An Investigation on Inter-Calibrating EMI/GF-5 With TROPOMI/S5p in Ultraviolet–Visible Spectra

Inter-calibration is a general method to harmonize the in-orbit spectral radiance of the target instrument with the reference instrument in the visible and infrared spectra. This study proposes and applies an inter-calibration method to the space-based Ultraviolet and visible spectra (UVIS) grating radiometers. Based on pixel pairs derived from simultaneous nadir overpasses (SNO), the in-orbit spectral radiances from the environmental trace gas monitoring instrument onboard the Chinese high-resolution remote sensing satellite GaoFen-5 (EMI/GF-5) and the TROPOspheric Monitoring Instrument onboard the Sentinel-5 Precursor satellite (TROPOMI/S5p) are compared through the double difference (DD) method, where the individual window channel method is also adopted for comparison. Uncertainties from SNO thresholds, instrument spectral specification, and radiative transfer simulation are estimated. The SNO thresholds, including time, distance, viewing angle, and scene uniformity of 300 s, 3 km, 0.01, and 0.01, are determined to get collocations for UVIS instruments. Several factors, including instrument line shape (ILS), instrument spectral accuracy, reference solar spectrum, surface albedo, and atmospheric polarization, can influence the radiometric difference obtained by the DD method. Accurate SAO2010 solar spectrum and wavelength accuracy can substantially reduce the dependence of radiometric difference on wavelength. Errors introduced by scalar approximation and parameterized ILS can be neglected. The surface albedo greatly affects the simulated radiance at wavelengths longward of 330 nm. The results show that radiometric differences obtained by using the DD method are consistent with those derived from the individual window channel method, with a discrepancy of 0.4%. An ocean–land calibration difference of 1.5% is found between EMI and TROPOMI at the UVIS band. Inter-calibration using the DD method can minimize the spectral variations in full spectral radiance inter-comparison for grating hyperspectral instruments, and the radiometric difference can be estimated comprehensively. The wavelength dependence can also be easily derived from the results of the DD method. The DD method is recommended for radiometric comparison for UVIS hyperspectral spectrometers.