Tutorial: reading the solar plots¶

Goal: lay the Stonyhurst and Carrington heliographic grids over a stack, find the position of a feature on the Sun from them, see where the sunspot table of the same report comes from, and read the diagnostics of a time-lapse (registration, brightness, derotation).

You need: a Helia that has helia stack --report (see the CHANGELOG), and for the grid a capture whose time is known: a SER (its camera time stamps are used), or any other capture with --date. The grid pictures are from a real SER, 15_20_57.ser (H-alpha, 1920 by 1080, 100 frames, 2022-07-17); the time-lapse pictures are from the real pair of the time-lapse tutorial. The positions are computed as in Sunspots, groups and the Wolf number and Solar derotation.

1. Stack with the report on¶

helia stack 15_20_57.ser -o ha.fits --preset ha --keep 20% --autocrop --report ha-report

The report notes the time it used. For a capture with no time of its own (an AVI, a video) give it: --date 2022-07-17T13:42:57 is the UTC time, YYYY-MM-DDThh:mm:ss. A file named by the camera program's local clock needs the offset taken off: this owner's files are in British Summer Time, so 15_32_09 is 14:32:09 UTC. Without any time the report says so and leaves the grids out:

report:  not drawn: grid-stonyhurst: the time of the capture is unknown, so P, B0 and L0 cannot be found (a SER with timestamps, or a FITS with DATE-OBS, gives it)

New files: grid-stonyhurst.png, grid-carrington.png, heliographic-grid.csv, solar.json and solar-reference.png (the stack the grids are drawn on).

2. P, B0 and L0¶

The title of each figure gives the three angles for the middle of the capture, here 2022-07-17 14:21:03.657 UTC (the SER's own time stamps), from the series of Meeus, Astronomical Algorithms, chapter 29 (good to about 0.01 degree):

• P = +4.73 degrees: the position angle of the Sun's rotation axis, east of celestial north. The pole is tilted that far from the celestial north direction.
• B0 = +4.55 degrees: the heliographic latitude of the middle of the disc, which is how far the north pole is tipped towards us. At +4.55 the north pole is just visible and the south pole hidden, and the equator curves slightly below the middle.
• L0 = 54.19 degrees: the Carrington longitude of the middle of the disc.

3. The Stonyhurst grid¶

Parallels and meridians every 15 degrees, measured from the central meridian (CM, the orange vertical line): 15W, 30W to the west of it and 15E to the east, 15N and 30S for latitude. The orange horizontal line is the equator (EQ).

1. Find a feature. The small dark feature just above and left of the centre, near pixel (468, 252), lies between the 30N and 45N parallels and a little east of the central meridian: the same arithmetic as the grid's gives 37 degrees north and 4 degrees east of the central meridian. Read positions from the lines, or compute them in Python from solar.json (the disc geometry and the three angles).
2. Solar north and east are the arrows in the corner. North is along the axis; east is to the left of the picture when north is up, as on the sky. Here the arrows are plain: the picture's orientation was not given, so solar north is assumed to be up. If your image has celestial north up (an equatorial mount that does not flip), add --celestial-north-up: the grid is then turned by P, 4.73 degrees. If it is rotated by other means (a diagonal, a camera turned), --north-angle DEG turns it; if it is a mirror image (a diagonal without a flip), --mirror. These are the same options as counting sunspots; a wrong orientation turns the whole grid, so check it against a feature you know, such as a prominence on the limb.
3. The grid ends at the limb: only the visible hemisphere is drawn, and the labels of meridians within 14 degrees of the limb are left out because they crowd.

4. The Carrington grid¶

The same grid labelled in Carrington longitude, a system fixed to the rotating Sun (it turns once in 27.2753 days, the Carrington rotation). The meridians are the multiples of 15 degrees: 15, 30, 45, 60, ... The orange one is the multiple nearest the middle of the disc: 60 here, which is 5.8 degrees more than L0, so it lies 5.8 degrees west of the central meridian (Carrington longitude grows towards the west).

