View all publications

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

7.1 Sector overview

The land-use, land-use change and forestry (LULUCF) sector represented the removal of approximately 30.3% of all New Zealand's greenhouse gas emissions in 2003. Net removals from the LULUCF sector in 2003 totalled 22,861.60 Gg CO2 equivalent and are 7.0% above net removals in 1990 (Figure 7.1.1).

Figure 7.1.1 LULUCF sector net removals from 1990 to 2003

Year Gg CO2 equivalent
1990 21,366.19
1991 20,096.63
1992 17,659.17
1993 15,236.12
1994 14,154.42
1995 14,645.86
1996 14,917.93
1997 16,451.70
1998 19,298.36
1999 21,106.78
2000 22,818.93
2001 23,186.69
2002 23,326.73
2003 22,861.60

7.1.1 A history of LULUCF in New Zealand

Prior to the first human settlement by Polynesians in about 1250 AD, approximately 75% of New Zealand's land area was estimated as natural forest. The forest area had reduced to about 60% by the mid-nineteenth century and was further reduced to the current approximate 23% coverage by subsequent European settlement, the latter due largely to deforestation and clearance for pastoral grazing land. Deforestation (subsequent to human settlement) is estimated to have resulted in vegetation C losses of 3,400,000 Gg C (Scott et al., 2001). Establishment of pastures probably slightly increased mineral soil C, however some losses of C due to erosion are also possible (Tate et al., 2003).

Government controls on forest clearance (deforestation) were first imposed in the late nineteenth century. However the continuing demand for timber and agricultural land resulted in on-going forest removal. By the 1970's, growing public concern lead to stronger conservation measures by Government. Large-scale forest clearance for agricultural land ceased and New Zealand's domestic timber supply came largely from mature planted forests. Further Government administrative changes in 1987 resulted in reservation of about 5 million hectares (18% of New Zealand's total land area) of Crown-owned natural forests. Currently, New Zealand has 6.4 million hectares of natural forest. Commercial timber harvest from private natural forest was restricted to that sourced under sustainable forest management plans and permits by a 1993 amendment to the Forests Act 1949. The amendment still exempted West Coast Crown-owned forests and forests on specific Maori-owned lands. However, further Government controls resulted in the cessation of logging of the West Coast Crown-owned forests in March 2002. Timber harvested from privately owned natural forests and from forests on exempted Maori lands has continued at a low level since the 1993 controls were imposed. Current proposed legislative changes will continue to exempt the Maori lands although logging has further reduced in these forests.

New Zealand has a substantial estate of planted forests, mainly comprised of Pinus radiata, created specifically for timber supply purposes and has well-established data on this estate's extent and characteristics. These forests have removed and stored substantially more CO2 over the period 1990 to 2003 than has been emitted through forest harvesting of both the combined planted and natural forests. The average new planting rate (land reforested or afforested) over the last 30 years has been, on average, 44,900 hectares per year. In the period 1992 to 1998 new planting rates were high (averaging 69,000 hectares per year). Since 1998 the rate of new planting has declined and in 2003, 19,900 hectares of new forest was established. Between 1990 and 2003 it is estimated that 660,000 hectares of new forest has been established as a result of afforestation and reforestation activities.

Having a large planted forest resource enables New Zealand to sustainably manage its Crown and privately owned natural forest. Less than 0.1 percent of New Zealand's total forest production is now harvested from natural forests.

7.1.2 Methodological issues for LULUCF in New Zealand

Six broad categories of land are described in GPG-LULUCF. The categories are consistent with the 1996 IPCC Guidelines and the requirements of Articles 3.3 and 3.4 of the Kyoto Protocol. The land categories are:

  • Forest land - all land with woody vegetation consistent with national thresholds used to define forest land. It also includes systems with vegetation that currently fall below, but is expected to exceed, the threshold of the forest land category.
  • Cropland - arable and tillage land, and agro-forestry systems where vegetation falls below the thresholds used for the forest land category.
  • Grassland - rangelands and pasture land that is not considered as cropland. It also includes systems with vegetation that falls below and is not expected to exceed, without human intervention, the threshold used in the forest land category.
  • Wetlands - land that is covered or saturated by water for all or part of the year (e.g. peat land) and that does not fall into the forest land, cropland, grassland or settlements categories. Natural rivers and lakes are unmanaged sub-divisions on wetlands.
  • Settlements - all developed land, including transportation infrastructure and human settlements unless they are already included in other categories.
  • Other land - bare soil, rock, ice, and all unmanaged land areas that do not fall into any of the other five categories.

A current lack of land use and land-use change data consistent with the IPCC land categories, and covering the period 1990 through to 2003, limits reporting in this sector. Research is being conducted on the carbon pools and fluxes in New Zealand's soils and natural forests through the Carbon Monitoring System (CMS) plots. The focus of the CMS activities is specifically to fulfil New Zealand's requirements for the UNFCCC LULUCF inventory and Kyoto Protocol. Details of the monitoring system are included in Annex 3.2.

In previous inventories, New Zealand has only included information on net changes in the living biomass from forestry land use and emissions and removals from the planting of forest on grassland. To improve the completeness of the LULUCF sector reporting, New Zealand has modified an existing national land cover dataset to generate a Tier 1 calculation for the other LULUCF land categories. This calculation also enables a comprehensive key category analysis including the LULUCF sector and to trial the LULUCF CRF tables. New Zealand has also included a Tier 1 calculation of the country-specific model used for planted forestry to support the modelled values.

New Zealand has not yet developed land use data for 1990. The 1990 data will be developed as part of the New Zealand Carbon Accounting System (NZCAS) described in Annex 3.2. The Tier 1 calculation is provided for the years 1997 onwards.

