the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Feb 2023
Declining Sea Ice and Its Relationship with Arctic Cyclones in Current and Future Climate Part I: Current Climatology in CMIP6 Models
Abstract. The Arctic climate system is changing rapidly. These large changes will have implications in the Arctic and beyond. One of the main components of the Arctic climate system are Arctic cyclones. The strong coupling between the sea ice and Arctic cyclones makes it an important topic in the warming climate. In this study, an ensemble of CMIP6 model output was utilized from 1985–2014, to determine how well the chosen models depict Arctic cyclones and their relationship with sea ice. A comprehensive climatology of Arctic cyclones and sea ice concentrations (SIC) was provided based on selected models from5 CMIP6 and the results were compared to the ERA5 product. The model results did closely match reanalysis data in depicting the observed sea ice trend. However, we found that the model results struggled to reproduce the strongly coupled relationship between the declining sea ice and Arctic cyclones. The local cyclogenesis in the Arctic was shown to be underestimated, which led to an overall underestimation of Arctic cyclones in the CMIP6 model results. The results also showed differences between model results and ERA5 with regard to cyclone intensities. The the magnitude and sign of the intensity differences varied based10 on the nominal resolution of the model, their surface roughness parametrization and cyclogenesis location.
- Preprint
(67954 KB) - Metadata XML
-
Supplement
(20 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on wcd-2023-2', Anonymous Referee #1, 29 Mar 2023
The comment was uploaded in the form of a supplement: https://wcd.copernicus.org/preprints/wcd-2023-2/wcd-2023-2-RC1-supplement.pdf
-
RC2: 'Referee Comment on wcd-2023-2', Anonymous Referee #2, 08 Apr 2023
Review notes for the manuscript to WCD: “Declining sea ice and its relationship with Arctic cyclones in current and future climate Part 1: Current climatology in CMIP6 models” by Valkonen et al.
General comments
This paper presents an Arctic cyclone and sea ice climatology for 1985 – 2014 utilizing a chosen ensemble of CMIP6 simulations. The authors further use ERA5 reanalysis data to address how closely the model succeeds in resembling characteristics of and relationships between sea ice and Arctic cyclones. The authors find that the sea ice trend is well reproduced in the models, whereas the coupling between sea ice and Arctic cyclones are less represented. Specifically, the authors find the models to struggle with representing local cyclogenesis and cyclone intensities; the former leading to an underestimation of Arctic cyclones compared to ERA5. This study tries to complete existing studies where the relationship between Arctic cyclones and sea ice is addressed, both in models and in observational data, in the current and in a changing “new” Arctic. However, in a main attempt to use CMIP6 to investigate model performance of Arctic cyclone characteristics and the relationship between cyclones and sea ice (which has been partly touched upon in previous studies as the authors point out in the introduction) and further trying to discuss causalities, in my opinion the authors fail to address this properly. This concerns me a little. The main reason to this is that the authors focus on the impact of sea ice on cyclogenesis, but never discuss the two main processes Arctic cyclones impose on the sea ice (see specific comments below; processes that can have different effect on the sea ice depending on the time scale and the local region). Thus, quantifying biases between models and ERA5 and addressing the causalities for the relationship without quantifying the two main processes lead to that a large (important) part of this paper is missing, in my opinion. The authors discuss biases they find mostly wrt the intensity metrics’ and how for example the biases in the depth of the cyclones (models overestimate) are related to the location of local cyclogenesis (underestimation of local cyclogenesis) – this is interesting and new findings, but need to be complemented with additional analysis. Additionally, even though the figures are well done, they do not always support the main findings of this paper. Therefore, I suggest a major revision before any possible acceptation of the paper. See comments below for clarifications.
