UCLA

Abstract: Although tropical forests differ substantially in form and function, they are often represented as a single biome in global change models, hindering understanding of how different tropical forests will respond to environmental change. The response of the tropical forest biome to environmental change is strongly influenced by forest type. Forest types differ based on functional traits and forest structure, which are readily derived from high resolution airborne remotely sensed data. Whether the spatial resolution of emerging satellite-derived hyperspectral data is sufficient to identify different tropical forest types is unclear. Here, we resample airborne remotely sensed forest data at spatial resolutions relevant to satellite remote sensing (30 m) across two sites in Malaysian Borneo. Using principal component and cluster analysis, we derive and map seven forest types. We find ecologically relevant variations in forest type that correspond to substantial differences in carbon stock, growth, and mortality rate. We find leaf mass per area and canopy phosphorus are critical traits for distinguishing forest type. Our findings highlight the importance of these parameters for accurately mapping tropical forest types using space borne observations.