astro-ph0701912/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3: \shorttitle{Evolution in average SN Ia properties with redshift}
4: \shortauthors{Howell}
5: 
6: \begin{document}
7: %\onecolumn
8: \title{Predicted and observed evolution in the mean properties of Type~Ia supernovae
9:   with redshift}
10: \author{D. Andrew Howell, Mark Sullivan, Alex Conley, Ray Carlberg}
11: \affil{Department of Astronomy and Astrophysics, University of
12:   Toronto, 50 St. George St., Rm. 101, Toronto, ON, Canada M5S 3H4}
13: \begin{abstract}
14:   Recent studies indicate that Type Ia supernovae (SNe Ia) consist of
15:   two groups --- a ``prompt'' component whose rates are proportional
16:   to the host galaxy star formation rate, whose members have broader
17:   lightcurves and are intrinsically more luminous, and a ``delayed''
18:   component whose members take several Gyr to explode, have narrower
19:   lightcurves, and are intrinsically fainter.  As cosmic star
20:   formation density increases with redshift, the prompt component
21:   should begin to dominate.  We use a two-component model to predict
22:   that the average lightcurve width should increase by 6\% from
23:   $z=0-1.5$.  Using data from various searches we find an
24:   8.1\%$\pm$2.7\% increase in average lightcurve width for
25:   non-subluminous SNe Ia from $z=0.03 - 1.12$, corresponding to an
26:   increase in the average intrinsic luminosity of 12\%.  To test
27:   whether there is any bias after supernovae are corrected
28:   for lightcurve shape we use published data to mimic the effect of
29:   population evolution and find no significant difference in the
30:   measured dark energy equation of state parameter, $w$.  However,
31:   future measurements of time-variable $w$ will require
32:   standardization of SN Ia magnitudes to 2\% up to $z=1.7$, and it is
33:   not yet possible to assess whether lightcurve shape correction works
34:   at this level of precision.  Another concern at $z=1.5$ is the
35:   expected order of magnitude increase in the number of SNe Ia that
36:   cannot be calibrated by current methods.
37: \end{abstract}
38: 
39: \keywords{cosmology: observations --- cosmology: cosmological
40:   parameters --- supernovae: general --- surveys}
41: 
42: \section{Introduction}\label{intro}
43: Type Ia supernovae (SNe Ia) are the most important standardized
44: candles for cosmology, and have been used to discover dark energy and
45: the accelerating universe \citep{1998AJ....116.1009R,
46:   1999ApJ...517..565P}.  This was facilitated by the realization that
47: supernovae with broader lightcurves are intrinsically brighter, while
48: those with narrow lightcurves are dimmer \citep{1993ApJ...413L.105P}.
49: Various schemes exist to correct SN Ia luminosities based on their
50: lightcurve shape \citep[e.g.][]{1996ApJ...473...88R,
51:   2003ApJ...594....1T, 2006ApJ...647..501P,
52:   2007A&A...466...11G,2007ApJ...659..122J} --- here we use the
53: ``stretch'' method, in which the time axis of a template lightcurve is
54: multiplied by a scale factor $s$ to fit the data
55: \citep{1997ApJ...483..565P}.
56: 
57: Some properties of SNe Ia have been found to correlate with
58: environment --- brighter supernovae with broader lightcurves (high
59: $s$) tend to occur in late-type spiral galaxies
60: \citep{1995AJ....109....1H}, while dimmer, fast declining (low $s$)
61: supernovae are preferentially located in an older stellar population,
62: leading to the conclusion that the age of the progenitor system is a
63: key variable affecting SN Ia properties \citep{2001ApJ...554L.193H}.
64: The fact that supernovae occur at a much higher rate in late type
65: galaxies, and that the SN Ia rate is proportional to the core-collapse
66: rate \citep{2005A&A...433..807M} is another indication that age plays
67: an essential role.  Following this previous work,
68: \citet{2005ApJ...629L..85S} model SNe Ia as consisting of two
69: populations --- a ``prompt'' component whose rate is proportional to
70: the star formation rate of the host galaxy, and a second ``delayed''
71: component whose rate is proportional to the stellar mass of the
72: galaxy.  \citet[hereafter S06]{2006ApJ...648..868S} tie all of these
73: results together using data from the Supernova Legacy Survey (SNLS),
74: finding that slow declining, brighter SNe Ia come from a young
75: population and have a rate proportional to star formation on a 0.5 Gyr
76: timescale, while dimmer, faster declining supernovae come from a much
77: older population with a rate proportional to the mass of the host
78: galaxy.
