the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Relevance of warm air intrusions for Arctic satellite sea ice climatologies
Philip Rostosky
Gunnar Spreen
Abstract. Winter warm air intrusions entering the Arctic ocean can strongly modify the microwave emission of the snow/ice system due to temperature induced snow metamorphism and ice crust formations e.g., after melt-refreeze events. Common microwave radiometer satellite sea ice concentration retrievals are based on empirical models using the snow/ice emissivity and thus can be influenced by strong warm air intrusions. Here, we carry out a long-term study analyzing 41 years of sea ice concentration observations from different algorithms to investigate the impact of warming events on the retrieved ice concentration. Our results show that three out of four analyzed sea ice concentration retrievals underestimate the sea ice concentration during warm air intrusions which increase the 2 m air temperature above -5 °C.
This can lead to sea ice area underestimations in the order of 104 to 105 km2. If the 2 m temperature during the warm air intrusions cross -2 °C, all retrieval methods are impacted. Our analysis shows that the strength of these strong warm air intrusions increased in recent years, especially in April. Within the scope of future climate change, it is expected that such warm air intrusions will occur even more frequent and also earlier in the season and thus the influence of these warm air intrusions on sea ice climatologies will become more important in future.
Philip Rostosky and Gunnar Spreen
Status: open (until 03 Jul 2023)
-
RC1: 'Comment on tc-2023-69', Anonymous Referee #1, 31 May 2023
reply
Relevance of warm air intrusions for Arctic satellite sea ice climatologies
Rostosky and Spreen
This paper investigates the impact of Arctic warm air intrusions (WAIs) on four passive microwave sea ice concentration products. To complete this analysis, the authors developed a methodology to detect WAI events from ERA-5 reanalysis temperatures and classify their intensity into three categories. Results show that the most extreme warm air intrusions reduce sea ice concentrations in most data products leading to an underestimation of sea ice area within the affected area of 2-4%. Further, the authors demonstrate a non-significant (due to high interannual variability) increase in the frequency of WAI events with a peak occurrence of the most severe events in April.
I think this paper makes a good contribution to the literature and makes the case that these WAI events have a significant impact on passive microwave sea ice concentration retrievals, the effects of which are expected to increase due to climate change. The analysis is thorough and appropriate for publication in The Cryosphere. I do think the paper would be improved with minor revisions to clarify some points of the methodology as I describe in my comments below.
General Comments
“… climate data records (CDR) and provide a stable time series of more than 40 years.” (L68): Other SIC data sets that are not CDRs also provide stable, long time series. The big thing that differentiates the NSIDC CDR from, for example, NASA-generated SIC products is that as a CDR, the NSIDC product does not involve any manual corrections that cannot be reproduced exactly by the code. This is the important distinction that defines how a CDR is different from any other long data time series. For this study, I don’t think that the distinction between CDRs and other time series of SIC is as important as the text would indicate.
“… the NSIDC CDR benefits from including the Bootstrap algorithm…” (L324-325): I was waiting for this point to be stated throughout reading the whole paper, but it does not appear until the very last paragraph and is only one sentence. I think a slightly longer explanation for the reason why the NSIDC product is performing better than the other is necessary for the readers especially since the NASA Team algorithm (from within the NSIDC CDR) performs very differently than the NSIDC CDR throughout your analysis. Specifically explain (1) that the NSIDC CDR SICs are primarily sourced from the Bootstrap algorithm during the WAIs and (2) some details on how the Bootstrap algorithm avoids the extreme sensitivity to the WAI events seen in the other algorithms (e.g., daily dynamic tie points, etc.). Point 2 would explain the “why” for point 1. This discussion should be moved to Section 5, and not be introduced in the conclusions.
I suggest that the authors consider moving the algorithm details from appendix A1 and A2 to the methodology (section 3). As it is now, neither section 3 nor the appendices are complete descriptions of the method, and some information is repeated in both places.
Specific Comments
L99: Here you state that you use the daily maximum 2m air temperature, however, throughout the rest of the paper you only refer to 2m air temperatures. These are not the same. I suggest changing your wording when mentioning 2m air temperatures to clarify that it is the daily max 2m air temperatures instead.
Figure 2 and L147-157: There is ~10 days difference between the max loss days between the NSIDC and other three algorithms. Why is this the case when the peak temperature is 19 April? The max loss dates from case 1 (Figure 1) are much closer (only 1 day difference). Can you add some commentary to this paragraph explaining why case 2 has a much longer time difference in max loss days?
Comments on Appendix A1: (L347) How many days are between the “new and previous warm air intrusions”? If I understand the procedure correctly (L346-349), there can only be one WAI event detected in any 5-day period? Is that correct? If that is not the case, I need more clarification as the schematic in Figure A1 is not very detailed. Why is a 65% SIC threshold (L343) chosen for the ice edge mask. Is it arbitrary or based on some other knowledge? How often does the SIC need to drop below 50% to be masked as a polynya (or how many times is “frequently”? L344)?
Comments on Appendix A2: (L364) I assume you are accumulating the area representative of the daily SIC difference from the background SIC over the duration of the WAI event, but you don’t specifically state how you compute the effective area loss shown here. Please state specifically how you compute effective area loss. (L369) Is there an average length of WAI events? How was 10 days after the peak warming day chosen? I think revising such that the appendices and the methods are in one section can help with clarifying the above questions.
Figure A3: What determines the black portions of the SIC curve where the SIC is below the background value but not included in the WAI event? How is that defined? It’s not explained in text. Also, convert the temperature scale to °C to be consistent with the text and other figures.
Technical Corrections
L20: expand the ASI acronym
L33: Warm air intrusions enter the Arctic region, not the Arctic Ocean.
L54: typo – snow/ice
L58: typo – influences
L60: The punctuation around the references to papers and figures here is confusing. Please revise.
L85: typo – daily gridding
L105-107: This sentence is important, but confusing as written. Please revise to clarify.
L264: extent should be extend
L289: typo – Another
L309: revise, “…Arctic amplification have these warm air intrusions increased…”
Figure 3: Can you make the area notation consistent with the rest of the paper (e.g., 10^4 km^2, etc.)
Citation: https://doi.org/10.5194/tc-2023-69-RC1
Philip Rostosky and Gunnar Spreen
Philip Rostosky and Gunnar Spreen
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
123 | 47 | 8 | 178 | 4 | 3 |
- HTML: 123
- PDF: 47
- XML: 8
- Total: 178
- BibTeX: 4
- EndNote: 3
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1