Comparative framework and Options for harmonization of definitions

Forest

At the international level, three widely used definitions of forest have been adopted by UNFCCC, CBD and FRA (Box 1.1). However, the CBD has not included the term `forest' in its Art. 2 (use of terms), and the definition used in this analysis is taken from the Report (of the Ad Hoc Technical Expert Group (AHTEG) on Forest Biological Diversity UNEP/CDB/SBSTTA 2001).

The most widely used definition is the one formulated in the FRA process. For instance, AHTEG refers to it as the `basic' definition (see annex 3).

Box 1.2 Definitions of Forests

The key features included in the various definitions are presented schematically in Table 1.3 where in the binary section `1' indicates the presence of a parameter in the definition and `0' its absence. The features are discussed in more detail in the next sections.

Table 1.3 Parameters of Definitions of `Forest'

Threshold Defining Stand Characteristics

The definitions of the UNFCCC, CBD and FRA are compatible. All of them are based on land use and tree cover. Regarding tree cover, all definitions set thresholds for minimum area, tree height and canopy cover. The CBD and FRA definitions have the same numerical values for thresholds, and they apply universally in all countries. The UNFCCC thresholds differ from these in that the Parties to Kyoto Protocol can establish the numerical values drawing on their national definitions within the indicated ranges.

There is not necessarily a great need for further harmonization because the national thresholds applied under the UNFCCC definitions are largely harmonized under the FRA process. A number of countries have either adjusted their national definitions or converted their national data to make them comparable with the FRA definition, and this process is expected to continue. It is also worth noting that the UNFCCC process requires that Parties must report on changes in their forest stock as they reported in the past to FRA, i.e. applying not their national definitions but the FRA ones.

On the other hand, it should be noted that, in the context of the UNFCCC, the countries have an interest to ensure that the choice of threshold values (e.g. adoption of FRA definitions) does not lead to exclusion of significant carbon stocks from carbon accounting. It is therefore necessary to ensure that the entire classification and accounting system is able to accommodate the change without causing distortions. For instance, accumulation of carbon stocks in areas not classified as forest should be taken into consideration under other concepts such as UNFCCC's `revegetation' or FRA's `other wooded land' and `other land'.

Definition of `Temporary'

A pivotal feature of both the FRA and UNFCCC definitions of forest is that temporarily unstocked forest areas are classified as forest provided that their land use remains forestry¹⁰. There are a number of reasons why the term `temporary' should be qualified.

Many lands which for legal or administrative reasons are classified as forest lands falling under forestry land use may not be covered with trees in a near future (or ever). On the other hand, there may be other ways than legal provisions or administrative decisions to ensure that the tree cover will be re-established and that forestry continues to be the land use. For example, existence of a management plan to reforest the land (soon) could be considered a qualifier, or that the tree cover is expected to expand to more than 10% of the crown cover and reach a minimum of 5 meters in height, if the area is brought under protection and not further disturbed by human intervention. However, the practicality of such options as qualifiers in connection with the term `temporary' would have to be carefully assessed.

The term `temporary' is present both in the UNFCCC and FRA definitions, but only the latter one is explicit on what it means; it is defaulted at roughly 10 years. The choice of default period is necessarily somewhat artificial given the wide range of conditions where forests regenerate. The reason why the UNFCCC lacks a definition is probably the highly varied national conditions for which an agreement on one definition would be difficult to reach.

The UNFCCC and FRA definitions are harmonized in the sense that data can theoretically be converted from one threshold to another. In practice, however, this may be difficult to achieve because data are usually insufficient to establish reliable conversion factors with regard to the period during which regeneration should occur. The other option is to standardize the default period¹¹, but it is unclear whether this is feasible or even desirable. The main benefit would be better convertibility of data which must be weighed against potential disadvantages such as disrupting established and agreed reporting patterns.

The CBD definition of forest does not include the concept `temporary', possibly because the context where it is applicable, i.e. unstocked areas, are not referred to in the definition. The rationale for this could be reviewed together with a discussion on `unstocked areas'.

Unstocked Areas

The FRA and UNFCCC definitions state that temporarily unstocked areas, are considered forest. The CBD definition does not explicitly mention these, but makes a reference to `young forest'. Lacking an explicit definition, it is unclear whether young forests are equal to unstocked forest or not. It would probably be rare to have a virtually clean unstocked forest without any seedlings, and this would in most cases be a temporary situation. A forest is not considered `stocked' before it has reached the thresholds set (10% and 5 meters). Before that it would be temporarily unstocked containing tree seedlings, i.e. it would be a young forest (natural or planted). If this interpretation is correct, the various definitions would be compatible.

The omission of unstocked forest from the CBD definitions has a number of connotations related to what extent unstocked forests house forest biodiversity and whether such areas are considered part of a forest ecosystem. On one hand, emergence of unstocked areas on a temporary basis is part of forest development dynamics, be the forest managed or unmanaged. On the other hand, a definition that would allow inclusion of (large) unstocked areas could be criticized from biodiversity standpoint (e.g. vastly reduced biodiversity at least temporarily).

The difference can be considered fundamental, and the possibility to agree on a common approach in this regard could be explored. The most promising option would be to modify the CBD definition of forest to explicitly include temporarily unstocked areas. Excluding unstocked areas from the FRA and UNFCCC definitions is difficult to justify, since they are firmly anchored in the internal logic of these processes. The context where the CBD definition is used is probably still more flexible, and an adjustment of the definition may be acceptable.

If this kind of consistency between definitions is not considered feasible or desirable, another option is to harmonize them through improving comparability. This could be achieved by introducing a separate category of unstocked forest within the FRA and UNFFFC definitions of forest. From an inventory standpoint this would be rather easy to accommodate, and the cost implications would be modest.

Another aspect of unstocking is degraded forest land which is discussed in Annex 1, Section 3.

Expansion of Existing Definitions of Forest

The current definitions of forest have been criticized for lacking environmental and social criteria, and overemphasizing production aspects (e.g., World Rainforest Movement 2002). One of the key issues is the debate on what characteristics of a `naturally' developed forest should be present in an area qualified as forest (see Annex 1, Section 1). Forest plantations, in particular, have been criticized for being `too simplistic ecosystems' to be considered a forest.

