Saturday 18 May 2024

High-altitude pyramid column arcs and kaleidoscope sky (6 May,China)

 Every year from March to May seems to be the pyramid column arcs' season for China. In the 2024 season, the most eye-catching display occured on 6 May, ChangSha, HuNan Province. At altitude of about 71.3°, the Pyramid column arcs, PC and Wegener crossed each other, bringing a spectacular picture of the sky.

(张晖)ZHANG Hui's stack image through a stack of 112 photos spanning 10 minutes. BGR and USM processes make arcs clearer.

Below are the results of a preliminary simulation using ZHANG Jia Jie 's program.

Well, now we are already looking forward to the second half of the year. Unlike March to May, June to August seems to be the peak time for pyramid plate arcs and 28° arcs for China. Please wish us luck.

JI Yun

Wednesday 6 March 2024

Two below-sun Tape arcs in German Alps, 9 January 2024

by Lasse Nurminen:
One of the great things in the modern day observing is the possibility to explore the web for observations around the world. It gives just so much more to this hobby.
Once again I was browsing a German website´s forum ( about halos. There are several threads with links to halo displays visible in a network of cameras widely spread over the mountainous area in Central Europe ( In a post from Ruben Jakob ( one of the links at first looked like only a well formed infralateral arc, but something was disturbing me. Just as if there was a faint arc next to it, what could it be?

After a little thinking I thought it must be something related with the lower Tape arc/Parry infralateral. I took a couple of screenshots of the changing view with the available 10 minute steps and run them through a little Photoshop process to make the faint halos better visible. After that I had no other ideas but to let Mr. Riikonen to take a look at the situation in hands.

by Marko Riikonen:
Nurminen mailed me some photos of this 9 January 2024 display in Balderschwang, Germany. His hunch of something going on there was right. As shown by the title photo, two below-sun Tape arcs – one tangent to 46° infralateral arc and one to supralateral arc – are visible. The former is well established: although still a high rarity for any halo hunter to bag, over the years a representative collection of specimens have accumulated from ice fog displays. The latter is a much more furtive animal. It's been photographed in a couple of spotlight displays, but the only previous celestial light source record is the display that Herman Scheer photographed on 18 October 2010 in Sonnblick Observatory, Austria ( However, the 9.5 degrees sun elevation meant the two arcs were superposed, contributing about equally to the joint arc's intensity. And given in spotlight beam we have photos of only either lower arc, here is what raises the Balderschwang display on pedestal: it is the first record in which both arcs are visible individually at the same time.

The webcam took a photo every 10 minutes. The display started at 14:30 (title photo, sun 15.1°) and was seen until 16:10 (sun 4.4°) after which the sun got too low to shine anymore to the ice fog. At the end of the post the full sequence is shown. The most interesting stuff was in the first hour and above are three stages from this interval with (a bit wider angle) simulations using HaloPoint: 14:40 when the two arcs are distinct and nicely separated (arrows); 15:20 when they are already quite merged; and 15:30 when a practically perfect superposition is seen.
There are 8 arcs to this family, but normally we regard as Tape arcs only those that are solely made by Parry orientation. This reduces it to four arcs – two obliquely from the sun on 46° infralateral arc, two on supralateral arc – into which number also belong the two arcs in Balderschwang photos. As to the naming, the situation is unclear, at least on the level of individual arcs. Tape himself has used terms Parry infralateral and Parry supralateral (also some usages of 46° Parry by others can be found in the internet). But it appears that Tape only associated two of the four arcs with these names. If we look into the monolithic A General Setting for Halo Theory (1999) by Tape and Können, we see that while the other arc in the Balderschwang display would be called the Parry infralateral, the other one is "unnamed" (page 1581, left column, last paragraph). And given there have been really no serious suggestions stemming from the generic "Tape arc", the task of naming this quartet members appears to be still before us. As the arcs divide into above and below light source pairs, the names could play out like this:

upper Tape/Parry supralateral (arc)
upper Tape/Parry infralateral (arc)
lower Tape/Parry supralateral (arc)
lower Tape/Parry infralateral (arc)

