Improved consistency in solar-induced fluorescence retrievals from GOME-2A with the SIFTER v3 algorithm

Abstract. Space-based observations of solar-induced fluorescence (SIF) provide valuable insights into vegetation activity over time. The GOME-2A instrument, in particular, facilitates SIF retrievals with extensive global coverage and a record extending over 10 years. SIF retrievals, however, are sensitive to calibration issues, and instrument degradation complicates the construction of temporally consistent SIF records. This study introduces the improved Sun-Induced Fluorescence of Terrestrial Ecosystems Retrieval (SIFTER) v3 algorithm, designed to obtain a more accurate and reliable long-term SIF record from GOME-2A for the 2007–2017 period, building upon the previous SIFTER v2. The SIFTER v3 algorithm uses newly reprocessed level-1b Release 3 (R3) data, which provides a more homogenous record of the reflectances by eliminating spurious trends from changes in level 0 to level 1 processing. This improved consistency supports detailed analysis and correction of the reflectance degradation across the SIF retrieval window (734–758 nm). To address the reflectance degradation accurately, SIFTER v3 incorporates an advanced in-flight degradation correction that accounts for time, wavelength, and scan-angle dependencies throughout the entire record. Additionally, algorithm revisions have consistently reduced the retrieval residuals by around 10 % and reduced sensitivity to water vapor absorption by better capturing the atmospheric and instrumental effects. A revised latitude bias adjustment resolves unrealistic values of GOME-2A SIF over desert areas. The SIFTER v3 dataset demonstrates improved robustness and consistency, both spatially and temporally, throughout the 2007–2017 record, and aligns closely with independent GPP measurements from the global FluxSat and FLUXCOM-X products.