Annex A. Sequestration by forests and land use change

1. THE PROCESS OF CARBON SEQUESTRATION BY FORESTS AND LAND USE CHANGE

   Carbon dioxide is removed from the atmosphere by photosynthesis in plants (carboxylation) and stored in organic matter (biomass). It is returned to the atmosphere by respiration of the living biomass in plants and by decomposition of dead organic matter in soils (oxidation). At any one time there is a reservoir of fixed carbon in living plant material and in dead soil organic matter. The carbon reservoir on land is about half the dry mass of living and dead organic matter. Carbon sequestration is said to occur when that reservoir is increasing. By definition, sequestration will occur only when photosynthesis exceeds the sum of plant and soil respiration.

   The carbon reservoir in trees is much greater than that in herbaceous plants. Thus, planting trees on former treeless land will increase the reservoir of carbon in vegetation. Sequestration occurs while a stand of trees is growing and ceases when the stand is mature and no longer increasing in biomass. Growing trees to sequester carbon therefore gives a one-time benefit over a limited time period. By contrast, growing trees for energy, as a substitute for fossil fuels, gives a continuous offset-benefit, in addition to the one-time benefit of carbon sequestration in the biomass stands.

   The carbon reservoir in soil organic matter depends on the balance between the annual input of dead plant material (litter) and the annual loss of organic matter by decomposition. When an ecosystem is mature and in a stable (equilibrium) state, litter inputs and decomposition losses are in balance and there is no change in the amount of soil organic matter. Obviously, ecosystems differ in the amount of soil organic matter they contain when they reach a steady state. At one extreme, there are peatlands, where the respiration rate per unit of organic matter is very slow (because of anoxia) and so the mass of soil organic matter at equilibrium (when respiration loss equals inputs) is very large, although the annual input of dead organic matter may be small. At the other extreme, there are hot, dry areas where soil respiration proceeds rapidly, inputs of dead organic matter are small and the reservoir of soil organic matter is small.

   Any disturbance of vegetation or of the soil itself is likely to alter the balance between litter input to the soil and decomposition loss and so cause a change in the reservoir of carbon in soil organic matter. If the reservoir size decreases, it is a source of carbon. If the reservoir size increases, it is a carbon sink - that is, sequestration occurs. In general, any physical disturbance of the soil (cultivation, draining etc) is likely to accelerate soil respiration and create a carbon source, whereas any measures taken to increase vegetation productivity (eg by applying fertilizers or planting more productive vegetation) are likely to create a soil carbon sink (ie cause carbon sequestration). Losses of carbon from soils normally occur much faster than gains. Thus, cultivating former natural lands will cause a loss of carbon over 10-20 years before a new equilibrium is reached, whereas planting former cultivated land with trees will cause an increase in carbon over 50-200 years.

   It should be stressed that carbon sequestration in trees or soils is reversible. The benefit gained during the period of forest growth and/or increase in soil organic matter could be reversed by a subsequent change in land use, or by fire, pest outbreaks or climate change itself. Thus, not only is the period of carbon sequestration of limited duration, it also results in a burden to maintain the land use which originally created the enlarged carbon reservoirs (the forests and organically rich soils) and a risk that the carbon could be inadvertently returned to the atmosphere, making the situation worse in future.

2. CURRENT KNOWLEDGE AND RESEARCH IN HAND

   Parties to the Framework Convention of Climate Change are asked to report carbon sources (emissions) and sinks (removals) due to land use change and forestry, using IPCC guideline calculations. The categories are:
   5A. changes in forest and other woody biomass stocks - essentially the sinks resulting from increases in the biomass of new or existing forests;
   5B. forest and grassland conversion - mostly the sources of carbon resulting from deforestation;
   5C. abandonment of managed lands - essentially the sink created when vegetation regrows naturally on former agricultural land;
   5D. carbon emissions and removals from soils - which covers the soil sources and sinks in all the above three categories; and
   5E other.

