[Prev][Next][Index][Thread]

(meteorobs) Dr. Frank's small comets under fire



(forwarded  with permission)

--------------------

Harald Schenk
Sheboygan, WI


>PLANETARY SCIENCE:
>Tiny Comets' Spots Called Artifacts
>
>Richard A. Kerr
>
>At a geophysics meeting last spring, space physicist Louis Frank had the
>whole community seeing spots. He
>flashed stunning satellite images on the screen, showing dark spots
>silhouetted against the ultraviolet (UV) glow
>of Earth's upper atmosphere. Those spots, he said, restored to grace his
>discredited theory that Earth is pelted
>with fluffy, house-size snowballs 30,000 times a day. Frank had proposed
>these small comets back in 1986
>(Science, 10 June 1988, p. 1403), but no one believed him because the idea
>conflicted with so much other
>evidence. The new images didn't win converts to small comets either, but
>they did persuade a number of
>researchers that something unusual is going on in Earth's upper atmosphere.
>
>Now the tables have turned once more. An analysis of UV images from a
>similar camera on the same satellite
>suggests that the mysterious spots are nothing more than instrument
>artifacts. In a paper to appear in the 15
>December issue of Geophysical Research Letters (GRL), space physicist George
>Parks of the University of
>Washington, Seattle, and his colleagues present analyses of their images
>from the Ultraviolet Imager (UVI)
>aboard the Polar satellite. They conclude that "there is no scientific
>evidence from UVI that snowballs pelt Earth."
>The spots are not clouds of water left from high-altitude impacts of small
>comets, they say, but simply artifacts
>produced inside the camera--so much snow on a UV television.
>
>"Parks's paper is absolutely devastating," says longtime small-comet critic
>Alexander Dessler, a space physicist
>at the University of Arizona, Tucson. But Frank, of the University of Iowa
>in Iowa City, counters that the problem
>lies with Parks's analytical methods, which he says dismiss real spots as
>mere mirage. Meanwhile, echoing past
>chapters in the saga of small comets, Parks removed his name from a paper
>with Frank, and other researchers are
>grumbling that Frank is keeping his raw data to himself. With Frank once
>again the only person who can see dark
>spots, even his newfound supporters are taking a cautious tack. "It would
>take some independent person to try
>to referee this," says Robert Meier of the Naval Research Laboratory in
>Washington, D.C., who spoke up for
>spots at the spring meeting. "That's probably the only way to resolve it."
>
>Frank's spring announcement--billed as a triumphant vindication in the
>popular press--was largely based on data
>from his UV camera aboard Polar. Although they rejected small comets, some
>space physicists were convinced
>that the spots, at least, were real (Science, 30 May, p. 1333), in part
>because the same specks apparently turned
>up in Parks's UVI. That camera records UV images in much the same way
>Frank's does, just in a narrower part of
>the UV spectrum.
>
>Bolstering the case even more, Frank said that water from some of the same
>spots could also be seen in images
>from the visible-light camera (VIS) on Polar. The VIS, which Frank also
>runs, picks up the glow of a water
>fragment called hydroxyl. Indeed, he and his Iowa colleague John Sigwarth
>found that the VIS and Parks's UVI
>captured the same spot in five cases. (The two groups exchanged data, making
>Parks one of the few outside
>Frank's Polar team who has his raw image data.) A preprint of a paper on
>those simultaneous detections--with
>both Frank and Parks as co-authors--was circulated last spring and helped
>persuade other researchers that the
>spots were real. That paper, together with several others on Polar
>observations and small comets, appears in the 1
>October GRL.
>
>But Parks is no longer a co-author. He grew uneasy with Frank's methods of
>analysis and calibration, especially
>after discovering that the UVI had also recorded dark spots during
>calibration tests in the laboratory, when the
>only thing it was looking at was a UV light. A dark spot of a given size--as
>detected by a computerized detection
>scheme--was as likely to show up in a calibration image as in an image made
>from orbit. Parks doesn't know
>exactly what causes the spots, but he believes they must be inherent to the
