Differences in Dobson Spectrophotometer and Filter Ozonometer Measurements of Total Ozone

Abstract

This paper reviews the results of the first systematic parallel measurements with a Dobson-type spectrophotometer and an M-83 filter ozonometer. Measurements were performed at the University of Sofia during a period of one full year. The M-83 is a working instrument in about one-third of the world's ozone network, and is predominantly used in the U.S.S.R. Observations taken every 15 min on clear sunny days, as well as the monthly and seasonal values, show that the filter ozonometer gives a reading of about 6% less ozone than the Dobson when observations are taken at sun zenith angles <57° (relative path length of sunlight through the atmosphere ? < 1.5). When the sun zenith angle is greater than 60° (? > 2.0) the filter ozonometer readings are 20?30% higher than the Dobson readings. When visibility is limited (V ≤ 5 km), the filter ozonometer readings of ozone are 9?14% higher than the Dobson readings. Differences in the monthly ozone values between adjacent pairs of nearby stations using filter or Dobson instruments give results which agree with those from the simultaneous observations in Sofia. A nomogram is presented giving values of ? as a function of the latitude and the time of the year. Using this nomogram and applying the results of the study it can be concluded that the ozone reported from filter ozonometers stations north of ?40N would be ?20% higher than the actual ozone content during the autumn-winter period. During the summer, up to about 70N, the filter data show less ozone. The causes of the significant differences shown by the filter ozonometer M-83 are probably due to the use of filters with a very wide transmission band (>400 Å). The range of error could be partially decreased if measurements with the filter ozonometer were taken only symmetrically around noon and only when the sun's zenith angle is between 50° and 30° (1.3 < ? < 2.0). It is concluded that the ozone diagrams used for the filter ozonometer should be recomputed, including greater detail on the influence of changes of the effective wavelength (and corresponding changes of the ozone absorption coefficients) as functions of the sun's zenith angle.