This was the first in a series of meetings to discuss what is and can be done with the data for the Sloan Digital Sky Survey. The next meeting will probably be held at the University of Chicago sometime in October 1998. Some information about this meeting is given below.
| Topic | Presentor | Approximate Time |
|---|---|---|
| Status of the First Light data | Brian Yanny | ~30 min. |
| Status of the Photometric System | Allyn Smith | ~20 min. |
| White dwarfs in the SDSS | Allyn Smith | ~10 min. |
| Metallicities of G Stars from SDSS Colors | Heidi Newberg | ~30 min. |
| Use of the Sloan Data Simulations for Science | Chris Stoughton | ~30 min. |
| Star-Galaxy Separation | Jon Loveday | ~15 min. |
| The Angular Correlation Function | Douglas Tucker | ~15 min. |
| Searching for Dwarf Spheroidals in the Galactic Halo | Douglas Tucker | ~10 min. |
| Photometric Redshifts | Tim Mckay | ~15 min. |
| Weak Lensing | Tim Mckay | ~30 min. |
| Galaxy-Galaxy Lensing | Albert Stebbins | ~15 min. |
| Name | Participation | |
|---|---|---|
| Asantha R. Cooray | asante@hyde.uchicago.edu | |
| Scott Dodelson | dodelson@mail-astro-theory.fnal.gov | |
| Phil Fischer | philf@astro.lsa.umich.edu | |
| Lucy Fortson | gates@tyrone.uchicago.edu | |
| Evalyn Gates | gates@tyrone.uchicago.edu | |
| Rocky Kolb | rocky@fnal.gov | |
| Jonathan Loveday | loveday@oddjob.uchicago.edu | |
| Tim Mckay | mckay@mich.physics.lsa.umich.edu | |
| Ana Paula Miceli | anapaula@fnal.gov | |
| Aronne Merrelli | merrelli@fnal.gov | |
| Heidi Newberg | heidi@fnal.gov | |
| Dennis Nicklaus | nicklaus@fnal.gov | |
| Angela Olinto | olinto@oddjob.uchicago.edu | |
| John Peoples | peop@fnal.gov | |
| Jean Quashnock | jmq@oddjob.uchicago.edu | |
| Gordon Richards | richards@oddjob.uchicago.edu | |
| Constance Rockosi | cmr@oddjob.uchicago.edu | |
| Erin Sheldon | esheldon@umich.edu | |
| Allyn Smith | jasmith@sdss1.physics.lsa.umich.edu | |
| Albert Stebbins | stebbins@fnal.gov | |
| Chris Stoughton | stoughto@fnal.gov | |
| Douglas Tucker | dtucker@fnal.gov | |
| Andrew Waltman | awaltman@umich.edu |
callibrated objects sit on disks of fits files corrected frames archived to tapes or robot in their place are atlas images + sky frame together they can be used to reconstruct full frame --------------------- May dark run 2 nights: PSF poor star galaxy poor astrometry good 600,000 stars r' < 23.5 June dark run: 80 square degrees, nearly filled rectangle 211 < RA < 253 Should have ~800 QSO < 19 PSF much better determined Good star/galaxy separation Several examples of frames/bands/deblender Deblender works at about 1 arcsec, relies on 2-fold symmetry 25% of objects are blended ------------------------------------ Showed luminosity functions in different bands Equal numbers of stars/galaxies at 20th mag. After that galaxies drop, artifact of identification. Should be rise in galaxies at faint mags. ------------------------------------- Showed color diagrams; look reasonable ------------------------------------ plot stars vs frame number. Get blip at Pal 5 globular cluster. Early science: look for correlations in star counts --------------------- working on zero point calibration between columns, appear to be off by tenth of mag between columns.In a given column, offset is only few hundredths of magnitude. ------------------ standard fits readers should work. Documentation at 1] message 44 on sdss-obs mailing list 2] http://www.astro.princeton.edu/~rhl/photomisc each object has many different descriptors ------------------- Every bright object has mask file, which at present is difficult to decode. Eventually this will be available
In ~stebbins/bin/ look for sdssOBJfileReadCat a way to read the files
CALIBRATION
Std Stars / 40 inch calibration observations at Flagstaff
---------------------
many holes in red and blue ends
plugged some of the holes but most extreme red and blue holes
remain. Similarly with other colors.
Getting good coverage over 5 different months
---------------------
dust ring at flagstaff lead to 6/100 mag error if image
is right on top of it.
