests will be remeasured at the end of the commitment period (2012), allowing carbon stock changes over the commitment period to be measured.

New Zealand has a long-term research programme that underpins forest carbon inventory and modelling. This work aims to improve carbon modelling, including partitioning in species other than Pinus radiata, plantation understory carbon and biomass decay rates.

The specific improvements expected from this research effort include:

  • establishment of the carbon regression between LiDAR and field measured pre-1990 planted forest plots
  • determination of how well LiDAR data is co-located with field measurements, and the effect of mislocation on the carbon regression between LiDAR plots
  • improved knowledge of decay rates associated with forest harvesting residues
  • improvement of the Douglas fir biomass database and the ability to model growth and carbon allocation using the Forest Carbon Predictor modelling system
  • improvement of the planted forest growth model and carbon allocation model for Pinus radiata. This will be achieved through peer-reviewed research to better understand decay rates of roots, dead wood and litter, drivers of wood density and improvements in the wood density model used in the carbon allocation model.

Planned future improvements also include further soil data collection for land under planted forest, and further measurement of paired sites for validation.

7.3 Cropland (CRF 5B)

7.3.1 Description

Cropland in New Zealand is separated into two subcategories: annual and perennial. In 2008, there were 334,159 hectares of annual cropland in New Zealand (1.2 per cent of total land area) and 88,541 hectares of perennial cropland (0.3 per cent of total land area). Annual crops include cereals, grains, oil seeds, vegetables, root crops and forages. Perennial crops include orchards, vineyards and shelterbelts except where these shelterbelts meet the criteria for forest land. A summary of land-use change within the cropland category is provided in Table 7.3.1.1.

The amount of carbon stored in, emitted, or removed from permanent cropland depends on crop type, management practices and soil and climate variables. Annual crops are harvested each year, with no long-term storage of carbon in biomass. However, the amount of carbon stored in woody vegetation in orchards can be significant with the amount depending on the species, density, growth rates, and harvesting and pruning practices.

In 2008, the net removals by cropland were 23.7 Gg CO2-e. Net removals from cropland have increased 53.6 Gg CO2-e (179.0 per cent) from the 1990 level when net emissions were 29.9 Gg CO2-e (see Table 7.3.1.1).

Table 7.3.1.1 New Zealand’s land-use change within the cropland category in 1990 and 2008
Cropland land-use category Net area in 1990 (ha) Net area in 2008 (ha) Change from 1990 (%) Net emissions/removals
(Gg CO2-e)1990 2008
Change from 1990 (%)
Cropland remaining cropland 417,913 417,825 –0.02 25.9 3.1 –88.0
Land converted to cropland 246 4,875 +1,881.7 4.0 –26.7 –767.5
Total 418,159 422,700 +1.1 29.9 –23.7 –179.3

Notes: 1990 and 2008 areas are as at 31 December. Net emission/removal values are for the whole year indicated.

The change within the cropland category from being a small source of emissions in 1990, to a small CO2 sink in 2008, reflects the net increase in the area of land converted to perennial cropland from other land-use subcategories with lower carbon emission factors, including 5,179 hectares from annual cropland and 4,249 hectares from high-producing grassland.

The carbon stored in cropland in 2008 was 116,849.3 Gg C. The growth in carbon stock since 1990 is 728.5 Gg C, equivalent to removals of 2,671.0 CO2-e. A summary of the carbon stock change by carbon pool within the cropland category is shown in Table 7.3.1.2.

Table 7.3.1.2 New Zealand’s carbon stock change by carbon pool within the cropland category in 1990 and 2008
Carbon pool Carbon stock in 1990
(Gg C)
Carbon stock in 2008 (Gg C) Change from 1990 (%)
Living biomass 6,655.7 6,844.9 +2.8
Dead organic matter NE NE NA
Soil 49,162.6 49,701.9 +1.1
Total 55,818.3 56,546.8 +1.3

Notes: 1990 carbon stock values are as at 31 December 1989, and 2008 values are as at 31 December 2008. Dead organic matter is not estimated (NE) as there is insufficient information to provide a basic approach with default parameters to estimate carbon stock change in this pool (IPCC, 2003).

Cropland remaining cropland

In New Zealand, there were 417,825 hectares of cropland remaining cropland in 2008. This is made up of 334,007 hectares of annual cropland remaining annual cropland, 78,639 hectares of perennial cropland remaining perennial cropland and 5,179 hectares of annual cropland converted to perennial cropland between 1990 and 2008.

The only emissions for cropland remaining cropland are a result of the land-use change from annual cropland to perennial cropland. This land-use change resulted in net removals of 34.5 Gg CO2-e in 2008. Over the period 1990 to 2008, removals totalled 300.4 Gg CO2-e.

These removals are mostly because the carbon accumulation in living biomass in perennial cropland is higher than for annual cropland (63 t C ha–1 versus 5 t C ha–1). Net change in living biomass between 1990 and 2008 for annual cropland converted to perennial cropland was 90.6 Gg C (332.3 Gg CO2-e of net removals); net change in soils over the same period was –8.7 Gg C (31.9 Gg CO2-e of net emissions). Net change in dead organic matter is not estimated as no Tier 1 defaults are available for estimating carbon stocks or change in this pool (IPCC, 2003).

In 2008, emissions from liming of cropland remaining cropland accounted for 37.6 Gg CO2-e of net emissions. Net emissions from cropland liming have increased 14.6 Gg CO2-e (63.2 per cent) from the 1990 level of 23.0 Gg CO2-e.

Land converted to cropland

Between 1990 and 2008, the net area of land converted to cropland was 4,875 hectares. Most of this (4,249 hectares) was high-producing grassland converted to perennial cropland. The net carbon change associated with land-use change to cropland was 32.1 Gg C (net removals of 117.8 Gg CO2-e).

The net change in carbon stocks for the 152 hectares of land converted to annual cropland between 1990 and 2008 was a gain of 0.478 Gg C (net removals of 1.8 Gg CO2-e). This is made up of a loss of 0.057 Gg C from living biomass (emissions of 0.2 Gg CO2-e) and a gain of 0.535 Gg C from soil organic matter (removals of 2.0 Gg CO2-e).

The net change in carbon stocks for the 4,723 hectares of land converted to perennial cropland between 1990 and 2008 was 31.6 Gg C (net removals of 116.0 Gg CO2-e). This is made up of a gain of 35.2 Gg C from living biomass (removals of 129.1 Gg CO2-e), a gain of 0.117 Gg C from soil organic matter (removals of 0.4 Gg CO2-e) and a loss of 3.7 Gg C from dead organic matter (emissions of 13.6 Gg CO2-e).

Between 1990 and 2008, emissions of nitrous oxide resulting from cultivation of land converted to cropland were 0.001 Gg N2O (0.2 Gg CO2-e).

7.3.2 Methodological issues

Emissions and removals for the living biomass and dead organic matter have been calculated using IPCC Tier 1 emission and removal factors, and activity data as described in section 7.1.2 – Representation of land areas. Emissions and removals by the soil pool are estimated using a Tier 2 method. This is described in section 7.1.2 – Soils.

