This seems like a very unique and powerful dataset. I am not an expert in phytoplankton taxonomy or pigment analysis, so I cannot review the methods of those approaches, but i think a potential user could do that due diligence. I have some minor technical comments to be corrected for clarity and accuracy.

Line 9: A missing word in here: “Weekly sampling resolution allowed to address….”

Figure 1: the black station labels are difficult to see in the deep depths in blue. Can you change the color of the sampling stations?

Figure 2: It is not that easy to discern the differences in the sizes of the circles to indicate the # of stations. If the authors feel this is an important point to make, perhaps they could use different colors to indicate the # of stations visited.

Line 111: The N species have superscripted numbers, but they should be subscripts

Line 125: Perhaps the authors could include here how long the samples were typically held before being analyzed in the laboratory

Line 139: “manufacture’s” should be “manufacturer’s”

Figure 11: The words and labels on this figure are probably too small

Line 283: I am not sure what “relaying on chlorophyll a concentrations” means

Line 298: This first sentence needs to be edited. “enables to differentiate” is incorrect grammar.

Figure 14 caption. For the last sentence, I would say “All surface stations were grouped by season.” (also add a period at the end of the sentence.

Line 337: make it “coast-offshore” (i.e., delete the space). But later (line 338), it is “coastal-offshore”. Be consistent.

Figure 16: The words and labels on this figure are probably too small

General comments

This manuscript presents the DYPHYRAD dataset, a multi‑year, high‑frequency time series of phytoplankton and environmental variables from the Eastern English Channel (Strait of Dover). The dataset itself is clearly valuable: weekly sampling at approximately kilometre scale in a very dynamic coastal system, with a rare, consistent approach across the full-size range of the phytoplankton community. This dataset clearly addresses a critical need for long-term observations in a very dynamic and important region, filling major gaps in existing networks like REPHY and SOMLIT. The approach of using consistent methodology across the full phytoplankton size range over multiple years is particularly valuable for the community. However, as it is written now, the manuscript has some fundamental problems and does not yet meet ESSD standards

Major Comments

I recommend rejecting the paper for now, for two main reasons which are linked:

• Too much overlap with another published paper
• Not enough value beyond the data repository

The manuscript's primary issue is that it presents itself as an abbreviated or "light" version of the scientific analysis already published in Ocean Science (Hubert et al., 2025; https://doi.org/10.5194/os-21-679-2025). The Ocean Science paper tells the full scientific story: it provides the scientific context, analyzes long-term trends in detail, and offers an ecological interpretation. While this ESSD manuscript avoids extended interpretation, it mirrors the Ocean Science paper in scope and narrative, and therefore fails to stand on its own.

This lack of an independent purpose became particularly evident during my review. Most of the scientific questions I initially had regarding data patterns and potential interpretations were already fully addressed in the Ocean Science paper. This highlights the core problem: the current manuscript raises scientific questions that it cannot explore in depth, because that work is already published. As a result, it feels redundant.

Also, the manuscript must cite the Ocean Science paper. Not doing so is a major problem. This is not just about giving credit; it's essential so that other scientists can understand the full context. It lets people reading the data paper see how the data has been used, and it helps people reading the science paper find the detailed documentation they need.

The paper does not give enough of the deep methodological details, technical validation, and data descriptions that are the key parts of a true data paper. The purpose of ESSD is to publish papers on datasets so they can be easily found, understood, and reused. This requires much more detail than what you find in a simple data repository. As it is now, the manuscript is missing this detail. For example:

• Methods: It says that different Cytosense instruments were used over the 11 years but gives no information on how they were compared or intercalibrated. This is absolutely critical to know if we are to trust a long-term time series.
• Data Processing: There is no description of the steps taken to process the data, the quality control flags used, how uncertainty was estimated, or the specific gating strategies for the flow cytometry data.
• Validation: There is no technical validation comparing different measurement approaches (e.g., Fluoroprobe chlorophyll estimates vs. extracted pigments, flow cytometry abundance vs. microscopy counts, or sensor performance across different environmental conditions).

Without this information, it is very difficult for another researcher to trust and reuse the data. The paper basically just says "this dataset exists," but the SEANOE repository already tells us that.

While the dataset is excellent, this manuscript doesn't justify being a separate paper yet. I believe that with a major rethinking and rewriting, it could become a suitable submission. This requires a shift in perspective: from "how we used this data" to "how you can use this data." I see two possible paths for the authors.

Option 1: Reformat as a "Pure" Data Paper

The manuscript must be expanded to include the precise, practical information a future user would need to assess quality, understand limitations, and trust the data for their own projects. This includes:

• Expand the Methods and Data Description:
  • Sampling: Station coordinates, sampling protocols, depth sampling…
  • Instrumentation: a detailed report on the intercalibration between the different instruments used (i.e., CytoSense).
  • Data Processing: A complete workflow from the raw files to the final data, including software used, code/scripts (if possible), the QC steps, and a full description of the quality flags applied to the data.
• Add a "Data and Potential Applications" section: Instead of a scientific discussion, this section should briefly mention the results of Hubert et al. (2025) as one example of how the data can be used. Then, it can suggest other potential uses (e.g., for satellite algorithm validation, data assimilation in models, or studies of short-term events).
• Reference Hubert et al. (2025): The authors must cite their Ocean Science paper in the introduction as the first scientific paper to come from this dataset.