---------------------
no important differences as focus changes
---------------------
As object moves across chip, no errors [as opposed to
early runs on MT]
---------------------
On 40'', can't find some of the standard stars; should
be fixed soon. Ross 530, finding the wrong star [fixed].
---------------------
Comparisons of 24'' and 40'' agree very well
---------------------
Reilluminizing MT changed zero points by ~.2 mags.
Cleaning "whenever needed." Want it done regularly, once
a month or every other month.
---------------------
Tim McKay: variations seen in z band. Do you see this in standard
star?
Smith: See variations day to day, and from morning to evening.
Haven't looked carefully yet
Looking for WD luminosity function, good way to date age of galactic
disk. [Now gives 9.7 Gyr]. In u-g-r locus WD fall close to rest of
stars. In g-r-i subdwarfs fall out [g-r ~ 2]. Most important for
aging at L/L_solar = 10^{-5} are inseparable from ordinary stars
using Sloan colors.
Showed ways of star/galaxy separation using colors. Galaxies always "on top" in u-g-r. G stars with same metallicities line up in u-g-r: method of determining metallicities from SDSS colors. Models suggest that knowledge of colors to a few percent gives metallicities to .5 decs. USNO data shows fainter --> lower metallicity Should be able to determine metallicity of every G star in [u-band] survey to within .5 dec. Will sample thin disk, thick disk, and halo stars to see if metallicities change.
Described simulations, will soon be documented,released --------- Test: choose a shape,scale length, mag,axis ratio, position angle For 100 objects, compare "observed" signal with input, find few percent error in mag for exponential disk,20% with deV bulge if scale length is 5 pixels. --------- Used to test pipeline. Can we measure PSF magnitudes? Test deblender. --------- QUASHNOCK: Put in weird shaped galaxies, what comes out? STOUGHTON: Haven't done it. McKAY: Can you take custom object and run it through PHOTO? STOUGHTON: Yes, uses JPGTest. STEBBINS: Where to put the files? YANNY: June data is 400 GBytes STEBBINS: Might get 100GB disk to do analysis on. Discussion of how such a disk would be managed,placed? STEBBINS: FNAL Theory group might get involved in this.
Comparison of Ssx magnitudes with PHOTO, quite a big spread and even a bit of bias. All done withMay data, should get better. Part of this discrepency is also due to different mags [Kron vs Petrosian] Star likelihood on May data is completely uncorrelated (PHOTO and Ssx), again because PSF is not well-known. ------------------------ Investigated different algorithm. Flux in annuli vs magnitude. Even on May data, with 1.5'' seeing, do much better, can separate down to 21st mag. New algorithm does about as well on the May data as PHOTO does on June data. STOUGHTON: Does this mean for target selection 1.5'' seeing is good enough? Will quantify all this with June data.
Use DD/DR - 1 N^2 operation. Can go to counts-in-cell statistics for large angles 2 Advantages: -- statistics -- 55'' fiber separation loses close pairs; adaptive tiling will help here. At small distances (less than a Mpc) the correlation function is affected by the fiber separation --------------------------------- Want to study clustering vs morphology, say vs galaxy color and concentration index. Clustering of faint galaxies can be done in Southern strip. DODELSON: Stebbins, Frieman and I are writing code to get w. Would be nice to use Chris' simulations to test it, especially accounting for mask. Also need luminosity function to get accurate window function in order to compare with theory.
Milky Way has 9 DS satellites. Armandropf et al preprint : took POSS-II images and wrote software to find DS. Might also use adaptive kernel technique for searching. MCKAY: Might check on binned sky images. Can look in 3 colors TUCKER: Most evident in bluer bands.
Most work being done by Connolly/Szalay; meeting 8-24/25. Get \delta z \simeq 0.03 No other survey w/ comparable numbers w/ 5 colors ---------------------- Want to obtain redshifts for high z galaxies to calibrate estimator. LCRS galaxies, Johnson color data , observed fields w/ previously measured redshifts. Ultimately get spectra for 1500 objects down to r ~ 22, need 4m telescope. ---------------------- How to do it: 1] fit polynomial to colors Did this @ Michigan ; got the right numbers for LCRS galaxies. 2 runs on Schmidt in 5 SDSS colors, get 2600 galaxies with colors. Data useful in of itself, i.e. for looking at stellar locus. 2] template fitting [Connolly] Goal of JHU meeting to get official SDSS color-z technique. YANNY: Problem if no detection in u-band MCKAY: Only need u at low redshift, where you will probably get them anyway. QUASHNOCK: How far out will the relation be valid?
