The
2002 |
Less
than 18 months since the June Solstice
Eclipse, Africa fell
beneath the very next path of totality.
It would have been
wonderful to see it. But back in early 2001, when I was
planning to get to one or the other of these consecutive
totalities, the smart choice was definitely for June 21st of
that year. It was an eclipse with favorable conditions in
nearly every way. The weather conditions were to be much,
much better, the solar cycle was near its maximum (which
creates the most spectacular and brilliant coronae and multi
prominences) and duration was considerably longer. I
couldn't afford both, and tried to choose wisely. I'm still
comfortable with that choice, for this was a singularly
beautiful eclipse trip (read
my report HERE),
with about the least stressful observing conditions I've yet
experienced (there were essentially no weather problems at
all). Also it was one marvelous corona! I'll not soon forget
the lovely people and places to see in Zambia. Altogether a
good choice.
Later that day I received the best
news of all from my collaborator of several eclipses,
Jonathan Kern (we'll speak a lot more about him below).
Jon's very web-savvy, too, and found a way to telnet big
files from his site in Africa. That afternoon I'd gotten a
complete sequence from his new experimental digital camera
setup. How much fun to have such high quality images
directly from the field so soon after the event! His wife,
Jan, had taken the sequence, you'll see their equipment next
below. |
I'm
very lucky to have
Jonathan Kern as a friend and a most knowledgeable fellow
coronophile. This description is really the Kern's report,
with me the "tag-along for post production," and also
greatly in spirit. (That's
another reason I took my time adding this new eclipse page
to our site -- to give them first dibs on the
attention.) I've described
Jon's amazingly fine work quite a few times on these pages
before (for example, HERE,
and HERE,
and more recently, HERE),
as we've collaborated on many of the better coronal images
you will find here. I've known him since 1972, when we
corresponded about the use of so-called "radial gradient
neutral density filters" for total eclipse reconnaissance.
The idea of grading one's camera equipment, to greatly
reduce the light close to the sun, then grade the added
density outwards radially to full transmission in the outer
portions isn't a new concept. I've encountered reports as
far back as the '50s and early '60s, by serious
investigators who were willing to pioneer new techniques in
coronal imaging. Some used rotating masks which shortened
the total exposure near the limb. Others tried stationary
out of focus obstructions and the best "rad-grad"
filters.
Jan is seen here reading the local newspaper eclipse reports full of observing tips telling people how to enjoy the astonishing experience with minimum risk to their eyes. It's a part of the fun to bring back some of these published stories from places along the path. Sometimes near hysteria is evident -- the magazines and papers essentially warn people "not to look -- stay indoors and watch it on television!" Stoo-pid, stoo-pid, although completely safe (take a nap, even safer), and many naive people take it to heart, missing nature's greatest spectacle. Better is to explain where the factual dangers lie, give advice for pinhole projection, or where to get inexpensive protective solar eclipse viewers, and invite everyone to observe with the naked eye during those precious seconds and minutes of totality. That is, if the day is clear. Unfortunately, this eclipse occurred during rainy weather in the region. While above there are sunny bright skies the day before the eclipse, here's the way things looked on eclipse day.
Those who have chased several eclipses will empathize greatly (droll understatement) with the "hole in the chest" or sinking feeling as you face the realities above. What was going on above Messina that morning? I thought Jon and Jan would be pleased to see what the clouds had been like throughout the region. There was no single good image I could find online, but working from five other shots, I could composite, draw in the country borders, add Messina's location, remove perspective distortions, and tweak the details, come up with a jpeg to send them right away. They were on the very north edge of a cloud pattern which grew far worse along the path to the lower right. (i.e. Kruger Park was socked in.) The clouds at Messina were "deja vu" close to what I'd experienced from Bucharest in August of '99: a thin finger of cumulus moved slowly, spread apart and eventually opened up a hole at the critical time. Talk about cliffhangers!
|
I
mentioned that the Kerns
were able to telnet me their excellent set of digital images
Jan had captured only a few hours before. I grabbed them
when I woke the next day, and began work on the housekeeping
initial tasks right away (you center each shot to match all
the others, crop away unused outer sky, make a first stab at
adjusting levels, color tweaks, the object being
consistency). With digital originals there were tradeoffs.
The exposures looked like slides: all the bright details
were blocked up, ditto the shadows
(i.e.,
maximum digital FF's and minimum 00's).
The shutter speeds were not spaced by the usual one stop (a
factor of 2), but 50% greater (factor of 3). So there wasn't
the long overlapping of details that good compositing
methods are based upon. Yet the images were very sharp
visually, had an immediacy like CCD's of DV Camcorders,
rather than the small "reality-distancing" of even the
finest film emulsion. This made the prominence and limb
details much easier to see, even though CCD's are still
lower in resolution than the finest films.
These three images will give you an idea of the fine work the Kerns had done. These are what I had in my G4 well within 18 hours after they had been taken. As for the rad-grad negative scans from the next day:
The left image is an example of some of the weird color and exposures we began with, caused by scanning the unusually dense negatives, which were overdeveloped in Jo-burg. The right is from a shorter exposure, so it had milder problems, most of which by this stage I'd slowly and carefully removed digitally. Well, the color is still "off", and it doesn't yet look fully promising. Again, the goal was to achieve a consistent series of frames here which would be compatible with the digital frame composite. After both of those slow tasks were complete, we had these two excellent, if somewhat differing representations of totality. I telnetted smaller sizes of each back to Jon right away, so he could relax about the progress going on here, while he and Jan finally enjoyed the last couple of days of their travels -- the major pressure was over.
