|Pyramidal Crystal Plate Arcs - A display imaged by Alistair Fraser (Kootenay Lake site) at Nelson, B.C. Canada March 13, 2009.. Image ©Alistair Fraser, shown with permission.
Most pyramidal crystal displays have circular halos (1,2,3). Their crystals are aerodynamically rather rounded and are poorly aligned.
This display is unusual. Its dominant features are 'plate arcs' produced by pyramidal crystals aligned like plate crystals. There is a difference - ordinary plate crystals can be within 0.5 degrees of horizontal, pyramidal plate aligned crystals are lucky if aligned within 10-20 degrees.
The features closest to the sun on each side are 18° plate arcs or parhelia. Above the sun is a 23° upper plate arc.
Farther out from each side of the sun are two more arcs that on an unsharp masked image appear slightly more distant from the sun than the 23° upper plate arc. If so, they could be 24° plate arcs but this is by no means certain.
In analysing the display it is worth noting what is not present. There are no inner 9° arcs or if an upper one is present it is only a trace. If the 23° plate arc identification is correct then there is no 22° halo.
The halos argue for crystals with some degree of plate like habit to account for their alignment. The absence (or near absence) of 9° arcs or a 22° halo indicate that the crystals had only a thin central hexagonal prism section. 9° arcs have ray paths where the entrance face is a central prism side face [path 3-26, see face numbers] and the 22 degree halo requires two side faces.
The HaloSim ray tracing simulation is for a solar elevation of 37.8°. The crystals - as shown - had a uniform distribution of tilts of ±25°. The uniform distribution was chosen to avoid the sharp V-shaped 18° parhelia generated when the crystals have only small tilts.
The fit is 'reasonable'. The upper 23° plate arc is possibly too sharp and intense. The arc results from rays entering the crystal top face and in principle increasing the height of the upper pyramidal section could weaken it. However, this is only partially effective and it introduces unwanted side effects.
Reasonable? Pyramidal crystal simulations often get 'reasonable' after only a few crystal population and tilt iterations. From then on it is hard to significantly improve them. There could be factors at work that we do not understand.