Full field of view temporal calibration for the directional polarimetric camera onboard the GaoFen-5 (02) satellite...

The directional polarimetric camera (DPC) onboard GaoFen-5(02) satellite provides critical observational data for global aerosol and cloud monitoring. Due to wide field of view (FOV) design and lack of onboard calibration systems, it is necessary to monitor the spectral radiometric degradation over full FOV using vicarious calibration techniques. In this study, two vicarious calibration methods based on natural targets are employed to acquire the radiometric calibration coefficients over entire FOV. The Rayleigh scattering calibration method is used for the shorter visible bands (443, 490, and 565 nm) and majority FOV results are obtained after filtering out glint-contaminated regions, while the desert field calibration method is applied to the longer visible and near-infrared bands (670 and 865 nm) and near full FOV calibration results are successfully achieved. However, uncertainties in input auxiliary parameters, including atmospheric properties and surface characteristics over oceans and deserts, inevitably introduce random errors into radiometric coefficients. Furthermore, seasonal fluctuations in solar elevation and surface reflectance tend to induce spurious periodic fluctuations in instrumental radiometric response. Here, a spatiotemporal fusion model is adopted to achieve high-precision radiometric coefficients across FOV of DPC, through analyzing temporal and spatial variations of radiometric calibration coefficients at different FOV positions, correcting seasonal fluctuations and reconstructing monthly radiometric calibration coefficients. Then, the radiometric calibration coefficients before and after reconstruction are compared with cross-calibration results of DPC and a particulate observing scanning polarimeter (POSP), another polarimetric payload on the same platform equipped with onboard calibration systems. Results show that calibration uncertainties are notably reduced after reconstruction, with RMSE decreasing from 0.0129 to 0.0107 at 443 nm band. Moreover, the spatial coverage of reconstructed results is also significantly improved within FOV. Finally, radiometric coefficients within FOV for DPC/GaoFen-5(02) are reconstructed based on vicarious calibration results during its first 34 months of in-orbit operation (from March 2022 to December 2024). The results indicate that DPC/GaoFen-5(02) experiences obvious degradation in blue to green visible bands, with more pronounced degradation near nadir region than at edges.The radiometric response in near-infrared bands remains relatively stable. For 443, 490, 565, 670, and 865 nm bands, the relative radiometric response variations within FOV are 2.6%, 4.5%, 2.6%, 5.2%, and 5.2% during early on-orbit period, respectively. By December 2024, the FOV average radiometric response coefficients decrease from 0.985 to 0.953, 0.998 to 0.974, 0.992 to 0.975, 0.976 to 0.970, and 0.974 to 0.972, and relative radiometric response variations increase to 8.2%, 6.3%, 4.2%, 5.3%, and 5.3%, respectively. Compared to DPC sensor on GaoFen-5 satellite, DPC/GaoFen-5(02) sensor shows more stable radiometric performance and provides high-quality observational data for future atmospheric remote sensing applications.