The SDSS-II Supernova Survey: Parameterizing the Type Ia Supernova Rate as a Function of Host Galaxy Properties
Mathew Smith, Robert C Nichol, Benjamin Dilday, John Marriner, Richard Kessler, Bruce Bassett, David Cinabro, Joshua Frieman, Peter Garnavich, Saurabh W Jha, Hubert Lampeitl, Masao Sako, Donald P Schneider, Jesper Sollerman
aa r X i v : . [ a s t r o - ph . C O ] A ug The SDSS-II Supernova Survey: Parameterizing the Type IaSupernova Rate as a Function of Host Galaxy Properties
Mathew Smith , , Robert C Nichol , Benjamin Dilday , , John Marriner ,Richard Kessler , , Bruce Bassett , , , David Cinabro , Joshua Frieman , , ,Peter Garnavich , Saurabh W Jha , Hubert Lampeitl , Masao Sako ,Donald P Schneider , Jesper Sollerman [email protected] ABSTRACT
Using data from the Sloan Digital Sky Supernova Survey-II (SDSS-II SN Survey), we measurethe rate of Type Ia Supernovae (SNe Ia) as a function of galaxy properties at intermediate redshift.A sample of 342 SNe Ia with 0 . < z < .
25 is constructed. Using broad-band photometry andredshifts we use the P´EGASE.2 spectral energy distributions (SEDs) to estimate host galaxystellar masses and recent star-formation rates. We find that the rate of SNe Ia per unit stellarmass is significantly higher (by a factor of ∼
30) in highly star-forming galaxies compared topassive galaxies. When parameterizing the SN Ia rate (SNR Ia ) based on host galaxy properties,we find that the rate of SNe Ia in passive galaxies is not linearly proportional to the stellar mass,instead a SNR Ia ∝ M . is favored. However, such a parameterization does not describe theobserved SN Ia rate in star-forming galaxies. The SN Ia rate in star-forming galaxies is well fitby SNR Ia = 1 . ± . × − M . ± . + 1 . ± . × − ˙ M . ± . (statistical errors only),where M is the host galaxy mass (in M ⊙ ) and ˙ M is the star-formation rate (in M ⊙ yr − ). Theseresults are insensitive to the selection criteria used, redshift limit considered and the inclusionof non-spectroscopically confirmed SNe Ia. We also show there is a dependence between thedistribution of the MLCS light-curve decline rate parameter, ∆, and host galaxy type. Passivegalaxies host less luminous SNe Ia than seen in moderately and highly star-forming galaxies,although a population of luminous SNe is observed in passive galaxies, contradicting previousassertions that these SNe Ia are only observed in younger stellar systems. The MLCS extinctionparameter, A V , is similar in passive and moderately star-forming galaxies, but we find indicationsthat it is smaller, on average, in highly star-forming galaxies. We confirm this result using theSALT2 light-curve fitter. Subject headings:
Cosmology:observations — distance scale — Galaxies:evolution — supernovae:general Astrophysics, Cosmology and Gravity Centre (ACGC),Department of Mathematics and Applied Mathematics,University of Cape Town, Rondebosch, 7701, SA Institute of Cosmology and Gravitation, University ofPortsmouth, Portsmouth, PO1 3FX, UK African Institute for Mathematical Sciences, 6-8 Mel-rose Road, Muizenberg 7945, SA Las Cumbres Observatory Global Telescope Network,6740 Cortona Dr., Suite 102, Goleta, CA 93117, USA Department of Physics, University of California, SantaBarbara, Broida Hall, Mail Code 9530, Santa Barbara, CA 93106-9530, USA Center for Particle Astrophysics, Fermilab, P.O. Box500, Batavia, IL 60510, USA Department of Astronomy & Astrophysics, The Uni-versity of Chicago, 5640 S. Ellis Ave, Chicago, IL 60637,USA Kavli Institute for Cosmological Physics, The Univer-sity of Chicago, 5640 S. Ellis Ave, Chicago, IL 60637, USA South African Astronomical Observatory, P.O. Box 9,Observatory 7935, South Africa Department of Mathematics and Applied Mathemat- . Introduction Type Ia supernovae (SNe Ia) have been exten-sively studied because they provide accurate rel-ative distances on cosmological scales. Measure-ments of SNe Ia have indicated that the expan-sion of the universe is currently accelerating (Riesset al. 1998; Astier et al. 2006; Wood-Vasey et al.2007; Kessler et al. 2009a; Lampeitl et al. 2010a),leading to the introduction of a “Dark Energy”component in our model of the Universe.SNe Ia are thought to arise from carbon-oxygenwhite dwarfs that accrete mass from a companionstar and approach the Chandrasekhar mass limit,resulting in a thermonuclear explosion (Hoyle &Fowler 1960; Branch et al. 1995; Yungelson & Livio1998). However, there is still significant debateon the details; e.g. the explosion mechanism, theaccretion process, and the progenitor companionstar, which may be a giant star, a main sequencestar, or a secondary white dwarf (H¨oflich et al.2003). A measurement of the delay time ( i.e. , thetime between the formation of the binary systemand its thermonuclear explosion), constrains thepossible progenitor systems (Greggio 2005). Thedelay time distribution can be determined obser-vationally by comparing the observed SNe Ia ratesin galaxies with different star-formation histories(Gal-Yam & Maoz 2004).It has been observationally determined thatSNe Ia are distinctly more common in galaxyhosts with recent star-formation activity (Oem-ler & Tinsley 1979). Recent work has determinedthat the SNe Ia rate per unit stellar mass de-pends on host galaxy morphology and (B-K) color ics, University of Cape Town, Rondebosch, 7701, SA Wayne State University, Department of Physics andAstronomy, Detroit, MI, 48201, USA Department of Physics, University of Notre Dame, 225Nieuwland Science Hall, Notre Dame, IN 46556, USA Department of Physics and Astronomy, Rutgers, theState University of New Jersey, 136 Frelinghuysen Road,Piscataway, NJ 08854, USA Physics & Astronomy, University of Pennsylvania, 209South 33rd Street, Philadelphia, PA 19104 Department of Astronomy and Astrophysics, Pennsyl-vania State University, 525 Davey Laboratory, UniversityPark, PA 16802, USA The Oskar Klein Centre, Department of Astronomy,AlbaNova, Stockholm University, SE-106 91 Stockholm,Sweden (Mannucci et al. 2005) and that the SN Ia rate inlate-type galaxies is a factor ∼
20 higher than inE/S0 galaxies. SNe Ia are seen locally to be rarerin galaxy bulges than spiral arms (Wang et al.1997) and more common in blue galaxies than red(Mannucci et al. 2005). The population associatedwith star-formation suggests that the SN Ia ratecontains a population with a cosmologically shorttime delay, while the observation of SNe Ia in veryold systems indicates the existence of a populationwith large time delay (Cappellaro et al. 1999).Scannapieco & Bildsten (2005) and later Man-nucci et al. (2006) and Sullivan et al. (2006) pro-posed a “two-component” SN Ia rate, consistingof a prompt component, dependent on recent hostgalaxy star-formation, and a delayed componentdependent on galaxy stellar mass. The overall SNIa rate is thus the sum of these two components,and can be further generalised as a function of thegalaxy star-formation rate and stellar mass. Ob-servations strongly favor a two-component modelover a single component model (Sullivan et al.2006) and since the cosmic star-formation rate in-creases with redshift, we expect that the promptcomponent will become a larger fraction of theSN Ia population with increasing redshift. Maozet al. (2010), using cluster rate measurements,suggest a universal delay time distribution, in-dependent of environment and parameterized bySNR Ia ∝ t − . ± . .Several attempts have been made to constrainthe functional form of the SNe Ia rate. Sullivanet al. (2006), using 124 SNe Ia from the SNLSsurvey, found that, for passive galaxies, the SNeIa rate is consistent with a linear relationship withhost galaxy stellar mass. Recently, Li et al. (2010),used a sample of 274 SNe Ia from the LOSS sur-vey, to consider how the size and morphology ofthe host galaxy affects the SNe Ia rate. They fa-vor a power-law relationship between galaxy stel-lar mass and the SNe Ia rate (SNuM) with expo-nent approximately one half independent of bothgalaxy morphology and color. Maoz et al. (2010)find evidence for both a “prompt” (age <
420 Myr)and “delayed” component ranging between 2 . z = 0 . z <
2, respec-tively. Dilday et al. (2008) determined the mostaccurate SNe Ia rate at intermediate redshift usingthe first year SDSS-II SN dataset to z ≤ .
12 andextended the analysis to z < . § § § § § § § § § § §
2. The SDSS-II SN Survey
In this work, we use the full sample from theSDSS-II SN Survey (Frieman et al. 2008). Thisprovides one of the largest samples of SNe Ia cur-rently available.The SDSS-II SN Survey was a three year rollingsearch that produced a sample of spectroscopicallyconfirmed SNe Ia with well-measured multi-colorlight-curves at intermediate redshift ( z < .
4) us-ing the SDSS 2.5m telescope (York et al. 2000;Strauss et al. 2002; Gunn et al. 2006) at ApachePoint Observatory with a wide field CCD camera(Gunn et al. 1998). Observations were made in theSDSS ugriz filters (Fukugita et al. 1996), alternat-ing between the northern and southern “strips”of the field designated as “Stripe 82” (Stoughtonet al. 2002), bounded by − ◦ < α (J2000) < ◦ ,and − . ◦ < δ (J2000) < . ◦ . Adverseweather and bright moonlight resulted in an aver-age observation of each strip once every four nightswith typical limiting magnitudes of g ∼ . r ∼ . i ∼ . § . < z < .
25, where theefficiency of the survey is high (Dilday et al. 2010).This homogenous sample is comprised of 197 spec-troscopically confirmed SNe Ia, with a further 87having a host galaxy spectroscopic redshift. Allobjects are selected using a well defined selectioncriteria, and have well-measured light-curves thatare consistent with a SNe Ia template, based onthe Bayesian light-curve fitting method of Sakoet al. (2008). The selection criteria used to createthis sample is discussed in §
3. Incompleteness Corrections
There are two major sources of inefficiency inthe SDSS-II SN pipeline that lead to potential bi-ases in the spectroscopically confirmed SN sample:detection efficiency and spectroscopic incomplete-ness. The detection efficiency was primarily mag-nitude limited and is amenable to calculation bysimulation. The spectroscopic selection and anal-ysis depends on many factors that are difficult toquantify. We adopt a strategy of augmenting thesample of spectroscopically confirmed SNe Ia witha sample of photometrically classified SNe Ia, iden-tified by their light-curve shape and color and cor-recting for detection efficiency.
The SDSS-II SN Survey prioritized spectro-scopic follow-up observations of SN candidates us-ing a Bayesian classification method (Sako et al.2008) . However, the final ranking and decisionson spectroscopic follow-up priorities were based onthe telescope’s capabilities, local weather condi-tions and the SN position on the sky, thus leading to a spectroscopic sample whose selection criteriaare difficult to describe quantitatively. To pro-duce a homogeneous sample of SN Ia candidates,we therefore seek a sample selection that avoidsthe uncertain and time-varying spectroscopic tar-get selection process. We also seek a sample withhigh quality light-curves and low levels of contami-nation. However, we must also ensure that the ma-jority of detected SNe Ia pass this criteria, so thatour results are not dominated by the efficiency cor-rections.We adopt a two-stage process. In the first stage,we use photometry obtained during the SN Iasearch and the Bayesian classification method, toapply very loose cuts that are intended to reducethe large number of non-SN Ia transient objectsthat are classified as candidates by the SDSS-IIsearch pipeline, while retaining any SN Ia thatcould possibly survive our subsequent quality cuts.This sample is then analyzed by the more accurateSMP photometry and fit by the MLCS2k2 light-curve fitter to obtain a sample of probable SN Ia.The criteria used for our two-stage process is de-scribed as follows.Firstly, as part of the SDSS-II SN opera-tions, every transient object with more than twoepochs was selected to be a candidate, after knownAGNs, variable stars and pipeline artifacts wereremoved. There are ∼ ,
000 such candidates.The Bayesian classification technique, used in theSDSS-II SN search operations, fits SNe Ia, Ib/cand II template light-curves to each candidate,producing a probability, p T , for a candidate tobelong to each class ( T ) of SNe. This methodassumes that each candidate is a SN of some par-ticular type, but has been shown, nevertheless,to be accurate in differentiating between differentSN types (Sako et al. 2008; Kessler et al. 2010).This Bayesian classification technique was appliedto each candidate, and the following criteria wasused to select viable SN Ia candidates: • At least 3 search discovery epochs, • p Ia > . • If the candidate has more than 5 search pho- An updated version of this method is given in Sako et al.(2011). However, as our goal is to replicate the follow-upstrategy of the SDSS-II SN Survey, it is not used in thiswork. .Additional cuts were considered, including us-ing the photometric redshift from the nearest hostgalaxy to constrain the light-curve, but were re-jected, as it is significantly harder to model theSDSS-II SN survey selection function with thosecuts. Our criteria select 1762 candidates, includ-ing 88% of the spectroscopically confirmed SNe Ia.Of the 12% of confirmed SNe Ia that fail this selec-tion criteria, 27% (17) were only observed on oneor two occasions, 70% (45) do not satisfy the p Ia criteria, and 3% (2) are best-fit by a 2005gj-liketemplate (including SN 2005gj itself).The selection criteria described above, usesphotometry obtained during the SN Ia search toproduce a sample of candidates containing the vastmajority of the spectroscopically confirmed SNeIa, whilst removing the vast majority of non-SNIa transient objects. In the second stage of ourselection criteria, this sample was then analyzedusing the more complete and more accurate SMPphotometry and fit using the MLCS2k2 light-curvefitter (Jha et al. 2007; Kessler et al. 2009b) to en-sure each candidate has a well covered light-curveand is well fit by an SN Ia event. The selectioncriteria are the same as were used by Kessler et al.(2009a) and Dilday et al. (2010), namely,1. At least 5 photometric observations (all atdifferent epochs) between −
20 and +60 daysrelative to peak light in the rest-frame of theSN,2. At least one epoch with signal-to-noise ratio > g , r , and i (not necessarily thesame epoch in each passband),3. At least one photometric observation at least2 days prior to maximum brightness in theSN rest frame,4. At least one photometric observation at least10 days past maximum brightness in the SNrest frame, SN 2005gj (Aldering et al. 2006; Prieto et al. 2007) is apeculiar SN, with a flat light-curve after maximum. In ad-dition to removing SN 2005gj-like SNe, this criterion alsoremoves AGN and other non-transient events from our sam-ple.
