1: 

2: 

3: 

4: \documentclass[useAMS,usenatbib]{mn2e}

5: 

6: \usepackage{mathptmx,subfigure,graphicx,lpic, color}

7: 

8: 

9: %%%%%%%%%%%%%%%%%%%%%%%%%%%%

10: 

11: \newcommand\aj{{AJ}}%

12:           % Astronomical Journal

13: \newcommand\araa{{ARA\&A}}%

14:           % Annual Review of Astron and Astrophys

15: \newcommand\apj{{ApJ}}%

16:           % Astrophysical Journal

17: \newcommand\apjl{{ApJ}}%

18:           % Astrophysical Journal, Letters

19: \newcommand\apjs{{ApJS}}%

20:           % Astrophysical Journal, Supplement

21: \newcommand\ao{{Appl.~Opt.}}%

22:           % Applied Optics

23: \newcommand\apss{{Ap\&SS}}%

24:           % Astrophysics and Space Science

25: \newcommand\aap{{A\&A}}%

26:           % Astronomy and Astrophysics

27: \newcommand\aapr{{A\&A~Rev.}}%

28:           % Astronomy and Astrophysics Reviews

29: \newcommand\aaps{{A\&AS}}%

30:           % Astronomy and Astrophysics, Supplement

31: \newcommand\azh{{AZh}}%

32:           % Astronomicheskii Zhurnal

33: \newcommand\baas{{BAAS}}%

34:           % Bulletin of the AAS

35: \newcommand\jrasc{{JRASC}}%

36:           % Journal of the RAS of Canada

37: \newcommand\memras{{MmRAS}}%

38:           % Memoirs of the RAS

39: \newcommand\mnras{{MNRAS}}%

40:           % Monthly Notices of the RAS

41: \newcommand\pra{{Phys.~Rev.~A}}%

42:           % Physical Review A: General Physics

43: \newcommand\prb{{Phys.~Rev.~B}}%

44:           % Physical Review B: Solid State

45: \newcommand\prc{{Phys.~Rev.~C}}%

46:           % Physical Review C

47: \newcommand\prd{{Phys.~Rev.~D}}%

48:           % Physical Review D

49: \newcommand\pre{{Phys.~Rev.~E}}%

50:           % Physical Review E

51: \newcommand\prl{{Phys.~Rev.~Lett.}}%

52:           % Physical Review Letters

53: \newcommand\pasp{{PASP}}%

54:           % Publications of the ASP

55: \newcommand\pasj{{PASJ}}%

56:           % Publications of the ASJ

57: \newcommand\qjras{{QJRAS}}%

58:           % Quarterly Journal of the RAS

59: \newcommand\skytel{{S\&T}}%

60:           % Sky and Telescope

61: \newcommand\solphys{{Sol.~Phys.}}%

62:           % Solar Physics

63: \newcommand\sovast{{Soviet~Ast.}}%

64:           % Soviet Astronomy

65: \newcommand\ssr{{Space~Sci.~Rev.}}%

66:           % Space Science Reviews

67: \newcommand\zap{{ZAp}}%

68:           % Zeitschrift fuer Astrophysik

69: \newcommand\nat{{Nature}}%

70:           % Nature

71: \newcommand\iaucirc{{IAU~Circ.}}%

72:           % IAU Circulars

73: \newcommand\aplett{{Astrophys.~Lett.}}%

74:           % Astrophysics Letters

75: \newcommand\apspr{{Astrophys.~Space~Phys.~Res.}}%

76:           % Astrophysics Space Physics Research

77: \newcommand\bain{{Bull.~Astron.~Inst.~Netherlands}}%

78:           % Bulletin Astronomical Institute of the Netherlands

79: \newcommand\fcp{{Fund.~Cosmic~Phys.}}%

80:           % Fundamental Cosmic Physics

81: \newcommand\gca{{Geochim.~Cosmochim.~Acta}}%

82:           % Geochimica Cosmochimica Acta

83: \newcommand\grl{{Geophys.~Res.~Lett.}}%

84:           % Geophysics Research Letters

85: \newcommand\jcp{{J.~Chem.~Phys.}}%

86:           % Journal of Chemical Physics

87: \newcommand\jgr{{J.~Geophys.~Res.}}%

88:           % Journal of Geophysics Research

89: \newcommand\jqsrt{{J.~Quant.~Spec.~Radiat.~Transf.}}%

90:           % Journal of Quantitative Spectroscopy and Radiative Transfer

91: \newcommand\memsai{{Mem.~Soc.~Astron.~Italiana}}%

92:           % Mem. Societa Astronomica Italiana

93: \newcommand\nphysa{{Nucl.~Phys.~A}}%

94:           % Nuclear Physics A

95: \newcommand\physrep{{Phys.~Rep.}}%

96:           % Physics Reports

97: \newcommand\physscr{{Phys.~Scr}}%

98:           % Physica Scripta

99: \newcommand\planss{{Planet.~Space~Sci.}}%

100:           % Planetary Space Science

101: \newcommand\procspie{{Proc.~SPIE}}%

102:           % Proceedings of the SPIE

103: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%

104: 

105: 

106: 

107: 

108: 

109: 

110: \title[Supernova rate from SDSS-I]{The Rate of Type Ia Supernovae at

111:   $z \approx 0.2$ from SDSS-I

112: Overlapping Fields}

113: 

114: \author[Horesh et al.]

115: {A.~Horesh$,^{1}$

116: D.~Poznanski,$^{2}$

117: E.~O.~Ofek$^{3}$ and

118: D.~Maoz$^{1,4}$\\

119: \\

120: $^{1}$School of Physics and Astronomy and Wise Observatory, Tel Aviv University, Tel

121: Aviv 69978, Israel\\

122: $^{2}$Department of Astronomy, University of California, Berkeley, CA 94720-3411, USA\\

123: $^{3}$Division of Physics, Mathematics and Astronomy, California

124: Institute of Technology, Pasadena, CA 91125, USA\\

125: $^{4}$Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5,

126: Firenze 50125, Italy\\}

127: 

128: \begin{document}

129: 

130: 

131: 

132: \maketitle

133: \label{firstpage}

134: \begin{abstract}

135: 

136:   In the course of the Sloan Digital Sky Survey (SDSS-I), a large

137:   fraction of the surveyed area was observed more than once due to

138:   field tiling overlap, usually at different epochs. We utilize some

139:   of these data to perform a supernova (SN) survey at a mean

140:   redshift of $z=0.2$. Our archival search, in $\sim 5 \%$ of the

141:   SDSS-I overlap area, produces $29$ SN candidates clearly associated

142:   with host galaxies.  Using the Bayesian photometric classification

143:   algorithm of Poznanski et al., and correcting for classification

144:   bias, we find $17$ of the $29$ candidates are likely Type Ia SNe.

145:   Accounting for the detection efficiency of the survey and for host

146:   extinction, this implies a Type Ia SN rate of $r_{\rm

147:     Ia}=\left(14.0^{+2.5+1.4}_{-2.5-1.1}\pm 2.5\right)\times

148:   10^{-14}~h_{70}^{2}~{\rm yr}^{-1}~L^{-1}_{\odot,g}$, where the

149:   errors are Poisson error, systematic detection efficiency error, and

150:   systematic classification error, respectively. The volumetric rate

151:   is $R_{\rm Ia}=\left(1.89^{+0.42+0.18}_{-0.34-0.15}\pm

152:     0.42\right)\times 10^{-5}~{\rm yr}^{-1}~h_{70}^{3}~{\rm

153:     Mpc}^{-3}$. Our measurement is consistent with

154:   other rate measurements at low redshift. An order of magnitude

155:   increase in the number of SNe is possible by analyzing the full

156:   SDSS-I database.

157: 

158: \end{abstract}

159: 

160: 

161: \begin{keywords}

162: Supernovae: general

163: Cosmology: observations, miscellaneous

164: Surveys

165: \end{keywords}

166: 

167: \section{Introduction}

168: 

169: Supernovae (SNe) play a central role in galaxy evolution and cosmic

170: metal production. Measuring the rates at which SNe explode is thus an

171: important step for understanding the chemical evolution of the

172: universe.  In recent years, efforts have intensified to measure the

173: low-redshift Type Ia SN rate both in field environments (Cappellaro et

174: al. 1999; Hardin et al.  2000; Madgwick et al. 2003; Blanc et al. 2004,

175: Botticella et al. 2008, Dilday et al. 2008) and in galaxy clusters

176: (Gal-Yam et al. 2002; Maoz \& Gal-Yam 2004; Sharon et al. 2007;

177: Mannucci et al. 2008; Graham et al. 2008; Sand et al. 2008).  

178: However, due to small SN numbers, there are

179: still significant uncertainties in low-redshift SN rates.

180: 

181: 

182: In this paper, we demonstrate that a large archival repository of SNe,

183: one that is potentially useful for a low-redshift rate measurement

184: using a large number of SNe, exists in the data from the first phase

185: of the Sloan Digital Sky Survey (SDSS; York et al. 2000). We use a

186: small fraction of these data to detect and compile a sample of SNe Ia

187: and to derive the SN Ia rate at low-redshift. The techniques we use in

188: this paper are also of relevance for future projects such as the

189:  Panoramic

190: Survey Telescope and Rapid Response System (Pan-STARRS; Kaiser 2004)

191: and the Large Synoptic Survey Telescope (LSST; Tyson 2002).

192: These projects will survey huge areas in a relatively short time, and

193: will produce large samples of SNe for which spectral

194: classification will not be possible, due to their large numbers.

195: 

196: In \S2 we describe the SDSS data we use.  The pipeline used to process

197: these data and detect SNe is presented in \S 3.  Detection efficiency

198: and photometric calibration are discussed in sections $4$ and $5$. \S

199: $6$ presents our preliminary results, including a first SN sample, its

200: classification, and a calculation of the SN Ia rate. We compare our

201: results to previous measurements in \S7, and summarize in

202: \S 8.

203: 

204: 

205: \section{SDSS Imaging Data}

206: 

207: The SDSS imaged about one quarter of the sky in five bands ($u$, $g$,

208: $r$, $i$, $z$, centered at 3551\AA, 4686 \AA, 6165 \AA, 7481 \AA, 8931

209: \AA; Fukugita et al. 1996). Images were photometrically (Tucker et al.

210: 2006) and astrometrically (Pier 2003) calibrated by the SDSS pipeline

211: (Lupton et al. 2001). The data products of the SDSS (images and object

212: catalogs) were made available\footnote{http://www.sdss.org} in a

213: series of Data Releases (see Adelman-McCarthy et al. 2008 for a

214: description of the latest data release, DR6).

215: 

216: 

217: With the objective of covering the survey area once, imaging was

218: performed by scanning the sky in great circles. Each scan was along a

219: $2.5$-degree-wide strip, where each strip was divided into numerous

220: ``fields''. However, dividing the celestial sphere on to rectangular

221: planes causes the rectangles to overlap, especially close to the poles

222: of the survey scan coordinate system (see e.g. Fig. 1). In addition,

223: adjacent strips have an intentional overlap for the purpose of

224: photometric and astrometric quality checks. The fact that different

225: strips were imaged at different times raises the possibility of using

226: the overlap regions to detect transient events.

227: \setcounter{figure}{0}

228: \begin{figure}

229: \includegraphics[width=0.48\textwidth]{f1.eps}

230: \caption{Examples of overlap between SDSS-I fields centered around

231:   RA=14:56:49, Dec=+10:57:48. The dashed polygon is an overlap between

232:   two adjacent strips which is used for SDSS quality checks and the

233:   solid polygon is an overlap resulting from the mapping of the

234:   celestial sphere on to rectangles.}

235: \end{figure}

236: 

237: 

238: Each final SDSS field is an image of $2048\times1361$ pixels. The

239: image pixel scale is $0\farcs 396$ with a median point spread function

240: (PSF) of $1\farcs 4$ in the $r$ band. An exposure time of $53.9$ sec

241: was used to image all fields, resulting in point source AB magnitude

242: $95\%$ repeatability limits of $22.0,~22.2,~22.2,~21.3$ and $20.5$, in

243: the $u,~g,~r,~i,~z$ bands, respectively.

