3 Forest management

3.1 Specific activities considered, and effects not covered

The "forest management" activity is used to describe the silvicultural regime employed for plantations in New Zealand. When considering Article 3.4 the term is implicitly used in conjunction with forest established prior to 1 January 1990. Human-induced factors affecting both the total carbon stock at harvest and the long-term average stock over a rotation, include species used, initial stocking, fertilisation, pruning/thinning regime, and rotation length. Livestock displacement effects on emissions, methane sinks in forest soils, and nitrous oxide emissions, will all be significant, but as noted in the Introduction are more naturally captured under Annex A and Article 3.1 of the Protocol.

In terms of carbon accounting, plantation forest management is far more important than managed indigenous forests which are a very small part of commercial forestry in New Zealand. The 1996 Revised IPCC Guidelines (IPCC, 1997) state "Natural, unmanaged (for wood products) forests are not considered to be an anthropogenic source or sink, and are excluded from the calculations". But it is recognised that few forests are truly undisturbed, and even undisturbed forests may be carbon sinks over many centuries (Schulze et al 2000). In principle, implementation of Article 3.4 could take such changes into account if they are human-induced.

Selective logging on a sustainable yield basis may not affect carbon stocks, although ceasing this activity is likely to result in a slight increase in stocks, but this will occur slowly in most cases. Roundwood removals from indigenous forests in 1997 were 0.7% of the total wood extraction (MAF Statistics 1997) and since then have declined substantially. The implications of possum damage to indigenous forest is discussed in section 7.3 and in the rest of this section the indigenous forest estate in New Zealand will be excluded.

3.2 Effects on non-CO2 greenhouse gases

The biggest impact on non-CO2 greenhouse gases of establishing new forest is the removal, or partial removal, of grazing animals. If this results in a net decrease in animal numbers, i.e. does not represent a displacement of animals to more intensive management over a smaller area, then the reduction in methane emissions can be significant. Nitrous oxide emissions are likely to be reduced because of lower nitrogen inputs, drier conditions and better drainage of forests compared to pasture.

Conversion to forestry may increase methane removal by soils. A summary of methane removal in soils is given in section 7.6.

In terms of forestry per se, non-CO2 impacts would appear to be of minor importance compared to the carbon stock changes in the forest biomass. Following conversion from pasture to forest, the carbon equivalent of non-CO2 greenhouse gas reductions will be in the same direction as that due to carbon stock increases. However, the effect of genuinely reduced non-CO2 emissions, i.e. those which are not merely displaced to other areas, will continue to accumulate over the long-term whereas the terrestrial carbon stocks will saturate.

Pasture planted in pines can continue to be grazed for a substantial part of the rotation. This is common practice and is generally beneficial to the forestry operation. Livestock are excluded for the first 3-5 years (to prevent tree damage) but, even at age nine, pasture productivity is still some 76% of open pasture (Knowles et al 2000). Averaged over the first rotation, forests on ex-farm sites could be expected to carry perhaps a quarter of the stock units of unplanted pasture. Second rotation sites are usually not grazed.

3.3 Definitional issues

There are no definitional issues unique to this activity. If a baseline is required, it would be possible to define that from previously published analyses of the carbon stocks in plantations, and the assumptions used in those scenarios.

3.4 Socio-economic factors influencing the activities considered

There is obviously no opportunity to increase the area of pre-1990 forests (although this area can be reduced by deforestation). There is, however, an opportunity to increase the average carbon density of the existing area. The rotation length of most P. radiata plantations could be increased if there was sufficient economic incentive to do so. Of lesser importance is the ability to hold higher stocking rates throughout a rotation and to dispense with pruning. Local socio-economic factors have little influence over management decisions in pre-1990 forests, since the vast majority are owned by major international forest companies whose management decisions are based on large areas and the trends in demand for wood with particular characteristics. A carbon trading scheme, which provides an incentive to enhance carbon density in existing forests, could influence that situation.

A system of carbon credits for forest establishment (under Article 3.3) will tend to increase land prices. These land price changes are likely to affect all rural land, including the land underneath the pre-1990 forests. Thus owners of such forests may see the asset value of their land rise, without a commensurate increase in revenue. Their rate of return would decline, making investment in existing forestry companies less desirable, and the New Zealand forest industry less competitive compared to countries with lower land prices. There is also a concern that widespread planting, both nationally and internationally, driven by carbon credits, may devalue forestry as an investment because of the expectation of large volumes of wood coming on-stream in future years.

Implementation of Article 3.3 without Article 3.4 creates an unbalanced situation with respect to the management of forests. A pre-1990 forest estate would incur debits for any parts deforested but, without Article 3.4, there is no mechanism to obtain credits for enhanced carbon sequestration that may occur in other parts of the estate.

Any system that places a value on carbon in standing stocks of trees will tend to alter management regimes. Forests are more likely to be maintained for higher stockings throughout the rotation, and rotations to be lengthened. Pruning is likely to be reduced, and there may be a swing to long-rotation high-volume species like Douglas-fir. This will impact on the current forestry infrastructure, which has in part become dependent on a supply of clearwood from pruned radiata pine stands.

3.5 Applicability of a land-based accounting system

There is no problem implementing a land-based accounting system. Most forests have good maps, where Net Stocked Area is calculated to the nearest 0.1 ha, and matched to legal land ownership titles. GIS and GPS systems are widely used.

Accounting systems for Article 3.4 could be based on inventories (i.e. stocks) rather than flows. This is in accordance with the wording of Article 3.3, and is a transparent and verifiable method that does not rely to the same extent on simulation models, or their validation. Forest inventory techniques are routinely employed in New Zealand for normal commercial operations and, with little modification, can be adapted to assess total biomass carbon rather than stemwood volume.

Even if simple inventory techniques are employed to assess carbon stocks (and thereby changes in carbon stocks), a scheme that involves continuous monitoring of "Kyoto-compliant forests" may eventually be overwhelmed by transaction costs. This is because the benefits of carbon sequestration are limited by the maximum carbon density (MgC ha-1)4 that vegetation can support, whereas a perpetual monitoring scheme implies perpetual costs. To overcome this dilemma, Forest Research has proposed an accounting system that may involve as few as three inventories. Carbon credits are exchanged up to a maximum level, which is the long-term average carbon density of the forested landscape. The credits represent the increase in carbon density resulting from afforestation, and there must be an obligation to maintain forest cover in perpetuity.

Implementation of Article 3.4 could, in principle, simplify a land-based accounting scheme for forests by treating areas planted before or after 1990 similarly. However, any requirement to discount credits claimed under Article 3.4, as incorporated in some policy proposals, would require the distinction between pre- and post-1990 plantings be maintained throughout the accounting system. In that case the necessary carbon accounting would not be simplified and would have to cover a larger area.

3.6 What data, publications and reviews are available

Excellent data are available on the growing stock and harvest removals from the pre-1990 forest estate. Combined with national carbon yield tables, the growing stock (stem volume) can be converted to carbon content in live and dead trees and the forest floor. The national carbon inventory is prepared annually for the plantation forest estate, separating the stocks in pre- and post-1990 forests (e.g. Marshall et al 2000). Usually the carbon stocks are separated in broad categories e.g. stem, crown, and below-ground (roots), but finer resolution is possible if required. Soil carbon has been excluded from the majority of calculations of forestry carbon budgets in New Zealand.

Forest Research staff, both existing and former, have been researching the greenhouse implications of forestry for well over a decade and have established considerable international credibility in this area. New Zealand forestry methods, data and analysis have been reported in peer reviewed publications and international conferences and used in policy discussions. New Zealand has a highly simplified plantation forest system that is well researched and intensively monitored for normal commercial reasons. We are therefore able to quantify the carbon implications of our commercial forestry with high confidence.

3.7 Sampling techniques and issues

Plantation forests have replaced a variety of pre-existing land cover types but, since 1990, National Exotic Forest Description (NEFD) data show that more than 80% of new P. Radiata forests have been planted on improved or unimproved pastures. Current national reporting of carbon accumulation in plantation forests includes accumulation on the forest floor, but does not consider changes in the mineral soil. This reflects the fact that full biomass sampling on every site is costly, but forest inventory is a cheaper and well-established practice with widely accepted techniques for estimating stem volume from easily measured parameters such as height and diameter. Data (and models) are not as reliable for younger trees, given that there has been no need for this information in the past. Some improvements could be made to measure non-stem above-ground components, but this is unlikely to significantly improve stock change estimates.

There are concerns about the degree to which spatial variability in soil carbon may affect estimates of total stocks. Results from one site suggest that the spatial variability in stocks can exceed the temporal variability over a single rotation (Beets et al 2001). Site preparation practices are much less intensive now than in the past, but practices such as mounding could cause sampling problems, although are unlikely to affect carbon stocks. Results from sampling only near-surface (< 10cm), soils pre- and post-harvesting, are likely to be influenced by the harvesting method (e.g. ground-based logging tends to be more disruptive than hauler systems) and there is some evidence that total soil carbon to greater depths is less variable than near surface carbon (Beets et al 2001).

Using the soil Carbon Monitoring System(CMS), a national-scale comparison of soil carbon values for a range of pastures and pine sites, indicates lower mineral soil carbon under pine than pasture (Davis et al 1999). Assuming that these soil carbon values represent a steady state for the different land uses would imply that recent conversion to forestry has incurred soil carbon losses of 1.3 to 2.1 Tg y-1 ( Tate et al 2000a). If earlier afforestation occurred on land with lower initial soil carbon levels, this approach may overestimate the recent soil carbon loss rate. Alternative approaches have used studies of paired sites and direct measurement of changes over time, but less data is available in these cases.

