CCNet DIGEST 4 January 1999

    Michael Paine <>

    Rainer Arlt <>


    Mark Davis <>

    Ron Baalke <>

    Andrew Yee <>

    NASA Science News <>

    Andrew Yee <>


From Michael Paine <>

Dear Benny,

Disappointing news for the end of 1998. I have received a reply to a
letter I wrote to Bruce Scott, the new "Minister Assisting the Minister
for Defence" in October. The reply was from the "Chief of Staff, Office
of the Minister for Defence" and, predictably from a tunnel-visioned
bureaucrat, it dismissed the issue with the sentence "Asteroid detection
is currently not a priority for Defence" (if it was I wouldn't have
raised the issue!!).

On the positive side the Shadow (Opposition) Minister for Science,
Martyn Evans, has been supportive and has indicated he plans to raise
it as a budget issue in the Senate.

for details.

Best wishes for 1999

Michael Paine
New South Wales Coordinator
The Planetary Society Australian Volunteer Coordinators


From Rainer Arlt <>

The BIBLIOGRAPHIC METEOR DATABASE is vailable on the WWW now at

Almost 10000 articles, books, and notes about meteors and related
topics are stored in the database, dating from 1794 to 1998. The web
page allows you to search through the files by a tree of keywords, an
authors index and an input mask for direct character string searches in
titles, authors, and keywords. Note that the topics you search for may
not be  present in the title; this is why a tree of keywords was
created in,

which satisfies your query despite the topic not being present in the
title of the article.

The Bibliographic Meteor Database was created in 1982 by Paul Roggemans
(Belgium), the first computerized version was mainly made by Ghislain
Plesier (Belgium), and the work was continued by Rainer Arlt (Germany),
who are very much indebted to Mr. Dale (Belgium), Mr. Vanderbecq
(Belgium), Jurgen Rendtel (Germany), Jeff Wodd (Australia), Dr.
Alexandra Terentjeva (Russia), Evelyne Blomme (France), Ludwig Cluyse
(Belgium), Albert de Kersgieter (Belgium), Dirk Laurent (Belgium), Jean
Meeus (Belgium), Ann Schroyens (Belgium), Christian Steyaert (Belgium),
Sabine Bohmer (Germany), and Manuela Trenn (Germany) for their help and

I would be most grateful for any comments, suggestions or hints on
Rainer Arlt,, 1998 Dec 29.


NEAR Major Engine Burn Completed
January 3, 1999

At noon, EST, January 3, the NEAR mission team conducted a 24-minute,
large bipropellant engine burn, to increase the spacecraft's speed for
a rendezvous with Eros in February 2000. Preliminary indications are
that the burn was successful. NASA's Deep Space Network, which is
tracking the NEAR spacecraft, is expected to confirm the accuracy of
the burn early Monday morning, January 4.

The burn increased NEAR's speed by 2,100 mph (940 meters per second) to
catch up to the faster-moving Eros asteroid, which overtook NEAR during
the Dec. 23rd flyby. At the time of the burn, the spacecraft was
565,650 miles (910,100 kilometers) from Eros

Once accuracy figures from the burn are received, plans will be
finalized for a small hydrazine engine burn to correct any deviation
from the spacecraft's intended location. This burn is expected to take
place in one to two weeks. Periodic trajectory correction maneuvers
will be executed by the Mission Operations Center as deemed necessary
to keep the spacecraft on course during its yearlong journey to the

For now, NEAR continues on its orbit around the sun, traveling at about
43,000 mph (19 kilometers per second) as it gains on asteroid Eros.


From Mark Davis <>


The major shower of January is the Quadrantid meteor shower (QUA),
named after the ancient constellation Quadrans Muralis. It is expected 
to reach a maximum on January 3rd at about 23h UT, Universal Time. This
is very near the time of full moon. The radiant is at 230 degrees, ie.
RA 15h20m, Dec +49, which is just past the halfway mark on a line from
the end star of the handle of the Big Dipper to the upraised arm of

The ZHR, Zenithal Hourly Rate, is about 120 meteors per hour with the
naked eye. These meteors are at an average speed of 41 km per second.
The duration of the shower is from January 1st to 5th, but this shower
is noted for an extremely narrow peak with some studies finding rates
higher than about 60 meteors per hour for only about 16 hours. Hence,
whether or not we are in darkness or daylight at the maximum time
becomes rather critical for seeing the best rates.