   So far, over 20 countries have reported sinks in category 5A - mostly from forest plantations, but some also claim an increase in the biomass (standing stock) in existing forests. Thus, in the Second Communication to the UNFCCC, New Zealand claimed a sink in category 5A of 5.6 million tonnes of carbon per year (MtC/yr), Japan 23.0 MtC/yr, France 10.4 MtC/yr, Australia 6.0 MtC/yr and the UK 2.1 MtC/yr. Fewer countries have estimated carbon losses from deforestation - exceptions being Australia, which estimated a source of 33.5 MtC/yr, France a source of 3.7 MtC/yr and Japan a source of 0.2 MtC/yr. A few countries have small estimates for the sink created by the regrowth of vegetation on abandoned lands (category 5C) and even fewer countries have reported changes in soil carbon (category 5D).

   The UK has one of the most comprehensive budgets of carbon sources and sinks resulting from land use change and forestry, including soils. Table 1 lists the land use changes and 'natural' processes which are having a significant effect on the stores of carbon in biomass and soils in the UK, approximately for the year 1990. Using databases on soils, land cover, historic land use change and models of the change in carbon stocks when land is subject to different kinds of disturbance (afforestation, cultivation, urbanization etc), estimates have been made of the carbon sources and sinks due to different factors for the whole UK (Table 2). These sources and sinks are divided in Table 2 between those which are reported to the UNFCCC and those which are not. Note that, overall the UK terrestrial sources exceed the sinks, with an emission of 7.4-3.1 = 4.2 MtC/yr for categories reported under the UNFCCC or 8.8- 5.9 = 2.9 MtC/yr for all categories - but with an uncertainty of about ± 30%.

   In the UK, the carbon sink due to the accumulation of carbon in the 2 million ha of forests that have been planted since the 1920s (2 MtC/yr in trees and 0.6 MtC/yr in litter and soils) is much less than the carbon lost from soils as a result of cultivation and urbanization (6.4 MtC/yr) - much of the latter is the legacy of improving hill pastures and intensive agriculture (Table 2). Also, it will be noted that the gains and losses to/from soils are large compared with those to/from vegetation and that substantial changes are occurring as a result of 'natural' processes which are not reported to the UNFCCC.

   Table 2 shows an uncertainty of ±15% in the estimate of sequestration of carbon in forest biomass (2 MtC/yr). This is much less than the uncertainty attached to the other sources and sinks, especially those involving soils. The reason is that reliable data exist on changes in forest areas, and changes in carbon stocks in forests can be derived reasonably accurately using carbon accounting models. The basic input to these models is the change in stem-wood biomass, which is forecast for all productive forest areas using 'yield tables' and inventories. Several countries, including the UK, use dynamic carbon accounting models to estimate the change in carbon in forests (eg Canada, Germany, Finland and New Zealand) while others use an assessment of present stocks and fluxes based on forest inventories. The greatest source of uncertainty in most estimates is the fraction of biomass in roots and branches. Most forest carbon models also make some (less certain) estimate of the change in soil organic matter resulting from changes in the balance of litter input to the soil and soil respiration after planting trees. Finally, many models estimate changes in the store of carbon in wood products harvested from forests over several rotations.

Estimates of changes in soil carbon resulting from land use change can be made when the 'equilibrium' carbon (organic matter) content of soils is known for each land use category, when a land use change matrix can be constructed for the whole country based on successive surveys, and a rate of change in soil carbon can be assigned to each cell in the land use change matrix. Limited information from long-term plots (eg at Rothamsted) give estimates of the rapid loss of soil carbon when land is first cultivated and the slow build up in carbon when cultivated land is transferred back to permanent vegetation cover. However, considerable uncertainty exists and research is underway to more accurately assess changes in soil carbon. The problems are most acute when dealing with carbon rich peaty soils. One of the tools that is now available is the direct measurement of carbon dioxide fluxes over areas of land, using micrometeorological methods.

Every country will have a unique history and mix of land type and land use change and so could, in theory, develop a unique list of carbon sources and sinks due to changes in biomass and soil carbon, as in Table 2 (but with additional categories in many countries, such as deforestation, fire loss etc. ). As mentioned above, many countries have estimated a forest biomass sink but have no estimates for the other categories, many of which may be sources.

3. POTENTIAL SCOPE FOR THIS APPROACH TO BE TAKEN IN THE UK AND INTERNATIONALLY

   3.1 The UK

   Clearly, in the UK, there is some scope to enhance the carbon sinks on the left side of Table 2 and to diminish the carbon sources on the right side. However, it should be realized that all the quantities in Table 2 are small compared with the 150-160 MtC/yr emitted in the UK by burning fossil fuels.

