This paper investigates whether JWST can measure the wavelength dependence of stellar granulation—the stochastic brightness variations caused by hot material rising and cooler material sinking in stellar photospheres. Granulation introduces noise on timescales of minutes to hours, potentially biasing exoplanet transit measurements and transmission spectra. The authors examine two planet-hosting stars, the G7V star LTT-9779 and the hotter F6V star WASP-121, to measure granulation amplitudes and characteristic timescales and determine how the amplitudes vary across the near-infrared. The analysis uses publicly available JWST/NIRISS SOSS phase-curve observations spanning 20.6 hours at a cadence of 16.5 seconds for LTT-9779 and 36.9 hours at a cadence of 38.5 seconds for WASP-121. The authors independently reduced the observations using exoTEDRF, correcting for detector effects, background contamination, spectral-trace motion, and instrumental variability. They then fitted and subtracted planetary transits, eclipses, phase variations, and relevant stellar and instrumental trends. From the residual light curves, the authors calculated power spectral densities and fitted models combining a Harvey-like granulation background, white noise, and a Gaussian envelope representing possible solar-like oscillations. Bayesian sampling constrained the model parameters. The analysis compared three main prescriptions: two single-component models with different high-frequency slopes and a two-component model that also accommodates excess low-frequency power. Solar irradiance observations served as a reference. The authors recommend the shorter-timescale component of the two-component model as their granulation estimate. For LTT-9779, this component yields a root-mean-square amplitude of 53.8 ppm, with uncertainties of +4.1 and −4.0 ppm, and a characteristic timescale of 4.8 minutes, with uncertainties of +1.5 and −1.3 minutes. For WASP-121, the corresponding amplitude is 59.3 ± 4.3 ppm, and the timescale is 6.3 minutes, with uncertainties of +2.1 and −1.8 minutes. Solar-calibrated scaling relations predict amplitudes and timescales of approximately 42 ± 2 ppm and 3.2 ± 0.4 minutes for LTT-9779, and 57 ± 5 ppm and 6 ± 1 minutes for WASP-121. WASP-121 therefore agrees well with these predictions, whereas LTT-9779 exhibits a higher amplitude and a somewhat longer timescale. The wavelength-resolved analysis uses six channels covering 0.62–2.82 micrometers, with timescales fixed to their broadband estimates. All three model prescriptions show an overall decline in granulation amplitude toward longer wavelengths. For example, the single-component model with a high-frequency slope of two yields amplitudes decreasing from 59.0 to 45.2 ppm for LTT-9779 and from 66.6 to 50.7 ppm for WASP-121 between channels centered at 0.735 and 2.270 micrometers. Three-dimensional magnetohydrodynamic simulations reproduce the general trend but do not match all measured amplitudes or spectral slopes. The authors also measure oscillation-envelope peak frequencies of 3382 microhertz for LTT-9779 and 1649 microhertz for WASP-121. The latter identification remains tentative because the frequency falls below scaling predictions and may be affected by instrumental contamination. The study demonstrates that JWST spectrophotometry can characterize the wavelength dependence of stellar flicker, potentially improving exoplanet analyses and providing diagnostics of stellar convection. It also reconciles some discrepancies with published Gaussian-process measurements by applying consistent definitions of amplitude and timescale. Important limitations remain. Instrumental peaks require masking, the longer-timescale component cannot be securely attributed to granulation, and model selection generally favors the simpler single-component description. Uncertainties in the wavelength-resolved amplitudes do not account for uncertainty in the fixed timescales. In addition, the available simulations do not match the targets’ stellar parameters closely enough to support quantitative interpretation. Finally, fitted characteristic timescales should not be interpreted directly as the lifetimes of individual granules.
- JWST/NIRISS observations of LTT-9779 and WASP-121 demonstrate that stellar granulation and its wavelength dependence can be measured, helping characterize noise that can bias exoplanet transit measurements and transmission spectra.
- Method: The authors removed planetary signals and stellar/instrumental trends, then fitted residual power spectra with granulation, white-noise, and possible stellar-oscillation components.
- Recommended granulation estimates: LTT-9779 has an RMS amplitude of 53.8 (+4.1/−4.0) ppm and a characteristic timescale of 4.8 (+1.5/−1.3) minutes; WASP-121 has 59.3 ± 4.3 ppm and 6.3 (+2.1/−1.8) minutes.
- Scaling relations: WASP-121 agrees well with solar-calibrated predictions, whereas LTT-9779 shows a higher amplitude and somewhat longer timescale than predicted.
- Wavelength dependence: All three model prescriptions show granulation amplitudes declining toward longer wavelengths across six channels covering 0.62–2.82 micrometers.
- Simulations reproduce the general wavelength trend but not all measured amplitudes or slopes; mismatched stellar parameters prevent reliable quantitative interpretation.
- Oscillations: Envelope peak frequencies are 3382 microhertz for LTT-9779 and 1649 microhertz for WASP-121, but the latter identification is tentative and may reflect instrumental contamination.
- Model limitations: Although the authors recommend the shorter-timescale component of a two-component model, model selection generally favors simpler single-component models, and the longer-timescale component cannot be securely attributed to granulation.
- Additional caveats: Instrumental peaks require masking, wavelength-resolved amplitude uncertainties exclude uncertainty in fixed timescales, and fitted timescales are not individual granule lifetimes.
- Consistent definitions of amplitude and timescale help reconcile discrepancies with published Gaussian-process measurements.
Summary1. Scientists used the James Webb Space Telescope to study tiny changes in the light of two stars.
2. They measured granulation, which is a pattern of bright and dark patches on a star.
3. These light changes happen over a few minutes and look smaller in light with longer wavelengths.
4. Understanding these changes helps scientists study planets that pass in front of stars.
5. Computer models explained some results, but telescope noise and other puzzles mean scientists still have more to learn.
Definitions- telescope: A tool that helps us see things far away in space.
- star: A huge, hot ball of glowing gas, like our Sun.
- granulation: Bright and dark patches made by hot gas rising and cooler gas sinking on a star.
- wavelength: The distance between the tops of two waves of light.
- planet: A large, round world that travels around a star.
- computer model: A computer's way of copying how something works to help us understand it.
- noise: Unwanted changes in a measurement that can hide what scientists want to see.
JWST Reveals How Stellar Flicker Changes Across Wavelengths
When astronomers study an exoplanet passing in front of its star, they look for tiny changes in starlight. But the star itself is not a perfectly steady lamp. Its surface churns as hot material rises and cooler material sinks, producing a shifting pattern of bright and dark regions called granulation. These changes create a subtle flicker that can complicate measurements of a planet’s atmosphere.
A new study investigates whether the James Webb Space Telescope can measure how this flicker changes with wavelength. Using observations of two planet-hosting stars, the researchers find that granulation generally becomes weaker toward longer near-infrared wavelengths. Their results offer both a way to better understand stellar convection and a route toward more reliable exoplanet measurements.
Why Stellar Granulation Matters
Granulation produces stochastic, or irregular, brightness variations on timescales of minutes to hours. That overlaps with the timescales astronomers care about when measuring planetary transits. If stellar variability is not properly accounted for, it can distort estimates of transit depth and potentially introduce misleading features into transmission spectra—the wavelength-dependent measurements used to investigate planetary atmospheres.
The researchers examined LTT-9779, a G7V star, and WASP-121, a hotter F6V star. Comparing these different stars allowed them to test expectations for granulation amplitudes and timescales while exploring whether JWST could resolve the effect across its near-infrared observations.
Separating Stellar Flicker From Planetary Signals
The team used publicly available observations from JWST’s Near Infrared Imager and Slitless Spectrograph in Single Object Slitless Spectroscopy mode, known as NIRISS SOSS. These phase-curve observations followed changes in the combined light from each star and its planet over extended periods:
- LTT-9779 was observed for 20.6 hours, with measurements every 16.5 seconds.
- WASP-121 was observed for 36.9 hours, with measurements every 38.5 seconds.
The researchers independently processed the data using exoTEDRF, correcting detector effects, background contamination, movement of the spectral trace, and instrumental variability. They then fitted and removed planetary transits, eclipses, orbital phase variations, and relevant stellar and instrumental trends.
To study the remaining brightness fluctuations, they calculated power spectral densities, which show how variability is distributed across frequencies. Bayesian sampling constrained models combining a Harvey-like granulation background, white noise, and a Gaussian envelope representing possible solar-like oscillations. Solar irradiance observations provided a reference.
What the Measurements Reveal
The analysis compared two single-component prescriptions with different high-frequency slopes and a two-component model that could also accommodate excess low-frequency power. The authors recommend the shorter-timescale component of the two-component model as their granulation estimate, although model selection generally favors the simpler single-component description.
For LTT-9779, the recommended granulation amplitude is 53.8 parts per million, with uncertainties of +4.1 and −4.0 parts per million. Its characteristic timescale is 4.8 minutes, with uncertainties of +1.5 and −1.3 minutes. For WASP-121, the amplitude is 59.3 ± 4.3 parts per million, and the timescale is 6.3 minutes, with uncertainties of +2.1 and −1.8 minutes.
Solar-calibrated scaling relations predict approximately 42 parts per million and 3.2 minutes for LTT-9779, compared with 57 parts per million and 6 minutes for WASP-121. WASP-121 therefore agrees well with expectations. LTT-9779 shows stronger variability and a somewhat longer characteristic timescale than predicted.
A Clear Wavelength Trend
The researchers divided the observations into six wavelength channels spanning 0.62–2.82 micrometers, fixing the characteristic timescales to their broadband estimates. All three model prescriptions showed an overall decline in granulation amplitude toward longer wavelengths.
For example, the single-component model with a high-frequency slope of two yielded amplitudes falling from 59.0 to 45.2 parts per million for LTT-9779, and from 66.6 to 50.7 parts per million for WASP-121, between channels centered at 0.735 and 2.270 micrometers. These values belong to that specific model and should not be confused with the recommended broadband estimates.
Three-dimensional magnetohydrodynamic simulations reproduced the general downward trend, but not every measured amplitude or spectral slope. Because the available simulations do not closely match the target stars’ properties, they cannot yet support a precise quantitative interpretation.
Promising Results, With Important Caveats
The study also measured possible oscillation-envelope peaks at 3382 microhertz for LTT-9779 and 1649 microhertz for WASP-121. The latter remains tentative: it falls below scaling predictions and may be affected by instrumental contamination.
Other limitations matter, too. Instrumental peaks required masking, and the longer-timescale component cannot be securely attributed to granulation. The wavelength-resolved amplitude uncertainties exclude uncertainty in the fixed timescales. Crucially, fitted characteristic timescales are statistical descriptions of variability, not direct measurements of individual granules’ lifetimes.
Despite these caveats, the study demonstrates that JWST can characterize the wavelength dependence of stellar flicker. It also reconciles some differences with published Gaussian-process measurements by using consistent amplitude and timescale definitions. As exoplanet studies pursue increasingly subtle atmospheric signals, understanding the changing light of the host star is becoming an essential part of the measurement.