79: 
80: S06 and \citet{2006MNRAS.370..773M} predict that the
81: SNe Ia whose rates are proportional to star formation will start to
82: dominate the total sample of SNe as cosmic star formation
83: increases with redshift.  Since these SNe are intrinsically
84: brighter, the mean luminosity of the population should increase with
85: redshift.  Here we use the two-component SN Ia model of 
86: \citet{2005ApJ...629L..85S} and the stretch distribution for
87: each component from SNLS data (S06) to quantify the expected magnitude
88: of this effect.  We then compare the predicted evolution in lightcurve
89: stretch to SN distributions from the SNLS and the Higher-z SN Search
90: \citep{2007ApJ...659...98R}.
91: 
92: 
93: \section{Predicting Evolution}\label{method}
94: \citet{2005ApJ...629L..85S} parameterize the supernova rate in a
95: galaxy as $$\rm{SNR(t)}=\it A M(t) + B \dot{M}(t),$$
96: where $M(t)$ is
97: the total stellar mass in the galaxy, $\dot{M}$ is the star formation
98: rate, and A and B are constants.  These authors 
99: use the supernova
100: rates in galaxies of different morphologies and colors to derive
101: values for $A$ and $B$.  S06 use an alternate method --- they fit
102: galaxy models to SNLS $u^*g\arcmin r\arcmin i\arcmin z\arcmin$ host
103: galaxy photometry to derive masses and star formation rates.  Then,
104: using the cosmic star formation history of
105: \citet{2006ApJ...651..142H}, S06 predict the rate of SNe from each
106: component versus redshift in their Fig. 10.  Note that here we adopt
107: the same definition of $\dot{M}(t)$ as S06 --- it is the mass turned
108: into stars and does not include mass loss from supernovae.
109: 
110: \begin{figure}
111: \epsscale{0.7}
112: \plotone{f1_color.eps}
113: \caption{Stretch distributions of SNe from each component of the two
114:   component model.  {\em Top:} 
115: Prompt ($B$-component) SNe Ia.  {\em Bottom:} Delayed ($A$-component)
116: SNe Ia.  The distributions were derived from those in S06, as
117: described in the text.  Best-fit Gaussians are shown.  The
118: $A$-component Gaussian is centered at $s=0.945$, with $\sigma=0.077$.
119: The $B$-component Gaussian is centered at $1.071$ with $\sigma=
120: 0.063$.}
121: \label{abdist}
122: \end{figure}
123: 
124: 
125: The prompt and delayed SN Ia components have different stretch
126: distributions (S06).  
127: To determine the stretch of each SN, here we use
128: the SNe from S06, though we fit a new lightcurve template to the data
129: using the SiFTO method \citep{2007arXiv0705.0367C,2007conley}.
130: Because stretches are always defined relative to the $s=1$ template
131: lightcurve, stretch values should only compared within a publication.
132: However, the stretches derived here are approximately 4\% larger than
133: those in \citet{2006A&A...447...31A}, largely due to the use of a
134: narrower $s=1$ template.
135: 
136: All SNe from passive galaxies (i.e. those with no measurable star
137: formation rate) were assigned to the $A$ component.  Star forming
138: galaxies have SNe Ia from both components, so the $A$ distribution
139: from passive galaxies was scaled by mass and subtracted from the
140: distribution of SNe Ia from star forming galaxies, leaving the $B$
141: distribution (as in S06).  The resulting distributions, and Gaussian
142: fits are shown in Figure~\ref{abdist}.  Note that to conserve the
143: total number of SNe one should add the SNe subtracted from the $B$
144: distribution back to the $A$ distribution -- this is unnecessary for
145: our purposes because the relative heights of the gaussians are
146: normalized as a function of redshift in the next step.
147: 
148: To estimate the expected stretch evolution with redshift, we take the
149: observed $A$ and $B$ distributions and scale them to the predicted
150: relative values with redshift from Fig. 10 of S06.  Increasing cosmic
151: star formation with redshift produces a larger fraction of SNe from
152: the prompt component.  Stellar mass as a function of redshift is
153: determined by integrating the star formation history from the earliest
154: times, so the total stellar mass, and the number of SNe from the A
155: component, decreases with increasing redshift.  The net result is that
156: in the $A+B$ model the mean stretch increases from 0.98 at $z=0$ to
157: 1.04 at $z=1.5$.