Many things can produce non-random orientations. Use stars to measure instrumental distortions, then remove these from galaxies. Need to account for shear -- all size change shape the same -- and smear -- change of shape depends on size. SDSS desgined to have little distortion, little shear. To measure PSF, PHOTO spits out (Q,U) for every object [weighted momnts of brightness]. Stability in space and time is very important. In each column, there is variation across chip in amplitude on the order of 5%. --- NEWBERG:How does this affect star/galaxy separation? --- PSF changes as a function of time, but changes are correlated across whole camera. ---------------------------- Even though SDSS does not go deep, statistics are fine because of large area. Clusters at z ~ 0.1 will give most of the lensing. ---------------------------- Science: 1] clusters Johnstone has developed simulations to test codes and predict signals. Results: 10 sigma detections of objects with velocity dispersions of 800 km/sec assuming 1'' PSF (Gaussian). Elliptical objects circularized by PSF, amount called "dilution factor." If galaxy size is .7'', dilution factor becomes large at seeing > 1.5''. Various weighting schemes give different dilutions but similar S/N. Use adaptive wieghting schemes to get same dilution for all S/N. Advocate using this adaptive weighting for other things, i.e. galaxy shapes. Might be able to do this with atlas images. If not, PHOTO could do it. YANNY: If you want PHOTO to change, tell us soon. Will get through May/June data in next few months. If there is signal, can correlate with known object or with dark object. All these subject to large scale systematics.
can determine mass profile of galaxies: halo profile. This is less prone to systematic effects since instrumental effects on small scales are small and there are so many galaxies that they wash out random instrumental effects. Used Joffre's software SExtractor to get prob of given object being star. Looked at 14000 galaxies from May data. Expect distribution of angles to be uniform. Plot -- for galaxies separated by 2-4'' -- maximum separation of observed cumulative distribution and random cumulative distribution. They don't agree but the result is opposite what a "signal" would produce. Hard to know how to analyze this. For galaxies separated by 6-24'' there is no difference between random and observed.
Fellow SDSS Participants
Now that SDSS data exists there are many people in the Chicago area who are starting to look at it as well as trying to glean some science from this data. We are organizing an informal get together at Fermilab on (somewhat along the lines of the meetings going on in Princeton) where people may discuss what they are doing with the data, and what they have found, both in terms of science and data quality. This would be a one-day meeting, aimed mostly at Chicagoland residents, but of course open to anyone in the collaboration.
The specifics are
At the moment all we would like from you is an indication if you think you might attend and also if there is something you would like to present - or want to hear presented. Please reply by e-mail to "ssc@simone.fnal.gov". Expect a 2nd announcement with more specifics in a week or two which will also be sent to the "sdss-general" mailing list. A web site with URL
will be set up soon for up-to-the-minute information on the agenda.
Fellow SDSS Participants
This is the 2nd and final announcement for the first Sloan Science in Chicago meeting, date and time above. Further details can be found at the web site
The meeting is informal and should contain both presentations and round-table discussion. Where you think it might be useful we would encourage anyone to bring up any unfinished work or incomplete ideas. A rough (and evolving) agenda is as follows
Status of the First Light data Brian Yanny
Use of the Sloan Data Simulations Chris Stoughton
for Science
Star-Galaxy Separation Jon Loveday
Using the Sloan to Find Dwarf Douglas Tucker
Spheroidals in the Galactic Halo
Photometric Redshifts Tim Mckay
Weak Lensing Tim Mckay
Galaxy-Galaxy Lensing Albert Stebbins
A few people have expressed interest in talking about stars - but have not quite come out of the closet. Any additions before the meeting will noted on the web page.
SPECIAL NOTE: The good news is that we are able to use the rather classy new 1 East conference room. The bad news is that absolutely no food or drink are allowed in this room. Don't despair though - since the cafeteria is only steps away one can easily pop out for a quick cup of coffee.
The list of participants is as follows
Scott Dodelson
Phil Fischer
Evalyn Gates
Don Lamb
Jonathan Loveday
Tim Mckay
Heidi Newberg
Dennis Nicklaus
Angela Olinto
John Peoples
Jean Quashnock
Gordon Richards
Constance Rockosi
Bob Rosner
Erin Sheldon
Allyn Smith
Albert Stebbins
Chris Stoughton
Douglas Tucker
Andrew Waltman
If you are planning to attend and your name is not on the list please send us an e-mail. Our address for this or any other questions you might have is