7.1.2.1 Representation of land areas

To estimate land areas in each LULUCF land category, New Zealand has used an analysis of two existing land-cover maps of New Zealand - the Land Cover Databases 1 and 2 (respectively, LCDB1 and LCDB2) (Thompson et al, 2004). The LCDB1 and LCDB2 are an example of the wall-to-wall mapping of Approach 3 as described in GPG-LULUCF. The LCDB's were not specifically developed for use in UNFCCC reporting, however they have been used by New Zealand as they are the only national land cover/land use spatial databases available that provide current information and can be reasonably mapped to the LULUCF land. The land categories to be mapped and monitored through NZCAS will be designed specifically for reporting under the UNFCCC and Kyoto Protocol, and will replace LCDB data in the inventory in the future.

LCDB1 was completed in 2000 using SPOT satellite imagery acquired over the summer of 1996/97. LCDB2 was released in July 2004 and used Landsat 7 ETM+ satellite imagery acquired over the summer of 2101/02. During processing of LCDB2, the classifications in LCDB1 were checked, corrected and integrated into the LCDB2 database. The inventory uses the corrected LCDB1 data. There are 43 land cover/land use classes mapped in LCDB 1 and 2, and these are mutually exclusive and additive to 100% of the surface area of New Zealand. Additional information on the processing to generate LCDB1 and LCDB2 is included in Annex 3.3. To complete the UNFCCC inventory, the land cover classes in the LCDB1 and LCDB2 were mapped to the applicable LULUCF categories (Table A3C.1).

Table 7.1 shows a simplified land-use change matrix developed from LCDB1 and LCDB2 for the years 1997 and 2002. More details of the conversions between categories and sub-divisions are included in each land use category in the NIR. Land use change matrices including sub-divisions of land categories and the annual changes interpolated for the periods 1997-1998, 1998-1999, 1999-2000, 2000-2001 and 2001-2002 are included in the worksheets accompanying the LULUCF sector. A land use change matrix for 2002-2003 was generated by extrapolating previous annual trends.

Table 7.1.1 Land use change matrix between LCDB1 (1997) and LCDB2 (2002)

  Land area categories from LCDB1 (1997) (kha)
F C G W S O Total
LCDB2 2002 F 10097.57 0.02 130.88 0.00 0.06 0.06 10228.59
C 0.22 412.91 4.32 0.00 0.00 0.01 417.44
G 5.11 0.00 14360.63 0.00 0.06 0.21 14366.01
W 0.00 0.00 0.67 531.24 0.04 0.01 531.96
S 0.56 0.02 5.03 0.00 214.84 0.00 220.46
O 0.19 0.00 0.06 0.00 0.00 1056.84 1057.09
Total 10103.65 412.95 14501.58 531.24 215.00 1057.13 26821.56
Net 124.94 4.49 -135.57 0.72 5.46 -0.03 0.00

F= forest land, C=cropland, G=grassland, W= wetland, S=settlement and O=other lands.

7.1.2.2 Inventory carbon pools

Greenhouse gas inventory for land-use categories involves estimation of changes in carbon stock from five carbon pools i.e. aboveground biomass, belowground biomass, dead wood, litter and soil organic matter, as well as emissions of non-CO2 gases from such pools. For UNFCCC inventory reporting purposes, the pools are grouped into changes in living biomass (aboveground biomass and belowground biomass), changes in dead organic matter (dead wood and litter) and changes in soil organic matter.

7.1.2.3 Land-use factors and carbon stock

Changes in carbon pools within a land-use category and between categories use a number of variables of the stocks and growth in the living biomass, and the effect on land-use factors on soil carbon stocks. The variables appear in the Tier 1 equations for all land use categories. For transparency, these factors are tabulated in Tables 7.1.3.1, 7.1.3.2 and 7.1.3.3.

The types of land-use and management factors affecting soil carbon stocks are defined in GPG-LULUCF and include: 1) a land use factor (FLU) that reflects C stock changes associated with type of land use, 2) a management factor (FMG) that for permanent cropland represents different types of tillage and 3) an input factor (FI) representing different levels of C inputs to soil.

New Zealand is using a country-specific reference soil carbon stock value of 83 t C ha-1 for 0-30 cm depth. This is very similar to the default IPCC values provided in GPG-LULUCF for warm temperate moist climate in Table 3.2.4 (a range of 34-88 t C ha-1). The New Zealand value is calculated from the measured soil C in New Zealand grassland soils of 105 t C ha-1 (Tate et al, 2003), divided by the stock change factors for high producing grassland i.e. 105 t C ha-1 / 1 / 1.14 / 1.11 = 83 t C ha-1. New Zealand has always applied the default inventory time period of 21 years in calculating the Tier 1 estimates.

Table 7.1.3.1 Land-use factors used across land-use categories

Land use Stock change factors selected from GPG-LULUCF (GPG-LULUCF tables 3.3.4 and 3.4.5)
  FLU FMG FI

Planted forest

1

1

1

Natural forest

1

1

1

Annual cropland

0.71

1.0

1.11

Perennial cropland

0.82

1.16

0.91

High producing grassland

1

1.14

1.11

Low producing grassland

1

1.14

1

Other land

1

1

1

Table 7.1.3.2 Living biomass carbon stocks in land use prior to conversion

Land use Value Source/Reference

Natural forest

182 t C ha-1

364 tonnes dm ha-1 (Hall et al, 1998) - carbon fraction of dry matter (0.5).

Planted forest

222 t C ha-1

1st rotation, 28 years old (Wakelin, 2004).

Annual cropland

0 t C ha-1

Annual crop is harvested. GPG-LULUCF only considers perennial crops (Table 3.4.8).

Perennial cropland

63 t C ha-1

GPG-LULUCF Table 3.3.2. Temperate (all moisture regimes).

High producing grassland

1.35 t C ha-1

2.7 tonnes dm ha-1 (GPG-LULUCF Table 3.4.2, warm temperate - wet climate) - carbon fraction of dry matter (0.5).

Low producing grassland

0.8 t C ha-1

63 t C ha-1

1.6 tonnes dm ha-1 (GPG-LULUCF Table 3.4.2, warm temperate - wet climate) - carbon fraction of dry matter (0.5).

