Temperature and slant path effects in Dobson and Brewer total ozone measurements

There is a worldwide tendency to replace Dobson spectrophotometers in ground‐based total ozone (TOZ) measurements by more advanced Brewer spectrophotometers. Ensuring the homogeneity of these data sets is of utmost importance if changes in TOZ of a few percent over long time periods are to be diagnosed accurately. Previous studies have identified a seasonal bias of a few percent between midlatitude Brewer and Dobson measurements. At Arosa (Switzerland), two Dobson and three Brewer instruments have been colocated since 1998, providing a unique data set of quasi‐simultaneous observations, invaluable to study systematic differences between these measurements. The differences are partially attributed to the seasonal variability in atmospheric temperatures and ozone slant paths (OSP). The sensitivity to the temperature dependence of the ozone absorption cross section is calculated for each operational Brewer spectrophotometers at Arosa by using different high‐ and low‐resolution reference spectra appropriately weighted with the instruments' slit functions, whereas the information on the primary standard instruments is used for all the Dobson instruments. The Brewer retrieval algorithm reveals a higher sensitivity to the reference spectra applied than the Dobson. When adopting the Bass and Paur (1985) or Malicet et al. (1995) ozone absorption spectra with their specific temperature dependence, and correcting for the OSP effect, the seasonal bias between Dobson and Brewer TOZ measurements is reduced to 0.6%. Conversely, these differences increase when using the spectral data of Burrows et al. (1999). This finding illustrates that the accuracy of ground‐based spectrophotometric TOZ measurements is limited by the uncertainty in the ozone cross sections measured by different internationally leading laboratories.