The authors have improved their manuscript since the previous submission and some of the open questions have been addressed. But the discussion in several of the sections remain confusing. I strongly encourage the authors to review the manuscript again for clarity to ensure that important points are stated in simple language at the appropriate locations in the text. In several cases it only became clear what the authors meant when I read discussions later in the manuscript. I have identified a few of these instances (see below), but my review in this respect is not complete. I feel that a few sentences early in this paper outlining the analysis approach will help the reader to connect the content in the various sections.
Some of my comments about the earlier version of this paper related to conclusions about the stray light characteristics that I felt were inadequately substantiated. In this revised version it seems as though the authors have the evidence at their fingertips to demonstrate they are accurately describing the internally scattered stray light characteristics of TropOMI. But actual evidence in this latest version of the manuscript is still lacking. The observed MgII index and the stray light region contents are two independent pieces of information about internally scattered stray light. The instrument parameter that connects the two is the stray light kernel. There are many potential stray light kernels that could equally explain one or the other of the observed phenomena. The authors will have a persuasive argument regarding their chosen SL kernel changes if they can demonstrate it explains both the MgII index and stray light row changes simultaneously. Perhaps the authors believe they have accomplished this, but I do not see the clear "2+2=4" arguments in the text. It may be that only minor adjustments are needed to achieve this objective.
There should also be a discussion of the problems remaining in this new product. If the authors believe the resulting Level 1 product is not suitable for ozone trend analysis, they should state so and explain why. Don't leave it to the more careful readers to figure this out for themselves.
Line 125-129
The flow of the discussion here is awkward. I find myself reading a rereading the sentences trying understand the authors' points. A clearer approach would be to first discuss all the potential explanations for the behavior of the MgII line depth, including solar irradiance change, then eliminate the ones that are not plausible. Walk the reader through the logic rather than hiding the key points in clauses at the end of sentences.
Line 143
What are the authors intending to say in this sentence? Addressed by whom or by what? The correction algorithm? This paper? The detector? It is unclear what this sentence means in the context of the preceding discussion.
Line 165
I recommend that the authors find a term other than "far-field" to describe Band 1 stray light originating from Band 2. Far-field or out-of-field refers to the field angle, which is a spatial dimension. Its use in describing the spectral dimension of stray light is confusing. Please use "out-of-band stray light" or more generically "spectral stray light" to describe contributions in the spectral dimension (also in Line 330).
Lines 170-175
The authors discuss the motivation for altering the stray light kernels: non-zero signal content in the stray light region after SL correction, and over-correction in Band 1. The authors provide detailed explanations for the stray light kernel assessment in Appendix B, but only for the spatial kernel adjustment. The stray light region contents are far more sensitive to the shape of the spatial kernel than to the shape of the spectral kernel, and therefore the latter cannot be tuned on the basis of the stray light region data. Since spectral stray light should be the main focus of this paper (it is the primary cause of their ozone profile retrieval errors), it would be best if the authors clearly and explicitly present the methodology for modifying the spectral axis shape of the kernel.
Section 2.2
The authors fail to describe how the dynamic SL correction is actually implemented. This is not an irrelevant detail because there are several possible approaches and not all will equally explain the observed measurements. Is the time-dependent behavior based on the MgII lines or is it based on the stray light region contents? Is the kernel modified by broadening its width or are the tails merely increased by adding a uniform background to the kernel? Are the spectral and spatial kernels adjusted identically or independently. These choices will significantly affect the residual signals.
Section 2.3
The authors explain that there are signal errors remaining even after the standard signal corrections. They describe the correction as a CKD image, which requires a bit more explanation for readers (or eliminate it if it provides no useful information). It is difficult to understand from the discussion exactly what artifacts are being corrected; how are they identified and how are they corrected. What is the magnitude of the correction? It is only in the discussion surrounding Figure 9 later in the paper that the reader discovers these additional signal corrections have little effect on the results. The authors should say so in this section. As currently written, the reader is left wondering what these errors are and the role they play in the observed TropOMI radiometric behavior.
Lines 201-204
The before/after comparison of the MgII index and Ca K line index shown in Figure A2 suggests the post-correction signal variation at 280 nm may be close to true solar variation, but it is hardly convincing evidence. It would be better if the authors could directly compare the post-correction MgII index observed by TropOMI with a predicted index (adjusted for the TropOMI bandwidth) based on an external source . Furthermore, this evidence and its discussion should not be relegated to an appendix. In my opinion, it is central to the stray light discussion and the authors' assertion that TropOMI stray light characteristics are changing in orbit.
Section 3
In this section the authors describe an approach to radiometric corrections that is rather difficult to follow. It lacks a high-level discussion of the authors' motivation for choosing this soft calibration approach. For example, it appears that the soft calibration of TropOMI is based on ozone climatology, not just initially but also throughout the mission. I presume this means the authors have given up on producing radiances suitable for trend-quality ozone retrievals. If this is the case, the authors should say so clearly and unequivocally here and in the abstract. If it is not the case, the authors need a clearer discussion in this section to explain how radiance trends are preserved.
The soft calibration approach also addresses an apparent non-linear response of the instrument. Is this an existing operational correction that you are merely updating to be more accurate, or is this an entirely new correction? Though this may have been previously described in another publication, these unusual corrections need some high-level description and justification. Figures 7 and D2 indicate there is a residual non-linearity in the instrument system. What is the cause of this non-linearity? Detector? Stray light overcorrection? Incorrect wavelength registration? Something else? How this non-linearity is addressed depends a lot on its cause. For example, a detector non-linearity will not have a wavelength dependence, nor does it typically have much time-dependence. Stray light errors are not addressed well by broad addition or subtraction of signal because they depend on neighboring pixel signals. Indeed, a comparison of Figures 9 and 10 suggests that some of the non-linearity was caused by stray light, but much of the non-linearity remains between versions. Even if the authors do not know the exact cause, they should be able to speculate how these errors might arise. The statements by the authors in lines 309-312 that such discussion is outside the scope of this study do not absolve the authors of their responsibility to convince the readers they are addressing the data problems in an appropriate manner.
Section 3.4 and Figure 8
It is clear, upon reading Section 4, that the soft calibration presented here applies to the current operational processing rather than the proposed future operation processing. This point should be stated more clearly. Perhaps the authors should refer to the actual product version numbers rather than "operational" and "updated" to indicate old and new.
Lines 314-317
This is a confusing paragraph. Exactly which corrections were applied in the reprocessing? Based on the subsequent text in this section and the contents of Figure 9, it seems that the authors are trying to say that they are recomputing the soft cal. correction for different combinations of instrument corrections applied in the L1B processing. If this is what the authors mean, just state so in plain and simple language. What does the sentence "The soft calibration procedure is instead kept the same ..." ? Instead of what?
Lines 330
The authors refer to a spectral cut off of the stray light convolution kernel. If such a cutoff was described in Section 2.2.1 it was not done so in a clear manner that allows the reader to identify what feature of the kernel the authors are referring to here. Are the authors saying the narrower kernel reduces the contributions to Band 1 from Band 2? If so, please say that instead of calling it a cut off.
Lines 339
Please identify the quantity being discussed. The time difference between two orbits is a constant and cannot become smaller, so the authors must be referring to something else. Correction magnitude?
Line 370
The sentence beginning "We do not observe ..." is awkward. I do not understand what the authors are trying to say. Perhaps removing the word "instead" will help. I've noticed several places throughout the manuscript where this word has been unnecessarily introduced.
Lines 372-373
The authors here conclude on the basis of Figures 12b and 12d that there is a reduction in anomalies. This point is not at all clear from the figures. By some criteria it could be argued the new version actually looks worse. I recommend revising the figures to more clearly support your conclusion. Or perhaps the authors just need to revise their tropospheric ozone conclusions.
Lines 385-387
Please make it clear that this statement refers to the current operational product.
Grammatical
Line 127: "a part for the spectral region"
Line 134: "part of to the instrument"
Line 238: "and the ISRF also updated"
Line 264: "approach as described in (?)"
Line 331: "between the orbits is reduced of around"
Line 350: "slightly increases of around 0.2% .... it decreases of 1-2%" See also Line 365, 394.
Line 374: "as it can be seen from"
Line 375: "they decrease of few percents"
Line 388: "updated with adjustments regarding" |
Attached