FIGURE 1:
Single 19-sec exp. vs. 4-hr composite (384-sec exp.)
Working notes by Craig DeForest, 20-Jan-1999
(These notes led to an observing campaign and ultimately a paper that has been submitted to the Astrophysical Journal.)
Just how high do plumes go? They're traceable from the surface to about 10 solar radii (DeForest et al., 1997) above the surface. This altitude is well above altitude at which the plasma becomes collisionless, and perhaps as high as the alfvénic point in the acceleration profile of the solar wind; but numerous attempts (eg Reisenfeld et al., 1998, Poletto et al., 1996, Woo and Habbal, 1997 [ see rebuttal by Paetzold and Bird, 1998 ]) have failed to find any signature of plumes at greater, interplanetary distances.
Several ideas have been forwarded about why the plumes do not persist. Several instability models suggest turbulent mixing of the plume and interplume plasma; and timeseries analysis of Ulysses data is probably obviated by the time variability of the plumes themselves.
The plumes disappear from LASCO C-3, not because they fade into a uniform brightness contour, but because they fade into the background photon shot noise of the f-corona. Thus, there is a significant chance that the plumes could be seen more clearly with very long C-3 exposures.
Because plumes last on the order of a day, very long exposures are possible to examine them. The limiting factor is the Sun's rotation rate, which sets a timescale of about four hours before the plumes' configuration begins to change via parallax.
I tested this idea by adding the signals from several LASCO C-3 exposures. I wrote a script that examined the entire LASCO image catalog, seeking the four hours of mission time (after the start of science analysis in March 1996) with the most accumulated seconds of non-polarized, non-filtered, full-frame C-3 exposure time.
The script found a number of different days:
total exp Date of time (secs) interval Data quality 443 980503 Good 429 971107 Bad -- overcompressed. 401 980502 More good data like 980503 367 971109 355 971110 351 971108 295 971021 286 971022, 971223
I focused my analysis on 971107 (which contains a few hours with as much as 300 seconds each!) and 980503. 980503 contains very good data, a 19 second C3 exposure with non-draconian H-compression once every 12 minutes throughout the day. (This amounts to about a 4% "duty cycle" of shutter time). I ran a preliminary analysis using the first four hours of the day, and a slightly more careful analysis using a later six-hour period.
Single 19-sec exp. vs. 4-hr composite (384-sec exp.)
Figure 1 shows a single normalized, cosmic-ray-removed,
background-subtracted image from the C3 sequence. It was prepared using
the prep-c3 pipeline I use for LASCO data, and despiked with
zunspike. Next to it
is a composite image combining 20 similar frames over a four-hour period.
Note the vastly improved signal to noise ratio.
Unwrapped, radially filtered image of the northern hemisphere from Figure 1
Figure 2 is "unwrapped" version of Figure 1-A, as in DeForest et al 1997, Figure 11. The plumes are visible about 2/3 of the way out of the image, considerably farther than in previous studies. The diagonal feature tilted up to the left in the middle of the image is a halo CME that is expanding out from a disk event on the 2nd of May (the previous day). The image has been smoothed with a tall, skinny kernel, to reduce noise.
Because of the still-unsatisfying signal-to-noise ratio of Figure 2, I made another montage using 6 hours' worth of data. Because of the CME, I made the second montage from a different time (to show that the CME feature moves). 4-6 hours is the longest time that can reasonably be used, because solar rotation causes the most out-of-the-plane plumes to shift by approximately 1/3 - 1/2 of their width in that time.
In order to convince myself that the corona doesn't change much in six hours, I compared frames from two hours apart through the range 09:00 - 15:00 on that day. Figure 3 shows four coronagraph images that look similar, but on close comparison contain a myriad of small differences. Still, the bulk structures -- including plumes -- maintain their integrity over the entire sequence.
Four single frames taken two hours apart each,
showing small differences in the coronal configuration.
I unwrapped the sum image of the entire sequence represented in Figure 3. I used ZUNWRAP to
do the unwrapping, with an azimuth range of [-90,90] degrees, a
minimum radius of 5 R0 and a maximum radius of 35 R0. The pixel
averaging kernel was set to 1.5 destination-plane pixels, to reduce
aliasing. I scaled each raster of the unwrapped image linearly to
keep the average intensity of the [-45,+45] degree region constant versus
radius. The resulting image, at left of Figure 4, shows
plumes that are clearly visible at 20 R0; some of the plume structures may be
present all the way out at 30 R0. Other features are also present: the warped grid
near the top of the image is compression noise caused by the H-compression in the
images: the boundaries of individual compression cells are always in the same place,
so that the compression noise adds coherently during the summing.
Unwrapped 6-hour-average images
To reduce the compression noise and clarify the expansion structure of the plumes, I again convolved the unwrapped image with a Gaussian kernel, with a HWHM of 20 pixels in the vertical direction and 1.5 pixels in the horizontal direction. The smoothed version is at the right hand side of Figure 4. The plumes appear visible farther out, perhaps because of the increased signal - noise ratio, but also perhaps because of the vertical extent of the smoothing kernel. The diagonal stripe at upper right of the montage image is the same CME as was imaged in Figure 2; but in the intervening 10 hours it has moved out to the outer part of the corona.
DeForest, C.E, et al., 1997: Solar Physics 175 393-410.
Pätzold, M., M.K. Bird: 1998: Geophys. Res. Lett. 25, 1845.
Poletto, G, S. Parenti, G. Noci, S. Livi, S.T. Suess, A. Balogh, and D.J. McComas, 1996: Astron. Astrophys. 316, 374-383.
Reisenfeld, D. et al., 1998: Presentation to the fall mtg. of the AGU.
Woo, R, and S.R. Habbal, 1997: Geophys. Res. Lett. 24, 1159.