Subject:
|
Re: The "geography" of local space
|
Newsgroups:
|
lugnet.space
|
Date:
|
Tue, 9 Nov 1999 11:17:21 GMT
|
Viewed:
|
1174 times
|
| |
|
|
Hi, Paul,
Paul Baulch wrote:
> John J. Ladasky Jr. wrote in message <3825442B.5A4DD282@my-deja.com>...
> > Paul, I hate to sound like a Microsoft salesdroid, but you need more memory • for
> > your computer (or, at least a larger download cache). For the genetic • database
> > queries I perform for my paying work, I regularly obtain downloads of this
> > size. The computer here at work is a Mac G3 with 256 MB of physical RAM. • I'm
> > using the Netscape browser, V4.7. I have allocated 20 MB of physical RAM • for
> > the browser, and it maintains a 20 MB cache.
>
> Well, what can I say, but:
> 1) This was two years ago on a machine with 32meg of memory and a version of
> Nutscrape that crashed all the time and generally sucked.
> 2) The VizieR results themselves were not downloaded as plaintext, but HTML.
> My poor 32meg machine had Buckley's chance of downloading it all, regardless
> of how big I made the cache, I'm afraid.
> 3) Who paid for your work machine? My employer has hundreds of computers
> other than mine to worry about ;-)
Trust me, Stanford has THOUSANDS of computers other than mine to worry about!
Oh, and it's not my personal workstation, though I handle most of its technical
operation. We have about one computer per two people in our research group.
> > The average distance between star systems in the immediate vicinity of the
> > Sun is around 2.0 parsecs. In the interest of creating an accurate map, I
> > would exclude any star whose distance error exceeded 0.2 parsecs -- one • tenth of
> > the average interstellar distance. Under optimal conditions, Hipparcos'
> > parallax errors were about one milli-arc second. This means that I would • flat-out
> > reject any star farther away than 20 parsecs.
>
> If you want it to be accurate, then why, pray tell, will you be excluding
> stars whose parallax with errors still puts them _completely_ within twenty
> parsecs? For instance, there's an apparent mag 5.9 star of type K4 whose
> parallax puts it between 6.2 and 7.9 parsecs - pretty darn close if you ask
> me, right slap dang in the middle of your map. But with an error of 1.5
> parsecs!!! "Nope! 's not there!" :-)
In my pre-doctoral days, I worked at a medical instrumentation company called
Becton Dickinson Immunocytometry Systems. The main product at BDIS was this
device called a flow cytometer. It drew cells in a liquid suspension into a
thin stream, past an observation point where a laser intersected the cells.
The cells passed by the laser *almost* one at a time, in single file. But
occasionally, two or more cells would pass through the observation point at the
same time. We called these coincidence events. Concidence events would give
us specious data, which we took great pains to identify and discard.
Take a closer look at the commentary on the Hipparcos main pages. They're
finding it necessary to manually reassess the data for "points" that have high
parallax errors, and other data anomalies. Quite a number of these have turned
out to be double stars that the automated software erroneously tried to analyze
as single stars. A few even appear to be pairs of stars that are not
gravitational doubles, but that were superimposed during one or more
measurements. In at least two cases, it appears that scattered light from a
slightly off-axis star entering the detector was misinterpreted by the
computer. In other words, there are coincidence events in the Hipparcos data.
Therefore, I think that some of these dim, large-error stars might really *not*
be "there." Oh, I'm sure that there is a glowing object in that general
direction, but I would be very uncertain about its *distance* -- more
uncertain, even, than the standard error computed by the automated software
would suggest. Take a look at this data point for a supposedly nearby red
dwarf, HIP 15689. In fact, it's the 13th-closest Hipparcos object to the Sun
with a parallax of 227.45 milli-arcseconds, placing it a mere 4.40 parsecs
away:
http://adc.gsfc.nasa.gov/viz-bin/VizieR-5?-out.add=.&-out.add=_RAJ2000,_DEJ2000&-source=I/239/hip_main&recno=15675
Not only is its parallax error high (61.79 milli-arc seconds), its positional
error is similarly horrible -- 66 and 88 milli-arcseconds in right ascension
and declination, respectively.
Now, take a random star and look it up on another cool search engine that I
found, called Simbad:
http://simbad.harvard.edu/Simbad
This appears to be a kind of thesaurus for the various star catalogs. Enter a
star's name in one catalog, and it will give you back the names of the same
star in other catalogs. You will typically find about twenty corresponding
entries, as it is in many catalogs. HIP 15689 only has seven corresponding
entries, and the prominent note "star in double system." Notably, it's missing
from the Gliese catalog, which focuses on nearby stars. Why is HIP 15689 not
found in more star catalogs? Could it be that more than one survey is not
exactly sure what it's looking at in that particular direction?
