At the line where we switch between time scales, the lines tend to "jump".
This suggests that they are plotting different values, or that values are missing.
I suspect, however, that the problem may be caused through the smoothing effect. The smoothed line in the top half is using different values to the smoothed line in the bottom half - so we can't expect them to join up.
To overcome this problem, could you do an experiment with non-smoothed lines, and see if that keeps the lines joined up as the switch between the two time scale factors?
At the line where we switch between time scales, the lines tend to "jump".
This suggests that they are plotting different values, or that values are missing.
I suspect, however, that the problem may be caused through the smoothing effect. The smoothed line in the top half is using different values to the smoothed line in the bottom half - so we can't expect them to join up.
To overcome this problem, could you do an experiment with non-smoothed lines, and see if that keeps the lines joined up as the switch between the two time scale factors?