Comprehensive validation of the FY-4B/AGRI sea surface temperature product over the china seas and adjacent marginal seas: accuracy, spatiotemporal biases, and multi-dimensional error characteristics

The accuracy of FY-4B/AGRI sea surface temperature (SST) over the China Seas has not been systematically evaluated. Using iQuam and Himawari-9/AHI data from June 2024 to May 2025, we conducted a comprehensive validation. The product shows excellent performance: monthly RMSE 0.650–0.863 °C, R2 > 0.966 in autumn–spring, and high consistency with Himawari-9 (RMSE 0.795–0.971 °C, R > 0.908). Errors exhibit latitudinal dependence, larger coastal biases, a positive bias in cold waters (<15 °C) and negative in warm waters (>28 °C)—a pattern persisting across FY-4A/4B, suggesting an algorithmic rather than hardware origin—plus diurnal and seasonal variations. Summer accuracy decline (R2 as low as 0.691) is primarily driven by high water vapour (TCWV > 40 kg/m2) rather than cloud cover alone. We recommend direct use in mid-latitude open oceans and in winter–spring–autumn. In low-latitude humid regions and coastal areas, bias correction with quality flags is advised, with particular caution warranted for summer observations.