In order to investigate whether extratropical cyclones may have almost negligibly short warm fronts in some area of the northern hemisphere, we examined statistically the ratio of lengths of warm fronts to those of cold fronts accompanied by extratropical cyclones. The FGGE IIIb data was used and the period analyzed was between 1 December 1978 and 28 February 1979 and a parameter GGθ (the directional derivative of the gradient θ along its gradient) proposed by Renard and Clarke (1965) was used to draw fronts.
It was shown that warm fronts are generally shorter than cold fronts around Japan and the east coast of the North America and that warm fronts are longer than cold fronts over Europe and the west coast of the North America. This geographical difference of the lengths of fronts are closely related with developing rates of cyclones. Around Japan and the east coast of the North America a number of cyclones develop with large pressure decreases. Most of them have shorter warm fronts. On the other hand, most cyclones over the Europe and the west coast of the North America develop weakly and often decay there and tend to have longex warm fronts and shorter cold fronts. It suggests that cyclones have different structures in different areas of the globe and that they should be classified into various types other than the Bjerknes model.