It has proved difficult to agree on such distinctions as well as on appropriate classification criteria, and the current definitions of forest do not yet include references to them. However, most attempts to define the `naturalness' of forests refer to indicators such as species composition and stand structure which, in principle, can be added as attributes to existing definitions. This may, however, lead to significant costs in data collection, since it would involve considerable field work.

Social criteria are more difficult to incorporate in definitions of forest in a manner that would make them practical to use. For instance, considerations such as "equitable sharing of benefits from forests", are difficult to operationalize, since the benefits are often intangible and it is difficult to make them comparable. `Equitable sharing' is also a highly value-laden and context-specific concept. A possible option would be to address social issues mainly through more detailed and comprehensive conceptual frameworks (e.g. criteria and indicators) rather than through such basic and concise instruments as core definitions.

Summary of Options

(1) Threshold values for stand characteristics used by UNFCCC are fixed. However, under GPG a case could be made for countries to voluntarily adopt the FRA definitions as also applied by CBD.

(2) Assess the need to add a qualifier for the term `temporary' in the UNFCCC and CBD definitions of forest.

(3) Consider including `temporarily unstocked areas' in the CBD definition of forest to make it essentially the same as the FRA definition. Alternatively, make the CBD definition comparable with the FRA definition, by distinguishing `temporarily unstocked areas' as a separate class of forest in FRA.

(4) Assess the need and justification for creating a sub-class of non-forest under FRA, `degraded former forest land' to make it consistent with ITTO definition.

(5) Assess the feasibility of incorporating of social considerations in the definitions of forest vs. addressing these under such comprehensive frameworks as Criteria & Indicators for SFM.

Forest Land

None of the international sets of forest-related definitions include an explicit formulation of the term `forest land' or `forestland' (e.g. Lund 2002). The UNFCCC definition of `forest management' includes the term `forest land', but it is not defined (Box 1.1). Related UNFCCC terms including `cropland management' and `grazing land management', which define other land uses.

Box 1.3 Definitions of Forest and Other Land Management

Relationship with `Forest'

Confusion arises from the fact that the UNFCCC definitions of afforestation, reforestation and deforestation (ARD) include three other related, but undefined terms, i.e. `forested land' `non-forested land' and `land without forest'. For example, afforestation is defined as "the conversion of land that has not been forested for at least 50 years to forested land through planting, seeding and/or human-induced promotion of natural seed sources". In this context, it is unclear whether `forested' refers to fully established stands or whether young forests which are not yet firmly established are included or not.

The simplest approach to make the definitions compatible would be to agree that the UNFCCC terms `forested land' and `forest land' are synonymous with `forest'. Similarly, `non-forested' and `land without forest' would be equivalent of `non-forest'. This would eliminate the ambiguity regarding young forests. It is also suggested that only one term be used, possibly `forest', or that `forest land' and `forest' are considered synonymous. `Non-forest' would be the symmetrical reverse of `forest'. However, the issue of dividing non-forest into other wooded land and other land should be considered in this context.¹²

Relationship with Other Land Use Classes

Further confusion arises from the fact that the term `forest land' appears to be referring to land use, in a similar fashion as `cropland management' and `grazing land management'. However, the situation is further complicated by the fact that these definitions do not exclude other land uses from an area. In other words, the available definitions do not provide clear guidance how to determine the land category for an area where a combination of different land uses is practiced which is widely spread and often also a policy objective.

The problem derives from the ambiguity in the UNFCCC definition of forest which does not explicitly address the issue of combined land use. The FRA definition of forest represents a feasible approach where it is stated that `other predominant land uses' should not be present in an area considered a forest. On the other hand, the UNFCCC definition may deliberately avoid references to land use because its main interest is carbon stock.

However, the possibility of overlap in the UNFCCC land classes makes the definitions incompatible with FRA and CBD definitions of forests, which do not allow it. In particular, they require that non-forest land uses should not be predominant in an area considered as forest. The FRA definition specifically excludes orchards, agroforestry and urban forests, and the CBD definition states that the area should not be primarily under agricultural or other specific non-forest land use.¹³

The difference may become an issue in the implementation of the Kyoto Protocol. Currently, the KP definitions assign any system of practices on land on which agricultural crops are grown to the activity `cropland management'. If trees form part of such a system, they may thus be excluded from forest¹⁴. This would also make such lands ineligible under Article 12 of the UNFCCC. So far, the difference has not been an issue because the current UNFCCC definition applies only to Annex 1 countries, where forests are mainly boreal and temperate, and the distinction between forest and other land uses is usually relatively clear.

However, this difference must be explicitly addressed, when formulating a definition of forest to be applied under the CDM of the Kyoto Protocol. In developing countries, forestry is often combined with other land uses (agroforestry, silvopastoral systems etc.). Under the CBD and FRA definitions, they would not qualify as forest if other uses are predominant, whereas according to the current KP definition they would be classified as forests as long as the tree formations in these areas (fruit trees, oil palms etc.) meet the minimum criteria. The discussion to find a suitable approach is underway but still remains unresolved. If the future definition of forest under the CDM will include combinations of forest and other land uses without qualification related to predominant or primary use, it will be incompatible with the FRA and CBD definitions.

On the other hand, harmonization of these definitions could be achieved rather easily. It would suffice to split the land use class `other land' applied by FRA into three classes: one would denote areas where trees are combined with other land uses in `stand-like' formations within `other land'. The second class would include `scattered trees', i.e. patches below 0.5 ha (FRA minimum size). This corresponds to the existing FRA definition `trees outside forest areas'. It is debatable whether a lower boundary such as 0.05 ha (UNFCCC minimum size) should be established. It may be an impossible condition to meet, because such high-resolution data are rarely available. It may be sufficient to establish a new class without a lower boundary. The remaining area under `other land' would constitute the third class of `no trees present'.

