Assessing the variability and uncertainty of two-flux FAPAR measurements in a conifer-dominated forest

The estimation of the Fraction of Absorbed Active Radiation (FAPAR) and its uncertainties is key for understanding global carbon balances. This work investigates the variability and associated uncertainties of in-situ two-flux FAPAR observations attributed to changes in phenological and meteorological conditions. Specifically, we assessed influences of illumination conditions with solar zenith angle and the ratio of diffuse-to-total incident radiation, wind speed, leaf color and snow coverage on the variability of two-flux FAPAR. We assumed FAPAR acquired under diffuse light conditions to be closest to true FAPAR as it is not influenced by the solar zenith angle. To reveal the uncertainty of the two-flux FAPAR measurements, we investigated the difference (Delta(two-flux)) between FAPAR acquired under diffuse light conditions and two-flux FAPAR acquired during a certain environmental condition (e.g. large solar zenith angle). A positive (negative) value obtained from this difference was interpreted as an indication for an underestimation (overestimation) of true FAPAR by the two-flux FAPAR estimate, as found in previously investigations with canopy radiative transfer models (RTM). Therefore, permanent PAR measurements were carried out 2015-2017 in a sub-alpine, spruce-dominated forest in Southern Germany using a Wireless Sensor Network (WSN). FAPAR observations exhibited considerable seasonal variability (0.89 to 0.99 +/- 0.03) despite the dominance of evergreen spruces. The in-situ observations confirm significant overestimation of FAPAR by up to 0.06 under solar zenith angles above 60 degrees and by up to 0.05 during the presence of colored autumn leaves, similarly to the results obtained from previous studies with canopy RTMs. Additionally, our results indicate an effect of wind speed which we consider crucial at sites where high wind speeds occur more frequently. Overall, this study shows the potential of permanent WSN monitoring activities to ensure multi-year FAPAR observations with associated uncertainty information that are demanded to validate satellite-derived FAPAR products in forest ecosystems.