Young Supernova Experiment

• Young Supernova Experiment is a large-scale, multi-color time-domain survey that captures early and rare supernova events using coordinated Pan-STARRS and ZTF observations.
• It employs a fast-rise detection strategy with deep griz imaging to study shock breakout, flash ionization, and early explosion physics, critical for cosmology.
• The survey delivers high-precision, well-calibrated transient data and training sets for advanced machine learning classification, paving the way for next-generation time-domain surveys.

The Young Supernova Experiment (YSE) is a large-scale, untargeted, multi-color time-domain optical survey designed to discover and characterize young, red, and rare supernovae (SNe) as well as other extragalactic transients. Utilizing the Pan-STARRS telescopes and coordinated with Zwicky Transient Facility (ZTF), YSE delivers high-cadence, deep, four-band ($griz$) imaging over a wide area, capturing the earliest phases of stellar explosions and generating well-calibrated low-redshift samples fundamental for cosmology, transient science, and the preparatory work required for the next generation of time-domain surveys such as the Vera C. Rubin Observatory and the Nancy Grace Roman Space Telescope (Jones et al., 2020, Aleo et al., 2022).

1. Survey Design and Observational Strategy

YSE operates primarily with the Pan-STARRS1 (PS1) 1.8 m telescope and, as commissioning completes, the Pan-STARRS2 (PS2) twin. Each is equipped with a 1.4 Gpixel camera offering a 7 deg² field of view. The baseline survey cadence is every three days in two $griz$ filters, with alternating $g+r$/$g+i$ in dark time and $r+i$/$r+z$ in bright time. Typical single-epoch 5$\sigma$ depths are $g, r, i \approx 21.5$ mag and $z \approx 20.5$ mag (empirically up to $g\approx22.2$ mag), yielding sensitivity to transients out to $z \approx 0.5$ for SLSNe, and $z \approx 0.3$ for SNe Ia (Aleo et al., 2022).

Coordination with ZTF (1.2 m Palomar Schmidt; 47 deg² camera, $g$, $r$ filters, $\sim2$ d cadence) enables sub-night and sub-two-day sampling. The effective combined YSE+ZTF cadence is median $\sim1.98$ d, essential for discovering fast transients and constraining explosion times with pre-discovery non-detections.

A fast-rise strategy is employed by flagging candidates with $>$1 mag brightening between YSE and ZTF epochs for rapid follow-up. Daily monitoring of cluster fields (e.g., Virgo) enables discovery of faint stellar outbursts and pre-explosion activity down to $M_i\sim-10$ mag (Jones et al., 2020).

2. Science Goals and Survey Scope

YSE is designed to address several central objectives:

1. Early-Time Supernova Physics: Detect SNe within hours to days of explosion, facilitating paper of shock breakout, flash ionization, companion interaction, and circumstellar mass loss.
2. Rare and Red Transients: Expand the observational parameter space for faint ($M\gtrsim -16$), fast, and extremely red transients (including Ca-rich events, luminous red novae, fast blue optical transients, gravitationally lensed SNe, and SLSNe) by leveraging deep $i$ and $z$ imaging.
3. Cosmology: Provide a homogeneously calibrated anchor sample of several hundred SNe Ia at low redshift ($z\sim0.1$), with relative calibration better than 3 mmag, crucial for reducing systematic uncertainties in $H_0$ and $w$ in future cosmological analyses (Jones et al., 2020).
4. Active Galactic Nuclei and Tidal Disruption Events: Monitor tens of thousands of galaxies for nuclear variability, enabling discovery of TDEs and changing-look quasars.
5. Rubin/Roman Pathfinder: Build a comprehensive training set of multi-band transient light curves with selection functions and cadence closely matched to the next-generation time-domain surveys (Jones et al., 2020, Aleo et al., 2022).

3. Data Products, Photometric Classification, and Early Results

YSE Data Release 1 (YSE DR1) provides forced PSF photometry for 1,975 transients with accompanying host associations, redshifts (spectroscopic and photometric), and classifications using both spectroscopic and advanced photometric methods. The photometric classifier ParSNIP, a physics-enabled variational autoencoder with latent parameterization and a random-forest discriminator, achieves 82% overall accuracy on three-class (Ia, II, Ib/Ic) validation (94% completeness, 89% purity for SNe Ia). SuperRAENN, a recurrent autoencoder with random forest trained on interpolated light curves, achieves 75% tertiary accuracy and 90% Ia-CC binary separation (Aleo et al., 2022).

Data products also include light-curve parameters, host galaxy cross-matches, and simulated light-curve injections using the SNANA framework, supporting classification algorithm development and robust derivation of volumetric SN rates.

Early YSE discoveries include prompt-phase SNe Ia (e.g., 2020pf, 2020fci), flash-ionized SNe II (e.g., 2020pni, 2020tlf), SNe Iax, super-Chandrasekhar SNe Ia, rare SNe IIb, and TDEs. The median discovery is $-6.4$ d before maximum light (Jones et al., 2020, Aleo et al., 2022).

4. Supernova Yields, Demographics, and Rates

Survey simulations predict $\sim$5000 SN detections per year (at full 1500 deg² coverage), subdivided as $\sim$3920 SNe Ia, $\sim$277 SNe Ib/c, $\sim$705 SNe II, and $\sim$67 SNe IIn for S/N$>$5 in at least three epochs (Jones et al., 2020).

Yearly yields from DR1 for magnitude-limited ($r<18.5$ mag) and volume-limited ($D<250$ Mpc) samples, with spectroscopic classification completeness up to 97%, are as follows:

Sample Type	SNe Ia	SNe II	SNe Ib/Ic	SLSNe
Magnitude-limited	0.682	0.239	0.074	0.006
Volume-limited	0.438	0.438	0.123	—

$R(c) = N_c / N_{\text{tot}}$; uncertainties are calculated via multinomial confidence intervals (Aleo et al., 2022).

DR1’s photometric sample (ParSNIP labels) is 71% SNe Ia, 23% SNe II, 6% SNe Ib/Ic. The sample extends to $z\lesssim0.5$ (median $z\approx0.14$), with spectroscopic reach to $z\sim0.3$. Relative rates align with other untargeted transient surveys (ZTF BTS, ASAS-SN, LOSS) within uncertainties.

5. Diagnostic Experiments: Young SNe as Physics Laboratories

YSE's early detections of stripped-envelope SNe (IIb/Ib/Ic) are critical for probing nonthermal acceleration physics. Radio synchrotron emission in $\sim1$ yr-old SNe exhibits an electron energy index $p_\text{obs} \simeq 3.0\pm0.2$, significantly steeper than standard DSA predictions ($p_\mathrm{DSA}=2.0$ for $r=4$). This steepening diagnoses the pre-acceleration “injection” regime: radio emission arises from electrons below $E_\mathrm{break}\sim100$ MeV ($\gamma_{\rm br}\sim200$) where acceleration is inefficient (Maeda, 2012).

The Young Supernova Experiment (in the sense of probing acceleration physics using early multi-wavelength data) employs:

• mm/sub-mm (ALMA): Multi-epoch (t=20, 50, 100, 200 days) monitoring across $\sim$100 GHz to track spectral breaks and infer $E_\mathrm{break}$ from SED features.
• cm (Radio): Constrains SSA turnover for magnetic field and shock radius.
• X-ray (Chandra): Observes IC and synchrotron emission to calibrate the transition from inefficient to efficient electron acceleration. Temporal and spectroscopic coverage (t$\sim$300, 500 days, 0.3–8 keV) probes slope hardening and slow decay ($F_X \propto t^{-0.3}$).

Key fitting parameters ($E_{\rm break}$, $p_{<}$, $p_{>}$, $\eta_e$, $A_*$, $\epsilon_B$) are determined through multi-band $\chi^2$ minimization, directly constraining DSA injection and acceleration efficiency for cosmic ray electron populations in young shocks (Maeda, 2012).

6. Legacy and Role in Next-Generation Surveys

YSE serves as a Rubicon for time-domain survey methodology, providing a contiguous, multi-band, well-calibrated data set for training alert brokers and photometric classifiers in filters closely matched to Rubin ($grizy$) and Roman (NIR). The magnitude-limited, low-$z$ SN Ia anchor population is designed to replace legacy calibration samples, reducing systematic cross-survey uncertainties to $\sim$2–3 mmag.

Synergy with ZTF (joint fields, $\sim$1–2 d cadence) and preparations for data handoff and cadence supplementation with Rubin positions YSE as both scientific and logistical pathfinder. Deep stacked imaging (to $griz\approx23.6$–$24.2$ mag after several years) will support pre-explosion progenitor searches and facilitate studies of dust-enshrouded or ultra-faint transients (Jones et al., 2020).

YSE DR1 has already contributed a substantial increase in low-$z$ SNe Ia samples, improved transient anomaly detection pipelines, and validated machine-learning classifiers (ParSNIP, SuperRAENN) foundational for the LSST era (Aleo et al., 2022).

7. Broader Scientific Applications

YSE DR1 supports:

• Cosmology: Robust anchor samples of SNe Ia for $H_0$ and $w$ measurements.
• Transient Astrophysics: Early-phase/fast SNe, classification of rare SNe (e.g. Ia–CSM, Iax, Ca-strong, LRNe, LBVs), and TDEs.
• Survey Methodology: Testbeds for cadence optimization, alert-broker architecture, and photometric pipeline scalability in preparation for future extremely large archives (Aleo et al., 2022).
• Core-Collapse SN Demographics: Direct measurement of relative SN rates without the need for targeted host-galaxy catalogs.

YSE’s design, execution, and data releases exemplify the integration of optimized high-cadence survey strategy, advanced machine learning classification of time-domain events, and the broader community utility of large, homogeneous transient datasets.