See attached PDF

This is a nicely written paper that demonstrates the authors understand the primary issues associated with spectral calibration.

I have only a few comments.

Line 33:

Do the authors mean to say 200 km swath rather than what appears to be an area? Please provide the single-dimension width. 

Line 115:

Assuming that the camera images the sphere aperture (i.e. the sphere is not being used as an irradiance source), how can changing the diameter of the output port alter the sphere radiance (it will alter its irradiance)?

Line 117:

The description of 'calibration curves' lacks clarity.  If you're going to talk about fits you need to describe exactly what you're fitting.

Line 186:

Please be more specific what you mean by 'percent of total detected light.'  Is this the modeled SL for a simulated input scene?  Is the percentage measured relative to the useful signal at each wavelength?  What does 'total detected light' mean?

Line 262:

No mention is made of a radiometric correction to the ISRF.  The iterative oversampling described in Section 5.1 yields empirical scale factors to 'stitch' the multiple pixels together.  The scale factors remove differences in radiometric response between the adjacent spectral pixels.  These differences include PRNU, which you want to remove, but also spectral response variations (because the pixels are at different wavelengths), which you do not want to remove.  This latter variation needs to be included in the ISRFs before they can be used in any forward model calculation.  I would expect any description of the 3D lookup table to mention how this variation is reintroduced.  If the authors have concluded the effect is negligible then they should state so in the text.

Review of amt-2020-513

General: Article presents the analysis methodology and results for characterizing the instrument spectral response and stray light and calibrating the wavelength scale of the airborne simulator for the forthcoming MethaneSAT, which will detect large point sources and characterize basin-scale, diffuse methane emissions.  A good, short description is given for the airborne instrument and test/characterization equipment.  The presentation is very detailed and thorough.  The figures are clear and support the text.  The article is worthy of publication concerning an instrument that will presumably be used to validate the satellite mission.  However, there are a few areas that would benefit from clearer language.

• Spectral pixels vs. central wavelength: From line 11 the term central wavelength is taken to be that of the test laser source, where spectral pixel is a coordinate on the 2-D detector (spatial pixel being the other coordinate). However, Fig. 14 seems to use both central wavelength and spectral pixel inter-changeably. The abscissa of the plots is labeled Central wavelength, but the caption mentions spectral pixels.
• Median filtering, Lines 245-256: It is not clear to what dimension of the table in line 250 the median filter is applied.
• Line 264: Again confused by the reference to “all spectral pixels” when I thought the ISRF was only measured/derived for specific “central wavelengths” where the laser source operated.
• Table 2: The footnote #6 is missing, but referenced in the last column.
• Line 343: From what was presented in the article it is not definitive that the possible inhomogeneous illumination of the slit produced asymmetric spectral response functions.

Line 115: How was saturation defined / measured?  How uniform is the saturation level across the spatial and spectral dimensions?

Line 147: Why did the laser power need to be increased at long wavelengths if a gain correction was already applied?

Fig 13:  Some ISRFs are not monotonic with relative wavelength.  Is this concerning and/or does it have a physical explanation?  Also, are asymmetries understood?

Fig 17:  How small do residuals need to meet science objectives?

Line 318:  In addition to retrieving changes, can any in-flight measurements be made to update ISRF in flight?