How to overcome different limitations in estimating plant diversity via spectral diversity?

Technological advances in optical remote sensing, which measures the electromagnetic radiation reflected by an object at various wavelengths, allow for efficient and relatively inexpensive collection of baseline data related to biodiversity. In particular, spectral diversity—the variability in remotely sensed spectral reflectance data obtained from plant communities—has emerged as a valuable proxy for different facets of biodiversity, such as plant taxonomic, phylogenetic and functional diversity. However, successful estimation of plant diversity using spectral diversity is negatively impacted by several factors, including: i) limited or coarse spatial resolution of remote sensing data, ii) changes in remotely sensed reflectance data over time, and iii) weak linkages between species counts and spectral diversity in agricultural landscapes. To overcome these limitations, we present three novel spectral diversity approaches: i) a subpixel spectral diversity approach, ii) a multi-temporal spectral diversity approach, and iii) an object-based spectral diversity approach. Here, we provide different case studies using these three spectral diversity approaches to quantify plant diversity in two distinct grassland ecosystems: an agricultural landscape in the Swiss Alps and a tallgrass prairie in Oklahoma, U.S.