Evaluating the potential of oxygen isoscapes for tropical timber tracing

Independent verification of timber origin is needed to enforce legislation aimed at combatting illegal tropical timber trade. A potential technique is tracing back the stable isotope signal preserved in wood samples, but the scarcity of reference data currently hampers its operationalization. This can be overcome by creating isoscapes. Here we develop continental isoscapes (at 0.5° resolution) for five tropical timbers based on wood δ18O ratios and assess their potential for timber tracing. We compiled a pantropical database of δ18O measurements from 712 trees in 20 countries. We tested effects of δ18O in rainfall, potential evapotranspiration (PET), temperature and precipitation on wood δ18O and used these to develop isoscapes based on quantile regression forests. A first indication of the tracing potential of these isoscapes was tested in leave one out cross validation (LOOCV) analyses. Across the five isoscapes, ranges in wood δ18O values (10th-90th percentile) averaged 3.9 ‰ and δ18O differences increased with distance. Yet local variability in wood δ18O was substantial compared to large-scale variability. The LOOCV analysis showed that the actual origin was included in the probable origin for 59–79 % of the cases. The area of probable origin was large, however, suggesting a low spatial precision of assignment. This study finds limited support for a potential to use wood oxygen isoscapes for tropical timber tracing within continents. Necessary future steps in timber isotope tracing include improving regional representation, conducting similar analyses for other isotopes, rigorous testing of species differences and conducting blind sample tests.