The strongest and most persistent temperature inversion in the world occurs over the interior of Antarctica in winter. At Vostok (78.45 S, 106.87 E, 3420 m) the average inversion for the six months May to October is 25 oC. Climatological information on the distribution of inversion strength is sparse, since there are or have been only 21 radiosonde stations in Antarctica of which only 3 are in the interior, the region of greatest interest. Two methods have been proposed to map the inversion strength. The widely-quoted map of Phillpot and Zillman is based on the difference between summer and winter temperatures and an assumption about the upper atmosphere. The other approach uses a regression equation to relate inversion strength to surface temperature. Both of these methods suffer from a lack of verification data.
By using the full output from the UK Met. Office Global Climate Model I can obtain a complete map of the (model) inversion. A restricted set of output, surface temperature only, is then used to generate maps of the inversion using the two methods, which I compare to the model inversion. I show that both methods describe the modelled inversion strength quite well, with a standard deviation of error over the continent of about 2.5 oC, though the patterns of error between the two methods are different. In certain areas the methods break down and by using model results we can identify these regions. The methods seem particularly susceptible to errors in the regions of the Ross ice shelf and the Lambert glacier with peak errors of up to 5 oC between the "true" model inversion and that produced by the two methods. This may affect the surface wind patterns that have been deduced using the assumed inversion strengths.