Recent studies have found a similarity between the orbit of this shower
and that of Comet 96P/Machholz 1, and indeed, possible relationships
with both the Delta Aquarid and daytime Arietid meteor showers. Another
possibility may be a relationship with Comet 1491 I. Because orbits
change over time (in this case, thousands of years) due to outside
factors, such as perturbations by Jupiter, it is difficult to always
match meteor showers to parent bodies.

Is this shower favorable this year? Unfortunately not. The moon is bad,
so the shower will be adversely affected. However, due to the high
rates that can be associated with this shower, it is still worth

Upcoming Meetings...

May 11-13, 1999:
The Leonid Meteoroid Storm & Satellite Threat Conference is being held
in Manhattan Beach, California. For more information, please contact or check out the website at:
Papers are solicited in many areas, including UV, optical, IR and radar
observations of the 1998 Leonid storm; dynamics, composition,
occurrence of the Leonid meteoroids; and orbital and meteoroid
dynamics: 1997-2000.

July 26-30, 1999:
The Asteroids, Comets, Meteors 1999 Conference is being held at Cornell
University, near Ithaca, in New York State. Details are available at
their website: You can also leave
your name and address, to be contacted with more information. Although
this is a professional conference, a number of North American amateurs,
including a number from NAMN, are planning to attend.

September 23-26, 1999:
The 1999 International Meteor Conference (IMC), the annual conference
of the International Meteor Organization, is being held in Frasso
Sabino, Italy. The cost, including conference, lodging, and meals, is
approximately $200 U.S. For more information, see the IMO website at

Mark Davis,
Mt. Pleasant, South Carolina, USA
Coordinator, North American Meteor Network

And check out:
NAMN home page:


From Ron Baalke <>


Announcement of Opportunity and Special Competition for FY 1999

Planetary Environments:
In order to provide insights into the possibility of life beyond our
own planet, research is also needed to characterize the environments of
planets in the solar system and beyond and to understand the
commonalities of their formation and evolution. Examples of relevant
topics include:

* studies of the formation of Earth, other planets and their satellites;
* remote sensing of planets and their atmospheres;
* studies of interstellar grains and meteorites to establish criteria for
  the presence of biogenic substances;
* studies of interstellar and cometary chemistry, particularly of
  biologically relevant molecules;
* the relationship between interstellar organic molecules and the origin
  of life; and
* research on the biogeochemical effects of microbes on their
  environments on Earth to better design tests for life on other
* Methods and Capabilities for LExEn Research


From Andrew Yee <>

University of Iowa
University News Services
100 Old Public Library
Iowa City IA 52242
(319) 384-0009; fax (319) 384-0024


UI's Louis Frank uses mathematical data analysis to show small comets
are real

IOWA CITY, Iowa -- Asserting that critics have been looking at the
wrong data, University of Iowa space physicist Louis A. Frank has
published a new paper supporting his "small comet" theory that about 20
snow comets weighing 20 to 40 tons each disintegrate in the Earth's
atmosphere every minute.

The paper, which appears in the Jan. 1, 1999 issue of the American
Geophysical Union's (AGU) Journal of Geophysical Research-Space
Physics, uses an automated mathematical formula to filter out
electronic instrument noise from data gathered by NASA's Polar
satellite. The result, says Frank and his UI colleague John B.
Sigwarth, is a "hands-off" analysis showing that "instrumental effects
were not major contributors" to the images of atmospheric holes. Using
the mathematical formula, the two researchers found that the
atmospheric holes photographed by the Polar satellite cameras:

* Increase in number when photographed from lower altitudes.
* Increase in number when photographed during local-morning time
* Appear larger in size in satellite images when photographed
  from lower altitudes.
* Vary in number, depending upon the season.

"What critics of the small comet theory were analyzing was instrument
noise," Frank says. "If you strip away the noise from the data, as they
properly should have done, what remains clearly validates the reality
of atmospheric holes. Our most recent paper is the only comprehensive
paper on this topic and shows, without reasonable doubt, that the
atmospheric holes are indeed a real phenomenon."

