the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Version 8 IMK/IAA MIPAS measurements of CFC-11, CFC-12, and HCFC-22
Gabriele P. Stiller
Thomas von Clarmann
Norbert Glatthor
Udo Grabowski
Sylvia Kellmann
Michael Kiefer
Alexandra Laeng
Andrea Linden
Bernd Funke
Maya Garcia-Comas
Manuel Lopez-Puertas
Abstract. The Michelson Interferometer for Passive Atmospheric Sounding (MIPAS) on Envisat provided infrared limb emission spectra, which were used to infer global distributions of CFC-11, CFC-12, and HCFC-22. Spectra were analysed using constrained non-linear least squares fitting. Changes with respect to earlier data versions refer to the use of version 8 spectra, the altitude range where the background continuum is considered, details of the regularisation and microwindow selection, and the occasional joint-fitting of interfering species, new spectroscopic data, the joint-fit of a tangent-height dependent spectral offset, and the use of 2D temperature fields. In the lower stratosphere the error budget is dominated by uncertainties in spectroscopic data, while above measurement noise is the leading error source. The vertical resolution of CFC-11 and CFC-12 is 2–3 km near the tropopause, about 4 km at 30 km altitude and 6–10 km at 50 km. The vertical resolution of HCFC-22 is somewhat coarser, 3–4 km at the tropopause and 10–12 km at 35 km altitude. In the altitude range of interest, the horizontal resolution is typically limited by the horizontal sampling of the measurements, not by the smearing of the retrieval. Horizontal information displacement does not exceed 150 km, which can become an issue only for comparisons with model simulations with high horizontal resolution or localised in-situ observations. Along with the regular data product, an alternative representation of the data on a coarser vertical grid is offered. These data can be used without consideration of the averaging kernels. The new data version provides improvement with respect to reduction of biases and improved consistency between the full and reduced resolution mission period of MIPAS.
- Preprint
(2448 KB) - Metadata XML
-
Supplement
(4760 KB) - BibTeX
- EndNote
Gabriele P. Stiller et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2023-172', Anonymous Referee #1, 01 Sep 2023
General CommentsThe manuscript provides an overview of the new Version 8 MIPAS retrievals of CFC-11, CFC-12, and HCFC-22. The paper is very detailed, but I found it easy to read and that it provides excellent information both for readers interested in the retrieval process and potential users of the data. A comprehensive error analysis is also provided. In my opinion the manuscript is very useful to the community and can be published almost as is. I have provided some of my thoughts while reading through it that the authors can choose to consider if they wish.Specific Commentsp. 5: Figure 1: I find this figure a little difficult to follow. Clearly the retrieval involves a complicated flow of data. I don't know a better way of presenting it, but maybe the authors can think of something.Sec. 3.1: Is the 2D temperature field used here? I see it mentioned in the CFC-11 section but not here.p.6 l.140: "CFC-12 is retrieved on a vertical grid as follow:...":I see altitude here, presumably the altitude-pressure relationship is taken from a preceeding retrieval step? I don't see it mentioned unless I missed it.p.6 l.142: "This implies that the involved inverse problem is ill-posed...":I understand what the authors mean, and it is certainly true if a single spectral point was used, but the use of multiple spectral points can result in a problem that is well-posed with a spacing finer than the measurement spacing.Sec. 3.4: Presumably here "nominal geolocation of the limb scan" is calculated based off of the ~30 km tangent point? If you were to calculate horizontal displacement relative to the geolocation of each individual's spectra tangent point is it essentially 0? I'm mostly wondering if it would be better to compare the MIPAS measurements to models by sampling the model at each spectra's tangent location rather than the nominal location.Sec. 3.6: My understanding is that essentially another iteration of the retrieval is performed with an adjusted grid to get the ML profile, I assume this is only done for CFC-12 and not for any of the input data? Can the change in averaging kernel of temperature for example have an influence on this coarse grid product?Sec. 4.5: Can you speculate on which changes in the retrieval had the largest impact on the differences?Citation: https://doi.org/
10.5194/amt-2023-172-RC1 -
AC1: 'Reply on RC1', Gabriele Stiller, 13 Oct 2023
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2023-172/amt-2023-172-AC1-supplement.pdf
-
AC1: 'Reply on RC1', Gabriele Stiller, 13 Oct 2023
-
RC2: 'Comment on amt-2023-172', Anonymous Referee #2, 11 Sep 2023
General comments
The manuscript of Stiller et al. (2023) describes the new Version 8 (V8) of MIPAS CFC-11, CFC-12 and HCFC-22 measurements. The changes performed in the retrieval processes between the previous version (V5) and V8 are very well detailed and justified, and the improvements observed in the resulting data sets are substantial. The data are well characterized by their averaging kernels and uncertainties, and furthermore the authors introduce a new representation of their products on a coarser grid which will be very useful for the users.
Therefore, providing a description of these new V8 data sets is important for the scientific community and well in the scope of AMT. I recommend the publication in AMT, and I list some questions and suggestions below that might help clarifying a few points.
Specific comments:
- p.3 Section 2 Description of MIPAS modes: It was not clear to me why while the NOM modes have two versions (FR and RR before and after Jan 2005 respectively), the MA has only one mode. I had to find some MIPAS documentation to see that the MA only started in Jan 2005 (right ?), so maybe put this information in the text here, for MIPAS non-expert.
- p.4 CFC-12, l.96: “v5 CFC-12 might have a slight low bias”; and “Zhou et al. found good agreement”: Can you be more specific and give numbers? To check that v8 is an improvement, ideally one would like to see some validation of v8 in this manuscript. If not possible but the authors (unless you could consider it?), providing numbers (and sign of the bias) could help to check more quantitatively that there will be an improved agreement with the new version.
- p. 5, figure1: I find it hard to follow. Maybe you could use different colors for the arrows concerning the 3 different species, or consider making 3 different schemes?
- p.10-14: Fig. 2 to 7: I would merge these 6 figures into 2 figures (Fig. 2 to 4; and Fig. 5 to 7). In Figs. 5-7, the legend describing what is the black lines/points is missing. Of course, the same remarks for the other species.
- Sect 3.4: Table 4: Maybe a Figure as in von Clarmann 2009a would be nicer than numbers in a Table? Unless the authors think the exact numbers would be used by the community? They can decide. Same remark for Table 8 and 12.
- p. 13 l.182-183: “The horizontal smearing…is always narrower than the… horizontal distance between the nominal geolocations of two subsequent limb scans”: can you give this horizontal distance in the text?
- p. 13, Sect. 3.5:
1) The authors describe the effect of new version vs v5 in term of profiles (biases, avoided negative vmr). How the uncertainty and the vertical and horizontal smoothings change from v5 to v8? Are they also improved or similar? Same question for the other species.
2) The authors give in the text some numbers for systematic differences between the 2 versions. Would it be possible to give some visualization of these changes? Some maps v5 vs v8 or a map of differences? It could be in a Supplement if the authors do not want to increase too much the size of the paper. Same remark for the other species.
3) Is V8 closer to other reference measurements (does the bias given in this section go in the good direction with similar values than observed bias in the literature?) What is the impact of the new spectroscopy for CFC-12? Since it is the dominant systematic source of uncertainty, does it affect the bias here? Or is it only the regularization that matters more for this species?
4) I find it hard to see an improvement in the step between FR and RR by simply looking at Fig 9 and at Fig 12 of Kellmann et al. (2012). Because this is an important improvement for users and a strong motivation to change from v5 to v8, would it be possible to show this result more clearly? E.g. some examples of time-series with an automatic step detection and some numbers given for the bias before / after Jan 2005 for the v5 and v8? Same remark for CFC-11 (p.25,l. 185).
- p.17, Fig. 10: could be merged with Figs. 17 and 24.
- p. 18, l. 222-227: Similar remark as above for CFC-12 p.4: the authors should give numbers of the observed bias in v5. And/or perform new validation here using reference measurements.
- p. 24, l.268-269: use FR, HR and MA as elsewhere (instead of “high resolution measurements,…) ?
- p. 25, Sect.4.5:
1) Same remark as for CFC-12: hard to visualize how much the data set is improved without additional maps (e.g. the reference of Fig.8 is given, but we don’t have the same fig. for v5 in order to compare).
2) Same remark for the step reduction between FR and RR as above (hard to see only by looking at Kellmann et al., 2012).
3) What is the part of the improvement that is due to the change of spectroscopy? What is the main driver of the improvement in the bias between FR and RR?
-p. 36, Sect. 5.5:
1) The maximum in the upper troposphere due to the old spectroscopy is also present in ACE-FTS and ground-based data using the same spectroscopy ?
2) Again, hard to see the improvement without appropriate v5 map/figure.
- p. 37, Conclusions, l. 396-397: “Several implausible features in the V5 data…. Have disappeared or… reduced”. As said above for the 3 individual results sections, I think that giving concrete examples of these features present in v5 in figures or maps, then with the same examples “resolved” using v8, would help the visualization of the improvements.
Citation: https://doi.org/10.5194/amt-2023-172-RC2 -
AC2: 'Reply on RC2', Gabriele Stiller, 13 Oct 2023
The comment was uploaded in the form of a supplement: https://amt.copernicus.org/preprints/amt-2023-172/amt-2023-172-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Gabriele Stiller, 13 Oct 2023
Gabriele P. Stiller et al.
Gabriele P. Stiller et al.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
244 | 95 | 24 | 363 | 19 | 14 | 13 |
- HTML: 244
- PDF: 95
- XML: 24
- Total: 363
- Supplement: 19
- BibTeX: 14
- EndNote: 13
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1