THE CARNEGIE SUPERNOVA PROJECT: SECOND PHOTOMETRY DATA RELEASE OF LOW-REDSHIFT TYPE Ia SUPERNOVAE

The current cosmological paradigm (Λ cold dark matter) suggests that the universe is composed of ∼4.4% baryonic matter, ∼21.3% cold dark matter, and ∼74.2% dark energy (e.g., Dunkley et al. 2009). The luminosity distances of Type Ia supernovae (SNe Ia) provided the first strong evidence that the expansion rate of the universe is currently accelerating (Riess et al. 1998; Perlmutter et al. 1999), and that dark energy having negative pressure is the major contributor to the overall energy budget of the universe. These findings have been further substantiated through the detection of the late-time-integrated Sachs–Wolfe effect (Giannantonio et al. 2008) and X-ray cluster distances (e.g., Allen et al. 2007).

To date, much effort and resources have been directed toward characterizing the nature of dark energy, resulting in more than 1000 SNe Ia being observed, extending from the smooth Hubble flow out to a redshift (z) of 1.7. The current data have revealed that the equation-of-state parameter of the dark energy, w = P/(ρc²), is consistent with a cosmological constant, w = −1 (Astier et al. 2006; Wood-Vasey et al. 2007; Riess et al. 2007; Freedman et al. 2009; Sullivan et al. 2011). Advancements in our understanding of dark energy will come from future experiments that seek to measure the time derivative of w. The success and credibility of these experiments will require percent-level accuracy in SN distances, something that has yet to be demonstrated. To achieve this will require both a deeper theoretical understanding of the progenitors and explosion mechanism as well as, on the observational side, a significant sample of well-calibrated low-z SNe Ia monitored over a wide wavelength range.

The Carnegie Supernova Project (CSP) is the umbrella designation of a five-year (2004–2009) National Science Foundation (NSF) funded observational campaign consisting of low- and high-z components. The overriding goal of the low-z component was to construct an atlas of ∼100 high signal-to-noise ratio (S/N) SN Ia light curves in a homogeneous and well-defined photometric system in the optical and near-infrared (IR). These data will serve as a fundamental reference to anchor future Hubble diagrams that will contain thousands of high-z SNe Ia observed by future all-sky surveys. The objective of the high-z portion of the CSP was to construct a rest-frame i-band Hubble diagram in order to minimize the effects of dust reddening. At the end of the latter observing program, 71 SNe Ia were observed at cosmological distances out to z ≈0.7. A subset of these objects was combined with our initial low-z sample to construct the Hubble diagram presented by Freedman et al. (2009).

By the end of the fifth and final year (2009 June), the low-z part of the CSP had obtained photometric follow-up observations of 129 SNe Ia. In the first of a series of papers, we presented optical (uBgVri) and near-IR (YJHK_s) light curves of 35 SNe Ia (Contreras et al. 2010), along with a detailed analysis (Folatelli et al. 2010). A highlight of the Folatelli et al. study is the discrepancy found between the value of the reddening parameter, R_V, as determined through the comparison of color excesses to the method of leaving R_V as a free parameter and minimizing the dispersion in the Hubble diagram. When excluding the two most highly reddened objects from the first sample, a comparison of colors or color excesses suggests agreement with the canonical Milky Way value, R_V ≈ 3.0. However, when one minimizes the dispersion in the Hubble diagram with R_V treated as a free parameter, lower values of R_V ≈ 1–2 are obtained. This discrepancy implies that there is an intrinsic color variation of SNe Ia that correlates with luminosity, but is independent of the light-curve decline-rate parameter Δm₁₅.¹⁴ To address this apparent inconsistency demands an expanded sample of well-observed multi-band SNe Ia.

In this paper, we present the second CSP release of SN Ia light curves. The sample consists of 50 objects observed between 2005 and 2009, with a subset of 45 objects having accompanying near-IR observations. The light curves presented here are a significant addition to a growing sample of low-z SNe Ia that have been observed in well-understood and stable photometric systems (Wood-Vasey et al. 2008; Hicken et al. 2009; Contreras et al. 2010; Ganeshalingam et al. 2010). A third and final paper containing optical and near-IR light curves of an additional ∼20–25 objects will complete the full data release of low-z SNe Ia observed by the CSP.

