Most of the definitions collected by Lund (1999) characterize deforestation as the long-term or permanent removal of forest cover and conversion to a non-forested land use. For example, deforestation is the "permanent removal of forest cover and withdrawal of land from forest use, whether deliberately or circumstantially." Similarly, the IPCC Guidelines emphasize the conversion of forests (to pasture, cropland, or other managed uses): "Conversion of forests is also referred to as 'deforestation,' and it is frequently accompanied by burning."
It is important to note that a common forestry definition of reforestation is not the antonym (mirror) to the common definition of deforestation. In forestry practice, reforestation commonly refers to any act involving reestablishment of trees regardless of whether it follows harvesting or a long period of deforestation, but harvesting is not equated with deforestation. For example: "Clear-cutting (even with stump removal), if shortly followed by reforestation, is not deforesting" (FAO, cited in Lund, 1999).
Nevertheless, many of the issues that arise with regard to certain definitions of reforestation also apply in defining deforestation. The period of changed land use required for reestablishment of trees to be treated as reforestation must be compatible with the period of changed land use that describes deforestation. Including a waiting period in the definition of deforestation (e.g., "the act of removing forest cover and not beginning to regenerate forest within X years") introduces accounting needs that can extend beyond a specific commitment period. For example, areas cleared of forest during the first commitment period would not be confirmed as deforested for some years after the clearing, which may extend beyond 2012.
Many definitions of deforestation include natural (non-anthropogenic) events such as landslides, volcanism, and so forth. Where these events are unambiguously not direct results of human activities, the Protocol and Decision 9/CP.4 are clear: If they do not meet the "direct human-induced LULUCF activities" requirement, they would not be included as deforestation under Article 3.3.
Not all natural events are so easily partitioned, however. In particular, several difficult definitional issues arise where significant forest losses occur as a result of fire (and, in some cases, landslides). Forest fires can be the result of natural events (notably lightning) or the direct (or indirect) result of human activities, including prescribed burning (and their escape), accidental fires, and arson. Even if the cause of the fire can be unambiguously attributed to prescribed burning, arson, or escape from deliberate fires (including camp fires and cigarettes), the act may or may not be deforestation in the sense defined above. In many-but not all-instances, forest cover loss is followed by natural (or aided) regeneration of new forest. To the extent that the regrowth is complete, the net release of carbon to the atmosphere may be completely recovered over time by the new forest. There is, however, an asymmetry in the rates (Kurz et al., 1995): The net release of carbon to the atmosphere is rapid, whereas the recapture takes decades or more. From a scientific perspective, however, asymmetrical accounting of changes in carbon stock over time arises if only one of the paired activities (forest loss and regrowth) is accounted as deforestation or reforestation, respectively.
Given that distinguishing between natural and anthropogenic factors that influence the vulnerability of the land to disturbance will be difficult, a key question for policymakers is whether the accounting should include only direct human activities (reforestation) or both the human activities and the event (fire) that makes the opportunity for reforestation possible.
In some cases, there are definitional and carbon accounting issues concerning drawing a clear boundary between natural phenomena and human-induced activities, when, for example, significant forest losses occur as a result of fires or disturbances such as pest outbreaks. In cases involving lands under Article 3.3 or 3.4 where fires or pest outbreaks occur in a forest, a question is whether accounting should, inter alia: (i) count neither the loss nor subsequent uptake of carbon (which reflects the actual net change in carbon stocks on those lands and exchange of carbon with the atmosphere in the long term, but creates problems in continuing to account for the area burnt/defoliated as lands under Article 3.3 or 3.4); (ii) count both the loss and subsequent uptake of carbon (which reflects the actual net change in carbon stocks on those lands and exchange of carbon with the atmosphere, but creates an initial carbon debit for the Party concerned); (iii) count only the loss of carbon (which would overestimate the actual losses of carbon stocks, not represent the exchanges of carbon with the atmosphere, and create future accounting problems), or (iv) count only the subsequent uptake (which would fail to reflect the actual changes in carbon stock and would not represent the exchanges of carbon with the atmosphere, and would provide carbon credits for the Party concerned).
In cases involving lands that do not fall under Articles 3.3 or 3.4, where fires or pest outbreaks trigger land-use change, the consequences are similar to deforestation. If similar vegetation cover is allowed to regenerate, such disturbances may not lead to a long-term change in carbon stocks.
Will forest protection practices (such as fire management systems that seek to prevent wildfires or ameliorate the damage they cause through suppression efforts) be rewarded (in a Protocol sense)? Will Parties be penalized (under the Protocol) for failing to manage wildfires? These issues remain to be resolved through negotiation by the Parties.
A further complication is the fact that fires in many ecosystems do not result in complete tree mortality, even though there are significant changes in the carbon balance of the affected areas. Individual trees and small patches of trees and understory can survive. In some forest types (e.g., eucalypt and some pine forests), even the most intense fires do not cause significant mortality of mature trees. Significant quantities of carbon may be released, but most trunks and large branches survive and regrow full canopies within a few years. Although measurement of these changes under the stock change method as recommended by Article 3.3 is possible, in practice the areas of forest meeting the deforestation requirement of Article 3.3 would be contentious and difficult to assess.
A less difficult situation arises when land-use change takes place subsequent to or as a result of fire. In many parts of the world, major forest fires (whether ignited by humans or naturally caused) are taken as the cue for land-use change to agricultural or pastoral uses. In these cases, the combination of fire and land-use change should be accounted as deforestation in a comprehensive and symmetric accounting system.
For several reasons, accounting procedures that always treat changes in carbon stocks that are attributable to fire as debits may be inconsistent with long-term objectives of increasing carbon stocks. First, fire has been used in some ecosystems as a forest management tool for enhancing forest growth over the long term, even though it might result in some carbon reductions over the short term (for example, during a commitment period). Second, fire is a natural part of many forest ecosystems, and management to eliminate fire can be futile or counterproductive. Natural or prescribed ground fires in such ecosystems may result in preservation of carbon stocks at the landscape scale by creating firebreaks-thus reducing the likelihood of widespread, intense fires.
In the long term and over large areas, there is a balance between carbon loss resulting from fires and carbon sequestered in regrowth, providing there is no change in the fire regime. This balance underlies the recommendation in the IPCC Guidelines to treat wildfires as carbon-neutral. Although this approach is pragmatic, there is some evidence that fire regimes are changing in parts of the world (e.g., Kurz and Apps, 1999). In some cases, these changes may be a result of deliberate human manipulation; in others, it is an indirect consequence of human actions-possibly including climate change. Moreover, because of asymmetry in the rates and timing of carbon loss from fire and its re-accumulation through regrowth, the complete balance is not achieved if fire regimes vary in response to changes in El Niņo patterns or rapid climate change.
Other reports in this collection