the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 01 Apr 2026
Antarctic Sea Ice Variations and Their Linkages with Global Extreme Weather Events
Abstract. Antarctic sea ice is a critical indicator of global climate dynamics, yet its post-2015 accelerated retreat and links to extreme weather/key climate modes remain insufficiently characterized. To address these gaps, we analyse 2010–2024 multi-source data (NSIDC sea ice extent (SIE), CMA extreme weather records, NOAA ENSO indices) via spatiotemporal decomposition, non-linear modelling, and mechanistic dissection. Antarctic SIE showed a "stable-then-decline" trend: minimal variability 2010–2014 (peak: 20.16 × 10⁶ km², Sep 2014), followed by unprecedented post-2015 retreat to a Feb 2023 historic minimum (1.85 × 10⁶ km², ~20 % below 2010–2014 summer mean). The Amundsen–Bellingshausen Sea and Antarctic Peninsula were main retreat zones, with 2020–2024 declines (10–15 %/20–25 % winter/summer) exceeding East Antarctica (5–10 %) and Ross Sea (8–12 %). We identified tiered negative covariance between SIE and extreme weather, presumably co-modulated by the El Niño-Southern Oscillation (ENSO) and Southern Annular Mode (SAM): strong associations (R² ≥ 0.6, P < 0.001) for extreme heat (400 % frequency increase) and cold waves (700 % increase), potentially via albedo–circulation feedback; moderate associations (0.5 ≤ R² < 0.6, 0.001 ≤ P < 0.01) for floods/rainstorm/typhoons, likely from ACC heat transport changes and convective propagation; weak associations (R² < 0.4, P ≈ 0.05) for blizzards, possibly due to spatially constrained El Niño–SAM effects. La Niña-positive SAM may have amplified these linkages (e.g., 400 % extreme heat increase in 2021–2023 triple La Niña), while El Niño-negative SAM suppressed them. These findings advance polar-low latitude coupling understanding, aiding extreme weather prediction and IPCC AR6-aligned adaptation.
- Preprint
(3423 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-6301', Anonymous Referee #1, 28 May 2026
-
AC1: 'Reply on RC1', Tang Shulin, 19 Jul 2026
Author Response to Anonymous Referee #1
Manuscript: egusphere-2025-6301
Revised title: Antarctic Sea-Ice Anomalies and Global Hazard-Event Records: An Exploratory Analysis, 2010–2024
Dear Anonymous Referee #1,
We sincerely thank you for your careful, constructive, and encouraging assessment of our manuscript. Your comments identified the need to remove causal overstatement, address the limitations of the short observational record, strengthen the statistical safeguards, and recast the proposed pathways as conceptual rather than demonstrated causal mechanisms. We have therefore undertaken a substantial rewrite rather than a limited textual revision.
The revised manuscript has been reframed as an exploratory observational analysis of statistical co-variation. The title, abstract, introduction, methods, results, conceptual framework, discussion, and conclusions have all been rewritten. The revised line numbers cited below refer to the clean revised manuscript. Our point-by-point responses follow the summary of major revisions.
Major revisions
• Causal interpretation removed. The manuscript now consistently uses terms such as statistical association, co-variation, synchronization and conceptual pathway, and explicitly states that the results do not establish direct causal attribution.
• Statistical analysis strengthened. The exponential curve is retained only as a descriptive comparison. Linear fitting is now the primary model-form check, supplemented by Pearson correlation, Spearman rank correlation, detrended-residual correlation, leave-one-out sensitivity analysis and Benjamini–Hochberg false-discovery-rate adjustment.
• Data object and scope clarified. The China National Climate Center dataset is described as a record of major hazardous weather and climate events, not a complete objective census of all global extremes. Regional SIC maps are used only as contextual illustrations, whereas basin-wide annual mean SIE is the statistical predictor.
• ENSO and SAM roles redefined. ENSO is treated as a first-order quantified climate background. SAM is discussed only as a literature-supported potential co-modulating factor because it is not independently included in the quantitative tests.
• Mechanistic discussion rebuilt. The former causal mechanism narrative and unsupported numerical claims have been removed. The new framework separates plausible physical pathways, event-level statistical association and common modulation, and explicitly identifies the analyses needed for future attribution.
Point-by-Point Response to Anonymous Referee #1
R1.1 Overall interpretation, causality and confounding
Reviewer comment: The paper is well written and the experimental design is relatively clear, but the claims are overstated. The analyses are not sufficient to infer causality, particularly given the short temporal record and the absence of a more comprehensive treatment of confounding climate drivers.