244: 

245: 

246: 

247: 

248: \section{Supernova Survey Data Pipeline}

249: 

250: To deal with the vast amounts of data in the SDSS database, we

251: developed a largely automatic pipeline for downloading individual

252: subsets of overlapping field images, and processing them one at a

253: time. Our pipeline consists of three independent modules for download,

254: registration, and detection, executed in that order.

255: 

256: As a first step, we compiled a list of overlapping SDSS fields. To do

257: so, we downloaded the coordinates of all the fields in the SDSS DR4

258: database. By applying a polygon intersection algorithm, which assumes

259: planar geometry, to the list of coordinates, we constructed a list of

260: the overlapping regions of each SDSS field. Each image set, consisting

261: of a first-epoch image (the ``reference'' epoch) and its overlapping

262: second epoch images, was individually downloaded for further

263: processing by our pipeline. In the present paper, we search for SNe in

264: the region $220^{\circ}<{\rm RA}<240^{\circ}$, and $-1^{\circ}<{\rm

265:   Dec}<64^{\circ}$. This region is not far from the pole of the SDSS

266: coordinate system (${\rm RA}=275^{\circ}$, ${\rm Dec}=0^{\circ}$),

267: resulting in a large overlap area of $92~{\rm deg}^{2}$, obtained from

268: $460~{\rm deg}^{2}$ of SDSS images.

269: 

270: 

271: Our survey search method is based on image subtraction. We note that

272: an alternative method is to search for SNe in the SDSS catalog using

273: different criteria, e.g., SN colors (see Poznanski et al. 2002).

274: However, a SDSS catalog SN search has some disadvantages. For example,

275: a blind color search will be affected by color contamination

276: originating from SN host galaxies. In addition, lacking a direct

277: access to the SDSS pipeline makes it difficult to estimate the survey

278: detection efficiency function.

279: 

280: 

281: 

282: We chose to limit our SN search to the $g$ and $r$ bands since they

283: are the deepest bands in the SDSS. Furthermore, the scanning order of

284: each field in the SDSS is $r,~i,~u,~z,~g$. Therefore, the $r$- and

285: $g$-band exposures of the same field in a given scan have the largest

286: time separation, i.e., there is a $\sim 5$ minute difference between

287: exposures of the same field in these bands. This time difference is

288: critical for identifying and excluding solar-system objects from among

289: the SN candidates.

290: 

291: 

292: 

293: Two computers were used for running our pipeline. One computer was

294: used for continuous downloading of images from the SDSS database. In

295: parallel, the registration and detection modules (see below) were run

296: on a computer with a Pentium IV 3.4 GHz processor and 2 GB of memory.

297: The download rate and the processing rate both dictated a net data

298: flow rate of about 1 deg$^2$ per day. In practice, software, hardware,

299: and communications problems resulted in a lower rate, and guided our

300: decision to stop the current search after about 90 deg$^2$.

301: 

302: 

303: 

304: \subsection{Image registration}

305: 

306: The registration module aligns the overlapping images in each set to

307: their reference image and produces a difference image in which SN

308: candidates are searched for by the detection module. Both the $g$ and

309: $r$ image sets, once downloaded, are registered separately by the

310: registration module. An overlapping image is first registered to its

311: reference image, based on their world coordinate system (WCS)

312: coordinates, using the $wregister$ IRAF (Tody 1986) task. The

313: overlapping areas are then cut out from both the reference and

314: registered images.  Using each field's photometric parameters,

315: extracted from the SDSS

316: database\footnote{http://casjobs.sdss.org/casjobs}, we next linearly

317: match the background level and the zeropoint of the registered image

318: to those of the reference image. At this point, we try to achieve a

319: better image registration, which is crucial for image subtraction, by

320: matching the positions of objects which appear in both images. We

321: first detect objects in the overlapping image segments by applying the

322: Source Extractor (SExtractor) program (Bertin \& Arnout 1996) to both

323: segments. Next, by cross correlating the object positions, a more

324: precise registration between the two segments is obtained using the

325: $geomap$ and $geotran$ IRAF tasks, allowing for offsets in the $x$ and

326: $y$ axes and a rotation angle between the two images. In order to

327: avoid poor statistics in the matching process, this latter alignment

328: is performed only if there are at least seven matching objects.

329: Otherwise this stage is skipped.

330: 

331: 

332: 

333: 

334: Next, the image with the smaller PSF full width at half maximum (FWHM)

335: is degraded by convolving it with a ${\rm 2D}$ Gaussian kernel,

336: $G(x,y)\propto {\rm exp}[-(x^2+y^2)/2\sigma^2]$, in order to match the

337: PSF of the second image. The kernel is found from the parameters of

338: the two image PSFs, listed as ``psfWidth'' in the SDSS catalogue

339: ``Field'' Table. This simplistic PSF matching approach is dictated by

340: the small number of objects (generally not point sources) in the

341: overlapping regions, which prevents the application of more

342: sophisticated PSF-matching algorithms (e.g. Alard \& Lupton 1998;

343: Alard 2000).

344: 

345: 

346: Following subtraction of two registered images, the absolute values of

347: the difference image is formed, so that all residuals are positive. In

348: order to smooth out residuals due to imperfect alignment, the

349: difference image is smoothed by convolving it with a 2D Gaussian,

350: three pixels wide $(1\sigma)$.

351: 

352: 

353: \subsection{SN candidate detection}

354: 

355: 

356: The residuals are detected in the difference image by applying

357: SExtractor to the image. Since the final difference image is positive

358: definite (see $\S3.1$), it has a one sided noise distribution. We

359: chose to apply a $6\sigma$ detection threshold in the detection

360: process. The value of the detection threshold is calculated using the

361: Poisson fluctuations of the background counts in both the reference

362: and registered images.  The residuals detected by SExtractor are

363: automatically examined in more detail to screen for various non-SN

364: detection contaminations, as described next.

365: 

366: 

367: 

368: 

369: We first search for variable stars within our candidate list. Using

370: SExtractor, we obtain a list of objects in both the registered and

371: reference images. If an object is detected in both images at the same

372: position where a residual was detected in the difference image, we

373: query the SDSS catalogue for objects at that position. If an object,

374: catalogued as a star, exists at that position, the candidate is

375: considered a variable star and is rejected from our candidate list.

376: Similarly, we reject candidates spectroscopically identified as

377: quasars.

378: 

379: 

380: 

381: 

382: We next explore the possibility that a residual is the result of poor

383: image registration. We search, using SExtractor, for positive

384: residuals in two new difference images: the reference minus registered

385: image, and vice versa. If in each of the two images a residual is

386: detected near the position of a candidate, we compare the difference

387: between the photon counts of the two residuals to our detection

388: threshold. In contrast to our original detection in the absolute value

389: of the difference image, we now require the photon count difference to

390: have at least $3\sigma$ significance. In a final test for improper

391: alignment, a stamp of $41\times41$ pixels around the position of each

392: residual is cut out of the registered and reference images. The two

393: image stamps are re-registered using the $xregister$ task in IRAF,

394: based on cross correlation. A new difference image is produced using

395: the new re-registered images. If no residuals are detected in the new

396: subtracted stamp image, the candidate is discarded.

397: 

398: 

399: 

400: 

401: The remaining candidates are subjected to another test, aimed at

402: determining whether or not they are moving objects. We first query the

403: SDSS catalogue to check if the target has been flagged as a moving

404: object. We also compare the position of each candidate in the $g$ and

405: $r$ bands, assuming that it was detected in both bands. If the

406: candidate position has changed by more than $2$ pixels, it is also

407: considered a moving object and is excluded from the candidate list.

408: All stages up to this point are performed automatically, with no human

409: intervention.

410: 

411: 

412: The remaining candidates are saved, together with their subtraction

413: images, for visual inspection, performed by a single person (AH). The

414: inspection helps reject false positives of various types, such as

415: artifacts and 

416: residuals due to poor PSF matching, poor image alignment, cosmic rays,

417: and saturated objects (see Fig. 2). About 99\% of the candidates found by the

418: automatic pipeline are discarded as false positives in the visual

419: inspection stage.

420: \setcounter{figure}{1}

421: \begin{figure}

422: \centering

423: \includegraphics[width=0.21\textwidth, height=0.2\textwidth]{f2a.eps}

424: \includegraphics[width=0.2\textwidth, height=0.2\textwidth]{f2b.eps}

425: \caption{Examples of false positives which were rejected in

426:   the visual inspection stage. The left panel shows a residual in the

427:   subtraction image due to image misalignment. The residual in the

428:   right panel is due to poor PSF matching.}

429: \end{figure}

430: 

431: 

432: \section{Detection efficiency functions}

433: 

434: Estimating the efficiency of our SN detection process is critical for

435: deriving a reliable SN rate. 

436: We have planted a sample of fake SNe in the SDSS images,

437: whose recovery fraction provides an estimate of the detection

438: efficiency as a function of SN magnitude. 

439: 

440: The first step in producing the fake SN sample was choosing the SN

441: hosts. We compiled the $g$ magnitude and the photometric redshift

442: (Csabai et al. 2003; Oyaizu et al. 2008) of each galaxy in the

443: overlapping sets of images from the SDSS catalog.  Under a simplifying

444: assumption that the SN rate is proportional to stellar $g$-band

445: luminosity, we selected a random subset of galaxies weighted by

446: luminosity. To each of these selected hosts we then assigned a fake

447: SN. The SN was assigned a random $g$-band absolute magnitude in the

448: range of $-19.5$ to $-7.5$. The absolute magnitudes were converted to

449: observed magnitudes using a distance modulus based on the SN host

450: photometric redshift (assuming a Hubble parameter of $H_{0}=70~{\rm km

451:   ~s}^{-1}{\rm Mpc}^{-1}$, a mass density in units of the critical

452: density $\Omega_{\rm m}=0.3$, and cosmological constant

453: $\Omega_{\Lambda}=0.7$). The SN $r$-band magnitude was randomly chosen

454: to be in the range of $-0.5$ and $+1.5$ of the $g$ band magnitude, a

455: range motivated by calculating synthetic $g$ and $r$ magnitudes from a

456: set of observed spectra of SNe Ia (Nugent 2002; Poznanski et al. 2002;

457: Poznanski, Maoz, \& Gal-Yam 2007).

458: 

459: 

460: 

461: The fake SNe were added to the real images as part of the data

462: processing, prior to image registration, as follows. First, we

463: randomly chose the image in which the SN was to be planted, i.e.,

464: either the reference image or the registered one. We then cut out a

465: region around the fake SN host of size $1.5$ times the host's $90\%$

466: light radius. SExtractor was applied to the host stamp image,

467: producing a list of $10\%,~20\% ,...,~100\%$ light radii of the host.

468: The radial distance of the fake SN from its host center was chosen

469: randomly from among these annuli, assuring that the locations of the

470: artificial SNe roughly follow their galaxy host light. The final

471: position of the fake SN with respect to the host was at a randomly

472: chosen position angle. The SN was then planted in the selected image

473: using the IRAF task $mkobjects$.

474: 

475: 

476: 

477: The fake SN sample underwent the same processing as the real data,

478: including the visual inspection stage, ensuring it reflected

479: faithfully the actual detection efficiency. By spreading the fake SNe

480: among all the overlapping fields, we also took into account the fact

481: that the efficiency may vary from field to field.

482: 

483: 

484: 

485: Our detection efficiency functions in the $g$ and $r$ bands are shown

486: in Figure $3$. We find that our efficiency level is $\sim 60\%$, at

487: best. This is probably due to the poor quality of image subtraction

488: when the PSF matching and/or the registration are not perfect. For

489: example, bright hosts often leave large residuals at their centers in

490: the difference images. Therefore, a real SN that is close to its host

491: center, may be mistaken for an artificial residual due to poor

492: registration, by either the pipeline or the human inspector. This

493: effect seems to be independent of magnitude, even for bright SNe, due

494: to the fact that such SNe will tend to be hosted by nearby, and hence

495: bright, galaxies. It is also evident that the detection efficiency in

496: the $r$ band starts declining at brighter magnitudes, compared to the

497: $g$ band, probably due to the relative faintness of galaxies in the

498: $g$-band.

499: \setcounter{figure}{2}

500: \begin{figure}

501:   \centering

502:   \subfigure{\includegraphics[width=0.5\textwidth]{f3a.eps}}

503: 

504:   \subfigure{\includegraphics[width=0.5\textwidth]{f3b.eps}}

505: \caption{Detection efficiency functions in the $g$ (top) and $r$ (bottom) bands. Error bars represent

506:   $1\sigma$ Poisson Errors.}

507: \end{figure}

508: 

509: 

510: \section{Photometric Calibration}

511: 

512: The SN candidates which pass the visual inspection are further

513: explored. In order to obtain the magnitudes of these candidates, we

514: download the images, in which a candidate resides, in the remaining

515: ($u$, $i$, $z$) SDSS bands. Since the images of a field in different

516: bands do not fully overlap, we first align them according to the image

517: in which the candidate was detected (in either the $g$ or the $r$

518: band). We perform this alignment for the reference field images and

519: for the registered field images separately. Then, the reference and

520: registered field images in each band are processed by the same

521: registration module used to originally process the $g$ and $r$ bands

522: (see $\S3$).  After a final difference image is obtained in all five

523: bands, $41 \times 41$-pixel images are cut out around the candidate.

524: 

525: 

526: 

527: These images are used to perform aperture photometry of the

528: candidates. The counts are summed in an aperture of radius $2\sigma$,

529: where $\sigma$ is $1/2.35$ of the FWHM of the average PSF of a field

530: (taken from the SDSS catalog ``Field'' table). With the zeropoint,

531: airmass, and extinction parameters, also listed in the SDSS catalog,

532: we convert the candidate counts to magnitudes. To each magnitude we

533: then apply an aperture correction, that accounts for the flux outside

534: the aperture.  From aperture photometry on bright and isolated SDSS

535: stars, we find mean aperture corrections of $0.21, 0.22, 0.23, 0.25,

536: 0.25$ in the $u,~g,~r,~i,~z$ bands, respectively. With these

537: corrections, our final magnitudes for bright stars also match those in

538: the SDSS catalogue.  We corrected the candidate magnitudes for

539: Galactic extinction according to Schlegel et al. (1998).

540: 

541: 

542: To obtain realistic error estimates for our magnitude measurements, we

543: have performed the same analysis on a large sample of artificial SNe.

544: The artificial SNe were blindly planted in several images in the same

545: manner as done for obtaining our detection efficiency functions (see

546: $\S4$). We expect the photometric errors to be larger than the usual

547: Poisson errors due to inaccurate registration, varying backgrounds,

548: and residuals from the host galaxy subtraction. We planted $\sim 400$

549: artificial SNe in each of eight magnitude bins (see

550: Table 1). In each bin, and for each band, we calculate the root-mean

551: square (rms) of the difference between the measured magnitudes and the

552: original magnitudes assigned to the artificial SNe, and adopt it as

553: the systematic photometric error. The results are listed in Table~1.

554: \begin{table}

555: \begin{center}

556: \caption{Photometric Errors}

557: \smallskip

558: \begin{tabular}{ccccccccc}

559: \hline

560: \noalign{\smallskip}

561: Band & \multicolumn{8}{c}{$1 \sigma$ Magnitude Errors}\\

562:      & $19$ & $19.5$ & $20$ & $20.5$ & $21$ &

563:      $21.5$ & $22$ & $22.5$\\

564: \cline{2-9}

565: \noalign{\smallskip}

566: 

567: \noalign{\smallskip}

568: $u$  & $0.10$ & $0.11$ & $0.17$ & $0.16$ & $0.17$ & $0.26$ & $0.36$ & $0.45$ \\

569: $g$  & $0.10$ & $0.13$ & $0.13$ & $0.12$ & $0.17$ & $0.17$ & $0.24$ & $0.24$ \\

570: $r$  & $0.11$ & $0.11$ & $0.17$ & $0.13$ & $0.15$ & $0.32$ & $0.30$ & $0.29$ \\

571: $i$  & $0.11$ & $0.14$ & $0.16$ & $0.22$ & $0.24$ & $0.63$ & $0.55$ & $0.50$ \\

572: $z$  & $0.18$ & $0.19$ & $0.26$ & $0.60$ & $0.62$ & $0.69$ & $0.86$ & $0.61$ \\

573: 

574: \noalign{\smallskip}

575: \hline

576: %\smallskip

577: \end{tabular}

578: \end{center}

579: \end{table}

580: 

581: 

582: 

583: 

584: 

585: 

586: \section{Results}

587: 

588: \subsection{SN sample}

589: 

590: 

591: Our final candidate list for the $92~{\rm deg }^{2}$ of overlap area

592: searched includes $47$ transient candidates which we denote as

593: SISN01 to 47, where SISN stands for SDSS-I SN. Among the candidates,

594: $25$ are clearly associated with a detected host galaxy, but offset

595: from the nucleus (if there is one). A further $11$ are, to within SDSS

596: resolution, at the centers of their hosts. A final $11$ candidates are

597: ``hostless'', i.e., cannot be unambiguously associated with any

598: detected galaxy. Our criterion for hostlessness is

599:  being separated by both $>5''$ and more than two times the

600: $90\%$ light radius from any galaxy.

601: 

602: 

603: 

604: We first turn our attention to estimate the sample contamination by SN

605: ``impostors''. Based on the SDSS limiting magnitudes (see section $\S

606: 2$), we estimate that a true SN Ia, at maximum light, will be

607: undetectable at redshifts $z>0.35$. Although there are exceptions,

608: most core-collapse SNe are less luminous than SNe-Ia. Candidates with

609: spectroscopic or photometric host redshifts with a $1\sigma$ lower

610: limit above $z>0.35$ are therefore excluded. Six of the seven

611: candidates excluded by this criteria are at their host centers, and

612: are thus likely to be active galactic nuclei (AGN), rather than SNe.

613: The seventh candidate, SISN$47$, which is not at the center of its host,

614: might be a valid SN candidate falsely rejected due to an error in the

615: photometric redshift of its host. However, if it were

616: included in our final sample it would have no effect on the total number of

617: Type Ia SNe, being classified as a likely core-collapse SN (see Table 2).

618: 

619: 

620: 

621: The hostless candidates could be either real SNe which reside in

622: galaxies below the SDSS limiting magnitude, or they can be impostors

623: such as quasars, slow-moving asteroids, and variable stars. To

624: estimate the expected fraction of SNe hosted by galaxies fainter than

625: the SDSS limiting magnitude, we use the Blanton et al.  (2003) galaxy

626: luminosity function in the $r$ band. The fraction of the stellar

627: luminosity in galaxies with a luminosity $L<L_{{\rm lim}}$ is

628: \begin{equation}

629: P(L<L_{{\rm lim}})=\frac{\int\limits_{0}^{L_{\lim}}L\phi(L)dL}{\int\limits_{0}^{\infty}L\phi(L)dL},

630: \end{equation}

631: where $\phi$ is the luminosity function. At $z=0.2$, the mean redshift

632: of SNe probed by our search (see $\S 6.3$, below), the SDSS $r$-band

633: flux limit correspond to an absolute magnitude of $M_{r}=-17.8$ mag,

634: and $P(L<L_{{\rm lim}})=23\%$. Again assuming that SNe track the

635: stellar luminosity, we therefore expect $23\% \times40 \approx9$

636: candidates in galaxies below the SDSS limiting magnitude, consistent

637: with the $11$ hostless candidates we find. Conversely, this also

638: argues that most of the hostless candidates are likely real SNe, as

639: otherwise a large deficit of SNe in low-luminosity galaxies would be

640: implied.  Nonetheless, due to the lack of redshifts for the hostless

641: candidates, we are unable to determine with great confidence which of

642: those candidates are real SNe and what are their types. We therefore

643: exclude the $11$ hostless candidates from our sample for the purpose

644: of the SN rate calculation.  The exclusion of the hostless candidates

645: is accounted for in the luminosity-normalized SN rate calculation by

646: using the luminosity density which originates from galaxies above the

647: SDSS limiting magnitude (see $\S6.3$).

648:  

649: 

650: 

651: Kuiper belt objects (KBOs) might also play a role as SN impostors.

652: However, they are normally found near the ecliptic, while all our SN

653: candidate have ecliptic latitude $\beta >15^{\circ}$. Moreover,

654: according to their magnitude distribution (Bernstein et al. 2004)

655: their expected number in our survey, even near the ecliptic, is at

656: least an order of magnitude lower than the observed number of SN

657: candidates. We have checked for asteroids near the positions of all

658: candidates in the Minor Planet

659: Center\footnote{http://www.cfa.harvard.edu/iau/mpc.html} and Nasa Jet

660: Propulsion Laboratory\footnote{http://ssd.jpl.nasa.gov/sbfind.cgi}

661: databases. No known asteroids were found within $5'$ of any

662: candidates. Our final main SN sample thus consists of $29$ candidates

663: with hosts, among which five are nuclear and therefore may be AGNs.

664: Figures $4,~5,~6$ show sections of the reference, registered, and

665: difference images for each candidate in the final, hostless and

666: high-$z$ samples, respectively.  

667: \setcounter{figure}{3}

668: \begin{figure}

669: %\begin{minipage}[t]{.45\textwidth}

670: \centering

671: \begin{lpic}{f4a(0.35)}

672: \setlength{\lpbgboxsep}{0.2mm}

673: \lbl[W]{15.5,51;SISN02}

674: \lbl{164,47,-121;{\huge{\color{white}$ \uparrow$}}}

675: %\lbl{164,47,315;{\huge{\color{white}$ \uparrow$}}}

676: %\lbl[W]{160,48,10;{\bf{\Large$\uparrow$}}}

677: \end{lpic}

678: \begin{lpic}{f4b(0.35)}

679: \setlength{\lpbgboxsep}{0.2mm}

680: \lbl[W]{15.5,51;SISN03}

681: \lbl{164,47,-112;{\huge{\color{white}$ \uparrow$}}}

682: \end{lpic}

683: \begin{lpic}{f4c(0.35)}

684: \setlength{\lpbgboxsep}{0.2mm}

685: \lbl[W]{15.5,51;SISN05}

686: \lbl{164,47,-99;{\huge{\color{white}$ \uparrow$}}}

687: \end{lpic}

688: \begin{lpic}{f4d(0.35)}

689: \setlength{\lpbgboxsep}{0.2mm}

690: \lbl[W]{15.5,51;SISN06}

691: \lbl{164,47,-111;{\huge{\color{white}$ \uparrow$}}}

692: \end{lpic}

693: \begin{lpic}{f4e(0.35)}

694: \setlength{\lpbgboxsep}{0.2mm}

695: \lbl[W]{15.5,51;SISN07}

696: \lbl{164,47,-106;{\huge{\color{white}$ \uparrow$}}}

697: \end{lpic}

698: \begin{lpic}{f4f(0.35)}

699: \setlength{\lpbgboxsep}{0.2mm}

700: \lbl[W]{15.5,51;SISN10}

701: \lbl{164,47,-110;{\huge{\color{white}$ \uparrow$}}}

702: \end{lpic}

703: \begin{lpic}{f4g(0.35)}

704: \setlength{\lpbgboxsep}{0.2mm}

705: \lbl[W]{15.5,51;SISN11}

706: \lbl{164,47,-90;{\huge{\color{white}$ \uparrow$}}}

707: \end{lpic}

708: 

709: 

710: %\end{minipage}

711: 

712: \caption{Final sample of SN candidates - For each candidate the reference (left),

713:   registered (center), and difference (right) images are shown. Images

714:   are $16''$ on a side. Arrows indicate north, with east to the left

715:   when facing north.}

716: \end{figure}

717: 

718: 

719: \setcounter{figure}{3}

720: \begin{figure*}

721: \begin{minipage}[t]{.45\textwidth}

722: \centering

723: 

724: 