   The current forest biomass carbon sink of 2 MtC/yr (ie the amount of carbon being sequestered by forests in trees, about 1% of fossil fuel emissions) could be maintained if the UK forest area continued to expand by about 30,000 ha per year and all harvested areas were restocked. If the existing forest area remained unchanged, the sink of 2 MtC/yr would decrease to zero by about 2020 as the whole UK forest area reached an approximate steady state. The sink would increase only if the rate of forest area expansion was faster than 30,000 ha/yr, which is unlikely without large shifts in land prices and incentives for agriculture and forestry.

   In order to sequester the 150-160 MtC/yr emitted from fossil fuels in the UK about 51 million ha of plantation forest would need to be planted - over twice the land area of the UK. The annual emission from an average car (about 1.1 tC/yr) is absorbed by about 0.4 ha of conifer forest or about 40 widely-spaced broadleaved trees (every year averaged from planting to maturity).

   The soil sink created by set-aside agricultural land in the UK could be expanded and enhanced by increasing the area and not employing rotational cultivation. The large soil source of 6.4 MtC/yr (with a large uncertainty) created by historic cultivation of organically rich soils will diminish over time and will be lessened by preserving areas of natural vegetation and discouraging hill land improvement for agriculture. The carbon losses from peatlands could be diminished by preserving wetlands, avoiding any peatland disturbance (for forestry, agriculture or building) and stopping peat extraction.

   The best estimate is that current trends in land use and forest planting will bring about an overall decrease in the UK soil carbon sources (emissions) by 2-3 MtC/yr over the next 20 years (ie the source of 7.4 MtC/yr in Table 3 is likely to decrease), while the sinks due mainly to forestry will be maintained around their current level (ie the sink of -3.1 MtC/yr in Table 2 will not change greatly).

   3.2 New Zealand and Europe

   As mentioned, countries differ hugely in the size of their forest carbon sink. At one extreme, New Zealand claims a current and future forest sink which exceeds its total fossil fuel emissions and even offers carbon credits for trading, while the Netherlands has a tiny forest carbon sink.

   In continental Europe, 14 countries have calculated their national forest carbon balance and all claim a sink due to new planting and/or an increase in the standing stock of biomass. In countries like Finland and Sweden, the forest sink may be equivalent to 40-50% of the national fossil fuel emissions whereas in most of the smaller nations with high population densities the sink is 1-15% of national fossil fuel emissions. The European average forest sink is about 9.5% of European emissions.

   As in the UK, there are limited opportunities in continental Europe to increase the current forest sink without major changes in land use policy. The sink due to increasing forest biomass occurs because timber harvesting is currently less than annual timber increment in the forest; this sink will saturate and so fall to zero.

   3.3 Global Assessment

   The 1995 IPCC report gives estimates of the global potential to sequester carbon by forest planting ('forestation' and agroforestry) based on the areas of land likely to be available (with socio-economic and population constraints), feasible planting rates, likely growth rates without intensive management and modelled fluxes of carbon in trees and soils (Table 3).

   For most high and mid latitude countries, the area of land actually available was equated with that technically suitable, or about 215 million ha. The rather large areas of land available in Canada was comprised of poorly stocked forest lands and marginal agricultural lands. In low latitude countries, the available land was only 6% of land deemed suitable, because of the needs of the growing population. The assumed global planting rates of 8.3 million ha/yr for plantations and 1.4 million ha/yr for agroforests were not unrealistic - and in fact were less than the actual rates in 1980-1990 in the tropics and in China.

   Globally, it was concluded that 345 million ha of land could be planted between 1995 and 2050, which would sequester 38 GtC (thousand million tonnes), with about 80% of this accumulating in aboveground tree biomass trees and 20% in soils and belowground biomass. The greatest fraction of this potential sequestration could occur in the tropics (60%).

   This level of sequestration (38 GtC) would absorb approximately 7.5% of the total fossil fuel emissions over the 1995-2050 period, according to the IS92a emissions scenario. Starting at zero in 1995, the annual rate of carbon sequestration would climb approximately linearly to about 1 GtC/yr by 2050.

   The IPCC also concluded that a further 20-50 GtC could be retained in vegetation in the tropics by encouraging natural regeneration and slowing the assumed IS92a rate of deforestation. Together, forestation, agroforestry, regeneration and slowed deforestation might offset 12-15% of the IS92a fossil fuel emissions from 1995 to 2050. Uncertainties were not estimated, but were known to be greatest for assumptions of land availability.

