Recent work suggested that near infrared (NIR) spectroscopy may be able to estimate extractive concentration in western redcedar (WRC) heartwood. This would facilitate the screening of breeding stock that will produce durable heartwood. Increment cores from 50 trees were scanned and analysed for extractives content to evaluate existing models. Extractives data were not accurately predicted by the developed models. This was likely due in part to the much smaller range of extractives in the increment core data set. Subsequent partial least squares (PLS) models based on the increment core data set confirmed that NIR is not sensitive enough to small variations in extractives to differentiate wood with small or medium differences in extractives content.
The present work also examined the ability of PLS models to predict decay resistance from NIR spectra. The developed models were not able to accurately predict weight loss caused by Coniophora puteana in a soil block test. Correlations between extractives data and decay resistance data were very weak. This weakness may have been exacerbated by the leaching and biodegradation steps that would have reduced the concentration of thujaplicins in the samples.