However, because supra and infra are synonymous for upper and lower, there is a bit of the feel of those "logical monstrosities" that Tape writes about in the Nomenclature Woes chapter of his book Streetlight Halos (second part, page 170). Moreover, naming of any halos from double oriented crystals should be such that it takes into account also the corresponding arcs that arise from raypaths involving a reflection from horizontal crystal face and that's where the above suggestion again performs poorly by extending what are already quite long-winded names. What's the solution? Alphabet letters would seem to be one:

Tape A81✔️+33...-27

counter-Tape A'861
Tape B71✔️-33...+27
counter-Tape B'731

counter-Tape A''831

counter-Tape B''761

Tape C15✔️+63...-3

counter-Tape C'164
Tape D14✔️-63...+3
counter-Tape D'135

counter-Tape C''134

counter-Tape D''165

Tape E31

counter-Tape E361

Tape F61
counter-Tape F631

Tape G16

counter-Tape G163

Tape H13
counter-Tape H136

The arcs made exclusively by Parry crystals run from A to D. The two arcs in the Balderschwang display come out as Tape C and Tape B. Arcs E to H are better known as made by plate oriented crystals and include circumzenith arc and circumhorizon arc, but should these occur solely from Parry crystals I agree with Alexander Haussmann that they be recognized also as Tape arcs (Haussmann points this out, as pertaining to circumzenith arc, in the comments section of my post Perhaps surprisingly, these specific situations appear still undocumented.
The table is just a sample from one of alternative nomenclature suggestions for halos from double oriented crystals that I have been working on. While the lettered versions seem to me the best way to go with Tape arcs, the above sample is not necessarily my favourite. For example, letters could be assigned to each arc differently, such as running from the most common arc to the least common arc. This kind of ordering may not be so clear-cut for all halo families, though. If we would apply it to the Tape arcs in the above table, swapping letters B and C would probably get it right. I hope to get the nomenclature article done with and posted here some day.
None of the reflection arcs in the right side of the table corresponding to Tape A-D have been observed – yet. They don't come out well with regular hexagons. Triangular(ish) shape, which is common in ice fogs, is better as demonstrated by the HaloPoint simulations below, arrows marking the reflection arcs. These tentatively named "counter" arcs may get overwhelmed when they occur on infra- and supralateral arcs, but the good thing about snow gun generated ice fogs is that every now and then Parry orientation dominates: there are displays in which Tape arcs are seen without infra- and supralateral arcs (example:
Of the two upper oblique Tape arcs we are all well familiar with the one that appears on supralateral arc, Tape A in the language of the table above. It is "the" Tape arc that accounts for the great majority of observations of the four arcs. The upper arc that is associated with infralateral arc, Tape D, has been observed only in spotlight displays. There is a reason for this: its light elevation range tops at 3 degrees above the horizon, rendering it improbably hard to see in solar and lunar displays. We have some representative photos of this arc (example: but no particularly good co-occurrences with arc A. Looks like there are only two such displays. First was the halo complex that Marko Mikkilä, Jarmo Moilanen and I watched in Rovaniemi on the night of 23/24 November 2015 ( The second is this winter's crop, a display that I photographed on the night of 10/11 November 2023 in Rovaniemi, shown below as two versions of a 11x15s stack (arrows mark Tape D) and a simulation. The two upper arcs in these displays are the corresponding arcs to the lower arcs in Balderschwang display, but in swapped positions due to light elevation difference. By the standard of Balderschwang, lightsaber wielders still have their work cut out to capture a better defined instance of the upper pair.


Saturday 30 September 2023

An attempt at high cloud 44° parhelion

High cloud 44° parhelion is still waiting for its capturer. A chance to try for one came on September 28 this year in Kontiolahti, Finland, when a cirrus floe containing a short-lived blinding parhelion moved out towards the appropriate position.

I had already camera snapping photos at 10s interval, but changed it to 4 seconds when I estimated the leading edge of the floe had reached about 44 degree distance from the sun. The b-r image above is 84 frame stack from the time it took for the floe to pass the 44-46 degree azimuth and shows indeed a dark spot indicative of red color. But it looks to be too far out for 44° parhelion when compared to the 46° halo in the image. Probably we are looking here just at a segment of 46° halo or 46° lateral arcs crossing, or both, but proper measurements using starfield may still be worth doing. 

Even if the result seems negative, it thought to share this to pay attention to situations where high cloud 44° parhelion might be a prospect. Below are other relevant images and some more info.

Bgr and usm versions.

At 13 degree sun elevation 44° parhelion is inside 46° halo, which is not the case with the spot in the observed display. The other half of the simulation has column and Parry population as the display had halos also from these crystals. Made with HaloPoint.

The short stage during which bright parhelion appeared in a cirrus floe. Photo interval is 10s.

Here is shown the first (top) and last (bottom) photo of the 84 frame stack (middle) from the period that it took for the floe (delineated) to pass 44-46 degree mark.

Uncropped versions of the 84 frame stack, br and usm. Notice the lacking of uppervex Parry. With regular hexagons at 13 degree sun it should be of equal brightness with the uppercave. In stacking, when sun is tracked at this solar elevation range, uppervex gets unfocused and uppercave focused relative to each other due to their sun elevation dependent movement which are in the opposite directions. But as the stack has only 0.5 degree sun elevation change (from 13.3 to 12.8 degrees) the effect should be negligible. Thus the absence of uppervex (or almost absence, there seems to be a ghostly suggestion of it in the image) tells of tabular crystals. A simulation with 0.7 1 1 0.7 1 1 shape modified crystal in HaloPoint reproduces the situation and this is what has been used in the simulation above (0.2 uniform deviation for the two faces). If somebody knows other similar Parry displays showing evidence of tabulars, I am happy to hear. I think I knew one, but have forgotten about it.

In Germany they had this display much better:

Marko Riikonen

Friday 28 July 2023

Simulation-like 18° plate arcs and puzzling 22° arcs from airplane, China

On May 24, 2023, I witnessed a very high-quality case of 18° plate arcs while onboard a flight over Southwestern China. Due to my busy work schedule, it took me nearly two months to review and analyze the batch of photos I captured. Clearly, there were more surprises contained in them.

The 18° plate arcs are common in China (in the southern region during the summer), but high-quality occurrences are rare worldwide. In recent years, Chinese sky enthusiasts have explored and contributed numerous images of 18° halos. Some were quite bright, but without exception, they were all quite blurry, likely due to wobbly crystals and diffraction. Blurriness of these arcs might be a common feature since, in previous cases, they are almost always accompanied by the 23° plate arc, indicating the presence of thicker ice crystals, which may have difficulty maintaining low tilts in crystal orientations.

Therefore, when I witnessed this pair of remarkably clear V-shaped 18° plate arcs, I was surprised that such small tilts even exist under natural condition. The arcs were amazingly bright, outshining all other halos minus the sun pillar. In fact, I only noticed the 18° plate arcs and the sun pillar with my naked eyes. Other halos in the display were all picked up later photographically.

(The display lasted for less than a minute, during which I managed to capture six photos. In hindsight, I regret not taking more photos by tolerating the glare and adjusting the shooting angle less frequently to avoid reflections from the aircraft window. However, with the limited six photos I have, I was able to align and stack them to obtain a more complete halo phenomenon. Taking more photos might have further enhanced the final result.)

6-frame stack. Solar altitutde 9.8°.

Now let's move on to the other halos in the display. The lower 20° plate arc and the lower right 24° plate arc are clearly present in some of the photos. The 9° plate arc appears to be absent. In simulations, one can invoke nearly triangular ice crystals to weaken the arc till it gets overwhelmed by the bright sun pillar. However, it's also possible that the absence of the 9° plate arc is due to uneven distribution of ice crystals since the lower left 24° plate arc never showed up. The 35° plate arcs appear strong in simulations, but they were unfortunately obstructed by the wing and engines during the display.

What caught my attention even more are the arcs at the 22° positions. Initially, I thought the pair of tilted arcs outside the 18° plate arcs were 22° parhelia, distorted due to wide-angle lens effects. However, simulations show that the 22° parhelia should be closer to the sun and shouldn't tilt to such an extent.

So, what are these tilted arcs in the photos? They are neither regular crystal halos nor something generated by (3 0 -3 2) exotic crystals or cubic ice crystals. Their presence perplexes me.

B-R, BGR and other post processing work indicate these tilted arcs are actually the brightest spots on a pair of much longer, vertical, slightly sunvex arcs, resembling the appearance of the Lower Schulthess Arcs, also known as Lower Reflected Lowitz Arcs or Subparhelic Arcs.

BGR(background removed) version. 

B-R processed version. The vertical extent of the arcs becomes clear. Segments of 18°, 20° and 24° halos can also be seen.

The behaviour of Subparhelic Arcs/Reflected Lowitz Arcs/Schulthess Arcs has been a long-standing issue. It seems challenging to explain why they mostly appear as single or double arcs, rarely as triple arcs as shown in simulations. But in this case, there is an additional puzzle: why do these arcs have significantly enhanced bright spots near the 22° parhelia position?

I don't want to explain all the discrepancies with simulations as "uneven clouds" or other coincidental factors. During the one-minute long display, it remained remarkably stable - the Lower Schulthess Arcs/Lower Reflected Lowitz Arcs/Subparhelic Arcs were always present, and the aforementioned 22° bright spots were also consistently there, even though the clouds had undergone some changes.

I wish to hear the opinions of fellow researchers worldwide on this matter. I am hesitant to claim it as a puzzle, but I am not exactly sure what mechanisms created these arcs. After initial attempts at simulation, I found that Lowitz raypaths 8-2-1-6 and its multi-reflection derivatives such as 8-2-1-2-1-6 etc. produce a pair of arcs that resembles what I saw. If we artificially disallow conventional Lowitz raypaths and allow only the above mentioned multi-reflection paths, then Lowitz-oriented plates with moderate tilt such as 15° produce a relatively ok match. (Compared to the odd radii in the photos, the match is much poorer. Lowitz arcs' mismatch with simulations has been a long-standing issue and it seems this case is no exception).

Simulation showing only the Lowitz raypaths 8-2-1-6 and its multi-reflection derivatives such as 8-2-1-2-1-6 etc.

Comparison between the simulation and the real scene. Not the best match, but close.

To explain why only multi-reflection raypaths are present, it might be speculated that the ice crystals are extremely thin. Nicolas Lefaudeux's '100 hit' theory inspired me. Perhaps certain extremely thin plates allow for multiple internal reflections between basal faces, creating the observed halo while excluding the usual Lowitz arcs.

However, explaining why only Lower Schulthess Arcs/Lower Reflected Lowitz Arcs are present, and not the upper and middle ones, is much more difficult. Perhaps a specific mechanism exists that gets the Lowitz oriented crystals azimuthally locked so that its face 6 faces the observer, while still allowing the main axis to azimuthally rotate within a limited range of angles, thus selectively producing these two arcs? (Fully locking the main axis cannot produce satisfactory arcs. I have tested this). If such a thing is possible, this pair of arcs might deserve a specific name to distinguish them from Schulthess Arcs/Reflected Lowitz Arcs/Subparhelic Arcs. However, I still feel this is a far fetched explanation since there is no available literature indicating the existence of such a 'alternative' Lowitz orientation.

As I haven't obtained the posting access to HaloVault, my colleague JIA Hao has kindly helped me translate and post these observations (originally in Chinese). Also special thanks to ZHANG Jiajie for having multiple fruitful discussions with me on this topic. We look forward to more excellent explanations from the global community. 

JI Yun

Thursday 18 May 2023

Digging up old odd radius display for 28d spotting

Hello everyone,

after several talks with Alec, I decided to make some digging into my old odd radius halos displays to search for rare halos and find out some more 28d halo (and 13d halos too).

Here are some examples of what I found so far that might be interesting to submit here to your assessment. I hope you will find those cases interesting.

For some of them, I still have all the raw files of the sequence, for others, I kept few raw images, but for all of them (exposed here at least) I made various time-lapses to keep sequence in B-R rendering, colour rendering, stacked with 4 or 8 images each, like the one I presented earlier in this blog.

