The size of the 300 mbar north circumpolar vortex, and its partition into western and eastern hemispheres and quadrants 0 degrees 90 degrees W, 90 degrees W-180 degrees, 90 degrees E-180 degrees, and 0 degrees-90 degrees E, is estimated by planimetering the area poleward of contours in the main belt of westerlies on the polar stereographic mean-monthly analyses of the Institute of Meteorology of the Free University of Berlin. During 1963-1997 the trend in annual vortex size is a significant -1.4%/decade (vortex contraction), but with the trend in the western hemisphere portion of the vortex a significant -2.1%/decade compared to -0.7%/ decade in the eastern hemisphere portion, signifying a movement of the vortex farther into the eastern hemisphere. Thus, while on average 52% of the vortex was in the eastern hemisphere during 1963-1997, this value increased by a significant 0.4%/decade during the period, resulting in a substantial increase in vortex eccentricity. In winter and spring the vortex has moved mainly into the Asian quadrant 90 degrees E-180 degrees, whereas in summer the vortex has become elliptical with less of a decrease in size of quadrants 90 degrees E-180 degrees and 0 degrees-90 degrees W, and in autumn there has been a decrease in size of quadrant 90 degrees W-180 degrees relative to the other quadrants. The standard deviation of the annual quadrant-size anomalies has decreased significantly during 1963-1997, indicating a decrease in the annual displacements of the vortex center from its mean location as the eccentricity increases and the vortex contracts. On the basis of a 41-station northern hemisphere radiosonde network, during 1963-1997 the annual 850-300 mbar temperature increased by a significant 0.12 degrees C/decade in the northern hemisphere as a whole, and by a significant 0.17 degrees C/decade in the western half and by 0.06 degrees C/decade in the eastern half of the hemisphere. The resulting association of a 1%/decade decrease in 300 mbar vortex size, with about an 0.1 degrees C/decade increase in 850-300 mbar temperature, differs by a factor of 2 from what would be expected hydrostatically, a difference that would be eliminated by an 0.4 mbar/decade increase in midlatitude surface pressure during 1963-1997.