Frank says that the mathematical formula applied to the data screened
out possible causes of electronic noise such as longer wavelength
radiation, energetic electrons and uneven sensitivity -- or "hot spots"
-- among camera instrument pixels. Significantly, he found mid-January
1998 data containing no atmospheric holes and used it as a baseline

"The period in mid-January during which no atmospheric holes were
detected provided an excellent opportunity to have a very effective
calibration series of images which were equivalent to an extensive
post-launch laboratory calibration. These in-flight calibration images
were extremely important in establishing the instrument noise
performance without the presence of atmospheric holes and with the
actual temperatures and operating voltages for the instrument. These
images verify the accuracy of our computations of random hole rates,"
he says.

In a 1998 study, Frank and Sigwarth analyzed 1981 data collected by the
Dynamics Explorer 1 satellite and compared it to data gathered by Polar
in 1997, finding a mid-January lull in both sets of data. Despite the
fact that observations of seasonal variations in atmospheric holes were
made 16 years apart by different spacecraft carrying different cameras,
criticism remained. Several papers refuting the theory were presented
at the spring 1998 AGU meeting, one of them suggesting that
measurements made by another satellite show that the atmosphere some 15
to 35 miles above the Earth is much drier than the small comet theory
would suggest.

In December 1997 Frank presented a study at the AGU fall meeting
showing that dark spots (called "atmospheric holes" because of their
appearance on film) captured in June 1997 on Polar photographs decrease
in size and number as the satellite's altitude and distance from the
holes increases. Earlier, Frank had created a stir at the May 1997 AGU
meeting when he revealed a series of Polar satellite photographs,
ranging from a picture of a small comet the size of a two-bedroom house
disintegrating thousands of miles above the Atlantic Ocean to an image
of light emitted by the breakup of water molecules from a small comet
less than 2,000 miles above the Earth. Frank and Sigwarth, who
co-discovered the small comets and designed and built the three Visible
Imaging System (VIS) cameras aboard Polar, offered the pictures as
proof of their theory.

Frank first announced the small comet theory in 1986 after examining
images recorded in photographs taken by Dynamics Explorer 1. Frank and
his colleagues had designed and built a special camera to take pictures
of the northern lights, including the first images of the complete ring
of the northern lights from above the North Pole. But some of the
images contained unexplained dark spots, or atmospheric holes. After
eliminating the possibility of equipment malfunction and numerous other
explanations, Frank and Sigwarth concluded that the atmospheric holes
represented clouds of water vapor being released high above Earth's
atmosphere by the disintegration of small comets composed mostly of

They calculated that more than 25,000 comets enter the atmosphere each
day. At that rate, the steady stream of comets would have added about
one inch of water to the Earth's oceans every 20,000 years -- enough to
fill the oceans over billions of years. The theory was immediately
controversial, with people asking why such objects hadn't been observed
previously. Frank countered that not only their small size --
20-to-30-feet in diameter -- makes observation difficult, but also that
water striking the upper atmosphere glows very faintly as compared to
the bright glow of metal and rock in solid meteors. The controversy
re-ignited after the 1996 launch of Polar, carrying two sensitive
visible light cameras and one far-ultraviolet light camera, made it
possible to photograph the small comets with greater resolution.

For further information, see:

* Small comet web site:

* Small comet website press release:

* Full text of JGR paper:

* New small comet photos:


From NASA Science News < >

28 Dec. 1998 January's Chilly Meteors - The 1999 Quadrantids

One of the year's most intense meteor showers, the Quadrantids, begins
tonight.  The shower stretches from Dec. 28 through Jan. 7 with a sharp
maximum on Jan. 3, 1999.  The Quadrantids are the only major annual
meteor shower whose source, presumably a comet or an asteroid, remains
unknown.  Readers are invited to observe the upcoming shower and to
submit their data for analysis by scientists studying the structure and
origin of the Quadrantid meteoroid stream.



From Andrew Yee < >

ESA Science News

14 Dec 1998

Rosetta Passes Important Design Review

The Rosetta comet rendezvous mission has passed another significant
milestone. According to ESA's usual practice for major projects, a
Rosetta Mission System Design Review took place at ESTEC in The
Netherlands on 10 December 1998. During the review an independent team
of engineers and ESA officials closely scrutinised all the elements of
the mission, including the ground stations, the spacecraft, the payload
of scientific instruments and the launcher.