725: \begin{lpic}{f4h(0.35)}

726: \setlength{\lpbgboxsep}{0.2mm}

727: \lbl[W]{15.5,51;SISN12}

728: \lbl{164,47,-110;{\huge{\color{white}$ \uparrow$}}}

729: \end{lpic}

730: \begin{lpic}{f4i(0.35)}

731: \setlength{\lpbgboxsep}{0.2mm}

732: \lbl[W]{15.5,51;SISN13}

733: \lbl{164,47,-110;{\huge{\color{white}$ \uparrow$}}}

734: \end{lpic}

735: \begin{lpic}{f4j(0.35)}

736: \setlength{\lpbgboxsep}{0.2mm}

737: \lbl[W]{15.5,51;SISN14}

738: \lbl{164,47,-95;{\huge{\color{white}$ \uparrow$}}}

739: \end{lpic}

740: \begin{lpic}{f4k(0.35)}

741: \setlength{\lpbgboxsep}{0.2mm}

742: \lbl[W]{15.5,51;SISN17}

743: \lbl{164,47,-115;{\huge{\color{white}$ \uparrow$}}}

744: \end{lpic}

745: \begin{lpic}{f4l(0.35)}

746: \setlength{\lpbgboxsep}{0.2mm}

747: \lbl[W]{15.5,51;SISN18}

748: \lbl{164,47,-93;{\huge{\color{white}$ \uparrow$}}}

749: \end{lpic}

750: \begin{lpic}{f4m(0.35)}

751: \setlength{\lpbgboxsep}{0.2mm}

752: \lbl[W]{15.5,51;SISN20}

753: \lbl{164,47,-93;{\huge{\color{white}$ \uparrow$}}}

754: \end{lpic}

755: \begin{lpic}{f4n(0.35)}

756: \setlength{\lpbgboxsep}{0.2mm}

757: \lbl[W]{15.5,51;SISN23}

758: \lbl{164,47,-95;{\huge{\color{white}$ \uparrow$}}}

759: \end{lpic}

760: \begin{lpic}{f4o(0.35)}

761: \setlength{\lpbgboxsep}{0.2mm}

762: \lbl[W]{15.5,51;SISN25}

763: \lbl{164,47,-117;{\huge{\color{white}$ \uparrow$}}}

764: \end{lpic}

765: \begin{lpic}{f4p(0.35)}

766: \setlength{\lpbgboxsep}{0.2mm}

767: \lbl[W]{15.5,51;SISN26}

768: \lbl{164,47,-88;{\huge{\color{white}$ \uparrow$}}}

769: \end{lpic}

770: \begin{lpic}{f4q(0.35)}

771: \setlength{\lpbgboxsep}{0.2mm}

772: \lbl[W]{15.5,51;SISN27}

773: \lbl{164,47,-116;{\huge{\color{white}$ \uparrow$}}}

774: \end{lpic}

775: \begin{lpic}{f4r(0.35)}

776: \setlength{\lpbgboxsep}{0.2mm}

777: \lbl[W]{15.5,51;SISN28}

778: \lbl{164,47,-101;{\huge{\color{white}$ \uparrow$}}}

779: \end{lpic}

780: 

781: \end{minipage}

782: \begin{minipage}[t]{.45\textwidth}

783: \centering

784: 

785: 

786: \begin{lpic}{f4s(0.35)}

787: \setlength{\lpbgboxsep}{0.2mm}

788: \lbl[W]{15.5,51;SISN29}

789: \lbl{164,47,-90;{\huge{\color{white}$ \uparrow$}}}

790: \end{lpic}

791: \begin{lpic}{f4t(0.35)}

792: \setlength{\lpbgboxsep}{0.2mm}

793: \lbl[W]{15.5,51;SISN30}

794: \lbl{164,47,-90;{\huge{\color{white}$ \uparrow$}}}

795: \end{lpic}

796: \begin{lpic}{f4u(0.35)}

797: \setlength{\lpbgboxsep}{0.2mm}

798: \lbl[W]{15.5,51;SISN31}

799: \lbl{164,47,-116;{\huge{\color{white}$ \uparrow$}}}

800: \end{lpic}

801: \begin{lpic}{f4v(0.35)}

802: \setlength{\lpbgboxsep}{0.2mm}

803: \lbl[W]{15.5,51;SISN32}

804: \lbl{164,47,-120;{\huge{\color{white}$ \uparrow$}}}

805: \end{lpic}

806: \begin{lpic}{f4ca(0.35)}

807: \setlength{\lpbgboxsep}{0.2mm}

808: \lbl[W]{15.5,51;SISN33}

809: \lbl{164,47,-115;{\huge{\color{white}$ \uparrow$}}}

810: \end{lpic}

811: \begin{lpic}{f4cb(0.35)}

812: \setlength{\lpbgboxsep}{0.2mm}

813: \lbl[W]{15.5,51;SISN35}

814: \lbl{164,47,-96;{\huge{\color{white}$ \uparrow$}}}

815: \end{lpic}

816: \begin{lpic}{f4cc(0.35)}

817: \setlength{\lpbgboxsep}{0.2mm}

818: \lbl[W]{15.5,51;SISN36}

819: \lbl{164,47,-122;{\huge{\color{white}$ \uparrow$}}}

820: \end{lpic}

821: \begin{lpic}{f4cd(0.35)}

822: \setlength{\lpbgboxsep}{0.2mm}

823: \lbl[W]{15.5,51;SISN40}

824: \lbl{164,47,-117;{\huge{\color{white}$ \uparrow$}}}

825: \end{lpic}

826: \begin{lpic}{f4ce(0.35)}

827: \setlength{\lpbgboxsep}{0.2mm}

828: \lbl[W]{15.5,51;SISN41}

829: \lbl{164,47,-114;{\huge{\color{white}$ \uparrow$}}}

830: \end{lpic}

831: \begin{lpic}{f4cf(0.35)}

832: \setlength{\lpbgboxsep}{0.2mm}

833: \lbl[W]{15.5,51;SISN44}

834: \lbl{164,47,-98;{\huge{\color{white}$ \uparrow$}}}

835: \end{lpic}

836: \begin{lpic}{f4cg(0.35)}

837: \setlength{\lpbgboxsep}{0.2mm}

838: \lbl[W]{15.5,51;SISN45}

839: \lbl{164,47,-126;{\huge{\color{white}$ \uparrow$}}}

840: \end{lpic}

841: 

842: \end{minipage}

843: \caption{[continued]}

844: \end{figure*}

845: 

846: 

847: 

848: %--------------------------------------

849: 

850: \setcounter{figure}{4}

851: \begin{figure}

852: \centering

853: \begin{lpic}{f5a(0.35)}

854: \setlength{\lpbgboxsep}{0.2mm}

855: \lbl[W]{15.5,51;SISN01}

856: \lbl{164,47,-121;{\huge{\color{white}$ \uparrow$}}}

857: \end{lpic}

858: \begin{lpic}{f5b(0.35)}

859: \setlength{\lpbgboxsep}{0.2mm}

860: \lbl[W]{15.5,51;SISN15}

861: \lbl{164,47,-114;{\huge{\color{white}$ \uparrow$}}}

862: \end{lpic}

863: \begin{lpic}{f5c(0.35)}

864: \setlength{\lpbgboxsep}{0.2mm}

865: \lbl[W]{15.5,51;SISN16}

866: \lbl{164,47,-88;{\huge{\color{white}$ \uparrow$}}}

867: \end{lpic}

868: \begin{lpic}{f5d(0.35)}

869: \setlength{\lpbgboxsep}{0.2mm}

870: \lbl[W]{15.5,51;SISN19}

871: \lbl{164,47,-97;{\huge{\color{white}$ \uparrow$}}}

872: \end{lpic}

873: \begin{lpic}{f5e(0.35)}

874: \setlength{\lpbgboxsep}{0.2mm}

875: \lbl[W]{15.5,51;SISN21}

876: \lbl{164,47,-95;{\huge{\color{white}$ \uparrow$}}}

877: \end{lpic}

878: \begin{lpic}{f5f(0.35)}

879: \setlength{\lpbgboxsep}{0.2mm}

880: \lbl[W]{15.5,51;SISN24}

881: \lbl{164,47,-115;{\huge{\color{white}$ \uparrow$}}}

882: \end{lpic}

883: \begin{lpic}{f5g(0.35)}

884: \setlength{\lpbgboxsep}{0.2mm}

885: \lbl[W]{15.5,51;SISN34}

886: \lbl{164,47,-88;{\huge{\color{white}$ \uparrow$}}}

887: \end{lpic}

888: \begin{lpic}{f5h(0.35)}

889: \setlength{\lpbgboxsep}{0.2mm}

890: \lbl[W]{15.5,51;SISN37}

891: \lbl{164,47,-96;{\huge{\color{white}$ \uparrow$}}}

892: \end{lpic}

893: \begin{lpic}{f5i(0.35)}

894: \setlength{\lpbgboxsep}{0.2mm}

895: \lbl[W]{15.5,51;SISN38}

896: \lbl{164,47,-118;{\huge{\color{white}$ \uparrow$}}}

897: \end{lpic}

898: \begin{lpic}{f5j(0.35)}

899: \setlength{\lpbgboxsep}{0.2mm}

900: \lbl[W]{15.5,51;SISN42}

901: \lbl{164,47,-96;{\huge{\color{white}$ \uparrow$}}}

902: \end{lpic}

903: \begin{lpic}{f5k(0.35)}

904: \setlength{\lpbgboxsep}{0.2mm}

905: \lbl[W]{15.5,51;SISN43}

906: \lbl{164,47,-122;{\huge{\color{white}$ \uparrow$}}}

907: \end{lpic}

908: 

909: \caption{Same as Fig. 4 but for the hostless candidate sample.}

910: \end{figure}

911: 

912: 

913: %----------------------------

914: \setcounter{figure}{5}

915: \begin{figure}

916: \centering

917: 

918: \begin{lpic}{f6a(0.35)}

919: \setlength{\lpbgboxsep}{0.2mm}

920: \lbl[W]{15.5,51;SISN04}

921: \lbl{164,47,-113;{\huge{\color{white}$ \uparrow$}}}

922: \end{lpic}

923: \begin{lpic}{f6b(0.35)}

924: \setlength{\lpbgboxsep}{0.2mm}

925: \lbl[W]{15.5,51;SISN08}

926: \lbl{164,47,-96;{\huge{\color{white}$ \uparrow$}}}

927: \end{lpic}

928: \begin{lpic}{f6c(0.35)}

929: \setlength{\lpbgboxsep}{0.2mm}

930: \lbl[W]{15.5,51;SISN09}

931: \lbl{164,47,-110;{\huge{\color{white}$ \uparrow$}}}

932: \end{lpic}

933: \begin{lpic}{f6d(0.35)}

934: \setlength{\lpbgboxsep}{0.2mm}

935: \lbl[W]{15.5,51;SISN22}

936: \lbl{164,47,-115;{\huge{\color{white}$ \uparrow$}}}

937: \end{lpic}

938: \begin{lpic}{f6e(0.35)}

939: \setlength{\lpbgboxsep}{0.2mm}

940: \lbl[W]{15.5,51;SISN39}

941: \lbl{164,47,-96;{\huge{\color{white}$ \uparrow$}}}

942: \end{lpic}

943: \begin{lpic}{f6f(0.35)}

944: \setlength{\lpbgboxsep}{0.2mm}

945: \lbl[W]{15.5,51;SISN46}

946: \lbl{164,47,-88;{\huge{\color{white}$ \uparrow$}}}

947: \end{lpic}

948: \begin{lpic}{f6g(0.35)}

949: \setlength{\lpbgboxsep}{0.2mm}

950: \lbl[W]{15.5,51;SISN47}

951: \lbl{164,47,-94;{\huge{\color{white}$ \uparrow$}}}

952: \end{lpic}

953: 

954: \caption{Same as Fig. 4 but for the high-$z$ candidate sample, i.e. the

955:   seven candidates rejected based on

956:   their high photometric host redshifts.}

957: \end{figure}

958: 

959: %-----------------------------

960: 

961: \subsection{Supernova classification}

962: 

963: %Spectroscopic classification of our SN candidates is, of course,

964: %impossible in this archival survey. We therefore adopt a photometric

965: %classification method, the Supernova Automated Bayesian Classification

966: %(SN-ABC) routine of Poznanski, Maoz, \& Gal-Yam (2007a). This method

967: %compares the SN candidate magnitudes to a sample of SN spectral

968: %templates of different types, ages, redshifts, and extinctions. The

969: %routine uses as a prior the host redshift probability distribution

970: %function of each candidate. Six candidates in our sample have

971: %spectroscopic host redshifts. For the rest of the candidates we

972: %retrieve photometric redshifts from the SDSS catalog. We then assume

973: %Gaussian redshift probability distribution functions using the host

974: %spectroscopic or photometric redshifts and their errors.

