FALL 2005 SDSS-II PROJECT SCIENTIST REPORT

Jim Gunn

SDSS sky coverage to date

ImagingImaging

SpectroscopySpectroscopy

Toward SDSS-II: Where we are

1. The DA system

2. Supernovae

3. SEGUE

4. Legacy

5. Calibration

The New DA System

FEATURES:

1. New Power-PC based VME single-board machines, one for eachlogical dewar—10 in all. Approximately 40x faster, muchlarger memory

2. New Linux host machine and file server, ~3TB of storage.Data now is transferred to the host as soon as it is takenand stored in a standard Linux file system where it can beaccessed immediately for QA and diagnostics.

3. The data are transferred over the new microwave link toFNAL and thence to Princeton for processing and storage.We no longer use tapes. Backups are made on removabledisks on the mountain.

The New DA System : Status

Installation work began in middle August.

After a few bitten fingernails, We observed (!) with it in the earlySeptember run, and have been since.

Small problems continue to be solved, but we have lost NOdata to DA problems.

The near-real time capability has already been used effectivelyto diagnose problems

The computing folks at FNAL deserve many heartfelt kudos forthe quick and excellent development, as do the observers for finding workarounds to the problems which have surfaced in real time.

SDSS SN Science Goals● Obtain ~200 high-quality SNe in the `redshift desert’,

z~0.05-0.35: continuous Hubble diagram● Probe Dark Energy in z regime arguably less sensitive to

evolution than, and complementary to, deeper surveys● Study SN Ia systematics with high photometric accuracy● Search for additional parameters to reduce Ia dispersion● Determine SN/SF rates/properties vs. z, environment● Rest-frame u-band templates for z >1 surveys ● Database of Type II and rare SN light-curves (large

survey volume with multi-band coverage)

SN Program: 2.5m Imaging● Repeat imaging of ~2.5ox120o deg. region along celestial equator (SDSS

stripe 82) for three 3-month runs (Sep-Nov. 05-’07)● Alternate every other night between strips 82N and 82S, giving dense

sampling, early detection, high-quality light-curves● Why stripe 82? repeated imaging in SDSS I allows veto catalog of variables; more accurate photometry for calibration; deeper template from co-add Use Fall months: complement Legacy, SEGUE programs Wide follow-up available from N and S hemispheres● Frame subtraction in SDSS gri allows SN Ia selection in <24 hrs

Fall 2004: Early Science & Test Run● Imaging: 20 nights of SDSS 2.5m scheduled every other night late

Sept.-mid Nov., covered half the survey area: ~1/2 the nights were useable.

● Follow-up spectroscopy: ARC 3.5m, HET 9.2m● Follow-up imaging (during/after run): NMSU 1m, ARC 3.5m● Science Goal: ~10 well-measured SN Ia light-curves with confirmed

spectroscopic types and redshifts. ● Yield: 16 confirmed Ia’s: 0.05<z<0.32 with z = 0.15, 5 Type II, 1 luminous Type Ic; 8 confirmed Ia’s found before peak;

+~10 more likely (spectroscopically unconfirmed) Ia’s ● Engineering goals met:● Rapid processing and selection of candidates in g,r using prototype

compute cluster on-mountain ● Coordinated follow-up observations ● Studied detection efficiency and photometric accuracy under varying

conditions (including moon)

SN Ia z=0.0513

3 epochs of ARC 3.5m

spectroscopy

SN2004ie

• Development Highlights:

Frame subtraction: improved diagnostics, remapping, masked

pixels, PSF determination (from PHOTO), convolution, object

finding, using co-added template images, added i band

Database: improved veto & star catalogs

`real-time’ efficiency tests with artificial SNe in data stream

Multi-color SN target selection using multi-band light curve

fitting with and without host photo-z or spectro-z: pre-typing

Web interface for human scanning of SN candidates

Public webserver for confirmed candidates

Hardware: 10 faster dual processors at APO: process gri in ~18 hrs • More extensive follow-up program• More experienced team w/ a few new faces

Improvements for 2005 Run:

SDSS II SN Follow-up 2005● Spectroscopy: SN typing, redshift, (multi-epoch spectrophotometry for improved K-corrections & sub-typing)● NIR imaging: extinction/reddening and low-z light curves● Optical imaging: follow high-z light curves beyond SDSS limit

● Spectroscopy: ARC 3.5m (31 half-nights), HET (>60 hrs), MDM 2.4m (~37 nights), Subaru (share 6 nights), WHT (6 nights), Supernova Factory (low-z targets); SALT proposed● NIR imaging: Carnegie Supernova Project (selected targets)● Optical imaging: NMSU 1m, MDM, UH 88in (6.5 nights), VATT (7 nights), WIYN (3 nights shared), INT (1 night), Liverpool Telescope (4 hours)

SN 2005 ff

z = 0.07, confirmed at WHT

Very raw Hubble diagram for 50 Ia’s

● Preliminary rough photometry, not corrected for extinction or brightness-decline relation, yet

= .25 mag!