The review came at the end of several weeks of very intensive
discussion focused around a number of severe constraints which the
mission team will have to overcome. They include:

* the spacecraft's thermal design -- how Rosetta will cope with high
temperatures close to the Sun and much lower temperatures beyond the
asteroid belt

* the available mass for the spacecraft and its scientific payload,
based on the lifting capacity of the Ariane-5 rocket

* providing sufficient electrical power supply from the spacecraft's
solar panels in the dark depths of the Solar System

* the very challenging construction, assembly and test schedule, for
the January 2003 launch date.

The Review Board concluded that there is a high confidence on the
success of the mission and its objectives. Work will now continue on
detailed design activities, with the aim of starting hardware
manufacturing and engineering model testing next year.

The CCNet is a scholarly electronic network. To subscribe, please
contact the moderator Benny J Peiser at < >.
Information circulated on this network is for scholarly and educational
use only. The attached information may not be copied or reproduced for
any other purposes without prior permission of the copyright holders.
The electronic archive of the CCNet can be found at



From Louis Friedman <>

Mr. Peiser:

Thank you for your "open" letter to me about the predictions of the orbit
for 1997XF11.  There are two points – the first is technical accuracy;
which is discussed in the scientific literature and which The Planetary
Society plays no role in evaluating.  We are satisfied that the process
of evaluating the technical and scientific work is proceeding properly.
I certainly do not know of any evidence of cover up or data being

Contrary to what you have stated, there are not two groups of
scientific conclusions about the orbit prediction.  No analysis has
suggested that there is anything more than an infinitesimal
(essentially zero) probability of an impact in the next century from

The second point is Dr. Chapman's position as a columnist with The
Planetary Report. We deeply appreciate his long time voluntary role as
our News & Review columnist. We respect his right to opinions and
reporting – based on his deep scientific involvement and professional
standing in the field of planetary science, even, sometimes, when we
don't agree with some of those opinions or characterization.  We review
his column (as we do all material in The Planetary Report) with the
high standards of our editorial staff, in consultation with our
Editorial Advisory Board and professional colleagues. On this matter,
Dr. Chapman's analysis is in the scientific mainstream, and, as such,
it really isn't very controversial. His opinions about the procedures
involved in reporting and publicizing the prediction, may not be
exactly our own, but we accepted his right, and scientific credibility,
to express them in his column. My own opinion is that there has been
too much personalizing of this whole subject about the 1997XF11
prediction, and rather than augment it, will do my best to minimize it.
It is a tempest in a teapot.

We appreciate your interest and welcome suggestions from our members.


Louis Friedman


From Paolo Farinella <> wrote:

Dear Benny,

I think that in your letter to the Planetary Society you misreported
the point of view of those scientists (I am among them) who think that
1997 XF11 shouldn't have been announced as a short-term potential
hazard in March this year.

You say:

> While it was apparent at the time of the initial announcement that XF11
> could come exceptionally close to Earth, little attention was paid to
> the object until the report on 11 March 1998 that it would in fact do
> so in 2028. Since then, the experts have been deeply split about
> this asteroid, and no unanimity was reached with regards to the future
> risk this object might pose to Earth. After three months of
> observations, astronomers involved in orbit calculations were divided
> into two groups: those who believed that any risk of impact in the next
> century could be ruled out altogether, and those who maintained that
> the limited data available at that time did not allow for such
> unequivocal conclusions.

Actually I belong to a third `group' (the vast majority of the experts, I
think) who claimed to different things: (1) that an impact in 2028
could be safely ruled out from the first three months of observation;
(2)  that as a consequence the short-term (~1 century) hazard posed by
this object is much lower than the collective hazard level posed by the
undiscovered NEAs of similar size (the so-called `background level'). I
think that, after many bitter discussions, Brian Marsden now agrees
with (1); and I have never seen a calculation by him or anybody else
showing that (2) is false. Nor I think that there are other issues
worth being presented to the public in this context.
Best regards and season greetings,

Paolo Farinella



From Brian G. Marsden (

"NASA should fund experts to make valid impact-probability calculations
[rather than] conclude that all it needs to do is fund more
observations ... when in fact we really need better early

This statement, which appeared in the latest issue of The Planetary
Report, prompts me to make a few remarks about (4179) Toutatis, 1998 XB
and 1997 XF11, three of everybody's favorite NEOs of 1998.