Cropland remaining cropland

As well as annual cropland remaining annual cropland, and perennial cropland remaining perennial cropland, the estimates for cropland remaining cropland include area and carbon stock changes for annual cropland converted to perennial cropland.

Living biomass

To estimate carbon change in living biomass for annual cropland converted to perennial cropland, New Zealand is using Tier 1 defaults for biomass carbon stocks at harvest. The value being used for annual cropland is 5 t C ha–1. This is the carbon stock in living biomass after one year as given in GPG-LULUCF, Table 3.3.8 (IPCC, 2003). The Tier 1 method for estimating carbon change assumes carbon stocks in biomass immediately after conversion are zero, that is, the land is cleared of all vegetation before planting crops (5 t C ha–1 is removed).

To estimate growth after conversion to perennial cropland, New Zealand uses the biomass accumulation rate of 2.25 t C ha–1 yr–1. This value is based on 63 t C ha–1, the default value for perennial cropland in a temperate climate (all moisture regimes) from GPG-LULUCF, Table 3.3.2 (IPCC, 2003), spread over (or divided by) 28 years, which is the maturity period New Zealand has chosen for its lands to reach steady state.

The activity data available does not provide information on areas of perennial cropland temporarily destocked; therefore no losses in carbon stock due to temporary destocking can be calculated.

Dead organic matter

New Zealand does not report estimates of dead organic matter in this category. The notation NE (‘not estimated’) is used in the common reporting format tables. There is insufficient information to provide a basic approach with default parameters to estimate carbon stock change in dead organic matter pools in cropland remaining cropland (IPCC, 2003).

Soil carbon

Soil carbon stocks in cropland remaining cropland are estimated using a Tier 2 method as described in section 7.1.2 – Soils.

The Tier 2 value for soil carbon in annual cropland at steady state is estimated to be 118.27 t C ha–1, with a standard error of 22.47 (Table 7.1.2.10). While the estimate appears high, New Zealand soils generally do contain higher soil carbon levels than similar soils in other countries (Tate et al, 1997) and than default IPCC values (Scott et al, 2002).

The Tier 2 value for soil carbon in perennial cropland at steady state is estimated to be 114.91 t C ha–1, with a standard error of 13.22 (Table 7.1.2.10).

Soil carbon change for annual cropland converted to perennial cropland is estimated using a Tier 2 method with the change in soil carbon reflecting a linear rate of change over 20 years (the IPCC default method) from the steady state value for annual cropland (118.27 t C ha–1) to the steady state perennial cropland value (114.91 t C ha–1).

There are large standard errors associated with the soil carbon stock values due to the small size of the datasets. There are two reasons for the small datasets; historically, little focus has been placed on collecting soil data under this land use as it represents only 1.6 per cent of New Zealand’s total land area (1.2 per cent being annual cropland and 0.3 per cent being perennial cropland). In addition, recent improvements to the dataset to remove potential bias from the emphasis of historical soil survey on undisturbed pedons resulted in the invalidation of the majority of cropland records as they could not be definitely identified as being within the cultivated area (Fraser et al, 2009). Planned future improvements in the soils area include soil data collection for identified gaps.

Liming

The calculation of carbon dioxide emissions from the liming of cropland soil is based on equation 3.4.11 in GPG-LULUCF (IPCC, 2003) as outlined in section 7.1.2 – Liming. The total amount of agricultural lime (limestone) applied is provided by Statistics New Zealand (New Zealand’s official statistics agency). This is split into lime applied to cropland and grassland based on analysis of agricultural lime use by land use and farm type from the 2007 Agricultural Census. This analysis indicates that, each year, around 6 per cent of agricultural lime used in New Zealand is applied to cropland. The amount of lime applied to cropland is then converted to carbon emissions using a conversion factor of 0.12 from GPG-LULUCF, section 3.3.1.2.1.1 (IPCC, 2003).

Non-CO2 emissions

Biomass burning

This is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables. Agricultural residue burning is reported in the agriculture sector.

Land converted to cropland

Living biomass

New Zealand uses a Tier 1 method to calculate emissions for land converted to cropland. The Tier 1 method multiplies the area of land converted to cropland annually by the carbon stock change per area for that type of conversion.

The Tier 1 method assumes carbon in living biomass and dead organic matter immediately after conversion is zero, that is, the land is cleared of all vegetation before planting crops. The amount of biomass cleared when land at steady state is converted is shown in Tables 7.1.2.4 and 7.1.2.5.

The Tier 1 method also includes changes in carbon stocks from one year of growth in the year conversion takes place, as outlined in equation 3.3.8 of GPG-LULUCF (IPCC, 2003).

To estimate growth after conversion to annual cropland, New Zealand uses the IPCC default biomass accumulation rate of 5 t C ha–1 for the first year following conversion (GPG-LULUCF, Table 3.3.8, IPCC, 2003). After the first year, any increase in biomass stocks in annual cropland is assumed equal to biomass losses from harvest and mortality in that same year and, therefore, after the first year there is no net accumulation of biomass carbon stocks in annual cropland remaining annual cropland (IPCC, 2003, section 3.3.1.1.1).

To estimate growth after conversion to perennial cropland, New Zealand uses the biomass accumulation rate of 2.25 t C ha–1 yr–1. This value is based on the 63 t C ha1, the default value for perennial cropland in a temperate climate (all moisture regimes) from GPG-LULUCF, Table 3.3.2 (IPCC, 2003), being gained over 28 years, which is the maturity period New Zealand has chosen for its lands to reach steady state.

Dead organic matter

New Zealand reports only losses in dead organic matter associated with the previous land use for this category. The losses are calculated based on the carbon in dead organic matter at the site prior to conversion to cropland. It is assumed that immediately after conversion dead organic matter is zero (all carbon in dead organic matter prior to conversion is lost). There is insufficient information to estimate gain in carbon stock in dead organic matter pools after land is converted to cropland (IPCC, 2003). Consequently, where there is no dead organic matter losses associated with the previous land use, the notation key NE (‘not estimated’) is used in the common reporting format tables.

Soil carbon

Soil carbon stocks in land converted to annual and perennial cropland are estimated using a Tier 2 method as described in section 7.1.2 – Soils. In the absence of country- and land-use specific data on the time rate of change, the IPCC default of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original and new land use.

Non-CO2 emissions

Nitrous oxide emissions from disturbance associated with land-use conversion to cropland

Nitrous oxide emissions result from the mineralisation of soil organic matter with conversion to cropland. New Zealand uses the method outlined in GPG-LULUCF, equations 3.3.14 and 3.3.15, to estimate these emissions. The inputs to these equations are:

  • change in carbon stocks in mineral soils in land converted to cropland: this value is calculated from the land converted to cropland soil carbon calculations
  • EF1: the emission factor for calculating emissions of N2O from nitrogen in the soil. New Zealand uses a country-specific value of 0.01 kg N2O – N/kg N (Kelliher and de Klein, 2006).
  • C:N ratio: the IPCC default ratio of carbon to nitrogen in soil organic matter (1:15) is used (IPCC, 2003).