Response: We agree. The central interpretation has been changed from a causal or regulatory claim to an exploratory assessment of co-variation. The revised title now explicitly identifies the study as an “Exploratory Analysis”. The abstract states that the results indicate statistical co-variation rather than direct causal attribution. The Introduction now distinguishes basin-wide SIE metrics from regional processes, describes the global-event question as unresolved, and states that current evidence does not demonstrate that Antarctic sea-ice retreat independently drives multiple types of global hazardous events. The Methods and Conclusions repeat these limits. We also acknowledge the short record, shared trends, individual anomalous years, event-recording limitations, ENSO and potential SAM co-modulation as sources of uncertainty.
Revised manuscript: Title and Abstract, Lines 1–24; Introduction, Lines 40–71; Methods, Lines 98–139; Conclusions, Lines 470–477.
R1.2 Additional statistical safeguards
Reviewer comment: The short temporal record and incomplete treatment of confounding climate drivers make the regression results insufficient for strong conclusions.
Response: To reduce the risk of overinterpretation, we added several robustness checks. The revised analysis now reports linear and exponential fits, Pearson and Spearman correlations, correlations between separately detrended residuals, leave-one-out sensitivity ranges and FDR-adjusted P values. These tests do not resolve causal identification, but they help distinguish associations that persist across model forms from results that are strongly influenced by common trends or individual years. The revised text explicitly treats these statistics as supporting evidence for robustness, not as evidence of causality. ENSO is used only to characterize a first-order quantified climate background, whereas SAM is not included in the quantitative regression or partial-correlation tests; neither should therefore be interpreted as having been causally controlled.
Revised manuscript: Methods, Lines 125–139; Results and Table 1, Lines 193–226; ENSO/SAM qualifications, Lines 62–71 and 290–306.
R1.3 Conceptual rather than causal pathways
Reviewer comment: The causal language should be softened, the limitations of the correlation pathways should be acknowledged, and the pathways should be reframed as conceptual. The paper should make clear that it identifies patterns consistent with teleconnection mechanisms rather than proving direct causality.
Response: We have fully adopted this recommendation. Section 3.3 is now titled “Conceptual pathways linking Antarctic sea-ice anomalies and recorded event frequency”. The framework separates four elements: Antarctic sea-ice anomalies, plausible physical pathways, event-level statistical association and common modulation. Figure 8 explicitly states that the stronger/intermediate/weaker categories denote only statistical association strength and do not imply direct causality. The text further states that the framework is intended for future regional diagnostics, lagged tests and numerical modelling, not as a completed attribution result.
Revised manuscript: Section 3.3, Lines 288–436; framework interpretation, Lines 307–323; Figure 8 caption, Lines 357–363.
R1.4 Recommendation to reframe and resubmit
Reviewer comment: If the paper can be turned into a more cautious, exploratory and correlation-based synthesis paper, revision and resubmission or major revision could be considered.
Response: The manuscript has been reconstructed in precisely this direction. It is now presented as an exploratory observational analysis of recorded major-event frequencies, with explicit separation between statistical results, conceptual interpretation and unresolved causal attribution. Prediction and attribution claims in the original abstract and conclusions have been removed.
Revised manuscript: Throughout the revised manuscript, particularly Lines 7–24, 54–79, 120–139, 288–323 and 437–477.
We thank you again for your constructive assessment. Your comments led us to reframe the study, strengthen the statistical evaluation, and state the inferential limits more clearly. We respectfully submit the revised manuscript for renewed consideration.
Sincerely, Shulin Tang and Yaozong Zhang
-
AC1: 'Reply on RC1', Tang Shulin, 19 Jul 2026
-
RC2: 'Comment on egusphere-2025-6301', Anonymous Referee #2, 09 Jun 2026
GENERAL COMMENTS
This study looks at the linkages between Antarctic sea ice extent and a wide range of extreme events globally. The main results as stated in the abstract are 1) a summary of regional changes in sea ice 2) statistical relationships between Antarctic sea ice and global extremes and 3) the modulation of this relationship by ENSO and SAM. There exist robust physical arguments why certain extremes may be modulated by Antarctic SIE - for example, local storms due to heat fluxes, or conditions in Southern Hemisphere continents due to modulation of the jet stream- and why SIE might modulate global temperature via the albedo feedback effect, and these form the foundation of hypotheses in the paper. The extent to which there is evidence for these linkages is a matter of great topical interest.
However, it is a more extraordinary claim that the number of global (rather than, say, regional or southern hemisphere) extremes might be modulated to any meaningful extent by Antarctic SIE, and the evidence presented in this paper is far insufficient to support this claim. My main concerns about the analysis and conclusions are:
- The analysis (Section 3.2) is correlations between global extremes indices and SIE using an exponential regression model. There are only 14 years of data and the justification for the exponential regression model is weak; I agree relationships would not be linear necessarily, but why exponential? This is an insufficient foundation for the claims made.