As the optional adjustment of the FRA classification would only split existing classes, it would not affect comparability with the CBD definition of forest. To some extent it might even be an improvement from the CBD's standpoint defining forest as function and ecosystem-oriented. The optional new classes could have distinguishable characteristics that may merit a separate different treatment as another, separate group of ecosystems from `forest'.

The cost implications for countries reporting to FRA could be significant. It is not easy to distinguish combinations of land uses with remote sensing. Using current methods, it is also difficult to distinguish differences of tree cover classes with reasonable precision.

Figure 1.2 Relationship between Current FRA Land Classes and Proposed New Sub-classes Combining Forestry with Other Land Uses

Summary of Options

(1) Assess whether UNFCCC terms `forested land' and `forest land' can be considered synonymous with `forest' and, if so, which term(s) should be used in the future.

(2) Clarify the method of classifying lands with a combined land use under the UNFCCC definitions and assess whether the UNFCCC approach can be aligned with the FRA classification by e.g. dividing the FRA land class `other land' into sub-classes.

The work related to clarifying and streamlining definitions could be carried out under the process of developing Good Practice Guidance.

Non-forest

Definitions

`Other wooded land' and `Trees outside forests' are terms found only in the FRA set of definitions (Box 1.1). The need for harmonization may arise from a desire to increase compatibility with the UNFCCC terminology which defines land areas eligible for `revegetation'. Were the terms compatible, the FRA data would directly benefit the UNFCCC in terms of monitoring changes in carbon stock.

The land area, which revegetation applies to, has not been explicitly defined. Reference to vegetation that "does not meet the definitions of afforestation and reforestation contained here (reference to UNFCCC definitions)" suggests that areas which, after revegetation, qualify as forest are excluded. However, the available definitions leave it unclear whether it applies to lands under `cropland management' and `grazing land management'.

Box 1.4 Definitions of Non-forest Land, Trees Outside Forests, and Change in Vegetation

The proposals put forward by various countries suggest that revegetation would not apply to any of these lands, but to another, yet undefined land category outside of them. For instance, the Australian Government has suggested to consider the establishment of widely spaced trees, trees in windbreaks and shelterbelts, trees in alley planting, salt bush tea tree and oil mallee as potential activities under revegetation (UNFCCC 2000b). The Government of Iceland has proposed growing of lupines, planting of grass and associated fertilization as eligible activities (UNFCCC 2000a).

The FRA definition of `other wooded land', on the other hand, does not define a land use. It is therefore not clear whether agriculture or grazing can be practiced on `other wooded land'. The FRA definition `other land' includes agricultural lands and meadows outside `forest' and `other wooded land', but it does not provide guidance as to land uses in `other wooded land'.

Another difference between the FRA definitions of `other wooded land' and the areas under UNFCCC definition of land eligible for `revegetation' is land cover. The FRA definition includes all `woody' vegetation such as trees, shrubs and bushes. In the UNFCCC definition there are no restrictions regarding vegetation growing on land qualifying under the term revegetation (or devegetation).

Further, the UNFCCC definition of land eligible for revegetation sets a minimum limit of 0.05 ha for an eligible area, whereas the FRA applies 0.5 ha. In `other land', this could be covered under the FRA definition `trees outside forests' but in `other wooded land', a new class would have to be created.¹⁵

The relationship between the FRA land classification and land eligible for `reforestation' is illustrated in Figure 1.3 suggesting that harmonization of these definitions would be difficult because several new sub-classes should be created. The biggest hurdle is probably that the FRA land classes `other wooded land' and `other land' should be split according to land use, which would be a new practice. If harmonization of these definitions is considered useful, the first step would therefore be to find a common understanding on treatment of land use. Taking into account the fact that revegetation would probably be relevant only in a limited number of countries, a full harmonization may not be warranted because of the high cost involved as all countries should brake down their data accordingly.

Another approach would be to aim for partial harmonization. All the UNFCCC definitions require that changes are `human induced'. This feature is currently not distinguished under FRA definition of `other wooded land'. It would probably be possible and useful to distinguish between `natural' `other wooded land' (e.g., mountain shrubs or dry savanna) and `human-made' `other wooded land' (e.g., fallow or degraded land). For consistency, it would probably also be useful to distinguish between patch sizes, i.e. 0.5 ha (FRA minimum) and 0.05 ha (UNFCCC minimum). The respective set of sub-classes are illustrated in Figure 1.4 in the area denoting `other wooded land'.

Figure 1.3 Relationship between Current FRA Land Use Classes and Land Areas Qualifying for `Revegetation' under UNFCCC

Figure 1.4 Relationship of Current FRA Land Classes with the Proposed Sub-classes under `Other Wooded Land' and `Other Land'

Option for Further Work

(1) Assess the feasibility of harmonizing the FRA land uses classes and lands qualifying under UNFCCC definition of revegetation.

Changes between Forest and Non-forest

Definitions

The two principal sets of definitions for deforestation, reforestation and afforestation are provided by FRA and UNFCCC. The CBD applies FRA definitions of reforestation and afforestation, but refers to wordings presented in FAO Working Paper 33 (FAO 2000b). ITTO has developed its own definitions of afforestation and reforestation. FRA has also developed a definition for natural regeneration and natural expansion of forests. The compatibility and potential for harmonization of the various definitions were extensively discussed in the first Expert Meeting on Forest-related Definitions (Proceedings 2002).

Afforestation and Natural Expansion of Forest

Afforestation, as applied by FRA 2000, is the conversion of non-forest into forest as a result of direct human action through planting or seeding. FRA does not make any qualification regarding the means of afforestation (through seeding or planting). The Expert Meeting recommended that FAO consider expanding the FRA definition of afforestation to include assisted succession to trees not involving direct seeding or planting. The ITTO definition refers only planting, and the option of including seeding and other assisted measures as a means of afforestation could be considered.