Specific comments
- As the title “declining sea ice and its relationship with Arctic cyclones” already suggest a link between SIC and frequency of Arctic cyclones, me as a reader would expect this link to be more discussed in the paper. The authors multiple times mention the “one way” relationship, i.e., a potential for more local cyclogenesis and higher cyclone frequency at the marginal ice zones due more baroclinic zones following the declining sea ice (how the declining sea ice affects local cyclogenesis: discussed briefly at L240). The authors focus on the time scale between years (if I understand correctly) and claim that an increase in the cyclone frequencies is due to the declining sea ice trend (where the cyclone characteristics not well captured by models). However, there is also a two-way relationship between SIC and cyclones which the authors do not mention at all. This is, Arctic cyclones also give rise to local changes on the sea ice, which depends on the timescale, local location, cyclone intensity and local sea ice conditions. There is an attempt in addressing the role of cyclones for the Arctic climate and the relationship to sea ice changes in the paragraph starting at L32, but the main processes cyclones impose on the sea ice are missing. These two processes – thermodynamical (ice phase changes and relation to surface energy fluxes) and dynamical (wind induced) processes – show different effects on sea ice changes depending on cyclone tracks and the time scales (shorter, immediate impact or with longer timescales of weeks after a cyclone has passed). I suggest the authors to read e.g., https://doi.org/10.1029/2022GL100051 from Aue et al. 2022. There, the overall impact of cyclones on the sea ice in the Atlantic sector is discussed. The main findings are that dynamical processes dominate at the direct impact of the cyclone (sea ice loss; wind causing sea ice openings, but also thermodynamical sea ice loss in the warm sector of the cyclone), whereas thermodynamical processes dominate weeks after the cyclone has passed. These “post-” processes are mainly in the cold sector, with positive sea ice changes as the sea ice refreezes in the openings of the Barents Sea. Also, the abstract of this current manuscript is missing the most important links to the processes that cyclones impose on sea ice (when talking about relationships: “The model results did closely match reanalysis data in depicting the observed sea ice trend. However, we found that the model results struggled to reproduce the strongly coupled relationship between the declining sea ice and Arctic cyclones”). This is all fine, but proof to the second sentence is partly missing in the paper. Discussing the relationships between SIC and cyclones need to include both-way links. Thus, when these two important processes are not quantified (or even discussed in the paper), causalities are hard to address (if only thinking about the role of sea ice loss on local cyclogenesis). I suggest the authors to consider the two-way relationship and add more analysis to the current manuscript.
- The authors also compare the negative trends of SIC and the relation with cyclone frequencies. I would be happy to see more discussion about how does the changed “new Arctic” with thinner and lower sea ice concentration affect the impact of cyclones compared to the “old Arctic”? Again, a two-way relationship (SIC --> cyclones and cyclones --> SIC changes) is suggested here.
- In the introduction, I am also missing the discussion between cyclones, moisture transport and extreme surface temperatures (as discussed in e.g., in the context of drivers for warm spells in Messori et al. 2018: https://doi.org/10.1175/JCLI-D-17-0386.1), as well as present some case studies of how cyclones affect the sea ice e.g., Boisvert et al. 2016 (showing the potential of a local thermodynamical ice loss due to anomalous energy flux towards the surface, but also sea-ice retreat induced by dynamical forcing by the cyclone winds contribute to the observed sea ice loss: https://doi.org/10.1175/MWR-D-16-0234.1). I find some attempt in the manuscript to highlight the different findings between the occurrence of Arctic cyclones and sea ice changes from L69 and a case study at L73 (emphasizing the increase in SIC after the cyclone passed), but no discussion of the reasons behind or the two main processes cyclones impose on the sea ice and/or the temporal time scales considered. I suggest the authors to complete the introduction as well as the discussion parts with the missing parts.
- The authors quantitatively describe biases between models and ERA5. I was wondering if the authors know any studies where the representativity or limitation of ERA5 in depicting cyclones is shown?
- An interesting finding of this current paper is the differences they find in the model intensities compared to ERA5. I was firstly wondering if the authors could quantify which of the intensity metrices they describe are more important? If the cyclone frequencies are underestimated by models but the central pressure is also underestimated together with an overestimation of the radius (models find stronger and deeper cyclones with larger area); which one of the intensities metrices are more important (depth, frequency, area, or energy flux?). In the process of quantifying the biases, I would like to see a discussion where relative importance of each metrics is discussed. Secondly, the authors elaborate on the reasons behind the differences between metrices (e.g., deeper and larger cyclones in models), and emphasise mainly on the representation of the SLP field and the resolution of the models. I find these results quite interesting. What do the authors think about other reasons behind the differences? For example, how well are small-scale processes such as diabatic heating represented by the models (which in a warming climate could lead to a potential of a larger moisture content and thus strengthening of cyclones through PV production via diabatic heating might be even more important in the future?) See e.g., Dominic Bühler, Stephan Pfahl (2017) https://doi.org/10.1175/JAS-D-17-0041.1 about extratropical cyclones that explains processes linked to PV diagnostics and cyclone intensification. Of course, these are mainly extratropical cyclones and the current manuscript discusses Arctic cyclones – but maybe still something to consider also here? Also, other processes, such as winds from Greenland over Arctic sea ice could lead to local cyclone genesis- again processes that might not be that well captured/parameterized in models and thus cyclones are underestimated in models? I suggest authors to elaborate a bit more about the reasons behind the differences.
