Discussion

The observed initiation altitudes (76$\pm$6 km MSL) of sprites are roughly consistent with conventional breakdown predictions, although the measured charge moment changes appear to be insufficient for sprite initiation at these altitudes if the atmosphere were homogeneously layered. Significant variations in the conductivity may arise from electromagnetic pulse fields creased by a horizontal lightning discharge (Valdivia et al., 1998). Variations in air density may arise from gravity waves (Pasko et al., 1997a), though recent observations show no obvious connection between gravity waves and sprite locations (Sentman et al., 1999).

The plan location of the sprites could not be determined precisely, but could only be estimated based on previous measurements of sprite terminal altitudes. Thus, the altitudes of sprite features were somewhat uncertain. As was shown in Figure 2.9, the charge moment threshold is sensitive to the initiation altitude. Because of the uncertainty in altitude, one can not make a definitive statement as to the degree of inhomogeneity which is required for conventional breakdown onset.