All three definitions of afforestation are compatible in the sense that they require human action, crossing of the forest/non-forest threshold and ceasing of other predominant land uses. They differ in that the Kyoto Protocol requires that the land has not been forested within the previous 50 years, whereas the FRA and ITTO definitions do not include any such qualification. In order to contribute to harmonization and simplification of definitions, the Expert Meeting recommended that the UNFCCC consider, in the second or subsequent commitment period, dropping the requirement for a 50-year non-forest condition for afforestation. This would eliminate the need for a separate definition of reforestation and bring the UNFCCC afforestation figures into closer agreement with the FRA results.

The treatment of young forests is compatible in the UNFCCC and FRA definitions. However, the UNFCCC definition explicitly includes young forests, whereas FRA definition considers as afforested only young forest stands that have been successfully established, but may not yet have crossed the applicable thresholds. The difference is minor, and harmonization of the definitions could be considered. The ITTO definition is problematic in the sense that afforestation is a `planted forest', but ITTO does not provide a definition for `forest'. ITTO may therefore consider adopting one of the existing definitions or formulate one, which is compatible with them.

The UNFCCC does not provide a definition for natural expansion of forest. This is logical in the sense that eligible activities include only those that are directly human-induced. ITTO, on the other hand, could consider developing a similar definition to make its set of definitions more complete.

Box 1.5 Definitions of Changes between Forest and Non-forest

Reforestation

The FRA definition of reforestation implies active establishment (through seeding or planting) of forest on land previously forested but temporarily below the forest threshold due to harvesting or disturbances. Natural regeneration on forest lands is defined and accounted separately, which makes the FRA definition in the case of assisted natural regeneration compatible with the UNFCCC requirement that changes must be human induced. Full consistency would require FRA to separate assisted and unassisted regeneration.

The ITTO definition is compatible with these definitions as it states that reforestation takes place `after removal of forest cover'. `Re-establishment' does not indicate whether reforestation is human-induced or not, and an adjustment may be considered. As discussed above, the most problematic issue regarding compatibility is that the ITTO definition does not clearly refer to an established definition of forest.

Lands undergoing reforestation or natural regeneration (according to FRA) continue to be forest throughout. Neither of these transition processes involves a change in land-use class. The UNFCCC definition defines reforestation as conversion of land that was forested but had been converted to non-forested land. For the first commitment period, domestic reforestation is restricted to land that did not contain forest on 31 December 1989. Reforestation, as defined by the UNFCCC, is accounted as afforestation under FRA 2000 since the land was not previously forested. The current definitions of reforestation by FRA and the UNFCCC are therefore incompatible from a land-use point of view. The definitions cannot be reconciled. The ITTO definition is close to that of FRA, and the two could be harmonized with minor adjustments.

The terms afforestation and reforestation have not yet been defined under Article 12 of the UNFCCC referring to the CDM. If different thresholds are used from those under Article 3.3, this could have major implications for land area reported as afforested or deforested. The requirement of meeting sustainable development objectives will also introduce additional conditions. Credits for afforestation and reforestation activities that do not meet sustainable development objectives, as defined by the Party, are likely to be excluded.

Deforestation

A key feature of UNFCCC definition of deforestation is that the process is directly human-induced. The FRA definition, on the other hand, does not distinguish natural loss of forest from that caused by human action. While both definitions are consistent with the logic of their respective frameworks, the difference makes them incompatible with each other. The Expert Meeting recommended that FAO differentiate direct human-induced deforestation and permanent forest loss due to other causes. This would make the FRA data compatible with the needs of the UNFCCC.

Both definitions refer to non-temporary (long-term or permanent) change from forest to non-forest. The FRA defaults the time period for a `temporary' unstocked state at usually a maximum of ten years, while the UNFCCC leaves it undefined.

Summary of Options

(1) Expand the FRA definition of afforestation to include assisted regeneration not involving direct seeding or planting

(2) Drop the requirement for a 50-year non-forest condition for afforestation in the UNFCCC definition to be applied from the second commitment period onwards

(3) Consider harmonizing the treatment of young forests in the FRA and UNFCCC definitions of afforestation

(4) Consider addition a definition of natural expansion of forest in the ITTO set of definitions that is compatible with the other existing definitions

(5) Consider developing the ITTO definitions by including a reference to an established definition of `forest', and making minor adjustments to increase compatibility with FRA and UNFCCC definitions

(6) Differentiate direct human-induced deforestation and permanent forest loss due to other causes in the FRA definition of deforestation

Forest Degradation

Definitions

Definitions of forest degradation have been formulated by FRA and CBD. The ITTO proposal¹⁶ has adopted the CBD definition but expands it with an additional qualifier. Forest improvement describes the reverse process of forest degradation. Other terms for this purpose may, however, be considered, such as aggradation, amelioration, melioration, rehabilitation, unsustainable management etc. The Expert Meeting discussed extensively the option to develop the concept and its definition (Proceedings 2002). Related concepts, `forest fragmentation' and `forest improvement', are discussed in Annex 1, Sections 4 and 5.

Box 1.6 Definitions of Forest Degradation

Table 1.4 Parameters of Various Definitions of `Forest Degradation'

Compatibility of Existing Definitions

The FRA definition defines `degradation' as reduction in the capacity of the forest supply of goods and services. The CBD and ITTO definitions share this approach, but a notable difference in the CBD definition is that the reference point is a `natural' state of forest. Significant deviation (defined in various ways) from such a state is considered degradation. The FRA and ITTO definitions, on the other hand, do not refer to a reference point, but imply any reduction in the supply of goods and services as degradation.

The draft definition formulated by IPCC for the UNFCCC tries to capture the essence of the degradation process through an operational definition using canopy cover as an indicator. Any reduction in indicator and value is considered `degradation'.

The FRA, CBD and ITTO definitions are largely compatible, and a generic common definition could be formulated without greatly disrupting the existing use of them. The IPCC draft definition is geared towards operational use, and it is difficult to make it compatible with the others.

Composite Index

The principal difficulty of applying existing definitions arises from the fact that forests produce a multitude of various goods and services, and many of these are produced simultaneously on the same piece of land. Due to trade-offs, efforts to increase production of one output may reduce the availability of another. Owing to this interdependence, one of the key issues is to define how to deal with trade-offs between outputs, i.e., whether a reduction in the supply of one good or service can be offset by an increase in the supply of another. For instance, monoculture plantation could increase wood supply but lead to loss of biodiversity.