>UV camera. "If you're going to try to
>understand your instrument," he says, "you had better look at it in the lab."
>
>Frank agrees that there is plenty of noise in the images, but he says he can
>distinguish spots from artifacts by
>comparing UV and visible images. When he finds a spot in the same place at
>the same time in both images, he
>assumes it's real: "What I use to verify [spots] is that there has to be
>simultaneous events in the VIS camera." He
>also looks for a distinctive effect caused by the Polar spacecraft's slight,
>unintended wobbling, which makes any
>camera recording a real object "see double" and produce a pair of spots.
>
>That approach has not convinced Parks that the cameras were both imaging a
>true spot. He has not yet
>completed his own correlation of the UV and visible data sets, but "my
>preliminary analysis suggests there's no
>simultaneity. It's just an accidental coincidence." In other words, both
>cameras happened to show noise at the
>same place at the same time. "I asked him to remove my name," he says,
>"because I didn't agree with his
>interpretation and the way he was doing the analysis."
>
>Parks also did a computerized search for doubled spots in his UV images. He
>found plenty, but he also found that
>close pairs of spots are oriented at random, rather than in the direction
>the wobble would give them. That
>suggests that these doubled spots are also just an accident. "There's no
>scientific justification for anybody
>looking at these data and extracting the kind of information Lou [Frank]
>is," Parks says. "Ours is a standard
>technique; this is what experimental physics is about. He's just looking at
>an image and picking out an event."
>
>Frank has rebuttals for all these concerns. First, he says that lab
>calibrations aren't as good as those done in
>flight: "We have lab calibration images, but we feel it's much more
>favorable to calibrate them in flight."
>Unfortunately, dark spots won't show up against the black of space--so the
>space calibrations can't be used as a
>control for dark spots. Frank also insists that Parks's statistical analysis
>isn't up to the job. Using the same
>approach as Parks, "I never got anything out of my images either," Frank
>says. "There's just so much noise in the
>data that [Parks's] approach didn't tell me whether they were there or not."
>
>To observers, the whole debate may create a sense of déjà vu, because this
>is not the first time Frank has seen a
>signal where others, including the originators of the data being analyzed,
>saw noise. Back in 1987, Frank's original
>dark spots, seen in images taken by the Dynamics Explorer satellite, were
>put down as artifacts by a team led by
>Bruce Cragin, now at CES Network Services in Farmers Branch, Texas. They
>noted, among other things, that the
>spots were the same size in the images no matter how far away the actual
>events would have been. Then in 1989,
>Frank and his colleagues reported dark spots in UV images from the Viking
>satellite. But John Murphree of the
>University of Calgary in Canada, the principal investigator for the Viking
>imager, found similar spots in lab
>calibration images--and removed his name from the preprint being circulated.
>Now, in the wake of Parks's
>withdrawal, even Meier says he's "trying to disengage a little bit."
>
>Given this history, it may not be easy to resolve what the dark spots really
>are. The solution will only come, many
>researchers say, if outsiders are free to analyze the raw data, but Frank
>has a reputation "for being slow to
>distribute his data," as one close colleague puts it. Other than Parks, no
>one has seen more than a smattering of
>the visible-light camera data, which are now vital to Frank's argument.
>Parks has a proposition: "Let's get this raw
>data out on the Internet and let the scientific community be the judge."
>Frank responds that anyone who wants
>data should just ask. Then, perhaps, more researchers will be able to see
>spots--or not.
>
>Volume 278, Number 5341 Issue of 14 November 1997, pp. 1217 - 1218 
>©1997 by The American Association for the Advancement of Science. 
>
>
>


------------------


James Richardson
Graceville, Florida
richardson@digitalexp.com

Operations Manager / Radiometeor Project Coordinator
American Meteor Society (AMS)
http://www.serve.com/meteors/


References: