The measurement of trace gases through remote sounding techniques has led to a better understanding of the processes controlling the structure and variability of the stratosphere. Differences between measurements over space and time are due to atmospheric variability and instrument errors: thus, comparison of measurements can be used to test our knowledge of both. Comparisons of measurements over long time periods are used to identify trends. Balloon-borne infrared emission radiometer instruments have been used to make measurements of midlatitude stratospheric HNO3 spanning a period of twelve years. The timing of the measurements is notable, since they occur before and well after the eruption of Mt. Pinatubo, which significantly perturbed HNO3 levels, complicating prior trend analyses. No significant differences are found between the HNO3 retrievals, although large measurement uncertainties preclude any conclusion concerning trends. Comparisons of measurements that are closely spaced in space and time are useful for satellite validation, where one aims to reduce the effect of atmospheric variability on the estimation of systematic and random errors. A novel technique for the estimation of systematic error, which differentiates between additive and multiplicative bias, is introduced. In a comparison of measurements by the ACE-FTS and Aura MLS instruments, significant multiplicative biases are identified and described. In order to validate the reported random errors (RREs) of measurements, satellite validation studies often focus on measurements in the tropical stratosphere, where variability is weak. The scatter in tropical measurements can then be used as an upper limit on instrument precision. In an analysis of tropical measurements by the ACE-FTS, scatter is found to be roughly consistent with the RREs for H2O and CO. The scatter in measurements of O3, HNO3, and N2O, while larger than the reported random errors, is roughly consistent with the variability simulated in the Canadian Middle Atmosphere Model. This work implies that the random error of the ACE-FTS measurements is smaller than the weak natural variability of the tropical stratosphere.

The measurement of trace gases through remote sounding techniques has led to a better understanding of the processes controlling the structure and variability of the stratosphere. Differences between measurements over space and time are due to atmospheric variability and instrument errors: thus, comparison of measurements can be used to test our knowledge of both. Comparisons of measurements over long time periods are used to identify trends. Balloon-borne infrared emission radiometer instruments have been used to make measurements of midlatitude stratospheric HNO3 spanning a period of twelve years. The timing of the measurements is notable, since they occur before and well after the eruption of Mt. Pinatubo, which significantly perturbed HNO3 levels, complicating prior trend analyses. No significant differences are found between the HNO3 retrievals, although large measurement uncertainties preclude any conclusion concerning trends. Comparisons of measurements that are closely spaced in space and time are useful for satellite validation, where one aims to reduce the effect of atmospheric variability on the estimation of systematic and random errors. A novel technique for the estimation of systematic error, which differentiates between additive and multiplicative bias, is introduced. In a comparison of measurements by the ACE-FTS and Aura MLS instruments, significant multiplicative biases are identified and described. In order to validate the reported random errors (RREs) of measurements, satellite validation studies often focus on measurements in the tropical stratosphere, where variability is weak. The scatter in tropical measurements can then be used as an upper limit on instrument precision. In an analysis of tropical measurements by the ACE-FTS, scatter is found to be roughly consistent with the RREs for H2O and CO. The scatter in measurements of O3, HNO3, and N2O, while larger than the reported random errors, is roughly consistent with the variability simulated in the Canadian Middle Atmosphere Model. This work implies that the random error of the ACE-FTS measurements is smaller than the weak natural variability of the tropical stratosphere.

A high-resolution infrared Fourier Transform Spectrometer (FTS) has been operational at the Toronto Atmospheric Observatory (TAO)since May 2002. An optimal estimation retrieval technique is used to analyse the observed spectra and provide regular total and partial column measurements of trace gases in the troposphere and stratosphere as part of the Network for the Detection of Atmospheric Composition Change. The quality of these results were assessed through two ground-based validation campaigns, comparisons with three satellite instruments, and comparison with a three-dimensional chemical transport model. The two ground-based campaigns involved two lower-resolution FTS instruments: the University of Toronto FTS and the Portable Atmospheric Research Interferometric Spectrometer for the Infrared. The first campaign took place over the course of four months and is the longest side-by-side intercomparison of ground-based FTS instruments, to date. The second campaign was more focused and involved all three instruments measuring over a two-week period. Simultaneous measurements of O3, HCl, N2O, and CH4 were recorded and average total column differences were all < 3.7% in the extended campaign, and < 4.5% in the focused campaign. Satellite-based comparisons were done with the SCanning and Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY), the Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE-FTS), and the Optical Spectrograph and InfraRed Imager System (OSIRIS). Total column CO, CH4, and N2O compared with SCIAMACHY all had average differences < 10% with results from the TAO-FTS being as good as, or better, than that of other instruments. Validation with the ACE-FTS showed that average partial columns of O3, NO2, N2O, CH4, and HCl were within 10% while observations of CO and NO each had an average bias of about 25%. Comparisons of monthly average partial column O3 and NO2 with OSIRIS were highly correlated (R = 0.82-0.97) with monthly mean differences of < 3.1% for O3 and < 2.6% for NO2. Finally, comparisons with the GEOS-Chem chemical transport model revealed that the model consistently over-estimates tropospheric columns of CO and C2H6 observed at TAO. It was determined that the enhanced CO values were partially due to the North American emissions specified in the model, but more work must be done in the future if the source of this discrepancy is to be fully explained.

A high-resolution infrared Fourier Transform Spectrometer (FTS) has been operational at the Toronto Atmospheric Observatory (TAO)since May 2002. An optimal estimation retrieval technique is used to analyse the observed spectra and provide regular total and partial column measurements of trace gases in the troposphere and stratosphere as part of the Network for the Detection of Atmospheric Composition Change. The quality of these results were assessed through two ground-based validation campaigns, comparisons with three satellite instruments, and comparison with a three-dimensional chemical transport model. The two ground-based campaigns involved two lower-resolution FTS instruments: the University of Toronto FTS and the Portable Atmospheric Research Interferometric Spectrometer for the Infrared. The first campaign took place over the course of four months and is the longest side-by-side intercomparison of ground-based FTS instruments, to date. The second campaign was more focused and involved all three instruments measuring over a two-week period. Simultaneous measurements of O3, HCl, N2O, and CH4 were recorded and average total column differences were all < 3.7% in the extended campaign, and < 4.5% in the focused campaign. Satellite-based comparisons were done with the SCanning and Imaging Absorption spectroMeter for Atmospheric CHartographY (SCIAMACHY), the Atmospheric Chemistry Experiment Fourier Transform Spectrometer (ACE-FTS), and the Optical Spectrograph and InfraRed Imager System (OSIRIS). Total column CO, CH4, and N2O compared with SCIAMACHY all had average differences < 10% with results from the TAO-FTS being as good as, or better, than that of other instruments. Validation with the ACE-FTS showed that average partial columns of O3, NO2, N2O, CH4, and HCl were within 10% while observations of CO and NO each had an average bias of about 25%. Comparisons of monthly average partial column O3 and NO2 with OSIRIS were highly correlated (R = 0.82-0.97) with monthly mean differences of < 3.1% for O3 and < 2.6% for NO2. Finally, comparisons with the GEOS-Chem chemical transport model revealed that the model consistently over-estimates tropospheric columns of CO and C2H6 observed at TAO. It was determined that the enhanced CO values were partially due to the North American emissions specified in the model, but more work must be done in the future if the source of this discrepancy is to be fully explained.