It is currently Fri, 21-09-18, 2:15 GMT

All times are UTC




Post new topic Reply to topic  [ 4 posts ] 
Author Message
PostPosted: Tue, 21-10-14, 23:59 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 418
Location: Tuscany, Tyrrhenian Sea
Hi! I'm diggin' into the ViZier database to collect planetary nebula data (not reflection nebula, SNR or HII regions, for now) in order to collate them and then getting a planetary nebula sheet as complete as possible before to create a .DSC. Now, because the distance directive is mandatory, the tables shows this latter scattered here and there and often with two/three values accordingly with the different enquire by astronomers, also quoting other data already collected. The noob question concern the criterium of choice about this value:
1) simple aritmetic mean?
2) to prefer the values of the latest enquiry?
3) Other?

Thanks in advance.

P.S.
In some case also the position axis follows the rule of two or more values.

_________________
Never at rest.
Massimo


Top
 Profile  
 
PostPosted: Wed, 22-10-14, 7:32 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4576
Location: Hamburg, Germany
fenerit wrote:
Hi! I'm diggin' into the ViZier database to collect planetary nebula data (not reflection nebula, SNR or HII regions, for now) in order to collate them and then getting a planetary nebula sheet as complete as possible before to create a .DSC. Now, because the distance directive is mandatory, the tables shows this latter scattered here and there and often with two/three values accordingly with the different enquire by astronomers, also quoting other data already collected. The noob question concern the criterium of choice about this value:
1) simple aritmetic mean?
2) to prefer the values of the latest enquiry?
3) Other?

Thanks in advance.

P.S.
In some case also the position axis follows the rule of two or more values.



If you really want to to a good job, you need to decide according to the method used for the distance determination. This was also mandatory when I selected the proper distances and other crucial parameters for my large galaxies.dsc file. Using Perl, I always added a respective code word to each entry:

Here are the corresponding hashes from the Perl script:

# hash with distance methods used, key = $dist_flag
%method = ("0" => " # distance uncertain!\n",
"1" => " # method: SBF\n",
"2" => " # method: T-F\n",
"3" => " # method: V_cmb\n",
"4" => " # method: B200\n",
"5" => " # method: ZCAT\n",
"6" => " # method: NED-1D average\n",
"7" => " # method: other\n",
"8" => " # method: NED (batch)\n"
);
%colormethod = ("0" => " # method: B-V measured\n",
"1" => " # method: B-V via Hubble Stage (T) data\n",
"2" => " # uncertain color!\n"
);


In addition you may pick the 3 best methods and finally use these only for calculating averages. All this is easy with Perl, as you know.

A good example for being clear-cut about the used method is the NED-1D data base of galaxy distances:

http://ned.ipac.caltech.edu/Library/Distances/

Good luck,
Fridger

_________________
Image


Top
 Profile  
 
PostPosted: Wed, 22-10-14, 16:01 GMT 
Offline
User avatar

Joined: Mon, 03-09-07, 23:01 GMT
Posts: 418
Location: Tuscany, Tyrrhenian Sea
t00fri wrote:
Here are the corresponding hashes from the Perl script:


I've seen that within "deepsky.pl", main set-up script even to learn Perl.
For PNs my question maybe is more simple.
For example, the ESO Catalogue of Galactic Planetary Nebulae (Acker+, 1992) on VizieR shows distances in kpc following this methods:

(cit.)
C = from cluster membership
D = from dust
E = from local extinction study
K = from kinematical studies
M = mean value from a compilation of individual distances
S = from spectroscopic parallax of binary companions
W = from wind
X = from a comparison of tangential and radial expansions
(blank) = statistical distance

so that for, i.e. NGC 6369 aka PNG 002.4+05.8 we have:

(cit.)
002.4+05.8 E -> 1.500 kpc
002.4+05.8 K -> 2.000 kpc
blank (statistical)
002.4+05.8 0.330kpc 84..4108
002.4+05.8 0.420kpc 82...363
002.4+05.8 0.600kpc Ma84
002.4+05.8 0.660kpc CKS91
002.4+05.8 1.080kpc 75..9005
002.4+05.8 1.100kpc 76..9034
002.4+05.8 1.200kpc 78..1992
002.4+05.8 1.500kpc 71..9065

The third column is the "Reference frame used for positions" (cit.) but no more than this verbose hint is said.

