Interference errors in infrared remote sounding of the atmosphere

R. Sussmann (1) , T. Borsdorff (1)

1 IMK-IFU - Institut für Meteorologie und Klimaforschung - Atmosphärische Umweltforschung

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

Domaines: Océan, Atmosphère

R. Sussmann, T. Borsdorff. Interference errors in infrared remote sounding of the atmosphere. Atmospheric Chemistry and Physics Discussions, 2006, 6 (6), pp.13027-13073. ⟨hal-00302363⟩