Abiotic and biotic stress can negatively influence forest health. While approaches based on plant functional traits (PFT) have been used for assessing tree health, collection of PFT using traditional field techniques is logistically challenging across large spatial scales, revealing limitations in using PFT approaches in a management context. Recent technological advances in optical monitoring methods have enabled the development of digital forestry; yet application of these approaches in identifying specific responses to abiotic and biotic stress has lagged behind. One approach that contains considerable information related with plant chemistry and physiology is hyperspectral data. Foliar reflectance is responsive to stress and thus can be used to detect different stress events. Here, our research objectives were to determine the ability of hyperspectral data to 1) estimate PFT responses to stress and 2) classify different abiotic and biotic stress factors. In a greenhouse environment, we exposed one-year-old black walnut (Juglans nigra) and red oak (Quercus rubra) saplings to multiple stress factors, alone and in combination, including water limitation, nutrient deficiency, and salinity stress and infection of fungal associates(Geosmithia morbida and Fusarium solani). We collected reference measurements of numerous leaf traits (photosynthetic, water, and defense related traits) and paired them with spectral collections.
Results/Conclusions
Preliminary results showed that spectral models can reliably estimate most black walnut and red oak leaf functional traits in response to multiple different stressors. Photosynthetic and water related PFT were responsive to variation in water availability, while foliar chemical changes were responsive to pathogen infections. Success in classification approaches for the different stressors depended on if the stress events were alone or in combination. This research highlights the ability of hyperspectral reflectance to 1) estimate foliar functional traits affected by diverse stressors and 2) classify different stress combinations. Continued work in relating high-dimensional spectral information with plant stress under controlled environments will improve management efforts in forest systems and provide the foundation for transferring information to larger spatial scales (e.g., UAV and manned aircrafts) for management in forest plantation systems.