Additional aims of the CSP included the construction of substantial samples of Type II SNe (∼100) and Type Ib/c SNe (∼25). By the end of our observational program, these goals were successfully achieved, and a majority of the data are currently being analyzed. The CSP was not limited to only photometric follow-up observations; a notable commitment was also made to obtain contemporaneous optical spectroscopy. In a companion paper currently in preparation an analysis of spectrophotometry of ∼80 SNe Ia monitored by the CSP will be presented. Additionally, a second analysis paper combining the CSP's first 85 published low-z SNe Ia will be presented in the near future.

The organization of this paper is as follows. Section 2 briefly describes the observations, Section 3 discusses results obtained from a calibration program to measure the optical and near-IR transmission functions of the Henrietta Swope 1 m and Irénée du Pont 2.5 m telescopes equipped with CSP filters, Section 4 presents the final light curves, Section 5 compares two peculiar objects, and Section 6 contains our conclusions.

A complete overview of the CSP including the goals of the project, the facilities at Las Campanas Observatory (LCO) used to conduct photometric and spectroscopic data, and the observing procedures are given by Hamuy et al. (2006). A detailed and up-to-date description of the data-reduction procedures, the host-galaxy subtraction techniques, and computation of definitive photometry in the natural system of the LCO facilities is given in the first data release of SN Ia photometry (Contreras et al. 2010). It is not our intention to repeat the material contained in these previous publications, so in this section we provide only a brief summary of our methodology.

The CSP was a follow-up program of southern and equatorial SNe discovered by various sources ranging from dedicated backyard amateurs to professional searches. Five observational campaigns, each nine months in duration, were conducted between 2004 and 2009. The primary criteria used to select an object for monitoring were (1) that it was caught young, preferably before maximum light, or within a few weeks after explosion in the case of Type II plateau objects, and (2) that the peak magnitude was ∼18 or brighter. The workhorse of the CSP was the Henrietta Swope 1 m telescope equipped with both a direct imaging CCD camera that we refer to as SITe3 and a near-IR imager called RetroCam. Additional near-IR imaging and limited optical observations were obtained with the du Pont 2.5 m telescope equipped with the Wide Field IR Camera (WIRC) and a direct optical imager referred to as Tek 5. Typically a year or more after the last observation of any given SN, deep "template" images of its host galaxy were obtained with the du Pont telescope. These high S/N templates are used to subtract away the host-galaxy light at the position of the SN in each science image. This procedure is essential for obtaining precise photometry, particularly if the SN occurred in the inner region of the host galaxy or in a spiral arm.

Template subtraction has been performed on the science images of all objects presented here except SN 2007if (near-IR) and SN 2009dc (optical and near-IR). As these two events are super-Chandrasekhar candidates, we recognize the interests within the community to promptly make their photometry public. Given the position of both objects in their respective host galaxies, we suspect that the photometry suffers negligible background contamination. Nevertheless, in the near future templates of these objects will be acquired, and the subsequent template-subtracted photometry will be available on the CSP Web site.

Photometry of each SN is computed differentially with respect to a local sequence of stars. Absolute photometry of each local sequence is determined relative to photometric standard stars which were observed over a minimum of three photometric nights. The master list of optical photometric standard stars used to calibrate the local sequences is composed of stars in common with the Landolt (1992) and Smith et al. (2002) catalogs. In the near-IR, absolute photometry of each local sequence is computed relative to Persson et al. (1998) standards.

The CSP has elected to publish photometry on the natural system of the Swope and du Pont telescopes. The reasons for this are highlighted by Contreras et al. (2010), and boil down to providing the photometry in a form that simplifies the process of combining it with data sets obtained by other groups. Armed with spectral templates and well-defined system-response functions, CSP photometry can be transformed or "S-corrected" (Stritzinger et al. 2002) to any desired user-defined photometric system in a straightforward manner.

The SNe contained in this data release were observed from mid-2005 to early 2009. A mosaic consisting of a V-band image of each SN is presented in Figure 1. General properties related to each object are provided in Table 1. Specifically, this includes (1) the coordinates, (2) identification, morphology, and redshift of the host galaxy, (3) a reference to the IAUC or CBET reporting the discovery, and (4) the name of the discovery individual(s) or group(s).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Basic photometric and spectroscopic information on each object is compiled in Table 2, consisting of (1) an estimate of the decline-rate parameter, Δm₁₅ (see Section 4 for details), (2) the number of nights each SN was observed, (3) the number of nights that standard-star fields were observed under photometric conditions, and (4) the SN spectral subtype. Values of Δm₁₅ were computed using the SNooPy Python package (SNooPy; Burns et al. 2011), while the SuperNova IDentification code (Blondin & Tonry 2007) provided aid in determining spectral subtypes. Finally, the last column of Table 2 lists the epoch of the spectrum (relative to the time of B-band maximum) used to determine the spectral subtype.

Examination of the light-curve properties and optical spectra reveals that 36 of the objects in this data release are normal SNe Ia, 8 are of the low-luminosity SN 1986G (Phillips et al. 1987) or SN 1991bg (Filippenko et al. 1992a; Leibundgut et al. 1993) variety, 2 are of the slow-declining SN 1991T-like subtype (Filippenko et al. 1992b; Phillips et al. 1992), 2 are considered possible super-Chandrasekhar SNe (e.g., see Howell et al. 2006), and 2 objects, SN 2006bt and SN 2006ot, are uniquely peculiar (Foley et al. 2010).

Optical and near-IR photometry of each local sequence in the standard Landolt (BV; Landolt 1992), Smith (u'g'r'i'; Smith et al. 2002), and Persson (JHK_s; Persson et al. 1998) photometric systems are given in Table 3. The Y-band photometry is calibrated relative to the (Y − K_s) and (J − K_s) color relation given by Hamuy et al. (2006, see Appendix C, Equation (C2)). The quoted uncertainty accompanying the magnitudes of each star corresponds to the weighted average of the instrumental errors computed from several measurements obtained over the course of multiple photometric nights (see Columns 5 and 6 of Table 2).

In the past, photometry of local SNe Ia has often been obtained with poorly quantified system (telescope + instrument + filters) response functions. This problem makes it difficult to precisely combine low- and high-z SN Ia samples, and contributes to the systematic error budget of cosmological parameters. To maximize the potential of our low-z sample, we have devised a calibration program to measure with 1% accuracy the response of the CSP bandpasses on the Swope (optical + near-IR) and du Pont (near-IR) telescopes. In this section, we describe the photometric calibration program and present the scanned optical response functions. The near-IR scanned response functions will be presented in a forthcoming data paper. A more complete discussion of the technical aspects of this experiment, including a schematic of the setup, is presented by Rheault et al. (2010).

3.1. Experimental Setup

3.2. Calibration Procedure

3.3. Calibration Results

Two calibration runs were conducted at LCO during 2010. During the first of these runs (2010 January) the uBgVri bandpasses of the SITe3 CCD imager on the Swope telescope were measured twice per filter, point-to-point by wavelength, except for the i-band filter which was scanned only once. In a second run (2010 July) a second scan was obtained of the Swope i bandpass, and the near-IR bandpasses of both RetroCam on the Swope telescope and WIRC on the du Pont were scanned. The data from the second run are currently being analyzed and will be presented in a future publication. The repeatability of the uBgVr measurements from the first calibration run was ⩽1%; by this we mean that there is a ⩽1% point-to-point (by wavelength) uncertainty in the filter-response profiles. In other words, the ratios of the two scans of each individual filter fall within an envelope ±1% wide.

Plotted in Figure 2 are the uBgVri bandpasses measured from our calibration program multiplied by a CTIO atmospheric transmission function, including telluric absorption features. Also included are the tracings previously published by Contreras et al. (2010, see their Figure 8). In addition to measuring the bandpasses, the response of the Swope telescope was measured with and without the SITe3 Direct CCD camera. The scanning without the CCD camera was performed with a photodiode placed at the focal plane of the telescope. These measurements yielded the optical throughput of the telescope which, when combined with the total throughput measurements, allowed the instrument response function (CCD + dewar window) to be deduced.

Zoom In Zoom Out Reset image size

Under close inspection of the scanned bandpasses it was evident that in all cases the wings did not reach zero transmission. The cause of this effect was determined to be white light that leaked into the calibration setup. In general the offset produced by the white light was reasonably constant across the bandpasses; thus, to produce the final corrected curves the same constant value was subtracted from each bandpass trace.

As a direct result of our measurements, we managed to resolve two issues with the bandpasses published by Contreras et al. (2010). First, we now have an accurate measurement of the u filter response function which, as discussed by Contreras et al. (2010), was a source of considerable uncertainty. Second, on 2006 January 14 UT the V filter (called "LC-3014") was damaged and subsequently replaced with another V filter ("LC-3009"). After several nights of use, it was determined that the replacement filter had a significantly different color term compared to the original. This filter was therefore replaced with a third filter ("LC-9844") on 2006 January 25 UT, which was used for the remainder of the CSP campaigns. During the first photometric calibration run a portion of the broken LC-3014 filter and the temporary replacement, LC-3009, were also scanned using the monochromator while mounted on an optical bench. A comparison of these tracings indicates that the LC-9844 and LC-3014 V-band filters, though similar, are not identical (see Figure 2). Details regarding the color terms for these three V-band filters are provided in Appendix B.

Final optical and near-IR light curves of 50 SNe Ia in the natural system are displayed in Figure 3. The corresponding optical and near-IR photometry of each object is provided in Tables 4 and 5, respectively.¹⁵,¹⁶ The quoted uncertainties in these tables correspond to the sum in quadrature of the instrumental error and the nightly zero-point error. In general, the light curves provide excellent temporal coverage and exhibit photometric precision at the 0.01–0.03 mag level.

Zoom In Zoom Out Reset image size

View extended figure (47 panels)

The light curves of the "normal" and "SN 1991T-like" spectral subtypes (see Table 2) and SN 2006ot are overplotted in Figure 3 with light-curve fits (solid red lines) computed by SNooPy (Burns et al. 2011) using the "max model" (see Stritzinger et al. 2010, Equation (5)). From the SNooPy fits we obtain estimates of the peak magnitudes, the time of B-band maximum, T(B_max), the decline-rate parameter Δm₁₅ (Burns et al. 2011), and a covariance matrix of these parameters. Currently, SNooPy is unable to satisfactorily fit template light curves to peculiar SNe Ia like SN 2006bt (Foley et al. 2010), the fast-declining low-luminosity SN 1986G/SN 1991bg-like objects, or the possible super-Chandrasekhar mass SNe. In these cases, we interpolated the observed photometry with the use of spline functions, from which the peak magnitudes may be estimated, and the values of T(B_max) and Δm₁₅ were derived directly from the B-band spline fit. Overplotted on the observed photometry in Figure 3 are the resulting spline fits (red dashed lines). Due to insufficient photometric coverage, we were unable to accurately determine estimates of Δm₁₅ for the SN 1991bg-like objects SNe 2006bd and 2006hb, nor for the possible super-Chandrasekhar SN 2007if.

The Δm₁₅ values derived from either the SNooPy template light-curve fits or the spline fits are listed in Table 2, and are plotted in a histogram in Figure 4. We note that because a single value of Δm₁₅ is used in SNooPy to parameterize all the filters when performing the "max model" fit to an SN Ia, the values reported here can differ slightly from the value one would obtain by measuring it directly from the decline of just the B-band light curve. Also included in Figure 4 are the Δm₁₅ values for the SNe Ia presented in the first CSP data release of SNe Ia (Contreras et al. 2010).

Zoom In Zoom Out Reset image size

We find that for the objects presented in this paper, 33 have pre-maximum optical coverage, and of these, 15 have observations beginning at least five days before T(B_max). Of the 45 SN Ia with near-IR observations, 27 have observations that commence prior to T(B_max), with a subset of 13 beginning at least five days earlier than maximum B-band light.

In a recent paper, Foley et al. (2010) called attention to one of the objects in our sample, SN 2006bt, which showed photometric and spectroscopic characteristics unlike those of any other known SN Ia. While possessing a broad, slow-declining (Δm₁₅ = 1.1) B light curve similar to a typical luminous SN Ia, SN 2006bt also displayed certain photometric and spectroscopic properties that are more similar to those of fast-declining, low-luminosity events. In particular, the i-band light curve showed only a weak secondary maximum, and the maximum and pre-maximum optical spectra displayed strong Si ii λ5972 absorption relative to Si ii λ6355 and a depression at ∼4100 Å due to Ti ii, all of which are properties characteristic of low-luminosity SN 1991bg-like objects. As mentioned earlier, our attempts to fit the light curves of SN 2006bt with SNooPy met with failure because of the combination of slow-declining light curves with a weak secondary maximum in the i band. The i light curve of SN 2006bt is also peculiar in reaching a primary maximum at essentially the same epoch as B maximum, whereas for normal slow-declining SNe Ia, i maximum occurs several days before B maximum. Nevertheless, Foley et al. (2010) showed that the luminosity of SN 2006bt was quite typical of normal, slow-declining SNe Ia.

Interestingly, a second object in our sample, SN 2006ot, appears to be closely related to SN 2006bt. This is illustrated in Figure 5 where the uBgVriYJH light curves of both SNe are overplotted. Also included in this figure are SNooPy fits to the CSP photometry of the normal, slow-declining (Δm₁₅ = 1.0) SN 2006af (Contreras et al. 2010). Once again, in the observations of SN 2006ot we find the combination of broad, slow-declining light curves with a weak secondary maximum in i. The morphology of the Y and J light curves of SN 2006ot is also peculiarly flat and consistent with the more limited CSP coverage obtained for SN 2006bt. Figure 6 shows a comparison of the spectra of SNe 2006bt and 2006ot near maximum light and approximately three to four weeks after maximum. Again, spectra of the normal SN 2006ax are included for comparison. The peculiar nature of the maximum-light spectrum of SN 2006bt compared to SN 2006ax is apparent in this figure, in particular the strength of the Si ii λ5972 absorption compared to λ6355, and the broad Ti ii absorption at ∼4100 Å. The maximum-light spectrum of SN 2006ot is different yet again, appearing in some ways intermediate between the spectra of SNe 2006bt and 2006ax (e.g., in the ratio of the equivalent widths of the Si ii λ5972 and λ6355 features). Most striking is the strength, width, and blueshift of the Si ii λ6355 line, the minimum of which yields an expansion velocity of ∼14500 km s⁻¹. At three to four weeks past maximum, the spectra of SNe 2006bt and 2006ot closely resemble each other, while the spectrum of the normal SN 2006ax differs significantly at wavelengths shortward of 6000 Å.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

According to the maps of Schlegel et al. (1998), interstellar reddening of SNe 2006bt and 2006ot due to dust in the Milky Way was small (A_V = 0.17 mag and A_V = 0.05 mag, respectively). Foley et al. (2010) did not find evidence for Na i D absorption in the spectrum of SN 2006ot at the redshift of its host galaxy, nor is Na i D detectable in our maximum-light spectrum of SN 2006ot. Hence, it is unlikely that the light of either SN suffered large dust extinction. The difference in host galaxy recession velocities corrected to the cosmic microwave background (CMB) reference frame (v_CMB = 9737 km s⁻¹ for SN 2006bt and v_CMB = 15695 km s⁻¹ for SN 2006ot) implies a brightness difference of ∼1.0 mag if both objects were unreddened and shared the same peak luminosity. This is only 0.1 mag different from the shift of ∼0.9 mag required to register the light curves of both SNe (see Figure 5). Thus, the peak luminosities were likely quite similar. Interestingly, the host of SN 2006bt was an S0/a galaxy and NED lists the morphology of the host of SN 2006ot as Sa, implying a relatively old progenitor (Foley et al. 2010).

It is not at all obvious how SNe 2006bt and 2006ot fit into the overall scheme of SNe Ia, but it is clear that they can easily be confused with normal events unless high-quality photometry and/or spectroscopy is obtained. This is particularly true at high redshifts where rest system i or near-IR light curves (where the photometric differences are greatest) are almost never obtained. Foley et al. (2010) concluded that contamination of SNe Ia samples by similar objects will both increase the scatter of the Hubble diagram and systematically bias measurements of cosmological parameters. Fortunately, they do not appear to be very common in the local universe as only two such events are among the 87 SNe Ia with photometry published to date by the CSP.

We have presented the second data release of SN Ia photometry obtained by the CSP. Combined with the first sample (Contreras et al. 2010) and photometry from Schweizer et al. (2008) and Stritzinger et al. (2010), the CSP data set currently consists of 87 objects. The high-quality and dense temporal coverage, extending over nine filters from the optical to the near-IR, provides a sample that is well suited for tackling the issues related to accurately estimating host-galaxy reddening (e.g., Folatelli et al. 2010).

With the addition of the 50 low-z objects presented in this paper, we are in a position to extend upon our first analysis and, among other things, to provide a detailed investigation of the fastest declining SNe Ia. The CSP sample now consists of 16 rapid decliners (Δm₁₅ > 1.5). This sample is nearly evenly split between the morphologies of the near-IR maximum of those objects that peak in the near-IR before T(B_max), and those that peak after. Krisciunas et al. (2009) suggest that objects whose near-IR light curves peak after T(B_max) appear to be 0.5–0.8 mag less luminous in the optical and near-IR than those events whose near-IR light curves peak before T(B_max). It will be particularly interesting to re-examine this issue using the CSP sample, as well as to compare other properties of the light curves and spectra of this SN Ia subtype.

When completely reduced and analyzed, the CSP optical and near-IR light curves of SNe Ia already in hand will constitute a fundamental reference for cosmological studies. However, at present, the full utility of the CSP sample as a reference for observations at high redshift is compromised by the fact that the median redshift of the sample is ∼0.02. At this redshift, the root-mean-square (rms) error in distance modulus due to peculiar velocities is ±0.13 mag or 6% in distance and dominates the observed Hubble diagram dispersion (Neill et al. 2007; Folatelli et al. 2010). This problem can be minimized by extending observations to SNe further out in the smooth Hubble flow. We have therefore proposed and been granted NSF funding to carry out a second stage of the CSP with the goal of producing optical and near-IR light curves of 100–200 SNe Ia in the redshift range 0.03 < z < 0.08. With such a sample we expect the rms error due to peculiar motions to drop to ± 3% or less in distance.

The CSP extends special thanks to the mountain staff of the Las Campanas Observatory for their assistance throughout the duration of our observational program, and to Jim Hughes for his superb support of our network of computers. This material is based upon work supported by the NSF under grants AST-0306969, AST-0908886, AST-0607438, and AST-1008343. The Dark Cosmology Centre is funded by the Danish National Research Foundation. M.H. acknowledges support by CONICYT through grants FONDECYT Regular 1060808, Centro de Astrofisica FONDAP 15010003, Centro BASAL CATA (PFB–06), and by the Millennium Center for Supernova Science (P06–045-F). A.V.F. is grateful for the financial support of the NSF and the TABASGO Foundation. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

With the scanned system-response functions presented in Section 3.3, we proceed to compute photometric zero points following the prescription described by Contreras et al. (2010, see their Appendix C).

The photometric zero point for bandpass X, with a response function denoted by S_X(λ), can be expressed as

$$\begin{equation} {\rm ZP}_{X}=2.5\log \left[\int _0^{\infty } S_{X}(\lambda)\, F_{o}(\lambda) \frac{\lambda }{hc}\, d\lambda \right]+m_{0}. \end{equation} \tag{ A1 }$$

Here F_o(λ) is the spectral energy distribution (SED) of a fiducial source and m_o is the observed magnitude of this source through the bandpass X.

To compute the B- and V-band zero points we adopt the SED of Vega (α Lyr) as our fiducial source. In doing so, we use the Bohlin & Gilliland (2004) SED of Vega (version 5) obtained from CALSPEC,¹⁷ and the Fukugita et al. (1996) standard magnitudes B_std = 0.03 and V_std = 0.03. As we aim to compute zero points in the natural system of the Swope telescope, B_std and V_std are transformed to the Swope natural system using the following equations:

$$\begin{equation} B_{\rm nat} = B_{\rm std} - {\rm CT}_{B} \times (B_{\rm std} - V_{\rm std}), \end{equation} \tag{ A2 }$$

$$\begin{equation} V_{\rm nat} = V_{\rm std} - {\rm CT}_{V} \times (V_{\rm std} - i_{\rm std}), \end{equation} \tag{ A3 }$$

where CT_X are the color terms. Color terms for the CSP bandpasses were given by Hamuy et al. (2006) and Contreras et al. (2010); the former values correspond to the average for the first of five yearly observing campaigns, while the slightly different numbers given in the latter reference are based on the first four observing campaigns. In Table 6, the final five-year average color terms are presented. The second column corresponds to the unweighted average after applying a 5σ clipping algorithm. Listed in the third column as a consistency check are the median values calculated without the 5σ clipping, with uncertainties corresponding to 1.49 multiplied by the median absolute deviation. As is seen, the differences between these two methods of computation are minimal. We therefore adopt the unweighted averages determined with the 5σ clipping.

In Equation (A3), the value of "i_std" was computed synthetically using the SED of Vega and the "USNO 40" i' bandpass. This yields i'_std = 0.382 mag. Executing the transformations yields Vega magnitudes in the Swope natural system of B_nat = 0.03 and V_nat = 0.009.

Armed with B_nat and V_nat, and the B and V (LC-9844) scanned bandpasses, zero points were computed directly using Equation (A1). In addition, zero points for the temporary (LC-3009) and replacement (LC-3014) V-band filters discussed in Section 3.3 were also computed, and the results are listed in Table 7.

We note that due to the limited use of the temporary LC-3009 V filter, we are unable to accurately determine the color term of this filter from observations of standard-star fields. To remedy this problem, the color term was estimated synthetically using an atlas of Landolt (1992) spectrophotometic standards (Stritzinger et al. 2005), and the corresponding known B_std and V_std magnitudes (see Appendix B for further details).

To compute the zero points of the scanned ugri bandpasses, BD+17°4708 was chosen as the fiducial source. Similar to the steps described above, the standard u_std = 10.56, g_std = 9.64, r_std = 9.35, and i_std = 9.25 mag of BD+17°4708 (Fukugita et al. 1996) were transformed to the Swope natural system using the color term in Table 6 and the following transformation equations:

$$\begin{equation} u_{\rm nat} = u_{\rm std} - {\rm CT}_{u} \times (u_{\rm std} - g_{\rm std}), \end{equation} \tag{ A4 }$$

$$\begin{equation} g_{\rm nat} = g_{\rm std} - {\rm CT}_{g} \times (g_{\rm std} - r_{\rm std}), \end{equation} \tag{ A5 }$$

$$\begin{equation} r_{\rm nat} = r_{\rm std} - {\rm CT}_{r} \times (r_{\rm std} - i_{\rm std}), \end{equation} \tag{ A6 }$$

$$\begin{equation} i_{\rm nat} = i_{\rm std} - {\rm CT}_{i} \times (r_{\rm std} - i_{\rm std}). \end{equation} \tag{ A7 }$$

These transformed magnitudes were then compared to synthetic magnitudes computed with our scanned bandpasses and the Bohlin & Gilliland (2004) SED of BD+17°4708 to compute the zero points via Equation (A1); the results are listed in Table 7.

To test the accuracy of the scanned B and V bandpasses, "synthetic" color terms were computed by comparing synthetic photometry derived using the Stritzinger et al. (2005) set of Landolt (1992) spectrophotometric standards to the published Landolt values. If the scanned bandpasses are correct, the synthetic color terms derived from this experiment should match the observed values reported in Table 6.

Plotted in the left panel of Figure 7 are the differences between the synthetic B-band magnitudes computed with the scanned bandpass and the published Landolt values versus synthetic B − V color. Applying a color cut to exclude stars bluer than B − V = 0 or redder than B − V = 0.9 mag yields the best-fit color term of CT_B = +0.065. This is reasonably consistent with the value of CT_B = +0.069 given in Table 6.

Zoom In Zoom Out Reset image size

Presented in the right panel of Figure 7 is a similar comparison for the three V bandpasses used by the CSP. The blue points correspond to values derived with the scan of the broken LC-3014 filter, while the red points are from the scan of the temporary LC-3009 filter, and the green points are the values obtained with the scan of the replacement filter LC-9844. Adopting the previously mentioned color cut yields best-fit color terms of CT_V = −0.063, CT_V = −0.044, and CT_V = −0.065 for the LC-3014, LC-3009, and LC-9844 bandpasses, respectively. These values are in excellent agreement with the color terms derived from the broadband photometry of the Landolt standard-star observations (see Table 6).

The results of this experiment confirm the accuracy of the scanned B and V bandpasses. Unfortunately, due to a lack of spectrophotometric measurements of the Smith et al. (2002) standard stars, it is not possible to accurately apply this test to the scanned ugri bandpasses.

As the V-band LC-3009 color term has a non-negligible difference with respect to the two other V-band filters used, we re-examined the small amount of photometry obtained with LC-3009. During the time the LC-3009 filter was used, 12 different SNe Ia were observed during one to three epochs. Nine of these objects are included in Contreras et al. (2010). For completeness we returned to these nights in question and recomputed the photometry in the natural system of the Swope telescope. We note that for these objects, eliminating the observations taken with the LC-3009 filter has little effect on the quality or coverage of the V-band light curves, except for the case of SN 2006D whose first three pre-maximum epochs were obtained with LC-3009. Updated V-band photometry for these three epochs is available on the CSP Web site.