   3.4 Private Sector Projects

   In April 1995 the Berlin Conference of Parties to the UNFCCC gave support to a pilot programme of 'Activities Jointly Implemented' to promote individual projects between nations which reduced greenhouse gas emissions, including forestry offset projects. About 24 national offices have been established and about 15 projects have been approved in the forestry/agricultural sector. The USA, Norway, Sweden, the Netherlands and Australia have been most active in establishing forestry projects in Costa Rica, Russia, Mexico, Indonesia, Panama and Belize. Notably, the Dutch Electricity Generating Board, set up the FACE Foundation in 1990 (Forests Absorbing Carbon dioxide Emissions) supporting projects around the world which will sequester an estimated 31 Mt of carbon over the next 100 years.

   To date, UK companies and organizations have not made major commitments to forestry offset projects. However, AES Electric UK has initiated a project in Brazil to offset emissions from a coal fired power station at Barry, Wales; Tesco Plc carried out a pilot project to sell green petrol to plant trees in Uganda; and BP is a stakeholder in the Noel Kempff Mercado Climate Project in Bolivia. Also, companies have been set up to service any future industry requirements, including SGS Forestry, Carbon Storage Trust, Edinburgh Centre for Carbon Management, and Future Forests.

It is clear that individual companies see a marketing benefit in offsetting all or a large fraction of their emissions by planting trees in other (mostly tropical) countries. By so doing, they are making a prior claim on the available land. However, individual projects (rather than national programmes) pose problems of 'leakage' - ensuring that planting trees at one place does not promote deforestation elsewhere- and of establishing a baseline - ensuring that there has been sequestration which would not have occurred anyway. These problems are additional to those regarding equity, socio-economic development, biodiversity and security of the carbon sequestered.

4. THE KYOTO PROTOCOL AND IPCC SPECIAL REPORT

   In Article 3.3 of the Protocol there is the following agreement to include forestry...

   'The net changes in greenhouse gas emissions from sources and removals by sinks resulting from direct human-induced land use change and forestry activities, limited to afforestation, reforestation and deforestation since 1990, measured as verifiable changes in stocks in each commitment period, shall be used to meet the commitments [in 2008-2012] in the Article of each Party..'

   ..and in Article 3.4 there is the following declaration to consider land use emissions and removals more generally in future...

   'The Conference of Parties... shall...decide upon modalities, rules and guidelines as to how and which additional human-induced activities related to changes in greenhouse gas emissions and removals in the agricultural soil and land use change and forestry categories shall be added to, or subtracted from, the assigned amount for Parties included in Annex I [the reduction targets], taking into account uncertainties, transparency in reporting, verifiability, the methodological work of the IPCC...'

   Article 3.3 requires the definition of 'Kyoto forests', which are the small forest area subject to direct human-induced afforestation, reforestation or deforestation since 1900. The carbon sink or source is then the average annual increase or loss of carbon in Kyoto forests in the 'commitment period', 2008-2012. In the UK, if it assumed that the forest area expands in future at the same rate as in 1996, then 395,000 ha of new forests will be planted from 1990 to 2012, which will sequester an average of 0.5 MtC per year from 2008 to 2012. The UK Kyoto forest sink in 2008-2012 will therefore be about a third of the actual national forest sink (about 1.5MtC/yr, assuming it is maintained at 1990 levels; Table 2).

   In the UK, the forest sink is created almost entirely by afforestation on former non-forest land. In other countries it is not so straightforward to define what constitutes a forest (eg a designation or the presence of a minimal cover of trees) what is meant by reforestation, deforestation and human-induced change. It has been estimated that the EU Kyoto forest sink could vary over tenfold depending on definitions.

   These and other questions that need to be answered to 'operationalize' Article 3.3 of the Kyoto Protocol are the subject of a Special Report being prepared by IPCC for the year 2000, along with the even more difficult task of defining the scientific issues to be considered if soils and other land use changes are to be included in future commitment periods. The difficulty is to avoid unintended (perverse) outcomes or spurious sink claims, to be equitable to Parties and to have a verifiable system.

