loading page

Sun-Induced Fluorescence's Correlation to Carbon-Flux Increases When Raw Data is Adjusted to Account for Vegetation Biochemistry and Structure.
  • Edgard Bontempo,
  • Dalton Valeriano
Edgard Bontempo
National Space Research Institute (INPE), National Space Research Institute (INPE)

Corresponding Author:edgardbontempo@gmail.com

Author Profile
Dalton Valeriano
National Space Research Institute (INPE), National Space Research Institute (INPE)
Author Profile

Abstract

The quantification and monitoring of photosynthesis are essential to understand the global carbon cycle and vegetation’s responses to climate. Among the different remotely-sensed photosynthesis-related variables, Sun-Induced chlorophyll a Fluorescence (SIF) is especially promising since it results directly from photochemical energy conversion but uncertainties still complicate its interpretation. Recent studies have pointed to the influences of vegetation biochemistry and structure on radiative transfer as the main confounding factors for the use of SIF as a photosynthesis proxy. Leaf-level fluorescence research has shown that such influences may be removed by adjusting the raw fluorescence signal to the emitting leaf’s spectra and we suggest that this can be upscaled to the landscape level. In this study we present and test new Spectrally-Adjusted SIF formulations (SASIFs), along with previously proposed SIF modifications and other acknowledged photosynthesis productivity proxies, against carbon-flux data from vegetation of diverse structure. Accordingly, we used Gross Primary Productivity (GPP) data spanning periods from two to seven years, from 27 FLUXNET sites classified into different Land Cover Classes (LCCs) as defined by the International Geosphere-Biosphere Programme (IGBP). The data tested against GPP was calculated with GOME-2 SIF data, MODIS reflectance and spectral vegetation indices, and it included: NIRV, SIF from the red and the far-red frequency peaks, SIF normalized by the cosine of the Sun’s zenith angle, SIF-yield, new SASIFs and FLUXCOM GPP. The relationships between all variables and FLUXNET GPP were tested using time-series decomposition, site- and LCC-specific Kendall’s rank correlation tests and linear mixed model analysis. Results show that one of our new SASIFs has the best overall correlation to FLUXNET GPP among all tested data. Our LCC-specific analysis demonstrates the influences of biochemistry, phenology, temporal resolution and vegetation structure on the relationships between the tested variables. Results support the idea that chlorophyll fluorescence can be complemented with reflectance data improving our ability to monitor vegetation productivity and predict climate-driven changes to standing biomass in spite of their particular limitations.