the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 14 Nov 2025
The variability of Antarctic fast ice extent related to tropical sea surface temperature anomalies
Abstract. While numerous studies have examined the influence of meteorological variables on fast ice, the mechanistic linkages between fast-ice variability and large-scale climatic oscillations have remained inadequately explored. Empirical Orthogonal Function (EOF) analysis is applied to circumpolar Antarctic fast-ice extent data (March 2000–February 2018) to investigate seasonal-scale teleconnections between fast ice anomalies and tropical sea surface temperature (SST) variability. Results demonstrate a fluctuating but increasing trend in fast-ice extent during austral winter and spring, with spatially predominant anomalies concentrated in the West Antarctic. A physical pathway linking tropical and subtropical SST anomalies to fast-ice variability is elucidated through multiscale interactions: SST anomalies modulate outgoing longwave radiation (OLR), subsequently perturbing the 200 hPa geopotential height field and triggering atmospheric Rossby wave trains. These planetary waves propagate from the tropics towards the Antarctic coastal zone, generating anomalies in the Southern Annular Mode (SAM) and other patterns of mean-sea-level pressure (MSLP) and surface wind field. These atmospheric adjustments directly regulate fast-ice mechanical formation/disintegration processes while indirectly influencing thermodynamic ice evolution through air temperature modifications. Tropical SST anomalies predominantly exciting planetary waves during austral autumn, whereas the subtropical South Pacific SST dipole mode emerges as the primary forcing mechanism during austral winter and spring. Seasonal variations in atmospheric forcing on fast ice are identified. By tracing how remote SST forcing propagates through atmospheric wave dynamics to influence regional fast-ice conditions, this study advances process-level understanding of tropical and subtropical impacts on the Antarctic.
- Preprint
(3260 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5304', Anonymous Referee #1, 02 Dec 2025
-
RC2: 'Comment on egusphere-2025-5304', Anonymous Referee #2, 06 Jan 2026
Reviewer comments:
Text in grey italics are direct words from the article.
General Comments:
This research establishes a teleconnection relationship between equatorial sea surface temperatures and Antarctic landfast ice using a principal component analysis. This article and the relationships established within would contribute to our understanding of the drivers of Antarctic landfast ice variability. However, this manuscript needs revisions, specifically with regards to the interpretation of the Empirical orthogonal functions and principal components and what the represent. I have outline specifics and other revisions necessary.
Figures – Pannels of figures need to be individually labelled i.e. Figure 1a such that in text referencing is clearer. EX. Line 121 left panels of Figure 1 should be Figure 1a/c/e/g. This should occur both on the figures and in the text referring to the figures.
Interpretation of PCs and EOFs are incorrect. Interpretations of both PCs and EOFs are referred to as trends in fast ice anomalies in time (PCs) and space (EOF). EOFs describe the spatial patterns of variability while the PCs indicate how strong the EOF is expressed each year. Adjustment to the interpretations of the PCs and EOFs is needed.
Specific Comments:
Line 31 – Fast ice is the short version. It is ok to use throughout the article but there should be a mention of the full “Landfast ice” somewhere.
Line 33 - The maximum width of Antarctic fast-ice zone is approximately 200 km, covering an area of up to 600,000 km2. This sentence is not clear as the 200 km is referencing a single width a while the 600,000 is referencing the seasonal maximum circumpolar summed value. Could be improved by adding specifics such as “Fast ice has been observed to extend 200 km from the Antarctic coastline. Fast ice reaches its annual maximum extent during October with an average extent of ~601,000 km2.”
Line 51 – I am unfamiliar with the term “fast ice shrisk” and could not find any reference to this. Possible typo. If not, please define this as it is not common vernacular.
Line 54/55 - The atmospheric forcing factors also influence the variability of Antarctic fast ice across time scales ranging from microscale to synoptic, regional, seasonal, and interannual. “regional” is not a time scale.
Line 55/56 - Missing a reference for wind speed and direction. Also, Leonard et al. 2021 which you cite above also says storms and cyclones affect fast ice variability.
