Here are our replies to each specific comments. If more details are needed, let us know (a warning in red indicates that one cannot put in a revised version a as supplement, and it will in any case have to wait for the comments of a second referee due May 25th). 

My main recommendation would be to add a table to summarise all the artefacts and corrections, the ranges of wavelength it is affecting, their origin, and the correction applied.

The recommended table on artefacts and mitigation will be added (it is mainly based on Langevin et al., 2024, but this information is indeed also of interest for the post-launch adjustment of the spectral and radiometric calibration

• Line 73: Stacking always makes the signal better. Maybe this precision is not needed here.
  • Comment accepted, new formulation : “(up to 450 for single spectra)”
• Lines 107 and 113: the shift it 5.4nm but the correction is 5.2nm, why? Is it a typo?
  • Typo indeed.2 nm replaced by 5.4 nm (the value from Haffoud et al. 2024)
• Lines 156-159: Any threshold values for when it occurs? How much is the underestimation? It seems a lot since it is a log scale.
  • For very deep absorption bands (e.g. H2O 1.9 µm), the underestimation is indeed by a factor of 10. However, section 2.1 addresses the post-launch evolution of the spectral calibration, which relies on the location of spectral features for which the 4A/OP model results are reliable), not on their strengths. New formulation of this paragraph:

The radiative transfer simulations are in good overall agreement with the MAJIS observations. However, the 4A/OP synthetic spectra underestimate the radiance by factors of up to 10 within strong water absorption bands (e.g. at 1.9 µm). This can result from an issue with the selected BRDF or from contamination by aerosols, which has not been considered. There is no impact on the post-launch spectral calibration which is the focus of this section as a wider BRDF or a contamination by aerosols results in a larger radiance in nearly saturated absorption bands without changing their wavelength.

• Line 189: what are the size of these windows and what size where tested?
  • New formulation :

« The same calculations have been performed with 5 sliding windows widths (40, 50, 60, 70, 80 spectels) for assessing the robustness of the spectral residuals »

• Figure 3: Depth of absorptions at 1.1 and 1.4 is underestimated. It is not really discussed is such term in the MS. Around 2.3, the “noise” is about PRISMA or MAJIS? It was not clear to me in the text.
  • Our view is that it is the radiances from the 4A/OP model which are underestimated, most likely due to a dust scattering contribution not considered in the model, not the depth of absorption which is underestimated by MAJIS. This is confirmed by the relatively good agreement on 1.4 µm and 1.9 µm band depths with MAJIS and with PRISMA as both instruments are similarly impacted by a contribution by dust scattering. The comparison with PRISMA is definitely relevant for the post-launch spectral calibration of MAJIS (section 2). However, as discussed in section 4.2.6, no good match with PRISMA in time and location was identified, which made it difficult to use the comparison of absolute radiances and absorption feature strengths from PRISMA and MAJIS for updating the radiometric calibration. 
  • “noise”: As shown in Figure 3, The SNR is higher for MAJIS than for PRISMA in deep absorption bands. New formulation:

« for wavelengths larger than 1.6 µm the PRISMA data becomes noisy for deep absorption bands”. 

• Figures A1 and A2 are very small and difficult to read.
  • These two figures have been reduced to 15 panels (3 lines of 5) instead of 26 panels (3 lines of 7 + 1 line of 5), increasing by 40% the width and by 33% the height of each panel for the same figure size so as to make them more easy to read. They are complementary to Fig. 5 in the text which provides full details for all 4 tested wavelength ranges at one position. The main purpose of including these two figures as an appendix is to show that reliable results have been obtained at all positions along the slit, so that further reduction of the number of panels is not recommended.
• Table 2: this table is very useful. Doing the same for all other instruments would make things easier for readers.
• Lines 1066-1068: why TROPOMI is mentioned here? It is not clear.
  • The mention of TROPOMI is made here in the context of the cross-calibration of the IR channel (2.28 µm to 5.54 µm), with IASI providing the most comprehensive overlap (from 3.7 to 5.54 µm) while TROPOMI provides data at 2.32 and 2.36 µm, close to the short wavelength end of the MAJIS IR channel range. A new formulation is proposed to make the complementarity clearer in this context:

IASI is an instrument on-board the METOP satellites of Eumetsat. The spectral range of IASI (3.7 µm to 15.5 µm) overlaps more than half of the spectral range of the IR channel of MAJIS (2.28 to 5.57 µm). As indicated in section 4.2.4, comparisons with TROPOMI provided complementary cross-calibration information for the lowest wavelengths of the IR channel, which are not covered by IASI.

• Figure 21: add a legend on the figure
  • The figure has been modified adding a label for each curve with the same color.

We fully concur with the first comment: the MAJIS spectral calibration definitely benefited from the comparison of MAJIS earth spectra to that obtained by Earth observation instruments. In particular, this made it possible to reliably monitor the small spectral shifts observed after launch and to confirm that the expected spectral resolution (FWHM) had been achieved by MAJIS. The very good match between radiances observed by MAJIS and IASI from 3.4 to 5.54 µm significantly exceeded our expectations and was very positively received by the Earth atmosphere community. 

The second recommendation ("convolved" instead of "convoluted" has already been implemented.