- Secondly, even if the significance of this regression model is being treated appropriately, correlation is not causation. Statements such as ‘the reduction in SIE is likely a significant driver’ are used frequently; the premise seems to be that SIE drives an impact and that impact is modulated by ENSO. However, a more plausible argument based on the literature and our prior understanding is that, to first order, ENSO modulates both SIE and the many global extremes indices listed. The ENSO-Sea Ice literature is insufficiently cited.
I also have severe concerns about the literature review and mechanistic discussion. First, and most importantly, six of the references do not seem to exist. If I had realised this before reviewing the paper I would not have proceeded to review in such detail as it implies AI hallucination. These are listed below. Regardless of this, and giving the authors the benefit of the doubt, several statements are made which are not supported by the cited literature, and the literature which is cited has many omissions. Please see my specific comments. Furthermore, there is a mechanistic discussion of the linkages when statistical significance is found, but these are written as if they assume causation. Even where literature exists, it is sometimes misrepresented (e.g. Arctic papers are cited but the Antarctic is being discussed) and many of the arguments are hurried. Conclusions in the mechanistic discussion are overly confident; it makes statements as fact, rather than plausible hypothesis, and the way the literature is cited is misleading and attempts to lend more support to the current hypothesis than exists. One example is lines 364-366: “This warming effect propagates through atmospheric circulation to mid- and low-latitudes, resulting in a 37% increase in global extreme heat frequency in 2024 compared to the 2010–2020 average (Jha et al., 2025; Ji et al., 2023)” (similar to d). I think the word ‘resulting’ implies that the previous authors examined the propagation of the heat, whereas they only presented the change in heat frequency. The mechanistic section is also rather long and could benefit from more focus.
In summary, this paper is not appropriate for publication. First, while the relationships presented between global extremes and Antarctic SIE are novel, the interpretation of statistical relationships based on so few data points requires extreme care and the authors have not done this. Second, the existing literature is substantially misrepresented, as outlined above. If the authors wish to re-frame and re-submit this manuscript, they need to revisit the literature carefully, including on polar-midlatitude linkages, with careful attention to which pole and which region hypotheses have been made for. I would then suggest selecting the indices for which strong physical arguments exist for a physical linkage and focussing on these, and considering carefully the evidence for the role of ENSO/SAM as a driver rather than a modulator. Narrowing the scope would enable the authors to add analysis to support their claims i.e. if it’s claimed that Antarctic SIE has driven the extremes, would a lag be expected? Would there be ways of identifying the ‘propagation of the warming effect’? They would also be able to concentrate on the mechanistic arguments for those plausible linkages.
REFERENCE LIST ISSUES
Ayres et al: incorrect DOI (correct DOI below)
Crosta et al: incorrect DOI
Cvijanovic, I., & Caldeira, K. (2015). Antarctic sea ice and climate sensitivity. Environmental Research Letters, 10(9), 094011. https://doi.org/10.1088/1748-9326/10/9/094011: DOI points to different article, named article does not seem to exist
Dou and Zhang: incorrect DOI
He et al: citation of AGU abstract, not appropriate
Hrudya et al: DOI points to different paper, named paper doesn’t seem to exist
Jeromsen et al: not relevant, not used in text
Ji et al: Global amplification….. DOI doesn’t work, paper doesn’t seem to exist
Komatsu et al, ‘Antarctic sea ice loss amplifies polar boundary layer convection…’: does not seem to exist
Osborne et al: not relevant, not used in text
Rantanen et al “Arctic amplification…” DOI does not work, article does not seem to exist
Zhang et al ‘SAM negative phase…’: DOI does not work, article does not seem to exist.
SPECIFIC COMMENTS
There is a stylistic choice which is confusing, name the use of quotation marks throughout: three examples are “small perturbations trigger large responses” L144; “stable-then-declining’ L157; “decreasing SIE leads to increased frequency of extreme events” L230. Quotation marks should generally be used only for actual quotations, and then referenced. From the style guide: “Quotations can also be used to denote an unfamiliar or newly coined term or phrase. They may also be used to introduce a term but only once at the first instance.” I do not think this generally applies here. Also from the style guide: “Italic font may be used for emphasis, although this should be used sparingly (e.g. data were almost consistent).” I suggest that the “stable-then-declining” could be highlighted in italics but otherwise all the text appearing in quotations should just be normal text.
The regional results are emphasised but do not seem to be used in the correlation analysis.