5. THE ENVIRONMENTAL SIDE EFFECTS

   5.1 Biodiversity

   In general, the diversity of plants and animals tends to decrease along the spectrum of land use from primary forest, regenerated forest, plantation forest to agricultural land. Conversely, biodiversity is likely to increase when converting agricultural land. to most types of forest. However, plant and animal species differ in their preferred habitats and the type, number and relative abundance of species depends on the type and diversity of habitats that are created. In tropical regions, biodiversity may be greatest in a landscape with a mixture of primary forest, regenerated forest and some plantations of native species.

   Trade offs between carbon storage and maintaining biodiversity are greatest when creating large areas of productive managed forest, especially monocultures of exotic species. Rapid carbon sequestration means high productivity, which demands high light interception, which suppresses the ground flora and limits other life forms; rapid forest establishment means that there are no long periods of recovery which provide habitats following natural disturbance; harvesting at the time which maximizes timber yield prevents the development of special habitats which occur in old growth forests; and plantations of single tree species are likely to have more limited and particular types of fauna and flora than natural mixed species stands. There are, however, management options to address the tradeoffs between production/carbon sequestration and biodiversity, such as altering felling unit sizes, edge lengths and creating a multi-aged patchwork of stands.

   5.2 Water use

   Afforestation can be expected to increase water use (the sum of transpiration and evaporation of water intercepted by the canopies) on land which previously had crops, grass or natural short vegetation. The reason is not increased transpiration - which can be less for trees than for short vegetation - but increased interception loss, especially if tree canopies are wet for a large proportion of the year. Interception losses are greatest from forests which have large leaf areas throughout the year. Thus, they tend to be greater for evergreen than deciduous forests and may be expected to be larger for fast-growing forests with high rates of carbon storage than for slow-growing forests. Consequently, afforestation with fast-growing conifers on non-forest land commonly decreases the flow of water from extensively forested catchments and can cause water shortage during droughts. Also, the replacement of natural broadleaved woodlands with conifer plantations is likely to increase water use by increasing interception loss. However, for short periods, afforestation activities can also increase peak flows as a result of tilling, draining and building roads and trails.

   In the dry tropics, forest plantations often use more water than short vegetation because the trees access water at greater depth as well as evaporate more intercepted water. Newly planted forests can use more water (by transpiration and interception) than the annual rainfall by mining groundwater. Extensive afforestation in the dry tropics can therefore have a serious impact on supplies of groundwater and river flows. However, it is less clear whether replacing natural forests with plantations, even with exotic species, increases water use in the tropics, when there is no change in rooting depth or stomatal behaviour of the tree species. In the dry zone of India, water use by Eucalyptus plantations is similar to that of indigenous dry deciduous forest - the annual water use of both forests is similar to the annual rainfall.

   In many regions of the world, decreased water use following deforestation can cause water tables to rise, bringing salt to the surface. In such situations, high water use by trees can be of benefit.

   5.3 Soil Quality

   When forests are converted to cultivated cropland, there is ample evidence that soil carbon and sometimes exchange capacities decrease. However, if the cleared land is not cultivated or degraded, soil carbon amounts may not change over decades and may depend more on the productivity of the vegetation (and hence nutrient supply and climate) than on vegetation type. Thus, the assumption, adopted in the 1996 IPCC guidelines, that the conversion of forest to non-forest decreases soil organic matter it is not always true.

   When cultivated lands, or soils previously low in organic matter, are afforested there can be substantial increases in amount of soil organic matters (eg in the Brecklands of East Anglia or following shifting agriculture in the tropics). However, if the land was previously peat or rich in organic matter, drainage and disturbance to establish trees accelerates decomposition, and the loss of carbon will offset and maybe eventually exceed the additional store of carbon created by growing trees.

   The introduction of forest operations such as harvesting and site preparation in areas that remain as forest, temporarily alter soils, but may do not have any long-term effects on soil organic matter amounts.

   5.4 Climate feedbacks

   Relatively high water use by forest compared with non-forest lands transfers more water to the atmosphere with potential effects on the local climate if the forest areas are very extensive. These effects are likely to be most important in the tropics. Thus, in the dry zone of India, forests over about 50 km² in extent are calculated to significantly increase humidity, lower temperature and increase local rainfall.

   The reverse effect, of decreased local rainfall and increased temperature as a result of extensive deforestation has also been calculated, most notably for Amazonia, where the predicted change in local climate could make conditions unsuitable for the subsequent regeneration of many rainforest species.

