Far-red sun-induced chlorophyll fluorescence shows ecosystem-specific relationships to gross primary production: An assessment based on observational and modeling approaches

Sun-induced chlorophyll fluorescence (SIF) is a radiation flux emitted from chlorophyll molecules and is considered an indicator of the actual functional stata of plant photosynthesis. The remote measurement of SIF opens a new perspective to assess actual photosynthesis at larger, ecologically relevant scales and provides an alternative approach to study the terrestrial carbon cycle. Recent studies demonstrated the reliability of measured SIF signals and showed significant relationships between SIF and gross primary production (GPP) at ecosystem and global scales. Despite these encouraging results, understanding the complex mechanisms between SIF and GPP remains challenging before SIF can be finally utilized to constrain estimates of GPP. In this study, we present a comprehensive assessment of the relationship between far-red SIF retrieved at 760 nm (SIF760) and GPP, and its transferability across three structurally and physiologically contrasting ecosystems: perennial grassland, cropland and mixed temperate forest. We use multi-temporal imaging spectroscopy (IS) data acquired with the Airborne Prism EXperiment (APEX) sensor as well as eddy covariance (EC) flux tower data to evaluate the relationship between SIF760 and GPP(Ec). We use simulations performed with the coupled photosynthesis-fluorescence model SCOPE to prove trends obtained from our observational data and to assess apparent confounding factors such as physiological and structural interferences or temporal scaling effects. Observed relationships between SIF760 and GPPEc were asymptotic and ecosystem-specific, i.e., perennial grassland (R-2 = 0.59, rRMSE = 27.1%), cropland (R-2 = 0.88, rRMSE = 3.5%) and mixed temperate forest (R-2 = 0.48, rRMSE = 15.88%). We demonstrate that asymptotic leaf level relationships between SIF760 and GPPEc became more linear at canopy level and scaled with temporal aggregation. We conclude that remote sensing of SIF provides a new observational approach to decrease uncertainties in estimating GPP across ecosystems but requires dedicated strategies to compensate for the various confounding factors impacting SIF-GPP relationships. Our findings help in bridging the gap between mechanistic understanding at leaf level and ecosystem-specific observations of the relationships between SIF and GPP. (C) 2015 Elsevier Inc. All rights reserved.