> > I just tried the query "parallax > 50 milli-arc seconds" AND "parallax
> error < 8 milli-arc > seconds" over at VizieR. There are 864
> stars that meet these criteria! Another search > shows that 345 of
> these have Johnson B-V values between +0.54 and +1.3 -- meaning
> > that their spectral types are between F8 and K7. For you non-astronomers • out
> > there, this range of spectral types encompasses the yellow to orange stars • like
> > our own Sun, that we think are most likely to possess habitable planets.
>
> Except for the multitude of stars you omitted due to your strange criteria -
> ones that _must_ be within those twenty parsecs according to the Hipparcos
> data.
Well, I'll admit that I've missed a few. But they're the ones that will most
likely require non-automated means of confirmation. Any astro grad students
out there?
> I think that the best and most accurate star map for, say, 20 parsecs, would
> be obtained by selecting all stars in the catalog for whom (parallax minus
> error) > 50 milliarcseconds. If its less, it means it might be further than
> 20ps, so your map can omit it by pretending it "actually happened to be
> further out". But if the star _must_ be within a certain range, it's _much_
> more accurate to show the star than make up some weird excuse as to why its
> absent.
Well, I hope I've convinced you that it's not a weird excuse. It's based on
personal experience with glitches in data collected by automated systems.
> > Remember, too, that there are hundreds of red dwarf stars, unknown and
> > unmapped, inside this 20-parsec sphere. So, if someone wanted to use their
> > imagination and create a fairly dense cluster of dim stars nearby, it • wouldn't
> > be totally implausible.
>
> *LOL* Priceless!!! Most of the nearest Hipparcos stars you've excluded fall
> within this very category.... :-)
> See for yourself. The dimmer, nearer stars in the Hipparcos catalogue show a
> strong tendency to have larger parallax errors.
Indeed they do -- as you get closer to the Hipparcos limiting magnitude of +12
or so, the errors climb accordingly. Still, after looking at a table of the
nearest 1000 or so objects, there seem to be two distinct groups. About half
of the magnitude +10 to +12 objects have parallax errors < 5 milli-arc
seconds. Then there's a second group that has parallax errors >20 milli-arc
seconds. I am really suspect of this latter category.
Then there will be objects with visual magnitudes dimmer than +12, that
Hipparcos will have missed entirely. Even nearby M-class dwarf stars may be
this dim. I was thinking of these when I said "use your imagination." Quite
a few such stars inside the FOUR-parsec radius are listed at:
http://cfa-www.harvard.edu/~thenry/TOP25.htm
Consider Lindsay Braun's mention of the Sun's proposed companion, "Nemesis."
Suppose that Nemesis was an M7 dwarf star with an absolute magnitude of +15,
located at a distance of 0.1 parsec. It would have a visual magnitude of +10.
Could we have missed it? Possibly. We might have even missed it with
Hipparcos, if the software wasn't prepared to accept two measurements of an
object with a *really* high apparent motion as a single object.
> I mean, let's face it..... we're speculating about who's exploring these
> systems in the future, using information that is still horribly incomplete.
> I think that we can be a little lenient when it comes to exactly where the
> stars are.... especially when Hipparcos has allowed us to create a star map
> of hundreds of thousands of stars that we_know_ to be within 500 parsecs.
> That's the important thing - they are _there_.
Still, it's fun to try to get it right, no?
Is there anyone out there who's really interested in doing a Lego scenario
based around Nemesis, or some other red dwarfs? Maybe the UFO or Insectoid
races have figured out how to eke out a living from tidally-locked terrestrial
worlds orbiting close to tiny suns. Maybe they spread around the galaxy
un-noticed, by hopping from star to star that nobody else wants.
--
John J. Ladasky Jr., Ph.D.
Department of Structural Biology
Stanford University Medical Center
Stanford, CA 94305
--
|
|
Message has 2 Replies:
 | | Re: The "geography" of local space
|
| (...) Whoa! Citation! Can I send this one to tthe Humanities Index? (>Insert the noise of a CV being violently padded here<) (...) Absolutely. Maybe it makes us feel a little closer to reality, like we're not just playing deities but playing (...) (25 years ago, 13-Nov-99, to lugnet.space)
| | | Re: The "geography" of local space
|
| John J. Ladasky Jr. wrote in message <382802B2.3FE87E39@m...ja.com>... (...) would (...) minus (...) than (...) _much_ (...) its (...) on (...) John, it's certainly a very good theory. After considering what you say, I think that I would consider (...) (25 years ago, 14-Nov-99, to lugnet.space)
|
Message is in Reply To:
| | Re: The "geography" of local space
|
| John J. Ladasky Jr. wrote in message <3825442B.5A4DD282@m...ja.com>... (...) for (...) database (...) I'm (...) for (...) Well, what can I say, but: 1) This was two years ago on a machine with 32meg of memory and a version of Nutscrape that crashed (...) (25 years ago, 9-Nov-99, to lugnet.space)
|
34 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
