the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Jul 2025
Drivers of Phytoplankton Bloom Interannual Variability in the Amundsen and Pine Island Polynyas
Abstract. The Amundsen Sea Embayment experiences both the highest ice shelf melt rates and the highest biological productivity in West Antarctica. Using 19 years of satellite data and modelling output, we investigated the long-term influence of environmental factors on the phytoplankton bloom in the Amundsen sea (ASP) and Pine Island polynyas (PIP). We tested the prevailing hypothesis that changes in ice shelf melt rate could drive interannual variability in the polynyas’ surface chlorophyll-a (chla) and Net Primary Productivity (NPP). We found that the interannual variability and long-term change in glacial meltwater may play an important role in chla variance in the ASP, but not for NPP. Glacial meltwater does not explain the variability in both chla and NPP in the PIP, where light and temperature are the main drivers. We attribute this to potentially greater amount of iron-enriched meltwater brought to the surface by the meltwater pump downstream of the PIP, and the coastal ocean circulation accumulating and transporting iron towards the ASP.
- Preprint
(5339 KB) - Metadata XML
-
Supplement
(3838 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-3149', Anonymous Referee #1, 18 Aug 2025
-
RC2: 'Comment on egusphere-2025-3149', Anonymous Referee #2, 02 Oct 2025
Review of Liniger et al.
This manuscript uses remote sensing data products, together with some model output, to investigate whether changes in glacial melt can account for variability in surface chlorophyll-a and net primary productivity in two Antarctic polynyas from 1998-2017 (Amundsen Sea Polynya and Pine Island Polynya). The authors report observational support for a positive relationship between surface chla and glacial melt (but no relationship with NPP) in ASP, and no significant relationship with either variable in PIP. Instead, local processes at PIP seem to impact the bloom phenology. The authors investigate and discuss plausible mechanisms that may be operating distinctly in each region.
The authors provide a speculative discussion on why the relationship between glacial melt and chla differs between the two polynyas and why chla and NPP appear decoupled in ASP. The methods and statistical analyses are appropriate, the manuscript is logically structured, and the figures are generally insightful. Ultimately, the study points towards interesting signals and empirical results and is therefore appropriate for publication. However, there are general concerns about the mechanistic interpretation of some of the signals, as well as queries about the quality and reliability of the chla data product in this region. There are a few further specific suggestions to improve the manuscript.
General comments
Discussion of Chla and NPP relationships
The decoupling of NPP and chla in the ASP is a central feature of the results but is not given much attention in the subsequent discussion. The explanation that it is due to the vertical mixing that may concurrently be promoted by glacial meltwater is somewhat unsatisfying. In the first instance, there is presumably a spatial separation between these processes – the meltwater plume will promote mixing at or near the glacier face, but to what extent is that enhanced mixing present across the rest of the polynya area? In contrast one might expect the meltwater to enhance stratification once it settles at a level of neutral buoyancy. Has this proposed enhancement of vertical mixing, and its spatial extent, been described elsewhere in the literature? Secondly, is your explanation that deeper mixed layers limit light availability and reduce NPP relative to chla consistent with the VPGM algorithm? In what way does that algorithm take mixed layer depth into account, and is it likely to capture variations in mixed layer depth due to glacial melt in this region? This is presumably testable by looking more closely at the chla:NPP relationship directly, rather than through the lens of their relationship with TVF.The claim related to the possible role of phytoplankton community composition needs to be described in greater detail.
Chla product and uncertainty
Ocean colour is influenced by absorption from pure water, dissolved compounds, phytoplankton, and suspended sediments. Globally tuned chla algorithms do not always perform optimally in optically complex waters. In the context of this study, glacial meltwater could impart an optical signature potentially affecting the accuracy of chla estimates. Could the authors comment on whether the chla algorithm used is expected to handle such conditions, and how confident they are in its performance in the study region? Some additional analysis looking at the uncertainty in the chla fields or comparison with other available chla products (that use different algorithms) would be useful to gauge the potential impact of additional optical influences on the results.What influence does the number of visible days in the region have on the results? Is there reason to be confident that, in this region, a fraction of the primary production is not missed prior to the return of sufficient light for ocean colour to be detected? See a couple of recent papers that have noted possible distinctions between what the satellite sees and what growth is taking place, both with respect to the solar angle and the sea ice cover (McLish and Bushinsky, 2023; Douglas et al., 2024).
Oceanographic context
The introduction could be strengthened by providing some background on the regional circulation and major water masses influencing the study area and how this differs between the two polynyas. Similarly, the description of the meltwater pump was somewhat lacking in detail. Expanding this section would help better frame and qualify the later discussion.Additionally, there is some confusing use of terminology that could be clarified. In particular, “ice shelf meltwater” and “glacial meltwater” seem to be used interchangeably throughout the manuscript, while “subglacial discharge” is used only once in the discussion but is not defined anywhere. Suggest using consistent terminology and provide a clear distinction between terms.
Figure presentation
Many of the maps are too small, leading to overlapping of labels, and obscuring of data with overlaid markings. Please make maps larger, especially for Figures 4 and 6.Data availability and reproducibility
Please make sure to provide all details needed to locate and access the versions of the data products used, rather than simply links to the general websites. DOI’s should be provided where available.Specific comments
L79: what does “natural” mean in this context?
L147-153: Given the NPP dataset is central to the main results of the manuscript, I suggest including a few lines on how NPP was derived in the model. This would also be useful for the later discussion.
L274: Why highlight the relationship between chla in ASP and TVFasp and Dotson ice shelf but not Crosson ice shelf?
