JWST Can Find RR Lyrae in a Nearby Galaxy—But Timing Them Is Another Matter

A heartbeat in the infrared

RR Lyrae stars are the metronomes of old stellar populations. In this WLM sample, these aging, metal-poor Population II stars pulse on periods of hours to about a day, and their intrinsic brightness rises and falls predictably. That makes them dependable distance markers—cosmic rulers astronomers use to measure how far away galaxies are. A new study asks whether the James Webb Space Telescope can use them as well as Hubble does.

The target is WLM, a dwarf irregular galaxy about 3 million light-years away. Past estimates have placed it near 0.98 megaparsecs; the new RR Lyrae analysis puts it at 0.93 ± 0.02 megaparsecs. The work has been accepted for publication in The Astrophysical Journal and is available as an arXiv preprint.

WLM: a small galaxy with an overlapping view

To test JWST's ability, the team used archival observations from both telescopes. Hubble's Advanced Camera for Surveys imaged an inner field of WLM for 25 orbits in 2015, using the F475W and F814W filters. JWST's Early Release Science program observed central and outer regions with JWST's near-infrared camera, NIRCam, in the F090W and F150W filters. The outer NIRCam field overlaps the Hubble field, letting astronomers compare the same stars through both telescopes.

The team extracted individual 14-minute JWST exposures from the archive and ran point-spread-function photometry with DOLPHOT—software designed to measure stars by fitting the telescope's blur pattern, even in crowded fields. They then searched for stars whose brightness changed from epoch to epoch and used a Markov Chain Monte Carlo (MCMC) template-fitting technique—a statistical method that samples many possible light-curve parameters—to derive periods, amplitudes, and mean magnitudes. The Hubble data yielded 101 bona fide RR Lyrae stars: 85 fundamental-mode RRab type and 16 of the higher-frequency first-overtone RRc type.

Finding the stars is easier than clocking them

JWST performed well as a finder. The team identified 125 RR Lyrae candidates in the NIRCam data. In the overlap region, 44 stars appeared in both data sets, and a statistical comparison suggested about 48 RR Lyrae are actually present there. The detection probability was 96 percent for Hubble and 92 percent for JWST for these baselines.

One JWST-only variable candidate appears to be real: it failed Hubble's crowding quality cut, so Hubble's variable catalog missed it, while JWST's higher angular resolution allowed recovery. Three Hubble RR Lyrae were missed by JWST because their brightness changes fell below the team's RMS (root-mean-square) variability threshold; loosening that threshold would have admitted more false positives, so the researchers prioritized purity over completeness.

Measuring periods was harder. The JWST observations spanned only about 15 hours, and because NIRCam operations discourage constant filter switching to reduce wear, each filter was observed in a separate block. Many sources—especially longer-period RRab stars—had little more than one full cycle, and some had less. The period-finding algorithm could recover short-period RRc stars, but for RRab stars the power spectrum flattened at long test periods. The template fitting struggled to lock on for both classes, even the shortest-period RRc stars, because the F150W baseline was less than one full cycle. When the team fixed the JWST fits to periods measured by Hubble, the fits became robust. The limitation was the observing baseline, not the stars.

A distance that leans on Gaia

The Hubble data produced a distance modulus—a logarithmic measure of distance—of 24.85 ± 0.05, corresponding to 0.93 ± 0.02 megaparsecs. That is closer than several recent previous WLM measurements, such as the 24.93 to 24.98 values from Lee et al. and Yan et al. The difference is less than a 2σ tension and is not yet a crisis.

The team attributes the offset mainly to calibration. Their Hubble period–Wesenheit–metallicity relation is tied to the Gaia distance scale, and previous work has found that including Gaia can shift distances about 3 percent closer. Could uncertain metallicity—the abundance of elements heavier than helium—explain the difference instead? Probably not. To match the older weighted-average distance of 24.96, the RR Lyrae would need a mean [Fe/H] of about −2.49 ± 0.21, lower than expected for this galaxy's RR Lyrae population.

The first NIRCam calibration, with wide error bars

Using 35 RRab stars in the overlap, anchored to the Hubble distance, the team fit a period–Wesenheit–metallicity (PWZ) relation in JWST's F090W and F150W filters. The PWZ relation combines pulsation period, a reddening-corrected brightness called Wesenheit, and metallicity to predict distance. This is the first such calibration in NIRCam bands. It is also preliminary: the zero-point offset is α = −1.05 with uncertainties of +0.21/−0.26, and the slope is b = −1.17 with uncertainties of +0.96/−1.26. The slope is poorly constrained, and the errors are much larger than in the Hubble calibration. A larger sample and better metallicity measurements will be needed to sharpen it.

NIRCam's tiny zero-point drift

The team also checked how stable NIRCam's brightness measurements were from one epoch to the next. The median systematic zero-point offset—the baseline brightness calibration—was at most 0.004 magnitudes, negligible compared with the template-fitting uncertainties in this study. But the offsets followed a pattern: within each set of consecutive observations, they rose slightly and then reset at dithers (small pointing shifts between exposures). The behavior is more consistent with detector burn-in—charge traps filling up and briefly reducing signal—than with detector persistence, whose timescales in Hubble's infrared camera are much longer than the pattern seen here. Future high-precision time-series work may need to account for it.

Give JWST a longer listen

JWST's infrared vision can find old stars in crowded fields, including one that Hubble missed. But for variable stars, observing strategy matters. The WLM data show that short, continuous blocks are poor for measuring longer RR Lyrae periods. Future programs should split observations into multiple epochs spread over a longer baseline, even if that adds overhead. More galaxies will help too: Leo P, for example, already has JWST observations spanning 36 hours over four days, more than twice the WLM JWST duration.

The RR Lyrae in WLM are not just points on a plot. They are clocks marking the oldest stars in a small galaxy. The new work shows that JWST can hear their infrared heartbeat—but to read the time accurately, astronomers will need to give it a longer listen.