Papers which are cited (e.g. Cvijanovic, Rantanen, Ji) [where they exist] are a) rather a selective subset of a large body of literature on midlatitude linkages and b) cover the Arctic impacts on midlatitude extremes, not the Antarctic influence on global extremes. I do not know the literature exhaustively, but missing citations include the following, which may point to further papers
- https://doi.org/10.1175/JCLI-D-21-0918.1 Ayres et al, The Coupled Atmosphere-Ocean Response to Antarctic Sea Ice Loss
- https://doi.org/10.1175/JCLI-D-23-0524.1 Hay And Kushner (2024), The Relative Importance of Antarctic Sea Ice Loss within the Response to Greenhouse Warming
L16-17 and L46-47: statements are made about about decreases in SIC, it’s unclear if these are new results or cited from the literature, since they appear in the abstract and introduction
L55-63: This section conflates a) well-evidence links between regional extremes and sea ice loss and b) globally distributed extreme events. This is problematic.
L64-66: “Existing research confirms a link between Antarctic sea ice changes and global extreme weather events” this is not supported by the cited literature
L70-72: The paper points to Josey et al (2024) as a reference for western tropical Pacific temperatures and typhoon generation frequency but that paper does not cover these affects
L86: West Antarctic rainstorm: This is written with no explanation
L92: ‘overall negative correlation between sea ice and extreme climate events’: it’s not explained what this correlation is, and there is no paper cited to justify the statement that CMIP models represent it.
L120: “This data” what data?
L125: this equation makes it clear that the percentages cited throughout are indeed ‘percentage changes in SIE’ and not sea ice concentration changes. This is important and should be made clear.
L323: Cite ENSO -sea ice literature here.
Note that my comments on the mechanistic discussion (Section 3.3) are not exhaustive since by this point I had more substantial concerns about the manuscript.L526: “this study is the first to quantify the regulatory effect of the “El Niño + negative SAM phase”. Many papers consider the combined impact of ENSO and SAM on sea ice, which is relevant here. For example; https://doi.org/10.1175/JCLI-D-22-0679.1 Wang et al, The Impacts of Combined SAM and ENSO on Seasonal Antarctic Sea Ice Changes, and citations within on the combined SAM/ENSO impact on atmospheric circulation. The fourth paragraph of this paper is helpful.
Citation: https://doi.org/10.5194/egusphere-2025-6301-RC2 -
AC2: 'Reply on RC2', Tang Shulin, 19 Jul 2026
1 / 7
Author Response to Anonymous Referee #2
Manuscript: egusphere-2025-6301
Revised title: Antarctic Sea-Ice Anomalies and Global Hazard-Event Records: An Exploratory Analysis, 2010–2024
Dear Anonymous Referee #2,
We sincerely thank you for your careful, critical, and detailed assessment of our manuscript. Your comments identified fundamental problems in the original submission, including unsupported global-event claims, the short observational record, inadequate treatment of common climate drivers, weak justification of the exponential model, conflation of regional evidence with global event records, and serious reference-list problems. We have therefore undertaken a substantial rewrite rather than a limited textual revision.
The revised manuscript has been reframed as an exploratory observational analysis of statistical co-variation. The title, abstract, introduction, methods, results, conceptual framework, discussion, and conclusions have all been rewritten. The revised line numbers cited below refer to the clean revised manuscript. Our point-by-point responses follow the summary of major revisions.
Major revisions
• Causal interpretation removed. The manuscript now consistently uses terms such as statistical association, co-variation, synchronization and conceptual pathway, and explicitly states that the results do not establish direct causal attribution.
• Statistical analysis strengthened. The exponential curve is retained only as a descriptive comparison. Linear fitting is now the primary model-form check, supplemented by Pearson correlation, Spearman rank correlation, detrended-residual correlation, leave-one-out sensitivity analysis and Benjamini–Hochberg false-discovery-rate adjustment.
• Data object and scope clarified. The China National Climate Center dataset is described as a record of major hazardous weather and climate events, not a complete objective census of all global extremes. Regional SIC maps are used only as contextual illustrations, whereas basin-wide annual mean SIE is the statistical predictor.
• ENSO and SAM roles redefined. ENSO is treated as a first-order quantified climate background. SAM is discussed only as a literature-supported potential co-modulating factor because it is not independently included in the quantitative tests.
• Mechanistic discussion rebuilt. The former causal mechanism narrative and unsupported numerical claims have been removed. The new framework separates plausible physical pathways, event-level statistical association and common modulation, and explicitly identifies the analyses needed for future attribution.
• References re-audited. The problematic, unverifiable, irrelevant or misapplied references identified by Reviewer 2 have been corrected or removed, and Antarctic-specific literature recommended by the reviewer has been added.
2 / 7
Point-by-Point Response to Anonymous Referee #2
R2.1 Global-event claims exceed the available evidence
Reviewer comment: It is an extraordinary claim that the number of global, rather than regional or Southern Hemisphere, extremes might be modulated to a meaningful extent by Antarctic SIE, and the evidence presented is far insufficient.
