Characterization of chlorophyll fluorescence, absorbed photosynthetically active radiation, and reflectance-based vegetation index spectroradiometer measurements

Abstract

ABSTRACT Spectroscopy plays a key role in Earth observations, especially for studies involving vegetation function and structure. These measurements are critical in the context of carbon cycle monitoring from leaf to global scales. Reflectance-based vegetation indices (RIs) have been used extensively in remote sensing studies from the unpiloted aerial vehicle, aerial, and space-based platforms to model quantities related to productivity, such as gross primary production (GPP), while more recently chlorophyll fluorescence (CF) measurements are increasingly exploited to track GPP. CF and RI measurements vary in magnitude, depend on different portions of the spectrum, and are derived from unique equations; thus, instrument uncertainty manifests distinctly for these measurements. Although this is well known, it is often unexamined in experiments and analyses. We use a portable spectroradiometer to make measurements of reflectance-based vegetation indices (RIs) and chlorophyll fluorescence (CF) in order to characterize how measurements of RIs and CF compare to one another. In particular, we examine fluorescence (F) and fluorescence yield (F Yield) under a light-emitting diode grow light (LED), solar-induced fluorescence (SIF), solar-induced fluorescence yield (SIFYield), absorbed photosynthetically active radiation (APAR), and reflectance-based vegetation indices (the normalized difference vegetation index (NDVI), the chlorophyll/carotenoid index (CCI), and the photochemical reflectance index (PRI)) and include maximized propagated uncertainty of the spectroradiometer for each measurement. We show that RIs have a significantly lower propagated error relative to the mean (0.01% to 0.28%) than CF measurements (0.01% to 1.28%) and that while fine resolution spectrometer CF measurements are outside the noise of the instrument and have potential to provide relative measurements of productivity, show why this instrument having fine spectral resolution and sampling is more effective for measurements of APAR and RIs. We also demonstrate that F and F Yield measurements have low propagated uncertainty and propose that future studies of plant function using this spectrometer/LED technique and the full range of spectra be undertaken. Finally, measurements of SIF, F, and APAR can provide estimates of SIFYieldand F Yield in the same order of magnitude, but further examination is required to determine how these measurements compare under a range of illumination and environmental conditions and how they might compare to PRI.