Speaker
Description
Plants mediate - through their traits - between the environment and ecosystem functioning. Plants' optical spectra can directly reflect some of these traits. A series of studies have assessed trait estimates or plant-environment relationships from hyperspectral data. Yet these relationships largely differed in biomes, functional aspects, application context, or for different functional groups. Thus, we still miss a universal and overarching understanding of what the main variations of optical data are and what they reflect in terms of functional traits or environmental conditions.
Here we assess the global plant spectral properties with respect to their shared information content with in-situ traits and environmental variables. In detail, we reduced the optical dimensions of a large, global and standardized spectral dataset across leaf and canopy scales to 10 principle components (PCs), and then related these - via models - to a range of traits (individually measured and database extracted) and to environmental variables.
We find up to 88% of spectral variation (capturing for leaf 99.2 % and for canopy 99.58 %) explained by database traits and the climate variables. Importantly, how much these optical PCs are associated with traits and climate, is independent from how much they explain the spectrum. In other words tiny spectral variations reflect important aspects of plant function. E.g. the 9th PC is explained by pigments. Since traits explain a larger fraction than the environment, the optical signal appears to vary much within the same climatic conditions. Generally, the information content for optical PCs decreases from leaf to canopy, area to mass-based traits (for canopy), directly measured to database-derived.
These findings are based on the largest aggregation of leaf and canopy spectra, and have the potential to improve our understanding about which aspects of plant function are reflected by in-situ, and optical data at different scales.
| Status Group | Postdoctoral Researcher |
|---|
