Rejection criteria for open-path Fourier transform infrared spectrometry during continuous atmospheric monitoring

Over 32,000 interferograms measured during open-path Fourier transform infrared (OP/Fr-IR) measurements at dairy and hog farms were evaluated for anomalies. Five types of anomalies could be distinguished: a reduction in the interferogram intensity because of weather-related optical misalignment; an increase in the amplitude of interferograms measured with too short a path-length that leads to a non-linear detector response; a periodic interference caused by wind-induced vibrations; the presence of spikes in the interferogram; and an increase in the noise level of the interferogram (and hence of the spectrum) because of the effect of electrical interference. Prior to testing for the presence of anomalous data, each interferogram is subjected to a high-pass filter. A noise level index is then calculated from the wings of the interferogram and interferograms are rejected if the value of this parameter is too high. When the criteria developed in this project are applied, OP/Fr-IR spectra may be measured at 1-min intervals over a period of several days. (C) 2006 Published by Elsevier B.V.