GEOSIF: A continental-scale sub-daily reconstructed solar-induced fluorescence derived from OCO-3 and GK-2A over Eastern Asia and Oceania

• Published in: Remote sensing of environment Vol. 311; p. 114284
• Main Authors: Jeong, Sungchan, Ryu, Youngryel, Li, Xing, Dechant, Benjamin, Liu, Jiangong, Kong, Juwon, Choi, Wonseok, Fang, Jianing, Lian, Xu, Gentine, Pierre
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.09.2024
• Subjects: Afternoon depression; GK-2A; NIRv; OCO-3; SIF; VPD; SIF; OCO-3; GK-2A; Afternoon depression; NIRv; VPD
• ISSN: 0034-4257
• DOI: 10.1016/j.rse.2024.114284

The diurnal solar-induced chlorophyll fluorescence (SIF) sampling capability of OCO-3 can provide crucial insights into ecosystem function at the sub-daily scale. However, potential applications of OCO-3 SIF have suffered from its inherent spatiotemporal discontinuity. In this study, we addressed the discontinuous observation coverage of OCO-3 SIF by utilizing information coming from the continuous geostationary satellite observations from Geostationary Korea Multi-Purpose Satellite-2A (GK-2A). We generated and comprehensively evaluated a continental-scale hourly reconstructed SIF over the Eastern Asia and Oceania. To do this, we trained an extreme gradient boosting (XGBoost) model using OCO-3 SIF and GK-2A observations including four band Nadir BRDF Adjusted Reflectance (NBAR) (blue, green, red, and near-infrared), shortwave radiation, and vapor pressure deficit (VPD) using the data from August 2019 to July 2021. The reconstructed SIF data showed robust agreement with OCO-3 SIF across diverse ecosystems, different hours of the day, and varying observation geometries (R2 = 0.68–79). We found large feature importance of near-infrared reflectance, red reflectance, and shortwave radiation, which together explained 84.6% of SIF prediction. VPD played an increasing role under high temperature conditions. The reconstructed SIF effectively captured the afternoon depression of photosynthesis across diverse ecosystems, ranging from 63.9% to 88.9%, which was consistent with the original OCO-3 SIF. Our results identified a more pronounced afternoon depression in the physiological SIF yield than in the canopy structural proxy. In addition, diurnal changes in both canopy structural and physiological components of SIF showed a stronger relationship with VPD than that of temperature. These findings highlight the benefits of the synergistic use of new-generation satellite observations to improve our understanding of large-scale diurnal ecosystem dynamics and its environmental drivers. [Display omitted] •Generated sub-daily scale reconstructed SIF by combining OCO-3 and GK-2A observations.•Demonstrated the role of spectral and environmental forcing in SIF prediction using explainable AI approach.•Revealed afternoon depression in canopy structure and physiological SIF components.