Biodiversity-mediated ecosystem functioning in the Amazon: a remote sensing approach

The megadiverse system of the Amazon contributes to many local and global ecosystem processes with potential to trigger an irreversible shift in the functioning of our planet. Yet, Amazon’s biodiversity is complex and remains mostly understudied. Biodiversity distribution patterns likely affect the functioning of this crucial system, yet large scale systematic assessments are still lacking. Herein we examine how can optical remote sensing contribute to understanding biodiversity patterns in the Amazon. We use the spectral diversity approach to map the heterogeneity of spectral signatures as a proxy for heterogeneity in canopy composition using Sentinel-2 imagery over the entire basin. Our map is able to reproduce patterns of primary, secondary and deforested areas, and within the forest areas, the patterns of spectral richness and spectral turnover resemble those expected for the Amazon system, with rapid turnover in species composition closer to the waterways and more stable plant community compositions away from the waterways. We then examine whether the locations with higher or lower spectral variability correspond to different community and trait compositions. We examine the relationship between spectral richness and turnover and the presence of hyperdominant trees in the Amazon. We also examine the multivariate trait space including Specific leaf area (SLA), Leaf dry matter content (LDMC), Leaf nitrogen content (LNC), and Leaf Phosphorous content (LPC) along the gradient of spectral variability to find that trait syndromes vary along the gradient of spectral diversity. These results show that biodiversity biodiversity gradients among the Amazon may explain differences in ecosystem functioning and that approaches such as the spectral heterogeneity may start to shed some light into such relationships, especially over entire and important ecosystems like the Amazon. Further examinations of these processes and relationships are therefore required and will contribute to our better understanding of the feedbacks between biodiversity and earth system processes.