2.3.7. Accounting for Uncertainty
Uncertainty in accounting for land-use change and forestry activities includes
not only measurement uncertainty but uncertainty in defining and quantifying
baselines (when they are used) and uncertainty related to the interpretation
of the Protocol's requirements, including the definitions of key terms. Discussion
of the magnitude of these uncertainties and approaches for minimizing them appear
in Section 2.2 and in subsequent chapters related to specific
Articles. This section briefly describes options for incorporating uncertainty
into the accounting framework.
All GHG emissions and removals reported under the Kyoto Protocol will be subject
to uncertainty to varying degrees. Dispersed emission sources, for example,
face many of the same measurement challenges as LULUCF activities. Uncertainty
enters the accounting system in a different way under the accounting framework
adopted for LULUCF, however, than it does for sources that form part of the
baseline for the Kyoto Protocol. In the latter case, systematic and random errors
will be present in the 1990 emissions baseline and the emissions inventory during
the commitment period. As long as consistent methods are used to estimate emissions
in both periods, the potential to introduce bias into the accounting system
will be minimized. On the other hand, LULUCF activities for which Annex I countries
gain credit will only enter as a credit during the commitment period for most
Annex I Parties. As a result, systematic errors are not offset through subtraction
of the same error during the baseline period. Similarly, any change in measurement
methods would affect only net emissions during the commitment period, without
a compensating change in emission baselines (and thus assigned amounts under
the Kyoto Protocol).
Accounting rules can be used to adjust for data limitations and uncertainty.
For example, conservative estimates of carbon benefits can be applied by including
all pools expected to have reduced carbon stocks and only a selection of pools
expected to have increased carbon stocks; only the pools that are measured and
monitored would be claimed as a carbon benefit (Sathaye et al., 1997).
Similarly, measurement uncertainty could be accounted for by adjusting estimated
fluxes, based on the uncertainty in this estimate, in the direction that is
likely to understate removals and overstate emissions. This approach would provide
an incentive to reduce uncertainties to the extent that it is cost-effective
to do so, but it would not require expensive monitoring of carbon pools that
do not significantly affect the overall carbon balance of a site.