Response: We agree that the original global-modulation claim was not supported. The revised manuscript no longer claims that Antarctic SIE controls or meaningfully modulates the global occurrence of extreme events. The statistical outcomes are now defined as the recorded annual frequencies of major hazardous weather and climate events under a consistent reporting scheme, rather than an objective census of all global extremes. The Introduction states that regional mechanisms must not be extrapolated into global causal claims and describes the global relationship as an open question. The seven categories are retained for transparent comparison, but the interpretation is explicitly hierarchical: extreme heat and cold waves show the more robust statistical associations; floods, rainstorms and typhoon-storm events show intermediate exploratory associations requiring cautious interpretation, with the detrended relationships for floods and rainstorms weakening markedly; tornadoes and blizzards are weak or unstable and are not used as primary physical evidence.
Revised manuscript: Introduction, Lines 36–61; data limitations, Lines 86–107; Results, Lines 219–226 and 227–287; Discussion, Lines 383–436.
R2.2 Exponential model and short sample
Reviewer comment: Section 3.2 uses correlations between global-extreme indices and SIE through an exponential regression model. There are only 14 years of data, and the justification for choosing an exponential relationship is weak.
Response: We agree that the exponential model cannot serve as the sole foundation for interpretation. The revised manuscript states explicitly that the dataset contains 15 annual observations, from 2010 through 2024 inclusive. The exponential fit is now used only as a descriptive curve for comparison with a possible nonlinear form and does not imply an assumed exponential climate response. We added a linear fit for every event category and use the linear-fit P values as the primary multiple-testing set. We further added Pearson correlation, Spearman correlation, detrended-residual correlation, leave-one-out sensitivity and FDR correction. Results are interpreted only as statistical associations.
Revised manuscript: Methods, Lines 120–139; Results, Lines 193–226; Table 1A–B, Lines 205–218.
R2.3 Correlation is not causation; ENSO may affect both SIE and event indices
Reviewer comment: Statements such as “the reduction in SIE is likely a significant driver” assume that SIE drives an impact and ENSO modulates it. A more plausible first-order argument is that ENSO modulates both SIE and many of the global extreme-event indices. The ENSO–sea-ice literature is insufficiently cited.
Response: We agree and removed the driver language. ENSO is now treated as a first-order quantified climate background that may co-vary with both Antarctic sea ice and event-generating conditions. SAM is treated only as a literature-supported potential co-modulating factor because it was not independently included in the quantitative tests. The revised framework therefore does not place sea ice at the beginning of a one-way causal chain. We also added relevant ENSO–SAM–Antarctic sea-ice literature, including Wang et al. (2023), Liu and Zhu (2024) and Wang et al. (2025). Because the 15-year annual record does not permit comprehensive multivariable causal identification, we state this limitation directly rather than claiming that ENSO or SAM has been statistically controlled.
3 / 7
Revised manuscript: Introduction, Lines 62–71; Data sources, Lines 92–97; Section 3.3.1, Lines 289–323; Conclusions, Lines 453–477.
R2.4 Literature review, reference validity and mechanistic overstatement
Reviewer comment: Several references do not appear to exist; other references are misrepresented or refer to the Arctic while the Antarctic is discussed. The mechanistic discussion assumes causation, makes unsupported statements as fact and is overly long. The statement that warming propagation “resulted” in a 37% increase in global extreme heat misrepresents the cited work.
Response: We acknowledge the seriousness of this concern. We conducted a reference-by-reference audit of the items identified by the reviewer. Unverifiable, irrelevant, inappropriate or misapplied references were removed; incorrect entries were corrected; and Antarctic-specific studies were added. The unsupported quantified statements concerning albedo change, radiative absorption, local warming rate, baroclinicity, jet displacement, event duration and the alleged 37% causal increase were deleted. The former mechanistic narrative was replaced by a shorter conceptual-pathway discussion. Each event category is now discussed in conditional language, with explicit attention to regional controls, shared climate backgrounds and the analyses required for future testing.
Revised manuscript: Introduction, Lines 54–71; Section 3.3, Lines 288–436; reference list, Lines 498–572.
R2.5 Narrowing the scope and focusing on physically plausible linkages
Reviewer comment: The manuscript should select indices for which strong physical arguments exist, focus on those linkages, and consider ENSO/SAM as possible drivers rather than merely modulators.
Response: We narrowed the interpretive scope rather than deleting the systematically compiled event categories. Retaining all seven categories enables a transparent comparison and shows that the evidence is not uniform. Only extreme heat and cold waves are presented as relatively strong statistical associations. Floods, rainstorms and typhoon-storm events are classified as intermediate exploratory associations that require cautious interpretation; the detrended relationships for floods and rainstorms weaken markedly. Tornadoes and blizzards are labelled weak or unstable and are not used as primary support for the conceptual framework. ENSO is now treated as a shared first-order climate background, and SAM is not described as a regulatory variable quantitatively identified by this study.
