Using Frequency Distributions of OSIRIS Ozone Observations to Characterize Stratosphere-Troposphere Exchange

Satellite observations from limb-scanning instruments such as OSIRIS provide high vertical resolution of chemical species (e.g. ozone) in the atmosphere, but more coarse resolution in the horizontal because of the satellite's high flight speed. The low horizontal resolution means small-scale events, such as incidences of stratosphere-troposphere exchanges (STE) of air, cannot be adequately characterized by individual observations, nor are they well represented by averages over many observations spanning a large geographical range.

We use the temporal and spatial geolocation coordinates of the OSIRIS vertical scans to map each OSIRIS ozone observation to specific coordinates in a potential vorticity - potential temperature (pv-theta) framework. Such mapping is not uncommon, however, presenting chemical species on, for instance, isentropic levels at pv-equivalent latitudes assumes conservation of both theta and pv on the timescales of interest. This assumption is not valid when considering STE events. We therefore treat all observations falling into a specific gridbox of the pv-theta space as a probability density function whose distribution characteristics, particularly the lowest and highest centiles, can provide global information on regions of STE.

Our main area of interest is the mixing in the upper troposphere - lower stratosphere (UTLS) layer, and we show seasonal climatologies for this level using nine years of OSIRIS ozone data. In the lower troposphere, our characterisations in pv-theta space provide ozone distribution information for regions below levels that OSIRIS can "see" directly, thereby enhancing the value of the satellite measurements.