ESSD exemplars you can follow:

• Acri et al. (2020): A long-term (1965–2015) ecological marine database from the LTER-Italy Northern Adriatic Sea site: plankton and oceanographic observations; 5194/essd-12-215-2020
• Lefebvre and Devreker (2023): How to learn more about hydrological conditions and phytoplankton dynamics and diversity in the eastern English Channel and the Southern Bight of the North Sea: the Suivi Régional des Nutriments data set (1992–2021); 10.5194/essd-15-1077-2023
• Sosinski et al. (2025): Long-term monitoring of hydrological dynamics and phytoplankton biomass indicator in three shellfish ecosystems of the English channel (2000–2024); 10.5194/essd-17-4737-2025

Option 2: Create a Different "Hybrid" Paper

If the authors want to keep a scientific story, it must be completely different from the one in Ocean Science. It would need a new scientific question and a different analysis. For example, they could focus on a methodological study, like a detailed comparison of how the high-frequency sensors performed in this turbid coastal water, or an analysis of short-term (daily, tidal) changes, which is not what they did in the Ocean Science paper.

Conclusion

I recommend rejection of the manuscript in its current form because its identity as a "light version" of existing research does not align with the core mission of ESSD.

However, the dataset is of high quality and importance. I would strongly encourage the authors to undertake a significant reconceptualization. By transforming this manuscript into a pure, technically detailed data descriptor that complements rather than condenses their Ocean Science paper, they could make a valuable and welcome contribution to the data community. I would be very supportive of a resubmission that follows this path.

---

Specific Comments

• I really appreciate the authors' transparency about using different instruments over the years. This is very important information for anyone who wants to use the data.
• Phaeocystis globosa is a very important species in this region. It would be really useful if the authors could add a short section that focuses specifically on its dynamics using this dataset.
• The statistical reporting in the text is inconsistent. Sometimes the mean is reported (often without a measure of variance like SD, SE, or CI), other times the mean with min/max, and elsewhere the median. A consistent statistical reporting style (e.g., median [min – max or IQR] or mean ± CI) must be adopted throughout the manuscript. See this article https://www.nature.com/articles/nmeth.2659
• It needs to be much clearer that all the data is from the sub-surface This is a critical detail. I suggest mentioning this in the Abstract and making it very clear in the Methods section.
• All the figures are too small, and it is very hard to read the axis labels. Please make them much bigger.
• Please consider using different color palettes for different variables to make the plots easier to understand.
• For the plots showing the transect, please add a label on the axis to show which side is "inshore" and which is "offshore."
• I find that showing both the time-series plots (like Fig. 7) and the boxplots (like Fig. 8) for the same data is a bit repetitive. Maybe you could choose just one format to make the paper clearer?

L18: “Berglund et al., 2007”

Please consider citing a more general, foundational reference for this statement, such as a paper by Falkowski  Falkowski, P. G. and Raven, J. A. (2007) Aquatic Photosynthesis. Princeton University Press, Princeton.

L28: “Phaeocystis globosa”

Please add the authority to the species name: Phaeocystis globosa Scherffel.

L30: “ (Lancelot et al., 1994, 1998; Medlin and Zingone, 2007; Li et al., 2022).”

To improve conciseness, consider replacing these multiple citations with a single, comprehensive recent review, such as Phaeocystis: A Global Enigma; 10.1146/annurev-marine-022223-025031.

L30-33: ...these blooms are considered undesirable due to their tendency to form dense gelatinous colonies...

It would be relevant to mention that this species is listed as a Harmful Algal Bloom (HAB) species in the IOC UNESCO database (https://www.marinespecies.org/hab/aphia.php?p=taxdetails&id=160538).

L34: ” (Peperzak and vanWezel, 2023).”

Please consider starting a new paragraph here.

L85: “Sampling Strategy”

The 'Sampling Strategy' section lacks information on the vertical dimension. Figure 2 suggests some variability in sampling depth, which could lead to vertical variability in the data. This needs to be addressed.

L91: “sub-surface sampling”

Please define 'sub-surface sampling' more precisely (e.g., specific depth or depth range).

L96: “from surface water samples”

Does 'surface water' correspond to a specific depth (e.g., ~1 m)? Given the riverine inputs, some stratification is expected. Please clarify how depth variations were handled.

L99-101: “high-frequency dynamics during periods of particular interest”.

Please specify what made these "periods of particular interest." For example, were they during expected blooms?

L106: “Sub-surface values were calculated as the average of measurements taken between 1 and 2 meters deep”.

By averaging data from 1-2m, do you risk catching only the influence of river plumes, especially in inshore areas?

L111: nutrients (NO2−, NO3−, SiO2 and H3PO4)

I agree with Reviewer #1’s comment on this.

L111: the sub-surface

Again, does 'sub-surface' always mean 1m? Was sampling always at the same depth? This needs to be clarified.

