Pw. Mote et al., AN ATMOSPHERIC TAPE-RECORDER - THE IMPRINT OF TROPICAL TROPOPAUSE TEMPERATURES ON STRATOSPHERIC WATER-VAPOR, JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES, 101(D2), 1996, pp. 3989-4006
We describe observations of tropical stratospheric water vapor q that
show clear evidence of large-scale upward advection of the signal from
annual fluctuations in the effective ''entry mixing ratio'' q(E) of a
ir entering the tropical stratosphere, In other words, air is ''marked
,'' on emergence above the highest cloud tops, like a signal recorded
on an upward moving magnetic tape, We define q(E) as the mean water va
por mixing ratio, at the tropical tropopause, of air that will subsequ
ently rise and enter the stratospheric ''overworld'' at about 400 K. T
he observations show a systematic phase lag, increasing with altitude,
between the annual cycle in q(E) and the annual cycle in q at higher
altitudes, The observed phase lag agrees with the phase lag calculated
assuming advection by the transformed Eulerian-mean vertical velocity
of a q(E) crudely estimated from 100-hPa temperatures, which we use a
s a convenient proxy for tropopause temperatures, The phase agreement
confirms the overall robustness of the calculation and strongly suppor
ts the tape recorder hypothesis, Establishing a quantitative link betw
een q(E) and observed tropopause temperatures, however, proves difficu
lt because the process of marking the tape depends subtly on both smal
l- and large-scale processes, The tape speed, or large-scale upward ad
vection speed, has a substantial annual variation and a smaller variat
ion due to the quasi-biennial oscillation, which delays or accelerates
the arrival of the signal by a month or two in the middle stratospher
e. As the tape moves upward, the signal is attenuated with an e-foldin
g time of about 7 to 9 months between 100 and 50 hPa and about 15 to 1
8 months between 50 and 20 hPa, constraining possible orders of magnit
ude both of vertical diffusion K-z and of rates of mixing in from the
extratropics. For instance, if there were no mixing in, then K-z would
be in the range 0.03-0.09 m(2) s(-1); this is an upper bound on K-z.