Effect of environmental conditions on sun-induced fluorescence in a mixed forest and a cropland

Due to the large carbon dioxide (CO2) fluxes between terrestrial ecosystems and the atmosphere, dynamics of photosynthesis can have significant effects on atmospheric CO2 concentrations and lead to large uncertainties in ecosystem C budgets. Remote sensing approaches using sun-induced chlorophyll fluorescence (SIF) hold the potential to directly assess ecosystem photosynthesis. However, many challenges remain linked to using the SIF emission signal to estimate gross primary production (GPP). The goal of this study was to gain a better understanding of the relationships between GPP and SIF over different time scales (minutes to years) and under varying environmental conditions. Two different ecosystems were investigated, a cropland and a mixed forest, with continuous eddy covariance flux measurements. Continuous tower-based SIF retrievals were performed in 2015 and 2016 at both ecosystems. In both ecosystems, SIF was found to be more affected by environmental conditions than GPP. Annual cycles for GPP and SIF differed at the mixed forest due in part to the influence of the different footprint size of the two independent measurements. Diurnal cycles in GPP and SIF corresponded well under unstressed conditions and followed the incoming photosynthetic photon flux density (PPFD). However, depressions in SIF were found at both sites either at midday or in the afternoon during the growing season. At the cropland site, reductions in SIF occurred at high PPFD (PPFD > 1470 μmol m−2 s−1, R2 = 0.62) and high VPD (VPD > 1590 Pa, R2 = 0.35). Whereas at the forest site, reductions in SIF were linked to high VPD (VPD > 1250 Pa, R2 = 0.25), but not to high PPFD (R2 = 0.84). The depression in SIF was also associated with an increase in non-photochemical quenching, as indicated by the photochemical reflectance index (R2 = 0.78), thus showing the complementarity between SIF and non-photochemical quenching as different energy pathways. Our results show the importance of characterizing the influence of different environmental conditions on SIF-GPP relationships for specific ecosystems to reliably estimate GPP from remote sensing measurements