HARP Quicklook Movie Format

Michael Turmon, JPL, April 2013

This brief note explains the format of HARP quicklook movies, which are derived from the hmi.Mharp_720s data series. The movies can give you a quick overview of a HARP's extent. They are made monthly.

The wiki entry has more on the data series itself.


Overview of Movie Layout

Here is a still image from a rather complex HARP movie (click to enlarge).

HARP Annotations

The movies have 1024x1120 pixel frames; they have been reduced 4x in each direction from the 4096x4096 originals by simple downsampling. The disk is gray, and pixels labeled as "active" within the gray disk are black.

The colored blobs, which we call patches when we're referring to just a single image, represent the content of each HARP in that image. (Clumps of black pixels outside the HARPs are magnetically active, but not concentrated enough to combine into a HARP.) The color of the patch is related to its HARP ID; the key at right helps to map colors to numerical ID. By numerical coincidence, two HARPs can share the same color, as with 1996 and 1979 here.

The white bounding box surrounding the patch encircles all within-HARP pixels, so you can tell HARPs apart even if they happen to have the same color. The HARP ID is in the corner of the bounding box, and the list of all HARP IDs is in the legend on the right side.

Note, the HARP is defined to be the content of the colored patches within the bounding box, not the bounding box itself.

NOAA ARs are shown in yellow: listed, in order from left to right, in the text overlay, and shown in proper location on the disk as a yellow cross, with accompanying number moved along the equator to avoid overplotting. (The offset from the equator is proportional to latitude, for the same reason.)

The upper left corner has T_REC, in TAI units, the lower left corner has HARP counts within the image, and the right corners hold active region lists, separated into northern and southern hemispheres.

Relation to Data Series Contents

The movie directly reflects various keywords within hmi.Mharp_720s (keywords are written LIKETHIS). Of course, T_REC and HARPNUM are on the movie. The colored patch corresponds directly to the bitmap data segment of the HARP (i.e., the colored region consists of those pixels within the bitmap segment of value 32 or greater).

The corner of the white bounding box is at location CRPIX1, CRPIX2 in the original 720s imagery (e.g., hmi.M_720s), and the white bounding box is CRSIZE1 by CRSIZE2 pixels. Finally, the latitude/longitude bounding box of the HARP (LATDTMIN, LATDTMAX, LONDTMIN, LONDTMAX) is tightly enclosed by the white pixel bounding box here.

As noted, the time of the "region created" event corresponds to T_FRST1, and the last solid-border HARP outline (the last observation of the HARP) corresponds to T_LAST1. The first and last instances of padding (dotted boxes) correspond to T_FRST and T_LAST. Placeholder and "use past" regions are signalled by dotted outlines, and H_FAINT = 1; placeholders additionally have a missing value for NPIX, the on-HARP pixel count. Similarly, times where merges were done are shown by the + character in the movies, and H_MERGE = 1 in keywords.

Details Regarding the Image Overlays

Below is the same image, with further annotations.

HARP Annotations, Detail

Regions that are new in this frame are shown with a white square hanging outside the corner of the white bounding box (HARP 2014 is new in this frame), and an ! character in the right-hand legend. The time of this event corresponds to T_FRST1 in the keyword list for hmi.Mharp_720s. A one-day padding period extends before and after the time interval where the HARP is observed. Padding is indicated via a dotted white bounding box, and by a ( or ) character in the legend.

The dotted white bounding box can occur for other reasons. A HARP that is lost and later re-appears has a gap (usually just a handful of frames) which is filled by "placeholder" frames. These frames are part of the HARP, but have empty bitmap segments. HARP 2008 in the far south of this frame is a placeholder. This is indicated by ? in the legend.

HARP 1974 illustrates an intermediate case, in which the HARP can only be detected because we expect to see it based on prior images. This is indicated by solid-horizontal, dotted-vertical lines, and a ~ character in the legend. In this case, the bitmap will not be empty, because on-HARP pixels were found.

Sometimes, as in the southern blue patch here (HARP 1990), the white bounding box encloses several unconnected areas separated by quiet Sun. This happens when, at a later time, additional flux emerged, and these areas were merged. A small white square will appear inside the upper-right corner of the white bounding box on the frame where the merge happened (this is usually not of interest).

Text Insets

All four corners of the movie have text insets.

The NOAA AR numbers are listed, in left-to-right order (not numerical). (Note, the ordering is not by longitude, but by column in the image, so your eye can easily tell the ordering.) Each number corresponds to a yellow cross and a tiny text label on the image overlay.

More on the NOAA AR overlays

The positions of NOAA ARs are shown as yellow crosses on the disk. These positions are only supplied once per day (nominally at 24:00 UT), so for other times, positions are linearly interpolated in latitude/longitude coordinates. Occasionally, the positions change too fast to be consistent with solar rotation, or exhibit other motions; this is present in the original data. Where correction is possible, the local copy of the NOAA database is often corrected.

NOAA ARs can appear to hover near the limb because only the visible content is taken in to account when computing the center. Also, there is a pause in motion of six hours that occurs before the start and after the end of the lifetime of each NOAA region, due to the lack of any data beyond the endpoints to use in interpolation. That is, if the region first appeared in the NOAA catalog at 24:00 UT on February 14, it would appear as a stationary yellow cross between 18:00 UT and 24:00 UT, after which it would drift westward as expected. This can be helpful in spotting new NOAA ARs.