Figure 6. Exponential growth of the DC current measured for leader B (UCL, upper plot) and leader A (UUL, bottom).
4.3 Leader pulses alternation and synchronism
One interesting feature that can be observed in Figures 2 and 4 is that although leaders A and B have their luminosity and current pulses increasing in time as the downward stepped leader approaches, they are not synchronized at the beginning (up to image 18 in Figure 2 and up to t = −114 µs in Figure 4).
It is also possible to observe from the high-speed video (in Supplementary Information – videos 1 and 2) that the downward leader is highly branched (Figure 1b) and the branches alternate luminosity during their downward propagation. The upward leaders A and B respond to different downward propagating branches and, as the branches alternate in propagation and intensity, so do leaders A and B accordingly. However, during the last 100 µs, the alternation ceases, all downward leader branches intensify and consequently leaders A and B synchronize and pulse together. This period coincides with the rapid exponential growth of the DC base current (Figure 6, upper plot). It seems that the proximity of the downward leader branches is such that differences in leader branches propagation and intensities are not driving the upward leaders differently anymore.
4.4 Leader charge density and charge transfer
From the current and 2D leader length measurements it was possible to estimate how the transferred charge and the linear charge density of the leaders vary with time. In Figure 7 the continuous lines show the charge transferred by the upward leaders A and B and the triangles their linear charge density during propagation. The charge and the charge density during the return stroke is not plotted for the UCL leader but its effects are seen in the charge and charge density plots of the UUL leader and will be addressed in the next section.
The charge densities of the leaders do not stay constant but increase rapidly with time during the propagation of the upward leaders (a similar result was obtained by Chen et al. 2013 for triggered lightning). This means that for both UUL and UCL the charge in the leader increases much faster than the length of the leader, which has a constant speed as shown in Figure 5. Note also that the step like character of the charge variation is due to the current pulses during the leader propagation.