Nitrous oxide emissions from disturbance associated with land-use conversion to croplands are minor in New Zealand (0.1 tonnes in 2008, and 0.7 tonnes N2O in total since 1990). As New Zealand is only reporting emissions to the nearest tonne and the N2O emissions for 2008 are less than 1 tonne in 2008, the notation key NE (‘not estimated’) is used in the common reporting format tables.

Biomass burning

Biomass burning with land conversion to cropland is not thought to be a significant activity in New Zealand, and there is no activity data available that would indicate otherwise. The notation key NO (‘not occurring’) is used in the common reporting format tables.

7.3.3 Uncertainties and time-series consistency

Uncertainties are analysed as uncertainty in activity data and uncertainty in emission factors. The combined effect of uncertainty is estimated at ±96.1 per cent in annual cropland and ±94.9 per cent in perennial cropland (95 per cent confidence interval).

As shown in Table 7.3.3.1, while uncertainty in activity data is low, the uncertainty in the IPCC default variables dominates the overall uncertainty in the estimate provided by New Zealand. However, uncertainty in activity data used in the inventory will be greater than assessed for the LCDB alone as error is introduced from extrapolation of land-use change out to the end of 2008. Extrapolation of land-use change rates is needed as mapping is not repeated annually (only two data points (1 January 1990 and 1 January 2008) of mapped activity data are used).

Table 7.3.3.1 Uncertainty in New Zealand’s 2008 cropland estimates
Variable Uncertainty at a 95% confidence interval (%)
    Land-use subcategory Annual cropland Perennial cropland
Activity data uncertainty    
    Uncertainty in land area ±9.9 ±8.8
Emission factor uncertainty    
    Uncertainty in biomass accumulation rates ±75.0 (IPCC, 2003,
Table 3.3.2)
±75.0 (IPCC, 2003,
Table 3.3.2)
    Uncertainty in soil carbon stocks ±19.0 ±11.5
    Uncertainty in liming emissions ±40.0 ±40.0
    Combined emission factor uncertainty ± 40.3 ± 84.1
Implied uncertainty in N2O emissions NA NA
Total combined uncertainty ± 41.5 ± 84.6

Notes: Not applicable (NA) is shown for the uncertainty in N2O emissions, which are minor for this land-use category and reported as not estimated (NE) in the common reporting format tables. See section 7.3.2 for details. The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.3.4 Category-specific QA/QC and verification

In the preparation of this inventory, the data for these emissions underwent Tier 1 quality checks.

7.3.5 Category-specific recalculations

The impact of recalculations on net CO2-e emissions estimates for the cropland category is shown in Table 7.3.5.1. Recalculations were carried out for this category as a result of new activity data from a modified mapping process as described in section 7.1.2 – Representation of land areas.

The annual growth in biomass for land converted to perennial cropland has been increased (from 2.1 to 2.25 t C ha–1 – see Table 7.1.2.5) and the use of this factor has been improved. Previously, ‘land converted’ was only in this state for one year. The current method has ‘land in a conversion’ state for 28 years, before it moves to the ‘land remaining land’ category.

Further, the soil carbon stock factors have been improved (see section 7.1.2 – Soils). The previous reference soil carbon stock value was 83 t C ha–1. The current value is 117.66 ± 12.56 t C ha–1 (Table 7.1.2.10), and the stock values for annual and perennial cropland are 118.27 ± 22.47 t C ha–1and 114.91 ± 13.22 t C ha–1 respectively. The previous approach used the reference soil carbon stock for both annual and perennial cropland.

The amount of agricultural lime (limestone) applied was also updated following the final results from the 2007 Agricultural Census being released.

Table 7.3.5.1 Recalculations for New Zealand’s net emissions and removals from the cropland category in 1990 and 2007
  Net emissions and removals Change from the 2009 submission
2009 submission (Gg CO2-e)  2010 submission (Gg CO2-e)  (Gg CO2-e)  (%)
1990 –477.7 29.9 +507.6 +106.3
2007 –510.3 17.5 +527.8 +103.4

7.3.6 Category-specific planned improvements

As outlined above, there are plans to improve the soil dataset to reduce the uncertainty in the cropland estimates. Further detail on this is included in section 7.1.2 – Soils.

The use of historic activity data for cropland is to be investigated. This would allow for improved estimates of the area in the land converted to cropland and cropland remaining cropland categories. As there is no one dataset on land use prior to 1990, the initial work will focus on what datasets are available for all land uses before 1990 and how these datasets might be combined.

The biomass value used for perennial cropland is also under review. Indications are that the IPCC default value of 63 t C ha–1 is too high for New Zealand’s perennial cropland that is dominated by grape and kiwifruit vines.

7.4 Grassland (CRF 5C)

7.4.1 Description

In New Zealand, grassland covers a range of land-cover types. In this submission, three subcategories of grassland are used: high producing, low producing, and with woody biomass.

High-producing grassland consists of intensively managed pasture land. Low-producing grassland consists of low-fertility grasses on hill country, areas of native tussock or areas composed of low shrubby vegetation both above and below the timberline. Grassland with woody biomass, a new land-use subcategory in 2008, consists of grassland areas where the cover of woody species is less than 30 per cent and does not meet, nor have the potential to meet, the New Zealand forest definition due to either the current management regime (eg, periodically cleared for grazing) or the characteristics of the vegetation (eg, shrubland). A summary of land-use change within the grassland category is provided in Table 7.4.1.1.

Land-use research indicates that, under business-as-usual grassland farming operations, areas of woody shrublands do not become forest over a 30- to 40-year time frame (Trotter and Mackay, 2005). This is the case as long as the farmer’s intention is to manage the land as grassland for grazing animals. As soon as it is evident that the farmer modifies land management that encourages sustained growth of woody vegetation, such as removing stock, then these areas will be mapped as forest. A description of the land-management approaches that result in sustained growth of woody vegetation is contained in the mapping interpretation guide (Dougherty et al, 2009).

In 2008, there were 5,813,712 hectares of high-producing grassland (21.6 per cent of total land area) 7,701,148 hectares of low-producing grassland (28.7 per cent of total land area) and 1,056,975 hectares of grassland with woody biomass (3.9 per cent of total land area).

In 2008, the net emissions from grassland were 3,557.0 Gg CO2-e. Net emissions from grassland have increased by 1,814.6 Gg CO2-e (104.1 per cent) from the 1990 level of 1,742.4 Gg CO2-e.

Land converted to grassland was identified as a key category (level and trend) for 2008.

Table 7.4.1.1 New Zealand’s land-use change within the grassland category in 1990 and 2008
Grassland land-use category Area in 1990 (ha) Area in 2008 (ha) Change from 1990 (%)  Net emissions/removals
(Gg CO2-e)
Change from 1990 (%)
1990 2008
Grassland remaining grassland 15,044,271 14,475,052 –3.8 627.6 707.4 +12.7
Land converted to grassland 1,899 96,783 4,996.5 1,114.8 2,849.6 +155.6
Total 15,046,170 14,571,835 –3.2 1,742.4 3,557.0 +104.1

Notes: 1990 and 2008 areas are as at 31 December. Net emission/removal estimates are for the whole year indicated.