5. MLCS2k2 light-curve fit probability > . ,6. MLCS2k2 light-curve decline rate parameterof ∆ > − . ,7. − ◦ < α (J2000) < ◦ .Excess color in SNe Ia is interpreted byMLCS2k2 as extinction by dust in the host galaxy,parameterized using Cardelli et al. (1989), where E ( B − V ) = A V /R V . For this analysis, we adopta value of R V = 2 . A V priorin the fitting process of P ( A V ) = e − A V /τ , with τ = 0 .
33, as described in Kessler et al. (2009a).For comparison, R V = 3 . R V ∼ . M = 0 . Λ = 0 . ∼ .
01 at low redshift(0 < z < . The cuts on MLCS2k2 light-curve fit probability and ∆(5,6) have a negligible effect on the size of our sample com-pared to the sampling cuts (1,2,3,4) § . < z < .
25. This leaves 379 SNe Ia. We find217 (57%) are spectroscopically confirmed SNe Ia,while 94 (25%) are unconfirmed but have hostgalaxy spectroscopic redshifts and 68 (18%) haveno spectroscopic redshift information. The num-ber of candidates that satisfy each stage of ourselection criteria is shown in Table 1. While themajority of our sample has been spectroscopicallyconfirmed, a significant fraction of candidates areonly photometrically classified. However, Dildayet al. (2010) conservatively estimated that thereis a 3% probability for non-SNe Ia to satisfy ourselection criteria, and the total estimated contam-ination by non-SNe Ia’s is 2%.Table 2 lists the number of SNe Ia that passour selection criteria for several redshift ranges, in-cluding the proportion of each sample that is spec-troscopically confirmed as SNe Ia. As expected,the proportion of spectroscopically confirmed SNedecrease with increasing redshift, but it remainsabove 50% out to z = 0 . Having defined a homogeneous sample of SNeIa candidates with 0 . < z < .
25, we need toknow the SDSS-II SN detection efficiency, ǫ ( z ). Adetailed analysis of the efficiency was given in Dil-day et al. (2008, 2010), differing here only in theuse of MLCS fitted photometric redshifts to selectthe fake SNe that pass our redshift cut. SimulatedSNe Ia, with a range of sky positions, time of peakbrightness, redshifts, decline-rate parameters, ex-tinction and host galaxy position, and realistic er-rors were added directly to the image data andwere processed by the SDSS-II SN pipeline (Sakoet al. 2008). The proportion of SNe Ia that satisfythe criteria defined in § A V and ∆. This additional correctionproduces results that are consistent with our nom-inal result, with differences of much less than 1 σ .We thus consider the survey efficiency to be a func-tion solely of redshift, but note that other analyseshave not considered the effect of this assumptionon their results.We have now defined a uniformly selected sam-ple of 379 SNe Ia candidates with 0 . < z < . able 1Number of candidates passing each stage of § No. of Candidates Spectroscopically confirmedAll SDSS-II SN candidates 19046 513After Bayesian LC fit 1762 449Passing Sample Selection 843 3190 . < z < .
25 379 217
Table 2Number of candidate SNe Ia as a function of redshift
Redshift Limit Total Spectroscopically confirmed Host Redshift Photo-z only0 . < z < .
10 21 19 (90 . . . . < z < .
15 88 73 (83 . . . . < z < .
20 214 144 (67 . . . . < z < .
25 379 217 (57 . . . . < z < .
30 559 272 (48 . . . . < z < .
40 800 312 (39 . . . the survey’s efficiency is small (Dilday et al. 2010)compared to the statistical precision of our dataand it is not necessary for us to include the uncer-tainty in our analysis. We now turn to considerthe host galaxies of these SN events.
4. Host Galaxy Determination and De-rived Quantities
Here we describe the method used to identifythe host galaxies and determine their characteris-tics, such as stellar mass and recent star-formationrate, for the 379 SNe Ia identified in §
3. We alsooutline the comparison field sample used to de-scribe the underlying galaxy population in our red-shift range.
Repeat imaging of SDSS Stripe 82 has enabledthe coaddition of images into a deep stacked image(Abazajian et al. 2009). The stack ranges from 20to 40 individual images (depending on sky posi-tion) in all 5 SDSS filters ( ugriz ) and is roughly 2magnitudes deeper than a single epoch SDSS im-age. To determine the host galaxy for each SN inour sample, we match the SN positions with SDSSgalaxies detected in this deep stacked image withina 0 .
25 arcminute radius. We require that the hostgalaxy has an SDSS model magnitude (Stoughtonet al. 2002) in the range 15 . < r < . §
3, 342 have a valid hostgalaxy identification, of which 197 (58%) are spec-troscopically confirmed to be SNe Ia, and 87 (25%)are spectroscopically unconfirmed but have a hostgalaxy spectroscopic redshift. The remaining 58objects are classified to be SNe Ia through theirphotometry alone. Of the 37 candidates that lacka valid host galaxy, 29 (78%) have a host galaxycandidate with r > . .
25 arcminute radius.
Having identified the host galaxy position andmagnitudes for 342 SNe Ia candidates, we nowdetermine their stellar mass and recent star-formation rate.There are several methods to infer galaxy prop-erties from broad-band photometry. A simple cut7n the color of the galaxy can be used to infer itsspectral type (Strateva et al. 2001) and the UVflux can provide an estimate of the recent star-formation rate (Donas et al. 1987). These simplemethods are able to differentiate between galax-ies with markedly different levels of star-formationactivity, but struggle with galaxies with similarcolors because multi-band photometry is not used(Baldry et al. 2006). Therefore, we fit our multi-band photometry to a set of Spectral Energy Dis-tributions (SEDs) and use the best-fit templateto determine the galaxy parameters. This tech-nique is widely used for photometric redshift esti-mates (Bolzonella et al. 2000; Le Borgne & Rocca-Volmerange 2002; Oyaizu et al. 2008).
The method used here is consistent with that ofSullivan et al. (2006), who studied the SN Ia rateas a function of host galaxy properties at high red-shift, allowing our results to be compared withinthe same framework. A discussion on how the dif-ferent redshift ranges covered by this analysis andthat of Sullivan et al. (2006) may affect our hostgalaxy derived properties is given in Appendix E.We use the SEDs produced by the P´EGASE.2galaxy spectral evolution code (Fioc & Rocca-Volmerange 1997; Le Borgne et al. 2004). Thesetemplates have been used extensively in the liter-ature to constrain the evolution of galaxies, par-ticularly at high redshift (Glazebrook et al. 2004;Grazian et al. 2006). We use the set of 8 evo-lutionary tracks listed in Table 1 of Le Borgne& Rocca-Volmerange (2002) (excluding the star-burst template), and assume a Kroupa (2001) Ini-tial Mass Function (IMF). In these scenarios, star-formation rate is determined using the relation-ship SFR = ν × M gas , where ν ranges from 0.07 to3 .
33 Gyr − , and M gas is the density of gas in solarmasses. Extinction due to dust is modelled inter-nally, with a King (1980) profile used for the Ellip-tical template, and a plane-parallel slab geometryis used for the spiral and irregular templates. Eachof the 8 evolutionary scenarios is evolved over 69time steps, each one corresponding to a differentgalaxy age, making a total of 552 template SEDs.These SEDs are convolved with the SDSS filterresponses (Fukugita et al. 1996) and fitted to thegalaxy fluxes (calculated from model magnitudesafter correcting for Galactic dust absorption from Schlegel et al. (1998) and AB-system offsets) usingthe Z-PEG photometric redshift code (Le Borgne& Rocca-Volmerange 2002). We keep the redshiftof the SN host galaxies fixed to the spectroscopicredshift (from either the SN or host galaxy) orthe photometric redshift determined by MLCS2k2.Applying a redshift constraint eliminates the coloruncertainty due to the cosmological redshift. Asdust is included internally in the SEDs no dust cor-rection is applied in the fitting process. We assumea default ΛCDM cosmology (Ω M = 0 . , Ω Λ = 0 . χ minimization using all 5 SDSS filters. The to-tal stellar mass of each galaxy is determined byintegrating the star-formation history of the best-fitting SED and subtracting the mass of stars thathave died. We characterize recent star-formationwith a mean star-formation rate, since the instan-taneous star-formation rate is difficult to estimatewithout high-resolution spectroscopic data. Weuse the result of Sullivan et al. (2006), who foundthat averaging the star-formation rate over a pe-riod of 0 . χ ≤ χ + 1. We consider errors on thegalaxy fluxes from the coadded image, with a min-imum error as given in Blanton & Roweis (2007).The stellar mass and recent star-formation rate forthe 342 host galaxies used in this analysis is givenin Table 9. To determine how our SN sample relates tothe underlying galaxy population in our redshiftrange, we require a sample of galaxies that is rep-resentative of the general galaxy population. Forthis sample, we use galaxies detected in the deepstacks described in § . < r < .
0. Thus cut also removes the pos-sibility of variable limiting magnitudes across theimage.We determine the stellar masses and recent8tar-formation rates for each galaxy in this sam-ple using the same method as for the host galaxysample except that the redshift is a free param-eter to be determined by the Z-PEG fit. We re-quire that the fitted redshift must lie in the red-shift range 0 < z <
2. The additional freedomallowed in determining the redshift for the fieldgalaxies can result in large error bars on the de-rived photometric redshift, stellar mass and star-formation rate estimates. In extreme cases, therecan be two or more distinct best-fit template solu-tions, resulting in more than one photometric red-shift estimate and spectral type. In these cases,the galaxy is excluded from our analysis becausethe spectral classification and derived galaxy prop-erties are ambiguous. To match the host galaxypopulation, we consider the ∼ ,
000 galaxieswith 0 . < z < . The comparison field sample is magnitude lim-ited, and thus becomes increasingly incomplete athigher redshifts, with only the brightest galax-ies observed at higher redshifts. Galaxies witha given absolute magnitude (and spectral type)will pass the apparent magnitude selection cri-teria (15 . < r < .
0) at different redshifts,which may be less than the full survey range(0 . < z < . V max method (Schmidt 1968; Felten1976). Using the best-fitting SED for each fieldgalaxy, we calculate its absolute magnitude andk-correction, and determine the redshift limits atwhich it would satisfy 15 . < r < .
0. When-ever the redshift range is less than the total sur-vey range (0 . < z < . V survey /V max , where V max is the co-moving vol-ume for which each galaxy will remain within oursurvey’s magnitude limits, and V survey is the co-moving volume of the SDSS-II SN survey, i.e. fora redshift range, 0 . < z < .
25 and constantfor each galaxy in our sample. 83% of the fieldgalaxies in our sample have redshift limits largerthan that of the SDSS-II SN survey, and are notaffected by this correction. The remaining 17% offield galaxies are on average weighted by a value of4 .
98. Since this form of incompleteness will affectboth the comparison field sample and host galaxysample, this incompleteness correction is applied to both, although only 3 of the 342 host galaxiesin our sample are affected by this correction.
