Decadal variations in the tropical tropopause: Implications for stratospheric water vapor

Although water vapor concentrations in the stratosphere are very low (a few parts per million), variations in stratospheric moisture content have important climate ramifications. Recent research indicates that stratospheric water vapor concentrations have increased over the past few decades, which may be related to lower-stratospheric cooling and depletion of stratospheric ozone. The reality of, and the possible causes of, the water vapor trend are poorly understood. However, the trend may be related to changes in the characteristics of the tropical tropopause, because, in the tropical lower stratosphere, flux through the tropical tropopause is the major source of stratospheric water vapor. A warming of the tropical tropopause could help explain the observed stratospheric humidity trends.

Newell and Gould-Stewart [JAS, 1981] hypothesized that cross-tropopause water vapor flux may be limited to the coldest regions of the tropical tropopause, where temperatures (and therefore water vapor saturation mixing ratios) are low enough to explain the very low stratospheric water vapor concentrations. This "stratospheric fountain" mechanism calls for preferential periods (northern hemisphere winter) and regions (deep convective regions above the western Pacific/Indian Ocean warm pool) of cross-tropopause transport. This hypothesis was based on examination of monthly mean 100 hPa data from tropical radiosonde stations. However, several recent studies (Dessler, GRL, 1998; Vömel and Oltmans, GRL, 1999; Zhou et al., JGR, accepted) have reexamined the fountain hypothesis and suggested that it may not be necessary to account for observed stratospheric water vapor levels.

We have recently created a new radiosonde-based dataset to document the climatology and variability of various characteristics of the tropical tropopause. In this presentation, we will examine the variability of the tropopause water vapor saturation mixing ratio in the region 15N-15S during 1958-1997. Two questions will be addressed: (1) Does the observed variability in tropopause saturation mixing ratio help explain the observed trend in lower stratospheric humidity? We find that it does not; a cooling of the tropical tropopause is observed. (2) Does the stratospheric fountain hypothesis depend on either the use of 100 hPa data, the use of monthly mean data, or the data period analyzed? We find that all of these factors are important. Daily tropopause-level (not 100 hPa level) saturation mixing ratios as low as stratospheric mixing ratios occur in all seasons, and at most locations throughout the deep tropics, so that a preferred fountain region/season is not necessary.