Treetops at risk : challenges of global canopy
- Authors
- Publication Date
- Jan 01, 2013
- Source
- Horizon / Pleins textes
- Keywords
- Language
- English
- License
- Unknown
- External links
Abstract
The structural organization of a forest canopy is an important descriptor that may provide spatial information for vegetation mapping and management planning, such as attributes of plant species distributions, intensity of disturbances, aboveground biomass or carbon stock. A variety of airborne and satellite images characterize forest stands from above the canopy, providing the advantage of a rapid exploration of extensive and sometimes inaccessible zones. Unfortunately, this approach has limited applicability in wet tropical regions, because most optical and radar signals that deliver medium to high spatial resolution data will saturate at intermediate levels of biomass ranges (ca. 150200 t.ha−1) or leaf area index values (Gibbs et al. 2007). As a consequence, while forest vs. non-forest classifications are nowadays routinely performed from such data, variations in stand structure and biomass within forests of fairly closed canopy remain almost undetectable with classical techniques, and the forest treetops seen from above appear as a homogeneously undulating green carpet. However, rainforest structure varies substantially from place to place either naturally (as the soil, composition or forest dynamics vary) or from anthropogenic degradations. Detecting, characterizing and mapping these variations over vast areas are critical to emerging policies, such as the REDD+ agenda (Maniatis and Mollicone 2010), whereby participating countries will monitor their carbon stock variation. Although promising technology such as LiDAR (light detection and ranging) has great potential, they remain very expensive to systematically assess large expanses of tropical forests (but see Asner et al. 2010). We propose as a cost-effective alternative canopy grain texture analysis from very-high-resolution air- or space-borne images, which proved efficient for retrieving and mapping stand structure parameters including aboveground biomass over vast poorly documented areas of tropical forest.