Line 101-109 – Because a lot of the emphasis and frankly the results are reliant on the EOF analysis and the preparation of the data prior to the EOF analysis, it would be nice to have more details regarding how the fast ice data was prepared prior to the EOF analysis (EX. What resolution, you mentioned anomalies, any detrending, that sperate EOF analysis is done on each season). If the data was not detrended prior to EOF analysis EOF1 is potentially representative of the linear trend of fast ice extent during each season where if the PC has a positive trend longitudes within the EOF with a positive value had a similar positive trend in fast ice extent where negative eigen values represent a negative trend.
Line 114-117 – Since you define the month triplets and use them in Table 1 instead of winter, summer etc. the text and table should use the same reference scheme. i.e. just use MAM or austral summer.
Line 112-Section 3.1 – The left panels of Figure 1 are not representative of trends in fast ice anomalies. These are principal components and given the y axes are from -3 to 3 I believe the data was standardized prior to PCA. As stated elsewhere, information regarding the data preparation prior to EOF analysis would be key to understand what the result mean. This makes the PCs unitless. Same with the EOFs which are dimensionless. In this analysis the EOF describe the where and the relative effect of the PC is expressed (The special pattern of the variability) while the PC described the strength of the EOF expression interannually. The interpretation of these data and figures needs to be changed to accurately represent the figures.
Section 3.1 – The along the front of the Ross ice shelf there is very little fast ice. Interpretation of variability in the Ross Sea should be done with caution.
Line 152-155 – It is unclear how the data shown in Figure 3 was calculated. Please explain the calculation, indicate what the total linear trend is for each season, and how these values vary across longitudes.
Section 3.2 – This section contains a lot of information. The information needs to be reduced to the important information relevant to the take home message. Or it needs to be organized differently such that the reader understand which variable is important where and during which season. At present all the information is there but is not easily trackable.
Figure 7-9b – What is the area defined as the fast ice region. The regression was done over multiple years I assume and the fast ice edge varies between years.
Figure 7-9b – the left y-axis is not latitude; it is the PC value.
Technical Corrections:
Line 46 – on line 45 you say “Secondly” then the next sentence is “Second”. I assume the “Second” should be “Third”. This would cause the “Third” on line 48 to be “Fourth”.
Line 110 – Double period to end the sentence.
Line 116 – asutral winter should be austral winter.
Citation: https://doi.org/10.5194/egusphere-2025-5304-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 260 | 94 | 29 | 383 | 25 | 21 |
- HTML: 260
- PDF: 94
- XML: 29
- Total: 383
- BibTeX: 25
- EndNote: 21
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This research paper explores the global mechanisms that lead to variations in landfast ice around the Antarctic continent. This is an interesting submission and one which could make a strong contribution, but one which also needs revision. I outline below how such a revision should be undertaken.
As a broad comment, the focus of the investigation is on the variability breakdown via EOF analysis. As some background at an early stage in the paper, it would be very valuable to preset time series of the TOTAL fast ice area for the four seasons and their trends.
Lines 43-46: In these opening lines in connection with the importance of the study it would be beneficial to reference the very recent studies of …
Luke G. Bennetts and Nathan J. Teder, 2025: Fast ice: The last line of defence for weakened Antarctic ice shelves. Nature Geoscience, 18, 574-575, and
Nathan J. Teder, Luke G. Bennetts, Phillip A. Reid, Robert A. Massom, Jordan P. A. Pitt, Theodore A. Scambos and Alexander D. Fraser, 2025: Large-scale ice-shelf calving events follow prolonged amplifications in flexure. Nat. Geosci., 18, 599-606.
Line 50: The year of publication is 2023, not 2024.
Line 51: ‘shrink’ (sp.)
Line 53: The text here cites ‘Crocker & Wadhams, 1989a’ (and also below). However, no related ‘1989b’ is cited or in References. From the context here I suspect the authors may have wanted to also cite the paper …
Crocker GB, Wadhams P (1989) Modelling Antarctic fast-ice growth. J. Glaciol. 35: 3-8 doi: 10.3189/002214389793701590.
If this is correct please to make appropriate changes.