The carbon stored in grassland in 2008 was 1,784,245.8 Gg C. The change in carbon stock since 1990 is a loss of 62,615.8 Gg C, equivalent to emissions of 229,591.3 Gg CO2-e. A summary of the carbon stock change by carbon pool within the cropland category is shown in Table 7.4.1.2.

Table 7.4.1.2 New Zealand’s carbon stock change by carbon pool within the grassland category in 1990 and 2008
Carbon pool Carbon stock in 1990 (Gg C) Carbon stock in 2008 (Gg C) Change from 1990
Living biomass 87,676.2 83,375.7 –4.9
Dead organic matter 10,662.1 9,290.1 –12.9
Soil 1,748,523.3 1,691,580.1 –3.3
Total 1,846,861.7 1,784,245.8 –3.4

Note: 1990 carbon stock values are as at 31 December 1989, 2008 values are as at 31 December 2008.

Grassland remaining grassland

In New Zealand, there were 14,475,052 hectares of grassland remaining grassland in 2008. This is split into the three subcategories and changes between subcategories, as shown in Table 7.4.1.3.

Table 7.4.1.3 New Zealand’s land-use change between grassland subcategories from 1990 to 2008
Grassland land-use change category Net area in 2008 (ha)
High producing remaining high producing 5,735,734
High producing to low producing 18
High producing to with woody biomass 8,190
Low producing remaining low producing 7,641,422
Low producing to high producing 1
Low producing to with woody biomass 25,912
With woody biomass remaining with woody biomass 1,018,495
With woody biomass to high producing 14,933
With woody biomass to low producing 30,347

Note: The areas of land converted to another land use are cumulative net values for land-use change since 1 January 1990, as at 31 December 2008.

The emissions and removals associated with land-use change to and from grassland with woody biomass dominate as the area of change is large and there are significant levels of carbon in the living biomass and dead organic matter components of grassland with woody biomass (29 t C ha–1).

In 2008, emissions from burning of grassland remaining grassland were estimated as 1.345 Gg of CH4 (28.2 Gg CO2-e) and 0.010 Gg of N2O (3.1 Gg CO2-e). This is a decrease in CH4 emissions of 25.5 per cent from the 1990 level of 1.8 Gg CH4 (37.8 Gg CO2-e), and a decrease of 23.1 per cent in N2O emissions from the 1990 level of 0.013 Gg N2O (4.0 Gg CO2-e).

In 2008, emissions from liming of grassland remaining grassland accounted for 572.6 Gg O2-e of net emissions. Net liming emissions from cropland have increased 221.8 Gg CO2-e (63.2 per cent) from the 1990 level of 350.8 Gg CO2-e.

Land converted to grassland

Between 1990 and 2008, 96,783 hectares of land was converted to grassland. The total net carbon change associated with land converted to grassland was 17,013 Gg C (equating to net emissions of 62,381 Gg CO2-e).

Much of New Zealand’s grassland is grazed, with pastoral agriculture being the main land use. Most New Zealand agriculture is based on extensive pasture systems, with animals grazed outdoors year-round. Increased profitability of pastoral farming relative to other land uses has seen a recent trend for conversion of forest to pasture (deforestation).

The majority (98 percent) of land converted to grassland was land previously in forestry. The 94,906 hectares of forest land converted to grassland since 1990 comprises an estimated 34,170 hectares of natural forest deforestation, 49,188 hectares of pre-1990 planted forest deforestation and 11,548 hectares of post-1989 forest deforestation. (For more information on deforestation, see sections 7.1.1 and 7.2.1.) The net effect of this land-use change between 1990 and 2008 was 17,051 Gg C, equating to net emissions of 62,521.8 Gg CO2-e.

7.4.2 Methodological issues

Emissions and removals for the living biomass and dead organic matter have been calculated using a mix of IPCC Tier 1 emission and removal factors and country-specific factors. Emissions and removals by the soil pool are estimated using a Tier 2 method as described in section 7.1.2 – Soils, and the activity data used is described in section 7.1.2 – Representation of land areas.

Grassland remaining grassland

For grassland remaining grassland, the Tier 1 assumption is that there is no change in carbon stocks (GPG-LULUCF, section 3.4.1.1.1.1, IPCC, 2003). The rationale is that, where management practices are static, carbon stocks will be in an approximate steady state, that is, carbon accumulation through plant growth is roughly balanced by losses. New Zealand has reported NE (‘not estimated’) in the common reporting format tables where there is no land-use change at the subdivision level because no estimate of removals or emissions is able to be calculated. However, there is a significant amount of area (79,401 hectares) converted from one grassland subcategory to another. The carbon stock changes for these land-use changes are reported under grassland remaining grassland.

Living biomass

To calculate carbon change in living biomass where there is a change in the subdivision level (eg, high-producing grassland converted to low-producing grassland) it is assumed the carbon in living biomass immediately after conversion is zero, that is, the land is cleared of all vegetation. In the same year, carbon stocks in living biomass increase by the amount given in Table 7.1.2.5 – Annual growth in biomass for land converted to another land use. The values given in Table 7.1.2.4 for high-producing and low-producing grassland are Tier 1 defaults. The value given for grassland with woody biomass is a country-specific factor based on Wakelin (2004).

Dead organic matter

New Zealand does not report estimates of dead organic matter for high-producing grassland or low-producing grassland as GPG-LULUCF states there is insufficient information to develop default coefficients for estimating the dead organic matter pool (IPCC, 2003). The notation key NE (‘not estimated’) is used in the common reporting format tables.

For grassland with woody biomass, an estimate of change in dead organic matter is available from Wakelin (2004), and estimates of dead organic matter with conversion to and from this land use are given in the common reporting format tables.

Soil carbon

Soil carbon stocks in grassland remaining grassland are estimated using a Tier 2 method as described in section 7.1.2 – Soils.

The soil carbon values for the three grassland subdivisions at steady state are given in Table 7.4.2.1.

Table 7.4.2.1 New Zealand’s soil carbon stock values for the grassland subcategories
Land-use Soil carbon stock (t C ha–1)
High-producing grassland 114.93 ± 3.56
Low-producing grassland 117.66 ± 12.56
Grassland with woody biomass 111.57 ± 4.29
Liming

The calculation of carbon dioxide emissions from the liming of grassland soil is based on equation 3.4.11 in GPG LULUCF (IPCC, 2003) as outlined in section 7.1.2 – Liming. The total amount of agricultural lime (limestone) applied is provided by Statistics New Zealand (New Zealand’s official statistics agency). This is split into lime applied to cropland and grassland based on analysis of agricultural lime use by land use and farm type from the 2007 Agricultural Census. This analysis indicates that, each year, around 94 per cent of agricultural lime used in New Zealand is applied to grassland. The amount of lime applied to grassland is then converted to carbon emissions using a conversion factor of 0.12 from GPG-LULUCF, section 3.3.1.2.1.1 (IPCC, 2003).

Non-CO2 emissions

Biomass burning

Only non-carbon dioxide emissions from wildfires in grasslands are reported for the LULUCF sector. Emissions from the burning of crop stubble and controlled burning of savanna are reported in the agriculture sector, and carbon dioxide emissions from natural disturbance events are not reported because the subsequent regrowth is not captured in the inventory (GPG-LULUCF, section 3.2.1.4.2, IPCC, 2003). In both these cases, the notation key IE (‘included elsewhere’) is used for controlled burning in common reporting format Table 5(V).