Curve is concordance cosmology, not a fit to the data.

2005 Run: Progress Report

Sept. 1- Oct. 15 (half the ‘05 run): 38 nights were scheduled, data taken on 25 of those nights; over half of the strip area (60 deg in RA) was covered on 18 nights. Conditions range from `SDSS survey quality’ (photometric, dark, good seeing) to mixed clouds and moon. On-mountain data processing has kept up, despite occasional glitches (brand new SDSS DA system; occasional crashes of new SN compute cluster). Rotating team of ~15 scanners have examined 88,363 objects that appeared in subtracted frames and had passed software cuts (which removed fast-moving asteroids, known stars, artifacts, etc), and found 14,749 of those potentially interesting --> 5324 distinct SN candidates which were run through light-curve fitting code to select spectroscopic targets. Results: 80 SN candidates targeted for spectroscopy, 52 spectroscopically confirmed SNe Ia (including 7 `probable’ Ia’s), a handful more unconfirmed but likely Ia’s, 2 confirmed type II, 1 confirmed luminous Ibc (hypernova); 4 CBET `circulars’ released; z-range: 0.05-0.37. Avg. of 6 SDSS observations per confirmed SN so far (+ other optical follow-up). Very high Ia efficiency.

http://www-sdss.fnal.gov:8000/sdssdp/supernova-dp/sdsssn.html

SEGUE Spatial structure, kinematics, chemical properties of the old stellar populations in the Galaxy:

1. The halo: velocity dispersion ~130 km/sec, density rough

power law, roughly spherical, metal-poor.

2. The thick disk: velocity dispersion ~40km/sec, scale height

~1 kpc, intermediate metallicity, origin veryuncertain

3. The old thin disk: velocity dispersion ~20km/sec, scale height

~200 pc, high metallcity.

Chemical history, most metal-poor stars, streams and othermerger remnants.....

How was the Galaxy put together?

SEGUE IMAGING, Galactic Coordinates

~4000 sq. deg.; Approximately 1/3 done, head start in 2004.

SEGUE SPECTROSCOPY~400 plates, 240,000 stars, head start in 2004, very little spectroscopic data yet this year. (Supernovae). Will begin in earnest in November, and share time with Legacy in the spring.plates are in a long-exposure/short-exposure pair.Spectroscopic Categories:

White dwarfs: colors, 25 per plate pairCool white dwarfs: color, proper motion, per plate pairBHB/A: colors, 150 per plate pairF turnoff, metal poor: colors, 150 per plate pairG main sequence: color range, sparse sample, 375 per

pairK giants: color, proper motionM dwarf :

regular, subdwarf, high velocity: colors and PM (50 per pp)

AGB red candidates: color, 15 per plate pair

SEGUE Spectroscopic targets in color space

SEGUE Target Selection StatusStill some evolution in the target selection, reflecting some uncertainties in the success of the current algorithms.

Most target categories are now firm, at least for targeting athigh galactic latitude.

Primary problem is that a lot of data must be acquired and analyzed before we are sure of success, and the software thatmeasures the parameters that we are interested in requirelarge samples to develop. Convergence seems nigh, however,for the high latitude samples.

At low latitudes, work remains on the photometry, and generallyagreed NOT to attempt to select targets which require sophisticated color information. Probably G dwarfs and blind K giant (simplecolor and PM only) will be only targets at low latitude.

SEGUE Software Status

Version 5 of the spectroscopic pipeline, which includes much improved spectrophotometry and stellar radial velocities, is readyto run and test.

Parameter estimation code is still under development, and itmust be made production quality and incorporated into theprocessing pipeline (almost certainly as a separate pipeline step)

Timescale approximately 1 year; data releases both internal andexternal before then will be ad hoc results of interim code.

Interim, `test' reductions are keeping up with data acquisition.

LEGACY

Still about 400 square degrees of imaging (winter/early spring),and about 400 plates to finish in the north, primarily to fill in thegap in the center of the survey but also small missing areaselsewhere.

We will begin Legacy observing in December.

Calibration Two new efforts: Apache Wheel and ubercalibration

Apache wheel is a system of fast (~8x normal scan rate)binned scans which create a network of well-

calibratedphotometry across the survey.

Ubercalibration uses all overlaps in the survey, includingstrip-stripstripe-stripeoblique crossing scans-normal scansApache wheel scans-normal scansApace wheel—Apache wheel scans

to create a global calibration. In the northern contiguous patchof the survey, it has been demonstrated to provide photometrywith 1% errors in g, r, i, z; 2% in u in a prototype application, withoutbenefit of Apache Wheel.

SDSS sky with Apache Scans

Calibration: StatusApache Wheel:

Observations: Probably done, if all current data which issuspected good actually is.

Photometric Pipeline modified appropriately andtested on binned data. Preparing now to reduce all Apache data,timescale ~ 1 month.

Ubercalibration:

As soon as Apache data are available, tests on full extant dataset will be done; test version of code is ready, still need toreduce level of manual control.

Integrating ubercalibration outputs into official data productsis still being explored.