A widely reported empirical relationship links satellite vegetation indices (VIs) and gross primary productivity (GPP) estimated by eddy covariance across diverse ecosystems. This observed covariation underpins the use of VIs as inputs to light-use-efficiency and greenness-based carbon-cycle models (for example as proxies for fPAR). However, canopy carbon assimilation results from both slowly evolving structural properties and rapidly changing physiological processes. The extent to which VI–GPP covariance reflects structural scaffolding versus transient physiological function under variable environmental conditions remained unresolved prior to this study.
The authors synthesized half-hourly eddy covariance measurements from 328 globally distributed FLUXNET sites. To reduce angular and illumination effects on remote-sensing signals, they employed an angle-normalized Enhanced Vegetation Index fixed to nadir view and a solar zenith angle of 30 degrees (EVI_SZA30). Analysis was performed across 54,720 high-frequency temporal windows at half-hourly resolution, enabling paired comparisons of satellite-derived index values and flux-tower GPP at sub-daily timescales.
To mechanistically separate structural from functional drivers of observed GPP (GPP_EC), the study applied a nonlinear light-response curve model to each temporal window. This decomposition partitioned actual GPP into two components:
This approach allowed the authors to quantify how well the selected VI (EVI_SZA30) tracked structural capacity versus high-frequency physiological variation.
At the macroscopic scale, the correlation between EVI_SZA30 and observed GPP_EC was substantial (R2 = 0.554). Crucially, decomposition revealed that this macroscopic covariance is primarily driven by the index's capacity to track structural photosynthetic capacity: EVI_SZA30 explained P_c with R2 = 0.538. By contrast, EVI_SZA30 showed limited sensitivity to high-frequency physiological variation captured by intrinsic quantum yield, with R2 = 0.038 for alpha. These results indicate that, across the global FLUXNET sample and the temporal windows analyzed, the satellite index signal corresponds mostly to structural aspects of canopy capacity rather than to rapid physiological down-regulation.
The authors highlight marked differences in water-limited biomes such as open shrublands and woody savannas. In these systems, intense environmental stress triggers rapid stomatal regulation and other physiological responses that can down-regulate photosynthesis on sub-daily timescales, while canopy physical structure remains comparatively stable. Under such conditions a structural–physiological decoupling emerges: the correlation between EVI and photosynthetic capacity (P_c) may be stronger than the correlation between EVI and actual, instantaneous GPP_EC. This decoupling emphasizes limitations of single overpass satellite observations to capture fast physiological dynamics.
Because sun-synchronous polar-orbiting satellites sample at discrete times, they are intrinsically limited in capturing sub-daily physiological down-regulation (for example midday photosynthetic depression driven by stomatal closure). The authors suggest that monitoring paradigms could benefit from continuous, high-frequency observations available from geostationary (GEO) platforms to better resolve transient ecosystem function and to bridge the gap between structural proxies and instantaneous physiological state. Incorporating higher-frequency remote sensing could improve detection of diurnal or stress-driven physiological responses that are poorly represented by widely used VIs tied to structural properties.
All results reported here derive from the presented synthesis and light-response decomposition across the 328 FLUXNET sites and 54,720 half-hourly windows. Specific methodological and supplementary details are provided in the preprint but are not reproduced in full here. The manuscript is a preprint posted on bioRxiv and has not been peer reviewed; the authors declare no competing interests. The suggestion to leverage geostationary observations is an interpretation informed by the observed temporal limitations of polar-orbiting overpasses described in the analysis.