Is it possible to take up a global, simple mean, or the statistical values must be avoided/treated with (statistical) caution and therefore not suited for a simple mean between "E" and "K"? The method "M" (mean value from a compilation of individual distances) doesn't seem to concern statistics, otherwise it would have replaced the "blank" of the last rows. Avoiding statistical distances will reduce remarkably the number of known PNs with enough data to be converted in .DSC.

_________________
Never at rest.
Massimo


Top
 Profile  
 
PostPosted: Thu, 23-10-14, 8:36 GMT 
Offline
Site Admin
User avatar

Joined: Fri, 31-08-07, 7:01 GMT
Posts: 4576
Location: Hamburg, Germany
fenerit wrote:
For PNs my question maybe is more simple.
For example, the ESO Catalogue of Galactic Planetary Nebulae (Acker+, 1992) on VizieR shows distances in kpc following this methods:

(cit.)
C = from cluster membership
D = from dust
E = from local extinction study
K = from kinematical studies
M = mean value from a compilation of individual distances
S = from spectroscopic parallax of binary companions
W = from wind
X = from a comparison of tangential and radial expansions
(blank) = statistical distance


so that for, i.e. NGC 6369 aka PNG 002.4+05.8 we have:

(cit.)
002.4+05.8 E -> 1.500 kpc
002.4+05.8 K -> 2.000 kpc
blank (statistical)
002.4+05.8 0.330kpc 84..4108
002.4+05.8 0.420kpc 82...363
002.4+05.8 0.600kpc Ma84
002.4+05.8 0.660kpc CKS91
002.4+05.8 1.080kpc 75..9005
002.4+05.8 1.100kpc 76..9034
002.4+05.8 1.200kpc 78..1992
002.4+05.8 1.500kpc 71..9065

The third column is the "Reference frame used for positions" (cit.) but no more than this verbose hint is said.

Is it possible to take up a global, simple mean, or the statistical values must be avoided/treated with (statistical) caution and therefore not suited for a simple mean between "E" and "K"? The method "M" (mean value from a compilation of individual distances) doesn't seem to concern statistics, otherwise it would have replaced the "blank" of the last rows. Avoiding statistical distances will reduce remarkably the number of known PNs with enough data to be converted in .DSC.


As much as I understand, the letter codes above are also to indicate the various methods used. The best and most time consuming approach is that you try to find out (from the original scientific papers) what is done in each of these methods and what are their inherent systematical errors.

Systematical errors do NOT decrease upon averaging like the statistical ones do! Statistical errors decrease like 1/sqrt(N) , where N is the number of statistically independent measurements. Systematical errors usually are related to hardware issues. E.g residual misalignments of instruments would count as a systematical error.

Hence if a measurement has resulted in a "value +- error", you should --first of all-- try to find out whether the quoted error is mainly systematical or mainly statistical in nature.
That would then dictate how to treat a set of data like your distance values above.

Proper statistical errors are themselves so-called stochastic variables, fluctuating according to a Normal (i.e. Gaussian) distribution. In such case statistical errors from different sources a,b,c... may be added in square like sigma_total = sqrt(sigma_a^2 + sigma_b^2 + ...) etc.

Most importantly, the center values may be averaged which then would get you closer to the "true" value.

As to systematical errors, think of the trivial example that a clock was improperly set or an instrument was misaligned by some amount. If you make a number of subsequent measurement with the same misalignd equipment, the values would obviously NOT improve by averaging. Such misalignments would generate the same shifts in each measurement!

In any case for a thorough job you will have to consult the various original papers in the light of what I just wrote. Most of these papers should be freely available in the net. If some paper is not freely available, I can presumably look it up for you. I am entitled through my lab to read (but NOT to distribute!) those scientific astronomical papers.

Here I just came across a tutorial web page that also discusses (in more detail) what I wrote above. What I called "statistical" errors, the author calls "random" errors. In my field these are always called statistical, though.

http://www.rit.edu/cos/uphysics/uncerta ... systematic

There are notably many examples and the text is kept simple! I hope it's also correct throughout ;-) .

Fridger

_________________
Image


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 4 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Group