158: 
159: One caveat is that in the $A+B$ model there is no time dependence for
160: the $A$ component.  SNe Ia from 10 Gyr old progenitors are just as
161: likely as SNe Ia from 3 Gyr old progenitors.  If 10 Gyr-old SNe Ia are
162: actually more rare, the $A+B$ model will overpredict the number of
163: $A$-component SNe at $z=0$, as they result from stars formed during
164: the high star formation rate in the early universe (see discussion in
165: S06).  As an alternative to the $A+B$ model we tested the two
166: component SN Ia delay time distribution from
167: \citet{2006MNRAS.370..773M}, which has an exponential decrease in
168: supernova probability from the delayed component with time.  The
169: drawback of this model is that the probability distribution is
170: somewhat arbitrary.  Also, rather than the 50-50 split between prompt
171: and delayed SNe chosen by \citet{2006MNRAS.370..773M}, here we scale
172: each component by the $A$ and $B$ values measured by S06.  This gives
173: similar results to the $A+B$ model, predicting a shift in mean stretch
174: from 0.98 to 1.02 from $z=0-1.5$.  
175: 
176: \section{Comparison to Observations}
177: In Figure~\ref{gaussreal} we compare the predicted stretch
178: distributions from the $A+B$ model to the observed stretch
179: distributions in three redshift bins from the low redshift data used
180: by \citet{2006A&A...447...31A}, the SNLS data in S06, and the data of
181: the higher-z supernova search \citep{2007ApJ...659...98R}.  All
182: lightcurves have been refit here using the same method.  We also
183: tested the data against the modified \citet{2006MNRAS.370..773M}
184: model, but we did not find it to be a better predictor of SN evolution
185: with redshift (Table~\ref{table}). 
186: 
187: \begin{deluxetable}{ccccc}
188: \tablewidth{0pt} 
189: \tablecolumns{5}
190: \tablecaption{$\chi^2$ and KS test: data and models\label{table}} 
191: \tablehead{
192: \colhead{} & \multicolumn{2}{c}{$\chi^2$} & \multicolumn{2}{c}{KS} \\
193: %\cline{1-2} \cline{3-4}
194: \colhead{z} & \colhead{A+B} &\colhead{MM} & \colhead{A+B} & \colhead{MM} 
195: } 
196: \startdata
197: 0-0.1 & 0.81 & 0.63 & 15\% & 39\% \\
198: 0.1-0.75 & 0.64 & 0.83 & 30\% & 28\% \\
199: 0.75-1.5 & 0.60 & 0.84 & 52\% & 35\% \\
200: \enddata 
201:   
202: \tablecomments{Cols. 2-3: the $\chi^2$ per
203:     degree of freedom between the data and the predictions of the A+B
204:     and modified Mannucci (MM) models.  Cols. 4-5: KS-test probability that
205:     the data is drawn from each model.  Bins with zero counts were
206:     assigned an error of 1.15, possibly underestimating the $\chi^2$.
207:     The KS test is also imperfect because probabilities were
208:     derived for a single Gaussian, not the sum of two Gaussians as
209:     used here.  In both cases the two model components were fixed by
210:     the A and B numbers in S06.}
211: \end{deluxetable}
212: 
213: 
214: Each survey has different selection effects -- the most serious for
215: the current study is Malmquist bias, the tendency to discover only the
216: brightest members of a group near the detection limit of a
217: magnitude-limited survey.  To minimize the effect, for each of the
218: high redshift searches we only consider supernovae from a reduced
219: volume so that none of the supernovae used are near the magnitude
220: limit.  The SNLS regularly discovers SNe Ia out to $z>1$, but here we
221: use only the subset with $z \leq 0.75$, where Malmquist bias is
222: minimal \citep{2006A&A...447...31A}.  Similarly, we only use
223: \citet{2007ApJ...659...98R} SNe with $z<1.5$, where the authors report
224: their sample is complete \citep{2004ApJ...613..200S}.  Lowering the
225: redshift cuttoff to $z=1.2$ does not change the average stretch for
226: the highest-z sample, but it reduces the sample size from 20 to 13,
227: and thus decreases the significance of the results.