I will start with a sharp display of odd radius circular halos I got last year, in April, at the morning.
I'm not usually an early bird so I can miss some nice displays each year. For this one I got the chance to have the display still available to start a capture from my roof window. The center of my house is a stair tower which offer a nice blocking roof for the spot I am from. Therefore, with the shaprness of the display, the 9° ring was particularly well visible on B-R rendering before the sun comes out of the roof (because the diffusion of the light on the lens add noise up to the 9° ring area). There is a first image, B-R rendered from a sole raw image, to give a view of the sharpness of the event:
Stacking with the above one as first image, with the 25 following images makes those odd radius halos more clear: Well, in the light of recent search for 28+° ring, this was one of my first tries even if I was expecting, like for the 13° halo, to find it when the rings are not sharp. But if you don't try, you won't find anything.
So there is a processing the the previous stack to enhance sharp halo first then eventually larger borders ones if any.

There is a folder on my drive, so you can take a look at the time-lapses of the day. As usual, better download it rather than viewing it as a Youtube video, to have a better video compression (as is the original one) Many years ago, I got a nice odd radius display at sunset, during a day of November 2014. The view is not ideal, as I placed my camera on the edge of a window, look south, with a wide angle rectangular lens, hence the distortion. (EoS 1200D + Sigma 8-16mm, set at 8mm). I was so amazed by the display I sent it to Nicolas Lefaudeux to have his expertise on the matter. And then he told me I caught the 28° halo, again (but I don't recall why this 'again') ;-) There is the display, B-R rendered, and unsharp masked.

Unfortunatly, I did not keep any raw file of this event.
There is a folder where are all the remainings.

And any old how, here are some I cannot say it is, but looks like there might be something there.... or not: And while browsing at all of those, I realized I got at least a dozen of 13° halos, with at least one with the Moon. Which could make a post after this one, for the record.

And a last one, not for the 28d spotting, but only because I like it a lot:
Now, a couple hours after starting this post, I think I may go get some rest ;-).

Greetings from France.
Nicolas R.

Monday 1 May 2023

Time Machine: High Cloud Hastings Arc in China, 2012.02.22

The Hastings arc is among the rarest of all halos, even more so in high clouds. On Feb 22, 2012, XU Guodong was blessed with an outstanding display in Mohe, and became possibly the world's very first person to photograph a high cloud Hastings arc.

The display started in the morning and lasted for at least two hours till noon. XU happened to be on a road trip thus unable to document the event continuously from a fixed location. Most of his photos were taken at two stages of the display, when the sun was at around 15° and 22° respectively.

During the first stage, the typical sun-side Parry elements such as the Helic arc and Tape arcs were not particularly strong. However, the Hastings arc was fairly prominent and very easy to distinguish in unprocessed images. Had XU been equipped with the necessary halo knowledge, he would've recognize the arc at the scene with naked eyes.

Early stage of the display when the sun was at around 15°. 4-frame mosaic. Slightly enhanced. See if you can spot the Hastings arc.

Background subtracted version of the mosaic. The Hastings arc stands out nicely.

The original image to illustrate how prominent the Hastings was.

The display remained strong as the sun rose to 22° elevation, when XU made the second stop of his road trip. At this stage the Tape arcs improved a bit while the Helic arc disappeared. The Hastings arc, together with the Wegner arc, somewhat weakened but was still easily discernible in unprocessed images.

2-image mosaic. Slightly enhanced. Sun at around 22°.

Background subtracted version. Though weaker than during the first stage, the Hastings extends further towards the sun.

While the sky around the sun was jam-packed with great stuff, the opposite side was also very busy. The highlight absolutely goes to the loop-shaped Tricker arc. XU was very impressed by how the loop gradually shrinks in size as the sun rises.

Tricker arc (probably some Greenler and Trankle too) during XU's first stop. Sun at 15°.

Tricker arc during XU's second stop. 3-image mosaic. Sun at 22°. Note how the faint Subhelic arc goes above the 120° parhelia and then touches the upper end of the Tricker loop. The blue spot is quite strong too.

This display bears great significance in China's modern halo history. 10 years have passed and it still is unchallenged, and will likely remain so for a very long time.

Jia Hao