A recent review of the affects of afforestation on soil carbon (Polglase et al 2000), based on studies using paired plots representing the two land uses (space-for-time substitution), chronosequence studies or repeated sampling over time at one site, demonstrated a wide variation in both the direction and magnitude of change. The main reason was the wide variation in climates, soil types, plantation productivity, previous land use and methodologies. Overall, there was a definite trend for initial soil carbon loss (up to about 40 y from planting), followed by soil carbon accumulation in older stands.

Figure 3.1. Soil carbon content by depth increment from intensive paired-site studies (Scott et al 1999).

wpeD38.jpg (24856 bytes)

The error bar represents the standard error and "Total" refers to the sum of O horizons and mineral soil carbon to 0.5m depth (0.3 m for Ngaumu).

Results from several published (Scott et al 1999, Ross et al 1999, Saggar et al 2001a) and unpublished (KR Tate, personal communication) New Zealand studies of the pasture-to-pine transition also show a downward trend in mineral soil carbon (Figure 3.1). Using paired site data (8 sites sampled to a depth of 0.3 m), and average annual changes in plantation forest area from 1990-1999, the estimated decline in soil carbon under pine ranges from 0.4 to 1 Tg y-1.

Studies to a depth of 0.1 m appear to show lower loss rates. E.g. a study on farm-forestry sites (Giddens et al 1997) suggested little or no change for some paired sites (0-0.1 m depth) and an unpublished analysis of over 30 pasture vs pine paired sites suggests that a loss of 4 MgC y-1 occurs in the upper 0.1 m of soil over 20 years with little or no loss subsequently (Murray Davis personal communication). This latter result would correspond to total loss of soil carbon due to recent afforestation of about 0.2 TgC y-1 in the upper 0.1 m and is not inconsistent with the range estimated by Scott et al (1999) given the smaller depth sampled.

Overall, these reductions in mineral soil carbon may be approximately balanced by the carbon that accumulates in the forest floor (Scott et al 1999); further losses of soil carbon from forest harvest effects may occur, according to results from direct measurement (Arneth et al 1998) and modeling (Tate et al 1999). Over the longer-term, soil carbon should stabilize as the readily decomposable organic matter derived from pasture inputs declines and more recalcitrant organic matter from the pine increases. Tree roots penetrate deeper than the 30 cm horizon normally sampled for soil carbon measurements and release carbon to the soil and atmosphere when they die (e.g. after harvest) and decay. Phillips and Watson (1994) observed that 20% of root biomass in 14-year old P. radiata was deeper than 20 cm. However, root biomass is a small fraction of total below-ground carbon.

In summary, there is consistent evidence for some loss of soil carbon after establishment of pine forest on pasture, possibly limited to the first 20 years. While this loss appears to be small compared to above ground carbon stocks, its magnitude is not likely to be further resolved until a better understanding of the processes controlling the changes is achieved and more data becomes available.

3.8 What process models are available to complement data

An existing model, DRYMAT, has recently been extended to incorporate environmental variables, such as solar radiation, rainfall, evapotranspiration, site fertility, severity of defoliation from diseases, and UMCY5. This model is being used to predict growth based on knowledge of biological processes. Once this or similar models are adequately tested they could be used to study the effect of current spatial variation in climate as a surrogate for longer term climate change on forest productivity and carbon sequestration. Such models are also available for scenario analyses.

An additional model in use, "C_change", is a hybrid of the physiological model (DRYMAT) attached to empirical growth models that are widely used by industry.

3.9 What regional and national scale inventories are available

The current data on the exotic plantation estate is collected by annual surveys of forest owners, and compiled in the National Exotic Forest Description and (every 5 years) into national and regional wood supply forecasts (MAF, 2000). The area estimates of plantation extent could be verified via remote sensing techniques currently proposed for the Land Cover Database (LCDB), but this may not be able to differentiate between stands established before or after 1990.

3.10 What is known of 1990 carbon stocks and current trends

The most recent estimate of 1990 above-ground carbon stocks in all managed forests is 113.2 TgC (Marshall et al. 2000). Even though statistics, methods and researchers have mostly changed since 1990, the figure is almost identical to that first proposed by Maclaren & Wakelin in 1991 (their figure was 113 TgC). No way has yet been devised to calculate error limits, but the closeness of the results engenders confidence.

