This spectral model provides an accurate method of analyzing low frequency components directly in terms of their physical parameters and, by their removal, a residual time series for subsequent spectral analysis without many of the normal assumptions involved in low frequency removal. Typical methods of removing trends, or very low frequency components, are empirical, generally produce spectral distortion in the residual time series, and require subjective decisions as how to treat the ends of, and gaps between, data segments. In some instances, especially where the time series are very transient or nonstationary, or where low frequencies are to be studied, this residual series should greatly improve the ability to accurately interpret the spectrum and its variation with season and with transient processes. Future efforts should be directed to analyzing other segments of the Viking record, better fitting methods for the large gaps, accounting for the residual autocorrelation by using generalized least square techniques, exploring higher harmonics, comparing these analyses with numerical model results and especially with data from future missions. Of even more importance is insuring that similar measurements are collected from future missions since it is possible that future pressure measurements could be inferior to the Viking measurements. It would seriously compromise the understanding of the Martian climate and its interannual variation if we fail to obtain measurements of similar accuracy and utility in the future.
This work was, in part, supported by the National Aeronautics and Space Administration under the Planetary Atmospheres Program and by the National Science Foundation.