63 t C ha-1assumed for grassland woody vegetation.

Table 7.1.3.3 Annual growth in living biomass for land converted to another land use

Land use Value Source/Reference

Natural forest

NE /

T1=4.3 t C ha-1

New Zealand's natural forests are assumed to be approximately in steady-state (Tate et al., 2000).

Tier 1 - GPG-LULUCF 3A.1.5 and 3A.1.8 (Gw=3.5 tonnes dm ha-1 (an average of the conifer (3.0) and broadleaf (4.0) values), R = 0.24, Cfrac = 0.5).

Planted forest

IE /

T1=8.9 t C ha-1

Tier 2 - Included in C_Change modelling (Beets et al, 1999, Wakelin, 2004).

Tier 1 - GPG-LULUCF 3A.1.6 and 3A.1.8 (Gw=14.5 tonnes dm ha-1 (pinus), R = 0.23, Cfrac = 0.5).

Annual cropland

5 t C ha-1

GPG-LULUCF Table 3.3.8 (- temperate all moisture regimes.

Perennial cropland

2.1 t C ha-1

GPG-LULUCF Table 3.3.8 (- temperate all moisture regimes.

High producing grassland

6.75 t C ha-1

13.5 tonnes dm ha-1 (GPG-LULUCF Table 3.4.9, warm temperate - wet climate), Cfrac = 0.5.

Low producing grassland

3.05 t C ha-1

6.1 tonnes dm ha-1 (GPG-LULUCF Table 3.4.9, warm temperate - dry climate), Cfrac = 0.5.

7.2 Forest land (CRF 5A)

New Zealand has not adopted a final national definition of forest to be used in reporting for the Kyoto Protocol. This definition requires New Zealand to state the minimum area, length (and thus width) of land areas categorised as forest land. This definition will be used when spatial information from the NZCAS is calculated and available for reporting. To enable New Zealand to complete the 2003 inventory and include information under the updated LULUCF guidance and CRF tables, the categories of forest land used in the LCDB1 and LCDB2 were applied. This is an area of one (1) hectare and a width of 100 metres. This definition of land mapped as forest land is not New Zealand's national definition to be applied under the Kyoto Protocol when the NZCAS is operational.

The LCDB1 and LCDB2 also include a shrubland vegetation cover category, which does not exist as a LULUCF category. Land cover classes within the shrubland category were assigned to the grassland category as they do not meet the forest land criteria that New Zealand expects to formally adopt for the Kyoto Protocol (namely, 1 hectare, 30 per cent canopy cover, 5 metres height and 100 metres width). The allocation of classes is shown in Table A3C.1.

New Zealand has adopted the definition of managed forest land as provided in the IPCC Guidelines and 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". New Zealand's natural forests are considered managed forests due to their primary management for ecological, biodiversity, and social functions.

7.2.1 Description

Forest land accounted for net CO2 removals of 23,818.43 Gg CO2 in the 2003 inventory. This figure includes removals from the growth of planted forests, emissions from the conversion of land to planted forest and from the very small amount of harvesting of natural forests.

Natural forest is a term used to distinguish New Zealand's indigenous forests from planted production forests. Natural forests are managed for a range of conservation, biodiversity and recreation purposes. 99.9% of New Zealand's wood needs are met from planted forests and the Crown does not harvest any timber from New Zealand's natural forests. The natural forest harvest reported in the inventory refers to harvest of forests on land granted to Maori (New Zealand's indigenous people) under the South Island Landless Natives Act 1906. These are currently exempt from the indigenous forestry provisions of the Forests Act that apply to all privately owned indigenous forests and required a sustainable forest management plan or permit prior to any harvesting. Approximately 50,000 ha are in the SILNA. There is no specific data to estimate growth in these forests. The NZCAS will provide data for similar forests in similar locations to the SILNA forests.

Removals of CO2 in natural forest are not calculated. This is because country-specific data are not available (refer the development of the CMS plots in Annex 3.2) and the growth rates provided for mixed broadleaf/coniferous forest in GPG-LULUCF are considered too high. Preliminary results are that New Zealand's natural forests are approximately in steady-state or a possible small sink of carbon i.e. changes in vegetation carbon stock lie between 0.3 to -2.5 Tg C yr-1 (Tate et al., 2000). Results from the CMS monitoring will enable New Zealand to provide an estimate.

7.2.2 Methodological issues

Forest land remaining forest land (Tier 2)

New Zealand has calculated the removal of CO2 by planted forests and emissions from planted forests through harvesting using a modelling approach and country-specific data. Compared to many forest ecosystems, total biomass in New Zealand's planted forests is relatively straightforward to estimate. Approximately 90% of the forest area is planted in Pinus radiata, a forest is usually composed of trees of a single age class and all forests have relatively standard silviculture regimes applied. The methodology applied for the UNFCCC inventory is:

  • A survey of forest growers is undertaken annually to estimate the area of forest by age, species, silvicultural regime and location.
  • Based on the results of this survey and stem wood volumes, the C_Change model (Beets et al., 1999) is used to estimate forest biomass and carbon at one point in time. C_Change was previously known as the CARBON/DRYMAT model.
  • The Forestry Oriented Linear Programming Interpretor (FOLPI) (Garcia, 1984; Manley et al., 1991) is also used to time-shift the estate forwards to forecast future forest growth and forest management, including harvesting.
  • The models also time-shift historic estimates of CO2 removals and emissions backwards using the latest available data.

Planted forest survey data: the results of the National Exotic Forest Description (NEFD) survey as at 1 April 2003 are used to calculate removals and emissions provided in the 2003 inventory. This latest information brings in new forest area data along with data on new planting, restocking and harvesting in 2003 (MAF, 2004).