The left image from the stack of
digital exposures seems particularly good toward the
middle-inner portions, including very decent prominence
details. Stacking images carefully this way cancels out any
noise or grain effects, and can be used to enhance visible
sharpness, if you perform the layer alignments at high
screen magnifications, as I've done. The radial gradient
image, as for our last several eclipse composites, at first
glance looks slightly less real, as the darkening of the sky
from the denser portions of the filter is very unlike what
the eye sees. With a near solar maximum corona (which tends
to be round) you might not notice it so much (see
'99
and '01).
This time we no longer have quite a maximum corona, as the
sunspot activity has begun its decline in the 11 year cycle
(well, 22 years, if you count magnetic polarity reversals).
It does look very reminiscent of the '91
Baja eclipse, though. The next will be even less round,
until at minimum activity in a few years we again will view
a long narrow lozenge shape, wide equatorial streamers to
each side (with north at the top), and only short polar
brushes both above and below. |
Now
that looks really lovely
to us, very naked-eye indeed, doesn't it? This is about what
you'd see under clear skies (those
of you who have experienced a few can better judge the
similarities), perhaps using a
small pair of binoculars. Look at your monitor from several
feet, through a pair of cardboard tubes, to simulate that
kind of impression. We've been told this may be the best
eclipse image yet made. Perhaps. But -- in eclipses there is
always a good portion of hype, it's just the nature of the
spectacle. So every eclipse becomes "the Eclipse of the
Century!" "The Most Spectacular Eclipse Ever Witnessed!"
"The Last Great Solar Eclipse!", and so on, ad nauseum. We
do keep trying to improve our images (many teensy
improvements do add up), and learn a lot in the effort. I
think it would be improbable if cool persistence didn't
result in a gradual improvement for many reasons, increments
of several kinds, even if modest.
Finally,
here's a first for you --
something we've certainly never seen before, although with
digital image tools, I can't help but to ask: "Now, why
hasn't anyone tried this
already?" Instead, remember you saw it here first! ;^) For
each eclipse computer plots are prepared, mapping out the
actual expected lunar edge in great detail (mountains and
valleys seen edge-on). The venerable series of Eclipse
Circulars prepared for years by the USNO, often included
this information. Eventually the Naval Observatory supplied
a separate booklet containing all the needed plots, and a
method to use them for every eclipse now, or in either
direction over centuries. The data came from a careful,
difficult survey of the actual edge data of our Moon,
gathered photographically over many years. And the final
detailed report (which
you could purchase from the USNO's printing office--perhaps
they still have copies of it)
was a heavy, bulky soft cover book. The nightmares to
convert it later to machine readable form were enormous. My
late friend and celestial mechanician there, LeRoy Doggett,
described some of the process -- yikes
(you
don't want to know)!
Isn't that a striking combination (please do click it for
the large view)? I spotted right off that Jon's North point
was extremely accurate (he
reported that he'd let the image of the sun drift, and
rotated his camera back for best horizontal match to the
drift path). It took a mere .6 degrees CW rotation to
match photo with chart
("good enough for government
work!" ;^). You can see right away how very close the
fit is between predictions and actual limb shape. The
mountain and valley shapes line up dramatically. In the
spots where the match is less exact you'll detect that there
were prominences along that position. Those are relatively
bright enough to overexpose this image section, leading to
false valleys. The bright pink of the prominences, even with
all the stretching they've undergone above, is still a good
clue to the origin of these deeper features. There is also
progress being made in improving the prediction data, so the
"fit" in future years should be even better. --Wendy Carlos |
Before we go, let me try to address a topic that comes up fairly often in your e-mails sent to our site, and in threads and speculations on the Web. This regards the methodology of making composite coronal images. I've been collecting a lot of material to demonstrate some of the principles involved, and have created a good start for 2-4 new web pages that will explain what little I've been able to learn in three decades of eclipse imaging. The pages have become far more difficult to present some subtle ideas clearly, and are taking much longer than I assumed when starting out. Ratz! I'll keep at it. Eventually you'll see them here. In the meantime, let me show you just two of the example images I created for those pages, since they came out of my two week's work with the latest images the Kern's took in Messina. You should compare these carefully with the composites above.
The left side example is from the somewhat earlier
Pellett method, which I first encountered in a long article
in Sky & Telescope magazine. This appeared at the start
of the year 1998 (the January issue?), just prior to the
Caribbean eclipse that February. I've mentioned in several
places why I think a fully algorithmic approach doesn't work
too well. This one in particular has several features which
I opinionatedly find decidedly "wrong-headed". The inner
through middle coronal details are not actually seen, just
the addition of several difference masks created via a
radial spin unsharp masking technique. Once these very light
gray images are added together all the coronal features are
presented schematically, and stand out in bold relief. That
can be most useful for solar astronomers, but it has really
little to do with a naked eye view of a total solar
eclipse. --Wendy Carlos You may also enjoy looking at several other good eclipse websites which cover this totality from other points of view. I've mentioned Jon Kern, and he's posted all of his negatives, in large sizes, too, for several recent eclipses, 1998, 1999 and 2001. He includes additional information on his special radially graded filters, too. |
|
Wendy Carlos 02Eclipse
Page