This section describes the range of factors that can affect terrestrial carbon stocks. It focuses on the relationship between human-induced LULUCF activities and other factors. Section 2.3.4 discusses a range of options for addressing these factors in the accounting system.
Terrestrial ecosystems are dynamic. Carbon stocks change over the full range of time and space scales, for many different reasons. Although carbon stock changes can be measured directly with a variety of techniques (see Section 2.4), attributing a given change in carbon stocks to a particular cause can be much more challenging. Yet identifying the reason for a change in carbon stocks may be crucial to determining whether it should be included in the Kyoto accounting system, for two reasons. First, stock changes that are part of a natural cycle will not have a long-term effect on climate if increases during one period or at one location are offset by decreases at another time or place (recognizing that the 5-year commitment period defined by the Kyoto Protocol samples a period that is short compared to many natural cycles of disturbance and recovery). Second, policymakers may decide that stock changes from certain causes should not be credited or debited because doing so would be at odds with the underlying objectives of the Kyoto Protocol.
ARD activities under Article 3.3 and activities that might be added to the Kyoto Protocol under Article 3.4 must be "human-induced." The significance of this limitation can be understood by considering changes in carbon stocks at the global scale that are presumably not a result of human activity. Most pronounced is the annual cycle of increasing terrestrial carbon stocks (decreasing CO2 concentrations) during the Northern Hemisphere spring and summer and decreasing carbon stocks during the autumn and winter (see Chapter 1). To the extent that this cycle is balanced on an annual basis between carbon uptake and release, there is no impact on average atmospheric GHG concentrations, and there would be no net change in carbon stocks observed during a 5-year commitment period. The Mauna Loa record, however, exhibits significant interannual variability around the long-term trend of increasing concentrations. These variations, which are associated in part with natural climate variations such as the El Ni�o Southern Oscillation (Keeling et al., 1989), have resulted in residual carbon stock changes exceeding 4 Gt C during some 5-year intervals since 1980 (Keeling and Whorf, 1998)-which is similar in magnitude to the annual emissions from Annex I Parties. Thus, at a global level, significant carbon stock changes that are unrelated to human activities and unlikely to reflect long-term changes in carbon sequestration can be expected over a 5-year commitment period.
At the national, regional, or local scale, many natural factors could lead
to significant carbon stock changes over a 5-year commitment period. These factors
include natural cycles of disturbance and recovery (e.g., from fires, pests,
and diseases). To the extent that such disturbances are random, their effect
will tend to average out as larger spatial and temporal scales are considered.
However, decisions about accounting for these stock changes, both during commitment
periods and from one period to another, could result in an accounting imbalance-for
example, if disturbances are not accounted for but recovery from disturbance
is considered a management activity.
Although land-use change and forestry require human decisions and action, in some cases the decisions and actions may be largely driven by acts of nature. For example, a patch of forest may be destroyed by landslides or flood damage, leaving no option to regenerate forest on that site in the near future. Although a decision may be made to convert the area to cropping or grazing, this change arguably would not be human-induced because regeneration was not possible. In other cases, the argument for attributing the loss of forest and subsequent land-use change to human activity may be stronger. For example, a wildfire might destroy a tropical forest. Although the forest would regenerate naturally over the next few decades, local land managers or settlers may be opportunistic and convert the land to gardens or plantations.
In managed forests, natural occurrences such as wildfires and pest infestations may have contributing human factors. Indeed, management practices may specify that certain fires should be allowed to burn for ecological reasons. Management practices may also exacerbate the damage from fires, pests, or disease. Arguably, failure to prevent or suppress such events, when prevention or suppression is possible, amounts to human inducement. Thus, non-human-induced occurrences may be limited to places and events that are entirely removed from human activity. Management might be taken as a token of human inducement, and events in managed forests might be considered human-induced per se. The IPCC Guidelines provide for forests to be designated as managed or unmanaged, with stock changes on unmanaged forests excluded from the accounting framework. Alternatively, any event that is not directly attributable to an activity under Articles 3.3 or 3.4 might be considered to be outside the accounting system. This approach, however, could result in a significant divergence between carbon stock accounts and carbon on the ground, which could increase the difficulty of monitoring and verification.
Several commentators have suggested that only intentional activities should be regarded as human-induced. Intent could be very difficult to ascertain, however. It may not be possible to determine, for example, whether a fire was caused by a lightning strike, arson, or mere carelessness. Even if such a distinction were possible, there is no obvious policy rationale for doing so because Parties have a limited ability to control any of these events.
Even if such definitional issues can be resolved, in practice distinguishing
natural stock changes from those that are directly or indirectly human-induced
may be difficult. Suppose, for example, that a natural forest is placed under
management to enhance its growth. Quantifying the change in stock that is directly
attributable to management practices would be difficult; some growth would have
occurred naturally, and some will be the result of CO2 or nitrogen fertilization.
Although the stock increase attributable to management may not be directly quantifiable,
inferring it from the measured stock increase may be possible by using a comparison
"control" plot or a model to estimate the stock increase that would have occurred
without this management practice.
Policymakers should understand that the phrase "human-induced" has no scientific meaning, and guidance will be needed to ensure that Parties are consistent in their characterization of activities as natural or human-induced. Even with clear definitions, in practice distinguishing between changes resulting from human activity and changes caused by natural events may be difficult.
Other reports in this collection