Slow-streak photographic recordings of the lowest approximately 4--5 m of lightning channels in four triggered flashes are presented to document examples of microscale tortuosity variations (~10 cm or smaller geometric features). These recordings have both high spatial resolution (centimeter scale) and modest temporal resolution (~1.5 ms) such that the channel geometry is accurately rendered formost return strokes and their ensuing M-component events. The M-component events, verified as such from ancillary fast-streak data records, often show a significant enhancement of the microscale tortuosity relative to the previous return stroke channel. Channel length increases (two-dimensional) corresponding to the tortuosity enhancements are found to be as large as 55%, though the mean for 10 evaluations is 23%. In contrast, two M-component events show a marked overall reduction of the microscale tortuosity relative to an immediately preceding sequence of several discharge pulses along the same meter scale channel. Ensuing return strokes occurring along common but previously distorted channels also show a tendency for overall tortuosity reduction or ''straightening.'' These tortuosity variations appear to be unrelated to the presence of the vaporized trigger wire residue.

Three primary mechanisms of tortuosity enhancement are examined: Z pinch magnetohydrodynamic instability (MHD), discharge generated turbulence, and ambient planetary boundary layer turbulence. The Z pinch instability is assessed quantitatively and may be a contributor through either the high current phase or integrated over a continuing current period. Three variants of discharge generated turbulence also could be contributors, but a critical test of these mechanisms cannot be made with the present data set. Ambient planetary boundary layer turbulence appears unable to account for the observed tortuosity enhancements. An inductive mechanism of tortuosity reduction is also examined as a possible explanation for the disappearance of individual acute kinks or loops as well as overall channel tortuosity reductions. This mechanism cannot explain M-component related tortuosity reductions, but it might be a viable explanation for return stroke straightening given the much larger current derivatives involved. A full accounting of microscale tortuosity variations in lightning discharge channels remains to be achieved. ¿ American Geophysical Union 1995