The Spectral Species Concept in Living Color

Abstract

Biodiversity monitoring is an almost inconceivable challenge at the scale of the entire Earth. The current (and soon to be flown) generation of spaceborne and airborne optical sensors (i.e., imaging spectrometers) can collect detailed information at unprecedented spatial, temporal, and spectral resolutions. These new data streams are preceded by a revolution in modeling and analytics that can utilize the richness of these datasets to measure a wide range of plant traits, community composition, and ecosystem functions. At the heart of this framework for monitoring plant biodiversity is the idea of remotely identifying species by making use of the 'spectral species' concept. In theory, the spectral species concept can be defined as a species characterized by a unique spectral signature and thus remotely detectable within pixel units of a spectral image. In reality, depending on spatial resolution, pixels may contain several species which renders species-specific assignment of spectral information more challenging. The aim of this paper is to review the spectral species concept and relate it to underlying ecological principles, while also discussing the complexities, challenges and opportunities to apply this concept given current and future scientific advances in remote sensing.

Keywords: airborne sensors; biodiversity; ecoinformatics; hyperspectral images; plant optical types; remote sensing; satellite imagery; vegetation communities.

Figure 1

Some potential scenarios that can happen in optical remote sensing of vegetation canopies. The graph shows sources of variation in the relationship between species and optical types. Plants of different species might belong to different optical types, but many other situations can also be found. Optical types can be related to information of interest (e.g., species or plant traits) or to irrelevant pattern (e.g., shadows, depending on the research question). Scenario (a) represents a stand with individuals of only one single species, with a similar reflectance. In scenario (b) individuals of two species have a similar reflectance; hence they would be grouped in the same spectral species. This is further complicated once mixing individuals belonging to the same taxon but to different optical types (c) or individuals of multiple species belonging to different optical types that do not follow the species boundaries (d). What many would hope for is that plants of different species belong to different optical types, which may happen (e). Finally, the same plant individual can consist of different optical types showing different spectral properties in for example, young versus old leaves, shadow and light, or differences in health conditions. This intra‐individual mixing property will be related to all of the previous cases (f)–(h). Note that a stand or individual can pass through several of these scenarios in time (intra and interannual variability).

Figure 2

Box 1 Figure ‐ The spectral species algorithm phases. The original image was acquired with the CAO AToMS imaging spectrometer during an airborne campaign over the CICRA experimental site (Amazonian Peru) (https://www.amazonconservation.org/about/mission-vision/cicra-station/). The first image (a) corresponds to the RGB representation of an imaging spectroscopy subset. A standardized PCA is applied on (a) and a reduced set of components is selected (b) to maximize signal corresponding to biological patterns on forested areas and discard noisy components. Spectral species are defined for each pixel by applying an unsupervised k‐means clustering on the spectral space defined by selected components (c). In this phase, a field survey recognition based on in situ data is crucial to define the number of singular spectral signatures (spectral species) expected. The spectral species map is divided into elementary spatial units and the spectral species inventory is performed for each spatial unit, by further calculating Shannon's H and Bray‐Curtis metrics to derive (d) alpha‐ (ranging here from minima to maxima from black to blue, green and red) and (e) beta‐diversity (in which colors represent differences among spectral species) maps, respectively.