Line 56: Paper is missing from References. Maybe you are referring to Shuki Ushio’s 2006 paper ‘Factors affecting fast-ice break-up frequency in Lützow-Holm Bay, Antarctica.’ Annals of Glaciology, 44, 177-182, doi: 10.3189/172756406781811835.
Lines 72-74: On this important point, include references to ….
Pezza, A.B. et al., 2012: Climate links and recent extremes in Antarctic sea ice, high-latitude cyclones, Southern Annular Mode and ENSO. Climate Dyn., 38, 57-73, doi: 10.1007/s00382-011-1044-y,
Irving, D. & co-authors 2016. 'A new method for identifying the Pacific-South American pattern and its influence on regional climate variability', J. Climate 29, 6109–6125.
Line 74: Another paper missing from the References! I imagine that ‘Clem et al., 2017’ is referring to
Kyle R. Clem, James A. Renwick and James McGregor, 2017: Large-scale forcing of the Amundsen Sea low and its influence on sea ice and West Antarctic temperature. Journal of Climate, 30, 8405-8424, doi: 10.1175/JCLI-D-16-0891.1.
Lines 121-124: Strongly suggest discussing the EOFs first, and then the PCs. The EOFs are the more fundamental aspect. Similarly, flip the two columns in Figs. 1 & 2.
It is important to note that the sign of EOFs is arbitrary. If the sign is changed so does the sign of the PC and, e.g., ‘positive trends’ become ‘negative trends’. Broadly speaking the four EOF1s have similar structures, but also large differences in the key regions around 180 deg and from -90 deg to -10 deg. I presume the sign of the patterns was taken because they ‘look similar’, but it would be of interest to calculate the spatial cross-correlations of the four patterns. This would maybe be a path to make a more physically meaning statement about the relative changes.
Lines 130-…: Related to the point above, the EOFs essentially identify the regions of (high) variability, rather than ‘positive anomalies’ etc. Please word these passages more appropriately.
Also, at line 130 (and 139, …) change ‘significant’ to something like ‘sizeable’. Only use the former word when referring to results of a statistical test of the null hypothesis.
Line 144-145: Here the text speaks of the lack of statistical significance of the trends of the PC2 series. Similar test should be conducted for the PC1 series. One might guess that the winter and spring trends could be significant, placing any discussion on this on firmer ground.
Lines 152-155: I am not sure that Fig. 3 is particularly useful in the paper, especially at the trends in PC2s are not significant (and a similar comment probably applies to most/all of the PC1s). Suggest deleting it.
Line 167-169: The SST structure in MAM is suggestive of ENSO. But note that in the far eastern Pacific the significant regressions move off the Equator, and into the NH. The other three seasons in Fig. 4 show only scattered ‘significant’ areas, and probably less that the 5% of the globe that you would expect by chance. The regions of significance around Antarctica in JJA and SON are regions covered by sea ice. What does ‘SST’ mean here?!
Interesting to note that the eastern Pacific SSTA Shigeru Aoki in the 2017 paper (lines 77-79 above) showed strong ENSO structure with greater anomalies just to the SOUTH of the Equator in April and December (Figure 4 in the paper). (Also see the comment made in the paper at line 292-294.)
Lines 176-179: The role of the ‘Amundsen Sea Low’ here is very interesting. Perhaps emphasise the points being made here by referencing study of Fogt, R. L., A. J. Wovrosh, et al., 2012 - (The characteristic variability and connection to the underlying synoptic activity of the Amundsen-Bellingshausen Seas Low. J. of Geophysical Research, 117, D07111, doi: 10.1029/2011JD017337) and commenting that the region is strongly tied to the location of the maximum cyclone system density and minimum cyclone central pressures, and is tied up with the SAM.
Lines 183-185: Figure 7a is not mentioned or referred to in the paper. However, its very interesting structure shows a strong wavenumber three. This mode has a large influence on subantarctic conditions (e.g., Irving & co-authors 2015: A novel approach to diagnosing Southern Hemisphere planetary wave activity and its influence on regional climate variability. J. Clim., 28, 9041-9057). Some additional text on this aspect would be very valuable.
Line 309: Unbalanced parentheses.