2.4.4.5. Future Remote-Sensing Systems
Many remote-sensing systems are expected to be implemented over the next several
years, and plans exist for many more between now and the first commitment period.
These systems include missions recently launched by the United States (including
public and private systems), India, Brazil and China, Europe, Japan, and others.
A detailed description of the plans for all of these systems is beyond the scope
of this section.
Several points are important in the context of the first commitment period
(and beyond), however. Currently, there is no published national or international
commitment to ensure that remote-sensing measurements adequate to determine
areas of forest clearing and regrowth will be collected during the first commitment
period. Systems that currently provide such information-such as Landsat, SPOT,
IRS, and CBERS-have current operational lifetimes that nominally end before
the first commitment period begins. Ensuring that remote-sensing data adequate
to this task will be available will require public and possibly private investment
and a commitment by governments to ensure its availability. The situation with
respect to remotely sensed data for ecosystem metabolism models and fire detection
is less uncertain. Under current plans, these data will be included in joint
U.S.-European meteorological missions that are projected to orbit during the
first commitment period.
2.4.5. Models
Bookkeeping models have been used for years to calculate the effects of land-use
change on terrestrial carbon storage (Moore et al., 1981; Houghton et
al., 1983; Woodwell et al., 1983). The same models may be used to
calculate annual emissions and accumulations of carbon at any scale-from the
project level to national, regional, and global levels. Two types of information
are required: rates of land-use change (e.g., areas cleared annually for croplands,
cubic meters of wood annually harvested) and per-hectare changes in carbon stocks
following a change in land use (including changes in living vegetation, logging
debris, harvested products, and soil carbon). The IPCC Guidelines are based
on simplifications of these models. These models generally consider only fluxes
of carbon resulting from land-use change, including forestry. They do not include
the indirect effects of human-induced changes, such as climate, CO2, or nitrogen
deposition.
A similar set of bookkeeping models that include natural disturbances such
as fire, insects, or wind-throw have been used to determine the carbon balance
for large areas of forest (e.g., Kurz and Apps, 1995, 1999). These models require
a third type of information-namely, areas annually disturbed by these factors,
as well as by harvesting.
Process-based models do include the effects of environmental variables on the
carbon dynamics of ecosystems; such models have some potential for estimating
natural as well as human-induced changes in carbon stocks. These models also
have the potential to extrapolate site-based measurements temporally and spatially;
to attribute changes in carbon stock to direct human-induced and naturally occurring
change; and to estimate emissions or consumption of trace gases such as NOX,
N2O, and CH4-although the current capacity to do so is not sufficiently accurate
for emissions trading purposes. Before these models can be used to assess changes
in carbon stocks, they must be thoroughly documented, evaluated against independent
data sets (e.g., from long-term forestry or agronomic experiments), and, if
necessary, calibrated for local conditions. Such models, when correctly parameterized,
can be effectively used at the time scale of the commitment period, but their
use over longer periods may be limited.
The range of available models enables simulation of carbon dynamics at various
scales, from plot level (<1 ha) to global level. As the scale increases, the
spatial resolution of the model decreases and the input data become more aggregated.
Some models deal with single ecosystem components such as soil (e.g., RothC;
Falloon et al., 1998), whereas others simulate whole ecosystems (e.g.,
Century; Parton et al., 1987). Other models operate at the biome level
and are most often applied at the continental and global scale (Meentenmeyer
et al., 1985; Heimann et al., 1998). To estimate and project changes
in carbon stock for particular locations or specific combinations of environmental
and management conditions, detailed biogeochemical process models can be employed.
Such models have high data input requirements and may need particular calibration
or parameters for individual locations. Model drivers include parameters such
as soil clay content and topographic position, as well as climate data and land
management information. Several reviews describe many of these models in detail
(e.g., McVoy et al., 1995; Smith et al., 1997c; Heimann et
al., 1998).
The use of models to estimate changes in carbon stock may result in a loss
of transparency because the modeling process can be very complex. If models
are used, they must be documented and archived in the form in which they were
used (including source code and input files), and the validation data must form
part of the reporting requirement. With a view to using models for subsequent
inventories, it is suggested that initial inventories gather data to evaluate
model outputs in addition to measures that are required solely for inventory
purposes. The incremental cost of doing so is small, and the data can be used
for cross-validation.
|