A new statistical approach can pinpoint where timber was grown more accurately than the methods currently used, particularly when reference data for a given tree species is thin. The study, published in the journal Annals of Forest Science and led by Jakub Truszkowski, shows that combining chemical information from several tree genera improves the prediction of harvest locations.
Tracing the origin of wood is a practical problem for customs officials, regulators and buyers trying to confirm that timber was legally sourced. Existing methods rely on comparing a sample's chemical signature against a reference database for the same species. Where that database has gaps, the predictions become unreliable. The new model addresses this by drawing on patterns shared across genera, filling in what a single-species approach cannot.
The team tested the model on four genera, Betula, Fagus, Pinus and Quercus, using 1,020 tree samples collected from 12 Eastern European countries. Compared with a model built on a single genus, the multi-genus version reduced the location error for oak (Quercus) by 38 percent. For pine (Pinus), it raised the sensitivity of verifying origin by 49 percent. The gains were largest in settings with limited data and were similar to existing methods when abundant reference samples were available.
The cost of testing also matters for enforcement. According to the study's figures, trace element analysis runs roughly 20 to 120 dollars per sample, while isotope ratio analysis costs roughly 200 to 500 dollars. A more accurate model could make cheaper testing go further. The authors note that it could be useful where direct sampling is blocked by conflict, citing restrictions on Russian timber under sanctions as an example.
The researchers suggest the approach could later be extended to tropical timber and other commodities. They caution, however, that its performance outside temperate Eastern European forests remains uncertain and has not yet been established.
Credit. Based on reporting by Alan Morgan for Science. Is this your work? Write to us.
15
