32
(Liang et al. 2016). Hundreds of rigorous biodiversity experiments have been
designed and conducted to tease apart effects of changing numbers of species (richness) from effects of changing identities of species (composition) (O’Connor et al.
2017; Grossman et al. 2018; Isbell et al. 2018). Complementarity among diverse
plant species that vary in their functional attributes and capture and respond to
resources differently is the primary explanation for increasing productivity with
diversity (Williams et al. 2017). Nevertheless, both the nature of biodiversity-ecosystem function (BEF) relationships and their causal mechanisms remain variable
and scale dependent in natural systems. In the Nutrient Network global grassland
experiments, in which communities have assembled naturally, the relationship
between diversity and productivity is variable (Adler et al. 2011). In tropical forest
plots around the globe, at spatial extents of 0.04 ha or less, the biodiversity-productivity relationship is strong. However, as scales increase to 0.25 or 1.0 ha, the relationship is no longer consistently positive and can frequently be negative (Chisholm
et al. 2013). These varied relationships at contrasting spatial scales may result from
nonlinear, hump-shaped relationships between biodiversity and ecosystem function
across resource availability gradients as the nature of species interactions and their
level of complementarity shift (Jaillard et al. 2014). RS methods—including imaging spectroscopy and LiDAR—that can detect both the diversity and the structure
and function of ecosystems (Martin, Chap. 5; Atkins et al. 2018) can discern these
relationships across spatial extents and biomes in natural systems. They thus have
high potential to enhance our understanding of the scale and context dependence of
linkages between biodiversity and ecosystem function (Grossman et al. 2018).
2.9 Incorporating Spectra into Relationships
Between Biodiversity and Ecosystem Function
Detection of spectral diversity, in particular, offers the potential to contribute to the
quantification of BEF relationships at large scales (Schweiger et al. 2018) and is
thus worth discussing in more detail. The variability captured by spectral diversity
in a given ecosystem depends on the way the spectral diversity is calculated, as
well as its spatial and spectral resolution (Sect. 2.7.5; Gamon et al., Chap. 16).
From a functional perspective, spectral profiles measured at the leaf level depend
on the chemical, structural, morphological, and anatomical characteristics of leaves
(Ustin and Jacquemoud, Chap. 14). Variation in spectra and spectral diversity can
be used to test hypotheses about how specific traits influence ecosystem function,
community composition, and other characteristics of ecosystems, when using
spectral bands or spectral indices with known associations with specific plant traits
(Serbin and Townsend, Chap. 3). Moreover, spectral bands and indices can be
weighted based on prior information about the relative contribution of individual
traits to specific ecosystem characteristics. However, while the absorption features
J. Cavender-Bares et al.
(Liang et al. 2016). Hundreds of rigorous biodiversity experiments have been
designed and conducted to tease apart effects of changing numbers of species (richness) from effects of changing identities of species (composition) (O’Connor et al.
2017; Grossman et al. 2018; Isbell et al. 2018). Complementarity among diverse
plant species that vary in their functional attributes and capture and respond to
resources differently is the primary explanation for increasing productivity with
diversity (Williams et al. 2017). Nevertheless, both the nature of biodiversity-ecosystem function (BEF) relationships and their causal mechanisms remain variable
and scale dependent in natural systems. In the Nutrient Network global grassland
experiments, in which communities have assembled naturally, the relationship
between diversity and productivity is variable (Adler et al. 2011). In tropical forest
plots around the globe, at spatial extents of 0.04 ha or less, the biodiversity-productivity relationship is strong. However, as scales increase to 0.25 or 1.0 ha, the relationship is no longer consistently positive and can frequently be negative (Chisholm
et al. 2013). These varied relationships at contrasting spatial scales may result from
nonlinear, hump-shaped relationships between biodiversity and ecosystem function
across resource availability gradients as the nature of species interactions and their
level of complementarity shift (Jaillard et al. 2014). RS methods—including imaging spectroscopy and LiDAR—that can detect both the diversity and the structure
and function of ecosystems (Martin, Chap. 5; Atkins et al. 2018) can discern these
relationships across spatial extents and biomes in natural systems. They thus have
high potential to enhance our understanding of the scale and context dependence of
linkages between biodiversity and ecosystem function (Grossman et al. 2018).
2.9 Incorporating Spectra into Relationships
Between Biodiversity and Ecosystem Function
Detection of spectral diversity, in particular, offers the potential to contribute to the
quantification of BEF relationships at large scales (Schweiger et al. 2018) and is
thus worth discussing in more detail. The variability captured by spectral diversity
in a given ecosystem depends on the way the spectral diversity is calculated, as
well as its spatial and spectral resolution (Sect. 2.7.5; Gamon et al., Chap. 16).
From a functional perspective, spectral profiles measured at the leaf level depend
on the chemical, structural, morphological, and anatomical characteristics of leaves
(Ustin and Jacquemoud, Chap. 14). Variation in spectra and spectral diversity can
be used to test hypotheses about how specific traits influence ecosystem function,
community composition, and other characteristics of ecosystems, when using
spectral bands or spectral indices with known associations with specific plant traits
(Serbin and Townsend, Chap. 3). Moreover, spectral bands and indices can be
weighted based on prior information about the relative contribution of individual
traits to specific ecosystem characteristics. However, while the absorption features
J. Cavender-Bares et al.