A comparison of some methods for reduction of pressure to sea level over Australia

Four methods for reduction of pressure from station level to sea level were quantitatively compared against the current Australian method, using the criteria of (a) minimising fictitious horizontal gradients of sea-level pressure arising as an artifact of the reduction method and, (b) maintaining the spatial continuity of the sea-level pressure analysis. The comparisons were made over four periods of about 15 days in different months, at four synoptic times per day. The two methods that performed best, relative to the current method, shared the characteristic of using the temperature at a standard level above the earth’s surface, rather than a surface (screen) temperature, to estimate the mean virtual temperature necessary for reduction to sea level. None of the methods studied was clearly superior in all circumstances. The methods that used a temperature at or close to the time of observation for pressure reduction tended to perform relatively better in the more mountainous areas in the vicinity of the Great Dividing Range, while the methods that used a monthly mean temperature tended to perform relatively better in the more plateau-like areas of inland Western Australia and the Northern Territory. The two best-performed methods were slightly modified versions of methods suggested by Gibbs in 1952, and by Benjamin and Miller in 1990. Both deserve consideration as replacements for the current Australian method.