In the CCNet Digest for Oct. 15, Duncan Steel alluded to Grzegorz
Sitarski's recent Acta Astronomica paper on Toutatis. I fully concur
with Duncan's statement about the epic orbital studies Grzegorz has
done in his characteristically quiet but competent way over several
decades. Actually, I don't agree that it is necessary to postulate the
action of nongravitational forces on Toutatis in order to link the
observations since its 1989 discovery with the two 1934 Uccle
prediscovery observations that were identified by former MPC Associate
Director Conrad Bardwell soon after Toutatis was found. As with Rob
McNaught's 1993 Apr. 27 precovery measurement of comet C/1995 O1
(Hale-Bopp) and the wealth of 1995-1996 postdiscovery data, the problem
is that the enormous weight of the postdiscovery observations (with
their associated errors, mainly the systematic errors arising from an
imperfect reference-star system) can cause the isolated earlier
observations to show departures from the solution that are
significantly larger than would be expected. By increasing the relative
weight of the prediscovery/precovery data (either by giving them a
weight of ten, say, rather than one, or by literally decimating the
recent data), those earlier data can be made to fit without doing any
injustice to the representation of the later data.

Be that as it may, the whole point of Grzegorz's analysis is that,
despite the well-known fact that Toutatis will come to a distance of
little more than 0.01 AU from the earth in 2004, there is no
possibility that this 5-km object will strike the earth in the near
future. The reason an impact is impossible is that the object's nodes
are currently more than 0.4 AU from the earth's orbit. This is true
even though the 0.5-deg orbital inclination to the ecliptic means that
the orbits of Toutatis and the earth are only 0.006 AU apart. By the
end of the 22nd century, however, the descending node will be
substantially closer than it is now--though still more than 0.1 AU from
the earth's orbit. Coupled with the approaching node is the fact that
the motion of Toutatis is extremely chaotic, mainly because the object
is only 1 AU from Jupiter at aphelion and in 3:1 resonance with that
perturber.  The chaos is such that the position of the object in its
orbit starts to become quite uncertain, perhaps already by the 24th
century, after which the nodal distances become uncertain too.  The
outcome of this is, not only that one cannot currently say whether or
not an earth impact will occur in, say, the 26th century, but that one
cannot say whether or not the circumstances will make it meaningful to
consider whether an impact is then even possible. 

Understandably, the phrase "impact-probability calculations" appears
nowhere in Grzegorz's paper.  Because of the extreme chaos, it is
obviously impossible to make such calculations for the 26th century and
beyond. This would be the case whether or not one considers
nongravitational forces, whether or not one remeasures the 1934 plates,
whether or not one includes Steve Ostro's radar data in the orbit
solution. Of course, one could compute, perhaps, that the impact
probability through the 22nd century is 10^{-9772} (to pick a number at
random).  But in such a case one could worry that the impact
probability might be as large as 10^{-97}, say, because a close
encounter with an unconsidered asteroid just might change Toutatis'
orbital inclination by 0.5 deg and reduce that 0.006 AU (and
decreasing) distance between the orbits all the way down to 0.00004 AU
and less...  Whatever Clark Chapman may think, I suspect our successors
will deem it prudent to continue to observe Toutatis from time to time
over the next several centuries. A project they might in fact undertake
would be to place a transponder on Toutatis--or maybe even colonize
it--the better to monitor the rake's progress.

In another masterly study of real celestial mechanics, Andrea Milani
(in CCNet Digest for Dec. 21) has examined the prospects for close
passages of 1998 XB by the earth (and Venus) over an interval of 25
millennia. One almost wonders why this object should even come up with
regard to consideration of the dreaded impact-probability calculations.
As Andrea says, neither node is currently much within 0.2 AU of the
earth's orbit. Furthermore, because of the 14-deg inclination, the
minimum distance between the orbits of 1998 XB and the earth is more
than 0.1 AU--and it would take a near collision with a Jupiter-sized
asteroid (yet the aphelion distance is only 1.2 AU!) to make this drop
precipitously. But Andrea shows that the slow combination of
perturbations by Venus and the earth can raise the specter of a
possible earth impact some eight millennia from now. As with Grezgorz
and Toutatis, Andrea is of course well aware of the impossibility of 
estimating an actual probability of an earth impact with 1998 XB at
that time. 

