Longitude-dependent decadal changes of total ozone in boreal winter months during 1979-92

A statistical analysis shows that the decadal change of zonally asymmetric total ozone (Total Ozone Mapping Spectrometer data) has a distinct spatial similarity with the decadal change of the 300-hPa geopotential patterns during December-February of 1979-92 in the Northern Hemisphere: Regions of ozone increase correspond with regions of geopotential decrease and vice versa. An area of strong ozone decrease above Europe is surrounded by areas of ozone increase over the North America-North Atlantic region and the eastern Europe-Siberia region, in both December and January. In February the picture is changed by the appearance of a region of ozone increase above Europe. In all three months an area of ozone decrease exists in the east Asia-Pacific region, with one region in January and two regions in December and February. In all cases the centers of decadal ozone change as well as their month to month alterations are significantly anticorrelated with those of the 300-hPa geopotential. These statistical results are investigated with a linear quasigeostrophic transport model. Taking into account 14 layers between 500 and 1 hPa the contributions of horizontal and vertical advection to ozone transport were estimated separately. In December and January the complexity of the spatial distribution of ozone change is mainly due to the contribution of advection by wavenumbers 3 and 4, but wavenumbers 1 and 2 also contribute a comparable amount. The January ozone decrease above Europe has different causes: in northeast Europe the contribution from wavenumbers 1 and 2 dominates, whereas in central Europe that from horizontal advection by wavenumbers 3 and 4 prevails. In February the spatial structure is mainly determined by wavenumbers 1 and 2 alone, with equal contributions from horizontal and vertical advection. In the vertical distribution in midlatitudes the essential contributions come from the height region between the upper troposphere and the ozone layer maximum. December's and January's structures are similarly determined by wavenumbers 3 and 4 with unchanged vertical phase. Here the contribution of the horizontal advection to the total decadal ozone change occurs in a higher-altitude range (100-50 hPa) and that of the horizontal together with the vertical advection in a lower-altitude range (200-100 hPa). In February the wavenumbers 1 and 2 determine to a large extent the height distribution of the ozone trend in the midlatitudes by the contribution of the horizontal advection.