975: 

976: 

977: 

978: 

979: 

980: %The SN-ABC returns the probability of a candidate being a Type Ia SN,

981: %$P({\rm Ia})$, as opposed to being a core-collapse SN. We consider

982: %candidates which have $P({\rm Ia})>0.9$ as likely type Ia SNe. We

983: %chose this $P({\rm Ia})$ threshold in order to minimize the

984: %misclassification of Type II SNe as Type Ia (see more details below).

985: % Out of the $29$ candidates in our

986: %sample, $16$ are classified as Ia's, with $\chi^{2}$ values in the

987: %range $0.3-10.6$. Table $2$ list the candidate properties, including

988: %SN types and host redshifts.

989: 

990: 

991: 

992: 

993: %We next perform an analysis similar to that of Poznanski et al.

994: %(2007b) in order to determine the degree of classification uncertainty

995: %and bias, i.e., what is the most probable real range in the fraction

996: %of SNe Ia, given the fraction that was classified as such. Using the

997: %SN spectral templates, we create a sample of fake SNe of the four most

998: %prevalent types: Ia, IIP, IIn, and Ib/c. The distribution of redshifts

999: %is according to the redshift distribution of the galaxy sample used

1000: %in the detection efficiency simulations (see $\S4$).  We note that

1001: %while Type Ia's and IIn's are observable up to a redshift of $z\approx

1002: %0.35$, given the SDSS limiting magnitude, types Ib/c and IIP are

1003: %observable only out to $z\approx 0.15$.  The fractions of simulated

1004: %SNe that are classified or misclassified as Type Ia by the SN-ABC

1005: %routine are $0.85,~0.04,~0.28$, and $0.13$ for types Ia, IIP, IIn, and

1006: %Ib/c, respectively. Since we do not know the real fraction of Type Ia

1007: %SNe at low redshift, we create a sample of possible fractions of the

1008: %different types, e.g., $30\%~{\rm Ia}, ~50\%~{\rm IIP},~ 10\%~{\rm

1009: %  IIn},~ 10\%~{\rm Ib/c}$, using steps of $10\%$. For each set of

1010: %fractions, we calculate the binomial probability of finding the

1011: %observed $16$ SNe Ia out of $29$, using the fractions of SNe

1012: %classified as Type Ia that we calculated for the four SN types. Each

1013: %of the combinations is given a weight according to the number of

1014: %permutations with the same fraction of SNe Ia. The outcome of this

1015: %calculation is a probability distribution for the real number of SNe

1016: %in our sample, given that $16$ SNe were classified as Type Ia. From

1017: %this probability function we derive that the most probable ``true''

1018: %number of SNe Ia in our sample is $N_{{\rm Ia}}=17^{+3.8}_{-3.1}\pm

1019: %3.8$, where the first error is the $68\%$ range Poisson error, and the

1020: %second one is a systematic $68\%$ range error due to the uncertainty

1021: %in classification.

1022: 

1023: 

1024: 

1025: 

1026: Spectroscopic classification of our SN candidates is, of course,

1027: impossible in this archival survey. We therefore adopt a photometric

1028: classification method, the Supernova Automated Bayesian Classification

1029: (SN-ABC) routine of Poznanski, Maoz, \& Gal-Yam (2007a). This method

1030: compares the SN candidate magnitudes to a sample of SN spectral

1031: templates of different types, ages, redshifts, and extinctions and

1032: returns the probability of a candidate being a Type Ia SN, $P({\rm

1033:   Ia})$, as opposed to being a core-collapse SN. The routine uses as a

1034: prior the host redshift probability distribution function of each

1035: candidate. Six candidates in our sample have spectroscopic host

1036: redshifts. For the rest of the candidates we retrieve photometric

1037: redshifts from the SDSS catalog. We then assume Gaussian redshift

1038: probability distribution functions using the host spectroscopic or

1039: photometric redshifts and their errors.

1040: 

1041: 

1042: 

1043: 

1044: 

1045: %The SN-ABC returns the probability of a candidate being a Type Ia SN,

1046: %$P({\rm Ia})$, as opposed to being a core-collapse SN. We consider

1047: %candidates which have $P({\rm Ia})>0.9$ as likely type Ia SNe. We

1048: %chose this $P({\rm Ia})$ threshold in order to minimize the

1049: %misclassification of Type II SNe as Type Ia (see more details below).

1050: % Out of the $29$ candidates in our

1051: %sample, $16$ are classified as Ia's, with $\chi^{2}$ values in the

1052: %range $0.3-10.6$. Table $2$ list the candidate properties, including

1053: %SN types and host redshifts.

1054: 

1055: 

1056: 

1057: 

1058: We next perform an analysis similar to that of Poznanski et al.

1059: (2007b) in order to determine the degree of classification uncertainty

1060: and bias, i.e., what is the most probable real range in the fraction

1061: of SNe Ia, given the fraction that is classified as such. Using the SN

1062: spectral templates, we create a sample of fake SNe of the four most

1063: prevalent types: Ia, IIP, IIn, and Ib/c. The distribution of redshifts

1064: is according to the redshift distribution of the galaxy sample used in

1065: the detection efficiency simulations (see $\S4$).  We note that while

1066: Type Ia's and IIn's are observable up to a redshift of $z\approx

1067: 0.35$, given the SDSS limiting magnitude, types Ib/c and IIP are

1068: observable only out to $z\approx 0.15$. We find that, in our redshift

1069: range, the value of $P({\rm Ia})$ is not indicative of the actual

1070: probability of an object being a SN Ia. This is due to the fact that

1071: while SNe Ia usually are assigned very high values ($85\%$ get $P({\rm

1072:   Ia})>0.9$) for core-collapse SNe the results are less obvious. About

1073: half the type IIn and Ib/c SNe, and a fifth of the type IIP SNe get

1074: misclassified, with values of $P({\rm Ia})>0.5$. As a consequence we use

1075: $P({\rm Ia})$ as a quality indicator, rather than a probability,

1076: cutting the sample at $P({\rm Ia})>0.9$ where the contamination by

1077: core-collapse SNe is minimal. The fractions of simulated SNe that are

1078: classified or misclassified as Type Ia by the SN-ABC routine using the

1079: above threshold are $0.85,~0.04,~0.28$, and $0.13$ for types Ia, IIP,

1080: IIn, and Ib/c, respectively. Applying SN-ABC together with a $P({\rm

1081:   Ia})>0.9$ threshold on the $29$ candidates in our sample results in

1082: $16$ candidates which are classified as Ia's, with $\chi^{2}$ values

1083: in the range $0.3-10.6$. Table $2$ lists the candidate properties,

1084: including SN types and host redshifts.

1085: 

1086: 

1087: 

1088: Since we do not know the real fraction of Type Ia SNe at low redshift,

1089: we create a sample of possible fractions of the different types, e.g.,

1090: $30\%~{\rm Ia}, ~50\%~{\rm IIP},~ 10\%~{\rm IIn},~ 10\%~{\rm Ib/c}$,

1091: using steps of $10\%$. For each set of fractions, we calculate the

1092: binomial probability of finding the observed $16$ SNe Ia out of $29$,

1093: using the fractions of SNe classified as Type Ia that we calculated

1094: for the four SN types. Each of the combinations is given a weight

1095: according to the number of permutations with the same fraction of SNe

1096: Ia. The outcome of this calculation is a probability distribution for

1097: the real number of SNe in our sample, given that $16$ SNe were

1098: classified as Type Ia.  From this probability function we derive that

1099: the most probable ``true'' number of SNe Ia in our sample is $N_{{\rm

1100:     Ia}}=17^{+3.8}_{-3.1}\pm 3.8$, where the first error is the $68\%$

1101: range Poisson error, and the second one is a systematic $68\%$ range

1102: error due to the uncertainty in classification.

1103: \begin{table*}

1104: \scriptsize

1105: %\begin{minipage}{\textwidth}

1106: %\begin{landscape}

1107: \begin{center}

1108: \caption{SN candidate sample}

1109: \smallskip

1110: %\tiny

1111: \begin{tabular}{ccccccccccccc}

1112: \hline

1113: \noalign{\smallskip}

1114: Id & RA & Dec & $\beta ({\rm deg})$ & $u$ & $g$ & $r$ & $i$ & $z$ &

1115: $E(B-V)$ & Redshift & $P(\rm{Ia})$ & Sample\\

1116: %\hline

1117: \hline

1118: %\noalign{\smallskip}

1119: \noalign{\smallskip}

1120: SISN01 & 15:44:10.57 & +51:46:30.57 & 67.7 & 22.43 & 20.72 & 20.91 & 21.04 &

1121: ---- & 0.014 & ---- & 0.98 & hostless \\

1122: SISN02 & 15:45:04.35 & +40:11:38.27 & 57.7 & 20.94 & 20.95 & 21.54 & 21.80 &

1123: 21.65 & 0.016 & 0.13 & 0.02 & main \\

1124: \bf SISN03 & 14:56:43.00 & +42:29:17.73 & 55.5 & 22.40 & 21.47 & 21.49 & 21.22 &

1125: 21.59 & 0.019 & 0.20 & 0.98 & main\\

1126: SISN04 & 14:57:06.28 & +46:37:25.51 & 59.0 & 21.61 & 21.31 & 21.64 & 21.07 &

1127: 22.38 & 0.015 & 0.61 & 0.42 & high-z\\

1128: \bf SISN05 & 14:56:48.96 & +10:57:47.53 & 26.5 & 22.84 & 21.18 & 21.24 & 21.18 &

1129: 21.41 & 0.034 & 0.16 & 0.99 & main\\

1130: \bf SISN06 & 14:56:49.05 & +36:36:56.41 & 50.3 & 22.05 & 20.36 & 20.56 & 20.74 &

1131: 21.65 & 0.015 & 0.20 & 1.00 & main\\

1132: \bf SISN07 & 14:42:59.88 & +62:42:48.38 & 69.1 & 22.44 & 20.59 & 20.39 & 20.69 &

1133: 20.65 & 0.015 & 0.16 & 0.98 & main\\

1134: SISN08 & 14:44:51.65 & +06:55:48.55 & 21.8 & 20.36 & 20.39 & 20.50 & 20.71 &

1135: 21.25 & 0.033 & 0.76 & 0.01 & high-z\\

1136: SISN09 & 14:46:19.33 & +53:47:14.39 & 63.7 & 22.66 & 21.04 & 20.87 & 21.60 &

1137: 21.22 & 0.010 & 0.65 & 0.93 & high-z\\

1138: SISN10 & 14:49:18.82 & +52:33:07.73 & 63.0 & 23.66 & 21.54 & 21.35 & 21.66 &

1139: 21.23 & 0.015 & {\bf 0.07} & 0.02 & main\\

1140: \bf SISN11 & 15:00:35.10 & +00:48:53.50 & 17.2 & 23.63 & 20.91 & 20.97 & 20.74 &

1141: 21.69 & 0.049 & 0.20 & 0.99 & main\\

1142: SISN12 & 15:00:58.34 & +29:04:11.96 & 43.8 & 22.57 & 20.98 & 20.50 & 20.89 &

1143: 21.25 & 0.018 & 0.04 & 0.19 & main\\

1144: \bf SISN13 & 15:02:35.07 & +28:56:59.42 & 43.9 & 21.79 & 19.72 & 19.30 & 19.79 &