I do have one small criticism of Andrea's account, and this is that I
wish to point out that the fact that the semimajor axis of 1998 XB is
as small as 0.906-0.908 AU was already quite clear on Dec. 14, when MPC
Associate Director Gareth Williams published such a value on MPEC
1998-X37, using the observations made during Nov. 25-Dec. 11. 
Furthermore, with regard to Al Harris' remark about the "revised
orbit", it seems not to be widely understood that the problem with the
early orbit determination is that there were in fact two discrete
solutions. Double solutions are quite the rule for objects like 1998 XB
that are discovered at elongations of only 90 degrees from the sun. 
Usually, of course, it is rather obvious which of the solutions should
be accepted.  What was remarkable in the case of 1998 XB is that the
two solutions should be so similar. Gareth's first published orbital
solution for 1998 XB, which appeared on MPEC 1998-X20 on Dec. 5 using
observations extending to earlier that day, gave a semimajor axis of
1.021 AU.  This orbit was the result of a least-squares fit to the 41
available observations with a mean residual of 0.45 arcsec.  At the
same time, Gareth had also computed the other solution, with semimajor
axis 0.903 AU, finding that this least-squares result represented the
same observations with a mean residual of 0.59 arcsec. Although the
fits were essentially indistinguishable, Gareth rather understandably
adopted the solution showing the smaller mean residual. The whole
point, surely, was to provide an ephemeris that would yield further
observations that would enable the physically correct solution to be
isolated. By the time the observations extended to Dec. 9 it was
starting to become apparent that the wrong initial choice had been
made. The switch to the correct solution then followed automatically.

But while Andrea finds the long-term orbital evolution interesting, and
I find interesting the fact that the two early orbit solutions were so
similar, others were drawn to 1998 XB by its presumed great size, or
more specifically, its low absolute magnitude.  And Al correctly points
out that the earlier enthusiasm on this issue (CCNet Digest, Dec. 15)
was premature. The use of the physically correct initial orbit solution
changes the initial calculation of the absolute magnitude H from 14.2
to 14.7. Furthermore, in calculating the absolute magnitude, the MPC
has a practice of essentially giving each observed magnitude unit
weight. But one should understand that these magnitudes are provided,
almost as an afterthought, by astrometrists who, for the most part,
have no particular interest in photometry. Except for the few
astrometrists who also specialize in photometry, the magnitudes they
give are just numbers that come out of the computer reduction, with no
consideration of the color band to which the CCD is sensitive, the
reliability and appropriateness of the magnitudes catalogued for the
reference stars, etc.  Only four of the six early observers of 1998 XB
chose to provide any magnitude observations at all. As later
observations came along, extending through Dec. 26, the unit-weight
process changed H to 15.4. If zero weight is given to the data by the
two most prolific observers, however, H changes all the way to 16.0. 
As it happens, those same two observers were already the most prolific
through Dec. 5, and if zero weight had been given to their data in the
initial calculation (and the correct orbit used), the value of H would
already then have been 15.9.  

It is good that Al drew attention to the problem (which is much more
general than in the case of 1998 XB, of course), but I'm sorry he did
not mention the role of the Czech astronomer Petr Pravec in coming up
with the solution. Petr has been concerned with this problem for some
years now, and he is working with the MPC on finding a satisfactory
general solution. This solution will require the preparation of a table
giving, for each observatory, both a weight and a standard adjustment,
possibly also with a dependence on the time (which could be a real
nightmare, as observers change their habits). The incorporation of such
a table will still allow the MPC to calculate the H values in a very
automatic manner, something that is essential, given that several
hundred orbits (sometimes more than a thousand) are typically being
computed (or recomputed) every day.

I hope the above remarks make it clear that it is fully appropriate for
NASA and other grant-giving groups to fund ever more observations, both
astrometric and photometric (as well as the processing of those
observations). On the other hand, the calculation of impact
probabilities is surely of quite limited value.

If the reader is still unconvinced, let him or her read on.