Systematic uncertainties in our derived galaxyproperties can arise from many sources includingthe wavelength coverage of the SDSS filters, ourdecision to use the P´EGASE.2 SEDs, our choice ofIMF, the accuracy of the photometric redshifts forthe comparison field sample, the accuracy of theP´EGASE.2 stellar mass estimates, and the abil-ity of P´EGASE.2 to accurately recover the stellarmasses and star-formation rates for a sample ofsimulated galaxies. All these systematic errors arediscussed further in Appendix A, B, C, D and E.In Appendix A, we show that the P´EGASE.2SEDs primarily use the color of a galaxy as a proxyto infer its spectral type. The reddest galaxies areclassified as passive galaxies, with the bluest galax-ies considered highly star-forming. Moderatelystar-forming galaxies are distributed between pas-sive and highly star-forming galaxies, spanning alarge range of color.In Appendix B, we investigate the accuracy ofthe P´EGASE.2 photometric redshift estimates forour field sample. We find a mean offset in red-shift of ∆ z = 0 .
03, with the photometric redshiftestimate being smaller than the known spectro-scopic redshift. This redshift error results in anerror in stellar mass of ∆ log M = 0 .
22 M ⊙ . InAppendix C we show the effect that applying thisoffset to our data would have on the results pre-sented in § .
12 M ⊙ yr − .9owever, Brinchmann et al. (2004) measure the“instantaneous” (present day) star-formation rateinstead of our “recent” star-formation rate, whichis averaged over the last 0.5 Gyr, so the two quan-tities are not directly comparable.In Appendix E we consider how the rest-wavelength coverage of the SDSS filter set affectsour stellar mass estimates. With increasing red-shift, the SDSS filters will sample a different restwavelength ranges. This can be particularly im-portant for systems with a variety of stellar pop-ulations, such as merging galaxies. To examinethe sensitivity to our wavelength coverage, we re-peat the determination of stellar masses and star-formation rates using only three or four of the fiveSDSS filters. We find an increased scatter in theresults, but no overall bias in the stellar mass orstar-formation rate estimates. This is particularlyencouraging, because it suggests that the compari-son of our galaxy properties with those of Sullivanet al. (2006) will not be affected by the differentcosmological redshifts of the two surveys.
5. Host Galaxy Properties In § §
4, determined a host galaxyfor each object. Having estimated their stellarmass and recent star-formation rate, we now ana-lyze these derived properties, and how they relateto the supernova rate.Figure 2 shows the distribution of our hostgalaxy sample in stellar mass and star-formationrate (SFR). Galaxies are shown in three cat-egories, highly star-forming (blue), moderatelystar-forming (green), and passive (red). Thehighly and moderately star-forming galaxies areseparated by their specific star-formation rate(sSFR): the star-formation rate per unit stellarmass (Guzman et al. 1997; Brinchmann & El-lis 2000; Brinchmann et al. 2004; Sullivan et al.2006). We followed Sullivan et al. (2006) in choos-ing log sSFR = − . − .
5. The average stellar mass of apassive galaxy is log M = 10 .
52 M ⊙ , considerablymore massive than star-forming galaxies, whichaverage log M = 9 .
91 M ⊙ , consistent with otherobservations of the local universe (Taylor et al.2009).Of the 342 galaxies in our sample, 80 (23%)are classified as passive galaxies, 139 (41%) havemoderate levels of star-formation activity and 123(36%) are highly star-forming.In Figure 2, we note a “ridge line” of galaxies,which are classified as moderately star-forming,but have the lowest possible values of sSFR al-lowed. A dashed-dotted line is shown on Fig-ure 2 to highlight this population of galaxies. 78%of these galaxies are best-fit by the lenticular S0(scenario), with the remaining 22% being best-described by the elliptical template. In compari-son 52% of the remaining moderately star-forminggalaxies are best-fit by the S0 scenario. In Ap-pendix A we show the color-magnitude diagram,and conclude that these galaxies lie at the edgeof the distribution of the moderately star-forminggalaxies but appear to be distinct from the passivegalaxies. Thus, we do not remove these galaxiesfrom our analysis. We will show later through aMonte-Carlo approach that removing these galax-ies from our sample do not affect our major con-clusions.
6. SN Ia Rate
We now turn to looking at how the supernovarate depends on the galaxy properties of total stel-lar mass and recent star-formation for passive andstar-forming galaxies.
According to the standard model of galaxy for-mation, passive galaxies are primarily comprisedof old, low mass stellar systems that evolve with-out forming new stars. It is reasonable to supposethat the SN Ia population in passive galaxies couldonly occur as a result of a process with a delay timethat is long compared to the age of the galaxy. If10ig. 2.— The distribution of stellar mass & SFR for the 342 SN host galaxies. Highly star-forming galaxies areshown as blue diamonds and passive galaxies as red circles. Moderately star-forming galaxies are plotted ingreen, with light green triangles indicating the “ridge line” of galaxies discussed in the text and Appendix A,and the remaining population plotted as dark green squares. The dashed-dotted line highlights this split.Passive galaxies have SFR = 0 but are shown here as randomly distributed in the range − < log SFR < − /V max correction, and the efficiency cor-rection applied to the host galaxy sample. Eachhost galaxy is weighted by 1 /ǫ , the survey effi-ciency at the redshift of the SNe given the yearit was observed. The efficiency correction ranges between 1.4 and 2.6, with a mean weighting of1.9 for each host galaxy. By dividing the numberof host galaxies by the corresponding number offield galaxies, and including a correction for thesurvey’s observing period, we can determine howthe rate of SNe Ia varies as a function of the stellarmass of their host galaxy. Figure 3 shows the SN Iarate for both the passive and star-forming galaxysamples. It is clear that the rate of SNe in all typesof galaxies depends on the stellar mass. We alsosee that the relationship between the SN Ia rateand stellar mass is different for passive galaxiesand star-forming galaxies in the SDSS data.The data are fit to a linear function in log-space,corresponding to a power law dependence of SNrate on stellar mass as shown on Figure 3. A lineardependence on stellar mass would result in a slopeof unity. The error bars shown and fitting errors11ig. 3.— SN Ia rate as a function of host galaxy stellar mass. The values for star-forming (passive) galaxiesare shown in blue (red). The data points, for passive galaxies, from Sullivan et al. (2006) are shown as opencircles. The best-fitting lines for passive and star-forming galaxies are shown as dashed and dotted-dashedlines, respectively. Also shown is a fit (solid line) to the passive galaxies where the line slope is assumed tobe one.include statistical errors only. The uncertainty ingalaxy stellar mass is discussed in § . ± .
15 (with χ -statistic ( χ ) = 2 .
30 for 6 degrees of freedom)compared to a value of 0 . ± .
08 ( χ = 9 . n star-forming )for star-forming galaxies. The value for passivegalaxies is incompatible (at the 2 . σ level) witha linear relationship, as favored by Sullivan et al.(2006), who found a slope of 1 . ± .
12 using theSNLS data at higher redshift.Figure 3 also shows the results for passive galax-ies from Sullivan et al. (2006) as open circles. Wesee that the SDSS galaxy sample contains fewerSNe Ia in high mass passive galaxies than SNLSand more SNe Ia in low mass passive systems.While the two analyses should be directly compa-rable, the galaxy population is expected to evolvebetween z ∼ .
75 and z ∼ .
2. However, itdoes not seem that galaxy evolution can explain these differences as more massive galaxies shouldbe found in the local universe. In addition, we notethat the SDSS analysis finds a larger slope for star-forming galaxies compared to passives, while theopposite is seen in the SNLS data, who find val-ues of n star-forming = 0 . ± .
08 and 0 . ± . . The data in Figure 3 indicate that the SNIa rate depends on galaxy stellar mass, but alsothat the rate depends on whether the galaxy isactively forming stars. A two-component modelwas considered by Mannucci et al. (2006), andScannapieco & Bildsten (2005), who modelledthe SN Ia rate of a galaxy to consist of a “de-layed” component, with a long delay time that12s driven by the stellar mass of the galaxy, anda “prompt” component, with short delay timesthat is caused by the formation of new stars. Themodel assumes that the “delayed” component isproportional to the stellar mass independent of thegalaxy age and star-formation history, and thatthe “prompt” component time scale is short com-pared to changes in the star-formation rate. Theseassumptions result in an expression whose param-eters can be determined from data as was done bySullivan et al. (2006). In detail, the SNR Ia can bewritten as;SNR Ia ( t ) = A × M ( t ) + B × ˙ M ( t ) , (1)where SNR Ia ( t ) is the explosion rate of SNe Ia attime t , M ( t ) is the stellar mass of a galaxy, ˙ M ( t )is the rate of change of stellar mass, and A and B are constants determined from the data and haveunits SNe yr − M − ⊙ and SNe yr − (M ⊙ yr − ) − ,respectively. We assume ˙ M ( t ) is equal to the star-forming rate (SFR) (averaged over the previous0.5 Gyr) as discussed in § t , our SN rate mea-surements apply only to the current era and wewill suppress the dependence on t . This modelis commonly known as the “A+B” model for thesupernova rate and assumes that the SN Ia rateis linearly dependent on both the stellar mass ofa galaxy and its star-formation rate. However, in § Ia = A × M n M + B × ˙ M n SFR , (2)where n M , n SFR , A and B are constants to bedetermined from the data. Since passive galaxieshave ˙ M = 0, we can apply the results of § M = 0 . ± .
15. The straight line fit tothe passive galaxies yields log A = − . ± .
68 or A = 1 . +4 . − . × − SNe yr − M − ⊙ . If we assumen M ≡
1, we find a value of log A = − . ± .
08 or A = 2 . +0 . − . × − , which differs at 2 . σ withthe value of A = 5 . ± . × − found using theSNLS dataset.While the above parameterization of the SNIa rate uses the stellar mass and the recent star-formation rate, other galaxy properties can be con-sidered, such as the metallicity, age and level of extinction. Gallagher et al. (2005) find qualitativeevidence suggesting that the progenitor age is apossible source of diversity in SNe Ia properties.However, there is a degeneracy between the ageof a galaxy, and its metallicity, which is extremelydifficult to break using broad-band photometry.We thus confine ourselves to considering the stellarmass and star-formation rates of our host galaxiesin this analysis, but note that with improved stel-lar population models, a larger wavelength cov-erage and galaxy spectra, it may be possible tobreak this degeneracy. Using SDSS-II SNe, Guptaet al. (2011) attempt to break this degeneracy byusing multi-wavelength photometry to better con-strain the ages of their SN Ia host galaxies whileD’Andrea et al. (2011) and Konishi et al. (2011)use spectral features to determine the metallicitiesof their host galaxies. We now consider the star-forming galaxies todetermine B and n SFR . We bin the host galaxyand comparison field sample in star-formationrate, and as in § /V max correction for boththe host galaxy and comparison field samples. TheSN Ia rate is shown (blue diamonds) as a func-tion of SFR in Figure 4. We want to determinethe excess SN Ia rate due to recent star-formationactivity assuming that the term proportional tostellar mass is the same for star-forming and pas-sive galaxies. The portion due to the stellar massterm is calculated using Equation 2, and shown onthe figures (green points) as are the SN Ia rates af-ter the stellar mass term has been subtracted (redpoints). The left panel of Figure 4 uses our best-fit line with slope n M = 0 .
67 while the right paneluses the fit where the slope is fixed at n M ≡ SFR = 0 . ± .
07 and log B = − . ± .
04 ( B = 1 . +0 . − . × − SNe yr − (M ⊙ yr − ) − )with χ = 1 .
58 for 6 degrees of freedom whenn M = 0 .
67. When n M ≡ SFR = 0 . ± .
08, and log B = − . ± . B = 1 . +0 . − . × − SNe yr − (M ⊙ yr − ) − ) with χ = 1 .
52 for 6 degrees of freedom. The lack of13ensitivity to the value of n M = 0 .
674 follows be-cause the stellar mass term is always small com-pared to the star-forming term.Our best-fit to Equation 2, is thereforeSNR Ia = 1 . +4 . − . × − M . ± . +1 . +0 . − . × − ˙ M . ± . . (3)As noted previously, the analysis of Sullivanet al. (2006) in the redshift range 0 . < z < . M ≡ SFR ≡
1, we find,SNR Ia = 2 . +0 . − . × − M +1 . +0 . − . × − ˙ M . (4)For comparison, Sullivan et al. (2006) find val-ues of A = 5 . ± . × − SNe yr − M − ⊙ and B = 3 . ± . × − SNe yr − (M ⊙ yr − ) − . Ourvalue of A is 2 . σ lower, while the values of B areinconsistent at 3 . σ , indicating that recent star-formation activity plays a more significant role indetermining the overall SNe Ia rate for our sam-ple. This result is consistent with models of howgalaxies evolve through cosmic time. Observationssuggest, that at high redshift ( z = 0 . § A, § B, § C, § E).
Thus far we have used only the passive galaxiesto determine the A term and then used the star-forming galaxies to determine the B term, whilekeeping A fixed. A more sophisticated methodis to constrain the parameters simultaneously us-ing all galaxy types, thus making optimal use ofthe data. We bin the host galaxy and comparisonfield sample in the stellar mass and star-formationplane, and correct for incompleteness. By dividingthe number of host galaxies in each bin by the cor-responding number of field galaxies, we are able todetermine the SN Ia rate in each bin of stellar mass and star-formation rate. We consider several vari-ations on Equation 2. First, we consider the casewhere B ≡ i.e. the SN Ia rate is purely depen-dent on stellar mass, and n M is a free parameter.In this case, we find A = 1 . ± . × − andn M = 0 . ± . § B = 0 but assuming n SFR ≡ χ from 347 for 42 degrees of freedomto 142 for 41 degrees of freedom, and yields val-ues of n M , A , and B consistent with those foundin § SFR to vary and findn
SFR = 1 . ± .
