Independent research report by Biswajit Jana · Live report · ORCID · Complete research portfolio
AI-generated artist's concept — not a real photograph. See the report for actual JWST NIRSpec/G395H data.
An ultra-hot Jupiter on a 1.27-day orbit, tidally distorted by its host star. This repo audits a public JWST NIRSpec/G395H phase-resolved emission dataset in both the directly measured planet/star flux-ratio space and the derived brightness-temperature space. Detector-band phase curves are kept separate, while the temperature contrast remains explicitly non-bolometric.
Open the full report — the live GitHub Pages version. You can also open index.html locally in a browser, or serve it with python -m http.server from this directory.
-
System parameters — from the NASA Exoplanet Archive TAP service (
pscompparstable). -
Phase-resolved emission spectrum — Zenodo v8 JWST NIRSpec/G395H planet/star flux ratio and brightness temperature as functions of wavelength (349 channels, 2.7-5.2 microns) and orbital phase (36 out-of-transit/eclipse bins), from Evans-Soma, Sing et al. (2025), released publicly on Zenodo (10.5281/zenodo.20651891).
-
Analysis —
scripts/analyze_spectrum.pyconstructs separate NRS1 and NRS2 flux-ratio phase curves, fits descriptive first harmonics, and retains the phase-window brightness-temperature comparison as a secondary spectral view. Every fit is labelled with its diagonal-error limitation. Run it:pip install -r requirements.txt python scripts/analyze_spectrum.py
index.html the report webpage
data/ seven pinned JWST NIRSpec/G395H Zenodo v8 products
scripts/analyze_spectrum.py detector phase curves + temperature audit
scripts/validate_data.py cross-platform source identity and shape checks
figures/ generated figure, CSV phase curves, summary CSV/JSON
tests/ analytic, regression, failure-mode, and output tests
tests/test_analysis.py checks the per-wavelength weighted-mean
function against a hand-computed case (including that a zero-error
point gets zero weight rather than dominating the mean), and reruns
the full pipeline on the real downloaded phase curve, verifying it
still reproduces the numbers this README documents — including the
explicit "not bolometric" label staying attached to the scalar
temperature. Runs automatically on every push via GitHub Actions; run
locally with:
pytest tests/ -vThe detector-band flux curves are the primary result. Descriptive first- harmonic fits peak 2.934° (NRS1) and 2.358° (NRS2) before eclipse. Their conjunction estimates are 3915/158 ppm for NRS1 and 4910/620 ppm for NRS2 (eclipse/transit). These same-observation estimates closely track the published broadband values and hotspot offsets; they are a reproducibility cross-check, not independent confirmation. Reduced χ² values of 14.36 and 13.07 show that a first harmonic plus diagonal formal errors is inadequate for parameter inference, so this repo does not attach detection significance or physical uncertainty to those descriptive offsets.
The secondary temperature view gives wavelength-averaged monochromatic values of 2722.1 ± 1.7 K for dayside-facing bins and 1237.5 ± 3.3 K for nightside-facing bins, with a 1493.2 ± 3.8 K contrast. These tiny formal errors must not be interpreted as uncertainties on hemisphere effective temperatures.
Detector-band averages assume independent wavelength-channel errors, although the release provides no inter-channel covariance. The large harmonic-fit reduced χ² values empirically warn against using those formal errors for inference. The phase bins exclude the transit and eclipse themselves, so conjunction values are model interpolations across gaps.
Brightness temperature is wavelength dependent—different channels probe different opacities and pressures — so the ± 3-4 K quoted above is the statistical precision of averaging monochromatic values across this bandpass, not a physical uncertainty on a bolometric hemisphere temperature. The paper's own per-detector nightside values show this directly: 926 ± 12 K on one NIRSpec detector (2.70-3.72 μm) versus 1122 ± 10 K on the other (3.82-5.15 μm) — a genuine ~200 K difference between two broad bands. A separate NIRISS/SOSS phase-curve analysis (Splinter et al. 2025), which explicitly models the part of the spectrum this bandpass doesn't cover, derives bolometric effective temperatures of Tday = 2717 ± 17 K and Tnight = 1562 ± 19 K — the physically meaningful numbers for an energy-budget calculation, which this repo's wavelength average is not intended to replace.
One point in the brightness-temperature uncertainty products is zero; it is excluded from the relevant inverse-variance average rather than treated as infinitely precise. The flux-ratio uncertainty matrix contains no zero values.
- Delrez, L. et al., 2016. WASP-121 b: a hot Jupiter close to tidal disruption transiting an active F star. Monthly Notices of the Royal Astronomical Society, 458(4), pp.4025-4043.
- Evans, T.M. et al., 2017. An ultrahot gas-giant exoplanet with a stratosphere. Nature, 548, pp.58-61.
- Evans, T.M. et al., 2018. Detection of H2O and Evidence for TiO/VO in an Ultra-Hot Exoplanet Atmosphere. The Astrophysical Journal Letters, 822, L4.
- Evans-Soma, T.M., Sing, D.K. et al., 2025. SiO and a super-stellar C/O ratio in the atmosphere of the giant exoplanet WASP-121b. Nature Astronomy, 9(6), pp.845-861 (arXiv:2506.01771).
- May, E.M. et al., 2023. A JWST NIRSpec Phase Curve for WASP-121b: Dayside Emission Strongest Eastward of the Substellar Point and Nightside Conditions Conducive to Cloud Formation. The Astrophysical Journal Letters, 943(1), L17 (arXiv:2301.03209).
- Splinter, J. et al., 2025. Precise Constraints on the Energy Budget of WASP-121b from its JWST NIRISS/SOSS Phase Curve (arXiv:2509.09760).
- Zenodo record 10.5281/zenodo.20651891, "WASP-121b JWST NIRSpec/G395H data products."
- NASA Exoplanet Archive, https://exoplanetarchive.ipac.caltech.edu/.