Revised manuscript: Results, Lines 219–287; Section 3.3.1, Lines 289–323; synthesis, Lines 426–436.
R2.6 Lag expectation and identification of propagation
Reviewer comment: If Antarctic SIE is claimed to have driven the extremes, would a lag be expected? Are there ways to identify the propagation of the warming effect?
Response: The revised manuscript no longer claims that SIE drove the recorded events. We did not add a lagged-regression claim because the annual sample contains only 15 observations; applying one- or two-year lags would further reduce the effective sample and multiply the number of tests. Instead, we explicitly identify lagged-correlation diagnostics, wave-activity-flux analysis, reanalysis-based circulation diagnosis and numerical sensitivity experiments as necessary next steps for testing timing and propagation. This limitation is stated in the abstract, the conceptual framework, the extreme-heat discussion and the conclusions.
Revised manuscript: Abstract, Lines 21–24; framework, Lines 319–323; extreme-heat discussion, Lines 351–355; Conclusions, Lines 470–477.
4 / 7
Reference-list issues raised by Reviewer 2
R2.R1 Ayres et al.: incorrect DOI.
Response: Corrected. The revised reference is Ayres, Screen, Blockley and Bracegirdle (2022), Journal of Climate, 35, 4665–4685, DOI 10.1175/JCLI-D-21-0918.1. It is used to support the coupled atmosphere–ocean response to Antarctic sea-ice loss and its possible links with Southern Hemisphere circulation adjustment.
Revised manuscript: Introduction, Lines 28–30 and 54–61; References, Lines 499–500.
R2.R2 Crosta et al.: incorrect DOI.
Response: Removed. The associated claim concerning marginal-zone uncertainty was not essential to the revised analysis and has been deleted.
Revised manuscript: The former Crosta citation and related statement no longer appear.
R2.R3 Cvijanovic and Caldeira (2015): DOI points to another article and the named article appears not to exist.
Response: Removed in full. The revised manuscript does not use this reference or the direct-driver statement it was cited to support.
Revised manuscript: The former citation in Section 3.2 has been deleted; revised Section 3.2 begins at Lines 175–178 with exploratory wording.
R2.R4 Dou and Zhang: incorrect DOI.
Response: Removed. The ENSO–sea-ice discussion has been rewritten using verifiable Antarctic-specific literature.
Revised manuscript: Revised ENSO discussion, Lines 62–71 and 289–306.
R2.R5 He et al.: AGU abstract is not an appropriate reference.
Response: Removed. No conference abstract is used to support the revised mechanism discussion.
Revised manuscript: The former citation no longer appears.
R2.R6 Hrudya et al.: DOI points to another paper and the named paper appears not to exist.
Response: Removed in full.
Revised manuscript: The reference and all associated claims have been deleted.
R2.R7 Jeromsen et al.: irrelevant and not used in the text.
Response: Removed from the reference list.
Revised manuscript: The reference no longer appears.
R2.R8 Ji et al.: DOI is invalid and the paper appears not to exist.
Response: Removed. The sentence claiming that atmospheric propagation resulted in a 37% increase in global extreme-heat frequency has also been deleted.
Revised manuscript: Revised extreme-heat discussion, Lines 331–355.
5 / 7
R2.R9 Komatsu et al., “Antarctic sea ice loss amplifies polar boundary layer convection…”: appears not to exist.
Response: Removed in full. The revised oceanic pathway is supported only by references retained after the audit.
Revised manuscript: Revised cold-wave and Southern Ocean discussion, Lines 364–382.
R2.R10 Osborne et al.: irrelevant or unused.
Response: Removed. The exponential model is no longer justified by a general optimisation reference; it is explicitly described as a descriptive comparison rather than a mechanistic model.
Revised manuscript: Methods, Lines 125–139.
R2.R11 Rantanen et al., “Arctic amplification…”: DOI invalid or misapplied.
Response: Removed. The unsupported quantitative statements about a 15% reduction in baroclinicity and a specific jet-displacement amplitude have also been deleted.
Revised manuscript: Revised extreme-heat discussion, Lines 331–355.
R2.R12 Zhang et al., “SAM negative phase…”: DOI invalid and the paper appears not to exist.
Response: Removed. The revised manuscript cites verified, directly relevant studies for SAM, Antarctic sea ice and Southern Hemisphere circulation, including Wang et al. (2023), Eabry et al. (2024), Schroeter et al. (2023), and Zhang et al. (2024) where appropriate.
Revised manuscript: Introduction, Lines 62–71; Section 3.3, Lines 289–306 and 364–436; References, Lines 521–523, 552–553, 559–560 and 567–569.