   5.5 Acidification

   When forests are planted on former non-forest lands in areas with high concentrations of polluted cloud water or of 'reactive' pollutant gases in the air (HNO₃, HN₃ and HCl), they increase the transfer of those pollutants from the atmosphere to the ground. This transfer can result in acidification of the soil and waters, and increased aluminium in waters, especially in areas with base-poor soils, having detrimental effects on salmonid fish, invertebrates, vegetation and perhaps the trees themselves. The risk of enhanced acidification is a well-known constraint on afforestation in parts of Europe, North America and Japan and may be a consideration in any country with high levels of acidifying air pollutants.

6. THE SOCIO-ECONOMIC SIDE EFFECTS

   Globally, wood consumption may exceed supply by 2050, assuming constant per capita wood use and a 2% annual volume growth increment. The greatest shortfall will be in the tropics, owing to the projected increase in population, driving the conversion of forest to agricultural land. Temperate zone needs are likely to be met, but there may be shortages of boreal industrial roundwood, depending on the forest response to climate change. Any increase in establishing forest plantations may help to overcome this shortfall. Increasing the growing stock of wood may also provide the opportunity to use that wood as a biofuel.

   Forests are valued for many goods and services, reflected in recent forest policy and institutional changes in many countries. In contrast, carbon is a single commodity and any form of carbon in trees and soils will suffice, as long as it is securely stored. If care is not taken, the promotion of carbon over other forest values could impoverish those who benefit from the diversity of products and non-carbon services. This problem will be greatest where livelihoods are at stake and where interests of forest-dependent people are poorly represented. In developing countries, land with a high carbon storage potential is often of high value to the most vulnerable groups, such as small farmers and the rural landless.

7. THE ECONOMIC COSTS

   In the UK, initial estimates of the cost of afforestation are £50-80/t carbon (tC). Costs in developing countries may be substantially less. The 1995 IPCC report gives estimates of the costs of implementing the 1995-2050 programme of global forestation and agroforestry outlined in Table 3. The unit costs were estimated to be $US 5-8/tC, depending on the region, based on country-specific marginal cost estimates for Brazil, USA, India, Central America and Thailand. The cumulative costs of mounting the global programme to sequester 7.5% of IS92a projected fossil fuel emissions to 2050 was $203 billion. Programmes to slow deforestation were much cheaper.

8. PRACTICALITY AND TIMESCALE

   In the UK, it is certainly practical to maintain the current forest carbon sink by further expanding forestry by up to 30,000 ha/yr for another decade, expanding the existing productive forest estate by about 15%. Beyond that, there may be increasing concern about the effects on water yield, landscape, biodiversity and acidification. It seems unlikely that the existing forest carbon sink in the UK could be substantially increased (ie by expanding the forest area by more than 30,000 ha/yr) without shifts in public opinion and land use policy. As mentioned, in the short term, carbon benefits may be more easily obtained by reducing the current rate of carbon loss from soils and peatlands (ie reducing the sources on the right of Table 2) than by accelerating the expansion of the forest area.

   Table 3 presents the global programme considered by the IPCC to be practical over the period 1995 to 2050, with realistic rates of planting. However, the IPCC estimates do not consider the impacts of projected changes in land use and climate change itself. They are therefore optimistic - less land may be available than assumed, and warming and shifts in rainfall may lead to losses of soil carbon and poor forest growth in some regions which now support thriving forests. DETR studies of the impacts of HadCM2 and HadCM3 climate projections suggest that many forest areas in the tropics and at warm temperate latitudes may no longer be able to support forests after about 2080 because of low rainfall and high temperatures. This prospect raises the issue posed at the beginning, that the creation of a large carbon store in trees and soils creates a risk of a large future emission if those stores are lost.

9. VERIFICATION

   The issue of verification is one being addressed in the IPCC Special Report on 'operationalizing' Article 3.3 of the Kyoto Protocol and advising the way forward on Article 3.4.

   The first requirement is to verify the area change in forests. In the UK, this is straightforward because there are negligible areas subject to deforestation (ie conversion of forest to non-forest land). It is sufficient then simply to determine the annual increase in forest area which can be assumed to be the area receiving planting grants. It is assumed that all clear-felled forests are replanted. But in countries where there is both afforestation and deforestation, it becomes imperative to have a spatially explicit map of the forest areas being considered. If the forest is defined in terms of canopy cover, it becomes possible to use remotely sensed imagery to determine the areas, but if the forest is defined in terms of land designation, then maps or ground sampling is required.

