Wed, Aug 17, 2022: 8:30 AM-8:45 AM
518A
Background/Question/MethodsOptical remote sensing permits modelling of variables related to forest biomass, which is a critical determinant of carbon stocks and fluxes. While spectral identity and dominance (i.e., averaged spectra within plots) have been successfully linked to plant functional composition, measures of spectral diversity have also been used recently to assess plant diversity. However, it remains unclear whether the link between remote sensing and C content is driven largely by tree composition or tree diversity. Here, we examine the relationship between hyperspectral reflectance and aboveground C content in forests, testing the relative importance of tree composition and diversity in mediating this relationship. We use airborne hyperspectral data in combination with field data on trees and their crowns within 64 field plots in southern Québec, Canada. We calculated (i) spectral composition by means of ordination, and (ii) spectral diversity using the convex hull volume of the first three ordination axes. From field data, we calculated variables characterizing the composition and diversity of canopy trees and C content. We applied two structural equation models based on partial-least squares to test both direct effects of spectral composition or diversity on C storage of trees, and indirect effects via on-the-ground tree composition or diversity, respectively.
Results/ConclusionsWe found that spectral composition in forest plots is a better predictor of C content than spectral diversity. Spectral composition is related to C content largely indirectly, via changes in taxonomic, functional, and phylogenetic composition along the elevational gradient (a transition from deciduous to coniferous species). In contrast, tree diversity did not mediate the spectral diversity – C content relationship, and the moderate direct effect of spectral diversity on C content suggests that spectral diversity captures information beyond what is measured by field-based tree diversity indices. Overall, our findings support the hypothesis that spectral identity or composition is more important than spectral diversity for estimating C content, and that hyperspectral remote sensing can be effectively used as a surrogate of taxonomic, functional, and phylogenetic composition of tree communities with strong links to C storage.
Results/ConclusionsWe found that spectral composition in forest plots is a better predictor of C content than spectral diversity. Spectral composition is related to C content largely indirectly, via changes in taxonomic, functional, and phylogenetic composition along the elevational gradient (a transition from deciduous to coniferous species). In contrast, tree diversity did not mediate the spectral diversity – C content relationship, and the moderate direct effect of spectral diversity on C content suggests that spectral diversity captures information beyond what is measured by field-based tree diversity indices. Overall, our findings support the hypothesis that spectral identity or composition is more important than spectral diversity for estimating C content, and that hyperspectral remote sensing can be effectively used as a surrogate of taxonomic, functional, and phylogenetic composition of tree communities with strong links to C storage.