One of the grand challenges in biodiversity and ecosystem science is characterizing the relevant variation in ecosystem properties that are associated with different levels of biodiversity (i.e., species, genetic, plants, arthropods, microbes, etc.)? In a practical sense, the key research needs are for the rapid and repeatable collection of functional “trait” data that can characterize environmental drivers of ecosystem processes, such as foliar biochemistry, metabolic rates, and vegetation response to stress, pests and pathogens and environmental change. Single observation data are not sufficient: we need to know how functional traits of plants and ecosystems vary comprehensively across space, time and environmental gradients. Imaging spectroscopy has recently emerged as potential method to characterize this variability. Here I outline the concepts behind the use of spectroscopy to measure multiple aspects of ecosystem function across a broad range of applications from individual leaves to the canopy. I address the fundamental science and key challenges behind the use of spectroscopy in ecosystem studies, specifically translating leaf optics from reflectance to traits encompassing the range of scales important to biodiversity science.
Results/Conclusions
Spectroscopic models can be developed to predict multiple traits at both the leaf and canopy levels. Some retrievals are possible due to direct estimation using well-characterized reflectance/absorptance properties, while other traits are predictable as emergent properties of their correlation with related features. There are also numerous traits that are predictable from spectra, but for which the chemical and spectral data are still needed to identify the physical basis for the retrieval. A key question remaining is what is the reasonable domain of inference from spectra to ecosystem processes? For example, research in aspen has shown that canopy spectra better predict belowground dynamics than canopy chemistry or genotype identity. However, belowground processes are not “visible” from canopy spectra, and thus emerge as a function of canopy chemistry, which is visible. Generally, spectroscopy works because it can capture a wider range of variation over more sites than possible from intensive chemical sampling. I present a framework for linking traits, spectra, genetics, environment and plant developmental stage (phenology) to understand the drivers of spectral variation among plants. This framework will result in increased understanding of spatial patterns of functional diversity on Earth, opening opportunities for a broad range of basic and applied science.