- Cyclone postprocessing method is for me a bit unclear (L144): the authors say that it is done following a previous study. However, not having read that paper, it is unclear for me if the postprocessing includes the sub-selection of cyclones based on their duration and spatial location or what it may include. Please rewrite these sentences for clarification.
- Sometimes the authors are mentioning twice the method of a certain parameter, e.g., the cyclone intensities on L150 and L153 – please modify and rephrase so that the message comes through. For example, rewrite on L153: “additionally to the ACE, other metrices for determining cyclone intensities were used. These metrices include cyclone central pressure, …” Also, the end of the paragraph needs to be re-written or moved to the begin of the paragraph (where the ACE is discussed).
- Regarding the ACE, for me it is a bit unclear what kind of intensity metrics it actually is. To my understanding from the manuscript, the ACE tells about the transfer of momentum and heat between the surface and the cyclone. However, it is unclear how it is calculated from the mean squared wind speeds (as given in L159). Please clarify.
- I was wondering why the NAO index was calculated by the authors and not using the daily NAO index provided by NOAA (https://www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/teleconnections.shtml)?
- Just a clarification question: are the cyclone counts provided in Fig 2 shown for the (lon,lat) points of the cyclone tracks identified by the minimum SLP, whereas the cyclone area considers a larger area (defined by the last closed isobar)?
- Suggestion: it would be nice to see cyclone frequencies over the whole northern hemisphere for the two seasons (for reference to the reader) – not only those selected to be within the Arctic region (Fig 2). Compare Fig. 4 in Wernli and Schwierz 2006 (Journal of the Atmospheric Sciences). However, the frequencies shown in Fig. 2 of this paper in review shows similar frequencies for the Arctic as their Fig. 4. If the authors decide not to include an additional panel in Fig2, I suggest that the other figure is discussed for comparison and “validation” of the authors cyclone data.
- For the readers convenience and to support the main findings of the paper (underestimation of cyclones in models compared to ERA5), I would suggest adding the difference plot in cyclone frequencies (ERA5 minus in models) for the two seasons already in the main paper. Suggest that one representative CMIP6 model is chosen.
- In section starting on L380, the relationships between SIC changes and cyclone frequencies are stated, however, no mention of correlation analysis or similar mathematical methods for defining this relationship are presented. Also, I am confused what the “time range” is referring to on the y-axis of Fig. 10. As the impact on cyclones on the SIC changes depending on the time scale considered (on lag times after a cyclone has passed), I suggest the authors to add this information as well.
Technical corrections
- E.g., at L15: “high North”: wondering if this is an appropriate notation. I assume the authors aim to say “the high Arctic”?
- Unclear what “describe the causalities between the two” in Point (3) at L91 refer to – to the relationship between Arctic cyclones and sea ice, or why we see the differences between CMIP6 models and ERA5?
- The reason for the selection of the CMIP6 models are presented in two separate sentences (L100, L105) – not sure which one is more dominant. Please rewrite / write more concise.
- The section about the “history” of ERA5 (from L117) seems a bit too much, in my opinion. If the authors wish to include this in the paper, maybe consider to mention (some) of it in the introduction. Here, I would just mention the data, resolution of the data as well as discuss its representativity and/or possible limitations.
- Reference typo on L134: (Wernli and Schwierz, 2006).
- Figure 1 caption should say “the Arctic”
- Typo on L147: “recorded” instead of “recorder”
- Abbreviation “ACE” is mentioned on L151 before being explained on L157. Please change
- Extra “;” on L177.
- Please rephrase the panel titles, e.g., in Fig. 3. This is true for almost every figure in this manuscript when the titles are given directly from the variable names in the models (it looks like it).
- On L196 I assume the given cyclone frequencies are average frequencies over the time period considered? Where are the trends in cyclone frequencies per season shown in the Figures/Tables (discussed on L200 onward)? If the authors decide to include the linear trends as lines or values, I would suggest adding them in all of the wide-basin figures.
- “Figure” missing in the brackets at L238.
- Remove the second “is” in L290.
- Number “6” on L431 a bit odd. I assume the authors refer to the Figure 6?
Citation: https://doi.org/10.5194/wcd-2023-2-RC2 - AC1: 'Author Response wcd-2023-2', Elina Valkonen, 18 May 2023
Data sets
Cyclone Catalogs | Northern Hemisphere cyclone track data Elina Valkonen, Elizabeth Cassano, and John Cassano https://arcticdata.io/catalog/view/urn%3Auuid%3A9f9a2d63-d53d-4def-bddb-324293904f9c
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 373 | 175 | 18 | 566 | 31 | 17 | 15 |
- HTML: 373
- PDF: 175
- XML: 18
- Total: 566
- Supplement: 31
- BibTeX: 17
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