I almost hesitate to bring up again the matter of 1997 XF11, but it
seems that my earlier messages concerning this object (notably in the
June 8 and July 27 CCNet issues) still have not got through to
everyone. On the morning of March 11 Gareth and I noticed that the
particular orbit I had published five days earlier on MPEC 1997-E13
gave an unprecedentedly close approach to the earth on 2028 Oct. 26. 
This miss distance of 0.00031 AU was a FACT (thus worthy of an
exclamation point), later fully confirmed by at least five other
groups, including Grzegorz and Andrea, who all obtained (using the same
data) values between 0.00023 and 0.00090 AU, significantly less than
half the distance of the moon. I immediately made some tests on
the uncertainty of the miss distance and found that the 88-day arc of
observations was also fully consistent with the object's passage at the
moon's distance, although these quick tests made it seem unlikely that
the object would pass at a distance much greater than that of the moon.
While we found this intriguing, it also immediately told us that we
probably did not have much to worry about. A careful and detailed
statistical analysis carried out by Karri Muinonen over the course of
several months (i.e., much longer than the few minutes I spent on the
topic) concluded that the data then available indicated a 72-percent
chance that 1997 XF11 would come within the moon's distance.  This is
the most interesting result in the papers Clark mentions as having been
presented at the meeting in Madison in October. The real 1997 XF11 was,
curiously, right in the tail of the distribution.

At this point on March 11, Gareth and I considered what we should do
with the information. Since we appreciated that (a) anybody else could
come to the same conclusion from published data (the March 3-4 McDonald
Observatory observations were actually irrelevant from this point of
view, but we were going to publish these on March 13 anyway), and (b)
further observations, both astrometric and photometric, were very
desirable, we decided to issue a request for further data on an IAU
Circular, backed up with an information page in the WWW that included
ephemerides showing where 1997 XF11 would have been in 1990 and in
earlier years.  We also considered saying nothing at all about the
object until the presentation I was scheduled to make at the NASA
meeting to be held in Houston on March 17.  We rejected this second
course because (a) we were almost sure that someone else would remark
on the 2028 encounter in the mean time anyway (for, after all, the
March 3-4 observations themselves would be available on March 13), and
(b) the object was progressively getting more difficult to observe, and
a delay of a week could make the difference between success and failure
(as it happened, the very last observations made of the object were
only on March 23).

But let us suppose that we had taken the second course, and I had
indeed dramatically dropped the bombshell at the end of my March 17
presentation in Houston.  Bear in mind that six or seven of my principal
antagonists were to be in the audience, which altogether totaled about 20.
What would have happened?  Certainly, the dreaded i.p. phrase would
have been uttered, perhaps with the assertion by one of my antagonists
that this was surely as high as 0.1 percent...  But bear in mind, too,
that rather than spend March 12-16 trying to stave off both the press
and my antagonists, I should have had ample time to perfect my
presentation. Remember that all I needed was a few hours of spare time
in order to come up with the "2037" calculation (June 8 CCNet), and a
few hours more to come up with the even better "2040" calculation (July
27 CCNet).  No new computer programs had to be written to do this. 
Given a weekend that included both a Friday the Thirteenth (with a full
moon!) and the Ides of March, please do not doubt that I could have
come up with this post-2028 twist, so that I could then spring the
whole scenario on the assembled group on St. Patrick's Day! Of course,
we all know what the response would be: "Give us i.p.'s!"  Well, as far
as I am aware, we still don't have i.p.'s!  But it is hard to argue
with a single trajectory that could have been taken by 2-km-wide 1997
XF11 to bring it--if unimpeded--only 3000 km from the center of the
earth on 2040 Oct. 26! Remember, too, that we should not then have had
salvation in the 1990 observations. Indeed, I should not even have
known myself that those crucial 1990 images had impressed themselves on
those Palomar films... 

Maybe there would be no precovery films.  The assembled throng could
then argue about i.p.'s until it was blue in the face, but the savior
would not arrive until the first observer pointed his or her telescope
in the direction of 1997 XF11 in January 2000...  Would everybody in
the room have to take an oath not to speak to the press assembled in
Houston for the annual Lunar and Planetary conference? 

So what WAS the impact probability for 2-km-wide 1997 XF11 during, say,
2033-2045? Sure, as with the possible eventual impacts of Toutatis and 
1998 XB, we are dealing with chaotic dynamics, and chaotic dynamics are
not conducive to probabilistic arguments. Since we would be dealing
with something only 40 years hence, rather than several hundreds or
thousands of years, there may be some hope for a meaningful result, but
I have not seen it yet. Yes, we know that 1997 XF11's descending node
has to cross the earth's orbit during that time. We also know that,
from the 1997-1998 data alone, 1997 XF11's revolution period, currently
1.73 years, would end up somewhere between 1.53 and 1.99 years after
the 2028 close approach. In his paper, Karri actually showed the
spreading out of the possible trajectories after 2028, but he stopped
his computations already at the end of 2034. The limits on the
revolution period mean that the only possible encounter of interest
during the time interval considered by Karri is in October 2033 (i.e.,
three times a period of 1.667 years), when I don't think 1997 XF11
could come within, say, 8000 km of the earth's surface: surely a miss
like that would not startle anyone...  This 2033 approach does show in
Karri's plot, but his sampling of data was not sufficient to show that
this date was more significant than other Octobers (and some Mays).