To estimate the non-carbon dioxide emissions for wildfire in grassland remaining grassland, activity data is sourced from the National Rural Fire Authority database that has data from the year ending 31 March 1992. The average area burnt between April 1992 and April 2008 from this database is used as the estimate of area burnt for 1990 to 1991, as the estimates for this period are inaccurate due to the incomplete coverage in data collection. The April year data is then converted to calendar years for use in the inventory (Wakelin et al, 2009).

New Zealand-specific proportions of biomass burned during wildfire are used in the inventory. This is set at 100 per cent for high- and low-producing grassland and at 70 per cent for grassland with woody biomass (Wakelin, 2004). The biomass quantity for high-and low-producing grassland is a weighted value based on IPCC defaults (GPG-LULUCF, Table 3.4.2) and New Zealand-specific values (Payton and Pearce, 2001) compiled by Wakelin (Wakelin et al, 2009). Different biomass quantity values are used for wildfire and controlled burning of grassland with woody biomass. The different values reflect the fact that grassland with woody biomass burnt for land conversion is of a lesser stature than other scrubland (type burnt by wildfire) (Wakelin, 2004).

Land converted to grassland

Living biomass

New Zealand uses a Tier 1 method to calculate emissions for land converted to grassland. The Tier 1 method multiplies the area of land converted to grassland annually by the carbon stock change per area for that type of conversion.

The Tier 1 method assumes carbon in living biomass immediately after conversion is zero, that is, the land is cleared of all vegetation at conversion. The amount of biomass cleared when land at steady state is converted is shown in Tables 7.1.2.4 and 7.1.2.5. The Tier 1 method also includes changes in carbon stocks from one year of growth in the year conversion takes place as outlined in equation 3.3.8 of GPG-LULUCF (IPCC, 2003).

Dead organic matter

For land conversion to high- and low-producing grassland, New Zealand reports only losses in dead organic matter. The losses are calculated based on the carbon in dead organic matter at the site prior to conversion to grassland. It is assumed that immediately after conversion dead organic matter is zero (all carbon in dead organic matter prior to conversion is lost). There is insufficient information to estimate changes in carbon stock in dead organic matter pools after land is converted to high- or low-producing grassland (IPCC, 2003). Therefore where there is no dead organic matter losses associated with the previous land use the notation key NE (‘not estimated’) is used in the common reporting format tables.

For land converted to grassland with woody biomass, there is a country-specific value for carbon in dead organic matter. Where land is converted to grassland with woody biomass, dead organic matter accumulates at 3 t C ha–1 over 28 years (the maturity period New Zealand has chosen for land to reach steady state) (Wakelin, 2004).

Soil carbon

Soil carbon stocks in land converted to grassland are estimated using the Soil Carbon Monitoring System, a Tier 2 method that uses New Zealand-specific land-use and soil pedon data, as described in section 7.1.2. In the absence of country- and land-use specific data on the time rate of change, the IPCC default of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original land use and the new land use.

Non-CO2 emissions

Biomass burning

Biomass burning on land converted to grassland is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables.

7.4.3 Uncertainties and time-series consistency

Uncertainties can be analysed as uncertainty in activity data and uncertainty in variables such as emission factors, growth rates and the effect of land-management factors. The combined effect of uncertainty in each of the grassland subcategories is estimated to be approximately ±94 per cent (95 per cent confidence interval). As shown in Table 7.3.3.1, while uncertainty in activity data is low, uncertainty in the IPCC default variable (GPG-LULUCF, Table 3.4.2, IPCC, 2003) dominates the overall uncertainty in the estimate provided by New Zealand. However, uncertainty in activity data used in the inventory will be greater than assessed for the land-use maps alone. Error is introduced from extrapolation, as mapping is not repeated annually. Only two data points (1 January 1990 and 1 January 2008) of mapped activity data are used.

Uncertainties in liming are estimated as ±40.0 per cent based on sampling and survey respondent error. Uncertainties in N2O are also estimated as ±40.0 per cent.

Table 7.4.3.1 Uncertainty in New Zealand’s 2008 estimates for the grassland category
Variable Uncertainty at a 95% confidence interval (%)
    Grassland subcategory High producing Low producing With woody biomass
Activity data uncertainty      
    Uncertainty in land area ±7.0 ±6.0 ±9.5
Emission factor uncertainty      
    Uncertainty in biomass     accumulation rates ±75.0 (IPCC, 2003, Table 3.3.2) ±75.0 (IPCC, 2003, Table 3.3.2) ±75.0 (IPCC, 2003, Table 3.3.2)
    Uncertainty in soil carbon stocks ±3.1 ±10.7 ±3.9
    Uncertainty in liming emissions ±40.0 ±40.0 ±40.0
    Combined emission factor     uncertainty ± 47.1 ± 60.5 ± 834.2
Implied uncertainty in N2O emissions ±94.0 ±94.5 ±94.0
Total combined uncertainty ± 47.6 ± 60.8 ± 834.3

Notes: Nitrous oxide emissions are calculated as a proportion of carbon stock changes with the same uncertainty as for CO2, and therefore do not add to the combined uncertainty value. The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.4.4 Category-specific QA/QC and verification

Carbon dioxide emissions from the ‘grassland remaining grassland’ and ‘land converted to grassland’ categories are key categories (level and trend and level assessment respectively). In the preparation of this inventory, the data for these emissions underwent Tier 1 quality checks.

7.4.5 Category-specific recalculations

The impact of recalculations on net CO2-e emission estimates for the grassland category is shown in Table 7.4.5.1 below. The largest net emission/removal recalculation change for this category is the result of reporting the impacts of deforestation in the ‘land converted to grassland’ subcategory. In the previous submission, deforestation emissions were not explicitly separated from harvesting emissions, and were reported in either land remaining forest land or grassland converted to forest land categories.

Some differences between the previous and current submissions have resulted from new activity data from the improved mapping process as described in section 7.1.2 – Representation of land areas.

The annual growth in biomass for grassland now includes above- and below-ground biomass for the three grassland subcategories. Previously, only above-ground biomass carbon stock changes were estimated.

Grassland with woody biomass has been added as a new grassland subcategory and therefore grassland categories better reflect land types/uses within New Zealand.

The soil carbon stock factors have been improved (see section 7.1.2 – Soils). The previous reference soil carbon stock value was 83 t C ha–1. The current value is 117.66 ± 12.56 (Table 7.1.2.10), and the stock values for the three grassland subcategories are 114.93 ± 3.56, 117.66 ± 12.56 and 111.57 ± 4.29 t C ha–1 for high-producing, low-producing and woody biomass grasslands respectively. The previous approach used the reference soil carbon stock for the first two grasslands subcategories.

Table 7.4.5.1 Recalculations for New Zealand’s net emissions from the grassland category in 1990 and 2007
  Net emissions Change from the 2009 submission
2009 submission (Gg CO2-e)  2010 submission (Gg CO2-e)  (Gg CO2-e)  (%)
1990 863.9 1,742.4 +878.5 +101.7
2007 1,063.7 13,618.3 +12,554.6 +1,180.3

7.4.6 Category-specific planned improvements

The use of historic activity data for grassland is to be investigated. This would allow for improved estimates of the area in the land converted to grassland and grassland remaining grassland categories. As there is no one dataset on land use prior to 1990, the initial work will focus on what datasets are available for all land uses before 1990 and how these datasets might be combined.

The carbon stock for the grassland with woody biomass land-use subcategory will be improved once a sampling framework has been established and plot measurements made. Additional work to improve the mapping of this land-use class will also be carried out.

7.5 Wetlands (CRF 5D)

7.5.1 Description

New Zealand has 425,000 kilometres of rivers and streams, and almost 4,000 lakes that are larger than 1 hectare. Damming, diverting and extracting water for power generation, irrigation and human consumption has modified the nature of these waterways and can deplete flows and reduce groundwater levels. Demand for accessible land has also led to the modification of a large proportion of New Zealand’s vegetated wetland areas in order to provide pastoral land cover. Just over 10 per cent of wetlands present prior to European settlement remain across New Zealand (McGlone, 2009).

Section 3.5 of GPG-LULUCF defines wetlands as “land that is covered or saturated by water for all or part of the year (eg, peat land) and that does not fall into the forest land, cropland, grassland or settlements categories”. This category can be further subdivided into managed and unmanaged wetlands according to national definitions. The definition includes reservoirs and flooded land as managed subdivisions, and natural rivers and lakes as unmanaged subdivisions. Flooded lands are defined in GPG-LULUCF as “water bodies regulated by human activities for energy production, irrigation, navigation, recreation, etc, and where substantial changes in water area due to water regulation occur. Regulated lakes and rivers, where the main pre-flooded ecosystem was a natural lake or river, are not considered as flooded lands”. As the majority of New Zealand’s hydroelectric schemes are based on rivers and lakes where the main pre-flooded ecosystem was a natural lake or river, they are not defined as flooded lands. 10 As no other areas of New Zealand’s wetlands qualify as ‘managed’ under the GPG-LULUCF wetlands definition, all of New Zealand’s wetlands have been categorised as ‘unmanaged’, even though, more broadly, it can be said that all land in New Zealand is under some form of management and management plan (see section 11.4.1).

New Zealand’s wetlands are mapped into two subcategories: ‘wetland – open water’, which includes lakes and rivers, and ‘wetland – vegetated’, which includes herbaceous vegetation that is periodically flooded, and estuarine and tidal areas. New Zealand has mapped its vegetated wetlands using existing LCDB data. Areas of open water have been mapped using hydrological boundaries defined by Land Information New Zealand (LINZ).

There were 644,135 hectares of wetlands in 2008 in New Zealand, an increase of 70 hectares since 1990. This category is 2.4 per cent of the total New Zealand land area.

In 2008, the net emissions from wetlands were 0.8 Gg CO2-e. Net emissions from wetlands have increased 0.8 Gg CO2-e from the 1990 level, when net emissions were 0.01 Gg CO2-e. These emissions are the result of marginal land conversions to wetlands, mainly from the subcategory forest land converted to wetlands, as carbon stored in forest land is lost instantly on land conversion, and there is no Tier 1 method for estimating carbon gain in biomass for wetlands.

Wetlands were not a key category in 2008.

Table 7.5.1.1 New Zealand’s land-use change for the wetlands category in 1990 and 2008, and associated CO2-equivalent emissions
Wetlands land-use category Net area in 1990 (ha) Net area in 2008 (ha) Change from 1990 (%) Net emissions/ removals (Gg CO2-e) Change from 1990 (%)
1990 2008
Wetlands remaining wetlands 644,054 643,920 –0.02 NE NE NA
Land converted to wetlands 11 215 +1,854.6 0.011 0.829 –7,433.3
Total 644,065 644,135 +0.01 0.011 0.829 –7,433.3

Notes: 1990 and 2008 area values as at 31 December. Net emission values are for the whole year indicated. Net emissions from the wetlands remaining wetlands land-use category are not estimated (NE); see section 7.5.2 for details.

7.5.2 Methodological issues

Wetlands remaining wetlands

Living biomass and dead organic matter

A basic method for estimating CO2 emissions in wetlands remaining wetlands is provided in Appendix 3A.3 of GPG-LULUCF. The appendix covers emissions from flooded land and extraction from peat land. Recultivation of peat land is included under the agriculture sector.

Due to the current lack of data on biomass carbon stock changes in wetlands remaining wetlands, New Zealand has not prepared estimates for change in living biomass or dead organic matter for this category, as allowed for in the IPCC GPG-LULUCF, chapter 1.7.

Soil carbon

Soil carbon stocks in wetlands remaining wetlands are estimated using a Tier 2 method as described in section 7.1.2 – Soils.

The soil carbon stock at equilibrium state is estimated to be 104.62 t C ha–1, with a standard error of 19.92 (Table 7.1.2.10).

The high level of uncertainty associated with this estimate is due to the small size of the dataset. Historically, little focus has been placed on collecting soil data under this land use as it represents 2.4 per cent of New Zealand’s total land area, and the historical emphasis of soil data collection has been on productive land uses. There is further uncertainty associated with the estimate for wetlands because organic soils are often associated with vegetated non-forest wetlands, and organic soils are currently excluded from the Soil Carbon Monitoring System calculations. However, the effect of this on emissions from soil carbon is probably negligible as organic soils cover approximately 0.9 per cent of New Zealand’s total land area (Tate et al, 2005), and there is very little conversion to and from wetlands (only 70 hectares of change between 1990 and 2008).

Non-CO2 emissions

Biomass burning

Biomass burning on wetlands remaining wetlands is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables.

Land converted to wetlands

Living biomass and dead organic matter

New Zealand uses a Tier 1 method to calculate emissions for land converted to wetlands (GPG-LULUCF, equation 3.5.6, IPCC, 2003). A key assumption is that all land converted to wetlands becomes flooded land. The Tier 1 method assumes carbon in living biomass and dead organic matter present before conversion is lost in the same year as the conversion takes place and that carbon stock in living biomass and dead organic matter following conversions are equal to zero.

Soil carbon

Soil carbon stocks in land converted to wetlands are estimated using a Tier 2 method as described in section 7.1.2 – Soils. In the absence of country- and land-use specific data on the time rate of change, the IPCC default method of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original land use and wetlands for any given period.

Non-CO2 emissions

Non-CO2 emissions from drainage of soils and wetlands

New Zealand has not prepared estimates for this category as allowed for in IPCC GPG- LULUCF, chapter 1.7. The drainage of soils and wetlands is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity.

Biomass burning

Biomass burning on land converted to wetlands is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables.

7.5.3 Uncertainties and time-series consistency

Uncertainties are analysed as uncertainty in activity data and uncertainty in emission factors. The uncertainties for wetlands are estimated as ±77.7 per cent based on the uncertainty in mapping and in carbon stocks lost during conversion to wetlands (GPG-LULUCF, section 3.5.2.1.1.4, Table 3.5.2). While uncertainty in activity data is low, uncertainty in the IPCC default variables dominates the overall uncertainty in the estimate provided by New Zealand. However, uncertainty in activity data used in the inventory will be greater than assessed for the LCDB alone. Error is introduced from extrapolation as mapping is not repeated annually. Only two data points (1 January 1990 and 1 January 2008) of mapped activity data are used.

Table 7.5.3.1 Uncertainty in New Zealand’s 2008 estimates for the wetlands category
Variable Uncertainty at a 95% confidence interval (%)
Activity data uncertainty  
    Uncertainty in land area ±9.9
Emission factor uncertainty  
    Uncertainty in biomass accumulation rates ±75.0
    Uncertainty in soil carbon stocks ±19.0
    Combined emission factor uncertainty ±97.2
Total combined uncertainty ± 97.7

Note: The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.5.4 Category-specific QA/QC and verification

In the preparation of this inventory, the activity data and emissions factor for soil carbon change underwent Tier 1 quality checks.

7.5.5 Category-specific recalculations

The impact of recalculations on net CO2-e emission estimates for the wetlands land-use category is shown in Table 7.5.5.1. Recalculations were carried out for this category as a result of new activity data from the improved mapping process as described in section 7.1.2 – Representation of land areas.

The carbon stock in soils at steady state has also been recalculated since the last submission. Details of this process are described in section 7.1.2 – Soils.

Table 7.5.5.1 Recalculations for New Zealand’s net emissions from the wetlands category in 1990 and 2007
  Net emissions Change from the 2009 submission
2009 submission (Gg CO2-e)  2010 submission (Gg CO2-e)   (Gg CO2-e)  (%)
1990 0.7 0.0 –0.7 –98.5

2007

0.7 4.2 +3.5 +482.7

7.6 Settlements (CRF 5E)

7.6.1 Description

This land-use category, as described in GPG-LULUCF chapter 3.6, includes “all developed land, including transportation infrastructure and human settlements of any size, unless they are already included under other land-use categories”. Settlements include trees grown along streets, in public and private gardens, and in parks associated with urban areas.

There were 206,288 hectares of settlements in 2008 in New Zealand, an increase of 2,747 hectares since 1990. This category was 0.8 per cent of New Zealand’s total land area in 2008. The largest area of change to settlements between 1990 and 2008 was from high-producing grassland, with 1,675 hectares of high-producing grassland converted to settlement between 1990 and 2008.

In 2008, the net emissions from settlements were 20.0 Gg CO2-e. These emissions are from the subcategory land converted to settlements.

Settlements were not a key category in 2008.

Settlements remaining settlements

In 2008, there were 203,272 hectares of settlements remaining settlements. Carbon in living biomass and dead organic matter are not estimated for this land-use category. The carbon stock in soil for this land use is assumed to be in steady state.

Land converted to settlements

There were 3,016 hectares of land converted to settlements between 1990 and 2008. The change in carbon stocks for this land-use change between 1990 and 2008 was estimated to be a loss of 125.0 Gg C (emissions of 458.2 Gg CO2-e).

Table 7.6.1.1 New Zealand’s land-use change within the settlements category in 1990 and 2008, and associated CO2-equivalent emissions
Settlements land-use category Net area as at 1990 (ha) Net area as at 2008 (ha) Change from 1990 (%) Net emissions/removals (Gg CO2-e) Change from 1990 (%)
1990 2008
Settlements remaining settlements 203,408 203,272 –0.1 NE NE NA
Land converted to settlements 133 3,016 +2,167.7 6.585 20.046 +204.4
Total 203,541 206,288 +1.3 6.585 20.046 +204.4

Notes: 1990 and 2008 area values as at 31 December. Net emission values are for the whole year indicated. Net emissions for the settlements remaining settlements land-use category are not estimated (NE) as New Zealand has insufficient activity data for this subcategory; see section 7.6.2 for details.

7.6.2 Methodological issues

Settlements remaining settlements

Living biomass and dead organic matter

A basic method for estimating CO2 emissions in settlements remaining settlements is provided in Appendix 3A.4 of GPG-LULUCF. The methods and available default data for this land use are preliminary and based on an estimation of changes in carbon stocks per tree crown cover area or carbon stocks per number of trees as a removal factor (IPCC, 2003). New Zealand does not have this level of activity data and is therefore unable to estimate emissions for this subcategory. The reporting of settlements remaining settlements is optional (GPG-LULUCF, chapter 1.7, IPCC, 2003).

Soil carbon

Soil carbon stocks in settlements remaining settlements are unable to be estimated using the Tier 2 method as described in section 7.1.2 – Soils, as there is no soil data for this land use. Soil data has not been collected for this land use as it represents only 0.8 per cent of New Zealand’s total land area, and the historical emphasis of soil data collection has been on productive land uses. In the absence of either land-use specific data or an IPCC default, the Soil CMS model intercept value (117.66 tonnes C ha–1) was used as the default.

Land converted to settlements

Living biomass and dead organic matter

New Zealand has applied a Tier 1 method for estimating carbon stock change with land conversion to settlements (GPG-LULUCF, equation 3.6.1, IPCC, 2003). This is the same as that used for other areas of land-use conversion (eg, land converted to cropland). The default assumptions for a Tier 1 estimate are that all living biomass and dead organic matter present before conversion are lost in the same year as the conversion takes place and that carbon stocks in living biomass and dead organic matter following conversion are equal to zero.

Soil carbon

Soil carbon stocks in land converted to settlements are estimated using a Tier 2 method as described in section 7.1.2 – Soils. In the absence of country- and land-use specific data on the time rate of change, the IPCC default of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original land use and settlements for any given period.

Non-CO2 emissions

Biomass burning

Biomass burning on land converted to settlements is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables.

7.6.3 Uncertainties and time-series consistency

Uncertainties are analysed as uncertainty in activity data and uncertainty in emission factors. The uncertainties for settlements are estimated as ±75.3 per cent based on the uncertainty for Tier 1 grassland carbon stocks (GPG-LULUCF, Table 3.4.2, IPCC, 2003). While uncertainty in activity data is low, uncertainty in the IPCC default variables dominates the overall uncertainty in the estimate provided by New Zealand. However, uncertainty in activity data used in the inventory will be greater than assessed for the LCDB alone. Error is introduced from extrapolation as mapping is not repeated annually. Only two data points (1 January 1990 and 1 January 2008) of mapped activity data are used. In addition, mapping is not specific to IPCC categories.

Table 7.6.3.1 Uncertainty in New Zealand’s 2008 estimates for the settlements category
Uncertainty source Uncertainty at a 95% confidence interval (%)
Activity data uncertainty ±9.8
Emission factor uncertainty ±56.3
Total combined uncertainty ± 57.2

Note: The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.6.4 Category-specific QA/QC and verification

In the preparation of this inventory, the activity data and emissions factor for soil carbon change underwent Tier 1 quality checks.

7.6.5 Category-specific recalculations

The impact of recalculations on net CO2-e emission estimates for the settlements land-use category is shown in Table 7.6.5.1. Recalculations were carried out for this category as a result of new activity data from the improved mapping process as described in section 7.1.2 – Representation of land areas.