The Expert Meeting discussed the option that a generic, composite index for degradation, based on a weighted combination of indicators and/or their changes over time, could be a template for international application. A negative change in any indicator (beyond a certain threshold value) would represent an element of degradation. Weighting would be justified, as various negative changes would not have an equal impact on forest functions. In principle, the use of such a composite index implies that both negative and positive changes could be accommodated to be determined based on a combined impact. This was, however, not spelled out in the proceedings.

In some cases the supply of goods and services from a managed forest can be better than from a `naturally' developed forest if all the economic, social and ecological functions are accounted for (e.g., planted forest on degraded marginal land) (Holmgren, pers. comm.). From this standpoint the key indicator would be the change in the supply of goods and services rather than the status of forest. The naturally developed forest could still be adopted as a reference point for assessing such a change.

In practice, there are a number of problems associated with this approach. Above all, many of the goods and services cannot be measured using the same, neutral yardstick. For instance, trying to determine trade-offs between environmental services and production of timber is a value-based choice. Many of the services (e.g. providing aesthetic values) are also intangible.

A further problem is that the degradation in the supply of environmental services (e.g. biodiversity) may materialize long after the activities causing degradation (or improvement) have taken place. For instance, forest species do not respond immediately to deterioration in the forest environment. Indicators based solely on the current supply of goods and services may not be able to detect these changes.

Another problem is the reference point. One option is to use the structure of `naturally' developed forest as benchmark. However, it is unclear what would constitute a forest in this context, i.e. whether it is (a) a single stand, or (b) a group of stands of varying ages and development stages. In the latter case it is unclear how the concept should be applied to a single stand, if at all. It is also difficult to define which point of forest succession should be adopted as a reference point (early/mature/climax), as well as to determine at what stage of succession the stand(s) to be analyzed are. Acceptable deviations from the reference point would also be difficult to determine. For instance, should replacement of a broadleaved stand with a coniferous stand earlier than would happen in `natural' succession be considered degradation?

Provided that the problems discussed above can be satisfactorily solved, the development of a comprehensive composite index could be considered. It should probably include indicators for the current as well as the potential supply of goods and services. For instance, the current supply of goods and services could serve as the main indicator, and change of forest structure could be used as a secondary indicator which is allowed to vary within a particular range (see Figure 1.5).

If this proves unfeasible, another approach could be to apply the generic definitions and develop individual indices for various aspects of degradation/improvement. Assessment on their combined effect would be done on a case-by-case basis.

Figure 1.5 Elements of Composite Index Comprising Supply of Goods and Benefits and Characteristics of Forest Structure

Other Considerations

Regarding the time frame the FRA definition indicates that degradation involves a `long-term' reduction of potential supply. The term is included to avoid all logging operations to be considered degradation, as it would often lead to improvement of the stand in the long run (e.g., thinnings). Similarly, the damage caused by a fire would represent degradation in the short-term, but a long-term improvement in some ecosystems (Holmgren, pers. comm.). The CBD 2001 definition leaves the time frame open.

In this context, an important concept is the resilience of different forest ecosystems, which the ITTO definition refers to. The available definitions may be inadequate because they do not take into account the relative levels of resilience or buffering in different forest types.

The `long-term reduction of potential supply' is almost equivalent to an indication that the resilience of forest has been exceeded (see above). However, the difference is that if the resilient capacity of a forest is exceeded, the result is an irreversible change. `Long-term' may refer to a situation where the supply of goods and services is reduced for a long period time, but the resilience of the forest has not necessarily been broken down. In other words, reference to resilience implies more dramatic and permanent changes than an indication of `long-term' changes. Harmonization could be approached by adjusting the FRA definition to include the concept of resilience.

Neither the FRA definition nor the CBD definition indicates spatial scale for assessments. The options include individual forest management units (FMU)¹⁷, groups of small holdings, landscape, ecosystem, administrative district, national or some other level. From the operational standpoint, the FMU is an important concept and its operations should be assessed as a whole. If assessment is made only in harvested areas of an FMU, temporal observations would record reduction in stocking levels or canopy cover which could be interpreted as degradation. However, management of the entire FMU could still be sustainable, including areas with no harvesting.

Another difference between the FRA and CBD definitions is that the latter refers only to degradation that it human induced, whereas the FRA does not make this distinction. This makes the FRA definition more complete, since degradation is not always human-induced, as it can also take place for natural reasons (e.g., nutrient leaching, desertification). In principle, this distinction is required under UNFCCC, but in practice making the distinction may be difficult (Annex 1). Given the theoretical and practical problems of distinguishing between these two cases two feasible options could be considered: (1) not defining this attribute, or (2) qualifying all degradation as `human-induced'.

Summary of Options

(1) Consider harmonizing the FRA, CBD and ITTO definitions of forest degradation

(2) Consider incorporating the concept of resilience in the existing definitions of forest degradation

(3) Explore options to determine on what spatial scale degradation should be assessed

(4) Analyze the feasibility of developing a composite index for forest degradation paying special attention to the justification/feasibility of (a) compensating degradation in the supply with one good or service with improvement in the supply of another one, and (b) adopting `naturally' developing forest as a reference point for forest degradation

Managed Forests

Definitions

Definitions for managed forests are provided by FRA and ITTO (Box 1.7). The UNFCCC process provides a definition for forest management, which is closely related to the other two. Table 1.5 denotes the presence of parameters in alternative definitions. The IPCC (1996) refers to an inverse concept "unmanaged forest", which contains definitional elements.

Table 1.5 Parameters of Definitions of Managed Forest and Forest Management

Box 1.7 Definitions of Managed Forests and Forest Management

All three definitions related to managed forests are based on the approach that management is a purposeful action with the objective of fulfilling specific objectives or functions. The ITTO definition states that management results in a modified forest structure or species composition, whereas the FRA and UNFCCC definitions do not define an outcome or add any qualifier to it.