One way of arguing is to say that the minimum distance between the
orbits of 1997 XF11 and the earth, currently 150,000 km, diminishes to
30,000 km in 2028, a reduction of 4000 km per year. This reduction
continues beyond 2028, modified somewhat by how close 1997 XF11
approaches the earth in 2028 (which is unknown, without the 1990 data).
The minimum miss distance therefore takes about three years to cross
the diameter of the earth. Thus it is reasonable that I should find one
deep impact (2040), a graze (2037) and another half-dozen possible
passages within a couple of earth radii. During the middle of the
three-year period, any impact would be quite central. One can therefore
attempt a statement like: during at least one year during the decade in
question, the impact probability may have been as high as the ratio of
the diameter of the earth to the circumference of its orbit, divided by
the object's revolution period in years. That number is close to
10^{-5}, which is, I think, larger than the annual impact probability
by an unknown object 2-km across or larger. However, since I really
have little use for impact probabilities, please don't argue with me
about these rough estimates. 

In conclusion, let me offer a cautionary story for those who are
inclined to solve all problems by statistics.  Consider the calculation
of the quantity N = n^2 + n + 41 for the integers n = 0, 1, 2, and up. 
The result is N = 41, 43, 47, etc.  So what is the probability that N
is not a prime number?  Well, these three values of N are prime, and
the reader can easily verify that the trend persists.  Statistically, I
suppose one can argue that, in this range of numbers, one integer out
of four or five is prime. So one could argue that, with each passing n,
the probability that N is not prime diminishes as 0.2, 0..04, 0.008,
etc.  So almost before one realizes it, the probability becomes
10^{-6}, 10^{-12}, 10^{-18}, etc.  If you wish, you can liken the
passage of n to the passage of time covered by the observations of an
NEO, with then the probability of nonprime N equivalent to earth-impact
probability for the NEO at some set time in the future. The statistical
plodder will probably notice that primes become less frequent as n
increases, but the formula still somehow manages to find them. The
thoughtful geometer, however, suddenly appreciates the seemingly
unrelated geometrical fact that, given n^2, (n+1)^2 follows from the
formula n^2 + n + (n+1). Bingo! If n+1 = 41, then N is 41^2 = 1681,
decidedly not a prime. So, whereas the plodder was thinking of
probabilities like 10^{-30} or so--maybe even "essentially zero"--all
of a sudden the probability of impact becomes 2.4 percent!  By the time
n = 100 rolls around, the probability is up to 14 percent, increasing
to 42 percent at n = 1000 and a whopping 59 percent at n = 10,000.  Far
from being a remote possibility, the nonprimeness of N, and the impact
of the NEO, is rapidly becoming a dead certainty!

Don't get me wrong. Statistics has its uses in attempting to correlate
quantities having no obvious causal relationship.  But the computation
of orbits and the deduction of consequences therefrom also involve
hearty doses of geometry and dynamics.  And except in the most simple
circumstances, dynamics means chaos. Most of the impact danger is from
asteroids that orbit the sun in a quasi-stable state.  An eventual
impact is the product of countless previous perturbations, generally
involving many earlier approaches to the earth.

A year ago, 1997 XF11 was catalogued as the 104th potentially hazardous
asteroid in a series that began in 1932. Now there are 159 PHAs, the
tremendous increase in 1998 being principally due to the great success
of the LINEAR program.  But it is important to examine whether PHAs
really can pose a threat to the earth in the near future. To my
knowledge, 1997 XF11 is unique in that it is the only object under
observation for more than a day or so that actually was a demonstrated
threat only decades hence. My study is not completely exhaustive, but
I have found only one other specific case--of a rather smaller
object--that may conceivably become a problem around 2090.

Brian G. Marsden

CCCMENU CCC for 1999