228: 
229: As an additional protection against selection bias, we only consider
230: SNe with $s \geq 0.7$.  SNe Ia with $s<0.7$ are both dim and
231: spectoscopically peculiar, like SN 1991bg \citep{1992AJ....104.1543F},
232: and have not yet been detected at $z>0.2$, probably because of a
233: combination of Malmquist bias and spectroscopic selection bias
234: \citep{2001ApJ...554L.193H,2005ApJ...634.1190H} -- as redshift
235: increases, and the angular size of the host galaxy decreases, and it
236: becomes more and more difficult to spectroscopically identify such
237: faint supernovae when blended with their bright, often elliptical,
238: hosts.  This cut removes 3 SNe Ia from the low-z sample [other low-z
239: SNe are already removed because we only consider Hubble-flow SNe Ia,
240: with $z>0.015$, to be consistent with \citet{2006A&A...447...31A}].
241: 
242: In all cases we use only SNe Ia with at least 4 lightcurve points, and
243: at least one detection before 10 rest-frame days after maximum light in the
244: $B$-band, so that stretch is accurately measured.
245: 
246: \begin{figure}
247: \plotone{f2_color.eps}
248: \caption{Actual stretch distributions compared to predictions from the
249:   $A+B$ model.  In each case the $A+B$ model evaluated at the median
250:   redshift of the distribution is shown.  {\em Top:} SNe Ia from
251:   $z=0-0.1$ (median $z=0.026$; N=50).  {\em Middle:} SNLS SNe Ia from
252:   $z=0.1-0.75$ (median $z=0.55$; N=99).  {\em
253:     Bottom:} \citet{2007ApJ...659...98R} SNe Ia from $z=0.75-1.5$
254:   (median $z=1.12$; N=20), with
255:   $s$ errors $< 0.2$ (requiring $s$ error $\leq 0.1$ reduces the
256:   sample to 16 and increases the average stretch from 1.06 to 1.07).  
257:   The vertical line gives the mean stretch for each distribution.
258:   The top axis converts a stretch difference into a magnitude
259:   difference using $\alpha=1.5$.}
260: \label{gaussreal}
261: \end{figure}
262: 
263: 
264: Figure~\ref{gaussreal} shows that the average observed stretch
265: increases with redshift, from $s=0.98\pm 0.02$ at a median redshift of
266: $z=0.03$, to $s=1.02\pm 0.01$ at $z=0.56$, and $s=1.06\pm 0.02$ at
267: $z=1.12$.  Simultaneously the percentage of SNe Ia with $s<0.9$
268: decreases from 24\% to 15\% to 1.4\%.  The KS test gives a 2\%
269: probability that the lowest and highest redshift sample are drawn from
270: the same distribution.  The predicted distributions from the $A+B$
271: model are overplotted.  The observed trends match the predictions of
272: the empirically-based models --- with increasing redshift fewer
273: low-stretch SNe Ia are observed, and the mean SN Ia stretch increases.
274: We find the same result when this analysis is repeated with the SALT
275: \citep{2005A&A...443..781G} and SALT2 \citep{2007A&A...466...11G}
276: lightcurve fitters.    These results are also consistent with the
277: findings of \citet{2006A&A...447...31A}, that the low-z sample had an
278: average stretch 97\% that of SNLS SNe.
279: 
280: 
281: \citet{2006A&A...447...31A} estimate distances from SNe Ia using 
282: $$\mu_B = m^*_B -M + \alpha (s-1) - \beta c,$$ where $\mu_B$ is the 
283: distance modulus, $m^*_B$ is the peak $B$-band magnitude, $c$ is a
284: color, and $M$, $\alpha$ and $\beta$ are parameters fit by minimizing
285: residuals on the Hubble diagram.
286: \citet{2006A&A...447...31A} found $\alpha = 1.52$, so a
287: drift in average stretch of $0.08 \pm 0.026$ from $z=0.03$ to $z=1.12$
288: results in a 12\% drift in average intrinsic SN Ia luminosity
289: over this redshift range.
290: 
291: \section{Effect on Cosmological Studies}
292: Evolution in the SN population will not necessarily bias cosmological
293: studies, since SNe are only used in this way after correction for
294: lightcurve shape.  However, we can no longer assume that any
295: deficiencies in lightcurve width correction schemes will average out
296: under the assumption the distribution of SNe is similar over 
297: all redshifts.  If there is a systematic residual between
298: low stretch and high stretch SNe when they are stretch
299: corrected, this could cause a bias in the determination of
300: cosmological parameters as the population evolves.
