Tracing Amazon timber provenance using stable oxygen isotope

• Autor(a) principal: Batista, Ana Claudia Gama
• Data de Publicação: 2025
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Biblioteca Digital de Teses e Dissertações da USP
• Texto Completo: https://www.teses.usp.br/teses/disponiveis/64/64135/tde-12092025-081803/
• topic: Environmental governance; Exploração ilegal de madeira; Governança ambiental; Illegal logging; Isótopo de oxigênio; Machine learning; Machine learning; Modelagem espacial; Oxygen isotope; Rastreabilidade de madeira; Spatial modeling; Wood traceability

The Amazon Rainforest, the most biologically diverse ecosystem on the planet, plays a crucial role in global biogeochemical cycles and climate regulation. However, its biodiversity is increasingly threatened by deforestation and illegal logging, highlighting the urgent need for effective conservation strategies. This study explores the application of stable isotope analysisparticularly oxygen isotopesas a forensic tool for wood traceability, combining geostatistical techniques and machine learning algorithms to enhance the accuracy of origin tracking and environmental monitoring. Chapter 1 contextualizes the ecological importance of the Amazon, emphasizing its role in the oxygen cycle and biodiversity conservation. Despite regulatory frameworks such as the Forest Code and Sustainable Forest Management policies, illegal logging continues, with nearly 40% of timber being extracted without authorization. Stable isotope analysis thus emerges as a promising tool to improve traceability and combat documentation fraud. Chapter 2 presents a standardized protocol for the collection and isotopic analysis of ¹O in wood, developed by CENA/USP, focusing on the evaluation of variability along the tree-ring radius. A total of 258 trees from 21 locations across the Amazon were analyzed. The samples, segmented from pith to bark, underwent -cellulose extraction and were analyzed in two laboratories with interlaboratory calibration to ensure comparability. ANOVA results indicated low variability in ¹O values, attributed to environmental stability in the region. The 75% point along the radius, corresponding to the heartwoodsapwood transition, was considered the most representative. This protocol strengthens wood traceability and has relevant applications in isotopic ecology and forensic science. Chapter 3 examines inter- and intra-site variations in ¹O values from 693 -cellulose samples of 12 species collected at 13 Amazonian sites. The results show that macroclimatic factorsespecially latitude, Isorix, and potential evapotranspiration (PET)exert a greater influence on the ¹O signature than local microclimatic conditions. These findings reinforce the potential of stable isotopes for both paleoclimate reconstructions and the refinement of traceability models. Chapter 4 focuses on the development of ¹O isoscapes using kriging and random forest models. Based on the analysis of 258 trees from 21 locations, the study demonstrates that the random forest model outperforms kriging in predicting the spatial distribution of ¹O, capturing isotopic gradients associated with regional patterns of precipitation and evapotranspiration. These models offer a promising approach for improving the forensic tracking of wood origin. In conclusion, the integration of isotopic techniques with geospatial modeling significantly enhances wood traceability, supporting its application in conservation policy and forest governance. Expanding isotopic databases and incorporating additional environmental variables are essential to refining predictive models and contributing to the sustainable management and protection of Amazonian natural resources