Ozone estimates derived from Dobson direct sun measurements: effect of atmospheric temperature variations and scattering.

We have performed an analysis of the impact of the temperature sensitivity of the ozone absorption coefficients on estimates of total atmospheric ozone obtained by the Dobson spectrophotometer operating in direct sun mode. In general, the higher the mean ozone temperature the greater will be the tendency to overestimate the ozone amount. The spreads in ozone residuals over the temperature models we investigated were 3%, 4%, and 6% for the A, C, and D line pairs, respectively, whereas for coupled line pairs the spread was only about 2%. The A-C-D triplet showed a very small temperature effect, the spread being probably less than 2%. For the A-D system, currently recommended by the WMO, the computed spread was 2.4 +/- 0.5%. A Monte Carlo model was applied to investigate the potential impact of scattered radiation entering the system. The effect has been computed for various conical fields of view. For clear sky conditions with no aerosols present, the error introduced appears to be less than 1%. When a tropospheric aerosol model was inserted, however, significant errors were observed. For the models we studied aerosol attenuation resulted in overestimates of total ozone up to 8%, but the impact of scattered radiation was to reduce the overestimate, and, in some cases, the scattering and attenuation effects may balance for a realistic Dobson system. Both effects increased from the A to the C to the D line pairs. The results indicate that line pair coupling reduces the combined error due to both sources to less than 1%.