The carbon stock in soils at steady state has also been recalculated since the last submission. Details of this process are described in section 7.1.2 – Soils.

Table 7.6.5.1 Recalculations for New Zealand’s net emissions from the settlements category in 1990 and 2007
National Inventory Report Net emissions (Gg CO2-e) Change from the 2009 submission
2009 submission 2010 submission (Gg CO2-e)  (%)
1990 estimate 97.2 6.6 –90.6 –93.2
2007 estimate 97.2 102.7 +5.6 +5.7

7.7 Other land (CRF 5F)

7.7.1 Description

Other land is defined in GPG-LULUCF section 3.7 as including bare soil, rock, ice, and all unmanaged land areas that do not fall into any of the other five land-use categories. It mostly consists of steep, rocky terrain at high elevation, often covered in snow or ice. This category is 3.3 per cent of New Zealand’s total land area.

In 2008, the net emissions from other land were 26.9 Gg CO2-e. Net emissions from other land are 15.4 Gg CO2-e (133.0 per cent) higher than the 1990 level of 11.5 Gg CO2‑e. These emissions are from the land converted to other land category. Other land was not a key category in 2008.

The LCDB analysis shows that most of the land converted to other land between 1990 and 2008 was from the forest land category (Table 7.1.2.8). Between 1990 and 2008, 314 hectares of natural forest, 292 hectares of pre-1990 planted forest and 68 hectares of post-1989 forest were converted to other land. This is likely to be from erosion of forested land. The net effect of this land-use change was a loss of 138.0 Gg C, equivalent to emissions of 506.0 Gg CO2-e.

Table 7.7.1.1 New Zealand’s land-use change within the other land category in 1990 and 2008
Other land land-use category Net area as at 1990 (ha) Net area as at 2008 (ha) Change from 1990 (%) Net emissions/removals (Gg CO2-e) Change from 1990 (%)
1990 2008
Other land remaining other land 894,577 888,112 –0.7 NA NA NA
Land converted to other land 32 956 +2,887.5 11.5 26.9 +133.0
Other land 894,609 889,068 –0.6 11.5 26.9 +133.0

Notes: 1990 and 2008 area values as at 31 December. Net emission values are for the whole year indicated. Net emissions for other land remaining other land are not applicable (NA) as change in carbon stocks and non-CO2 emissions are not assessed for this category; see section 7.2.2 for details.

7.7.2 Methodological issues

Other land remaining other land

The area of other land has been estimated based on LCDB2. The method used is described more fully in section 7.1.2 – Representation of land area.

Living biomass and dead organic matter

All of New Zealand’s land area in the other land category is classified as ‘managed’. New Zealand considers all land to be managed as all land is under some form of management plan, regardless of the intensity and/or type of land-management practices. No guidance is provided in GPG-LULUCF for estimating carbon stocks in living biomass or dead organic matter for other land that is managed, and, as a result, the change in carbon stocks and non-CO2 emissions and removals are not assessed for this category.

Soil carbon

Soil carbon stocks in other land remaining other land are unable to be estimated using the Tier 2 method described in section 7.1.2 – Soils, as there is no soil data for this land use. Soil data has not been collected for this land use as it represents just 3.3 per cent of New Zealand’s total land area, and the historical emphasis of soil data collection has been on productive land uses. To estimate soil carbon in other land, the IPCC Tier 1 default of 88 t C ha–1 is used. This is the default for Moist Temperate High Activity Clay (HAC) and is the value that relates to the largest soil–climate category for New Zealand (Scott et al, 2002). It provides a balance between the knowledge that New Zealand soils generally do contain higher soil carbon levels than similar soils in other countries (Tate et al, 1997) and the recognition that carbon levels in bare soils are likely to be less than in soils under other land uses.

Land converted to other land

Living biomass and dead organic matter

New Zealand uses a Tier 1 method to calculate emissions for land converted to other land (GPG-LULUCF, equation 3.7.1, IPCC, 2003). This is the same as that used for other areas of land-use conversion (eg, land converted to cropland). The Tier 1 method assumes carbon in living biomass and dead organic matter present before conversion is lost in the same year as the conversion takes place and that carbon stock in living biomass and dead organic matter following conversions are equal to zero. There is no Tier 1 method for calculating carbon accumulation in living biomass or dead organic matter for land converted to other land.

Soil carbon

Soil carbon stocks in land converted to other land prior to conversion are estimated using a Tier 2 method as described in section 7.1.2 – Soils. In the absence of country- and land-use specific data on the time rate of change, the IPCC default method of a linear change over a 20-year period is used to estimate the change in soil carbon stocks between the original land use and other land for any given period.

Non-CO2 emissions

Biomass burning

Biomass burning on land converted to other land is a relatively minor activity in New Zealand, and there is insufficient information to reliably report on this activity. The notation key NE (‘not estimated’) is used in the common reporting format tables.

7.7.3 Uncertainties and time-series consistency

Uncertainties are analysed as uncertainty in activity data and uncertainty in emission factors. The uncertainties for other land are estimated as ±75.3 per cent based on the uncertainty in carbon stocks lost during the conversion to other land, for example, GPG-LULUCF Table 3.4.2. While uncertainty in activity data is low, uncertainty in the IPCC default variables dominates the overall uncertainty in the estimate provided by New Zealand. However, uncertainty in activity data used in the inventory will be greater than assessed for the LCDB alone. Error is introduced from extrapolation as mapping is not repeated annually. Only two data points (1 January 1990 and 1 January 2008) of mapped activity data are used. In addition, mapping is not specific to IPCC categories.

Table 7.7.3.1 Uncertainty in New Zealand’s 2008 estimates for the other land category
Uncertainty source Uncertainty at a 95% confidence interval (%)
Activity data uncertainty ±9.9
Emission factor uncertainty ±40.7
Total combined uncertainty ± 41.9

Note: The activity data and combined emissions factor uncertainty are weighted values and have been calculated using equation 5.2.2 from GPG-LULUCF, IPCC (2003).

7.7.4 Category-specific QA/QC and verification

In the preparation of this inventory, the activity data and emissions factor for soil carbon change underwent Tier 1 quality checks.

7.7.5 Category-specific recalculations

The impact of recalculations on net CO2-e emissions estimates for the other land category is shown in Table 7.7.5.1. Recalculations were carried out for this category as a result of new activity data from the improved mapping process as described in section 7.1.2 – Representation of land areas.

Table 7.7.5.1 Recalculations for New Zealand’s net emissions from the other land category in 1990 and 2007
  Net emissions Change from the 2009 submission
2009 submission (Gg CO2-e) 2010 submission (Gg CO2-e) (Gg CO2-e) (%)
1990 26.7 11.5 –15.2 –56.8
2007 40.6 88.0 +47.4 +116.7

 


9  A probable limit of error (PLE) refers to the confidence limits expressed as a percentage of the estimated mean. For example, a PLE of 10 per cent at the 95 per cent probability level implies that there is a 95 per cent chance that the true mean is within 10 per cent of the estimated mean.

10  For example, the Clyde Dam was created from the damming of the Clutha River in the South Island, creating Lake Dunstan.