301: 
302: \begin{figure}
303: \plotone{f3.eps}
304: \caption{Cosmological fits done with the \citet{2006A&A...447...31A}
305:   sample (dashed), and with SNe Ia drawn from the same sample, but
306:   using only $s<1$ SNe Ia at $z<0.4$ and $s\geq 1$ SNe Ia at 
307:   $z\geq 0.4$ (solid).  Combining with Baryon Acoustic Oscillation results
308:   \citep{2005ApJ...633..560E} produces the thick lines.
309:   The results are consistent at the one sigma level.}
310: \label{lowhighs}
311: \end{figure}
312: 
313: To test an extreme case of evolution, we fit the equation of state
314: parameter for dark energy, $w$ and the matter density, $\Omega_M$
315: using the data from \citet{2006A&A...447...31A}, (dashed lines in
316: Fig.~\ref{lowhighs}) and compared it to a fit using the same data, but
317: retaining only $s<1$ SNe Ia at $z<0.4$ and only $s\geq 1$ SNe Ia at
318: $z\geq 0.4$ (solid lines).  The strongly evolving subset gives estimates
319: for $w$ and $\Omega_M$ consistent with the full set, although
320: the errors are larger because there are fewer SNe in the subset.
321: 
322: One worry with an evolving population is that stretch-magnitude or
323: color-magnitude relations may evolve, i.e. $\alpha$ or $\beta$ derived
324: at one redshift may not be appropriate at another.  The values derived
325: here for the strong evolution subset and the full set are consistent,
326: but again a strong test awaits a larger data set.
327: 
328: Measuring changes in $w$ with time will require much stricter control
329: of potential evolutionary effects.  Possible biases depend critically on the
330: exact nature of the evolution, the experimental design, the cosmology, and how the time 
331: variable component of $w$ is parameterized --- e.g. $w\prime $, $w_1$,
332: or $w_a$
333: \citep{2002PhRvD..65j3512W,2004MNRAS.347..909K,2006APh....26..102L}.
334: However, a good rule of thumb is that to keep systematic errors 
335: significantly below statistical errors for a mission such as SNAP, the
336: corrected magnitudes of SNe Ia should not drift by more than 0.02 mag
337: up to $z=1.7$ \citep{2004MNRAS.347..909K}.  Unfortunately there are
338: not enough well measured SNe Ia in the literature to determine whether
339: or not stretch correction works to this level of precision.  
340: After stretch correction, the rms scatter of SNe Ia around the Hubble 
341: line is $\sim 0.2$ mag in the best cases \citep{2007A&A...466...11G}.
342: Therefore $\sim 100$ SNe Ia are required in each of several stretch or
343: redshift bins to determine whether biases remain at the 0.02 level 
344: after correction.  Such precise tests will soon be possible by
345: combining data from large surveys underway at low and high redshift.
346: 
347: \section{Discussion}
348: We have shown that there is some evolution in the average lightcurve
349: width, and thus intrinsic luminosity of SNe Ia from $z=0$ to $z=1.5$,
350: although significant evolution is found only over a large redshift
351: baseline.  This evolution is consistent with our predictions from the
352: $A+B$ model -- as star formation increases with redshift, the
353: broader-lightcurve SNe Ia associated with a young stellar population
354: make up an increasingly larger fraction of SNe Ia.
355: 
356: Though we have taken steps to minimize the effects of Malmquist bias, 
357: it is possible that residual 
358: effects play some role in increasing average stretch with redshift.
359: However, the net effect on cosmology studies is the same no matter 
360: the underlying cause.  In either case, there is increased pressure on
361: the light curve shape calibration method to correct for the evolution in 
362: SN Ia properties with redshift.
363: 
364: \citet{1999ApJ...517..565P} found that SNe Ia still give evidence for
365: an accelerating universe even if SNe Ia are not corrected for stretch.
366: This was possible because the difference in a universe with dark
367: energy and one with $\Omega_M=1$ is large --- 0.25 magnitudes at
368: $z=0.7$, whereas the population evolution seen here implies that average SN
369: Ia magnitudes should increase by 0.07 mag over the same redshift
370: range.  However, discriminating between dark energy models requires
371: much more precise control of SN Ia magnitudes over a larger redshift
372: baseline. 