1. A spot keeps its Carrington longitude from one day to the next, while the disc turns about 13.2 degrees a day under the central meridian, so L0 falls by that much a day. A group at longitude 40 reaches the central meridian when L0 has fallen to 40, in (54.19 - 40) / 13.2 = 1.1 days; one at longitude 75 crossed it 1.6 days ago and is now in the west.
2. Compare captures. Two stacks of the same region taken days apart show it at different places on the Stonyhurst grid and at the same place on the Carrington one.
3. helia sunspots gives each spot's lat, CMD (longitude from the central meridian) and Carrington lon, the same numbers you read here.

5. The sunspot overlay and table¶

With --sunspots (a white-light capture with the whole disc in view) the report also carries what helia stack --sunspots writes beside the output, integrated: sunspots.png (the flattened disc with each spot's penumbra and umbra outlined and the groups numbered), sunspots.csv and sunspots.json (one row per spot), and a table of the groups with their latitude, longitude from the central meridian, Carrington longitude and corrected area in millionths of the hemisphere. The grids use the same time and orientation as the detection. A real overlay, from the 2025-09-17 white-light capture of the sunspots tutorial:

(This report was not made from the 2022 white-light capture: it is the only photospheric capture in the set, and it is dusty and cut by the frame, so the detection finds dust. The sunspots tutorial explains the numbers.)

6. Time-lapse diagnostics¶

helia timelapse takes --report DIR too. The pair of the time-lapse tutorial, stacked and registered with the brightness evened out and the Sun derotated to the later frame:

helia timelapse 15_32_09.avi 15_37_19.avi -o tl --date 2026-10-20 --utc-offset 1 \
    --deflicker match --derotate --movie none --report tl-report -- --preset ha

Two frames is the smallest series, so each curve is a line between two points; with a long series they are the curves you watch. The three figures:

Registration. Top: the shift each frame needed to land on the reference (the second frame): frame 1 had drifted (17.3, -5.9) pixels, the telescope's drift in five minutes. A slope across many frames is drift; a jump is a pointing change; a red point is a frame that could not be registered. Bottom: the confidence of each frame's limb fit (green, 0 to 1) and the peak-to-sidelobe ratio of the phase correlation that refined the centre (purple, right axis; 183 here: a sharp lock). A frame with both low is the one to look at.

Brightness. The mean level of the inner disc as measured (orange) and after the gain that --deflicker match applied (blue): the two frames differ by 4.2% (the standard deviation over the mean) and 0% after. A series whose brightness falls steadily is the Sun sinking in the sky or a thickening haze; steps are cloud or a camera gain change.

Derotation. How far the Sun's differential rotation (Howard, here) moved the surface to bring each frame to the common time: the largest displacement over the disc. Five minutes is 0.18 pixels at this scale, so the plot says that for a pair this close, derotation is not worth doing; it grows linearly with the time from the target, and over hours (about 13 degrees of longitude a day at the equator) it is several pixels, which is when it matters. The numbers are in timelapse-frames.csv (derotated_px).

7. In the app, through the API, redrawing¶

In the app, the grids and (with the Find chip of Sunspots) the sunspot table are in the Science section of the Stack step, drawn from the capture's time stamps (a capture with none has no grids); the time-lapse step has Save diagnostics next to Save time-lapse. Through the API, "report": true and, for a capture with no time of its own, "date": "2022-07-17T13:42:57" in a job's config serve the grids; "report": true in a POST /timelapse writes the diagnostics into OUTPUT/report/ and returns report, the page's path. Redraw one at any size:

helia plot grid-stonyhurst ha-report -o grid-big.png --size 1600x1200
helia plot timelapse-registration tl-report

When it goes wrong¶

• grid-stonyhurst: the time of the capture is unknown. Give --date (UTC), or use a SER with time stamps, or a FITS with DATE-OBS.
• The grid is rotated against the Sun. The orientation was not given: see --celestial-north-up, --north-angle and --mirror above.
• grid-*: no solar data. There is no fitted disc in the stack (a crop that cuts the limb, or a capture of something else).
• timelapse-derotation: no derotation ran. Add --derotate (it needs the time of every frame; --date and --utc-offset give it for AVIs).