Measurements and data analysis from a balloon-borne Fourier transform spectrometer

The Middle Atmosphere Nitrogen TRend Assessment (MANTRA) campaigns consist of a high-altitude balloon launch supported by a ground-based campaign. Campaigns were held in Vanscoy, SK, when stratospheric zonal winds change from easterly to westerly: a time of year called turnaround. An investigation of the climatology and predictability of turnaround is presented. The study determined that the timing of turnaround cannot be predicted from stratospheric zonal wind speeds earlier in the summer. The climatology showed that there is a launch window which contains August 26 through September 5, wherein stratospheric zonal wind speeds are low enough to facilitate a launch.; The University of Toronto's Fourier transform spectrometer (U of T FTS) was converted from a ground-based FTS to a balloon- and ground-based FTS through an extensive electronics and software redevelopment. Its functionality was shown by its successful flight on the MANTRA 200=1 balloon payload, recording two spectra on each detector, and its ground-based data, recorded during the 2001 campaign. Its balloon and ground-based data were compared with other ground-based instruments on-site and the Microwave Limb Sounder (MLS) satellite instrument.; The U of T FTS participated in a ground-based intercomparison campaign in Toronto with the Toronto Atmospheric Observatory Fourier transform spectrometer (TAO-FTS) and the Portable Atmospheric Research Interferometric Spectrometer for the Infrared (PARIS-IR). The resolutions of these three instruments are significantly different. The two lower-resolution instruments (PARIS-IR and the U of T FTS) were found to measure 4-day average total columns of O 3, HCl, N2O and CH4 to within 3.5% of the TAO-FTS total columns. The largest errors were produced by the total column retrievals of the stratospheric species (O3 and HCl). In order to achieve this 3.5% agreement for the stratospheric species, the instrument line shape of the U of T FTS and PARIS-IR instruments had to be taken into account. Much of the remaining error can be attributed to the averaging kernels of the lower-resolution instruments.