The NEFD survey provides estimates of the forest area and merchantable stem wood volume (via yield tables) by crop-type and age. A crop-type is an aggregate of forest stands that are similar, in respect to species, silviculture and location. Each crop-type has a yield table that provides estimated volumes of stem wood per hectare by age. The total forest area after harvest for the year ending March 2003 is based on (a) the latest area estimates provided by the 2003 NEFD, (b) an estimate of the area to be planted during the year, (c) an estimate of the area harvested during the year. The total estate area for the years before 2002 has been estimated through back-calculations using this latest NEFD area data combined with new planting and harvesting time-series information. The area of new land planting is based on MAF statistics. These estimates are revised and recalculated annually as provisional estimates are replaced by confirmed actual areas.

To simplify the subsequent modelling, all crop-types are then aggregated to form a single, national area weighted crop-type (broken down by year of planting) and associated area-weighted national yield table. Estimates are calculated for March years and are reported as three year averages as per the IPCC approach. The assumption is that the stem wood removed at harvest for both permanent and planted forests is oxidised in the year of harvest.

Modelling: the C_Change model estimates carbon stock per hectare, by component and annual age-class, from stem wood volume data (box 7.1). The second of the two models, FOLPI, is a linear programming model that optimises the management of the forest estate across time while maximising the discounted harvest volume. The model simulates actual rates of planting and harvesting where time-series data exists.

Box 7.1 Process steps in the C_Change model (Beets et al, 1999)

  1. Stem wood volume is converted to an oven-dry biomass weight.
  2. The dry weight of non-stem wood components (bark, branches, foliage, cones, stumps, roots, floor litter and understorey) is calculated from stem wood volume using allometric equations. These allometric equations take account of age, stocking and site fertility.
  3. Total forest biomass is converted to carbon weight. Carbon is taken as being 50% of biomass.

Several simplifications have been made in the C_Change model. Firstly it is assumed that all forests grow in a medium wood density region in New Zealand i.e. the trees have average wood density. Secondly, the model takes the weighted national crop-type as being wholly Pinus radiata when in fact around 10% of the estate is made up of other species such as Douglas-fir (5%) (Pseudotsuga menziesii), other exotic softwoods (2%) and exotic hardwoods (3%).

The FOLPI model uses the biomass and carbon stocks at one point in time to give total carbon stocks for each modelled year and changes in carbon stocks between those years. Among the outputs of the FOLPI model are the LULUCF inventory results for 1990 to 2003. These results include:

  • Stem wood volume harvested from the planted estate, hence CO2 emitted in that harvest.
  • Total stock of estate carbon after harvesting in each year (accounting also for the decay of non-stem wood carbon left after harvesting).

The removal of carbon (net of harvest) is calculated from the total stock values. The gross removal of carbon is then calculated by adding the harvested stem wood carbon back into the net carbon removal figures. This gives the change in carbon stock between last year's harvested forests and this year's unharvested forests.

Emissions from natural forest harvesting: estimates of any harvesting from natural forests are provided by the MAF. Stem wood volumes are converted to oven-dry weight using a factor of 0.5 (accounting for wood moisture) and then expanded to include non-stem wood biomass using a factor of 2.04. As for planted forest harvesting, emissions from harvesting natural forests are potential maximums rather than actual emissions as much of the carbon contained in the harvested stem wood ends up as wood products produced from the harvested timber rather than being oxidised at the time of harvest.

Land converted to forest land

Data on the amount of land clearance for new forest planting are sourced from the MAF. The information includes the proportion of new forest planting that occurs on grassland with woody vegetation that falls below and is not expected to exceed, without human intervention, the threshold used to define forest land. Data are available from 1993 to the present and it is assumed that prior to 1993, the proportion was 20%. It is assumed that 25% of the vegetation biomass is burnt on-site and that the remainder is left to decay.

The quantity of on-site biomass for both grassland woody vegetation and natural forest, used in the land conversion calculations (worksheets 5a-d), is now based on the (provisional) results of research (adapted from Hall et al., 1998). The values reported (136 t dm/ha for grassland with woody vegetation and 364 t dm/ha for mature natural forest) are based on a national area-weighted average for biomass per hectare for a range of the principal classes.

The non-CO2 emissions estimated for burning of both forest biomass and grassland with woody vegetation are likely to be over-estimated as not all biomass is typically consumed in a fire event. For mature forests, the typical fuel load combusted may only be 10 - 40% of on-site biomass. For grassland with woody vegetation, the upper value could be somewhat higher (depending on seasonal and other factors), though combustion is again unlikely to be complete for wildfire events. The current assumptions reflect a lack of data on the percentage of fuel load combusted.

Wildfire burning: only non-CO2 emissions from wildfires are reported in the inventory (consistent with the IPCC default method). Emissions from wildfires are based on fire reports collected by the National Rural Fire Authority. These reports show the area of forest and grassland with woody vegetation burnt. It is assumed that all forest burning occurs in natural forest. In planted forests, fires occur infrequently and fire-damaged trees are usually salvaged and appear in harvest statistics. Some of the areas reported in the Fire Authority statistics involve land clearing and it is not specified whether this is for agricultural or forestry purposes. This implies that there may be double counting between these figures and those allocated to land clearing for new forest planting.

Forest removals and emissions calculated via a Tier 1 approach

The LCDB1 and LCDB2 analysis used to complete the LULUCF inventory allows calculating a coarse Tier 1 estimate for the categories 'forest land remaining forest land' and 'land converted to forest land'. These estimates are reported in the NIR to support the modelling approach.

Living biomass: for the category 'forest land remaining forest land', the calculation follows the Tier 1 procedure outlined in GPG-LULUCF equation 3.2.4 using parameters from Tables 3A.1.5 for natural forest, 3A.1.6 for plantation forest and root-shoot ratios from Table 3A.1.8. The values chosen for the carbon stocks and growth rates (Gw and R) are documented in Table 7.1.3.2.

Dead organic matter: in the Tier 1 calculation, the average transfer rate into the dead wood pool equals the transfer rate out of the dead wood pool. The net change is zero (GPG-LULUCF).

