Land Use, Land-Use Change and Forestry

Other reports in this collection

3.4.2. Monitoring Land-Cover Change

Scenarios that create ARD land on the basis of a wide range of activities, including harvest/regeneration cycles and natural disturbances followed by regeneration (as in Land Cover or FAO scenarios), will result in a much larger area of ARD land. The data requirements for area determination under such scenarios may be met through approaches that are based on monitoring land-cover change, such as remote sensing. Forest definitions with a high canopy cover threshold likely would lead to more accurate identification of ARD activities with remote sensing than definitions with a low canopy cover threshold because sparse forest conditions are more frequently confused with some types of agriculture or scrub/shrub ecosystems in assessments based on the use of remotely sensed imagery. For the same reason, if regenerated areas with minimal crown cover are considered forest (e.g., FAO scenario), remote-sensing techniques would be unable to reliably detect them. If the forest definition is dependent on a biomass assessment (which is possible in the Flexible scenario), forest areas could not be accurately estimated with current remote sensing systems. In this case, field inventory would almost certainly be required. Table 3-8 summarizes the potential of current remote-sensing systems to detect ARD activities.

Table 3-8: Potential of current remote sensing systems to detect ARD activities.

Definitional
Scenario
Ability to
Detect ARD
Notes

FAO
Low
a
IPCC
Low
a
Land Use
Low
Land Cover
Moderate
a
Flexible
Low
b
Degradation/Aggradation
Low
c
Biome
Moderate
a,d

a Considerable misclassification may exist between sparse or young forest and agriculture or other vegetation types, leading to errors when forest cover thresholds are low.
b Biomass estimates will be of low accuracy.
c Transitions from one cover class at maturity to another within forest are difficult if not impossible to assess accurately.
d Highly variable, depending on defined threshold.


When ARD activities are defined as a change between forest and non-forest, most such activities will be readily discernible using forest/non-forest classifications at least in 1990 and 2012. Some definitions of ARD, however, could mean that land that is in forest status in both 1990 and 2012 might still be included in Article 3.3 (Section 3.4.1), so simple forest classification at the two endpoints would be insufficient to determine ARD land. Thus, information beyond a forest/ non-forest classification at different points in time is required. This information could be gathered through combinations of remote sensing and ground-based sampling. The information-gathering process includes two main tasks:

In many countries, reliable data on forest area, age, and species distribution result from detailed forest inventories. These inventories are usually conducted on periods of 10 to 20 years, however. Thus, assessment of changes between 1990 and 2012 will require application of approximation procedures that yield estimates based on assumptions or auxiliary information about the continuation or alteration of a trend.



Other reports in this collection