Abstract.A one-dimensional chemistry model is applied to study the stable hydrogen (D) and stable oxygen isotope ( 17 O, 18 O) composition of water vapour in stratosphere and mesosphere. In the troposphere, this isotope composition is determined by "physical" fractionation effects, that are phase changes (e.g. during cloud formation), diffusion processes (e.g. during evaporation from the ocean), and mixing of air masses. Due to these processes water vapour entering the stratosphere first shows isotope depletions in D/H relative to ocean water, which are ∼5 times of those in 18 In the stratosphere the known mass-independent fractionation (MIF) signal in O 3 is in a first step transferred to the NO x family and only in a second step to HO x and H 2 O. In contrast to CO 2 , O( 1 D) only plays a minor role in this MIF transfer. The major uncertainty in our calculation arises from poorly quantified isotope exchange reaction rate coefficients and kinetic isotope fractionation factors.

No abstract

Abstract. We report the first high-precision measurements of δ 18O and Δ 17O at high southern latitudes that can resolve changes in the isotopic composition of water vapor in the lowermost stratosphere and upper troposphere. A strong increase of δ 18O with decreasing mixing ratio above the tropopause is evident in the data. Since also the water vapor mixing ratio decreases above the tropopause, the effect seen in the isotope data can be explained by mixing of moist air from the tropopause with dry stratospheric air. However, the source of this dry stratospheric air is not known; both fast transport from the extratropical tropopause or mixing with air from the dehydrated polar vortex are likely. The magnitude of the Δ 17O-anomaly (departure from mass-dependent fractionation (MDF)) was below 2 per mil for each datapoint, and a zero anomaly in lower level stratospheric water vapor is possible. Various transport histories for the stratospheric data are discussed based on the mixing ratio and isotope data.

A new continuous-flow system for the analysis of the complete stable isotopic composition of water vapor has been developed. The sample size is reduced to only 120 microg (identical with 120 nL of liquid substance) of water, yielding precisions of about 0.7, 1.3 and 7 per thousand for delta17O, delta18O and delta2H, respectively. The total time for the analysis of a sample is about 150 min including purging times. Oxidized steel surfaces can be a source of memory effects which can be corrected for. The system is predestined for atmospheric applications in the tropopause region, as the sample can be directly introduced into the system from a cryogenic trap.