Soil carbon: in the Tier 1 calculation it is assumed that when forest remains forest, the carbon stock in soil organic matter does not change. For land converted to forest, New Zealand has followed the Tier 1 method outlined in GPG-LULUCF 3.2.2.3. For Tier 1, the initial soil C stock is determined from the same reference soil C stocks used for all land uses, together with stock change factors (FLU, FMG, FI) appropriate for the previous land use. The stock change factors used by New Zealand's Tier 1 calculation are listed in Table 7.1.3.1. New Zealand has used 83 tonnes C ha-1 for reference C stock (Tate et al, 2003).

Decreases in the carbon stock of forest land from harvest were calculated according to GPG-LULUCF equation 3.2.9 - 'annual other losses of carbon'. This equation was used in preference to the equation for commercial felling (equation 3.2.7) to keep calculations consistent with the LCDB analysis used for other land categories i.e. equation 3.2.7 is based on the extracted roundwood volume rather than the area of forest harvested or disturbed (equation 3.2.9). The annual area of deforestation was calculated from the total area of LCDB1 forest classes that had changed to harvested forest in LCDB2, divided by the five years. Carbon stocks in living biomass for planted and natural forest were those documented in Table 7.1.3.2. Under the Tier 1 methodology, it is assumed that all above ground biomass is lost.

Decreases in living biomass carbon stocks associated with land converted to forest were also included in the Tier 1 calculation. For the Tier 1 estimate, it was assumed that the land category of 'low producing grassland converted to planted forest' was equivalent in area to the clearance of grassland with woody vegetation for forest planting. The value of living biomass carbon stocks in perennial crop land prior to conversion provided in GPG-LULUCF (63 t C ha-1) was used as an approximation of the C stock in grassland woody vegetation biomass. This value is very similar to the value used in the Tier 2 modelling of 68 t C ha-1 from Hall et al (1998) (136t dm ha-1 * 0.5 (carbon fraction of dry matter)).

Table 7.2.1 compares the results from the modelled and Tier 1 approaches for the 2003 inventory. The results for the living biomass stock are similar (11.0% different) however the different approaches mean that the two estimates will always differ. The primary reasons for this are the different methodologies used in assessing planted forest estate and planting rates i.e. comparing annual NEFD survey versus remote sensing and extrapolation of previous interpolated planting trends, the lack of forest age data from the LCDB analysis, mapping of LCDB classes to LULUCF categories and the selection of GPG defaults for Gw and R compared to country-specific modelling data.

Table 7.2.1 Comparison of the Tier 2 and Tier 1 approaches for forest land in 2003

  Area (kha) Living biomass stock (Gg C)
  Modelled/ actual Tier 1 Difference Modelled/ actual Tier 1 Difference

Forest land remaining forest land (planted)

1878

2046

8.9%

6977

7742

11.0%

Land converted to planted forest land

18

26

44%

-237

-291

22.8%

7.2.3 Uncertainties and time-series consistency

A process of using the models to time-shift the forest estate forwards to represent future forest growth and forest management, and backwards to improve historical estimates, is performed to minimise errors. As the estimation of carbon stocks is continuously being improved, both past and future years are re-calculated.

Attempts have been made to quantify the uncertainties in the CO2 removal estimates for planted forests. However, it is difficult to quantify the overall error due to the assumptions implicit in the models. Some uncertainties within the C-Change (CARBON/ DRYMAT) model are well characterised (Hollinger et al., 1993). These include ±3% for wood density, ±15% for carbon allocation and ±5% for carbon content. Combining the uncertainties indicates that the proportional error in the carbon sequestration estimates is likely to be at least ±16%. The total national planted area is considered to be accurate to within ±5% (MAF, 2003) and the yield tables are assumed to be accurate to within ±5%.

A sensitivity analysis was conducted using the above accuracy ranges for total planted area and commercial yield, and a proportional uncertainty error of ±16%. The C-Change (CARBON/ DRYMAT) model runs indicate that the precision of the carbon stock estimates could be of the order of ±25%. As part of the Carbon Accounting System, research has been commissioned to better quantify uncertainty. No uncertainty estimates are currently available for emissions from harvesting of natural forests.

Including the forest land category in the overall inventory uncertainty analysis (Annex 7) shows that removals from forest land are 6.7% of New Zealand's total emissions and removals in 2003. This is the third largest source of uncertainty in New Zealand's total for a given year. Forest land introduces 2.2% uncertainty into the trend in the national total from 1990 to 2003. This is the second largest impact on the trend after CO2 emissions from the energy sector.

7.2.4 Source-specific QA/QC and verification

The information presented in the NIR and the variables chosen for calculation were reviewed by officials of the MFE and the MAF. Calculated estimates were visually assessed for obvious errors in calculations. For forest land, the Tier 2 modelled values were compared against a Tier 1 value. Land-use change matrices were used to ensure that the allocation of land between categories produced a consistent national total area of land.

One of the primary input datasets used is the National Exotic Forest Description (NEFD). The NEFD is New Zealand's official source of statistics on planted production forests and as such is subject to formalised data checking procedures. Each NEFD report is reviewed by a technical NEFD Committee prior to publication. In addition broad comparisons of forest areas reported in the NEFD reports are made with independent sources of information such as the Land Cover Database estimates and the annual results of Statistics New Zealand's Agricultural Production Survey.

The 2003 planted forests removals and emissions have been compared for consistency with the 2002 estimates (Wakelin S, Carbon Inventory of NZ's Planted Forests [calculations revised as at April 2005], Forest Research, Contract Report, April 2005).

7.2.5 Source-specific recalculations

New proportions of area by NEFD regime are used to weight the carbon yield in the 2003 inventory. Area data and carbon yields underlying the models in recent reports are similar to those used previously.

Backcasting and recalculation of 1990-2002 values have been included as in previous years i.e. previous estimates have been replaced with those from the current model. The process of using the forest models to time-shift the forest estate forwards to represent future forest growth and forest management, and backwards to improve historical estimates, is performed to minimise errors. As the estimation of carbon stocks is continuously being improved, both past and future years are re-calculated. The difference in net managed forest CO2 removals between this and the previous inventory is generally of the order of less than 2%.

