Because of control measures from international agreements, it is predicted that previous decreases in ozone should lessen and eventually turn around as a recovery of ozone. It is important to statistically determine from available ozone data that a change in the rate of downward trend in ozone occurs and that there is an overall turnaround in the downward trend. For this purpose, characteristics of a statistical trend model that allows for a change in trend (either a flattening or a turnaround) at some specified date are investigated. This model permits the use of data from before as well as after the change date to be used in the trend analysis and affords the opportunity for detection of a change in the downward ozone trend as an early signal in addition to detection of positive trend (recovery) after the change. Total Ozone Mapping Spectrometer data and ground station total ozone data from the northern midlatitudes are used to estimate the number of years of data required to detect a change in trend and to detect a positive trend recovery, under two different assumptions about future ozone trends. Results show that for midlatitudes (30¿--60¿) a positive change in trend of reasonably assumed magnitude can be detected within ~7--8 years from the change date, whereas detection of a positive trend recovery can require roughly 15--20 years for southern midlatitude zonal average data and 20--25 years for northern midlatitude zonal average data. In addition, preliminary trend analysis of recent ground station total ozone data is performed to illustrate estimation using the proposed trend model, with some informative results.