This difference evokes the question, where to draw the line between managed and unmanaged stands. The ITTO definitions suggest that management is something that requires active intervention, i.e. altering natural forest development. The definition is rather broad covering a whole range of management objectives, which implies that unmanaged forests occur in rather rare circumstances. Strictly speaking, non-managed forests are limited to areas, which are in no way influenced by human intervention. Accordingly, managed forests are not limited to production forests, as protected areas are subject to management as well. For example, semi-natural forest meadows would not be able to maintain their status without active management. Even limited interventions, such as boundary demarcation, fire protection (e.g. against human-induced fire) are part of management.

The same principles can be derived from the FRA or UNFCCC definitions, but they, in principle, go one step further. Since no outcome is defined, a mere decision not to intervene, could be considered forest management in the sense that it involves a conscious choice between alternative development paths. In other words, making a choice is the key issue. On the other hand, this interpretation would be restricted to situations, where forest management is an option. In inaccessible areas, where "active" management is not an option, no management can be practiced, and, consequently, no choices are made. The difference may be significant and should be clarified.

All of the above-mentioned definitions disqualify anthropogenic influence that occurs without a specific purpose of forest management (Annex 1, Ch. 6). For instance, setting a forest accidentally in fire would not qualify as management. The IPCC definition applicable to carbon accounting, on the other hand, is less clear in this regard. It states that "If they [forests] are being used as a source of fuelwood, or are being affected in other ways by ongoing human activities, they should be accounted for" [i.e. considered managed]. This formulation may be based on the notion that countries that are signatory to the Kyoto Protocol are responsible for protecting their forest resource against any kind of deterioration caused by humans. For instance, lack of proper fire protection should not be an excuse for not recording a reduction of carbon stock in a country's carbon accounts.

The ITTO and UNFCCC definitions include references to sustainability and maintenance of all forest functions, whereas the FRA definition does not mention these explicitly. For clarity, incorporation of these features in the FRA definition could be considered.

Summary of Options

(1) Consider clarifying the distinction between managed and unmanaged forest especially as regards non-intervention, and accidental anthropogenic influence.

(2) Consider incorporating a reference to sustainability into FRA definition of managed forests.

Forest Classification

The concepts of biome, forest type and forest ecosystem are interrelated in the sense that they serve to classify forest vegetation in various manners. The most commonly used term is forest type, which provides a basis for forest classification systems in most countries. The option to use biome as a basis for forest classification has also been discussed, mainly in conjunction with the implementation of Kyoto Protocol. It has been suggested that biome-based classifications would be particularly suitable for developing differentiated definitions or thresholds of forest. Forest ecosystem has not yet been applied as a basis for forest classification in international processes.

Biome

Among the international processes and instruments analyzed under this study the CBD is the only one that has provided a definition for `biome'. The expression `domain' used in the FRA ecological zoning is another similar concept (Box 1.).

Box 1.8 Definitions of Biome and Domain

A number of other formulations are available, and there is also half a dozen of other words or expressions denoting similar concepts (e.g., formation, major life form, major life zone, major community, ecoregion, ecofloristic zone, etc.). The distinguishing features highlighted in these definitions vary considerably, and the expressions are, perhaps unavoidably, often so vague that the difference in wordings has less significance for classification than the interpretation given to them by the one whom is doing it. The level of classification is also unclear; one definition may consider a forest type as an example of a biome, whereas for another definition the same forest type belongs to a hierarchically lower class (Rakonczay 2002).

However, as regards harmonization, the differences between these definitions are not yet a major issue, because they are not applied in any major international process. The key issue regarding biomes is whether it is necessary to apply different definitional thresholds in different forest conditions. This has been debated in particular in conjunction with the Kyoto Protocol. If a differentiated approach is adopted, the question is then, whether it should be done on the basis of biomes, or the current classifications based largely on forest type, or at all.

The Expert Meeting made a recommendation regarding the latter issue. It was concluded that biomes are probably less useful as a basis for different definitional thresholds than forest or vegetation types. The concept of biome is not necessarily compatible with the actual land use and the way forests are managed and utilized. As the territory of many countries includes several biomes, using biome-specific definitions would increase, rather than decrease, the reporting burden. Socio-economic parameters and land-use systems cross-cut the limits of biomes, which is another complicating factor.

Forest Type

Forest type has been defined by several processes, including CBD and ITTO while the FRA applied a related concept `ecological zone' (Box 1.9).

The CBD definition of forest type refers to "groups of forest ecosystems of generally similar composition" as a basic unit of a forest type. In the ITTO definition it is "a community of trees and associated plant species with uniform physiognomy". The FRA system of ecological zones (Box 1.10) is close to these definitions, even though the approach is broader encompassing formations without trees. In the FRA system the basic classification is done on the basis of "zones or areas with broad, yet relatively homogenous vegetation formations".

Similarities in the physiognomy (structure) of vegetation formation are distinguishing features of both the FRA and ITTO definitions. The CBD definition includes structural elements (tree and undercanopy species composition, crown closure) as well as productivity as a separate aspect. However, productivity is largely a function of ecological and climatic conditions, and it is to large extent reflected in the physiognomy, which reduces the significance of the difference.

Both the ITTO and FRA definitions indicate similarity of ecological conditions as another distinguishing feature between classes. In the FRA definition a specific reference is made to climatic conditions based on temperature and rainfall. This attribute, however, is missing in the CBD definition. Another slight difference is that FRA and ITTO definitions refer to `natural' formations, whereas this is not explicitly mentioned in the CBD definition.

The definitions are still rather broad, and there is ambiguity as to what level of hierarchy they represent, and what detail of classification they require. Sometimes forest type is confused with biome or used interexchangably with it (Rakonczay 2002). On the other hand, the concept of forest type is necessarily so diffuse that this is perhaps unavoidable. Even if there were a commonly agreed definition of forest type, it would probably not be able to provide unambiguous guidance on determining the number of classes, distinguishing characteristics between classes, etc.