05 with χ = 142 for 40 degreesof freedom, a negligible improvement.We thus conclude that our data is consistentwith a linear dependence on star-formation rate.Our fiducial result using bivariate fitting isSNR Ia = 1 . ± . × − M . ± . +1 . ± . × − ˙ M . ± . . (5)This is in good agreement with the values foundin § § The results from § § § §
5, sSFR is a way ofdistinguishing between galaxy types, with galax-ies with low values of sSFR being primarily largegalaxies that are using a small fraction of theirtotal stellar mass to form new stars, while thosewith larger levels of sSFR are starburst galaxies,or galaxies that are using a significant fraction oftheir stellar mass to form new stellar systems.Figure 5 exhibits the rate of SNe Ia per unitstellar mass in star-forming galaxies as a functionof sSFR. The rate increases with sSFR, reaching14ig. 4.— SN Ia rate as a function of host galaxy star-formation rate.
Left panel:
Green points indicate theexpected rate of SNe Ia due to the stellar mass of each galaxy, using the values of n M and A as determinedin § i.e. the difference between the blue and green values). A best-fitting line (dashed), and best-fittingline with unit slope (solid) is also shown. Right panel:
Identical, except a value of n M ≡ ≃ − ∼
30 for starburst galaxiescompared to passive galaxies. The measurementsof this work are in excellent agreement with thosefound at higher redshift (Sullivan et al. 2006) andin the local universe (Mannucci et al. 2005), indi-cating that this relationship holds for all redshiftsthat have been studied. The SDSS data, how-ever, has a point that appears to disagree withthe other data and the generally linear trend ofincreasing SN Ia rate with sSFR. This point cor-responds to the galaxies highlighted in § We have studied how the SN Ia rate is relatedto the host galaxy properties for the SDSS sam-ple. However, as discussed in § A and B parameters will necessarily be smaller thanfor the full sample. The results for n M and n SFR ,however, should be comparable.From Table 3 we see that the spectroscopicallyconfirmed sample is fit by n M = 0 . ± . M = 1. This value of n M may bedue to the lower proportion of passive galaxies inthis sample and a bias against more luminous andthus massive galaxies in the spectroscopic selec-tion. This bias is caused by a targeting againstprobable SNe Ia that occur in the centres of lu-minous galaxies, making them difficult to identifyspectroscopically. As the redshift limit consideredis decreased, resulting in a more complete sample,the value of n M is stable and shows no trend to-wards one (although the errors increase rapidly asthe sample size is reduced). The value for log A when we assume n M ≡ M is a free parameter, it ishighly degenerate with log A ). Table 3 also showsthat the value of n SFR is not influenced by the in-clusion of non-spectroscopically confirmed SNe Ianor the redshift range.Table 3 also shows how our selection criteriaaffects the dependence that the SN Ia rate has onthe star-formation rate. We showed in § § SFR ∼
1. A value of log B ∼ − .
85 is valid for all redshiftranges considered.Finally, we study the results of § > − . ∼
30) than thatseen in passive galaxies.The fit parameters shown in Table 3 do notevolve across our redshift range. Since any evolu-tion would be unexpected due to the small rangein cosmic time covered by our analysis, it is reas-suring to note that our results are insensitive tothe redshift interval that is chosen. The only pa-rameter that significantly changes with redshift isthe proportion of passive galaxies found. This maysimply reflect observations that SNe Ia in passivegalaxies are fainter than their star-forming coun-terparts (see § A V when determining the sam-ple of SNe Ia, by using a flat prior and a positiveprior ( A V ≥ M and n SFR we have performed linear fits to the log-log plots,but it is also possible to fit to the power law formdirectly. These fits are consistent with our linearfits to the logarithms. We have also consideredvarious bin sizes for each stage of our analysis andfind that our results are unaffected. We also con-sidered the possibility that our results may dependon a specific, anomalous year with the SDSS-II SNsurvey or may vary as a function of position onthe sky. We split the host galaxy sample by bothyear and position but found no variation on ourfinal results. Finally, we considered the effect ofmodelling ǫ , the survey efficiency, as a function ofredshift, A V and ∆, to account for the observa-tion that passive galaxies host fainter, higher ∆,SNe than star-forming galaxies ( § able 3Effect of our selection criteria on the results described in § § § Parameter Nominal Result Confirmed g Phot-ID h z < . z < . z < . . . . . . . M . ± .
15 0 . ± .
27 0 . ± .
41 0 . ± .
20 0 . ± .
26 0 . ± . n star-forming . ± .
08 0 . ± .
11 0 . ± .
12 0 . ± .
09 0 . ± .
15 0 . ± . A a − . ± . − . ± . − . ± . − . ± . − . ± . − . ± . A b − . ± . − . ± . − . ± . − . ± . − . ± . − . ± . SFR a . ± .
07 0 . ± .
11 0 . ± .
15 1 . ± .
10 1 . ± .
12 1 . ± . SFR b . ± .
08 1 . ± .
11 0 . ± .
15 1 . ± .
12 1 . ± .
16 1 . ± . B c − . ± . − . ± . − . ± . − . ± . − . ± . − . ± . B d − . ± . − . ± . − . ± . − . ± . − . ± . − . ± . B e − . ± . − . ± . − . ± . − . ± . − . ± . − . ± .
18% Passive galaxies 23 . . . . . . f . ± . . ± . . ± . . ± . . ± . . ± . f . ± .
95 66 . ± . . ± . . ± . . ± . . ± . a n M free, in units of SNe yr − M − ⊙ b n M ≡
1, in units of SNe yr − M − ⊙ c n M and n SFR free, SNe yr − (M ⊙ yr − ) − n M free and n SFR ≡
1, SNe yr − (M ⊙ yr − ) − n M ≡ SFR ≡
1, SNe yr − (M ⊙ yr − ) − × − per unit mass per year g Considering solely spectroscopically confirmed SN Ia h Considering solely photometrically typed SN Ia tent with our fiducial result.In order to study the robustness of our results,we have considered the effect of altering our selec-tion criteria, in Table 3. We have shown that theinclusion of non-spectroscopically confirmed SNeIa in our sample and varying our redshift rangeconsidered does not significantly change the val-ues of
A, B, n M and n SFR
For each host galaxy in our sample we have de-termined a value for its stellar mass and recentstar-formation rate. Each of these measurementshas an associated error that may allow galaxies tomove between bins, and thus affect our fitted pa-rameters. This may be especially important forthe sample of galaxies with ambiguous classifica-tion, highlighted in § §
5, and investigates if theyare likely to be predominantly passive in nature.To determine how the SED uncertainties affectour overall conclusions, we determine the valueof each parameter for each realization, and fit aGaussian (which is observed to provide a good fit)to each distribution. This provides an estimatefor the central values and systematic uncertaintyin each case. Table 4 gives the values and associ-ated errors for the parameters determined in thiswork for each of these four systematic tests.Both the MC and BP analysis provide valuesfor the parameters determined that are consistentwith those found as our main result, as describedin § § § M determined bythe MC analysis is larger than our default result,it is still inconsistent with n M = 1 at the 3 . σ levelwhen the stellar mass is allowed to vary, and 2 . σ when galaxies are allowed to move from passive tostar-forming.We observe that the SN Ia rate per unit stellarmass in passive galaxies is consistent in all fourcases considered. This implies that the sample ofgalaxies with ambiguous classifications (as notedin §
5) do not affect our overall conclusions. Inthe MC where the stellar mass and star-formationrate are allowed to vary, these galaxies are ableto move from moderately star-forming to passivewhere their error bars allow. However, we notethat in this case, the SN Ia rate in passive galaxiesis in fact lower than the observed value, suggestingthat these galaxies are not passively evolving, andmay have non-zero star-formation rates.The systematic error bars determined by theMC and BP tests are sub-dominant to the statis-tical uncertainties obtained in § § § § § § M ∼ . . σ level. Theexcess rate of SNe Ia in star-forming galaxies islinearly proportional to the star-formation rate inall cases considered. We find some evidence for alower SNe Ia rate per unit stellar mass in passivegalaxies, than determined in § . σ . Throughout this analysis we have compared ourresults to that of Sullivan et al. (2006). We finda different dependence on stellar mass for the SNIa rate, but agree that there is a strong depen-dence on the recent star-formation rate. In agree-ment with the results of Sullivan et al. (2006), wefind that the SN Ia rate per unit stellar mass isgreater in highly star-forming galaxies comparedto passive galaxies, with an approximately lineardependence on sSFR except for one potentiallyanomalous point in the SDSS data. Different as-sumptions about the dependence of the SN Ia rateon stellar mass do not significantly alter our con-clusions about its dependence on the recent star-formation rate.A summary of how our results compare to thosefound by other studies is given in Table 5. Scan-napieco & Bildsten (2005); Mannucci et al. (2005)also investigated the possibility that the SN Iarate may be a two-component model, assumingn M = n SFR = 1. These analyses updated SNrates from Cappellaro et al. (1999), determiningvalues of A = 3 . +1 . − . × − SNe yr − M − ⊙ , forSNe in E/S0 galaxies (which can be crudely asso-ciated with passive galaxies in this analysis) andeither B = 1 +0 . − . × − SNe yr − (M ⊙ yr − ) − or B = 2 . ± . × − SNe yr − (M ⊙ yr − ) − , depend-ing on whether the z ≤ . B − K )galaxies is used to model the SN Ia rate in star-19 able 4Effect of our SED uncertainty on the results described in § § § Parameter Nominal Result MC (variable M) MC (variable M and SFR) BP (split) BP (not split)n M . ± .
15 0 . ± .
07 0 . ± .
08 0 . ± .
11 0 . ± . n star-forming . ± .
08 0 . ± .
03 0 . ± .
04 0 . ± .
07 0 . ± . A a − . ± . − . ± . − . ± . − . ± . − . ± . A b − . ± . − . ± . − . ± . − . ± . − . ± . SFR a . ± .
07 0 . ± .
001 0 . ± .
10 0 . ± .
06 0 . ± . SFR b . ± .
08 0 . ± .
001 0 . ± .
09 1 . ± .
06 1 . ± . B c − . ± . − . ± . − . ± . − . ± . − . ± . B d − . ± . − . ± . − . ± . − . ± . − . ± . B e − . ± . − . ± . − . ± . − . ± . − . ± . f . ± . . ± .
01 3 . ± . . ± . . ± . f . ± .
95 92 . ± . . ± . . ± . . ± . a n M free, in units of SNe yr − M − ⊙ b n M ≡
1, in units of SNe yr − M − ⊙ c n M and n SFR free, SNe yr − (M ⊙ yr − ) − n M free and n SFR ≡
1, SNe yr − (M ⊙ yr − ) − n M ≡ SFR ≡
1, SNe yr − (M ⊙ yr − ) − × − per unit mass per year forming galaxies. The value of A is consistentwith our result (2 . +0 . − . × − SNe yr − M − ⊙ ),when n M ≡ B =1 . +0 . − . × − SNe yr − (M ⊙ yr − ) − when n M ≡ SFR ≡ M and ˙ M does not providethe best-fit for our dataset. Scannapieco & Bild-sten (2005); Mannucci et al. (2005) assumed thatthe SN Ia rate depends linearly on M and ˙ M . Dil-day et al. (2008) using a sample of low redshift SNe( z < .
12) from the SDSS-II SN Survey (and over-lapping with this work) combined with other pub-lished work, used the global star-formation rate asdetermined by Hopkins & Beacom (2006) (whichmay over-estimate the total mass density) to de-termine A = (2 . ± . × − SNe yr − M − ⊙ and B = (0 . ± . × − SNe yr − (M ⊙ yr − ) − ,which are in agreement with those found in thisanalysis. Li et al. (2010) study how the SN Iarate is related to the size, color and morphologyof the host galaxy for a sample of local SNe. Theyshow that the SN Ia rate is not linearly related tothe stellar mass of the host galaxy, instead pre-ferring a relationship, SNR Ia ∝ M ∼ . , indepen-dent of host galaxy morphology and color. Theirresult for elliptical galaxies is in excellent agree-ment with our results for passive galaxies, favor- ing a SN Ia rate proportional to M . ± . . How-ever, our results differ for star-forming galaxies,where we find that an SNR Ia ∝ M . ± . isfavored. Li et al. (2010) also consider the casewhere SNR Ia ∝ M for elliptical galaxies, finding avalue of A = 4 . +0 . − . × − SNe yr − M − ⊙ (in ourframework), which is consistent with our result.
