ABSTRACT
Tropical forests (TFs) are globally important due to their extremely high species diversity and major role in biogeochemical cycles. Spatial and temporal sampling of these forests in the œeld is greatly restricted due to prohibitive costs and inaccessibility at the ground and canopy level, and extrapolating plot data to broader spatial scales is problematic as species distributions and ecosystem processes vary at different spatial scales due to overlapping factors, such as climate, soil, and past disturbance. Remote sensing plays an important role in mapping TFs over larger areas than available from the ground, especially with the recent use of new sensors, processing and modeling techniques, and integration with œeld data [1]. In particular, a new generation of hyperspectral, hyperspatial, and hypertemporal sensors is making rapid and innovative advances in understanding TF composition, structure, and function over broad spatial and temporal scales [1]. These advances include canopy chemistry by hyperspectral sensors [2-6]; analysis of individual tree crowns (ITCs) [7,8] and canopy components, such as lianas [9] by hyperspatial (<4 m) sensors; and new insights into canopy chemistry, physiology, and phenology by hypertemporal sensors [4,10,11].
