Weeding down a big target list
Here is a case study on how to weed down a big list of potential targets using semi-automatic searches. This is exactly how I weeded down this list of Lynds Dark Nebula.
- 1 1. get master LDN list.
- 2 2. get rid of the lower opacity clouds.
- 3 3. get rid of the ones i know are covered by spitzer.
- 4 4. get number of references from simbad.
- 5 5. get hits in spitzer archive.
- 6 6. get information from lee and myers paper.
- 7 7. sort by spitzer availability.
- 8 8. make a sorted cut
- 9 final product at this point
- 10 next step
1. get master LDN list.
this link on wiki: http://adsabs.harvard.edu/abs/1996yCat.7007....0L go to Online Data: http://vizier.cfa.harvard.edu/viz-bin/VizieR?-source=VII/7A ask it to give us plain text ("| separated values"), all objects, ldn number, ra and dec (1950 and 2000), glon and glat (galactic lat and long), area, and opacity class, plus any other columns you want.
take plain text and import into excel, such that columns import cleanly. i had to rename the file to something.txt and then ask excel to import it, with "|" delimiting the columns. you have then i think 1794 objects.
2. get rid of the lower opacity clouds.
i think it shows up already sorted by ldn number. sort by opacity class, reverse numerical order. move everything other than opacity class 5 or 6 down to the bottom of the table. keep the opacity class 5 or 6 near the top. i entered some blank lines which i colored bright red to remind myself easily where the border was between opacity classes. this gets rid of 1247 of the objects.
3. get rid of the ones i know are covered by spitzer.
i happen to know there are two big surveys using Spitzer, called GLIMPSE and MIPSGAL. both are surveying the galactic plane, the former using IRAC and the latter using MIPS. The GLIMPSE survey is older than MIPSGAL. their website is here: http://www.astro.wisc.edu/sirtf/ here, you can learn that they surveyed the inner 65 degrees (of longitude) of the galaxy, plus or minus 1 degree of latitude.
sort the lynds target list by ldns number. this ends up being rougly galactic long sorted.
what i did next was ask xls to do conditional formatting, and (a) color the cells in the galactic longitude column to be light blue if the value was less than 65 or greater than 295 (and less than 360). and (b) color the cells in the galactic longitude column to be light blue if the value was greater than -1 and less than 1. anything that is then blue in both the glat and glon cells is an object that appears in mipsgal and glimpse. this gets rid of just 28 of the objects; i was expecting it to get rid of many more. i entered some blank lines which i colored bright red to remind myself easily where the border was between glimpse and non-glimpse coverage.
i have now 518 objects left in the pool of potential targets. sort this by ldn number. extract the ldn number to its own spreadsheet and use xls to create a text file of just ldn numbers for these 518 objects. in order to do this, i make a separate worksheet, copy the ldn numbers to column b, in column a, make the first cell "LDN" and then "fill down" to fill out the rest of the column with all "LDN". i make column A right-adjusted, column B left-adjusted, and save this worksheet to a text file.
4. get number of references from simbad.
go to simbad http://simbad.u-strasbg.fr/simbad/sim-fid tell it ("output options") that you want plain text output. query by list (second down on that page). pick the plain text list of targets file you just created. it will take several seconds to come back. save the file of results as a plain text file.
import this results text file back into xls, as a new worksheet. copy the column with the number of bibliographic references into a new column in the master spreadsheet. if it doesn't match the number of rows you had from before, you did something wrong, and you should go back and check things.
5. get hits in spitzer archive.
take the input file you just made for simbad, reformat it so that each line is surrounded by quotes, e.g. "ldn 4 " rather than just ldn 4 without the quotes and break it into pieces of 100 targets per file (you'll have 6 files). The top of each of these files should look like this:
COORD_SYSTEM: Equatorial # Equatorial (default), Galactic, or Ecliptic EQUINOX: J2000 # B1950, J2000 (default), or blank for Galactic NAME-RESOLVER: Simbad # NED or Simbad (default) RADIUS: 5 # RADIUS is in unit arcmin #Name
start leopard. query by fixed target list, load the first of your files you just created. (here is an example batch search list: media:simbadlist.txt) this will take a while for each leopard search file. when it is done searching, it will ask if you want to load the AORs for each program it found. just tell it "cancel" for now. leopard automatically generates a file in the same directory as your target list, with the same root filename but with a time/date stamp appended. because of a leopard bug, you need to restart leopard after each target search.
this header tells leopard to search for things within 5 arcmin. it's going to tell us everything (IRS, IRAC, MIPS) that it thinks overlaps. it's not perfect, but it will get us close enough for weeding purposes.
make a new column for these results in the master xls file. look at the summary files, and if there is an IRAC/MIPS hit, enter that information in the xls file.
6. get information from lee and myers paper.
http://adsabs.harvard.edu/abs/1999ApJS..123..233L this is the one looking at optically-selected cores to see if they thought there was a YSO inside. add this to a column in the spreadsheet.
7. sort by spitzer availability.
start with ones without any spitzer data (as opposed to ones with just irac or just mips). this leaves 270 obj.
8. make a sorted cut
sort those by lee&myers, then number refs (decr numbers), then area (incr number). just 9 yeses from lee&myers. 14 without lee&myers but with >10 pubs (worth going to look at those pubs to see if those pubs think there is something there).
-> 23 objects to go look up the pubs, look up the POSS image, see how big/interesting they look. LDN number: 31 129 219 425 518 543 769 778 981 1036 1041 1082 1100 1125 1139 1143 1235 1257 1340 1407 1598 1685 1744
final product at this point
still need to add notes here re: pub checking and image checking for those objects
--Dewolf 10:42, 18 January 2008 (PST) I have set up a PowerPoint with images of each of the final potential targets from Luisa's Excel file. It includes visual opacity and visibility info, as well as the number of Simbad references. We had a snow day (6th one since Thanksgiving!) again today, so I had plenty of time to work on this. Media:LDN Targets.ppt
--Johnson 13:50, 21 January 2008 (PST) Attached is a chart that shows the SIMBAD references for each of the objects in Dr. Rebull's short list: 31 129 219 425 518 543 769 778 981 1036 1041 1082 1100 1125 1139 1143 1235 1257 1340 1407 1598 1685 1744. Media:SIMBADsearch.doc Talk with you soon. --chj
--Rebull 14:16, 21 January 2008 (PST) but really from jan 20: cris sent me this in email:
I don't know how accurate this would be - but here is how I came up with the attached graphic. I opened the fits file of the three "best" (in my opinion) LDNs with Image J. I did a rectangular selection of what appeared to be the darkest (visually) area of each LDN (after I first enhanced the contrast a bit. I then did a histogram of these areas to find the Mean pixel value - thinking that it may show the optically densest area of the LDN. If this is true - then 1143 appears to be the best candidate - at least in terms of optical density. What do you think?
I think you're right, this is probably a good way to get at how opaque the cloud is. here's the graphic:media:cris_opacity.png