7. SNe Properties
We have studied how the rate of SNe Ia is re-lated to the host galaxy properties and have seenthat the rate of SNe is dependent the galaxy stel-lar mass and star-formation rate. We extend thisanalysis to consider how the SN Ia light-curve pa-rameters are related to the host galaxy properties.Previously studies by Hamuy et al. (1995, 2000)and Sullivan et al. (2006), for example, found thatbright SNe Ia are preferentially seen in young stel-lar environments, and Hamuy et al. (1996) showedthat there is a strong correlation between the light-curve decline rate and the host galaxy morphol-ogy. The homogeneity of the SDSS-II SN sampleprovides an ideal opportunity to determine SN Ialight-curve parameters as a function of the galaxystar-forming rate.SNe Ia have two key observables that affecttheir use as cosmological probes; their light-20 able 5A comparison of the results of § § Analysis Redshift range n M a n SFR b A a B b covered SNe yr − M − ⊙ SNe yr − (M ⊙ yr − ) − This work 0 . < z < .
25 0 . ± .
15 0 . ± .
07 1 . +4 . − . × − . +0 . − . × − This work 0 . < z < .
25 fixed = 1 fixed = 1 2 . +0 . − . × − . +0 . − . × − Sullivan et al. (2006) 0 . < z < .
75 1 . ± .
12 0 . ± .
06 - -Sullivan et al. (2006) 0 . < z < .
75 fixed = 1 fixed = 1 5 . ± . × − . ± . × − Mannucci et al. (2005) c,d low redshift fixed = 1 fixed = 1 3 . +1 . − . × − . ± . × − Dilday et al. (2008) e z < .
12 fixed = 1 fixed = 1 2 . ± . × − . ± . × − Li et al. (2010) f z < .
05 fixed = 1 - 4 . +0 . − . × − - a As derived in § b As determined in § c Results taken from Mannucci et. al. (2005) and Scannapieco & Bildsten (2005) d Approximating E/S0 galaxies for passive galaxies to determine the value of A and using the rate of SNe Ia in blue ( B − K )galaxies to determine the value of B e Using the global star-formation rate as determined by Hopkins & Beacom (2006) f Considering Elliptical galaxies as passive galaxies curve decline rate / peak brightness relationship(Phillips 1993) and their color. In MLCS2k2, therelationship between the peak luminosity of a SNIa and the shape of its light-curve, is parameter-ized through the ∆ parameter, where smaller ∆values correspond to brighter SNe Ia. The ob-served color excess of SNe Ia is modelled as thelevel of extinction in the V band, through theparameter A V . ∆ Parameter as a Functionof Host Galaxy Type
Figure 6 shows the distribution of the MLCS2k2∆ parameter for the SNe Ia found in passive andstar-forming host galaxies (shown both separatelyand as a combined dataset), after correcting forefficiency as described in § § . .
14, compared to the star-forminggalaxies, which have lower mean value ∆ = − . . . × − for the KS testand 3 . × − for the AD-test and conclude that the histograms arise from two different pop-ulations. Our result confirms previous findings(Sullivan et al. 2006; Lampeitl et al. 2010b) thatSNe in star-forming galaxies are brighter thantheir passive counter-parts and that SNe in passivegalaxies exhibit a broader range of ∆ values whencompared to their star-forming counterparts.To investigate further, we split the star-forminggalaxy sample into moderately and highly star-forming datasets (as described in § . × − for moderately star-forming, and2 . × − for highly star-forming galaxies, withcomparable values for the AD test. This showsthat SNe Ia in star-forming galaxies differ fromtheir passive counterparts, even for moderate lev-els star-formation. We find a KS test probabilityof 0 .
04 (with AD test value 0 . top panel ) is compared to those foundin star-forming galaxies. The distributions for moderately star-forming and highly star-forming galaxies areplotted in the second and third panels respectively. ( bottom panel: ) Star-forming galaxies are plotted as acumulative histogram, with the distributions for moderately star-forming and highly star-forming galaxiescombined.star-forming rate, we might have observed a sig-nificant difference between the high star-formingand moderately star-forming distributions. A V Parameter as a Func-tion of Host Galaxy Type
Determining the color of SNe Ia is important forcosmological parameter estimation. Recently, Sul-livan et al. (2010); Kelly et al. (2010) and Lampeitlet al. (2010b) found evidence that SNe in differ-ent environments may follow different color laws.Here we consider how the distribution of color, ex-pressed by the MLCS2k2 A V parameter, varies asa function of host galaxy sSFR. To examine thisrelationship we apply a flat prior in MLCS2k2, al-lowing A V to take all values (both positive andnegative), so that we are not sensitive to assump-tions about the distribution of A V values. The useof a flat prior changes the number of SNe Ia thatpass our selection criteria from 342 to 338. Theeffect of our choice of prior is discussed further in § A V for SNe inpassive hosts and star-forming galaxies. The dis-tributions for moderately star-forming and highlystar-forming galaxies are plotted separately, alongwith the case where the two star-forming datasetshave been combined. We use the efficiency correc-tion described in § /V max correctiondetermined in § A V of 0 . .
27, compared to a mean of 0 . .
15 for star-forming galaxies. For the individualstar-forming galaxy populations, we find means of0 .
43 and 0 . .
16 and 0 . § .
163 and 0 .
049 respectively, suggesting no ev-idence that the distributions may be drawn fromdifferent parent distributions. These results areconsistent with Lampeitl et al. (2010b), who used22he SDSS spectroscopically confirmed SNe fittedwith the SALT2 light-curve fitter, and saw no sig-nificant difference in the distribution of the SALT2color parameter, c , for SNe Ia’s in passive and star-forming galaxies.As before, we split the star-forming dataset into moderately and highly star-forming galaxies.We fit KS test probabilities of 0 .
42 when pas-sive and moderately star-forming datasets are con-sidered, 8 . × − for passive and highly star-forming, and 2 . × − between the two star-forming datasets, and comparable AD test statis-tics. We conclude that whilst there is no evidenceof a difference in the A V distributions between thepassive and moderately star-forming datasets, thehighly star-forming sample has a different distri-bution in A V , as shown in Figure 7, with SNe Ia inhighly star-forming galaxies on average exhibitingsmaller values of A V .This analysis assumes that the observed val-ues of A V are good approximations to the trueunderlying values. However, while the majorityof SNe in our sample have well measured light-curves, resulting in accurate measurements of A V ,many of the SNe Ia in our sample have low S/Nmeasurements. In such cases, since the underlyingdistribution of A V for SNe Ia is observed to be ex-ponentially declining, the measured value of A V ,for an individual SNe, is more likely to be scat-tered towards a higher value of A V than a lowervalue. This would result in a higher proportion ofSNe with high A V measurements compared to adistribution of SNe Ia with high S/N light-curves.To rigorously account for S/N variations in theobservations, the underlying ∆ and A V distribu-tions are determined using the method describedin D’Agostini (1995) and Appendix D of Kessleret al. (2009a). To quantify the uncertainty in theunderlying distributions, 60 data-sized simulationswere analyzed in the same way as the data. Thespread in the mean and RMS of the extracteddistributions are taken to be the uncertainties inthese quantities. To avoid pathologies from poorlymeasured photometric redshifts, only SNe Ia witha spectroscopic redshift (either from the SNe orhost galaxy) are used. The resulting incomplete-ness was modelled in simulations and found tohave a negligible impact on the results. Basedon the spread in the distribution moments (bothmean and RMS), we estimate that the distribu- tion of A V in highly star-forming galaxies differsfrom that of passive and moderately star-forminggalaxies at 3 . σ and 3 . σ , respectively.For our sample, SNe Ia have similar S/N valuesat maximum brightness, for all host galaxy types.SNe Ia in passive galaxies have a mean S/N of36 (with RMS of 26) compared to means of 33and 42 and variances of 17 and 30 for SNe Ia inmoderately star-forming and highly star-forminggalaxies, respectively.Finally, we use the SALT2 light-curve fitter(Guy et al. 2007, 2010) to determine if our re-sults are dependent on light-curve fitting tech-nique. SALT2 uses a color term, c , as a measureof the color of an individual SN Ia, but does notexplicitly attribute it to dust extinction. We re-cover the underlying distribution of c for SNe inour sample, following the technique of D’Agostini(1995), and find that the both the mean and RMSof the color distribution for SNe Ia in highly star-forming galaxies is different from those in passiveand moderately star-forming galaxies at 2 . σ and3 . σ , respectively.A physical understanding of this difference isunclear. Sargsyan et al. (2010) suggest that theremay be an increased amount of dust observed instar-burst galaxies, whilst Salim et al. (2005) ar-gue that the dust content is smaller, and coversa smaller range, in low mass, highly star-forminggalaxies.The similarity between passive and moderatelystar-forming galaxies suggests that much of thespread in color could arise from variations in theexplosion process or effects of the local SN envi-ronment (Maeda et al. 2011). If host galaxy dustwere responsible for the difference we would ex-pect to see less extinction in passive galaxies sincethey have lower dust levels (Calzetti 1998). In § § § § § § A V prior used in the MLCS2k2 light-curve fits af-fect our conclusions.Tables 6 and 7 show the KS test probabilitiesdescribed in § § A V for SNe found in passive galaxies ( top panel ) is compared to those foundin star-forming galaxies. The distributions for moderately star-forming and highly star-forming galaxies areplotted in the second and third panels respectively. The bottom panel is the sum of the two middle panels,showing the combined distribution for star-forming galaxies. A flat prior is used in the light-curve fitting.teria. We consider the “standard” A V prior dis-cussed in § A V prior, as used in § A V is forced to positive. We alsoconsider the effect of varying our redshift range,and considering only spectroscopically confirmedSNe Ia.From Table 6 we see that in all cases considered,there is a very low probability that the distribu-tion of ∆ from SNe in passive galaxies matchesthat seen in star-forming galaxies. Similarly, theevidence that highly and moderately star-forminggalaxies are different is consistently weak. Table 7,shows the KS test probabilities for the A V distri-butions. As in § A V in passive galaxies differs fromthat seen in star-forming galaxies. There is evi-dence that the distribution of A V in highly star-forming galaxies does not match that of passiveand moderately star-forming galaxies.From Figure 6, we observe that fainter, higher∆, SNe are preferentially found in passive galaxies.However, the survey efficiency considered for thisanalysis, is a function of redshift, and does not distinguish between SNe Ia of differing intrinsicbrightness. To account for this, we consider howmodelling the survey efficiency as a function ofboth A V and ∆ affects our conclusions, and findthat our results are unaffected by this additionalcorrection.In § A V (or c ) for SNe Ia as a function of hostgalaxy type are independent of light-curve fittingtechnique considered. SALT2 parameterizes therelationship between the peak luminosity of a SNIa and the shape of its light-curve through the x parameter. We recover the underlying distribu-tion of x for SNe in our sample, and find thatboth the mean and RMS of the x distribution forSNe Ia in passive galaxies differs from those foundin moderately and highly star-forming galaxies at4 . σ and 9 . σ , respectively, confirming the resultsof §
8. Conclusions
We have studied how the SN Ia rate and light-curve properties depend on the host galaxy stel-lar mass and star-formation rate. By augment-ing the SDSS spectroscopically confirmed SNe Iawith SNe identified by only their light-curves, wehave constructed a large, homogeneous and well-understood sample of 342 SNe Ia in the redshiftrange 0 . < z < .
25. Our sample has low con-tamination and is unbiased with respect to spec-troscopic selection effects and survey conditions.The efficiency of the SDSS-II SN Survey is wellmeasured in this redshift range, allowing us tostudy the overall SN Ia rate as a function of thesehost galaxy properties. We summarize below themain conclusions of this work: • We find that the SN Ia rate in passive galax-ies is not linearly proportional to the stellarmass, but instead favoring SNR Ia ∝ M . ,as illustrated in Figure 3. This result dif-fers from that of Sullivan et al. (2006), athigher redshift, who favor a linear relation-ship, but is in good agreement with theconclusions of Li et al. (2010), who favorSNR Ia ∝ M . ± . for elliptical galaxiesin the local Universe. • For star-forming galaxies we find that theSN Ia rate as a function stellar mass differsfrom that of passive galaxies, instead favor-ing SNR Ia ∝ M . . This result differs fromthat of Li et al. (2010), who found that theSN Ia rate as a function of stellar mass isindependent of host galaxy morphology andcolor. • We show that the SN Ia rate per unit stel-lar mass is a strong function of specific star-formation rate (sSFR), with SNe Ia beingpreferentially found in highly star-forming orstarburst galaxies, compared to their passivecounterparts (Figure 5). This relationshipis consistent with those found by Mannucciet al. (2005) and Sullivan et al. (2006), lo-cally and at high redshift, respectively, im-plying that this relationship does not evolvewith redshift. • We demonstrate that the excess SN Ia ratein star-forming galaxies is well fit by a linearrelationship proportional to the recent star-formation rate, as shown in Figure 4. Thecomponent related to recent star-formationis the dominant contributor to the SN Ia ratein these galaxies. • We find that a bivariate fitting techniqueconfirms that SNe Ia in this sample sat-isfy a SN Ia rate of the form SNR Ia =1 . ± . × − M . ± . + 1 . ± . × − ˙ M . ± . (statistical errors only). Thisparameterization is a generalization of theA+B model and provides a better fit to theSDSS-II SN data than assuming a SN Iarate linearly dependent on stellar mass andstar-formation rate. • We have tested the effect of our selectioncriteria on these results, and find that theexclusion of photometrically classified SNeIa’s, and variations in redshift range, do notsignificantly alter our results. • We confirm the striking difference in light-curve shape between passive and star-forming galaxies. Specifically, brighter,slowly declining SNe (with smaller ∆ val-ues for MLCS2k2) are seen preferentiallyin star-forming galaxies while faint, quicklydeclining SNe (with high ∆ values) are pref-erentially found in passive galaxies as shownin Figure 6. • We see no difference in the distribution ofthe extinction parameter, A V , between pas-sive and star-forming galaxies as illustratedin Figure 7. We find no evidence that thedistribution of A V in passive galaxies differsfrom that of moderately star-forming galax-ies, but find evidence that the distributionis different for highly star-forming galaxies,which favor lower mean values of A V . Weuse the method described in Appendix Dof Kessler et al. (2009a) to determine theunderlying distribution of A V for variousgalaxy types, and show that the distribu-tion of A V in highly star-forming galaxiesdiffers at the 3 σ level from that of passiveand moderately star-forming galaxies. Wefind that the choice of A V prior used in the25 able 6Effect of our selection criteria on the results described in § ∆ Selection No. Hosts KS-test for ∆ distributionPassive / Star-forming High / Mod.Std A V Prior 342 2 . × − . A V Prior 338 6 . × − . A V Prior 364 3 . × − . . × − . . × − . z < .