Land previously reported as non-forest scrubland is classified as grassland as per GPG-LULUCF definitions.

7.2.6 Source-specific planned improvements

New Zealand's development of the CMS and the NZCAS will enable New Zealand to complete the time-series in the LULUCF inventory, and reduce the uncertainty by using country-specific emission and removal factors and UNFCCC category specific activity data. Details of the research are included in Annex 3.2.

7.3 Cropland (CRF 5B)

7.3.1 Description

The 2003 inventory is the first inventory where New Zealand has reported emissions and removals from cropland remaining cropland and from land converted to cropland. In 2003, the net CO2 removals were 583.87 Gg CO2. Cropland is a key category for New Zealand.

Cropland includes all annual and perennial crops as well as temporary fallow land. Annual crops include cereals, oil, seeds, vegetables, root crops and forages. Perennial crops include orchards, vineyards and plantations except where these lands meet the criteria for forest land.

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

7.3.2 Methodological issues

Emissions and removals have been calculated using IPCC Tier 1 emission and removal values and activity data from the LCDB analysis described in section 7.1 and Annex 3.3. To align with the methodologies provided in GPG-LULUCF, cropland was partitioned into annual and perennial cropland for the UNFCCC inventory.

Cropland remaining cropland

Living biomass: as per GPG-LULUCF, the change in biomass is only estimated for perennial woody crops. For annual crops, increase in biomass stocks in a single year is assumed equal to biomass losses from harvest and mortality in that same year - thus there is no net accumulation of biomass carbon stocks.

Values for the biomass accumulation rate (2.1 t C ha-1 yr-1) in perennial vegetation and biomass carbon loss (63 t C ha-1) are from GPG-LULUCF Table 3.3.2. New Zealand is using the values for a temperate climate (all moisture regimes) as this is the default regime most applicable to New Zealand. The LCDB analysis cannot provide information on areas of perennial vegetation temporarily destocked, therefore no losses in C stock can be calculated.

Dead organic matter: the GPG-LULUCF state that there is not sufficient information to provide a basic approach with default parameters to estimate carbon stock changes in dead organic matter pools in cropland remaining cropland. The notation NE is used in the CRF tables.

Soil carbon: to provide a Tier 1 estimate, New Zealand uses the IPCC default method for mineral soils (equation 3.3.3 of GPG-LULUCF). Mineral soils comprise 99.93% of New Zealand soils. This equation compares the soil organic carbon stock in the inventory year, with the soil organic carbon stock in T years prior to the inventory. New Zealand uses the IPCC default value of 21 years for T. The soil organic carbon stock is calculated from a reference carbon stock multiplied by the three land use and management factors shown in Table 7.1.3.1.

Changes in soil carbon stock are caused by changes in the land use and management factors (FLU, FMG and FI). Within the cropland category, the LCDB does not provide sufficient information to determine whether there has been a change in land use and management in the 20 years prior to the inventory. Therefore for cropland remaining cropland the values for FLU, FMG and FI are considered to be constant and there is no net change in carbon stocks in soils e.g. (83*0.82*1*1.16)-(83*0.82*1*1.16))*Area)/20 = 0. The values for FLU, FMG and FI are from Table 3.3.4 in GPG-LULUCF.

Land converted to cropland

Living biomass: the Tier 1 method is the same approach for all conversions and provided in equation 3.3.8 of GPG-LULUCF. The calculation is based on multiplying the annual area of land converted to cropland, by the carbon stock change per area for that type of conversion and including changes in carbon stocks from one year of cropland growth.

At Tier 1, carbon stocks in biomass immediately after conversion are assumed to be zero i.e. the land is cleared of all vegetation before planting crops. To complete the Tier 1 analysis, New Zealand has selected from default parameter values provided in GPG and country specific values where possible. These are shown in Tables 7.1.3.2 and 7.1.3.3.

Dead organic matter: GPG-LULUCF states that there is not sufficient information to provide a basic approach with default parameters to estimate carbon stock change in dead organic matter pools in land converted to cropland. The notation NE is used in the CRF tables.

Soil carbon: New Zealand has followed the method outlined in GPG-LULUCF (3.3.2.2.1.1). For Tier 1, the initial soil C stock is determined from the same reference soil C stocks used for all land uses, together with stock change factors (FLU, FMG, FI) appropriate for the previous land use (refer section 7.1.2.3 in this inventory).

N2O emissions: these emissions are from mineralisation of soil organic matter resulting from conversion of forest land, grassland, settlements or other land to cropland. New Zealand uses the method outlined in GPG-LULUCF equations 3.3.14 and 3.3.15. The input parameters 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 emissions factor for calculating emissions of N2O from N in the soil. The global default value of 0.0125 kg N2O-N/kg N is used.
  • C:N ratio: the default ratio of C to N in soil organic matter (15) is used.

7.3.3 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 etc. It is the uncertainty in the IPCC default variables that dominates the overall uncertainty in the estimate provided by New Zealand. The combined effect of uncertainty in cropland is estimated at ±75% (95% CI).

Variable Uncertainty (95% CI)

Uncertainty in cropland remaining cropland

± 75%

LCDB1 (user accuracy 93.9%)

± 6%

LCDB2 (assumed to be equal to LCDB1)

± 6%

Uncertainty in biomass accumulation rates

± 75% (GPG-LULUCF table 3.3.2)

Uncertainty from land converted to cropland

± 75%

Carbon stocks in previous land use

± 75%

Estimated uncertainty in land management factors

± 12% (GPG-LULUCF table 3.3.4)

7.3.4 Category-specific QA/QC and verification

No specific QA/QC and verification were used for cropland. Sector-level procedures are described in section 7.2.4 - forest land.

7.3.5 Category-specific recalculations

There are no recalculations for this category.

7.3.6 Category-specific planned improvements

No specific improvements are planned for cropland. Sector-level improvements resulting from the NZCAS are described in section 7.2.6 - forest land.

7.4 Grassland (CRF 5C)

7.4.1 Description

Grasslands in New Zealand can vary greatly in their degree and intensity of management, ranging from the extensively managed rangelands of the South Island high country, to low producing grasslands with woody vegetation cover, to the intensively managed dairy pasture in the Waikato and Taranaki regions. Grasslands generally have vegetation dominated by perennial grasses, with grazing as the predominant land-use, and are distinguished from forest land by having a woody vegetation cover of less than the threshold used in the forest definition. In 2003, the net emissions from grassland were 688.02 Gg CO2. These emissions are from the sub-category 'land converted to grassland'.

7.4.2 Methodological issues

Grassland remaining grassland

Living biomass: in GPG-LULUCF, the Tier 1 assumption is no change in living biomass. The rationale is that in grassland where management practices are static, biomass carbon stocks will be in an approximate steady-state where carbon accumulation through plant growth is roughly balanced by losses. New Zealand has reported NA in the CRF tables because the activity occurs but there are no removals or emissions associated with it.

Dead organic matter: no estimate is calculated as GPG-LULUCF state that not enough information is available to develop default coefficients for estimating the dead organic matter pool. For Tier 1 and 2 methods, changes in dead organic matter and inorganic carbon stocks should be assumed to be zero.

Soil carbon: to provide a Tier 1 estimate, New Zealand uses the IPCC default method for mineral soils (equation 3.4.8 of GPG-LULUCF). As noted in previous sections, mineral soils cover 99.93% of New Zealand (Tate et al., 2004). The LCDB analysis used in the 2003 inventory does not provide sufficient information to determine whether there has been a change in land-use and management in grassland for the 20 years prior to the inventory. Therefore for areas of grassland remaining grassland, the values for FLU, FMG and FI are considered to be constant and consequently the calculation shows there is no net change in carbon stocks in soils.

Liming of grassland: the calculation for CO2 emissions from the liming of grassland soils is included in worksheet 5.5. The calculation is based on the total amount of limestone sold (provided by Statistics New Zealand) and a carbon conversion factor from limestone to carbon. New Zealand uses the IPCC (1996) default value of 0.12 for carbon conversion.

Land converted to grassland

Living biomass: New Zealand has applied the GPG-LULUCF Tier 1 method where the amount of carbon removed is estimated by multiplying the area converted annually by the difference between average carbon stocks in biomass prior to and following conversion, and accounting for carbon in biomass that replaces cleared vegetation. Pre-conversion stocks and annual growth figures are tabulated in Tables 7.1.3.2 and 7.1.3.3. Carbon stocks in biomass immediately after conversion are assumed to be zero.

Dead organic matter: no Tier 1 methodology is provided in GPG-LULUCF.

Soil carbon: land conversion to grassland can occur from all land-uses. In New Zealand, the primary change into grassland is from forest land to grassland. New Zealand uses the methodology outlined in GPG-LULUCF (3.4.2.2.1.1). For Tier 1, the initial (pre-conversion) soil C stock is determined from a reference soil C stock together with stock change factors (FLU, FMG, FI) appropriate for the previous land-use as well as for grassland use. The stock change factors used by New Zealand are tabulated in Table 7.1.2.1.

7.4.3 Uncertainties and time-series consistency

It is the uncertainty in the IPCC default variables that dominates the overall uncertainty in the estimate provided by New Zealand. The combined effect of uncertainty in grassland is estimated at ±75% (95% CI).

Table 7.4.3.1 Uncertainty in emissions and removals from grassland

Variable Uncertainty (95% CI)

Uncertainty in grassland remaining grassland

± 75%

LCDB1 (user accuracy 93.9%)

± 6%

LCDB2 (assumed to be equal to LCDB1)

± 6%

Uncertainty in biomass accumulation rates

± 75% (GPG-LULUCF table 3.4.2)

Uncertainty from land converted to grassland

± 75%

Carbon stocks in previous land use

± 75%

Estimated uncertainty in land management factors

± 12% (GPG-LULUCF table 3.3.4)

7.4.4 Category-specific QA/QC and verification

No specific QA/QC and verification were used for grassland. Sector-level procedures are described in section 7.2.4 - forest land

7.4.5 Category-specific recalculations

There are no recalculations for this category.

7.4.6 Category-specific planned improvements

No specific improvements are planned for grassland. Sector-level improvements resulting from the NZCAS are described in section 7.2.6 - forest land.

7.5 Wetland (CRF 5D)

7.5.1 Description

GPG-LULUCF defines wetlands as "land that is covered or saturated by water for all or part of the year (e.g. peat land) and that does not fall into the forest land, cropland, grassland or settlements categories. It includes reservoirs as a managed subdivision and natural rivers and lakes as unmanaged subdivisions". New Zealand has categorised LCDB land cover classes for lakes, rivers and estuarine open water into the unmanaged wetlands category (Annex 3.3). Other LCDB classes e.g. herbaceous freshwater vegetation, commonly associated as a wetland in New Zealand, have been categorised as grassland following the GPG-LULUCF definitions. In 2003, the net emissions were 0.72 Gg CO2. These emissions are from the subcategory 'land converted to wetland'. Wetlands are not a key category for New Zealand.

7.5.2 Methodological issues

Wetland remaining wetland

A methodology for this category is not covered in GPG-LULUCF but is addressed in Appendix 3a.3 'Wetlands Remaining Wetlands: Basis for future methodological development'. The appendix covers emissions from peat land and flooded land. GPG-LULUCF defines flooded lands as "water bodies regulated by human activities for energy production, irrigation, navigation and recreation etc., and where substantial changes in water area due to water level regulation occur. Regulated lakes and rivers, where the main pre-flooded ecosystem was a natural lake or river, are not considered as flooded lands".

New Zealand has not reported emissions from flooded land because of a lack of data i.e. the LCDB does not separate out regulated water bodies where substantial changes in water area occur, and because the majority of New Zealand's hydro electric schemes are based on rivers and lakes where the main pre-flooded ecosystem was a natural lake or river. The CRF tables for LULUCF do not require Parties to prepare estimates for this category (footnote 3, CRF Table 5.)

Land converted to wetland

New Zealand has applied the GPG-LULUCF Tier 1 methodology for estimating the carbon stock change due to land conversion to flooded land (GPG-LULUCF equation 3.5.6). This method assumes that the carbon stock of land prior to conversion is lost in the first year following conversion. The carbon stock of the land prior to conversion is documented in Table 7.1.3.2. In Tier 1, it is assumed that the carbon stock after conversion is zero.

GPG-LULUCF does not provide guidance on carbon stock changes from soils due to land conversion to flooded land. Emissions of non-CO2 gases from land converted to flooded land are covered in Appendix 3a.3 of GPG-LULUCF but are not reported (note 3, CRF Table 5).

7.5.3 Uncertainties and time-series consistency

Uncertainties are estimated as ±75% based on the uncertainty for Tier 1 grassland carbon stocks (GPG-LULUCF Table 3.4.2) lost during conversion to wetland.

7.5.4 Category-specific QA/QC and verification

No specific QA/QC and verification were used for wetland. Sector-level procedures are described in section 7.2.4 - forest land.

7.5.5 Category-specific recalculations

There are no recalculations for this category.

7.5.6 Category-specific planned improvements

No specific improvements are planned for wetland. Sector-level improvements resulting from the NZCAS are described in section 7.2.6 - forest land.

7.6 Settlement (CRF 5E)

7.6.1 Description

This land-use category is described in GPG-LULUCF as including "all developed land, including transportation infrastructure and human settlements of any size, unless they are already included under other land-use categories". Settlement includes trees grown along streets, in public and private gardens, and in different kinds of parks where the parks are associated to urban areas. In 2003, the net emissions from settlements were 97.2 Gg CO2. These emissions are from the subcategory 'land converted to settlement'. Settlement is not a key category for New Zealand.

New Zealand has categorised the applicable LCDB land cover classes into the settlement category (Annex 3.3). The LCDB analysis showed that there was 214.84 kha of settlement remaining settlement from 1997 to 2002 with a net gain of 5.46kha (Table 7.1.1). The largest single category change in area was from high producing grassland converted to settlement, averaging 1.0 kha yr-1.

7.6.2 Methodological issues

Settlement remaining settlement

GPG-LULUCF provides a basic method for estimating CO2 emissions and removals in settlements remaining settlements 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. New Zealand does not have this level of activity data available. The CRF Tables for LULUCF do not require Parties to prepare estimates for this category (note 3, CRF Table 5.)

Land converted to settlement

The fundamental equation for estimating change in carbon stocks associated with land-use conversions is the same as applied for other areas of land use conversion e.g. land converted to cropland and grassland. The carbon stock of the land prior to conversion is documented in Table 7.1.3.2. The default assumptions for a Tier 1 estimate are that all living biomass present before conversion to settlements will be lost in the same year as the conversion takes place, and that carbon stocks in living biomass following conversion are equal to zero.

7.6.3 Uncertainties and time-series consistency

Uncertainties are estimated as ±75% based on the uncertainty for Tier 1 grassland carbon stocks (GPG-LULUCF table 3.4.2).

7.6.4 Category-specific QA/QC and verification

No specific QA/QC and verification were used for settlement. Sector-level procedures are described in section 7.2.4 - forest land.

7.6.5 Category-specific recalculations

There are no recalculations for this category.

7.6.6 Category-specific planned improvements

No specific improvements are planned for settlement. Sector-level improvements resulting from the NZCAS are described in section 7.2.6 - forest land.

7.7 Other land (CRF 5F)

7.7.1 Description

'Other land' is defined in GPG-LULUCF as including bare soil, rock, ice, and all unmanaged land areas that do not fall into any of the other five land-use categories. The other land category is included in New Zealand's land area for checking overall consistency of land area and tracking conversions to and from other land. In 2003, the net emissions from other land were 31.6 Gg CO2. These emissions are from the subcategory 'land converted to other land'. Other land is not a key category for New Zealand.

7.7.2 Methodological issues

Other land remaining other land

Change in carbon stocks and non-CO2 emissions and removals in unmanaged 'other land remaining other land' do not need to be assessed under GPG-LULUCF. No guidance is provided in GPG-LULUCF for other land that is managed.

Land converted to other land

Living biomass: the fundamental equation for estimating change in carbon stocks associated with land-use conversions is the same as applied for other areas of land use conversion e.g. land converted to cropland and grassland. The carbon stock of the land prior to conversion is documented in Table 7.1.3.2. The default assumptions for a Tier 1 estimate are that all living biomass present before conversion to other land will be lost in the same year as the conversion takes place, and that carbon stocks in living biomass following conversion are equal to zero.

Soil carbon: New Zealand uses the IPCC methodology outlined in GPP LULUCF (equation 3.7.3). For Tier 1, the initial (pre-conversion) soil C stock is determined from reference soil C stocks together with stock change factors (Table 7.1.2.1) appropriate for the previous land use. New Zealand uses a reference C stock of 83 t C ha-1. Soil carbon stocks in the inventory year are zero for land converted to other land.

7.7.3 Uncertainties and time-series consistency

Uncertainties are estimated as ±75% based on the uncertainty in carbon stocks lost during the conversion to other land e.g. GPG-LULUCF Table 3.4.2.

7.7.4 Category-specific QA/QC and verification

No specific QA/QC and verification were used for other land. Sector-level procedures are described in section 7.2.4, forest land.

7.7.5 Category-specific recalculations

There are no recalculations for this category.

7.7.6 Category-specific planned improvements

No specific improvements are planned for other land. Sector-level improvements resulting from the NZCAS are described in section 7.2.6, forest land.