TITAN SN~Ia DR1 Dataset

• The TITAN SN~Ia dataset is a systematically calibrated collection of over 10,000 confirmed Type Ia supernovae, including about 3,000 cosmology-grade light curves targeting low-redshift analysis.
• It employs advanced intra-chip and inter-chip calibration techniques that reduce RMS uncertainties to 3–5 mmag, ensuring robust and precise photometric measurements.
• Accessible via the ATLAST package, the dataset underpins cross-survey validation and serves as a foundational low-z anchor for future dark energy and Hubble constant investigations.

The TITAN SN~Ia dataset, formally the "Type Ia supernova Trove from ATLAS in the Nearby universe" (TITAN), represents the first public release (DR1) of a uniquely large, systematically calibrated collection of low-redshift Type Ia supernova (SN Ia) light curves from the ATLAS (Asteroid Terrestrial Last Alert System) time-domain sky survey. With over 10,000 spectroscopically confirmed SNe Ia and approximately 3,000 cosmology-grade light curves prepared for precise cosmological analyses, TITAN DR1 aims to provide a reference dataset for low-$z$ SN cosmology featuring robust calibration, extensive validation, and thorough propagation of all relevant systematic uncertainties (Marlin et al., 26 Dec 2025).

1. Data Set Scope, Instrumentation, and Coverage

TITAN DR1 comprises:

• More than 10,000 spectroscopically confirmed SNe~Ia identified by ATLAS.
• "Cosmology-grade" light curves (well-sampled, host-galaxy redshifts measured) totaling approximately 3,000 SNe~Ia.

The dataset is volume-limited to $z\lesssim 0.1$, set by the ATLAS photometric detection threshold ($m\approx 20$), with near-all-sky coverage:

• Northern telescopes (Haleakala, Mauna Loa): $\textrm{Dec} \geq -50^\circ$
• Southern telescopes (South Africa, Chile): $\textrm{Dec} \lesssim +40^\circ$
• Nightly full-sky monitoring cadence

ATLAS utilizes four 0.5-m f/2 Wright–Geminian telescopes, each equipped with a $10,\!560\times10,\!560$ px STA-1600 CCD, read out at $1\times1$ binning. The typical point spread function has FWHM $3.7''$–$5.6''$. Photometric observations are performed in two broad bands:

• ATLAS-cyan ($c$): $$4200\ \text{Å} \lesssim \lambda \lesssim 6500\ \text{Å}$$
• ATLAS-orange ($o$): $$5600\ \text{Å} \lesssim \lambda \lesssim 8200\ \text{Å}$$

2. Photometric Calibration and Cross-Calibration Workflow

The central calibration objective is to place ATLAS forced photometry on an absolute AB system with mmag-level systematic control. The workflow, designed to ensure transferability and uniformity, involves:

2.1 Reference Catalogs and Stellar Samples

• The DES Y6 tertiary star catalog (17 million stars over 5,000 deg$^2$, 1.8 mmag spatial uniformity, anchored to CALSPEC C26202 at 1% absolute flux) serves as the prime calibration reference.
• Baseline ATLAS calibration uses Refcat2 (incorporating PS1/Gaia/APASS/Skymapper).
• Within the DES footprint, three stellar samples are constructed:
  1. "Color-blind" (uniform in DES $g-i$ color, matched to Refcat2)
  2. "Blue" (DES $g-i \leq 0.2$)
  3. "Non-Refcat2" (stars in DES Y6 absent in Refcat2; selected to mimic SN~Ia host galaxies)

• Stars only observed in Gaia within Refcat2 are excluded to remove Gaia-only zeropoint biases.

2.2 Intra-Chip Zeropoint Offsets ($\Delta ZP_\mathrm{pixel}$)

For each chip, the pixel-level residual is evaluated as:

$$\Delta ZP_\mathrm{pixel}(x, y) = m_\mathrm{obs}(x, y) - \mathrm{Med}\left[ m_\mathrm{obs}(x, y') \right]\,,$$

where $m_\mathrm{obs}(x, y$) is the observed magnitude, and the median is computed over many dithered observations.

• The 10560$\times$10560-px focal plane is binned into $50\times50$ pixel cells, convolved with a Gaussian kernel of $\sigma\approx540$ px to produce per chip-filter correction maps.
• RMS across 10-pixel bins is reduced from $\sim8$ mmag (pre-correction) to $\sim4$ mmag (post-correction).

2.3 Inter-Chip Zeropoint Offsets ($\Delta ZP_\mathrm{chip}$)

For each chip and filter $f$, the color-transformed residual:

$$\Delta_{i}(f;x,y) = \left( m^{\mathrm{ATLAS}}_{f,i} - m^{\mathrm{DES}}_{y2,i}\right) - f^{\mathrm{synth}}_{f\to y2}\left(m^{\mathrm{DES}}_{x1,i} - m^{\mathrm{DES}}_{x2,i}\right)$$

where $f^{\mathrm{synth}}$ is a 3rd-order polynomial mapping based on synthetic photometry of NGSL/CALSPEC standards.

