Robustness of atmospheric trace gas retrievals obtained from low-spectral-resolution Fourier transform infrared absorption spectra under variations of interferogram length

Abstract

This study examines the sensitivity of atmospheric trace gas column retrievals from ground-based Fourier transform interferometer measurements to variations in the number of points in the recorded interferograms. Shortening an interferogram can be part of standard FTIR data processing and typically occurs with a convolution operation on the interferogram. Shortening will alter the leakage pattern in the associated spectrum, and we demonstrate that the removal of a relatively small number of points from the interferogram edges creates a beat pattern in the difference of the associated spectra obtained from the original and shortened interferograms. For low-spectral-resolution interferometers, the beat pattern may exceed the spectral noise level, and, if this occurs, the number of points in the interferogram can become a strong influence parameter for the retrieved trace gas. In a case study with formaldehyde retrievals obtained from low-resolution spectra in Sodankylä and Kolkata, we show that the retrieval software does not accurately model the leakage pattern and that interferogram shortening has a large effect on atmospheric gas column retrievals that exceeds the estimated retrieval uncertainty of approximately 15%-30%. This sensitivity of the retrieval algorithm to the length of the underlying low-resolution interferogram can be reduced by applying a nontrivial apodization such as Norton-Beer apodization. For the Sodankylä case study the correlation between formaldehyde columns obtained from low- and high-resolution measurements increased from 0.72 (without apodization) to 0.93 (Norton-Beer strong apodization).