Espectroscopia VIS-NIR aplicada à identificação de espécies em campo e avaliação da tolerância à seca em diferentes ecossistemas amazônicos

Le résumé fourni par la source

Recognized as the world’s largest center of tropical tree diversity, the Amazon still has significant gaps in our understanding of the ecology and taxonomy of its species. Given the intensification of climate change and environmental degradation, the development of methodologies to accelerate the collection of ecological and taxonomic information has become urgent and is essential for expanding biodiversity monitoring and informing conservation strategies. This thesis integrates reflectance spectroscopy and plant ecophysiology. The study investigates the potential of VIS–NIR spectroscopy to minimize taxonomic identification errors in forest inventories, assesses the variation in the turgor loss point (TLP) across different Amazonian ecosystems, and tests leaf reflectance methods to accelerate the large-scale quantification of this hydraulic trait. The study was conducted in three contrasting Amazonian ecosystems: terra-firme forest, campinarana, and igapó floodplain forest, located in the central and eastern regions of the Amazon basin. In the first chapter, the potential of reflectance spectroscopy for identifying tree species in the field was evaluated using different plant tissues (outer bark, inner bark, and fresh leaves). The spectral models demonstrated high accuracy in species discrimination. The results demonstrate the applicability of spectroscopy as a tool to optimize forest inventories in highly diverse environments. In the second chapter, the variation in TLP across tree species distributed across the three ecosystems was investigated, as well as its relationship with leaf, wood, and nutrient functional traits. The results indicated that TLP did not differ significantly among ecosystems, showing wide variation among species. In the third chapter, the potential of VIS–NIR spectroscopy to estimate TLP and pressure–volume (P–V) curve parameters was evaluated. Models based directly on leaf spectra performed poorly in predicting TLP. In contrast, models applied to data obtained during the leaf dehydration process performed well in estimating water potential and relative water content, allowing for the reconstruction of P–V curves and the indirect determination of TLP. Overall, the results demonstrate that reflectance spectroscopy has high potential for both species identification and the estimation of physiological attributes, contributing to overcoming operational limitations in ecological studies in the Amazon. The integration of spectral and ecophysiological approaches expands the possibilities for investigating the diversity and functioning of tropical forests, providing insights for monitoring and understanding species’ responses to environmental conditions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.