Impact of radiation variations on temporal upscaling of instantaneous Solar-Induced Chlorophyll Fluorescence

Abstract

Solar-Induced Chlorophyll Fluorescence (SIF) has been increasingly used as a novel proxy for vegetation productivity. Several space-borne instruments can retrieve SIF at varying overpass time, which complicates the interpretation as SIF is driven by absorbed Photosynthetically Active Radiation (PAR) at the acquisition time. To facilitate comparisons across sensors, satellite-based SIF is upscaled to daily averages with a length-of-day correction factor (DC). In conventional DC calculations, the light intensity over a day is approximated geometrically by the cosine of the Solar Zenith Angle (SZA), neglecting changes in atmospheric extinction and topographic effects. Here, we use reanalysis PAR data for DC calculations to evaluate the impact of atmospheric extinction and diffuse radiation individually. We find that the simple SZA approach is a reliable approximation for flat surfaces, where the overall atmospheric impact on DC is less than 10% as large individual effects on direct and diffuse PAR partially compensate each other. At longer time-scales, a sampling (clear sky) bias might exist due to cloud-filtering of satellite data. We find that in the Amazon the true monthly mean PAR can be 25% lower than the one for cloud-filtered days, potentially inducing seasonal SIF biases on the same order. An additional factor impacting PAR during a day is topography. For complex terrain, direct light in the DC expression requires a correction for surface slopes. For example in the San Gabriel Mountains, California, USA, the modified DC is changed by as much as 500% for strongly tilted surfaces. This modification is especially important for satellite instruments with fine spatial resolutions, where surface slopes are not averaged out and can have a substantial impact on reflectance and SIF. Overall, our refined DC-corrections and averaging strategy can help satellite SIF interpretation as well as intercomparisons over a wide range of spatio-temporal scales and overpass times.