Mapping the Cloud-Driven Atmospheric Dynamics & Chemistry of an Isolated Exoplanet Analog with Harmonic Signatures

TL;DR

We apply damped Fourier harmonic analysis to time-resolved JWST NIRISS and NIRSpec spectra of the highly variable exoplanet-analog SIMP J0136, mapping how cloud layers (notably forsterite and iron) modulate flux across atmospheric pressures from $\lesssim 0.1$ bar to several bars. Using a cloudless $T_{eff}=1150$ K atmospheric contribution function, we translate harmonic parameters into vertical maps, revealing two interacting atmospheric layers and pressure-dependent variability, with odd harmonics ($k\ge 3$) tracing deep, equatorially concentrated cloud modulation. The results show anti-correlations between the forsterite cloud region and H$_2$O/CO absorption, and correlations with CH$_4$ at certain depths, consistent with cloud-driven temperature shifts and disequilibrium carbon chemistry; high-altitude CH$_4$ emission signatures imply auroral heating possibly from electron precipitation. Overall, the work demonstrates a novel vertical diagnostic of cloud-chemistry-temperature coupling in L/T-transition planetary-mass atmospheres and highlights the potential of multi-rotation JWST time-resolved spectroscopy for atmospheric Doppler imaging and dynamical studies of exoplanet analogs.

Abstract

Young planetary-mass objects and brown dwarfs near the L/T spectral transition exhibit enhanced spectrophotometric variability over field brown dwarfs. Patchy clouds, auroral processes, stratospheric hot spots, and complex carbon chemistry have all been proposed as potential sources of this variability. Using time-resolved, low-to-mid-resolution spectroscopy collected with the JWST/NIRISS and NIRSpec instruments, we apply harmonic analysis to SIMP J013656.5+093347, a highly variable, young, isolated planetary-mass object. Odd harmonics (k = 3) at pressure levels ~ 1 bar, corresponding to iron and forsterite cloud formation, suggest a potential North-South hemispheric asymmetry in the cloudy, and likely equatorial, regions. We use the inferred harmonics, along with 1-D substellar atmospheric models, to map the flux variability by atmospheric pressure level. We identify distinct time-varying structures in the near-infrared that we interpret as planetary-scale wave (e.g., Rossby or Kelvin)-associated cloud modulation. We detect deviations from bulk (composite) variability in water (S/N = 14.0), carbon monoxide (S/N = 13.0), and methane (S/N = 14.9) molecular signatures. Forsterite cloud modulation is anti-correlated with overlying carbon monoxide and water abundances and correlated with deep methane absorption, suggesting complex interaction between cloud formation, atmospheric chemistry, and temperature structure. Furthermore, we identify distinct harmonic behavior between methane and carbon monoxide absorption bands, providing evidence for time-resolved disequilibrium carbon chemistry. At the lowest pressures (< 100 mbar), we find mapped methane lines transition from absorption to emission, supporting evidence of high-altitude auroral heating via electron precipitation.

Figures (6)

• Figure 1: NIRISS/SOSS (left) and NIRSpec/PRISM (right) spectra for SIMP$\,$J0136. Top Row: Time-averaged spectra normalized to maximum emission. Bottom Row: Per pixel (purple) and spectrally and temporally binned (magenta) signal-to-noise ratio. NIRISS/SOSS and NIRSpec/PRISM data were previously presented in Akhmetshyn2025 and McCarthy2025Nasedkin2025, respectively.
• Figure 2: NIRISS/SOSS (left) and NIRSpec/PRISM (right) variability maps (dynamics spectra) for SIMP J0136. Variability maps use 0.05 $$m spectral binning and demonstrate the normalized spectral variability over the periods of observation. Integration times for each temporal bin are 118 s (NIRISS/SOSS) and 129 s (NIRSpec/PRISM). Top Row: Observed variability maps. Middle Row: Best-fit variability maps with lightcurves inferred using harmonic model described in § \ref{['ssec:models:wave_model']} and Bayesian nested sampling. Bottom Row: Residuals between observed and best-fit variability maps. NIRISS/SOSS's residuals ($0.00 \pm 0.215$ %) approximate white noise while the NIRSpec/PRISM residuals ($0.00 \pm 0.075 \%$) contain seemingly structured harmonics signals of $k \gtrsim 3$, likely due to unmodeled dynamics or systematics. NIRISS/SOSS and NIRSpec/PRISM data were previously presented in Akhmetshyn2025 and McCarthy2025Nasedkin2025, respectively.
• Figure 3: NIRISS/SOSS comparison of retrieved harmonic periods and amplitudes as a function of wavelength. Each data point denotes an inferred harmonic mode with either 2 or 3 modes per wavelength (contingent on the best-fit). Data points' sizes are inverse to their relative uncertainties. As seen on the left y-axis: Top Row: Period vs. Wavelength. It can be seen that $k=3$ ($\sim0.8$ h) harmonics correspond to the deepest pressure levels ($\gtrsim1~$bar). Bottom Row: Amplitude vs. Wavelength. Color denotes the dominant spectral absorption feature in each bin: Continuum (purple), H$_{2}$O + CH$_{4}$ (magenta), H$_{2}$O (red), CO (orange), CH$_{4}$ (yellow). The teal, dashed horizontal lines denote 2.4 h ($k=1$), 1.2 h ($k=2$), and 0.8 h ($k=3$) periodicities. The black line corresponds to the maximum flux from the Sonora Bobcat Marley2021 contribution function seen at each wavelength and the associated pressures as seen on the right y-axis. The gray shaded region represents $10^{-3}$ of the maximum contribution function flux.
• Figure 4: NIRSpec/PRISM comparison of retrieved harmonic periods and amplitudes as a function of wavelength. Each data point denotes an inferred harmonic mode with either 2 or 3 modes per wavelength (contingent on the best-fit). Data points' sizes are inverse to their relative uncertainties. As seen on the left y-axis: Top Row: Period vs. Wavelength. It can be seen that $k=2$ ($\sim1.2$ h) harmonics correspond to $\sim1~$bar while $k=3$ ($\sim0.8$ h) harmonics correspond to the deepest pressure levels ($\gtrsim1~$bar). Bottom Row: Amplitude vs. Wavelength. Color denotes the dominant spectral absorption feature in each bin: Continuum (purple), H$_{2}$O + CH$_{4}$ (magenta), H$_{2}$O (red), CO (orange), CH$_{4}$ (yellow). The teal, dashed horizontal lines denote 2.4 h ($k=1$), 1.2 h ($k=2$), and 0.8 h ($k=3$) periodicities. The black line corresponds to the maximum flux from the Sonora Bobcat Marley2021 contribution function seen at each wavelength and the associated pressures as seen on the right y-axis. The gray shaded region represents $10^{-3}$ of the maximum contribution function flux.
• Figure 5: NIRISS/SOSS (left) and NIRSpec/PRISM (right) vertical variability maps for SIMP J0136. (Top Row) Composite maps including variation in % deviation across all observed wavelengths. (Middle Row) S/N vertical map. Pressure levels $\gtrsim 1~$bar have the highest S/N corresponding to higher observational S/N in wavelengths $\lesssim 1.8~$m (Bottom Row) S/N-weighted composite variability maps. Stratification into $\gtrsim2$ interacting layers can be observed. As rotational period is $\sim 2.4$ h, longitudinal wrapping can be observed within each observation. Visually, atmospheric evolution can be seen over the $\sim 35~$h between observations.