Understanding Boswellia papyrifera tree secondary metabolites through bark spectral analysis

Decision makers are concerned whether to tap or rest Boswellia Papyrifera trees. Tapping for the production of frankincense is known to deplete carbon reserves from the tree leading to production of less viable seeds, tree carbon starvation and ultimately tree mortality. Decision makers use traditional experience without considering the amount of metabolites stored or depleted from the stem-bark of the tree. This research was designed to come up with a non-destructive B. papyrifera tree metabolite estimation technique relevant for management using spectroscopy. The concentration of biochemicals (metabolites) found in the tree bark was estimated through spectral analysis. Initially, a random sample of 33 trees was selected, the spectra of bark measured with an Analytical Spectral Device (ASD) spectrometer. Bark samples were air dried and ground. Then, 10 g of sample was soaked in Petroleum ether to extract crude metabolites. Further chemical analysis was conducted to quantify and isolate pure metabolite compounds such as incensole acetate and boswellic acid. The crude metabolites, which relate to frankincense produce, were compared to plant properties (such as diameter and crown area) and reflectance spectra of the bark. Moreover, the extract was compared to the ASD spectra using partial least square regression technique (PLSR) and continuum removed spectral analysis. The continuum removed spectral analysis were performed, on two wavelength regions (1275-1663 and 1836-2217) identified through PLSR, using absorption features such as band depth, area, position, asymmetry and the width to characterize and find relationship with the bark extracts. The results show that tree properties such as diameter at breast height (DBH) and the crown area of untapped and healthy trees were strongly correlated to the amount of stored crude metabolites. In addition, the PLSR technique applied to the first derivative transformation of the reflectance spectrum was found to estimate the concentration of the metabolites reliably at higher coefficient of determination. The most influential maximum slope positions of the spectrum obtained through PLSR analysis of the petroleum ether extract (crude metabolites) and the pure compounds (incensole acetate and boswellic acid) were found to coincide and concentrate in the region between 1383-1406 nm and 18611896 nm. However, analysis on these two individual specific region absorption features relationship with the bark extract, using the continuum removed approach, was not as robust as the PLSR analysis. This reveals that the ability to estimate metabolites in the stem-bark of B. papyrifera using spectral analysis opens a new approach on how to manage B. papyrifera tree. Development of such technique provides quick and reliable information for decision makers to decide on when to tap or for how long to rest the trees. However, to implement the technique, further research need to be conducted to determine how B. papyrifera tree metabolites vary due to tapping. In addition, research to determine the lowest metabolites quantity threshold, useful for management of the tree, needs to be conducted. (C) 2015 International Society for Photogrammetry and Remote Sensing, Inc. (ISPRS). Published by Elsevier B.V. All rights reserved.