Combining wood traits as a promising timber origin verification and its application in the Brazilian trade chain

Tracing the geographic origin of wood remains a major challenge in the fight against illegal timber in tropical countries. Methods using anatomical, chemical, isotopic and DNA markers have been successfully tested, but methods are either expensive, unpractical in the field or time consuming. By analyzing 17 Neotropical populations of Cedrela spp., we investigated the potential to improve wood provenance identification by combining wood traits measured with field adaptable equipment. Using machine learning models, we demonstrate that Ca, K, S, Al, wood density and tree growth rates predict wood origin with over 80 % accuracy at the regional scale and 63 % accuracy at the site level. Climate and soil conditions are the primary drivers of wood traits, particularly Ca, highlighting its value as a "fingerprint." National and international efforts to build robust reference databases are needed. Our cost-effective and reliable method for tracing wood origins can be a powerful tool to aid law enforcement in fighting illegal timber trade.