Atmospheric Measurement Techniques (Aug 2024)

A new non-linearity correction method for the spectrum from the Geostationary Inferometric Infrared Sounder on board Fengyun-4 satellites and its preliminary assessments

  • Q. Guo,
  • Q. Guo,
  • Q. Guo,
  • Q. Guo,
  • Y. Liu,
  • X. Wang,
  • X. Wang,
  • X. Wang,
  • W. Hui,
  • W. Hui,
  • W. Hui

DOI
https://doi.org/10.5194/amt-17-4613-2024
Journal volume & issue
Vol. 17
pp. 4613 – 4627

Abstract

Read online

Non-linearity (NL) correction is a critical procedure to guarantee that the calibration accuracy of a spaceborne sensor approaches a reasonable level (i.e., better than 0.5 K). Unfortunately, such an NL correction is still not used in spectrum calibration from the Geostationary Interferometric InfraRed Sounder (GIIRS) onboard the Fengyun-4A (FY-4A) satellite. Different from the classical NL correction method where the NL coefficient is estimated from out-band spectral artifacts in an empirical low-frequency region, originally with prelaunch results and updated under in-orbit conditions, a new NL correction method for a spaceborne Fourier transform spectrometer (including GIIRS) is proposed. In particular, the NL parameter μ, independent of different working conditions (namely the thermal fields from environmental components), can be determined from laboratory results before launch and directly utilized during in-orbit calibration. Moreover, to overcome the inaccurate linear coefficient from the two-point calibration that influences the NL correction, an iteration algorithm is established to make both the linear and the NL coefficients converge to their stable values, with relative errors less than 0.5 % and 1 %, respectively, which is universally suitable for NL correction of both infrared and microwave sensors. Using the onboard internal blackbody (BB), which is identical to the in-orbit calibration, the final calibration accuracy for all the detectors and all the channels with the proposed NL correction method is validated to be around 0.2–0.3 K at an ordinary reference temperature of 305 K. Significantly, the relative error in the classical method NL parameter immediately transmitting to that of the linear one in theory, which inevitably introduces some additional errors around 0.1–0.2 K for the interfering radiance no longer exists. Moreover, the adopted internal BB with higher emissivity produces better NL correction performance in practice. The proposed NL correction method is scheduled for GIIRS implementation on board the FY-4A satellite and its successor after modifying their possible spectral response function variations.