See in the middle, the seeing indices go from one end of the scale (5,4) to the other (1,1) - but only for 1 hour, before reverting to (5,4) again while the arcsec column doesn't deem to even notice.
I know there are lots of variables, but none of the other columns seem to suggest quite the tectonic disturbance in the force that meteoblue seems to be expecting.