Abstract
<jats:p>Geostationary satellites offer unprecedented high-frequency terrestrial monitoring, yet rigorous atmospheric correction pipelines for the newly deployed FY-4B/AGRI remain critically limited. We developed a robust 6SV-based operational land surface reflectance (LSR) retrieval pipeline for FY-4B/AGRI, dynamically driven by CAMSRA-EAC4 atmospheric forcing. Comprehensive validation across 25 AERONET/SKYNET sites throughout 2023 confirmed highly reliable reanalysis forcing fields (water vapor/ozone R = 0.964/0.928; AOD R = 0.712). For rigorous MODIS cross-validation, we applied land-cover-specific spectral band adjustment factors (SBAF) and forward-simulated MODIS reflectance explicitly to match FY-4B/AGRI’s exact instantaneous sun-view geometries. The red and near-infrared (NIR) bands demonstrated strong quantitative consistency, achieving global correlations of R = 0.743 and 0.584, mean biases of 0.032 and 0.012, and RMSEs of 0.058 and 0.077, respectively. The broad blue band exhibited expectedly lower correlations driven by intense Rayleigh scattering. Retrieval accuracy proved highly biome-dependent, with exceptional fidelity over homogeneous barren/shrubland targets (RMSE 0.025–0.05) but larger radiometric uncertainties in complex, heavily polluted urban landscapes. Substituting reanalysis data with in-situ AOD yielded strong consistency (R > 0.9, RMSE < 0.007) over stable targets, validating the LUT inversion architecture's outstanding resilience. Radiometric agreement remained stable at moderate view zenith angles (< 50.0°) but degraded at high-VZA edge sites when solar zenith angles exceeded 60.0°. Crucially, its 15-minute sampling cadence systematically exploited transient clear-sky windows and successfully reconstructed stabilized NDVI and EVI seasonal trajectories, effectively mitigating severe cloud-cover data loss inherent to conventional polar-orbiting sensors. This framework establishes a physical baseline for high-frequency terrestrial monitoring utilizing China’s FY-4B/AGRI geostationary satellite.</jats:p>