Evaluating the utility of hyperspectral data to monitor local-scale β-Diversity across space and time

Everest, J.J., E. Van Cleemput, A.L. Beamish, M.J. Spasojevic, H.C. Humphries, and S.C. Elmendorf. 2025. Evaluating the utility of hyperspectral data to monitor local-scale β-Diversity across space and time. Remote Sensing of Environment 316: 114507. Available at https://doi.org/10.1016/j.rse.2024.114507.

Niwot’s long-term monitoring plots arrayed across the Saddle. Image credit: Joseph Everest & NEON

Abstract

Plant functional traits are key drivers of ecosystem processes. However, plot-based monitoring of functional composition across both large spatial and temporal extents is a time-consuming and expensive undertaking. Airborne and satellite remote sensing platforms collect data across large spatial expanses, often repeatedly over time, raising the tantalising prospect of detection of biodiversity change over space and time through remotely sensed methods. Here, we test the degree to which in situ measurements of taxonomic and functional β-diversity, defined as pairwise dissimilarity either between sites, or between years within individual sites, is detectable in airborne hyperspectral imagery across both space and time in an alpine vascular plant community in the Front Range, Colorado, USA. Functional and taxonomic dissimilarity were significantly related to spectral dissimilarity across space, but lacked robust relationships with spectral dissimilarity over time. Biomass showed stronger relationships with spectral dissimilarity than either taxonomic or functional dissimilarity over space, but exhibited no significant associations with spectral dissimilarity over time. Comparative analyses using NDVI revealed that NDVI alone explains much of the variation explained by the full-range spectra. Our results support the use of hyperspectral data to detect fine-scale changes in vascular plant β-diversity over space, but suggest that methodological limitations still preclude the use of this technology for long-term monitoring and change detection.