Development of GCOM-C/SGLI ground-area-based chlorophyll content estimation: a computationally efficient algorithm free from LAI and vegetation classification

The Chlorophyll (Chl) content estimated from satellite observations often employs inversion of canopy radiative transfer models using other satellite estimates [leaf area index (LAI) and vegetation classification (VC)] and the leaf-carotenoid-content assumption, because its direct retrieval from the light absorption is difficult due to the rapid decrease in the Chl-absorption band reflectance reaching its lower limit at low Chl concentration as Chl increases in forests and grasslands. Such estimation largely depends on the accuracy of the input data, such as LAI, VC, and carotenoids, and thus cannot account for unexpected changes in VC and carotenoids. This study aims to develop a Chl-content estimation method directly from Global Change Observation Mission-Climate (GCOM-C) satellite observations, independent of other satellite estimates and assumptions. This study demonstrated that the SGLI green band (Band 6, 565 nm center wavelength) enables robust estimation due to its small but broad Chl sensitivity and lack of carotenoid sensitivity. In contrast, another green band (Band 5, 530 nm), which has the sensitivity to carotenoids, was not suitable for this purpose because carotenoid content varies across species and seasons and thus requires its assumption or estimation. In short, the green wavelength and its width were critically important. Using a green band (Band 6) and an NIR band (Band 11), one general estimation model of ground-area-based total Chl content (R²=0.87 and RMSE=0.55 g m⁻²) and three specific models for broad leaves, needle leaves, and grasses were created (R²=0.93, 0.91, and 0.94; RMSE=0.39 g m⁻², 0.44 g m⁻², and 0.36 g m⁻², respectively). The ground-area-based total Chl content estimation was independent of PFTs and LAI, whereas the leaf-area-based estimation required them. The general model, which only requires SGLI’s two-band reflectance, offers computational efficiency and near-real-time detection even in areas of unexpected change driven by natural and anthropogenic disturbances, because it is independent of the carotenoid assumption and other satellite-estimated results, such as VC (e.g., land cover classification) and LAI. Besides, it has less dependence on soil moisture, which affects vegetation background reflectance, and water area fraction in a vegetated pixel. Accordingly, SGLI ground-area-based Chl estimates are independent of those factors. Such independent ground-based estimates of Chl content can provide new insights into vegetation studies.