Global plant biodiversity assessment and monitoring depends on large scale, long-term, and spatially complete data. These data can only be acquired with remote sensing. However, careful tests based on standardized field surveys are needed before applying remotely sensed metrics in biodiversity research and conservation. Here, we use open-source imaging spectroscopy and plant inventory data collected across 30 sites covering all major biomes in the US by the National Observatory Network (NEON) to test the degree to which spectral diversity predicts plant biodiversity on the ground.
Results/Conclusions
We found that spectral dissimilarity among plant inventory plots measuring 20 m x 20 m predicts dissimilarity in species composition, or plant beta-diversity, across all sites and ecosystems. In contrast, the spectral alpha-diversity—plant diversity relationship depended on ecosystem characteristics, including canopy density and plant to pixel size. Spectral beta-diversity is particularly relevant for remote sensing of biodiversity because beta-diversity is derived at the plant community scale, which is the scale captured by airborne and spaceborne imaging spectrometers across ecosystems. Assessing spectral beta-diversity in space and over time might allow identifying areas containing unique plant assemblages that may be of significant conservation value, and detecting changes in community composition that can provide early-warning signs of ecosystem transition. Such ecosystem assessments from a distance will provide the information to promote targeted field campaigns and to develop strategies for mitigating detrimental ecosystem change in areas that need them the most.