Specific comments from Reviewer 2
R2.S1 Use of quotation marks for emphasis
Response: The three examples identified by the reviewer—“small perturbations trigger large responses”, “stable-then-declining”, and “decreasing SIE leads to increased frequency of extreme events”—have been removed or rewritten as normal prose. Quotation marks are no longer used as a general emphasis device.
Revised manuscript: Methods, Lines 120–139; Results, Lines 149–169 and 175–226.
R2.S2 Regional results are emphasized but not used in the correlation analysis
Response: We now state explicitly that SIC maps are illustrative regional background only, whereas annual mean basin-wide SIE is used in the statistical analysis. No regional SIC result is presented as direct evidence for global-event attribution.
Revised manuscript: Data sources, Lines 81–85; Results, Lines 159–169; Figure 2 caption, Lines 171–174; Conclusions, Lines 438–445.
R2.S3 Selective and Arctic-focused literature; missing Antarctic studies
Response: The Arctic or misapplied references have been removed. We added the two Antarctic studies recommended by the reviewer: Ayres et al. (2022) and Hay and Kushner (2024), together with additional Antarctic-specific literature. The revised Introduction distinguishes evidence for regional Southern Hemisphere responses from unsupported global causal generalization.
Revised manuscript: Introduction, Lines 47–61; References, Lines 499–500 and 529–530.
6 / 7
R2.S4 Original Lines 16–17 and 46–47: unclear whether SIC decreases are new results or literature statements
Response: The revised manuscript separates these functions. The abstract reports the study’s representative spatial result. The Introduction cites previous studies for regional Antarctic sea-ice variability. The Methods state that SIE is used for interannual correlation analysis and SIC only for representative spatial illustration.
Revised manuscript: Abstract, Lines 13–18; Introduction, Lines 31–42; Data sources, Lines 81–85.
R2.S5 Original Lines 55–63: conflation of regional links and globally distributed events
Response: The revised Introduction now separates established or plausible regional Southern Hemisphere responses from the unresolved question of global recorded-event frequency. It explicitly warns against extrapolating local mechanisms into global causal claims.
Revised manuscript: Introduction, Lines 36–61.
R2.S6 Original Lines 64–66: unsupported statement that existing research confirms a link with global extreme weather
Response: The statement has been removed. It is replaced by language stating that the global recorded-event relationship remains an open question and that current evidence mainly supports potential teleconnections or physical links to specific regional extremes.
Revised manuscript: Introduction, Lines 43–61.
R2.S7 Original Lines 70–72: Josey et al. (2024) does not support western tropical Pacific temperature and typhoon-frequency claims
Response: The unsupported typhoon claim and numerical percentage have been deleted. Josey et al. (2024) is now used only to support the anomalously low Antarctic sea ice in 2023 and the associated regional sea-ice background; it is no longer cited for western tropical Pacific SST or typhoon-generation frequency.
Revised manuscript: Introduction, Lines 36–40; Results, Lines 155–158; revised reference entry, Lines 533–534.
R2.S8 Original Line 86: unexplained “West Antarctic rainstorm”
Response: This phrase and the associated unsupported regional claim have been removed.
Revised manuscript: The revised aims at Lines 72–79 contain no such claim.
R2.S9 Original Line 92: unexplained negative correlation and unsupported claim about CMIP models
Response: The statement has been removed. The revised text no longer asserts that CMIP models reproduce a general negative correlation between Antarctic sea ice and global extremes.
Revised manuscript: Introduction and aims, Lines 43–79.
R2.S10 Original Line 120: ambiguous “This data”
Response: The data sources and statistical object are now named explicitly: NSIDC SIE/SIC, China National Climate Center monthly major-event records, and NOAA/NCEP NEPI and NCPI indices. A dedicated subsection explains event aggregation and reporting limitations.
Revised manuscript: Data sources and event compilation, Lines 81–107.
7 / 7
R2.S11 Original Line 125: percentages are SIE changes, not SIC changes
Response: Clarified. The equations now define SIEDN as the daily SIE anomaly percentage relative to the 1981–2010 daily SIE median, and SIEYN as the annual mean SIE anomaly percentage. SIC is separately described as a spatial illustrative field.
Revised manuscript: Data sources, Lines 81–85; equations and definitions, Lines 108–119; Table 1 note, Lines 208–218.
R2.S12 Original Line 323: ENSO–sea-ice literature should be cited
Response: Added and integrated. The revised manuscript cites Wang et al. (2023) for combined ENSO–SAM effects on seasonal Antarctic sea ice, Wang et al. (2025) for the 2021–2023 three-year La Niña background, and Liu and Zhu (2024) for a modelled Antarctic sea-ice–ENSO pathway.
Revised manuscript: Introduction, Lines 62–71; Section 3.3.1, Lines 289–306; References, Lines 538–539 and 559–563.
