World Forest ID data analyst, Dr. Jakub Truszkowski, has co-authored a paper presenting an innovative model using Stable Isotope Ratio Analysis (SIRA) data from seven oak species to identify location of harvest for ‘unknown origin’ oak timbers. The advanced methodology bolsters the toolset available for determining timber harvest locations, aiding authorities in tackling the illegally harvested timber trade and enforcing anti-deforestation legislation. By providing detailed and reliable geographic origin determination at the sub-national level, the research demonstrates potential to scale, supporting the protection of forest land and threatened ecosystems.

The work represents a significant step forward in the scientific determination of timber origins, reinforcing World Forest ID in its mission to use innovative technologies to safeguard forests.

Novel Analytical Pipeline

The paper introduces a new SIRA data analysis pipeline designed to infer the harvest location of an individual tree from a large, continuous area. By using Gaussian Processes to estimate isoscapes from reference wood samples and integrating species distribution data, the method offers probabilistic predictions rather than binary outcomes, enhancing confidence in sub-national timber provenance estimates.

Enhanced Accuracy and Probabilistic Modeling

The method can pinpoint the harvest location within 520 km of the true origin of the sample. Incorporating species distribution data improves the accuracy by up to 36%, showcasing the advantages of incorporating different data layers.

Active Learning for Optimized Sampling

The model includes an active learning tool which helps World Forest ID to identify optimal locations for future sample collection, maximizing the improvement to model performance of marginal samples. Strategic sampling reduces the variance of isoscapes by up to 86% more than random sampling, significantly improving return on investment in expeditions.