Consistency of time series and trends of stratospheric ozone as seen by ozonesonde, SAGE II, HALOE, and SBUV(/2)

[1] We examined the consistency of time series for stratospheric ozone obtained by ozonesondes, SAGE II, HALOE, and SBUV(/2) in the northern extratropics and find excellent agreement for interannual variability and trends of ozone for 50–12.5 hPa over Europe. Ozonesonde, SAGE, and HALOE data agree well for 200 hPa–12.5 hPa. Ozone in the past few years is 6% below values in the early 1980s for 80–20 hPa, while values for 200–125 hPa are similar to those in the early 1980s. Small differences in monthly mean ozone and trends between ozonesondes and SAGE are caused by spatial and temporal sampling differences. The quality of SBUV(/2) Version 8 ozone data was evaluated by comparison to the homogeneous SAGE data. Most biases between SBUV(/2) and SAGE data are within ±10% except for large negative biases of SBUV(/2) at 50 hPa in the tropics. There is less homogeneity in the SBUV(/2) time series for the upper stratosphere, with negative biases for NOAA-11 after 1997 at 3–5 hPa and positive biases for NOAA-9 and 16 at 2–5 hPa, compared to earlier SBUV data, implying that the later data may lead to errors in calculated ozone trends. SBUV(/2) ozone trends agree well with SAGE trends in the midlatitude lower stratosphere for 1984–2000, but SBUV(/2) trends are as much as 4%/decade more positive than SAGE trends in the upper stratosphere. We find that SAGE sampling does not influence zonal mean trends, implying that SAGE/SBUV(/2) trend differences are related to data quality and to the different coordinate systems in which ozone is measured.