20 196 8 . × − . z < .
16 103 3 . × − . Table 7Effect of our selection criteria on the results described in § A V Selection No. Hosts KS-test for A V distributionPassive / Star-forming Passive / Mod. Passive / Highs Mod. / HighsStd A V Prior 342 0 .
295 0 .
889 6 . × − . × − Flat A V Prior 338 0 .
163 0 .
416 7 . × − . × − Positive A V Prior 364 0 .
036 0 .
923 3 . × − . × − Confirmed 197 0 .
320 0 .
843 0 .
022 3 . × − Phot-ID 145 0 .
573 0 .
390 0 .
271 0 . z < .
20 196 0 .
542 0 .
606 0 .
034 1 . × − z < .
16 103 0 .
612 0 .
032 0 .
240 8 . × − light-curve fitting does not affect our con-clusions. We find the same results using theSALT2 model to extract the distribution ofcolor, c , thus providing evidence that the dif-ference in the distribution of A V (or c ) is amodel-independent feature. • We perform a rigorous test of the P´EGASE.2SEDs and the Z-PEG fitting technique andfind a systematic offset in the photometricredshift estimates produced for our compar-ison field sample. If a simple correction is ap-plied to both the redshifts and stellar massesof our field sample, we find that our conclu-sions are unchanged.The process used to determine our sample of hostgalaxies and their derived properties allows us todirectly compare our conclusions with those of Sul-livan et al. (2006), who studied the SN Ia rateat higher redshifts. Sullivan et al. (2006) foundthe same trends with star-formation rate, but witha different relationship parameterizing the stellarmass into the SN Ia rate. It is unlikely that thesedifferences are due to an evolution of the galaxy population but may reflect that SNe Ia are primar-ily triggered by recent bursts of star-formation ina galaxy, causing uncertainties in the contributiondue to the stellar mass.
Acknowledgements . Color-Magnitude Diagram In § χ values. Figure 8 shows the color-magnitude diagram for the host galaxiesused in this analysis. The sample is split into passive, moderately and highly star-forming as described in thecaption for Figure 2. The subset of 55 galaxies with low levels of sSFR ( − . < sSFR < .
5) are plotted aslight-green triangles. We see that the P´EGASE.2 SED primarily determine the level of star-formation activityin a galaxy based on its color, with passive galaxies being the “reddest” and brightest galaxies through tothe “bluest” galaxies being classified as highly star-forming. From Figure 8, the population of objects withlow sSFR but classified as moderately star-forming are observed to lie between the passive and the othermoderately star-forming galaxies in color-magnitude space, making their classification understandable, ifambiguous. Figure 8 uses absolute magnitudes and colors, but the same conclusions can be drawn whenapparent magnitudes are considered. We note that, 43 of the 55 (78%) “ridge line” galaxies are best describedby the lenticular (S0) scenario, with the remaining 12 being best-fit by the elliptical galaxy template, possiblyhighlighting the uncertainty in the nature of lenticular galaxies. In comparison, only 41 of the 79 (52%)remaining moderately star-forming galaxies are best-fit by an S0 template.To further investigate the nature of the “ridge-line” galaxies, we study the how distribution of ∆ for SNein these galaxies compare to those in moderately star-forming and passive galaxies, since, as described in § . × − that they are drawn from the same parent distribution. This compares to a probability of 0 . B. The P´EGASE.2 Photometric Redshifts
One of the key systematic uncertainties in this analysis concerns the accuracy of the derived propertiesof the comparison field sample used, and in particular the photometric redshift estimates produced by theP´EGASE.2 SEDs. These redshift estimates will affect not only the number of field galaxies, but also theirassociated stellar masses. The photometric redshift estimates have been tested at high redshift by Sullivanet al. (2006), but have not been extensively used in the local Universe.To test the accuracy of the photometric redshifts, we use the host galaxy sample described in §
4, andwhose properties are listed in Table 9. This sample covers the magnitude range of the comparison fieldsample and is large enough to statistically determine if the photometric estimates are accurate. Figure 9shows the difference between the photometric redshift estimates and the known spectroscopic redshift forthis sample as a function of both redshift and apparent magnitude. We find a mean difference of 0 .
03 inredshift, with the photometric redshifts being smaller than the spectroscopic redshift. From Figure 9, thereis no evidence of this offset being dependent on either the redshift or apparent magnitude of the host galaxy,although the scatter does increase with apparent magnitude.This observed offset in the photometric redshift will also lead to an incorrect value for the galaxy’s stellarmass. To quantify this, we consider the derived stellar mass when the redshift is held fixed, compared to thatwhen it is allowed to float in Figure 10, resulting in the offset described above. An offset of log M = 0 . u-r ) versus absolute magnitude for the host galaxy sample used in this analysis. Galaxiesclassified as passive are plotted as red circles, with highly star-forming galaxies shown as blue diamonds.Moderately star-forming galaxies are plotted in green, with light-green triangles indicating the “ridge-line”of galaxies discussed in § C. The Effect of the Observed Offset on the Conclusions of this Work
The offset between the photometric redshifts produced by the P´EGASE.2 SEDs and the spectroscopicredshifts for the host galaxy sample implies that the distribution of galaxies in the comparison field sampleused in this analysis do not accurately reflect the distribution of galaxies in our redshift range. This mayaffect the results of §
6. Here we attempt to quantify this systematic uncertainty.In Appendix A, we observed that there is a strong dependence between the color of the host galaxy andthe best-fitting P´EGASE.2 template determined by the Z-PEG code. This is true for colors determinedboth by using absolute and apparent magnitudes. The reddest galaxies (in u − r ) are well-fit by a passivetemplate, through to the bluest galaxies, which are considered to be highly star-forming. Thus, it appearsthat, we can approximately describe the level of star-formation inferred by the P´EGASE.2 templates aspurely a function of observed quantities, and not affected by the offset described in Appendix B. We henceassume, for this analysis, that the offset in the photometric redshifts determined from the P´EGASE.2 SEDsin our redshift range purely affect the inferred stellar masses and not the star-formation rates. We notethat this approximation will only be valid for galaxies spanning a narrow redshift range, as large relativek-correction terms can lead to a color dependence, affecting the relationship determined in Figure 8.In Appendix B we showed that there is no evidence that the difference in redshift ( | z photo − z spec | )29ig. 9.— Top:
The difference between the photometric redshift estimates derived from the P´EGASE.2SEDs and the spectroscopic redshift as a function of redshift.
Bottom:
Same as above, as a function of hostgalaxy apparent magnitude. Individual galaxies are plotted in grey, with red points indicating the valuesdetermined when the sample has been binned. The black, dashed line indicates no difference, while the blue(dashed-dotted) line indicates the mean difference.and difference in stellar mass ( | log M ( z photo ) − log M ( z spec ) | ) are dependent on either redshift or apparentmagnitude. We thus assume that the photometric redshifts derived from the P´EGASE.2 SEDs and associatedstellar masses can be offset by the values determined in Appendix B. Table 8 shows the effect that correctingthe redshifts and stellar masses of the comparison field sample has on several of the key parameters discussedin this work, when the differences found in Appendix B are applied.From Table 8, it is clear that the number of field galaxies in the redshift range, 0 . < z < .
25, isdramatically reduced when this corrections is applied, but there is also an increase in stellar mass of eachgalaxy, resulting in the total stellar mass of the field sample being increased from when no correction isapplied. Consequently, the SN Ia rate per unit stellar mass per year in passive galaxies is decreased (at the2 . σ level, when only statistical errors are considered) when the correction is made. This is still in goodagreement with other measurements Sullivan et al. (2006); Mannucci et al. (2005).When considering the effect that the offset in redshift and stellar mass has on the exponents consideredfor the SN Ia rate, we see that the value of the slope determined in § . σ .30 able 8Table showing how the SN rate parameters determined in this paper are altered whenthe offsets in redshift and stellar mass, as determined in Appendix B, are applied to thecomparison field sample Parameter Original Result Mean OffsetNo. Field Galaxies a a,b .
67 9 . c . ± .
45 2 . ± . M d . ± .
150 0 . ± . SFR e . ± .
078 0 . ± . SFR (when n M = 1) f . ± .
081 1 . ± . SFR (when n M = 1) g . ± .
074 1 . ± . a After the magnitude cut (15 . < r < .
0) and redshift cut (0 . < z < . b In units of 1 × M ⊙ c The SN rate per unit stellar mass per year in passive galaxies, as describedin § × − per unit stellar mass per year d The SN rate per galaxy per year for passive galaxies as a function of logstellar mass, as described in § e As d , except for all star-forming galaxies combined, as described in § f The SN rate per galaxy per year for star-forming galaxies as a functionof log star-formation rate, after assuming a component proportional to thestellar mass, as described in § g As f , only assuming a component proportional to the values determinedin d , as described in § § . σ level), although as in the main result, there is a clear difference between the rate inpassive galaxies when compared to that in star-forming galaxies, indicating the need for a SN Ia rate that isdependent on more than stellar mass. Finally, we consider the results of § D. The P´EGASE.2 Stellar Mass and SFR Estimates
In Appendix B we compared the P´EGASE.2 photometric redshifts to a similarly distributed sample ofgalaxies with spectroscopic redshifts, and determined a bias in both redshift and stellar mass. Here weconsider how our derived properties from the P´EGASE.2 SEDs compare to those determined from thespectral features of a sample of SDSS-I galaxies. Kauffmann et al. (2003) and Brinchmann et al. (2004) usedthe 4000˚A break and the Balmer absorption line index ( Hδ A ) to measure the stellar masses and instantaneousstar-formation rates for galaxies in the SDSS-I DR4 spectroscopic catalog (Adelman-McCarthy et al. 2006).While a comparison between the stellar mass and star-formation rates determined by the P´EGASE.2templates and the results of Kauffmann et al. (2003) and Brinchmann et al. (2004) may have limitations(this sample consists of only the brightest galaxies in our host galaxy sample, the resolution of the SDSS-Ispectra is not optimal, and this method relies on the same underlying physics as the P´EGASE.2 templates),it provides a useful validation of the P´EGASE.2 measurements. We use ∼ ,
000 galaxies from the SDSS-Icatalog, limiting ourselves to the redshift range considered in this analysis. Several spectral measurementsare available; we use the dust-corrected stellar mass (median value) and total star-formation rate (medianof the likelihood distribution), and determine estimates from the P´EGASE.2 SEDs using the spectroscopicredshifts and model magnitudes.Figure 11 shows the relationship between the spectroscopic and photometric estimates of stellar mass andstar-formation rate. The total stellar mass is well recovered, with a mean offset of only ∆ log M = 0 . . .
115 and variance 0 . . E. Rest Wavelength Coverage
Throughout this analysis we have compared our observations at z < .
25 to those at higher redshift(Sullivan et al. 2006). While our methodology is identical to that used by Sullivan et al. (2006), both32nalyses use filter sets that cover the same observed wavelength range, and thus the P´EGASE.2 SED fitsare carried out over different rest-wavelength ranges. To test how this difference may affect our conclusions,specifically the comparison to Sullivan et al. (2006), we consider how carrying out our P´EGASE.2 fits usinga reduced number of filters affects the derived stellar mass and star-formation rates.The analysis of Sullivan et al. (2006) covers a bluer part of the rest-wavelength spectrum than our analysis.Thus, to produce a combination of filters that closely mimics their work, we use a reduced number of filters,removing the reddest bands from the fitting process. Specifically we investigate the cases where only the ugr and ugri filters are used.Figures 12 and 13 show the difference, as a function of apparent magnitude, between the stellar massesand star-formation rates derived by the P´EGASE.2 SEDs when various filter combinations are used. Asthe number of filters is reduced, and thus the number of data points used in the P´EGASE.2 fits is reduced,the scatter between the stellar mass and star-formation rate distribution is increased. However, in the casewhere 4 filters are considered, no significant offset is seen, with a mean difference of log M = − .