L112-113: The samples were transferred to sterile 50 mL PVC flasks, then analyzed...

Were the nutrient samples filtered? If so, what was the filter type and pore size? If not, the data could be influenced by intracellular nutrients, especially during blooms. Please clarify this critical step. Also, were the analyses performed in triplicate?

L140-141: The instrument can also deliver total chlorophyll a estimates...

Was an inter-comparison performed to check the agreement between the different methods of estimating total Chl-a?

L142: Only four spectral groups can be addressed at a time.

Why is this? Is it a limitation of the instrument? Please clarify.

L169-170: (FWS and SWS).

These acronyms are not defined.

L177-180: OraPicoProk or Synechococcus-like cells, RedPico or picoeukaryotes... These acronyms are a bit difficult to follow. Please consider using more intuitive names if possible.

L187-188: ...the following four instruments were used...

Were any intercalibration exercises performed to ensure data comparability between the four different Cytosense instruments?

L190: 3.5 Quality control

This section is too brief. The manuscript is missing details on data processing. Please specify the software used (e.g., R, Python) and consider making the analysis scripts available in the data repository to ensure full reproducibility.

L191: Only data with a good quality code

What constitutes a "good quality code"? Please explain the criteria.

L191: according to Argo quality control (www.somlit.fr/codes-qualite/; Wong et al., 2019).

The linked website is in French. Please provide a brief summary of the quality control procedures in English in the manuscript.

L201: mean: 8.43 °C,

When reporting a mean, please also provide a measure of variance (e.g., SD, SE, or CI). Also, please clarify why the mean is used here while the median is used elsewhere.

L211: Over these 11 years, SST have increased, particularly in winter.

Is this observation supported by a statistical test (e.g., a linear regression)?

L241: which explains the dynamics

"Explains" is a very strong word. It would be more accurate to say "is consistent with."

L244-247: During this period, phosphorus (P), silicate (Si), and nitrogen (N) are sequentially depleted...

This is a very interesting point and should be expanded. Is this sequence of nutrient depletion (P -> Si -> N) consistent every year? Changes in this pattern could explain some of the phytoplankton community dynamics. A plot of nutrient ratios (e.g., Si:N vs N:P) could illustrate this very effectively.

L249: The years with the highest median

Why switch to the median here, when the mean was used previously? Please be consistent.

L258-259: Winter silicate concentration decreased until 2017 and then increased again...

Is there a reason for this trend? Does it correlate with any changes observed in the phytoplankton community in your dataset?

L267-269: Stations nearest the coast had highest values. Only years 2013 and 2021 showed a spring bloom that spreads across the entire transect.

What might explain this? Do these stations/years have the highest nutrient concentrations?

L274-276: ...phytoplankton growth is limited by reduced light availability

What is the role of temperature in limiting winter growth? This should also be considered.

L289-291: The timing and intensity of spring blooms can be influenced by limited light availability or competition with diatoms.

Your dataset can be used to test this hypothesis. Is there evidence in your PAR data for light limitation? Do your phytoplankton data show that diatoms bloom before Phaeocystis? It would be stronger to support this statement with your own data rather than only a citation.

L297: 4.2.2 Phytoplankton spectral (pigmentary) groups...

It would be helpful to start this section with a brief overview of the typical phytoplankton community composition in this region.

L302-304: Phaeocystis was the main Haptophyte in our study area...

Was this confirmed by the parallel microscopy work mentioned in the acknowledgements (i.e., Skouroliakou et al., 2022, 2024)?

L309-310: The start of the bloom for “brown algae” (mainly diatoms here) is in winter,

A diatom bloom starting in winter is unusual. Please specify the timing (e.g., early or late winter). Also, the manuscript should explicitly define the months corresponding to each season (winter, spring, etc.).

L320-323: It represented between 28 and 90 % of the total chlorophyll a...

This is a very wide range. Can the authors use their data to explain the conditions that lead to either low or high dominance?

L344-349: ...RedNano dominated the community by almost 90 % corresponded mostly to the well-known P. globosa bloom. Flow cytometry can discriminate its different life stages. but not shown here.

Which fraction did RedNano dominate? The assertion that the abundant RedNano group is P. globosa is too strong without direct microscopic or genetic confirmation. Please rephrase this more cautiously (e.g., 'is consistent with' or 'likely represents'). The manuscript states that flow cytometry can discriminate Phaeocystis life stages but the data is 'not shown here'. This is a major missed opportunity. Presenting this data would significantly increase the impact of the dataset. Could the authors explain why this was omitted or consider including it?

Fig1: I agree with Reviewer #1; please modify the size and color of the points.

Fig2: I agree with Reviewer #1. The differences in point size should be more pronounced. Consider filling the points with a color gradient to better show the differences. The graph should also be wider. Alternatively, using geom_tile might be a good way to display this data.

Fig8 and other boxplot plots: For your information, you can easily control the order of the x-axis in R using factor(). This would avoid the need to put numbers in front of the season names.

Table A2: In Table A2, the number of NA values for some parameters is very high (e.g., >50% for CTD Fluorescence). Could the authors provide a brief explanation for