Early Stages of Brown Rot Fungi Detection: NIR Spectroscopy Integrated with Chemometrics

Abstract

The southern U.S. is an important region for the forestry industry, playing a crucial role in the economy and the environment, being an important region of loblolly pine (Pinus taeda L) timber production in the U.S. Unfortunately, the occurrence of natural disasters like a hurricane can not only increase costs associated with salvaging operations of storm-damaged timber but also increase the vulnerability of forests to pests and diseases. One of these is the brown rot fungi (BRF), which are well known to attack cellulose, leaving behind a residue composed of lignin. It is the most common and damaging type of fungal decay of wood in-service. Hurricanes often cause a large amount of timber and damaged wood, creating an ideal environment for fungal decay. Landowners are interested in salvaging their timber, but the utilization of this downed timber is going to rely on the grade of the timber that is damaged and decayed, as well as the length of time between the disturbance and harvest. Current methods used to determine fungal decay in wood can be expensive, time-consuming, and ineffective for detecting early stages of decay.

One alternative that has gained attention in recent years is near-infrared spectroscopy (NIRS) technology, which can rapidly and simultaneously measure the properties of wood affected by fungal decay without destroying the sample. The NIRS can be used in combination with chemometrics to establish the relationship between near infrared (NIR) spectra and a tedious traditional method, that usually involves chemistry and enables the prediction of a key property of interest. Principal component regression (PCR) and partial least squares regression (PLS) are two popular approaches used in the chemometric field. Several models have been created to predict chemical properties of loblolly pine, but little research has been focused on using NIRS to evaluate the chemical changes in loblolly pine caused by BRF using degradation.

The NIRS coupled with chemometric models might work as a powerful tool that could benefit the forestry sector enabling managers to make salvage decisions based on real-time decay statues and prioritize healthy timber. This work explored the feasibility of defining the relationship between chemical composition of loblolly pine wood with NIRS to predict brown rot decay. Two decay experiments were conducted with seven weeks and fifteen weeks durations. Wet chemistry analysis was performed on each sample to determine acid-soluble lignin and acid-insoluble lignin. The calibration of the chemometric models were performed with the sum of these two fractions using the Spectrum Quant+ software. Finally, a validation step was performed for each model to evaluate its predictive performance.

Results indicated that the reference data explained more than 70% of the variation in the estimated values during the calibration process. Validation results demonstrated that the models could predict changes in wood lignin with an SEP of ± 3.32% for PCR model and ± 1.29% for PLS model. In the future, the NIRS chemometric models could serve as a valuable quality control tool, benefiting the forestry sector by enabling landowners to make salvage decisions based on real time decay status and prioritize healthy timber