373: 
374: Most theoretical studies addressing possible SN Ia evolution have
375: focused on metallicity.  Although theorists have proposed various
376: mechanisms that could conceivably alter the properties of SNe Ia as
377: the average metallicity changes with cosmic time
378: \citep{1998ApJ...495..617H,2000ApJ...530..966L, 2001ApJ...557..279D,
379:   2003ApJ...590L..83T} there is no consensus
380: regarding which effects are important or even the sign of these
381: effects.  Observational studies have found no evidence that
382: metallicity affects the properties of SNe Ia
383: \citep{2000AJ....120.1479H,2000ApJ...542..588I,2005ApJ...634..210G,2007ellis}.
384: Instead, it is more likely that age differences between the two
385: populations (and thus almost certainly the mass of the secondary star)
386: play a role in the evolution of the observed stretch distribution with
387: redshift \citep{2001ApJ...554L.193H}.
388: 
389: Prompt SNe Ia are thought to be brighter because they produce more $^{56}$Ni.
390: If the Chandrasekhar-mass model describes most SNe Ia, they must then
391: produce less intermediate mass elements, assuming that the amount of
392: unburned material is negligible in normal SNe Ia.  We therefore
393: predict that high redshift SNe Ia will have less Ca and Si.  
394: This is confirmed by the most intensive study of
395: high redshift SN Ia spectra \citep{2007ellis}.
396: 
397: One concern raised by these findings is that pathological SNe Ia such
398: as SN 2001ay \citep{howell2004}, SN 2002cx
399: \citep{2003PASP..115..453L}, SN 2002ic \citep{2003Natur.424..651H},
400: and SNLS-03D3bb \citep{2006Natur.443..308H},
401: which do not obey typical lightcurve shape correction schemes, are
402: associated with star formation.  Since star formation density
403: increases by a factor of
404: ten from $z=0$ to $z=1.5$ \citep{2006ApJ...651..142H},
405: at high redshift these pathological supernovae will be an order of
406: magnitude more common.  Thus the conventional wisdom that all high
407: redshift supernovae will have counterparts at low redshift 
408: \citep{2001astro.ph..9070B} only holds if sample sizes are much larger
409: than those currently used for cosmology.  Only 20 SNe~Ia have
410: published lightcurves at $z>1$
411: \citep{2007ApJ...659...98R,2006A&A...447...31A}, and only $\sim $50
412: SNe Ia at $z<0.1$ have sufficient data to be cosmologically useful.
413: Fortunately, thus far it has been possible to identify these outliers
414: so that they do not affect cosmological analyses, but future studies
415: requiring increased precision must be vigilant of the effects of
416: an evolving SN Ia population.
417: 
418: \acknowledgements
419: We thank Richard Ellis, Peter Garnavich, Julien Guy, Eric Linder, Peter Nugent, and an
420: anonymous referee for helpful comments and the Canadian NSERC for support.
421: 
422: \bibliographystyle{apj}
423: %\bibliography{astro}
424: 
425: \begin{thebibliography}{41}
426: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
427: 
428: \bibitem[{{Astier} {et~al.}(2006)}]{2006A&A...447...31A} 
