Improved SAGE II Water Vapor Inversion

Archived versions of SAGE~II water vapor data exhibit some problems, including strong sensitivity to enhanced aerosol and seasonal variability unlike other stratospheric data sets as well as an apparent dry bias near the hygropause. A modification to the water vapor inversion technique has been developed which addresses these problems. Water vapor is measured with the SAGE~II instrument using a single filter-based channel, nominally at 935~nm. It is hypothesized that the spectral response of the filter underwent a significant change in the first few years after launch, broadening and shifting to longer wavelengths. This is in accordance with evidence that the physically similar { {NO_2}} channel, nominally at 448~nm, also underwent a change in wavelength after launch. The { {NO_2}} retrieval was reworked and resulted in significant improvement in { {NO_2}} in version 6.1. A similar change in the water vapor channel would affect the apparent strength of the water vapor feature as well as the contributions of other components such as molecular and aerosol scattering. It was found that a single constant filter adjustment is adequate to yield reasonable agreement with a water vapor climatology from HALOE for the period of 1986 through the present except immediately following the Pinatubo eruption when aerosol interference remains a problem. Better agreement is obtained by allowing slightly different aerosol clearing models for the post-El Chichon period and the post-Pinatubo period. Water vapor data in this newest version show marked improvement over earlier versions.