Box 1.9 Definitions of Forest Type

Box 1.10 FAO Global Ecological Zoning

The main issue may therefore not be the harmonization of various, more or less vague definitions, but reaching an agreement on the key features of the classification to be adopted, as well as an assessment of the potential gain of more detailed harmonization. One option, therefore, is to adopt an existing system as a starting point. While the number of different classifications is large, few of them have been used systematically to gather and process large quantities of global data. The FRA classification by ecological zones is the most comprehensive and widely used system, and the option to adopt it as a basis for further development should be carefully considered. This approach was also recommended by Rakonczay (2002) in case a decision is made to adopt differentiated definitions.

Table 1.6 Parameters of Definitions of Forest Type

At the country level, the existing systems are well established, and have been developed to match the local conditions. It may therefore be advisable not to aim for harmonization of these systems with any international system, but rather to ensure that the information systems are able to provide data that is internationally required. This appears to be a highly feasible approach, as indicated by the reclassification of forest types used at the country level according to the FRA ecological zones.

Forest Ecosystem

The scientific community has formulated numerous definitions of ecosystem. Of the four international processes analyzed in this report, the CBD is the only one providing a definition for it (Box 1.11).

Box 1.11 Definition of Forest Ecosystem

The CBD definition of ecosystem indicates that it can be applied at different levels of definitional hierarchy. It is also a complex approach, as it involves dynamic interaction between biotic and abiotic environment, including humans. Despite the fact that ecosystem is a sound concept in ecological terms, it is difficult to apply it as a basis for forest classification. Classification systems must be based on concepts, which are rather rigid and display as little ambiguity as possible. Concepts such as forest type are more robust, even if in ecological terms they may not capture all the elements of the state and change of forest. In addition, the fact that different interpretations of the ecosystem concept are fully justified in particular contexts makes it difficult to reach an agreement on a globally applicable definition which could be operationalized through data collection. For practical purposes, it may be sufficient to adopt an appropriate classification of forest types serving as a rough proxy for classifying ecosystems.

Differentiated vs. Universal Definitions

It is of interest to analyze what gaps differentiated definitions are intended to fill, and whether the current systems - including the definitions they apply - could be adjusted to meet the new requirements. The main conclusions reached by Rakonczay (2002) in this regard include:

• Even in a differentiated approach, the same basic parameters of definition of forest would apply, namely: minimum area, minimum tree height at maturity, and minimum canopy cover (or another appropriate measure of density).

• The values of the above parameters influence the amount of land that is classified as forest. The effect of these parameters appears to be greatest towards the margins of the natural distribution of forests, and in areas highly impacted (disturbed/fragmented) by humans.

• No strong arguments have been identified either in favour or against defining forest on a biome-specific basis. Left unchanged long enough, any reasonable definitional scenario will detect major trends in transitions between forest and non-forest areas. As long as the system is based on a set of threshold criteria, its sensitivity (its ability to detect small changes) can be increased by reducing the spatial assessment units than by changing the definition.

It is also noteworthy that a change in a definition always entails a significant cost increase. Reporting burden during the transition period from one definition to another would increase, inventories may be come more complicated and comparability with historical data is often difficult to maintain. There is also a risk that such a change would create loopholes and perverse incentives (Rakonczay 2002).

Alternative approaches include development of more accurate measurements. Reduction of spatial assessment units is one of the means for increased accuracy to detect small changes (Rakonczay 2002). This option, of course, also entails a significant cost increase, but it should be compared with the costs of other approaches.

Another option is to develop an overall classification and assessment system (Rakonczay 2000). Instead of focusing assessments on the extent of forest and non-forest land, the scope could be enlarged to detect changes within forest and non-forest land. This approach could achieve the same objectives as differentiated definitions (Figure 1.6).

The UNFCCC has proposed a method for assessment of carbon stock within non-forest land. Revegetation (and devegetation) is a concept intended to capture an increase of carbon stocks in case the change does not qualify as reforestation or afforestation. Revegetation is a particularly useful concept in marginal areas of natural forest vegetation (e.g. areas subject to desertification), which has been pointed out as one of the problem areas for a non-differentiated definition of forest. Provided that appropriate monitoring methods are developed, revegetation may enable assessments that are sufficiently accurate for the purposes of UNFCCC. The FRA concepts `other wooded land' and `trees outside forest' serve for the same purpose, and introduction of new sub-classes could make them more useful than at present.

Regarding changes within forest land, the existing classification could also be developed to detect changes in zones which are highly disturbed or fragmented by human activities. These are claimed to be another problem area for a non-differentiated definition of forest. There is a concern that, in strict definitional terms, an area may remain forest, even if substantial degradation has taken place (Rakonczay 2002). Detecting such changes may be possible by developing the terminology and assessment methods related to `forest degradation' in a manner suggested earlier.

Figure 1.6 Detection of Changes in and between Forest and Non-forest Land

Option 1

Option 2

More detailed measurements in areas adjacent to borderline

The problem regarding carbon accounting is that the CDM has opted to exclude from its scope other changes than those related to afforestation and reforestation. In the CDM approach, the detection of changes is strictly based on changes between forest and non-forest areas. However, if changes in carbon stock within forest areas is considered the problem that has to be addressed under the UNFCCC, the option of enlarging the scope of eligible activities to forest degradation and aggradation would be studied. Changing an existing and well-established definition of forest is a complex and intricate process involving significant costs.

Summary of Options

(1) Consider harmonizing the various definitions of forest type using the FRA classification of ecological zones as a starting point.

(2) Clarify the difference between forest type and biome in terms of their level in the definitional hierarchy.

(3) Review the possibility of developing the terms `revegetation/devegetation' and `forest degradation/improvement' as an alternative to introducing a differentiated definition of forest.

Natural Forest vs. Forest Plantation

Definitions

Definitions of forest plantations vs. natural forests have a whole range of connotations which are subject to debate at an international level which are beyond the scope of this analysis. The key issue underlying the debate is the criticism that plantations are claimed not to fulfil the functions of a `normal' or `natural' forest. The available definitions are reviewed only briefly concentrating on their relationship with other definitions. Definitions of plantation forests are provided by FRA, ITTO and CBD (Box1.12). A definition of natural forest and semi-natural forest, which is a related concept, is provided by FRA.