R2.S13 Mechanistic discussion comments were not exhaustive
Response: We treated this as a request for comprehensive reconstruction rather than isolated edits. Section 3.3 has been rewritten throughout. Unsupported quantitative statements and deterministic causal wording from the original manuscript have been removed, regional controls are specified for each event category, and the framework is explicitly presented as a conceptual interpretation.
Revised manuscript: Section 3.3, Lines 288–436; Figure 8 caption, Lines 357–363.
R2.S14 Original Line 526: unsupported claim of being the first to quantify the “El Niño + negative SAM” effect
Response: The novelty-priority claim has been deleted. The revised manuscript states that combined ENSO–SAM phases are already known to affect seasonal Antarctic sea ice and cites Wang et al. (2023). It also states repeatedly that SAM is not independently quantified in this study.
Revised manuscript: Introduction, Lines 62–71; Data sources, Lines 92–97; Section 3.3.1, Lines 298–306; Conclusions, Lines 453–460.
We thank you again for your detailed and critical assessment. Your comments led us to reframe the study, strengthen the statistical evaluation, rebuild the conceptual framework, and complete a reference-by-reference audit. We respectfully submit the revised manuscript for renewed consideration.
Sincerely, Shulin Tang and Yaozong Zhang
-
AC3: 'Comment on egusphere-2025-6301', Tang Shulin, 19 Jul 2026
Note for RC1
Dear Anonymous Referee #1,
Thank you for your careful and constructive assessment of our manuscript. In response to your comments, we have substantially revised the paper rather than making only limited textual changes. The revised manuscript is now framed as an exploratory observational analysis of statistical co-variation. Causal and predictive claims have been removed, the limitations arising from the short observational record and shared climate drivers are stated explicitly, additional robustness analyses have been included, and the former causal-mechanism narrative has been reformulated as a conceptual pathway framework.
Please find attached our detailed point-by-point response to all comments raised by Referee #1. The response document also identifies the corresponding locations of the revisions in the revised manuscript.
Sincerely,
Shulin Tang and Yaozong ZhangNote for RC2
Dear Anonymous Referee #2,
Thank you for your detailed and critical review. We have carefully addressed both the major scientific concerns and the specific comments raised in your report. The revised manuscript no longer claims that Antarctic sea-ice extent directly controls global extreme-event occurrence. The exponential model is retained only as a descriptive comparison, while linear fitting and additional robustness analyses have been added. We have also clarified the roles of ENSO and SAM, substantially narrowed the mechanistic interpretation, distinguished regional sea-ice evidence from global event records, and completed a reference-by-reference audit in which unverifiable, irrelevant, or misapplied references and unsupported numerical claims were corrected or removed.
Please find attached our detailed point-by-point response to all comments raised by Referee #2, including responses to the reference-list issues and the specific line-by-line comments.
Sincerely,
Shulin Tang and Yaozong ZhangCitation: https://doi.org/10.5194/egusphere-2025-6301-AC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 325 | 159 | 21 | 505 | 29 | 34 |
- HTML: 325
- PDF: 159
- XML: 21
- Total: 505
- BibTeX: 29
- EndNote: 34
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study examines Antarctic sea ice variability between 2010 and 2024, and it’s relationship to a vast number of global climate and weather processes, from storms and heat extremes to large scale systems such as the SAM and ENSO. The authors use a combination of varied sea ice data products, and statistical tests including an exponential regression model, to show the relationships. The authors find that alongside the previously observed phase changes in sea ice over the observed time period, there were extremely high correlations with extreme events globally. Results would be of interest to anyone researching the theoretical connections between sea ice and the global climate. Overall, the paper was well written with a relatively clear experiment design, however, the claims of the paper are overstated.
The manuscript presents evidence for statistical associations between declining Antarctic sea ice extent and increasing extreme weather events. However, the analyses do not appear sufficient to infer causality. In several sections, the interpretation of the regression results may overstate the strength of the conclusions, particularly given the short temporal record and the absence of a more comprehensive treatment of confounding climate drivers.
The paper has strong merit in recognising these relationships, I particularly like the schematic, and from my extensive understanding of how Antarctic sea ice may interact with the global climate, the mechanisms and relationships are not scientifically impossible, but it would need further analysis or reframing from more certain language around causality. Specifically, softening of casual language, further acknowledgment of the limitations of the correlation pathways, reframing of pathways to be more conceptual, making it clear that the study identifies patterns consistent with teleconnection mechanisms rather than proving their direct causality. The ways in which Antarctic sea ice change interacts with the global climate is an important research topic and this work adds value, but the paper is not scientifically sound enough for publication at this time.
If the paper can be turned into a more cautions, exploratory and correlation-based synthesis paper, then I suggest Revise and Resubmit, or major revisions, at the editor’s discretion.