As I stood there, watching the display and letting the camera snap away photos, I was wondering why the glitter that was the circumnadir arc looked so broad. Photos confirmed this impression. A comparison of long series of shots taken in fixed configuration showed that in the beginning there was briefly a normal looking circumnadir arc (the first photo in the collage above) but then it thickened by extending towards the lamp. Actually, photos reveal it quickly stretches all the way to the lamp in a manner similar to divergent 22 tangent arc.
I was flummoxed and asked Nicolas to take a look when he was visiting in Rovaniemi March next year. He told what should have been obvious: it's 46 contact arc. At the same time also some kind of subparhelic arcs (Schulthess arcs) appear, so it all fits.
But at least one question still hovers in the air: why the 46 contact arc is divergent, with not much at all the classical form to talk of? Only in the second image of the sequence there is what appears to be a classical instance of the 46 contact arc, with some shape and red inner edge. After that it turns into divergent form.
CORRECTION (13 March 2020): in the second image it is not the classical instance of 46 contact arc. It is way too high up. Anyway, the 46 contact arc hypotheses still probably holds up for the explanation of the stretching of circumnadir arc towards the lamp.