Content area
The SAGE-I, II ozone profiles are compared with coincident SBUV and MLS measurements to investigate (1) the systematic differences between SAGE-I and SAGE-II, (2) the aerosol interference on SAGE ozone retrievals, and (3) how the long term trends in SAGE ozone would be affected by such effects. The reference height in SAGE-I ozone profiles are inferred to be systematic lower than that in SAGE II by approximately 300m. A sunrise/sunset difference in SAGE ozone, approximately 4-6%, is found in the tropical upper stratosphere above 5 mb. This could result from (1) a high beta angle effect, (2) anomalous SAGE ozone values in January, and (3) atmospheric tide.
SAGE II/MLS comparisons between September 1991 and December 1993 indicate that SAGE ozone retrievals underestimate the aerosol contribution at 0.6 $\mu$m. For the situation of the aerosol layer optical depth at 1.02 $\mu$m in each UARS layer to be less than $2\times10\sp{-3}$, the sensitivity of SAGE ozone to aerosol is approximately $3\times10\sp{10}$ cm$\sp{-3}/10\sp{-3}$ aerosol layer optical depth. This suggests that an approximately 3% aerosol contribution at 0.6 $\mu$m is being interpreted as ozone.
The ozone long term trends from SAGE I/II are recalculated after correcting the reference height error in SAGE I and removing the aerosol interference from SAGE-II ozone values. In the upper stratosphere the maximum ozone loss, approximately $-1.0\sim1.2$%/year, is located near 2 mb at $50\sp\circ\sim60\sp\circ$ for both hemispheres. A minimum downward trend of $-0.4\sim-0.6$%/year is found in the tropics. The high beta angle effect and the anomalous ozone values in January, if not removed, could induce a downward ozone trend bias by $0.1\sim0.2$%/year in the tropical upper stratosphere.
Below $30\sim50$mb, the downward ozone trends, approximately $-$0.8 to $-$3%/year, generally increases with latitude and decreasing altitudes. The largest ozone loss of approximately $-$3.0%year is found at 100 mb in the regions 65$\sp\circ$S poleward. The maximum downward ozone trend bias induced by aerosol contamination, approximately $-0.8\sim1.0$%/year, is found in the equator near 46 and 68 mb. This could explain most of the difference between the SAGE II and TOMS column ozone trends between October 1984 and May 1991.