More than a quarter of global forest loss is related to raw material production, including deforestation. Although illegal logging provides wood for products such as furniture and window sills, independent verification of the origin of the wood in the final product is difficult. Now researchers have found a new way to identify the tree.
The chemical composition of wood varies geographically. Scientists from Wageningen University and the Netherlands Research Company analyzed wood samples taken from about 1,000 different trees in Cameroon, Congo and Gabon in central Africa, as well as parts of the Indonesian island of Borneo, and determined that it can be detected. their origin at the subnational scale. The researchers see the analysis as a step toward a global tree-tracking tool.
"The main motivation [for conducting the study] was to further improve small-scale detection," said data scientist Laura Boeschoten, "but ultimately to reduce the illegal timber trade." Boishoten is the lead author of the study, and his Ph.D. Wageningen candidate. Current timber tracing methods do not consistently limit the origin of samples to an area of less than 100 kilometers, the distance required to accurately identify illegally logged timber.
Rail elements for wooden rails
The team sampled three economically important tropical forests: azobe and tali in central Africa and red meranti in Borneo. They recorded the GPS coordinates of trees located in logging concessions in state-designated forest areas where logging is managed by private entities.
The researchers dissolved each sample and measured the amounts of about 40 elements, including magnesium, calcium and lanthanum, developing a fingerprint of each logging concession. They then developed a task model that used machine learning to determine the most likely sample origin. The assignment model is trained on a random subset of the database, after which the test sample is assigned to the most likely origin.
Because the chemical composition of trees varies by region, machine learning can distinguish the major elements present in a given region. For example, Peter Zuidemann, a forest ecologist at Wageningen University and study co-author, said samples showed that western Cameroon had high molybdenum concentrations, while eastern Republic of Congo had higher sodium concentrations and higher silicon concentrations. North of Borneo.
The researchers found that the samples could be traced back to their sub-national origins in Central Africa with 86% and 98% accuracy, and to forests of origin in Borneo with 88% accuracy. When the samples were submitted by an independent third party for actual analysis, the models identified their subnational origin with 70% to 72% accuracy.
One of the limitations of this tool is that it only asks, "Does this tree come from a specific area?" can answer the question. but "Where did this tree come from?" Zuidema said: Future research should improve the tool to make it more open.
Implement legislative objectives
Similar multivariate analyzes are available for other commodities, including asparagus, bananas, and tea. "But this is the first time it has been applied to wood," says Zuidema.
The analysis can be an important tool given the European Union's (EU) anti-forestry law, which requires companies to prove that products sold in the EU do not cause deforestation or forest degradation under the laws of their country of origin. . . The law regulates products such as beef, cocoa, coffee, palm oil, rubber, soybeans and timber.
"I really see high potential for this technique and every new discovery is a big step in the right direction," said Victor Declercq, head of research for World Forest ID at the Royal Botanic Gardens, Kew, England. But he also admitted that the chemistry of the wood was unclear. "Trace element amounts can vary between years and tree rings, and more research is needed to see how consistent the signals are," he said.
—Rishika Pardikar (@rishpardikar), science writer
