Supernova Survey

SDSS Supernova Survey

The Sloan Digital Sky Survey (SDSS) uses a wide-field, 2.5-meter telescope at Apache Point Observatory to survey the sky (for general information about the SDSS survey, see www.sdss.org ). The SDSS II Survey, a 3-year extension of the original SDSS, began operations in July 2005 and will finish in July 2008. One of the three components for SDSS II was a time-domain survey, involving repeat imaging of the same region of sky every other night. A primary scientific motivation for the time-domain survey was to detect and measure light curves for a large number of supernovae. Through repeat scans of the SDSS Southern equatorial stripe (about 2.5 deg wide by ~120 deg long) over the course of three 3-month campaigns (Sept-Nov. 2005-7), we discovered and measured multi-band lightcurves for ~500 spectroscopically confirmed Type Ia supernovae (look here for a gallery of images) in the redshift range z=0.05-0.4. In addition, we harvested a few hundred light curves for SNe Ia that we did not have the opportunity to spectroscopically confirm, but for which we are measuring host-galaxy redshifts after the fact. The survey also discovered on the order of 100 spectroscopically confirmed core-collapse supernovae (supernova types Ib/c and II). All SN candidates were made publicly available instantly via the web and via VOEventNet. All confirmed and likely SNe were announced rapidly in Central Bureau for Electronic Telegrams circulars. This sample, with excellent photometric calibration, should provide new constraints on the dark energy and insights into systematics of SNe Ia as calibrated standard candles. It complements recent surveys that have targeted both lower and higher redshift regimes. The collaboration carried out rapid spectroscopic follow-up on other telescopes for the vast majority of high-quality SN Ia candidates, with the aim of confirming SN type and determining host galaxy redshift. In addition, we carried out multi-epoch spectrophotometry for a subsample, with the aim of deriving improved SN K-corrections and studying the spectroscopic diversity of the SN Ia population.

A brief synopsis of the survey campaigns is given below. For more details about the SDSS SN Survey, see Frieman, et al., Astronomical Journal, vol. 135, p. 338 (2008) and Sako, et al., Astronomical Journal, vol. 135, p. 348 (2008).

Fall 2007 Campaign

The third of 3 time-domain campaigns was carried out from Sept. 1 to Nov. 30, 2007, alternating between the northern and southern halves of SDSS stripe 82 on the southern celestial equator. The Fall 2007 campaign yielded approximately 171 spectroscopically confirmed Ia's and handfuls of other supernova types; confirmed supernovae were announced in CBAT telegrams. Analysis of this sample is on-going.

Fall 2006 Campaign

The second of 3 time-domain campaigns was carried out from Sept. 1 to Nov. 30, 2006, alternating between the northern and southern halves of SDSS stripe 82 on the southern celestial equator. The light curve fits and information from likely host galaxy candidates (where available) was used to select candidate supernovae for follow-up spectroscopy, which was carried out at HET, NTT, ARC, NOT, MDM, Subaru, KPNO, and Keck. In addition to the SDSS telescope, follow-up imaging in SDSS filters was obtained for a subsample of confirmed Ia's using the NMSU 1m, MDM, VATT, the ARC 3.5m, WIYN, and Maidanek. The Fall 2006 campaign yielded approximately 193 spectroscopically confirmed Ia's and handfuls of other supernova types; confirmed supernovae were announced in CBAT telegrams. Analysis of this sample is on-going.

Fall 2005 Campaign

The first of 3 time-domain campaigns was carried out from Sept. 1 to Nov. 30, 2005, for the most part alternating between the northern and southern halves of SDSS stripe 82 on the southern celestial equator. A new compute cluster was installed at APO with 10 dual processors to carry out photometric reductions (using the PHOTO pipeline), carry out frame subtraction in g, r, and i filters using co-added template images from previous years of imaging on the same region, and positionally match detected objects in the different filters. This on-mountain processing typically took about 20 hours; information on new objects was transferred to Fermilab for web-based manual scanning by the collaboration. Interesting objects were matched over multiple epochs and fit with preliminary light curves based on spectroscopic templates. The light curve fits and information from likely host galaxy candidates (where available) was used to select candidate supernovae for follow-up spectroscopy, which was carried out at HET, ARC, WHT, MDM, Subaru, and Keck. In addition to the SDSS telescope, follow-up imaging in SDSS filters was obtained for a subsample of confirmed Ia's using the UH 88in, the NMSU 1m, MDM, VATT, WIYN, and the ARC 3.5m. In addition, the Carnegie Supernova Project followed up a subsample in optical and NIR filters. The Fall 2005 campaign yielded approximately 130 spectroscopically confirmed Ia's, another 13 spectroscopically likely Ia's, and handfuls of other supernova types; confirmed supernovae were announced in CBAT telegrams. Analysis of this sample is on-going.

For PR about the 2005 campaign, see Astronomy magazine on-line and KICP Research Highlight page.

Fall 2004 Campaign

To prepare for SDSS II, we carried out an early science and engineering run of the SDSS Supernova Survey in Fall 2004. Approximately 20 nights of imaging of one half of SDSS stripe 82 were scheduled between Sept. 21 and Nov. 15, with a cadence of roughly every other night. At APO, a prototype, recycled compute cluster with 7 dual processors automatically processed the data: producing corrected frames with the PHOTO pipeline, carrying out frame subtraction between new images and template images of the same region (using high-quality SDSS data from previous years), and inserting detected candidates into a database. For this run, we processed two filters (g and r) through frame subtraction and required positional matches between the two (the g and r exposures for a given field are separated by about 5 minutes, so this removes fast-moving asteroids). This processing took approximately 48 hours for a full night of SDSS imaging on the prototype cluster. For SDSS II, a new cluster that processes 3 filters (gri) in 24 hours is now in place.

For this season, we carried out spectroscopic follow-up using the DIS spectrograph on the ARC 3.5m telescope and queue-scheduled target of opportunity observations with the Hobby Eberly Telescope. Additional imaging in SDSS gri was carried out using the NMSU 1m and the ARC 3.5m. Preliminary results from this test run are summarized in our contribution to the Texas Symposium proceedings on astro-ph, April 2005.