L304-343: Table 1 and Figure 4 show opposing relationships between chla and TVF for ASP and PIP. While the relationship is positive for ASP it is negative for PIP (particularly for Cosgrove where the relationship is quite strong). How do you interpret this difference?
L333. Comments for Figure 4;
- Suggest using black crossing to indicate insignificant correlations rather than significant ones. At present, it is difficult to read the magnitude of the correlations because the colours are obscured by the black crosses.
- Consider repeating the labels from Figure 1 to guide the reader (at least for the ice shelves)
- Ensure longitudinal labels are legible and not overlapping
L421-422: “IRT and OWP are significantly related in the PIP.” Is this also true for ASP? Where is this relationship shown?
L443-444: Did you do any pretreatment of the data, e.g. mean centering and normalisation? Please specify or alternatively, argue for why you did not do this.
L454-456: The loadings (vectors) for OWP and IRT are very similar in ASP and PIP, both in their projections onto Dim1 and Dim2 and in their magnitudes. The main difference between the two polynyas with regards to OWP and IRT lies in the variance explained by Dim1. I suggest using this difference to support the statement that “...physical conditions might play a stronger structuring role…” rather than how they project on Dim1 and Dim2.
L457-458: which is in line with the earlier correlation analysis showing opposing relationships between chla and TFV between the two polynyas.
L476: Comment for Figure 7:
General: Please explain in more detail in the main text how to interpret this figure, and PCAs in general, for the uninitiated.
- Figure 7 and the accompanying text heavily relies on the use of acronyms. You might consider providing a legend next to panel b for ease of interpretation.
- Please ensure labels are not overlapping
L546: “settling depth” is unclear. Do you mean the depth of neutral buoyancy?
Minor/technical comments
L50: Reference Figure 1.
L276-277: Delete duplicate sentence.
L278-279: This has already been stated. Delete.
L446: Please make sure you define all acronyms. “BD” is currently not defined.
L454: As above, please define “BM”.
L498: change “… and the modelling…” to and models.
L519: delete “related”.
References
McClish, S., Bushinsky, S.M., 2023. Majority of Southern Ocean Seasonal Sea Ice Zone Bloom Net Community Production Precedes Total Ice Retreat. Geophysical Research Letters 50, e2023GL103459. https://doi.org/10.1029/2023GL103459
Douglas, C.C., Briggs, N., Brown, P., MacGilchrist, G., Naveira Garabato, A., 2024. Exploring the relationship between sea ice and phytoplankton growth in the Weddell Gyre using satellite and Argo float data. Ocean Science 20, 475–497. https://doi.org/10.5194/os-20-475-2024
Citation: https://doi.org/10.5194/egusphere-2025-3149-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 753 | 57 | 21 | 831 | 30 | 12 | 18 |
- HTML: 753
- PDF: 57
- XML: 21
- Total: 831
- Supplement: 30
- BibTeX: 12
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of "Drivers of Phytoplankton Bloom Interannual Variability in the Amundsen and Pine Island Polynyas" by Guillaume Liniger et al.
The manuscript presents a valuable study of the phytoplankton blooms and their drivers in the Amundsen and Pine Island Polynyas. Satellite-derived Chl-a and NPP maps were used to characterize phytoplankton abundance and primary productivity in the years 1998 – 2017. Overall, the manuscript is well-written, well-organized, and the main points are clearly articulated. The determination of phenology metrics followed the standard methodologies described in the literature, and the use of Principal Component Analysis (PCA) and the Mann-Kendall test demonstrates good statistical practices. I especially appreciate the application of non-parametric statistical metrics in this study.
I see, however, a few issues that when fixed, could improve the final version of the paper. I present them in the points below:
The study compares the Amundsen and Pine Island Polynyas, highlighting several differences that appear to arise from variations in local topography, sediment resuspension, and currents (as mentioned in lines 519-520). While these factors were discussed, they were overlooked in the study area's section. To improve the brief description and make it easier for readers to follow the discussion, it would be beneficial to add the following: (a) the bathymetry of the area, which is an important aspect in the analysis of sediment resuspension; (b) contrasts between the polynyas regarding circulation patterns; (c) a brief description of the glaciers analyzed in the study with a particular focus on the differences between them; and (d) information on phytoplankton composition, which determines the demand for nutrients, sensitivity to iron shortages (notably different for diatoms and small flagellates), and the potential for using recycled nutrients. Recent research has indicated changes in the phytoplankton community structure on Antarctica's shelf, including a decline in diatoms sensitive to iron shortages, so I would expect at least a brief characterization of these communities.
The role of ligands, which are mentioned late in the discussion, seems significant for the availability of iron to phytoplankton. Information in lines 568-570 seems to suggest a possible feedback loop between the biological activity, ligands and the bioavailability of iron, which could be an interesting aspect to consider when analyzing bloom cycles. It might be worth adding a short comment on this topic in the model description around line 190.
There are at least two GlobColour L3 chl-a products that differ by the averaging method. It would be helpful to provide an ID or DOI number for the dataset. On a similar note, it would be interesting to see a discussion on the strong connection between net primary production (NPP) and chl-a, as chl-a is a key parameter for estimating NPP.
Under high-nitrate low-iron conditions, literature reported significant variations in the carbon-to-chlorophyll (Cphyto:Chl) ratio from those assumed globally, due to phytoplankton adaptations to iron shortages. Additionally, low-light conditions can alter the carbon-to-chlorophyll ratio. It would be worth including these elements in the discussion as a potential source of uncertainty. Might these differences explain the significant correlation with chl-a in Figure 3 and the lack of correlation with NPP at the same time?
Lastly, a small editorial note: lines 274-277 contain a repeated sentence.