1145: 20.07 & 0.023 & 0.14 & 1.00 & main\\

1146: SISN14 & 15:03:47.16 & +04:56:04.76 & 21.3 & ---- & 21.96 & 20.97 & 21.19 &

1147: 20.82 & 0.040 & 0.44 & 0.39 & main\\

1148: SISN15 & 15:07:33.29 & +45:33:49.61 & 59.1 & 22.21 & 21.75 & 21.31 & 21.29 &

1149: 21.41 & 0.021 & ---- & 0.23 & hostless\\

1150: SISN16 & 15:08:45.09 & -00:38:53.27 & 16.3 & 23.41 & 21.07 & 20.51 & 20.33 &

1151: 21.06 & 0.065 & ---- & 0.74 & hostless\\

1152: \bf SISN17 & 15:08:37.26 & +45:29:16.72 & 59.2 & 22.98 & 22.20 & 21.31 & 21.06 &

1153: 21.31 & 0.025 & 0.09 & 0.93 & main\\

1154: \bf SISN18 & 15:12:23.54 & +02:40:28.03 & 19.8 & 22.44 & 21.18 & 21.15 & 21.01 &

1155: ---- & 0.041 & 0.23 & 1.00 & main\\

1156: SISN19 & 15:12:44.37 & +06:20:37.23 & 23.3 & 21.95 & 21.52 & 20.97 & 20.65 &

1157: 20.17 & 0.034 & ---- & 0.57 & hostless\\

1158: SISN20 & 15:14:43.66 & +04:39:55.11 & 21.9 & ---- & 21.31 & 21.12 & 20.91 &

1159: 20.81 & 0.051 & 0.10 & 0.19 & main\\

1160: SISN21 & 15:16:38.95 & +05:57:26.38 & 23.2 & 20.82 & 20.13 & 19.90 & 19.78 &

1161: 19.87 & 0.042 & ---- & 0.35 & hostless\\

1162: SISN22 & 15:16:56.21 & +47:10:04.82 & 61.4 & 21.10 & 21.47 & 20.86 & 21.22 &

1163: 22.26 & 0.030 & 0.94 & 0.00 & high-z\\

1164: SISN23 & 15:17:32.35 & +04:30:02.51 & 21.9 & 21.22 & 19.43 & 19.31 & 19.55 &

1165: 18.72 & 0.047 & 0.00 & 0.08 & main\\

1166: SISN24 & 15:17:33.01 & +39:19:52.54 & 54.6 & 22.24 & 20.81 & 20.71 & 21.09 &

1167: 21.36 & 0.017 & ---- & 0.95 & hostless\\

1168: \bf SISN25 & 15:22:29.67 & +38:46:35.4 & 54.5 & 22.67 & 21.68 & 21.56 & 21.36 &

1169: 21.72 & 0.017 & 0.25 & 0.99 & main\\

1170: \bf SISN26 & 15:33:57.56 & -00:48:42.1 & 17.9 & 21.50 & 20.27 & 20.25 & 20.48 &

1171: 21.16 & 0.116 & {\bf 0.12} & 1.00 & main\\

1172: \bf SISN27 & 15:33:07.47 & +29:54:54.7 & 47.2 & 23.18 & 22.15 & 21.03 & 21.22 &

1173: 21.19 & 0.03 & 0.08 & 0.95 & main\\

1174: SISN28 & 15:33:40.44 & +09:16:37.1 & 27.6 & 20.62 & 20.57 & 20.72 & 20.59 &

1175: 20.96 & 0.041 & 0.05 & 0.02 & main\\

1176: \bf SISN29 & 15:37:37.48 & -00:38:37.3 & 18.2 & 22.25 & 22.32 & 21.02 & 20.90 &

1177: 20.97 & 0.098 & {\bf 0.16} & 1.00 & main\\

1178: SISN30 & 15:43:11.48 & -00:23:55.6 & 18.8 & 21.70 & 20.70 & 20.38 & 20.14 &

1179: 19.96 & 0.096 & 0.04 & 0.04 & main\\

1180: SISN31 & 15:45:27.58 & +26:27:58.8 & 44.9 & 19.88 & 18.61 & 18.07 & 18.16 &

1181: 17.82 & 0.049 & {\bf 0.03} & 0.01 & main\\

1182: \bf SISN32 & 15:45:46.02 & +35:37:06.7 & 53.4 & 21.73 & 20.02 & 19.55 & 19.98 &

1183: 20.59 & 0.029 & 0.06 & 0.97 & main\\

1184: \bf SISN33 & 15:46:03.12 & +22:58:53.0 & 41.6 & 23.01 & 22.47 & 21.17 & 21.21 &

1185: 20.94 & 0.055 & {\bf 0.12} & 1.00 & main\\

1186: SISN34 & 15:46:15.29 & -00:37:05.8 & 18.8 & 21.76 & 19.83 & 19.25 & 18.91 &

1187: 18.94 & 0.103 & ---- & 0.06 & hostless\\

1188: SISN35 & 15:46:48.36 & +03:26:30.7 & 22.8 & 20.68 & 20.30 & 20.83 & 20.76 &

1189: 20.89 & 0.094 & 0.10 & 0.83 & main\\

1190: SISN36 & 15:49:35.40 & +39:59:10.9 & 57.8 & 22.28 & 21.17 & 20.49 & 20.15 &

1191: 20.55 & 0.012 & 0.56 & 0.86 & main\\

1192: SISN37 & 15:51:22.86 & +04:19:46.6 & 23.9 & 21.33 & 20.71 & 20.65 & 20.89 &

1193: 22.35 & 0.079 & ---- & 0.96 & hostless\\

1194: SISN38 & 15:52:32.30 & +25:38:47.8 & 44.6 & 22.58 & 20.58 & 20.02 & 20.35 &

1195: 21.72 & 0.060 & ---- & 0.99 & hostless\\

1196: SISN39 & 15:52:55.62 & +03:40:15.3 & 23.3 & 23.30 & 20.55 & 20.27 & 20.62 &

1197: ---- & 0.153 & 1.18 & 0.97 & high-z\\

1198: SISN40 & 15:54:12.31 & +24:15:30.8 & 43.3 & 23.52 & 21.85 & 21.11 & 21.24 &

1199: 22.42 & 0.048 & 0.08 & 0.59 & main\\

1200: SISN41 & 15:54:52.01 & +21:07:10.8 & 40.4 & 24.77 & 19.03 & 17.90 & 17.75 &

1201: 17.48 & 0.055 & 0.04 & 0.00 & main\\

1202: SISN42 & 15:54:26.70 & +03:41:09.8 & 23.4 & 21.85 & 20.00 & 19.65 & 19.57 &

1203: 19.69 & 0.154 & ---- & 0.64 & hostless\\

1204: SISN43 & 15:55:53.55 & +31:23:24.4 & 50.3 & 21.44 & 21.08 & 21.21 & 21.16 &

1205: 21.10 & 0.025 & ---- & 0.34 & hostless\\

1206: \bf SISN44 & 15:58:38.83 & +05:15:48.3 & 25.2 & 22.12 & 19.95 & 19.22 & 19.84 &

1207: 20.05 & 0.056 & {\bf 0.07} & 1.00 & main\\

1208: \bf SISN45 & 15:59:11.36 & +46:17:49.3 & 64.3 & 21.66 & 19.75 & 19.39 & 19.97 &

1209: 20.94 & 0.016 & 0.15 & 1.00 & main\\

1210: SISN46 & 15:43:30.40 & -01:11:51.3 & 18.1 & 21.64 & 21.00 & 21.52 & 21.85 &

1211: 21.05 & 0.117 & 0.56 & 0.79 & high-z\\

1212: SISN47 & 15:39:04.47 & +03:48:51.4 & 22.7 & 21.73 & 21.22 & 20.98 & 21.04 &

1213: 20.20 & 0.061 & 0.51 & 0.04 & high-z\\

1214: \noalign{\smallskip}

1215: \hline

1216: \smallskip

1217: \end{tabular}

1218: \end{center}

1219: %\end{landscape}

1220: %\end{minipage}

1221: Notes: Candidates with names in boldface are classified as

1222: Type Ia in the final sample.

1223: Magnitudes are before correction for Galactic

1224: extinction, derived from the reddening listed in the $E(B-V)$

1225: column. 

1226: Photometric errors are

1227: according to Table 1. Redshifts in boldface are

1228: spectroscopic. Candidates with no redshift are apparently

1229: hostless. The $P(Ia)$ values of the candidates in the hostless and high-$z$

1230: samples were calculated assuming a uniform redshift probability

1231: distribution in the range $0<z<0.35$. $\beta$ is the ecliptic latitude.

1232: \end{table*}

1233: %end{landscape}

1234: 

1235: 

1236: 

1237: 

1238: \subsection{Supernova rate}

1239: 

1240: We now derive the luminosity-normalized SN Ia rate, $r_{{\rm Ia}}$, in SNu$_{{\rm band}}$

1241: units\footnote{1 SNu$_{band}=$ SN $(100~{\rm

1242:     yr}~10^{10}{\rm L}^{band}_{\sun})^{-1}$}. The rate

1243: is calculated using

1244: \begin{equation}

1245: r_{{\rm Ia}}=\frac{N_{{\rm Ia}}}{\displaystyle \sum_{i} \int{\eta_{i}(z)j_{{\rm lim}}(z)dV}},

1246: \end{equation}

1247: where 

1248: $N_{{\rm Ia}}$ is the number of SNe Ia, 

1249: $dV$ is a comoving volume element, $j_{{\rm lim}}$ is the

1250: luminosity density originating from galaxies which are above the SDSS

1251: limiting magnitude, and $\eta_{i}(z)$ is the effective visibility time

1252: (or ``control time'') of the

1253: $i$-th image set, i.e., the time during which the SN is detectable.

1254: The integration is over the cosmological volume in each set, and the

1255: summation is over image sets. Given $\epsilon (m)$, the detection

1256: efficiency function as a function of magnitude $m$,

1257: \begin{equation}

1258: \eta_{i} (z)= \int{\epsilon[m_{{\rm eff}}(t)]\frac{dm_{{\rm eff}}}{dt}dt},

1259: \end{equation}

1260: where, $m_{{\rm eff}}(t)$ is the effective SN light curve determined by the

1261: time difference, $\Delta t_{i}$, between the reference and registered images in each

1262: set $i$,

1263: \begin{equation}

1264: m_{{\rm eff}}(t)=-2.5{\rm log}\left(10^{-0.4m(t)}-10^{-0.4m(t+\Delta

1265: t_{i})}\right).

1266: \end{equation}

1267: 

1268: 

1269: 

1270: The mean redshift which we probe in this work, which depends on the

1271: visibility time and thus on the efficiency function, is given by

1272: \begin{equation}

1273: \left<z\right>=\frac{\int{\eta(z)z~\frac{dV}{dz}dz}}{\int{\eta(z)\frac{dV}{dz}dz}}.

1274: \end{equation}

1275: 

1276: 

1277: In order to calculate the luminosity density, $j(z)$, we again use the

1278: Blanton et al. (2003) galaxy luminosity function. We convert their

1279: luminosity function, which is given for the SDSS bands shifted to

1280: $z=0.1$, back to the rest-frame SDSS bands. We also account for

1281: luminosity evolution using their luminosity evolution parameter $Q$,

1282: thus obtaining the luminosity density as a function of redshift.

1283: Integrating over the luminosity function up to the limiting magnitude

1284: at each redshift provides $j_{{\rm lim}}(z)$.

1285: 

1286: 

1287: 

1288: The SN Ia rate must also be corrected for host extinction. Riello \&

1289: Patat (2005) performed Monte Carlo simulations in which they modeled

1290: the dust distribution in host galaxies and accounted for various

1291: bulge-to-disc ratios and total optical depths. They found that the

1292: factor, $f$, by which SN Ia rates need to be corrected, is

1293: $1.27<f<1.91$, for Milky-Way-like dust. A similar, though lower,

1294: correction factor was derived by Neill et al. (2006), who derived the

1295: type Ia SN rate at $z\approx0.5$. Neill et al. (2006) considered both

1296: Gaussian and exponential host extinction distributions in their

1297: detection efficiency simulations. They found a correction factor of

1298: $1.10<f<1.37$. Based on these studies, we adopt an intermediate

1299: correction factor of $f=1.25$ to our SN rates.

