The First Hours of a Supernova Are Its Most Revealing—and Rubin Alone Will Miss Them

A supernova announces itself with a flash that can last only hours. Shock breakout, flash spectroscopy, precursor outbursts—these are the moments when a dying star's final structure and explosion physics are most exposed. But they are also the moments easiest to miss. The Vera C. Rubin Observatory's Legacy Survey of Space and Time, or LSST, will scan the southern sky every ~3 days. That is fast by past standards, but not fast enough for signals that rise and fade in a day.

A three-day blink in an hours-long story

LSST's Wide Fast Deep survey will observe ~18,000 square degrees over 10 years in six filters, with a cadence of about three days. Its 8-meter telescope and 9.6-square-degree camera will discover enormous numbers of supernovae. Yet the paper argues the survey's cadence, not its depth, is the limiting factor for nearby supernovae. The youngest phases—hours after first light—require follow-up surveys to 'shadow' LSST: re-observe nearby galaxy fields within hours to a day.

The arXiv preprint is a review and planning study, not a new observational result. The authors compiled known early supernova signatures by timescale and brightness, used simple depth and cadence metrics, and used Transient Name Server statistics from 2020–2024 with declination <20 degrees to approximate the LSST footprint. They then estimated how many nearby supernovae LSST and a DECam-like shadowing survey could catch. One example: DECam on the 4-meter Blanco telescope, with a 3-square-degree field, reaching ~22.5 mag. LSST transient detection was conservatively assumed at ~24 mag.

Precursors: the star's final months and years become visible

One target is precursor emission—brightening days to years before explosion, likely from late-stage mass loss in massive stars. LSST's depth could detect or set limits on precursor emission from hundreds of core-collapse supernovae, including at least 50–100 Type IIn annually and at least 5 Type Ibn. For normal Type II supernovae, if SN 2020tlf is a guide, LSST could sample ~150 per year and constrain the fraction with such outbursts to less than 10% within a few years.

The reach depends on luminosity, but some precursors could be seen out to ~600 Mpc.

Catching the flash and the early excess

After explosion, shock-breakout cooling and interaction with circumstellar material can produce early light-curve excesses and 'flash' spectra—narrow emission lines from ionized material close to the star. These last hours to days. A shadowing survey at ~22.5 mag could search for SN 2023ixf-like excesses across dozens of nearby core-collapse supernovae each year. Rapid spectroscopy of at least 150 Type II supernovae per year within ~160 Mpc could constrain the incidence of flash features to less than 10% within a few years, and tie that rate to precursors.

For Type Ia supernovae—thermonuclear explosions used as cosmic distance markers—the paper estimates ~250 classified per year within ~200 Mpc. That could yield hundreds of early light-curve excesses over LSST's lifetime, plus a robust measurement of extreme high-velocity features above 25,000 km/s. Early spectra can also reveal unburned carbon, a clue to how the white dwarf burned.

The infrastructure problem: 80 hours can be an eternity

The paper flags a practical obstacle. Full-frame processed LSST images are subject to an 80-hour embargo by the funding agencies. That blocks visual inspection and forced photometry at the very times when early signals matter. The authors request a waiver for fields around nearby galaxies, within roughly 100–200 Mpc. They also advocate for a Nearby Galaxy Transient Broker: a single platform that merges transient streams from LSST and other surveys, adds forced photometry and non-detections, pulls archival light curves and space-based imaging, and triggers follow-up. Such a system would help identify the youngest transients and coordinate observations.

What could go wrong—and what the numbers mean

The rates in the paper are forecasts, not LSST discoveries. They rely on current supernova statistics, assumed survey depths, and the simplifying assumption that current nearby discovery rates translate to the LSST era. The authors also warn that early classification is ambiguous: the very young SN 2023bee was first classified as a Type Ic before a spectrum a day later showed it was a Type Ia. Faint, nonterminal transients such as luminous blue variable eruptions and luminous red novae can mimic young supernovae. Distinguishing them requires rapid spectroscopy, plus historical variability at the transient site and host-galaxy context. Space-based facilities—Hubble, Swift, ULTRASAT, Einstein Probe, and future missions—are important but limited in response time and access.

From chance encounters to a statistical probe

Despite those caveats, the prize is clear. If shadowing surveys and an integrated broker work as envisioned, the first hours of a supernova—once caught only by luck—could become a routine, quantitative probe of stellar evolution and explosion physics. LSST will provide the depth and the discovery engine. The challenge is to add the cadence, the spectroscopy, and the infrastructure to see what happens before the story has already begun.