429: {Astier}, P. {et~al.} 2006, \aap, 447, 31 
430: 
431: \bibitem[{{Branch} {et~al.}(2001)}]{2001astro.ph..9070B} 
432: {Branch}, D., {Perlmutter}, S., {Baron}, E., \& {Nugent}, P.  2001, ArXiv  
433: Astrophysics e-prints 
434: 
435: \bibitem[{{Conley} {et~al.}(2007{\natexlab{a}})}]{2007arXiv0705.0367C} 
436: {Conley}, A., {Carlberg}, R.~G., {Guy}, J., {Howell}, D.~A., {Jha}, S.,  
437: {Riess}, A.~G., \& {Sullivan}, M.  2007{\natexlab{a}}, ArXiv e-prints, 705 
438: 
439: \bibitem[{{Conley} {et~al.}(2007{\natexlab{b}})}]{2007conley} 
440: {Conley}, A. 2007{\natexlab{b}}, in preparation 
441: 
442: \bibitem[{{Dom{\' i}nguez}, {H{\" o}flich}, \&  
443: {Straniero}(2001)}]{2001ApJ...557..279D} 
444: {Dom{\' i}nguez}, I., {H{\" o}flich}, P., \& {Straniero}, O.  2001, \apj, 
445: 557,  279 
446: 
447: \bibitem[{{Eisenstein} {et~al.}(2005)}]{2005ApJ...633..560E} 
448: {Eisenstein}, D.~J. {et~al.} 2005, \apj, 633, 560 
449: 
450: \bibitem[{{Ellis} {et~al.}(2007)}]{2007ellis} 
451: {Ellis}, R. 2007, in preparation 
452: 
453: \bibitem[{{Filippenko} {et~al.}(1992)}]{1992AJ....104.1543F} 
454: {Filippenko}, A.~V. {et~al.} 1992, \aj,  104, 1543 
455: 
456: \bibitem[{{Gallagher} {et~al.}(2005)}]{2005ApJ...634..210G} 
457: {Gallagher}, J.~S., {Garnavich}, P.~M., {Berlind}, P., {Challis}, P., 
458: {Jha},  S., \& {Kirshner}, R.~P.  2005, \apj, 634, 210 
459: 
460: \bibitem[{{Guy} {et~al.}(2007)}]{2007A&A...466...11G} 
461: {Guy}, J. {et~al.} 2007, \aap, 466, 11 
462: 
463: \bibitem[{{Guy} {et~al.}(2005)}]{2005A&A...443..781G} 
464: {Guy}, J., {Astier}, P., {Nobili}, S., {Regnault}, N., \& {Pain}, R.  2005, 
465:  \aap, 443, 781 
466: 
467: \bibitem[{{Hamuy} {et~al.}(1995)}]{1995AJ....109....1H} 
468: {Hamuy}, M., {Phillips}, M.~M., {Maza}, J., {Suntzeff}, N.~B., {Schommer},  
469: R.~A., \& {Aviles}, R.  1995, \aj, 109, 1 
470: 
471: \bibitem[{{Hamuy} {et~al.}(2003)}]{2003Natur.424..651H} 
472: {Hamuy}, M. {et~al.} 2003, \nat, 424, 651 
473: 
474: \bibitem[{{Hamuy} {et~al.}(2000)}]{2000AJ....120.1479H} 
475: {Hamuy}, M., {Trager}, S.~C., {Pinto}, P.~A., {Phillips}, M.~M., 
476: {Schommer},  R.~A., {Ivanov}, V., \& {Suntzeff}, N.~B.  2000, \aj, 120, 
477: 1479 
478: 
479: \bibitem[{{H\"oflich}, {Wheeler}, \&  
480: {Thielemann}(1998)}]{1998ApJ...495..617H} 
481: {H\"oflich}, P., {Wheeler}, J.~C., \& {Thielemann}, F.~K.  1998, \apj, 495, 
482: 617 
483: 
484: \bibitem[{{Hopkins} \& {Beacom}(2006)}]{2006ApJ...651..142H} 
485: {Hopkins}, A.~M. \& {Beacom}, J.~F.  2006, \apj, 651, 142 
486: 
487: \bibitem[{{Howell}(2001)}]{2001ApJ...554L.193H} 
488: {Howell}, D.~A.  2001, \apjl, 554, L193 
489: 
490: \bibitem[{{Howell} \& {Nugent}(2004)}]{howell2004} 
491: {Howell}, D.~A. \& {Nugent}, P.  2004, in Cosmic Explosions in Three  
492: Dimensions, 151 
493: 
494: \bibitem[{{Howell} {et~al.}(2006)}]{2006Natur.443..308H} 
495: {Howell}, D.~A. {et~al.} 2006, \nat, 443, 308 
496: 
497: \bibitem[{{Howell} {et~al.}(2005)}]{2005ApJ...634.1190H} 
498: {Howell}, D.~A. {et~al.} 2005, \apj, 634, 1190 
499: 
500: \bibitem[{{Ivanov}, {Hamuy}, \&  {Pinto}(2000)}]{2000ApJ...542..588I} 
