Effects of spectral resolution and SNR on the vegetation solar-induced fluorescence retrieval using FLD-based methods at canopy level

Abstract

The accuracy of solar-induced chlorophyll fluorescence (SIF) retrieval is dependent on the specifications of the sensor used. Here, we present a quantitative accuracy assessment of the SIF retrieval for different spectral resolution (SR) and signal-to-noise ratio (SNR) at five potential bands (Hα, O2-B, Fe, O2-A, and KI). Firstly, the sensitivities of the SIF retrieval to the correction factors in the FLD algorithm or the radiance noise increased rapidly when the SR decreased, especially for the O2-B band. Then, the dominant factor influencing the SIF retrieval was the SNR for the narrow bands, while was the SR for the broad band. Finally. the accuracy assessment of the SIF retrieval for the potential space-based hyperspectral sensors, including GOSAT TANSO-FTS, SCIAMACHY, OCO-2, Metop GOME-2, and FLEX FLORIS, showed a large error (>35%) for GOSAT, and relative small errors from 5% to 20% for other sensors.