the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 17 Feb 2026
MAJIS performances in the infrared during the JUICE 2024 Earth fly-by: comparisons with IASI measurements and sensitivity to trace species
Abstract. The JUpiter ICy moons Explorer spacecraft (JUICE) successfully performed a Lunar and Earth gravity assist maneuver on 19 and 20th August 2024, which provided an excellent opportunity to test its instruments before its arrival at Jupiter in 2031. Here we focus on the evaluation of the performances of the infrared channel of the Moon and Jupiter Imaging Spectrometer (MAJIS) based on Earth observations acquired over part of the Pacific Ocean at tropical latitudes. We specifically compare MAJIS observations with co-located ones from the Infrared Atmospheric Sounding Interferometer (IASI). The two instruments overlap in the spectral range 3.6 and 5.56 µm. Having removed spectra contaminated by clouds or ocean glint, we find an excellent match between the absolute radiance of the two instruments. We argue that most of the differences can be explained by natural variability of water vapour content. Once this effect is taken into account, our results suggest that the radiometric calibration of MAJIS is better than 10 %, rising to ~15 % in the range 5.25–5.56 µm. We then compare MAJIS to synthetic spectra generated with a radiative transfer algorithm, the OPerational version of the Automatized Atmospheric Absorption Atlas (4A/OP). Both the comparison of MAJIS to IASI and to synthetic spectra reveal a small spectral shift of MAJIS spectra beyond 4 µm, of the order of 4 nm. The use of 4A/OP allows us to highlight the detection and sensitivity of MAJIS spectra to several trace species: methane (CH4), nitrous oxide (N2O), ozone (O3) and carbone monoxide (CO). Based on the residuals between MAJIS and synthetic spectra, we estimate the signal-to-noise ratio to lie in the range 200–300, meeting nominal expectations. These excellent performances point to promising future jovian atmospheric observations.
- Preprint
(3927 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-805', Anonymous Referee #1, 16 Mar 2026
-
RC2: 'Comment on egusphere-2026-805', Anonymous Referee #2, 02 Jun 2026
General comments: This is an interesting cross-comparison for the first in-flight characterisation of the MAJIS IR channel against a reference sounder (IASI), applied through an established methodology (4A/OP). The data is well studied and the study merits publication after revision. My main concern is that the authors tends to state the central calibration results more firmly than the actual data support: The comparison uses limited data subset rather than statistics over the full data set, The differences are attributed mainly to water vapor when other effects such as those from calibration bias, pixel-position effects and a high-angle emissivity assumption are omitted. I recommend publication after revision, contingent on the four major comments below.
Major comments
1. The calibration conclusions come from on one example (C15), + some limited C10/C11 yet there are 36 co-locations, each with tens–hundreds of MAJIS spectra. Why do authors do not want to show the mean and scatter of full data set; (MAJIS−IASI)/IASI as a function of wavelength across all 36 co-locations, with error bars? Without this, “better than 10%” remains unsupported. When the authors show the difference, they can add a %-difference plot or curve secondary axis to Fig. 10b.
2. The “best case” is C15 (L176), itself flagged as glint-affected and corrected; C15 is then reused in Section 5 for the SNR and trace-species analysis despite being among the poorer matches. Using a contaminated swath as the radiometric reference, and again as the trace-species testbed, should be justified or replaced with cleaner swaths.
3. The fact that C10/C11 agree well without correction while C15 needs +5% does not confirm the influence of the water-vapour story.The water-vapour attribution, and the +5% bias, should be presented as cpossible hypotheses , not as established. The abstract and conclusions currently overstate this.
4. The flyby geometry is ~50° emission angle, yet 4A/OP uses a fixed emissivity of 0.98 across the whole range. At ~50° the sea-surface MWIR emissivity departs measurably from 0.98 and becomes angle- and wavelength-dependent, and the reflected term grows, wavelength-varying effect of the same magnitude attributed to water vapour. An angle-dependent emissivity should be tested before the residual is assigned to water vapour.Minor comments
Abstract
- Abstract can be slighly misleading since it may imply IASI and MAJIS share similar local times, locations and footprints, whereas all differ and effectively only three IASI spectra are truly comparable. Please state this clearly.
- The “Excellent” is contradictory with “better than 10%” Considering it is an instrument designed for other objectives yes the comparison is indeed impressive but either name the benchmark or soften the sentence.
In Abstract “~15%” beyond 5.25 µm is menstioned here, however in the text and conclusions “~15–20%” — is stated. - For the trace species the authors use uniform capability. The detection significances show each gas is detectable above the noise. The retrieval precisions (Table 2) are a different, theoretical quantity: the minimum column uncertainty achievable if noise were the only error. Usefulness depends on comparing that precision against each gas's natural variability: CH₄ is unusable, CO marginal. The abstract and Section 5 should therefore not present a uniform "sensitivity to trace species.
- Minor: “carbone monoxide” → “carbon monoxide.”
Introduction
- State briefly here the spectral resolutions of both MAJIS and IASI (currently mentioned only later), and note the observations are close but not exactly co-located in local time.
- For description o Majis observations the authors move too fast and rely heavily on forward references to companion papers (Poulet 2025, Langevin 2026, Oliva 2026); some data and figures must stand on its own.
Section 3 — IASI characteristics
- Drifts into describing MAJIS co-located measurements; comaprison of the footprint, local-hour mismatch and comparability should be addressed may be in other sections.?.
Section 4 — Timing & co-location
- Provide a table of the time and local time of each observation actually used (currently deferred to other papers).
- The 06–10 h MAJIS scenes; are they within 3 h of IASI, or 2.5 h as in Table 1? Reconcile, and re-examine the loosened criterion.
- ±2.5 h at these local times is a large geophysical gap; quantify its expected radiance impact (linked to Major comments).
Section 4.1 — Cloud filtering
- Excluding clouds is the good choice (cloud variability would affect the signal); the paper need only state this rationale in one sentence early also (also Table 1).
- For clouds VIS+IR is used on the MAJIS side (0.7 + 4.64 µm); IASI has limited range. Does it have a visible channel ? Figure 5 mentions an onboard camera. how is it used? Authors can discuss further?
- The <5% IASI cloud fraction is not a radiometric floor; The authors can quantify its actual radiometric impact and show it is negligible against the 10% claim.
Section 4.2 — Glint
- Justify the 3.9 µm glint proxy and the >301 K threshold; this explanation arrives very late (L155) and should move forward, next to Fig. 6 (which is vey helpful).
- Glint is itself an additional uncertainty Authors shall make it explicit.
- 11 retained observations are all daylight (L141)....are not they?
- Fig. 7 / L146: can clouds be separated so the glint pattern is clearer? Explain the dark slit/line along the edge of some swaths (C11, C14, C15).
Section 4 — Figure 2
- It can be confusing for reader to see data is over exposed and saturated, nevertheless still have lower radiance values, and he cloud signal in addition the glint. Plesea explain how a a swath can be glint-bright and saturation-dark at once , mechanism and how saturation is distinguished from clouds?
Section 4.4 Comparing IASI and MAJIS spectra
- Explain why the IASI footprint is an ellipse rather than a circle (in figure we see a circle, off-nadir projection foreshortening...
- The 2–5% emission-angle correction (ratio of two 4A/OP spectra, L180) is another error source can be considered it in the error budget.
- Clarify whether the “raw comparison” (L183) is before or after these corrections; the wording is confusing.
Section 4 — Spectral shift & residuals - L172: the 4 nm shift beyond 4 µm is asserted but not shown — provide a demonstrating figure, or state clearly it is documented in Langevin (2026) and confirm the spectra used here already include the correction.
- L183: where does the 0.15 W/m²/sr/µm difference apply (below 4.57 µm?), and what are the peaks above 4.5 µm — water vapour? Explain.
- L210: why 5.25 µm and not 5 or 4.75 µm? Why not anchor on specific water-vapour lines? The “<10% above 5–5.25 µm” claim is a bit arbitrary oplease revise these numbers carefully (and see Major 1 for the % plot).
- Can the water-vapour effect be modelled out with 4A/OP rather than invoked qualitatively?
Section 5 — Synthetic spectra & trace species
- The SNR (200–300) is derived from the variance of (MAJIS−4A/OP), which contains model and measurements error and it is not an instrument-noise metric. The authors can extend the discussion or estimate noise from a cleaner source (e.g. repeated-pixel variance).
Conclusions
- The 4 nm shift is not demonstrated in the paper.
- The water-vapour explanation is a hypothesis, not established (see Major 3)
- The % error figures and band-by-band claims should be revised (see Major 1).
- Trace gases: retain the detection result but temper the retrieval-precision claims (
Citation: https://doi.org/10.5194/egusphere-2026-805-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 724 | 255 | 63 | 1,042 | 115 | 145 |
- HTML: 724
- PDF: 255
- XML: 63
- Total: 1,042
- BibTeX: 115
- EndNote: 145
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The manuscript under consideration is a relevant contribution and generally well written. In some details, the presentation is difficult to follow. I would recommend publication after revision of these shortcomings.
Comments in detail:
I would suggest a slight rephrasing of the title: "MAJIS infrared performance during the JUICE 2024 Earth flyby: comparison with IASI measurements and sensitivity to atmospheric trace species".
page 4: "We did not perform any retrievals in this paper" - a few lines below the authors state "For CO, CH4 and N2O, we .... scaled these profiles in order to obtain a best fit". These are somewhat contradictory statements.
page 5, cloud filtering: The MAJIS cloud filtering via a 4.64 um brightness T threshold appears not to be the optimal choice. Would the addition of a VIS/NIR brightness value further improve the recognition of clouds? (See Fig 3, where in the right panel more filigrane cloud structure tend to be lost.)
page 13, line 193ff: if the discrepancy is induced by differences in H2O as claimed by the authors, I would expect that the quite pronounced spectral patterns shown in Fig 10 at wavelengths beyond 4.6 um correspond to H2O spectral signatures. Adding some further simulations using 4A/OP should allow to verify this claim more explicitely.
page 15, line 239: "potential variations in meteorological conditions". This seems a persisting problem for the whole comparison, especially when it comes to the H2O profile. I wonder whether the ERA5 hourly data combined with 4A/OP runs for estimating the resulting effects on radiances would help to further substantiate the conclusions and perhaps narrow down the residual uncertainty of the MAJIS radiance calibration.
Conclusions: Here, a coincidence interval of 2 hours is mentioned, while the reader would expect (based on section 4.3 and table 1) 2.5 hours?
Further down, the conclusion discusses in length (lines 286 to end) items which have not at all been treated in the paper. Either this part should be removed or an additional section needs to be added which makes the claims of the conclusions explicit.
Data availability: I find an embargo period of about 5 years after data collection of the flyby data (this would imply data accessibility three years after publication of this study) inacceptable, roughly as distant from any contemporary FAIR principles as Jupiter is from the sun. It would appear to me that the minimum requirement from the viewpoint of a study publication would be a release of at least the subset of MAJIS spectra actually used in the form of an electronic supplement.