A critical analysis of Stratospheric Aerosol and Gas Experiment ozone trends

Abstract

A latitudinally dependent reference height error in Stratospheric Aerosol and Gas Experiment (SAGE) I ozone profiles is inferred from two separate analyses: first, by essentially comparing the mean height of the stratospheric ozone layer between SAGE I and SAGE II measurements and second, by examining SAGE I/II differences above the ozone peak using solar backscattered ultraviolet (SBUV) profiles as transfer standards. Both approaches yield a reference height error in the tropics of approximately 200 m and midlatitude errors of approximately 400 m. Comparisons between SAGE II and SBUV measurements also suggest that SAGE II ozone concentrations are too small at altitudes above approximately 40 km for a few days following high beta angle conditions and that anomalous differences, related to SAGE II, exist in several Januaries. When these periods are filtered out, the SAGE I measurements are corrected for the reference height errors, and the SAGE II measurements are corrected for aerosol interference [Cunnold et al., 1996a], the SAGE I/II ozone trends are found to be in good agreement with the SBUV and total ozone mapping spectrometer trends and with the Hohenpeissenberg ozonesonde results. The SAGE I/II deduced ozone trend from 1979 to 1991 shows an almost negligible ozone loss in the tropical upper stratosphere but an increasingly negative trend, with latitude maximizing at − 1.2%/yr at 2 mbar at 60°S (the highest latitude reported). In the lower stratosphere, decreases of approximately 1%/yr are reported at 68 mbar, and the column ozone decreases at midlatitudes are almost all related to losses below 30 mbar. In the tropics, SAGE I/II suggests stratospheric ozone losses of approximately 0.2%/yr. The solar cycle ozone dependence in the upper stratosphere is less than that previously reported in SBUV analyses, with a local maximum response in the ozone concentration of approximately 3% at midlatitudes at approximately 41 km altitude. In the lower stratosphere, SAGE results confirm previous reports of a sensitivity of approximately a 1.5% change over the solar cycle but indicate that this change is mostly occurring just below the ozone concentration peak.