Disentangling Changes in the Spectral Shape of Chlorophyll Fluorescence: Implications for Remote Sensing of Photosynthesis

Abstract

Novel satellite measurements of solar‐induced chlorophyll fluorescence (SIF) can improve our understanding of global photosynthesis; however, little is known about how to interpret the controls on its spectral variability. To address this, we disentangle simultaneous drivers of fluorescence spectra by coupling active and passive fluorescence measurements with photosynthesis. We show empirical and mechanistic evidence for where, why, and to what extent leaf fluorescence spectra change. Three distinct components explain more than 95% of the variance in leaf fluorescence spectra under both steady‐state and changing illumination conditions. A single spectral shape of fluorescence explains 84% of the variance across a wide range of species. The magnitude of this shape responds to absorbed light and photosynthetic up/down regulation; meanwhile, chlorophyll concentration and nonphotochemical quenching control 9% and 3% of the remaining spectral variance, respectively. The spectral shape of fluorescence is remarkably stable where most current satellite retrievals occur (“far‐red,” >740nm), and dynamic downregulation of photosynthesis reduces fluorescence magnitude similarly across the 670‐ to 850‐nm range. We conduct an exploratory analysis of hourly red and far‐red canopy SIF in soybean, which shows a subtle change in red:far‐red fluorescence coincident with photosynthetic downregulation but is overshadowed by longer‐term changes in canopy chlorophyll and structure. Based on our leaf and canopy analysis, caution should be taken when attributing large changes in the spectral shape of remotely sensed SIF to plant stress, particularly if data acquisition is temporally sparse. Ultimately, changes in SIF magnitude at wavelengths greater than 740 nm alone may prove sufficient for tracking photosynthetic dynamics.

Cropland Carbon Uptake Delayed and Reduced by 2019 Midwest Floods

“…SIF has been shown to be a robust proxy for GPP (Frankenberg & Berry, ; van der Tol et al, ). While nonlinear relationships between GPP and SIF have been observed at leaf and flux‐tower scales under certain conditions, for example, strong incoming light (Magney et al, ; Verma et al, ), linear relationships have been generally observed at ecosystem and regional scales (Frankenberg et al, ; Li et al, ; Magney et al, ; Sun et al, ). The robust linear relationship at increasing scales is likely because satellite measurements are primarily measuring an integrated canopy average of low Photosynthetically Active Radiation (PAR).…”

Towards a Harmonized Long‐Term Spaceborne Record of Far‐Red Solar‐Induced Fluorescence

“…We also examine sensitivity to SIF reference shapes by substituting simulated estimates obtained from the SCOPE model (Van Der Tol et al, ; GOME2A‐PKv1, GOME2A‐PKexp1, and GOME2A‐PKtrop) with observed shapes obtained from leaf fluorescence spectra (Magney, Frankenberg, et al, ). Leaf‐level data account for reabsorption by chlorophyll across red and into the far‐red wavelengths, and thus reduced magnitude up to 740 nm compared to photosystem‐level emissions assumed by SCOPE (e.g., see Figure ).…”

“…Although the intersensor wavelength range is relatively small (~30 nm), absolute fluorescence values vary by more than a factor of 2 in this region (Joiner et al, ; Köhler, Guanter, et al, ; Sun et al, ). Reference SIF shapes derived from leaf‐level studies suggest that far‐red fluorescence spectra, and thus wavelength conversions, are roughly consistent across species (Magney, Frankenberg, et al, ). In principle, a look‐up table approach could be used to determine a multiplicative factor to convert between different wavelengths evaluated across sensors/retrievals based on a reference spectrum derived from an average across multiple plant species.…”

“…Additional complications arise in that each sensor provides only a single daily overpass (wide‐swath sensors at higher latitudes are an exception). This prevents accurate estimation of daily‐averaged SIF because the SIF yield varies with changing light conditions (e.g., Magney, Frankenberg, et al, ; Yang et al, ). Overpass times differ between sensors (morning for GOME, GOME‐2, and SCIAMACHY; midday for GOSAT, OCO‐2, TROPOMI), which further confounds sensor intercomparison and cross calibration (e.g., Figure a), and adds uncertainty when analyzing vegetation photosynthetic responses across optical vegetation indices due to differences in diurnal stress patterns (Zhang, Xiao, et al, ).…”

“…Another important question regarding SIF retrievals is whether the reference SIF emission shape is correlated with other spectral functions used in the retrieval in a way that affects the retrieval accuracy and precision leading to location‐dependent biases. Leaf‐level spectral measurements show relatively weak variability across species and conditions in the SIF emission spectrum near the far‐red peak (~740 nm), and high variability across the red peak (~680 nm; Magney, Frankenberg, et al, ). The shorter far‐red spectral window used for TROPOMI is stable with respect to changes in the shape of the fluorescence emission spectrum, as suggested by leaf‐level spectral measurements.…”