1300: 

1301: 

1302: 

1303: 

1304: The derived SN Ia rates in the

1305: $g$ and $r$ bands, using Eq. $2$ and also correcting for host

1306: extinction, are

1307: \begin{eqnarray}

1308: r^{r}_{{\rm Ia}}&=&\left(11.5^{+2.5+1.1}_{-2.5-0.9}\pm 2.5\right)\times

1309: 10^{-2}~h_{70}^{2}~{\rm SNu}_{r}, \nonumber \\

1310: r^{g}_{{\rm Ia}}&=&\left(14.0^{+2.5+1.4}_{-2.5-1.1}\pm 2.5\right)\times

1311: 10^{-2}~h_{70}^{2}~{\rm SNu}_{g} \nonumber

1312: \end{eqnarray}

1313: at a mean redshift of $\left<z\right>=0.20$. The first error is due to

1314: the Poisson fluctuations in the SN number. The second is a systematic

1315: error due to the uncertainty in the detection

1316: efficiency function, calculated by

1317: using the efficiency function upper and lower $(1\sigma)$ limits. The

1318: third error is the systematic classification error.

1319: 

1320: 

1321: 

1322: For comparison with previously published rates, we convert our

1323: luminosity-normalized rates also to a volumetric rate. We do so by

1324: replacing Eq. 2 with

1325: \begin{equation}

1326: R_{{\rm Ia}}=\frac{N_{{\rm Ia}}}{\displaystyle \sum_{i} \int{\eta_{i}(z)\frac{j_{{\rm

1327:           lim}}(z)}{j_{{\rm total}}(z)}dV}},

1328: \end{equation}

1329: where $j_{{\rm total}}$ is the total luminosity density. The resultant

1330: volumetric rates are then

1331: \begin{eqnarray}

1332: R^{r}_{Ia}(0.2)&=&\left(1.75^{+0.40+0.17}_{-0.32-0.14}\pm 0.40\right)\times

1333: 10^{-5}~{\rm yr}^{-1}~h_{70}^{3}~{\rm Mpc}^{-3}, \nonumber \\

1334: R^{g}_{Ia}(0.2)&=&\left(1.89^{+0.42+0.18}_{-0.34-0.15}\pm

1335:   0.42\right)\times 10^{-5}~{\rm yr}^{-1}~h_{70}^{3}~{\rm Mpc}^{-3}. \nonumber

1336: \end{eqnarray}

1337: The difference between these two volumetric rates is due to the

1338: different evolution of the luminosity density in each band. This

1339: difference is an inherent weakness of deriving volumetric rates from

1340: luminosity-normalized rates, but in our case the difference is smaller

1341: than any of the other sources of uncertainty.

1342: 

1343: 

1344: 

1345: 

1346: 

1347: 

1348: \section{Comparison with previous measurements}

1349: 

1350: In this section, we compare our rate measurements to previously

1351: reported low-redshift SN rates. Most of these measurements (e.g.,

1352: Cappellaro et al. 1999; Hardin et al. 2000; Blanc et al. 2004) were

1353: given in $B$-band SNu units. These rates were then

1354: converted to volumetric rates using the luminosity density at the

1355: relevant redshift. However, various luminosity functions were used to

1356: convert to volumetric rates. For example, Blanc et al. converted their

1357: rate and the rates of Cappellaro et al.  (1999), Hardin et al.

1358: (2000), and Madgwick et al. (2003), using the 2dF redshift survey

1359: luminosity density (Cross et al. 2001). In contrast, Botticella et al.

1360: (2008) fit a set of luminosity density measurements (Norberg et al.

1361: 2002; Bell et al.  2004; Blanton et al.  2003; Faber et al. 2007;

1362: Tresse et al.  2007), with a smooth function of redshift, and used it

1363: to perform the conversion to volumetric rates. We will repeat the Botticella et al. (2008) conversion of volumetric rates of previously published

1364: luminosity-normalized rates (see Table 3), using their redshift-dependent luminosity

1365: density,

1366: \begin{equation}

1367: j_{B}(z)=(1.03+1.76\times z)\times 10^{8}~L_{\sun}^{B}~{\rm Mpc}^{-3}.

1368: \end{equation}

1369: 

1370: 

1371: 

1372: Dilday et al. (2008) have recently reported a Type Ia SN rate

1373: from the SDSS-II Supernova Survey (Frieman et al. 2008). The majority

1374: of SNe found in this survey, in contrast to our survey, have been

1375: confirmed spectroscopically (resulting in a lower redshift range being

1376: probed). The SN Ia rate measured by Dilday et al.,

1377: based on $17$ SNe at $z\sim 0.09$ is higher than our rate measurement

1378: by a factor of $\sim 1.5$ but consistent within the errors. Table 3

1379: shows these various low-redshift rate measurements and Fig. $7$ shows a compilation of rate measurements to $z<0.5$.

1380: \begin{table*}

1381: 

1382: %\smallskip

1383: %\centering

1384: \begin{minipage}{\textwidth}

1385: \begin{center}

1386: \caption{Comparison of low-redshift Type Ia SN rate measurements.}

1387: \begin{tabular}{@{}ccllll@{}}

1388: \hline

1389: \noalign{\smallskip}

1390: $<z>$ & $N_{{\rm Ia}}$ & \multicolumn{3}{c}{$R_{Ia}$} & Author \\

1391: \cline{3-5}

1392: \noalign{\smallskip}

1393:      & & $h_{70}^{2}~{\rm SNu}_{B}$ & $h_{70}^{2}~{\rm SNu}_{g}$ & $10^{-5}~{\rm yr}^{-1}~h_{70}^{3}~{\rm

1394:        Mpc}^{-3}$ &  \\

1395: \hline

1396:  $\sim0$ & $70$ & $0.18\pm 0.05$ & & $1.85\pm 0.5$ & Cappellaro et

1397:  al. (1999)$^b$ \\

1398:  \noalign{\smallskip}

1399:  $0.09$ & $17$ & & $0.235^{+0.07}_{-0.06}$ & $2.9^{+0.9}_{-0.7}$ & Dilday et al. (2008)$^a$ \\

1400:  \noalign{\smallskip}

1401:  $0.098$ & $19$ & $0.196\pm 0.098$ & & $2.4\pm 1.2$ & Madgwick et al. (2003)$^b$ \\

1402:  \noalign{\smallskip}

1403:  $0.13$ & $14$ & $0.125^{+0.044+0.028}_{-0.034-0.028}$ & &

1404:  $1.58^{+0.56+0.35}_{-0.43-0.35}$ & Blanc et al. (2004)$^b$ \\

1405:  \noalign{\smallskip}

1406:  $0.14$ & $4$ & $0.22^{+0.17+0.06}_{-0.10-0.03}$ & &

1407:  $2.8^{+2.2+0.7}_{-1.3-0.4}$ & Hardin et al. (2000)$^b$ \\

1408:  \noalign{\smallskip}

1409:  ${\bf 0.2}$ & ${\bf 17}$ & ${\bf 0.14^{+0.03+0.01}_{-0.03-0.01}\pm 0.03}$ & ${\bf 0.14^{+0.03+0.01}_{-0.03-0.01}\pm 0.03}$ &

1410:  ${\bf 1.89^{+0.42+0.18}_{-0.34-0.15}\pm 0.42}$ & {\bf This work}  \\

1411:  \noalign{\smallskip}

1412:  $0.25$ & $1$ & & & $1.7\pm 1.7$ & Barris \& Tonry (2006) \\

1413: 

1414: \noalign{\smallskip}

1415: \hline

1416: \end{tabular}

1417: \end{center}

1418: \end{minipage}

1419: 

1420:   Notes: $^a$ Luminosity-normalized rate derived from a

1421:   volumetric rate. $^b$ These rates have been converted to volumetric

1422:   rates using the redshift-dependent luminosity density function from Botticella et al. (2008).

1423: 

1424: \end{table*}

1425: 

1426: 

1427: \setcounter{figure}{6}

1428: \begin{figure}

1429: \includegraphics[width=0.35\textwidth, angle=-90]{f7.eps}

1430: \caption{Type Ia SN rates from different authors

1431:   (see legend). Error bars are from quadrature additions of the various

1432: errors listed in Table~3.}

1433: \end{figure}

1434: 

1435: 

1436: Our rate measurement is consistent with other low-redshift rate

1437: measurements, and ranks with the most accurate among them. 

1438: However, as mentioned

1439: above, some of the previously published volumetric rates, which appear

1440: in Table 3, would change, depending on the luminosity density that is

1441: used to derive them. 

1442: A remaining uncertainty in such comparison arises from the fact that

1443: the rates in SNu of Cappellaro et al. (1999),

1444: Madgwick et al. (2003), Blanc et al. (2003), and Hardin et al. (2000) are multiplied by the luminosity density at the {\it

1445:   mean} redshift of each survey, thus not taking into account that the

1446: effective volume of the survey is a function of redshift.

1447: 

1448: 

1449: In order to compare our luminosity-normalized rate in SNu$_{g}$ units

1450: also to rates given in SNu$_{B}$ units, we adopt the Lupton et al. (2005)

1451: conversion between $B$ and $g$ band magnitudes,

1452: \begin{equation}

1453: B=g+0.2271+0.313(g-r).

1454: \end{equation}

1455: Using the mean luminosity-weighted color, $g-r=0.53$, of the SDSS

1456: galaxy subset sample described in \S4, together with Eq. 8, implies a factor

1457: of $1.03$ increase going from ${\rm SNu}_{g}$ units to ${\rm SNu}_{B}$

1458: units. Using this factor, our rate measurement in ${\rm SNu}_{B}$

1459: units is $r_{{\rm Ia}}=\left(14.4^{+2.6+1.4}_{-2.6-1.1}\pm 2.6\right)

1460: \times 10^{-2}~h_{70}^{2}~{\rm SNu}_{B}$. As seen in Table~3, this

1461: again agrees with previous measurements.

1462: 

1463: 

1464: 

1465: 

1466: 

1467: \section{Summary}

1468: 

1469: 

1470: We have conducted a low-redshift photometric SN survey using archival

1471: data from SDSS-I overlapping fields.  Based on the number of Type Ia

1472: SNe that we find, $N_{{\rm Ia}}=17^{+3.8}_{-3.1}\pm 3.8$, and keeping

1473: track of the various sources of error and bias, we have derived a SN

1474: Ia rate of $r^{g}_{{\rm Ia}} = \left(14.0^{+2.5+1.4}_{-2.5-1.1}\pm

1475:   2.5\right)\times 10^{-2}~h_{70}^{2}~{\rm SNu}_{g}$, or a volumetric

1476: rate of $R^{g}_{{\rm Ia}} = \left(1.89^{+0.42+0.18}_{-0.34-0.15}\pm

1477:   0.42\right)\times 10^{-5}~{\rm yr}^{-1}~h_{70}^{3}~{\rm Mpc}^{-3}$.

1478: Our derived rates are consistent with previous measurements, but rank

1479: with the most accurate ones. However, SN Ia rates at low-redshift,

1480: including ours, still suffer from several sources of uncertainty. In

1481: our case, the uncertainty is mostly due to small numbers. The

1482: derivation of volumetric rates using different luminosity functions

1483: and different extinction corrections is another source of ambiguity

1484: when comparing different measurements at similar redshifts, and when

1485: comparing observations with model predictions.  Nevertheless, we have

1486: shown that there is a vast amount of archival SDSS data that can be

1487: used for studying SNe at a low cost.  The full SDSS-I SN sample, once

1488: mined, would include several hundreds of SNe, comparable to the $\sim

1489: 500$ expected from SDSS-II, (a survey designed specifically for

1490: finding SNe). Assuming a similar fraction of Type Ia SNe as we found,

1491: both the Poisson and binomial classification uncertainties for such a

1492: large sample would be reduced to the $\sim 5\%$ level, while the

1493: uncertainty due to the detection efficiency function will remain the

1494: same.  Although SNe found by archival search methods, such as ours,

1495: can not be studied spectroscopically, a full SDSS-I sample could be

1496: useful for improved investigations of SN rates as a function of galaxy

1497: type and environment.

