Stepwise changes in stratospheric water vapor?

The sparse data available of stratospheric water vapor since the 1950s suggests a positive long-term trend that cannot be explained by the methane increase and what is known about temperature trends around the tropical tropopause, which constrain the amount of water entering the stratosphere. Here, we discuss the 1991-2005 time series of stratospheric water (and methane) measurements from the Halogen Occultation Experiment (HALOE). The high sampling, global coverage and measurement of methane render HALOE data ideal to check the data for self-consistency and to pinpoint the time of changes in entry mixing ratios. In addition to the well-known 'drop' in October 2000, the HALOE data at 10 hPa and less suggest a steep increase in entry mixing ratios shortly before the beginning of the HALOE measurements. Model calculations using simple representations of the stratospheric age of air spectrum in the tropics show that the very dry phase may be explained by a range of scenarios: A long (several years) dry phase followed by a step increase with amplitude 0.3 ppmv; a shorter (>= 1 year) dry pulse with amplitude 0.6 ppmv; or steep linear trends over about 2 years with total increases similar to the step scenarios. The drop in October 2000 coincides with anomalously large eddy heat fluxes in the Southern hemisphere and low tropopause temperatures, but no such relation is found for the situation around 1991. The coincidence with the eruption of Mt. Pinatubo is discussed. The evidence for the results presented here is circumstantial, but they would imply that decoupling between stratospheric water trends and tropical tropopause temperatures can occur on short timescales.