501: {Ivanov}, V.~D., {Hamuy}, M., \& {Pinto}, P.~A.  2000, \apj, 542, 588 
502: 
503: \bibitem[{{Jha}, {Riess}, \&  {Kirshner}(2007)}]{2007ApJ...659..122J} 
504: {Jha}, S., {Riess}, A.~G., \& {Kirshner}, R.~P.  2007, \apj, 659, 122 
505: 
506: \bibitem[{{Kim} {et~al.}(2004)}]{2004MNRAS.347..909K} 
507: {Kim}, A.~G., {Linder}, E.~V., {Miquel}, R., \& {Mostek}, N.  2004, \mnras, 
508: 347,  909 
509: 
510: \bibitem[{{Lentz} {et~al.}(2000)}]{2000ApJ...530..966L} 
511: {Lentz}, E.~J., {Baron}, E., {Branch}, D., {Hauschildt}, P.~H., \& 
512: {Nugent},  P.~E.  2000, \apj, 530, 966 
513: 
514: \bibitem[{{Li} {et~al.}(2003)}]{2003PASP..115..453L} 
515: {Li}, W. {et~al.} 2003, \pasp, 115, 453 
516: 
517: \bibitem[{{Linder}(2006)}]{2006APh....26..102L} 
518: {Linder}, E.~V.  2006, Astroparticle Physics, 26, 102 
519: 
520: \bibitem[{{Mannucci}, {Della Valle}, \&  
521: {Panagia}(2006)}]{2006MNRAS.370..773M} 
522: {Mannucci}, F., {Della Valle}, M., \& {Panagia}, N.  2006, \mnras, 370, 773 
523: 
524: \bibitem[{{Mannucci} {et~al.}(2005)}]{2005A&A...433..807M} 
525: {Mannucci}, F., {della Valle}, M., {Panagia}, N., {Cappellaro}, E., 
526: {Cresci},  G., {Maiolino}, R., {Petrosian}, A., \& {Turatto}, M.  2005, 
527: \aap, 433, 807 
528: 
529: \bibitem[{{Perlmutter} {et~al.}(1999)}]{1999ApJ...517..565P} 
530: {Perlmutter}, S. {et~al.} 1999, \apj, 517, 565 
531: 
532: \bibitem[{{Perlmutter} {et~al.}(1997)}]{1997ApJ...483..565P} 
533: {Perlmutter}, S. {et~al.} 1997,  \apj, 483, 565 
534: 
535: \bibitem[{{Phillips}(1993)}]{1993ApJ...413L.105P} 
536: {Phillips}, M.~M.  1993, \apjl, 413, L105 
537: 
538: \bibitem[{{Prieto}, {Rest}, \&  {Suntzeff}(2006)}]{2006ApJ...647..501P} 
539: {Prieto}, J.~L., {Rest}, A., \& {Suntzeff}, N.~B.  2006, \apj, 647, 501 
540: 
541: \bibitem[{{Riess} {et~al.}(1998)}]{1998AJ....116.1009R} 
542: {Riess}, A.~G. {et~al.} 1998, \aj, 116, 1009 
543: 
544: \bibitem[{{Riess}, {Press}, \&  {Kirshner}(1996)}]{1996ApJ...473...88R} 
545: {Riess}, A.~G., {Press}, W.~H., \& {Kirshner}, R.~P.  1996, \apj, 473, 88 
546: 
547: \bibitem[{{Riess} {et~al.}(2007)}]{2007ApJ...659...98R} 
548: {Riess}, A.~G. {et~al.} 2007, \apj, 659, 98 
549: 
550: \bibitem[{{Scannapieco} \& {Bildsten}(2005)}]{2005ApJ...629L..85S} 
551: {Scannapieco}, E. \& {Bildsten}, L.  2005, \apjl, 629, L85 
552: 
553: \bibitem[{{Strolger} {et~al.}(2004)}]{2004ApJ...613..200S} 
554: {Strolger}, L.-G. {et~al.} 2004, \apj, 613, 200 
555: 
556: \bibitem[{{Sullivan} {et~al.}(2006)}]{2006ApJ...648..868S} 
557: {Sullivan}, M. {et~al.} 2006, \apj, 648, 868 
558: 
559: \bibitem[{{Timmes}, {Brown}, \&  {Truran}(2003)}]{2003ApJ...590L..83T} 
560: {Timmes}, F.~X., {Brown}, E.~F., \& {Truran}, J.~W.  2003, \apjl, 590, L83 
561: 
562: \bibitem[{{Tonry} {et~al.}(2003)}]{2003ApJ...594....1T} 
563: {Tonry}, J.~L. {et~al.} 2003, \apj, 594, 1 
564: 
565: \bibitem[{{Weller} \& {Albrecht}(2002)}]{2002PhRvD..65j3512W} 
566: {Weller}, J. \& {Albrecht}, A.  2002, \prd, 65, 103512 
567: 
568: \end{thebibliography}
569: 
570: %\begin{thebibliography}{38}
571: %\end{thebibliography}
572: 
573: \end{document}
574: 
575: 
576: 
577: