Foliar functional and genetic variation in a keystone Hawaiian tree species estimated through spectroscopy

Abstract

Imaging spectroscopy has the potential to map closely related plant taxa at landscape scales. Although spectral investigations at the leaf and canopy levels have revealed relationships between phylogeny and reflectance, understanding how spectra differ across, and are inherited from, genotypes of a single species has received less attention. We used a common-garden population of four varieties of the keystone canopy tree, Metrosideros polymorpha, from Hawaii Island and four F1-hybrid genotypes derived from controlled crosses to determine if reflectance spectra discriminate sympatric, conspecific varieties of this species and their hybrids. With a single exception, pairwise comparisons of leaf reflectance patterns successfully distinguished varieties of M. polymorpha on Hawaii Island as well as populations of the same variety from different islands. Further, spectral variability within a single variety from Hawaii Island and the older island of Oahu was greater than that observed among the four varieties on Hawaii Island. F1 hybrids most frequently displayed leaf spectral patterns intermediate to those of their parent taxa. Spectral reflectance patterns distinguished each of two of the hybrid genotypes from one of their parent varieties, indicating that classifying hybrids may be possible, particularly if sample sizes are increased. This work quantifies a baseline in spectral variability for an endemic Hawaiian tree species and advances the use of imaging spectroscopy in biodiversity studies at the genetic level.

Keywords: Genetic diversity; Hawaii; Leaf spectra; Metrosideros polymorpha; Plant evolution; Spectroscopy.

Fig. 1

a Mean brightness-normalized reflectance (represented as a percentage) and b coefficient of variation (CV) of reflectance values for the four M. polymorpha Hawaii Island varieties and Oahu incana. See Fig. SI 2 for reflectance prior to brightness-normalization. c Spectral separability of all Hawaii Island and Oahu genotypes. Spectral separability was calculated for each wavelength (Eq. 1). Higher values indicate less spectral overlap. See Table 3 for average SSI values

Fig. 2

Boxplots of leaf traits for the Hawaii Island varieties glaberrima (G), polymorpha (P), newellii (N), and incana (I) as well as Oahu incana (OI). Hawaii Island varieties with traits that differed at a significance of p < 0.05 as determined by ANOVA and Tukey HSD are noted with an asterisk. Incana from Hawaii Island and Oahu were also compared using ANOVA, and significant differences are likewise noted with an asterisk. Boxplots denote quartile ranges, with the lower and upper bounds of the box indicating the 25th and 75th percentile. Middle lines in the box represent the median of the data, and the whiskers end at the group minimum and maximum. Outliers are shown as points

Fig. 3

Mean brightness-normalized reflectance (represented as a percentage) of a the F1 hybrids glaberrima-incana (GI) and its parents, glaberrima (G) and incana (I); b the hybrid incana-polymorpha (IP) and its parents, incana (I) and polymorpha (P); the hybrid newellii-polymorpha (NP) and its parents, newellii (N) and polymorpha (P); d the hybrid glaberrima-polymorpha (GP) and its parents, glaberrima (G) and polymorpha (P). See supplementary information Figure SI 4 for CV of F1 hybrids

Fig. 4

Boxplots of nine leaf traits for all F1 hybrids measured and their parents. The left figure a represents glaberrima (G), the F1 hybrid glaberrima-incana (GI), incana (I), the hybrid incana-polymorpha (IP), and polymorpha (P). The right figure b displays newellii (N), the hybrid newellii-polymorpha (NP), polymorpha (P), the hybrid glaberrima-polymorpha (GP), and glaberrima (G). Genotypes with traits that differed at a significance of p < 0.05 as determined by ANOVA and Tukey HSD are noted with an asterisk. Only groups of the F1 hybrid and their parents (Table 1) were compared using ANOVA and Tukey HSD. None of the traits differed significantly among GI, I, and G according to ANOVA