Box1.12 Definitions of Natural and Plantation Forest

The FRA definition of natural forests is based on an exclusion, `forests excluding plantations', which suggests the difficulty of formulating an appropriate definition, and probably explains why other processes have limited their definitions to concern only plantations.

The three definitions of plantation provided above are very similar. All of them refer to forest stands established either through planting and seeding. A minor difference is that the FRA definition includes a qualifier regarding introduced and indigenous species which are lacking in other definitions. The CBD definition provides site characteristics.

In many cases the difference between natural forest and forest plantation is not clearcut, as planted seedlings can be mixed with naturally regenerated seedlings. This is often the case in the European forests, where the line between semi-natural and plantation forests can be difficult to determine. Semi-natural forests often have stand characteristics, which resemble those of a `naturally' developed forest (e.g. Buchwald 2002).

Table 1.7 Parameters of Definitions of Forest Plantations

This suggests that the method of re-establishing a tree stand may have less significance for forest functions than the characteristics of the stand that ultimately develops. AHTEG (UNEP/CBD/SBSTTA 2001) pointed out that a gradient exists among plantation forests from even-aged, single species monocultures of exotic species with a fiber production objective to mixed species, native to the site with both fiber and biodiversity objectives. This gradient would probably also reflect the capability of the plantation forest to maintain `normal' local biological diversity.

If this argument is accepted, it may be inferred further that the way in which forest stands are established would not have to be a starting point for classification from the biodiversity point of view¹⁸. If a distinct class is considered necessary, it could be established based on the characteristics of established stands. On the other hand, the concepts `afforestation' and `reforestation' are closely tied to the forest plantation concept.

Regarding the various definitions analyzed in this study, the distinction between plantations and natural forest is particularly pertinent to forest degradation and improvement. The mere establishment of plantations has been considered forest degradation, especially if they replace natural forests (e.g. World Rainforest Movement 2002). The argument is that they should therefore be either excluded from a definition of forest, or at least should be treated separately from natural forests in terms of assessing their status, supply of goods and services, etc.

On the other hand, if a composite index of degradation could be developed, the less desirable characteristics of plantations (monoculture, uniform stand structure, etc.) could be taken into consideration as components of this index. The negative changes could be weighted against the positive contributions that plantations may have (e.g., increase in fiber production or soil protection) which would make the classification based on plantation/natural forest less relevant. However, as discussed previously, this approach faces a number of theoretical and practical problems.

Summary of Options

(1) Explore the possibilities to combine stand characteristics as a descriptor of forest plantations in addition to the method of establishment.

(2) Assess whether the sustainability of forest plantations could be evaluated using a composite index or using the C&I frameworks.

Low Forest Cover

Definition Approaches

The Intergovernmental Forum on Forests (IFF) identified countries of Low Forest Cover (LFC) as being of special concern. The Tehran process initiated in 1999 has been promoting the development of appropriate relevant definitions (FAO 2002). In response, UNEP and IUFRO carried out an analysis on options to define low forest cover (Lund 1999). It suggested that one of the most interesting alternatives is classification of countries based upon a combination of variables. These include, inter alia, ratios of:

- forest/total land area

- forest and other wooded/total land area

- existing/original forest area

- actual/potential forest area

- forest per capita

Many of the countries, which potentially qualify as low forest cover countries, are situated in dry zones. In these areas, the indicator values are very sensitive to changes of the thresholds set for the definition of forest. The accuracy of estimating forest area in marginal natural conditions is another key issue.

Two possible approaches have been discussed to increase the sensitivity and accuracy of assessments: (i) differentiated definitions based on biome or forest type, (ii) development of technical means to improve accuracy of classifications based on universal definitions. Regarding differentiated definitions, it is not clear whether they increase accuracy of measurements (Rakonczay 2002). In addition, problems with technical feasibility and cost may hinder efforts to pursue this strategy.

An alternative approach is to retain universal definitions, but increase the ability of technical means to assess the status of vegetation in marginal areas. With respect to FRA, enhancing the ability to detect changes in the FRA classes `other wooded areas' and `trees outside forest' would probably increase the accuracy significantly in LFC environments. There is a cost factor involved but, as indicated earlier, it may be less expensive and technically more feasible than using differentiated definitions. The development of the concept `revegetation' under the UNFCCC process is another key approach.

Alternative approaches to define low forest cover are in some respects a separate issue not directly related to harmonization of definitions of the core terms of international instruments. Exploring further options for LFC definition(s) (e.g. development of combined indexes), should continue to be part of a concerted, comprehensive effort to develop compatible and harmonized definitions. How vegetation cover is assessed in marginal areas is linked to development of definitions not only under the FRA but also within the UNFCCC (e.g. as regards the term `revegetation'). To ensure that the definitions emerging from these processes are mutually compatible and consistent, it is necessary that all relevant parties take part in this process.

Summary of Options

(1) Assess the options to increase accuracy of assessments on vegetation cover in marginal (dry) natural conditions in conjunction with the work carried out under other relevant processes, especially the FRA and UNFCCC.

(2) Assess the feasibility of composite indices for characterizing low forest cover countries.

¹⁰ This may not necessarily be true with the Kyoto Protocol.

¹¹ Not necessarily one default value.

¹² This could possibly be taken up by GPG.

¹³ Lands under silvo-pastoral systems where grazing is a complementary activity would be considered forest by the FRA definition.

¹⁴ The countries may decide how they deal with such situations in reporting.

¹⁵ The internal consistency of the FRA definitions for `trees outside forests' and `other wooded land' may need a review.

¹⁶ ITTO is still in the process of finalizing the guidelines where degraded and secondary forests are defined.

¹⁷ There are alternative interpretations for how a forest management unit is defined (cf. e.g., ISO 1998).

¹⁸ For the Kyoto Protocol, the method would remain a relevant aspect.