TreeSpec: Quantifying the functional role of canopy structural and chemical traits for productivity across experimental tree diversity gradients using a hyperspectral imaging system
Final Report Abstract
Tree canopy structural and physiological traits are important to help understand how tree species diversity influences ecosystem function. Tree canopy complexity and height make measurement of these traits difficult, requiring the use of novel tools for data acquisition. Unmanned aerial vehicles have been used to date in agricultural systems with a main focus on crop monocultures, though the potential exists to employ these tools to forest systems, which few studies have done thus far. In the TreeSpec project, we tested the potential of UAV hyperspectral imaging to measure important canopy properties of young experimental forest stands, and utilizing three forest diversity experiments within Germany, we aimed to test what role structure and physiological traits play in influencing productivity. Overall, we found UAV hyperspectral imaging is a useful tool to estimate canopy %N and LAI for young tree plantations. We also found that NDVI estimated using the UAV camera data, which represents a measure of how green and healthy forest canopies are, was strongly correlated with a ground based measure of Leaf Area Index, raising the possibility of more rapid measurement of LAI, a key canopy property related to light interception and productivity of tree communities. In addition, NDVI was also found to be a very useful tool for capturing variation in tree species responses to drought, with dominant fast growing species exhibit strong decreases in NDVI that were reflective of ground-based observations during a strong drought in 2018. Our experience also suggested that UAV hyperspectral imaging is still maturing as a technology. Specifically, we experienced an important technical issue that challenged our team to solve in order to generate robust data. We hypothesize the issue was related to the camera getting warm while repeatedly collecting images and thus occasionally failing to fire. This suggests that this technology is still evolving and maturing and demonstrates the value of field testing of potentially promising instruments and methodologies. Finally, although we planned to go beyond remote sensing indices and partial least squares regression, and also employ radiative transfer models to develop detailed information on what aspects of canopy structure drive changes in reflectance observed by the hyperspectral sensor, working with such models requires detailed knowledge and expertise that is still the domain of seasoned remote sensing experts.
Publications
- 2018. Synthesis and future research directions linking tree diversity to growth, survival, and damage in a global network of tree diversity experiments. Environmental and Experimental Botany, 152, 68-89
Grossman, J. J et al.
(See online at https://doi.org/10.1016/j.envexpbot.2017.12.015) - 2018. UAV photogrammetry of forests as a vulnerable process. A sensitivity analysis for a structure from motion RGB-image pipeline. Remote Sensing, 10(6), p.912
Frey, J., Kovach, K., Stemmler, S. and Koch, B.
(See online at https://doi.org/10.3390/rs10060912)