   The second requirement is to quantify a change in carbon stock in an area of forest between two dates. This is most commonly done by extrapolating inventory data or growth and yield models - knowing the species and site type. Some countries, including the UK and most Nordic countries have good, detailed forest inventories, which may be adequate to obtain reasonably accurate estimate of changes in carbon stock. Other countries, such as Canada and the USA would need to intensify their existing forest inventory systems. The US inventory currently takes over a decade to complete.

   Any claims of carbon sinks involving soils would need a comprehensive, detailed soils maps, estimates of soil carbon in each soil type, an historic land use change matrix for the whole country or region, and measurements of the rate of change in soil carbon when land is transferred from one land use to another. Few countries have this information at present, and the absence of historic land use change statistics (which are needed because of the long timescale of soil carbon changes) may make it difficult to quantify current soil carbon sources and sinks with any satisfactory level of accuracy. It may be recalled that, even in the UK, estimates based on detailed soils and land use change data have an uncertainty of ± 50% (Table 2).

10. NEXT STEPS/FURTHER WORK REQUIRED

    The IPCC Special Report will be influential in setting the agenda for including land use and forestry as means of sequestering carbon and so meeting commitments to limit the rise in atmospheric CO₂. The UK inventory shows the need to include soils in order to prevent unintended outcomes. This will demand better soils inventory data, better estimates of rates of change in soil carbon following land use change, improved matrix analysis of soils and land use statistics and the use of new remote sensing and flux measurement methods. The challenge will be to develop a range of carbon accounting systems for different country situations.

REFERENCES

Armstrong Brown, S., Rounsevell, M. D. A., Annan, J. D., Phillips, V. R. and Audsley, E.: 1996. Agricultural policy impacts on UK carbon fluxes. In: Instruments for Global Warming Mitigation: the Role of Agriculture and Forestry. Workshop in Trento, Italy, 1996. Soil Survey and Land Research Centre, Cranfield University, Silsoe , Bedfordshire, UK.

Brown, S., Sathaye, J., Cannell, M. G. R. and Kauppi, P. 1996. Management of forests for mitigation of greenhouse gas emissions pp 773-797. In: Climate Change 1995. Impacts, adaptations and mitigation of climate change: scientific-technical analyses. Contribution of Working Group II to the Second Assessment Report of the IPCC. Eds R. T. Watson, M. C. Zinyowera, R. H. Moss and D. J. Dokken. Cambridge University Press.

Cannell, M. G. R. and Milne R. 1995 Carbon pools and sequestration in forest ecosystems in Britain. Forestry 68, 361-378.

Cannell, M. G. R., Milne, R., Hargreaves, K. J., Brown, T. A. W., Cruickshank, M. M., Bradley, R. I., Spencer, T., Hope, D., Billett, M. F., Adger, W. N. and Subak, S. 1999 National inventories of terrestrial carbon sources and sinks: the UK experience. Climatic Change (in press).

Dewar, R. C. and Cannell, M. G. R. 1992 Carbon sequestration in the trees, products and soils of forest plantations: an analysis using UK examples. Tree Physiology 11, 49-71.

Intergovernmental Panel on Climate Change (IPCC). 1997,Greenhouse Gas Inventory Reporting Instructions. Volumes 1, 2 and 3. IPCC Technical Support Unit, Hadley Centre, Meteorological Office, Bracknell, UK.

Milne, R and Brown, T. A. W. 1997. Carbon in the vegetation and soils of Great Britain'. J. Environmental Manage. 49. 413-433.

Nabuurs, G. J., Paivinen, R., Sikkema, R. and Mohren, G. M. J. 1997. The role of European forests in the global carbon cycle - a review'. Biomass and Bioenergy, 13, 345-358.

Price, C and Willis, R. 1993. Time, discounting and the valuation of forestry's carbon fluxes. Commonwealth Forestry Review 72, 265-271.

Scole, D. L., Justice, C. O., Townshend, J. R. G. and Janetos, A. C. 1997. A land cover change monitoring program: strategy for an international effort.' Mitigation and Adaptation Strategies for Global Change 2, 157-175.