02, wheregalaxies are determined to be slightly less massive when only four filters are used. The star-formation ratesare well recovered, with a mean difference of log SFR = − . σ = 0 . . z -band is omitted.When only three filters are used to determine the derived parameters, the observed scatter increases asexpected for both the stellar mass and star-formation rate distributions. A mean difference of log M = − . σ = 0 .
23 is seen. For the star-formation rate distribution, a scatter with mean differencelog SFR = − .
12 and σ = 0 .
23 is found. ten galaxies (11 . ugr filters are used in the P´EGASE.2fits. Four galaxies (1 . i and z filters does not seem to significantly cause an offset in theP´EGASE.2 fits to higher or lower stellar masses or star-formation rates, it appears that the rest-wavelengthcoverage does not affect our comparisons to Sullivan et al. (2006).33ig. 10.— Top:
The difference between the stellar mass derived when the redshift is allowed to float in theZ-PEG code compared to that when the redshift is held fixed, resulting in the offset described in Figure 9,as a function of redshift.
Bottom:
Same as above, as a function of stellar mass, as derived when the redshiftis held fixed. Individual galaxies are plotted in grey, with red points indicating the values determined whenthe sample has been binned. The black, dashed line indicates no difference, while the blue (dashed-dotted)line indicates the mean difference. 34ig. 11.—
Top:
Log stellar mass derived from the P´EGASE.2 templates for a sample of ∼ ,
000 galaxiesfrom SDSS-I, compared to the estimates obtained from spectral features by Kauffmann et al. (2003)
Bottom:
Same as above, comparing P´EGASE.2 star-formation rates to those of Brinchmann et al. (2004). Contoursenclose 99% (dark blue), 95% (purple), 90% (red) 68% (yellow) and 35% (orange) of the data, respectively.35ig. 12.—
Top:
The stellar mass derived from the P´EGASE.2 SEDs when all five SDSS filters ugriz are usedin fit compared to when only ugri are used as a function of apparent magnitude.
Bottom:
Same as above,except only three filters ( ugr ) are used. Individual galaxies are plotted in grey, with red squares indicatingthe values determined when the sample has been binned. The black, dashed line indicates no difference,while the blue (dashed-dotted) line indicates the mean difference.36ig. 13.—
Top:
The star-formation rate derived from the P´EGASE.2 SEDs when all five SDSS filters ugriz are used in fit compared to when only ugri are used as a function of apparent magnitude.
Bottom:
Sameas above, except only three filters ( ugr ) are used. Individual star-forming galaxies are plotted in grey, withred points indicating the values determined when these galaxies have been binned. Green points (plottedat ∆ log(SFR) = 0) indicate those that are determined to be passive in both cases, while purple diamonds(shown here at ± .
6) are those which are determined to be star-forming in only one scenario. The black,dashed line indicates no difference, while the blue (dashed-dotted) line indicates the mean difference.37 able 9List of SN Ia Events used in this paper and their associated host galaxy properties. designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]762 2005eg 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
71 -10.64 sn779 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
93 -9.18 gal822 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
46 -10.64 lc911 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
81 -9.49 gal1008 2005il 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
15 -10.64 gal1032 2005ez 03 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn1241 2005ff 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn1371 2005fh 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn1415 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal1580 2005fb 03 h m . s − ◦ ‘ . “ ± . +1 . − . − . < − . < .
04 -8.70 sn1658 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
52 -9.05 lc1740 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal2031 2005fm 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
59 -9.38 sn2057 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
68 -9.64 gal2162 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal2246 2005fy 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
69 -10.64 sn2308 2005ey 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn2330 2005fp 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
37 -9.69 sn2372 2005ft 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
32 -10.64 sn2440 2005fu 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
04 -9.38 sn2561 2005fv 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
33 -10.64 sn2635 2005fw 03 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
75 -9.18 sn2639 N/A No Host Detected 0.2150 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
07 -9.80 lc2864 N/A 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
17 -99.00 gal2916 2005fz 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
07 -99.00 sn2992 2005gp 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
56 -9.90 sn3049 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
26 -9.64 gal3080 2005ga 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
54 -10.64 sn3087 2005gc 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
52 -9.20 sn3256 2005hn 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
44 -9.49 sn3331 2005ge 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
42 -10.64 sn3377 2005gr 03 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
34 -8.81 sn3451 2005gf 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
42 -10.64 sn3452 2005gg 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
18 -9.01 sn3506 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
06 -9.01 lc3592 2005gb 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
15 -9.37 sn3746 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
19 -9.18 lc3901 2005ho 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
59 -9.50 sn4019 N/A 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
33 -9.87 gal4065 N/A No Host Detected 0.1306 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
46 -8.81 gal4651 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
06 -10.64 gal4676 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
49 -9.87 gal4690 N/A 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal5103 2005gx 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
25 -9.05 sn5199 N/A 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
38 -9.18 lc5340 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
27 -99.00 sn5350 2005hp 20 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
34 -8.63 sn5395 2005hr 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
26 -8.63 sn5486 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
92 -9.59 gal5533 2005hu 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
67 -9.18 sn5549 2005hx 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -99.00 sn5550 2005hy 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
82 -8.54 sn5635 2005hv 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
69 -8.91 sn5702 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
34 -8.81 lc5735 N/A No Host Detected 0.2277 ± able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]5751 2005hz 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
18 -10.64 sn5785 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal5792 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
14 -9.37 lc5890 N/A 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
05 -10.64 gal5916 2005is 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
40 -10.64 sn5959 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
60 -10.28 gal5963 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
87 -9.39 gal5994 2005ht No Host Detected 0.1870 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
84 -9.18 sn6295 2005js 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn6304 2005jk 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
07 -9.77 sn6406 2005ij 03 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
64 -10.03 sn6422 2005id 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
71 -9.05 sn6479 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
74 -99.00 gal6491 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
43 -10.64 sn6530 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal6558 2005hj 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
23 -9.18 sn6614 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
29 -9.69 gal6773 2005iu No Host Detected 0.0903 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
24 -8.55 sn6861 N/A No Host Detected 0.1900 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
96 -10.28 gal6933 2005jc 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
09 -8.75 sn6936 2005jl No Host Detected 0.1810 ± ± h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
07 -10.03 gal7099 N/A No Host Detected 0.2184 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
18 -9.49 gal7147 2005jh 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
26 -99.00 sn7243 2005jm 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
44 -8.63 sn7335 2005kn 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
27 -8.56 sn7444 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
56 -9.49 gal7473 2005ji 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
15 -8.90 sn7512 2005jo 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
05 -8.55 sn7847 2005jp 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
29 -10.57 sn7876 2005ir 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
24 -9.25 sn8213 2005ko 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
61 -99.00 sn8254 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
59 -9.05 gal8280 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
19 -9.49 gal8297 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
08 -8.57 lc8495 2005mi 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
20 -9.87 sn8555 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
71 -9.37 gal8719 2005kp 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
03 -8.63 sn8742 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
74 -9.16 gal9052 N/A No Host Detected 0.2361 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc9467 2005lh 21 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
54 -10.64 sn9954 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
64 -9.56 gal10106 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
44 -9.90 sn11092 N/A No Host Detected 0.0843 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
97 -10.03 gal12780 2006eq 21 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
38 -99.00 sn12804 N/A 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < . < .
32 -9.18 gal12843 2006fa 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn12853 2006ey 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
07 -9.52 sn12855 2006fk 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
18 -9.64 sn12856 2006fl 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
01 -9.29 sn12860 2006fc 21 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < . < .
45 -10.28 sn able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]12874 2006fb 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
42 -9.69 sn12898 2006fw 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
32 -9.77 sn12930 2006ex 20 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
40 -9.65 sn12936 N/A No Host Detected 0.1918 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
84 -9.16 sn12964 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
29 -9.33 gal12977 2006gh 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
78 -8.75 sn13015 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
17 -9.16 lc13016 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
02 -9.18 lc13025 2006fx 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
33 -9.65 sn13038 2006gn No Host Detected 0.1040 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
52 -9.18 sn13045 2006fn 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
27 -10.57 sn13064 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc13070 2006fu 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
99 -9.29 sn13072 2006fi 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
50 -8.75 sn13135 2006fz 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
00 -99.00 sn13152 2006gg 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
21 -9.50 sn13174 2006ga 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
47 -10.64 sn13224 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
13 -10.64 gal13254 2006gx 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
85 -9.18 sn13305 2006he 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
90 -9.18 sn13323 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
04 -9.16 gal13354 2006hr 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
04 -9.68 sn13411 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
11 -9.29 sn13432 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
88 -9.69 gal13506 2006hg 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
79 -9.65 sn13511 2006hh No Host Detected 0.2376 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
10 -9.01 sn13615 N/A 20 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc13641 2006hf 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
26 -8.79 sn13703 N/A No Host Detected 0.2354 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
78 -9.41 sn13736 2006hv 22 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
30 -9.29 sn13740 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
31 -8.81 lc13768 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -8.81 lc13796 2006hl 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
81 -9.49 sn13813 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc13835 2006hp No Host Detected 0.2477 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
16 -9.87 sn13896 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
16 -8.91 lc13904 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
08 -10.19 lc13907 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal13908 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal14019 2006ki 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
71 -9.05 sn14024 2006ht 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
35 -9.94 sn14108 2006hu 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
71 -9.80 sn14113 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
56 -10.64 lc14157 2006kj 03 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
95 -9.87 sn14206 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
07 -9.01 lc14212 2006iy 22 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
03 -10.28 sn14231 N/A 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
17 -9.49 lc14284 2006ib 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
52 -10.57 sn14303 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
05 -99.00 lc14317 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
50 -10.28 gal14331 2006kl 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
69 -9.05 sn14377 2006hw 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
95 -10.16 sn14421 2006ia 02 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]14437 2006hy 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
70 -10.64 sn14444 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc14451 2006ji 20 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
19 -8.96 sn14463 N/A No Host Detected 0.2024 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc14481 2006lj 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn14525 N/A 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
63 -9.59 gal14549 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
28 -10.64 lc14561 N/A 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
86 -10.57 lc14750 N/A 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -9.87 gal14782 2006jp 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn14784 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
54 -9.87 gal14816 2006ja 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn14846 2006jn 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
96 -10.19 sn14871 2006jq 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
12 -9.18 sn14965 N/A 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
69 -9.05 lc14979 2006jr 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
77 -9.29 sn14984 2006js 20 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
16 -9.39 sn15033 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal15055 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
38 -9.37 lc15129 2006kq 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
19 -10.03 sn15132 2006jt No Host Detected 0.1440 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
52 -9.68 sn15160 N/A 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
63 -9.39 lc15161 2006jw 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
47 -9.96 sn15171 2006kb 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
05 -9.64 sn15201 2006ks 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn15203 2006jy 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
19 -10.64 sn15219 2006ka 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
68 -10.64 sn15222 2006jz 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn15229 2006kr 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < . < .
18 -9.05 sn15234 2006kd 01 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
93 -9.69 sn15251 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
54 -10.57 lc15254 2006oy 20 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
07 -9.49 sn15259 2006kc 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
40 -9.90 sn15260 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
07 -10.16 lc15286 N/A No Host Detected 0.2089 ± -9.0000 N/A N/A N/A sn15303 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
80 -9.56 gal15343 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
36 -10.64 gal15354 2006lp 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn15362 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
38 -10.64 gal15365 2006ku 23 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
58 -10.64 sn15369 2006ln 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
17 -9.32 sn15381 N/A No Host Detected 0.1620 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
98 -9.37 sn15425 2006kx 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
14 -10.57 sn15433 2006mt 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
45 -10.64 sn15443 2006lb 03 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
93 -9.59 sn15448 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
14 -9.69 lc15453 2006ky 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
25 -9.18 sn15459 2006la 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
12 -8.81 sn15461 2006kz 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
47 -10.03 sn15466 2006mz 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
06 -9.69 sn15467 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
24 -9.18 sn15496 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
20 -9.24 lc15508 2006ls 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
70 -9.18 sn15583 2006mv 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
13 -9.29 sn15587 N/A 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
29 -9.59 gal able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]15648 2006ni 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn15674 2006nu No Host Detected 0.1970 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal15719 N/A 02 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
01 -10.16 lc15722 N/A 03 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < . < .
60 -10.64 gal15748 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal15806 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
45 -10.64 gal15823 N/A 20 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
75 -10.03 gal15850 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal15868 2006pa 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
89 -9.29 sn15872 2006nb 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
06 -9.49 sn15892 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
06 -9.94 gal15897 2006pb 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn15901 2006od No Host Detected 0.2053 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal16021 2006nc 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
63 -9.90 sn16032 2006nk No Host Detected 0.2040 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
34 -9.77 gal16073 2006of 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
42 -9.38 sn16100 2006nl 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
36 -10.57 sn16103 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal16111 N/A No Host Detected 0.2246 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
76 -9.87 gal16185 2006ok 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
72 -10.64 sn16199 N/A 22 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < . < .
30 -9.49 lc16259 2006ol 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
95 -10.64 sn16276 2006om 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -9.56 sn16302 N/A 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
74 -8.62 lc16462 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal16466 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
88 -9.29 sn16467 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
23 -99.00 gal16768 N/A 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
08 -9.77 gal17168 2007ik 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
59 -9.05 sn17186 2007hx 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
06 -8.67 sn17206 N/A 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
15 -99.00 gal17332 2007jk 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
18 -10.64 sn17366 2007hz 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
74 -9.94 sn17389 2007ih 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
53 -9.59 sn17434 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
08 -9.47 gal17435 2007ka 01 h m . s +00 ◦ ‘ − . “ ± . +0 . − . − . < . < .
22 -8.63 sn17497 2007jt 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
04 -9.39 sn17568 2007kb 20 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
44 -9.68 sn17629 2007jw 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
72 -10.64 sn17647 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
44 -8.54 lc17695 N/A 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
34 -9.16 lc17745 2007ju 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
03 -8.83 sn17746 2007jv 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
68 -99.00 sn17748 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
72 -9.18 gal17801 2007ko 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
91 -10.64 sn17811 2007ix No Host Detected 0.2132 ± ± h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -10.19 lc17875 2007jz 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
03 -10.64 sn17880 2007jd 02 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
48 -9.69 sn17884 2007kt 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
79 -9.68 sn17899 N/A 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
57 -9.01 lc17906 N/A No Host Detected 0.1847 ± h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]18030 2007kq 00 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
84 -8.91 sn18083 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
08 -9.69 lc18241 2007ks 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
30 -10.19 sn18283 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
26 -10.64 lc18298 2007li 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18323 2007kx 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
06 -9.29 sn18362 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
38 -10.57 lc18375 2007lg 00 h m . s +00 ◦ ‘ − . “ ± . +0 . − . − . < − . < .
08 -10.57 sn18405 N/A 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
08 -9.69 lc18415 2007la 22 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18456 2007lk 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
38 -10.64 sn18466 2007lm 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
88 -10.16 sn18486 2007ln No Host Detected 0.2403 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
06 -9.40 sn18604 2007lp 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18610 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
20 -9.18 sn18612 2007lc 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
28 -10.64 sn18650 2007lt 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
14 -9.05 sn18651 N/A 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc18697 2007ma 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
74 -9.87 sn18740 2007mc No Host Detected 0.1570 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18751 2007ly 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18768 2007lh 01 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
39 -10.57 sn18787 2007mf 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
74 -10.64 sn18804 2007me 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
01 -9.49 sn18807 2007mg 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
06 -99.00 sn18809 2007mi 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18835 2007mj 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18855 2007mh No Host Detected 0.1278 ± ± ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
21 -10.28 sn18909 2007lq 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn18927 2007nt 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
21 -10.57 sn18940 2007sb 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
84 -9.49 sn18965 2007ne 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < .
27 -10.57 sn19008 2007mz 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
84 -9.87 sn19023 2007ls No Host Detected 0.2430 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
90 -9.69 sn19101 2007ml No Host Detected 0.1870 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
44 -9.18 sn19155 2007mn 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
63 -99.00 sn19209 N/A 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
03 -9.33 lc19220 2007ox 22 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -10.64 sn19230 2007mo 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < .
08 -10.64 sn19282 2007mk No Host Detected 0.1864 ± h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn19353 2007nj 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < . < .
73 -10.28 sn19399 N/A 03 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
57 -10.64 lc19459 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc19525 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
34 -9.86 gal19531 N/A 03 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
34 -9.18 lc19543 2007oj 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
40 -9.24 sn19545 N/A 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc19616 2007ok 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
35 -9.86 sn19658 2007ot 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
18 -8.81 sn19708 N/A 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc able 9— Continued designation host position redshift stellar mass SFR a sSFR a SN b SN ID IAU α ( J δ ( J M ⊙ ] [log M ⊙ /yr ] [ yr − ]19769 N/A 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
11 -10.03 lc19772 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal19775 2007pc 21 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
04 -99.00 sn19787 N/A 00 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal19821 N/A 23 h m . s +01 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
39 -99.00 lc19825 N/A 02 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc19833 N/A 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
41 -9.69 gal19913 2007qf 22 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
80 -9.05 sn19968 2007ol 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < .
20 -99.00 sn19969 2007pt 02 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
14 -9.38 sn19990 2007ps 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn19992 2007pb 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
99 -8.55 sn20033 N/A 01 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc20039 2007qh 00 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
95 -10.64 sn20048 2007pq 22 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn20064 2007om 23 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn20084 2007pd 23 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
44 -9.24 sn20088 N/A 00 h m . s +00 ◦ ‘ . “ ± . +0 . − . . < . < .
71 -10.64 sn20090 N/A 20 h m . s − ◦ ‘ . “ ± . +1 . − . . < . < .
50 -8.56 lc20097 2007rd 20 h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
39 -9.20 sn20111 2007pw 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn20171 N/A No Host Detected 0.2399 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 gal20350 2007ph No Host Detected 0.1295 ± h m . s − ◦ ‘ . “ ± . +0 . − . − . < . < .
65 -10.03 sn20376 2007re 21 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 sn20491 N/A No Host Detected 0.2279 ± h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
69 -9.87 sn20718 2007rj 01 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
70 -10.57 sn20721 N/A 21 h m . s − ◦ ‘ . “ ± . +0 . − . . < . < .
20 -9.24 gal20744 N/A 23 h m . s +00 ◦ ‘ . “ ± . +0 . − . − . < − . < − .
00 -99.00 lc20768 2007qq 02 h m . s − ◦ ‘ . “ ± . +0 . − . − . < − . < − .
24 -99.00 sn20821 2007rk 03 h m . s +01 ◦ ‘ . “ ± . +0 . − . . < . < .
46 -10.16 sn21033 2007qy No Host Detected 0.2290 ± a Passive galaxies are represented by -99.00 b Redshift used in host galaxy template fitting based on SN spectra (sp), galaxy spectra (gal) or light-curve (lc) EFERENCES
Abazajian, K. N., et al. 2009, ApJS, 182, 543Adelman-McCarthy, J. K., et al. 2006, ApJS, 162,38Aldering, G., et al. 2006, ApJ, 650, 510Astier, P., et al. 2006, A&A, 447, 31Baldry, I. K., et al., 2006, MNRAS, 373, 469Blanton, M. R., & Roweis, S. 2007, AJ, 133, 734Bolzonella, M., Miralles, J.-M., & Pell´o, R. 2000,A&A, 363, 476Branch, D., Livio, M., Yungelson, L. R., Boffi,F. R., & Baron, E. 1995, PASP, 107, 1019Brinchmann, J., et al. 2004, MNRAS, 351, 1151Brinchmann, J., & Ellis, R. S. 2000, ApJ, 536, L77Calzetti, D. 1998, ArXiv Astrophysics e-printsCappellaro, E., Evans, R., & Turatto, M. 1999,A&A, 351, 459Cardelli, J. A., Clayton, G. C., & Mathis, J. S.1989, ApJ, 345, 245Chakravarti, I. M., Laha, R. G., & Roy, J. 1967,Handbook of Methods of Applied Statistics,Vol. I (USE: John Wiley and Sons)D’Agostini, G. 1995, Nuclear Instruments andMethods in Physics Research A, 362, 487Dahlen, T., et al. 2004, ApJ, 613, 189D’ Andrea, C., et al., 2011, Submitted to ApJDilday, B., et al. 2008, ApJ, 682, 262Dilday, B., et al. 2010, ApJ, 713, 1026Donas, J., Deharveng, J. M., Laget, M., Milliard,B. & Huguenin, D. 1987, A&A, 180, 12Felten, J. E. 1976, ApJ, 207, 700Fioc, M., & Rocca-Volmerange, B. 1997, A&A,326, 950Frieman, J. A., et al. 2008, AJ, 135, 338Fukugita, et al., 1996, AJ, 111, 1748 Gal-Yam, A., & Maoz, D. 2004, MNRAS, 347, 942Gallagher, J. S., et al. 2005, ApJ, 634, 210Giavalisco, M., et al. 2004, ApJ, 600, L103Glazebrook, K., et al. 2004, Nature, 430, 181Graur, O., et al. 2011, arXiv:1102.0005Grazian, A., et al. 2006, A&A, 449, 951Greggio, L. 2005, A&A, 441, 1055Gunn, J. E., et al. 1998, AJ, 116, 3040Gunn, J. E., et al. 2006, AJ, 131, 2332Gupta, R., et al 2011, Submitted to ApJGuy, J., et al. 2007, A&A, 466, 11Guy, J., et al. 2010, A&A, 523, A7Guzman, R., et al. 1997, ApJ, 489, 559Hamuy, M., et al.1995, AJ, 109, 1Hamuy, M., et al. 1996, AJ, 112, 2391Hamuy, M. et al. 2000, AJ, 120, 1479H¨oflich, P., et al. 2003, Stellar Candles for the Ex-tragalactic Distance Scale, 635, 203Holtzman, J. A., et al. 2008, AJ, 136, 2306Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651,142Howell, D. A., Sullivan, M., Conley, A., & Carl-berg, R. 2007, ApJ, 667, L37Hoyle, F., & Fowler, W. A. 1960, ApJ, 132, 565Jha, S., Riess, A. G., & Kirshner, R. P. 2007, ApJ,659, 122Kasen, D., & Woosley, S. E. 2007, ApJ, 656, 661Kauffmann, G., et al. 2003, MNRAS, 341, 33Kelly, P. L., Hicken, M., Burke, D. L., Mandel,K. S., & Kirshner, R. P. 2010, ApJ, 715, 743Kessler, R., et al. 2009a, ApJS, 185, 32Kessler, R., et al. 2009b, PASP, 121, 1028Kessler, R., et al. 2010, PASP, 122, 141545ing, I. R. 1980, ApJ, 241, 474Konishi, K., et al, 2011, Submitted to ApJKroupa, P. 2001, MNRAS, 322, 231Lampeitl, H., et al. 2010a, MNRAS, 401, 2331Lampeitl, H., et al., 2010b, ApJ, 722, 566Le Borgne, D., & Rocca-Volmerange, B. 2002,A&A, 386, 446Le Borgne, et al., 2004, A&A, 425, 881Li, W., et al., 2011, MNRAS, 412, 1473Mannucci, F., Della Valle, M., & Panagia, N. 2006,MNRAS, 370, 773Mannucci, F., et al, 2005, A&A, 433, 807Maeda, K., et al. 2011, MNRAS, 413, 3075Maoz, D., Sharon, K., & Gal-Yam, A. 2010, ApJ,722, 1879Li, W., Filippenko, A. V., Della Valle, M., & Pana-gia, N. 2011, MNRAS, 412, 1508Neill, J. D., et al. 2006, AJ, 132, 1126Nobili, S., & Goobar, A. 2008, A&A, 487, 19Oemler, Jr., A., & Tinsley, B. M. 1979, AJ, 84,985Oyaizu, H., et al., 2008, ApJ, 674, 768Phillips, M. M. 1993, ApJ, 413, L105Prieto, J. L., et al. 2007, ArXiv e-prints, 706Riess, A. G., et al. 1998, AJ, 116, 1009Sako, M., et al. 2008, AJ, 135, 348Sako, M., et al. 2011, Accepted for publication inApJSalim, S., et al. 2005, ApJ, 619, 39Sargsyan, L. A., Weedman, D. W., & Houck, J. R.2010, ApJ, 715, 986Scannapieco, E., & Bildsten, L. 2005, ApJ, 629,L85Schlegel, D. J., Finkbeiner, D. P., & Davis, M.1998, ApJ, 500, 525 Schmidt, M. 1968, ApJ, 151, 393Sollerman, J., et al. 2009, ApJ, 703, 1374Stephens, M. A. 1974, Journal of the AmericanStatistical Association, 69, 730Stoughton, C., et al. 2002, AJ, 123, 485Strateva, I., et al. 2001, AJ, 122, 1861Strauss, M. A., et al. 2002, AJ, 124, 1810Sullivan, M., et al. 2006, ApJ, 648, 868Sullivan, M., et al. 2010, MNRAS, 406, 782Taylor, E. N., et al. 2009, ApJ, 694, 1171Wang, L., Hoeflich, P., & Wheeler, J. C. 1997,ApJ, 483, L29+Wood-Vasey, W. M., et al. 2007, ApJ, 666, 694York, D. G., et al. 2000, AJ, 120, 1579Yungelson, L., & Livio, M. 1998, ApJ, 497, 168Zheng, C., et al. 2008, AJ, 135, 1766