Figure 3.2. Annual sequestration - total plantation estate, new land planting

graph2.gif (8075 bytes)

As at 1st April 1991, some 65.5% of the radiata pine resource was 15 years or younger (MOF Stats 1993). Given that the average rotation age of radiata pine is 28 years, the resource could be described as "immature". Growing stock averaged 181 m3 ha-1, but the average for a normalised radiata pine forest (from NEFD yield tables 1995) is 248 m3 or 37% more. As a first approximation then, pre-1990 stocks could be expected to increase to a sustainable long-term average of about 155 TgC - an increase of about 42 TgC. This is borne out by a recent report (Marshall et al., 2000), which used estate modelling to provide future estimates for the pre-1990 component.

Table 3.1. Change in above-ground carbon stocks of plantation forests (TgC) using the medium scenario of new-land planting

 

 1990 2000 2008 2010 2013 2020
Pre-1990 113.2 161.9 170.9 165.2 155.4 130.0
Post-1990 0 7.4 44.6 57.4 75.7 134.2
ALL 113.2 169.3 215.5 222.6 231.1 264.2

Changes in the total above-ground carbon stocks in plantation forestry and the corresponding annual sequestration rates are quite variable from year to year reflecting the past history of planting rates. Recent values and projections for the next 40 years are shown in Figures 3.2 and 3.3, and Table 3.1 gives corresponding values.

Figure 3.3. Total above-ground Carbon Stocks in TgC, for the total plantation estate, under a medium new land planting scenario, separated into pre- and post-1990 forests and showing the harvesting of pre-1990 forest.

Figure 3.3. Total above-ground Carbon Stocks in TgC

Values given in Table 3.2 provide complementary estimates of below-ground carbon. Rates of soil carbon change are based on post-1990 afforestation and although these cover a wide range there is general agreement that the value is negative. Because the accumulation rate of above-ground carbon is highly variable, and the uncertainties in soil carbon changes are large, no figure is given for the total rate of carbon stock changes in forested land.

Table 3.2. Estimated 1990 Carbon stocks and Recent Trends for Managed Forest Land

Estimated for 1990

Recent rates of change per yeara
Area (Mha) Vegetation C stock (Tg C) Soil C stock
(Tg C)		 Total C stock
(Tg C)		 Area
(Kha y-1)		 Vegetation C Stock (Tg y-1) Soil C stock (Tg y-1) Total C stock
(Tg y-1)
1.31 1132 983 211 +44 3.0 to 7.0
see Fig 3.2		 -0.4 to -1.04 see section 3.7

1May be covered already under Article 3.3. At this stage, no litter or slash is included in any LR estimates. Litter may account for additional ~ 1 TgC y-1 sequestered.
2Based on Marshall et al (2000).
3Does not include forest floor.
4Based on data for pasture-pine conversion for 8 paired sites, covering a range of soil types, sampled to 0.3 m.

It should be noted that changes in soil carbon can apply to forests that would be included under Article 3.3 as well as under Article 3.4. However, as shown in the following section, the dominant effect of including Article 3.4 in the forestry sector is to decrease the likely credits due to changes in above-ground carbon caused by age class distribution changes.

3.11 Projections for the first commitment period where feasible

Although there is a "long-term average" gain in pre-1990 stocks, that gain may be not be realised during the critical "commitment periods" (e.g. 2008-2012 and possibly 2013-2017). The delay of 18 years between 1990 and the start of the first commitment period (2008) changes the pre-1990 forests to a state of "over-maturity" - i.e. where older age-classes dominate.

In other words inclusion of Article 3.4 would have a negative effect on potential credits by including pre-1990 forests where the net effect is a significant decrease in above-ground carbon stocks, as shown below.

Table 3.3 Change in carbon stocks of plantation forests over the Commitment Period

Change over Commitment Period

Pre-1990

-15.5

Post-1990

31.1

ALL

15.6

(TgC)

3.12 Data and research needed to reduce uncertainties

There is a paucity of long-term research sites that can be used for the type of evaluation that includes soil carbon. It is of course very difficult to verify what would have happened to carbon stocks in the absence of a particular activity, but the C-change model has been found to provide excellent estimates of above-ground stock changes. Nevertheless, the empirical modelling system used to estimate carbon changes, at the levels of both forest stand and forest estate, is not conducive to calculating error limits.

There appear to be no measurements of carbon stocks, or their changes, in New Zealand landfills or buildings, and very few overseas measurements. Thus inferences about changes in wood product pools are largely based on simple models of flows and assumed product lifetimes or decay rates. It is not clear to what extent this approach would be classed as "verifiable" in terms of either Article 3.3 or 3.4.

4 Mg = 106 g = 1 tonne
5 Upper mid-crown yellowing, a nutritional imbalance in P. Radiata associated with high potassium and or low magnesium which is widespread in NZ plantation forests, particularly after periods of below normal rainfall.

Previous PageTable Of ContentsNext Page

Contact for Enquiries

MAF Information Services
Pastoral House
25 The Terrace
PO Box 2526
Wellington, NEW ZEALAND

Fax: +64 4 894 0721
Contact this person

 
Print this page Print this page