Total ozone variability and trend estimates from ground-based and satellite observations in the southern subtropics

International audienceDespite its low abundance in the atmosphere, ozone is an important component that prevents solar UV from reaching the Earth's surface. It contributes to the radiative balance, and hence has impact on climate change. The long-term evolution of stratospheric ozone depends on changes of both stratospheric and tropospheric constituents such as ozone-depleting substances, greenhouse gases, water vapor, and aerosols. It also depends on changes in the troposphere and in the stratosphere caused by natural variability and anthropogenic forcing (WMO, 2006). As reported by many authors, air penetrates into the stratosphere primarily through the tropical tropopause, while the tropical stratosphere is a region of major exchange with mid-latitude regions through the subtropical barriers (Portafaix et al., 2003; Bencherif et al., 2003; Semane et al., 2006; Bencherif et al., 2007). The present contribution reports on a comparative study on total ozone observations obtained from ground-based and satellite measurements over subtropical sites. The study examines the climatology and variability of total ozone. It is based on more than 15 years of continuous observations at four sites in the southern subtropics: Reunion Island (21°S, 55.5°E), Bauru (22°S, 49°O), Irene (25.5°S, 28.1°E) and Springbok (29.6°S, 17.9°E). Depending on the observational periods of each instrument, ground-based total ozone measurements are compared with satellite datasets, in terms of daily and monthly variations. The study focuses on comparisons with total ozone values measured from space by TOMS (Total Ozone Mapping Spectrometer), OMI (Ozone Monitoring Spectrometer) and by IASI (Infrared Atmospheric Sounding Interferometer) on the MetOp satellite. The obtained monthly-mean time-series of total ozone are analyzed by a multi-regression model named Trend-Run (Bencherif et al., 2006; Bègue et al., 2010). It is a trend model based on the linear multi-regression principal, i.e., the ozone signal is broken into a sum of several forcings (seasonal cycles, QBO, ENSO, IOD, Solar cycles, ...) that explain most of its variability. The trend values are then derived from the residual terms as a linear function