Tropical climates - The Global Climate System: Patterns, Processes, and Teleconnections - page 82

grass for cattle grazing and/or crops. A major release of CO 2 occurs as forests
are replaced by forms of low biomass vegetation. Commonly, after a few
years of crop or grazing, increased sequestration of carbon from CO 2 occurs
when shrubs, small trees, and dense grasses become established as initial
secondary succession (SS1). Passage of another 5-12 years usually results in
trees again becoming dominant with a relatively dense cover resulting in the
establishment of an intermediate secondary succession forest (SS2). Change
in land use from SS1 to SS2 results in further sequestration of carbon derived
from CO 2 . If SS2 is left undisturbed for an additional 10-15 years, an
advanced SS forest that has an appearance similar to the original moist mature
forest often appears (SS3). The SS3 forests have developed a multi-canopy
and have many trees almost as tall as the original moist mature forest, but
their species complexity and total biomass is less than the original forest.
Even after 30-50 years of re-growth in a good physical environment (and
much longer than 100 years or never in poorer nutrient environments)
succession vegetation has less biomass than the original forest.
The LULC sequence from mature forest to crop/pasture to SS1 to SS2 to
SS3 to mature forest conditions again is often interrupted. At any SS stage,
human activity can intercede by reintroducing crops, grazing economies, and
agroforestry. The reintroduction of crops and grazing in a SS forest creates a
net increase in CO 2 released and a net decrease in sequestered carbon. The
introduction of agroforestry in a SS forest makes little change in carbon gain
or loss to the atmosphere.
Remote sensing of LULC changes: implications to
carbon sequestration
The most important LULC classes in Amazonia to study biomass changes,
from highest sequestered carbon/high biomass to lowest, are moist mature
forest/rainforest, SS3, SS2, SS1, agriculture (grazing/crop), bare or nearly
bare (e.g. bare soil, road, very sparse grass), and water. Accurate classifica-
tion of these LULC features in Amazonia using remote sensing was histori-
cally difficult to achieve. Spectral data acquired from low-resolution satellite
sensors can classify huge areas, but primarily identify mature forest, succes-
sion vegetation, bare, and water, often with accuracy problems. Low-resolution
sensors provide a general insight into the location and extent of deforestation/
afforestation over large areas, but they do not provide data suitable to identify
LULC changes in the detail required to assess the amount and trends of carbon
sequestration and CO 2 released to the atmosphere or to help explain rates and
nature of SS growth.
High-resolution satellite and airborne sensors can acquire spectral data at
1-meter resolution or less. However, their use is costly and too much analysis
is needed to provide subregional information useful for LULC change