作者: R. Sussmann , T. Borsdorff
DOI:
关键词: Spectroradiometer 、 Smoothing 、 Spectral line 、 Satellite 、 Noise (video) 、 Computer science 、 Remote sensing 、 Interference (wave propagation) 、 Depth sounding 、 State vector
摘要: More and more profiles of atmospheric state parameters are being retrieved from remote soundings in the infrared spectral domain. Classical error analysis, which was originally applied to microwave sounding systems, distinguishes between "smoothing errors," "forward model parameter "retrieval noise errors". We show that for "interference errors", have not been treated up now, can be significant. Interference errors originate "interfering species" introduce signatures into measurement overlap with features used retrieval target species. This is a frequent situation spectra where vibration-rotation bands different species often overlap; it case region. paper presents full theoretical formulation interference errors. It requires generalized vector including profile entries all interfering leads averaging kernel matrix made classical kernels plus here defined kernels". The latter together climatological covariances order quantify To illustrate methods we apply them real significant impact on standard CO retrievals ground-based mid-infrared solar absorption spectra. also demonstrate how minimize overall error, trade-off minimizing smoothing error. approach this infrared-active species, includes than two dozen gases relevant climate ozone. And holds kind i.e., ground-based, balloon-borne, airborne, or satellite spectroradiometers.