The lag-correlation analysis reveals that, while the intraseasonal variability of the midlatitude HT is associated with tropospheric wave activity, that of the tropical HT is mostly associated with stratospheric wave activity, and is negatively linked to the subpolar HT. This seesaw structure in tropical-subpolar HT, which is not observed in the Southern Hemisphere winter, results from planetary scale waves in the Northern Hemisphere winter.
A budget analysis for potential temperature showed that the midlatitude HT rises only on the poleward side of the jet, and this rise is due primarily to eddy heat fluxes. The effect of eddy momentum flux on the midlatitude HT is rather minor. Furthermore, it is found that the rise of the midlatitude HT is due to the combined effect of dynamic warming below, and dynamic cooling above the climatological tropopause.
The impact of stratospheric static stability on midlatitude HT is investigated with the modified Eady model, where HT is defined as the neutral height corresponding to the short wave cut-off. It is found that the HT increases with increasing (decreasing) static stability in stratosphere (troposphere). In addition, the effect of the tropospheric vertical wind shear on HT is dependent upon the stratospheric wind shear. These results indicate that, although midlatitude HT is mainly modulated by tropospheric motion, the effect of the stratosphere is not negligible. This finding is used to interpret the observed interannual variability of HT.