A Unified Brightness Temperature Features Analysis Framework for Mapping Mare Basalt Units Using Chang’e-2 Lunar Microwave Sounder (CELMS) Data
Maxim L. Litvak
Round 1
Reviewer 1 Report
I have read paper "A Unified Brightness Temperature Features Analysis Framework for Mapping Mare Basalt Units Using Chang’E-2 LunarMicrowave Sounder (CELMS) Data" and think that presented study is appropriate for the publication in Remote Sensing journal. Authors presented the analysis of microwave sounder (installed onboard Chang’E-2 mission) measurements to understand how it is sensitive for the detection of basalt and ejecta units. The FeO/TiO2 abundance map of Mare Fecunditatis (proposed by Kramer et al.) was selected as a test area for this purpose. It was shown that cluster analysis of CELMS data could distinguish units found by Kramer et al. Moreover it could create CELMS own geological map of Mare Fecunditatis.
Before acceptance for the publication I would like to clarify some questions:
(1) As far as I understood from the other publications the brightness temperature maps suffer from various systematic effects and the difference dTb (Day - Night) maps is a good way to minimize them. That is why I recommend to focus efforts on the analysis of dTb maps. The analysis of Tb is OK for the particular region, but it cannot be easily expanded to the other lunar areas.
(2) CELMS is operating at different frequencies. Each frequency means different penetration depth. Authors explained that manipulating with frequencies one can test layering structure from the top down to the 2 m depth. For the comparison with Kramer et al. I recommend to analyze the frequency which provides the similar depth. Taking into account that Kramer et al used Gamma-ray spectrometer and UV spectrometer one can consider the depth down to tens of cm or 19.35 GHz/37 GHz in CELMS data. My impression is that Figure 8 confirms this point.
(3) It is seen that Tb map at 3GHz is quite different from the Tb maps at other frequencies and is different from the Kramer's map (Figure 2). I could assume that distribution at depth >1 m is significantly different from near subsurface distribution. I think it deserves special discussion and could be considered as most important result of the paper.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The submitted study by Li et al. is very interesting and in part innovative in trying to ‘make sense’ of the great potential of the CELMS data within the microwave region of the spectrum.
I cannot, unfortunately, comment on the mathematical approach part of the paper since my background is in geology and remote sensing from a comparative and mainly qualitative angle.
Taking this into consideration, I have found several weaknesses in the paper, apart from language, but this could be easily fixed (I offer several suggestions).
The main problem of this study is the absence of the framing of the findings in a geological context outside of the comparison with Kramer’s (2008) map. I am very familiar with CELMS’ data and the images in Fig. 2 were the base of some of my publications. However, as a geologist, I concentrate on the differences between channels and day/night data, called by the authors dTb. My interpretation (the authors would call it “artificial”) of these is that at 3 GHz (a3) the dTb is mostly influenced by the dielectric properties of the surface and subsurface materials, the metal titanium in its various mineralogical associations, like ilmenite. At higher channel frequencies (d3) the contribution of surface rocks during the day fades.
The clearest example of this phenomenon is the Ihtm, the high-titanium mare patch squeezed in the north between the main craters’ ejecta. At the lowest frequency (a) the differences between noon and midnight are small, in the range of ~6 K, as per your figure. Given the error margin (not given!), this is relatively marginal. Still, by comparing a1 and a2 we can see that the largest differences in Tb during the day correspond broadly to Ihtm and during the night to the lower Ti area (Chtr). Why? This question is not addressed in the paper. What makes one area ‘hotter’ during the day and the other during the night? The answer again ought to be sought in the ‘dual’ characteristic, or better, sensitivity of the microwave radiation: dielectric properties (FTA) of the surface materials and Rock Abundance.
The task of the scientist is to disentangle these two elements to see which is predominant and when.
To conclude, the approach of this study is good, and I am sure the comparative techniques pulling different mathematical and analytical approaches are equally valid and interesting. What is missing is a geology/analysis part. Also, given the confidence in the results by the authors, it would have been useful for them to offer an ‘interpretation kit’ in the end, such as: “if you combine noon data from Ch1 with night from C4 and subtract C3 (I am making this up, of course!), then the resulting map will approximate compositions A, B, C for the corresponding geological areas”. This would be cool. As it stands the study appears as just an exercise in comparative techniques.
Major Revision suggested
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