1498: 

1499: 

1500: 

1501: 

1502: 

1503: 

1504: 

1505: \section*{Acknowledgments}

1506: Funding for the SDSS and SDSS-II has been provided by the Alfred P.

1507: Sloan Foundation, the Participating Institutions, the National Science

1508: Foundation, the U.S. Department of Energy, the National Aeronautics

1509: and Space Administration, the Japanese Monbukagakusho, the Max Planck

1510: Society, and the Higher Education Funding Council for England. The

1511: SDSS Web Site is http://www.sdss.org/. We thank the anonymous referee

1512: for comments that improved the presentation.

1513: 

1514: 

1515: 

1516: \begin{thebibliography}{99}

1517: 

1518: \bibitem[\protect\citeauthoryear{Adelman-McCarthy et 

1519: al.}{2008}]{2008ApJS..175..297A} Adelman-McCarthy J.~K., et al., 2008, 

1520: ApJS, 175, 297 

1521: 

1522: 

1523: \bibitem[Alard

1524: \& Lupton(1998)]{1998ApJ...503..325A} Alard, C., \& Lupton, R.~H.\ 1998, \apj, 503, 325

1525: 

1526: \bibitem[Alard(2000)]{2000A&AS..144..363A} Alard, C.\ 2000, \aaps,

1527:   144, 363

1528: 

1529: 

1530: \bibitem[Barris

1531: \& Tonry(2006)]{2006ApJ...637..427B} Barris, B.~J., \& Tonry, J.~L.\ 2006, \apj, 637, 427

1532: 

1533: 

1534: \bibitem[\protect\citeauthoryear{Bell et al.}{2004}]{2004ApJ...600L..11B} 

1535: Bell E.~F., et al., 2004, ApJ, 600, L11 

1536: 

1537: 

1538: \bibitem[Bernstein et al.(2004)]{2004AJ....128.1364B} Bernstein, G.~M.,

1539: Trilling, D.~E., Allen, R.~L., Brown, M.~E., Holman, M.,

1540: \& Malhotra, R.\ 2004, \aj, 128, 1364

1541: 

1542: 

1543: \bibitem[Bertin

1544: \& Arnouts(1996)]{1996A&AS..117..393B} Bertin, E., \& Arnouts, S.\ 1996, \aaps, 117, 393

1545: 

1546: 

1547: 

1548: \bibitem[\protect\citeauthoryear{Blanc et 

1549: al.}{2004}]{2004A&A...423..881B} Blanc G., et al., 2004, A\&A, 423, 881 

1550: 

1551: %\bibitem[\protect\citeauthoryear{Blanc 

1552: %\& Greggio}{2008}]{2008NewA...13..606B} Blanc G., Greggio L., 2008, NewA, 13, 606 

1553: 

1554: 

1555: 

1556: \bibitem[\protect\citeauthoryear{Blanton et 

1557: al.}{2003}]{2003ApJ...592..819B} Blanton M.~R., et al., 2003, ApJ, 592, 819 

1558: 

1559: 

1560: \bibitem[\protect\citeauthoryear{Botticella et 

1561: al.}{2008}]{2008A&A...479...49B} Botticella M.~T., et al., 2008, A\&A, 479, 49 

1562: 

1563: 

1564: \bibitem[Cappellaro et

1565: al.(1999)]{1999A&A...351..459C} Cappellaro, E., Evans, R., \& Turatto, M.\ 1999, \aap, 351, 459

1566: 

1567: 

1568: \bibitem[\protect\citeauthoryear{Cross et al.}{2001}]{2001MNRAS.324..825C} 

1569: Cross N., et al., 2001, MNRAS, 324, 825 

1570: 

1571: 

1572: \bibitem[Csabai et al.(2003)]{2003AJ....125..580C} Csabai, I., et al.\

1573: 2003, \aj, 125, 580

1574: 

1575: \bibitem[Dahlen et al.(2004)]{2004ApJ...613..189D} Dahlen, T., et al.\

1576: 2004, \apj, 613, 189

1577: 

1578: 

1579: \bibitem[\protect\citeauthoryear{Dilday et al.}{2008}]{2008ApJ...682..262D} 

1580: Dilday B., et al., 2008, ApJ, 682, 262 

1581: 

1582: 

1583: 

1584: \bibitem[Faber et al.(2007)]{2007ApJ...665..265F} Faber, S.~M., et al.\

1585: 2007, \apj, 665, 265

1586: 

1587: 

1588: \bibitem[Frieman et al.(2008)]{2008AJ....135..338F} Frieman, J.~A., et al.\

1589: 2008, \aj, 135, 338

1590: 

1591: \bibitem[Fukugita et al.(1996)]{1996AJ....111.1748F} Fukugita, M.,

1592: Ichikawa, T., Gunn, J.~E., Doi, M., Shimasaku, K.,

1593: \& Schneider, D.~P.\ 1996, \aj, 111, 1748

1594: 

1595: \bibitem[Gal-Yam et al.(2002)]{2002MNRAS.332...37G} Gal-Yam, A., Maoz, D.,

1596: \& Sharon, K.\ 2002, \mnras, 332, 37

1597: 

1598: \bibitem[Graham et al.(2008)]{2008AJ....135.1343G} Graham, M.~L., et al.\ 

1599: 2008, \aj, 135, 1343 

1600: 

1601: \bibitem[Hardin et

1602: al.(2000)]{2000A&A...362..419H} Hardin, D., et al.\ 2000, \aap, 362,

1603: 419

1604: 

1605: \bibitem[Kaiser(2004)]{2004SPIE.5489...11K} Kaiser, N.\ 2004, \procspie,

1606: 5489, 11

1607: 

1608: 

1609: \bibitem[Madgwick et al.(2003)]{2003ApJ...599L..33M} Madgwick, D.~S.,

1610: Hewett, P.~C., Mortlock, D.~J., \& Wang, L.\ 2003, \apjl, 599, L33

1611: 

1612: 

1613: %\bibitem[Mannucci et al.(2006)]{2006MNRAS.370..773M} Mannucci, F., Della

1614: %Valle, M., \& Panagia, N.\ 2006, \mnras, 370, 773

1615: 

1616: 

1617: \bibitem[Mannucci et al.(2008)]{2008MNRAS.383.1121M} Mannucci, F., Maoz,

1618: D., Sharon, K., Botticella, M.~T., Della Valle, M., Gal-Yam, A.,

1619: \& Panagia, N.\ 2008, \mnras, 383, 1121

1620: 

1621: 

1622: \bibitem[Maoz

1623: \& Gal-Yam(2004)]{2004MNRAS.347..951M} Maoz, D., \& Gal-Yam, A.\ 2004, \mnras, 347, 951

1624: 

1625: 

1626: \bibitem[Neill et al.(2006)]{2006AJ....132.1126N} Neill, J.~D., et al.\

1627: 2006, \aj, 132, 1126

1628: 

1629: 

1630: \bibitem[Neill et al.(2007)]{2007AIPC..924..421N} Neill, J.~D., et al.\

1631: 2007, American Institute of Physics Conference Series, 924, 421

1632: 

1633: 

1634: \bibitem[Norberg et al.(2002)]{2002MNRAS.336..907N} Norberg, P., et al.\

1635: 2002, \mnras, 336, 907

1636: 

1637: 

1638: \bibitem[\protect\citeauthoryear{Nugent, Kim, 

1639: \& Perlmutter}{2002}]{2002PASP..114..803N} Nugent P., Kim A., Perlmutter S., 2002, PASP, 114, 803 

1640: 

1641: 

1642: 

1643: \bibitem[\protect\citeauthoryear{Lupton et al.}{2001}]{2001ASPC..238..269L} 

1644: Lupton R., Gunn J.~E., Ivezi{\'c} Z., Knapp G.~R., Kent S., 2001, ASPC, 

1645: 238, 269 

1646: 

1647: 

1648: \bibitem[Lupton (2005)]{} Lupton, R. H., 2005, AJ, submitted

1649: 

1650: %\bibitem[Ofek et al.(2007)]{2007ApJ...659L..13O} Ofek, E.~O., et al.\ 2007,

1651: %\apjl, 659, L13

1652: 

1653: 

1654: \bibitem[Oyaizu et al.(2008)]{2008ApJ...674..768O} Oyaizu, H., Lima, M.,

1655: Cunha, C.~E., Lin, H., Frieman, J., \& Sheldon, E.~S.\ 2008, \apj, 674, 768

1656: 

1657: 

1658: \bibitem[Pier et al.(2003)]{2003AJ....125.1559P} Pier, J.~R., Munn, J.~A.,

1659: Hindsley, R.~B., Hennessy, G.~S., Kent, S.~M., Lupton, R.~H.,

1660: \& Ivezi{\'c}, {\v Z}.\ 2003, \aj, 125, 1559

1661: 

1662: 

1663: \bibitem[\protect\citeauthoryear{Poznanski et 

1664: al.}{2002}]{2002PASP..114..833P} Poznanski D., Gal-Yam A., Maoz D., 

1665: Filippenko A.~V., Leonard D.~C., Matheson T., 2002, PASP, 114, 833 

1666: 

1667: 

1668: \bibitem[Poznanski et al.(2007)]{2007AJ....134.1285P} Poznanski, D., Maoz,

1669: D., \& Gal-Yam, A.\ 2007, \aj, 134, 1285

1670: 

1671: \bibitem[Poznanski et al.(2007)]{2007MNRAS.382.1169P} Poznanski, D., et

1672: al.\ 2007, \mnras, 382, 1169

1673: 

1674: \bibitem[Riello

1675: \& Patat(2005)]{2005MNRAS.362..671R} Riello, M., \& Patat, F.\ 2005,

1676: \mnras, 362, 671

1677: 

1678: \bibitem[Sand et al.(2008)]{2008AJ....135.1917S} Sand, D.~J., Zaritsky, D., 

1679: Herbert-Fort, S., Sivanandam, S., \& Clowe, D.\ 2008, \aj, 135, 1917 

1680: 

1681: \bibitem[Scannapieco

1682: \& Bildsten(2005)]{2005ApJ...629L..85S} Scannapieco, E., \& Bildsten, L.\ 2005, \apjl, 629, L85

1683: 

1684: 

1685: \bibitem[Schlegel et al.(1998)]{1998ApJ...500..525S} Schlegel, D.~J.,

1686: Finkbeiner, D.~P., \& Davis, M.\ 1998, \apj, 500, 525

1687: 

1688: 

1689: \bibitem[Sharon et al.(2007)]{2007ApJ...660.1165S} Sharon, K., Gal-Yam, A.,

1690: Maoz, D., Filippenko, A.~V., \& Guhathakurta, P.\ 2007, \apj, 660,

1691: 1165

1692: 

1693: %\bibitem[Smith et al.(2007)]{2007ApJ...666.1116S} Smith, N., et al.\ 2007,

1694: %\apj, 666, 1116

1695: 

1696: 

1697: \bibitem[\protect\citeauthoryear{Tody}{1986}]{1986SPIE..627..733T} Tody D., 

1698: 1986, SPIE, 627, 733 

1699: 

1700: 

1701: \bibitem[Tonry et al.(2003)]{2003ApJ...594....1T} Tonry, J.~L., et al.\

1702: 2003, \apj, 594, 1

1703: 

1704: 

1705: \bibitem[Tresse et

1706: al.(2007)]{2007A&A...472..403T} Tresse, L., et al.\ 2007, \aap, 472,

1707: 403

1708: 

1709: 

1710: \bibitem[Tucker et al.(2006)]{2006AN....327..821T} Tucker, D.~L., et al.\

1711: 2006, Astronomische Nachrichten, 327, 821

1712: 

1713: \bibitem[Tyson(2002)]{2002SPIE.4836...10T} Tyson, J.~A.\ 2002, \procspie,

1714: 4836, 10

1715: 

1716: \bibitem[York et al.(2000)]{2000AJ....120.1579Y} York, D.~G., et al.\ 2000,

1717: \aj, 120, 1579

1718: 

1719: 

1720: \end{thebibliography}

1721: 

1722: 

1723: 

1724: 

1725: \end{document}

1726: 

1727: