the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Assimilation sensitivity of satellite-derived surface melt into the Regional Climate Model MAR: case study over the Antarctic Peninsula
Abstract. The study of the recent variability and the future projections of the poles’ climate currently relies on polar-oriented Regional Climate Models (RCMs). However, RCMs are subject to biases and systematic errors that impact the results of their simulations. Remote Sensing (RS) data can help to reduce these ambiguities by providing indirect observations to the modeled estimates. Using the behavior of radiofrequency signals with regard to the presence of water in a snowpack, passive and active microwave instruments such as AMSR2, ASCAT, and Sentinel-1 are used to detect melt at the surface of the snowpack. In this paper, we investigate the sensitivity of the RCM “Modèle Atmosphérique Régional” (MAR) to the assimilation of surface melt occurrence estimated by RS datasets. The assimilation is performed by nudging the MAR snowpack temperature to match the observed melt state by satellite. The sensitivity is tested by modifying parameters of the assimilation: (i) the depth to which MAR snowpack is warmed up or cooled down (corresponding to the penetration depth of the satellites) to match with satellite, and (ii) the quantity of water required into the snowpack to qualify a MAR pixel as melting or not, and (iii) by assimilating multiple RS datasets. The data assimilation is performed over the Antarctic Peninsula for the 2019-2021 period. The results show an increase in the melt production (+66.7 % on average, or +95 Gt) going along with a small decrease in surface mass balance (SMB) (-4.5 % on average, or -20 Gt) for the 2019–2020 melt season. The model is sensitive to the three parameters tested but with different orders of magnitude. The sensitivity to the assimilated dataset is reduced by using multiple datasets during the assimilation and discarding the remote observations that are not coherent. For the other two parameters, the penetration depth has more impact on the assimilation than the quantity of liquid water used as melt threshold. The first one is especially sensitive for the sensors with a shorter penetration depth. In the first centimeters, a densification due to a refreeze can impact the melt production and cause an overestimation of the melt production. For the second threshold, the impact is more important on the number of melt days rather than the melt production itself. The values tested for the quantity of liquid water required into the snowpack to qualify a MAR pixel as melting or not (0.1 or 0.2 % of the snowpack mass being water) are lower than during typical melt days (~1.2 %) and impact results mainly at the beginning and end of the melt period when lower values are reached. Such an assimilation will allow an uncertainty estimation of MAR’s melt production, as well as identifying potential issues at the snowpack surface processes.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(5979 KB)
-
Supplement
(3831 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5979 KB) - Metadata XML
-
Supplement
(3831 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Modèle Atmosphérique Régional(MAR) to the assimilation of wet-snow occurrence estimated by remote sensing datasets. The assimilation is performed by nudging the MAR snowpack temperature. The data assimilation is performed over the Antarctic Peninsula for the 2019–2021 period. The results show an increase in the melt production (+66.7 %) and a decrease in surface mass balance (−4.5 %) of the model for the 2019–2020 melt season.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1371', Anonymous Referee #1, 03 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-1371/egusphere-2022-1371-RC1-supplement.pdf
- AC1: 'Reply on RC1', Thomas Dethinne, 18 Apr 2023
-
RC2: 'Comment on egusphere-2022-1371', Anonymous Referee #2, 08 Mar 2023
This study presents the results of an assimilation study using satellite-derived melt to nudge the MAR snow-temperature, to improve simulations of surface melt. The study builds on Kittel et al., 2022 that already presented this procedure, but for a different region, and adds extra satellite nudging products and discusses its sensitivity.
The paper is detailed, presents sophisticated results and its application might be important for improving contemporary simulations of surface melt and can likely be of benefit to the cryospheric community and beyond. However, it is poorly written (I would advise a native English speaker to look at the text, or a stronger contribution from the co-authors) hence several things are unclear, it is chaotically structured, and needs some serious revamping to warrant publication in TC. I have numerous technical and structural comments, but also some major criticisms that I will discuss one by one below.
I hope the authors can adress my comments, improve the readability and then I am willing to review the manuscript again after major revisions.
Major comments
- I miss the applicability of this assimilation technique for Antarctica. In Antarctica melt-rates are relatively unimportant in the contemporary climate, and for future simulations this assimilation technique obviously does not work. What will this technique provide us with? Some extra words/paragraph, either in the introduction, or in the conclusions, or both, should be spent on this to improve the relevance of this paper.
- Slightly related to point 1, I miss a recommendation based on the results of this study. Would the authors advise to use this technique on all future simulations, or is the main aim to provide better uncertainty estimates? I advise the authors to take a stronger stance on what is the main take-home message of the study.
- I miss a thorough evaluation of the actual modelled surface melt. There are several AWS on the AP that close the SEB (Jakobs, C. L. et al., (2020)) and enable a much more detailed and independent evaluation of simulated melt production. In turn, these can then be used to actually provide (a part of) the uncertainty calculation that the authors hint at in the last sentence of the abstract, which would really improve the papers conclusions and applicability (see point 1).
Minor (line by line) comments
First of all, the paper is poorly written and many things unclearly explained, and I found it really difficult to understand the methodology. I will provide some suggestions and point to some of the parts that I do not understand, but overall I advise the authors to edit the paper with an experienced English speaking (or experienced) academic writer, as I believe the inexperience in academic writing is the main cause of the problems.
Secondly, several sections should be restructured or rewritten, I will provide some suggestions below.
L2-5: unclear. Too much detailed and lengthy information for an abstract, can be considerably shortened by just writing something like the following: “However, RCMS are subject to biases, which Remote Sensing (RS) products can help solving. Here, we assimilate several satellite products that detect surface melt into the RCM MAR…” etc.
L10-11: This seems ambigiuous, are the previous two methods not also assimilations?
L14-15: Way too detailed for an abstract. Shorten
L17: A refreeze of what?
L22-23: Good to end the abstract with this (but I expect you to end the conclusions section likewise). Can you extend slightly on this?
Abstract overall: Please shorten and simplify the abstract!
L25: Here you mention both polar ice sheets, and in the following sentence you immediately move to Greenland. This transition can be improved.
L27: Here you should distinguish between grounded ice mass loss and actual mass loss, especially if later in the introduction you want to emphasize the importance of surface melt (i.e. hydrofracture and grounded ice acceleration)
L28: Why is surface melting not yet a big concern now?
L36: RCMs are not yet introduced, rephrase
L38: what do you mean with “other independent sources of uncertainties”. Vague!
L39-41: How does this assimilation technique compare to other common techniques such as reanalyses?
L42: What do you mean with sequentially? Reword
L45: what is a complex surface hydrology?
L46: isn’t it more like 10km scale?
L47: rephrase
L49: “multiple” comes out of the blue and confuses me, rephrase
L50-51: Is it, or will it be, a promising technique? Outside of Kittel 2022, there is not really any other study doing this right?
L52: Again, vague, and repeat of the previous.
Figure 1: It’s George VI, not Georges
L54-55: you already mentioned this in the beginning of the introduction.
This paragraph needs to be rewritten or completely removed; most info is repeated or obsolete. Your paper is about assimilating melt, so spend time on explaining melt and why it is important to improve melt simulations.
L67-72: This paragraph is all over the place, again repeating previously introduced information. Rephrase it and make it more concise by just writing: “Here, we assimilate different satellite observations of melt in the RCM MAR, etc”. The Methods section is to explain the actual details, pros and cons of the products.
L78: Introduce what a binary melt mask is, I did not know.
L78: Three sensors? Sensors on the satellites? The three satellites? Unclear.
L79: radiometer is a new word, should be introduced. Are all satellites equipped with microwave radiometers?
L84: this is not correct. Liquid water can't be melting. Rephrase to something like: "Here,we relate subsurface liquid water with subsurface melting". Although I am still confused how this works, how do you distinguish between percolated surface melt water and subsurface meltwater, or meltwater that has not yet refrozen after a previous melt event?
L85: I am unfamiliar with remote sensing so have no idea what you mean by acquisition capabilities
L86: specify “small scale”
L97: Rephrase to “A dry snowpack has a lower emissivity than a wet snowpack”
Equation 1: Is it epsilon^*, or is * a multiplicator? Anyhow epsilon is not defined in the text.
L108: I don’t understand, what’s “dominant melt”? rephrase
L111: Please group the three satellite production per subsection.
L143: “pixel-wise”? huh? Typo?
Figures 2-4: Is there a way to combine these in one graph, or something? They seem rather obsolete to this study (using 3 figures to show something that’s not main result of the paper).
Section 2.3.1: So, what threshold do you finally chose? This is unclear.
L235: It’s unclear for me what you are presenting here. Are you evaluation your melt assimilation simulations, or are you repeating evaluations from previous studies? It the latter, this entire paragraph is obsolete.
Section 3: The evaluation should be more detailed. AWS observations exist that are used in a SEB model so explicitly calculate melt. This can be perfectly used to evaluate the model, especially the later sensitivity experiments, and assess the models performance in simulating surface melt production. See for instance Jakobs et al., 2020.
L280: rephrase.
Results overall: Several sentences in the results section are better suited in the introduction or methods section; please increase the focus of this section on the actual results.
L295: What’s the global zone? And what evolution?
Section 4.1: This section is completely unclear to me and contains several unphysical explanations (e.g. a cold snowpack producing melt??). And, I don’t understand Figure 10. What are the curves? Not all curves are explained in the legend and as most of them overlap I also can’t distinguish them at all. Improve the figure and try to extend the caption.
L306: the penetration depth of what? Be a bit more explicit.
L311-313: Rephrase. What do you mean here?
L337-338: How is that calculated? As the melt season starts in November of the previous year?
L355- 368: I think this can all be moved to the Methods section (and in fact it already contains several things already introduced in the methods).
L372: is this the sum of three ice shelves, or the whole AP? This is unclear. Also elsewhere in the next paragraphs. Be very consistent with these numbers.
L388: uncomplete sentence.
L390-391: rewrite.
References:
Jakobs, C. L., Reijmer, C. H., Smeets, C. J. P. P., Trusel, L. D., Van De Berg, W. J., Van Den Broeke, M. R., and Van Wessem, J. M. (2020). A benchmark dataset of in situ Antarctic surface melt rates and energy balance. Journal of Glaciology, 66(256), 291–302. doi:10.1017/jog.2020.6. issn:00221430.
Citation: https://doi.org/10.5194/egusphere-2022-1371-RC2 - AC2: 'Reply on RC2', Thomas Dethinne, 18 Apr 2023
- I miss the applicability of this assimilation technique for Antarctica. In Antarctica melt-rates are relatively unimportant in the contemporary climate, and for future simulations this assimilation technique obviously does not work. What will this technique provide us with? Some extra words/paragraph, either in the introduction, or in the conclusions, or both, should be spent on this to improve the relevance of this paper.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-1371', Anonymous Referee #1, 03 Mar 2023
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-1371/egusphere-2022-1371-RC1-supplement.pdf
- AC1: 'Reply on RC1', Thomas Dethinne, 18 Apr 2023
-
RC2: 'Comment on egusphere-2022-1371', Anonymous Referee #2, 08 Mar 2023
This study presents the results of an assimilation study using satellite-derived melt to nudge the MAR snow-temperature, to improve simulations of surface melt. The study builds on Kittel et al., 2022 that already presented this procedure, but for a different region, and adds extra satellite nudging products and discusses its sensitivity.
The paper is detailed, presents sophisticated results and its application might be important for improving contemporary simulations of surface melt and can likely be of benefit to the cryospheric community and beyond. However, it is poorly written (I would advise a native English speaker to look at the text, or a stronger contribution from the co-authors) hence several things are unclear, it is chaotically structured, and needs some serious revamping to warrant publication in TC. I have numerous technical and structural comments, but also some major criticisms that I will discuss one by one below.
I hope the authors can adress my comments, improve the readability and then I am willing to review the manuscript again after major revisions.
Major comments
- I miss the applicability of this assimilation technique for Antarctica. In Antarctica melt-rates are relatively unimportant in the contemporary climate, and for future simulations this assimilation technique obviously does not work. What will this technique provide us with? Some extra words/paragraph, either in the introduction, or in the conclusions, or both, should be spent on this to improve the relevance of this paper.
- Slightly related to point 1, I miss a recommendation based on the results of this study. Would the authors advise to use this technique on all future simulations, or is the main aim to provide better uncertainty estimates? I advise the authors to take a stronger stance on what is the main take-home message of the study.
- I miss a thorough evaluation of the actual modelled surface melt. There are several AWS on the AP that close the SEB (Jakobs, C. L. et al., (2020)) and enable a much more detailed and independent evaluation of simulated melt production. In turn, these can then be used to actually provide (a part of) the uncertainty calculation that the authors hint at in the last sentence of the abstract, which would really improve the papers conclusions and applicability (see point 1).
Minor (line by line) comments
First of all, the paper is poorly written and many things unclearly explained, and I found it really difficult to understand the methodology. I will provide some suggestions and point to some of the parts that I do not understand, but overall I advise the authors to edit the paper with an experienced English speaking (or experienced) academic writer, as I believe the inexperience in academic writing is the main cause of the problems.
Secondly, several sections should be restructured or rewritten, I will provide some suggestions below.
L2-5: unclear. Too much detailed and lengthy information for an abstract, can be considerably shortened by just writing something like the following: “However, RCMS are subject to biases, which Remote Sensing (RS) products can help solving. Here, we assimilate several satellite products that detect surface melt into the RCM MAR…” etc.
L10-11: This seems ambigiuous, are the previous two methods not also assimilations?
L14-15: Way too detailed for an abstract. Shorten
L17: A refreeze of what?
L22-23: Good to end the abstract with this (but I expect you to end the conclusions section likewise). Can you extend slightly on this?
Abstract overall: Please shorten and simplify the abstract!
L25: Here you mention both polar ice sheets, and in the following sentence you immediately move to Greenland. This transition can be improved.
L27: Here you should distinguish between grounded ice mass loss and actual mass loss, especially if later in the introduction you want to emphasize the importance of surface melt (i.e. hydrofracture and grounded ice acceleration)
L28: Why is surface melting not yet a big concern now?
L36: RCMs are not yet introduced, rephrase
L38: what do you mean with “other independent sources of uncertainties”. Vague!
L39-41: How does this assimilation technique compare to other common techniques such as reanalyses?
L42: What do you mean with sequentially? Reword
L45: what is a complex surface hydrology?
L46: isn’t it more like 10km scale?
L47: rephrase
L49: “multiple” comes out of the blue and confuses me, rephrase
L50-51: Is it, or will it be, a promising technique? Outside of Kittel 2022, there is not really any other study doing this right?
L52: Again, vague, and repeat of the previous.
Figure 1: It’s George VI, not Georges
L54-55: you already mentioned this in the beginning of the introduction.
This paragraph needs to be rewritten or completely removed; most info is repeated or obsolete. Your paper is about assimilating melt, so spend time on explaining melt and why it is important to improve melt simulations.
L67-72: This paragraph is all over the place, again repeating previously introduced information. Rephrase it and make it more concise by just writing: “Here, we assimilate different satellite observations of melt in the RCM MAR, etc”. The Methods section is to explain the actual details, pros and cons of the products.
L78: Introduce what a binary melt mask is, I did not know.
L78: Three sensors? Sensors on the satellites? The three satellites? Unclear.
L79: radiometer is a new word, should be introduced. Are all satellites equipped with microwave radiometers?
L84: this is not correct. Liquid water can't be melting. Rephrase to something like: "Here,we relate subsurface liquid water with subsurface melting". Although I am still confused how this works, how do you distinguish between percolated surface melt water and subsurface meltwater, or meltwater that has not yet refrozen after a previous melt event?
L85: I am unfamiliar with remote sensing so have no idea what you mean by acquisition capabilities
L86: specify “small scale”
L97: Rephrase to “A dry snowpack has a lower emissivity than a wet snowpack”
Equation 1: Is it epsilon^*, or is * a multiplicator? Anyhow epsilon is not defined in the text.
L108: I don’t understand, what’s “dominant melt”? rephrase
L111: Please group the three satellite production per subsection.
L143: “pixel-wise”? huh? Typo?
Figures 2-4: Is there a way to combine these in one graph, or something? They seem rather obsolete to this study (using 3 figures to show something that’s not main result of the paper).
Section 2.3.1: So, what threshold do you finally chose? This is unclear.
L235: It’s unclear for me what you are presenting here. Are you evaluation your melt assimilation simulations, or are you repeating evaluations from previous studies? It the latter, this entire paragraph is obsolete.
Section 3: The evaluation should be more detailed. AWS observations exist that are used in a SEB model so explicitly calculate melt. This can be perfectly used to evaluate the model, especially the later sensitivity experiments, and assess the models performance in simulating surface melt production. See for instance Jakobs et al., 2020.
L280: rephrase.
Results overall: Several sentences in the results section are better suited in the introduction or methods section; please increase the focus of this section on the actual results.
L295: What’s the global zone? And what evolution?
Section 4.1: This section is completely unclear to me and contains several unphysical explanations (e.g. a cold snowpack producing melt??). And, I don’t understand Figure 10. What are the curves? Not all curves are explained in the legend and as most of them overlap I also can’t distinguish them at all. Improve the figure and try to extend the caption.
L306: the penetration depth of what? Be a bit more explicit.
L311-313: Rephrase. What do you mean here?
L337-338: How is that calculated? As the melt season starts in November of the previous year?
L355- 368: I think this can all be moved to the Methods section (and in fact it already contains several things already introduced in the methods).
L372: is this the sum of three ice shelves, or the whole AP? This is unclear. Also elsewhere in the next paragraphs. Be very consistent with these numbers.
L388: uncomplete sentence.
L390-391: rewrite.
References:
Jakobs, C. L., Reijmer, C. H., Smeets, C. J. P. P., Trusel, L. D., Van De Berg, W. J., Van Den Broeke, M. R., and Van Wessem, J. M. (2020). A benchmark dataset of in situ Antarctic surface melt rates and energy balance. Journal of Glaciology, 66(256), 291–302. doi:10.1017/jog.2020.6. issn:00221430.
Citation: https://doi.org/10.5194/egusphere-2022-1371-RC2 - AC2: 'Reply on RC2', Thomas Dethinne, 18 Apr 2023
- I miss the applicability of this assimilation technique for Antarctica. In Antarctica melt-rates are relatively unimportant in the contemporary climate, and for future simulations this assimilation technique obviously does not work. What will this technique provide us with? Some extra words/paragraph, either in the introduction, or in the conclusions, or both, should be spent on this to improve the relevance of this paper.
Peer review completion
Journal article(s) based on this preprint
Modèle Atmosphérique Régional(MAR) to the assimilation of wet-snow occurrence estimated by remote sensing datasets. The assimilation is performed by nudging the MAR snowpack temperature. The data assimilation is performed over the Antarctic Peninsula for the 2019–2021 period. The results show an increase in the melt production (+66.7 %) and a decrease in surface mass balance (−4.5 %) of the model for the 2019–2020 melt season.
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
447 | 178 | 18 | 643 | 45 | 13 | 6 |
- HTML: 447
- PDF: 178
- XML: 18
- Total: 643
- Supplement: 45
- BibTeX: 13
- EndNote: 6
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
Quentin Glaude
Ghislain Picard
Christoph Kittel
Anne Orban
Xavier Fettweis
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5979 KB) - Metadata XML
-
Supplement
(3831 KB) - BibTeX
- EndNote
- Final revised paper