论文部分内容阅读
Hyperspectral data have important research and application value in the fields of meteorology and remote sensing. With the goal of improving retrievals of atmospheric temperature profiles, this paper outlines a novel temperature channel selection method based on singular spectrum analysis (SSA) for the Geostationary Interferometric Infrared Sounder (GIIRS), which is the first infrared sounder operating in geostationary orbit. The method possesses not only the simplicity and rapidity of the principal component analysis method, but also the interpretability of the conventional channel selection method. The novel SSA method is used to decompose the GIIRS observed infrared brightness temperature spectrum (700?1130 cm?1), and the reconstructed grouped components can be obtained to reflect the energy variations in the temperature-sensitive waveband of the respective sequence. At 700?780 cm?1, the channels selected using our method perform better than IASI (Infrared Atmospheric Sounding Interferometer) and CrIS (Cross-track Infrared Sounder) temperature channels when used as inputs to the neural network retrieval model.