• The chip- and filter-averaged zeropoint offset $\Delta ZP_\mathrm{chip}(f)$ is determined by maximum likelihood estimation over all calibration stars.
• RMS on these corrections reduces from $\sim17$ mmag pre-correction to $\sim3$ mmag post-correction.

2.4 Transmission-Function Color Dependence

Calibration residuals as a function of DES $(g-i)$ for each chip/filter show a slope $A_f$, attributed to deviations in assumed filter throughputs. Correction is implemented by shifting the filter central wavelength by $\Delta \lambda_\mathrm{filt}$, typically

• Cyan (chips 0–8): $\Delta \lambda_c \approx +28...+87\ \text{\AA}$
• Orange: $\Delta \lambda_o \approx -21...+27\ \text{\AA}$

After this chromatic correction, residual color-dependent systematics are suppressed to $\lesssim$5 mmag across SN~Ia colors.

3. Validation Regimes and Systematic Uncertainties

Comprehensive validation of the calibration chain includes the following elements:

3.1 Tertiary Star Validation

Full calibration (intra-chip, inter-chip, wavelength shift) yields:

• Median zeropoint offset for the "non-Refcat2" sample $\sim0$ mmag
• Scatter (σ) reduced from $\approx0.028$ mag to $\approx0.019$ mag

3.2 CALSPEC and DA White Dwarfs

Synthetic magnitudes computed from HST CALSPEC and DAWD standards are compared to corrected ATLAS fluxes. For chip 6 cyan:

• Pre-correction: $$\langle m^{\mathrm{synth}} - m^{\mathrm{obs}} \rangle = +0.015$$ mag; slope $=-0.0399$ mag/(g–i)
• Post-correction: $\langle ...\rangle = -0.0026$ mag; slope $=-0.0004$ mag/(g–i)

3.3 SN~Ia Cross-Matched Distance Moduli

Using SALT3-DESY5 light-curve fits, standardized $\mu$ distances are compared for 63 DEBASS, 35 YSE, and 474 ZTF DR2 SNe:

• DEBASS–TITAN: $\Delta\mu = -0.005 \pm 0.012$ mag
• YSE–TITAN: $\Delta\mu = +0.039 \pm 0.025$ mag
• ZTF–TITAN: $\Delta\mu = -0.038 \pm 0.007$ mag (consistent with ZTF DR2 offset per Newman et al. 2025)

3.4 Systematic Uncertainty Budget

Systematic errors per filter, after all corrections:

Source	Systematic (mmag)
Intra-chip	$\sim$3
Inter-chip	$\sim$3
Chromatic $\lambda$-shift	$\sim$5
Absolute scale (CALSPEC)	$\sim$6
Combined	$5$–$10$

A conservative $10$ mmag systematic error floor is included for SNANA light-curve analyses.

4. Data Structures, Tools, and Accessibility

ATLAST, a python software package, enables users to apply all calibration corrections and to read/write the forced-photometry light-curve data. Each SN data product includes:

• MJD (observation date)
• Filter (cyan or orange)
• Calibrated AB magnitude (post-corrections)
• Magnitude error (statistical + systematics)
• Optional: chip ID, $(x,y)$ pixel coordinates

Access options:

• TITAN documentation: https://titan-snia.github.io
• ATLAST software and calibration maps: https://github.com/SterlingYM/ATLAST
• ATLAS forced photometry server: https://fallingstar-data.com/forcedphot/

Users are advised to apply the pixel, chip, and $\lambda$-shift corrections in sequence using ATLAST, and include the $10$ mmag SNANA error floor. Citation: Murakami, Y. S., Marlin, E. G., et al. 2026, in preparation.

5. Role in SN~Ia Cosmology and Compatibility

TITAN’s RMS-controlled calibration ($\lesssim3$–$5$ mmag), extensive sample size, and full documentation position it to strengthen cosmological analyses relying on low-$z$ SNe~Ia anchors. Key metrics:

• Distance modulus calibration systematic $\lesssim0.01$ mag per filter
• After light-curve fitting, intrinsic scatter $\lesssim0.12$ mag
• Hubble diagram: $\sim$3,000 low-$z$ SNe anchor cosmological fits

TITAN DR1 exhibits cross-survey agreements at $\lesssim0.04$ mag with DEBASS and YSE, and replicates the known ZTF DR2 offset. When augmented by higher-redshift samples (e.g., DESY5, Roman, LSST), TITAN can contribute to competitive constraints on cosmological parameters ($w$, $w_a$, $H_0$) with reduced inter-survey correlation systematics.

A plausible implication is that the controlled, independently validated low-$z$ reference sample will facilitate joint cosmological analyses across present and forthcoming SN datasets (Marlin et al., 26 Dec 2025).

6. Summary and Prospects

TITAN DR1 delivers a public, thoroughly validated, systematically controlled set of $\sim$3,000 cosmology-grade, low-redshift SN~Ia light curves. It combines wide sky and redshift coverage, mmag-level calibration stability, and fully documented methodology and access tools. This resource enables robust calibration transfer, facilitates cross-survey standardization, and establishes a foundational anchor for forthcoming dark energy constraints and Hubble constant measurements (Marlin et al., 26 Dec 2025).