Minimum Noise Fraction Analysis of TGO/NOMAD LNO Channel High-Resolution Nadir Spectra of Mars

Round 1

Reviewer 1 Report

Review of remotesensing-2735923: Minimum Noise Fraction analysis of TGO/NOMAD LNO channel high-resolution nadir spectra of Mars

This paper investigates the usefulness of MNF in enhancing the quality of LNO data and to what extent such improvements help understand Mars via those data. The paper starts with a controlled investigation using synthetic data, followed by trial applications to the real LNO data. Specific techniques designed for thresholding and applying MNF to real LNO data were presented. The paper proceeds to suggest that MNF is very unlikely to help detection of faint trace gas (e.g., methane) but demonstrates appealing possibilities in mapping global/regional widespread species.

The investigation provides a possibility for better exploitation of the LNO data and an explorative basis for developing future spectral filtering methods, and the case studies are appealing. To me, the paper is a bit lengthy but it is acceptable given the extensive work and case studies presented. I have only a few suggestions/comments below:

The authors can briefly elaborate on the scientific goals of LNO to the readers. What do we expect to find from LNO observations?

In section 3.2.1 the authors suggested a Ne selection technique based on the overlapping region of u_b+-sigma_b and u_y+-sigma_y with uncertainties. The authors may need to explain further how the uncertainties of <u_b+-sigma_b> and <u_y+-sigma_y> are computed, since readers may have an impression that sigma_b and sigma_y already represent the uncertainties.

In section 3.2.2, p8 ln 321-322. Why do the authors need to average over regions of the same radius?

In section 4.3.1 (p13, ln 507-529) the authors presented how different Ne and latitudes affect the absorption depth. In one case, the authors suggested that the spectra shape is “reliably reconstructed” (ln 526). These have to be assessed visually through Fig. 7. I would suggest the authors add a partially quantitative assessment to support their analysis. Such as giving the approximate band depths of different cases and that of the simulated spectra.

I’d also give similar suggestions to sections 5.2 and 5.3. The authors can consider providing the scatter plots of Fig. 9A vs 9B and 9A vs 9D, each data point on the plot corresponds to a 1 deg by 1 deg data and perform a line-fitting (in e.g., Excel). In the perfect case, one should get a neat diagonal line (i.e., slope of 1, intercept of zero, and R-squared of 1.0). Such plots help determine if the data is over/underestimated relative to the MCD, how good is MNF in reducing the spread, and if there are systematic trends in the uncertainties. For section 5.3 the plot would be Fig. 10C vs 10B but of course, only areas with data are concerned.

The text is easy to read. To me, it is a bit lengthy but acceptable since the analysis is extensive. In general, some sentences can be simplified or shortened by using e.g., a more straightforward tone.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper presents interesting results aimed at an enhancement of SNR in the spectra measured in the Martian environment. The MNF approach based on the Principal Components technique gives a possibility to improve the retrievals of atmosphere and surface composition from the measured spectra. It can be useful for development of remote sensing methods

Remarks:

1) The MNF method was applied to the CO column and CO2 ice band depth retrievals (pages 18-19). It is advisable to point out the initial error of the retrievals without MNF-processing and after using of this technique.

2) The fonts in Figures 1, 2, 5, 8, 9 are too small.

The paper is good structured and contains detailed analysis of the MNF-processing adaptability to the spectra measured by NOMAD Instrument.

The article can be published after small revision.

Small correction is need

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

I think this work attempts to enhance the observed SNR of the TGO-NOMAD LNO using Minimum Noise Fraction (MNF). I think the results are more interesting and worth publishing. However, I am skeptical about the effectiveness of the application to actual observational spectra. I.e., to what extent is this SNR enhancement algorithm able to reduce the observational noise, and what noise can be effectively reduced and what can't? Does the author have any further detailed suggestions for using this method? 

Author Response

Please see the attachment

Author Response File: Author Response.pdf