Hyperspectral leaf reflectance of grasses varies with evolutionary lineage more than with site

Abstract To predict ecological responses at broad environmental scales, grass species are commonly grouped into two broad functional types based on photosynthetic pathway. However, closely related species may have distinctive anatomical and physiological attributes that influence ecological responses, beyond those related to photosynthetic pathway alone. Hyperspectral leaf reflectance can provide an integrated measure of covarying leaf traits that may result from phylogenetic trait conservatism and/or environmental conditions. Understanding whether spectra‐trait relationships are lineage specific or reflect environmental variation across sites is necessary for using hyperspectral reflectance to predict plant responses to environmental changes across spatial scales. We measured hyperspectral leaf reflectance (400–2400 nm) and 12 structural, biochemical, and physiological leaf traits from five grass‐dominated sites spanning the Great Plains of North America. We assessed if variation in leaf reflectance spectra among grass species is explained more by evolutionary lineage (as captured by tribes or subfamilies), photosynthetic pathway (C 3 or C 4 ), or site differences. We then determined whether leaf spectra can be used to predict leaf traits within and across lineages. Our results using redundancy analysis ordination (RDA) show that grass tribe identity explained more variation in leaf spectra (adjusted R 2 = 0.12) than photosynthetic pathway, which explained little variation in leaf spectra (adjusted R 2 = 0.00). Furthermore, leaf reflectance from the same tribe across multiple sites was more similar than leaf reflectance from the same site across tribes (adjusted R 2 = 0.12 and 0.08, respectively). Across all sites and species, trait predictions based on spectra ranged considerably in predictive accuracies ( R 2 = 0.65 to <0.01), but R 2 was >0.80 for certain lineages and sites. The relationship between Vc max , a measure of photosynthetic capacity, and spectra was particularly promising. Chloridoideae, a lineage more common at drier sites, appears to have distinct spectra‐trait relationships compared with other lineages. Overall, our results show that evolutionary relatedness explains more variation in grass leaf spectra than photosynthetic pathway or site, but consideration of lineage‐ and site‐specific trait relationships is needed to interpret spectral variation across large environmental gradients.