Estratégias ecológicas associadas aos padrões de sazonalidade da fenologia foliar na Amazônia central

Abstract

Understanding the mechanisms governing tree phenology and mortality in humid tropical forests is crucial for predicting these ecosystems' responses to climate change. This doctoral thesis investigates the complex interactions among environmental gradients, species' functional traits, and trees' internal dynamics in shaping foliar phenological patterns and signaling senescence and death in Central Amazonian terra firme forests. The first chapter addresses the role of spatial gradients in soil water availability on leaf phenology. Through a 52-year analysis of phenological data from hundreds of trees, we demonstrate that water table depth is a key determinant of the frequency, duration, and timing of leaf production. Distinct hydrological environments promote contrasting leaf exchange strategies, even in predominantly evergreen forests. These subtle variations, including the coexistence of mature, old, and new leaves, suggest complex phenological plasticity that confers resilience to local environmental constraints. Our results enhance the representation of phenology in ecological models and contribute to understanding the functionality and diversity of Amazonian forests in a changing environmental scenario. Secondly, the thesis explores the link between tree functional traits and the timing of new leaf flushing. We hypothesize that traits related to water and resource use efficiency, such as embolism resistance (P50, P12) and wood density, are key predictors of new leaf production seasonality. We identified two distinct phenological strategies in evergreen trees: a drought avoidance/resource acquisition strategy, with flushing during wet periods to mitigate hydraulic function losses, and a drought resistance/resource conservation strategy, with flushing during drier periods. Network analysis of traits revealed a modular structure in the correlation among hydraulic, structural, and resource acquisition traits, highlighting the integration of multiple functional dimensions in determining plant strategies. These findings are fundamental to understanding how water and resource use strategies shape the seasonality of leaf productivity in the central Amazon. Finally, we investigate tree mortality from the perspective of physiological self-regulation and early warning signals. We applied statistical methods of resilience indicators to a 52-year phenological dataset to test whether variations in leaf flushing patterns can signal impending death. We show that an increase in variance and critical slowing down in leaf phenology dynamics precede tree death by approximately seven and a half years. This suggests a progressive decline in physiological regulation and vitality, indicating that tree death is not solely driven by external disturbances but can also be shaped by internal factors such as aging and senescence. Additionally, deciduous tree species that died by falling exhibited greater losses in resilience compared to evergreen and standing dead trees. This systemic approach offers unprecedented insights into the use of leaf phenology dynamics to predict tree death and tree homeostasis and paves the way for a deeper understanding of tree aging and death.